An Arena for Higher Powers Numen Book Series Studies in the History of Religions

Series Editors

Steven Engler (Mount Royal University, Calgary, Canada) Richard King (University of Kent, UK) Kocku von Stuckrad (University of Groningen, The Netherlands) Gerard Wiegers (University of Amsterdam, The Netherlands)

Volume 150

The titles published in this series are listed at brill.com/nus An Arena for Higher Powers

Ceremonial Buildings and Religious Strategies for Rulership in Late Iron Age Scandinavia

Olof Sundqvist

LEIDEN | BOSTON Cover illustration: Stone coffin, Sanda, Gotland. Photo: Bengt A. Lundberg. ATA, Riksantikvarieämbetet

Library of Congress Cataloging-in-Publication Data

Sundqvist, Olof. An arena for higher powers : ceremonial buildings and religious strategies for rulership in late Iron Age Scandinavia / by Olof Sundqvist. pages cm. — (Numen book series : studies in the history of religions : ISSN 0169-8834 ; Volume 150) Includes bibliographical references and index. ISBN 978-90-04-29270-3 (hardback : alk. paper) — ISBN 978-90-04-30748-3 (e-book : alk. paper) 1. Scandinavia—Religion. 2. Religion and politics—Scandinavia—History. 3. Scandinavia—Kings and rulers—Religious life. 4. Scandinavia—Antiquities. I. Title.

BL863.S85 2015 293—dc23

2015034775

issn 0169-8834 isbn 978-90-04-29270-3 (hardback) isbn 978-90-04-30748-3 (e-book)

Copyright 2016 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change.

This book is printed on acid-free paper. Contents

Preface ix List of Figures xi Abbreviations xiii

Part 1 Prolegomena

1 Introduction 3 1.1 Religious Ruler Ideology in Previous Research 7 1.2 Towards a Model for Analysing Religious Strategies for Rulership at Ceremonial Buildings 14 1.3 Approach and Methods 20 1.4 Outline and Hypotheses 29

Part 2 The Regional-Contextual Approach

2 The Three Areas of Investigation and their Political Leaders 35 2.1 The Kings in the Mälaren Area 35 2.2 The Earls in Trøndelag 46 2.3 The Chieftains of Iceland 53

3 Religious Strategies for Political Authorization—A Case Study 63 3.1 The Kings Called the Ynglingar 63 3.2 The Háleygja Kin (Hlaðajarlar) 80 3.3 The Þórsnesingar 86 3.4 A Contextual Interpretation 90

4 Ceremonial Buildings and Sanctuaries 95 4.1 The State of Research 95 4.2 The Problem of Terminology 104 4.3 Uppsala and other Sanctuaries in the Mälaren Area 110 4.4 Lade and Mære in Trøndelag 132 4.5 Hof Buildings in Iceland 147 vi contents

Part 3 General Features

5 The Religious Leadership of the Sanctuaries and the Economic System 163 5.1 The Debate about Cultic Leadership and the Terminology 164 5.2 Chieftains (hǫfðingjar, hersar, goðar) as Cultic Leaders of the Sanctuaries 167 5.3 Kings and Earls as the Cultic Leaders of Sanctuaries in Trøndelag and Svetjud 186 5.4 The Financial System of the Public Cult at the Sanctuaries 192 5.5 Ideological Motives behind the Rulers’ Interest in Public Cult and Sanctuaries 197

6 The Ritual Structure of the Sanctuaries and Cosmic Symbolism 199 6.1 Cosmic Symbolism: Some Theoretical and Methodological Issues 200 6.2 The Location, Architecture and Decoration of Ceremonial Buildings 205 6.3 The Interiors: High-seat, High-seat Posts and Hearths 219 6.4 The Exterior Ritual Structure of the Sanctuaries 249

7 The Relation between Rulers and Cultic Images 264 7.1 Miniatures and Medallions 265 7.2 The Three Statues of Gods in the “Uppsala Temple” 266 7.3 Cultic Images in the Sanctuaries of Trøndelag and Northern Norway 272 7.4 Cultic Images in the Sagas of Icelanders and the Image from Eyrarland 282 7.5 Cultic Images and Religious Strategies for Rulership: Hypotheses 287

8 Sacred Sites, Ritual Restrictions and Religious Strategies for Rulership 290 8.1 Sanctifications of Places and Ritual Restrictions (Taboos) 291 8.2 The Enclosed Sacred Space 298 8.3 Violations against the Sanctity of the Sacred Space: The Example of Egils saga 303 8.4 The Ruler as the Protector of the Sanctified Space 305 8.5 Burning or Breaking Down the Ceremonial Building: An Attack against the “World” of the Ruler 307 contents vii

8.6 The Renovation and Ending of Ceremonial Buildings and Sanctuaries 311

9 Ceremonial Feasts as Ritual Strategies 316 9.1 Hákonar saga góða and the Ceremonial Feasts at Lade and Mære 316 9.2 Calendrical Religious Feasts 320 9.3 The Ruler as Sacrificer 325 9.4 Smearing Sacrificial Blood on the Sacrificial Platforms 331 9.5 Ceremonial Meals and the Custom of Eating Horse Meat 344 9.6 Ritual Drinking and the Reciting of Ritual Formulae 351 9.7 Gift-Giving Ceremonies 359 9.8 Excursus: The Central Role of Noble Women at the Ceremonial Banquets 363

10 Rulers, Rings and Rituals—From Svetjud to Iceland and Back Again 376 10.1 The Forsa Ring from Hälsingland—A Sacred “Oath-Ring” at a Sanctuary? 377 10.2 “Oath-Rings” and hof Sanctuaries in the Sagas of Icelanders 387 10.3 Ritual-Rings and Sanctuaries in Svetjud 391 10.4 The Custom of Swearing Oaths on Rings 398

Part 4 Regional Diversity

11 Golden Objects as Media of Power 407 11.1 Gold Foil Figures 407 11.2 The Gold Bracteates at the Hall of Helgö 418 11.3 The Golden Medallion from Inderøy 419 11.4 The “Golden Door Ring” of the hof Sanctuary of Lade 421 11.5 Golden Arm- and Neck-Rings 424

12 Ceremonial Buildings, Great Burial Mounds, and the Worship of Dead Rulers 430 12.1 The Distribution of Great Burial Mounds in the Investigation Areas 431 12.2 Great Burial Mounds as Part of the Aristocratic Sanctuaries 433 12.3 Great Burial Mounds as Symbols of Genealogy and Land Rights 444 viii contents

12.4 The Cult of Heroes at the Great Mounds 453 12.5 The Cult at the Burial Mounds of Freyr and Hǫlgi 466

13 The Arenas for Royal Funeral Feasts and Inauguration Rituals 476 13.1 King Ǫnundr’s Funeral Feast 476 13.2 Libations at Funeral Feasts 481 13.3 The Place for Drinking Beer at the Funeral Feast 485 13.4 The High-Seat and Inheritance 488 13.5 Royal Inaugurations, High-Seats and the Great Mounds 492 13.6 A Royal Rite of Passage 498

14 (Inter-)Regional Chief Sanctuaries as a Religious Strategy for Rulership 503 14.1 Cultic Organization 503 14.2 The Chief Sanctuaries of Trøndelag and Svetjud 508 14.3 The Cult of Many Gods at the Chief Sanctuaries as Strategy of Power 516

15 Valhǫll, Ceremonial Buildings and Warrior Lords 521 15.1 The Martial and Eschatological Valhǫll Mythology 521 15.2 The Warrior Halls in Late Iron Age Scandinavia and Valhǫll 523 15.3 The Distribution of Warrior Lords with Warrior Bands in Scandinavia 527 15.4 The Valhǫll Mythology as Part of Ruler Strategies in Svetjud and Trøndelag 533

Part 5 Epilegomena

16 Conclusions 539

Bibliography 547 Index 617 Preface

The original aim of the present research project, which was financed by Riksbankens Jubileumsfond in Sweden (2008–2011), was to investigate the pre- Christian and medieval cultic buildings in Scandinavia from an ideological perspective. The specific purpose was to study how rulers used these buildings to express and reinforce their political power. The study involved a long-term perspective, AD 550–1200. The purpose of this long-term perspective was to investigate continuity and change in the context of the conversion from the pre-Christian halls and cultic buildings to their replacement by churches in Scandinavia. The project followed to a great extent the initial project plan, but some limitations were defined as work progressed. The empirical study was delimited to the Lake Mälaren area, Trøndelag and Iceland. The project also shifted focus somewhat. In the present monograph, the regional perspective has gained greater weight than the diachronic dimensions of the study. Focus has been placed on the pre-Christian contexts, specifically on the issue of regional similarities and differences/variations. The diachronic dimension of the study, i.e. the development from pre-Christian to Christian contexts, was published in an article at Brill 2011, with focus placed on the Mälaren region.1 I am indebted to many scholars for giving me valuable advice and construc- tive criticism during the preparation of the present book. First and foremost I express my gratitude to Professor Bo Gräslund (Uppsala), who has been standing by my side from the first manuscript of this work to the final production. Also my former “Doktor Vater” Professor Anders Hultgård (Uppsala) has offered me a great deal of good advice and encouragement, as well as my friend and colleague Professor Michael Stausberg (Bergen), who also convinced me to publish the results of my project in an international publication. My dear colleagues, Associate Professor Torun Zachrisson (Stockholm), Professor Anne- Sofie Gräslund (Uppsala), Associate Professor John Ljungkvist (Uppsala) and Associate Professor Per Vikstrand (Uppsala), have all contributed with valuable advice and comments, particular regarding archaeology and onomastics. They have indicated additional materials, described excavation sites, answered questions and discussed problems. I am also indebted to some other scholars who also have generously shared their profound knowledge, especially when working at the Centre for Advanced Study (CAS) in Oslo (2007–2008), but also later; in particular Professor Gro Steinsland (Oslo), Professor Jon Viðar

1 See Sundqvist 2011a. x preface

Sigurðsson (Oslo) and Professor Jens Peter Schjødt (Aarhus). Gro and Jon Viðar gave valuable comments on the entire manuscript at an earlier stage. I am very thankful for that. Some parts of this text have also been discussed in the higher seminars at the University of Gävle (with valuable comments from e.g. Professor Peder Thalén and Professor Jari Ristiniemi) as well as in the higher seminars at Stockholm University (with good advice and reflections from e.g. Professor Peter Jackson, Associate Professor Andreas Nordberg, Associate Professor Erik af Edholm, Associate Professor Marja-Liisa Keinänen; Asso- ciate Professor Susanne Olsson; Doctor Ferdinando Sardella; Doctor Stefan Larsson and Doctor Niklas Foxeus). I am also grateful to Doctor Merete Røskaft (Trondheim), who showed me some of the central places and cultic sites when travelling in Trøndelag, Alan Crozier (Lund) for correcting my English, Maarten Frieswijk (Leiden) and Judy Pereira (Boston) for editing the text, and Professor Anders Andrén (Stockholm) for providing me with photos. Many thanks also to Professor Neil Price (Uppsala) and Professor Terry Gunnell (Reykjavík) for answering my questions. For all shortcomings and faults that still may remain in the text I am alone responsible. The project (and partly the printing of the book) was made possible by grants from Riksbankens Jubileumsfond, while language correction of the manuscript was financed by the University of Gävle. Their support is hereby gratefully acknowledged. I am also grateful to Brill, for publishing my work.

Olof Sundqvist Fårösund, July 2015 List of Figures

0 The three investigation areas 6 1 Tjängvidestenen (G 110), Alskog. Gotland. Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet 79 2a Excavated areas in Old Uppsala. Ljungkvist & Frölund 2015 MS 98 2b Southern plateau in Old Uppsala. Photo: Anders Andrén 98 3 The cultic house at the Viking Age chieftain farm of Borg, in Östergötland, Sweden. Illustration Richard Holmberg 99 4 A cultic house at the central place of Uppåkra, outside Lund, Scania. Illustration Loic Lecareux 100 5 The Stenkvista runic stone (Sö 111) Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet 123 6 The Altuna stone, Uppland (U 1161). Photo: Pål-Nils Nilsson. ATA, Riksantikvarieämbetet 123 7a The hall at Helgö. After Herschend 1998 128 7b Gold foil from Helgö. SHM 128 8a The church of Mære, Trøndelag. Photo: Olof Sundqvist 142 8b Gold foils from Mære, Norway. Photo Åge Hojem. NTNU Vitenskaps museet, Trondheim 142 8c Gold foil from Borg, Lofoten, Norway. Tromsø Museum— Universitetsmuseet 142 9a Detail from the Gotlandic picture stone from Buttle Änge no. I. After Herschend 1998, based on Lindqvist 1941–42 226 9b The picture stones from Buttle Änge no. I. Photo: Carl Gustaf Rosenberg. ATA, Riksantikvarieämbetet 226 10 Stone coffin, Sanda, Gotland. Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet 226 11 A “high-seat” amulet (Sw kubbstol) from Birka. SHM 228 12a One Viking Age tricorn outside the cultic hall building at Helgö. Photo: Anders Andrén 256 12b The plateau of the cultic hall at Helgö. Photo: Anders Andrén 256 13 A small Viking Age bronze statue discovered in Rällinge, Södermanland. SHM 267 14 Three small images from Lunda, Södermanland: Lunda 1 (SHM); Lunda 2 (SHM); Lunda 3 (SHM) 271 15 A cultic image from Eyrarland, Iceland 283 16a Gotlandic picture stone from Lärbro (Hammars I). Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet 335 xii list of figures

16b Detail of the picture stone from Lärbro (Hammars I). Photo: Olof Sundqvist 335 17a The ring from Forsa in Hälsingland, Sweden. Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet 378 17b Detail from the ring from Forsa. Photo: Bengt A Lundberg. ATA, Riksantikvarie-ämbetet 378 18 The door ring from Delsbo church, Hälsingland, Sweden. Photo: Unknown. ATA, Riksantikvarieämbetet 381 19a Þórr’s hammer made of silver from Läby, Uppland. SHM 385 19b Þórr’s hammer from Mickels, När, Gotland. SHM 385 20 Fire-steel-shaped pendants from Borg, Östergötland. SHM 385 21 The Häckelsäng ring, found in Hamrånge parish in Gästrikland. Gävleborgs museum 397 22 A picture stone from Tängelgårda, Gotland. SHM 399 23a Gold foil figures from Hauge-Tu, Rogaland, Norway. The Museum in Stavanger 410 23b Gold foil from Krokek, Östergötland. SHM 410 24a and 24b Iron door rings from Uppåkra, Scania, Sweden. Photo: Lunds universitets Historiska museum 423 25 The Sparlösa stone (Vg 119), Västergötland. Photo: Bengt A Lundberg ATA, Riksantikvarieämbetet 424 26 The small figurine from Kymbo, Västergötland. SHM 428 27a The three royal burial mounds from Old Uppsala. Photo: Marie-Louise Sundqvist 434 27b The burial mound at Alstadhaug, Trøndelag. Photo: Olof Sundqvist 435 27c The burial mounds at Bertnem, Trøndelag. Photo: Olof Sundqvist 435 28 The distribution of theophoric place-names in the Mälaren region. Made by Fideli Sundqvist and based on Per Vikstrand 2001 517 Abbreviations

ATA Antikvarisk-topografiska arkivet, Riksantikvarieämbetet, Historiska museet och Kungl. myntkabinettet ÄVgL Äldre Västgötalagen BL Borgarthings-Lov DMS Det medeltida Sverige DR Danmarks runeindskrifter DS Diplomatarium Suecanum E Edsöresbalken EL Eidsivathings-Lov FL Frostathings-Lov GL Gotlandslagen (Gutalagen) Goth Gothic GuL Gulathings-Lov HMS Historiska Museet Stockholm Hs Hälsinglands runinskrifter Ísl. Fornr. Íslenzk Fornrit Kg Konungabalken KrL Kristoffers Landslag Kk Kyrkobalken KVAAH Kungl. Vitterhets-, historie-, och antikvitetsakademiens handlingar KVHAA Kungl. Vitterhets-, historie-, och antikvitetsakademien Lat Latin MGH Monumenta Germaniae Historica Mh Manhelgdsbalken MHG Middle High German NGL Norges Gamle Love No Norwegian OE Old English OFris Old Frisian Ög Östergötlands runinskrifter OGut Old Gotlandic OHG Old High German OIr Old Irish Öl Ölands runinskrifter ON Old Norse ONP Ordbog over det norrøne prosasprog xiv abbreviations

OÖD Ordbok över folkmålen i övre Dalarna OPIA Occasional Papers in Archaeology OSax Old Saxon OScan Old Scandinavian OSw Old Swedish PG Proto-Germanic PN Proto-Nordic R Rättlösabalken (in ÄVgL and YVgL) Rättegångsbalken (in ÖgL and UL) RAÄ (raä) Riksankvarieämbetet RSw Runic Swedish SHM Statens Historiska Museum Skj Den Norsk-Islandske Skjaldedigtning SkP Skaldic Poetry of the Scandinavian Middle Ages Sm Smålands runinskrifter Sö Södermanlands runinskrifter SRS Scriptores rerum Suecicarum medii aevi SSGL Samling af Sweriges Gamla Lagar Sw Swedish U Upplands runinskrifter UL Upplandslagen VA Vita Anskarii Vg Västergötlands runinskrifter part 1 Prolegomena

⸪

chapter 1 Introduction

The great Icelandic writer and chieftain Snorri Sturluson (1179–1241) narrates in his Ynglinga saga (a part of his masterpiece Heimskringla, c. 1230) the story of King Ingjaldr of Svetjud and the funeral feast (erfi) he gave for his father King Ǫnundr at the pre-Christian royal site of Uppsala in present-day Sweden.1 According to Snorri, Ingjaldr prepared a hall in no wise smaller or less stately than the hall called Uppsalr, and called it “the Hall of Seven Kings”. In this hall seven high-seats were erected. Then the king sent messengers through all of Svetjud, inviting seven kings, earls and other prominent men. All but one of the kings arrived at Uppsala, where they were given seats in the new hall. Snorri than describes in general terms the rituals usually performed at a royal funeral feast in pre-Christian contexts. It was the custom, for instance, when a funeral feast was prepared to honour a departed king or earl, that the one who prepared the feast and was to be inducted into the inheritance was to sit on the step before the high-seat until the beaker called the bragafull was brought in; and then he was to stand up to receive it and make a vow, then quaff the beaker, whereupon he was to be inducted in the high-seat which his father had occupied. Then he had come into his rightful inheritance as his successor. This duly took place here; and when the beaker was brought in, King Ingjaldr stood up, seized a large drinking horn, and made the vow that he would increase his dominion to twice its size in every direction, or else die. Then he emptied the beaker. Later the same night Ingjaldr said to his retainers that they should bring fire into the new hall and set it aflame. There the six kings burned and all their troop, and those who came out were quickly killed. After that, King Ingjaldr put under himself all those realms that the six kings had had, and he took tribute from them. This text cannot be regarded as a source for historical reconstruction in a general sense. Several fictitious, or literary elements may be at stake in this account, for instance, Ingjaldr’s project of making a kingdom under a single monarch by means of burning the other district kings in the new hall. Actually we do not really know whether King Ingjaldr ever lived in Svetjud in ancient times, or if he held a funeral feast for his father Ǫnundr at Uppsala. Anyhow, there are some details in the text which seem to be based on more solid

1 Ynglinga saga 36, Heimskringla I, Ísl. Fornr. 26, 66–67. This survey is based on Hollander’s translation. This text will be quoted in full and analysed in detail in ch. 13.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_002 4 CHAPTER 1 historical grounds, such as the description of rituals, ritual objects, constructions and buildings used by Ingjaldr to claim authority and power in Svetjud. Most significant is the use of an impressive ceremonial building called Old Norse salr ‘hall’ and the high-seat called hásæti. They seem to have been the foci of the royal inheritance rituals according to other sources (see further ch. 6 and 13 below). It is also striking that Ingjaldr erected the hall building at Uppsala, the most important political and religious centre of the Svear in the Late Iron Age (c. 550–1050/1100 AD). Snorri’s text harmonizes here with significantly more reliable sources. Archaeological evidence indicates, for instance, that a great hall building, fifty metres in length, was erected in Old Uppsala, Sweden, during the Merovingian Period and the Early Viking Age. At other aristocratic settlements in Late Iron Age Scandinavia, traces of great hall buildings, high-seats and ceremonial beakers have also been found.2 It is likely that they all played a crucial role for the contemporary rulers and their construction of power. The aim of this study is to investigate religious strategies for gaining political authority and legitimacy at ceremonial buidlings in some parts of Late Iron Age Scandinavia (c. 550–1050/1100 AD). The major task is to analyse the role played by religion in political undertakings among the pre-Christian ruling elites at such buildings from the Merovingian Period (550–750 AD) to the Viking Age (c. 750–1050/1100 AD).3 As seen in the case of King Ingjaldr, religious and ideological motives were part of performative strategies that also included a deliberate choice of localities, architecture and rituals, intended to symbolize and legitimize political power. Several studies in the history of religions have previously shown that myths and rituals in general were important for Late Iron Age rulers in Scandinavia

2 References to these excavations are presented in chapters 4, 6 and 9 below. 3 The chronological system used in this study is as follows: The Roman Iron Age (c. 0–375 AD), the Migration Period (c. 375–550), the Merovingian Period (in Sweden called the Vendel Period; c. 550–750 AD), and the Viking Period (c. 750–1050/1100 AD) which is followed by the Medieval era. The Roman Iron Age and the Migration represent in present study the Middle Iron Age, while the Merovingian Period and the Viking Age represent the Late Iron Age. The introduction of Late Iron Age at c. 550 is used in, for instance, Norsk Arkeologisk Leksikon, (eds.) Lotte Hedeager & Einar Østmo 2005, 115, 241. The border AD 550 has come into renewed light through the climate catastrophe of global proportions that took place in the year 536, with subsequent impact the 10 following years. It has been described and discussed by Bo Gräslund & Neil Price 2012. It coincides approximately with the changes in material culture around AD 550 that were previously observed by archaeologists—changes that are especially obvious in the prehistory of Sweden and Norway. The term “Middle Iron Age” was suggested to me by professor Bo Gräslund. Introduction 5 when it came to expressing positions of power.4 Few of these studies, however, have focused on the role of ceremonial buildings in this context.5 In this book I will study how rulers used halls and cultic houses to construct, express and maintain their political power. I will explore how rulership was built in a concrete sense, and focus on the material dimension of ruler ideology, such as the monumentality of the buildings, the ritual structures and objects appearing at the sites of the elites.6 I will also concentrate on ceremonies and performative rituals executed at these places and structures. By performative rituals I mean that the focus is on the impact and effect of the actions. Attention is therefore directed towards the intention behind the ritual action, what it accomplishes, and how it affects the individuals who have taken part in it.7 When historians of religions previously investigated the relation between power and religion, they often proceeded from the theory of a Scandinavian “sacral kingship”.8 In these studies scholars often reconstructed a religious ruler ideology which was supposedly common to all of Scandinavia (see further below). In this work I will apply a regional approach, so as to be able to account for the specific historical, cultural and social contexts. The focus is mainly on three regions, the Lake Mälaren area in Sweden, Trøndelag in Norway, and Iceland (see fig. 0). These areas have been selected since Late Iron Age halls, cultic houses and sanctuaries are well documented from all of them, by means of both written sources and archaeological finds.9 The analytic concept of “hall” refers in this study to a multifunctional building where cultic activities also took place, while “cultic house” designates a smaller, more specific cultic building. The concept of “ceremonial building” comprises both “hall” and “cultic house”. Occasionally the term “sanctuary” will be applied. It refers to the entire cultic site, including the ceremonial buildings as well as the ritual structures and sacred places located outdoors beside such buildings.10 There are also outdoor sanctuaries without ceremonial buildings

4 See e.g., Steinsland 1991; 2000; Sundqvist 2002. 5 Admittedly, some archaeologists and historians of religions have also touched upon the ideological dimensions of the ceremonial buildings. See e.g., Herschend 1993; 1998; Nordberg 2003; Söderberg 2005. 6 Cf. Miller 2005; 2010; Meskell 2005. 7 A classic example of a performative speech act is “I now pronounce you as man and wife”. The words mean an actual change. On performative rituals and acts, see e.g., Austin 1962; Tambiah 1979; Bell 1997, 68–69; Rappaport 2000, 124–126; Raudvere 2002, 32–34; Grimes 2006, 390–391. 8 On this theory, see e.g., McTurk 1975–76; 1994; Schjødt 1990; Sundqvist 2002, 18–38. 9 The three regions and the sources will be presented in more detail in chapters 2–4. 10 See further ch. 4 below on these concepts. 6 CHAPTER 1

figure 0 The three investigation areas. in Late Iron Age Scandinavia. They will not primarily be focused on in this study. The present investigation focuses on the following questions:

– who controlled, owned and managed the halls, cultic houses and sanctuaries, and what relations did they have with the ruling elites in the three areas? – where were the ceremonial buildings and sanctuaries located and what did they look like, and in what sense can they be connected to the leading groups in society? – which ritual objects found, and ritual activities deduced, in and outside these buildings, may have had a connection to the religious strategies for legitimacy, authorization and empowerment?

My hypothesis is that ceremonial buildings and sanctuaries during the pre- Christian periods in all three investigation areas were multifunctional arenas for the elite. These places served as a ritual space where the rulers could dem- Introduction 7 onstrate and manifest their power, in both a concrete and a symbolic sense, and as a place where they took tribute from their subjects. Since the political structure and other contextual aspects partly differed in the three regions, the religious strategies for gaining legitimacy and authorization at the sanctuaries also varied to some extent in these areas. Conditions in the Mälaren area and Trøndelag resembled each other more than circumstances in Iceland. The present work is to a considerable degree based on textual studies. Since the questions posed in this study concern Late Iron Age ceremonial buildings and material aspects of ruler ideology, archaeological sources and place-names are equally important source materials. In what follows, previous research on religious ruler ideology in Scandinavia will be surveyed (i.e. the theory of sacral kingship), also noting some methodological problems and theoretical issues. I will then present my own perspectives, approaches, essential concepts and hypotheses. The concept of “religious strategies for rulership” will be discussed below. A thorough presentation of the pre- Christian ceremonial buildings, their history of research, and the terminology, will be given in ch. 4, as well as the evidence of them in the sources referring to the investigation areas.

1.1 Religious Ruler Ideology in Previous Research

1.1.1 The Theory of Sacral Kingship When historians of religions have studied the religious ruler ideology in ancient Scandinavia they have often applied the comparative and phenomenological concept of “sacral kingship”. This concept has a long history in research. It was coined and elaborated by James G. Frazer in his famous book The Golden Bough (1890). Inspired by ideas presented by Wilhelm Mannhardt,11 Frazer constructed a theory based on the annual cycle of nature, dying and resur- recting gods, and the supernatural power of the kings visible in sources from the Mediterranean area. According to Frazer, the king was considered as both a priest and a divinity. He took part in a sacred marriage (hieros gamos) with the fertility goddess, played the role of the god in a ritual drama, and was even killed or sacrificed in order to guarantee the well-being of the community. As soon as The Golden Bough became known in Scandinavia, some Swedish scholars absorbed Frazer’s ideas about sacral kings, and applied them to materials on the ancient kingship of the Svear. Henrik Schück, for instance, tried to

11 Mannhardt 1904–1905. 8 CHAPTER 1 reconstruct the cult of Freyr at Uppsala, frequently referring to Frazer’s theory.12 According to Schück, the king of the Svear was not only regarded as a priest or a mediator between the people and the divine world. He was equal to a god and considered a son of Freyr. Frazer’s idea that the king celebrated a hieros gamos with the fertility goddess was also quickly absorbed by those who applied the sacral kingship theory to Scandinavia.13 Also essential for the establishment of this theory in Scandinavian contexts was Vilhelm Grønbech’s Vor folkeæt i oldtiden (1909–1912).14 In this study Grønbech elaborated the idea of the king’s luck. According to this, the ancient Scandinavians believed that their chieftains and kings had an exceptional ability to bestow blessings on their people, such as harvest, prosperity and peace. This theory was particularly maintained by German scholars. They argued that a legitimate claimant to the ancient Germanic kingship was taken from the stirps regia, i.e. the royal kin, by the members of the þing-assembly.15 Two criteria then had to be fulfilled. A candidate must claim divine descent and he must be a carrier of royal luck, “Königsheil” (mana). Each member of the royal kin was thus believed to have a “Geblütsheiligkeit”, which legitimized him for the position of king (“das germanische Geblütsrecht”). The divine king represented the gods vis-à-vis his people. He granted good crops, victory in battle and the maintenance of law and order. On the other hand, in the cult of divinities he represented the people. The people’s offerings therefore had to be mediated by the king. From 1904 to the beginning of the 1960s there was thus a broad consensus that the ancient Scandinavian (and Germanic) kingship was “sacral”.16 However, when Walter Baetke published his work Yngvi und die Ynglinger in 1964, this entire issue was reconsidered. Employing radical source criticism, Baetke argued that the fundamental features of the sacral theory were not visible in the reliable primary sources. They could only be seen in the uncertain medieval Icelandic saga literature. Baetke’s critical line of reasoning gained support from some philologists and historians of literature,17 while scholars with a comparative perspective held to the idea of sacral kingship in

12 Schück 1904. 13 See e.g., Schück 1904; von Friesen 1932–34; F. Ström 1954. 14 Grønbech 1997 (1909–12). 15 See e.g., Kern 1954 (1914), 3–45; Schramm 1937, 14, 28, 140–178; Hauck 1950; Höfler 1952; 1959; Hoffmann 1976, 1–22; de Vries 1956–57, §§242, 274–276. On this theory, see also ch. 13 below. 16 Åke V. Ström (1959), for instance, stated thus: “That kingship in Old Scandinavia was entirely sacral, is nowadays considered as a matter of fact.” 17 E.g. Kuhn 1971. Cf. Lönnroth 1963–64; 1968; 1986. Introduction 9 pre-Christian Scandinavia.18 In the last four or five decades it seems as if the sacral kingship theory mostly has lost its importance for research into ruler ideology in Scandinavia. Recent works on Old Norse religion, on the other hand, are sometimes still influenced by this theory and the terminology it comprises. However, the discussion has in several senses also been renewed.19

1.1.2 Methodological Problems and Theoretical Issues When discussing the religious ruler ideology in ancient Scandinavia, historians of religions have thus often used the phenomenological category of “sacral kingship”. As noted above, there has been heated debate as to whether sacral kingship existed in pre-Christian Scandinavia or not. This debate has, however, mostly concerned source criticism. Few debaters have reflected on the wider methodological and theoretical problems involved in the attempt to apply this category to ancient Scandinavian contexts.20 In what follows, I will survey some problems that I have noticed in previous research, and point out some reorientations that will be applied in the present investigation.

1.1.3 Global Comparisons Based on Secondary Sources and Looking for Similarities Only One problem is that the concept of sacral kingship was first formulated by scholars who combined evolutionary theories and comparative methods. Its elaboration was based mainly on sources from the Mediterranean area and the ancient Near East. When it was applied to ancient Scandinavian society, similarities to Near Eastern kingship were thus sought, while unique cultural and historical developments were overlooked. Scholars thus made global comparisons in which they were looking for similarities only.21 By means of the abstract concept of “sacral kingship”, comprising sets of features, they also “colonized” the ancient Scandinavian culture with certain aspects which sometimes were not clearly visible in the sources. I will give an example. During the first half

18 F. Ström 1968; 1981; 1983; Å. V. Ström 1975, 266f. 19 See e.g., Schjødt 1990; Steinsland 1991; 2000; Sundqvist 2002; 2004b. 20 Baetke’s reaction against the sacral kingship theory was based not so much on general methodological grounds as on source criticism. It was only in a footnote that Baetke criticized the methods used by those who supported the theory: “Auf die religionsgeschichtlichen Spekulationen, die er [I. Lindquist] auf der Grundlage gewisser (sehr zweifelhafter) Theorien von Frazer (The Golden Bough), Dumézil und andern entwickelt, kann ich an dieser Stelle nicht näher eingehen. Ich halte sie für ziemlich phantastisch.” Baetke 1964, 37, note 1. 21 For similar criticism, see Rydving 2006; 2011a. 10 CHAPTER 1 of the twentieth century, certain specialists in ancient Scandinavian religion were anxious to find traits of a hieros gamos in the pre-Christian Scandinavian kingship.22 Since the “sacral theory” claimed that the divine king celebrated a symbolic or ritual marriage with the fertility goddess, they also expected to find evidence of such a marriage in the Scandinavian sources. Of course they also found it there. However, it is obvious that they pushed the data too far in order to do that. The interpretations produced by these scholars were thus strongly influenced by the general research concept, while the information from the sources was repressed. These interpretations have recently been reassessed. By means of a careful analysis of the sources, Gro Steinsland has shown that the marriage between the god and a mythical being in Scandinavia should not be interpreted as a “sacred marriage” similar to the Near Eastern version.23 In the Scandinavian context, the god was related to a chaotic female, described as a giantess. Even if this union led to the production of a prototypical king, it also indicated a dark fate for the king. It should be noted that Gro Steinsland in her thesis took a sceptical position against the general features of sacral kingship in the Scandinavian context.24 Britt-Mari Näsström too has been sceptical of the transfer of the hieros gamos motif from the cults of the ancient Mediterranean to Scandinavian conditions, but mostly because it sometimes has been based on false assumptions about the former religious systems: “The hypothesis according to which the king marries the goddess, whereupon she sacrifices him, is thus based upon false conclusions about the Phrygian rites.”25 According to Näsström (who is a specialist in both Roman and ancient Scandinavian religion) these scholars26 lacked relevant facts about the Phrygian cult, “a circumstance which invalidates a hypothesis along these lines”. In my opinion the problem that Näsström puts her finger on is bound to the fact that the comparisons made by some scholars not were based on primary sources only, but also on secondary sources (literature), such as references from handbooks and other syntheses.

1.1.4 The Static and Abstract Character, and the Method of Using Texts Only In early studies, the pattern of the sacral kingship concept was thus more or less imposed upon the Scandinavian material. Scholars looked almost

22 Schück 1904, 248–306; von Friesen 1932–34; F. Ström 1954, 7ff., 32ff. 23 Steinsland 1991, passim; 2000, 57ff.; 2011b, 16f. See also her new criticism against the “sacral kingship theory” in Steinsland 2011a, 16, note 3. 24 Steinsland 1991, 310ff.; 2011b, 57. 25 Näsström 1995, 200. 26 E.g. Phillpotts 1920; F. Ström 1954. Introduction 11 desperately for evidence to fill out the pattern, which was sometimes borrowed from totally different areas and periods. Because of this wide synchronous method, they often made broad generalizations based on weak data and they also overlooked possible regional and temporal variations not only in Scandinavia, but even in all Germanic areas. Their method was built on a tacit theoretical assumption that religion, ideology and culture were homogeneous over the entire Scandinavian or Germanic area. It also seems as if they felt that time had no influence on the development of the ruler ideology. When, for instance, Folke Ström only found evidence of the “pagan notion” of the king’s luck (Königsheil) in the medieval Scandinavian prose (and slight evidence in skaldic poetry), he supplied this lack of data by bringing attestations from Ammianus Marcellinus’ text about the Burgundian rulers written in the fourth century, which actually had a quite different context from the former materials.27 This wide synchronous method was repeated by Åke V. Ström on the same issue, where he brought together information from Ammianus Marcellinus with some utterances made by King Gustav I (Vasa) at the “Riksdag” in Västerås 1527, recorded by Peder Swart.28 In this case too, the contexts of the sources were quite different. It thus seems as if the application of the “sacral model” to the Germanic sources has sometimes resulted in reconstructions which totally overlook possible regional and diachronic variations. Folke Ström’s and Åke V. Ström’s approaches to the notion of king’s luck are good examples here. It is also typical that scholars who applied the sacral model mainly investigated written sources, and often disregarded other types of

27 Ammianus Marcellinus XXVIII 5,14. F. Ström 1968. According to Baetke (1964) and Lönnroth (1963–64) the notion of a pre-Christian king’s luck in the Scandinavian sources was affected by the medieval Christian notion of the king’s divine grace (rex dei gratia) or “charisma”. Criticism of this position is voiced by Hallberg 1966; 1973. See further ch. 12 below. It should be noticed that Folke Ström in previous investigations sometimes took regional variations into account. In, for instance, his Diser, nornor, valkyrjor: fruktbarhetskult och sakralt kungadöme i Norden (1954) he made a distinction between the cult of dísir in Western and Eastern Scandinavia. In this study he nevertheless interprets the skaldic poem Ynglingatal with support of the universal theory of sacral kingship and thus reproduced “the general pattern” of this institution in his interpretation of this lay. Ström was actually aware about the problem of continuity, change and regional variation within the Norse religion. In his handbook Nordisk hedendom, for instance, he stated that Norse religion, may have varied over time and space, and that we are facing a serious problem because of the uneven distribution of the literary sources, which almost all have a West-Nordic background. F. Ström 1985, 7f. Cf. Nordberg 2012, 126f. In my opinion, Ström was anyhow one of those scholar who overlooked possible local and temporal variations related to ruler ideology, by using unilaterally the pattern of the sacral kingship theory. 28 Å. V. Ström 1975, 268f. 12 CHAPTER 1 sources. This fact also contributed to the static character of the reconstructions, since the written materials alone do not allow chronological analyses of pre-Christian conditions in Scandinavia. In the present study I will argue that the sources indicate that the religious strategies for authorization and empowerment at ceremonial buildings in Late Iron Age Scandinavia may have varied in different regions and contexts.

1.1.5 The Decontextualized Perspective When scholars focused on sacral kingship in their compara tive and phenom enological treatments, often only the “religious” aspects were studied, while “profane” elements were left out. In that way, the religious structure of kingship was detached from its specific historical, social and cultural settings. For instance, the specific political structures of different parts of Scandinavia were seldom thoroughly analysed in these studies, and concepts such as “kingship” were not clearly defined. This leads to serious problems, since the specific social and political contexts most likely had consequences for the formulation of the religious ideology of the rulers in different places. We may take a concrete example. In more recent studies, historians have argued that the Late Iron Age societies of Norway and Svetjud were decentralized, and the power was split between many chieftains, who only occasionally gathered in unstable federations under overlords or High Kings.29 In both Norway and Svetjud, extended territorial domination may have developed at a rather late date, and in Iceland we did not have any kingship at all. Several questions appear here. On what social level did the ideology of “sacral kingship” actually appear? Did it exist in minor chieftainships, or was it only connected to territorial kingships, or the institutions of High Kings? Such questions have rarely been addressed by scholars using the sacral kingship model.30 When discussing the religious strategies construed and performed at sanctuaries in Svetjud, Trøndelag and Iceland below, these questions will be taken into consideration.

1.1.6 A Biased Focus on Male Leaders When the concept of sacral kingship has been applied in previous research, there has been a bias towards male leaders. In more recent studies the religious leadership of mighty women has also been observed.31 It has been noted, for

29 See e.g., Jón Viðar Sigurðsson 2008a, 22–31; Krag 2008, 645–651; Lindkvist 2008; P. H. Sawyer 1991, 3–10. 30 See, however, Steinsland 2011a; 2011b; Jón Viðar Sigurðsson 2011; Sundqvist 2011a; 2012b. 31 See e.g., Steinsland 1985a; 1985b; Enright 1996; Näsström 2001b; 2009; N. Price 2002; Raudvere 2003; Dillmann 2006; Sundqvist 2007. Introduction 13 instance, that queens and other female leaders (ON húsfru, gyðja, vǫlva) may have played important ritual roles during cultic feasts in the aristocratic halls. Thus these females too used religious strategies to gain authority. These powerful women may be attested notably in the archaeological materials,32 but they seem to appear in other types of sources as well.33 In chapters 5 and 9 I will argue that female cultic leaders played a significant ritual and ideological role within the contexts of halls, cultic buildings and sanctuaries.

1.1.7 Lack of Interest in the Material Dimension of Religious Ruler Ideology When proceeding from the sacral kingship theory in the analysis of ancient Scandinavian rulership, certain themes were often focused on, while other aspects were overlooked. Much energy was placed on the establishment of the ruler’s relationship with the gods and divine power. The king was, for instance, regarded as divine or/and as an offspring of the gods, possessed supernatural powers and controlled the fate of the world etc.34 The material dimension of the religious ruler ideology was rarely observed. Without doubt this dimension was crucial when constructing authority.35 The present study will focus on the role of ceremonial buildings in the context of rulership, as well as other ritual structures, objects and “things”, such as the high-seat, ritual rings, ceremonial food and drinks, as well as sacrificial gear and blood.

1.1.8 The Hereditary Taint of the Concept Concepts such as sacral kingship are tools. The user of these tools is the one who gives them their specific meaning. Hence, in principle, anyone could use any concept he or she wants and give it its specific meaning for a certain context. In any case, we know that the general concepts used in research have a history, which has also given them certain loaded values and connotations. Sometimes they reflect more neutral values, while in other cases they may be associated with problems and criticized ideas. Looking at previous research in the general history of religions, it seems as if the concept of “sacral kingship” has this latter hereditary taint. It has been associated with evolutionary thinking, problematic theories of diffusion, and unscrupulous universalism. This attitude is revealed, for instance, in an article written by Roberta Frank, where she attacks scholars who have interpreted certain motifs in skaldic poetry as

32 N. Price 2002; A.-S. Gräslund 2008. 33 Sundqvist 2007, 56ff. 34 See above and Sundqvist 2002, 27. 35 Cf. e.g., Meskell 2005, 51–71 and Miller 2005, 15–20. 14 CHAPTER 1 essential ingredients in the “sacral kingship package”, i.e. the Romantic theories associated with Frazer and Mannhardt.36 The concept has thus in certain contexts developed into a communicative obstacle.

1.2 Towards a Model for Analysing Religious Strategies for Rulership at Ceremonial Buildings

There is thus a clear need to reconsider the entire issue of rulers, religion and ideology in ancient Scandinavia. The application of the sacral kingship concept to the Scandinavian context suffers from a number of methodological flaws, and its “patternism” has sometimes led to a disregard of the evidence of the sources themselves. Hence, several reorientations must be made. Firstly, when using the comparative approach to this theme, not only similarities but also differences or variations should be sought.37 We must look for unique aspects of the religious ruler ideology and not only discuss the general features which have been associated with the sacral model. The comparisons should also be based on historical studies of primary sources. Secondly, we must go beyond the intrinsic static character of the model, in order to take temporal and regional variations into consideration.38 Not only texts but also other

36 Frank 2007, 178. 37 In present work the concepts “regional differences” and “regional variations” are regarded as synonymous, even if their actual semantics vary somewhat. I do not mean that there were “different religions” in ancient Scandinavia, but regional differences or variations within Old Norse religion. Cf. Nordberg 2012, 132–136. In the present study, ancient Scandinavian religion refers to the religion which existed among those people who spoke ancient Scandinavian/Old Norse languages during the Late Iron Age in this area. See further Sundqvist 2007, 9f. 38 This approach has recently been applied to ancient Scandinavia, but in another context. See mainly Rydving 1990; 1993; 2006. See also Jonathan Z. Smith (2000; 2004) who has questioned the old comparative approach in the general study of religion and argued that comparisons are problematic and unscientific, since specific rules are lacking. He and other scholars have also stated that comparative studies have sometimes included an inadequate attention to differences, diachronic dimensions, and contexts. For a radical post-modern and post-colonial position on this issue, see e.g., Hewitt 1996; Wiebe 1996; Gilhus & Mikaelsson 2003. This position is criticized by Segal 2001; 2005. In my opinion William E. Paden (1994; 1996) and some of the authors in the volume A Magic Still Dwells (eds. Patton & Ray 2000) have taken a reasonable stance. They argue in support of a new comparativism, where both similarities and differences are taken into consideration, as well as cultural contexts. This approach has also been applied to the theme of religious ruler ideology. Gregor Ahn, for instance, has studied ruler ideology in ancient Iran. He states: “Die neuere Forschung betont diesem Ansatz gegenüber zunehmend die Eigenarten Introduction 15 types of sources must be applied in this enterprise. Thirdly, since the formation of “sacral kingship” or ruler ideology concerns the relationship between religion and political structures, decontextualized analysis must be avoided. This implies an increased demand for general cultural competence on the part of the scholar. Even if the phenomenological category of “sacral kingship” in many senses is problematic, it has at least been useful as a heuristic instrument, for instance, when framing a research area and posing questions to the materials. If this concept is to be dismissed, it must be replaced by another analytic concept which is more adjusted to conditions in pre-Christian Scandinavia, and also to the object of the present study, i.e. the ruler-ideological dimensions of ceremonial buildings and sanctuaries.39 In my opinion an alternative term could be proposed, namely “religious strategy/strategies for rulership, (or political authorization)”.40 This concept can be applied to different social, gender and religious contexts. It can, for instance, be applied to a chieftain (hǫfðingi, hersir, goði) in a local settlement district, as well as to an earl ( jarl) or a king (konungr) ruling over people in a regional dominion, or even to a “federation

der jeweiligen Ausprägung des Königstums in den einzelnen Kulturen und fordert ein bei weitem feineres Differenzierungssensorium sowohl für den interkulturellen Vergleich als auch für Entwicklungsstufen innerhalb ein und desselben Kulturraums.” Ahn 1992, 18. 39 Alternative concepts to “sacral kingship” have been presented in previous research on pre-Christian Scandinavian religion. The concept of ruler ideology, for instance, has been discussed in this context. See Steinsland 2011a. Cf. Jón Viðar Sigurðsson 2011, 70. In the research project “Ideology and Power in Viking and Middle Ages” at CAS in Oslo 2007–2008, led by Steinsland, where also I took part, the concept of “religious ideology of rulership” was often applied. See also Sundqvist 2011a. Steinsland (2011a, 4) defined ideology as a “system of thoughts which expresses a world-concept and which permeates culture as a whole”. The concept referred to “the intentional use of myths, rites or narratives by individual groups with the aim to legitimise a certain type of rulership” (Steinsland 2011a, 4). For good overviews of the concept of ideology, see e.g., Geertz 1993 (1973); Heywood 1998; Freeden 1996; 2003; Žižek 2012 (1994). See also the articles in Žižek (ed.) 2012 (1994), Mapping Ideology. I will not use “ruler ideology” as a central concept or as a kind of super structure in the present study. Instead I will apply the term “religious strategy/-ies for rulership” since it fits my regional and contextual perspectives well. See also Sundqvist 2012b, 234–236. The concept of “ideologic strategies” has been used in political science. See e.g., Schweitzer 1962. On the concept “strategy”, see further below. 40 In present study the concept “rulership” refers to “a position in which one rules or has sovereignty over others”, while the term “ruler” refers to “a person who rules or governs; a person who exercises dominion or controlling power over others”. A ruler could appear on different social levels and be of different genders. It could be a king, a queen, an earl, a local chieftain, or a powerful female etc. See Sundqvist 2002, 11. 16 CHAPTER 1 king”, i.e. an overlord or a “king of kings”, who controlled several groups living in a wide area. It can also be used when investigating ideological aspects of powerful women (dróttning, húsfrú, gyðja), appearing on different social levels in society. Another advantage of this concept is that it can easily be used in the plural. Thus in one and the same cultural context, such as ancient Scandinavia, we may speak about different religious strategies for rulership. Most important for the present study is that this concept is more suitable when applied to the theme of ceremonial buildings than the term “sacral kingship”. The noun authorization is related to the term authority (see below). It refers to the action or fact of authorizing or being authorized. In addition, the related term ruler-empowerment will occasionally be applied as well as the concept of legitimacy. The latter concept was coined as a scientific tool by the sociologist Max Weber and has since played an important role in political sociology and political science.41 Scholarly treatments have often emphasized the subjective aspect of legitimacy; for there is no such thing as objective legitimacy. Whether the authority, the institution, the action and so forth are to be considered legitimate always depends on the beliefs of the acting subjects. Legitimacy may be derived from different sources, such as law, charisma, tradition, history, democracy, religion, etc. These sources of legitimacy may be widely shared by different groups in a society as common values and qualities, but they may also lead to conflicts. Religion may legitimate non-religious as well as religious actions and institutions. For instance, the ruler’s divine character may legitimate both his role as military commander and his function as an officiant in the public cult. Closely connected to the concept of “legitimacy” is the concept of “authority”. In previous scholarly traditions authority was considered as an entity. Weber introduced an ideal typology with three categories: traditional, charismatic and legal-rational authority.42 In agreement with later generations of scholars I view authority from a more discursive and contextual perspective.43 Authority is always established by means of the premises that are prevalent in a specific context. It may be achieved by means of law, charisma, tradition, history, democracy (see above), but also in many other ways. When using the concept of strategy I refer to a more or less deliberate plan of action or policy designed to achieve a major or overall aim.44 Tentatively

41 Weber 1971 (1922). See also Adams & Mikelson 1987; Kehrer 1998. 42 Weber 1971 (1922). Cf. Waida 1987. 43 See e.g., Bloch 1989; Lincoln 1994. 44 Such strategies are of course embedded into the ruler’s understanding and interpretation of the situation. The assessment of the ruler’s agency was also embedded in their Introduction 17 there are four possible religious strategies for raising the political leaders above common people, gaining legitimacy and political authority appearing at the ceremonial buildings. It should be noted that there are no watertight bulkheads between the strategies; they actually overlap each other. They could appear in different combinations at different contexts. When the individual strategies appear in combinations they may actually support each other in order to achieve the major aim, i.e. to gain legitimacy, political authority or empowerment:

(1) By means of the specific relation to the mythic world. Legitimacy and/or political authority may be achieved by means of the ruler’s close relation to the mythic world. The character of this relationship varies in different contexts. The following are some types of relations that may be tentatively discerned, although it is possible that additional types of relations exist: (a) the ruler may be considered divine by his or her own means and power; (b) the ruler may be seen as an incarnation of the deity or the mythic being, and represent him/ her in ritual contexts. In such cases the ruler may be worshipped as a mythical being, although his/her power actually derives from the divinity or the mythical being; (c) the ruler may be considered as a descendant of a mythical being; (d) the ruler may be the deputy or delegate of the divinity on earth, i.e. just a weaker variant of types a–c; (e) the ruler may be the most important mediator in the relationship between the human and the mythical worlds, i.e. he or she may act as the (highest) cult leader (see below). As noted, the list consists of a hierarchy, where the “sacredness” of the ruler seems to be more intensified in the upper half (a–c), while the lower part (d–e) reflects relations which put less stress on the “divine status” of the ruler.45 The main purposes of the ruler’s close relation to the mythic world are in all cases to mark him/her off from the rest of the community and to give his/her authority a divine sanction. The ruler’s close relation to the mythic world may be expressed in several senses at ceremonial buildings and sanctuaries. Certain narratives pro- claiming these ideas, such as myths or legends, may be recited under ritual observation, for instance by cultic leaders in these buildings and places. Such performances probably had a striking effect on the audience. There were also other ways of communicating these ideological notions. The seat of the ruler in the ceremonial building, for instance, may be construed as an interface with

cosmology or religious worldview. In some contexts the rulers may have been more tactical than strategical, when aiming at short-term aims, or when having to improvise for example when facing situations of crisis. 45 For a similar typology, see Köhler 2004. 18 CHAPTER 1 the Other World, i.e. a liminal and ritual place intended for people who had a specific relation to the divinities. There may also be other ritual objects (e.g., cultic images) or structures, appearing either inside or outside these types of buildings, which may indicate that the ruling family had a certain relationship to the divine world. The sanctuary including the ruler’s seat, may, for instance, be deliberately constructed as a reflection of the cosmos, or arranged in accordance with the mythical topography. Thus the sanctuary emphasizes that the ruler was situated close to the centre of the world (axis mundi), perhaps even inside the mythical world, where divine beings dwelled and could be encountered.

(2) By performing central roles in (religious) rituals. The ruler’s legitimate right to exert power in different fields of society is also related to his/her roles in cults, (religious) rituals and ceremonies. These roles, as noted above, may be connected to his/her specific relation to the mythic world. Since the ruler is the mediator between the human and the mythic world, he or she is also the one who must perform the sacrifices and so on. The rituals of the ruler can be described as “political rituals”, i.e. they comprise those ceremonial practices that specifically display and promote the power of the ruler.46 They do not simply give form to the power, they actually construct it, and orchestrate the ruler’s position in society and the cosmos. Coronations, consecrations and inaugurations, for instance, are important rituals for such constructions. These types of performative actions often take place in a ceremonial building at the cultic site. Most likely the properties of such “sacred rooms”, and also the architecture of the rooms/places themselves, contribute to creating the dignity of these “political rituals”. The sanctuary is thus the perfect arena and setting for such activities, which may include processions, ambulatory rituals, drinking ceremonies, as well as gift-giving ceremonies and the like.

(3) By using (religious) symbols. In the inauguration ritual the ruler may receive certain (religious) symbols which have a similar intention to the rituals of the ruler, i.e. to raise the ruler above his/her fellows, and/or to relate him/her to the mythical world. These symbols of power may consist of certain clothes, insignia and other paraphernalia. They may also be constituted by specific geographic places and sites, cultic buildings and halls, ritual structures and seats, monumental graves and so on. Ceremonial buildings, for instance, are sometimes located (symbolically) at elevated places in the landscape, indicating

46 Bell 1997, 128–135. Introduction 19 dignity and sovereignty, but also a position in the universe. The monumental architecture of these buildings, the interiors, including the ritual objects, can thus be concrete or symbolic expressions of power. Like the “political rituals” these buildings likewise have the function of displaying and demonstrating the political power, and they construe borders between the ruler and the rest of the society.47 These symbols attempt to demonstrate that the values and forms of social organization are naturally derived from the way the world or the cosmos is organized. Therefore symbols (and rituals) are more effective than coer- cive force when securing a particular social order and hierarchy.

(4) By controlling the cultic organization or institution. In order to carry out the strategies shown above (1–3) efficiently, the ruler must control the cultic organization in different ways. The ruler may, for instance, be considered as the owner or the manager of the cultic sites and ceremonial buildings. Thus his or her involvement is necessary for performing the public cults. He or she may also be the one who controls and organizes the cultic leaders at these sites, and sometimes he or she may even be regarded as the highest cultic leader. In such cases the ruler may have a clear view of the activities taking place during public gatherings, and give priority to ones that may be advantageous to the political power. The ruler may also, in different ways, encourage or even force the members of the cultic community to take part in the common religious feasts at the sanctuaries. The organization of cult may also be seen as one of the ruler’s duties. He/she may only be legitimate as long as he/she maintains and secures the cultic activities.

Even if this model of religious strategies for rulership is mainly adjusted to ancient Scandinavian condition, it may be used in cross-cultural studies, where both resemblances and differences are sought. The advantage of it is that it can take both regional and temporal variations into consideration, not only between two or more cultures, but also within one and the same culture. Therefore it is very useful in the present study, where a contextual-regional approach will be applied to the ideological contexts of ceremonial buildings and sanctuaries in pre-Christian Scandinavia. It will be shown below that some of the strategies described in the model can be attested there, but with some regional and contextual variations.

47 Bell 1997, 129–135, 155–159; Gilhus & Mikaelsson 2003, 132ff. 20 CHAPTER 1

1.3 Approach and Methods

As noted above, previous research on ruler-ideological aspects of ancient Scandinavia has been impaired by some methodological and theoretical problems, especially when proceeding from the phenomenological concept and theory of “sacral kingship”. Therefore a new approach was suggested above, which includes some general reorientations, for instance, (1) a change from global comparisons (looking for similarities only) to a “new comparativism” (with a focus on both resemblances and differences/variations); (2) a change from phenomenological generalizations to more localized contextualizations.48 When previous generations of scholars investigated ancient Scandinavian religion, ruler ideology and sacral kingship, they often proceeded from written sources, without considering issues of time and space.49 While archaeologists often observed regional differences and processes of change, the constructions of historians of religions have often been static and abstract.50 In contrast to these previous attempts, the present study has the ambition to bring contextual and processual perspectives into the analyses, where regional variation, development over time and social-political differences will be taken into consideration.51 Therefore ceremonial buildings and their religio-political contexts will be studied and compared in three different regions, where both similarities and differences/variations will be noted:

48 See also Rydving 2006. 49 See e.g., Å. V. Ström 1959. 50 See the criticism against such perspectives in Rydving 2005; 2006. 51 My perspective in this study is thus in agreement with some intentions visible in the new theoretical debate about the general history of religions. In this debate it has been noted that religious beliefs and expressions may vary within one and the same culture, because of contextual factors such as social aspects, gender, regional differences, historical circumstances, text genres, ritual contexts and sitituations etc. etc. See e.g., Hultkrantz 1956; T. Olsson 1985; Paden 1994, 55f.; Gilhus & Mikaelsson 2003, 53–56; Andrén, Jennbert & Raudvere 2006, 13f. The question of diversity in terms of gender, social status, locations and spaces is also noted in the book More than Mythology (ed. Raudvere & Schjødt 2012), see particularly Nordberg 2012; Schjødt 2012; and Sundqvist 2012b. My theoretical base is also to some extent inspired by some perspectives recently developed by Kim Knott (1998; 2009; 2014) and others on “religion, space and locality” (e.g., T. Blomkvist 2010). A local perspective on religion refers to the scientific point of procedure where religion is studied in terms of the specific conditions and contexts of a locality. This means, for instance, that a researcher investigates a certain religious centre or tries to explain how the religious compositions and interactions are constituted in a town or a region. Introduction 21

‧ The area of Lake Mälaren, in eastern central Sweden. ‧ Trøndelag and northern Norway. ‧ Iceland.

The reason these regions have been selected is that they are well documented in both written and archaeological sources on ceremonial buildings and that they reflect a variety of social-political structures and religious ruler strategies.52 They are also associated with interesting place-names, which have the ability to illuminate some of the questions posed in this study. They may therefore be considered as important regions during both the Viking Age and the Early Middle Ages.

1.3.1 Methodological Considerations: “Direct” and “Indirect” Sources When investigating religious strategies for political authorization appearing at Late Iron Age ceremonial buildings and sanctuaries in the Mälaren region, Trøndelag and Iceland, some methodological considerations must be made in the use of sources. A fundamental problem in this research is the lack of contemporary written sources. Many of the traditions relating to religious practice and sanctuaries in, for instance, Viking Age Norway or Svetjud are fixed in texts composed hundreds of years after the conversion of Scandinavia. The authors of these sources were thus not adherents of the ancient religion. Most of them were Christians living in Iceland in the High Middle Ages. The distance in both time and space between the texts and the actual events, rituals, beliefs and so on creates methodological problems for the historian of religions. When discussing sources, simplified categories such as pre-Christian and Christian sources are often used. The conversion of Iceland and Norway is usually dated to c. 1000. This date has been the crucial point when estimating the source value of individual texts. Yet the process of Christianization was very complicated and so are the sources. Christians and pagans existed side by side for a long period and influenced one another, and the sources sometimes reflect features of acculturation or syncretism.

52 The reason I have chosen not to take southern Scandinavia and Denmark into consideration in this study is twofold: (a) there are few written sources informing about ceremonial buildings from this area; (b) this study needs some kind of delimitations. Since much interesting archaeological evidence referring to ceremonial buildings has been found in southern Scandinavia, I will nonetheless use these sources as comparative materials. 22 CHAPTER 1

By proceeding from the principle of “insiders” and “outsiders”, a more reasonable theoretical and methodological framework for treating the sources may be offered.53 The sources of ancient Scandinavian religion may be divided into two major groups: direct and indirect sources.54 Archaeological finds are direct sources. They were created by people who themselves were part of ancient Scandinavian society and culture. These materials inform us about the Late Iron Age culture and religion in a direct way from an “insider’s” perspective, and have not been distorted by tradition and ideology as the medieval written sources. However, source criticism and other disciplines beside archaeology, such as anthropology, toponymastics, philology, history of religions must help to interprete and give meaning to them. Post-holes, for instance, are mute and must be interpreted by the archaeologist. A ceremonial building is thus an interpretation argued for with all kinds of source-critical aspects undertaken “along the road” and information from auxiliarian disciplines. Similarly, some surviving skaldic and Eddic poetry as composed in the Viking Age by “insiders” are direct sources, although they were not written down until the Middle Ages. The world view and ethics in this poetry are obviously not Christian, and the medieval Icelanders regarded them as pagan.55 They can thus be considered as oral art forms from the past. The formalistic language in these poems, rhymes (for instance the skothending and aðalhending), kennings, and their metres could have allowed them to retain their original shapes for centuries. Especially skaldic poetry can be regarded as based on a firm oral tradition before they were written down. Else Mundal states thus: “The written skaldic stanzas give us in many ways direct access to the oral forms behind.”56 Some of these poems must therefore be treated as direct sources when reconstructing the old religion, even if they appear in medieval manuscripts. The skaldic poems are, however, often complicated in a linguistic sense, and some of the kennings make them hard to understand and/or interpret. In many cases we would not be able to interpret them without Snorri’s texts and other medieval sources. In this study they will nevertheless be considered to have a

53 On the problems of “insiders” and “outsiders” in the study of religion, see Riesebrodt 2007 and Stausberg 2009. 54 See Hultgård 1996, 25–28; 2003b, 436–37. Jens Peter Schjødt (2012, 263–265) uses the concepts “contemporary” and “medieval sources”. My division of sources is actually a simplification of the actual case. There are sources which fit badly into this categorization. The Eddic poems, for instance, could perhaps be classified as a third group, since the dating of them often is so uncertain. For good and more detailed overviews of the sources, see e.g., Steinsland 2005, 35–66 and Schjødt 2008, 85–107. Cf. Sundqvist 2002, 39–62. 55 Meulengracht Sørensen 1991a, 225. 56 Mundal 2008, 1. Introduction 23 high source value. Also most of the Eddic poetry must be considered as direct sources, even if there are some exceptions.57 It must essentially be seen as memorised oral literature, “although it must have been more open for improvi- sation and changes than skaldic poetry”.58 Most of the Eddic poems have thus been changed during the oral transmission, but their fundamental content may have been mostly pre-Christian. Place-names and runic inscriptions are further direct sources, since they derive directly from the Late Iron Age society. They give information about the old rituals and belief systems from an insider’s perspective. However, scholars working with these source categories must also face some serious source critical problems.59 Anyhow, these sources contain linguistic materials that sometimes have escaped distortion by secondary traditions. The indirect sources are made up of prose narratives and descriptions transmitted by people who were not adherents of the ancient religion. Their authors lived outside the Viking Age society and their writings give only indirect information about the ancient Scandinavian cult system. Oral traditions appearing in these sources have often been reworked by rhetorical embellishment, artistic inspiration, ethnographic clichés, and subjectivity. Most of the sources belong to this category. They comprise the Sagas of Icelanders from the Middle Ages, Kings’ Sagas, medieval laws from Sweden, Norway and Iceland, and ecclesiastical texts in Latin, for example, written by Adam of Bremen and Rimbert, Arabic reports, and finally as historiographical accounts from the first century AD to the Middle Ages, such as those of Tacitus and Jordanes. The mythical and historical writings of Snorri Sturluson (1179–1241) must also be considered as indirect sources, since they were written by a Christian more than two hundred years after the official conversion of Iceland.60 Snorri’s texts have therefore been felt to be almost useless when discussing ancient Scandinavian religion and society.61 This type of source-critical approach has deep roots in the study of Old Norse philology and religion as well as in the general study of history.62 Sophus Bugge was one of the first to claim that the Old Norse myths and heroic narratives transmitted by Snorri were

57 Individual Eddic poems may have a medieval background, but many of them were probably composed during the Viking Age. See Mundal 2004, 222. 58 Mundal 2008, 2. Cf. Clunies Ross 1994, 20–33. 59 See e.g., Vikstrand 2001; McKinnell & Simek 2004; Düwel 2008. 60 On Snorri’s person and authorship, see e.g., Vésteinn Ólason 2008. 61 For example Baetke 1964. 62 See e.g., Lauritz Weibull 1948; Curt Weibull 1964; Baetke 1950; 1951; 1964; Holtsmark 1964; Olsen 1966; Düwel 1985. 24 CHAPTER 1 influenced by medieval Christian thinking.63 Eugen Mogk argued that Snorri in his Edda composed “mythological novellas” by reworking the old traditions.64 These novellas had no value when reconstructing pre-Christian religion except in cases where Snorri cited old poetic traditions. Even if Mogk exaggerated his argument, his scepticism was instrumental in establishing the modern treatment of Snorri’s texts.65 The philologist and historian of religions Walter Baetke, for instance, argued that Snorri revised the mythical traditions in accordance with learned medieval models and his own imagination. For example, Snorri frequently used medieval euhemerism when describing ancient myths. This method was a common feature of ecclesiastic apologetics, where the pagan deities were considered as ancient human rulers and thus not “real gods”. Because of this approach and other rewritings of the traditions, Baetke considered Snorri’s descriptions of myth to be unreliable when reconstructing pagan religion.66 A similar stand was also taken by the Norwegian philologist Anne Holtsmark and the German scholar Gerd Wolfgang Weber.67 The historian of literature Margaret Clunies Ross has also emphasized the euhemeristic explanations used by Snorri when dealing with pre-Christian myths and religion.68 Opposite to, for instance, Saxo’s interpretations, where paganism was explained as the work of Satan, Snorri linked Old Norse religion, myth and poetry to the culture of the classical world as well as Western European societies with his euhemerism. A critical attitude against Snorri is also reflected in more recent research concerning his historical writing in Heimskringla. It has been stated that Snorri never looks for “sociological explanations of the course of events, merely for human ones. He is not at all reluctant to adapt his sources as he sees fit and even write whole new chapters of his own in order—rightly or wrongly—to clarify connections and make his narrative more alive.”69 Even if Snorri’s writings have not satisfied the stern criteria of source criticism, they should not be completely discarded as sources of knowledge for Viking Age religion, mythology and culture. Some of them incorporate oral

63 Bugge 1881–89. 64 See for instance Mogk 1923; 1932. 65 Cf. Lindow 1988, 285. 66 Baetke 1950; 1951; 1964. 67 E.g. Holtsmark 1964 and Weber 1987 and 1994. 68 Clunies Ross 2006 and 2008. Heinrich Beck (2013) argues that the medieval writers also applied other models than euhemerism and idolatry when they were engaged in transmission of their own mythology. Snorri, for instance, sought to establish correlations between the message of the Bible and Gylfaginning or Ynglinga saga. 69 Jónas Kristjánsson 1988, 175. Cf. Vésteinn Ólason 1998, 59. Introduction 25 traditions which reach back to the Viking Age.70 Some traditions have been reworked by Snorri and other medieval writers. This reworking is no reason to dismiss them altogether. Snorri and the medieval writers probably had access to more direct sources than the modern historian has, and most likely they were better suited to interpret them.71 Even if Snorri’s texts and the sagas themselves are dated to the thirteenth century, their meaning and content may be much older. By using a comparative method which combines source criticism and philology, ancient structures may be detected in the materials—structures which can be interpreted by setting them in a wider context of sources—and developed alongside knowledge which we have from the auxiliary disciplines, that is, archaeology, runology and toponomastics.72 When such comparison is not possible (due to scarcity of direct sources) the absence of similar ideas and customs in contemporary Christian culture is here taken as a support for the reliability of medieval statements on ancient Scandinavian tradition.

1.3.2 Some Methodological Problems The present study will mainly take its point of departure from three texts belonging to the category of indirect sources when treating the pre-Christian ceremonial buildings, religion and rulers: (1) Adam of Bremen’s Gesta Hammarburgensis ecclesiae pontificium IV, 26–27 for Uppsala and Svetjud; (2) Snorri Sturluson’s Hákonar saga góða 13–18 for Lade, Mære and Trøndelag; and (3) Eyrbyggja saga 3–4 for Þórsnes and Iceland. These texts will be presented in more detail in chapter 4 below and estimated from a source-critical point of view. In this work the spatial and temporal dimension of the sources

70 On the transition process from oral tradition to literacy in Old Norse society, see e.g., Clunies Ross 1994; Quinn 2000; Gisli Sigurðsson 2004; Brink 2005; Mundal 2008; T. M. Andersson 2008; Lincoln 2012 and 2014. 71 Preben Meulengracht Sørensen (1991b, 244) made the following observations when considering Snorri’s narrative about Hákon the Good and sacrifices in Trøndelag: “History is not the data extracted from the sources, but the syntheses created out of the data, and historical writings are not the reestablishment of past reality, but the creation of historical probabilities. In this perspective the medieval narrative historiographic texts should be considered primary sources, since their authors had better access to authentic sources than the modern historian and were presumably better suited to interpret them.” 72 See e.g., Steinsland 2000, 16–18, 101–29; 2005, 35–66. Several contemporary scholars have defended the value of Snorri’s writings and other medieval prose texts as sources for ancient Scandinavian religion and culture. See particularly Georges Dumézil 1986 (1948), 61–83). Cf. Hultgård 1993 and Dillmann 1997; 2008. 26 CHAPTER 1 is important.73 Therefore, archaeological finds and place-names also play a significant role, since they unquestionably have their provenance in the three regions observed. The archaeological sources may also be connected to specific periods. A methodological problem in the present study is the relationship between the different source categories.74 Different kinds of sources, such as texts (and different text genres), archaeological finds and place-names, may admittedly together illuminate different aspects of the ceremonial buildings/cultic sites and ideological strategies in different investigation areas. However, even if these different types of sources on one level correlate to common phenomena at a certain place, they also reflect various discourses. The contexts and the creation of the texts and the archaeological objects/monuments are almost never identical. They may for instance reflect different social conditions. There may also be chronological gaps between them. On the other hand, they may sometimes function as complementary to each other. By means of the archaeological finds we can find chronological layers, something which is hard to find in the written sources. The texts, for their part, may contribute to the interpretations by creating more concrete hypotheses. One problem in the present study is that the different source types used in the investigation have an uneven distribution. In the Mälaren area there are few contemporary written sources in the native language referring to pre-Christian ceremonial buildings. The most important written sources for this region are written in Latin. There are, however, plenty of archaeological finds reflecting ceremonial buildings in the middle of Sweden. In western Scandinavia the position of source types has the opposite distribution. In Trøndelag, and particular in Iceland, we have good access to Old Norse traditions informing us about halls and cultic houses. However, we have less archaeological evidence of such buildings there compared to the area of the Svear. The uneven position of source types must thus be taken into consideration when discussing and answering the questions posed in this study.

1.3.3 Materiality and Some Further Methodological Issues and Considerations Recently religious studies, as well as the human sciences more broadly, have shown a growing interest in what the study of materiality may offer for the

73 On spatial studies of religion, see e.g., Knott 2014. 74 Cf. Andrén & Jennbert & Raudvere 2006. Introduction 27 understanding of religion.75 Since the present study is concerned with sanctuaries, ceremonial buildings, ritual structures and objects, i.e. different “things” that constitute the materiality of religion, some aspects related to this new approach should not be overlooked. When treating “things” previously in the history of religions, the focus has often been on the abstract meanings and ideas behind them. In recent anthropology and religious studies the living religion and its material dimensions are seriously taken into account from an emic perspective, for instances, at places “where powder is power, where costumes allow access to other planes of existence, and where legal documents may not primarily concern reason or argument”.76 Instead of dismissing the informants’ accounts as imagination or elaborate metaphors, scholars have tried to interpret them from a novel theoretical understanding, often called “the ontological turn”, “the material turn” or “the stance to materiality”. Rather than immediately assuming that “things” signify, represent, or stand for something else, the new approach is to take the materials encountered in the field as they present themselves.77 Materiality is thus the way the gods or the mythical world, for instance, are manifested and defined in the first place; how they are present.78 The sanctuaries, the images and the ritual objects are religion; they do not have to be reduced to semiotic signs or representations of abstract religious ideas for the believer.79 The new approach thus aims at considering materiality directly, not vicariously through the quest for immateriality. According to this line of reasoning, it is important to notice that immateriality can only be expressed through materiality. The sharp dichotomy between materiality and immateriality must actually be dissolved and refused in the first place. There are no pre-objectified forms. Birgit Meyer, David Morgan, Crispin Paine and S. Brent Plate summarize the material study of religion thus in the journal Religion:

Materializing the study of religion means asking how religion happens materially, which is not to be confused with asking the much less help- ful question of how religion is expressed in material form. A materialized study of religion begins with the assumption that things, their use, their valuation, and their appeal are not something added to a religion,

75 See e.g., Miller 2005 and 2010; Meskell 2005; Henare et al. 2007; Meyer et al. 2010; Engelke 2012; Stausberg 2012; Carp 2014. 76 Henare et al. 2007, 1. 77 Henare et al. 2007. 78 Cf. Engelke 2012, 213. 79 Cf. Miller 2005. 28 CHAPTER 1

but rather inextricable from it . . . Religion is not a pure realm of ideas or beliefs that are translated into material signs. The material study of religion avoids reifications that identify ideas or dogmas or individual people as the irreducible core of religion.80

In the present study the methodological and theoretical parts of the “stance to materiality” will be followed to some extent. It must be considered as a fact that religion includes objects, artefacts, images, gestures, words, ways of speaking, i.e. the materiality of religious life. For some people in Late Iron Age society the sanctuaries, the burial mounds, cultic images, the sacrificial blood, and the ritual structures were probably experienced in a direct way, without reflecting on symbolic dimensions, deeper meanings or abstract aspects. For them the dichotomy between materiality and immateriality never existed. Most likely, however, some people who had access to religious and cultural knowledge also interpreted the ritual objects at the sanctuaries as semiotic signs reflecting more abstract religious and ideological aspects.81 That religion, culture and objects in ancient Scandinavia often were symbolically encoded may be attested in several ways. One example is the language and diction in the Viking Age skaldic poetry, and the metaphors called ON heiti ‘appellations’ and kenningar ‘kennings’.82 When studying and interpreting religious strategies for rulership at ceremonial buildings in ancient Scandinavia we cannot reject the possibility that some sanctuaries, ritual constructions and objects were intentionally made as signs referring to more abstract ideas, such as myth and cosmology, at least by people who had access to cultural and religious traditions and probably sometimes also were the producers of these sanctuaries. Most of the people representing the elite probably had the ability to interpret the materiality of religion in that way. For ordinary people, however, these ritual structures were probably perceived in a more direct way. Perhaps ontologies related to things could

80 Meyer et al. 2010, 209. 81 That some people apprehend “a reality” behind materiality is actually accepted by the “materiality scholars”. Daniel Miller (2010, 73), for instance, writes thus: “So one can correctly label Hinduism as polytheistic, monotheistic and even atheistic, partly because each is seen as appropriate to the capacity of certain kinds of people to apprehend the ‘reality’ behind mere materiality. Ordinary people are ordinary partly because they require more materiality. Great sages are great partly because they can apprehend more immateriality.” 82 On kennings, see e.g., Meissner 1921; Marold 1983; Jónas Kristjánsson 1988, 87–90; Bergsveinn Birgisson 2007, 40–72. Introduction 29 differ for one and the same individual on different occasions and in various contexts. The things themselves may have dictated a plurality of ontologies. In the present study, hence, I will argue that sanctuaries and images may represent or refer to a mythic world and deities, although the materiality of religion, in the first place, probably gave form to the immaterial ideas. Most likely the ritual structures and ceremonial buildings of the elite were symbolically encoded in some way, since the entire Scandinavian culture was imbued with metaphors and “kennings”. These symbolic dimensions were also part of the religious strategies used for legitimacy. Whether all people approaching these sanctuaries noticed them is, however, uncertain. It was most likely due to their gender, age, social status and also if they had earlier encountered such milieus.

1.4 Outline and Hypotheses

Since this study has a thematic outline, some repetitions of source instances are inevitable. As far as possible, however, such disturbing reiterations will be avoided. In what follows I will present the outline, the general hypothesis as well as some sub-hypotheses. In Part 2, chapter 2, the social, political, and historical contexts of the three investigation regions—the Mälaren area, Trøndelag and Iceland—will be discussed in more detail. I will argue there that the social and political structure differed in these areas, and it also varied over time. In both the Lake Mälaren area and Trøndelag the power structure was originally decentralized. However, as time went, a process of centralization and power concentration took place. During the early Viking Age overlords, warrior lords, High Kings, and powerful earls appeared in these regions, with ambitions to exercise domination and power over several local chieftains. In Iceland, however, the process towards a concentration of power never took place during the Viking Age. The early Icelandic society and its “constitution” were built on a decentralized system, where power was split between many minor chieftains. Why the settlers of Iceland never established a society with a king, a military or administrative centre is uncertain. In my opinion, these regional differences between the Mälaren area, Trøndelag and Iceland had consequences for the formation of the religious strategies for political authority and legitimacy. As indicated by a case study presented in chapter 3, the religious strategies for rulership differed in these regions in a substantial sense. The kings called Ynglingar and the Earls of Lade in Svetjud and Norway claimed, according to Old Norse sources and Latin texts, that their families had a divine descent from Freyr or Óðinn, while the Icelandic chieftain family called Þórsnesingar more humbly argued that they 30 CHAPTER 1 were the deity Þórr’s “dearest friend”, and therefore took care of his cult. The kings and the earls among the Svear and the Þrœndir thus had more bombastic terms than the Icelandic chieftains when expressing ideological matters. The centralizing tendencies which were observed in the societies of the Mälaren area and Trøndelag had thus generated a type of mythic strategy which was not as visible in the more decentralized Icelandic chieftain society. Chapter 4 will concentrate on ceremonial buildings, and the evidence of such buildings in the sources relating to the three investigation areas. A preliminary statement will be formulated, i.e. that more monumental Iron Age buildings (including rich contextual finds made of, for instance, gold) appeared in Svetjud and Trøndelag compared to the ones occurring in Viking Age Iceland. Monumentalization of power was thus more accentuated in the former regions (i.e. the Lake Mälaren area and Trøndelag). The general hypothesis for the present study is that the social-political conditions and other contextual aspects in the three investigation areas affected the use of religious strategies for gaining legitimacy and political authority observable in connection with the halls, cultic houses and sanctuaries in each area. There is a link between the political and historical conditions in these areas and the formation of religious strategies for rulership as they appear at public cultic sites. It is argued, therefore, and will be shown in Part 4 (ch. 11–15) of this study, that the ideological dimensions of the sanctuaries and the strategies for legitimacy and empowerment in Svetjud and Trøndelag often resemble each other, while the situation in Iceland to some degree diverges from the other regions. Some religious strategies at sanctuaries in the former regions included a monumentalization of power and rulership and were also related to the idea that certain royal or noble families of the Svear and Þrœndir claimed that they were of divine descent and thus raised above other people. Such strategies and mythic-cosmic notions were constituted by, for instance, gold foil figures and golden bracteates (with images of mythical beings), which often appear in the context of the ruler’s high-seat in the ceremonial buildings. Certain monumental graves may also be associated with such a strategy. Thus certain impressive (symbolic) objects or structures related to some of the ceremonial buildings and sanctuaries of Svetjud and Trøndelag emphasized a close relationship between the rulers and the mythical beings. In Svetjud and Trøndelag the halls and cultic houses were also used by the kings and earls as arenas for more specific royal rituals, which were never performed in Iceland. The royal inheritance ceremony performed by Ingjaldr, as presented in the introduction to this chapter, is an example of one such performative ritual. These actions could also be seen as ritual strategies for gaining legitimacy and constructing authority at the highest level in the Scandinavian societies. There Introduction 31 were also certain martial hall milieus in Norway and Sweden (and Denmark), where warrior lords with warrior bands appeared quite frequently. These halls were, for instance, decorated with weapons. Such martial hall buildings were less common in Iceland, since the institution of the comitatus did not exist there. In this connection, it is suggested that the implicitly martial ideology of the eschatological Valhǫll mythology was very important when constructing power in Svetjud and Trøndelag, but less essential among chieftains in Iceland. Such mythical traditions may have their Sitz im Leben in these hall contexts. In Part 4, but also in ch. 5 (Part 3), it is argued that regional differences also are observable in connection with the organizational dimension of ceremonial buildings. Inter-regional chieftainly sanctuaries, comparable to the ones in Uppsala, Lade, and Mære, never existed in Iceland. In my opinion this may be due to the lack of kings and earls in Iceland. In Svetjud and Trøndelag it seems as if there were ceremonial buildings and sanctuaries at several social levels in society, which were controlled and used by chieftains, earls and kings in order to gain legitimacy. Some of them were local, while others had a regional or inter-regional significance. In Iceland the ceremonial buildings were erected on the local farms and most of them played a role only for local groups. It was thus the local chieftains who organized public cult there at the local ceremonial buildings. There were no supreme leaders of Iceland, and no central power to organize cult at central sanctuaries for all Icelanders. I completely agree with Bjarni F. Einarsson when he states that: “Icelandic Iron Age society, with no king and no centralized power, did not need all-encompassing religious buildings or large monumental burials. It was the family of the farmstead that was of central importance.”83 In Part 3 (ch. 5–10) of this study, I will suggest that the strategies for political legitimacy at ceremonial buildings and sanctuaries in the three areas not only diverged, but to a large extent also resembled each other. In my opinion, the members of the lower elite in Norway and Svetjud also had religious functions as leaders of public cult. Some of these chieftains went to Iceland, because of the political situation and other conditions in their homelands. In Iceland they appeared as cultic leaders and erected ceremonial buildings in a similar way as they had done previously back in their homelands. We may reckon with several reasons for building sanctuaries. One reason may be related to economic aspects. It seems as if rulers and cultic leaders in all three investigation areas, and at different social levels in society, took tribute from the people who

83 Bjarni F. Einarsson 2008, 178. Terry Gunnell (2001, 14) has also emphasized some differences in the constitution of ceremonial buildings and rituals between Norway, Denmark and Sweden, on the one hand, and Iceland, on the other hand. 32 CHAPTER 1 attended the public cultic feasts at the sanctuaries. But mostly the sanctuary was used by the elite as a place where they could gain legitimacy by religious means, such as appearing on a stage which reflected a mythic-cosmic symbolism. I will give examples of such strategies in Part 3 below.

To sum up, when discussing the religious strategies for political legitmacy and authority at ceremonial buildings and sanctuaries in the present study, a regional-contextual approach is applied. These buildings and sanctuaries will be investigated in three regions, namely Svetjud, Trøndelag and Iceland. In this investigation similarities between the three regions will be sought in Part 3 (ch. 5–10), and regional differences in Part 4 (ch. 11–15). In Part 2 (ch. 2–4), the regional-contextual approach will be discussed further. The main sources will also be presented there. part 2 The Regional-Contextual Approach

∵

chapter 2 The Three Areas of Investigation and their Political Leaders

The purpose of the present chapter is to outline and discuss the social, political and historical contexts of the three investigation regions: the Lake Mälaren area, Trøndelag and Iceland. First I will discuss the development of political power during the Late Iron Age in the societies of the Svear and Þrœndir. In both the Mälaren area and Trøndelag a tendency towards power concentration and political centralization is visible in the sources. A central power emerged in both regions, which gradually led to more institutionalized kingdoms in the Late Viking Age or the Early Middle Ages. This development will be compared with evidence referring to Viking Age Iceland, where the social and political structure seems to have been quite decentralized until the breakdown of the Free State in 1262/64. In Iceland many chieftains ruled, while power in the Mälaren area and Trøndelag was focused on kings and earls.

2.1 The Kings in the Mälaren Area

Very little is known about the development and structure of the Late Iron Age society of the Svear. There are some mentions of the Svear in classical and ecclesiastical sources as well as in Anglo-Saxon traditions. Otherwise Old Norse texts, archaeological finds and place-names must form the base of information when seeking knowledge of these matters. Only a rudimentary image of the Late Iron Age Mälaren society can therefore be outlined. However, some general tendencies connected to the political, religious, ideological and social structures are accessible through some written sources.1

2.1.1 The Svear and their Society The Svear lived in an area called Svetjud (OSw Svethiudh is a compound constructed of Svear + *thiudh ‘people’ (cf. ON Svíþjóð)).2 We do not know the

1 See e.g. Rimbert’s Vita Anskarii, Adam’s Gesta Hammaburgensis Ecclesiae Pontificium, and perhaps Ynglingatal. 2 The name of the people “Svear” appears as early as in Tacitus’ Germania 44 (c. 100 AD). He calls them Suiones (= PG *sueonez > OSw svear). On Svetjud and the Svear, see e.g. P. H.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_003 36 chapter 2 geographic extent of this area, but it has been suggested that it comprised the Mälaren region, i.e. Uppland, Södermanland, Västmanland and parts of Närke. This area was delimited by a wide belt of forest towards the south and the south-west. Beyond these woodlands was Götaland, i.e. the land where the people called Götar lived. The unification of Svealand and Götaland probably took place late, perhaps not before the Late Viking Age or the Early Middle Ages (the eleventh or twelfth century).3 There are signs of an organized society in the Mälaren region during the Late Iron Age. Trade, handicrafts and import of high-status goods provide evidence of chieftainships as early as the Bronze Age. During the Middle and Late Iron Age the economy was based on cattle-breeding and cultivation. From ancient times Scandinavians were well-known for their skins and furs; Jordanes, for example, wrote about the Suetidi (the Svear) who supplied Romans with sap- phire furs.4 Adam also comments on their rich goods.5 Archaeological finds from the Migration Period show vigorous iron production in the northern part of the area.6 Handicrafts, especially gold-working, were highly developed. Lake Mälaren was surrounded by extensive arable lands. Fishing was important not only along the coasts, but also in lakes, rivers and streams. Hunting was essential for the food supply, but also for the production of fur pelts. In the Early Middle Ages, but probably also earlier, markets for the exchange of products produced in different areas were common. During the Late Iron Age there was strong economic growth in Svetjud, visible in trade centres, such as Helgö and Birka, and long-distance trade.7 This development led to an increase in population, observable in cemeteries and

Sawyer 1991; Sundqvist 2002; T. Andersson 2004; 2005; Arrhenius 2004b; Ljungkvist 2006; N. Blomkvist et al. 2007. See also the distinction between Svíþjóð and Svíaveldi in Snorri’s Heimskringla, as described by T. Andersson 2004. 3 Adam of Bremen distinguishes Götar from Svear, but paradoxically he also includes the area of Götar in Sueonia. See e.g. Nyberg 1984, 312–323. 4 Alia vero gens ibi moratur Suehans, quae velud Thyringi equis utuntur eximiis. Hi quoque sunt, qui in usibus Romanorum sappherinas pelles commercio interveniente per alias innumeras gentes transmittunt, famosi pellium decora nigridine. Hi cum inopes vivunt, ditissime vestiuntur. Getica 21. Jordanes is probably referring to arctic fox furs. See T. Zachrisson 2011b, 123f. 5 Omnia enim instrumenta vanae gloriae, hoc est aurum, argentum, sonipedes regios, pelles castorum vel marturum, quae nos ammiratione sui dementes faciunt, illi pro nihilo ducunt. Adam IV,21. 6 Hyenstrand 1974, 153f. 7 See Holmqvist et al. 1961–64, 29; K. Lamm 1999; Clarke & Ambrosiani 1993, 65ff. The Three Areas Of Investigation And Their Political Leaders 37 settlements.8 Local centres can be attested in this area through the presence of certain graves and place-names. Towns arose gradually during the Viking Age, e.g. Birka and Sigtuna. They were often important for the region and its economic life.

2.1.2 Society and Political Structure The political and social structure of the society of the Svear has been the subject of much research.9 When debating the existence of kingship, Iron Age Svetjud has sometimes been regarded as a centralized state.10 These assumptions have since been discussed.11 The theory, based on Beowulf and Ynglingatal, that “Sweden” became “Sweden” when the Svear defeated the Götar in the sixth century, is today more or less rejected. The emergence of a more stable kingdom is later and must be described as a long and unclear process.12 Historians have argued that the Middle and Late Iron Age society in the Mälaren region seems to have been a society split into several areas and ruled by more or less independent leaders residing at local centres.13 Archaeological materials, for instance, indicate the presence of an upper stratum at different locales in this region, perhaps reflecting several petty kingships.14 The boat graves in Uppland and Västmanland all reflect superior strata.15 Badelunda in Västmanland, for instance, has graves containing the bodies of females who may have had ruling functions, dating from c. 300 to 1050.16 Similar rich finds, but with males buried in boats, may be found at Vendel and Valsgärde.17 There are also plenty of great mounds with rich grave-goods all over the Mälaren region dating from the Late Iron Age.18 Finds show that sites such as Uppsala, Adelsö, Birka and Sigtuna played a highly significant role in the Late Iron Age.

8 It has recently been argued that the increase in population started as early as Roman Iron Age. See e.g. B. Gräslund 2004. 9 See an overview in e.g. Lindkvist 1995, 10–12. 10 See Lindquist 1944; Baetke 1964, 169f. 11 E.g. Lindkvist 1995; 2008; 2009; P. H. Sawyer 1991. 12 See e.g. P. H. Sawyer 1991; Hyenstrand 1996; Lindkvist 2008. Recently scholars such as T. Andersson (2004; 2005) and B. Arrhenius (2004b), have opposed a late dating of kingship in Svetjud. They argue that it should be dated at least to the Late Iron Age. 13 See e.g. Lindkvist 1995; P. H. Sawyer 1991. 14 See Hyenstrand 1996; Ljungkvist 2006. 15 See e.g. Hyenstrand 1996. 16 Nylén & Schönbäck 1994, 128. 17 Ljungkvist 2006, 65–72. 18 See e.g. Lindqvist 1936, 14f., 18–36; Hyenstrand 1974, 116f.; Ljungkvist 2006; Bratt 2008. 38 chapter 2

In the upper stratum of society, among the nobility, there were probably many officials supporting the petty kings and chieftains, such as cultic leaders (e.g. goðar) and warriors (rinkar, þegnar, karlar). Below the elite there were free farmers, craftsmen, sailors, and also unfree people, called thralls (þræll sg.). Stefan Brink has argued that the prehistoric society in Svetjud not only was divided in a vertical sense, but also in a horizontal, spatial one.19 The spatial or horizontal division of society was based on small settlement districts (Sw bygd), which in some sense were tied to some larger, regional formations (Sw land), which in its turn had ties with a social formation of an inter-regional kind (Sw rike). For each and every one of these settlement districts, regional formations, and inter-regional units, there were one or more focal sites which may be characterized as “central places”, i.e. “sites or small settlement structures that have had some function or significance exceeding the particular site or settlement, in other words, some kind of ‘power’ over a wider area”.20 The local rulers and chieftains of the Svear thus seem to have resided at the central places in the settlement districts during the Late Iron Age. Not only archaeological finds but also place-names help to identify such sites. Tuna names, for instance, refer to ancient centres in the settlement districts and they are often attached to boat graves with rich finds.21 It has been suggested that the distribution of the tuna names shows that such sites most likely played an administrative role in a central-place complex.22 Originally some tuna names may have just referred to farms of local chieftains, while others may have had a function as a central place within the administrative district.23 Specialists in place-names have even suggested that the Late Iron Age system of central places (including habita- tions) named Tuna, surrounded by one or more sites with sacred names, was transformed into royal sites at the end of the prehistoric period.24 “The clearly pre-Christian nature of this system can be regarded as an indication of an indigenous king in the Late Iron Age in the Mälaren Valley” (on cultic organization, see further ch. 5 and 14 below).25

19 Brink 1997, 404f. 20 Brink 1996a, 237. See also Vikstrand (2011), where several problems with previously elaborated models of social and cultic organization are taken into consideration. 21 The word tuna means ‘fence, enclosure’ or ‘fortification’. It sometimes appears in compound names with a theophoric name as the first element, e.g. Fröstuna, Torstuna and Ultuna. T. Andersson 1991; Brink, 1996a; Vikstrand 2000; SOL. 22 Hellberg 1984–85. Cf. Brink, 1999c. 23 Brink 1996a, 263f. 24 See mainly Vikstrand 2000. 25 Vikstrand 2000, 232. The Three Areas Of Investigation And Their Political Leaders 39

There is actually evidence indicating that the people living in the different settlement districts of the Mälaren region comprised a political unit including a common organization during the Late Iron Age. These regions, e.g. Uppland, Västmanland, Södermanland, may have had a common law, or better in prehistoric society: common legal customs with a law-speaker. These regions probably had common assembly places. The names of the “folklands” in the Mälaren area, including the concept of hund, may also indicate regional units and a common organization. The medieval “folklands” (OSw folkland sg.) of the Svear were called OSw Tiundaland, Attundaland and Fiædhrundaland.26 They were made up of several hund or hundare. On linguistic grounds, Thorsten Andersson has suggested that the hund division goes back to the Early Viking Age or perhaps even earlier.27 The younger term hundare appears in the Late Viking Age runic inscription at Vallentuna church (U 212) (c. 1050).28 This system may thus reflect some kind of common military organization in Late Iron Age Svetjud, also indicating a political unit among the Svear people.29 At the top of such system we may assume that some type royal institution was in charge (see further ch. 15).30 The term husaby, visible in place-names, may also be associated with a common royal organization in the Mälaren region with roots in the Late Iron Age.31 Admittedly the concept of husaby itself is probably secondary,32 denoting a medieval royal estate or farm belonging to the medieval crown lands called “Uppsala Öd” (Uppsala auðr). The connection of husabyar with Iron Age mounds and other ancient monuments, however, indicates that at an earlier stage these places may also have functioned as seats of rulers.33 It is of course possible that royal power confiscated these important chieftain farms during the Late Viking Age and turned them into crown lands. But it is also possible

26 See T. Andersson 1982; 1999a; 2000. 27 T. Andersson (2004, 8) writes: “Hund-indelningen, som ligger under hundaresindelningen som ett äldre skikt, bör ha genomförts under tidig vikingatid eller ännu tidigare.” See also T. Andersson 1999a, 11f. 28 The coastal regions were divided into Sw skeppslag (OSw skiplagh, skiplæghi). This area was called Roden. 29 The term hund denoted a district and a military unit ‘(group, army of) one hundred men’. 30 T. Andersson argues convincingly for early kingship in Svetjud. See T. Andersson 2004; 2005. Cf. Arrhenius 2004b. 31 Brink 1990b, 58; 1996a, 248–250; 1999c; 2000a; 2000b. 32 During the Middle Ages Husby in Långhundra, Uppland, for instance, was called Husaby Ærnavi (husabyernaui 1314). In more ancient periods the place was most likely called †Ärnavi. Perhaps it was an old cultic site. Elmevik 1995. Cf. Brink 1999c; 2000a; 2000b. 33 See Brink 2000b, 70f. 40 chapter 2 that they were connected even earlier to an ancient bona regalia system,34 and thus may have constituted a well-organized pre-Christian kingship in the Mälaren area. Written sources indicate, however, that during the eleventh century there were still several petty kings and chieftains ruling over local districts in Svetjud. In the passage about the Uppsala cult Adam of Bremen tells us that all groups of the Svear, from all districts (omnium Sueoniae provintiarum), gathered at Uppsala. Kings (reges) and peoples (populi) had to send their gifts to the “temple” there.35 In my opinion, this passage indicates that an early central religio- political power used the public cult and the important cultic site of Uppsala to take tribute from the chieftains and the people living in the settlement districts of Svetjud. I will develop this argument below in Part 3 and Part 4.

2.1.3 The Pre-Christian Kings and the Royal Site of Uppsala The early society in Svetjud thus seems to have been divided into several groups ruled by more or less independent rulers. The sources indicate however that, at times, these rulers and groups were united in a loose alliance under a federate leader or inter-regional High King, called “the king of Svear” (rex sueonum).36 It seems as if such a king or overlord could occasionally exercise domination and power (ON ríki) over several groups and later also over a territory. As time went, this alliance of groups was stabilized and a kingdom developed in the region of Lake Mälaren.37 It seems as if this kingdom was a highly structured political organization. The ninth-century “Svea kings” mentioned by Rimbert

34 See Hyenstrand 1974, 103–118, Cf. N. Blomkvist et al. 2007, 167. 35 Adam, IV,27. 36 Evidence of high kings of the Svear may be seen in the expression or title rex sueonum in the ninth century Vita Anskarii 19, 26. Adam of Bremen (1075) calls the tenth-century king Eiríkr inn sigrsæli, Hericus igitur rex Suedorum, II,38. See also scholion 24, which mentions Hericus rex Sueonum. The term kunungr (ON konungr, OSw kununger, konunger) is evidenced in Viking Age runic inscriptions (Ög 136; Sm 42; U 11; Vg 40). Peterson 1994, 28. We do not know what this concept exactly referred to in these early sources; Brink 1990a, 55. Neither is the etymology clear. It has been suggested that the word konungr is derived from the stem of ON konr ‘(distin guished) man, son’ (from PG *kunja- ‘family, kin, descent’), with the patronymic suffix ‑ung-, and thus means ‘son of a distinguished man’. Hellquist 1957, 496f. De Vries (1977, 326) has ‘Mann von göttlicher Abkunft’. In such cases the etymology indicates a strategy of legitimacy, whereby the ruler had to be descended from a noble or divine family. When I use the analytical concept of “king” in the present study I refer to a person who was at the top of the ruling elite. 37 See e.g. Vikstrand 2000. The Three Areas Of Investigation And Their Political Leaders 41 in Vita Anskarii (c. 870), for instance, were dependent on an assembly.38 Before the assembly, they summoned the foremost men of their realm to hear their advice.39 The Latin term princeps ‘first, foremost, principal man of the ruler’ corresponds perhaps to OSw høfþinger ‘chieftain’. These chieftains were probably regarded as members of the ruler’s council. One of them was called Hergeir. He was a counsellor of the king (consiliarius regis), and he was also called prefect (praefectus) of Birka, perhaps equivalent to ON jarl ‘earl’.40 This information from Rimbert may thus be seen as an indication of overlordship in early Viking Age Svetjud. The Svea king’s power seems, however, to have been restricted by the will of the assembly, as well as the will of council, i.e. the petty kings, earls or chieftains. The Viking Age kingdom of the Svear was a partly seaborne and loose empire, connected to the Baltic Sea.41 According to Wulfstan’s description of a sea voyage from the ninth century from Hedeby to Truso, several provinces connected to this sea belonged to the Svear: “And then after Bornholm we have the land, first Blekinge, Möre, Öland and Gotland being the land to port, and these lands belong to the Svear.”42 Perhaps the Svear had some kind of seaborne hegemony in the Baltic Sea, occasionally with territorial control. Rimbert mentions that the ninth-century kings of the Svear demanded and took tribute from the people of Curonia.43 Hence, the rulers were often some kind of war leaders supported by a warrior band (ON hirð) and perhaps also seaborne retinues (OSw leþunger; ON leiðangr) (see further ch. 15). Their political power was based on warfare, pillage and demanding tribute. It thus seems as if the Viking Age kings of the Svear were well-organized and that some features of “state institutions” existed in these early kingdoms, such as assembly, council and counsellors. Different types of written sources indicate that the Late Iron Age kings of the Svear also had a political, royal

38 Vita Anskarii 26–27. This thing institution most likely restricted the king’s power. According to Rimbert the Svea king Olof told Ansgar: “It is our custom that the control of public business of every kind should rest with the whole people and not with the king” (Sic quippe apud eos moris est, ut quodcumque negotium publicum magis in populi unanimi voluntate quam in regia constet potestate). Vita Anskarii 26. Trans. Robinson. 39 Vita Anskarii 27. 40 Vita Anskarii 11. Cf. Norr 1998, 159f., 165; T. Zachrisson 2011a, 100f. 41 Lönnroth 1977, 7–16; Lindkvist 2008, 669. 42 Þonne æfter Burgenda lande wæron us þas land þa synd hatene ærest Blecingaeg & Meore & Eowland & Gotland on bæcbord, & þas land hyrað to Sweon. The Old English Orosius. (Ed.) Bately, 16. Trans. Page. 43 Vita Anskarii 30. 42 chapter 2 and cultic centre in Old Uppsala.44 For instance, the Viking Age skalds quite often relate the rulers of the Svear (e.g. the royal family called the Ynglingar) to Uppsala,45 and the medieval saga authors describe them as Uppsalakonungr who sat on konungsstóll at Uppsala.46 Similar information appears in independent medieval sources from eastern Scandinavia, where we meet the expressions kunungr i Upsalum and upsala kunungr.47 Archaeology supports the assumption that Old Uppsala was a royal centre during the Late Iron Age.48 This site is one of the most complex archaeological monuments in Scandi navia. The finds are sufficient for it to be considered a high-ranking central place during the entire Late Iron Age. The royal mounds (including rich finds), boat graves, traces of a hall at Kungsgården and house terraces, the long row of posts, appearing both south and north of the place, exclusive handicrafts

44 See e.g. Lindkvist 2008, 669. About Uppsala, see B. Gräslund 1993; Arrhenius 1995; Duczko 1996a; Sundqvist 2002; 2013; Ljungkvist 2006; 2013; Göthberg 2008; Frölund & Göthberg 2013; A.-S. Gräslund 2013; Vikstrand 2013; T. Zachrisson 2013; Beronius Jörpeland et al. 2013. 45 In the ninth-century skaldic poem Ynglingatal the Norwegian skald Þjóðólfr says that Aun st. 18 (13) and Aðils st. 21 (16) died at Uppsala, while Óláfr trételja st. 26 (21) is mentioned as sá áttkonr frá Uppsǫlum “that relative from Uppsala”. Dómarr burned at Føre (við Fýri brann) st. 9 (6), i.e. a place close to Uppsala. About Vanlandi it says: sá brann á beði Skútu “at the banks of [possible] Skutån” st. 6 (3). Skúta is most likely a small tributary of the River Fyris located in the parish of Skuttunge (OSw Skuttungi), just a couple of miles north of Old Uppsala. Cf. Calissendorff 1986, 35, 55; Vikstrand 2004b. References from both Ynglingatal and Ynglinga saga are cited from Bjarni Aðalbjarnarson’s edition of Heimskringla I in Ísl. Fornr. 26. The numbers of stanzas in Ynglingatal are counted from this edition with the numbers of Noreen’s edition in brackets. There is actually a new edition of Ynglingatal and some skaldic poems appearing in the Kings’ Sagas (see SkP I–II). This edition appeared for me unfortunately at such a late stage in the writing process that it was not possible for me to use it in this book properly. On Ynglingatal, see SkP I, 3–60. 46 According to Snorri Sturluson, several Svíakonungar of the Ynglinga dynasty were connected with Uppsala and the surrounding area. See e.g. Ynglinga saga (c. 1230), in Heimskringla I, Ísl. Fornr. 26, 25, 28f., 31f., 41, 45–49, 52, 57–59, 63, 65–67, 72–74. So also historical kings, see e.g. Óláfs saga Helga, in Heimskringla II, Ísl. Fornr. 27, 109f. As early as c. 1120 Ari inn fróði explicitly called an ancient Svea king: Aðísl at Uppsǫlum. See Íslendingabók, Ísl. Fornr. 1, 27. 47 The judicial text Landamæri I mentions a kunungr i Upsalum. SSGL 1, 67f. Landamæri II has the expression Upsalakunungr. The provincial Gotlandic narrative, Guta saga, which also is independent of Icelandic traditions, tells about upsala kunungr. Guta saga. (Ed.) Peel, 6. 48 See e.g. Lindqvist 1936; Duczko 1993b; 1996a; 1996b; 1996c; 1996d; 1997; Ljungkvist 2005; 2006; 2008a; Ljungkvist et al. 2000; Rahmqvist 1986; Göthberg 2008. A history of research and bibliography about Uppsala has been written by Duczko 1993a. The Three Areas Of Investigation And Their Political Leaders 43 indicate the presence of a ruling stratum at least from the sixth century up to at least the Early Viking Age. Excavations at the church plateau in 2005 indicate that Late Iron Age settlements also existed under the church.49 Huge post-holes and hearths were discovered, indicating a large house dated to the Migration Period. Remains of (pig) bones were also found in two hearths, probably reflecting sacrificial meals from the same period.50 Archaeological analyses show that the site has been settled since the Roman Period. It was an important site from at least the Migration Period and probably up to the Middle Ages, though it may have experienced some troughs during the Late Viking Age, when the kings of the Svear converted to Christianity. Arguments from settlement history, however, support the notion that Uppsala remained a royal site in the Viking Age and the Early Middle Ages. During the twelfth century Old Uppsala was regarded as crown land and one of the largest villages in the region. Sigurd Rahmqvist’s analysis indicates that the western part of Old Uppsala was a royal demesne even before the Middle Ages. The ancient name of the crown lands, Uppsala Öd (ON Uppsala auðr ‘the wealth of Uppsala’), suggests that Uppsala was the very core and the centre of these estates.51 Old Uppsala maintained its symbolic significance during the Early Middle Ages. In 1164 it became the archbishopric of the new church province in Sweden. The remains of the royal saint Eiríkr were placed in the newly built cathedral there.52

2.1.4 Religious and Political Changes: The First Christian Kings The first known missionary to visit Svetjud was Ansgar.53 He came to the trading town of Birka, on the island Björkö in Mälaren, sometime around 829 or 830. Most likely his mission was a double one, to Christianize the pagan Svear and to establish a relationship between Emperor Louis and the king of the Svear. It seems as if these intentions failed, as no stable Christian congregation was established in Svetjud, and most likely the king’s relation to the empire did not endure long.54 Gradually, however, Christianity was rooted in Götaland,

49 Göthberg 2008. 50 Frölund 2007. 51 Rahmqvist 1986. 52 During the summers of 2011, 2012, 2013, 2014 and 2015 new excavations were conducted in Old Uppsala, resulting in many new and important finds. Since these finds only partly have been published yet, I will only occasionally refer to them. See, however, Ljungkvist 2013. 53 See Vita Anskarii. 54 Archaeological evidence indicates, however, that some kind of a local Christian congregation may have maintained in Birka also after Ansgar. Birka may have served as a centre of innovation from where Christianity slowly was spread in the Lake Mälaren area. See e.g. A.-S. Gräslund 2001, 130. 44 chapter 2 where we can see a tendency towards Europeanization of the institutions and organizations earlier than in Svealand.55 At the end of the tenth century the pagan elite in the Mälaren region was challenged by a Christian kingship, apparently based in the provinces of the Götar. An episcopal see was founded early in Skara, approximately 1013–1014. A number of Romanesque churches were erected in Götaland. Such churches occurred later in the Mälaren region. One exception however is Sigtuna, where the Christian royal power had an outpost and an episcopal see (founded 1060) quite early. Several churches were built in Sigtuna, from which the process of Christianizing Uppland was launched. Sigtuna became a point of support for Christian kingship, and from there alliances were made with local rulers in the Mälaren region. The pagan cult in Uppsala survived, however, until the end of the eleventh century. It is possible that the Christian influence on people living in the southern part of Lake Mälaren and the conversion of the kings made Uppsala’s base as an over-regional central-place narrow.56 Most likely, some type of pagan elite remained in Uppsala during the eleventh century, perhaps some local representatives of the old royal institution. Perhaps this elite had some influence on central and northern Svetjud. As soon as they were converted, however, a wooden church was erected in Uppsala. Sometime in the middle of the twelfth century a great Romanesque stone church was built there, and in 1164 the archiepiscopal see was founded in Old Uppsala. Thus Old Uppsala regained its role as an inter-regional central place and cultic site. The first Christian kings of Svetjud were probably King Eiríkr inn sigrsæli Bjarnarson (r. c. 970–995) and his son King Óláfr sœnski (“Skötkonung”) Eiríksson (r. 995–1022). Adam refers to Eiríkr as Hericus rex victor, and in native traditions he is called inn sigrsæli ‘the Victorious’.57 These titles indicate that he was a warrior lord. According to Adam, Eiríkr converted to Christianity but relapsed into paganism.58 It was probably Eiríkr inn sigrsæli who founded Sigtuna during the 970s. Sigtuna was organized mainly as a Christian ruler’s administrative seat. This town reflects a new type of kingship, i.e. a central power influenced by ecclesiastical organization.59

55 Lindkvist 2008, 671. 56 Göthberg 2008, 107. 57 Adam II,30, 34f.; Schol. 24. Cf. Hallencreutz 1984, 372; P. H. Sawyer 1991, 31. 58 Adam II,35, 38f., 52. 59 Tesch 1990a; 1990b; Clarke & Ambrosiani 1993, 73; Lindkvist 1996, 225. The Three Areas Of Investigation And Their Political Leaders 45

In Adam’s description Eiríkr’s son, King Olaph (ON Óláfr sœnski), was a very pious Christian, and hostile to paganism.60 For instance, he was eager to destroy the pagan temple at Uppsala. Adam states that Óláfr founded the bishopric in Skara. He let Unwan consecrate Thurgot as the first bishop there.61 Scholars have previously argued that this passage shows that Óláfr lost power in Svealand and moved to Västergötland.62 On the other hand, Óláfr may have effected a diplomatic compromise.63 He could continue to be the king of the Svear although he had cut his connections with Uppsala and the old cult. His influence in Uppland remained. It is actually quite evident that Óláfr was also the king of the Christian Svear living in the southern parts of the Mälaren area.64 Numismatic evidence supports this and even connects him with Sigtuna. It seems as if King Óláfr had a royal mint there. On the coins produced in Sigtuna, Óláfr is called rex svevorum “the king of the Svear” and rex Situm “king in Sigtuna”.65 According to Icelandic traditions the Uppsala lineage ended with King Steinkell (r. 1060–1066); Óláfr sœnski’s sons Anundr Jacob (r. 1022–1050) and Eymundr (r. 1050–1060) were the last kings of this family.66 Adam is somewhat uncertain regarding Steinkell’s relationship to Eymundr, and states that he either was the nephew/grandson or the stepson of the old king (nepos an privig nus regis).67 According to Hervarar saga he was the son-in-law of Eymundr (hann atti dóttur Eymundar konungs).68 This text says that Steinkell’s father was Rǫgnvaldr the Old, perhaps an earl from Götaland.69 Sources indicate that Steinkell had interests in Uppland; according to Adam he supported the foundation of the diocese of Sigtuna, with Adalvard the Younger as bishop.70 When

60 Adam II,58. On the baptism of Óláfr, see Palme 1959, 69–72; Hellström 1996, 15–41. 61 Adam II,58. 62 Cf. P. H. Sawyer. 1991, 32. 63 Cf. Hallencreutz 1997, 125–127. 64 In the Old Norse contemporary poem, Óláfsdrápa sœnska (c. 1020), the skald Óttarr praised Óláfr as a king of the Svear: “Host-Baldr defends field, few kings have such ability. Óláfr gladdens the eagle [with corpses]. The lord of the Svear is outstanding.” Skj B1, 267. My trans. 65 Malmer 1996, 99–111. 66 Cf. Óláfs saga helga, in Heimskringla II, Ísl. Fornr. 27, 152, 156. 67 Adam III,15–16. 68 Hervarar saga . (Ed.) Turville-Petre, 70. Cf. (ed.) Jón Helgason, 159. 69 Rǫgnvaldr is mentioned by the skald Sigvatr in a contemporary poem and several times by Snorri. See e.g. Heimskringla II, Ísl. Fornr. 27, 28, 85, 88ff. 70 Adam IV,29. Cf. P. H. Sawyer 1991, 34f. Hallencreutz interpreted this passage as an expression of Adam’s historico-theological dualism. Hallencreutz 1993, 56. 46 chapter 2

Adalvard wanted to burn down the temple at Uppsala, Steinkell stopped him and declared that it would be punished with immediate death and that he, Steinkell, would be driven from the kingdom for bringing malefactors into the country.71 Steinkell’s son, King Ingi (r. 1080–1110) was a devoted Christian, and had probably his base in Götaland.72 Perhaps it was during his reign the pagan cult at the sanctuary of Uppsala finally was destroyed.

The political and social structure of the Late Iron Age society of the Svear was initially decentralized. Gradually the people living in the settlement districts, the lands and the regions of the Mälaren area, were united in an unstable inter- regional federation under a common ruler, and a loose empire or kingdom emerged.73 This federation leader, called the “king of Svear”, probably had a specific relation to Uppsala, where a political, royal and cultic centre early emerged. Perhaps Adelsö and Fornsigtuna (and later Sigtuna) became royal sites too during the Viking Age. A more stable and institutionalized Christian kingdom first emerged in the eleventh or twelfth centuries in the Mälaren area. At that time Svealand and Götaland also constituted a united kingdom.

2.2 The Earls in Trøndelag

Compared to ancient Svetjud, the source situation of Late Iron Age and medieval Trøndelag is good. There are several Old Norse traditions referring to this area. Most of them were written down in the thirteenth and fourteenth centuries, but some poetic traditions are contemporary with the Viking Age society there. In addition, there is archaeological and toponomastic evidence from Trøndelag, which helps to reconstruct the contextual aspects of this study.

2.2.1 The Region and the Society Trøndelag refers to the region situated around the Trondheimsfjord, in the central part of Norway.74 The mouth of the river and the coastal areas around it called Fosen originally belonged to Nordmøre and Naumadalen.75 During the Viking Age fishing was important for people living in the outer parts of these

71 Adam IV,30. 72 Lindkvist & Ågren 1985, 62. 73 Cf. T. Andersson 2004; 2005. 74 This general presentation of Trøndelag is based on e.g. Sandnes 1967; 1987; Sveaas Andersen 1977, 58ff.; Røskaft 1997; 2003; Brendalsmo 2006; Krag 2007; 2008. 75 See Røskaft 1997, 228ff.; Brendalsmo 2006, 36ff. The Three Areas Of Investigation And Their Political Leaders 47 areas. For a majority of the people living in Trøndelag, however, farming was the major industry/activity, not least in the inner parts of the Trondheimsfjord, but also in the southern parts of the region, at Gauldalen, Orkdalen and Neadalføret. The rich Late Iron Age farms of the province are located at places where farming conditions were good. In the mountains hunting was also common. The name Trøndelag is a compound including the folk-name Þrœnda- (gen. pl. of Þrœndir) and the word lǫg ‘law’.76 Today Trøndelag consists of an extensive region called Nord- and Sør-Trøndelag. In more ancient times Trøndelag was more limited and equivalent to the area called Frostuþingslǫg, i.e. the areas situated close to the Trondheimsfjord. The Old Norse names Þrœndalǫg and Frostuþingslǫg indicate that the organization of society and the affinity of the people living in this region were connected to a common law (ON lǫg). Most likely the common judicial assembly was held at Logtu (Lagatún),77 on the Frosta peninsula, in the central parts of the fjord (the river). The age of the thing-organization is, however, uncertain, but most scholars think that it at least goes back to the Viking Age.78 In the Viking Age, Trøndelag consisted of eight “shires” or “folks” (ON fylki sg.): four situated in Inn-Trøndelag (Eynafylki, Sparbyggvafylki, Verdœlafylki, Skeynafylki) and four in Ut-Trøndelag (Strindafylki, Stjórdœlafylki, Gauldœlafylki, Orkdœlafylki).79 Most likely the farmers connected to each fylki were ruled by a chieftain. These chieftains resided at great farms situated at central places in the landscape (No sentralgårder).80 At some of these central places local assemblies were held, as well as public cultic activities. In Inn-Trøndelag such

76 A more ancient name of the province is Þróndheimr. See Nyman 2002, 143. 77 The first element is pl. gen. of lǫg ‘law’. The second element is tún ‘courtyard, enclosed field’, see Røskaft 2003, 210. 78 See e.g. Sveaas Andersen 1977, 59. It has, however, been discussed whether Frostathing or Eyrathing was the most ancient thing of the Þrœndir, and whether the inner parts or the outer parts of Trøndelag were the central settlement districts. According to Jørn Sandnes (1967) the inner parts of Trøndelag were most ancient and thus Frosta was the old central place for all Þrœndir. In the Late Viking Age, however, Eyra became the new assembly place, because of the new political power appearing at Lade, i.e. the earls of Lade and the central royal power. In historical time the Eyrathing became a konungsþing. It was also located close to the new ecclesiastic centre of Niðaróss. See also Krag 2007. 79 On the fylki system, see e.g. Sveaas Andersen 1977, 58f.; Røskaft 2003, 168ff.; Brendalsmo 2006, 40f. It has also been discussed whether Romsdal, Nordmøre and Naumadalen were also included in the judicial “Trøndelag”. On this discussion, see e.g. Krag 2007, 269. 80 For a careful analysis of the central places and the political power in Trøndelag, see Røskaft 2003. 48 chapter 2 locals often were designated with names containing the element haugr.81 The fylki system with the central places was probably very important for the organization of the local society. The smallest unit of society in Trøndelag was the settlement districts called ON bygðir or héruð. The society was thus organized in segmentary fashion. The people living in the settlement districts were grouped together in the “shires” (fylki), which in turn were grouped together to form a “law” (lǫg), i.e. Þrœndalǫg. The term “law” (lǫg), thus intimates not only the authoritative discourse that regulated social practice, but also the territory in which that discourse prevailed and the social group whose community was based upon their submission to it, i.e. the Þrœndir (on the cultic organization of Trøndelag, see ch. 5 and 14). It has been suggested that the political structure of Trøndelag during the Viking and the Early Middle Ages should be described as a “farmer democracy” (No bondedemokrati), where the ruler only was a primus inter pares among the farmers.82 This theory had its base in Snorri’s description in Heimskringla of political conditions in Trøndelag. Today most scholars argue that this society was differentiated in both a social and economic sense.83 The farmers in this region were not a homogeneous group. The concept of hǫldr in the Frostathings-Lov designates a small farmer elite group who owned land (cf. ON óðal). These free men had strong political power on a local level of the society. There were also other free men who did not own land. They were called árborinn menn (cf. bóndi in the Gulaþingslǫg). There were also unfree people called þrælar ‘slaves’ at the very bottom of the society.84

2.2.2 The Earls of Lade In addition to the local rulers in Viking Age Trøndelag, there were also regional elites who attempted to gain both economic and political control over other social groups living in there. During the tenth century the Earls of Lade (ON Hlaðajarlar) seem to have been the most important political actors in the region.85 They had also great influence in the coastal regions in northern

81 See also Røskaft 2003. 82 On this theory, see Røskaft 1997, 228; 2003, 33f.; Krag 2007. Magnus Olsen (1926) even argued that Trøndelag formed a small kingdom in the Migration and the Merovingian Periods. 83 See e.g. Røskaft 2003, 33f. 84 Røskaft 2003, 40ff. 85 See e.g. Krag 2007; Jón Viðar Sigurðsson 2008a, 27ff. The concept of jarl (OSw iærl, iarl) is often associated with the rulers of Lade in Norway, but it has been used in other parts of The Three Areas Of Investigation And Their Political Leaders 49

Norway, and in Hálogaland, where they originally came from, more precisely Andøya. There was also an alliance between the earls of Lade and the Danish royal power during the Late Viking Age, with a short break during the reign of Earl Hákon. Lade (ON Hlaðir or Laðir) was the central place and main residence of the Earls in the tenth century. It was situated at the mouth of the river Nið (Nidelva), not far from the present Trondheim (ON Niðaróss). For purposes of communication it had a very good location. The name may be connected to the verb (h)laða ‘to load’ and perhaps it designates ‘the place where things were loaded [on boats]’.86 Thus Lade originally was connected to trade and communications.87 Most likely Lade also was an important ruler site. Documents from the eighteenth and nineteenth centuries report that ancient grave mounds and monuments were located there, along the waterside. When Gerhard Schøning visited the place in 1774 he observed several ancient monuments at “the church of Lade”: “Norden for Hlade, paa den slette Mark, har været en rund, med store Steene omkringsat Plads, som rimeligst har været en Domplads.”88 “Strax sønden for [Kirken] har man oppløiet en Hob Kull og Steene, tildels smukt udarbeidede.”89 Klüwer noted at the beginning of the nineteenth century “nogen næsten jævnede Kæmpehouge, hvori er fundne Vaaben, Ringe og Spænder af Guld, [dette] er ogsaa Alt hvad der nu fines tilbage af det gamle navnkundige Hlade.”90 These burial mounds are not preserved today. Only a few archaeological finds have been made there, such as a Merovingian point

Scandinavia, too. It has, for instance, been suggested that earls ruled Götaland in the tenth and early eleventh century. Foote & Wilson 1980, 135, critically considered by Gahrn 1988, 139–150. In historical sources referring to later periods, we hear that Earl Birger lived and ruled in Östergötland around the 1250s. RSw iarl is also attested in two Swedish Viking Age inscriptions, though both refer to Danish and Norwegian rulers. Sm 76 and U 617. Peterson 1994, 26. The etymology of iærl is ‘noble man’ (Sw ‘förnäm man’). Hellquist 1957, 418f. See also Ebel 2000, 29ff.; Düwel 2000, 32ff. The primary meaning of the concept jarl is thus ‘a distinguished, or noble man’, but in Viking Age Scandinavia it seems as if it was almost restricted to men of high rank, who might be independent rulers or subordinate only to kings. In present study I translate jarl to ‘earl’, even if the modern concept actually has somewhat different connotations compared to the ancient term. Cf. Cleasby & Vigfusson 1957 (1874), 323. On the concept of jarl, see also P. H. Sawyer 1991; T. Zachrisson 2011a. 86 An alternative interpretation is ‘Etwas aufgestapeltes’. Sandnes 2001; cf. Sandnes & Stemshaug 2007, 278. 87 Sveaas Andersen 1977, 226ff. 88 Gerhard Schøning II:6. 89 Gerhard Schøning II:4. 90 Klüwer 1823, 44. 50 chapter 2 of a spear as well as a couple of swords.91 A bucket-shaped clay tub from the Migration Period was also discovered at Lade. Beside these finds there is little archaeological evidence indicating what status and function the place had in the Late Iron Age.92 In Old Norse prose, however, Lade is described as an important centre for the people living in Ut-Trøndelag. According to Snorri, King Haraldr Finehair made Lade his most important seat after conquering Trøndelag in 900.93 Meanwhile, the Háleyjgjajarl Hákon Grjótgarðarsson also came to Trøndelag from Ørlandet and King Haraldr made him earl over Strindafylki.94 The king gave Hákon Grjótgarðarsson the ruler site at Lade.95 He and his successors, the Hlaðajarlar, made Lade into their main residence.96 This account made by Snorri has been questioned by historians. However, it is generally accepted that Earl Sigurðr, Earl Hákon and his sons resided at Lade.97 Snorri’s narratives about Lade, and the earls living there, may partly be supported by more ancient saga traditions.98 According to the Old Norse prose Lade is also described as the most important pre-Christian cultic site of Ut-Trøndelag. It is most likely that the Earls of Lade controlled the important trading route at the mouth of the river Nið (Nidelva).99 From a local perspective they were probably more powerful than the central royal power in the region. Most likely these rulers had some connection with the foundation of the “kaupang” at Nidelva, i.e. Niðaróss. According to the saga traditions King Óláfr Tryggvason

91 See Sandnes 2001. 92 Some bones were found in 1915, which may be related to the farm. Brendalsmo 2006, 523f. 93 Haraldr konungr fór þá aptr til Þrándheims ok dvalðisk þar um vetrinn. Jafnan síðan kallaði hann heimili sitt í Þrándheimi. Þar setti hann inn mesta hǫfuðbœ sinn, sem Hlaðir heita. Haralds saga ins hárfagra 9, in Heimskringla I, Ísl. Fornr. 26, 100. 94 Haralds saga ins hárfagra, in Heimskringla I, Ísl. Fornr. 26, 99. 95 Haralds saga ins hárfagra ch. 38. 96 Haralds saga ins hárfagra, in Heimskringla I, Ísl. Fornr. 26, 142. Hákonar saga góða, in Heimskringla I, Ísl. Fornr. 26, 163, 168, 171f.; Haralds saga gráfeldar, in Heimskringla I, Ísl. Fornr. 26, 272. Óláfs saga Tryggvasonar, in Heimskringla I, Ísl. Fornr. 26, 299, 308f. 97 See e.g. Sveaas Andersen 1977, 61, 107; Sandnes 2001; Krag 2000, 49; 2008; Røskaft 2003, 10ff. 98 See Fagrskinna, Ísl. Fornr. 29, 65, 145, 166. This narrative about Lade is somewhat different from Snorri’s version. See also Sverris saga 56, 78, 227. This saga describes how King Magnús Erlingsson (r. 1161–84) had his men stay out at Lade when he was in Niðaróss. It was thus still regarded as crown land in the twelfth century. See Brendalsmo 2006, 522. 99 Røskaft 2003, 106. The Three Areas Of Investigation And Their Political Leaders 51 founded the site c. 995.100 Archaeological evidence indicates, however, that a settlement existed there even before King Óláfr’s reign.101 At some point in the Late Viking Age the farm at Lade was confiscated by the royal power. A stone church was erected at the farm at the end of the twelfth century. It became the fylkiskirkja of Strindafylki. There are also traditions mentioning that a monastery was located at this site.102 We may conclude that Lade was both a religious and a political centre in the period from the tenth century to the twelfth century. It may also have been the trade centre in the area, before the “kaupang” at Niðaróss was established.

2.2.3 Political and Religious Changes in the Viking Age It has been much debated when the creation of Norway (“rikssamlingen”) took place.103 In the thirteenth-century sagas we get the impression that Haraldr Finehair (r. 872–930) played a major role in this event. Snorri Sturluson made a systematic description of Haraldr’s conquest in Heimskringla, where the king conquered territory after territory.104 Finally at the battle of Hafrsfjǫrðr, Haraldr won a great victory over the kings of the Vestlands, who were his last opponents. This image of historical events cannot be corroborated by contemporary sources, such as skaldic verses. Today historians argue that the kingdom Haraldr took over included only the central and southern parts of the Vestland region.105 He never succeeded in taking the areas around Viken under his possession, since they were controlled by Danish kings. In the Nordvestland and further north Haraldr had no more than a formal overlordship. The earls in these regions, who often were subordinate to him only in name, had the actual power there. Earl Hákon Grjotgardsson of Lade, for instance, was in many senses Haraldr’s equal. The regions that were unified around 900 were actually the lands situated along the rich trade route, from the south to the north.

100 According to the saga traditions it was Óláfr Tryggvason who founded the site, see e.g. Óláfr saga Tryggvason 70, in Heimskringla I, Ísl. Fornr. 26, 318: Óláfr konungr fór liði sínu út til Niðaróss. Þá lét hann reisa þar hús á Niðarbakka ok skipaði svá, at þar skyldi vera kaupstaðr . . . 101 See Røskaft 2003, 106f. 102 Brendalsmo 2006, 521–525 103 An overview of the early research on “rikssamlingen” in Norway is given by Sveas Andersen 1977, 40ff. For a more recent account, see e.g. Krag 2000, 44–80. Cf. Jón Viðar Sigurðsson 2011 and most recently Lincoln 2012, 52–62 and 2014. 104 See Haralds saga ins hárfagra, Heimskringla I, Ísl. Fornr. 26, 94–149. 105 See Krag 2000, 44ff.; 2008, 647. 52 chapter 2

In these areas an inner network was created, by family ties, within the highest social circles of chieftains. However, the political structures in the new kingdom were weak and the areas around Viken, in the south, were still during the tenth century under the influence of Danish kings. It was only during the reign of Earl Hákon of Lade (r. 970–995) that Norwegian rulers became more independent in relation to the Danish kings. As in other Scandinavian areas, the shift in religion was very important for the political and social development of Norway in the Viking Age. As early as the tenth century Christianity had reached most parts of the coastal regions. Hákon the Good (r. 934–961) was the first king who tried to establish Christianity in Norway, although he met some resistance from the conservative farmer elite in Trøndelag. It was therefore not until King Óláfr Tryggvason (r. 995–1000) and Óláfr Haraldsson (r. 1015–1030) that the whole country was converted.106 The canonization of King Óláfr contributed to legitimize his successors to the throne and it also helped to establish the Christian kingship in Norway and unify the country. In previous research, the relationship between local aristocracy in Trøndelag and central royal power has been much focused on.107 Heimskringla reports that the pre-Christian chieftains and farmers of Trøndelag were quite independent of the central royal power in the tenth century. The Þrœndir of Inn- Trøndelag, for instance, resisted Hákon the Good’s attempt to Christianize the region. Even if King Hákon had good relations with Earl Sigurðr of Lade, the latter and his son Earl Hákon remained pagans. Earl Hákon, who was a highly conservative and confident pagan, had a great conflict with the Christian sons of King Eiríkr blóðøx (King Hákon’s brother). Earl Hákon’s sons (Eiríkr and Sveinn), however, converted to Christianity at the beginning of the eleventh century. Gradually the central royal power and the official Church gained more political and religious influence over several groups in Trøndelag. The opposition between them and the elite of the Þrœndir seems to have remained, however. It is visible in the two systems of organizing the churches in the province. One system was connected to the fylkiskirkjur and it was controlled by the bishop and the king. The other system was connected to the hǿgendiskirkjur, and represented the local elite. These two competing systems were maintained until at least the thirteenth century.108

106 See however Sæbjørg Walaker Nordeide (2011 and 2012) who by means of archaeological finds has established a more differentiated image of the Christianization of Trøndelag. 107 See e.g. Helle 1974; Røskaft 2003, 20. 108 Brendalsmo 2006, 147–226, 285. The Three Areas Of Investigation And Their Political Leaders 53

During the Early Middle Ages Trøndelag was developed into one of the most important parts of Norway. In Fagrskinna it is described as hǫfuð Nóregs,109 and Historia Norwegie calls it patria principalis.110 The ceremony of taking the king (konungstekja) now took place at the Eyrathing. Perhaps this role of Trøndelag was inspired by the fact that St Óláfr’s remains were placed in the shrine of Niðaróss (today Trondheim), after his death at Stiklastaðir in 1030. According to tradition, a lot of miracles took place in connection with the saint’s relics. Soon Niðaróss became both an ecclesiastic and a royal centre. As early as 1075 Adam of Bremen refers to Niðaróss as the most important ecclesiastic centre (metropolis civitas) of Norway.111 Several churches and monasteries were built there as well as royal settlements. One of the most important churches was Kristkirken, which later was made into the cathedral of Niðaróss (finished in the thirteenth century). In 1152–1153 Niðaróss became the archbishop seat and the centre of Norwegian church province.112

It seems as if the people of Trøndelag, the Þrœndir, were gathered early on into a political unit with a common law. The political structure of the early Viking Age society seems to have been decentralized and the political power was split among many hands. There were several independent local chieftains ruling in the settlement districts. During the tenth century a regional elite appeared with ambitions to control the whole province, namely the Earls of Lade. Most likely they were supported by the kings from the south-western Norway. The descendants of King Haraldr converted to Christianity in the middle of the tenth century. Gradually the central royal power and the official Church gained more political and religious impact in Trøndelag. In the twelfth century, however, they were still opposed by some elite groups of the Þrœndir, who claimed local independence from central power.

2.3 The Chieftains of Iceland

Iceland was settled in the period 870–930, which is usually called “the settlement period” (landnámatíð). There are several medieval texts describing this and other important processes of Early Iceland, such as the political development, the constitution of the “Free State”, and the introduction of Christianity.

109 Fagrskinna, Ísl. Fornr. 29, 212. 110 Historia Norwegie, (ed.) Ekrem & Mortensson 2003, 74. 111 Adam IV,33. 112 On Niðaróss, see e.g. Brendalsmo 2006, 94ff. and the literature quoted therein. 54 chapter 2

The most important sources for these processes and developments are Landnámabók, Ari inn fróði’s Íslendingabók, the law-book Grágás, the Sagas of Icelanders, and the Bishops’ Sagas. In addition to these written sources we also have place-names and archaeological finds illuminating the contexts of the present study.

2.3.1 The Settlement (landnám) Thanks to recent archaeological investigations, we know almost exactly when Iceland was settled. Ari inn fróði Þorgilsson’s information (in his Íslendingabók) that Iceland was first settled in 870,113 is almost confirmed by the new excavations. Just before the first settlers arrived there was a volcanic eruption. The ash from this eruption has been dated to 872 ± 2. According to archaeological data, there are no traces of settlement below the ash layer, only above.114 One of the most reliable texts informing about the early settlements of Iceland is Landnámabók.115 It tells that most of the settlers came from Norway (the coastal areas), and the British Isles (mostly from Ireland). Some of them also came from Denmark and even Sweden. Recent DNA analyses confirm this mixture of people.116 Landnámabók mentions 417 settlers (landnámsmenn)117 who were in charge of the settlement process. Of these, 404 were men while 13 were women. Landnámabók informs us about the places where the new farms were founded, kinship ties, and sometimes also the causes behind the emigrations. The settlers were often wealthy farmers and local chieftains. Many of them owned their own ship and they transported property such as livestock to Iceland, but also family members and slaves. The number of people who came to Iceland has been estimated at about 10,000–20,000.118 The landnámsmenn brought with them domestic animals, sheep, cattle, goats, pigs and horses. It seems as if this stock of animals was not large enough to sustain the new settlers. Therefore fishing became important when people arrived in Iceland. Consequently we find the highest numbers of new settlers in the coastal areas, mostly in south-western and western Iceland, where the

113 Íslendingabók, Ísl. Fornr. 1, 4. 114 Jón Viðar Sigurðsson 2008b, 571. 115 Landnámabók exists in several medieval witnesses: Sturlubók (= S) made by Sturla Þórðarson; Hauksbók (= H) by Haukr Erlendsson; Melabók (= M); Skarðsárbók (= Sk); Þórðarbók (= Þ) and also Þ (M): Lesháttur Þ sem telja má runninn frá M. 116 See Jón Viðar Sigurðsson et al. 2005, 93ff. 117 The concept landnámsmenn is a technical term for the leading immigrants to Iceland, including both men and women. On the amount of landnámsmenn, see Jón Viðar Sigurðsson 2008a; 2008b. 118 See Vésteinn Ólason 1998, 23. The Three Areas Of Investigation And Their Political Leaders 55

North Atlantic Drift (Gulf Stream) gave propitious conditions for fishing, whale hunting and living in general.119 The single farm was the main feature of the settlement pattern. No villages or towns developed there during the Middle Ages. Many farms were moved or even completely abandoned in the early phase, until the settlers had found the most favourable places to live at. There are several causes behind the emigration from Norway to the Atlantic islands. One major cause often emphasized is King Haraldr Finehair’s politics at the end of the ninth century.120 He challenged the long-standing tradition of local independence, and sought control over the greater part of the country. According to the thirteenth-century sources, King Haraldr also levied property taxes on men who traditionally not had paid any land taxes because they owned their lands, i.e. he disturbed the age-old customs of the family-based landholding, called ON óðal. Many landowners reluctant to accept Haraldr’s demands therefore left Norway, and went to Iceland, but also to the Shetlands, the Orkneys, the Hebrides, England, Scotland and Ireland. When they came to their new lands they settled quite extensive areas and distributed land to family members, relatives, friends and protégés. One example of such a set- tler is Skalla-Grímr Kveld-Úlfsson (the skald Egill Skalla-Grímsson’s father) in Egils saga, who was forced to leave Norway because of his conflicts with King Haraldr, and therefore settled at Borgarfjǫrðr in western Iceland.121 He built his farm at Borg, in the Mýrar area, and distributed land to his fellows there. Landnámabók reports in laconic statements about Ǫrn í Arnarnesi. He was a renowned man and a friend of Geirmundr Hjǫrsson. He left Rogaland in Norway because of King Haraldr’s overbearing power (hann fór af Rogalandi fyrir ofríki Haralds konungs). He took land in Arnarfjǫrðr, in western Iceland.122 Probably this short passage is more reliable than most of the stories about Skalla-Grímr in Egils saga. In Landnámabók we find roughly 20 similar statements about people who left Norway because of Haraldr’s politics.123 Even if the composer(s) of Landnámabók and the saga authors exaggerated King Haraldr’s greed and influence on the Icelandic landnámsmenn,124 there is most

119 See Byock 2001, 26, 31; Jón Viðar Sigurðsson 2008b, 572. 120 See the overview in Jón Viðar Sigurðsson et al. 2005, 95f., but also in Byock 2001, 82ff. 121 Egils saga, Ísl. Fornr. 2, 5–80. Cf. Landnámabók S29–30, Ísl. Fornr. 1, 68ff. 122 Landnámabók S134; H106, Ísl. Fornr. 1, 176. 123 Jón Viðar Sigurðsson et al. 2005, 95. 124 It should be noted that Ari makes no mention of such factors in Íslendingabók. However, it is not unlikely that some people left Norway because of disputes with the king. See Vésteinn Ólason 1998, 24. 56 chapter 2 likely some kind of truth behind this information.125 Of course there were also other factors behind the emigration from Norway. Population growth and lack of agricultural lands in Norway may have played a certain role in this process.126 Also prospects of good fishing and hunting in the new lands in the North Atlantic may have attracted some of the emigrants, as well as good conditions for stock breeding and cultivation of crops.127 The early settlers, of course, had several problems when arriving in Iceland. They had to make the limited habitable area of the empty land prosper. However, they managed to establish a society with a general assembly as early as the tenth century, and soon it functioned as a single island-wide community. Viking Age Iceland was in many senses a decentralized, stratified society, with a blend of pre-state features and organized state institutions.128

2.3.2 Early Leadership Early Icelandic chieftains were usually called goðar (sg. goði),129 and their authority or office was called goðorð (‘god-dignity’, cf. mannaforráð ‘power, rule over people’¸ ríki ‘might, power’).130 A holder of such office was called goðorðsmaðr or goði. The term goði is derived from ON goð ‘god’, thus indicating a cultic function of these leaders (cf. Goth gudja ‘priest’, OHG *goto ‘ein Mann der zu einem Gott gehört’) (on the religious aspects of goðar, see ch. 5 below).131 Besides his religious assignments the goði also performed other societal functions, such as a lawman, but also acted as a more general political leader.132 However, it seems as if they rarely appeared as warrior chieftains. There were quite a few goðar in the settlement districts of Iceland (see below). In the present study we will focus on one particular “goði family”, more precisely the Þórsnesingar, who lived at Þórsnes, situated on the Snæfellsnes

125 Byock 2001, 84. 126 Jón Viðar Sigurðsson et al. 2005, 96f. 127 Jón Viðar Sigurðsson et al. 2005, 97ff. 128 Byock 2001, 11. 129 It is an old scholarly tradition to refer to the goðar as “chieftains”, though the correspondence may not be exact. See Byock 2001, 13. 130 On goðar in general, see mainly Ebel 1998; Jón Viðar Sigurðsson 1999; Dillmann 2006, 312ff. See also Vésteinn Ólason 1998, 26ff.; Byock 2001. For the religious aspects of goðar, see Sundqvist 2003a; 2003b; 2007. 131 Wesche 1937; de Vries 1956–57; Kuhn 1978. 132 Eyrbyggja saga 12 (Ísl. Fornr. 4, 20) states: Hann [Arnkell] var manna mestr ok sterkastr, lagamaðr mikill ok forvitri. Hann var góðr drengr ok umfram alla menn aðra þar í sveit at vinsældum ok harðfengi; hann var ok hofgoði ok átti marga þingmenn. The Three Areas Of Investigation And Their Political Leaders 57 peninsula in western Iceland, between Vigrafjǫrðr and Hofsvágr. This family is described briefly in Landnámabók,133 and more extensively in Eyrbyggja saga,134 which tell of the chieftain and landnámsmaðr Þórólfr Mostrarskegg, who settled the areas around Þórsnes. Several of Þórólfr’s sons and grandsons carried the title goði (Hallsteinn Þorskafjarðingar-goði; Þorgrímr goði Þorsteinsson þorskabíts; Snorri goði Þorgrímsson) and most likely he had a similar type of authority himself. (I will consider Þórólfr and his family at greater length below.) Most likely many landnámsmenn who arrived in Iceland and became chieftains and cultic leaders (goðar) there may have had similar functions in their homeland, Norway. It has been shown that several Icelandic goði families were descendants of Norwegian hersar, i.e. chieftains.135 The chieftains of Iceland were thus recruited from the lower elite of Norway. The power of the goðar was restricted in Viking Age Iceland. They had no possibilities to limit the access of other farmers to natural resources, and they were unable to control a region’s surplus production. The goðar were thus no commanding nobility (see further ch. 15), but rather leaders of interest groups, composed of þingmenn drawn from the farmers.136 A þingmaðr was the chieftain’s legally recognized follower. The chieftain thus made personal alliances with the farmers, often motivated by shared and reciprocal self-interests. The goði was socially and politically nothing more than a primus inter pares.137 It is possible that the goðar systematically demanded hospitality from the farmers who owed them alliance by means of veizlur ‘(ceremonial) feasts’, but probably also the chieftains invited the farmers to such banquets in order to improve their friendship with them.138 The authority or chieftaincy (goðorð) of the goðar was regarded as a private possession, usually inherited in the family, but it could also be acquired in other ways as well (see ch. 5).139

133 Landnámabók, S86, H74, M25. Ísl. Fornr. 1, 124–127. 134 Eyrbyggja saga, Ísl. Fornr. 4, 5ff. 135 See e.g. Phillpotts 1912–13, 276ff.; Vésteinn Ólason 1998, 29f. 136 Byock 2001, 13. 137 Vésteinn Ólason 1998, 28. 138 Jón Viðar Sigurðsson 1999. 139 There is, however, an early legal document that states that the goðorð, should be seen as “authority and not property”. Grágás I, (ed.) Finsen 1852, 206. In the Christian period it therefore was not subject to tithes. It is quite clear, however, that the sagas mention that it normally was inherited, and could be sold, lent, and even divided. See Turville-Petre 1964, 260. 58 chapter 2

2.3.3 The Political Structure in the Icelandic Commonwealth (the Free State) In the perspective of the present study, it is important to note the decentralized political structure of the early power in Iceland. There are no tendencies of any development towards a “centralized state” in Iceland during the Viking Period, comparable to the kingdoms and earldoms in contemporary Svetjud and Trøndelag. The political development between c. 1120 and 1262/64 may admittedly be characterized by a concentration of power. Around 1200 only about seven families controlled most of Iceland. However, no central royal power appeared on the scene as long as the Free State existed. After 1262/64 the Norwegian monarchy took control over all goðorð in Iceland, and the fall of the Free State was a fait accompli.140 A more central political power was thus very slow to develop in Late Viking Age Iceland, and the political system that developed before the thirteenth century was characterized by the inability of the chieftains to create permanent power structures.141 It was only at the beginning of the thirteenth century that a few chieftains were able to establish control over more than one region in the country. These chieftains never developed into overlords, but acknowledged the Norwegian king as their ruler. The decentralized structure of early Icelandic society and political power may be connected to the so-called “constitution”, which emerged shortly after the settlement period (landnámatíð). According to Ari all land was claimed (albyggt) within 60 years from the first landnám, that is, c. 930.142 About that time the General Assembly (alþingi) was also founded at Þingvellir, close to Þingvallavatn, in western Iceland. However, before that the first settlers had established assemblies at Þórsnes and at Kjalarnes. It has been much debated whether the assembly at Kjalarnes was only local or applied to all Iceland, and thus a forerunner of the general assembly.143 It is generally accepted that the main elements of “the constitution of the Free State” and a common code of law, were introduced at the first General Assembly of Þingvellir c. 930.144 The most important source for the “constitution” is the description in the Konungsbók version of Grágás (c. 1250). When the constitution was founded at the first assembly of Þingvellir, Iceland had 36 chieftaincies (goðorð).

140 Jón Viðar Sigurðsson 1999; 2008b, 573. 141 Orri Vésteinsson 2001, 338. 142 Íslendingabók 3, Ísl. Fornr. 1, 9. 143 See the information about the first assembly at Kjalarnes in Íslendingabók 3, Ísl. Fornr. 1, 8. See discussion about this assembly in Jón Viðar Sigurðsson et al. 2005, 137. 144 Maurer 1873; Sigurður Nordal 1942; Jón Jóhannesson 1956. The Three Areas Of Investigation And Their Political Leaders 59

Around 965, however, when the country was divided into four quarters, it was expanded with three new chieftaincies in the Northern Quarter (Norðlendingafjórðungur).145 In order to maintain the balance between all quarters, the Eastern, Western and Southern Quarters (Austfirðingafjórðungur, Vestfirðingafjórðungur, Sunnelendingafjórðungur) were also given three additional goðar each, resulting in a total of 48 chieftaincies represented at the General Assembly. Hence, each quarter had 12 chieftains represented at this annual meeting. In each quarter spring assemblies (sg. várþing) were also held. In the Eastern, Western and Southern Quarters there were three such assemblies, while the Northern Quarter had four spring assemblies.146 All 48 chieftains constituted together the Law Council (lǫgrétta). Each chieftain had two “assembly men” (þingmenn) as advisors. Hence the total number of men in the Law Council was 144. In addition to them the Lawspeaker (lǫgsǫgumaðr) and later the two bishops of the country took part in the Law Council. The power was thus distributed among many hands in this system.

2.3.4 Christianization In Íslendingabók Ari mentions that Christianity was preached and accepted as the official religion in Iceland during King Óláfr Tryggvason’s reign.147 The king sent a priest called Þangbrandr to Iceland who taught Icelanders about Christianity and baptized them. Ari also states that the final decision concerning the conversion was taken at the General Assembly in 999 or 1000. In this process the Lawspeaker Þorgeirr Ljósvetningagoði Þorkelsson played a significant role. In Ari’s opinion, it was because of him that the conflict between the Christians and the heathens at the assembly resulted in a peaceful outcome.148 He let the Christian party win, although some concessions were made to the pagans. For instance, everybody could continue to eat horse meat and perform infanticide by exposing unwanted children in the woods, according to ancient customs. People were also still allowed to perform sacrifices (blót) to the old gods as long as they did so in secret.

145 Whether these systems with 36 or 39 chieftaincies really existed in Viking Age Iceland is much debated. See e.g. Jón Viðar Sigurðsson 1999. 146 Jón Viðar Sigurðsson 2008b, 573. 147 Íslendingabók 7, Ísl. Fornr. 1, 14ff. 148 For the peaceful nature of the introduction of Christianity in Iceland, see Maurer 1855–56; Jón Hnefill Aðalsteinsson 1998; Orri Vésteinsson 2001; Jón Viðar Sigurðsson 2003. 60 chapter 2

It has been much debated why the conflict at the General Assembly had this outcome.149 In this debate the role of the goðar has been discussed. About 150 years ago the German legal historian Konrad Maurer argued that the system of the goðar had a pagan and religious origin.150 He based his argument very much on the etymology of the concept goði (see above). Maurer’s thesis has been contradicted with the claim that if the system of goðar had a pagan and religious origin, it would never have survived the introduction of Christianity at the General Assembly 999/1000.151 This argument has its turn been challenged. Recently scholars have thought that it was precisely because of the goðar’s control over the pagan cult in Iceland that it was possible to introduce the new religion via a resolution at the General Assembly.152 In religious questions the goðar were dominant over their subjects, i.e. the farmers and the þingmenn. Thus the farmers had to accept the decisions of the chieftains. It is well known that many goðar became Christian priests when Iceland converted to Christianity in the year 1000 (or 999). They erected private churches on their farms in much the same way that the pagan chieftains had erected hof buildings there previously. Ulrich Stutz argued that the Eigenkirchenwesen was built on a pagan Eigentempelwesen in Iceland.153 In his theory the system of goðar was also involved. The goðar thus maintained their traditional authority over different aspects of society in the twelfth and thirteenth centuries. They

149 See different positions in Jón Viðar Sigurðsson 2003, 35. 150 Maurer 1855–56. 151 Ólafur Lárusson 1960 152 Jón Viðar Sigurðsson 1999; 2003; 2011. 153 See Stutz & Feine 1989. Ulrich Stutz argued that the system of proprietary churches was built on a continuation from a pre-Christian Germanic “Eigentempelwesen”. The Sagas of Icelanders often mention that the pre-Christian chieftains themselves erected ceremonial buildings (ON hof ) on their farms. They controlled these buildings and the activities that took place there and exacted taxes from the farmers for cultic activities performed there (tolla gjalda, hoftoll til leggja), e.g. Eyrbyggja saga 3–4; Kjalnesinga saga 2. According to Stutz, the system of proprietary churches was based on these ideas. His entire theory, however, was questioned by Olaf Olsen: as outlined in ch. 4 below, he argued that we have no clear evidence of pre-Christian Germanic temples at all. See e.g. Olsen 1966. Cf. Schäferdiek 1986, 560. Olsen and others also questioned the idea that the system of proprietary churches was unique for Germanic peoples. On the contrary, there are clear signs that this system also existed in Slavic and Celtic regions, as well as in Byzantium. E.g. Olsen 1966, 80. Today scholars argue that the system of proprietary churches had its roots in Late Antiquity, see Landau 1982, 400. According to Susan Wood (2006, 11, 92–108) the specific power developed among Roman provincial landowners played an essential role for the development of this system. She states, however, that it is hard to find its exact origin. The Three Areas Of Investigation And Their Political Leaders 61 probably constituted a small-scale local elite, who exerted some power over Christian cult, politics and trade. The influence of the goðar in ecclesiastic matters in Iceland was later opposed by the Church. The bishopric of Skálholt was founded in 1056, and in 1106 the see of Hólar was established. Before 1104, Iceland, like the rest of Scandinavia, belonged to the archdiocese of Hamburg-Bremen, and from 1104 to 1152–53 that of Lund. In 1152–53 the archbishopric in Niðaróss was founded in Norway. This archdiocese included the Norse settlements in North Atlantic islands. One important reason for founding the see at Niðaróss was the intentions of the reform movement (libertas ecclesiae), i.e. to free the church from secular influence and to set the Scandinavian churches more directly under the authority of Rome. The archbishop of Niðaróss demanded complete control over the Ice- landic Church. In order to achieve this, alliances were forged with the Icelandic bishops. The Niðaróss bishop gained support from Þorlákr Þórhallsson, who became bishop of Skálholt in 1178. After a conflict with one of the leading chieftains of Iceland, Jón Loptsson, however, Þorlákr relinquished these claims, and did not raise them again. He continued as bishop of Skálholt until his death in 1193, when Páll Jónsson (of the powerful Oddaverjar family) was elected as his successor.154 In Hólar the bishop Guðmundr Arason was elected in 1201. He was a strong supporter of the reform movement and established bonds with Niðaróss. Because of this he had many conflicts with the chieftains. The conflicts between the Church and the chieftains in Iceland were thus related to the new centralization tendencies of the Church. According to Icelandic tradition, power and religious concerns had to be connected to the chieftains. The authority of these concerns thus had to be divided among many hands. It was not until the middle of the thirteenth century, and in connection with the breakdown of the Free State, that the Icelandic Church was transformed into a bishop’s church, which better suited the general structure of the papal church created in contemporary Europe.

In this chapter I have argued that the political structure and society in the different parts of Late Iron Age and medieval Scandinavia varied. Both in the Mälaren region and in Trøndelag the power structure was originally decentralized. However, as time went, a concentration of power and a centralization process took place. During the early Viking Age overlords, High Kings and powerful earls appeared in these regions, with ambitions to exercise domination and power (ON ríki) over several local chieftains. At times such overlords constituted loose kingdoms. In the second half of the Viking Age more stabilized

154 Jón Viðar Sigurðsson 2008b, 576. 62 chapter 2 and territorial kingdoms/earldoms emerged in Svetjud and Trøndelag. Most likely these political structures were influenced by Continental and Christian ideas. In Iceland, however, the process towards a central power structure never took place during the Viking Age, but only in the twelfth or thirteenth century. The early Icelandic society and its “constitution” were built on a decentralized system, where power was divided among many men. The goðar exercised some power over the farmers who had declared allegiance to them, but they were actually more like leaders of interest groups, where the farmers also ben- efited from the decisions made by their chieftains. These chieftains of Iceland were often descendants of Norwegian hersir families and thus recruited from the lower elite of Norway. Why the settlers of Iceland never established a society with a king, a military or administrative centre is uncertain, but most likely the Icelandic constitution may have played a certain role in this development. chapter 3 Religious Strategies for Political Authorization— A Case Study

As noted in chapter 2, the social and political structure differed in the three investigation areas, and it also varied over time. In the following chapter I will delve into the question of the religious strategies for rulership in these three areas, although the investigation here will be limited to the mythical- narrative dimension and a case study. I will investigate and compare traditions connected to three representative ruling families in the investigation areas, namely the Swedish-Norwegian Ynglingar, the Norwegian Earls of Lade, and the Icelandic chieftain family called the Þórsnesingar. These traditions appear in Old Norse skaldic poetry, Kings’ Sagas, and Sagas of Icelanders, and also in Latin texts. None of the descriptions may be regarded as strictly “historical”, but we may at least reach medieval opinions in them about these pre-Christian rulers and their religious-mythic strategies. By means of the skaldic poetry we may also grasp some attitudes prevalent in Viking Age society. It will be argued that the kings called the Ynglingar and the Earls of Lade had a quite similar religious strategy, while the Icelandic chieftains used another strategy. It is suggested that these differences are related to the variations in the social-political structures in the three areas. This conclusion has formed the basis for the general hypothesis of this study.

3.1 The Kings Called the Ynglingar

In my opinion, there was a fundamental ideological and cosmological system of thinking in the uppermost aristocratic milieus of the Late Iron Age Mälaren region. This system of thinking may be seen in the traditions describing the ancient royal family called the Ynglingar. This family appears in different types of sources (see below). According to them, they originally lived in the area around Lake Mälaren, but then emigrated from there, and settled in Vestfold, southern Norway. Perhaps some of these traditions were recited in the halls of the Mälaren region during the ceremonial feasts and when drinking memorial cups to departed rulers and ancestors. However, we have no clear evidence of this.

In what follows I will concentrate on the Ynglinga family and study their mythic strategies for political authority and legitimacy. It will be argued that the real core of this strategy was the notion that these kings descended from the gods. It will also be argued that this mythic-cosmological strategy of the Ynglingar might very well reflect historical conditions in Late Iron Age Svetjud.

3.1.1 Ynglingatal The oldest traditions about the “Ynglingar” appear in the skaldic poem Ynglingatal. This poem has been preserved to the present via manuscripts of Snorri Sturluson’s prose text Ynglinga saga (in Heimskringla) (c. 1230), where it is quoted.1 According to Snorri, it was composed by Þjódólfr inn fróði ór Hvini, who was King Haraldr Finehair’s skald sometime towards the end of the ninth century.2 It is composed in Þjódólfr’s native tongue, and recounts the reigns of twenty-nine rulers. Ynglingatal has twenty-seven stanzas. Each stanza briefly describes the deaths, burials and sometimes burial places of the kings. It seems as if the poem is made up of three distinct units.3 The introductory eight stanzas concern mythical and/or heroic kings living in the Mälaren region. There are then thirteen stanzas about legendary kings of the Svear, with names beginning with a vowel. Finally, the last six stanzas describe six possibly historical Norwegian kings living in the areas around Vestfold. It has been commonly held among scholars that Snorri’s information concerning the dating of Ynglingatal is reasonable, that is, c. 890.4 This dating, however, was challenged by Claus Krag in 1991.5 He argued that there are medieval Christian values and ideas present in the poem indicating an anachro- nism. I have previously presented my objections and those of other scholars

1 There are three major witnesses of Ynglinga saga. Two of them are later copies of the medieval parchment codices Kringla and Jǫfraskinna. (1) K—AM 35 fol. is a copy of Kringla made by Ásgeirr Jónsson (c. 1700), and provided with corrections made by Árni Magnússon. (2) J1— Manuscript AM 37 fol. is the oldest and best copy of Jǫfraskinna, unfortunately defective, made by Bishop Jens Nielssön (c. 1567–68). J2—AM 38 fol. is a copy of Jǫfraskinna made by Ásgeirr Jónsson 1698. (3) F—Codex Frisianus, AM 45 fol., was written by an Icelander c. 1325. Cf. Noreen 1925, 195–197. Wessén 1964, v–vii; Bjarni Aðalbjarnarson 1979 (1941), 2. 2 Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 4. 3 See Wessén 1964. John McKinnell (2010) argues that the poem is made up of three or four distinct blocks, connected by two “linking” figures, which probably reflects different oral sources. 4 See e.g. Åkerlund 1939. 5 This opinion was not new in the research on Ynglingatal, both Bugge (1894, 108–153) and Neckel (1908, 389–421) have made similar statements earlier, but Krag (1991 and 2009) emphasized it further. Religious Strategies For Political Authorization—a Case Study 65 against Krag’s arguments for a late dating,6 and I will not repeat them here. In my opinion, there is no reason to abandon the traditional dating of Ynglingatal. Ynglingatal is usually regarded as a genealogical poem. It was composed in honour of King Rǫgnvaldr, a minor ruler in Vestfold, Norway, son of Óláfr Guðrøðarson (called digrbeinn and/or Geirstaðaálfr in later texts, see below) and a relative of King Haraldr, and it includes twenty-eight of his forefathers. The name Ynglingar is not attested in the poem, and does not occur before Ari’s list of this family (Þessi eru nǫfn langfeðga Ynglinga . . .) in Íslendingabók, dated to c. 1130.7 It thus seems as if the medieval reception regarded this poem as a genealogy of one family. In his thesis, Inn i skaldens sinn, Bergsveinn Birgisson rejects the idea that Ynglingatal reflects a “historical genealogy” of the family of “Ynglingar”.8 He conceives the poem as a níð since the kings mentioned in it died ignoble deaths. According to Bergsveinn Ynglingatal was composed by a skald who was acting within the hirð of King Haraldr,9 and the polemics in it were directed against Danish royal power ruling in the area around Viken in southern Norway, and also against Swedish kings. I agree with Bergsveinn that the poem may not reflect one “historical genealogy” and that some of the rulers die a humili- ating death. In my opinion, however, not all of the kings in the poem have these disgraceful deaths and thus an interpretation of all of Ynglingatal as a níð is somewhat misleading.10 I think that Ynglingatal may have been composed from several genealogical traditions deriving from eastern Scandinavia and elsewhere.11 Þjódólfr revised these traditions to suit his purpose. In my opinion, however, the poem was early conceived as a genealogy of one family. Ari obviously made such interpretation of it in his Íslendingabók,12 and

6 See e.g. Sundqvist 1995; 2002, 43–52; 2005b; 2007, 83–85. Also critical of Krag’s thesis are, e.g. Fidjestøl 1994; Dillmann 2000a; Skre 2007b; Bergsveinn Birgisson 2007; McKinnell 2010; and Marold in SkP I, 5f. 7 Íslendingabók, Ísl. Fornr. 1, 27f. 8 Bergsveinn Birgisson 2007. Krag (1991) likewise argues that the Ynglinga dynasty was a medieval fabrication. 9 A similar interpretation has been made by Lönnroth 1986. Svanhildur Óskarsdóttir (1994, 768) characterized the poem as grotesque rather than ironic. 10 See the discussion in McKinnell 2010, 34ff. and Marold in SkP I, 7f. 11 See Sundqvist 2002, 47–52; 2007, 86–88. 12 See Ísl. Fornr. 1, 27f. A genealogical account of these kings also appears in Historia Norwegie (c. 1160–1175). See (eds.) Ekrem & Boje Mortensen 2003. In this Latin text several kings are presented as cuius filius “his son”, emphasizing that it was a genealogy concerning one family. So too does Snorri in his Ynglinga saga. In ch. 12, for instance, he states that “Sveigðir succeeded to the realm after his father” (Sveigðir tók ríki eptir fǫður sinn). Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 27. 66 chapter 3 probably also Eyvindr Finnsson drew such conclusions when he composed his genealogical poem Háleygjatal (c. 990) by taking inspiration from Ynglingatal.13 Most likely Þjódólfr also intended to create a genealogical poem when he composed Ynglingatal. There are some references in the text which indicate this.14 In stanza 9 (6), for instance, King Dómarr is called Fjǫlnis niðr “Fjǫlnir’s relative”. Hence, Þjódólfr demonstrated that Dómarr was related to Fjǫlnir, who appears at the very top of the list, in stanza 4 (1). In stanza 13 (10) the brothers and kings Alrekr and Eiríkr are designated with the kenning Dags fríendr “Dag’s kinsmen”. The kenning thus intimates that the brothers were related to King Dagr, who is mentioned earlier in the poem, in stanza 11 (8). When King Álfr in stanza 14 (11) is called Dǫglingr “descendant from Dagr”, Þjódólfr emphasized that he too was related to King Dagr. About King Óláfr trételgja, stanza 26 (21) says that he was “that offspring of well-known men from Uppsala” (sá áttkonr frá Uppsǫlum). In this study I therefore call the kings appearing in Ynglingatal “Ynglingar”, even though I am aware of the possibility that the “genealogy” in it is only a construction made by Þjódólfr, and should not be considered historical.15 Þjódólfr’s intention was most likely to make his lay as a genealogical poem over one family.

3.1.2 The Provenance of the Traditions of the Ynglingar The question of Ynglingatal’s background is very important in the present study. Ynglingatal can only be used as a primary source for the rulers of the Svear if it is based on traditions with an eastern Scandinavian origin. Claus Krag thinks, however, that these traditions originated elsewhere.16 He argues that most of the content of Ynglingatal was built on late western Scandinavian traditions, i.e., the same traditions which also formed the unreliable fornaldar sögur. According to him, the connection between the name Ynglingar and the Uppsala dynasty was nothing more than a twelfth-century construction. Þjódólfr probably composed Ynglingatal somewhere in southern Norway. In all likelihood he knew of vital traditions about the five or six last generations of Norwegian rulers. This part of the poem has a partly historical background. It is uncertain whether the link between the rulers of the Svear and the Norwegian rulers is historical, even if it cannot completely be excluded.17 By connect-

13 See F. Ström 1981, 446; McKinnell 2010, 33f. 14 Cf. J. Turville-Petre 1978–79, 61. 15 Since Bergsveinn’s arguments are very convincing I will hold this question open for the future. 16 Krag 1991, 219f. Cf. also Baetke 1964, 125–139. 17 According to most scholars, the connection between the Svea kings and the Norwegian rulers is not historical. See Baetke 1964, 72–85; Faulkes 1978–79, 97; J. Turville-Petre Religious Strategies For Political Authorization—a Case Study 67 ing the Norwegian pedigree to the well-known Uppsala dynasty, it has been argued, Þjódólfr was able to glorify the petty rulers of Vestfold.18 At any rate, Þjódólfr seems to have had access to older traditions about the Ynglingar when he composed the poem. This is indicated by expressions he used in the poem such as “often I have asked the wisest men [about where Dómarr’s burial place was located]”;19 “I have been informed [about Dagr’s death]”;20 and “I learned [that King Aðils’ life was to finish by means of a vile witch]”.21 These terms could of course be literary or conventionally expressions without any factual historical background. There are however other arguments supporting the notion that Þjódólfr built his poem on information from older (poetic) traditions about the Svea kings and their deaths. In what follows, the arguments for this and for an eastern origin of these traditions are set out. Elias Wessén referred to the names in Ynglingatal which reflected Upplandic conditions, as support for the eastern origin of the Ynglinga traditions. Personal names with Ing-, Ingi-, equivalents of ON Yngvi, for example, in st. 10 (7), were more common in eastern Scandinavia than in Norway during this period.22 They appeared especially in the realm of the Svear. Lars Hellberg has proposed that there is a group of place-names in the Mälaren region (Sweden), Ingeby (four places), Ingespjuta and Ingeberga, which can be associated with the names Ynglingar and Yngvi. These place-names have as the first element OSw Inge (< *ingi). According to him, this element is derived from a designation *ingvi (linguistically speaking the eastern Scandinavian equivalent of the ON personal name Yngvi) which was regarded as the Svea ruler’s honorific (tígnarnafn).23 Inge is thus explained as a designation for the king of the

1978–79, 52f.; McKinnell 2010, 28f. Birgit Arrhenius (2004a) has, however, recently argued that grave gifts in Óttarr’s mound, in Vendel parish, indicate a real connection between Norwegian chieftains and the elite of the Svear. 18 This is contested by Bergsveinn Birgisson (2007), who sees the poem as a nið. See above. 19 Ok ek þess opt/ of yngva hrør/ fróða menn/ of fregit hafðak, . . . Ynglingatal 9 (6). 20 Frák, at Dagr/ dauðaorði,/ frægðar fúss, . . . Ynglingatal 11 (8). 21 Þat frák enn,/ at Aðils fjǫrvi/ vitta véttr/ of viða skyldi . . . Ynglingatal 21 (16). See further Wessén 1964, xiv. Wessén states that the Swedish part of the poem was built on eastern traditions. Also Bergsveinn Birgisson (2007, 341, 415f.) argues that there were traditions among the Svear about ancient rulers’ deaths, which Þjóðólfr had access to when he created his poem. 22 Wessén 1924, 59, 64–67. 23 In Ynglinga saga 10 and 17 Snorri writes: “Freyr was also called Yngvi. The name of Yngvi was for a long time afterwards kept in his line as an honour-name (tígnarnafn). His race was thereafter called Ynglingar. . . . Everyone in their line was always called Yngvi or Ynguni, and all of them Ynglingar.” Heimskringla I, Ísl. Fornr. 26, 24, 34f. 68 chapter 3

Svear, strictly ‘the Ingvaeon’ (singular of the tribal designation Ingvaeones, PG *ingṷianiz). He was also a descendant of the mythical *Ing (OSw *Ingifrø, ON Yngvifreyr). A place-name such as Ingeby in Södermanland should thus be interpreted as “ ‘the hamlet’ belonging to the Svea king”.24 Perhaps a boundary mark called Ingefreds sten on Öland should be mentioned in this context. It is attested on old cadastral maps from the seventeenth century, and according to Hellberg this mark was called *Ingifrøys stæin(n) (OSw *Ingifrøs sten) in ancient periods.25 In addition, place-names occurring in the poem, such as Uppsalir (Uppsala),26 Fýri,27 Lófund,28 Ræningi (Ræningr),29 and Vendil30 can with greater or lesser certainty be identified with toponyms in the Mälaren region and thus support an eastern origin of the Ynglinga tradition.31 The specialist on toponymy Per Vikstrand has made a distinction between close-horizon and remote-horizon names in Ynglingatal.32 The latter do not say anything about the provenance of the poem. Uppsala and Fýri, for instance, were known throughout Scandinavia. The name Skúta, in stanza 6 (3) (á beði Skútu), a close- horizon name, was probably unknown outside the local community. It may refer to Skutån, a tributary of the River Fyris, located in the parish of Skuttunge, in the “hundred” of Bälinge, about ten kilometres north of Uppsala, or to a lost name in the parish of Vendel, indicated by the name Skottbro (j skotbroo 1457 (C7 f. 6 v.); skotbro 1471 (6/2 Stockholm RApp); skwtbro 1473 (C8 f. 19 v.) which means “the bridge over the creek *Skuta”.33 This name indicates the existence of a Swedish tradition behind at least some parts of Ynglingatal. Archaeological evidence likewise supports an eastern origin of the Swedish part of Ynglingatal. Sune Lindqvist compared the burial customs mentioned in the poem with archaeological finds and concluded that “the part about the Norwegian kings is in good accord with the burial in mounds (without burning)

24 Hellberg 1975; 1984–85 and 2014. 25 Hellberg 1986b. On Ing-, see Krause 1944. 26 B. Gräslund 1993. Cf. Brink 1999b, 38f., 48f., particularly note 15; Vikstrand 2013. 27 Jansson 1987, 90; Gustavson 1986, 11. 28 The expression á Lófhundi, may refer to an ancient Lóhund, an older form of Lyiahundare, today Lyhundra in eastern Uppland. See Noreen 1925, 239; Wessén 1964, 69. 29 The expression á Rœningi in Ynglingatal 25 (20) (J1, K, cf. rœninge in J2 and reiningi in F) corresponds to the Aspa inscription’s a rauniki, in Södermanland. See ch. 13 below. 30 On the problems of Vendil. Vikstrand 2004b. 31 Noreen 1925; Lindqvist 1936, 315–317; Wessén 1964. 32 Vikstrand 2004b. 33 Vikstrand 2004b, 377. Skuttunge probably means “the habitat of the people called *skutungar”. The name *skutungar means “those who live by the *Skuta”. Vikstrand 2004b, 373. Religious Strategies For Political Authorization—a Case Study 69 which was customary in Norway during the Viking period”. In the Swedish generations of Ynglingatal there are several stanzas showing “an intimate familiar- ity with Swedish customs and beliefs—about the burial customs which seem to have prevailed among the Swedish royal dynasties ever since the period of migrations”, i.e., the cremation customs.34 Stanza 26 (21), for instance, mentions that the fire swallowed King Óláfr the Woodcutter’s body by the lake. Also Vanlandi (st. 6 (3)) and Dómarr (st. 9 (6)) seem to have been cremated after death in places around Uppsala, i.e., Skutån and Föret. The most important monuments in the Mälaren region in the Early Merovingian Period are the Old Uppsala mounds and the mound in Vendel (Ottarshögen), which reflect cremation funerals of people belonging to the highest strata of Scandinavian society.35 The prerequisites for a mighty royal family are evident in the Uppsala area in the sixth–seventh centuries. In my opinion, evidence indicates that the poem Ynglingatal was composed by the Norwegian skald Þjódólfr sometime around AD 900. He probably had access to traditions of eastern Scandinavian origin. These traditions had flourished among the Svear at least during the ninth century and they were his sources for the Swedish generations. It seems reasonable to suppose that Þjódólfr received the accounts of the Swedish ancestors in the form of an active poetic tradition.36 Runic inscriptions confirm that both poetry and genealogical knowledge existed in eastern Scandinavia during the Viking period.37 For instance, the Rök inscription (Ög 136) from ninth-century Östergötland contains an eight-line stanza in fornyrðislag. It tells about ÞioðrikR, whoever he was: “He lived nine generations ago.” This demonstrates an interest in genealogical matters. Even if Þjódólfr partly built his poem on eastern traditions, he could have revised them and added ideas and ideological aspects reflecting the contemporary society of western Scandinavian. The six last stanzas of the poem obviously refer to cultural conditions prevailing in Early Viking Age Vestfold.38 In any case, I am inclined to view Ynglingatal as a primary

34 Lindqvist 1936, 302–307, 351. According to Ljungkvist (2005 and 2008a), the Uppsala mounds should be dated to the Merovingian Period. 35 Cf. Ljungkvist 2008a; Bratt 2008. 36 J. Turville-Petre 1978–79, 55. 37 See e.g. the runic inscriptions from Norra Sandsjö in Småland (Sm 71), where Ärnvard names himself and five paternal ancestors, and Malsta in Hälsingland (Hs 14), where Frömund names himself and six paternal ancestors. See e.g. Jansson 1987, 100f.; Hübler 1996; Sundqvist 2002, 151ff.; T. Zachrisson 2002. 38 Wessén (1964, xvii) noted that the naming custom among the Norwegian (or Danish) kings is different from that of the Swedish kings. This indicates that the last six stanzas were based on other traditions than the rest of the poem. 70 chapter 3 source for religious and ideological aspects of Svetjud (and southern Norway) during the Late Iron Age as well. When Þjódólfr in his poem gave the rulers of Svear certain attributes, he probably was convinced that these nobles in eastern Scandinavia were described in such a manner. I am therefore inclined to take these attributes as indications of a religious-mythic ruler strategy of the pre-Christian Svea kings.

3.1.3 Divine Descent The “genealogy” in Ynglingatal begins with the “earthly rulers”, including Fjǫlnir, Sveigðir, Vanlandi, Vísburr etc. This is rather peculiar, since the medieval versions of the Ynglinga genealogy trace the family to a divine origin.39 In Ynglinga saga, for instance, the list begins with Njǫrðr – Yngvi-Freyr – Fjǫlnir – Sveigðir etc., while Íslendingabók has Yngvi Tyrkjakonungr – Njǫrðr Svíakonungr – Freyr – Fjǫlnir – Sveigðir etc. Scholars have found it difficult to accept that Ynglingatal does not include these divine generations (i.e., Njǫrðr and (Yngvi-)Freyr). Some have thought that Ynglingatal as we know it is incomplete. They have suggested that it has lost some introductory stanzas which might have carried information about the Ynglinga kings’ connection to the gods.40 Others believe that the poem was preserved in oral tradition in the form we know it today, but was accompanied by some explanatory remarks, Begleitsprosa, before it was written down in the twelfth and thirteenth centuries.41 Still others think that the poem has its original shape, and that there were in fact no pre-Christian traditions connecting the Ynglinga kings with the gods. According to Walter Baetke, the divine descent of the Ynglingar was medieval speculation and cannot be supported by Ynglingatal. It can only be observed in the medieval versions of the genealogy. Ari and Snorri were elabo rating on pre-Christian traditions about Scandinavian royal families and were creating the fictions that these families were of divine descent in imitation of medieval Frankish and Anglo-Saxon genealogies.42 It is not necessary, however, to reconstruct any lost stanzas of Ynglingatal in order to find indications of the ruler’s mythical descent. Alrekr and Eiríkr are called Freys afspringr, ‘Freyr’s offspring’; Egill is týs ǫ́ttungr, ‘descendant of the god’; Aðils is Freys ǫ́ttungr, ‘Freyr’s descendant’ and Ingjaldr is goðkynningr, ‘of divine descent’. In the “Norwegian section” of the poem the

39 Sundqvist 2002, 156f. 40 Cf. Wessén 1924, 57; Åkerlund 1939, 125–128. F. Ström 1954, 34f. Critically considered by Baetke 1964, 89–103; Steinsland 1991, 182; Krag 1991, 85f. 41 Cf. Beyschlag 1950. Lönnroth 1986, 75. Steinsland 1991, 177f. 42 Baetke 1964, 70–103. Religious Strategies For Political Authorization—a Case Study 71 entire dynasty is called þróttar Þrós niðkvísl, ‘the kin-branch of the powerful (potent) man [i.e., the god]’. The Svea king Dyggvi is called allvaldr Yngva þjóðar, ‘the ruler over Yngvi’s [the Ingvaeon’s] people’ (st. 10 (7)). We do not know if Þjóðólfr was referring to a heiti/title of a ruler, a god or a mythical ancestor with the name Yngvi.43 If he regarded Yngvi as a mythical being, it would appear that the entire group was being lauded for its divine descent. Similar ideas are attested elsewhere in Scandinavia and the Germanic area.44 Originally, not only royal houses but also groups and peoples could have been praised as the descendants of gods. This strategy may have been gradually monopolized by noble families.45 Expressions and denominations of the kings found in Ynglingatal, therefore, clearly indicate that members of the Ynglingar were praised for their divine origin.

3.1.4 The Father of the Ynglingar In general, Freyr is regarded as the divine father of the Ynglingar, since e.g., Alrekr and Eiríkr in Ynglingatal are called “Freyr’s offspring”.46 This view has been challenged by the archaeologist Svante Norr. He states: “when we look only at Ynglingatal, the traditional view of Frey as ancestor god is far from obvious. If one god should be singled out as ancestor the argument for Odin is stronger.”47 In my opinion, these conclusions are not completely convincing. Norr proceeds from stanza 31 (26) and the expression niðkvísl . . . þróttar Þrós, which refers to the “Norwegian” King Óláfr. He intimates that Þrór is Óðinn and that the expression should be interpreted as ‘the lineage of the strong Óðinn’. This interpretation has been made by previous scholars.48 Þrór occurs as a heiti of Óðinn in Grímnismál,49 and in the list of Óðinn names in Gylfaginning.50 Most likely Þrór was an epithet of Freyr as well, for it is related to the verb þróaz, ‘to grow, increase’.51 In Skírnismál (For Scírnis) 38 the form Þroskr refers

43 Cf. Krause 1944; Sundqvist 2002, 161f. 44 See e.g. Germania 2 and Vǫluspá 1. 45 Cf. Faulkes 1978–79, 93f. 46 E.g. Jungner 1919, 79–84. Cf. Wessén 1924, 25f., 53–80; F. Ström 1954, 57. 47 Norr 1998, 86–89. In a recent publication Steinsland (2011b, 58) is also critical of the traditional view: “This chapter agrees that Óðinn had a strong position as god of kings and earls in the late Viking Age, but warns that Freyr’s relationship to rulers is not as clear-cut as has been assumed.” 48 See e.g. Finnur Jónsson in Skj. B1, p. 13; and most recently McKinnell 2010, 34. 49 Þrór þingom at. . . . “Þrór (I am called) at the thing. . . .”. Grímnismál st. 49. 50 Gylfaginning 20. 51 De Vries (1977, 623) interpreted þróast ‘wachsen, gedeihen’. Cf. Heggstad et al. 1993, 456. 72 chapter 3 to Freyr when he is having a tryst with Gerðr.52 The name Þrór, however, is not restricted to gods. It occurs as the name of a dwarf in Vǫluspá 12 and in Gylfaginning 14. Hence, Þrór cannot be interpreted as Óðinn or any other particular god. Freyr is the only god explicitly mentioned in Ynglingatal in connection with the kings’ ancestry.53 The medieval reception of the old traditions likewise held Freyr as the ancestor of the Ynglingar. Snorri, for instance, made use of an authentic tradition in Ynglinga saga when he separated Njǫrðr and Freyr from Óðinn’s kin and made the Vanir deities into ancestors of the Uppsala family.54 According to Ari and Historia Norwegie, Freyr/Froyr was one of the progenitors of the Ynglingar.55 Nothing is said about Óðinn there. Saxo also regards Frø (i.e., Freyr) as the Svea people’s special cult god. He tells us that the great men of the Svear were regarded as the sons of Frø:

The most valiant of the Svear were Ari, Haki. . . . Indeed, they were kinsmen [or close friends] of the divine Frø and faithful confederates of the gods.56

Place-names, including the element/name Frö-, further indicate that Freyr was an important god among the Svear.57 Some of these names are attached to the organization of society and indicate centrality. One example is Frösåkers härad, Uppland, where Frösåker most likely is an ancient central place, indicated by the field name Frössberg in the parish of Harg.58 Several names with

52 Wessén (1964: 77) argued that the mythical name Þrór in Ynglingatal referred to Freyr, since he was the forefather of the Ynglingar. Þrór was probably also a male name. It appears in Swedish runic inscriptions. See Källström 2010b. Marold suggests that Þrór in Ynglingatal refers to ‘an unidentified divine being’. SkP I, 56. 53 It is very uncertain whether we should interpret týss ǫ́ttungr (st. 14) as “descendant of Týr”, as suggested by e.g. Norr 1998, 85 and Marold, SkP I, 31f. Cf. Noreen 1925, 203, 235; Wessén 1964, 66. 54 In the “Prologue” of Snorra Edda, there is another version: “And Óðinn took with him a son of his whose name was Yngvi, who became king in Sweden, and from him are descended the family lines known as the Ynglingar.” It has been argued that the “Prologue” was never a work by Snorri, but a late addition. It has also been shown that genealogical ideas in Snorra Edda were clearly influenced by British regnal lists. Cf. von See 1988, 18–30. See also Faulkes 1978–79; J. Turville-Petre 1978–79. 55 Íslendingabók, Ísl. Fornr. 1, 27; Historia Norwegie IX. (Ed.) Ekrem & Mortensen, 74f. 56 At Sueonum fortissimi hi fuere: Ari, Haki. . . . Qui quidem Frø dei necessarii erant et fidissimi numinum arbitri. Saxo 8.3.11. Text Friis-Jensen. Trans. Fisher. See also Saxo. 3.2.13; 9.4.1.; 6.5.10. 57 Wessén 1923, 1ff.; Hellberg 1986a, 54; T. Andersson 1992c, 247ff.; Vikstrand 2001, 55ff., 72ff. 58 Vikstrand 2001, 69f. Religious Strategies For Political Authorization—a Case Study 73

Frö- may directly or indirectly be connected to tuna places, i.e., ancient administrative centres in the settlement districts. Frustuna (< *Frøstuna), in Frustuna parish, Södermanland, is one example.59 Thus, the place-names in the Mälaren area support a connection between ruling power, organization of society, and the god Freyr. If we combine the information in Ynglingatal with information found in other sources, it seems reasonable to assume that Freyr was regarded as the father of the Ynglingar, as early as in the Viking Period. The designation of Freyr as blótgoð svía also supports the assumption, although it is evidenced in late sources.60

3.1.5 The Mythical Mother of the Ynglingar Ynglinga saga says about Freyr that: “His wife was called Gerðr, the daughter of Gymir, and their son was called Fjǫlnir.”61 This is the only text to mention the god Freyr and the giantess Gerðr as Fjǫlnir’s parents. Skírnismál (For Scírnis) describes the union between Freyr and Gerðr, but nothing is said about a child. Snorri gives similar information in Gylfaginning.62 In my opinion, it is not completely safe to place Gerðr as a mythical mother at the upper end of the Ynglinga genealogy, since this is attested only in Ynglinga saga. Perhaps it was Snorri’s own conclusion that Fjǫlnir was the son of Gerðr and Freyr.63 Without doubt, however, there was most likely a mythical mother in the traditions about the Ynglingar, although we do not know for sure who she was. Many scholars have argued that a hieros gamos occurred as an important element in the ancient Scandinavian royal ideology.64 Gro Steinsland suggested that this type of holy marriage was very specific in the Norse traditions and took place between a god and a giantess.65 The Ynglingar, for instance, were the products of the marriage of Freyr and the giantess Gerðr. According to Steinsland, the odd marriage between a god and a giantess brought forth something new, a prototypical ruler. It created a bridge between mythical

59 Cf. K. A. Holmberg 1969; Hellberg 1986a, 50–54; T. Andersson 1991, 201f.; SOL, 328ff. 60 See e.g. Óláfs saga Tryggvasonar, Flateyjarbók I, 339 and Ǫgmundar þáttr dytts, Ísl. Fornr. 9, 114. 61 Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 24. Trans. Hollander, modified. 62 Cf. Gylfaginning 37. 63 See Sundqvist 2002, 167. Cf. Hultgård 1994, 78. 64 See ch. 1. 65 Steinsland 1991; 2000. Scholars have supported her theory, e.g. Herschend 1996; Munch 2003b; Wickström 2004. 74 chapter 3 and historical levels, with consequences.66 Other ruler families in ancient Scandinavia, according to Steinsland, were also the products of a marriage between a god and a giantess. In Háleygjatal, for instance, the Earls of Lade originated from the union between Óðinn and the giantess Skaði.67 Other scholars have been sceptical of Steinsland’s theory.68 Some have denied the hieros gamos between the god and giantess in general, on the grounds that the normal hierogamy is between the god/ruler and the goddess.69 That a prototypical king is always a product of a meeting between a god and a giantess has also been questioned.70 The Skjǫldungar of Denmark, for instance, according to Snorri, were descendants from Óðinn’s son Skjǫldr and the goddess Gefjun.71 The mythical father is not always a god; some rulers in Norway were descended from Fornjótr, who perhaps was a giant (Hversu Noregr byggðisk).72 Steinsland’s idea that the abnormal exogamy leads to the ruler’s ignoble death has also been contested. Lotte Motz states: “All kings must die, and some die in an honourable and others in a dishonourable manner.”73 In my opinion the reliable sources tell us only that the Uppsala family reckoned its origin from the gods. It was most likely Freyr, the blótgoð svía, who was praised as the Ynglingar’s particular god. It is harder to identify the mother. The only information we can reach from Old Norse sources, however, is that Gerðr was Freyr’s consort, and according to Snorri they had a son called Fjǫlnir. Steinsland’s theory is thus at least plausible. In what follows (particularly in ch. 11) I will apply it when interpreting other sources (e.g., the gold foils), i.e., that the Ynglingar were regarded as Freyr’s and Gerðr’s offspring.

66 Steinsland 1991; 2000. In a more recent article, Steinsland (2011b, 21ff., 58) hesitates regarding her former view of Freyr and Gerðr as the divine parents of the Ynglingar, since the Ynglinga genealogy might be seen as a construct from thirteenth century (cf. Bergsveinn Birgisson 2007). 67 The text is quoted below. Adam mentions that the cruel Hákon had sprung from Yngvar and from a giant. Haccon iste crudelissimus, ex genere Inguar et giganteo sanguine descendens, . . . Adam II,25. This passage may, on the other hand, be a secondary interpolation. 68 E.g. La Farge 1994; Hultgård 1994; Clunies Ross 1994; Motz 1996; Lönnroth 1997; Krag 2001; Sundqvist 2002; Frank 2007; Cöllen 2011. 69 Hultgård 1994. 70 Motz 1996. 71 Heimskringla I, Ísl. Fornr. 26, 14f. 72 Flateyjarbók I, 21–24. The name Fornjótr could be explained ‘Ancient-Jutlander (or Giant)’, ‘Early-User, Early-Destroyer’, ‘One-who-enjoys-sacrifices’, or ‘Ancient-Screamer’. Lindow 2001, 119. 73 Motz 1996, 79. Religious Strategies For Political Authorization—a Case Study 75

3.1.6 Mythical Marriage and the Ruler’s Death As has been noted above, expressions in Ynglingatal indicate that the Ynglinga kings were related to the mythical world by the notion of divine descent. The god Freyr was regarded as the father of the family. Some statements and expressions referring to mythical females in Ynglingatal indicate that the rulers were also related to the Other World in other ways. It seems as if these females had an erotic relationship with the Ynglinga kings. They were also in different ways connected to the kings’ deaths. Most likely these themes reflect aspects of a religious-mythic ruler strategy.74 In Ynglingatal 12 (9), for instance, a wedding between King Agni and a mythical female has been observed. Crucial for this interpretation is the expression Skjalfar rǫ́ð.75 The first word is the singular genitive of a mythic female name Skjálf. The second word rǫ́ð is probably the plural of ON ráð neutr. ‘decision, decree, plan, venture’, but has also been interpreted as ‘marriage, intercourse’.76 The expression may thus be interpreted as “Skjálf’s plan or marriage”. In the context of the stanza, it seems as if Skjálf’s “plan or wedding” included the hanging of her own husband:

Þat telk undr, ef Agna her Skjalfar rǫ́ð at skǫpum þóttu, þás gœðing með gullmeni Loga dís77 at lopti hóf, hinns við taur78

74 Cf. e.g. F. Ström 1954; Steinsland 1991. 75 For an overview of the interpretations of this kenning, see Evans 1981, 91, 94f., 105. 76 Cf. Noreen 1892, 208–221; Åkerlund 1939, 89; Steinsland 1991, 203; SkP I, 23f. 77 K and F have loga, while J has lolka. It is probable that J is wrong in this case. Cf. Skj. A1, 9. 78 Taur(r) may be a place-name somewhere in Stockholm. Scholars have argued on good grounds that Ynglingtal’s við taur probably means “at Södertörn”, south of Stockholm. See e.g. Elmevik 1986. Cf. Dillmann 2000b. Historia Norwegie also mentions a place- name, Agnafit, in connection with the hanging of Agni. Most likely it refers to a place (perhaps Norrström) in Stockholm. See Óláfs saga Helga 7, in Heimskringla II, Ísl. Fornr. 27, 8. See also Wessén 1964, 63. Alternative interpretations have also been made. Noreen interpreted taur(r) as an appellation ‘guldhalsband’ and compared it with Lat torques, 76 chapter 3

temja skyldi svalan hest Signýjar vers.79

The wedding motif may also be emphasized by the kenning and designation of Skjálf, loga dís. This poetic expression has been interpreted as a genitive plural of a plural tantum ON *log ‘marriage’ (cf. OIr luige ‘oath’; Goth liugan ‘to marry’; OFris loga ‘to get married’). The second element in the kenning, dís, means either ‘woman’ or ‘female mythical being’. Thus the kenning may be interpreted as “marriage woman, female mythical being”.80 The medieval reception of Ynglingatal regarded Skjálf as the wife of Agni. Historia Norwegie, for instance, states that Agni was killed by his own wife.81 Ynglinga saga mentions that the hanging of Agni took place in the evening after a banquet, perhaps the wedding feast.82 Skjálf’s identity has been debated. Folke Ström argued that she is identical to the goddess Freyja, i.e., the divine mother of the royal Uppsala family.83 According to him, this tradition concerns the universal ideology of sacral kingship, where the divine king after a hieros gamos was sacrificed to his own consort, that is, the fertility goddess. A similar pattern, according to Ström, was also present in the traditions about King Dómaldi. In Ynglingatal 8 (5) Dómaldi was sacrificed/slain (sóa) by the Svear, because they were eager for crops (árgjarn). In order to find the

torquis. Noreen 1925, 201, 227f. Also D. A. H. Evans connected taur(r) with Lat torques or thesaurus ‘necklace’. Evans 1981, 92f. Cf. McKinnell 2010, 29 and SkP I, 23f. 79 “I think that it is strange/ if Agni’s men/ thought that/ Skjálf’s marriage plans were convenient,/ when “the wedding dis” [Skjálf]/ raised aloft/ the ruler [King Agni]/ with the golden neck-ring,/ he who at Taur,/ had to tame/ Signý’s husband’s [Hagbarðr’s]/ cold horse [= the gallows].” My trans. 80 This interpretation is supported by Noreen 1925, 226f.; F. Ström 1954, 40f.; Evans 1981, 105. The kenning could also be interpreted plainly as Loga dís ‘Logi’s sister’, i.e. the inter pretation Snorri made. See e.g. Wessén 1964, 62f. Some scholars have felt that the kenning is almost impossible to explain, see e.g. Åkerlund 1939, 89; Marold 1983, 114, 121; SkP I, 24. 81 Istum uxor sua iuxta locum Agnafit propriis manibus interfecit suspendendo ad arborem cum catena aurea “[Agne], whose wife dispatched him with her own hands by hanging him on a tree with a golden chain near a place called Agnafit.” (Ed.) Ekrem & Mortensen, 76f. 82 Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 37–39. Folke Ström (1954, 40) interpreted the banquet as a wedding feast. 83 F. Ström (1954, 39ff.) supported this interpretation by references to Wessén’s onomastic investigation, since Wessén regarded Skjálf as a divine mother of the Skilfingar, i.e. perhaps a name of the ancient ruling family at Uppsala. See e.g. Wessén 1924, 54. The connection between Skjálf and the Skilfingar is, however, much debated. See e.g. Noreen 1925, 201, 225; Åkerlund 1939, 89; Baetke 1964, 125–139; Evans 1981. Steinsland 1991, 203ff.; Gade 1985. Religious Strategies For Political Authorization—a Case Study 77 recipient of this sacrifice, Ström consulted Historia Norwegie, where it says that Dómaldi “was hanged by the Svear as a sacrificial offering to Ceres to ensure the fruitfulness of the crops”.84 Ström stated that Ceres was an interpretatio romana of Freyja. The divine king had failed to maintain a positive relationship with this fertility goddess, i.e., to bring prosperity to his land, and was therefore sacrificed to her.85 In my opinion, Ström’s interpretation leans too much on the general paradigm of the sacral kingship theory as produced by Frazer and others. The complete pattern, including both the hieros gamos motif and the royal sacrifice, was never discovered in the textual traditions relating to one single king, but instead by compiling different traditions referring to several kings. When it comes to the Agni traditions, for instance, the royal sacrifice is not clear. The traditions only say that Agni was hung by his wife Skjálf. In the Dómaldi tradition, on the other hand, the marriage motif is totally missing. It is hard to deny, however, that the earthly kings in Ynglingatal are often related to female mythical beings, as Ström suggests. These females either kill or take the kings to the realm of death. It seems as if the rulers had intercourse with them there. In stanza 10 (7) Dyggvi’s death occurs in symbolic language with sexual undertones.

Kveðkat ek dul, nema Dyggva hrør glitnis Gnǫ́ at gamni hefr, þvít jódís Ulfs ok Narfa konungmann kjósa skyldi, ok allvald Yngva þjóðar Loka mær of leikinn hefr.86

84 [Cujus filium] Domald Sweones suspendentes pro fertilitate frugum Cereri hostiam obtulerunt. Historia Norwegie IX. (Ed.) Ekrem & Mortensen, 74f. 85 Baetke (1964) opposed this interpretation as well as the existence of divine kings in Scandinavia. According to him, the ruler gave good crops and fertility among cattle only when he maintained the sacrifice. In a later essay F. Ström (1968) argued that Domaldi was sacrificed because he had lost his “luck” (Königsheil). This theory has also been contested. See e.g. Lönnroth 1986. Cf. Sundqvist 2002, 243ff. and the literature cited there. 86 “I do not tell lies,/ that Glitnir’s woman/ has Dyggvi’s corpse/ for pleasure./ The “horse- dis” of/ Ulfr and Narfi/was choosing/ a king-man./ Loki’s maiden/ had the ruler/ of Yngvi’s people/ to play with.” My trans. based on Wessén 1964. 78 chapter 3

The identity of the woman in the stanza is debated.87 The kenning jódís Ulfs ok Narfa may however facilitate the identification, especially as the poet also explicitly refers to this being as Loka mær, “Loki’s daughter”. It must be Hel.88 She takes Dyggvi’s body “to play with” and “for pleasure”. Dyggvi’s departure from this life is thus described through sexual allusions.89 Dyggvi comes to the realm of death for an erotic tryst with Hel. Gro Steinsland has argued that the theme of Eros and death was crucial to ruler ideology in general in ancient Scandinavia. Death is represented as a hieros gamos, a holy wedding between the dead ruler and a female representative from the realm of death. This theme can also be seen in several stanzas of Ynglingatal.90 In the Norwegian parts of Ynglingatal 29 (24), for instance, a mythical female Hveðrungr’s maid [i.e., Hel]91 invited the third chieftain [i.e., Halfdan] to a tryst92 beyond the world.93 Similar death symbolism may also occur in iconography. Picture stones from Gotland, dated to about AD 700–900 and appearing in aristocratic milieus, depict a man on a horse being welcomed by a woman holding a cup. The pictures are often interpreted as the scene of the dead arriving in Valhǫll, welcomed by a valkyria; she gives the dead warrior a horn with mead. The ships on these stones are usually interpreted as the craft which take the dead to the realm of death. One example is the Tjängvide stone, which contains both the ship scene and the Valhǫll scene (fig. 1).94 Erotic elements have been noticed; for instance, the shape of the stone has been interpreted as phallic. But perhaps the further interpretations of the sword hilts of the men in the ships or on the horse as erect penises are exaggerated.95 At any rate, we can accept that these picture stones can involve instances where death is described in erotic terms.96

87 See e.g. Sundqvist 2002, 285f. 88 Marold 1983, 120 and SkP I, 19f. 89 Steinsland 1991, 2000; 2011b, 23. Cf. F. Ström 1954, 38–49; McKinnell 2010, 31. For a quite different interpretation, see Bergsveinn Birgisson 2007, 297–325. 90 Steinsland 1991; 1994; 2000, 57ff. 91 Cf. Wessén 1964, 76. Marold 1983, 131. 92 Cf. Steinsland (1991, 185f.), who suggests that til tings in this context has the connotation of tryst. Cf. SkP I, 52. 93 Ok til þings/ þriðja jǫfri/ Hveðrungs mær/ ór heimi bauð, . . . 94 See also e.g. Ardre, and Halla Broa, Gotland. Lindqvist 1941–42; Nylén & Lamm 1987, 68f., 91. 95 Steinsland 1994. 96 These ideas appear in sources referring to other parts of ancient Scandinavia. See e.g. Hákonar saga hins gamla, Flateyjarbók II, 98. Sonatorrek (st. 10), Skj. B1, 35. Cf. Steinsland 1994. Religious Strategies For Political Authorization—a Case Study 79

Figure 1 Tjängvidestenen (G 110), Alskog. Gotland. Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet

In my opinion, the relation between the Svea kings and the mythical females in Ynglingatal, and in other sources, comprises clear features of a mythic-cosmic ruler strategy (i.e., propaganda). The ruler was supplied in death, as in life, with riches. He not only had weapons, food and drinking horns for his feast in the Other World, but also partners for intercourse. In the realm of death the Ynglinga kings came to a place where not only banquets but also erotic relations were common. The departed leader continued to be wealthy and successful in the cosmic world, so much so that even mythical beings were ready to have intercourse with him. Perhaps this was only one aspect of the Valhǫll mythology glorifying the life beyond, as typically represented in Viking Age warrior society (see further ch. 15 below).

In my opinion a certain mythic-cosmic ruler strategy occurred during the Late Iron Age in Svetjud (and southern Norway). This strategy appears in the poetic traditions about the Swedish-Norwegian kings called the Ynglingar. According to these traditions, it seems as if these rulers were regarded as descendants of the god Freyr. Most likely they also counted descent from a mythical mother. Her identity is uncertain, but Gerðr is at least a good suggestion. The Ynglinga 80 chapter 3 kings are related to the mythical world in other ways, too. It seems as if the kings’ deaths in Ynglingatal were sometimes related to “mythical marriages”, where the rulers were sexually united with mythical females.

3.2 The Háleygja Kin (Hlaðajarlar)

According to Snorri Sturluson, King Haraldr organized Norway by appointing one earl in each fylki. Snorri states, for instance, that Earl Hákon Grjótgarðsson was made earl in c. 900 over Strindafylki, in Trøndelag.97 The king also gave him the ruler site of Lade to reside at. This family was therefore called the Earls of Lade (Hlaðajarlar). Whether we can rely on Snorri regarding these matters is uncertain. However, it is agreed among historians that the Hlaðajarlar took control over an extensive area in the tenth century, stretching from Trøndelag northward. During the reign of Earl Hákon Sigurðarson (who died in 995), this family probably controlled almost all of coastal Norway. The contemporary poem Vellekla, for instance, says that Earl Hákon ruled over the land of sixteen earls. There are several praise poems composed in honour of Earl Hákon, dating back to the end of the tenth century, for instance Háleygjatal (c. 985), Hákonardrápa (c. 990), and Vellekla (990). These poems were created within a specific historical context where a conflict between paganism and Christianity was raging in Norway. The pagan ruler Earl Hákon succeeded in destroying the attempt of the Danish kingdom to regain sovereignty over Norway in the battle of Hjǫrungavágr, c. AD 986. His victory provided the stimulus and background for these pagan poems.98 In the following analysis I will restrict my presentation to some mythic-cosmic motifs in them which play a fundamental role for the religious strategies for rulership. These themes concern the divine descent of Hákon and his relationship to a female mythical being. They seem- ingly correspond well with the ones seen in Ynglingatal. A thorough investigation shows, however, that they were partly constructed, styled and outlined in a slightly different manner from the traditions stemming from the Ynglingar.

3.2.1 A Mythical Descendant In the genealogical poem Háleygjatal, composed by Eyvindr Finnsson (skál- daspillir), the forefathers of Earl Hákon are counted back to the divine generations. Only nine whole and seven half stanzas of the poem are preserved. They

97 Haralds saga ins hárfagra, in Heimskringla I, Ísl. Fornr. 26, 99. 98 F. Ström 1981. Religious Strategies For Political Authorization—a Case Study 81 appear in e.g. the manuscripts of Fagrskinna, Snorra Edda and Heimskringla. The poem begins as follows:99

(1)Viljak hljóð at Hǫ́ars líði, meðan Gillings gjǫldum yppik, meðan hans ætt í hverlegi galga farms til goða teljum. . . . (3) Þann skaldblœtr skattfœri gat ása niðr við jarnviðju; þás þau meir í Manheimum skatna vinr ok Skaði byggðu.

(4) Sævar beins, ok sunu marga Ǫndurdís við Óðni gat.

Snorri comments on these stanzas thus: “Njǫrðr married a woman who was called Skaði. She would not have intercourse with him, and later married Óðinn. They had many sons. One of them was called Sæmingr. . . . Earl Hákon the Mighty reckoned his pedigree from Sæmingr.”100 According to these passages,

99 “(1) I desire silence/ for Har’s ale [Óðinn’s mead = poetry]/ while I raise/ Gilling’s payment [= the mead],/ while his descent/ in pot-liquid [the mead = poetry]/ of gallows-cargo [=Óðinn]/ we trace to gods. . . . (3) That distributor/ of gifts [= Earl Sæming],/ Æsir-kin got,/ [the one who is] worshipped by skalds [=Óðinn],/ with giantess [= Skaði];/ the time when they,/ the men’s friend [=Óðinn] with Skaði,/ in Manheim. (4) . . . of the bone of the sea [the stone’s],/ and sons many,/ the ski-goddess,/ had with Óðinn.” My trans. This “restored text” is taken from Skj. B1, 60–62. For the actual witnesses, see Skj. A1, 68ff. See also SkP I, 195–213. 100 Njǫrðr fekk konu þeirar, er Skaði hét. Hon vildi ekki við hann samfarar ok giptisk síðan Óðni. Áttu þau marga sonu. Einn þeira hét Sæmingr. . . . Til Sæmings talði Hákon jarl inn ríki langfeðgakyn sitt. Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 21f. 82 chapter 3

Earl Hákon saw Óðinn as his divine forefather. Some earls in Háleygjatal are, however, said to be descended from Freyr or Yngvi. Earl Hákon Grjótgarðsson, for instance, is called Freys ǫ́ttungr “Freyr’s descendant”. Snorri also stated in other passages that the Lade Earls descended from Yngvifreyr or Ingunarfreyr. Some scholars have therefore argued that Óðinn’s position at the top of the genealogical list is a sign of western European influence, i.e., the Anglo-Saxon royal lists.101 It has also been argued that Eyvindr applied a learned euhemeristic approach since he stated that Óðinn and Skaði lived in Manheimar, instead of Goðheimar, which would have been their natural home.102 Like the Anglo-Saxon authors, Eyvindr also interpreted the old deities as human beings. This has likewise been regarded as a sign of Western and Christian impact on the poem. In my opinion, it is not necessary to see any Anglo-Saxon influence on Háleygjatal. Even if Snorri made a euhemeristic interpretation of the Óðinn- Skaði myth, it is far from certain that Eyvindr also had this perspective. Admittedly Eyvindr stated that Óðinn and Skaði lived in Manheimar. But we do not know how Eyvindr conceived the mythological topography and how he designated the different places located there.103 The idea that Óðinn was regarded as the ancestor of the Lade Earls may very well be based on old native traditions from Trøndelag or Hálogaland.104 The skaldic poem Vellekla, for instance, reports that Earl Hákon was called Yggs niðr, “Óðinn’s relative”. In Hákonardrápa, Hákon’s conquest of the land is described symbolically as a marriage between the earthly ruler and the wild land, here represented as a mythical female (see below). In this meta- phoric language Hákon is also identified with Óðinn. A similar symbolism is also evident in Háleygjatal 15.105 It should be noted that Eyvindr in the introduction to Háleygjatal mentioned that Óðinn begat Sæmingr (the mythical father of the Lade Earls) on Skaði. According to Gro Steinsland, this foundation myth resembles the one found in Ynglingatal (and Yngling saga). The ruling families described in these poems both derived from a peculiar marriage between a god and a giantess.106 I agree with Steinsland that this type of marriage is probably evident in the

101 Bede, for instance, made Woden (Óðinn) into an ancestor of ancient British kings in genealogies. See Bede, I,15. See also J. Turville-Petre 1978–79, 63; Faulkes 1978–79, 96. 102 Cf. Bjarni Aðalbjarnarson in Heimskringla I, Ísl. Fornr. 26, xxxii. Faulkes 1978–79, 97f. 103 On interpretations of Manheimar in this context, see Steinsland 2011b, 30 and SkP I, 200. 104 Cf. Wessén 1924, 34f.; F. Ström 1981, 447; Steinsland 2011b, 32f., 39. 105 See Sundqvist 2007, 101ff. 106 See Steinsland 1991. It is somewhat uncertain whether we should classify Skaði as a giantess. See Näsström 1995, 51f.; Lindow 2001, 268ff.; Simek 2006, 286f. In a genealogical sense she is undoubtedly a giantess, since she is the daughter of Þjazi. The denomination of her in Háleygjatal, jarnviðja, also indicates that she was a giantess. Steinsland 2011b, 29. Religious Strategies For Political Authorization—a Case Study 83 genealogy of the Lade Earls. However, it is not completely safe to place the giantess Gerðr as a mythical mother at the upper end of the “Ynglinga genealogy”, even if it is a good suggestion (see above). Anyhow, there is no doubt that both Ynglingatal and Háleygjatal clearly indicate that both the “Ynglinga kings” and the Lade Earls claimed that their families were of divine descent on the male side.

3.2.2 Mythical Wedding with the Land In some stanzas of Hallfrøðr Óttarsson’s Hákonardrápa Earl Hákon’s victory at Hjǫrungavágr is expressed in metaphorical terms with a mythical symbolism which includes Óðinn:

(3) Sannyrðum spenr sverða snarr þiggjandi viggjar barrhaddaða byrjar biðkvǫ́n und sik Þriðja.

(4) Því hykk felygjanda frakna, ferr jǫrð und menþverri, ítra eina at láta Auðs systur mjǫk trauðan.

(5) Rǫ́ð lukusk, at sá síðan snjallráðr konungs spjalli átti einga dóttur Ónars, viði gróna.

(6) Breiðleita gat brúði Báleygs at sér teygða stefnir stǫðvar hrafna stála ríkismǫ́lum.107

107 “(3) The war-ship’s brisk lord/ with the veracity of the sword/ entices under him Þriði’s [=Óðinn’s] beloved,/ whose hair is the foliage of pine trees./ (4) Therefore I think that the thrower of the spear/ is very unwilling to leave Auðr’s glorious sister [= Earth, the land] alone;/ the land prostrates herself /under the ring-waster./ (5) The outcome of the union/ was afterwards this quick-witted intimate friend of kings/ gained possession in marriage of Ónarr’s/ forest-grown only daughter./ (6)The commander of ships [= ravens of the harbour]/ was able to attract to himself/ Báleygr’s [=Óðinn’s] broad-faced bride [= the land, Norway]/ with the help of the sword’s powerful discourse.” Text and trans. F. Ström 1981, 452f. Cf. Skj. B1, 147f. Only nine half-stanzas are preserved of this poem. It has been argued that the sixth stanza should be the first in this sequence. See F. Ström 1981. 84 chapter 3

In this poem the earl’s conquest of Norway is described symbolically as a marriage between the earthly ruler and the wild land, here represented as a mythical female. The ruler must tame his land/bride with the sword. The historical event at Hjǫrungavágr is transposed to a mythical scene creating a cosmic symbolism.108 It seems as if Hákon here is identified with Óðinn. He celebrates a wedding with this mythical woman, i.e., the personification of the country. Óðinn is here called by his heiti name Þriði.109 The land/mythical female is invoked with the kenning “Þriði’s beloved, whose hair is the foliage of pine trees”, emphasizing the correspondence between earthly and mythical levels. A similar symbolism appears in the kennings “Ónarr’s forest-grown only daughter” and “Báleygr’s (Óðinn’s) broad-faced bride” also referring to the land of Norway as a mythical female.110 This symbolism also occurs in other skaldic poems. In Háleygjatal (st. 15) we find a metaphor reflecting how Earl Hákon takes possession of the land: “[Hákon,] under whose arms Val-Týr’s bride, all the way eastward to Agdermen’s dwelling, now lies.”111 The erotic undertones in this stanza are also quite obvious.112 There have been some objections to the idea that this mythical marriage was part of a genuine Scandinavian ideology. Dag Strömbäck has shown that the symbolism in Hákonardrápa has similarities to the Christian notions of Jerusalem as the bride of God seen in the Old Testament and “the idea of Christian bishop’s ‘marriage’ with his church, in which his episcopal ring is usually regarded as symbolic”.113 Folke Ström, on the other hand, has referred to pre-Christian Irish traditions which mention that the king celebrated a symbolic wedding feast (OIr banais ríghi) with his country at the inauguration ceremony.114 In Ireland, the land was represented by a goddess, sometimes called Medb (Medhbh). According to Ström a similar kind of hieros gamos appears in Hákonardrápa where Earl Hákon (identified as Óðinn) ritually married a local

108 Cf. F. Ström 1981; 1983; Steinsland 1991; 2011b, 30ff. 109 Falk 1924, 30. 110 Ónarr, according to Skáldskaparmál, is a mythical being of the giant race. He is Nátt’s second husband and the father of Jǫrð. Báleygr is a heiti of Óðinn. Falk 1924, 4. 111 þeims alt austr/ til Egða býs/ brúðr Val-Týs/ und bœgi liggr. Trans. F. Ström 1981, 448. Cf. Skj B1, 62. 112 Fragments of this symbolism appear in some of the most ancient skaldic poems. In Bragi’s Ragnarsdrápa 5 the earth (land) is depicted as Hergauts vina “Óðinn’s mistress, lady-friend”. The land is called Svǫlnis ekkja “Óðinn’s woman, wife” in Þjóðólfr’s Haustlǫng (st. 15) and in a lausavísa (21) by Egill it is designated Bors niðjar beðja “Óðinn’s bedfellow”. 113 Strömbäck 1975, 70, note 1. 114 E.g. F. J. Byrne 1973, 16f.; MacCana 1970, 117–121. Critically considered by Maier 1994, 163f. Religious Strategies For Political Authorization—a Case Study 85 fertility goddess called Þórgerðr Hǫlgabrúðr.115 This hieros gamos was part of an old local ruler ideology used by the conservative Earls of Lade in Trøndelag and Hálogaland. There are reasons to believe that Hákonardrápa reflects structures of an ancient mythic-cosmic ruler strategy in Trøndelag. That the wedding symbolism referred to aspects of inauguration seems likely. By transferring the historical event of Hákon’s conquering of Norway to a mythical scene, and by identifying the ruler with Óðinn and the land with a female mythical being, the perspective is moved from the microcosmos to the macrocosmos. The local perspective thus fades and the ruler’s activity acquires a cosmic dimension. Placing the earl in a mythical context was an effective means in homage poetry and was an instrument of political propaganda for Earl Hákon.116 The mythical wedding in Hákonardrápa is thus styled differently from the symbolic marriage in Ynglingatal. In Ynglingatal, the deaths of the rulers are described as a tryst between the earthly ruler and a mythical female, while Hákonardrápa depicts Hákon’s victory and inauguration as a sexual union between the earl/Óðinn and the land/mythical female. Since the contexts of Ynglingatal and Hákonardrápa varied, as did the themes they were built on (death and inauguration respectively), the mythical strategy has acquired different conceptions in these poems.117 At any rate, there are some striking similarities between the mythic-cosmic ruler strategy found in Ynglingatal and the poems praising Earl Hákon. Both the Ynglingar and the Earls of Lade counted themselves as descendants of the gods, Freyr and Óðinn respectively. It also seems as if the members of both families related themselves sexually to female mythical beings.

115 Roberta Frank (1978, 62ff.) has suggested that the hieros gamos myth behind Hákonardrápa concerns Freyr and Gerðr. This idea is rejected by F. Ström (1983, 72) since the poem obviously refers to Óðinn. In a later study Frank (2007) is sceptical of the hieros gamos theory when applied to Old Norse skaldic traditions. 116 Cf. F. Ström 1981; 1983. See also Steinsland 1991. 117 A similar strategy to that seen in Hákonardrápa also appears in connection with the kings of the Svear and Danes. Around 1018 Óttarr svarti praised King Óláfr (sœnski) for seizing a kingdom: “The ruler takes for certain battle-Freyr’s [Óðinn’s] desirable wife [Jǫrð =land], loveless, in the east. The all-powerful commander of men has a good life.” Skj. B1, 267. The verb taka has here the double meaning ‘conquer, have sex with’. See Frank (2007, 182f.) for this example and skaldic references to Knútr the Great. 86 chapter 3

3.3 The Þórsnesingar

It appears as if the traditions about the Icelandic chieftain family of the Þórsnesingar display another type of religious strategy for rulership compared to the ones found in Ynglingatal and Háleygjatal. These traditions are mainly written down in Eyrbyggja saga and to some extent also in Landnámabók.118 Eyrbyggja saga (c. 1245–1265) exists in a large number of manuscripts, but no definitive edition has yet appeared.119 The most important editions have been edited by Guðbrandr Vigfusson and Einar Ól. Sveinsson. In the present study I follow the text produced in Íslenzk Fornrit, i.e., the edition of Einar Ól. Sveinsson. The author of Eyrbyggja saga is unknown, but we can sometimes identify his sources. Scattered through the prose in this saga, there are some separate stanzas ascribed to different skalds. The majority of them are not genuine, however. A few of them may be regarded as Viking Age sources for some passages in the saga. The early chapters of the saga, which are of interest here, borrow from Landnámabók. Landnámabók may thus also be considered as an important source for the saga. On the surface the saga may be regarded as a historical record, tracing the lives of several generations of Þórólfr’s kinsmen, i.e., the Þórsnesingar, from the late ninth century to the early eleventh century. But even if there are kernels of historical truth in some of the stories they are undoubtedly surrounded by much apocryphal matters. Hermann Pálsson and Paul Edward’s characterization of the saga as “fictionalized history” is most striking.120 Other scholars have argued, however, that there might be certain elements in it, which may inform about quite ancient structures and ideas.121 Eyrbyggja saga tells about Þórólfr Mostrarskegg, who left Norway and settled in the Snæfellsnes peninsula in western Iceland.122 It is stated that he originally was called Hrólfr and was a chieftain of considerable standing (hǫfðingi mikill). Since he was “a close friend of Þórr” (mikill vinr Þórs), and watched over a Þórr’s hof (varðveitti . . . Þórshof) in Mostur Island, the people called him Þórólfr. Once Þórólfr protected an outlaw and thereby came into a conflict with King Haraldr. Therefore he held a great sacrificial feast and asked his friend Þórr for

118 On this saga, see Perkins 1994. 119 Einar Ól. Sveinsson divides the manuscripts into three different groups: Vatnshyrnuflokkur (V), B-flokkur (B) and Melabókarflokkur. See Eyrbyggja saga, Ísl. Fornr. 4, lvii–lxiv. See also Perkins 1994. 120 Hermann Pálsson & Paul Edward 1989, 13. 121 Hollander in Perkins 1994, 56. 122 Eyrbyggja saga, Ísl. Fornr. 4, 6–10. See also Landnámabók S85, H73, M26, Ísl. Fornr. 1, 124–126. Religious Strategies For Political Authorization—a Case Study 87 advice about what he should do, whether to make peace with the king or leave the country. He was advised to go to Iceland. Þórólfr had a figure of Þórr carved on one of his high-seat posts. When he arrived in Iceland he threw these posts overboard and declared that he would settle at any place where Þórr chose to send the posts. The posts reached land at a ness in Breiðafjǫrðr. Þórólfr called that place Þórsnes, and a river close to that place was called Þórsá and the whole area was consecrated with a fire ritual. Not far from there he established his farm Hofstaðir, where he also erected a hof, and placed figures of gods and other sacred objects in it (see further chapter 4 below). In his old age Þórólfr married a woman called Unn. They had a son called Steinn. Þórólfr dedicated the boy to his friend Þórr and gave him the name Þorsteinn. That boy was called Cod-Biter (þorskabítr), and he inherited his father’s farm. He married a woman called Þóra. She gave birth to a son, who was sprinkled with water and given the name Grímr. “Þorsteinn dedicated this boy to Þórr, calling him Þorgrímr, and said he should become a hofgoði.”123 Þorgrímr’s son was Snorri the goði. He too was a cultic leader and cared for a hof located at Helgafell. Eyrbyggja saga and Landnámabók also mention another son of Þórólfr, i.e., Hallsteinn Þorskafjarðingagoði.124 In the Landnámabók version Hallsteinn’s close relation to Þórr is also emphasized.125 When he settled in Þorskafjǫrðr he sacrificed to Þórr in order to get his high-seat posts. Soon a huge tree was washed ashore at his site. This tree was used to make high-seat posts. It thus seems as if several male members of the Þórólfr family were goðar and devoted to the cult of Þórr. Evidence in Eyrbyggja saga indicates that Þórólfr himself was regarded as a goði, and it seems as if his office was inherited by his offspring (see ch. 5).

3.3.1 A Very Good Friend of Þórr The religious ruler strategy of the Þórsnesingar is not associated with a divine descent of the family. The chieftains in this family had another type of relationship to the god. Eyrbyggja saga says that Þórólfr was “a very good friend of Þórr” (mikill vinr Þórs), and that he regarded the god as his dearest friend (ástvinr).126 The Sagas of Icelanders report that other goðar and magnates in Iceland had this close relation to a chosen god. Hrafnkels saga Freysgoða mentions that the goði Hrafnkell had Freyr as his friend (vinr). “Hrafnkell loved no other god

123 . . . þann svein gaf Þorsteinn Þór ok kvað vera skyldu hofgoða ok kallar hann Þorgrím. Ísl. Fornr. 4, 19. 124 Eyrbyggja saga, Ísl. Fornr. 4, 136. See also Landnámabók S85, Ísl. Fornr. 1, 126. 125 Landnámabók S85, H73, M26, Ísl. Fornr. 1, 124–126. 126 The naming custom in Eyrbyggja saga is most likely fictitious. See Vikstrand 2009. 88 chapter 3 more than Freyr”.127 The chieftain Þorkell the High from Þverá made a prayer to Freyr in Víga-Glúms saga ch. 9: “Freyr, who for a long time has been my confidant (fulltrúi) and received many gifts from me and well rewarded them . . .”128 It has been argued that the linguistic usage in these texts was based on biblical expressions and that the saga authors had the Christian relationship between man and God as a model.129 It has also been suggested that, for instance, the concept of fulltrúi,130 which sometimes expresses an individual devotion to a pagan deity in Old Norse prose,131 is based on Christian ideas.132 This position has however been contradicted. Other scholars have felt that even if these terms were first applied to pre-Christian conditions by authors of the twelfth and thirteenth centuries, it is likely that memories of personal devotion to the old deities were passed on by oral tradition into later centuries.133 Some scholars argue that a personal relationship between man and god in Iceland is evidenced in the tenth-century skaldic poetry.134 In the Icelandic skald and chieftain, Egill Skalla-Grímssonar’s Sonatortek, for instance, we read thus:

Átta ek gott við geirs dróttinn, gerðumk tryggr at trúa hánum, áðr vinátt vagna rúni, sigrhǫfundr, um sleit við mik.135

127 Hrafnkell elskaði eigi annat goð meir en Frey, . . . Hrafnkels saga Freysgoða, (ed.) Jón Helgason, 2. Cf. Ísl. Fornr. 11, 99. 128 Freyr, sagði hann, er lengi hefir fulltrúi minn verit, ok margar gjafar at mér þegit ok vel launat, . . . Víga-Glúms saga 9, Ísl. Fornr. 9, 34. 129 See e.g. Maier 2003, 36. He refers to Exodus 33,11; John 11,11; James 2,23. 130 ON fulltrúi m. ‘confidant, true or completely trustworthy friend’. 131 See e.g. Eiríks saga rauða 8, Ísl. Fornr. 4, 224. 132 See e.g. Zernack 1998. 133 Hultgård 2008, 213. 134 See e.g. Å. V. Ström 1975, 198f.; 1990, 374f. On “friendship” between man and god in pre- Christian Scandinavia, see also Jón Viðar Sigurðsson 2010, 122ff. 135 “I was on good term with the lord of the spear [Óðinn], I grew trustful, believing in him [Óðinn], until the friend of chariots [?], the prince of victory [Óðinn], broke friendship with me.” Text and trans. E.O.G. Turville-Petre 1976. Cf. Skj. B1, p. 37. See also Skj. A1, p. 43. Religious Strategies For Political Authorization—a Case Study 89

Whether this stanza really indicates that Egill previously had a close relation to “the lord of the spear [Óðinn?]” is, however, somewhat uncertain.136 In any case, I agree with those scholars who accept that a personal devotion to a pagan god may have existed in more ancient time. For there are several instances where the phrase trúa á einum “believe in someone or something” seem to indicate a personal relation between man and pagans gods.137 Some of them appear in Eddic lays, which probably were composed in Iceland. In Harbarðzljóð 34, for instance, Hárbarðr states: “Then I should believe in you [Þórr]” (Ec mynda þér þá trúa). According to Åke V. Ström this phrase confirms a personal piety in Norse religion: “here we really have ‘une foi personelle’ in pagan time”.138 In Hyndlolióð 10 we read:

Hǫrg hann mér gerði, hlaðinn steinom, nú er griót þat at gleri orðit; rauð hann í nýio nauta blóði, æ trúði Óttarr á ásynior.139

The expression “Óttarr has always trusted (or believed) in Asynior” may be an ancient expression even if the poem is quite young. At least the cultic context of this expression indicates a pagan setting. Anyhow, it seems as if Óttarr was a devoted worshipper of the goddesses called Asynior (ON ásynior pl.). Thus it is not impossible that the idea that the Þórsnesingar in western Iceland considered themselves as close friends of Þórr may be based on old oral traditions. The close personal relationship to a god may even have legitimated the Þórsnesingar’s central position in public cult. It seems as if some members of this family were goðar, who protected the hof and organized the common cult and sacrificial feasts.140 This type of strategy is quite different from the one found among the Ynglingar and the Lade Earls. Relating to the tentative taxonomy presented in ch. 1 above, the West Icelandic chieftains may be regarded as the most important mediator in the relationship between

136 See e.g. Zernack 1998, 244. 137 See e.g. Å. V. Ström 1975, 198f.; 1990, 374f. Helge Ljungberg (1947) interpreted the linguistic construction trúa á einum as influenced by Christian thinking. 138 Å. V. Ström 1990, 375. 139 “He’s made an altar for me, faced with stone,/ now that stone has turned to glass;/ he’s reddened the new altar with ox blood,/Óttarr has always trusted in Asynior.” Trans. C. Larrington. 140 Even the concept of goði (Goth. gudja, OHG *goto ‘ein Mann der zu einem Gott gehört’) may indicate such personal devotion to a particular deity. De Vries 1956–57, §277. 90 chapter 3 the human and the mythical worlds, i.e., they perform as the (highest) cultic leaders. They are in some sense symbolically close to the divine world, although they are not regarded as divinities or praised for having a divine origin and the like. Instead they are regarded as a “close friends” of the god, with a cultic relationship, and not a genealogical one.141

3.4 A Contextual Interpretation

It thus seems as if the religious ruler strategy of the Icelandic chieftain family was different from the ones used by the Ynglingar and the “Lade Earls”. The kings and earls of Svetjud and Norway seem to have applied a more “bombastic” or “monumental” religious-ideological strategy than the chieftains in Iceland. They argued that their families were of divine descent, while the Þórsnesingar just stated that they were “friends” of the god. The reasons for these differences may be related to various aspects, such as the heterogeneous source materials, but also other contextual matters. In what follows I will consider some possible explanations. The families investigated in this study represent three different types of political structures or systems, namely kingdom, earldom and chieftaincy. They also represent different geographical areas. The Ynglingar may be connected to the Mälaren region and Vestfold, the Lade Earls to Trøndelag, and the Þórsnesingar to the area around Snæfellsnes, at Breiðafjǫrðr in western Iceland. In Svetjud and Norway we may see a concentration of power and some tendencies towards a more centralized political system during the Late Iron Age, with the appearance of some great royal and noble dynasties.142 In Iceland power

141 See Jón Viðar Sigurðsson 1999, 187. It should be noted that goðar also existed in Norway, Svetjud and Denmark and thus this ruler strategy may have existed there too. See e.g. Sundqvist 2007, 30–34. In late sources Earl Hákon, for instance, was also regarded as Þorgerðr Hǫlgabúðr’s best trusted friend (fulltrúi), see Þáttr Þorleifs jarlaskálds in Flateyjarbók I, 213f. But the genealogical relation to the gods seems to be exclusive for certain noble families who mainly resided in Svetjud, Norway and Denmark. Occasionally Icelanders claimed in euhemeristic accounts that they were related to these families. See e.g. Íslendingabók Ísl. Fornr. 1, 27f. This, however, is not a common strategy among Icelandic goðar, as we see from Landnámabók, Íslendingasögur and Íslendingaþættir. Jón Viðar Sigurðsson (2011, 83) writes: “Ari is the only Icelandic chieftain claiming such noble origin. Why he created this pedigree is still an unsolved puzzle, but for Ari, as well as for the Oddaverjar, it was probably important to stress that the Breiðfirðingar were of royal blood.” 142 See e.g. Lindkvist 2008; Krag 2008. Religious Strategies For Political Authorization—a Case Study 91 was split among many hands during “the settlement period” and “the saga age” (c. 870–1050), and there was a large number of Icelandic chieftains.143 It was not until the second half of the Commonwealth period (c. 1050 to 1262/64) that the number of chieftaincies was reduced and we may see a concentration of power.144 The early Icelandic chieftains often traced their genealogies back to a Norwegian hersir, i.e., a type of local chieftain who belonged to the lower elite.145 It has often been argued that such a chieftain was regarded as primus inter pares. He had to secure his own positions by creating alliances and economic wealth, so he could support a small group of men at his farm and a larger group in times of conflict. Such alliances were created by friendship and mutual giving of gifts. It was probably not possible to establish a “political power” in Iceland comparable to the one held by the great dynasties in Norway, Svetjud and Denmark. Iceland had only a thinly spread population and there were no members of royal and noble families present to make such claims. Perhaps there were also strong feelings of independence and pride among those settlers who had emigrated from Norway. Kings and earls were regarded by them as, at best, “a necessary evil”.146 During the Early Viking Age, Norway consisted of over twenty great and perhaps even a hundred minor chieftaincies.147 At the end of the ninth century the number of chieftaincies was reduced. The kingship of Haraldr included the breakthrough of a new type of political structure. It seems as if Haraldr and his sons had ambitions to control all Norway.148 As a reaction to this ambition, some of the minor Norwegian chieftains moved to Iceland, where they gradually established the system of chieftaincies and the new constitution. It is probable that the pre-Christian kingdom in Svetjud also had a rather highly structured political organization.149 It is uncertain, however, when the Svear first constituted a political unit. Early society in Svetjud seems to have been divided into several groups under independent chieftains. A concentration of

143 Jón Viðar Sigurðsson (1999; 2008a, 63f.) has argued that there were a large number of chieftains in Iceland, possibly between 50 and 60 during the “saga age”, and new chieftaincies were created continuously. Cf. Byock 2001, 3f., 13f., 63–80. 144 Jón Viðar Sigurðsson 1999, 39–83; 2008a; 2008b. 145 Cf. Phillpotts 1912–13, 276ff.; Vésteinn Ólason 1998, 29f. 146 Vésteinn Ólason 1998, 27–30. 147 Jón Viðar Sigurðsson 2008a, 22ff. 148 It should be mentioned, however, that King Haraldr’s kin only controlled Vestlandet c. 900. Danish kings had the power over Østlandet and the Earls of Lade controlled Trøndelag and northern Norway. See e.g. Krag 2008 and Jón Viðar Sigurðsson 2008a, 28. 149 Lindkvist 2008, 669. 92 chapter 3 power in this area may have started as early as the sixth or seventh century, when the three royal mounds were erected at Uppsala.150 It is likely that, at times, the chieftains and groups were united in a loose alliance under a federation leader or High King.151 Such a king or overlord could occasionally exercise power over several groups and later also over territory. As time passed, this alliance of groups was stabilized and kingship developed in the Mälaren region, with a royal and cultic site in Uppsala.152 Thus, the political structures in Svetjud and Norway resembled each other in the Viking Age, while Iceland was different. In my opinion, there is a link between the political and historical conditions in Norway, Svetjud and Iceland and the formation of religious strategies for rulership in these three areas. The kings of Norway and Svetjud as well as the Earls of Lade had certain characteristics which few others possessed in these societies.153 By means of religious symbols, descent and wealth, as well as monumental buildings and structures, these royal and noble lines raised themselves over other influential families. They claimed, for instance, that their families were of divine descent. Perhaps such a strategy in a previous historical phase was related to most of the chieftains in Norway and Svetjud, but when the powerful “kingdoms” and “earldoms” emerged, such characteristics were more exclusively used by kings, earls and the very top elite.154 It is also possible that the notion of the rulers’ divine descent only was applied by some of the greatest dynasties in Late Iron Age Scandinavia, such as the Skjǫldungar (of Denmark), the Ynglingar, the Earls of Lade and perhaps the Earls of Orkney.155 Since no kings or earls from these dynasties played

150 Ljungkvist 2006. 151 See Sundqvist 2002, 69. On the concept of kunungR see ch. 2. 152 See e.g. Rahmqvist 1986; 2000; Sundqvist 2002; Ljungkvist 2006; Göthberg 2008 and ch. 2 above. 153 Cf. Jón Viðar Sigurðsson 2008a, 29f. 154 Cf. Jón Viðar Sigurðsson 2008a, 30. In Óláfs saga Helga 94 Snorri states about King Óláfr’s daughter Ingigerðr of Svetjud: “She is of royal birth in all her kin and of Uppsvía lineage, which is the noblest of all in Norðrlǫnd, for it is descended from the very gods” (er komin frá goðunum sjálfum). Heimskringla II, Ísl. Fornr. 27, 152. 155 According to Skjǫldunga saga (Danasaga Arngríms Lærða, ch. 1) the royal family Skjǫldungar of Denmark (Daniæ) were called sons of Óðinn. See Ísl. Fornr. 35, 3f. The god’s son Skjǫldr (Scioldus) ruled over Danes, while his other son, Yngvi (Ingo), was the ruler over Sveciæ. . . . adeptum esse imperium Daniamqve (qvæ tum tamen eo nomine caruerit) Scioldo, Sveciam Ingoni filiis assignasse. Atqve inde à Scioldo, qvos hodie Danos, olim Skiolldunga fuisse appellatos, ut et Svecos ab Ingone Inglinga. The Earls of Orkney were believed to descend from the mythical being (or giant) called Fornjótr and his descendants Religious Strategies For Political Authorization—a Case Study 93 a central role in early Icelandic society, this strategy rarely appeared there.156 Perhaps this religio-political strategy was even discouraged in the early Commonwealth society of Iceland, since it was associated with oppressive rulers who dominated in the old homeland, the Norwegian kingdom. Perhaps also “the Icelandic constitution” discouraged political centralization. Most likely the historical and socio-political contexts affected the formation of the religious strategies for gaining political authority in the three areas. Probably other contextual aspects were involved too, but the suggestions presented above may at least have played a certain role.

In the present chapter a case study has been presented where the specific religious strategies for rulership were investigated in relation to some noble kindreds as narrated in written sources. It included the royal family named Ynglingar from the Mälaren region, the Earls of Lade from Trøndelag, and the chieftain family named Þórsnesingar from Iceland, i.e., representatives from the three regions addressed in the monograph. The kings and earls from the Mälaren region and Trøndelag claimed to have a divine origin. This strategy was not seen among the chieftains in Iceland. They argued, however, that they were the deity Þórr’s dearest friend. The relationship between rulers and mythical beings in the former areas could be conceived as “genealogical”, while the latter (found in Iceland) was “cultic”. The conclusion was that the centralizing tendencies which were observed in the societies of the Mälaren region and Trøndelag had generated a type of ideological strategy, which included the monumentalization of power and rulership and the notion of divine descent. Such a strategy was not as visible in sources referring to the more decentralized Icelandic chieftain society.

Kári, Frosti, and Snær. See Orkneyinga saga, Ísl. Fornr. 34, 3. Cf. Steinsland (2011b, 48ff.) who emphasizes the hieros gamos motif at the top of this genealogy. 156 Vésteinn Ólason (1998, 29f.) has made a similar interpretation. Gro Steinsland (2011a, 6) has recently argued: “Kinship was probably more important for the elite than for other classes of population—note that for example Germanic titles of rulers are derived from kin-groups. The importance of kin for the elite can also be seen in the many genealogies: on the level of kings and earls ancestors were counted back to the gods and other non- human powers. By contrast, chieftains do not claim descent from mythical forces; they were linked to the gods in different ways, for example via initiation.” As noted above, Ari traced his family back to the Ynglingar and a divine descent in his euhemeristic account. But this is an exception in the Icelandic context. Óttarr in Hyndlolióð is descended from the gods (er frá goðom qvómo) although he was born into a hersir family (hersborinn). On the other hand, the poem states that he also was a relative of the royal Skjǫldungar, Ynglingar and Skilfingar. 94 chapter 3

This conclusion constitutes the fundamental hypothesis of the entire monograph. The hypothesis of the present monograph is that the political- structural differences in the three regions (including some other historical- contextual variations) must have had consequences for the ruler strategies associated with cultic places and ceremonial buildings. Strategies in Sweden and Norway (among earls and kings) would therefore be different from those found in Iceland (among chieftains). Before we test this hypothesis in Parts 3 and 4 of this study, I will present and discuss the sources in general terms and previous research on ceremonial buildings. chapter 4 Ceremonial Buildings and Sanctuaries

Both written and archaeological sources, as well as place-names, provide information about ceremonial buildings and sanctuaries in the three investigation areas, i.e., the Mälaren Valley, Trøndelag and Iceland. In this chapter these sources will be critically evaluated and discussed. The presentation is regional, and it will be noted whether different types of ceremonial buildings seem to have existed in each region, such as banqueting halls and more specific cultic houses. Some general contexts of these buildings will be noted, and some comparisons between the three regions will also be made. Great ceremonial buidlings can be detected in all the investigation areas, but their monumentalization is much more emphazised in Svetjud and Trøndelag. These features harmonize well with the ruler strategies discussed in the last chapter (ch. 3). Before we go into the question of evidence, sources, and appearance, I will survey the research history of pre-Christian ceremonial buildings and sanctuaries, as well as the problems of terminology.

4.1 The State of Research

4.1.1 The Existence of pre-Christian “Cultic Buildings” The question whether the ancient Scandinavians had specific cultic buildings or not has occupied scholars for more than 150 years. As early as 1835, Jacob Grimm argued, with support from classical sources, that pre-Christian cultic actions among the Germanic people were mostly performed outdoors, at sacred groves.1 Eventually, other scholars stated that “temples” were also visible in the Old Norse sources and place-names referring to pre-Christian Scandinavia.2 According to these sources, the ancient Scandinavians had buildings called hof, hǫrgr, goðahús and blóthús. Images and sacred objects were kept in these houses, and it was also there that the Scandinavians made sacrifices to their gods and celebrated their religious feasts. Following an excavation at Hofstaðir, Mývatnssveit, in northern Iceland in 1908, it was confirmed that the descriptions of the “temples” in the Sagas of Icelanders were reliable, and could be

1 J. Grimm 1835. 2 E.g., Keyser 1847; Thümmel 1909.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_005 96 chapter 4 supported with archaeological evidence.3 For almost sixty years there was close consensus among scholars that the ancient Scandinavians had “temples” for their cultic activities.4 However, Olaf Olsen’s source-critical investigation, Hørg, hov og kirke (1966), constituted a radical reconsideration. Olsen argued that the descriptions of the “temples” (hof sg.) in the Old Norse traditions could not be supported by the archaeological evidence discovered at so-called hof-lands (hovtomter) in Iceland. He believed that the written descriptions were projections of the medieval authors’ and saga scribes’ experiences of Christian churches. Olsen proposed that the “temples” excavated in Iceland, such as the one at Hofstaðir, were in fact the dwelling-houses or halls of the chieftains, and as such multifunctional: ceremonial feasts would probably have taken place there, but also many non-religious activities. In other parts of Scandinavia too, cultic feasts were celebrated in the multifunctional halls situated on the magnates’ farms. According to Olsen, we have no reliable evidence of specific pre- Christian “temples” in ancient Scandinavia, except in (Old) Uppsala. The pre-Christian cultic activities were in general performed outdoors, at “natural sanctuaries” (naturhelligdomme), such as trees, wells, stones and cairns. Occasionally a type of simple building was erected at such sites, called ON hǫrgr. Olsen argued that these structures may have been some kind of precursors to the earliest Norwegian stave-churches. He also concluded his thesis with the statement that we have no evidence for cultic place continuity in ancient Scandinavia, i.e., that the earliest churches were situated at pre- Christian cultic sites: “vi . . . må hævde, at teorien om, at de ældste kirker fortrinsvis er pleceret på hedenske kultsteder, ikke hviler på et forsvarligt grundlag.”5 According to him, (Old) Uppsala cathedral was the only certain example in Scandinavia where we have such cultic place continuity.6 Olsen’s conclusions have been contradicted by several scholars. It has been argued, for instance that Olsen generalized too much when he stated that cultic place continuity did not exist in ancient Scandinavia.7 His results regarding this specific issue were actually only based on materials from Denmark, i.e., sources which need not be valid for other areas in Scandinavia. Both in the

3 Daniel Bruun & Finnur Jónsson 1909. 4 E.g., M. Olsen 1926; Ohlmarks 1936; Gehl 1941; de Vries 1956–57. 5 O. Olsen 1966, 275. 6 O. Olsen 1966, 236ff. Sune Lindqvist’s theory that the post-holes under the present church in Old Uppsala reflect Adam of Bremen’s “temple” is today considered obsolete. See Nordahl 1996. 7 See e.g., Sandnes 1987, 145. Ceremonial Buildings And Sanctuaries 97

Mälaren area and in Trøndelag there is clear evidence that early churches occasionally were erected at pre-Christian cultic sites.8 It has also been stated that Olsen was inconsistent in his source criticism, for instance when he rejected the indications of hof as a cultic building in the Eddic and skaldic poetry, but accepted hǫrgr as a pagan structure basing his statement on equivalent source materials.9 Despite this criticism, Olaf Olsen’s dissertation exerted a strong influence on research in the following decades, also amongst archaeologists discussing Late Iron Age halls. The archaeologist Frands Herschend, for instance, argued that the halls could be interpreted as multifunctional buildings attached to nobility and royalty.10 According to him, the hall was the room for leadership in an economic as well as a military sense already in the fifth century. The finds discovered in them, for instance gold foil figures (Swedish guldgubbar) and ceremonial glass, indicate that also cultic acts and rituals were occasionally performed there. Such halls had been erected at aristocratic central places of varying local, regional and trans-regional significance, such as Dejbjerg, Gudme and Lejre in Denmark, Uppsala, Vallhagar (Gotland) and Slöinge in Sweden and Huseby (Tjølling) and Borg (Lofoten) in Norway (fig. 2a and 2b). Historians of religions were also inspired by Olsen’s study, arguing that the hof buildings mentioned in the Kings’ Sagas and Sagas of the Icelanders should be interpreted as multifunctional halls.11 The most recent scholarship, however, indicates that the debate concerning specific cultic buildings must be resumed. New archaeological finds suggest that more exclusive cultic houses may have existed in Late Iron Age Scandinavia. At several excavation sites in Denmark and Sweden, traces of

8 See e.g., Lidén 1969; Sandnes 1987; A.-S. Gräslund 1992; Jørgensen 2009. See also the self- criticism by O. Olsen 1996: “Jeg ville være mere åben over for muligheden af, at man på centrale hedenske kultpladser har rejst kirker som udtryk for ecclesia triumphans . . . Jeg ville have brugt ordet ‘magtkontinuitet’ om stormændenes kirkebyggeri på deres gårde. Her er det magten og ikke kulten, der er det centrale element.” 9 See Sandnes 1964; 1987, 145, who also refers to Rostvik’s (1967) critical investigation of hǫrgr. 10 Herschend (1993: 182; 1998: 16) has pointed out certain features connected to these buildings: they belong to big farms; they consist of one room with a minimum of posts; they are singled out by their position on the farm; their hearths are not used for cooking, nor do they facilitate handicrafts; artefacts found in these houses are different from those found in the dwelling part of the main house on the farm. 11 See e.g., Hultgård 1996. Terry Gunnell (2001) has argued that the hall was temporarily transformed into a sacred building at particular times of year, for example during calendrical feasts. See further ch. 8 and 9 below. 98 chapter 4

Figure 2a The great Merovingian period hall in Old Uppsala, c. 50 metres at length. Ljungkvist & Frölund 2015 MS.

Figure 2b Southern plateau in Old Uppsala, where the hall was erected. Photo: Anders Andrén. Ceremonial Buildings And Sanctuaries 99 smaller buildings have been discovered, occasionally adjacent to huge halls. The context of finds testifies that these buildings had ritual functions. Of specific importance is the small house (5.5 × 7 m) on the Viking Age chieftain farm at Borg, in Östergötland, Sweden (see fig. 3). This house appeared in a context of finds that make a cultic interpretation most plausible.12 The building had stood in a stone-paved courtyard, where approximately 75 kg of unburned bones were discovered at a rock. These bones have been interpreted as the remains of animal sacrifices. Beside the rock was a deposit containing almost hundred amulet rings. At the central place of Uppåkra, outside Lund, Scania, a possible cultic house (13.5 × 6 m) was explored (see fig. 4).13 This house had

Figure 3 The cultic house at the Viking Age chieftain farm of Borg, in Östergötland, Sweden. The building was erected on a stone-paved courtyard where some 75 kg unburned bones were discovered in association with a rock. Beside the rock there was a deposit of almost 100 amulet rings. Illustration: Richard Holmberg.

12 Nielsen 1996; 2006. 13 L. Larsson & Lenntorp 2004, L. Larsson 2006a; 2006b; 2011. 100 chapter 4

Figure 4 A cultic house at the central place of Uppåkra, outside Lund, Scania, a possible cultic house (13.5 × 6 m) was explored. This house was founded during the Roman Iron Age, and was rebuilt in several phases up until the Viking Period. The finds indicate that this house may have been a more exclusive and specific cultic building. Illustration: Loic Lecareux. first been built during the Roman Iron Age, and was rebuilt in several phases up until the Viking Period. The peculiar finds, for instance the gold foil figures, a glass bowl and a gilded silver beaker, indicate that this house may have been a more exclusive and specific cultic building.

4.1.2 Large Residences and Small Cultic Houses: Some Views in Recent Research In a very important article, the Danish archaeologist Lars Jørgensen argued that a certain “pattern in terms of the organization of the central, prestigious area on the one hand, and of the closest hinterland on the other hand” could be detected at aristocratic residences (central places), and classical sites in southern Scandinavia, such as Gudme, Uppåkra, Sorte Muld, Tissø, Lejre, Toftegård and Järrestad.14 At all these sites the main building (the actual

14 Jørgensen 2009, 349ff. Cf. Christensen 1991, 1997; Söderberg 2005 Ceremonial Buildings And Sanctuaries 101 residence) occurs in a complex with a smaller building situated beside the former building. In certain cases (e.g., Tissø, Toftegård and Järrestad, which all belong to a later generation of central places), the smaller building is surrounded by fencing which is often built together with the main building. According to Jørgensen, there seems to be a functional difference between the large main building and the smaller hall building. In the smaller building a number of sacral activities took place, while more profane activities such as receptions and feasts took place in the large residence building. There are, however, some indications that in the course of the sixth and seventh centuries several functions were moved from the smaller hall building to the large main building.15 In the large hall of Tissø, for instance, the space was divided into one convivial area and one private area during the seventh to eleventh centuries, while sacrifices were still performed at the fenced-in cultic house. In addition to these buildings there may also be other houses and structures at aristocratic residences which may have been related to cultic activities. Open- air sacrificial places may also occur at such complexes (see below).16 Frands Herschend has discussed the general development of the Iron Age hall building from another perspective. He argues that the embryonic halls during the Roman Iron Age were rather small freestanding buildings like, for instance, house 16 at Vallhagar, Gotland, and house VII at Rönnerum, Öland.17 This type of building was the hall owner’s arena and an interface between public and private spheres. Some rituals and ceremonies related to the local leaders were performed in it and it thus played a significant role for creating authority in society. During the Migration Period the feasting hall (which Herschend calls triclinium) was developed into a room inside a large residential building with several rooms, as at Borg, Lofoten. The hall now became a part of the permanent living quarters of the ruling family. Later, during the Late Viking Age, some feasting rooms were moved into freestanding buildings again. This can be seen at e.g., Tissø, Järrestad and perhaps Lockarp in southern Scandinavia. These small buildings had subtracted the original function of the hall; it is no longer the hall owner’s arena and interface between public and private spheres. It is given “to the deity and its servants”. These buildings could therefore be regarded as a kind of “temples”, since the sacred dimension of the hall has been isolated. Thus we may see a stricter separation of profane and sacred rooms during this phase.

15 Jørgensen 2009. 16 On Tissø, see Jørgensen 1998; 2002; 2009. 17 Herschend 1993; 1998; 1999; 2009, 251–260, 369–377. 102 chapter 4

Herschend’s interpretation has encountered some resistance from initiated archaeologists. Lars Larsson, for instance, has stated that the development outlined by Herschend may fit in a general sense, but there are many exceptions to this pattern. “The development identified by Herschend, from a special hall building to a residential building, did not take place at Uppåkra. The building was rebuilt several times, but not extended.”18 During seven centuries (c. 200– 900) the cultic house at Uppåkra thus retained the same fundamental layout and did not undergo any dramatic changes, even though the society changed. Lars Larsson concludes that houses linked to ritual or ceremonial activities differ considerably in Middle and Late Iron Age Scandinavia. “Some, like the building at Uppåkra, are extremely solid while others, as at Sanda [Uppland], are described as lightly built. Except for Tissø, and Lejre, it seems difficult to identify similarities relating to common traditions and rules.”19 It is thus hard to discern general patterns for the ceremonial buildings. In my opinion, we must stick to the fact that two types of ceremonial buildings may be discerned in Late Iron Age society. First we have the multifunctional and aristocratic halls erected at central places such as Slöinge, Uppsala, Huseby in Tjølling, and Borg in Lofoten. In addition to these buildings there were probably also smaller cultic houses, such as the ones found in Borg, Östergötland and Uppåkra. These smaller houses most likely had a more specific religious function, and were occasionally erected just beside the large Iron Age hall buildings, as seen in Gudme, Lejre, Tissø and Järrestad.

4.1.3 Outdoor Cultic Sites The cultic buildings and halls were erected at places where there were also outdoor ritual structures. The small cultic buildings were sometimes surrounded by fencing which often was built together with the main building. These fenced enclosures must have constituted an “open-air” ritual space. At the cultic buildings and halls there could be different types of layout and outdoor ritual structures, including different types of graves and burial mounds, altars, and erected posts (axis mundi symbols) etc. Some ritual places could have been constituted by natural locales and elements, such as groves, trees, mounds, hills, lakes, bogs, etc.20 At some of these cultic sites there were no buildings. Lars Jørgensen has presented some “open-air cultic sites” in Denmark from the Middle and Late Iron Age.21 On a hillside at Lærkefryd, near

18 L. Larsson 2011. 19 L. Larsson 2011, 201. 20 Vikstrand 2001, 20f.; 2011, 315f. 21 See Jørgensen 2009, 347ff. Ceremonial Buildings And Sanctuaries 103 the village of Jørlunde, in north-eastern Zealand, gold bracteates (fifth century) were found in the nineteenth century. In the bog Rappendam Mose close to Lærkefryd many wagon wheels and parts had also been found, as well as a human skeleton and animal bones from the Late Pre-Roman Iron Age and the Early Roman Iron Age. It was suggested that Lærkefryd succeeded Rappendam Mose as a public cultic site. At the place where the bracteates were discovered, a roughly 30 cm thick black culture layer with small stones was also found. An excavation in 1992 unearthed 47 Roman denarii, 15 pieces of hack gold, five gold finger rings, gilt bronze fibulae and pendants, hack silver in the form of a chopped-up horse harness, a sword hilt, an axe, an Arab dirhem etc. Some of the larger objects showed signs of ritual destruction, for instance the sword hilt, jewellery and harness parts. A majority of the finds were dated to the period from the third to the sixth century. There were also some finds from the Merovingian Period and the Viking Age, but they were sparsely represented. Jørgensen suggested that this was a type of sacrificial site comparable to the open-air offering place on a hill at the Tissø complex. The black clay and some fire pits indicated that ritual meals were celebrated at Lærkefryd. Jørgensen found no traces of buildings, however. A similar type of “open-air cultic sites” was discovered at Smørenge on the island of Bornholm, with finds from the first millennium AD.22

As noted above, previous research has mostly concentrated on the question whether the ancient Scandinavians had specific cultic buildings or not. There seems to be consensus today that religious rituals were performed both indoors, in multifunctional halls as well as specific cultic houses, and at outdoor cultic sites. In the present study the focus will not only be on the character of the ceremonial buildings and cultic sites as such, but also on the religious and ideological dimensions and their wider contexts. The crucial issue is how these ritual arenas—the ceremonial buildings including outdoor cultic places related to such buildings—were used by powerful people and elite groups when constructing authority.23 As far as I know, no such study has been done in previous research from the perspective of history of religions in a more extensive and elaborated way.

22 Jørgensen 2009, 347ff. 23 The “outdoor cultic places” will thus not be focused on as such in this investigation. 104 chapter 4

4.2 The Problem of Terminology

Certain terms in the Old Norse written sources, such as hǫll, salr, veizlustofa, hirðstofa and skáli have sometimes been associated with the Iron Age hall buildings found at excavations in Scandinavia.24 In Old Norse texts it seems as if the term hǫll was mostly applied to royal contexts in Norway, Svetjud and Denmark, while the banqueting buildings at chieftains’ sites in Iceland were usually called stofa or skáli.25 A royal hall (or hall room) in Norway, Svetjud and Denmark could also be described as a veizlustofa.26 In connection with the smaller and more specific cultic buildings, scholars have searched for other Old Norse concepts which may correspond to the archaeological finds. These smaller buildings have sometimes been regarded as equivalents to the hǫrgr appearing in the textual sources, while the fenced and ritual areas around them have been interpreted as a hof.27 Even if these interpretations are not totally unreasonable,28 they have weak support in the ancient Icelandic literature. The idea that hǫrgr designates a house can only be seen in a few (problematic) texts. The Codex Regius manuscript of Snorra Edda, for instance, states that the gods built a hall (salr). This was the hǫrgr that belonged to the goddesses. This building (hús) was called Vingólf.29 However, if we go to the manuscript called Codex Upsaliensis, the hǫrgr in the same passage designates a structure inside the hall, not the building itself.30 The other instance where hǫrgr refers to a house is in Kong Sverrers Christenret. According to this text, it is forbidden to build a house and call it a hǫrgr.31 The interpretation of this reference has also

24 See Brink 1996a; 1999b; Meulengracht Sørensen 2003; Sundqvist 2005c; 2007; 2009a. 25 Cf. Carstens 2012 and 2014. 26 See e.g., Fagrskinna, Ísl. Fornr. 29, 300. 27 See e.g., Jørgensen 1998; 2002; Andrén 2002, 315f.; Söderberg 2005, 109, 195f. In the Old Norse traditions more specific ritual buildings, (“cultic houses”), seem to be designated hof, goðahús and blóthús, while hǫrgr refers to an open-air or outdoor “sanctuary” or “altar”. The Old Norse terms vé and perhaps *al seem to be more comprehensive concepts corresponding to the operational terms “sanctuary” or “shrine”. See e.g., Vikstrand 2001; Sundqvist 2005c; 2007; 2009a. For a critical discussion on *al as ‘sanctuary’, see Brink 1992; Vikstrand 2001, 192; Elmevik 2004; Melefors 2014. On Goth alhs, see Jackson 2014a. 28 See e.g., Rostvik 1967; Vikstrand 2001, 211, 260. 29 Annan sal gerðu þeir, þat var hǫrgr er gyðjurnar áttu, ok var hann allfagr. Þat hús kalla men Vingólf. Gylfaginning (Cod. Reg. Edda). (Ed) Faulkes 1988, 15. 30 Annan sal gerðu þeir er hǫrgr var í . . ., Gylfaginning (Cod. Ups. Edda). (Ed) Heimir Pálsson 2012, 24. 31 ef maðr . . . læðr hauga eða gerer hus oc kallar horgh. Here quoted from Kong Sverrers Christenret 79. NGL 1, 430. Ceremonial Buildings And Sanctuaries 105 been regarded as problematic in previous research.32 In religious contexts, the term hǫrgr is mostly given the meaning ‘outdoor sanctuary’ or ‘stone altar’.33 The term that most frequently refers to ceremonial buildings in written sources is ON hof (OSw hov). Old Norse dictionaries do not give a thorough description of the character of these houses, however. According to Norrøn ordbok, for instance, it means ‘heidnisk gudehus, hov, tempel’.34 Since the meaning of hof seems to be unclear, I have devoted a previous study to a scru- tiny of its semantics in Old Norse texts and other sources.35 This investigation indicated that hof (hov) had varying meanings in the sources.36 In some cases, it designated a large hall building erected on the chieftain’s farm. Ceremonial banquets and sacrificial feasts were occasionally performed and celebrated in these buildings, at least in some parts of them. In other sources hof seems to have designated a more specific and sometimes smaller religious building, located some distance apart from the central place, or chieftain’s farm, where the banqueting hall was situated. Perhaps it was not the locality or size of the building that determined the nomenclature hof. It could just as well have been certain rooms, sacred objects, specific architecture or the presence of certain cultic leaders, or the ritual activities that took place there, which created such

32 See e.g., O. Olsen 1966, 106. 33 See e.g., Vǫluspá 7; Hyndlolióð 10; see also La Farge & Tucker 1992, 133. 34 Heggstad et al. 1993. 35 See Sundqvist 2007; 2009a. 36 In the field of linguistics it has been stated that the word hof originally had the meaning ‘height, hill’ and only later developed the meaning ‘magnate farm’ or ‘ceremonial building’. There are equivalent terms in Old High German, Old Saxon and Old Frisian, such as hof, which means ‘farm, house, court’; Old English hof ‘enclosure, dwelling, building, house, temple, sanctuary’. See e.g., T. Andersson 1986; 1992a; 1992b; 1992c; Brink 1996a; Vikstrand 2001, 253; de Vries 1977, 246; SOL, 131f. Theophoric place-names testify, however, that hof as early as the Middle and Late Iron Age could have had a sacred meaning, for instance Frøyshov (‘Freyr’s hof’), Norderhov (*N(j)ærðarhof) (‘*N(j)ærð’s hof’) and Ullinshov (‘Ullinn’s hof’). Helleland 1996, 136f. Certain Old Norse poetical traditions also support the idea that hof often had a sacral meaning. In the Eddic poems, for instance, it refers to mythical structures where the deities dwelled. See for instance, Vǫluspá. In the earliest Christian skaldic poetry, however, it is apparent that hof refers to a house. See e.g., Austrfararvísur Skj. A1, 233–240; B1, 220–225. In the thirteenth-century prose traditions, hof often designates a pre-Christian huge hall building. These buildings seem to have had religious function. Inside such buildings were kept ritual objects and images of gods. See e.g., Hákonar saga goða 14–18, in Heimskringla I, Ísl. Fornr. 26, 167ff.; Kjalnesinga saga 2, Ísl. Fornr. 14, 7; Eyrbyggja saga, Ísl. Fornr. 4, 8f. and Landnámabók S85, H73, M26, Ísl. Fornr. 1, 124–126. Sometimes hof may have referred to a small and more exclusive cultic house. See e.g., Víga-Glúms saga, Ísl. Fornr. 9, 34. For a more thorough survey, see Sundqvist 2009a. 106 chapter 4 a designation. It is also possible that the meaning of the term varied at different places in Scandinavia and was also altered over time. Scholars who have searched for terms from the ancient Scandinavian languages to describe the archaeological finds discovered at cultic sites have sometimes not taken into consideration the problems the linguistic materials involve. The varying and occasionally unclear meaning of hof makes it hard to connect this term to a specific type of find or structure found at archaeological excavations. Because of these conditions it is very important to formulate operational concepts which may be applied as analytical instruments, when analysing ceremonial buildings and sanctuaries in the present study.

4.2.1 Operational Concepts The most common operational term for cultic buildings in comparative studies of religion and phenomenological treatments is “temple” (from Lat templum (cf. Greek τέμενος) ‘a space marked out’).37 This analytic concept refers to a sanctified and sacred building or house of gods, which is roofed and situated inside a ritual area.38 The architecture of temples and cultic sites often comprises a cosmic symbolism, representing a celestial prototype. The symbolic objects and images related to the temple may thus reflect the mythical topography, mythical scenes or divine beings. Occasionally the temple (or symbols situated in the temple) therefore is regarded as the omphalos of the world, the cosmic navel, the centre or pillar where earth and heaven meet (axis mundi). It can then be conceived as the interface, the door or the threshold to the mythical world, i.e., the place where humans may meet mythical beings.39 The activities in the temple are often connected to religious rituals, such as sacrifices, where the purpose is to communicate with the Other World. These rituals are led by religious specialists and cultic leaders, such as priests and priestesses. By means of inauguration rituals the temple was demarcated from the profane sphere. Therefore ritual restrictions (taboos) must be observed in this ritual

37 A Latin Dictionary gives the following meanings of this world: ‘a space marked out’; hence, in partic., in augury, ‘an open place for observation, marked out by the augur with his staff’. In addition to this meaning the dictionary also gives following: 1) ‘an open, clear, broad space’, 2) ‘a consecrated or sacred place, a sanctuary’, in part., ‘a place dedicated to some particular deity, a fane, temple, shrine’. A Latin Dictionary, (eds.) Lewis & Short 1879. Cf. Georges 1951II, 3050. 38 van der Leeuw 1956, 448ff.; Ringgren & Ström 1984 (1957), 28; Heiler 1961, 131ff.; Widengren 1969, 328ff.; Hinnells 1995, 517; Eliade 1991 (1949), 6ff.; 1987 (1957), 58–65; 1974 (1949), 371ff. 39 See e.g., Eliade 1991 (1949), 12f., 16f.; Widengren 1969, 344. Ceremonial Buildings And Sanctuaries 107 restricted area, and sometimes only the religious specialists have access to this room or space. Ritual objects are also kept there, such as the images, sacrificial tools and ceremonial equipment. These objects may not be used by common people. The synthetic concept “temple” has been applied to many cultures all around the world, but perhaps particularly to sources describing ceremonial buildings in the Mediterranean area. “Solomon’s Temple” in Jerusalem is one example. It was built on a Canaanite pattern, and included a long building which was opened on one of its sides.40 Inside the Temple, the building was divided into three parts; a vestibule, a room for worship and the place usually called the “Holy of Holies”, i.e., the room reserved to Yahweh, and the Ark of Covenant.41 In the second Temple, during post-exile period, this room was only opened to the High-Priest, alone, once a year, on the Day of Atonement.42 According to the Old Testament, the Temple in Jerusalem was modelled on a divine prototype.43 There is evidence that in their early history the Indo-Europeans (including the ancient Scandinavians) did not build temples or make images of their deities.44 Ethnographical observations made by Greeks and Romans support these assumptions. According to Herodotus, Persians sacrificed on the highest mountains and in an open space.45 Strabo describes feasts and sacrifices performed under the open sky, on cliffs facing the sea with caverns underneath.46 Ovid reports that Romans worshipped their gods in consecrated groves: “Under the Aventine there lay a grove black with the shade of holm-oaks; at the sight of it you could say, ‘There is a spirit here.’ ”47 It seems, however, as if the Greek and Roman people, as well as the Persians, gradually started to build temples. Well-known is the temple of Apollo in Delphi. The ὀμφαλός (omphalos) or ‘navel’ situated in the holy area (ἄδυτον or aduton) symbolized the centre of the world.48 In Rome the term Lat templum refers to an inaugurated rectangular area, which had been marked out by an augur ‘sage, diviner’, by means of divine signs, in order to form the sanctuary according to the heavenly

40 Ringgren 1989, 90f.; cf. de Vaux 1980, 312ff.; Smith 1987, 47–73. 41 1 Kings 6; Fritz 2002. 42 Lev. 16:15; Ezek. 9:7. 43 1 Chron. 28:19. 44 For this aniconic tendency, see e.g., Widengren 1969, 339–346. 45 Herodotus I:131–132. 46 E.g., Strabo XI 7,5. See also Widengren 1965, 131, 161. 47 lucus Aventino suberat niger ilicis umbra, quo posses viso dicere, ‘numen inest’. Fasti III, 295f. Trans. Frazer. Widengren 1969, 340; cf. van der Leeuw 1956, 446f. 48 S. R. F. Price 1993, 135f. 108 chapter 4 prototype. The earthly templa were considered as loci effati, i.e., they were relieved of sins. A templum often had a building with a large room (cella). The sacrificial altar (ara) was situated in front of the building. The Roman term harmonizing best with the concept of temple is actually Lat aedes sacra ‘temple, house of god(s)’.49 In the mystery cults too, temples and roofed sanctuaries were used in cultic contexts. In the Mithras cult, for instance, underground sanctuaries (Lat spelaeum, templum, crypta, fanum) were used, with an assembly hall, a central gangway with benches on both sides, and at the very front, round sculptures were situated together with the relief representing Mithras killing the bull (tauroctonus).50 The sanctuary mirrored the cosmic cave created by Mithras in primordial time.51 These Mithraic sanctuaries are attested in the Germanic area.52 There is no doubt that some type of ceremonial buildings also existed in Late Iron Age Scandinavia. Since the analytic concept of “temple” (from Lat templum) was coined in a specific cultural context and arouses strong associations of Greek-Roman Antiquity, and the Classical world (including e.g., Israel, Mesopotamia, Egypt, China, Persia and India), and in addition, the empirical situation still is very unclear in ancient Scandinavia (see above), I will use more neutral and tentative terms in this study when describing such buildings. As noted above, archaeology indicates that two types of ceremonial buildings existed in the pre-Christian context, i.e., (1) the multifunctional and aristocratic banqueting buildings, and (2) the (smaller) more specific cultic houses. In this work I will designate the multifunctional buildings “halls”, while the more specific ritual structures will plainly be called “cultic houses”/“cultic buildings”.53 In the hall building there might have been certain places or areas which were regarded as ritual spaces more permanently, such as the high-seat (hásæti). Otherwise the hall was probably only temporarily transformed into a sacred building at particular times of year, for example during the calendrical feasts.54 The cultic house/cultic building was more permanently regarded as a sacred place.

49 Tomlinson 1987, 386ff.; Beard 2002, 22f., 87ff., 121ff., 196ff., 253ff.; Rüpke 2001, 180ff.; Golzio 2002. 50 Clauss 1990, 51ff.; Turcan 1993, 74ff. 51 Porphyry, De antro nympharum, 6. 52 Vermaseren 1956–1960 II, 47–143; Clauss 1992, 96ff. 53 In my opinion, we need more information from archaeology about these small cultic buildings, before we may describe them as “temples”. 54 Cf. Gunnell 2001. Ceremonial Buildings And Sanctuaries 109

I will sometimes also use a more neutral umbrella concept “ceremonial building”, comprising both terms “hall” and “cultic house”.55 When discussing the entire public and permanent cultic site, including the ceremonial buildings, outdoor ritual structures and places, I will use the term “sanctuary” (or “shrine”). When using this term I refer to a place where the deities and divine powers dwell or manifest themselves, and where human beings recurrently encounter these powers. Such places are public and may be used by groups of varying size, from small to large communities. These terms thus carry two meanings, one cultic and one social. A “sanctuary” is a sacred place, which sometimes is set apart from the space of ordinary existence, dedicated to a deity, used for cultic activities, and it may include a cosmic symbolism.56 There are outdoor sanctuaries without ceremonial buildings. They will not primarily be focused on in this study. Non-public and more “private” sanctuaries will not be in focus in the present work either.57 Instead the searchlight will be on the public ceremonial arenas with buildings, and particular the aspects of ruler strategies related to such ritual spaces. Thus a special focus will be put on those cultic places which were controlled by the individual rulers and chieftains. Some more communal sanctuaries will also be observed.

55 Historians of religions usually make a distinction between rite, ritual and ceremony. Spiro and Rydving regard ritual as “the generic term for any kind of cult behavior, regardless of its degree of elaboration or complexity”, while they understand rite as “the minimum significant unit of ritual behavior”. They describe ceremony as the smallest configuration of rites constituting a meaningful ritual whole” and ceremonial “the total configuration of ceremonies performed during any ritual occasion”. See Spiro 1982, 199; Rydving 2011a, 37. The concept of “ceremonial building” refers to a building where rites, rituals and ceremonies sometimes take place. Cf. Torun Zachrisson & Margareta Kempff Östlind 2006, who use the Swedish term “ceremoniella rum” in similar contexts. 56 For a discussion of these operational concepts, see e.g., Rydving & Kristoffersson 1993; Nordberg 2010; 2011; 2014. See also Per Vikstrand’s (2001, 20–29; 2011, 315f.) general descriptions of sacral places. 57 In the field of archaeology it has been emphasized during the last few decades that the terms “private” and “public” may be problematic in such contexts, since a lot of things that we today relate to the public sphere actually took place during the Viking Age in the banqueting hall of the ruler or chieftain, which was located at his private farm. Anne- Sofie Gräslund (2001) writes thus: “Det är emellertid tveksamt om man bör använda begreppen [privat och offentlig] för vikingatiden med den innebörd de senare kommit att få. Sannolikt var dessa två sfärer mycket mer inflätade i varandra, åtminstone så till vida att mycket av det som skedde på gården, särskilt i stormansgårdarnas hallbyggnader, i hög grad tillhörde det offentliga livet.” I agree with Gräslund’s opinion about these terms, but still I will occasionally use them, even though I am aware of the fact that public sanctuaries in ancient Scandinavia almost always were located on private farms. 110 chapter 4

We will now turn to the evidence of ceremonial buildings and sanctuaries in the three investigation areas. The purpose of this survey is to discuss the sources and their value for historical reconstructions. I will also make some comparisons between the buildings found in Lake Mälaren area, Trøndelag and Iceland and their material contexts. We will start with the “temple” of Uppsala in Svetjud.

4.3 Uppsala and other Sanctuaries in the Mälaren Area

4.3.1 Adam of Bremen and the “Temple” in Uppsala The written sources relevant for pre-Christian ceremonial buildings and sanctuaries in the Mälaren area are few and frail. Fortunately, there is one important text preserved on this topic, namely Adam of Bremen’s description of the “Uppsala temple” appearing in his Gesta Hammaburgensis Ecclesiae Pontificium. This book was written around 1075/1076.58 It is regarded as the main historical source for eleventh-century Svetjud.59 The relationship between the manuscripts of the text is very complicated.60 The oldest manuscript is usually called “A2”, or the “Leiden manuscript” (Cod. Voss. Lat 40 123). It is dated to c. AD 1100. The first three books in this manuscript are imperfect, although Book IV, where the description of the Uppsala cult appears is complete. All notes (scholia) attached to the chapters about the Uppsala cult are present in this manuscript. The relationship between them and the main text is complicated.61 Book IV is an ethno-geographical description of Scandinavia. The successful mission activity in this area of the Hamburg-Bremen diocese is described, but also the missionary work remaining to be done is mentioned. The temple Ubsola (Uppsala) is referred to as the “final hindrance opposing the victory of Christianity” in the land of the Svear.62 The passage about Uppsala is introduced thus in chapter 26:

That folk [the Svear] has a very famous temple called Uppsala (Ubsola), situated not far from the city of Sigtuna (Sictona). In this temple, entirely

58 On Adam and his work, see e.g., Adam of Bremen, Historien om Hamburgstiftet och dess biskopar, 1984. 59 P. H. Sawyer 1991, 16–19. 60 Nyberg 1984, 302–307. 61 Hultgård 1997, 9–15. 62 Hallencreutz 1997. Ceremonial Buildings And Sanctuaries 111

decked out in gold, the people worship the statues of three gods, in such wise that the mightiest of them, Thor (Þórr), occupies a throne in the middle of the room intended for ceremonial banquets. Wodan (Óðinn) and Fricco (Freyr) have places on either side. The significance of these gods is as follows: Thor, they say, presides over the air, which governs the thunder and lightning, the winds and rains, fair weather and crops. The other, Wodan—that is, the Furious—carries on war and imparts to man strength against his enemies. The third is Fricco, who bestows peace and pleasure on mortals. His likeness, too, they fashion with an immense phallus. But Wodan they chisel armed, as our people are wont to represent Mars. Thor with his sceptre apparently resembles Jove. The people also worship heroes made gods, whom they endow with immortality because of their remarkable exploits, as one reads in the Vita of Saint Ansgar they did in the case of King Eric.63

In scholion 138 we get some information about the surroundings of the temple:

Near this temple stands a very large tree with widespreading branches, always green, winter and summer. What kind it is nobody knows. There is also a spring at which the pagans are accustomed to make their sacrifices, and into it to plunge a live man. And if he is not found, the people’s wish will be granted.64

63 XXVI: Nobilissimum illa gens templum habet, quod Ubsola dicitur, non longe positum ab Sictona civitate. In hoc templo, quod totum ex auro paratum est, statuas trium deorum veneratur populus, ita ut potentissimus eorum Thor in medio solium habeat triclinio; hinc et inde locum possident Wodan et Fricco. Quorum significationes eiusmodi sunt ‘Thor’, inquiunt, presidet in aere, qui tonitrus et fulmina, ventos ymbresque, serena et fruges gubernat. Alter Wodan, id est furor, bella gerit hominique ministrat virtutem contra inimicos. Tercius est Fricco, pacem voluptatemque largiens mortalibus. Cuius etiam simulacrum fingunt cum ingenti priapo. Wodanem vero sculpunt armatum, sicut nostri Martem solent; Thor autem cum sceptro Iovem simulare videtur. Colunt et deos ex hominibus factos, quos pro ingentibus factis immortalitate donant, sicut in Vita sancti Ansgarii legitur Hericum regem fecisse. Translation of Adam’s text by Francis J. Tschan 2002, but slightly modified here. 64 Scholion 138: Prope illud templum est arbor maxima late ramos extendens, semper viridis in hieme et aestate; cuius illa generis sit, nemo scit. Ibi etiam est fons, ubi sacrificia paganorum solent exerceri et homo vivus inmergi. Qui dum non invenitur, ratum erit votum populi. This scholion as well as scholia 139–141 are preserved in the oldest manuscript called A 2 (“the Leiden manuscript”). It is dated to c. 1100. See Hultgård 1997, 15. The last sentence in scholion 141 (Hoc sacrificium fit circa aequinoctium vernale) is, however, only attested late in the tradition of manuscripts, more precisely in manuscript A 3 from 1434. Hultgård 112 chapter 4

Scholion 139 presents the temple in more detail:

A golden chain goes round the temple. It hangs over the gable of the building and sends it glitter far off to those who approach, because the shrine stands on level ground with mountains all about it like a theatre.65

In chapter 27 Adam turns to the cultic activities at Uppsala and he describes them in detail:

For all their gods there are appointed priests to offer sacrifices for the people. If plague and famine threaten, a libation is poured to the idol Thor; if war, to Wodan; if marriages are to be celebrated, to Fricco. It is customary also to solemnize in Uppsala, at nine-year intervals, a general feast of all provinces of Svetjud. From attendance at this festival no one is exempted. Kings and people all and singly send their gifts to Uppsala and, what is more distressing than any kind of punishment, those who have already adopted Christianity redeem themselves through these ceremonies. The sacrifice is of this nature: of every living thing that is male, they offer nine heads, with the blood of which it is customary to placate gods of this sort. The bodies they hang in the sacred grove that adjoins the temple. Now this grove is so sacred in the eyes of the heathen that each and every tree in it is believed divine because of the death or putrefaction of the victims. Even dogs and horses hang there with men. A certain Christian informed me he had seen seventy-two miscellaneous bodies suspended there. Furthermore, the incantations customarily chanted in the ritual of a sacrifice of this kind are manifold and unseemly; therefore, it is better to keep silence about them.66

1997, 30. In this study I consider the notes (scholia) 138–141 as parts of Adam’s text, even if it is somewhat uncertain. 65 Scholion 139: Catena aurea templum circumdat pendens supra domus fastigia lateque rutilans advenientibus, eo quod ipsum delubrum in planitie situm montes in circuitu habet positos ad instar theatri. 66 XXVII: Omnibus itaque diis suis attributos habent sacerdotes, qui sacrificia populi offerant. Si pestis et fames imminet, Thor ydolo lybatur, si bellum, Wodani si nuptiae celebrandae sunt, Fricconi. Solet quoque post novem annos communis omnium Sueoniae provintiarum sollempnitas in Ubsola celebrari. Ad quam videlicet sollempnitatem nulli prestatur immunitas. Reges et populi, omnes et singuli sua dona transmittunt ad Ubsolam, et, quod omni pena crudelius est, illi, qui iam induerunt christianitatem, ab illis se redimunt cerimoniis. Sacrificium itaque tale est: ex omni animante, quod masculinum est, novem capita offeruntur, quorum sanguine deos [tales] placari mos est. Corpora autem Ceremonial Buildings And Sanctuaries 113

In scholion 141 Adam mentions how many days the sacrifices were solemnized. He also reports about the nature and number of the victims, as well as when the feast took place:

Feasts and sacrifices of this kind are solemnized for nine days. On each day they offer a man along with other living beings in such number that in the course of the nine days they will have made offerings of seventy-two creatures. This sacrifice takes place about the time of vernal equinox.67

When surveying all Adam’s notices on Uppsala, including the scholia, it seems as if the cultic site there comprised several ritual elements and cultic places. Besides the temple with the three statues of gods and the holy grove where the sacrificial objects were hung, a specific holy tree and a well are mentioned. In the latter place the pagans are accustomed to make their sacrifices. One scholion also says that the shrine stands on level ground with mountains all about it. Whether Adam in this case is referring to the royal mounds in Old Uppsala is debated.

4.3.2 The Criticism of Adam’s Account Scholars have evaluated Adam’s text differently.68 Some decades ago, historians of religions argued that this text was reliable since it was contemporary with the events it described.69 Indeed, Adam built his narrative on second-hand

suspenduntur in lucum, qui proximus est templo. Is enim lucus tam sacer est gentilibus, ut singulae arbores eius ex morte vel tabo immolatorum divinae credantur. Ibi etiam canes et equi pendent cum hominibus, quorum corpora mixtim suspense narravit mihi aliquis christianorum LXXII vidisse. Ceterum neniae, quae in eiusmodi ritu libationis fieri solent, multiplices et inhonestae, ideoque melius reticendae. 67 Scholion 141: Novem diebus commessationes et eiusmodi sacrificia celebrantur. Unaquaque die offerunt hominem unum cum ceteris animalibus, ita ut per IX dies LXXII fiant animalia, quae offeruntur. Hoc sacrificium fit circa aequinoctium vernale. 68 As early as the 1920s Elias Wessén (1924, 131) wrote: “Omdömena om hans [Adam’s] vederhäftighet och tillförlitlighet äro emellertid mycket olika. Den ene forskaren är benägen att ta för gott allting, som kan stödja Adams auktoritet; den andre möter hans framställning med mycket stor misstro.” 69 See e.g., de Vries 1956–57, §290. See also Nils Lid (1942, 86), who stated: “Ein må gå ut frå at dei einskilde ting Adam fortel om, har eit faktisk grunnlag.” Folke Ström (1985, 79) wrote thus: “I en berömd och ofta citerad skildring har Adam av Bremen återgett ett ögonvittnes berättelse om de offer och ceremonier som utspelades i Uppsala vart nionde år”, while Anne Holtsmark (1992, 17) argued: “Detta är en autentisk skildring av gudadyrkan i Norden i slutet av den hedniska tiden.” 114 chapter 4 information, but his informants were eyewitnesses. One of them was the Danish King Svein Estridsson (Sveinn Ástriðarson). He had lived among the Svear for a long period (perhaps around the 1030s and 1040s) and was probably most familiar with their customs, at least at that time. More recent research is more sceptical.70 Scholars now argue that Adam’s description of Uppsala is permeated with rhetorical adornments and missionary strategies. In the book Uppsalakulten och Adam av Bremen this attitude is emphasized. According to the editor of the book, Anders Hultgård, the text is a result of a literary process in which the informants’ reports have been reworked by rhetorical embellishment, ethnographic clichés and subjectivity. There are, for instance, two classical narrative genres present in Adam’s account, called evidentia and ekphrasis (descriptio). Both aimed at sharpening his description. Elements from Adam’s own imagination were added to make the narrative vivid and clear, for instance in the description of the temple with the golden chain. Even if Hultgård’s criticism is sharp, he reaches a fairly positive conclusion regarding Adam’s text.71 It has also been argued that the missionary strategy of Hamburg-Bremen diocese may have affected Adam’s description of Uppsala and the religion of the Svear. Carl Fredrik Hallencreutz, for instance, has shown that Adam applied an Augustinian perspective when describing the pagan cult at Uppsala. This included a historico-theological dualism, in which the civitas dei, i.e., Christianity, was opposed by the paganism of Uppsala (civitas terrena).72 This tendency must be taken into consideration when evaluating the text. The Augustinian perspective adds secondary elements to Adam’s account. Just before describing the cult at Uppsala, he has this to say about Svetjud: “There are Amazons, and Cynocephali, and Cyclops who have one eye on their fore- heads; there are those Solinus calls Himantopodes, who hop on one foot, and those who delight in human flesh as food, and as they are shunned, so may they also rightfully be passed over in silence.”73 These examples show that Adam’s text must be treated with great care. Adam had a specific purpose with his text. He wanted to legitimize the German

70 E.g., Hallencreutz 1997; Hultgård 1997; Göthberg et al., 2010. 71 “Om man skalar av den retoriska apparaten och de polemiska schablonerna, och om man bortser från vissa missuppfattningar och medvetna tillrättalägganden av sagesmännens berättelser, äger vi i Adams notiser om Uppsalakulten en värdefull religionshistorisk källa.” Hultgård 1997. 72 Hallencreutz 1997, 122–124. 73 Ibi sunt Amazones, ibi Cynocephali, ibi Ciclopes, qui unum in fronte habent oculum; ibi sunt hii quos Solinus dicit Ymantopodes, uno pede salientes, et illi, qui humanis carnibus delectantur pro cibo, ideoque sicut fugiuntur, ita etiam iure tacentur. Adam, IV: 25. This quotation is mine and not taken from Hallencreutz. Ceremonial Buildings And Sanctuaries 115 mission in Svetjud. Influences from the Bible are also present in Adam’s text. The description of the gold-decorated temple, for instance, reminds us of the portrayal of Solomon’s temple in the Old Testament.74 There are thus elements in Adam’s text which may be regarded as unreliable. The most radical criticism of Adam’s text has been formulated by the historian Henrik Janson. He states that the passage about the temple of Uppsala is not aimed at pagan rituals, but as an allegorical satire against the Gregorians and the affiliated Church (called gallicana ecclesia), which according to Janson existed at Uppsala in the mid-eleventh century.75 In the middle of the eleventh century the archdiocese of Hamburg-Bremen was in a bitter feud with the pope in Rome, Gregory VII. When Adam wrote about pagani in Sweden he thus did not mean people of the ancient Scandinavian religion, but the members of the affiliated Roman Church. The temple in Uppsala with its triclinium actually referred to the Lateran palace and its triclinium majus. The three gods mentioned by Adam were fictive and functioned as figurations of the sins that the pope and the curia comprised: Wodan (Óðinn) represented fury (furor), Thor (Þórr) pride or arrogance (superbia) while Fricco (Freyr) personalized fornication and pleasure (voluptas).

4.3.3 Adam and the Problem with Christian Runic Stones in Old Uppsala In my opinion, most of the recent source-critical contributions on Adam’s text have been good for the research in this area. That theology, mission strategies and rhetorical aspects affected his description is quite plausible. But some of the criticism has gone too far, for instance Janson’s position. Janson denies that a pre-Christian cult at all existed in Svetjud during the second half of the eleventh century. According to him, there are no sources supporting such a cult in this area at that time. On the other hand, there are plenty of runic inscriptions from this period testifying that Christianity was adopted by the Svear in Uppland, with its centre in Old Uppsala. I agree with Henrik Janson that the erection of runic stones in Uppland during the eleventh century probably reflects the propagation of Christianity in this province. But it is no obvious fact that Uppsala was the centre of these activities, as suggested by Janson.76 As several scholars have noted, there are

74 1 Kings 7,47–50. See Hultgård 1997, 17. 75 Janson 1997; 1998. 76 Janson 1998, 21f., 116f. Runic inscriptions appear mostly in the southern part of Uppland. The northern, eastern and western parts have just a few. Even if the Uppsala area constitutes one nucleus in the distribution of runic inscriptions in Tiundaland, the frequency of inscriptions in Attundaland, e.g., Vallentuna härad, Sollentuna härad and 116 chapter 4 surprisingly few runic inscriptions in Old Uppsala parish compared with the neighbouring parishes.77 In Sveriges runinskrifter nine runic stones (U 978–986) are mentioned from Old Uppsala parish,78 while of the neighbouring parishes Vaksala has eighteen, Ärentuna thirteen and Bälinge as many as thirty-seven runic stones. In Sveriges runinskrifter five of the runic stones are said to be (or have been) located right beside the church and the vicarage: U 978, U 979, U 980, U 981† and U 986†.79 It is exactly in this area that scholars have expected to find the cultic site.80 The well-preserved U 978 is today mortared into the southern apse wall of the church, on the outside. We do not know its original site. According to Peringskiöld it had previously been placed on the altar, as a table, inside the church. The stone is decorated with a cross and is dated to the first half of the eleventh century.81 The stone U 979 lacks runes. It is raised against the western gable, outside the church. According to Peringskiöld it was previously the threshold of the main entrance of the church. It was decorated with a ship which had a Christian cross as a sail. The picture has, however, been worn away.82 U 980 is now preserved only in fragments, three pieces. It was previously mortared into the oven in the vicarage bakery. We do not know its original site. Peringskiöld’s assumption that the stone “warit uprest någonstädes ute på backen wid Kongsgårds Högarna” is according to Wessén and Jansson unprov- able (“ett löst antagande”).83 The stone is depicted in Bautil (1750) standing in front of the mounds, but this was just a guess. In fact the stone was broken in pieces when it was discovered on the ground at the vicarage at the beginning of the eighteenth century. On stylistic grounds it can be dated to the second half of the eleventh century.84 U 981† does not exist today. We are dependent on the information from Johannes Bureus and Johan Rhezelius living in the seventeenth century.

Seminghundra härad, is much higher. See A.-S. Gräslund 1987. Cf. Thompson 1975, 163; R. Palm 1992, 98–100. 77 Cf. Sundqvist 2002, 112ff; Lerche Nielsen 2003; A.-S. Gräslund 2013. 78 Besides the nine stones there is one recently discovered runic stone from Ulva kvarn. See Lerche Nielsen 2003. 79 Wessén & Jansson 1940–58, Vol. 4, 126–144. 80 A.-S. Gräslund 1997, 108. 81 A.-S. Gräslund 1997, 108. Wessén & Jansson 1940–58, Vol. 4, 126–129. U 978 can be dated to c. 1025–1050 (Pr 2). See also A.-S. Gräslund 2013. 82 Wessén & Jansson 1940–58, Vol. 4, 129–132. 83 Wessén & Jansson 1940–58, Vol. 4, 132. 84 A.-S. Gräslund 1997, 108. U 980 can be dated to c. 1075–1100 (Pr 4). A.-S. Gräslund 2013. Ceremonial Buildings And Sanctuaries 117

According to them it was located in the wall of the vicarage bakery.85 Michael Lerche Nielsen has recently argued, on good grounds, that U 981† never existed.86 According to Lerche Nielsen “Bureus inscription [U 981†] is a con- tamination of two inscriptions: U 980 and U 986†.”87 U 986† is now lost, but a picture from the seventeenth century indicates that it originally stood by the brook Samnan, situated on the royal manorial lands.88 It was thus not located at Kungsgården as suggested by Sveriges runinskrifter. None of these runic stones, therefore, are known with certainty to have been raised in the area around the cathedral, the vicarage or the central parts of the royal estate during the eleventh century. They may have been brought to this place when the Romanesque church was erected.89 Old Uppsala thus cannot be viewed as an “epicentre” for the raising of runic stones in Uppland, as Janson does.90 The statement made by the runologist Michael Lerche Nielsen harmonizes better with reality: “Gamla Uppsala är anmärkningsvärt fattigt på runstenar.”91 There are also many other obscurities related to Janson’s hypothesis.92 One fundamental problem with his theory is the question why Adam must criti- cize the pope and Rome in an ulterior way, i.e., in an allegorical form. Adam actually worked among like-minded people, who had the same attitude as he had against Rome. His supreme leader, Archbishop Liemar, and the royalists demonstrably rejected the pope at a synod in Worms 1076.93 Adam actually had nothing to fear when criticizing Gregory and Rome. Janson’s suggestion must therefore be regarded as far-fetched and unconvincing.

85 Wessén & Jansson 1940–58, Vol. 4, 135. 86 Lerche Nielsen 2003. 87 Lerche Nielsen 2003, 205. 88 Källström 2011. 89 Sundqvist 2002, 112ff. 90 Janson 1998, 22. 91 Lerche Nielsen 2003, 195. 92 Göthberg, Lovén & Dahlbäck state: “Jansons omtolkning har svagheter. Viktigt är att han inte pekar på några paralleller till att Adam skulle ha anklagat konkurrerande kyrklig organisationer för att bedriva människooffer. Vidare är det oförklarat varför kung Stenkil skulle ha haft en så svag ställning att han inte kunde låta Adalvard d.y. få Uppsala till biskopssäte efter att biskop Osmund lämnat landet.” Göthberg et al. 2010, 34. A number of other researchers have raised critical arguments against Janson’s theories of Adam’s text, see e.g., Nilsson 2000; Gahrn 2000; Hultgård 2001b; Sundqvist 2002. 93 See Nilsson 2000. 118 chapter 4

4.3.3 “Pagans” and “Christians” in Eleventh-century Old Uppsala According to Anne-Sofie Gräslund, who has investigated the distribution of runic stones and their relationship to agrarian districts and cemeteries, the raising of cross-marked stones served to consecrate pagan cemeteries so that Christians could be buried there. In the northern area more runic stones are connected with cemeteries of Late Iron Age character than in the southern area. This implies that pagan cemeteries were used longer in the northern parts. Christianity advanced gradually towards northern Uppland.94 Gräslund concludes her study thus: “Det är en antydan om att man använde de hedniska gravfältet här [in the northern parts of Uppland] längre än i det södra området, något som väl skulle passa in i bilden av kristendomens gradvisa framträngande mot nordväst i Uppland.”95 In the southern parts of Uppland where runic stones were raised, churches were also erected. In the northern parts pagan burial practices continued at the same time.96 The archaeologist Anders Broberg argues that the conversion did not take place until c. 1150 in the northern parts and that traditional cemeteries still were in use there in the late twelfth century. The formation of parishes was not completed until the end of the thirteenth century.97 It should also be noted that the burial site at Valsgärde, situated a couple of kilometres north of Old Uppsala, shows continuity in burial customs from the fifth century up to the end of the eleventh century.98 There are, for instance, boat graves and chamber graves (graves 25 and 28, and perhaps grave 26) located in the north-western part of the cemetery, which seem to belong to the late eleventh century or even the early twelfth century. It is possible that some people living at Valsgärde were still pagans during this period, or that after conversion they only nominally defined themselves as Christians. Thus they maintained some traditional and indigenous customs and beliefs.99 Cremation burials from the Late Iron Age were found east of the Old Uppsala historical site during excavations in summer 2012. In one of the tombs, one Saxon coin dated to the period between 1015 and 1060 was discovered. The coin had a pierced hole, suggesting that it worn for a time as a pendant or a symbol.100 The coin must have been placed in the grave

94 A.-S. Gräslund 1987, 258f. 95 A.-S. Gräslund 1987, 259. 96 A.-S. Gräslund 1985. 97 Broberg 1991. 98 Arwidsson 1980, 45; Schönbäck & Thunmark-Nylén 2002; Ljungkvist 2008b. 99 Cf. Ljungkvist 2008b, 45f. 100 The coin was decorated with a Christian cross and thus may have acted as a Christian symbol. What speaks against this interpretation is that the coin appears in a cremation Ceremonial Buildings And Sanctuaries 119 at least one or two decades later than the date of issue. This tomb reveals that people in Old Uppsala at least until the middle of the eleventh century were buried in accordance with the ancient pre-Christian practice.101 Likewise, cremation graves in five smaller piles “in the cemetery beside the Kings’ Mounds” (SHM 23316) indicate pre-Christian burial customs. They were examined by Greta Arwidsson and Gunnar Ekholm in 1944. In two of these cremation tombs double combs (Sw dubbelkammar) were found, suggesting that the graves could be from the Late Viking Age, eleventh century or even younger.102 Old Norse sources confirm the idea that parts of Svetjud were Christianized only at a late date.103 The story about Blotsveinn in the U-version of Hervarar saga might reflect a pagan reaction somewhere in central Uppland during King Ingi Steinkelsson’s reign 1080–1110.104 This account should not be interpreted literally. Some details about Blotsveinn, for instance, are not historically reliable. On the other hand, we cannot rule out that the story reflects a real conflict between King Ingi and those people who still remained at and defended the “Uppsala temple”.105 The tradition about King Ingi and Blotsveinn also appears in Orkneyinga saga, which is usually dated to the early thirteenth century (see further ch. 9 below). We also have information in the legend of St Botvid (Vita sancti Botvidi) about a pagan revolt as late as the 1120s in the Mälaren area.106 This vita is preserved in a thirteenth-century manuscript, but most likely it was formulated as early as the twelfth century.107 Other written sources likewise support the existence of pagan cult in Svetjud in the late eleventh century.108

grave, that is, a form of burial which undoubtedly perceived as a pre-Christian custom. But we cannot rule out that a syncretistic situation existed in Old Uppsala during the mid eleventh century. See further below. 101 Personal communication from Hans Göthberg, October 2012. See also A.-S. Gräslund 2013. 102 A.-S. Gräslund 2013. 103 Cf. L. Lönnroth 1996, 141–158. 104 Hervarar saga. (Ed) E. O. G. Turville-Petre, 70f. Cf. (Ed) Jón Helgason, 160. Lönnroth 1996, 150f. 105 “Blotsven har förmodligen aldrig regerat vårt land . . . Däremot är det inte alls omöjligt att berättelsen om honom trots allt bygger på minnen av någon autentisk konflikt mellan den kristne kung Inge och de hedniska krafter som ännu höll fast vid Uppsalatemplet.” Lönnroth 1996, 151. 106 SRS I, 377ff. 107 Schmid 1931, 109; N. Blomkvist et al. 2007, 186. 108 See DS, no. 25; SRS II, 389–404. See N. Blomkvist et al. 2007, 186. 120 chapter 4

That Christian runic stones were erected in the surroundings of Uppsala in the middle of the eleventh century109 actually harmonizes with Adam’s account. Adam mentions that no one is exempted from attendance at this festival and states as follows: “Kings and peoples all and singly send their gifts to Uppsala and, what is more distressing than any kind of punishment, those who have already adopted Christianity have to buy themselves out of those ceremonies.”110 It seems as if Christians and pagans lived side by side in Svetjud and Uppsala during that period. Adam’s statement perhaps suggests that the leaders of the old religion may have had a far-reaching religious- political tolerance. In place of coercion, there was a legal opportunity for Christians to be exempted from the old cult.111 Perhaps Christians could visit or even stay at the old cultic place in Uppsala and its surroundings. As the runic stones show, there were many Christian people in the neighbourhood of Old Uppsala during the mid eleventh century.112 The old cult in Uppsala probably did not have as many followers as during the first decades of this century. Many people who lived in central Uppland had already converted to Christianity around 1050. Some were probably mixed in faith and lived in a syncretistic state. The cross-adorned Saxon coin that was discovered in a cremation grave may be an expression of this. Perhaps many people were Christians only in a nominal sense. They continued their pre-Christian customs, just as we hear that people in Iceland and Norway did during the transition period, according to the Old Norse literature.113 In all probability there was still a pagan elite in Old Uppsala in the middle of the eleventh century who had political power and authority to charge a fee of Christians who wanted to be released from the old cult.

4.3.4 Sources Supporting Adam’s Account of Uppsala There is no doubt that Adam’s descriptions sometimes were influenced by tendencies, clichés and stereotypes, as assumed by Anders Hultgård and

109 According to Torun Zachrisson (1998, 126–164, 224f.), the Christian runic stones in Uppland were raised by people belonging to a Christian community mainly resident in Sigtuna. 110 Adam IV,27. The Latin text is quoted above. 111 Cf. Hultgård 1997, 29; Göthberg et al. 2010, 34. 112 A.-S. Gräslund 2013. 113 There are many Icelandic stories of those who wavered in the faith during the transition period. In Landnámabók (S218, Ísl. Fornr. 1, 250) we hear about Helgi the Lean, who was very mixed in the faith. He believed in Christ, but invoked Þórr when travelling and facing difficult business (Helgi var blandinn mjǫk í trú; hann trúði á Krist, en hét á Þór til sjófara ok harðræða). Another who wavered in the faith was Óláfr Tryggvason’s poet Hallfreðr, whose contemporary poetry from 990–1000 expresses the faith conflict. See Skj. B1, 147–163. See also Ljungberg 1938. Ceremonial Buildings And Sanctuaries 121 others. In my opinion, however, Adam’s intention was to describe the pagan cultic place of Uppsala as it appeared in the middle of the eleventh century. Adam had first-hand information regarding Svetjud, from several people who had lived there. The Danish king Sveinn Ástriðsson, for instance, was exiled from Denmark and stayed in Svetjud sometime during Anund Jacob’s reign (c. 1020–1050).114 Adam explicitly states that Sveinn informed him about conditions there.115 Adam is specifically influenced by Sveinn’s perspective on the political relationship between Denmark and Svetjud.116 He also had information from Sveinn concerning geographical and religious matters in Svetjud.117 Most likely Sveinn was also one of the most important informants regarding the eleventh-century Uppsala cult.118 Bishop Adalvard the Younger of Sigtuna was also an important informant for Adam regarding conditions in Svetjud. Most likely they knew each other well from the Chapter of Bremen.119 Adam refers explicitly to his information, for instance, regarding a failed coloniza- tion of Svear in “Kvänland”.120 It has been suggested that Adalvard is Adam’s main informant concerning the sanctuary of Uppsala. When Adam states that “a certain Christian” informed him that “he had seen seventy-two miscellaneous bodies suspended there [in the sacrificial grove of Uppsala]” it may very well have been Adalvard.121 Adam was informed that Bishop Egino of Scania together with Adalvard the Younger planned to visit the temple of Uppsala in the 1060s, in order to do missionary work there. They even wanted to burn it down. King Steinkell stopped them because they would have been killed and he himself would have been dethroned (a regno).122 Most likely this information came directly from Adalvard to Adam. In my opinion, we must therefore assume that pagan cult still took place in some kind of building in Old Uppsala in the 1060s. There are also essential aspects of Adam’s description of Uppsala which may be supported by information found in independent direct sources, such as place-names and archaeological finds. Adam mentions, for instance, that the three gods Óðinn, Þórr and Freyr had their idols in the “temple” of Uppsala. Place-names actually confirm that these gods were worshipped in Svetjud.

114 Adam II,73–74. 115 Adam I,61, II,30, 35 and 38. 116 Adam I,48 and 52, I, 30, 38–39, 73 and 78. 117 For geography Adam IV,21 and religio-political matters Adam II,73. 118 Hallencreutz 1984, 359f. 119 Hallencreutz 1984, 360. It should be noted that Adam also had some information from some of Adalvard’s companions in Sigtuna. See Adam, scholion 142. 120 Adam, scholion 123. 121 See Adam IV,27. Cf. Hallencreutz 1984, 360. 122 Adam IV,30. 122 chapter 4

In the Mälaren area there are at least seventeen place-names containing the name Þórr, sixteen with Freyr and twelve with Óðinn.123 In my opinion, it is no coincidence that Adam mentions these particular gods, since they were the most worshipped ones in the surroundings of Uppsala. Archaeological finds may also support the existence of cult images or symbols representing these gods in pre-Christian Svetjud. A small Viking Age bronze statue discovered in Rällinge, Södermanland, for instance, has a phallic shape and might represent Freyr (see fig. 13 below). The phallic shape resembles Adam’s description of Freyr in the “temple”: “His likeness, too, they fashion with an immense phallus (cum ingenti priapo)” (see ch. 7 below). Adam says that Thor (Þórr) has as his attribute a sceptre and thus he resembles Jove (Thor autem cum sceptro Iovem simulare videtur).124 Þórr’s attribute was probably not a sceptre in the Mälaren region during the Viking Age. In this passage Adam has applied an interpretatio romana when comparing Þórr with Jove. The most important symbol of Þórr was his hammer. It probably appears on two Viking Age runic stones from Södermanland, the Åby and Stenkvista stones (see fig. 5). On the Altuna stone (eleventh century), Uppland, we may see the mythical scene when Þórr with his hammer tries to catch the Miðgarðsormr (see fig. 6). On this image the deity’s feet are pushed down through the bottom of the boat, in exactly the same manner as Snorri describes in Gylfaginning 48: “Then Þórr got angry and summoned up his ás- strength, pushed down so hard that he forced both feet through the boat, . . .”125 There are plenty of Þórr’s hammer symbols in Uppland (see fig. 19A). Some of these symbols are also found in Old Uppsala, for instance several Viking Age Þórr’s hammer-amulets.126 Thus we have good evidence that at least Þórr was worshipped in Old Uppsala and Svetjud. Old Norse traditions, appearing for example in the Kings’ Sagas, also support Adam’s idea that pre-Christian Uppsala was an important cultic site as well as a common assembly place among the Svear. Snorri Sturluson writes in Óláfs saga helga 77 (1230):

123 See Vikstrand 2001, 55ff., 115ff., 141ff., 167ff. It has been argued that the god Ullr belonged to a more ancient stratum of deities. See e.g., T. Andersson 1992c, 256. 124 Adam IV,26. 125 Gylfaginning 48: Þá varð Þórr reiðr ok færðisk í ásmegin, spyrndi við svá fast at hann hljóp báðum fótum gǫgnum skipit . . . Cf. also Meulengracht Sørensen 1986. 126 Personal communication from Lena Jörpeland, Hans Göthberg and others in January 2013. One Viking Age Þórr’s hammer symbol was found in June 2015 in the area called Kungsgården at Old Uppsala, close to the place where the hall building was found. Personal communication from the excavator John Ljungkvist. Ceremonial Buildings And Sanctuaries 123

Figure 5 The Stenkvista runic stone, Södermanland (Sö 111). Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet.

Figure 6 The Altuna stone, Uppland (U 1161). Photo: Pål-Nils Nilsson. ATA, Riksantikvarieämbetet.

At the time when heathendom still prevailed in Svetjud, it was an old custom there that the main sacrifices were held at Uppsala in the month of gói.127 Sacrifices were to be made at that time for peace and victory for the king. People from all Svíaveldi were to resort there.128

127 Sometime from 15 February to 15 March. 128 Í Svíþjóðu var þat forn landssiðr, meðan heiðni var þar, at hǫfuðblót skyldi vera at Uppsǫlum at gói. Skyldi þá blóta til friðar ok sigrs konungi sínum, ok skyldu menn þangat sœkja um allt 124 chapter 4

In Snorri’s accounts of pre-Christian Uppsala, ceremonial buildings are often mentioned. According to him, there were both huge hof buildings (hof mikit) and great halls (ON hǫll, salr) erected at this site.129 One problem is that these texts were composed hundreds of years after the conversion in Iceland. Thus the distance in both time and space between the texts and the existence of these buildings restricts their source value. On the other hand, there is nothing to indicate that Snorri relied on Adam’s description of the cultic site in Uppsala. As far as I know, he never mentions Adam’s name.130 Consequently, even if Snorri’s texts often have a mythical-legendary character, their information about ceremonial buildings in Uppsala must be seen as independent of Adam’s descriptions. Some of these Old Norse traditions may be quite ancient. In Ynglinga saga 29 (c. 1230), for instance, Snorri tells that King Aðils came to Uppsala in connection with the sacrifices honouring the mythical beings called the dísir (konungr var at dísabloti). He rode a horse around the hall of the dís (um dísar- salinn). The horse stumbled and fell; the king landed on his head and broke his skull. He died there and was buried in a mound in Uppsala. This tradition appears in Ynglingatal (c. 890), but neither the dísablot nor the dísarsalr are mentioned there.131 However, it is said that Aðils fell from the horse and died in Uppsala. Snorri’s idea that Aðils was performing pagan rituals at a ceremonial building, on the other hand, finds support in other older texts. Historia Norwegie (c. 1160–75) states that Aðils “gave up the ghost after falling from his horse before the temple of Diana, while he was performing the sacrifices made to idols”.132 Dianæ edes may be seen as an interpretatio romana for “the dís’ [perhaps Freyja’s] temple”.133 Thus we have a fairly old tradition about a ceremonial building in Uppsala consecrated to the dís. That the cult of the dísir

Svíaveldi. Óláfs saga Helga, in Heimskringla II, Ísl. Fornr. 27, 109. Trans. Hollander. See also Ynglinga saga 34. 129 See e.g., Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 23, 41, 65f. 130 Adam’s text was known in fourteenth-century Iceland. One passage of it is translated into Old Norse in AM 415 4to (c. 1310) and in Flateyjarbók (c. 1390). But Snorri was probably not acquainted with it. Cf. Lassen 2011, 122f. 131 On my view of the dating, provenance and source value of Ynglingatal, see Sundqvist 2002, 41–52. The designation dísarsalr also occurs in Hervarar saga 7, (ed.) E. O. G. Turville-Petre, 30. 132 . . . ante edem Diane, dum ydolorum sacrificia faceret, equo lapsus expirauit. Historia Norwegie IX, (eds.) Ekrem & Mortensen 2003, 76ff. 133 The author of Historia Norwegie was probably acquainted with Adam’s work. See Ekrem & Mortensen 2003. Most likely, however, Snorri never used Historia Norwegie for his presentation of Aðils. Ceremonial Buildings And Sanctuaries 125 was important in Uppsala can be seen in the name of the fair there, Disthingen (OSw Disaþing).134 It was held during the spring and might be identical with the calendrical feast mentioned by Adam (at the vernal equinox) and Snorri (in the month of gói). Place-names also indicate that the dísir were worshipped at specific sanctuaries in the eastern Scandinavian area. The name Disevid in Östergötland, for example, has been interpreted as ‘the vi sanctuary of the dísir’.135 Most recent research discusses whether the term templum in Adam’s text really refers to an exclusive religious building. Evidence indicates that Adam was actually referring to a multifunctional hall building when using this term in this context. This building was not only intended for religious rituals and symbols, but also for other purposes, such as a dining room at banquets, a room for political-judicial meetings and a gathering place for the hirð. The room inside the “temple”, where the idols were placed, is called triclinium by Adam. In classical and mediaeval Latin it has both the meaning of ‘dining-room’ and ‘room for ceremonial banquets’.136 It should also be noted that the Latin term used for designating the sacrificial rituals in the ceremonial building, libare, actually refers to drinking ceremonies. It thus seems as if Adam’s “temple” resembles the banqueting halls described in the medieval Old Norse texts. In Hauks þáttr hábrókar, for instance, a hall (hǫll) located in Uppsala is mentioned. In this hall King Eiríkr made libation sacrifices to the image of the god called Lytir.137 There is archaeological evidence for at least one banqueting hall in prehistoric Uppsala. On Södra Kungsgårdsplatån, just north of the church, post-holes and a stone structure belonging to a Merovingian Period hall (50 × 12 m) were found (see fig. 2A and 2B).138 Also on the northern plateau (Norra Kungsgårdsplatån) a large three-aisled building was recently found.139 It was smaller than the hall on Södra Kungsgårdsplatån, but also dated to the Merovingian Period. It seemed to have been a very exclusive workshop.140 We may recognize the

134 See e.g., Upplandslagen (1296 AD). SSGL 3: 274−275. See Naumann 1984. 135 See Sundqvist & Vikstrand 2014. This name is debated, however. See Elmevik 2005. 136 See Dillmann 1997, 65–69, 72. 137 Óláfs saga Tryggvasonar, Flateyjarbók I, 579−580. The text is quoted in ch. 9 below. 138 Nordahl 1993; 1996; Hedlund 1993; A.-S. Gräslund 1997. Scholars have also argued that the last element in the name Uppsala refers to ‘banqueting halls’ (pl.). E.g., B. Gräslund 1993; Brink 1999b, 38−39, 48−49, note 15; Herschend 2001, 39ff. For a recent treatment of the name Uppsala, see Vikstrand 2013. New excavations at Old Uppsala in the summer of 2011 indicate that the hall was 50 metres in length. 139 Personal communication from the archaeologist and chief excavator at Old Uppsala, Hans Göthberg, August 2011 and e-mail from John Ljungkvist, October 2011. 140 Personal communication from John Ljungkvist. 126 chapter 4 pattern of a small building beside a large residence, as seen in Gudme and other places in southern Scandinavia (see 4.1). Since the large hall building at Kungsgården is dated to the Merovingian Period and was abandoned in the Early Viking Age, it cannot be identical with the cultic building (templum) mentioned by Adam. It nevertheless indicates that the monumentalization process of power in Uppsala took place in the Merovingian Period.141 The large hall on the southern plateau was probably embellished with spiral decorations, made of iron.142 It seems further that these iron spirals were deposited in the post-holes of the house after the building was burned down in ritual forms c. 800. These decorations and the ritual deposits indicate that the house had a special function. Even the discovery that the interior of the house was whitewashed shows that it was special. These aspects indicate that this house was used in ceremonial or feasting contexts. Post-holes and a stone packing indicate that a large building also was erected on the spot of the present church. A fragment of a mould for a brooch (of the type JP 51), found in the first clay fill under the church, dated to the tenth century, indicates that activities took place there during the Viking Age.143 Else Nordahl therefore cautiously suggested that the building should be dated to the Viking Age.144 Recent excavations indicate that the house is older.145 The traces of large buildings in Uppsala are located in the area of the old royal demesne (Kungsgården). These facts indicate that the ruling power had some kind of influence on these houses. Also not far from the royal demesne are the famous royal grave mounds, dating back to the sixth or seventh century. They were thus situated in the immediate context of the great Merovingian Period buildings at Uppsala and must therefore be considered as an element in the sanctuary complex. Recent archaeological findings indicate that Old Uppsala may have involved a kind of temenos, i.e., a consecrated area. The long row of foundations for standing posts found during the period 1996 to 2013, both south and north of the site, may well have been intended to ritually separate the cult area.146 Within

141 Cf. Ljungkvist 2013. A seventh century pendant of gold was found close to the hall in June 2015. Personal communication John Ljungkvist. 142 Personal communication from John Ljungkvist, January 2013. 143 Nordahl 1996, 63, 73. 144 Nordahl 1996, 60–62. 145 See Göthberg 2008. He has also shown that the houses indicated by means of ground- penetrating radar (see Alkarp & Price 2005) cannot be confirmed by the archaeological investigations made at the church in 2005. 146 Beronius Jörpeland et al. 2013; Sundqvist 2013, 89f. and personal communication Per Frölund 2012 and Jonas Wikborg 2013 and 2014. This long row of posts could also indicate the existence of a procession road. Ceremonial Buildings And Sanctuaries 127 this area, at least one ceremonial building and several great burial mounds were located during the Merovingian Period and Early Viking Age. Perhaps the large amount of deposited amulet rings (roughly 80), knives and bones, which were found at the Fire Station in Old Uppsala in the summer of 2012, indicate that the consecrated area was large and also extended to the east. These new findings occurred in connection with a settlement with a large Late Iron Age burial ground (with over 90 cremation graves).147

Adam of Bremen’s text is a result of a literary process whereby the information from his informants has been reworked. The text has clearly been influenced by Christian polemics and rhetorical embellishments and therefore cannot be relied upon in every detail. Nevertheless, since some of his information about a cultic building and sanctuary in Uppsala harmonizes with independent Old Norse written sources, it should not be completely disregarded. Archaeology indicates that at least one feasting hall was erected in Uppsala during the Merovingian Period and Early Viking Age. It is unclear, however, whether any archaeological evidence can be connected with Adam’s “temple”, but pre-Christian burial customs can be attested in Old Uppsala, at least up until c. 1050.

4.3.4 Other Ceremonial Buildings in the Mälaren Region Attested by Archaeology Archaeological excavations show that aristocratic halls in the Mälaren region in general were used as a kind of ceremonial buildings. On Helgö (Lillön), Ekerö parish, Färentuna härad, Uppland, a three-aisled hall was erected on a terrace (Foundation I A, at Building Group 2) in an E–W orientation (fig. 7a and 12B).148 It was built during the Migration Period and remained there during several phases up to the Viking Age. It was 20–21 metres long and about 8 metres wide. The find distribution at this house indicates that it was divided into two rooms, one “hall-room” and one “chamber”.149 In the “hall-room area”, which was 8 × 8 metres, several interesting finds were made, such as gold foil figures (fig. 7b), fragments of a bracteate and exclusive imported glass. They indicate that cultic activities took place at this area. In or near this building a Buddha figurine, an Irish crosier and a Coptic scoop were also found, which all may have had a symbolic significance for the users of these objects. In the south-western

147 Personal communication from Lena Beronius Jörpeland and Hans Göthberg 2013. 148 See Herschend 1995; 1998; K. Lamm 2004; J. P. Lamm 2004; Arrhenius 2011; Lund Hansen 2011. For an overview of Helgö in general, see K. Lamm 1999. 149 Herschend 1995. Arrhenius suggests that the house was divided into four or five rooms. See Arrhenius 2011, 20. 128 chapter 4

Figure 7a The hall at Helgö. Find distribution in the hall on Foundation Ia at Helgö. (a) Fragments of filigree glass claw and cone beakers = open triangle. (b) Guldgubbar = open square; weapons = filled square. (c) Knives = filled triangle; loom- weights and wetstones = a dot. (d) Density map of the kind of distribution. The first isarithmic curve marks areas with at least 1 find per m2. The most dense square metre contains 8 finds, and the entire house (125 m2) contains 90 finds. The ‘H’ marks a presumed position of the high seat. After Herschend 1998.

Figure 7b Gold foil from Helgö. SHM. Ceremonial Buildings And Sanctuaries 129 corner of the banqueting room there was a concentration of finds. This area has been interpreted as the place of the high-seat.150 Most of the gold foil figures and the expensive glass were found there. In front of the assumed “high-seat” was also the hearth, which may have had a ritual significance during the ceremonial banquets which took place in the hall. In my opinion, there is no doubt that the hall at Helgö had a cultic significance. Besides the hall, an open-air ritual site was found at the foot of a bare rock. Arrowheads, ceramics and food were deposited there under a Viking Age triangular stone setting, i.e., a “tricorn” (Sw treudd) (fig. 12A).151 This sacrificial place was used during Merovingian Period too. There were also other large residences (longhouse VI and hall III A2) during the Migration and Merovingian Periods located beside the cultic hall (I A) at Helgö, indicating a situation similar to what is seen at Gudme, Funen and Uppsala. During the Viking Age there were two hall buildings, I A and V A. According to Birgit Arrhenius, several finds at these buildings indicate that they should be interpreted as cultic. She argues, for instance, that the traces of a rectangular platform in V A should be interpreted as something equivalent to the Old Norse seiðhjallr, used during divination rituals (ON seiðr).152 She also suggests that the post which stood in post-hole 49 in hall building I A was a large “phallus post”.153 In addition to this, Arrhenius proposes that some kind of ceremony also took place at the early boat-shaped building III A1 in Building Group 2. The boat kept in this building was used during sacrifices to the gods, especially Freyja. It was filled with sacrificial objects and then carried in procession to the lake. In my opinion, some of Arrhenius’s suggestions must be considered very hypothetical. The place-name Helgö (OSw *Hælghø) nevertheless indicates that this place undoubtedly was an important cultic site.154 This name has been interpreted as ‘the holy island’ or ‘the island where peace prevails’.155 It has been argued

150 See Herschend 1995. 151 See especially Torun Zachrisson’s (2004a; 2004b; 2010) important studies of Helgö. See also Anders Andrén’s (2004) interesting interpretation of the triangular stone-settings (Sw treuddar). 152 See Arrhenius 2011. Whether there ever were permanently constructed seiðhjallar in ancient Scandinavia is in my opinion uncertain. It does not have clear support in reliable written sources. The practice of seiðr was a crisis ritual performed only occasionally when the situation needed it, and the seiðhjallar mentioned in the reliable texts seem to have been temporary structures. See Sundqvist 2012a. 153 Arrhenius 2011, 23. 154 See e.g., T. Andersson 1999b; Vikstrand 2001, 244. Cf. Calissendorff 1964. 155 Cf. Calissendorff 1991; Vikstrand 2001; SOL, 123. 130 chapter 4 that the name indicates that certain activities took place at this island, such as trade and cult, which were both protected with peace. Sanda in Central Uppland (see ch. 5 below) and the island of Birka in Mälaren (see ch. 6 below) also have remains of Late Iron Age halls, appearing in contexts of finds indicating cult.156 Similar indications of Iron Age halls (sometimes on terraces) have also been found at several other central places in the Mälaren region, for instance at Hovgården on Adelsö, at Fornsigtuna, Granby-Hyppinge in Vallentuna, Karsvik in Bromma, Runsa in Upplands Väsby, Barksta in Västmanland, Valsgärde and Vendel in Uppland.157 In these cases it is mainly the construction and/or the location of these buildings which indicates that they may have had a possible cultic or ceremonial function in connection with banquets. At the hall of Husby, Glanshammar, Närke, dated to the Migration and Merovingian Periods, there was a demarcation or palisade around the building.158 This demarcation, and the gold foil figure discovered to the west of the hall, support the idea that this building may have had a ritual function. Perhaps a hall was also located at the same spot as the monastery in Eskilstuna. Adjacent to a row of stone-edged post-holes two gold foil figures were discovered, dated to c. 700.159 A gold foil figure from the eighth or ninth century was also found at the hall building of Svintuna, in Östergötland.160 There is also slight evidence of more specific cultic houses in the Mälaren region.161 To the north of the Middle and Late Iron Age hall in Lunda, Södermanland, remains of a small building (3 × 6 m) were found together with two phallic figures.162 A third figure was found in the yard south of the hall. According to the archaeologists the small house should be interpreted as a specific cultic house, comparable to the structures found at Tissø and Järrestad. In the yard to the north of the hall, several pits were discovered, probably intended for ritual cooking. Þórr’s hammers, amulets and miniature scythes

156 Åqvist 1996; Holmquist Olausson 2001; Holmquist Olausson & Kitzler Åhfeldt 2002. 157 Brunstedt 1996; Damell 1991; 1993; A. Carlsson 1997; Hedman 1989; Ringsted 2005; Karlenby 2005; Norr & Sundkvist 1995; Arrhenius 2000; Skre 2007a; Olausson (ed.) 2011 and Olausson 2014. 158 It should be noted that several halls (“house II” and “house III”) may have succeeded each other. Ekman 2000, 25. 159 See Sune Zachrisson 1963 and 1998, 13f. The archaeologist Torun Zachrisson (the daughter of Sune Zachrisson) has orally suggested that there are reasons to interpret the finds in Eskilstuna as a hall. 160 Herschend 1998. 161 Indications of a more specific three-aisled cultic house may have been found at Alby in Botkyrka parish, Södermanland. See Hjulström 2008. 162 G. Andersson et al. 2004.; Skyllberg 2008. Cf. L. Larsson 2011. Ceremonial Buildings And Sanctuaries 131 were also found in this yard. A grove was situated on a hill approximately 200 metres to the west of this ritual area. Archaeological finds show that rituals were performed there too.

4.3.5 Hov-, Harg- and Vi-places Place-names may contribute information about pre-Christian ceremonial buildings and cultic places in the Mälaren region. Names containing words such as OSw *al- (*alh-), *hargher (hǫrgr), hov (hof), lunder (lundr), sal (salr), staver (stafr) and vi (vé) may designate cultic sites, especially when they have theophoric first elements.163 In the context of ceremonial buildings the word hov (hof ) is the most interesting term in the place-names.164 Per Vikstrand has suggested that the hov-names in the Mälaren region refer to cultic buildings, which all appear in a very specific local context.165 They do not designate the chieftain’s hall located at a central place. Instead, they are related to minor farms situated separately and at some distance from the major settlements. The hall buildings often appear at places with names including the term ‑tuna, such as Fornsigtuna (and Sigtuna), Eskilstuna, Tuna in Vendel, and Svintuna. The names in ‑tuna seem to designate places of central importance in the Late Iron Age Mälaren Valley.166 It is also possible that names including sal may refer to halls.167 Place-names including the term vi (cf. ON vé) also occur frequently in the Mälaren region.168 Old Swedish vi is a noun derived from a Germanic adjective *wīha- ‘holy’.169 When it is used as a place-name, it probably has the meaning ‘holy place’ or ‘sanctuary’. Most interesting is the cultic site of Lilla Ullevi, in Bro parish, Uppland.170 The name Ullevi (*Ullarvi) may be interpreted as ‘the sanctuary of the god Ullr’. At this site a ritual area was discovered covered with stones. In this area a lot of extraordinary finds were made, such as amulet rings, objects for ritual usage, a ceremonial meat fork, silver objects, arrow points etc.

163 For a critical investigation of these terms, see Vikstrand 2001; 2002, 129ff.; 2004a. 164 It should be noted that the word hof could also mean ‘farm’ or ‘elevation, height’. The latter meaning is preserved in Norwegian dialects. Sandnes 1992, 261; Vikstrand 2002, 132. See also below. 165 Vikstrand 2001, 256ff., 267ff. There are nine places called Hov (simplex) in the Mälaren region, seven in Uppland, one in Södermanland; one in Västmanland. On the word ON hof as a ceremonial building in different contexts, see also Sundqvist 2009a. 166 See Vikstrand 2000. 167 See the discussion in Vikstrand 2013. 168 In this region there are forty-nine place-names which include the term vi. 169 Vikstrand 2001, 298ff.; 2002, 129f. 170 Bäck et al. 2008.; Hållans Stenholm 2010; and on the place-name, see Vikstrand 2010. 132 chapter 4

No ceremonial building was found, however. Also very interesting is the Viking Age site of Götavi in Närke.171 The last element of the name, -vi, indubitably designates a holy place.172 Archaeological finds also indicate that this site was an important outdoor cultic place during the Viking Age.

To sum up, the combination of a monumental ceremonial hall and royal burial mounds in Uppsala indicates that this site was an important sanctuary in the Late Iron Age. In my opinion, there are no reasons to rule out that this place had a similar function in the middle and at the end of the Viking Age, even if the archaeological evidence does not clearly indicate this. If Uppsala had lost its significance during this period, one must ask why it was so important to establish an episcopal see there during the 1120s, and later make this place into the archiepiscopal see in 1164. My only answer is that Uppsala had an enormous symbolic significance for the Svear during the entire Viking Age. The symbolic capital that Uppsala constituted was also useful for the church. Some information in Adam’s description may thus be based on authentic conditions, for instance that Uppsala was an important sanctuary for the Svear. Written sources, place-names and archaeological finds indicate that pre- Christian halls and cultic buildings also existed at other places in the Mälaren area. At certain places, such as Helgö, we may see the pattern of a large residence situated beside a small hall building. Perhaps there was a functional difference between these buildings (see above).

4.4 Lade and Mære in Trøndelag

Unlike the situation in eastern Scandinavia, there are plenty of indigenous written sources mentioning pre-Christian ceremonial buildings in Trøndelag and the ritual activities which took place there. In these descriptions pagan rulers often play a central role in the rituals. Some of these accounts are also found in contemporary Viking Age sources, such as the skaldic poetry. But the most extensive descriptions of ceremonial buildings and sanctuaries appear in the problematic medieval Old Norse prose. Even if these texts are often influenced by the medieval reception, they should not be rejected a priori, but

171 See Svensson 2010. 172 The first element of the name Götavi is problematic and may refer to the name of a people (Götar), the name of a god *Gauti, or a (sacred) well (from the verb gjuta). See Vikstrand 2010. Ceremonial Buildings And Sanctuaries 133 critically evaluated by comparisons with more direct sources when available. For there are other sources informing about ceremonial buildings and sanctuaries in this area, for instance archaeology and place-names. Compared to the Mälaren area, however, the witnesses from archaeology are more restric- tive in this region. But we have some places such as Mære in Inn-Trøndelag and Ranheim in Ut-Trøndelag with many interesting finds. In what follows I will also take Borg in Lofoten into consideration, since this place may have been related to the command of the Earls of Lade. They derived originally from Hálogaland, between Trøndelag and Lofoten.

4.4.1 Snorri’s Description of Ceremonial Buildings in Lade and Mære One of the most extensive description of ceremonial buildings in Trøndelag appears in Hákonar saga góða 13–18, a part of Snorri Sturluson’s Heimskringla.173 Snorri is thought to have written Heimskringla about 1230, and it is preserved in several witnesses. Two of them are later copies of the medieval parchment codices Kringla and Jǫfraskinna. A third one, Codex Frisianus, was written by an Icelander c. 1325 (see ch. 3, note 1 above). Snorri writes thus:

Sigurðr Hlaðajarl was most ardent heathen worshipper, as had been Hákon his father. Sigurðr jarl maintained all sacrificial feasts there in Trøndelag on the king’s behalf. It was ancient custom that when sacrifice was to be made, all farmers were to come to the hof and bring along with them the food needed while the feast lasted. At this feast all were to take part in the drinking of ale. Also all kinds of livestock were killed in connection with it, horses also; and all the blood from them was called “hlaut”, and “hlautbolli”, the vessel holding that blood; and “hlautteinar”, the sacrificial twigs. These were fashioned like sprinklers, and with them were to be smeared all over with blood the “stallar” and also the walls of the hof, within and without, and likewise the men present were to be sprinkled with blood. But the meat of the animals was to be boiled and to serve as food at the banquet. Fires were to be lighted in the middle of the hof floor, and kettles hung over them. The sacrificial beaker was to be borne around the fire, and he who made the feast and was chieftain was to bless the beaker as well as all the sacrificial meat. Óðinn’s toast was to be drunk first—that was for victory and power to the king—then Njǫrðr’s and Freyr’s, for good harvest and peace. Following that many used to

173 Heimskringla I, Ísl. Fornr. 26, 166–73. On Heimskringla, see Whaley 1999. 134 chapter 4

drink a “bragafull”. Men drank toasts also in memory of departed kinsfolk—that was called “minni”.174

Snorri also states that Earl Sigurðr was generous and that he once made a great sacrificial feast at Lade, defraying all expenses himself. He supports his narrative with a stanza from Kórmakr Ǫgmundarson’s Sigurðardrápa (960 AD) where Earl Sigurðr is praised. In the frame story Snorri recounts how King Hákon the Good had had a Christian upbringing in England. He was unenthusiastic about the pagan cult in Norway, and wanted to convert the Norwegians gently. The farmers of Trøndelag opposed him and at the Frostaþing required the king to take part in the rituals: “The farmers said that it was their wish that the king should make sacrifice to procure for them good crops and peace, as his father had done.”175 Later, during the autumn, the king came to the annual sacrificial feast at Lade:

In the autumn, at the beginning of winter, there was a sacrificial feast at Lade and the king attended it. Before that, if present at a place where heathen sacrifice was made, he was accustomed to eat in a little house apart, in the company of few men. But the farmers remarked about it that he did not occupy his high-seat when there was the best cheer among the people. The earl told him that he should not do that; and so it came that the king occupied his high-seat [on this occasion]. But when the first beaker was served, Sigurðr jarl proposed a toast, dedicating the horn to Óðinn, and drank to the king. The king took the horn from him and made

174 Trans. Hollander (somewhat modified). Íslenzk Fornrit has this text: Sigurðr Hlaðajarl var inn mesti blótmaðr, ok svá var Hákon, faðir hans. Helt Sigurðr jarl upp blótveizlum ǫllum af hendi konungs þar í Þrœndalǫgum. Þat var forn siðr, þá er blót skyldi vera, at allir bœndr skyldu þar koma, sem hof var, ok flytja þannug fǫng sín, þau er þeir skyldu hafa, meðan veizlan stóð. At veizlu þeiri skyldu allir menn ǫl eiga. Þar var ok drepinn alls konar smali ok svá hross, en blóð þat allt, er þar kom af, þá var kallat hlaut, ok hlautbollar þat, er blóð þat stóð í, ok hlautteinar, þat var svá gǫrt sem stǫkklar, með því skyldi rjóða stallana ǫllu saman ok svá veggi hofsins útan ok innan ok svá støkkva á mennina, en slátr skyldi sjóða til mannfagnaðar. Eldar skyldu vera á miðju gólfi í hofinu ok þar katlar yfir. Skyldi full um eld bera, en sá, er gerði veizluna ok hǫfðingi var, þá skyldi hann signa fullit ok allan blótmatinn, skyldi fyrst Óðins full—skyldi þat drekka til sigrs ok ríkis konungi sínum—en síðan Njarðar full ok Freys full til árs ok friðar. Þá var mǫrgum mǫnnum títt at drekka þar næst bragafull. Menn drukku ok full frænda sinna, þeira er heygðir hǫfðu verit, ok váru þat minni kǫlluð. Heimskringla I, Ísl. Fornr. 26, 167f. 175 Bœndr segja, at þeir vilja, at konungr blóti til árs þeim ok friðar, svá sem faðir hans gerði. Heimskringla I, Ísl. Fornr. 26, 170. Ceremonial Buildings And Sanctuaries 135

the sign of the cross over it. Then Kárr af Grýtingi said “Why does the king do that? Doesn’t he want to drink of the sacrificial beaker?” Sigurðr jarl made answer, “The king does as all do who believe in their own might and strength, and dedicated his beaker to Þórr. He made the sign of the hammer over it before drinking.” People said no more about it that evening. Next day when people had seated themselves at the table, the farmers thronged about the king, saying that now he must eat the horse meat. That, the king would not do under any condition. Then they asked him to drink the broth from it. He would not do that, either, and they came near to making an attack on him. Sigurðr jarl said he would help them come to an agreement, asking them to cease their tumult; and he asked the king to gape with his mouth over the handle of the kettle on which the smoke of the broth from the horse meat had settled, so that the handle was greasy from it. Then the king went up to it and put a linen cloth over the handle, and gaped with his mouth over it. Then he went back to his high- seat, and neither party was satisfied with that.176

Later, at Yule, the king visited Mære, where the chieftains and farmers held ceremonial meals. On this occasion he was forced to eat a few pieces of horse- liver. He also drank the toasts without making the sign of the cross.

4.4.2 Criticism of Snorri’s Text Snorri’s text has been debated in recent decades. Olaf Olsen, for instance, was very sceptical of Snorri’s description of pagan cult.177 Although he accepted

176 Um haustit at vetri var blótveizla á Hlǫðum, ok sótti þar til konungr. Hann hafði jafnan fyrr verit vanr, ef hann var staddr þar, er blót váru, at matask í litlu húsi með fá menn. En bœndr tǫlðu at því, er hann sat eigi í hásæti sínu, þá er mestr var mannfagnaðr. Sagði jarl, at hann skyldi eigi þá svá gera. Var ok svá, at konungr sat í hásæti sínu. En er it fyrsta full var skenkt, þá mælti Sigurðr jarl fyrir ok signaði Óðni ok drakk af horninu til konungs. Konungr tók við ok gerði krossmark yfir. Þá mælti Kárr af Grýtingi: ’Hví ferr konungrinn nú svá? Vill hann enn eigi blóta?’ Sigurðr jarl svarar: ’Konungr gerir svá sem þeir allir, er trúa á mátt sinn ok megin ok signa full sitt Þór. Hann gerði hamarsmark yfir, áðr hann drakk.’ Var þá kyrrt um kveldit. Eptir um daginn, er menn gengu til borða, þá þustu bœndr at konungi, sǫgðu, at hann skyldi eta þá hrossaslátr. Konungr vildi þat fyrir engan mun. Þá báðu þeir hann drekka soðit. Hann vildi þat eigi. Þá báðu þeir hann eta flotit. Hann vildi þat ok eigi, ok var þá við atgǫngu. Sigurðr jarl segir, at hann vill sætta þá, ok bað þá hætta storminum, ok bað hann konung gína yfir ketilhǫdduna, er soðreykinn hafði lagt upp af hrossaslátrinu, ok var smjǫr haddan. Þá gékk konungr til ok brá líndúk um hǫdduna ok gein yfir ok gekk síðan til hásætis, ok líkaði hvárigum vel. Heimskringla I, Ísl. Fornr. 26, 171f. Trans. Hollander, somewhat modified. 177 See Olsen 1966, 59ff. 136 chapter 4

Snorri’s statement that pre-Christian ceremonial meals and drinking feasts were celebrated indoors, he regarded as uncertain some details in Snorri’s account, such as the description of the hof, the cultic objects, and the ritual actions that took place there. Ernst Walter likewise felt that Snorri’s description was suspicious, especially parts of the ceremonial drinking in chapter 17 (see further ch. 9).178 Klaus Düwel argued in the same vein as Olsen and Walter in his impressive study Das Opferfest von Lade (1985). The purpose of his study was to expose the structure of Snorri’s text and the storytelling. He also scrutinized the semantic development of certain religious terms, such as the concepts hlautteinn and hlaut, and finally he also investigated the sources Snorri may have used when describing the pre-Christian cult. In his conclusion, Düwel stated that the actions connected with the sacrifices at Lade had no Germanic origin, that is, that Snorri had no pre-Christian sources for these rituals. Rather, when describing how blood was sprinkled on the altar in the hof there, Snorri modelled his account on ancient Hebrew rituals described in Exodus 24, which was translated into Old Norse in the text called Stjórn.179 During the Early Middle Ages, the sacrificial rituals of the Old Testament were considered pagan, and pagan religion was thought to be the same in all places: this led Snorri to use this material in his account of the ancient cult in Trøndelag. Düwel also argued that Snorri projected customs common in medieval guilds back to heathen times, for instance, the ceremonial minni- drinking. In his analyses of the religious terminology Düwel concluded that in most cases Snorri either misunderstood these concepts or mixed them up with Christian ideas with no basis in pre-Christian culture, such as the term minni. Düwel also stated that Snorri’s purpose was actually not to describe the pagan cult, but to show how violent the missions of the Norwegian royal power had been in contrast to the situation in Iceland. Düwel concluded therefore that the description of sacrificial rituals in Hákonar saga góða is not reliable and thus cannot be used as a source for pre-Christian religion (see further ch. 9 below).

178 Walter 1966. 179 Compare the following phrases: (1) Stjórn 100 (Ed.) Unger, 305; cf. Exodus 24,6): Moyses tok halft fornarblodit ok hellti þui i kerit. en halft blod hellti hann yfir alltarann. (2) Hákonar saga goða 14 (in Heimskringla I, Ísl. Fornr. 26, 167f): Þar var ok drepinn alls konar smali ok svá hross, en blóð þat allt, er þar kom af, þá var kallat hlaut, ok hlautbollar þat, er blóð þat stóð í, ok hlautteinar, þat var svá gǫrt sem stǫkklar, með því skyldi rjóða stallana ǫllu saman ok svá veggi hofsins útan ok innan ok svá støkkva á mennina, en slátr skyldi sjóða til mannfagnaðar. Ceremonial Buildings And Sanctuaries 137

In a general sense the criticism by Olsen, Walter, Düwel and Baetke180 is well- founded. Düwel was, for instance most likely right in his conclusions about the terms hlautteinn and hlaut. However, sometimes this source criticism has been somewhat simplified and without nuances, especially when stating that Snorri had no access to ancient sources on these rituals. Preben Meulengracht Sørensen gave, in my view, a more balanced account.181 He agreed with Olsen and Düwel that some of the details in Snorri’s hof descriptions cannot be verified and thus it is not a trustworthy image of the pagan cult, as previous historians of religions sometimes claimed. He also agreed with these scholars that these texts must be related to Snorri’s Christian view of history. However, this does not mean that Snorri was not at pains to convey a reliable image of the pre-Christian sacrificial cult. Meulengracht Sørensen also wrote thus concerning Düwel’s critical opinion of the terminology that Snorri applied in his text:

At de centrale religiøse ord ikke med sikkerhed kan føres tilbage til hedensk tid, er ikke et tilstrækkeligt argument for, at de begreber, de udtrykker, ikke har været hedenske. Det samme gælder de parallelste- der, som Düwel påviser i den kristne litteratur. De er ikke enbetydende med, at Snorri har lånt sin fremstillings enkeltdele derfra. Begge forhold viser kun, at Snorri har udtrykt sig ved hjælp af kristen tankegang og terminologi.182

Without doubt Meulengracht Sørensen puts his finger on a crucial weakness in Düwel’s thesis. During the last few decades several other scholars have also treated Snorri’s text more positively. By means of careful analyses of the Old Norse materials, and using a comparative method, Anders Hultgård has been able to show that some elements in Snorri’s description may be of pre-Christian origin, such as the cultic formula til árs ok friðar.183 François-Xavier Dillmann has likewise shown that some aspects of Snorri’s account about banqueting may be built on old traditions going back to the Viking Age.184 I myself have argued that some rituals described by Snorri may be genuine.185

180 See e.g., Baetke 1951, 28. 181 Meulengracht Sørensen 1991b, 238f. 182 Meulengracht Sørensen 1991b, 239. 183 Hultgård 1993; 2003c; 2007. 184 Dillmann 1997. 185 Sundqvist 2005e; 2013b. 138 chapter 4

4.4.3 Snorri’s Sources In order to evaluate Snorri’s descriptions of the pagan cult in Trøndelag more carefully, we must identify his sources. In Hákonar saga góða 14 he quoted a stanza from Sigurðardrápa (c. AD 960) composed by the skald Kormákr, which describes ancient feasts held in Trøndelag. This stanza was made in honour of Earl Sigurðr:

Hafit maðr ask né eskis afspring með sér þingat fésæranda at fœra fats. Véltu goð Þjaza. Hver myni vés við valdi vægja kind of bægjask, þvít fúr-Rǫgni fagnar fens. Vá Gramr til menja.186

Most scholars interpret the expression valdr vés as “the ruler of the sanctuary”, i.e., Earl Sigurðr. Klaus Düwel has rejected this interpretation and suggested that the sequence valdr vés vægja should be interpreted as “der Beherrscher des Thingplatzes der Schwerter (= Schilde) = der Kriger Sigurðr”—that is, “the warrior Sigurðr”.187 According to Düwel, there is nothing supporting an assertion that Snorri knew about ancient traditions of the ritual feasts (blótveizlur) in Trøndelag; he had only support for the idea that Sigurðr was regarded as generous. Düwel’s position, however, has been contradicted. Skilled philologists such as Ottar Grønvik and François-Xavier Dillmann have put forward strong arguments in favour of the old interpretation of valdr vés as “the ruler (protector) of the sanctuary”.188 Grønvik relates vægja to kind ‘offspring’ and interprets the former word as a hapax legomenon of a noun vægir (m.) ‘godfearing man, that is, a man who is full of fear (for the deity)’.189 Vægja (gen. pl. of vægir) is interpreted as a nomen agentis of the verb vægja ‘turn aside’ (Norwegian

186 “You do not need to bring with you/ neither basket nor tankard/to the generous man./ The gods deceived Þjazi./ All men should avoid opposing/ the keeper of the sanctuary,/ because he makes the chieftains glad./ The ruler battled for the gold.” Heimskringla I, Ísl. Fornr. 26, 168. See also Skj. B1, 69f. My trans. 187 Düwel 1985, 14–17. 188 Norwegian: “véets (templets, hovets) høvding, hersker”. Grønvik 1989, 82–90. See also Dillmann 1997. 189 Norwegian: “gudfryktig mann, d.e. mann fylt av vægje, age (for guddommen)”. Ceremonial Buildings And Sanctuaries 139

‘vike til side’). Accordingly, the half-stanza is interpreted as: “Which offspring of godfearing men will oppose the ruler (protector) of the sanctuary?’190 In my opinion, Grønvik’s interpretation is quite convincing. It seems as if Snorri at least had some support for his description of ceremonial meals at cultic sites in Trøndelag, by means of the tenth-century stanza in Sigurðardrápa. The first half-stanza mentions that nobody needed to bring food to the banquets arranged by Sigurðr, because of his generosity. Since the other half of the stanza indicates that Sigurðr was regarded as a protector of the sanctuary, we may suppose that the first half of the stanza actually referred to a religious feast, i.e., a blótveizla. The manager and agent of this feast was thus the earl himself. Other skaldic poems may likewise support Snorri’s information that Norwegian earls and kings took care of the public cult in Trøndelag.191 According to Vellekla 15–16 (c. 990), which was quoted by Snorri, Earl Hákon of Lade also continued to perform cultic duties as his father had done. He thus restored the sanctuaries which had been destroyed by the sons of Eiríkr. A similar cultic role is also applied to King Hákon the Good in Hákonarmál 18 (c. 960). When the king died the skald praised him as follows: “Then it was made known how well that king had protected sanctuaries (þyrmt véum) when all those who have domination and power [the gods] bade Hákon welcome.” It thus seems as if Hákon became an apostate or turned to a tolerant attitude towards paganism. This poem too was quoted by Snorri.192 Snorri thus had other sources, besides Kormákr’s stanza, when writing about the pagan cult in Trøndelag. Some of his sources consisted of older prose traditions. Ágrip (1190 AD), for instance, tells us that the Þrœndir in Mære required King Hákon the Good “to sacrifice as other kings used to do or we throw you out of the country, if you do not follow us in this thing”.193 The story about the horse liver also occurs in this text, but in another version and in much sparser terms.194 Fagrskinna, written sometime between Ágrip and Heimskringla, also reproduces this tradition briefly. At the assembly in Mære “the people of

190 Norwegian: “Hvilken ætling av agefylte (gudfryktige) menn/ vil vel strides med (sette seg opp mot) templets herre?” In the first half-stanza, Sigurðr’s generosity at feasts is praised and the stanza thus indicates a pre-Christian tradition of ceremonial banquets in Lade led by Earl Sigurðr. 191 See also Jón Hnefill Aðalsteinsson 1998, 62ff 192 Heimskringla I, Ísl. Fornr. 26, 193–197. 193 . . . báðu hann blóta sem aðra konunga í Nóregi, ‘ella rekum vér þik af ríki, nema þú gerir nekkvern hlut í samþykki eptir oss.’ Ágrip, Ísl. Fornr. 29, 8. 194 . . . at hann biti á hrosslifr, ok svá, at hann brá dúki umb ok beit eigi bera, en blótaði eigi ǫðruvís. Ágrip, Ísl. Fornr. 29, 8. 140 chapter 4

Trøndelag gave him [King Hákon] two things to choose between. Either he should sacrifice as kings used to do before and in such a way maintain old law/custom for good crops and peace, or else they would throw him out of the country”; “because of this threat and out of love for his friends the king yielded to their demands and sacrificed.”195 Compared to Snorri’s version, Fagrskinna presents the events at Mære as more of matter-of-fact statements. The author of Fagrskinna has a sober style and avoids much of the wonders and the supernatural elements that Snorri was so fond of.196 Several other Old Norse texts report that both Lade and Mære had well- known pre-Christian hof sanctuaries during the Viking Age.197 According to them, these buildings seem to have had a function and significance similar to the hof sanctuaries in Snorri’s accounts. Most likely it was widely known knowledge among learned medieval Icelanders that Lade and Mære were pre- Christian cultic sites.

Snorri Sturluson’s Hákonar saga góða and other medieval prose thus indicate that Lade and Mære were important pre-Christian cultic sites with ceremonial buildings. Snorri’s description of these sites and the cultic activities that took place there has been partly distorted by Judaeo-Christian elements. In my opinion, Snorri had some ancient traditions available when describing these sites, primarily from skaldic poetry, but also from prose traditions. Archaeological sources also indicate that some of his information is quite sound (see below). Therefore, we should not discard all of his information about ceremonial buildings. We will return to detail in Snorri’s accounts and other texts below, in the thematic analysis.

4.4.3 Ceremonial Buildings and Sanctuaries Attested by Archaeology Since so many written medieval sources mention the ceremonial building at the ruler site of Lade, one might expect this place to display a lot of archaeological finds. There are, however, few finds there indicating cult and power.

195 . . . ok á því þingi gørðu Þrœndir konunginum tvá kosti, at hann skyldi blóta eptir vanða enna fyrri konunga ok fylla svá en fornu lǫg til árs ok friðar, elligar mundu þeir reka hann af ríkinu, ef hann vildi . . ., þá gørði hann eptir bœn þeira ok blótaði. Fagrskinna, Ísl. Fornr. 29, 80. 196 The diplomatic role of Earl Sigurðr in Hákonar saga góða does not appear in the two older prose versions and may be seen as a supplement by Snorri himself. Perhaps Snorri’s positive description of the earls was connected to his friendship with Earl Skúli (of Lade), whom he met in Tønsberg sometime around 1220. Steinsland 2000, 112. 197 See e.g., Landnámabók S 297, H 258, Ísl. Fornr. 1, 307f.; Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 193 (AM 310 4to) and 278ff.; Brennu-Njáls saga 87, Ísl. Fornr. 12, 210. Ceremonial Buildings And Sanctuaries 141

Apart from some Late Iron Age weapons there is not much indicating high status at this site.198 Gerhard Schøning’s and Klüwer’s reports from the eighteenth and nineteenth centuries indicate that ancient grave mounds and monuments were situated there when they visited the place (see ch. 2 above). These graves and monuments are not visible today. Anyhow, the written sources and the fact that Lade in the Middle Ages was regarded as crown land support the notion that the Hlaðajarlar really controlled a public sanctuary at this site. The fact that a fylkiskirkja for Strindafylki was erected at this site in the eleventh or the twelfth century also supports this conclusion, since cultic place continuity is attested at other places in Trøndelag (see below). Mære in Inn-Trøndelag is mentioned as a pre-Christian cultic site several times by Snorri and in other Old Norse sources.199 The important text Landnámabók, for instance, which may have been composed as early as the twelfth century, describes a Viking Age cultic leader (hofgoði) and his ceremonial building (hof ) situated in Mære.200 Not only written sources but also archaeological finds produce good evidence of a pre-Christian ceremonial building at Mære. In 1966–1967 archaeological excavations were conducted under the floor of Mære church. These excavations were led by Hans-Emil Lidén.201 The church, which is dated to the last part of the twelfth century, is placed on the top of a hill, east of the Borgenfjord, called “Mæreshaugen” (fig. 8a). In the prehistoric period Mæreshaugen was an island in the fjord, which completely surrounded Inderøya.202 Under the stone church, vestiges of an early wooden church were discovered. It was surrounded by a churchyard, with about sixty graves. At the church the excavators also found remains of two pre- Christian buildings. The oldest one was a building with wattle-and-daub walls and dated to the Migration Period. The function and character of this building is uncertain. The later one was a Viking Age building with huge post-holes. The form of this house, its dimensions and orientation are unknown, however. The only traces of it were remains of an assumed “palisade wall”, lumps of burnt clay with twigs and branches, fire-cracked stones and post-holes with remains of timber. The post-holes formed irregular circles with diameters varying from 80 to 100 cm. In the post-holes some 19 (or 23)203 gold foil figures were found, which probably should be dated to the Early Viking Age (fig. 8b). Most of them

198 See Sandnes 2001; Brendalsmo 2006, 523f. 199 See e.g., Heimskringla I, Ísl. Fornr. 26, 317f. 200 Landnámabók S 297, H 258, Ísl. Fornr. 1, 307f 201 See Lidén 1969; 1996; 1999. 202 Røskaft 2003, 140. 203 See Lidén 1999, 47, note 17. 142 chapter 4

Figure 8a The church of Mære, on the top of a hill, east of the Borgenfjord, called “Mæreshaugen” in Trøndelag. Photo: Olof Sundqvist.

Figure 8b Gold foils from Mære, Norway.

Figure 8c Gold foil from Borg, Lofoten, Norway. Ceremonial Buildings And Sanctuaries 143 were related to a structure which was interpreted as the place of the high-seat.204 On these figures a couple was depicted, standing against each other, a man and a woman. Whether these symbols shall be regarded as sacrificial objects or some kind of mythical insignia for aristocratic families is uncertain (see ch. 11 below). At this site other finds were made which may likewise be associated with pre-Christian ceremonial feasts, such as pieces of pottery and glass, quantities of animal bones and some characteristic cooking stones. Place-names may also support the idea that Mære was an important pagan cultic site in Inn-Trøndelag. Close to the church of Mære there is a place called Hælgjin, ‘the holy meadow’ or rather ‘the holy meadows’.205 The composite name Mære (ON Mærin) has the same last element vin ‘meadow’ as Hælgjin, while the first element mærr, according to one interpretation, means ‘famous’, referring to the famous and holy meadow at this site, where specific rituals were performed during the religious gatherings.206 It may therefore be suggested that the Viking Age house at Mære was some type of pre-Christian ceremonial building. Whether this house was a multifunctional hall or a cultic house is almost impossible to say. Lidén suggested that the Viking Age building should be related to the ceremonial buildings called hof in the written sources. As noted above, the term hof in the written sources may have semantic variations in different contexts. Nevertheless, the place indicates high status with the gold finds. The medieval farm was considered as crown land in the Early Middle Ages.207 The finds at Mære also indicate that we have a clear example of cult-place continuity, that is to say, the early Christian church was erected on the pagan cultic site.208 Lidén even suggested that the Viking Age building where the gold foil figures were found was turned into a church after the conversion. Christian graves were found associated with this building, which could not be related to the wooden church. As far as I know, there is little archaeological evidence of other ceremonial buildings in Trøndelag similar to Mære. In the summer of 2010, however, “a well-preserved cultic site” was found at Ranheim in Sør-Trøndelag. According

204 This interpretation was later reconsidered by Lidén (1999, 2, 38f.). He rejected the high- seat interpretation. 205 Dillmann 1997, 71, not 43. Critically considered by Sandnes 1987, 151. 206 Dillmann 1997, 63f. The adjective mærr means ‘fræg [berømt], gjæv, gild’. See Heggstad et al. 1993. Sandnes & Stemshaug (2007) state that the first element is uncertain. It could be an ON adjective meaning ‘herleg, gjæv’ or perhaps a noun mæri meaning ‘grense’. 207 Røskaft 2003, 143f. See ch. 5 below. 208 Whether this pre-Christian ceremonial building was used as a church after conversion is debated. Cf. Lidén 1969, 20ff., 30ff.; 1996; Olsen 1969, 26f. 144 chapter 4 to the excavator, Preben Rønne, this site consisted of a flat, roughly circular stone cairn about 15 metres in diameter and 1 metre high.209 This cairn was interpreted as something equivalent to an Old Norse hǫrgr and dated to c. AD 400. Just beside the cairn was an almost rectangular building with strong foundations, and a processional avenue marked by two stone rows. It was argued that this house was an Old Norse hof sanctuary and dated to the period AD 895 to 990. The posts belonging to the house had been pulled out and all of the wood removed, and the cairn had been carefully covered with stones and clay. Afterwards, the whole site had been entirely covered with earth, presumably when the people there converted to Christianity.210 One very early hall building has been discovered at Hovde, Ørland, in Sør- Trøndelag.211 This place is located strategically at the strait between Ørland and Agdenes, i.e., the entrance to the Trondheimsfjord. Around this strait there are many Iron Age burials. The three-ailed hall was found associated with several other buildings during excavations in the summer of 1997. Two phases of settlement were identified at this site. The first was dated to Pre-Roman Period and displayed a farm consisting of three long-houses. House D was 27.9 m long. It was distinguished from the other houses, since it was built of very large posts, up to 50 cm in diameter, and it had a clay floor. It had several fireplaces stretching over the whole length of the building and no stable for the animals was found. Because of these features the house was interpreted as a hall, i.e., a building used for extraordinary purposes. All three houses (C, D and E) were enclosed by a ditch and a fence. In the next phase, dated to Roman Iron Age and Migration Period, it seems as if House A took over the functions of House D from the previous phase. The houses in the second phase were also enclosed, this time by a fence consisting of a row of parallel posts. At these houses too, several fireplaces and cooking pits were found. One fireplace was dated to the period AD 390–550, perhaps indicating when the farm was abandoned. The long continuity of the farm and houses with certain functions indicates that this place was once very special. North-east of Hovde is a place called Hov and another place called Balsnes, perhaps ‘the isthmus of Baldr’.212 The former name indicates that pagan cult took place in this area. Middle and Late Iron Age houses have also been found at other places close to Hovde. The medieval church,

209 Rønne 2011. 210 Since the interpretation of this place still is under debate, I will treat this cultic site with care. Personal communication from the archaeologist Sæbjørg Walaker Nordeide, April 2012. 211 Grönnesby 1999; 2000. 212 Balsnes was written Baldrsnes in 1342. See Sandnes & Stemshaug 2007, 83. Ceremonial Buildings And Sanctuaries 145 for instance, was erected at Viklem. About eighty post-holes were found there and forty cooking pits and hearths, dating back to the period between AD 400 and 800. The post-holes were remains of long-houses, but also a pit-house.213 Beside the church there is also a Late Iron Age royal burial mound. Most likely pre-Christian ceremonies took place at this site. Interesting Middle and Late Iron Age long-houses have also been found at other places in Trøndelag. Beside the three royal burial mounds at Bertnem in Overhalla, Nord-Trøndelag, some type of long-houses were erected (see fig. 27C). These houses are dated to Migration and Merovingian Period, but we do not know much about them.214 Important Late Iron Age houses have also been found at, for instance, Melhus in Gauldal, and at Skei in Steinkjer.215 The most interesting parallel to the ceremonial building in Mære is found at Borg at Vestervågøy, i.e., the second largest of the Lofoten islands. At this place five gold foil figures were found associated with a monumental hall building. This place is also interesting in the context of the Lade Earls, since this family is said to have originated from the north of Norway, more precisely from Hálogaland. In the excavations at Borg in the 1980s and 1999–2002 a settlement was found dated to c. AD 200–1000.216 The investigations were concentrated on a house structure I:1, where a huge house called I:1b (64 m long and 7–8 m wide) was erected during the Migration Period. Around AD 600 an even larger house called I:1a was raised, 83 m long and 7.5–9 m wide. Both houses were three-aisled wooden buildings and both comprised certain rooms which were interpreted as a “hall-room”. A lot of finds were made at structure I:1. Some finds were exclusive and may have indicated cultic actions, such as two pottery jugs and the seventeen or eighteen glass beakers imported from the Continent and the British Isles. Most interesting in the context of the Trøndelag sanctuaries is the five gold foil figures found in the younger “hall-room” (Room C) of Borg I:1a, depicting the motif of a couple (fig. 8c). These objects seem to be related to post-holes and the high-seat area. The assumed high-seat was located in the north corner of the room, where a bronze vessel was also found, as well as the gold pointer, the Tating ware jugs and most of the glasses. In this hall-room (14 m long) there were also some unusual fireplace structures, which may indicate that ritual activities took place there during the Early Viking Age. It was argued that this room had two main functions; during most of the year

213 Berglund 2003, 35. 214 Løken 1992; Berglund 2003, 34. 215 Smedstad 1993, 40ff.; Berglund 2003, 36. 216 See mainly Johanson & Munch 2003; Herschend & Mikkelsen 2003; Munch 2003a and 2003b. 146 chapter 4 it was an ordinary living room, but on special occasions it functioned as a banqueting hall. Around 950–1000 the house was abandoned. It has been supposed that a church was erected at Borg in the twelfth century.217

4.4.3 Hov-, Vang- and Haug-places Several central places and farms (No sentralgårder) in Trøndelag have names containing the element hov (ON hof).218 It has been stated that the word ON hof in Norway originally meant ‘height, hill’ and only later developed the meaning ‘magnate farm’ or ‘ceremonial building’.219 In Norwegian dialects, for instance, it still has the meaning ‘height, small hill, or mound’.220 But in some cases an interpretation ‘elevation, height’ seems to be excluded for topographic reasons, as at Hov in Ørland, Trøndelag.221 In such cases the hov-names may have designated a cultic place or building. Theophoric place-names in other parts of Norway testify that hov could have had a sacred meaning, for instance Frøyshov (‘Freyr’s hof’), Norderhov (*N(j)ærðarhof) (‘*N(j)ærð’s hof’) and Ullinshov (‘Ullinn’s hof’).222 The site designated Hov in Vingrom at Lillehammer, Oppland, is also interesting in this context. During excavations in the 1990s a Late Iron Age house (some 15 metres in length) was discovered and in a test-pit a gold foil figure was found with several strike-a-lights. A recent excavation has revealed additional finds of gold foil figures.223 Totally 29 foils were deposit in connection with four postholes, perhaps traces of a high-seat construction. The house at Hov was most likely a ceremonial building. It was located on a slope overlooking the lake Mjøsa. Thus we may conclude that hof/hov names in Norway could sometimes be referred to as cultic. There are no compound hof/hov names with theophoric first elements in Trøndelag. This does not rule out that hof/hov could be interpreted in a cultic sense there. In some parts of Ut-Trøndelag we may see a common pattern where the hov-farms constituted the very core of the ancient settlement districts.224 According to Merete Røskaft, the word hof/hov in place-names was

217 Elsewhere in Norway, Iron Age halls have also been found at Borg, Forsand in Rogaland, see Løken 2001; and at Huseby, Tjølling in Vestfold. See Skre 2007a. 218 See M. Olsen 1926, 230f.; Røskaft 2003, 56; Sandnes & Stemshaug 2007, 225. 219 See e.g., T. Andersson 1986; Sandnes 1987; 1992; Brink 1996a; Vikstrand 2001, 253; de Vries 1977; SOL, 131f. 220 Sandnes & Stemshaug 2007, 225. 221 Vikstrand 2002, 132. 222 See Helleland 1996, 136f.; Sandnes & Stemshaug 2007, 225. 223 See Haraldsen 1998; Lidén 1999, 43; L. Larsson 2011, 201. 224 Røskaft 2003, 53–76, 210. Ceremonial Buildings And Sanctuaries 147 used during the entire Late Iron Age in this area. Originally it had the plain meaning ‘farm’, but during the Viking Age it seems as if it gained the meaning ‘the farm par excellence’ or ‘the farm where the banqueting hall was located’.225 In my opinion, this term could also be interpreted ‘ceremonial building’. The place-name Vang appears in Trøndelag. It has also been interpreted as cultic. Magnus Olsen, for instance, has suggested that a special kind of meadow (vangr) often was located beside the hof sanctuaries in Norway as an equivalent to the kirkevangen, i.e., the meadow where the people gathered beside the church for different kinds of activities.226 This interpretation has more recently been accepted.227 Place-names, including the term haugr, have also been referred to as “cultic” in Inn-Trøndelag. According to Magnus Olsen, they designated central farms with ancient burial mounds, where the religious and political communities regularly gathered.228 It is uncertain, however, whether pre-Christian ceremonial buildings were raised at these sites. During the Middle Ages fylkiskirkjur were erected there.

The best evidence of the combination of ceremonial buildings/sanctuaries and rulers in Trøndelag may be seen at Lade, and particularly Mære. These two places, together with Borg at Vestervågøy, will play a significant role in the following discussion. Whether the building at Mære should be classified as a banqueting hall or cultic house is uncertain. The building at Borg is one of the most prominent halls found in all Scandinavia.

4.5 Hof Buildings in Iceland

When historians of religions229 have discussed pre-Christian ceremonial buildings in Iceland, they have often taken their point of departure from Eyrbyggja saga and the story of Þórólfr Mostrarskegg Ǫrnólfsson and his sanctuary (hof ) at Hofstaðir in western Iceland:

Þórólfr established a great farm at Hofsvágr which he called Hofstaðir, and had a large hof built there with its door in one of the side walls near

225 Røskaft 2003, 53–76, 210. 226 M. Olsen 1926, 218. 227 Sandnes 1992, 15f.; Sandnes & Stemshaug 2007, 479. 228 M. Olsen 1926, 262ff. 229 See e.g., Gunnell 2001 and Sundqvist 2005c; 2009a. 148 chapter 4

the gable. Just inside the door stood the high-seat pillars with the so- called holy nails fixed in them, and beyond that point the whole building was considered a sanctuary. Inside the main hof was a structure built much like the choir in churches nowadays, and in the middle a raised platform (stalli) like an altar. On this platform lay a solid ring weighing twenty ounces, upon which people had to swear all their oaths. It was the business of the hofgoði to wear this ring on his arm at every public meeting. There was a sacrificial bowl on the platform (stalli) too, with a sacrificial twig shaped like a priest’s aspergillum for the blood of animals killed as offerings to the gods to be sprinkled from the bowl. This blood was called hlaut. Inside the choir-like part of the building (afhús) the figures of gods were arranged in a circle right round the platform (stalli). Every farmer had to pay tax to the hof. Another of their duties was to support the hofgoði in his mission, just as farmers nowadays have to support their chieftains. It was the goði’s business to see to the hof and maintain it properly at his own expense, as well as hold sacrificial feasts.230

It has been debated whether this description of Þórólfr’s hof should be considered reliable source for the history of religions.231 It seems as if some information in this account does not carry an independent source value and may be dependent on Snorri Sturluson’s description of ceremonial buildings in Trøndelag in Hákonar saga góða.232 Several scholars have also argued that the author (or editor) of the saga used concepts which most likely must be regarded as late constructions. The sacrificial terms hlaut and hlautteinn in the sense of

230 Hann setti bœ mikinn við Hofsvág, er hann kallaði á Hofsstǫðum. Þar lét hann reisa hof, ok var þat mikit hús; váru dyrr á hliðvegginum ok nær ǫðrum endanum; þar fyrir innan stóðu ǫndvegissúlurnar, ok váru þar í naglar; þeir hétu reginnaglar; þar var allt friðarstaðr fyrir innan. Innar af hofinu var hús í þá líking, sem nú er sǫnghús í kirkjum, ok stóð þar stalli á miðju gólfinu sem altari, ok lá þar á hringr einn mótlauss, tvítøgeyringr, ok skyldi þar at sverja eiða alla; þann hring skyldi hofgoði hafa á hendi sér til allra mannfunda. Á stallanum skyldi ok standa hlautbolli, ok þar í hlautteinn sem stǫkkull væri, ok skyldi þar støkkva með ór bollanum blóði því, er hlaut var kallat; þat var þess konar blóð, er svœfð váru þau kvikendi, er goðunum var fórnat. Umhverfis stallann var goðunum skipat í afhúsinu. Til hofsins skyldu allir menn tolla gjalda ok vera skyldir hofgoðanum til allra ferða, sem nú eru þingmenn hǫfðingjum, en goði skyldi hofi upp halda af sjálfs síns kostnaði, svá at eigi rénaði, ok hafa inni blótveizlur. Eyrbyggja saga 4, Ísl. Fornr. 4, 8f. Trans. Hermann Pálsson & P. Edwards (slightly modified). 231 See e.g., Hultgård 1993, 228ff., 232ff.; 1996, 36ff.; Perkins 1994, 56f.; Jón Hnefill Aðalsteinsson 1998, 35ff. 232 See e.g., Düwel 1985, 10f.; Hultgård 1993, 228; Perkins 1994, 53. Ceremonial Buildings And Sanctuaries 149

‘sacrificial blood’ and ‘sacrificial twig’, for instance, cannot be found in Viking Age terminology. In the pre-Christian vocabulary these two terms referred to ‘lot’ or ‘lot twig’. They thus passed through a semantic development and changed their meaning sometime during the Late Viking Age and Early Middle Ages (see above and ch. 9 below).233 Nor has the description of the ceremonial building and its interior been considered reliable. The author adds details in his description, which in a striking way resembled the structure and ritual objects of a church.234 According to Olaf Olsen, for instance, the medieval author of Eyrbyggja saga had the Christian church in mind when describing the hof at Þórsnes.235 When he described the afhús, he added “much like the choir in churches nowadays” (sem nú er sǫnghús í kirkjum). This comment, according to Olsen, proved that the scribe actually was describing a church in this passage. The criticism against Eyrbyggja saga’s description of Þórólfr’s ceremonial building with all its details is partly sound. But it should also be stated that the archaeological evidence of ceremonial buildings today is completely different from the situation in the 1960s, when Olsen wrote his book. Since the archaeological knowledge about ceremonial buildings has increased dramati- cally during the last few decades, we must also reinterpret and re-evaluate the testimony about such buildings appearing in the sagas. New archaeological evidence deriving from several places in Scandinavia indicates, for instance, that choir-like rooms or separate buildings connected to or situated beside the feasting halls may have functioned as the sanctuary part of these houses (see above). This could definitely be associated with the afhús mentioned in Eyrbyggja saga.236 Archaeologists have also recently shown that a small cultic house was erected at Hólmur in south-eastern Iceland (see below). It should also be emphasized that the author of Eyrbyggja saga did not invent the story of Þórólfr by himself. Most likely he proceeded from older traditions when writing his own text. Such traditions appear in Landnámabók (probably compiled as early as the twelfth century), where the basic features of this story may be found.237 Þórólfr’s ceremonial building is mentioned here too: “There he [Þórólfr] erected his home and built a large ceremonial building

233 Düwel 1985, 25ff.; Hultgård 1993, 230. 234 O. Olsen 1966, 26, 187. 235 O. Olsen 1966, 26, 187ff. Cf. Perkins 1994, 57. 236 See G. Andersson et al. 2004, 14f. It is interesting to note that the term afhús, according to ONP I, 104 (A Dictionary of Old Norse Prose), means ‘separate/remote building/room, outhouse, side-room’. 237 See Landnámabók S85, H73, Ísl. Fornr. 1, 124–125. 150 chapter 4

(hof mikit), and consecrated it to Þórr, and now the place is called Hofstaðir.”238 This indicates that some traditions about Þórólfr and his ceremonial building existed in Iceland at least as early as the second half of the twelfth century, and probably also earlier, since Þórólfr is mentioned as a landnámsmaðr in Ari’s Íslendingabók (c. 1120–32).239 There are some descriptions of Icelandic hof buildings in the Old Norse prose traditions, which may have inspired the author of Eyrbyggja saga when describing Þórólfr’s hof. The most important text is probably the one found in Úlfljótslǫg, which, for instance has been preserved in the Hauksbók redaction of Landnámabók (the text is presented in ch. 5 below). This description of the interior of the hof partly resembles the account on the hof at Þórsnes in Eyrbyggja saga, where we also hear about a ring placed on an altar (stalli). The connection between the goði and the hof also appears in these both texts. In addition, the financial system of the sanctuaries (gefa toll til hofs) mentioned in Úlfljótslǫg harmonizes with the information found in Eyrbyggja saga, where we hear that “every farmer had to pay tax to the hof”. Olaf Olsen was, however, sceptical of the historical value of Úlfljótslǫg as well. According to him, this text is only a learned medieval construction of the pre-Christian Law, cult and society.240 In my opinion, there are some elements connected to the description of the ceremonial building in this text and in Eyrbyggja saga which may be based on old oral traditions. I will come back to them in the thematic discussion below and evaluate some details in these texts, but I will show here that archaeological evidence and place-names indicate that the term hof in Iceland really may have designated a quite large building where public religious rituals took place.

4.5.1 The Feasting Hall at Hofstaðir, Mývatnssveit—Archaeological Evidence According to Eyrbyggja saga, Þórólfr’s hof at Þórsnes was a large house (mikit hús) with a door in one of the side walls near the gable. Other descriptions of hof buildings in the Sagas of Icelanders report that such sanctuaries were sometimes large and located on the chieftains’ farms (see ch. 6 below). Archaeology indicates that the medieval scribes were well-informed when describing the size of pre-Christian hof buildings at chieftain farms in Iceland. At the chieftain farm of Hofstaðir, Mývatnssveit, in north-eastern Iceland, a

238 þar reisti hann bæ sinn ok gerði þar hof mikit ok helgaði Þór; þar heita nú Hofstaðir. See Landnámabók S85, H73, Ísl. Fornr. 1, 125. 239 See Íslendingabók, Ísl. Fornr. 1, 10. 240 O. Olsen 1966. See also ch. 5 below. Ceremonial Buildings And Sanctuaries 151

Viking Age hall building was found in 1908 by the archaeologist Daniel Bruun (with Finnur Jónsson).241 It measured 36 metres long and it was 8 metres wide. Since this large building was located at the place called Hofstaðir ‘the hof place’ it was interpreted as a hof, i.e., a ‘ceremonial building’.242 It was argued that this house resembled the large ceremonial buildings called hof in the Sagas of Icelanders, e.g., Þórólfr’s hof mentioned in Eyrbyggja saga. Not only the name but also the large size and the structure of the house indicated that it was a hof, i.e., a “pagan temple”. The hall was divided into two parts. There was first a banqueting room, marked AB. There was also a separate room at the north end of the hall (marked C). It was interpreted by Daniel Bruun and Finnur Jónsson as a shrine or the inner sanctuary, where the images of deities were kept (afhús). There were also two further rooms at the western side of the building, marked D and E. In addition there was a strange pit or orbital ruin (marked G), outside the hall to the south, full of ash and animal bones. It was interpreted as a place where rubbish was placed after the ceremonial feasts in the “temple”. For several decades there was general consensus in research that the building at Hofstaðir should be interpreted as a pre-Christian Scandinavian “temple”.243 In the middle of the twentieth century, however, some doubts were expressed about this interpretation. It was argued that this settlement differed little in form from other Iron Age dwelling houses and farmsteads in Scandinavia.244 These doubts were even more emphasized by Olaf Olsen, when he re-excavated the pit (G), to the south of the hall in 1965.245 He argued that Hofstaðir was a quite ordinary chieftain farm. The hall there was sometimes used as a place for ceremonial banquets, but mostly it was employed for other more ordinary and profane activities. The pit (G) was interpreted as a cooking pit, with a size of 4.60–5.40 metres. It was used for ritual cooking in connection with religious feasts and gatherings. Olsen rejected the idea that the room at the north end of the hall (C) could be interpreted as an “inner sanctuary” (afhús). He argued that the choir-like structure at the northern gable had a door only on the outside, and thus could not be compared with the afhús mentioned in Eyrbyggja saga.

241 Bruun & Finnur Jónsson 1909. 242 Cf. Orri Vésteinsson 2007. 243 See e.g., Thümmel 1909; M. Olsen 1926; Ohlmarks 1936; Gehl 1941; de Vries 1956–57, §269. 244 See Rousell 1943, 220f. 245 O. Olsen 1966. 152 chapter 4

Results from the new excavations, conducted between 1991 and 2002,246 indicate that the adjacent room (C2), built in phase II (c. 980–1030 AD), also had a connection directly into the hall,247 and thus could be interpreted as some kind of an afhús. According to Gavin Lucas, the function of this room is very unclear, because of the context of the finds.248 Only a polished bone pin was retrieved there. This room/house was rebuilt during phase V (1300–1477) called C1. Its function is likewise unknown, but its massive design indicates something unusual.249 During the new excavation, a small “hall” (D1) was discovered in the south- western part of the hall.250 This “hall” was built in phase II (c. 980–1030), and probably had a direct passage into the major hall (AB). The interior of this structure covers an area of 8 × 3.2 m and its roof was supported by paired posts. The room also had a hearth, which suggests that people spent some time there. It was interpreted as a domestic room. The new excavation indicates that the large “bowed, aisled hall” (AB) was erected in phase I, c. 940 and abandoned in phase II, around 1030.251 This hall building, in Scandinavian style, was divided into three rooms. The northern room had two great cooking pits, which had been cleaned out repeatedly. Most likely food processing occurred in this room (see further ch. 6 below). The central room seems to have been used for sleeping, eating, and other activities. Perhaps some small craft activities took place at the southern end of this room, since many of the beads were found in this area. The central hearth in this room was used for warmth and cooking. Most likely the high-seat was located opposite this hearth. This must have been the most important space for ceremonial feasting and could perhaps be described as the “hall-room”. The southern room of the hall building may have been used as food storage. The new archaeological analysis of the Hofstaðir bone material has generated some sensational results, which also have relevance for a cultic interpretation of the place. A minimum of 23 cattle skulls were retrieved in two clusters from outside the walls of the hall at structure A2 and D (the minor hall).252 These skulls show evidence of specialized butchery and prolonged

246 On the new excavations, see e.g., Adolf Friðríksson & Orri Vésteinsson 1997; Orri Vésteinsson 2001; 2007; Lucas & McGovern 2008; Lucas 2009. 247 See Lucas 2009, 124. 248 See Lucas 2009, 125ff., 388, 391. 249 See Lucas 2009, 155. 250 See Lucas 2009, 131ff., 388, 391. 251 See Lucas 2009, 62ff., 112ff., 373ff., 386ff. 252 See Lucas 2009, 236ff. Ceremonial Buildings And Sanctuaries 153 display on the outside of a structure. Differential weathering indicates that these specimens were displayed face outwards, and that they remained exposed to weathering for months or years after the soft tissue had decayed. Two skulls were found in the wall demolition deposits along the exterior walls of the great hall (AB). They were found lying face down, approximately 75–80 cm apart. Their location and wear suggest that the skulls originally were placed at short intervals on the turf walls or roof along the entire length of the big hall. They were thus exposed and visible to people approaching the hall at least during certain periods of the years, most likely spring and summer. Another one was recovered from the wall demolition in area E, perhaps indicating the same original setting. The two “skull dumps” or clusters at A2 and D appear to be the deposit of skulls collected when the demolition of the great hall took place (in phase III, i.e., 1030–1070 AD), and thus probably do not provide direct evidence for their original placement. Most likely this collection and deposition of skulls reflect some kind of ritual of abandonment or closure in connection with the Christianization of Iceland. Investigations of the skulls indicate that the latest deaths occurred near the year AD 1000, and that the animals were ritually killed and beheaded in a dramatic context. The investigation indicates that ritual slaughter of cattle in prime age, on a seasonal basis, took place at Hofstaðir, just before or during the conversion of Iceland.253 At about the same time, a complete female sheep was killed by a blow between the eyes and its unbutchered body was left beside the cattle skulls at A2. The unusual deposit is best understood as ritual, rather than in a strictly agricultural context. Perhaps this ritual slaughter of the sheep should be related to the abandonment of the aisled hall during phase III, i.e., contemporary with the gathering and deposit of the skulls at this structure. The pit to the south of the hall has been reinterpreted as a pit-house (G) during the new excavations. This house also seems to be older than the hall and is dated to the late ninth century AD.254 The pit-house was probably used for textile production and in addition it perhaps had a domestic function. There was also a smithy (A4, A5) during phase I–II, located at the south-eastern side of the hall. The presence of small amounts of metalworking debris from the floor of these structures indicates this.255 According to the archaeologists, the farmstead and the hall of Hofstaðir had many functions. Most likely it had a political significance, since its

253 See mainly Lucas & McGovern 2008. 254 Lucas 2009, 93ff., 140, 388, 393. See also Adolf Friðríksson & Orri Vésteinsson 1997; Bjarni F. Einarsson 2008, 152. 255 See Lucas 2009, 127ff., 388, 391; McDonnel & Maclean 2009 (in Lucas 2009), 271ff. 154 chapter 4 central location and the monumental size of the house (according to the recent excavation 38 metres long) probably functioned as manifestations of power.256 Perhaps the bloody slaughter, the ritual decapitation and display of the cattle (bull) heads had a similar function. Thus we can interpret it as a chieftain settlement which also had public cultic functions, as e.g., at Helgö or Borg in Lofoten. Activities of an everyday character were also carried out at Hofstaðir. The household economy was based on the production of piglets, goat-milk products, and young sheep in preference to wool. It seems as if desirable food was produced there. Also forging and other crafts took place there. Most likely, it was only in connection with the seasonal ritual feasts that the hall was transformed into a sanctuary and perhaps also occasional dwelling space for the people who gathered at Hofstaðir to take part in the religious ceremonies. Hofstaðir was probably an important chieftain site in the society of the Mývatn region, but it was not the first farm settled there. Sites such as Sveigakot and Hrísheimar were settled as early as the last quarter of the ninth century. Hofstaðir’s significance only starts around 940. Perhaps the farm was established there in order to control the flow of coastal resources into the Mývatn region. The hall was erected high up on the homefield, rendering it visible in the immediate landscape, thus expressing some kind of symbolic power.257 Perhaps this farm competed with other farms in the region, such as Reykjahlíð, to be the most powerful chieftain farm there. It seems, however, as if Hofstaðir failed in that competition. In one sense Hofstaðir may have been very important during its existence, namely as a cultic site and a place of rich feasting.258 The settlement of Hofstaðir was abandoned in phase III (1030–1070), which is marked by several acts of closure, such as the ritual depositions of skulls and the sheep in A2 and D. A new settlement was established in the south- western area of the levelled farm mound, where a medieval church (eleventh century) was also built. This abandonment of the feasting hall should probably be related to the conversion of Iceland.259 The context of the hall at Hofstaðir differs in some respects from the halls in the Mälaren area and Norway. While the halls in the Mälaren area, such as the one at Helgö, have a long structural continuity, the hall at Hofstaðir

256 Cf. Lucas 2009, 252. The normal house size in Iceland was 12 to 16 metres. See Lucas 2009, 376. 257 Lucas 2009, 402. 258 Lucas 2009, 404ff. 259 See Lucas 2009, 407f. Ceremonial Buildings And Sanctuaries 155 had a short existence. Furthermore, Hofstaðir lacks the monumental graves which are so typical in the Swedish hall context.260 The find context of the hall building at Hofstaðir also differs from the halls and cultic buildings found in the Mälaren area and Norway. The status of the material culture is poor at Hofstaðir compared to the halls in these latter areas. Hofstaðir lacks the dis- tinctive luxury features that are so typical in the contexts of the halls at e.g., Helgö and Borg in Lofoten, such as gold foil figures, imported glass or ceramic table vessels. There are few precious metals other than two silver items found there. Also, the small size of the hearth at Hofstaðir has been mentioned in this context.261 As noted by the archaeologist Orri Vésteinsson, the slow development of centralized power in Iceland may partly be supported by archaeological evidence. “The earliest sites, farmsteads and burials, are all characterised by material poverty and there were no monumental grave mounds or other such symbolism to indicate overlordship.”262 In the context of Icelandic society, however, the hall of Hofstaðir seems to have been in a class of its own. Early long-houses and halls, such as the ones found at Aðalstræti, Granastaðir, Grelutóttir, Eiriksstaðir, Vatnsfjörður, and Havitárholt measure between 60 and 120 square metres, while Hofstaðir is about 270 square metres. In my opinion, this indicates that those who built this hall were chieftains with great ambitions.

4.5.2 More Specific Cultic Houses in Iceland Different types of sources concerning Icelandic conditions thus produce the image that hof referred to a multifunctional hall at the chieftain’s residence. Other sources indicate, however, that the term hof also designates a more specific ceremonial building, one whose location is detached from the chieftain’s farm, his hall and the major settlements.263 For instance, Hrafnkell Freysgoði’s ceremonial building (called both hof and goðahús) in Hrafnkels saga Freysgoða was isolated at distance from Hrafnkell’s farm, below the banks towards the river and above the cliff called Freyfaxahamarr.264 The images of the deities were kept in this house. There is no mention, however, that banquets took place there. According to Víga-Glúms saga, a similar situation existed at Þverá, Eyjafjörður, northern Iceland. The farm of the chieftain (goði) at Þverá was

260 Cf. Lucas 2009, 400. 261 Lucas & McGovern 2008, 20. 262 Orri Vésteinsson 2001, 339. 263 Cf. Vikstrand 2001, 266 264 Hrafnkels saga Freysgoða, Ísl. Fornr. 11, 99, 124; cf. (Ed) Jón Helgason 1968, 2f., 28f 156 chapter 4 located at some distance from “Freyr’s sanctuary” (hof Freys) at Hripkelsstaðir.265 Nothing indicates that this house should be interpreted as a multifunctional hall or festival building. Instead, the dedication of the sanctuary to Freyr intimates that the author interpreted it as a specific religious building. When Víga-Glúmr visited the hof at Djúpadalr in order to swear an oath that he was innocent of Þórvaldr’s death, Þórarinn arrived with a hundred men in order to check that everything was done in a decent manner. Yet only five men followed Víga-Glúmr inside the hof.266 This indicates that the author thought that this cultic house was rather small, and had a different character from the banqueting halls.267 In the Sagas of Icelanders there are some other terms beside the concepts hof and goðahús, which also refer to ceremonial buildings, such as the term blóthús.268 If this compound word is genuine it strongly indicates that more exclusive religious buildings existed in pre-Christian Iceland. It should be noted that the term hús refers to both ‘house’ and ‘room’ and thus could designate a specific part of the multifunctional hall building. The context of narrative in the sagas indicates, however, that at least the medieval authors believed that the term blóthús sometimes referred to a separate building.269 Vatnsdœla saga 26 mentions that the sons of Ingimundr visited Hrolleifr in the middle of the night. When they arrived at Hrolleifr’s farm at Áss they saw “firewood piled against the wall on both sides of the gable. They also saw a little hut standing in front of the door, and a gap between it and the door of the main building.” Þorsteinn said. “That must be the blóthús” (Þetta mun vera blóthús).270 This story arouses strong associations with the ritual structure of ceremonial buildings e.g., at Tissø at Zealand and Järrestad in Scania, where we have small cultic houses beside the main buildings, i.e., the halls. One hesitates, however, when reading the scribe’s description of this building. It seems as if he/she projected his/her own opinion about private farm churches back to pre-Christian times.271 Anyhow, it is possible that the compound blóthús is old.

265 Víga-Glúms saga, Ísl. Fornr. 9, 34 266 See Ísl. Fornr. 9, 85f. 267 In Continental Germanic sources there is evidence of small and more specific ceremonial buildings. In Indiculus superstitionum et paganiarum (MGH LL III II:1, 222f.), dated to the eighth century, a kind of small houses are mentioned which were also regarded as sanctuaries (De casulis id est fanis). This information appears in a list of examples describing non-Christian customs and phenomena. Sundqvist 2005c. 268 Grønbech 1997 II (1909–12), 136f.; O. Olsen 1966, 111. 269 Fritzner 1954. 270 Vatnsdæla saga, Ísl. Fornr. 8, 68f. See also Droplaugarsona saga 4, Ísl. Fornr. 11, 146. 271 See Grønbech 1997 II (1909–12), 137; O. Olsen 1966, 111f. Ceremonial Buildings And Sanctuaries 157

There are also archaeological indications of more specific and small pre- Christian cultic houses in Iceland. At Hólmur close to Hornafjörður (Höfn) in south-east Iceland, a specific cultic house (“blót house”) was discovered by excavations in the summers of 1997 and 1999 close to a Viking Age burial situated on a mound or a hill (15 × 25 m. and approx. 2 m high).272 Hólmur seems to have been a Viking Age farmstead and the cult house and the grave were located on the small mound about 250 metres north-east of this settlement in an exposed position. The house is small (2 × 2.16 metres) and has a NW–SW orientation with its entrance facing north-west. It had two roof-supporting posts and there was an oven in the south-east corner. The house was sunken and must be considered as a typical pit-house. Several objects were found inside the house, such as two pieces of whalebone vertebrae with cutting abrasions. On the floor in front of the oven, many objects were found, for instance, a worked disc-shaped piece of soft rock (round disc). In addition to this, charcoal, cremated bones and fire-cracked stones were also frequently discovered. All finds are quite common at settlements in Iceland, except the round stone disc. Around the house there was an area containing a vast amount of fire- cracked stones, which were related to ritual activities. Besides the grave, several post-holes were also found outside the house. A couple of them seem to have formed a kind of entrance or gate to the mound. Crude iron was placed in the post-holes. Hearths and cooking pits were also found outside the house. Certain objects were discovered in them, such as gaming pieces, loom weights, burnt and unburnt bones and fire-cracked stones. The cult house was established during the first phase of the landnám period, c. 870–880, and it was used during the entire Viking Age.

4.5.3 Hof Places in Iceland—The Evidence of Place-names Place-names which incorporate the names of the Norse deities as the first element, as well as names containing the terms hof and hǫrgr, may indicate pre-Christian sacred places in Iceland. Most interesting are the place-names containing the word hof and the compound hofstaðir. These names are extremely richly represented in Iceland. Svavar Sigmundsson argued that the evidence of pre-Christian activities at these sites is specifically strong when churches were later erected there.273 Per Vikstrand has noted that the hof names in Iceland often appear in simplex form. According to him, this suggests a more elaborate meaning than just ‘elevation, height’ for these names.

272 See Bjarni F. Einarsson 2008. 273 Svavar Sigmundsson 1992, 247ff. 158 chapter 4

Vikstrand does not rule out that hof may refer to some kind of ceremonial building or a feasting hall in Iceland too.274 On the basis of the finds made at Hofstaðir, Mývatnssveit, Orri Vésteinsson has recently argued that the hall there was built as an attempt to create a political and religious centre based on the resistance to growing Christian influence in Iceland.275 The pre-Christian cult there was maintained until 1030, although the legal decision of conversion was taken in AD 1000 at the Alþingi. By means of a topographical analysis, he argued that the farms with designations containing the element hof in Iceland were never part of the original settlement cluster, but rather should be seen as secondary formations. They were established after the farming communities had been formed. The intention was to create central places in the already existing settlements as resistance nodes against Christianity. According to Orri, the word hof in place-names refers to the feasting hall which was the arena for different types of socio-political activities. The religious, pagan connotation of the term is, however, first seen in the late tenth century, when these halls became the centres of resistance to the new religion.

In the present chapter I have argued that recent discussion of sanctuaries in pre-Christian Scandinavia has mostly concerned the question whether specific cultic buildings (“temples”) existed or not. Admittedly, some archaeologists have also touched upon the functional and ideological dimensions of hall buildings. Remarkably, however, no extensive investigations have yet been attempted by historians of religions, taking both political and religious contexts into consideration, as well as the historical development and regional variations. The present study will therefore establish the ideological and religious contexts of the ceremonial buildings and sanctuaries in Scandinavia during the Late Iron Age. This chapter has also shown that written sources, place-names and archaeological finds indicate that pre-Christian ceremonial buildings existed in all three investigation areas of this study. Most likely there were both multifunc-

274 Vikstrand 2001; 2002, 132. Olaf Olsen also accepted that the hof names in Iceland may have connoted religious aspects. However, this word should not be interpreted as ‘temple’, i.e. a building or place intended exclusively for religious activities: “Ligeledes kunne det påvises, at ordet hof ikke nødvendigvis må oversættes ved ‘temple’, men antagelig kan opfattes som betegnelse for en gård, på hvilken der er foregået kulthandlinger, vel især blótgilder, for en store gruppe mennesker end gardens egne beboer.” O. Olsen 1966, 192. 275 Orri Vésteinsson 2007 Ceremonial Buildings And Sanctuaries 159 tional hall buildings and more specific cultic houses. The latter type is, however, less visible in the archaeological sources from these areas, compared to the situation in the southern parts of Scandinavia. The find contexts of ceremonial buildings in the Mälaren region and Trøndelag are sometimes distinguished by exclusive objects such as gold foil figures and imported glass. In connection with the cultic sites of these areas, monumental graves were sometimes erected (see ch. 12). Compared to this situation, the material context of ceremonial buildings and sanctuaries in Iceland is poor. The archaeological finds made at Hofstaðir nevertheless indicate that those who built this hall either were chieftains or had ambitions to be powerful leaders in the district where they lived.

In Part 2 of the present study, some contextual matters on the issue of the ceremonial buildings and ruler strategies in Late Iron Age Scandinavia have been outlined, such as the sources and their value for historical reconstructions, as well as the regional-contextual approach. In chapter 2 the social and political structure and historical development in the three investigation regions (the Mälaren area, Trøndelag and Iceland), were examined. It can be concluded that a tendency towards centralization of power from the Migration Period onwards in the Mälaren area could be discerned, with kings and earls at the top of society. A similar centralizing tendency was also observable in Norway and Trøndelag from the early Viking Age onwards. As a result of this power concentration involving kings and earls, Norwegian magnates from the lower elite and farmers emigrated from their homeland to Iceland. There the Icelandic Free State was established, which had a decentralized power structure, without royalty. In Iceland power was thus split among many local chieftains, who only had limited possibilities to exploit their subjects. A concentration of power did not appear there until the thirteenth century. Chapter 3 has presented a case study where the religious strategies for political authority and legitimacy of three ruling families were investigated in some written sources. This case study included the royal family named Ynglingar from the Mälaren region, the Earls of Lade from Trøndelag, and the chieftain family named Þórsnesingar from Iceland, i.e., representatives from the three regions addressed in the present monograph. The kings and earls from the Mälaren region and Trøndelag claimed to be of divine origin. This strategy was not seen among the chieftains in Iceland. These chieftains argued, however, that they were the deity Þórr’s dearest friend. The conclusion was that the centralizing tendencies which were observed in the societies of the Mälaren region and Trøndelag had generated a type of monumental and bombastic 160 chapter 4 ruler strategy which was not as visible in the more decentralized Icelandic chieftain society. This conclusion constitutes the fundamental hypothesis of the entire monograph. The hypothesis of the present monograph is that the political-structural differences in the three regions must have had consequences for the formation of ruler and power strategies associated with cultic places and ceremonial buildings. Religious strategies for rulership in Sweden and Norway (among earls and kings) would therefore be different from those found in Iceland (among chieftains). As we shall see below in Part 4, the investigation of the source material shows that this hypothesis can partially be confirmed, but perhaps not as significantly as expected. The similarities between the three regions proved to be striking, especially when it comes to ritual and symbolic strategies. We therefore now turn to these similarities in Part 3. part 3 General Features

⸪

chapter 5 The Religious Leadership of the Sanctuaries and the Economic System

It has been suggested in archaeology that sacrificial practice, cultic leadership and religious organization changed in southern Scandinavia during the Migration Period.1 Collective sacrifices in bogs and lakes outside settlements ceased. A new type of military leader appeared. These leaders “obtained enough power to centre the manifestations and ceremonies upon themselves”.2 The new religious strategies for political authority and legitimacy were inspired by notions and behaviour circulating in Continental Europe. The public cult was tied to the political power governing the society and it was performed at the rulers’ residences. There are several attestations indicating this ritual strategy in Scandinavian contexts.3 In what follows, I will turn to some questions related to the ritual ruler strategies as they appear in connection with ceremonial buildings in the three investigation areas. In this chapter I will mostly concentrate on the religious leadership of public sanctuaries. I will argue that the political leaders on different social levels in all three regions controlled the cultic activities at these sanctuaries. These leaders were considered as the protectors, managers or owners of such places. In Svetjud and Trøndelag kings and earls also performed in similar roles at the central sanctuaries which had a regional or an inter-regional significance, such as Uppsala and Lade. In these areas sanctuaries appear both at aristocratic centres (places controlled by individual rulers) and at centres of the community (i.e. places controlled by a cooperative of chieftains) (see further below and ch. 14).4 In the final parts of this chapter it will be argued that one possible driving force behind the rulers’ and chieftains’ interest in public cult was related to economy. Other forces may be related to these political leaders’ religious duties and roles in cult. Before turning to these issues, however, I will survey previous research related to cultic leadership in ancient Scandinavia.

1 Fabech 1991; 1994. Cf. Näsman 1994. 2 Fabech 1991, 302. Critically considered by T. Zachrisson 1998, 117f.; Hedeager 1999. 3 Golden bracteates (AD 400–500) and especially gold foil figures (AD 500–800), for instance, appear often at places which may be described as rulers’ seats with hall buildings or cultic houses. See further ch. 12 below. 4 Cf. Vikstrand 2001, 410ff.

5.1 The Debate about Cultic Leadership and the Terminology

The issue of cultic leaders in Late Iron Age Scandinavia has rarely been discussed in previous research. Besides surveys in handbooks of ancient Germanic and Scandinavian religion,5 and references in some philological and onomastic studies,6 only a few scholars have exclusively investigated ancient Scandinavian cultic leaders thoroughly.7 Previous discussion has been polar- ized; two lines of interpretation can be discerned, represented by scholars from two different fields of study: philology (onomastics) and the history of religions. Some historians of religions, for instance Folke Ström, argue that the ancient Scandinavians lacked a professional priesthood. According to him, there were no priests who exclusively had the assignment to serve the divine powers.8 The political ruler, the king or chieftain, made contact with the deities at the public sanctuaries on behalf of the people at the sacrificial feasts and in other rituals at the cultic sites. This type of religious leadership has often been associated with the notion of sacral kingship. Historians of religions who investigate the entire Germanic area sometimes argue that priests existed in some parts of it. Jan de Vries, for instance, states that in the southern area profane and sacred leadership functions were separated. The situation in the northern area seems to be different.

Bei den Südgermanen haben die Priester neben der weltlichen Obrigkeit gestanden. [. . .] Die Quellen, die wir für Skandinavien besitzen, scheinen darauf hinzuweisen, daß die Trennung der weltlichen und priesterlichen Funktionen hier nicht, oder jedenfalls nur sehr spät, stattgefunden hat.9

Other scholars think that the Scandinavians had specialized priests. These are mainly philologists, especially specialists in onomastics. Klaus von See argues that the ON term goði refers to an exclusively priestly office. Only in Iceland, in his opinion, where the historical situation was very special, did the goði office develop into a leadership including several functions, such as law and other

5 De Vries 1956–57, §§274–281. 6 E.g., T. Andersson 1992a; Brink 1996a; Vikstrand 2001. 7 E.g., Phillpotts 1912–1913; Wesche 1937; Kuhn 1978, 231–242; Sundqvist 1998; 2003a; 2003b; 2007; 2010b. 8 F. Ström 1985, 72–93. See also Phillpotts 1912–1913; Dumézil 1958; 1959; 1973; 2000; H. E. Davidson 1994 (1993), 68, 88; Hultgård 1997, 19f.; Näsström 2001a, 76. The archaeologist Olaf Olsen (1966, 55) has a similar point of view; see also Kuhn 1978, 242. For the early discussion on this topic, see Phillpotts 1912–1913, 264f. 9 De Vries 1956–57, §278. The Religious Leadership of the Sanctuaries 165

“secular” aspects. Klaus von See also states that originally the Germanic people strictly distinguished between religious and judicial aspects.10 This issue has been debated throughout the twentieth century and beyond. John Kousgård Sørensen states that a priestly class, in Danish “præstestand”, existed in Late Iron Age Scandinavia. He focused on the term OScand ‑vé(r), ‑vi(r), ‑væ(r) (PN *wīhar), which he interpreted as ‘priest’. According to him, compound nouns including this term reflected a differentiated hierarchical priesthood.11 Other specialists on toponomastics have observed terms in the place-name materials which could refer to exclusive religious specialists: e.g., ON goði, *vífill, and perhaps *lytir. In general the specialists in toponymy apply designations such as “priests” (“Priester”),12 “pagan priests”,13 “priesthood” (“Termini der Priesterschaft”),14 “priest office” (prästämbete)15 and “pagan vicarage” (hedniska prästgårdar)16 to ancient Scandinavian contexts. One exception is Per Vikstrand, who uses the concept of “designations of cultic leaders” (kultfunktionärsbeteckningar).17 The difference between historians of religions and philologists/specialists in toponymy is probably due to the different evaluation of the source categories and different methodologies. Historians of religions rely mainly on historical and narrative sources, while philologists put their trust in the linguistic material and etymology.18 The difference of opinion is also due to confusion of terminology. It seems as if the philologists have used the common category “priest” without any definition or deeper reflection. Nor have historians of religions defined what they mean by “priest”. In this discipline, however, suggestions as to how operational concepts such as “priest” and “priesthood” can be conceived are sometimes encountered.19 In an article published in 1998 I criticized this terminology. On the basis of classical phenomenological treatments, I proposed analytic definitions of the categories “priest” and “priesthood” and tested them on the Scandinavian materials.20 I arrived at the conclusion that general features of priests or

10 Von See 1964, 78f., 105–112. Cf. Maurer 1873; Heusler 1911. 11 Kousgård Sørensen 1989, 5–33. 12 Laur 2001, 223. 13 Brink 1996a, 266. 14 T. Andersson 1992a, 518. 15 Strid 1999, 103. 16 Hellberg 1976, 11; 1986a, 61ff. 17 Vikstrand 2001, 386, 427. 18 Cf. Vikstrand 2001, 396. 19 See e.g., Sabourin 1973; Widengren 1969. 20 For definitions of “priest” and “priesthood”, see e.g., Sabourin 1973; Widengren 1969; Sundqvist 1998, 78f. 166 CHAPTER 5 priesthood were vague in early Scandinavia. There is weak evidence for initia- tions into or formal training for a religious office.21 The cultic officials do not seem to constitute an organized and hierarchical form of institution, appearing as an independent social stratum in society. Neither was there a priestly institution normalizing world view or ritual practice. Since the concept of “priest” (from Greek πρεσβύτερος ‘the older’, πρέσβυς ‘old person’) was formed and developed in a Christian context I argued that it is more correct to use more neutral concepts, such as cultic leaders, cult performers or religious officials.22 The general categories to be used in the study of religion have been much discussed. During the first half of the twentieth century, the phenomenologists of religion developed instrumental tools and operational concepts for the discipline by means of a comparative method. By using a synchronic approach they elucidated structures in the religious material which were supposed to appear transhistorically and transculturally. In that manner certain categories were elaborated, such as “myth”, “ritual”, “priest” and “temple”. In recent decades the comparative perspective and the phenomenology of religion have been much criticized.23 This criticism rejects universal theories and concepts produced by the comparativists; the phenomenologists, it says, have observed only the similarities and ignored the particularities and differences. The critics argue that the generalized categories are rarely based on solid empirical material and that the old comparativists often proceeded from a priori assumptions. When construing a category such as “temple” or “priest”, for instance, the cognitive conception is strongly influenced by Christian and western thinking. The use of such concepts in non-Christian contexts always causes a risk of serious misinterpretation.24 In my opinion this criticism has been sound in parts. On the other hand, I think that some common analytic concepts are needed to enable interdis- ciplinary communication and transcultural comparisons of religious phenomena, even if they must always to a certain extent be subjective. We must, however, always keep in mind that the concepts applied do not always correspond to the native categories. The distinction between etic and emic levels elaborated in anthropology is very enlightening on that particular point.25 All operative concepts applied must therefore always be well defined. We must

21 Traces of initiation may be seen in mythical traditions, however. See mainly Schjødt 2008. Cf. Sundqvist 2009b; 2010a. 22 For a more detailed argumentation, see Sundqvist 1998, 76ff. and also 2003a; 2003b; 2007. 23 E.g., J. Z. Smith 2000, 23–44; 2004; Flood 1999; Gilhus & Mikaelsson 2003, 61ff., 79ff., 83ff. 24 Cf. Hewitt 1996, 16; Rüpke 1996, 241. If our tools (i.e. our terms) “do not do the job properly, then they must be honed or replaced”. Jensen 2014, 7. 25 E.g., Lett 1996, 382f. The Religious Leadership of the Sanctuaries 167 also avoid value-loaded terms that apparently have their background in western or Christian contexts or problematic evolutionistic discourses. This criticism of concepts has also involved the notion of “priests”. Jörg Rüpke, for instance, has formulated a critique of Joachim Wach’s typology of ten categories of religious authority, including concepts such as priest, priesthood, prophet, shaman, healer, magician and diviner. According to Rüpke these terms are taken from specific cultural contexts but are used transculturally. He states that: “these types are defined by certain functions and performances that sometimes are embarrassingly missing from the culture just analysed.”26 As an alternative Rüpke attempts to describe cultic leaders as agents of control within systems of symbols. They may thus be defined as controllers. These controllers may monopolize certain ritual activities. They may control theological products and dogmas, as well as holy scriptures or oral traditions. Writing may be a way to preserve verbal conceptualizations of symbol systems. The controllers could therefore also monopolize the systems of documentation, i.e. writing systems. In my view, however, concepts such as controller, theology and dogma fit badly into the ancient Scandinavian context. The term “cultic leader” is proposed instead. It refers to a person who was temporarily responsible for certain public religious functions in society and at the cultic sites.27 The cultic leader had other societal duties beside his or her religious tasks. He or she also functioned as a general political leader. The term “religious specialist” designates an exclusive religious office in the present study, i.e. when a more intensified and permanent specialization had taken place.28 We will now turn to the formation of cultic leadership at the sanctuaries in the three investigation areas.

5.2 Chieftains (hǫfðingjar, hersar, goðar) as Cultic Leaders of the Sanctuaries

Different types of sources indicate that chieftains controlled the cultic sites and the ceremonial buildings in the three investigation areas. These chieftains were sometimes designated goðar (OSw *gudhar). This term (ON goði sg.) is a derivation of the noun goð ‘god’, indicating that this denomination in some

26 Rüpke 1996, 241. 27 Cf. Gunnell 2001. 28 This analytic term can be related to Rüpke’s concepts of profession and professionalization. 168 CHAPTER 5 sense referred to a cultic function and/or office.29 Sometimes these chieftains also carried designations with no religious references, such as hǫfðingjar and hersar. In what follows I will argue that most of these authorities also performed in other societal functions beside their religious assignments and thus should be described as cultic leaders.

5.2.1 The Relation between the Chieftains and Sanctuaries in Trøndelag Medieval prose texts and other sources report that chieftains in Trøndelag and Norway had a close relation to the pre-Christian public sanctuaries and also organized cult there.30 Some of them seem to have been designated goðar.31 According to these sources, they totally controlled the ceremonial buildings or even regarded them as their own property. It seems as if they could do whatever they wanted with them. When they moved, for instance, they dismantled these buildings and brought the most essential parts of them with them to the new land where they planned to settle. Landnámabók, for instance, states thus:

Þórhaddr the old was a hofgoði in Trondheim in Mære. He wanted to travel to Iceland, and before he went, he took down the hof and brought all the soil of the sanctuary and the high-seat posts. He came to Stǫðvarfjǫrðr and proclaimed the Mære Peace (Mærina-helgi) over the entire fjord area and nothing was allowed to be killed there, except the animals on the farm.32

29 See e.g., de Vries 1956–57, §277. 30 Most likely the chieftains of Norway combined religious and other societal leadership functions. Phillpotts 1912–1913, 267ff. 31 The earliest evidence of a chieftain designated goði (PN gudija) in Norway is probably a runic inscription from the farm of Nordhuglo, on the island of Huglo (south of Bergen), in Hordaland, dated to the fifth century. Spurkland 2005, 48f. The word gudija, according to Krause (1966, 146f.), is a jan-stem and it may be related to a Goth gudja (m.) ‘priest’ (cf. ON goði, i.e. an an-stem) attested in the Gothic Bible. It is a translation of the Greek ἱερεύς ‘priest’, ‘sacrificer’ in Wulfila’s Bible (cf. sa auhumista gudja for ἀρχιερεύς ‘arch- priest, high-priest’). The term is thus old, and it seems to be spread in several parts of the Germanic area. Sundqvist 2003b. 32 Þórhaddr enn gamli var hofgoði í Þrándheimi á Mæri. Hann fýstisk til Íslands ok tók áðr ofan hofit ok hafði með sér hofsmoldina ok súlurnar; en hann kom í Stǫðvarfjǫrð ok lagði Mærina-helgi á allan fjǫrðinn ok lét øngu tortíma þar nema kvikfé heimilu. Landnámabók S 297, H 258, Ísl. Fornr. 1, 307f. My trans. The Religious Leadership of the Sanctuaries 169

This short passage gives us some important information.33 It reports that Þórhaddr had full control over his hof sanctuary, and even dismantled it when he travelled to Iceland. The close relation between him and his sanctuary is indicated by his title hofgoði ‘sanctuary chieftain’. Most likely Þórhaddr was regarded as a political leader or chieftain as well. The Skarðsárbók and Þórðarbók versions of Landnámabók report that he was described as a (great) chieftain (hǫfðingi (mikill).34 Archaeology may partly support the information in Landnámabók (see ch. 4 above). Under the medieval church of Mære remains of a Viking Age building were found, which has been interpreted as a hof sanctuary. Gold foil figures were discovered in the building (fig. 8B). They constitute a strong indication of cultic actions and the presence of a political-religious ruling power at this site at the beginning of the ninth century. Perhaps the gold foils should be related to a wealthy family who had great ambitions to gain power in Inn-Trøndelag, in other words, on a regional level as well.35 Since Mære probably was no magnate farm during the Early Viking Age, it has been suggested that the family who controlled the sanctuary lived at Egge.36 At this place rich Late Iron Age burials have been found. Later Mære turned into a communal sacred site, organized by several chieftain families (see below and ch. 14). Landnámabók combined with the archaeological evidence informs us that chieftains (goðar) played an important role in the public cult in Trøndelag during the Early Viking Age. There was, however, a radical change in this situation when the central royal power started to exert influence in this province, i.e. in the late ninth century.37 Sources report that local chieftains migrated

33 This laconic information may have quite a high historical source value. See e.g., Dillmann 1997. 34 Landnámabók, Ísl. Fornr. 1, 307, note 12. 35 The Old Norse medieval traditions report that kings and earls appeared in Sparbyggvafylki, during the Early Viking Age. Snorri mentions, for instance, an earl called Ketill jamti (son of Ǫnundr jarl). He lived in Sparbyggvafylki, during the Early Viking Age. He fled to Jämtland when King Eysteinn Upplendingakonungr conquered the land c. 780. Hákonar saga góða 12, in Heimskringla I, Ísl. Fornr. 26, 164f.; Óláfs saga helga 137, in Heimskringla II, Ísl. Fornr. 27, 241. Snorri also reports about a king of this fylki, who fell in battle just before Haraldr Finehair’s appearance. Haralds saga ins hárfagra 7, in Heimskringla I, Ísl. Fornr. 26, 99. 36 See the discussion in e.g., Stenvik 1996; Røskaft 1997, 237; 2003, 138f.; Lidén 1999, 45. 37 It is possible that hofgoðar were important in the Late Viking Age as well. One manuscript of Heimskringla mentions that during Guðbrandr’s life (c. 1000 AD) there was a hofgoði of the Dalesmen who was called Þórðr ístrmagi. In the other manuscripts Þórðr is called hǫfðingi. See Phillpotts 1912–1913, 271. 170 CHAPTER 5 to Iceland with their sanctuaries, in a similar fashion to Þórhaddr the Old.38 The reason for doing this, according to these narratives, was the harsh reign of King Haraldr Finehair.39 One example is the chieftain Hrólfr Mostrarskegg, who brought his hof sanctuary from Mostr Island in south-western Norway to western Iceland after a conflict with King Haraldr (see ch. 3 and 4 above). Sources indicate thus that there was no formal religious organization or priesthood, which owned land in the Early Viking Age Norway or other parts of Scandinavia. In the central settlement districts land was owned by private persons or families. If a chieftain erected a ceremonial building on his land, it was his property alone, and he could do whatever he wanted with it. These sanctuaries were thus related to aristocratic centres and chieftains’ farms. Þorbjǫrn hersir in Fjalafylki, for instance, took care of (inf. varðveita) the hof sanctuary at his farm in Gaular,40 while Guðbrandr hersir cared for his hof in Guðbrandsdalir at the end of the tenth century.41 The people living in the settlements could however be invited to take part in the ceremonial feasts at these sanctuaries. Perhaps they paid tribute to do this (see below). The close relationship between the Þrœndir chieftains, hof sanctuaries and cultic activities may also be indicated by archaeological finds attested at farms designated hov in Trøndelag, i.e. places where pre-Christian sanctuaries were possibly erected. The best example is probably the farm Hove in Åsen, in Strindafylki, where we have great monuments and exclusive finds indicating the presence of political power.42 Hove is surrounded by other old farms such as Husby and Vang and perhaps they originally constituted one farm unit. All three sites have Viking Age burial fields and at Vang a great burial mound was erected. The name Vang has been interpreted as a sacred or cultic place- name.43 It may refer to a special kind of sacred meadows which was often located beside the hof sanctuaries in pre-Christian Norway (cf. Hovsvangen (Oppl.) and Gudvangen (SogFj)). During excavations at Hove in 1981–84 several finds were made indicating pre-Christian cultic activities.44 Several pits with

38 See e.g., Landnámabók SH7–8; H11; S289 H 250; S310 H270. For more complete documentation, see Strömbäck 1928b; Birkeli 1932, 24ff. 39 It seems, however, as if Þórhaddr went to Iceland more or less voluntarily. The expression hann fýstist til Íslands “he wanted [to travel] to Iceland” indicates this. Jón Viðar Sigurðsson et al. 2005, 96. 40 . . . at hofi því, er Þorbjǫrn . . . hafði varðveitt. Landnámabók S 368, Ísl. Fornr. 1, 368. 41 Brennu-Njáls saga 87–88, Ísl. Fornr. 12, 208–220. Óláfs saga Helga 112, Heimskringla II, Ísl. Fornr. 27, 183–188; Flateyjarbók II, 189. 42 Farbregd 1986; Røskaft 2003, 66ff.; Nordeide 2011, 104ff.; 2012. 43 M. Olsen 1926, 218. Cf. Sandnes 1992, 15f.; Sandnes & Stemshaug 2007, 479. 44 Farbregd 1986. The Religious Leadership of the Sanctuaries 171 charcoal and burned stone suggest that ritual cooking took place there, from c. 100 BC to AD 1000. In connection with these pits post-holes were also discovered which not were related to house foundations. Perhaps they could be associated with ritual structures. There are other hov-places in this area which may have been connected to local cultic sites and also ceremonial buildings, such as Hov in Ålen, Hov in Soknedalen, Hov in Orkdal, and Hove at Byneset.45 These farms all had a central position in the local settlements and most likely they functioned as public gathering places for the local judicial assembly and pre-Christian cult. It seems as if local magnates owned these farms and from them they controlled the social-political activities in the settlement districts during pre-Christian periods.46 At several of these places early medieval churches were also erected. Several of them may be regarded as private farm churches (hǿgendiskirkjur). A religious and political power continuity may probably be discerned at these sites. It can be concluded that rulers of different dignity thus had sanctuaries and ceremonial buildings erected on their own farms in Trøndelag during the Viking Age. Some of these farms were also central places for the entire fylki district. At Melhus, in Gauldœla fylki, for instance, many finds have been made on the farm, mainly dating to the Late Iron Age. The finds indicate a high social status, for instance an expensive sword and a buckle. Viking Age burial mounds are also attested at this farm.47 During archaeological investigations at Melhus and Skjerdingstad (just south of Melhus), several three-aisled buildings dating from the Iron Age have been found. Some of them had fireplaces, and one had three fireplaces, indicating that no animals were held there.48 The medieval written sources report that ceremonial banquets were celebrated at Melhus. It is mentioned, for instance, that Earl Hákon took part in banquets at this farm.49 It is plausible that banqueting halls were erected at Melhus during the Viking Age. It seems as if Snorri believed that Melhus was an important religious centre in tenth-century Trøndelag. He states, for instance, that Ásbjǫrn of Meðalhús (= Melhus) was one of the chieftains, who organized public

45 Røskaft 2003, 53–76. 46 Røskaft 2003, 209. 47 Bjørk Birgisdottir & Strøm 2002. 48 Rønne 2002. 49 In Óláfs saga Tryggvasonar, Snorri mentions that Earl Hákon took part in banquets (vei‑ zlur) at Melhus. See Heimskringla I, Ísl. Fornr. 26, 293ff., 316. According to Ágrip, it seems as if Earl Hákon controlled activities at Melhus. This text also mentions Skerdingstad. See Ágrip, Ísl. Fornr. 29, 16. Cf. Ágrip, (ed.) Driscoll, 22. 172 CHAPTER 5 sacrifices in Trøndelag.50 Cooking pits discovered at Melhus indicate that ritual activities really took place at this site during the Late Iron Age.51

5.2.2 Collaboration between Þrœndir Chieftains in Cultic Matters during the Tenth Century It seems as if the normal way of organizing cult in Early Viking Age Trøndelag was that independent and individual chieftains took care of cultic matters at local sanctuaries erected on their farms. This decentralized organization structure may have been challenged during the Viking Age. The Old Norse prose traditions report that the pagan chieftains of Trøndelag during the tenth century collaborated concerning cultic issues, especially when the traditional religion was threatened by the Christian royal power. Sometimes this cult was practised at places which may be regarded as communal centres. Perhaps this cult was part of the resistance movement. There is a quite lengthy story in Snorri Sturluson’s Hákonar saga góða 13–18 which concerns cultic conditions in Trøndelag during the first attempts at a royal mission there in the tenth century (see the text in ch. 4 above). According to Snorri, King Hákon the Good was the first king who wanted the gospel to be preached in Norway. When he arrived in Trøndelag he met great resistance from the Þrœndir farmers, who were led by the fylki chieftains. At the Frostathing, the farmers said that it was their wish that the king should make sacrifice to procure for them good crops and peace, as his father had done. But when the king attended a sacrificial feast at Lade at the beginning of winter, he did not drink the ceremonial toasts to the pagan gods, or eat the sacrificial horse meat. A dispute between the king and the great farmers arose, but Earl Sigurðr managed to end the tumult. Later in the autumn when the jól feast was prepared in Mære, eight chieftains (hǫfðingjar), “who had most to do with the sacrifices in the whole Trøndelag”, arranged a meeting between them.52 They decided that the four from the outer districts were to destroy the Christianity there was, and the four of the inner districts were to force the king to sacrifice. At Mære the chieftains, together with a great number of farmers, met King Hákon and Earl Sigurðr during the jól banquet. They thronged about the king and asked him to sacrifice, or else they would force him to. Earl Sigurðr mediated between them, and in the end King Hákon ate a few bits of horse liver. He also drank all the toasts the farmers poured for him.

50 Heimskringla I, Ísl. Fornr. 26, 172. 51 Røskaft 2003, 98. 52 . . . er mest réðu fyrir blótum í ǫllum Þrœndalǫgum . . . Heimskringla I, Ísl. Fornr. 26, 172. The Religious Leadership of the Sanctuaries 173

According to Snorri’s text it seems as if the tenth-century Þrœndir were well organized when it comes to cultic matters. The eight chieftains who cared for public sacrifices in Trøndelag collaborated with each other in order to prevent the spread of Christianity in Trøndelag. In previous research it has been argued that this passage in Hákonar saga góða must be regarded as a literary construction made by Snorri. It is claimed that the images of these chieftains and their enterprises actually are projections back to Viking Age of medieval power and societal conditions in Trøndelag.53 In my opinion, this is partly true; Snorri probably made some literary constructions when describing the conflict between the Þrœndir and King Hákon. However, we cannot rule out that Snorri based some of his story on old traditions, which may have carried more historical aspects (see e.g., the section “Snorri’s sources” in ch. 4 above).54 He might, for instance, have heard that pagan chieftains were concerned about religious matters, and that they sacrificed and organized cultic feasts at their farms in Viking Age Trøndelag e.g., at Meðalhús (= Melhus), Varnes (Værnes), and Ǫlvishaugr (Alstadhaug).55 He may also have been informed that their cultic activities were expressions of an organized resistance against the Christian and royal power during second half of the tenth century. In my opinion, it is plausible that such information may have reached Snorri during his journey in Norway, for instance, when he met Earl Skúli of Lade in Tønsberg sometime between 1218 and 1220.56 Kings’ Sagas older than Snorri’s text indicate an organized resistance among Þrœndir against King Hákon and his Christian mission. In Ágrip (c. 1190), for instance, it is stated that the Þrœndir burned down the churches that King Hákon had erected and killed the priests he had appointed. It is also stated that they rose against the king at Mære and required the king to sacrifice as other kings used to do, or they would throw him out of the country.57 Since the king noted that the farmers stood hard against him and also had support from their chieftains, he accepted their wish in order to tie bonds of friendship with them. It is thus obvious that the scribe of Ágrip intended to describe how the Þrœndir were well-organized and also had support from their chieftains. A similar image of the event in Mære is

53 See e.g., Røskaft 2003, 83ff. 54 It has been argued that Snorri was a thirteenth-century historian who tried to render the history as truly as possible. He therefore based his narratives on such traditions as he deemed reliable. Cf. Meulengracht Sørensen 1991b. 55 Archaeological finds actually indicate such cultic organization in Viking Age Trøndelag (see above). 56 Cf. Steinsland 2000, 112. 57 Ágrip, Ísl. Fornr. 29, 8. 174 CHAPTER 5 also produced in Fagrskinna.58 Perhaps these medieval Kings’ Sagas incorporated an old oral tradition, which reported that a cooperative of chieftains cared for the public cult at the communal sanctuary of Mære in the tenth century before the conversion (see ch. 14 below).

5.2.3 The goðar and their hof Sanctuaries in Iceland Several Old Norse prose texts report that many minor chieftains (hǫfðingjar, hersar, goðar) and great farmers left Trøndelag and Norway and migrated to Iceland during King Haraldr’s reign at the end of ninth century and the beginning of the tenth century.59 Some of them had been involved in cultic matters in their homeland, for instance, the hofgoði Þórhaddr the Old. It is stated that these chieftains and cultic leaders sometimes brought parts of the sanctuaries from Norway to Iceland, such as the high-seat posts and the “temple-soil”. When arriving in the new land, they consecrated the land where they intended to build their farms. It seems as if the intention was to establish some kind of sacred or ritual landscape. On the new farms they also erected hof buildings, where they placed the ritual objects, such as the high-seat posts. The intimate relation between the chieftain (goði) and his sanctuary (hof ) may be seen in the compound hofgoði, which was quite a common title among Icelandic leaders.60 There are several instances in the Old Norse prose describing the close relationship between chieftains (goðar) and hof sanctuaries in Iceland. One of the most important texts is probably the one found in Úlfljótslǫg.61 According to the Icelandic tradition, the lawman Úlfljótr brought this pre-Christian “Law” from Norway to Iceland in connection with the establishment of the Icelandic Law at the General Assembly (alþingi) at Þingvellir.62 It has been preserved in the Hauksbók redaction of Landnámabók,63 Þáttr Þorsteins uxafóts (fourteenth

58 Fagrskinna, Ísl. Fornr. 29, 80. 59 Landnámabók states that several chieftains of the Trøndelag region migrated to Iceland during Haraldr’s reign. See e.g., Hrafn enn heimski and his son Jǫrundr goði in Landnámabók, Ísl. Fornr. 1, 340f., S 338 H 296; Kolgrímr hinn gamli in Landnámabók, Ísl. Fornr. 1, 65, H 22; the brothers Eyvindr vápni and Refr enn rauði in Landnámabók, Ísl. Fornr. 1, 289, S 267 H 229; Bárðr blǫnduhorn in Landnámabók, Ísl. Fornr. 1, 342, S 340 H 298. Phillpotts (1912–1913) noted that many chieftains who were designated as goðar in Iceland actually came from old hersir families in Norway. 60 See Fritzner 1954; Heggstad et al. 1993. 61 See Jón Hnefill Aðalsteinsson 1998, 35–56. 62 This is mentioned already by Ari in Íslendingabók, Ísl. Fornr. 1, 7. 63 Landnámabók, H 268, Ísl. Fornr. 1, 313ff. On the dating of this text, see Hallberg 1979, 3. The Religious Leadership of the Sanctuaries 175 century)64 and the Vatnshyrna version of Þórðar saga hreðu (thirteenth–fourteenth centuries).65 These three texts have different versions of the “Law”. According to the passage from Hauksbók (Landnámabók) it says that a ring weighing two ounces or more should lie on the stalli in every chief hof building, and every chieftain (goði) should have this ring upon his arm at all public law-motes at which he should be at the head of affairs, having first reddened it in the blood of a neat which he himself had sacrificed there. Further down in the text it says that then was the land divided into Quarters, and it was decided that there should be three assemblies in each Quarter and three hof buildings in each Assembly Commune, i.e. Assembly District or Community, and that men should be selected according to wisdom and righteousness to have ward of the hof, and they were to nominate Courts of Judges at the assembly and to regulate the proceedings of lawsuits, and therefore were they called goðar; and every man should pay toll to the hof as now they pay tithes to the Church.66 The historical source value of this text has been much debated.67 Olaf Olsen, for instance, has questioned several details in this description, such as the organization and the number of the thing-assemblies and hof sanctuaries in Iceland, as well as the information about sanctuary dues (hoftollr), and the oath-rings.68 I will come back to some of these issues below. Here I will only concentrate on one aspect in Olsen’s criticism, namely, the connection between chieftains (goðar) and hof sanctuaries, and the appointment of these cultic leaders. According to Olaf Olsen, the information that the goðar were “selected according to wisdom and righteousness to have ward of the hof ” must be considered as nothing more than a myth which had nothing to do with historical reality.69 The goðar were never selected by the people as “temple superinten- dents” (“tempelforstandare”). It was people who derived from the foremost

64 Flateyjarbók I, 249. 65 Printed in Kjalnesinga saga, Ísl. Fornr. 14, 231f. On these texts, see O. Olsen 1966, 34ff. 66 Hauksbók states thus in Old Norse: Baugr tvíeyringr eða meiri skyldi liggja í hverju hǫfuðhofi á stalla; þann baug skyldi hverr goði hafa á hendi sér til lǫgþinga allra, þeira er hann skyldi sjálfr heyja, ok rjóða hann þar áðr í roðru nautsblóðs þess, er hann blótaði þar sjálfr . . . Þá var landinu skipt í fjórðunga, ok skyldu vera þrjú þing í fjórðungi, en þrjú hǫfuðhof í þingsókn hverri; þar váru menn valðir til at geyma hofanna at viti ok réttlæti; þeir skyldu nefna dóma á þingum ok stýra sakferli; því váru þeir guðar kallaðir. Hverr maðr skyldi gefa toll til hofs, sem nú til kirkju tíund. Landnámabók, H 268, Ísl. Fornr. 1, 313–315. 67 See e.g., Maurer 1869; 1874; Phillpotts 1912–1913; O. Olsen 1966; Jón Hnefill Aðalsteinsson 1998, 35–56. 68 O. Olsen 1966, 34–49. 69 O. Olsen 1966, 42f. 176 CHAPTER 5 and noblest families in Iceland who had the power position and the opportunity to play a central role at the judicial courts and in public cult. Usually they erected the hof buildings on their own farms. Olsen’s argument is, in my opinion, well founded. People from the lower strata of society in Iceland could probably never be “selected as goði” or “appointed to a hof ” just because they were wise and righteous. Neither are there any signs in the sources that people were initiated into the office of goðorð after a long period of religious and judicial education. The cultic function and the title as goði was probably originally inherited within certain families who belonged to the elite of Iceland.

5.2.4 The Inheritance of Cultic Leadership The free people of Iceland thus did not select someone for the office of goðorð or appoint a candidate to take care of the hof merely because of that person’s wisdom and righteousness, as suggested in Úlfljótslǫg. These qualifications were of course expected from a legitimate leader, as were other qualities, such as generosity and ability to obtain support from friends.70 There were probably additional aspects which qualified someone for the status of goði. According to some traditions, it seems as if the role as cultic leader (goði) in Iceland, including the charge over the hof, was inherited within the chieftain family.71 These traditions intimate that the dignity of cultic leader was brought from Norway to Iceland. Landnámabók and Eyrbyggja saga, for instance, report that the chieftain Þórólfr Mostrarskegg continued as a cultic leader on his new farm Hofstaðir after he left Mostr in Norway and settled at Breiðafjǫrðr in western Iceland (see ch. 3 and 4 above). In Landnámabók we do not get much information about Þórólfr’s identity or what kind of office he occupied. It says only that he was the son of Ǫrnólfr fiskreki and that he lived in Mostr. It is also mentioned that “he was much devoted to offering up sacrifices” (blótmaðr mikill), i.e. that he was a very religious man, and that he “believed in Þórr” (trúði á Þór). According to Eyrbyggja saga, Þórólfr was a chieftain of considerable standing (hǫfðingi mikill). It also says that he was “a close friend of Þórr” (mikill vinr Þórs), and watched over a Þórr’s hof (varðveitti . . . Þórshof) at his farm in Mostr Island. No text explicitly states that he occupied the office as goði, even though he was a very religious man. Evidence in both Eyrbyggja saga and Landnámabók indicates, however,

70 On “the wise chieftain” and other qualifications of chieftains, see Jón Viðar Sigurðsson 1999, 84–150. 71 That cultic assignments were inherited in a family is known from many ancient cultures. See Sabourin 1973, 102–135. The Religious Leadership of the Sanctuaries 177 that Þórólfr occupied a cultic office, which also included the charge of the hof. This office was passed on within his family to his sons and grandchildren.72 Both Landnámabók and Eyrbyggja saga state that several male members of his family were titled goðar, i.e. they were regarded as some kind of cultic leaders. Eyrbyggja saga 3, for instance, mentions that Þórólfr had a son called Hallsteinn, who was born in Norway. Chapters 5 and 6 tell how he went to Iceland together with Bjǫrn Ketilsson. Hallsteinn considered it a slur on his manhood that he should have land granted him by his own father Þórólfr at Hofstaðir, so he crossed over to the other side of Breiðafjǫrðr to a place called Hallsteinsnes at Þorskarfjǫrðr, and staked his claim there. Later in the text Hallsteinn was called goði af Hallsteinsnesi.73 Similar information is also recorded in Landnámabók, where he is called Þorskafjarðargoði.74 This text says that he sacrificed to Þórr, since he wished that the god could send him his high-seat posts. After some days a big tree came ashore, which Hallsteinn used for high-seat posts.75 In my opinion, it is possible to interpret the tradition thus: Hallsteinn was qualified to occupy a goðorð since he was a son of the chieftain and cultic leader Þórólfr. Even though it not explicitly mentioned, Eyrbyggja saga intimates that Þórólfr’s other son, Þorsteinn þorskabítr Þórólfsson, was also a kind of a goði, who took care of the hof at Hofstaðir after his father’s death. The context of the narrative suggests that Þorsteinn lost the full custody over this sanctuary. According to chapter 9, Þorsteinn had a great conflict with Þorgrímr Kjallaksson and Ásgeirr á Eyri. Þorgrímr and Ásgeirr threatened to desecrate the sacred field at Þórsnes, which led to an open battle between the Kjalleklingar and the Þórsnesingar.76 There had been deaths on both sides and many wounded. People decided to send for Þórði gelli, who at that time was a leading chieftain at Breiðafjǫrðr. He was considered the likeliest man to make peace between the fighting parties. Þórði decided that no compensation was to be paid for the killings and woundings at Þórsnes. Since Þórsnes now was defiled by the spilling of blood in enmity it was considered useless for holding assemblies, so a new place had to be chosen for this purpose. In order to ensure friendship between the fighting parties and establish reconciliation, Þórði declared that

72 Several scholars have regarded Þórólfr as a goði, see e.g., Wessén 1924, 170; Baetke 1942, 133f.; Strömbäck 1975, 41f. Cf. DuBois 1999, 65f.; Sundqvist 2007, 25ff. 73 Eyrbyggja saga 48, Ísl. Fornr. 4, 136. 74 Landnámabók M 25 S85, Ísl. Fornr. 1, 124ff. He is also mentioned by Ari in Íslendingabók, Ísl. Fornr. 1, 10. 75 Landnámabók S123, H95, Ísl. Fornr. 1, 163f. 76 Eyrbyggja saga, Ísl. Fornr. 4, 14ff. This tradition also occurs in Landnámabók S85, Ísl. Fornr. 1, 125f. 178 CHAPTER 5 from then on Þorgrímr Kjallaksson was to bear half of the coast of maintain- ing the hof sanctuary at Hofstaðir, and that he and Þorsteinn were to share the sanctuary dues and support of the farmers equally between them.77 Þorgrímr was also supposed to back Þorsteinn in all his lawsuits and safeguard the sanctity of whatever place Þorsteinn might choose for the new assembly. Þorgrímr was from then on designated Þorgrímr goði.78 The text clearly indicates that Þorsteinn þorskabítr, before the conflict with the sons of Kjallakr, was the only person who occupied the cultic office at Hofstaðir. He probably inherited this office from his father Þórólfr, as Hallsteinn had done. After the conflict with Þorgrímr, however, Þorsteinn was forced to share the chieftaincy (goðorð) with him. The interpretation that Þorsteinn acquired the chieftaincy (goðorð) and the management of the hof at Hofstaðir by means of the inheritance from his father harmonizes with the continuation of the saga, for it is mentioned that Þorsteinn later built a great farm at Helgafell, where he erected a hof. At his new farm, Þorsteinn and his wife had a son called Grímr. It seems as if Grímr inherited his father’s cultic role and the charge of the sanctuary: “Þorsteinn dedicated this boy to Þórr, calling him Þorgrímr, and said he should become a hofgoði.”79 Eyrbyggja saga goes on to say that Þorgrímr Þorsteinsson, in his turn, had a son called Snorri.80 He inherited his father’s farm and the sanctuary at Helgafell. “He was now in charge of the hof and was thus called goði (Hann varðveitti þá hof; var hann þá kallaðr Snorri goði).” We may note that the scribe of the saga used the verb varðveita ‘be in charge of, take care of’ when expressing the relation between the cultic leader and his sanctuary.81 This expression was also used when describing Þórólfr’s relation to his sanctuary in Mostr; i.e. Þórólfr watched over the Þórr’s hof (varðveitti . . . Þórshof) located there. The author of Eyrbyggja saga intimates that several members of Þórólfr’s family for several generations occupied a religious office which included the charge of hof buildings in the areas around Breiðafjǫrðr. It also seems as if

77 Þorgrímr Kjallaksson skyldi halda uppi hofinu at helmingi ok hafa hálfan hoftoll ok svá þing‑ menn at helmingi, . . . Eyrbyggja saga, Ísl. Fornr. 4, 17. 78 Var hann af því kallaðr Þorgrímr goði. Eyrbyggja saga, Ísl. Fornr. 4, 18. 79 Eyrbyggja saga, Ísl. Fornr. 4, 19. 80 Eyrbyggja saga, Ísl. Fornr. 4, 27. 81 Fritzner (1954) translates the verb varðveita ‘tage vare paa noget for at det ikke skal komme til Skade’. He also suggests a translation of the specific expression varðveita hof ‘siges den Mand der har det paa Gaarden, vedligeholder det og besørger Gudstjensten’. The Religious Leadership of the Sanctuaries 179 the sons and grandsons of Þórólfr were devoted to the cult of Þórr.82 In my opinion, the idea of making Þórólfr and his descendants into a kind of “priest family” who cared for local sanctuaries at Breiðafjǫrðr was not a literary construction made by the scribe. Most likely he built this idea on an older tradition. In Landnámabók too we may note that several descendants of Þórólfr were titled goði. Þórólfr’s son Hallsteinn was called Þorskafjarðingargoði, while his grandchild was called Þorgrímr goði Þorsteinsson þorskabíts, and his great grandchild Snorri goði Þorgrímsson.83 The sources report that the goðorð could also be acquired in other ways as well. It could be shared between two persons, received as a gift or even purchased. Hrafnkels saga Freysgoða, for instance, mentions that Þorkell leppr Þjóstarsson had passed over his position of authority (mannaforráðr) to his brother Þorgeirr when he went abroad. Before that he had been a goðorðs‑ maðr. The text indicates that he originally inherited the title goði from his father. In the course of events, Þorgeirr offers to pass back the goðorð and the position of authority to his brother Þorkell for a period, but then, he suggests, they could share it equally between them. Þorkell, however, refuses to take this offer from his brother, since he felt that Þorgeirr was the more accomplished of the two for this assignment.84 Eyrbyggja saga likewise says that Þorgrímr and Þorsteinn shared the goðorð at Þórsnes. We must not forget that these statements appear in medieval texts. Most scholars accept, however, that the goðorð could be handled in such way.85 The evidence from the saga literature thus suggests that chieftains who were designated hofgoðar and goðar erected and were in charge of the hof buildings in Iceland. Some of them seem to have occupied a cultic office in their old homeland as well, for instance Þórhaddr and Þórólfr. They actually brought some parts of the old sanctuary with them when they emigrated from Norway to Iceland. The ceremonial buildings were usually built at the farms of the chieftains. Some sagas indicate that the office of hofgoðar/goðar and the care of the sanctuary were inherited within the family. Such an office could also be acquired in other ways. It could, for instance, be received as a gift or purchased. The sources indicate that the Icelandic goðar should not be seen as persons

82 Some of Þórólfr’s sons, like their father and grandfather, carried the name of the god as a first element in their names, e.g., Þorsteinn and Þorgrímr. Whether this naming custom within goði families really reflects ancient conditions is much debated. See e.g., T. Andersson 1992a; Vikstrand 2009. 83 Landnámabók, S86, H74, M25. Ísl. Fornr. 1, 124–127. 84 See e.g., Hrafnkels saga Freysgoða, Ísl. Fornr. 11, 112ff. 85 See e.g., Byock 2001, 14. 180 CHAPTER 5 who exclusively occupied a priestly office and permanently appeared in cultic roles.86 Most of them functioned as general political-judicial leaders as well, and also had many secular leadership assignments in society. They could thus be described as both hofgoði and lagamaðr (see ch. 2). Some of them, however, were closely related to a deity. It could be Þórr,87 for instance, but it could also be Freyr.88 Sometimes the chieftains’ sanctuaries therefore were dedicated to one of these gods.

5.2.5 Female Cultic Leaders (gyðjur) in Iceland The Sagas of Icelanders also report that female leaders, designated gyðjur (gyðja sg., a female equivalent of the term goði) or hofgyðjur, could be in charge of hof buildings in Iceland. Vápnfirðinga saga, for instance, describes the hof‑ gyðja called Steinvǫr. She was in charge of a major hof (varðveitti hǫfuðhofit) at the farm called Hof in Vápnfjǫrðr, eastern Iceland. To her all the farmers had to pay sanctuary tribute (hoftoll).89 We cannot rule out that Steinvǫr also had a political position in Vápnfjǫrðr since she took care of the sanctuary tribute. In order to execute this assignment she was supported by the chieftain Brodd-Helgi, which indicates that she at least did not have military power. Old Norse texts also mention the name of other gyðjur and hofgyðjur, who probably played important roles at sanctuaries in pre-Christian Iceland. In Landnámabók and Vatnsdæla saga, for instance, Þuriðr gyðja Sǫlmundardottir is mentioned.90 She was connected to the farm of Hof in Vatnsdalr, where a hof building had been erected by Ingimundr inn gamli. Þorlaugr gyðja Hrólfsdóttir, according to Landnámabók, was related to the hof sanctuary at Reykjardal in south-western Iceland,91 while Þuriðr hofgyðja Véþórmsdóttir and her brother Þórðr Freysgoði Ǫzurarson were connected to sanctuaries situated in Bakkárholt.92 Magnus Olsen argued that the gyðjur exclusively appeared in the fertility cult that was intended for Freyr and that these female cultic leaders

86 Cf. Gunnell 2001. 87 Besides Þórólfr’s family, we may refer to Þorgrímr goði, who had a hof sanctuary at Kjalarnes in south-western Iceland. In that sanctuary Þórr was the most worshipped god. Kjalnesingasaga, Ísl. Fornr. 14, 6f. 88 See e.g., the traditions about the chieftain Hrafnkell Freysgoði in Hrafnkels saga Freysgoða, Ísl. Fornr. 11, 98f. Landnámabók mentions a chieftain called Þórðr Freysgoði Ǫzurarson. Landnámabók H 276, Ísl. Fornr. 1, 321. 89 Vápnfirðinga saga, Ísl. Fornr. 11, 33. 90 Landnámabók S 180 H 147, Ísl. Fornr. 1, 223; Vatnsdæla saga, Ísl. Fornr. 8, 95. She is called Gyða in Vatnsdæla saga. Ísl. Fornr. 8, 71f. 91 Landnámabók S 41 H 29, Ísl. Fornr. 1, 79f.; See also M. Olsen 1926, 251f. 92 See Landnámabók H 276, Ísl. Fornr. 1, 321. It is uncertain where Bakkárholt was located. The Religious Leadership of the Sanctuaries 181 there were regarded as this deity’s spouse (“ektefelle”).93 The weak evidence can, in my opinion, neither refute nor corroborate this. But perhaps we can at least relate Þuríðr hofgyðja to Freyr, since she was the sister of Þórðr Freysgoði, whose family was called Freysgyðlingar ‘priestlings of Freyr’. In the region of their home there is a place called Freysnes ‘Freyr’s Headland’.94 A lausavísa made by the skald Þorvaldr Koðránsson (tenth century) provides early and probably reliable evidence of an ON gyðja called Friðgerðr in a cultic context in Iceland.95 According to Kristni saga, Þorvaldr arrived at her farm in Hvammr in western Iceland together with a bishop. They preached the faith while Friðgerðr was in the hof sanctuary sacrificing (enn Friðgerðr var meðan í hofinu ok blótaði).96 Friðgerðr’s son Skeggi laughed at the missionaries. Þorvaldr then uttered this verse:

Fórk með dóm enn dýra; drengr hlýddi mér engi; gǫ́tum háð at hreyti hlautteins, goða sveini, en við enga svinnu aldin rýgr við skaldi (þá kreppi goð gyðju) gall of heiðnum stalla.97

This lausavísa renders a vivid eyewitness description of the gyðja Friðgerðr and her son Skeggi. They seem to be upset, since the skald and the bishop are disturbing them with their mission. Skeggi is called hlautteins hreytir, “the one who cast lots”, and goða sveinn “the servant of the gods”. These expressions

93 See M. Olsen 1926, 247ff. Olsen referred to the late tradition about Gunnarr Helmingr who played the role as Freyr when travelling around in Svetjud. He had a female companion who also appeared to be a “Frøis-prestinne”. Flateyjarbók I, 337ff. Cf. Ǫgmundar þáttr dytts, Ísl. Fornr. 9, 112ff. The text does not state, however, whether this female was a gyðja or not. She is just called Freys kona “Freyr’s wife”. 94 See E. O. G. Turville-Petre 1964, 261. 95 This stanza is preserved in Kristni saga and the Flateyjarbók version of Óláfs saga Tryggvasonar. 96 Kristni saga 2, (ed). Kahle, 9. Cf. Kristni saga 2, trans. Grønlie, 36. 97 Skj. B1, 105; A1, 110. “I preached the precious faith,/ no man paid heed to me;/ we got scorn from the sprinkler/—priest’s son—of blood-dipped branch./ And without any sense,/old troll-wife against poet/—may God crush the priestess—/ shrilled at the heathen altar.” Trans. Grønlie. Grønlie interprets hlautteins hreytir as “the sprinkler of blood-dipped branch”. I follow Düwel’s (1985, 30f.) interpretation here “the one who cast lots”. 182 CHAPTER 5 may not have been chosen by mere accident. Perhaps they refer to the activities that the skald really observed during his stay at Hvammr. The stanza also says that Friðgerðr was shouting from the pagan “altar/platform” (gall of heið‑ num stalla). Even if the ritual context is not fully clear, it seems as if Friðgerðr was performing sacrifices at the “altar/platform” (stallr/stalli), while Skeggi was casting lots.98 The term stallr/stalli indicates that they were at a sanctuary, perhaps inside a hof building. The prose in Kristni saga supports this interpretation. The version found in Flateyjarbók also states that Friðgerðr was sacrificing during Þorvaldr’s visitation.99 The sources thus suggest that the gyðja and her son performed both sacrifices and divination rituals at a ceremonial building, while Þorvaldr prayed about Christ for the people in Hvammr.

5.2.6 The Cultic Leaders of the Public Sanctuaries in Svetjud The sources thus indicate that the chieftains called goðar and their female counterparts called gyðjur protected, or even owned, the cultic sites and the ceremonial buildings in Norway and Iceland. These chieftains and leaders thus had a designation indicating cultic functions. Swedish place-names and runic inscriptions indicate that an equivalent office of ON goðar called OSw *gudhar also existed in the Mälaren area.100 It seems as if also they were closely related to public sanctuaries.101 The close connection between the *gudhi office and public sanctuaries may be evidenced in the place-names around Lake Mälaren. The farm name Gudby (< Gudhaby), in Fresta parish, in Uppland, for instance, has been interpreted as “the *gudhi’s farm”.102 The place-name specialist Lars Hellberg has suggested that the area around Gudby was the cultic centre for the people living

98 For a more thorough examination of the stanza, see Düwel 1985, 30f. 99 . . . Þorvalldr talde tru firir henni en hon blotade j mote. Flateyjarbók I, 270. 100 See e.g., Hellberg 1986a; Elmevik 1966; 1990; 2003; Brink 1996a; Vikstrand 2001. 101 It has been argued that gyðjur also appeared in the Mälaren region. Some specialists on onomastics have argued that this term (OSw *gydhia) is represented in the place-names of this area. Lars Hellberg (1986a) has suggested that the name Gödåker in Tensta parish, about 20 km north of Uppsala, could be interpreted as an Old Swedish *Gydhiuaker, ‘prästinnans åker’. He admits that this interpretation, from the linguistic point of view, is not without problems when referring to the earliest attestations of the name (1492, Gydakrom, Gedaker). These name forms also appear in quite late medieval documents from c. 1500. Hellberg’s interpretation was, however supported by Lennart Moberg (2002), who stated that the interpretation ‘den hedniska prästinnans åker’ was plausible. Several distinguished specialists on onomastics have, however, rejected this interpretation. See T. Andersson 1992a; Vikstrand 2001, 388f. 102 Hellberg 1986a, 61. The Religious Leadership of the Sanctuaries 183 in the settlement districts or “small regional formation” called *Valand.103 Directly north of the village, there is a settlement (farm) designated Vallensjö. According to Hellberg it is derived from the name *Valændasior ‘Valandsborna’s Lake’. The first element includes a genitive form of the inhabitant designation valændar (pl.), which is derived from the settlement name *Valand. This lake, which today is drained, probably played a significant role for all people living in *Valand. Hellberg argues that this lake was regarded as a holy lake and that the *gudhi in the neighbour village may have been the religious specialist/cultic leader (Sw hednisk präst) for all valændar, when sacrifices were performed at this place.104 The *gudhi in Gudby may also have been related to another public cultic site in the neighbourhood. Close to Gudby is also the farm Sanda, where a three- aisled Viking Age hall was discovered at a settlement with roots from AD 500. On a crest about 50 metres to the south-east of the hall, a square stone structure was found. It contained crushed pottery and some 40 “miniature sickles” (miniatyrskäror). Cecilia Åqvist has interpreted this structure as equivalent to an ON hǫrgr.105 Below the crest several hearths were found, including “double hearths” (“dubbelhärdar”). They comprised one rectangular and one oval part containing bones of lamb. In this area below the crest other miniature sickles were also discovered. The hall, the hearths, and the square stone structure most likely constituted a ritual space. It is quite possible that the *gudhi in Gudby may have played a certain role at this cultic site too, perhaps as an administrator and the leader of the public cult. But it is also possible that the magnate at Sanda organized the cult at his farm by himself in a similar way as chieftains did at farms in Trøndelag and Iceland. In the parish of Sorunda, south of Stockholm, in Södermanland, there is also a place called Gudby “the *gudhi’s farm”. Gudby’s neighbour to the north is Tyrved (< Tøravi). The first element of this name contains the genitive form of the inhabitant designation in the genitive *Tørar, i.e. the inhabitants of the area called Tør. The second element is the substantive vi. There are several large burial mounds indicating that Sorunda was an important area during the Late Iron Age.106 Perhaps Tøravi was a communal religious centre there for all *Tørar, with the religious specialist/cultic leader living in the neighbour village Gudby.107 There is an interesting parallel between Gudhaby—Tøravi and

103 Hellberg 1976; Vikstrand 2001, 387f. 104 Cf. Brink 1997, 430. 105 Åqvist 1996, 111. 106 See Bratt 2008, 126ff. 107 Hellberg 1976, 10f.; Vikstrand 2001, 316ff., 387f., 426. 184 CHAPTER 5

Gudhaby—*Valændasior noticed by Per Vikstrand. He states: “I båda fallen uppträder Gudhaby i kombination med ett ortnamn som antyder en koppling till en geografiskt avgränsad folkgrupp—tørar respektive valændar.”108 It thus seems as if the Old Swedish *gudhar sometimes took care of public cult for local or minor regional groups in the Mälaren region, such as the tørar and the valændar. They probably also watched over the public cultic sites of community type and the ritual structures there. In my opinion, these cultic leaders were probably not professional and specialized priests, in the strict sense of the term (i.e. “religious specialists”).109 Nor were they organized in a hierarchical priesthood institution appearing as an independent social stratum in society, normalizing cosmology or ritual practice. Admittedly the *gudhi carried a title indicating a sacred sphere, but sources suggest that it designated an office involving several leadership functions, including both religious and profane aspects.110 Most likely the *gudhi was some kind of a general chieftain in the settlement districts or regional formation, but at the same time he was also the local cultic leader. He thus assumed many functions in society, much the same as the pre-Christian goðar are described in the medieval Icelandic prose. He could also bear various titles at one and the same time. The Glavendrup inscription (DR 209) from Denmark is an example of this phenomenon:

. . . auft | ala . sauluakuþa | uial(i)þshaiþuiarþanþia | kn . . .

. . . in memory of Alle, gode of the Sølver, honour-worthy thegn of the uia-host. . . .111

Alle was not only the ODa goþi of the Sølver, he was also an honour-worthy thegn, i.e. a free man and successful ‘warrior, champion’ of high rank.112

108 Vikstrand 2001, 388. 109 See Sundqvist 2003a; 2003b; 2007. 110 See for instance Sundqvist 2007, 41f. 111 Moltke 1985, 226. 112 John Kousgård Sørensen explained saulua as “sal-præsternes” and linked it to kuþa “gode”, i.e. “sal-præsternes gode”. uia was associated with the word “priest”, but not as an accusative of Goth weiha m., but as an element *wīhaR which occurs in proper nouns translated as “pagan priest”. He suggested “sal-præsternes gode, præsternes hæderværdige tegn”. Kousgård Sørensen 1989, 17–26. Moltke 1985, 226. M. L. Nielsen (1998, 197) argued that saulua and uialiþs in the inscription “still have to be regarded as elements in the not fully interpreted appositions to Alli”. See also Düwel 2008. McKinnell et al. (2004, 118) suggest Ála sölva “Áli the pale”, but they seem to have forgotten the sequence kuþa in their interpretation. The Religious Leadership of the Sanctuaries 185

If we accept Jan Paul Strid’s interpretation of the runic inscription of the Karlevi stone (Öl 1), in Öland (c. 1000), we may also there meet a multifunctional leader designated *gudhi (ODa goþi).113 “Sibbi the goþi” is honoured after death with a complete dróttkvætt stanza, where he is praised as a great and upright sea warrior, who ruled “over land in Denmark”.114 Sibbi is here honoured with the kenning dólga Þrúðar draugr “executor, performer of the goddess of battles”, i.e. warrior or warlord. The milieu of place-names around the runic stone is interesting. The place-name Karlevi may be interpreted as “the sanctuary of the karlar”, i.e. the sanctuary used by the warriors, or “the sanctuary of all free men”.115 Not far from Karlevi is Torslunda. This place-name indicates a holy grove dedicated to the god Þórr. Perhaps Sibbi took part in the cultic activities at these cultic sites when he was not a warlord at sea. Sibbi may thus have fulfilled the double role of cultic leader and military commander.116 The inscription at Gursten (Sm 144), Småland, is also interesting in this context. The place-name derives from OSw Godhastæinn, i.e. ‘the stone of the *gudhi’ or ‘the part of †Sten that belonged to the *gudhi’.117 It is possible that a lost hamlet or farm close to Gursten was called †Sten, a name that could have denoted the hillfort situated in the vicinity. Otto von Friesen dated the inscription to the late ninth century.118 Kinander read and interpreted it thus:

(1) sunuR na | ut smiþ | a kata (2) uifrþaR . . . un . . . (3) kuþaskaki faþi

Sunr naut smíða Káta, Vé fríðar (s)un(ar). Guða-Skeggi fáði.

The son received Kåte Vifridsson’s forgings. Guda-Skägge carved (the inscription).119

The first part of the sequence kuþaskaki in line 3 can be interpreted either as the personal name Gudhi/Guþi in the genitive case or as a genitive form

113 Strid 1999, 101ff. See also Jacobsen & Moltke in DR 411: “kuþa kan opfattes . . . som akk. af subst. goþi, en gode; i så fald kunde lakunen fx. udfyldes uia- el. frus-kuþa, hellig- dommenes gode el. Frøsgode.” Brink (1999a, 431) accepts Strid’s interpretation. See also Hellberg 1976, 9, note 14. Critically considered by Marold 2000, 276. 114 See Jansson 1987, 134; Düwel 2008. 115 Hellberg 1986b; Strid 1999, 102. Per Vikstrand suggested a new interpretation: ‘(alla) de fria (öländska) männens vi’ or ’allmänt vi’. See Vikstrand 2001, 345; 2007, 206. 116 Strid 1999, 102. For another interpretation, see Vikstrand 2007, 206. 117 Hellberg 1979, 129. Cf. Brink 1999a, 427; Strid 1999, 104. 118 Von Friesen 1914. 119 My translation based on Ragnar Kinander, Sm 144, 292–296. 186 CHAPTER 5 of a common OSw noun (appellation) *gudhi. This personal name is not well documented.120 We may assume that the concept *gudhi here denotes a person with religious leadership roles.121 Perhaps Skeggi’s father was a *gudhi or he came from a family of *gudhar.122 Perhaps this family previously had military functions at the hillfort (Godhastæinn), beside their religious assignments. Whether Guða-Skeggi himself was a cultic leader is uncertain. The picture of the eastern Scandinavian *gudhar/goþar extracted from runic inscriptions thus harmonizes with the information about ON goðar in the medieval Icelandic prose texts. In these sources the goðar seem to act in multifunctional leadership roles as general chieftains. Most likely the cultic leaders in Svetjud (and Denmark) called *gudhar (ODa goþar) should also be regarded as general chieftains. It seems as if these chieftains could be related to local or regional sanctuaries, where they also organized the public cultic activities. Most of these sanctuaries were probably located at the chieftains’ farms, but some could perhaps be regarded as communal cultic sites.

It thus seems as if some kind of chieftains and political leaders took care of public sanctuaries in all three investigation regions. These chieftains were often designated with titles indicating a cultic function such as the ON goði (OSw *gudhi) and the female counterpart gyðja. My investigation indicates, however, that these chieftains in general should be considered as “cultic leaders”, i.e. they had other societal duties beside their religious tasks.

5.3 Kings and Earls as the Cultic Leaders of Sanctuaries in Trøndelag and Svetjud

It is entirely certain that the chieftains designated with the title goði (or the gyðja) never monopolized the public cult in Scandinavia. According to Old Norse narrative sources the king (konungr), and the earl ( jarl) in Svetjud and Trøndelag could appear in similar religious roles and take care of sanctuaries.123 It seems as if the kings and earls of Svetjud and Norway were related to specific chief sanctuaries at rulers’ farms, which had a regional or an inter-regional significance for the Svear and the Þrœndir, for instance Uppsala and Lade. There

120 It may appear in a few inscriptions. Sm 96; Vg 187; U 579. Ingrid Sanness Johnsen states that the proper nouns Góði and Guði, as well as the appellative *gudhi/goði are theoreti- cally plausible. Johnsen 1968, 162. 121 von Friesen 1933. 122 Källström 2007, 243. 123 Phillpotts 1912–1913, 267. The Religious Leadership of the Sanctuaries 187 was also, however, one “regional sanctuary” in Trøndelag which seems to have been a “cultic centre of the community type” (if we use Vikstrand’s terms),124 namely Mære. The cult there seems to have been organized by a cooperative of Þrœndir chieftains, at least during the second half of the tenth century. The organization of cult and cultic leadership at the inter-regional and regional sanctuaries of Uppsala, Lade and Mære, will be treated in chapter 14 in Part 4. In what follows I will just argue that the kings and earls in Trøndelag and Svetjud sometimes appeared as protectors and cultic leaders of the sanctuaries. They arranged great ceremonial feasts there, sometimes with support from other subordinated cultic leaders.

5.3.1 Kings and Earls in Trøndelag In one essay Preben Meulengracht Sørensen discussed the pre-Christian cultic leadership at the sanctuaries of Trøndelag.125 According to him, the Norwegian king and the earl were not involved as leaders or organizers of the public cult. Religious activities were completely controlled by the free farmers and their chieftains in this region during the Viking Age. In contrast, I will argue that the sources indicate that the king and the earl had important cultic roles at some sanctuaries as organizers and cultic leaders. In Hákonar saga góða 13–18 (see the text in ch. 4), for instance, it seems as if Earl Sigurðr and King Hákon were expected to take care of the public sacrificial feasts in Trøndelag, and even perform central ritual roles during these feasts. At the Frostathing the farmers said that it was their wish that King Hákon should make sacrifice to procure for them good crops and peace, as his father had done. Snorri also states that Earl Sigurðr sometimes defrayed all expenses himself for the sacrificial feasts at Lade. The idea that the earl played important roles in the religious sphere and commissioned the sacrifices may, for instance, be supported by the contemporary skaldic poem Sigurðardrápa (960 AD), which was quoted by Snorri and thus must be regarded as one of his major sources for the current passage (see ch. 4 above). Other skaldic poems may also support Snorri’s information that Norwegian earls and kings took charge of public cult in Trøndelag in a more general way. According to Vellekla 15–16 (990), which was quoted by Snorri, Earl Hákon of Lade maintained cultic duties as his father had done. He thus restored the sanctuaries which had been destroyed by the sons of Eiríkr. We read that Hákon “allowed the men of Þórr to uphold the plundered hof-lands and shrines of the gods”. By means of these actions and the cult, which seemed to be organized by Hákon, prosperity returned to the country. A similar cultic

124 Vikstrand 2001, 410–417. 125 Meulengracht Sørensen 1991b. 188 CHAPTER 5 role is also ascribed to King Hákon the Good in Hákonarmál 18 (c. 960) (see ch. 4 above).

5.3.2 In Svetjud Skaldic poetry reports that the Viking Age kings of the Svear had much the same duties as their Norwegian colleagues. In the ninth-century poem Ynglingatal, King Yngvi of the Svear is called vǫrðr véstalls, which may be interpreted as “the custodian of sanctuary altar”.126 The term vé has the meaning ‘cultic place, sanctuary’.127 Most likely there were ritual structures and/or buildings at such places, as indicated by the composition véstallr ‘sanctuary stand, altar’.128 The word vǫrðr signifies ‘watchman, keeper, custodian, guardian’. A similar type of epithet to vǫrðr vestalls also appears in connection with Norwegian rulers. Earl Sigurðr, according to Sigurðardrápa, was called vés valdr, for instance. Runic inscriptions from Sweden testify that such designations also really existed in eastern Scandinavian contexts. In the monumental inscription of Rök from Östergötland (Ög 136) (c. 800 AD), for example, the sequence sibi uiauari appears. It may be interpreted as Sibbi vīaværi, i.e. “Sibbi, the custodian of the sanctuaries”.129 Sibbi was not only a great ruler in the central settlement districts of Östergötland; he was also closely connected to certain vi sanctuaries. In a Viking Age runic inscription from Västergötland the name Vīurðr appears, which may be interpreted as a byname or designation of a ruler, ON *vévǫrðr m. ‘sanctuary custodian’.130 In several Swedish runic inscriptions the personal name (or byname) Vīseti (ON Véseti) appears. It may be connected with the appellation ON *véseti m. ‘the one who dwells at (or sits) or superintends a sanctuary’.131 Such nomenclature probably originally designated rulers who performed the double roles of political and cultic leader at the vi sites.

126 On this kenning, see Sundqvist 2002, 196ff.; 2007, 94. 127 The term vé is probably connected to the PG adjective *wīha- ‘holy’, preserved in Goth weihs, OHG wīh ‘holy’, and in Germ. Weihnachten. Vikstrand 2001, 298. 128 The manuscripts diverge regarding the second element of this word. The manuscript Kringla has vestallz ‘sacred stand’, which must be considered the most reasonable reading. See e.g., Finnur Jónsson 1931 (1913–16), 629; Åkerlund 1939, 91; Wessén 1964, 24, 64; Bjarni Aðalbjarnarsson, in Heimskringla I, Ísl. Fornr. 26, 42; Marold 1987, 66, 88. For another opinion, see Adolf Noreen 1925, 231. In Old Norse the word stallr (m.) has the basic meaning ‘stand’. In the compound word véstallr it acquires a sacral significance. Cf. Düwel 1985, 38; Hultgård 1993, 230ff. See also the poetic expression of heiðnum stalla “over the pagan altar”. Skj. A1, 110; B1, 105. 129 For interpretations of the sequence uiauari, see e.g., Peterson 1994, 73. 130 Vg 73. See Peterson 2007, 257. 131 See Peterson 2007, 257. The Religious Leadership of the Sanctuaries 189

Different types of sources indicate that the rulers of eastern Scandinavia protected, or even owned, the cultic sites and the ceremonial buildings there. Also archaeological sources indicate that rulers had a special relation to public cultic places and sanctuaries in the Late Iron Age Mälaren region. Several halls are located at central places or rulers’ farms with archaeological finds and structures indicating very high social status.132 At the halls of Uppsala and Adelsö/Hovgården there are contemporary royal and great burial mounds containing rich finds.133 Some of these graves suggest that the people buried in them belonged to the highest social stratum in Svetjud.134 At the halls of Vendel, Valsgärde and Uppsala there are boat graves indicating extremely high status; the same goes for extraordinary finds discovered at the hall of Helgö, such as gold foil figures, bracteates and exclusive glass. The theophoric place- names Fröstuna, Ultuna, and Torstuna also indicate that political power, organization of society, and cult, were closely interrelated in pre-Christian Svetjud. As mentioned above in chapter 2, the names in ‑tuna seem to designate places of central importance during the Late Iron Age, probably abodes for distinguished and powerful families. It is no coincidence that in the Mälaren area we often find banqueting halls and/or gold foil figures at the tuna sites, such as Fornsigtuna (and Sigtuna), Eskilstuna, Tuna in Vendel, Ultuna and Svintuna. A similar picture emerges when we consider the conditions of medieval owner ship. The Iron Age halls in the Mälaren region are frequently situated at places which in the Middle Ages reappear as crown lands and royal demesnes. This pattern can be seen, for instance, at Uppsala, Adelsö, Fornsigtuna, and Husby in Glanshammar.135 Some of these places were probably old manorial estates, which may have been taken over or confiscated by the king during the Viking Age and made into crown lands.

5.3.3 Rulers and Subordinated Cultic Leaders/Religious Specialists at the Chief Sanctuaries: A Hypothesis According to the narrative sources it seems as if the rulers, at different levels in the society, appeared with important ritual functions in public cults. They should therefor be referred to as cultic leaders. Evidence from toponymy

132 See e.g., Damell 1993. 133 The royal mounds of Uppsala used to be dated to the Migration Period. John Ljungkvist (2005; 2006; 2008) has, however, given good reasons for a later dating of the mounds, i.e. the Early Merovingian Period. The hall and other great buildings in Uppsala may be dated to the Late Roman Iron Age, Migration and the Merovingian Periods. Cf. Göthberg 2008, 77. 134 Arrhenius 1995. 135 See Rahmqvist 1986; DMS 1.2: 180; DMS 1.7: 218–236; Damell 1991; Ekman 2000. 190 CHAPTER 5

(such as the terms *gudhi, *vivil and perhaps *lytir), however, indicates the existence of religious specialists. How should this contradiction be explained? Is it possible to give a coherent picture taking into consideration both perspectives? Per Vikstrand has produced one suggestion. He argues that some kind of religious specialists may have been protected by and worked under the “great men” in the same way as other craftsman used to do.136 There is some support for this line of reasoning in available sources. It seems as if the rituals performed at the great public feasts were very comprehensive and complicated. Perhaps the rulers (such as kings and earls) needed some kind of specialists for different aspects and separate rites during the public sacrifice, although it was the ruler who had the overarching responsibility for the public sacrificial feasts.There are, however, very few authentic attestations of sacrificial rituals and the cultic performers are rarely mentioned (except the rulers).137 In Hávamál, st. 144, we find some formulations alluding to the sacrificial rituals:

veiztu, hvé biðia scal, veiztu, hvé blóta scal? veiztu, hvé senda scal, veiztu, hvé sóa scal?138

Perhaps the verbs biðia, blóta, senda and sóa reflect different aspects of the sacrificial ceremony. Klaus Düwel states regarding these terms: “Es wird sich nicht um synonyme Opferausdrücke handeln, sondern jeder Begriff dürfte eine bestimmte Art von Opferhandlung betonen”.139 It is probable that the sacrifice included ritual slaughter (ON sóa), performed outdoors at an Old Norse hǫrgr or stallr, i.e. structures of stones and wood. The Eddic poem Hyndlolióð st. 10 alludes to some customs performed at the hǫrgr. The stones had been reddened by ox blood during the sacrifice and turned into glass (nú er griót þat at gleri orðit) by repeated sacrificial fires (see ch. 9 below). The verb sóa is usually translated ‘to sacrifice, to slaughter’.140

136 Vikstrand 2001, 396. Such relation could be compared with the Greek and Roman patron- and-client-relationship. See e.g., Jackson 2014b. 137 Cf. Hultgård 1993. See further ch. 9 below. 138 “Do you know how to ask, do you know how to sacrifice, Do you know how to dispatch, do you know how to slaughter?” Trans. C. Larrington. 139 Düwel 1970, 234. Cf. Hultgård 1993, 221f. 140 Fritzner 1954 and de Vries 1977, 528. The Religious Leadership of the Sanctuaries 191

It appears twice in a ritual context in Hávamál st. 144–145 with this meaning.141 Also in Ynglingatal (st. 8 (5)) it gives clear religious-cultic associations, as it is combined with the adjective árgjarn ‘eager for a good year’.142 Some parts of the victims were given (senda; cf. gefa) to the gods and perhaps hung in a grove, an Old Norse lundr. When depicting the cultic feast at Uppsala Adam of Bremen states in Book IV,27: “The sacrifice is of this nature: of every living thing that is male, they offer nine heads, with the blood of which it is customary to placate gods of this sort. The bodies they hang in the sacred grove that adjoins the temple. Now this grove is so sacred in the eyes of the heathen that each and every tree in it is believed divine because of the death or putrefaction of the victims.” (see the Latin text in ch. 4 above). Adam’s information about sacrificial rituals in holy groves seems to be reliable. Evidence from ancient Scandinavia and other parts of Germanic area attests such cults.143 Some parts of the meat were cooked and consumed during the feast in the hall as a communion meal, where also the libations were made. The story referred to above, about Earl Sigurðr and Hákon the Good, from Hákonar saga góða ch. 14–18, reflects probably such a ceremony. In the texts these feasts are usually called Old Norse blótveizla (sg.). Perhaps the term blóta fits well in this context.144 For this reason the complicated sacrificial rituals during the great ceremonial feasts held at regional or inter-regional central places and royal farms most probably needed some kind of specialists. In my opinion, however, these religious officials were not permanently employed as “priests”. The terms in toponymy (e.g., *gudhi and *vivil) does not give sufficient empirical support for an exclusive religious office in ancient Scandinavia. Most likely these officials— who probably were recruited from chieftain families—had functions similar to the bailiffs and the king’s and earl’s representatives at the royal demesnes. They managed the crown lands and central places, and supported the “great men” in different spheres of society, both in political and judicial matters, as well as in religious duties. During the sacrificial feast, for instance, they had more specific religious functions and performed as cultic specialists, however, they were all subordinated the ruler (i.e. the king or the earl) who was the ultimate leader and organizer of the feast. These religious functions gave them a lot of social prestige and they were during the feasts in such contexts designated with honor titles, which originally (in most cases) were inherited within their families, such

141 In Hávamál st. 145 the skald states: Betra er óbeðit, enn sé ofblótið,/ ey sér til gildis giof;/ betra er ósent, enn sé ofsóit. In Hávamál st. 109 sóa plainly means ‘to slain or slaughter’. 142 See Sundqvist 2002, 248ff. 143 Hultgård 1996, 38f. and Sundqvist 2002, 128ff. 144 Düwel 1970, 226f. and 234; Sundqvist 2002, 249, note 61. 192 CHAPTER 5 as goði, þulr and *vífill. These specialists may to some extent be regarded as professional practitioners of the cult. The “great men” probably rewarded them for these assignments, for instance, by means of precious gifts and protection etc. This could perhaps be described as a patron-and-client relationship based on a redistribution or an exchange system which was favorable for both sides, i.e. “clientelism”. It could be suggested thus that in certain aristocratic milieus the king or the earl bore the ultimate responsibility for the sanctuaries and the sacrificial feasts held at these sites. He may have functioned as a commission agent for the feasts and he was certainly regarded as the custodian of the holy places (see ch. 8). When performing the great sacrificial ceremonies at the gatherings, he sometimes needed support from other cultic leaders/specialists, who were subordinated him during these rituals. Such cultic system and the co-operation between chieftains at (inter-)regional sanctuaries will be treated further in ch. 14.

5.4 The Financial System of the Public Cult at the Sanctuaries

As noted above, sources suggest that political leaders on different social levels, in all three investigation areas, managed and controlled the public sanctuaries. It seems as if these cultic structures or buildings were often erected on the farms of the rulers. This feature is probably present in all of Late Iron Age Scandinavia. But what was the driving force behind the erection of such public sanctuaries? One possible motive could be economic aspects and income for the rulers, because it seems as if tribute was paid to the sanctuaries in all three areas.

5.4.1 Tribute to the Sanctuaries of Svetjud and Trøndelag Adam of Bremen mentions that the Svear celebrated a common cultic feast in Uppsala, at nine-year intervals. At this feast people paid tribute to the temple:

It is customary also to solemnize in Uppsala, at nine-year intervals, a general feast of all the provinces of the Svear. From attendance at this festival no one is exempted. Kings and peoples all and singly send their gifts to Uppsala and, what is more distressing than any kind of punishment, those who have already adopted Christianity have to buy themselves out of those ceremonies.145

145 See the Latin text in ch. 4 above. Adam’s indignation has been interpreted as an indication of the accuracy of his statements. Hultgård states: “Adams upprördhet antyder att hans uppgifter är riktiga, även om man kan tolka denna möjlighet för de kristna att friköpa sig The Religious Leadership of the Sanctuaries 193

Adam tells us that kings (i.e. chieftains) and people were obliged to visit and send their gifts (dona transmittunt) to the “temple of Uppsala” during the major calendrical sacrifices. 146 Adam states that no one was granted exemption from these ritual feasts. This information may be compared with a tradition reproduced by Snorri Sturluson in Óláfs saga helga ch. 77, where the Uppsala sacrifices are likewise described: “At the time when heathendom still prevailed in Svetjud, it was an old custom there that the main sacrifices were held at Uppsala in the month of gói. Sacrifices were to be made at that time for peace and victory for the king. People from all Svíaveldi were to resort there.” Snorri uses the expression skyldi þá blóta; skyldu menn þangat sœkja; skyldi þar . . . vera þing etc. The verb skylda should be translated as ‘to be obliged, bound’.147 Both Adam and Snorri intimate that the assembly gathered at the same time, in February–March, as the sacrificial feasts at Uppsala. It is not unlikely that the ruler who was in charge of the sanctuary took tribute on this occasion.148 Perhaps this ruler was identical with the Uppsala king in the Early Viking Age, i.e. before the royalty converted to Christianity. In the eleventh century it might have been a pagan and conservative bailiff who “cared” for the “temple” in Uppsala. It is possible that the term Uppsalaauðr (OSw øþer) ‘Uppsala wealth’ was originally related to this tribute system. In Ynglinga saga 10, Snorri mentions that the Svear sent tribute to Uppsala, where the (euhemer- ized) god Freyr received lands and cattle. He also states: “This was the origin of Uppsalaauðr” (see ch. 12 below). Old Norse sources referring to Trøndelag indicate that participation in the pagan cultic feasts in the sanctuaries was compulsory. In Hákonar saga góða 14, Snorri states concerning the blótveizlur in this region:

It was ancient custom that when sacrifice was to be made, all farmers were to come to the hof and bring along with them the food they needed while the feast lasted. At this feast all were to take part in the drinking

från de religiösa festerna på ett annat sätt. Det är enligt min mening snarare ett utslag av en tolerant religionspolitik från sveahövdingarnas sida.” Hultgård 1997. 146 It should be noted that taxes and tributes in early medieval Denmark were always designated as “gifts”. See Poulsen 2012 and T. Zachrisson 2013. 147 Norrøn ordbok has “leggje på ein som skyldnad el. plikt, drive, tvinge, nøyde (e-n el. e-m) . . . ; binde ein til noko; leggje noko på ein; . . .”. Heggstad et al. 1993, 390. Fritzner has “drive, tilskynde, tvinge en til noget saa at han ikke vel kan unddrage eller undslaa sig der- for, undlade at gjøre det, paalægge en noget som hans pligt eller skyldighed, undergive en noget som hans skyldighed, som en byrde forpligtelse, der paahviler ham, forskylde noget, gjøre eller vise sig fortjent dertil”. Fritzner 1954, Vol. III, 401f. 148 Cf. Wessén 1924, 196; Brink 1990b, 56. 194 CHAPTER 5

of ale. Also all kinds of livestock were killed in connection with it, horses also; . . . .

In the passage above the verb skylda also frequently appears: “when sacrifice was to be made (þá er blót skyldi vera), all farmers were to come to the hof (at allir bœndr skyldu þar koma) and bring along with them the food needed while the feast lasted (ok flytja þannug fǫng sín, þau er þeir skyldu hafa, meðan veizlan stóð). At this feast all were to take part in the drinking of ale (at veizlu þeiri skyldu allir menn ǫl eiga).” Some of these expressions indicate that the farmers were obliged to take part in the rituals at the hof. It is possible that the public cults and ceremonial feasts performed by the Hlaðajarlar also involved economic aspects. They may very well have been part of an ancient system of redistribution. In exchange for “the ruler’s cult”, festivals, banquets and feasts, the people paid tribute or offered service to the earls. The occasion when Earl Sigurðr held his generous feast was something unusual and opposite to custom (i.e. the tradion where the farmers paid for the cultic feast themselves). There are also archaeological indications of this economic system at other cultic sites in the Mälaren region and Trøndelag. It has been argued that bread played an important role at the ceremonial feasts of Helgö.149 Rotary querns have been found close to the hall as well as corn placed in a round store-house (Rutenberge). Moreover, at the stone setting at this site, bread was deposited, probably as a gift to the gods. Close to the hall of Sanda too, wheat grains and rotary querns have been discovered. Archaeologists have argued that the seed that was ground there must have been brought in from somewhere else, since arable land in this area was limited.150 Perhaps it came as tribute from the farmers who intended to participate in the cultic feast of Sanda. Perhaps the grain at Helgö was also collected as tribute in connection with a cult.151 Gold foil figures have been discovered at the ceremonial buildings of Mære, Borg (Lofoten), Helgö, Husby in Glanshammar, Svintuna and the assumed hall at Eskilstuna. Several scholars have interpreted them as “temple coins”.152 In order to participate in the religious feasts and rituals people would have had to pay with these golden objects. In my opinion there are problems with this interpretation. The fact that some of these objects were discovered in post-holes indicates rather that they were sacrificed in connection with the

149 T. Zachrisson 2004a; 2004b; Hansson 2011. 150 T. Zachrisson 2004a; 2004b. 151 See T. Zachrisson 2004a; 2004b. 152 See e.g., Watt 1999; 2004; 2007. Critically considered by Ratke & Simek 2006. The Religious Leadership of the Sanctuaries 195 foundation of the house or that they had ideological implications (see ch. 11 below).153 Whether the foils should be regarded as “temple coins” is thus uncertain.

5.4.2 Tribute called hoftollr in Iceland A financial system similar to the one in the Mälaren region and Trøndelag seems to have existed in relation to the local sanctuaries in Iceland. The medieval sagas and Úlfljótslǫg report that the free landholding farmers (bœndir, sg. bóndi) and the legally recognized followers (þingmenn, sg. þingmaðr) of a chieftain (goði) in Iceland had to pay duties to their cultic leader (i.e. the goði) for the maintenance of the hof and the cultic activities which were supposed to take place there. This tribute was called hoftollr, and must be regarded as some type of sanctuary taxes or dues. The word tollr m. is often translated ‘duty, charge, fee, due’ (cf. tolla v. ‘pay’). Expressions such as gefa toll (hoftoll) or tolla gjalda til hofs often occur in the sagas. In the general description of Þórólfr’s hof mentioned in Eyrbyggja saga 4, for instance, we read: “Every farmer had to pay tax (tolla gjalda) to the hof. Another of their duties was to support the hof‑ goði in his mission, just as “thing-followers” (þingmenn) nowadays have to support their chieftains. It was the goði’s business to see to the hof and maintain it properly at his own expense, as well as to hold sacrificial feasts.” The farmers thus supported the hofgoði economically and politically. In return the hofgoði had to care for the sanctuary and hold sacrificial feasts. The system described in this text favoured the chieftains and the farmers in a reciprocal sense.154 The information found in Eyrbyggja saga, that the farmers had to pay dues for the maintenance of the hof, may originally derive from the gefa toll til hofs in Úlfljótslǫg (see above). The expression “as now they pay tithes to the Church” has made scholars suspicious. Olaf Olsen, for instance, argued that the clerical financial system of tithes (kirkjutíund) in Iceland was not based on a pre- Christian tradition, as intimated in this text.155 Clerical tithes were introduced there in 1096. Before this system was introduced the Christian priests had no regularly established compensation for their service (tollr); they had to rely

153 K. Lamm in J. P. Lamm 2004, 51. The posts have however been replaced several times in both the case of Uppåkra and Helgö. Hence, the gold foils could have ended up in the most recent of these rebuildings. These buildings are thus not constructed just once at the same spot, but several times which complicates the matter. 154 Many Icelandic sources report that the farmers had to pay duties to the sanctuaries and the hofsgoðar. Egils saga ch. 84, for instance, mentions Oddr, who was chieftain (hǫfðingi) in Borgarfjǫrðr, in western Iceland. He was also hofsgoði and cared for the sanctuary. All men paid him for that (. . . hann var hofsgoði ok réð fyrir hofi því, er allir menn guldu hoftoll til. . . . Egils saga, Ísl. Fornr. 2, 293). See further Wessén 1924, 196ff. 155 O. Olsen 1966, 43ff. Cf. Males 2010, 158, 219–222. 196 CHAPTER 5 entirely on occasional or even accidental gifts from the church owners or on alms.156 There was thus no “toll system” which the Church could take over from the pre-Christian cultic organization in Iceland. The information about a pre- Christian hoftollr was, according to Olsen, a learned postulation, construed as a projection back to pagan times, of the dues or tithes given to the Church.157 In my opinion, Olsen’s view is partly reasonable. The Church did not take over a financial system from the pre-Christian cultic organization in Iceland or elsewhere in Scandinavia. Tithes had a long history in Continental Europe, before Christianity was ever established in the Nordic countries. From Olsen’s line of reasoning, we cannot exclude, however, that a pre-Christian system existed in Iceland, with compulsory fees to the sanctuaries. There are some sources indicating that similar systems existed in other parts of pre-Christian Scandinavia. As noted above, such system seems to have existed in Uppsala if we rely on Adams’s information. Evidence from pre-Christian Continental Europe likewise indicates a similar financial system related to pagan sanctuaries among, for instance, the West Slavs. According to Saxo, there were ceremonial feasts among the Slavic tribes at Arcona. In the temple there was a cult image; Saxo states that each year every man and woman paid one coin to the cult of this statue, and one third of the war booty was given. All Slavs worshipped and paid tribute to this image. It also received gifts from neighbouring kings with no regard for the violation of Christian principles (e.g. King Sveinn of Denmark paid reverence to it).158 Saxo’s text thus intimates that this financial system was more or less compulsory for the West Slavs. The comparative materials cannot, of course, give reliable evidence about Icelandic conditions or whether a specific financial system of sanctuaries existed there. However, one thing may be stated at least; it is most plausible that cult involved economic aspects in most places in pagan northern Europe. Reliable sources indicate further that tribute, at least occasionally, was given to the pagan sanctuaries in Iceland, for their maintenance and activities. One trustworthy indication of this may be found in Ari Þorgilsson’s Íslendingabók

156 Olsen (1966, 47) supported this statement with a reference to Grágás I, 1, 20f. 157 “Hovtolden kan—som så meget andet i stykket om Úlfljóts love—være et vidste havde eksisteret før tiendens gennemførelse.” O. Olsen 1966, 48. Cf. Males 2010. 158 Nummus ab unoquoque mare uel foemina annuatim in huius simulacri cultum doni nomine pendebatur. Eidem quoque spoliorum ac predarum pars tertia deputabatur, ...... Hanc itaque statuam totius Sclavuie pensionibus cultam finitimi quoque reges absque sacrilegii respectu donis prosequebantur. Quam inter ceteros etiam rex Danorum Sueno . . . Saxo 14.39.7–8. Archaeological evidence from Groß Raden (ninth century), Ralswiek and Feldberg (eighth century), Wollin (tenth century) and Parchim (eleventh–thirteenth centuries) supports the existence of “sacrosanct temples” and “cult halls”. Written sources mention cults at Rethra and Szczecin. Slupecki 1993. The Religious Leadership of the Sanctuaries 197

(c. 1122–32), where it is stated that Grímr geitskǫr explored the whole of Iceland before the Assembly (alþingi) was established: “Each person gave a penny to him, and he gave that money afterwards to the hof buildings.”159 It is almost impossible to find out whether the system of hoftollr existed in pre-Christian Iceland, with compulsory fees regularly paid to the sanctuaries and the hofgoði. There is no clear evidence for this in the more reliable written sources. Since “sanctuary fees” seem to appear in other parts of Scandinavia, as well as among the West Slavs, it is reasonable to believe that some kind of tribute was paid to the hofgoðar in Iceland too. Perhaps the farmers brought animals or meat to sanctuaries during the public feasts. Some parts were consumed during the ceremonial meals, while other parts were given to the one who organized the feast. The evidence from Hofstaðir in Iceland may partly support such an assumption (see ch. 4 above).

5.5 Ideological Motives behind the Rulers’ Interest in Public Cult and Sanctuaries

It is quite possible that not only economic interests motivated the rulers to erect sanctuaries at their farms and organize public cult there. It may have been a matter of prestige and political power (see ch. 6 below), perhaps even linked with the leaders’ property right to the ceremonial building and the sanctuaries. Most likely this role of the political leader was also a strategy for gaining political authority, empowerment and legitimacy. The rulers were believed to have a specific relationship to the Other World; the kings and the earls were sometimes regarded as descendants of the gods, while the chieftains stated that they were “best friends” with the deities in cultic contexts (see ch. 3 above). Because of these genealogical and cultic relations, the farmers and other common people in general accepted or even expected that the ruler should be the one who took care of the sanctuaries, since he/she was the one who had these unique links to the Other World. Therefore they were also willing to pay tribute to the managers of the sanctuaries. The sources reflecting the transition period in Svetjud and Norway indicate quite clearly that the people, who still were pagans, expected and even required the first Christian kings to take part in the traditional and calendrical sacrificial feasts and to perform sacrifices on their behalf (see chapter 9 below). There was actually a reciprocal relationship and exchange system between the rulers and the farmers; the elite used the ceremonial feasts at the

159 En hónum fekk hverr maðr penning til á landi hér, en hann gaf fé þat síðan til hofa. Íslendingabók, Ísl. Fornr. 1,7. 198 CHAPTER 5 sanctuaries located on their farms as repayment for the tribute, while the farmers needed the cult of the ruler—he who had a close relationship to the gods—as a means of metaphysical protection and perhaps also as entertainment, including skaldic performances. The cult at the political leaders’ farms was a means of consolidating social structures. The rulers’ interest in public cult and sanctuaries could, on the other hand, also be regarded as a sign of a society with little specialization and division of labour.160 The Viking Age rulers appeared with a wide spectrum of leadership functions. A more stratified society with specialized occupations did not develop until the Middle Ages.161

It can be concluded that political leaders, on different social levels, in all three investigation areas, Svetjud, Trøndelag and Iceland, could be regarded as a kind of cultic leaders. In one or another sense these leaders had influence on or even controlled the public sanctuaries and the cultic activities that took place there. Often ceremonial buildings seem to have been erected on the farms of these chieftains and rulers. One reason for organizing the common cult at a public sanctuary could be economic interests. It seems as if tribute was paid to the rulers who organized and performed the cultic feasts at the cultic sites. This feature too appeared in sources from all three investigation areas. At the central sanctuary of Uppsala and perhaps at Lade, the income could be quite large, since all Svear and Þrœndir were according to these sources more or less obliged to attend the feast at their respective sanctuary. But there could also have been ideological motives behind the rulers’ interest in public cult and sanctuaries. Since these political leaders were believed to have specific genealogical and/or cultic links to the Other World, common people accepted and even expected that they were the ones who took care of the sanctuaries and the cult performed there. In Svetjud and Trøndelag it seems as if the earls and kings also constructed and took care of inter-regional sanctuaries and also organized cultic feasts, important for all Svear or all Þrœndir. These types of sanctuaries and feasts do not occur in the Icelandic context. The issue of regional differences related to cultic organization will be treated more thoroughly in chapter 14 below. Now we will turn the structure of the cultic sites and ceremonial buildings and their possible cosmic symbolism.

160 Sundqvist 2003a and 2007. 161 Lindkvist 1996, 221. chapter 6 The Ritual Structure of the Sanctuaries and Cosmic Symbolism

Ceremonial buildings and sanctuaries were thus not only a source of income for the ruling elite in the three investigation areas. They were also regarded as arenas where these groups could express political power. In this chapter I will argue that the ritual structure of ceremonial buildings and sacred sites in Svetjud, Trøndelag and Iceland sometimes displayed common features. In all three regions there were certain symbolic-ritual elements or objects which quite regularly appear in association with the ceremonial buildings. These elements played an important role during cultic activities, but were also strategies for gaining political authority and legitimacy. The ceremonial buildings and the sanctuaries must sometimes have been impressive in an immediate sense for people approaching them. Certainly they constituted the materiality of religious ideology and provided agency for the political leaders who owned or controlled them. They represented in a concrete way the presence of divine and political power in the human world.1 Some of these ritual structures found at sanctuaries may also have been regarded as symbols which reflected mythical or cosmic aspects, at least for those who had access to religious and cultural knowledge.2 Most of these ritual structures also occur in connection with “mythical buildings” and “mythical landscapes” mentioned in the Old Norse traditions. In my opinion, there may have been a micro-macrocosmic relationship between the real ceremonial buildings and the “mythical buildings”. The real cultic place—the house and the high-seat—was seen as a reduplication of the mythical landscapes. In that sense the cultic place was also regarded as a threshold to the Other World, i.e. a meeting place for humans and divine beings. The owner and user of such

1 Cf. Meskell 2005, 51. 2 Cf. Miller 2010, 73. The operational concept of cosmology (from Greek κόσμος (kósmos) ‘world, order’ and λόγος (lógos) ‘word, doctrine’) usually refers to ideas about the origin, the structure, and the order of the universe. It thus concerns the spatial-geographic structure and organization of the world. In a wider sense, the concept also refers to ways of thinking, human acts, and social phenomena. Cf. Bolle 1987; Zinser 1993; Lincoln 1986; 1999; Auffahrt 2001. For a critical discussion of the concept of cosmology among traditional people living in the North, see Eidlitz Kuoljok 1999, 9ff.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_007 200 CHAPTER 6 place had enormous power since he or she appeared at a sacred place that was related to the gods. The specific character of these sites was perhaps necessary for creating dignity and power. These conditions may be seen as a symbolic- ritual or cosmic strategy for gaining legitimacy, which may have been common to all three areas.

6.1 Cosmic Symbolism: Some Theoretical and Methodological Issues

The idea that ancient Scandinavian ceremonial buildings and sanctuaries sometimes reflect mythic or cosmic aspects is related to a general theory in the comparative study of religions, mainly elaborated by Mircea Eliade.3 He argued that temples, cultic places and towns in several societies mirror the structure and geography of the mythical world. Sanctuaries such as the temple of Jerusalem were imitations of a transcendent model. A similar symbolism is also evident in the early Christian basilicas and later in medieval cathedrals. From patristic times they were conceived as imitating the Heavenly Jerusalem, and at the same time they also reproduced Paradise or the celestial world. This kind of symbolism remained in the thought of Christendom. In the Byzantine church, for instance, the four parts of the interior symbolize the four cardinal directions. The interior church is conceived as a microcosmos of the universe. “The altar is paradise, which lay in the East. The imperial door to the altar was also called the Door of Paradise. During the Easter week, the great door to the altar remains open during the entire service; the meaning of this custom is clearly expressed in the Easter Canon: ‘Christ rose from the grave and opened the doors of Paradise unto us.’ The West, on the contrary, is the realm of dark- ness, of grief, of death, the realm of the eternal mansions of the dead, who await the resurrection of the flesh and the Last Judgement. The middle of the building is the earth.”4 Eliade also emphasized that the early churches and other sanctuaries were conceived as the centres of universe (axis mundi sg.), and regarded as hierophanies, i.e. manifestations of the sacred. At these centres the human and the mythical world meet. These sacred places are thus held to be the point of junction between heaven, earth and hell.

3 See e.g., Eliade 1974 (1949), 367–387; 1987 (1957), 58ff.; 1991 (1949), 6ff. Eliade built his theory on other scholars, such as Uno Holmberg [Harva] 1917; 1922; Edsman 1944. 4 Sedlamayr quoted in Eliade 1987 (1957), 61f. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 201

In modern theoretical debate, Eliade’s ideas about “the sacred” have sometimes been regarded as problematic.5 He has been criticized for following an extraordinary theological reasoning about the “religious man” (homo religio- sus), which cannot be verified. He has also been attacked for his unconditional universal perspectives and lack of consideration for specific cultural contexts. When making comparisons he always looks for similarities and disregards differences. A number of scholars also disagree with Eliade’s apolitical interpretation of cosmic myths.6 Some even argue that his theories about cosmic pillars as centres of the universe are not based on sound empirical grounds.7 In any case, most historians of religions agree that cultic sites often reflect a specific symbolism which may include references to the cosmic-mythical world.8 During the last few years Eliade’s ideas have been applied to Scandinavian conditions, not so much by historians of religions as by archaeologists and specialists in place-names, i.e. researchers who explore the ancient landscape.9 The archaeologist Lotte Hedeager, for instance, has argued in an innovative article that the process by which foreign objects from the outside world were acquired and transformed into “prestige objects” at Gudme, “the home of gods” on Funen, Denmark, was embedded in a cosmological system.10 In order to understand this cosmology, Hedeager searched for information in Old Norse mythical traditions found in, for instance, The Poetic Edda and Snorra Edda (c. 1220). According to Hedeager, the topographic model of Gudme was deliberately based on the paradigmatic structure of Ásgarðr as it is described in these Norse texts. The large hall of Gudme was arranged as a counterpart to Glaðsheimr (sic!), i.e. the mythical home of Óðinn.11 The small hall, located south of the great hall, represented the mythical building Vingólf, i.e. the hall of the goddesses. Beside the hall buildings at Gudme, there was also a workshop area. A similar structure may also be found at Ásgarðr, where a forge was

5 See e.g., Paden 1992, 84f.; 1996; McCutcheon 1997, 42ff., 74–100 and passim; Flood 1999, 5f.; Knott 2014, 492. 6 See Smith 1987, 17ff.; McCutcheon 1997, 42ff. 7 Smith 1987, 1–23. 8 See e.g., Widengren 1969, 328ff., 339ff.; Brereton 1987; Smith 1987. 9 See e.g., Hedeager 2001; 2011; Brink 2001, 90; Vikstrand 2001, 26f.; Andrén 2004; 2006; 2014; T. Zachrisson 2004a; 2004b; Jørgensen 2009, 332. But see also e.g., Drobin 1991; Drobin & Keinänen 2001; Nordberg 2003; 2008; Warmid 2004; Sundqvist 2004a; 2007. 10 Hedeager 2001; 2011, 137–163. On the place-name, see Kousgård Sørensen 1985, 133ff.; 1992, 233. Critically considered by Brink 2011. 11 According to Grímnismál 8, Glaðsheimr is not the house but rather the place where Óðinn lives. This place is the location of Valhǫll. See Simek 2006, 112. It is unclear whether Snorri refers to the house or the place when mentioning Glaðsheimr in Gylfaginning 14. 202 CHAPTER 6 located at Iðavǫllr. The lake situated in the western part of the Gudme settlement, and the springs connecting Gudme lake with Gudbjerg to the west and Galbjerg to the north, were related to the mythical wells called Urðr’s well and Mímir’s well. Hedeager concluded her analysis by stating that Gudme was “a replication of Asgard, its divine counterpart”. Hedeager’s investigation is interesting in several senses, and her cross- disciplinary approach of applying theories developed in the study of comparative religion to archaeological materials is very fruitful. She also points to important aspects which were overlooked by Eliade, namely that the sanctuary also had political and economic functions. Gudme was a site for the elite. The cosmic symbolism at this site can therefore be related to political motifs and be seen as part of the mechanism that creates authority. But there are also some problems related to Hedeager’s attempt. My major objection is her uncritical and too literary way of a reading of the Old Norse mythical texts, and particularly her method of using these texts directly as sources for the Late Roman Iron Age and Migration Period world view at Gudme.12 As far as I can see, she does not evaluate the individual texts to which she refers. When she construes Old Norse cosmology, for instance, she does not make a distinction between the testimony contained in Snorri’s prose and the poetic traditions. For Hedeager, Snorri’s words are an accurate reflection of pre-Christian myths. She does not take into consideration the fact that Snorra Edda is a later literary and Christian adaptation of older traditions. Snorri himself arranged his material in accordance with literary models and rhetorical devices that were common among learned writers in the Middle Ages. For a historian of religions, therefore, Snorri’s text is more unreliable as a source of ancient Scandinavian mythology than the information contained in the Eddic lays and skaldic poetry, which may have been composed in a Viking Age context. An example: Hedeager argues that the temple called Vingólf, ‘the friendly house’, was situated close to Óðinn’s hall in the mythical topography. This information appears only in Snorri’s text. In fact, Vingólf is never mentioned in the old poetic traditions. Historians of religions have argued that Snorri invented this name himself.13 Some scholars suggest that it is very young (“sehr jung”) and should be read as Víngólf, ‘wine house’, i.e. another name for Valhǫll.14 It is thus possible

12 For a more thorough and extensive criticism of Hedeager’s theory, see Sundqvist 2011b. 13 Simek 2006, 363. 14 See Braune 1889. It has been argued that the name could be read and interpreted as Vingólf ‘hall of friends’, i.e. the location of the warrior paradise, einherjar. Simek 2006, 363. See also Finnur Jónsson (1890), who criticized Braune’s interpretation and thought that the term was very old. He argued that it should be interpreted as ‘det venlige, hyggelige The Ritual Structure of the Sanctuaries and Cosmic Symbolism 203 that it was Snorri’s own idea that Vingólf was a separate house located somewhere near Valhǫll. His information on this specific place cannot be regarded as reliable within a Viking Age cosmological context, and is even less reliable within the Late Roman Iron Age and Migration Period world view.15 Nor does Hedeager go into the textual-criticism problems related to these texts. When scrutinizing the different manuscripts of Snorra Edda, one can see that the information on Vingólf varies. In the Snorra Edda manuscript known as the Codex Regius, for instance, it is said that the goddesses’ building, Vingólf, was a hǫrgr: Annan sal gerðu þeir, þat var hǫrgr er gyðjurnar áttu, ok var hann allfagr. Þat hús kalla men Vingólf (Cod. Reg. Edda).16 However, if we refer to the manuscript called Codex Upsaliensis, the hǫrgr in the same passage designates a structure inside the hall (salr), not the building itself (Annan sal gerðu þeir er hǫrgr var í . . ., Cod. Ups. Edda).17 It is thus unclear what type of building Snorri had in mind when mentioning Vingólf, as the medieval witnesses diverge in their descriptions. There are many other complex issues connected with the reconstruction of the ancient Scandinavian cosmology that Hedeager does not take into consideration. According to Snorri, the ash tree Yggdrasill had three roots, one of which went to Mímir’s well, one to Urðr’s well, and one to the well of Hvergelmir, which was in the realm of the dead, Niflheimr (Hel).18 This statement, i.e. that three roots extended to three wells, is not found in the Eddic lays. Perhaps this reflects Snorri’s own systematization. Most historians of religions believe that there was only one tree and one well, both of which went under different names.19 In addition, the dichotomy and dualism between

hus’ and that it referred to the goddesses’ temple. Kauffmann (1892) interpreted Vingólf as ‘the hall of lovers’, i.e. the place where the shield-maiden makes the deathless hero happy. 15 Snorri actually gives conflicting information about Vingólf. In Gylfaginning 3 Vingólf is an alternative name for Gimlé (Gimlé eða Vingólf ), i.e. the heavenly place where good people will live after Ragnarǫk (see Gylfaginning 3 and 5; cf. Vǫluspá 64). This information is not compatible with the idea that Vingólf was the temple of the goddesses. See also Snorri’s Edda (Ed) Faulkes 1988, 176. 16 Gylfaginning 14. (Ed) Faulkes 1988, 15. 17 Gylfaginning 13 in The Uppsala Edda. (Ed) Heimir Pálsson 2012, 24. In the Codex Upsaliensis the name was Vinglóð, which makes the issue even more confusing. 18 Gylfaginning 15. 19 De Vries 1956–57, §583; F. Ström 1985, 97; Nordberg 2003, 171; Simek 2006, 205. 204 CHAPTER 6

Ásgarðr and Útgarðr as identified by some scholars (including Hedeager),20 has recently been questioned by other scholars.21 Hedeager’s reconstruction of the pre-Christian cosmology is thus, in my opinion, sometimes uncertain and not based on the methodology conventionally applied by philologists and historians of religions. There are additional problems attached to her assertion, which are also related to the sources. Extensive accounts of cosmology are not preserved in southern Scandinavian traditions. They appear only in Old Norse texts written down in thirteenth- century Iceland. We are uncertain of their prevalence outside of Iceland, or whether they ever existed in Roman Iron Age Denmark. In my opinion, there was never a centralized institution or organized priesthood which could have formulated normative world views and ritual practices over all of Scandinavia. On the other hand, there might have existed mythical and ritual themes which were spread over wide geographic areas and were stable over time. Some notions may have been transmitted by aristocratic groups who made contact with each other via marriages, gift-giving systems and trade. Perhaps we may ascertain such mythic-cosmic aspects embedded in the landscape by using a method where we not only apply the problematic mythic texts but also compare the ritual structure at Gudme with conditions at other cultic sites reflected in different types of sources. By comparing several known Iron Age cultic sites appearing in different types of source materials, we may see some recurrent elements and features which constitute certain configurations.22 In my opinion, these configurations may reflect parts of a “common cultural model”, which was used when constructing a cultic site. These empirically visible configurations may also in some senses relate to a common cosmology, including general ideas of how mythical and human cultic sites were structured and conceived. Most likely materiality gave form to the mythical world in the first place,23 but subsequently ideas of the mythical sanctuary could have been projected on ritual structures in the real world and thus served as a prototypical model of a cultic site. In what follows, the ritual structures and features of ceremonial buildings in the investigation areas will be presented by means of three categories or themes: (1) the locations, architectures and decorations of these buildings; (2) their interiors; and (3) the external ritual structure of the ceremonial

20 See e.g., Gurevich 1969; Meletinskij 1977; Hastrup 1985, 147. 21 See e.g., Clunies Ross 1994, 48–56; Brink 2004, 292–298; Ney 2006; Vikstrand 2006. See also the critical remarks of Sebastian Cöllen 2011 and especially by Løkka 2010. 22 Sundqvist 2004a; 2007; 2011b; cf. Söderberg 2005; Jørgensen 2009. 23 Cf. Miller 2005; Henare et al. 2007; Engelke 2012. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 205 buildings, i.e. the entire sanctuary. In each category I will discuss whether the ritual structure of these buildings and places was constituted according to a common model. I will also relate the results of these investigations to structures appearing in the mythical landscape. It will be argued that the structure of the mythic-cosmic world was deliberately projected on to real sanctuaries in the investigation areas. By projecting cosmic symbols and structures on the human ceremonial buildings, these buildings were given form and meaning.24 This cosmic symbolism had several functions. It guaranteed the presence of the divine powers at these structures. In a ruler-ideological perspective it was also important to show that the political rulers who owned and controlled these sacred places were closely related to the divinities and the mythic landscape. When entering his feasting hall at the sanctuary the ruler appeared on a stage, which was surrounded by mythic-cosmic symbols, indicating that his power came from a source beyond this world.

6.2 The Location, Architecture and Decoration of Ceremonial Buildings

In the history of religions it has often been noticed that religion is often integrated with political factors. The construction and location of ceremonial buildings are therefore not only determined by religious motifs, but also by a deliberate choice of localities, architecture and rituals, intended to symbolize, authorize and legitimize political power. I will first concentrate on the localities and architectures of ceremonial buildings in the three investigation areas, and look for possible references to and/or similarties with the cosmic-mythic world.

6.2.1 The Elevated and Imposing Ceremonial Buildings in Svetjud In scholion 139 Adam of Bremen presents the “Uppsala temple” in detail: “A golden chain goes round the temple. It hangs over the gable of the building and sends it glitter far off to those who approach, because the shrine stands on level ground with mountains all about it like a theatre.” This information has been viewed with scepticism by scholars. They have suggested that the temple of Solomon mentioned in the Old Testament may have influenced Adam’s description, for instance, when he describes the golden chain of the

24 It is a well-known fact that myths often function as a norm or prototype for human con- duct and behaviour. See e.g., Malinowski 1948. Cf. Honko 1972; Segal 1998. 206 CHAPTER 6 cultic building.25 Anne-Sofie Gräslund argued that Adam’s description of the “temple” may have been affected by “house-shaped” reliquaries, which had a chain between the gables, by which they were carried.26 I agree with Gräslund and others that Adam was influenced by such secondary elements and ideas when describing the ceremonial building in Uppsala, but there are some elements in Adam’s description which may be regarded as more genuine. For instance, the idea that the ceremonial building at Uppsala was visible from quite a long distance to people approaching the sanctuary may be reliable. Archaeology indicates that the halls and cultic houses of Svetjud were topographically located in the wide valleys with clay sediment around Lake Mälaren, where conditions for agriculture are good. Most of them were erected at central places near lines of communications, especially waterways. The halls were built on artificially made terraces or in topographic positions expressing loftiness.27 The hall buildings were thus elevated in the landscape and visible from public roads. A good example is the large buildings in Uppsala. At “Kungsgården” to the north of the church, two plateaus are built of clay. The southern plateau is about 2.5 m high; on it stood the hall during the Merovingian Period (see fig. 2b). In the summer of 2011 a Merovingian Period three-aisled building was also found on the northern plateau, perhaps an exclusive workshop. A third presumed large building, situated on the “church plateau”, was in an exposed position in the surrounding landscape. Every traveller, particularly from the north, must have seen these buildings and the mighty royal mounds from far away. The elevated position of the ceremonial buildings may also be seen at Adelsö, Birka and Fornsigtuna. The terrace carrying the presumed hall at Hovgården (Adelsö) was about 1–2 m high and located furthest out on an isthmus at Lake Mälaren.28 The hall at Birka was also exposed on the upper terrace at “the Garrison”.29 This house must have been easily visible from the fairway at Björkö and must have been an impressive sight for people sailing by. The topographic position was also essential in Fornsigtuna.30 The dominating plateau with the foundations of the house, the “King’s hall”, was situated furthest out on a cape close to the fairway at the Håtuna bay. The plateau carrying the largest hall, “the Garrison”, was situated in a strategic position, somewhat separate

25 See Hultgård 1997, 17. 26 A.-S. Gräslund 1997, 108f. 27 Cf. Herschend 1998, 39−40. 28 Brunstedt 1996. 29 Holmquist Olausson & Kitzler Åhfeldt 2002. 30 Damell 1991; 1993. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 207 from the rest of the settlement.31 The halls at Helgö are likewise situated on terraces and thus visible from afar (see ch. 4 above and fig. 12B below). The dwellings of the political and religious elite were thus associated with an elevated position in Svetjud. The majestic impression of these houses was also reinforced by means of architecture and size. These houses were large, had three aisles, and a roof carried by powerful posts erected in pairs. In Old Uppsala the hall on the southern plateau was about 50 m long and 12 m wide. One of the post-holes (G) was 1.47 m deep and at the surface some 50–90 cm in diameter.32 This indicates that the house was not only long but also high. Two of the post-holes (9 and 12) discovered by Sune Lindqvist under the present church had a diameter of about 85 cm. No. 12 was as deep as 1.25 m.33 It has been suggested that these holes may have belonged to a building, which thus must have been high. One of the largest halls in the Mälaren region was found in Lunda. It was 50 m long and in a topographic position which must have overawed those who approached the building. One of the halls in Fornsigtuna (the Garrison) was huge as well, some 40 m long and 10 m wide. There are not so many detailed descriptions of the architecture of the ceremonial buildings in the medieval written sources referring to Svetjud. However, most of the textual evidence emphasizes that the pre-Christian buildings in Uppsala were large and imposing. In Ynglinga saga 10, Snorri says, for instance: “Freyr erected a huge hof at Uppsala and made his chief residence there.”34 That Uppsala had one or several large and impressive halls is mentioned several times in Old Norse and medieval Latin texts. Ynglinga saga 36 says that Ingjaldr built a hall when his father Ǫnundr died: “He had made ready a hall in no wise smaller or less stately than the [hall] Uppsalr, which he called the Hall of Seven Kings.”35 There he held a great funeral feast (erfi) for his father.

31 At many other sites in this region we may find possible hall buildings on platforms and high positions in the landscape, see e.g., Granby-Hyppinge in Vallentuna, see Skre 2007a, 230; at Karvik in Bromma, see Ringsted 2005; Skre 2007a, 230; the cultic house at Alby in Botkyrka parish, Södermanland, see Hjulström 2008; at Husby, Glanshammar, see Ekman 2000; at Vendel (Tuna) in Uppland, see Arrhenius 1998, 185; and at Runsa in Uppland, Olausson (ed.) 2011 and Olausson 2014. The hall at Barksta, Västmanland, was also very high. Karlenby 2005. 32 Hedlund 1993, 66. 33 Nordahl 1996, 47–50. 34 Freyr reisti at Uppsǫlum hof mikit ok setti þar hǫfuðstað sinn, . . . Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 23. Trans. Hollander, modified. 35 Hann lét búa sal einn, engum mun minna eða óvegligra en Uppsalr var, er hann kallaði sjau konunga sal. Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 66. Trans. Hollander, modified. 208 CHAPTER 6

The traditions about King Aðils mention a great hall or maybe a cultic building at Uppsala. When describing Rolf’s visit to Aðils in Uppsala, Snorri used the expression konungs hǫll,36 while Arngrímur Jónsson’s translation of Skjǫldunga saga refers to a domus magnifica, “magnificent house”,37 and Saxo to Atisli forte penates “Aðils’ strong house, hall”.38 Snorri also mentioned that Óðinn “erected a large hof in Fornsigtuna and made sacrifices according to ancient customs among the æsir”.39 The thirteenth-century authors thus believed that great and impressive ceremonial buildings were erected in Svetjud in pre- Christian periods. Archaeological evidence indicates that they surely were right on that point. The halls and cultic buildings in Svetjud were probably not only large, they were also decorated. When Adam described the “temple” in Uppsala he stated that it was “all fitted out with gold” (quod totum ex auro paratum est). Scholion 139 mentions that “a gold chain surrounds the temple” (catena aurea templum circumdat). It hangs, shining, from the gables of the building, visible to all who approach it. Adam’s description, however, may have been “enhanced” (not to say distorted) by literary elements. Anders Hultgård has argued that the text may have been affected by two classical literary figures, evidentia and ekphrasis (Latin descriptio). Both sharpen a description. They would, for example, exag- gerate the impression of a work of art, or of a building. Their purpose was to make the narrative vivid and clear, and to encourage the reader’s imagination. On the other hand, it is also conceivable that ceremonial buildings were in fact decorated with gold and precious metals. There is archaeological evidence for this: for instance, molten gold was discovered at one of the post-holes of the Migration Period hall at Gudme. According to Frands Herschend, the posts at Gudme might have been decorated with gold, but a fire had caused the gold to melt and run down into the post-hole.40 The large Merovingian Period hall on the southern plateau in Old Uppsala was also decorated with spirals and perhaps heads of slaughtered animals.41

36 Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 56–59. Snorra Edda, Skáldskaparmál ch. 44. 37 Skjǫldunga saga, Ísl. Fornr. 35, 30f. 38 Saxo 2.6.3. 39 [Óðinn] gerði þar mikit hof ok blót eptir siðvenju Ásanna. Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 16. 40 See Herschend in A.-S. Gräslund 1997, 108−109 and Petersen 1994, 36. 41 Personal communication from Per Frölund. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 209

6.2.2 Strategic Topographic Positions of Ceremonial Buildings in Trøndelag The cultic sites of Trøndelag, which may have included cultic buildings and halls, are mostly situated along the water systems of the Trondheimsfjord, where several lines of communications ran. The cultic site of Lade, with the assumed hof, for instance, was situated at the mouth of the river Nið (Nidelva) on a peninsula reaching out to the Trondheimsfjord. It thus had a central location in relation to important communications on the water systems reaching into the inner parts of Trøndelag. Exactly where the assumed ceremonial building was located is not known. Presumably it was situated in the area around the present church. It had an exposed position in the landscape in more ancient periods. No extensive archaeological investigations have been made there, however.42 The ceremonial building of Mære was located on the top of a hill called “Mæreshaugen”, east of the Borgenfjord, a shallow fjord arm separating Inderøya from the mainland (fig. 8A). The slope down to the fjord is gradual but on the north and east sides of the hill the ground drops steeply to flat land, which today is cultivated but in prehistoric time was bog. It has been argued that the sea surrounded “Mæreshaugen” in the prehistoric period, which thus constituted an island in the fjord.43 The ceremonial building was placed on the south-east side of the summit, very near the highest point, 67 metres above the fjord, at a place which slopes both southwards and westwards. This building must have been impressive from the sailing route down in the fjord, manifesting its loftiness. Its location must be regarded as central in relation to the surrounding settlements in Sparbyggvafylki. It was also a central place for the whole region, as the Borgenfjord is connected to the Trondheimsfjord, and thus it is possible by means of the water system, to reach the coastal areas and the settlements in Ut-Trøndelag from Mære. The Iron Age hall at Hovde, Ørland (c. 290–550) was located at the strait between Ørland and Agdenes, the entrance to the Trondheimsfjord from the sea.44 This site had a strategic topographic position on the flat mainland far- thest out towards the sea. The buildings there were clearly visible from the sailing route on the sea and at the mouth of the Trondheimsfjord.45 Also visible from the sailing route was the cultic site at Ranheim, with the assumed hǫrgr and hof. It was located close to the Trondheimsfjord, roughly 9 km east

42 See further Brendalsmo 2006, 320. 43 Lidén 1969, 5; Røskaft 2003, 140. 44 Grönnesby 1999. 45 E-mail from Geir Grönnesby, 16 June 2009. 210 CHAPTER 6 of Trondheim.46 The farm with the assumed cultic site at Værnes most likely had a strategic location as well, between the Trondheimsfjord and the mouth of Stjørdalselva, i.e. between important lines of communication.47 Similar conditions may also be seen at the farm and the assumed cultic site of Melhus. The modern name Melhus is a reduced form of an Old Norse name Meðalhúsar ‘the middle farm’ (‘midtgarden’).48 It has been argued that this name is secondary to a more original name of the farm. When the farm was split up, the new name Meðalhúsar appeared, since it was situated between two other farms. The name itself may thus indicate centrality in the settlement district. The locations of the three fylki churches—in Alstadhaug, Sakshaug and Haug—have all been assumed to have been pre-Christian cultic sites before Christianity arrived in this region.49 Alstadhaug and Sakshaug especially have impressive situations in the landscape, high up on hills, and visible from the sailing route far down in the Trondheimsfjord. The hof places where ceremonial buildings were probably erected seem in general to have had central locations in the settlement districts of Trøndelag. They are situated in fertile plains at the bottoms of valleys where rivers run.50 For instance, the farm of Hov in Ålen has this central position by the river Gaula in Gauldalen. It was placed in one of the fertile plains around the river. Hov in Soknedalen also has this central position in the settlement district. It is located in the upper parts of the river Sokna, where fertile plains surround it. Hove in Meldal has a similar location at the bottom of the valley and right beside the river Orkla. This farm belonged to the central settlement of the Orkladaføret.51 The farm Hov in Orkdal is strategically situated at the mouth of Orkla and just beside the Trondheimsfjord. Finally, Hove in Selbu is located in the bog plains by the lower parts of the river Nea, and close to the lake Selbusjøn, the natural centre of the settlement. Ceremonial buildings with impressive topographic situations are known from Late Iron Age high-status settlements elsewhere in Norway. Borg in Lofoten, for instance, has a great topographic location.52 The impressive main building (I:1a) with the “hall room” (C), dating to the Viking Age, was situated

46 Rønne 2011, 4. 47 Røskaft 2003, 112. 48 Sandnes & Stemshaug 2007, 309. 49 M. Olsen 1926, 262ff. Cf. Brendalsmo 2006, 138. 50 Røskaft 2003, 54. 51 Most likely the name Meldal (ON Medaldalr) should be seen in this context. Røskaft 2003, 63. 52 Munch et al. (eds.) 2003. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 211 on the top of a low ridge, at the highest point of the site.53 At exactly the same spot, the large Merovingian Period house (I:1b) had been erected earlier. The tradition of building halls on plateaus cannot be seen in the archaeological sources from Trøndelag or northern Norway, however. Neither the cultic house at Mære nor the Iron Age hall at Hovde was built in this way. In neither case, however, was this necessary in order to express loftiness or an exposed position, since the topographic conditions were otherwise so favourable. This is also the case of Borg in Lofoten. As far as I know, only one hall found in Norway was built on a plateau, namely the hall at Huseby, Tjølling, close to Kaupang in southern Norway.54 This plateau was 36 m long and 13 m wide, and had bowed long sides. It was situated on the crest of a rocky hillock and it was widely visible.

6.2.3 Large Ceremonial Buildings in Trøndelag? As noted above, there is not much archaeological evidence of ceremonial buildings in Trøndelag. However, remains of an assumed Viking Age ceremonial building with huge post-holes have at least been found at Mære. The form of this house, its dimensions and orientation are unknown, however. The only traces of it were remains of a “palisade wall”, lumps of burnt clay with twigs and branches, fire-cracked stones and post-holes with remains of timber. The post-holes formed irregular circles with diameters varying from 80 to 100 cm.55 The dimensions of the post-holes indicate that this house was quite high. Since it also was situated on a hill it must have been impressive to people approaching it. The three-aisled hall building at Hovde, in Ørland, was erected as early as the Pre-Roman Period. This building (House D) was 27.9 m long in its first phase. It was distinguished from the other houses at this site, since it was built of large posts, up to 50 cm in diameter, and it had a clay floor. Most likely this building too seemed quite high and impressive to people coming sailing in the strait between Ørland and Agdenes. In the next phase (200–500) it seems as if House A took over the functions of House D from the previous phase. It had about the same dimensions as house D.56 Opposite the hall building at Ørland, the assumed hof sanctuary at Ranheim was quite small. The Ranheim house measured 5.3 × 4.5 metres.57

53 Herschend & Kaldal Mikkelsen 2003. 54 Skre 2007a, 223ff. 55 Lidén 1969, 12. 56 Grønnesby 1999; 2000. 57 Rønne 2011, 7. 212 CHAPTER 6

In other parts of Norway, more impressive banqueting halls have been discovered. The most interesting parallel to the ceremonial buildings of Trøndelag is the hall building at Borg, Vestervågøy, since it also included gold foil figures.58 The Migration Period hall (I:1b) was 64 m long and 7–8 m wide, while the Merovingian and Viking Period house (I:1a) was roughly 83 m long and 7.5–9 m wide. This latter hall is the most impressive multifunctional hall building hitherto found in Scandinavia. Elsewhere in Norway, large Iron Age halls have also been found, for instance at Huseby, Tjølling in Vestfold. It was 35 m long and 7.9 to 11.7 m wide and located on a terrace.59 The written sources do not provided much guidance about the exterior character of the ceremonial buildings of Trøndelag. Snorri Sturluson comments on the hof at Lade in, for instance, Hákonar saga góða 17, where he describes it as a large banqueting hall where ceremonial feasts took place. Óláfs saga Tryggvasonar (manuscript AM 310 4to), of Oddr munk, reports that Earl Hákon had a great hof sanctuary (mikla hof) at this site, with plenty of divine images.60 Njáls saga 87 states that the greatest hof sanctuary in Norway was located at Lade. If we combine the information found in Old Norse traditions, we get the impression that the Viking Age hof of Lade was a large hall, intended for great public gatherings and ceremonial feasts. Some sources cannot be regarded as historically reliable concerning these matters. In Egils saga 11, for instance, there is a description of a great, but temporary banqueting house, located on the island of Torgar in Hálogaland, where Þórólfr Kveld-Úlfsson held a feast for Haraldr Finehair:

There was a large barn (kornhlaða) that Þórólfr had fitted out with benches so that the drink could be served there, because he did not have a room large enough to accommodate that number of people [800 people in total]. Shields were mounted all around the building. The king sat in his high-seat, and when the upper and lower benches were both filled . . .61

Preben Meulengracht Sørensen is probably right when he argues that this description of the banqueting building has a literary function in the saga narrative.62 It indicates that Þórólfr was eager to demonstrate his own power against

58 See mainly Johanson & Munch 2003; Herschend & Mikkelsen 2003; Munch 2003a and 2003b. 59 Skre 2007a. 60 Óláfs saga Tryggvasonar eptir Odd munk Snorrason. Ísl. Fornr. 25, 193. 61 Egils saga, Ísl. Fornr. 2, 28f. Transl B. Scudder. 62 Meulengracht Sørensen 2003. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 213 the king, by showing that he had more men than the king at his disposal in his large (temporary) hall building. In this case the saga description of the feasting house may not be used as a historical source, since it has a literary function. The archaeological evidence from Borg nevertheless indicates that Viking Age ceremonial buildings may have been quite large in Norway. It seems as if the size and character of hall buildings in Trøndelag and elsewhere in Norway varied a lot. The large hall in Borg, Lofoten, was most likely an enormous status symbol. Whether such impressive ceremonial buildings also existed in Trøndelag is uncertain, but the archaeological finds at Hovde, for instance, indicate that a similar type of three-aisled hall buildings also existed there in the Middle Iron Age. Even though this house was not as big as the one in Borg, it must have been awesome for people passing through the strait at the mouth of the Trondheimsfjord. The large post-holes of the Viking Age house at Mære, and its location on the top of the hill, indicate that this house too must have appeared splendid to those approaching it.

6.2.4 Large Ceremonial Buildings at Strategic Locations in Iceland The Sagas of Icelanders report that hof buildings in Iceland were sometimes large and situated at the chieftains’ farms. According to Eyrbyggja saga, for instance, Þórólfr’s hof at Þórsnes was a large house (mikit hús) with a door in one of the side walls near the gable (see ch. 3 and 4 above). It seems as if this building had some kind of a spatial division. Some parts of it, such as the space beyond the high-seat posts and the afhús, were regarded as friðstaðr ‘protected areas’, i.e. a sanctuary, while other parts seem to have been regarded as more profane space. Perhaps this house could be classified as a multifunctional hall located on the chieftain’s farm itself. The laconic description of Þórólfr’s hof in Landnámabók likewise indicates that his sanctuary was large: “There he [Þórólfr] erected his home and built a large ceremonial building (hof mikit), and consecrated it to Þórr, and now the place is called Hofstaðir.”63 Other sagas report that some hof buildings were large. Kjalnesinga saga 2 (thirteenth century), for instance, mentions that the chieftain (goði) Þorgrímr had a hof erected in his courtyard (í túni sínu) in south-western Iceland.64 This house was 100 feet long and 60 feet wide (þat var hundrað fóta langt, sextugt á breidd). This big building too could be interpreted as a feasting hall. This was also the case for Ingimundr inn gamli’s hof at Vatnsdalr, in north-western

63 See Landnámabók S85, H73, Ísl. Fornr. 1, 125. See also above. 64 Kjalnesinga saga, Ísl. Fornr. 14, 7. 214 CHAPTER 6

Iceland, since Vatnsdœla saga 15 states: “He built a great hof a hundred feet long.” It was built at his homestead and “he called the farm Hof”.65 Whether the saga authors are reliable when describing the great size of the hof sanctuaries is uncertain. As noted in chapter 4, there is not much archaeological evidence of large ceremonial buildings in Iceland. Apart from the hall building (38 m in length) on the chieftain farm of Hofstaðir at Mývatnssveit, there are few finds indicating large long-houses in Iceland. The normal halls/ long-houses in Iceland fall between 12 and 16 m. The hall at Aðalstræti, Reykjavík, for instance, was about 16 m in length. Also very small is the late tenth- century hall found at Sveigakot, Mývatn. This hall is a typical Scandinavian Viking Age hall, but in miniature, with concave long walls, a three-aisled design and a centrally placed hearth.66 The floor area was 7.3 m by 3.2 m. The fact that Icelandic halls/long-houses in general are small should be taken into consideration when assessing the house at Hofstaðir as a monumental or large house.67 The internal area of Hofstaðir II was almost twice as big as the areas in the houses of Aðalstræti, Grelutóttir, Granastaðir, Eiriksstaðir, Vatnsjǫrður, Goðatættur II, Sveigakot and Hvítarhólt IX.68 The Late Iron Age halls in Svetjud and Trøndelag are much grander than the hall buildings of Iceland. The archaeological evidence indicates, at any rate, that the picture painted by the medieval saga scribes of large halls and cultic buildings in the Viking Age was not totally distorted. There are also some more realistic descriptions in the sagas of the architecture and size of Viking Age feasting halls in Iceland. It seems as if the medieval transmitters had sometimes witnessed such buildings themselves. One such description is rendered in Fóstbræðra saga 13, which is dated to the second half of the thirteenth century. The text tells us that Þorgeirr Hávarsson was sent on a mission to Iceland by King Óláfr Haraldsson c. 1020, in order to avenge one of the king’s retainers who had been hurt by a man called Þórir. Þorgeirr arrived at Reykjahólar in western Iceland and engaged a carpenter called Vélagr and together they built a feasting hall (skáli): “[they] each built one side of the skáli. It was planked on the long walls, but did not have any other plank-walls; those planks lasted right up to the time when the first Bishop Magnús was at Skálholt.”69 This skáli was thus built c. 1020 and it

65 . . . hann reisti hof mikit hundrað fóta langt . . . Bœr sjá skal heita at Hofi. Vatnsdæla saga, Ísl. Fornr. 8, 42. 66 Orri Vésteinsson 2009, 495. 67 Lucas 2009, 376f. 68 Lucas & McGovern 2008, 21 and Lucas 2009, 376. 69 . . . ok gerði sínum megin hvárr þeira skálann. Skálinn var um endilangt þilinn, en eigi ǫðrum þiljum; þau þili héldusk allt til þess, er Magnús byskup var at staðnum í Skálaholti inn síðari. Fóstbræðra saga. Ísl. Fornr. 6, 184. Trans. Meulengracht Sørensen 2003. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 215 remained at least until Magnús Gizurarson became bishop at Skálholt in 1216. In the Hauksbók redaction of Fóstbræðra saga, dated to the early fourteenth century, the hall is described with the present tense of the verb: “it is nine- teen ells and forty ells long; it still stood there when Bishop Árni inn síðari was consecrated bishop at Skálholt.”70 The present tense of the verb (er) indicates that the scribe saw the building with his own eyes and apparently measured it.71 (Árni Helgason was bishop 1304–1320.) The hall at Reykjahólar was about 24 metres long and 11.5 metres wide, i.e. quite a large building comparable to Hofstaðir I.

6.2.5 The Great Hall at Hofstaðir, Mývatnssveit Hofstaðir was probably an important chieftain site in the society of the Mývatn region in the tenth and eleventh centuries. The hall building discovered there was quite big (38 m in length), at least when we compare it with other halls in Iceland. It was erected high up on the homefield, rendering it visible in the immediate landscape, thus expressing some kind of symbolic power.72 The cattle skulls found outside the walls of the hall were originally placed at short intervals on the turf walls or roof along the entire length of the building.73 They were exposed and visible to people approaching the hall building, at least during certain periods of the year, most likely spring and summer. They must have had a great suggestive power. The farm was strategically located as a nodal point between Mývatn and the coast. It also had a good location in relation to other major routes to the north and the west, as the farm is situated next to Laxá, which is one of the major rivers flowing from the lake to the sea. Perhaps Hofstaðir controlled the flow of coastal resources into Mývatn district. A church was erected at Hofstaðir during the Early Middle Ages. It was situated at the new settlement on a mound 150–200 metres south-west of the Viking Age hall. It indicates that Hofstaðir remained as a place with central significance in the Middle Ages. The locations of the hall buildings found in Iceland, however, are not always central. The small hall of Sveigakot is one example. The place there was inhabited early (c. AD 880) in an extensive marshland south of the great inland lake Mývatn, and it is described as “marginal and poor”.74 Orri Vésteinsson suggests that inhabitants here may have been unfree people of non-Norse/British

70 Hann er nítján alna ok fjogurra tuga alna langr; hann stóð enn, er Árni biskup inn síðari var vígðr til Skálaholts. Fóstbræðra saga. Ísl. Fornr. 6, 184. 71 Meulengracht Sørensen 2003, 266. 72 Lucas 2009, 402. 73 See Lucas 2009, 236ff. 74 Orri Vésteinsson 2009, 500. 216 CHAPTER 6 ethnicity.75 Sveigakot may very well have been controlled by either the rich farm at Baldursheimur or the farm at Grænavatn. It is interesting to note that the small pit-house (2 × 2.16 metres) at Hólmur close to Hornafjörður (Höfn) in south-east Iceland, which has been interpreted as cultic house (“blót house”), has an extraordinary location. It was situated in an exposed position on a mound or a hill (15 × 25 m and about 2 m high) some 250 metres north-east of the Viking Age farm at Hólmur.76 Perhaps the location of this house also had a symbolic significance.

6.2.6 The Large Glittering Feasting Hall(s) in the Mythical World As noted above, pre-Christian ceremonial buildings in Svetjud, Trøndelag and sometimes in Iceland occupied prominent positions in the landscape. The prominence of these buildings was sometimes reinforced by their magnificent size and architecture, including the rich decoration. It seems as if these buildings were decorated with precious metal and artistic adornments, but also with skulls from butchered or sacrificed cattle. Some of these features also occur in relation to buildings and structures called hof and hǫrgr sanctuaries appearing in the mythical traditions. Vǫluspá 7, for instance, says that the Æsir met at Iðavǫllr and built a hǫrgr and a hof:

Hittuz æsir á Iðavelli, þeir er hǫrg oc hof há timbroðo; afla lǫgðo, auð smíðoðo, tangir scópo oc tól gorðo.77

Even if we do not get a detailed description of the mythic sanctuaries in this stanza, the verb hátimbra indicates at least that the hof (and perhaps also the hǫrgr) was built high.78 This information may be related to the large ceremonial buildings found by archaeologists, which sometimes seem to have been built very high. The size and depth of the holes for the roof-carriers indicate this. The idea that the mythic sanctuaries were high and even large is reinforced by Snorri. In Gylfaginning (ch. 14) (the Codex Regius version) he prob-

75 Orri Vésteinsson 2009. 76 See Bjarni F. Einarsson 2008. 77 “The Æsir met at Iðavǫllr/ they built a hǫrgr and a high hof/ they set up their forges, smithed precious things,/ shaped tongs and made tools.” Trans. Larrington. 78 See Heggstad et al. 1993 on hátimbra ‘tømre høgt’. Vikstrand (2001, 264) writes: “Detta ska dock inte nödvändigtvis uppfattas som att hovet och hargen var timmerkonstruktioner, utan verbet timra används här måhända i en mer allmän betydelse ‘bygga’.” The Ritual Structure of the Sanctuaries and Cosmic Symbolism 217 ably alludes to the hof and the hǫrgr of Vǫluspá 7, when he mentions that the gods decided to build a gold-glittering hof and next to it a hǫrgr called Vingólf:

It was their [the gods’] first act to build the hof that their thrones stand in, twelve in addition to the throne that belongs to All-father [Óðinn]. This house is the best that is built on earth and the biggest. Outside and inside it seems like nothing but gold. This place is called Glaðsheimr. They built another hall, this was the hǫrgr sanctuary that belonged to the goddesses, and it was very beautiful. This building is called Vingólf.79

Most likely Snorri also had another poem in mind when describing the “best and biggest hof on earth” built by the gods at Glaðsheimr, namely Grímnismál. According to the list of godly residences in this Eddic lay, the gilded building, which was situated at Glaðsheimr (‘bright home’ or ‘joy-home’),80 seems to be identical with Valhǫll, i.e. the hall where the Æsir deity Óðinn gathered the warriors slain in battle (einherjar) around him.81 In stanza 8 we may read thus:82

Glaðsheimr heitir inn fimti, þars en gullbiarta Valhǫll víð of þrumir; . . .83

79 Var þat hit fyrsta þeira verk at gera hof þat er sæti þeira standa í, tólf ǫnnur en hásætit þat er Alfǫðr á. Þat hús er bezt gert á jǫrðu ok mest. Allt er þat útan ok innan svá sem gull eitt. Í þeim stað kalla menn Glaðsheim. Annan sal gerðu þeir, þat var hǫrgr er gyðjurnar áttu, ok var hann allfagr. Þat hús kalla menn Vingólf. 80 Simek 2006, 112. Cf. Lindow 2001, 145. 81 The name Valhǫll should probably be interpreted as ‘the hall of the slain warriors’. It has been argued that this name is late, even though it appears in tenth-century sources such as Hákonarmál, Eiríksmál and Grímnismál. The word hǫll in the sense of ‘banqueting hall’ was not used in name-giving in Scandinavia. Hǫll is probably a loan-word from OE heall or OSax, OHG halla, which means ‘banqueting hall’. This does not entail that the mythological conception behind the term Valhǫll must be a late construction. Most likely the old expression for this conception was Óðins salr ‘Óðinn’s hall’, which appears in Eiríksmál, for instance. It also appears in place-names, such as Onsala and Odensala (both <Óðinssalir). Brink 1996a; 1999b; Nordberg 2003. 82 It should be noted that Snorri in Gylfaginning 2 describes Valhǫll as a “high hall” (háva hǫll). Its roof was covered with gilded shields like tiles (var lagt gyltum skjǫldum svá sem spánþak). 83 Glaðsheimr a fifth is called,/ there gold-bright Valhǫll/ rises peacefully, seen from afar; . . . 218 CHAPTER 6

It is interesting to note that “the gold-bright Valhǫll” was raised peacefully and “seen from afar”. It could be compared with, for instance, Adam of Bremen’s information about the “temple” in Uppsala, which may be quoted here once again: “A golden chain goes round the temple. It hangs over the gable of the building and sends it glitter far off to those who approach, because the shrine stands on level ground . . .” We have already noted that archaeological finds indicate that ceremonial buildings, at least in Sweden and Norway, had an elevated position in the cultural landscape, and thus could be seen from afar. Archaeological evidence also supports the assertion that such buildings occasionally were adorned with gold and other objects, like Valhǫll, for instance the Migration Period hall at Gudme, and the Merovingian Period house at Uppsala. In the mythical traditions there are also other hall buildings, such as Breiðablik, Glitnir and Valaskjálf.84 Whether they should be seen as separate mythic buildings or different designations of one cosmic hall is uncertain. Breiðablik ‘the far-shining one’ and Glitnir ‘the shining one’ have names which indicate an affinity to Valhǫll, which according to Grímnismál was considered as ‘gold-bright’. The name Valaskjálf may also be related to Valhǫll. According to Snorri, Óðinn owns this hall.85 The first element of the name could also be associated with the son of Óðinn, Vali. It is also possible that these names refer to different cosmic halls. In any case, some of these mythical names seem to emphasize important aspects of ceremonial halls, namely, that they should be decorated and visible from afar.

It seems as if the hall buildings and cult houses in Svetjud, Trøndelag and sometimes in Iceland occupied a prominent position in the landscape. The sovereignty of these buildings was reinforced by their magnificent size and architecture. Central locality, elevated position, size and architecture were probably associated with ideological aspects and a deliberate strategy of being observed and associated with the divine world. Most of these features appearing in relation to ceremonial buildings may be seen also in connection with essential “mythical houses” in the divine world. Perhaps these recurrent features and correspondences indicate an “ideal state”, the existence of a “common model”, and/or a cosmic symbolism of the cultic sites in the three regions (see further in ch. 15 below). Without doubt the ceremonial buildings in all three investigation areas were important agencies and/or symbols of power for the elite. The “bombastic expressions”, the monumental sizes and splendid

84 See e.g., Grímnismál 6, 12, 15; Gylfaginning 16, 21, 31. See also Simek 2006, 44, 113, 346. 85 Gylfaginning 16. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 219 localities, however, are not emphasized as much in Iceland as in Svetjud and Trøndelag. The tradition of building halls on platforms is mainly attested in the Mälaren region, and not found in Trøndelag or Iceland.

6.3 The Interiors: High-seat, High-seat Posts and Hearths

The interiors of the ceremonial buildings were probably often equally as impressive as the exterior of these houses. The halls had large rooms and high roofs. In these buildings there were several ritual structures and objects which repeatedly occur in the sources referring to Svetjud, Trøndelag and Iceland, such as the high-seats, high-seat posts, cultic images and hearths. Some of them probably had implications for the religious strategies for rulership, and they also appear as essential objects or places in the mythical world. As argued above, some of these elements may reflect a common tradition related to ceremonial buildings. In what follows I will first survey one of the most important written sources describing the interior of the ceremonial buildings. I will then concentrate on the high-seat, high-seat posts and hearths appearing inside ceremonial buildings. I will explore the evidence for them in the three investigation areas and their ritual function and ideological implications there. I will also discuss whether these “ritual elements” appear in the mythical world with similar functions and values. Cultic images will be treated in chapter 7 below and ritual rings (including oath-rings) in chapter 10.

6.3.1 The Interiors of Ceremonial Buildings as Described in the Written Sources It is unclear whether the medieval saga authors and scribes made a clear distinction between the Old Norse terms hof, salr and stofa. According to their descriptions, it seems as if hof was a pre-Christian ceremonial building intended for sacrificial feasts. It was also a place where cultic images were kept. The interiors of the buildings called salr or stofa often resemble the situation in the hof buildings in these accounts, but without cultic images, platforms and some other pagan paraphernalia. In both hof and stofa high-seats and hearths appear. Both had ritual and ideological dimensions. It seems as if the feasting halls called stofa were also used by Christian rulers. One of the best general descriptions of the interior of a Late Viking Age feasting hall (veizlustofa) is found in Fagrskinna 79 (c. 1220). The compiler or author of Fagrskinna has a general perspective when describing the interior of the royal halls as they appeared in the second half of the eleventh century. According to this account 220 CHAPTER 6

(based on manuscript *B),86 conditions were similar in Norway, Denmark and Svíaveldi:

It was an ancient custom in Norway, Denmark and Svíaveldi that where royal farms and banqueting halls (veizlustofur) were erected, there was a door at each gable of the house. The king’s high-seat (hásæti) was in the middle of the long bench, facing south. The queen was then supposed to sit at the king’s left hand. This was then called ǫndvegi (= ‘high-seat’) and the most honourable guests were assigned to sit there nearest to the high- seat, and then from there along the walls out [to the gables], females and men. The one who was less respected sat beside the door. And the most distinguished guest, who was old and wise, was called “the advisor of the king”. At that time it was a custom among kings to bring old wise men/ counsellors (spekingar) with them, in order to acquire knowledge about ancient customs and narratives about their parents. This man used to sit on the long bench opposite the king, on that seat called “the lesser seat” (óœðra ǫndvegi). He had females at his right hand, and at his left hand males were sitting. It was customary among the chieftains to carry the beer around the hearth and to make toasts to their “high-seat man”. It was conceived most honourable to sit and drink opposite the king’s toast.87

The value of this text as a source for Viking Age conditions can obviously be debated. It is inserted in a passage depicting how King Óláfr kyrri Haraldsson

86 There are two manuscripts of Fagrskinna. The manuscript *A burned in Copenhagen 1728, but was copied by Ásgeir Jónsson three times in AM 52 fol., AM 301 4to and AM 303 4to. *A is dated to the fourteenth century. Also manuscript *B burned in Copenhagen 1728, but was copied by Ásgeir Jónsson once (in OsloUB 371 fol.) and by Eyjólf Björnsson twice (AM 51 fol. and AM 302 4to). One page of this manuscript is preserved NRA 51. The manuscript *B is dated to c. 1250. See Fagrskinna, Ísl. Fornr. 29, lxi. On Fagrskinna, see Kolbrún Haraldsdóttir 1994. 87 Þat var forneskjusiðr í Nóregi ok í Danmǫrk ok í Svíaveldi, þar sem váru konungsbú ok veizlustofur, váru dyrr á hvárumtveggja enda stofunnar, en konungs hásæti var á miðjan lengbekk, þann er vissi móti sólu. Sat þá dróttning á vinstri hǫnd konungi. Var þat þá kallat ǫndvegi ok enn ágætasti sess út í frá hvárn tveggja veg til kvenna ok karla, er næst var ǫnd- vegi, en hinn óvegligastr, er næstr var durum. Enn gǫfgasti maðr, sá er var gamall ok vitr, var kallaðr konungs ráðgjafi, sem í þat mund var konungum títt at hafa með sér gamla spekinga til þess at vita forna siðu ok dœmi sinna forellra. Þessi maðr skyldi sitja á annan langpall gagnvert konungi, ok hét þat et óœðra ǫndvegi. Váru konur á hœgri hǫnd hónum, en á vinstri hǫnd sátu karlar. Var þá títt hǫfðingjum at bera ǫl um eld ok drekka á ǫndvegismann sinn. Þótti þat vera mest virðing at sitja fyrir konungs ádrykkju. Fagrskinna, Ísl. Fornr. 29, 300. My trans. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 221

(r. 1066–1093) moved the high-seat (hápall) from the middle of the long wall to the middle of the “crossway stool” (á miðjan þverpall) at the gable.88 The distance of time between Óláfr kyrri’s reign and the scribe of Fagrskinna is, on the other hand, just a little more than a hundred years. Consequently, it is possible that Fagrskinna’s description of the hall is reliable. The scribe also quotes a stanza contemporary with King Óláfr in support of his description. This stanza was composed by Arnórr Þorðarson jarlaskáld in honour of Earl Þorfinnr, and it describes the drinking rituals which were supposed to be observed by the person sitting opposite the ruler.89 Since the content of the poem supports parts of the prose account, the latter may arguably be considered a trustworthy source for interiors of Late Viking Age halls in general.90 Most likely the description of the interior of ceremonial buildings in Fagrskinna influenced other medieval authors and scribes writing about feasting halls (veizlustofur). Snorri Sturluson, for instance, may have been inspired by this description when presenting Óláfr Haraldsson’s hirðstofa ‘retainer hall’ at the royal residence (konungsgarðr) in Niðaróss.91 In this description, he says that the hall had doors at both ends, and that the king’s high-seat (hásæti) was situated in the middle of the hall. Opposite the king’s high-seat was the “lesser high-seat” (ǫðru ǫndugi), where the king’s marshal sat, and next to him the noble guests. On those occasions when ale was consumed, fires were lit in the middle of the floor. All these statements resemble Fagrskinna’s description of feasting halls. It is quite possible that Snorri partly borrowed from this text when describing the interior of the feasting halls. Perhaps he also borrowed from it when describing the hof buildings in Trøndelag.

6.3.2 The High-seat: Concepts and Position Judging by Fagrskinna, it seems as if the symbolic and ritual centre of the hall was the high-seat (ON hásæti, ǫndvegi; OSw høghsæte; OE hēah-setl; OHG hohsedal, hohsidil, hohgisetu). It is generally accepted that it was regarded

88 Similar information appears in Óláfs saga kyrra 2, in Heimskringla III, Ísl. Fornr. 28, 204. See Birkeli 1932; Steinsland 1991, 77. 89 Finnur Jónsson interpreted it thus: “Jeg tiltalte (drak til) krigeren,/ da jeg sad ligeoverfor ham;/ fyrsten drak altid hver vinter/ fyrsten drak altid hver vinter/ det brave mandskab til.” Hétk, þás (hvern vetr) sǫ́tum,/ hrafns verðgjafa, (jafnan/ lið drakk gramr á góðar)/ gag- vart (skipa sagnir). See Skj. B1, 316. 90 The passage in Fagrskinna probably has a specific relevance to conditions in Trøndelag. Before the passage about the banqueting halls (veizlustofur), a stanza made by Steinn Herdísarson is quoted where Óláfr kyrri is called þengill Þrœnda ‘the king/chieftain of the Þrœndir’. See Fagrskinna, Ísl. Fornr. 29, 299. See also Skj. B1, 282. In the prose immediately after this stanza the increasing significance of Niðaróss is emphasized. 91 See Óláfs saga helga 57, in Heimskringla II, Ísl. Fornr. 27, 72. 222 CHAPTER 6 as a place of honour, and intended for the owner of the hall: the king or chieftain.92 The high-seat was probably elevated in some way, so that the ruler actually looked down on the other people in the room. The other participants seem to have been seated according to rank and gender at specific places in the ceremonial building, but on a lower level than the hall owner. The terms fótpallr ‘foot bench’ or fótskǫr ‘foot step’, which are associated with the high- seat in some sources, indicate that this seat stood on some sort of dais or platform.93 The elevated position of this seat is even marked in the first element of the compound term hásæti; the adjective há- means ‘high’. According to the medieval texts, the high-seat was a social and political fixture with ideological implications. In addition to that it played a central role in religious and judicial contexts. It was thus a symbol of tradition, power and authority. A person who had the right to occupy such a seat was called hásætismaðr, ǫndvegishǫldr or ǫndvegismaðr. There has been debate in previous research concerning the placing of the high-seat in the hall and whether hásæti and ǫndvegi were identical or not. I will not go too far into that debate, just summarize it, and then make some comments on it in the context of ceremonial buildings in the investigation areas. Based on the information in Fagrskinna and other Old Norse texts, it was argued early on that the pre-Christian high-seat (whether it was called hásæti or ǫndvegi) was situated in the middle of the northern long bench.94 On the opposite side of the high-seat was “the lesser seat” (óœðra ǫndvegi) intended for the highest-ranking guest, or the most honourable person after the ruler of the house. In the middle of the eleventh century this custom was changed when the high-seat was moved to the eastern gable. Hjalmar Falk opposed some of these ideas and stated that a distinction must be made between ǫndvegi and hásæti. The former term was applied to the place of honour in farmers’ houses, while the latter term referred to the seat of rulers in royal halls.95 A similar proposal was made by Emil Birkeli. According to him the concept and

92 See e.g., Birkeli 1932; Holmqvist 1962; Drescher & Hauck 1982; Steinsland 1991; Herschend 1997; 1998; H. Beck 2000; Sundqvist 2014. According to Old Norse prose texts, high-seats (ǫndvegi) also appeared in the houses of minor chieftains and farmers. See Birkeli 1932, 29ff. 93 See e.g., Ynglinga saga 36, in Heimskringla I, Ísl. Fornr. 26, 66; Haralds saga in Hárfagra 8, in Heimskringla I, Ísl. Fornr. 26, 100; Saga Magnúss góða 23, in Heimskringla III, Ísl. Fornr. 28, 37; Flateyjarbók I, 42. 94 Keyser 1868; cf. Gudmundsson 1894; Nikolaysen 1881–91. 95 Falk 1911–16. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 223 phenomenon of hásæti was developed and emancipated from the old ǫndvegi. It was also accommodated in a new and more aristocratic context.96 The old term ǫndvegi referred originally to the seat of honour in farmers’ houses, which was always situated in the north-eastern corner. In this context Birkeli also discussed the etymology of the term ǫndvegi. Usually it has been interpreted as a compound word with the preposition ǫnd ‘opposite’ (‘rett imot’) as the first element and vegi (derived from vegr) ‘side, edge, margin’ (‘side, retning, kant’) as the second element.97 The compound word should be translated as ‘the seat opposite [the sun]’ (‘sætet rett imot [solen]’). In Birkeli’s opinion, however, the first element ǫnd should be translated ‘soul’ and the second element vegi (vegr) as ‘road’. The term ǫndvegi should thus be interpreted as ‘the road of the souls’ (‘åndeveien’), i.e. the place where dead ancestors have their place in the house. The ǫndvegi in the north-eastern corner of the house was thus originally a sacred sphere where the ancestor cult took place.98 Birkeli has gained weak support for his theory about the meaning of ǫndvegi.99 His ideas about the location of the high-seat have also been questioned. Wilhelm Holmqvist stated, for instance, that Falk and Birkeli were leaning too much on written evidence concerning this issue.100 According to Holmqvist, the location of the high-seat must be related to building technology. In a hall or a long-house, with an open fireplace in the centre of the room, it is natural to place the high-seat in the middle of the long wall. During the Middle Ages a new type of (smoke) oven was introduced in houses, which led to the corner location of the high-seat. Frands Herschend is also critical of the reasoning in previous research, and has proposed an alternative schematic development of the position of the prominent seats in the house, hásæti or/and ǫndvegi.101 His proposal was based on new empirical data gained from recent archaeological finds in Iron Age halls. It is quite possible that Falk and Birkeli were on the right track when they argued that the aristocratic and royal hásæti was developed from the ǫndvegi of farmers’ houses,102 but it is probably impossible to prove this theory from existing sources. In the written sources from the thirteenth century, however,

96 Birkeli 1932, 11ff., 16ff. 97 Birkeli 1932, 8. Cf. de Vries 1977, 687. 98 Birkeli 1932, 48. 99 Steinsland 1991, 77. 100 Holmqvist 1962. 101 Herschend 1998, 27ff. 102 See e.g., Steinsland 1991, 76. 224 CHAPTER 6 it seems as if hásæti and ǫndvegi sometimes were used as synonymous terms.103 If we look at the Sagas of Icelanders, for instance, these two words are often used for one and the same place.104 Moreover, Johan Fritzner translated both hásæti and ǫndvegi as ‘Høisæde’.105 What we may conclude is that the term hásæti (OSw høghsæte) was used for important structures in ceremonial buildings in all three areas (Svetjud, Trøndelag and Iceland). Whether the term ǫndvegi was used in eastern Scandinavia or is specific for western Scandinavia is uncertain.

6.3.3 Evidence of High-seats in Svetjud The saga literature mentions that high-seats were placed in the halls of Svetjud.106 Some of these accounts may be considered trustworthy sources for interiors of Viking Age halls in general. However, for conditions in the Mälaren region in particular, the source value of the Old Norse written evidence is more limited, because of the distance in both time and space between the actual events they describe and the time of fixation in written text. On the other hand, pre-Christian high-seats may be archaeologically attested in the halls of the Mälaren region. Inside the cultic hall building (IA) at Helgö there was an elevated position in the corner of the southern aisle of the banqueting room. It has been interpreted as the place of the high-seat (fig. 7A).107 Most of the gold foil figures and the expensive glass were found there, indicating that it was a ritual space. They were clustered in an area 3 × 4 m in size and close to the southern wall.108 In front of it was the hearth, which probably also had ritual significance. At the “Garrison” in Birka there were high-status objects concentrated in the western parts of the hall, e.g., a dragon’s head, glass beakers and sword handles. Archaeologists have suggested that the “high-seat” was located in this place. This “high-seat” was likewise situated close to the hearth.109 In the Migration Period hall at Vallhagar (house 16), Gotland, features of both “the high-seat” and the opposite “lesser high-seat” have been found. They are

103 Birkeli (1932, 6f.) stated, however, that Snorri only used the concept hásæti in royal contexts in his Heimskringla. Perhaps Snorri made a distinction between hásæti and ǫndvegi. 104 See e.g., Eyrbyggja saga 33 (Ísl. Fornr. 4, 91f.), where these two concepts refer to one and the same seat. 105 Fritzner 1954. 106 See e.g., Ynglinga saga 36–37, in Heimskringla I, Ísl. Fornr. 26, 65ff., but also in more historical accounts such as Fagrskinna 79, Ísl. Fornr. 29, 300 and Óláfr saga helga 79, in Heimskringla II, Ísl. Fornr. 27, 113. 107 See Herschend 1995; 1998, 28. 108 See also K. Lamm 2004, 51; Lund Hansen 2011. 109 Holmquist Olausson 2001, 15; Holmquist Olausson & Kitzler Åhfeldt 2002, 16. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 225 located in the middle of the long walls. Drinking goblets have been found at these “seats”, and the hearth was located between them.110 It is possible that the high-seat in the younger hall of Lunda (c. 450–650) was also close to the hearth. The central post-holes which flanked it there were distinguished from the other post-holes both in size and in the richness of the finds. They had a diameter between 0.74 and 1.68 metres and were 0.46 and 0.78 metres deep. In these holes a lot of objects were deposited, such as a bead, an amber-stone, a comb and a fishing hook. The other post-holes of the building had very few finds. The excavators suggest that the high-seat was in this area, close to the northern wall. According to them, the finds deposited there indicate that some kind of cult took place.111 The high-seat area has also been identified at the Migration Period hall of Högom, Medelpad.112 High-seats may also be seen in the Late Iron Age iconography from eastern Scandinavia. On one Gotlandic picture stone from Buttle Änge no. I the interior of a hall is depicted with two seats located opposite each other (see fig. 9a and 9b). According to Frands Herschend, they must be interpreted as a “high- settle” and a “lesser high-seat”. The right one is higher than the left seat, since it is standing on a dais. It was interpreted as the “high-seat”.113 A “high-seat” and a “lesser high-seat” are also evidenced on the stone coffin from Sanda, Gotland, dated to the eleventh century (see fig. 10).114 Two persons are sitting opposite each other in these seats, in some sort of a building with roof and walls. The one who is sitting in the right seat, probably a “high-seat”, seems to be offered a drinking vessel, or some other object, by a third person. Below this motif, three men (or deities) are walking in a procession carrying different attributes.115 This motif emphasizes that the upper image may also concern some kind of ritual or ceremony. This coffin is also decorated with a runic inscription at the very top. This inscription (G 181) does not give any clues to interpret the images below it, since it only says: Rōðvīsl auk Farbiorn auk Gunnbiorn “Rodvisl and Farbjörn and Gunbjörn”.116

110 Herschend 1998, 26. 111 Skyllberg 2008, 29. 112 Ramqvist 1987; 1991. 113 Herschend 1998, 26. 114 Jungner 1930; Lindqvist 1941–42; Nylén & Lamm 1987; Drescher & Hauck 1982, 258. 115 Cf. Jungner 1930. 116 Thorgunn Snædal 2002, 76. 226 CHAPTER 6

Figure 9a Detail from the Gotlandic picture stone from Buttle Änge no. I. After Herschend 1998, based on Lindqvist 1941–42)

Figure 9b The picture stones from Buttle. Carl Gustaf Rosenberg.

Figure 10 Stone coffin, Sanda, Gotland. Photo: Bengt A Lundberg. ATA, RIKSANTIKVARIEÄMBETET 226.

6.3.4 Religious Implications of the High-seats It thus seems as if high-seats occurred in the ceremonial buildings of Svetjud and eastern Scandinavia. It has been argued that when the ruler entered the high-seat, he probably made a specific contact with the divine world.117 The high-seat was a sacred place and the high-seat posts were perhaps even

117 Cf. Olivecrona 1942; Hellström 1971, 221; Steinsland 1991, 66–86; Herschend 1997, 9, 49–59; Sundqvist 2002, 268ff.; 2014. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 227 regarded as some kind of cosmic pillars (axis mundi sg.) or an interface to the cosmic world (see below). The ritual deposits of gold foil figures at the position of the assumed high-seats in the halls of Helgö and Svintuna (Östergötland) may be regarded as good evidence of such notions (see further ch. 12 below). The religious significance of the high-seat is only slightly evidenced in the written sources referring to eastern Scandinavia. Adam of Bremen, for instance, describes the interior of the “temple” at Ubsola thus: Thor in medio solium habeat triclinio “Thor occupies a seat in the middle of the dining-room.” Latin solium connotes ‘high-seat, throne, insignia’.118 The prototype for this high-seat must have been intended for earthly rulers. There might have been some kind of micro-macro symbolism or ritual relationship between the real high-seats in the halls of the Mälaren region and the divine high-seat of Þórr in the sanctuary of Uppsala, as described by Adam.119 When the rulers entered the high- seats they actually imitated the prototypical actions performed by the gods in the mythical time. At the same time they were in contact with the divine world when sitting in these seats (see ch. 7). It is also possible that the entering of the high-seat could be plainly seen as a performative ritual intended to authorize and legitimize the ruler (see ch. 13 below). The ritual and ideological dimension of the high-seats among the Svear is also attested in the Old Norse sources. In these texts the high-seat plays an important role in connection with drinking rituals and rites of passages performed by rulers (see further ch. 9 and ch. 13 below). The religious significance of high-seats in the Mälaren area may also be attested in another sense. Among the objects in the tenth-century hoard from Eketorp, Närke, was a silver miniature “high-seat” which had been used as a pendant or amulet.120 This type of miniature is often interpreted as a representation or attribute of the divinities. They also give us an image of how “real high-seats” were shaped and decorated. In a Birka skeleton grave, no. 844, there was a similar amulet to the one in Eketorp from the Viking Age (fig. 11).121 Nine “Klotzstuhl-Amulette” have been discovered in Scandinavia.122 Two of them were in the coffin graves nos. 632 and 968 at Birka and dated to the Viking Age. One was also found at Ihre, Hellvi parish, Gotland. In addition Karl Hauck

118 Lat solium “der hohe erhobene Sitz, der Thron, . . . Regale”. Georges 1951 (1913), 2711. Habel (1959, 371) has “Sitz, Stuhl, Bank, Tribüne, Katheder, Thron; Hochwarte; Sarg”. 119 See e.g., Nordberg 2003, 170ff. 120 Ekelund 1956; Stenberger 1971 (1964), 731; Drescher & Hauck 1982, 267–269. 121 Drescher & Hauck 1982, 267–269. 122 Drescher & Hauck 1982, 248–256. 228 CHAPTER 6

Figure 11 A “high-seat” amulet (Sw. kubbstol) from Birka. Statens Historiska Museum. has drawn attention to iconographic evidence of “Klotzstühle”. One appears on the runic stone from Sanda, Gotland, mentioned above.123 According to Hauck these materials depict “Götterthrone”. Their symbolism is attested on bracteates as early as the Migration Period. It was thus firmly rooted in a pagan context. Most interesting is the high-seat-amulet found recently in Lejre, in Denmark.124 It is often called “Óðinn from Lejre” and dated to 10th century. Whether this amulet should be interpreted as an image of Óðinn in Hliðskjálf can, however, be disputed.

6.3.5 High-seats in Trøndelag and their Religious-Ideological Implications There are plenty of texts indicating the existence of ancient high-seats in Trøndelag.125 There is also evidence to show that these seats had a religious and ideological significance. Snorri’s description of the ceremonial feasts in Lade in Hákonar saga góða 17 tells us that King Hákon the Good wanted to sit in a lesser house during the sacrificial feast (blótveizla), but the farmers were not satisfied with this and wondered why he was not sitting in the high-seat, when there was so much joy among his people.126 Earl Sigurðr told them that he would not do it this time. In the end the king was more or less forced to take his place in the high-seat.127 Between the lines Snorri intimates that the king had no other choice than to take the seat inside the large ceremonial building. When he was seated he had to lead the ceremonial drinking ( full) to Óðinn.

123 Drescher & Hauck 1982, 258. 124 Christensen 2009 and 2010. 125 See e.g., Óláfs saga helga 57, in Heimskringla II, Ísl. Fornr. 27, 72; Landnámabók S 297, H 258, Ísl. Fornr. 1, 307f.; Brandkrossa Þáttr, Ísl. Fornr. 11, 188f.; Flateyjarbók II, 333; Óláfs saga Kyrra 2, in Heimskringla III, Ísl. Fornr. 28, 204; Fagrskinna ch. 79, Ísl. Fornr. 29, 300, which may also be relevant for Viking Age Norway. 126 . . . er hann sat eigi í hásæti sínu, þá er mestr var mannfagnaðr . . . Hákonar saga góða 17, in Heimskringla I, Ísl. Fornr. 26, 171. 127 Var ok svá, at konungr sat í hásæti sínu. Ísl. Fornr. 26, 171. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 229

The next evening King Hákon was compelled to take part in the ritual meal and eat the horse-meat (hrossaslátr) and then enter the high-seat. Even though Snorri’s account has been criticized by scholars, there is no reason to suspect that he misunderstood the significance of the high-seat. There are other Old Norse texts indicating the existence of high-seats in this region. Landnámabók, for instance, mentions that the chieftain and hofgoði Þórhaddr from Mære dismantled the hof and brought all the “temple soil” and the high-seat posts (súlurnar) to Iceland (see ch. 5 below). These posts had most likely been related to the high-seat of the old hof in Mære. A Viking Age ceremonial building, perhaps a hof, has been detected in Mære, Trøndelag. When Hans-Emil Lidén investigated this house, he discovered four post-holes, which do not logically fit in relation to the rest of the building.128 Whether these post-holes belonged to a building or stood within the building is uncertain. They constituted a square measuring about one square metre. They were located at the long wall, not far from the fireplace, and the gold foil figures were found beside them. Inspired by information from Old Norse texts, such as Landnámabók, Lidén first interpreted the four post-holes as remains of the high-seat, or more precisely, the high-seat posts (ǫndvegissúlar). In a later publication he believed that three of the four post-holes belonged to posts supporting the wall.129 As far as I know, no other possible “high-seats” with gold foil figures have been found in Late Iron Age houses in Trøndelag. The two gold foil figures discovered at Hevne were probably located at the old farm, but no post-holes or traces of houses were found there. At the hall in Hovde neither gold foil figures nor remains of a high-seat were discovered. The closest parallel in Norway to the finds at Mære is probably the structures and finds excavated at Borg, Vestervågøy. In the Merovingian and Viking Age “hall-room” (Room C) at this site (I:1a) five gold foil figures depicting the motif of a couple were found (fig. 8C).130 These objects seem to be related to post-holes (actually one post- hole) and an area interpreted as the place of the high-seat. The assumed high- seat was located in the north corner of the room, where other finds were also made which may indicate that ritual activities took place there during the Early Viking Age. Another close parallel is the important hall-building with 29 gold foils found in Hov in Vingrom, Oppland. The gold foils were found in connection with four post-holes, which may be interpreted as traces of a high-seat.

128 Lidén 1969. 129 Lidén 1999, 2, 38f. 130 Munch 2003a and 2003b. 230 CHAPTER 6

Fragments of glass vessels, weapons, seaxes, knives, pearls and a large amount of strike-a-lights were also discovered there.131

6.3.6 High-seats in Iceland There are plenty of Old Norse texts mentioning the existence of high-seats in Iceland.132 It is quite clear that this place in the house, according to the sagas, had an ideological significance. In several passages it seems to have been a place of honour. According to Njáls saga 118, for instance, Njáll was carried into Ásgrímr’s house from his horse and placed in the high-seat, while Ásgrímr himself took another seat.133 This was done as an act of honour for the old man. It also seems as if the high-seat was a sacred and protected place, where people could find peace. Perhaps it was even conceived as an interface to the Other World. According to Vatnsdœla saga, the chieftain Ingimundr of Vatnsdalr was wounded by Hrolleifr’s spear. He went inside and took his place in his high-seat (settisk í ǫndvegi sitt). When his sons came home, Ingimundr was sitting dead in his high-seat (í ǫndvegi sínu ok var dauðr).134 Eyrbyggja saga relates how Þórólfr bægifótr came home one night after arguing with his son Arnkell goði. He sat down in his high-seat (í ǫndvegi) but did not speak to anyone. When the people went to bed he was still sitting there. The next morning they discovered that he was still sitting there, dead.135 In the year 1148 Bishop Magnús Einarsson made a visitation in Iceland. One night fire broke out in the house. The bishop took his place in his high-seat:

The men wanted the bishop to run out [of the house], but he answered that he did not need to escape his seat, since this was the will of God. When the fire was over he sat there wholly unburned in his high-seat.136

These men probably believed that the high-seat opened the door to the Other World. Perhaps they also wanted to die in full regalia (“in Amt und Weiden”) as they sat in their high-seat.

131 Oral communication from Heid Gjøstein Resi, Kulturhistorisk museum, Oslo University, via the reviewer of my manuscript. 132 See e.g., Eyrbyggja saga 13, Ísl. Fornr. 4, 23f.; Vatnsdæla saga 22, Ísl. Fornr. 8, 61. 133 Ásgrímr tók Njál af hestinum ok bar hann inn ok setti hann í hásæti. Brennu-Njáls saga 118, Ísl. Fornr. 12, 296. Cf. Grettis saga 19, Ísl. Fornr. 7, 61–71. See also Birkeli 1932, 42f. 134 Vatnsdæla saga, Ísl. Fornr. 8, 61f. 135 Eyrbyggja saga 33, Ísl. Fornr. 4, 91f. 136 villdu menn at byskup hefdi þar vt farit en hann sagdi sier hæfdi ei at ryma sitt sæti þvi þetta værj guds vilie. Sat hann vpp obrunen med aullu j sinu hasæti þa af var elldurinn. Gottskalks Annaler 1148, Islandske Annaler, 321. Cf. Hellström 1971, 220f. My trans. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 231

The relation between ceremonial buildings and high-seats/high-seat posts may also be attested by archaeology in Iceland. At the hall building (AB) of Hofstaðir, Mývatnssveit, there was a central hearth in the middle room. The excavators suggested that the high-seat was located there, opposite the hearth.137 Certain finds support this assumption. Several primary foundation deposits were made in post-holes in this area, such as the seventeen stone gaming pieces placed in a post-hole [2589], a human tooth and a crucible fragment in post- hole [2876] and a spindle whorl and a knife in post-hole [2754]. Gavin Lucas argued that the location of these deposits probably was not coincidental.138 These deposits may have been placed there because they were used for the high-seat posts. Iconography and images may support the existence of high-seats in pre- Christian Iceland. The Viking Age image from Eyrarland at Akureyri in northern Iceland is sitting on a chair which has been interpreted as a “high-seat” (see fig. 15). Since the figure has an object interpreted as a hammer (or cross) in his hand, it has been suggested that the image represents Þórr in his high-seat. This interpretation has, however, been debated.139

6.3.7 The High-seat Posts Not only the high-seat, but also the high-seat posts (ǫndvegissúlar) seem to have been regarded as important religious symbols and/or ritual objects. They constituted an architectural part of the ceremonial buildings.140 Most likely they were used as roof-carriers in the central area of such buildings.

137 See Lucas 2009, 387, 394. 138 Lucas 2009, 394f. 139 See e.g., Drescher & Hauck 1982, 263ff.; Perkins 2001, 82ff. 140 Landnámabók, S297 H258, Ísl. Fornr. 1, 307f.; Eyrbyggja saga, Ísl. Fornr. 4, 6–10; Vatnsdæla saga 15, Ísl. Fornr. 8, 42. As far as I know there is only one exception where such posts are related to another type of house. According to Landnámabók, Ingólfr’s high-seat posts seem to have been placed in his ‘fire house’ (eldhús) at some time: “The high- seat posts are still there in the fire house” (þar eru enn ǫndugissúlur þær í eldhúsi) S 8–9, Ísl. Fornr. 1, 45. Most likely this information must be related to the transmitter’s own experience. After Christianization the high-seat posts were taken out of their original pagan context, since the old hof was abandoned and/or destroyed, and the pagan cult was rejected. Thus they were placed in the ‘fire house’, as a memory of Ingólfr and his landnám. In this context they gained a new symbolic meaning and function. The fact that Ingólfr’s farm at Kjalarnes was called Hof (Landnámabók, S 14, H 14, Ísl. Fornr. 1, 50f.), indicates that these high-seat posts originally were placed in a hof sanctuary. Cf. Birkeli 1932, 25. 232 CHAPTER 6

The term ǫndvegissúlar also intimates that they were connected to the high- seat (ǫndvegi) in some sense.141 High-seat posts are mainly attested in written sources referring to Norway and Iceland, but some archaeological evidence suggests that they also occurred in Svetjud. When studying the literary sources it is clear that the high-seat posts sometimes were considered as ritual objects which could be separated from the ceremonial building and the high-seat. When travelling to Iceland from Norway, for instance, these posts were brought on the ships and used in a type of divination ritual or in order to get good luck.142 Landnámabók states that the first landnámsmaðr Ingólfr and his foster-brother Hjǫrleifr left their home in Norway. Before leaving, Ingólfr made a great sacrifice in order to find out about his future, and to achieve luck with his enterprises.143 He was advised to take off to Iceland.144 It is mentioned that Hjǫrleifr never sacrificed. Anyhow he followed Ingólfr to Iceland. As soon as Ingólfr arrived at the coast of Iceland, he cast his high-seat posts overboard, in order to get luck.145 Ingólfr also declared that he would settle down where the posts came ashore.146 He sent out his slaves to find the high-seat posts. But it was not until the third winter after their arrival that they could be found at an isthmus in south-western Iceland. Although the conditions for living were not so propitious on this spit of land, Ingólfr decided to settle there since the high-seat posts and the supernatural powers had pointed out this place to him. According to Landnámabók, many men followed Ingólfr’s example and settled in Iceland.147 They also brought their high-seat posts from the sanctuaries in Norway and cast them overboard when approaching Iceland. For instance, Loðmundr inn gamli,148 Hrollaugr Rǫgnvaldsson149 and Þórðr skeggi Hrappsson followed this custom.150 People from Trøndelag also performed

141 For discussion, see Birkeli 1932, 23ff.; Holmqvist 1962, 292. 142 Cf. Strömbäck 1928b; Baetke 1942; Clunies Ross 1998, 142ff. 143 . . . at blóti miklu ok leitaði sér heilla um forlǫg sín . . . Landnámabók, SH 7, Ísl. Fornr. 1, 42. 144 Fréttin visaði Ingólfi til Íslands. Landnámabók, SH 7, Ísl. Fornr. 1, 42. 145 . . . skaut hann fyrir borð ǫndugissúlum sínum til heilla. . . Landnámabók, SH 8, Ísl. Fornr. 1, 42. ON heill (f. and n.) means ‘luck’ or ‘good omen’. Heggstad et al. (1993, 176): ‘lykke, (godt el. vondt) varsel, teikn’ or ‘ting som gir lykke’. 146 During the taking of land Hjǫrleifr was killed by one of his slaves. 147 The story about Ingólfr functions as a paradigmatic example of good behaviour and may be considered as a dǿmisaga or an exemplum. See Wellendorf 2009; 2010. 148 Landnámabók, S289 H250, Ísl. Fornr. 1, 302. 149 Landnámabók, S289 H270, Ísl. Fornr. 1, 316f. 150 Landnámabók, H 11, Ísl. Fornr. 1, 124f. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 233 such rituals. Landnámabók mentions that, for instance, the hofgoði Þórhaddr from Mære took down the hof and brought all the soil of the sanctuary and the high-seat posts to Iceland (see ch. 5 above). Eyrbyggja saga 4 presents a similar story (see ch. 3 and 4 above). It thus seems as if the theme of casting the high-seat posts was developed into a topos in both Landnámabók and the saga narratives about the landnám.151 The management of the high-seat posts and the ritual occupation of land in the saga accounts occur in a clearly pre-Christian context. Most scholars have also argued that these narratives reflect genuine pre-Christian customs and rituals from the landnám period.152 This opinion has recently been questioned by Jonas Wellendorf, because there are a couple of narratives in the Hauksbók redaction of Landnámabók with a content similar to the accounts of high-seat posts, but with a clearly Christian context.153 According to Dag Strömbäck, these Christian stories reflected a “cult action continuity”. He interpreted them as a sign of the great significance and bearing capacity the pagan custom also had for the early Christians.154 Wellendorf opposes this way of reasoning and suggests that the accounts of high-seat posts might be secondary to the Christian accounts found in Hauksbók. In order to support his argument, Wellendorf presents two medieval ecclesiastic narratives, which may have functioned as models for these

151 See e.g., Wellendorf 2009, 1018f.; 2010. There are some traditions of the casting of high-seat posts which do not turn up in Landnámabók, for instance in Laxdæla saga 5, Ísl. Fornr. 5, 9 and Kormáks saga 2, Ísl. Fornr. 8, 205. 152 Dag Strömbäck (1928b) interpreted this pagan ritual as a way of transporting the sacred power and good fortune (heill) from the old land to the new. One of the purposes was to make the landvættir, the ruling spirits of the land, well disposed to the human intruders into their domain. Margaret Clunies Ross (1998, 150) opposed Strömbäck and stated that “the rituals described in the sagas and in Landnámabók as having been carried out by the first settlers in order to secure their land claims were directed, not at the landvættir as the first owners of the land, but at other humans who might appropriate what individuals wanted for themselves.” Emil Birkeli (1932, 24ff.) connected the high-seat and its posts with the ancestor cult. See also Gunnell 2001. 153 According to Landnámabók (H 21, Ísl. Fornr. 1, 59ff.), Ásólfr alskik came to Iceland to preach about Christ. When he died the congregation wanted to build a church at his funeral place. Illugi, who was sent abroad to find the building material, was forced to cast all church wood overboard with a prayer when returning to Iceland. Three days later the wood came ashore at Kirkjunes close to Akranes, where they built the church. There is one similar story related to the Christian man Ørlygr Hrappsson in Landnámabók H 15, Ísl. Fornr. 1, 53ff. 154 “. . . denna hedniska seds livskraft och stora betydelse . . .”. Strömbäck 1928b. 234 CHAPTER 6 accounts in Landnámabók, namely Ægidius vita and Declans vita.155 The cult of Ægidius was popular in Norway during the Middle Ages. He was included and celebrated in the liturgy of Niðaróss. “The Legend of Ægidius” was also translated into Norse.156 It tells that Ægidius travelled to Rome and saw a couple of doors which were carved and decorated with pictures of the apostles. He liked them very much and wished that he could take them home to his monastery at Nîmes in Provence. The pope therefore gave them to him. When taking these doors home to Provence, Ægidius cast them into the River Tiber. People thought that he had gone mad. But when he arrived at his monastery, people told him that two fine decorated doors had been found at the banks. According to Wellendorf, this narrative has much in common with the account of Þórólfr Mostrarskegg’s high-seat posts. In both stories the numinous objects find the predestined place in a miraculous way. And in both accounts these objects are decorated with “sacred beings”, the apostles and Þórr respectively. In my opinion, there are also some important differences. In the accounts of Landnámabók and Eyrbyggja saga, the ritual with the posts was performed as a divination in order to find a place to settle in a new land. In these narratives the high-seat posts also play a part in a consecration ritual, where the intention was that the new and unoccupied land should be transformed from chaos to cosmos.157 At the place where the high-seat posts came ashore the landnámsmenn went with fire around the area which was supposed to be occupied and settled.158 The purpose of these permormative rituals was to symbolically transform the land into cosmos, and to create a sacred area suitable for a settlement and particularly a sanctuary.159 In “The Legend of Ægidius” the doors arrived at a place which was already cosmicized (speaking in Eliade’s terms), settled and even regarded as sacred. Thus the very core of this legend was quite different from the Icelandic traditions. The “Legend of Declan” admittedly has some similarities to the accounts of the high-seat posts. This text explains why the Irish monk and saint, Declan, settled at Ardmore, near Waterford. Declan had a little magic bell that he had received from God. Once he was out at sea this bell came sailing on a stone. He followed it and decided that where the bell reached land he would settle

155 Bibliotheca hagiographica latina. Subsidia hagiographica, 93 and 2116. Quoted from Wellendorf 2009. 156 The Norse text is quoted in Wellendorf 2009, 1021. 157 See Eliade 1987 (1949), 29ff.; 1991 (1949), 10ff. 158 See Strömbäck 1928b, 203ff. 159 It was also a way to secure the territory by a symbolic means of showing one’s dominance over other men. Clunies Ross 1998, 150. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 235 down and found a bishopric. It came ashore on a small island off the Irish coast, where he also established his episcopal see. This story thus has a great deal in common with the narratives about the high-seat posts. One problem is, however, that Declan was not celebrated in the Liturgy of Niðaróss. Nor was this story translated into Norse. It should be mentioned that Wellendorf never intended to discuss the his- toricity of the ritual of casting high-seat posts; he was mostly interested in the literary motif. In any case, his result has consequences for historians of religions, since it undermines the historical source value of Landnámabók. In my opinion, there are nevertheless reasons to believe that the high-seat posts played a significant role for the pre-Christian religion, but perhaps not in the way that is sometimes described in the Old Norse traditions. The connection between these pillars and gold foils attested by archaeology supports such an assumption (see further below).

6.3.8 Ritual Implications of the High-seat Posts The high-seat posts seem to have been connected to hof sanctuaries in the medieval accounts. Sometimes these accounts also intimate that the posts were erected in a part of the sanctuary which was most sacred, where other sacred objects were also buried. There is, for instance, an interesting description of high-seat posts and a hof sanctuary in Vatnsdæla saga. It is stated that the chieftain Ingimundr the Old, who was brought up on the island of Hefniey in Hálogaland, once received a gift from King Haraldr, a weight (hlutr) on which Freyr was stamped in silver.160 This weight had previously belonged to the rich and wealthy King Kjǫtvi. Ingimundr carried this weight in his purse, but once he lost it. During a seiðr séance at Hefni, a Sámi enchantress assured him that he would find it again at the place where he would build his farm in Iceland. Later Ingimundr settled in Vatnsdalr, north-western Iceland. There he erected a great sanctuary (hof mikit). When he dug the holes for the high-seat posts he found the lost hlutr ‘weight, object, thing’.161 A compressed version of this story is preserved in Landnámabók, where the “Freyr-image” and the high- seat posts are also mentioned.162 This tradition is very interesting in the context of modern archaeology. As noted above, sacred objects were deposited in and around the holes where the high-seat posts were erected. Sometimes these posts seem to have constituted parts of sanctuaries. At the hall of Hofstaðir in Mývatnssveit, as we saw above,

160 . . . á hlutnum markaðr Freyr af silfri . . . Vatnsdæla saga 9–10, Ísl. Fornr. 8, 26, 29f. 161 Vatnsdæla saga 15, Ísl. Fornr. 8, 42. 162 Þat var Freyr ok gǫrr af silfri. Landnámabók, S 179, Ísl. Fornr. 1, 218. 236 CHAPTER 6 objects such as seventeen stone gaming pieces, a human tooth, and a crucible fragment were deposited in holes close to the central hearth. They may have been placed there since these holes were used for the high-seat posts. Also in the halls of Borg, Lofoten, gold foils (perhaps representing Freyr and Gerðr) have been found in holes of the assumed high-seat posts, and at Lunda, Södermanland, deposited objects were found in a similar context (see ch. 4 above). It thus seems as if the transmitter of the Ingimundr tradition must have known that small sacred objects were placed in the soil around the high- seat posts of the hof sanctuaries and other ceremonial buildings during pre- Christian periods.163 The story about Ingimundr, recounted in Landnámabók and Vatnsdæla saga, in combination with the archaeological evidence, indicates that the high-seat posts really had a religious significance during the Viking Age, perhaps when the ceremonial building and the high-seat were consecrated (see ch. 8 below). There are also other aspects related to the high-seat posts which indicate that they played an important role in the pre-Christian religion and rulership. In Eyrbyggja saga 4 and Landnámabók (S85, H73) we read that images of Þórr were carved on high-seat posts which Þórólfr brought from Norway (see above). When he erected them in his hof at Hofstaðir, in Iceland, he fixed nails into them which were called “gods-nails” (reginnaglar), according to Eyrbyggja saga. The information about the reginnaglar has been regarded as genuine in this text. The first element of this compound, regin ‘(pagan) gods’, indicates that the word is old.164 The term reginnagli appears also in a Christian poem, although in a rather obscure context.165 Two kennings for nails found in the Þulur (AM 758 I, 4to) confirm that the compound indeed had a pagan religious significance; regingaddi ‘gods-sting’ and veraldarnagli ‘world-nail’. The latter term may indicate that such nails were hammered into the high-seat posts, since these posts were conceived as axis mundi symbols. At this place men could encounter the mythical world. It has been argued that the gold foils found by archaeologists at post-holes in halls were originally attached to the roof-supporting posts. Lars Larsson, for

163 Lidén 1969, 19. Most likely the hlutr mentioned in the saga(s) may be considered as some kind of amulet, perhaps a gold foil figure. However, no gold foil figures have been found in Iceland, see ch. 11 below. 164 Olrik 1910; Simek 2006, 262f.; Bertell 2003, 127. 165 Þás þú rekr/ fyr reginnagla/ bóka-máls/ bœnir þínar. “Hvor du fører dine bönner frem for præsterne” Glælognskviða 10, Skj, B1, 301. Finnur Jónsson’s Lexicon Poeticum (1931, 460) has ‘bókamál’, ‘bogsprogets gudanagel’, betegner vist ‘bogsprogets’, d.v.s. den gudelig læres ‘hovedstøtte’, dermed menes ‘helgener’. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 237 instance, suggests that many of the foils found at the cultic house of Uppåkra “had been fastened on the posts as decorations”.166 According to Birgit Arrhenius, a similar situation may be seen at the hall IA at Helgö. In post-hole 49, situated in room 3 (i.e. the “hall room”), she argues that a “phallus post” was placed. Arrhenius proposes that figural gold foils could have been used to decorate this wooden pillar, rather as the foils were used at Uppåkra.167 Perhaps the term “gods-nails” (reginnaglar) attested in the sagas should be related to the gold foils in such contexts. It could might as well refer to other ritual objects. The information in Eyrbyggja saga and Landnámabók that images of Þórr were carved on high-seat posts actually has very few parallels in the sagas.168 One example may be found in Fóstbrœðra saga 23, where Gríma is said to have owned a large chair, on the back of which a large image of Þórr was carved (var skorinn Þórr).169 Whether this image was carved on high-seat posts is quite uncertain, however. The decoration of some early medieval episcopal thrones found in Sweden suggests, however, that high-seat posts were decorated with mythical images. Representations of powerful animals are, for instance, cut on the elbow- rests of the episcopal throne (late twelfth century) that once stood in Suntak church, Västergötland. A similar decorated episcopal throne (thirteenth century) appears in the stave church of Heddal, Øst-Telemark. The wood carvings from the thirteenth century stave church of Hegge in Norway may also give us an impression of what the high-seat posts could have looked like.170 The Grimstad figure from the Roman Iron Age may likewise convey an idea of the appearance of the earliest high-seat posts.171 Perhaps archaeological evidence from the Mälaren region also indicates that high-seat posts could be decorated. In the hall (house 52) discovered at the ruler site of Lunda, Södermanland, three small cultic images have been found (see ch. 7). Gunnar Andersson and Sara Fritsch suggest that these figures were only miniature images of big cultic sculptures, shaped like posts and connected to the high-seat of the hall.172 They argue that figures A and B give a “post-like” impression. Inside the big hall at Lunda four extraordinary post-holes were discovered. They have a different character from the other post-holes where

166 L. Larsson 2006a, 251. 167 Arrhenius 2011, 23, 31. She also proposes alternative contexts for the gold foil figures. 168 But perhaps such high-seat posts should be associated with the ON term skurðgoð. 169 Fóstbrœðra saga, Ísl. Fornr. 6, 245. 170 Van der Sanden & Capelle 2001, 13f. 171 Van der Sanden & Capelle 2001, 19f. 172 G. Andersson et al. 2004, 144ff. 238 CHAPTER 6 posts carrying the roof were placed, not so much in terms of size—all the roof- carrying post-holes were big and deep—but because different types of objects were ritually deposited in these holes, such as loom weights and objects made of amber. No such finds were discovered in the other post-holes. According to Andersson and Fritsch, the decorated cultic posts of the high-seat were placed in these four particular post-holes.

6.3.9 Hliðskjálf—A High-seat in the Mythical World It thus seems as if high-seats, including high-seat posts, were regarded as places of honour in the cultic houses and banqueting halls of the investigation areas. They were intended for the owner of the hall: the king or chieftain. The Old Norse traditions report that high-seats also occurred in the mythical buildings. These traditions also suggest that the high-seat had both religious and ideological implications in hall contexts. When entering a high-seat, the god or mythical being was, for instance, enriched with supernatural abilities. The mythical references to high-seats also indicate that a micro-macrocosmic symbolism may have existed between them and real high-seats in the ceremonial buildings of the investigating areas. The most famous one is probably Óðinn’s “throne”, called Hliðskjálf,173 which according to Snorri was located in the hall called Valaskjálf.174 It is quite striking that this high-seat was particularly intended for Óðinn, the sovereignty god, and regarded as a seat of honour. The name Hliðskjálf is not mentioned in any Eddic lays, but it appears in the prose introductions to the Grímnismál and Skírnismál (For Scírnis). In both cases it refers to a seat from where the gods have a view over the whole earth. In this seat the gods also have experiences which may play a certain role for the future. The prose introduction to Skírnismál reports that the god Freyr once sat in Hliðskjálf:175

Freyr, the son of Njǫrðr, had seated himself in Hliðskjálf and looked into all the worlds. He looked into Jǫtunheimr and saw there a beautiful girl, as she was walking from her father’s hall to the storehouse. From that he caught great sickness of heart.176

173 This name is sometimes also written Hliðskiǫlf. See Steinsland 1991, 66ff.; Simek 2006, 152. 174 Gylfaginning 17. 175 Many scholars have felt that Skírnismál is quite a young Eddic lay, see Steinsland 1991, 30ff. 176 Freyr, sonr Niarðar, hafði sezc í Hliðsciálf oc sá um heima alla. Hann sá í Iotunheima, oc sá þar mey fagra, þá er hon gecc frá scála fǫður sins til scemmo. Þar af fecc hann hugsóttir miclar. Trans. Larrington. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 239

It has been argued that this prose was an integral part of the poem from the very beginning.177 This is quite plausible. When reproducing this mythical account in Gylfaginning (manuscript R), Snorri made Hliðskjálf into a central element in his narrative, for instance. It indicates that at least he regarded this motif from Skírnismál as quite genuine, even if he added some details and moral aspects to the episode:

It happened one day that Freyr had gone into Hliðskjálf and was looking over all worlds, and when he looked to the north he saw on a certain homestead a large and beautiful building, and to this building went a woman, and when she lifted her arms and opened the door for herself, light was shed from her arms over both sky and sea, and all worlds were made bright by her. And his punishment for his great presumption in having sat in that holy seat was that he went away full of grief. And when he got home he said nothing, he neither slept nor drank; no one dared to speak with him.178

It is interesting to note that Hliðskjálf is called “holy seat” (helga sæti) and that it was intended for Óðinn only, i.e. the ruler of the gods. According to Snorri, Freyr was punished for sitting there. This is partly expressed in another passage of Gylfaginning, where it is also emphasized that Hliðskjálf makes Óðinn into an all-seeing and all-knowing god. This place is thus in some sense numinous and endows the one who is seated there with supernatural abilities, such as the ability to see into other worlds. In chapter 17 Snorri writes thus:

There also is a great place called Valaskjálf. This place is Óðinn’s. The gods built it and roofed it with pure silver, and it is there in this hall that Hliðskjálf is, the high-seat of that name. And when All-father (Óðinn) sits on that seat he can see over all worlds.179

177 Steinsland 1991, 66. 178 Þat var einn dag er Freyr hafði gengit í Hliðskjálf ok sá of heima alla, en er hann leit í norðrætt þá sá hann á einum bœ mikit hús ok fagrt, ok til þess húss gekk kona, ok er hon tók upp hǫndum ok lauk hurð fyrir sér þá lýsti af hǫndum hennar bæði í lopt ok á lǫg, ok allir heimar birtusk af henni. Ok svá hefndi honum þat mikla mikillæti er hann hafði sezk í þat helga sæti at hann gekk í braut fullr af harmi. Ok er hann kom heim, mælti hann ekki, hvárki svaf hann né drakk; engi þorði ok krefja hann orða. Gylfaginning 37. Trans. Faulkes. 179 Þar er enn mikill staðr er Valaskjálf heitir. Þann stað á Óðinn. Þann gerðu guðin ok þǫkðu skíru silfri, ok þar er Hliðskiálfin í þessum sal, þat hásæti svá heitir. Ok þá er Alfǫðr sitr í því sæti þá sér hann of allan heim. Gylfaginning 17. Trans. Faulkes. 240 CHAPTER 6

Similar information is also repeated in the prose introduction to Grímnismál, which states that “Óðinn and Frigg sat in Hliðskjálf and looked into all the worlds”.180 It seems thus as if Hliðskjálf, according Snorri and the prose introduction to Grímnismál, was a high-seat where the ruler of the gods, Óðinn, must be seated.181 Sometimes other gods also enter this seat, such as Freyr and Frigg. When sitting there the gods are enforced with numinous abilities and cosmic visions; for instance, they can see over all worlds. There are some kennings indicating that the connection between Óðinn and Hliðskjálf is old. In a tenth-century lausavísa, Hallfreðr vandræðaskáld calls Óðinn Hliðskjálfar harri “the ruler of Hliðskjálf”, while Þórǫlfr skáld half a century later calls the god Hliðskjálfar gramr “the leader of Hliðskjálf”.182 It is interesting to note that in both kennings Hliðskjálf is related to the old designations of rulers, gramr and harri. Hliðskjálf thus seems to have been a pre- Christian symbol of authority and power.183 There has also been debate as to whether Hliðskjálf really should be described as a high-seat.184 As noted above, Snorri explicitly states that it was a high-seat. Several scholars have argued, however, that Hliðskjálf should be described as “Óðinn’s watch-tower”.185 Anne Holtsmark, for instance, accepted this interpretation. She argued that Snorri transformed the meaning of the name from “watch-tower” to “high-seat” under influence of the medieval royal throne.186 According to her, Snorri may also have been inspired by Christian iconography, where Christ is depicted as seated on a throne.187 In my opinion, Holtsmark’s interpretation is not convincing. As we have noted above, the high-seat, with its connotations of religious authority, was an

180 Óðinn oc Frigg sáto í Hliðsciálfo oc sá um heima alla. On this tradition, see de Vries 1956–57, §398. 181 When describing Ásgarðr he also adds in Gylfaginning 9: “There is a place called Hliðskjálf, and when Óðinn sat on the high-seat there, he saw over all worlds and every man’s activity and understood everything he saw.” Þar er einn staðr er Hliðskiálf heitir, ok þá er Óðinn settisk þar í hásæti þá sá hann of alla heima ok hvers manns athæfi ok vissi alla hluti þá er hann sá. Gylfaginning 9. It has been noted that Snorri is slightly inconsistent here, when he says that it was the place, and not the high-seat, that was named Hliðskjálf. See Steinsland 1991, 68; Simek 2006, 152. 182 Hallfreðr’s lausavísa is in Skj. B1, 158 and Skj. A1, 168, while Þórólfr’s stanza is in Skj. B1, 388. 183 Steinsland 1991, 69. 184 I am following Steinsland (1991, 69f.) in this overview. 185 According to Eric Elgqvist (1944), Hliðskjálf should be interpreted ‘the watch-tower of the band (lið)’. 186 Holtsmark 1964, 41. 187 Holtsmark has gained supported from Gerd Wolfgang Weber 1972. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 241 institution with roots in Late Iron Age society. Snorri did not need Christian iconography to relate Óðinn to a high-seat.188 The high-seat is actually a very natural position for the ruler of the gods to be seated when appearing in mythical halls, such as Valaskjálf. There is evidence in Viking Age skaldic poetry relating Óðinn to a specific high-seat. A mid-tenth-century lausavísa, made by Kormákr Ǫgmundarson, says that a dead man will soon drink mead together with the gods. This will take place in “the high-seat of Óðinn”:

myndak ǫl at Óðins i ǫndvegi drekka . . .189

This is repeated with a similar expression some 150 years later in the skaldic poem Krákumál 29, which says that Óðinn had sent dísir from Valhǫll to invite the dead man to his home. The dead man says that he will gladly drink beer in the high-seat together with the Æsir (glaðr skalk ǫl með ǫ́sum í ǫndvegi drekka).190 According to Frands Herschend we may have archaeological support for such a notion.191 He points out similarities between the dead ruler’s position in grave no. 8 (c. AD 800) in Valsgärde, and the general locality of the high- seat in the hall building. The dead ruler in the grave is in his symbolic high-seat together with all his equipment, weapons, the cup and cauldron, the drinking- horn, gaming table and counters, the fireplace and so on. A ceremonial feast is being celebrated, in the same manner as he used to do during his life in the hall. Perhaps this feast is taking place in Valhǫll, together with Óðinn. The “watch-tower theory” seems unnecessary when relating Hliðskjálf to a wider literary context about Óðinn and his high-seat. Nor does this theory suit well when it is related to the role of Óðinn in the ancient Scandinavian pantheon, i.e. his function as the ruler of the gods. The role as guard in the watch-tower is better suited to one of the servants in Ásgarðr. The ruler himself (i.e. Óðinn) is supposed to be seated in his high-seat in the hall.192

188 Cf. Steinsland 1991, 70. 189 “I will soon drink mead/ in the high-seat at Óðinn.” Skj. B1, 84; A1, 90. 190 Skj. B1, 656; A1, 649. 191 Herschend 1997, 49–59. 192 Philologists have questioned the “watch-tower theory” on linguistic grounds. Vilhelm Kiil (1960 and 1962) has suggested that Hliðskjálf originally was a noun (*hliðskjálf ) related to the concept seiðhjallr, i.e. the platform on which a seeress carried out her magic acts. Place-names including the term skiálfr perhaps reflect seats where seiðr took place. Cf. Vikstrand 1996. 242 CHAPTER 6

We may thus conclude that Hliðskjálf in the mythical traditions should most likely be related to Óðinn’s high-seat. When entering this seat Óðinn acquired a numinous ability and a cosmic vision. He was able to see all worlds in the universe, and thus he was also informed about everything that was going on in every place. It is possible that some mythical ideas about the high-seat were projected on the real high-seats of the investigation areas. These ideas may have enforced the religious and ideological character of these seats, and made them into a place where the real and divine worlds were joined. When the rulers entered their high-seats they were enriched with certain supernatural gifts, which perhaps were necessary when leading the cultic acts in the hall. The association of Óðinn with the high-seat may also have emphasized these rulers’ close relation to this sovereign god.

6.3.10 The Hearth—Archaeological Evidence of Hearths in the Investigation Areas The sources indicate that not only the high-seats but also the hearths constituted an important ritual and ideological element in the context of the interiors of ceremonial buildings. According to the interior description found in Fagrskinna 79, the fire and the hearth played a central ritual role during drinking ceremonies in the eleventh-century feasting halls (veizlustofur) in Svetjud, Norway and Denmark. Since the fireplace was located close to the high- seat(s), these kinds of rituals also emphasized the central role of the ruler during these feasts. Fagrskinna says thus: “It was customary among the chieftains to carry the beer around the hearth and to make toasts to their ‘high-seat man’ ” (see above). This type of ambulatory rituals (circumambulationes) related to hearths also occurs in other medieval descriptions of halls, referring to early Christian contexts.193 Snorri states, however, that this ritual was an old custom (siðr forn) in Norway. In Hákonar saga góða 14 (see also ch. 4 above) he also sets this type of ritual in a pre-Christian context:

Fires were to be lighted in the middle of the hof floor, and kettles hung over them. The sacrificial beaker was to be borne around the fire, and he who made the feast and was chieftain, was to bless the beaker as well as all the sacrificial meat. Óðinn’s toast was to be drunk first—that was for

193 See Óláfs saga kyrra 2: Þat var siðr forn í Nóregi, at konungs hásæti var á miðjum langpalli. Var ǫl um eld borit. Heimskringla III, Ísl. Fornr. 28, 204. Cf. Óláfs saga helga 57: Ef gǫfgir menn kómu til konungs, var þeim vel skipat. Við elda skyldi þá ǫl drekka. Heimskringla II, Ísl. Fornr. 27, 72. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 243

victory and power to the king—then Njǫrðr’s and Freyr’s, for good harvest and peace.

It thus seems as if Snorri regarded the ambulatory ritual around the hearth as an old pre-Christian custom, which was integrated with the sacrificial meals and cult performed for the pagan gods. In this account, the sacrificial meat was cooked on the fire of the hof. Archaeological evidence from the three investigation areas indicates that Snorri’s accounts are not totally distorted on this point. It seems as if hearths constituted an important element in the ceremonial buildings found there. They often appear in contexts of finds, which suggest that they had symbolic and religious functions beside the more practical aspects, such as giving heat and light.194 Most likely they played a central role during the ceremonial feasting (see ch. 9). Their locations, close to the high-seat, indicate that they also had important ideological implications. In the ceremonial building discovered at Helgö, for instance, a central hearth (no. 43) was discovered. It was located in the hall-room, but not exactly in the middle of it. It seems as if it had a dis- located position, slightly closer to the middle pair of roof-carrying posts, and close to the position of the assumed high-seat in the south-western corner of the room.195 Gold foil figures were found around the fireplace, as well as knives, loom weights and whetstones.196 As noted above, gold foils may be related to religious rituals, myths and ruler strategies. Also the fact that the assumed high-seat was situated close to the fireplace arouses similar associations. At Vallhagar on Gotland, the hearth was located in between the places of the two high-seats, in the middle of the hall (house 16). This position indicates that it played an important ceremonial role for the rulers at this site.197 The context of the finds also indicates ritual activities. Drinking vessels have been found close to it. Perhaps some kind of rituals were performed there, similar to the ceremonies described in the written sources. At several halls in the Mälaren area, hearths appear close to the places interpreted as high- seats. At the Garrison in Birka, for instance, the hearth was situated in the area where most high-status objects were deposited, i.e. the site of the assumed

194 One of Herschend’s (1993, 182) five criteria for halls is that “the hearth was not used for cooking or for handicraft”. In my opinion, at least ritual cooking may sometimes have taken place at the hearths inside the ceremonial buildings. But most likely such cooking often also took place outdoors at cooking pits etc. 195 Herschend 1995, 224. 196 See Herschend 1995, 1998, 28ff.; K. Lamm 2004, 51ff. 197 Herschend 1998, 25f. 244 CHAPTER 6 high-seat. A similar structure is also evident at the large hall of Lunda. It should be noted that outdoor cooking pits and hearths appear at several halls in the Mälaren area, for instance, at Lunda and Helgö. At such pits ritual cooking may have thus taken place outdoors, and not at the hearth inside the ceremonial building. Hearths have also been found inside the assumed ceremonial buildings found in Trøndelag and northern Norway. Ritual cooking took place, for instance, at the Viking Age ceremonial building at Mære. Quantities of animal bones and some characteristic cooking stones indicate this.198 Most interesting in this context is the fireplace structure of the Late Iron Age hall in Borg, Lofoten. The central fireplace was probably in daily use for cooking and for heating and light.199 In hall-room C there were also two fireplaces. One of them was an oblong pit and one was round. Both of them were filled with fire-cracked stones and pure charcoal, but none of them had ashes or burnt bones. This indicates a difference in use compared to the central fireplace. Gerd Stamsø Munch suggests that the reason why no ashes and burnt bones were found in the long hearth is connected to its use for cooking the slaughtered animal at feasts. “The fire would have been heated intensely before the meat was put into place, covered with red-hot stones and finally with sods, the effect being to bake the meat slowly in fire which did not produce ashes and burnt bones. The long-fire was probably even cleaned afterwards—this seems particularly important if it was reserved for the preparation of very special meals.”200 She also suggests that something similar may have been the case for the round hearth, and concludes that both of them may have been used for cultic activities. The round hearth is located very close to the corner, where the high-seat was assumed to be, and where the gold foil figures and other prestige objects were detected. This fact suggests that the hearths in room C may have played a religious-ideological function for the ruler of Borg. Together with the other ritual objects in room C (the high-seat, the gold foils etc.) it legitimized the chieftain’s political as well as his religious power. During the religious feasts it may have played a specific ritual and symbolic role, which emphasized the central position of the ruler. Certain ceremonies, such as ambulatory rituals, may have been performed at the round hearth in room C, during the ritual drinking, as suggested by the written evidence. Hearths have also been found in the ceremonial buildings of Iceland. Several fireplaces were found in the Viking Age hall (AB) of Hofstaðir, Mývatnssveit.201

198 Lidén 1969; 1996. 199 Herschend & Mikkelsen 2003; Munch 2003b. 200 Munch 2003b. Most likely this was a type of seyðir, see Vikstrand 1993, 79. 201 See Lucas 2009, 62ff., 112ff., 373ff., 386ff. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 245

In the northern room there were two great cooking pits which had been cleaned out repeatedly. Most likely food processing occurred in this room. The majority of the carbonized cereal grain came from this room, which supports this interpretation. Outside the north-western entrance fire-cracked rocks/ stones were found, which may have derived from the use of these cooking pits. There was also a hearth in the central room of the building.202 During phase II this hearth (2776) consisted of a rectangular stone kerb with a stone base and had a size of 0.7 × 1.2 m. It was located in the areas where the high-seat may have been situated and where several objects were deposited. This hearth was used for heat and light, but may also have had religious and ideological functions during the feasts. No ashes were found in situ during the new excavation, but Daniel Bruun reports charcoal and bones in and around it, indicating that it was used for cooking. Whether it was used for ritual or everyday cooking is hard to tell, however. In any case, it was in this area the ceremonial feasts may have taken place. The hearths in structures A2 and D (the minor hall) from phase II are also interesting in this context. When the hall was abandoned, cattle skulls were placed in two clusters outside the walls of these structures. In A2 there was a hearth (3039) consisting of a sub-rectangular cut (0.6 × 1 m), lined with stones and mixed with ashes.203 There was also a rectangular barrel pit (3100), 1.64 m in diameter, and 0.9 m deep with vertical sides and a flat base. Both suggest that this room was used for food preparation and storage. But the barrel had been removed at some point and the pit backfilled with metalworking waste.204 This action may have been symbolic in some sense. The excavators suggest that the hearth and the pit were used for preparation of food, perhaps in a ritual context, because the floor contained an extremely high proportion of burnt bone. The cattle skulls and a slaughtered sheep car- cass related to this room support this interpretation. There was also a hearth in the small hall (D). Hardly any bones were found there, however. It thus appears unlikely that preparation or cooking food occurred at that hearth. In any case, the cattle skulls suggest that this room may have had a special function. Hearths have also been found in other halls in Iceland, for instance at Aðalstræti, Granastaðir, Grelutóttir, Eiriksstaðir, Goðatættur II, Vatnsfjörður and Havitárholt IX.205 An oven and hearths were also discovered at the more specific and small cultic house (“blót house”) at Hólmur close to Hornafjörður

202 Lucas 2009, 115ff. 203 Lucas 2009, 121f. 204 Lucas 2009, 390. 205 Lucas 2009. 246 CHAPTER 6

(Höfn) in south-east Iceland.206 Perhaps ritual food was cooked at this house, but it is unlikely that ceremonial feasts with a lot of participants were celebrated inside it. It is possible that the ritual feasting took place outdoors or in a banqueting hall in the farm at Hólmur, 250 metres south-west. Around the cult house there was an area containing a vast amount of fire-cracked stones. There were also hearths and cooking pits which were related to religious activities. In some of them certain objects were discovered, such as gaming pieces, loom weights, burnt and unburnt bones and fire-cracked stones. Archaeology suggests that hearths appear in connection with possible ceremonial buildings in all three investigation areas. They appear in contexts of finds which indicate that they may have had a ritual and ruler-ideological significance. The medieval texts may provide some further information about these hearths.

6.3.11 The Mythical Dimension of the Fire It has often been argued that ambulatory rituals (circumambulationes) were made around objects which were regarded as sacred. The purpose of these kinds of rituals is often to demarcate the sacred object from the profane outer world.207 It is also possible that this type of ritual had an apotropaic function, to protect the sacred object from evil forces. Whatever function this kind of rituals actually had, the textual and archaeological evidence presented above indicates that the fires and the hearths were regarded as sacred objects during the Viking Age, and perhaps ambulatory rituals were performed in relation to them. The sacred character of the fire is actually explicitly stated in Kjalnesinga saga 2, where a vígðan eldr ‘consecrated fire’ is mentioned. It was located in Þorgrímr’s hof at Kjalarnes in western Iceland:

In front of Þórr [the image of the god] was a platform, which was very nicely decorated and covered with iron on the top. On this place there had to be a fire, which could never be put out. It was called the consecrated fire.208

This passage has been debated. Some scholars have argued that the information about the fire on top of the platform derives from the imagination of the

206 See Bjarni F. Einarsson 2008. 207 De Vries 1956–57, §212. 208 Frammi fyrir Þór stóð stallr með miklum hagleik gerr ok þiljaðr ofan með járni; þar á skyldi vera eldr, sá er aldri skyldi slokkna; það kǫlluðu þeir vígðan eld. Kjalnesinga saga 2, Ísl. Fornr. 14, 7. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 247 scribe.209 In my opinion, it is quite possible that consecrated and sacred fires were situated in the ceremonial buildings, even if it is certain that they were not placed on iron platforms. Perhaps the compound ON vébrandr (‘heilig brand’, ‘kyndel’, ‘fakkel’) in some way originally was related to such fire or hearth.210 There is some weak evidence from the investigation areas indicating the existence of a cult of fire.211 Most interesting in this context are two kennings for “fire” from the ninth-century skaldic poem Ynglingatal: sævar niðr, “the relative of sea” and Fornjóts sonr, “Forniot’s [the earth’s] son”. Claus Krag has argued that they must be young and dated to the twelfth century, since they presuppose Christian medieval teaching on the four elements. This doctrine of Empedocles was disseminated in Christian Europe through Bede’s commentary on the Bible and a Latin translation of Plato’s Timaeus. The poem with these kennings must have been composed after this doctrine had become known in Scandinavia approximately in the twelfth century.212 Krag’s arguments are not convincing. Ideas about the relations of the different elements to each other could very well have already been part of the existing culture of pre-Christian people in Scandinavia, as elsewhere. They are not restricted to Christian cultures. There are expressions similar to sævar niðr in Vedic literature, for instance.213 The Indic fire god Agni is sometimes called apām napāt “the child of water”.214 It has been argued that this is an ancient formula connected to a fire god among those people who spoke Indo-European languages.215 Fornjóts sonr is probably identical with the mythical being called Logi ‘flame, fire’, who appears in Hversu Noregr byggðist.216 He is the son of the primordial ancestor Fornjótr and appears with his “fire aspect” in late literary revisions of the mythical traditions, for instance, in the tale of Þórr and

209 So e.g., Finnur Jónsson 1898; cf. de Vries 1956–57, §272. 210 Alexanders saga (ed.) van Weenen, 41 (c. 1250). Brandr Jónsson was a priest and later bishop of Hólar and in one of the two manuscripts of this saga he is said to be the trans- lator. Perhaps he translated cruenta faces in the Latin original Alexandreis (made by Gautier (Walter) of Châtillon) to dreyrugir vébrandar; see Fritzner 1954 and Heggstad 1993. Cleasby & Vigfusson (1957 (1874), 76) states that ON brandr could be translated to ‘a brand, firebrand; even used synonymous with “hearth”’. 211 See Huth 1943. Critically considered by de Vries 1956–57, §256 and Simek 2006, 83f. 212 Krag 1991, 47–58, 60–67. 213 See Krause 1930. Cf. Bjarni Fidjestøl 1994, 195; Dillmann 2000a, 108; Sundqvist 2002, 44. 214 Oldenberg 1970 (1917), 117ff. 215 See Bader 1986, 39–123; Dillmann 2000a, 108. It has been argued that the Indo-European tribes had an old fire cult. See e.g., Å. V. Ström 1975, 54. 216 Flateyjarbók I, 21ff. See also Orkneyinga saga 1, Ísl. Fornr. 34, 3ff. Cf. Lindow 2001, 213f. 248 CHAPTER 6

Útgarða-Loki.217 Even if this tale is young, the very core of it may consist of an old riddle.218 These mythic traditions does not justify assuming a cult of fire or the existence of an Scandinavia fire god in the Late Iron Age. It is not impossible, however, that the fire/hearth in the ceremonial building was consecrated and considered sacred during the religious feasts in the three investigation areas. Perhaps this was evident when the sacrificial meals were being prepared. According to the Old Norse mythical sources, it seems as if the hearth and the kettle in Valhǫll played an essential role. Grímnismál and Gylfaginning report that the slain warriors (einherjar) were served with meat from the boar Sæhrímnir, which was prepared by the cook Andhrímnir ‘the one exposed to soot’ in the cauldron Eldhrímnir ‘the one covered in soot from the fire’. In Grímnismál 18 it says thus:

Andhrímnir lætr í Eldhrímni Sæhrímni soðinn, flesca bezt, enn þat fáir vito, við hvat einheriar alaz.219

The martial life in Valhǫll, as described in Grímnismál and Gylfaginning, resembles the situation in the real aristocratic warrior halls of Svetjud and Norway (see ch. 15 below), where ritual meals were prepared for the retainers in cauldrons hanging over the sacred fires.220 The cauldron Eldhrímnir may be interpreted as a mythical reduplication of a real cultic cauldron. It is interesting to note that the verb used for boiling the boar Sæhrímnir in this cauldron is sióða ‘to seethe, to boil’, i.e. a term used when preparing the sacrificial meals in real cultic contexts. This too indicates that the actions which take place in Valhǫll may be interpreted as mythical reflections of real ceremonial banquets (see ch. 9 and 15 below).

The descriptions of Valhǫll and other mythical buildings resemble in general the image we have of the real cultic buildings and feasting halls in the investiga-

217 Gylfaginning 45 and 46. 218 Simek 2006, 192. 219 “Andhrímnir has Sæhrímnir boiled/ in Eldhrímnir,/ the best of pork; but few know/ by what the Einherjar are nourished.” Trans. Larrington. 220 Cf. Nordberg 2003, 85ff. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 249 tion areas.221 Like the real ceremonial buildings, Valhǫll was decorated and visible from afar to those approaching it. With no doubt the image of the mythical hall originally was based on real buildings. Inside the mythical hall there were seats for the gods, and in addition there was a high-seat which belonged to Óðinn. When entering this seat Óðinn acquired a numinous ability and a cosmic vision. Most likely the real high-seats were also regarded as symbolic centres and interfaces between the real world and the cosmic world. During the sacrifices and the religious feasts, the ruler and/or house-owner therefore had to enter the high-seat, in order to create a relation between this world and the divine world. In both the human feasting halls and the mythical halls the hearth seems to have been an essential place, where different types of rituals were performed in connection with the ceremonial meals.

6.4 The Exterior Ritual Structure of the Sanctuaries

Sometimes we also find information about the exterior ritual structure of the ceremonial buildings in the sources referring to Svetjud, Trøndelag and Iceland. In connection with these buildings we may find, for instance, sacred trees (or groves), holy wells, springs or lakes, holy mountains and sacred meadows. There are also outdoor areas with pits for ritual cooking and sacrifices. These ritual elements sometimes appear in different types of combinations with each other, for instance, a ceremonial building, may be related to a sacred tree and/or a holy well. Such a structure is occasionally found in mythical traditions as well, where we may see, for instance, a cosmic tree and a well beside an essential hall building. In my opinion, these mythical references and structures indicate that a micro-macrocosmic symbolism was related not only to the ceremonial building, but also to the outdoor sanctuary surrounding this building. The similarities between the real cultic sites and the mythical landscape also indicate that sanctuaries were organized according to a common cultural or cosmic model. For some people, however, these ritual structures were plainly the material form of human and divine power.

221 As will be noted below, the martial dimension of the eschatological Valhǫll mythology was probably more relevant for warrior lords with warrior bands appearing in Svetjud and Norway. The institution of warrior lords probably did not exist in Iceland. See ch. 15 below. 250 CHAPTER 6

6.4.1 Uppsala and Ásgarðr Detailed information about the structure of the outdoor sanctuary may be found in the written sources referring to Svetjud. Adam of Bremen, for instance, commented on the surroundings of the Uppsala “temple” in scholion 138:

Near this temple stands a very large tree with widespreading branches, always green, winter and summer. [. . .] There is also a spring at which the pagans are accustomed to make their sacrifices.

This description resembles information from mythical traditions about Yggdrasill and its wells. According to Snorri, Yggdrasill was the biggest and the best of all trees. Its branches spread out over the whole world and extended across the sky. One of its roots reached down towards Mímir’s well, where Óðinn deposited his eye in order to drink of the knowledge the well contained. Another root was located among the gods at Urðr’s well and the third root extended down to the well Hvergelmir.222 Even though Snorri described three different wells, they are usually interpreted as different names of one and the same mental figure, i.e. the cosmic well.223 The tree and the well(s) are also mentioned in Eddic lays as being related to each other. Grímnismál 44 tells that the ash Yggdrasill “is of all trees best”. The detail that the tree was always green and stood beside a well is included in Vǫluspá 19:

Asc veit ec standa, heitir Yggdrasill, hár baðmr, ausinn hvítaauri; þaðan koma dǫggvar, þærs í dala falla, stendr æ yfir, grœnn, Urðar brunni.224

222 Gylfaginning 15. 223 E.g., Simek 2006. 224 “An ash I know there stands,/ Yggdrasill is its name/ a tall tree, showered/ with shining loam./ From there come the dews/ that drop in the valleys./ It stands forever green over/ Urðr’s well.” Trans. Larrington. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 251

Eddic lays also describe how the tree extended its branches wide.225 Like the tree in Uppsala, the cosmic tree in the mythic traditions seems to stand beside an important building. According Grímnismál 25–26 there was a tree called Læraðr situated beside Valhǫll (á hǫllo Heriafǫðrs).226 Sometimes the mythic traditions emphasize that a well was situated beside Hár’s [Óðinn’s] hall, i.e. Valhǫll. In Hávamál 111 we hear this from Óðinn:

Mál er at þylia þular stóli á, Urðar brunni at; sá ec oc þagðac, sá ec oc hugðac, hlýdda ec á manna mál; of rúnar heyrða ec dæma, né um ráðom þǫgðo, Háva hǫllo at, Háva hǫllo í; heyrða ec segia svá: . . .227

The similarities between the description of Uppsala and the mythical traditions are striking, and have led to debate in previous research. Some scholars have argued that the “temple”, the tree and the well in Adam’s account of Uppsala are actually descriptions of a mythical landscape. These scholars feel that Adam misunderstood the mythical traditions surrounding Valhǫll, Yggdrasill and the well of Urðr and confused it with reality. Adam’s text therefore cannot be accepted as a trustworthy source on Viking Age sanctuaries, they say.228 In my opinion, these connections may lead to another conclusion. In what follows, I will argue that the mythical references in Adam’s description render a cultic reality. The temple, the tree and the well may have been deliberately arranged as a correspondences of notions about the mythical landscape.

225 Fjǫlsvinnsmál 13–14. 226 Cf. Gylfaginning 39. 227 “It is time to declaim from the sage’s high-seat/ at Urðr’s well;/ I saw and was silent, I saw and I considered,/ I heard the speech of men;/ I heard talk of runes nor were they silent/ about good counsel;/ at the High One’s hall, in the High One’s hall;/ thus I heard them speak:. . .” Trans. Larrington. I cannot agree with Grønvik (1999) that this stanza refers to an earthly place. See Sundqvist 2010a. 228 Bruun and Jónsson 1909; Alkarp 1993; 1997; 2007. Anders Hultgård argues that the description in the scholion basically is poetic-mythical. He believes that the motif may be based on a common mythical heritage. See Hultgård 1997, 26. He does not comment on the possibility that the cultic place in Uppsala may be based on a mythical model. On the tree in Uppsala, see also Läffler 1911. 252 CHAPTER 6

This is not unique to Uppsala, as it can be seen at cultic places in other parts of Scandinavia and on the Continent as well. In my opinion, these cosmic references to Uppsala were significant for the rulers. Uppsala was not only the site of the famous temple, it was also a political and economic centre. The rulers of Uppsala used mythical traditions about the cosmos and the divine world in order to gain legitimacy and power. Such a mythical or cosmic strategy may also be seen at other cultic sites in the investigation areas of the present study, as well as in all Germanic Europe.

6.4.2 Cultic Trees and World Pillars The existence of holy trees and wells is very common in texts referring to Germanic sacred sites. In the ecclesiastical polemics against pagan customs in Continental Europe, for instance, the expression arbor et fons seems to announce a pagan cultic place. It appears in texts from the Council of Tours in AD 567 to the capitularia in the age of Charlemagne.229 In England too, the Church struggled against the pagan worship of trees and wells.230 There is also extensive material indicating that people living in northern Europe had cul-

229 A document from the Council of Tours in 567 describes how converted people who still perform pagan rituals should be treated. Some of them worshipped mountains, trees and wells. Concilium Turonense, MGH, Legum, Sectio III, Concilia, Tomus I, 133. In a letter from Gregory I to Queen Brunhilde 597 it is stated that the cult of tree does not exist any longer. Gregorii I. Papae, MGH, Epistolarum, Tomi II, Pars III, 7. See also Clemen, Fontes, 1928, 30. A Langobardic law, compiled during King Liutprand’s reign in 727, stipulates the fines for those who worshipped trees and wells. Liutprandi Leges, Anni XV, Cap. 84. Leges Langobardorum 643–866, F. Beyerle (ed.) 1962 (1947), 139. See also Clemen, Fontes, 1928, 38. At Concilium Germanicarum, led by Bonifatius in the year 743, the grounds for the Carolingian mission strategies were worked out. For instance, regulations against pagan customs were formulated. Some of these regulations were presented in the text Indiculus superstitionum et paganiarum. In this text we read in ch. 6 De sacris silvarum, quae nimidas vocant “About forest sanctuaries, which they call Nimidas” and ch. 11 De fontibus sacri- ficiorum “About sacrificial springs”. MGH, Legum Sectio II, Capitularia Regum Francorum, Tomus I, 222f. See also Clemen, Fontes, 1928, 42f. In Capitulatio de partibus Saxoniae (769) the cult of wells and trees was forbidden. Sacrifices at groves in a heathen fashion and the custom of making meals in honour of the demons were also rejected. MGH, Legum, Sectio II, Capitularia Regum Francorum, Tomus I, 69. See also Boudriot 1928, 34f., 38ff.; Homann 1976, 107ff.; Nilsson 1992, 27ff. 230 In ecclesiastic regulations formulated during the reign of King Edgar 959–975 it is stated that every priest (preosta) must encourage Christianity and fight against heathen practice (hæðendōm) and condemn cult at wells, trees and stones (forbeode wil-weorðunga . . . tre- owum & on stānum). ALIE 1840, 248. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 253 tic trees or trunks, symbolically representing the world-tree or cosmic pillar.231 The most famous world-pillar in Germanic contexts is probably the great tree trunk called Irminsul, which was worshipped among the Saxons at Eresburg.232 Rudolf of Fulda (c. 865), writes:

They [the Saxons] worship green trees and wells. They also worship a large tree trunk, erected under the naked sky; in native language they call it Irminsul, which in Latin means world-pillar, since it supports everything [in the world].233

Rudolf’s text is not a direct source, since it is built on older accounts and there seem to be secondary elements in it, such as the interpretation of the name Irminsul.234 However, his information that Irminsul was a big tree-trunk or pillar seems to be reliable, since the name Irmin-sul (Ermin-sul) means ‘the large pillar’.235 It is thus possible that the pillar or tree trunk Irminsul was an important cultic object at the chief sanctuary of the Saxons. Most likely it referred to some cosmic symbolism.236 Cultic trees may be seen in Uppsala, not only in Adam’s text.237 According to the U-version of Hervarar saga, a sacrificial tree, called blóttré, stood at the assembly place of the Svear, perhaps in Uppsala (see the text in ch. 9 below).

231 About tree cult and cosmic pillars in Eurasia, see Holmberg (Harva) 1922; Drobin & Keinenen 2001. 232 See in particular Palm 1948. Cf. Springer 2000; Drobin & Keinänen 2001; Sundqvist 2004a; 2007; Tolley 2009. 233 Frondosis arboribus fontibusque venerationem exhibebant. Truncum quoque ligni non parvae magnitudinis in altum erectum sub divo colebant, patria eum lingua Irminsul appellantes, quod Latine dicitur universalis columna, quasi sustinens omnia. Translatio S. Alexandri, MGH, Scriptorum, Tomus II, 676. See also Clemen, Fontes, 1928, 60f. My trans. 234 Palm 1948, 86ff.; see however Springer 2000. Widukind, who was a monk in Corvey, records in his chronicle Rerum gestarum Saxonicarum (c. 967) a cultic image made as a pillar which was devoted to the god Irmin: . . . effigie columpnarum . . . Hirmin . . . dicitur, (ed.) Lohmann & Hirsch 1935, 20f., see also note 3 there. The information that Irmin was a god or forefather is considered uncertain. See e.g., Palm 1948, 89ff.; Simek 2006, 175f.; Maier 2000b; Drobin & Keinänen 2001, 140f. 235 Cf. ON jǫrmun-, OE eormen- ‘large’, ‘enormous’, ‘elevated’ ‘wide’; in jǫrmungandr, ‘the big [enormous] stave’, i.e. the Miðgarðsormr; in OE eormengrund ‘wide world’; and in a heiti of Óðinn, Jǫrmunr. See e.g., Fritzner 1954, Vol. II, 244; Clark Hall 1916, 93; Palm 1948, 92f.; Maier 2000b; Drobin & Keinänen 2001, 141. It is unlikely that Irminsul should be interpreted as ‘pillar of the god Irmin’. See Simek 2006, 175. 236 See e.g., Simek 2006, 175f.; Maier 2000b. 237 See Palm 1942, 101f.; 1948, 98ff. 254 CHAPTER 6

When the Svear dethroned King Ingi Steinkelsson (r. 1080–1110), who refused to perform sacrifices on their behalf, his brother-in-law Sveinn stepped forward and told the Svear that he would perform sacrifice on their behalf if they would give him the kingdom. All assented to that proposal. “A horse was led to the meeting place, dismembered and distributed for eating, and the sacrificial tree (blóttré) was reddened with the blood.”238 This text is usually regarded as an uncertain source on pre-Christian religion, but the information about a cultic tree harmonizes with Adam’s text. There is also archaeological support for cultic trees in Western Sweden. During excavations under the church of Frösön in Jämtland, numerous animal bones were discovered at a mouldered birch stump. Sixty per cent of the bones belonged to wild animals, mostly bears, and 40 per cent were domestic animals. The bones and the stump were dated to the tenth century.239 It should be noted that the church was called Hoffs kirkio in a document from 1408. Still today, a place south-east of the church is called Hov. It has been argued that the Hov names in Jämtland refer to buildings where cult practices were carried out.240 Thus, we cannot rule out the possibility that we have a structure at Frösön similar that in Uppsala; a cultic house situated close to a cultic tree. Many scholars have associated the ritual trees with the “guardian trees” (in Swedish called vårdträd and in Norway tuntre) which appear mainly in late folklore. Such trees were often situated on farms.241 A “guardian tree” could be of different species, for instance, oak, lime, ash or maple. These trees were related to the farm and the family living there and they concerned the fate of the family members. Such trees could protect the people living on the farm, against fire and other dangers. Rituals were performed at the tree, for instance during childbirth. These customs were probably old and they remind us about the great oak, barnstokkr ‘the child trunk’, which was placed in the middle of the Vǫlsunga hall, according to Vǫlsunga saga.242 In this case too we may find the combination of a feasting hall and a sacred tree, closely related to each

238 Hervarar saga. (Ed) Jón Helgason, 160. My trans. Cf. Hervarar saga, (ed.) E. O. G. Turville- Petre, 71. 239 Iregren 1989; Magnell & Iregren 2010. See also Näsström 2001a; 2002, 113. 240 Vikstrand 1993. 241 Mannhardt 1904–1905 (1874–75) I, 51–54; Palm 1948, 56ff.; F. Ström 1961; Tillhagen 1995, 35ff., 61ff.; Steinsland & Meulengracht Sørensen 1998, 30f.; Vikstrand 2001, 279f. 242 Vǫlsunga saga, 6f. According to Hrólfs saga kraka, a tree was situated in the middle of King Aðils’ hall in Uppsala. See Nordberg 2003, 164, note 1. Cf. Ninck 1935, 340. See also Gunnell’s (2001) interesting cosmic interpretation. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 255 other. This structure harmonizes well with examples presented above, even if it appears in a source that is usually considered to be of low value as evidence. As mentioned above, cultic trees may not always have been real trees, as at Frösön. According to Anders Andrén, the triangular stone settings in Swedish called treuddar ‘tricorns’ (or ‘tree settings’) which have been found at several places in Sweden and Norway should be interpreted as symbolic expressions of the world-tree Yggdrasill, with its three roots.243 They are mainly dated to the Late Iron Age and they appear in different find contexts. One Viking Age tricorn is situated just outside the cultic hall building at Helgö (see fig. 12a and 12b).244 It contained ritual depositions, such as crucibles, bread, arrowheads, pottery and an Arabic coin dated to 819. Under the stone setting, in the middle, a single post-hole was found, dating to the Merovingian Period. It too had remains of ritual depositions. Perhaps a cosmic pillar of the Irminsul type stood there before the stone setting was built. Torun Zachrisson has argued that the clay found around the post-hole represented the mythical-cosmic white mud called hvíta-aurr, which according to Vǫluspá 19 was poured (ausinn) on Yggdrasill.245 It is thus possible that both Helgö and Frösön have a ritual pattern reminding us of Uppsala: beside the cultic house there is a sacred or cosmic tree. Tricorns also appear in Trøndelag. At Skei, Steinkjer, for instance, several tricorns are situated close to a Late Iron Age burial field. Iron Age houses have been discovered there as well as cooking pits.246 An interesting large tricorn also appears at the assumed cultic site of Lunderhaugen in Brønny municipal- ity, beside an Iron Age burial mound (35 metres in diameter). In the middle of this structure four big stones were placed and several deposited objects were found there. One of the stones was interpreted as a phallus symbol. This place was used from 500 BC until at least AD 785.247 Andrén’s interpretation of tricorns as symbols of the cosmic trees has been applied to stone settings in other parts of Norway, more precisely in Rogaland. Bjørn Myhre has argued that the stone settings at the farms of Ullandhaug and Hauge-Tu, for instance, could be interpreted as cosmic symbols, representing the mythical world-tree.248 The tricorn at Hauge-Tu was situated close to a place where gold foils and a house

243 Andrén 2004, 406; 2014, 27ff. Tricorns have mainly been found in southern and central Sweden and in southern Norway, see Myhre 2005. 244 T. Zachrisson 2004a; 2004b. 245 T. Zachrisson 2004a; 2004b. 246 Smedstad 1993. 247 Beverfjord & Binns 1994. 248 Myhre 2005. 256 CHAPTER 6

Figure 12a One Viking Age tricorn outside the cultic hall building at Helgö (p. 242). Photo: Anders Andrén.

Figure 12b The plateau of the cultic hall at Helgö. Photo: Anders Andrén. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 257 foundations has been found, thus indicating the combination of ceremonial building and cosmic tree.249 In Norway, the onomastic evidence indicates that individual trees were regarded as sacred, as in the case of Eik, the vicarage in Sogndal, with the neighbouring farm Frøyland and the church on the farm at Bø. Another case is Eig in Lund, with the neighbouring farm Hove, as well as Eig in Bjerkreim, the neighbouring farm of Hovland.250 In the latter two cases we may perhaps see the configuration of a cultic tree located close to a ceremonial building (hof). There are also some indications of sacred trees in Iceland. According to Geirmundar þáttr heljarskinns, which is preserved in Sturlunga saga,251 a mountain ash (rowan) was situated at a sacred grove at Skarði. It was circumscribed with ritual restrictions and taboos. If a cow browsed at this grove, her milk had to be thrown away.252 Such folk beliefs may have been based on old traditions and the worship of sacred trees. It seems as if the place was still regarded as sacred after the conversion, because a church was erected at Skarði sometime during the Middle Ages.253 Some scholars have argued that the “high-seat posts” which were brought from Norway (including Trøndelag) to Iceland represented the cosmic pillar.254 The information in Eyrbyggja saga 4 that reginnaglar ‘gods-nails’ had been hammered into the high-seat posts at Þórólfr’s hof supports this notion. The kenning veraldarnagli ‘world-nail’ found in the Þulur suggests that such nails were hammered into the high-seat posts, since these posts were conceived as axis mundi symbols. These nails may have represented the Pole Star, the holding-point for the axis in the sky (see above). Also interesting in this context is the tradition about Hallsteinn Þorskafjarðargoði mentioned in Landnámabók (see ch. 5 above). According to this text, this chieftain sacrificed to Þórr, since he wished that the god could send him his high-seat posts. After some days a big tree came ashore, which he used for high-seat posts. Perhaps this tradition intimates that posts must come from a “cosmic tree”.

249 Sundqvist 2015. 250 M. Olsen 1915, 275; Brink 2001, 99. On the place-name Eik, see also Sandnes & Stemshaug 2007, 131. 251 Sturlunga saga (Ed.) Jón Jóhannesson, 19. 252 Magnús Már Lárusson 1965, 375. 253 . . . í þeim sama stað, er nú stendr kirkja at Skarði. Sturlunga saga 1946, 19. 254 See Davidson 1975, 178; Drobin & Keinänen 2001, 145ff.; Simek 2006, 262f; Nordberg 2003, 157ff., 170ff. 258 CHAPTER 6

6.4.3 Sacred Wells, Springs and Lakes When describing the cultic site at Uppsala Adam of Bremen states in scholion 138: “There is also a spring at which the pagans are accustomed to make their sacrifices, and into it to plunge a live man. And if he is not found, the people’s wish will be granted.” Scholars have been searching for this sacred spring in Old Uppsala. Recently the so-called “Odens brunn”, situated beside the Middle Mound, has been dated to the twelfth century by means of dendro- chronological analysis and interpreted as a sacrificial well in the cult of St Eiríkr.255 We cannot rule out the possibility that it has a pre-Christian background. David Damell has suggested that Adam’s well is identical with the little lake called Myrby träsk close to Old Uppsala, where a high content of phosphorus was detected.256 In 1921 Lennart von Post investigated the so-called “Minnurs brunn” (RAÄ 261) at the vicarage and stated that it was made by nature.257 It has also been suggested as a possible sacrificial well in Uppsala. Historians of religions have been critical of Adam’s description of the ritual practice at the sacrificial well in Uppsala. Anders Hultgård, for instance, argued that Adam’s description may have been affected by the medieval ordeals, where persons who had to prove their innocence were lowered into cold water (purgatio aquae frigidae).258 If they sank, they were regarded as innocent, but if they floated it proved they were guilty.259 Whether or not a cultic well existed in Uppsala was never discussed by Hultgård. There is much evidence to support the existence of cultic activities at wetlands in southern Scandinavia, from different times, for instance at Thorsbjerg, Vimose, Nydam, Kragehul, Illerup and Skedemosse.260 Especially sacrifices at springs are well documented, as at Röekillorna, Löderup parish, Gårdlösa,

255 Oral communication with the archaeologist Frölund 2013. See also: http://arkeologiup- plandsmuseet.wordpress.com/2013/10/11/gamla-uppsala-nytt-ljus-over-gammal-offer- brunn-2/. The well was first investigated by Martin Olsson (1912) and it appeared to be lined with wood. The wood was dated in the 1950s by means of dendrochronology; as the date obtained was 1659, it was rejected as a possible sacrificial well. Hult de Geer 1950; Alkarp 1993, 43. The recent investigation indicates that it must have been in use as a sacrificial well at least in the 1100s (in a Christian context). 256 Damell 1980. Cf. Duczko 1993b, 21. 257 In Lindqvist 1936, 82. 258 Hultgård 1997, 28. 259 The origin of these ordeals is much debated. See Nilsson 1991. 260 See e.g., Hagberg 1967–1977; Ilkjær 1990; 2000; Fabech 1994. Traces of ship parts have also been found in bogs and lakes in the Mälaren region, for instance at Örsmossen, Tensta parish, Lake Söderbysjön, Nacka parish, and Rickebasta, Alsike parish. These finds were probably traces of rituals, perhaps sacrifices. G. Larsson 2007, 238ff. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 259

Smedstorp parish, both in Scania, and at a spring (a tributary of the Ätran) situated at Hjuleberg farm, Abild parish, Halland,261 but also at many other places in Continental Europe.262 Cult at springs is also documented from Uppland. At Gödåker, in Tensta parish, for instance, there is a spring beside a burial field with finds, dating back to the Roman Iron Age. In this spring several sacrificial objects were detected.263 Charlotte Fabech has argued that sacrifices in water and wetlands existed in Scandinavia during the Stone, Bronze and Early Iron Ages, but disappeared at the end of the Migration Period.264 Torun Zachrisson, on the other hand, has stated that Fabech’s theory cannot be applied to the Mälaren area. Deposits in water continued there after the Migration Period.265 Lotte Hedeager has shown that ritual deposits in wetland also existed in the Viking Age at several places in Scandinavia.266 Most interesting is the little lake at Tissø, Zealand.267 The name Tissø has been interpreted as ‘the lake dedicated to Týr’ or perhaps ‘the lake dedicated to the god’.268 In this lake, several Late Iron Age finds have been made which have been interpreted as sacrificial objects, such as weapons and other artefacts. At the settlement beside the lake, a large ceremonial hall was raised. Palisaded enclosures were situated close to the south-western part of the hall. They enclosed a small structure that was interpreted as a cultic house. A similar pattern is also in evidence at the Late Iron Age farm of Järrestad, Scania, with the hall, an enclosed area with a small cultic building, and beside this hall a wetland area with sacrificial wells and springs. Ritual deposits have been found in the wetland area.269 A similar ritual structure may also be found at Late Iron Age central places in the Mälaren region. At Tuna in Vendel, a feasting hall was situated on a platform, just beside the famous boat graves. North of Tuna there is a farm name Torsmyra. Place-name specialists have interpreted it as ‘the bog dedicated to the god Þórr’.270 Perhaps wetland sacrifices were performed there. As far as

261 Stjernquist 1970; 1989; 1997; Carlie 2000. 262 Some of these springs are, however, from Roman-Celtic areas. See e.g., Capelle 2000; Müller-Wille 2002. 263 Stenberger 1971 (1964), 398ff.; Häringe 1991. 264 Fabech 1994, 169. 265 T. Zachrisson 1998, 117f. 266 Hedeager 1999; see also Stjernquist 1997, 14. 267 Jørgensen 2002. 268 Brink 2001, 97; cf. Holmberg 1986. Since the last element in the name Tissø has an initial s, the first element could be interpreted as an appellative *tí ‘god’ or the name Týr. 269 Söderberg 2003, 293ff.; 2005, 211ff.; 238, 252ff. 270 Vikstrand 2000. 260 CHAPTER 6

I know, however, no excavations have been conducted. South-west of Tuna there is also a well location called Odensbrunn, i.e. ‘Óðinn’s well’. The well is mentioned by both Bureus and Rhezelius, and in “Rannsakningar” it is reported as a sacrificial well. Rhezelius renders the name in a reduced form Onsbrönn, which indicates that it is old and authentic.271 Whether this place reflects a pre-Chris- tian sacrificial well is most uncertain, since no finds have been made there. It is not impossible, however, to imagine that the topography of Vendel, as well as that seen in Uppsala, was structured according to a certain cultural model, indicating that a sacrificial bog (or well) was situated close to a cultic hall. The sources report that sacrificial wells and springs also occurred in Iceland. Sometimes they appear in relation to ceremonial buildings. According to Kjalnesinga saga 2, there was a large ceremonial building (hof mikit) at Hof in Kjalarnes, south-western Iceland. Outside the door of this building there was a sacrificial fen/bog (or well/spring) (fen), called Blótkelda.272 Human sacrifices were thrown (steypa) into this fen during the sacrificial feasts.273 The information that humans were sacrificed in the fen has been doubted by scholars.274 According to other scholars, there is no reason to reject the particular idea that a sacred fen was located beside the ceremonial building at Kjalarnes. The old place-names at this site, Hof and Blótkelda, support this assumption.275 Perhaps we find the same structure at Hofteigur, Jökuldalur in Norður-Múlasýsla. At this site too we see a name indicating the presence of a ceremonial building (the compound Hofteigur) in combination with the name Blótkelda.276 There is also a place called Blótkelda at Möðrudalur, Fjöll, in Norður-Múlasýsla, indicating cult at sacred fens or springs in Iceland. There are other compound place-names in Iceland indicating cult in sacred springs. Most interesting is the place-name Goðakelda, situated not far from the feasting hall at Hofstaðir, Mývatn.277 This name could be interpreted as ‘the fen/

271 Vikstrand 2001, 127. 272 ON blótkelda can be translated ‘a fen near the heathen temple’ (Cleasby & Vigfusson 1957 (1874)), but also ‘offerkjelde’ (Heggstad et al. 1993). 273 En mǫnnum, er þeir blótuðu, skyldi steypa ofan í fen þat, er úti var hjá dyrunum; þat kǫlluðu þeir Blótkeldu. Ísl. Fornr. 14, 7. 274 See Hultgård 2002. 275 See Ólafur Lárusson 1942, 77; Magnús Már Lárusson 1965. Also the fact that sacrificial finds have been made in fens and bogs in other parts of Scandinavia may support these cultic conditions (see above). 276 See Ólafur Lárusson 1942, 77. Cf. Svavar Sigmundsson (1992, 249) for a more critical account. 277 Magnús Már Lárusson (1965, 59) refers to this name. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 261 spring of the gods’.278 Whether this name, and some of the other names presented above, really reflects pre-Christian cultic conditions is, however, somewhat uncertain.279 Svavar Sigmundsson states that it would be dubious to assume “a direct link between heathen worship and the many natural features all around the country with placenames incorporating goð and blót. Some of the goð names could equally be derived from the masculine noun goði (chieftain), and the blót names are so widespread that the assumption of any ancient derivation is called into question.”280 Restrictions against cultic actions connected to water appear in some sources referring to the transition period in Norway. According to “Viðbætir við Óláfs sǫgu hins helga”, for instance, which is included in Flateyjarbók, King Óláfr Haraldsson abolished the worship of rocks, woods, water, and trees when doing mission work in Norway,281 while Gulathings-Lov kristinréttur hinn nýi reports that it was forbidden to believe in gnomes and perform rituals at rapids.282 In the Norwegian medieval laws there are otherwise few regulations against cultic activities performed at sacred fens or springs. The lack of information about this may be a consequence of the possibility that this custom was taken over by the medieval Church.283 It should be noted, however, that there are scholars who think that the cult at sacred springs was introduced by the Church in Scandinavia, after the conversion there.284 This type of cult was related to the worship of local saints in Europe. When the cult of local saints was taken to Scandinavia, the spring cult was also introduced. There are some saintly springs in Scandinavia showing some curious features which may indicate that they were objects of cult even before conversion and the introduction of Christianity in this area. For instance, the custom of making food offerings in a spring dedicated to St Óláfr may indicate traces of cult continuity.285 Some

278 On the first element Goða- in place-names, see Ólafur Lárusson 1942, 76. 279 See Magnús Már Lárusson 1965, 59; Näsström 2001a; 2002, 120. Cf. Svavar Sigmundsson (1992) sceptical attitude to the goð-names. 280 Svavar Sigmundsson 1992, 254. 281 . . . aull blot braut hann nidr ok aull god . . . ok mork onnur blotskapar skrimsl bædi hamra ok hǫrga skoga votn ok tre ok oll onnr blot bædi meiri ok minni. Flateyjarbók III, 246. Cf. Åke V. Ström 1975, 215. 282 . . . trua a landvættir at se j lundum æða haugum æða forsom . . . NGL 2, 308. See also Nilsson 1992, 35. 283 See Näsström 2001a; 2002, 120ff. 284 See Arwidsson 1965 and Näsström 2001a; 2002, 120ff.; cf. Döring 2003. 285 This appears in information from 1777 at Olofskällan in Sankt Olof’s parish, in Skåne. See Arwidsson 1965, 55. 262 CHAPTER 6 of these “Christian springs” appear in the context of folk religiosity, which may also indicate that these cults built on ancient traditions.286 Perhaps the combination of a hall building and a sacred lake could be seen in southern Norway. Close to the hall at Huseby, close to Kaupang in the middle of the Tjølling area, there is a lake called Vittersen. Stefan Brink has argued that the first element in the Norwegian name contains the word vítr, véttr, vettr, or vættr, referring to a ‘supernatural being, spirit or god’.287 The name thus refers to a sacred lake, which was located not far from hall found at Huseby (see ch. 4 above).

6.4.4 The Cultic Site as a Reduplication of Ásgarðr One thing that is not mentioned in Adam’s text, but has been revealed by recent excavations, is the posts erected at intervals in a straight line, found both south and north of Old Uppsala, which may be interpreted as ritual demarcations.288 Similar demarcations have also been found near halls and/or cultic houses at Helgö in Uppland, Husby-Glanshammar in Närke, Hovde in Sør-Trøndelag, Tissø in Zealand and at Järrestad in Scania (see further ch. 8). In the mythic world there were fences called ON garðr which enclosed different places.289 According to Snorri, one of these fences was called Miðgarðr and it was built as a fortification round the world against the hostility of giants.290 The concept indicates that it was located in the middle of the world. In this place lived people, the gods and other divine beings. The specific enclosed place where the gods resided was called Ásgarðr.291 This was the location of Valhǫll as well as Yggdrasill and the well of Urðr. That some of these notions also existed in eastern Scandinavia may be attested by an earth-embedded rock inscribed with runes at Fyrby, Södermanland (Sö 56), which says that Hastæin and Holmstæin were the most rune-skilled men in Miðgarðr (mænnr rynasta a Miðgarði).292

286 Näsström 2001a; 2002, 120ff.; 2003. For Norwegian evidence of sacred wells and springs, see Bø 1965. 287 Brink 2007, 56. 288 See ch. 4 above. 289 La Farge & Tucker (1992, 80f.) explain ON garðr thus: ‘yard’, ‘enclosure’ and ‘an enclosed space’. Cf. Heggstad et al. 1993, 139; Vikstrand 2006; Løkka 2010; Cöllen 2011, 221. 290 Gylfaginning 8. Faulkes (1988, 171) interpreted Miðgarðr as ‘the rampart surrounding the world of men and protecting it from giants’. Miðgarðr also detones the enclosed area where men lived, i.e. ‘dwelling place in the middle’ (Simek 2006, 214), ‘the home of humans’ or ‘central enclosure’ (Lindow 2001, 228f.). 291 Gylfaginning 9. 292 Jansson 1987, 138. The Ritual Structure of the Sanctuaries and Cosmic Symbolism 263

It is thus not impossible to imagine that the religious topography and cult at Uppsala and other cultic sites in the investigation areas, at least for certain people, (i.e. those who had access to religious and cultural knowledge), were seen as a reduplication of the cosmic world. Perhaps real places and symbols in Uppsala represented the world-tree Yggdrasill and the mythical well(s). The ceremonial building in Uppsala and the activities that took place there might very well have reflected notions about the life in Valhǫll. Like Ásgarðr, Uppsala was at least for a period probably enclosed by a symbolic fence. All these ritual attributes in Uppsala were probably markers indicating an axis mundi, the centre of the Universe. But perhaps it is an exaggeration to say that the setting of Uppsala as a whole was a reduplication of Ásgarðr. It was rather a place with features reflecting a widespread cosmological concept, including several mythical notions and ideas.

To conclude, Adam’s description of Uppsala is thus, if not in detail then at least in essence, based on reliable information. Since cultic halls recurrently appear also at other places in combination with specific trees and/or wells, springs, lakes or bogs, this could, in my opinion, also have been the case in Viking Age Uppsala (even if clear archaeological evidence is lacking). I suggest that this structure reflects a widespread cosmological tradition, which seems to have existed in all three investigation areas as well as elsewhere in Germanic Europe. The mythical symbolism seen at these places had several functions. On a religious level, it indicated that these sites were thresholds to the divine world. At such places, man could meet the divine powers by performing rituals such as sacrifices. It also had ideological and power implications for the rulers of these aristocratic sites. When the ruler appeared, the scene needed to be set with specific properties. Religious symbols, such as representations of the cosmic hall, tree and well, created the appearance that the ruler’s authority came from a realm beyond politics, society and the natural world. Aspects of such symbolism and mythic-cosmic strategy for political authority are observable at rulers’ sites with sanctuaries in Scandinavia, and perhaps also on the Continent. Even for those who never perceived the symbolic dimensions of the sanctuary when approaching it, the impressive ritual structures, such as the hall, the mounds and the long row of pillars in Merovingian Period Uppsala, could nevertheless provide agency for the political leaders there and elsewhere, but in a more direct way. Also these structures may have belonged to a common tradition. chapter 7 The Relation between Rulers and Cultic Images

In the previous chapter I argued that several elements related to the ceremonial buildings in the three investigation areas may reflect a mythical symbolism. So too do the cultic images, which, according to the written sources, were placed inside these buildings. Anthropomorphic figures giving form to and/ or representing gods or mythical beings have been made in almost all cultures of the world since the Stone Age. They also appear very early in Germanic contexts.1 In the history of religions, a distinction has sometimes been made between the German concepts of Idole “idol” (from Lat idolum) and Götterbild “image of gods”.2 The former term refers to a symbolic representation of an abstract and multivalent religious content, while the latter term, “image of gods”, designates a figurative representation of a specific deity, which was used in cultic contexts.3 In this study I will apply the concept “cultic images”, which here refers to both anthropomorphic figures representing specific gods (i.e. “Götterbilder”) and more abstract expressions of divine power. Since this study also takes the materiality turn into consideration, it must be emphazised that the cultic images were not always perceived as symbols or representations of the divinity. For some people these images were plainly the way the gods were being present, how they were defined in the first place, and how they could be recognized.4 In the native language the concepts tréguð and skurðgoð were applied to such objects, but also terms such as trémaðr, líkneski (probably a translation loan of Latin simulacrum) as well as stafr and hlutr. The authors who described the ancient Scandinavian religion in Latin used concepts such as idolum ‘idol’ and simulacrum ‘image, idol’. I will present some cultic images which mainly appear in the context of ceremonial buildings in the three investigation areas. These images may be seen in different sources, both literary and archaeological materials. Some of them were miniatures and made of metal, while others were made of wood.

1 Van der Sanden & Capelle 2001. 2 Maier 2000a, 329. 3 See also the distinction between “an aniconic image” and “an iconic image” in e.g., Nordberg 2010 and 2011. 4 Cf. Engelke 2012, 213.

Since the religious-political leaders always strove to establish a close relationship to the divine world in different ways, they needed to surround themselves with images of the gods when entering the public arenas. By acting close to these images in the sanctuaries, they could display this relationship in a concrete sense to the participants. These images thus became important elements when constructing power in the ceremonial buildings by means of mythic- ritual strategies. They also depicted the gods in positions and acts which may be seen as ideals or prototypes of the earthly rulers , for instance, the authoritative and liminal act of being seated in the high-seat (see 6.3.4 above).

7.1 Miniatures and Medallions

Some of the most important evidence of the relationship between religious aspects and rulers in the context of ceremonial buildings probably comes from gold foil figures. These foils include a type of small images that have been associated with the mythical sphere. Some scholars have interpreted them as sacrificial objects or votives, while other scholars relate them to so-called “temple coins”. Most likely they were also important when constructing religious legitimacy, as they have often been found close to the high-seat, i.e. the place of the ruler and the owner of the hall. The gold foil figures have been found in halls and cultic houses at central places in the Mälaren region, Trøndelag, and northern Norway. However, they have not been found in Iceland. This may be related to chronology, since the foils are usually dated to the period 550 to 800/900, and may have been obsolete when Iceland was colonized during the Viking Age. In my opinion, the geographic distribution of the foils may also be related to other aspects as well. I will argue in chapter 11 that these objects were related to a specific religious strategy for rulership, which appeared in the upper elite and noble social contexts in Sweden and Norway, but never in Iceland. This is also the case for the Migration Period golden bracteates, which may likewise be related to a certain social stratum. These finds occasionally also appear in hall contexts and they certainly belong to a period which preceded the settlement of Iceland (see ch. 11). Amulets, miniatures, Þórr’s hammer symbols, ritual rings, and Þórr’s hammer rings have also been found in the context of ceremonial buildings and high-status milieus. They can be dated to several periods of the Late Iron Age, including the Viking Age (see ch. 10). They were also important means for constructing power at such places. In the context of this study the Viking Age miniature “high-seats” or “Klotzstuhl-Amuletten” (“Götterthronen”) are particularly interesting. They have been discovered in several parts of Scandinavia (see ch. 6). 266 CHAPTER 7

7.2 The Three Statues of Gods in the “Uppsala Temple”

According to Adam of Bremen, the Svear honoured the statues of the three gods in the “Uppsala temple”. Thor (Þórr), occupied a throne in the middle of the banqueting room while Wodan (Óðinn) and Fricco (Freyr) had places on either side (see ch. 4 above). It is reasonable to suggest that Adam used the Latin expressions statuas trium deorum, idolum and simulacrum to refer to cultic images made of wood.5 In other contexts of his book these terms seem to refer to wood images. In one passage Adam says that an “idol” (idolum) of Thor (Þórr) was placed at the thing place of the Svear.6 Wolfred, from England, smashed it with an axe, but was killed by the pagans for that deed. This cultic image was most likely made of wood, since Wolfred used an axe when destroying it. Adam also mentioned that a well-known “idol of Fricco (Freyr)” (Fricconis simulacrum) was located in Västergötland, not far from the church of Skara. When Bishop Egino visited this place he smashed it.7 It seems as if this statue too was made out of wood, since it “was broken into pieces” (in frusta concidit). Uncertain Old Norse sources refer to pagan cult figures made of wood among the Svear. According to the Þáttr af Ǫgmundi dytts ok Gunnari helminga a cultic image (skur(ð)goð, líkneski) of Freyr is said to have been in a sanctuary among the Svear.8 In Hauks þáttr hábrókar an image of the deity Lytir was brought to the hall of Uppsala.9 Even if such sources may not be trusted, the Old Norse concepts skurðgoð and tréguð indicate the existence of images made of wood among pre-Christian Nordic people.10 Adam’s statement about the statues of the three gods Þórr, Óðinn and Freyr in the Uppsala sanctuary should not be dismissed as pure fantasy. Place-names such as Odenslund(a), Torslunda and Fröslunda support a cult of these gods in pre-Christian Uppland (see ch. 4 above). Archaeological finds may also support the existence of cult images in pre-Christian Svetjud. A small Viking Age bronze statue discovered in Rällinge, Lunda parish in Södermanland, has a phallic shape and might represent Freyr (see fig. 13). The phallic shape resem-

5 These terms are used by Adam for the cultic images at the Uppsala sanctuary, see e.g., Adam II,58; IV,26–27. Hultgård suggests that the concepts of idolum ‘idol’ and simulacrum ‘image, idol’ belonged to Adam’s polemic and anti-pagan terminology. See Hultgård 1997, 22. 6 . . . ydolum gentis nomine Thor stans in concilio paganorum . . . Adam II,62. 7 . . . [Egino] Ibi [sc. In Scaramensi] etiam opinatissimum Fricconis simulacrum in frusta concidit. Adam IV,9. Adam also states that the bishops Adalvard and Egino travelled around in the communities of the Götar and smashed pagan “idols”. Adam IV,30. 8 Flateyjarbók I, 337–339. Cf. Ǫgmundar þáttr dytt, Ísl. Fornr. 9, 112ff. 9 Flateyjarbók I, 579f. 10 See e.g., de Vries 1956–57, §270. Johan Fritzner (1954), for instance, translated the term tréguð as ‘Gudebillede af Træ’. The Relation Between Rulers and Cultic Images 267

Figure 13 A small Viking Age bronze statue discovered in Rällinge, Lunda parish in Södermanland. Staten Historiska Museum. bles Adam’s description of Freyr in the “temple”. It has even been proposed that the small statue from Rällinge was shaped as miniature image of the big Freyr statue located in the “Uppsala temple”.11 It should be noted, however, that statues of Freyr were probably placed at other sanctuaries as well, for instance the well-known “idol of Freyr” at Skara, which perhaps was situated outdoors. Thus, images of the deity Freyr might have been quite common in eastern Scandinavia (see also on Lunda below). The walrus-ivory figure (miniature) from Lund in southern Sweden (height 4.7 cm) (from around the second half of the tenth century to the twelfth century) has also been related to Adam’s description of the cultic images in “the Uppsala temple”.12 According to Ivar Lindquist this figure represents Þórr.13 This figure has, for instance, the same characteristic features as Thor (Þórr) in Adam’s account, i.e. the deity from Lund is seated on “log-chair” (Sw kubbstol) similar to the way Þórr is seated on a throne in the “Uppsala temple”. Lindquist also compared the Lund figurine with a large chair described in Fóstbrœðra saga. An image of Þórr was carved on this chair.14 This chair was placed á stufugólf mitt. This statement harmonizes with Adam’s account: “Thor (Þórr), occupies a throne in the middle of the banquet room” (Thor in medio solium habeat triclinio). Lindquist also argued that the Lund figurine resembled a portrait, usually interpreted as Þórr, depicted on a runic store at Norra Åby, Södermanland (Sö 86). The archaic circular eyes, for instance, are very similar in both images. Finally he also argued that the ring-and-dot pattern on the back of the Lund figurine, shaped like a T, represented Þórr’s hammer (compare with the image of Þórr in Fóstbrœðra saga 23). These arguments, if taken

11 So e.g., Näsström 2001a; 2002, 201. 12 Perkins 2001, 63. 13 Lindquist 1963. 14 Fóstbrœðra saga, Ísl. Fornr. 6, 245. See also Perkins 2001, 56, 63. 268 CHAPTER 7 together, may indicate that the Lund figurine represents Þórr.15 The Lund statuette may thus give a suggestion of how an image of Þórr may have been shaped. Medieval laws also indicate that cultic images existed in Svetjud during the transition period. In the Law of Uppland (1296) we have the following expression: “Nobody shall worship pagan idols (?) (affguþum) or believe in groves and stones.”16 The OSw term affguþ most likely refers to an image representing a pagan god in this context, since it is related here to other cultic objects, such as the sacred groves and the stones. We must further assume that these practices and customs were still occurring when the laws were written down at the end of thirteenth century.17 Perhaps similar prohibitions on venerating pagan images are mentioned in the Guta Law: haita a . . . haþin guþ “invoke heathen gods (i.e. images)”. There is also evidence of outdoor cultic images in other reliable written sources, which may refer to customs used among eastern Scandinavian people. When describing the Rūs people (perhaps Svear) on the Volga Ibn Fadlān mentioned that after anchoring their ships, they made sacrifices at a tall piece of wood set up in the ground. “This piece of wood has a face like the face of a man and is surrounded by small figurines behind which are long pieces of wood set up in the ground.”18 These cultic images seem to be made out of tall pieces of wood, perhaps carved slightly to look like a man with a face. But they could also be regarded as aniconic symbols of deities, i.e. non-figurative images of gods, such as the Early Iron Age “Pfahlgötter”, which have been discovered at for instance Oberdorla.19 It has been suggested that the main image in Ibn Fadlān’s account represents Óðinn, “the leader of the tribe of deities known as Æsir”.20 It has also been argued that this figure may represent Freyr, since he was “particularly associated with the Swedes”.21 In my opinion, it is almost impossible to identify these representations of gods. Anyhow, the passage produces strong evidence that the Rūs used cultic images made out of wood during the Viking Age when sacrificing. Place-names in eastern Scandinavia, including those with the Old Swedish term staver ‘stave’ (Old Latvian and Lithuanian stãbas, stabs ‘pillar’ or ‘cultic

15 Perkins 2001, 67. 16 . . . ængin skal affguþum blotæ. ok ængin a lundi ællr stenæ troæ. UL Kk 1 pr, SSGL 3, 11f. My trans. 17 Nilsson 1992, 38. 18 A translation of this text is reproduced by Montgomery 2000, 9ff. 19 Hasenfratz 1997, 16; Van der Sanden & Capelle 2001, 19, 38, 86. 20 See Montgomery 2000, 10 and the literature cited there. 21 Foote & Wilson 1980, 389. E. O. G. Turville-Petre (1964, 255–56) argues that a sacrifice of an ox or a bull was most appropriate to Freyr, who seems also to have been thought of as a bull, while his sister was thought of as a cow. The Relation Between Rulers and Cultic Images 269 image’), may refer to a type of outdoor cult image or object. Staver appears in theophoric place-names, e.g., Nälsta (Nærthastaff, 1354) in Spånga, Uppland, ‘the staff or cultic image connected with the divinity *Niærdh’.22 The row of erected posts, both to the south and north of Uppsala, could perhaps be related to the Old Swedish term staver and cultic images, since ritual depositions were made in connection with some of them.23 These posts may, on the other hand, have had other functions (see ch. 6 above). There are in general very few cultic images made of wood preserved from Late Iron Age Scandinavia. One such image was however found on the periphery of the Mälaren area. Near Sundborn in the province of Dalarna a very simple figure was dug up in a peat bog. The modelling of this figure is very far from the description that Adam gives of the images, according to him, were placed in the Uppsala sanctuary. This figure is 122 cm long and it is a virtually non- carved product, recently dated to the Middle Ages.24 “A few strokes of an axe sufficed to indicate a head (with eyes and a mouth), shoulders, breasts (?), and knees.”25 Perhaps it should be related to the cult of Sami deities, since Sami lived in this area. There is also a quite exceptional figure from Nordmyra, Njutånger in Medelpad, also found in a peat bog. This figure, probably dating from the first half of the first millennium, has a “teardrop-shaped head resting on a very long neck that ends in a sphere which in turn rests on a hemisphere”.26 This figure may very well have been connected to a structure in a building, or to a vehicle; the hemisphere at the bottom of the figure indicates this. Representations of the gods may also be found on two dimensional images. The divine triad appearing in “Adam’s temple” may, for instance, be visible on the Skog tapestry, Hälsingland, and the picture stone of Sanda, Gotland (see fig. 10 above).27 External evidence thus corroborates that Adam’s description of the three images of gods in the Uppsala sanctuary may be based on sound information. According to Adam, these statues were erected in the triclinium, ‘the banquet room’ of the “temple”.28 The terminology used by Adam (libatio and libare) indicates that drinking rituals were performed during the ritual banquets there, for instance, when honouring the three gods represented in the images

22 Hellberg 1986a; Vikstrand 2001, 292–297. 23 Beronius Jörpeland et al. 2013; Sundqvist 2013. Oral information Jonas Wikborg. See also: www.ArkeologiGamlaUppsala.se, last accessed 1 May 2014. 24 Personal communication, Torun Zachrisson. 25 Van der Sanden & Capelle 2001, 22f. 26 Van der Sanden & Capelle 2001, 58ff. 27 See Anjou 1935; Jungner 1930. 28 Adam IV, 26. 270 CHAPTER 7

(see ch. 9 below). It is most likely that the rulers of Uppsala had some specific relation to these deities during such rituals. According to Ynglingatal, several of kings of the Ynglinga family were regarded as Freyr’s or simply the deity’s offspring (Freys ǫ́ttungr, Freys afspringr, týss ǫ́ttungr). Perhaps the images in the Uppsala sanctuary thus constituted ideological expressions in the public space, such as the close (genealogical) relationship between the ruler family and the divine world. It is also possible that the cult of Óðinn, Þórr and Freyr in this sanctuary was a deliberate strategy used by the rulers in order to gather different cultic groups and communities in Uppsala during the cultic feast, i.e. groups who worshipped these particular gods (see further ch. 14).

7.2.1 The Three Phallic Images from Lunda There are also other cultic images found in the Mälaren region which clearly can be connected to halls, cultic houses and aristocratic contexts.29 At the ruler site of Lunda, Södermanland, three small images were recently discovered (see fig. 14). Two of them (B and C) were found at a small house (house 53), north of the Migration Period hall building (house 52). This small building has been interpreted as a specific cultic house, comparable with the cultic houses found at Tissø, Järrestad, and Borg in Östergötland. The third image was located in the yard, to the south of the hall. This area too must be regarded as ritual space, since other “cultic finds”, such as ceremonial glass, have been discovered there. These figures are small, 2–3 cm high, and vary in appearance. Figure A is made of bronze, with a gilded head. Figure B is cast in bronze alloy and completely gilded, while C is made out of a gold rod and carved. Figure C, which is smaller than A and B, was deposited in a post-hole related to the small “cultic house”, while B was found close to the northern wall of this house. Despite the variation in appearance, the three figures were probably intended to represent one and the same god, since they all display the same characteristic features and attributes. For instance, all three are phallic, and their “gesture language” seems to be identical. The downturned arms with palms placed on the stomach and exposed frontwards may be an expression of dignity and respect.30 Also, the fact that they are gilded suggests that they were one and the same god, even if they may have been made by different craftsmen in different periods. In my opinion, these figures should in the first instance be

29 See G. Andersson et al. 2004; Skyllberg 2008. 30 Watt 2007, 141. The Relation Between Rulers and Cultic Images 271

Figure 14 At the ruler site of Lunda, Södermanland, three small images were recently discovered. Lunda A, Södermanland; Lunda B, Södermanland; Lunda C, Södermanland.

related to Freyr, although other deities are also possible.31 The most characteristic feature of them is their phallic nature. Since we have the important information from Adam of Bremen that the image of Freyr in “the Uppsala temple” was phallic (cuius [Fricco’s] etiam simulacrum fingunt cum ingenti priapo) it seems most plausible to interpret these figures as representations of the deity Freyr.32 Other sources indicate that the cult of Freyr was spread in this area

31 These phallic images have been interpreted as representations of the god Freyr by Anders Hultgård, Gunnar Andersson and Sara Fritsch (Hultgård 2003a; G. Andersson et al. 2004, 136ff.). Andersson and Fritsch also present alternative interpretations. Since figures A and C have visible belts they might represent the god Þórr. The belts on these images may refer to Þórr’s megingiǫrð ‘belt of power’, mentioned in the Old Norse mythical traditions. Gylfaginning 20; Skáldskaparmál 18. The weakness of this interpretation is that these images do not carry Þórr’s most central and important attribute, the hammer Mjǫllnir. Therefore, it is in my opinion not convincing to interpret these images as Þórr. As an alternative to the god Freyr, Andersson and Fritsch suggest that the images represent the god Ullr. Ullr’s name has often been related to Gothic wulþus ‘glory’ or ‘lustre’. In Wulfila’s Bible this term is used to translate Greek δόξα ‘glory, splendour, grandeur’. According to this interpretation Ullr’s name must be interpreted ‘the one who transmits lustre, the splendid one’. Since some of the figures at Lunda were gilded it would therefore be natural to connect these figures to Ullr. In my opinion, the etymology of the name is too uncertain to make this interpretation convincing. Many alternative interpretations of the name have also been proposed. For a short overview of the etymology of the name Ullr, see Vikstrand 2001, 166. 32 That a phallic cult existed in Viking Age Scandinavia could also be attested by means of the short narrative called Vǫlsa þáttr, Flateyjarbók, where ceremonies related to a horse penis are mentioned (see ch. 9 below). See also Hultgård 2003a. 272 CHAPTER 7 during the Late Iron Age. Many place-names in this area, for instance, have Frö- as the first element (see ch. 14).33 It has been argued that the three figures at Lunda actually were miniature images of big cultic sculptures, shaped like posts and connected to the high- seat of the hall.34 A possible parallel to this is the assumed phallic pillar in the hall on Helgö ‘the holy island’. One post-hole there had a very striking construction. Birgit Arrhenius describes it thus: “At the edge of the hole to the west, two large saddle querns were placed beside each other on a filling of gravel and clay mixed with dark soil. Both stones were turned upside-down, i.e. with the depressed grinding surface facing downwards and the vaulted underside facing upwards . . . In this way the post together with the two stones would have had a shape very close to a Roman phallus, to be seen e.g., on phallus amulets . . . My proposal here is that there was a large phallus post positioned almost in the centre of the hall.”35 In addition Arrhenius suggests that some of the gold foils found in this hall could have decorated this pillar. If Arrhenius’ interpretation is correct, we may see a cult of Freyr at Helgö, since the motifs of the foils (see above and below) and the phallic post could both be related to this deity.

7.3 Cultic Images in the Sanctuaries of Trøndelag and Northern Norway

The Old Norse texts often report that cultic images were also placed in the hof sanctuaries of Trøndelag and elsewhere in Norway. These texts also indicate that such images were made out of wood. In several cases we hear that rulers, such as Earl Hákon, had such images in their sanctuaries. It thus seems as if they were important for establishing a religious-political position. Many of these texts are, however, related to the Christianization of Norway, particularly to the mission actions of King Óláfr Tryggvason and King Óláfr Haraldsson. These stories are often related to certain topoi which were common in the hagiography and legendary sagas of the saints, and they must therefore be treated with great care if used as historical sources. In what follows I will pre sent some of these traditions, and discuss their value in the context of rulership.

7.3.1 In the Sanctuaries of Lade and Mære Snorri Sturluson reports in Heimskringla that several cultic images were erected in the hof sanctuaries of both Lade and Mære in the late tenth century. In Óláfs

33 Vikstrand 2001, 55ff. 34 G. Andersson et al. 2004, 144ff. 35 Arrhenius 2011, 23. The Relation Between Rulers and Cultic Images 273 saga Tryggvasonar 59, for instance, he describes how King Óláfr Tryggvason dismantled Earl Hákon’s hof at Lade, and removed all the property and all the decorations from it and from the image(s) of the god(s).36 When Snorri described the interior of the same hof sanctuary in Lade in Hákonar saga góða 14–17, he did not mention any cultic images. But indirectly he intimated that several gods were worshipped there since he stated: “They toasted Óðinn for victory and dominion (til sigrs ok ríkis) and Njǫrðr and Freyr for good crops and peace (til árs ok friðar).” In this case it was Earl Sigurðr who owned and ruled over the sanctuary at Lade. Some Old Norse texts, which may be older than Heimskringla, produce a similar image. According to one of the manuscripts of Óláfs saga Tryggvasonar eptir Odd munk Snorrason (AM 310 4to) (c. 1190), for instance, Earl Hákon had a hof sanctuary at Lade, which he had dedicated to several deities.37 There are some younger traditions mentioning that King Óláfr desecrated the sanctuary including the cultic images in Lade. Óláfs saga Tryggvasonar en mesta 150 (c. 1300), for instance, states thus:

After these things King Óláfr sailed up the Frith to Lade where he commanded the hof to be broken up (lét brjóta ofan hofit). He seized all the valuables that were in it, together with the vestments of the idols (alt skraut af goðonum), and he took from the door the large gold ring (tók gullhríng mikinn or hofshurðinni) that Earl Hákon had caused to be made. The hof and idols were ordered to be burnt together with the deities (brenna alt samman hofit ok guðin).38

This story seems to be based on Snorri’s description of the Lade sanctuary, although it is possible that the scribe or compiler of this saga also had other traditions available. In Óláfs saga Tryggvasonar, Snorri described how King Óláfr Tryggvason also attended the sacrificial feast at the hof of Mære. This passage also mentions several images, but Þórr seems to be the most important one in the Mære sanctuary. This sanctuary was probably protected by the chieftains together at the end of the tenth century:

36 . . . lætr brjóta ofan hofit ok taka allt fé ok allt skraut ór hofinu ok af goðinu/goðunum. Óláfs saga Tryggvasonar 59, in Heimskringla I, Ísl. Fornr. 26, 308. The reading af goðinu appears in manuscript Kringla, while the manuscripts of Frisianum and Jǫfraskinna say af goði- num “from the gods”. 37 Ok þat hit mikla hof er þar stóð hafði hann eignat ótalligum guðum. Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 193. 38 Óláfs saga Tryggvasonar en mesta I, Fornmanna sögur 1, 302. My trans. 274 CHAPTER 7

King Óláfr now entered the hof, accompanied by a few men and some of the farmers. And when the king came to where the gods were, he found Þórr sitting there as the most honoured of the gods, adorned with gold and silver. King Óláfr lifted up the gold-adorned rod he held in his hands and struck Þórr, so he fell from the podium (stalli). Then the king’s men ran up and shoved all the gods from their podiums (stallar).39

When it comes to Þórr there are striking parallels between Adam’ description of the “Uppsala temple” and Snorri’s account of the Mære sanctuary. Thor (Þórr) is, according to Adam, potentissimus eorum, “the most powerful of them”, while Snorri depicts this deity as “the most honoured of all the gods” ([Þórr] var mest tígnaðr af ǫllum goðum).40 It is quite clear that Snorri built this story in Óláfs saga Tryggvasonar on older traditions. According to Theodoric the Monk’s Historia de antiquitate regum Norwagiensium (XI) (c. 1180), for instance, Earl Hákon had gathered images in a house dedicated to the pagan deities at Mære. These images were burnt together with the house by King Óláfr Tryggvason. Snorri’s most important source for this passage was probably Oddr Snorrason’s Saga Óláfs Tryggvasonar (c. 1190) originally written in Latin. The Latin version is lost, but an Old Norse translation of it is preserved. Oddr depicted the image of Þórr inside the hof in Mære thus: “Þórr was in the middle of the sanctuary and he was highly thought of.”41 In this version too, King Óláfr insulted the pagan farmers by smashing their cultic images. In one version of this story found in Óláfs saga Tryggvasonar (Flateyjarbók) (1390) the author describes the image of Þórr in the sanctuary of Mære as sitting in the seat of a carriage: Þor sat j midiu . . . sa var umbunadr Þors, at hann sat j kerru.42 That the image of Þórr sometimes included a carriage is also mentioned in Njáls saga 88. In this text the outlaw Víga-Hrappr entered the sanctuary (hof ) owned by Earl Hákon and the hersir Guðbrandr in Guðbrandsdalr: “Then he saw Þórr’s carriage and he took from him [Þórr] another ring” (Þá sér

39 Óláfr konungr gengr nú í hofit ok fáir menn með honum ok nǫkkurir af bóndum. En er konungr kom þar, sem goðin váru, þá sat þar Þórr ok var mest tígnaðr af ǫllum goðum, búinn með gulli ok silfri. Óláfr konungr hóf upp refði gullbúit, er hann hafði í hendi, ok laust Þór, svá at hann fell af stallinum. Síðan hljópu at konungsmenn ok skýfðu ofan ǫllum goðum af stǫl- lunum. Óláfs saga Tryggvasonar 69, in Heimskringla I, Ísl. Fornr. 26, 317. Trans. Hollander. 40 Adam IV, 26; Heimskringla I, Ísl. Fornr. 26, 317. Cf. Dillmann 1997, 67. 41 Ok Þórr var í miðju hofinu ok hafði mest yfirlát. Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 280. See also Meulengracht Sørensen 1991b, 239ff.; Steinsland 2000, 114ff. 42 Flateyjarbók I, 320. The Relation Between Rulers and Cultic Images 275 hann kerru Þórs ok tekr af honum annan hring).43 This implies an image of Þórr sitting in a carriage in the sanctuary. This idea is probably based on old traditions. We may see that one of the mythological tags used by the tenth-century skald Kormákr Ǫgmundarson in his Sigurðardrápa (c. 960) was “Þórr is sitting in his carriage” (sitr Þórr í reiðu).44 This poem may be related to Trøndelag since it was composed in honour of Earl Sigurðr of Lade. Narratives about cultic images in tenth-century Trøndelag are also repeated in other late prose traditions. These traditions include some unique narratives which may not be found in e.g., Heimskringla. In some instances it is mentioned that cultic images of Freyr were placed in the hof sanctuaries in Trøndelag. According to Óláfs saga Tryggvasonar in Flateyjarbók, for instance, one image of Freyr (skurgoð þat er gert er eftir Frey . . . likneski Freyss) is said to have appeared in Þrandheim, i.e. Trøndelag.45 King Óláfr Tryggvason had set out to desecrate it. When he landed near the hof, he saw a stud of horses and was told that they belonged to Freyr. The King mounted the stallion as if to insult the god, while his men took the mares. They rode in triumph to the hof, where Óláfr derisively seized the image. He took it down from its stand (hio nidr godin af stǫllunum),46 put it under his arm, closed the sanctuary (hof) and rode away from the place. A similar tradition is also mentioned in Óláfs saga Tryggvasonar by Oddr munk Snorrason. In this text too, the king desecrated the image of Freyr at a hof in Trøndelag: hljóp konungr í hofit ok braut þar í sundr Frey.47 Several sources report that Earl Hákon of Lade worshipped Þorgerðr Hǫlgabrúðr. Images of this mythical being were placed in sanctuaries.48 Some of them seem to have been located in Trøndelag. In Flateyjarbók’s version of Óláfs saga Tryggvasonar, the statue of Þorgerðr is mentioned.49 It is stated that Þorgerðr had been Earl Hákon’s best friend and that she was worshipped in Trøndelag. A sanctuary of Þorgerðr is also mentioned in the Færeyinga þáttr (preserved in Flateyjarbók) and in Færeyinga saga 23.50 It seems as if it was

43 Brennu-Njáls saga, Ísl. Fornr. 12, 214. 44 Skj. B1, 69; A1, 79. 45 Flateyjarbók I, 401. 46 Flateyjarbók I, 401. 47 Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 264. 48 E.g., Flateyjarbók I, 144, 213f., 407–409. 49 Óláfs saga Tryggvasonar, Flateyjarbók I, 401, 407–409. 50 This sanctuary is described as having glass windows in Færeyinga saga: Þar var fjǫlði goða. Glergluggar váru margir á húsinu, svá at hvergi bar skugga á. Kona var þar innar húsit um þvert, ok var hon vegliga búin. Jarl kastaði sér niðr fyrir fætr henni ok lá lengi; ok síðan stendr hann upp ok segir Sigmundi at þeir skulu færa henni fórn nǫkkura ok koma silfri því 276 CHAPTER 7 located somewhere in Trøndelag. Njáls saga 88 mentions that Earl Hákon and the chieftain Guðbrandr had a great hof in Guðbrandsdalr, where they kept a triad of cult images representing the deities (or mythical beings) Þórr, Þorgerðr and Irpa.

7.3.2 Norwegian Chieftains and their Cultic Images There are also some other saga traditions reporting cultic images located in the sanctuaries of minor chieftains in Trøndelag. Óláfs saga Tryggvasonar en mesta 203, for instance, mentions a minor chieftain called Sveinn who lived in the northern part of Trøndelag. He had a hof on his farm, with many images of gods, the most worshipped one being Þórr.51 Sveinn’s son was Finnr, and he was baptized by King Óláfr. Finnr once came to his father’s farm and went to the sanctuary (gekk till hofsins). He cut the cultic images from their platforms (sky- fði goðin af stallonum), he stripped and plundered them of all that was valuable (reytti ok ruplaði af þeim alt þat er fémætt var). Þórr was struck with three blows (sló Þór 3 hǫgg), before he fell. Finnr placed a cord round his neck and dragged him down to the seashore. He sailed to the king (Óláfr Tryggvason) and displayed Þórr to him, chopped him up into splinters and then burnt him. One of the most extensive and detailed descriptions of a Norwegian pagan cultic image appears in the story about the hersir and Þórr worshipper Dala- Guðbrandr’s pagan insurgency, which is rendered in several texts.52 After the failure of this revolt, St Óláfr Haraldsson himself came to Sil and Hundþorp in Guðbrandsdalr, in order to convert the farmers. He was told that the farmers at this site still worshipped Þórr. When the farmers were told that King Óláfr was approaching Hundþorp it was suggested that Þórr should be taken out from their hof (berum út Þór ór hofi váru) in order to scare the king and his men.53 The image was described thus by Snorri in Óláfs saga helga:

“. . . he has a hammer in his hand and is of great size and hollow inside, and he stands on a kind of pedestal when he is outside. There is a profu- sion of gold and silver upon him. He receives four loaves of bread every day and also fresh meat.” . . . Then they saw a great crowd of farmers come up to the place of meeting who carried between them a big figure of a

á stólinn fyrir hana; . . . See Færeyinga saga, Ísl. Fornr. 25, 50f. See also a similar version in Færeyinga þáttr, Flateyjarbók I, 144f. We may see a similar description of Þorgerðr’s sanctuary in Jómsvíkingadrápa 32. 51 Óláfs saga Tryggvasonar en mesta II, Fornmanna sögur 2, 153, 162f. 52 E.g., Óláfs saga helga, Heimskringla II, Ísl. Fornr. 27, 183–190; Flateyjarbók II, 188–192. 53 Óláfs saga helga 112, in Heimskringla II, Ísl. Fornr. 27, 184. The Relation Between Rulers and Cultic Images 277

man all glistening with gold and silver. And when the farmers already at the place of meeting saw it, they all sprang up and bowed down before this idol. Then it was set down in the middle of the assembly.54

The king was not impressed, however. He had the image struck so it fell in pieces, and “out jumped mice as big as cats, and adders, and snakes”.55 This phenomenon is recurrent in the hagiography and the mission literature, the idea that the devil and his animals have their dwelling inside the images of pagan gods.56 The story also ends with the conversion of the recalcitrant farmers and St Óláfr’s admonitions that they must not decorate logs and stones with gold and jewelleries.

7.3.3 The Þórr Statue at Rauðr’s Sanctuary and the Influence of Hagiography The saga authors often describe how pagans worshipped cultic images. Sometimes the images seem to have been animated. Most likely the Christian writers wanted to demonstrate that these images were possessed by the devil.57 There is a story in Óláfs saga Tryggvasonar en mesta called Rǫgnvalds þáttr ok Rauðs, which illustrates this well.58 On an island located just off Hálogaland, north of Trøndelag, there was a man called Rauðr, who kept a sanctuary (hof ) with a cultic image representing Þórr.59 Rauðr was a chieftain and great sacrificer (hinn mesti blotmadr), and he performed a great sacrifice upon the image of Þórr to give it strength there in its sanctuary (magnade med myklum blotskap likneski Þors þar var j hofinu), so that a demon talked out of the image (at fean- dinn mælti vid hann or skurgodinu). The image could also move, and was seen to walk outdoors in the daytime with Rauðr, who often took it with him about

54 . . . ok hefir hann hamar í hendi ok mikill vexti ok holr innan ok gǫrr undir honum sem hjallr sé, ok stendr hann þar á ofan, er hann er úti. Eigi skortir hann gull ok silfr á sér. Fjórir hleifar brauðs eru honum færðir hvern dag ok þar slátr við. . . . Þá sá þeir mikinn fjǫlða búanda fara til þings ok báru í milli sín mannlíkan mikit, glæst allt með gulli ok silfri. En er þat sá bændr, þeir er á þinginu váru, þá hljópu þeir allir upp ok lutu því skrímsli. Síðan var þat sett á miðjan þingvǫll. Óláfs saga helga 112–113, in Heimskringla II, Ísl. Fornr. 27, 187f. Trans. Hollander. 55 . . . ok hljópu þar út mýss, svá stórar sem kettir væri, ok eðlur ok ormar. Óláfs saga helga, in Heimskringla II, Ísl. Fornr. 27, 189. 56 See Näsström 2001a; 2002, 200; cf. Lassen 2011, 215ff. 57 Cf. Ǫgmundar þáttr dytt (Ísl. Fornr. 9, 112ff.), where the statue of Freyr is animated. 58 This text has also been preserved independently. See Perkins 2001, 27ff. 59 See Flateyjarbók I, 291ff. Cf. Óláfs saga Tryggvasonar en mesta I, Fornmanna sögur 1, 294f. On this tradition see Perkins 2001, 27ff. 278 CHAPTER 7 the island. At the end of this story the pagan god lost a contest with King Óláfr Tryggvason and Rauðr converted to the new faith. This story cannot be used as a source for pre-Christian religion and rulers. However, it seems as if it was partly based on older traditions.60 Rauðr is probably mentioned in Rauðs þáttr in ramma, as preserved in Snorri’s Óláfs saga Tryggvasonar 78–80. According to Snorri, Rauðr was much given to making sacrifices and he was a great magician (blótmaðr mikill ok mjǫk fjǫlkunnigr).61 He also put up stiff resistance to King Óláfr’s mission. Snorri’s version states that Rauðr lived at Goðey (or Goðeyjar in the plural), which is located at Sálpti in Hálogaland. This place can be identified with modern Godøy in Saltfjorden, near Bodø. There are no outstanding archaeological finds on this island, but the spit of land immediately to the north-east that goes under the name Goðeynes had a considerable Iron Age settlement. One burial mound there was 20 metres in diameter, while another was 35 m in length and had the shape of a ship.62 The name Godøy could very well be interpreted as ‘the island of the gods’.63 If this interpretation is accepted we have a gen. pl. goða- ‘the gods’ as the first element, where the last vowel has fallen because of its hiatus position. Per Vikstrand does not rule out the possibility that the first element of Goðey could be a gen. sg. of an ON goði ‘cultic leader’.64 If so, the name could be interpreted as ‘the cultic leader’s island’/‘the religious specialist’s island’. Hence, in my opinion, the very core of this story may be related to some oral traditions, mentioning that a chieftain of Goðey or the Goðeyjar resisted Christianity. Perhaps this chieftain also was regarded as a strong advocate of paganism. There are some stories about a chieftain called Hróaldr in Oddr’s Saga Óláfs Tryggvasonar, which probably were Snorri’s sources, and which may indicate the existence of an old tradition.65 Whether this tradition originally included ideas about cultic images is uncertain. It seems as if the images in the late traditions had something of a literary function and illustrate the demonic aspect of the pre-Christian religion. But even if the narrative in Rǫgnvalds þáttr ok Rauðs has a literary and Christian tendency, the terminology and expressions may be pre-Christian, such as magna með miklum blótskap ‘to strengthen (with supernatural power) by means of much sacrifice and/or pagan worship’.66 It

60 See Perkins 2001, 35. 61 Óláfs saga Tryggvasonar, Heimskringla I, Ísl. Fornr. 26, 324. 62 O. Grimm 2011, 178f. 63 M. Olsen 1942, 61; Vikstrand 2001, 401f. 64 Vikstrand 2001, 401, note 307. 65 Perkins 2001, 163. 66 Cf. Perkins 2001, 58. The Relation Between Rulers and Cultic Images 279 is thus reasonable to envisage that a pagan cultic image was believed to be endowed with supernatural power during the period of transformation. The stories about Óláfr Tryggvason’s and St Óláfr Haraldsson’s “mission actions” in the sanctuaries of Lade, Mære and elsewhere in Trøndelag and Norway are influenced by hagiography, and they have several parallels in the genre of legendary sagas of the saints and the “conversion þættir”.67 They do not have much source value, as some of them obviously reflect clerical imagination. Yet some of these traditions appear in some of the oldest Old Norse prose materials, for instance Oddr’s text. Of course these motifs may be regarded as topoi connected to the medieval ecclesiastic literature there too.68 However, some ideas in these stories may also build on old oral traditions, such as the information about the existence of cultic images at the sanctuaries of Mære and Lade, for instance. In order to argue more convincingly for such a position we must turn to other source categories.

7.3.4 Staff and Podium: The Evidence from the Norwegian Laws Even if the Old Norse sources describing cultic images in sanctuaries in general are late and partly must be regarded as uncertain, we cannot rule out the possibility that images of deities really existed in Trøndelag and elsewhere in Norway during the Viking Period. In the Norwegian medieval laws we have certain restrictions and prohibitions connected to the use of certain cultic practices, which sometimes seem to include cultic images. The most extensive source describing such prohibitions may be found in Den ældre Eidsivathings- Christenret, which was used in south-eastern Norway. This law rules that nobody may have in their house such things as belong to paganism:

Nobody shall in their house have staff or podium, wizardry or “blot” or anything else belonging to the pagan custom.69

The staff (ON stafr) and the podium (ON stalli/stallr) may very well refer to the custom of having a cultic image placed on a sacrificial altar/platform in the house. This information actually harmonizes very well with Snorri’s description in Óláfs saga Tryggvasonar 59 of the interior of the hof in Mære, which says that Þórr sat on a stalli. Several other traditions connected to sanctuaries in Norway also mention the existence of ritual structures called stalli/stallr.

67 Cf. E. O. G. Turville-Petre 1964, 247; Perkins 2001, 33. 68 Cf. Lassen 2011. 69 Engi maðr skal hafa i husi sinu staf eða stalla. vit eða blot. eða þat er til heiðins siðar uæit. Den ældre Eidsivathings Christenret NGL 1, 383. 280 CHAPTER 7

The word stallr or stalli (m.) is translated by Johan Fritzner as ‘Stillads hvor- paa Afgudbilleder ere oppstillede, Alter’. In the medieval prose it thus refers to the pre-Christian sacrificial altar, as well as to the podium or platform for the cultic images. The basic meaning of the word must be ‘a stand upon which something could be placed’.70 Since we have the compound véstallr ‘sacred stand’ in the early poetic language we may assume that some type of altars or ritual platforms really existed in the Viking Age.71 Also the kenning referring to Óðinn, stalla vinr ‘the friend of the altars/podiums’ in Egill Skalla-Grímsson’s Berudrápa (c. 970) indicates the same.72 That the term was associated early on with some kind of altar may be supported by Fjǫlsvinnsmál 40, where we have the expression “to sacrifice at place of the altar” (blóta á stallhelgom stað). In this case the word *stallheilagr is an adjective with the meaning ‘altar-holy’ or ‘hallowed by an altar’.73 In the law the staff (ON stafr) seems to be related to the podium. Fritzner translates stafr (m.) as ‘noget som benyttedes ved afgudernes dyrkelse’.74 Here we must make a comparison with the information given by Ibn Fadlān about the tall pieces of wood set up in the ground, which may represent aniconic images of gods. The stafr ‘staff’ in the house may very well be an aniconic representation of the deity. The word staff may also be related to rods in enclosures around sacred sites and assembly places.75 This idea is based on information appearing in Den ældre Frostathings-Lov in relation to vébǫnd ‘sacred bonds’.76 This was a kind of band attached to the rods of such enclosure. It is uncertain whether these rods played any specific role in public cult. Since the word stafr appears in connection with the word stalli/stallr in the Den ældre Eidsivathings- Christenret, the interpretation ‘cultic image’ seems more plausible, at least in this context. Perhaps something similar to the stafr described in the written sources can be seen in the archaeological materials. In the assumed small hof found at Ranheim in Trøndelag, there were four post-holes arranged as a square. They

70 Heggstad et al. 1993, 407. 71 See Finnur Jónsson 1931, 629. Cf. Düwel 1985; Hultgård 1993; 1996, 50. 72 See Skj B1, 42; A1, 48; Finnur Jónsson 1931, 619. 73 La Farge & Tucker 1992, 247. 74 Fritzner 1954. See also T. Andersson 1992b, 91–93; Brink 1990a, 478; 1999c, 428. The cult image representing a deity called staver must be distinguished from the memorial stave which was raised to a dead person. In some inscriptions RSw stafr may refer to a “memorial stave”. Cf. Williams 1996b, 62; Peterson 1994, 61; Jansson 1987, 121f., 137. 75 Sanmark 2004, 164. 76 Heggstad et al. 1993, 487. The Relation Between Rulers and Cultic Images 281 had smaller diameters than the post-holes used in the outer walls and they had not been dug to the same depth. According to Preben Rønne, they did not play any part in supporting the roof. He suggests that these holes originally were posts or staves which also included images of gods.77 Den ældre Eidsivathings-Christenret also reports on “blót”: matblot. eða lær- blot gort i mannzliki. af læiri eða af dæigi . . . The terms matblót and leirblót refer to human-shaped images made of dough and clay. The exact function of these images is not mentioned in the law. It has been suggested that they were connected to magical practices.78 In my opinion, they might equally as well be related to the religious sphere. These figures may very well represent deities and/or mark the presence of them in the ritual room. The terms matblót and leirblót indicate that they were used in the worship of gods, perhaps during sacrificial rituals, because the verb blóta means ‘to sacrifice, or to worship through sacrifice’. Perhaps such images were placed on the stalli/stallr. Most likely other Norwegian laws also indicate the existence of cultic images.79 Den ældre Gulathings-Lov 29 mentions that “sacrifices are also banned for we are not permitted to worship any heathen god, or [any] mound, or [in any] heathen sanctuaries (hǫrgr)”.80 Perhaps the implicit meaning of the first part of this regulation is that sacrifices to an image representing a pagan deity were forbidden. A similar formulation is repeated in Kong Sverrers Christenret, which is directly derived from Den ældre Gulathings-Lov. In this law a prohibition against sacrifices to vættir ‘female creatures of supernatural kind’ is added.81 It is possible that the vættir were also represented by images, and thus such figures implicitly are indicated in the law. Also in Den ældre Frostathings-Lov, sacrifices to pagan gods were prohibited.82 Den ældre Eidsivathings-Christenret outlawed the keeping of any objects connected to paganism,83 while Den ældre Borgarthings-Christenret prohibited the population from putting their trust in any “objects (beings) that they sacrificed to (blótskap)”.84 The term blótskap has several meanings: ‘avgudsdyrkning, blotning’, ‘noko ein blotar til, avgud’,

77 Rønne 2011. 78 Cf. Sanmark 2004, 167. 79 See e.g., Nilsson 1992, 34ff.; Sanmark 2004, 158f. 80 Blot er oss kviðiat at ver scolom eigi blota heiðit guð. ne hauga. ne horga. Den ældre Gulathings-Lov NGL 1, 18. 81 Blott er os oc kuiðiat at vér skulum æigi blota hæiðnar vetter. oc æigi hæiðin guð ne hauga ne horgha. NGL 1, 430. 82 FL III:15, NGL 1, 152. 83 EL I:24, NGL 1, 383. 84 A guð skulu menn væll trua, en æigi a boluan eða a blot skapp. BL I:16, NGL 1, 351. 282 CHAPTER 7

‘avgudsoffer’.85 In the context of the law this concept may thus include the meaning cultic images.

7.3.5 Óðinn at the Church of Hustad As far as I know, there is weak archaeological evidence of pre-Christian cultic images in Trøndelag, except the gold foils, bracteates and golden medallion (see ch. 11 below). The decorated stone preserved in the vestry of the Hustad church at Inderøy, Eyna fylki, is at least one possible representation of a pagan god. Hustad church is one of the oldest stone churches in all Norway (c. 1160). The decorated stone was originally discovered in the churchyard in 1961. There is a cross decorated on the top of it. In addition there is also a figure below, which has been interpreted as a representation of Óðinn, since the face of the figure has only one eye. Other interpretations are also possible, for instance that the figure represents Christ.86 This stone is only one part of a large stone which must have constituted a great monument. It is today 90 cm high and 40 cm wide. The original dimensions are impossible to determine since the stone was damaged by dynamite when discovered. Perhaps the combination of a cross and an Óðinn representation indicates syncretistic beliefs or tolerance from both pagans and Christians during the phase of transition in the Early Middle Ages (c. 900–1000).87 The church at this farm was a “hǿgendiskirkja” and the people living there manifested their economic wealth by erecting a stone church. That Christianity reached this district early on may also be evidenced by the stone cross found at the farm of Rol, close to Hustad, where it was situated on a burial mound. This cross belongs to the very first phase of Christianity in Norway (c. 950–1050) and it is quite unique for Trøndelag.88 Perhaps the Óðinn representation may originally have been situated at a pagan sanctuary, which was controlled by the pagan chieftain of Hustad.

7.4 Cultic Images in the Sagas of Icelanders and the Image from Eyrarland

Several Sagas of Icelanders report that cultic images were erected in the hof buildings of chieftains’ farms in Viking Age Iceland. Eyrbyggja saga 4 mentions that such images of gods (goðar) were arranged in a circle right

85 Heggstad et al. 1993, 56. 86 Nordeide 2005; 2010, 39, 270ff.; 2012. 87 Cf. Røskaft 2003, 137; Nordeide 2005; 2010, 39, 270ff.; 2012. 88 Nordeide 2005; 2010, 39, 270ff.; 2012. The Relation Between Rulers and Cultic Images 283

Figure 15 A cultic image from Eyrarland, Iceland. round the platform (stalli) inside choir-like part (afhús) of the chieftain Þórólfr Mostrarskegg’s sanctuary (hof ) at Hofstaðir in western Iceland. Þórólfr had also high-seat posts in his new hof and Þórr was carved on one of them. It thus seems as if he was surrounded with images of gods when taking part in the ceremonies of his hof sanctuary. Þórr also seems to have been an important 284 CHAPTER 7 god for the chieftain at Þórsnes.89 In Kjalnesinga saga 2 we hear that Þórr was placed in the middle of the hof of the chieftain (goði) Þorgrímr (þar stóð Þórr í miðju) at Kjalarnes, and other images of gods were situated on either side of him. In front of Þórr there was a platform (stalli/stallr) with a consecrated fire (vígðan eldr). These reports of hof interiors from Iceland are probably to some extent dependent on Snorri’s descriptions of hof buildings in Trøndelag. There is one difference, however. The images in the Icelandic hof buildings are not placed upon the platform, but beside or around it. Other sagas mention that cultic images were placed inside sanctuaries in Iceland. In Hrafnkell Freysgoði’s sanctuary (hof, goðahús) at Aðalból in eastern Iceland, there were several statues of gods (goðar).90 Most likely one of them represented Freyr. In Harðar saga ok Hólmverja 19 a sanctuary with an image of Þorgerðr is mentioned. This sanctuary was erected at Bláskógar, and it was called hof Þorgerðar Hǫlgabruðar. Hence it seems as if one deity played a central role in the Icelandic ceremonial buildings. This deity may have had a specific relation to the owner of the building.91 According to the very young Fljótsdæla saga 26, images of several gods and goddesses were erected in the hof of Bersi inn spaki. They were placed in some sort of high-seats situated on a platform. In the “lesser high-seat” Freyr and Þórr were seated,92 while Freyja and Frigg were seated in the high-seat.93 According to this account, the deities are alive and they are challenged by Helgi, who was a Christian man. Helgi knocks the deities from their platforms down on to the floor.94 This late story is rather fantastic and follows a common pattern in the clerical literature. In my opinion, it cannot be relied upon when discussing pre-Christian ceremonial buildings in general. But the information that the deities are sitting in their high-seats may build on old traditions. As noted above, small Viking Age figures, found at several places in Scandinavia, depict anthropomorphic miniature figures sitting on thrones. The information in the Sagas of Icelanders produces some evidence of the custom of having cultic images in the ceremonial buildings of the chieftains in Iceland. Usually one deity seems to be the major god/goddess of the sanctuary,

89 According to Landnámabók (S85, H73, Ísl. Fornr. 1, 125), Þórólfr’s hof building was dedicated to Þórr (helgaði Þór). It is also mentioned that he had an image of Þórr on his high- seat post(s). 90 Hrafnkels saga freysgoða 2, 6, Ísl. Fornr. 11, 99, 124. 91 See also Víga-Glúms saga, Ísl. Fornr. 9. 92 Í ǫndvegi á hinn óæðra bekk sátu þeir í samsæti Freyr ok Þór. Fljótsdæla saga, 11, 295. 93 Þar sátu þær Frigg ok Freyja. Fljótsdæla saga, 11, 295. 94 . . . steypir þeim af stǫllum ok á golf fram. Fljótsdæla saga, 11, 295. The Relation Between Rulers and Cultic Images 285 which appears to have a particular relationship with the chieftain. In order to get more solid and trustful information we must turn to the archaeological evidence.

7.4.1 The Cultic Image from Eyrarland and other Images Discovered by Archaeologists Archaeological evidence may support the existence of cultic images in pre- Christian Iceland. The best-known object is probably the eleventh-century figurine from Eyrarland at Akureyri in northern Iceland, mentioned above (see fig. 15). This figure is 6.7 cm high and cast in bronze. It depicts a man with a moustache and beard, sitting on a chair and wearing a conical hat. The figure is holding an object between his legs, which is often interpreted as a hammer (or a cross). Many scholars have suggested that this image represents Þórr with his hammer. This interpretation has, however, been debated.95 Richard Perkins has made a very interesting interpretation of the Eyrarland statuette, where he argues that Þórr here is blowing in his beard.96 By means of this magical- religious action he raises wind. The idea of Þórr as a wind-raiser may be found in the text Rǫgnvalds þáttr ok Rauðs, which has been preserved independently, but is also incorporated in Óláfs saga Tryggvasonar en mesta. In this saga Þórr raises wind by “sounding the voice of his beard” (þeyta skeggroddina/ skeg- graustina). When King Óláfr is approaching the sanctuary of Rauðr, with his sailing ships, Þórr performs this ritual act and raises a strong head wind for the king and his companions. The purpose was to prevent the Christian king’s mission in the area. According to Perkins, images of Þórr “sounding the voice of his beard” were common among Viking Age seafarers. Small Viking Age anthropomorphic figures holding and/or blowing in their beards may attest this phenomenon, such as the small image made of walrus ivory found in Lund, an amber statuette from Feddet, Roholte parish, in Zealand, a bronze figure from Chernaia Mogila, Chernigov, Ukraine, and a figurine made of whalebone from Baldursheimur, Suður-Þingeyjarsýsla, Iceland.97 In my opinion, it is possible that the figure from Eyrarland depicts the god in an important ritual role as he is sitting in his high-seat. Perhaps he is using his hammer to consecrate the victim, perhaps he is sounding the voice of his beard. It is possible, however, that this statuette and the other small images were miniatures of large wooden statues of gods. These statues may have been

95 See e.g., Drescher & Hauck 1982, 263ff.; Perkins 2001, 82ff. 96 Perkins 2001. 97 Cf. Lindquist 1963; Perkins 2001, 53ff. 135f. 286 CHAPTER 7 placed in the hall buildings and the hof sanctuaries of Viking Age Iceland as described in the Sagas of Icelanders. Perhaps some of these statuettes were seated in high-seats, as suggested by Fljótsdæla saga and Adam of Bremen. As far as I know, however, no real anthropomorphic wooden images have been found at hall buildings discovered in Iceland. Some abstract symbols of deities have, on the other hand, been discovered in such contexts. A silver pendant, with incised decoration in the form of a possible cross and a Þórr’s hammer, was found at Hofstaðir in Mývatnssveit.98 It was discovered in the hall room (AB) and must be dated to the first phase I (940–980 AD). There is also an iron hook, which may be interpreted as a Þórr’s hammer, found in area A/C at Hofstaðir.99

7.4.2 Images on the Walls of the Hall During the Viking Age, it seems as if mythical motifs were carved on the walls of halls and sanctuaries at the chieftains’ farms in Iceland. This custom may be attested in literary sources. Most famous is the decoration on the walls of the chieftain Óláfr pái Hǫskuldsson’s “fire-hall” (eldhús) at Hjarðarhólt in western Iceland, mentioned in Laxdæla saga.100 Tales were carved there on the panelling of the fire-house (um sǫgur þær, er skrifaðar váru á eldhúsinu). When Óláfr’s daughter Þuríðr married Geirmundr, the well-known skald Úlfr Uggason took part in the wedding feast. During this feast he composed a poem called Húsdrápa (c. 985) about Óláfr Hǫskuldsson and the tales carved on the walls in his fire-house. This drápa is not preserved in the saga, but 11 verses are known from the manuscripts of Snorra Edda, quoted in Skáldskaparmál.101 The subjects in the poem are all mythological. Several stanzas concern the funeral of Baldr. The other topics treated in the surviving verses are Heimdallr’s conflict with Loki and Þórr’s fishing for the Miðgarðsormr. We may thus assume that similar images and carvings referring to mythical events were also made in hof sanctuaries and perhaps also in ordinary dwelling houses in Iceland.102 According to Gísla saga Súrssonar 15, Þorgrímr Þorsteinsson goði arranged a feast in his drinking hall at the end of autumn, in order to celebrate the coming of the winter-nights (vetrnætr). During this feast he also sacrificed to Freyr

98 Batey 2009 (in Lucas 2009), 253f. 99 Batey 2009 (in Lucas 2009), 265. 100 Laxdæla saga 29, Ísl. Fornr. 5, 80. 101 Skj. B1, 128ff. See also Skáldskaparmál, Faulkes 1998, 157. 102 Cf. de Vries 1956–57, §271; E. O. G. Turville-Petre 1964, 248. The Relation Between Rulers and Cultic Images 287

(blóta Frey). It is stated that Þorgrímr decorated the hall with tapestries for this occasion. We cannot rule out that these tapestries had mythical motifs.103 Laxdæla saga is the only text to mention carvings on the walls of the feasting hall. Some doors belonging to the Norwegian stave churches may give an idea of how such carvings of pagan sanctuaries were shaped, for instance, the one found at Hylestad church in Setesdal, Norway. It probably refers to a scene from the traditions of Sigurðr fáfnisbani. Perhaps tapestries with mythic decorations also covered the walls of such buildings as suggested by Gísla saga. The tapestry from the Oseberg ship, for instance, may have occurred in such contexts. Some motifs on it have been related to mythical and ritual themes, such as the hanging scene.104 The Viking Age or early medieval tapestries from Skog in Hälsingland and Överhogdal in Härjedalen could also be related to pre-Christian mythic motifs.105 Austfararvísur mentions that húsbúnaði ‘wall-hangings (?)’ hang in the hall of St Óláfr.

7.5 Cultic Images and Religious Strategies for Rulership: Hypotheses

The existence of cultic images found in the Late Iron Age halls of the three investigation regions suggests that the rulers wanted to remind the cult communities in a concrete way that the political authority had a specific relation to the gods and the mythical world. As noted above, the kings and earls in Svetjud and Trøndelag claimed that they were descended from the gods, while the chieftains in Iceland regarded themselves as “best friends” with a specific deity. The former relationship should, in my opinion, be regarded as genealogical, while the latter relationship rather should be classified as cultic. This close relationship, whether it was genealogical or cultic, was nevertheless emphasized when the authority was seen together with the cultic images on the public scene in the sanctuary. As noted above, some of the images represented the gods as they were seated in the high-seat. When Adam described the statue of Thor (Þórr), for instance, this god was placed in a high-seat (solium) in the “Uppsala temple”, and so too were Þórr, Freyr, Freyja and Frigg in Bersi’s hof sanctuary, according to Fljótsdæla saga. Small statuettes and miniatures discovered in Sweden,

103 Þorgrímr ætlaði at hafa haustboð at vetrnóttum ok fagna vetri ok blóta Frey ok býðr þangat Berki bróður sínum ok Eyjólfi Þórðarsyni ok mǫrgu ǫðru stórmenni. . . . Þá er þeir Þorgrímr bjuggusk um ok skyldu tjalda húsin . . . Gísla saga Súrssonar 15, Ísl. Fornr. 6, 50. 104 See Steinsland 2005, 297. 105 Anjou 1935 and A.-S. Gräslund 2001, 101ff. 288 CHAPTER 7

Denmark and Iceland corroborate that the cultic images of the real sanctuaries really were shaped in that fashion. When the gods were seated in the high-seat they actually appeared in prototypical ritual roles, which had to be imitated by the earthly rulers.106 As noted in chapter 6, Óðinn was endowed with specific abilities when entering his high-seat Hliðskjálf. He could, for instance, see all worlds in the universe. Some of these abilities were necessary for him when exercising his power. When taking part in the ceremonial feast in the hall during the sacrificial feasts, the earthly political leader had to be seated in the high-seat in a similar way to the deity, especially when drinking the ceremonial cups and eating the sacrificial food in the hall or the cultic house (see ch. 4 and 6 above, and ch. 9 below). Perhaps he was also endowed with numinous abilities and cosmic visions when seated there. Most likely the political leader also represented the god in this position. Thus when entering the high-seat he achieved the ultimate prerequisite for making the proper decisions when exercising power. The cultic images may also have played another ideological role during the sacrificial feasts. The ruler was the one who performed some essential parts of the rituals when sacrificing to the gods on the behalf of the people. Sources report, for instance, that he was also involved in drinking ceremonies which were performed in honour of the gods (see ch. 4 above and ch. 9 below). These deities were represented and marked present by the images at the sanctuaries. The images constituted communication channels to the Other World when people needed divine advice or changes of state, for instance, in nature, or in relations between persons and groups in society. In his role as sacrificer, the ruler was the one who communicated with the deities (see ch. 9 below), who were concretely present or represented by the images in the public cultic places. By means of these figures the executor of the sacrifices could create a relationship or communication with the divine world. The rulers thus appeared in a considerable and crucial ritual role in front of the participants during the sacrificial ceremonies when interacting with these images. This public role of the ruler was of course crucial when establishing power structures. It seems as if images of deities and the other mythical objects which surrounded the rulers in the hall buildings or cultic houses in all three areas also had another ideological function. The sanctuary included (at least for some people with cultural knowledge) a cosmic symbolism and reflected the mythical world. This symbolism was even emphasized when the cultic images were also

106 The immaterial prototypical notion that the deity was seated in the high-seat derived, of course, in the first place from the materiality of religion, i.e. the situation when the ruler was placed in his seat. The Relation Between Rulers and Cultic Images 289 placed there on or around the platform or carved on the high-seat posts. When entering the ceremonial building, the king, the earl or the chieftain appeared on this symbolic scene, where his close relationship to the gods was emphasized.107 Very much as Clifford Geertz has described “the theatre state” in nineteenth- century Bali,108 the political leaders of Svetjud, Trøndelag and Iceland were situated in a dramaturgical performance on a stage surrounded with ritual props and objects representing mythical places, objects and beings. To the participants in the halls or the cultic houses it must have seemed that the political leader was in some way part of the divine world, when he appeared on such a scene with references representing the mythical world. This experience was very important for the rulers in these areas.

107 On dramaturgical aspects of the ritual space in ceremonial buildings, see Gunnell 1995; 2001. 108 See Geertz 1980; 1993 (1973), 334ff. See also the example of the Roman imperial cult in Asia Minor, as described by Sr. F. Price 1994. The emperor there was located between human and divine in a way which helped the Greek cities to accept their subjection to the external power of Rome. Chapter 8 Sacred Sites, Ritual Restrictions and Religious Strategies for Rulership

In Old Norse texts religious phenomena are sometimes associated with the adjective heilagr ‘holy’, ‘holy, inviolable’ (cf. helga ‘sanctify, make holy, consecrate’, helgi ‘holiness, sanctity’; see also Proto-Germanic *hailagaz).1 A place which was regarded as heilagr belonged to or was intended for the deities and therefore had to be treated with reverence. At such places certain ritual restrictions and taboos2 were to be observed. Therefore they sometimes were demarcated from the profane space and enclosed with symbolic walls or bonds. The rules of the sanctuaries were also important for the ruler strategies. The sacred site, which was controlled or even owned by the ruler, could not be violated. It was the ruler’s assignment to sanctify, protect, restore and perhaps also to ritually end these places. In skaldic poems rulers are therefore called vǫrðr véstalls “the protector of the sacred stand” or vés valdr “the ruler of the sacred place”. If these customs or regulations of the sacred site were disturbed, not only the sanctity of the place was offended, but also the power, values, and the “world” of ruler was violated. It was therefore of the utmost importance for a ruler or a chieftain to protect and perhaps periodically re-establish the sanctity of the ceremonial buildings and the sanctuaries. This type of performative and

1 ON heilagr is translated ‘heilag, heilag, ukrenkjeleg’ in Heggstad et al. 1993, 176. The basic meaning of this adjective is actually disputed. Most likely heilagr contains a sacred compo- nent of meaning, but its semantics do not exactly correspond with the analytic concept of holy which has been developed in Liberal Protestant discourses by e.g. Nathan Söderblom (1913; 1914) and Rudolf Otto (1917). On the problems with the meaning of ON heilagr, see e.g., Baetke 1942; Vikstrand 2001, 226ff.; Nordberg 2010; 2011. See also the verb helga ‘helge, gjere heilag, vie, gjere ukrenkjeleg, fredlyse’ and the noun helgi ‘heilagskap, det å vere heilg el. ukrenkjeleg’, Heggstad et al. 1993, 180. 2 The concept of “taboo” refers to ritual prohibitions or restrictions associated with phenomena considered sacred and/or dangerous. Places, objects, persons, times, actions and words may be permanently or temporary circumscribed by such rules. The term taboo derives from Polynesian tapu (Tongan tabu; Hawaiian kapu) ‘marked thoroughly’, ‘that which is forbidden’. Cf. Marett 1909; Söderblom 1913; 1914; Lehmann 1930; Radcliffe-Brown 1979 (1939); Steiner 1956; Douglas 1966; van Baaren 1972; Widengren 1969; Wagner 1987; Schmidt 1998. On the concept of “taboo” in Germanic contexts, see Sundqvist 2005d.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_009 Sacred Sites, Ritual Restrictions and Religious Strategies 291 ritual strategies may be seen in all three investigation areas, and on different social levels, i.e. among kings, earls and chieftains.

8.1 Sanctifications of Places and Ritual Restrictions (Taboos)

In Old Norse the act of sanctifying land for oneself is called at helga sér land.3 By this act the claimed land gained sanctity and was granted by divine authority. It seems as if it also was possible to sanctify a place or a building and then dedicate it to a certain deity (see below).4 The purpose of these performative rituals was primarily to claim land rights, but also (in a religious sense) to separate the sacred area from the profane sphere. The sanctified land or place had to be freed from evil powers and dedicated to the deity. Some places in the new land also had to be sites or areas where men could encounter the mythical world during cult. Therefore certain ritual restrictions and taboos had to be observed there.5 These performative rituals executed by the chieftain in public were also intended to show that from now on it was his rules/restrictions the group or community must follow.

8.1.1 In Trøndelag and Iceland Acts to sanctify lands and places are well preserved in the medieval written sources referring to pre-Christian Iceland.6 As noted above, Landnámabók reports that the chieftain and hofgoði Þórhaddr from Mære took down his hof and brought all the soil of the sanctuary and the high-seat posts to Iceland. When he came to Stǫðvarfjǫrðr in Iceland he proclaimed the Mærina-helgi ‘Mære sanctity’ over the entire fjord area and stated that nothing was allowed to be killed there, except the animals on the farm (see the text in ch. 5 above).

3 Strömbäck 1928b; Baetke 1942, 106ff., 153f.; Clunies Ross 1998, 149ff. 4 Consecration and dedication rituals related to sacred places are known from many cultures. See Widengren 1969; Eliade 1974 (1949); 1987 (1957); 1991 (1949); Smith 1987. The medieval Church likewise had rites for dedicating churches. Descriptions of such rituals exist in, for instance, the Romano-Germanic Pontifical called the “Mainz Pontifical” (PRG 33 and PRG 40), composed c. 950–962 in Mainz. A ritual close to this Roman-Germanic rite of dedication of churches was probably also used in Scandinavia during the Early Middle Ages. See e.g. P. Ström 1997. 5 According to Mircea Eliade (1987 (1957), 32), the consecration of a territory or a place should be seen as a repetition of the cosmogony; “to organize a space is to repeat the paradigmatic work of the god.” Possibly the renovations of sanctuaries in Viking Age Scandinavia could be seen as repetition of cosmogony. See below. 6 See e.g. Strömbäck 1928b; Clunies Ross 1998, 122–157. 292 Chapter 8

These ritual limitations associated with the concept of Mærina-helgi must originally have been related to the sanctuary in Þórhaddr’s home land, i.e. the hof sanctuary at Mære in Trøndelag, and perhaps the sacred meadows situated close to it, called Hælgjin, ‘the holy meadow’ or rather ‘the holy meadows’.7 The information that Þórhaddr brought the high-seat posts and “the soil of the sanctuary” indicates that he also consecrated a new ceremonial building when he arrived in Iceland, even if it is not explicitly stated in the text. It is interesting to note that it was the chieftain Þórhaddr himself who consecrated the sacred space and proclaimed its extent when settling in Iceland. He also decided the rules that prevailed at such places. Most likely these regulations were based on ancient traditions and customs deriving from Mære. Similar consecration rituals and regulations related to sacred lands and places are reported in Eyrbyggja saga 4. According to this saga, the Þórr worshipper Þórólfr Mostrarskegg emigrated from the island of Mostr, Sunnhǫrðaland, in Norway, to the isthmus of Þórsnes in western Iceland, where he threw overboard the high-seat posts with “Þórr images” from his old sanctuary and declared that he would settle at any place in Iceland where Þórr chose to send the posts ashore. The posts were found at Þórsnes. Then Þórólfr carried fire round the land he claimed. He settled at Hofsvágr and erected a hof on his farm called Hofstaðir. Just inside the door of the hof the high-seat posts were placed, and beyond that point the whole building was considered a sanctuary (var allt friðarstaðr). Not far from this farm on the headland, there was a mountain which Þórólfr regarded as sacred. Therefore, no one was allowed even to look at it without first having washed himself. No living creature on this mountain, neither man nor beast, was to be harmed until it left of its own accord. Þórólfr called this mountain Helgafell ‘the holy mountain’ and he believed that he and his kinsmen would enter it when they died. He held his courts on the point of the headland, and there he started the district assembly. This place was so holy (svá mikill helgistaðr) that he would let no one desecrate it, either with bloodshed or with excrement.8 We may assume that Þórólfr had applied similar consecration rituals, restrictions and taboos to the sanctuary, assembly site and sacred landscape of Mostr, his old home, before his emigration to Iceland. The customs he applied thus derived from his Norwegian homeland. According to Eyrbyggja saga, it is obvious that Þórólfr himself was the one who cared for the sacred space and also ensured that the restrictions connected to this space were followed.

7 Dillmann 1997, 71, note 43. Critically considered by Sandnes 1987, 151. 8 In Landnámabók too, the ritual restrictions connected to the mountain Helgafell are mentioned. Landnámabók S 85, H 73, M 25, Ísl. Fornr. 1, 124f. Sacred Sites, Ritual Restrictions and Religious Strategies 293

Most of these rituals and restrictions are also mentioned in the narrative about Þórólfr preserved in Landnámabók. This text explicitly states that Þórólfr built and dedicated the new hof to Þórr (gerði þar hof mikit ok helgaði Þór).9 By performing these rituals the local group was informed that they must follow Þórólfr’s rules. In Eyrbyggja saga (but not in Landnámabók) it is recorded that Þórólfr carried fire round the land he claimed. This ritual was probably also performed in order to sanctify his new homeland. It is possible that this land was related in one or another sense to the hof building. This may be supported by other Old Norse texts. Landnámabók, for instance, tells of the goði and new set- tler called Jǫrundr, who erected a hof at Svertingsstaðir, to the west of Fljót in north-western Iceland. “There was an unclaimed piece of land east of Fljót, between Krossár and Jǫldusteinn, and Jǫrund carried fire around it and dedicated it to the hof sanctuary.”10 Perhaps other rituals were performed in such contexts as well, such as swidden agriculture, the erection of wooden poles to stake one’s claim, shooting an arrow over the land, ritually depositing symbolic and highly personal objects at the boundaries of the land etc.11 Margaret Clunies Ross argues that some of these rituals were associated with various gods in the Norse pantheon: “Óðinn is often represented as dedicating enemy armies to himself by shooting an arrow over them, while Þórr is associated with the protection of the household and family property through high-seat posts and other stave-like objects in various Old Norse myths and sagas.”12 It should be noted that the claimed lands and places were often dedicated to specific deities. In Landnámabók we read, for instance, that Þórólfr promised

9 Landnámabók M25, Ísl. Fornr. 1, 125. 10 Bjórr lá ónuminn fyrir austan Fljót milli Krossár ok Jǫldusteins; þat land fór Jǫrundr eldi ok lagði til hofs. Landnámabók S 346, M 10, Ísl. Fornr. 1, 350f. Trans. Hermann Pálsson & P. Edwards. 11 For evidence of these customs in Old Norse texts, see e.g. Clunies Ross 1998, 151 and the ritual of depositing objects at boundaries of the new land, see T. Zachrisson 1998, 197ff. This ritual is attested in Svarfdæla saga 11: Þorsteinn svǫrfuðr fór nú millum fjalls ok fjǫru ok lagði undir sik allan dalinn ǫðrum megin ár. Hann ferr til fjalls ok gerir þar kennimark, sem hann kom framast, ok braut þar í sundr kamb sinn ok kastaði niðr kambsbrotunum ok lætr eptir silfr í þremr stǫðum, hálfa mǫrk í hverjum stað, ok er sá rimi kallaðr at Kambi; ok nefnir Þorsteinn sér vátta, ok fyllir hann með því dalinn sér til vistar ok gaf af sér nafn ok kallaði Svarfaðardal. Ísl. Fornr. 9, 151f. This evidence is not mentioned by Strömbäck. 12 Clunies Ross 1998, 151f. 294 Chapter 8 to dedicate his entire land claim to Þórr (at helga Þór allt landnám sitt) and call it after him.13 The information that the chieftains proclaimed ritual restrictions at sanctified land and places is probably based on old traditions.14 In the Old Norse sources, ritual restrictions and taboos are often also related to shrines called vé ‘sacred place, sanctuary’.15 To kill a man in a sacred space (víg í véum) was strongly forbidden and punished with curses and outlawry. Some of these regulations refer to Norwegian contexts. Landnámabók mentions that one of King Haraldr’s escorts, Vali enn sterki, killed on sanctified ground and therefore was outlawed (hann vá víg í véum ok varð útlægr).16 A person who killed on sanctified ground was sometimes metaphorically described as a “wolf”.17 In Egils saga 49, Queen Gunnhildr’s brother, Eyvindr, was considered a “wolf” after killing at a vé sanctuary: “Because Eyvindr had committed murder at a sacred place he was declared a defiler [actually a wolf (vargr)] and had to go into outlawry at once.”18 This episode took place at a hof sanctuary at Gaular, north of Sogn, in Norway. King Eiríkr also participated in the sacrificial feast there. During the night a quarrel broke out that ended in assault. Eyvindr stabbed Þorvaldr with his short-sword. The other men leaped to their feet “but none

13 Landnámabók S 85, H 73, Ísl. Fornr. 1, 124. See also Landnámabók M 8, Ísl. Fornr. 1, 344f.: Ásbjǫrn helgaði landnám sitt Þór ok kallaði Þórsmǫrk. 14 Anglo-Saxon and Frankish sources, as well as texts from the Carolingian Period, indicate that sacred places and objects in Continental Europe and England were connected to certain restrictions during the Early Middle Ages. Lex Frisionum (800 AD), for instance, reports how persons guilty of sacrilege against cultic sites and sacred objects were treated by the pagan community. If someone destroyed a sanctuary and carried away the sacred objects, he was taken to the shore of the sea which would be covered by the flood. There he had his ears cut off and was sacrificed to the gods of the temple he had desecrated (Additio XI, De honere templorum). In Section 5,1 it is stated that the person who violated the temple (qui fanum effregerit) could be killed without wergild. See Clemen, Fontes, 1928, 59. Even if these passages appear in a Christian law they have been regarded as authentic and old. They obviously have a background in pre-Christian traditions. Cf. Nilsson 1992; Maier 2003, 117. For more examples, see Sundqvist 2005d. 15 Cf. PG *wīhaz; Old Saxon wīh-dag ‘holy day’, Old English wīg-bed ‘altar’ (literary ‘holy bed’), or German Weihnacht ‘Christmas’ (literally ‘holy night’), and Goth weihen ‘to consecrate’ (ON vígja ‘consecrate, hallow, bless’). See Baetke 1942; de Vries 1956–57, §§240–244 (critically); Heggstad et al. 1993, 498; Vikstrand 2001, 298; Sundqvist 2005d, 251f. On the word vé/vi in place-names, see e.g. T. Andersson 1992a; 1992b; 1992c; Vikstrand 2001. 16 Landnámabók. S 72, H 60, Ísl. Fornr. 1, 102f. See also S 376, H 331, Ísl. Fornr. 1, 374. 17 F. Ström 1942, 73ff., 128ff.; Näsström 2001a, 94; 2002, 138. 18 Eyvindr hafði vegit í véum, ok var hann vargr orðinn, ok varð hann þegar brott at fara. Ísl. Fornr. 2, 125. Sacred Sites, Ritual Restrictions and Religious Strategies 295 of them was armed inside [the building], because hof sanctity (hofshelgi) prevailed there”.19 It thus seems as if the area inside the hof was also hedged with taboo rules. It is interesting to note that this holy space in this text is designated vé ‘sacred place’. Most likely it was a certain type of killing that was forbidden on the sanctified grounds. The story of Þórhaddr (mentioned above) says that nothing was allowed to be killed in the area where Mærina-helgi ruled, except the animals on the farm. Most likely it was also permitted to kill consecrated sacrificial animals there. Manslaughter, on the other hand, was forbidden. After the fight between the Þórsnesingar and the Kjallakssynir at the sacred field of Þórsnes, the sanctified grounds were considered to be defiled by the spilling of blood in enmity (þá var vǫllrinn óheilagr af heiptarblóði), according to Landnámabók. It was then decided that the assembly should be moved to the eastern part of the headland. This became a very sacred place too.20 The first element of the compound heiptarblóð, heipt refers to ‘fiendskap, hat, ovund, misunning’.21 Thus this pollution of the sacred place consisted of the spilling of blood in enmity, a phenomenon caused by hatred and violence between men.

8.1.2 Consecration of the High-Seat As noted above, Þórhaddr and Þórólfr brought the high-seat posts from the old sanctuaries to Iceland. Eyrbyggja saga indicates that these posts were erected in the new hof sanctuary at Hofstaðir. In these posts certain sacred objects were placed, called reginnaglar ‘nails of the gods’, indicating their religious significance. It is also stated that images of Þórr were carved on them. The peculiar story about Ingimundr the old, transmitted in Vatnsdæla saga and Landnámabók, indicates that sacred objects also were deposited in the holes of the high-seat posts when hof sanctuaries were erected, perhaps as an act of dedication (see ch. 6 above). Archaeology indicates the same.22 As noted in chapters 5 and 6, the gold foil figures and other symbolic objects from the Merovingian Period and the Early Viking Age were sometimes deposited in the post-holes or other places at ceremonial halls and cultic houses. Sometimes

19 . . . en menn váru allir vápnlausir inni, því at þar var hofshelgi . . . Egils saga, Ísl. Fornr. 2, 125. Cf. Vatnsdæla saga 17, Ísl. Fornr. 8, 48: Eigi er þat siðr at bera vápn í hofit, ok muntu verða fyrir goða reiði, ok er slíkt ófœrt, nema bœtr komi fram; Landnámabók (S 179, Ísl. Fornr. 1, 220): Hrafn . . . hafði sverð gott; þat bar hann í hof; því tók Ingimundr af honum sverðit. 20 Þá var þat ráð tekit at fœra brutt þaðan þingit ok inn í nesit, þar sem nú er; var þar þá helgis- taðr mikill . . . Landnámabók S 85, H 73, Ísl. Fornr. 1, 126. 21 Heggstad et al. 1993. 22 See particularly Carlie 2004. 296 Chapter 8 they were even placed in the holes of the assumed high-seat posts. This may be attested, for instance, at Borg in Lofoten. Two of the foils found there had obviously been placed there intentionally. They lay under a stone, deep down in the last phase of the post-hole. Gerd Stamsø Munch interpreted this thus: “It is suggested that the plaques were deposited by the post at dedication of the room for cult activities at Borg, a ceremony so important that it was necessary that the plaques were still present when the post was subsequently replaced.”23 Similar finds have been made around the central post-holes at the ceremonial houses of Mære, and at Hofstaðir (Mývatnssveit) (ch. 6 above and 13 below). Perhaps these objects should be seen as traits of a consecration of the sanctuary and particularly the high-seat, which was regarded as the ritual centre of such a building. Similar rituals may have been performed in the Mälaren area. Gold foils found in the post-holes of the hall of Helgö may attest to such rituals (see ch. 11 below). Weapon deposits at the aristocratic hall of Birka may be evidence of the same.24 Other ritual objects discovered in the post-holes of the hall of Lunda may have played a similar ritual role (ch. 6 above).

8.1.3 Consecration of Sacred Space in Svetjud It is possible that the sanctification of territory and sacred places mentioned in the Old Norse literature about Iceland also occurred in eastern Scandinavia.25 The Guta saga, for instance, begins like this:

Gotland was first discovered by a man named Þieluar. At that time the island was so bewitched that it sank by day and rose up at night. That man, however, was the first that brought fire to the island, and afterwards it never sank again.26

23 Munch 2003b, 261. 24 Holmquist Olausson & Kitzler Åhfeldt 2002; Carlie 2004. 25 See e.g. Strömbäck 1928b. It is possible that archaeological evidence from eastern Scandinavia may indicate that the act of sanctifying land for oneself (at helga sér land) also existed there. Torun Zachrisson (1998, 194–200 and passim) has convincingly argued that the ritual of depositing symbolic objects (sometimes made of gold or silver) was used when claiming and consecrating land for oneself. These types of ritual are archaeologically attested in Uppland and Gästrikland in Sweden. They are also mentioned in Old Norse sagas. Cf. Hedeager 2011, 172ff. 26 Gutland hitti fyrsti maþr þann, sum Þieluar hit. Þa var Gutland so eluist, et þet dagum sank ok natum var uppi. En þann maþr quam fyrsti eldi a land, ok siþan sank þet aldri. (Ed) Peel 1999, 2f. Sacred Sites, Ritual Restrictions and Religious Strategies 297

It seems here as if the chaotic land was cosmicized (speaking in Eliade’s terms) by means of a “fire ritual”, which perhaps also is mentioned in Eyrbyggja saga (see above). This ritual act was performed by Þieluar, the mythical founder of Gotland.27 Guta saga also indicates that ritual restrictions were observed at sacred sites in Gotland. The text mentions that a sacrifice was performed at Vi (perhaps a sacred place situated close to present-day Visby). Later, however, Botair built a church there. The people of the island wanted to burn it down. Then Botair went up on the top of the church and said: “If you want to burn it, you will have to burn me along with this church.” Botair was supported by Likkair from Stenkyrka. He said: “Do not persist in burning the man or his church, since it stands at Vi below the cliffs.” As a result of this, the church was allowed to stand unburnt. The name Vi probably refers to a pre-Christian sacred place designated with the Old Gutnish word vi (cf. OSw vi, ON vé), i.e. a place where no violence was allowed. The idea that sacred sites (ON vé, OSw vi) in eastern Scandinavia were circumscribed with certain regulations and rules may also be attested in older sources. According to the Oklunda runic inscription (N288) (ninth century), in north-eastern Östergötland, protection against violence and asylum could be found at one such vi site:

kunar : faþirunaRþisaR : insa flausakaR : sutiuiþita . . . uifin : þittafaþi

Gunnarr faði runaR þessaR. En sa flo sakR. Sotti vi þetta . . . Vi-Finnr þetta faði.

Gunnarr cut these runes. And he fled from penalty. Sought this sanctuary (vi) . . . Vi-Finn carved this (?).28

This inscription has been much debated.29 Scholars have often used the Oklunda inscription combined with Swedish provincial laws in a retrospective

27 On the possible hallowing with fire ritual in Guta saga and the problems related to such interpretation, see Peel 1999, xx. 28 S.B.F. Jansson 1987, 37; von Friesen 1933, 152. 29 Cf. e.g. von Friesen 1933, 152f.; Salberger 1980; Ruthström 1988; Nilsson 1991; Peterson 1993; Lönnqvist & Widmark 1997; T. Zachrisson 1998, 47f.; Vikstrand 2001, 323ff.; Gustavson 2003. 298 Chapter 8 way to support the idea of pre-Christian asylum at sanctuaries.30 This method has been questioned.31 Bertil Nilsson, a specialist in canon law, does not rule out the possibility that the Oklunda inscription may have been influenced by Christian ideas.32 Asylum and the concept of peace existed very early as a Christian institution and they also appeared in Decretum Gratiani of 1140.33 In the early medieval Swedish laws, churches are often conceived of as asylum sites.34 On the other hand, the context of the inscription tells that Gunnarr took refuge at a pre-Christian vi “sanctuary”. This vi was probably a sacred grove, if we rely on the place-name, Oklunda. Also close to this place is Lundby. Both place- names refer to the holy site mentioned in the inscription, i.e. the vi or lund in Vikbolandet.35 In addition the name Vi-Finn (RSw Vī-Finn) may indicate a pre-Christian cultic context.36 Maybe it was he, the ruler of the sanctuary, who ensured Gunnarr’s protection there. Uncertain medieval sources suggest that rulers and/or cult leaders protected those who sought peace at sanctuaries in Svetjud. The Þáttr af Ǫgmundi dytts ok Gunnari helminga, in Flateyjarbók, says that Gunnarr took refuge in a sanctuary (hofstaðr, goðahús) in Svetjud.37 There Freyr’s wife ruled, i.e. perhaps a female cultic leader who supervised the worship of the Vanir god. She protected Gunnarr there.

8.2 The Enclosed Sacred Space

The sources thus report that the sanctuaries in Svetjud, Iceland, Trøndelag and elsewhere in Norway had ritual spaces which were considered as protected

30 Cf. Frense 1982, 92; Ruthström 1988, 68f., 73f.; Lönnqvist & Widmark 1997, 154f. 31 Cf. Sjöholm 1988, 18–32, 50f.; Nilsson 1989, 24–27; 1991, 487. 32 “Slutligen måste ändå frågan ställas, hur pass förkristen ristningen är till sitt innehåll, och om inte det förfarande som den omvittnar till dels åtminstone kan ha influerats av konti- nentalt och/kyrkligt rättstänknde.” Nilsson 1991, 488. 33 Nilsson 1991, 474–484. 34 According to Gutalagen, Fardhem, Tingstäde and Atlingbo were considered as asylum churches. Similar regulations appear in the Östgötalagen (D VII:1, E III:2). See also Gutalagen, Af mandrapi. (Ed) Pipping, 15–18: þa riþi þengat sum hann bandu will hafa dragit. Cf. Ruthström 1988; Holmbäck & Wessén 1979 (1933–46), Vol. 4, 212f., 255f. 35 Cf. Franzén 1982, 113; Vikstrand 2001, 325. 36 Since uifin lacked the mark of the nominative it has been argued that it must be considered as an accusative object. See e.g. Salberger 1980; Ruthström 1988, 74. Peterson (1993, 33–40) argued in favour of the traditional reading and interpretation of the sequence uifin þitta faþi, i.e. with uifin (RSw Vī-Finn) as a nominative form. 37 Flateyjarbók I, 337. See also Ǫgmundar þáttr dytts. Ísl. Fornr. 9, 111ff. Sacred Sites, Ritual Restrictions and Religious Strategies 299 ground. These areas most likely included the ceremonial buildings, but also ritual objects, structures and places located outdoors beside these buildings. It seems as if such areas sometimes were enclosed in certain ways. The term ON hofsgarðr ‘sanctuary-yard’ may indicate how such ritually restricted yards were conceived. The room inside the ceremonial building and the sanctuary- yard were thus regarded as sanctified and protected space and they could not be violated. At such areas the hofshelgi prevailed. Most likely sacred meadows could also be enclosed in some way during the religious and judicial gatherings. There are some traditions from Trøndelag related to judicial customs, which also may be valid for attitudes related to sanctuaries.

8.2.1 The Sacred Bonds of the Frostathing In Den ældre Frostathings-Lov, used at Frosta in Trøndelag, we meet the term vébǫnd ‘sacred bonds’.38 This law reports that such bonds, cords or ropes were used in connection with the place (originally at Logtu) where the political- judicial assembly met during the Middle Ages:

It is old law that a bailiff from each of the various shires shall set up enclosures [actually “sacred bonds”] here at the thingstead (scolo gera vébönd her á þingvelli). And the enclosure shall have sufficient space so that all who are appointed to the law court shall have ample room to sit within it.39

The enclosures were intended to demarcate the area the so called ‘law court’ or ‘lay assessors’ were supposed to sit within. In more ancient periods this space was probably conceived as a vé ‘sacred place’. The first element (vé) in the term vébǫnd indicates this. In the ritual space inside the enclosures, the þinghelgi ‘assembly sanctity’ or friðr ‘peace’ (cf. also grið40) thus prevailed. The “sanctity” or “peace” of the assembly place (vé) thus could not be violated. It seems as if this space was quite extensive, at least during the Early Middle Ages. Den ældre Frostathings-Lov rules that forty men shall be named from each “shire” ( fylki) of “Inner Trondheim” and sixty men from each “shire” of “Outer Trondheim”. All these “lay assessors” were supposed to

38 Heggstad et al. (1993) explains it as “band som gjerder inn ein fredheilag stad, særleg om inngjerdinga omkring lagretta på tinget”. Most likely sanctuaries had some type of enclosures, equivalent to the phenomenon designated by the ON term vébǫnd. 39 Den ældre Frostathings-Lov, NGL 1, 127. Trans. Larson. 40 Cleasby & Vigfusson (1957 (1874), 215) explains grið as thus: ‘in pl., metaph. a truce, peace, pardon; friðr is the general word, grið the special, deriving its name from being limited in time or space (asylum)’. On þinghelgi, see de Vries 1956-57, §§ 148 and 242. 300 Chapter 8 fit in this sanctified space within the vébǫnd. The assembly that gathered on this ground could thus be interpreted as a social microcosmos, i.e. there were men representing each “shire” ( fylki) of both Inn-Trøndelag and Ut-Trøndelag. Together they symbolically represented the community which stood under the Frostuþingslǫg or Þrǫndalǫg, i.e. ‘the law of Þrœndir’.41 There are other sources describing the nature of “court circles” in Norway during the transition period. In Egils saga, for instance, there is a detailed description of one at Gulathing, in Sogn, south-western Norway:

There was a flat plain where the court was established, and hazel staffs were set in a circle, with a cord around them. This was known as “sacred bonds” (vébǫnd). Within the circle sat the court, twelve men from the Firðafylki, twelve from Sygnafylki and twelve from Hǫrðafylki.42

The spatial structure of the court circle (the “vé area”; cf. the term dómhringr) was situated on smooth or levelled (sléttr) ground. It thus seems as if this place was prepared by human hand, i.e. people had worked to remove all irreg- ularities and impediments that may have been naturally present there. On this ground twelve men representing each “shire” ( fylki) gathered. Together they symbolically represented the community which was related to the Gulathing.43 In order to separate the sacred space of the court from the profane sphere, staffs made of hazel were used at the Gulathing. This wood had strongly ritual and apotropaic associations in Norway as well as among other Germanic people.44 On these staffs the “sacred bonds” (vébǫnd) were connected. We hear from other Germanic peoples that the place of the judge was marked off in a similar fashion. The Gothic term staua ‘staff’, for instance, denotes ‘court’ and ‘judge’. The Old High German poem Muspilli (c. 870) describes the court (mahal-stat) as kimarchôt ‘enclosed’,45 and in the Frankish law text Lex Ribuaria (c. 630) we read that an oath must be sworn “vor des Königs Gerichtshügel

41 The assembly place of Frostathing was located at Logtu, on the Frosta peninsula in the central part of the Trondheimsfjord. Røskaft 2003, 110. 42 En þar er dómrinn var settr, var vǫllr sléttr ok settar niðr heslistengr í vǫllinn í hring, en lǫgð um útan snœri umhverfis; váru þat kǫlluð vébǫnd; en fyrir innan í hringinum sátu dómendr, tólf ór Firðafylki ok tólf ór Sygnafylki, tólf ór Hǫrðafylki . . . Egils saga, Ísl. Fornr. 2, 154. Trans. Scudder. 43 Lincoln 1994, 62. 44 Lincoln 1994, 62. 45 De Vries 1956–57, § 265. Sacred Sites, Ritual Restrictions and Religious Strategies 301 im Kreise und im Haselnußgehege, das ist innerhal der Zweige”.46 The sacred place within the enclosures could perhaps be associated with the term vangr ‘(sacred) plain, field’, which sometimes was related to the hof sanctuaries in Trøndelag and Norway. Perhaps the place-name Fåvang (Fǫðvangr) in Ringebu, Gudbrandsdalen, indicates this. It has been suggested that the first element *fǫð- may be related to Gothic faþa ‘to make boundaries’. Fǫðvangr may thus be interpreted as ‘the enclosed plain’.47 Other suggestions have been made too.48

8.2.2 The Enclosed Sacred Area in Uppsala and Elsewhere Perhaps something equivalent to the enclosed sacred area constituted by the “sacred bonds” may also be seen at assembly sites in eastern and southern Scandinavia. The entire sanctuary of Old Uppsala may have been enclosed or demarcated by means of erected wooden posts. They appear in a straight line at regular intervals roughly 400 metres to the south of the sanctuary in an east–west direction, but also in a straight line approximately 900 metres to the north of Old Uppsala in a south–north direction. The ritual area of Uppsala, or temenos,49 could be identical with the delubrum mentioned in Adam’s text, scholion 139. The post aligments could, on the other hand, also be markers of a procession road.50 At the outdoor cultic place (the bare rock) of Helgö, beside the ceremonial building, remains of a wooden palisade were discovered running along the slope. This was called VII by Wilhelm Holmqvist,51 and consisted of a line of stones including post-holes. Several finds were made at this structure, such as a piece of “ring-gold”, a decoration of a belt, sixteen amulet rings etc. Most likely this palisade enclosed the ritual space at the rock and perhaps the amulet rings were placed on the enclosure poles.52 Some type of fence also enclosed the

46 Si quis pro hereditate vel pro ingenuitate certare coeperit post malo ordine, cum 6 in ecclesia coniuret, cum 12 ad stappulum regis in circulo et in hasla, hoc est in ramo, cum verborum contemplatione coniurare studeat. Lex Ribuari 69.5. Text and trans. (ed.) Eckhardt, 186f. See also Clemen, Fontes, 1928, 31. 47 Olsen 1915, 137ff. 48 See e.g. Sandnes & Stemshaug 2007, 163. 49 The term temenos actually refers to the ancient Greek temple areas. The word τέμενος is translated in Liddell and Scott (1888) as ‘a piece of land cut off’ . . . ‘a piece of land dedicated to a god, the sacred precincts’. 50 Beronius Jörpeland et al. 2013; Sundqvist 2013, 89f. See also www.ArkeologiGamla Uppsala.se, last accessed 1 May 2014, and oral information from Jonas Wikborg. On processions in general, see Hultgård 2001a. 51 See T. Zachrisson 2010, 84. 52 T. Zachrisson 2004b, 151f., 162; 2010, 84. 302 Chapter 8 cultic site of Lilla Ullevi in Uppland. Post-holes, cooking-pits and hearths indicate this.53 At the hall of Husby, Glanshammar, dated to the Migration and Merovingian Periods, there was similarly a demarcation or palisade.54 A gold foil figure was discovered on this site. Together with the wall it supports the idea that this hall and place may have had a religious function. The Iron Age hall at Hovde, Ørland, in Sør-Trøndelag, was enclosed by ditches and fences.55 During the second phase there was a fence consisting of a row of parallel posts fencing the house. Several fireplaces and cooking pits were also found associated with this area. Perhaps the fence here also enclosed a ritual area. In southern Scandinavia we likewise find ritual enclosures around cultic buildings and sacred sites. At Tissø, Zealand, a large ceremonial hall was erected in the Late Iron Age.56 Palisaded enclosures were found at the south-western part of the hall; they enclosed a small structure that was interpreted as a cultic house. A similar pattern is also in evidence at the Late Iron Age farm of Järrestad, Scania, with the hall, an enclosed area with a small cultic building, and beside this hall.57

8.2.3 When was the Ritual Space Restricted by Rules? The question when the holy landscapes, enclosed areas, the cultic houses and the hall buildings were regarded as sacred is crucial. In my opinion it seems logical that the ritual restrictions inside a cultic house and an enclosed area were permanent. At such a ritual space certain restrictions must always be observed. But perhaps the feeling of holiness was emphasized there during the calendrical feasts (see ch. 9 below), and when certain rites of passage were celebrated. When it comes to the multifunctional halls, Terry Gunnell has offered a reasonable answer to this problem.58 According to him, these buildings were temporarily transformed from their everyday character into sacred buildings, especially during religious feasts. When the nomenclature hof is used in the Icelandic texts it is actually the temporarily sacred shape or status of the hall that the saga authors are describing: “the central hof was seen by people simul- taneously as being both farm, and a symbolic pagan microcosmos (somewhat like a theatre stage which has the gift of being two places at once in the minds

53 Bäck et al. 2008; Hållans Stenholm 2010, 51. 54 It should be noted that several halls (“house II” and “house III”) may have replaced each other. Ekman 2000: 25. 55 Grönnesby 1999; 2000. 56 Jørgensen 2002. 57 Söderberg 2003, 293ff.; 2005, 211ff.; 238, 252ff. 58 Gunnell 2001. Sacred Sites, Ritual Restrictions and Religious Strategies 303 of the audience, at least during the duration of performance).”59 In my opinion, this reasoning does not exclude that certain areas or objects in the hall, such as the high-seat, the posts and certain fireplaces, could have had a more permanent sacred status as the space of cultic houses and the ritual enclosed areas. At these places ritual restrictions always prevailed. Thus enclosed areas, the cultic houses and the hall buildings were all protected places. On the other hand, it is obvious that certain ritual restrictions and taboos also were related to time.60 Most likely the communal thing-places were regarded as sacred only as long as the grið ‘peace’ or þinghelgi ‘assembly-sanctity’ prevailed, and as long as the vébǫnd were set up around the assembly. The adjective heilagr is sometimes also used in a temporal sense in connection with certain religious and calendrical feasts. In Sigvatr Þórðarson’s Austrfararvísur (st. 4–7) the Christian poet visited a farm in western Svetjud while they were celebrating a pagan feast called álfablót there, probably in a hof building (see ch. 9 below). The pagan people sent the poet and his fellows away. They declared it was holy (days) (heilagr) there. At the doorway of the building they were met by a female (húsfreyja): “No farther can you enter, you wretch,” she said, “we are heathen people, and I fear the wrath of Óðinn (hræðumk ek við Óðins reiði).”61 Most likely the Christian poet Sigvatr and his friends were sent away because of the ritual restrictions that prevailed during the holy time of álfablót, and perhaps also because these Christian foreigners were not allowed to touch the holy space inside the hof.62

8.3 Violations against the Sanctity of the Sacred Space: The Example of Egils saga

It thus seems as if the enclosed and the sacred places in the investigation areas were protected with “peace” ( friðr or grið), which could not be violated. The “peace” at the Frostathing was extended temporarily to a wider space than only the assembly place during the thing-days. Violation of this “peace” was strongly forbidden. Den ældre Frostathings-Lov reports thus:

59 Gunnell 2001, 18. 60 See Sundqvist 2005d, 252f. 61 Skj. B1, 221f. 62 Cf. F. Ström 1985, 89; Maier 2003, 115. For examples of taboos related to time, see Sundqvist 2005d, 253. 304 Chapter 8

Now it shall be told what peace those men shall enjoy who are chosen to attend the Frostathing. Those who are appointed [to attend] and all others who go to attend the thing shall be in each other’s peace (scal í griðum við annan) while on the journey to the Frostathing and until they have returned to their homes. And if any one shall violate this peace (gengir á þau grið), he has forfeited his land, his chattels, and his right to peace among men (oc friði sínum við alla men); and let him never return to the kingdom.63

Violation of the grið or friðr ‘peace’ (cf. þinghelgi ‘assembly-sanctity’) was thus punished with outlawry and considered a capital crime in medieval Trøndelag and elsewhere in Norway. There is a passage in Egils saga 56 where King Eiríkr and Queen Gunnhildr committed this crime. The saga reports that Egill travelled from his home at Borg in Iceland to the Gulathing in Norway (c. 934), in order to collect his legal share of the inheritance after Bjǫrn hǫlðr Brynjólfsson, his wife Ásgerðr’s father. He had received reports that his brother-in-law Berg- Ǫnundr, with support from King Eiríkr and Queen Gunnhildr, had seized all the lands that Bjǫrn had owned. Berg-Ǫnundr was married to Bjǫrn’s other daughter, Gunnhildr Bjarnadóttir. This journey was not without danger, since Egill had previously been condemned to outlawry by King Eiríkr. During the judicial process at Gula, Egill and his brother-in-law Berg-Ǫnundr were standing in the court-circle, inside the sacred bonds (vébǫnd), disputing about the inheritance. The court also sat in this circle (see above). When things were turning out in favour of Egill, Queen Gunnhildr commanded her brother Askmaðr and the royal retainers to attack the court in order to prevent injus- tice. The men ran to the court and cut the ropes where “the sanctuary” had been staked out, broke the hazel poles and drove the court away, since no one was carrying arms there. The violation of the sacred space seems to be the turning point of this narrative.64 According to the saga, which was composed during the first half of the thirteenth century, this evil deed caused a rebellion against King Eiríkr and Queen Gunnhildr in the shires of Vík and Trøndelag. Soon the king together with his wife and their children had to flee the land. Even if this episode is literary and fictitious, the narrative reveals an attitude towards crimes committed against sacred space, which may have existed in both the Viking Age and the Early Middle Ages. This attitude also appears in a stanza quoted in the saga, which is believed by most scholars to have been composed by the historic Egill

63 Den ældre Frostathings-Lov I,5. NGL 1, 128. Trans. Larson. 64 Cf. Lincoln 1994, 69ff. Sacred Sites, Ritual Restrictions and Religious Strategies 305 himself around 935. Here Egill cries out his disgust against King Eiríkr’s [and Queen Gunnhildr’s] crime:

Svá skyldu goð gjalda, gram reki bǫnd af lǫndum, reið sé rǫgn ok Óðinn, rán míns féar hánum; folkmýgi lát flýja Freyr ok Njǫrðr af jǫrðum, leiðisk lofða stríði landáss þann er vé grandar.65

In this stanza Egill invokes the gods and asks them to avenge the deeds of the king. King Eiríkr is here called “he who destroys sanctuaries” (þann er vé grandar). This expression probably refers to the violation of his retainers against the sacred bonds (vébǫnd) at the Gulathing, but it could also refer to any evil deed the king had done against the pagan sanctuaries in Norway. The expression granda vé means plainly ‘to harm or injure a sanctuary’. In any case, the verse and the prose together give strong indications of how sacred space at both sanctuaries and thing-places was conceived during the transition period.

8.4 The Ruler as the Protector of the Sanctified Space

The sacred bonds (vébǫnd) thus separated the sacred space from its profane surroundings at the thing-place in both Frosta and Gula, and probably at many other “sacred” assembly places in Scandinavia. All those who entered such areas had to leave their weapons behind, and the people standing inside the bonds were treated as inviolable. The crime committed by King Eiríkr and Queen Gunnhildr in violating this space was thus very serious in the eyes of contemporaries. They acted in a way diametrically opposed to what a just and decent ruler was expected to do. As noted above, a just ruler in Trøndelag, according to one Viking Age poem, was called vés valdr ‘the ruler (protector) of the vé sanctuary’. Similar conditions probably existed in other places too. According to the Viking Age poem Ynglingatal, for instance, one pagan Svea

65 “Thus may the gods repay him (Eiríkr) for the seizure of my property; may the powers drive the prince from his lands; may Óðinn and the gods be wrathful. Freyr and Njǫrðr make the tyrant flee from his estates; may the god of the land hate the oppressor of princess, who destroys temples”. Text and trans. E. O. G. Turville-Petre 1976. 306 Chapter 8 ruler was called vǫrðr véstalls ‘the custodian of the sacred stand (cultic site)’. The chieftain or local ruler Sibbi, mentioned in the Rök inscription, was considered as vīaværi ‘the custodian of the vi sanctuaries’ (see ch. 5). The sacred space was thus supposed to be protected by the chieftain, earl or the king, and it could not be violated. The chieftains Þórhaddr of Mære and Þórólfr of Mostr even proclaimed ritual rules which had to be obeyed when entering the sacred space in Iceland (see above). They thus cared for and protected their sanctuaries, assembly places and sacred lands with taboos.66 Landnámabók records that Þorgrímr Kjallaksson and his brother-in-law threatened to desecrate the holy assembly place at Þórsnes by discharging their bodily needs there. The Þórsnesingar (Þórólfr’s sons) could not tolerate such desecration of this holy place. It ended with a great fight where several men were killed.67 The major point in the episode of Egils saga is that King Eiríkr’s and Queen Gunnhildr’s violation of the consecrated sphere led to a process where they both de-authorized their own power by breaking the law.68 They acted in the opposite way to a just ruler, who must protect the court-circle, since it was regarded as a sacred space. The rule and order that prevailed at such places must have been in the interests of the ruling power. The sacred space of both sanctuaries and thing-places constituted the most important arenas where the rulers could manifest their own power, by showing their authority, wealth and glory in public and in ceremonial circumstances. The ritual restrictions and taboos were therefore actually in favour of the ruler at such gatherings. When he appeared on the public scene he did not need to feel physically threatened by enemies. No weapons were allowed to be carried at such places. In Egils saga 25, for instance, we read that “it is a custom here to meet the king unarmed”.69 In this passage King Haraldr was attending a ceremonial feast (var á veizlu) at Voss in Hǫrðaland, which took place in a hall (stofa), probably a ritually protected place. In the warlike society of Viking Age Norway it must have felt safe for the leaders to inhabit places where ritual and religious restrictions prevailed, i.e.

66 Eyrbyggja saga 3 reports that Hrólfr (Þórólfr) watched over a Þórr’s hof (varðveitti . . . Þórshof ) in Mostr, before he emigrated to Norway. Eyrbyggja saga, Ísl. Fornr. 4, 6. 67 Þat þolðu eigi Þórsnesingar, er þeir vildu saurga svá helgan vǫll. Landnámabók S 85, H 73, Ísl. Fornr. 1, 126. 68 King Eiríkr is even called a lǫgbrigðir by Egill in one stanza quoted in the saga. Egils saga, Ísl. Fornr. 2, 165. 69 Þat mun hér vera siðr, at menn gangi vápnlausir fyrir konung, . . . Egils saga, Ísl. Fornr. 2, 63. Sacred Sites, Ritual Restrictions and Religious Strategies 307 the sacred places. If someone attacked the ruler with weapons at such a place, he or she not only offended this person and his office, but also the religious and judicial rules which prevailed at such space. These regulations were established and sanctioned by old custom (siðr). Indeed, any such violation had to be considered a capital crime, since it was directed against the society and perhaps even the cosmos itself. If the ruler was attacked when attending the sanctified grounds, for instance in the ceremonial building or at the sacred meadow, the anger of the mythical powers was also aroused. Such a violation could lead to disturbances in the cosmos, because the sanctuary was regarded as the mythic landscape on earth. Therefore the taboos and ritual restrictions at cultic sites in Trøndelag could be seen as a strategy connected to the rulers and political power, not only in this region, but also elsewhere in Norway, Iceland and Svetjud.

8.5 Burning or Breaking Down the Ceremonial Building: An Attack against the “World” of the Ruler

By appearing in places which were considered sacred and ritually protected against violence, the ruler not only gained physical security. Most likely his person also gained an aura of sanctity, when entering such space. The sanctuary was regarded as a microcosmos and a threshold to the Other World, that is, a meeting place for humans and divine beings (see ch. 6 above). Violations against the ceremonial building or the sacred objects situated there were actually regarded as an attack against the entire kingdom, or even against the entire “world” of the ruler and the cosmos. Since the ruler was considered as the owner of the cultic site, he was also the one who had to protect these places and objects from enemies, because they represented such important values. Frands Herschend has argued that a recurrent pattern seen in the archaeological materials and in written sources, such as the Old English poem Beowulf, is the fighting at and destruction of feasting halls. The keeping of a hall is a manifestation of the hall owner. When competing for power it was therefore important for the hall owner’s enemies to destroy this building, since it was the most important symbol of rulership. Herschend thus states:

In Beowulf, king Hroðgar flatters himself on having built a most splendid hall in the settlement where he lives and the act of building the hall plays an important role, being the inauguration of his kingship. Likewise the initial armistice at Finnsburg (Beowulf, vv. 1086–7) contained an agree- 308 Chapter 8

ment saying that a hall should be put in order for the thane of the slain Danish king Hnæf and his retainers. Clearly one cannot be a leader without a hall, however temporary. . . . sagas like the fight at Finnsburg have their obvious counterparts in the archaeological remains of the Late Iron Age. It seems fair to propose the hypothesis that the struggle for power among the leading families was to a certain degree a matter of fighting each other with the purpose of destroying each other’s hall. Smashing rather than plundering was the keynote of this kind of political rather than economic warfare.70

The destroying of halls has even been described as ritualized warfare. The enemies did not want to steal or plunder, as Herschend noted, even though there may have been many expensive objects in these houses.71 It is quite possible that the act of destroying or burning down the hall referred to a performative ritual action of breaking down the entire world or cosmos of the enemy chieftain and challenging his rules and authority. Perhaps the episodes mentioning the desecration of sacred sites in the Sagas of Icelanders could be interpreted in a similar way. King Eiríkr’s and Queen Gunnhildr’s attack against the sacred area of the Gula assembly could be seen as symbolic ritual act against the power of the local and regional authorities, who were about to give Egill victory at the trial. A similar purpose may have been behind Þorgrímr’s defiance at the holy assembly of Þórsnes, with the intention to breake down the authority of the Þórsnesingar. In the texts referring to the conversion period it is quite clear that the destruction of ceremonial buildings in Svetjud, Trøndelag and Iceland was a symbolic and strategic action intended to exterminate the values and world views of the enemies. Snorri Sturluson reports in Heimskringla how hof sanctuaries in late tenth-century Trøndelag were destroyed by King Óláfr Tryggvason when he embarked on his mission work. The king’s enemies were the fylki chieftains who wanted to preserve the old customs. In Óláfs saga Tryggvasonar, for instance, Snorri describes how King Óláfr Tryggvason dismantled the hof of Lade and removed all the property and all the decorations from it and from the cultic image(s) (af goðinu/af goðinum).72 He took a large ring from the door of

70 Herschend 1998, 36f. 71 Nordberg 2003, 115; Andrén 2014, 98ff. 72 . . . lætr brjóta ofan hofit ok taka allt fé ok allt skraut ór hofinu ok af goðinu [goðunum]. Óláfs saga Tryggvasonar 59, in Heimskringla I, Ísl. Fornr. 26, 308. Sacred Sites, Ritual Restrictions and Religious Strategies 309 the hof, whereupon he had the building burned down.73 This story relates that not only the hof was desecrated but also other ritual objects related to it, such as the decorations, the images, and the door ring. It also says that the farmers were very upset about the king’s treatment of the sanctuaries and sent war arrows to every fylki. King Óláfr maintained the same strategy when he attended the sacrificial feast at the hof of Mære, in the inner reaches of Trøndelag, where he struck the statue of Þórr off the podium and killed the leader of the farmers, Járn-Skeggi.74 The desecration of the sanctuary and the death of their chieftain made the farmers weak, and they finally accepted Christianity. Similar conversion stories appear frequently in the Old Norse texts. When Christian kings, such as King Óláfr Tryggvason, arrived at pagan cultic sites, the hof buildings were burned down (brenna hof ), and the hǫrgr sanctuaries were broken (brjóta hǫrga). According to the version of Óláfs saga Tryggvasonar in Flateyjarbók, King Óláfr “had the hof buildings burnt down and the hǫrgr sanctuaries broken and had churches erected instead”.75 He also destroyed the cultic images.76 By such symbolic and performative actions, the missionaries hoped that the pagans would lose their old beliefs and instead turn to the Christian God. Latin sources likewise record that the Christian missionaries wanted to burn down the pagan sanctuaries.77 Adam of Bremen mentions that Adalvard

73 Síðan lét Óláfr konungr brenna hofit. Heimskringla I, Ísl. Fornr. 26, 308. 74 Óláfs saga Tryggvasonar 69, in Heimskringla I, Ísl. Fornr. 26, 317. Cf. Oddr’s Saga Óláfs Tryggvasonar, Ísl. Fornr. 25, 279f. 75 . . . þa let hann brenna hof en briota hǫrga ok ræisa j stadinn kirkjur . . . Flateyjarbók I, 285. 76 . . . let ok eyda oll skurgod edr brenna edr briota . . . Flateyjarbók I, 285. This passage is actually supported with a skaldic stanza dating back to the twelfth century and composed by Hallar-Steinn. In his poem Rekstefja King Óláfr is praised since he condemned the many hǫrgr sanctuaries and had cultic houses burned ( firna mǫrg ok hǫrga/ brenna blóthús láta). He also ordered the pagan cultic images to be destroyed (bað hann heiðin goð meiða). Skj. B1, 527; A1, 545. 77 Several sources indicate that during the conversion period the Church deliberately offended traditional customs all over Europe. See Maier 2003, 116f.; critically considered by v. Padberg 1995. Bede (Hist. I, 30–32; II, 15; III, 30), for instance, reports that the pagan Anglo-Saxons had sanctuaries which were violated by the Christians. According to him (Hist. II, 13), the former pagan cult leader (primus pontificum) Coifi (seventh century) desecrated a sanctuary called Godmunddingaham, east of York, and broke several traditional rules in order to display his new belief: “For it was not before lawful for a priest of the sacrifices (pontificem sacrorum) either to wear harness or to ride on other than mare. Girded therefore with a sword about his loins he took a spear in his hand, and mounting the king’s war horse set forth against the idols. Which sight when men saw they thought he had been mad. But for all that he stayed not to profane the temple, as 310 Chapter 8 the Younger came to Svetjud, avid to preach the Gospel and in a short time lead the people in Sigtuna and Uppland to the Christian faith. He and Bishop Egino agreed that they should go to the pagan temple of Uppsala and offer Christ some fruit of their labours there, for they would willingly undergo every kind of torture for the sake of destroying that house which was the seat of pagan superstition. For if it was torn down, or preferably burned, the conversion of all the people might follow (see ch. 4 above).78 Adam also mentioned the violation of the pagan images elsewhere in Sweden. Wolfred, from England, for instance, smashed an image with an axe, but was killed by the pagans for that deed,79 while Bishop Egino smashed the well-known “idol of Fricco (Freyr)”, which was located in Västergötland, not far from the church of Skara.80 As mentioned above, it was not only the pagan ceremonial buildings which were destroyed, but also other cultic structures and objects appearing at the sanctuaries were torn down. King Óláfr Haraldsson abolished the worship of rocks, woods, water, and trees when doing mission work in Norway.81 It is interesting to note that missionaries in Continental Europe cut down trees and cosmic posts, which were objects of pagan worship.82 These objects may very well have been expressions or symbols of the pagan world view and cosmos. The act

soon as he approached near unto it, casting thereupon the spear which he held in his hand; and much rejoicing for the acknowledgement of the worship of the true God he commanded the company, which was there with him, to set fire to and destroy the temple together with all its courts.” This text was written more than hundred years after the event and it seems to be imbued with a Christian tendency. Cf. Kuhn 1978, 234f. In any case, the restrictions associated with sacred sites may build on authentic information. Cf. Å. V. Ström 1975, 272; v. Padberg 1995, 167ff.; Simek 2003, 237. 78 Adam IV, 30. These plans were never put into action, however. When the pious King Stenkil (Steinkell) heard about them, he kept the bishops from any such undertaking, and declared that they would immediately be punished with death and he would be driven from the kingdom. Probably everyone in Svetjud would also relapse into paganism. 79 Adam II, 60. 80 Adam IV, 9. (Latin text is quoted in 7.2 above). Adam also states that the bishops Adalvard and Egino travelled around the communities of the Götar and smashed pagan “idols”. Adam IV, 30. 81 Viðbætir við Óláfs sǫgu hins helga, Flateyjarbók III, 246. 82 There is evidence indicating that the Church deliberately desecrated holy trees devoted to the pagan gods. Willibald described in Vita Bonifatii (c. 768) how the holy oak of Jupiter (. . . appellatur robor Iobis . . .) in Geismar, Hessen, was cut down by the Christians (Clemen, Fontes, 1928:45). Jupiter was here probably an interpretatio romana for Donar (Þórr) (e.g. Simek 2006, 63f.; von Padberg 1995, 148; critically considered by Palm 1948, 49–74). The motive was probably to show that the pagan god Jupiter/Donar was without Sacred Sites, Ritual Restrictions and Religious Strategies 311 of cutting down these trees symbolically implied that the pagan “world” was destroyed. Perhaps also the authority of the ruler, who was supposed to protect the sanctuary, also was broken in connection to such hostile actions.

8.6 The Renovation and Ending of Ceremonial Buildings and Sanctuaries

The idea that the conservative Earls of Lade wanted to display themselves as custodians of the pagan cultic sites appears in several sources. Besides the kenning and designation of Earl Sigurðr as vés valdr there are references indicating that his son Earl Hákon also sought to acquire a similar image. According to the sources he seems to have been a devoted pagan worshipper, as his father and grandfather had been. Both Fagrskinna and Heimskringla return to this theme several times. It is quite clear from Snorri’s Haralds saga Gráfeldar that the pagan cult in Norway was destroyed by the Christian sons of Eiríkr blóðøx.83 Eiríkr and his wife Gunnhildr converted to Christianity while living in exile in England. Haraldr Gráfeldr Eiríksson and his brother therefore broke down the pagan sanctuaries and sacrificial cult when they returned to Norway. Snorri writes thus: “But when they [the sons of Eiríkr] entered upon the government in Norway they had no success in converting the inhabitants, and all they accomplished was to destroy the heathen fanes and to break up the sacrifices (þá brutu þeir niðr hof ok spilltu blótum), and this brought them much enmity. There came bad seasons in their time . . .”84 During Hákon Hlaðajarl’s reign, however, heathendom was restored. He commanded that all people throughout his dominions were to maintain the hof sanctuaries and the sacrifices in them (at menn skyldu halda upp hofum ok blótum). Hákon is therefore praised in the skaldic poem Vellekla 15–16 by Einarr Helgason thus:

[15] Ǫll lét senn hinn svinni [16] Ok herþarfir hverfa, sǫnn Einriða mǫnnum Hlakkar móts, til blóta, herjum kunn um herjuð rauðbríkar fremsk rœkir hofs lǫnd ok vé banda; ríkr, ásmegir, slíku; áðr veg jǫtna vitni nú grœr jǫrð sem áðan

power and that nothing happened when his tree was cut down. For the pagan population this act must have been wretched and execrable. 83 Haraldr Saga Gráfeldar, in Heimskringla I, Ísl. Fornr. 26, 198–224. Cf. Fagrskinna, Ísl. Fornr. 29, 95–103, 112. 84 Haraldr Saga Gráfeldar, in Heimskringla I, Ísl. Fornr. 26, 203. 312 Chapter 8

valfalls um sjá allan – aptr geirbrúar hapta – þeim stýra goð– geira auðrýrir lætr áru garðs Hlórriði farði. óhryggja vé byggja.85

The author of Fagrskinna comments on these stanzas while quoting them,86 stating that Earl Hákon was very rich and performed sacrifices more strictly than before. Soon the year’s crops of grain and herring improved and the earth germinated and bore fruit.87 These texts display a specific ideological message. A legitimate ruler must protect the sanctuaries and maintain the sacrificial cult in order to create the requisite base for a good relationship to the gods.88 By means of cultic activities the ruler gains divine guidance and the land can then blossom. He is then able to create luck for his people. Vellekla’s description of the paradise-like conditions in Norway under Earl Hákon’s reign and the restoration of the old sanctuaries strikingly resembles the mythical world after Ragnarök as described in Vǫluspá.89 In Vellekla, thus, Earl Hákon’s deeds gained cosmic dimensions. By

85 (15) “The wise man [Hákon] at once allowed the men of Þórr to uphold the plundered hof lands and shrines of the gods well known to the people; and then the god of the fence of the spears (i.e. the warrior) carried the wolf of slaughter (i.e. the sword) over the mountains and all the sea—the gods guide that man [Hákon].” (16) “And the sons of the Æsir (i.e. the gods), needful to men, return to the sacrifices; the mighty tender of the red board of the meeting of Hlǫkk (i.e. Hákon) wins fame by such a thing. Now the soil flourishes as before—again the destroyer of the wealth of the spear- bridge [Hákon] allows the merry messengers of the gods to inhabit the sanctuaries.” Text and trans. E. O. G. Turville-Petre 1976, 60f. modified. This poem was reproduced in several manuscripts. Cf. Skj. A1, 122–131. See also Finnur Jónssons’s reconstruction in Skj. B1, 117–124 and SkP I, 280–329. 86 St. 15’s second helming does not occur in Fagrskinna. 87 Hákon var ríkr ok tók at efla blót með meiri freku en fyrr hafði verit. Þá batnaði brátt árferð ok kom aptr korn ok síld, grøri jǫrðin með blómi, . . . . Fagrskinna, Ísl. Fornr. 29, 111. Snorri makes similar comments on these stanzas: . . ., þá bauð hann þat um ríki sitt allt, at menn skyldu halda upp hofum ok blótum, ok var svá gǫrt. . . . Inn fyrsta vetr, er Hákon réð fyrir landi, þá gekk síld upp um allt land, ok áðr um haustit hafði korn vaxit, hvar sem sáit hafði verit. Óláfs Saga Tryggvasonar, in Heimskringla I, Ísl. Fornr. 26, 241–243. 88 See F. Ström 1981, 450. 89 Sér hón upp koma ǫðro sinni/ iorð ór ægi iðiagrœna; falla forsar, flýgr ǫrn yfir,/ sá er á fialli fisca veiðir. . . . Muno osánir acrar vaxa, / bǫls mun allz batna, Baldr mun koma. Vǫluspá 59 and 62. Cf. F. Ström 1981, 451. Helmut de Boor (1930) argued that the terminology of Vǫluspá resembled the vocabulary of the poems devoted to Earl Hákon. Perhaps Vǫluspá had some connection to the circle of skalds connected to the tenth-century earls of Lade. Cf. Schier 1981, 415ff.; Steinsland & Meulengracht Sørensen 1999, 36ff. Sacred Sites, Ritual Restrictions and Religious Strategies 313 re-establishing the pagan cult and restoring the sanctuaries which had been damaged by the Christians, Earl Hákon transformed the misery in Norway to a mythical scene identified as the “perfect new world” after Ragnarök, where “soil flourishes as before” and “the gods inhabit the sanctuaries”. Earl Hákon is seen in this context as a saviour and a creator of prosperity and wealth. These mythical allusions are clearly used to create authority and legitimacy. As long as the ruler maintains the sacrificial cult and protects the cultic sites and the sanctified grounds, the divine powers and ideal condition of the mythical world may be manifested among humans. Earl Hákon’s restoration of the pagan cultic sites may therefore be seen as a repetition of the cosmogony.

8.6.1 A Periodic Renovation of Cultic Sites as a Repetition of the Cosmogony The context of these traditions of Earl Hákon’s renovation of sanctuaries in Norway is the struggle between pagans and Christians during the conversion period. But it is possible that pagan rulers even before the encounter with Christianity restored or reconstructed the pagan sanctuaries periodically. Several halls and cultic houses seem to have been rebuilt in that way on the very same site as the foregoing building.90 The banqueting hall on Södra Kungsgårdsplatån in Uppsala was rebuilt several times during the Merovingian Period.91 The ceremonial building (the compact hall I A) of Helgö was also rebuilt during several phases from the Roman Iron Age up to the Viking Age on the very same spot.92 The same pattern can be seen at the outdoor sanctuary at the bare rock beside this building. It was restored in six phases between c. 200 and 830.93 This can be compared with the cultic house of Uppåkra. It was also restored in six phases from c. AD 200 until c. 800.94 This house did not change shape during this long process, but retained its archaic architecture during the entire period.95 The hall of Husby, Glanshammar, Närke, dated to the Migration and Merovingian Periods, was rebuilt in several sequences (“house II” and “house III”) and remained in its old location during the entire period,96 and the same applies to the hall at Lunda, Södermanland.97 The Iron

90 See Hållans Stenholm 2012, 173ff. 91 Hedlund 1993; Nordahl 1993; 1996. 92 See Herschend 1995; 1998; K. Lamm 1999; 2004. 93 T. Zachrisson 2010, 84. 94 L. Larsson & Lenntorp 2004, 9f. 95 L. Larsson 2006c, 15. Cf. T. Zachrisson 2010, 84. 96 Ekman 2000, 25. 97 G. Andersson et al. 2004.; Skyllberg 2008. 314 Chapter 8

Age hall building at Hovde, Ørland, in Sør-Trøndelag, was also rebuilt several times.98 There are also several layers of settlements at the cultic house of Mære.99 The hall of Borg in Lofoten had at least two phases. This is also the case with the Viking Age hall of the chieftain farm at Hofstaðir, Mývatnssveit, in north-eastern Iceland.100 Perhaps the periodic renovation of ceremonial buildings and sanctuaries could be seen as a repetition of the cosmogony. As the world and cosmic process must be renewed by means of calendrical sacrifices, the cultic sites must also be restored and sanctified periodically. The responsibility for this renewal and consecration lay with the ruler. When the ruler is called vés valdr or vǫrðr véstalls this implies that he not only was seen as the one who must take care of and protect the sanctuary, but that he also periodically had to reconstruct and consecrate it. By re-establishing and restoring the old sanctuaries the ruler transformed the earthly fane into a mythical stage identified as the “perfect new world”, where “soil flourishes as before” and “the gods inhabit the sanctuaries” as described in Vellekla.

8.6.2 The Ritual Ending of a Sacred Place It has recently been noted that certain rituals also were performed when a sacred place or ceremonial building was closed or terminated.101 Probably the actions mentioned above, performed by the chieftains Þórhaddr and Þórólfr when leaving Norway for Iceland, could be categorized as a ritual ending of a sacred building. They dismantled their hof sanctuaries and took the high- seat posts and some earth from under the old sanctuaries with them (see above). The ritual ending of a sacred place or ceremonial building is probably also attested in archaeology. Anne Carlie has noted that some sacred objects were occasionally placed in the post-holes when the house had already been demolished. Such rituals have been noticed, for instance, at the cultic site of Järrestad in Scania.102 Sometimes the building was buried, as in the case of the cultic house at Borg, Norrköping.103 During the eleventh century the walls were knocked down and the foundation was covered with gravel. About 100 metres to the east the church was then erected. At Ranheim in Trøndelag too, rituals were performed when the cultic site was ended. The posts belonging

98 Grönnesby 1999; 2000. 99 Lidén 19969; 1996; 1999. 100 See e.g. Johanson & Munch 2003; Lucas & McGovern 2008; (ed.) Lucas 2009. 101 See e.g. Carlie 2004, 193ff. 102 Carlie 2004, 194. 103 Nielsen 1996, 102. Sacred Sites, Ritual Restrictions and Religious Strategies 315 to the assumed hof were pulled out and all of the wood removed. The assumed hǫrgr was carefully covered with stones and clay. The entire cultic site was then entirely covered with earth. These actions took place during the transition period in the tenth century.104 As noted above, some rituals were also performed at Hofstaðir when the great hall (AB) was dismantled in phase III (1030–1070) (see ch. 4 above). Several acts of closure were performed, such as the ritual depositions of the skulls. In Uppsala the large hall building was likewise ritually ended at the beginning of the ninth century. After it burned down the spiral decorations were intentionally placed in the post-holes (see ch. 4 above). Most likely the ruler of these places played a central role in these rituals. Since he was designated vés valdr or vǫrðr véstalls he also took charge of ending them ritually. With these performative rituals he may have signalled to the local or regional group as well as the competing magnates that he no longer had any claims to authority at this place or area.

Sources thus indicate that the rulers consecrated lands and places with a complex of rituals which were performed in public. In connection with that, they proclaimed the ritual rules of these sacred lands and places. By means of these performative rituals they demonstrated that it was their rules and restrictions the group or community must follow. The rulers were also the custodians of the sacred places. If the ruler (at least the king and the earl) was successful with these ritual assignments he could create a good relationship with the divine powers. This relation also had a cosmic dimension, in that the land he ruled over could be transformed into a “paradise-like world”. The ruler could also ritually end or close a cultic site, by removing certain important objects from the place, such as the high-seat posts. He could also re-establish and consecrate sanctuaries which had been desecrated. These ritual roles were performed in public and rendered the ruler designations such as vǫrðr véstalls. In my opinion this ritual strategy was applied to chieftains, earls and kings in the investigation areas. In the next chapter we will turn to other types of rituals performed by the rulers when appearing in sacred space and ceremonial buildings.

104 Rønne 2011. Chapter 9 Ceremonial Feasts as Ritual Strategies

Social anthropologists and historians of religions have emphasized that rites often support an established social order. Important changes in society are frequently marked by rituals, especially aspects which concern political power. The rituals help to maintain balance between individuals, and they accentuate social values and develop feelings of social dependence. Some of them also emphasize social boundaries and indicate that some individuals are more important than others.1 Even if such theories have been criticized for their tendency to be ahistorical,2 some of the ideas behind them may find fruitful application in the present study of ancient Scandinavian ceremonial buildings and sanctuaries. Sources indicate that rituals sanctioned a particular social structure and that the ceremonies legitimated the social position of the rulers. Many studies during the last two decades have been occupied with the rituals that took place in the Germanic and Scandinavian Iron Age halls and cultic houses.3 I will mainly focus on those rituals visible in the evidence from Svetjud, Trøndelag and Iceland, which were connected to rulers and their strategies for gaining legitimacy. My hypothesis is that the rulers in all three investigation areas played a central ritual role during the sacrificial feasts and ceremonial banquets in the halls and cultic houses. Some of these rituals were related to gift-giving systems, where the religious-political authorities had the opportunity to manifest their wealth and generosity. All these rituals may have been crucial for the elite when constructing power and authority.

9.1 Hákonar saga góða and the Ceremonial Feasts at Lade and Mære

One of the most extensive descriptions of pre-Christian cultic feasts in Scandinavia appears in Snorri Sturluson’s Hákonar saga góða 13–18 (see the text in ch. 4 above). According to this text, such feasts were celebrated at hof sanctuaries in Viking Age Trøndelag. Earl Sigurðr was a great sacrificer. He was responsible for the sacrificial feasts (blótveizlur) in Lade as a deputy for

1 See e.g. Radcliffe-Brown 1952, 153–55; cf. Platvoet 1995; Bell 1997, 128ff.; Rappaport 2000. 2 See Bell 1997, 27–29. 3 See e.g. Herschend 1993; 1997; 1998; Brink 1996a; 1999b; Enright 1996; Lönnroth 1997; Näsström 2001a; 2002; Nordberg 2003; Sundqvist 2002; 2006; 2013 and 2014.

King Hákon the Good and his father King Haraldr. Snorri depicts in general terms the interior of a hof, the objects and actions there, for instance, the ritual slaughter of animals and the treatment of the sacrificial blood. He also tells us that the beaker ( full) was to be carried around the fire, and that the chieftain who gave the feast was to bless the beaker (signa fullit) as well as all the sacrificial food (ok allan blótmatinn). When the toasts were drunk to the gods, ritual formulae were proclaimed, such as til árs ok friðar. This description of ceremonial feasts in Hákonar saga góða has been much debated (see ch. 4 above). The sharp criticism of Snorri’s text has partly been justified, but in some senses it failed. Recently scholars have modified it and produced, in my opinion, a sounder view of Snorri’s account, where both possibilities and problems of source criticism are noted (see ch 4 above). In my opinion, several elements in Snorri’s text are supported by other reliable sources and may thus be genuine. In what follows I will discuss some details in this description and try to estimate what parts of it may build on more ancient traditions. I will first present those elements in Snorri’s description which I consider inauthentic as regards pre-Christian rituals. Than I will point out aspects I believe to be authentic. My purpose here is to illuminate these notions with sources from the investigation areas. These comparisons are intended to show that these elements existed in all these areas.

9.1.1 The Sacrificial Terminology There are several elements in Snorri’s description of the pre-Christian ceremonial feasts which seem to be obscure and inauthentic. He writes, for instance, thus:

At this feast all were to take part in the drinking of ale. Also all kinds of livestock were killed in connection with it, horses also; and all the blood from them was called “hlaut”, and “hlautbolli”, the vessel holding that blood; and “hlautteinar”, the sacrificial twigs.

In his book Das Opferfest von Lade, Klaus Düwel analysed the semantic development of the words hlaut, hlautteinar and hlautbolli used in Snorri’s text. Düwel argued that in most cases Snorri either misunderstood these concepts or mixed them up with Christian ideas with no basis in pre-Christian culture.4 Düwel is no doubt right in his criticism. It is very doubtful that the term hlaut designated ‘sacrificial blood’ in pre-Christian language and that hlautte inn and hlautbolli meant ‘sacrificial twig’ and ‘sacrificial bowl’. The meanings

4 Düwel 1985, 21–38. 318 Chapter 9 of these terms are not attested in skaldic poetry or the Eddic lays. Most likely hlaut originally referred to a ‘lot’ used in divination rituals.5 Snorri is perhaps the first one to use the term hlaut in the sense of ‘sacrificial blood’. Most likely the scribes of Eyrbyggja saga and Kjalnesinga saga took over this meaning when applying this term in their descriptions of hof buildings. When describing the “pagan use” of the sacrificial twigs (hlautteinar) Snorri writes: “These were fashioned like sprinklers, and with them were to be smeared all over with blood the ‘stallar ’ [altars] and also the walls of the hof, within and without, and likewise the men present were to be sprinkled with blood.” It is quite possible that Snorri was influenced by ecclesiastical ceremonies in this description.6 In the medieval church holy water was sprinkled over the congregation. Perhaps examples in the Old Testament may also have played some role in this context, since it is said there that Moses sprinkled blood on his people.7 An indication of Christian impact on this text may also be seen in the comparison that Snorri made between the hlautteinar and the stǫklar ‘holy water twigs’. The Old Norse term stǫkkull (sg.) designates the Latin aspergillum, an object that was often used in the holy water rituals of the medieval Church. Hence the information about the use of the hlautteinar in the contexts of hof buildings is not completely reliable (see however below on the blood rituals).8 According to Hákonar saga góða 14, the chieftain who held the feast was to bless the beaker as well as all the sacrificial food (. . . signa fullit ok allan blótmatinn). Scholars have been suspicious about this ritual. Ernst Walter, for instance, convincingly argued that Old Norse signa was a loan-word from the Christian concept, Latin signare (signo) ‘to make a sign (cross-sign)’ (see ch. 4).9 Whether the loan-word signa indicates that the content of the text is late and not built on pre-Christian notions is uncertain, however, for this term appears in Eddic and skaldic poetry, as well as in an eleventh-century runic

5 Hultgård (1996, 53) referred this term to the Gothic hlauts (a masculine a-stem) which reproduces the Greek concept κλῆρος (klēros) ‘lot, part’. It seems as if it is this meaning which is evident in the first element of the concept hlautviðr ‘lot-twig’ in Vǫluspá 63, for instance. 6 See Hultgård 1993, 230; 1996, 45. See also Simek (2006, 151) who comments on the term hlaut teinn thus: “the concept would appear to have its origins more in the imagination of the Christian authors of the High Middle Ages, who created a heathen counterpart to the sprinkler (asperges) of Christian liturgy.” 7 See Exodus 24:4–9. Cf. Düwel 1985, 34ff. 8 A more original use of the term hlautteinn is attested in a lausavísa (c. AD 980) by Þorvaldr Koðransson, where it has the meaning ‘lot-bough, lot-twig’. Skj. B1, 105. Cf. Skj. A1, 110. The term hlautteinn should probably be associated with the concepts blótspánn or hlautviðr, i.e. the ‘twig(s)’ used in divination rituals. 9 Walter 1966. Ceremonial Feasts as Ritual Strategies 319 inscription from Uppland, namely U 942. The runologist Henrik Williams therefore argues that it is possible that this word was borrowed before conversion and thus may have been incorporated into the pagan religious terminology, in e.g. Sigrdrífomál 8: full scal signa.10 We can thus not rule out that the chieftains consecrated the beaker and the food (even if the terminology is uncertain). In chapter 17, the verb signa occurs in connection with the ritual of making the krossmark and hamarsmark over the beaker. Snorri says there that the king during the sacrificial ritual at Lade took the horn and made the sign of the cross (krossmark) over it (since he was a good Christian). The Old Norse kross mark is probably (as Walter supposed) derived from the Christian expression signum crucis, while hamarsmark appears for the first time in Hákonar saga góða and thus cannot be accepted as authentically pre-Christian. Ernst Walter was probably on the right track when he stated that this must be related to Christian terminology. The expressions gerði krossmark yfir and gerði hamars mark yfir seem to be derived from Christian signum crucis facere, and thus the latter expression cannot be regarded as authentically pre-Christian.11 According to Snorri’s text, Earl Sigurðr adduces another reason why the king made the sign of a hammer over the beaker: “The king does as all do who believe in their own might and strength.” The formula trúa á matt sinn ok megin “to believe in one’s own might and strength” has long been accepted as reflecting a genuine pre-Christian belief and basic attitude of the pagans. The alliteration of the expression seems to suggest that it was ancient.12 Recently scholars have argued that this notion is a topos often used in Christian literature. Pagans and barbarian were permeated by conceit and arrogance and thus believed in their own strength.13 Snorri’s explanation (put in the mouth of Earl Sigurðr) probably does not reflect an authentic pre-Christian belief. It is quite clear that Snorri’s description of ceremonial meals in Hákonar saga góða must be regarded as a late literary reconstruction. The application of several terms and expressions seems to be confused as well as the action of making the sign with the hammer over the horn, as noted above. Olsen’s, Walter’s and Düwel’s criticism were with no doubt well-founded on these points (see also ch. 4 above). Therefore Snorri’s text must be treated with cau- tion. This does not mean that everything in his account must be dismissed. In what follows I will present some themes in Snorri’s description which I believe existed in all three investigation areas during the Viking Age.

10 Williams 1996b, 79. 11 Cf. Walter 1966, 365f.; Hultgård 1993, 229. 12 See e.g. de Vries 1956–57. 13 See Weber 1981; von See 1988, 94f.; Hultgård 1993, 229; Simek 2006, 208. 320 Chapter 9

9.2 Calendrical Religious Feasts

Calendrical rituals and feasts exist in most cultures around the world.14 They give “socially meaningful definitions to the passage of time, creating an ever renewing cycle of days, months, and years . . . Calendrical rituals occur periodically and predictably, accompanying seasonal changes in light, weather, agricultural work, and other social activities.”15

9.2.1 Trøndelag The feasts in Trøndelag depicted in Hákonar saga góða 13–18 could be defined as calendrical. They were performed “at the beginning of winter” (um haus tit at vetri (vetrnóttum)) and at Yule (ON jól).16 This information seems to be built on an old and firm oral tradition.17 Such feasts may be attested in written sources from all three investigation areas. Equivalents to the concept jól ( júl) (neutr. pl.) (‘Yule-feast’, ‘Gastmahl, Festmahl’; cf. the kenning Hugins jól ‘Fest, Trinkgelage des Raben’) can also be seen in other Germanic areas, such as Gothic fruma Jiuleis, and Old English gēol (neutr.), gēola (mask.), giūli (mask.), and iūla (mask.).18 According to Snorri, there seem to have been three calendrical religious feasts celebrated in Trøndelag (and Svetjud) during the pre- Christian era. These cultic feasts coincided temporally with the quarters of the year.19 The feasts at the winter-nights (vetrnætr) took place in October, the midwinter sacrifices (miðsvetrarblót) in January, and the feast to greet the summer ( fagna sumri) took place in April. Perhaps there was also a Midsummer feast in July (miðsumarsblót).20 In Snorri’s account it seems as if rich farmers in northern Norway also arranged sacrificial feasts at their farms three times every year. In Óláfs saga Helga ch. 117 he writes about Sigurðr Þórisson, who lived at Ǫmd at Þrándarnes (Trones):21

While heathendom still prevailed he was accustomed to have three sacrifices made every year, one at the beginning of winter, one in midwinter,

14 Cf. Widengren 1969, 243ff.; Bell 1997, 102ff. 15 Bell 1997, 102. 16 Heimskringla I, Ísl. Fornr. 26, 171f. 17 See e.g. de Vries 1956–57, §§305–314; Näsström 2001a, 106ff.; 2002, 156ff.; Sundqvist 2002; Nordberg 2006b. 18 See e.g. Hultgård 2000, 100 and Simek 2006, 379f. 19 See e.g. Nordberg 2006b, 76; Nordberg & Sundqvist 2008, 13. 20 See Heimskringla I, Ísl. Fornr. 26, 20, 314ff.; Heimskringla II, Ísl. Fornr. 27, 194. 21 Perhaps the large island of Hinney in the northern Lofoten Islands. Ceremonial Feasts as Ritual Strategies 321

and the third at the beginning of summer. And when he had adopted Christianity he still persisted in this way of holding feasts. So he had a great banquet in autumn for friends, then a jól feast in winter to which he invited a great many people; and at Easter too he had numerous guests.22

Earlier in Óláfs saga Helga 107 Snorri wrote about the calendrical sacrificial feasts which were still performed on the winter nights (i.e. October) in Inn-Trøndelag during St Óláfr’s reign. The Christian king was informed that all toasts were then brought to the Æsir deities, and they followed old heathen custom. Cattle and horses had been slaughtered and the pedestals (of images) reddened with their blood, and sacrifices were performed in order to improve the harvest.23 Later, during the winter, the king was told that the people from the inner reaches of the Trondheimsfjord had assembled in great numbers at Mære and that there were big sacrifices made at midwinter for peace and a good season (blótuðu þeir þá til friðar ok vetrarfars góðs).24 The king was upset about this and arranged a meeting with the chieftain of the farmers called Ǫlvir and accused the Þrœndir of arranging midwinter sacrifices (miðs vetrarblót). Ǫlvir told the king that the farmers were not guilty of that. They just celebrated the jól feasts, including social drinking.25 After Easter the king despatched a man called Þoraldi to a meeting. He was the king’s steward on his estate at Haugr. He told the king that the people living in the interior parts of Trøndelag were pure heathen in their belief, even though a few there were baptized:

22 Hann var því vanr, meðan heiðni var, at hafa þrjú blót hvern vetr, eitt at vetrnóttum, annat at miðjum vetri, þriðja at sumri. En er hann tók við kristni, þá helt hann þó teknum hætti um veizlur. Hafði hann þá um haustit vinaboð mikit ok enn jólaboð um vetrinn ok bauð þá enn til sín mǫrgum mǫnnum, þriðju veizlu hafði hann um páska ok hafði þá ok fjǫlmennt. in Heimskringla II, Ísl. Fornr. 27, 194. Trans. Hollander. 23 Þat haust váru sǫgð Óláfi konungi þau tíðendi innan ór Þrándheimi, at bœndr hefði þar haft veizlur fjǫlmennar at vetrnóttum. Váru þar drykkjur miklar. Var konungi svá sagt, at þar væri minni ǫll signuð Ásum at fornum sið. Þat fylgði ok þeiri sǫgn, at þar væri drepit naut ok hross ok roðnir stallar af blóði ok framit blót ok veittr sá formáli, at þat skyldi vera til árbótar. Heimskringla II, Ísl. Fornr. 27, 177. 24 Heimskringla II, Ísl. Fornr. 27, 178. 25 ’Hǫfðum vér,’ segir hann, ’jólaboð ok víða í heruðum samdrykkjur. Ætla bœndr eigi svá hneppt til jólaveizlu sér, at eigi verði stór afhlaup, ok drukku menn þat, herra, lengi síðan. Er á Mærini mikill hǫfuðstaðr ok hús stór, en byggð mikil umhverfis. Þykkir mǫnnum þar til gleði gott at drekka mǫrgum saman.’ Heimskringla II, Ísl. Fornr. 27, 179. 322 Chapter 9

It is their custom to perform a sacrifice in the autumn to welcome winter, a second at midwinter, and a third in summer to welcome its arrival.26

The king took then his ship and sailed up to Mære where he had Ǫlvir and many other men killed. He converted the people to Christianity, placed priests there, and finally erected and consecrated the church there. There are other and more ancient sources indicating that the calendrical festivals mentioned by Snorri existed in Norway during the transition period. Ágrip, for instance, indirectly supports the existence of four pre-Christian calendrical feasts. It describes how Óláfr Tryggvason substituted ritual drinking for pagan sacrifices and libation rituals at Christmas and Easter, St John’s Mass (summer) and at St Michael’s Mass (fall).27 A praise poem called Haraldskvæði, composed by Þorbiǫrn Hornklofi about AD 900, gives strong evidence of the calendrical Yule feast (ON jól) in southern Norway (see below).

9.2.2 Svetjud and Iceland According to Ynglinga saga 8, there were three cultic feasts in Svetjud every year. A sacrifice was to be made at the beginning of winter to ensure a good season, one in midwinter for good crops, and a third in summer for victory.28 It has been argued that this description too is based on genuine traditions.29 It is uncertain, however, whether any of the three feasts is identical with the sacrifice at the vernal equinox at Uppsala, mentioned in scholion 141 of Adam’s Gesta.30 Adam’s feast may, on the other hand, correspond to the Disthingen referred to in Upplandslagen and the feast of the month of gói which appears in Snorri’s Óláfs saga helga 77. There is evidence for a name of an ancient lunar month called Distingstungel in local dialects in central Sweden.31 A cult of the dísir at Uppsala is also attested in the Aðils traditions (see ch. 4 above). This feast thus probably has a historical background, in the same way as the Yule feast ( jól). The Yule feast

26 En þat er siðr þeira at hafa blót á haust ok fagna þá vetri, annat at miðjum vetri, en it þriðja at sumri, þá fagna þeir sumri. Heimskringla II, Ísl. Fornr. 27, 180. Trans. Hollander. 27 . . . felldi blót ok blótdrykkjur ok lét í stað koma í vild við lýðinn hátíðadrykkjur jól ok páskar, Jóansmessu mungát ok haustǫl at Míkjálsmessu. Ágrip, Ísl. Fornr. 29, 22. 28 Þá skyldi blóta í móti vetri til árs, en at miðjum vetri blóta til gróðrar, it þriðja at sumri, þat var sigrblót. Heimskringla I, Ísl. Fornr. 26, 20. 29 de Vries 1956–57, §§305–314; cf. Nordberg 2006b, 76–78. 30 Scholion 141 says that the sacrifice was at the vernal equinox (hoc sacrificium fit circa aequinoctium vernale), i.e. the beginning of the spring planting. This sentence is uncertain, as it is only attested in manuscript A3, dated to 1434. Hultgård 1997, 30. 31 OÖD I, 321. Ceremonial Feasts as Ritual Strategies 323 was celebrated at the first full moon after the first new moon following the winter solstice, while the Disthingen took place at the third full moon.32 Since these feasts took place during the winter, some of the rituals were probably performed inside halls and cultic buildings, where the rulers also took part in the ceremonies. In Austrfararvísur a calendrical feast called álfablót was celebrated in western Svetjud (see ch. 8 above and below). Most likely calendrical feasts also took place in Iceland during the Viking Age. The sacrificial feasts at the winter-nights (vetrnætr) are well-attested in the sources. According to the Sagas of Icelanders, they took place in the autumn, probably October. Eyrbyggja saga 37, for instance, mentions that Snorri goði gave a great feast at his home in western Iceland (Snæfellsnes) and invited a lot of friends during the autumn, at the winter nights (at vetrnóttum). The guests were served with ale and there was drinking.33 Some sagas report that the object of sacrifice during these feasts was Freyr. In Gísla saga Súrssonar 15, for instance, we may read thus:

Þorgrímr decided to hold a feast at the end of autumn to celebrate the coming of the winter-nights. There was to be a sacrifice to Freyr, and he invited his brother Bǫrkr and Eyjólfr Þorðarson and many other men of distinction.34

The events of this saga take place in north-western Iceland. A sacrificial feast at the winter-nights is also mentioned in chapter 10 of Gísla saga. It seems as if chieftains and magnates prepared and perhaps paid the cost of these feasts:

In those days it was the custom to celebrate the coming of winter by holding feasts and a “winter-night sacrifice”. Gísli no longer sacrificed after he left Viborg, but he still held feasts and showed the same magnanimity as before.35

32 Nordberg 2006b. 33 Annat haust eptir at vetrnóttum hafði Snorri goði haustboð mikit ok bauð til vinum sínum. Þar var ǫldrykkja ok fast drukkit. Eyrbyggja saga, Ísl. Fornr. 4, 98. 34 Gísla saga Súrssonar 15, Ísl. Fornr. 6, 50. Trans. S. Regal. The Icelandic text is quoted in 7.5 above. 35 Þat var þá margra manna siðr at fagna vetri í þann tíma ok hafa þá veizlur ok vetrnáttablót, en Gísli lét af blótum, síðan hann var í Vébjǫrgum í Danmǫrku, en hann helt þó sem áðr vei zlum ok allri stórmennsku. Gísla saga Súrssonar 15, Ísl. Fornr. 6, 36. Trans. S. Regal. 324 Chapter 9

Víga-Glúms saga 6 states that the “winter-night sacrifice” (probably October) was intended for the dísir at Vigfúss’s farm in Vǫrs (Voss), Norway. It was thus called dísablót.36 This information does not harmonize with the conditions at Uppsala in Svetjud, where the dísablót seems to have been celebrated in the winter or early spring.37 It seems, however, as if the “winter-night sacrifice” sometimes concerned the dísir, at least in western Scandinavia and Iceland. Several sources indicate this. Egils saga tells how King Eiríkr Bloodaxe and his wife Gunnhildr visited Atløy in Norway for the dísablót during the autumn.38 It was a splendid banquet with much drinking inside the house. The context of Þáttr Þiðranda ok Þórshalls likewise indicates that a sacrificial feast in honour of the dísir was celebrated at the “winter-nights” (ueitzlan uar buin at uetrnattum).39 The term Þorrablót indicates a sacrificial feast held at midwinter in Iceland, probably during the month called Þorri. Vatnsdæla saga mentions sacrifices performed “towards summer” among heathen people in Iceland: “Now his mother, Ljót, will sacrifice towards summer, as she usually does according to their custom . . .”40 Archaeology also supports the existence of calendrical feasts in pre- Christian Iceland. The ritual slaughter of cattle in prime age at the hall building of Hofstaðir, Mývatnssveit, was performed on a seasonal basis.41 Thus calendrical feasts took place there, with gatherings of many people. A lot of meat was consumed and perhaps the lavishness of it all was also an expression of power claims.42 The remains of the sacrifices discovered under Frösö church, Jämtland (close to Trøndelag), have been re-analysed. Radiocarbon analyses of the bones indicate that pagan sacrifices were maintained there during the Late Viking Age. Seasonal analyses indicate that the sacrificial rituals took place in late autumn, early spring, and possibly around the summer solstice. These results harmonize quite well with the information about calendrical sacrificial feasts mentioned in the written sources.43

36 Þar var veizla búin at vetrnóttum ok gǫrt dísablót . . . Víga-Glúms saga 6, Ísl. Fornr. 9, 17. 37 See de Vries 1956–57, §311; Näsström 2001a; 2002, 167. 38 Egils saga, Ísl. Fornr. 2, 105–111. 39 Flateyjarbók I, 420. Cf. F. Ström 1954. 40 . . . nú mun Ljót, móðir hans, blóta í mót sumri, sem hon er vǫn at þeira sið . . . Vatnsdæla saga, Ísl. Fornr. 8, 67. Trans. Wawn. 41 See mainly Lucas & McGovern 2008. 42 Lucas 2009, 404ff. 43 Magnell & Iregren 2010. Ceremonial Feasts as Ritual Strategies 325

9.3 The Ruler as Sacrificer

Hákonar saga góða 13–18 tells us that Sigurðr Hlaðajarl was “the most ardent heathen worshipper” (var inn mesti blótmaðr), as Hákon his father had been. We are also told that Sigurðr maintained all sacrificial feasts there in Trøndelag on the king’s behalf. When the Christian king Hákon the Good appeared in Trøndelag and did not want to take part in the traditional sacrifices with the farmers, they became very upset. At the Frostaþing they asked him to participate in the rituals. Later in this story the king was more or less forced to take part in the sacrificial meals at the sanctuaries of Lade and Mære and dedicate toasts to the native deities worshipped there. It thus seems as if the Viking Age rulers of Trøndelag not only organized the cult (see ch. 5); they were also expected to perform the sacrifice on behalf of the people during the public religious feasts. Perhaps they consecrated the beaker and the food, for instance. Similar information is attested in older texts, too.44 Ágrip (1190 AD), for instance, tells us that the farmers of Trøndelag (Þrœndir) in Mære demanded that King Hákon the Good “sacrifice as other kings used to do or we throw you out of the country, if you do not follow us in this thing”.45 Fagrskinna, written sometime between Ágrip and Heimskringla, also reproduces this tradition in short words. At the assembly in Mære “the Þrœndir gave him [King Hákon] two things to choose between. Either he should sacrifice as kings used to do before and in such a way maintain old law/custom for good crops and peace; or else they would throw him out of the country”, and “because of this threat and out of love for his friends the king yielded to their demands and sacrificed”.46 The three prose texts in Fagrskinna, Ágrip and Heimskringla thus suggest that the people expected the king to take part in the religious feast. According to Ágrip and Heimskringla he had to eat the sacrificial meat (i.e. horse liver) in order to become a legitimate king. By means of these rituals he ratified ancient law. According to Fagrskinna and Heimskringla, he had power over good crops and peace (ár ok friðr) for the people. It thus seems as if he was a necessary ritual link to the gods during these feasts, according to these texts.47

44 On this theme, see particularly Jón Hnefill Aðalsteinsson 1992; 1998, 57ff. 45 . . . báðu hann blóta sem aðra konunga í Nóregi, “ella rekum vér þik af ríki, nema þú gerir nekkvern hlut í samþykki eptir oss.” Ágrip, Ísl. Fornr. 29, 8. 46 . . . ok á því þingi gørðu Þrœndir konunginum tvá kosti, at hann skyldi blóta eptir vanða enna fyrri konunga ok fylla svá en fornu lǫg til árs ok friðar, elligar mundu þeir reka hann af ríkinu, ef hann vildi ) ...... , þá gørði hann eptir bœn þeira ok blótaði. Fagrskinna, Ísl. Fornr. 29, 80. 47 See Jón Hnefill Aðalsteinsson 1992; 1998. 326 Chapter 9

In my opinion, these traditions from the transition period indicate quite clearly that the political rulers were expected to perform sacrifices when participating in the calendrical ceremonial feasts in Viking Age Trøndelag. In other cases, when the king did not attend these feasts, probably the earl or the local chieftain performed some central rituals.48 These political leaders were probably considered to have a specific relationship to the Other World, whether genealogical or cultic (see ch. 3 above). This is probably one reason why they were expected to lead and execute these religious ceremonies.

9.3.1 In Svetjud Snorri’s ideas that the pagan rulers in Norway were expected to perform rituals and be involved in public cult were thus not taken out of thin air. He had several sources for these notions (see ch. 4 above). These ideas are actually attested in other parts of Late Viking Age Scandinavia and in other types of sources. Scholion 140 of Adam of Bremen’s text, for instance, says:

When not long ago the most Christian king of the Svear, Anunder, would not offer the demons the prescribed sacrifice of the people, he is said, on being deposed, to have departed “from the presence of the council, rejoicing” that he had been “accounted worthy to suffer reproach for the name of Jesus.”49

According to the context it seems as if these events took place at the public feast in Uppsala. It should be noted that this scholion appears in the oldest manuscript of Adam’s text, namely A2, which usually is dated to about 1100 (see ch. 4 above). It was thus made only 25 years after Adam wrote his auto- graph on contemporary conditions in Uppsala. It is likely that conditions at the sanctuary of Uppsala changed during the early decades of the eleventh century. After King Óláfr sœnski Eiríksson converted to Christianity there was probably no king of the Svear who performed the central rituals there. The traditional order was thus disturbed. This may be the reason why kings are so absent in Adam’s account of the Uppsala cult. This scholion indicates that the people who gathered in Uppsala during the eleventh century still expected the ruler to perform the major sacrifices on

48 See Meulengracht Sørensen 1991b. 49 Nuper autem cum rex Sueonum christianissimus Anunder sacrificium gentis statutum nollet demonibus offerre, depulsus a regno dicitur a conspectu concilii gaudens abisse, quoniam dignus habebatur pro nomine Iesu contumeliam pati. This scholion was probably written by Adam himself, as it occurs in the oldest manuscript A2. Ceremonial Feasts as Ritual Strategies 327 their behalf. At such great sacrificial feasts in Uppsala, however, there must have been several cultic leaders involved in the different ritual actions (see ch. 14 below). Most likely the king was seen as a kind of leader of these specialists and perhaps a necessary symbolic link to the gods. King Anunder was thus driven from the public feast and assembly at Uppsala, since he would not perform as a cultic leader at the sacrifices. This image agrees well with some information given in the U-version of Hervarar saga (c. 1300), which says that the Svear removed King Ingi Steinkelsson from office, since he would not make sacrifices on their behalf at the assembly which probably was held at Uppsala:

At the assembly that the Svear held with King Ingi, they gave him two choices. Either he would maintain the ancient law or he had to give up the kingdom. King Ingi said that he was not ready to give up that faith, which was the right one. The Svear shouted and threw stones at him and drove him off the assembly place. Sveinn, the king’s brother-in-law, remained at the thing. He told the Svear that he would make sacrifice on their behalf if they would give him the kingdom. All assented to that proposal. Sveinn was taken as king over all Svetjud. A horse was led to the meeting place, dismembered and distributed for eating, and the sacrificial tree was reddened with the blood. All Svear thus abandoned Christianity, re-introduced the sacrifice and drove Ingi off and he travelled to Västergötland. Blótsveinn was king over the Svear for three winters.50

Because of the change of style, most scholars accept that the chronicle of the Svea kings has been added to the original version of Hervarar saga.51 It may be part of a lost historical source moulded from legendary and mythical elements. Some parts of this tradition are also preserved in Orkneyinga saga, which tells of the Christian King Ingi of Svetjud who fought against paganism.

50 Á þingi nokkuru, er Svíar áttu við Inga konung, gerðu þeir honum tvá kosti, hvárt hann vildi heldr halda við þá forn lǫg eða láta af konungdómi. Þá mælti Ingi konungr ok kveðst eigi mundu kasta þeiri trú, sem rétt væri. Þá æptu Svíar ok þrǫngdu honum með grjóti ok ráku hann af lǫgþinginu. Sveinn, mágr konungs, var eftir á þinginu. Hann bauð Svíum at efla blót fyrir þeim, ef þeir gæfi honum konungdóm. Því játa þeir allir; var Sveinn þá til konungs tekinn yfir alla Sviþjóð. Þá var fram leitt hross eitt á þingit ok hǫggvit í sundr ok skipt til áts, en roðit blóðinu blóttré. Kǫstuðu þá allir Svíar kristni, ok hófust blót, en þeir ráku Inga konung á brott, ok fór hann í vestra Gautland. Blót-Sveinn var þrjá vetr konungr yfir Svíum. Hervarar saga. (Ed) E. O. G. Turville-Petre, 70f. Cf. (Ed) Jón Helgason, 160. My trans. 51 For instance Schück 1910, 14–15; Reifegerste 1989, 196–200. 328 Chapter 9

The farmers, however, chose another king, called Blótsveinn (the brother of the Queen) who retained sacrifices.52 This tradition may have been transferred to Iceland by Markus Skeggjason, who was probably Ingi’s skald before being made Lawman at Þingvellir in 1084.53 Blótsveinn’s existence has been questioned and his story has been regarded as a medieval exemplum, that is, a homiletic story.54 Yet Blótsveinn (in Old Swedish Blodhsven) is also mentioned independently of Icelandic traditions in The Legend of St Eskil (Legenda sancti Eskilli). Here, the sacrifices are said to occur in Strängnäs.55 It thus seems as if the Late Viking Age cult community in Svetjud expected the king to perform or at least participate in the common sacrifices at the Uppsala sanctuary. In my opinion, there is no reason to doubt this information. Perhaps the sacrifice was only valid when the king took part in the rituals. A plausible interpretation is that the king was the most important mediator between the human and the divine world during the sacrifices at this sanctuary. His ritual role in the cult was thus necessary for the community, and per se it also legitimated his high position in other social contexts. Medieval written sources thus support the idea that pagan rulers were involved in the sacrificial feasts in eastern Scandinavia, which included ceremonial meals and drinking rituals. This idea may also be indicated in other types of sources, sources which are undoubtedly pre-Christian. Some runic stone monuments of the Blekinge group may give information about political leaders in cultic functions. They were carved not before AD 550 and perhaps not after AD 750.56 Stentoften (DR 357), Gummarp (DR 358) and Istaby (DR 359), have a proper noun Haþuwulfr. On Stentoften and Istaby there are also two persons called Hariwulfr and Heruwulfr. The Björketorp stone (DR 360) can be associated with this group, although it lacks the name Haþu- wulfr. It has almost exactly the same curse as Stentoften. As the name Haþuwulfr appears on three stones in this area scholars believe that he was a local ruler in Blekinge some time around AD 600.57

52 Kom því svá, at bœndr tóku sér annan konung, þann er þá helt til blóta, Sveinn, bróðir drott ningar, ok var kallaðr Blót-Sveinn. Orkneyinga saga, Ísl. Fornr. 34, 89–90. Snorri was also acquainted with this tradition and mentioned Blót-Sveinn. See Heimskringla III, Ísl. Fornr. 28, 263. 53 This theory was first suggested by Henrik Schück 1910, 17. Cf. Reifegerste 1989, 198f.; P. H. Sawyer 1991, 37; Foote 1993, 24. 54 Lönnroth 1996, 150f. 55 SRS II: 1, 391–399. 56 On datings of these inscriptions see Krause 1966, 203ff., Moltke 1985, 137–147; Williams 2001. 57 E.g. Krause 1966, 203–214; Williams 2001; Düwel 2008, 21f. Ceremonial Feasts as Ritual Strategies 329

The interpretation of the Stentoften inscription was debated throughout the twentieth century. I have examined it earlier,58 and will here just refer to Lillemor Santesson’s interpretation. She read and interpreted lines 1–3 thus:

niuhAborumR niuhagestumR hAþuwolAfRgAfj

With nine bucks, with nine stallions Haþuwulfr gave good growth.59

Haþuwulfr performed a sacrifice where he offered nine bucks and nine stallions. In that way he gave the people a good crop. Perhaps he also arranged a great sacrificial feast at his hall building or at an outdoor sanctuary, with plenty of meat for all the people. Santesson’s interpretation has many advantages compared to previous attempts, and many scholars accept it.60 Her interpretation can be supported on several grounds, also from the perspective of history of religions. For instance, the number nine corresponds to the sacrifices at Uppsala and Lejre mentioned by Adam and Thietmar.61 As in Uppsala, Lejre and Lade, the runic inscription indicates that horses were sacrificed in Blekinge too.62 It should also be noted that the Proto-Nordic word jāra in this inscription (here marked with the j-rune, i.e. an ideograph (Begriffsrune)) is probably equivalent to the Old Norse ár that appears in Snorri’s text in connection with sacrifices, meaning ‘good crops’ (see below).63 Stentoften thus supports the idea that a ruler performed a sacrifice, or at least that he commissioned it.

58 Sundqvist 1997. 59 Santesson 1989; 1993. 60 E.g. Williams 1990, 36; 2001; Düwel 1992, 348–353; 2008, 21f.; M. Stoklund 1994, 166–168; Sundqvist 1997; Antonsen 2002; Schulte 2006. Some scholars have also been sceptical of Santesson, especially regarding the interpretation of the sequence gestumR, see e.g. Reichert 2003. The runologist Michael Schulte (2006) has, however, defended Santesson’s interpretations against these recent objections. 61 Adam of Bremen IV, 27; Thietmar of Merseburg (Chronicon) I, 17. See Sundqvist 1997, 164–168. 62 See quotation from Adam of Bremen IV, 27 in ch. 4 above. Thietmar mentions that horses were sacrificed in Lejre: Est unus in his partibus locus, caput istius regni Lederun nomine, in pago, qui Selon dicitur, ubi post VIIII annos mense Ianuario, post hoc tempus, quo nos theo phaniam Domini celebramus, omnes convenerunt, et ibi diis suimet LXXXX et VIIII homines et totidem equos, cum canibus et gallis pro accipitribus oblatis, immolant . . . ’. Chronicon I, 17. 63 See Hultgård 1993; 2003c; 2007. 330 Chapter 9

This runic inscription indicates that persons belonging to the ruling elite officiated in the public cults and performed ceremonial acts. The changes in sacrificial custom that have been pointed out by archaeologists may reflect changes in social organization,64 from tribal chieftains to “petty kinglets” and finally more permanent ruling families. The family of Haþuwulfr seems to be a dynasty, according to the naming customs.65 The new ruling elite used religious ceremonies when manifesting their power and sovereignty. They probably institutionalized public cults which they tied to themselves and their residences.

9.3.2 Iceland Sources report that the Viking Age chieftains in Iceland, i.e. the goðar, also had central ceremonial functions during public meetings, such as law assemblies and cultic feasts. According to the Hauksbók redaction of Landnámabók, the chieftain (goði) was to wear a ritual ring on his arm at such meetings, especially when he himself was at the head of the affairs (the text is paraphrased in ch. 5). This ring probably had an important religious significance since it otherwise was placed on the ritual platform in every chief hof. The text also reports that the ring was first reddened in the blood of a neat which the chieftain himself had sacrificed at the meeting. The connection between chieftains (goðar) and sacrifices may be indicated by skaldic poetry dating back to the tenth century. In a lausavísa (8), composed by Hallfreðr vandræðaskáld, Óðinn is called hrafnblóts goði ‘the goði of the raven sacrifice’.66 Eyrbyggja saga 3–4 also suggests that the chieftain Þórólfr Mostrarskegg in Iceland (and Norway) performed central roles at the hof sanctuary and executed parts of the sacrificial rituals (see ch. 3 and 4 above). Similar information is reported about the chieftain called Þorgrímr goði in Kjalnesinga saga 2. He is described as blótmaðr mikill ‘a great sacrificer’, who had a hof sanctuary built on his farm at Kjalarnes in south-eastern Iceland.67 Most likely the information in Kjalnesinga saga and Eyrbyggja saga is to some extent dependent on Hákonar saga góða 14, when describing the chieftain and his role as a sacrificer. They do not add much new information in relation to them. On the other

64 See Fabech 1991; 1994. 65 See Sundqvist & Hultgård 2004. 66 Skj B1, 158; cf. A1, 168. 67 It is also mentioned that Þorgrímr goði was hard against those who did not sacrifice: Þorgrímr goði gaf mikinn gaum at þeim mǫnnum sem ekki vildu blóta. Kjalnesinga saga 3, Ísl. Fornr. 14, 9. Ceremonial Feasts as Ritual Strategies 331 hand, Landnámabók, which probably is older than Snorri’s text, also describes Þórólfr as a blótmaðr mikill ‘a great sacrificer’. Other Icelandic traditions similarly report that chieftains (goðar) were involved in the sacrificial rituals at the public sanctuaries. Gísla saga Súrssonar 15 mentions that Þorgrímr Þorsteinsson goði arranged a feast in his drinking hall at the end of autumn to celebrate the coming of the winter-nights. During this feast he also sacrificed to Freyr (blóta Frey).68 It has been suggested that this is one of the oldest Sagas of Icelanders.69 Thus we may assume that the information about the cultic role of the chieftain was based on a quite reliable oral tradition, since it also harmonizes with information from other sources.

Snorri’s information that King Hákon was supposed to perform some central rituals during the sacrificial feasts in Viking Age Trøndelag has many parallels in sources referring to the ritual functions of rulers in all three investigation areas. They were thus not only organizers of cultic feasts, but also cultic performers during the sacrifices. Most likely they were regarded as a necessary ritual link between the human and divine world, and mediators in giving blessings to the people, such as good crops and peace. The close genealogical and/or cultic relation to the Other World made them suitable for this position and function.

9.4 Smearing Sacrificial Blood on the Sacrificial Platforms

Snorri states in Hákonar saga góða that sacrificial blood (hlaut) was smeared on the platforms (stallar) and on the walls of the hof. As noted above, the term hlaut is dubious in this context. But this does not mean that the ritual itself needs to be suspect. In Óláfs saga Helga 107, Snorri presents similar information without mentioning the term hlaut. He states that sacrificial feasts were still performed in Inn-Trøndelag during St Óláfr’s reign and “that cattle and horses had been slaughtered and the platforms reddened”.70 Klaus Düwel was very critical of these passages in Snorri’s text.71 He argued that the ritual actions connected with the sacrifices in these texts had no Germanic origin, that is, that Snorri had no pre-Christian sources for these

68 Gísla saga Súrssonar 15, Ísl. Fornr. 6, 50. The text is quoted above. 69 Jónas Kristjánsson 1988, 258. 70 . . . at þar væri drepit naut ok hross ok roðnir stallar af blóði ok framit blót ok veittr sá formáli, at þat skyldi vera til árbótar. Heimskringla II, Ísl. Fornr. 27, 177. 71 Düwel 1985, 32ff. 332 Chapter 9 rituals. Rather, when describing how blood was sprinkled on the altars in Trøndelag, Snorri modelled his account on ancient Hebrew rituals described in Exodus. In Snorri’s time, parts of the Old Testament had been translated into Old Norse in a text called Stjórn. In this text there was a passage about the Old Jewish sacrifice in chapter 100 (i.e. Exodus. 24):

Moyses tok halft fornarblodit ok hellti þui i kerit. En halft blod hellti hann yfir alltarann.

According to Düwel this text clearly resembled the passage about the sacrificial blood in Hákonar saga góða 14. Düwel argued, further, that during the Early Middle Ages, the sacrificial rituals of the Old Testament were considered pagan, and pagan religion was thought to be the same in all places; this led Snorri to use this material in his account of the ancient rituals in the Trøndelag. It is possible that Snorri took some expressions from the description in Stjórn when writing about the sacrifices in Trøndelag. It seems, however, as if the word stallr/stalli (m.) is a genuine term in pre-Christian Scandinavian sacrificial contexts (see ch. 7 above). The compound véstallr ‘sacred stand’, for instance, appears in the early poetic language.

9.4.1 Pouring Sacrificial Blood on Ritual Objects and “The Temple Soil” Snorri’s statement in Hákonar saga góða that the sacrificial blood from the animals was smeared on the sacred objects, called stallar, may thus not be totally distorted. His statement that the walls of the hof were smeared with such blood may likewise be reliable. Independent sources indicate that sacred objects and cultic structures were reddened with sacrificial blood in western Scandinavia. In Hyndlolióð 10 the goddess Freyja exclaimed thus about one of her adorants: “He’s made an altar (hǫrgr) for me, faced with stone. Now that stone has turned to glass. He’s reddened the new altar with ox blood.” Admittedly Hyndlolióð is attested in a late manuscript from the second half of the fourteenth century, and usually the poem is regarded as young and dated to the twelfth century.72 Yet the content indicates that the poet had good information about ancient customs and rituals. It seems quite unlikely that a medieval Christian skald would have created from his own imagination the idea that the stones of the hǫrgr had been reddened by the blood from the ox. These stones had probably acquired a glaze as an effect of the many sacrificial fires on the blood flowing

72 This manuscript is called Flateyjarbók Gks 1005, fol. (F). For a recent treatment, see Cöllen 2011, 192–247. Ceremonial Feasts as Ritual Strategies 333 over them.73 It is possible that some kind of burnt offering took place there, where the victim was sent as a gift to the gods. Similar information to that in Hyndlolióð also appears in other texts. According to the U-version of Hervarar saga, a holy tree called blóttré, situated in Svetjud, was reddened with the blood from a sacrificial horse (see above). Whether the word blóttré in this text really refers to a tree is somewhat uncertain. One possibility is that this term designates the wood of the ritual platform or altar (ON véstallr, stallr, stalli), which was reddened with the sacrificial blood.74 According to Ynglinga saga 15, the Svear smeared the stallar in Uppsala with sacrificial blood. This text reports that they sacrificed their king called Dómaldi for better crops (blóta til árs), and that they attacked and killed him and reddened the ritual platforms (rjóða stalla) with his blood.75 This account has no historical background, but it is built on the ninth-century poem Ynglingatal 8 (5), which definitely renders a pagan world view. The poem reports that the Svear carried swords against their own ruler and stained the earth red (at fold ruðu) with his blood. They sacrificed him, eager for crops (árgjǫrn . . . of sóa skyldi). No stalli/stallr is mentioned in this stanza. The poem indicates, however, that the earth of the place was reddened with sacrificial blood. A passage referring to a possible sacrifice made to the álfar in Kormáks saga also points in the same direction. The injured hero Þorvarðr was recommended to pour the blood of a bull on to one of the hills/mounds (rjóða blóð graðungsins á hólinn) inhabited by álfar and to prepare a meal for them from the meat of the bull in order to heal his injury.76 As we noted in ch. 5, several chieftains who dismantled their hof buildings in Norway when travelling to Iceland brought some of the “temple-soil” to the new land. One man who did this was the hofgoði, Þórhaddr the Old, according to Landnámabók. According to Eyrbyggja saga 4, Þórólfr also brought “temple- soil” from Norway to Iceland, i.e. the soil from under the ritual platform or altar (svá moldina undan stallanum). The information from the written sources indicates that the soil below the ritual platform/altar (sg. stalli/stallr) was holy since the sacrificial blood had poured down on it.77 The written sources also

73 Hultgård 1996, 32. 74 See Hultgård 1993, 237. 75 Ynglinga saga 15, in Heimskringla I, Ísl. Fornr. 26, 32. Cf. Óláfs saga Helga 107, in Heimskringla II, Ísl. Fornr. 27, 177. 76 Kormáks saga 22, Ísl. Fornr. 8, 288. The text is quoted and discussed in ch. 12 below. 77 Archaeology may also produce similar evidence. Sacred objects, such as gold foils and fragments of glass, have been detected in the earth close to certain post-holes in the hall 334 Chapter 9 report that ritual objects which were placed on the stallar were reddened with blood during the sacrificial rituals, for instance the “oath-ring”.78 In several non-Christian religions, the custom of pouring blood on sacrificial altars is common, for instance in Greek religion, as well as in the Jewish burnt offering.79 In the Greek sacrifice the blood was collected in containers and then poured or sprinkled on the altar, on the hearth or in sacrifice pits. Small animals were cut up on the altar itself.80 Also interesting is the Roman sacrificial practice whereby blood reddened the altar in front of the temple.81 Lucretius, for instance, gives a vivid picture of “altars washed with the stream- ing blood of beasts”.82 Some of the blood was caught in containers. Virgil mentions that when Aeneas’ companions sacrificed to Hecate, they “put knives to the animals’ throats and catch the warm blood in vessels”.83 There is even evidence indicating that the priest tasted the blood. The animal was then butchered, cooked and sometimes eaten by the worshippers.

9.4.2 Were the Stallar Outdoors? It is quite possible that the killing of animals in Scandinavia likewise took place outdoors at a ritual structure. It has been suggested that such outdoor structures may be depicted in Late Iron Age iconographic sources from eastern Scandinavia. On the Gotlandic picture stone from Lärbro (Hammars I), one scene shows a man lying on a platform or structure (fig. 16a and 16b). Another man is standing beside this platform. He is pricking the man lying on the platform with a spear. According to Anders Hultgård, this structure could be

buildings (see ch. 11 below). Perhaps the earth where they were found was once regarded as sacred “temple soil”. Cf. Gunnell 2001. 78 According to the Hauksbók redaction of Landnámabók, the oath-ring was placed on the platform. This ring was reddened in the sacrificial blood of a neat by the goði. Landnámabók, H 268, Ísl. Fornr. 1, 313. 79 Burkert 1983; Hultgård 1996, 45. On the blood at sacrifices, see especially Ekroth 2002, 242–276. 80 Burkert 1983. Fine descriptions of Greek sacrifices and sacrifices can be find in Hesiod (Theogony, v. 535–57) and especially in Homer (Odyssey I, Book III, 432–63, 70–72), where blood also plays a role. According to Gunnel Ekroth (2005), only some of the blood was sprinkled on the altar at the thysia sacrifices, while most of it was consumed as food by the worshippers. Unlike Hebrew religion, there is no evidence for blood taboo or prohibition for eating blood in Greek religion. 81 See e.g. Ogilvie 1994, 41–52; Beard 2002, 36ff.; Näsström 2005. 82 Lucretius, V, 1200. 83 Vergil, Aeneid VI, 248–249. On Roman sacrifices, see also Cato: De agri cultura 83, 132, 134, 139–41. Ceremonial Feasts as Ritual Strategies 335

Figure 16a Gotlandic picture stone from Lärbro (Hammars I). Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet.

Figure 16b Detail of the picture stone from Lärbro (Hammars I). Photo: Olof Sundqvist.

interpreted as a stallr or stalli.84 Hultgård hesitates, however, as to whether this cultic structure should be interpreted as an altar. Perhaps they had another

84 “Das Gestell, das auf dem Bildstein abgebildet ist, könnte dem kultischen Gegenstand stalli oder stallr nahekommen, den Snorri und der Verfasser der Eyrbyggjasaga erwäh- nen.” Hultgård 1993, 234f. 336 Chapter 9 cultic function than being the place where the sacrificial objects were presented and given to the deity. This ritual platform seems anyhow to be located at a cultic site out-doors, beside a tree, where a victim is hanging. One puzzling thing in Snorri’s account of the sacrificial ritual at Lade is that the sacred platforms are referred to in the plural, stallar. Whether these stallar were located inside or outside the hof is not mentioned. In both Landnámabók and Eyrbyggja saga the stalli/stallr, in the singular, seems to be inside the hof. Archaeological evidence indicates that altars or ritual platforms could be placed either inside a house or outdoors. Inside the small cultic house (5 × 7 m) in Borg, Östergötland, for instance, there was a foundation built up with strong and flat stones.85 It was located along the eastern wall, at the end of the path, which split the room in two halves. Just to the south of the foundation two amulet rings were discovered. The foundation has been interpreted as a véstallr.86 The traces of a rectangular platform inside the hall building V A at Helgö, mentioned by Arrhenius, could perhaps also be interpreted as reflections of a véstallr. In my opinion, such an interpretation is better than the proposal seiðhjallr (see ch. 4 above). An outdoor ritual structure or platform (perhaps equivalent to a véstallr or hǫrgr) might be evidenced at Lilla Ullevi in Uppland.87

9.4.3 Platforms (stallar), Bloody Sacrifices and Divination Rituals It is quite possible, however, that both sacrificial and divination rituals88 were performed under a continuous ceremony at the platforms (stallar) of the ceremonial buildings. As noted in ch. 5 above, Þorvaldr Koðránsson’s lausavísa (tenth century) provided us with early and probably reliable evidence of a gyðja called Friðgerðr and her son Skeggi, who performed such rituals at a stalli in Hvammr in western Iceland. In Kristni saga, which renders this stanza, Friðgerðr was in the hof sanctuary sacrificing (enn Friðgerðr var meðan í hofinu ok blótaði), when the skald Þorvaldr approached them.89 According to the stanza, Friðgerðr was shouting from the pagan “altar/platform” (gall of heiðnum stalla), while her son probably was casting lots. He is called hlautteins hreytir, “the one who cast lots”, and goða sveinn “the servant of the gods” in this context. The sources indicate that the gyðja Friðgerðr and her son performed both

85 A.-L. Nielsen 1996; Lindeblad 1997. 86 A.-L. Nielsen 1996, 98; Vikstrand 2001, 263. 87 On Lilla Ullevi, see Bäck et al. 2008; Hållans Stenholm 2010. 88 On divination rituals in Old Norse sources in general, see particularly Dillmann 2006 (who also discusses their connection to sacrifices, see pp. 109ff.). On the category of divination, see Hultkrantz 1976. 89 Kristni saga 2, (Ed) Kahle, 9. Ceremonial Feasts as Ritual Strategies 337 sacrifices and divination rituals in a ceremonial building, while Þorvaldr prayed to Christ for the people in Hvammr (see ch. 5 above).90 Perhaps the combination of sacrifices and divinations in ceremonial buildings is also unconsciously indicated by Snorri. In Hákonar saga góða we meet the term hlaut. It is likely that this word had nothing to do with ‘sacrificial blood’ in the pre-Christian language as Snorri suggests (see above). The word hlaut probably meant ‘lot’ in that language. On the other hand, Snorri’s confusion of terms may indicate that bloody sacrifices were combined with the casting of lots. Such a combination is evidenced in another text produced by Snorri. In Ynglinga saga 18 a divination ritual was connected with a boar sacrifice: Gekk hann þá til sonarblóts til fréttar . . .91 Some scholars argue that auguries were regularly associated with sacrifices.92 Klaus Düwel rejects this and states that the link between these rituals exists only in Snorri’s texts.93 The pre-Christian auguries performed at Birka, and mentioned by Rimbert, for instance, are not explicitly connected with sacrifices.94 On the other hand, Rimbert never commented in detail on these rituals. Probably no Christians were allowed to participate in them. The divination at Birka might thus have been combined with sacrifices.95 It has been suggested that the Old Norse composite form blótspánn supports the combinations of these actions.96 There is evidence besides Snorri’s texts showing auguries in combination with sacrifices.97 Hauks þáttr hábrókar, for instance, preserved in Flateyjarbók, tells us about interesting rituals connected with a god Lytir, performed in the hall at

90 See Eyrbyggja saga for similar rituals. Ísl. Fornr. 4, 7. 91 Heimskringla I, Ísl. Fornr. 26, 35. The word sonarblót (n.) ‘boar sacrifice’ can be associated with sonargǫltr ‘sacrificial boar’ attested in Helgaqviða Hiǫrvarðzsonar, prose passage after stanza 30. Sonarblót can also be associated with Hyndlolióð 38, where the expression sonardreyri “blood of the sacrificial boar” appears. Cf. Bjarni Aðalbjarnarson, in Heimskringla I, Ísl. Fornr. 26, 35f.; Frense 1982, 244f.; La Farge & Tucker 1992, 244; Heggstad et al. 1993, 401. 92 E.g. de Vries 1956–57, §288, 297f. 93 Düwel 1985, 26f. 94 See Vita Anskarii 18, 26 and 27. The texts are quoted in Sundqvist 2002, 215ff. 95 For a more thorough argumentation, see Sundqvist 2002, 214–224. 96 Düwel (1985, 26f.) is sceptical of this argument and states: “Blót- in der Komposition darf nicht als Zeugnis für eine feste Verbindung zwischen Opferhandlung und Losbefragung herangezogen werden. In der großen Zahl der blót- Komposita hat blót meist die Funktion, das zweite Glied als etwas ‘Heidnisches’ zu charakterisieren (vgl. die Übersetzungen der blót- Komposita bei Cl.-Vigf.).” 97 E.g. Eyrbyggja saga, Ísl. Fornr. 4, 7: Þórólfr Mostrarskegg fekk at blóti miklu ok gekk til fréttar við Þór, . . . 338 Chapter 9

Uppsala. The text is usually regarded as having a low source value, as it is from a late fourteenth-century manuscript and contains fictional elements, but at the same time it has aspects that may be very old. The story tells how King Haraldr Finehair had sent two men, Haukr and Vígharðr, to Bjarmaland, in order to bring furs to the king. They were blessed on their journey by the “witch” Heiðr. King Eiríkr, Haraldr’s enemy, discovered this and sent two retainers, Bjǫrn and Salgarðr, to follow Haraldr’s men. Before they went on with their journey, they needed support from the divine powers. Eiríkr held a feast (væizla) at Uppsala.

Thereafter he brought two carriages to that place where he sacrificed (blotade) to that god who is called Lytir. It was a custom that the carriage had to stay there during the night and that he [Lytir] went to it in the morning. But now Lytir did not appear as he was accustomed to do. And the king was told that Lytir was reluctant to go there. The carriage stood there two nights, but he did not show up. Then the king made much greater sacrifices (myklu meire blot) than before and the third morning they noticed that Lytir had come. And the carriage was so heavy (hofugr), that the draught horses (eykirnir) broke down before they arrived at the hall (til hallarinnar). The carriage was placed in the middle of the floor of the hall (hallargolf ). And the king approached and welcomed Lytir and said he wished to offer him a toast. And he made a full goblet for him (drekka full til hans) and wanted to decide on a matter (male skifta) and take advice (radizst) about the journey . . .98

It has been suggested that Lytir is another name for Freyr.99 An appellation *lytir may occur in some Swedish place-names, e.g. Lytisbærg, at Östra Husby, Östergötland, Lytislunda, in Österrekarne, Södermanland and Litslunda, in Lillhärad, Västmanland.100 It is derived from OSw luter, loter ‘lot which is used in divination’ (cf. ON hlaut) or a Proto-Nordic *hluti-wīhar “Losdeuter, Wahrsager, Opferpriester”.101 The appellative may thus refer to a cultic leader/ specialist who performed divination rituals. In the account of the god Lytir, sacrifices were first made to him. When Lytir indicated that he was present, the deity’s wagon was pulled, perhaps in a procession, to the hall by draught horses (eykr sg.). Carriage and horses may

98 Hauks þáttr hábrókar, Flateyjarbók I, 579f. My trans. 99 Cf. Strömbäck 1928a; de Vries 1956–57, §516; Elmevik 1966; 1990; 2003; Vikstrand 2001, 390ff. 100 Elmevik 1966; 1990; 2003. 101 Elmevik 2003. Cf. Elmevik 1966; 1990. Critically considered by Vikstrand 2001. Ceremonial Feasts as Ritual Strategies 339 have ritual implications. It is said that the carriage became heavy when the god was in it. Auguries were performed before it—King Eiríkr asked questions of the god (i.e. his image) after he had given him the libation sacrifice in the hall. The story may therefore reflect the practice of augury combined with different types of sacrifices. Thus it is possible that divination rituals were performed by the rulers when making sacrifices at the platforms (stallar) of the sanctuaries. These rites were probably performed in order to obtain divine guidance in crisis situations and when something dangerous had to be done. The action to be taken by the ruler thus acquired divine legitimacy.

9.4.4 Bloody Sacrifices and Decapitations—Archaeological Evidence from Hofstaðir New archaeological evidence indicates that ritual slaughter, including bloody sacrifices at ceremonial feasts, took place at the chieftains’ hall buildings at places designated hof in Iceland. Analysis of the bone assemblage from Hofstaðir, Mývatnssveit, in north-eastern Iceland, has generated some sensational results.102 Cattle skulls were retrieved in two clusters from outside the walls of the hall (see ch. 4). These skulls show evidence of specialized butchery and prolonged display on the outside of a structure. “Butchery marks include depressed fracture of the frontals caused by a heavy and immediately fatal crushing blow between the eyes, and (where the base of the skull is preserved) a powerful shearing blow, which would have beheaded the animal.”103 According to Gavin Lucas and Thomas McGovern, this decapitation was an integral part of the sacrificial ritual. So too was the display of the cattle heads at the hall of Hofstaðir. Most important is that this sacrificial ritual involved dramatic and violent elements. The cattle were stunned by a blow between the eyes before being decapitated, with an axe. With good skill, such an act would have produced a fountain of blood, because the heart was still beating as the head came off:

The most likely reconstruction from the forensics of the skulls requires at least a two-person team, one of whom struck the animal between the eyes (effectively killing it and certainly stunning it into momentary immobility) while the second swung a fairly broad-bladed axe at the neck or base of the skull for a beheading stroke. There is a definite concentration of beheading blows from the right side towards the left of the

102 Cf. Orri Vésteinsson 2007; Lucas & McGovern 2008. 103 Lucas & McGovern 2008. 340 Chapter 9

animals’ skulls. If the axe-wielder was right handed, this means this person was probably standing to the right of the sacrificial animal. If the team got their timing right, the beheading stroke would produce a blood fountain as the animal’s heart would still be beating.104

According to Lucas and McGovern, this is a ritual and violent practice intended to dramatize the death. This interpretation could be supported with sources referring to the other parts of the investigations areas. For it seems as if decapitation of the victims (including bloody elements) was a widespread ritual practice at sacrifices in several parts of Scandinavia. Adam of Bremen, for instance, suggests that it was only the heads of the sacrificial animals (and humans) that were sacrificed to the gods at the temple of Uppsala. He writes thus: “The sacrifice is of this nature: of every living thing that is male, they offer nine heads.”105 The term capita should probably be understood in this context in its concrete sense as ‘heads’.106 Perhaps it was only these parts that were given to the gods, while the meat was consumed during the ceremonial feast. This idea could also be supported by means of archaeological evidence from Jämtland. Under the medieval church at Frösön ‘the island dedicated to the god Freyr’, Viking Age animal bones were found around a birch stump.107 These bones were interpreted as remains from animal sacrifices. Most of the bones belonged to the heads of the animals. Perhaps these heads were given to the gods at a sacred tree or a holy grove. The ceremonial meal may have taken place in the Viking Age ceremonial building beside this outdoor cultic site, indicated by the place-name Hov. A similar situation may be found at Borg, Östergötland. At this site a small house was discovered, which was interpreted as a cultic building. Outside this building 75 kilos of unburned bones were discovered, together with 98 amulet rings.108 The bones, which belonged to horses, dogs and pigs, were interpreted as remains of sacrifices. Most of the bones consisted of skulls and jaws from animals. It is thus not impossible to assume that the custom of offering heads of animals was widespread in Scandinavia. This evidence supports the interpretation that the skulls from Hofstaðir could also be seen as sacrificial gifts to the gods, while the meat was consumed during ceremonial feasts celebrated in

104 Lucas & McGovern 2008, 23. 105 . . . ex omni animante, quod masculinum est, novem capita offeruntur . . . Adam IV, 27. 106 For this interpretation, see Hultgård 1997, 32. 107 Iregren 1989; Magnell & Iregren 2010. Cf. Hultgård 1997, 33. 108 Nielsen 1996. Ceremonial Feasts as Ritual Strategies 341 the chieftain’s hall (see below). It also indicates that sacrifices involved bloody rituals as described in the written sources above.

9.4.5 The Meaning and Function of Sacrificial Blood It is not easy to draw a conclusion from Snorri’s description in Hákonar saga góða on the meaning and function of the blood in sacrifices. Since the information on this kind of blood in the sources is so fragmentary many scholars have made a synthetic interpretation, where all available sources have been used. In my opinion, we must be aware of the fact, however, that the meaning and function of blood could have varied in different contexts and in different types of sacrifices. In previous research this type of blood was seen as means for opening a connection and communication with the gods at the communion sacrifice. According to Jan de Vries, a sacred communion between men and deities was established, when reddening and sprinkling the ritual objects and walls with blood: “Der Sinn des Besprengens war der, daß zwischen Mensch und Gottheit eine helige Gemeinschaft gestiftet warden sollte”.109 Also Ulf Drobin emphasizes the communion aspect, where the victim’s blood actually represented the deity’s blood in cultic contexts. This blood was symbolically related and transformed to the mead, which was consumed by the worshippers during ceremonial feasts. Both the blood and the mead were in cultic contexts intermediary of the divine knowledge. Therefore the sacrificial blood was also useful in divination rituals (see above).110 In my opinion the sacrificial blood could also have another function, namely to create a dramatic and emotional effect during the public ceremony. This interpretation could be related to some aspects emphasized in the field of ritual theory. In her book Ritual. Perspectives and Dimensions, Catherine

109 Jan de Vries 1956–57, §289. 110 Drobin 1991. We have actually some support for this notion in classical sources. In Strabo’s Geography some customs by the Cimbri is mentioned, i.e. a Celtic or Germanic tribe living in Jutland, Denmark, during the first century common era. Strabo states that the Cimbri had female cultic leaders or prophetess who crowned the prisoners of war with wreaths and then lead them to a vessel. They had raised a platform which the priestess would mount and then, bending over the kettle, would cut the throat of each prisoner after he had been lifted up. From the blood that poured forth into the vessel some of the priestesses would draw prophecy (manteia), while still others would split open the body and from an inspection of the entrails would utter a prophecy of victory for their own people. Geography VII, 2.3–4. The source value of this text has been questioned. It could be a “Greuelmärchen” which circled around Roman soldiers. On the other side the information that the blood was poured into a vessel harmonizes with the use of the hlautbolli in Snorri’s text, i.e. the bowl for sacrificial blood. See Sundqvist 2003b. 342 Chapter 9

Bell devotes one chapter to characteristics of ritual-like activities, such as for- malism, traditionalism, invariance, rule-governance, sacral symbolism and performance.111 In the perspective of present study the last characteristic may be interesting. In recent years, much attention in theoretical discussion has focused on what ritual has in common with theatrical performance, dramatic spectacles, sports and other public events. There are several overlapping features here which all refer to the qualities of performance, such as a communication on multiple sensory levels involving visual imagery, dramatic sounds and gestures, specific clothes, and certain props. The performance dimension of ritual (including feelings, imagery and expressions) also suggests active rather than passive roles for all ritual participants, not only for the cultic leaders. Close to performance is the performative dimension of ritual, which refers to what ritual is able to create, effect, or bring out for the participants. The ritual performance has brought about certain shifts and changes, constructing a new situation and a new reality. By means of the ritual, for instance, a boy is now perceived as a man. Another feature of performance is also that the framing with certain props, costumes and special way of talking etc., indicate that these activities are something other than day-to-day reality, i.e. something set off, which perhaps takes place in a microcosmic portrayal of the macrocosm. In my opinion, the blood in the ancient Scandinavian sacrificial rituals may have had the function to create the dramatic frame, which affected the participators experience of the actions, and made them aware that they witnessed something other than routine reality, i.e. something exceptionally. When the cultic leader at Hofstaðir, on Iceland, swung his broad-bladed axe at the neck of the animal, as suggested by Lucas and McGovern, he produced a blood fountain as the animal’s heart would still be beating (see below).112 This was probably done in public when the cultic community was gathered and it surely must have affected the participators’ experience of the action. The excavators state that this type of violent butchery was unusual and indicates something special, and that the action must be interpreted as something completely set off. Most likely the blood rituals mentioned in Snorri’s account and other sources had the same dramatic function and alluded to sensory aspects of the participators and engaged them emotionally in the ritual drama, for instance the pouring of blood onto the sacred objects and reddening the walls with the sacrificial blood. This was probably done in a dramatic way, with ritual gestures, and certain type of recitations etc.

111 Bell 1997, 72–76, 138–169. 112 Lucas and McGovern 2008. Ceremonial Feasts as Ritual Strategies 343

Perhaps the blood in these rituals also had the performative function to change the situation for the participators, i.e. to construe new conditions and a new reality. Inspired by René Girard’s theory of sacrifice as a form of scapegoating,113 Lucas and McGovern suggest that the main reason why the rituals at Hofstaðir were performed with such unnecessary violence and bloodshed was expiation and the removal of tensions and conflicts within the community. The extreme violent action was thus a way to draw away from interpersonal conflicts and channel it out in other ways. The violent bloody ritual changed conditions for the community, especially in periods of social stress. Since, the archaeological records from Hofstaðir show that these rituals were season based and performed during the summer, I suggest that this calendric ritual rather may be seen as a repetition of the cosmogony made in order to strengthen the cosmic process. According to Snorri’s version in Gylfaginning of the creation myth, the slaying of the primordial giant Ymir made by the divine brothers Óðinn, Vili and Vé in order to create cosmos, involved a bloody slaughter.114 The blood of Ymir killed all giants (i.e. the evil forces) except Bergelmir, by means of drowning. The blood was also used when making the seas and the lakes. This mythic action may have been repeated in the sacrificial contexts as suggested by Bruce Lincoln.115 During the ritual performance the sacrificial blood alluded to Ymir’s blood. The bloodshed implied a definite change of condition for the participators; from now and on the world was renewed and the seasonal changes were guaranteed as well as the crops and general welfare etc. For many people, who did not have access to the religious and cultural knowledge, this symbolic meaning of the blood was not perceived. Perhaps the treatment of blood was only a signal of the sacred act as such. These dramatic acts were executed by representatives of the local elite, who probably gained much prestige by performing these rituals. In connection to these sacrificial rituals the rulers also arranged feasts in their halls, including ceremonial meals as described in Hákonar saga góða. According to this and some other traditions the custom of eating horse meat seems to be crucial there.

113 Girard 1977 and 1986. 114 Gylfaginning 6–7. 115 Lincoln 1986, 41–64. 344 Chapter 9

9.5 Ceremonial Meals and the Custom of Eating Horse Meat

9.5.1 Trøndelag and Norway Hákonar saga góða 18 reports that when King Hákon the Good visited Mære at Yule ( jól), he was forced to eat a few pieces of horse liver. In chapter 17 he was treated by the farmers in a similar way when visiting Lade during the sacrificial feast (blótveizla). There too they forced him to take part in the ceremonial meals, including the consumption of horse meat. This tradition is attested in older texts, but in other versions.116 Ágrip (AD 1190), for instance, reports in much more sparse terms thus: “It is said that he [Hákon the Good] had to bite the horse liver, but wrapped it in cloth so that he would not bite it directly. He would worship in no other way . . .”117 Fagrskinna also renders this tradition briefly, but without mentioning the horse liver.118 The theme of killing horses in sacrificial contexts actually appears in the general description of hof buildings in Hákonar saga góða 14. There it is said that all kinds of livestock were killed in connection with the sacrificial feasts, horses included. It seems reasonable to assume that the sacrificial meat consumed during the feast consisted of horse meat. Snorri also returns to this theme in Óláfs saga Helga 107 when describing the sacrificial feasts held in Inn-Trøndelag in the early eleventh century. It says there that King Óláfr Haraldsson was told that cattle and horses (væri drepit naut ok hross) had been slaughtered and the platforms reddened with their blood.119 Most likely Snorri was well-informed about the pre-Christian ritual slaughter of horses, and the eating of horse meat during the sacrificial feasts in Trøndelag. Almost all early Norwegian Christian laws prohibited the custom of eating horse flesh.120 This was probably connected to the association of horse meat with pre-Christian sacrificial rituals. According to Den ældre Gulathings-Lov, eating of horse meat was to be punished by a fine of three marks, unless it took place during Lent in which case the punishment was permanent outlawry.121

116 See Jón Hnefill Aðalsteinsson 1992; 1998, 57ff. On the medieval Christian attitudes concerning the consumption of horse meat and prohibitions against these customs, see Sanmark 2004, 216ff. 117 Ágrip, Ísl. Fornr. 29, 8. The text is quoted above. 118 Fagrskinna, Ísl. Fornr. 29, 80. The text is quoted above. 119 Heimskringla II, Ísl. Fornr. 27, 177. 120 See Sanmark 2004, 222. 121 GuL 20. NGL 1, 11f. The prohibition on eating horse flesh in GuL appeared in a chapter that forbade the eating of meat during the seasonal fast. Anders Hultgård (1996, 42) has argued that the fact that horse meat was forbidden at all times of year suggests that horses previously were consumed in ritual contexts. Ceremonial Feasts as Ritual Strategies 345

A similar punishment also appears in the Borgarthings-Lov. Anyone eating horse, dog or cat was punished with permanent outlawry. The law made an exception for people who were lost in the wilderness, without food for seven days.122 The regulations prescribed in the Frostathings-Lov indicate that similar rules also existed in Trøndelag.123 The prescriptions stipulated in the Borgarthings-Lov are interesting, since they clearly indicate that eating these animals (horse, dog or cat) was regarded as pagan practice. The law states thus:

. . . when someone eats one of these [animals] without necessity then he has forfeited fortune and peace, land and property, and must leave for a heathen land and never return to where Christians are.124

The phrase “to leave for a heathen land” ( fare a land hæiðit) is significant. It also appears in the outlawry clauses regarding those who refused to baptize a child, neglected to pay tithe, or performed illegal marriages and divorces. All these practices were thus not in keeping with Christian customs.125 Archaeology supports the idea that pre-Christian ceremonial feasts, including meals with meat, took place in Trøndelag at cultic sites with halls and cultic houses. Sometimes we may assume that rulers were involved in these rituals. At the Late Iron Age houses at Mære, where the gold foil plaques were also found, other finds were made which may be associated with pre-Christian ceremonial feasts in aristocratic settings. For instance, pieces of pottery and glass were discovered as well as quantities of animal bones (including horse bones) and some characteristic cooking stones mixed with charcoal. The cooking stones and burnt debris seem to have been connected to the latter Viking Age building, which Lidén interpreted as a pre-Christian hof.126 About the same time as the gold foils were deposited at this building, it is highly likely that ceremonial meals were also prepared and celebrated there. At the chieftain farm of Hove in Åsen, Ut-Trøndelag, there is a small ridge called Hovshaugen, where several finds were made indicating that pre-Chris- tian cultic meals were celebrated there over a long period.127 Several pits with

122 BL I:5. NGL 1, 341f. 123 FL II:43. NGL 1, 145. See also EL I:29. NGL 1, 384. 124 En ef maðr etr æinhuern þænn lut nauðsynia laust þa hæfir han firergort fe oc friði lande oc lausum æyri. fare a land hæiðit oc kome aldri þær sem krisnir menn ero. BL I:5. NGL 1, 342. 125 See further Sanmark 2004, 222f. 126 Lidén 1969, 15; 1999. 127 Farbregd 1986. Cf. Nordeide 2010, 104ff.; 2012. 346 Chapter 9

charcoal, burned stones, and animal bones (sheep and goats) were found there. It was suggested that these pits were used for ritual cooking and perhaps sacrifices. Perhaps these meals sometimes took place at some buildings of the farm Hove. This cultic site was used for a very long time, from c. 100 BC to AD 1000. Also at the central farm Melhus, in Gauldœla fylki, we find traces of ritual meals in the context of rulers and settlements. A great many cooking pits were discovered at the churchyard of Melhus, with dates stretching over a long period. The oldest pits may be dated to Early Iron Age, but most of them were dated to the period between 900 and 1300. Their function is debated. Bones were found in some of the pits, for instance remains of pigs but also of other domesticated animals. The archaeologist Anne Stalsberg suggests that many of these structures are remains from the building of the stone church. The pre-Christian remains situated on the plain below the yard, however, may reflect an outdoor gathering place of both political and religious significance.128 Ceremonial feasts at aristocratic halls may be attested in northern Norway. According to Gerd Stamsø Munch, the hearths in the hall-room C (I:1a) of Borg, Lofoten, were used for ritual baking of the slaughtered animal meat (see ch. 6 above). These meals were probably an important element in the construction of authority at the ceremonial building of Borg.

9.5.2 Horse and Cattle Sacrifices and Ceremonial Meals in Iceland The custom of celebrating cultic meals in the context of rulers is actually well- known from all three investigation areas. According to trustworthy Old Norse traditions, horses were sacrificed and consumed during ceremonial meals in Iceland during the tenth century: Ari inn fróði mentions in Íslendingabók (1122–32) that people during the transition period were allowed to follow the ancient law about consuming horse meat:

. . . regarding the exposure of children and the consumption of horse flesh the old laws were to be valid (skyldu standa en fornu lǫg). People would be allowed to sacrifice in secret if they wished to, but hit by the “lesser outlawry” if this was made known by witnesses. After a few winters, however, this kind of heathendom was removed, just like the other [kind].129

This information about consuming horse flesh is probably directed at customs which occurred in connection with pre-Christian sacrifices. Most likely

128 Stalsberg in Røskaft 2003, 99, note 72. 129 Íslendingabók, Ísl. Fornr. 1, 17. My trans. Ceremonial Feasts as Ritual Strategies 347 this information is reliable, since Ari said that he personally retrieved it from “old intelligent people”. These people may have lived with one foot in paganism and one in Christianity.130 The investigation at Hofstaðir indicates that the ritual slaughter of cattle had a seasonal basis (see above). The evidence suggests recurring ritual activities over a period of years, and not a single mass killing event.131 During these calendrical feasts a lot of meat was consumed in the hall. Perhaps the richness of these feasts was an expression of power claims. Lucas describes this phenomenon as “competitive feasts”, “occasions where the intention is to display or show off the success of a household or community”.132 The aim may have been both to strengthen the solidarity within the group and to acquire status. The skulls hung out in the open air on the walls/roof may be regarded as “trophy displays” from these grandiose feasts. Such feasts could be seen as a strategy for rulership in Viking Age Iceland

9.5.3 Sacrificial Meals in Eastern Scandinavia As noted above, ceremonial meals, including horse meat, are also attested in reliable sources from eastern Scandinavia. In the Guta Law (in the section “Af blotan”) invocations with food and drinking were prohibited when they did not follow Christian customs.133 Guta Saga relates that sacrifices (Old Gutnish blotan) were performed in more ancient periods. People believed in groves and grave howes, holy and ancient sites, as well as in heathen images. The whole island held the highest sacrifice on its own account: “But smaller assemblies held a lesser sacrifice with cattle, food, and drink. Those involved were called ‘boiling companions’, because they all cooked their sacrificial meals together.”134 The Old Gutnish appellation suþnautr ‘comrade-in-sacrifice’ and the verb siauþa ‘cook, boil’ belong to an ancient Germanic sacrificial terminology and should be related to the Gothic word sauþs ‘sacrifice’ (cf. ON seyðir ‘cooking-fire, cooking-place, cooking-pit’). The word fileþi is usually translated ‘cattle, beast’, but perhaps this term also connoted ‘horse’.135 The description of sacrifices in the Guta Saga thus constitutes a close parallel to the rituals mentioned in

130 Cf. Dillmann 1997, 70, note 28. 131 See mainly Lucas & McGovern 2008; Lucas 2009 and ch. 4 above. 132 Lucas 2009, 404. 133 . . . haizl nequara þa miþ mati eþa miþ dryckju senni sum ai fylgir cristnum siþi. Guta Lag, (ed.) Pipping, 6–7. 134 . . . En smeri þing hafþu mindri blotan miþ fileþi, mati ok mungati, sum haita suþnautar, þy et þair suþu allir saman. Guta saga (ed.) Peel, 4f. 135 Guta Saga, Peel 1999, 29. 348 Chapter 9

Hákonar saga góða. The composition of this saga is usually dated to sometime between 1220 and 1275.136 Viking Age burial customs were performed in Gotland until the thirteenth century. The archaeologist Lena Thunmark-Nylén states: “Not until about the time of the Guta Law (c. 1220), when the last pagan customs were strictly forbidden and criminalized, do we lose contact with the remaining non-Christians.”137 This statement hints that the information in the Guta Law and the Guta saga may build on reliable traditions.

9.5.4 Horse Sacrifices and Ritual Meals in the Royal Milieus at Uppsala Adam of Bremen intimates that sacrificial meals were also celebrated at Uppsala during the Viking Age. The term commessationes (sg. commesatio) probably refers to such meals.138 It is also interesting to note that Adam mentions that horses were sacrificed in Uppsala. When describing the sacrificial grove he states that “even dogs and horses hang there with men” (Ibi etiam canes et equi pendent cum hominibus).139 Even if the information about human sacrifices is regarded as an exaggeration by Adam, made in order to demonize the old religion in Uppsala,140 the idea about horse sacrifices may be true. It harmonizes well with the information we have about horse sacrifices at Mære mentioned in Ágrip and Heimskringla. Perhaps horse meat was also consumed during the ceremonial feasts at Uppsala, as in Lade and Mære. Several other sources report that horses were used as sacrificial animals in Svetjud, especially in aristocratic contexts. The U-version of the Hervarar saga has already been mentioned. It narrates that Blótsveinn had to participate in a horse sacrifice and eat horse meat when he accepted the kingship of Svear. The excavations at Uppsala and its surroundings have also uncovered traces of ceremonial practices involving horses: cremated fragments of horses were discovered, among other animal bones, in Uppsala’s Eastern and Western Mounds (from the early and second parts of the sixth century respectively).141 A horse has been found south of Prästgården, located together with boat grave 1, dated to the ninth century.142 A lot of bones, connected with the Viking Age settlement area at Norra gärdet, just north of Old Uppsala church, were also discovered.

136 Wessén 1960, 602–604. Guta Saga, Peel 1999, xlix–liii. 137 Thunmark-Nylén 1989, 232. 138 Adam IV, 27 and scholion 141. 139 Adam IV, 27. 140 Hultgård 1997. 141 See Lindqvist 1936, 197 and 200. On horse rituals and ceremonies in eastern Scandi- navia, see for instance Petré 1984, 117–118; Sundkvist 2001, 66–69; Sundqvist 2002, 225–40. 142 Nordahl 2001, 23–24. Ceremonial Feasts as Ritual Strategies 349

Some were horse bones, but bones from cattle, sheep, pigs, and even frogs were also found. The archaeological and osteological analyses support the hypothesis that some of these bones were house offerings: in House 1 (post-hole A32), for instance, a dog jaw bone, a horse jaw bone, and other bones from a horse were found and interpreted as foundation sacrifices.143 Deposited horse bones have also been found in association with the row of posts (see above). During the excavations in the summer 2012 a lot of sacrificial bones were also found in the eastern part of Old Uppsala (see ch. 4 above). Sacrifices of horses, possibly from the Late Iron Age, have been discovered at several places in eastern Scandinavia, some of which appear to reflect aristocratic milieus.144 Horse bones found in a water-hole outside the ring fort at Eketorp in Öland, have been associated with offerings connected to communal meals.145 At this site there was also a ceremonial building, dating to a period between the fifth and the seventh centuries.146 A similar cult including ceremonial meals probably existed at Skedemosse, in Öland where horse bones, gold hoards, and other discoveries have been made and dated to the Middle and Late Roman Iron Age.147

9.5.5 Sacrificial Feasts at the Aristocratic Hall of Helgö Indications of sacrificial feasts may also be found outside the aristocratic hall of Helgö (Late Iron Age). South of the hall at Foundation II a system of hearths and cooking pits has been found. It has been argued that these hearths were not related to ordinary cooking or craft,148 but rather to ritual handling of animals in connection with sacrificial ceremonies. One of the pits was huge, almost 6 by 3 m wide and 1 m deep and filled with burnt stones, sherds of pottery, and animal bones. It has been interpreted as an ON seyðir, i.e. a cooking pit intended for ritual preparation of meat.149 There is also an interesting triangular stone setting or tricorn (Sw treudd) located just east of the hall on a stony ledge (called Foundation IV) at Helgö (see fig. 12A).150 This stone setting is dated to the Early Viking Age, placed on top of a rich culture layer that was up to 1 m

143 Ljungkvist et al. 2000, 44–51; Carlie 2004, 127–129. 144 Hagberg 1967–77. 145 Backe et al. 1993, 327–342; Götherström 2002. 146 Herschend 1998, 184. 147 Hagberg 1967. Cattle and sheep and humans have been deposited continuously in the lake Skedemosse starting around 300 BC and ending with the skeleton of an old woman dated to AD 998–1118 and that of a horse, a mare from the same time. Horses have been part of the rituals from the year AD 1. Monikander 2010, 94f. 148 Arrhenius 1961, 96f. 149 Wernersson 1996; Vikstrand 2001, 242f. 150 T. Zachrisson 2004a; 2004b; 2010; Hansson 2011; Frölich 2011. 350 Chapter 9 thick and clearly stratified. The layer included all kinds of deposited objects, such as surgical instruments, crucibles, bread, arrowheads, whetstones, birch- bark containers, ice nails, pottery, flint and an Arabic silver coin. Sherds from a number of vessels were also found at the south end of the setting, beside a bare rock. The sherds included an amphora of Rhineland origin. Different types of saddle and rotary querns were also found there, as well as baking ovens and bread. They were also related to pre-Christian sacrificial rituals. On the stony ledge some amulets were discovered as well, such as miniature spears, Þórr’s- hammer pendants, fire-steel pendants, miniature shears and amulet rings, usually considered as symbols of ancient Scandinavian deities. In my opinion, the sacrificial rituals performed at the stony ledge must be related to the ritual activities that took place in the hall (I A), such as the communal meals and banquets. The sacrificial rituals performed at Helgö were probably very complex in nature, and were performed at several places. The ritual slaughter and the handing over of victims to the gods may have taken place outdoors at the rock and the ledge, while the communal meal was celebrated in the hall. Without doubt it was the hall owner, i.e. the ruler of Helgö, who played a central role in the rituals that took place both inside and outside the hall building. Not only the drinking vessels but also the great amount of animal bones (calf, pig and hare), and all the finds under the stone setting (such as the bread) may reflect the complicated rituals which were performed at this sacred site during the calendrical feasts and on other occasions. In connection with these ceremonies there may have been a team of cultic leaders taking part in the preparation of the sacrificial feast. Most likely the ruler or chieftain was the supervisor of these actions. His leading ritual role was emphasized during the ceremonial feast when he sat in the high-seat inside the hall. Helgö thus constitutes an extraordinarily good example of sacrifices, banquets and sacrificial meals performed under the surveillance of the ruler, in the context of a ceremonial building.

9.5.6 Lunda At the large Iron Age hall building at Lunda there are also finds which may be associated with drinking rituals and ceremonial meals in an elite milieu, such as broken glass originating from Frankish beakers. As at Helgö and Birka, the Lunda hall had a hearth which may have played a certain role in ambulatory rituals and ceremonies placing the hall owner (i.e. the ruler) in a central position, close to the assumed high-seat. Most likely the ruler of Lunda sat in this high-seat when celebrating the ceremonial feast. The high-seat in this hall also emphasized the central position of the ruler during the rituals. Ceremonial Feasts as Ritual Strategies 351

As at Helgö, ritual activities took place both inside and outside the hall at Lunda. In the yard north of the hall there were several cooking pits, most likely intended for cooking ritual meals. In one huge cooking pit (3 × 2 m in area and 1.1 m deep) unburned bones of pigs were found. The archaeologist Eva Skyllberg relates these bones to the god Freyr.151 In my opinion, she might be right. In the Old Norse mythical traditions it seems as if Freyr had a special relation to the pigs. One of his attributes was the pig called Gullinbursti.152 It was also close to the pit with pig bones that the three phallic images were discovered. Most likely they represent Freyr. Both pigs and Freyr may also be related in different senses to the Yule feast ( jól). That Freyr was connected to the Yule feast is well-established in research.153 Economic arguments may also relate pigs to this winter feast. Pigs rummage in the earth during the autumn, looking for beechnuts, truffles, roots and other things. When there is frost in the ground it is time to slaughter them. This is the reason why the traditional swine slaughter took place in late autumn or early winter in Sweden. In my opinion, there are thus good reasons to believe that religious feasts were celebrated at the cultic hall of Lunda, perhaps during the winters. The great hall building, the exclusive ceremonial glass and the consumption of pig indicate that such feasts were controlled by the ruling power. Most likely these meals were a way to construct authority for the political power at Lunda.

9.6 Ritual Drinking and the Reciting of Ritual Formulae

In Hákonar saga góða, Snorri describes the pre-Christian drinking rites ( full, minni) connected with the sacrificial feasts at the hof of Lade. In my opinion, there is no doubt that such rituals were crucial during the pre-Christian religious meetings in the ceremonial buildings of Svetjud, Trøndelag and Iceland. That certain religious terms connected to such rituals, such as minni, may not be traced back to pre-Christian times is not a sufficient argument for reject- ing the idea that the phenomenon itself is ancient. Sources of different types clearly indicate that ceremonial drinking and meals were important in the aristocratic halls and cultic buildings in all three investigation areas.

151 Skyllberg 2008, 34. 152 Gylfaginning 48; Skáldskaparmál 7. 153 See Nordberg & Sundqvist 2008. 352 Chapter 9

9.6.1 Norway The compound word blótdrykkja—‘drinking bout at pagan sacrifice’154—seems to refer to ancient drinking rituals in Norway. It is attested in Ágrip, which describes how Óláfr Tryggvason abolished the old custom and replaced it with Christian drinking ceremonies at Christmas and Easter, St John’s Mass, and St Michael’s Mass.155 The expression drekka jól—‘celebrate Yule with drinking- ceremonies’—appears in royal contexts in a praise poem to Haraldr Finehair, Haraldskvæði, composed by Þorbjǫrn hornklofi around 900. The poem reads:

Úti vill jól drekka ef skal einn ráða fylkir enn framlyndi ok Freys leik heyja; ungr leiddizk eldvelli ok inni sitja, varma dyngju eða vǫttu dúns fulla.156

The stanza is preserved in Fagrskinna and in Snorri’s Haralds saga ins hárfagra.157 The source value here is high, but it is uncertain whether Freys leikr should be interpreted as “sport of Freyr” (that is, the “battle”) or as “Freyr’s sacrifice” (compare Old Norse leikr with Old English lāc, ‘play, sport: strife, battle: sacrifice, offering: gift, present: booty: message’).158 Hans Kuhn, who has discussed the kenning in detail, supports a cultic interpretation.159 The combination of the expressions jól drekka and Freys leik heyja, as well as the Old English lāc attested in the sense of ‘sacrifice, offering’, point in that direction. Another plausible interpretation has been offered by Heinrich Beck, arguing that leikr in this context should not be interpreted metaphorically, but abso- lutely, as ‘Waffentanz’ or ‘Waffenspiel’. Representations of dancing warriors on the helmet from Sutton Hoo, for instance, strengthen this interpretation.

154 See Fritzner 1954: ‘drikkelag vid hedensk offring’. 155 Ágrip, Ísl. Fornr. 29, 22. 156 “He will drink his Yule feast in the open, if he alone shall decide, this ambitious prince, and perform the sport of Freyr; when young he tired of basking by the fire, of sitting indoors and the snug women’s bower, and of mittens filled with down.” Text and trans. E. O. G Turville-Petre 1976, 14. Cf. Skj. A1, 25; B1, 23 and SkP I, 99. 157 Fagrskinna, Ísl. Fornr. 29, 61. Heimskringla I, Ísl. Fornr. 26, 112. 158 Hall 1916, 179. See discussion in SkP I, 100. 159 Kuhn 1978, 277–79. Ceremonial Feasts as Ritual Strategies 353

This dance may ultimately derive from cult.160 Anders Hultgård argues in the same vein when he suggests that leikr perhaps referred to plays and rituals connected to the pre-Christian Yule.161 Þorbjǫrn’s stanza could thus be used as reliable evidence, in the context of the practices of rulers, of pre-Christian libation rites. In a mythologized account in Hákonarmál stanza 16, Hákon is invited to Valhǫll to drink beer with the gods (þigg þú at ǫ́sum ǫl), reflecting again the relationship between rulers and ritual drinking.162 Further, ceremonial beer drinking at feasts is attested in saga texts referring to pre-Christian Norway. Egils saga narrates the visit of King Eiríkr blóðøx and his wife Gunnhildr to Atløy for the dísablót.163 The saga describes it as an outstanding banquet with much drinking inside the house (veizla in bezta ok drykkja mikil inni í stofunni). When Egill and his men are finally allowed to enter the house (stofa), Ǫlvir, perhaps a cultic leader,164 is allowed to sit beside the king and the queen in the high-seat. Beer is carried out to drink memory toasts (minni) and for each toast a horn has to be emptied.165 Egill is angry with Bárðr, who at first forbade him and his men to take part in the sacrificial feast. The episode is supported by a stanza composed by Egill, dated to the beginning of the tenth century:

Sǫgðuð, sverriflagða, sumbleklu, herkumbla – því telk – brjótr, þars blétuð – bragðvísan þik – dísir; leynduð alls til illa ókunna hjǫrrunna, ilt hafið bragð of brugðit, Bárøðr, hugar fári.166

160 Beck 1968. 161 Hultgård 2000, 103. Cf. Steinsland 2011b, 37. Critically considered by Fulk in SkP I, 100. 162 Skj. B1, 59. See also the mythical references where the dead drink beer with Óðinn in his high-seat, in ch. 6 above. 163 Egils saga, Ísl. Fornr. 2, 106–111. 164 Sundqvist 2002, 77; 2003a; 2003b; 2007. 165 Síðan var þeim borit ǫl at drekka; fóru minni mǫrg, ok skyldi horn drekka í minni hvert. Egils saga, Ísl. Fornr. 2, 108. Düwel does not accept the “memory cup” of minni in this passage. Düwel 1985, 51–55. 166 “You said, warrior, to your guests, that beer was lacking, when you held sacrifices for the Dísir. I think you are fraudulent, host, you hid your evil nature. The hostility you housed against the men you did not know. Bárðr, you have made a foul play.” Skj. B1, 42f. My trans. based on Finnur Jónsson. Cf. Egils saga, Ísl. Fornr. 2, 108. 354 Chapter 9

This saga text, combined with the tenth-century lausavísa, provides us with reliable evidence of pre-Christian ceremonial drinking connected with the dísablót in Norway. These drinking customs took place in a building called stofa, perhaps a banqueting hall at the royal farm.

9.6.2 Svetjud and Iceland The medieval sources report that ceremonial drinking also occurred in the pre- Christian ceremonial buildings of Svetjud. Often the rulers of the Svear played a significant role during these ceremonies. In these texts it seems as if these types of rituals were not simply a public expression of power, they actually established it. In Ynglinga saga 37, for instance, it is stated that King Granmarr of Södermanland invited King Hjǫrvarðr of Denmark to a ceremonial feast (til veizlu). In the evening it was customary “to drink a full beaker” (drekka . . . full). During the banquet the rulers sat in high-seats opposite one another: “The high-seat of King Hjǫrvarðr was placed opposite that of Granmarr.”167 The text implies that during the drinking rituals everyone present in the drinking hall paid a great deal of attention to the kings. As in the description of the hall in Fagrskinna (see ch. 6 above), the very placement of the high-seats indicates this. Whether this type of drinking ceremonies (drekka . . . full) had religious connotations is uncertain, however. But, there are signs in other sources that drinking rituals occasionally had religious elements and were parts of the sacrificial feasts in eastern Scandinavia. According to Adam of Bremen libation rituals were made to the three images in the ceremonial building of Uppsala (templum):

If plague or famine is nigh, a libation is made to the idol of Thor, if war, to Wodan, if marriage is to be solemnized, to Fricco.

In this quotation the Latin verb libo (libare) must be interpreted in its more precise meaning, i.e. ‘to perform a libation sacrifice to the deity’.168 It thus seems as if the worship of the gods in the banqueting room (triclinium) was intimately associated with the drinking rituals, perhaps with the ruler sitting in his high-seat. Also in connection with the “unseemly” songs that were sung at these feasts, “libations” (libatio sg.) were performed: “For the rest, the trivial songs that they use in this sort of libation rites are manifold and shocking,

167 Hásæti Hjǫrvarðs konungs var búit gagnvart hásæti Granmars konungs . . . Heimskringla I, Ísl. Fornr. 26, 68. 168 For this interpretation, see Dillmann 1997, 66. Ceremonial Feasts as Ritual Strategies 355 so it is best to keep quite about them.” It seems likely that these ritual songs had sexual connotations, since Adam disapproved of them so much. Perhaps the “unseemly songs” and the libation sacrifices in the banqueting room of Uppsala were related to a fertility cult.169 Drinking customs in ceremonial buildings may be attested by archaeological finds in the Mälaren area. Large amounts of glass have been discovered at house group 2 at Helgö, for instance.170 No less than 96 percent of the 1,600 glass fragments found on the island were discovered at this house group. Interestingly, there was a particular concentration in the ceremonial hall I A at Foundation I, more specifically at the place interpreted as the “high-seat”. The glass was clearly high-status and comprised, for instance, fragments of filigree glass claw and cone beakers. It has been estimated that these glass fragments belonged to at least 50 beakers and bowls of different types dated to the period from the Roman Iron Age to the Viking Age. It cannot be excluded that these beakers were used for ceremonial purposes, for instance in connection with ritual banquets.171 The warrior hall discovered in Birka yielded finds similar to those in the hall at Helgö. In the western part of that house broken glass from exclusive beakers was found. Most likely the high-seat was located at this place, where the king sat during his visits to Birka (see ch. 6 above). Prohibitions in normative Christian texts from eastern Scandinavia directed against pagan customs and ideas may also give some support to the idea that libations were essential at pre-Christian sacrificial feasts. In the medieval Guta Law (in the section “Af blotan”) we read that invocations with food and drinking were prohibited when they did not follow Christian customs.172 These pagan customs may also be related to some information in the medieval Guta saga. The Old Gutnish term mungat ‘ale, beer, feast’ refers to a specially-brewed ale for times of celebration, such as religious feasts. Perhaps it was also drunk on

169 There is more evidence of this in Gotland, where smaller assemblies held lesser sacrifice with cattle, food and drink. Guta saga, (ed.) Peel, 4f. In this text these ceremonies are described with the Old Gutnish noun blotan (f.) ‘sacrifice’ (cf. ON blót ‘pagan worship, sacrifice’). The one who performs such rituals is called sauþnautr ‘person with whom one boils meat, comrade-in-sacrifice, boiling-companion’ (cf. siauþa ‘cook, boil’; cf. Gothic sauþs ‘sacrifice’). 170 Lund Hansen 2011. 171 Holmqvist 1980, 56–65; T. Zachrisson 2004a; 2004b; 2010; Lund Hansen 2011. 172 Þet ier nu þy nest et blotir iru man num mier firi buþin Oc fyrnsca all þaim sum haiþnu fylgir Engin ma haita a huathci a hult eþa hauga. eþa haþin guþ. huatki a vi eþa staf garþa . . . nequara þa miþ mati eþa miþ dryckiu senni sum ai fylgir cristnum siþi. . . . Gutalagen, af blotan, (ed.) Pipping: 6−7. On these texts, see T. Blomkvist 2002. On religious drinking rituals, see Drobin 1991; Nordberg 2003, 175ff. 356 Chapter 9

Gotland in connection with the ritual dedications to the deities, for a good harvest, for the dead and so on, as described in Hákonar saga góða. Sources narrate that ritual drinking also occurred in Iceland during the calendrical feasts. Such rituals were often arranged by the local chieftains. Gísla saga Súrssonar 15, for instance, reports that Þorgrímr decided to hold a feast at the end of autumn to celebrate the coming of the winter-nights (at vetrnóttum ok fagna vetri). During this feast sacrifices to Freyr were made (blóta Frey) and there was a lot of drink to be served (drykkja skyldi vera . . .). At the same time Gísli held a great feast at his farm, with copious drinking.

9.6.3 Blóta til árs ok friðar According to Snorri’s Hákonar saga góða, certain formulae were recited in connection with the ritual toasts during the calendrical feasts in the hof sanctuary of Trøndelag. In chapter 14 he mentions that Óðinn’s toast was to be drunk first, then Njǫrðr’s toast and Freyr’s toast, for good harvest and peace (drekka . . . Njarðar full ok Freys full til árs ok friðar). It appears that the ruler himself had to perform some of the drinking rituals and perhaps also recite the formula. Actually, the farmers in Trøndelag expected the king to perform the sacrifices in order to gain “good harvest and peace” (til árs ok friðar) for his people at the assembly of Frostaþing: “The farmers said that it was their wish that the king should make sacrifice to procure them (blóti til árs þeim) good crops and peace, as his father had done.”173 Fagrskinna reports on the same event in Trøndelag: “that he should sacrifice as was the custom for previous kings and in that sense ratify ancient laws for good crops and peace (til árs ok friðar).”174 In past decades, the expression ár ok friðr ‘prosperity and peace’ received much attention in research. For a long time it was regarded as a genuine pre- Christian formula. Vilhelm Grønbech, for instance, argued that it was one of the ritual expressions recited during the pre-Christian ceremonial meals.175 Elias Wessén suggested that it was related to the cult of Freyr.176 Recently, however, several scholars have argued that it has no pre-Christian background. Wolfgang Lange, for instance, stated that: “friðr ist zwar ein altes Wort, aber die ältesten Belege für die Formel stehen ausnahmslos erst in der christlichen

173 Bændr segja, at þeir vilja, at konungr blóti til árs þeim ok friðar, svá sem faðir hans gerði. Heimskringla I, Ísl. Fornr. 26, 170. 174 . . . skyldi blóta eptir vanða enna fyrri konunga ok fylla svá en fornu lǫg til árs ok friðar. Ísl. Fornr. 29, 80. 175 See Grønbech 1997 (1912). 176 Wessén 1924, 177ff. See also Baetke 1942, 3–39; F. Ström 1954; 1985; de Vries 1956–57, §§ 149, 184, 292, 461. Ceremonial Feasts as Ritual Strategies 357

Dichtung.”177 Klaus Düwel is also sceptical of the pre-Christian origin of this expression.178 Klaus von See regards ár ok friðr as Christian: “Sehr wahrschein- lich stammt sie aus der christlichen Missionssprache’.179 Anders Hultgård has considered whether the expression is rooted in medieval Christian literature (such as the Roman and the Eastern Churches’ early liturgies, homiletics, hagiographical writings) and concluded that there are no prototypes for this formula in medieval Christian literature: “Es läßt sich also feststellen, daß die Formeln mit ár ok friðr Vorstellungen ausdrücken, die kon- genial mit der alt-skandinavischen Religion sind.”180 Hultgård, who suggests that the medieval Church adopted the expression from pre-Christian religion, has gained support from other scholars.181 In a more recently published essay Hultgård has also found similarities between ON ár and the cultic expression in ancient Iranian traditions (Avest. yāiriia- in e.g. Yasht 8), indicating that this expression might belong to an Indo-European heritage.182 In my opinion, Hultgård’s arguments are well-founded. The expression ár ok friðr is probably pre-Christian. It appears in a clear pagan cultic context in both Hákonar saga góða and Fagrskinna. In many other sources this formula appears in the context of pre-Christian rulers. One of the best pieces of evidence of this from eastern Scandinavia is the Stentoften inscription (DR 357), lines 1–3, which was quoted at the beginning of this chapter. The ideographic rune called jāra ‘good year crops’ found in line 3 is usually interpreted as equivalent to Old Norse ár. The Stentoften inscription may thus support the argument that Late Iron Age rulers performed sacrifices in order to ensure a good crop (ár) for their peoples.183 The expression ár ok friðr was most likely taken over by the Church in some parts of Scandinavia. In Norway, for instance, Den ældre Gulathings-Lov states that when celebrating All Saints’ Day and Christmas, the beer should be consecrated “in order to thank Christ and St Mary. For prosperity and peace” (til Krist þakka. oc sancta Mariu. til árs. oc til friðar).184 This law is introduced in the following words:

177 Lange 1958, 119. 178 Düwel 1985, 61–69 179 von See 1988, 84–87. 180 Hultgård 1993, 251; 2003. 181 See Grønvik 1996, 169; Dillmann 1997, 58f. 182 Hultgård 2003, 304–306. 183 Hultgård 1993; 2003; Sundqvist 1997, 135–174; Düwel 2008, 21f. 184 NGL 1, 6. 358 Chapter 9

This is the beginning of our Law, we shall turn to the east and pray to the Holy Christ [that he may procure] good crops and peace.185

In the Frostathings-Lov, fast days are dedicated “for prosperity and peace and for good health to all men”.186 This instance produces solid evidence that the formula was used in Trøndelag during the transition period. Other evidence from the transition period in Norway indicates that ár ok friðr was ancient and taken over by the Church. Glælognskviða, composed by Þórarinn loftunga in the first decades of the eleventh century, says of St Óláfr:

hann of getr af goði sjǫlfum ár ok frið ǫllum mǫnnum.187

The ancient expression has most probably been set in a new Christian context by the skald. The priest and skald, Sigvatr Þórðarson, designates King Óláfr jǫfurr ársæll in the poem Knútsdrápa 3.188 This stanza was quoted in Ágrip 31. The scribe of Ágrip noted that when King Sveinn and his mother Alfífa took over the reign of Trondheim, famine and misery prevailed: aldregi var ár á þeira dǫgum.189 The expression ár ok friðr also appears in medieval sources that derive from eastern Scandinavia: Gutalagen, which was barely influenced by Old Norse tradition, says: “he [God] grants us prosperity and peace, victory and health” (hann unni os ar oc friþ sigr oc hailsu)—and it is probable that Christianity in Gotland adopted the pre-Christian formula.190

The expression ár ok friðr was thus probably known in all three investigation areas.191 Most likely it has a pre-Christian origin but during the transition period was taken over by the Church. According to Hákonar saga góða, this

185 Þat er upphaf laga varra at ver scolom luta austr oc biðia til hins helga Crist árs oc friðar. NGL 1, 3. 186 . . . til árs oc friðar oc til heillsu allum monnum . . . NGL 1, 141. 187 “He, by himself, receives from God, good year and peace for all men.” Skj B1, 301. My trans. This is actually the earliest attestation of the formula in the Old Norse traditions. 188 Skj B1, 232. 189 Ágrip, Ísl. Fornr. 29, 31. 190 Gutalagen,(ed.) Pipping, 3. 191 It is superfluous to mention that the formula ár ok friðr was known in medieval Iceland. All references to it in the Old Norse sources suggest this. Ceremonial Feasts as Ritual Strategies 359 formula was recited in connection with the ritual toasts during the calendrical feasts in the ceremonial building. In this text it appears that the king himself had to perform some of the drinking rituals and perhaps also recite this formula. The king had to sacrifice in order to produce “good harvest and peace” for his people. It seems as if the concept of ár ‘good year’s crop’ is often associated with rulers, at least in Sweden and in Norway.192 Sources suggest that the pagan ruler, by means of sacrifices, was able to produce such good conditions. The formula was thus an important ritual means for creating authority in ceremonial buildings.

9.7 Gift-Giving Ceremonies

The sources thus indicate that rulers held sacrificial feasts including ritual drinking in the halls in all three investigation areas. The sources also suggest that the rulers during these banquets had the opportunity to demonstrate their wealth and to display their generosity in a concrete sense. According to Hákonar saga góða 14, for instance, Sigurðr Hlaðajarl is described as very generous to the farmers, since he paid all the expenses for the cultic feast by himself. Snorri supports his statements here with a stanza taken from Kormákr Ǫgmundarson’s Sigurðardrápa (960 AD). In this stanza it seems as if the feast was held at a sanctuary (vé), which was supervised by Earl Sigurðr. It is introduced thus: “You do not need to bring with you/ neither basket nor tankard/ to the generous man.” In the Sagas of Icelanders we see a similar attitude, where chieftains invited their subjects to the calendrical ceremonial feasts in order to show their generosity and splendour. Gísla saga Súrssonar 15, for instance, reports that Þorgrímr decided to hold a feast at the end of autumn to celebrate the coming of the winter-nights. During this feast sacrifices were made to Freyr and the hall was decorated with tapestries, while the floor was strewn with rushes. The text mentions that Þorgrímr invited his brother Bǫrkr and Eyjólfr Þorðarson to this feast, but also “many other men of distinction”. There were one hundred and twenty men in all and a lot of drink was served. It thus seems as if wealth and generosity were key concepts for the Viking Age rulers when constructing authority. In the Sagas of Icelanders, individual Norwegian-Icelandic chieftains are therefore often described as auðmaðr ‘a rich man’ (cf. auðigr maðr), rausnarmaðr ‘a generous man’ or stórlátrmaðr

192 Cf. Hultgård 2003. 360 Chapter 9

‘a generous man’.193 Distributing gifts and holding great banquets were thus important means for rulers to tie subjects to themselves all over Scandinavia. The notions of this gift-giving system are expressed in the Eddic lay Hávamál 41–42, for instance, where we read:194

Vápnom oc váðom scolo vinir gleðiaz, þat er á siálfom sýnst; viðrgefendr oc endrgefendr erost lengst vinir, ef þat bíðr at verða vel. Vin sínom scal maðr vinr vera oc gialda giof við giof . . .;195

The historian Jón Viðar Sigurðsson has studied this system in relation to chieftains in Iceland. He argues that gifts were used to establish or renew ties of friendship or loyalty. It was a reciprocal redistribution system. One could only accept a gift if one was able to give another gift in return. Otherwise one did not have the right to own or use the gift. Jón Viðar concludes: “Because of the strong obligation to reciprocate, gifts were a good instrument by which the various types of chieftain could bind supporters to them.”196 In Iceland the gifts consisted mainly of weapons, horses, oxen, jewellery, clothes and fabrics.197 The feasts in the hall, however, served virtually the same purpose as the gifts. In many cases the chieftains held great feasts and invited

193 Laxdæla saga, for instance, says about the chieftain Ingjaldr Sauðeyjargoði: “He was a wealthy man who liked to throw his weight around” (. . . hann var auðigr maðr ok mikill fyrir sér). Laxdæla saga 14, Ísl. Fornr. 5, 28. Þórólfr Mostrarskegg, according to Eyrbyggja saga, ran his farm in fine style (as a rich and generous man) and had a large number of men with him (Þórólfr gerðisk rausnarmaðr mikill í búi ok hafði fjǫlmennt með sér, . . .). Eyrbyggja saga 4, Ísl. Fornr. 4, 10. Þorsteinn Ingimundarson from Hof was generous to his neighbours with the goods from his estates (Þorsteinn frá Hofi var stórlátr af búi sínu við heraðsmenn; . . .) Vatnsdæla saga 31, Ísl. Fornr. 8, 84. 194 The classical study of gift-giving systems in archaic societies was written by Marcel Mauss (1997). Mauss introduced his study by quoting Hávamál. For such a system in Iceland, see Jón Viðar Sigurðsson 1999. This system could perhaps also be related to the so called patron-and-client-relationship (often called clientelism) mentioned in ch. 5 above, and more known from the classical areas, see e.g. Jackson 2014b. 195 “With weapons and gifts friends should gladden/ one another, that is most obvious;/ mutual givers and receivers are friends for longest,/ if the friendship is going to work at all./ To his friends a man should be a friend/ and repay gifts with gifts; . . .” Trans. Larrington. 196 Jón Viðar Sigurðsson 1999, 91; 2010, 25ff.; 2011, 77f. 197 Jón Viðar Sigurðsson 1999, 92; 2010, 25ff.; 2011, 77f. Ceremonial Feasts as Ritual Strategies 361 a lot of people. At the end of these feasts they sometimes distributed gifts to the participants, in order to enhance their reputation. Thus both feasts and gifts were means to construct authority and bind people to certain interests of the rulers. Such gift-giving could also be practised as a kind of payment to loyal subjects and could be performed in ritual forms in the hall. Egils saga 55 narrates that after helping King Aðalsteinn to defeat King Óláfr (rauði), Earl Aðils and the Scots in England, Egill Skalla-Grímsson came to the hall of the king when he was sitting and drinking, and celebrating his victory. The king took a gold ring from his arm, put it on the point of his sword, went to the floor and stretched the sword over the fire to Egill. This episode is supported by a stanza made by Egill himself (see further ch. 11). In the Anglo-Saxon poem Beowulf, the rulers distributed rich gifts to their retainers in the hall during the feasts. Sometimes the lord was therefore called OE bēag-gifa (-gyfa) “ring-giver”. He is also called OE manna mildust “the most generous of men”.198 Occasionally the noble females also took part in these gift-giving ceremonies (see below). Among the Franks too, the rulers were expected to distribute gifts from the booty. This is attested by an account recorded by Gregory of Tours. When Clovis was still a pagan he plundered churches, and all the objects from the booty were later distributed among his retainers in Soissons.199 These references indicate old systems of redistribution in several parts of ancient Europe, where the ruler tied loyal people to himself by means of gifts. There are no equivalent sources from eastern Scandinavia describing the gifts and drinking ceremonies in the hall as exhaustively as in Egils saga and Beowulf. However, skaldic poems, runic inscriptions and archaeological finds may give some indications. In Ynglingatal the Svea king Vanlandi is called menglǫtuðr ‘spendthrift of jewellery’ or ‘generous person’, while King Agni is called gœðingr ‘magnate’ or ‘a man with a lot of property’. It is possible that such epithets were ascribed to rich rulers who had manifested their wealth in their halls and generously distributed gifts to their retainers. It must be mentioned, though, that Ynglingatal was composed by a Norwegian skald. These attributes of kings may thus have nothing to do with eastern Scandinavian conditions. However, runic inscriptions from eastern Scandinavia also indicate that rulers had to be generous and care for their retainers. According to Santesson’s interpretation of the seventh-century runic inscription from Stentoften, Haþuwulfr arranged a great sacrificial feast, perhaps with plenty of meat for all the people (see above). This event was extraordinary in several senses and was therefore carved into the runic monument. Similar attitudes

198 Cf. Gustavson et al. 1992, 166. 199 Gregory of Tours, Historiae Francorum, II, 27. (Ed.) Arndt & Krusch, 88f. 362 Chapter 9 also appear in Viking Age inscriptions from the Mälaren area.200 The Turinge stone inscription (Sö 338) tells us that the chieftain Thorstein and his brother “treated their retainers well” (heldu sina huskarla vel). According to this inscription Thorstein was a landmaðr, i.e. some kind of chieftain who also had a lot of landed property.201 In the inscription on the Väppeby stone (U 703), at Veckholm, the dead man is praised for being generous with food and eloquent (mandr matar goðr ok malsrisinn). Most likely he was eloquent when he appeared in his hall. Similar information appears in the inscription on the Gådi stone (U 739) where Holmbjǫrn praises himself for being “generous with food and eloquent”. Qualities of this type are also praised on the runic stone from Hagstugan (Sö 130) where the magnate Domare is honoured for being mildan orða ok matar goðan. The sources thus indicate that the rulers of the Mälaren region—chieftains or kings—during the banquets had the opportunity to demonstrate their generosity and to create alliances and friendship, by giving their men food and gifts, probably in ritualized forms in the hall. Evidence of this system may also be seen in a late Viking Age runic inscription on a bone discovered in the block called Trädgårdsmästaren in Sigtuna:

kunukr : iaR : matr : bestr : han : a f mest : han iaR : þekili

Kunungr er matar bæztr. Hann a fe mæst. Han er þækiligr,

The king is generous with food. He is the richest [man]. He is full of favour.

On the back the inscription says:

mari . selti . rif . a fe mest . aff asa

Mari (?) seldi rif. A fe mest . . .

Mari (?) gave the rib. He is the richest . . .202

This inscription may have been carved by coincidence on a bone when Mari took part in a banquet organized by the king in one of the halls discovered in

200 See Jansson 1985, 63f., 131ff.; 1987, 158ff., 127; Herschend 1994, 187ff. 201 Palm 2004, 239. 202 Gustavson et al. 1992, 166. My translation based on Gustavson. Ceremonial Feasts as Ritual Strategies 363

Sigtuna.203 Since the bone is dated to around 1100 it is possible that the term kunungr refers to King Ingi the Elder. It thus seems as if Christian kings and magnates also held great banquets in halls during the transition period. The bone from Sigtuna is an indication of such events.

To conclude, several notions in Snorri’s description of pre-Christian ceremonial feasts may thus be attested in other independent sources, such as skaldic poetry, Latin texts, runic inscriptions and archaeological finds. According to these sources, the rulers in all three investigation areas were supposed to take part in and perform certain rituals when attending such feasts, for instance the drinking ceremonies and food customs. These ceremonies took place in some kind of cultic building or banqueting hall at the ruler’s farm or the communal assembly building (as in Mære), and it seems as if the ruler was supposed to be seated in his high-seat (see ch. 6) while performing drinking rituals. The rituals related to the sacrificial blood created a dramatic frame, and most likely these actions, including the cultic feast with gift-giving ceremonies, gave much prestige to the ruler. Snorri’s text and other evidence also indicate that the sacrificial ceremonial feasts were comprehensive and complicated. They included a complex of individual rites, such as the ritual killing, the treatment of sacrificial blood, the divination, the ritual cooking, the drinking ceremonies, the recitation of ritual formulae, and the distribution of gifts etc. The rulers of the sanctuaries needed cultic leaders/religious specialists for different aspects and separate rites during the public sacrifice, although it was the ruler who had the overarching responsibility for the public sacrificial feasts. These rituals may have been performed at separate places of the sanctuary, e.g. in connection with stone structures, at holy groves and in the hall or the cultic house (see ch. 5–6). In any case, the ruler played a very central ritual role in these ceremonial feasts. They can undoubtedly be regarded as a ritual strategy for authorization.

9.8 Excursus: The Central Role of Noble Women at the Ceremonial Banquets

In Hákonar saga góða no females appear as cultic leaders during the sacrificial feasts. Other Germanic sources often state that noble women took part in the ceremonial banquets performed in halls.204 The most informative text

203 For the halls in Sigtuna see Tesch 2001, 28. 204 See particular Enright 1996; but also Sundqvist 2005a; 2007. 364 Chapter 9 on such ceremonies in the context of halls is the Old English poem Beowulf, written sometimes between eighth and eleventh century.205 According to this poem King Hroðgar owned a hall (OE heall or sele) called Heorot.206 In this hall the retainers and other guests were seated according to social rank on the long benches. It was the assignment of Queen Wealhtheow to pour mead into the participants’ cups.207 This too was done in ritual forms and according to rank. By means of the formalized drinking ceremonies Wealhtheow creates a social balance between the king and the retainers in the hall. She is therefore called freoðuwebbe ‘weaver of peace’.208 In Beowulf it appears that the queen, in addition to her role in the cup-offering ritual, also gave gifts and advice to the retainers in more or less ceremonial forms in the hall. For instance, she gave Beowulf a necklace, a coat of mail, and much advice.209 These ritual gifts may have been part of an old redistribution system. By means of the gifts, the queen and king won loyal men, manpower and raw materials. Beowulf thus reflects ceremonies in the warrior hall, which may be connected to the ruler strategies for gaining legitimacy. This text should not, perhaps, be treated as a historical source, but it probably reflects a mentality which was common among aristocrats and comitatus retinues in the Late Iron Age society of northern Europe and in the Early Middle Ages (see further ch. 15 below). Similar ceremonies are also described in medieval Old Norse texts. Orkneyinga saga 86 (dated to c. 1200) mentions that Earl Rǫgnvaldr of Orkney sailed to France and came to a seaport which was named Nerbón (Narbonne). At this place a noble woman called Queen Ermingerðr was ruling over people. She and her kinfolk invited the earl and the best of his retainers to a splendid banquet (til veizlu virðili grar). “There was lavish entertainment at the feast, and nothing that might be done in his honour was denied the earl. The earl was sitting feasting one day when the queen came into the hall escorted by a group of ladies and carrying a serving-bowl. She was in her finest clothes, with her hair falling loose as is customary with virgins, and a golden tiara upon her forehead. She served the earl, while her companions began to entertain them with music.”210 This scene,

205 On the dating of Beowulf, see e.g. Klaeber 1950, cvii–cxiii.; Wrenn & Bolton 1996; Mitchell & Robinson 1998. 206 Beowulf 1015; 1640ff. The text is edited by Klaeber 1950. On Beowulf and its connection to Late Iron Age Scandinavia, see Bo Gräslund’s (2015 MS) interesting forthcoming study. 207 Beowulf 612ff. On this text, see e.g. Enright 1988; 1996. On her role as gift-giver, see Beowulf 1215ff. 208 Beowulf 1942. See e.g. Enright 1988; 1996, 21f. See also Klaeber 1950, 335. 209 Beowulf 1188–1231. 210 Ok er þeir kómu til veizlunnar, var þar inn bezti fagnaðr ok engi hlutr til sparaðr, sá at jarl sins sómi væri þá meiri en áðr. Þat var einn dag, er jarl sat at veizlunni, at drottning gekk Ceremonial Feasts as Ritual Strategies 365 which probably was influenced by the interpretation of the Norse scribe, seems to have a ritual context. For instance, the noble lady entered the hall in a kind of ceremonial procession, in which the ladies also took part. The queen’s golden beaker (borðker . . . af gulli), her dress and the golden tiara in her hair also indicate a formal or ritual context. She too performs the action of pouring wine into the earl’s cup, while the ladies were playing, singing or dancing (leka) before them. By performing these ceremonies the queen displayed her hospitality and generosity towards the guests. At the same time she demonstrated her splendour, wealth and grace to the other people in the hall.211

9.8.1 The Ritual Role of the Housewife (húsfreyja) in Vǫlsa Þáttr Whether royal women took part in such “hall ceremonies” in the three investigation areas is of course uncertain. The sources do, however, mention some other women in these areas who may have played certain ritual roles at religious feasts in halls and other places. Such ladies were sometimes called húsfreyja sg. (OSw husfröja) ‘housewife’ or ‘the mistress of the house’.212 They were a kind of female leaders on the farm, who watched over the economy and household, as well as storehouses and domestic servants.213 The keys to the house and chests were symbols of their power. They were always carried visible at her girdle and sometimes they were artistically decorated. It seems as if the housewives were also involved in the religious ceremonies at the farm, perhaps as some kind of cultic leaders.

inn í hǫllina ok margar konur með henni; hon hafði borðker í hendi af gulli. Hon var klædd inum beztum klæðum, hafði laust hárit, sem meyjum er títt at hafa, ok hafði lagt gullhlað um enni sér. Hon skenkti jarli, en meyjarnar léku fyrir þeim. Orkneyinga saga, Ísl. Fornr. 34, 210. The text is based on the vellum Flateyjarbók codex written around 1390. Trans. Hermann Pálsson & P. Edwards. 211 See also Orkneyinga saga, ch. 52, Ísl. Fornr. 34, 112, and Sundqvist 2007, 60ff. In this context a quotation from Archbishop Hincmar of Reims in the ninth century must also be mentioned. In his De Ordine Palatii ch. 5 we read thus: “Für die rechte Ausstattung des Hofes und besonderes für die Jahresgaben der Vasallen—abgesehen von der Versorgung mit Speise und Trank und dem Unterhalt der Pferde—waren vornehmlich die Königin und unter ihr der Kämmerer zuständig.” Quoted from A.-S. Gräslund 2006, 164. 212 See Fritzner 1954. 213 It thus seems as if the society was not only ruled by men. Women played roles as leaders at several levels in society. See e.g. Jesch 1996, 28ff., 63ff., 193ff. Admittedly men controlled the public space during the Viking Age, but in certain spheres women may have performed as leaders, e.g. the public cult. There are also women who transgressed gender boundaries in the sources. On the position of women in pre-Christian society, see e.g., Jesch 1996; B. Sawyer 2000; Kristoffersen 2000, 127ff.; A.-S. Gräslund 2001, 65–89; Bagge 2005, 21ff.; Ney 2004; Quinn 2005; Hultgård 2005; Sundqvist 2005a; 2007; 2015. 366 Chapter 9

There is a burlesque narrative called Vǫlsa þáttr preserved in Flateyjarbók (c. 1380), which describes a húsfreyja at a remote place in northern Norway at the beginning of the eleventh century.214 It tells of certain rites performed at a farm in connection with the autumn slaughter. This ritual centred on the penis of the slaughtered stallion, called ON Vǫlsi. The housewife had ritually prepared the penis (vingull) with onion and herbs, and then wrapped it in linen cloth and put it into a chest. By means of this treatment the strength of the penis grew so that “it could stand great and strong for the housewife if she wished that”. She honoured the Vǫlsi as if it was her god (guð sinn). Every evening when all were gathered in the main house of the farm ( j stofu) she brought the penis to the high-seat, where her husband was sitting. She placed it on his lap and recited a stanza:

Aukinn ert Vǫlsi ok upp of tekinn líni gæddr, en laukum studdr; þiggi mǫrnir þetta blæti . . .215

After this the Vǫlsi was passed from hand to hand, and everyone who received it uttered a verse, often obscene, and always accompanied with the puzzling refrain: þiggi Mǫrnir/þetta blæti. When King Óláfr came to the farm he threw the Vǫlsi on the floor to the dog, and finally he succeeded in converting the people living there. The credibility of this text has been discussed.216 The narrative is put in a frame which concerns St Óláfr and his efforts to Christianize the Norwegians. The description of the old religion may thus have been deliberately made bizarre. But there are some elements in the story which seem to be very ancient. For instance, the magically charged words “linen” and “onion” (linalaukar) are carved on a bone knife found in a female grave discovered at Fløksand, Norway, and dated to the Late Roman Iron Age. This inscription has been associated

214 Vǫlsa þáttr, Flateyjarbók II, 331–335. 215 Text from Skj. B2, 237. “Enlarged are you, Vǫlsi, and taken up, enriched with linen, supported by leeks, may mǫrnir accept this sacrifice, . . .” My trans. 216 On Vǫlsa þáttr and its source value see e.g. F. Ström 1954, 22ff.; E. O. G. Turville-Petre 1964, 256ff.; Steinsland & Vogt 1981; Mundal 2001; Näsström 2001a, 100ff.; Raudvere 2003, 134ff.; Kaplan 2011, 201f. Ceremonial Feasts as Ritual Strategies 367 with the ritual treatment of the horse penis by the housewife in Vǫlsa þáttr.217 There are other elements in this text which may reflect ancient customs. When the king throws the Vǫlsi on the floor, the housewife performs an interesting ritual. As she is chanting a verse, she also heaves herself above the frame and ridge of a door (hefi mik of hjarra ok of hurðása) to see if she can rescue the sacred sacrificial object (ef ek borgit fæ blætinu helga) (i.e. the Vǫlsi). According to Ibn Fadlān (tenth century), a possible Scandinavian slave-girl performs a similar ritual during a chieftain burial. She had voluntarily accepted to follow her master to the Other World. Before she was ritually killed, according to Ibn Fadlān, she did this: “they brought the slave-girl to a thing that they had constructed, like a door-frame. She placed her feet on the hands of the men and was raised above that door-frame. She said something and they brought her down . . . I quizzed the interpreter about her actions and he said, “The first time they lifted her, she said, ‘Behold, I see my [dead] father and my [dead] mother.’ ”218 By means of this ritual she had the ability to see into the Other World, where her parents were already dwelling. There are thus probably traces of old genuine folklore and rites in this story, which may have survived until the conversion period.219 E. O. G. Turville-Petre suggested that Vǫlsi and Mǫrnir (also called vingull and beytill) are one and the same, namely the emblem of Freyr.220 The phallus was not only the emblem, however, but also the embodiment of the god. Vǫlsi- Mǫrnir was thus, according to Turville-Petre, both the sacrificial victim and the recipient of the sacrifice. He relates this idea of sacrifice to Hávamál 138 and the words of Óðinn: “wounded with a spear and given myself to myself”. Since Mǫrnir in other poetic contexts refers to ‘sword’, i.e. a natural phallic symbol and perhaps a representation of Freyr, this interpretation is plausible. It should be noted, however, that Mǫrnir is female plural of Mǫrn, and that the name could refer to the partner(s) of the fertility god, namely the fertility goddess(es) or the giantess(es).221 This interpretation would also make sense of the phrase: þiggi Mǫrnir/þetta blæti. The ritual in Vǫlsa þáttr thus symbolized a hieros gamos, where Vǫlsi represented Freyr, and Mǫrnir his female cov- enanters. In cultic contexts Mǫrnir could be represented by the housewife and other female participants.

217 See e.g., Heizmann 1995. 218 Translation Montgomery 2000. See also Wikander 1979. 219 Cf. Hultgård 2003a. 220 E. O. G. Turville-Petre 1964, 258. 221 Cf. F. Ström 1954; 1985 and Steinsland & Vogt 1981; Steinsland 2005. 368 Chapter 9

Whether this housewife performed in cultic roles in more public arenas is not clear, however. It is therefore uncertain whether this piece of evidence is relevant for the topic of the present study, i.e. ceremonial buildings, sanctuaries and religious strategies for rulership.

9.8.2 Noble Housewives in Svetjud We have a very interesting piece of evidence of a Viking Age housewife in the direct source materials from the Mälaren area, which may suit better in the context of this investigation. In a runic inscription, dated to the eleventh century, from Hassmyra, Fläckebo parish, in Västmanland (Vs 24) a housewife (RSw hīfrøyja) is praised by her husband Holmgautr.222 Parts of the inscription are formulated as a praise poem in fornyrðislag which has counterparts in the Old Norse skaldic poems:223

buonti x kuþr x hulmkoetr x lit x resa x ufteR x oþintisu x kunu x seno x kumbr x hifrya x til x hasuimura x iki betr x þon x byi raþr roþbalir x risti x runi x þisa x sikmuntaR x uaR . . . sestR x kuþ

Boandi goðr Holmgautr let ræisa æftir Oðindisu, kunu sina. Kumbr hifrøya til Hasvimyra æigi bætri, þan byi raðr. Rauð-Ballir risti runir þessar. Sigmundar var [Oðindisa] systir goð.

The good farmer Holmgautr had this [stone] raised in memory of his wife Oðindisa. A better housewife will never come to Hassmyra to run the farm. Rauð-Ballir carved these runes. She was a good sister to Sigmund.224

222 It should be noted, however, that the ON húsfreyja and RSw hīfrøyja have different first elements hús ‘house’ and OSw hi or hæ ‘family’. Both words (húsfreyja and hīfrøyja) are nevertheless usually translated ‘housewife’. 223 See e.g., Jesch 1994. 224 Text and interpretation may be found in Vs 24, Jansson 1964, 69–76. Trans. to English in Jesch 1996, 65. Ceremonial Feasts as Ritual Strategies 369

The information that the housewife “runs” or “rules over” the farm [the village or settlement] (byi [boi] raðr) is important.225 It indicates that Oðindisa (Ōðin- Dīsa) had a leading position in a more public sphere, perhaps also outside the walls of the building. The Runic Swedish verb rāða may be interpreted as ‘rule over’.226 The text also says that Oðindisa’s husband was “a good farmer” (boandi goðr). This expression is intended to show that Oðindisa was married into a good family. The information that she was a sister of Sigmund also intended to indicate her high position in the local community and that she originally derived from an influential family. There is slight evidence indicating that the housewife of Hassmyra was involved in cultic matters. The compound name Oðindisa is only attested in this inscription. In a formal sense the iō-stem dis should not appear with a weak inflection in this case.227 According to the onomastic specialists Roland Otterbjörk and Thorsten Andersson this may have a simple explanation.228 The female name was probably originally Disa.229 The name of the god Oðin- has been added to her secondarily, as a byname prefix.230 The name Oðindisa indicates that she had a specific relation to the god Óðinn. Perhaps it even suggests that in her youth she appeared in ritual functions in the cult of Óðinn. At least she was faithful to the old Æsir gods during the conversion period. Old Norse sources indicate that other housewives in pagan Svetjud were involved in religious rituals at sanctuaries, especially during the calendrical feasts. According to Austrfararvísur (c. 1020),231 quoted by Snorri Sturluson in

225 The runologist Thorgunn Snædal writes to me in a e-mail dated 20 November 2006 that the u-rune in this sequence actually not is stung, as Jansson states. She proposes that the passage should be read bui raþr and normalized to boi ræðr, i.e. that Oðindisa ruled over the farm, the animals, the people, the stores, the crew etc. Cf. ON búsforráð ‘management of household affairs’ Cleasby & Vigfusson 1957 (1874), 86. On ON bú, see Fritzner 1954, Vol. I, 204ff. See also Peterson (1994, 8), however, who states that the sequence could be read byi and interpreted as a dative of a RSw byr ‘gård, by’. 226 Peterson 1994, 43. 227 Peterson 1981, 149. 228 Otterbjörk 1983, 110; T. Andersson 1992a, 512; 1993, 45f.; Peterson 2007, 171. I would like to thank the runologist Jan Axelson for these references. 229 OSw dis means ‘woman, mythical female being’; cf. ON dís. 230 According to Thorsten Andersson, Oðindisa belongs to a group of “Theophore Personennamen”: “Da die schwache Form des Namens auffallend ist, erscheint die Erklärung ansprechend, daß es sich hier um den Namen Disa mit dem Götternamen als sekundärem Beinamenpräfix handle.” T. Andersson 1992c, 512; 1993, 45. Cf. Vikstrand 2009. 231 Skj. B1, 220–225. For an English translation, see Page 1995, 48ff. 370 Chapter 9 his Óláfs saga helga 91 (c. 1230),232 the Christian skald Sigvatr Þorðarson, with some companions, made a journey to Svetjud at the beginning of the eleventh century. One evening they came to a farm called Hof. Whether Hof in this case refers to a place-name is uncertain. It is possible that it should be interpreted as an appellation hof, i.e. a building where cultic activities took place. The poem says “the door was shut” (hurð vas aptr). It also says that Sigvatr and his fellows were sent away, since the people there declared that it was “holy” (heilagr) there. The text mentions that they went to another farm. There was a housewife (húsfreyja) in the doorway: “No farther can you enter, you wretch!”, she said, “we are heathen people, and I fear the wrath of Óðinn” (hræðumk ek við Óðins reiði). The poem also mentions that “she was holding álfablót there in the house”, i.e. she was preparing sacrifices to the mythical beings called álfar. According to some scholars, the calendrical feast called álfablót was identical with the pre-Christian Yule (ON jól).233 It seems as if the Yule feast had some connections to the cult of Óðinn.234 The Æsir-god is called Jólnir (or Jóli) in Þórsdrápa st. 12 and in Ágrip ch. 1.235 Also the poetic kenning Hugins jól ‘the (Yule) feast of Huginn’, must in one or another way be related to Óðinn or more exactly to his raven. During such feasts ceremonial drinking took place inside the ceremonial buildings. Most likely, however, álfablót, was a religious feast celebrated during the autumn, separate from the Yule feast, which took place at midwinter or the winter solstice. It is possible that housewives had leading cultic roles at such feasts and also gained social prestige for this.

9.8.3 Gyða from Møre in Norway The Old Norse texts also report that the housewives sometimes were initiators, organizers and leaders of great ceremonial banquets in Norway. These feasts were often performed in ritual circumstances, where drinking ceremonies also took place in the hall buildings. According to Egils saga 64, a noble housewife (húsfreyja) and widow, called Gyða, arranged such banquets. She was the sister of Arinbjǫrn hersir, and thus she belonged to the elite of the local community.236 She was also the mother of Friðgeirr, who was a landholder (lendr maðr) on the island called Hǫð [Hareidlandet] outside Møre. Since Friðgeirr was very young, Gyða helped him to run the farm (hon var at ráðum með syni sínum). As in the Hassmyra inscription, this text uses the verb ráða ‘to rule’ or ‘to run’. When

232 Heimskringla II, Ísl. Fornr. 27, 135ff. 233 See Celander 1955; cf. F. Ström 1985, 89ff. 234 See Nordberg & Sundqvist 2008, 20f. 235 Ágrip, Ísl. Fornr. 29, 3 and Skj. B1, 142. 236 Egils saga, Ísl. Fornr. 2, 199ff. Ceremonial Feasts as Ritual Strategies 371

Egill and his men arrived on the island, Gyða arranged an expensive banquet, with plenty of drinking (var þar drykkja mikil ok dýrlig veizla). During the evening Gyða approached Egill. It is in precisely this context Gyða is designated húsfreyja. A couple of days later she arranged a new banquet at her farm. Egill and his men went into the hall room (í stofu) and started drinking. There were many people and a great feast (var þar þá mikil veizla). Egils saga thus reports that a housewife could be involved in public banquets at halls in Norway. This housewife appeared as the arranger of the feast, and perhaps she also performed as some kind of ceremonial leader when she approached Egill and was designated húsfreyja. It is very hard to estimate the source value of Egils saga when it comes to rituals taking place in the Viking Age. The prose must be considered as indirect secondary materials, but some of the stanzas on which the prose partly is based may have been composed by Egill himself, sometime during the first half of the tenth century. They could thus be considered as direct or contemporary sources. There are several stanzas said to be made by Egill in chapter 64, and one mentions also the boy Friðgeirr. No stanzas support, however, the scene where Friðgeirr’s mother invited Egill to the ceremonial feast and the ritual drinking. It is thus uncertain whether the information about the ritual feast was based on an old oral tradition.

9.8.4 Unnr (Auðr) in djúpúðga Ketilsdóttir from Hvammr, Iceland Old Norse prose reports that powerful females in Iceland appeared in similar roles to that of Gyða. One is Unnr (Auðr) in djúpúðga Ketilsdóttir, who appears in Laxdœla saga,237 Íslendingabók and Landnámabók.238 She was the daughter of the rich and powerful man called Ketill hersir flatnefr from Raumsdœlafylki in Norway, and married to Óláfr hvíti Ingjaldsson. Unnr followed her father to Scotland in order to escape the tyranny of King Haraldr. After her father, husband and son were killed, Unnr built a ship and gathered up all her valuables, her kin and followers, and sailed to Iceland. Unnr (Auðr) claimed land at Hvammr in western Iceland at Breiðafjǫrðr, and in Íslendingabók she is described as a landnámskona.239 Like other early immigrants to Iceland, she performed divination rituals with her high-seat posts on her arrival.240

237 Laxdæla saga, Ísl. Fornr. 5, 6–15. In Íslendingabók and Landnámabók she is called Auðr in djúpúðga. On this tradition, see also Herschend 1998, 137ff.; Hedeager 2011, 118ff. 238 Íslendingabók, Ísl. Fornr. 1, 6, 26; Landnámabók, S 12–14, 94–110, H 12–14, 81–84, Ísl. Fornr. 1, 50f., 136ff. 239 Íslendingabók, Ísl. Fornr. 1, 26. 240 Clunies Ross (1998, 147ff.) emphasizes that certain rituals in these contexts could be specifically for females. 372 Chapter 9

The description of her landnám is almost mythologized in the saga and it emphasizes what a powerful and important woman she was. While looking for land to settle at Breiðafjǫrðr she stopped at places which afterwards were known by names associated with her. Dǫgurðarnes is the place where she stopped for breakfast, Kambnes is the place where she lost her comb, and Krosshólar where she erected her cross.241 For Unnr was baptized and a Christian woman. Like a powerful landnámsmaðr, she distributed land to her kinfolk and followers after she had settled at Hvammr. Both Laxdœla saga and Landnámabók report that Unnr (Auðr) arranged ceremonial banquets. According to Laxdœla saga, she once went to great expense with the feast (hafði mikinn fékostnað fyrir veizlunni), when her kinsman Óláfr feilan married Álfdís. She invited many prominent people from dis- tant districts. The saga describes the leading and ceremonial role of Unnr, and how rich and generous she was:

Unnr then entered the hall, followed by a large group of people. When the hall was filled, everyone was impressed by the magnificence of the feast.242

Unnr was very old when the wedding feast was held, and therefore she announced that her inheritance should be handed over to her grandson Óláfr when she was dead. During the feast Unnr died as well and the feast then continued, in commemoration of both Óláfr’s marriage and Unnr’s death. Óláfr now owned the farm at Hvammr and its property. When the feast concluded he gave generous parting gifts to his most respected guests.243 It is quite obvious that Unnr was the one who previously distributed gifts at banquets at Hvammr. In the version of Auðr’s (Unnr’s) death in Landnámabók it is clear that she also had this role.244 In this text no wedding feast is mentioned. It just says that she invited her kinsmen to an expensive feast (bjó dýr liga veizlu). When the feast had been going for three nights she distributed gifts to her friends (valði hon gjafir vinum sínum) and gave them good advice (réð þeim helræði). She said that the feast must continue for three more nights, but that this feast was going to be her “inheritance feast” (erfi).

241 Laxdæla saga, Ísl. Fornr. 5, 9; Landnámabók, S 97, Ísl. Fornr. 1, 139. 242 Síðan gekk Unnr inn í skála ok sveit mikil með henni. Ok er skálinn var alskipaðr, fannsk mǫn num mikit um, hversu veizla sú var skǫrulig. Laxdæla saga, Ísl. Fornr. 5, 12. 243 . . . gefr Óláfr stórmannligar gjafar þeim mǫnnum, er þar váru mest virðir, áðr á brott fóru. Laxdæla saga, Ísl. Fornr. 5, 13. 244 Landnámabók, S 110, Ísl. Fornr. 1, 146f. Ceremonial Feasts as Ritual Strategies 373

According to the sources, Unnr thus acted as a male magnate during the ceremonial feasts in the hall buildings. She arranged and invited people to these feasts, she entered the hall in some kind of ritual procession, and also took part in the drinking ceremonies. Unnr also distributed gifts to loyal people and friends during these assemblies. Unnr was thus some kind of leader in the local community, and therefore she also appeared in these roles at public gatherings.

9.8.5 Archaeological Evidence of Powerful Females with Possible Ceremonial Functions There are plenty of other texts indicating that housewives and noble females could be powerful on different levels in the societies of the investigation areas. The housewife Ásgerðr Asksdóttir, for instance, was a landnámskona with plenty of power. She took land in the south of Iceland, between Seljalandsmúli and Markarfljót and Langanes.245 Like Gyða and Unnr/Auðr, she was a widow. It seems as if widows could appear in social roles which otherwise were intended for men.246 But even married women could be the initiators of banquets which had a more public character. One example is Ásta, who invited all the magnates in the settlements to a banquet (veizla) and a “welcome-beer” (fagnaðarǫl) for her son Óláfr Haraldsson. She did that without asking her husband Sigurðr sýr.247 In this context the women designated baugrýgr should also be mentioned. Norrøn ordbok translates the term ‘kvinne som har rett til bot el. skyldnad (plikt) til å betale bot’.248 The concept appears in Norwegian laws.249 In a judicial sense such a woman was considered to be a man. She could, for instance, carry weapons and take part in the judicial assembly.250 We also hear about powerful widows in Viking Age runic inscriptions from Uppland, Södermanland and Öland. It seems as if they ruled over their properties independently of men.251 Archaeological evidence indicates that such powerful women also were involved in the public cult at cultic sites in the three investigation areas. In Old Uppsala the skeleton of a middle-age female was discovered with expensive grave goods in a boat grave (grave 36) dated to the ninth century.252 She was

245 See Landnámabók, S 341, H 299, M 6, Ísl. Fornr. 1, 343. 246 See Jón Viðar Sigurðsson et al. 2005, 119f. 247 Óláfs saga helga, in Heimskringla II, Ísl. Fornr. 27, 40. 248 Heggstad et al. 1993, 45. 249 See e.g., Den ældre Gulathings-Lov 275, NGL 1, 92. 250 On these females called baugrýgr, see Klos 2007. 251 See B. Sawyer 2000, 8. 252 Nordahl 2001, 46ff. 374 Chapter 9 dressed in fine clothing made of materials of the highest quality. She had a rich set of jewellery and a knife with a silver-decorated sheath. In addition she had two oval brooches and one equal-armed brooch. Most interesting, in the light of the present study, was the small pendant of bronze. It was made in the shape of a woman holding a beaker in front of her. This type of pendant is usually called “valkyria” and they have been found at several places in the Mälaren Valley. They indicate a ceremonial role of females, i.e. the important function of pouring wine and beer for the guests during the ceremonial meals. It is quite possible that the noble lady in the boat grave had this ritual function in the Early Viking Age cult of Uppsala. The Merovingian Period hall of Uppsala was situated just a few hundred metres from this grave. Ceremonial banquets probably took place in this hall. Most likely another hall was erected in the vicinity during the Viking Age, where the lady in the boat grave perhaps played a role during the ceremonial feasts. At Valsgärde, about 2–3 kilometres to the north of Old Uppsala, several rich Viking Age burials of females were found. They included some expensive glass and metal vessels. Perhaps some of these vessels were used at ceremonial banquets.253 A Late Iron Age hall building was situated just beside the burial field at Valsgärde.254 It is quite possible that the women buried in these graves had a similar role to that of Wealhtheow in Beowulf during the ceremonial banquets. A chamber grave (Bj 523) in Birka, dated to the tenth century, may lead us to a similar conclusion. In this grave a female was buried with two drinking horns with silver mounts.255 At Birka too we have evidence of a Viking Age hall where ceremonial banquets took place. Grave 85 at Valsgärde contained the cremated remains of an old woman, one horse, two dogs and a cock as well as a pig jaw. In her grave there was also a staff with a bone animal head. This animal-headed staff reminds us of the animal heads at Oseberg in Norway, where two females were buried. In the Oseberg grave, several Viking Age beast sculptures were found. It has been argued that they had a religious function.256 It is quite possible that the animal-headed staff in Valsgärde 85 also had a religious function.257 Most likely the lady there was a powerful woman. Probably she also played important roles in the public cult of Valsgärde. Anne-Sofie Gräslund writes thus: “Concerning the woman in

253 A.-S. Gräslund 1999. 254 See Norr & Sundkvist 1995; Ljungkvist 2006, 70. 255 A.-S. Gräslund 2006. 256 Grieg 1928, 65f.; Ingstad 1993; A.-S. Gräslund 2008, 78. 257 A.-S. Gräslund 2008, 77ff. On ritual staffs in general, see N. S. Price 2002; Heide 2006a; 2006b; Tolley 2009. Ceremonial Feasts as Ritual Strategies 375

Vgde 85, in all probability a husfru, my impression is that her religious power was strong and thereby, of course, also her social status.”258 Gräslund supports her assumption with the animal-headed staff, but also with the bones of the various animals found there. They may have had a symbolic meaning in relation to the burial. Female ceremonial leaders may also be found at Tuna in Badelunda, Västmanland.259 At this place rich grave goods were found in a Roman Iron Age chamber grave, such as a necklet, two arm-rings and one finger-ring made out of gold. There were also two bronze vessels of Hemmoor type and glass. In this chamber grave a noble woman was buried. Perhaps the glass and the vessels were used for serving wine in ritual contexts. At the same place eight Merovingian Period and Viking Age boat graves were discovered, all of them containing females. Close to the burial field there is a place called Närlunda.260 The first element in this name is a genitive form of the name of the deity *Njärd. Is it possible that the females in Tuna were related to Närlunda and that they were the cultic leaders in the public worship of *Njärd? Archaeological evidence from Norway indicates that rich females may have been involved in cult and used ritual strategies to gain legitimacy during the Late Iron Age. Besides the rich Oseberg ship grave, the great mounds at the farms of Hauge and Tu in Rogaland may also be related to aristocratic ladies who performed cultic functions during religious assemblies.261

258 A.-S. Gräslund 2008, 77. 259 Nylén & Schönbäck 1994; Vikstrand 2000. 260 Vikstrand 2000. 261 Kristoffersen 2000; 2006; Sundqvist 2015. Chapter 10 Rulers, Rings and Rituals—From Svetjud to Iceland and Back Again

When the authors described the inner parts of the hof sanctuaries in the Sagas of Icelanders, they often commented on the “oath-rings”, which were placed on the altars (stalli, stallr sg.).1 In these texts such rings were often called stal‑ lahringr, hringr or baugr. Certain rituals were performed in connection with them, which seem to be related to the sacrifice, but also to judicial matters. Oaths, for instance, were sworn at them. Such oaths were called baugeiðr or lǫgeiðr. Sources indicate that these rings also had other connotations for the Early Icelanders. It seems as if they sometimes were seen as symbols of honour and dignity related to the chieftains and cultic leaders called goðar.2 When appearing in public functions at sacrifices or judicial assemblies, the goði had to carry the ring in his hand or on his arm. In opposition to some previous researchers,3 I will argue that the testimony from the saga authors regarding these ritual rings is built on old oral traditions. For there are clear indications in more direct sources that such rings served this cultic and judicial purpose. I will also argue that this context of ritual rings was not restricted to Iceland, but can be seen in sources deriving from other parts of the investigation areas, particularly Svetjud. In this area too we can see ritual rings, which seem to appear in the context of pre-Christian sanctuaries and ceremonial buildings. In my opinion these rings were not only important religious and judicial objects there, but also essential media or symbols of status and sovereignty. I will start with Forsa, Hälsingland, where a very interesting iron ring with a runic inscription has been preserved. Much of the discussion in this chapter will be concentrated on this ring, since, in my opinion, it is one

1 The concept of “oath-ring” derives originally from C. J. Thomsen [& Petersen] 1886. Thomsen designated all types of Bronze Age rings as “oath-rings”. We do not know what kind of function these Bronze Age rings actually had. In the present study the term “oath-ring” refers to such rings as had a religious or judicial significance at assembly places and sanctuaries during the Late Iron Age. On this concept, see also Beck 1986, 538ff.; Jacob-Friesen 1986; Brink 1996b, 46; Habbe 2005, 134–144. 2 In some contexts and for certain people these rings provided agency to the carrier in a more direct way, without being perceived as symbols. Cf. e.g., Miller 2005 and Engelke 2012. 3 See e.g., O. Olsen 1966; Kabell 1975.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_011 Rulers, Rings and Rituals 377 of the most important pieces of evidence of the construction of Viking Age authority in the context of sanctuaries in eastern Scandinavia.

10.1 The Forsa Ring from Hälsingland—A Sacred “Oath-Ring” at a Sanctuary?

The ring from Forsa in Hälsingland, Sweden, is forged and decorated with runes (see fig. 17a). It is about 43 cm in diameter, and has for a long time been in the possession of the church in Forsa. When it was first observed in scholarly contexts in the eighteenth century it was nailed to the door between the porch and the interior of the old church which was demolished in 1840. Today the ring is kept in the Forsa congregation archives. For a long time there was a close consensus among scholars that the runic inscription on this ring was a medieval clerical legal enactment from the twelfth century. Recently, however, it has been argued that the ring should not be looked upon as a clerical record.4 It has been stated that the content of the inscription is a Viking Age enactment, more precisely from the ninth century.5 The function of the ring has also been reinterpreted. In previous research it was stated that the ring should be seen as a “church-door ring” (Sw kyrkdörrsring). Today scholars argue that the Forsa ring was an “oath-ring”, i.e. a pre-Christian ring with ritual or symbolic value. It was used by the people of Hälsingland when swearing oaths during the common judicial assembly. This ring was kept at an important cultic site in Hälsingland. In what follows I will discuss the dating and the function of the ring. I will argue that those scholars who believe that this ring is old and had a ritual or symbolic function in a pre-Christian religious-judicial context are on the right track. In connection with this I will discuss the shape of the ring.

10.1.1 The Problem of Dating the Ring Several well-known runologists and philologists have been working with the Forsa ring and its runic inscription.6 There was strong agreement among these scholars that the inscription could not be dated before the twelfth century and that it concerned an ecclesiastic law. In 1940 Elias Wessén made a normalized

4 Liestøl 1979; Ruthström 1990; Brink 1996b; 2010; Widmark 1999; Källström 2007; 2010a; Sundqvist 2007; Williams 2008. 5 Brink 1996b. Liestøl (1979) holds the door open for the possibility that the ring should be dated to the tenth century. 6 E.g., Bugge 1877; von Friesen 1933; Johnsen 1968. 378 Chapter 10

Figure 17a The ring from Forsa in Hälsingland, Sweden. Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet.

Figure 17b Detail from the ring from Forsa. Photo: Bengt A Lundberg. ATA, Riksantikvarieämbetet. text of the inscription, which has often been quoted. It was based on Sophus Bugge’s reading from 1877.7 Gerhard Hafström’s interpretation in the article about the Forsa ring in Kulturhistoriskt lexikon för nordisk medeltid (1959) has been normative and is very close to Wessén’s interpretation:

Oxa tvisgildan ok aura tva staf at fyrsta lagi, oxa tva ok aura fiura at aðru lagi, en at þriðia lagi oxa fiura ok aura atta staf, ok allt æigu i værr, ef hann hafsk ækki rett fyrir, svað lærðir æigu at liuðretti, ova vas innt fyrr ok hælgat. En þæir gærðu sik þetta, Anundr a Tarstaðum ok Ofagr a Hiortstaðum. En Vibiorn faði.

En tvegill oxe och två öre till biskopsstaven för första mässfallet; två oxar och fyra öre till biskopsstaven för andra mässfallet;

7 Bugge 1877; Wessén [& Holmbäck] 1940, livff. Rulers, Rings and Rituals 379

men för tredje mässfallet fyra oxar och åtta öre; och all egendom i kvarstad, om han icke gör rätt för sig. Det som de lärde tillkommer enligt landets lag, det blev förr stadgat och stadsfäst. Men de gjorde sig detta, Anund i Tåsta och Ofeg i Hjortsta. Men Vibjörn ristade.8

The most important argument for those who stated that the inscription had a medieval-clerical background was provided by the two Runic Swedish terms lærðir and staf, which appeared in Wessén’s normalized version of the inscription. The first word was interpreted as ‘a learned person’, i.e. ‘a priest’, while the other term was a metonymic expression for ‘the bishop’ or the ‘the office of the bishop’. The “bishop-staff/rod” was a symbol of the office of the bishop during the Middle Ages.9 The ring should thus be related to a medieval context and concerned a clerical-judicial record. Some scholars felt that the inscription was about fines for omissions inside the church, or as Hafström interpreted it, about fines when no service was held (Sw mässfall). These fines were paid to the bishop. The enactment on the ring thus reflected a phase when the Church was organized in Hälsingland. Recently some scholars have argued that the medieval dating of the inscription is disputable, specifically if the forms of the runes and the language of the inscription are taken into consideration.10 The runes appearing on the ring are of the type called “short-twig runes” or “stub runes” (Sw kortkvistrunor).11 They usually occur in older runic inscriptions, such as the one on the Rök stone, Östergötland (Ög 136) and the Tjängvide stone from Alskog, Gotland (G 110), which are both dated to the Early Viking Age.12 Most runic inscriptions with “short-twig runes” are from the ninth or the tenth centuries. In this context the Forsa ring has been explained as an odd phenomenon, an unusual late inscription with “short-twig runes”. The form of the runes is thus an obstacle to those who date the Forsa ring to the twelfth century. A similar problem per- tains to some word forms in the inscription. The verb fāði ‘cut, painted’, for instance, and the particle svāð ‘so that, if’ are archaic terms, which normally

8 Hafsström 1959, 530ff. 9 In Östgötalagen and Västgötalagen there are expressions such as staf ok stol which designate this. See Schlyter 1887, 594f. 10 See Liestøl 1979; Ruthström 1990; Brink 1996b. 11 Johnsen 1968, 121ff. 12 See also Johnsen 1968, 109–167. 380 Chapter 10 do not occur in twelfth-century inscriptions.13 Both these terms are attested in the Rök inscription and they usually belong to runic inscriptions from the ninth and tenth centuries. Also, the use of the os-rune (Sw åsrunan) as a sign for a nasalized a-sound is assumed to be an ancient feature. In Uppland the sound value of this rune was changed to an o-sound in the middle of the eleventh century.14 It has been argued that the preservation of the ancient runic forms and archaic words in the inscription on the Forsa ring can be explained by the peripheral location of Hälsingland. This argument is doubtful, however. There are at least a dozen runic inscriptions in Hälsingland dated to the Late Viking Age which closely resemble contemporary runic inscriptions in, for instance, Uppland.15 Most likely the runic masters of Hälsingland had a good knowledge of “common” or “modern” Upplandic runic and word forms during the eleventh century. There are also other arguments which support an early dating of the inscription. In 1974 the runologist Aslak Liestøl visited Forsa church and read the inscription there.16 He discovered that the reading of the sequence lirþiR (lærðir) was wrong. It should actually be read liuþiR, and be normalized to a plural of the word liūðR (ON lýðr; OSw *liuþer) ‘people’. The whole sentence was reinterpreted as “that which the people are entitled to demand according to the people’s right (hence, the law of the land)”.17 By means of this new reading, the strongest argument for a medieval interpretation of the inscription disappeared, namely, that the enactment concerned the Christian “priests”. Liestøl also argued that the shape of the Forsa ring differed typologically from “normal” church-door rings. It lacked the stable hold of such rings (see e.g. the door ring from Delsbo church, fig. 18). The door-hold that the Forsa ring is placed in is a secondary construction.18 The hold always hides some parts of the runic inscription when the ring is moved in it. The ring must thus originally have had another function than as a church-door ring. Following on from Liestøl’s results, the historian of law and philologist Bo Ruthström made a new interpretation of the beginning of the inscription in

13 See further Holmbäck & Wessén 1979 (1933–46) III, lvif. See also Peterson 1994, 11, 67. It should be mentioned that the verb RSw fā appears in other runic inscription from eleventh-century Hälsingland; Hs 14 Malsta; Hs 21 Jättendals kyrka. 14 Wessén 1940, lvi; Liestøl 1979. On the problems with the sound value of the os-rune and its use as a criterion for dating runic inscriptions, see Williams 1990. 15 Brink 1996b. 16 See Liestøl 1979. 17 Svað liuðir æigi at liuðretti. Liestøl 1979, 17ff. 18 Liestøl 1979, 22. Rulers, Rings and Rituals 381

Figure 18 The door ring from Delsbo church, Hälsingland, Sweden. Photo: Unknown. ATA, Riksantikvarieämbetet.

1990. He argued that it must be read t vis kil[t]an. This could be normalized and interpreted as at vis gildan “for the restoration of a vi [cultic site] in a valid state” (“för att återställa vi [kultplatsen] i gillt skick”).19 The enactment thus concerned the treatment of the pre-Christian cultic site called RSw vī ‘sanctuary’. He interpreted the word gildan as a verbal noun ‘restoration in a valid state’ (“återställande i gillt skick”). Ruthström’s new interpretation was important in several respects. It made it possible to show what type of crime the scheme of fines in the inscription concerned, namely the mismanagement of a cultic site.20 By means of this interpretation it was also possible to omit the problematic term “tvegill” which previous researchers had identified in the inscription. This term is not attested in the Scandinavian languages. With support from Liestøl’s and Ruthström’s new interpretations, Stefan Brink argued in an article from 1996 that the inscription on the Forsa ring must go back to the ninth century.21 According to Brink, it concerned a pre-Christian legal enactment for the attendance of a cultic site. The designation of cultic sites OSw vi (ON vé) was most likely well known in Hälsingland during the pre-Christian period. There are at least five place-names from this province where this element appears.22 At such a place this ring could have had its real Sitz im Leben. Brink interpreted the entire inscription in this way:

19 Ruthström 1990, 45. 20 Gun Widmark (1999, 118) has produced alternative readings to Ruthström’s suggestion uis kil[t]an. She argues that a staff is missing on the k-rune and reads the sequence at uisbilan at vispillan ‘vid skada på vi’. 21 Brink 1996b. 22 Brink 1990a; 1990b. English translation of the runic inscription in Brink 2008. 382 Chapter 10

: uksatuiskilanaukauratuâstafatfurstalaki : uksatuâaukaurafiurataþrulaki : : inatþriþialakiuksafiuraukauratastaf : aukaltaikuiuarRif anhafskakiritfuriR : suaþliuþiRakuatliuþritisuauasintfuraukhalkat : inþaRkirþusikþitanunrâtarstaþum : aukufakRâhiurtstaþum : inuibiurnfaþi :

Oxa at vis gil[d]an ok aura tva staf at fyrsta lagi, oxa tva ok aura fiura at aðru lagi, en at þriðia lagi oxa fiura ok aura atta staf; ok allt æigu i værr, ef hann hafsk ækki rett fyriR, svað liuðiR æigu at liuðretti, sva vas innt fyrr ok hælgat. En þæiR gærðu sik þetta Anundr a Tarstaðum ok OfagR a Hiortstaðum. En Vibiorn faði.

One ox and two aura [in fine] [to?] staf [or] aura staf [in fine] for the restoration of a cult site (vi) in a valid state for the first time; two oxen and four aura for the second time; but for the third time four oxen and eight aura; and all property in suspension, if he does not make right. That the people are entitled to demand, according to the law of the people that was decreed and ratified before. But they made [the ring, the statement or?], Anund from Tåsta and Ofeg from Hjortsta. But Vibjörn carved.23

Brink felt that the term liuðrettr ‘the law of the people’ referred to the law of the province of Hälsingland, which during this period only consisted of some common customs, legal rules and enactments detached from their context, as we may see in the inscription on the Forsa ring. He argued further that the ring had its real function at the provincial assembly (Sw landsting) at Kungshögen in Hög parish, which probably was the political and religious central place of Hälsingland during the Viking Age and the Middle Ages.24 At this place this ring was probably a symbol of power and perhaps an “oath-ring” comparable

23 Brink 1996b and 2008, 28f. 24 There are two arguments for this: (1) Anundr and Ofegh, who are mentioned in the inscription, and are said there to have “made” the ring and the legal enactment, seem to have lived in Tåsta and Hjortsta. These villages are located in the parish of Hög. (2) According to an old clerical legend written down around the eighteenth century, the Rulers, Rings and Rituals 383 with the ones seen in the Old Norse sources. People thus swore their holy oaths on this ring at the assembly. He also argued that it was Anundr and Ofegh who made this ring and formulated the legal enactment inscribed on this object. They were some kind of judicial and cultic leaders. One problem remains when interpreting the inscriptions as pre-Christian, namely the term staf (perhaps a dative, without any sign of the case inflection ‑i). Previously the runologists interpreted it as “to the bishop”. The term Old Swedish staf (staffr, staver; cf. ON stafr) is present in the pre-Christian vocabulary and designates there a cultic object located at the cultic site. It appears, for instance, in the place-name Nälsta (< *Niærdharstaver) ‘the staff of the deity *Niærdh’. Such cultic staffs are also attested in the reliable written sources. Ibn Fadlān, for instance, mentions the Rūs (probably Svear), who sacrificed at some kind of staffs or posts which were placed in the ground (see ch. 7 above). Brink suggests that the staf in the inscription on the Forsa ring may be interpreted in a metonymic sense, as referring to the one who used the staff as a sign of dignity, i.e. the chieftain or the leader responsible for the cultic site.25 There are, however, no other examples in the runic inscriptions or Old Norse poetry where RSw stafr or ON stafr in a figurative sense refers to a leader of society or cult. The meaning of staf in this inscription is quite enigmatic. It cannot be excluded, however, that it should be related to a ritual object appearing at the cultic site.26

10.1.2 The Three-Branched Figure and the Fire-Steel-Shaped Ring Despite the problem with the term staf, there are, in my opinion, reasons to believe that the Forsa ring should be dated to the Viking Age. The runological and philological arguments presented above support such an assumption.27 In addition to them there are also some arguments which may be presented from the perspective of history of religions. These arguments have rarely been noted

Forsa ring was nailed to the door of Hög Church before it was transferred to the Forsa Church. Brink 1996b, 40. See also Hafström 1959. 25 Brink 1996b, 37. Ruthström (1990, 49ff.) argues that staf refers to the fence which was set up around the sanctuary. 26 On a cultic stav in place-names, see Vikstrand 2001, 292ff. 27 Widmark (1999, 123) emphasizes that both vocabulary and syntax are archaic in this inscription and could very well be dated to the ninth century. She suggests that the three vertical lines which introduce the inscription could be interpreted as i-runes, with a magic intention. It could also support a pre-Christian background for the inscription and the ring. Magnus Källström (2010a) has rejected this interpretation and argued that these signs only tell where the reading should start. See also Bugge 1877, 7. 384 Chapter 10 before in this discussion. In what follows I will present some of them, which concern the shape and appearance of the Forsa ring. The Forsa ring has at one side a three-branched figure, which at first glance may be taken as a Christian cross or a lily (see fig. 17b). But it could also be interpreted in other ways. There are two runic stones in Södermanland with iconographic motifs which remind us of the figure on the Forsa ring. On the Åby stone (Sö 86) and the Stenkvista stone (Sö 111) from the settlement of Rekarne (Rekarnebygden) there are large hammers depicted in the middle above the runic coil (see fig. 5). According to the historian of religions Anders Hultgård, these motifs may directly be associated with the pre-Christian religion and the hammer of the deity Þórr, called Mjǫllnir.28 The motifs on the runic stones differ to some degree from the figure on the Forsa ring. The images of the stones are more T-shaped than three-branched. Other sources from the investigation areas, however, indicate that Þórr’s hammers were sometimes three-branched at the top of the handle. The assumed Þórr figure from Eyrarland, Iceland, for instance, holds a three-branched object in his hand which is usually interpreted as Þórr’s hammer (see fig. 15 above).29 The hammer of the figure on the Altuna stone, Uppland, also seems to be three-branched (see fig. 6). There is great consensus among researchers that this figure represents the god Þórr. On this stone he is struggling against the Miðgarðsormr. This motif is known from literary sources.30 Similar three-branched Þórr’s hammer symbols appear on a ring from Hilda, Innvik, Norway, a Þórr’s hammer made of silver from Läby, Uppland, one from Mickels, Gotland, and intimated on several silver pendants with Þórr’s hammers (see fig. 19a and 19b).31 In the Mälaren area, particularly, small Þórr’s hammers made of iron are common, of similar form, sometimes placed on larger rings, known as Þórr’s hammer rings. When the Forsa ring is observed together with the assumed hammer as a whole, a very characteristic form emerges. It remind us undeniably of the fire- steel-shaped pendants which have been found at several places in Scandinavia, for instance in Borg, Östergötland (see fig. 20).32 These pendants, according to the archaeologists, have a form symbolism which belongs to the Viking Age.33 They often appear together with the miniatures representing Þórr’s hammer.

28 Hultgård 1992, 89ff. On the hammer as a symbolic expression of ancient Scandinavian religion, see A.-S. Gräslund 1983–84; 1985; Hultgård 1999. 29 But see also Perkins 2001. 30 See e.g., Hymiskvida and Gylfaginning 48. See also Meulengracht Sørensen 1986. 31 For a thorough survey of the Þórr’s hammer symbols, see Perkins 2001, 82–149; A.-S. Gräslund 2001, 55ff. 32 Liestøl 1979, 21ff. 33 On the dating of these objects, see Jonsson 2006. Rulers, Rings and Rituals 385

Figure 19a Þórr’s hammer made of silver from Laby, Uppland. SHM

Figure 19b Þórr’s hammer from Mickels, Nar, Gotland. SHM

Figure 20 Fire-steel-shaped pendants from Borg, Östergötland. SHM

The relationship between Þórr, the hammer and the fire-steel is in my opinion quite natural: Þórr used his hammer Mjǫllnir when making thunder and fire.34 According to the Old Norse traditions, Þórr’s hammer had also other functions. It was used, for instance, when consecrating or protecting various phenomena and objects. In one Eddic poem the hammer consecrates a wedding couple; it was placed in the lap of the bride.35 According to Gylfaginning, the hammer was also used at the funeral pyre of Baldr: “Then Þórr stood by and consecrated the pyre with Mjǫllnir.”36 The “may Þórr consecrate” formula appears

34 Bertell 2003, 68ff., 207f. 35 See e.g., Þrymsqviða st. 30: “ ‘Bring in the hammer to sanctify the bride, lay Mjǫllnir on the girl’s lap, consecrate us together by the hand of Várr’ ” (‘Berið inn hamar, brúði at vígia,/ leggit Miollni í meyiar kné,/ vígit ocr saman Várar hendi!’). See also Bertell 2003, 93ff. 36 Þá stóð Þórr at ok vígði bálit með Mjǫllni. Gylfaginning 49. See also how Þórr consecrated the goatskin with Mjǫllnir in the myth of Útgarða-Loki: tók hamarinn Mjǫllni ok brá upp ok vígði hafrstǫkurnar. Gylfaginning 44. 386 Chapter 10 in several Viking Age runic inscriptions.37 On the Glavendrup stone (DR 209) from Funen (tenth century), for instance, it says thus: “may Þórr consecrate these runes.”38 Another version of this formula may be seen on the Jursta stone (Sö 140) from Södermanland: “may Þórr protect [this monument].”39 On this stone a symbol of the hammer is also depicted. When the symbol of the hammer appears on the Forsa ring and the runic stones from Södermanland, they probably have the function of protecting the holy objects (the rings and the monuments) as well as the runic inscriptions. It is actually no big surprise that a Þórr’s hammer symbol appears associated with a ring. There is often a relationship between Þórr and “oath-rings” in several Sagas of Icelanders.40 It thus seems as if Þórr was related to the swearing of oaths, in a similar way to the god Ullr.41 With support from Irish annals it has been stated that a Viking king in Dublin had a ring which he called “Þórr’s ring” ( fail Tomair). It was captured by King Maelseachlainn in the year 994.42 In my opinion, the fire-steel shape and the Þórr’s hammer image indicate that the Forsa ring had a Viking Age background and that it played an essential role in the cult of Þórr. There are additional arguments from the history of religions which may corroborate that the Forsa ring was a pre-Christian cultic object. In what follows, I will try to shed new light on the religious and ideological function of the Forsa ring by focusing on the context where this ring and other similar ritual rings may have appeared. I will particularly focus on the ritual space where the rings may have appeared.

37 Hultgård 1998, 726ff. 38 Þōrr vīgi þāsi runaR. DR 209. A similar formula is also attetested in DR 110, DR 220 and Vg 150. 39 siþi þur eller þur siþi. See Hultgård 1992. Henrik Williams (1996a, 302ff.) has, however, suggested that siþi may be interpreted as a noun ‘seiðr man’ and that the inscription is polemic. A Christian is here attacking, in runic cryptography, the foremost among the pagan cultic figures in the ancient Scandinavian religion with the expression: “Tor [är] en sejdkarl.” Anders Hultgård has however rejected William’s suggestion and argued that such inter-religious polemics never occur in other Viking Age runic inscriptions. See Hultgård 1998, 734. Marco Bianchi (2010, 125) has presented an entire new reading of the sequence: i siþiuþu ī Svēþiūðu. If it is correct, both Hultgård’s and Williams’ interpretations must be rejected. 40 See e.g. Eyrbyggja saga 4, Ísl. Fornr. 4, 6–10; Kjalnesinga saga 2, Ísl. Fornr. 14, 7. 41 See Nordberg 2006a, 407. 42 See e.g. Marstrander 1915a; 1915b; de Vries 1956–57, § 272. The text is quoted in Kabell 1975. Whether Tomair may be related to the god Þórr is somewhat debated. See Kabell 1975, 38f. Rulers, Rings and Rituals 387

10.2 “Oath-Rings” and hof Sanctuaries in the Sagas of Icelanders

It was Stefan Brink who first suggested that the Forsa ring originally was an ancient “oath-ring” which was kept at the general assembly and cultic site of Hög in Hälsingland.43 When making this interpretation he supported his argumentation with information from the Sagas of Icelanders. In these sagas a certain type of sanctuaries are described, called hof. In connection with them there were officials known as goðar. During the sacrificial feasts these goðar appeared as some kind of cultic leaders. One of the sacred attributes used by these officials was a ring (stallahringr, hringr or baugr), which the goði either carried in his hand or wore on his arm during the religious and judicial meetings. Otherwise this ring was kept on the platform or altar (stalli, stallr) of the hof sanctuary. The ring was sometimes used during the sacrificial rituals. It was reddened in the blood of the animal which had been sacrificed at the sanctuary. This ring was also used in judicial contexts. Oaths, for instance, were sworn on this ring. They were called baugeiðr or lǫgeiðr. The most important description of “oath-rings” may be found in the Úlfljótslǫg (the text is paraphrased in ch. 5 above). According to this text, a ring weighing two ounces or more had to lie on the ‘stalli’ in every chief’s hof building, and every chieftain (goði) had to wear this ring on his arm at all the assemblies he headed, having first reddened it in the blood of a neat which he himself had sacrificed there. Every man who was there to transact any business, as by law provided by the Court, first had to take an oath on that ring and name for the purpose two or more witness in evidence. He was to say:

I take oath upon the ring, a lawful one so help me Freyr and Njǫrðr and the Almighty God, to this end that I shall in this case prosecute or defend or bear witness or give award or pronounce doom according to what I know to be most right and most true and most lawful, and that I will deal lawfully with all such matters in law as I have to deal with while I am at this assembly.44

43 Brink 1996b. 44 “. . . at ek vinn eið at baugi, lǫgeið; hjálpi mér svá Freyr ok Njǫrðr ok hinn almáttki áss, sem ek mun svá sǫk þessa sœkja eða verja eða vitni bera eða kviðu eða dóma, sem ek veit réttast ok sannast ok helzt at lǫgum, ok ǫll lǫgmæt skil af hendi leysa, þau er undir mik koma, meðan ek em á þessu þingi.” Ísl. Fornr. 1, 315. 388 Chapter 10 altar or platform in the middle of the room, called stalli. On this platform lay an open ring,45 weighing twenty ounces,46 upon which people had to swear all their oaths. It was also the business of the hofgoði to wear this ring on his arm or hold it in his hand at every public meeting. Similar information appears in Kjalnesinga saga 2, which mentions that a ring made of silver was placed on the platform in Þorgrímr goði’s hof sanctuary at Kjalarnes. When there was an assembly the hofgoði had this ring in his hand. People also swore oaths on this ring at all litigations.47 Víga-Glúms saga reports that the chieftain (goði) Glúmr was supposed to swear an oath upon a ring, affirming that he was innocent of the manslaughter of Þorvaldr krókr.48 This had to be done at three hof sanctuaries in Eyjafjǫrðr. When the men were assembling at Djúpadalr, six men entered the hof. Then the saga says in general terms that whoever had to take the “hof-oath” placed his hand on the silver ring, which was stained red with the blood of the cattle sacrificed, and which was supposed to weigh not less than three ounces.49 Then Glúmr spoke and made his oath:

I name Ásgrimr to bear witness, and Gizurr in the second place to bear witness that I take the “hof-oath”, on the ring, and I say it to the Æsir-god. When Þorvaldr krókr got his death-blow, I was far away and did not take

45 That some symbolic (oath-)rings had to be open is attested elsewhere. A ring was discovered at the small cultic house in Järrestad, in eastern Scania, which is located beside a multifunctional hall (Söderberg 2005). This ring had a diameter of 9.5 cm. It was forged and made of round-iron and it had a secondary made break into it. This feature could be associated with the Icelandic “oath-rings” which could not be closed and which had to be carried by the goðar during sacrificial ceremonies. It could also be associated with the ritual restrictions which the Roman priest of Jupiter, the flamen dialis, had to follow. He was not allowed to carry closed rings. Aulis Gellius, Noctes Atticae X, 15, 7. In this context I must also refer to the interesting Pietroassa ring found in Romania and usually dated to the fourth century. Those parts of the ring which have been preserved have a runic inscription. According to the runologist Helmer Gustavson (2004), it says “Goternas odal är helig”. In the nineteenth century this ring was stolen from the museum of Bucharest and cut into pieces. There is a drawing of it from 1867 indicating that it was an open ring, exactly like the Icelandic “oath-rings”. 46 . . . ok lá þar á hringr einn mótlauss, tvítøgeyringr . . . One ounce is approximately 27 grams. The ring thus weighed around half a kilo. 47 Á þeim stalli skyldi liggja hringr mikill af silfri gerr; hann skyldi hofgoði hafa á hendi til allra mannfunda; þar at skyldu allir menn eiða sverja um kennslamál ǫll. Kjalnesinga saga, Ísl. Fornr. 14, 7. 48 Víga-Glúms saga 25, Ísl. Fornr. 9, 85f. 49 Sá maðr, er hofseið skyldi vinna, tók silfrbaug í hǫnd sér, þann er roðinn var í nauts blóði, þess er til blóta væri haft, ok skyldi eigi minna standa en þrjá aura. Ísl. Fornr. 9, 85f. Rulers, Rings and Rituals 389

part in that action, when the point and edge was reddened with blood. Now let those men who are skilled in such matters, and who stand by, look to my oath.50

The same oath was also sworn at the hof buildings of Gnúpafell and Þverá. Since Glúmr committed perjury, he was later forced to leave his homeland at Þverá. Also interesting is the information in Droplaugarsona saga 6, which says that the magnate (mikill maðr) Sveinungr concluded an oath at the “altar- ring” together with two men, who probably were witnesses: Vann þá Sveinungr eið at stallahring ok tveir menn með honum, . . .51 The compound expression stallahringr indicates that the “oath-ring” had a special relationship to the stalli/stallr, i.e. the platform or altar, which was situated in the hof sanctuary. The expression vann eið at stallahring also appears in Eyrbyggja saga 16 in connection with a goði.52 These descriptions in the Sagas of Icelanders have been debated. The Danish archaeologist Olaf Olsen argued that they do not say anything at all about pre- Christian religion and ancient customs in Iceland.53 Neither was Úlfljótslǫg, according to him, based on ancient tradition. The oath formula, for instance, with the expression “the Almighty God” (hinn almáttki áss), gave a Christian impression.54 Aage Kabell argued in the same vein as Olsen. He rejected the historical value of the descriptions of the oath-rings in the sagas and claimed that they were made up by imaginative medieval scribes.55 Kabell also felt that the other texts referring to oath-rings in some way were corrupt. This distrustful attitude towards the historical source value of the Sagas of Icelanders and Landnámabók has been a “leitmotiv” in twentieth-century saga research. In principle all information about pre-Christian religion has been rejected by some scholars. In my opinion, this critical attitude has been pushed too far,56 because there is information in the more direct sources which may confirm some of the information in the sagas. There are, for

50 . . . ek nefni Ásgrím í vætti, annan Gizur í þat vætti, at ek vinn hofseið at baugi, ok segi ek þat Æsi, at ek vark at þar ok vák at þar ok rauðk at þar odd ok egg, er Þorvaldr krókr fekk bana. Líti nú á eið, þeir er spekimenn eru ok við eru staddir. 51 Droplaugarsona saga ch. 6, Ísl. Fornr. 11, 153. 52 Eyrbyggja saga ch. 16, Ísl. Fornr. 4, 30. 53 O. Olsen 1966. 54 O. Olsen 1966, 49. 55 Kabell 1975, 33. 56 Several scholars have objected to the hyper-critical position in research, see e.g., Brink 1996b, 49f.; Jón Hnefill Aðalsteinsson 1998; Jón Viðar Sigurðsson 1999, 35ff.; 2011. 390 Chapter 10 instance, some passages in the Eddic poetry where the custom of swearing oaths on rings may be confirmed. Hávamál 110 mentions a baugeiðr ‘ring- oath’, and Atlaqviða in grœnlenzka 30 states that oaths were sworn on Ullr’s ring (eiða opt um svarða . . . oc at hringi Ullar). There is also a note in the Anglo-Saxon Chronicle dated to 876 which is of great interest in this context. It states that western Scandinavian Vikings swore oaths on a sacred ring and promised that they would not continue to ravage in the realm of King Alfred.57 This source must be regarded as historically reliable. According to some reliable Old Norse texts, oaths were also sworn on other sacred objects, such as the edge of the shields, but also on the side of a ship and the back of a horse. The Eddic poem Vǫlundarkviða st. 33, for instance, states thus:

Eiða scaltu mér áðr alla vinna, at scips borði oc at scjaldar rǫnd, at mars bægi oc at mækis egg, . . .58

The round shield could perhaps be related in a symbolic sense to a ring. Sometimes oaths were sworn on sacred stones.59 In Hænsa-Þóris saga 12 a wedding feast was to be celebrated in western Iceland. When the guests were sitting at the tables, the bride-groom Hersteinn ran up and went to a stone, which was situated inside the banqueting building. He put one of his feet on it and declared: “I make this promise, that before the Alþingi has passed in summer, I will have Arngrímr goði deemed outlawed . . .”60 After Christianization in Iceland, oaths were probably sworn on the Holy Bible. In Njáls saga 143, for instance, we read that after the Christian Law was accepted at the Alþingi, Eyjólfr appointed witnesses “and swore an oath on the Book, a ‘lawful-oath’, and declared before God” when he put up the case for defending Flosi at the

57 Her hiene bestæl se here into Werham Wesseaxna fierde, and wiþ þone here se cyning friþ nam, and him þa aþas sworon on þam halgan beage, þe hie ær nanre þeode noldon, þæt hie hrædlice of his rice foren. This text is quoted by Kabell 1975, 36. In the Latin translation of this text, made about 100 years later, the ring is called sacrum armillum. See the Peterborough Manuscript 876 [875], The Anglo-Saxon Chronicle, (ed.) Swanton 1997. 58 “ ‘First you shall give me all these oaths,/ by the side of the ship and the rim of the shield,/ the back of the horse and the edge of the blade.” Trans. Larrington. See also Beck 1986, 537f. 59 See Habbe 2005, 141f. 60 “Þess strengi ek heit”, sagði hann, “at áðr alþingi er úti í sumar, skal ek hafa fullsekðat Arngrím goða . . .” Rulers, Rings and Rituals 391 assembly.61 Grágás records that oaths were sworn on the cross: “I name witnesses to witness that I swear an oath on the cross, a lawful oath, and declare before God . . .”62 These customs may be seen as rituals of continuity, which had their origins in the pre-Christian context, but were accommodated after Christianization to the new circumstances in Iceland.

10.3 Ritual Rings and Sanctuaries in Svetjud

There is thus good written evidence from Iceland that ritual rings appeared in ceremonial buildings during the Viking Age and that different types of rituals were related to them. But is it possible to relate “oath-rings” or other ritual rings with ceremonial buildings also in eastern Scandinavia and particularly in Hälsingland? According to Carl Löfving, there is nothing to indicate this.63 Löfving argues that traits of more specific cultic buildings at pagan sanctuaries are not known in Scandinavia. The pre-Christian cult was performed outdoors or in the chieftain’s hall. This is a problem for those who argue that the Forsa ring should be dated to the Viking Age, according to Löfving, since it is impossible to explain how the ring can have been preserved in such good shape during this long period.64 In my opinion, there is much support for the theory that the Late Iron Age halls found at several places in Scandinavia really were regarded as protected religious spaces.65 It was in such spaces the ceremonial banquets were celebrated. As noted above, archaeologists have shown that these halls often have continuity over time, which indicates that they could contain objects of symbolic value for a long period. At the assembly place of Hög a large Iron Age building has recently been found, as well as an outdoor sacrificial place dated to the Viking Age.66 We will now turn to some other Late Iron Age ceremonial buildings and sanctuaries in the Mälaren area and see if it possible to find ritual rings at these places. These buildings and places are mainly attested in the archaeological record. The following survey is far from complete. I will only highlight some

61 . . . at ek vinn eið at bók, lǫgeið, ok segi ek þat guði . . . Ísl. Fornr. 12, 392. 62 . . . nefna ser vatta. Nefni ec i þar vætti. At ec vin eið at krossi log eið ok segi þat Guþi. . . . GrK ch. 25. See also GrK ch. 41–42. Quoted from Habbe 2005, 144 and 241. 63 Löfving 2005, 113; 2010. 64 Stefan Brink (2010) has pointed out, on well-founded grounds, several flaws in Löfving’s way of reasoning. 65 See Herschend 1993; 1997; 1998; 2009. 66 Personal communication from the archaeologist Katarina Eriksson in November 2011. 392 Chapter 10 representative examples. Two types of rings will be discussed: “Þórr’s hammer rings” and “amulet rings”. There are no general definitions for them, but archaeologists usually make some distinctions. Krister Ström defines Þórr’s hammer rings as “en cirkelformad ring av järnten, sluten med ett öppningsbart lås och med uppträdda hängen av järn”.67 In addition to this the “Þórr’s hammer rings” usually appear in graves.68 Gunnar Andersson states that the size is important when making a distinction between these rings: “På boplatser förekom- mer också mindre amulettringar som ibland har morfologiska likheter med de större torshammaringarna.”69 As Andersson states, there are no great mor- phological differences between these types of rings. It is only the size that can determine whether a ring should be categorized as a “Þórr’s hammer ring” or an “amulet ring”. In what follows, I will occasionally include both these types of rings in the analytic concept of “ritual rings”, since they are so hard to keep apart. Specific attention will be paid to fire-steel-shaped rings, since the Forsa ring had this shape.

10.3.1 Borg in Östergötland Borg in Östergötland is a very interesting place in this context, since it is located so close to the border of the Svear. At this place a Viking Age building (5 × 7 m) (house 5) was discovered, which was interpreted as a cultic house (see fig. 2). Close to the eastern long wall there was a foundation built of strong flat stones. It was located just in front of the alley which divided the room into two sections. Just to the south of the foundation two amulet rings were discovered. Elsewhere this building was empty. It has been suggested that the flat stones could be interpreted as a pre-Christian altar, i.e. something equivalent to the structure called ON véstallr in the written sources.70 In addition, outside this building, in the courtyard, a total of 98 “amulet rings” were found below a hillock (see fig. 20). The building was placed on paving in the courtyard, where 75 kg of unburned animal bones were also found, from horses, dogs, pigs and sheep. These bones can probably be considered as remains of sacrificial animals. The rings found at this place often have a fire-steel shape which reminds us of the shape of the Forsa ring.71 Ann-Lili Nielsen writes: “All the amulet

67 K. Ström 1970, 1. It should be noted that the amulets on these rings do not always look like hammers. 68 Bäck et al. 2008, 45. 69 G. Andersson 2005, 63f. 70 Nielsen 1996, 98; Vikstrand 2001, 263. 71 For a complete investigation of fire-steel-shaped rings in the Mälaren region, see Jonsson 2006. Rulers, Rings and Rituals 393 rings found at Borg had been forged together from two separate parts and had thereby been shaped like fire-steels . . .”72 Some of them have pendants which resemble Þórr’s hammers and some have miniature axes. This could also be related to the decoration of the Forsa ring. Some of the rings have a diameter of 8 centimetres.73 According to Landnámabók (Úlfljótslǫg), the “oath-ring”, which appeared on the altar (stalli) in the hof sanctuary weighed at least two eyrir, i.e. 54 grams. According to Víga-Glúms saga, “oath-rings” had to weigh at least three eyrir, i.e. 81 grams. If this information is taken into consideration, we cannot rule out that the small rings which appeared at the “altar” foundation in the cultic house at Borg may have been some kind of oath or cultic rings of a type similar to those referred to in the Old Norse literature.

10.3.2 Helgö in Uppland Not far from the Late Iron Age hall and cultic building on Helgö, in foundation III, a fire-steel-shaped amulet ring was found (2174).74 On this ring there was an axe pendant. This ring had a diameter of 4.8 cm. According to Wilhelm Holmqvist, this was a symbol of the god Freyr.75 In my opinion, it could just as well have represented Þórr. Fragments of two fire-steel-shaped rings were found in foundation I, with a diameter of 4 cm.76 Perhaps there is also a fire- steel-shaped ring at the ceremonial building in foundation II, with a diameter of 5.7 cm (2433).77 Just below the bare rock, on the stony ledge and the triangular stone setting, many Merovingian Period and Viking Age iron amulet rings and Þórr’s hammer rings were found, together with miniature spears, fire-steel pendants, miniature shears etc. Most likely they were deposited there. Torun Zachrisson emphasizes the location of the finds, close to a rock (compare with Borg above, where the rings were also found close to a rock).78 She interprets them as cultic objects deposited at a place which may have been considered as a cosmic centre. Perhaps some rings were related to a ritual enclosure of this site. Holmqvist likewise associated these rings with cultic activities.79 Ritual iron rings were discovered at several other places on Helgö.80

72 Nielsen 2006, 382f. 73 Björkhager et al. 1997, 70ff. 74 Holmqvist & Arrhenius 1964, 85; cf. Arrhenius 2011, 35. 75 Holmqvist 1969, 175ff.; 1980, 62. 76 Holmqvist & Arrhenius 1964, 148. 77 Holmqvist & Arrhenius 1964, 148. 78 T. Zachrisson 2004a, 156; 2004b, 367; 2010, 84. 79 Holmqvist 1980, 56ff. 80 On these rings, see Arrhenius 1961, 147ff.; Lundström & von Heland 1964, 85ff; Hallinder 1970, 61ff. 394 Chapter 10

10.3.3 Lilla Ullevi in Uppland Also most interesting is the cultic site of Lilla Ullevi, in Bro parish, Uppland, where several Merovingian Period and Viking Age amulet rings and fire steels have been discovered.81 No ceremonial building was found there. Torun Zachrisson has recently suggested, on the other hand, that the structure, which the excavators identified as a “harg” (ON hǫrgr), should be interpreted as a symbolic gable of a hall.82 The total number of amulet rings was 65, of which 32 were complete. The size of the rings varied from 2 cm in diameter to 13 cm in diameter. Some of the large rings had smaller amulet rings or other amulets hanging on them. They had a clincher or a staple, perhaps used as a holder, so that the ring could be attached to a wooden object, a post etc. Most of them were detected at the structure called “harg” and at the “fence” (which the excavators called “the vébǫnd”), which demarcated the ritual space of the site. They were intentionally deposit in these areas. Five objects were interpreted as fire- steels. One of them was found at the “harg” structure. Most likely Lilla Ullevi was an important cultic site in central Uppland. The name Ullevi may support this assumption, as it could be interpreted as ‘the sanctuary dedicated to the god Ullr’.83 Perhaps some of the rings had a ritual and ideological significance in the public cult that was performed there.

10.3.4 Old Uppsala Both written sources and archaeological finds indicate that Old Uppsala was an important cultic site for the Svear during the Late Iron Age. Adam of Bremen, for instance, mentions that a cultic building (templum) was erected there (see ch. 4 above). To the north of the church a Merovingian Period ceremonial building has been discovered. Not far from it, Viking Age boat graves were found in the vicarage (Sw Prästgården). In one of the graves a Þórr’s hammer ring was discovered.84 The ring had a diameter of 16–17 cm and it could be opened at two places. On the ring there were several pendants, a coil, discs and a hammer. The ring was very rusty and found in three pieces. It was placed on the top of boat grave no. 1, just below a packing of stones. A similar Viking Age ring was found at Valsgärde, grave 4, just a couple of kilometres north of Old Uppsala. There was also a Merovingian Period Þórr’s hammer ring in grave 6. It is noticeable that these ritual rings appear at cultic sites where ceremonial

81 Bäck et al. 2008; Hållans Stenholm 2010. 82 Torun Zachrisson presented this interpretation at a conference in Reykjavík in November 2010. 83 Vikstrand 2010. 84 Nordahl 2001, 22f. Rulers, Rings and Rituals 395 buildings were most likely erected. The shape of the Þórr’s hammer ring of Uppsala resembles in one sense the Forsa ring, with its three-branched figure. But there are also some differences. The Þórr’s hammer ring of Uppsala was a grave gift, while the Forsa ring has been preserved in other contexts. It is possible, however, that the Þórr’s hammer ring of Uppsala in a previous phase played a ritual role in the context of the sanctuary of the site. During the summer of 2012 excavations about 300 metres to the east of the church discovered several hoards of amulet rings. Eighty amulet rings, including Þórr’s hammer symbols, were found. These finds emphasize the religious significance of this site.

10.3.5 Adelsö and Other Places with Ritual Rings in the Mälaren Area There are several other places where ritual rings, fire-steels and pendants have been found in the context of possible ceremonial buildings at Late Iron Age settlements in the Mälaren area. Under the floor of one of the long-houses, built on a foundation just beside the assumed banqueting hall at Adelsö, Uppland, a miniature of a fire-steel was found.85 Close to the hearth in the same house, three linked miniature sickles were found. These terraced buildings seem to have been in use from 700 to 900. The place appears to have been a royal site, but some cultic activities may have also been performed there. One fire-steel-shaped ring (approx. 7 × 8 cm), with three smaller fire-steel-shaped rings hanging on it, was also found at Adelsö in 1942.86 Information about the find context is missing, unfortunately. At the farm of Sanda (RAÄ 147) in central Uppland, three Viking Age fire-steels and four amulet-rings were likewise found in the context of a hall and an outdoor ritual space area, including an assumed equivalent to an ON hǫrgr.87 Most of the rings were found at the assumed hǫrgr. This farm was definitely an important cultic place in the local context during the Merovingian and Viking Periods. Late Iron Age amulet rings have also been found in the context of settlements at several places in Uppland and Södermanland.88 A Þórr’s hammer ring was discovered at a settlement called Viby gård (RAÄ 308) in Sollentuna parish, Uppland.89 It is quite

85 Brunstedt 1996, 22. 86 Jonsson 2006. 87 Åqvist 1996. 88 See, for instance, Alaby (RAÄ 164), Botkyrka parish, Uppland; Säby (RAÄ 167) and Valsta (RAÄ 165), Norrsunda parish, Uppland; Pollista (RAÄ 228), Övergran parish, Uppland; Mälby (RAÄ 327), Tillinge parish, Uppland; and Frövik (RAÄ 495), Fors parish, Södermanland. Survey in Bäck et al. 2008, 45ff. 89 Beronius Jörpeland & Sander 1996, 37. 396 Chapter 10 unusual for such a ring to appear outside grave contexts. At Sanda in Vansö, Södermanland, two large fire-steel-shaped “neck-rings” (approx. 16 cm in diameter) were found. In one of them there were holes for hanging pendants.90 Some of the fire-steel-shaped rings found in the Mälaren area were made of silver, such as the ones discovered at Eketorp, Edsberg, Närke.91 Such silver rings are more common on Gotland.

10.3.6 Häckelsäng, Hamrånge Parish, Hälsingland There is no reason to doubt that ritual rings played an important role at ceremonial buildings and sanctuaries in the Mälaren area. Most likely these rings were kept in these buildings or at other ritual structures. Probably similar buildings with ritual rings also existed in Hälsingland, i.e. in the province where the Forsa ring belongs. In this case we must find support from the place- names, for instance, names including the term hov. In certain contexts they may designate ceremonial buildings where ceremonial feasts were celebrated during the pre-Christian era. In some cases this term may also be interpreted as a designation of ground with a meaning ‘height, hill’.92 In Hälsingland there are two hov-names, one in Arbrå and one in Söderarla. Perhaps these places had some kind of ceremonial buildings during the Viking Age. Both places are located in the central parts of the Late Iron Age settlements. None of them have hills in the landscape which could have generated a name with the meaning ‘height, hill’. It is more reasonable to interpret these names in a cultic sense.93 There is another word in the place-names which may be even more interesting in the context of Hälsingland, namely the term vi (ON vé) ‘sacred place’. There are five certain vi-names in this province.94 Whether Byberg in Forsa parish should be derived from an old Viberg is uncertain. Both Ruthström and Brink felt that the inscription on the Forsa ring concerned a cultic place called vi since this term appeared in the inscription. According to Brink, the Forsa ring was kept at the central place of Hög, i.e. Kungshögen, which most likely was an important assembly place and cultic site during the Viking Age. A large Iron Age building has recently been discovered there (see above). Perhaps parts of the cultic and judicial activities took place inside it, or at the stone structure situated beside it. That ritual rings were kept at cultic sites in this region may be supported by the Häckelsäng ring, found in Hamrånge parish

90 K. Ström 1970, 40. 91 Jonsson 2006. 92 Vikstrand 2001, 253ff. 93 Brink 1990b. 94 Brink 1990b. Rulers, Rings and Rituals 397 in Gästrikland, not far from the border with Hälsingland (fig. 21).95 This ring was discovered in 1887 in connection with road construction. Since 1920 it has been kept at the Länsmuseet in Gävle. The ring is made out of bent iron rod, and it has a diameter of 30.8 cm. The ends are hook-shaped so they may grip each other. The inside of the ring has hook-bent iron points, alternately turned towards each other, alternately away from each other. One of them has a hole, which may have made it possible to put a ring into it. The man who discovered the ring remembers that he also saw other smaller rings beside the Häckelsäng ring. He also remembers a hammer made out of iron weighing about 1 kg. This is lost, however, but there is a drawing. According to the archaeologist Hanna Rydh, the ring should probably be dated to the Viking Age, and most likely it had a symbolic-religious significance. There are several Late Iron Age burial fields at Häckelsäng and in Hamrånge parish. At Hamrånge church, there is a place called Vi, which most likely is the assembly and cultic site of the settlement.96 It is not impossible that the ring played an important role at this site during the Viking Age. The situation at Hamrånge resembles the context of the Forsa ring, i.e. a ritual ring which appears at a cultic place called Vi. The survey of the archaeological find sites in the Mälaren area and its envi- rons clearly indicates that ritual rings and fire-steels often appeared at places which could be regarded as pre-Christian cultic sites or ceremonial buildings. Some of the rings with a fire-steel shape could perhaps be related to the cult of the god Þórr, for instance, the Häckelsäng ring and the Forsa ring. I agree with Brink that the Forsa ring may have been kept at a site in Hög parish, perhaps at

Figure 21 The Häckelsäng ring, found in Hamrånge parish in Gästrikland. Permission from Gävleborgs museum.

95 Rydh 1920; Bellander 1938, 131f.; 1944, 77f. 96 Vikstrand 2000. 398 Chapter 10

Kungshögen, which was both a sanctuary and an assembly place. At this sacred place the ring played an important role in ritual contexts, perhaps at sacrifices and when swearing oaths. In Norway too, Þórr’s hammer rings are quite common. A ring with amulets, for instance, has been found at Hilda, Innvik.97 Viking Age golden rings were also found in the “temple treasure” at Hoen Øvre Eiker, Buskerud.98 Other types of ritual rings appear in the context of sanctuaries in Trøndelag (see ch. 11 below).

10.4 The Custom of Swearing Oaths on Rings

The sources thus indicate that ritual rings appeared at the cultic sites of all three investigation areas, Iceland, Trøndelag, and Svetjud. I have previously argued that pre-Christian rulers protected and took care of such sanctuaries. This may be seen in reliable sources. Like the goðar in the Sagas of Icelanders, the *gudhar in Svetjud may have acted with judicial functions (see ch. 5 above). The persons mentioned in the inscription on the Forsa ring, Anundr in Tåsta and Ofegh in Hjortsta, may very well have been a kind of Viking Age *gudhar in Hälsingland, with both religious and judicial functions. In this sense these officials may have had functions comparable to the goðar mentioned in the Sagas of Icelanders. Maybe they carried the “oath-ring” in their hands during religious ceremonies, as a sign of dignity and religious power. A picture stone from Tängelgårda, Gotland, may display such cultic leaders/specialists, in a procession carrying “oath-rings” in their hands (see fig. 22). It seems as if these rings contributed to the construction of authority. But do we have any evidence from the Mälaren area indicating that oaths were sworn on such rings in the sanctuaries as the Icelandic sources report? The custom of swearing oaths (ON sverja eið, selja eiða) is well attested from all Germanic areas and cannot be conceived as a local Icelandic tradition.99 The origin of the word oath, ON eiðr (Goth aiþs; OHG eid; OE āð) is debated.100 According to Elof Hellquist, for instance, it is an old loan-word from OI ōeth,101 but if so it must have been borrowed by the Germanic languages at an early

97 Perkins 2001. 98 Østmo & Hedeager 2005, 174ff. 99 Cf. de Vries 1956–57; Frense 1982, 229ff.; H. Beck 1986; Köbler 1986. 100 See H. Beck 1986. 101 Hellquist 1957, 173f. Rulers, Rings and Rituals 399

Figure 22 A picture stone from Tängelgårda, Gotland. SHM. date.102 The verb ON sverja (OSw sværia; Goth swaran; OE swerian; OSax swerian), which is connected with the oath (ON sverja eið; OE āðswyrd), also seems to be old and spread in several Germanic languages.103 The custom of swearing oaths on ritual objects is also attested from all the Germanic area.104 One source reports that the Saxons swore oaths on swords.105 In Annales Fuldenses an unknown writer says that in the year 873 the Danish kings Siegfried and Halfdan sent an envoy to Emperor Leo (Ludwig) asking for peace. They met at the River Eider. There they concluded peace by swearing oaths on weapons.106 The passage is usually conceived as having a high source value.107 When such oaths were taken, rings or ring swords were customarily used as ritual objects. I have already referred to the information in the Anglo-Saxon Chronicle about the Vikings who swore oaths on a holy ring in 876. Also in the ancient Greek world, oaths and other contractual rituals were sworn on rings and other sacred objects. The gods were often evoked as witnesses. Such rituals were often circumscribed by a complex of actions, including animal sacrifices and libations (σπονδή sg.).108 Often the ring was broken into two pieces. These pieces were then given to the partners of the contract.

102 Eddic poetry talks about the ‘oath-breaker, perjurer’ (eið-rofa). See e.g. Brot af Sigurðarqviðo 16. Helreið Brynhildar 5. In Hávamál st. 110 a baugeiðr ‘ring-oath’ is mentioned. Atlakviða in grœnlenzka 30 narrates that an oath was sworn on Ullr’s ring. 103 Cf. H. Beck 1986. 104 H. Beck 1986; Frense 1982, 229ff.; Jacob Friesen 1986; Köbler 1986; Habbe 2005, 134ff. 105 Köbler 1986, 540. 106 Iurabant etiam iuxta ritum gentis suae per arma sua, . . . Annales Fuldenses, MGH, Scriptorum, Tomus I, 386. 107 Frense 1982, 235. 108 Burkert 1985, 250ff. 400 Chapter 10

It was called σύμβολον (symbolon) ‘sign’ and when used in plural ‘the halves of a bone or a coin, which two persons broke between them, each keeping one piece’.109 This word is related to the modern term “symbol”. The ritual oath testimony was usually performed in the temple.110 In ancient Scandinavian contexts too, it seems as if the ring was sometimes broken in pieces during contractual rituals.111 Most likely the swearing of oaths on rings kept at sanctuaries was practised in Svetjud as well. The archaeological finds made at cultic sites such as Borg, Helgö and Lilla Ullevi, as well as written sources, support such assumption. The Swedish medieval provincial laws report that oaths were sworn at the assembly place during the inauguration of kings.112 In connection with these contractual rituals, certain oath formulae were recited. One possible pre-Christian oath formula has been preserved in Äldre Västgötalagen.113 The section Af mandrapi states: sva se mær guþ hol. This is conventionally translated “may God be gracious to me”.114 Leopold Läffler noted that in a Chris tian context the word “God”, OSw guþ (nom. masc.), should take an adjective hulder and not hull. In the form hull the adjective hulder (nom. sg.) appears in three cases: (1) nom. sing. fem.; (2) nom. pl. neut.; (3) acc. pl. neut. The third case must be rejected, as guþ is the subject. This word, guþ “God”, is in general masculine. But when meaning “pagan god”—which often appears in the plural—it is frequently attested in the neuter form.115 The first case may therefore also be rejected. Äldre Västgötalagen’s sva se mær guþ hol should thus

109 Liddell & Scott 1888, 759f. See also Braarvig 1997; Gilhus & Mikaelsson 2003, 125. 110 Burkert 1985, 253. 111 The Eddic poem Rígsþula st. 23 mentions that the married couple Karl and Snør split a ring, bauga deildu. According to Åke V. Ström this was a traditional wedding ritual. Å. V. Ström 1975, 241. The ring perhaps appears in contractual rituals related to wedding contexts in other Eddic poems, as well as in Kings’ Sagas. See e.g. Skírnismál st. 21–22; Hávamál st. 110; Oddr Munk’s Óláfs saga Tryggvason ch. 29, Ísl. Fornr. 25, 226f.; and the discussion in Steinsland 1991, 148–153. 112 The custom of swearing oaths (OSw eþæ swæriæ) is attested in medieval Swedish provincial laws. In Äldre Västgötalagen (1225) it is stated in Rättlösebalken that the coming king must swear to be faithful (trolekæn sværiæ) to all the Götar, and obey their law when he comes to All-Göta Assembly. ÄvgL. R 1, SSGL 1, 36f. Upplandslagen (1296) mentions that the people who followed the king during the ceremonial royal progress called Eriksgatan must swear oaths (eþæ swæriæ) that they will be loyal to the coming king. UL Kg 1, SSGL 3, 52ff. 113 Läffler 1905. 114 ÄVgL, Af mandrapi ch. 1, §2. SSGL 1, 10f. Cf. Frense 1982, 256–259. Holmbäck & Wessén 1979 (1933–46) Vol. 5, 16–18. 115 Cf. Fritzner 1972, Vol. IV, 134. Rulers, Rings and Rituals 401 be translated “may the [pagan] gods be gracious to me”. This interpretation may be supported by the fact that younger law texts such as Magnus Erikssons Landslag have sua se mik guþ hulder. The formula in Äldre Västgötalagen shows that we may sometimes have survivals of oath formula from the pre-Christian period in eastern Scandinavia.116 Perhaps this formula should be related to the oath formula mentioned in Úlfljótslǫg, which was spoken over the ring in the hof sanctuary: hjálpi mér svá Freyr ok Njǫrðr ok hinn almáttki áss (or áss hinn almáttki).117 In Víga-Glúms saga a similar formula is attested in the context of a hof and a ring, . . . ok segi ek þat Æsi . . . Even if the examples from the sagas are somewhat uncertain, they suggest that ritual oath formulae were sworn in all Scandinavia, in the context of ritual rings. Perhaps the enigmatic introduction to the Guta saga should also be related to such rituals and oath formulae:

Alt ir baugum bundit. Boland al þitta varþa, ok faum þria syni aiga.118

It has been emphasized that the poetic expression “everything in ring is bound” should be related to a religious-judicial context.119 In this chapter I have argued that ritual rings really appeared in Late Iron Age sanctuaries and ceremonial buildings in both Iceland and Svetjud. Most likely they appeared in similar contexts in Trøndelag too, although I have only occasionally referred to that area in this chapter. These rings were used by political leaders in ceremonial contexts, for instance when making sacrifices and in religious processions, but also during contractual rituals, such as oath-swearing. These rings were thus essential media or symbols of status and sovereignty. I have also argued that the Forsa ring from Hälsingland may have played an important role in religious and ritual contexts as an “oath-ring”. The small three-branched figure on the ring was interpreted as a Þórr’s hammer. Together with the form of the runes and the language of the inscription, it indicates that the ring had a Viking Age background. The fire-steel shape of this ring likewise points in the same direction. Like the western Scandinavian goðar, eastern Scandinavian cultic leaders or religious specialists may have

116 Sjöholm 1988, 259, is critical of Läffler’s argument. 117 Landnámabók H 268, Ísl. Fornr. 1, 314f. and note 6. Critically considered by O. Olsen 1966, 34ff. Cf. Kabell 1975. But see also Brink 1996a. 118 “Everything in ring is bound. Inhabited this land shall be; we shall beget sons three.” Guta saga, (ed.) Peel, 2f. 119 See Blomkvist & Jackson 1999. 402 Chapter 10 used such rings in ceremonial contexts, at sacrifices and when swearing oaths. These rings were kept at the sanctuary or the ceremonial building, which was protected by the cultic leaders. It is possible that the Forsa ring was kept at such a place before it ended up on the door of Forsa church. Most likely some religious specialists, cultic leaders and men skilled in law used it during the pre-Christian assemblies at Hög as a sign of dignity. Perhaps it was Anundr of Tåsta and Ofegh of Hjortsta who played this important role in the Viking Age community of Hälsingland.

The source investigation of ceremonial buildings and religious strategies for rulership in Part 3 clearly indicated some similarities between the three regions, the Mälaren area, Trøndelag and Iceland. Evidence indicates that general political-judicial leaders, chieftains or rulers, owned, controlled or managed the ceremonial buildings and sanctuaries in all three areas. One driving force behind this situation was related to economic factors. The farmers brought tribute to the orginizers of the religious feast at the sanctuary, i.e. the rulers. As compensation they could take part in public cult and gain metaphysical protection. Ceremonial buildings and sanctuaries were, however, not only sources of income for the rulers. They were also arenas for expressing power. In all three regions, these buildings were often seen in strategic locations adjacent to transport routes. They were often built on terraces or in topographical positions expressing an elevated status. The majestic impression this gave was also reinforced by the architecture and size of the building. The halls were also impressive inside, with open rooms and high ceilings. There were also places in the hall that were specially destined for the rulers and magnates. During the rituals, they would sit in the high-seat, looking down on those who sat in the long benches. Sources indicate that these rulers and magnates controlled the cultic actions taking place at the hall. During ceremonial feasts, he could manifest his generosity and express his wealth in a concrete way. The political leaders from all three regions thus used different types of public rituals in connection with these ceremonial buildings, in order to gain legitimacy and power, such as banquets, gift-giving, and judicial oath ceremonies (including the use of oath-rings). By means of certain paraphernalia, such as the high-seat, cultic images and oath-rings, they were able to link themselves closely to the gods and the mythical world. The investigation has revealed that certain cultic places in the three regions were based on a common model. Adjacent to the hall there was sometimes a smaller, more specific cultic house and an outdoor place of worship. These cultic places were sometimes marked with centre symbols, such as a tree or Rulers, Rings and Rituals 403 a spring/well. This model may correspond to a mythical structure; Valhǫll, Yggdrasill and Urðarbrunnr. It is possible that a micro-macrocosmic relationship existed at these sites, which may have been used to legitimize power. When the ruler appeared in the hall or outside it, he was in a space that had mythical dimensions. According to the basic hypothesis of the present study, however, the political-structural differences in the three regions also created a regional diversity concerning ceremonial buildings, and strategies for gaining legitimacy. Ritual and symbolic strategies related to such buildings in the centralized societies of Svetjud and Trøndelag were different from those found in Iceland, where power was split among many hands. In the following Part 4, evidence supporting this hypothesis will be presented and discussed.

Part 4 Regional Diversity

∵

chapter 11 Golden Objects as Media of Power

Sometimes objects of gold appear in the context of Late Iron Age ceremonial buildings in some parts of the investigation areas, i.e. the Mälaren region and Trøndelag. It has been argued that some of these objects reflect mythical images or religious symbolism. Since gold also was an expression of ruling power, we may assume that these objects may be related to a religious- ideological strategy for gaining authority. Some of these objects only appear during the Migration Period and the Merovingian Period and thus could not be expected to be seen in Iceland. I will nevertheless argue that the golden objects with a presumed religious meaning or symbolism may have been an important element and medium related to some specific religious strategies which were exclusive to the very upper stratum in Scandinavia, i.e. the noblest royal families in Denmark, Norway and Sweden.

11.1 Gold Foil Figures

One of the most important testimonies to the religious significance of the ceremonial buildings in Late Iron Age Scandinavia is that gold foil figures (Sw guldgubbar) often appear in relation to them. These foils include a type of small images that have been associated with the mythical sphere. They often appear close to the presumed high-seat, which suggests that these images not only had a religious significance, but also had implications of rulership and power. They have been discovered in hall buildings at several central places in the Mälaren region and Trøndelag, but none have so far been found in Iceland.1 The reason for this most likely has to do with chronology. The gold foil figures are usually dated to the period 550 to 800, so when Iceland was colonized c. 870 these objects were probably not produced anymore.2 In my opinion there may have been another reason why they appear so frequently in the

1 The most impressive find of gold foil figures has been made at Sorte Muld in Bornholm. More than 2,700 gold foil figures have been discovered there since 1985. Watt 1991; 2004; 2007. 2 It has been suggested, however, that even if no gold foil figures were produced after 800 they may have been used during the Viking Age. The investigations at Slöinge in Halland and Borg in Lofoten indicate this. See Lundqvist 1996; 1997; 2003; Munch 2003a and 2003b; Carlie 2004, 174ff.; Watt 2007, 133.

Mälaren region and Trøndelag. Some of these foils may have been related to a specific religious strategy of the highest-ranking royal families in Scandinavia in these regions. They may signal that the ruling couple were closely related to the divine sphere, i.e. they descended from the gods and/or mythical beings (see ch. 3 above). Since they were often placed close to the high-seat, they also indicated this intimate connection in a concrete and physical way. In what follows, I will first present the distribution of the gold foil figures, their find contexts and their appearance in the investigation area. Then I will take up previous interpretations, and finally present my own opinion about them. I will argue that some of these object with images represented scenes with deities/mythical beings, such as Óðinn, Þórr, Freyr, Skaði and Gerðr. Some of them may also represent scenes with more local mythical beings such as Hǫlgi and Þorgerðr. In the context of ceremonial buildings these objects underlined the very close genealogical relation which existed between the mythical world and the high-ranking ruling families of the Mälaren region and Trøndelag.

11.1.1 The Distribution of Gold Foil Figures, their Find Contexts and Appearance Gold foil figures have been found at several places in the Mälaren region and Trøndelag.3 At Helgö in Mälaren a total 26 foils were found. Most of them were related to Foundation I, where the compact cultic hall building IA was also situated (fig. 7A and 7B).4 No fewer than 12 gold foil figures were clustered within an area of 3 × 4 m close to the long wall and immediately south-west of the central hearth. Some of them were placed in post-holes, such as gold foil figure no. 1186, which appeared in post-hole no. 48. Since fragments of filigree glass claw and cone beakers were also concentrated in this area, this has been interpreted as the place of the high-seat.5 A similar structure may be seen at other central places of the Svear. The gold foil figure found in Krokek parish (fig. 23b), Östergötland, for instance, was discovered at a medieval royal manor called Svintuna, which was located on the route of the royal progress through the

3 See the distribution map in J. P. Lamm 2004. 4 K. Lamm 2004, 51ff. 5 Herschend 1995; 1998, 30. Arrhenius (2011, 23) argues that some foils could also have been related to post-hole no. 49, which is situated close to hole no. 48. In her opinion, a “phallus pillar” was placed in this hole. In this area fragments of glass were also found. She suggests that the foils either decorated the “phallus pillar” or were placed in a glass beaker containing liquid and when floating around there a very vivid effect was created. Golden Objects As Media Of Power 409 kingdom, the Eriksgata.6 The figure at Svintuna was found in a Merovingian Period or early Viking Age building which has been interpreted as a hall.7 The archaeologist Arthur Nordén assumed that it was deposited in front of or beneath the place of honour beside the long southern wall.8 The foil depicts a pair standing close to each other. Two gold foil figures were also found at the monastery of the St John’s Order in Eskilstuna. During an excavation at this monastery Sune Zachrisson discovered them in a Merovingian Period or Viking Age manorial building, possibly a hall.9 Both of them depict a pair of figures. At this place we have a clear example of cultic site continuity.10 One gold foil figure (split into two pieces) was also discovered in connection with an aristocratic hall at Husby, Glanshammar, in Närke. This foil likewise depicts a pair of figures. A gold foil with a couple was also found at Ultuna, Uppland, in a grave.11 Gold foil figures have been found at other places in Sweden too.12 Most extraordinary are the finds of 113 foils at the central place of Uppåkra.13 Most of them were found in relation to the famous cultic building there. A total of 57 gold foil figures have also been found at the halls (House II and House III) of Slöinge, Halland.14 The gold foil figures of the Mälaren area all display a common motif,15 embossed double figure, often interpreted as a man and a woman (see fig. 23b and fig. 7b above). They are turned towards each other and they are very well dressed. The man is usually bare-headed, and the woman is sometimes wearing a mask. In many cases they are hugging and perhaps kissing each other. The man is usually standing to the left while the female is to the right. The overall motif is stereotyped, but there are variations in the details such as poses, hairstyle, clothing, or attributes, such as jewellery or symbols. In several cases these foils appear at buildings which have been assumed to be halls or cultic

6 The term Eriksgata refers to the ceremonial road that the medieval king travelled clock- wise to each provincial assembly, from Uppland to Västergötland and back again. During this journey he was acclaimed as king. 7 Nordén 1938; Herschend 1998, 184. 8 Nordén 1938, 228. 9 S. Zachrisson 1963 and 1998, 13f. Personal communication Torun Zachrisson. 10 See A.-S. Gräslund 1992. 11 See J. P. Lamm 2004, 65. 12 Gullmarsberg and Hög Edsten, Bohuslän; Slöinge, Halland; Gårdlösa, Ravlunda, and Vä, Skåne; Bolmsö and Brahekyrkan (Visingsö), Småland; Eketorp, Öland. J. P. Lamm 2004. Recently 29 gold foils were also found at Västra Vång, Blekinge. 13 J. P. Lamm 2004, 61f. 14 Lundqvist 1996. 15 See mainly J. P. Lamm 2004, 44ff. 410 chapter 11

Figure 23a Gold foil figures from Hauge-Tu, Rogaland. Permission from Siv Kristoffersen and the Museum in Stavanger.

Figure 23b Gold foil from Krokek, Östergötland. SHM. Golden Objects As Media Of Power 411 buildings. Sometimes they are found at the place where the high-seat is assumed to have been located. Gold foil figures have also been found at a couple of places in Trøndelag and northern Norway. At the remains of a Viking Age ceremonial building under the church of Mære 19 (or 23) gold foils were found (fig. 8b above).16 Couples were depicted on these plaques, with the motif of a man and a woman standing opposite each other. They varied somewhat in size, but most of them were about 1 square centimetre. Eight were undamaged, while the rest were buckled and with the impress indistinct. Of these, three can be abstracted as a distinct group. They were probably impressed with the same stamp. The others vary, but on all of them the man is shown with a short tunic and hair, whilst the woman has a long tunic and long loose hair. On four of the images the couple are holding a staff between them. On the majority the couples are surrounded by a pearl border. According to Lidén, the gold foil figures from Mære should be dated to the Viking Age. He based this dating on the clothes of the figures, as well as on the context of the finds. Most scholars argue, however, that the foils were made during the Migration and Merovingian Periods, but were perhaps still in use during the Viking Age.17 Gold foil finds have been used at other places in Trøndelag as well as elsewhere in Norway. Two gold foil figures (cut-out human figures) were found at Hemne (Hevne), in Sør-Trøndelag in the early nineteenth century. Unfortunately, the archaeological context of these finds is unclear.18 Gold foil figures have also been found along the Norwegian coastline, from the south to the north.19 The most interesting parallel to Mære in the Norwegian context is probably Borg at Vestervågøy.20 At this place five gold foil figures were found in the 14-metre-long salr room in the hall-building (fig. 8c above). All of them were found at a place interpreted as remains of the high-seat in the northern corner, not far from an 8-metre-long “fireplace area”, which ended in a round hearth to the north-east. Excavation of the northern post-hole (1) revealed that the post had been replaced twice. One of the foils came from

16 Lidén 1969, 15ff.; 1999, 47, note 17. 17 See e.g., Watt 2007. 18 J. P. Lamm 2004, 43; Brendalsmo 2006, 421. 19 For instance at Hauge (Klepp) in Rogaland, Borg, and Kongsvik in Nordland, and in the inner parts of the country at the central place of Hov, Vingrom Oppland, Borge, Østfold and Sandsvær, Buskerud. See J. P. Lamm 2004; Watt 1999; 2004; 2007. These finds have in general been made at settlements, but the figures found at Borge and Sandsvær appeared in graves. These objects too may have been used by the noblest families of Norway in the Merovingian Period and the Early Viking Age. 20 Munch 2003a and 2003b. 412 chapter 11 the top of this layer. Two were found just outside it. They were probably also connected with this post-hole. The last two figures were placed under a stone, deep down in the last phase of the post-hole. They were believed to have been placed there intentionally. All five gold foil figures depict the couple motif. Two of them were probably made with the same die, while the others were quite heterogeneous. One is also unique in that the couple is seen in semi- profile. Three of them are very tiny, only 7–8 mm high. Gold foils have also been found in southern Norway. On the farm of Hauge, Klepp, in Rogaland sixteen foils were discovered (fig. 23a). Some foils were also found at Hov, in Vingrom, Lillehammer.21 The ones at Hov were related to a possible cultic building.

11.1.2 Previous Interpretations The gold foil figures found in the Mälaren region and Trøndelag thus almost all display a common motif, of a double figure or a couple, often interpreted as a man and a woman. Several interpretations have been proposed, but most scholars see them as mythical and/or ritual representations. Magnus Olsen, for instance, suggested that the couples depicted on the 16 gold foil figures found at the farm of Hauge, Klepp at Jæren in Norway (fig. 23a) may reflect the mythical sacred marriage (hieros gamos), which according to the Eddic poem Skírnismál (For Scírnis) was celebrated between Freyr and Gerðr.22 He also stated that these figures were used as sacrificial objects in a fertility cult. He thus made a myth-and-ritual interpretation of these foils and their iconography. Hans-Emil Lidén followed the same vein as Magnus Olsen and interpreted the gold foil figures from Mære as some kind of votive objects which had been used as sacrifices in a fertility cult. Like Olsen, he argued that the couples depicted on the gold foil figures may reflect the mythical sacred marriage celebrated between Freyr and Gerðr.23 When investigating the gold foil figures at Helgö, Jan Peder Lamm likewise agreed with Magnus Olsen that these figures probably refer to the mythical tale of Freyr and Gerðr preserved in Skírnismál, but he also made some modifications.24 He suggested that the figures portrayed human beings in costumes and masks, who either participated in the religious ceremony or

21 See Lidén 1999, 43. 22 Olsen & Shettling 1909; Olsen 1909. A similar interpretation was made by Holmqvist (1960), who argued that the foils depicted “dancing gods”. 23 Olsen 1909; Olsen & Shettling 1909. 24 J. P. Lamm 2004, 46. Ratke & Simek (2006) are critical of the Freyr-and-Gerðr interpretation, since the poem Skírnismál reflects a high-medieval tradition, and thus no authentic pre-Christian myth. Cf. Ratke 2009. Golden Objects As Media Of Power 413 personified the gods. The disproportionate size and almost grotesque styliza- tion of the heads supports this interpretation.25 The historian of religions Gro Steinsland has tried to modify Olsen’s and Lidén’s interpretations of the gold foil figures.26 She toned down the fertility aspect of these foils and argued that the couple motif should be attached to religious ruler ideology prevalent in Scandinavia during the Late Iron Age.27 According to Steinsland, the images represent the peculiar marriage between a god and a giantess, also expressed in Viking Age skaldic poetry such as Háleygjatal and Ynglingatal (actually elaborated more in Ynglinga saga).28 The product of this odd marriage was a human son. This son represented the prototypical ruler: the king or the earl. In Háleygjatal the divine couple was constituted by Óðinn and Skaði. Their son was Sæmingr, i.e. the ancestor of the Earls of Lade. In Ynglinga saga Freyr and Gerðr had a son called Fjǫlnir, who was the first king in the Ynglinga dynasty. According to Steinsland, the motif on the gold foil figures reflected this peculiar mythical marriage. The prototypical marriage myth behind these pictures is the one between Freyr and Gerðr, found for instance in the Eddic poem Skírnismál. This myth also had ideological implications for the rulers, since the royal kin originated from these mythical parents. Steinsland argues, further, that when the gold foil figures with this motif were deposited at the high-seat of the ceremonial building in Mære the intention was to emphasize the status of the building and the family who controlled this house. Their function was to support the ruling power by means of the hieros gamos myth.29 She cautiously suggests that these objects may have been used in connection with the inauguration rituals of the ruler, but the question of the ritual contexts of these finds mostly remains open.30 In an article from 1996, Hans-Emil Lidén maintained his interpretation of the gold foil figures as sacrificial or votive objects in a fertility cult. He was also critical of Steinsland’s ruler-ideological interpretation of these finds.

According to available sources, Mære was never a chieftain site in the pre- Christian period. Neither the archaeological materials nor the written sources indicate that. Most likely the chieftain farm of the district was located at Egge, just outside Steinkjer, where we have a large burial field dating from the Early, Middle and Late Iron Age. According to Lidén, the chieftains of Egge, in collaboration with other chieftains of Trøndelag, controlled the ceremonial building at Mære and organized the sacrificial feasts at this site. This cultic site of Mære was thus an assembly place for many minor chieftains and local groups without a ruler’s farm.31 In his book from 1999 he also stated that his earlier interpretation of a high-seat in Mære must be rejected. Therefore Steinsland’s theory could hardly suit the historical context of this site, according to Lidén. The Norwegian archaeologist Gerd Stamsø Munch, in opposition to Lidén, has supported Steinsland’s general interpretation, when discussing the gold foil figures found at Borg in Lofoten. She added, however, that these foils should be interpreted as house offerings.32 This ritual action was performed in order to consecrate the room or this part of the building. She argued that gold foil figures with the couple motif usually appear at places which were both political and cultic centres in the Late Iron Age, and followed Steinsland’s theory that this motif on the foils was connected to ruler ideology. According to her, the scene on the gold foil figures reflected the hieros gamos between Óðinn and Skaði as seen in Háleygjatal, i.e. the union between the mythical parents of the Earls of Lade.33 She also suggested that the chieftains of Borg originally belonged to the Lade family, since they derived from Hálogaland. Therefore we may find gold foil figures at the hall of Borg. Recently it has been argued that the gesture language on these gold foil figures must also be taken into consideration when interpreting them.34 By relating the foils to late classical and medieval parallels in iconographic and literary sources, the Danish archaeologist Margrethe Watt, for instance, suggests that some figures indicate gestures with a legal or social background.35 The action of one member of the couple putting his hands around the other person’s wrist has parallels in the medieval literature and iconography. It refers to the judicial act of binding or attaching someone to someone else, as in a marriage ceremony. It therefore seems reasonable to interpret these figures as iconographi- cal reflections of marriage scenes. In that sense, Watt’s suggestion supports

31 Lidén 1996, 66f. 32 Munch 2003b. 33 This idea was also presented by Gro Steinsland 1991. 34 See e.g., Ratke & Simek 2006, Ratke 2008; Watt 2007. 35 Watt 2007. Golden Objects As Media Of Power 415

Steinsland’s interpretation. Watt argues, however, that the figures should not be related to any particular mythical being.36

11.1.3 Gold Foil Figures as a Strategy for Gaining Legitimacy Hans-Emil Lidén’s objection to Steinsland’s interpretation in the historical context of Mære is, in my opinion, not a reason to reject her theory in general. Even if this cultic site lacked a chieftain farm and was considered as a communal sanctuary, it is quite possible that some regional kings or earls frequently visited this place, since it was so important for the farmers in Inn- Trøndelag. This idea is actually attested in the medieval literary sources. King Hákon the Good and Earl Sigurðr, for instance, visited this pagan cultic site in the middle of the tenth century.37 It is quite possible that regional rulers, kings and earls sometimes also visited or even controlled this site earlier, during the Merovingian and Early Viking Period. Such rulers are mentioned in Old Norse traditions.38 Perhaps they used gold foil figures in order to gain legitimacy. It is quite possible that something similar took place at other chieftains’ sites in the Mälaren region and Trøndelag/northern Norway where gold foil figures appear. They may be seen as symbolic markers indicating that a member of a noble family (of divine descent) sometimes visited these sites and took part in the ceremonies which took place at the high-seats. In my opinion, Steinsland’s theory is well founded. She combines the evidence from the written sources with the archaeological finds in a convincing manner when proposing that the foils had an ideological function for rulers. Even if Watt is sceptical about relating the motif on the gold foil figures to any specific mythical pair, her interpretation of the gestural language may support Steinsland’s hieros gamos theory in general, i.e. that the couple on these figures are celebrating a marriage. The stereotypical expression may also indicate that the motif is mythical. There is also an argument supporting the view that the motif on these figures in this context sometimes represents the mythical couple Freyr and Gerðr, at least in eastern Scandinavia. The distribution of theophoric place-names in the Mälaren region indicates that the cult of the god Freyr was important in the south-western part of the area, i.e. south Fjädrundaland, south Västmanland and Södermanland (see ch. 14). This area

36 “Uden at være helt sikker, er jeg tilbøjlig til at mene, at guldgubber med parfigurer skal opfattes som ‘neutrale’, d.v.s. ikke nødvendigvis gengivende bestemte navngivne personer fra den nordiske mytologi.” Watt 2007, 144. 37 See Hákonar saga góða 18, in Heimskringla I, Ísl. Fornr. 26, 172. 38 On kings and earls in Sparbyggvafylki, during the Early Viking Age. See e.g., Heimskringla I, Ísl. Fornr. 26, 99, 164f.; Heimskringla II, Ísl. Fornr. 27, 241. See also Þórhaddr in 5.2.1 above. 416 chapter 11 corresponds quite well to the places where gold foil figures have been found in this region. The Rällinge statuette and the Lunda figures, which most likely represent Freyr, may also be connected to this area (see ch. 7 and 14). In my opinion, Steinsland’s interpretation of the foils is thus quite sound, as long as we relate these objects to this geographic area.39 The mythical wedding on these foils may thus sometimes be related to Freyr and perhaps Gerðr. It is possible, however, that the couple on the foils from Trøndelag represented another divine couple. Perhaps the couple on the foils from Mære could be related to Óðinn and Skaði, as suggested by Gerd Stamsø Munch and Gro Steinsland, since these mythical beings, according to Háleygjatal, were regarded as the divine parents of the Lade Earls. One could also relate the couple depicted on the gold foil figures to Þorgerðr Hǫlgabrúðr and Hǫlgi, who may have been considered as the original mythical ancestors of the noble kin from Hálogaland. This assumption presupposes that the Lade Earls had some influence on the sanctuary at Mære as early as the Early Viking Age. This idea must be considered somewhat uncertain. If gold foil figures were still used at the end of the ninth century this theory is at least possible.40 It is possible too that in an earlier period there may have been some other noble ruling families in Mære and Trøndelag, who regarded Freyr and Gerðr as their mythical ancestors. Later Óðinn and Skaði were regarded as the mythical parents of the Háleygja kin. As noted above, most scholars argue that the gold foil figures could be regarded as sacrificial objects, since they were found at places which may be considered as cultic sites.41 Margrethe Watt has a different opinion. She admits that these objects have been discovered in important settlement complexes, where religious ceremonies also took place.42 However, in her opinion they played a role as “temple money” in this context. In order to take part in the religious feasts and ceremonies the participants would have had to pay with these objects. In my opinion, this interpretation is somewhat doubtful. At Helgö, for instance, some foils were discovered in post-holes. This indicates rather that they were sacrificed or ritually deposited in connection with the foundation of the house or that they had ideological implications. The gold foil figure no. 1186 at Helgö, for instance, was found in one of the post-holes of the

39 There are some problems of source criticism connected to Steinsland’s theory, that Freyr and Gerðr were regarded as the parents of the Ynglingar, see ch. 3 above. 40 Some scholars have actually argued that gold foil figures were still used in the Viking Age. See Lundqvist 2003; Munch 2003a; Carlie 2004, 174ff.; Watt 2007, 133. 41 See e.g., M. Olsen 1909; Lidén 1969; Steinsland 1990; 1991. 42 Watt 2007. Critically considered by Ratke & Simek 2006. Golden Objects As Media Of Power 417 hall.43 Two gold foil figures from Borg in Lofoten were likewise found in post- holes.44 At the Migration Period hall of Gudme gold foils were found in the post-holes of the roof-carriers,45 and at the Merovingian Period and Viking Age hall (houses III and II) at Bergagård in Slöinge such figures were found in post- holes located in the north-west corner, perhaps the place of the high-seat.46 The relationship between the foils and the presumed high-seat in the hall of Helgö indicates that they had an ideological function for the ruler. The figures most likely were not used as visible ruler attributes, since they are so tiny. The figures there varied between 7 × 7 mm and 11 × 11 mm.47 These objects should rather be interpreted as ritual and symbolic objects, intended to reinforce the holiness of a certain place. In my opinion, they were placed at the high-seat in the hall as a concrete statement of the close relation that existed between the ruler and the divine world. The iconographic motif may also have emphasized the divine origin of the ruler, his family (i.e. the genealogical relationship), and the sacredness of his chair in the hall. The foils may thus have been used in an inauguration ritual of the new ruler or in connection with the ritual consecration of his ceremonial building or high-seat. In my opinion, gold foil figures found in the ceremonial buildings in the Mälaren region and Trøndelag (including northern Norway) may thus be seen as objects and symbols related to the noblest ruling families’ and royalties’ specific genealogical relation to the mythic beings.48 It is possible that this type of religious strategy never existed among the chieftains of Iceland. Therefore no gold foil figures have been found there. It is also possible, however, that the distribution of these archaeological finds reflects temporal variations in Scandinavian religious ruler strategies. The usage of gold foil figures in aristocratic contexts may have disappeared in the middle of the Viking Age. Since Iceland was first settled after 870, we therefore never find such objects there. On the other hand, some scholars argue that gold foil figures were still used at the end of the ninth century, even if the production of them ceased earlier.49 If these scholars are right, the interpretation that gold foil figures represent a more specific ideological strategy is much more convincing.

43 K. Lamm 2004: 51. 44 Munch 2003a, 249. 45 Thrane 1998. 46 Lundqvist 1996; 2003; Carlie 2004, 178f. 47 J. P. Lamm 2004, 46. 48 Steinsland 1991. 49 See Lundqvist 1996; 1997; 2003; Munch 2003a. Cf. Carlie 2004, 174ff.; Watt 2007, 133. 418 chapter 11

11.2 The Gold Bracteates at the Hall of Helgö

Like the gold foil figures, the gold bracteates appear in both the Mälaren region and Trøndelag, but not in Iceland.50 The bracteates are usually dated to the Migration Period (c. 450–550), and one would thus not expect to find them in Iceland. In my opinion these finds too may be related to the specific ruler- ideological strategy that existed in Svetjud and Trøndelag (and in southern Scandinavia). Sometimes they may display mythic motifs, which may reflect the rulers’ close relation to the divine sphere. Like gold foil figures, these objects may be found in relation to ceremonial buildings at the rulers’ sites. The provinces around the Mälaren region are not particularly rich in finds of gold bracteates. One exception is the hoard found at Söderby, in Danmark parish, with four identical B-bracteates (IK 176), five identical D-bracteates (IK 522) and the newly discovered large D-bracteate (IK 583).51 Three fragmentary gold bracteates (IK 270, IK 271 and IK 399) have also been found at the settlement of Helgö. One of them (IK 270) was very large with an original diameter of roughly 90 mm. According to Morten Axboe, these bracteates can be attributed to the fifth or the sixth century.52 Two of the Helgö bracteates were discovered at Building Group 2, Foundation I, associated with the compact cultic hall building IA. IK 270 lay by itself at some distance from the gold foil figures and the area where it has been suggested that the high-seat was situated. IK 271, on the other hand, appeared at the place where most of the gold foils were deposited. Since these bracteates are fragmentary it is almost impossible to interpret the motifs on them with certainty. Axboe proposed that IK 271 may be a Nordic C-bracteate, although he admits that this suggestion is conjectural. Fragments of a third bracteate (IK 399) were found in Building Group 3. The motif on them, when put together, can be identified as the neck, chest and foreleg of a C- or F-bracteate animal. The C-bracteates generally contain an image of a large human head above a four-legged animal, often interpreted as a horse. Other animals, such as birds and fish, may also be seen. It has been suggested that this type of image represents Óðinn in a mythical scene. Karl Hauck suggested that Óðinn here is curing Baldr’s horse, as mentioned in the Second Merseburg Charm.53 Other scholars have argued that these images represent Óðinn as a shaman riding on

50 More than 900 bracteates have been found in Sweden, Norway, Denmark, England and Germany. See Hauck 1985–1986; 1999; Andrén 1991; Axboe 2007. 51 J. P. Lamm et al. 1999; Axboe 2004. 52 Axboe 2004; 2011. 53 Hauck 1985, 99ff., 106; 2002, 81ff. Golden Objects As Media Of Power 419 his horse to the Other World, followed by his ravens, i.e. his helping spirits.54 It has been argued that the medallions had a specific religious-political function in society and were used by elite groups as expressions of power. The images on them thus represent not only the god, but also the ruler who saw himself as descending from the divinity. This mythic-cosmic notion may have been inspired by ideas occurring in the Roman emperor cult.55 Together with the archaeologist Anders Kaliff, I have argued that the motif on the C-bracteates has striking similarities to the iconography associated with the late classical and Roman god Mithras.56 The bracteates remind us of the scene where Mithras kills the bull. This does not mean that it is Mithras represented on them. The bull-killing motif may merely have inspired the artist, who transformed it into the Germanic style and world of ideas. This may be why the four-legged animal resembles both an ox and a horse. Perhaps the intention was to depict the known mythical motif of Óðinn and Sleipnir, or Óðinn and Baldr’s horse, yet the prototype, the Mithraic scene, still exerted an influence when it was made. For the purposes of this work I will only stress that the motifs of the bracteates found at Helgö may have comprised mythical themes. Perhaps the rulers at Helgö wore these gold bracteates around their necks during the rituals when sitting in the high-seat. These objects were more visible to the “audience” than the small gold foil figures. The two bracteates from Helgö may thus have played an ideological role for the rulers of this site during the Migration Period, perhaps connected to a mythic-ritual strategy. As far as I know, two bracteates have also been found in Trøndelag. One C-bracteate (IK 231) was found in Dalum in Sør-Trøndelag. Another interesting bracteate (IK 282) was found at Hov in Snåsa, Nord-Trøndelag. It is possible that this place may have been an ancient central place, with a ceremonial building designated ON hof.57 There is also another type of finds discovered in Trøndelag, which may be related to the iconography of the golden bracteates and gold foil figures, namely the golden medallions.

11.3 The Golden Medallion from Inderøy

Not far from Mære at Vika on Inderøy, in Inn-Trøndelag, a golden medallion was found in a grave in 1937 with a similar couple motif to the one seen on

54 Hedeager 1997a, 273f. 55 Axboe 1991, 191ff. 56 Kaliff & Sundqvist 2004. Critically considered by Fischer 2005; Düwel 2008. 57 Røskaft 2003, 72f. 420 chapter 11 the gold foil figures. It has been argued that this medallion is Germanic but belongs to a certain group of objects where the intention of the artists was to imitate Roman medallions from the end of the fourth century.58 There are in total thirteen examples of these objects in Scandinavia and five were discovered in Norway. A characteristic feature is that these medallions are decorated on both sides. The front of the Inderøy medallion displays a full-length portrait of a man to the right and a woman to the left, turned towards each other in profile. Each is holding one hand on a tree or a branch with twigs which is placed in between them. The man is holding a ring in his left hand, while the woman is holding a staff. The reverse of the medallion displays a male rider, who is also holding a ring in his hand. Kristen Møllenhus argued that this medallion probably was produced in Trøndelag, although it was made as an imitation of Roman portraits of the Emperor and the Empress.59 The tree between the couple represented “the tree of life”. The wreath in the hand of the male was interpreted as “the wreath of victory” belonging to the goddess Victoria, i.e. an attribute which is frequently used on portraits of the Emperor. Møllenhus never managed to trace the Roman original of this image, which is a weakness in his argumentation. The Inderøy medallion is also unique since it depicts the couple as a full-length portrait. On the Roman medallions only the head/bust of the Emperor is seen. Gro Steinsland has opposed Møllenhus’s interpretation of the medallion.60 According to her, we must relate the couple motif on it to the motif seen on the gold foil figures found at Mære. Both of them represent the hieros gamos motif, known from pre-Christian Scandinavian myth, cult and rulership. Steinsland admits that the medallion and the gold foil figures display differences in styles and forms, and perhaps also differences when it comes to function. Anyhow, the motif may have arisen from the same tradition of ideas. In my opinion, we may assume that the couple on the Inderøy medallion represent divinities. The style, the attributes, the poses and the fact that this full-length portrait is placed on gold indicate this. It is difficult, however, to relate them to any specific god or goddess/mythical beings. We do not know whether Óðinn and Skaði, Freyr and Gerðr, or even Hǫlgi and Þorgerðr Hǫlgabúðr were worshiped in Trøndelag during the fourth and fifth centuries. Since the medallion was discovered in a grave it most likely had another function than the gold foil figures. The latter finds have often been unearthed at

58 Møllenhus 1953; cf. Steinsland 1991, 157ff.; 2005, 407. 59 Møllenhus 1953. 60 Steinsland 1991, 158. Golden Objects As Media Of Power 421 ceremonial buildings at aristocratic settlements, and it seems as if they were related to the religious strategy of the ruler. It is nevertheless interesting to note that a mythical couple in general seems to have played a central role in the Late Roman or Early Migration Age cult in Inn-Trøndelag. In any case, the attributes of ring and staff on the Inderøy medallion may perhaps indicate that these figures are holding certain insignia or symbols, which may be related to a religious strategy for rulership.61

11.4 The “Golden Door Ring” of the hof Sanctuary of Lade

Besides golden medallions, bracteates and gold foil figures there were other golden objects in ancient Scandinavia which may have played a certain ideological function in the context of the uppermost elite and ceremonial buildings, namely the golden door rings of the hof sanctuaries. In the written sources we hear about a golden ring placed on the door of a sanctuary in Trøndelag. Snorri Sturluson mentions in his Óláfs saga Tryggvasonar that Earl Hákon had a large golden ring on the door of the hof sanctuary of Lade. During his mission in Norway King Óláfr Tryggvason visited this place: “Then he sailed into the fjord to Lade and had the hof sanctuary there dismantled and all the property and all the decorations removed from it and from the idol. He took a large gold ring from the hof gates which Earl Hákon had had constructed, whereupon he had the temple burned down.”62 It is also mentioned that King Óláfr had this great gold ring sent to Queen Sigríðr of Svetjud, as an expensive gift. It was discovered that the ring was not made out of gold and the queen felt that she had been fooled. It has been argued that this ring story was a late addition to the tradition, since it does not appear in Oddr Snorrason’s Óláfs saga Tryggvasonar, which is the major source for Snorri’s text.63 That ritual rings were placed on the doors of ceremonial buildings is, on the other hand, intimated in other more reliable and direct sources. At the Middle and Late Iron Age site of Uppåkra, outside Lund, a more specific cultic building was found beside a possible hall building. The cultic house was 13 × 6 metres and probably rather high (fig. 4). Several finds indicating ritual

61 Cf. Steinsland 1991, 160. 62 Síðan siglir hann inn á Hlaðir ok lætr brjóta ofan hofit ok taka allt fé ok allt skraut ór hofinu ok af goðinu. Hann tók gullhring mikinn ór hofshurðunni, er Hákon jarl hafði látit gera. Siðan lét Óláfr konungr brenna hofit. Heimskringla I, Fornr. 26, 308. 63 Kabell 1975, 35. 422 chapter 11 activities were discovered there.64 At this house an iron door ring measuring 15 cm in diameter was found in a post-hole (see fig. 24a).65 It was decorated with four beads (Sw vulster) and a staple made out of band-iron, which embraced the ring. Another iron ring was found at the house, which was 23 cm in diameter and also had four beads (see fig. 24b). Such beads are common on Romanesque church doors. The church-door ring from Delsbo, Hälsingland, had this kind of decoration (fig. 18).66 In general the Christian rings have only three beads, symbolizing the Holy Trinity.67 There is some information in the Old Norse literature about rings which were nailed to the doors of the pagan ceremonial buildings. That ritual rings were placed on the doors of specific buildings may also be attested from iconographic sources. The Sparlösa stone (Vg 119), from ninth- century Västergötland, is decorated with images and a runic inscription. On one side of the stone, at the top, there is an image depicting a building, which seems to be a hall or perhaps a cultic house (fig. 25). The most salient aspect of this building is a door with a great ring, which seems to be nailed there. Several scholars have argued that the door with the ring must have a ritual- symbolic significance.68 The assumed hall with the ring arouses associations with a stanza in the Eddic poem Rígsþula. When the god Rígr is out wandering in stanza 26 he comes to a hall (at sal) where the substantial couple Faðir and Moðir (i.e. the parents of Jarl) live. He walks towards the door at the southern part of the building: var hurð hnigin,/ hringr var í gætti. Researchers have interpreted these bars in different ways. Ursula Dronke, for instance, suggests this: “The door was taken down. The handle-ring was in the side-niche.”69 According to her, Rígr did not have to knock before he entered the building, since the door was taken down. He passed the ring, which was placed in the side-niche instead of the door. It is also possible to interpret these bars in another sense. The door could have been closed, but not locked (hnigin) and in the middle of the door case (gætti) the ring was placed, i.e. in the middle of the door. The use of the preposition í makes this interpretation possible.70 There is also information in the written sources that pre-Christian ritual rings, which had been nailed to the doors of the hof sanctuaries during the Viking Age, were taken down after Christianization and instead placed on

64 L. Larsson & Lenntorp 2004. 65 L. Larsson & Lenntorp 2004, 25ff. 66 Brink 1996b, 39ff. 67 Ödman 2003, 95; L. Larsson & Lenntorp 2004, 27. 68 See e.g., Hyenstrand 1996, 153ff. 69 Dronke 1997, 168, 227. 70 I must thank Jón Viðar Sigurðsson for this suggestion. Golden Objects As Media Of Power 423

FigURES 24a and 24b Iron door rings from Uppåkra. Photo: Lunds universitets Historiska museum. 424 chapter 11

FigURE 25 The Sparlösa stone (Vg 119), from ninth-century Västergötland. ATA, Riksantikvarieämbetet. Photo: Bengt A Lundberg. the church doors.71 It is attempting to assume that the Forsa ring was also originally placed on the door of a pagan building and after Christianization was placed on the church door of Hög or Forsa. The problem is, however, that the door opening in which the Forsa ring is placed is a secondary construction (see ch. 10 above). The finds at Uppåkra and the image of the Sparlösa stone nevertheless indicate that ritual rings appeared at the doors of pre-Christian sacred buildings in Norway (as narrated by Snorri in Óláfs saga Tryggvasonar) and elsewhere. When these rings were made out of gold, they may have been symbols of royal dignity or earldom.

11.5 Golden Arm- and Neck-Rings

Perhaps Earl Hákon’s golden ring at the sanctuary of Lade should be related to some other symbols of dignity found elsewhere in Scandinavia, namely the golden arm- and neck-rings, which may appear in the context of ceremonial buildings and royal persons. Karl Hauck stated about sixty years ago that such golden rings (Lat torques, monile, diadema, corona, annulus and armilla) were signs of honour among almost all Germanic tribes.72 Written sources

71 Sveinbjörn Rafnsson 1983, 3; L. Larsson & Lenntorp 2004, 29. 72 Hauck 1954. Cf. Steinsland 1991, 153ff.; Habbe 2005, 141. Golden Objects As Media Of Power 425 report that neck-rings were used in royal contexts in Svetjud and elsewhere. Ynglingatal 12 (9), for example, mentions that King Agni was hung by his golden neck-ring by his own wife:

Þás gæðing með gullmeni Loga dís at lopti hóf . . .73

Karl Hauck is probably on the right track when he states that the Old Norse term gullmen ‘neck-ring of gold’ in Ynglingatal refers to a royal attribute. Another royal attribute related to the Ynglinga dynasty was King Aðils’ golden ring (gullhringr) called Svíagríss. Snorra Edda (Skáldskaparmál 44) tells how Hrólfr kraki set fire to a hall when he visited his sister Yrsa at Uppsala. Yrsa, who was married to Aðils, came with a horn full of gold and Aðils’ ring Svíagríss and gave it all to her brother Hrólfr. The Danes went to their horses and rode down to Fýrisvellir. Aðils and his men followed Hrólfr and the retainers on their horses. To make them stop, Hrólfr spread out the gold he had stolen on the road so that Aðils’ men had to pick it up. Hrólfr threw the ring on the ground; when King Aðils bent to pick it up, Hrólfr said:

Svínbeygt hefi ek nú þann er ríkastr er með Svíum74

Snorri says: “This is the reason why gold is called the seed of Kraki or of Fýrisvellir”,75 and he refers to a stanza from Eyvindr skáldaspillir (lausavísa st. 8):

Bá́rum, Ullr, of alla, ímunlauks, á hauka fjǫllum Fýrisvalla fræ Há́konar ævi.76

73 “. . . when ‘the wedding dis’ [Skjalf]/ raised aloft/ the ruler [King Agni]/ with the golden neck-ring”. Heimskringla I, Ísl. Fornr. 26, 38. My trans. Cf. SkP I, 22–25. 74 “Now I have made him the most powerful of the Svear to make a Swine-bend.” Skáldskaparmál 44. 75 Af þessi sǫk er gull kallat sáð Kraka eða Fýrisvalla. Skáldskaparmál 44. Cf. Skáldskaparmál 32 on kennings for gold, and Ynglinga saga 29, in Heimskringla I, Ísl. Fornr. 26, 57. 76 “Battle-leek Ull [sword warrior], we used to wear on hawk-fells [arms] the seed of Fýrisvellir [= gold] throughout Hákon’s life.” Text (Ed.) Faulkes, Edda (Skáldskaparmál, 1), 59. Translated by Faulkes, slightly modified. See also SkP I, 226f. 426 chapter 11

Skalds in tenth-century Norway were thus acquainted with “the Aðils tradition” and the idea that the king carried a specific golden ring.77 Were it otherwise, it would have been impossible to understand the kenning for gold, Fýrisvalla fræ. Norwegian kings from the Ynglinga dynasty were also associated with similar royal rings. When King Haraldr is designated with a kenning hringberendr ‘the ring carrier’ in the skaldic poem Haraldskvæði st. 1, we may conclude that it was a common practice that kings either wore rings on their arms or around their necks, or else held them in their hands during ceremonies. A survival of the noble ring custom may also be seen in Egils saga 55. When Egill had helped King Aðalsteinn to conquer King Óláfr rauði, Earl Aðils and the Scots, he visited Aðalsteinn’s hall with his band.

. . . and when the king saw Egill arrive, he gave an order to clear the lower bench for his men, and told Egill to sit in the high-seat there, facing him. . . . And after they had been sitting there like that for a while, the king unsheathed his sword, took a fine, large ring from his arm and slipped it over the point of his sword, then stood up and walked across the floor and handed it over the fire to Egill. Egill stood up, drew his sword and walked out on the floor. He put his sword through the ring and pulled it towards him, then went back to his place. The king sat down in his high-seat.78

It seems as if this text reflects an ancient ritual action in the hall of King Aðalsteinn. In addition this prose text may be supported by a stanza which may have been composed by Egill himself.

Hrammtangar lætr hanga hrynvirgil mér brynju Hǫðr á hauki troðnum heiðis vingameiði; rítmæðis knák reiða,

77 Hedeager (2011, 224ff.) argues that the story about Aðils is old and even influenced by Hunnic traditions. 78 . . . ok er konungr sá, at Egill var inn kominn, þá mælti hann, at rýma skyldi pallinn þann inn óæðra fyrir þeim, ok mælti, at Egill skyldi sitja þar í ǫndvegi gegnt konungi. . . . ok er þeir sátu svá um hrið, þá dró konungr sverðit ór slíðrum ok tók gullhring af hendi sér, mikinn ok góðan, ok dró á blóðrefilinn, stóð upp ok gekk á gólfit ok rétti yfir eldinn til Egils. Egill stóð upp ok brá sverðinu ok gekk á gólfit; hann stakk sverðinu í bug hringinum ok dró at sér, gekk aptr til rúms síns, konungr settisk í hásæti. Egils saga, Ísl. Fornr. 2, 143f. Trans. B. Scudder. Golden Objects As Media Of Power 427

ræðr gunnvala bræðir, gelgju seil á galga geirveðrs, lofi at meira.79

Both the golden arm-ring and the high-seats mentioned in Egils saga should be seen as signs of honour, and only intended for the powerful kings and the most outstanding chieftains. The story of Earl Hákon’s large golden ring at the sanctuary of Lade points in the same direction. Late Iron Age golden arm- and neck-rings have been found in Sweden, Norway and Denmark, but not in Iceland.80 Occasionally golden rings have been discovered in the context of cultic sites with ceremonial buildings. At Tissø, in Denmark, for instance, a golden neck-ring, weighing 1.8 kilos and dated to c. AD 900, was discovered in the workshop area.81 It is interesting to note that this neck-ring was found at a place which included both a hall and a cultic house. In my opinion, the reason these rings sometimes had this important role in cultic contexts is that they symbolized and represented aspects of the divinity and the mythical world. Karl Hauck’s investigation indicates that not only kings used rings as honourable objects. They were also carried by the gods. Hauck refers, for instance, to a Migration Period wooden image found in Rude Eskildstrup, Zealand, which has been interpreted as a representation of a divinity.82 It has a ring around its neck. Also interesting in this context is the golden figure from Slipshavnskogen, Funen, which is almost 7 cm high. It has been described as either an image of god or an aristocrat wearing a double ring around his neck.83 Since it is naked I think that it represents a divinity. The small figurine from Kymbo, Västergötland, also has a double ring made of gold (see fig. 26).84 It has been suggested that it reflects a hanged person. I agree

79 “The god of the armour (the king) hangs/ a jangling snare (a ring) upon my clutch/ the gibbet of hunting-birds (arm)/ the stamping-ground of hawks/ I raise the ring, the clasp that is worn/ on the shield-splitting arm,/ on to my rod of the battle-storm (sword)/ in praise of the feeder of ravens (warrior, i.e. the king).” Egils saga, Ísl. Fornr. 2, 144. Trans. B. Scudder. 80 See e.g., Hårdh 1996, 132–134; Utne Nilsen 2002. See also Kristján Eldjárn 1956 and the new edition Kristján Eldjárn & Adolf Friðriksson 2000. 81 Jørgenssen 2002. 82 Hauck 1954; Van der Sanden & Capelle 2000, 53. 83 G. Andersson et al. 2004, 79ff. 84 T. Zachrisson 2003; G. Andersson et al. 2004, 82ff. 428 chapter 11

Figure 26 The small figurine from Kymbo, Västergötland. with Torun Zachrisson’s suggestion, that it may represent Óðinn, hangatýr, ‘the hanging god’.85 In the written sources too, we see that the images of gods could be decorated with rings, mostly arm-rings. Saxo, for instance, mentions a statue representing Óðinn, where the god had heavy golden rings on his arms.86 The Old Norse mythic traditions mention the golden arm-ring Draupnir, which belonged to Óðinn.87 From it another eight equally heavy rings drip every ninth night. In the traditions about Earl Hákon’s and Guðbrandr’s hof in Guðbrandsdalr the cultic images were decorated with rings. Njáls saga 88 reports that Þorgerðr had a huge golden bracelet in her hand or ‘on her arm’.88 The historical value of this episode must be questioned. But other sources state that Þorgerðr had a ring as an attribute on statues representing her in Trøndelag.89 In this context, we must also mention Freyja’s necklace, which according to late Old Norse traditions was called Brísingamen.90 The rings appearing in the context of sanctuaries could very well be seen as symbols of dignity. Since the images of gods also carried similar objects, it is possible that these rings in the earthly sanctuary context also referred to a divine aspect, and therefore also gave the leaders using them as attributes in the ceremonies a divine dignity and religious sanction for their social position.

85 T. Zachrisson 2003. See also the small figurine with a double neck-ring from Viksø, Zealand. It has been interpreted as a representation of a goddess. This statue is much older, however, perhaps from 500 BC. 86 Cuius etiam brachiorum lineamenta confertissimo armillarum pondere perstringebant. Gesta Danorum 1.7.1. 87 See Gylfaginning 48; Skáldskaparmál 33. 88 Hann sá Þorgerði hǫldabrúði sitja, ok var hon svá mikil sem maðr roskinn; hon hafði mikinn gullhring á hendi ok fald á hǫfði. Hann sviptir faldinum hennar, en tekr af henni gullhringinn. Þá sér hann kerru Þórs ok tekr af honum annan hring. Hann tók inn þriðja af Irpu . . . Brennu-Njáls saga 88, Ísl. Fornr. 12, 214. 89 See e.g., Færeyinga saga 23, Ísl. Fornr. 25, 50f. and below. 90 Þrymskviða 13; Gylfaginning 34; Skáldskaparmál 20. Golden Objects As Media Of Power 429

For some people these impressive golden rings, which were carried by the rulers or placed on cultic images, were perceived as agencies of power in a more direct way, without making symbolic interpretations.

Golden objects, such as gold foil figures, golden bracteates, medallions and golden rings of different types, most likely played an important role for rulers when constructing authority in the ceremonial buildings of the Mälaren region and Trøndelag in the Late Iron Age. Such objects are not found in Iceland. These regional differences may be related to chronology, since gold bracteates and gold foil figures were mainly used before the time Iceland was colonized. In my opinion, some of these objects (mainly the gold foil figures) were also in some sense related to the somewhat bombastic religious strategy applied by rulers in Svetjud and Trøndelag, namely the idea that the rulers were regarded as descendants of the gods. This strategy was probably not applied by the chieftains in Iceland. We will now turn to another category of finds visible in the Mälaren region and Trøndelag (but not in Iceland), which also may be related to this religious strategy for authorization, namely the great burial mounds. chapter 12 Ceremonial Buildings, Great Burial Mounds, and the Worship of Dead Rulers

There are plenty of great burial mounds from the Late Iron Age situated beside the ceremonial buildings in the Mälaren area and Trøndelag.1 Recently scholars have emphasized their ideological dimensions.2 Besides being the last dwelling place for wealthy people, they were also important means for the elite to distinguish themselves from other groups in these regions.3 The monumental impression of these graves was thus used as a visual marker of the presence of power in the landscape. Most likely they also had other functions. Some of them may have served as meeting places for the judicial assembly and quite often they occurred at central places which were later considered as royal demesnes, i.e. places which played an organizational role for society. In his classical study Uppsala högar och Ottarshögen (1936) Sune Lindqvist argued that these graves often functioned as thing mounds. Different assemblies and ceremonies took place on these mounds, including the inauguration of the ruler.4

1 In archaeology a mound which is at least 20 metres in diameter at the base is usually called a “great burial mound” (Sw storhög). A mound which is at least 30 metres in diameter at the base is usually called a “royal burial mound” (Sw kungshög). See Hyenstrand 1974. When using these terms scholars usually refer to mounds belonging to the Iron Age, while monumental Bronze Age mounds are excluded. Silver 1996. For a critical discussion of these concepts, see Bratt 2008, 43ff. Surveys of the existence of great burial mounds in the Mälaren area and Trøndelag may be found in Bratt 2008 and Stenvik 1996. 2 Bjørn Ringstad (1987; 1991), for instance, has discussed the ideological implication of great burial mounds from western Norway in the Migration Period, while Peter Bratt (2008) focused on large mounds and power structures in the Mälaren basin during the Iron Age. On similar interpretations of great burial mounds, see also e.g. Ramqvist 1987; 1995; Müller-Wille 1992; T. Zachrisson 1994; Gansum 1995; 2002; 2004; 2013; Silver 1996. 3 Snorri wrote in Ynglinga saga (regarding Svetjud): “For notable men burial mounds were to be thrown up as memorials.” Heimskringla I, Ísl. Fornr. 26, 20. Cf. Heimskringla I, Ísl. Fornr. 26, 4. It is quite clear that Snorri was acquainted with the tradition in Svetjud of erecting large mounds after departed magnates. 4 See Lindqvist 1936, 3–17. Cf. Ringstad 1991, 143; Brink 2000a; 2000b; Bratt 2008, 141; Gansum 2013.

It thus seems as if great royal mounds should be seen as markers or symbols of power. They could be described as a ritual strategy for gaining legitimacy. In the present chapter I will argue that the royal mounds and other monumental graves constituted important ritual and ideological elements in the context of some of the most important chief sanctuaries in the Mälaren area and Trøndelag.5 They could be regarded as the materialized or concrete expression of the specific religious strategy for rulership which is seen in sources referring to these regions, i.e. the idea that the living ruler was of noble or even divine descent. The dead ancestor in the mound had a divine status or was considered as a mythical being (or hero). The monument may thus have been considered as a concrete visible and/or symbolic expression of the genealogy of the living ruler, including both its mythical and its historical aspects. Therefore the monument itself was considered as a sacred place and arena where offerings and ritual performances were made to these ancestors. The monumental graves located at the sanctuary were thus important ritual, religious and symbolic means or strategies for constructing power in the Mälaren region and Trøndelag. This element, however, is not visible in the Icelandic context, where great burial mounds are lacking. In my opinion these conditions may thus be related to the monumental ideological strategy, which may be seen in Svetjud and Trøndelag, but not in Iceland.

12.1 The Distribution of Great Burial Mounds in the Investigation Areas

During the Iron Age great burial mounds appear frequently in Sweden and Norway. According to available records there are at least 268 great burial mounds in the Mälaren region.6 Most of them (79%) have a diameter of 20–29 metres. There are also some large royal mounds with a base over 40 metres in diameter. But they make up only 7.5% of the total (i.e. 20 mounds). The largest mounds are located in the central part of the Late Iron Age settlements, for instance, Uppsa kulle and Norsborgshögen in Södermanland, Uppsala högar and Nordians hög in Uppland, and Anundshögen, Ströbohögen and Östens hög, Västmanland. The great burial mounds are often situated at lines of communication, especially water routes. They are often located at burial fields of important settlements and places close to ancient chieftain sites designated

5 A similar argument was put forward by Magnus Olsen as early as 1926. 6 See Bratt 2008, 29ff., 119. See also Lindqvist 1936; Hyenstrand 1974; Müller-Wille 1992; Ljungkvist 2006. 432 chapter 12 husby/husaby, i.e. settlements which the medieval royal power acquired by means of confiscations, compensations, ransoms, and gifts.7 At Åshusby in Norrsunda parish, Uppland, for instance, stands one of the biggest royal mounds of Sweden, Nordians hög (60 × 9 m), and Husby in Vendel is the location of the mound called Ottarshögen (37 × 6 m). There is thus often a spatial relationship between the medieval crown lands and the rich Late Iron Age burial mounds.8 Great burial mounds also appear in southern and northern Sweden, especially along the coasts. The tradition of erecting great burial mounds was introduced in the Mälaren region during the Early Merovingian Period and it ceased during the tenth century, when, for instance, Skopintull at Adelsö was erected.9 In Norway the great and royal burial mounds are spread in the area from the southernmost parts of the country up to Trøndelag in the north, and they can be dated to the period from the Pre-Roman Iron Age until the Viking Age.10 Most significant are the great burial mounds in Vestlandet and in Vestfold.11 The Oseberg ship burial in Vestfold, dated to the ninth century, contains two females with very rich finds (see above). Usually these mounds are located at old farms and settlements. Great burial mounds also appear in Trøndelag. In Inn-Trøndelag alone there are at least 143 registered Iron Age great burial mounds (more than 20 metres in diameter), which are mostly situated at farms.12 They appear frequently around Verdal. Some of the most interesting ones are the three great burial mounds from Bertnem, at Namsen in Overhalla, Nord-Trøndelag (fig. 27C). No great burial mounds have been found in Iceland. Haakon Shetelig wrote as follows about the burial customs there: “I ytre former er graverne all- tid uanselige, dekket av røiser sammenlagt av sten og dekket med gresstorv, dysser som de kalles på Island, ofte innskrenket til lave stenlegninger som neppe hever sig oven marken; . . . og det eiendommelige for Island er at vi der savner de mer anselige monumenter, større gravhauger som vi har i alle strøk av Norge fra vikingetiden.”13 Some burial mounds occur in Iceland, such as

7 See Lindqvist 1936; Rosén 1964. and Brink 2000a; 2000b. 8 See Brink 2000a; 2000b. 9 Rydh 1936. On the datings, see Bratt 2008. 10 On the Norwegian great burial mounds, see mainly Ringstad 1987; 1991; Gansum 1995; Stenvik 1996, 84ff.; Bratt 2008, 25ff., 38f. 11 Ringstad 1987; 1991. 12 Stenvik 1996, 85. 13 Shetelig 1937, 208. Ceremonial Buildings, Great Burial Mounds 433

Skalla-Grímr’s mound at Borgarnes, but they are quite small compared to the great burial mounds found in Trøndelag and Svetjud. Skalla-Grímr’s mound is only 5 metres in diameter at the base and about 1 metre in height.14 Sometimes the Sagas of Icelanders and Landnámabók also mention that people were buried in mounds.15 Whether they should be regarded as great burial mounds is uncertain. At any rate, no archaeological finds support the existence of such monuments in Iceland.

12.2 Great Burial Mounds as Part of the Aristocratic Sanctuaries

As noted above, the great and royal burial mounds in the Mälaren region and Trøndelag could be seen as a ritual-symbolic strategy used by elite groups. Whether they appeared at cultic sites is somewhat debated. The archaeologist Peter Bratt, for instance, has recently argued, in a detailed and impressing investigation, that the burial mounds in the Mälaren region rarely coincide with cultic sites.16 The cultic sites are often situated on the periphery of settlements with mounds, rather than at the sites of the mounds themselves. Bratt supports his argumentation mostly with theophoric place-names. In my opinion, his documentation is too narrow in this case. If we include the halls of these settlements, where the ceremonial feasts took place, we may find a stronger connection between burial mounds and cultic sites, at least in the foremost aristocratic milieus. There is, for instance, a physical relationship between banqueting halls and great mounds at several important central places, such as Uppsala, Tuna, in Vendel and at Hovgården, Adelsö. Since these important central places (which probably included cultic and ceremonial functions) have names which do not contain obvious theophoric or cultic elements, Bratt’s conclusion is somewhat uncertain, at least when it comes

14 Kristján Eldjárn (1956, 206) writes: “Það virðist leyfilegt að álykta, að stórir haugar hafi verið fátíðir eða óþekktir á Íslandi. Haugar hafa að vísu verið, en mjög lágir, líklega sjaldan meira en 1 m á hæð og 5 m í þvm., oft minni.” Cf. Kristján Eldjárn & Adolf Friðriksson 2000. 15 See e.g., Laxdæla saga, Ísl. Fornr. 5, 13, 72; Gísla saga, Ísl. Fornr. 6, 58; Brennu-Njáls saga, Ísl. Fornr. 10, 193; Landnámabók, H60, Ísl. Fornr. 1, 103ff.; See also Sveinbjörn Rafnsson 1974ff.; Gansum 2004, 102f. Sveinbjörn argues that many “mound narratives” (högberät- telser) are literary constructions in Landnámabók. They were made up because of the judicial right to land or in an apologetic manner against pagan practice. 16 Bratt 2008, 137–141. 434 chapter 12 to the prime aristocratic milieus.17 If we include other important cultic sites and central places outside the Mälaren region we may find the combination of cultic buildings/halls and (great) burial mounds at several important cultic sites, such as Uppåkra, Scania; Lejre, Zealand; and Borre, Vestfold.18 When it comes to the “lower elite”, however, it is possible that the cultic places often appeared at some distance from the settlements and great burial mounds as suggested by Bratt. In what follows I will point to some patterns occurring at aristocratic settlements in Svetjud and Trøndelag, which indicate that the great burial mounds were regarded as constituting ritual elements of important public sanctuaries.

Figure 27a The three royal burial mounds from Old Uppsala. Photo: Marie-Louise Sundqvist.

17 Bratt (2008, 185) is aware of these conditions when he writes: “The case of (Gamla) Uppsala does show that cult and Thing could be located in one and the same place, but this may have been reserved for the foremost ruling dynasties.” 18 L. Larsson & Lenntorp 2004; Christensen 1991; 1997; 2001; on Borre, personal communication T. Gansum. Ceremonial Buildings, Great Burial Mounds 435

Figure 27b The burial mound at Alstadhaug, Trøndelag. Photo: Olof Sundqvist.

Figure 27c The burial mounds at Bertnem, Trøndelag. Photo: Olof Sundqvist. 436 chapter 12

12.2.1 Three Great Burial Mounds in a Row and a Ceremonial Hall: A Common Ritual Pattern? In the manuscripts of Adam of Bremen’s text there is a scholion which describes the whole sanctuary of Uppsala. It says that the shrine (delubrum) was set on a plain and has hills (montes) round about it in the form of an amphitheatre. Most likely Adam built his description directly on the information he had from his informants and eye-witnesses. However, Adam revised their information by means of an intepretatio romana when describing the sanctuary as an amphitheatre.19 This information is thus somewhat uncertain. In my opinion, it seems likely that Adam had in fact heard about “mounds” located close to the “temple”, since this description harmonizes quite well with the real situation at Old Uppsala during the Merovingian Period and the Early Viking Age. At Högåsen (RAÄ 123) in Old Uppsala there are three royal burial mounds (approx. 60–68 metres in diameter) and a large number of lesser mounds (fig. 27a), at a short distance from the place where the Merovingian Period hall was found.20 The “Eastern Mound” and the “Western Mound” have been excavated and found to contain cremation graves. For a long time these graves have been dated to the Migration Period (c. AD 500), but recently the archaeologist John Ljungkvist has suggested on good grounds that they could be connected to the Early Merovingian Period (c. AD 550–650).21 The preserved grave-goods indicate that the people buried in them were very rich.22 If we take Adam’s scholion and the information extracted from the archaeology into consideration, it seems as if the monumental graves and the ceremonial hall building in Uppsala constituted parts of one and the same sanctuary. According to Ljungkvist’s chronology, the royal mounds must have been erected almost contemporary with the Merovingian Period hall at Kungsgården or soon before. Perhaps the supposed building under the church also was contemporary with these graves or erected a short time before them. Most likely the people who built the hall(s) had some kind of relation to the mounds. Perhaps they regarded the dead persons buried there as their kinsmen. The closest parallel to Uppsala in the Mälaren area may be found at the Viking Age remains of Hovgården (Adelsö), Uppland. At both places there are

19 See e.g., Hultgård 1997. 20 Lindqvist 1936; Duczko 1996a; 1996d; Bratt 2008, 119f. On the concept of royal mounds, see Silver 1996. 21 See Ljungkvist 2005; 2006, 144; 2008a. 22 See Lindqvist 1936; Arrhenius 1995; Duczko 1996d. Ceremonial Buildings, Great Burial Mounds 437 halls,23 combined with royal and great burial mounds containing extensive finds.24 At these burial mounds there are also “thing mounds” with plateaus, i.e. places where assemblies could possibly have been held in ancient periods.25 The locations of these thing mounds in relation to the burial mounds are surprisingly similar at these sites. At both Uppsala and Hovgården they constitute a complex with three mounds in a straight row, and a thing mound somewhat separated from the burial mounds, at the end of the line.26 A similar ritual structure is partly also visible at Högom, Medelpad. There we can see four great mounds from the Migration Period associated with an Iron Age hall.27 Three of these mounds were excavated in 1949 to 1960. In one of them a wooden chamber (5 × 2 m) was found with the remains of a man, an imported sword, axe, spear, expensive glass, and beakers with gold and silver decoration. These finds indicate that the man belonged to the uppermost social stratum in this region. Below burial mound no. 3, well-preserved parts of a house foundation were discovered. It was interpreted as a hall 40 m long and 7 m wide in the middle. The archaeologist Per H. Ramqvist identified one central part of the foundation and some post rows as a well built stage- like structure. Near it, he argued, the high-seat was situated.28 Most artefacts were concentrated in this area of the house. Also under the burial mound no. 4 there were remains of a house.29 Three royal mounds in a row are also visible at Bertnem in Nord-Trøndelag (fig. 27c). They are about 40–45 metres in diameter and 5–6 metres high. These

23 Brunstedt 1996, 20, 23ff.; Damell 1993 and A. Carlsson 1997. 24 The burial mound Skopintull (RAÄ 48) is dated to the Viking Age. It displays rich finds belonging to a male. It is possible that he was a king or a royal. See Rydh 1936, 104ff., 211; Bratt 2008, 231ff. The other burial mounds at Adelsö have not been excavated. 25 Bratt 2008, 145f. 26 There is also a similar pattern at Fornsigtuna. Close to the halls in Fornsigtuna a great mound was erected during the Iron Age. This mound did not contain a grave, however. Three or more great burial mounds in a row may be found at ten places in the Mälaren region. This tradition is attested from the Migration Period or Early Merovingian Period to the Viking Age. See Bratt 2008, 122ff., 185. 27 Baudou 1989; Ramqvist 1987; 1995. 28 Ramqvist 1995, 37. 29 Evert Baudou (1989) argued that ancestor worship in general was practised at grave mounds, which were intentionally placed over houses. Finds connected to these mounds indicate that cultic activities took place there. Good examples of this phenomenon could be seen at the burial fields of Högom, Medelpad, and Trogsta, Forsa parish, Hälsingland. Some of the mounds at these sites did not even contain graves. Baudou states that burial mounds were erected over houses in south-western Norway as well, for instance, at Ullandhaug in Rogaland. On this phenomenon, see also Hållans Stenholm 2012, 195ff. 438 chapter 12 prominent mounds have not been excavated. It has been suggested, however, that they should be dated to the Migration Period. There are also traces of a chieftain house at a settlement situated close to these mounds which has been excavated and dated to this period. Most likely this settlement is contemporary with the mounds.30 Perhaps the ritual structure of the burial mounds at Bertnem was based on a common cultural model, which was spread among elite groups in Svetjud and Trøndelag in the the Late Iron Age (including Viking Age).31

12.2.2 Other Monumental Graves and Cultic Sites in the Mälaren Region Extremely rich and spectacular graves associated with hall buildings may also be seen in Vendel. The funeral field there is divided into three groups: the southern, the middle and the north cemetery groups.32 The southern group is connected to the famous boat graves, the great burial mound of Vendla and two smaller burial fields. The burial finds comprise high-status objects. Most of them were found in the boat graves. There are 15 boat graves in all. Close to the hall at Vendel (Tuna) is the great burial mound of Vendla, which actually comprises two interlinked mounds. It also contained high-status finds.33 A couple of kilometres from the hall is Ottarshögen at Husby, containing rich materials.34 Also at Valsgärde there are boat graves close to a large Late Iron Age banqueting hall. A majority of the graves at this site were, however, more common cremation graves.35 This burial field displays continuity from the Pre- Roman Iron Age to the Viking Age. There are five chamber graves, which are older than the boat graves. The rich boat graves were built from the end of the sixth century up to the eleventh century. In the boat graves there are only men, one individual from each generation. There are also extremely rich female graves at Valsgärde. Grave 57, from the Merovingian Period, and graves 85 and 94, from the Viking Age, were clearly high-status female graves.36 Grave 57 is located at the highest point of the cemetery. The location of the monument clearly indicates that this woman was particularly honoured (see ch. 9 above).

30 Farbregd 1980; Ramqvist 1987, 114f.; Müller-Wille 1992, 4; Østmo & Hedeager 2004, 37; Bratt 2008, 39. 31 Müller-Wille 1992, 2ff. 32 See Seiler 1997; 2001. On boat graves see Stolpe & Arne 1912; Arne 1932; Müller-Wille 1970. 33 Ljungkvist 2006, 65ff. 34 Lindqvist 1936, 162ff. 35 Arwidsson 1942; 1954; 1977; 1980; A.-S. Gräslund 1997; Schönbäck & Thunmark-Nylén 2002; Ljungkvist 2006, 72. 36 A.-S. Gräslund 2008. Ceremonial Buildings, Great Burial Mounds 439

Great burial mounds also appear occasionally at places designated with theophoric and cultic place-names in the Mälaren region, such as Torslunda (Tierp parish) and Ärnavi (Husby-Långhundra parish). Even if this is not a rule, it intimates that some kind of connection sometimes existed between these monumental graves, the cultic sites and perhaps also the assembly sites.37

12.2.3 Ceremonial Halls and Great Burial Mounds in Ut-Trøndelag As in the Mälaren area, it seems as if (great) burial mounds recurrently appear at the pagan sanctuaries of Trøndelag, where ceremonial buildings were also erected.38 Late Iron Age burial mounds were erected, for instance, at the very important sanctuary of Lade. In 1774 Gerhard Schøning documented several burial mounds and ancient monuments at the church, but at that time these mounds were partly excavated or destroyed by ploughing (“udgravede, eller tildeels udpløiede”) (see ch. 2). Whether any of these burial mounds could be considered as “great burial mounds” is uncertain. Lorentz Diderich Klüwer adds some information on Lade in 1823, which may indicate that such mounds really existed (“nogen næsten jævnede Kæmpehouge”) (see ch. 2). Scholars therefore usually accept that great burial mounds were erected at the farm of Lade in the pre-Christian period.39 There are other places in Ut-Trøndelag where great burial mounds were erected at possible cultic sites. Written sources indicate that cult and ceremonial drinking took place at, for instance, Melhus, in Gauldœla fylki, during the Viking Age (see ch. 5). Some archaeological finds may also reflect an outdoor ritual area right beside settlements. Even if there are no burial mounds at this farm today, Gerhard Schøning reported about Melhus in the 1770s: “Paa den flade Mark, neden eller norden for Melhuus, ligge 2de runde Kiæmpe-Haue, under den Elve-Mæl, paa hvilken Kirken staaer; men neden for bemeldte Flade, længer hen mod Vest eller Nord, ligge adskillige andre, baade runde og anseelige Mænd eller familier, og dette er overensstemmende med Historien, som oppnævner os en del av dem, som der har boet og regnes blant de mæg- tigste Høvdinger og anseeligste Mænd i Thrøndelagen.”40 A similar statement was made by Klüwer at the beginning of the nineteenth century. During the archaeological investigations in 2002 these burial mounds could be confirmed.

37 Bratt (2008, 149) states: “Min analys av de sakrala ortnamnens relation till storhögsenhet- erna visar att närmare 25 % av dessa har ett direkt eller nära samband med heliga platser.” 38 For a complete investigation of the relationship between pagan cultic sites and great burial mounds in Trøndelag, see Brendalsmo 2006, 137ff. 39 See e.g., Brendalsmo 2006, 137. 40 Gerhard Schøning II:235. Quoted from Bjørk Birgisdottir & Strøm 2002, 19. 440 chapter 12

They could also be dated to the Viking Age.41 Most likely Melhus was both a chieftain farm and a public cultic site during this period. Like Lade, Melhus became the site of a “shire church” ( fylkiskirkja) in the twelfth century, and thus continued to be an important cultic site for the people of this fylki. There are some other possible cultic sites in Ut-Trøndelag where great burial mounds from the Late Iron Age occur. Since some of them are designated with place-names including the ON term hof/hov, they may have had ceremonial buildings. The burial mounds at these sites indicate the presence of wealthy and powerful families. These families were probably not as powerful as the Earls of Lade and the rulers of Melhus, yet they may have tried at some time to gain a certain position by means of public cult and great burial mounds. At the farm of Hove in Åsen, Strindafylki, for instance, finds have been made indicating cult (see ch. 5). At the neighbouring farm of Vang there is a great burial mound.42 This mound is 35 metres in diameter at the base and 3 metres high.

12.2.4 Alstadhaug, Sakshaug and Haug: Three Pre-Christian Cultic Sites in Inn-Trøndelag? It is most striking that three “shire churches” of Inn-Trøndelag were erected at farms which were designated with the term haugr ‘(burial) mound’, namely Alstadhaug (<Ǫlvishaugr) in Skogn, Sakshaug (< Sǫrshaugr, or Saurshaugr) on Inderøy, and Haug (< Haugr) in Verdal. It is very unusual that farm names in Norway contain this element. Magnus Olsen propounded an interesting theory about these names.43 According to him, the places designated with -haugr constituted the oldest central places of the “shires” ( fylki) of Trøndelag, where the cultic and political communities regularly gathered. He argued that the names Alstadhaug and Sakshaug referred to the burial mounds of two ancient rulers in Trøndelag, called Ǫlvir and Saur. In more ancient time each fylki constituted a kingdom, according to Olsen. The “good” kings were worshipped at their mounds during the common fylki assemblies at these sites. Olsen stated that these names indicate that (Inn-)Trøndelag was a well-organized society as far back as the Early Iron Age, with political and religious centres located at ancient burial mounds. He also made comparisons to names of the burial mounds found at Østlandet, where he identified a similar pattern. There he found four Fylkishaugr: Flesaker in Veldre, Flikkeshaug in Vestre Toten and Flekshaug in Norderhov and Flesaker in Øvre Eiker. He admitted that the first element was ambiguous. It could either refer to a genitive form of fylki ‘shire’ or

41 Bjørk Birgisdottir & Strøm 2002. 42 Farbregd 1986, 44–50; Røskaft 2003, 66f.; Brendalsmo 2006, 577. 43 M. Olsen 1926, 257–288. Ceremonial Buildings, Great Burial Mounds 441 fylkir ‘king’, but he preferred the latter interpretation. He thus interpreted these names as ‘the royal burial mound’ (‘kongshaugen’ or ‘haugen hvor en konge ligger (eller menes å ligge) begravet’). Also interesting in this context is the farm- name Hauge in Klepp, Rogaland, which may refer to the royal burial mound called Krosshaug. At this farm 16 gold foils were also found, indicating that this farm too was a cultic site, perhaps with a ceremonial building.44 Alstadhaug (ON Ǫlvishaugr) is located at Skogn in Skeynafylki (fig. 27b).45 The first element in the place-name contains Ǫlvir- in the genitive case, which may derive from a proper name Ǫlvir or an appellation and designation of a cultic leader called *ǫlvir (< *alu-wīhaz ‘Priester eines alu- (alh) Heiligtums’).46 The last element haugr obviously refers to the great burial mound located at the churchyard. This mound is dated to the Early Iron Age.47 The medieval name Ǫlvishaugr is mentioned in Hákonar saga góða 18. Snorri says that Blótólfr of Ǫlvishaugr was one of the eight chieftains who had most to do with the sacrifices in the whole of Trondheim. Thirteenth-century sources report that Ǫlvishaugr was an ancient political and religious centre. This information may to some extent be corroborated by archaeology. There are several Iron Age burial mounds at this place, including stray finds, such as swords, remains of a shield boss and a bronze bowl. These finds indicate high status and the presence of chieftains. Whether cultic activities took place is difficult to say. Sometime in the twelfth century, a Romanesque church was erected at Alstadhaug, and it became a fylkiskirkja. The name Sakshaug (medieval Saurshaugr), on Inderøy, in Eynafylki, most likely refers to “Saur’s (or *Sorr’s < *Sarwa’s) burial mound”,48 but this mound does not exist today. Whether a great burial mound ever was located at this site is somewhat debated.49 There are several small Iron Age burial mounds documented on this farm, as well as burial cairns. Other finds have been made, such as swords and bowl-buckles made of bronze. These finds do not give any clues to the general status and function of the farm. Written sources indicate, however, that Sakshaug had a political and religious significance during the pre-Christian period. A peculiar story is related in Hákonar saga góða 12 about

44 Cf. Særheim 1989; 2015, 53; Kristoffersson 2006; Sundqvist 2015. 45 See e.g., M. Olsen 1926, 265ff.; Sandnes & Stemshaug 2007, 69; Røskaft 1997, 232; 2003, 119ff. 46 De Vries 1932; 1956–57, §278; 1977, 687; Kousgård Sørensen 1989; See also Sundqvist 2003a, 113; 2007, 34f. 47 Stenvik 1996, 86. 48 See Sandnes & Stemshaug 2007, 379. Cf. M. Olsen 1926, 265ff.; Røskaft 2003, 131. Other interpretations have been suggested, see Stenvik 1996, 86. 49 Stenvik 1996, 86. 442 chapter 12

King Eysteinn from Oppland, who ravaged in Trøndelag and conquered both Eynafylki and Sparbyggvafylki, He put his dog Saur as ruler there in the high- seat, which was situated on the mound called Saurshaugr (i.e. Sakshaug). This story is of course some kind of migratory legend, where the name of the dog has been associated with the old name of the mound. In any case, it indicates that people during the thirteenth century associated Saurshaugr/Sakshaug with ancient ruling power and pre-Christian ceremonies connected to the mound.50 Finally, there is also a place called Haug (< Haugr) in Verdølafylki, close to Stiklastaðir. The name was probably related to a great burial mound. This mound does not exist today and some scholars think it never did.51 There is an Iron Age burial field with some mounds on this farm.52 Some of them are quite big. One measures about 25 metres in diameter and it is 3.5 metres high. Some of the finds from Haugr indicate high social status. In one boat grave a sword and the point of a spear were discovered. Most likely this farm was rich in the Late Viking Age and the Early Middle Ages. Haugr is mentioned a couple of times in Heimskringla. According to Óláfs saga Helga 109, the farm at Haugr was owned by the Christian king, and the people living there seem to have been Christians.53 Historical sources report that the farm was regarded as crown lands all through the Middle Ages. Whether there was a pre- Christian cultic building or banqueting hall in Haugr is uncertain. There are perhaps some indications of a Late Viking Age ceremonial building at Haugr in Magnúss saga in góða 14. Snorri reports that King Magnús was at a banquet (veizla) in Haugr together with Kálfr Árnason and Einarr Þambarskelfir Einriðason.54 It seems as if the king sat in the high-seat, with Kálfr on one side and Einarr on the other side. Most likely this banquet took place in some kind of hall at Haugr. Of course this story is just based on Snorri’s imagination. At least in the Middle Ages, Haugr was a religious centre with a fylkiskirkja. It was a stone church, probably built in the twelfth century. In the fourteenth century, however, this church burned down, and the church of Stiklastaðir took over its role as a fylkiskirkja.

50 See further M. Olsen 1926, 265–270. Based on the medieval information Magnus Olsen suggested that Sakshaug was also a cultic and political centre during pre-Christian era. Most scholars today think that this assumption is not possible, since the status of the available sources is so poor. See e.g., Røskaft 2003, 132. 51 See further Stenvik 1996, 86f. 52 Røskaft 2003, 128. 53 Heimskringla II, Ísl. Fornr. 27, 179f. 54 Heimskringla III, Ísl. Fornr. 28, 24. Ceremonial Buildings, Great Burial Mounds 443

12.2.5 Burial Mounds as Sacred Sites: Some Indications in Medieval Laws It thus seems as if great burial mounds in Trøndelag and Svetjud quite often were situated at pre-Christian cultic places. This idea may also to some extent be supported by prohibitions mentioned in the medieval laws from Sweden and Norway. According to these laws it seems as if pagan worship occurred at such places. Gutalagen, for instance, states as follows:55

Nobody shall invoke groves and mounds, or heathen gods (or idols), sanctuaries or stave-enclosures.56

The verb OGut haita ‘invoke’ (cf. OSw heta, hæta, haita, ON heita, Lat vocare), with the postpositive preposition a, indicates that it was the groves and mounds themselves which were regarded as divine and thus worshipped.57 It is most likely that the noun OGut haugr refers to ‘burial mound, grave howe’.58 This law thus suggests that cultic activities really did take place at pagan burial mounds. This may also be supported by the Guta Saga, which states: “Prior to that time, and for a long time afterwards, people believed in groves and grave howes, holy places and ancient sites, and in heathen idols.”59 There were thus medieval traditions in Gotland mentioning that people in earlier times venerated mounds (haugr sg.).60 Most likely Norwegian medieval laws also indicate the existence of cult at burial mounds.61 Den ældre Gulathings-Lov ch. 29 mentions that: “heathen sacrifices are also banned for we are not permitted to worship any heathen god, or [any] mound, or [in any] heathen sanctuaries.”62 Perhaps the implicit meaning of the term haugr is ‘burial mound’ and that sacrifices to such structures were forbidden. A similar formulation is repeated in the Ecclesiastical Law of King Sverrir (Kong Sverres Christenret), which is directly derived from

55 The oldest manuscript (ms.A) is dated to c. 1350. It goes back to an original dated to c. 1220. Holmbäck & Wessén 1979 (1933–46) Vol. 4, lxiv ff. 56 Engin ma haita a huathci a hult eþa hauga. eþa haþin guþ. huatki a vi eþa stafgarþa. Gutalagen, Af blotan. (Ed.) Pipping, 7. Cf. GL I: 4, SSGL 7, p. 14. My trans. 57 Cf. Lindqvist 1962–1963, 57–60. 58 Cf. Peel 1999, 68. 59 Cf. Peel 1999, 4f. 60 Cf. Sveinbjörn Rafnsson 1974, 197. 61 See e.g., Nilsson 1992, 34ff.; Sanmark 2004, 158f. 62 Blot er oss oc kviðiat at vér scolom eigi blota heiðit guð. ne hauga. ne horga. NGL 1, 18. Trans. Larson. 444 chapter 12

Den ældre Gulathings-Lov.63 Thus Norwegian laws likewise seem to have regulations against sacrifices performed at burial mounds. They therefore support the assumption that these burial mounds were regarded as cultic places in Norway during the transition period.

12.3 Great Burial Mounds as Symbols of Genealogy and Land Rights

The sources thus seem to indicate that great burial mounds in Svetjud and Trøndelag were regarded as ritual elements at aristocratic sanctuaries.64 In what follows I will argue that the reason they had this status may be related to a specific ruler-ideological strategy, which was applied by rulers in these regions. In my opinion, some mounds were regarded (at least among certain people with cultural and religious knowledge as well as among others) as a materialized medium or symbol of noble descent. The top figure of the genealogy was a forefather or a foremother, who was buried in the mound. This figure may even have had a divine status and been regarded as a kind of hero.65 Qualities such as noble or divine descent seem to have been crucial for a legitimate ruler, and also for land rights.

12.3.1 Genealogies and Great/Royal Burial Mounds As noted in chapter 3, information about native genealogies compiled for pre-Christian Swedish and Norwegian noble families has come down to us in Viking Age skaldic poetry and runic inscriptions. It has been argued that grave mounds may also have expressed such concerns in Svetjud and Norway.66 The archaeological dating of the three mounds and Gullhögen at Old Uppsala and their placing in a row seem to reflect a dynasty of Uppsala rulers who reigned in the second half of the sixth century and in the seventh century, though it may not be a straight lineage. An ossuary was discovered in the Eastern Mound,67 containing human bones of a young person.68 Among the finds there were fragments of flat sheet-metal mouldings (pressbleck) that

63 Blott er os kuiðiat at ver skulum æigi blota hæiðnar vetter. oc æigi hæiðin guð ne hauga ne horgha. NGL 1, 430. 64 The expression hǫlga grǫf in Eddic poetry also supports the idea that the mound was considered as a sanctuary. See Baetke 1942, 129f, 137. 65 See Widengren (1969, 185f.) on hero-cult. Cf. Ringgren & Ström 1984, 299. See also below. 66 Cf. Ringstad 1991, 144f.; T. Zachrisson 1994. 67 All the finds in this mound are presented by Lindqvist 1936, 168–177, 341f. 68 Duczko 1996a, 413. Ceremonial Buildings, Great Burial Mounds 445 may have been attached to a boy’s leather helmet. Birgit Arrhenius therefore suggests that this mound was erected over a young, high-ranking person.69 According to Arrhenius and Sjøvold, a jawbone probably belonged to a boy.70 If their assumptions are correct, this constitutes strong indications of dynastic kingship at Uppsala. Osteological investigations of the bone fragments in the Eastern Mound show, however, that some human bones in this mound probably belonged to a woman aged about 20–30.71 The Uppsala mounds may in fact have belonged to one kin or dynasty. Behind the mounds there was a lot of labour and organization.72 To raise a royal mound, 30 metres in diameter at the base, took 10,000 man-days.73 Very few could afford such graves. They were thus made in order to raise this family above other chieftain families as well as ordinary people. They may also have functioned as “genealogical markers”, i.e. they enumerate the forefathers of a certain family with claims to kingship at Uppsala. Other places in the area of the Svear also had mounds, perhaps with similar functions.74 As mentioned above, Adelsö has three great mounds in a row, like Uppsala, which may have expressed genealogical concerns.75 Similar conditions may have existed at Högom in Medelpad and Bertnem in Nord-Trøndelag, where we also find great or royal burial mounds situated in a row. A genealogical strategy may perhaps be attested in the act of burying dead relatives on top of the graves of early ancestors, i.e. overlying graves (Sw överla- grade graver). This phenomenon is seen at the cemetery of Ormknös on Björkö in Lake Mälaren. There are three great mounds on the crest of the ridge. Birgit Arrhenius has investigated one of these. This mound, which is 21 metres in diameter and 2.7 metres high, seems to have been constructed in two phases. The first mound was built in the early Roman Iron Age, about 17 metres in diameter and 1 metre high, covering a single grave, which has been plundered. In the Viking Age this mound seems to have been enlarged to its current size. Arrhenius states that it has the impression of being “erected as a sort of marking

69 Arrhenius 1995, 321, 328–330. 70 Arrhenius & Sjøvold 1995. 71 Sten & Vretemark 1999. 72 Ringstad 1991, 145f.; Silver 1996. 73 Silver 1996. 74 Cf. Ringstad 1991, 144f.; Hyenstrand 1974, 116f. 75 Peter Bratt expresses it thus: “To build great mounds in a row may have been a way for the leading elite to symbolically make visible a genealogy of splendid and powerful predeces- sors, and in doing so claim the elite’s time-honoured right to power in society.” Bratt 2008, 124, 185. 446 chapter 12 with the earlier ancestors”.76 A parallel to the mound at Ormknös is perhaps the Gullhögen mound in Husby-Långhundra, Uppland. Torun Zachrisson points to several other cases where mounds overlie older graves, e.g. Valsta in Norrsunda parish (RAÄ 59), Uppland, Sweden.77 Perhaps these graves are visible signs of genealogies within the landscape.

12.3.2 The Names of the Great Burial Mounds It seems as if the great burial mounds sometimes were named after the noble or mythical ancestors or heroes. By relating himself to the noble or mythical ancestor buried in the mound, the living ruler gained legitimacy. Some important great burial mounds of Trøndelag seem to have been named after the persons who were buried there. For instance, the great burial mound at Alstadhaug (in medieval documents Ǫlvishaugr) at Skogn, Skeynafylki in Inn-Trøndelag, may have been the dwelling place of the great ancestor called Alvir/Ǫlvir. Perhaps the mound was also a symbol and visible marker demonstrating that the descendents of Ǫlvir were the only legitimate rulers of the area. The name Sakshaug (in medieval documents Saurshaugr), on Inderøy in Eynafylki, may indicate something similar. It could be interpreted as “Saur’s burial mound”.78 Also in the Mälaren region, great burial mounds may have names after departed kinsman. Anundshögen, for instance, is a royal burial mound (68 m in diameter at the base and 9 m high) situated in Badelunda parish, Västmanland, and dated to AD 500–1000.79 The name is attested in the fourteenth century as vidher Anunda høgh (1392).80 It should be noted that the name appears with an Old Swedish genitive ‑a(r), which may indicate that the name is old. Perhaps there was a belief in the Viking Age or the Middle Ages that an ancestor of the ruling family in Badelunda called Anundr was buried in the mound. It is also possible that the name of the mound is secondary and related to the Late Viking Age runic stone (eleventh century) situated beside the mound.81 The inscription mentions a person called Anundr. In the vicinity of the runic stone, fourteen standing stones (bautasteinar) have been discovered in a straight row. The phrase “all these stones” in the inscription most likely refers to them and

76 Arrhenius 1990, 74. English translation in T. Zachrisson 1994, 228. 77 T. Zachrisson 1994; 1998, 98. See also Hållans Stenholm 2012, 107ff. 78 M. Olsen 1926, 265ff. Cf. Røskaft 2003, 131.; Sandnes & Stemshaug 2007(1997), 379. Other interpretations have been suggested, see Stenvik 1996, 86. 79 On this mound, see e.g., Bratt 2008, 125, 134f., 264. 80 SOL, 23. 81 S. B. F. Jansson 1987, 125. Ceremonial Buildings, Great Burial Mounds 447 it is probable that these stones lined the processional road called Eriksgatan which passed by this monument. There are many other monuments at this site, such as five stone ships (skeppssättningar). It is most likely that different types of ceremonies also took place there, at least in the Viking Age. During the Middle Ages Anundshögen was the assembly place for Siende härad.82 It is thus possible that different types of religious and judicial activities were related to this place. The main monument there represented the great ancestor. Whether he was called Anundr or something else is unknown. There are also other great burial mounds in the Mälaren region which seem to be designated with names referring to ancient ancestors. An interesting example is the royal burial mound called Ottarshögen (sixth century) at Husby, Vendel parish, Uppland. It has been argued that the name of this royal burial mound refers to the dwelling place of the Migration Period Ynglinga king called Óttarr vendilkráka.83 According to Ynglinga saga, Óttarr was killed at Vendil in Denmark.84 There is evidence to indicate that Óttarr actually could be connected with Vendel in Uppland. In two independent records from the seventeenth century this burial mound is called Utters hög and Otters hög. The information stems from people who probably could not have been influenced by the name Óttarr vendilkráka attested in Ynglinga saga. Snorri’s text became known in Sweden through Worm’s 1633 edition of Ynglinga saga.85 In 1677 the name is mentioned in an official inquiry into antiquities, in the form Otters hög.86 The information was submitted by the vicar of Vendel church. More important is the testimony in court two years earlier (1675) at proceedings held in Vendel. At this trial, two women were accused of sorcery, and one of them stated that Utters hög in Vendel was one of their places for rendezvous with the devil.87 It thus seems as if the name was rooted in a popular tradition

82 The great mound at Aspa löt, in Ludgo parish, was also probably located beside a thing place. The runic inscription Sö 137 indicates this. S. B. F. Jansson 1987, 123ff. This structure may be seen at some other places in the Mälaren region, see Bratt 2008, 142ff. and Sanmark & Semple 2008. 83 Heimskringla I, Ísl. Fornr. 26, 53–55. 84 “Then he steered west in the Öresund (Eyrarsund), and then south to Jutland ( Jótland) and entered the Limfjord (Limafjǫrðr). Then he harried in the Vendil District (á Vendli). . . . in Heimskringla I, Ísl. Fornr. 26, 54. Also Historia Norwegie (IX,25) places Vendil in Denmark. Cui successit in regnum filius suus Ottarus, qui a suo equiuoco Ottaro, Danorum comite, et fratre eius Fasta in una prouinciarum Danie, scilicet Wendli, interemptus est. 85 Linderholm 1919; Lindqvist 1936, 37–47; Vikstrand 2004b. 86 See Rannsakningar I, 24. 87 Linderholm 1919, 36ff.; Vikstrand 2004b, 379f. 448 chapter 12 in Vendel at that time.88 In addition, there is an ancient nickname referring to the people living in Vendel parish (Sw sockenboöknamn): “Vendels kråkor”.89 This nickname may support the idea that Óttarr vendilkráka really could be related to Vendel in Uppland.90

12.3.3 Great Burial Mounds as Markers of óðal and Territorial Requirements It is therefore quite possible that the great burial mounds can be interpreted as concrete or symbolic expressions of the ruler family’s genealogy and their noble descent from a legendary ancestor, a hero or mythical being. Several scholars have argued that the burial mounds not only symbolized genealogical aspects, but also inheritance of óðal, ‘inherited land property, family estate’ and land rights.91 The medieval laws make a clear distinction between the owner of the land and the óðalsmaðr.92 For the latter it was a prerequisite to be able to count his forefathers back five or six generations or to the one who was buried in the mound during the heathen period. Norwegian laws from the Middle Ages inform us how such rights of inheritance were tested. In Bergen in 1316, for instance, Håkon V proclaimed a law about óðal, that the witnesses should always “count the ancestry back to the mound and heathen time” (langfæðga tall. till haughs ok till heiðni).93 This is also illustrated by a document from Sunnmøre, in south-western Norway, dated to the fourteenth century.

Herbjørn ligger i haugen i Drifsvig. Aatte til eigu och edel Helgabostad och Holm och Starrin. Barder Nessjakonnig ligger i haugen paa Bardarstad; hand aate all Roma och alla Strandarvig och Gudøj och Hundeidavig. Eilifr jarl ligger i Eilifshaug i Gudø. Thorsten Bláfotr ligger indeni Hundeid. Rafn

88 SOL, 246; Vikstrand 2004b, 380. 89 Stjerna 1905; Linderholm 1919; Lindqvist 1936; Vikstrand 2004b, 380f. 90 Ynglingatal 20 (15) just mentions that Óttarr died in Vendil: . . . þann hergamr/ hrægum fœti,/ víts borinn,/ á Vendli sparn. It could refer to a place in Uppland as well as in Denmark and does not give any clues. 91 See e.g., Taranger 1913; 1934–1936; Sveinbjörn Rafsson 1974, 196ff.; T. Zachrisson 1994; Bratt 2008. The burial mounds have also been described as territorial markers (territoriemarker- ing eller revirmarkering). See Ringstad 1991, 143f. 92 See e.g., Magnus Lagabøters landslov, VI “Landbrigden”, ch. 16. (Ed.) Taranger, 107. Cf. Ringstad 1991, 144. 93 Kong Haakon Magnussöns Retterbod om Odelslösning. NGL 3, 121. The term langfæðgatal means, according to Fritzner (1954), ‘Slægtregister som opregner ens langfeðgar [i.e. “Personer som følge hinanden Søn efter Fader i same Slægt-række”]’. Ceremonial Buildings, Great Burial Mounds 449

som ligger i Hval i Heimdal. Grimr ligger i haugen paa Hjortdal paa Sundmør.94

The great burial mounds of Sunnmøre thus had the function of a concrete marker of territorial requirements. Magnus Lagabøters landslov (1273) says that if someone discovers goods in the ground (“jordgravet gods”), the king should have one third of them, but the one who is closest to haugóðalsmaðr, i.e. the person who can prove that he is descended from the people buried in the mound, should have one third of the goods. One third goes to the owner of the land. Sometimes the mounds are not mentioned in these laws. In those cases the heir should be able to count his forefathers in five or six generations. Den ældre Gulathings-Lov states that the litigants disputing óðal, ancestral land property, had to prove their claims by enumerating their forefathers:

Now when the doom is set, [the claimant] shall enumerate his ancestors, the five who have owned the land, and the sixth who had it both in own- ership and in óðal.95

Then witnesses were brought into court. The one who had the best witnesses won the lawsuit. The óðal right was proved by referring to the ancestors who had possessed the estate in six successive generations.96 According to the

94 Quoted from Ringstad 1991, 144. 95 Nu er þar domr settr. þeir scolo telia til langfeðra sinna. v. er átt hava. en sa hinn sette er bæðe atte at eign oc at oðrle. Den ældre Gulathings-Lov. NGL 1, 87. Trans. Brink. Cf. Gurevich 1992, 194. The same law states: Nu scal þær iarðer telia. er oðrlom scolo fylgia. Su er ein er ave hever ava leift. NGL 1, 91. Cf. Robberstad 1950; 1967. Critically considered by Gelting 2000. 96 The enumeration of five kinsmen in Den ældre Gulathings-Lov partly resembles information found in the Eddic poem Hyndlolióð, which may reflect Norwegian conditions. According to this poem, Óttarr rehearses and enumerates his five immediate paternal and maternal ancestors, in order to win the lawsuit and inherit his kinsmen’s heritage, here called fǫðurleifð ‘patrimony’. Perhaps Óttarr’s fǫðurleifð actually was his father’s kingship. Information in st. 8 indicates such assumptions: “Let’s dismount to argue about this! We should sit down, and talk about the lineage of princes (iofra ættir), about those men who are descended from the gods ( frá goðom qvómo)”. When the concept of descent of humans from gods comes into play in Old Norse sources is usually in royal or aristocratic contexts. Genealogies including divine ancestors appear in Viking Age skaldic poems. In Ynglingatal the god Freyr was regarded as the ancestor of the Ynglingar (see ch. 3 above). Genealogical knowledge including mythic elements was clearly one aspect of pre- Christian royal ideology, and it is possible that it played a crucial role when inheriting kingship as has been suggested in the case of Óttarr. Several scholars support the idea that Hyndlolióð concerns royal ideology. See e.g., Fleck 1970, Steinsland 1991, Schjødt 2008 and 450 chapter 12

Frostathings-Lov, however, the land must have been inherited through four generations of kinsmen when óðal is claimed.97 Since the Ældre Gulathings- Lov is older than Frostathings-Lov, historians have argued that the earlier rig- orous demands on óðal were somewhat eased at the beginning of the Middle Ages.98 These examples show that the laws, long after Christianity was introduced in Norway, were still influenced by the old óðal system, in which the pre-Christian burial mounds also played an essential role. Most likely the great burial mounds played a similar role in the óðal system of the Mälaren region during the Late Iron Age. The Runic Swedish term ōðal is attested on several runic stones there. On a rock at Nora, in Danderyd parish, Uppland (U 130), an inscription tells us about Bjǫrn Finnvidsson who had the rock carved with runes in memory of his brother Olev: “This farm is their ōðal and family inheritance, the sons of Finnvid at Älgesta (Er þessi byr þæira oðal ok ættærfi, Finnviðar suna a Ælgiastaðum).”99 The genealogical and judicial concept of ættærfi “family inheritance” appears here almost as a synonym to ōðal. RSw ōðal also occurs in the runic inscription at Eneby, Runtuna parish, Södermanland (Sö 145).100 Two important runic inscriptions from eleventh-century Hälsingland and Småland indicate a connection between genealogies and land rights, namely the Malsta stone (Hs 14) and the Sandsjö stone (Sm 71). By enumerating the five forefathers of He-Gylve in the Malsta inscription, Romund is declaring his legal right to inherit the estates that his father had possessed.101 Most likely judicial claims of ōðal ‘inherited land’ or ættærfi ‘family inheritance’ connected to genealogy are also attested in the Sandsjö inscription.102 This inscription counts six generations of kinsmen. The royal burial mounds as expressions of genealogy probably played a similar role throughout Norway and Svetjud also when royal patrimony was to be inherited. It proved a ruler’s nobilitas, i.e. that he was a konungsefni, and his

Cöllen 2011. On the age of Hyndlolióð, see e.g., Steinsland 1991, 247ff. It has been debated whether Hyndlolióð can be regarded as a uniform poem. Some scholars think that the poem is composed from two independent sources. E.g. Gurevich 1992, 190–199. Steinsland has convincingly shown that this lay is uniform. Steinsland 1991, 248–251. Cf. Cöllen 2011, 192ff. 97 Engum manni verðr iörð at óðali fyrr en .iij. langfeðr hafa átt oc kemr undir hinn .iiij. samf- leytt. “Land does not become any man’s óðal until three of his ancestors have owned it and it comes to the fourth in unbroken descent.” Frostathings-Lov, NGL 1, 237. 98 Lindkvist 1979, 142. 99 S. B. F Jansson 1987, 97. Cf. T. Zachrisson 1994, 222f.; Sundqvist 2002, 173. 100 Peterson 1994, 37. Cf. T. Zachrisson 1994, 223; Sundqvist 2002, 173. 101 Brink 1994; Sundqvist 2002, 151ff. 102 Sundqvist 2002, 151ff.; T. Zachrisson 2002, 39f. Ceremonial Buildings, Great Burial Mounds 451 land rights, i.e. his right to the fǫðurleifð and óðal.103 That kingship was considered as leifð or óðal can be seen in Late Viking Age kennings and poetic expressions. Following skalds state thus: • Sigvatr (c. 1030): Haralds áttleifð = the Kingdom of Norway.104 • Sigvatr (c. 1030): jǫfra erfðir = ‘the kings’ inheritance lands’, i.e. the Kingdom of Norway.105 • Valgarðr á Velli (eleventh century): óðal = [King Haraldr’s] Kingdom of Norway.106

Also according to medieval sources, the Norwegian rulers counted the kingdom (konungs dómr) and the land as their óðal.107 Snorri states, for instance, that Óláfr Tryggvason was entitled by óðal birth to the kingdom (er óðalborinn var til konungdóms).108 Similar expressions occur in the context of Óláfr Haraldsson. When he returned to Norway and spoke to his stepfather Sigurðr, his mother Ásta and foster-father Hrani, Óláfr mentioned that foreigners “dispose of the possession which my father, and his father, and one after the other of our kinsmen owned, and to which I am entitled (ok em ek óðalborinn til). Nor are they satisfied with that, but have appropriated what has belonged to our kinsmen who are descended in direct line from Haraldr Finehair (at lang- feðgatali erum komnir frá Haraldi inum hárfagra). . . . Now I shall disclose to you what has been in my mind for a long time, which is that I mean to regain my paternal inheritance (at heimta fǫðurarf minn); . . .”109 It should be noted that important terms such as óðalborinn, langfeðgatal and fǫðurarfr appear in this passage. They indicate a connection between royal genealogies and the land/kingdom as paternal inheritance.110 That Snorri’s description of these conditions is based on sound information is indicated by some words in the Norwegian laws, which say this about the king and his land: “. . . þui at hans

103 Cf. Gurevich 1992, 198; Sundqvist 2002; Cöllen 2011, 192–247. 104 Skj. B1, 232. 105 Skj. B1, 235. 106 Skj. B1, 361. 107 T. Zachrisson 1994, 221. 108 Heimskringla II, Ísl. Fornr. 27, 47. 109 Heimskringla II, Ísl. Fornr. 27, 43f. 110 The Eddic poem Rígsþula tells us that Rígr (perhaps Heimdallr) taught Jarl, i.e. sovereignty, runes, gave him his own name—declared he had a son. “That son he bade appropriate ancestral plains—ancestral plains, long-dwelt-in countrysides (þann bað hann eignaz óðalvǫllo, óðalvǫllo, aldnar bygðir).” Rígsþula 36. Trans. Dronke. 452 chapter 12

æign oc oðall er alt landit.”111 The kings of Norway thus counted the kingdom and the land as their óðal, and their ancestors’ mounds were the visible proof of this.112 Just like any óðal farm, the royal estates also had their mounds in the yards.113 Most likely a similar system was applied in Svetjud. The royal mounds in Uppsala, for instance, may have been related to the rights of a certain noble family to acquire the early crown lands called “Uppsala öd”. It is thus possible that some great burial mounds in Norway and Svetjud can be interpreted as symbolic expressions of the ruler family’s genealogy and their noble descent. Perhaps they also symbolized the ruler’s right to acquire certain land related to his office.

12.3.4 Burial Mounds, Genealogies and Inheritance of Land in Iceland The term óðal was probably not used in Icelandic judicial contexts. It does not appear in Grágás, at least as it is preserved in the thirteenth-century texts.114 Many scholars, on the other hand, have connected the Icelandic concept of aðalból ‘head farm’ to allodial landholding.115 Unlike Norwegian óðal, an Icelandic aðalból was not the exclusive possession of a patrilineal descent group. Jesse Byock writes as follows: “Icelandic heirs seem to have been mostly drawn from immediate kinsmen related through either the mother or the father”,116 but adds that an aðalból “tended to remain under the management of a family member, or at least it could be neither sold nor rented without specific agreement by the legally responsible male heirs”.117 It is easy to understand that the idea of óðal never existed in Iceland in the same fashion as in Norway. When the landnámsmenn came to Iceland they had no allodial lands which were marked with ancient great burial mounds. Sveinbjörn Rafnsson has discussed this issue as it is expressed in Landnámabók.118 He states that the purpose of Landnámabók was to verify the actual twelfth- and thirteenth-century possessions of land. Sveinbjörn argues that originally the tales of burial mounds linked with yeomen’s rights were few in Landnáma.119 Over time, family landownership changed. Therefore new

111 NGL 2, 403. 112 Sveinbjörn Rafnsson 1974, 196f.; T. Zachrisson 1994, 221. 113 Taranger 1934–1936, 124. 114 Clunies Ross 1998, 129; Byock 2001, 270f. 115 See e.g., Sveinbjörn Rafnsson 1974, 142ff. Critically considered by Clunies Ross 1998, 125. 116 Byock 2001, 271. 117 Byock 2001, 271. 118 Sveinbjörn Rafnsson 1974, 196ff., 234. 119 For a survey of tales of burial mounds in Landnámabók, see Sveinbjörn Rafnsson 1974, 198. Ceremonial Buildings, Great Burial Mounds 453 owners needed tales of settlements (landnám) to explain their possession of land. They therefore resorted to tales of the burial mounds and interpolated them into Landnáma, often as a correction to or in polemic against an older Landnáma text. Many of these tales should therefore be regarded as fictions. Anyhow, we must accept that this custom was also known in Iceland.

The great burial mounds in Svetjud and Trøndelag may thus have been the materialized and symbolic expression of a great genealogy, a memory of a famous ancestor and/or land rights related to the ruler family. Perhaps some of the small burial mounds in Iceland were related to similar ideas. In my opinion, this custom was more restricted in Iceland, however, since allodial land had no long continuity there.

12.4 The Cult of Heroes at the Great Mounds

In both the Mälaren region and Trøndelag the royal or great burial mounds seem to have constituted an important ritual element of the central and important assembly sites and sanctuaries. One reason that monumental bar- rows and rich boat graves sometimes are situated inside the ritual area of these sanctuaries may be connected to the cult of dead royal ancestors. There are signs in the written sources that the dead rulers were sometimes honoured with a common and public cult. Most likely these royals were even deified or regarded as some kind of heroes or mythical beings and expected to ensure good harvests etc.120

120 See e.g., Birkeli 1938, 89; Lid 1942, 97ff.; F. Ström 1954, 34; de Vries 1956–57, §184. Cf. Nordberg 2013, 280ff. and 382. Geo Widengren (1969, 185f., 419) describes the hero cult thus: “Die Trennungslinie zwischen Göttern und Menchen wurde in der Antike nicht sonderlich scharf gezogen. Zwischen der Gottheit und dem Menschen stand der theíos anér, der göttliche Mensch, der sich aus der menschlichen Sphäre zur göttlichen erhob. Besonders die Heroen gehörtenzu dieser Klasse von Übergangswesen, mehr als Menschen, aber gänzlich Götter. Der Heroenkult war örtlich gebunden und an die Gräber geknüpft, wo ihre sterblichen Überreste verwahrt wurden und wo sie ihre Wirkung aus- übten. Ihr Kult war sehr volktümlich. Zu den Gräbern der Heroen kan man, um Hilfe in allerei Angelegenheiten zu suchen, vor allem aber gegen Krankheiten. So sind die Heroen Heilsgestalten geworden. Ihnen brachte man blutige Opfer, Gewänder und Früchte dar. Typisch für den Heroenkult waren jedoch Mahlzeiten, gefeiert in den Heiligtümern der Heroen, die über ihren Gräbern errichtet waren.” For recent treatments on Greek hero cult, see e.g., Whitley 1995; Ekroth 2002; Hållans Stenholm 2012. 454 chapter 12

12.4.1 King Eiríkr in Vita Anskarii Adam of Bremen states in his Gesta when describing the Svear and the cultic site of Uppsala: “They also pay reverence to gods derived from men, to whom they ascribe immortality because of their heroic deeds.”121 It is quite obvious that Adam took this information from Rimbert’s Vita Anskarii, since he adds: “as one reads in the Vita of Saint Ansgar they did in the case of King Hericus (Eiríkr).”122 Vita Anskarii thus mentions that when Ansgar arrived in Birka on his second tour the people of the Svear had fallen victim to delusions. One man at this place claimed that he had participated in a council of the gods. According to this man, the gods were sceptical about the new Christian god. Instead they stated: “if you desire to have more gods and we do not suffice, we will agree to summon your former King Eiríkr to join us so that he may take his place among your gods.”123 The consequence of this was great activity and a “temple” (templum) was erected for King Eiríkr. Rimbert states:

For they [the Svear] had resolved to have a temple in honour of the late king, and had begun to render votive offerings and sacrifices to him as to a god.124

It is quite possible that Rimbert applied a euhemeristic method or ethnographic clichés when stating that the Svear worshipped deities derived from human beings. This kind of reasoning was very common in the Judaeo-Christian polemics against pagan polytheism. Anders Hultgård compares Rimbert’s (and Adam’s) information with some sequences in the apocryphal text Epistula Jeremiae (which was included in Septuaginta and Vulgata), and Thietmar’s Chronicon, where similar polemic aspects and wording appear.125 The former text says that the pagan gods are false and the idols (simulacra) of the pagan people are made out of wood, tree or metal. They are nothing but human constructions (quia non sunt dii sed opera manum hominum).126 Thietmar uses the

121 Colunt et deos ex hominibus factos, quos pro ingentibus factis immortalitate donant, . . . Adam IV,26. 122 “. . . sicut in Vita sancti Ansgarii legitur Hericum [Eiríkr] regem fecisse. Adam IV,26. 123 Porro, si etiam plures deos habere desideratis, et nos vobis non sufficimus, Ericum quondam regem vestrum nos unanimes in collegium nostrum asciscimus, ut sit unus de numero deorum. Vita Anskarii 26. 124 Nam et templum in honore supradicti regis dudum defuncti statuerunt et ipsi tanquam deo vota et sacrificia offerre coeperunt. Vita Anskarii 26. 125 See Hultgård 1997, 21. 126 Epistula Jeremiae st. 50–51, 72. Ceremonial Buildings, Great Burial Mounds 455 same cliché or topos when he writes about the Slavic sanctuary in Riedegost. In this sanctuary there are gods made by human hands (dii manu facti).127 Rimbert’s information, on the other hand, may not be totally out of order, for there are indications in the Old Norse skaldic poetry that kings after their deaths were considered as mythical beings or deities. The panegyric poem Hákonarmál (c. 960), for instance, mentions that King Hákon the Good fell at the Battle of Fitjar. When the king died with his retinues on the battlefield one valkyria exclaimed: Vex nú gengi goða.128 According to Edith Marold, this phrase may be interpreted as “die Schar der Götter wächst”, i.e. the pantheon was increased when Hákon died.129 Marold states that an apotheosis took place, i.e. the glorification and elevation of the king to the rank of god. Perhaps the king was merely transformed to one of the einherjar, i.e. a dead warrior who dwelled in Valhǫll.130 The term gengi goða should probably be translated as ‘the retinue of the gods’, which may be synonymous with the einherjar. Both interpretations, however, harmonize well with other expressions in Hákonarmál, such as “when all those who have domination and power [the gods] bade Hákon welcome” (es Hǫ́kon bǫ́ðu heilan koma rǫ́ð ǫll ok regin).131 When Hákon arrives at Valhǫll it is stated: “You may share the ale of Æsir” (þigg þú at ǫ́sum ǫl); and stanza 21 says that Hákon “to the heathen gods fared” ( fór með heiðin goð). I partly agree with Marold when she concludes: “Hákon kommt also nicht nur nach Walhall und zu Odin, er wird vielmehr unter die Götter aufgenommen.”132 It is also possible that Hákon after death was just considered as a mythical being designated einheri. Since the notion of a deified/mythified king probably appears in Hákonarmál, we cannot, in my opinion, exclude that Rimbert’s text (and Adam’s notice) contains authentic information when saying that the dead King Eiríkr was worshipped with sacrifices. But still it seems quite odd that this cult took place at a “temple”, as most translations declare.133 In the ancient Scandinavian context we hear that dead kings were honoured with sacrifices at

127 Thietmar’s Chronicon VI: 23. 128 Skj. B1, 58. 129 Marold 1972, 25. Fulk has “the gods’ force grows now”. SkP I, 185. 130 Hultgård (2011, 321) argues that the expression should rather be interpreted thus: “that the retinue (gengi) of the gods will increase in number and quality when Hákon with his host is invited to join the divine powers.” 131 Hákonarmál st. 18. Skj. B1, p. 59. Trans. Foote, in Jónas Kristjánsson 1988, 96f., modified. 132 Marold 1972, 24. See also the literature quoted in note 18. 133 See e.g., C. H. Robinson’s translation. According to Olaf Olsen (1966, 85), this information from Rimbert is quite dubious. He refers to the verb used by Rimbert in this passage, namely statuere (statuo). When Rimbert in other contexts mentions the activity of 456 chapter 12 their burial mounds (see below). If we focus on the Latin text we may note that Rimbert used the term templum, which in classical contexts meant ‘a space marked out; hence, in partic., in augury, an open place for observation, marked out by the augur with his staff’, ‘an open, clear space, a circuit’, ‘a consecrated or sacred place, a sanctuary’. In addition to these meanings the dictionaries also give the following: ‘a place dedicated to some particular deity, a fane, temple, shrine’.134 The term templum could thus have designated a cultic place under the open sky, where a burial mound was also erected. One important aspect is probably that cultic activities took place there, as Rimbert’s statements clearly indicate. The verb statuere means ‘to establish, charter, determine’ besides the basic meaning ‘to erect, to build’.135 The question is whether Rimbert’s expression templum . . . statuerunt in a medieval context could be interpreted as “that they [the Svear] established a cultic place under open sky”, where (implicitly) a burial mound was also erected. Such an interpretation would fit the ancient Scandinavian context perfect in a purely factual sense.

12.4.2 King Óláfr Geirstaðaálfr An Old Norse tradition, usually called Óláfs þáttr Geirstaðaálfs, indicates that a pre-Christian king in Norway, could be regarded as a mythical being (álfr) and venerated with sacrifices at his burial mound after his death.136 This king was called Óláfr Guðrøðarson and he was King Hálfdan the Black’s half-brother and Haraldr Finehair’s uncle. In some versions of the þáttr he appears with two by-names, digrbeinn ‘the Big-Legged’ and Geirstaðaálfr ‘the álfr from

building or erecting churches he normally uses the verbs fabricare, fiere, facere and perfi- cere. See Vita Anskarii 11, 24, 28, 31 och 32. 134 Lewis & Short. 1879. Ahlberg et al. (1964, 830) has ‘ritualenligt bestämt område e. rum’, ‘invigd plats’, ‘helgedom, tempel, kapell’. Also in Georges’ dictionary (1951, 3049) we find a similar explanations. 135 Ahlberg et al. 1964, 830. Cf. Georges 1951, 2790. 136 The þáttr has survived in many text versions published in The Great Saga of St Olaf (“Den store saga om Olaf den helige”) (all dated to the 14th century): Den selvstændige tot (AM 75e); Flateyjarbók; Bæjarbók; AM 61; Bergsbók. There is also a more comprehensive version of the þáttr in The Legendary Saga of St Olaf (“Legendariska Olavssagan”) (13th century). Most known is the version in Flateyjarbók, where the text is titled Óláfs þáttr Geirstaðaálfs. On the preservation of the þáttr in the medieval manuscripts and the different text variations, see Røthe 1997 and Wellendorf 2003. In present study I mainly follow the version from Flateyjarbók II, 6–9 and 134–135. Occasionally I refer also to the other versions of the þáttr. Ceremonial Buildings, Great Burial Mounds 457

Geirstaðir’.137 In the version of Flateyjarbók the latter byname appears. This historical king (?), who belonged to the family of the Ynglingar, lived in a farm called Geirstaðir in Vestfold during the first half of the ninth century. In the first part of the account it is stated that King Óláfr Guðrøðarson was very well liked by the people. During his reign there were always good years’ crops. But one night the king had a dream. In this dream he saw a black ox coming, killing half of the people. Nobody could interpret the dream. The king himself therefore tried to understand it. He stated that the dream indicated that after a long period of good years’ crops, a plague would befall the people with many deaths, including the death of himself and his retainers. After interpreting the dream, he told his men to erect a burial mound. He also said that they must bring expensive gifts to this mound, depose them there and when he was dead they must also bury him there. But they should not sacrifice to him when he was dead or to other dead people. The mound was erected when King Óláfr was still alive, the plague came, and many people died. Finally the king and his retainers also died and they were put into the mound. The people, however, acted contrary to the king’s advice and sacrificed to him and called him Geirstaðaálfr:

. . . var þa þat rad tekit at þeir blotudu Olaf konung til árs ser ok kolludu hann Geirstada álf.

. . . then they took that advice and sacrificed to King Óláfr for a good years’ crops and they called him Geirstaðaálfr.138

The second part of the þáttr takes place later during Óláfr Tryggvason’s and Óláfr Haraldsson’s reigns (i.e. c. 995–1028) and is mostly related to the strange events which took place at the burial mound of the Óláfr Geirstaðaálfr. In this part the farmer Hrani is playing a central role. After a notice in a dream, he opened the burial mound and took a belt with a knife, a sword and other objects (e.g. a golden ring) from the Geirstaðaálfr.139 These precious objects were later given to queen Ásta and her son St Óláfr. There is also another passage about the close relationship between Óláfr Geirstaðaálfr and St Óláfr.140

137 In, for instance, The Legendary Saga of St Olaf 2 (ed. Johnsen, 1) it is stated thus: oc var kallaðr digrbæinn eða geirstaða alfr. 138 Flateyjarbók II, 6f. My trans. 139 . . . briota hauginn . . . iij. gripe þa er þu kyss . . . taka gullhring af þeim manni . . . knif ok bellti . . . taka suerd . . . Flateyjarbók II, 8. 140 This part of the tradition is placed in Óláfs saga hins Helga, Flateyjarbók II, 134–135. 458 chapter 12

When St Óláfr and his retainers went past Geirstaðaálfr’s mound once, a man asked him: “ ‘tell me, Lord, were you buried here?’ The King answered: ‘never did my soul have two bodies, and it never will have, neither now nor on the day of resurrection, and if I say anything else, then the common faith is not truly implanted in me.’ ”141 The text continues: “Then the courtier said: ‘people have said that when you came to this place before you exclaimed; “here we were, and here we go.” ’ The king answered. ‘I never said that and I never will.’ The King was deeply disturbed at heart; he pricked his horse and sped from the place as fast as he could. It was easy to see that King Óláfr wished to uproot and blot out this heretical superstition.”142 This þáttr is with no doubt permeated with literary medieval ideas and Christian topoi.143 The motif of Óláfr Guðrøðarson as “a noble heathen”, for instance, is an example of such topos.144 Before dying (in the first part of the þáttr) King Óláfr calls his people to avoid sacrifices to his burial mound when he is dead (i.e. pagan customs). This is a bit strange, since Christianity and the new cult not yet had reached Norway at that time. He thus acts as a true Christian, though he had never heard the gospels. This is a well-known theme in medieval texts treating great pagan ancestors and beloved rulers from the past.145 The þáttr also reflects the common principle of interpretation called interpretatio christiana typologica or the “principle of prefiguration”. Events and characters in Jewish history and in the history of other “pagan religions” are seen as precursors of the true God’s revelation. The tradition about the pagan king has thus been mixed up with the medieval veneration of St Óláfr.146 This Christian elaboration about St Óláfr does not preclude that some elements of the story of King Óláfr Guðrøðarson may be based on authentic pre-Christian beliefs and customs.147 In this story, the dead king seems to be regarded as an álfr, i.e. a mythical being, who was worshiped with sacrifices in

141 Flateyjarbók II, 135. Trans. E. O. G. Turville-Petre 1964, 194. 142 Flateyjarbók II, 135. Trans. E. O. G. Turville-Petre 1964, 194. 143 See Baetke 1964, 40–47; Røthe 1997, 28 and Wellendorf 2003, 166. 144 Wellendorf 2003, 164. 145 On the topos of the “noble heathen”, see Lönnroth 1969 and Males 2010 for similar notions. 146 Wellendorf (2003, 166) concludes his study thus: “Bag totten ser vi altså en forfatter som ikke går af vejen for at benytte hedenske trosforestillinger, autentiske eller fabrikerede, i ideologisk øjemed for at bestyrke Óláfr Haraldssons position som Norges evige konge.” 147 I agree with Gunnhild Røthe (2007, 48) when she states that “this literary detail [that the king at Geirstað is identified as St Óláfr’s heathen predecessor] does not preclude the possibility that the story might be a source showing the apotheosis of a dead person.” See also E. O. G. Turville-Petre 1964, 194f. Ceremonial Buildings, Great Burial Mounds 459 order to secure fertility and the crops for the people.148 In the Eddic lays the álfar are described as a category beside the Æsir and the Vanir, indicating that they since ancient times had a prominent position in the old religion.149 We often meet the alliterated expression ásar ok álfar in oblique forms in these poems.150 Sacrifices to the álfar (álfablót) are known from several sources, such as Sigvatr Þórðarson’s Austfararvísur (see ch. 9 above). Even if these stanzas have been regarded as somewhat suspicious, the reference to the álfablót is usually perceived as reliable.151 However, this passage does not say much about the character and function of this sacrificial feast and the mythical beings called álfar. Another reference to the cult of álfar appears in Kormáks saga, where the injured hero Þorvarðr was recommended to pour the blood of a bull on to one of the hills (or burial mounds) (hóll einn) inhabited by álfar and to prepare a meal for them from the meat of the bull in order to heal his injury.152 Some scholars argue that this text reflects a memory of sacrifices performed at burial mounds153 to dead human beings called álfar,154 while others think that it has been somewhat distorted by secondary elements.155 Evidence in Austfararvísur indicates anyhow that sacrifices to the álfar (i.e. the álfablót) may belong to the pre-Christian religion. This evidence harmonizes also with the information about sacrifices to the dead king (the álfr) in Óláfs þáttr Geirstaðaálfs. There are some few examples of other rulers who were designated álfr after death, for instance Hálfdan hvítbeinn in Ynglingatal (see below).156 It is, however, much uncertain whether the worship of álfar, in a general sense, was developed from a belief of the spirits of dead people

148 Birkeli 1938, 87–89. Cf. de Vries 1956–57, §184; E. O. G. Turville-Petre 1964, 231; Røthe 1997, 34; Steinsland 2005, 248, 342ff. 149 Cf. F. Ström 1985, 199; Lindow 2001, 109f.; Gunnell 2007 and Nordberg 2013, 248. 150 See e.g., Þrymsqviða 7 and Locasenna 13. 151 On álfablót, see e.g., F. Ström 1985, 89; Holtsmark 1992, 58; Lindow 2001, 54; Simek 2006, 8. For critical assessments, see e.g., Lönnroth 1996 and Vikstrand 2001, 265f. 152 Hon [Steingerðr] segir: “Hóll einn er heðan skammt í brott, er álfar búa í; graðung þann, er Kormákr drap, skaltu fá ok rjóða blóð graðungsins á hólinn útan, en gera álfum veizlu af slátrinu, ok mun þér batna.” Kormáks saga 22, Ísl. Fornr. 8, 288. 153 Whether the concept hóll could be interpreted as ‘burial mound’ is uncertain. See Nordberg 2013, 248. 154 De Vries (1956–57, §184) states: “Die in Hügel hausenden Wesen, die in Lebensgefahr Hilfe bringen können, sind sicherlich als Totengeister zu verstanden.” 155 Lindow 2001, 54 and Simek 2006, 8. 156 See de Vries (1956–57, §184) for some examples. 460 chapter 12 and an ancient ancestor worship as suggested by Emil Birkeli and others.157 It has carefully been suggested that Óláfr was attributed Geirstaðaálfr since he both in life and death was closely related to the fertility beings called álfar and therefore was an object of cult.158 The by-name had thus nothing to do with an apotheosis of the dead king. This interpretation is of course possible. It seems, however, as if the people gave Óláfr his by-name Geirstaðaálfr right after his death when the people started to perform sacrifices to him. This is clearly expressed in one version of the þáttr (AM 75e) found in The Great Saga of St Olaf (which is older than the version in Flateyjarbók) where it explicitly is stated that the people changed his by-name from digrbeinn to Geirstaðaálfr after his death: þeir breittu nu nafne könngs og kaulludu hann Olaf Geirstada älf.159 This intimates that he was deified at the moment of death and then became an álfr.160 The þáttr also states that it was first after death the king received sacrifices, which indicates that he was elevated to the rank of gods or mythical beings at that moment.

157 I certainly agree with Nordberg (2013, 249) when he states: ”Tillnamnet Geirstaðaálfr kan dock knappast ses som bevis för att alverna som kategori generellt skulle ha utvecklats ur de dödas andar.” There are two lines of interpretation regarding the background of these mythical beings: Some scholars relate the worship of álfar to a more general pre- Christian ancestor worship. E.g. Birkeli 1938 and Steinsland 2005, 247, 344ff. Others reject the existence of such cult and interpret the álfar as developed from ancient fertility beings. F. Ström 1954, 34 and 1985, 199. For a qualified overview and discussion on the existence of ancestor worship in ancient Scandinavia, see Nordberg 2013, 239–302. Nordberg suggests that the term ancestor veneration would suit the research on ancient Scandinavian religion better than the concept ancestor worship. 158 Nordberg 2013, 249, 280f. 159 Great Saga of St Olaf (eds.) Johnsen & Helgason, 730. Røthe (1997), Wellendorf (2003) and Hultgård (2014) makes the same interpretation. 160 Anders Hultgård (2014, 34) has also produced a somewhat different suggestion in addition to the common interpretation that Óláfr after death was elevated to the rank of gods and became object of a sacrificial cult: “Det kan dock enligt min mening även vara så att en lokal fruktbarhetsgud sekundärt kom att knytas till en hövding eller kung med namnet Olav. Hur det än förhåller sig med detta, utgör berättelsen om Olav Geirstadalv ett exempel på en i förkristen tid dyrkad lokalgudom.” Cf. Hultgård (2001c, 578): “Der Beiname Geirstaðaálfr eines vorchrist. Herrschers in Vestfold deutet auf die Existenz des Kultes einer Lokalgottheit (álfr) hin, bennant nach Geirstad, dem Haupsitz der Herschenden Familie”. Nordberg (2013, 281) presents a similar interpretation: “Snarare kom dessa kun- gar och stormän att övergå till en form av lokala eller regionala platsrådare (genius loci), och lika nära som de stod till en allmän kategori anfäder, stod de förmodligen till andra typer av lägre gudomligheter med snarlika funktioner, men utan mänsklig bakgrund.” Ceremonial Buildings, Great Burial Mounds 461

The byname Geirstaðaálfr appears in texts which are older than Flateyjarbók, namely Ynglinga saga (c. 1230) and the þáttr called Af Upplendinga konungum preserved in Hauksbók (compiled by Haukr Erlandsson before 1334).161 But neither Ynglinga saga nor Af Upplendinga konungum say anything about sacrifices at King Óláfr’s burial mound. The byname Geirstaðaálfr also appears in The Legendary Saga of St Olaf (13th century),162 which is based on The Oldest Saga of St Olaf (12th century) (which only is preserved in fragments). Ynglingatal (c. 890) knows about this king, but not his byname.163 It says in stanza 26 only that King Óláfr had his site at Geirstaðir (á Geirstaðum). In this stanza some interesting information about this king appears, however. It says that King Óláfr’s family belonged to niðkvísl . . . þróttar Þrós (see ch. 3 above). The name Þrór has been related to both Óðinn and Freyr, and the expression could thus be interpreted as ‘the lineage of the strong Óðinn (or Freyr)’, i.e. a family who descended from the gods. Thus Ynglingatal intimates that King Óláfr was according to ancient tradition related to a mythical sphere. King Óláfr’s great-grandfather, King Hálfdan hvítbeinn, was furthermore called brynjálfr ‘mail-álfr’ in Ynglingatal after being buried at Skíringssalr in Vestfold,164 but this expression may be a general kenning and not a specific designation of the deified king.165 It was previously argued that the historical King Óláfr Guðrøðarson was buried at the great burial mound of Gokstad in Vestfold dating back to the Viking Age.166 Such interpretations are nowadays usually not accepted, since it has been discovered that the grave probably is from the tenth century, i.e. too late to fit the time of Óláfr’s reign.167 Anyhow, we cannot dismiss the possibility that the traditions about Óláfr Geirstaðaálfr reflects a memory of a cult of deified rulers, comparable with the Greek cult of heroes.168 These dead rulers/ mythical beings were honoured with sacrifices at their great burial mounds in order to give prosperity (til árs) (on the formula til árs ok friðar, see ch. 9 above). It should be admitted that such cult has some affinities with the medieval

161 Heimskringla I, Ísl. Fornr. 26, 4, 79, 82–84; Hauksbók, (ed.) Eiríkur Jónsson & Finnur Jónsson, 456f. 162 As noted above the name also appears in The Legendary Saga of St Olaf 2, (ed.) Johnsen, 1. 163 On the dating of Ynglingatal, see e.g., Sundqvist 2002, 43–47 and McKinnell 2010 and SkP I, 5f. 164 See de Vries 1956–57, §184. 165 Birkeli 1938, 87; Simek 2006, 8. It is possible that we should relate Óláfr Geirstaðaálfr’s name also to his mother Álfhildr and his grandfather Álfarinn from Álfheimar. See Heimskringla I, Ísl. Fornr. 26, 79ff. Cf. Gunnell 2007. 166 See e.g., Brøgger 1916; Birkeli 1938, 88. 167 Myhre 1992a, 43; 1992b, 276. 168 See e.g., Birkeli 1938, 89; de Vries 1956–57, §184; Nordberg 2013, 280ff. and 382. 462 chapter 12 cults of royal saints, which often were local and related to the burial places and reliquaries of the holy king.169 The Greek cult of heroes indicates that such ritual practices were not restricted to Christian contexts, however. Walter Burkert argued that the pre-Christian cult of heroes was transformed to the Christian cult of saints in the Mediterranean area.170 It is not impossible that something similar also took place in Scandinavia. When the Christian cult of St Óláfr appeared it could easily be recognized by people, who were used with a traditional hero cult, such as the one performed to Óláfr Geirstaðaálfr.

12.4.3 King Hálfdan Svarti and His Burial Mounds There are also other traditions mentioning that the bodies of departed Viking kings in Norway were venerated at royal mounds. In Hálfdanar saga svarta ch. 9 Snorri describes a competition between people from different parts of southern Norway (Hringaríki, Raumaríki, Vestfold and Heiðmǫrk) concerning King Hálfdan svarti’s (‘The Black’s’) body. They decided to quarter the body, each part of which was buried in a mound in a different region:

He [Hálfdan] had been most blessed with good years’ crops among all kings (hann hafði verit allra konunga ársælstr); and people were so affected by his death that when they learned of his demise and that his body was being taken to Hringaríki in order to be interred there, men of influence from Raumaríki, Vestfold and Heiðmǫrk came and prayed, all of them, to take the body with them to be buried in their lands; for it was thought that he who got possession of it could expect good seasons (ok þótti þat vera árvænt þeim, er næði). They reached an agreement in this wise, that the body was assigned to four places: the head was laid in a mound in Steinn in Hringaríki, but each of the others carried away their share and interred them in burial mounds in their homelands, and all are called the Mounds of Hálfdan.171

The account of King Hálfdan the Black is also preserved in other texts, but in different versions.172 In Fagrskinna’s redaction A of this story, the body was

169 See Sundqvist 2002, 347–364. 170 “The hero cult has often been compared to the Christian cult of saints; and without doubt there is a direct continuity as well as a structural parallel here” Burkert 1985, 207. Cf. Widengren 1969, 419ff. 171 Heimskringla I, Ísl. Fornr. 26, 92f. 172 The version in Þáttr Halfdanar svarta (Flateyjarbók I, 566f.), resembles Snorri’s story. On the different sources where this tradition is preserved, see Dillmann 2000c, 436; 2008 and Lincoln 2014. Lincoln even offers a synoptic table. Ceremonial Buildings, Great Burial Mounds 463 dismembered into three parts.173 The lost manuscript *A is dated to the fourteenth century.174 In the older redaction B (based on manuscript *B)175 this theme is lacking; it says only that Hálfdan drowned at Rǫnd í Rykinsvík and that he was taken to Steinn in Hringaríki and buried in a mound there ( fœrðr til Steins á Hringaríki ok þar heygðr). Neither does Nóregs konunga-tal, dated to the last decades of the twelfth century (and probably the oldest version of Hálfdan’s death), mention any ritual dismemberment. It says only that Hálfdan drowned under the ice, and that he was buried in the “warlike Hringaríki” (ok es hann í haug lagiðr á herskǫ Hringaríki).176 Ágrip is also silent regarding the dismemberment. It only mentions that he was taken to Steinn in Hringaríki and buried in a mound there.177 Historia Norwegie mentions that Hálfdan drowned at a lake called Rand, but it does not say anything about his burial or burial mound.178 The manuscripts of Hálfdanar saga svarta in Heimskringla, Fagrskinna A, and Flateyjarbók (Þáttr Hálfdanar svarta) mention that the people believed that Hálfdan was ársæll both in life and in death, and for this reason all the “shires” wanted his body.179 According to Flateyjarbók, sacrifices were made to

173 Fagrskinna *A, Ísl. Fornr. 29, 365f. 174 Dillmann 2000c, 436. See also ch. 6 above on the manuscripts of Fagrskinna. 175 The manuscript *B is dated to 1250. See Fagrskinna, Ísl. Fornr. 29, lxi, and 58. 176 Skj. B1, 575. This poem is preserved in e.g. Flateyjarbók II. See Skj. A1, 579 and SkP II, 761ff. 177 Ágrip, Ísl. Fornr. 29, 3. 178 Historia Norwegie X, (ed.) Ekrem & Mortensen, 80f. 179 There has been a great debate whether compound words with the adjective suffix ‑sæll, as well as the terms gæfa, gipta and hamingja, constitute evidence of the belief in an innate and immanent power of pagan Scandinavian rulers. German scholars in particular have adopted Grønbech’s (1997 (1909–12) Vol. I, 135–162) ideas and applied the concept of “Königsheil” (or “Mana”) to legitimacy and sacral kingship. E.g. Kern 1954 (1914), 16–18, and note 34. Cf. de Vries 1956–57, §274; Gehl 1939, 64–68, 142–147. Several scholars, using a radical source-critical approach, have rejected the opinion that these words reflect the notion of a pagan “Königsheil”, e.g. Baetke 1951, 47–54; 1964, 12–68; L. Lönnroth 1963–64; 1968; Ejerfeldt 1971. In reaction to Baetke’s and Lönnroth’s viewpoints, Peter Hallberg (1966; 1973) stated that the saga authors used these words and concepts as if they were “deeply rooted in a native Scandinavian world of ideas”. In more recent studies uncertainty still prevails as to the pre-Christian origin of these concepts. Claus Krag (1991, 69) denies that they are reflections of pre-Christian ideas. Gro Steinsland (1991, 312; 2000, 54, 65–67, 87, 94 f., 129, 131) states that the relevant texts of sacral kingship focus not on the luck but on the ill-luck of the kings, an aspect not observed in earlier research. Jens Peter Schjødt (1990) maintains that the issue of the king’s luck (“kraft”) is not limited to source criticism, but has to be dealt with on a comparative basis. This statement, on the other hand, raises the question of whether a comparative method can safely be used when indigenous sources are not sufficient. Sverre Bagge (1994) felt that Snorri’s description 464 chapter 12 the mound: “Each chieftain brought his part (of the body) to his home and they built a burial mound in each fylki, which they called the Hálfdan mound. Many people who had that belief performed sacrifice, until it was forbidden by kinsmen.”180 The worship of a dead ruler’s relics at his grave resembles the cult of royal saints. Hagiographical traditions may have influenced the medieval historians who wrote the late stories about Hálfdan. In the oldest texts, this motif is lacking. On the other hand, the theme of (ritual) dismemberment appears in mythical traditions, such as the cosmogonic myth of Ymir.181 This mythic theme is probably old and need not have been affected by Christian traditions. It has, for instance, parallels in several Indo-European traditions, and some scholars think that it belongs to a common heritage.182 Ritual dismemberment of a royal (mythic) person’s body also has parallels in an Indo-European context. Plutarch, for instance, mentions how Romulus was dismembered in the temple of Vulcan; thereafter the pieces were distributed to the sena- tors, who carried them away hidden in their robes.183 This narration has been interpreted as both a cosmogonic and a sociogonic myth.184 The theme of dismemberment and venerating the dead body of a royal person need not be considered as a Christian notion. I thus agree with François-Xavier Dillmann when he cautiously suggests in his thorough analysis of the Hálfdan traditions that the motif of dismemberment might be old and could be related to the ritual repetition of the cosmogony.185

of hamingja could not be characterised as a Christian notion. Some scholars, such as Bruce Lincoln (2014, 80f.), have recently stated that “the importance of this question [the interpretation of the concept hamingja] has been overemphasized”. In my opinion, the interpretation of these concepts (gæfa, gipta and hamingja as well as the suffix ‑sæll) is very crucial when discussion the pre-Christian royal ideology in ancient Scandinavia. However, I do not think that these concepts support the idea of the living ruler as a carrier of a specific intrinsic luck (“Mana”) or that the living ruler was perceived of as a divinity in pre-Christian contexts with the ability to bring blessings to his people. As the Stentoften inscription indicates (see ch. 9 above), a living chieftain could only bring good crops as long as he performed sacrifices and acted just in his cultic office. See Baetke 1964. 180 En huerr hofdinge hafde sinn hluta heim med ser ok letu uerpa haug j hueriu þui fylke ok eru þeir kalladir Halfdanarhaugar. ok hellt vid blot ok atrunat af morgum monnum adr en þat var bannat af frændum hans. Flateyjarbók I, 566f. My trans. 181 E.g. Gylfaginning 6–8. Grímnismál 40–41. See Dillmann 2008. 182 See e.g., Lincoln 1986. 183 Plutarch’s Lives I, “Romulus” 27, 4–6. 184 Cf. Puhvel 1975; Lincoln 1986, 41–64; Näsström 1995, 193. 185 Dillmann 2008. Ceremonial Buildings, Great Burial Mounds 465

Other scholars have made similar conclusions. According to Bruce Lincoln, the dismemberment of Hálfdan’s body repeats the events of creation. Hálfdan’s corpse suffers the same fate as Ymir’s: “A new world, with new conditions of existence, took shape as a result of these two deaths and more particularly as a result of the way these bodies were treated, including a radical division between central territory associated with all that is good and peripheries associated with hostile others.”186 Thus, Hálfdan’s death and the dismemberment of his body were in literary contexts related to the founding of the State of Norway. In my opinion the late versions of the Hálfdan narrative may contain an old oral tradition, which indicates that dead pagan rulers were sometimes venerated at their burial mounds for their ability to bring goodness to the living people.

12.4.4 The Cult of Dead Magnates in Iceland In his treatment of the ancestor cult in ancient Scandinavia Emil Birkeli stated that sacrifices were made to dead magnates in Iceland as well.187 One example of this practice was seen, according to him, in the traditions about Auðr in djúpúðga. Auðr appears in Landnámabók, more precisely the version rendered in Sturlubók.188 She also appears in Laxdæla saga, but there she is called Unnr.189 According to Sturlubók, Auðr was a Christian woman who settled in western Iceland, where she took possession of land at Breiðafjǫrðr, around the farm called Hvammr. She had a cross erected at a place called Krosshólar ‘Cross Hills’ where she used to pray. After her death Auðr’s kinsmen relapsed into paganism,190 and they started to worship the hills at Krosshólar. They erected a hǫrgr sanctuary there, where they also sacrificed. They believed that they would go into the hills when they died.191 In my opinion this story does not give a clear evidence of a cult of Auðr after her death. It only tells that her kinsmen made a hǫrgr sanctuary where Auðr previously had erected a cross, and that they sacrificed at this place. Nor do we

186 Lincoln 2014, 77. 187 Birkeli 1938, 90f. Cf. Lid 1942, 100. 188 Landnámabók, S13, 95–110; H14,82–84; Ísl. Fornr. 1, 50f., 136–147 and in Íslendingabók, Ísl. Fornr. 1, 6. 189 Laxdæla saga, Ísl. Fornr. 5, 3–13. 190 This information appears in Landnámabók, S110, Ísl. Fornr. 1, 147: Eptir þat spilltisk trúa frænda hennar. 191 Hon hafði bænahald sitt á Krosshólum; þar lét hon reisa krossa, því at hon var skírð ok vel trúuð. Þar hǫfðu frændr hennar síðan átrúnað mikinn á hólana. Var <þar> þá gǫr hǫrg, er blót tóku til; trúðu þeir því, at þeir dæi í hólana, . . . Landnámabók, S97, Ísl. Fornr. 1, 139f. 466 chapter 12 know whether Auðr was buried at Krosshólar. In Sturlubók we read that Auðr was “buried at the high water mark as she had ordered, because having been baptized, she did not wish to lie in unconsecrated earth”.192 The only thing we can say with the support of this text is that the people believed that they would go into the hills when they died. Birkeli has no support for his statement that “den ‘átrúnaðr’ som tok til på haugen, var knyttet til dyrkningen av Aud personlig”.193 Nothing in Sturlubók actually supports this claim.

12.5 The Cult at the Burial Mounds of Freyr and Hǫlgi

In my opinion it is thus possible that Late Iron Age rulers of the uppermost elite in Norway and Svetjud were deified and honoured with cult after their death. This cult may sometimes have taken place at their great or royal burial mound, close to the cultic building/ceremonial hall. It seems as if the dead rulers could produce welfare for living people in several ways. Whether a similar cult appeared in Iceland is uncertain. In what follows I will give some examples from semi-mythical traditions suggesting that divine and royal ancestors had such cult at their great burial mounds in Svetjud and Trøndelag.

12.5.1 Freyr in Uppsala Some euhemeristic and semi-mythical medieval prose texts report that different cultic activities took place at the mounds of Uppsala. Ynglinga saga 10, for instance, relates that Freyr, “the king of Svear”, erected a great ceremonial building (hof ) at Uppsala and made his chief residence there. The people of the Svear sent tribute to Uppsala and Freyr also received lands and cattle. This was the origin of Uppsala auð (‘Uppsala wealth’, i.e. the crown goods of the Svear), which have been kept up ever since that. In Freyr’s days there originated the so-called Fróða friðr ‘Peace of Fróði’ and there were good crops in all lands at that time. The Svear attributed that to Freyr. Snorri also tells us:

Freyr took sick; and when the sickness gained on him, his followers hit upon the plan to let few men see him, and they threw up a great burial mound with a door and three windows. And when Freyr was dead they carried him secretly into the mound and told the Svear that he was still alive, and kept him there for three years. But all the tribute they poured

192 . . . ok var grafin í flæðarmáli, sem hon hafði fyrir sagt, því at hon vildi eigi liggja í óvígðri moldu, er hon var skírð. Landnámabók, S110, Ísl. Fornr. 1, 146f. Trans. Hermann Pálsson & Paul Edwards. 193 Birkeli 1938, 90. Ceremonial Buildings, Great Burial Mounds 467

into the mound—gold by one window, silver by another, and copper coin by the third. Thus good seasons and peace endured.194

It thus seems as if Freyr received treasures in his mound at Uppsala when he was dead in order to maintain “peace and prosperity”. This information from Ynglinga saga was repeated and elaborated in other Old Norse texts, for instance Óláfs saga Tryggvasonar in Flateyjarbók.195 This text says that Freyr was regarded as a great king in Svetjud. When he died he caused his people great sorrow. They put him into a mound and offered gold, silver and copper coins, and thought that he still was alive and thus could give them ár ok friðr. Because no living person wanted to join Freyr in the mound, they made two wooden statues of Freyr and placed one in the mound. The other they sent to Trondheim and this statue became the starting point for the local cult of Freyr there. These traditions are usually regarded as fictional and late stories about the cult of Freyr in Uppsala.196 They are impaired by euhemeristic interpretations of the pre-Christian gods, where the deities are styled as human beings.197 When the medieval authors depicted the pagan gods as more or less earthly figures, the intention was to draw a negative picture of the devotes, i.e. making them so stupid that they not even could see the difference between real gods and “historical” people.198 Therfore Snorri’s text is usually considered as unreliable in the context of history of religions. In addition, there is a conflicting tradition about Freyr’s death in Gylfaginning, mentioning that Freyr was killed by Surtr at Ragnarǫk.199 The motif that it was kept secret that Freyr was dead for three years also appears in Saxo Grammaticus’ text.200 In Saxo’s text this motif is related to the Danish king Frotho III of the Skjǫldunga kin, sometimes interpreted as

194 Freyr tók sótt, en er at honum leið sóttin, leituðu menn sér ráðs ok létu fá menn til hans koma, en bjoggu haug mikinn ok létu dyrr á ok þrjá glugga. En er Freyr var dauðr, báru þeir hann leyniliga í hauginn ok sǫgðu Svíum, at hann lifði, ok varðveittu hann þar þrjá vetr. En skatt ǫllum helltu þeir í hauginn, í einn glugg gullinu, en í annan silfrinu, í inn þriðja eirpen- ningum. Þá helzk ár ok friðr. See Ynglinga saga Prologus: “Freyr was put to rest in a burial mound at Uppsala” (Freyr hafði heygðr verit at Uppsǫlum). Heimskringla I, Ísl. Fornr. 26, 4, 24. Trans. Hollander. 195 Óláfs saga Tryggvasonar, in Flateyjarbók I, 400–405. 196 See e.g., Lindqvist 1936, 112, 248f. 197 See e.g., G. W. Weber 1994, 9–12. Critically considered by Schjødt 2009. 198 Cf. Weber 1994, 3f. and Schjødt 2009, 569. 199 Gylfaginning 51. Cf. Vǫluspá 53; Lokasenna 42. 200 Saxo 5.16.3. 468 chapter 12 identical with Freyr.201 During his reign the “Peace of Frotho” (cf. Fróða friðr) was established.202 This peace also included a long series of good years’ crops which made the farmers prosperous. The peace was maintained as long as nobody removed a golden ring lying on the hearth at the royal hall.203 When Frotho was dead, he was carried around in a vehicle for three years, in order to keep the information about his death secret. As long as they thought that he was alive they continued to pay tribute to him to maintain the peace of Frotho. Thus this narrative resembles the traditions about Freyr’s death, with the tribute and the maintained peace. Whether this motif is built on an ancient Scandinavian tradition is somewhat uncertain. There are similar stories in the classical literature. Herodotus reports on a deity of the Getae people (Thracian Dacians) called Zalmoxis.204 He was a god of death, and hid himself for three years in the underworld, while the people mourned his death. However, on the fourth year he appeared again to the Getae and taught them about the immortality of the soul. It has been argued that the Zalmoxis myth may have influenced the Scandinavian tradition about Freyr’s death.205 In my opinion, the Greek and Scandinavian traditions do not resemble each other in detail.206 In Ynglinga saga Freyr does not return to his people after three years as Zalmoxis does, for instance. Therefore these traditions need not necessarily be linked. Even if the accounts of Freyr’s death at Uppsala and his burial in a mound have the character of euhemeristic fictions, they may reflect a misty memory of sacrifices to a dead deified ancestor or hero, which were performed at his mound in Uppsala.207 As noted above, it is quite likely that Freyr was regarded as the specific sacrificial god of the Svear (blótgoð svía) and it seems as if he had a specific relation to the royal Svea kin. Some of the rulers of the Ynglinga family, according to Ynglingatal, were called Freys afspringr and Freys ǫ́ttungr, i.e. Freyr’s offspring. Snorri’s account of Freyr may have developed from a tradition mentioning that an ancestor of the royal family received cult at his mound.

201 There is support for this conclusion in Ynglinga saga 10, where Fróða friðr is mentioned in connection with Freyr. See also e.g. de Vries 1956–57, §§460f.; Å. V. Ström 1975, 143, 148; Näsström 2001b, 88f. 202 See also Vellekla 18; Grt. 1–6; Ynglinga saga 10–11; Skáldskaparmál 40. 203 See Simek 2006, 95. 204 Herodotus 4,94–96. 205 See e.g., Neckel 1920, 119; see also de Vries 1956–57, §460 and the literature referred to there. 206 Cf. Schier 1968; Eliade 1972; Polomé 1995; Schjødt 2009. 207 The tradition of the Danish king Frotho mentions that he was also honoured with ceremonies, such as the reciting of skaldic poems, at his burial mound. See Saxo 6.1.1. Ceremonial Buildings, Great Burial Mounds 469

This human ancestor may have been deified after death and related to, or confused with, the mythical forefather Freyr, who also secured good seasons and peace (ár ok friðr) for the people.

12.5.2 Hǫlgi and Þorgerðr Hǫlgabrúðr as Local Tutelary Gods of the Háleygja Kin (Lade Earls) A similar story to the one about Freyr in Uppsala may also be found in the traditions related to the mythical beings/local heroes/heroines called Hǫlgi and Þorgerðr Hǫlgabrúðr. It seems as if they were worshipped especially in Háleyjgjaland and Trøndelag. Þorgerðr was also worshipped in other parts of Norway and Iceland.208 Many sources connect Earl Hákon with Þorgerðr.209 It seems as if he worshipped this deity and dedicated sanctuaries to her in Trøndelag. Óláfs saga Tryggvasonar, as it is rendered in Flateyjarbók, states that Þorgerðr had been Earl Hákon’s best friend ( fulltrúi) and that she and Freyr were worshipped by the Þrændir. King Óláfr Tryggvason arrived in this part of the country, where he burnt the statues of Freyr and Þorgerðr.210 This late story intimates that the shrine of Þorgerðr had previously been owned and controlled by Earl Hákon. Exactly where Þorgerðr’s sanctuary was located is not mentioned, but it has been argued that this tradition reflects a historical event at Lade, from the transition period when the sanctuary there burned down.211 A cultic relationship between Earl Hákon and Þorgerðr is also mentioned in Færeyinga saga 23 (c. 1210).212 This relation also appears in the traditions about the battle between Earl Hákon and the Jómsvíkingar around 986. According to Jómsvíkinga saga (c. 1200), Earl Hákon realized that his own life was threatened. He then invoked his female protector (fulltrúi) Þorgerðr Hǫlgabúðr and offered her all kinds of victims, including human sacrifices, but she did not listen.213 In the end he offered her his own son Erlingr, who was seven years

208 Brennu-Njáls saga 88 for instance, mentions Earl Hákon’s and Guðbrandr’s hof in Guðbrandsdalr, where the deities Þórr, Þorgerðr and Irpa were worshipped. In Harðar saga ok Hólmverja 19, a sanctuary of Bláskógar in Iceland is mentioned, called hof Þorgerðar Hǫlgabruðar. 209 E.g. Flateyjarbók I, 144, 213f., 407–409. 210 . . . þa brend bæde likneskin Þorgerðar ok Freys . . . Óláfs saga Tryggvasonar, Flateyjarbók I, 409. 211 De Vries 1956–57, §562; cf. Steinsland 2005, 293. 212 See Færeyinga saga, Ísl. Fornr. 25, 50f. The text is quoted in ch. 7 above. See also a similar version in Færeyinga þáttr, Flateyjarbók I, 144f. 213 Jómsvíkingasaga ch. 34. (Ed.) Petersens, 115. Cf. Jómsvikinga Þáttr, Flateyjarbók I, 191. Jómsvíkinga drápa, tells that the earl sacrificed his son to Þorgerðr Hǫlgabúðr. Skj. B2, 7. 470 chapter 12 old, and that she accepted. The fight went on and suddenly a great and unusual hailstorm arose. It worked to the advantage of Earl Hákon, and his enemies, Earl Sigvaldr and the Jómsvíkingar were put to flight.214 In the nineteenth century it was often argued that the intimate cultic relationship between Earl Hákon and Þorgerðr Hǫlgabúðr reflected historical conditions. It was also suggested that Þorgerðr’s father or lover Hǫlgi (Helgi or Hølgi) was regarded as the eponymous ancestor of the Háleygja kin, i.e. the mythical father of the Earls of Lade. Gustav Storm, for instance, believed that Þorgerðr was the original divine “tribe mother” of this family, but this notion was replaced by a more “distinguished” ( finere) genealogical tree for the earls when Eyvindr skáldaspillir in Háleygjatal made the earls into descendants from Óðinn and Skaði.215 A more sceptical attitude towards the traditions about Hǫlgi and Þorgerðr was seen in the twentieth century. Walter Baetke, for instance, stated that Þorgerðr was a late fabulous being with no basis in the pre-Christian Scandinavian religion.216 Other scholars likewise felt that Þorgerðr was a late literary figure who did not have a place in older traditions. Since the contemporary skalds of the Earls of Lade, e.g. Kormákr, Eyvindr skáldaspillir, Einarr skálaglamm and Tindr Hallkelsson, never mentioned Þorgerðr in their poetry, she cannot have played any significant role for this family during the tenth century.217 In her doctoral thesis, Gro Steinsland opposed these sceptical opinions about Þorgerðr. According to her, there may very well be allusions to Þorgerðr in the contemporary skaldic poetry. Steinsland referred, for instance, to Tindr Hallkelsson’s poem (drápa) on Earl Hákon where the name Gerðr appears,

Cf. Heimskringla I, Ísl. Fornr. 26, 286 and Fagrskinna. Ísl. Fornr. 29, 129, 131, 138. In the latter texts nothing is said about a sacrifice. 214 Scholars have in general been sceptical of these traditions about Þorgerðr’s role at the battle of Hjǫringavágr and many have classified them as Christian inventions. See e.g., Baetke 1951, 45; Weibull 1948, 349–358; Näsström 1996. Cf. de Vries 1956–57, §562, note 2; F. Ström 1954, 52; Hultgård 2002. The relation between Earl Hákon and Þorgerðr Hǫlgabúðr is also evident in the story about the skald Þorleifr jarlaskáld rendered in Þáttr Þorleifs jarlaskálds in Flateyjarbók I, 213ff. In addition to the saga literature we may perhaps also find references to Þorgerðr in Saxo’s Gesta Danorum, i.e. the being called Ladgerd (Latin Lathgertha). See Chadwick 1950, 408 and Näsström 1995, 159. 215 Storm 1885. Cf. Steinsland 2011b, 33ff. 216 Baetke 1950, 32 and 45. 217 Olrik & Ellekilde 1925–1957, 458f. Olrik and Ellekilde argued that Þorgerðr belonged to the lower mythology (“den lavere mytologi”). According to them, the worship connected to her was described as a primitive fisher and seaman cult. Motz (1997, 473ff.) identified Þorgerðr as a giantess who controlled the elements of nature. Ceremonial Buildings, Great Burial Mounds 471 and interpreted it as an abbreviation of Þorgerðr. According to her, the skald referred to the symbolic erotic relationship which existed between the earl (who perhaps incarnated Hǫlgi) and this female mythical being.218 Steinsland also argued that Þorgerðr’s byname Hǫlgabrúðr, which is sometimes attested as Hǫlga troll, Hølda brúðr, Hørða brúðr, hørga brúðr, Hǫrda-troll, holga-troll or horda-troll, indicates that this being was considered as a giantess. The word troll emphasizes this. The close relationship between Hǫlgi and Þorgerðr thus resembles the odd marriage which took place between the god Óðinn and the giantess Skaði attested in Háleygjatal, as well as the relationship between the god Freyr and the giantess Gerðr in Skírnismál.219 According to Steinsland, the descent from Þorgerðr and Hǫlgi represents a forerunner of the Óðinn- Skaði genealogy of the Háleygja dynasty.220 In my opinion, both Storm and Steinsland may be right. There may have existed an old tradition in Trøndelag and Hálogaland reporting that both Hǫlgi and Þorgerðr were considered as some kind of mythical ancestors of the Háleygja kin.221 This idea may partly be supported by a stanza in Þorbjǫrn hornklofi’s poem Haraldskvæði, where the expression Hǫlga ættar appears.222 It seems there as if it refers to the Háleygja kin. This poem is usually dated to the time around 900 and is thus older than Háleygjatal. In Njáls saga 113 we may find some interesting genealogical information about King Háleygr, who probably is identical with Hǫlgi.223 There it says that the mother of Grímr was Hervǫr. Her mother was Þorgerðr, the daughter of King Háleygr of Hálogaland. The eponymous Háleygr (Hǫlgi) must be regarded as the forefather of Háleygjar, i.e. the family from which Earl Hákon originated. Whether Þorgerðr had the character of a giantess, as Steinsland suggests, must be regarded as uncertain. The name forms containing the element troll appear in late traditions.224 This designation may thus be imbued with

218 Steinsland 1991, 221; cf. Chadwick 1950, 400. 219 Steinsland’s theory was partly an elaboration of an idea originally suggested by Folke Ström (1983). Cf. Motz 1997, 473ff. 220 Steinsland 1991, 225. 221 Cf. Storm 1885; Steinsland 1991, 220–226. 222 Hafnaði Holmrygjum/ ok Hǫrða meyjum/ hverri enni heinversku/ ok Hǫlga ættar,/ konungr enn kynstóri/ es tók konu ena dǫnsku. “Han vraket holmrygr og hordemøyer alle fra Hedmark og av Holges ætt; kongen den ættstore som ektet den danske kvinne. Skj B1, 24. It should be noted that F has holga and J1 has halþa. Skj A1, 27. Cf. SkP I, 108f. 223 Brennu-Njáls saga 113, Ísl. Fornr. 12, 285. See Storm (1885, 128) for the relationship between Hǫlgi and Háleygr. 224 Storm 1885, 126f. 472 chapter 12

Christian attitudes to the old pagan goddess.225 In the oldest texts, the compound always has the last element ‑brúðr, meaning plainly ‘bride, wife, woman’.226 But it should be admitted that Snorri lists the name Hǫlgabrúðr among designations of giantesses.227

12.5.3 Hǫlgi and the Cult at his Burial Mound The references to Hǫlgi and Þorgerðr in the Old Norse texts thus indicate that the traditions about both these mythical beings may be old. Other sources indicate an old tradition about them. In Skáldskaparmál 45, for instance, it is stated that sacrifices were made at Hǫlgi’s royal burial mound. Silver and gold were offered to him and placed in his mound. In the version of Codex Regius we read thus:

They say that the king known as Hǫlgi, after whom Hálogaland is named, was Þorgerðr Hǫlgabrúðr’s father. Sacrifices were offered to them both, and Hǫlgi’s mound was raised with alternately a layer of gold or silver— this was the money offered in sacrifice—and a layer of earth and stone.228

Hǫlgi’s burial mound is also evidenced in a skaldic poem made by Skúli Þorsteinsson (c. 1020), referring to the battle of Svolder. This verse was also quoted by Snorri when treating kennings referring to gold:

Þá er ræfrvita Reifnis rauð ek fyrir Svǫlð til auðar, herfylgins bar ek Hǫlga haugþǫk sama baugum.229

The kenning for gold, Hǫlga haugþak ‘Hǫlgi’s mound-roof’, indicates that there were old traditions about Hǫlgi’s (rich) burial mound.

225 Näsström 1996, 107f. 226 Heggstad et al. 1993, 65. 227 Skáldskaparmál 75, v. 424. See also Motz 1997, 475. 228 Svá er sagt at konungr sá er Hǫlgi kallaðr, Hálogaland er við nefnt, var faðir Þorgerðar Hǫlgabrúðar. Þau váru bæði blótuð ok var haugr Hǫlga kastaðr, ǫnnur fló af gulli eða silfri (þat var blótféit) en ǫnnur fló af moldu ok grjóti. Skáldskaparmál 45. 229 “When I reddened Reifnir’s roof fire [shield-fire = sword]/ Off Svold to gain wealth,/ I amassed warlike/ Hǫlgi’s mound-roof [gold] in rings.” Quoted from Faulkes 1998, 60. Cf. Skj B1, 284, A1, 306. Trans. Faulkes. Ceremonial Buildings, Great Burial Mounds 473

In Skáldskaparmál 45 Snorri stated that that sacrifices were offered to both Hǫlgi and Þorgerðr. Snorri also mentioned that Hǫlgi’s mound was raised with “alternately a layer of gold or silver—this was the money offered in sacrifice”. It seems thus as if Snorri believed that Hǫlgi received sacrificial gifts directly in his mound, rather as “King Freyr” received treasures in his mound at Uppsala. According to the Codex Upsaliensis version of Skáldskaparmál, Hǫlgi’s mound was also made for Þorgerðr (ok var haugr gerr at þeim).230 Father and daughter were thus placed in the same burial mound and sacrifices were performed to them both there. Based on this information, Gunnhild Røthe argues that Þorgerðr Hǫlgabrúðr originally was a human foremother of the Háleygjar family.231 According to Røthe, the veneration of Þorgerðr developed from a cult based at her mound. This cult was later located in the ceremonial buildings as described in the medieval sources. In these buildings she was represented by cultic images and statues. Earl Hákon venerated her as his fulltrúi ‘best trusted friend’, but she was also important for all members of the Háleygjar family, where she was conceived as a fylgja i.e. a “tutelary spirit”. Originally she was an ordinary human being who was not an object of cult until after her death. Røthe proceeds from the idea that the cultic place called hǫrgr in the sources could be identical with a “burial mound”. In a manuscript of a medieval Norwegian Law called AM 78 4° it was explained that it was forbidden to build a mound or house and call this a hǫrgr (at han læðr hauga eða gerer hus oc kallar horgh).232 According to Røthe, this prohibition indicates thus that a burial mound could be regarded as a cultic place called hǫrgr.233 Røthe states that the connection between Þorgerðr and her mound is indicated in the name Hǫrgabrúðr ‘the mound bride’.234 Þorgerðr was thus worshipped at her burial mound. Þorgerðr’s descendents, i.e. the members of the Háleygja family, may also have been worshipped at mounds. For instance, the name of Earl Hákon’s great-grandfather, Grjótgarðr ‘stone enclosure’ indicates such a notion. According to Snorri Sturluson, King Herlaugr and his brother King Hrollaugr built a mound of stone (grjót) and earth. Herlaugr supplied the mound with food and drink and entered it alive with his men when King Haraldr was

230 See Codex Upsaliensis of Snorra Edda, edited by Heimir Pálsson 2012, 244. 231 Røthe 2007. 232 See NGL 1, 430. 233 Røthe 2007, 50. Cf. O. Olsen 1966, 105f. 234 Þorgerðr is called Hǫrgabrúðr, for instance, in one of the manuscripts of Njáls saga, i.e. Kálfalækjarbók. See Ísl. Fornr. 12, 214, note 2. See also Harðar saga Hólmverja ch. 19; Flateyjarbók I, 214. 474 chapter 12 approaching him.235 Like the story about Freyr in Uppsala, this tradition indicates that some kind of rituals were performed to the dead people “living” in the mound. There are some weaknesses in Røthe’s argumentation. It is highly uncertain whether a burial mound could be designated hǫrgr in Old Norse. It is possible that the term hǫrgr in manuscript AM 78 4° may only refer to ‘house’ (hus) and not to ‘mound’ (haugr). Her assumption that Þorgerðr was worshipped at a mound called hǫrgr, further, is based on Þorgerðr’s byname Hǫrgabrúðr. This name might be secondary to the name Hǫlgabrúðr, which seems to be the most common name-form in the oldest texts. Røthe’s suggestion that Þorgerðr originally was a human woman is based on more solid ground. In Njáls saga 113 we have the genealogical information that Þorgerðr was daughter of a king called Háleygr of Hálogaland (i.e. Hǫlgi).236 Snorri too believed that she was the daughter of King Hǫlgi. In one manuscript Þorgerðr was buried in a mound like her father, thus indicating that she originally was regarded as a human. This may of course be an expression of learned euhemerism on Snorri’s part. A connection between Þorgerðr and a burial mound also appears, however, in Harðar saga Hólmverja 15. In this text Þorgerðr is described as the sister of the haugbúi Sóti. If Sóti was regarded as a mound-dweller then his sister could also be regarded as such. This tradition indicates that Þorgerðr was believed to be an ordinary woman, most likely a queen or a princess, who was linked to a burial mound. All these traditions about Hǫlgi and Þorgerðr may perhaps be related to the idea mentioned in several sources that rulers in Norway sometimes were deified after their death. Such rulers were worshipped with sacrifices at their burial mounds. Even if several imaginative aspects were gradually added to the traditions about Hǫlgi and Þorgerðr, it seems as if they indicate a pattern visible in other sources. Evidence thus indicates that there really was a pre-Christian cult related to these local mythical beings (heroes/heroines) at a great burial mound.237 Whether Hǫlgi and Þorgerðr were considered as father and daughter or husband and wife is hard to know. The term bruðr has the primary meaning ‘bride, woman’ but not ‘daughter’. In my opinion, this means that their relationship should be described as the one existing between

235 Haralds saga ins Hárfagra, in Heimskringla I, Ísl. Fornr. 26, 99f. 236 Þorgerðr, dóttir Háleygs konungs af Hálogalandi. Brennu-Njáls saga 113, Ísl. Fornr. 12, 285. Cf. Storm 1885, 128. 237 Cf. Storm 1885; de Vries 1956–57, §562; Ström 1983; Steinsland 1991, 220–226; Hultgård 2001c, 578. Ceremonial Buildings, Great Burial Mounds 475 husband and wife.238 Most likely they were also regarded as the protectors (tutelary spirits) of the noble kin of Hálogaland.239 It also seems plausible that Hǫlgi at some stage in history was regarded as the eponymous hero of this kin, although a competing genealogy related to the Earls of Lade made Óðinn the divine ancestor of this family. Based on these assumptions, it is not unreasonable to presume that Earl Hákon had a certain relation to Þorgerðr. But I must admit that the descriptions of this relation in the medieval sources most likely are results of literary constructions.

It can be concluded that there was a pre-Christian veneration of dead and deified rulers in Svetjud and Trøndelag, which took place at the great burial mounds of the cultic sites of e.g. Uppsala, and Alstadhaug. This veneration included both foremothers and forefathers. Invoking the dead ruler at his mound was a way of reaching the divine powers, i.e. a means of receiving blessings from the ruler in the Other World.240 The great burial mounds were thus regarded as the materialized expressions of the close relation that was believed to exist between the ruler and the mythical beings. The living ruler was in a genealogical sense related to the deified ancestor or hero dwelling in the mound. This religious strategy including the notion of divine descent and the monumentalization of rulership is clearly indicated at sanctuaries in Svetjud and Trøndelag, but not in Iceland. In what follows I will discuss the inauguration ceremonies performed at sanctuaries in Svetjud and Trøndelag. In these performative rituals the great burial mound may have played a certain role as a ritual arena.

238 See also Røthe 2007, 46. 239 Else Mundal’s (1974) and Gunnhild Røthe’s (2007) suggestion that Þorgerðr was the foremother and fylgja of the Háleygjar family seems very plausible in my view. 240 Kaliff 1997, 75ff. and 2006. chapter 13 The Arenas for Royal Funeral Feasts and Inauguration Rituals

As noticed in the introduction to this book, some traditions in Old Norse sources refer to the funeral feast, often called erfi (n) ‘funeral feast, heir, inheritance’ (cf. ON arfr m, arfi m, erfa vb). During this type of feasts ritual drinking was often performed.1 It seems as if these feasts took place in all three investigation areas. In the Old Norse prose texts there are also some traditions referring to performative rituals executed at such feasts, which seem to have been more exclusive for kings and earls in Svetjud, Norway and Denmark. One of them is the royal funeral feast. In Ynglinga saga Snorri describes such ceremonies in general terms. These rituals also included an inauguration of the new king. Since these performative rituals took place in a hall, close to the high- seat, they are also relevant in this study, as they can clearly be conceived as a ritual strategy for gaining authority and legitimacy. Closely related to these rituals is the royal custom of sitting on a (grave) mound (sitja á haugi) mentioned in the Norse sources.2

13.1 King Ǫnundr’s Funeral Feast

In Ynglinga saga 36 Snorri tells the story of King Ingjaldr and the funeral feast he gave for his father King Ǫnundr at Uppsala:

King Ingjaldr had a great banquet prepared at Uppsala for the purpose of honouring King Ǫnundr, his father, with a funeral feast (erfi). He had

1 There are few studies on this topic, although one exception is an essay by Ottar Grønvik (1982) dealing mostly with etymology and linguistic aspects. Cf. de Vries 1956–57, §292f.; Cahen 1921, 51f., 86–91, 150–161; Düwel 1985, 75–82. See however Sundqvist 2002, 259ff. General drinking ceremonies have been treated more frequently in the research. The classical work is by Cahen 1921. Cf. de Vries 1956–57, §292f. L. Carlsson 1965, 125–131; Å. V. Ström 1975, 234–239; Düwel 1985, 48–118; Grønvik 1990; Drobin 1991; Hultgård 1996, 42–44; Rydving 1996; Dillmann 1997, 59–61; Sundqvist 2002, 191ff. 2 This is a revised version of the chapter “The ruler’s funeral feast (erfi)” in my dissertation, Sundqvist 2002.

made ready a hall in no wise smaller or less stately than [the hall] Uppsalr, which he called the Hall of Seven Kings (sjau konunga sal). In it were erected seven high-seats. King Ingjaldr sent messengers through all of Svetjud, inviting kings, earls, and other prominent men. To this funeral feast (erfi) came King Algautr, Ingjaldr’s father-in-law, and king Yngvarr of Fjaðryndaland, and his two sons, Agnarr and Álfr. Also King Sporsnjallr of Næríki and King Sigverkr of Áttundaland. Only King Granmarr of Suðrmannaland did not come. There, the six kings were assigned seats in the new hall. One high-seat that King Ingjaldr had had erected remained empty. All the host that had come there were given seats in the new hall. But for his bodyguard and all his people King Ingjaldr had made room in the [old hall called] Uppsalr. It was custom at that time, when a funeral feast was prepared to honour a [departed] king or earl, that the one who prepared the feast and was to be inducted into the inheritance, was to sit on the step before the high-seat (á skǫrinni fyrir hásætinu) until the beaker ( full) called the bragafull was brought in; and then he was to stand up to receive it and make a vow (strengja heit), then quaff the beaker, whereupon he was to be inducted in the high-seat (í hásæti) which his father had occupied. Then he had come into the [rightful] inheritance to succeed him. So was done here; and when the beaker was brought in, King Ingjaldr stood up, seized a large drinking horn (dýrshorni), and made the vow that he would increase his dominion to double its size in every direction, or else die. Then he emptied the beaker. Now when everyone was drunk, King Ingjaldr told Fólkviðr and Hulviðr, the sons of Svipdag, to arm themselves and their men when evening approached, as was planned. They went out to the new hall and put it to the torch; the hall blazed up, and the six kings and all their followers were burned [inside]. Those that tried to come out were quickly cut down. Thereupon King Ingjaldr took possession of all realms these kings had ruled, and levied tribute on them.3

The text tells further about the banquet that King Granmarr of Södermanland held for King Hjǫrvarðr of Denmark. It resulted in a companionship between them. King Ingjaldr thus felt threatened and therefore collected a force intend- ing to proceed against these kings. Then they fought a hard battle, but Ingjaldr had to return to Uppsala. When Granmarr and Hjǫrvarðr made a royal progress at their estates in the island called Selaön (Sili) in Lake Mälaren, Ingjaldr fell upon them one night with his army. They surrounded the house and burned

3 Heimskringla I, Ísl. Fornr. 26, 66f. Trans. Hollander, modified. 478 chapter 13 them inside with all their followers. He also subjected all their dominions and set chieftains over them. Later, the same fate struck Ingjaldr himself. When visiting a hall at Ræningr he was surrounded by King Ívarr and his force. So he and his wife Queen Ása made the decision which has since been famous. They got all their men dead drunk and then set fire to the hall. The hall burned down with all people inside. The story of Ingjaldr is the most elaborate about any single ruler in Ynglinga saga. In Ynglingatal there is one stanza about Ǫnundr’s death and burial,4 and one about Ingjaldr himself.5 None of them mention Ǫnundr’s erfi feast. The latter stanza tells only that Ingjaldr committed suicide and let himself be burnt inside a house at Ræningr. Historia Norwegiæ does not add much more information about these kings.6 Ingjaldr inn illráði is also enumerated among Ari’s forefathers in his langfeðgatal, but we do not get any detailed information about him there either.7 Although Snorri’s story about Ingjaldr’s erfi feast cannot be verified by other sources, some scholars have been inclined to regard it as partly built on old oral traditions, perhaps with roots in eastern Scandinavia. Elias Wessén, for instance, thought that the naming custom in Snorri’s text, e.g. Fólkviðr and Hulviðr, is authentically Swedish, and thus indicates that the tradition had an eastern origin.8 He argued that names with -viðr are unknown in Iceland and rare in Norway. They are on the other hand very common in the region of

4 “Ǫnundr was stopped by the mass of falling stones at Himinfjǫll. The bastard’s hatred was very heavy for the enemy of the Estonians, and the one who killed Hǫgni was covered under the bones of the earth.” Varð Ǫnundr/ Jónakrs bura/ harmi heptr/ und Himinfjǫllum,/ ok ofvæg/ Eistra dolgi/ heipt hrísungs/ at hendi kom,/ ok sá frǫmuðr/ foldar beinum/ Hǫgna hrørs/ of horfinn vas. My trans. based on Wessén 1964. Cf. SkP I, 42. 5 “And at Ræningr the one who sends smoke gushing trod down the still living Ingjaldr, when the housebreaker on stockinged feet of flame stepped right through the man of divine descent. And this fate most fitting seemed to all Svear for scion of kings: to die first in fiery death and end first his own brave life.” Ok Ingjald/ ífjǫrvan trað/ reyks rǫsuðr/ á Ræningi,/ þás húsþjófr/ hyrjar leistum/ goðkynning/ í gǫgnum steig.// Ok sá yrðr/ allri þjóðu/ sanngǫrvastr/ með Svíum þótti,/ es hann sjalfr/ sínu fjǫrvi/ frœknu fyrstr/ of fara vildi. Trans. Foote, in Jónas Kristjánsson 1988, 95 and Hollander. Cf. SkP I, 44. 6 It has been suggested that Snorri took his information for this account from the original Skjǫldunga saga or possibly Þáttr af Upplendingakonungum. The lacuna in Arngrim’s manuscript may possibly have covered this story. Lindqvist 1936, 270–273; Beyschlag 1950, 80–83; Bjarni Guðnason 1963, 90–95. Ingjaldr is mentioned in the Sǫgubrot af Skjǫldungasaga (Ísl. Fornr. 35, 56) and he appears among Haraldr Finehair’s forefathers in Ágrip af Sǫgu Danakonunga (Ísl. Fornr. 35, 325). 7 Íslendingabók, Ísl. Fornr. 1, 27. 8 Wessén 1964, xi, 72. The Arenas for Royal Funeral Feasts and Inauguration Rituals 479 the Svear. There are also authentic place-names in the Ingjaldr account that can be identified with toponymy in eastern Sweden, such as Myrkvafjǫrðr, i.e. Mörköfjärden in Södermanland, and Sili (OSw Sila 1327), i.e. Selaön in Lake Mälaren.9 At Selaön both place-names and evidence from archaeology reflect an organized Late Iron Age society, including royal farms. Close to each other are Tuna, Husby and Karleby, which indicate power and centrality. Other names indicate pre-Christian beliefs and cultic practices, e.g. Fröberga, Fröslunda, Kilfröslunda, Lytislunda, Ullunda and Odensicke.10 Snorri’s information that King Granmarr of Södermanland and King Hjǫrvarðr took banquets (taka veizlu) on their farms (at búum sínum) on the island Selaön (Sili) in Mälaren, is not historical, of course. On the other hand, it is reasonable to assume that such royal banquets may have taken place there during the Late Viking Age and the Early Middle Ages. Snorri’s comment that Uppsala was an important royal site, with at least one great hall building, harmonizes with archaeological evidence (see ch. 4 above). This information is probably also built on more reliable oral traditions from the east. The strongest argument for the suggestion that the Snorri’s account was based on eastern traditions is a quite obscure place-name mentioned in Ynglingatal’s account of Ingjaldr, which could not have been common knowledge for the ninth-century Norwegian skalds. The place-name included in an expression of Ynglingatal’s Jǫfraskinna version, á rauningi, corresponds directly to the sequence a rauniki (a Rauningi) quoted from a Viking Age runic stone from Aspa löt in Södermanland. This name is related to Rönö hundred (OSw Røna hundare) and the parish name Runtuna (OSw Røntuna), which both appear in the same area of Södermanland.11 Nevertheless there are several fictious and literary elements at stake in the Ǫnundr-Ingjaldr-account, which do not have support in direct sources. It seems as if Snorri wanted to describe how the decentralized Svetjud, with several district kings, was unified under one monarch, in a similar way as King Haraldr unified Norway to one kingdom.12 Some scholars have argued that Ynglinga saga in this passage was subjected “to typification on the basis of the ideology of early medieval kingship” and that King Ǫnundr was depicted as an

9 Cf. Brate 1914, 86; Wessén 1964, 72. 10 Brink 1999a, 427–428; 1999b, 44–46; 2001, 93–94. 11 S. B. F. Jansson 1987, 105. It has been suggested that it is a corrupt form of Rällinge in the parish of Helgarö, Södermanland, but this is not plausible. Wessén 1964, 72f. It probably refers to a name Røning(i) in Södermanland, most likely designating an area equivalent to Rönö hundred. Personal communication from Vikstrand. 12 Wessén 1964, 71. 480 chapter 13 archetype of the good medieval ruler.13 He cleared new land and built many roads. He established great provinces and had a mansion (bú) in each province (stórherað). He went around the land on circuit (at veizlum), arranged assemblies (þing) and held great sacrifices (blót) at Uppsala.14 Snorri’s statement that Ǫnundr established great provinces and mansions in each province is not reliable. Such a political process must have taken several generations to build up. Bruce Lincoln has recently empasized the literary function of the Ingjaldr account not only in Ynglinga saga, but for Heimskringla as a whole text. It heralds the story about King Haraldr in Haralds saga ins hárfagra. The text thus constructs a clear parallel between Ingjaldr and Haraldr, “two men who began their careers as district kings, but who, by virtue of vows and subsequent ruthless action, transformed themselves into full-fledged monarchs, conquering and uniting their nations . . . Heimskringla’s portrait of Ingjald’s viciousness and treachery, and also its description of the oath he swore at the start of his reign, seems to have been carefully crafted to resonate with later events: not only Harald Fairhair’s state-founding conquest—that much is obvious—but also Norway’s designs on Iceland in the mid-thirteen century. Thus, through the story of Ingjald, the text argued that national unification under a single monarch was a project that could be accomplished only by violence and treachery, and that a king capable of this project was a wolf-hearted miscreant, situated halfway between man and beast.”15 It is thus true that Snorri’s account cannot be trusted as factual history, since it to a great extent has an ideological and literary function in Heimskringla. Even if Snorri’s account of Ingjaldr has been revised through medieval perspectives, methods, and “patterns”,16 it seems to be at least partly based on older traditions, perhaps deriving from the Mälaren area. This tradition included authentic Swedish place-names and personal names, for instance. Also the existence of a great banqueting hall in Uppsala, and dated to the Merovingian Period, may indicate that the Ingjaldr story in some details is old and based on eastern oral traditions. In my opinion also the descriptions of some rituals may reflect Late Iron Age and Early Medieval conditions, and thus be based on older traditions.

13 Norr 1995; 1998, 111–151. 14 Ynglinga saga 33, Heimskringla I, Ísl. Fornr. 26, 62f. 15 Lincoln 2014, 99–101. 16 On such medieval aspects and the relationship between Ingjaldr in Ynglinga saga and Haraldr in Haralds saga ins hárfagra, see most recently Lincoln 2014, 95–103. The Arenas for Royal Funeral Feasts and Inauguration Rituals 481

13.2 Libations at Funeral Feasts

Snorri describes in general terms the memorial drinking rites (ON minni, full) connected with the erfi feasts. These rituals have been hotly debated in research. The most common view among scholars is that the practice had its origin in the pre-Christian period but continued into the Middle Ages.17 Maurice Cahen argued that minni (n.)18 replaced the older word full (n.),19 sometime between the tenth and the thirteenth centuries: “Minni s’est substi- tué à full dans tous les sens et dans toutes les expressions de la phraséologie.”20 When the pagans toasted their gods they called it Óðins full, Freys full, while Christians drank Krists minni and Mikjáls minni. There was thus a cultic continuity in the libation rituals, but not in the terminology: “Le rite n’a pas varié, seul le nom a changé.”21 Klaus Düwel opposed this line of reasoning; when the saga writers were describing pre-Christian drinking ceremonies, they actually had in mind customs common to the medieval guilds.22 According to Düwel, full is an old word but the meaning is unclear in the earliest sources, while the meaning of minni—‘drink, toast in memory of a person’—first appears clearly in the Middle Ages (1200) in Old Norse prose literature.23 The meaning of minni ‘drink, toast in memory of a person’ appears clearly first during the Middle Ages (AD 1200) in Old Norse prose literature.24

17 E.g. Cahen 1921; Bø 1960; B. Stoklund 1960; L. Carlsson 1965, 125; Grønvik 1990; Dillmann 1997. On mythical aspects of ritual drinking, see particularly Drobin 1991. Cf. Rydving 1996. 18 ON minni (n.) ‘1) hukommelse, evne til at mindes; 2) minde, hvad man mindes; 3) fyldt bæger som naar man drak nogen til etc.’ Fritzner 1954, Vol. II, 701–703. 19 ON full (n.) ‘fyldt bæger’. Fritzner 1954, Vol. I, 500. 20 Cahen 1921, 179. 21 Cahen 1921, 179. 22 Düwel states: “Für die detallierte Beschreibung von Trünken auf die Götter als einer fes- ten Einrichtung im Opferablauf rechne ich mit einer Übertragung, einer Rückprojektion der christlichen Sitte des Minnetrinkens ins Heidentum.” Düwel 1985, 52–60, 119. Cf. O. Olsen 1966, 60; Holtsmark 1992, 32. 23 Düwel 1985, 70–118. Simek (2006, 217) follows a similar line of reasoning: “The custom of ‘drinking in memory of someone’ only reached Scandinavia in the High Middle Ages from Germany, and therefore it should not be seen as a heathen custom.” Other scholars have partly moderated Düwel’s criticism. See e.g., Hultgård 1993; Dillmann 1997; Sundqvist 2005f. and 2013b. 24 Snorri distinguishes between minni and full in Hákonar saga góða. He uses full for the gods (Óðins full and Njarðar full) and minni for kinsmen (Menn drukku ok full frænda sinna, þeira er heygðir hǫfðu verit, ok váru þat minni kǫlluð). Heimskringla I, Ísl. Fornr. 26, 168. But Snorri is not consistent. In Óláfs saga helga he uses minni for the gods: at þar væri minni ǫll signuð Ásum at fornum sið. Heimskringla II, Ísl. Fornr. 27, 177. Cf. Cahen 482 chapter 13

Düwel is probably right when arguing that the term minni with the meaning ‘drink, toast in memory of a person’ never appeared in pre-Christian Scandinavia. This word probably arrived in Scandinavia in a Christian context. But even if the term minni is young and has a Christian origin, it is very likely that some type of drinking ceremonies were held when celebrating dead kinsmen and the gods in ceremonial buildings during the Viking Age.25 The term full or bragafull could very well have been applied in such contexts.

13.2.1 Drinking ærvisøl in Svetjud Terms and expressions in Swedish provincial laws confirm that drinking ceremo nies attached to funerals existed in eastern Scandinavia during the Middle Ages. In Gutalagen there is the term erfis gierþ ‘graföl’, and in Västgötalagen, Magnus Erikssons Landslag and Magnus Erikssons Stadslag there is the term ærvisøl Sw ‘arföl, gästabud som af arfvingen anställdes till den dödes minne’.26 In modern Swedish the expressions arvöl and gravöl are still in use. That these drinking customs connected to funerals also existed in more ancient periods is most likely. One piece of evidence for this may be found in the Arab traveller Ibn Fādlan’s account of the people called Rūs. He described the funeral of a rich and important man among them, which took place near the Volga in the tenth century. He mentions that the people drank a fermented drink called nabîdh on the day that a slave woman was killed and cremated together with her master.27 Perhaps it reflects ancient customs in eastern Scandinavia, as the people called Rūs are often regarded as the Svear.28

13.2.2 Funeral Feasts (erfi) in Norway The concept of ‘funeral-beer, inheritance-beer’ (ON erfiǫl, ǫl, erfðaǫldr, sáluǫl, sjaund, sjaundargerð) appears in medieval law texts from Norway. Den ældre Gulathings-Lov 23 says: “And when men are dead, and the heir will make beer after [them].”29 In Christian era this beer was also called a sáluǫl “soul-beer” and a priest was to participate in these ceremonies.30 Medieval sources referring

1921, 177–179. The pre-Christian origin of Germanic memory-drinking in general is not discussed by Düwel 1985. 25 Cf. Meulengracht Sørensen 1991b, 239; Hultgård 1993, 233. 26 Schlyter 1887, 138, 777. Cf. Cahen 1921, 86–91. 27 Ibn Fadlān. (Ed.) Wikander, 65–69. Cf. Foote & Wilson 1980, 408ff.; Montgomery 2000. 28 The origin and evolution of the name Rūs is however debated. On the scholarly contro- versy, see Montgomery 2000, 1ff., 23. 29 En hvervetna þess er menn verða dauðer. oc vill ervingi ol efter gera. Den ældre Gulathings- Lov, NGL 1, 14. 30 Den ældre Gulathings-Lov. NGL 1, 14f. Cf. Düwel 1985, 80–82. The Arenas for Royal Funeral Feasts and Inauguration Rituals 483 to early medieval conditions thus show that libations were made during the inheritance feast. Such a ceremony ought to take place soon after death.31 It could be held on the same day as the funeral, or on the seventh or the thirtieth day after death or even later. Common people in Norway did not need to hold funeral feasts; it was required only for rich farmers and rulers, as it carried legal implications for the inheritance.32 According to several runologists, an early ceremonial funeral feast (erfi), including beer ceremonies, perhaps occurs also in an aristocratic context in the Tune inscription (fifth century), Norway. James E. Knirk recently read and interpreted it thus:

A: ekwiwazafter · woduri / – – dewita(da(halaiban : worahto : ? . . . B: – ????woduride : staina · / – – – þºrijozdohtrirz}dalidun / B: arbijaºsijostezarbijano

I Wiwaz, in memory of Woduridaz, ‘provider of bread’, wrought [the runes]. ??? the stone to/for Woduridaz. Three daughters prepared the funeral celebration (or: carried out the inheritance), the most beloved/most devout/most ??? of heirs.33

It is possible that Marstrander is right when he claims that WoduridaR in this inscription was a local ruler, since he is called wita(da(halaiban ‘he who provides (takes care of) bread’, ‘he who secures bread’ (cf. OE hlāf-weared, hlāford ‘lord’).34 WoduridaR was possibly also a cult officiator, since his name may be related to Óðinn.35 }dalidun arbija might be interpreted as dāliðun arbija

31 Cf. Ólafur Briem 1956; Bø 1960; B. Stoklund 1960. 32 Bø 1960. 33 Knirk 2006, 333. Cf. Krause 1966, 162–167; McKinnell & Simek 2004, 162; Spurkland 2005, 37ff. 34 Witanda- is a present participle with a meaning ‘the one who secures the bread’. Marstrander 1930, 327–338. Cf. Krause 1966, 165; Grønvik 1981, 91, 94, 148, 166. On OE hlāford, see Clark Hall 1916, 159; Enright 1988, 188f.; 1996, 20f.; Spurkland 2005, 38. 35 The name WōdurīdaR may correspond to an ON *Óð-ríðr. The first element is equivalent to the ON adjective óðr ‘frenzied, raging, raving: mad with fear, angry, insane, violent’ and the second element is ríðr ‘rider’, thus possibly a heiti of Óðinn ‘the frenzied (violent) rider’. Cf. Spurkland 2005, 38. 484 chapter 13

“arranged/prepared the funeral feast/beer decently” (“Erbmahl, Erbtrunk”.36 Wolfgang Krause stated that the name of the carver should be interpreted as a proto-Nordic Wīwar, which may contain an old appellative *wīwar.37 This term is a construction of the root *weik- ‘to separate’ and related to Gothic weiha ‘arch-priest, high-priest’ and the Old Norse verb vígja ‘to consecrate’. According to Krause, it could mean ‘the consecrated one’ or ‘the one who consecrates’ (‘der Geweihte’ or ‘der Weihende’).38 Ottar Grønvik, who has made a thorough study of the Tune inscription, suggested that WiwaR inherited the property of the magnate and cult officiator WoduridaR, who was his grandfather. WiwaR together with his three aunts arranged the funeral feast.39 It contained several religious-juridical ceremo nies, such as ritual toasts and the delivery of the poem of inheritance. WiwaR became the major legal heir and the new cult leader. Grønvik argued that the name WiwaR on the Tune stone should be read with a short vowel in the root syllable, *Wiwar, and be interpreted as “ ‘den som vier (helger, signer)’ og betegne en slags kultleder”.40

13.2.3 Drekka erfi in Iceland There are also some Sagas of Icelanders that suggest that the custom of drinking funeral beer (drekka erfi) was firmly rooted in the early Icelandic society. Laxdœla saga reports that people drank erfi for the Christian female magnate Unnr “depth-of-thought” (Unnr in djúpúðga): var nú drukkit . . . erfi Unnar.41 This is also implicitly mentioned in the Sturlubók version of Landnámabók, where Auðr (i.e. Unnr) said when the wedding feast of Óláfr feilan, her grandchild, had been celebrated, that “the feast would go on for another three days and that it would be her funeral feast”.42 The expression drekka erfi occurs also in Flóamanna saga 13.43 Also very interesting is a passage in Eyrbyggja saga 54, which says that the great farmer Þóroddr and his companions were drowned

36 See Krause 1966, 166. Cf. Grønvik 1981; Marstrander 1930. 37 Krause 1966, 164f. 38 Krause 1966, 165. Cf. Peterson 2004, 19, who also presents alternative interpretations. See also Vikstrand 2009, 10f. 39 Grønvik 1981, 162–191. Krause argued that Woduridaz was a goði who kept a hof at Tune where Óðinn was worshipped. Before he died he was an ársæll chieftain, called *vitand- hleif i.e. ‘he who takes care of bread’, ‘bread-securer’. Krause 1966, 166f. 40 Grønvik 1987, 54f.; cf. 1981. 41 Ísl. Fornr. 5, 13. On the great erfi feast after Hǫskuldr, see Laxdæla saga, Ísl. Fornr. 5, 73ff. 42 . . . at þá skyldi standa veizlan enn þrjár nætr; hon kvað þat vera skyldu erfi sitt. Landnámabók, S110, Ísl. Fornr. 1, 146. 43 Flóamanna saga, Ísl. Fornr. 13, 257. Fritzner 1954, Vol. I, 349. The Arenas for Royal Funeral Feasts and Inauguration Rituals 485 off Nes. Soon after this event his family arranged a funeral feast, at which they used the jól ale for the drinking ceremonies (var þá tekit jólaǫl þeira ok snúit til erfisins). On the first evening of the feast, when all guests were seated, Þóroddr and his companions came into the room drenched to the skin. “Everyone welcomed Þóroddr and his men, and thought this a happy omen because in those days it was believed that drowned people had been well-received by the sea- goddess, Rán, if they came to their own funeral feast.”44 It has been argued that this text indicates that the erfi feast, in previous periods, clearly was set in a pre-Christian religious context.45 The detail about Rán has no counterpart in any other tradition about pre-Christian religion, however. I must therefore agree with Thomas A. DuBois when he states that this account “seems created to underscore the conflict of old and new understandings of death during the era of conversion”.46 Anyhow, we may conclude that drinking ceremonies were performed during the pre-Christian funeral feasts in all three investigation areas, Svetjud, Trøndelag and Iceland. Probably such customs were more common in elite milieus, for instance, among kings, chieftains and rich farmers. The Eddic poem Guðrúnarhvǫt 8, which probably is quite ancient, indicates such customs when it says: “that you are drinking a funeral feast for us all” (at þú erfi at ǫll oss dryccir).

13.3 The Place for Drinking Beer at the Funeral Feast

The Kings’ Sagas and the Sagas of Icelanders report that the pre-Christian drinking ceremonies took place in specific buildings for the purpose. Accord ing to these medieval texts these buildings could be referred to as drinking halls. Snorri states in Ynglinga saga 36 that King Ingjaldr “made ready a hall in no wise smaller or less stately than [the hall] Uppsalr, which he called the Hall of Seven Kings (sjau konunga sal)”. In this hall seven high-seats were erected for minor kings representing the provinces of Svetjud, Fjaðryndaland, Áttundaland, Næríki and Suðrmannaland. This story is impaired by fictional elements (see above), but it is nevertheless interesting that Snorri thought that

44 Menn fǫgnuðu vel Þóroddi, því at þetta þótti góðr fyrirburðr, því at þá hǫfðu menn þat fyrir satt, at þá væri mǫnnum vel fagnat at Ránar, ef sædauðir menn vitjuðu erfis síns; . . .” Eyrbyggja saga 54, Ísl. Fornr. 4, 148. 45 See Ólafur Briem 1956. 46 DuBois 1999, 90. 486 chapter 13 the royal erfi feast should take place in a specific building which was equipped with high-seats for the ruling elite. In Ynglinga saga, Snorri associated the term ON salr in the place-name ON Uppsalr with a hall building. This is one of few times the name is written in the singular. It is usually attested in the plural, Upsalir, Upsala or Upsǫlum.47 The meaning of the place-name element Old Swedish sal in Uppsala is in fact hotly debated. Snorri’s idea that ON -salr at Uppsala refers to a hall building(s), has gained support from philologists (Stefan Brink) and archaeologists (e.g. Bo Gräslund and Frands Herschend). Since features of a Merovingian Period hall has been discovered at Uppsala, it has been suggested that the meaning of the Old Swedish place-name refers to it, and that sal is used in the sense of ‘banqueting hall’.48 The most common and traditional interpretations of sal in the name Uppsala among specialists (e.g. Jöran Sahlgren and Mats Wahlberg) is ‘occasional dwelling-place’, ‘simple building’, perhaps some kind of meadow-barn.49 There is, however, strong evidence that ON salr in skaldic poetry often denotes ‘banqueting hall’.50 In Eddic poetry its meaning is more ambivalent,51 but in other Germanic languages equivalents of the word denote ‘hall of banquets

47 The singular form appears in Ynglinga saga, in Heimskringla I, Ísl. Fornr. 26, 66 and perhaps in Vg 119. 48 B. Gräslund 1993, 179–184; Brink 1999a, 38f., 48f., particularly note 15; Herschend 2001, 39ff. 49 E.g. Sahlgren 1953, 26–32; Wahlberg 1994, 15. See also SOL, 267, 338, which refers to both meanings. 50 In e.g. Sigvatr’s Austrfararvísur st. 16, the expression hilmis salr is used as a synonym for the word hǫll in a royal context. Also in Eiríksmál the word salr appears as a synonym for hǫll, but in a mythical-royal context. 51 ON salr does not always refer to a hall in the Eddic poetry, but to a house consisting of one room, thus also used for a house of a poor man, eigi . . . taugreptan sal “own . . . a tow- raftered hut” (e.g. Hávamál st. 36). It might also represent an abode of gods, giants and other beings, e.g. í Suttungs sǫlom (Hávamál st. 104) or just a ‘ground’ (cf. Lat solum) as in Vǫluspá st. 4 á salar steina “on the stones of the ground”. Eddic poetry thus allows for wider connotations of the word. However, it often has the meaning of “banqueting hall”. In Vǫluspá st. 37 the biórsalr ‘beer hall, drinking hall’ must be a banqueting hall, like the salr in Skírnismál 16. References in Vǫlundarkviða st. 7, 16 and 30 and in Þrymskviða st. 27, to endlangr salr “the length of the hall” seem to connote a long building. Hávamál st. 152 has a salr which must be interpreted as ‘hall’. On salr in Gylfaginning 14, see ch. 4 above). Gylfaginning 35 reports that Frigg lived in a place called Fensalir. Cf. Brink 1996b, 255–258; Brink 1999a, 27–30. For the expression eigi . . . taugreptan sal in particular, see Herschend 2001, 39–60. The Arenas for Royal Funeral Feasts and Inauguration Rituals 487 or gatherings’.52 Whether sal in the name Uppsala refers to a banqueting hall is, however, uncertain. Per Vikstrand states that further research must be done before we can convincingly interpret this place-name.53 We may nevertheless assume that ritual drinking and other erfi ceremonies, really took place in banqueting halls in Svetjud during the Late Iron Age. Archaeological investigations in the central parts of Sweden may support this assumption. As mentioned in chapter 9, fragments of expensive ceremonial glass have been found at Late Iron Age halls, for instance, at Helgö, Birka and Lunda. In Norway too, ritual drinking and aristocratic erfi ceremonies most likely took place in large halls. Traces of ceremonial feasting are attested at the hall of Borg, Lofoten. The information in Hákonar saga góða 14 about the ceremonial drinking in memory of departed kinsfolk (i.e. minni ‘memorial toast’), at the hof sanctuary of Lade in Trøndelag may also be related to such notions. Most likely the erfi feast in Iceland took place in similar milieus and buildings. Eyrbyggja saga 54, for instance, reports that that the erfi feast of Þóroddr was held in a room called eldskáli. The medieval saga author even gave information about some details of this building: “Þóroddr and his men walked across the main room (setaskáli), which had two doors, and into the living- room (eldskáli). They ignored the greetings people gave them and sat down at the fire. The people ran out of the living-room, but Þóroddr and his men stayed on until the fire began to burn very low, then went away.”54 According to the medieval scribe, the eldskáli seems thus to be inside the setaskáli. Norrøn ordbok gives the following meaning for the Old Norse word setaskáli, ‘der ein sit i lag med gjestene sine, seteskåle’,55 while ON eldhús is explained as ‘eldhus, koke- og soverom’.56 This eldhús may also be the entertainment room of the main house of the farm in Iceland.57 Most likely it was in this room the ceremonial drinking was performed during the erfi feast, during both the Viking Age and the Early Middle Ages. According to Laxdæla saga 7, there was also a banqueting house (skáli) at Unnr’s farm called Hvammr. It is stated that people were drinking in this house

52 OE sele refers to “hall, house, dwelling, prison”. Clarke Hall 1916, 260. OHG, MHG sal ‘die alte Bezeichnung des germ. Einraumhauses, auch des bäuerlichen’. Kluge 1957 (1883), 617. 53 Vikstrand 2013. 54 Þeir Þóroddr gengu eptir endilǫngum setaskálanum, en hann var tvídyrðr; þeir gengu til eldaskála ok tóku einskis manns kveðju, settusk þeir við eldinn, en heimamenn stukku ór eldaskálanum, en þeir Þóroddr sátu þar eptir, þat til er eldrinn var fǫlskaðr; þá hurfu þeir á brott. Ísl. Fornr. 4, 148. 55 Heggstad et al. 1993, 362. 56 Heggstad et al. 1993, 91. 57 See also Meulengracht Sørensen 2003. 488 chapter 13 during the evening when the combined wedding and funeral feast (brullaup Óláfs ok erfi Unnar) was celebrated.58 Most likely this house (skáli) resembled the setaskáli or eldskáli mentioned in Eyrbyggja saga 54. When funeral feasts were held in memory of great magnates it was probably necessary to hold such feasts in large buildings in all parts of Scandinavia. It should be noticed, however, that in Old Norse sources the Icelandic farmer’s banqueting house called skáli was set in direct opposition of the king’s feasting building called hǫll in Norway and Svetjud.59

13.4 The High-Seat and Inheritance

As noted in chapter 6 above, several Old Norse sources indicate that the high- seat was the ritual-symbolic centre during the ceremonies performed in the hall building. The rituals which took place at the high-seat had both judicial and religious dimensions.60

13.4.1 Judicial Aspects of the High-Seat In Snorri’s account of Ǫnundr the high-seat of the banqueting house has a central judicial function. Snorri states that after drinking the libations, the heir should sit “in the high-seat which his father had occupied. Then he will have come into the [rightful] inheritance to succeed him.” Fagrskinna has similar expressions when describing the inheritance feast (veizluerfi) that Sveinn Forkbeard (tjúguskegg) held for his father Haraldr Gormsson Bluetooth in Ringsted, Zealand. At the same time and place was held the erfi of Earl Strút- Haraldr (Earl Sigvaldr’s father). Sveinn invited many noble men and also the Jómsvíkingar. In this account the scribe of Fagrskinna (manuscript *B) distinguished between the erfi custom in general and the particular erfi feast of Haraldr and Strút-Haraldr. In the general part it is stated that the heir should sit in the seat of the person who had left the inheritance. The text goes as follows:

If the funeral feast was held according to old custom, it had to take place in the same year as the death of the man from whom the inheritance came. And he who had arranged the funeral feast was not to sit in that man’s seat from whom he inherited until the men had drunk the funeral beer. And the first night, when the men came to the funeral feast, many

58 Laxdæla saga, Ísl. Fornr. 5, 12f. 59 Carsten 2014, 151. 60 Cf. Birkeli 1932; Steinsland 1991, 66–86; Sundqvist 2014. The Arenas for Royal Funeral Feasts and Inauguration Rituals 489

drinking cups should be filled, in that way that memorial cups (minni) are drunk nowadays; and they devoted that cup ( full) to their most powerful kinsmen or to Þórr or some other of their gods, as long as they were pagan; and in the last place there should be filled the bragafull and then he who held the funeral feast had to make a promise at the bragafull and so should all who participated in the funeral feast. And then the heir would enter into the seat of the man who had left the inheritance. And then the inheritance was fulfilled and the praise was accomplished after the dead man, but not before that.61

Not until the heir had drunk the bragafull and made the oath was he allowed to enter the high-seat. Snorri was also acquainted with this tradition about King Sveinn and reproduced a similar ritual of the erfi feast in his Óláfs Saga Tryggvasonar 35:

King Sveinn arranged a great feast, requesting the presence of all the chieftains in his realm. He intended to honour his father Haraldr with a funeral feast, and enter into his inheritance. . . . On the first day of the banquet, before King Sveinn ascended the high-seat of his father, he drank to his memory and made the vow that before three years had passed he would invade England . . .62

61 En þá er erfin váru gǫr eptir fornum sið, þá skyldi þat skylt at gøra þau á því ári, er sá hafði andazk, er erfit var eptir gǫrt. En sá er gøra léti erfit, skyldi eigi fyrr setjask í þess manns sæti, er hann erfði, en menn drykki erfit. Et fyrsta kveld, er menn kómu til erfis, skyldi skenkja upp mǫrg full með þeima hætti sem nú eru minni, ok eignuðu þau full enum ríkustu fræn- dum sínum eða Þór eða ǫðrum guðum sínum, þá er heiðni var, en síðast skyldi upp skenkja bragafull ok þá skyldi sá, er erfit gørði, strengja heit at bragafulli ok svá allir þeir at erfinu væri, ok stíga þá í sæti þess, er erfðr var, ok skyldi þá fullkominn vera til arfs ok virðingar eptir enn dauða, en eigi fyrr. Fagrskinna, Ísl. Fornr. 29, 124f. My trans. There are two versions of this story in the manuscripts (*A and *B). I follow the older version *B, which is given in Ísl. Fornr. 29. On the witnesses, see ch. 6 above. 62 Sveinn konungr gerði mannboð ríkt ok stefndi til sín ǫllum hǫfðingjum þeim, er váru í ríki hans. Hann skyldi erfa Harald, fǫður sinn. . . . Fyrsta dag at veizlunni, áðr Sveinn konungr stigi í hásæti fǫður sins, þá drakk hann minni hans ok strengði heit, áðr þrír vetr væri liðnir, at hann skyldi kominn með her sinn til Englands . . . Óláfs Saga Tryggvasonar 35, in Heimskringla I, Ísl. Fornr. 26, 273f. See also Jómsvíkingasaga (exists in five versions). Flateyjarbók I, 179–181. These sagas, as well as the general part of the erfi in the Ingjaldr account, probably depend on the Fagrskinna redaction A. Cf. Düwel 1985, 76–79. 490 chapter 13

Snorri’s and Fagrskinna’s texts thus intimate that the funeral feast concerned not only the dead ruler and the inheritance but also the inauguration of the successor. After the libations and the oaths, the heir was to take his place in the high-seat. The ceremony was then complete in both the legal and the religious sense. There are several attestations of the heir being placed in the high-seat at inheritance ceremonies in the medieval Scandinavian literature. However, most of them describe medieval conditions, especially in Norway. Den ældre Gulathings-Lov says: “Now a man is dead and the heir must be seated in the high-seat.”63 Den ældre Frostathings-Lov states: “A person should dispose his estate as long as he is able to sit in the high-seat, both woman and man.”64 Similar ideas are expressed in other medieval sources referring to Norway.65 Flateyjarbók tells about King Hákon Hákonarson: “then he [Hákon] could go there and require from them inheritance and the high-seat and instead be seated there himself.”66 In Iceland too, it seems as if the high-seat played a ritual role during the inheritance. The high-seat of the murdered chieftain Ingimundr of Vatnsdalr, in north-western Iceland, plays an important role in Vatnsdœla saga. Þorsteinn told his brother that they should not be seated in their father’s high-seat, neither at home nor in the house of men, until their father was avenged.67 The sons of Ingimundr were at home that winter and sat on the benches of lower rank (sátu á inn óœðra bekk).68 When the father had been avenged, Þorsteinn said that they could now enter their father’s high-seat.69 Emil Birkeli was thus partly right in his statement that it was in a concrete sense the hásæti was inherited. The heir inherited not only his predecessor’s

63 Nu er maðr dauðr. arve scal i ondvege setiazt. Den ældre Gulathings-Lov. NGL 1, 51. 64 Siálfr scal hverr ráða fé síno meðan hann má sitia í öndvegi síno. svá cona sem carlmaðr. Den ældre Frostathings-Lov. NGL 1, 213. 65 Cf. Ældre Eidsivathings-Lov: Sitiæ skal han i andueghe ok kona hans i hia honum. NGL 1, 404. Nyere Lands-Lov states: Nu er maðr dauðr. Þa skal arfe i anduegi setiazt. NGL 2, 86. Cf. Birkeli 1932, 44–46; Frense 1982, 195–197; Steinsland 1991, 78–81. 66 Þa mætti hann ganga til ok krefia (þa) arfs ok or haasæti enn setiaz sialfr i stadinn. Flateyjarbók, III, 12. 67 “Þat sýnisk mér ráð, at vér setimsk eigi í sæti fǫður várs, hvárki heima né í mannboðum, meðan hans er óhefnt.” Vatnsdæla saga 23, Ísl. Fornr. 8, 63. 68 Vatnsdæla saga 24, Ísl. Fornr. 8, 65. 69 “. . . nú þykki mér ráð, at vér fœrim sess várn í ǫndvegi fǫður várs.” Vatnsdæla saga 27, Ísl. Fornr. 8, 71. The Arenas for Royal Funeral Feasts and Inauguration Rituals 491 dignity and property, but also the high-seat. It is therefore interesting to note that the high-seat in Old English was called yrfestōl ‘hereditary seat, home’.70

13.4.2 Religious Aspects of the High-Seat It is possible that the high-seat also reflects some religious aspects (see ch. 6 and 11 above). Most important in this context are the gold foils found at the assumed high-seats of hall buildings in both Svetjud and Norway. These religious-ritual objects indicate that these seats were considered as important and even sacred in some sense. Perhaps these seats were regarded as a place where the real and divine worlds were encountered. The association of the sovereign god Óðinn with the high-seat in the mythic world emphasizes that this place was intended for rulers. It is therefore possible that the human heir of a ruler inherited some religious functions, responsibilities and abilities when he was placed in the high-seat during the erfi feast. After these ceremonies at the high-seat he was thus installed in a cultic office. That the ceremony had religious connotations may be attested by Fagrskinna’s description of the general inheritance feast. The libation mentioned there was not only made to the noblest kinsmen, but also to Þórr or some other gods as long as the people were pagans. The interference of the pagan gods in this ceremony may also indicate that this custom was old. Snorri’s text about King Sveinn states, however, that the memory cups were drunk to Christ (drekka Krists minni) and the Archangel Michael (Mikjáls minni), but also to the forefathers.71

It thus seems as if the inheritance ceremonies mentioned in Ynglinga saga 36 and Fagrskinna 20, i.e. the placing of the claimant in the high-seat, also involve a type of performative inauguration ritual of the new ruler. After sitting on the step before the high-seat, the libations and the oaths (vows), the heir was to take his place in the high-seat. The ceremony was then complete in both the legal and the religious sense. Snorri’s statement in Ynglinga saga 36, “it was custom at that time, when a funeral feast was prepared to honour a [departed] king or earl”, indicates that some ceremonies were only practised by the uppermost social strata, i.e. the kings and earls in Svetjud, Norway and Denmark. It is thus possible that the inaugurations described in Ynglinga saga and Fagrskinna did not occur among chieftains in Iceland.

70 Clark Hall 1916, 372. 71 Óláfs Saga Tryggvasonar 35, in Heimskringla I, Ísl. Fornr. 26, 273f. 492 chapter 13

13.5 Royal Inaugurations, High-Seats and the Great Mounds

Medieval sources often report that Christian kings and earls in Norway and elsewhere were placed in the high-seat at their investiture.72 The most detailed version of this ceremony can be seen in Magnús Lagabøter’s Hirðskrá (c. AD 1270),73 which describes a ceremony called konungstekja74 that could only be performed at the val-thing. Only a rightful claimant (kongsefni) was taken as king;75 after a common service, all men processed to the thing place, where there were several high-seats for the nobles. One of the seats in the middle was the most splendid one, and the claimant was supposed to sit on the step below it. When he was chosen king, he was raised by the nobles and placed in the high-seat.76 The last part of the procedure corresponds to Snorri’s and Fagrskinna’s accounts of pre-Christian ceremonies. Genealogical knowledge is sometimes related to the ritual induction in the high-seat in early Christian contexts. When King Magnús appointed Sveinn Úlfsson (also called Sveinn Ástriðarson) to earl (c. 1047), Sveinn had to sit on the foot-board (á fótskǫrinni) in front of the king who sat in his high-seat (í hásæti), according to Snorri.77 Then the king declaimed in public, that Sveinn “has the birth to be a chieftain” (Hefir hann ætt til þess at vera hǫfðingi). Thereupon the king stood up and gave a sword and a shield to Sveinn, and last he set a helmet on Sveinn’s head and bestowed the title of earl on him. Therupon the king led the earl to share his high-seat with him (Síðan leiddi konungr jarl til hásætis með sér). In this context Snorri also describes Sveinn’s paternal and maternal ancestors: “Sveinn was the name of the son of Earl Úlfr, the son of Þorgils. Sveinn’s mother was Ástríðr, the daughter of Sveinn Forkbeared. She was the sister of King Knútr the Powerful by the same father and of Óláfr Eiríksson of Sweden by the same mother. Their mother was Queen Sigríðr the Haughty, the

72 See e.g., . . . þá leiddi Sigurðr konungr Rogeir jarl til hásætis með sér ok gaf hónum konungs nafn. Fagrskinna, Ísl. Fornr. 29, 318. . . . þá tok Sigurðr konungr í hǫnd hertoganum ok leiddi hann upp í hásæti. Magnússona saga 8, Ísl. Fornr. 28, 247. Síðan leiddi konungr [Magnúss] jarl [Sveinn] til hásætis með sér. Magnúss saga ins góða 23, Heimskringla III, Ísl. Fornr. 28, 37 f. . . . sattis hedherlica a Konungx stool vidh Vpsala “The Legend of St Erik”, SRS II, p. 318. 73 Hirðskrá 5–11, NGL 2, 395–399. 74 Bøe 1964. 75 See Flateyjarbók I, 239–242; Sverris saga 16 (Cod. AM 327 40), (ed.) Indrebø, 17–19; NGL 2, 309. 76 Hirðskrá 5 says: skulu up standa byskupar oc lender men hirðstiorar oc loghmen. oc hæfia konong up i hasæti. NGL 2, 396. 77 Magnúss saga ins góða 23, Ísl. Fornr. 28, 37f. The Arenas for Royal Funeral Feasts and Inauguration Rituals 493 daughter of Skǫglar-Tósti.”78 This information is also mentioned in Fagrskinna, however, the induction into the high-seat is missing there.79 This makes Snorri’s description of the ritual somewhat uncertain. Genealogy was, on the other hand, declaimed in such contexts in other parts of medieval Europe, in for instance Celtic speaking areas. Alexander III of Scotland was crowned in 1249. “[As] he sat at the ancestral place of crowning, surrounded by his French- speaking court, a Gaelic-speaking Scot knelt before him and read the king’s genealogy back through the ninth-century Cinaed mac Alpine, and the fifth- century Fergus mac Erca, to Scota, daughter of Pharaoh”.80 According to David Dumville such customs were based on old tradition in Celtic contexts.81

13.5.1 Placing the Future Ruler in the High-Seat on the Mound In Old Norse sources referring to the pre-Christian period it is sometimes said that the high-seat was not only placed in the ceremonial hall, but also outdoors, on a thing mound or a burial mound, perhaps in order to make the king’s authority more public. The expression sitja á haugi refers to rulers who exercised their power on a grave mound or thing mound. This custom occurs especially in Norway, particularly in the areas around Trøndelag and Naumudalr.82 According to Axel Olrik, it was actually only kings who sat on the mounds: “. . . netop betragter det som et kendetegn på kongværdighed at sitja á haugi, . . .”83 Snorri Sturluson’s description of King Hrollaugr of Naumudalr in Haralds saga ins hárfagra is famous. He went up on the burial mound ( fór upp á haug) on which the kings were supposed to sit. There he had a king’s high- seat prepared for himself (ok lét þar búa konungs hásæti), and seated himself on it (see the complete text below). Some of these stories have a quite fictional image. Hákonar saga góða 12 mentions that King Eysteinn from Oppland ravaged successfully in Trøndelag. He put his dog Saur as ruler over the Þrœndir at Saurshaugr (“Saur’s mound”), i.e. modern Sakshaug on Inderøy, in Eynafylki, Trøndelag. Snorri states:

78 Magnúss saga ins góða 22, Ísl. Fornr. 28, 36. Trans. Hollander. 79 Fagrskinna, Ísl. Fornr. 29, 218f. 80 Dumville 1979, 73. 81 Dumville (1979, 73) bases his argument on Duncan 1975, 555. Duncan, in his turn, refers to Chron. Fordun (Chronica Gentis Scotorum), I, 293–295. 82 Olrik 1909; de Vries 1956–57, §246; Ringstad 1991, 142. 83 Olrik 1909, 3. 494 chapter 13

A high-seat was prepared for him, and he sat on a hill, as kings do, and lived on the Inner Island, residing in the place called Saurshaugr (“Saur’s mound”).84

Even if this story is fictional it is likely that Snorri built on older traditions when he stated that kings sat in their high-seats on the royal mound. Heroic and mythical traditions also indicate that the rulers exercised their power from a (grave) mound (ON haugr). Saxo mentions that King Høther, after being defeated by Baldr (Balderus), travelled around in Svetjud. He was now filled with sorrow and he was blamed for not doing his duties. Previously he used to give his people advice from the top of a mound (consueuerat autem in editi montis uertice consulenti populo plebiscita depromere), but now he neglected this important societal function.85 The Fornaldar sögur likewise mention that the kings used to sit on their mounds as an expression of their power and authority. In chapter 2 of Vǫlsunga saga King Rerir sat on the mound (sat á haugi) when Hrímnir’s daughter arrived at his home. Gautreks saga 8 tells how King Gautrekr mourned his dead wife and spent all his days on her grave mound (Hann sat á haugi hvern dag). According to Þrymskviða 6, Þrymr was sitting on a (burial) mound when Loki arrived in the land of giants:

Þrymr sat á haugi, þursa dróttinn, greyiom sínom gullbǫnd snøri . . .86

Þrymr must here be considered as a ruler of the giants, since he is called “the lord of the ogres” (þursa dróttinn). Some traditions also intimate that the inauguration rituals, and the acta contraria ceremonies, were related to the high-seat on the burial mound or the thing mound. A very interesting custom is mentioned, for instance, by Snorri in Haralds saga ins hárfagra, where he describes two petty kings of Naumudalr (just north of Trøndelag) in Norway. When they heard that King Haraldr was falling upon them, Herlaugr went into a mound with eleven companions and a lot of food. He ordered the mound to be closed up. The other brother acted as follows:

84 Heimskringla I, Ísl. Fornr. 26, 164. 85 Saxo, 3.3.3. 86 “Þrymr sat on a (grave) mound,/ the lord of the ogres, plaiting golden collars for his bitches; . . .” Trans. Larrington. The Arenas for Royal Funeral Feasts and Inauguration Rituals 495

King Hrollaugr went up on the mound on which the kings were wont to sit. There he had a king’s high-seat prepared for himself, and seated himself on it. Then he had down pillows laid on the footstool where it was the custom of earls to sit. Thereupon King Hrollaugr rolled himself down from the king’s high-seat and onto the earl’s seat and gave himself the title of “earl”. Then he went to meet King Haraldr and gave to him all his realm, offering to become his follower and informing him about the procedure he had taken.87

A similar story occurs in Ágrip af Nóregskonunga sǫgum 15. It is older than Haralds saga ins hárfagra, and perhaps Snorri was influenced by it when he wrote Heimskringla. Ágrip narrates that King Hersir of Naumudalir mourned his dead wife. But as he could not kill himself in his capacity of king, he had to degrade himself to the office of jarl.

Then he went up on a mound and rolled himself down from it and said then that he had rolled himself out of the king’s name, and he hanged himself in the title of jarl, . . . .88

In connection with this passage the author of Ágrip refers to the skaldic poem Háleygjatal composed by Eyvindr Finnsson (skáldaspillir) in the tenth century. The stanzas mentioning this story are unfortunately lost today.89 The

87 Hrollaugr konungr fór upp á haug þann, er konungar váru vanir at sitja á, ok lét þar búa konungs hásæti ok settisk þar í. Þá lét hann leggja dýnur á fótpallinn, þar er jarlar váru vanir at sitja. Þá veltisk Hrollaugr konungr ór konungshásætinu ok í jarlssæti ok gaf sér sjálfr jarlsnafn. Heimskringla I, Ísl. Fornr. 26, 99f. Trans. Hollander. According to several scholars, these degradation-rituals were based on an ancient tradition, as well as the opposite act to place the new king in his high-seat or on the rock. In connection to Snorri’s text Dillmann (2000c, 442) states thus: “. . . cette coutume a été expliquée, avec une grande vraisemblance, comme une manifestation du pouvoir à la fois juridique et législatif de ces rois, sur l’arrière plan du caractère sacré de royauté chez les anciens Scandinaves”. Vestergaard (1990, 122) argued in the same vein as Dillmann: “The rulers of the Viking Age were persons of royal descent who were ‘taken’ kings by popular courts and placed upon a stone or highseat . . . However, they did not have power over their seats. Kings could be removed, . . .”. 88 Ok hann fór þá á haug nekkvern ok veltisk fyrir ofan ok kvaðsk þá hafa velzk ór konungs nafni ok hengði sik síðan í jarls nafni, . . . . Ágrip, Ísl. Fornr. 29, 18. 89 In all probability these customs are ancient. However, similar judicial-ceremonial acts, known as acta contraria, are known from Roman Law, and also in Christian contexts. Holmgren 1937b, 24f.; Olivecrona 1942. 496 chapter 13

information that the king was called Hersir has been observed in research and taken as an indication of a late addition to the tradition.90 These stories about acta contraria ceremonies do not reflect historical events. But the custom whereby the king threw himself down from the high- seat may correspond with the oldest preserved medieval law in Sweden, Äldre Västgötalagen (c. 1225), where the Old Swedish term vræka ‘heave, evict, throw, reject’ occurs:

Sveær egho konong at taka ok sva vrækæ.91

Svear have the right to take and also reject the king.

The opposite act of vræka was thus “taking a king” (taka konung). Evidence from the fourteenth century shows that when the Svear, in a concrete sense, took a man as king they elevated him on the Mora stone south of Uppsala.92 This ceremony is described in Erikskrönikan (Rimkrönikan), dated to 1320–1350 but preserved only in fifteenth-century manuscripts,93 Kristoffers Landslag (1442)94 and Olaus Magnus (1555).95 A note from 1434 intimates that each royal election was documented by an inscription on a little stone that was placed upon the Mora stone.96 According to a picture in Olaus Magnus’ book the king was lifted up on to the stone. This is also attested in written documents.97 There is a house, built in 1770, which contains Mora stenar. Nine stones are preserved, two of which may be genuine documentation stones.98

90 Driscoll (2008, 92) writes thus: “It is highly unlikely that Hersir was ever the name of any particular king and there is no other record of any king bearing this name.” 91 ÄVgL R 1. SSGL 1, 36. 92 For a more detailed and critical description of this ceremony, see Sundqvist 2002, 306–333. 93 See Holmgren 1937a, p. 4. 94 KrL, Kg III, SSGL 12,14. 95 . . . lapis ingens, & rotundus . . . Morasten dictus. . . . [. . .] . . . super latum, & ingentem lapidem Rex electus stans, . . . . Historia De Gentibus Septentrionalibus I:31, 52; XI:45, 401. According to Olaus Magnus Morasten was surrounded by twelve smaller rocks: . . . est lapis camp- estris amplus, ab incolis perpetuo tempore Morasten appellatus, in circuitu XII. continens lapides paulo minori forma humi firmatos. . . . [. . .]. Historia De Gentibus Septentrionalibus VIII: 1, 243. 96 The notary public made a notation regarding the “investiture stone” of Erik of Pomerania, which was placed on the Mora rock. See Holmgren 1954, 5. 97 In the annotations to Vadstenadiariet the elevation of King Kristoffer on the Mora rock in 1441 is mentioned. Holmgren 1937b, 21. See also Libellus Magnipolensis, SRS III: 1, 196. Cf. Rosén 1939, 384f. See also Olaus Magnus above. 98 Cf. Holmgren 1954; Liedgren 1966; Lönnerholm 1993. The Arenas for Royal Funeral Feasts and Inauguration Rituals 497

The stone of Mora was probably moved from its ancient site in the second half of the fifteenth century. King Gustav Vasa had a search made for it but could not find it.99 It is not improbable, however, that the Uppsala burial mounds, and/or the Mora stone, in more ancient times, were regarded as a kind of high-seat for all Svetjud. According to several traditions the Danish kings were taken from the royal kin at the thing at Viborg. As in Sweden, the king was lifted up on to a stone at Danaerigh (Danaerygh, Danaerugh), where he was also given his royal name.100 The royal authority of Danish kings is sometimes called Hleiðrarstóll ‘the seat of Lejre’.101 A similar custom of placing the new king on a stone or a seat is found in sources referring to medieval Europe.102 To sum up, Fagrskinna’s and Snorri’s descriptions of the royal funeral feasts may include a type of royal inauguration. It seems as if this ritual action could be performed in the ceremonial hall buildings, but also outdoors at rocks, thing mounds or the burial mound of some ancestor or relative. In cases where the high-seat was placed on the burial mound or thing mound, the purpose was probably to make the ceremony more public for a wider crowd. Perhaps there was a religious argument behind this action too, especially when the high-seat was placed on a burial mound. The high-seat in itself was the link to the ancestors, who had previously sat in this seat. By placing the high-seat on the burial mound, this belief was emphasized even more, since the present ruler was in a concrete sense sitting on the (divine) forefather. When the high- seat was placed inside the hall the link to the (divine) ancestors was perhaps established by means of the gold foils depicting these mythical forebears which were placed in the holes of the high-seat posts. During the performative inauguration ritual, legitimacy was thus sought from both mythical and worldly regions.

99 Holmgren 1954; Liedgren 1966; Olsson & Berglund 1993. 100 Videntes autem Jutones Dan strennuum uirum et fortem et uirtuosum duxerunt eum ad lapidem, qui dicitur Danærigh, posueruntque eum super lapidem, inponentes ei nomen regis. . . . Chronicon Lethrense II, 45. Lecturi regem ueteres affixis humo saxis insistere suffragiaque promere consueuerant, subiectorum lapidum firmitate facti constantiam ominaturi. Quo ritu . . . rex creatus. . . . Saxo 1.2.1. huilcken vdvelgelse skeede udi Judland hos en stor Steen paa den Hede som effter den Hendelse er kaldet Danerliung. Huitfeldt I, 3. Cf. Hoffmann 1976, 170–190. 101 Olrik 1909, 8. 102 See Graber 1919, 91. 498 chapter 13

13.6 A Royal Rite of Passage

Some scholars are sceptical about these medieval descriptions and state that “there is no evidence of pre-Christian inauguration rituals in Scandinavia”.103 They feel that the descriptions mentioned in the Kings’ Sagas only refer to medieval customs. Other scholars, however, have argued that the medieval ceremony of placing the king in the high-seat was a memory of pre-Christian customs. Elisabeth Vestergaard, for instance, has suggested that “the placing of the future ruler, with the consent of the people, upon the stone or in the high-seat endowed him with regal qualities and royal power. No other ritual was necessary for granting the qualities that belonged to the seated person.”104 In my opinion, Vestergaard’s stance is quite sound. The custom of placing the future ruler on a stone or in a high-seat is most likely older than the medieval sources. However, I cannot agree with her statement that no other rituals were performed to grant the qualities that belonged to a king. Most likely there was a complex of rites and ideas associated with the inaugurations, which are not accessible for us today because of the state of the sources. Some scholars have searched for such rituals and notions in the mythical traditions. Jere Fleck, for instance, suggested that some Eddic poems, especially Rígsþula, reflect an ancient ritual education, containing genealogies, magic spells and runic knowledge, which a prince had to receive before entering upon the office of sacred kingship.105 Jens Peter Schjødt partly followed Fleck, as he stated that the king received his divine character and numinous knowledge through an inauguration, described as a rite de passage or rather an “initiation”.106 He concludes that the pre-Christians Scandinavians certainly had specific ideas about “what a future king or chieftain had to master in the field of numinous knowledge and numinous powers and something of how these could be acquired”, even if the material about consecration of kings is sparse. In my opinion, it is quite natural for royal funeral and inheritance ceremonies to be attached to inaugurations as intimated in the Ingjaldr account. Such performative rituals often have the character of rites de passage, common in cultures all around the world. Arnold van Gennep identified three successive and distinct factors in such rituals: separation, margin, and aggregation.107

103 Cf. P. H. Sawyer 2001, 385. 104 Vestergaard 1990, 121. 105 Fleck 1970. 106 Schjødt 2008, 373ff. 107 Van Gennep 1960 (1909). Cf. Turner 1967 (and 1979); 1969; Widengren 1969; Eliade 1975; Schjødt 2008. The Arenas for Royal Funeral Feasts and Inauguration Rituals 499

Victor Turner proceeded from van Gennep’s framework, but went further and focused on an aspect of these rituals that had previously been neglected, the marginal or liminal period. In initiation rituals the neophytes are—during a well-marked liminal period—removed, hidden, without rank or insignia etc. “The subject of passage ritual is, in the liminal period, structurally, if not physically, ‘invisible’ ”.108 He is “betwixt and between”, neither here nor there. The neophyte is thus regarded as symbolically dead. He is leaving his former being, but he has yet not achieved what he will become. Only after the “exhibitions”, “actions” and “instructions” (of the “knowledge”), i.e. the ritual actions that comprised the initiation, can he become what he is going to be.109 If we go back to Snorri’s account we can notice that when the future king “sit[s] on the step” in the hall, (the skǫr), the most sacred object of his father’s office and power is displayed for him, namely the high-seat. In this marginal position he can be perceived of as passing through the liminal phase of a rite de passage. He is no longer a “prince” and he has not yet become the new king or ruler. He cannot sit in his former seat on the long bench (which followed a hierarchic order),110 yet the dead ruler’s high-seat is still empty. Not until he drinks the bragafull and swears the oath can he be a righteous ruler and allowed to enter the seat. These rituals transform his identity so he can become the heir of his father’s possessions and power. Perhaps he also acquired some numinous knowledge during this ritual process, although evidence in these particular sources is lacking.

13.6.1 Toasts, Oaths, and Power Continuity The combination of making vows or promises (strengja heit) at the ritual drinking, the bragarfull, bragafull ‘oath-goblet; the ruler’s cup; cup for the most distinguished (ancestor)’,111 which is referred to in both Ynglinga saga

108 Turner 1967, 95. 109 Turner states: “By and large, this threefold classification holds good for initiation rites all over the world. Sacra may be communicated as (1) exhibition, ‘what is said’; (2) actions, ‘what is done’; and (3) instructions, ‘what is said’ . . . Among the ‘instructions’ received by neophyte may be reckoned such matters as the revelation of the real, but secularly secret, names of deities or spirits believed to preside over the rites . . .”. Turner 1967, 102f. Cf. Turner 1969 and 1979. 110 Birkeli 1932, 6f., 31–36. 111 ON bragarfull (also bragafull) “beger som ved festlige sammenkomster pleiede tømmes under afleggelse af et høitideligt løfte”. Fritzner 1954 and Simek 2006, 42. Cf. OE brego ‘ruler, chief, king, Lord’. Clark Hall 1916, 50. Perhaps the terms bragarfull and bragafull should be distinguished from each other with different meanings: “Bragafull [‘coupe de Bragi’]: leçon commune à l’ensemble des mss pour cette première occurrence du 500 chapter 13 and Fagrskinna, may thus be part of the combined erfi and inauguration ceremony. Medieval laws describe how oaths were sworn during the investitures of kings and earls in Sweden and Norway. According to Äldre Västgötalagen the Swedish king had to swear fidelity to All-Götar and promise to uphold the law.112 Hirðskrá states that Norwegian kings and earls swore oaths during the ceremony called konungstekja.113 These ceremonies could perhaps be regarded as survivals from the pre-Christian period. Libations (bragafull/bragarfull) combined with oaths (strengja heit, heit- strengingar) occur in Old Norse prose traditions. These rituals seem to be performed in banqueting halls, especially at Yule Eve (iólaaptann). The term bragr ‘the most outstanding’ indicates that this libation (bragarfull) was made in honour of the “high-seat man”, i.e. the one who was leading the ceremonial drinking in the hall building. Hervarar saga ok Heiðreks narrates that Arngrímr and his twelve sons were celebrating Yule ( jól) at the farm of Bólmr ey (Bolmsö in present-day Småland). As custom required, oaths were made at the libations (heit strengja at bragarfulli). The son of Arngrímr, Hjǫrvarðr, promised that he either would marry the daughter of King Ingjaldr of Svetjud, or remain unmarried. During the spring the brothers travelled to Uppsala and Hjǫrvarðr stepped forward to the king’s table, where he related his promise. He was then informed that another man, namely Hjálmarr, also wished to marry the daughter of the king. This led to the fall of both Hjǫrvarðr and Hjálmarr.114 Whether the customs described in this saga reflect pre-Christian rites is very uncertain. There are some similar descriptions of libations (bragarfull) combined with oaths on Yule Eve (iólaaptann) in the Eddic poem Helgakviða Hiǫrvarðzsonar.115 Similar customs also occur in Hervarar saga ok Heiðreks (redaction R).116 All the

composé, mais plus loin le ms. J1 donne la leçon bragar-full (en revanche, au chap. XIV de l’Histoire de Hakon le Bon, tous les mss donnent Bragafull). Dans la leçon bragar-full, le premier élément est le génitif sing. du mot bragr (‘le premier, le meilleur’, cf. anglo-saxon brego ‘souverain, roi’), en sorte que bragar-full doit posséder le sens de ‘coupe du person- nage le plus éminent’.” Dillmann 2000c, 412. Cf. Düwel 1985, 84–89. 112 ÄVgL R 1. SSGL 1, 36. 113 NGL 2, 395f. Cf. Heimskringla III, Ísl. Fornr. 28, 37f. 114 See Hervarar saga ok Heiðreks, ch. 2, (ed.) E. O. G. Turville-Petre, 2ff. 115 See also the prose passage before st. 31. On this ritual complex, see in particular H. Beck 1965, 177ff. 116 The R-version of the saga mentioned that King Heiðrekr’s large boar was used when making promises and oaths. According to some other manuscripts of this saga (redaction H and U), Heiðrekr worshipped Freyr (Freyja in U) and sacrificed a boar to the god during the annual feast in February, in order to procure good growth (gefa Frey til árbotar). The king claimed that this boar was so holy that it could give advice in all cases. On Yule Eve, The Arenas for Royal Funeral Feasts and Inauguration Rituals 501 texts are late and hard to evaluate as sources for pre-Christian religion. They may nevertheless show that the information about the rituals at Ǫnundr’s erfi is not totally isolated in the Old Norse text traditions. In Ynglinga saga King Ingjaldr made the vow (strengði . . . heit) that he would increase his dominion (auka ríki) to twice its size in every direction, or else die.117 Then he emptied the cup. This ritual is directly connected with his new office and the fortune of his forthcoming reign. Fagrskinna says that Sveinn vowed that before three winters were gone he would travel to England and either kill King Aðalráðr or drive him out of his land, while Earl Sigvaldr of Jómsborg vowed that he would kill Earl Hákon in Norway.118 This story, too, features vows associated with the deeds to be accomplished in the forthcoming reign.119 It is probable that the connection of vows and memorial drinking to dead ancestors and gods referred to succession and the continuity of power and luck. The anxious transition from one reign to another was probably a crucial phase of the political process and it had to be resolved in the most appropriate way. The toasts made to the family’s mythical ancestor(s) might be regarded as expressions of the specific links which existed between the ruler, the god(s) and the dead forefathers. Some scholars would probably consider them to be an “individual consecration” (Individualweihe) to a certain god or gods.120 It has previously been argued by scholars that ancient Scandinavian inauguration rituals consisted of a ritual education, which besides royal genealogy also consisted of runic knowledge and magic spells (see above). It is thus possible that some types of “esoteric knowledge”, such as information about genealogy, including divine ancestors, played some role at this moment of the inauguration. 121 As noted above, genealogies, burial mounds and high-seats were important at inheritance processes in general. The genealogy could have been part of the “exhibitions”, “actions” and “instructions” which often occurred in the liminal phase in general of initiation rituals.122 Remains of such genealogies may be

the boar had to be led to the king, so that men could put their hands on his bristles and make promises (logdu menn þá hendur yfir bust honum ok streingdu heit). See the text in Hervarar saga ok Heiðreks, (ed.) Helgason, 129 and the comments in Hervarar saga ok Heiðreks, (ed.) E. O. G. Turville-Petre, 78. See also H. Beck 1965, 177. 117 Heimskringla I, Ísl. Fornr. 26, 67. 118 Fagrskinna, Ísl. Fornr. 29, 125. 119 Cf. Lincoln (2014) who also brings in the medieval perspective of these rituals as they were presented in Heimskringla. See above. 120 Cf. Höfler 1952. Grønvik 1983. 121 Cf. Schjødt 2008. 122 Gro Steinsland states in a recent article: “Kinship was probably more important for the elite than for other classes of population—note that for example Germanic titles of rulers 502 chapter 13 visible in Ynglingatal and Háleygjatal (see ch. 3) as well as in the accounts on the early Christian inaugurations mentioned above.

The erfi ceremony described in Ynglinga saga and Fagrskinna concerned both the heir and the dead ruler. After making toasts to the dead ruler, the forefathers and the gods and the swearing of oaths and holy vows, the heir entered the high-seat, situated either in the hall or on the mound. The memory toasts and vows in these texts could be seen as constructions of power continuity and legitimate succession. Perhaps the recitation of genealogies also took place during these performative rituals. In my opinion, it is not impossible that the rituals described in these medieval sources reflect ceremonies and ideas relevant for the Viking Age societies of Denmark, Svetjud and Norway. The high-seat played a central part in this ritual system; it was the link between the rulers and their reigns. It was also regarded as the ritual entrance to both rulership and the Other World. By means of this performative ritual the prince was transformed into a ruler in public. When sitting in the high-seat during his life, he executed his functions as a leader in a judicially legitimate sense. Entering the high-seat also comprised a complex of religious ideas. The ruler had, for instance, the ability to make contact with the gods and divine ancestors from there, which was one of the reasons why he had to be in his high-seat during the sacrifices at the ceremonial feasts. The performative rituals associated with the high-seat at these feasts could also be seen as ritual strategies for gaining legitimacy and constructing authority and royal power. Whether such strategies also occurred among chieftains in Iceland is uncertain.

are derived from kin-groups. The importance of kin for the elite can also be seen in the many genealogies: on the level of kings and earls ancestors were counted back to the gods and other non-human powers.” Steinsland 2011a, 6. Chapter 14 (Inter-)Regional Chief Sanctuaries as a Religious Strategy for Rulership

As we saw in chapter 5, there were sanctuaries on different levels in Svetjud and Norway. Beside the local public sanctuaries there were also regional or inter-regional sanctuaries in these areas, which were important for all Svear and Þrœndir. Examples of such chief sanctuaries could be seen in Uppsala, Lade and Mære. Most likely they were controlled by the kings, earls or a cooperative of chieftains and farmers (as in the case of Mære, which may have had a more communal character). Such (inter-)regional sanctuaries are not attested in the Icelandic context. In what follows I will discuss the cultic organization of the inter-regional and regional sanctuaries in Svetjud and Trøndelag. I will suggest that they were not only used for religious purposes, but also for political reasons. During the Viking Age pagan kings and earls endeavoured to be the supreme cultic leaders of their realms. In my opinion the (inter‑)regional sanctuaries could therefore be seen as a ruler strategy. By practising a cult of many gods in one and the same sanctuary, the kings and the earls managed to gather many individuals and cultic groups, for instance in Uppsala and Lade. In the wide circle of gods at the sanctuary, the close relation between the ruler and the mythical world was emphasized.

14.1 Cultic Organization

14.1.1 Trøndelag and Svetjud Previous research has discussed the pre-Christian cultic organization in ancient Scandinavia. Based on cultic place-names, Magnus Olsen suggested that religious activities in Norway (including Trøndelag) appeared on different social levels.1 On the lowest level we may find a private cult, which took place at individual farms. This type of cultic site was often designated by “marknavn” (e.g. hǫrgr), and the cult there rarely had a public character. On a higher level there was a public cult which took place at sanctuaries designated with names such Hof (-hof) and Vangr. These cultic sites were connected to a type of pagan cultic parishes (“hedenske kultsogn”), which resembled the medieval

1 M. Olsen 1926, 257ff.

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_015 504 Chapter 14 church parishes. There was thus a public sanctuary centrally located in every “pagan parish”, where people regularly gathered to celebrate pagan feasts and perform sacrifices. There was also a type of chief sanctuaries above this level, which concerned many people living in a wider area, such as the fylki ‘shire’. At these places not only cultic activities took place; administrative-political matters were also dealt with there. In Trøndelag these cultic-political central places were identical to the sites where the shire churches (fylkiskirkjur) were later erected (see ch. 12). Olsen’s theory about the pre-Christian religious organization in Norway and Trøndelag is quite obviously influenced by the medieval and clerical cultic organization. The terminology, for instance “hedenske kultsogn”, bears a clear Christian impression.2 In a more general sense, however, Olsen’s theory is quite sound. Most likely there was both a private and public cult in pre-Christian Norway and Trøndelag.3 The public cult was probably also hierarchical in some sense; there were sanctuaries for more local settlement districts and there were chief shrines which were important for an entire shire (fylki). The “fylki sanctuaries” were situated at the central farms (No sentralgårdar), while some of the “local sanctuaries” appeared at the hof sites, particularly in Ut-Trøndelag.4 Most likely there were also regional sanctuaries, which were important for all people living in Trøndelag, e.g. Mære. Some public central sanctuaries may have even been of great significance for all of Norway, for instance Lade. Specialists in place-names, such as Lars Hellberg, have argued that Svetjud also had a similar hierarchical cultic organization to that in Norway.5 On the lowest level there were local sanctuaries, which were important either for a district equal to the medieval “hundred” (Sw hund, hundare, skeppslag) or for the “parish” (Sw socken).6 There were also regional or inter-regional sanctuaries, which were important for all the Svear, such as Old Uppsala. As noted above, few scholars accept today that the Mälaren area had prehistoric districts equal to medieval parishes.7 Hellberg’s suggested that pre-Christian cultic organization nevertheless harmonizes quite well with Stefan Brink’s more recent proposal of spatial or horizontal division of society in prehistoric Svetjud. People lived in settlement districts (OSw byghþ sg.) of various sizes.

2 On the criticism of “pagan parishes”, see Brink 1990a; 1996a. 3 See e.g. Ström 1985, 76–93; Steinsland 2005, 265, 274ff., 327ff. See however the discussion of the concepts “private” and “public” in ch. 4 above. 4 Røskaft 2003, 53–76, 210. 5 See e.g. Hellberg 1986a, 66ff. Cf. Ström 1985, 76–93; Strid 1999, 95. 6 Hellberg 1986a, 66. 7 See e.g. Brink 1990a; Strid 1999, 92ff. (Inter-)Regional Chief Sanctuaries 505

These local districts were loosely linked to some larger regional formation called “land” (OSw land) (e.g. Uppland, Västmanland and Södermanland)8 and perhaps an inter-regional area called “realm” (OSw rike), including all Svetjud (Svíþjóð, Svíaríki or Svíaveldi).9 Brink writes: “For each and every one of these settlement districts, as well as all the social formations on the regional and inter-regional levels, there were one or more focal sites that may be characterized as central places (of different kinds).”10 At some of these central places public cultic activities also took place. Per Vikstrand has made important contributions to the discussion of pre- Christian cultic organization. He argued that the sacral place-names in the Mälaren region represent three different types of locations: (1) “a normal Iron Age settlement”; (2) “the centre of the community type” and (3) “the aristocratic centre”.11 For the present study it is important to make a distinction between “the cultic centre of communal type” and “the aristocratic cultic centre”. In my opinion, both types were controlled by rulers. We may conclude that cultic organization in both Norway and Svetjud was hierarchical in some sense. Most likely there were both local and regional (or inter-regional) sanctuaries in these societies. Some of them could be described as “aristocratic” and others as “communal” sacred sites. At both types, rulers on different social levels were in charge and controlled the ritual actions which took place there.

14.1.2 The Organization of Sanctuaries in Iceland The cultic organization in pre-Christian Iceland has also been discussed.12 This organization is sometimes described in the Old Norse medieval sources. According to the Hauksbók version of Úlfljótslǫg, the land was divided into quarters (fjórðungar), and it was decided that there should be three assemblies (þing) in each quarter and three major hof sanctuaries (hǫfuðhof ) in each assembly commune, i.e. assembly district or community (þingsóknir) (see ch. 5 above). Úlfljótslǫg also states that a chieftain (goði) was appointed to each major hof sanctuary (hǫfuðhof). If other sources are taken into consideration, such as Ari’s Íslendingabók, however, it seems as if there were four assembly

8 Also smaller lands such as Arland, Soland and *Valand in Uppland may have a similar background. 9 Cf. Brink 1997, 404–406; T. Andersson 2004. 10 Cf. Brink 1997, 405. 11 Vikstrand 2001, 410–417. 12 See e.g. V. Finsen 1888; Taranger 1890; O. Olsen 1966; Jón Viðar Sigurðsson 1999. 506 Chapter 14 districts in the northern quarter, but three in each of the others.13 So perhaps there were twelve public hof sanctuaries in the northern quarter. It has been argued that there must therefore have been thirty-nine public sanctuaries with goðar in Iceland after 965, besides the smaller private ones, or sanctuaries of ease.14 But when the General Assembly was established in 930, there may have been just thirty-six such sanctuaries, before the land was divided into quarters, as Úlfljótslǫg intimates.15 Olaf Olsen was sceptical of the description of this cultic organization in Úlfljótslǫg. Most of all he was critical of the concept of hǫfuðhof.16 This term is not mentioned in Ari’s text, and appears first in the late saga literature.17 He also argued that we have many places with names including the term hof in Iceland. Totally there are forty-four (?) known hof-names in Iceland, and in Norway the number is even higher. There is, however, no single name including the compound hǫfuðhof, neither in Iceland nor in Norway. Olsen says that the term hǫfuðhof may be a literary construction by the medieval scribes, inspired by the Christian term hǫfuðkirkja. He argued that the thirteenth-century scribes were acquainted with the fact that there were many more hof places in Iceland than goðar. Therefore they had to explain this situation somehow. The idea that the goðar only were related to hǫfuðhof was one way for them to solve this problem smoothly. In my opinion, the term hǫfuðhof may very well be a late construction inspired by Christian terminology. However, the phenomenon itself, where the chieftains (goðar) were related to certain public hof sanctuaries, may be old. Whether there were thirty-six or thirty-nine of these local “public sanctuaries”, or any other number in pre-Christian time, is hard to ascertain. But we may conclude that they were quite numerous, because of the evidence from toponymy of hof names. Most likely they were situated at the chieftains’ farms and spread out in the settlement districts over all Iceland, in all quarters, as we may see on the distribution map of the hof settlements. How systematic the organization was in the pre-Christian period and if there was a hierarchy between different sanctuaries is impossible to say.

13 See Íslendingabók 5, Ísl. Fornr. 1, 12. 14 See e.g. E. O. G. Turville-Petre 1964, 243. 15 On the number of chieftaincies in Early Iceland, see e.g. Jón Viðar Sigurðsson’s (1999) critical survey. 16 Olsen 1966, 40ff. Cf. Taranger (1890, 263) who also believed that hǫfuðhof derived from hǫfuðkirkja. 17 It is also stated in the introduction to Þingskapa þáttr in Grágás that hǫfuðhof were based on ancient goðorð. See Grágás I, 1, 38. Cf. V. Finsen 1888, 72; Olsen 1966, 41. (Inter-)Regional Chief Sanctuaries 507

The sources clearly indicate, however, that public cultic activities regularly took place at some of these hof sanctuaries. The archaeological evidence from the Hofstaðir settlement at Mývatnssveit, for instance, indicates that ritual slaughter on a seasonal basis took place at this hof farm.18 Several finds and the size of the hall building indicate also that not only people living on the farm took part in these feasts, but also people from settlement districts around this place. We may assume that such public activities must have been organized and performed by some local authority, perhaps a chieftain, who also took care of the ceremonial building at Hofstaðir. Whether this chieftain carried the title goði is hard to say, although we cannot rule out this possibility. In any case, the cult at Hofstaðir was local and restricted mostly to the people living in the settlement districts of the Mývatn area. Public cult thus most often took place at certain local hof sites situated at chieftain farms in central parts of the settlement districts around Iceland. This cult was organized by the local chieftains, perhaps called goðar. Whether all Icelanders had a common regional or inter-regional cult at a chief sanctuary, comparable with Uppsala for the Svear or Mære and Lade for the Þrœndir, is more uncertain.19 The decentralized political structure of Icelandic society probably never gave any favourable conditions for such chief sanctuaries. Admittedly, there was the title allsherjargoði, ‘goði for the whole assembly’ or ‘goði for the althing’ mentioned in Landnámabók,20 which may indicate a central religious institution for all Iceland. Based on information in the Þórðarbók redaction of Landnámabók,21 some scholars have argued that the allsherjar- goði was supposed to consecrate the Alþingi at Þingvellir.22 Whether other cultic activities were performed during this communal assembly is uncertain. There is actually nothing in the source materials to indicate that Þingvellir was an important pre-Christian cultic place, nor is there any indication of other regional or inter-regional public cultic sites in Iceland.

18 See mainly Lucas & McGovern 2008. 19 Cf. Lid 1942, 82. 20 Son hans var Þormóðr, er þá var allsherjargoði, er kristni kom á Ísland. Landnámabók, SH 9 H10, Ísl. Fornr. 1, 46f. 21 Þorsteinn Ingólfsson lét setja fyrstr manna þing á Kjalarnesi, áðr alþingi var sett, við ráð Helga bjólu ok Ørlygs at Esjubergi ok annarra viturra manna, ok fylgir þar enn sǫkum (þess), því goðorði alþingishelgun. Þorkell máni lǫgsǫgumaðr var son Þorsteins Ingólfssonar. Þormóðr var son Þorkels, er þá var allsherjargoði er kristni kom á Ísland. Þorgeirr Ljósvetningagoði var þá lǫgsǫgumaðr. See Þórðarbók redaction of Landnámabók, Ísl. Fornr. 1, 46f., note 3. 22 See Lid 1942, 82; Jakob Benediktsson in Ísl. Fornr. 1, 46f., note 3; ONP I, 327. 508 Chapter 14

It thus seems as if the pre-Christian cultic organization in Svetjud and Trøndelag was based on a spatial hierarchical division, with public sanctuaries of varying local, regional or inter-regional significance. In contrast, the cultic organization of Iceland lacked this division. It was based on local sanctuaries at chieftains’ farms in the settlement districts. It also seems as if sanctuaries in Svetjud and Trøndelag could be controlled by individual rulers at the aristocratic centres. Some of them, however, might have had a more communal character. Whether this type of communal sanctuaries existed in Iceland is uncertain. The hof sanctuaries there were usually situated at chieftains’ farms.23 We do not know whether the communal assembly place at Þingvellir was regarded as a cultic site.

14.2 The Chief Sanctuaries of Trøndelag and Svetjud

Several scholars have argued that regional or inter-regional sacrificial feasts were held in Scandinavia.24 It seems as if these feasts took place at certain chief sanctuaries. These central sanctuaries of (inter-)regional significance appear at important places in Denmark, Svetjud and Norway. According to Jan de Vries, such “Hauptheiligtümer” could be found at Lejre (Hleiðr) on Zealand, Odense on Funen, Viborg in Jutland, Lund in Scania, Uppsala in the Mälaren area, Lade in Trøndelag, and Gaular for the people living at Firðir, Fjalir and Sogn. Skíringssalr was the central sanctuary for the people on the Oslofjord.25 At this site common sacrificial feasts were celebrated. There is some interesting information about common religious feasts in Guta saga, including the highest sacrifice for the whole island (land alt) of Gotland:

The whole island held the highest sacrifice on its own account, with human victims, otherwise each third held its own. But smaller assemblies held a lesser sacrifice with cattle, food, and drink. Those involved were called ‘boiling-companions’, because they all cooked their sacrificial meals together.26

23 Cf. O. Olsen 1966. 24 See e.g. de Vries 1956–57, §§327ff.; F. Ström 1985, 76ff. 25 de Vries 1956–57, §327. Cf. Phillpotts 1912–1913. 26 Land alt hafþi sir hoystu blotan miþ fulki. Ellar hafþi huer þriþiungr sir. En smeri þing hafþu mindri blotan miþ fileþi, mati ok mungati, sum haita suþnautar, þy et þair suþu allir saman. Peel 1999, 4f. (Inter-)Regional Chief Sanctuaries 509

The text does not tell where these “highest sacrifices” took place, however. Perhaps they were held at the communal assembly place of Roma, where the medieval general þing was held for all Gutar. The name Roma simply means ‘open place, field’.27 But the pre-Christian cultic site there may have been designated Gutnal (Gutna + al), which has been interpreted as ‘the sanctuary of the Gutar’.28 This term appear in Guta saga, as first element in gutnalþing ‘the Gotlanders’ general assembly’.29 In one Low German translation of Gutalagen (manuscript dated to 1401) it is stated that this assembly took place at Roma (czu gutnaldhing das is czu Rume).30 It is not impossible, however, that the highest sacrifices of the Gotlanders took place in Visby, i.e. the medieval religious centre. The first element of this compound name, Vis-, may be interpreted as the genitive of an appellative vi ‘pre-Christian sanctuary’.31 The Guta saga states: “Then, later there was a sacrifice at Vi.”32 This place was situated “below the cliffs” (firir niþan klintu) and is usually identified with present-day Visby or a site nearby.33 It seems as if the organization of the public sanctuaries and the question of cultic leadership in Svetjud and Norway varied. According to Old Norse narrative sources, the political leader, the king (konungr), the earl (jarl) or even the chieftain (hǫfðingi, hersir, goði) could appear in religious roles there and take care of the sanctuaries (see ch. 5 above).34 The uppermost elite, however, were probably related to some specific sanctuaries which had a regional or an inter- regional significance for the Svear and the Þrœndir, for instance Uppsala and Lade. Yet there was also one “regional sanctuary” in Trøndelag, which seems to have been a “cultic centre of the community type”, namely Mære. The cult there seems to have been organized by a cooperative of Þrœndir chieftains at least during the second half of the tenth century. In what follows I will discuss

27 I. Olsson 1996, 58; Peel 1999, 29. 28 Cf. I. Olsson 1996, 137. The interpretation of the element *al as ‘sanctuary’ is critically considered by Brink 1992; Vikstrand 2001, 192–196; and Elmevik 2004. The common interpretation of Gutnal as ‘the sanctuary of the Gutar’ is critically examined by Melefors 2014 (see also below). He states that it rather consists of a Gutn- ‘utflöde’ and -al ‘sänka i ter- rägnen’. For a discussion of where the old assembly place was located at Roma, see Myrberg 2009. For an intresting discussion on the semantics of Goth alhs, see Jackson 2014a. 29 Peel 1999, 6f. The form gutnalþing is problematic because of the -n-. The form guta alþing is expected. See Peel 1999, 33. 30 This text is quoted in Peel 1999, 34. 31 I. Olsson 1996, 23ff. 32 Þa eptir þan tima var blotan i Vi. (Ed.) Peel 1999, 8f. 33 Olsson 1996, 23ff.; Peel 1999, 43ff. 34 Phillpotts 1912–1913, 267. 510 Chapter 14 the organization of cult and cultic leadership at the inter-regional and regional sanctuaries of Uppsala, Lade and Mære.

14.2.1 The Organization of Cult at the Regional and Communal Public Sanctuary of Mære As noted in chapter 5 above, Old Norse traditions indicate that the chieftains of tenth-century Trøndelag collaborated concerning religious matters, especially when they were threatened by Christian royal power. Perhaps this cultic organization was initiated when King Haraldr and the Earls from Hálogaland started to have interests in Trøndelag at the end of the ninth century. It seems, for instance, as if the chieftains together were involved in the common cultic feasts during the tenth century, which were celebrated at the central sanctuary in Inn-Trøndelag, i.e. the one at Mære.35 This shrine seems to have been considered a communal “chief sanctuary” for the Þrœndir living in the shires (fylki). That Mære was important for more than one shire is indicated in several ways. It was, for instance, located on the border between two shires, at a neutral place, and at some distance from the farm of the fylki chieftain.36 Pre-Christian Mære diverges from other cultic sites in Trøndelag, which were usually located at central farms. Compared to them, Mære lacks the social and economic criteria typical of the central farms and it stands out mostly as an exclusively religious site.37 The gold foil figures found at the assumed cultic building under the church indicate that this place in the Merovingian Period and Early Viking Age was related to a family with great ambitions in Inn-Trøndelag. Most likely the leaders who cared for the cult in Mære during that period were situated at a farm very close to the cultic site. Perhaps this ruling family resided at Egge, a site with a burial field with rich finds from Iron Age.38 In the late tenth century, however, it seems as if the cult in Mære was a common affair for all chieftains of both Ut- and Inn-Trøndelag.39 This idea is indicated in a passage by Snorri Sturluson. In Óláfs saga Tryggvasonar 65–72 Snorri describes the missions of King Óláfr Tryggvason in Trøndelag.40 During

35 During Þórhaddr’s days (ninth century) Mære was probably a local sanctuary controlled and managed by the local hofgoði. 36 Stenvik 1996, 88. 37 See e.g. Stenvik 1996; Lidén 1999. These scholars do not think that Mære was a chieftain farm (sentralgård) before the Middle Ages. Probably there was some kind of settlement there also during the Viking Age. See Røskaft 2003, 143f. 38 Stenvik 1996; Røskaft 1997, 237; 2003, 138f.; Lidén 1999, 45. 39 Cf. Phillpotts 1912–1913, 274. 40 Heimskringla I, Ísl. Fornr. 26, 315–318. (Inter-)Regional Chief Sanctuaries 511 his stay the king said that he wished to journey to the place where the Þrœndir celebrated their greatest sacrifice, and there see their customary faith (it mesta blót yðvart, ok sjá þar siðu yðra). He was informed that he then should attend the midsummer sacrifice (miðsumarsblót) at Mære. When King Óláfr Tryggvason came to the inner reaches of the Trondheimsfjord and to Mære at midsummer, all the chieftains who had opposed Christianity had arrived with the rich farmers. At the assembly Járn-Skeggi (Skeggi) Ásbjarnarson of Upphaugr in Yrjar (today called Ørland located in Ut-Trøndelag) was the leader of the Þrœndir. He said: “It is our wish, sir king, that you make the sacrifice as other kings have done here before you.” The king replied by killing Járn-Skeggi and desecrating the hof building and cultic images there. It is peculiar that Járn-Skeggi of Yrjar in Ut-Trøndelag represented the Þrœndir at the hof building of Mære, and not a local chieftain. It seems as if Snorri believed that the chieftains together managed the communal chief sanctuary of Mære and together organized cult in Trøndelag. This story could of course be seen as a projection of the thirteenth-century conditions back to events taking place around the year 1000, and I agree that it must be partly a literary construction. It is quite obvious, for instance, that Snorri’s descriptions of King Óláfr’s missions follows a recurrent pattern; first the king meets the farmers at an assembly, where he tells them about Christ, then he asks them to convert or else they must fight the king. In these stories the king is always successful. It is a matter of fact, however, that Snorri had good sources when he described the missions of King Óláfr Tryggvason.41 His most important source for this text was probably Saga Óláfs Tryggvasonar, written in Latin by the twelfth- century Icelandic monk Oddr Snorrason at the monastery in Þingey. The Latin version is lost, but an Old Norse translation of it is preserved. This text is full of miracles and wonder tales, but there are also some passages which have a more historical character. One such passage is Oddr’s statement that the major sacrifice of the Þrœndir was held at Mære (ok á Mærini í Þrándheimi var hǫfuðblót).42 Another historical aspect in Oddr’s narrative may be his description of the chieftains, who together organized and protected the pagan cult in Trøndelag and at Mære.43 In order to eradicate the sacrificial cult there, the Christian king had to kill a representative of these pagan chieftains and desecrate the hof sanctuary. As in Snorri’s version, Oddr mentions that it was Járn-Skeggi of Yrjar

41 Steinsland 2000, 113ff. 42 Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 279. This story is also told in Flateyjarbók I, 320. See also Meulengracht Sørensen 1991b, 239ff.; Steinsland 2000, 114ff. 43 See Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 278ff. 512 Chapter 14 who was killed beside the sanctuary.44 Ágrip and Fagrskinna also vaguely indicate that the cult at Mære during the tenth century was common to all Þrœndir, and perhaps organized by the chieftains together. In the story of King Hákon the Good’s conflict with the Þrœndir at Mære, as rendered by Fagrskinna, sacrificing men (blótmenn) participated on the farmers’ side.45 It seems as if these men were identical with the Þrœndir chieftains. There is actually one source which explicitly states that the chieftains of Trøndelag together organized the cultic feasts at Mære. Óláfs saga Helga 109 reports that a man called Þóraldi was King Óláfr’s steward (ármaðr) on his estate (konungs bú) at Haugr in Verdælir sometime at the beginning of the eleventh century. King Óláfr asked him if it was true that the Þrœndir in Inn- Trøndelag still performed heathen sacrifices. Þóraldi answered him that nearly all the people in the interior of the Trondheim district were pure heathens in their belief, and that only a few were baptized:

It is their custom to perform a sacrifice in the fall to welcome winter, a second at midwinter, and a third in summer to welcome its arrival. In this, the people of Eynir and those of Sparbyggvar, of Verdælir, and of Skeynir participate. There are twelve men who take it upon themselves to arrange the sacrificial feasts; and this spring it falls to Ǫlvir to make ready the feast. He is now busily engaged upon that at Mære, and all provisions necessary for that feast are brought there.46

The chieftain Ǫlvir came from Egge, in Sparabú, close to Mære. That he took care of the sacrificial feast is not sensational. In any case, we may conclude from this quotation that in the twelfth- and thirteenth-century saga traditions there was some information among Icelandic authors about chieftains (“twelve men”) living in Trøndelag around 1000, who together arranged the common cultic feasts that were held at the sanctuary of Mære. According to them this sanctuary of Mære had a regional significance. Since the gap between some of these texts and the actual events is not more than 150–200 years, these

44 . . . var Járn-Skeggi drepinn hjá hofinu . . . Óláfs saga Tryggvasonar eptir Odd munk Snorrason, Ísl. Fornr. 25, 280. 45 Fagrskinna, Ísl. Fornr. 29, 80. 46 “. . . En þat er siðr þeira at hafa blót á haust ok fagna þá vetri, annat at miðjum vetri, en it þriðja at sumri, þá fagna þeir sumri. Eru at þessu ráði Eynir ok Sparbyggvar, Verdælir, Skeynir. Tólf eru þeir, er fyrir beitask um blótveizlurnar, ok á nú Ǫlvir í vár at halda upp veizlunni. Er hann nú í starfi miklu á Mærini, ok þangat eru til flutt ǫll fǫng, þau er til þarf at hafa veizlunnar.” Heimskringla II, Ísl. Fornr. 27, 180. (Inter-)Regional Chief Sanctuaries 513

traditions must be regarded as important sources for the cultic context in Trøndelag around the year 1000. Perhaps information about this was orally transmitted during this period to Snorri and the medieval saga authors.

14.2.2 The Earls’ Sanctuary at Lade The written sources report that the Hlaðajarlar had an important sanctuary erected at the residence and farm called Lade. This farm was located at the mouth of the river Nið, in the region of Ut-Trøndelag, i.e. the area which was early in coming under the control of the earls. According to Snorri Sturluson, this farm was originally in King Haraldr’s possession and it was his chief estate in Trøndelag (see ch. 2 above). It seems, however, as if King Haraldr granted this farm to the Hlaðajarlar who originally came from Hálogaland. Old Norse traditions report that the earls had a hof on their own farm at Lade. Some of these texts mention that this hof was one of the greatest sanctuaries in all Norway. Njáls saga 87, for instance, says that Earl Hákon owned a hof in Guðbrandsdalr, south of Trøndelag, together with the hersir Guðbrandr of the Dale. In this passage we read that the biggest hof in Norway was located at Lade.47 One of the manuscripts of Oddr munk’s Óláfs saga Tryggvasonar (A = AM 310 4to) produces a similar image. It records that Earl Hákon had his main residence at Lade. There he had a great hof sanctuary which was dedicated to numerous gods.48 This was a chief sanctuary, which was the greatest in all Norway.49 Snorri Sturluson likewise writes in several passages that Earl Hákon had a hof in Lade.50 He also related this hof to Hákon’s father Earl Sigurðr Hlaðajarl. In Hákonar saga góða 14, for instance, Snorri states that Sigurðr maintained the cult in Trøndelag and that he also organized and invited generously to sacrificial feasts at Lade. When King Hákon the Good visited Trøndelag the farmers expected that he should also perform in ritual roles at the public sacrifices there (see the text in ch. 4 above). It seems as if Snorri identified the king of Norway as the supreme cultic leader and the ultimate protector of the sacrificial cult at the sanctuaries, also in Trøndelag. It is

47 Hann [Guðbrandr í Dala] var inn mesti vin Hákonar jarls; þeir áttu hof báðir saman, ok var því aldri upp lokit, nema þá er jarl kom þangat; þat var annat mest hof í Nóregi, en annat á Hlǫðum. Brennu-Njáls saga 87, Ísl. Fornr. 12, 210. 48 Ok þat hit mikla hof er þar stóð hafði hann eignat ótalligum guðum. Ísl. Fornr. 25, 193. 49 Manuscript A (AM 310 4to) states thus: Þar var ok hǫfuðhof þat er mest var í Nóregi. Also in manuscript S (Stockh. perg 4to nr. 18) it is stated that the sanctuary in Lade was great: þar stóð mikit hof. Ísl. Fornr. 25, 193. 50 See e.g. Óláfs saga Tryggvasonar 59–60, in Heimskringla I, Ísl. Fornr. 26, 308ff. 514 Chapter 14 interesting to note that in Snorri’s text about the cult in Lade no “cultic leaders/ specialists” are mentioned other than the earl and the king. Snorri supports his statements about Earl Sigurðr’s generous sacrificial feasts at Lade with an interesting stanza from Kormákr Ǫgmundarson’s Sigurðardrápa (AD 960). Judging by this stanza, religious feasts were held at a sanctuary (vé), which was superintended by Earl Sigurðr. The earl is here called vés valdr, i.e. “the keeper, owner, ruler, or custodian of the vé sanctuary.” The Old Norse word valdr (valdi) refers to ‘ruler, owner’ (‘som har magten (over noget), styrer, besidder),51 and vé to ‘sanctuary’. Most likely this sacred site was located at Lade, as Snorri assumed. We may thus conclude that the Earls from Hálogaland owned a great and famous hof sanctuary at Lade. This sanctuary, at least during Earl Hákon’s reign, probably had an inter-regional significance. Together with Skíringssalr in Vestfold it was one of Norway’s most important cultic sites. At Lade prominent guests from all Norway as well as chieftains and farmers from Trøndelag celebrated religious feasts together, under the surveillance but also the hospitality of the earls.

14.2.3 The Svea King and the Inter-Regional Sanctuary at Uppsala The most important sanctuary of the Svear seems to have been located in Uppsala. Adam of Bremen’s text describes how the Svear celebrated a common cultic feast there every ninth year; no one was granted exemption from that feast. He states: “It is customary also to solemnize in Uppsala, at nine-year intervals, a general feast of all the provinces of the Svear. From attendance at this festival no one is exempted. Kings and peoples all and singly send their gifts to Uppsala . . .” (see ch. 4 and 5 above). Snorri similarly writes in Óláfs saga helga 77 that the main sacrificial feast of the Svear was held at Uppsala. Sacrifices were to be made at that time for peace and victory for the king, and people from all Svíaveldi were to resort there. The written sources thus clearly indicate that Uppsala was the most important inter-regional sanctuary in Viking Age Svetjud According to Adam, there were priests (sacerdotes) appointed at the “temple of Uppsala”, who offered the people’s sacrifices at the calendrical feast.52 In his description it seems as if the priests were some kind of permanently employed cultic functionaries: “One of the priests who used to serve the

51 Finnur Jónsson 1931, 589. 52 Omnibus itaque diis suis attributos habent sacerdotes, qui sacrificia populi offerant. Adam IV,27. (Inter-)Regional Chief Sanctuaries 515 demons at Uppsala turned blind without gaining help from any deities.”53 Several researchers have been sceptical about Adam’s information on that point.54 The term sacerdotes evokes the differentiated priesthood of the Roman Catholic church. Adam may be indulging in an interpretatio romana or Christiana, in that he interpreted Scandinavian culture and religion through his Roman-Christian glasses. A centralized differentiated priesthood with religious specialists who were exclusively devoted to their cult probably never existed in pre-Christian Svetjud.55 On the other hand, there may have been some kind of “religious specialists” at the calendrical religious feasts, who were responsible for specific cultic functions at some sanctuaries. The rituals performed at the public cultic feasts in Uppsala, for instance, were probably very comprehensive and complicated (see ch. 5 above). Perhaps some “specialists” were needed for different aspects and separate rites during the public sacrifices there (see ch. 5 and 9 above). As noted above, there are terms and designations preserved in Swedish place-names, carrying meanings supporting this notion, for instance OSw *gudhi, *lytir, and *vivil. Whether these “specialists” were permanently employed “priests” at the Uppsala sanctuary is uncertain. Most likely they had many functions similar to the bailiffs’ and the king’s representatives’ assignments at other royal farms and central places in Svetjud. They took care of the royal property on behalf of the king. Since they were representatives of the leading stratum of society, they also temporarily played central roles at the cultic feasts. As compensation they were protected by the king. The relationship between the kings and “the religious specialists”/ “cultic leaders” at Uppsala could perhaps be described as a kind of patron-and-client-system, i.e. the reciprocal order admitting both clients and patrons to gain benefit from the other’s support. In my opinion, it was the king who had the overarching responsibility for the public sacrificial feasts there. It was also he who performed certain important rituals at these gatherings. It actually seems as if Adam of Bremen was acquainted with the idea that it was the king who officiated at the sacrifices in the common cult in Uppsala. Scholion 140 says that King Anunder was driven from the public feast and assembly (concilium) at Uppsala, since he would not perform as a cult leader at the sacrifices (see ch. 9). This image also agrees with the information given in Old Norse literature. According to the U-version of Hervarar saga (c. 1300), for instance, the Svear removed King Ingi Steinkelsson (1080–1110) from office

53 Quidam e sacerdotibus, qui ad Ubsolam demonibus astare solebat nequiequam iuvantibus diis factus est cecus. Adam IV,28. 54 See e.g. Hultgård 1997, 19−20 and Sundqvist 2003a; 2003b; 2007. 55 See Sundqvist 2007. 516 Chapter 14 since he would not make sacrifices on their behalf at the assembly which probably was held at Uppsala.56 Like Anunder, Ingi had converted to the Christian faith. These sources indicate that the king, according to the tradition, had connections to the cult in the Uppsala sanctuary. However, since some of these kings had converted to Christianity they did not take part in the sacrifices there. During the eleventh century the Christian Svea kings appeared at other places, such as Sigtuna. Perhaps some of the pagan bailiffs and old royal representatives took care of the Uppsala sanctuary as deputies for the absent king during the final phase of the pre-Christian religion. It is possible that it was these officials that Adam confused with priests. In my opinion there are arguments supporting the notion that the pagan Svea rulers were closely connected to the Uppsala sanctuary in the Late Iron Age and that it should be described as an “aristocratic cultic site”. The sanctuary was thus located on the royal estate. The most important argument is the strong evidence of archaeology indicating the continuity of outstanding elite groups in Uppsala from the Merovingian Period to the Early Viking Age (c. 550–800/900). Archaeological evidence related to the top elite is, however, lacking from the last part of the tenth century onwards. It is possible that this reflects some new economic and political situation, where the new Christian kings now appear at other places. In my opinion, Uppsala was an important pre-Christian sanctuary up until c. 1060, even if royals did not take part in the religious activities there. Adam’s text and other sorces clearly indicate that (see ch. 4 above).

Pre-Christian kings and earls in Svetjud and Trøndelag seem to have owned and controlled some central sanctuaries, such as Uppsala and Lade. These sanctuaries most likely had a regional and perhaps even an inter-regional significance. They were erected at aristocratic farms. At Mære in Inn-Trøndelag there was a regional sanctuary which was organized in another fashion, at least during the second half of the tenth century. It could be described as a “cultic centre of the community type”. This sanctuary may have been administered by a cooperative of Þrœndir chieftains.

14.3 The Cult of Many Gods at the Chief Sanctuaries as Strategy of Power

In the written sources we have descriptions of cultic sites where several deities were worshiped. The most famous of these is probably Adam’s description of

56 Hervarar saga, (ed.) Jón Helgason 1924: 160. See also Orkneyinga saga 35. (Inter-)Regional Chief Sanctuaries 517 the Uppsala sanctuary (see ch. 4 above). He states that the statues of Óðinn, Þórr and Freyr were worshipped in the temple there. In my opinion, it is no coincidence that Adam mentions these specific gods in his description: place- names and archaeological finds confirm that they were worshipped in this part of Sweden.57 Adam’s information about the gods may thus be reliable. The distribution of theophoric place-names in the Mälaren region is very interesting (see fig. 28). The Freyr names seem to dominate in the south-western part of the area, more exactly in south Fjädrundaland, south Västmanland, and Södermanland.58 Place-names with Þórr and Óðinn, on the other hand, appear throughout the entire area, with a minor dominance in the eastern parts. In the light of the place-names and Adam’s text, it thus seems as if the common cult at the Uppsala sanctuary was some kind of synthesis, where different cultic groups, living in different areas of the Mälaren region, were supposed to

Figure 28 The distribution of theophoric place-names in the Mälaren region. Made by Fideli Sundqvist and based on Per Vikstrand 2001.

57 See Vikstrand 2001, 55ff., 115ff., 141ff., 167ff. and ch. 4 above. 58 Vikstrand 2001, 57ff. 518 Chapter 14 assemble.59 It is possible that this may have been connected to a strategy followed by the regional ruler who resided at Uppsala. In order to gain control over the people living in the Mälaren territory, he promoted the Uppsala cult. In the ceremonial space of the sanctuary the ruler was surrounded by images of gods. These images had a symbolic value for several of the groups and individuals that gathered in Uppsala at the common cultic feast. Here the king could show that he had a good relation to all these deities. He also took charge of the sacrifices made to them at Uppsala. One problem with this hypothesis is that the place-names may reflect different periods from those we may isolate by means of Adam’s information and the written sources. It is most probable, further, that on a local level the cult was not restricted to just one deity. In polytheistic systems the principle is that the individual deities guarded different spheres of life and society. Some deities may have been connected to nourishing and reproductive functions, others to physical and military protection, while some watched over intellectual aspects.60 One individual thus needed several gods with different functions to manage everyday life. However, it is quite possible that a single god may have been a unifying symbol for an individual, a family, an occupational group or a cultic community, etc. For instance, Freyr is called blótgoð Svía “the sacrificial god of the Svear” in Óláfs saga Tryggvasonar,61 and Óðinn is designated Gauta-týr ‘the god of the Götar’ in the skaldic poem Hákonarmál (960).62 In medieval sources we hear about pre-Christian Icelandic goðar who had a specific relation to a certain god. The goði family called Þórsnesingar were related to Þórr in Landnámabók and Eyrbyggja saga,63 while the goði Hrafnkell was related to Freyr in Hrafnkels saga Freysgoða.64 This personal devotion to one

59 Cf. Hellberg 1986a, 68f. 60 Adam actually mentions that the deities in Uppsala watched over different spheres of life: Þórr governs thunder and lightning, winds and pouring rain, fine weather and fertility, Óðinn wages war and gives man courage in the face of the enemy, and Freyr distributes peace and sensual delight to mortals. Adam IV,26. The gods were also related to different spheres in the cosmos. The Æsir were probably oriented towards the sky, while the Vanir were bound to the earth. See e.g. Schjødt 1991; 2008. See also the theories of Georges Dumézil 1959; 1973. For similar structures in other ancient polytheistic systems, see Gilhus & Thomassen 2010, 10. 61 Flateyjarbók I, 339. 62 Skj B1, 57; A1, 64. 63 Landnámabók, S85, H73, M25, Ísl. Fornr. 1, 124ff.; Eyrbyggja saga 3–4, Ísl. Fornr. 4, 6–10. 64 See e.g. Hrafnkels saga Freysgoða, Ísl. Fornr. 11, 112ff. (Inter-)Regional Chief Sanctuaries 519 god is sometimes expressed in the term fulltrúi in the medieval texts (see ch. 3 above). Most likely such personal devotion also existed in the Mälaren area in the Viking Age. A person who wore a Þórr’s hammer around his neck in this region was probably devoted to the thunder god. Therefore, it is not unreasonable to think that the three deities in the “temple of Uppsala” actually were part of a ruler strategy, the purpose of which was to gather all local cultic communities and individuals at this sanctuary, in order to display the sovereignty of the leader by means of ritual expressions. On such occasions the ruler could also take tribute from his subjects, as suggested by Adam.

14.3.1 Gudme on Funen A possible parallel to this cultic organization and ruler strategy may be found at Gudme (

65 See e.g. Thrane 1998; Hedeager 2001; 2011; Jørgensen 2011. 66 Kousgård Sørensen 1985, 133ff.; 1992, 233. Cf. SOL, 97. For some critical considerations, see Brink 2011. He states that -hem could mean ‘an area’, or ‘a district’ as well. 67 For a more thorough argumentation of this, see Kousgård Sørensen 1985, 135. 520 Chapter 14

Most likely several gods were also worshipped at the central sanctuary of Lade. According to Snorri’s Óláfs saga Tryggvasonar 59, for instance, King Óláfr dismantled the hof there, and removed all the property and all the decorations from it and from the images of gods. Several other Norse prose texts give a similar image (see ch. 7 above). The prose texts frequently mention that individuals or families on different social levels were devoted to a certain god in Norway. Earl Hákon had Þorgerðr as his tutelary god at the end of the tenth century (see ch. 12 above), while Guðbrandr of Hundþorp, in Guðbrandsdalr, had Þórr as his favourite deity, still at the beginning of the eleventh century.68 Theophoric place-names around the Trondheimsfjord likewise indicate a cult of a specific deity at local sanctuaries, such as †Óðins(s)alr, used for Beitstad or Sør-Beitstad, Nord-Trøndelag.69 It is thus possible that the common cult of many gods at Lade could be seen as a ruler strategy to gather different cultic groups and individuals of Trøndelag, and elsewhere in Norway, at the central sanctuary of Lade.

68 Óláfs saga helga 112, in Heimskringla II, Ísl. Fornr. 27, 184. 69 Sandnes & Stemshaug 2007, 341. Magnus Olsen mentions several theophoric place-names in Trøndelag, e.g. Onsøien (*Óðinsin/ *Óðinsey) at Byneset in Orkdælafylki. M. Olsen 1915, 268. Since I cannot find some of them in Norsk stadsnamnleksikon I guess that his interpretations today sometimes are considered as unreliable. Chapter 15 Valhǫll, Ceremonial Buildings and Warrior Lords

As noted in chapter 6 above, the Old Norse sources do not only describe cultic buildings and feasting halls situated in the real world. There are also traditions describing ceremonial buildings situated in the mythical world, among the gods. These mythical descriptions have many affinities with the ritual structure of the real ceremonial buildings, indicating that they had their Sitz im Leben in such milieus. Most likely they represent some type of ideal ceremonial buildings, and thus they may contribute information relevant for this study. In what follows I will discuss the martial dimension of the ceremonial buildings. This dimension appears in both mythical and real contexts. In Valhǫll, for instance, the presence of einherjar ‘single combatants, champions (especially the dead warriors who dwell in Valhǫll)’ is striking.1 In my opinion, the concept of Valhǫll should be related to certain hall milieus in Norway and Sweden (and Denmark), where warrior lords with warrior bands probably appeared quite frequently. Such comitatus institutions and milieus were most likely less common in Iceland. For rulers in Svetjud and Trøndelag, the martial and eschatological Valhǫll mythology was very important when constructing power.2 This martial mythology had its setting in an ideal hall milieu for retinues. When it was recited in Viking Age Iceland, it probably appeared in other social and ritual contexts.

15.1 The Martial and Eschatological Valhǫll Mythology

The reports from Grímnismál and Gylfaginning3 suggest that Valhǫll refers to a large aristocratic banqueting hall which is located in the central parts of Ásgarðr at a place called Glaðsheimr. It is gold-bright (gullbiarta) and it “rises peacefully, seen from afar” (see ch. 6 above). Grímnismál 9 describes the hall further thus:

1 On the mythical conception of Valhǫll, see e.g. Neckel 1913; Ellis Davidson 1943; Nordberg 2003; Dillmann 2007; Hultgård 2011. 2 The concept of eschatology refers to both the afterlife (individual eschatology) and the end and renewal of the world (cosmic or universal eschatology). Hultgård 2011, 298. Cf. Widengren 1969. 3 Grímnismál 8–10, 18–26 and Gylfaginning 37–40.

Mioc er auðkent, þeim er til Óðins koma, salkynni at siá; scǫptom er rann rept, scioldom er salr þakiðr, bryniom um becci strát.4

It is said that Valhǫll has “spear-shafts for rafters” and is thatched “with shields”. Essential structures of the house are thus made of weapons. When the skald states that “mail-coats are strewn on the benches” the audience understands quite clearly that this house was intended for warriors. It is also stated in this poem that Hroptr (i.e. Óðinn) “chooses every day those dead in combat” (st. 8), i.e. the einherjar (einheri sg.). It seems as if these einherjar were brought to Valhǫll ‘the hall of the slain’5 by the valkyrjur ‘those who choose the slain’.6 In Valhǫll these warriors had a paradise-like life, where they spent their days in the battlefield, but in the evenings they were alive again. In Vafþrúðnismál (41) the giant Vafþrúðnir replies thus to one of Óðinn’s question:

‘Allir einheriar Óðins túnom í hǫggvaz hverian dag; val þeir kiósa oc ríða vígi frá, sitia meirr um sáttir saman.’ 7

In Valhǫll the valkyrjur offer the einherjar mead from the goat Heiðrun’s horn,8 and they eat the meat from the constantly renewed boar Sæhrímnir each day. This boar meat is prepared for the warriors in the cauldron called Eldhrímnir by the cook Andhrímnir (see ch. 6 above).9 At Ragnarǫk 800 einherjar will march out of each of the 540 gates of Valhǫll on the gods’ side, in order to fight against the demonic and destructive power of the wolf (i.e. Fenrisúlfr). Grímnismál 23 states thus:

4 “It is very easy to recognize for those who come to / Óðinn, to see how his hall’s arranged./ the hall has spear-shafts for rafters, with shields/ it is thatched. Mail-coats are strewn on the/ benches.” Trans. Larrington. 5 The term Valhǫll probably derives from the ON words valr ‘the corpses lying on the battlefield’ and hǫll ‘hall’. Simek 2006, 346ff.; Hultgård 2011, 301. 6 The term valkyrja (sg.) derives from valr and the verb kjósa ‘to choose’, thus meaning ‘those who choose the slain’. See Simek 2006, 349. 7 “All the einheriar fight in Óðinn’s courts/ every day/ they choose the slain and ride from battle;/ then they sit the more at peace together.” Trans. Larrington. 8 Grímnismál 25, 36; Gylfaginning 39. 9 Grímnismál 18; Gylfaginning 38. Valhǫll, Ceremonial Buildings and Warrior Lords 523

Fimm hundruð dura oc um fiórom togom, svá hygg ec at Valhǫllo vera; átta hundruð einheria ganga ór einom durom, þá er þeir fara at vitni at vega.10

Magnus Olsen suggested that the description of Valhǫll in this stanza actually reflected experiences gained by Scandinavians who had visited the Colosseum in Rome, with its many combats of gladiators.11 This theory has been doubted by more recent researchers. Rudolf Simek suggests that the “number of 800 times 540 = 432,000 einherjar mentioned in Grímnismál can possibly be traced back to Hellenistic influence”.12 He is also open for other interpretations. Andreas Nordberg states that the skald probably never referred to a cosmic number. His intention was just to show that the building was huge and that the number of einherjar was great.13 In the perspective of the present study it is important to emphasize that Valhǫll in the mythic descriptions is some kind of a large banqueting hall decorated with weapons, and intended for ceremonial feasts. Óðinn invites the slain warriors to this place. At Ragnarǫk these warriors will march out and fight on the gods’ side against the giants.

15.2 The Warrior Halls in Late Iron Age Scandinavia and Valhǫll

The situation reflected in the mythical world may be associated with conditions in some human halls discovered by archaeologists in eastern Scandinavia. One human three-aisled warrior hall has been found at the Garrison of Birka in Lake Mälaren, on the slope north-west of the stronghold called Borg.14 This Viking Age building was divided into two large rooms. There were fireplaces in both rooms, and the building had two entrances. Many finds were made in the northern room, such as spearheads, shield bosses, webs of rings from mail coats, as well as mountings and keys for chests. These finds indicate that this

10 “Five hundred doors and forty/ I think there are in Valhǫll;/ eight hundreds of warriors will go together from one door/when they go to fight the wolf.” It is possible that hundruð refers to a “long hundred” = 120. Trans. Larrington. 11 M. Olsen 1938 (1931). 12 Simek 2006, 347. 13 Nordberg 2003, 229. 14 Holmquist Olausson 2001; Holmquist Olausson & Kitzler Åhfeldt 2002; Carlie 2004, 172ff., 291. 524 Chapter 15 room was a dwelling place for the warriors of Birka. In the southern room many animal bones and knives were found, including fragments of glass and a dragon head made out of bronze. This room was interpreted as an assembly hall for the warriors, used for feasts and religious ceremonies. In the western part of this room the most exclusive objects were found. They were concentrated in the floor layer in a restricted area. For instance, the dragon head was found there, a sword handle, and most of the glass fragments. According to Lena Holmquist, this was the place of the high-seat, i.e. the place “where the king sat when visiting the Garrison”.15 The sacrificial depositions in the post- hole of the eastern roof carrier also support the interpretation that warriors occupied the hall at Birka. In this hole two spearheads made of iron were found, as well as one decorated ferrule made out of bronze for a sword sheath, one decorated Þórr’s hammer made of horn, two Arabic coins and about 40 sawn-off comb cases.16 A lot of weapons were also found at other hall buildings in the Mälaren region, for instance at the halls of Helgö and Fornsigtuna.17 An enormous amount of weapons were also discovered outside the cultic house at Uppåkra.18 It could not be ruled out that the weapons in these halls may have been used for decorating the inside of these buildings, which could be compared with the appearance of Valhǫll, as suggested by Grímnismál.19 That real warrior halls in Norway sometimes were decorated with weapons is also attested by Sigvatr Þórðarson’s Austrfararvísur st. 16. In this stanza the skald praises one of the retinue-halls of King Óláfr Haraldsson in the early eleventh century:

Búa hilmis sal hjǫlmum hirðmenn, þeirs svan grenna (hér sék) bens, ok brynjum (beggja kost á veggjum), þvít ungr konungr engi (ygglaust es þat) dyggra húsbúnaði á hrósa; hǫll es dýr með ǫllu.20

15 “. . . där kungen satt när han besökte Garnisonen.” Holmquist Olausson 2001, 15. 16 Holmquist Olausson 2001. So too Carlie 2004, 174, 291. 17 Herschend 1995; 1998; Damell 1991; 1993. 18 L. Larsson & Lenntorp 2004. 19 Cf. Nordberg 2003, 203f. 20 Finnur Jónsson translated the stanza thus: “Hirdmændene, som mætter ravnen, udstyrer kongens sal med hjælme og brynjer; her ser jeg der findes rigeligt af bægge dele på Valhǫll, Ceremonial Buildings and Warrior Lords 525

In Óláfs saga helga 57 Snorri states that Óláfr Haraldsson had a “retinue-hall” (hirðstófa) at his royal residence (konungsgarðr) in Niðaróss, Trøndelag. This hall had two doors at both ends, and the king’s high-seat (hásæti) was situated in the middle of the hall. Opposite the king’s high-seat was the “lesser high- seat” (ǫðru ǫndugi). This was the seat of the king’s marshal, and next to him the noble guests were seated. Ale was carried around the fire during the ceremonial drinking feasts and when noble guests visited the king.21 In my opinion, Andreas Nordberg is on the right track when he states that the social context of the motif of Valhǫll is found in the band of the warrior aristocracy.22 The warrior band consisted of free men, usually from noble families, who voluntarily subordinated themselves to a warrior lord in order to do military service. As long as they belonged to the leader’s band they also constituted part of his family and household. The chieftain or ruler expected unlimited support and loyalty from his men, and in return he supplied them with food, drinking, lodging and weapons. He rewarded them with ceremonial feasts, at which he also distributed gifts. These feasts took place in the ruler’s hall, which seems to have been a crucial element in the organization of warrior bands. Among these groups there was an eschatological belief that those warriors who were killed in battle were consecrated or sacrificed to Óðinn. After death they were taken to his residence called Valhǫll, which was situated in the world of the gods. There they resided in lustre, light, and warmth, with the gods and other warriors slain in battle, i.e. the einherjar. There is thus some kind of relationship between the mythical Valhǫll and aristocratic warrior halls. In the halls at Birka, Fornsigtuna and Helgö, for instance, where lots of weapons have been found, the mythical complex related to the Valhǫll may have played an important role. However, all cultic halls in the investigation areas may not have had such connections and associations. In my opinion, this is due to the regional and contextual differences and variations. The hall at Hofstaðir, Mývatnssveit, for instance, has almost no weapons or other signs of the presence of a warrior aristocracy. At such halls the myths connected with Valhǫll may have appeared in other social contexts and played a more literary function, than being embedded in a warrior ideology.

væggene,—ti ingen ung konge har prægtigere husprydelse at rose sig af; det er udenfor tvivl; hallen er i enhver henseende herlig.” Skj. B1, 224. Page (1995, 53) translated it thus: “Men of the king’s guard/ Who feat the wound-swan/ Array the prince’s hall with hel- mets,/ Mail-coats. Here they hang./ No other ruler/ Can boast richer wall-hangings./ No fear of that./ Glorious is this hall.” 21 Óláfs saga helga 57, in Heimskringla II, Ísl. Fornr. 27, 72. 22 Nordberg 2003, 85ff., 300. 526 Chapter 15

The community around Hofstaðir was occupied with livestocking and farming, not with war (see below). It is quite striking that the oldest skaldic poems referring to the Valhǫll motif appear in royal contexts in Norway, where the institution of warrior bands was essential (see further below). In the tenth-century poems Eiríksmál and Hákonarmál, history and myth are combined. There we get the image that Valhǫll was a banqueting hall situated in the divine world, but also intended for human fallen kings and warriors from Norway.23 Eiríksmál (st. 3–4) calls it Óðins salr ‘Óðinn’s hall’.24 In stanza 1 we hear that the einherjar must clear up the hall, since an outstanding person (vísi) is arriving (i.e. the fallen Norwegian king Eiríkr blóðøx). The valkyrjur must get up to strew the benches and to clean the drinking vessels and to bring wine (vín bera). In stanza 5 Sigmundr asks Óðinn why he is expecting Eiríkr rather than other kings. Óðinn answers that Eiríkr has reddened his sword in many countries (i.e. he was a great warrior lord). Sigmundr replies: “Why did you then take the victory from him, if you thought him able (snjallr)?” Óðinn answers by referring to the uncertainty of knowing when the grey wolf will attack the homes of the gods (i.e. Ragnarǫk). The other poem, Hákonarmál, describes Valhǫll as grœnna heima goða “the green home of the gods” (st. 13).25 King Hákon the Good of Norway is welcomed to this place after his last battle. His arrival is announced to Óðinn by the two valkyrjur, Gǫndul and Skǫgul. There he enjoys the peace and truce of the einherjar (einherja grið) and he drinks beer together with the Æsir (þigg þú at ǫ́sum ǫl). This poem tells us that the reason King Hákon was taken to Valhǫll is that he must support the gods when the wolf breaks free from his bonds at Ragnarǫk (see st. 20). Both Eiríksmál and Hákonarmál refer to the theme of a cosmic eschatology, i.e. that the Norwegian warrior kings will be in “the ranks of the good forces at the side of the gods during the great battle at the end of time”.26 This was probably a glorious task awaiting brave kings and warriors in the upper classes in Viking Age Norway. Perhaps the Eddic poem Grímnismál (see above) should also be related to royal contexts in Norway. According to Helmut de Boor, this poem was

23 On these important poems, see e.g. Nordberg 2003, 48ff.; Hultgård 2011, 302ff.; SkP I, 171ff. Even if these kings have been regarded as Christians in the medieval traditions, the poems were probably made by contemporary pagan skalds. These skalds intentionally placed these kings in pagan contexts, perhaps for political reasons. King Hákon may furthermore have turned to paganism in the last part of his life. On this problem, see Hultgård 2011. 24 Skj. B1, 164ff. 25 Skj. B1, 57ff. Cf. SkP I, 187. 26 Hultgård 2011, 322. Valhǫll, Ceremonial Buildings and Warrior Lords 527

composed by one of the skalds appearing in the circle around Earl Hákon at Lade.27 I will not rule out the possibility that the Valhǫll motif therefore was more relevant in the context of aristocratic halls in Norway and Sweden where warrior lords with warrior bands probably appeared quite frequently, than in Iceland, where such military institutions were less common or totally lacking. At least it had a different social context in Iceland compared to the setting it had in Viking Age Norway and Svetjud. In what follows I will look a little more closely at this warrior institution and outline its distribution in Scandinavia.

15.3 The Distribution of Warrior Lords with Warrior Bands in Scandinavia

Some of the rulers in the uppermost elite of Scandinavia could be described as warrior lords. They had often a close group of people tied to their own person.28 It was often young landless men, recruited from high-ranking families, who constituted the private bodyguard, the ON hirð (cf. OE hīred ‘household, family, retinue’, hīredman ‘retainer, follower’).29 The hirð had a hierarchical structure. The retainers were the ruler’s companions when he travelled from estate to estate, entertained in the aristocratic warrior hall with food, mead, games, religious ceremonies, declamations of poetry and narratives. Sometimes they were treated to gifts in such halls (see chapter 9 above).

15.3.1 Svetjud and Trøndelag (and Norway) The Viking Age Svea kings exercised their power by ambulating with their hirð around to different sites in the Mälaren region. It seems as if they there had access to, or even owned, several places with halls. As noted in chapter 5, Late Iron Age halls in this region are frequently situated at places which in the Middle Ages reappear as crown lands and royal demesnes, for instance at Uppsala, Adelsö, Fornsigtuna, and Husby in Glanshammar.30 Some of these places were probably old manorial estates, which during the Viking Age may have been taken over or confiscated by the king and made into crown lands. Most likely the kings had representatives or administrators at such sites. According to the medieval provincial laws, such administrators (managers or bailiffs) were called brytar (OSw bryti sg.) and the royal farms were designated

27 See de Boor 1930; cf. Meulengracht Sørensen 1991a; 1993, 74–108; Nordberg 2003, 213. 28 E.g. Enright 1996; Norr 1998, 136f. Sawyer & Sawyer 1993, 92–95; Jón Viðar Sigurðsson 2008a, 117–125. 29 Clark Hall 1916, 158. 30 See Rahmqvist 1986; DMS 1.2: 180; DMS 1.7: 218–236; Damell 1991; Ekman 2000. 528 Chapter 15 husabyar.31 When travelling around his land the ruler visited these husabyar and other royal estates. He claimed hospitality from the brytar for himself and his retinues.32 This system was called OSw gæstning (cf. ON veizla).33 Most likely it was in similar contexts that the halls of the warrior bands also played a certain role in Viking Age Svetjud. The system of gæstning is actually indicated in sources deriving from the Late Viking Age. It seems, for instance, that the Svea king had a bryti at the Viking Age settlement of Adelsö, very close to Birka. The runic inscription of Hovgården (U 11) from the eleventh century reads as follows:

raþu:runaR:ret:lit:rista:toliR:bry[t]i:i roþ:kunuki:toliR:auk:gyla: litu:ris···-:þaun:hion:eftiR···k:merki *srni···*hakun*baþ:rista

Rað þu runaR. Rett let rista ToliR bryti i Roð kunungi. ToliR ok Gylla letu ris[ta. . .], þaun hion æftiR [si]k(?) mærki . . . Hakon bað rista.

You read the runes! Tolir had them cut right, bryti to the king in Roden. Tolir and Gylla had (these runes) carved, this pair after themselves as a memorial . . . Hakon bade carve.

The interpretation of the inscription is debated.34 Erland Hjärne has claimed that Tolir not was a bryti in Roden (i.e. the eastern part of the Upplandic coastal area) but only at the royal demesne of Adelsö.35 Today several scholars argue that Roden reached into Lake Mälaren, to Färentuna härad, i.e. west of Stockholm, and thus also to Adelsö.36 Most likely Tolir and/or some of his precursors guarded over and cared for the activities that took place there at Adelsö on the king’s behalf, such as the military defence, but also the entertainment, drinking ceremonies and cultic activities. The feasts—including religious ceremonies—were most likely performed in the presumed banqueting hall,

31 Schlyter 1887, 97 and 414f.; Hjärne 1951; Lindkvist & Ågren 1985, 20, 38; Brink 2003. 32 The brytar were originally unfree people, but gradually they gained power. OSw bryti means “som utdelar maten till de öfvriga tjenarne”. Medieval Swedish laws also describe the royal office of OSw lænsmaþer, i.e. a person who administered great farms (Sw huvudgårdar) in a skeppslag, hundare, härad etc. Lindkvist & Ågren 1985, 20, 38. 33 Lindkvist 1995; 2009. 34 See Brink (2003, 106) for reading and interpretation. 35 Hjärne 1951, 81. Cf. Brink 2003, 106. 36 See T. Zachrisson 2011a. Cf. T. Andersson 2004, 10, note 3. Valhǫll, Ceremonial Buildings and Warrior Lords 529 located on a plateau there. Perhaps this hall was also used as a dwelling for the king’s retinues.37 Another possible example of a Viking Age royal bailiff, with a hall for retainers, may be related to the town of Birka, on Björkö in Lake Mälaren, close to Adelsö. Rimbert mentions in Vita Anskarii (AD 870) that several Svea kings visited Birka, for instance King Bjǫrn and King Óláfr.38 At Birka the king had a royal counsellor (consiliarius regis), who also was the prefect of the site (praefectus vici, loci). Perhaps this office had some connections to the fortress called Borg at Birka and the hall discovered at the Garrison there. As noted above, the hall was interpreted as a place of assembly for the warriors. In this hall the place of the high-seat was identified. It has been assumed that the king sat there when visiting the Garrison.39 Most likely the king and his retainers in Birka celebrated ceremonial feasts there in the same way that the einherjar held great feasts together with Óðinn in Valhǫll. In the Viking Age the rulers of the Svear plundered and took tribute from the peoples living around the Baltic Sea.40 In Óláfs saga helga, Snorri describes how Eiríkr Emundarson, king at Uppsala, held a levy (ON leiðangr) every summer and sailed to various lands.41 The Swedish medieval laws report a military organization called OSw leþunger, which was led by the king.42 In the medieval

37 It is possible that the aristocratic settlement at Adelsö was developed into crown lands (bona regalia) belonging to “Uppsala Öd” (Uppsalaauðr) as early as the Viking Age. The runic stone indicates that. We know for sure that it was a royal estate in the Middle Ages. A medieval letter from 1200 mentions that Adelsö (Alsnö) was regarded as that: (ad) man- sionem regiam Alsnu. There are still remains of the medieval royal palace at Adelsö. See e.g. Thordeman 1920; Ambrosiani & Eriksson 1993, 39ff.; Brunstedt 1996, 10, 14; DMS 1.7: 218–236; SOL, 17, 21; 38 See e.g. Vita Anskarii 11, 26. 39 Perhaps Hergeir also had the function of the cultic leader of the local sanctuary. During Ansgar’s first visit to Birka the recently converted and pious Hergeir was appointed as counsellor. At one assembly, when all delegates were gathered for a meeting, Hergeir was attacked, since he did not praise the pagan gods but held on to his new faith. It seems as if the participants were upset, since Hergeir neglected and mismanaged his office as bailiff and representative of the king, i.e. his cultic functions to maintain good relations with the pagan gods. Hergeir’s function at Birka could be compared to Earl Sigurðr’s role as cultic leader on behalf of the king at the sanctuary of Lade. 40 E.g. Lindkvist & Ågren 1985, 12f., 22–25; Sawyer & Sawyer 1993, 91f.; Lindkvist 1995; M. G. Larsson 1990. 41 Heimskringla II, Ísl. Fornr. 27, 115. 42 Nu biuþær kunungær liþ ok leþung utt. biuþær utt roþ och reþ. Þa skal næmpnæ hampn ok stampn. ok styriman. ok hasætæ allæ. UL Kg 10. SSGL 3, 94. See also UL’s Mh 11:3. Cf. Hafström 1970 (1959), 49–57. Holmbäck & Wessén 1979 (1933–46), 55–62. 530 Chapter 15 period it functioned as a taxation system. The leþunger may, however, have deep roots in the native culture. The term hund/hundare, which was associated with this system, is ancient.43 It seems as if Tacitus was acquainted with it among the Germanic peoples: “The number of these men [the retainers] is fixed— one hundred from each canton: and among their own folk this, ‘the Hundred,’ is the precise name they use; what was once a number only has become a title and a distinction”.44 In addition Tacitus states: “Beyond these tribes the states of the Suiones [the Svear], not on, but in, the ocean, possess not merely arms and men but powerful fleets.”45 This information does not necessarily refer to the leþunger. But Rimbert and many Viking Age runic inscriptions mention organized maritime expeditions.46 Beowulf mentions “sea-kings” in Swīorīce.47 Other sources also support the idea that great men of Svetjud organized naval attacks on foreign peoples and took tribute. Like the Svea rulers, the Viking Age Norwegian kings (konungar), earls (jarlar) and chieftains (hersar) had a hirð, i.e. a group of professional warriors.48 These retainers (hirðsveinar, sveinar) had to be supplied with food, equipment, and weapons. In addition they also had a part of the booty. Therefore an economic surplus was important for the Norwegian rulers, in order to maintain their military power and keep the hirð alive. By plundering and taking tribute, this surplus was achieved. In order to supply his hirð the rulers also had estates and farms (No hovedgård, sentralgård), which gave economic income. Such royal farms also appear in medieval Trøndelag.49 Merete Røskaft has recently presented most of them in her dissertation.50 In Inn-Trøndelag they

43 On the hund/hundare system, see primarily T. Andersson 2004; 2005. Mats G. Larsson (1993, 122–126) believes that the places where husabyar appeared were ancient gathering places and connected to the leþunger system. 44 definitur et numerus: centeni ex singulis pagis sunt, idque ipsum inter suos vocantur, et quod primo numerus fuit, iam nomen et honor est. Germania 6. Trans. Hutton. 45 Suionum hinc civitates, ipso in Oceano, praeter viros armaque classibus valent. Germania 44. 46 See e.g. VA 30. Some ledung expeditions in the eleventh century were organized by some sort of central leadership. The best-known ledung is “Ingvar’s journey”, attested on thirty-two runic inscriptions from Sweden and in Icelandic prose. Since this expedition was very complex and involved many people from different districts of eastern Sweden, one assumes that it needed an organization. Cf. M. G. Larsson 1990; 1997, 9–163 and his literature. Lindkvist 1995, 38–64. 47 Þone sēlestan sæ-cyninga,/ þāra ðe in Swīorīce sinc brytnade,/ mærne þēoden. Beowulf 2382f. 48 Jón Viðar Sigurðsson 2008a, 24ff. 49 Hallan 1981 (1954). 50 Røskaft 2003. Cf. Brendalsmo 2006. Valhǫll, Ceremonial Buildings and Warrior Lords 531 included, for instance, the farms at Alstadhaug, Sakshaug, and Haug, and in Ut-Trøndelag the farms at Gryting, Meldal, Stein(e), Melhus, Lade, Logtun, and Værnes. Sometime during the Viking Age or the Early Middle Ages these farms were confiscated by the king and turned into royal estates.51 Most likely banqueting halls were erected there, where the ruler could gather his retainers, celebrate ceremonial feasts and distribute gifts to them. Perhaps the Middle Iron Age hall discovered at Hovde, Ørland, had similar functions. It had a strategic location in a military sense, between Ørland and Agdenes, controlling the entrance to the Trondheimsfjord. As noted in chapter 2, there was a tendency towards power concentration in Norway during the Early Viking Age. At the same time, the hirð system gradually became something more exclusive for kings, earls and great warrior lords. Only these social groups could afford to have such band of warriors. In Norway, as in Svetjud, a royal levy system (ON leiðangr), a seaborne military organization led by the king, was developed during the Late Viking Age. It has been suggested that King Hákon Aðalsteinsfóstri first organized this system in Vestlandet in the tenth century, with a model borrowed from England.52 It was then introduced to Trøndelag and finally also in Østlandet, when the Danish influence was defeated. This system was not completely established until the late twelfth century. The relationship between the Norwegian king and the members of his hirð was characterized by loyalty and friendship. According to medieval Norwegian laws, the king must have at least four retainers (No følgesvenner) accompanying him everywhere, from morning to evening.53 They protected the king and they were therefore not allowed to drink beer or wine. The royal hirð was one of the most important elements for constructing power in Late Viking Age and Early Middle Age Norway and Svetjud. Jón Viðar Sigurðsson states that it even was the central point of the society.54 It was in this martial milieu the greatest feasts were celebrated in impressive halls controlled by warrior lords. In these halls all kinds of cultural activities took place, such as the reciting of poems and other literary traditions.

51 Hallan 1981 (1954); Røskaft 2003. 52 Jón Viðar Sigurðsson 2008a, 124. 53 Jón Viðar Sigurðsson 2008a, 123f. 54 “Kongens hird var midpunkten i det norrøne samfundet. Dit dro alle men med ambis- joner; lokale høvidinger, unge men fra eliten og skalder.” Jón Viðar Sigurðsson 2008a, 118. 532 Chapter 15

15.3.2 Iceland As noted in chapter 2, most chieftains who came from Norway to Iceland were recruited from the lower elite. There was some competition between these chieftains (goðar), but it seems as if they usually did not use weapons and violence in order to attain a favourable position in society. It has been suggested that there were mainly three ways to mark high social status in the Icelandic society: by means of kinship, wealth and generosity.55 These strategies are visible in the story of Hjalti Þórðarson’s erfi, related in Landnámabók:

Hjalti, son of Þórðr skálp, came to Iceland, took possession of Hjaltadal with Kolbeinn’s approval, and made a home at Hof. His sons were Þorvaldr and Þórðr, both of them men of some note. The most magnificent funeral feast ever to be held in Iceland was the one his sons celebrated in honour of their father; there were about 1440 guests, and all the important people were presented with gifts when they left.56

In a similar fashion, gifts were distributed to the guests at the erfi of Unnr (Auðr) djúpúðga Ketilsdóttir, according to Laxdæla saga and Landnámabók (see ch. 9 above). As a result of the great feasts and the distribution of gifts, the chieftains established friendship and alliances between each other, and with the farmers who had pledged allegiance to them. Even if warfare occasionally also took place in Iceland, the society is often characterized by its peacefulness and productive life.57 The general assembly (alþingi) at Þingvellir was probably an important factor for establishing the peace. Jón Viðar Sigurðsson writes as follows about the Viking Age Icelandic society and the chieftains there:

Det var freden som kjennetegnet det isländske . . . samfunnet. Dette fram- går klarest av høvdingetypen som utviklet seg på Island. Den var svært

55 Jón Viðar Sigurðsson 1999, passim; 2008a, 58ff. Cf. Vésteinn Ólason (1998, 28) who states, however: “The goði who socially and politically was primus inter pares, had to secure his own position by creating alliances and by building up an economic position sufficiently strong to be able to support a small group of fighting men in his home at all times and a larger number in times of conflict.” 56 Hjalti son Þórðar skálps kom til Íslands ok nam Hjaltadal at ráði Kolbeins ok bjó at Hofi; hans synir váru þeir Þorvaldr ok Þórðr, ágætir menn. Þat hefir erfi verit ágætast á Íslandi, er þeir erfðu fǫður sinn, ok váru þar tólf hundruð boðsmanna, ok váru allir virðingamenn með gjǫfum brutt leiddir. Landnámabók, S 207, H 174, Ísl. Fornr. 1, 238. Trans. Hermann Pálsson & P. Edwards. 57 Vésteinn Ólason 1998, 30. Valhǫll, Ceremonial Buildings and Warrior Lords 533

lite krigersk. I alle de utallige beskrivelsene av høvdingene og deres egens- kaper er deres våpeferdigheter ikke et tema, men derimot deres gavmild- het og klokskap. Dette står i en klar motsetning til beskrivelser av norske konger, norske høvdinger og jarler, . . . i desse vektlegges deres krigerske karaktertrekk.58

There were thus no rulers or warrior lords riding around with a hirð in Viking Age Iceland, nor was there a leiðangr.59 The Icelandic þingmenn related to a goði were probably something different from the professional warrior bands visible in the Norwegian and Swedish contexts. It was not until the twelfth century, when power was concentrated in a few families, that the chieftains started to surround themselves with bodyguards and retainers.60

15.4 The Valhǫll Mythology as Part of Ruler Strategies in Svetjud and Trøndelag

We may thus not expect to encounter great aristocratic warrior halls, decorated with weapons, in Iceland. There were no Icelandic warrior lords with warrior bands there who could use such buildings as their dwelling. The halls found by archaeologists in Iceland, such as Hofstaðir, do not display a lot of weapons, compared to, for instance, the Garrison at Birka. The social and historical background of the Valhǫll mythology should therefore not primarily be searched in these Icelandic hall buildings. In Svetjud and Trøndelag, on the other hand, the great kings and earls used military force to express their power. In these regions we can see warrior lords with hirð retinues ambulating around the country to certain royal estates. At these estates, martial banqueting halls, decorated with weapons, were erected. In these halls the lord had his “household”, including the retainers, around him. He supplied them there with food, drink, entertainment, lodging and weapons. He rewarded them with ceremonial feasts, at which he also distributed gifts. Mythical traditions were probably recited in these buildings, including the eschatological Valhǫll mythology. It is quite possible that this mythology also included an element of ideological strategy. As Anders Hultgård has formulated it, the myth of Óðinn welcom- ing the dead warriors in Valhǫll, to be part of the einherjar troops, “can be part

58 Jón Viðar Sigurðsson 2008a, 65. 59 Jón Viðar Sigurðsson 2008a, 125. 60 Jón Viðar Sigurðsson 2008a, 123. 534 Chapter 15 of a pronounced male warrior ideology associated with the uppermost strata of west Scandinavian society . . . The Valhǫll myth becomes ideology through its function of supporting the ruling families’ claim to power and prestige. The male aristocracy could use the eschatological Óðinn mythology to point out their close relation with the god in afterlife, and thus anticipate their elevated position and make use of it for their political agenda in this world. In earthly life rulers had a key position because of their power to choose the members of their retinue, the hirð; which could later grant the followers a glorious afterlife in the hall of Óðinn.”61 In my opinion, the martial ideology of the eschatological Valhǫll mythology is thus quite clear. This mythological conception had the specific function of encouraging warriors to be brave in combat and not fear for their life or afterlife. If they died on the battlefield, they were rewarded by coming to Óðinn and the “warrior paradise” Valhǫll, i.e. a pleasant place intended only for warrior lords and high-ranking retainers. In Valhǫll the einherjar sat in the benches beside Óðinn, who was seated in his high-seat Hliðskjálf, and they were served meat and mead by the valkyrjur. At Ragnarǫk, the einherjar also took part in the great cosmic struggle, i.e. when the good powers (the gods) fight the destructive chaotic giants. As Hultgård puts it, the individual eschatology here was connected with the cosmic eschatology; “an afterlife ideology of a wider outlook was created which would have strongly appealed to Scandinavian warrior elites of the Viking Period”.62 This martial ideology and mythology, associated with the ideal warrior hall, must have been useful for the rulers in Norway and Svetjud, who probably needed to be surrounded with brave warriors. It can probably also be related to the religious strategy of monumentalization of kings and earls outlined in previous chapters. In life the warriors sat in the benches and celebrated feasts together with their lord, who presided in the high-seat of the hall. This situation resembled the conditions in Valhǫll, where the einherjar feasted with Óðinn, who had his place in Hliðskjálf. Most likely there was a micro-macro-cosmic symbolism (perceived by most cultic participators) between the human warrior halls and Valhǫll. During the religious feasts, when mythical traditions were recited and drinking rituals were performed, the real warrior hall may have alluded to aspects of the mythical Valhǫll or even been transformed into a perceived Valhǫll. In these ritual contexts the high-ranking retinues and their leaders were thus connected with the cosmos and the mythical world, to the einherjar, Óðinn and Valhǫll.63 The close relationship between the warrior

61 Hultgård 2011, 325. 62 Hultgård 2011, 325. 63 See e.g. Drobin 1991; Nordberg 2003. Valhǫll, Ceremonial Buildings and Warrior Lords 535 lord and Óðinn was emphasized when the ruler entered the high-seat. Perhaps he was considered godlike, since he also was sometimes believed to have descended from this deity (see ch. 3 above).

Sources presented in this chapter, as well as in all of Part 4, indicate that the hypothesis of this study could be supported. In the case of religious organization connected to the ceremonial buildings, there were clear differences between Iceland, on the one hand, and Sweden and Norway, on the other hand. In Iceland, ceremonial buildings were controlled by local chieftains. No trans- regional cultic sites for all Icelanders appeared there. In the Mälaren region and Trøndelag there were cultic places intended for different levels of society. In addition to more local shrines, there were also some regional and even inter- regional cultic sites, such as Uppsala, Mære and Lade. These cultic places were ruled and exploited by the society’s uppermost elite. Some warlords in Svetjud and Norway had warrior bands which celebrated feasts in martial halls. Most likely these halls had some connection to the Valhǫll mythology. Such warrior halls did not appear in Iceland. When mythic traditions about Valhǫll were recited in Iceland the social, ritual and ideological setting was probably different compared to the one in the Scandinavian homelands. There were also other regional differences in the ritual and symbolic strategies applied by rulers at cultic sites. In Norway and Sweden, large burial mounds were often placed adjacent to halls. Such monuments were lacking in the cultic places in Iceland. There is evidence that Swedish and Norwegian kings and earls were worshipped with offerings at the burial monuments, as after their deaths they were seen as mythical beings. A similar cult involving dead chieftains was absent from Iceland. Another regional difference is the distribution of gold foils. They are often found in connection with halls in Sweden and Norway, but so far none have been found in Iceland. The reason for this is probably linked to chronology. When Iceland was settled c. AD 870, the tradition of depositing gold foils at the high-seat had subsided. There may also be another explanation. Gold foils with a couple depicted on them may have represented the mythical ancestors of the rulers. The belief that the royal families were closely related to the gods may be the reason why they were deposited at the high-seat. In this belief we may see a type of religious strategy of monumentalization of power and rulership, which had the function of supporting the royal families’ claim to power and prestige. This belief and strategy existed in Norway and Sweden, but does not seem to have occurred in Iceland.

part 5 Epilegomena

∵ chapter 16 Conclusions

This book discusses the significance of Late Iron Age ceremonial buildings and sanctuaries in Scandinavia as parts of religious strategies to create authority, legitimation, and empowerment for rulers. The major purpose has been to analyse how rulers established and expressed their power by means of these buildings, ritual structures, objects and different types of rituals which took place in them. In contrast to previous research, this study has emphasized the context and the process, examining regional variations and developments over time. A particular focus has been placed on three regions: the Lake Mälaren Valley, Trøndelag and Iceland. These areas have been chosen because they index different political systems: kingship, earldom and chieftainship. They also have plenty of written sources about pre-Christian ceremonial buildings and sanctuaries. Since the spatial and temporal dimension of the sources is crucial for my analysis, archaeological finds and place-names also play a significant role, not least because they can be related directly to these regions and can also be connected to specific periods. Based on the discussion in Part 2, it can be concluded that tendencies towards concentration of power took place in the Mälaren area from the Migration Period onwards. Gradually there developed a type of royal power with ambitions to exercise domination (ríki) over several groups of people and minor chieftains, with a religious-political centre in Uppsala. Similar tendencies towards concentration of power occurred in Norway, and Trøndelag, from the Early Viking Age, where powerful kings and earls emerged. As a consequence of this political development, chieftains and great farmers emigrated from Norway to Iceland and other places in the Atlantic. In Iceland the Commonwealth (the Free State) was established around 930, a decentralized political system. There were no tendencies of any development there towards a “centralized state” during the Viking Age, comparable to the kingdoms and earldoms in contemporary Svetjud and Trøndelag. A concentration of power did not take place until the end of the Viking Age in Iceland, and no central royal power appeared on the scene as long as the Free State existed. By means of a case study, presented in Part 2, it can also be suggested that the religious strategies for gaining legitimacy and power, in general, diverged in the three areas, especially in the ways the rulers related to the mythical world. Old Norse sources indicate that the royal family of the Svear, called the Ynglingar, and the Earls of Lade claimed divine descent for their families, while

© koninklijke brill nv, leiden, ��6 | doi ��.��63/9789004307483_017 540 chapter 16 the Icelandic chieftain family, the Þórsnesingar, asserted a different relation to the gods, which was not phrased in terms of kinship. In the sources, the chieftains from Þórsnes described themselves as “dearest friend” (ástvinr) of the god (i.e. Þórr), and they took care of the cult to the deity at the local sanctuary. Their relation to the gods should primarily be described not as genealogic, but as cultic. Based on this case study, it can be argued that the kingdoms and earldoms in Svetjud and Trøndelag in general had developed ruler strategies which included the monumentalization of the political power. Such strategies could not be found in the more decentralized society of Iceland. This observation leads to the general hypothesis proposed by the present work. According to this general hypothesis, the political and social differences in the three regions had consequences for the strategies to establish power and rulership appearing at the ceremonial buildings and the sanctuaries in the three investigation areas. The findings in Svetjud and Trøndelag resembled each other, but diverged from conditions in Iceland. A detailed investigation of sources shows that this hypothesis can be confirmed to some extent (see Part 4 and below), but not as clearly as expected. There are several similar strategies of empowerment that recur in all three areas under investigation. These similarities are outlined in Part 3 of this book, called “General features”. It can be stated, for instance, that political leaders at different levels in society attempted to control and take care of the ceremonial buildings and sanctuaries in the three areas. These buildings and ritual structures were therefore often erected at the rulers’ farms. There are several reasons behind these actions. One may be related to economic aspects. The religious gatherings at the sanctuaries were a suitable occasion for the ruler to take tribute from the participants. Since the rulers were believed to have a particular relation to the mythical world (whether it was genealogic or cultic) they were also allowed or even expected to take care of the sanctuaries, organize the cult at their farms and also take tribute from the farmers for these activities. Economic, political, and religious activities and interests were thus intertwined. It can also be concluded that the large ceremonial buildings, including the ritual structures located outdoors beside these houses, were not only important sources of income for the rulers in the three investigation areas. They were also important (symbolic) media of power and appropriate arenas for displaying sovereignty. These buildings and structures were often built at specific topographic positions, at central places, on important lines of communications where a lot of people passed by. They were often erected on natural heights in the landscape or on artificially made plateaus (in Svetjud), i.e. places giving expression to the idea of loftiness. The halls often had splendid decoration, magnificent dimensions and architecture, especially in Svetjud and in Conclusions 541

northern Norway. These ritual-symbolic designs are clearly part of the strategy of monumentalization of power and rulership visible especially in Svetjud and Trøndelag. It can also be concluded that the interiors of the ceremonial buildings were just as impressive as the exterior of these houses, with large rooms and high roofs. In these rooms there were ritual structures and objects, such as the high-seats, high-seat posts, cultic images, tapestries, and hearths. They could be perceived as a ritual-symbolic strategy for legitimacy. The high-seat, for instance, was regarded as a place of honour in all three regions, and intended for the owner of the hall: the king, the earl or the chieftain. It was elevated in some way, so that the ruler actually looked down on the other participants in the room, thereby creating a spatial hierarchy of power. The other participants seem to have been seated according to rank and gender at specific places in the ceremonial building, but on a lower level than the hall owner. The sources indicate that the high-seat was a social and political furnishing with ideological dimensions, i.e. a thing materializing hierarchical relationships. In addition, it played a central role in religious and judicial contexts. It was thus a medium of tradition, power and authority (see further below). By means of the investigation in Part 3 it may also be concluded that the ritual structure of ceremonial buildings and sacred sites in Svetjud, Trøndelag and Iceland displayed some common features. In all three regions, there were ritual-symbolic elements or objects which quite regularly appeared at the ceremonial buildings. This book presents some evidence that these ritual structures were (at least for certain groups who had access to religious and culture knowledge) regarded as symbols of mythical or cosmic significance. Most of them also occur in connection with mythical buildings and mythical landscapes mentioned in the Old Norse traditions. Evidence indicates a micro- macro-cosmic relationship between the real ceremonial buildings and the mythical buildings. It seems as if the cultic places, ceremonial buildings and the high-seat sometimes were organized according to a prototype or model. Such sanctuaries were also regarded as thresholds to the Other World, i.e. places where humans could encounter divine beings. The owner and user of such a place had an enormous power since he or she appeared on a sacred place related to the gods. On such occasions, the scene needed to be set with specific properties. Religious symbols such as representations of the cosmic hall, the world-tree and the mythical well created the appearance that the ruler’s authority derrived from a realm beyond politics, society and the natural world. They were part of the strategies applied by rulers to construct legitimacy by drawing on mythological concepts and narratives. Some people never noticed the symbolic dimension of these impressive halls and sanctuaries. The ritual 542 chapter 16 structures may have been perceived by them in a more direct way as powerful. Likewise, they provided agency to the elite just by their material appearance. The sources indicate that cultic images were placed in the ceremonial buildings and sanctuaries in the three regions. They also suggest that the rulers wanted to persuade the cultic communities in a concrete sense that they had a specific relation to the gods and the mythical world. When the ruler performed the sacrifices in public, he interacted and communicated with these images by means of recitations, divinations and other rituals. Some of the cultic images represented the gods as they were seated in the high-seat. When the gods were seated in the high-seat they appeared in prototypical ritual roles, which could be imitated by the earthly rulers. Evidence presented in Part 3 indicates that rulers in the three areas consecrated lands, holy places and ceremonial buildings with a complex of rituals which were performed in public. During these rituals, these magnates proclaimed certain ritual restrictions which prevailed at these ritual spaces. By performing these rituals, they gained prestige and demonstrated in public their close relationship to sacred objects, holy places and the otherworld in general, which could increase their political ascendancy. Rituals performed inside the ceremonial buildings also contributed to the construction of political power and authority. The examination in Part 3 clearly indicates that rulers in the three areas played a central ritual role during the public and calendrical sacrifices and ceremonial banquets in the halls and cultic houses. During the ritual drinking they were seated in the high-seat and perhaps they consecrated the beaker and recited ritual formulae. It seems that they also performed divination rituals as part of the sacrifices, in order to acquire divine legitimacy for the actions to be taken. Some of the rituals were also related to gift-giving systems, where the religious-political authorities had the opportunity to manifest their wealth and generosity. The political leaders built alliances and gained friends on such occasions. These rituals took place in the ceremonial buildings, where the mythic-cosmic references in them emphasized the dignity and religious nature of these actors and actions. Regional similarities may also be seen in the oath rituals which probably took place in such buildings or sanctuaries. Oath-rings were important agencies of power during the religious ceremonies. They were carried in the hands or worn on the arms of the rulers while performing the religious and judicial rituals in the sanctuaries or when participating in public processions with other cultic leaders. In Part 4 (“Regional diversity”) materials and themes supporting the general hypothesis were discussed. The investigation indicates that some mythical, ritual and symbolic strategies for legitimacy related to ceremonial buildings and sanctuaries in Svetjud and Trøndelag never turned up in Iceland. These Conclusions 543 strategies (sometimes including a monumentalization of power) appeared more exclusively among the uppermost elite groups of Svetjud and Trøndelag, i.e. kings, earls and high aristocrats. It can be stated, for instance, that gold foil figures depicting mythical motifs mediated this mythic-cosmic notion. They played an important role for rulers of Svetjud and Trøndelag when constructing authority in the ceremonial buildings. These golden objects are not found in Iceland. In my opinion, the couple motif on the foils conveyed the idea that the ruling family descended from a divine/mythical couple (as suggested by Gro Steinsland). The fact that the foils were deposited at the high-seat of the ceremonial hall or cultic building may corroborate that they were media of agency and thus a concrete part of the more specific ideological strategy which may be found in Svetjud and Trøndelag (and also in Denmark). It is possible, however, that the distribution of these foils actually reflects chronological variations in Scandinavia as regards when authority claims were made. The use of gold foils in aristocratic contexts may have disappeared in the middle of the Viking Age. Since Iceland was first settled after 870, no such objects were found there. But the presence of Viking Age objects of gold, often arm- or neck-rings found in Sweden and Norway and the absence of such on Iceland underline that the use of gold as a means for ideological strategies display geographical variations that continue in the Viking Age. The study has also shown that great burial mounds and other monumental graves are situated close to the ceremonial buildings in Svetjud and Trøndelag. It can be argued that a pre-Christian veneration of dead and deified rulers in these areas took place at these mounds and monuments. Invoking the dead ruler at his mound was a way of appropriating the super-natural agencies in terms of receiving blessings from the dead divine ruler (i.e. the hero) in the Other World. Such monuments, located at the sanctuaries beside the ceremonial buildings, can be regarded as the material dimension of a more specific rulership. The living family was also in a genealogical sense related to the deified ancestor or divine hero dwelling in the mound. The spectators were reminded of these ideas when viewing the pompous graves at the sanctuaries and the great ceremonies that took place there. These monumental ritual structures and actions are clearly indicated at sanctuaries in Svetjud and Trøndelag, but not in Iceland. It can also be concluded that certain complex initiation rituals performed in the hall buildings were specific for Svetjud and Norway (and also Denmark) because these performative rituals concerned only kings and earls, not chieftains. During the royal erfi ceremony, the future ruler performed several rituals there. After drinking toasts to the dead ruler, the forefathers and the gods, 544 chapter 16 and after the swearing of oaths and holy vows, the heir entered the high-seat. The commemorative toasts and vows could be seen as constructions of power continuity and legitimate succession. The high-seat played a central role in this ritual system; it was the link between the rulers and their realms. It was also regarded as the ritual entrance both to rulership and to the Other World. When sitting in the high-seat, the ruler executed his functions as a leader in a judicially legitimate sense. The entering of the high-seat also comprised a complex of religious ideas. The ruler, for instance, had the ability to make contact with the gods and divine ancestors from there, which was one of the reasons why he had to be in his high-seat during the sacrifices at the ceremonial feasts. The rituals associated with the high-seat at these feasts could also be seen as ritual strategies for gaining legitimacy and constructing authority and royal power, which thereby shared in divine agency. The inquiry in Part 4 indicated that the religious organization related to the ceremonial buildings and sanctuaries differed in the three regions. Only in Svetjud and Trøndelag do we find chief sanctuaries with an (inter-)regional significance such as Uppsala, Lade and Mære. They were related to aristocratic centres, but some may have had a more communal base, such as Mære in the Middle Viking Age. Most likely the kings and earls used the chief sanctuaries and the cultic organization in order to gain legitimacy or claim/invoke power, sometimes with support from subordinated cultic leaders/specialists. At these monumental sites they could gather people from the whole region and medi- ate their sovereignty by means of ceremonies, ritual objects and a cult of many gods. In Iceland, the situation was different. No chief sanctuary with significance for all Icelanders is known. The shrines there were located on chieftains’ farms in the settlement districts, and in general they only played an important role for local groups. There is nothing to indicate that common religious feasts were celebrated at the communal assembly of the Alþingi. The warrior lords in Svetjud and Trøndelag gathered their retainers in the aristocratic halls in order to strengthen their loyalty to the leaders by means of religious ceremonies. Such martial hall milieus were probably not common in Iceland. In connection with this, it can be suggested that the conception of Valhǫll, and the myths related to it, were formulated in the Scandinavian homelands among bellicose warrior lords rather than in Iceland, where such phenomena did not appear. The eschatological Valhǫll myth included elements of a “warrior ideology” associated with the uppermost strata of Norway, Svetjud and Denmark. Engaging the Valhǫll myth may have provided support for the royal families’ claim to power. The narrative that the warrior lord and his retinues in the afterlife were welcomed by Óðinn to his dwelling place emphasized the close connection between the elite groups and the mythical world. Conclusions 545

By means of this myth the warrior lord could also attract good and noble warriors to his hirð, and ensure them a glorious afterlife which included the hospitality of Óðinn in his dwelling of Valhǫll. The Valhǫll myth also had another ideological function: when the earthly retinues were seated in the warrior hall celebrating religious feasts, the warrior lord, who sat in his high-seat, could have been associated with aspects of Óðinn seated in Hliðskálf together with the einherjar in Valhǫll. The Valhǫll mythology was probably recited in other social and ritual contexts in Iceland, since war-lords and warrior bands never existed there. The strategies including a monumentalization of power visible at ceremonial buildings and sanctuaries developed gradually in the Mälaren area and Trøndelag during the Late Iron Age. At the same time the first tendencies to power concentration may be discerned in these regions, probably after tough competition between minor kings and chieftains. During this period ceremonial buildings and sanctuaries were erected at many local rulers’ farms. They became important arenas for competition, where power could be expressed, claimed or invoked by means of a set of religious and ritual strategies. In the Migration and Merovingian Periods several powerful petty kings and chieftains in the Mälaren area and Trøndelag may have used almost the same mythical narratives, ritual performances, symbolic and organizational devices at these cultic sites. Gradually the strategies including a monumentalization of power were monopolized by the uppermost families in these regions. Yet some elements were still identical at the sanctuaries of chieftains, earls and kings. Since the uppermost elite never emigrated to Iceland, some of strategies or actions for claiming and invoking power at sanctuaries never arrived there. They could only be seen in the old homelands of the Scandinavians.

Bibliography

Adam av Bremen. Historien om Hamburgstiftet och dess biskopar. Trans. Emanuel Svenberg. Stockholm. 1985 (1984) Adam of Bremen. 2002 (1959). History of the Archbishops of Hamburg-Bremen. Trans. F. J. Tschan. New York. 2002 (1959) ———. Magistri Adam Bremensis Gesta Hammaburgensis Ecclesiae Pontificium. Scriptores rerum germanicarum in usum scholarum. Ex Monumentis Germaniae Historicis. Editio Tertia. (Ed.) B. Schmeidler. Hanover – Leipzig 1917. Ágrip. In Ágrip af Nóregskonunga sǫgum. Fagrskinna—Nóregs konunga tal. Íslenzk fornrit 29. (Ed.) Bjarni Einarsson. Reykjavík 1985. Ágrip af Nóregskonungasǫgum. (Ed.) M. J. Driscoll. Viking Society for Northern Research. 2008. Ágrip af Sǫgu Danakonunga. In Danakonunga Sǫgur: Skjǫldunga Saga. Knýtlinga Saga. Ágrip af Sǫgu Danakonunga. Íslenzk fornrit 35. (Ed.) Bjarni Guðnason. Reykjavík 1982. Alexanders saga. AM 519a 4o in the Arnamagnæan Collection, Copenhagen. Manuscripta Nordica 2. (Ed.) A. de Leeuw van Weenen. Copenhagen 2009. ALIE = Ancient Laws and Institutes of England 1–2. London 1840. Ammianus Marcellinus. Vols. I–III. (Ed.) Loeb. Trans. J. C. Rolfe. London – Cambridge, Massachusetts 1956–1963. Annales Fuldenses et Bertiniani. Monumenta Germaniae Historica. Scriptorum, Tomus I. III. Edidit Georgius Heinricus Pertz. Hanover 1826. Atlakviða in grœnlenzka. See Edda. Aulis Gellius. Noctes Atticae. The Attic Nights. II. (Ed.) Loeb. Trans. J. C. Rolfe. London – Cambridge, Massachusetts 1955. Austrfararvísur. See Skj. Bede. Baedae Opera Historica. Vols. I–II. (Ed.) Loeb. Trans. J. E. King. London – Cambridge, Massachusetts 1962–63 (1930). Beowulf and the Fight at Finnsburg. (Ed.) Fr. Klaeber. Lexington, Massachusetts 1950. Beowulf with the Finnesburg Fragment. Exeter Medieval English Texts and Studies. (Eds.) C. L. Wrenn & W. F. Bolton. Exeter 1996 (1953). Beowulf. An Edition with Relevant Shorter Texts. Including “Archaeology and Beowulf” by L. Webster. (Eds.) B. Mitchell & F. C. Robinson. Oxford 1998. Berudrápa. See Skj. Biblia Sacra. Iuxta vulgatam versionem. Recensuit et brevi apparatu instruxit Robertus Weber osb. Tomus II. Stuttgart 1969. Brandkrossa Þáttr. In Austfirðinga sǫgur: Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins þáttr stangarhǫggs, Ǫlkofra þáttr, Hrafnkels saga Freysgoða, Droplaugarsona saga, Brandkrossa þáttr, Gunnars 548 Bibliography

þáttr Þiðrandabana, Fljótsdæla saga, Þorsteins saga Síðu-Hallssonar, Draumr Þorsteins Síðu-Hallssonar, Þorsteins þáttr Austfirðings, Þorsteins þáttr Sǫgufróða, Gull-Ásu-Þórðar þáttr. Íslenzk fornrit 11. (Ed.) Jón Jóhannesson. Reykjavík 1950. Brennu-Njáls saga. Íslenzk fornrit 12. (Ed.) Einar Ól. Sveinsson. Reykjavík 1971. Brot af Sigurðarqviðo. See Edda. C7 = Cadastre for Uppland, Södermanland, Västmanland and Närke. The Swedish National Archives. Stockholm. C8 = Cadastral Notes Concerning the Inheritance Left by King Karl Knutsson (1473), Ivar Axelsson and Magdalena Karlsdotter. The Swedish National Archives. Stockholm. Cato. Marcus Porcius Cato, On agriculture; Marcus Terentius Varro, On agriculture. (Eds.) Loeb. Trans. W. D. Hooper & H. B. Ash. London – Cambridge, Massachusetts, 1934. Capitulatio de partibus Saxoniae. Monumenta Germaniae Historica. Legum Sectio II. Capitularia Regum Francorum. Tomus I. Denvo edidit Alfredus Boretius. Hanover 1883. Chronicon Lethrense. Scriptores Minores Historiæ Danicæ. Medii Ævi. Vol. I. (Ed.) M. Cl. Gertz. København 1917–1918. Clemen, Carolus. Fontes historiae religionis germanicae. Fontes Historiae Religionum III. Berlin 1928. CIMRM = Corpus inscriptionum et monumentorum religionis. Vols. 1–2. (Ed.) M. J. Vermaseren. 1956–1960. Concilium Turonense. Monumenta Germaniae Historica. Legum Sectio III Concilia. Tomus I. Concilia Aevi Merovingici. Recensvit Fridericus Maasen. Hanover 1893. Die Gotische Bibel. Hrsg. W. Streitberg. Heidelberg. DMS = Det medeltida Sverige. Riksarkivet. Stockholm. 1972–. DR = Danmarks runeindskrifter. I–II. Ved L. Jacobsen & E. Moltke under medvirkning af A. Bæksted & K. M. Nielsen. Text. København 1941–42. Droplaugarsona saga. In Austfirðinga sǫgur: Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins þáttr stangarhǫggs, Ǫlkofra þáttr, Hrafnkels saga Freysgoða, Droplaugarsona saga, Brandkrossa þáttr, Gunnars þáttr Þiðrandabana, Fljótsdæla saga, Þorsteins saga Síðu-Hallssonar, Draumr Þorsteins Síðu-Hallssonar, Þorsteins þáttr Austfirðings, Þorsteins þáttr Sǫgufróða, Gull-Ásu-Þórðar þáttr. Íslenzk fornrit 11. (Ed.) Jón Jóhannesson. Reykjavík 1950. Droplaugar sona saga. Mit nachwort und Karten hrsg. von U. Ebel. Texte des skandinavischen Mittelalters 2. Metelen 1990. Edda. Die Lieder des Codex Regius. Nebst verwandten Denkmälern. Band 1. Text. (Ed.) G. Neckel, 5. verbesserte Auflage von H. Kuhn. Heidelberg 1983 (1914). ———. Translations to English. See also The Poetic Edda. Egils saga. In Egils saga Skallagrímssonar. Íslenzk fornrit 2. (Ed.) Sigurður Nordal. Reykjavík 1988. ———. Egil’s saga. Trans. B. Scudder. In The Sagas of Icelanders. New York. Bibliography 549

Eiríksmál. See Skj. Eiríks saga rauða. In Eyrbyggja saga, Grœnlendinga sǫgur: Brands þáttr ǫrva. Eiríks saga rauða. Grœnlendinga saga. Grœnlendinga þáttr. Íslenzk fornrit 4. (Ed.) Einar Ól. Sveinsson & Matthías Þórðarson. Reykjavík 1985. Epistula Jeremiae. See Biblia Sacra. Eyrbyggja saga. Trans. Hermann Pálsson & Paul Edwards. London 1989. ———. In Eyrbyggja saga, Grœnlendinga sǫgur: Brands þáttr ǫrva. Eiríks saga rauða. Grœnlendinga saga. Grœnlendinga þáttr. Íslenzk fornrit 4. (Ed.) Einar Ól. Sveinsson & Matthías Þórðarson. Reykjavík 1985. Færeyingja saga. In Færeyingja saga. Óláfs saga Tryggvasonar eptir Odd Munk Snorrason. Íslenzk fornrit 25. (Ed.) Ólafur Halldórsson. Reykjavík 2006. Fagrskinna. In Ágrip af Nóregskonunga sǫgum. Fagrskinna—Nóregs konunga tal. Íslenzk fornrit 29. (Ed.) Bjarni Einarsson. Reykjavík 1985. Fjǫlsvinnsmál. See The Poetic Edda (ed. Hollander). Flateyjarbók: En samling af norske konge-sagaer med inskudte mindre fortællinger. Band 1–3. (Eds.) Guðbrandur Vigfusson & C. R. Unge. Christiania 1860–68. Flóamanna saga. In Harðar saga: Harðar saga Grímkelssonar (Hólmverja saga), Bárðar saga Snæfellsáss, Þorskfirðinga saga (Gull-Þóris saga), Flóamanna saga, Þórarins þáttr Nefjólfssonar, Þorsteins þáttr uxafóts, Egils þáttr Síðu-Hallssonar, Orms þáttr Stórólfssonar, Þorsteins þáttr tjaldstoeðings, Þorsteins þáttr forvitna, Bergbúa þáttr, Kumlbúa þáttr, and Stjörnu-Odda draumr. Íslenzk fornrit 13. (Eds.) Þórhallur Vilmundarson & Bjarni Vilhjámsson. Reykjavík 1991. Fljótsdæla saga. In Austfirðinga sǫgur: Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins þáttr stangarhǫggs, Ǫlkofra þáttr, Hrafnkels saga Freysgoða, Droplaugarsona saga, Brandkrossa þáttr, Gunnars þáttr Þiðrandabana, Fljótsdæla saga, Þorsteins saga Síðu-Hallssonar, Draumr Þorsteins Síðu-Hallssonar, Þorsteins þáttr Austfirðings, Þorsteins þáttr Sǫgufróða, Gull-Ásu-Þórðar þáttr. Íslenzk fornrit 11. (Ed.) Jón Jóhannesson. Reykjavík 1950. For Scírnis (Skírnismál). See Edda. Fóstbræðra saga. In Vestfirðinga sǫgur: Vestfirðinga sogur. Gisla saga Surssonar. Fostbræðra saga. Þattr Þormoðar. Havarðar saga Ísfirðings. Auðunar þattr vestfirzka. Þorvarðar þattr krakunefs. Íslenzk fornrit 6. (Eds.) Björn K. Þórólfsson & Guðni Jónsson. Reykjavík 1943. G = Gotlands runinskrifter. Granskade och tolkade av S. B. F. Jansson och E. Wessén. Sveriges runinskrifter. 11–12. KVHAAH. Stockholm 1962–1978. Gautreks saga. Fornaldar sögur Norðurlanda IV. (Ed.) Guðni Jónsson. Reykjavík 1950. Gísla saga. In Vestfirðinga sǫgur: Vestfirðinga sogur. Gisla saga Surssonar. Fostbræðra saga. Þattr Þormoðar. Havarðar saga Ísfirðings. Auðunar þattr vestfirzka. Þorvarðar þattr krakunefs. Íslenzk fornrit 6. (Eds.) Björn K. Þórólfsson & Guðni Jónsson. Reykjavík 1943. 550 Bibliography

———. Gisli Sursson’s Saga. Trans. M. S. Regal. In The Sagas of Icelanders. New York. Glælognskviða. See Skj. Grágás: Islændernes lovbog i fristatens tid. (Ed.) Vilhjálmur Finsen. Copenhagen: Berling, 1852. (Newprint Odense 1974). Gregorii I. Papae. Registrum Epistolarum. Monumenta Germaniae Historica. Epistolarum Tomi II Pars III. Edidit Ludovicus M. Hartmann. Berlin 1899. Gregory of Tours. Gregorii Episcopi Turonensis. Historia Francorum. Monumenta Germaniae Historica. Scriptores rerum Merovingicarum. I. Edidit W. Arndt & Br. Krusch. Hanover. MCMLI. ———. The History of the Franks. Trans. L. Thorpe. London. 1974. Grettis saga Ásmundarsonar. In Grettis saga Asmundarsonar, Bandamanna saga. Íslenzk fornrit 7. (Ed.) Guðni Jónsson. Reykjavík 1950. Grímnismál. See Edda. Gutalag och Guta saga jämte ordbok. (Ed.) H. Pipping. København 1905–07. Guta saga. The History of the Gotlanders. Viking Society for Northern Research. University College London. Vol. XII. (Ed.) C. Peel. London 1999. Gylfaginning. See Snorri Sturluson, Edda. Hákonardrápa. See Skj. Hákonarmál. See Skj and SkP. Háleygjatal. See Skj and SkP. Haraldskvæði. See Skj and SkP. Harðar saga. In Harðar saga: Harðar saga Grímkelssonar (Hólmverja saga), Bárðar saga Snæfellsáss, Þorskfirðinga saga (Gull-Þóris saga), Flóamanna saga, Þórarins þáttr Nefjólfssonar, Þorsteins þáttr uxafóts, Egils þáttr Síðu-Hallssonar, Orms þáttr Stórólfssonar, Þorsteins þáttr tjaldstoeðings, Þorsteins þáttr forvitna, Bergbúa þáttr, Kumlbúa þáttr, and Stjörnu-Odda draumr. Íslenzk fornrit 13. (Eds.) Þórhallur Vilmundarson & Bjarni Vilhjámsson. Reykjavík 1991. Hauksbók. (Eds.) Eiríkur Jónsson & Finnur Jónsson. Copenhagen 1892–96. Haustlǫng. See Skj. Hávamál. See Edda. Helgaqviða Hiǫrvarðzsonar. See Edda. Helreið Brynhildar. See Edda. Herodotus. Vol. I. (Ed.) J. Loeb. Trans. A. D. Godley. London – Cambridge, Massachusetts. 1960. Hesiod. Theogony. Works & Day. Testimonia. (Ed.) Loeb. Trans. & (ed.) G. W. Most. London – Cambridge, Massachusetts. 2007. Hervarar saga. In Heiðreks saga: Hervarar saga ok Heiðreks konungs. (Ed.) Jón Helgason. København 1924. ———. In Hervarar saga ok Heiðreks saga. (Ed.) E. O. G. Turville-Petre. Viking Society for Northern Research. London 2006 (1956). Bibliography 551

Historia Norwegie. (Eds.) I. Ekrem & L. B. Mortensen. Trans. P. Fischer. Copenhagen. 2003. Hænsa-Þóris saga. In Borgfirðinga sögur: Hænsa-Þoris saga. Gunnlaugs saga orm- stungu. bjarnar saga Hitdælakappa. Heiðarviga saga. Gisls Þattr Illugasonar. Íslenzk fornrit 3. (Eds.) Sigurður Nordal & Guðni Jónsson. Reykjavík 1938. Homer. The Odyssey I. (Ed.) Loeb. Trans. A. T. Murray. Rev. G. E. Dimock. London – Cambridge, Massachusetts. 1995. Hrafnkels saga Freysgoða. In Austfirðinga sǫgur: Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins þáttr stangarhǫggs, Ǫlkofra þáttr, Hrafnkels saga Freysgoða, Droplaugarsona saga, Brandkrossa þáttr, Gunnars þáttr Þiðrandabana, Fljótsdæla saga, Þorsteins saga Síðu-Hallssonar, Draumr Þorsteins Síðu-Hallssonar, Þorsteins þáttr Austfirðings, Þorsteins þáttr Sǫgufróða, Gull-Ásu-Þórðar þáttr. Íslenzk fornrit 11. (Ed.) Jón Jóhannesson. Reykjavík 1950. ———. Nordisk filologi 2. (Ed.) Jón Helgason. København – Oslo – Stockholm 1968. ———. The Saga of Hrafnkel Frey’s Godi. Trans. T. Gunnell. In The Sagas of Icelanders. New York. Hrólfs saga kraka. Hrólfs saga kraka og Bjarkarímur. (Ed.) Finnur Jónsson. København. 1904. Hs = Hälsinglands runinskrifter (är under utarbetande). Huitfeldt, A. Danmarks Riges Krønike. Fra Kong Dan indtil Knud VI 1603. København. 1978. Hymiskviða. See Edda. Hyndlolióð. See Edda. Ibn Fadlān. See Montgomery, J. E. 2000 below. Indiculus Superstitionum et Paganiarum. Monumenta Germaniae Historica. Legum Sectio II. Capitularia Regum Francorum. Tomus I. Denvo edidit Alfredus Boretius. Hanover 1883. Islandske Annaler indtil 1578. (Ed.) G. Storm. Christiania. 1888. Íslendingabók. In Íslendingabók. Landnámabók. Íslenzk fornrit 1. (Ed.) Jakob Bene diktsson. Reykjavík 1986. Jómsvíkingasaga. (Ed.) C. af Petersens. København – Lund. 1882. Jordanes. Getica: Om goternas ursprung och bedrifter. Trans. A. Nordin. Stockholm. 1997. Kjalnesinga saga. In Kjalnesinga saga: Jǫkuls þáttr Búasonar. Víglundar saga. Króka- Refs saga. Þórðar saga hreðu. Finnboga saga. Gunnars saga Keldugnúpsfífls. Íslenzk fornrit 14. (Ed.) Jóhannes Halldórsson. Reykjavík 1959. Klüwer, Lorentz Diderich. Norske Mindesmærker, aftegnede paa en Reise igjennem en Deel af det Nordenfjeldske. Faksimile, Trondheim 1960. Christiania, 1823. Kormáks saga. In Vatnsdæla saga. Hallfreðar saga. Kormáks saga. Hrómundar þáttr halta. Hrafns þáttr Guðrúnarsonar. Íslenzk fornrit 8. (Ed.) Einar Ól. Sveinsson. Reykjavík 1939. Krákumál. See Skj. 552 Bibliography

Kristni saga. In Kristnisaga. þáttr þorvalds ens víðfǫrla. Þáttr Ísleifs biskups Gizurarsonar. Hungrvaka. Altnordische Saga-Bibliothek 11. (Ed.) B. Kahle. Halle 1905. ———. In Íslendingabók and Kristni saga. Viking Society for Norther Research. Text Series. Trans. S. Grønlie. London 2006. Landnámabók. In Íslendingabók. Landnámabók. Íslenzk fornrit 1. (Ed.) Jakob Bene diktsson. Reykjavík 1986. ———.The Book of the Settlements. Trans. Hermann Pálsson & P. Edwards. Winnipeg. Laxdœla saga. In Laxdœla saga. Halldórs þættir Snorrasonar. Stúfs þáttr. Íslenzk fornrit 5. (Ed.) Einar Ól. Sveinsson. Reykjavík1934. ———. The Saga of the People of Laxardal. Trans. K. Kunz. In The Sagas of Icelanders. New York. Legenda sancti Eskilli. See SRS II, 389–399. Leges Langobardorum 643–866. Bearbeitet von Franz Beyerle. Göttingen 1962 (1947). Lex Frisionum. See Clemen, Carolus. Lex Ribuari. See Clemen, Carolus. ———. Die Gesetze des Karolingerreiches 714–911. Hrsg. K. A. Eckhardt. Weimar 1934. Lokasenna. See Edda. Lucretius. On the Nature of Things. (Ed.) Loeb. Trans. W. H. D. Rouse. London – Cambridge, Massachusetts 1924. Magnus Lagabøters landslov. Trans. A. Taranger. Kristiania 1915. NGL = Norges Gamle Love indtil 1387. Vols. I–III. (Eds.) R. Keyser & P. A. Munch. Christiania 1846–49. Njáls saga. See Brennu-Njáls saga above. Ög = Östergötlands runinskrifter. Granskade och tolkade av E. Brate. Sveriges Runinskrifter. 2. KVHAAH. Stockholm. 1911–1918. Ǫgmundar þáttr dytts. In Eyfirðinga sǫgur: Víga Glúms saga. Ǫgmundar þáttr dytts. Þorvalds þáttr tasalda. Svarfdæla saga. Þorleifs þáttr jarlsskálds. Valla-Ljóts saga. Sneglu-Halla þáttr. Þorgríms þáttr Hallasonar. Íslenzk fornrit 9. (Ed.) Jónas Kristjánsson. Reykjavík 1956. Öl = Ölands runinskrifter. Granskade och tolkade av S. Söderberg & E. Brate. Sveriges Runinskrifter. 1. KVHAAH. Stockholm. 1901–1906. Óláfsdrápa sœnska. See Skj. Óláfs saga Tryggvasonar en mesta I, Fornmanna sögur 1–3. Óláfs saga Tryggvasonar eptir Odd munk Snorrason. In Fœreyinga saga. Óláfs saga Tryggvasonar eptir Odd munk Snorrason. Íslenzk fornrit 25. (Ed.) Ólafur Halldórsson. Reykjavík 2006. Olaus Magnus. Historia de gentibus Septentrionalibus. Farnborough 1971. Orkneyinga saga. In Orkneyinga saga: Legenda de sancto Magno. Magnúss saga skemmri. Magnúss saga lengri. Helga þáttr ok Úlfs. Íslenzk fornrit 34. (Ed.) Finnbogi Guðmundsson. Reykjavík 1965. Bibliography 553

———. The History of the Earls of Orkney. Trans. Hermann Pálsson and Paul Edwards. London 1981 (1978). Ovid. Ovid’s Fasti. (Ed.) J. Loeb. Trans. J. G. Frazer. London – Cambridge, Massachusetts. 1959 (1931). Plutarch. Plutarch’s Lives. Vol. I. (Ed.) J. Loeb. Trans. B. Perrin. London – Cambridge, Massachusetts. 1959 (1914). Porphyrios. De abstinentia. See Cumont 1896–1899 II, 39ff. ———. De antro nympharum. See Cumont 1896–1899 II, 42f. ———. The Cave of the Nymphs in the Odyssey. Trans. and (ed.) Seminar Classics 609. Arethusa Monographs, no. 1. Buffalo. 1969. Ragnarsdrápa. See Skj. Rannsakningar efter antikviteter. Band I. Uppland, Västmanland, Dalarna, Norrland, Finland. Häfte I. Text. KVHAA. Utg. I. Olsson et al. Stockholm 1960. RApp = Documents on Paper in the Swedish National Archives. Stockholm. Rerum gestarum Saxonicarum. See Widukind. Rígsþula. See Edda. Rimbert. Anskar. The Apostle of the North. 801–865. Trans. C. H. Robinson. London. 1921. Rimbert. Vita Anskarii auctore Rimberto. Accedit Vita Rimberti. (Ed.) G. Waitz. Hanover 1884. Saxo Grammaticus. Saxonis Gesta Danorum. (Eds.) J. Olrik & H. Ræder. Hauniæ 1931. ———. The History of the Danes, Book I–IX. Edition and Commentary Hilda Ellis Davidson. Translation Peter Fischer. Cambridge 2008 (1996). ———. Gesta Danorum. I–II. Ed. K. Friis-Jensen. Trans. P. Fischer. Oxford 2015. Schøning, Gerhard. Reise som giennem en Deel af Norge i de Aar 1773, 1774, 1775 paa Hans Majestæt Kongens Bekostning er giort og beskreven. Udgivet af Det Kgl. Videnskabers Selskab i Trondhjem i Anledning af dets 150-Aars Jubilæum. Bind I–II. Ved Karl Rygh. Trondhjem. 1910. Sigrdrífomál. See Edda. SkP I = Skaldic Poetry of the Scandinavian Middle Ages I. Poetry from the Kings’ sagas 1 : From mythical times to c. 1035. (Ed.) D. Whaley. Turnhout 2012. SkP II = Skaldic Poetry of the Scandinavian Middle Ages II. Poetry from the Kings’ sagas 2 : From c. 1035 to c. 1300. (Ed.) K.E. Gade. Turnhout 2009. Skáldskaparmál. See Snorri Sturluson, Edda. Skírnismál (For Scírnis). See Edda. Skj = Den Norsk-Islandske Skjaldedigtning 800–1400. AI–II, BI–II. (Ed.) Finnur Jónsson. Köbenhavn 1967–1973 (1912–15). Skjǫldunga Saga. In Danakonunga Sǫgur: Skjǫldunga Saga. Knýtlinga Saga. Ágrip af Sǫgu Danakonunga. Íslenzk fornrit 35. (Ed.) Bjarni Guðnason. Reykjavík. 1982. Sm = Smålands runinskrifter. Granskade och tolkade av Ragnar Kinander. In Sveriges runinskrifter 4. KVHAAH. Stockholm – Uppsala 1935–61. Snorri Sturluson. Edda. (Ed.) Finnur Jónsson. København 1900. 554 Bibliography

———. Heimskringla I–III. Íslenzk fornrit 26–28. (Ed.) Bjarni Aðalbjarnarson. Reykjavík 1979 (1941, 1945, 1951). ———. Heimskringla. History of the Kings of Norway. Trans. L. M. Hollander. Austin 1964. ———. Ynglingasaga. Nordisk filologi 6. (Ed.) E. Wessén. København – Oslo – Stockholm 1964. ———. Nordiska kungasagor I–III. Trans. K. G. Johansson. Stockholm 1991–1993. ———. Edda. Trans. A. Faulkes. London 1987. ———. Edda. Prologue and Gylfaginning. (Ed.) A. Faulkes. London 1988. ———. Snorres Edda. Trans. K. G. Johansson & M. Malm. Stockholm 1997. ———. Edda. Skáldskaparmál. 1: Introduction, Text and Notes. (Ed.) A. Faulkes. London 1998. ———. Skáldskaparmál. 2: Glossary and Index of Names. (Ed.) A. Faulkes. London 1998. ———. Histoire des rois de Norvège par Snorri Sturluson. Première partie. Traduit du vieil islandais, introduit et annoté par François-Xavier Dillmann. Paris. 2000. ———. The Uppsala Edda. (Ed.) Heimir Pálsson. London 2012. Sö = Södermanlands runinskrifter. Granskade och tolkade av Elias Wessén och Erik Brate. Sǫgubrot af fornkonungum. In Danakonunga Sǫgur: Skjǫldunga Saga. Knýtlinga Saga. Ágrip af Sǫgu Danakonunga. Íslenzk fornrit 35. (Ed.) Bjarni Guðnason. Reykjavík. 1982. Sǫgubrot af Skjǫldungasaga. See Sǫgubrot af fornkonungum. Sonatorrek. See Skj. SRS = Scriptores rerum Suecicarum medii aevi. I–III. (Eds.) E. M. Fant et al. Upsaliae 1818–1876. SSGL = Samling af Sweriges Gamla Lagar. Corpus Iuris Sueo-Gotorum Antiqui. Vols. 1–13. (Eds.) [H. S. Collin &] C. J. Schlyter. Stockholm 1827–1877. Stjórn. Gammelnorsk Bibelhistorie: Fra Verdens Skabelse til det Babyloniske Fangenskab. (Ed.) C. R. Unger. Christiania. 1862. Strabo. The Geography of Strabo. Vols. I–VIII. (Ed.) Loeb. Trans. H. L. Jones. London – Cambridge, Massachusetts. 1917–1932. Sturlunga saga I. (Eds.) Jón Jóhannesson, Magnús Finnbogoson, Kristján Eldjárn. Reykjavík. 1946. Svarfdæla saga. In Eyfirðinga sǫgur: Víga Glúms saga. Ǫgmundar þáttr dytts. Þorvalds þáttr tasalda. ———. Þorleifs þáttr jarlsskálds. Valla-Ljóts saga. Sneglu-Halla þáttr. Þorgríms þáttr Hallasonar. Íslenzk fornrit 9. (Ed.) Jónas Kristjánsson. Reykjavík 1956. Svenska landskapslagar. Tolkade och förklarade för nutidens svenskar av Åke Holmbäck och Elias Wessén. 1–5. Stockholm. 1979 (1933–46). Sverris saga. Sverris saga etter Cod. AM 327 4o. Utgjevi av Den Norske Historiske Kildeskriftkommission ved G. Indrebø. Kristiania. 1920. Bibliography 555

Tacitus, Cornelius. Germania. Originalets text med svensk tolkning jämte inledning och kommentar av A. Önnerfors. Stockholm 1960. ———. Germania. In Agricola, Germania, Dialogus. (Ed.) Loeb. Trans. M. Hutton. London – Cambridge, Massachusetts. 1992 (1914). The Anglo-Saxon Chronicle. Trans. and (ed.) M. Swanton. London 1997. The Earliest Norwegian Laws: Being the Gulathing Law and the Frostathing Law. Trans. L. M. Larson. New York 1935. The Great Saga of St Olaf. Den store saga om Olaf den helige. (Eds.) O. A. Johnsen & Jón Helgason. Utg. För kjeldeskriftfondet. Band 1–2. Oslo. 1941. The Legendary Saga of St Olaf. Olafs saga hins helga. Efter pergamenthandskrift I Uppsala universitetsbibliotek, Delagardieske samling nr. 8II utg. Av den Norske Historiske Kildeskriftkommision. (Ed.) O. A. Johnsen. (Det norske historiske kilde- skriftfonds skrifter 47.). Kristiania. 1922. The New English Bible. (Ed.) Collins. 1970 (1961). The Old English Orosius. Early English Text Society S.S. 6. (Ed.) J. Bately. London – New York – Toronto. 1980. The Poetic Edda. L. M. Hollander. Austin 1990 (1962). ———. A new translation by Carolyne Larrington. Oxford 1996. ———. Vol. 2. Mythological Poems. Trans. and (ed.) U. Dronke. Oxford 1997. Thietmar av Merseburg, Chronik. Thietmari Merseburgensis Episcopi Chronicon. Trans. and (ed.) W. Trillmich. Darmstadt 1957. Þjóðólfr inn fróði ór Hvini. Ynglingatal. KVHAAH. 28:2. Red. och översättning A. Noreen. Stockholm 1925. Þrymsqviða. See Edda. Translatio S. Alexandri. Monumenta Germaniae Historica. Scriptorum. Tomus II. Edidit Georgius Heinricus Pertz. Hanover 1829. U = Upplands runinskrifter. Granskade och tolkade av E. Wessén & S. B. F. Jansson. Sveriges runinskrifter. 6–9. KVHAAH. Stockholm 1940–58. Vafþrúðnismál. See Edda. Vápnfirðinga saga. In Austfirðinga sǫgur: Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins saga hvíta, Vápnfirðinga saga, Þorsteins þáttr stangarhǫggs, Ǫlkofra þáttr, Hrafnkels saga Freysgoða, Droplaugarsona saga, Brandkrossa þáttr, Gunnars þáttr Þiðrandabana, Fljótsdæla saga, Þorsteins saga Síðu-Hallssonar, Draumr Þorsteins Síðu-Hallssonar, Þorsteins þáttr Austfirðings, Þorsteins þáttr Sǫgufróða, Gull-Ásu-Þórðar þáttr. Íslenzk fornrit 11. (Ed.) Jón Jóhannesson. Reykjavík 1950. Vatnsdæla saga. In Vatnsdæla saga: Hallfreðar saga. Kormáks saga. Hrómundar þáttr halta. Hrafns þáttr Guðrúnarsonar. Íslenzk fornrit 8. (Ed.) Einar Ól. Sveinsson. Reykjavík. 1939. ———. The Saga of the People of Vatnsdal. Trans. A. Wawn. In The Sagas of Icelanders. New York. Vellekla. See Skj and SkP. 556 Bibliography

Virgil. Aeneid. In Eclogues, Georgics, Aeneid. (Ed.) Loeb. Trans. H. R. Fairclough. Cambridge, MA. 1916. Vg = Västergötlands Runinskrifter. Granskade och tolkade av H. Jungner och E. Svärdström. Sveriges Runinskrifter. 5. KVHAAH. Stockholm – Uppsala. 1940–1970. Víga-Glúms saga. Eyfirðinga sǫgur: Víga Glúms saga. Ǫgmundar þáttr dytts. Þorvalds þáttr tasalda. Svarfdæla saga. Þorleifs þáttr jarlsskálds. Valla-Ljóts saga. Sneglu-Halla þáttr. Þorgríms þáttr Hallasonar. Íslenzk fornrit 9. (Ed.) Jónas Kristjánsson. Reykjavík 1956. Vita Anskarii. See Rimbert. Vita Willehadi. See Clemen, Carolus. Vita Willibrordi. See Clemen, Carolus. Vǫlsunga saga. Fornaldar sögur Norðurlanda I. (Eds.) Guðni Jónsson & Bjarni Vilhjálmsson. Reykjavík 1943. Vǫlundarkviða. See Edda. Vǫluspá. See Edda. Vs = Västmanlands runinskrifter. Granskade och tolkade av S. B. F. Jansson. Sveriges runinskrifter 13. KVHAAH. Stockholm 1964. Widukind. Rerum Gestarum Saxonicarum. Libri tres. Scriptores Rerum Germanica- rum in Usum Scholarum ex Monumentis Germaniae Historicis (Separatim Editi). In Verbindung mit H.-E. Lohmann neu bearbeitet von P. Hirsch. Hanover 1935. Willibaldus. Vita Bonifatii. See Clemen, Carolus. Ynglingatal. See Snorri Sturluson. Heimskringla I. See also Þjóðólfr inn fróði ór Hvini Skj, and SkP.

Literature

Adams J. L. & T. Mikelson. 1987. Legitimation. In The Encyclopedia of Religion. Vol. 8. (Ed.) M. Eliade. New York – London, 499–509. Ahlberg, A. W. et al. 1964. Latinsk-Svensk Ordbok. Stockholm. Ahlsson, L.-E. 1992. Lånord i svenska runinskrifter. In Blandade runstudier 1. Runrön 6, 7–12. Ahn, G. 1992. Religiöse Herrscherlegitimation im Achämenidischen Iran: Die Vorausset zungen unddie Struktur ihrer Argumentation. Acta Iranica. Encyclopédie Permanente des Études Iraniennes Publiée par le Centre International d’Études Indo-Iraniennes. Troisième Série. Textes et Mémoires, Vol. XVII, Acta Iranica. 31 Leiden – Louvain. Åkerlund, W. 1939. Studier över Ynglingatal. Skrifter utgivna av Vetenskaps-societeten i Lund 23. Lund. Alkarp, M. 1993. Adam av Bremen och Gamla Uppsala. Uppsats CD. Inst. för arkeologi. Uppsala Universitet. Bibliography 557

———. 1997. Källan, lunden och templet—Adam av Bremens Uppsalaskildring i ny belysning. Fornvännen (92), 155–161. ———. 2007. . . . men där är också mycken galenskap—Adam av Bremen, arkeologin och Gamla Uppsala. In Kult, guld och makt: Ett tvärvetenskapligt symposium i Götene. Historieforum Västra Götaland, Serie B, Vetenskapliga rapporter och Småskrifter No. 4. (Ed.) Ingemar Nordgren. Skara, 198–220. Alkarp, M. & N. Price. 2005. Tempel av guld eller kyrka av trä? Markundersökningar vid Gamla Uppsala kyrka. Fornvännen (100), 261–272. Ambrosiani, B. & B. G. Erikson. 1991–96. Birka Vikingastaden. Vols. 1–5. Stockholm. Andersson, G. 2005. Gravspråk som religiös strategi: Valsta och Skalby i Attundaland under Vikingatid och tidig Medeltid. Riksantikvarieämbetet. Arkeologiska undersök- ningar. Nr. 61. Stockholm. ——— et al. 2004. Att föra gudarnas talan—figurinerna från Lunda. Riksantikvarie ämbetet. Arkeologiska undersökningar. Skrifter. Nr. 55. Stockholm. Andersson, T. 1982. Hund, hundare och härad från språklig synpunkt. Bebyggelsehistorisk tidskrift (4), 52–66. ———. 1986. Germ. Hof—Hügel, Hof, Heiligtum. In Sprache und Recht: Beiträge zur Kulturgeschichte des Mittelalters. Festschrift für Ruth Schmidt-Wiegand zum 60. Geburtstag. (Eds.) K. Hauck et al. Berlin, 1–9. ———. 1991. The Origin of the Tuna-Names Reconsidered. In People and Places in Northern Europe 500–1600: Essays in Honour of Peter Hayes Sawyer. (Eds.) I. Wood & N. Lund. Woodbridge, 197–204. ———. 1992a. Orts- und Personennamen als Aussagequelle für die altgermanische Religion. In Germanische Religionsgeschichte: Quellen und Quellenprobleme. (Eds.) H. Beck et al. Berlin – New York, 508–540. ———. 1992b. Kultplatsbeteckningar i nordiska ortnamn. In Sakrale Navne. NORNA- rapporter 48. (Eds.) G. Fellows-Jensen & B. Holmberg. Uppsala, 77–105. ———. 1992c. Haupttypen sakraler Ortsnamen Ostskandinaviens. In Der historische Horizont der Götterbild-Amulette aus der Übergangsepoche von der Spätantike zum Frühmittelalter. (Ed.) K. Hauck. Göttingen, 241–256. ———. 1993. Sakrala personnamn—eller profana: Klassifikations- och gränsdragnings- problem i det gamla nordiska personnamnsförrådet. In Personnamn i nordiska och andra germanska fornspråk. NORNA-rapporter 51. (Ed.) L. Peterson. Uppsala, 39–60. ———. 1999a. Hundare och det germanska hundratalet. Namn och bygd (87), 5–12. ———. 1999b. Helgö. §1. In Reallexikon der Germanischen Altertumskunde. Band. 14. (Eds.) H. Beck et al. Berlin – New York, 285–286. ———. 2000. Hundare. In Reallexikon der Germanischen Altertumskunde. Band 15. (Eds.) H. Beck et al. Berlin – New York, 233–238. ———. 2004. Svethiudh, det svenska rikets kärna. Namn och bygd (92), 5–18. ———. 2005. Svethiudh arkeologiskt förankrat. Namn och bygd (93), 13–15. 558 Bibliography

Andersson, T. M. 2008. From Tradition to Literature in the Sagas”. In Oral Art Forms and their Passage into Writing. (Eds.) E. Mundal & J. Wellendorf. Copenhagen, 7–17. Andrén, A. 1991. Guld och makt—en tolkning av de skandiniaviska guldbrakteaternas funktion. In Samfundsorganisation og Regional Variation—Norden i romersk jernalder og folkevandringstid. Beretning fra 1. nordiske jernaldersymposium på Sandbjerg Slot 11–15 april 1989. (Eds.) C. Fabech, & J. Ringtved. Jysk Arkæologisk Selskabs Skrifter XXVII. Århus, 245– 256. ———. 2002. Platsernas betydelse. Norrön ritual och kultplatskontinuitet. In Plats och praxis: Studier av nordisk förkristen ritual. Vägar till Midgård 2. (Eds.) K. Jennbert, A. Andrén & C. Raudvere. Lund, 299–342. ———. 2004. I skuggan av Yggdrasil: Trädet mellan idé realitet i nordisk tradition. In Ordning mot kaos—studier av nordisk förkristen kosmologi. Vägar till Midgård 4. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 389–430. ———. 2006. Warrior Culture, Hybridity, and Old Norse Cosmology. In Old Norse Religion in Long- Term Perspectives: Origins, Changes, and Interactions. Vägar till Midgård 8. (Eds.) Andrén, A. & K. Jennbert & C. Raudvere. Lund, 33–38. ———. 2014. Tracing Old Norse Cosmology. The world tree, middle earth, and the sun in archaeological perspectives. Vägar till Midgård 16. Nordic Acadeic Press, Lund. Andrén, A., K. Jennbert & C. Raudvere. 2006. Old Norse Religion. In Old Norse Religion in Long-Term Perspectives: Origins, Changes, and Interactions. Vägar till Midgård 8. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 11–14. Anjou, S. 1935. Uppsalatemplets gudabeläten avbildade på bonaden från Skog. Fornvännen (30), 257–264. Antonsen, E. H. 2002. Runes and Germanic Linguistics. Trends in Linguistics. Studies and Monographs 149. Berlin/New York. Åqvist, C. 1996. Hall och harg—det rituella rummet. In Religion från stenålder till medeltid. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter 19. (Eds.) K. Engdahl & A. Kaliff. Linköping, 105–120. Arne, T. J. 1932. Vendel före vendeltiden. Fornvännen 27, 1–22. Arrhenius, B. 1961. Vikingatida miniatyrer. Tor (7), 139–164. ———. 1990. Utgrävningen av den östligaste storhögen på gravfältet Ormknös, raä 111, Björkö,Adelsö sn, Uppland. Laborativ Arkeologi (4). Stockholm, 65–80. ———. 1995. 65–80. Regalia in Svealand in Early Medieval Times. Tor (27), 311–335. ———. 1998. Bebyggelsestrukturen i det förhistoriska Vendel och järnålderns mång- mannagårdar. In Centrala platser—Centrala frågor: Samhällsstrukturen under Järnåldern. En Vänbok till Berta Stjernquist. Uppåkrastudier. 1. (Eds.) L. Larsson & B. Hårdh. Stockholm, 179–187. ———. 2000. Tuna och Husby i Vendel. In En bok om Husbyar. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. Skrifter 33. (Ed.) Michael Olausson. Örebro, 93–100. Bibliography 559

———. 2004a. Ottarshögen och järnålderskolonisationen i Vendel. In Current Issues in Nordic Archeology. Proceedings of the 21st Conference of Nordic Archeologists, 6–9 September 2001, Akureyri Iceland. (Ed.) Garðar Guðmundsson. Reykjavík, 37–41. ———. 2004b. Det forntida Sveariket—myt eller en arkeologisk realitet. Kungl. Vitterhets Historie och Antikvitets Akademien. Årsbok. Stockholm, 203–227. ———. 2011. Helgö—pagan sanctuary complex. In Excavations at Helgö XVIII. (Eds.) B. Arrhenius & U. O’Meadhra. Stockholm, 11–43. Arrhenius, B. & T. Sjøvold 1995. The infant prince from the East Mound at Old Uppsala. Laborativ arkeologi (8). Stockholm, 29–37. Arwidsson, G. 1942. Valsgärde 6. Die Gräberfunde aus Valsgärde I. Uppsala. ———. 1954. Valsgärde 8. Die Gräber von Valsgärde II. Uppsala. ———. 1977. Valsgärde 7. Die Gräberfunde aus Valsgärde III. Uppsala. ———. 1980. Båtgravarna i Valsgärde. Vendeltid. (Ed.) A. Sandwall. Stockholm, 45–64. ———. 1965. Källa”. Kulturhistoriskt lexikon för nordisk medeltid. Vol. 10, cols. 53–57. Auffahrth, C. 2001. Kosmologie. I. Zum Begriff. Religion in Geschichte und Gegenwart: Handwörtbuch für Theologie und Religionswissenschaft. Band 4. (Eds.) Hans Dieter Betz et al. Stuttgart – Berlin – Köln, 1706–1707. Austin, J. L. 1962. How to do Things with Words (The William James Lectures 1955). (Ed.) J. O. Urmson. Cambridge, MA. Axboe, M. 2004. The Bracteates from Helgö. In Excavations at Helgö XVI: Exotic and Sacral Finds from Helgö. Kungl. Vitterhets Historie och Antikvitets Akademien. Stockholm, 35–40. ———. 2007. Brakteatstudier. Copenhagen. ———. 2011. The third gold bracteate from Helgö. In Excavations at Helgö XVIII. (Eds.) B. Arrhenius & U. O’Meadhra. Stockholm, 93–96. Baaren, T. P. van 1972. Vi människor: De skriftlösa folkens religion och kultur. Orig. title Wij mensen: Religie en wereldbeschouwing bij schriftloze volken. Trans. G. Carlstedt. Stockholm. Bäck, M. et al. 2008. Lilla Ullevi—historien om det fridlysta rummet. UV Mitt, rapport 2008:39. Arkeologiska undersökningar. Stockholm. Back-Danielsson, I.-M. 1999. Engendering performance in late Iron Age. Current Swedish Archaeology (7), 7–18. ———. 2002. (Un)masking gender—gold foil (dis)embodiments in Late Iron Age Scandinavia. In Thinking through the body: archaeology and corporeality. (Eds.) Y. Hamilakris et al. New York, 179–199. Backe, M., B. Edgren & F. Herschend. 1993. Bones Thrown into a Water-Hole. In Sources and Resources: Studies in Honour of Brigit Arrhenius. (Eds.) Greta Arwidsson et al. Pact 38. Journal of European Study Group on Physical, Chemical, Biological and Mathematical Techniques Applied to Archaeology, 327–342. 560 Bibliography

Bader, F. 1986. An I. E. Myth of Immersion-Emergence. The Journal of Indo-European Studies (14), 40–123. Baetke, W. 1942. Das Heilige im Germanischen. Tübingen. ———. 1950. Die Götterlehre der Snorra-Edda. Berichte über die Verhandlungen der sächsischen Akademie der Wissenschaften zu Leipzig. Philologisch-historische Klasse. Band 97. Heft 3. Berlin. ———. 1951. Christliches Lehngut in der Sagareligion: Das Svoldr-Problem. Zwei Beiträge zur Sagakritik. Berichte über die Verhandlungen der sächsischen Akademie der Wissenschaften zu Leipzig. Philologisch-historische Klasse. Band 98. Heft 6. Berlin. ———. 1964. Yngvi und die Ynglinger. Eine quellenkritische Untersuchung über das nordische Sakralkönigtum. Sitzungsberichte der sächsischen Akademie der Wissenschaften zu Leipzig. Philologisch-historische Klasse. Band 109. Heft 3. Berlin. Bagge, S. 1994. Snorres bilde av det tidlige norske kongedømmet. In Myte og Ritual i det førkristne Norden. Et symposium. (Eds.) J. P. Schjødt et al. Odense, 9–22. ———. 2005 (1998). Mennesket i middelalderens Norge: Tanker, tro og holdninger 1000– 1300. Oslo. Batey 2009. See Lucas 2009. Baudou, E. 1989. Hög—gård—helgedom i Mellannorrland under den äldre järnåldern. Arkeologi i norr (2), 9–43. Beard, M. 2002. Religions of Rome. Vol. 1. A History. Cambridge. Beck, H. 1964. Einige vendelzeitliche Bilddenkmäler und die literarische Überlieferung. Bayerische Akademie der Wissenschaften. Sitzungsberichte, phil.-hist. Kl., 1964, 6. München. ———. 1965. Das Ebersignum im Germanischen: Ein Beitrag zur germanischen Tier- Symbolik. Quellen und Forschungen zur Sprach- und Kulturgeschichte der germanischen Völker. Neue Folge, 16 (140). Berlin. ———. 1968. Waffentanz und Waffenspiel. Festschrift für Otto Höfler zum 65. Geburtstag. I. Band. Hrsg. von H. Birkhan und O. Gschwantler. Wien, pp. 1–16. ———. 1986. Eid (Sprachliches—Literarisches). In Reallexikon der Germanischen Altertumskunde. Band 6. (Eds.) H. Beck et al. Berlin – New York, 537–539. ———. 2000. Hochsitz. In Reallexikon der Germanischen Altertumskunde. Band 15. (Eds.) H. Beck et al. Berlin – New York, 14–16. ———. 2013. Snorri Sturlusons Mythologie—Euhemerismus oder Analogie? In Snorri Sturluson—Historiker, Dichter, Politiker. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 85. Red. H. Beck et al. Berlin/New York: de Gruyter, 1–22. Bell, C. 1997. Ritual: Perspectives and Dimensions. Oxford – New York. Bellander, E. 1938. Gästriklands järnåldersbebyggelse I. Fornlämningar och fynd. Från Gästriklands kulturhistoriska förenings meddelande. Årsbok 1938. Gävle, 7–174. Bibliography 561

———. 1944. Gästriklands järnåldersbebyggelse II. Bebyggelsehistoria. Från Gästriklands kulturhistoriska förenings meddelande. Årsbok 1944. Gävle, 5–95. Berglund, B. 2003. Hus og gard—den arkeologiske bakgrunnen. In Middelaldergården i Trøndelag. (Ed.) O. Skevik. Verdal, 25–62. Bergsveinn Birgisson, 2007. Inn i skaldens sinn: Kognitive, estetiske og historiske skatter i den norrøne Skaldediktningen. Bergen: Universitetet i Bergen. Beronius-Jörpeland, L. & Sander, B. 1996. Viby gård—livs- och vilorum: Arkeologiska förundersökningar och undersökningar, Uppland, Sollentuna socken, RAÄ 308 och 309. RAÄ UV-Stockholm Rapport 1996:114. Beronius Jörpeland, L., H. Göthberg, A. Seiler and J. Wikborg. 2013. Monumentala stol- prader i Gamla Uppsala. Fornvännen (108): 278–281. Bertell, M. 2003. Tor och den nordiska åskan. Föreställningar kring världsaxeln. Stockholm. Beverfjord, A. & Binns, K. S. 1994. Religion og regime i Brønnøys jernalder. SPOR (9.17), 16–20. Beyschlag, S. 1950. Konungasögur: Untersuchungen zur Königssaga bis Snorri. Die Älteren Übersichtswerke samt Ynglingasaga. Bibliotheca Arnamagnæana. VIII. Kopenhagen. Bianchi, M. 2010. Runor som resurs: Vikingatida skriftkultur i Uppland och Södermanland. Runrön 20. Uppsala. Birkeli, E. 1932. Høgsætet: Det gamle ondvege i religionshistorisk belysning. Stavanger. ———. 1938. Fedrekult i Norge: Et forsøk på en systematisk-deskriptiv fremstilling. Skrifter utgitt av Det Norske Videnskaps-Akademi i Oslo II. Hist.-Filos. Klasse. 1938. No. 5. Oslo. Bjarni Aðalbjarnarson. See Snorri. Heimskringla I–III. Bjarni F. Einarsson. 2008. Blót houses in Viking Age farmsted and cult practices. New findings from south-eastern Iceland. Acta Archaeologica (79), 145–184. Bjarni Guðnason. 1963. Um Skjöldungasögu. Reykjavík. Bjørk Birgisdottir, B. & L. Strøm 2002. Vikingen fra Medalhus. In SPOR 2002/2, 18–20. Björkhager, V. et al. 1997. Fynd. In Kungens gods i Borg: Om utgrävningarna vid Borgs säteri i Östergötland. (Eds.) K. Lindeblad & A.-L. Nielsen. Linköping, 62–102. Bloch, M. 1989. Ritual, History and Power: Selected Papers in Anthropology. London School of Economics. Monographs on Social Anthropology. No. 58. London. Blomkvist, N., Brink, S. & T. Lindkvist 2007. The kingdom of Sweden. In Christianization and the rise of Christian Monarchy. Scandinavia, Central Europe and Rus’ c. 900–1200. (Ed.) Nora Berend. Cambridge, 167–213. Blomkvist, T. 2002. Från ritualiserad tradition till institutionaliserad religion: Strategier för maktlegitimering på Gotland under järnålder och medeltid. Uppsala. ———. 2010. Lokala perspektiv på religion. Människor och makter. 2.0. En introduktion till religionsvetenskapen. Högskolan i Halmstad. http://hh.diva-portal.org. 562 Bibliography

Blomkvist, T. & P. Jackson. 1999. Alt ir baugum bundit: Skaldic Poetry on Gotland in a Pan- Scandinavian and Indo-European Context. Arkiv för nordisk filologi (114), 17– 29. Bø, O. 1960. Gravøl (Norge). In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 5. Malmö, cols. 450–452. ———. 1965. Källa. Noreg. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 10. Malmö, cols. 57–59. Bøe, A. 1964. Konge (Innleiing og Noreg). In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 9. Malmö, cols. 1–4. Bolle, K. W. 1987. Cosmology. An Overview. In Encyclopedia of Religion. Vol. 4. (Ed.) M. Eliade. New York, 100–107. Boor, H. de 1930. Die religiöse Sprache der Völuspá und verwandter Denkmäler. In Deutsche Islandforschung 1930. (Ed.) W. H. Vogt. Breslau, 68–142. Boudriot, W. 1928. Die altgermanische Religion in der amtliches kirchlichen Literatur des Abendlandes vom 5. bis 11. Jahrhundert. Untersuchungen zur allgemeinen Religions geschichte, Heft 2. Bonn. Braarvig, J. 1997. The Function of Religious Rites in Creating and Sustaining the ‘Meaning’ of Symbols: A Commentary on Aristotle’s De interpretatione 1. In In Search of Symbols. (Eds.) J. Braarvig & T. Krogh. Oslo, 44–67. Bratt, P. 2008. Makt uttryckt i jord och sten. Stora högar och maktstrukturer i Mälardalen under järnåldern, Stockholm Studies in Archaeology 46. Stockholm. Braune, W. 1889. Vingolf. Beitrage zur Geschichte der deutschen Sprache und Literatur 14, 369–376. Brendalsmo, J. 2006. Kirkebygg og kirkebyggere: Byggherrer i Trøndelag ca. 1000–1600. Oslo. Brereton, J. P. 1987. Sacred Space. In Encyclopedia of Religion. Vol. 12. (Ed.) M. Eliade. New York, 526–535. Brink, S. 1990a. Sockenbildning och sockennamn: Studier i äldre territoriell indelning i Norden. Acta Academiae regiae Gustavi Adolphi. LVII. Studier till en svensk ortnamnsatlas utgivna av Thorsten Andersson. 14. Uppsala – Stockholm. ———. 1990b. Cult Sites in Northern Sweden. In Old Norse and Finnish Religions and Cultic Place-Names. Scripta Instituti Donneriani Aboensis. XIII. (Ed.) T. Ahlbäck. Stockholm, 458–489. ———. 1992. Har vi haft ett kultiskt *al i Norden? In Sakrale Navne. NORNA-rapporter. 48. (Eds.) G. Fellows-Jensen & B. Holmberg. Uppsala, 107–121. ———. 1994. En vikingatida storbonde i södra Norrland. Tor (26), 145–162. ———. 1996a. Political and Social Structures in Early Scandinavia. A Settlement- Historical Pre-Study of the Central Place. Tor (28), 235–281. ———. 1996b. Forsaringen—Nordens äldsta lagbud. In Beretning fra femtende tvær- faglige vikingsymposium. (Eds.) E. Roesdahl & P. Meulengracht Sørensen. Århus, 27–55. ———. 1997. Political and Social Structures in Early Scandinavia II. Aspects of Space and Territoriality—The Settlement District. Tor (29), 389–437. Bibliography 563

———. 1999a. Social Order in the Early Scandinavian Landscape. In Settlement and Landscape: Proceedings of a Conference in Århus, Denmark, May 4–7 1998. Jutland Archaeological Society. (Eds.) C. Fabech & J. Ringtved. Højbjerg, 423–438. ———. 1999b. Fornskandinavisk religion—förhistoriskt samhälle: En bosättningshis- torisk studie av centralorter i Norden. In Religion och samhälle i det förkristna Norden: Et symposium. (Eds.) U. Drobin et al. Odense, 11–55. ———. 1999c. Nordens husabyar—unga eller gamla. In Et hus med mange rom: Vennebok til Bjørn Myhre på 60-årsdagen. Bind B. (Eds.) I. Fuglestvedt et al. Stavanger., 283–291 ———. 2000a. Husby. In Reallexikon der Germanischen Altertumskunde. Band 15. (Eds.) H. Beck et al. Berlin – New York, 274–278. ———. 2000b. Nordens husabyar—unga eller gamla? In En bok om Husbyar. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. Skrifter 33. (Ed.) Michael Olausson. Örebro, 65–73. ———. 2001. Mythologizing Landscape: Place and Space of Cult and Myth. In Kontinuitäten und Brüche in der Religionsgeschichte: Festschrift für Anders Hultgård zu seinem 65. Geburtstag am 23.12.2001. (Eds.) M. Stausberg, O. Sundqvist & A. van Nahl. Berlin – New York, 76–112. ———. 2003. ambátt, seta, deigja—þræll, þjónn, bryti: Termer för trälar belyser träldo- mens äldre historia. In Trälar: Ofria i agrarsamhället från vikingatid till medeltid. Skrifter om skogs- och lantrukshistoria 17. (Eds.) T. Lindkvist & J. Myrdal. Stockhom, 103–117. ———. 2004. Mytologiska rum och eskatologiska föreställningar i det vikingatida Norden. In Ordning mot kaos—studier av nordisk förkristen kosmologi. Vägar till Midgård 4. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 291–316. ———. 2005. Verba Volant, scripta manent? Aspects of Early Scandinavian Oral Society. In Literacy in Medieval and Early Modern Scandinavian Culture. (Ed.) P. Hermann. Odense, 77–135. ———. 2007. Skiringssal, Kaupang, Tjølling—the Toponymic Evidence. In Kaupang in Skiringssal. Kaupang Excavation Project. Publication Series. Vol. 1. Norske Oldfunn XXII. (Ed.) D. Skre. Oslo, 53–64. ———. 2008. Law and society—polities and legal customs in Viking Scandinavia. In The Viking World. (Ed.) S. Brink. London, 23–31. ———. 2010. Är Forsaringen medeltida? In Hälsingerunor (2010), 109–117. ———. 2011. Gudhem—the toponymic evidence (or rather challenge). In The Gudme/ Gudhem Phenomenon. (Eds.) O. Grimm & A. Pesch. Neumünster, –. Broberg, A. 1991. Religionsskifte och sockenbildning i Norduppland. Kyrka och socken i medeltidens Sverige. Studier till det medeltida Sverige 5. Riksantikvarieämbetet. (Ed.) O. Ferm. Stockholm, 49–79 Brøgger, A. W. 1916. Borrefundet og Vestfold-kongernes graver. Det norske Videns kapsselskapets Skrifter. II. Hist. Fil. Kl. No. 1. Kristiania. 564 Bibliography

Brunstedt, S. 1996. Alsnu Kungsgård. Forskningsprojekt Hovgården. Uppland. Adelsö socken RAÄ 46 m fl. Riksantikvarieämbetet. Arkeologiska undersökningar. UV Stockholm, Rapport 71/1 (2). Bruun, D. & Finnur Jónsson. 1909. Om hove og hovudgravninger på Island. Aarbøger for nordisk Oldkyndighed og Historie. II Række. 24. Bind. Kjøbenhavn, 245–316. Buchholtz, P. 1968. Schamanistische Züge in der altisländischen Überlieferung. Münster. Bugge, S. 1877. Runeskriften paa ringen i Forsa kirke i nordre Helsingland. Festskrift til det Kgl. Universitet i Upsala ved dets Jubilæum i September 1877 fra det Kgl. Fredriks Universitet i Christiania I. Kristiania. ———. 1881–89. Studier over de nordiske Gude- og Heltesagns oprindelse. Christiania. ———. 1894. Bidrag til den Ældste Skaldedigtnings Historie. Christiania. Burkert, W. 1983. Homo Necans: The Anthropology of Ancient Greek Sacrificial Ritual and Myth. Orig. title Homo Necans. Trans. P. Bing. Berkeley & Los Angeles. ———. 1985. Greek Religion. Orig. title Griechische Religion der archaischen und klassischen Epoche. Trans. J. Raffan. Cambridge, Massachusetts. Byock, J. 2001. Viking Age Iceland. London. Byrne, F. J. 1973. Irish Kings and High-Kings. London. Cahen, M. 1921. Études sur le vocabulaire religieux du vieux-Scandinave: La Libation. Collection Linguistique Publiée par la Société de Linguistique de Paris 9. Paris. Calissendorff, K. 1964. Helgö. Namn och Bygd (52), 105–152. ———. 1986. Ortnamn i Uppland. Stockholm. ———. 1991. Helgö: unikt handelscentrum eller vanlig bondgård. Fornvännen (86), 106–110. Capelle, T. 2000. Quell- und Brunnenopfer. In Opferplatz und Heiligtum. Kult der Vorzeit in Norddeutschland. Veröffentlichungen des Helms-Museums. Hamburger Museum für Archäologie und die Geschichte Harburgs. Nr 86. (Eds.) R. Busch, T. Capelle, F. Laux. Neumünster, 101–105. Carlie, A. 2000. Käringsjön—en offerplats för bondebefolkning eller ledarskikt? Studier kring en romartida kultplats och dess omland i södra Halland. Tor (30), 123–164. ———. 2004. Forntida byggnadskult: Tradition och regionalitet i södra Skandinavien. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter Nr. 57. Stockholm. Carlsson, A. 1997. Birkas kungsgård på Adelsö och Svearnas Fornsigtuna—två aristokratiska miljöer i Mälardalen. In “. . . gick Grendel att söka det höga huset . . .”: Arkeologiska källor till aristokratiska miljöer i Skandinavien under yngre järnålder. Rapport från ett seminarium i Falkenberg 16–17 november 1995. Hallands länsmuseers Skriftserie No. 9/GOTARC C. Arkeologiska skrifter. No. 17. (Eds.) J. Callmer & E. Rosengren. Halmstad, 83–88. Carlsson, L. 1965. Jag giver dig min dotter: Trolovning och äktenskap i den svenska kvin- nans äldre historia I. Rätthistoriska Biblioteket 8. Stockholm. Bibliography 565

Carp, R. M. 2014. Material Culture. In The Routledge Handbook of Rearch Methods in the Study of Religion. (Eds.) M. Stausberg & S. Engler. London, 474–490. Carsten, L. 2012. Posts of Power. The Sagas’ Use of the Past Exemplified by an Iron Age Building. In The 15th International Saga Conference: Sagas and the Use of the Past. Aarhus, 97–98. ———. 2014. Lordship, Symbols and Memory: the Great House at Runsa from a Philological Point of View. In Runsa Borg. Representative Life on a Migration Period Hilltop Site—a Scandiavian Perspective. Papers from the project Runsa borg, Uppland. No 2. (Ed.) M. Olausson. Östersund, 145–166. Celander, H. 1955. Förkristen jul: Enligt norröna källor. Göteborgs universitets årsskrift. Vol. LXI. 1955. 3. Stockholm. Chadwick, N. 1950. Þorgerðr Hölgabuðr and the Trolla þing. A note on the sources In The Early Cultures of North-West Europe. Cambridge, 395–417. Christensen, T. 1991. Lejre Beyond Legend—The Archaeological Evidence. Journal of Danish Archaeology (10), 163–185. ———. 1997. Lejrehallen. In “. . . gick Grendel att söka det höga huset . . .”: Arkeologiska källor till aristokratiska miljöer i Skandinavien under yngre järnålder. Rapport från ett seminarium i Falkenberg 16–17 november 1995. Hallands länsmuseers Skriftserie No. 9/GOTARC C. Arkeologiska skrifter. No. 17. (Eds.) J. Callmer & E. Rosengren. Halmstad, 47–54. ———. 2001. Lejre. §2. In Reallexikon der Germanischen Altertumskunde. Band. 18. (Eds.) H. Beck et al. Berlin – New York, 249–254. ———. 2009: Odin fra Lejre. I Årskrift fra Roskilde Museum, 6–25. ———. 2010: Odin fra Lejre. I Danefæ—Skatte fra den danske muld. (Eds.) P. O. Nielsen & M. Andersen. Kopenhagen, 183–186. Clark Hall, J. R. 1916. A Concise Anglo-Saxon Dictionary for the Use of Students. Cambridge. Clarke H. & B. Ambrosiani. 1993 (1991). Vikingastäder. Org. title Towns in the Viking Age. Höganäs. Clauss, M. 1990. Mithras: Kult und Mysterien. München. ———. 1992. Cultores Mithrae: Die Anhängerschaft des Mithras-Kultus. Heidelberger Althistorische Beiträge und Epigraphische Studien 10. Stuttgart. Cleasby R. & Guðbrandur Vigfusson. 1957 (1874). An Icelandic-English Dictionary. Oxford. Clunies Ross, M. 1994. Prolonged Echoes: Old Norse Myths in Medieval Northern Society. Vol. 1. Odense. ———. 1998. Prolonged Echoes: Old Norse Myths in Medieval Northern Society. Vol. 2. Odense. ———. 2006. The measures of Old Norse religion in long-term perspective. In Old Norse Religion in Long-Term Perspectives: Origins, Changes, and Interactions. Vägar till Midgård 8. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 412–416. 566 Bibliography

———. 2008. The creation of Old Norse mythology. In The Viking World. (Ed.) S. Brink. London, 231–234. Cöllen, S. 2011. Der Rätselhafte Gott: Heimdallr im Licht altnordischer Vorstellungen von Ahnen und Ordnung. Uppsala. Damell, D. 1980. Om en fosfatkarta över fornminnesområdet vid Gamla Uppsala. In Inventori in honorem: En vänbok till Folke Hallberg. (Ed.) Å. Hyenstrand. Stockholm, 64–67. ———. 1991. Fornsigtuna: En kungsgårds historia. Upplands Bro. ———. 1993. About Royal Manors from the Late Iron Age in Middle Sweden. Current Swedish Archaeology 1. (Eds.) M. Burström & A. Carlsson. Stockholm, 39–47. Davidson, H. R. E. 1943. The Road to Hel: A Study of the Conception of the Dead in Old Norse Literature. Cambridge. ———. 1975. Scandinavian Cosmology. In Ancient Cosmologies. (Eds.) C. Blacker & M. Loewe. London, 172–197. ———. 1994 (1993). The Lost Beliefs of Northern Europe. London – New York. Dillmann, F.-X. 1997. Kring de rituella gästabuden i fornskandinavisk religion. In Uppsalakulten och Adam av Bremen. (Ed.) A. Hultgård. Nora, 51–73. ———. 2000a. Chronique des études nordiques. Proxima Thulé. Revue d’études nordiques. Vol. 4. (Ed.) F.-X. Dillmann. Paris, 101–112. ———. 2000b. Ad noen geografiske navn i Heimskringla. In Oluf Rygh: Rapport fra symposium på Stiklestad 13.–15. Mai 1999. NORNA-rapporter. 70B. (Eds.) B. Sandnes et al. Uppsala, 43–54. ———. 2000c. See Snorri Sturluson in Sources. ———. 2006. Les magiciens dans l’Islande ancienne: Études sur la représentation de la magie islandaise et de ses agents dans les sources littéraires norroises, Acta Academiae regiae Gustavi Adolphi. XCII. Uppsala. ———. 2007. Valhǫll. In Reallexikon der Germanischen Altertumskunde. Band 35. (Eds.) H. Beck et al. Berlin – New York, 341–366. ———. 2008. Omkring Halvdan Svartes gravferd. Høvdingen: Om Snorre Sturlasons liv og virke. (Eds.) J. O. Askedal & K. J. Myrvoll. Oslo, 73–87. DMS. Det medeltida Sverige. 1960–. Stockholm. Douglas, M. 1966. Purity and Danger: An Analysis of Concepts of Pollution and Taboo. London. Drescher, H. & K. Hauck. 1982. Götterthrone des heidnischen Nordens. Frühmittel alterliche Studien. Jahrbuch des Instituts für Frühmittelalterforschung der Uni versität Münster. 1982. Band 16, 237–301. Driscoll, M. J. 2008. See Ágrip above. Drobin, U. 1991. Mjödet och offersymboliken i fornnordisk religion. In Studier i religionshistoria tillägnade Åke Hultkrantz professor emeritus den 1 juli 1986. (Eds.) L. Bäckman et al. Löberöd, 97–141. Bibliography 567

Drobin, U. & M.-L. Keinänen. 2001. Frey, Veralden olmai och Sampo. In Kontinuitäten und Brüche in der Religionsgeschichte: Festschrift für Anders Hultgård zu seinem 65. Geburtstag am 23.12.2001. (Eds.) M. Stausberg, O. Sundqvist & A. van Nahl. Berlin – New York, 136–169. Dronke, U. 1997. See The Poetic Edda above. DuBois, T. A. 1999. Nordic Religions in the Viking Age. Philadelphia. Duczko, W. 1993a. Bibliografi över Gamla Uppsala. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. OPIA. 7. Societas Archaeologica Upsaliensis. (Ed.) W. Duczko. Uppsala, 71–78. ———. 1993b. Introduktion till Gamla Uppsala. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. OPIA. 7. Societas Archaeologica Upsaliensis. (Ed.) W. Duczko. Uppsala, 9–37. ———. 1996a. Gamla Uppsala. In Reallexikon der Germanischen Altertumskunde. Band 10. (Eds.) H. Beck et al. Berlin – New York, 409–418. ———. 1996b. De litterära och arkeologiska kungarna. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. Vol. II. OPIA. 11. Department of Archaeology. Uppsala University. (Eds.) W. Duczko. Uppsala, 9–19. ———. 1996c. Kungsgården. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. Vol. II. OPIA. 11. Department of Archaeology. Uppsala University. (Ed.) W. Duczko. Uppsala, 37–51. ———. 1996d. Uppsalahögarna som symboler och arkologiska källor. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. Vol. II. OPIA. 11. Department of Archaeology. Uppsala University. (Ed.) W. Duczko. Uppsala, 59–94. ———. 1997. Gamla Uppsala—svearnas maktcentrum i äldre och nyare forskning. In “. . . gick Grendel att söka det höga huset . . .”: Arkeologiska källor till aristokratiska miljöer i Skandinavien under yngre järnålder. Rapport från ett seminarium i Falkenberg 16–17 november 1995. Hallands länsmuseers Skriftserie No. 9/GOTARC C. Arkeologiska skrifter. No. 17. (Eds.) J. Callmer & E. Rosengren. Halmstad, 71–81. Dumézil, G. 1958. La Rígsþula et la structure sociale indo-européenne. Revue de l’histoire des religions (154), 1–9. ———. 1959. Les dieux des Germains. Essai sur la formation de la religion scandinave. Paris. ———. 1973. Gods of the Ancient Northmen. Berkeley. ———. 1986 (1948). Loki. Paris. ———. 2000. Mythes et dieux de la Scandinavie ancienne. Édition établie et préfacée par F.-X. Dillmann. Paris. Dumville, D. N. 1979 (1977). Kingship, Genealogies and Regnal Lists. In Early Medieval Kingship. The School of History. University of Leeds. (Eds.) P. H. Sawyer & I. N. Wood.Leeds, 72–104. 568 Bibliography

Duncan, A. A. M. 1975. Scotland. The Making of Kingdom. Edinburgh. Düwel, K. 1970. Germanische Opfer und Opferriten im Spiegel altgermanischer Kultworte. In Vorgeschichtliche Heiligtümer und Opferplätze in Mittel- und Nordeuropa. Bericht über ein Symposium in Reinhausen bei Göttingen in der Zeit vom 14. bis 16. Oktober 1968. Abhandlungen der Akademie der Wissenschaften in Göttingen, Philologisch-historische Klasse 3, dritte Folge 74. Göttingen, 219–239. ———. 1985. Das Opferfest von Lade: Quellenkritische Untersuchungen zur germanischen Religionsgeschichte. Wiener Arbeiten zur germanischen Altertumskunde und Philologie 27. Wien. ———. 1992. Runeninschriften als Quellen der germanischen Religionsgeschichte. In Germanische Religionsgeschichte: Quellen und Quellenprobleme. (Eds.) H. Beck et al. Berlin – New York, 336–364. ———. 2000. Jarl. §2. Runologisches. Reallexikon der Germanischen Altertumskunde. Band 16. (Eds.) H. Beck et al. Berlin – New York, 33–34. ———. 2008 (1968). Runenkunde. 4. Auflage. Weimar. Döring, A. 2003. Quellen. In Reallexikon der Germanischen Altertumskunde. Band 24. (Eds.) H. Beck et al. Berlin – New York, 11–15. Ebel, E. 1998. Gode, Godentum. In Reallexikon der Germanischen Altertumskunde. Band 12. (Eds.) H. Beck et al. Berlin – New York, 260–263. ———. 2000. Jarl. Reallexikon der Germanischen Altertumskunde. Band 16. (Eds.) H. Beck et al. Berlin – New York, 29–33. Edsman, C.-M. 1944. Arbor inversa. Religion och Bibel 3, 5–33. Eidlitz Kuoljok, K. 1999. Moder jord och andra mödrar: Föreställningar om verkligheten bland folken i norr och vår syn på den. Stockholm. Ejerfeldt, L. 1971. Helighet, ‘karisma’ och kungadöme i forngermansk religion. Skrifter utgivna av Religionshistoriska Institutionen (Hum. Fak.). Genom C.-M. Edsman. Uppsala. Ekelund, G. 1956. Silverskatten från Eketorp. Från Bergslag och bondebygd (11). Örebro, 143–175. Ekman, T. 2000. Item Husaby in Niericia. In En bok om Husbyar. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. Skrifter 33. (Ed.) Michael Olausson. Örebro, 9–38. Ekroth, G. 2002. The Sacrificial Rituals of Greek Hero-Cults in the Archaic to the Early Hellenistic Period. Liège. ———. 2005. “Blood on the Altars? On the Treatment of Blood at Greek Sacrifices and the Iconographic Evidence.” In Antike Kunst (48), 9–29. Elgqvist, E. 1944. Skälv och Skilfingar: Vad nordiska ortnamn vittna om svenska expan- sionssträvanden omkring mitten av första årtusendet e.Kr. Lund. Eliade, M. 1972. Zalmoxis. The Vanishing God. Transl by W. R. Trask. Chicago. ———. 1974 (1949). Patterns in Comparative Religion. Orig. title Traité d’historie des réligions. Trans. R. Sheed. New York – Scarborough. Ontario. Bibliography 569

———. 1975. Rites and Symbols of Initiation. New York. ———. 1987 (1957). The Sacred and the Profane: The Nature of Religion. Orig. title Das Heilige und das Profane. Trans. W. R. Trask. San Diego – New York – London. ———. 1991 (1949). The Myth of Eternal Return: Or, Cosmos and History. Orig. title Le Mythe de l’éternel retour: Archétypes et répétition. Trans. W. R. Trask. Princeton. Ellis, H. R. See Davidson, H. R. E. Elmevik, L. 1966. Fsv. *lytir (*Lytir): Ett etymologiskt och religionshistoriskt bidrag. Namn och Bygd (54), 47–61. ———. 1986. Fornisl. við taur, á austanverðum Taurinum och det svenska ortnamnet Södertörn. In Festschrift für Oskar Bandle. Zum 60. Geburtstag am 11. Januar 1986. Beiträge zur nordischen Philologie. Band 15. (Ed.) von H.-P. Naumann. Basel – Frankfurt am Main, 11–17. ———. 1990. Aschw. Lytis- in Ortsnamen: Ein kultisches Element oder ein profanes? In Old Norse and Finnish Religions and Cultic Place-Names. Scripta instituti Donneriani Aboensis. XIII. (Ed.) T. Ahlbäck. Stockholm, 490–507. ———. 1995. Härnevi och Friggeråker. Namn och bygd (83), 67–77. ———. 2003. En svensk ortnamnsgrupp och en hednisk prästtitel. Ortnamnssällskapets i Uppsala Årsskrift. Uppsala, 68–78. ———. 2004. Till diskussionen om ett sakralt *al i nordiska ortnamn. In Namenwelten. Orts- und Personennamen in historischer Sicht. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 44. Red. A. van Nahl et al. Berlin – New York, 38–56. ———. 2005. Namnet Disevid—sakralt men inte kultiskt? In Text i arbete: Festskrift till Britt-Louise Gunnarsson den 12 januari 2005. (Ed.) Bäcklund. Uppsala, 54–60. Engelke, M. 2012. Material Religion. In The Cambridge Companion to Religious Studies. (Ed.) R. A. Orsi. Cambridge, 209–229. Enright, M. J. 1988. Lady with a Mead Cup: Ritual, Group Cohesion and Hierarchy in Germanic Warband. Frühmittelalterliche Studien (22), 170–203. ———. 1996. Lady with a Mead Cup: Ritual, Prophecy and Lordship in the European Warband from La Tène to the Viking Age. Dublin. Evans, D. A. H. 1981. King Agni: Myth, History or Legend? In Specvlvm norroenvm: Norse Studies in Memory of Gabriel Turville-Petre. (Eds.) U. Dronke et al. Odense, 89–105. Fabech, C. 1991. Samfundsorganisation, religiøse ceremonier og regional variation. In Samfundsorganisation og Regional Variation. Norden i romersk jernalder og folkevandringstid. Beretning fra 1. nordiske jernaldersymposium på Sandbjerg Slot 11–15 april 1989. Jysk Arkæologisk Selskabs Skrifter. XXVII. (Eds.) C. Fabech & J. Ringtved. Aarhus, 283–304. ———. 1994. Reading Society from the Cultural Landscape: South Scandinavia between Sacral and Political Power. In The Archaeology of Gudme and Lundeborg. (Eds.) P. O. Nielsen, K. Randsborg & H. Thrane. København, 169–183. 570 Bibliography

Falk, H. 1911–16. Hochsitz. In Reallexikon der Germanischen Altertumskunde. Band II. (Ed.) J. Hoops. Berlin, 538–539. ———. 1924. Odensheite. Skrifter utgivna av Videnskapsselskapet i Oslo, II. Historisk- filosofisk klasse. No. 10. Kristiania. Farbregd, O. 1980. Gravfeltt på Vang. Bøgda vår 1980. Oppdal. ———. 1986. Hove i Åsen. SPOR—2–1986. Trondheim, 42–46. Faulkes, A. 1978–79. Descent from the Gods. Mediaeval Scandinavia (11), 92–125. Fidjestøl, B. 1991. Skaldediktinga og trusskiftet. Med tanker om litterær form som historisk kjelde. In Nordisk hedendom: Et symposium, (Eds.) G. Steinsland et al. Odense, 113–131. ———. 1994. Anmeldelse—Claus Krag 1991. Maal og Minne (3–4), 191–199. Finnur Jónsson. 1890. Vingolf. Arkiv for nordisk filologi (5), 280–284. ———. 1898. Hofalýsingar í fornsögum og goðalíkneski. Árbók hins íslenzka fornleifa- félags. Reykjavík, 28–38. ———. 1931. Lexicon Poeticum Antiquæ Linguæ Septentrionalis: Ordbog over det norsk- islandske Skjaldesprog. Oprindelig forfattet af Sveinbjörn Egilsson. København. Finsen, V. 1888. Om den oprindelige Ordning of nogle af den islandske Fristats Institutioner. Videnskabernes Selskabs Skrifter. 6. Række, historisk og philosopisk Afd. Bd. II,1. Köbenhavn. Fischer, S. 2005. Review Kaliff, A. & O. Sundqvist. 2004. Oden och Mithraskulten: Religiös ackulturation under romersk järnålder och folkvandringstid. OPIA. 35. Institutionen för arkeologi. Uppsala universitet. Fornvännen (100), 68–71. Fleck, J. 1970. Konr – Óttarr – Geirroðr: A Knowledge Criterion for Succession to the Germanic Sacred Kingship. Scandinavian Studies (42), 39–49. Flood, G. 1999. Beyond Phenomenology: Rethinking the Study of Religion. London – New York. Foote, P. 1993. Icelandic Historians and the Swedish Image; Comments on Snorri and his Precursors. In Snorre Sturlasson och de isländska källorna till Sveriges historia. Sällskapet Runica et mediævalia. (Ed.) G. Dahlbäck. Stockholm, 9–42. Foote, P. & D. M. Wilson. 1980 (1970). The Viking Achievement: The Society and Culture of Early Medieval Scandinavia. London. Frank, R. 1978. Old Norse Court Poetry. The Dróttkvætt Stanza. Islandica. XLII. Ithaca – London. ———. 2007. The Lady of the Land in Skaldic Praise Poetry. In Myth in Early Northwest Europe. (Ed.) S. O. Glosecki. Tempe, Arizona, 175–196. Franzén, G. 1982. Ortnamn i Östergötland. Stockholm. Frazer, J. G. 1890. The Golden Bough: A Study in Comparative Religion. Vols. I–II. London. Freeden, M. 1996. Ideologies and Political Theories. A Conceptual Approach. Oxford. ———. 2003. Ideology. A Very Short Introduction. Oxford. Frense, B. 1982. Religion och rätt: En studie till belysning av relationen religion-rätt i förkristen nordisk kultur. Lund. Bibliography 571

Friðriksson, A. & O. Vésteinsson. 1997. Hofstaðir Revisited. Norwegian Archaeological Review (30), 105–112. Friesen, O. von. 1914. Gursten: En runsten och ett gårdsnamn. Namn och bygd (2), 1–17. ———. 1932–34. Har det nordiska kungadömet sakralt ursprung? En ordhistorisk utredning. Saga och Sed. Gustav Adolfs Akademiens Årsbok. Uppsala, 15–34. ———. 1933. De svenska runinskrifterna. In Runorna. Nordisk Kultur. 6. Stockholm – Oslo – København, 145–248. Fritz, V. 2002. Tempel Alter Orient und Altes Testament. In Theologische Realen zyklopädie, 46–54. Fritzner, J. 1954 (1883–96). Ordbog over det gamle norske sprog. Vols. 1–3. Omarb., forøget og forbedret. Oslo. ———. 1972. Ordbog over det gamle norske sprog. Vol. 4. Rettelser og tillegg ved Finn Hødnebø. Oslo – Bergen – Tromsø. Frölich, A. 2011. Helgö: A place for surgical treatment or sacrifice—or both. In Excava tions at Helgö XVIII. (Eds.) B. Arrhenius & U. O’Meadhra. Stockholm, 161–184. Frölund, P. 2007. Gamla Uppsala under äldre järnålder. In Land och samhälle i föränd- ring. (Eds.) E. Hjärthner-Holdar et al. Uppsala, 361–378. Frölund, P. & H. Göthberg. 2013. Gamla Uppsala före formationen—Ett bidrag till cebtralplatsens genes. In Gamla Uppsala i ny belysning. (Eds.) O. Sundqvist and P. Vikstrand with J. Ljungkvist. Religionsvetenskapliga studier från Gävle 9. Gävle, 9‒32. Gade, K. E. 1985. Skjalf. Arkiv för nordisk filologi (100), 59–71. Gahrn, L. 1988. Sveariket i källor och historieskrivning. Meddelanden från Historiska Institutionen vid Göteborgs Universitet, nr 36. Göteborg. ———. 2000. Review: H. Janson, Templum Nobilissimum. Adam av Bremen, Uppsala templet och konfliktlinjerna i Europa kring år 1075. Göteborg, Historiska institutionen, Göteborgs universitet, 1998. Fornvännen (95), 128–131. Gansum, T. 1995. Haugar og haugating: Rikssamlingen og det eldste Tønsberg. Tønsberg. ———. 2002. Fra jord til handling. In Plats och praxis: Studier av nordisk förkristen ritual. Vägar till Midgård 2. (Eds.) K. Jennbert, A. Andrén & C. Raudvere. Lund, 249–286. ———. 2004. Hauger som konstruksjoner—arkeologiske forventninger gjennom 200 år, Gotarc Serie B. Gothenburg Archaeological Thesis No. 33. Göteborg. ———. 2013. Haugar & Haugating. Haugar. Geertz, C. 1980. Negara: The Theatre State in Nineteenth-Century Bali. New York. ———. 1993 (1973). The Interpretation of Cultures. New York. Gehl, W. 1939. Der germanische Schicksalsglaube. Berlin. ———. 1941. Das Problem des germanischen Tempels. Zeitschrift für deutsches Altertum (78), 37–49. Gelting, M. 2000. Odelsrett—lovbydelse—bördsrätt—retrait lignager: Kindred and Land in the Nordic Countries in the Twelfth and Thirteenth Centuries. In Family, 572 Bibliography

Marriage and Property Devolution in the Middle Ages. (Ed.) L. I. Hansen. Tromsø 2000, 133–165. van Gennep. A. 1960. The Rites of Passsage. Orig. title Les rites de passage. Trans. M. B. Vizedom & G. L. Caffee. Chicago. Georges, K. E. 1951 (1913). Ausführliches Lateinisch-Deutsches Handwörterbuch. I–II. 9. Auflage. Nachdruck der achten verbesserten und vermehrten Auflage von H. Georges. Basel. Gilhus, I. S. & L. Mikaelsson. 2003. Nya perspektiv på religion. Orig. title Nytt blikk på religion: Studiet av religion idag. Trans. U.-S. Rask. Stockholm. Gilhus, I. S. & E. Thomassen 2010. Oldtidens religioner. Midtøstens og middelhavsområ- dets religioner. Oslo. Girard, R. 1977. Violence and the Sacred. Trans. P. Gregory. Baltimore. ———. 1986. The Scapegoat. Trans. Y. Freccero. Baltimore. Gisli Sigurðsson. 2004. The Medieval Icelandic Saga and Oral Tradition. Cambridge, MA. Golzio, K.-H. 2002. Tempel. Religionsgeschichtliche. In Theologische Realenzyklopädie, 42–46. Göthberg, H. 2008. Bosättning och kyrkogård vid Gamla Uppsala kyrka. Uppsala. Göthberg H., C. Lovén & G. Dahlbäck. 2010. Domkyrkan i Gamla Uppsala. In Uppsala Domkyrka. II. Domkyrkan i Gamla Uppsala. Nuvarande domkyrkans omgivningar. Sveriges kyrkor 228. (Eds.) R. Carlsson et al. Uppsala, 11–64. Götherström, A. 2002. The Value of Stallions and Mares during the Medieval Time in Upper Class Svealand: Molecular Sex Identifications on Horse Remains from Vendel and Eketorp. Journal of Nordic Archaeological Science 2002, 75–78. Graber, G. 1919. Der Einritt des Herzogs von Kärnten am Fürstenstein zu Karnburg: Eine Studie zur kärntischen Volkskunde. Akad. d. Wiss. Wien, Phil-hist. Kl., Sitz.-ber. 190. Band 95. Wien. Gräslund, A.-S. 1983–84. Tor eller Vite Krist? Några reflektioner med anledning av Lugnås-hammaren. Västergötlands fornminnesförenings tidskrift, 229–235. ———. 1985. Den tidiga missionen i arkeologisk belysning—problem och synpunkter. Tor (20), 291–313. ———. 1987. Runstenar, bygd och gravar. Tor (21), 241–262. ———. 1992. Kultkontinuitet—myt eller verklighet? Om arkeologins möjligheter att belysa problemet. In Kontinuitet i kult och tro från vikingatid till medeltid. Projektet Sveriges Kristnande. Publikationer 1. (Ed.) B. Nilsson. Uppsala, 129–150. ———. 1997. Adams Uppsala—och arkeologins. In Uppsalakulten och Adam av Bremen. (Ed.) A. Hultgård. Nora, 101–115. ———. 1999. Is there any Evidence of Powerful Women in Late Iron Age Svealand? In Völker an Nord- und Ostsee und die Franken. Akten des 48. Sachsensymposiums in Mannheim vom 7. bis 1. September 1997. Kolloquien zur Vor- und Frühgeschichte. Band 3. Mannheimer Geschichtsblätter, Neue Folge, Beiheft 2. (Eds.) U. von Freeden et al. Bonn, 91–98. Bibliography 573

———. 2001. Ideologi och mentalitet: Om religionsskiftet i Skandinavien från en arkeologisk horisont. OPIA. 29. Department of Archaeology and Ancient History. Uppsala University. Uppsala. ———. 2006. Prestige Graves—for men, women and children: Mortuary behaviour among the élites of the Mälar valley in the second half of the first millennium AD, with some remarks from a gender perspective. In Herrschaft—Tod—Bestattung: Zu den vor- und frühgeschichtlichen Prunkgräber als archäologisch-historische Quelle. Internationale Fachkonferenz Kiel 16–19. Oktober 2003. Universitätsforschungen zur prähistorischen Archäologie Band 139. Aus dem Institut für Ur- und Frühgeschichte der Universität Kiel. (Eds.) C. von Carnap-Bornheim, D. Krausse & A. Wesse. Bonn, 161–168. ———. 2008. Some powerful women in Valsgärde. In Valsgärde studies: The place and its people, past and present, OPIA 42. Uppsala, 65–82. ———. 2013. Kristna inslag i Gamla Uppsala och dess närområde. In Gamla Uppsala i ny belysning. (Eds.) O. Sundqvist & P. Vikstrand with J. Ljungkvist. Religion svetenskapliga studier från Gävle 9. Gävle, 113–134. Gräslund, B. 1993. Folkvandringstidens Uppsala: Namn, myter, arkeologi och historia. Kärnhuset i riksäpplet. Upplands fornminnesförening och hembygdsförbunds års- bok. Uppland 1993. (Eds.) K. Blent et al. Uppsala, 173–208. ———. 2004. Äldre ortnamn och forntida bebyggelse. Ortnamnssällskapet i Uppsala årsskrift (2004), 11–38. ———. 2015. MS. Beowulf. En historisk och arkeologisk undersökning. Uppsala. Gräslund, B. & N. Price. 2012. Twilight of the gods? The ‘dust veil event’ of AD 536 in critical perspective. Antiquity (86), 428–443. Grieg, S. 1928. Kongsgaarden. Osebergfundet. Bind II. Oslo. Grimes, R. L. 2006. Performance. In Theorizing Rituals. Issues, Topics, Approaches, Concets. (Eds.) J. Kreinath, J. Snoek and M. Stausberg. Leiden, 379–394. Grimm, J. 1835. Deutsche Mythologie. Göttingen. Grimm, O. 2011. Godøy sites on the west coast of Norway—gods’ islands, but ‘no home of the gods’. In The Gudme/Gudhem Phenomenon. (Eds.) O. Grimm & A. Pesch. Neumünster, 175–188. Grønbech, V. 1997 (1909–12). Kultur und Religion der Germanen. Orig. title Vor Folkeæt i Oldtiden. Trans. E. Hoffmeyer. Darmstadt. Grønnesby, G. 1999. Fortidens hus på Kvenild. SPOR, 1999–1, 10–11. ———. 2000. Langhus fra eldre jernalder på Hovde. Årbok for Fosen 2000, 41–52. Grønvik, O. 1981. Runene på Tunesteinen: Alfabet, Språkform, Budskap. Oslo – Bergen – Tromsø. ———. 1982. The words for ‘heir’, ‘inheritance’ and ‘funeral feast’ in early Germanic: An etymological study of ON arfr m, arfi m, erfi n, erfa vb and the corresponding words in the other Old Germanic dialects. Det Norske Videnskaps-Akademi. II. Hist.-Filos. Klasse Avhandlinger. Ny serie No. 18. Oslo. 574 Bibliography

———. 1983. Runeinnskriften på Rök-steinen. Maal og Minne (1983:3–4), 101–149. ———. 1987. Fra Ågedal til Setre: Sentrale runeinnskrifter fra det 6. århundre. Oslo – Bergen – Stavanger – Tromsø. ———. 1989. Kormaks vers om Sigurd jarls blot. In Festskrift til Finn Hødnebø 29. desember 1989. (Eds.) B. Eithun et al. Oslo, 82–90. ———. 1990. To viktige ord i Rök-innskriften: norr. gjalda vb og minni n. Arkiv för Nordisk Filologi (105), 1–40. ———. 1996. Fra Vimose til Ødemotland: Nye studier over runeinnskrifter fra førkristen tid i Norden. Oslo. ———. 1999. Håvamål: Studier over verkets formelle oppbyggning og dets religiøse innhold. Det Norske Videnskaps-Akademi. II Hist.-Filos. Klasse. Skrifter. Ny Serie No. 21. Oslo. Gudmundsson, V. 1894. Den islandske Bolig i Fristats-Tiden. København. Gunnell, T. 1995. The Origins of Drama in Scandinavia. Camebridge. ———. 2001. Hof, Halls, Goðar and Dwarves: An Examination of Ritual Space in the Pagan Icelandic Hall. Cosmos 17, 3–36. ———. 2007. How Elvish were the Álfar? In Constructing Nations, Reconstructing Myth: Essays in Honour of T. A. Shippey, (Ed.) A. Wawn with G. Johnson and J. Walter. Making the Middle Ages, 9. Turnhout, 111–30. Gurevich, A. Y. 1969. Space and Time in the Weltmodell of the Old Scandinavian Peoples. Mediaeval Scandinavia (2), 42–53. Gurevich, A. J. 1992. Historical Anthropology of the Middle Ages. (Ed.) J. Howlett. Oxford. Gustavson, H. 1986. Runstenarnas Uppsala. In Från Östra Aros till Uppsala: En samling uppsatser kring det medeltida Uppsala. Uppsala stads historia. VII. (Eds.) N. Cnattingius &. T. Nevéus. Uppsala, 10–34. ———. 2003. Oklundainskriften sjuttio år efteråt. In Runica—Germanica— Mediaevalia. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 37. (Eds.) W. Heizmann & A. van Nahl. Berlin – New York, 186–198. ———. 2004. Frågor kring en berömd runring. In Guldskatter: Rumänien under 7000 år. (Eds.) K. Slej et al. Stockholm, 86–89. ———. et al. 1992. Runfynd 1989 och 1990. Fornvännen (87), 164–174. Habbe, P. 2005. Att se och tänka med ritual: Kontrakterande ritualer i de isländska släkt- sagorna. Vägar till Midgård 7. Lund. Habel, E. 1959. Mittellateinisches Glossar. Paderborn. Hafström, G. 1959. Forsaringen. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 4. Malmö, cols. 530–534. ———. 1970 (1959). Land och lag. Stockholm. Hagberg, U. 1967–77. The Archaeology of Skedemosse I.–IV. The Royal Swedish Academy of Letters. History and Antiquities. Stockholm. Hallan, N. 1966. Kulisteinen og kristenrettsvedtaket på Mostertinget. In Du mitt Nordmøre, 21–28. Bibliography 575

———. 1981 (1954). Det eldste kronogodset i Trøndelag. In Bosetning og økonmi. Norske historikere i utvalg 5. Nye middelalderstudier. (Eds.) J. Sandnes & C. Krag. Oslo. Hållans Stenholm, A.-M. 2010. Lilla Ullevi—en kultplats. In Makt, kult och plats. Rapport från arkeologiska seminarier vid Stockholms läns museum. (Eds.) P. Bratt & R. Grönwall. Stockholm, 31–38. ———. 2012. Fornminnen: det förflutnas roll i det förkristna och kristna Mälardalen. Vägar till Midgård 15. Lund. Hallberg, P. 1966. Medeltidslatin och sagaprosa: Några kommentarer till Lars Lönnroths studier i den isländska sagalitteraturen. Arkiv för nordisk filologi (81), 258–276. ———. 1973. The Concept of gipta—gæfa—hamingja in Old Norse Literature. In Proceedings of the First International Saga Conference 1971. (Ed.) P. Foote. London, 143–183. ———. 1979. Den isländska sagan. Verdandis skriftserie. 6. Lund. Hallencreutz, C. F. 1984. Adam, Sverige och trosskiftet: Det missionsvetenskapliga per- spektivet. In Adam av Bremen: Historien om Hamburgstiftet och dess biskopar. Trans. E. Svenberg. Comm. by C. F. Hallencreutz, K. Johannesson, T. Nyberg, A. Piltz. (Ed.) B. Sahlin. Stockholm, 355–378. ———. 1993. När Sverige blev europeiskt: Till frågan om Sveriges kristnande. Stockholm. ———. 1997. Missionsstrategi och religionstolkning: Till frågan om Adam av Bremen och Uppsalatemplet. In Uppsalakulten och Adam av Bremen. (Ed.) A. Hultgård. Nora, 117–130. Hallinder, P. 1970. Iron. Excavation at Helgö III. (Ed.) W. Holmqvist. Stockholm. Hansson, A.-M. 2011. Bread offerings at Foundation IV, Helgö—further influence from the classical world? In Excavations at Helgö XVIII. (Eds.) B. Arrhenius & U. O’Meadhra. Stockholm, 53–77. Haraldsen, T. H. 1998. Gullgubben fra Vingom—på Frigg og Odins Hov? In Fåberg og Lillehammer: En lokalskrift utgitt av Fåberg Historielag. Fåberg, 136–146. Hårdh, B. 1996. Silver in the Viking Age: A regional-economic study. Lund. Häringe, K. 1991. Gödåker: Ett uppländskt gravfält från romersk järnålder. Uppsats C/20p. Institutionen för arkeologi. Uppsala universitet. Uppsala. Hasenfratz, H.-P. 1997 (1992). Die religiöse Welt der Germanen: Ritual, Magie, Kult, Mythus. Freiburg – Basel – Wien. Hastrup, K. 1985. Culture and History in Medieval Iceland: An Anthropological Analysiss of Structure and Change. Oxford. Hauck, K. 1950. Geblütsheiligkeit. In Liber Floridus Mittellateinische Studien, Paul Lehmann zum 65. Geburtstag. (Eds.) B. Bischoff & S. Brechter. St. Ottilien, 187–240. ———. 1954. Halsring und Ahnenstab als herrscherliche Würdezeichen. In Herrschaftszeichen und Staatssymbolik I. (Ed.) P. E. Schramm. Stuttgart, 145–212. ———. 1985–86. Die Goldbrakteaten der Völkerwanderungszeit. Mit Beiträgen von M. Axboe, K. Düwel, L. von Padberg, U. Smyra und C. Wypior. Münstersche Mittelalterschriften. Vols. 1–3. München. 576 Bibliography

———.1999. Goldblechfigürchen. In Reallexikon der Germanischen Altertumskunde. Band 14. (Eds.) H. Beck et al. Berlin – New York, 318–323. ———. 2002. Neues aus Uppåkra bei Lund, Südschweden: Zur Ikonographie der Goldbrakteaten, LXIV. In Central Places in the Migration and the Merovingian Periods: Papers from the 52nd Sachsensymposium Lund, August 2001. Uppåkrastudier 6. (Eds.) B. Hårdh and L. Larsson. Lund, 79–95. Hedeager, L. 1997a. Odins offer: Skygger af en shamanistisk tradition i nordisk folkevandringstid. Tor (29), 265–278. ———. 1997b. Skuggor ur en annan verklighet: Fornnordiska myter. Orig. title Skygger af en anden virkelighed. Trans. G. Rydström. Stockholm. ———. 1999. Sacred Topography. Depositions of Wealth in the Cultural Landscape. In Glyfer och arkeologiska rum—en vänbok till Jarl Nordbladh. Gotarc Series A Vol. 3. Department of Archaeology. Göteborg University. (Eds.) A. Gustafsson & H. Karlsson. Göteborg, 229–252. ———. 2001. Asgard Reconstructed? Gudme—A ‘Central Place’ in the North. In Topographies of Power in the Early Middle Ages. (Eds.) M. de Jong, F. Theuws & C. van Rhijn. Leiden – Boston – Köln, 467–507. ———. 2011. Iron Age Myth and Materiality: An Archaeology of Scandinavia AD 400– 1000. London & New York. Hedeager, L. & E. Østmo (eds.). 2005. Norsk Arkeologisk Leksikon. Oslo. Hedlund, G. 1993. Utgrävningen 1992. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. OPIA. 7. Societas Archaeologica Upsaliensis. (Ed.) W. Duczko. Uppsala, 64–69. Hedman, A. 1989. Gård och samhälle under yngsta järnålder—utgångspunkter för en reviderad bebyggelsearkeologi. In Mänsklighet genom millennier: En vänbok till Åke Hyenstrand. (Ed.) M. Burström. Stockholm, 93–99. Heggstad, L. et al. 1993 (1930). Norrøn ordbok. Oslo. Heide, E. 2006a. Spinning seiðr. In Old Norse Religion in Long-Term Perspectives: Origins, Changes, and Interactions. Vägar till Midgård 8. (Eds.) Andrén, A. & K. Jennbert & C. Raudvere. Lund, 164–170. ———. 2006b. Gand, seid åg åndevind. Bergen. Heiler, F. 1961. Erscheinungsformen und Wesen der Religion. Stuttgart. Heizmann, W. 1995. Fløksand. In Reallexikon der Germanischen Altertumskunde. Band 9. (Eds.) H. Beck et al. Berlin – New York, 216–221. Hellberg, L. 1975. Ortnamnen och den forntida sveastaten. In Ortnamn och samhälle— aspekter—begrepp—metoder—i Hanaholmens kulturcentrum vid Helsingfors den 25–27 april 1975. NORNA, Non-Publ. Materials, 92–110. Publ. in Namn och bygd (2013), 9–31, with an appendix written by Thorsten Andersson. ———. 1976. Vallensjö: Ett uppländskt kultcentrum. Ortnamnssällskapets i Uppsala Årsskrift. Uppsala, 5–14. Bibliography 577

———. 1979. Forn-Kalmar: Ortnamnen och stadens äldsta förhistoria. In Kalmar stads historia: 1. Kalmarområdets forntid och stadens äldsta utveckling: Tiden intill 1300-talets mitt. (Ed.) I. Hammarström. Kalmar. ———. 1984–85. Aktuell forskning om tuna-namnen 1–5. Upsala Nya Tidning 6/11; 1984; 20/11 1984; 4/12 1984; 26/2 1985; 12/3 1985. Also published in Uppländska namnstudier. Valda namnspalter ur Upsala Nya Tidning 1982–2002. Ortnamnssällskapet i Uppsala. (Eds.) K. Leibring, S. Nyström & M. Wahlberg. Uppsala. 2011, 30–42. Also publ. 2011 in Uppländska namnstudier. Valda namnspalter ur Uppsala Nya Tidning 1982–2002, 30–41. ———. 1986a. Hedendomens spår i uppländska ortnamn. Ortnamnssällskapets i Uppsala Årsskrift. Uppsala, 40–71. ———. 1986b. “ ‘Ingefreds sten’ och häradsindelningen på Öland. In Festschrift für Oskar Bandle zum 60. Geburtstag am 11. Januar 1986. Beiträge zur nordischen Philologie. Band 15. (Eds.) H.-P. Naumann et al. Basel – Frankfurt am Main, 19–29. ———. 2014. “Inge och Inga—ett omaka kortnamnspar.” Studia anthroponymica Scandinavica (32), 37–58. Helle, K. 1974. Norge blir en stat 1130–1319. Oslo. Helleland, B. 1996. Ullins vang og Njords hov: Heidne gudar på kyrkjebakken? In Fra hedendom til kristendom: Perspektiver på religionsskiftet i Norge. KULT-serien 11. (Ed.) M. Rindal. Oslo, 125–140. Hellman, E. 2011. Vad är religion? En disciplinteoretisk metastudie. Nora. Hellquist, E. 1957 (1922). Svensk etymologisk ordbok I–II. Lund. Hellström, J. A. 1971. Biskop och landskapssamhälle i tidig svensk medeltid. Skrifter utgivna av Institutet för rättshistorisk forskning. Serien I. Rättshistoriskt Bibliotek. 16 Stockholm. ———. 1996. Vägar till Sveriges kristnande. Stockholm. Henare, A., M. Holbraad, & S. Wastell. 2007. Introduction. Thinking Through Things. In Thinking Through Things: Theorising Artefacts Ethnographically. London, 1–31. Hermann Pálsson & Paul Edwards 1989. See Eyrbyggja saga. Herschend, F. 1993. The Origin of the Hall in Southern Scandinavia. Tor (25), 175–199. ———. 1994. Models of petty rulership: Two early settlements in Iceland. Tor (26), 163–191, ———. 1995. Hus på Helgö. Fornvännen (90), 221–228. ———. 1996. A Note on Late Iron Age Kingship Mythology. Tor (28), 283–303. ———. 1997. Livet i hallen: Tre fallstudier i den yngre järnålderns aristokrati. OPIA. 14. Institutionen för arkeologi och antik historia. Uppsala universitet. Uppsala. ———. 1998. The Idea of the Good in Late Iron Age Society. OPIA. 15. Institutionen för arkeologi och antik historia. Uppsala universitet. Uppsala. ———. 1999. Halle. In Reallexikon der Germanischen Altertumskunde. Band. 13. (Eds.) H. Beck et al. Berlin – New York, 414–425. 578 Bibliography

———. 2001. Journey of Civilisation: The Late Iron Age View of the Human World. OPIA. 24. Institutionen för arkeologi och antik historia. Uppsala universitet. Uppsala. ———. 2009. The Early Iron Age in South Scandinavia: Social Order in Settlement and Landscape. OPIA 46. Uppsala. Herschend, F. & D. Kaldal Mikkelsen. 2003. The main building at Borg (I:1). In Borg in Lofoten: A chieftain’s farm in North Norway. (Eds.) Munch, G. Stamsø et al. Trondheim, 41–76. Heusler, A. 1911. Das Strafrecht der Isländersaga. Leipzig. Hewitt, M. A. 1996. How New is the ‘New Comparativism’? Difference, Dialectics, and World-Making. Method & Theory in the Study of Religion (8–1), 15–20. Heywood, A. 1998. Political ideologies. An Introduction. London. Hinnells, J. R. 1995. Temples (Ancient Near Eastern). In A New Dictionary of Religions. London, 517–518. Hjärne, E. 1951. Från Alsnö och Sko: Två undersökningar i svensk medeltidshistoria. Uppsala. Hjulström, B. 2008. Patterns in diversity: Geochemical analyses and settlement changes during the Iron Age—Early Medieval time in the Lake Mälaren region, Sweden. Theses and papers in scientific archaeology 11. Stockholm. Hoffmann, E. 1976. Königserhebung und Thronfolgeordnung in Dänemark bis zum Ausgang des Mittelalters. Beiträge zur Geschichte und Quellenkunde des Mittelalters. Band 5. Berlin – New York. Höfler, O. 1952. Germanisches Sakralkönigtum. Band 1. Der Runenstein von Rök und die Germanische Individualweihe. Tübingen/Münster/Köln. ———.1959. Der Sakralcharakter des germanischen Königtums. In La Regalità Sacra. Leiden, 664–701. Holmberg, B. 1986. Den hedenske gud Tyr i danske stednavne. In Mange bække små: Til John Kousgård Sørensen på tresårsdagen 6.12.1985. (Eds.) V. Dalberg & G. Fellows- Jensen. København, 109– 126. Holmberg, K. A. 1969. De svenska tuna-namnen. Acta Academiae regiae Gustavi Adolphi. XLV. Studier till en svensk ortnamnsatlas utgivna av Jöran Sahlgren. 12. Uppsala. Holmberg-[Harva], U. 1917. Valhall och världsträdet. Finsk tidskrift för vitterhet, veten- skap, konst och politik (48), 349–377. ———. 1922. Der Baum des Lebens. Annales Academiae Scientiarum Fennicae B 16:3. Helsinki. Holmbäck, Å & E. Wessén. 1979 (1933–46). See Svenska landskapslagar. Holmgren, G. 1937a. Gamla Uppsala och Mora Äng i medeltidslagarnas valföreskrifter. Uppsala. Bibliography 579

———. 1937b. Taga och vräka konung. Fornvännen (19), 19–26. ———. 1954. När försvann Mora Sten?. Historisk Tidskrift (74), 1–43. Holmqvist, W. 1962. Högsäte. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 7. Malmö, cols. 290–292. ———. 1960. The dancing gods. Acta Archaeologica (31), 101–127. ———. 1969. Comments. Norwegian Archaeological Review (2), 22–24. ———. 1969. Helgö—den gåtfulla ön. Uddevalla. ———. 1980. Vikingar på Helgö och Birka. Stockholm. 62. Holmqvist, W. & B. Arrhenius & P. Lundström. 1961–64. Excavations at Helgö I–II. KVHAA. Stockholm. Holmquist Olausson, L. 2001. Birkas befästningsverk—resultat från de senaste årens utgrävningar. In Birkas krigare: 11 artiklar kring Birkas befästningsverk och dess krigare. (Ed.) M. Olausson. Arkeologiska Forskningslaboratoriet. Stockholm, 9–15. Holmquist Olausson, L. & L. Kitzler Åhfeldt 2002. Krigarnas hus. Arkeologiska under- sökningar av ett hallhus i Birkas Garnison. RAÄ 35, Björkö, Adelsö sn, Uppland 1998– 2000. Rapport 4. Arkeologiska Forskningslaboratoriet. Stockholm. Holtsmark, A. 1964. Studier i Snorres mytologi. Skrifter utg. av Det Norske Videnskaps- Akademi i Oslo, Hist.-Filos. Klasse, Nye Serie. No. 4. Oslo. ———. 1992 (1970). Fornnordisk mytologi: Tro och myter under vikingatiden. Orig. title Norrøn mytologi: Tro og myter i vikingtiden. Trans. H. Williams. Lund. Homann, H. 1976. Baumkult. In Reallexikon der Germanischen Altertumskunde. Band. 2. (Eds.) H. Beck et al. Berlin – New York, 107–110. Honko, L. 1972. The problem of defining myth. In The Myth of the State. Scripta Instituti Donneriani Aboensis 6. T. Ahlbäck. Stockholm, 7–19. Hübler, F. 1996. Schwedische Runendichtung der Wikingerzeit. Runrön. 10. Runologiska bidrag utgivna av Institutionen för nordiska språk vid Uppsala universitet. (Eds.) L. Elmevik & L. Peterson. Uppsala. Hult de Geer, E. 1950. Det kliar i fingrarna. Aftonbladet 27/3 1950, 7. Hultgård, A. 1992. Religiös förändring, kontinuitet och ackulturation/synkretism i vikingatidens och medeltidens skandinaviska religion. In Kontinuitet i kult och tro från vikingatid till medeltid. Projektet Sveriges Kristnande. Publikationer. 1. (Ed.) B. Nilsson. Uppsala, 49–103. ———. 1993. Altskandinavische Opferrituale und das Problem der Quellen. In The Problem of Ritual. Scripta Instituti Donneriani Aboensis. XV. (Ed.) T. Ahlbäck. Stockholm, 221–259. ———. 1994. Anmeldelser: Gro Steinsland 1991. Maal og Minne (1–2), 75–80. ———. 1996. Fornskandinavisk kult—finns det skriftliga källor? In Religion från sten- ålder till medeltid. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter nr 19. (Eds.) K. Engdahl & A. Kaliff. Linköping, 25–57. 580 Bibliography

———. 1997. Från ögonvittnesskildring till retorik: Adam av Bremens notiser om Uppsalakulten i religionshistorisk belysning. In Uppsalakulten och Adam av Bremen. (Ed.) A. Hultgård. Nora, 9–50. ———. 1998. Runeninschriften und Runendenkmäler als Quellen der Religions geschichte. Runeninschriften als Quellen interdisziplinärer Forschung. Abhand lungen des Vierten Internationalen Symposiums über Runen und Runeninschriften in Göttingen vom 4.–9. August 1995. (Ed.) K. Düwel. Berlin – New York, 715–737. ———. 1999. Hammer. § 2. Religionsgeschichtliches. In Reallexikon der Germanischen Altertumskunde. Band 13. (Eds.) H. Beck et al. Berlin – New York, 486–492. ———. 2000. 2000. Jul. In Reallexikon der Germanischen Altertumskunde. Band 16. (Eds.) H. Beck et al. Berlin-New York, 100–105. ———. 2001a. Kultische Umfahrt. In Reallexikon der Germanischen Altertumskunde. Band 17. (Eds.) H. Beck et al. Berlin – New York, 437–442. ———. 2001b. Recension Henrik Janson. Templum Nobilissimum. Adam av Bremen, Uppsalatemplet och konfliktlinjerna i Europa kring år 1075. Göteborg 1998 (Avhandlingar från Historiska institutionen i Göteborg, 21). Kyrkohistorisk årsskrift 2000, 262–263. ———. 2001c. Lokalgottheiten. In Reallexikon der Germanischen Altertumskunde. Band 18. (Eds.) H. Beck et al. Berlin – New York, 576–579. ———. 2002. Menschenopfer. In Reallexikon der Germanischen Altertumskunde. Band 19. (Eds.) H. Beck et al. Berlin – New York, 533–546. ———. 2003a. Phallusverehrung. In Reallexikon der Germanischen Altertumskunde. Band 23. (Eds.) H. Beck et al. Berlin – New York, 135–139. ———. 2003b, Religion. In Reallexikon der Germanischen Altertumskunde. Band 24. (Eds.) H. Beck et al. Berlin – New York, 429–457. ———. 2003c. Ár-‘Gutes Jahr unt Ernteglück’—ein Motivkomplex in der altnord- ischen Literatur und sein religionsgeschichtlicher Hintergrunt. Runica— Germanica—Mediaevalia. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 37. (Eds.) W. Hiezmann & A. van Nahl. Berlin – New York, 282–308. ———. 2005. Seherinnen. In Reallexikon der Germanischen Altertumskunde. Band 28. (Eds.) H. Beck et al. Berlin – New York, 113–121. ———. 2007. Kultens språk—rituella formler och rituell åkallan i fornskandianvisk religion. Kult, guld och makt. (Ed.) I. Nordgren. Göteborg, 123–132. ———. 2008. The religion of the Vikings. In The Viking World. (Ed.) S. Brink. London & New York, 212–218. ———. 2011. Óðinn, Valhǫll and the Einherjar. Eschatological myth and ideology in Late Viking Age. In Ideology and power in the Viking and Middle Ages. Scandinavia, Iceland, Ireland, Orkney and the Faeroes. (Eds.) Steinsland, G. et al. Leiden, 297–328. Bibliography 581

———. 2014. Kultplats och gudstro. Om framväxten av lokala gudomar på germanskt område. In Den heliga platsen. Handlingar från symposiet Den heliga platsen Härnösand 15–18 september 2011. (Eds.) E. Nyman et al. Sundsvall, 23–38. Hultkrantz, Å. 1956. Configurations of religious belief among the Wind River Shoshoni. Ethnos (21:3–4), 194–215. ———. 1976. Divinationsformer. En klassifikation. In Nordisk folktro. Studier tillägnade Carl-Herman Tillhagen 17 december 1976. (Eds.) B. af Klintberg et al. Nordiska museet. Stockholm, 49–70. Huth, O. 1943. Vesta. Leipzig & Berlin. Hyenstrand, Å. 1974. Centralbygd—Randbygd: Strukturella, ekonomiska och administra- tiva huvudlinjer i mellansvensk yngre järnålder. Acta Universitatis Stockholmiensis. Studies in North-European Archaeology. 5. Stockholm. ———. 1996. Lejonet, draken och korset: Sverige 500–1000. Lund. Ilkjær, J. 1990. Illerup Ådal. 1. Die Lanzen und Speere. Textband. Jutland Archaeological Society Publications XXV:1. Aarhus-Højbjerg. ———. 2000. Den første Norgehistorien. Illerupfunnet: Ny innsikt i skandinavisk romer- tid. Forord og etterskrift av Terje Gansum. Tønsberg. Ingstad, A. S. 1993 (1992). Oseberg-dronningen—hvem var hun? In Osebergdronningens grav: Vår arkeologiske nasjonalskatt i nytt lys. (Eds.) A. E. Christensen et al. Oslo, 224–256. Iregren, E. 1989. Under Frösö kyrka—ben från en vikingatida offerlund? In Arkeologi och religion: Rapport från arkeologidagarna 16–18 januari 1989. University of Lund. Institute of Archaeology. Report Series No. 34. (Eds.) L. Larsson & B. Wyszomirska. Lund, 119–133. Jackson, P. 2014a. Alhs och gudhūs. Kultplatsbeteckningar och översättningsanom‑ alier i den gotiska bibeln. In Den heliga platsen. Handlingar från symposiet Den heliga platsen Härnösand 15–18 september 2011. (Eds.) E. Nyman et al. Sundsvall, 39–48. ———. 2014b. “A Song Worth Fifty Cows: Graeco-Indo-Iranian Variations on the Etiology of Sacrifice”. In: Münchener Studien zur Sprachwissenschaft (68/1), 101–118. Jacob-Friesen, G. 1986. Eidring. § 2. Archäologisches. In Reallexikon der Germanischen Altertumskunde. Band 6. (Eds.) H. Beck et al. Berlin – New York, 250–253. Janson, H. 1997. Adam av Bremen, Gregorius VII och Uppsalatemplet. In Uppsalakulten och Adam av Bremen. (Ed.) A. Hultgård. Nora, 131–195. ———. 1998. Templum Nobilissimum: Adam av Bremen, Uppsalatemplet och konfliktlinjerna i Europa kring år 1075. Avhandlingar från Historiska institutionen i Göteborg. 21. Göteborg. Jansson, S. B. F. 1964. See Vs above. ———. 1985. Runinskrifter i Sverige. Stockholm. 582 Bibliography

———. 1987. Runes in Sweden. Orig. title Runinskrifter i Sverige. Trans. P. Foote. Stockholm. Jensen, J.S. 2014. What is religion? Durham. Jesch, J. 1994. In Praise of Ástríðr Óláfsdóttir. Saga-Book. Vol. XXIV. Part 1, 1–18. ———. 1996 (1991). Women in the Viking Age. Woodbridge. Jocobsen, L. & E. Moltke. 1941–42. See DR. Johanson O. S. & G. Stamsø Munch 2003. Introduction and summary. In Borg in Lofoten: A chieftain’s farm in North Norway. (Eds.) Munch, G. Stamsø et al. Trondheim, 11–18. Johnsen, S. 1968. Stuttruner i vikingtidens innskrifter. Oslo. Jón Hnefill Aðalsteinsson. 1992. A Piece of Horse-Liver and the Ratification of Law. In Snorrastefna: 25.–27. júlí 1990. Rit Stofnunar Sigurðar Nordals 1. (Ed.) Úlfar Bragason. Reykjavík, 81–98. ———. 1998. A Piece of Horse Liver: Myth, Ritual and Folklore in Old Icelandic Sources. Reykjavík. Jón Jóhannesson. 1956. Íslendinga saga I. Reykjavík. Jón Viðar Sigurðsson. 1994. Forholdet mellom verdslig og religiøs makt på Island i fristatsperioden. In Myte og ritual i det førkristne Norden: Et Symposium. (Eds.) J. P. Schjødt et al. Odense, 129–139. ———. 1999. Chieftains and Power in the Icelandic Commonwealth, The Viking Collection, 12. Odense: Odense University Press. ———. 2003. Kristninga i Norden 750–1200. Oslo. ———. 2008a. Det norrøne samfunnet: Vikingen, kongen erkebiskopen, og bonden. Oslo. ———. 2008b. Iceland. In The Viking World. (Ed.) S.Brink. London & New York, 571–578. ———. 2010. Den vennlige vikingen: Vennskapets makt i Norge og på Island 900–1300. Oslo. ———. 2011. Kings, Earls, and Chieftains: Rulers in Norway, Orkney and Iceland c. 900–1300. In Ideology and Power in the Viking and Middle Ages: Scandinavia, Iceland, Ireland, Orkney and the Faeroes. (Eds.) G. Steinsland et al. Leiden, 69–108. ——— et al. 2005. Ingólfr: Norsk–Islandsk Hopehav 870–1536. Førde. Jónas Kristjánsson. 1988. Eddas and Sagas: Iceland’s Medieval Literature. Trans. P. Foote. Reykjavík. Jonsson, E. 2006. Eldstålsformade ringar: En studie av när de tillverkades, och var och hur de hittas inom gränserna för dagens Sverige. Kandidatuppsats (C) i arkeologi. Institutionen för arkeologi och antikens kultur, Stockholms universitet. Stockholm. Jørgensen, L. 1998. En storgård fra vikingetid ved Tissø, Sjælland—en foreløbig præsen- tation. In Centrala platser—Centrala frågor: Samhällsstrukturen under Järnåldern. En Vänbok till Berta Stjernquist. Uppåkrastudier. 1. (Eds.) L. Larsson & B. Hårdh. Stockholm, 233–248. ———. 2002. Kongsgård—kusted—marked: Overvejelser omkring Tissøkompleksets struktur og funktion. In Plats och praxis—studier av nordisk förkristen ritual. Vägar till Midgård 2. (Eds.) K. Jennbert, A. Andrén & C. Raudvere, Lund, 215–247. Bibliography 583

———. 2009. Pre-Christian Cult at Aristocratic Residences and Settlement Com‑ plexes in Southern Scandinavia in the 3rd–10th centuries AD. In Glaube, Kult und Herrschaft: Phänomene des Religiosen im 1. Jahrtausend n. Chr. in Mittel- und Nordeuropa. Akten des 59. Internationalen Sachsensymposiums. Koll. Vor- und Fuhgesch. 12. (Eds.) U. von Freeden/H. Friesinger/E. Wamers. Bonn, 329–354. ———. 2011. Gudme-Lundeborg on Funen: As a model for northern Europe? In The Gudme/Gudhem Phenomenon. (Eds.) O. Grimm & A. Pesch. Neumünster, 77–90. Jungner, H. 1919. Uppsala- och Vendel-konungarnes mytiska ättefäder. Fornvännen (14), 79–102. ———. 1930. Den gotländska runbildstenen från Sanda: Om Valhallstro och hednisk begravningsritual. Fornvännen (25), 65–82. Kabell, A. 1975. Baugi und der Ringeid. Arkiv för Nordisk Filologi (90), 30–40. Kaliff, A. 1997 Grav och kultplats: Eskatologiska föreställningar under yngre bronsålder och äldre järnålder i Östergötland. Aun 24. Department of Archaeology, Uppsala University. Uppsala. ———. 2006. Gravhus, kulthus eller tempel? In Kulthus och dödshus: Det ritualiserade rummets teori och praktik. (Eds.) M. Anglert et al. Lund, 129–142. Kaliff, A. & O. Sundqvist. 2004. Oden och Mithraskulten: Religiös ackulturation under romersk järnålder och folkvandringstid. OPIA. 35. Institutionen för arkeologi. Uppsala universitet. Kaplan, M. 2011. Thou Fearful Guest. Addressing the past in four tales in Flateyjarbók. FFC CXLVIII. Helsinki. Källström, M. 2007. Mästare och minnesmärken: Studier kring vikingatida runristare och skriftmiljöer i Norden. Acta Universitatis Stockholmiensis. Stockholm Studies in Scandinavian Philology. Nr 43. Stockholm. ———. 2010a. Forsaringen tillhör 900-talet. Fornvännen (105), 228–232. ———. 2010b. Ett par norrländska namnproblem: nom. þrusun (Hs 12) och nom. aunhar (M 8). Studia Anthroponymica Scandinavica 28: 5–26. ———. 2011. Ett namnproblem hos Åsmund Kåresson: Runföljden k)anilf på den försvunna U 986 från Kungsgården, Gamla Uppsala. Studia anthroponymica Scandinavica (29), 5–20. Karlenby, L. 2005. Hall, hem och hird. Järnålderstida rumslig organisation i Barksta och Torp. In Tidens resenär: Arkeologiska händelser längs vägen mellan Köping och Kolsva. (Ed.) L. Karlenby. Uppsala, 111–124. Kauffmann, F. 1892. Vingolf. Zeitschrift fur deutsches Altertum (36), 32–41. Kehrer, G. 1998. Legitimation/ Legitimität. In Handbuch religionswissenschaftlicher Grundbegriffe. Band IV. (Eds.) H. Cancik et al. Stuttgart – Berlin – Köln, 46–47. Kern, F. 1954 (1914). Gottesgnadentum und Widerstandsrecht im früheren Mittelalter: Zur Entwicklungsgeschichte der Monarchie. Hrsg. von R. Buchner. Münster/Köln. Keyser, R. 1847. Nordmændenes Religionsforfatning i Hedendommen. Oslo. ———. 1868. The private life of the old Northmen. London. 584 Bibliography

Kiil, V. 1960. Hliðsljálf og seiðhjallr. Arikiv för nordisk filologi (75), 84–112. ———. 1962. The Norse Prophetess and the Ritually Induced Prostitution. Norveg (9), 159–174. Klos, L. 2007. Lady of the Rings: Järnålderns kvinnor mellan makt och kult. In Kult, guld och makt: Ett tvärvetenskapligt symposium i Götene. Historieforum Västra Götaland. Serie B. Vetenskapliga rapporter och Småskrifter. No. 4. (Ed.) I. Nordgren. Skara, 70–86. Kluge, F. 1957 (1883). Etymologisches Wörterbuch der deutschen Sprache. Berlin. Knirk, J. E. 2006. Tune: Runological. In Reallexikon der Germanischen Altertumskunde. Band 31. (Eds.) H. Beck et al. Berlin – New York, 332–335. Knott, K. 1998. Issues in the study of religion and locality. Method & Theory in the Study of Religion (10), 279–290. ———. 2009. From locality to location and back again: A spatial journey in the study of religion. Religion (39/2), 54–60. ———. 2014. Spatial Methods”. In The Routledge Handbook of Research Methods in the Study of Religion. (Eds.) M. Stausberg & S. Engler. London & New York, 491–501. Köbler, G. 1986. Eid. § 3. Rechtshistorisches. In Reallexikon der Germanischen Altertumskunde. Band 6. (Eds.) H. Beck et al. Berlin – New York, 539–242. Köhler, U. 2004. Sakralkönigtum. §2. Ethnolog. In Reallexikon der Germanischen Altertumskunde, Band 26. (Eds.) Heinrich Beck et al. Berlin – New York, 181–183. Kolbrún Haraldsdóttir, 1994. Fagrskinna. In Reallexikon der Germanischen Alter tumskunde. Band 8. (Eds.) H. Beck et al. Berlin – New York, 142–151. Kousgård Sørensen, J. 1985. Gudhem. Frühmittelalterliche Studien (19), 131–138. ———. 1989. Om personnavne på -vi/-væ og den før-kristne præstestand: Med nogle overvejelser over en omstridt passage i Glavendrup-stenens indskrift. Danske studier. København, 5–33. ———. 1992. Haupttypen sakraler Ortsnamen Sudskandinaviens. In Der historische Horizont der Gotterbild-Amulette aus der Ubergangsepoche von der Spatantike zum Fruhmittelalter. (Ed.) K. Hauck. Göttingen, 228–240. Krag, C. 1991. Ynglingatal og Ynglingesaga: En studie i historiske kilder. Studia humani- ora. 2 (Oslo: Universitetsforlaget, 1991). ———. 2000. Norges historie frem til 1319. Oslo. ———.2001. Trosskiftet og teorien om sakralkongedømmet. Collegium Medievale (14), 211–225. ———. 2007. Trøndelag. In Reallexikon der Germanischen Altertumskunde. Band 35. (Eds.) H. Beck et al. Berlin – New York, 268–272. ———. 2008. The creation of Norway. In The Viking World. (Ed.) S. Brink. London & New York, 645–651. ———. 2009. Ynglingatal str. 4—omskrivningen viljes borg som dateringskriterium. Maal og minne (02/2009), 129–146. Bibliography 585

Krause, W. 1930. Die Kenning als typische Stilfigur der germanischen und keltischen Dichtersprache. Schriften der Königsberger gelehrten Gesellschaft. 7. Jahr. Geistes wissenschaftliche Klasse. Heft 1. Halle (Saale). ———. 1944. Ing. Nachrichten von der Akademie der Wissenschaften in Göttingen aus dem Jahre 1944. Philol.-Hist. Klasse, Nr. 10. Göttingen. ———. 1966. Die Runeninschriften im älteren Futhark 1. Text. Mit Beiträgen von Herbert Jankuhn. Abhandlungen der Akademie der Wissenschaften in Göttingen, Philologisch-historische Klasse, 3. Folge 65. Göttingen. Kristján Eldjárn. 1956. Kuml og Haugfé: Úr heiðnum sið á Íslandi. Reykjavík. Kristján Eldjárn & Adolf Friðriksson. 2000. Kuml og Haugfé: Úr heiðnum sið á Íslandi. New edition. Reykjavík. Kristoffersen, S. 2000. Sverd og spenne: Dyreornamentik og sosial kontekst. Studia huma- nitatis Bergensis 13. Bergen. ———. 2006. Tu. In Reallexikon der Germanischen Altertumskunde. Band 31. (Eds.) H. Beck et al. Berlin-New York, 307–309. Kuhn, H. 1971. Kleine Schriften II. Aufsätze und Rezensionen aus den Gebieten der germanischen und nordischen Sprach-, Literatur- und Kulturgeschichte. Zweiter Band. Literaturgeschichte, Heldensage und Heldendichtung, Religions- und Sittengeschichte, Recht und Gesellschaft. Berlin – New York. ———. 1978. Kleine Schriften IV. Aufsätze und Rezensionen aus den Gebieten der germanischen und nordischen Sprach-, Literatur- und Kulturgeschichte. Vierter Band. Aufsätze aus den Jahren 1968–1976. Berlin – New York. La Farge, B. 1994. Review Gro Steinsland 1991. Skandinavistik (24:1), 55–59. La Farge, B. & J. Tucker. 1992. Glossary to the Edda. Heidelberg. Läffler, Fr. L. 1905. Ännu några ord om de hedniska edsformulären i Norden. Antiqvarisk tidskrift för Sverige (13:3), 1–11. Läffler, L. Fr. 1911. Det evigt grönskande trädet vid Uppsala hednatämpel. In Fästskrift till H. F. Feilberg. Svenska landsmål och svenskt folkliv. (Ed.) J. A. Lundell. Stockholm, 617–196. Lamm, J. P. 2004. Figural Gold Foils Found in Sweden: A Study Based on the Discoveries from Helgö. In Excavations at Helgö XVI. Exotic and Sacral Finds from Helgö. Kungl. Vitterhets Historie och Antikvitets Akademien. Stockholm, 41–142. Lamm J. P., H. Hydman & M. Axboe. 1999. “ ‘Århundradet s brakteat’—kring fyndet av en unik tionde brakteat från Söderby i Danmarks socken, Uppland. Fornvännen (94), 225–243. Lamm, K. 1999. Helgö. §§ 2–5. In Reallexikon der Germanischen Altertumskunde. Band. 14. (Eds.) H. Beck et al. Berlin – New York, 286–291. ———. 2004. In J. P. Lamm 2004 (see above). Landau, Peter 1982. Eigenkirchenwesen. In Theologische Realenzyklopädie 9. Berlin – New York, 399–404. 586 Bibliography

Lange, Wolfgang. 1958. Studien zur christlichen Dichtung der Nordgermanen: 1000–1200. Palaestra. Untersuchungen aus der Deutschen und Englischen Philologie und Literaturgeschichte. Band 222, Göttingen: Vandenhoeck und Ruprecht. Larsson, G. 2007. Ship and Society: Martime Ideology in Late Iron Age Sweden. Aun 37. Uppsala. Larsson, L. 2006a. Ritual building and ritual space: Aspects of investigations at the Iron Age central site Uppåkra, Scandia, Sweden. In Old Norse Religion in Long-Term Perspectives. Origins, Changes, and Interactions. Vägar till Midgård 8. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 248–253. ———. 2006b. Hall, harg eller hof. Ett kulthus i Uppåkra. Kulthus & gravhus. Det ritualiserade rummets teori och praktik. (Eds.) M. Anglert et al. Stockholm, 143–152 ———. 2006c. Rum, rymd och areal. Ett kulthus och dess närmiljö ur ett ceremoniellt perspektiv. Bebyggelsehistorisk tidskrift (52), 7–19. ———. 2011. A ceremonial building as a ‘home of the gods’? In The Gudme/Gudhem Phenomenon. (Eds.) O. Grimm & A. Pesch. Neumünster, 189–206. Larsson, L. & K.-M. Lenntorp. 2004. The Enigmatic House. In Continuity for Centuries: A Ceremonial Building and its Context at Uppåkra, Southern Sweden. Uppåkrastudier. 10. (Ed.) L. Larsson. Stockholm, 3–48. Larsson, M. G. 1990. Runstenar och utlandsfärder: Aspekter på det senvikingatida sam- hället medutgångspunkt i de fasta fornlämningarna. Acta Archaeologica Lundensia, series in 8o, 18. Lund. ———. 1993. Kung Eriks nyordning? Spår av samhällsorganisation i vikingatidens Uppland. In Kärnhuset i riksäpplet. Upplands fornminnesförening och hembygds- förbunds årsbok. Uppland 1993. (Eds.) K. Blent et al. Uppsala, 105–134. ———. 1997. Vikingar i österled. En samlingsutgåva av: Ett ödesdigert vikingatåg, Ingvar den vittfarnes resa 1036–1041. Väringar, Nordbor hos kejsaren i Miklagård. Rusernas rike, Nordborna och rysslands födelse. Stockholm. Lassen, A. 2011. Odin på kristent pergament: En teksthistorisk studie. Københavns Universitet. Laur, W. 2001. Germanische Heiligtümer und Religion im Spiegel der Ortsnamen: Schleswig-Holstein, nördliches Niedersachsen und Dänemark. Kieler Beiträge zur deutschen Sprachgeschichte. Band 21. (Eds.) F. Debus & W. Laur. Neumünster. Leeuw, G. van der. 1956 (1933). Phänomenologie der Religion. Neue Theologische Grundrisse. Tübingen. Lehmann, F. R. 1930. Die Polynesischen Tabusitten: Eine ethnosoziolog. u. religionswis- senschaftl. Untersuchung. Leipzig. Lerche Nielsen, M. 1998. Glavendrup. § 2 Runologisches. In Reallexikon der Germani schen Altertumskunde. Band 12. (Eds.) H. Beck et al. Berlin – New York, 196–200. ———. 2003. Från prästgården till drottning Disas Samna: Två problematiska runstenar från Gamla Uppsala. Fornvännen (98), 195–205. Bibliography 587

Lett, J. W. 1996. Emic/Etic Distinctions. In Encyclopedia of Cultural Anthropology. Vol. 2. (Eds.) D. Levinson & M. Ember. New York, 382–383. Lewis, C. T. & C. Short. 1879. A Latin Dictionary. New York. Lid, N. 1942. Gudar og gudedyrkning. In Religionshistorie. Nordisk kultur. XXVI. Oslo, 80–153. Liddell, H.G. & R. Scott. 1888. An Intermediate Greek-English Lexicon. Oxford. Lidén, H.-E. 1969. From Pagan Sanctuary to Christian Church: The Excavation of Mære Church in Trøndelag. Norwegian Archaeological Review (2), 3–32. ———. 1996. Utgravning i Mære kirke: Hvordan skal funnene tolkas. In Før og etter Stiklestad 1030: Religionsskifte—kulturforhold—politisk makt. Seminar på Stiklestad, 1994. (Ed.) Ø. Walberg. Verdal, 59–68. ———. 1999. En gullgubbe: Festskrift til Hans-Emil Lidén. (Eds.) A. M. Hoff & A. T. Hommedal. Øvre Ervik. Liedgren, J. 1966. Mora sten. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 11. Malmö, cols. 689–690. Liestøl, A. 1979. Runeringen i Forsa: Kva er han, og når vart han smidd? Saga och Sed: Gustav Adolfs Akademiens Årsbok. Uppsala, 12–27. Lincoln, B. 1986. Myth, Cosmos, and Society: Indo-European Themes of Creation and Destruction. Cambridge, Massachusetts – London. ———. 1994. Authority. Construction and Corrosion. Chicago – London. ———. 1999. Theorizing Myth. Narrative, Ideology, and Scholarship. Chicago – London. ———. 2012. Gods and Demons, Priests and Scholars. Critical Explorations in the history of religions. Chicago. ———. 2014. Between History and Myth. Stories of Harald Fairhair and the Founding of the State. Chicago. Lindeblad, K. 1997. Gården från yngre järnåldern—skede 3. In Kungens gods i Borg: Om utgrävningarna vid Borgs säteri i Östergötland. Red. K. Lindeblad & A.-L. Nielsen. Linköping, 31–44. Linderholm, E. 1919.Vendelshögens konunganamn i socknens 1600-talstradition. Namn och Bygd (7), 36–40. Lindkvist, T. 1979. Landborna i Norden under äldre medeltid. Uppsala. ———. 1995 (1988). Plundring, skatter och den feodala statens framväxt: Organisatoriska tendenser i Sverige under övergången från vikingatid till tidig medeltid. Opuscula Hisorica Upsaliensia. 1. Uppsala. ———. 1996. Kungamakt, kristnande, statsbildning. In Kristnandet i Sverige: Gamla källor och nya perspektiv. Projektet Sveriges Kristnande, Publikationer. 5. (Ed.) B. Nilsson. Uppsala, 217–241 ———. 2008. The emergence of Sweden. In The Viking World. (Ed.) Stefan Brink. London & New York, 668–674. 588 Bibliography

———. 2009. Klerkernas tid 800–1520. In Det svenska samhället 800–1720: Klerkernas och adelns tid. (Eds.) T. Lindkvist & M. Sjöberg. Lund, 31–202. Lindkvist, T. & K. Ågren. 1985. Sveriges medeltid. Stockholm. Lindquist, I. 1944. Kungadömet i hednatidens Sverige. Arkiv för Nordisk Filologi (58), 221–234. ———. 1963. Två vikingatida gudabeläten. In Kulturen (1962), 70–78. Lindqvist, S. 1934. En förnäm kvinnas grav på Valsgärde gravbacke. Finska fornminnes- föreningens tidskrift (40), 138–142. ———. 1936. Uppsala högar och Ottarshögen. KVHAAH. Stockholm. ———. 1941–1942. Gotlands Bildsteine. I–II. KVHAAH. Stockholm. ———. 1962–63. Forngutniska altaren och därtill knutna studier. Kungl. Humanistiska Vetenskaps- Samfundet. Årsbok. 1962–1963. Stockholm. Lindow, J. 1988. Scandinavian Mythology: An Annotated Bibliography. New York – London. ———. 2001. Norse Mythology. A Guide to the Gods, Heroes, Rituals, and Beliefs. Oxford. Ljungberg, H. 1938. Den nordiska religionen och kristendomen: Studier över det nordiska religionsskiftet under vikingatiden. Nordiska texter och undersökningar 11. Stockholm – København. ———. 1947. Tor: Undersökningar i indoeuropeisk och nordisk religionshistoria. Uppsala. Ljungkvist, J. 2005. Uppsala högars datering. Fornvännen (100), 245–259. ———. 2006. En hiar atti rikR: Om elit, struktur och ekonomi kring Uppsala och Mälaren under yngre järnålder. Aun 32. Uppsala. ———. 2008a. Dating two royal mounds of Old Uppsala: Evaluating the elite of 6th to 7th century Middle Sweden. Korrespondenzblatt: Römisch-Germanisches Zentralmuseum. Mainz. ———. 2008b. Vendel—Development and Change on a Burial Ground over 1300 Years. In Valsgärde studies: The place and its people, past and present, OPIA 42. Uppsala, 13–55. ———. 2013. Monumentaliseringen av Gamla Uppsala. In Gamla Uppsala i ny belysning. (Eds.) O. Sundqvist and P. Vikstrand with J. Ljungkvist. Religionsvetenskapliga studier från Gävle 9. Gävle, 33‒67. ——— et al. 2000. I maktens närhet: Två boplatsundersökningar i Gamla Uppsala. Raä 285, Norra Gärdet. Raä 547, Mattsgården. Gamla Uppsala socken, Uppland. Societas Archaeologica. Upsaliensis. Skrifter nr 1. Uppsala. Löfving, C. 2005. Forsaringen åter till medeltiden. In Hälsingerunor: En hembygdsbok. Bollnäs, 111–118. ———. 2010. Forsaringen är medeltida. Fornvännen (105), 48–53. Løken, T. 1992. Ullandhaug sett i lys av Forsandundersøkelsene. AmS-Småtrykk 26. Stavanger. ———. 2001. Oppkomsten av den germanske hallen: Hall og sal i eldre jernalder i Rogaland. Viking (LXIV), 49–86. Bibliography 589

Løkka, N. 2010. Steder og landskap i norrøn mytologi. En analyse av topografi og kosmologi i gudediktene av Den eldre Edda. Oslo. Lönnerholm, T. 1993. Mora sten och Mora stenar: En vägledning till ett märkligt nationalmonument. Utg. Riksantikvarieämbetet i samarbete med Lagga och Danmarks hembygdsföreningar. Stockholm. Lönnqvist O. & G. Widmark. 1997. Den fredlöse och Oklunda-ristningens band. Saga och Sed: Gustav Adolfs Akademiens Årsbok. 1996–97. Uppsala, 145–159. Lönnroth, L. 1963–64. Kroppen som själens spegel—Ett motiv i de isländska sagorna. Lychnos: Lärdomshistoriska samfundets årsbok. 1963–64. Stockholm, 24–61. ———. 1968. Filologi och ordräkning. Arkiv för nordisk filologi (83), 241–249. ———. 1969. The Noble Heathen: A Theme in the Sagas. In Scandinavian Studies (41/1), 1–29. ———. 1977. The Riddles of the Rök-Stone: A Structural Approach. Arkiv för nordisk filologi (92), 1–57. ———. 1986. Dómaldi’s Death and the Myth of Sacral Kingship. In Structure and Meaning in Old Norse Literature: New Approaches to Textual Analysis and Literary Criticism. The Viking Collection. 3. (Eds.) J. Lindow et al. Odense, 73–93. ———. 1996. En fjärran spegel: Västnordiska berättande källor om svensk hedendom och om kristningsprocessen på svenskt område. In Kristnandet i Sverige: Gamla källor och nya perspektiv. Projektet Sveriges kristnande. Publikationer. 5. (Ed.) B. Nilsson. Uppsala, 141–158. ———. 1997. Hövdingahallen i fornnordisk myt och saga. In “. . . gick Grendel att söka det höga huset . . .”: Arkeologiska källor till aristokratiska miljöer i Skandinavien under yngre järnålder. Rapport från ett seminarium i Falkenberg 16–17 november 1995. Hallands länsmuseers Skriftserie No. 9/GOTARC C. Arkeologiska skrifter. No. 17. (Eds.) J. Callmer & E. Rosengren. Halmstad, 31–37. Lucas, G. 2009. Hofstaðir: Excavations of a Viking Age feasting hall in North-Eastern Iceland. Intitute of Archaeology Monograph Series (1). Reykjavík. Lucas, G. & T. McGovern. 2008. Bloody slaughter: Ritual decapitation and display at the Viking settlement of Hofstaðir, Iceland. European Journal of Archaeology (10/1), 7–30. Lund Hansen, U. 2011. New analyses of the Helgö drinking-glass fragments. In Excavations at Helgö XVIII. (Eds.) B. Arrhenius & U. O’Meadhra. Stockholm, 97–142. Lundqvist, L. 1996. Slöinge—en stormansgård från järnåldern. In Slöinge och Borg: Stormansgårdar i öst och väst. Riksantikvarieämbetet arkeologiska undersökningar skrifter nr. 18. Linköping, 9–52. ———. 1997. Slöinge: om ett pågående projekt. In “. . . gick Grendel att söka det höga huset . . .”: Arkeologiska källor till aristokratiska miljöer i Skandinavien under yngre järnålder. Rapport från ett seminarium i Falkenberg 16–17 november 1995. Hallands länsmuseers Skriftserie No. 9/GOTARC C. Arkeologiska skrifter. No. 17. (Eds.) J. Callmer & E. Rosengren. Halmstad, 590 Bibliography

———. 2003. Slöinge 1992–1996. Undersökningar av en boplats från yngre järnåldern. Slöingeprojektet 2. Gotarc serie C. Arkeologiska skrifter No. 42. Göteborg. Lundström, A. & B. von Heland. 1964. Iron. In Excavations at Helgö II. Stockholm. MacCana, P. 1970. Celtic Mythology. London. Magnell, O. & E. Iregren. 2010. Veitstu hvé blóta skal? The Old Norse blót in light of osteological remains from Frösö Church, Jämtland, Sweden. Current Swedish Archaeology (18), 223–250. Magnús Már Lárusson. 1965. Källa. Island. Kulturhistoriskt lexikon för nordisk medeltid. Vol. 10. Malmö, cols. 59–60. Maier, B. 1994. Lexikon der keltischen Religion und Kultur. Stuttgart. ———. 2000a. Idole und Idolatrie. §2. Religionswissenschaftliche Aspekte. In Reallexikon der Germanischen Altertumskunde. Band 15. (Eds.) H. Beck et al. Berlin – New York, 329–330. ———. 2000b. Irminsul. § 2. Sprachliche Deutung. § 3 Religionsgeschichtliche. In Reallexikon der Germanischen Altertumskunde. Band 15. (Eds.) H. Beck et al. Berlin – New York, 505–506. ———.2003. Die Religion der Germanen: Götter, Mythen, Weltbild. München. Males, M. 2010. Mytologi i skaldedikt, skaldedikt i prosa. En synkron analys av mytologiska referenser i medeltida norröna handskrifter. Oslo. Malinowski, B. 1948. Magic, Science, and Religion and other Essays. London. Malmer, B. 1996. Sigtunamyntningen som källa till Sveriges kristnande. Kristnandet i Sverige. Gamla källor och nya perspektiv. Projektet Sveriges Kristnande. Publi kationer. 5. (Ed.) B. Nilsson. Uppsala, pp. 85–113. Mannhardt, W. 1904–05 (1874–1875). Wald- und Feldkulte I–II. Berlin. Marett, R. R. 1909. The threshold of religion. London. Marold, E. 1972. Das Walhallbild in Eiríksmál und den Hákonarmál. Medieval Scandinavia (5), 19–33. ———. 1983. Kenningkunst: Ein Beitrag zu einer Poetik der Skaldendichtung. Quellen und Forschungen zur Sprach- und Kulturgeschichte der germanischen Völker. Neue Folge. 80 (204). Berlin – New York. ———. 1987. Die norwegische Reichseinigung und die Prieslieddichtung. In Arbeiten zur Skandinavistik. 7. Arbeitstagung der Skandinavisten des Deutschen Sprach gebietes: 4.8.–10.8.1985 in Skjeberg/Norwegen. (Ed.) U. Groenke. Frankfurt am Main – Bern – New York – Paris, 59–97. ———. 2000. Karlevi. § 2. In Reallexikon der Germanischen Altertumskunde. Band 16. (Eds.) H. Beck et al. Berlin – New York, 275–280. Marstrander, C. J. S. 1915a. Tor i Irland. Maal og Minne, 80–89. ———. 1915b. Thor en Irlande. Revue Celtique (36), 246–247. ———. 1930. Tunestenen. Norsk tidsskrift for sprogvidenskap (4), 294–358. Maurer, K. 1855–1856. Die Bekehrung des norwegischen Stammes zum Christentume, in ihrem geschichtlichen Verlaufe quellenmässig geschildert. München. Bibliography 591

———. 1869. Die Quellenzeugnisse über das erste Landrecht und über die Ordnung der Bezirksverfassung des isländchen Freistaates. Abhandlungen der philosophisch- philologischen Classe der Königlich Bayerischen Akademie der Wissenschaften. 12. Bd., 1. Abth. München. ———. 1873. Zur Urgeschichte der Godenwürde. Zeitschrift für deutsche Philologie (4), 125–130. Mauss, M. 1997 (1950). The Gift: The Form and Reason for Exchange in Archaic Societies. Orig. title Essai sur le Don. Trans. W. D. Halls. London. McCutcheon, R. T. 1997. Manufacturing Religion: The Discourse on Sui Generis Religion and the Politics of Nostalgia. New York – Oxford. McDonnel & Maclean 2009. See Lucas 2009. McKinnell, J. 2010. Ynglingatal—A Minimalist Interpretation. Scripta Islandica— Isländska sällskapets Årsbok 60/2009, 23–48. McKinnell, J., R. Simek & K. Düwel. 2004. Runes, Magic and Religion: A Sourcebook. Studia Medievalia Septentrionalia. 10. Wien. McTurk, R. W. 1975–76. Sacral Kingship in Ancient Scandinavia. A Review of Some Recent Writings. Saga-Book. Vol. XIX. Parts II–III, 139–169. ———. 1994. Scandinavian Sacral Kingship Revisited. Saga-Book. Vol. XXIV. Part I, 19–32. Meissner, R. 1921. Die Kenningar der Skalden. Ein Beitrag zur skaldischen Poetik. Bonn – Leipzig. Melefors, E. 2014. Gutnal þing och Gutnalia än en gång. In Gutagåtor. Historiska problem och tolkningar. (Ed.) P. Stobaeus. Visby, 211–230. Meletinskij, E. 1977. Scandinavian Mythology as a System of Oppositions. In Patterns in Oral Literature. (Eds.) H. Jason & D. Segal. Paris, 251–260. Meskell, L. 2005. Objects in the Mirror Appear Closer Than They Are”. In Materiality. (Ed.) D. Miller. Durham – London, 51–71. Meulengracht Sørensen, P. 1986. Thor’s Fisching Expedition. In Words and Objects: Towards a Dialogue between Archaeology and History of Religion. (Ed.) G. Steinsland. Oslo, 257–278. ———. 1991a. Om eddadigtenes alder. In Nordisk hedendom: Et Symposium. (Eds.) Gro Steinsland et al. Odense, 303–319. ———. 1991b. Håkon den gode og guderne: Nogle bemærkninger om religion og centralmagt i det tiende århundrede– og om religionshistorie og kildekritik. In Fra Stamme til Stat i Danmark II. Høvdingesamfund of kongemagt, pp. 235–244. (Eds.) P. Mortenssen & B. M. Rasmussen. Århus, 235–244. ———. 1993. Saga and Society: An Introduction to Old Norse Literature. Odense. ———. 2003. The hall in Norse Literature. In Borg in Lofoten: A chieftain’s farm in North Norway. (Eds.) Stamsø Munch, G. et al. Trondheim, 265–272. Meyer, B. et al. 2010. The origin and mission of Material Religion. Religion (40:3), 207–211. 592 Bibliography

Miller, D. 2005. Materiality: An Introduction. In Materiality. (Ed.) D. Miller. Durham – London, 1–50. ———. 2010. Stuff. Cambridge. Moberg, L. 2002. Gödåker som språkligt problem. Namn och bygd (90), 45–52. Møllenhus, K.R. 1953. En germansk gullmedaljong fra Inderøy. In Det Kongelige Norske Videnskabers Selskab. Museet. Årsbok, 73–79. Mogk, E. 1923. Novellistiche Darstellung mythologischer Stoffe Snorris und seiner Schule. FF Communications No. 51. Helsinki. ———. 1932. Zur Bewertung der Snorra-Edda als religionsgeschichtliche und mytholo- gische Quelle des nordgermanischen Heidentums. Berichte über die Verhandlungen der Sächsischen Akademie der Wissenschaften zu Leipzig. Phil.-hist. Klasse 84. Band. 1932. 2 Heft. Leipzig. Moltke, E. 1985 (1976). Runes and their Origin: Denmark and Elsewhere. Orig. title Runerne i Danmark og deres oprindelse. Trans. P. Foote. Copenhagen. Monikander, A. 2010. Våld och vatten: Våtmarkskult vid Skedemosse under järnåldern. Stockholm Studies in Archaeology 52. Stockholm. Montgomery, J. E. 2000. Ibn Fadlan and the Russiyyah. Journal of Arabic and Islamic Studies (3), 1–25. Motz, L. 1996. Kingship and the Giants. Arkiv för nordisk filologi (111), 73–88. ———. 1997. The Faces of the Goddess. New York & Oxford. Müller-Wille, M. 1970. Bestattung im Boot: Studien zu einer nordeuropäischen Grabsitte. Offa. Berichte und Mitteilungen aus dem Schleswig-Holsteinischen Landesmuseum. 25/26. 1968/69. Neumünster. ———. 1992. Monumentale Grabhügel der Völkerwanderungszeit in Mittel- und Nordeuropa. Bestand und Deutung. In Mare Balticum: Beiträge zur Geschichte des Ostseeraums in Mittelalter und Neuzeit. Festschrift zum 65. Geburtstag von Erich Hoffmann. Kieler Historische Studien. Band 36. (Ed.) W. Paravicini. Sigmaringen, 1–20. ———. 2002. Offerplatser på kontinenten: Några exempel från förkristen tid. In Plats och praxis: Studier av nordisk förkristen ritual. Vägar till Midgård 2. (Eds.) K. Jennbert, A. Andrén & C. Raudvere. Lund, 135–166. Munch G. Stamsø. 2003a. Jet, amber, bronze, silver and gold artefacts. In Borg in Lofoten: A chieftain’s farm in North Norway. (Eds.) Munch, G. Stamsø et al. Trondheim, 241–252. ———. 2003b. Borg as a pagan centre. In Borg in Lofoten: A chieftain’s farm in North Norway. (Eds.) Munch, G. Stamsø et al. Trondheim, 253–263. Munch G. Stamsø et al. (eds.) 2003. Borg in Lofoten: A chieftain’s farm in North Norway. Trondheim. Mundal, E. 1974. Fylgjemotiva i norrøn litteratur. Skrifter for nordisk språk og litteratur ved universitetene i Bergen, Oslo, Trondheim og Tromsø. Oslo. Bibliography 593

———. 2001. Holdninga til erotikk i norrøn diktning. In Kjønn—erotikk—religion. Bergen Museums skrifter. (Eds.) E. Ådland & K. Bang, Bergen, 28–40. ———. 2004. Edda og skladedikting. In Handbok i norrøn filologi. (Ed.) O. E. Haugen. Bergen, 215–266. ———. 2008. Introduction”. In Oral Art Forms and their Passage into Writing. (Eds.) E. Mundal & J. Wellendorf. Copenhagen, 1–5. Myhre, B. 1992a. Diskusjonen om ynglingeættns gravplasser. In Osebergdronningens grav: Vår arkeeologiske natsjonalskatt i nytt lys. (Eds.) A. E. Christensen et al. Oslo, 35–50. ———. 1992b. Kronologispørsmålet og ynglingeættens gravplasser. In Osebergdron ningens grav: Vår arkeeologiske natsjonalskatt i nytt lys. (Eds.) A. E. Christensen et al. Oslo, 272–278. ———. 2005. Krossane på Ullandhaug, Døds-sjødno på Sele og Fem dårlige jomfruer på Norheim: Symboler for Yggdrasil—livets tre? In Frá haug ok heiðni (2005/3), 3–10. Myrberg, N. 2009. An Island in the Middle of an Island: On cult, laws and authority in Viking Age Gotland. In From Ephesos to Dalecarlia: Reflections on Body, Space and Time in Medieval and Early Modern Europe. (Eds.) E. Regner et al. Stockholm, 101–118. Näsman, U. 1994. Liv och död: Sydskandinaviska grav- och offerriter från 200 till 1000 e.Kr. In Myte og ritual i det førkristne Norden: Et Symposium. (Eds.) J. P. Schjødt et al. Odense, 73–94. Näsström, B.-M. 1995. Freyja—the Great Goddess of the North. Lund Studies in History of Religions 5. Department of History of Religions. University of Lund. Lund. ———. 1996. Torgerd Holgabrud och Pytonsanden. Religion och Bibel. Nathan Söderblom-Sällskapets Årsbok. 1996 (55). Uppsala, 104–115. ———. 2001a. Blot: Tro og offer i det førkristne Norden. Oslo. ———. 2001b. Fornskandinavisk religion: En grundbok. Lund. ———. 2002. Blot—tro och offer i det förkristna Norden. Stockholm: ———. 2003. Quellheiligtümer und Quellkult. §5. Norden. In Reallexikon der Germa nischen Altertumskunde. Band 24. (Eds.) H. Beck et al. Berlin – New York, 26–29. ———. 2005. Romersk religion. Lund. ———. 2009. Nordiska Gudinnor: Nytolkningar av den förkristna mytologin. Stockholm. Naumann, H. 1938. Der König und die Seherin. Zeitschrift für deutsche Philologie (63), 347–358. Naumann, H.-P. 1984. Disen. In Reallexikon der Germanischen Altertumskunde 5. (Eds.) H. Beck et al. Berlin – New York, 494–497. Neckel, G. 1908. Beiträge zur Eddaforschung: Mit Excursen zur Heldensage. Dortmund. ———. 1913. Walhall: Studien über germanischen Jenseitsglauben. Dortmund. ———. 1920. Die Überlieferungen vom Gotte Balder. Dortmund. Ney, A. 2004. Drottningar och sköldmör: Gränsöverskridande kvinnor i medeltida myt och verklighet ca 400– 1400. Stockholm. 594 Bibliography

———. 2006. The Edges of the Old Norse World-View. A Bestiary Concept. In Old- Norse Religion in Long-Term Perspective: Origins, Changes, and Interactions. Vägar till Midgard 8. (Eds.) A. Andren, K. Jennbert & C. Raudvere, Lund, 63–67. Nielsen, A.-L. 1996. Hedniska kult- och offerhandlingar i Borg. In Religion från stenålder till medeltid. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter nr 19. (Eds.) K. Engdahl & A. Kaliff. Linköping, 89–104. ———. 2006. Rituals and power. About small buildings and animal bones from late Iron Age. In Old Norse Religion in Long-Term Perspectives: Origins, Changes, and Interactions: an international conference in Lund, Sweden, June 3–7, 2004, (eds.) A. Andrén, K. Jennbert, C. Raudvere, Lund, 243–247. Nikolaysen, N. 1881–1891. Kunst og Haandværk. Kristiania. Nilsson, B. 1989. De Sepulturis: Gravrätten i Corpus Iuris Canonici och i medeltida nordisk lagstiftning. Bibliotheca Theologiae Practicae. Kyrkovetenskapliga studier 44. Stockholm. ———. 1991. Frids- och asylföreskrifter rörande den medeltida sockenkyrkan. In Kyrka och socken i medeltidens Sverige. Studier till det medeltida Sverige 5. Riksantikvarieämbetet. (Ed.) O. Ferm. Stockholm, 473–504. ———. 1992. Till frågan om kyrkans hållning till icke-kristna kultfenomen: Attityder under tidig medeltid i Europa och Norden. In Kontinuitet i kult och tro från vikingatid till medeltid. Projektet Sveriges Kristnande. Publikationer. 1. (Ed.) B. Nilsson. Uppsala, 9–47. ———. 2000. Recension: Uppsalatemplet granskat på nytt. Henrik Janson, Templum Nobilissimum. Adam av Bremen, Uppsalatemplet och konfliktlinjerna i Europa kring år 1075. (Avhandlingar från Historiska institutionen i Göteborg. 21.) Göteborg 1998. Historisk tidskrift (1/2000), 105–111. Ninck, M. 1935. Wodan und germanischer Schicksalsglaube. Darmstadt. Nordahl, E. 1993. Södra Kungsgårdsplatån. Utgrävningarna 1988–1991. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. OPIA. 7. Societas Archaeologica Upsaliensis. (Ed.) W. Duczko. Uppsala, 59–63. ———. 1996. . . . templum quod Ubsola dicitur . . . i arkeologisk belysning. Aun 22. Med bidrag av Lars Gezelius och Henrik Klackenberg. Uppsala. ———. 2001. Båtgravar i Gamla Uppsala: Spår av en vikingatida högreståndsmiljö. Aun 29. Med bidrag av Anita Malmius, Petra Molnar, Anna Kjellström och Bengt Schönbäck. Uppsala. Nordberg, A. 2003. Krigarna i Odins sal: Dödsföreställningar och krigarkult i fornnordisk religion. Stockholm. ———. 2006a. Ull und Ullin. In Reallexikon der Germanischen Altertumskunde. Band 31. (Eds.) H. Beck et al. Berlin – New York, 405–408. ———. 2006b. Jul, disting och förkyrklig tideräkning: Kalendrar och kalendariska riter i det förkristna Norden. Acta Academiae regiae Gustavi Adolphi. XCI. Uppsala. Bibliography 595

———. 2008. Döden är ett kosmiskt drama: Religiös symbolik i gravarkitektur på gravfältet i Sylta. In Hem till Jarlabanke: Jord, makt och evigt liv i östra Mälardalen under järnåldern och medeltid. (Eds.) M. Olausson et al. Lund, 248–272. ———. 2010. Kultplatser, helgedomar och heliga platser. In Makt, kult och plats. Rapport från arkeologiska seminarier vid Stockholms läns museum. (Eds.) P. Bratt & R. Grönwall. Stockholm, 16–22. ———. 2011. Vad är en kultplats? Fornvännen (106), 215–231. ———. 2012. Continuity, Change and Regional Variation in Old Norse Religion. In More than Mythology: Narratives, Ritual Practices and Regional Distribution in pre- Christian Scandinavian Religion. (Eds.) C. Raudvere & J. P. Schjødt. Lund, 119–151. ———. 2013. Fornnordisk religionsforskning mellan teori och empiri. Kulten av anfärder, solen och vegetationsandar i idéhistorisk belysning. Acta Academiae regiae Gustavi Adolphi. CXXVI. Uppsala. ———. 2014. Kultplatser och tolkningsperspektiv i fornnordisk religionsforskning. In Den heliga platsen. Handlingar från symposiet Den heliga platsen Härnösand 15–18 september 2011. (Eds.) E. Nyman et al. Sundsvall, 11–22. Nordberg, A. & O. Sundqvist. 2008. Förkristen jul: Kalender och kult. din—Tidskrift for religion og kultur (2008/4), 12–25. Nordeide, S. W. 2005. Fra Odin til Kvitekrist. SPOR (1), 18–21. ———. 2011. The Viking Age as a Period of Religious Transformation. The Christianization of Norway from AD 560–1150/1200. Studies in Viking and Medieval Scandinavia 2. Turnhout. ———. 2012. Kristningen av Trøndelag—ikke uten motstand? In Människor i vikingatidens Mittnorden: Föredrag vid de Mittnordiska Arkeologidagarna i Östersund 2010. (Ed.) O. Hemmendorff (red.) Östersund, 75–91. Nordén, A. 1938. Le problème des ‘bonhommes en or’. Acta Archaeologica (9), pp. ———. 1892. Mytiska beståndsdelar i Ynglingatal. Uppsalastudier tillägnade Sophus Bugge påhans 60 -åriga födelsedag den 5. januari 1893. Uppsala, 194–225. Noreen, A. 1925. See Þjóðólfr inn fróði ór Hvini. Norr, S. 1995. Ingjald Illráði—Nice Chap or False, Murderous Swede? Tor (27), 487–505. ———. 1998. To Rede and to Rown: Expressions of Early Scandinavian Kingship in Written Sources. OPIA. 17. Department of Archaeology and Ancient History. Uppsala University. Uppsala. Norr, S. & A. Sundkvist. 1995. Valsgärde Revisited: Fieldwork Resumed after 40 Years. Tor (27), 395– 417. Nyberg, T. 1984. Stad, skrift och stift: Några historiska inledningsfrågor. In Adam av Bremen: Historien om Hamburgstiftet och dess biskopar. Trans. E. Svenberg. Comm. by C. F. Hallencreutz, K. Johannesson, T. Nyberg, A. Piltz. (Ed.) B. Sahlin. Stockholm, pp. 295–339. 596 Bibliography

Nylén, E. & J. P. Lamm. 1987. Bildstenar. Stockholm. Nylén, E. & B. Schönbäck. 1994. Tuna i Badelunda: Guld—Kvinnor—Båtar. Med bidrag av M. Nockert.Västerås kulturnämnds skriftserie 27. Västerås. Nyman, E. 2002. Nidaros. §1. Namenkundliches. In Reallexikon der Germanischen Altertumskunde. Band 21. (Eds.) H. Beck et al. Berlin – New York. Ödman, A. 2003. Tankar kring två ringhantag från Uppåkra. In Fler fynd i centrum. Materialstudier i och kring Uppåkra. Uppåkrastudier. 9. (Ed.) B. Hårdh. Lund, 89–96. Ogilvie, R. M. 1994 (1969). The Romans and their Gods. London. Ohlmarks, Å. 1936. Isländska hov och gudahus. In Bidrag till nordisk filologi tillägnade E. Olsson. Lund, 339–355. Ólafur Briem. 1956. Arveøl. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 1. Malmö, cols. 267–268. Ólafur Lárusson. 1942. Kultminne i stadnamn. Island. Nordisk kultur: Religions historie. (Ed.) N. Lid. Stednavn. Stockholm – Oslo – København, 74–79. ———. 1960. Goði og Goðorð. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 5. Malmö, cols. 363–366. Olausson, M. (ed.) 2011. Runnhusa—bosättningen på berget med de många husen. Skrifter från projektet Runsa borg, Eds socken, Uppland nr 1. Stockholm. ———. 2014. The Great House at Runsa—A House of Representation. In Runsa Borg. Representative Life on a Migration Period Hilltop Site—a Scandiavian Perspective. Papers from the project Runsa borg, Uppland. No 2. (Ed.) M. Olausson. Östersund, 145–166. Oldenberg, H. 1970 (1917). Die Religion des Veda. Darmstadt. Olivecrona, K. 1942. Döma till konung. En rättshistorisk undersökning. Lund. Olrik, A. 1909. At sidde på höj: Oldtidens konger og oldtidens thulir. Danske studier (6), 1–10. ———. 1910. Irminsul og gudestøtter. Maal og minne, 1–9. Olrik, A. & H. Ellekilde 1925–1951. Nordens gudeverden. Copenhagen. Olsen, M. 1909. Fra gammalnorsk myte og kulthus. Maal og minne, 17–36. ———. 1912. Kung Orre: Et bidrag til gammel nordisk folketro. Maal og minne (1912), 1–26. ———. 1915. Hedenske kultminder i norske stedsnavne. Skrifter utg. av det norske videnskaps-akademi i Oslo, hist.-filos. Kl. 1914, 2:4. Oslo. ———. 1926. Ættegård og helligdom: Norske stednavn sosialt og religionshistorisk belyst. Instituttet for sammenlignende kulturforskning, ser. A: Forelesninger. 9A. Oslo. ———. 1938 (1931). Valhall med de mange dører. Norrøne Studier. Oslo. ———. 1942. Kultminne i stadnamn. 3. Norge. In Religionshistorie. Nordisk kultur. XXVI. Oslo. Olsen, M. & H. Sheteling. 1909. De to runestener fra Tu og Klepp paa Jærderen. Bergens Museums Aarbog (11). Bergen, 3–29. Bibliography 597

Olsen, O. 1966. Hørg, hov og kirke: Historiske og arkæologiske vikingetidsstudier. København. ———. 1969. In Lidén 1969. ———. 1996. “ ‘Hørg, hov og kirke 30 år efter’ ”. In Møtet mellom hedendom og kristendom i Norge. Red. H.-E. Lidén. Oslo, 121–128. Olsson, G. & S. Berglund. 1993. Mora sten och Mora stenar: En vägledning till ett märkligt nationalmonument. Utg. Riksanti kvarieämbetet i samarbete med Lagga och Danmarks hembygdsföreningar. Stockholm. Olsson, I. 1996. Gotländska ortnamn. Stockholm. Olsson, M. 1912. En forntida brunn i Gamla Uppsala. In Upplands fornminnesförenings tidskrift (28), 307–321. Olsson, T. 1985. Gudsbildens gestaltning: litterära kategorier och religiös tro. Svensk religionshistorisk årsskrift (1), 42–63. OÖD = Ordbok över folkmålen i övre Dalarna. 1–. (Eds.) L. Levander et al. Uppsala. ONP = Ordbog over det norrøne prosasprog/A Dictionary of Old Norse Prose. 1–. 1995–. (Eds.) H. Degnbol et al. København. Orri Vésteinsson. 2001. The Conversion of the Icelanders. In Europe around the Year 1000. (Ed.) P. Urbanczyk. Warszawa, 325–342. ———. 2007. ‘Hann reisti hof mikið hundrað fóta langt . . .’: Um uppruna hof-örnefna og stjórnmál á Íslandi á 10. öld. Saga: Tímarit Sögufélags. XLV:1. (Eds.) Guðmundar J. Guðmundsson & Páll Björnsson. Reykjavík, 53–91. ———. 2009. Ethnicity and class in the settlement-period Iceland. In The Viking Age: Ireland and the West. (Eds.) J. Sheehan & D. Ó Corráin. Dublin, 494–510. Østmo E. & L. Hedeager. 2005. Norsk arkeologisk leksikon. Oslo. Otterbjörk, R. 1983. Starka och svaga kvinnonamn i fornsvenskan. Namn och bygd (71), 101–114. Otto, R. 1917. Das Heilige—Über das Irrationale in der Idee des Göttlichen und sein Verhältnis zum Rationalen. München. Oxford Latin Dictionary. 1996. Oxford. von Padberg, L. E. 1995. Mission und Christianisierung: Formen und Folgen bei Angelsachsen und Franken im 7. Und 8. Jahrhundert. Stuttgart. Paden, W. E. 1992. Interpreting the Sacred: Ways of Viewing Religion. Boston. ———. 1994 (1988). Religious Worlds: The Comparative Study of Religion. Boston. ———. 1996. ‘Elements of a new comparativism’, Method & Theory in the Study of Religion 8–1, 5–14. Page, R. I. 1995. Chronicles of the Vikings. Records, Memorials and Myths. London. ———. 1992. Runor och regionalitet: Studier av variation i de nordiska minnesinskrift- erna. Runrön 7. Runologiska bidrag utg. av institutionen för nordiska språk vid Uppsala universitet. Uppsala. Palm, R. 2004. Vikingarnas språk 750–1100. Stockholm. 598 Bibliography

Palm, T. 1942. Uppsalalunden och Uppsalatemplet. Vetenskaps-Societeten i Lund. Årsbok. 1941. Lund, 79– 109. ———. 1948. Trädkult: Studier i germansk religionshistoria. Skrifter utgivna av Vetenskaps-Societeten i Lund, 33. Lund. Palme, S. U. 1959. Kristendomens genombrott i Sverige: Den stora omvälvningen i svensk historia: vikingatågen, missionen och det svenska rikets uppkomst. Stockholm. Patton K. C. & B. C. Ray (eds.). 2000. A Magic Still Dwells: Comparative Religion in the Postmodern Age. Berkeley – Los Angeles – London. Peel, P. 1999. See Guta saga. The History of the Gotlanders. Perkins, R. 1994. Eyrbyggja saga. In Reallexikon der Germanischen Altertumskunde, Band 8. (Eds.) H. Beck et al. Berlin – New York, 49–57. ———. 2001. Thor the Wind-Raiser and the Eyrarland Image. Viking Society for Northern Research & University College London. London. Petersen, P. V. 1994. “Excavations at Sites of Treasure Trove Finds at Gudme.”In Samfundsorganisation og Regional Variation. Norden i romersk jernalder og folkevandringstid. Beretning fra 1. nordiske jernaldersymposium på Sandbjerg Slot 11–15 april 1989. Jysk Arkæologisk Selskabs Skrifter. XXVII. (Eds.) C. Fabech & J. Ringtved. Aarhus, 30–40. Peterson, L. 1981. Kvinnonamnens böjning i fornsvenskan: De ursprungligen starkt böjda namnen. Anthroponymica Suecana. 8. Uppsala. ———. 1993. Namnformen uifin på Oklundahällen—nominativ eller ackusativ? Studia Anthroponymica Scandinavica (11), 33–40. ———. 1994 (1989). Svenskt runordsregister. Runrön. 2. Runologiska bidrag utgivna av Institutionen för nordiska språk vid Uppsala universitet. Uppsala. ———.2004. Lexikon över urnordiska personnamn. See: http://www.sofi.se. ———. 2007. Nordiskt runnamnslexikon. Uppsala. Petré, B. 1984. Arkeologiska undersökningar på Lovö. Part 4. Bebyggelsearkeologisk analys, Acta Universitatis Stockholmiensis Studies in North-European Archaeology 10, Stockholm: Almqvist och Wiksell. Phillpotts, B. S. 1912–1913. Temple-Administration and Chieftainship in Pre-Christian Norway and Iceland. Saga-Book of the Viking Society (8), 264–284. ———. 1920. The Elder Edda and Ancint Scandinavian Drama. Cambridge. Platvoet, J. 1995. Ritual in plural and pluralist societies: Instrument for analysis. In Pluralism and identity: Studies in ritual behavior. (Eds.) J. Platvoet and K. van der Toorn. Leiden, 25–51. Polomé, E. C. 1995. Freyr. In Reallexikon der Germanischen Altertumskunde. Band 9. (Eds.) Heinrich Beck et al. Berlin – New York, 587–594. Poulsen. B. 2012. If the Danish vikings knew land rent, what was it like? In Settlement and lordship in Viking and Early Medieval Scandinavia. The Medieval Countryside 9. (Eds.) B. Poulsen &. S. M. Sindbæk. Turnhout, 277–291. Bibliography 599

Price, N. S. 2002. The Viking Way: Religion and War in Late Iron Age Scandinavia. Aun. 31. Uppsala universitet. Uppsala. Price, S. R. F. 1993 (1985). Delphi and divination. In Greek Religion and Society. (Eds.) P. E. Easterling & J. V. Muir. Cambridge, 128–154. ———. 1994 (1984). Rituals and Power: The Roman Imperial Cult in Asia Minor. Cambridge. Puhvel, J. 1975. Remus et frater. History of Religions (15), 146–157. Quinn, J. 2000. From Orality to Literacy in Medieval Iceland”. In Old Icelandic Literature and Society. (Ed.) M. Clunies Ross. Cambridge, 30–60. ———. 2005. Women in Old Norse Poetry and Sagas. In A Companion to Old Norse- Icelandic Literature and Culture. (Ed.) R. McTurk. Oxford, 518–535. Radcliffe-Brown, A. R. 1952. Structure and Function in Primitive Society: Essays and Adresses. London. ———. 1979. Taboo. In Reader in comparative religion: An anthropological approach (fourth ed.). (Eds.) W. A. Lessa & E. Z. Vogt. ??, 46–56. Rahmqvist, S. 1986. Gamla Uppsala by—Upplands största. In Från Östra Aros till Uppsala: En samling uppsatser kring det medeltida Uppsala. Uppsala stads historia 7. Uppsala, 254–271. ———. 2000. Gamla Uppsala under medeltiden. In Myt, makt och människor: Tio uppsatser om Gamla Uppsala. (Ed.) G. Friberg. Stockholm, 47–49. Ramqvist, P. H. 1987. Mellannorrland under äldre järnålder: Några aspekter under äldre järnålder. Bebyggelsehistorisk tidskrift (14), 105–126. ———. 1991. Perspektiv på regional variation och samhälle i Nordens folkvandringstid. In Samfundsorganisation og Regional Variation: Norden i romersk jernalder og folkevandringstid. Beretning fra 1. nordiske jernaldersymposium på Sandbjerg Slot 11–15 april 1989. Jysk Arkæologisk Selskabs Skrifter. XXVII. (Eds.) C. Fabech & J. Ringtved. Aarhus, 305–318. ———. 1995. Resources, Population and Power in the Migration Period: An Analysis of the Archaeological Remains in Central Norrland. In Archaeology East and West of the Baltic: Papers from the Second Estonian-Swedish Archaeological Symposium. Sigtuna, May 1991. Estonian-Swedish Archaeological Symposium. 2. Theses and Papers in Archaeology N.S.A. 7. Department of Archaeology, Stockholm University. (Ed.) I. Jansson. Stockholm, 33–38. Rappaport, R. A. 2000 (1999). Ritual and Religion in the Making of Humanity. Cambridge Studies in Social and Cultural Anthropology. 110. Cambridge. Ratke S. 2009. Guldgubber—Einblicke in die Völkerwanderungszeit. Ph.D. dissertation, University of Bonn. Ratke S. & R. Simek. 2006. Guldgubber: Relics of Pre-Christian law rituals? In Old Norse Religion in Long- Term Perspectives: Origins, Changes, and Interactions: an international conference in Lund, Sweden, June 3–7, 2004. (Eds.) A. Andrén, K. Jennbert, C. Raudvere, Lund, 259–266. 600 Bibliography

Raudvere, C. 2002. Mellan liv och text: Ordets makt som litterärt motiv och rituell praktik i norrön litteratur. In Plats och praxis: Studier av nordisk förkristen ritual. Vägar till Midgård 2. (Eds.) K. Jennbert, A. Andrén & C. Raudvere. Lund, 31–68. ———. 2003. Kunskap och insikt i norrön tradition: Mytologi, ritualer och Troll domsanklagelser. Vägar till Midgård 3. Lund. Raudvere, C. & J. P. Schjødt (eds.). 2012. More than Mythology: Narratives, Ritual Practices and Regional Distribution in Pre-Christian Scandinavian Religion. Lund. Reichert, H. 2003. Vier Miszellen zum Urgermanischen und ‘Altrunischen’. Zeitschrift für deutsches Altertum 132, 355–356. Reifegerste, E. M. 1989. Die Hervarar Saga: Eine kommentierte Übersetzung und Untersuchungen zur Herkunft und Integration ihrer Überlieferungsschichten. Altnordische Bibliothek. 6. Köln – Leverkusen. Riesebrodt, M. 2007. Cultus und Heilsversprechen. Eine Theorie der Religion. München. Ringgren, H. 1989. Israels religion. Stockholm. Ringgren, H. & Å. V. Ström. 1984 (1957). Religionerna i historia och nutid. Stockholm. Ringstad, B. 1987. De store gravminnene—et maktideologisk symbol? Viking (50), 65–78. ———. 1991. Graver og ideologi: Implikasjoner fra vestnorsk folkevandringstid. In Samfundsorganisation og Regional Variation: Norden i romersk jernalder og folkevandringstid. Beretning fra 1. nordiske jernaldersymposium på Sandbjerg Slot 11–15 april 1989. Jysk Arkæologisk Selskabs Skrifter. XXVII. (Eds.) C. Fabech & J. Ringtved. Aarhus, 141–150. Ringsted, N. 2005. Karsvik platåhus—ett märkligt fornminne. Brommaboken: Bromma hembygdsförenings Årsskrift. Bromma. Robberstad, K. 1950. Frå gama log ny rett. Oslo. ———. 1967. Odelsrett. Norge. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 12. Malmö, cols. 493–499. Roesdal, E. 1982. Viking Age Denmark. London. Rønne, P. 2002. Hus på hus i Melhus. SPOR (17), 21–23. ———. 2011. Horg, hov and ve—a pre-Christian cult place at Ranheim in Trøndelag, Norway, in the 4rd–10th centuries AD. Adoranten (2011), 1–14. Rosén, J. 1939. Striden mellan Birger Magnusson och hans bröder. Studier i nordisk politisk historia 1302–1319. Lund. ———. 1964. Krongods. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 9. Malmö, cols. 434–438. Røskaft, M. 1997. Kultkontinuitet og tidlig kirkebygging i Trøndelag. In Helgonet i Nidaros: Olavskult och kristnande i Norden. (Ed.) L. Rumar. Stockholm, 225–239. ———. 2003. Maktens landskap: Sentralgårdar i Trøndelag ved overgangen fra vikingtid til kristen middelalder, ca. 800–1200, No. 39 Skriftserie fra Historisk institutt, Trondheim: Norges teknisk- naturvitenskaplige universitet. Bibliography 601

Rostvik, A. 1967. Har och harg. Acta Academiae Regiae Gustavi Adolphi. XLIV. Studier till en svensk ortnamnsatlas utgivna av Jöran Sahlgren. 11. Uppsala. Røthe, G. 1997. Fra Olav Geirstadalv til Olav den helige. En religionshistorisk kommentar til tåten om Olav Geirstadalv. In Gokstadhøvdingen og hans tid. (Ed.) T. Frost. Sandefjord, 25–42. ———. 2007. Þorgerðr Hölgabrúðr—the fylgja of the Háleygjar family. Scripta Islandica (57), 33–55. Rousell, A. 1943. Islands gudehove. In Forntida gårdar i Island. (Ed.) M. Stenberger. København, 215–223. Rüpke, J. 1996. Controllers and Professionals: Analyzing Religious Specialists. Numen. Vol. XLIII. (Eds.) H. G. Kippenberg & E. T. Lawson. Leiden, 241–262. ———. 2001. Die Religion der Römer: Eine Einführung. München. Ruthström, B. 1988. Oklunda-ristningen i rättslig belysning. Arkiv för Nordisk Filologi (103), 64–75. ———. 1990. Forsaringen—vikingatida vi-rätt? Arkiv för Nordisk Filologi (105), 41–55. Rydh, H. 1920. Den magiska ringen från Hamrånge. Meddelanden av Gästriklands Kulturhistoriska förening, 3–6. ———. 1936. Förhistoriska undersökningar på Adelsö. KVHAA. Stockholm. Rydving, H. 1990. Scandinavian–Saami Religious Connections in the History of Research. In Old Norse and Finnish Religions and Cultic Place-Names, Scripta instituti Donneriani Aboensis, XIII. (Ed.) Tore Ahlbäck (Stockholm: Almqvist & Wiksell, 1990), 358–373. ———. 1993. The End of the Drum-Time: Religious Change among the Lule Saami, 1640s–1740s. Acta Universitatis Upsaliensis. Historia Religionum 12. Uppsala. ———. 1996. Suttungamjödet, vishetens källa: Öl och mjöd i Skandinavien under vikingatid och medeltid. In Rus och rit: Om droger och religioner. Religionshistoriska forskningsrapporter från Uppsala. 9. (Ed.) H. Rydving. Uppsala, 155–162. ———. 2005. Review-article on Sundqvist 2002. Namn och bygd (93), 180–182 ———. 2006. Constructing Religious Pasts: Summary Reflections. In The Making of Christian Myths in the Periphery of Latin Christendom (c. 1000–1300). (Ed.) L. B. Mortensen. Copenhagen, 315– 321. ———. 2011a. The ‘Bear Ceremonial’ and Bear Rituals among the Khanty and the Sami. Temenos (46,1) (2010), 31–52. ———. 2011b. Le chamanisme aujord’hui: Constructions et deconstructions d’une illu- sion scientifique. Études mongoles et sibériennes, centrasiatiques et tibétaines 42 (1911), 1–13. Rydving, H. & R. Kristoffersson. 1993. Några samiska offerplatser. Fornvännen (88), 195–210. Sabourin, L. S. J. 1973. Priesthood: A Comparative Study. Leiden. 602 Bibliography

Særheim, I. 1989. Litt om eldre busetjingsnamn i Klepp. Frá haug ok heiðni (2), 213–224. ———. 2015. Stadnamn og sentralitet i eit jærsk jordbrukssamfunn frå jenralderen. In Et Akropolis på Jæren? Tinghaugplatået gjennom jernalderen. AmS-Varia 55. Eds E.S. Kristoffersen, M. Nitter og E. S. Pedersen. Stavanger, 49–62. Sahlgren, J. 1953. Från Morasten till Håga. Namn och Bygd (41), 1–45. Salberger, E. 1980. Oklunda-hällens runristningar. Östgötska runstudier: Scripta runica. (Ed.) E. Salberger. Göteborg 1, 2–23. Saler, B. 1993. Conceptualizing Religion: Immanent Anthropologists, Transcendent Natives, & Unbounded Categories. Studies in the History of Religions (Numen Bookseries), Vol. LVI. Leiden – New York – Köln. Van der Sanden, W. & T. Capelle (2001). Mosens Guder: Antropomorfe træfigurer fra Nord- og Nordvesteuropas fortid/Immortal Images: Ancient anthropomorphic wood carvings from northern and northwest Europe. Silkeborg. Sandnes, J. 1964. Gno. hörgr i gardsnavn. Maal og minne, 113–120. ———. 1967. Trøndelags eldste politiske historie. Historisk Tidsskrift (46), 1–20. ———. 1971. Ødetid og gjenreisning: Trøndsk busetningshistorie ca 1200–1600. Oslo. ———. 1987. Hedensk kulthus og kristen kirke. In Festskrift til Alfred Jakobsen. (Eds.) J. R. Hagland, J. T. Faarlund & J. Rønhovd. Trondheim, 144–152. ———. 1992. Haupttypen sakraler Ortsnamen Westskandinaviens Anhang: Karte der sakraler Ortnamen in Rogaland (Stavanger Amt nach M. Olsen). In Der historische Horizont der Götterbild-Amulette aus der Übergangsepoche von der Spätantike zum Frühmittelalter. (Ed.) K. Hauck. Göttingen. ———. 2001. Lade. In Reallexikon der Germanischen Altertumskunde. Band 17. (Eds.) H. Beck et al. Berlin – New York, 541–542. Sandnes, J. & O. Stemshaug. 2007. Norsk stadnamnleksikon. Oslo. Sanmark, A. 2004. Power and Conversion: A Comparative Study of Christianization in Scandinavia. OPIA. 34. Uppsala. ———. 2005. Dietary Regulations in Early Christian Norway. Viking and Medieval Scandinavia (1). Turnhout, 203–224. Sanmark, A. & S. Semple. 2008. Places of assembly : new discoveries in Sweden and England. In Fornvännen (103), 245–259. Santesson, L. 1989. En blekingsk blotinskrift: En nytolkning av inledningsraderna på Stentoftenstenen. Fornvännen (84), 221–229. ———. 1993. Eine Blutopferinschrift aus dem südschewedischen Blekinge: Eine Neudeutung der einleitenden Zeilen des stentofter Steines. Frühmittelalterliche Studien (27), 241–252. Sawyer, B. 2000 (1992). Kvinnor och familj i det forn- och medeltida Skandinavien. Occasional Papers on Medieval Topics. 6. Göteborg. Bibliography 603

Sawyer B. & P. Sawyer. 1993. Medieval Scandinavia: From Conversion to Reformation, circa 800–1500. The Nordic Series. Vol. 17. Minneapolis – London. Sawyer, P. H. 1991. När Sverige blev Sverige. Orig. title The Making of Sweden. Trans. B. Sawyer. Occasional Papers on Medieval Topics. 5. Alingsås. ———. 2001. Krönung (Scandinavia). In Reallexikon der Germanischen Altertumskunde. Band 17. (Eds.) H. Beck et al. Berlin – New York, 384–385. Schäferdiek, K. 1986. Eigenkirchen. In Reallexikon der Germanischen Altertumskunde 6. (Eds.) by Heinrich Beck et al. Berlin – New York, 559–561. Schier, K. 1968. Freys und Fróðis Bestaattung. In Festschrift für Otto Höfler zum 65. Geburtstag. I. Band. Hrsg. von Helmut Birkhan und Otto Gschwantler. Wien, 389–409. ———. 1981. Zur Mythologie der Snorra Edda: Einige Quellenprobleme. In Specvlvm norroenvm: Norse Studies in Memory of Gabriel Turville-Petre. (Eds.) U. Dronke et al. Odense, 405–420. Schjødt, J. P. 1990. Det sakrale kongedømme i det førkristne Skandinavien. Chaos. Dansk-Norsk tidsskrift for religionshistoriske studier (13), 48–67. ———. 1991. Relationen mellem aser og vaner og dens ideologiske implikationer. In Nordisk hedendom symposium Granavolden. (Ed.) G. Steinsland. Odense. ———. 2001. Óðinn—Shaman eller fyrstegud? In Kontinuitäten und Brüche in der Religionsgeschichte: Festschrift für Anders Hultgård zu seinem 65. Geburtstag am 23.12.2001. Ergänzungsbände zum Reallexikon der germanischen Altertumskunde 31. (Eds.) M. Stausberg et al. Berlin – New York, 562–577. ———. 2008. Initiation Between Two Worlds: Structure and Symbolism in Pre-Christian Scandinavian Religion. Trans. Victor Hansen. The Viking Collection. Vol. 17. Odense. ———. 2009. Freyr and Fróði and some Reflections on Euhemerism. Analecta Septentrionalia. Beiträge zur nordgermanischen Kultur- und Literaturgeschichte. (Eds.) Wilhelm Heizmann et al. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 65. Berlin – New York, 567–579. ———. 2012. Reflections on aims and methods in the study of Old Norse religion. In More than Mythology. Narratives, ritual practices and regional distribution in pre- Christian Scandinavian religions. Eds. C. Raudvere & J. P. Schjødt, J. P. Lund, 263–287. Schlyter, C. J. 1887. Ordbok till Samlingen af Sweriges Gamla Lagar: Glossarium ad Corpus Iuris Sueo-Gotorum Antiqui. Lund. Schmid, T. 1931. Eskil, Botvid och David: Tre svenska helgon. Scandia (4), 102–114. Schmidt, A. 1998. Tabu. In Handbuch religionswissenschaftlicher Grundbegriffe. (Eds.) H. Cancik et al. Band 4. Stuttgart, 160–162. Schönbäck, B. & L. Thunmark-Nylén. 2002. De vikingatida båtgravarna vid Valsgärde— relativ kronologi. Fornvännen (97), 1–8. 604 Bibliography

Schramm, P. E. 1937. Geschichte des Englischen Königtums im Lichte der Krönung. Weimar. Schulte, M. 2006. Ein kritischer Kommentar zum Erkenntnisstand der Blekinge Inschriften. Zeitschrift für deutsches Altertum 135, 399–412. Schück, H. 1904. Studier i nordisk litteratur- och religionshistoria. Vols. I–II. Stockholm. ———. 1910. Sveriges förkristna konungalängd. Uppsala universitets årsskrift. 1910. Program. 6. Uppsala. Schweitzer, A. 1962. Ideological Strategies. Political Research Quarterly (15:1), 46–66. See, K. von. 1964. Altnordische Rechtswörter: Philologische Studien zur Rechtsauffassung und Rechtsgesinnung der Germanen. Tübingen. ———. 1988. Mythos und Theologie im skandinavischen Hochmittelalter. Skandinavi stische Arbeiten 8. Heidelberg. Segal, R. A. 1998. Introduction. In The Myth and Ritual Theory: An Anthology. (Ed.) R. A. Segal. Oxford, 1–13. ———. 2001. In Defense of the Comparative Method. Numen (48–3), 339–373. ———. 2005. Classification and Comparison in the Study of Religion: The Work of Jonathan Z. Smith. Journal of the American Academy of Religion (73–4), 1175–1188. Seiler, A. 1997. Ett litet brandgravfält i näreten av brandbåtgravarna: Arkeologisk delun- dersökning. Raä 26, Vendel, Vendels sn, Uppland. In SIV—Svealand i vendel—och Vikingatid: Rapport från utgrävningarna i Vendel. Stockholm, 56–92. ———. 2001. I skuggan av båtgravarna: Landskap och samhälle i Vendel socken under yngre järnålder. Del 1–2. Theses and papers in archaeology B:7–8. Stockholm. Shetelig, H. 1937. Islands graver og oldsaker fra vikingetiden. Viking: Tidskrift for norrøn arkeologi. Bind I, 205–219. Sigurður Nordal. 1942. Íslenzk menning 1. Reykjavík. Silver, K. 1996. Kungshögarna i Sverige. In Arkeologi och miljöekologi i Gamla Uppsala: Studier och rapporter. Vol. II. OPIA. 11. Department of Archaeology and Ancient History. Uppsala University. (Ed.) W. Duczko. Uppsala, 53–58. Simek, R. 2003. Religion und Mythologie der Germanen. Darmstadt. ———. 2006 (1993). Dictionary of Northern Mythology. Orig. title Lexikon der germanischen Mythologie. Trans. A. Hall. Cambridge. Sjöholm, E. 1988. Sveriges medeltidslagar: Europeisk rättstradition i politisk omvandling. Rättshistoriskt Bibliotek. 41. Stockholm. Skre, D. 2007a. Excavations of the Hall at Huseby. In Kaupang in Skiringssal: Kaupang Excavation Project. Publication Series. Vol. 1. Norske Oldfunn XXII. (Ed.) D. Skre. Oslo, 223–247. ———. 2007b. The Dating of Ynglingatal. In Kaupang in Skiringssal. Kaupang Exca vation Project. Publication Series. Vol. 1. Norske Oldfunn XXII. (Ed.) D. Skre. Oslo, 407–429. Bibliography 605

Skyllberg, E. 2008. Gudar och glasbägare—järnåldersgården i Lunda. In Gestalter och gestaltningar: Om tid, rum och händelser i Lunda. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter 72. Stockholm. Slupecki, L. P. 1993. Die Slawischen Tempel und die Frage des sakralen Raumes bei den Westslawen in vorchristlichen Zeiten. Tor (25), 247–298. Smedstad, I. 1993. Gravminner. In Kulturminner i Stjørdal, Frosta, Levanger, Verdal, Inderøy, Stenkjer. (Eds.) I. Smedstad & S. Krüger. Stiklestad, 33–52. Smith, J. Z. 1987. To Take Place: Towards Theory in Ritual. Chicago – London. ———. 2000. In Comparison a Magic Dwells. In A Magic Still Dwells: Comparative Religion in the Postmodern Age. (Eds.) K. C. Patton & B. C. Ray. Berkeley – Los Angeles – London, 23–44. ———. 2004. Relating religion: Essays in the study of religion. Chicago & London. Söderberg, B. 2003. Järnålderns Järrestad. In Järrestad: Huvudgård i centralbygd. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter nr 51. (Ed.) B. Söderberg. Stockholm, 109–174. ———. 2005. Aristokratiskt rum och gränsöverskridande: Järrestad och sydöstra Skåne mellan region och rike 600–1100. Riksantikvarieämbetet Arkeologiska undersök- ningar Skrifter. No. 62. Lund. Söderblom, N. 1913. Holiness. In Encyclopedia of Religion and Ethics 6. Edinburgh, 731–741. ———. 1914. Gudstrons uppkomst. Stockholm. SOL = Svenskt ortnamnslexikon. (Ed.) M. Wahlberg. Uppsala. Spiro, M. E. 1982. Buddhism and Society: A Great Tradition and its Burmese Vicissitudes. Berkeley. Springer, M. 2000. Irminsul. § 1. Historisches. In Reallexikon der Germanischen Altertumskunde. Band 15. (Eds.) H. Beck et al. Berlin – New York, 504–505. Spurkland, T. 2005. Norwegian runes and runic inscriptions. Woodbridge, Suffolk. Stausberg, M. 2009. Interventionist practices and the promises of religion. On Martin Riesebrodt, Cultus und Heilsversprechen (2007). In Contemporary Theories of Religion. A Critical Compainion. London, 264–282. ———. 2012. “A new theory of religion? Introducing the review symposium on Manuel A. Vásquez, More Than Belief (2011)”. Religion (42), 597–608. Steiner, F. 1956. Taboo. London. Steinsland, G. 1985a. Husfruer, gyðjer og volver. In Kvinnenes kulturhistorie. Vol. I. Fra Antikken til år 1800. (Eds.) K. Vogt et al. Oslo, 126–130. ———. 1985b. Kvinner og kult i vikingtid. In Kvinnearbeid i Norden fra vikingetiden til reformasjonen. (Ed.) R. Andersen. Bergen, 31–42. ———. 1990. De nordiske gullblekk med parmotiv og norrøn fyrsteideologi. Et tolkningsforslag. In Collegium Medievale I. Vol. 3, pp. 73–94. Oslo. 606 Bibliography

———. 1991. Det hellige bryllup og Norrøn Kongeideologi: En analyse av hierogami-my- ten i Skírnismál, Ynglingatal, Háleygjatal og Hyndluljóð. Oslo. ———. 1994. Eros og død i norrøn kongeideologi: Kan mytisk herskerideologi kaste lys over forestillinger og riter knyttet til død, begravelse og gravkult. In Myte og ritual i det førkristne Norden: Et Symposium. (Eds.) J. P. Schjødt et al. Odense, 141–157. ———. 1997. Eros og død i norrøne myter. Oslo. ———. 2000. Den hellige kongen: Om religion og herskermakt fra vikingtid til middelalder. Oslo. ———. 2005. Norrøn religion: Myter, riter, samfunn. Oslo. ———. 2011a. Introduction: Ideology and Power in the Viking and Middle Ages: Scandinavia, Iceland, Ireland, Orkney and the Faroes. In Ideology and Power in the Viking and Middle Ages: Scandinavia, Iceland, Ireland, Orkney and the Faeroes. (Eds.) G. Steinsland et al. Leiden, 1–14. ———. 2011b. Origin Myths and Rulership: From Viking Age Ruler to the Ruler of Medieval Historiography: Continuity, Transformations and Innovations. In Ideology and Power in the Viking and Middle Ages: Scandinavia, Iceland, Ireland, Orkney and the Faeroes. (Eds.) G. Steinsland et al. Leiden, 15–68. Steinsland, G. & P. Meulengracht Sørensen. 1998. Människor och makter i vikingarnas värld. Stockholm. ———. 1999. Völuspå. Oslo. Steinsland, G. & K. Vogt. 1981. “ ‘Aukinn ertu Uolse og vpp vm tekinn’: En religionshistorisk analyse av Völsaþáttr i Flateyjarbók. Arkiv för nordisk filologi (96), 87–106. Sten, S. & M. Vretemark. 1999. Kungshögarna i Gamla Uppsala, Gamla Uppsala socken i Uppland. Invnr. 1410, 1411, 5308, 5308b, 13821 och 18927. Rapportserie från samlingsen- heten Statens Historiska Museum. Osteologisk Rapport. 1999:1. Stockholm. Stenberger, M. 1971 (1964). Det forntida Sverige. Stockholm. Stenvik, L. F. 1996. Gravminner og maktsentra. In Før og etter Stiklestad 1030: Religionsskifte—kulturforhold—politisk makt. Seminar på Stiklestad, 1994. (Ed.) Ø. Walberg. Verdal. Stjerna, K. 1898. Erik den Helige: En sagohistorisk studie. Lund. ———. 1905. Vendel och Vendelkråka. Arkiv för Nordisk Filologi (21), 71–80. Stjernquist, B. 1970. Germanische Quellenopfer. In Vorgeschichtliche Heiligtümer und Opferplätze in Mittel- und Nordeuropa: Bericht über ein Symposium in Reinhausen bei Göttingen in der Zeit vom 14. bis 16. Oktober 1968. Abhandlungen der Akademie der Wissenschaften in Göttingen. Phil.-hist. Kl. 3. Dritte Folge, 74. (Ed.) H. Jankuhn. Göttingen, 78–99. ———. 1989. Arkeologiskt material som belägg för religion: Tolkningen som problem. In Arkeologi och religion: Rapport från arkeologidagarna 16–18 januari 1989. University of Lund. Institute of Archaeology. Report Series No. 34. (Eds.) L. Larsson & B. Wyszomirska. Lund, 57–66. Bibliography 607

———. 1997. The Röekillorna Spring: Spring-Cults in Scandinavia Prehistory. Contri butions by Tage Nilsson and Ulrik Møhl. ACTA. Regiae Societatis Humaniorum Litterarum Lundensis. Skrifter utgivna av Kungl. Humanistiska Vetenskapssamfundet i Lund. LXXXII. Stockholm. Stoklund, B. 1960. Gravøl (Danmark). In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 5. Malmö, cols. 449–450. Stoklund, M. 1994. Myter, runer og tolkning. Myte og ritual i det førkristne Norden: Et Symposium. (Eds.) J. P. Schjødt et al. Odense, 159–174. Stolpe, H. & Arne. T. J. 1912. Graffältet vid Vendel. KVHAA. Stockholm. Storm, G. 1885. Om Thorgerd Hölgebrud. Arkiv för nordisk filologi (2), 124–135. Strid, J. P. 1999 (1993). Kulturlandskapets språkliga dimension: Ortnamnen. Stockholm. Ström, Å. V. 1959. The King God and his Connection with Sacrifice in Old Norse Religion. In La Regalità Sacra. Leiden, 702–715. ———. 1975. Germanische Religion. In Germanische und Baltische Religion. Die Religionen der Menschheit. Band 19, 1. Stuttgart – Berlin – Köln – Mainz, 15–306. ———. 1990. Personal Piety in Nordic Heathenism. In Old Norse and Finnish Religions and Cultic Place-Names, Scripta instituti Donneriani Aboensis, XIII. (Ed.) Tore Ahlbäck. Stockholm, 374–380. Ström, F. 1942. On the Sacral Origin of the Germanic Death Penalties. Lund. ———. 1954. Diser, Nornor, Valkyrjor: Fruktbarhetskult och sakralt kungadöme i Norden. KVHAAH. Filologisk-Filosofiska Serien. 1. Stockholm. ———. 1961. Heliga träd. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 6. Malmö, cols. 373–374. ———. 1968. Kung Domalde i Svitjod och ‘kungalyckan’. Saga och Sed: Kungl. Gustav Adolfs Akademiens Årsbok. Uppsala, 52–66. ———. 1981. Poetry as an Instrument of Propaganda: Jarl Hákon and his Poets. In Specvlvm norroenvm: Norse Studies in Memory of Gabriel Turville-Petre. (Eds.) U. Dronke et al. Odense, 440–458. ———. 1983. Hieros gamos-motivet i Hallfreðr Óttarssons Hákonardrápa och den nordnorska jarlavärdigheten. Arkiv för Nordisk Filologi (98), 67–79. ———. 1985 (1961). Nordisk hedendom: Tro och sed i förkristen tid. Stockholm. Ström, K. 1970. Om fynden av torshammarringar. Lic. avhandling framlagd för semina- riet i nordisk fornkunskap, Stockholms universitet. Ström, P. 1997. Paradisi recuperatio: Den romersk-germanska kyrkoinvigningens form och innebörd. Uppsala. Strömbäck, D. 1928a. Lytir—en fornsvensk gud? In Festskrift til Finnur Jónsson 29. maj 1928. København, 283–293. ———. 1928b. Att helga land: Studier i Landnáma och det äldsta rituella besittningsta- gandet. In Festskrift tillägnad Axel Hägerström den 6 september 1928 av filosofiska och juridiska föreningarna i Uppsala. Uppsala – Stockholm, 198–220. 608 Bibliography

———. 1975. The Conversion of Iceland: A Survey. Viking Society for Northern Research. Text Series. Vol. VI. Trans. P. Foote. London. Stutz, U. & H. E. Feine 1989. Forschungen zu Recht und Geschichte der Eigenkirche: Gesammelte Abhandlungen. Aalen. Sundkvist, A. 2001. Hästarnas land: Aristokratisk hästhållning och ridkonst i Svealands yngrejärnålder. OPIA 28. Uppsala: Institutionen för arkeologi och antik historia. Sundqvist, O. 1995. Recension—Claus Krag, Ynglingatal og Ynglingesaga: En studie i historiske kilder, Oslo: Universitetsforlaget, 1991. Svensk Religionshistorisk Årsskrift (5). (Ed.) D. Westerlund. Stockholm, 158–162. ———. 1997. Runology and History of Religions: Some Critical Implications of the Debate on the Stentoften Inscription. In Blandade runstudier II. Runrön. 11. Runologiska bidrag utgivna av Institutionen för nordiska språk vid Uppsala universitet. (Eds.) L. Elmevik & L. Peterson. Uppsala, 135–174. ———. 1998. Kultledare och kultfunktionärer i det forntida Skandinavien. Svensk religionshistorisk årsskrift (7). (Ed.) D. Westerlund. Stockholm, 76–104. ———. 2002. Freyr’s Offspring: Rulers and Religion in Ancient Svea Society. Acta Universitatis Upsaliensis. Historia Religionum. 21. Uppsala. ———. 2003a. The Problem of Religious Specialists and Cult Performers in Early Scandinavia. Zeitschrift für Religionswissenschaft (1/2), 107–131. ———. 2003b. Priester und Priesterinnen. In Reallexikon der Germanischen Altertumskunde. Band 23. (Eds.) H. Beck et al. Berlin – New York, 424–435. ———. 2004a. Uppsala och Asgård: Makt, offer och kosmos i forntida Skandinavien. In Ordning mot kaos—studier av nordisk förkristen kosmologi. Vägar till Midgård 4. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 145–179. ———. 2004b. Sakralkönigtum (Skandinavische Quellen). In Reallexikon der Germa nischen Altertumskunde. Band 26. (Eds.) H. Beck et al. Berlin – New York, 279–293. ———. 2005a. Kvinnliga kultledares religiösa och sociala position i forntida Skandinavien. Chaos. Dansk-Norsk tidsskrift for religionshistoriske studier (43), 9–29. ———. 2005b. Aspects of Rulership Ideology in Early Scandinavia—with Particular References to the Skaldic poem Ynglingatal. In Das frühmittelalterlische Königtum. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 49. (Ed.) F. R. Erkens. Berlin – New York, 87–124. ———. 2005c. Tempel. In Reallexikon der Germanischen Altertumskunde. Band 30. (Eds.) H. Beck et al. Berlin – New York, 327–338. ———. 2005d. Tabu. In Reallexikon der Germanischen Altertumskunde. Band 30. (Eds.) H. Beck et al. Berlin – New York, 249–260. ———. 2005e. Sagas, Religion and Rulership: The Credibility of the Description of Ritual in Hákonar saga goða. In Viking and Medieval Scandinavia 1. (Eds.) J. Quinn et al. Turnhout, 225–250. ———. 2006. Från vikingatida aristokratiska hallar till medeltida stormannakyrkor. Bebyggelsehistorisk tidskrift (52), 20–32. Bibliography 609

———. 2007. Kultledare i fornskandinavisk religion. OPIA 41. Institutionen för arkeologi och antik historia. Uppsala universitet. Uppsala. ———. 2009a. The question of ancient Scandinavian cultic buildings: with particular reference to Old Norse hof. Temenos (45/1), 275–294. ———. 2009b. The Hanging, the Nine Nights and the ‘Precious Knowledge’ in Hávamál 138–145: The Cultic Context In Analecta Septentrionalia. RGA-E-Band 65. Berlin & New York, 649–668. ———. 2010a. Om hängningen, de nio nätterna och den dyrköpta kunskapen i Hávamál 138–145: Den kultiska kontexten. Scripta Islandica—Isländska Sällskapets Årsbok (61), 68–97. ———. 2010b. Vi-platsernas väktare: Till frågan om kultledare i det förkristna Mälarområdet. In Makt, kult och plats. Rapport från arkeologiska seminarier vid Stockholms läns museum. (Eds.) P. Bratt & R. Grönwall. Stockholm, 90–95. ———. 2011a. An Arena for Higher Powers: Cultic Buildings and Rulers in the Late Iron Age and the Early Medieval Period in the Mälar Region. In Ideology and Power in the Viking and Middle Ages: Scandinavia, Iceland, Ireland, Orkney and the Faeroes. (Eds.) G. Steinsland et al. Leiden, 163–211. ———. 2011b. Gudme on Funen: A central sanctuary with cosmic symbolism? In The Gudme/Gudhem Phenomenon. (Eds.) O. Grimm & A. Pesch. Neumünster, 63–76. ———. 2012a. Var sejdhjällen (fvn seiðhjallr, hjallr) en permanent konstruktion vid kultplatser och i kultbyggnader? Fornvännen (107), 280–285. ———. 2012b. Religious ruler ideology in pre-Christian Scandinavia. A contextual perspective. In More than Mythology. Narratives, ritual practices and regional distribution in pre-Christian candinavian religions. Eds. C. Raudvere & J. P. Schjødt, J. P. Lund, 225–262. ———.2013. Gamla Uppsala som förkristen kultplats: en översikt och en hypotes. In Gamla Uppsala i ny belysning. (Eds.) O. Sundqvist & P. Vikstrand with J. Ljungkvist. Gävle, 69–112. ———. 2013b. Snorri Sturluson as a historian of religion—The credibility of the descriptions of pre-Christian cultic leadership and rituals in Hákonar saga góða. In Snorri Sturluson—Historiker, Dichter, Politiker. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 85. Red. H. Beck et al. Berlin/New York: de Gruyter, 71–92. ———. 2014. Religious and Ideological Aspects of Hall Interiors in Late Iron Age”. In Runsa Borg. Representative Life on a Migration Period Hilltop Site—a Scandiavian Perspective. Papers from the project Runsa borg, Uppland. No 2. (Ed.) M. Olausson. Östersund, 109–144. ———. 2015. Gårdarna Hauge och Tu i Klepp. En kultplats med kvinnliga kultledare? In Et Akropolis på Jæren? Tinghaugplatået gjennom jernalderen. AmS-Varia 55. Eds Ellen Siv Kristoffersen, Marianne Nitter og Einar Solheim Pedersen. Stavanger, 89–105. 610 Bibliography

———. 2015. M.S. Vapen, våld och vi-platser: Skändande av helgedomar som makt- strategi i det vikingatida Skandinavien. In Krig och fred i vendel- och vikingatida tra- ditioner. (Eds.) S. Olsson & H. Rydving. Bergen. Sundqvist, O. & A. Hultgård. 2004. The Lycophoric Names of the 6th to 7th Century Blekinge Rune Stones and the Problem of their Ideological Background. In Namenwelten. Orts- und Personennamen in historischer Sicht. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 44. (Eds.) A. van Nahl et al. Berlin – New York, 583–602. Sundqvist, O. & P. Vikstrand. 2014. Disevid och Distingen—spår av östnordisk diskult? In Den heliga platsen. Handlingar från symposiet Den heliga platsen Härnösand 15–18 september 2011. (Eds.) E. Nyman et al. Sundsvall, 153–178. Svanhildur Óskarsdóttir. 1994. Daið þér Ynglinga? In Sagnaþing helgað Jónasi Kristjánssyni sjötugum 10. April 1994. 2 vols. Ed Gísli Sigurðsson et al., 761–768. Svavar Sigmundsson. 1992. Áttrúnaður og örnefni. In Snorrastefna 25.–27. Júlí 1990. Ritstjóri: Úlfar Bragason. Reykjavík. Sveaas Andersen, P. 1977. Samlingen av Norge og kristningen av landet. 800–1130. Handbok i Norges historie. Bind 2. Oslo. Sveinbjörn Rafnsson. 1974. Studier i Landnámabók: Kritiska bidrag till den isländska fristatstidens historia. Lund. ———. 1983. Frásögur um Fornaldarleifar 1817–1823. Reykjavík. Svensson, K. 2010. Götavi—en vikingatida kultplats i Närke. In Makt, kult och plats. Rapport från arkeologiska seminarier vid Stockholms läns museum. (Eds.) P. Bratt & R. Grönwall. Stockholm, 58–68. Tambiah, S. J. 1979. A performative approach to ritual. Proceedings of the British Academy 65, 113–169. Taranger, A. 1890. Den angelsaksiske kirkes indflydelse paa den norske. Kristiania. ———. 1913. The Meanings of the Words Óðal and Skeyting in the Old Laws of Norway. In Essays in Legal History Read before the International Congress of Historical Studies Held in London in 1913. (Ed.) P. Vinogradoff. London, 159–73. ———. 1934–1936. Om Kongevalg i Norge i Sagatiden. Norsk Historisk Tidsskrift (30), 110–166 / 273–311. Tesch, S. 1990a. Bakgrund och metod. In Makt och människor i kungens Sigtuna: Sigtunautgrävningen 1988–90. 28 artiklar om de preliminära resultaten. (Ed.) S. Tesch. Sigtuna, 7–19. ———. 1990b. Stad och stadsgård. In Makt och människor i kungens Sigtuna: Sigtunautgrävningen 1988–90. 28 artiklar om de preliminära resultaten. (Ed.) S. Tesch. Sigtuna, 23–50. ———. 2001. Olof Palme, S:ta Gertrud och Sigtunas medeltida kyrkotopografi. In Biskopen i museets trädgård. En arkeologisk gåta. Sigtuna Museers skriftserie 9. (Eds.) S. Tesch & R. Edber. Sigtuna, 9–44. Thompson, C. W. 1975. Studies in Upplandic Runography. Austin, Texas – London. Bibliography 611

Thomsen, C. J. [& N. M. Petersen]. 1886. Ledetraad til nordisk Oldkyndlighet. København. Thordeman, B. 1920. Alsnö hus: Ett svenskt medeltidspalats i sitt konsthistoriska sammanhang. Stockholm. Thorgunn Snædal. 2002. Medan världen vakar: Studier i de gotländska runinskrifternas språk och kronologi. Runrön. 16. Runologiska bidrag utgivna av Institutionen för nordiska språk vid Uppsala universitet. Uppsala. Thrane, H. 1998. Overvejelse af kultinholdet i Gudmes bebyggelse. In Centrala platser—Centrala frågor: Samhällsstrukturen under Järnåldern. En Vanbok till Berta Stjernquist. Uppakrastudier 1. (Eds.) L. Larsson/B. Hardh. Stockholm, 249–261. Thümmel, A. 1909. Der germanische Tempel. Beiträge zur Geschichte der deutschen Sprache und Literatur (35), 1–123. Thunmark-Nylén, L. 1989. Samfund och tro på religionsskiftets Gotland. In Medeltidens födelse: Symposier på Krapperups Borg. Vol. 1. (Ed.) A. Andrén. Lund, 213–32. Tillhagen, C.-H. 1995. Skogarna och träden: Naturvård i gångna tider. Stockholm. Tolley, C. 2009. Shamanism in Norse Myth and Magic. FF Communications. Vol. CXLIV, No.296. Helsinki. Tomlinson, R. A. 1987. Temple: Ancient Near Eastern and Mediterranean Temples. In The Encyclopedia of Religion 14. (Ed.) M. Eliade. New York, 383–388. Turcan, R. 1993. Mithra et le Mithriacisme. Paris. Turner, V. 1967. The Forest of Symbols: Aspects of Ndembu Ritual. London. ———. 1969. The Ritual Process. Structure and Anti-Structure. London. ———. 1979. Betwixt and Between: The Liminal Period in Rites de Passage. In Reader in Comparative Religion: An Anthropological Approach. (Eds.) W. A. Lessa & E. Z. Vogt. New York. (Reprinted from The Proceedings of the American Ethnological Society, 1964), 234–243. Turville-Petre, E. O. G. 1964. Myth and Religion of the North: The Religion of Ancient Scandinavia. New York – London. ———. 1976. Skaldic Poetry. Oxford. Turville-Petre, J. 1978–79. On Ynglingatal. Mediaeval Scandinavia (11). Odense, 48–67. Utne Nilsen, M. 2002. Gull—edle skatter fra forntidens Midt-Norge. SPOR (2002/1, 17/33), 10–14. de Vaux, R. 1980 (1961). Ancient Israel: Its life and institutions. 1980. Vermaseren, M. J. 1956–60. See CIMRM above. Vésteinn Ólason. 1998. Dialogues with the Viking Age: Narration and Representation in the Sagas of the Icelanders Reykjavík. ———. 2008. Snorri Sturluson—tiden, mannen og verket. In Høvdingen: Om Snorre Sturlasons liv og virke. (Eds.) J. O. Askedal & K. J. Myrvoll. Oslo, 21–40. Vestergaard, E. 1990. A Note on Viking Age Inaugurations. In Coronations: Medieval and EarlyModern Monarchic Ritual. (Ed.) J. M. Bak. Berkeley – Los Angeles – Oxford, 119–124. 612 Bibliography

Vikstrand, P. 1993. Förkristna sakrala ortnamn i Jämtland. Namn och bygd (81), 49–84. ———. 1996. Kan ortnamnselementet skälv, fsv. skialf, skiælf, ha haft en sakral innebörd? In Från götarna till Noreens kor: Hyllningsskrift till Lennart Elmevik på 60-årsdagen 2 februari 1996. Skrifter utgivna genom Ortnamnsarkivet i Uppsala. Serie B. Meddelande. 11. (Eds.) E. Brylla et al. Uppsala, 201–208. ———. 2000. Konungen och helgedomen. In Oluf Rygh: Rapport fra symposium på Stiklestad 13.–15. Mai 1999. NORNA-rapporter. 70B. (Eds.) B. Sandnes et al. Uppsala, 43–54. ———. 2001. Gudarnas platser: Förkristna sakrala ortnamn i Mälarlandskapen. Acta Academiae regiae Gustavi Adolphi. LXXVII. Studier till en svensk ortnamnsatlas utgivna av Thorsten Andersson. 17. Uppsala. ———. 2002. Sacral place-names in Scandinavia. ONOMA (2002), 121–143. ———. 2004a. Sakrale Namen. In Reallexikon der Germanischen Altertumskunde. Band 26. (Eds.) H. Beck et al. Berlin – New York, 167–179. ———. 2004b. Skutá and Vendil. Two Place Names in Ynglingatal. In Namenwelt. Orts- und Personennamen in historischer Sicht. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde. Band 44. (Eds.) A. van Nahl, L. Elmevik & S. Brink. Berlin – New York, 372–387. ———. 2006. Ásgarðr, Miðgarðr, and Útgarðr: A Linguistic Approach to a Classical Problem. In Old- Norse Religion in Long-Term Perspective: Origins, Changes, and Interactions. Vägar till Midgård 8. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 354–357. ———. 2007. Bebyggelsenamnen i Mörbylånga kommun. Sveriges ortnamn. Ortnamnen i Kalmar län 7. Uppsala. ———. 2009. Förkristna sakrala personnamn. Studia anthroponymica Scandinavica 27 (2009), 5–31. ———. 2010. Ortnamn och den äldre järnålderns högstatusmiljöer. In Makt, kult och plats. Rapport från arkeologiska seminarier vid Stockholms läns museum. (Eds.) P. Bratt & R. Grönwall. Stockholm, 23–30. ———. 2011. Sakrala ortnamn—och centrala: Om den religiösa sfärens ställning i dimensionen centrum- pereferi under förkristen tid. In Navnemiljøer og samfund i jernalder og vikingetid: Rapport fra NORNAs 38. Symposium i Ryslinge 12.–15. Maj 2009. (Eds.) L. Eilersgaard Christensen & B. Jørgensen. Uppsala, 315–346. ———. 2013. Namnet Uppsala. In Gamla Uppsala i ny belysning. (Eds.) O. Sundqvist & P. Vikstrand with J. Ljungkvist. Religionsvetenskapliga studier från Gävle 9. Gävle, 135–160. Vries, J. de. 1932. Über Sigvats Álfablót-Strophen. Acta Philologica Scandinavica (7), 169–180. ———. 1956–57. Altgermanische Religionsgeschichte I–II. Grundriss der germanischen Philologie 12/1–2. Berlin. ———. 1977. Altnordisches etymologisches Wörterbuch. Leiden. Bibliography 613

Wahlberg, M. 1994. Uppsala gatunamn. Sveriges ortnamn. Ortnamnen i Uppsala Län. Uppsala kommun och Uppsala gatunamn. Del 5:1. Uppsala. ———. (ed.) 2003. Svenskt ortnamnslexikon. Uppsala. See also SOL above. Wagner, R. 1987. Taboo. In The Encyclopedia of Religion. Vol. 14. (Ed.) M. Eliade, New York – London, 233–236. Waida, M. 1987. Authority. In The Encyclopedia of Religion. Vol. 2. (Ed.) M. Eliade. New York – London, 1–7. Walter, E. 1966. Quellenkritisches und Wortgeschichtliches zum Opferfest von Hlaðir in Snorris Heimskringla (Hák. góð. c. 17). In Festschrift Walter Baetke: Dargebracht zu seinem 80. Geburtstag am 28. März 1964. Weimar, 359–367. Warmid, M. 2004. . . . á þeim meiði, er mangi veit, hvers hann af rótom renn. In Ordning mot kaos—studier av nordisk förkristen kosmologi. Vägar till Midgård 4. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 135–144. Watt, M. 1991. Guldgubberne fra Sorte Muld, Bornholm: Tanker omkring et muligt hedensk kultcentrum fra yngre jernalder. In Nordisk hedendom: Et symposium. (Eds.) G. Steinsland et al. Odense, 373– 386. ———. 1999. Gubber. In Reallexikon der Germanischen Altertumskunde. Band 13. (Eds.) H. Beck et al. Berlin – New York, 132–142. ———. 2004. The Gold-Figure Foils (Guldgubbar) from Uppåkra. In Continuity for centuries—A ceremonial building and its context at Uppåkra, southern Sweden. Acta Archaelogica Lundensia series in 8o, no. 48. (Ed.) L. Larsson. Lund, 167–222. ———. 2007. Kan man tyde guldgubbernes gestussprog? Kult guld och makt—ett tvärvetenskapligt symposium i Götene. (Ed.) I. Nordgren. Göteborg. Weber, G. W. 1972. Das Odinsbild des Altuna-Steins. Beiträge zur Geschichte der deutschen Sprache und Literatur (94), 323–34. ———. 1981. Irreligiosität und Heldenzeitalter. Zum Mythencharakter der altislän- dischen Literatur. In Specvlvm norroenvm: Norse Studies in Memory of Gabriel Turville-Petre. (Eds.) U. Dronke et al. Odense, 474–505. ———. 1987. Intellegere historiam. Typological Perspectives of Nordic Prehistory (in Snorri, Saxo, Widukind and others). In Tradition og historieskrivning: Kilderne til Nordens ældste historie. (Eds.) K. Hastrup & P. Meulengracht Sørensen. Århus, 95–141. ———. 1994. Euhemerismus. In Reallexikon der Germanischen Altertumskunde. Band 8. (Eds.) H. Beck et al. Berlin – New York, 1–16. Weber, M. 1971 (1922). The Sociology of Religion. Orig. title Religionssoziologie in Wirtschaft und Gesellschaft. Trans. E. Fischoff. London. Weibull, C. 1964 (1921). Om det svenska och det danska rikets uppkomst. In Källkritik och historia: Norden under äldre medeltiden. Stockholm, 7–52. Weibull, L. 1948. Nordisk historia: Forskningar och undersökningar I. Forntid och vikingatid. Stockholm. Wellendorf, J. 2003. Ideologi og trosforestillinger i Óláfs þáttr Geirstaðaálfs. Om jord- fundne genstande og rituelle højbrud. Nordica Bergensia (29), 147–169. 614 Bibliography

———. 2009. Om højsædestolper, klokker og døre. In Á austrvega: Saga and East Scandinavia. Preprint papers of the 14th International Saga Conference. Uppsala 9th–15th August 2009. Gävle, 1016–1023. ———. 2010. The Interplay of Pagan and Christian Traditions in Icelandic Settlement Myths. In Journal of English and Germanic Philology (109), 1–21. Wernersson, L. 1996. Helgö som religiöst centrum: En hypotes. Uppsats i fördjupnings- kurs i arkeologi. Stockholms universitet. Wesche, H. 1937. Beiträge zu einer Geschichte des deutschen Heidentums. Beiträge zur Geschichte der deutschen Sprache und Literatur. (Ed.) T. von Frings. Halle, 1–116. Wessén, E. 1923. Minnen av forntida gudsdyrkan i Mellan-Sveriges ortnamn. Studier i nordisk filologi. 14:1. Skriftern utgifna av Svenska litteratursällskapet i Finland. Helsingfors. ———. 1924. Studier till Sveriges hedna mytologi och fornhistoria. Uppsala Universitets Årsskrift. 1924. Filosofi, språkvetenskap och Historiska vetenskaper. 6. Uppsala. ———. 1940. See Holmbäck, Å & E. Wessén. 1979 (1933–46). ———. 1960. Gutasagan. In Kulturhistoriskt lexikon för nordisk medeltid. Vol. 5. Malmö, cols. 602–604. ———. 1964. See Snorri Sturluson. Wessén & Jansson 1940–58. See U above. Whaley, D. 1999. Heimskringla. In Reallexikon der Germanischen Altertumskunde. Band 14. (Eds.) H. Beck et al. Berlin – New York, 238–247. Whitley, J. 1995. Tomb cult and hero cult: the uses of the past in Archaic Greece. In Time, Tradition and Society in Greek Archaeology: Bridging the Great Divide. (Ed.) N. Spencer. London, 43–63. Wickström, J. 2004. Bröllopsmyten i Helgakviða Hjörvarzsonar: Ett exempel på fornskandinavisk härskarideologi i Codex Regius hjältediktning. Arkiv för nordisk filologi (119), 106–123. Widengren, G. 1965. Die Religionen Irans. Die Religionen der Menscheit, Bd. 14. Stuttgart. ———. 1969. Religionsphänomenologie. Berlin. Widmark, G. 1997. Tolkningen som social konstruktion: Rökstenens inskrift. In Runor och ABC: Elva föreläsningar från ett symposium i Stockholm våren 1995. Sällskapet Runica et mediævalia. Opuscula. 4. (Ed.) S. Nyström. Stockholm, 165–175. ———. 1999. Till tolkningen av Forsaringens inskrift. In Runor och namn: Hyllningsskrift till Lena Peterson den 27 januari 1999. Namn och samhälle 10. (Eds.) L. Elmevik & S. Strandberg. Uppsala, 21–28. Wiebe, D. 1996. Is the new comparativism really new? In Method & Theory in the Study of Religion (8–1), 21–29. Wikander, S. 1979. Araber—Vikingar—Väringar. Svenska Humanistiska Förbundet. 90. Lund. Bibliography 615

Williams, H. 1990. Åsrunan: Användning och ljudvärde i runsvenska steninskrifter. Runrön 3. Runologiska bidrag utgivna av Institutionen för nordiska språk vid Uppsala universitet. Uppsala. ———. 1993. Ó-namn: Nordiska personnamn med det privativa prefixet Ó-. In Personnamn i nordiska och andra germanska fornspråk. NORNA-rapporter. 51. (Ed.) L. Peterson. Uppsala, 95–107. ———. 1996a. Runstenstexternas teologi. In Kristnandet i Sverige: Gamla källor och nya perspektiv. Projektet Sveriges kristnande. Publikationer. 5. (Ed.) B. Nilsson. Uppsala, 291–312. ———. 1996b. Vad säger runstenarna om Sveriges kristnande?. In Kristnandet i Sverige: Gamla källor och nya perspektiv. Projektet Sveriges Kristnande. Publikationer. 5. (Ed.) B. Nilsson. Uppsala, 45–83. ———. 2001. Lister. §2. Runological. In Reallexikon der Germanischen Altertumskunde. Band 18. (Eds.) H. Beck et al. Berlin – New York, 509–512. ———. 2008. Runes. In The Viking World. (Ed.) S. Brink. London, 281–290. Wood, S. 2006. The Proprietary Church in the Medieval West. Oxford. Zachrisson, S. 1963. Från vendeltid till vasatid: Redogörelse för de byggnadsarkeolo- giska undersökningarna av Eskilstuna kloster och slott 1961−62. Eskilstuna stads museer. Årsbok 1963. Eskilstuna, 17−74. ———. 1998. Eskilstuna kloster, utgrävningarna 1961–62. In Johanniterna—vilka var de? Symposiet tisdagen den 3 november 1998. Projektet Eskilstuna historia. Stadsarkivet. Eskilstuna kommun. (Ed.) B.-E. Ohlsson. Eskilstuna, 11–23. Zachrisson, T. 1994. The Odal and its Manifestation in the Landscape. Current Swedish Archaeology (2), 219–239. ———. 1998. Gård, gräns, gravfält: Sammanhang kring ädelmetalldepåer och runstenar från vikingatid och tidig medeltid i Uppland och Gästrikland. Stockholm Studies in Archaeology. 15. Stockholm. ———. 2002. Ärinvards minne—om runstenen i Norra Sandsjö. In Om runstenar i Jönköpings län. (Eds.) J. Agertz & L. Varenius. Värnamo, 35–54. ———. 2003. Den hängde med halsringen. Om en figurin från järnåldern funnen vid Kymbo Storegården i Västergötland. Fornvännen (98), 89–97. ———. 2004a. Det heliga på Helgö och dess kosmiska referenser. In Ordning mot kaos—studier av nordisk förkristen kosmologi. Vägar till Midgård 4. (Eds.) A. Andrén, K. Jennbert & C. Raudvere. Lund, 343–388. ———. 2004b. The Holiness of Helgö. In Excavations at Helgö XVI. Exotic and Sacral Finds from Helgö. Kungl. Vitterhets Historie och Antikvitets Akademien. Stockholm, 143–175. ———. 2010. Helgö—mer än ett vi. In Makt, kult och plats. Rapport från arkeologiska seminarier vid Stockholms läns museum. (Eds.) P. Bratt & R. Grönwall. Stockholm, 79–88. 616 Bibliography

———. 2011a. Arkeologin bakom Rimbert: Om Hergeiirs och Gautberts kyrkor och borgen i Birka. Fornvännen (106), 100–112. ———. 2011b. Exotiska föremål och nya seder under mellersta järnåldern. In Förmodern globalitet: Essäer om rörelse, möten och fjärran ting under 10 000 år. (Ed.) A. Andrén. Lund, 109–130. ———. 2013. Gamla Uppsala—på nytt. In Gamla Uppsala i ny belysning. (Eds.) O. Sundqvist & P. Vikstrand with J. Ljungkvist. Religionsvetenskapliga studier från Gävle 9. Gävle, 161–204. Zachrisson, T. & M. Kempff Östlind. 2006. Ceremoniella rum. Bebyggelsehistorisk tidskrift (52), 5–6. Zernack, J. 1998. Fulltrúi. In Reallexikon der Germanischen Altertumskunde, Band 10. (Eds.) H. Beck et al. Berlin – New York, 243–245. Zinser, Ha. 1993. Kosmologie. In Handbuch religionswissenschaftlicher Grundbegriffe. Band III. (Eds.) H. Cancik, B. Gladigow & K.-H. Kohl. Stuttgart – Berlin – Köln, 446–448. Žižek, S. 2012 (1994). Introduction: The Spectre Ideology. In Mapping Ideology. London – New York, 1–33. ——— (ed.) 2012 (1994). Mapping Ideology. London – New York. Index

aðalból 452–453 Æsir 216, 459, 518 n60 Aðalból (Iceland) 284 Af Upplendinga konungum 461 Aðalsteinn (king) 361, 426 Agni (king) 361, 425 Adalvard the Younger (bishop) 45–46, 121, Ágrip af Nóregskonunga sǫgum 58, 139, 173, 309–310 322, 325, 344, 348, 352, 370, 463, 495, 512 Adam of Bremen Ahn, Gregor 14–15 n38 informants of 114, 121 Äldre Västgötalagen 400–401, 495, 500 use of term, templum 125 Alexander III, of Scotland 493 views of álfar/álfablót 303, 323, 370, 459–460 on Adalvard the Younger 309–310 Alfífa (paramour to King Knútr) 358 on animal sacrifices 340 allsherjargoði 507 on first Christian kings of Alstadhaug/Ǫlvishaugr (Skogn) Svetjud 44–45 burial mounds at 435 on Niðaróss 53 fylkiskirkja at 210, 440 on paying tribute 192–193 name of 441, 446 on Pope and Rome 115, 117 altars. see podiums/altars (stallar/stalli/stallr) on Uppsala sanctuary Altuna stone (U 1161) 122, 123, 384 Augustinian perspectives on 114 Alþingi 158, 390, 507, 544 burials mounds at 436 Alvir/Ǫlvir (ruler) 440, 446 cultic feasts at 113, 191, 192–193, ambulatory rituals 246 322, 326–327, 514 Ammianus Marcellinus 11 description of 110–113 amulet rings drinking rituals at 354–355 from Borg (Östergötland) 392 exterior of 208, 218 fire-steel-shaped 392 high-seats 227 from Helgö 393 idols of gods in 111, 266–267, from Lilla Ullevi 394 269–270 from Old Uppsala 395 influences on 205–206 from Sanda 395 as inter-regional 40 Andersson, Gunnar 237–238, 271 n31, 392 reliability of 263 Andersson, Thorsten 39, 369 sacred spring/wells at 258 Andrén, Anders 129, 255 sacred trees in 250–251 Anglo-Saxon Chronicle 390, 399 sources supporting 120–127 Anglo-Saxons 309–310 n77 on Þórr 274, 287 animal sacrifices Adelsö (Uppland) 206, 395, 445, 528 in Blekinge 329 Aðils (king) 124, 425 and decapitations 339–340 Ægidius vita 234 at Hofstaðir (Mývatnssveit) 153, 324, 339, Den ældre Borgarthings-Christenret 281, 342–343 345 horses 59, 133, 135, 139, 172, 194, 229, 254, Den ældre Eidsivathings-Christenret 279, 317, 327, 329, 333, 344–346, 348 280, 281 on Iceland 339 Den ældre Frostathings-Lov 280, 281, 299, at Lade 133, 193–194 303–304, 345, 450, 490 in Mälaren area 340, 348 Den ældre Gulathings-Lov 104–105, 261, 281, at Old Uppsala 348–349 344, 357–358, 443–444, 449–450, 482, outdoor performance of 190–191 490 as scapegoating 343 618 Index

Annales Fuldenses 399 Bell, Catherine 341–342 Ansgar 43 Beowulf 37, 307, 361, 364, 530 Anundr of Tåsta 378–379, 382–383, 398 Bergagård (Slöinge) 417 Anunder (king) 326–327, 515 Berg-Ǫnundr 304 Anundr (ruler in Västmanland) 446 Bertnem 435, 437–438 Anundr Jacob (r. 1022–1050) 45, 121 Berudrápa 280 Anundshögen 431, 446–447 Birgisson, Bergsveinn 65 Åqvist, Cecilia 183 Birka ár ok friðr expression 356–359 chamber graves at 374 archbishoprics 43, 53, 60 divination rituals at 337 archdiocese 61, 110, 114, 115 drinking rituals at 355 Ari inn fróði Þorgilsson 54, 58, 59, 70, 72, elevation of 206 196–197, 346–347 hall building at 523–534 arm rings 428–429 high-seat at 224, 227, 524, 529 Árni Helgason (bishop) 215 importance of hearths 243 Arnkell goði 230 warrior lords at 529 Árnórr Þorðarson jarlaskáld 221 weapons found at 524 Arrhenius, Birgit 128, 237, 272, 336, 445 Birkeli, Emil 222–223, 233 n152, 460, 490 Arwidsson, Greta 118–119 bishoprics Ásgarðr. see also Valhǫll myth; also Vingólf in Skálholt 61, 214–215 cultic site as reduplication of 262–263 in Skara 45 dualism between Útgarðr and 203–204 Björketorp stone (DR 360) 328 Gudme as replica of 201–202 Bjǫrn Finnvidsson 450 Ásgeirr á Eyri 177 blood, sacrificial Ásgerðr Asksdóttir 373 and decapitations 339–341 Åshusby (Norrsunda) 432 pouring over ritual objects of 332–334 Ásólfr alskik 233 n153 purposes of 341–343 Aspa löt stone (Sö Fv1948;289) 479 smearing of 331–332, 333 Ásta Guðbrandsdóttir 373 blood sacrifices asylum sites 299 and cultic objects 332–334 Atlaqviða in grœnlenzka 390 and decapitations 339–341 Austrfararvisur (Sigvatr Þórðarson) 303, and podiums/altars 331–334 369–370, 459, 524 blót house/blóthús 95, 104 n27, 156, 157, 216 authority/authorization, concept of 16 blót images 281 Axboe, Morten 418 blót place-names 260–261 axis mundi symbols 18, 102, 106, 200, 227, Blótkelda 260 236, 257, 263 blótskap 282–283 Blótsveinn (king) 119, 327–328, 348 Badelunda (Västmanland) 37, 375, 446 blótveizlur (sacrificial feasts) 138, 191, Baetke, Walter 8, 24, 70, 77 n85, 470 193–195, 316–317, 344 Bali 289 boat graves 118, 189, 373–374, 375, 394, 438, Baltic Sea 41 442, 453 banqueting hall. see feasting/banqueting bona regalia (crown lands) 40 halls bondedemokrati (farmer democracy) 48, 52 Bárðr from Atley 353 Boor, Hulmut de 526–527 Beck, Heinrich 24 n68, 352, 500 n115 Borg (Lofoten/ Vestervågøy) 411–412 Bede 247 ceremonial buildings at beer ceremonies 482–485 in general 145–146 Index 619

hearths in 244 mounds high-seat posts in 236 and ceremonial buildings 436–438, high-seats in 229 439–442, 543 rebuilding of 314 hero cults at size of 212, 213 of Eiríkr (king) 453–475 strategic positions of 210–211 of Hálfdan svarti 462–465 cultic feasts at, meals at 346 of Hǫlgi 469–476 gold foils from in Iceland 465–466 related to hearth 244 in Norway 456–465 related to post-holes 142, 145–146, of Óláfr Guðrøðarson 456–462 229, 236, 295–296, 417 overview 453–476 seen as temple coins 194 in Svetjud 454–456, 466–467 Borg (Norrköping/ Östergötland) 314 in Trøndelag area 469–476 amulet rings from 392 of Þorgerðr Hǫlgabrúðr 469–476 cultic house at 99, 99, 102, 336, 340 on Iceland 433 fire-steel-shaped pendants from in Mälaren area 384–385, 385 in general 431–433 Borgarnes (Iceland) 433 at Högom 437 Borgarthings-Lov 345 at Hovgården 436–437 Bǫrkr Þorsteinsson 359 at Husby (Vendel) 432, 438, Botair (in Guta saga) 297 447–448 bracteates 103, 265, 418–419 at Old Uppsala 434, 436 bragafull (libations) 3, 477, 482, 489, 498, as markers of óðal 448–452 500 names of 446–449 Bratt, Peter 433 in Norway 432 Braune, Wilhelm 202–203 n14 Østlandet 440 Bredgränden stone (U 942) 318–319 as part of sanctuaries 433–436 Breiðablik 218 rituals associated with 443–444 Brink, Stefan 262, 381–382, 387, 396, 486, in a row 436–438 504–505 and royal inaugurations 493–497 Brísingamen 428 symbols of genealogy 444–446 Broberg, Anders 118 as symbols of power 430–431 Brodd-Helgi 180 in Trøndelag area Bruun, Daniel 151 at Alstadhaug 435 Bugge, Sophus 23–24, 378 at Bertnem 435, 437–438 Bureus, Johannes 116–117, 260 overview 439–442 burial customs overlying graves at 445 on Iceland 432 for women 373–375, 438 in Old Uppsala 118–119 Burkert, Walter 462 in Ynglingatal 68–69 Buttle Änge 225, 226 burial sites Byock, Jesse 452 boat graves 118, 189, 373–374, 375, 394, 438, 442, 453 Cahen, Maurice 481 chamber graves 118, 374, 375, 438 Carlie, Anne 314 monumental graves casting of lots 337 in Mälaren area cemeteries, consecration of 118 overview 438–439 ceremonial buildings at Valsgärde 118, 241, 374–375, and burial mounds 433–434, 436–438 438 cultic houses. see cultic houses 620 Index ceremonial buildings (cont.) Christendom destruction of 307–311 cosmic symbolism in 200 for drinking ceremonies 485–488 and cross-signs 116, 118, 120, 134–135, 282, examples of 95–96, 97, 98, 100–102 318–319 existence of 95–102, 108 holy water rituals in 318 exteriors of 541 introduction of cult of sacred spring feasting/banqueting halls. see feasting/ by 261 banqueting halls and prohibition of pagan rituals 268, hall buildings. see hall buildings 279, 281, 344, 347, 355, 443–444, 473 high-seats in. see high-seats Christianization hof buildings/sanctuaries. see hof and destruction of cultic sites 308–310 buildings/sanctuaries of Götaland 44 in Iceland. see under Iceland of high kings of the Götar 44 interiors of 219–220, 541 of high kings of the Svear 43–45 in Mälaren area. see under Mälaren area of Iceland 21, 59–62, 158 political functions of 202, 218–219 of Norway 21, 52 regional differences in 31, 102 of Trøndelag area 52, 172, 282 and religious ruler strategies 290 of Svetjud 43–44 renovation of 311–314 of Mälaren area 43–44 ritual ending of 314–315 of Uppland 115, 118–120 as symbols of power 402–403, 540 Christians, continuing pre-Christian terminology of 104–106 customs 120 in Trøndelag area. see under Trøndelag Chronicon (Thietmar) 454 area chronological system 4 n3 use of term 109 church-door rings 380, 381, 422 warrior halls. see warrior halls churches. see also shire churches (fylkiskirkja) see also under names of specific sites as asylum sites 299 ceremonial meals door rings 380, 381 on Iceland 346–347 at Frösön 324 in Mälaren area 347–348 at Hofstaðir 215 on Helgö 349–350 at Hustad 282 in Lunda 350–351 private 60 in Trøndelag area 344–346 in Sigtuna 44 ceremonial sanctuary civitas dei 114 analytic concept of 5 civitas terrena 114 cosmic symbolism in 200–202, 263, Clunies Ross, Margaret 24, 233 n 152, 293 288–289, 541 Codex Frisianus 133 economic aspects of 31–32, 191–198, 402, Codex Regius (Snorra Edda) 104, 472 540 Codex Upsaliensis (Snorra Edda) 104 of inter-regional significance 508–509 Coifi 309–310 n77 as reduplication of Ásgarðr 262–263 Colosseum (Rome) 523 see also hof buildings/sanctuaries; outdoor comitatus 30–31, 364, 521 sanctuaries consecrations. see sanctification ceremonies cosmic eschatology 526, 534, 544–545 use of term 109 n55 cosmic symbolism 200–202, 263, 288–289, see also cultic rituals 541 chamber graves 118, 374, 375, 438 cosmology, Old Norse 201–204 chieftains, on Iceland. see under Iceland Council of Tours 252 children, exposure of 346 court circles 299–300 Index 621 cremation burials 118–119 cultic houses cross-signs 116, 118, 120, 134–135, 282, 318–319 analytic concept of 5, 108 cult continuity 261–262 at Borg (Östergötland) 99, 102, 336, 340 cultic feasts besides feasting halls 99, 100–102, calendrical 125–126, 149, 156 on Iceland 323–324, 343 at Hólmur 157, 216 in Mälaren area 322–323, 324 on Iceland 155–157 in Trøndelag area 320–322 at Lunda 270 chief sanctuaries and 508–510 at Uppåkra 99–100, 100, 313 described use of term 108 by Adam of Bremen 113, 191, 192–193, cultic images 322, 326–327 in general 264–265, 542 in Hávamál 190–191 animated forms of 277 by Snorri Sturluson 193–194 destruction of 308–311 in general 316–317 of Freyja 284 authenticity/inauthenticity of Freyr 111, 266–267, 271, 284 in 317–319 of Frigg 284 Christian impact on 318–319 of gods 264, 277–279, 280 in Ynglingatal 190 on Iceland 267, 285–287 drinking rituals at at Lade 272–273 in general 350, 351–356, 359, 370 at Mære 273–274 and formulaic ritual toasts 356–359 in Mälaren area 268–270 funeral feasts 481–488 Norwegian Laws on 279–282 gift-giving during 359–363 of Óðinn 111, 266, 282 for inheritance 488 prohibition of 279 inheritance feasts 483 and religious ruler strategies 287–289 at Lade 316 role during ritual feasts 288 at Mære 316 staver and 268–269 paying of tribute during 133, 192–196 on tapestries 287 role of cultic images in 288 of Þorgerðr Hǫlgabrúðr 275, 284 at royal funerals 476–492 of Þórr 111, 231, 237, 266, 267–268, 271 in Iceland 484–485 n31, 274, 276–277, 279, 283, 284, libations at 481–482 285–286, 287 in Norway 482–484 in Uppsala Temple 266–268, 269–270 for Ǫnundr 3, 476–480 see also decorations rituals performed at 3–4 cultic leaders/leadership. see also cultic feasts sacrificers at 325–331 concept of 15 n40 in general 288 on Iceland. see under Iceland in Iceland 330–331 ideological motives of 197–199 in Mälaren area 326–330 in Mälaren area. see under Mälaren area in Trøndelag area 277, 316–317, other social duties of 184–186, 191 325–326 as protector of cultic sites 188, 290–291, in Trøndelag area 138, 193–195, 316–317 305, 311, 315 at Uppsala 113, 191, 192–193, 322, 326–327, and re-establishing cultic sites 311–313 514 religious strategies of 15–19 women’s role in religious-political tolerance of 120 in general 13 research on 164 húsfreyja 365–368 and ritual ending cultic sites 315 noble 363–365 as sacrificer 277, 288, 316–317, 325–331 622 Index cultic leaders/leadership (cont.) De Ordine Palatii 365 n211 social prestige of 191–192 de Vries, Jan 164, 341, 508 subordinated 189–192 death, and Eros 78 in Trøndelag area. see under Trøndelag decapitations, and bloody area sacrifices 339–341 use of terms 167 Declans vita 234–235 and violations against cultic decorations sites 307–311 on Altuna stone (U 1161) 122, 123, 384 cultic objects on bracteates 418–419 and blood sacrifices 332–334 of ceremonial buildings/sites 208 destruction of 310–311 on Norra Åby stone (Sö 86) 384 see also cultic images; under specific on Sparlösa stone (Vg 119) 422, 424 objects on Stenkvista stone (Sö 111) 384 cultic organization on stone from Tängelgårda 398, 399 in Gudme 519–520 on stone in vestry of Hustad church 282 on Iceland 505–508, 544 on Tjängvide stone (G 110) 78, 79 in Norway 503–504, 510–514 Decretum Gratiani 298 in Trøndelag area 187, 510–514, 544 Delsbo (Hälsingland) 380, 381, 422 in Svetjud 544 Dillmann, François-Xavier 137, 138, 464 in Mälaren area 504–505 dísablót 124, 324, 353, 354 cultic place continuity 96–97, 141, 143 dísarsalr 124 cultic rituals Diser, nornor, valkyrjor: fruktbarhetskult och animal sacrifices. see animal sacrifices sakralt kungadöme (F. Ström) 11 n27 associated with Óðinn 293 dísir cult 11 n27, 124–125 associated with Þórr 293 dismemberment 462–465 blood sacrifices. see blood, sacrificial Disthingen 125, 322–323 complexity of 190–191 Distingstungel 322 for divination 336–339 divination rituals 129 dramatic dimensions of 342 divinations 336–339 drinking rituals. see drinking rituals divine descent importance of 4–5 of Earls of Lade 80–83, 287, 469, 473, 539 for inaugurations 492 of royal families in gereral 92–93 Norwegian Laws on 281 of rulers of Þrœndir 30 in Old Testament 136 of Ynglingar 29–30, 70–74, 270, 468, 539 prohibition of, by Christian laws 268, Dómaldi (king) 76–77 279, 281, 344, 347, 355, 437, 443–444 door rings 421–422, 423, 424 and religious ruler strategies 542 Draupnir 428 restrictions/taboos on 294–295, 297 drinking rituals at royal funerals 476–492 in general 350, 351–356, 359, 370 at royal inaugurations 492–499 and formulaic ritual toasts 356–359 sanctifications. see sanctification at funeral feasts 481–488 use of term 109 n55 religious aspects of 354–355 see also cultic feasts role of noble women in 363–364 culture, in Scandinavia, symbolic encoding special halls for 485–488 of 28 substituting pagan for Christian 322 Drobin, Ulf 341 Dalum (in Trøndelag) 419 Dronke, Ursula 422 Damell, David 258 Droplaugarsona saga 389 de Boor, Hulmut 526–527 DuBois, Thomas A. 485 Index 623

Dumville, David 493 Eriksgata 409, 447 Düwel, Klaus 136–137, 138, 190, 317, 331–332, Erikskrönikan (Rimkrönikan) 496 337, 358, 481–482 Ermingerðr (queen) 364 Eros, and death 78 Earls of Lade (Hlaðajarlar) Eskilstuna 130, 194, 409 cultic centres of euhemerism 24 at Lade. see Lade Exodus 332 at Mære. see Mære Eymundr (also Emundr, king) 45 divine descent of 80–83, 287, 469, 473, Eyrarland, figurine from 231, 283, 285–286, 539 384 mythical marriages 83–85, 413, 414, 416 Eyrbyggja saga as regional rulers 48–49, 50–51 on calendrical feasts 323 religious ruler strategies of 29–31, 80–85, criticism against 149 287–289 on cultic images in Iceland 283 Earls of Orkney 92 descriptions of oath-rings in 387 economic aspects, of ceremonial editions of 86 sanctuaries 31–32, 191–198, 402, 540 on erfi ceremonies 487 Eddic poetry, as direct sources 23 on funeral beer 484–485 Edward, Paul 86 on goðar as sacrificers 330–331 Egill Skalla-Grímsson 88–89, 280, 304–305, on high-seat posts 233, 234, 236–237, 257 353–354, 361, 371, 426 on podiums/altars 336 Egils saga 55, 212, 294, 300, 304–305, 306, on sanctification of land/ 324, 353, 361, 370–371, 426–427 places 292–294 Egino of Scania (bishop) 121, 310 on sanctuaries 150–151 Eigenkirchenwesen 60 sources for 86, 149–150 Eigentempelwesen 60 on temple-soil 333 Einarr Helgason (skald). see Vellekla Þórólfr bægifótr in 230 Einarsson, Bjarni F. 31 Þórólfr Mostrarskegg Ǫrnólfsson in einherjar 217, 248, 455, 522–523, 533–534 57, 86–87, 147–150, 176–177, 195, 213 Eiríkr (legendary king at Uppsala) 125, 338 Þorsteinn Þorskabítr Þórólfsson Eiríkr (with temple at Birka) 454–455 in 178–179 Eiríkr Bloodaxe Haraldsson (r. 931–933) use of hlaut in 318 304–305, 306, 308, 311, 324, 338–339, Eysteinn from Oppland (king) 441–442, 493 353, 526 Eyvindr (Gunnhildr’s brother) 294 n14 Eiríkr Eymundarson (also Emundarson) 529 Eyvindr Finnsson (skald) 65, 82, 495 Eiríkr inn sigrsæli Bjarnarson (r. c. see also Háleygjatal 970–995) 44 Eiríksmál 526 Fabech, Charlotte 259 Eketorp hoard 227, 228 Færeyinga saga 275, 469 Ekholm, Gunnar 118–119 Fagrskinna 139–140, 172–173, 219–222, 242, Eldhrímnir 248 311, 312, 325, 344, 352, 358, 462–463, 488, elevation, of sanctuaries 146, 206–207, 211 490, 491, 492, 493, 500, 501, 512 Eliade, Mircea 200–201, 234 Falk, Hjalmar 222, 223 Empedocles 247 farm names 440–442 Eneby stone (Sö 145) 450 farmer democracy (bondedemokrati ) 48, 52 Epistula Jeremiae 454 farmers erfi/inheritance and inauguration and cultic leadership 197–199 ceremonies 3, 207, 372, 476–492, elite groups of 48, 52 499–500, 502, 532, 543–544 on Iceland 148, 180, 195–197 624 Index farmers (cont.) Fornsigtuna 206, 207, 524 paying of sancturary taxes by 60 n153, Forsa ring 378 148, 150, 180, 195–197 in general 377 paying of tribute to rulers 133, 193–196 dating of 377–379, 391 relations with rulers 57, 60, 197–198, 402 as door ring 424 of Trøndelag area 133–135, 172, 173, 187, medieval-clerical background of 378 193–194 role in rituals feasting rooms 101 in general 397–398, 401 feasting/banqueting halls cult of Þórr 386 at Adelsö 528 runic inscriptions on 378–383, 396 cultic houses besides 99, 100–102, shape and appearance of 381–384 125–126, 149, 156 Fóstbræðra saga 214, 215, 237, 267 destruction of 307–308 four elements 247 development of 101 Frank, Roberta 13–14 in Iceland 214 Franks 361 role of hearths in 242 Frazer, James G. 7–8, 77 at Södra Kungsgårdsplatån in Freyja Uppsala 125, 313 cultic images of 284 written sources on 104, 219–220, 221, 242 necklace of 428 see also hall buildings Freyr (Frö) feasts. see cultic feasts association with pigs 351 female leaders. see women emblems of 367 fen lands 258–260 on Hliðskjálf 238–239 Fiædhrundaland 39 images of 111, 266–267, 271, 284 figurines importance of 72 from Eyrarland 231, 283, 285–286, 384 marriage to Gerðr 412, 415–416 holding and/or blowing in their as mythical father of Earls of Lade 82 beards 285 as mythical father of Ynglingar 29, 70, 71, from Kymbo 427, 428 72, 74, 75 from Lund 267, 285 personal devotion to 87–88 from Lunda 270–271, 271, 416 in place-names 517 from Rude Eskildstrup 427 sacrifices to 331, 355 from Slipshavnskogen 427 symbols of 393 see also statues worship of 8, 121–122, 180–181, 271–272, Finnr Sveinsson 276 415–416, 466–469, 518 Finnsburg 307–308 Friðgerðr (gyðja) 181–182, 336 fire, mythical dimension of 246–248 friðr (peace) 299, 303, 304 fire-steel-shapes 384–386, 385, 392–393, Friesen, Otto von 185 395–396, 401 Frigg 240, 284 Fjǫlsvinnsmál 280 Fritsch, Sara 237–238, 271 n31 Flateyjarbók 182, 261, 274, 275, 298, 366, 457, Fritzner, Johan 224, 280 490 Frö/Freyr place-names 72–73 Fleck, Jere 498 Fróða friðr (Peace of Frotho) 466, 468 Fljótsdæla saga 284, 286, 287 Frösön, church at 324 Flóamanna saga 484 Frösön 254, 255, 340 folklands 39 Frostathing/Frostaþing/Frostuþing fornaldar sögur 66–70, 494 Hákon the Good at 134, 325, 356 Fornjótr 247 sacred spaces and 299–301, 303–304 Fornjóts sonr 247 Frostathings-Lov 345, 358 Index 625

Frostuþingslǫg 300 goðar/goði Frotho III (king) 467–468 in general 56–57 full (memorial drinking rites) 481–482 chieftains as 168–169 fulltrúi 88, 519 cultic leadership of 57, 177–178, 306, funeral feasts 330–331 drinking rituals at 481–488 and aðalból 452–453 for Ǫnundr 3 functions of 148, 164–165, 174–176 rituals performed at 3–4 gift-giving ceremonies by 359–360 royal 476–492 hereditary nature of 176–179 fylkiskirkja/fylkiskirkjur (shire churches) judical function of 398 51, 52, 141, 147, 210, 440, 441, 442 and oath-rings 376, 387–390, 398 fylki (shire) 47–48, 171, 300, 308, 440–441 and power control 532–533 Fyrby stone (Sö 56) 262 as protector of cultic sites 148 recruited from lower Norwegian elite 57 Gådi stone (U 739) 362 relations with gods 87, 89, 539 garðr 262 religious origin of 60 Gautrekr (king) 494 as sacrificers 330–331 Gautreks saga 494 selection of 175–176 Geertz, Clifford 289 sharing of goðorð by 179 Geirmundar þáttr heljarskinns 257 women as 371–373 Geirmundr Hjǫrsson 55 see also *gudhar/*gudhi; gyðjur; hofgoðar/ genealogical knowledge hofgoði; Þórsnesingar family existance of 69 Godøy (Saltfjorden) 278 importance for the elite 93 n156 gods and land rights 448–453 images of and royal inaugurations 492–493 aniconic 280 in Ynglingatal 65–66, 70–71 possessed by the devil 277–279 Gennep, Arnold van 498 personal devotion to 88–89, 517–518 geography rings carried by 427 of Svetjud 35–36 worship of many 516–518 of Trøndelag area 46–47 worshipped Gerðr 73–74, 79, 83, 412, 415–416 in Iceland 88–89 Gesta Hammaburgensis Ecclesiae Pontificium by Svear 72, 121–122 (Adam of Bremen) see also under specific gods evaluations of 113–115 god’s nails (reginnaglar) 236, 237, 257 extracts from 110–113 gœstning 528 people contesting 114–115 gold foils purpose of 114–115 in general 236–237, 407–408 see also Adam of Bremen from Bergagård (Slöinge) 417 gift-giving ceremonies 359–363, 364 from Borg (Lofoten/Vestervågøy) Girard, René 343 related to hearth 244 Gísla saga Súrssonar 286, 287, 323, 331, 359 related to post-holes 142, 145–146, Glaðsheimr 201, 202, 217, 521 229, 236, 295–296, 411–412, 417 Glælognskviða (Þórarinn loftunga) 358 as temple coins 194 Glavendrup stone (DR 209) 184, 386 from Eskilstuna 194, 409 Glitnir 218 gesture language on 414–415 goð place-names 260–261, 278 from Gudme 417 goðahús 95, 104 n27, 155–156, 284, 298 from Hauge-Tu 255–256, 410, 412 Goðakelda 260 from Helgö 128 626 Index gold foils (cont.) graves. see burial sites in general 129, 227 The Great Saga of St Olaf 460 related to hearths 243, 408 Greeks 399 related to high-seats 224 Gregory of Tours 361 related to post-holes 296, 408, Gregory VII, Pope 115, 117 416–417 Gríma (Icelandic female) 237 as temple coins 194 Grimm, Jacob 94 from Hevne (Hemne) 229, 411 Grímnismál 72, 217, 218, 238, 240, 248, 250, from Hov (Vingrom) 229 521–523, 526–527 from Husby (in Glanshammar) 194, 302, Grimstad figure 237 409 Grjótgarðr (great-grandfather of Earl interpretation of Hákon) 473 and fertility cult 412, 413–414 Grønbech, Vilhelm 8, 356, 463 n179 and hieros gamos myth 412, 414, Grønvik, Ottar 138–139, 484 415–416 guardian trees 254 and religious ruler strategies 265, 413, gud place names 519 415, 417, 543 Guðbrandr of Hundþorp 170, 274, 276, 513, as temple coins 194–195, 414, 416 520 lack of, in Iceland 265 Guðbrandsdalr 170, 274, 276–277, 428, 513, from Mære 142 520 as indicator of cultic activities 169, Gudby (Fresta parish) 182–183, 184 510 Gudby (Sorunda parish) 183–184 motifs on 411 *gudhar/*gudhi/goþi related to post-holes 296 judical function of 398 as temple coins 194 in Mälaren area motifs on 409, 411, 412 other social duties of 184–186 from Svintuna (Krokek) 194, 408–409, 410 place-name evidence for 182–183, from Ultuna 409 185, 515 from Uppåkra 237 written sources on 184–186 The Golden Bough (Frazer) 7–8 other social duties of 184–186 golden objects place-name evidence for 182–183, 185 arm rings 428–429 wearing of ritual rings 398 bracteates 103, 265, 418–419 women as 374–375 door rings 421–422, 424 written sources on 184–186 medallions 419–421 see also goðar/goði; gyðjur neck rings 424–428 Gudme (Funen) as symbols of power 407, 429 archaeological finds at 208 see also gold foils cultic organization in 519–520 goþi. see goðar/goði/*gudhi/*gudhar gold foils from 417 Götaland 36, 44 as replica of Ásgarðr 201–202 Götar 36, 37, 44 Guðmundr Arason (bishop) 61 Gotland 508–509 Gulathing/Gulaþing (Sogn) 300, 305 Götterbilder (images of gods) 264 Gulathings-Lov kristinrettur hinn nýi 261 Götterthrone (thrones of gods) 228 Gummarp stone (DR 358) 328 Grágás 391 Gunnarr Helmingr 297–298 Granmarr (king) 354, 477–478, 479 Gunnell, Terry 31 n83, 97 n11, 254 n242, 289 Gräslund, Anne-Sofie 109 n57, 118, 206, n107, 302 374–375 Gunnhildr, Queen 304–305, 306, 308, 311, Gräslund, Bo 4 n3, 486 324, 353 Index 627

Gursten stone (Sm 144) 185 Hákonarmál (Eyvindr Finnsson) 187, 353, Gustav I/Gustav Vasa (king) 11, 497 455, 518, 526 Gustavson, Helmer 388 n45 Háleygjatal (Eyvindr Finnsson) 65, 74, 80, Guta Law/Gutalagen 347–348, 355, 358, 413, 470–471, 495 443, 482 Háleygr (king) 471, 474 Guta saga 296–297, 347–348, 355, 401, Halfdan (Danish king) 399 508–509 Hálfdan hvítbeinn (king) 459, 461 Gyða (from Møre) 370–371 Hálfdan svarti (king) 462–465 gyðjur Hálfdanar saga svarta 462, 463 in Iceland 180–182 hall buildings in Mälaren area 182 n101 archaeological evidence of 4, 21 n52, Gylfaginning (Snorri Sturluson) 71–72, 73, 125–127 122, 216–217, 239, 248, 343, 385, 521–522 descriptions of 523–525 see also Snorri Sturluson salr 3–4, 5, 104 use of term 108 Häckelsäng, burial sites at 397 see also feasting/banqueting halls; warrior Häckelsäng ring 396–397, 397 halls Hænsa-Þóris saga 390 Hallencreutz, Carl Fredrik 114 Hafström, Gerhard 378 Hallfrøðr Óttarsson (skald) 83–84 Hagstugan stone (Sö 130) 362 halls (salr) 3–4, 5, 104, 219 Hákon Grjótgarðarsson (earl) 50, 51, 82 Hallsteinn Þorskafjarðargoði 177, 257 Hákon Hákonarson (king) 490 Hälsingland 396–398 Hákon Hlaðajarl 311–313 see also Forsa ring; Häckelsäng ring Hákon Sigurðarson (also called Earl Hákon Hamburg-Bremen archdiocese 61, 110, 114, and Hákon Hlaðajarl) (r. 970–995) 115 50, 311–313 hammer, of Þórr 283, 285–286, 384, 385 cultic duties of 139, 187 Hamrånge Parish hof at Guðbrandsdalr 274, 276–277, 428, burial sites in 397 513, 520 Häckelsäng ring in 396–397, 397 hof at Guðbrandsdalr of 513 Haraldr Finehair Hálfdanarson (r. 872–930) hof at Lade of 273, 421, 513 50, 51, 55, 64, 80, 170, 212, 235, 306, 338, independence from Danish kings 52 352, 426, 451, 480, 510–513 mythical descent of 80–81 Haraldr Gormsson Bluetooth (r. 958–986) mythical marriage with land of 83–85 488–489 and Þorgerðr 469–470, 473 Haraldr Gráfeldr Eiríksson (r. 961–976) 311 worship of Þorgerðr Hǫlgabrúðr 275 Haralds saga Gráfeldar (Snorri Hákon the Good Haraldsson (also called Sturluson) 311 Aðalsteinsfóstri) (r. 934–961) 25 n71, 52, Haralds saga ins hárfagra (Snorri 134–135, 139, 172–173, 187–188, 228–229, Sturluson) 493, 494, 495 325, 344, 356, 415, 455, 512, 526, 531 Haraldskvæði (Þorbiǫrn Hornklofi) 322, Håkon V (king) 448–452 352–353, 426, 471 Hákonar saga góða (Snorri Sturluson) 133, Harbarðzljóð 89 136, 148, 172, 187, 193–194, 212, 228, 242, Harðar saga Hólmverja 284, 474 273, 316–319, 320, 325, 330–332, 337, 341, harg/hǫrgr place-names 131, 157 344, 351, 352, 358–359, 359, 441–442, harg/hǫrgr structures 183, 190, 209, 216–217, 487, 493, 513 309, 394, 395 see also Snorri Sturluson hásœti (high-seat) 3–4 Hákonardrápa (Hallfrøðr Óttarsson) 80, 82, Hassmyra stone (Vs 24) 368–369 83–85 HaþuwulfR 328–330, 361 628 Index

Hauck, Karl 227–228, 418–419, 424, 427 outdoor sanctuary on 301 Hauge-Tu (Klepp) podiums/altars on 336 as cultic site 441 ritual rings on 393 gold foils from 255–256, 410, 412 weapons found at 524 Haug/Haugr (Verdal) Hellberg, Lars 67, 182–183, 504 fylkiskirkja at 210, 440, 442 Hellquist, Elof 398–399 name of 442 Hergeir (prefect at Birka) 41 haugr place-names 48, 147, 440 Herlaugr of Naumudalr (king) 473, 494 Haukr Erlandsson 461 hero cults Hauks Þáttr hábrókar 125, 266, 337–338 of Eiríkr (king) 453–475 Hauksbók 174–175 in Greece 462 Hávamál 190–191, 251, 360, 367, 390 of Hálfdan svarti 462–465 hearths of Hǫlgi 469–476 and ambulatory rituals 246 in Iceland 465–466 ritual importance of in Norway 456–465 archaeological evidence for 243–246 of Óláfr Guðrøðarson 456–462 written sources on 242–243 in Svetjud 454–456, 466–467 Hedeager, Lotte 201–204, 259 in Trøndelag area 469–472 heilagr (holy) 290, 303, 370 of Þorgerðr Hǫlgabrúðr 469–476 Heimskringla (Snorri Sturluson) 3, 24, 133, Herodotus 107, 468 272–273, 308, 311, 325, 348, 442, 480, 495 Herschend, Frands 96–97, 101–102, 223, 225, see also Snorri Sturluson 241, 307–308, 486 heiti (appellations) 28 Hersir of Naumudalr (king) 495–496 Helgafell, sanctuary at 178, 292 Hersteinn (bridegroom) 390 Helgakviða Hiǫrvarðzsonar 500 Hervarar saga 45, 119, 253–254, 327, 333, Helgö (Uppland) 515–516 assumed phallic pillar on 272 Hevne (Hemme in Trøndelag) 229, 411 bracteates from 418 hierophanies 200 ceremonial buildings on hieros gamos myth 7–8, 10, 73–74, 76, 77, 78, elevation of 207, 256 84–85, 367, 412, 420 finds at 128 high-seat posts (ǫndvegissúlar) hearths in 243 in general 231–235 high status of 129, 189 and sacred trees 177, 257 high-seats in 224 high-seat-amulets 227–228 rebuilding of 313 high-seat (hásœti) cultic feasts on archaeological evidence for 224–226 drinking rituals at 355 at Birka 224, 227, 524, 529 meals at 349–350 consecration of 295–296 gold foils from elevated position of 222 motifs on 272 importance of 4 related to hearths 243 and inheritances 488–491 related to high-seat posts 237 judical function of 488–491, 541 related to high-seats 128, 129, 224, 227 location of 222–223 related to post-holes 296, 408, in mythical world, Hliðskjálf 238–242, 416–417 288 seen as temple coins 194 posts of interpretation of name 129–130 as axis mundi symbols 236, 257 outdoor sanctuaries on, tricorn at 255, decorations of 237 256 in form of phallus 129, 237 Index 629

and gold foil figures 236–237 existence of 96–97 religious significance of in Iceland 147–155, 156–157, 195–197. see in general 231 also Hofstaðir pre-Christian context of 233–235 interiors of 219 written sources on 232–233 in mythical traditions 216–217 ritual implications of sanctification of land/places in 292–295 archaeological evidence for 235– strategic positions of 210 236, 237–238 use of term 104–106 written sources on 235, 236 in written sources 124, 133, 140, 143, 150 in religious ruler strategies 287–288 hofgoðar/hofgoði 87, 141, 148, 164–165, religious significance of 168–169, 174, 178, 179–180, 195, 197 in general 226–227, 491, 541, 544 see also goðar/goði archaeological evidence for 227–228, hofshelgi 295, 299 228, 229, 231 hof/hov place-names 131, 146–147, 157–158, written sources on 227, 228–229, 230 170, 261, 440 and royal inauguration 493–497 hofsgarðr 295, 299 silver miniature of 227, 228 Hofstaðir (Mývatnssveit) use of term 222–224 animal sacrifices at 153, 324, 339, written sources on 220–221 342–343 see also lesser high-seat; thrones ceremonial buildings/sites at high-seat (ǫndvegi) 220, 222–224, 230, 231, archaeological evidence for 95–96, 241 150–155, 214, 244–245 Hincmar of Reims 365 n211 find context of 155 hirð 531–532 functions of 153–154 Hirðskrá (Magnus Lagabøter) 492, 500 hearths in 244–245 hirðstofa 104 high-seat at 231, 295 Historia de antiquitate regum Norwagiensium high-seat posts at 235–236 (Theodoric the Monk) 274 oath-rings 387–388 Historia Norwegie 53, 72, 76–77, 124, 478 rebuilding of 314 Hjálmarr inn hugumstóri 500 ritual ending of 314–315 Hjalti Þórðarson 532 strategic positions of 215 Hjärne, Erland 528 church at 215 Hjǫrliefr (foster-brother of Ingólfr) 232 cultic activities at 507 Hjǫrungavágr, battle of 80, 83–84 cultic feasts at, calendrical 347 Hjǫrvaðr Ylfingr (king) 354, 477–478, 479 cultic leadership of 176, 177–178 Hjǫrvarðr Arngrímsson 500 gold foils from, related to post-holes 296 Hlaðajarlar. see Earls of Lade sanctification of 292 hlaut (sacrificial blood or lot) 133, 136–137, silver pendant from 286 148–149, 317–318, 331, 337 Hofstaðir (Þórsnes) hlautbolli (vessel holding blood) 133, written sources on 147–150, 150–151, 213 317–318 høfþinger see hǫfðingi/ hǫfðingjar 41 hlautteinn (sacrificial twig) 136–137, hǫfðingi/ hǫfðingjar 167–168 148–149, 317–318 hoftollr 195–197 Hliðskjálf 238–242, 288 Högåsen (in Old Uppsala) 436 hlutr 264 hǿgendiskirkja/ hǿgendiskirkjur 52 hof buildings/sanctuaries. see also Hofstaðir; Högom (Medelpad) 436 also Lade; also Mære Hǫlgi/Helgi/Hølgi (king/mythical control of chieftains over 169 ancestor) 416, 469–476 descriptions of 137 hǫll (banqueting halls) 104, 125 630 Index

Holmbjǫrn (at Gådi) 362 Husby (Glanshammar) Holmgautr (farmer at Hassmyra) 368–369 ceremonial buildings at, rebuilding of Holmqvist, Wilhelm 223, 301, 393 313 Hólmur (Iceland) 157, 216, 245–246 demarcations at 130, 302 Holtsmark, Anne 24, 240–241 farm at 170 Holy Bible, oath swearing on 390–391 gold foils from 409 Hørg, hov og kirke (Olsen) 96 Husby (Vendel), burial sites at 432, 438, hǫrgr 95, 96, 97, 104–105, 131, 144, 473 447–448 horses Husby-Långhundra (Uppland) 446 in general 323–324 Huseby (Tjølling) 211, 262 penises of 365 n213 húsfreyja 303, 365–371, 375 sacrifice of/eating of 59, 133, 135, 139, 172, Hvergelmir 203 194, 229, 254, 317, 327, 329, 333, Hversu Noregr byggðist 247 344–346, 348 Hyndlolióð 89, 332–333, 449–450 n96 Høther (king) 493–494 hypothesis, of present study 6–7, 30–31, 94, Hov (Ålen) 171, 210 160, 189–192, 316, 403, 535, 540 Hov (Orkdal) 171 Hov (Snåsa) 419 Ibn Fadlān 268, 280, 367, 383, 482 Hov (Vingrom) 229–230 Iceland hov place-names. see hof/hov place-names animal sacrifices on 339 Hovde (Ørland) burial mounds on 432–433 ceremonial buildings at ceremonial buildings/sanctuaries on. see archaeological evidence for 144–145, also Hofstaðir 211 build by chieftains from rebuilding of 314 Trøndelag 174 size of 213 hearths in 245–246 strategic positions of 209 high-seats in 229, 231–323 outdoor sanctuary at 302 locations of 215–216 Hove (Åsen) 170–171, 345–346, 440 and paying of tribute/taxes 60 n153, Hove (Byneset) 171 148, 150, 195–197 Hove (Meldal) 210 for public cultic activities 506–507 Hove (Selbu) 210, 345 size of 213–214 Hovgården (Uppland) 436 chieftains (goðar/goði) on. see also Hrafnkell Freysgoði 155, 284 Þórsnesingar family Hrafnkels saga Freysgoða 87–88, 155, 179, in general 56–57 518 cultic leadership of 57, 177–178, 306, Hrani (farmer) 457–458 330–331 Hroðgar (king) 364 and aðalból 452–453 Hrólfr Mostrarskegg 170 functions of 174–175 see also Þórólfr Mostrarskegg gift-giving ceremonies by 359–360 Hrollaugr of Naumudalr (king) 473, 494–495 hereditary nature of 176–179 Hultgård, Anders 114, 120–121, 137, 208, 251 judical function of 398 n228, 258, 271 n31, 334–336, 353, 358, and oath-rings 376, 387–390, 398 384, 454, 460 n160, 533–534 and power control 532–533 human sacrifices 261 as protector of cultic sites 148 hund/hundare place-names 39 recruited from lower Norwegian husabyar 39, 528 elite 57 husaby/husby place-names 39–40, 432 relations with gods 87–89, 539 Index 631

religious origin of 60 Inderøy 419–421 as sacrificers 330–331 Inge/Ingi place-names 67–68 selection of 175–176 Ingi Steinkelsson (r. 1080–1110) 46, 119, 254, women as 371–373 327, 515–516 Christianization on 21, 59–62, 158 Ingimundr the Old (of Vatnsdalr) 180, 230, communal assembly on 158, 390, 507, 235, 295, 490, 515–516 544 Ing/Ingi, in personal names 67–68 cultic feasts on Ingjaldr of Svetjud (king) 3–4, 29–30, 207, calendrical 323–324, 343 476–480, 501 drinking rituals at 355, 359 Ingólfr (landnámsmaðr) 231 n140, 232 funeral feasts, drinking rituals inheritance feasts 483, 488–491 at 484–485 Inn i skaldens sinn (Birgisson) 65 gift-giving during 359–360 interpretatio romana 122 meals at 346–347 Ireland 84 cultic houses 155–157 Irminsul 252 cultic images on Íslendingabók (Ari Þorgilsson) 54, 59, archaeological evidence for 285–287 65, 70, 150, 196–197, 346–347, 371, from Eyrarland 285–286 505 places beside platform 284 see also Ari inn fróði Þorgilsson wall decorations 286–287 Íslenzk Fornrit 86 written sources on 283–285 Istaby stone (DR 359) 328 cultic organization on 174–176, 505–508, 544 Janson, Henrik 115, 116, 117 cultic springs/wells on 260–261 jarl, concept of 48–49 n85 General Assembly of Þingvellir 58–59 Járn-Skeggi Ásbjarnarson 511–512 gods worshipped by, personal devotion Järrestad (Scania) 259, 302, 314, 388 n45 to 88–89 Jerusalem 84 gold foils missing from 417 Jǫfraskinna 133 gyðjur in 180–182 Jómsvíkinga saga 469 hero cults in 465–466 Jón Loptsson 61 high-seats in, judical function of 490 Jónsson, Finnur 151, 202–203 n14 long houses in 214 Jordanes 36 neck/arm rings missing from 427 Jørgensen, Lars 100–101, 102–103 outdoor sanctuaries on, sacred trees Jursta stone (Sö 140) 386 in 257 population of 54–56 Kabell, Aage 389 power structure in 29, 58–59, 90–91, 93 Kaliff, Anders 419 private churches on 60 Karlevi stone (Öl 1) 185 reason for selection 5, 6, 20–21 kenningar (kennings) religious ruler strategies on 29–30, for fire 247 86–87, 287–289 formalistic language of 22 sanctification of land/places in 291–295 for gold 426, 472 settlement of 54–55 for nails 236 sources on 54 symbolic encoding in 28 Idole (idols) 264 Kinander, Ragnar 185 Illugi Halldórsson 233 n153 kinship 93 n156 images. see cultic images Kjalarnes 388 immateriality/materiality dichotomy 27–28 Kjallakssynir family 295 632 Index

Kjalnesinga saga 213, 246–247, 260, 284, 318, Landnámabók 330, 388 allsherjargoði in 507 Kjǫtvi (king) 235 Auðr (Unnr) en djúpauðga Ketilsdóttir Klotzstuhl-Amulette 227–228 371–372 Klüwer, Lorentz Diderich 49, 141, 439 on burial mounds 433 Knirk, James E. 483 distrustful attitude to 389 Knútsdrápa (Sigvatr Þórðarson) 358 on female cultic leaders 180 Kong Sverrers Christenret 281, 443 on funeral beer 484 Königsheil 8, 11, 463 n179 on goðar as sacrificers 330–331 Konrad Maurer 60 Hallsteinn Þorskafjarðargoði in 257 Konungsbók 58–59 on high-seat posts 232–233, 234–235, Kormákr Ǫgmundarson 134, 138, 241, 275, 236–237 359, 514 on high-seats 295 Kormáks saga 333, 459 on Mære 141 Kousgård Sørensen, John 165, 184 n112, 519 on oath-rings 393 Krag, Claus 64, 66–70, 247 on podiums/altars 336 Krákumál 241 on power control in Iceland 532–533 Krause, Wolfgang 484 purposes of 452 Kringla 133 on ritual restrictions and taboos on Kristkirken 53 sanctified space 294 Kristni saga 181–182, 336 on sanctification of land/places 291 Kristoffers Landslag 496 on sanctuaries 213, 229 Kungsgården (Old Uppsala) 436 on settlement Iceland 54–55 Kungsgårdsplatån (Old Uppsala) 125–126 as source for Eyrbyggja saga 86 Kymbo (Västergötland) 427, 428 on temple-soil 333 on Þórsnesingar family 57, 518 Lade on Þórhaddr the Old 168 animal sacrifices at 133, 193–194 on Þórólfr Mostrarskegg Ǫrnólfsson burial mounds at 439 149–150, 176–177 ceremonial buildings/sites at on Þorsteinn Þorskabítr Þórólfsson 179 archaeological evidence of 140–141 on Úlfljótslǫg 174 destruction of 308–309 Lange, Wolfgang 356–357 golden door-ring at 421 Lärbro picture stone 334–335, 335 high-seats in 228 Larsson, Lars 101, 236–237 strategic positions of 209 Laxdæla saga 286, 287, 371–372, 484, written sources on 134–135, 212, 487–488, 532 513–514 The Legend of Ægidius 234 as chief sanctuary 509 “Legend of Declan” 234–235 cultic feasts at The Legend of St Eskil (Legenda sancti in general 316 Eskilli) 328 drinking rituals at 351 The Legendary Saga of St Olaf 461 cultic images at 272–273 legitimacy, concept of 16 importance of 49–51 leikr 352–353 see also Earls of Lade leirblót 281 Læraðr 251 Lejre, high-seat-amulet found at 228 Lærkefryd (Denmark) 102–103 lesser high-seat (ǫðru ǫndugi) 221, 224–225, Läffler, Leopold 400 284 Lake Mälaren area. see Mälaren area leþunger 529–530 Lamm, Jan Peder 412 Lex Frisionum 294 n14 Index 633

Lex Ribuaria 300 high-seats in 229 libations (bragafull) 3, 477, 482, 489, 498, 500 Snorri on 135, 141 Lidén, Hans-Emil 141, 143, 229, 411, 412, strategic positions of 209, 210–211 413–414, 415 as chief sanctuary 414, 509, 510–513 Liemar, Archbishop 117 church of 142 Liestøl, Aslak 380 cultic feasts at líkneski 264 in general 316, 345 Lilla Ullevi (Uppland) 131, 302, 336, 394 calendrical 321–322 Lincoln, Bruce 343, 465, 480 meals at 345 Lindqvist, Ivar 267 cultic images at 273–274 Lindqvist, Sune 68–69, 207, 430 and Earls of Lade 416 Liturgy of Niðaróss 234–235 gold foils from 142 Ljungkvist, John 436 as indicator of cultic activities 169, Lofoten islands 145 510 Löfving, Carl 391 motifs on 411 Logi 247 related to post-holes 296 long houses seen as temple coins 194 in Iceland 214 Hákon the Good in 172, 173 in Trøndelag area 144, 145 place-names in 143 Louis I, Holy Roman Emperor 43 written sources on 510–513 Lucas, Gavin 152, 339–340, 342, 343, 347 Mærina-helgi 291–292, 295 Lucretius 334 A Magic Still Dwells (Patton & Ray) 14 n38 Lund Magnús (king) 442, 492 archdiocese of 61 Magnús Einarsson (bishop) 230 figurine from 267, 285 Magnus Erikssons Landslag 401, 482 see also Uppåkra Magnus Erikssons Stadslag 482 Lunda (Södermanland) Magnús Gizurarson (bishop) 214–215 ceremonial buildings at Magnús Lagabøter (lagabœtir) (r. 1263–1280) cultic houses 130, 270 492 hearths in 244 Magnus Lagabøters landslov 449 high-seat posts in 236, 237–238 Magnúss saga in góða (Snorri high-seats in 225 Sturluson) 442 large hall 207 Mälaren area rebuilding of 313 animal sacrifices in 340, 348 cultic feasts in bracteates from 418–419 drinking rituals at 350 burial sites in. see under burial sites meals at 350–351 calendrical feasts at 322–323, 324 figurines found at 270–271, 271, 416 ceremonial buildings in see also Rällinge archaeological evidence of 127, 128, Lunderhaugen (Brønny) 255 129 Lytir (god) 337–339 and burial sites 436–439 *lytir place-names 338, 515 decorations of 208 elevation of 206–207 Maelseachlainn (king) 386 hearths in 243–244 Mære high-seats in ceremonial buildings at archaeological evidence of archaeological evidence for 141, 169, 224–225, 227–228 211 written sources on 224, 226–227 destruction of 309 size of 207–208 634 Index

Mälaren area (cont.) materiality Christianization of 43–44 people’s apprehension of reality behind cultic feasts in 28–29 calendrical 370 in studies of religion 26–28 drinking rituals at 351, 355–356, 370 materiality/immateriality dichotomy 27–28 meals at 347–348 McGovern, Thomas 339–340, 342, 343 women’s role in 368–370 meals. see ceremonial meals cultic images in 268–270 medallions 419–421 cultic organization in 504–505 Melhus (Gauldœlr) 171–172, 210, 346, gift-giving ceremonies in 361–363 439–440 gold foils from 408–409, 410 methodology outdoor sanctuaries in of present study sacred springs/wells in 258–259 chronological system used in 4 n3 sacred trees in 254–255 hypothesis 6–7, 30–31, 94, 160, place-names in 67, 131 189–192, 316, 403, 535, 540 political organization of 39–40 investigatory questions 6 reason for selection 5, 6, 20–21 outline of 29 religious ruler strategies in 29–30, 63, problems, use of different source 70–80, 287–289, 518 types 26 ritual rings in 391–398, 400 selection of regions 5, 20–21 rulers (chieftains and kings) in and stance to materiality 26–28 judical function of 398 use of sources 14–15, 21–26 other social duties of 184–186 Meulengracht Sørensen, Preben 25 n71, 137, place-name evidence for 182–183, 187, 212–213 185 Meyer, Birgit 27–28 and power control 29 Miðgarðr 262–263 as sacrificers 326–330 Miller, Daniel 28 n81 warrior lords 527–533 Mímir’s well 202, 203 wearing of ritual rings 398 miniatures, from Lund 267, 285 women as 368–370, 374–375 minni (memorial drinking rites) 481–482, written sources on 184–186 487 sanctification of land/places in Minnurs brunn (Old Uppsala) 258 296–297 Mithras cult 108, 419 settlements in, central places in 38 Mjǫllnir (hammer of Þórr) 283, 285–286, societal strata in 37–38 384, 385 theophoric place-names in 517–518 Moberg, Lennart 182 n101 worship of Freyr in 415–416 Mogk, Eugen 24 see also Svear; Svetjud; under specific sites Møllenhus, Kristen 420 Malsta stone (Hs 14) 450 monasteries 130 mana 8, 11 Mora stone 496–497 Mannhardt, Wilhelm 7 Møre (Norway) 370–371 maritime expeditions 530 Morgan, David 27–28 Markus Skeggjason 328 Mǫrnir 367 Marold, Edith 455 Motz, Lotte 74 Marstrander, C. J. S. 483 mounds, burial. see under burial sites martial hall buildings 30–31 Munch, Gerd Stamsø 244, 296, 346, 414, 416 matblót 281 Mundal, Else 22 “the material turn” 27–28 Muspilli (Old High German poem) 300 Index 635

Myhre, Bjørn 255 cultic leaders/leadership, of mythical brides 83–84 woman 365–368, 370–371 mythical fathers cultic organization in 503–504, 510–514 not always gods 74 gold foils from 411 of Ynglingar royal family 29, 70, 71–72 hero cults in 456–465 mythical marriages kings/rulers of between Freyr and Gerðr 413, 415–416 hirð of 530–532 between Óðinn and Skaði 413, 414, 416 mythical status after death 455–460 of Earls of Lade 83–85, 413, 414, 416 and óðal 450–452 of Ynglingar family 75–80 and power control 12, 531–532, 533 see also hieros gamos myth population of, emigration to mythical mother, of Ynglingar royal Iceland 54–56 family 73–74 ritual rings in 398 mythical symbolism. see cosmic symbolism see also Trøndelag area mythical world feasting halls in 87 oath, origin of word 398–399 ritual demarcations in 262–263 oath-rings descriptions of 387 names Forsa ring 377–383, 386, 391 of burial mounds 446–449 goðar/goði and 387–390 of farms 440–442 open 388 n45 see also personal names; place-names/ and podiums 393 geographical names on stone from Tängelgårda 398, 399 Närke 36 swearing on 376 Näsström, Britt-Mari 10 oaths neck rings 424–428 formulae for 400 níð 65 swearing of Niðaróss 53, 60, 221 on crosses 391 Nielsen, Ann-Lili 392–393 in erfi/inauguration Nielsen, Michael Lerche 117 ceremonies 499–500, 544 Niflheimr 203 on Holy Bible 390–391 Nilsson, Bertil 298 on oath-rings 387–390 Njáls saga 212, 230, 274, 276, 390–391, 428, on ritual rings 376, 398–402 471, 513 on weapons 390, 399 Nora stone (U 130) 450 objects, symbolic encoding of 28 Nordahl, Else 126 óðal right 448–452 Nordberg, Andreas 460 n157, 523, 525 Oddr Snorrason 273, 274, 275, 421, 511–512, Nordisk hedendom (F. Ström) 11 n27 513 Nordmyra (Njutånger) 269 Odens brunn (in Old Uppsala) 258 Nóregs konunga-tal 463 Odensbrunn (Vendel) 260 Norr, Svante 71 Oðindisa (housewife) 368–369 Norra Åby stone (Sö 86) 267, 384 Óðinn Norway arm rings of 428 burial sites in 432 on bracteates 418–419 Christian kingship in 52 and cosmic eschatology 526, 534, Christianization of 21, 52 544–545 creation of 51 Hákon of Lade identified with 83–84 cultic feasts in, funeral feasts 482–484 images of 111, 266, 282 636 Index

Óðinn (cont.) ǫndvegi (high-seat) 220, 222–224, 230, 231, marriage to Skaði 413, 414, 416 241 as mythical father of Hákon of Lade 81–82 ǫndvegissúlar (high-seat posts) 177, 231–235, as mythical father of Ynglingar 71–72 257 mythical home of 201, 202, 217, 526, “the ontological turn” 27–28 533–534 Ǫnundr (king) funeral feast of 3, 207, mythical throne of 238–242 476–480 in place-names 517 óðal 48, 55 rituals associated with 293 óœðra ǫndvegi (lesser high-seat) 221 worship of 121–122, 518 Das Opferfest von Lade (Düwel) 136, 317 see also Valhǫll myth oral traditions 22, 23 ǫðru ǫndvegi, ǫndugi (lesser high-seat) 221, Orkneyinga saga 119, 327–328, 364 224–225, 284, 525 Ormknös (Björkö) 445 see also high-seats Ǫrn í Arnarnesi 55 Ofegh of Hjortsta 378–379, 382–383, 398, Oseberg ship burial 287, 432 402 Østlandet (Norway) 440 Oklunda stone (N 288) 297–298 Óttarr vendilkráka (king) 93 n156, 98, Óláfr (grandson of Unnr) 372 447–448, 449–450 n96 Óláfr rauði (king) 361 Ottarshögen (Uppland) 431, 438, 447–448 Óláfr Guðrøðarson 65, 456–462 Otterbjörk, Roland 369 Óláfr Haraldsson (r. 1015–1030)(Saint) 52, outdoor sanctuaries 53, 214, 221, 261, 272, 276–277, 279, 310, cultic activities performed in 96 321–322, 344, 366, 373, 451, 457–458, developing into temples 107–108 512, 524–525 examples of 102–103 Óláfr kyrri Haraldsson (r. 1066–1093) at Helgö 301 220–221 at Lilla Ullevi 302 Óláfr pái Hǫskuldsson 286 at Old Uppsala 301 Óláfr sœnski (“Skötkonung”) Eiríksson restrictions on 302–303 (r. 995–1022) 44–45, 326 ritual demarcations at 130, 262–263, Óláfr Tryggvason (r. 995–1000) 50–51, 52, 59, 301–302 272–275, 276, 278, 279, 285, 308–309, ritual slaughter at 190–191 322, 352, 421, 451, 457, 469, 510–511 sacred springs/wells in 257–261 Óláfr Þáttr Geirstaðaálfs 456–459 sacred trees in 177, 203, 250–255, 256, Óláfs saga Helga (Snorri Sturluson) 122–123, 257 193, 276–277, 320–322, 322, 331, 344, written sources on 250–251 442, 512, 514, 529 see also sacred spaces see also Snorri Sturluson outline, of present study 29 Óláfs saga Tryggvasonar 212, 273, 275, 276, Ovid 107 277, 278, 279, 285, 308–309, 421, 469, 489, 510, 518 pagan images. see cultic images see also Snorri Sturluson Paine, Crispin 27–28 Old Testament, sacrificial rituals in 136 Páll Jónsson 61 The Oldest Saga of St Olaf 461 Pálsson, Hermann 86 Olrik, Axel 493 Peace of Frotho (Fróða friðr) 466, 468 Olsen, Magnus 147, 180–181, 440, 455, pendants 503–504 fire-steel-shaped 384–385, 385 Olsen, Olaf 96, 135–136, 149, 151, 175–176, from Hofstaðir 286 195–196, 319, 389, 412 with Þórr hammer symbols 384 Ǫlvir 321–322, 512 Peringskiöld, Johan 116 Index 637

Perkins, Richard 285 with tuna 38, 131, 189 personal names with vang 147, 170 with Ing or Ingi 67 with vítr/véttr/vættr 262 with viðr 478 with vi/vé 131–132, 188, 297–298, 396 Pfahlgötter 268 Plate, S. Brent 27–28 phallus symbols platforms. see podiums/altars (stallar/stalli/ assumed pillar at Helgö 272 stallr) bronze statue from Rällinge as 122, Plato 247 266–267, 267 Plutarch 464 figurines found at Lunda 271 podiums/altars (stallar/stalli/stallr) posts in the hall as 129, 237 and blood sacrifices 331–334 at Old Uppsala 111 and divination rituals 336–339 stone from Lunderhaugen as 255 and oath-rings 393 Phrygian rites 10 outdoor/indoor 334–335 picture stones written sources on 279–280 Buttle Änge 225, 226 The Poetic Edda 201 from Gotland 78 Post, Lennart von 258 from Lärbro (Hammers 1) 334–335, 335 post-holes Tängelgårda 399 gold foils related to Tjängvide 79 from Bergagård (Slöinge) 417 Pietroassa ring 388 n45 from Borg (Lofoten/Vestervågøy) 142, pigs 351 145–146, 229, 236, 295–296, 411–412, pit-houses 145, 153, 157, 216 417 place-names/geographical names from Gudme 417 with blót 260–261 from Helgö 296, 408, 416–417 as direct sources 23 from Mære 296 as evidence for power cultic sites 131, 143, 146–147, 157–158 legitimization of 4, 8, 52, 205, 227, 244, in Sweden 182–183, 185 403 dísir worship 125 monumentalization of 30, 488–491 existence of priestly class 165 and religion 5, 204 worship of gods 121–122 symbols of 376, 402–403, 407, 428–429, with Frö/Freyr 72–73 430–431, 540 with goð 260–261, 278 power control with harg/hǫrgr 131, 157 in Iceland 29, 90–91, 93 with haugr 48, 147, 440 in Norway 12, 531–532, 533 with hof/hov 131, 146–147, 157–158, 170, in Trøndelag area 29, 47–49, 52, 261, 440 91–92 with hǫrgr 157–158 in Svetjud 12, 91–92, 527–529, 533 with hund/hundare 39 in Mälaren area 29 with husaby/husby 39–40, 432 pre-Christian cults with Inge/Ingi 67–68 and Christian prohibitions 268, 279, 281, in Ingjaldr account 479 344, 347, 355, 443–444, 473 with *lytir 338, 515 existence of 95–102, 108, 115 in Mære 143 sources for 22–23 in Mälaren area 67–68, 131 studies of with stafr 264, 383 and regional differences 14 n38 with staver 268–269 static character of 20 theophoric 517, 517–518 symbolic encoding of 28 638 Index pre-Christian cults (cont.) in religion 14 n38 see also cultic feasts; cultic images; cultic in religious ruler strategies 30–31, 90–93 objects; cultic organization; cultic regional similarities 31–32, 92 rituals regional variations. see regional differences present study religion chronological system used in 4 n3 material study of 26–28 hypothesis of 6–7, 30–31, 94, 160, political factors of 5, 204 189–192, 316, 403, 535, 540 see also Christendom; pre-Christian cults investigatory questions in 6 Religion (journal) 27–28 outline of 29 religious ruler strategies selection of regions 5, 20–21 in general 287–289 sources use in 14–15, 21–26 and cultic images 287–289 and stance to materiality 26–28 eros and death in 78 textual research for 7 and gold foils 265, 413, 415, 417, 543 use of sources in 14–15, 21–26 and golden medallion from Inderø 421 Price, Neil 4, 12–13, 126, 374 high-seats in 287–288 priesthood on Iceland (Þórsnesingar family) 29–30, concept of 165–167 86–87, 287–289 existence/non-existence of 164–166, 191, and legitimization of power 4–5, 15 204, 514–515 in Mälaren area (Ynglingar family) and landownership 170 29–30, 63, 70–80, 287–289, 518 priests, concept of 165–167 model for analyzing 17–19 primus inter pares 91 and mythical marriages 75–76, 79, 85 private, use of term 109 n57 regional differences in 30–31, 90–93 prohibition, of cultic rituals 268, 279, 281, regional similarities in 31–32, 92 344, 347, 355, 437, 443–444 role of sanctuary in 290 public, use of term 109 n57 studies on 5, 7, 20 in Trøndelag area (Earls of Lade) 29–31, queens. see women 80–85, 287–289 and Valhǫll myth 30–31, 533–535 Rafnsson, Sveinbjörn 452 see also cultic leaders/leadership; sacral Rällinge (Lunda), bronze statue from 122, kingship 266–267, 267, 416 Rerir (king) 494 Ramqvist, Per H. 437 rex dei gratia 11 n27 Ranheim (Trøndelag) rex sueonum 40 ceremonial buildings/site at Reykjahólar (Iceland) 214–215 archaeological evidence for 143–144, Rhezelius, Johan 116–117, 260 211 Rígr (god) 422 ritual ending of 314–315 Rígsþula 422, 498 strategic position of 209–210 Rimbert 40–41, 337, 454–455, 455, 529, 530 staffs at 280–281 rings Rappendam Mose 103 arm rings 428–429 Rauðr (chieftain) 277–278 neck rings 424–428 Rauðs þáttr in ramma 278 as symbols of power 428–429 reginnaglar (god’s nails) 236, 237, 257, 295 see also church-door rings; rings, ritual regional differences rings, ritual concept of 14 n37 amulet rings 392–395 organization of sanctuaries 31 breaking of 399–400 Index 639

door rings 421–422, 423, 424 placing ruler on high-seat on mound fire-steel-shaped 395–396 493–497 Häckelsäng ring 396–397, 397 swearing in 499–501 on Helgö 393 royal sacrifice 76–77 in Norway 398 Rude Eskildstrup (Zealand) 427 oath swearing on 376, 398–402 Rudolf of Fulda 252 oath-rings rulers. see cultic leaders/leadership descriptions in sagas 387 runic inscriptions Forsa ring 377–383, 386, 391 from Adelsö (U 11) 528 and goðar/goði 387–390 on Aspa löt stone (Sö Fv1948;289) 479 open 388 n45 on Björketorp stone (DR 360) 328 roles of 401–402 on bone from Sigtuna 362–363 as symbols of power 376 on Bredgränden stone (U 942) 318–319 Þórr’s hammer rings on coffin from Sanda (G 181) 225, 226 in general 392 as direct sources 23 and Forsa ring 384 on Eneby stone (Sö 145) 450 from Helgö 393 on Forsa ring 378–383, 396 from Old Uppsala 394–395 on Fyrby stone (Sö 56) 262 from Viby Gård 395–396 on Gådi stone (U 739) 362 rites on Glavendrup stone (DR 209) 184, 386 of passage 227 on Gummarp stone (DR 358) 328 use of term 109 n55 on Gursten stone (Sm 144) 185 ritual demarcations 262–263 on Hagstugan stone (Sö 130) 362 ritual feasts. see cultic feasts on Hassmyra stone (Vs 24) 368–369 Ritual. Perspectives and Dimensions (Bell) on Istaby stone (DR 359) 328 341–342 on Jursta stone (Sö 140) 386 ritual spaces. see sacred spaces on Karlevi stone (Öl 1) 185 rituals. see cultic rituals on Malsta stone (Hs 14) 450 Rǫgnvaldr (king) 65 on Nora stone (U 130) 450 Rǫgnvaldr of Orkney 364 from Norhuglo (Norway) 168 n31 Rǫgnvaldr the Old (earl from Götaland) 45 on Norra Åby stone (Sö 86) 267, 384 Rǫgnvalds Þáttr ok Rauðs 277, 278, 285 on Oklunda stone (N 288) 297–298 Rök stone (Ög 136) 69, 188, 306, 378 from Old Uppsala 115–117 Roman world, temples in 107–108 on powerful women 373 Romania 388 n45 on Rök stone (Ög 136) 69, 188, 306, 378 Rønne, Preben 144, 281 on Sandsjö stone (Sm 71) 69 n37, 450 Røskaft, Merete 146–147, 530–531 on Stenkvista stone (Sö 111) 122, 123 Røthe, Gunnhild 473–474 on Stentoften stone (DR 357) 328–329, royal funerals 476–492 358, 361 drinking rituals at 482–485 on Tjängvide stone (G 110) 378 libations at 481–482 on Tune stone (Norway) 483–484 Ǫnundr of 3, 476–480 on Turinge stone (Sö 338) 362 rituals performed at 3–4 on Väppeby stone (U 703) 362 see also royal inaugurations runic stones royal inaugurations 492–499 and cemeteries 118 and genealogical knowledge 492–493, Christian 120 501–502 see also runic inscriptions lack of evidence for 498–499 Rüpke, Jörg 167 640 Index

Rūs people 268, 383, 482 see also animal sacrifices; blood, Ruthström, Bo 380–381, 396 sacrificial; blood sacrifices Rydh, Hanna 397 sacrificial feasts. see cultic feasts Rydving, Håkan 14–15 n38 Saga Óláfs Tryggvasonar (Oddr Snorrason) 273, 274, 275, 278, 511–512, sacerdotes/sacerdos 514–515 513 sacral kingship Sahlgren, Jöran 486 in Germanic area 11 saints, cults of royal 462 and hierogamy 7–8, 10, 76, 77 Sakshaug/Sǫrshaugr (Inderøy) and “Königsheil” 8, 11, 463 n179 fylkiskirkja at 210, 440 material dimensions of 13 name of 441–442, 446 in Near East 9–10 salr (hall) 3–4, 5, 104, 219, 486 in Scandinavia 8–12 sanctification social level of 12 of land/places studies on in general 542 in general 5 in Iceland 291–295 biased focus in 12–13 in Mälaren area 296–298 and decontextualized ritual restrictions and taboos on perspectives 12, 15 294–295 global comparisons in 9–10 in Trøndelag area 294–295 history of 7, 13–14 Sanda (Gotland) static character of 10–12, 20 stone coffin from 225, 226, 228, 269 themes focused on 13 Sanda (Uppland) use of primary sources in 14–15 archaeological finds at 130, 183, 194 use of secondary sources in 10 ritual rings at 395–396 views on 7–9 Sandnes, Jørn 47 n79 see also cultic leaders/leadership Sandsjö stone (Sm 71) 69 n37, 450 sacred bonds (vébǫnd) 299–301, 303, 305 Santesson, Lillemor 329, 361 sacred lakes 262 Saur (dog) 422, 493–494 sacred spaces Saur (ruler) 440 enclosed 301–302 Saxo Grammaticus 72, 196, 428, 467–468, of the Frostathing 299–301 494 in Iceland 291–295 Schjødt, Jens Peter 22 n54, 413 n25, 463 n179, in Mälaren area 296–298, 301–302 498 ritual ending of 314–315 Schøning, Gerhard 49, 141, 439 ritual restrictions/taboos on 294–295, Schück, Henrik 7 302–303 Second Merseburg Charm 418 ruler as protector of 305–307 See, Klaus von 164, 358 in Trøndelag area 294–295, 299–301, 302 Shetelig, Haakon 432 violations against sanctity of 303–307 shields 390 see also outdoor sanctuaries shire churches (fylkiskirkja) 51, 52, 141, 147, sacred springs/wells 202, 203, 250–251, 210, 440, 441, 442 258–261 Sibbi (local ruler) 185, 188, 306 sacred trees 177, 203, 250–255, 256, 257, Siegfried (Danish king) 399 310–311 Sigvatr Þórðarson 303, 358, 369–370, 459, sacrificers, ruler as 277, 288, 316–317, 325–331 524 sacrifices Sigmundsson, Svavar 157, 261 to álfar/álfablót 459 Sigríðr of Svetjud (queen) 421 to Freyr 331, 355 Sigtuna 43–44, 45, 362–363 Index 641

Sigurðardrápa (Kormákr Ǫgmundarson) high-seats 228–229 134, 139, 187, 188, 275, 359, 514 informants used for 138–140 Sigurðr Hlaðajarl Hákonarson 50, 52, political conditions 48 133–135, 138, 172, 187, 188, 194, 228, 273, role of hearths in rituals 242 311, 316–317, 319, 325, 359, 415, 513 role of Kings and Earls in 187 Sigurðr Þórisson 320–321 ruler as sacrificer in 325, 331 Sigurðsson, Jón Viðar 360, 531, 532 sacrificial rituals 136–137, 148–149, Sigvaldr of Jómsborg (earl) 501 317–318, 337, 341 Simek, Rudolf 523 on dating of Ynglingatal 64 size, of ceremonial buildings 207–208 on divine descent of royal families Sjøvold, T. 445 70 Skaði 413, 414, 416 on drinking ceremonies 481 Skálholt 61, 214–215 on Guðbrandsdalr 276–277 skaldic poetry on Hákon Sigurðarson 81, 311, 513 as direct sources 22 on Hálfdan svarti 462 interpretation of 22 on Haraldr Finehair 50, 51, 80 language and diction in 28 on high-seats and inheritance 488 see also under specific poems on Hǫlgi and Þorgerðr 473 Skáldskaparmál 286, 472–473 on Ingjaldr (king) 476–477, 485–486 skáli (feasting hall) 104, 214–215, 487–488 on King Magnús 442 Skalla-Grímr Kveld-Úlfsson 55, 433 on Lade Skara, bishopric in 45, 267 in general 272–273 Skeggi (son of Friðgerðr) 181, 336 dismantling of 308–309 Skei (Steinkjer) 255 door ring 421 Skírnismál 73, 238–239, 412, 471 drinking rituals at 351 Skjǫldungar 92 sacrificial rituals at 336 Skog tapestry 269 on Mære 273–274, 279 Skúli of Lade (earl) 173 on Melhus 171–172 Skúli Þorsteinsson (skald) 472 on mythical father 71–72 skulls, at Hofstaðir 152–153 on mythical marriages 76, 413 skurðgoð 264, 266 on mythical mother 73 Skyllberg, Eva 351 on mythical throne 238–240 slaughter, ritual. see animal sacrifices on Óláfr Haraldsson 523–525 Slipshavnskogen (Funen) 427 on Óláfr Tryggvason 451, 510–511 Smith, Jonathan Z. 14 n38 on Ǫlvishaugr 441 Snorra Edda 104, 201, 202–203, 286 on Rauðr 278 Snorri Sturluson on ring of Aðils 425 informants of 133 on ritual toasts 358–359 views of on royal inheritances/inaugurations on ceremonial buildings/sites 124, 493–494, 499 207–208, 216–217, 221 on sacrificial rituals 331–332 on cult of Freyr 466–467 on Sigvatr Þórðarson 369–370 on cults in Trøndelag on Sigurðr Hlaðajarl 513–514 in general 133–135, 172 on Sveinn Forkbeard 489 ceremonial meals 344 on Sveinn Úlfsson (also called Sveinn cultic feasts 191, 193–194, 316–319, Estridsson) 492–493 320–321 on Þórr 122, 274 exterior of buildings 212 on Uppsala sanctuary 124–125, 193, gift-giving ceremonies 359 479, 514 642 Index

Snorri Sturluson (cont.) Stjórn 332 on Vingólf 202–203 stofa 104, 219 on Yggdrasill 250 stones. see under decorations; picture stones; writings of runic inscriptions as direct sources 23 Storm, Gustav 469, 471 reliability in reconstructing pagan Strabo 107 religion 3–4, 23–25, 173, 202–203 strategy, concept of 16–17 Söderby (Danmark parish) 418 Strid, Jan Paul 184 Södermanland Strindafylki 47, 50, 51, 80, 141, 170, 440 common law at 39 Ström, Åke V. 11, 89 place-names in 68, 73 Ström, Folke 11, 76–77, 84, 164 runic stones from 122 Ström, Krister 392 sanctuaries in 130 Strömbäck, Dag 84, 233 Solomon’s Temple (Jerusalem) 107, 205 Strút-Haraldr 488 Sonatortek (Egill Skalla-Grímsson) 88–89 Sturlunga saga 257 songs, ritual 355 Stutz, Ulrich 60 Sorunda (Södermanland) 183 Sundborn (Dalarna) 269 sources Sunnmøre (Norway) 448–449 critical approaches to 23–25 Sutton Hoo 352 for Eyrbyggja saga 86 Svear on Iceland 54 disposing of rulers by 515–516 lack of contemporary 21 folklands of 39 on pre-Christian cults 22–23 gods worshipped by 72, 121–122 on Svear 35 livelihood of 36 on Trøndelag area 46 and paying of tribute 192–193 used in rulers of other studies 10 cultic roles of present study 14–15, 21–26 archaeological evidence for 189 Sparlösa stone (Vg 119) 422, 424 written sources on 188 spekingar 220 high kings of St Botvid 119 in general 40–41 St John’s Order 130, 409 Christianization of 43–45 staff (stafr) 279, 280 inaugurations of 496–497 stafr place-names 264, 383 petty king 38, 40 stallar/stalli/stallr (podiums/altars). see and power control 527–529, 533 podiums/altars seaborne hegemony of 41 Stalsberg, Anne 346 society of 36–37 “the stance to materiality” 27–28 sources on 35 statues see also Svetjud; Uppsala sanctuary; phallic-shaped 122, 266–268, 267 Ynglingar family see also figurines Sveigakot (Iceland) 214, 215–216 staver place-names 268–269 Svein Estridsson (Sveinn Úlfsson/Sveinn Steinkell (r. 1060–1066) 45–46, 121 Ástriðarson) (r. 1047–1076) 114, 121, Steinsland, Gro 10, 73–74, 75–78, 82, 93 n156, 492–493 413, 414–416, 420, 470–471 Sveinn (minor chieftain) 276 Steinvǫr (hofgyðja) 180 Sveinn Álfífusonr, King 358 Stenkvista stone (Sö 111) 122, 123, 384 Sveinn Forkbeard (r. 997–1014) 488–489, Stentoften stone (DR 357) 328–329, 358, 361 491, 501 stirps regia 8 Sveinsson, Einar Ól. 86 Index 643

Sveinungr (magnate) 389 Tjängvide stone (G 110) 78, 79, 378 Sveriges runinskrifter 116, 117 toasting rituals 356–359 Svetjud toponymy. see place-names/geographical Christianization of 43–44 names cultic feasts in tréguð 264, 266 drinking rituals at 354–355 trémaðr 264 funeral feasts, drinking rituals at treuddar (tricorns) 255, 256 482 tribute/taxes cultic organization in 544 paying of economic growth of 36–37 on Iceland 60 n153, 148, 150, 195–196, geography of 35–36 195–197 hero cults in 454–456, 466–467 in Mälaren area 192–193 power structure in 12, 91–92, 527–529, at Old Uppsala 192–193 533 in Trøndelag area 133, 193–195 religious ruler strategies in 30–31 triclinia/triclinium (feasting halls) 101, 115, see also Svear; Uppsala sanctuary 125, 269, 354 Svíagríss 425 tricorns (treuddar) 255, 256 Svintuna (Krokek, Östergötland), gold foils Trøndelag, origin of name 47 from 194, 408–409 Trøndelag area Sweden, emergence of 37 bracteates from 419 burial sites in. see under burial sites taboos, on cultic rituals 294–295, 297 ceremonial buildings in Tängelgårda (Gotland), picture stone archaeological evidence for 140–141, from 398, 399 143–144 tapestries 287 and burial sites 439–442 taxes. see tribute/taxes elevation of 211 temenos 126 exterior of 212–213 temple coins 194–195, 414, 416 hearths in 244 Temple of Apollo (Delphi) 107 high-seat posts in 231–232 temples, concept of 106–108 high-seats in 228–229 temple-soil 333 size of 211–213 templum sources for 132–133 use of term strategic positions of 209–210 in general 107–108 chieftains in. see also Earls of Lade by Adam of Bremen 125 and control of sanctuary 168–170, textual research, for present study 7 170–172 Theodoric the Monk 274 migration to Iceland 174–175, 177 Thietmar of Merseburg 454–455 resistance against Hákon the thing/þing assembly 47 n79 Good 172–173 Thor. see Þórr as sacrificers 325–326 Thorstein (chieftain) 362 Christianization of 52, 172, 282 thrones 111, 217, 227, 237, 240 court circles in 299–300 see also high-seats cultic activities in Thurgot 45 in general 133–135 Timaeus (Plato) 247 chieftains role in 187–188 Tindr Hallkelsson 470 collaboration between chieftains in Tissø (Zealand) 101, 103, 259, 302 172–174 tithes 195–196 hero cults in 469–476 Tiundaland 39 paying of tribute during 133, 193–195 644 Index

Trøndelag area (cont.) Þórði gelli 177–178 cultic feasts in Þórðr Freysgoði Ǫzurarson 180–181 gift-giving during 359–360 Þorfinnr (earl) 221 meals at 344–346 Þorgeirr Hávarsson 214 written sources on 138, 193–195, Þorgeirr Ljósvetningagoði Þorkelsson 59 316–317 Þorgerðr Hǫlgabrúðr 275, 284, 416, 428, cultic images in 469–476 chieftains and 276–277 Þorgrímr Helgason (chieftain) 284, 330, at Hustad church 282 388 and Norwegian Laws 279–282 Þorgrímr Kjallaksson 177–178, 306, 308 at Rauðr’s sanctuary 277–279 Þorgrímr Þorsteinsson (chieftain) 87, 178, written sources on 275–276 286–287, 323, 359 cultic organization in 187, 510–514, 544 Þórhaddr the Old (from Mære) 168–169, 174, geography of 46 229, 233, 291–292, 295, 306, 314, 333 gold foils from 411 Þorkell leppr Þjóstarsson 179 long houses in 144, 145 Þorlákr Þórhallsson 61 outdoor sanctuaries in 255, 257 Þorlaugr gyðja Hrólfsdóttir 180 as patria principalis 53 Þórólfr bægifótr 230 power structure in 29, 47–49, 52, 91–92 Þórólfr Kveld-Úlfsson 212–213 reason for selection 5, 6, 20–21 Þórólfr Mostrarskegg Ǫrnólfsson religious ruler strategies in 30–31, 80–85, biography of 86–87, 176, 292–294 287–289 cultic leadership of 306, 330–331 sanctification of land/places in 294–295 cultic leadership of family of 57, 177–178 sources on 46 high-seat posts of 234 see also under specific sites hof at Hofstaðir (Þórsnes) in 147–150, Trondheimsfjord 209–210 283–284 Tuna (Badelunda) 375 passing mention 314 Tuna (Vendel) 259–260 see also Hofstaðir (Þórsnes) tuna place-names 38, 189 Þórr Turinge stone (Sö 338) 362 carriage of 274–275 Turner, Victor 498 compared with Jove 122 Turville-Petre, E. O. G. 367 hammer of 283, 285–286, 384, 385, 524 typology, of religous authority 165–167 images of 111, 231, 237, 266, 267–268, 271 n31, 274, 276–277, 279, 283, 284, Þangbrandr (priest) 59 285–286, 287 Þáttr af Ǫgmundi dytts ok Gunnari helminga importance of 283–284 (also called Ǫgmundar þáttr dytts) 266, as most honoured 274 298 personal devotion to 86–87 Þattr Þorsteins uxafóts 174–175 in place-names 517 Þing-assembly 8 relationship between hammer and Þinghelgi (assembly-sanctity) 299, 303, fire-steel 385–386 304 relationship with oath-rings 386 Þingvellir 58–59 rituals associated with 293 Þjódólfr inn fróði ór Hvini 64–67, 69–70 symbols of 393 Þóraldi (ármaðr) 321, 512 as wind-raiser 285 Þórarinn loftunga (skald) 358 worship of 121–122, 176, 177, 179, 518 Þorbiǫrn hornklofi (skald) 322, 352 Þorrablót 324 Þórðar saga hreðu 175 Þórr’s hammer rings Index 645

in general 384, 392 Uppsala högar och Ottarshögen from Helgö 393 (Lindqvist) 430 from Old Uppsala 394–395 Uppsala sanctuary from Viby Gård 395–396 animal sacrifices at 348–349 Þórsdrápa 370 archaeological evidence of 125–127 Þórsnesingar family as axis mundi 263 defending of sacred space by 306 becoming Christian 44–45 fight with Kjallakssynir family 295 burial customs in 118–119 gods worshipped by 283–284, 518 burial mounds at 373–374, 434, 436 as regional rulers 57 ceremonial buildings at 98 relations with gods 539 ritual ending of 315 religious ruler strategies of 29–30, as chief sanctuary 509 86–87, 287–289 cultic activities at 42–43, 112–113, 124–125 Þorsteinn Þorskabítr Þórólfsson 177–178, cultic images at 111, 266–267 490 decorations of 208 Þorvaldr Koðránsson (skald) 181–182, dynastic kingship at 444–445 336–337, 459 elevation of 206 Þrœndir 48, 52, 170 excavations at 43 cultic organization of 187, 510–513 exterior of 218 divine descent of 30 funeral feast at 3–4 law of 300 high-seats in 227 organization of 53, 173 interior of 110–111 resisting Christianization 52, 139, outdoor sanctuary at 301 172–173, 321, 512 outdoor sanctuary in see also Mære; Trøndelag area sacred springs/wells at 258 Þrœndalǫg 47, 300 sacred trees in 250–251 Þrymr 494 written sources on 111–112, 250–251, Þrymskviða 494 253–254 Þuriðr gyðja Sǫlmundardottir 180 paying of tribute to 192–193 Þuriðr hofgyðja Véþórmsdóttir 180–181 ritual demarcations at 262–263 ritual rings at 394–395 Úlfljótslǫg 150, 174–175, 176, 195, 387, 389, 401, runic inscriptions at 115–117 505–506 size of 207 Úlfr Uggason (skald) 286 surroundings 111–112 Ullr (god) 271 n31 views on Ultuna (Uppland) 409 of Adam of Bremen 110–113 Unnr (Auðr) 371–373, 484, 487–488, 532 of Snorri Sturluson 124 Uppåkra worship of Freyr at 466–469 cultic house at 99–100, 100, 102, 313 written sources on 514–516 door rings at 421–422, 423 see also Svear gold foil figures from 237 Uppsalaauðr 193 weapons found at 524 Uppsalakulten och Adam av Bremen (ed. Uppland Hultgård) 114 Christianization of 115, 118–120 Urðr’s well 202, 203, 251 common law at 39 Útgarðr 203–204 sanctuaries in 129–130 Upplandslagen (Law of Uppland) 268, 322 Værnes 173, 210, 531 Uppsala, interpretation of name 486–487 vættir (female supernatural creatures) 281 646 Index

Valaskjálf 218, 238, 239 vítr/véttr/vættr names 262 Valhǫll myth Vittersen lake 262 cosmic tree besides 251 vi/vé place-names 131–132, 188, 297–298, descriptions of 521–523, 525–526 396 on picture stones 78 Vǫlsa Þáttr 366–368 and religious ruler strategies 30–31, Vǫlsunga saga 254, 494 533–535 Vǫluspá 216, 250 and warrior ideology 533–534, 544–545 von Friesen, Otto 185 valkyria pendants 374 von Post, Lennart 258 Vallhagar (Gotland) 224–225, 243 von See, Klaus 164–165, 358 Valsgärde, burial sites at 118, 241, 374–375, 438 Vor folkeæt i oldtiden (Grønbech) 8 van Gennep, Arnold 498–499 Vǫlsi 367 Vang (Trøndelag) 147, 170, 440 Vǫlundarkviða 390 vang place-names 147, 170 Vries, Jan de 164, 341, 508 vangr, use of term 301 Vanir 459, 518 n60 Wach, Joachim 167 Vanlandi (king) 361 Waffentanz/Waffenspiel 352 Vápnfirðinga saga 180 Wahlberg, Mats 486 Väppeby stone (U 703) 362 wall decorations 286 Västgötalagen 482 Walter, Ernst 135–136, 318, 319 Västmanland 37, 39 warrior halls Vatnsdœla saga 156, 180, 214, 230, 235, 295, at Birka 523–534 324, 490 Valhǫll myth and 524–527, 533–534 vé, use of term 104 n27, 165, 188, 294–295, weapons found at 524 299–300, 305 warrior ideology 533–534, 544–545 vé place-names. see vi/vé place-names warrior lords 527–533 vébǫnd (sacred bonds) 299–301, 303, 305 warriors veizlustofur/veislustofa 104, 219–220, 221, as einherjar 217, 248, 455, 522–523, 242 533–534 Vellekla (Einarr Helgason) 80, 82, 139, 187, and Valhǫll myth 533–534 311–312 watch-tower-theory 241 Vésteinsson, Orri 155, 158, 215–216 water, cultic activity and 202, 203, 250–251, Vestergaard, Elisabeth 498 258–261 Vestfold, Oseberg ship burial 287, 432 Watt, Margrethe 414–415 Viby gård (Uppland) 395–396 Wealhtheow (queen) 364 viðr, in personal names 478 weapons, found at hall buildings 524 Víga-Glúmr 156, 388–389 Weber, Gerd Wolfgang 24 Víga-Glúms saga 155–156, 324, 388–389, 393, Weber, Max 16 401 Wellendorf, Jonas 233, 234, 235 Víga-Hrappr 274 Wessén, Elias 67, 72 n52, 116, 377–378, 478 Vigfusson, Guðbrandr 86 West Slavs 196 Vikstrand, Per 68, 131, 157–158, 183, 278, 487, wetlands, cultic activities at 258–260 505 Widengren, Geo 453 n120 Vingólf 201, 202–203, 217 Williams, Henrik 319 Vita Anskarii (Rimbert) 40–41, 454–455, wind-raisers 285 529 Wodan. see Óðinn Vita sancti Botvidi 119 Wolfred 310 Index 647 women Ynglingatal (Þjódólfr) burial site of 373–375 burial customs in 68–69 role of cultic feasts in 190 in cultic feasts 13, 363–365 cultic roles of rulers in 188 as cultic leaders dating of 64–65 in Iceland 371–373 on death of Ǫnundr 478 in Mälaren area 368–370 divination rituals in 337 in Norway 370–371 genealogy in 65–66, 70–71 in drinking rituals 363–364 kenning for fire in 247 in gift-giving ceremonies 364 mythical father of Ynglingar in 71–72 húsfreyja 365–371 mythical marriages in 75–76, 77–78, 413 transgressing gender boundaries 365 on neck rings 425 n213 obscure place-name mentioned in 479 worship on Óláfr Guðrøðarson 461 of Freyr 8, 271–272, 415–416, 466–469 origin of traditions in 66–70 of many gods 516–518 on Óttarr 447 of Óðinn 121–122, 518 personal names in 67 of Þórr 121–122, 176, 177, 179, 518 place-names in 67–68 Wulfstan 41 on rich rulers 361 on ruler as protector of sacred space 305 Yggdrasill 203, 250, 255 seen as níð 65 Ymir (giant) 343, 464, 465 smearing of sacrificial blood in 333 Ynglinga saga (Snorri Sturluson) 3–4, 64, on Sweden 37 70, 72, 73, 76, 124, 207, 322, 351–356, 413, Yngvi 461, 466–467, 480, 486, 501 as other name for Freyr 67, 70, 82 see also Snorri Sturluson as ruler honorific 67 Ynglingar family Yngvi (king) 188 divine descent of 29–30, 70–74, 270, 468, Yngvi und die Ynglinger (Baetke) 8 539 Yule feasts 320, 322–323, 351, 352–353, 370, mythical father of 29, 70, 71–72 500 mythical marriages of 75–80 mythical mother of 73–74 Zachrisson, Sune 409 religious ruler strategies of 70–80, 270 Zachrisson, Torun 255, 259, 394, 446 sources on 64 Zalmoxis myth 468