Arv Nordic Yearbook of Folklore 2020

ARV Nordic Yearbook of Folklore Vol. 76

Editor

ARNE BUGGE AMIUNDSEN OSLO, NORWAY

Editorial Board

Lene Halskov Hansen, København; Fredrik Skott, Göteborg; Suzanne Österlund-Poetzsch, Helsingfors (Helsinki); Terry Gunnell, Reykjavik

Published by THE ROYAL GUSTAVUS ADOLPHUS ACADEMY UPPSALA, SWEDEN

Articles appearing in this yearbook are abstracted and indexed in European Reference Index for the Humanities and Social Sciences ERIH PLUS 2011–

Editorial address: Prof. Arne Bugge Amundsen Department of Culture Studies and Oriental Languages University of Oslo Box 1010 Blindern NO–0315 Oslo, Norway phone + 4792244774 e-mail: [email protected] http://www.hf.uio.no/ikos/forskning/publikasjoner/tidsskrifter/arv/index.html Cover: Kirsten Berrum For index of earlier volumes, see http://www.gustavadolfsakademien.se/tidskrifter/tidskrift/arv

Distributor eddy.se ab e-post:[email protected] Box 1310, S-621 24 Visby Telefon +46(0)498 25 39 00 http://kgaa.bokorder.se

Printed in Sweden Exakta Print, Malmö 2020 Contents

Articles on Digital Humanities and Folklore Peter M. Broadwell & Timothy R. Tangherlini: Geist, Geest, Geast, Spøgelse: Challenges for Multilingual Search in Belief Legend Archives...... 7 Venla Sykäri: Digital Humanities and How to Read the Kalevala as a Thematic Anthology of Oral Poetry...... 29 Trausti Dagsson & Olga Holownia: Legends, Letters and Linking: Lessons Learned from Amassing and Mapping Folklore and Viewing as Part of 19th-Century Culture Creation...... 55 Katherine S. Beard: The Eitri Database: A Digital Humanities Case Study...... 75 Pia Lindholm: Finland-Swedish Folklore as a Versatile Online Tradition Source...... 93 Mari Sarv & Janika Oras: From Tradition to Data: The Case of Estonian Runosong...... 105

Other articles Anders Gustavsson: Nineteenth-Century Cholera Epidemics in Sweden from a Popular Perspective...... 119

Book Reviews Aarbakke, Thea: Forfattermuseumsfunksjonene (Lars Kaijser)...... 151 Cocq, Coppélie & DuBois, Thomas A.: Sámi Media and Indigenous Agency in the Arctic North (Alf Arvidsson)...... 156 Danielson, Eva: Skillingtryckarna (Gunnar Ternhag )...... 159 Frykman, Jonas & Löfgren, Orvar: Den kultiverade människan (Mats Lindqvist)...... 161 Gerndt, Helge: Sagen ‒ Fakt, Fiktion oder Fake? (Anders Gustavsson)...... 165 Gustafsson, Sofia: Järtecken (Anders Gustavsson)...... 166 Jarlert, Anders (ed.): Reformationen i Lund ‒ Malmö ‒ Köpenhamn (Anders Gustavsson)...... 168 Klintberg av, Bengt: Vänster hand och motsol (Ane Ohrvik)...... 169 Lindqvist, Katja (ed.): Kompetens i museisektorn (Teemu Ryymin).... 170 Løvlie, Birger et al. (eds.): Tru på Vestlandet. Tradisjon i endring (Anders Gustavsson)...... 172 Lundqvist, Pia: Ett motsägelsefullt möte (Anders Gustavsson)...... 173 Ramsten, Märta: De osynliga melodierna (Karin Hallgren)...... 175 Rasmussen, Tarald (ed.): Å minnes de døde (Ulrika Wolf-Knuts)...... 177 Resløkken, Åmund Norum: ’Ein lut av det nære levande livet’ (Torunn Selberg)...... 179 Roos, Anna Marie: Goldfish (Ingvar Svanberg) ...... 180 Schön, Ebbe: Ängel med bockfot (Anders Gustavsson)...... 182 Strand, Karin: En botfärdig synderskas svanesång (Inger Lövkrona)...... 183 Ternhag, Gunnar: Jojksamlaren Karl Tirén (Krister Stoor)...... 186 Geist, Geest, Geast, Spøgelse Challenges for Multilingual Search in Belief Legend Archives

Peter M. Broadwell and Timothy R. Tangherlini

Abstract We describe the challenges of devising a federated multilingual search engine for a series of otherwise unconnected tradition archives. The international project, Intelligent Search Engine for Belief Legends (ISEBEL), solves several challenges from this broader problem. First, the collections are made searchable through a single interface by implementing an Open Archive Initiative-Protocol for Metadata Harvesting (OAI-PMH) system for three target collections (Netherlands, Germany and Denmark). To address the problems of searching across multi lingual collections, we devise a hybrid system that relies on Neural Machine Translation using a Transformer architecture, and expert created tables of domain specific terms. By integrating these approaches, users can search from a single portal over a corpus of nearly seventy thousand legends in English. Results are returned from all of the language groups in the collection, while places are projected on a map. Keywords: Legend, multilingual search engine, language, computational folkloristics, archives

Introduction If legends collected in the late nineteenth century are any indication, rural Danish farming communities were the sites of fairly consistent – and alarm ing – supernatural intrusions. Witches threatened the production of milk and the churning of butter; ghosts and revenants rampaged through cemeteries and barns; the fields, dotted with boulders thrown by long vanished giants, were filled with hidden folk who, when angered, threatened to tear the roofs off farmhouses; basilisks banged about in old beer kegs; serpents slept in the stone walls separating fields; the local church, surrounded by even big- ger serpents, was also the stomping grounds of Satan; and brownies made life indoors unpredictable at best (Kristensen 1980). People venturing outside the village could be waylaid by robbers in the woods, drowned by river spirits, seduced by elves, or lured to their deaths in the ocean by merfolk. Standing between villagers and these threats were Lutheran ministers, cunning folk, wandering Norwegians, brave farmhands or a person’s own quick wits or dumb luck. Late nineteenth century Denmark’s belief landscape, in 8 Peter M. Broadwell and Timothy R. Tangherlini short, was just as filled with dangers and disruptions as were the political, physical and economic landscapes of the time (Gunnell 2005; Tangherlini 2005). Indeed, it is this interdependence between belief and the lived world that undergirds interpretive perspectives on legends that reveal their crucial role in the ongoing negotiation of cultural ideology in times of considerable change (Gunnell 2009; Taussig 2010; Tangherlini 2013a). Of course, these observations about Denmark also lead to a question: while all this was going on in Denmark, what was happening in neighboring countries? Certainly, nineteenth century southern Sweden seemed to be in the grip of similar, yet subtly different, supernatural threats, while Norway’s tall mountains and deep fjords played host to a distinctive set of violently threatening beings (Klintberg 2010; Christiansen 1977). Northern Germany and the Netherlands again shared some overlap in the belief landscape, but the particular characteristics of their versions of these supernatural threats, along with other unheard-of threats (at least from the perspective of Danish farmers) ensured slight differences in the belief topography (Bächtold- Staubli et al. 1927–42; Saatkamp & Schlüter 1995). Iceland, perhaps because of its island status, had a divergent set of beliefs, although aspects of its witchcraft and wizardry had some commonalities with earlier, potentially pan-Nordic, beliefs (Mitchell 1997; Mitchell 2011; Gunnell 2012). Understanding the areas of overlap, commonality, and potential influence between and across not only national but also linguistic borders, provides greater nuance to interpretations of the impact of beliefs on worldview in local communities (Dundes 1971; Dégh 2001). Yet developing research collections from disparate archival and print resources that can help answer these questions is a considerable challenge. The nineteenth and early twentieth centuries mark the heyday of broad national folklore collecting and archiving, at least in Northern Europe (Estonian Folklore Archives 2017). Although most national archival collections were initially envisioned in the context of emergent Romantic nationalist ideals and, as such, to support the imaginings of the “nation” as a coherent culturally homogeneous entity, the linguistic diversity of “nations” necessarily also meant that many of these archives included records in all of the languages of the nation in question, whether those languages were conceived of as dialects or as wholly independent (Bendix 1997; Hult 2003). Searching even in a fairly narrow “national” archive, thus, could tax the linguistic abilities of many researchers (Gunnell 2010). The Danish folklore archive, for instance, includes stories recorded not only in standard Danish, but also in the various Jutlandic dialects that can differ quite dramatically from standard language not only orthographically but in vocabulary and word use (Olrik 1910; Feilberg 1886; Tangherlini 2013a). This phenomenon is compounded many times over in archives such as those from the Netherlands, which include records in Frisian and numerous Dutch Geist, Geest, Geast, Spøgelse 9 dialects, or the various archives in Germany, Norway and Sweden that preserve the wide dialectical diversity of those areas. A broad ethno-historiographic approach to the study of folklore, and folk legend in particular, raises considerable problems related to the discovery of resources (Hudson 1966; Sturtevant 1966; Christiansen 1996). If one wants to trace, for example, the interaction between economic developments, changes in physical environment, and the impact of the shifting sands of politics on the daily lives of a rural population through the lens of folk legend, then one must be able to search and find variants of legends that provide insight into the storytelling of people from specific regions at specific times (Schmitt & Tangherlini 2018; Tangherlini 2005; Tangherlini 2013a). An interest in studying witchcraft and the local response to witches in con- tinental northern Europe, to use another example, might be motivated by an interest in the varying perceptions of the power of ministers and priests across the region, as well as the ability of others to play the role of mediator of the threat of witchcraft (Tangherlini 1998; Tangherlini 2000; Stokker 1995). Such a broad comparative approach might allow one to understand the culturally acceptable allomotifs for fulfilling various motifemic slots in analogous stories from separate regions, thereby providing a map of what a community viewed as a threat, and what they viewed as potential strategies for dealing with that threat, as well as their evaluation of the various power structures that permeated their daily lives (Dundes 1964; Tangherlini 2018; de Certeau 1984). Yet, to ensure that the conclusions of this type of regional study are not based on cherry-picking a small number of legends that, by design, confirm the research hypothesis, these studies would need to rest on research corpuses produced by applying the same search parameters across many different archives. These searches would also need to be delimited by time and be cognizant of region, but capacious enough to accommodate significant language variation and areas of collection that can cross the state borders that determined the scope of many earlier folklore archives (Trieschnigg et al. 2012). In the past, to carry out a search of this nature would require a huge amount of time, a great deal of travel, an enormous amount of patience, the goodwill of local experts, as well as a degree of insight into local laws governing access and use of archival resources. Even for archives limited to a single country, the search process would be arduous. Gunnell describes what a researcher would need to do to carry out a fairly simple search on the legend collections housed at the Arnamagnæan Institute in Reykjavik: “Anyone wanting to research particular international types or motifs simply had to plough through every single one of the various Icelandic collections in the Árnastofnun, which could take weeks or months” (Gunnell 2010: 154). This process would need to be repeated at each of the target archives. And, at the end of these explorations, one might need to make the circuit 10 Peter M. Broadwell and Timothy R. Tangherlini once again, adjusting one’s search strategies to what one learned in the first pass of the archives.

Comparative Archival Research and the Promise of Digitization A desideratum of folklore research has, ever since the inception of the field, been the ability to work comparatively across cultural and linguistic boundaries, an idea that informed the large indexing projects of the late nineteenth and early twentieth centuries (Uther 2004; Thompson 1966; Christiansen 1977). More regional approaches to indexing are always done with an eye toward this type of broad research (Klintberg 2010). While some of the early indexing efforts were clearly predicated on a desire to understand developmental trajectories of stories as well as the origins of various beliefs (Krohn 1926), a great deal of recent research has been interested in situating belief in place and time, and using this as a means for exploring how people use stories as a means for the ongoing negotiation of cultural ideology (Tangherlini 1994). Both approaches – and many lying on the continuum between them – rely on good collections and even better finding aids. National archives have by design been largely available for folklore research since their establishment. Yet, the exigencies of local needs bal- anced with the limited availability of archivist time to index and classify materials necessarily meant that the archives supported local questions first, and responded to more broad scale considerations second. The development of international indices for aspects of traditional expressive culture ensured at least, on some level, a degree of “interoperability” (Uther 2004). A researcher interested in stories about a particular topic or a particular ATU number could write to a series of archives and receive fairly comprehensive responses, although the responses would, more often than not, be in the format of the local archive, and might require a degree of knowledge of a particular collection’s structure to understand those responses. Similarly, there was no obvious manner in which to evaluate the thoroughness of the reply – did it represent the results of a comprehensive search or was it the result of a sampling of the archive’s holdings? Often, there was little support in this type of search for “query expansion”, where the original research terms could be expanded, based on local archival knowledge or other forms of domain knowledge, to include other related terms. The digital revolution held the promise of making archival research both more efficient and more thorough. For the digital revolution to be fully realized, however, it required a degree of technical knowledge and digitization resources such as high quality high-speed scanners that were frequently out of reach of all but the most privileged of archives. Yet nearly all archivists recognized that, if archival records were digitized and stored Geist, Geest, Geast, Spøgelse 11 with appropriate metadata, then a researcher would be able to access, at least in theory, the records needed to support her work by generating a search over the archive – the results would be returned much quicker than through the physical search of a paper or tape-based archive as described by Gunnell, or by requesting resources via the mail (Gunnell 2010). Even prior to the broad-based adoption of the internet in the late 1990s and the migration in the early twenty-first century of archival collections to databases queried through web-based search interfaces, the results of a search could be returned via email in a matter of hours or days, instead of weeks and months. Similarly, a researcher could expand queries quickly, learning from the results of very precise queries to generate searches that would substantially increase the recall over the archival collection (Singh et al. 2016). A great deal of archival thought and folkloristic expertise ensured that the migration from wholly analog records (i.e. paper; photographic and moving images recorded on film; sound recorded on cylinders, vinyl and magnetic tape; material artifacts) to largely digital records proceeded much more quickly than if the archives had to start from scratch. The accessibility and low cost of well-supported internet infrastructure bundles, such as the open source Linux-Apache-MySQL-PHP (LAMP) stack that included robust solutions for relational databases and web-based querying, also ensured a degree of commonality across many folklore archives. Since these archives generally consist of collections created by individual collectors or networks of collectors from interactions, either in person or through mailed surveys, with informants whose performances of informal expressive culture were the main targets of collection, one could formally represent the overall structure of these archives as a tripartite network of persons – places – things (Tangherlini & Broadwell 2014). In this model, persons can play multiple roles such as archivists/classifiers, collectors, or informants. Places can likewise have multiple roles, related to any of the persons, or to the things the persons perform (e.g. places mentioned in a legend), while things can be conceived of as covering any of the broad types of folklore stored in the archives. Importantly, this conceptualization of folklore – as informal cultural expressive forms circulating on and across social networks embedded in time and space – does not presuppose that folklore is the product of a deliberate collecting process, and can just as easily be applied to “born digital” folklore that may be archived asynchronously and indirectly. The demands of the digital folklore archive, then, are closely aligned with the archival practices of the past century. The more recent prolifera- tion of metadata standards throughout the library and archival world aligns remarkably well with the preexisting deep description of folklore materials in the various archives, and has allowed for the quick adaptation of digital approaches to folklore archiving that one encounters across most of Northern Europe, even in the face of severe budgetary constraints. This 12 Peter M. Broadwell and Timothy R. Tangherlini convergence on relatively similar systems, the ability to create meaningful queries that are not constrained by previous indexing regimes, and the ability to integrate search results with helpful user interfaces including various types of visualizations, has allowed for a rapid increase in the usability of the now-digital archive. The metadata for the Icelandic folklore collections of Jón Árnason, for instance, form the bedrock for the geographic navigation system of the Sagnagrunnur system developed by Gunnell and Dagsson, which can then be integrated with other folklore collections at the AMI (Gunnell 2010). Similar infrastructure has been used to support a geographically based user interface to search legends in Sweden and Norway through the Sägenkartan (Skott 2017). In the Netherlands, a longstanding effort of the Meertens Instituut has led to the development of perhaps the most comprehensive online folklore collection in Europe, with a clear and easy to use interface at the Volksverhalenbank (Meder et al. 2016). Fairly complex database models predicated on hypergraphs have allowed researchers at the Wossidlo Archive to create a series of research interfaces to explore the remarkable holdings of that archive, and to bring together otherwise disparate representations of a single story performance in WossiDiA (Bruder et al. 2015). Danish folklore collected by Evald Tang Kristensen has formed the basis for a number of experiments in computational folkloristics, including various interfaces for user search such as ETKspace and The Danish Folklore Nexus (Tangherlini 2013a; Tangherlini & Broadwell 2014) as well as flexible network classi fiers that leverage the power of existing classification systems, while taking advantage of the metadata associated with stories and storytellers in the people-places-things model of folklore noted above (Abello et al. 2012; Broadwell et al. 2017). Other analytic approaches based on the digital representation of the Tang Kristensen folklore collection include multi-scale statistical representation of legend repertoires (Tangherlini 1994), a series of geographically based studies of distribution patterns (Tangherlini 2010; Broadwell & Tangherlini 2016; Broadwell & Tangherlini 2017), topic modeling (Tangherlini & Leonard 2013; Broadwell & Tangherlini 2015), text reuse (Broadwell et al. 2018), and the impact of transportation infrastructure on Tang Kristensen’s collecting practices (Storm and Tangherlini 2018). While all of these studies have considerably altered not only the type of questions that one can address in folkloristics but also the manner in which one can address them, they are still largely constrained to a single archive, or a very small subset of archives (as in the case of the Swedish/Norwegian legend map noted above). In addition, nearly all of these archives restrict search to the language or languages represented by the records in their archival collections. In short, research that crosses the boundaries of multiple archives still requires one to visit each archive individually – albeit virtually – and create Geist, Geest, Geast, Spøgelse 13 archive-tailored searches. Results are often presented in a manner that pre- cludes easy download and few, if any, of the archives provide an API for more automated search of the archival records. Furthermore, laws governing access can, in certain instances, make it impossible for international researchers to access sophisticated local search interfaces. Consequently, despite the convergence of folklore archives on relational databases and the broad acceptance of international metadata standards for representing the data, there are several major barriers to realizing the goal of search across multiple archival resources, irrespective of language, and to work with those results in sophisticated analytical environments, be they geographic information systems (GIS), text mining environments such as Voyant (Sinclair & Rockwell 2016), or simply as a series of documents subjected to the standard philological and close reading methods of folklore. While one can, given a certain degree of patience, some limited programming background, and a fairly broad linguistic competence, find results for a search on “ghosts” in northern Europe by visiting the various online archives mentioned above and capturing those results locally, it is nearly impossible to work with those results in a “macroscopic” fashion (Tangherlini 2013b). Börner reminds us that the macroscope provides a “vision of the whole, helping us synthesize the related elements and detect patterns, trends, and outliers while granting access to myriad details. Rather than make things larger or smaller, macroscopes let us observe what is at once too great, slow, or complex for the human eye and mind to notice and comprehend” (Börner 2011: 60). To realize an international – or at least broadly regional – macroscopic approach to folklore collections, one needs to have several facilities built-in to a search engine: (i) the ability to search once and retrieve results from multiple archives; (ii) the ability to search in a single language, and receive high-confidence results from those searches across all of the collections, irrespective of the language of the target repositories; and (iii) the ability to work either in the search environment or locally with the returned results, ideally with various tools for visualization and statistical analysis (Ilyefalvi 2018). The international project ISEBEL (Intelligent Search Engine for Belief Legends) has addressed at least the first two of these challenges, and the lessons learned during the process of developing the infrastructure for multi-lingual search across disparate folklore archives can help chart the path forward for addressing the third challenge as well as other future projects (Meder 2018; Schmitt & Tangherlini 2018).

Multilingual Search for Belief Legends Search across collections of multi-lingual data poses a series of well-known challenges for which various strategies have been proposed (Aula 2009). In the study of folk legend, an example of an information need could be posed 14 Peter M. Broadwell and Timothy R. Tangherlini as follows: “I am interested in ghosts, and legends about ghosts, from folklore collections in German, Dutch, and Danish.” The problem presupposes the existence of searchable collections for each of the target languages, and some way to access those collections. Earlier, a search might require an exchange of letters or physical presence at each of the archives, as well as time spent learning about the archival practices in each of the countries. It would similarly presuppose linguistic competence in each of the target languages. Over the past three years, researchers associated with collections representing Denmark, Germany and the Netherlands have, under the umbrella of the aforementioned “Intelligent Search Engine for Belief Legends” (ISEBEL) project, explored ways in which to address this type of information need, recognizing that a series of disjoint keyword searches across the various collections, while a reasonable strategy, would not be likely to provide a coherent dataset for study. An important first step was agree- ment on a minimal data representation, expressed as an XML schema. Such a schema would facilitate integrating the collections on a “middle level” tuned to the search requirements of researchers interested in legends. While describing the inner workings of the entire ISEBEL system is beyond the scope of this paper, the XML schema allows the collections to be accessed via the Open Archive Initiative Protocol for Metadata Harvesting (OAI- PMH) in a standardized form, and to be indexed in a manner that facilitates rapid search over the concatenated collections (Van de Sompel et al. 2004). The concatenated collections can then be indexed and act as the target for search. This approach solves several problems at once: first, it obviates the need of researchers to visit multiple archives and create individual search strategies for each collection and, second, it allows each collection local control over what materials they offer for this type of concatenated search. This latter consideration is particularly important when collections have limitations on which materials researchers from outside of the archive are allowed to access. While the OAI-PMH structure thus solves several major barriers to developing research collections that cross both national and linguistic borders, it does not immediately address one of the most vexing challenges to working across languages, namely searching in one language and receiving appropriate results from all of the languages in the integrated collection. Indeed, it was this problem that motivated the earlier indexing work prevalent in folkloristics in the mid twentieth century. Search across collections of multi-lingual data poses a series of well-known challenges for which various strategies have been proposed, including translating the search string into the target languages at the time of search, running a search string in one language against translations of the target data which have been translated previously either into a common language or into the search language, Geist, Geest, Geast, Spøgelse 15 limiting the vocabulary allowed in a search to terms in a multilingual thesaurus, or some combination of those strategies along with the use of preexisting indices, such as the ML index (Trojahn et al. 2014; Christiansen 1977). Each strategy has its distinct advantages, as well as fairly significant disadvantages. Translating all of the legends in one collection into the languages of all the other collections is time consuming (computationally expensive) and, depending on the quality of the translation required, quite likely impossible. Given the potentially very large number of languages, including dialects and other low resource languages that may be found in folklore archives, the scale of the problem could very quickly become intractable. Translating search strings into the target languages might be possible, but given challenges in both syntax and semantics (i.e. the search program would need to guess at the intention of the searcher, and translate that intention into multiple languages) unlikely to succeed in a highly specialized domain such as folklore. To sidestep these problems, one could limit searches to some sort of a controlled vocabulary so that the translations of those terms were of high quality and of clear relevance to each of the collections. This approach necessarily channels the searcher into constrained searches that presuppose a range of information needs. Because of the scale of the collections, even for a modest pilot such as ISEBEL, hand translation is impossible, and some degree of machine translation is inevitable (Table 1). The goal of any such machine translation, then, is to provide the best possible representation of a language for the lowest computational cost.

Collection Total Number of Languages in Total number of records in ISEBEL collection keywords Verhalenbank (NL) 26866 Dutch, Frisian, 28820 numerous dialects WossiDia (DE) 11622 High German, 2079 Low German, numerous dialects ETKspace (DK) 31086 Danish, Jutlandic 1375 dialects Table 1. An overview of the size of the ISEBEL collections, including total number of records, languages represented in the collection, and total number of keywords associated with the collection.

ISEBEL’s Hybrid Approach to Multilingual Search Given the lack of resources for many of the languages represented by ISEBEL – a characteristic of many folklore collections – we devised a multipart strategy to facilitate multilingual search. The goal was to present the researcher with multiple avenues for search, including faceted search, 16 Peter M. Broadwell and Timothy R. Tangherlini where pre-existing collection-based indices could be turned on or off as appropriate, and keyword search, where the existing keyword indices for each corpus could be used either as a primary or secondary search param- eter. Importantly, the ISEBEL XML schema explicitly captures this metadata (collection-based keywords and indices) during the harvesting process. Even with this limited data, one could in theory mimic a multilingual search across all of the collections by simply entering terms for the query in each of the target languages. A search on the terms for “ghost” in German, Dutch and Danish – “Gespenst OR spook OR spøgelse – properly returns hits from each of the collections but requires the user to do her own multi-lingual query expansion. While words for ghosts are easily generated through dic- tionary look-up, more unusual terms, such as the Danish lindorm or the Dutch kabouter would likely not be found in most dictionaries. These two problems in multilingual search for folklore archives can be summed up as: (i) the need for a common search language and (ii) the need for a multi lingual thesaurus of domain specific terms. In the ISEBEL project, we address these two issues separately, and then combine the results in the concatenated database records that form the target for search. Generating domain specific terms for each corpus was a fairly straightforward task. We designated a domain expert for each target collection, who was then tasked with creating a list of terms that were likely to appear in the text of a legend but unlikely to be translated by a machine translation system. To illustrate this, the Danish term nisse is widely used in the Evald Tang Kristensen collection, appearing in 585 stories, but improperly tran slated by most machine translation systems as “elf” rather than the more appropriate “house elf” or “brownie.” After each corpus expert had created a list, they were then asked to translate these lists into English, with multiple translations stored in separate rows. Similarly, the Danish term lygtemand has multiple possible English entries including “fen fire”, “jack-o’-lantern”, and “sprite” to name but three, while the terms bjærgfolk and højfolk both resolve to the English term, “hidden folk” while højfolk can also be trans lated as “mound folk”. Once all of the lists were concatenated, unaligned terms from each of the other languages or dialects were translated into each of the collection languages if a corresponding term existed. These terms were then added to the indexing for each of the stories such that a Danish story that included the term bjærgfolk was indexed with the English term, “hidden folk.” A Danish story with the term højfolk would similarly include the “hidden folk” indexing, and a search on “hidden folk” would subsequently return both stories. Since German and Dutch stories are indexed via the combined domain specific term list, this search would also return records from each of those collections for Erdmännchen, Unterirdische, aardmannetje, verborgen volkje Ierdmännken, ierdmantsje, and forburgen folkje (See Table 2). The incorporation of the domain specific terms greatly Geist, Geest, Geast, Spøgelse 17 increases the accuracy of search, particularly given the specialized vocabulary of storytellers, while also simplifying the creation of a comprehensive search string. Due to the manner in which these additional terms are included in the search, the domain specific terms list can be updated easily, thereby increasing the accuracy of searches. Because of this feature, any archive that joins the ISEBEL consortium in the future will be required to generate a list of domain specific terms with their English language equivalents.

German Dutch Danish English Low German Frisian Erdmännchen aardmannetje bjærgmænd hidden folk Ierdmännken ierdmantsje Hexe toverkol heks witch Hex tsjoenster Wassermann meerman havmand merman Wasserman mearman Table 2. Some examples of the domain specific key terms from the ISEBEL alignment table.

Along with keyword indices and domain specific terms, ISEBEL makes use of neural machine translation (NMT) to produce “dirty translations” of all the stories in the harvested corpuses. These dirty translations are then indexed and exposed to the search engine. NMT, which represents the current state of the art in machine translation, received a considerable boost in the past two years with the development of what is known as the “Transformer” architecture. This architecture was developed specifically for machine translation tasks, where an input sequence must be accurately transformed into an output sequence. While earlier models of machine translation tended to be built on rules-based models of translation between two languages (direct translation) and, later, on abstraction based models (interlingua translation and transfer systems), current machine translation systems make use of very large corpuses of paired sentences (sourcetarget) for training purposes and take advantage of the power of multi-layered neural networks to learn, in an unsupervised manner, high dimensional feature spaces (Hutchins 1995). These new models rely on a mathematical approximation of “attention” that encodes how source and target words are likely to be associated with each other across different contexts as a means for training the encoders and decoders that comprise the Transformer (Sutskever et al. 2014; Cho et al. 2014; Bahdanau et al. 2014). For ISEBEL, we implement a modified version of the OpenNMT-py BPE (byte-pair encoding) pipeline, a “multi-headed attention” model that, along with the encoding of the input, greatly speeds the training of the model and the accuracy of its output (Vaswani et al. 2017; Sennrich et al. 2105). The BPE or “word piece” component is a powerful refinement to Transformer training practices that uses byte pairs (roughly equivalent to letter pairs) as its primary token vocabulary, rather than full words. This has the effect of shrinking the overall unique token count of the corpus, greatly reducing the amount of GPU RAM required. It also tends to produce more accurate 18 Peter M. Broadwell and Timothy R. Tangherlini translations. Part of this increased accuracy, particularly for the tasks related to creating indexable translations of legends, is that the BPE pipeline typically does a better job of recognizing untranslatable terms (e.g., proper names or supernatural creatures) and leaving them unaltered in the final translated output, whereas full word-based systems may drop these terms from the vocabulary entirely as a memory-saving measure due to their rar- ity. This latter attribute of word-based systems often makes it quite a chore to recover the original forms of dropped words and retain them in the output translations. BPE also tends to make the Transformer system better able to encode how minor orthographic and morphological changes might modify the “meaning” (in a very narrow, computational sense) of a word. Currently, every record in ISEBEL is translated into English via the byte- pair encoding Transformer approach described above. Training these translation models requires large “parallel” corpuses of equivalent sentence pairs in the source and target languages, most of which we obtained from the Open Parallel Corpus (OPUS) project (Tiedemann 2016). The availability of such open text sets for moderately and even under-resourced languages greatly accelerates the process of creating rich sourcetarget training sets, and provides a clear path forward for the addition of future folklore collections into the ISEBEL framework. As of this writing, fully pretrained models for translating to and from some of the most highly resourced joint translation counterparts of English (French, German, Chinese) are available for download, as are rudimentary trained Transformer models based on many of the languages and paired texts from the OPUS project. It is not always easy to adopt models others have trained, nor is this necessarily desirable when working with regional and historical corpora. Such models may be quite accurate for translating modern prose or speech as they have been trained on extremely large contemporary news, web, literature, and multi- media corpuses, but they often fail to perform adequately when translating, for example, a nineteenth century regional dialect. Simply training a model with the historical texts alone also is not effective due to their much smaller number. Often the best method – which we adopt for ISEBEL – is to train a model with numerous contemporary parallel corpora such as those from OPUS to establish the model’s linguistic “backbone,” and to augment this with additional materials from more period appropriate, publicly available translation resources such as the Bible, or even manually translated texts from our own corpuses. The Transformer architecture and NMT models’ voracious appetite for training data also make it possible and frequently desirable to train a single English-targeted model with corpuses from multiple closely related source languages, such as Danish and Swedish; the resulting model is often more accurate when translating from either source language to English than models that are trained separately. When possible, we refine our models by comparing the accuracy of Geist, Geest, Geast, Spøgelse 19 the NMT translations with hand generated “gold standard” translations – though not the same texts used to train the model – and, after discovering areas of systematic failure, update the model to solve these problems (Table 3). It was this type of an evaluative approach that led us to adopt the BPE method and to devise and incorporate the “domain specific term” search augmentation workflow described above, as we discovered a series of terms that the initial Transformer consistently failed to translate.

Original Story Prior translation Current Translation Gold Standard Phrase (Transformer) (Byte pair encoding) Så siger den fattige Then the poor man Then the poor Then the poor man Mand, at han havde says that he had a man says he had a says that he had a en Karl, der rimelig- farmhand who could farmhand who could farmhand who prob- vis kunde begrave bury her so that she reasonably bury her ably could bury her hende sådan, at hun wouldn’t come again. so that she shouldn’t so that she wouldn’t skulde ikke komme come again come back. igjen. Bjærgfolkene i den The in that The mound men in The mound dwellers bakke havde også hill also sometimes that hill also some- in that mound some- undertiden deres had their out. times had their linen times put their linen linned ude. out. out. Heksemesteren havde had told him The witch master The witch master had sagt ham, hvordan how things would go had told him how it told him what would det vilde gå til. would go happen Men en gang var But once , But once the mound But once the mound bjærgfolkene, der which lived in dwellers lived in the folk who lived in the boede i højen, blevne , was mound, they were mound were offended fornærmede på dem upon them insulted by them En karl traf en gang A farmhand once A farmhand once A farmhand once met et par snoge, der took a couple of took a couple of a pair of grass snakes spøgte, og slog da that songs that ghost, and that were slithering den ene ihjel. and then killed one killed one about, and he killed one of them Table 3. Examples of the various “dirty” translations of Danish phrases from a small selection of legends. In the original Transformer, untranslated words were replaced with . The quality of the translations using byte-pair encoding is significantly better than the earlier implementation. Incorporating the domain specific keywords provides even better search capabilities.

Initially, we had attempted to create a ni x nj translation system, where each language corpus was translated into the language of each of the other corpuses, but we soon ran into several performance barriers. Although we had access to fairly robust computing resources, the training of the language models took significant time even with the performance gains offered by multi-headed Transformers with byte-pair encoding. Also, because there are few resources for translating between many low resource languages such as Danish and Frisian, the accuracy of the translations was quite low. In contrast, there are many more parallel corpuses and even pre-trained models for 20 Peter M. Broadwell and Timothy R. Tangherlini translating from many of the languages in our archives to English. By using English as the target translation and search language, we were able to side step several of these considerable performance barriers, while providing the facility for researchers to search in a single language and retrieve results from many languages.

Examples of Search Queries on ISEBEL The complexity of searching across multiple archives, each containing a series of languages is illustrated by a simply query for stories about ghosts. Prior to the implementation of the ISEBEL multilingual query model, one would need to search on at least the terms Geist and Gespenst (German), Geest and Spauk (Low German) geest and spook (Dutch), geast and spûk (Frisian), and spøgelse (Danish). Even with the low inflectional complexity for these particular nouns, the search needs to be increased to account for plural forms of the words, either through the use of wildcards or through inflection of the search terms. An individual search for these items would also require sorting the resulting results for unique matches. These challenges are solved by the single search on the term “ghost” in ISEBEL, returning a research set of 3278 records from the various archives, presented in a manner that makes broad geographic distribution immediately apparent, as well as providing optional selections filters for the collections (Fig. 1). Examination of the results allows for the discovery of both intriguing overlaps and critical distinctions in the beliefs surrounding ghosts in the various regions. Since the interface provides for numerous paths to access the root stories, a researcher would be able to explore a comparison of Danish

LÅGUPPLÖST BILD Arne har efterlyst

Figure 1. A result on the search “ghost” across all of the ISEBEL data pro- viders. Geist, Geest, Geast, Spøgelse 21 ghosts located close to maritime areas, and the haunting of dead mariners by selecting stories from the map that cluster along the coastline. That searcher would quickly find a Danish story about the theft of boots from a stranded sailor (Kristensen 1894: no. 386), and the story of haunting by a sea captain in a remote Dutch seaside village (Franke 1934: 205–207), which could in turn lead to an exploration of the uneasy relationship between seaside pop- ulations and the unpredictable North Atlantic. Similarly, an inspection of the keywords associated with the various collections would quickly show a high correlation between ghosts and driving, presumably wagons. Using “drive” as a limiting search on the subcorpus results in a collection of one hundred and eighty-seven stories (121 from the Danish, 58 from the Dutch and 8 from the German) for more focused study (Figure 2). Future refinements to the interface will allow users to download all of the records from a search such as this with their corresponding metadata formatted according to the ISEBEL XML schema. Although the current ISEBEL interface still requires that the end user have some facility with the various target languages, the ability to search concisely in a single language and return a useful research corpus is a significant advantage over previous search strategies. A researcher may have the hypothesis that one of the main problems with living on a farm in northern Europe during the nineteenth century was not only the ubiquity of witches, but their propensity to steal cream or butter. Searching for these types of encounters with witches on ISEBEL shows the power of multi-term queries. A search on “witch”, for example, returns 5,909 results (Danish – 2238, Dutch and Frisian –3004; German – 668), while a search for “butter” returns 861 results. Recognizing that “cream” is often mentioned as a precursor for churning butter, a search for “butter” or “cream” returns 901 results, slightly fewer than the simple union of the results for butter or cream individually. A search for witch AND (butter OR cream) returns a far more useful 249 results. These can then form the basis of a research corpus,

Figure 2. Stories from the “ghost” search limited to “drive” (lower right inset), with closeups of the story LÅGUPPLÖST BILD view for a single story from Arne har efterlyst the Evald Tang Kristensen collection, here a story about a minister who binds not only a thief, but also a ghost. 22 Peter M. Broadwell and Timothy R. Tangherlini which can be expanded based on the various discoveries that one might make while working with the preliminary corpus. In the Danish legends, witches frequently use toads to secure limitless access to butter, whereas toads play little or no role in the work of “butter witches” in the Netherlands or Germany (cf. Kristensen 1980: vol 7, no. 613). Yet, in all traditions, the witch’s familial cat is broadly attested. The increasing importance of butter in the local economies of late nineteenth century Northern Europe may provide an interesting backdrop against which to begin interpreting these results, while also providing insight into the different conceptualizations of witches across this broad region (Tangherlini 2000).

Conclusions and Future Directions The incorporation of multilingual search into integrated folklore collections shows great potential for supporting research into the interactions between tradition groups that are geographically proximate yet linguistically distinct. Similarly, it offers the opportunity to explore the distribution of similar motifs or topics across broad geographic regions, while still providing access to the underlying resources to support an understanding of how motifs, topics or even phrases differ across various tradition groups. Importantly, these approaches allow researchers to access materials in a more comprehensive manner, even when the researcher’s linguistic expertise does not cover all of the linguistic complexity and diversity contained in the target corpuses. There are many potentially productive future avenues for both the concatenated search represented by ISEBEL or other harvester-based methods, and for the multilingual search methods presented here. Of particular interest – and an example of fairly low hanging fruit – is the integration of all the Nordic language folklore archives into a NordISEBEL, particularly given the linguistically proximate nature of the Nordic languages. This feature of the languages would allow for higher quality cross-language searches, while also allowing for low cost / high accuracy ni x nj language translations so that any one of the Nordic languages could be used for search. Thus a search for stories related to the Danish nisser would not require searching the dirty English translations of the Danish, Swedish and Norwegian stories using the presumed English equivalent(s) of nisser in each language, but rather would allow for the use of the Danish search term and would rely on much more accurate translations of the Swedish and Norwegian texts directly into Danish. A search in Swedish therefore would rely on dirty Swedish translations of the Danish and Norwegian, and so on. While generating the numerous models needed for this type of multilingual search (as noted above) is prohibitively expensive for a large number of languages, the existence of rich, historical training data between the Nordic languages Geist, Geest, Geast, Spøgelse 23 and their morphonological and syntactic proximity, would make this type of highly accurate approach easier to achieve. The methods presented here are agnostic to the type of data fed to the translation algorithms. While we have tried to tune the language models to late nineteenth century language by adding various resources that exist in parallel corpuses, there is no reason these models cannot be expanded to cover other vocabularies appropriate to the time(s) of collection. It is a bit of a truism that the richer the training materials, the more likely the translations will present a good target for search. Expanding the vocabulary with fairy tale collections, such as those of the Grimm brothers, which have been translated into many of the target collection languages, would improve search accuracy for fairy tale collections (Joosen 2014). Similar expansions could be created for ballads, proverbs, or any other number of genres in folklore collections. It is not difficult to imagine a consortium of folklore archives that, by each running a simple OAI-PMH data provider node, could share those parts of their collection they are allowed to share, while the consortium could support multi-lingual search and other sophisticated tools for the visualization, analy sis and download of those results. The requirements for joining would be quite low: the ability to install and run a simple OAI-PMH data provider node, an understanding of the ISEBEL XML schema, and the creation of a domain specific term list, with corresponding terms in English, either drawn from or added to the existing table of terms. ISEBEL could then, in partnership with the new consortium member, train or adapt an English language NMT translation model, and fairly quickly create a corpus for indexing, which would then be accessible for search alongside the other ISEBEL corpuses. In addition to centralized consortium search interfaces, one could easily create regional portals, offering different types of concatenations of the consortium partners’ data along with tools particularly suited for a particular region. For instance, a Nordic consortium might only harvest from Nordic partners while providing rich multilingual search, while a European consortium might harvest more broadly, yet provide concatenated search primarily in English. A consortium focused on the fairy tale might only harvest those records designated as fairy tales, with a search vocabulary augmented with terms specific to the genre and access to a sophisticated network model of the ATU and motif indices, while a consortium of North Atlantic fiddlers might only harvest fiddle music from North Atlantic consortium members with search methods optimized for music. The incorporation of straightforward download mechanisms, so that researchers could save the results of their searches, and the development of analytical tools that sit on top of the concatenated collections can be fairly easily addressed in future iterations of the user platforms, since all of the harvested data, by design, conform to the same data schema. 24 Peter M. Broadwell and Timothy R. Tangherlini Numerous emerging projects speak to the possible realization of these ambitious goals. SAMLA, a national project in Norway, aims to digitize the various Norwegian folklore archives of the nineteenth and early twentieth centuries (Kverndokk & Ågotnes 2020). An initiative supported by the Kungl. Gustav Adolfs Akademien för Svensk Folkkultur aims to bring researchers from across the Nordic region together to plan for NordISEBEL, while additional interest from archives in Estonia, Lithuania, Finland, and Hungary suggest that the time is ripe to develop consortia of archives to enable the type of international, multilingual search that lay at the very root of the folklore discipline back to its beginnings in the early nineteenth century.

Timothy R. Tangherlini Professor of Danish Literature and Culture Department of Scandinavian University of California, Berkeley 6303 Dwinelle Hall Berkeley, CA 94720-2690 USA e-mail: [email protected]

Peter M. Broadwell Digital Scholarship Research Developer Center for Interdisciplinary Digital Research, Stanford University Libraries Stanford University Stanford, CA 94305 USA e-mail: [email protected]

References Abello, James, Peter Broadwell, and Timothy R. Tangherlini 2012: Computational folkloristics. Communications of the ACM 55(7). Aula, Anne, and Melanie Kellar 2009: Multilingual search strategies. In: CHI’09 Extended Abstracts on Human Factors in Computing Systems. Bachtold-Staubli Hanns, Eduard Hoffmann-Krayer, and Ludtke Gerhard 1927–42: Handworterbuch Des Deutschen Aberglaubens. Berlin. Bahdanau, Dzmitry, Kyunghyun Cho, and Yoshua Bengio 2014: Neural machine translation by jointly learning to align and translate. arXiv:1409.0473 Bendix, Regina 1997: In Search of Authenticity : The Formation of Folklore Studies. Madison. Börner, Katy 2011: Plug-and-play macroscopes. Communications of the ACM 54(3). Broadwell, Peter M., and Timothy R. Tangherlini 2015: ElfYelp: Geolocated topic models for pattern discovery in a large folklore corpus. In: Proceedings of Digital Humanities 2014. Geist, Geest, Geast, Spøgelse 25 Broadwell, Peter M., and Timothy R. Tangherlini 2017: Ghostscope: Conceptual mapping of supernatural phenomena in a large folklore corpus. In: Maths Meets Myths: Quantitative Approaches to Ancient Narratives. Basel. Broadwell, Peter M., and Timothy R. Tangherlini 2016: WitchHunter: Tools for the geo-semantic exploration of a Danish Folklore Corpus. Journal of American Folklore 129(511). Broadwell, Peter M., Peter Leonard, and Timothy R. Tangherlini 2018: “‘Hvad der byggedes om dagen, blev revet ned om natten …’: Word Sequence Repetition in Danish Legend Tradition.” Svenska landsmål och svenskt folkliv 140. Broadwell, Peter M., David Mimno, and Timothy R. Tangherlini 2017: The Tell- Tale Hat: Surfacing the Uncertainty in Folklore Classification.Journal of Cultural Analytics. doi: 10.7910/DVN/SYZ1PZ Bruder, Ilvio, Holger Meyer, Alf-Christian Schering, and Christoph Schmitt 2015: Das Projekt WossiDiA: Digitalisierung des Wossidlo-Archivs. Digitales Kulturerbe 61. Cho, Kyunghyun, Bart Van Merriënboer, Dzmitry Bahdanau, and Yoshua Bengio 2014: On the properties of neural machine translation: Encoder-decoder approaches. arXiv:1409.1259 Christiansen, Palle Ove 1996: A Manorial World : Lord, Peasants and Cultural Distinctions on a Danish Estate, 1750–1980. Oslo. Christiansen, Reidar Thoralf 1977: The Migratory Legends. International Folklore. New York. Davies, Owen, and Willem de Blécourt 2004: Witchcraft continued: Popular magic in modern Europe. Manchester. De Certeau, Michel 1984: The Practice of Everyday Life. Berkeley. Degh, Linda 2001: Legend and Belief : Dialectics of a Folklore Genre. Bloomington. Dundes, Alan 1964: The Morphology of North American Indian Folk Tales. FF Communications 195. Helsinki. Dundes, Alan 1971: Folk ideas as units of worldview. Journal of American Folklore 84(331). Estonian Folklore Archives 2017: Archives As Knowledge Hubs: Initiatives and Influence: Abstracts Tartu. Feilberg, Henning F 1886: Bidrag Til En Ordbog Over Jyske Almuesmål. Kjøbenhavn. Franke, Simon: 1934. Legenden Langs De Noordzee. Zutphen. Gunnell, Terry A. 2012: Waking the Dead: Folk Legends Concerning Magicians and Walking Corpses in Iceland. In: News from Other Worlds: Studies in Nordic Folklore, Mythology and Culture. In Honor of John F. Lindow. Berkeley. Gunnell, Terry A. 2010: Sagnagrunnur: A New Database of Icelandic Folk Legends in Print. Folklore: Electronic Journal of Folklore 45. Gunnell, Terry A. 2009: Legends and landscape in the Nordic countries. Cultural and Social History 6(3). Gunnell, Terry A., ed. 2005: Legends and landscape. Reykjavik. Hudson, Charles 1966: Folk history and ethnohistory. Ethnohistory. Hult, Marte H. 2003: Framing a National Narrative: The Legend Collections of Peter Christen Asbjørnsen. Detroit. Hutchins, W. John 1995: Machine translation: A brief history. In: Concise history of 26 Peter M. Broadwell and Timothy R. Tangherlini the language sciences. Oxford. Ilyefalvi, Emese 2018: The theoretical, methodological and technical issues of digital folklore databases and computational folkloristics. Acta Ethnographica Hungarica 63(1). Joosen, Vanessa 2014: Grimms’ Tales Around the Globe: The Dynamics of Their International Reception. Detroit. Klintberg, Bengt af 2010: The Types of the Swedish Folk Legend. FF Communications 300. Helsinki. Kristensen, Evald Tang 1894: Gamle folks fortællinger om det jyske almueliv. Vol. 6: Vore fædres tankesæt og andsliv. Kolding. Kristensen, Evald Tang 1980: Danske Sagn Som De Har Lydt in Folkemunde: Udelukkende Efter Utrykte Kilder. København. Krohn, Kaarle Leopold, and Julius Krohn 1926: Die Folkloristische Arbeitsmethode. Instituttet for Sammenlignende Kulturforskning, Serie B, Skrifter 5. Oslo. Kverndokk, Kyrre and Hans-Jacob Ågotnes, PI 2020: SAMLA: National Infrastructure for Cultural History and Tradition Archives. Bergen. Meder, Theo 2018: ISEBEL: Intelligent Search Engine for Belief Legends. Volkskunde 119(1). Meder, Theo, Folgert Karsdorp, Dong Nguyen, Mariët Theune, Dolf Trieschnigg, and Iwe Everhardus Christiaan Muiser 2016: Automatic enrichment and classification of folktales in the Dutch folktale database. Journal of American Folklore 129(511). Mitchell, Stephen A. 2011: Witchcraft and Magic in the Nordic Middle Ages. Middle Ages Series. Philadelphia. Mitchell, Stephen A. 1997: Blåkulla and its antecedents: transvection and conventi- cles in Nordic witchcraft. Berlin. Olrik, Axel 1910: Dansk Folkemindesamling (Dfs): The National Collection of Folklore in Copenhagen. FF Communications 1. Helsinki. Saatkamp, Marielies, and Schluter Dick 1995: Over Heksen, Toverij En Bijgeloof in De Nederlands-Duitse Grensstreek. Westmunsterland Bd. 4. Enschede. Schmitt, Christoph, and Timothy R. Tangherlini 2018: Folklore Archives Online: Zur Sichtbarmachung, Auswertbarkeit und Interoperabilität einer dänischen und einer nordostdeutschen Sammlung. In: Jahrbuch für Europäische Ethnologie Dritte Folge 13–2018. München. Sennrich, Rico, Barry Haddow, and Alexandra Birch 2015: Neural machine translation of rare words with subword units. arXiv:1508.07909 Sinclair, Stéfan and Geoffrey Rockwell 2016: Voyant Tools. Web. http://voyant- tools.org/. Singh, Jaspreet, Wolfgang Nejdl, and Avishek Anand 2016: History by diversity: Helping historians search news archives. In: Proceedings of the 2016 ACM on Conference on Human Information Interaction and Retrieval. Skott, Fredrik 2017: Sägenkartan. http://www.sprakochfolkminnen.se/om-oss/kar- tor/sagenkartan.html#/places Stokker, Kathleen 1995: Between Sin and Salvation: The Human Condition in Legends of the Black Book Minister. Scandinavian Studies 67(1). Storm, Ida, and Timothy R. Tangherlini 2018: ‘En temmelig lang fodtur’: hGIS, Text Mining, and Folklore Collection in 19th Century Denmark. Human IT 14(2). Geist, Geest, Geast, Spøgelse 27 Sturtevant, William C. 1966: “Anthropology, history, and ethnohistory.” Ethnohistory. Sutskever, Ilya, Oriol Vinyals, and Quoc V. Le 2014: Sequence to sequence learning with neural networks. In: Advances in neural information processing systems. Tangherlini, Timothy R. 2018: Toward a generative model of legend: Pizzas, bridges, vaccines, and witches. Humanities 7(1). Tangherlini, Timothy R. 2013a: Danish Folktales, Legends, and Other Stories. Seattle. Tangherlini, Timothy R. 2013b: The Folklore Macroscope: Challenges for a Computational Folkloristics. Western Folklore 72. Tangherlini, Timothy R. 2010: Legendary Performances. Ethnologia Europaea 40(2). Tangherlini, Timothy R. 2005: The Beggar, the Minister, the Farmer, his Wife and the Teacher: Legend and Legislative Reform in Nineteenth-Century Denmark. In: Legends and Landscape. Reykjavik. Tangherlini, Timothy R. 2000: ‘How do you know she’s a witch?’: Witches, Cunning Folk, and Competition in Denmark. Western Folklore 59(3/4). Tangherlini, Timothy R 1998: ‘Who ya gonna call?’: Ministers and the Mediation of Ghostly Threat in Danish Legend Tradition. Western Folklore 57(2/3). Tangherlini, Timothy R. 1994: Interpreting Legend: Danish Storytellers and Their Repertoires. Milman Parry Studies in Oral Tradition. New York. Tangherlini, Timothy R., and Peter M. Broadwell 2014: Sites of (re) Collection: Creating the Danish folklore nexus. Journal of Folklore Research 51(2). Tangherlini, Timothy R., and Peter Leonard 2013: Trawling in the Sea of the Great Unread: Sub-corpus topic modeling and Humanities research. Poetics 41(6). Taussig, Michael T. 2010: The Devil and Commodity Fetishism in South America. Chapel Hill. Thompson, Stith 1966: Motif Index of Folk-Literature. Bloomington. Tiedemann, Jörg 2016: OPUS--Parallel Corpora for Everyone. Baltic Journal of Modern Computing. Trieschnigg, Dolf, Djoerd Hiemstra, Mariët Theune, Franciska Jong, and Theo Meder 2012: An Exploration of Language Identification Techniques in the Dutch Folktale Database. In: Proceedings of the Workshop on Adaptation of Language Resources and Tools for Processing Cultural Heritage (LREC 2012). Trojahn, Cássia, Bo Fu, Ondřej Zamazal, and Dominique Ritze 2014: State-of- the-art in multilingual and cross-lingual ontology matching. In: Towards the Multilingual Semantic Web. Berlin. Uther Hans-Jorg 2004: The Types of International Folktales. FF Communications No. 284–286. Helsinki. Van de Sompel, Herbert, Michael L. Nelson, Carl Lagoze, and Simeon Warner 2004: Resource harvesting within the OAI-PMH framework. D-lib magazine 10(12). Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł. and Polosukhin, I. 2017: Attention is all you need. In: Advances in neural information processing systems. 28 Peter M. Broadwell and Timothy R. Tangherlini

1 The ISEBEL project was funded for three years, 2017–2020, as part of the Transatlantic Platform – Digging into Data Challenge. Funding was provided by the National Endowment for the Humanities (USA), the DFG (Germany), and the NWO (Netherlands). The four co-PIs on the project were Timothy R. Tangherlini (USA), Theo Meder (Netherlands), Holger Meyer and Christoph Schmitt (Germany). Scientific personnel included QiQing Ding (Netherlands), Petra Himstedt-Vaid and Alf-Christian Schering (Germany), and Peter M. Broadwell (USA). The ISEBEL homepage provides information about the project (cf. http://search.isebel.eu/ about). 2 The ISEBEL schema can be found at: http://purl.org/ISEBEL/schema 3 For the ISEBEL user interface, we relied on CKAN, an open data portal, for a variety of reasons, most related to ease of customization, sophisticated search operators, and low cost (cf. https://ckan.org/). 4 Search on ISEBEL can be carried out at http://search.isebel.eu. An explanation of search operators can be found at http://purl.org/ISEBEL/searchop Digital Humanities and How to Read the Kalevala as a Thematic Anthology of Oral Poetry

Venla Sykäri

Abstract The quickly developing digital tools together with the digitization of large corpuses of archived material provide new tools for the research of early folklore editors’ methods and practices. In Finland, such need regards in particular the Kalevala (1849), which continues to have a great cultural significance for Finland, and for world literature in general. The Kalevala was created by Elias Lönnrot on the basis of the literary and ideological models of his time and the many long fieldwork trips that provided him with deep knowledge of the oral singing culture. Lönnrot chose to present the resulting source materials by incorporating them into a historically well-known literature genre, epic. With his methods of manipu- lating the narratives, creating characters, and inserting incantations and lyrical and ritual poetry within the weave, he nevertheless stretched the concept of epic. Generations of researchers have worked to shed light on the Kalevala as a literary work. However, the Kalevala’s relation to its oral source poetry has never been coherently illustrated, and few people therefore know what Kalevala in fact is. This paper presents a research project titled as “Open Kalevala” and its use of means provided by digital humanities to decode the elements that construct the Kalevala. Keywords: Kalevala, epic, Elias Lönnrot, oral poetry, digital humanities

Elias Lönnrot’s Kalevala was first published in 1835, and in a new, expanded form in 1849. The latter has gained the status of “Finnish national epic” (here, I will use the title “Kalevala” to refer only to the 1849 edition, but when needed, epithets old and new will be used). This national epic is the ultimate result of Elias Lönnrot’s long-term process of working on oral Finnish language runosong poetry (the old trochaic, alliterative, non-stanzaic eight-syllable form): carrying out long fieldwork trips and arranging, editing, and publishing the material noted down by him and others from oral singers. Lönnrot’s field notes, travel diaries, correspondence, manuscripts and publications have been preserved extensively. His sources are thus in principle well-known, as is the fact that Kalevala is a composite work. 30 Venla Sykäri However, just how the Kalevala is compiled is closely known only to a very limited group of specialists. Most readers of the Kalevala anticipate that the epic represents “old narratives sung in verse by Finnish and Karelian peasants”. While epic narratives are without doubt at the core of the work, verses, motifs and even longer passages from other types of oral poetry – lyrical poetry, charms, wedding songs, proverbs – intervene, and most verses and verse units have undergone changes in Lönnrot’s hands. Lönnrot never denied his interventions nor indicated that the Kalevala was based on narratives only. Yet even when earlier research has tracked the source referents in minute detail (esp. Niemi 1898; Krohn 1903–1910; Kaukonen 1939; 1944; 1956), it has not been comprehensively shown how the various pieces are woven together and what kind of units and songtypes they represent. On the other hand, the fact that most readers find the non-narrative interludes dull and jump over large parts of the book raises the question: Could the expectations be different; How could the reader be better equipped to read and understand the Kalevala? The aim of this article is to discuss how digital humanities can introduce the Kalevala’s structure and sources, and help the reader become familiar with the genres and themes of archived Finnish oral poetry. The research I will present derives from a multidisciplinary project titled Avoin Kalevala (“The Open Kalevala”), whose scope has been to take a grip of all the aspects of Lönnrot’s creative process in a digital critical edition of the Kalevala (http://kalevala.finlit.fi/). My own research in this project was carried out on the Kalevala’s relation to its source poetry, with a particular focus on two cycles of cantos on a central hero, Lemminkäinen. Based on the results of this analysis, and introducing the structure of the canto 11 as an example, the paper examines the central elements in the construction of the Kalevala. After pointing out the need for a comprehensive inventory of these elements, I will argue that in addition to the evident benefits in illustration and display, the digital transformation of the text can change the way we read it: the digital critical edition also allows the Kalevala to be read as a thematically orga nized anthology of oral poetry.

Background: the Kalevala as Elias Lönnrot’s Multi-Decade Project Elias Lönnrot (1802–1884) was from a humble background, but he was helped to get schooling and entered the Academy of Turku for education in classical humanistic subjects. Together with certain fellow students, he absorbed the prevailing Fennoman ideas and became interested in folk poetry. In 1827, the city of Turku burned down to the ground, and being unable to continue his studies, Lönnrot realized his wish to gain Digital Humanities and How to Read the Kalevala 31 more knowledge on folk poetry when he began his first large field tour in Finnish Karelia in April 1828. This trip gave him a first-hand experience of the oral tradition that was still lively in eastern Finland and over the border, in Russian Karelia. Later Lönnrot conducted studies in medicine, and his nomination as a district medical doctor in Kajaani, a town close to both Finnish and Russian Karelia, helped him to reach the villages beyond the border. By 1844, he had completed eleven large tours in a vast district, where eastern Finnish Savo-Karelia, Russian Archangel Karelia, as well as the Karelian Isthmus were central for his folkloristic pursuits. Lönnrot was not the first one to record oral poetry, nor the last. His interest in folk poetry was aroused by his predecessors’ publications, chiefly those of Z. Topelius and C. Gottlund, and his field activities were shared by some colleagues, in particular after the publication of the first version of the Kalevala (1835). In his editorial work, Lönnrot preferred versions he had noted down by himself. However the impact of the young student D. E. D. Europaeus and other collectors was significant for the enormous amount of versions and variants and new verses, which were in Lönnrot’s hands when he compiled the new Kalevala (1849). The Kalevala’s composite structure and the core narrative frame began to take form in a manuscript titled as Runokokous Väinämöisestä (The Collected Songs About Väinämöinen; later published with the name Alku- Kalevala, or the Proto-Kalevala; Lönnrot 1928), in November 1833. In this manuscript Lönnrot intertwined three initially separate manuscripts into one narrative continuum. One of these contained wedding songs and the other two different stories sung on Lemminkäinen and Väinämöinen, respectively. These 5052 lines, divided into 16 cantos, were the beginning of the Kalevala. With further addition of new epic songs, multiple parallel versions and complementary motifs, and in particular long pieces of incantations, that manuscript grew into 12,078 verses and 32 cantos of the first Kalevala edition published in 1835 (Kalevala Taikka Vanhoja Karjalan Runoja Suomen Kansan Muinoisista Ajoista; The Kalevala, or Old Karelian Poems about the Ancient Times of the Finnish People; published in two parts, the second coming out 1836). Even if Lönnrot published this first version of the Kalevala, a lot more material was waiting, and new field trips tempted him: as he had written to his colleague Cajander in December 1833, he “wouldn’t stop collecting poems before having a collection equivalent to half of the Homer” (Niemi 1902: 196, translation by V.S.). By the time Lönnrot finally put together the new Kalevala, expanded to 22,795 verses and 50 cantos, the material had become enormous. Several factors impacted this growth: 1) Lönnrot’s new field trips gave him a) new parallels to those songtypes that he already knew, b) new compilations, where known songtypes, themes 32 Venla Sykäri and motifs were mixed in new ways, as well as c) entirely new songtypes, motifs and a lot of new verses. 2) Between these two publications, Lönnrot travelled for the first time in South-Karelia, where he found a singing culture that was different from that he was familiar with in northern Finnish Karelia and Russian Archangel Karelia. In southern areas, female singers and lyrical themes, as well as lyrical epic songs, were much more current than long epic poems, which were often sung by men in northern areas. This led to the preparation of an edition of lyrical and ballad-based narrative poetry, the Kanteletar (Lönnrot 1840; 1841). Much of this material also found its way between the epic passages in the expanded new Kalevala. 3) Other field workers found some new narrative types as well as new poems that represented already known types in a more elaborated form. As all field collectors sent their material to Lönnrot, he had access toa vast quantity of parallel and complementary versions, which provided him with a still deeper knowledge of both those songtypes and verses that were widely shared and common, as well as of the boundless possibilities for variation and the creation of new verses in composition. With all this material, Lönnrot could have created an anthology of Finnish-Karelian runosong poetry, organized according to genres and typical songtypes, themes and motifs. But he had already chosen another option, the creation of an epic: an overarching narrative, which could absorb the material organized as parts of an invented narrative structure. This decision was exceptional in extension even in the light of the prevailing literature aesthetics and international models of his time, but it also reflected the flex- ibility of the singing culture and methods of oral composition that he by then knew so well.

Archived runosong material and research The Finnish Literature Society (Suomalaisen Kirjallisuuden Seura, SKS) was established to sustain Lönnrot’s collecting initiatives in 1831. One significant impact of the publishing of the Kalevala was a general ardour to collect ever more pieces of runosong poetry, and successively, other folklore genres, especially fairy tales. The finding of the same or similar songtypes in large areas led to intensive research of the distribution of this poetry. Uncovering the origin of each songtype became the scientific core of the world’s first folklore studies discipline, at the University of Helsinki (1898). Comparative research grew from and also resulted in further collecting activities, and materials became plenty. The present SKS archive house was built in 1890 as a response to the urgent need to preserve and organize the huge and fast-growing amount of materials in one place. Thus, all the runosong material noted down from oral singers (until 1939) is Digital Humanities and How to Read the Kalevala 33 archived in SKS. Most of this material was published in 33 large tomes between 1908–1948, with an additional part in 1997, under the title Suomen Kansan Vanhat Runot (SKVR), “The Ancient Poems of Finnish People”. These tomes, which contain approximately 89,000 texts and almost 27,000 pages (Harvilahti 2013: 224), are organized according to the area, genre and titles representing the songtypes, and each source text is given a reference number. The expression “songtype” (in SKVR runotyyppi) can name units of quite different size and cohesion; according to genres, it can either mean a relatively stable plot construction (epic), a target matter (incantations), or a motif (ritual and lyrical poetry). By “motif” I here translate the Finnish word aihelma and refer to a central idea that is expressed with a relatively stable unit of lines: on average these units contain four to six lines, but they can also be couplets or triplets or longer constructions. The unit index thus refers to narrative songtypes or incantation types according to their subject matter or motifs in lyrical and ritual poetry. With respect to the narrative songs and incantation texts, the smaller parts, episodes and motifs, are not yet analysed. All this material was digitized at the turn of the millennium (see Harvilahti 2013) and is today freely available in a digital data basis (https:// skvr.fi/) maintained by the SKS. The digital SKVR offers information and search possibilities related to the collectors, years when recorded, as well as districts, villages, and singers, in case these are marked. An important but not yet finalized and thus not fully published digital index of songtypes has been under work for long. It currently works through songtype/motif titles: it is possible to search the classification hierarchy by a title and see, for example, the distribution and number of variants analysed as representing a certain songtype/motif. One can also see which parts one particular textualized rendition is constructed of. However it is not yet possible to read a description of the songtype/motif. A new, more flexible user interface Octavo for searching the digital SKVR data basis is being developed at the University of Helsinki by Eetu Mäkelä (see Kallio & Mäkelä 2019), and has proved to be very effective for this research group. The Kalevala’s compilation and relation to its sources raised discussion among scholars soon after its publication (Hämäläinen 2020). The epic songs beyond the Kalevala were first researched by Julius Krohn in 1885, and successively, by Kaarle Krohn (9013–9010) in close detail. The first critical line-by-line analysis of the composition of the early versions of the Kalevala was carried out by folklorist A. R. Niemi in 1898. His work addressed the first manuscript from 1833 that is today referred to as the Sikermä-Kalevala, “Cluster-Kalevala” (Lönnrot 2007), the second manuscript from the end of 1833, which is referred to as Alku-Kalevala, or the “Proto-Kalevala” (Lönnrot 1928), as well as the first printed Kalevala from 34 Venla Sykäri 1835. Niemi relates the Kalevala’s lines to those in collectors’ manuscripts, which from a modern reader demands a very throughout knowledge of the source items. Half a century later, after the publication of the SKVR tomes, literary critic Väinö Kaukonen undertook a new line-by-line analysis of both the old (Kaukonen 1939; 1945) and the new Kalevala (Kaukonen 1956) with references to the SKVR numbers. Both the work of Niemi and that of Kaukonen are extremely useful for specialists, but they do not introduce nor discuss the source material itself: their target was to verify the closest source for each line of a written, literary work. Yet Kaukonen also illustrated Lönnrot’s work and working methods in his several other research publications (see esp. Kaukonen 1979; 1984; in English, see Kaukonen 1990). These early scholars did not of course yet have in their use the digital SKVR data basis, which has proved to be an indispensable tool for both research and illustration of the sources in the Open Kalevala edition. Kaukonen’s comprehensive literary critical perspective established an understanding of the Kalevala as literature. During the later decades of the twentieth century, many Finnish folklorists instead turned their interest towards the source itself, the oral poetry. The large and deep collections could be revived as a research field when the emergent performance-centred and ethnopoetic research paradigm provided new ways to contextualize how that poetry had lived in its contexts of use (e.g. Harvilahti 1992; Timonen 2004a; Tarkka 2005; 2013; Kallio 2013). Yet many folklorists have also addressed the Kalevala in their work. Professor Lauri Honko, a researcher of epic, was also a central figure in the folkloristic research of the Kalevala, its birth and reception (in English, see esp. Honko 1990; 2002). Analysis of the Kalevala’s reception and nationalistic uses has been, perhaps, a self-evident target for folklorists (e.g. Wilson 1977; Anttonen 2004; 2014). Professor Satu Apo has addressed Kalevala’s images and world view in many of her writings, for example, on gender representations and national stereotypes (e.g. Apo 1995; 2002; 2004; 2009; 2019). The annual publication of the Kalevala Yearbook by the Kalevala Society (https://kale- valaseura.fi/en/the-yearbook/), as well as several other collected volumes (e.g. Siikala et al. 2002; Siikala et al. 2004; Piela et al. 2008), continue to produce new insights. Beginning from the 1990s, Kalevala’s textualization processes were taken under closer folkloristic analysis also by Jouni Hyvönen and Niina Hämäläinen in their doctoral theses. In his work on the textualization of the charm episodes of the old Kalevala, Hyvönen (2001; 2005) shows that Lönnrot had been inspired by the eighteenth-century French Encyclopedists, in addition to his better known ethnographic, national romantic, and literary influences. Hämäläinen (2012; see also 2014) addressed Lönnrot’s family representations and his deployment of lyrical poetry as expression of Digital Humanities and How to Read the Kalevala 35 emotions, displaying how Lönnrot strongly participated in the discussion of the new family ideals through the scenes and characters he created. While these and several other research initiatives and interpretations have continued to shed light on the Kalevala, this knowledge has barely been opened to the public at large. Moreover, the gap between the Kalevala as a literary work and knowledge of the oral poetry that was used as its sources has never been coherently transmitted nor illustrated. Very few people thus concretely know what the Kalevala in fact is: what it means that the Kalevala is constructed of epic, lyric and ritual poetry, as well as charms, or what these elements beyond the Kalevala are and what they represented in the oral system of genres. This is a significant gap given the continuing cultural significance of the Kalevala for Finland, and indeed for world literature in general.

Avoin Kalevala – a Digital Critical Edition of the Kalevala The digital Avoin Kalevala edition (http://kalevala.finlit.fi/) is based on the goal that any reader of the Kalevala should be able to become familiar with the work’s different, parallel parameters through one source. These parameters include the relation between the written text and its oral sources, the historical time-frame with its intellectual, literary and ideological currents, as well as Lönnrot’s handprint on the plot, characters, and the overall mise en scène. The reader should also be able to read the text effortlessly, and to grasp the old and dialectal words, which Lönnrot deployed extensively and in morphologically manipulated forms, as one of his simultaneous interests was to develop the Finnish language and vocabulary. The initial idea for a digital critical edition of the Kalevala derives from folklorist Niina Hämäläinen’s research work on the textualization of the Kalevala. The implementation of this idea was first tested by Hämäläinen in a pilot digital edition on the personality of Aino, the maid that in the Kalevala drowned herself after being promised by her brother as wife to Väinämöinen, an old man and from her perspective non-desirable choice for husband (http://aino.finlit.fi/omeka/2017). On this basis, Hämäläinen planned a research project for a multi-disciplinary group. A three-year funding received from the Alfred Kordelin Foundation in 2017 enabled the enactment of the plan. Funding was complemented with the participation of the Finnish Literature Society (SKS), whose research department hosted the project in 2018–2020 and which has taken the responsibility for building the website and maintaining it in the future. The 2018–2020 research group was led by Hämäläinen, currently the executive director of the Kalevala Society, and consisted, in addition to her, of another folklorist (Venla Sykäri), a linguist (Reeta Holapainen until September 2019 and after 36 Venla Sykäri that Marika Luhtala), as well as a cultural historian (Juhana Saarelainen). The results were brought to the Omeka Open Access data basis by SKS’s specialist Maria Niku, who shared her time with this and other digital projects of the SKS. Based on this research group’s work, two peer-reviewed parts of the edition have been published in 2019 and 2020 (Hämäläinen et al. 2019; 2020). These cover a sample of the Kalevala’s text. They contain word explications and annotated critical commentaries to selected groups of poems, as well as critical commentaries to Lönnrot’s introductions to both the old and the new Kalevala. Commentaries analysing the oral source material are completed on the Aino cycle (poems 3–5 in the Kalevala) and the main part of the Lemminkäinen cycle (11–15, 26–28). Hämäläinen has also analysed Väinö Kaukonen’s research of the Kalevala’s lines with respect to canto 4 (Aino) and canto 22, which contains wedding poems (see also Hämäläinen 2020). These commentaries contain links to the source poetry texts in the SKVR data basis (see figure 1 below). Commentaries on the intellectual and literary influences and cultural history are realized on certain poems (1–2, 42–43, 50), which reflect Lönnrot’s ideological purposes in the framing of the work (1–2 and 50) or have been a central target of interpretation (42–43, The Sampo cycle). These latter commentaries, as well as those on Lönnrot’s introductions, contain direct links to Lönnrot’s correspondence, which is the subject of another digitizing project being undertaken in the SKS (http:// lonnrot.finlit.fi/omeka/). In addition to these basic functions, the edition contains an introduction, compact texts introducing more in-depth the various viewpoints offered to the Kalevala in this edition, instructions for use, bibliography, a possibility to search over all text, as well as a facsimile of Lönnrot’s handwritten manuscript. Hämäläinen has also realized a sample that displays changes from the old Kalevala to the new, in comparison. A third part of the edition is waited to be published in 2021. It is wished by the research group that one day the whole Kalevala could be analysed and commented (and further, at least the comments and background texts translated into English): funding applications for new research, further commentaries, larger articles, as well as illustrations are being processed. After proceeding from the main page, the website page is divided into three columns: to the left, the reader can find the original printed page of the Kalevala. The middle column hosts a transcription of the text, with the possibility to see basic word explanations by going over the word with the mouse. More information on selected words, expressions and linguistic forms appear by opening a text box. By ticking a “commentaries” box above the column, the text receives asterisk marks* (or numbers in some cases) that lead to commentaries, which open to the column at right. Digital Humanities and How to Read the Kalevala 37

1. The three columns of the Avoin Kalevala edition. The commentaries include direct links to the indicated source poetry (SKVR referents in grey).

Composition of the Kalevala’s Canto 11 My work carried out in the Open Kalevala edition from September 2018 until April 2020 chiefly comprised of the analysis of the Kalevala’s poems 11–15 and 26–28, whose primary protagonist is a young male hero referred to as Lemminkäinen. In the oral poetry saved to us, two basic epic songtypes on Lemminkäinen were popular: Lemminkäisen virsi, “Lemminkäinen’s song” which tells the story of the protagonist’s dangerous travel to a big feast where he was not invited, and Lemminkäisen surma, “The Death of Lemminkäinen”, in which the travel episodes end with his death and his mother’s attempts to gather the pieces of his body and revive the hero. Except for some excerpts and composite variants, the latter was found only in few villages of Archangel Karelia, while the former has been extremely popular, one of the runosong types that were performed well until the twentieth century (Siikala 2012: 268, 286). In the oral poetry of the main areas of the early nineteenth-century field- workers, Finnish North-Karelia and Russian Archangel Karelia, the character of Lemminkäinen had already become mixed with another young male hero, Kaukamoinen or Kaukomieli, and his story of a quarrel in a drinking party, which ended by him killing a man and escaping to an “Island”, where he slept with all its hundred maidens and wives. In Finnish North-Karelia, Lemminkäinen is rather exclusively the hero of this narrative songtype that combines elements of two or three different plot structures, while in Archangel Karelia, both names appear but certain parts are normally sung 38 Venla Sykäri only in the name of one of them. Nevertheless, Lemminkäinen was a hybrid person in the oral poetry, and being a ‘tradition dominant’, occasionally was mixed with or replaced several other names. It was therefore easy for Lönnrot and others to interpret him as one character. Along with Väinämöinen, Lemminkäinen was the first central epic hero whose poems Lönnrot begun to arrange in a linear continuum in 1833. Episodes from poetry on Lemminkäinen played decisive role in the developing core narrative frame of the Alku-Kalevala (in detail, Sykäri 2020). In the construction of the old Kalevala, the plots’ mythic elements invited Lönnrot to insert within the story line large numbers of hunting and healing charms, as well as charms on magic anticipation against witches and envi- ous spirits. For the new Kalevala, Lönnrot wrote a completely new beginning to precede the established plot elements related to Lemminkäinen: the canto 11 and the beginning of the canto 12. I will give a summary of the main contents of the canto 11 and indicate the most significant songtypes and motifs that Lönnrot employed when constructing this piece. My particular focus is on the genres used. As the text in the Avoin Kalevala data basis is in Finnish, I will here use as my reference the English translation by Keith Bosley (Lönnrot & Bosley 1989). However, any full translation that main- tains the original number of lines can be used to follow this rough analysis. I have translated the central epic songtype titles in English; the Finnish SKVR titles are given in parentheses, when they appear for the first time.

Lines 1–20: Introduction of the male hero To begin with, Lönnrot unites the names of three epic figures into one: Ahti, Kauko (Farmind) and “wanton loverboy”, which refers to Lemminkäinen (his current epithet is lieto, “wanton”). The first lines derive from the introduction to The Song of Ahti and Kyllikki (Ahdin ja Kyllikin runo). This is a rare songtype noted down only a couple of times in Archangel and Olonets Karelia. Other lines describing the protagonist are compiled of couplets and triplets coming each from different epic poem or fragment. At the end, Lönnrot gives the decisive description of the hero as having the bad habit of going with women; these verses are semantically in line with the events told at the end of the story of Kaukomieli (songtype: Maids of the Island; Saaren neidot), but such characterization is never made in the source poetry; Lönnrot’s lines, instead, are modified from an individual text, in which a mother groans about his son’s habit of going out by night.

Lines 21–60: Introduction of the female protagonist Kyllikki’s name comes from The Song of Ahti and Kyllikki, but there is no description of her in this songtype. To describe her Lönnrot utilizes Digital Humanities and How to Read the Kalevala 39 certain poems written down in North-Karelia, which present a mixture of epic songtypes. The first lines derive from the introduction to the Song of Marketta (Marketan runo), which represents ballad-based songs of a girl who is said to have stayed home long, turning down all suitors. The same central idea is in the next passage, from a mythic epic song The wooing of the Sky Lights (Taivaanvalojen kosinta). In this songtype a maid turns down the proposals of the Sun, the Moon and the Stars to marry their son, with the pretext that their work in the sky is heavy and lonely. As in the renditions deployed by Lönnrot, in the southern areas (Ingria, South-Karelia), this song can end in a denial to marry Estonia, a motif that is commonly met in another ballad-based narrative song, the Island’s maid (known as Saaren neito, in the SKVR Nurmen neito). In addition, the motif of “lack/plenty of all things” with or without comparison is in southern areas a common motif in lyrical Songs of Homeland.

61–78: A dialogue between Lemminkäinen and his mother Lemminkäinen informs his mother of his decision to set off to woo the Island’s maid, and his mother forbids him from going. The model for the dialogue is directly drawn from Lemminkäinen’s song, and all its verse components are found in oral poetry, yet the realization of this dialogue is entirely from Lönnrot’s pen. Mother’s first argument is that this Island’s maid’s kin is of higher societal status, and he would not be accepted among her kin. Lemminkäinen answers with a unit deriving from an individual lyric Men’s song motif written down by Lönnrot in South-Karelia: the young man compares himself with the suitor from a rich house and argues that he himself will “take” (the bride) with his more attractive physical appearance. This motif continues to build the image of Lemminkäinen as a sexually active man that attracts women. Lemminkäinen’s mother’s next argument is that the women will laugh at him. In addition to its sexual overtones, the women’s laughter is a common symbol for utter social shame. The motif of women’s laugh appears in epic poetry, within a threefold series of questions set by the mother to her home-coming, crestfallen son: was he laughed at by women? Normally this episode follows the return of Joukahainen after his song contest with Väinämöinen (Singing Contest; Kilpalaulanta), but it is once applied to Kaukomieli by singer Arhippa Perttunen of Latvajärvi. The couplet “How do I pay the women’s laughter / the women’s laughter, the wenches’ taunts?” appears as a motif in the epic song Kaleva’s son’s revenge (Kalevanpojan kosto). On the other hand, the statement “I know how to pay the women’s laughter” is a popular motif in male lyric songs. In these, often after saying that he would like to sing but doesn’t in fear of women’s laughter, the lyric I argues that he’ll pay women’s laughter by making them pregnant – by returning the shame with the shame of having a bastard child. 40 Venla Sykäri Lönnrot inserts the above lyric verse unit as Lemminkäinen’s answer. His mother’s response is then constructed by taking words from the epic song Maids of the Island, where the protagonist (normally Kaukomieli, who has fled from home to the Island after killing a man) sleeps with all the women of the Island, even the “purest” (which hints at nuns) and is thereafter chased away by the Island’s men. On the other hand, at the onset of Lemminkäinen’s song, his having slept with even the purest maids is in some areas expressed as the reason for not inviting him to the feast.

103–156 Lemminkäinen enters the island and makes merry with the woman By repeating and varying previously presented lines and adding new motifs deriving from the epic Lemminkäinen’s Song (departure) and Kaleva’s son’s revenge (getting hired as a herdsman) Lönnrot advances the narrative: Lemminkäinen has now entered the Island. Just as Kaukomieli (Farmind) in the songtype Maids of the Island, Lemminkäinen undergoes a dialogue with the maids and sleeps with all of them. That songtype never has the motif of women laughing at him, the connection is purely Lönnrot’s idea – but the motif has a connection with other, new motifs taken from the Kaleva’s son’s revenge.

157–186 Lemminkäinen hunts Kyllikki, who mocks her suitor The description of Kyllikki as “one lass who would not accept bridegrooms, take in good husbands” is modified from a heeling charm against rickets. The next four lines describing how Lemminkäinen “wore away a hundred boots, rowed a hundred oars in half” while hunting Kyllikki is directly from the main inspiration for Lönnrot’s construction of this canto 11: the epic Song of Ahti and Kyllikki. Kyllikki’s words belittling the suitor are put together from three different motifs representing lyric Young people’s songs, which express wishes related to courtship and marrying, for example descriptions of suitable and non-suitable partners. These lyric motifs – which in a performance could be freely combined with any other related entities – are clearly discernible even in the translation. The first motif of a groom who has sand in his pockets instead of money is (also) conventionally employed as a ritual mocking song in a wedding.

189–222 Lemminkäinen abducts Kyllikki This epic passage derives from one individual oral poem written down by Lönnrot, “The Maid’s robber” (Neiden rosvo). This is a personal rendition inspired by two ballad-based epic poems: The Incest (Rutsa), where a young man tempts a young woman to join him in his sledge, and later finds her to be his sister; and the Iivana Kojonen’s son (Iivana Kojosenpoika), where the protagonist, after wooing a maid, snatches her into his sledge and Digital Humanities and How to Read the Kalevala 41 takes the complaining maid home – where he coldly chops her into pieces and then sends these pieces prepared as food to the maid’s mother as a gift. The source text used by Lönnrot neither has the incest nor the murder, but presents instead a threat put by the abductor to the group of merrymaking maids not to tell to the snatched maid’s mother that he had been there. The content of the lines telling that Lemminkäinen would ‘sing’ (incant) their bridegrooms to war is modified from an unrelated songtype Borderline War (Rajasota).

223–288 Kyllikki moans and Lemminkäinen consoles her Kyllikki’s moaning seems to be inspired by the above mentioned songtype of Iivana Kojonen’s son, but the lines and motifs come from many different poems, for example a motif in wedding poetry, where the groom is warned that the bride’s male relatives will revenge any maltreatment. Lemminkäinen consoles Kyllikki first with a rare emotional motif from Men’s lyrical songs: “you in my arms as I eat / you in my hands as I walk / you at my side as I stand / you beside me as I lie”. After that Lönnrot puts in Lemminkäinen’s mouth a longer ballad-based mocking song (used to ritually mock the groom during the wedding ceremony) that appears in the form of an appraisal. The description of the groom’s many cows makes clear that they in fact do not exist; thereby there are no worries and no tiring work awaiting the maid. Lönnrot then adds to Lemminkäinen’s utterance a few lines that pick up the motif of his kin being of lesser status, as well as how he solves this problem with his sword skills. These lines also foresee his future adventures by hinting at his magic powers: the sword is described as being “one refined among demons / polished among gods”. Lönnrot ends this passage with the introductory formula typical of the epic song of Iivana Kojonen’s son in certain areas: the maid utters that she was born and grew up being idle, wishing that by marrying Kojonen she could continue to be idle. Lönnrot, by putting the word “luckless” within the first line and by continuing with the motif of accusing the groom as an idler, manipulates the verses here with this minor change towards his narrative goals.

289–314 Swearing the oaths The introductory phrase uttered by Kyllikki “if you want a maid like me / for an everlasting mate / for a hen under your arm” is typical of the Archangel Karelia singer Arhippa Perttunen and appears in many of his epic songs, where a maid is asked/promised/consented to a wife. These narratives are always connected to Väinämöinen, never to Lemminkäinen, as here. The central motif of this passage is the swearing of oaths, which derives from the Song of Ahti and Kyllikki: for Ahti not to go to war, and Kyllikki not to go to the village. The verb ‘visiting’ is paralleled by Lönnrot with ‘need of dance’; the source texts clearly hint at adultery. 42 Venla Sykäri 315–362 The way and arrival home The description of the couple’s travel to Lemminkäinen’s home, including a dialogue between Lemminkäinen and Kyllikki, is inspired by the song Iivana Kojonen’s son, but lines are constructed from various bits and pieces. In Lemminkäinen’s answer to his mother Lönnrot repeats once again the motif of paying the wives’ laughter. At the end, the lines of Lemminkäinen’s utterance turn towards motifs from wedding poetry, and the specific theme of the appraisal of the bride. His mother’s utterance is constructed of clearly identifiable wedding song motifs: first expressing the perspective of a mother-in-law, appraising the bride as a new, much waited household worker, and thereafter by a poetic appraisal motif “Pure is the bunting on snow / purer is the one you have”, ext. Lönnrot ends canto 11 by Lemminkäinen’s mother’s instructions to construct a new, better home. The central lines describing how he should construct ‘better thresholds’ and ‘new doors’ come from a lyric motif Reward of Mother’s Pains (Äidin vaivojen palkinto), where usually a male singer speaks of the mother’s pains in giving him birth and growing him up, and asks how he could reward these. The closing two lines addressed to Lemminkäinen are created by Lönnrot. They repeat the social point expressed by the mother earlier: the maid is “better than yourself, greater than your kin”.

Analysis of the elements of the poem 11 With the above rough account of the Kalevala’s canto 11 I wish to draw attention to its major ingredients; this account reflects the original Finnish analysis and references given to the source poetry (in Finnish) only in very broad lines. Yet I hope that it can illustrate those three central elements in the work of Lönnrot to which I will turn now: 1) the narrative frame, 2) compilation of the contents, and 3) his ideological fingerprints.

The narrative frame The songtype that frames the compilation of this canto – as well as the beginning of the next canto (12) – is The Song of Ahti and Kyllikki. To create the overall story line of this canto (11), Lönnrot used only nineteen lines from the beginning of two known renditions of this songtype (mixed from SKVR I 906; SKVR II 232). Based on Keith Bosleys’s English translation of the Kalevala they are as follows:

’Tis time to tell of Ahti and to lilt about a rogue. (He) wore away a hundred boots Digital Humanities and How to Read the Kalevala 43 rowed a hundred oars in half going for that maid hunting Kyllikki. Thereupon they swore their oaths made their pledges for ever before the God known to all beneath the Almighty’s face: Ahti would not go to war nor Kyllikki visiting

These ten lines are spread over the poem 11 in lines 1–2, 165–168, and 309–314. As already mentioned, the canto 11 (and the beginning of 12) was prepared by Lönnrot as a new expansion to the story of Lemminkäinen in the New Kalevala. It was inspired by the finding made by D. E. D. Europaeus in Olonets Karelia: he noted new exemplars of this songtype, which had earlier been written down only once by Lönnrot in Archangel Karelia. Lönnrot’s early text alone is pretty cryptic and did not yet give reason to identify it as an independent songtype, but the Olonets texts (and some later variants) shed light to its different story line, which Matti Kuusi (1963: 236–242) has analysed as belonging to the Viking Age stratum of songs, given its representation of a man and a woman as equal partners. A hero named Ahti chases a woman referred to as Kyllikki, and they become a couple by swearing oaths to each other in front of an icon: Ahti will not go to war and Kyllikki will not go ‘visiting’. In the continuation, the next lines tell that Kyllikki goes to ‘strange gates’. Ahti’s sister yells this to Ahti, who now has the reason to prepare for ‘war’ (a raid, most probably). A dialogue between the spouses is undergone about leaving vs. staying (in the Kalevala, these events make the first part of the canto 12). To create a canto of 402 lines with the frame given by 10 lines of a source narrative is rather extreme even in the Kalevala’s case, even if the ballad-based narratives Iivana Kojonen’s son and The Incest, for which Lönnrot had not found any other place in the Kalevala, provided their episodes for help. In general, to create the narrative structure of the larger chunks of his epic, Lönnrot followed the story lines of selected source narratives more or less closely. In some cases he uses one common, “full” rendition or a mixture of two or three versions as his master version. With respect to Lemminkäinen, this is the case with the Kalevala’s poems 26 and 27 that convey the narrative of Lemminkäinen’s travel and acts in his destination, the feast of Päivölä (in the Kalevala: Pohjola). However, just as in this case, Lönnrot often mixed several songtypes. He also seems to have been very attracted by individual versions, which are already mixed compilations of well known songtypes, names and motifs. With respect to Lemminkäinen, Lönnrot employed several songtypes that 44 Venla Sykäri are normally sung in the name of another hero, but were occasionally or even just once performed in the name of Lemminkäinen. An example of this is a large passage in the Kalevala’s canto 26, where Lemminkäinen travels and asks to be received in three different houses – this episode “belongs” with no other exception to the songtype Väinämöinen’s Knee Wound (Väinämöisen polvenhaava), and relates the story of Väinämöinen searching aid for his bleeding knee wound in three different houses. After the nearly 20 years’ field work experience and constant, - longi tudinal working on the material, Lönnrot knew well what was typical or common variation within the per se fluid boundaries of the songtypes, and what was clearly an exception. It is clear that Lönnrot’s purpose was not to render the songtypes in their most common form. He was inspired by mix- tures, exceptions and the unique, which may have stood as an analogy and justification for his own work. This is one of the reasons why digital tools make such a big difference in the analysis of the Kalevala: the mass analysis of texts under digital humanities techniques is well designed to bring out exactly the emergence of unique elements.

Compilation of the contents of the canto 11 In accordance with the chief narrative frame, the central content components of the canto 11 derive from themes relevant to young people of marriageable age: sexuality, appearance, wooing, partner choice, and marriage. Plot items come from several narrative poem types: from the male point of view, these are related to wooing, proposal of marriage, and sexual adventures, as well as bride abduction; on behalf of the girl, they are related to her rejection to marry (too early, not the preferred suitor). None of the epic texts or songtypes is employed with its entire story line; they rather provide episodes, verse material and models. The dialogue structures imitate those of the protagonist with his mother in Lemminkäinen’s song and Kaukomieli’s song, and that with the abducted Kyllikki in Iivana Kojonen’s son. Other suitable passages, and sometimes only one motif, are fitted in the narrative frame as they appear in oral versions – even in the translation, it is easy to distinct the units that were crystallized in oral expression (see e.g. the Wooing of the Sky lights, or the ritual appraisal formulas that make the mother’s utterance at the end). Moreover, Lönnrot deploys lyrical poems, similarly related to wooing and desired/non-desired grooms, in particular as his protagonists’ utterances to each other. These lyric motifs chiefly derive from the category of Young people’s songs. Wedding poems are also put to use, first in the form of ritual mockery (which could also be performed as narrative or lyric songs). To complete the bond made, typical motifs praising the bride are presented by the mother. Incantation motifs are very few in this canto, whose focus is on social life, and their primary role is to encourage the image of Lemminkäinen as Digital Humanities and How to Read the Kalevala 45 a self-confident, shameless bragger and thus to ground his adventures in the future. This image of Lemminkäinen is in line with the hybrid image of Lemminkäinen in the source poetry, even if never transmitted by such explicit description. The four-line unit of the lass who would not accept bridegrooms, modified by Lönnrot from a healing charm against rickets and employed to describe Kyllikki, is a good example of Lönnrot’s very throughout knowledge of all the genres of the runosong poetry, as well as of his wish to employ a maximal amount of the available verse units in the Kalevala – albeit if modified and entirely out of their normal co-text and context.

Lönnrot’s literary and ideological fingerprints As a way to structure and progress his narrative, Lönnrot uses common means both of oral epic and those of written literature. Dialogues are created with utterances preceded by descriptions of the speaker and the act of speaking: “Wanton Lemminkäinen said / fair Farmind uttered”; “His mother put this in words”; “He uttered a word, spoke thus”. These lines are common in oral epic, although Lönnrot extends the length and amount of dialogues. In addition to this, oral epic narratives employ short couplets or motifs of 3–4 lines for various introductory purposes. The lines might introduce the protagonist: “Who grew up in a lofty home / came up in one most graceful / sitting in her father’s rooms / where the back bench sagged”. The lines might indicate someone’s mental state: “(He) twisted his mouth, turned his head / and twisted his black whiskers”. The lines might indicate the passing of time: “A little time passed / barely half a month went by”. And the lines might indicate that someone is going or moving: the har- nessing of a horse, or whipping a horse, are typical motifs indicating this. Lönnrot uses all these devices provided by his source material. Nevertheless, Lönnrot often extends these units’ lengths and adds explanatory lines and long entries that describe the persons and explain what happens (for Lönnrot’s editorial techniques, see also e.g. Kaukonen 1979, Apo 2002, Hyvönen 2001; 2005, Hämäläinen 2012; 2014; DuBois 2000). He uses a lot of repetition in a way that is alien to oral poetry: while oral poetry eagerly repeats scenes (three times), it does not use repetitive structures of the kind Lönnrot uses. For example, in the canto 11 he repeats a description of the shame of being laughed at: first as the mother’s warning, then in Lemminkäinen’s response to his mother, then as an act taking place, followed by Lemminkäinen commenting on this act, and still further as Lemminkäinen’s words repeating the scene for providing an explanation for his long absence to his mother. Certain oral renditions of Lemminkäinen’s song do take advantage of repetition in first telling how the mother warns Lemminkäinen of the dangers of his planned trip, and in the succession, how Lemminkäinen meets these dangers (even by explicitly 46 Venla Sykäri saying that this is like my mother told me). This is a method of expanding the story by dwelling on important narrative episodes, and not a way to construct inter-personal relations or help a reader, who is not familiar with the traditional story world, to keep track of what’s going on. Nevertheless, this model of repetition provided by the oral song is again easy to see as an inspiration for Lönnrot, when he was constructing this canto. Yet the extent of repetition and explanatory lines in the Kalevala greatly surpasses the oral model. At the level of the text flow, repetition of verbal expressions and ideas is one of the clearest marks that Lönnrot was addressing in Kalevala a very different audience from that of the village people. His audience was the Swedish-speaking upper classes that were accustomed to literary texts and conventions, but barely understood Finnish and knew little of peasant life and living conditions, not to speak of the runosong tradition’s elliptic poetic language and expression. While such literary methods introduced to the texture of the Kalevala are easily discernible, Lönnrot’s ideological fingerprints can be less evident for someone not knowing well the oral source texts. One side of this is the structure framing the overall story world: the events are framed to have taken place in a pagan past, which has then ended with the coming of the more civilized world at the age of Christianity. The beginning and the end of the Kalevala are constructed according to this leading principle. It is often remarked that Lönnrot took off the Christian elements from the passages he used for the above-mentioned reason: the Kalevala’s stories represented ancient, pagan times. With regard to his cantos on Lemminkäinen, and with respect to the extensive use of charms, which nearly exclusively contain a mixture of pagan and Christian names and layers, he also used several references to God as well as Christian names (Mary and Jesus, chiefly). In these cases he organized the sequences to show that when the traditional methods or helpers are not powerful enough to help, the Christian actors give release to the problems. In the canto 11, the two Christian elements are however different. The oaths are woven “before the God known to all / beneath the Almighty’s face”, with the specific reference of the original lines to oaths woven ‘before the coppery icon’, and thus to Orthodoxy, left out. At the end, Lönnrot has in fact added the words praising God to the wedding appraisal motif. These were not connected to the specific verses he employs when the mother praises the bride, but such Christian words of praise were commonly used in similar traditional contexts. A more complicated level of ideologically oriented variation of plot and verse elements is found in the presentation of persons and their relations. As Niina Hämäläinen has researched in detail in her dissertation (2012), Lönnrot strongly participated in the epoch’s discussion regarding the new Digital Humanities and How to Read the Kalevala 47 family values and, in particular, the woman’s role as an educator of children. Depicting family relations, and in the case of the Finnish runosong poetry, where fathers are seldom mentioned, typically a mother’s relation to her daughter or son was thus central for Lönnrot when he created his characters and constructed the narratives. In the canto 11, due to the enhanced repetition and number of comments uttered, we find a powerful mother that controls and comments every step of her son. The same controlling aspect follows throughout Lönnrot’s story of Lemminkäinen. In canto 15, this is coupled with the image of the mother devoted to finding and healing her son with every possible means (see also Timonen 2004b). The bases for this is the rare mythic song The Death of Lemminkäinen, which depicts how the mother seeks with her magic powers to find her murdered son and to collect the pieces from the river of Tuoni and put life in him again. Lönnrot has enhanced the image of the self-sacrificing mother by introducing long stretches of healing charms, as well as from the Christian-inspired narrative history sung in the Archangel Karelia (in SKVR: Luojan virsi, The Creator’s song). Every possible opportunity is deployed by Lönnrot to verbalize that male and female roles and spaces are different: women’s space is home and men’s the wild nature and adventures. Through the mother’s utterances in this canto, Lönnrot also introduces forcefully the aspects of social status and shame. As mentioned above, there is one source rendition regarding Kaukomieli, in which the mother asks among other things her son, who returns home after killing a man in the feast, whether he was laughed at by women. The epic motif thus exists, but was brought by Lönnrot to a new significance and extent by persistent repetition and addition of other socially oriented images, like the question of social status here.

Conclusions The Kalevala is Lönnrot’s creative project whose active, processual construction phase lasted very long, over two decades. His fingerprints mark all levels of the epic from the dialectal details of the language to the out- lines, contents and inhabitants of the story world. But even if this story world is widely shaped by one man’s experiences, imagination and extraordinary immersion in his subject, it also reflects the ethnographic ideas, literary models, and ideological ideals of the nineteenth-century European and Finnish intellectual environment. All these perspectives are taken into account in the digital Open Kalevala edition, which illustrates them through annotated critical commentaries side-by-side with the Kalevala’s text. In addition to the analyses done by each individual researcher on a specific subject, the research project has enabled to draw from a long continuum of earlier research. 48 Venla Sykäri While critical editions typically analyse one author’s production, the subject of the Open Kalevala is layered. As shown above with a rough analysis of the Kalevala’s canto 11, it is important to understand Lönnrot’s role when interpreting the Kalevala or inferring arguments about its narratives or characters. However, tracking Lönnrot’s fingerprints and literary ambition should not remain the only possibility to read this major work, as it is also based on Lönnrot’s thorough knowledge of this oral poetry and its fluid life. The oral composers of runosong poetry created all their own versions of the shared songtypes. They employed existing motifs, shaped lines and created new images and arguments for their needs. Lönnrot was thus like the original composers in that he “mixed” verses and created new connections and lines, even if he did it in a different context and to a very different degree. Epic poetry is well known to be a complex, in Bakhtinian terms “secondary” genre (Bakhtin 1986), constructed of variable plot modules and motifs, and employing third person narrative, introductory verses and dialogue. By using methods of narrative framing, by creating central characters to whose doings, sayings and dialogues large portions of lyrical material and charms could be attached, all original poetic material was organized in a form of a progressing narrative, an epic. By using the elements of digital humanities the Open Kalevala exhibits these structures. The analysis of the Kalevala’s oral sources illustrates another significant part of Lönnrot’s work: the thematic display of folk song materials, types and genres. Canto 11 is a good example of this: it brings together a selection of songtypes and motifs related to courting, wooing and sexuality, and young people’s reflections of these in various songs. Many epic songtypes used as source poetry in this canto derive from the ballad-inspired type of runosong poetry, which itself reflects a change in the singing culture towards social themes. One of Lönnrot’s targets in the compilation work was to illustrate the forefathers’ world view and living conditions. Even if the Kalevala does not present the original songs in their ethnographic context, it does provide a way to become acquainted with oral poetry in thematically organized sequences, across genres. A good portion of the Kalevala’s cantos is constructed with thematic orientation: a healing episode in the narrative is completed by introducing healing charms, a protagonist’s hunting activity leads to hunting charms, for example. I would therefore like to describe the epic as also being a thematic anthology of folk poetry. In this idiosyncratic case, it is 1) a thematic anthology compiled of pieces of song texts (songtypes, motifs, verse couplets) representing all major genres of oral poetry (epic, lyric, charms, ritual songs), 2) fitted in a fictional narrative frame, 3) the events and turns of which are orchestrated by Lönnrot’s pen and voice. Generations of researchers have worked to understand and explain this layered construction. Digital humanities offer entirely different facilities Digital Humanities and How to Read the Kalevala 49 both to find new sources and connections, as well as to display and visualize the results of analysis. In addition to the affordances of the digital platform itself for display of results and data, and for search, other data basis can be linked to it. Current links include digitized archives of Lönnrot’s manuscripts and correspondence, and foremost the digitized runosong text corpus, the SKVR. The latter is the foundation for illustrating the analysis of the Kalevala’s relation to its source poetry in the Avoin Kalevala project. Readers of the Avoin Kalevala edition can familiarize themself with the oral singers’ presentations which ended up in the Kalevala’s text. Digital access to SKVR thus offers via the digital critical edition the possibility to lead the reader to Lönnrot’s sources – however, not only to the textually closest variant but far beyond: the commentaries can provide insights into the place and meaning of the individual texts within the singing culture and its genres. This way the digital edition – today within the limits of the so far analysed and commented parts of the work – can also become a vehicle for capturing at least glimpses of the field of runosong poetry in a way Lönnrot once did.

Venla Sykari Researcher (PhD, docent) Finnish Literature Society, SKS Hallituskatu1, PL 259 FI-00171 Helsinki Finland email: [email protected]

References Anttonen, Pertti 2005: Tradition Through Modernity. Postmodernism and the Nation-State in Folklore Scholarship. Studia Fennica Folkloristica 15. Finnish Literature Society, Helsinki. Anttonen, Pertti 2014: The Kalevala and the Authenticity Debate. In János M. Bak, Patrick J. Geary and Gábor Klaniczay (eds.), Manufacturing a Past for the Pre sent. Forgery and Authenticity in Medievalist Texts and Objects in Nineteenth- Century Europe, 56–80. Brill, Boston & Leiden. Apo, Satu 1995: Äidin poikia ja epäonnisia kosijoita. Sukupuolten ja perhesuhteiden kuvaus Kalevalassa. Teoksessa Satu Apo, Naisen väki. Tutkimuksia suomalaisten kansanomaisesta kulttuurista ja ajattelusta, 89−129. Hanki ja Jää, Helsinki. Apo, Satu 2002: Kertojan ääni Kalevalassa. Teoksessa Pekka Laaksonen & Ulla Piela (toim.), Lönnrotin hengessä, 108−122. Kalevalaseuran vuosikirja 81. Suomalais- en Kirjallisuuden Seura, Helsinki. Apo, Satu 2004: Laulaen vai kirjallisesti luoden? Uuden Kalevalan valmistumispro- sessi Elias Lönnrotin kuvaamana. Teoksessa Anna-Leena Siikala, Lauri Harvilahti ja Senni Timonen (toim.), Kalevala ja laulettu runo, 273−298. Suomalaisen Kir- jallisuuden Seuran Toimituksia 958. Suomalaisen Kirjallisuuden Seura, Helsinki. 50 Venla Sykäri Apo, Satu 2009: Uusin Lönnrot-myytti ja kansanrunouden editiohistoria. Elore 2/2009. https://doi.org/10.30666/elore.78808 Apo, Satu 2019: Vanhasta Kalevalasta uuteen. Muutoksia, jatkuvuuksia ja aikala- isten kannanottoja. Teoksessa Ulla Piela, Pekka Hakamies ja Pekka Hako (toim.), Eurooppa, Suomi, Kalevala. Mikä mahdollisti Kalevalan?, 107–130. Kale- valaseuran vuosikirja 98. Suomalaisen Kirjallisuuden Seura, Helsinki. Bakhtin, M. M. 1986: Speech Genres and Other Late Essays. University of Texas Press. DuBois, Thomas A. 1995: Finnish Folk Poetry and the Kalevala. Garland Publish- ing, Inc., New York and London. DuBois, Thomas A. 2000: The Narrator’s Voice in Kalevala and Kalevipoeg. Oral Tradition 15(1): 126−144. https://journal.oraltradition.org/wp-content/uploads/ files/articles/15i/7_dubois.pdf Frog 2010: Baldr and Lemminkäinen. Approaching the Evolution of Mythological Narrative through the Activating Power of Expression. A Case Study in Germanic and Finno-Karelian Cultural Contact and Exchange. Dissertation. UCL Eprints. London: University College London. Hämäläinen, Niina 2012: Yhteinen perhe, jaetut tunteet. Lyyrisen kansanrunon tek- stualisoinnin ja artikuloinnin tapoja Kalevalassa. Väitöskirja. Turun yliopisto, Turku. Hämäläinen, Niina 2014: Why is Aino not described as a black maiden? Reflections on the textual presentations by Elias Lönnrot in the Kalevala and the Kanteletar. Journal of Finnish Studies. Vol. 18, no. 1: 91−129. Hämäläinen 2020 (forthcoming). Säe säkeeltä. Väinö Kaukosen säetutkimukset Kalevalasta. Sananjalka. Hämäläinen, Niina & Holopainen, Reeta & Luhtala, Marika & Saarelainen, Juhana & Sykäri, Venla 2019: Avoin Kalevala. Kansalliseepoksen digitaalinen, kriittinen editio. Osa I. Suomalaisen Kirjallisuuden Seura, Helsinki. ISSN 2490-1318. http://kalevala.finlit.fi Hämäläinen, Niina & Luhtala, Marika & Saarelainen, Juhana & Sykäri, Venla 2020: Avoin Kalevala. Kansalliseepoksen digitaalinen, kriittinen editio. Osa II. Suoma- laisen Kirjallisuuden Seura, Helsinki. ISSN 2490-1318. http://kalevala.finlit.fi Harvilahti, Lauri 1992: Kertovan runon keinot: Inkeriläisen runoepiikan tuottamis- esta. Helsinki: Suomalaisen Kirjallisuuden Seura. Harvilahti, Lauri 2013: The SKVR Database of Ancient Poems of the Finnish People in Kalevala Meter and the Semantic Kalevala. Oral Tradition, 28/2 (2013):223–232. https://journal.oraltradition.org/wp-content/uploads/files/articles/28ii/07_28.2.pdf Honko, Lauri (ed.) 1990: Religion, Myth, and Folklore in the World’s Epics. Mouton de Gruyter, Berlin, New York. Honko, Lauri (ed.) 2002: The Kalevala and the World’s Traditional Epics. Studia Fennica Folkloristica 12. Finnish Literature Society, Helsinki. Hyvönen, Jouni 2001: Vanhan Kalevalan loitsujaksojen tekstualisointi. Lönnrotin toimitusstrategiat tutkimustradition jatkumolla. Elore 1/2001. https://doi.org/ 10.30666/elore.78304 Hyvönen, Jouni 2005: Vanhan Kalevalan Vipusessa käyntirunon tekstualisointi: Digital Humanities and How to Read the Kalevala 51 Elias Lönnrotin tekstualisointistrategiat. Folkloristiikan lisensiaattitutkielma. Helsingin yliopisto, Helsinki. Järvinen, Irma-Riitta 2010: Kalevala guide. Finnish Literature Society, Helsinki. Kallio, Kati 2013: Laulamisen tapoja. Esitysareena, rekisteri ja paikallinen laji länsi-inkeriläisessä kalevalamittaisessa runossa. Dissertation at the University of Helsinki, folklore studies. http://urn.fi/URN:ISBN:978-952-10-9566-5. Kallio, Kati & Frog with Sarv, Mari 2017: What to Call the Poetic Form – Kalevala- Meter or Kalevalaic Verse, regivärss, Runosong, the Finnic Tetrameter, Finnic Alliterative Verse or Something Else? RMN newsletter No 12–13, 139–161. https://www.helsinki.fi/sites/default/files/atoms/files/rmn_12-13_2016-2017.pd- f?fbclid=IwAR2T2OBrbiyg8qUTDv1aQZmpR1TeaZjZL5c4eUtF9io1BjtBV3b- jGMn_mpw Kallio, Kati & Mäkelä, Eetu 2019: Suullisen runon sähköisestä lukemisesta. Elore 2/2019: 26–41. https://doi.org/10.30666/elore.84570 Kaukonen, Väinö 1939: Vanhan Kalevalan kokoonpano I. Suomalaisen Kirjal- lisuuden Seuran Toimituksia 213. Suomalaisen Kirjallisuuden Seura, Helsinki. Kaukonen, Väinö 1945: Vanhan Kalevalan kokoonpano II. Suomalaisen Kirjal- lisuuden Seuran Toimituksia 213. Suomalaisen Kirjallisuuden Seura, Helsinki. Kaukonen, Väinö 1956: Elias Lönnrotin Kalevalan toinen painos. Suomalaisen Kir- jallisuuden Seuran Toimituksia 247. Suomalaisen Kirjallisuuden Seura, Helsinki. Kaukonen, Väinö 1979: Lönnrot ja Kalevala. Suomalaisen Kirjallisuuden Seuran Toimituksia 349. Suomalaisen Kirjallisuuden Seura, Helsinki. Kaukonen, Väinö 1984: Elias Lönnrotin Kanteletar. Suomalaisen Kirjallisuuden Seuran Toimituksia 386. Suomalaisen Kirjallisuuden Seura, Helsinki. Kaukonen, Väinö 1990: The Kalevala as Epic. In Lauri Honko (ed.), Religion, Myth, and Folklore in the World’s Epics, 157–179. Mouton de Gruyter, Berlin, New York. Krohn, Julius 1885: Suomalaisen kirjallisuuden historia. Ensimmäinen osa: Kale- vala. Weilin & Göös, Helsinki. Krohn, Kaarle 1903−1910: Kalevalan runojen historia. Suomalaisen Kirjallisuuden Seura, Helsinki. Kuusi, Matti 1963: Suomen kirjallisuus I: Kirjoittamaton kirjallisuus. Suomalaisen Kirjallisuuden Seura, Helsinki & Otava. Lönnrot, Elias 1829: Kantele taikka Suomen Kansan sekä Vanhoja että Nykysempiä Runoja ja Lauluja. Toinen osa. Suomalaisen Kirjallisuuden Seura, Helsinki. Lönnrot, Elias 1833: Runokokous Väinämöisestä. Käsikirjoitus. Lönnrot, Elias 1835: Kalevala Taikka Vanhoja Karjalan Runoja Suomen Kansan Muinoisista Ajoista. Suomalaisen Kirjallisuuden Seura, Helsinki. Lönnrot, Elias 1849: Kalevala. Suomalaisen Kirjallisuuden Seura, Helsinki. Lönnrot, Elias 1840, 1841/2000: Kanteletar elikkä Suomen kansan vanhoja lauluja ja virsiä. Suomalaisen Kirjallisuuden Seuran Toimituksia 3. Suomalaisen Kirjal- lisuuden Seura, Helsinki. Lönnrot, Elias 1862/2005: Kalevala 1862. Lyhennetty laitos. Tärkeimmillä selityk- sillä koulujen tarpeeksi varustanut Elias Lönnrot. Suomalaisen Kirjallisuuden klassikoita. Suomalaisen Kirjallisuuden Seura, Helsinki. Lönnrot, Elias 1928: Alku-Kalevala. Suomalaisen Kirjallisuuden Seuran Toimituk- sia 178. Suomalaisen Kirjallisuuden Seura, Helsinki. 52 Venla Sykäri Lönnrot, Elias 2007: Sikermä-Kalevala. Lemminkäinen. Väinämöinen. Naimakansan virsiä. Juminkeko, Kuhmo. Lönnrot, Elias & Bosley, Keith 1989: The Kalevala. Translated in English by Keith Bosley. Oxford University Press, Oxford. Niemi, A. R. 1898: Kalevalan kokoonpano I. Runokokous Väinämöisestä. Suoma laisen Kirjallisuuden Seura, Helsinki. Niemi, A. R. (toim.) 1902: Elias Lönnrotin matkat 1−2. Suomalaisen Kirjallisuuden Seuran Toimituksia 98. Suomalaisen Kirjallisuuden Seura, Helsinki. Piela, Ulla & Knuuttila, Seppo & Laaksonen, Pekka (toim.) 2008: Kalevalan kulttuurihistoria. Suomalaisen Kirjallisuuden Seuran Toimituksia 1179. Suoma laisen Kirjallisuuden Seura, Helsinki. Saarinen, Jukka 2018: Runolaulun poetiikka. Säe, syntaksi ja parallelismi Arhippa Perttusen runoissa. Väitöskirja. Helsingin yliopisto, Helsinki. Siikala, Anna-Leena & Harvilahti, Lauri & Timonen, Senni (toim.) 2002: Kalevala ja laulettu runo. Suomalaisen Kirjallisuuden Seuran Toimituksia 958. Suomalaisen Kirjallisuuden Seura, Helsinki. Siikala, Anna-Leena & Harvilahti, Lauri & Timonen, Senni (toim.) 2004: Kalevala ja laulettu runo. Suomalaisen Kirjallisuuden Seuran Toimituksia 958. Suomalaisen Kirjallisuuden Seura, Helsinki. Siikala, Anna-Leena 2012: Itämerensuomalaisten mytologia. Suomalaisen Kirjallisuuden Seura Toimituksia 1388, Tiede. Suomalaisen Kirjallisuuden Seura, Helsinki. Sykäri, Venla 2020: Toimitustasot/Kalevala ja kansanrunousaineisto/ Lemminkäinen Kalevalan eri versioissa. Teoksessa Niina Hämäläinen & Marika Luhtala & Juhana Saarelainen & Venla Sykäri 2020: Avoin Kalevala. Kansalliseepoksen digitaalinen, kriittinen editio. Osa II. Suomalaisen Kirjallisuuden Seura, Helsinki. ISSN 2490-1318. http://kalevala.finlit.fi Tarkka, Lotte 2005: Rajarahvaan laulu: Tutkimus Vuokkiniemen kalevalamittaisesta runokulttuurista 1821−1921. Suomalaisen Kirjallisuuden Seuran Toimituksia 1033. Suomalaisen Kirjallisuuden Seura, Helsinki. Tarkka, Lotte 2013: Songs of the Border People. Genre, Reflexivity, and Performance in Karelian Oral Poetry. Folklore Fellows’ Communications Vol. cxlxii, No. 305. Helsinki: Suomalainen Tiedeakatemia. Timonen, Senni 2004: Minä, tila, tunne: Näkökulmia kalevalamittaiseen kansan lyriikkaan. Suomalaisen Kirjallisuuden Seuran Toimituksia 963. Suomalaisen Kirjallisuuden Seura, Helsinki. Wilson, William A. 1976: Folklore and Nationalism in Modern Finland. Indiana University Press, Bloomington and London.

1 In this essay, I use the term runosong (Finnish: runolaulu) for the old Finnic singing tradition, instead of terms like Kalevala meter poetry, Kalevalaic singing, etc. The latter, anachronistic terms are well-established within research of the textualized artefacts, but avoided today by contemporary singers and researchers of the singing culture (in detail, see Kallio & Frog with Digital Humanities and How to Read the Kalevala 53

Sarv 2017). In an analysis of the Kalevala’s relation to its source poetry, the latter represents a conceptual contradiction. (For the runosong culture, see also the Wiki-inventory for living heritage: https://wiki.aineetonkulttuuriperinto.fi/wiki/Runosong). 2 This article is based on a conference presentation given at the SIEF2019 14th Congress at Santiago de Compostela, Spain (14–17 April 2019). I wish to thank all the panellists and the audience for vivid conversation, and in particular Susanne Österlund-Pötzsch who, in the succession, suggested that I submit an article on the theme for this special issue of ARV. While the research presented here was carried out for the Avoin Kalevala edition, funded by the Alfred Kordelin Foundation and the Finnish Literature Society (2018–2020), the manuscript was completed during the summer of 2020 with funding provided by Kone Foundation (201906994). My warmest thanks are due to Professor Nigel Fabb and Niina Hämäläinen, executive director of the Kalevala Society, who read drafts of the manuscript and helped me to improve the text. I am also grateful to Professor Arne Bugge Amundsen for his insightful comments and all his editorial assistance. 3 For more information in English on Lönnrot’s predecessors and Lönnrot, the Kalevala process and Lönnrot’s publications, as well as the Kalevala’s main structure, see Honko 1990; Kuusi 1990; Kaukonen 1990; Honko (ed.) 2002; also e.g. DuBois 1995; Wilson 1976; Järv- inen 2010. Detailed bibliography (mainly in Finnish) can be found at http://kalevala.finlit.fi/ kirjallisuus. 4 As the references made by Niemi are to the original manuscripts, I must admit that I could benefit from many details in his analysis only after having made my own. In general, much of the early research unfolded when I had done the same analysis work by myself. This illustrates well how closely the nineteenth-century and early twentieth-century researchers were acquainted with the manuscripts, compared to the present day generation. 5 Equivalents for the narrative and mythic elements present in the songs on Lemminkäinen have been found in many narrative cultures, in particular the Scandinavian and other North- ern mythologies as well as the Bible. For connections between Lemminkäinen and Baldr, see especially Frog 2010. 6 Later when the runosong poetry of Ingria had been discovered, including a common song of Kaukomieli and the drinking party, and Finnish researchers had become familiar with the Es- tonian tradition, Kaarle Krohn (1903–10) could analyse the various motifs of these songtypes in detail. 7 Matti Kuusi (1963: 392) remarks that the poor suitor who boasts about his non-existent riches is one of the favourite characters of the late medieval Swedish mocking songs (skämtvisor). 8 I have expressed these three dimensions in an oral presentation of the Avoin Kalevala in Finnish as: “Runolaulutekstien osista koostettu temaattinen antologia (runotyypit, aihelmat, säeparit), sovitettuna kuvitteelliseen kertomuskehykseen, jonka käänteitä ohjailee tekijän ääni ja kynä”.

Legends, Letters and Linking Lessons Learned from Amassing and Mapping Folklore and Viewing as Part of Nineteenth-Century Culture Creation

Trausti Dagsson and Olga Holownia

Abstract Focusing primarily on a database of Icelandic folklore material (Sagnagrunnur), we discuss how this database has been developed, expanded and reconfigured over the past 20 years. We also place Sagnagrunnur within a network of Icelandic projects that have aimed to collect, digitize and make available a significant collection of manuscripts and books related to the nineteenth-century Icelandic “culture creators”, and we further contextualize it by showing connections with international projects and databases. We also discuss the open-ended nature of digital projects and outline future steps. Keywords: digital folkloristics, Icelandic folktales, database, romantic nationalism, culture creation

Introduction The Icelandic noun grunnur means foundation and base. Sagnagrunnur (literally [the] base / foundation of stories/tales), a project that was initiated over two decades ago, has become the heart of a key digital project in Iceland relating to the collection of folklore. The name is also a play on gag- nagrunnur (database) and indeed, it has become a database that is has now reached beyond its original purpose and ambition in that it is finding new ways of expressing itself, simultaneously reaching into new territories, both at home and abroad, drawing to itself new contexts which provide not only new uses but also new means of understanding its ever-growing content. This article will describe the evolution of Sagnagrunnur, show its increasing connections with other Icelandic databases and digital archives and place it in a wider context of digital folkloristics as a whole. In the light, considerations will be made of possible extensions of the existing network of digital projects in Iceland and their potential integration with other existing databases of Nordic and mainland European folklore. 56 Trausti Dagsson and Olga Holownia Context The work on Sagnagrunnur (sagnagrunnur.com) started in 1999 and was led by professor Terry Gunnell at the University of Iceland. The main objective was to create a searchable database of all legends in published Icelandic folklore collections. The project built on earlier archival work by Gísli Sigurðsson and Rósa Þorsteinsdóttir of the Árni Magnússon Institute for Icelandic Studies in Reykjavík. Initiated in 1994, the aim was to make a record of the Institute’s extensive collection of audio recordings of oral traditions before it was published online as part of a wider database dealing with Iceland’s musical heritage, now available at ismus.is). The archive includes over 10,000 recordings of oral folk tales. The creation of an index of keywords that could be used for both the sound archive and a database of folk legends in print (as dreamt of by Bo Almqvist of UCD in Dublin) was a crucial step in the process. Arguably, it would provide the backbone of the new database (Gunnell, 2005, 2015). While work on Sagnagrunnur had been proceeding (with its major rede- velopment in 2013–2014), other developments had been taking place in Iceland. Two closely related Icelandic projects: Inventing Culture: Defining Sources of Theory and Inspiration and the Long-term Results of Culture Creation by Icelandic Intellectuals 1857–1874 (henceforth Creating Culture, sigurdurmalari.hi.is) and Íslenzkar þjóðsögur og æfintýri: Origin, Context and Collection 1864–2014 (jonarnason.is) were coming into being, closely connected to the wider on-line Encyclopedia of Romantic Nationalism in Europe (ERNiE, now available at romanticnationalism. net) led by Joep Leerssen of the University of Amsterdam and coordinated by SPIN: Study Platform on Interlocking Nationalism. The former project (2011–2014 and led by Karl Aspelund and Terry Gunnell) provided an extensive context for the creation of Icelandic cultural identity in the latter part of the nineteenth century, a project which had been had had the collection of Icelandic folktales as one if its key ingredients, along with the design of a national costume, the establishment of a national museum and the creation of “national drama” (which had folklore at its heart). The latter project (2015–2018), led by Rósa Þórsteinsdóttir, aimed to place more focus on one particular aspect of this movement, tracing the international context and the collection and publication process of Íslenzkar þjóðsögur og ævintýri (Icelandic Legends), edited by Jón Árnason (1918–1888) also referred to as the father of Icelandic folklore collection or even the Grimm of Iceland. In due course, this latter project will become part of yet another wider project entitled Grimm Ripples, which looks at the origin and cultural context of the collection of folk legends in the other neighbouring countries, considering why the legend should have led the way rather than the fairy tale. Legends, Letters and Linking 57 Spanning over 7 years, the two latter Icelandic projects, both related in one way or another to the collection of Icelandic folklore in the nineteenth century, have involved the digitization, transcription and annotation a significant number of original documents archived and stored at different Icelandic and international institutions. The both cases, the aim was to put all these materials under one digital roof, make them openly accessible and searchable, and link to existing digital resources both in Iceland and abroad. An important aspect of all these projects, like that of Sagnagrunnur, is that rather planning for them to exist and work in isolation, there has from the start been a strong emphasis on the idea of interconnection with other projects at home and abroad, building on previous knowledge (including technical expertise), and working closely with the key Icelandic cultural heritage institutions.

Sagnagrunnur: Indexing and Mapping Icelandic Legends From its inception, Sagnagrunnur was about (much needed) searchability and access. Unlike in many other countries, those researching Icelandic legends lacked any access to any comprehensive index of Icelandic legends, but the need was clear (not least if these legends were to be considered in international terms). Many of the published Icelandic folktale collections had certainly categorized the legends they contained on the basis certain mythical beings they mentioned or themes of the legends. In Íslenzkar þjóðsögur og æfintýri (1862–1864) collected by Jón Árnason, one can see, for example, categories such as álfar (e. elves), tröll (e. trolls) and a wide category which includes náttúrusögur (e. nature-stories). This categoriza- tion helped readers find legends of similar themes, but another problem was the difficulty in finding legends told by individual storytellers, or legends relating to a particular area which could only be done with the help of the name indices in the back of the collections. No references were made to similar legends abroad. The same problem continued in the large number of collections that followed, which even lacked reference to each other. By the time of the new six-volume edition of Jón Árnason’s Íslenzkar þjóðsögur og ævintýri published in the years 1954–1961, containing all of Jón Árnason’s collected material, things were slightly better. A motif and type index of the fairy tales in the collection had been created by Bo Almqvist based on the Aarne Thompson catalogue (Aarne & Thompson, 1928). While Almqvist had dreamed of creating a comprehensive index of migratory legends in the North Atlantic islands, like that he had compiled for Ireland (Almqvist, 2008), this had never reached completion. By the 1990s, while indexes of migratory motifs and legends in Norway, Sweden, Finland, Ireland and Scotland had already started appearing (Christiansen, 1958; Bruford, 1967; Almqvist, 1988; Jauhiainen, 1998; af Klintberg, 2010), Iceland was still a relatively uncharted wilderness. 58 Trausti Dagsson and Olga Holownia Sagnagrunnur aimed to enable users to search for legends told by specific informants or collected by a specific collector, as well as allowing considerations of where and when a legend was collected and how it might be related to other legends (with the help of the list of key words). The list of keywords used in the database of audio material were built on (in collaboration with Rósa Þorsteinsdóttir at the Árnastofnun), and these were shared with Timothy Tangherlini of UCLA, who was already working on his Danish database (Broadwell & Tangherlini, 2012, 2020). From the beginning, the dream was of connecting databases at home and abroad.

From FileMaker to A Relational Database In the making of Sagnagrunnur, Terry Gunnell enlisted the help of his students in Folkloristics at the University of Iceland who read through all the printed legends, adding metadata to a FileMaker Pro database (Gunnell, 2005). At that time, the collections had not yet been digitized yet and were only available in print. The aim was more to facilitate the finding of legends in libraries rather than encouraging reading them online. The metadata originally consisted of the following fields:

– The title of the legend in print – A bibliographic reference – The name, gender and home of the informant – The Name, gender and home of the sender/collector of the legend – The year in which the legend was recorded – ML, AT, MI (and other) international index numbers if known – All place names mentioned in the legend – Any dates mentioned in the legend – Keywords describing the content (drawn from the aforementioned list) – A summary of the legend or the whole legend in cases of shorter texts – Any poems appearing in the legend

Place names were typed down in a standard way registering the abbre- viated name of the county (Ice. sýsla) followed by a colon and then the name of the place, possibly followed by a name of the larger area indicating where the place is located, for example, a valley or a parish (e.g. Skag: Illugastaðir, Laxárdal, referring to Illugastaðir in the valley of Laxárdalur in Skagafjarðarsýsla county). Between 1999 and 2005, details of around 10,000 legends were added to the database, drawn from the main 21 printed collections. In 2008, the first online version was made available. The FileMaker database naturally had limitations. The first online version ‒ which is no longer available ‒ did not enable a search in any more than one field at a time. Its biggest issue, however, was its flat structure. Legends, Letters and Linking 59 This was not helped by the large amount of inconsistencies in the data that had arisen caused by a certain lack of standardization. The work on re- designing Sagnagrunnur in 2013˗2014 by one of the present authors, Trausti Dagsson, formed a key part of his MA project written under the supervision of Terry Gunnell (Trausti Dagsson, 2014). Using the entity-relationship database model (Pin-Shan Chen, 1976), the new data model placed the legends in one data table, and persons, places and key words others. Each of these latter tables has many-to-many relations to the legends table via intermediate tables that join the different data entities together. The conversion from the FileMaker database to a relational MySQL database was carried out via scripting. Specialized scripts were written the purpose of which was to read through the data from FileMaker converted to a comma separated (csv) spreadsheet file. One script aggregated the place names and personal names as well as the keywords, storing them in the right table in the relational database. Another script then read through all the legends, storing each the metadata of each legend in the database and linking place names, persons and key words to a corresponding entity in one of the related tables. This process ensured that each related entity was only stored once but due to the occasional misspelling of names and place names by the recorders, a number of duplicates occurred throughout the database. The work on correcting these is still ongoing, partly because many cases are ambiguous. A majority of the records were nonetheless corrected in the beginning by searching the database for entities bearing very similar names using the Levenshtein distance method to determine the closeness of two strings (Wichmann, Holman, Bakker, & Brown, 2010).

The Visual Interface For the redesigned database, a web-based interface was built which served both the purpose of browsing the data as well that of as maintaining it. This interface was built to provide tools for finding possible duplicates of related data and quickly combine two or more duplicates into one entity. The interface also provided an interactive map which could be used to locate place names geographically. At the end of this phase, about 40% of the place names mentioned in the database had been located geographically. Much of this work was done as a part the beforementioned MA project but continued after that thanks to funding from The Royal Gustavus Adolphus Academy for Swedish Folk Culture (Kungl. Gustav Adolfs Akademien för svensk folkkultur). The first version of the new interface was preserved mainly as an administrative tool while another interface was being developed to serve as a front-end interface for both researchers and the general public. The public version now featured an interactive map with categories based on keywords 60 Trausti Dagsson and Olga Holownia related to certain common supernatural beings such as ghosts, trolls, elves. Clicking on a category produced a map was populated with places mentioned in those legends which have a particular keyword. The more advanced version, intended more for researchers, featured a more powerful search format, allowing users to search for legends on the basis of on one or more of the following criteria:

– A particular printed collection – Keywords – A free-text search using the summaries of legends or titles – Legends told or collected by persons from a specific county or municipality – Legends told or collected by persons of a specific gender – Legends mentioning specific place

The search results then appear in the form of a list of found legends as well as a map showing all the places mentioned in the legends in question. A simple graph also appears illustrating the number of keywords relating to the legends found in the search and also the printed collections they are found in. These tools give the user an overview of results, before they proceed to greater detail (should they so wish). The view for each legend displays its metadata and related materials, such as the names of persons and places mentioned in the legends as well as keywords. This information also acts as links: by clicking on the name of a place or person, the user can proceed to a view displaying further metadata about the corresponding entity and those legends related to it. Relations between legends and persons can be of three types. The most common model is connecting recorders and informants, the informant being the storyteller who told the legend to a recorder who then sent it to the collector who is credited for the printed collection. In some cases, this was a little more complicated. In the case of Jón Árnason’s, Íslenzkar þjóðsögur og æfintýri, Jón did little collection himself, rather sending out calls for material, asking friends and acquaintances around Iceland to send in stories (see e.g.Gunnell, 2011; Ólína Þorvardardóttir, 1998). These people would then often appoint someone to record legends from nearby farms for them, material which was then often written up and sent to Jón. The relations model thus takes the form of informant-recorder-sender, the database taking this into account by enabling different types of relations to exist between persons and legends. Clicking on a person’s name in the database enables the user to see their dates of birth and death and a short biography, if such is available. One can also see a list of those legends told or collected by this person as well as a map displaying related places which will includes the person’s homes as well as those of any recorders that sent legends to the collector. In the case Legends, Letters and Linking 61 of persons being recorders of legends, the homes of their storytellers are also shown on the map as well as any places mentioned in their legends. This map thus gives an overview of the geographical space of those who recorded or told of legends. In some case, personal entities reflect no single person but rather an ambiguous reference to a person or a group of persons, as with “people from [the town] Grindavík” (Ice: Fólk úr Grindavík) which refers to an unknown group of people from a town but remains as a personal entity in the database to retain the link between the place and the two legends in which where this ambiguous entity is referred to as an informant.

Content of the Database Sagnagrunnur currently includes around 10.397 legends taken from 21 printed collections. These collections vary greatly in size and were collected in different time periods. The largest collection is Íslenzkar þjóðsögur og æfintýri, edited by Jón Árnason (originally published in the middle of the nineteenth century in two volumes, and later republished in its complete form in six volumes in 1954–1961) with 2,683 legends. Next in size is Íslenskar þjóðsögur og sagnir collected by Sigfús Sigfússon around the beginning of the twentieth century with 1,683 legends. The third largest collection is Íslenzkar þjóðsögur collected and edited by Ólafur Davíðsson with 1,136 legends. Following these three larger collections are various slightly smaller ones including Gríma hin nýja (981 legends), and Íslenskir sagnaþættir og þjóðsögur (692). The total number of persons named in the database is 3,165. Some 7,240 place names are mentioned, of which 4,349 have been connected to GPS coordinates, enabling them to be shown on a map. Over the last few years, the database has grown, making it more than just a database of legends. In 2016, a further database of all wonder tales found in printed sources was added, created by Aðalheiður Guðmundsdóttir, Professor of medieval Icelandic literature at the University of Iceland was integrated into the database. This new database includes metadata on 554 wonder tales, many of which come from the same sources as the legends and storytellers who are closely related to many of those people already mentioned in (Aðalheiður Guðmundsdóttir, 2016).

Documenting the Collection and Publication of Íslenskar þjóðsögur og æfintýri The year 2014 marked the 150th anniversary of the publication of the second volume of Jón Árnason’s collection work, and provide the opportunity to apply for funding to widen the scope of Sagnagrunnur, putting in into several new contexts, and not least into the contexts of other digital projects 62 Trausti Dagsson and Olga Holownia that were already up and running, but could usefully be expanded. A new project was launched with the aim of documenting the process of collection and publication Íslenzkar þjóðsögur og æfintýri which had come to be regarded as one of the most important ten works in the Icelandic literary canon. The enduring value of this collection for the Icelandic nation is that it was recently voted tenth most important work in the Icelandic literary canon, following the Passíusálmar: see http://www.ruv.is/kiljan/urslit (last viewed 29 May 2014). The core of the work involved the creation of a comprehensive digital archive that would link material found in different Icelandic archives and make it accessible online, enabling users not only to trace how the material was collected and edited for national consumption (Rósa Þórsteinsdóttir, Trausti Dagsson, & Holownia, 2019), but also for them to understand what was happening in the context of the creation of “national” cultural identity in the mid-nineteenth century, something that was already being charted by Joep Leerssen’s SPIN project in Amsterdam, and another Icelandic project relating to the figure of the painter Sigurður Guðmundsson (1833–1874) (see below). While the project had a ded- icated website (jonarnarson.is), from an early point, it was decided that Sagnagrunnur would be house the digital archive of materials (records and transcripts of letters and other manuscript material) since it already contained metadata about the material and those involved in Jón’s collection.

Linking Legends and Letters During four years of the project (2014–2018), Sagnagrunnur was expanded greatly with added metadata, including material about the collection of the material (Gunnell, 2011). The letters sent to Jón contain a valuable picture story of the communication that took place between Jón and the large number of individuals that he was in contact with during the collection and editing period. In Iceland, they frequently travel between Jón and his friends around the country, including priests and other important figures. In Copenhagen, there was his friend, Guðbrandur Vigfússon (1798–1867) whose role was to receive the manuscripts of the collection and send them on to Konrad Maurer (1823–1902) in Munich, who would be a contact person to the printer in Leipzig (Gunnell, 2010). For the Jón Árnason research project, 1,560 letters sent to and from Jón Árnason, and now stored at the National and University Library of Iceland, the Árni Magnússon Institute for Icelandic Studies, the Royal Danish Library and the Bodleian Library in Oxford have been digitized. There is also a collection of letters between Konrad Maurer and Guðbrandur Vigfússon. The letter collection as a whole consists of 3,000 letters but since the project focused mostly on letters from the period 1858–1864, the main period of collection, only the most relevant 794 letters have been transcribed. Other metadata stored for each letter Legends, Letters and Linking 63

1. Screenshot of the letters interface. includes the names of senders and receivers, and the place of dispatch as well as destination and date. The transcriptions of letters along with those scanned images we are allowed to display (with some exceptions when the permission was not granted) were then added to the online database einkaskjöl.is run by the National Library of Iceland, containing an archive of personal documents stored in the library, while the metadata was also stored as a new part of Sagnagrunnur, allowing relations to existing entities, places and persons to be created. In the visual interface, the user can therefore now view much richer information about persons in Jón Árnason’s network, seeing not only a list of legends and fairy tales collected or told by them, but also letters written to or by them. These letters can also be explored on an interactive map, which has a time-range slider, filtering material by time. As can be seen from the above, Sagnagrunnur is moving towards becoming a database of linked entities of different types relating to both folklore and personal networks, now including not only legends, wonder tales, places, and persons, but also letters. The future may well see it including other entity types. As noted above, one can see immediate links to the aforementioned Encyclopedia of Romantic Nationalism in Europe (romantic nationalism.net) (Leersen, 2018), something emphasized by the web site that grew up around the database as part of the jonarnarson.is project.

Digitized Books Another feature that came into being as a result of this project was the arrival of digitized versions of the folktales. Indeed, a key part of the jonarnarson. 64 Trausti Dagsson and Olga Holownia

2. Screenshot of a single legend, showing links to book pages and manuscript images. is project was not only the digitization of letters and original manuscript but also the digitization of a number of folk tale collections, which were also directly linked to Sagnagrunnur, as well as being made available on project the web site. The National and University Library of Iceland has now digitized around 2000 published books, making then all available on the web (see baekur.is). Jón Árnason’s Íslenzkar þjóðsögur og æfintýri was one of those works made available online, including both the original two volume edition and that which came later in six volumes. The scanned book pages have now been linked to the legend entries s in the database, thereby making it possible to read not only summaries but also complete legends as they appeared in both editions. The next step is to link entries to images of the original manuscripts of the legends which are also scanned for the project and being made available on the National Library’s website handrit.is. At present it is possible to view and read 2,600 legends as they appeared in print alongside how they appeared in their original manuscript format. That means most of the legends in Íslenzkar þjóðsögur og æfintýri.

The Website Providing a new framework (and entry point) for Sagnagrunnur is the new website that came into being as part of the aforementioned project. Íslenzkar þjóðsögur og ævintýri: Origin, Context and Collection 1864–2014 is a website which additional context in the form of provides bibliographical information about all participants, a contextual timeline relating to both local developments and those occurring abroad, links to additional visual material (such as illustrations from the English translations of Íslenzkar Legends, Letters and Linking 65 þjóðsögur og ævintýri [Icelandic Legends] and other art works inspired by this collection, and ), as well as links to the key manuscripts digitized for handrit.is, links to the other folktale collections made elsewhere at this time, and academic article dealing with various aspects of the project. The focus was on providing new entry points to the material in Sagnagrunnur and those digitized manuscripts hosted by the library. A good example of how the website and the database are integrated can be seen in a section dealing directly with collection and publication in which each name is linked directly to records in Sagnagrunnur, an embedded tengslanet (Eng. network) image then visualizing those links, enabling further searches of specific individuals. The section on the “Editorial Process” on the website, which highlights the correspondence between the editors then contains direct links to exchange of letters in Sagnagrunnur over the course of time (sagnagrunnur.arnastofnun.is). Finally, three different timelines (dealing with Jón Árnason’s life and works, the international cultural context and the Icelandic cultural context), allow viewers to position this project within the wider movements both in Iceland and abroad, something assisted by direct links to the Sigurður Guðmundsson digital archive (see below), and access to digitized versions of daily newspapers in Iceland from the period (the National Library’s timarit.is website). While they focus mostly on the publication of various folklore collections in neighbouring countries (accessible through HathiTrust Digital Library, Internet Archive and other portals), the timelines also touch on other disciplines (such as theatre and literature) and important dates in the process of “culture creation” in Iceland and elsewhere.

The Icelandic Context: Recreating an Earlier Sense of Cultural Identity through Digital Archives As noted above, the jonarnarson.is project is closely linked to another slightly earlier digital project entitled Inventing Culture: Defining Sources of Theory and Inspiration and the Long-Term Results of Culture Creation by Icelandic Intellectuals 1857–1874, which aimed to bring together and digitize a vast collection of documents produced by a particular group of nineteenth-century Icelandic intellectuals based in Reykjavík. At the centre of this project was Jón Árnason’s close friend Sigurður Guðmundsson málari (the painter), the acclaimed ‘father’ of national costume, the national theatre and the national museum, and a gifted storyteller (Gunnell, 2017a, 2017b). Sigurður and Jón, who were among the founding members of the so-called Kveldfélagið (the Evening Society) 1861–1874), played an instrumental role in what Terry Gunnell and Karl Aspelund, the principal inves- tigators of the project, have called “culture creation” (Karl Aspelund and Gunnell, 2017). 66 Trausti Dagsson and Olga Holownia Closely related and indirectly linked to Joep Leerssen’s Study Platform on Interlocking Nationalism SPIN) which in turn gave birth to the Encyclopedia of Romantic Nationalism in Europe (ERNiE: see romanticnationalism.net), Inventing Culture was conceived as a project which aimed to collect, scan and transcribe as much available material as possible relating to the work of Sigurður and his colleagues over the space of a very active 17 years in the cultural history of Iceland, spanning the period from the publication of Sigurður’s first polemical article on a traditional dress (1857) until his untimely death in 1874 (when the society came to an end). This, of course, was also the period in which Íslenzkar þjóðsögur og æfintýri appeared in print. In addition to transcripts of the minutes of The Evening Society and all of the papers of Sigurður Guðmundsson, including his letters, essays and note books, the Inventing Culture on-line database (this time using Wikimedia as a platform) contains Sigurður’s artwork and sketches as well as transcribed letters by other members of the Society, all of which come from the National Museum, the National Archives, The National Library, the Bodleian Library and the Danish Royal Library (see Holownia, 2015). As with the jonarnarson.is project, the website contains information (and links to) relating to cultural and historical context at home and abroad. It might be noted that many of the bodies noted above were once again directly also involved in the project, which was similarly funded by the Icelandic Science Fund. The bringing together of all of this material in an easily searchable format facilitates a close reading of the sources of the group’s culture-shaping ideas, mapping their inspirations and the development of their thinking within the four decades of Sigurður’s life. It effectively provides a wider context for the more narrowly focused project on Jón Árnason’s collection to which it is now directly linked. Indeed, as has been noted above, one of the main goals of all of the projects noted above was that they should not exist in an independent vacuum, but should be closely connected, and not only to materials in Iceland. While both Jón Árnason and Sigurður Guðmundsson lived on an island, they were both closely enmeshed in a range of movements that were taking place abroad.

International Context: Reaching and Linking Out As has been noted above, both of the latter projects which provide important cultural context for the material contained in Sagnagrunnur demonstrate the degree to which Icelanders were influenced by the different strains of Romantic Nationalism in Europe which naturally also gave birth to the post-Grimmian movement of collecting. All of those connections, inspirations and influences become ever more apparent with the help ofthe range digital tools and approaches that have been coming into being in the Legends, Letters and Linking 67 last few years and not least by means of the aforementioned SPIN project which gave birth to ERNiE the website, and more recently ERNiE the book (Leersen, 2018). One of key features of ERNiE is its focus on cultural networks, and the role of correspondence in these. the metadata from the letters in the Inventing Culture project has been into ERNIE with information about senders, receivers, places of dispatch and destination, and the same thing will be done with the digital letter collection produced by the jonarnarson.is project. Just as both Inventing Culture and jonarnarson.is projects have reached out, so too has Sagnagrunnur, which became the inspiration for the new digitization project conducted by The Institute of Language and Folklore in Sweden (se. Institutet för språk och folkminnen). Following the Sagna grunnur model (and format), the Swedish project has aimed at making the large folklore archives stored in Gothenburg and Uppsala available digitally, and has resulted in creation of the publicly available map-based website, Sägenkartan in which the focus has once again been on making folk legends available online via an interactive map [see Skott, Trausti Dagsson in this volume]. Shortly after the launch of Sägenkartan in 2017, two sister projects followed: Matkartan and Dialektkartan ‒ the first forming part of the Matkult project (www.isof.se/matkult) which aims to create a knowledge base focusing on Swedish traditional food culture, while the latter has centred on making available audio recordings illustrating the different dialects found in Sweden (Trausti Dagsson, Skott, & Wenner, 2018). Sägenkartan, however, did not stop at the Swedish border. Further material was provided by the Norwegian Folklore Collection (No. Norsk Folkeminnesamling), giving an example of what a larger Nordic legend map might look like (Trausti Dagsson, Skott, & Esborg, 2018). The plan is for legends collected from Swedish speaking informants in Finland to be added in the future, as well as a selection of Icelandic legends. The aim is for all of this material to become available on line as part of the even wider ISEBEL (Intelligent Search Engine for Belief Legends, www.isebel.eu) being organized by Theo Meder and Timothy Tangherlini (Meder, 2018; See also article in this volume). Indeed, the long-term aim is for ISEBEL to connect a range of databases containing legends in different languages, enabling a single-language search through all of them. At present ISEBEL includes metadata from The Dutch Folktale Database (Nederlandse Volksverhalenbank: www.verhalenbank.nl); the Danish Folktale Database (http://etkspace.scandinavian.ucla.edu/etkSpace); and The North-East German Folklore Database WossiDia (Das Digitale Wossidlo-Archiv: https://apps.wossidia.de/webapp/run). In the fall of 2020, the plan is to have Sagnagrunnur as a whole connected with ISEBEL, thereby making the Icelandic legends available for research, now within the wider northern- European context. 68 Trausti Dagsson and Olga Holownia Developments are nonetheless not restricted to outreach. As has been noted above, Sagnagrunnur has been consistently evolving and expanding, will continue to do so in the future. As noted at the start, from the beginning, plans were made for the database containing the audio recording collection of the Árni Magnússon Institute for Icelandic Studies, Ísmús, to be somehow connected to Sagnagrunnur, which is closely related in terms of contents and relevance for research. Established in close collaboration with the Icelandic National Library and its Musicology Museum, the database also includes data about musicians, songs and notation and has documented all the organs found in the churches around Iceland. Currently, the Árni Magnússon Institute for Icelandic Studies (which now hosts and main- tains Sagnagrunnur), and the National Library are working on a redesign of Ísmús, splitting it into two portals that will both use the same database. The material contained in the Musicology Museum will form one part, with folkloric data being kept in another, thereby enabling Sagnagrunnur and Ísmús to effectively join in the shape of a single folklore database. Indeed, as has been noted above, some collaboration took place at an early point with regard to standardizing the keywords used for both the printed legends and the audio recordings, and other relevant data to both databases will be added enabling further connection. Plans involve using use the latest developments in language technology from the University of Reykjavík, in collaboration with a private company called Tíró, to make use of speech recognition software on the recordings. Tíró has before developed speech recognition tools for the Alþingi ‒ The Icelandic Parliament and the X-ray division of The National University Hospital of Iceland to transcribe speeches of members of the Parliament and annotations of x-ray doctors (Inga R. Helgadóttir et al., 2019). Audio recordings of from elderly people in all parts of Iceland between 1960 and 1980 represent a particular challenge for such speech recognition technology (earlier used with modern recordings) since it involves varying degrees of audio quality and certain dialect variants. The hope is that this technology will produce transcriptions of a large part of the interviews, making them searchable, and immediately opening up new possibilities for folklore research.

Computational Research Connectability to other data is an important factor in any project like the ones discussed in this article, not only widening the context but also the approaches that can be taken, not only folkloristics but also in other fields within the social sciences and humanities. In the case of the projects noted here, this has been helped by the fact that several key research institutes working the preservation of historical data in Iceland have in recent years been digitizing large parts of their collections. Already mentioned above are the National Library of Iceland’s digitized collections now freely available Legends, Letters and Linking 69 online at baekur.is (a collection of digitized and OCR-ed books); timarit. is (digitized and OCR-ed journals and newspapers from early nineteenth century until today); einkaskjöl.is (a collection of letters and personal documents); and handrit.is (a digital collection of manuscripts) which also contains digital copies of early manuscripts made available by the Árni Magnússon Institute for Icelandic Studies, which houses Sagnagrunnur, jonarnason.is and ísmús.is. Alongside these are the National Archives of Iceland which has digitized those national censuses taken between 1703 and 1920 as well as a large number of church books and other official documents reaching back to 1600. Launched in 2012, Sarpur.is, run by the National Museum of Iceland houses digital collections from 50 Icelandic museums. Finally, there is the invaluable Íslendingabók database which enables one to examine genealogical information about any Icelander, and reaches as far back as the Middle Ages. While the ever-increasing digitization of historical documents raises ever more possibilities for computational historical research, it also underlines several problems, and not least the need for standardization, compatibility and cooperation. In Iceland, this applies especially to data containing historical person names and place names. Trying to trace an Icelander with a common name (such as Jón Jónsson) who was born in a certain year across multiple datasets can be an extremely difficult task and prone to potential error. The present version of Sagnagrunnur made use of a historical person- and place names index created by the National Library which was used to link data on handrit.is with the other databases created by the library. The personal names and places in Sagnagrunnur were then automatically linked to IDs in the library’s name index, metadata in Sagnagrunnur being linked to matching entities found in the index. Similar work is now being done as part of connecting Sagnagrunnur and ísmús.is (see above), which will result in some degree of interconnectability between Sagnagrunnur, the audio database and the datasets in the National Library datasets, the orga nizers of which are planning on enhancing standardization of their material to allow greater degrees of effective interconnectivity. Models developed in other related projects (such as Sägenkartan) can also be applied as a means of experimenting with text mining, tools such as Voyant possibly helping with both this and increased means of visualization. In many ways, it might be said that we are still just scraping the surface of the research potential within this material as ever more corridors are built between institutes, and ever more digital doors are opened through walls.

“The Dream Would Be To…” Arguably the folkloric material stored in various archives could be said to lend well to digital projects of the kind noted above (something helped by the fact that most of the material is now out of copyright). This might explain 70 Trausti Dagsson and Olga Holownia the fundamental role that the folklorists at the University of Iceland and the Árni Magnússon Institute in Reykjavík have played in the development of digital projects in Iceland. All Digital Humanities projects are, by nature, interdisciplinary, combining the knowledge and skills of subject specialists with that of those who know how to build the infrastructure for those projects. Regardless of how sophisticated tools and solutions are available, the key ingredient behind all DH initiatives is “the vision”. Terry Gunnell, Rósa Þórsteinsdóttir, Karl Aspelund and Bo Almqvist have been the “visionaries” in the projects discussed here. In this case, the vision started on one hand with Sagnagrunnur and the idea of placing Icelandic folklore within a wider, North European context. The same idea lies behind the projects relating to the figures of Jón Árnason and Sigurður Guðmundsson, except here the borders were somewhat wider, relating to the Romantic Nationalism movement as a whole. As this article shows, over the past 20 years, a range of different tools and platforms have been created to fulfil these visions, which have been directed towards re-connecting Icelandic culture with that of its neighbouring countries, at the same time underlining Iceland’s potential uniqueness. All the same, as Matthew G. Kirschenbaum has underlined (and as is suggested above), the idea of “finishing” any piece of digital work is naturally always questionable, since “‘completeness’, in a medium where the prevailing wisdom is to celebrate the incomplete, the open-ended, and the extensi- ble” (Kirschenbaum, 2009). This has certainly been the case with the projects noted above. In the course of developing new projects, tools have been redesigned and repurposed as the local has increasingly become the international. In a sense, the key hurdle that needs to be overcome with making this local material available to the outside world is language. When writing about the “Internet” in 2007, the Dutch scholar Theo Meder noted how easily collections of folktales could be naturally disseminated over the web, lending themselves to being published online, as they used to be passed on orally (Meder, 2007). It is reassuring to see that two projects he gives as examples of folktale databases (The Dutch and the Flemish Folktale Databases) have grown to inspire other projects. Stressing the importance of metadata, which was also a starting point for Sagnagrunnur. Meder nonetheless makes an important point which, considering the growing number of aggregated repositories and platforms around the world, is becoming more valid than ever: “[t]hese databases take digital archiving and retrievability a step further: they serve as advanced research instruments and can be consulted from all over the world.” As Meder adds, the single disadvantage of this material is in many cases they are not available in English (Meder, 2007, p. 491). Projects such as SPIN/ ERNiE and ISEBEL (both noted above) are taking a crucial step forward pulling aggregated metadata from a number of local databases and making them available in an international language. One might argue that they represent the new “vision”, in which machine Legends, Letters and Linking 71 aided translations can be made automatically of all the material that has been gathered in databases such as those that have been put together in Iceland in recent years, allowing Sagnagrunnur and its cousins to reach the other side of the world.

Trausti Dagsson Project coordinator The Árni Magnússon Institute for Icelandic Studies Laugavegur 13 IS-101 Reykjavík Iceland e-mail: [email protected]

Olga Holownia The British Library 96 Euston Road London, NW1 2DB United Kingdom e-mail: [email protected]

References Aarne, Antti, & Thompson, Stith 1928: The Types of the Folktale. Helsinki: Suoma lainen Tiedeakatemia. Almqvist, Bo 1988: Crossing the Border: A Sampler of Irish Migratory Legends about the Supernatural. India: Department of Irish Folklore, University College. Almqvist, Bo 2008: Midwife to the Fairies (ML 5070) in Icelandic Tradition. In T. Gunnell (Ed.), Legends and Landscape (pp. 273–322). Reykjavík: University of Iceland Press. Broadwell, Peter M., & Tangherlini, Timothy R. 2012: TrollFinder: Geo-Semantic Exploration of a Very Large Corpus of Danish Folklore. Paper presented at the Computational Models of Narrative: Proceedings of LREC, Istanbul. Turkey. Broadwell, Peter M., & Tangherlini, Timothy R. 2020: Geist, Geest, Geast, Spøgelse. Challenges for Multilingual Search in Belief Legend Archives. Arv. Bruford, Alan 1967: Scottish Gaelic Witch Stories: A Provisional Type List. 11. Christiansen, Reidar Thoralf (1958): The Migratory Legends: A Proposed List of Types with a Systematic Catalogue of the Norwegian Variants: Academia Scientiarum Fennica. Dagsson, Trausti 2014: Landslag þjóðsagnanna: Landfræðileg kortlagning á íslenskum sögnum. Reykjavík: University of Iceland. Dagsson, Trausti, Skott, Fredrik, & Esborg, Line. (2018). Nordisk sagnkart. In. Dagsson, Trausti, Skott, Fredrik, & Wenner, Lena 2018: Nätpublicering av dialekt inspelningar. Svenska landsmål och svenskt folkliv. Tidskrift för talspråksforskning, folkloristik och kulturhistoria, 140, 205–212. 72 Trausti Dagsson and Olga Holownia Guðmundsdóttir, Aðalheiður (2016): Ævintýragrunnurinn. Gunnell, Terry 2005: Sagnagrunnur: A New Database of Icelandic Legends in Print. Folklore: Electronic Journal of Folklore 45, 151–162. doi: 10.7592/FEJF2010.45. gunnell Gunnell, Terry 2010: Daisies Become Oaks: The Politics of Early Folklore Collection in Northern Europe. Folklore: Electronic Journal of Folklore, 121, 12–37. Gunnell, Terry 2011: Prestlærðir safnarar þjóðsagna á 19. öld. In Á. G. Ásgeirsdóttir, H. Björnsdóttir, & H. Ólafsdóttir (eds.), Þjóðarspegill 2011: Rannsóknir í félagsvísindum IX (pp. 598–605). Reykjavík: Félagsvísindastofnun Háskóla Íslands. Gunnell, Terry 2015: Sagnagrunnur: en kartlagt database over de islandske folke sagn. Saga och sed, 15–40. Gunnell, Terry 2017a: Sigurður Guðmundsson og þjóðsögurnar. In K. Aspelund & T. Gunnell (eds.), Málarinn og meningarsköpun: Sigurður Guðmundsson og Kvöldfélagið (pp. 415–433). Reykjavík: Þjoðminjasafn & Opna. Gunnell, Terry 2017b: Viðhorf landsmann til þjóðmenningar og þjóðsagnasöfnunar. In K. Aspelund & T. Gunnell (eds.), Málarinn og meningarsköpun: Sigurður Guðmundsson og Kvöldfélagið (pp. 235–247). Reykjavík: Þjoðminjasafn & Opna. Helgadóttir, Inga R., Nikulásdóttir, Anna Björk, Borský, Michal, Fong, Judy Y., Kjaran, Róbert, & Guðnason, Jón 2019: The Althingi ASR System. Proc. Interspeech 2019, 3013–3017. doi:10.21437/Interspeech.2019–1248 Jauhiainen, Marjatta 1998: The Type and Motif Index of Finnish Belief Legends and Memorates. Helsinki: Academia Scientiarum Fennica. Klintberg, Bengt af 2010: The Types of Swedish Folk Legends. Helsinki: Academia Scientiarum Fennica. Leersen, Joep (ed.) 2018: Encylopedia of Romantic Nationalism in Europe. Amster dam: Amsterdam University Press. Meder, Theo 2007: Internet. In D. Haase (ed.), The Greenwood Encyclopedia of Folktales and Fairy Tales (pp. 489–495). Meder, Theo 2018: ISEBEL Intelligent Search Engine for Belief Legends. Volks kunde, 119(1), 87–89. Pin-Shan Chen, Peter 1976: The entity-relationship model–toward a unified view of data. ACM Trans. Database Syst., 1(1), 9–36. doi: https://doi.org/10.1145/ 320434.320440 Þórsteinsdóttir, Rósa, Dagsson, Trausti, & Holownia, Olga 2019: Digital Collections Relating to 19th-Century Icelandic Intellectuals and Culture Creators. Paper presented at the DHN 2019, 4th Digital Humanities in the Nordic Countries, Copenhagen. Þorvardardóttir, Ólína 1998: Þjóðsögur Jóns Árnasonar: Tilraun til heimildaryni. In J. H. Aðalsteinsson & J. Jónsson (eds.), Þjóðlíf og þjóðtrú: ritgerðir helgaðar Jóni Hnefli Aðalsteinssyni (pp. 245–269). Reykjavík: Þjóðsaga. Wichmann, Søren, Holman, Eric W., Bakker, Dik, & Brown, Cecil H. 2010: Eval- uating linguistic distance measures. Physica A: Statistical Mechanics and its Applications, 389(17), 3632–3639. doi: https://doi.org/10.1016/j physa.2010.05.011 Legends, Letters and Linking 73 Databases: Danish Folktale Database: http://etkspace.scandinavian.ucla.edu/etkSpace Das Digitale Wossidlo-Archiv (The North-East German Folklore Database Wossi Dia): https://apps.wossidia.de/webapp/run Database on Iceland’s musical heritage: ismus.is Encyclopedia of Romantic Nationalism in Europe (ERNiE): romanticnationalism.net Inventing Culture: Defining Sources of Theory and Inspiration and the Long-term Results of Culture Creation by Icelandic Intellectuals 1857–1874, sigurdurmalari. hi.is ISEBEL (Intelligent Search Engine for Belief Legends): www.isebel.eu Íslenzkar þjóðsögur og æfintýri: Origin, Context and Collection 1864–2014: jonarnason.is Matkult: www.isof.se/matkult Nederlandse Volksverhalenbank (The Dutch Folktale Database): www.verhalenbank. nl Sagnagrunnur: sagnagrunnur.com and sagnagrunnur.arnastofnun.is SPIN – Study Platform on Interlocking Nationalisms: https://spinnet.eu

Sagnagrunnur was supported by The Royal Gustavus Adolphus Academy for Swedish Folk Culture (Kungl. Gustav Adolfs Akademien för svensk folkkultur. In- venting Culture: Defining Sources of Theory and Inspiration and the Long-term Results of Culture Creation by Icelandic Intellectuals 1857–1874 was organized in partnership between the University of Iceland, Árni Magnússon Institute for Icelan- dic Studies in Reykjavík, the Icelandic National Museum, the National and Univer- sity Library of Iceland and the National Archives. Íslenzkar þjóðsögur og ævintýri: Origin, Context and Collection 1864–2014 was organized was organized in partnership between the University of Iceland, Árni Magnússon Institute for Icelandic Studies in Reykjavík and the National and University Library of Iceland. Both projects were supported by Rannís – the Icelandic Centre for Research.

The Eitri Database A Digital Humanities Case Study

Katherine S. Beard

Abstract This case study presents Eitri (http://eitridb.com), a Norse artifacts database created as a companion to the author’s thesis work Hamarinn Mjǫllnir: The Eitri Database and the Evolution of the Hammer Symbol in Old Norse Mythology. Eitri is a tool that exemplifies a powerful promise of the digital humanities: that researchers and technology can continually improve one another in a virtuous circle. As researchers use technology to discern new insights and theories, they can be reincorporated back into their digital tools. These newly improved technologies can then help to generate further novel insights, continuing the virtuous circle on and on again. This article also explores the challenges of collecting and standardizing data, learned first- hand from building the Eitri database. Additionally, the analytical features of the Eitri database are explained in detail, including the map, timeline, property histogram, gallery view, and search features. Following this in-depth discussion of Eitri, the Beard Hammer Taxonomy is introduced, a new classification system for hammer amulets created with help from Eitri. The new taxonomy, now added back into Eitri, is then used to power an analysis of how hammer shapes changed over the course of the Viking Age. Finally, the future of Eitri in the digital humanities space is considered. Keywords: database, hammer, Thor, Þórr, Þórshamar, data-collection, histogram, data-analytics

Introduction Our modern understanding of the pre-Christian religions of the Viking Age is built upon only two categories of materials: primary literature and archaeology. As such, we must bring every technique and tool available to bear upon these materials, as they provide the only way to travel more than a millennium into the past and understand Viking Age beliefs. The field of digital humanities presents a novel and exciting approach to improving our understanding of primary source materials, particularly regarding archaeology. Indeed, a new technological approach was at the foundation of my master’s thesis work, Hamarinn Mjǫllnir: The Eitri Database and the Evolution of the Hammer Symbol in Old Norse Mythology, written at the University of Iceland (Beard, 2019). Hamarinn Mjǫllnir aimed to catalog and analyze every known reference to Þórr’s mythical weapon in both the extant literary 76 Katherine S. Beard corpus and the known archaeological record. To support this analysis, I created a new tool called Eitri (http://eitridb.com), which combined a comprehensive collection of hammer artifacts with modern analytical tools. Eitri is named for the dwarven smith Eitri, one half of the duo who forged Mjǫllnir in Skáldskaparmál (Snorri Sturluson 1998: 42). Its collection of hammer entries was amassed from many disparate sources throughout Northern Europe, including databases and published research. Eitri then layers a suite of analytical tools onto this data to filter, sort, organize, and visualize the available archaeological record into useful metrics for researchers. The creation of digital tools like Eitri in the digital humanities requires overcoming several key challenges. The first of these challenges is- inti mately familiar to many scholars in the field of Scandinavian folkloristics: the difficulties of data collection. The second problem is the limited avail- abilities of powerful analytical tools. Few “off-the-shelf” analytical tools are available to researchers in the humanities, and more general-purpose tools often come with a steep digital learning curve. Specialized datasets often require bespoke analytical solutions to unlock the full potential of their contained information. However, once these solutions exist, novel conclusions can be uncovered or validated by innovative digital research techniques. As I applied Eitri to my research, I discovered a virtuous circle between researcher and technology. Once created, my initial use of Eitri yielded insights into hammer shapes, which I was able to codify back into the technology as a new hammer shape taxonomy. Eitri then was able to support an analysis of how hammer shapes evolved over time throughout Viking Age Scandinavia using that new taxonomy.

Digital Humanities Challenges The analysis of Viking Age hammer symbology naturally began with a survey of the relevant archaeological evidence. Here I first experienced the daunt- ing challenges and limitations of performing analysis on the archaeological record. Thanks to several hundred years of archaeologists discovering hammer amulets and academics discussing Þórr’s hammer, a significant amount of data on the symbol has been collected across all of northern Europe. It soon became apparent that I would be required to manually amass data from a variety of disparate digital databases and library repositories, in many languages, stored in many data formats, and in many cases with limited or lost context. The first challenge associated with data collection is the variety of sources available. Many prominent universities and museums maintain their own individual digital data silos of their collections, and rarely share data. There is also no standardized way to query and collect data across different databases. Furthermore, some references to hammer symbols could only be found as textual references in published books and articles. In many The Eitri Database 77 cases, these references came with much more context and analysis than digital databases, but they require significantly more manual effort to find and extract. Finally, I often discovered duplicate records across data sources, requiring careful review that was often reliant on images or written descriptions to match records. The second challenge of a comprehensive data collection effort is language standardization. Databases and academic works have been published by authors all over Europe and beyond, in languages including English, Danish, Norwegian, Swedish, Icelandic, German, Polish, and more. The sheer number of different languages needing to be translated into one common language (in my case, English) creates a risk of data being lost in translation. Furthermore, this multilingual challenge creates an increased risk of transcription errors, especially when transcribing non-digital sources. A careful editorial review can mitigate these risks, yet this review can be costly in time and is not a flawless process. The third significant hurdle to a large data collection effort is data format standardization. Every source of data prioritizes different attributes and structures data differently. Some sources prioritize the collection of photographs and drawings of artifacts, while others simply describe their artifacts without an accompanying visual reference. Some sources may provide a single date for an artifact, while others instead suggest a date range. Different sources may use different units of measure when describing artifact dimensions or may not provide measurements at all. Written descriptions further frequently leave out important characteristics that a researcher may want to include in a potential analysis. For example, “hand-stamped” punch marks on hammers can be either circular or triangular in shape. If a hammer is recorded as simply “a stamped silver hammer,” with no accompanying image, it is impossible to explore the nature of the hand-stamps further, as the finer details of the artifact are lost. Finally, many artifacts that have been found were not recorded with relevant site context, or simply were recorded incorrectly. Such information may sadly be lost to the ages, for as Sæbjørg Walaker Nordeide notes in her study of Norwegian hammers, “most [hammers] were found 50–100 years ago or more, often accidentally, and not in modern, professional excava- tions” (Nordeide 2004: 219). Individually, these finds may not be useful as they lack a complete picture. However, in aggregate they may provide data that can prove helpful to scholars. Out of necessity, I have accepted that any database, including Eitri, is necessarily continually evolving and never truly finished. The many data issues, from hodgepodge data silos, to translation, data standardization, or lack of context, force upon the editor a responsibility to decide which sources are trustworthy and which attributes to include in their compilation and guarantee that no collection can be perfect or complete. These limitations do not mean that 78 Katherine S. Beard such data collection endeavors are fruitless, only that they are costly and intensive exercises that slow the advance and evolution of the digital humanities. The above-enumerated data collection challenges on Viking Age scholarship specifically and the digital humanities generally have had a significant impact on the field. As an example, scholars held an outdated understating of the archaeological record of hammer amulets for most of the twentieth century. Multiple notable and influential scholars believed that only approximately 50 hammer specimens had been found throughout Scandinavia dating between the ninth and eleventh centuries (Turville- Petre 1964: 83). However, even a cursory survey of today’s archeological data verifies that this number is out-of-date. In reality, the hammer count now exceeds 300, with new finds being discovered seemingly monthly. This number is partly attributable to the rise in hammer finds in the late twentieth and twenty-first century thanks to advances in metal detector technology and the increasing availability of centralized digital collections where loose finds can be easily recorded, like the Portable Antiquities Scheme in the UK (finds.org.uk). However, I believe that the difficulty of building an accurate accounting of the hammer symbol without modern technology is responsible for most of the blame for this inaccurate understanding of the archeological record. After collecting and standardizing all archaeological data relevant to a research endeavor, the next logical step of the digital humanities process is to analyze it. Unfortunately, the tools available for such research analyses are often cumbersome or limited. Existing databases that house archeological data, such as those maintained by research universities and museums, often provide only primitive analysis. Most of the databases I worked with only allowed a user to run queries and review a single record at a time. Of course, there is nothing inherently wrong with this methodology for a database that is primarily focused on collecting and storing data. However, search-and-retrieve workflows do not allow a researcher to analyze trends in aggregate or easily compare sets of records against one another, such as silver hammers found in Norway versus those found in Denmark. To conduct more modern analytical techniques that promise new insights, many scholars today are forced to rely on either general purpose off-the- shelf technologies or advanced technical expertise. Tools like Excel can certainly perform detailed analyses of artifact finds. However, it takes time to generate useful visualizations, and Excel has limited support for free-text artifact descriptions and nearly no support for images. In contrast, an enterprising academic with a programming background could write customized logic for exploring a database, but those kinds of technical skills are not widely available in the humanities. Despite the many pains cataloged above, perhaps the greatest inhibitor to the advance of the digital humanities is that scholars are forced to repeat this The Eitri Database 79 process over and over again. Many traditional scholarly analyses operate in independent vacuums, forcing each individual research publication to repeat- edly re-address the same problems. For example, one common approach to the digital humanities is for a scholar to collect data entries one at a time, often in a spreadsheet, and then laboriously generate conclusions from that private dataset. Without modern collaboration, academics in the humanities must repeat one another in isolation instead of working together as a community to collectively build and advance a solid foundation of data and knowledge. As I discovered and grappled with the many problems associated with data collection and analysis for my research on the Þórr’s hammer symbol, I realized that I would need to create a modern, digital solution in order to complete my thesis. To that end, I set out to create Eitri, a tool built specifically for my thesis research on the Þórr’s hammer symbol. Eitri offers both a centralized database of hammer artifacts and a suite of analytical tools to visualize data quickly and in a visually compelling manner. Eitri’s marriage of hard data and clear visual analytical tools with the field of folkloristics helped me to develop compelling new insights on the hammer symbol while also demonstrating the promise that the digital humanities can bring to Scandinavian studies.

Eitri Overview2 The foundation of Eitri, and indeed it’s most valuable asset, is the data that undergirds it. Eitri’s data foundation was constructed from a range of sources, including (but certainly not limited to) from the National Museum of Iceland’s Collection (sarpur.is), DIME – Digitale Metaldetektorfund (metaldetektorfund.dk), The Mu- seum of Norwegian Cultural History’s Archaeo logical Collection (unimus.no), The Swedish History LÅGUPPLÖST BILD Museum’s Collection (mis.historiska.se), The Na- Arne har efterlyst tional Museum of Denmark’s Online Collection Gallery (samlinger.natmus.dk), and The British Mu- seum’s Portable Antiquities Scheme (finds.org.uk). Eitri has collected a standardized set of artifact properties (when available) including country, current collection, find location, material, manufacture, dating, a textual description with find context, and images. Finally, Eitri stores all of its data in English. With a strong foundation in place, Eitri’s analytical features could then be realized. These features, Figure 1: An Eitri Histo detailed below, include a map, timeline, property gram, filtered by material histogram, gallery view, and search. silver (Eitri 2020). 80 Katherine S. Beard The most potent tool in Eitri is the property histogram, located along the left side of the application. Histograms are visualizations that use different width bars to display the frequency of data values. Eitri’s property histograms allow researchers to compare types of artifacts within or across specific categories. For example, a researcher could easily use the country histogram to filter down to hammers from Norway, then review the material and manufacture histograms to quickly understand how hammers were made in that country. The histogram endeavors to provide both an immediate impression of a chosen set of artifacts and a primary method for filtering data. The true power of the histogram, and of most of the other analytical tools of Eitri, is not merely in the features it provides itself. When a filter is applied in the histogram or elsewhere, that filter carries through to every visualization in Eitri. This feature empowers the scholar to narrow their focus and then compare their chosen category against everything else. To illustrate this feature, all of the following screenshots in this section will show all hammers in Eitri with a filter for silver hammers. The most visually impactful feature of Eitri is its dynamic map, which shows the geolocation of finds from the Eitri database. The map clusters hammers into groups as you zoom out, making it easy to understand both micro and macro geographical trends. If several hammers are found at a particular site, such as the 13 finds from the Truso site in modern-day Poland, a production area “[…] where goods were crafted from antler and bone, amber, glass, precious metals, bronze and iron,” the total number of hammers are included in a single point which can be expanded if clicked on for further

LÅGUPPLÖST BILD Arne har efterlyst

Figure 2: An Eitri map, filtered for silver hammers (Eitri 2020). The Eitri Database 81 exploration (Jagodziński 2010: 203). Mapping the location of artifacts provides a researcher with a useful vector for further inquiry, as when this information is considered alongside other types of data, like shape, material, or approximate time period, new conclusions can be drawn. The ability to generate a map is also useful for comparing data to another scholar’s work. For example, a scholar could contrast Stefan Brink’s map of Þórr place names (Brink 2007: 114) to a map of hammer finds across Scandinavia. The map function is useful, but not without flaws. Several hammers have ambiguous or missing find locations. When such finds lack context, or when its original provenance was lost when the item was initially collected, these artifacts do not appear on the map. Instead, what is known about the artifact’s original location is noted in their text description field. Located directly below the map in the Eitri application is the timeline. The timeline takes the form of a vertical histogram that displays the number of artifacts dated to 10-year increments. This timeline attempts to bring a sense of chronology to the archaeological record, thus making it easier for the user to spot patterns and trends over time. The timeline provides a general understanding of when the hammer symbol is thought to have appeared, when it gained prominence, and when the symbol began to decline in popularity according to the extant archaeological record.

LÅGUPPLÖST BILD Arne har efterlyst

Figure 3: A Linear Timeline of all hammer finds from 800-1100, filtered for silver hammers (Eitri 2020).

The unique nature of the archaeological record presents interesting visu alization challenges, best exemplified by artifact dating. Eitri associates artifacts with a date range to capture the unavoidable uncertainly associated with dating archaeological finds. As an example, ETR-00024, a hammer from Sejerby, Denmark, is dated to 800–900 by Nationalmuseet. To better understand dating ranges, the timeline has a toggle for two different views, “Linear” and “Weighted,” which change how artifacts are counted. Whereas most of the features in Eitri need little in-depth explanation, these two visualization options require a certain amount of clarification. The linear option 82 Katherine S. Beard displays a date in the timeline by adding up all of the hammers with dating that overlap that particular date. This view makes it easy to count how many artifacts may have existed on a specific date. The disadvantage of this approach is that it can hide how artifact popularity rises and falls over time do it its “blocky” appearance. For example, many artifacts are dated to between 800 and 1000 AD, but far more hammer amulets were likely in circulation around 900 than around 1000. The weighted timeline endeavors to compensate for the disadvantages of the linear visualization option for the timeline by better displaying the likely prominence of hammers over time. This logic behind the weighted timeline was directly influenced by the work of Angela Karlsberg in her thesis “Flexible Bayesian Methods for Archaeological Dating” (2006), which explores various mathematical approaches to modeling the deposition rate of finds across various strata of archaeological digs. The primary contribution to Eitri came from how the author represented the probability that a single artifact was deposited on a particular date within a single stratum. In brief, if an artifact is found in a stratum dated from 800–1000 AD, Karlsberg argues that it may be correct to assume that the years 800 or 1000 were less likely to be the actual date of when the artifact was deposited than a date closer to 900 AD, the middle of the date range. Therefore, Karlsberg proposes to represent an artifact’s deposition likelihood with a non-linear mathematical function (Karlsberg 2006: 71–72). As such, Eitri’s weighted visualization was created based on Karlsberg’s proposed sigmoidal distribution function (Karlsberg 2006: 74–78). The difference between this distribution and a linear distribution is most easily understood by comparing the below weighted timeline figure to the linear one above.

LÅGUPPLÖST BILD Arne har efterlyst

Figure 4: A Weighted Timeline of all hammer finds from 800-1100, filtered for silver hammers (Eitri 2020).

This screenshot shows the sigmoidal function in action: the weighting at 800 is small, rises in a curve, stays at peak weight for much of the date range, and falls dramatically at 1000 and later. The complexity in the artifact The Eitri Database 83 data associated with find dating shows the importance of bringing digital techniques and traditional humanities expertise together, as the insights provided by the humanities, such as the likely frequency of an artifact over a date range, can be used to inform technical design decisions. Eitri’s gallery view arranges images of all selected hammers into a grid. This view mode is useful for a researcher who wishes to identify a visual typology, search for a particular artifact based on its image, or simply browse collections of hammer images. The gallery is especially valuable when looking for visual similarities between hammer symbols after applying filters and can allow users to discard hammers quickly which do not fit a visual pattern. However, as stated previously, one of the weaknesses of the data collected in Eitri is that not all hammers have images associated with their particular entries. Despite this, including a tool for visual analysis allows for trend spotting and may give a scholar new ideas for where to continue their research exploration.

LÅGUPPLÖST BILD Arne har efterlyst

Figure 5: Gallery view of several silver hammers (Eitri 2020).

Hammer data lacking a well-defined category is recorded in the “text description” field of the database. Information found here can include descriptions of ornamental style, a find’s context (if known), measurements of the artifact, and any other relevant information. Because of the variety of data captured in this unstructured text field, Eitri also provides a general text search function. A user can search for any general search term, such as “hand-stamped,” and all relevant hammer entries that mention this word in any part of their description will appear. Though Eitri has many tools with high analytical value, it was designed to be usable regardless of a user’s technical expertise. Eitri opens for the first time with a tutorial to walk through the features of the application. 84 Katherine S. Beard After the tutorial completes, the entire hammer dataset is available to be reviewed and filtered as the researcher sees fit. As described above, all of Eitri’s features are clickable, dynamic, and interactive, working in concert to create a complete analytical application. When applied to the wellcu rated hammer data set, a scholar can interrogate the archaeological record in new and exciting ways. Eitri is designed to be a dynamic system, and if a type of data is consistently captured in the general text field, such as dimension data, a new property could be added easily to the database at a later date. In this way, Eitri’s features can be improved by the scholars who use and contribute to it, and the scholars who use Eitri can then advance their own research with new Eitri features. This cycle is the beauty of the digital humanities: data collection, analytics, and scholars working in concert all improve one another in a virtuous circle.

Eitri Case Study: A Novel Hammer Taxonomy Now armed with a comprehensive, standardized data set of hammers and the analytical tools to explore them, I was able to begin a modern, digital examination of the hammer’s archaeological record. As I began to explore Eitri’s data, I started to see patterns in hammer shapes. The gallery clearly showed a variety of different head shapes, shaft lengths, and suspension terminuses. Naturally, I soon wanted to understand these shape trends further: where different shapes were more prominent, the relationship between shape and material, and, most importantly, how shapes may have changed over time. However, Eitri was not yet able to help with that analysis. To properly analyze hammer shape, Eitri would need to be improved by adding a new hammer classification system, or taxonomy. This new classification methodology, named the Beard Hammer Taxon omy, aims to categorize hammers by the kinds of shapes that they commonly take. The most comprehensive prior hammer typology was created by Jörn Staecker, who divided hammer finds into twelve different categories in his publication, Rex regum et dominus dominorum: Die wikingerzeitlichen Kreuz- und Kruzifixanhänger als Ausdruck der Mission in Altdänemark und Schweden. These typologies sorted hammers by a series of characteristics, including material and decoration (see Staecker 1999: 223–233). Although the Staecker typology does take into account some hammer shape attributes, most notably whether an artifact is an instance of small hammers on a ring, Staecker types are primarily concerned with decoration and material, rather than shape. Also, since the publication of Staecker’s work in 1999, many more hammers have been uncovered and entered into public databases. The new Beard Hammer Taxonomy aims to resolve the deficiencies in categorizing hammer shape with existing typologies by categorizing hammers The Eitri Database 85 into three related categories: hammer type, shaft type, and suspension type. Each hammer can then be categorized across all of these three dimensions, providing a robust methodology to describe hammers that is also highly compatible with new digital techniques. Using Eitri’s histogram, scholars can easily filter by any combination of these characteristics as they like. Below are the three distinct categories:

A – Hammer Type B – Shaft Type C – Suspension Type 1. Ax-like 1. Short handle 1. Hammer(s) on a small loop 2. Rounded head 2. Long handle 2. Multiple hammers on ring 3. Flat head 3. Decorative handle shape 3. Suspension hole 4. V-shaped head 4. No suspension hole 5. Cross-like 5. Unknown

The first category is hammer head type, ranging from ax-like (A1) to cross- like (A5). Ax-like heads have a bladed head as a focal point, tapering into the hammer’s shaft. Rounded head hammers (A2) are similar to ax-like hammers; however, they connect to their helve rather than taper into it like the ax-like hammers. Flat head hammers (A3) are similar to rounded head, except their heads are leveled rather than rounded. The V-shaped head (A4) is the most archetypal type of hammer. It has a distinct triangular projection and shoulders that are most often squared off. The last category in the A list is the cross-like hammer (A5). The cross-like hammer is shaped like an upside-down letter T. This particular shape is differentiated from the typical Christian cross by its suspension, which is positioned at the terminus of the haft’s long end.

LÅGUPPLÖST BILD Arne har efterlyst

Figure 6: Hammer types defined by author. Graphic by author (2019). 86 Katherine S. Beard The second category (B) describes a hammer amulet’s haft or shaft. There is a clear binary distinction for most shaft types: short (B1) and long (B2). However, not all hammer shafts are straight. The third hammer shaft shape category is a catch-all: decorative (B3). Decorative shafts refer to any sort of flourishes or detail like animal/bird heads, humanoid faces, or even a decorative bar at the shaft’s terminus.

LÅGUPPLÖST BILD Arne har efterlyst

Figure 7: Handle shapes defined by author. Graphic by author (2019).

The final component of the taxonomy is suspension type, as hammer amulets are not only defined by their shape, but also by what they are (or are not) attached to. Often hammers are found with a small suspension loop attached (C1). Occasionally, this small loop will have other charms or hammers on it also, but C1 is defined by a loop that is smaller than the primary hammer it is attached to. The second category is large hammer rings (C2), which are often found with multiple charms, metal loops, and small bars. Hammers are also regularly found without the ring or chain they were originally attached to. The third suspension type (C3) describes hammers with suspension holes only. The presence of a suspension hole suggests that the amulet was once attached to some sort of ring, chain, or cord, since lost to time. The suspension hole also gives researchers clues as to the hammer’s intended orientation, which is especially important when determining if a cross-shaped amulet is a cross-shaped hammer or a regular Christian cross. Some hammer amulets have no suspension holes at all (C4). Possibly these were never suspended, instead carried in amulet pouches, pockets, or buried for ritual purposes. The last type (C5) is a catch-all for unknown suspension. Frequently C5 amulets are found with broken shafts, making it impossible to determine the suspension type correctly. The Eitri Database 87

LÅGUPPLÖST BILD Arne har efterlyst

Figure 8: Suspension types defined by author. Graphic by author (2019).

The Beard Hammer Taxonomy notes a hammer’s shape using an alpha-numeric system in the LÅGUPPLÖST BILD form of “A#.B#.C#.” Take, for example, a ham- Arne har efterlyst mer from Rømersdal (ETR-00016), currently on display at Nationalmuseet in Denmark. This particular hammer is recorded in Eitri as A4.B1.C1, which translates to a V-shaped head, short handle, and suspended on a small loop. With this new hammer taxonomy, Eitri can break down hammers by their head type, shaft length, and suspension loop and then filter using any combination of Figure 9: ETR-00016, hammer amulet from Rømersdal, Den- the three. Shape trends can also be studied using mark. CC BY-SA: National- Eitri’s map, timeline, or gallery view. museet, Arnold Mikkelsen. The creation of the Beard Hammer Taxonomy represents another example of the virtuous circle between technology and researcher made possible by the digital humanities. Without the data foundation in Eitri combined with the gallery feature, it would have been much more challenging to review enough hammers by hand to propose a meaningful taxonomy. Once the taxonomy was created, the entries in the hammer database categorized appropriately, and the new system added to Eitri’s histogram, Eitri then became even more useful for researching the hammer symbol. With the newly improved Eitri, now enhanced with the Beard Hammer Taxonomy, I could now begin an exploration that became the central 88 Katherine S. Beard conclusion of my thesis: that the shape of Þórr’s hammer changed recogniz- ably over time, morphing from ax-like amulets to the well-known V-shaped intermediate form, to cross-hammer amulet hybrids. The evolution of the hammer symbol, or the “spectrum theory” described in my thesis, postulates that the hammer symbol initially emerged out of an earlier tradition of ax-shaped amulets. One of the earliest hammer shapes that emerged was the flat-headed hammer type (A3), the first concentration of which appeared c. 500–800.

LÅGUPPLÖST BILD Arne har efterlyst Figure 10: Timeline of A3, or flat-headed, hammers (Eitri 2020).

As the popularity of hammer amulets spread, finds started to appear in the archaeological record possessing curved, yet still ax-like, heads typical of type A2. These hammers with rounded heads and squarer shoulders began to appear much more solidly in the Viking Age, c. 800–1000, illustrated by the Eitri timeline below.

LÅGUPPLÖST BILD Arne har efterlyst Figure 11: Timeline of A2, or round headed, hammers (Eitri 2020).

The most prominent types of hammers are those with V-shaped heads (A4), which emerged with a substantial spike c. 800–1000. These hammers comprise a plurality of hammer examples from the Viking Age, with 91 recorded in Eitri as of this writing. The first examples of this hammer shape type begin to appear c. 500–800, though those hammers share characteristics with several examples of single hammers on rings (Beard Taxonomy Shape C1) from southern Sweden (for example ETR-00093 and ETR-00110) and a collection of long-handled hammers (Beard Taxonomy Shape B2) from Kent, UK (ETR-00322). The V-shaped hammer type drops off dramatically from c. 1000–1100, possibly correlated with the advent of Christianity. The Eitri Database 89

LÅGUPPLÖST BILD Arne har efterlyst

Figure 12: Timeline of A4, or V-shaped head, hammers (Eitri 2020).

The V-shaped (A4), hammer shape is the only type to be definitively identified as a hammer in the archaeological record. A runic inscription on a hammer amulet found in Købelev, Denmark in 2014 (ETR-00005) bears an inscription on its head reading: Hmar x is or “This is a hammer,” illustrating that the creator or owner of this amulet understood it to be a form distinct from ax-like or cross-like amulets. The timeline data in Eitri supports this theory, with ax-like and flat-headed hammers emerging earliest, with cross- like hammers (A5) tending to appear at a much later date.

LÅGUPPLÖST BILD Arne har efterlyst

Figure 13: Timeline of A5, or cross-shaped, hammers (Eitri 2020).

As the above timeline illustrates, the first wave of cross-like hammers appears c. 800 and dropped off c. 1100. These types of amulets are generally dated to a similar timeframe as the Eyrarland Image from Eyjafjörður, Iceland, c. 950–1100 (ETR-00034). Eitri’s timeline makes clear that cross- like hammer shapes are the youngest of the various hammer shape traditions. The above case study of hammer shapes, powered by Eitri analysis, represents another loop of the virtuous circle between technology and researcher in the digital humanities. First, data was collected in service of answering a research query and analytical tools were provided to query the collected data. This initial effort enabled the creation of a novel hammer shape taxonomy, which was then reincorporated into Eitri. Finally, this hammer shape taxonomy, no in Eitri, was used to develop new insights into the way the hammer symbol shape changed throughout the Viking Age. It is easy to imagine new cycles of the virtuous circle in the future, as the above analysis could be further expanded to consider hammer shape by region, material, manufacture, or other segregators, leading to yet further insights to be incorporated into Eitri. 90 Katherine S. Beard Eitri’s Future in the Digital Humanities As mentioned above, the true power of the digital humanities cannot be unlocked without regular improvements to the technology that powers research. To that end, Eitri will continue to grow and develop across several dimensions. In the future, I plan on expanding Eitri’s scope to consider a broader view of Viking Age culture and context by analyzing and comparing promi nent symbols of the period. Symbols gain prominence in diverse regions at divergent points in time for a variety of reasons, and to understand these trends will require the full power of the digital humanities. Specifically, my analysis of symbols will require both artifacts and primary literature. Adding additional artifacts to Eitri, although costly in time and effort, is a fairly straightforward endeavor. New artifact data would be required to fit the same artifact properties captured for the hammer amulets as best as possible. Of course, as the dataset grows with new hammer finds and the inclusion of other types of non-hammer artifacts, more properties may be added to the histogram for future analysis. Nevertheless, the data collection methodology for artifacts and the myriad challenges described above will remain largely unchanged. Primary literature is a similarly powerful vector to understand where, when, and how often a symbol emerges and if it has lasting staying power. However, a digital humanities approach to literature presents a substantially different set of data challenges than artifacts. While it is easy to concep- tualize the data associated with an artifact, capturing data properties of a literary source is a more challenging endeavor. To do this, Eitri will rely extensively on tags, which track the locations in texts where specific references to symbols or other proper nouns are found. These tags can not only be used to identify where references to symbols appear, but also maintain their surrounding context. Adding tags to Eitri will also require enhance- ments to its suite of analytical tools, such as new ways to browse individual literary references or visualize them in aggregate with features like word clouds. Unlocking the power of the literary corpus and joining it to the conversation of the archeological record is an exciting possible future for the digital humanities. As noted above, Eitri is by no means limited to hammer data. It is my great hope that other scholars will find uses for Eitri by integrating their own non-hammer datasets. Analysis of other archeological finds, place names, maps, or literary references would not only be valuable to evaluate on their own, but could also unlock new types of research questions when compared to other data sets. Eitri, and other digital platforms like it, open up the possibility for novel forms of interdisciplinary collaboration that would operate on a much broader scale than the work of a single academic. The Eitri Database 91 Eitri proves to be a compelling case study on how the digital humanities can be applied to draw new insights from the limited primary sources from the Viking Age. The creation of Eitri’s comprehensive hammer database illuminated many problems with data collection, including disparate data sources, language standardization, data format standardization, and man- aging limited or missing context. This database, combined with analytical tools like the map, timeline, property histograms, and various data filters, enables a researcher to ask broad questions, identify data trends, and then explore those trends in detail. Furthermore, as I began to use Eitri to power my thesis analysis, I discovered a virtuous cycle between researcher and technology. My use of Eitri enabled me to construct a new Beard Hammer Taxonomy, which was then reincorporated back into Eitri. Now improved, Eitri was able to help understand how hammer shapes evolved throughout the Viking Age. This potential of the digital humanities, that improved technology yields new insights, leading to further improvements in technology, promises an exciting future for the combination of traditional scholarship with modern analytics. This author hopes that digital tools, like Eitri, and the virtuous cycle of improvement will be applied to other complex humanities data sets to unlock new layers of understanding of Viking Age Scandinavia that would otherwise be difficult or impossible to discern.

Katherine Suzanne Beard DPhil Candidate in Old Norse Literature Faculty of English University of Oxford 17 The Villas, Rutherway Oxford, OX2 6QY United Kingdom email: [email protected]

References Beard, Katherine. 2019. Hamarinn Mjǫllnir: The Eitri Database and the Evolution of the Hammer Symbol in Old Norse Mythology. University of Iceland. Brink, Stefan. 2007. How Uniform was the Old Norse Religion? In Learning and Understanding in the Old Norse Word: Essays in Honour of Margaret Clunies Ross, edited by Judy Quinn, Kate Heslop, and Tarrin Wills, 105–136. Turnhout: Brepols. DIME. 2019. Digitale Metaldetektorfund. https://metaldetektorfund.dk. DuBois, Thomas. 1999. Nordic Religions in the Viking Age. Philadelphia: University of Pennsylvania Press. Gunnell, Terry. 2015. Pantheon? What Pantheon? Concepts of a Family of Gods in Pre-Christian Scandinavian Religions. Scripta Islandica 66: 55–76. 92 Katherine S. Beard Historiska museet. 2019. The Swedish History Museum’s Collection. http://mis. historiska.se/mis/sok/start.asp. Jagodziński, Marek. 2010. Truso. Między Weonodlandem a Witlandem / Truso. Between Weonodland and Witland. Elbląg: The Archaeological and Historical Museum in Elbląg. Karlsberg, Angela. 2006. Flexible Bayesian Methods for Archaeological Dating. University of Sheffield. Nationalmuseet. 2019. Nationalmuseets Samlinger Online. https://samlinger. natmus.dk. Nordeide, Sæbjørg Walaker. 2004. Thor’s Hammer in Norway. In Old Norse Religion in Long-Term Perspectives, edited by Anders Andrén, Kristina Jennbert, and Catherina Raudvere, 218–223. Lund: Nordic Academic Press. Portable Antiquities Scheme. 2019. http://finds.org.uk. Sarpur. 2019. Menningarsögulegt gagnasafn. http://www.sarpur.is. Singh, Lisa, Mitchell Beard, Lise Getoor, and M. Brian Blake. 2007. Visual Mining of Multi-Modal Social Networks at Different Abstraction Levels. In Proceedings of the 11th International Conference Information Visualization, edited by Ebad Banissi, Remo Aslak Burkhard, Georges Grinstein, Liz Stuart, Theodor G. Wyeld, Gennady Andrienko, Jason Dykes, Mikael Jern, Anthony Faiola, Dennis Groth, Anna Ursyn, Andrew J. Cowell, and Ming Hou, 672–679. Washington: IEEE Computer Society. Snorri Sturluson. 1998. Edda: Skáldskaparmál I. Edited by Anthony Faulkes. London: Viking Society for Northern Research. Staecker, Jörn. 1999. Rex regum et dominus dominorum: Die wikingerzeitlichen Kreuz- und Kruzifixanhänger als Ausdruck der Mission in Altdänemark und Schweden. Lund Studies in Medieval Archaeology 23. Stockholm: Almqvist & Wiksell International. Turville-Petre, Gabriel. 1964. Myth and Religion of the North: The Religion of Ancient Scandinavia. London: Weidenfeld and Nicolson. UiO: Museum of Cultural History. 2019. Unimus: Universitetsmuseenes Samlings portaler. http://www.unimus.no.

1 Special thanks are in order for professional technologist Mitchell Beard. Eitri’s user interface would not be what it is today without his advice and programming expertise. For his academic work on information visualization, see Visual Mining of Multi-Modal Social Networks at Different Abstraction Levels (Singh, Beard, Getoor, and Blake 2007: 672–679). 2 This section is adapted from Hamarinn Mjǫllnir: The Eitri Database and the Evolution of the Hammer Symbol in Old Norse Mythology at the University of Iceland (Beard 2019). For further details of the methodology, features, and architecture of Eitri, see Hamarinn Mjǫllnir: The Eitri Database and the Evolution of the Hammer Symbol, freely available for download at https://skemman.is/handle/1946/32768 (Beard 2019). Finland-Swedish Folklore as a Versatile Online Tradition Source

Pia Lindholm

Abstract This article describes the journey of materials related to Finland’s Swedish-language folklore from the field to the archive, and from a printed anthology to becoming an online resource. The article explains why the Society of Swedish Literature, among others, began to collect information on Finland-Swedish traditions and customs, and how the materials came to be published, first in the comprehensive anthology entitled Finlands svenska folkdiktning and later in the online portal folkdiktning.sls.fi. How, concretely, were the collected materials processed for online publication, and what is the functionality of the digital version? The article also discusses how an online publication can be used to disseminate knowledge of folklore, as well as to increase the recognition of tradition-bearers and of those who went to great trouble in their time to collect the materials from the field. Keywords: folklore, oral tradition, Finland-Swedish, traditions, online resources, online platforms, digital publication, collection, archive

It is possible these days to enjoy the oldest representations of Finland- Swedish folk culture directly on the Internet. The portal folkdiktning.sls.fi gives access to all volumes of the anthology Finlands svenska folkdiktning [Finland’s Swedish Folklore], which is a nearly inexhaustible resource on Finland-Swedish traditions. Its contents include tales, legends, traditional medicine and phenomena, among many others, covering Finland’s entire Swedish-speaking area. For those interested in folk music, there are 3,000 digital tunes to discover, as well as a wealth of interesting information on the musicians and music transcribers, and the entire Finland-Swedish folk music tradition. These materials have been published, first in print and latterly in digital format, by the Society of Swedish Literature in Finland (SLS), an academic society established on 5 February 1885 in memory of Finland’s national poet Johan Ludvig Runeberg. The mission of SLS was, and continues to be, to collect, edit and publish testimonies of the origins and development of Swedish-language culture in Finland. Despite 135 years having passed 94 Pia Lindholm

1. Otto Andersson recording songs from Erik Lönnberg in Vittisbofjärd in 1912. Photo: Gösta Wid- bom. (SLS/Brages spelmanstävlingar) SLS 105b_120.

since the establishment of SLS, during which time the society has of course radically changed, the same basis for the society’s operations remains. Of course, the ways in which these operations are conducted have in many ways altered and found new forms. Today’s enormous technological advances make it possible, among other things, to make available and present materials to a huge audience via user-friendly web interfaces. The Internet plays a growing role in the work of archivists, and SLS’s operations – including making archival materials available and collecting new material – are increasingly conducted online. The digital efforts carried out by SLS have particularly intensified over the last decade. These days SLS purposefully works to provide online access to selected archival materials in their entirety, and this has been enabled by the national search engine Finna.fi, to which SLS has, at the time of writing, made more than 20,000 freely available posts. Currently the materials Finland-Swedish Folklore as a Versatile Online Tradition Source 95 searchable on sls.finna.fi represent only a fraction of all the data held in the SLS archives, but the proportion is steadily growing. Focusing on the Internet as a publication channel has also led to the conception of several new online resources, such as the letters of artist Albert Edelfelt, or the focus of this article: the digital version of Finlands svenska folkdiktning. Despite digital efforts having intensified in recent years and expanded the ways in which archival materials can be published, the actual idea of making archives available to the public is nothing new. Probably the first person to discuss the release of materials from the locked vaults of the SLS archive was the educator and cultural historian Ernst Lagus, who was one of the society’s founders and a board member from 1891 onwards (http:// www.blf.fi/artikel.php?id=9610).

Tradition-related materials at the service of nation-building When Finland fell into the hands of Russia in 1809, the idea arose that our land should aspire not only to autonomy, but to full independence. It therefore became of interest for our nation to start building its own forms of expression and national identity which, like the majority of the population, would primarily be Finnish. These endeavours turned the spotlight onto the Finnish vernacular and Finnish folklore, with the aim of discovering and awakening the Volksgeist or national spirit (Ekrem 2014:24). First and foremost, the search for the Finnish spirit was conducted in Karelia, where Elias Lönnrot collected Finnish folk poetry in the 1820s and 1830s. It is well known that these poems formed the basis for the celebrated epic, The Kalevala. Gradually, these nationalistic aspirations also began to reach the Swedish- speaking parts of the population. To prove that they, too, had a place in a future independent state, they wanted to show that along the west coast lived a people with their own culture and customs, and a long history. This laid the foundations for the thorough collecting activity that was initiated among the common populace, especially in remoter areas. Here, too, the aim was to unveil the national spirit, but also to preserve older forms of expression and traditions, which it was feared could be lost in the zeal to modernize (Ekrem 2014:32). The efforts involved many participants, including associations and students, from the mid-1800s onwards, later to be more concertedly orchestrated by SLS, after its establishment in 1885. As a result, a significant amount of tradition-related materials had already been amassed into the archives by the turn of the twentieth century. This was the heritage that Ernst Lagus was keen to bring to the attention of a broader public. The wellspring was a long, emotional speech given by Lagus to an SLS board meeting in 1908. His impassioned plea was that the materials collected with such great difficulty should not be hidden away 96 Pia Lindholm

2. Södernäs Emel chopping knarrin away in Esse in 1917. Photo: Valter W. Forsblom. Knarrin (“the creaks”) was a type of joint pain and stiffness that could affect both humans and animals. It could be cured by spells or it could be scared away. This picture and many more records of folk medicine are included in Finlands svenska folkdiktning VII:5 Folktro och trolldom. Magisk folkmedi-cin. (SLS/ Sjukdomsbesvärjelser och Omlagelser i Syd-Österbotten V) SLS 285_4.

and forgotten in the depths of the archives (Förhandlingar och uppsatser 22 1909:LXXXVI-LXXXVII). Lagus spoke for publication – and not just for the purpose of research carried out in Finland and the Nordic region, but for the delectation of interested members of the public. The edition was intended both to serve science and to become a tome that could be found on every home’s bookshelf. The board of SLS approved the scheme and a publication plan was formulated for the work that would later be named Finlands svenska folkdiktning (FSFD) (Ekrem 2014:71,126). FSFD was released as a part of SLS’s publication series between 1917 and 1975. This magnum opus encompassing more than 12,000 pages can be considered an encyclopaedia of Finland-Swedish folk culture, and it is subdivided into eight genres of folklore spread across 21 printed volumes (FSFD, vol. 1–21, 1917–1975). It embodies ancient folk traditions in the form of tales, legends, beliefs, proverbs, folk music and much more, collected right across Finland’s Swedish-speaking regions towards the end of the 1800s and the first decades of the 1900s. This extraordinary folk anthology has undergone a very colourful history, whose latest chapter was written only in 2019. Finland-Swedish Folklore as a Versatile Online Tradition Source 97

3. The five volumes of music included in the collection of Finnish folklore, Finlands svenska folkdiktning. Photo: Kai Martonen. (SLS/ Tjänstearkivet 2020) TA 2020 photo 132.

From book series to online publication The idea of publishing FSFD in digital format was brought up as early as 2008 by SLS archivists Carola Ekrem and Anne Bergman, exactly 100 years after Ernst Lagus’s fervent speech to the board of SLS. The SLS website had been overhauled in the early 2000s, and the society had set out to publish increasing quantities of its materials online. Once funds for initiating the digitization of the work had been allocated in 2008, SLS’s archivists decided to start with the folk music volumes in category VI Folk Dance, Instrumental Music (VI A1 Older Dance Tunes, VI A2 Recent Dance Tunes and VI A3 Wedding Music). This choice was based on a number of factors: that the materials in the folk music volumes were the most sought-after and utilized, and that the online version would open excellent opportunities to provide users with a great deal of information beyond that found in the printed volumes. The plans were ambitious: SLS wanted to create a portal that would allow users to listen to and download music, while also finding a great deal of interesting information on the folk music tradition, the folk musicians and the transcribers. Work began immediately. The first folk music volume was scanned and converted into code, the data was recorded in detail in several databases, the tunes were transformed into digital music using two separate programs, and information on musicians and transcribers was researched and uploaded. After two years’ intensive work, the first volume, VI A1, was published on 98 Pia Lindholm the SLS website, with facts about and digital versions of 1,100 minuets, reels and polonaises, as well as a wealth of knowledge on musicians, transcribers and the Finland-Swedish folk music tradition. Later, the site was supplemented with information on the materials in VI A2 and VI A3. A number of years later, the idea of publishing the whole of FSFD online began to take shape, and this led to a need to redesign the web platform so that it could encompass the work in its entirety. A whole lot of time was spent on designing the logic of the site, the ways in which the materials would be presented and the methods of searching the enormous quantities of data. The design had to take into account that it would contain a large amount of materials on different levels: not only the scanned and OCR-processed books in PDF format, each with its own type of content, but also the music which had previously been converted and published. It was not simple to come up with a solution that would do justice to each type of material. Work on the new platform began in earnest in 2018, and by late 2019 the latest milestone in the history of FSFD had been reached: the relaunch of the publication’s online portal, with its new layout and functionality, on a generic platform built by SLS’s own IT experts. Today, the platform encompasses all 21 volumes of the anthology. Users can search directly within the volumes, as well as downloading them. Free-text searches can be made using keywords, names or dates, among others. This means that one can search a specific geographical location, for example, and read what folk tales were told there; explore sayings from diverse Swedish-speaking districts in Finland; discover what people used to do to ward off bad luck or what games children used to play; or find out how mould was used and what its significance was, to name but a few. One can listen to digital versions of melodies and download them as musicxml files that can be reworked further using Sibelius software or other notation editors. There is also a lot of information on various types of tunes, and on folk musicians and music transcribers.

4. Screenshot of the new website for Finlands svenska folkdiktning, which was launched on 9 October 2019. Finland-Swedish Folklore as a Versatile Online Tradition Source 99 A web resource for the benefit and enjoyment of a broad audience First and foremost, the aim of the digital publication is that these interesting materials will reach an extensive audience in a format that is as easy to assimilate as possible. The books have been out of print and difficult to get hold of for a long time, so online access makes it easier for people to research or just enjoy their content. Positive user feedback has been received exactly on this aspect: people are pleased to have easy access to the materials. Positive responses have come not least from students and researchers of folkloristics. The digital format increases not only accessibility but also user-friendliness; it is a huge advantage to be able to conduct word searches, raking through the entire wealth of data with just a few clicks to find what is sought, rather than browsing through printed tables of contents and indices. The designers of the portal’s presentation made sure to observe the different types of materials, functions and levels contained therein, while still striving to make them accessible and easy to use. The search function plays a central role, as it searches not only through the digitized text but also the revised music files. For example, should one type in the search worddjävul (“devil”), it gives 537 matches comprising not only stories and beliefs related to the devil, but also tunes in which it appears. Clicking further on the results provides deeper access to the materials. Text-based results highlight the search term in excerpts from the scanned books, while clicking on the music-based results opens the processed music materials. A further goal could be that the scanned articles related to folk music, musicians and transcribers could also be seen at a glance in these shared search results. The site is a treasure-trove, particularly for persons with an interest in folk music. As I have stated above, it contains more than 3,000 tunes: minuets, reels, polonaises, waltzes, marches and many other types of melodies can be downloaded or directly played online. By downloading the tunes as musicxml files, musicians can use them as a basis for composing their own versions. Beside the music itself, the portal gives access to abundant additional information on the tunes, nearly 500 musicians and more than 50 transcribers, in the form of both articles and database entries.

Next steps for Finland-Swedish folklore The launch of the web platform is not the end of the story for FSFD. The next step will be to encode volumes V1 Older Folk Songs and V3 Singing Games in similar fashion to the previous music volumes, also making digital versions of the around 1,400 tunes therein. References will also be added for the scanned original collections. The online resource is further supported 100 Pia Lindholm by SLS’s other digital publications, including the materials found in SLS’s page on Finna (sls.finna.fi): among others, digital versions of music transcriptions by Greta Dahlström and Alfhild Forslin, the two most prominent Finland-Swedish music transcribers, who contributed hugely to FSFD’s musical volumes as editors and collectors. The original collections at sls. finna.fi make a fine complement to FSFD and give visitors direct access to transcribed music materials that are not included in the printed books – after all, the books only contain a selection of all the existing transcribed materials, which means that much interesting information remains to be discovered in the archives.

5. Greta Stenbäck (married Dahlström) and Alfhild Adolfsson (married Forslin) went out into the field together quite often. In 1924 they recorded ballads and fiddle tunes in various parts of Åboland. In the picture, Greta is noting tunes from Konrad Blomqvist on Äppelö in Houtskär. Photo: Alfhild Adolfsson (mar-ried Forslin). On this occasion one of the tunes Blomqvist played was the waltz recorded in the picture alongside. (SLS/Visor och spelmansmusik) SLS 367_18 and SLS 367_571.

As with most resources and websites, there is still room for improvement. One issue with the platform, which has also been noted in user feedback, is that downloading the books can be very slow. Naturally this depends on each user’s Internet connection, but the main reason is the size of the books themselves: some have more than 600 pages, which means that there are many large picture files that take time to download. The file size can be reduced, but only to a certain point before the picture quality and reading experience suffer. The size of the books also encumbers the search function. Finland-Swedish Folklore as a Versatile Online Tradition Source 101 Because there is so much data for the search engine to comb through, giving the results can take time. In other words, the portal is a nearly inexhaustible treasure chest of comprehensive and rich information, but the digital format has its drawbacks. Luckily, however, the pros outweigh the cons in this case.

The journey of materials from the field to digitization The ways in which the materials are presented and the kind of audience they can now reach could never have been foreseen by the early driving forces who collected them from the field or took part in having them printed. These early contributors were scholarship recipients, students and professors who worked hard to preserve the Swedish-speaking population’s traditions, tunes, beliefs, tales and legends. They were driven by their task of honour, which was to “collect folklore in Swedish, especially from the outer reaches of civilization” (Ekrem 2014:33-34). The tradition collectors’ meetings with hundreds of tradition-bearers on often toilsome journeys also deserve attention. Once the plan to publish FSFD was originally approved, efforts were made to collect materials that would fill “gaps” in the existing archives, with a view to completing the publications. Scholarships were granted to selected persons who were known to be competent field workers, able to find the missing elements. The editors of the various volumes in the FSFD anthology – many of whom also worked out in the field, collecting materials – also gradually became subject-matter experts (Ibid. 2014:125). Furthermore, there was a large number of other persons involved in the publication effort, which went on for nearly sixty years: copyists, proofers and other enthusiasts who ded- icated thousands upon thousands of hours to giving rise to these exceptional volumes. Considering the early work put in by so many people, it is especially gratifying to be able to further disseminate these fine materials and stories. The journey from the field, via the archive, to the digital web platform can be illustrated through the story of one of SLS’s most diligent scholarship recipients and most faithful employees. The Sipoo-born scholar, linguist and folklorist V. E. V. Wessman (1879–1958) first applied for a collecting scholarship from SLS in 1902. That sparked off a field-based career spanning several decades. Wessman was not only an assiduous and careful field worker but also went on to edit eight volumes of FSFD – more than any other editor. His editions were greatly admired, even outside of Finland’s national boundaries, and his contributions are considered to have been critical to the completion of FSFD. Wessman was an incredible asset for SLS. There are today 26 collections in the SLS archive that contain materials contributed by Wessman (Ekrem 2014:80-86). 102 Pia Lindholm

6. A loyal fieldworker of SLS, the teacher and linguistic scholar V. E. V. Wessman, pictured with his notepad and his trusty bicycle on 5 August 1905 after returning from a folkloristic research trip in Strömfors, Pyttis and Lappträsk. (SLS/ Fältinsamlaren V. E. V. Wessman) SLS 2186_12.

Wessman collected many representations of oral traditions and folklore, including tales, legends, superstitions, proverbs, sayings, names and dialects, and even songs and dance tunes. FSFD contains 21 melodies transcribed by Wessman, all of which are available in digital format, and thanks to the digital platform we can now enrich this with some of his own notes, as well as articles written about him and his work. One example is the following quotation included in Wessman’s unpublished biographical notes in the collection SLS 1113, found in the SLS archive, cited by Gun Herranen in her article “Med velociped, vaxdukshäfte and penna. V.E.V. Wessman som insamlare of folkloristiskt material i början of seklet” [“With bicycle, wax paper notebook and pen; V.E.V. Wessman as collector of folk materials at the turn of the century”], which can be found via folkdiktning.sls.fi:

“When I visited Lemlax in Pargas in 1915, the estate’s surveyor began to suspect that I was a German spy, as I wandered from cottage to cottage asking questions. In vain did I try to explain to him who I was, and that I was travelling around on a scholarship from the Society of Swedish Literature. He had never heard of such a society, or of any scholarships, and so forth. With great insolence he ordered me to stay put until the police came to investigate…” Finland-Swedish Folklore as a Versatile Online Tradition Source 103 This quotation gives an impression of what it could be like to work in the field. Being met with great mistrust was compounded by how arduous it could be to travel to the remote areas in which the materials were to be collected – not to mention the difficulty of paying for food and lodgings with the modest compensation received by the researcher. Naturally, there were also plenty of positive experiences; fine interpersonal meetings that could lead to lasting relationships between collectors and informants, as well as fascinating folkloristic findings that were recorded in the archives for the benefit of future generations. The articles, comments and references in the online portal allow users to place the materials in a context, often providing a more colourful picture or fascinating details about the informant or the materials themselves. In the best cases, one can find out how the materials were collected – how the meeting with the informant went, or what kind of person they were. These details and additional facts, among other things, are what make it such a joy to be able to make available these extraordinarily valuable folklore materials, which were once so painstakingly collected. The digital publication contributes to ensuring that these treasures not only survive but that awareness of them increases, while also providing visibility to the persons behind the materials – be it the transcribers or the tradition-bearers.

Pia Lindholm Archivist The Society of Swedish Literature in Finland Riddaregatan 5 00170 Helsinki Finland e-mail: [email protected]

Bibliography Printed sources Ekrem, Carola 2014: “Belysandet af vår allmoges andliga lif. Traditionsinsamlingen inom Svenska litteratursällskapet”. Ekrem, Carola, Pamela Gustavsson, Petra Hakala & Mikael Korhonen (eds.). Arkiv, minne och glömska. Arkiven vid Svenska litteratursällskapet i Finland 1885–2010. Helsinki: The Society of Swedish Literature in Finland.

Finlands svenska folkdiktning, vol. I−VII (21 volumes). SLS 1917−1975.

I Sagor [Tales] A. referatsamling [Collected Summaries]. Ed. by Oskar Hackman. 1. 1917 (SLSS 132) 2. 1920 (SLSS 151) B. Sagor i urval [Selected Tales]. Ed. by Anders Allardt. 1. 1917 (SLSS 136) 2. 1920 (SLSS 163) 104 Pia Lindholm II Sägner [Legends] Ed. by V.E.V. Wessman. 1. Kulturhistoriska sägner [Culture-Historical Legends]. 1928 (SLSS 201) 2. Historiska sägner [Historical Legends]. 1924 (SLSS 174) 3. Mytiska sägner [Mythical Legends]. 1931 (SLSS 226, 227) III Ordstäv [Sayings] Ed. by Väinö Solstrand. 1923 (SLSS 172) IV Gåtor [Riddles] Ed. by V.E.V. Wessman. 1949 (SLSS 327) V Folkvisor [Folk Songs] Ed. by Otto Andersson. 1. Den äldre folkvisan [Older Folk Songs]. 1934 (SLSS 246) 3. Sånglekar [Singing Games]. 1967 (SLSS 423) VI Folkdans [Folk Dances] A. Instrumentalmusik [Instrumental Music]. Ed. by Otto Andersson. 1. Äldre dansmelodier [Older Dance Tunes]. 1963 (SLSS 400) 2. Yngre dansmelodier [Recent Dance Tunes]. 1975. (SLSS 466) 3. Bröllopsmusik [Wedding Music]. 1964 (SLSS 402) B. Dansbeskrivningar [Descriptions of Dances]. Ed. by Yngvar Heikel. 1938 (SLSS 268) VII Folktro och trolldom [Popular Beliefs and Magic] 1. Övernaturliga väsen [Supernatural Beings]. Ed. by Gunnar Landtman. 1919 (SLSS 147) 2. Växtlighetsriter [Fertility Rites]. Ed. by Gunnar Landtman. 1925. (SLSS 184) 3. Människan och djuren [Man and Animals]. Ed. by V.E.V. Wessman. 1952 (SLSS 337) 4. Folktro om nature [Beliefs About Nature]. Ed. by V.E.V. Wessman. 1955 (SLSS 349) 5. Magisk folkmedicin [Magical Traditional Medicine]. Ed. by V.W. Forsblom. 1927 (SLSS 195) VIII Lekar och spel [Games] Ed. by V.E.V. Wessman. 1962 (SLSS 390) Herranen, Gun 1986: “Med velociped, vaxdukshäfte och penna. V.E.V. Wessman som insamlare av folkloristiskt material i början av seklet”. Historiska och litteraturhistoriska studier 61. Helsinki. Minutes of the Board Meeting of the Society of Swedish Literature in Finland, 19 November 1908. Förhandlingar och uppsatser 22, SSLS, Helsinki, 1909.

Digital sources Biografiskt lexikon för Finland:http://blf.fi/ (Ernst Lagus 13/5/2020) Finlands svenska folkdiktning: http://folkdiktning.sls.fi/ (13/5/2020) From Tradition to Data The Case of Estonian Runosong

Mari Sarv and Janika Oras

Abstract Runosong was a prevalent poetic–musical tradition among several Finnic ethnic groups in premodern times. Since its establishment in 1927 the Estonian Folklore Archives has aimed to gather and preserve all the knowledge we have about Estonian tradition, and as part of this effort started to compile a database of Estonian runosongs in 2003. That database is currently approaching completion, i.e. all the runosong texts collected and recorded in Estonia thus far will be available to researchers as well as for public use. Although it has taken a great deal of effort, not to mention time to build the system, to scan, tag, proofread and edit all the texts, the compilation of the database has been possible only through a serendipitous sequence of wise decisions and the constant work of previous generations of folklorists for over a hundred years. At the present moment the database provides access to approximately two-thirds of the runosong texts at the Estonian Folklore Archives, enabling researchers to query, map and download texts either in plain mode or with xml tags. Together with the database of runosongs created by the Finnish Literature Society, the databases cover the whole area of runosong tradition, and most of the texts ever recorded. Compiled according to the same principles, the two databases are easily usable next to each other. Further plans include reviewing the classification and typology of songs, adding tools for linguistic analysis, and integrating the sister data base of runosong melodies that is currently being reorganized into a web-accessible format. Keywords: folklore, runosong, database, Estonia, archives

What is Runosong? Runosong (in Estonian regilaul) is a shared tradition among some Finnic ethnic groups of the Finnish Gulf area, i.e. Estonians, Finns, Karelians, Ingrians and Votians. Runosong is thought to be formed on the basis of previous tradition(s) during the last phase of Finnic affinity (Late Proto-Finnic) in the first centuries of the Common Era, when the prosodic system of language proved to be suitable for the development of a syllabic meter with additional stress and quantity regulations. At around the same time significant changes in way of life also took place in Estonia and neighbouring areas – the establishment of permanent settlements can be traced to this period along with the adoption of continuous agriculture (Korhonen 1994; Rüütel 1998). 106 Mari Sarv and Janika Oras Runosong uses the stichic poetic form (no couplets and no rhyme, as with many oral epic traditions), the number of lines is not regulated, and motifs can be added on top of one other to prolong the singing endlessly. Runosong meter has a trochaic core. Placement of the stressed syllables used to be regulated according to their quantity: long stressed syllables in strong positions of the metric schema, and short stressed syllables in weak positions. In some regions the quantitative principle has been gradually lost due to the changes in language and its prosodic system, retaining, however, the trochaic core of the meter, as well as the importance of the stressed syllables in it and some typical lines that still follow the quantitative system (Sarv 2015). Another characteristic poetic feature is consistent use of a specific type of grammatical verse parallelism where the parallel concepts in two or more lines are to be interpreted as semantically equivalent, as representatives of a more general idea, irrespective of the semantic relations in the colloquial language. This kind of enlargement of word meaning in the context of parallelism makes possible the extensive use of alliteration and poetic synonyms, and alleviates the strictness of metrical norms (Sarv 2017). These poetic features (trochaic meter, parallelism, alliteration, use of poetic synonyms) together form a poetic system common to the whole area of runosong, despite the stylistic or genre differences. This poetic form, or some elements of it, regularly occur in other genres of folklore – in proverbs, riddles, sayings, charms, children’s songs, short songs in fairy tales, and even in colloquial speech forming a widespread and multifunc- tional poetic code in Finnic cultures (Kuusi 1994; Krikmann 1997). Older runosong melodies use only 3–6 notes of the diatonic scale, with most of them having a melody contour close to that of speech. Due to the cultural contacts the newest runosong melodies spread in the nineteenth and twentieth centuries have adopted western harmonic thinking (Lippus 1995; Rüütel 1998; Särg 2005). The runosongs of the Setos, a distinct group who live in the south-east border area of Estonia, form a distinct branch of the tradition with polyphonic singing and peculiar scales (Ambrazevičius & Pärtlas 2011). In runosongs, mythological motifs, plots and figures are intertwined with the realities of peasant life (family life, work, social relations, the environment), and fantasies. Most often runosongs are expressed in lyrical “I” mode, as if the singer would tell his or her story or express his or her thoughts. Narrative plots have partly also been sung in the third person, especially in Karelia where mythological topics dominate in the singing tradition. Within oral culture runosong has fulfilled several functions. Runosongs accompanied events linked to the calendar, and to the lifecycle, on the one hand to magi cally secure the continuity and sustainability of existence, and on the other hand to pass on knowledge and support ritual activities. Runosongs have also been a means to communicate either with people, animals, realia or abstract From Tradition to Data 107 phenomena (for example rain, fields, old coats, legs, feast days, illnesses) following the traditional animistic worldview. Songs have also served as a means to exert one’s social position, to express one’s feelings within the controlled framework of tradition, and to exercise one’s artistic creativity. (For an overview of the functioning and properties of runosong compared to the European ballad tradition, see Sarv 2009; for examples of sound recordings, texts, comments and English translations, see Tampere et al 2016.)

From Antipathy to National Treasure: “Cultured” Approaches to Runosong Before heading to the introduction of the Estonian runosong database we will provide a short overview of the collection history of runosongs, and the approaches leading to it. Recording runosongs meant the contact of oral and literary culture. The latter was brought to Estonia by foreign colonists who inhabited the area from the thirteenth century, under the auspices of spreading Christianity. They acquired land and the people living on it, and formed an upper class. The first notes on Estonian runosong were made by literate foreigners, mainly Germans, and range from derogatory and mocking, to the exaltation of the intact nature of Estonians’ souls. Herder’s writings and his publication “Stimmen der Völker in Lieder” (1807) increased the prestige of runosong as well as the oral singing tradition among educated Europeans with literary interests. Representing and implementing indigenous religion, runosongs were sometimes treated as part of Estonians’ superstition, and were not thought well of by the church (administered as it was by foreign clergymen until the end of the nineteenth century, when the first Estonian priests were ordained) (e.g. Laugaste 1975; Västrik 1998). Estonians who received an education or otherwise landed up in upper class, usually also changed their everyday language to German. However, as part of the national awakening, in the second half of the nineteenth century, people started to think about Estonian literary culture, and with the help of newspapers and other periodicals, to develop it. Because the stan dard of “real culture” came from the Germans, it was natural that “cultured” singing and poetry followed German, or at least Western, examples, meaning that runosong was easily considered non-culture. A notable example comes from a regional newspaper in 1857, when one of the leaders of the national movement, Johann Voldemar Jannsen, described runosongs sung in the taverns and on traditional swings mockingly, referring to the German style of choral singing as the desirable aim (Laugaste 2016 [1963]:178). In 1869 Jannsen organized the first Estonian song festival, following a similar model to that used by the German-speaking nations, to demonstrate that Estonians were equal to them. This event led to the tradition of big song festivals, which today take place every five years. 108 Mari Sarv and Janika Oras There were also views among the leaders of the national movement that resisted the adoption of German/European literary and music culture and spoke for developing Estonia’s own high culture, which would at least partly rely on folklore. In 1888, Lutheran priest and linguist Jakob Hurt organized a nationwide folklore collecting campaign through the news papers, with the aim of recording all the unwritten knowledge and poetry of the Estonians. His work, lasting 20 years until his death, clearly valued Estonian oral tradition and indigenous culture. In contrast to Jannsen, Hurt considered this knowledge a real justification to call Estonians a “cultured” nation, highlighting runosong as comparable to Greek or Roman poetry. This collection campaign, together with Mattias Johann Eisen’s similar campaign a little later, proved to be very popular, and become an important event in the process of nation building. Today both collections form the most valuable sources for Estonian folklore because they were recorded on the borderline between traditional and modern, between mainly oral and mainly literary culture. On the one hand this has provided us with a vast number of transcriptions, while on the other hand most of them were made by unprofessional collectors. In line with developing modernity, folklore, including runosong, had gradually lost its traditional function in society and was reconfigured or reinterpreted in conditions of modernity as a valuable body of knowledge on the distinct nature of the Estonians. As the idea that Estonian high culture should rely on folklore gained ground, a campaign to collect folk melodies was organized in the first decade of the twentieth century by Oskar Kallas, with the help of the first Estonian music students at the Conservatoire of St. Petersburg. In 1927, ten years after Estonia had declared independence, the Estonian Folklore Archives was established to gather and preserve all the major folklore collections and to make them available to the public, especially students, researchers, writers and composers. Since the establishment of the archives its collecting activities have mainly been performed or supervised by professional folklorists, with better and better recording facilities, although runosongs have been recorded in sharply decreasing numbers. Through the re-use of archival records, runosong, as something specific to Estonian indigenous culture, has acquired an increasingly important role on the cultural scene of Estonia, has found its way into song festivals and pop music, has become part of unofficial youth singing and the president’s annual reception.

The Organization and Systematization of Runosong Recordings in the Archives The work of folklorists in organizing collecting campaigns as well as in the archives has designed texts and metadata we have at our disposal. The tools and possibilities to manage and organize the records have changed with the From Tradition to Data 109 digital era allowing us quicker access to the texts as well as other ways to look at the material. The consistency of work throughout the collection, and across the archiving history, in systematically collecting and noting basic metadata alongside the songs has provided us with a corpus of scientific quality, enabling us to take the jump to the digital with ease. Jakob Hurt, being a member of the Russian Academy of Sciences, aimed to use scientific principles in his collecting work. By the time of Hurt’s campaign, the field of folkloristics had already developed the comparative- historical method, which focused on the study of variation with the aim of detecting the origins of the folkloric piece. Thus, for scientific reasons it was necessary that songs be written down exactly as sung, without amendments by collectors, and information about the location origin of the song also be recorded. From the very start of his campaign, Hurt insisted that collectors make note of the songs in the local dialect, exactly as presented, together with information about the singer, his or her place of residence and place of origin. The information on how, when, where and why singing took place was not of primary interest for the research of this time, and was recorded only occasionally. The lack of contextual data is to an extent alleviated by the existence of fieldwork diaries, which became a standard in the collecting activities of bursars and professional folklorists from the start of the twentieth century. When the Estonian Folklore Archives was established in 1927, the primary principle of its first head Oskar Loorits was to get the overview on the contents of the collections (Loorits 1932:8). The first task was to create content registers of the existing collections, and to make typewritten copies of selected texts (including all the runosongs, but not, for example, rhymed songs belonging to the newer layer of songs). During this work, existent metadata was systematized: for every item a parish of origin was detected, if possible, and an additional metadata layer of genre and folkloristic type was created. In typewritten copies a strict structure was established where this information was added in a standard order, separated with specific signs. Later on, the exemplars of typewritten copies were organized into dossiers in different ways: in the order of income, by location, and by folkloristic genre and type, and were available to archive visitors so that they could have quick access to the materials they were interested in, and to save the original manuscripts or recordings. Compiling handwritten registers and typewriting copies of original material lasted until the 1990s when digital workflows gradually started to emerge in archive management.

Establishing and Development of the Database of Estonian Runosongs The first stage of switching to digital archival processes in the mid-nineties included the switch to recording fieldwork materials in digital format; the 110 Mari Sarv and Janika Oras start of the systematic digitization of analogue sound recordings and photographs; and occasional digitization of texts, photographs, musical notations, sound recordings in course of preparing the publications of archival materials. All the files with digital copies (or digital originals) were gathered in a folder system to be accessible to researchers. The researchers, though, needed a more feasible system to access and select the materials for their research, thus several databases based on the collections at the Estonian Folklore Archives were initiated by different folklore research groups at different institutions, according to their research interests and profile (Järv 2016). Financial support for this work was often applied and gained by the state-funded Estonian Language and Cultural Memory program in its different stages, and other occasional funding programs. Because in Estonia the folkloristic research of this period was lined up mainly according to genres, these databases tended to be focused on specific genres, also functioning as an organizing tool for vast archival collections and taking over the role of card indexes. Among the group of runosong researchers at the Estonian Folklore Archives the idea to organize the digitized runosong texts into a database evolved in the second half of 1990s, with different options considered. In 1993 KVIS, an infosystem of Estonian cultural values, was established that was meant to gather information about materials in Estonian museums, but this proved to be unsuitable for work with text records. When we consulted corpus linguists at the University of Tartu they suggested that we should mark-up texts and metadata using the Text Encoding Initiative standard (Sarv 1996), but we lacked the necessary competence to set up the system and handle the texts. As an easily accessible alternative to texts in MS Word files, a database in MS Access was created to store the digitized texts and to provide researchers with a corpus covering the whole of Estonia. In 1998 another initiative from Finland intervened in this process. Similarly to Estonians, the Finns had come up with the idea of making their runosong collections available digitally. At this point optical character recog nition technology had developed to the state where it was meaningful to use it to convert printed texts to digital. For this exercise the Finns had the opportunity to use volumes from the academic publication Suomen Kansan Vanhat Runot (SKVR, “The ancient poems of the Finnish people”, 1908– 1997) containing most of the runosongs at the Finnish Folklore Archives in an edited, systematized, proofread and well-printed form. For economic reasons the digitization work was done in Estonia. Under the direction of professor Arvo Krikmann the necessary technical equipment and an appropriate team were compiled at the Estonian Literary Museum. In the course of the project all 34 volumes of the series were scanned and optically recognized, and non-standard characters were given Unicode numbers so they could be displayed correctly. On the basis of the standard layout style and From Tradition to Data 111 specific marks used throughout the publication series xml tags on text structure and metadata were automatically added with the help of an MS Word macro written by Arvo Krikmann (Klemettinen 2006). As a follow-up to this project, and partly with funding from Finland, all the machine-typed copies of runosongs at the Estonian Folklore Archives were scanned and optically recognized. The team of runosong researchers at the Estonian Folklore Archives (Janika Oras, Liina Saarlo, Mari Sarv) now had the task of converting the results into reliable and usable output. Since 1997 the work with digitization, establishing the database, and computational research on digitized runosongs has been performed in the framework of numerous projects applied for and led by Janika Oras. The most stable of the various sources has been a project for compiling academic publications and a database of Estonian runosongs within the framework of the state-funded Estonian Language and Cultural Memory program and its different stages, operating since 2004. In 2017 the project was reorganized to receive some of the permanent funding allocated by the Estonian Ministry of Education and Research for the development of research archives at the Estonian Literary Museum, sparing the team the annual application and reporting process. The wise decision of the founders of the Estonian Folklore Archives to typewrite song texts over 70 years on sheets with a standard size, with a header and a signature containing metadata in a standard sequence with standard marks and spaces, miraculously spared us from the time-consuming typing work that might otherwise have been required, and after scanning and optical recognition in fact we had the data in an almost machine-readable form. From the start of the collaboration with the Finnish Literary Society in digitizing the records it was decided wherever possible to use the same principles the Finns had used in compiling the SKVR database in building the database of Estonian runosongs in order to ensure the interoperability of the databases. That meant the metadata and texts were organized as xml files with a schema that easily matched the tradition of archival signatures in neighbouring archives. Metadata typically includes an archival reference; the location where the song was recorded (parish, and other administrative units from different eras); name and eventual additional information of the collector; name, age, and eventual additional information of the performer; year of collection; folkloristic genre (in addition to runosong the items may be classified according to other genres, for example as proverbs, games, charms, customs and beliefs, fairy tales, newer rhymed songs); folkloristic type(s) (attributed to the songs by the researchers and archivists of different periods); eventual title of song (or group of songs) in the manuscript attributed by either performer or collector. For the sake of a better quality of outcome, we decided to proofread all the texts for typing and scanning errors, inconsistencies and specific characters in manuscripts, and to mark 112 Mari Sarv and Janika Oras some parts of the texts (refrains, comments, etc.). All the volumes were also to be checked for missing texts and data. In addition to the raw version replicating the spelling and punctuation of the original manuscript as faithfully as possible, we decided to create an orthographically normalized version of the texts (of course retaining all the dialectal features), necessary for word queries and smoother reading for non-archivists. For the purposes of querying, parish names from the location data and names of collectors were to be normalized. A small team of proofreaders started in 2004, and working on this is on-going. Parallel to this, software development company Piksel has designed and developed a database system that enables storage, querying and downloading of text and data, also adding features and functionality when requested. The database (http://www.folklore.ee/regilaul/andmebaas/) was presented to the public in 2010, when it contained approximately one third of all the runosong texts held at the Estonian Folklore Archives. Since then a lot of work has been done proofreading, tagging/encoding, inserting new information, etc., in addition to which the database been supplemented by tools that allow inserted data to be corrected, and folkloristic typology to be revised. The database has been connected to the digital archival system (launched in 2012) of the Estonian Literary Museum (http://kivike.kirmus. ee/). In 2016 IT students at the University of Tartu (Joosep Hook, Mihkel Kaasik, Raina Liiva) developed additional functionality for mapping query results within the framework of a university software project. Although Finnic runosongs form a common tradition with many shared themes, motifs, formulae, etc., the folkloristic typologies developed by Finnish and Estonian folklorists are not synchronous with each other. After completion of their publication series, Finnish runosong researchers spent years reworking and unifying the typology of the runosongs in the Finnish archives. The type signs on Estonian typewritten copies reflect the varying folkloristic vision of the typology at the time of typing. The traditional type names were revised and reorganized during this period, meaning that the typology we have in our database is not consistent. Therefore, in parallel with proofreading we have started to revise the typology in the database, and to implement this the system required some further development. In the revised typology the larger thematic group of songs are identified (for example songs about singing, songs about sorrow), as well as the function according to when the song is known to have been sung (for example songs sung by drunkards, songs at harvesting). At the moment two parallel versions of the type names are available to researchers. Once the work is completed the initial type names can be hidden, and the folkloristic song types in the Finnish and Estonian databases can then be matched. From Tradition to Data 113 The Current State of the Runosong Database, Uses and Outlooks At the moment the database of Estonian runosongs contains more than 100,000 texts (approximately 10M words) from the Estonian Folklore Archives together with metadata: an archival reference as a minimum, but in most cases also information about the location, collector, performer and year of collecting, folkloristic genre, thematic group of songs, function and type name(s). Apart from pure runosong texts (which form the majority), the database also contains pieces of folklore from other genres that either contain the elements of runosong, have some connections with runosong, or give us information about songs or singers. The number of runosong texts in the database has reached approximately two-thirds of the whole volume of runosong texts in the archives. The database interface enables the user to query the data, to download the results in txt or xml format, and to see the results on a map. Queries enable users to choose one or more locations by larger units (county) and/ or by smaller units (parish), to choose the collector by normalized name morphology, to restrict the period of collection, to choose genre, thematic group, function, typename. In addition, users can make a text search from the longer units: text, metadata (so that the user also has the option to search from fields that cannot be queried, the most important of which is a precise location of text origin (for example village, household), name of performer or preliminary typename), and comments. Users also have the option to choose if the raw or normalized text version is presented, and the number of the results per page. The query results are presented in a form of table displaying the selected metadata: archival reference, exact location of origin, name of collector, and collection year. Clicking on the archival reference enables users to see the text in a pop-up window. There is also the option to display the query results on a map showing the number of texts per parish, or alternatively the per- centage of texts recorded from a given parish, to eliminate the influence of collecting density, which differs remarkably across Estonia. The map page also displays the number of occurrences in the form of a table, including results that are not shown on the parish map (if the texts are of unknown origin, if a larger area is given as place of origin, if the texts have been recorded in larger towns or abroad). Maps and tables can be downloaded in svg and csv format, or printed (as pdf for example). The query results can be downloaded as a whole or as selected texts in txt or xml format. Downloading is restricted to 1,000 texts per query; if access is required to the whole corpus, or to a larger number of texts, a special contract has to be signed. The interface for database compilers also contains the tools required to correct data files; to revise and comment on typology; to manage genres, 114 Mari Sarv and Janika Oras thematic groups, and functions; a couple of additional query fields; and to make downloads without restriction. All the texts in the database are also accessible through the Estonian Literary Museum’s digital archive system (http://kivike.kirmus.ee), where the scanned manuscript pages can also be consulted. The system also has a facility for virtual exhibitions or community portals where selected archival materials can be presented together with comments and illustrations, for example in the Laiuse parish portal (Saarlo 2018). In addition to runosongs fieldwork notes, sound recordings, information on singers and collectors, etc., can also be explored. The database has become a dependable tool for runosong researchers in the first instance, and researchers of other fields willing to look for information in runosongs: it already contains songs from the most valuable older runosong collections, it is of reliable quality (i.e. texts are carefully proofread), and the selected texts are easily downloadable. There have already been the first attempts to use the Estonian and Finn- ish databases in joint analysis (their similar design facilitates this kind of work) (Sarv 2018). The database is also actively used by people interested in regional culture. Folksongs offer valuable information on cultural milieu, on everyday life and feasts, attitudes, values, joys and sorrows of peasant life. As the folklore collections cover all of Estonia, almost everyone can find songs sung in their region, village, or even family. There are numerous folk groups, performers, musicians, and composers who are looking for sources for their performances. The runosong melodies recorded are scarce compared to the number of texts, the corpus of which offers endless possibilities to recon- figure a composition due to the property of runosong tradition that in principle allows us to combine any text with any melody. The folksongs from the database are also sometimes used to compile publications on regional culture. Further plans with the database depend on the further interests of researchers, and on funding. The last third of the texts will be proofread, tagged and added to the database. The next task is to revise the folkloristic typology of runosongs, which would enable queries of better quality and would further enhance collaboration with Finnish researchers. We have also been looking for the possibility to add a layer of linguistic analysis to the database. Because the songs used archaic dialectal language, the tools for language analysis developed for standard language would not be directly applicable to runosong. The additional layer would enable researchers to query the occurrences of words by lemmas, which at the moment is impossible because of the dialectal variability, and also because of how Estonian changes word forms and even roots in grammatical forms (declinations). Linguistic mark-up would also enable us to analyse linguistic variation in From Tradition to Data 115 folksongs and compare the language of runosong with the properties of colloquial language. There are also plans to develop better cohesion between the database and the digital archival system of the Estonian Literary Museum, and to allow for data sharing with the sister database of runosong melodies that is currently in the process of being reorganized into a web accessible format.

Acknowledgement This study was supported by the Estonian Ministry of Education and Research (IUT 22-4 Folklore in the process of cultural communication: ide- ologies and communities and EKKD65 Source documents in the cultural process: Estonian materials in the collections and databases of the Estonian Literary Museum), by the Finnish Academy (project no. 333138 Formulaic intertextuality, thematic networks and poetic variation across regional cultures of Finnic oral poetry), and by the European Regional Development Fund (Centre of Excellence in Estonian Studies).

Mari Sarv Senior researcher Estonian Folklore Archives Estonian Literary Museum Vanemuise 42 51003 Tartu Estonia e-mail: [email protected]

Janika Oras Senior researcher Estonian Folklore Archives Estonian Literary Museum Vanemuise 42 51003 Tartu Estonia e-mail: [email protected]

References Ambrazevičius, Rytis & Žanna Pärtlas 2011: Searching for the ‘natural’ origins of the symmetrical scales: traditional multipart setu songs. Journal of Interdisciplinary Music Studies 5. DOI: 10.4407/jims.2011.07.001. http://musicstudies.org/wp- content/uploads/2017/01/Ambrazevicius_JIMS_11050101.pdf (accessed 10 May 2020). 116 Mari Sarv and Janika Oras Järv, Risto 2016: The Singing Wolf Meets His Kin Abroad. Web-based databases of the Estonian Folklore Archives. Estudis de Literatura Oral Popular. Studies in Oral Folk Literature 5: pp. 29–44. https://revistes.urv.cat/index.php/elop/article/ view/1797/1720 (accessed 25 May 2020). Kallio, Kati & Frog & Mari Sarv 2017: What to Call the Poetic Form: Kalevala- Meter or Kalevalaic Verse, regivärss, Runosong, the Finnic Tetrameter, Finnic Alliterative Verse or Something Else? RMN Newsletter 12–13: pp. 139–161. www.helsinki.fi/sites/default/files/atoms/files/rmn_12-13_2016-2017.pdf (accessed 27 May 2020). Kinnep, Aime 2016: Laula, laula, suukene, liigu linnukeelekene. Regilaule Rakke ümbrusest: Koeru ja Simuna kihelkondade regilaulud. Rakke: Rakke Valla Hariduse Selts. Klemettinen, Pasi (ed.) 2006: ”Ei se synny synnyttämättä” Selvitys digitointiprojektin vaiheista ja työprosesseista. Suomalaisen Kirjallisuuden Seura. https://www. finlit.fi/sites/default/files/mediafiles/tutkimus/elias_loppuraportti.pdf (accessed 30 January 2019). Korhonen, Mikko 1994: The Early History of the Kalevala Metre. In Anna-Leena Siikala & Sinikka Vakimo (eds.), Songs Beyond the Kalevala. Transformations of Oral Poetry. Studia Fennica Folkloristica 2. Helsinki: Finnish Literature Society: pp. 75–87. Krikmann, Arvo 1997: Sissevaateid folkloori lühivormidesse I. Tartu: Tartu Ülikooli kirjastus (http://haldjas.folklore.ee/~kriku/LEX/ KATUS.HTM). Kuusi, Matti 1994 [1978]: Questions of the Kalevala Metre. In Anna-Leena Siikala & Sinikka Vakimo (eds.), Songs beyond the Kalevala. Transformations of Oral Poetry. Studia Fennica Folkloristica 2. Helsinki: Finnish Literature Society: pp. 41–55. Laugaste, Eduard 1975: Eesti rahvaluule. Tallinn: Valgus. Laugaste, Eduard 2016 [1963]: Eesti rahvaluuleteaduse ajalugu. Valitud tekste ja pilte. Tartu: EKM Teaduskirjastus. (http://www.folklore.ee/rl/pubte/ee/laugaste/ Laugaste1963.pdf). Lippus, Urve 1995: Linear musical thinking: A Theory of Musical Thinking and the Runic Song Tradition of Baltic Finnish Peoples. Studia musicologica Universitatis Helsingiensis 7. Helsinki: University of Helsinki. Loorits, Oskar 1932: Eesti rahvaluuleteaduse tänapäev. Olevase ülevaade ja tulevase töökava. Vanavara vallast. Õpetatud Eesti Seltsi kirjad 1. Tartu: Õpetatud Eesti Selts: pp. 7–34. Oras, Janika 2017: Favourite Children and Stepchildren. Elite and Vernacular Views of Estonian Folk Song Styles. Res Musica 9: 27–44. (https://resmusica.ee/wp- content/uploads/2017/11/rm9_2017_27-44_Oras.pdf). Palmiste, Arlet 2014: Laulud Sääskülast. Vahi: Ecoprint. Rüütel, Ingrid 1998: Estonian folk music layers in the context of ethnic relations. Folklore. Electronic Journal of Folklore 6. https://www.folklore.ee/folklore/vol6/ ruutel.htm (accessed 25 January 2020). Saarlo, Liina 2018: Laiuse pärimus. http://kivike.kirmus.ee/index.php?id=15& module=415&op=11 (accessed 9 February 2019). Särg, Taive 2005: Eesti keele prosoodia ning teksti ja viisi seosed regilaulus. From Tradition to Data 117 Dissertationes folkloristicae Universitatis Tartuensis 6. Tartu: Tartu Ülikooli Kirjastus (http://www.utlib.ee/ekollekt/diss/dok/2005/b17338311/sarg.pdf). Sarv, Mari 2018: Regional variation of Finnic runosongs on the basis of word frequencies. https://www.slideshare.net/Marii1111/regional-variation-of-finnic- folksongs (accessed 9 February 2019). Sarv, Mari 2017: Towards a Typology of Parallelism in Estonian Poetic Folklore. Folklore. Electronic Journal of Folklore 67: pp. 65−92. http://www.folklore.ee/ folklore/vol67/sarv.pdf (accessed 27 February 2019). Sarv, Mari 2015: Regional Variation in Folkloric Meter: The Case of Estonian Runosong. RMN Newsletter 9: pp. 6–17. https://www.helsinki.fi/sites/default/files/ atoms/files/rmn_09_2014-2015.pdf (accessed 28 May 2020). Sarv, Mari 2009: Stichic and Stanzaic Poetic Form in Estonian tradition and in Europe. Traditiones, Slovenian Journal of Ethnography and Folklore, 38 (1): pp. 161−171 (https://ojs.zrc-sazu.si/traditiones/article/view/1059/839). Sarv, Mari 1996: Kõik regilaulud arvutisse! ehk Kolkasse elama. Vaga vares. Pro Folkloristica IV: pp. 104−109 (http://www.folklore.ee/era/nt/PF4/). Tampere, Herbert & Erna Tampere & Ottilie Kõiva 2016: Eesti rahvamuusika antoloogia. Veebiväljaanne. Janika Oras & Kadi Sarv (eds.). Helisalvestusi Eesti Rahvaluule Arhiivist 3 (http://www.folklore.ee/pubte/eraamat/rahvamuusika/en/). Västrik, Ergo-Hart 1998: Clothed Straw Puppets in Estonian Folk Calendar Tradition: A Shift from Cult to Joke? Folklore. Electronic Journal of Folklore 7: pp. 38−78. https://www.folklore.ee/folklore/vol7/metsikx.htm (accessed 12 February 2020).

1 On the multiplicity of terms denoting Finnic song tradition see Kallio et al 2017. 2 For a discussion see Oras 2017:29–30. 3 The Estonian Folklore Archives (EFA) was established in 1927 as an independent institution under the Estonian National Museum. In 1940, with the Soviet invasion, national institutions were either disbanded or reorganized. Three archival institutions of the Estonian National Museum were reorganized into the State Literary Museum of the Academy of Sciences of the Estonian Soviet Socialist Republic, while EFA was renamed the department of folklore. After Estonia regained independence in 1991, the museum was renamed the Estonian Literary Museum, and the folklore archives regained its original name, the Estonian Folklore Archives. 4 Examples of regional publications, compiled by amateurs using runosong database: Palmiste 2014; Kinnep 2016.

Nineteenth-Century Cholera Epidemics in Sweden from a Popular Perspective

Anders Gustavsson

Abstract My focus is on the many cholera epidemics which hit Sweden from 1834 until 1873. I have studied how cholera epidemics affected the countryside from an ethnological, folkloristic and cultural historian’s point of view. Popular practices and ideas in difficult crisis situations are the subject, rather than top-down regulations. How did the population perceive the cholera and how was the disease treated on the local level when it broke out? The different epidemics share a common feature: the disease has been spread by shipping across the oceans and then along inland waterways and the shores of larger lakes. A cholera epidemic outbreak immediately raised questions regarding barriers against the immediate neighbourhood. Smoking with juniper or tar was used as a protection against cholera infection. This points to the opinion that the cholera contagion was airborne, namely a miasmatic view. In towns, the disease hit the socially weak areas where poverty, bad hygiene, and overcrowding reigned. This tendency was apparent in the countryside as well. Since the cholera hit local communities suddenly and many died within a short time, it is to be expected that strong fears appeared. There are informants who had survived cholera and left tales of shattering memories. In many cases, the themes in legends about cholera had roots going back to the Black Death in the fourteenth century. Keywords: cholera, contagion, epidemic, legend, miasma

In Arv: Nordic Yearbook of Folklore 2017 I studied a coastal area in Western Sweden concerning the establishment of public health services in the nineteenth century in the form of district physicians, pharmacists, and midwives, services which earlier had been available solely in cities and towns. Rural culture was confronted with a new officialdom present in cities and towns. The earlier folk healers had not previously met competition from medically trained doctors (Gustavsson 2017:51–90). I will maintain the focus on the nineteenth century as I study the various cholera epidemics which affected Sweden from 1834, lasting for a number of decades. Cholera is a bacterial intestinal disease with diarrhoea, fever and vomiting, and with a rate of mortality of some 50 per cent. The onset is sud- den and is accompanied by severe pain. The important losses of fluid lead to 120 Anders Gustavsson dehydration and the body contracts to become unrecognizable. Death might occur within hours (Tallerud 1999:113; Schiøtz 2017:237ff). The extensive Swedish cholera epidemic of 1834 led to the establishment of district physi cians in the countryside. In 1834, medical doctors might be sent out from towns to the countryside on temporary duty (Öberg 1988; Jacobsson 1989).

The Problem Being an ethnologist, folklorist and culture historian, I wish to study the impact of the cholera epidemic on the countryside. My focus is on popular practices and perceptions in severe crisis situations rather than upon regulations imposed from “above”. How was the cholera seen by the inhabitants and how was the disease handled on the local level? What was seen as the cause of the disease? What preventive measures might be available? Was isolation tried locally and/or between regions? How did the local population care for the sick and the dead? What happened on the psychological level regarding horror and worry? The epidemic was most widespread in the towns and came there first, as described in a number of town studies. This is what happened in Uppsala (Tallerud 2006), Stockholm (Zacke 1971) and Gothenburg (Öberg 1988). The countryside, however, has not been studied in the same detailed way, at least not from a popular point of view. The medical historian Sven-Ove Arvidsson has placed the Swedish cholera epidemics in a global context in his doctoral thesis in Stockholm “The Swedish Cholera Epidemics” by showing the major propagation paths leading to the various outbreaks in Sweden (Arvidsson 1972). The author and medical historian Berndt Tallerud’s surveys have also been important in my study (Tallerud 1999, 2006).

Sources The Swedish Folklore Archives in Gothenburg, Lund, and Uppsala have rich recorded material which was collected in the early 1900s from all parts of Sweden. It contains stories about local developments during various cholera epidemics. The informants were, most often, born in the late 1800s and most lived in the countryside as did most Swedes at the time (cf. Skott 2008; Gustavsson 2014). In 1853, that share was almost 90 per cent (Arvidsson 1972:53). There are some cases where stories are told by informants who had personal memories of a cholera epidemic, or had heard stories told by close relatives, about how the local environment had experienced terror, isolation, etc. I have also read local historical literature which tells stories of local events. I use stories from various regions in Sweden together with in-depth reviews on the Swedish west coast. Nineteenth-Century Cholera Epidemics in Sweden 121 Added to stories of what has actually happened in local communities as told in folklore records, there are also legends associated with the cholera epidemics. The basic ideas behind these legends can be traced back to the Black Death in the fourteenth century in both Sweden and Norway. Issues regarding old popular traditions being revised to fit new situations will be discussed along with questions regarding a possible basis in truth in old legends. This issue was raised recently by the distinguished German ethnologist Helge Gerndt, Munich (Gerndt 2020).

Public Regulations I use public regulations as background information to developments in local communities. There are government regulations from as early as 1831 and somewhat later, 1833, when the threat of cholera was approaching Sweden across the Baltic Sea from Russia. Regulations were issued regarding quarantine sites for ships arriving in Swedish harbours (Arvidsson 1972:106f). The first such quarantine site was situated on the island of Känsö in the Gothenburg archipelago. The required quarantine was ten days for ships from a town with an epidemic and five days for ships from a suspected town. Corresponding regulations were introduced in Norway in 1831, then in a union with Sweden. The cholera arrived there earlier than in Sweden, i.e. in 1832 and in 1833 and 1834 (Bente Alver et al. 2013:52ff). A Swedish quarantine site was established for travellers from Norway on Öddö (close to Strömstad) as early as in 1832. The Swedish quarantine regulations were modified through a new regulation in 1859. The older quarantine regulations regarding arriving vessels had not given the desired effects (Arvidsson 1972:139). In 1860 Norway established a Health Law which required that Health Boards be established in all local municipalities. Earlier, such com- missions had been temporary when an epidemic broke out. From 1860 on, these boards were charged with fighting epidemic diseases and with taking preventive measures (Alver et al. 2013:57, 248; Schiøtz 2017:239ff). In 1831, Health Boards were required in Sweden as well, in towns as in local municipalities. They were to be activated when epidemics broke out. Designated “cottage hospitals” were to be established in which diseased individuals could be isolated. In the countryside, it was quite common that this “cottage hospital” was where the first cases appeared. The Health Boards were to issue regulations regarding the care of the diseased so as to hinder the spread of the infection. Cleanliness was emphasized. “All roads, squares and streets between houses are to be brushed clean and kept free of all uncleanliness, infringements to be punished by fines” (Royal announcement November 12, 1831). This indicates an awareness of the role of contagion in the transmission of disease (more to follow below). Similar opinions are present in minutes from local Health Boards, such as 122 Anders Gustavsson the Torsåker municipality in the province of Gästrikland, where the Board minutes say: “remind forcefully that cleanliness regarding body, houses, yards and clothes, together with an orderly and sober way of life should be maintained at all times” (ULMA 21096). Other scientists have carried out studies specifically into the minutes of local Health Boards in some cities (Stockholm and Uppsala, as well as Falun, Kalmar and Karlskrona), but much less so in the countryside. I had to build upon local literature. The minutes which have been saved or copied show how conscientious the local Health Boards could be. They closed roads, bridges, and harbours to isolate the local community from their immediate surroundings. Guards were posted with considerable authority to stop people from entering or passing through the municipality. On September 11, 1853, the Möja chapel Assembly’s Health Board in the province of Uppland decided that “each individual is to watch over his bridge. In case somebody tries to use violence, this is to be met with violence. It is the disobedient person’s own fault if he loses his life and no penalties should ensue” (ULMA 28970). This statement refers back (almost word for word) to paragraph 94 in the Royal Regulation of July 9, 1831. Specific carriers of the sick and dead could be designated by the Health Boards. Coffins were to be built and preparations made for the establishment of cholera cemeteries.

District Physicians’ Efforts The District Physicians’ annual reports which are available on the net (http:// www.ep.liu.se/databas/medhist.sv.asp) illustrate what was happening in the countryside as seen by the physicians from 1850s onwards. There were physicians in the towns from 1834 with responsibility for the surrounding countryside. As early as in 1831 physicians might visit municipalities in the countryside to prepare for a threatening cholera epidemic. This was done by the Uddevalla physician Fredric Marin (1772–1834) who met with the newly established Health Boards on the islands of Orust and Tjörn in order to describe the disease and how to treat the diseased. He mentioned smoke ster- ilization of the localities and hot baths in designated bath tubs. He carefully described the medications which were thought to be effective against the cholera. They were to be purchased ahead of time in an Uddevalla pharmacy and distributed for free to the poor. At that time, there were no pharmacies in the countryside. A cottage hospital was to be established in each parish. When Marin visited the island of Klädesholmen on Tjörn on August 13, 1834, the cholera was raging and the medications had been used up. Marin ordered new medications to be sent immediately from Uddevalla (Pettersson 1989:103f). During the cholera outbreak on the islands of Orust and Tjörn the regional authorities sent three physicians in training to assist the town physician Marin in Uddevalla (Jacobsson 1989, Gullman 2003). Nineteenth-Century Cholera Epidemics in Sweden 123 There was a great shortage of physicians and they were liable to succumb to cholera and die themselves. In the Strömstad district physicians’ area, which included the whole of North Bohuslän, the mayor Johannes Hegardt sent several requests during 1834 to the governor in Gothenburg to get more physicians, but with no effect. The situation became extremely difficult when the district physician Simon Landeberg died during a visit to a farm in the Brastad parish close to the village of Lysekil (Andersson 1998:17). There are other cases in Sweden of district physicians dying when fighting cholera. During the next cholera outbreak in 1850–1851, two young physicians were stationed on Tjörn which had no permanent district physician until 1892. The first district physician was stationed on Orust in 1836 witha responsibility for Tjörn as well. Sven Kellberg was district physician 1850–1853. In his annual report for 1851 he noted that the cholera hit Tjörn, but not Orust. Roughly one third of the 200 diseased died. The Health Boards in the different parishes had closed the ferry line at Skåpesund, that being the border between Orust and Tjörn. Doctor Kellberg was not at all convinced that this would have stopped the spread of the infection to Orust. He writes that the Health Boards “activities often focussed on the useless and often ridiculous barriers, and here, like in many places elsewhere, they would not listen to the physician’s advice”. Living on Orust and being a physician, he could use the ferry to visit “the cholera patients on Tjörn and be in their close proximity and touching them . . . without any negative effects or spread of the disease”. Touching was not a problem if the physician believed in the miasma theory, that the disease was spread through the air. Kellberg’s report also shows that large quantities of cholera medications were purchased by the Health Board on the advice of the pharmacist who lived on Orust at the time. Doctor Kellberg’s arguments raises the question of how physicians of the time regarded the spread of diseases. Annelie Drakman’s doctoral thesis in history of ideas, Uppsala 2018, has made a careful analysis of Swedish physicians’ view of the human body and diseases (Drakman 2018). The cholera bacterium had not yet been identified, that happened as late as in 1883 through the physician Robert Koch (1843–1910). Instead, many physi cians regarded cholera as being spread through the air through damaging vapours emanating from dead bodies and “unhealthy places”. The disease was not spread person-to-person. Smoke from burning tar and juniper were thought to be helpful against airborne contagion. This is the miasmatic opinion. It recurs quite often in folklife records. The putrefactive matter was called “miasma” and could be identified through its horrible stench. Local sources of contagion could be stacks of refuse or manure, as well as swamps. Physicians advocated ditching of such areas. From the middle of the nineteenth century, the miasma theory was being replaced by the view of contagious spread of diseases. It emphasized 124 Anders Gustavsson contacts between people rather than airborne contagion. This is visible in local physicians’ statements. In 1849, the British physician John Snow showed that cholera was spread through water. From there on, cholera was seen as being spread through dismal hygienic conditions, primarily contaminated water, but also overcrowding. Body secretions were regarded as disease carriers. The district physician Niklas Olof Gammelin who served on the islands of Orust and Tjörn in Bohuslän 1853–1863 illustrated how rapidly the disease could be spread through physical contacts.

A skipper’s wife in Rågårdsvik with cholera... had a newborn baby. She did not dare to give her breast to the child for fear of contagion and asked an old woman to suckle her. This happened in the evening, and the old woman turned ill during the night and died in the morning, 12 hours after suckling.

Dirty streams close to sewers and manure stacks were regarded as key sources of contagion, particularly in densely populated areas. Because of this, cleanliness and disinfection came to be highly valued, at least by physicians and Health Boards. Washing and bathing became important means to remove dangerous secretions. The annual reports from district physicians show how people in the countryside offered a fair amount of resistance. They wanted to stay with well-established traditions regarding how to live and act together. During the latter part of the nineteenth century, physicians complained that overcrowding was self-inflicted during the winter by many people. Everybody lived in the kitchen and in the common room to save on heat. The district physician Uno Helleday described conditions on Orust and Tjörn:

Unhealthy habits and conditions. In family homes, even wealthy homes, only one room was heated during the winter, and this offers insufficient space for the large family and their servants. It is not possible to separate the diseased however necessary that might be. An epidemic during the cold part of the year makes this situation even worse, since it is virtually impossible to get fresh air, the windows being nailed shut.

The district physicians faced a need to spend a lot of time to build trust among the local population. They were steeped in a tradition of using uned- ucated local healers (Gustavsson 2017). These healers were not expected to treat cholera, however. In this field there was no controversy between physicians and local healers. Many people who lived in the countryside were also reluctant to seek a doctor for economic reasons as well as a high respect for educated persons and the officialdom of the town.

The Spread of Cholera to Sweden and the First Cases The medical historian Sven-Ove Arvidsson, who has been mentioned before, studied the global spread and distribution during the various epidemics Nineteenth-Century Cholera Epidemics in Sweden 125 during the nineteenth century. He has used government regulations, Health Board minutes, the annual reports from district physicians, and the cholera physicians’ journals. Cholera originated in India and after 1817 it spread across the borders. It came by sea from the east through Russia in 1830, arriving in the Nordic countries in 1831. Finland had an outbreak in 1831, and Norway in 1832, 1833 and 1834. Coastal towns were the first to be hit through arriving vessels. In Norway, Drammen was hit in 1832 and then the areas surrounding the Oslo fjord in 1833 and 1834 (Alver et al. 2013:53).

Fig. 1. The cholera epidemic 1826–1837 (after Arvidsson 1972:16).

Sweden escaped any outbreaks until 1834, but the worry made itself known amongst the leadership as early as in 1831. This is clear from the already mentioned royal regulations regarding cholera epidemics. The very first case hit Gothenburg through the sailor Anders Rydberg (52 years old) and his wife Anna Persdotter (55). He worked in a shipyard in Gothenburg which repaired open-sea vessels. The course of development was very rapid and painful through diarrhoea and cramps. The post-mortem report regarding this couple says:

Last Saturday, on June 26, at 5 o’clock in the morning, the man left his home in good health after having had a breakfast consisting of coffee and two shots of liquor. At work, he was hit by vomiting and diarrhoea, as well as leg cramps, between 7 and 8 o’clock. He became very weak and was carried to his home at 9 o’clock. The stomach pains ceased but he became muddle-headed and died at 3.30 in the afternoon. The wife had been in town in the morning without having anything to eat. She 126 Anders Gustavsson had set the table with food when her husband was carried home. She was terrified at the sight and fainted. She vomited immediately when trying to drink. She did not have any diarrhoea. The vomiting ceased in the afternoon. She was fully conscious until her death Sunday morning at 5 o’clock. Dr Lübeck cared for the diseased and prescribed medications. The couple were both known as sober and well-behaved individuals in spite of their poor conditions (Öberg 1988:66).

Many folklife records mention the rapid course and great pains through vomiting, cramps, and diarrhoea. In some cases the informants had witnessed the disease close-up. A man born in 1845 in the parish of Upphärad in the province of Västergötland told (in 1929) how his sister died of cholera in the 1860s.

She turned like a worm because of the cramps. They gave her massage for the cramps. “I’ll soon be well,” she said. But then the cramps hit her intestines – and then she died within thirty minutes (IFGH 1627:20).

The cholera spread from Gothenburg along the Göta Älv river. Two seamen died as early as July 31, 1834, in the village of Västerlanda close to the river roughly 50 kilometres north of Gothenburg. Soon thereafter, the cholera spread epidemically in the countryside parishes Västerlanda and Hjärtum. It spread further along the lake Vänern coasts. The epidemic spread along the waterways by way of the Göta Canal to the lakes Vättern, Hjälmaren, and Mälaren. The town of Vadstena was hit heavily, also by sea transport activities. In the Bohuslän coastal region, the fishing villages were hit hardest. Ninety out of 223 inhabitants died on the island of Gullholmen, 103 out of 535 inhabitants in the village of Grundsund, and one third of the around 400 inhabitants in the village of Klädesholmen on Tjörn. A total of 24,978 persons were taken ill in Sweden in 1834, roughly half of whom, 12,637, died. One third, 4,659, came from the Gothenburg and Bohuslän region. The city of Stockholm had 3,665 deaths (Arvidsson 1972:166). The next wave of cholera came from the east to Finland in 1848. The epidemic of 1849 hit the towns of Borgå and Helsingfors particularly hard. In 1850, Sweden was hit by another epidemic in early August in the town of Malmö. Fishermen spread the contagion, and the fishing village of Råå was hit particularly hard. The first cholera cases appeared in Gothenburg in late September. A sailor was taken ill on a vessel. The contagion was spread along the waterways to both the archipelago and along the Göta Älv river to villages along the lake Vänern coast. The following wave came to Finland from the Russian city of Petersburg in 1853. Denmark was hit from June 11 in Copenhagen. A total of 10,598 persons were taken ill, of whom 6,588 died. Norway also suffered the largest cholera epidemic in 1853 (Bente Alver et al. 2013:53). Towards the end of July 1853, cholera epidemics hit several coastal Nineteenth-Century Cholera Epidemics in Sweden 127 towns in the southern Skåne region of Sweden as well as in the archipelago of the Blekinge region. A quarter of all inhabitants in the fishing village of Borstahusen close to the town of Landskrona in Skåne were taken ill, and one in thirteen died. Many middle-aged married women were infected when they cared for the sick who were placed in the school building according to an informant who was born in 1835 (LUF M 4824:4). In August, both Stockholm and Gothenburg were hit. The disease was spread from Stockholm along the waterways of lake Mälaren. A number of towns and villages in the Bohuslän archipelago were infected from Gothenburg, particularly the fishing village of Edshultshall (more about this later) and further along both sides of Göta Älv river. The contagion then spread further along the lake Vänern coasts like during earlier epidemics. A total of 17,327 persons were taken ill in Sweden in 1853, and 8,775 died. Stockholm had the largest share with 2,875 deaths. In 1854, a new wave of cholera hit Stockholm as well as the island of Gotland in the Baltic Sea, this time coming from the Finnish archipelago of Åland. The disease spread from Stockholm along the northern coast. This epidemic was, however, clearly smaller than that of 1853. Of 3,038 people who fell ill, 1,212 died. Almost half of that number, 560 persons, came from Stockholm. In 1855, the cholera started towards the end of July, and spread widely across the town of Halmstad. Stockholm was declared infected on August 11. In 1856 a minor epidemic broke out in Sweden, with 278 deaths, 98 of which were in Stockholm, and 58 in Gothenburg. The cholera was spread by arriving vessels in Gothenburg and then by way of waterways along the Göta Älv river towards the province of Värmland. In 1934, a man born in Borgvik in 1839 still remembered how a ship came from Gothenburg to the harbour of Borgvik on the coast of lake Vänern.

It came loaded with rye and was to take on iron going back. It showed a black flag, and we understood that it carried a contagious disease. The manager at the mill, however, needed both the rye and transportation for the iron, so he was rowed out to the ship to get an explanation for the black flag. It was the cholera, he was told, but he didn’t care, and ordered the ship to unload. The ship hadn’t even been unloaded before the first deaths were reported, and the disease spread rapidly. Three physicians came to Borgvik, I remember that one of them came from Gothenburg. My uncle was one of the first to be infected, and his wife soon followed. I had to care for both of them as well as the animals in the barn. I didn’t need to bother about food because the diseased were not allowed to eat anything. … Every ten minutes, I was to give them a spoonful of water. … I was sixteen when this happened (IFGH 3401:7ff).

In 1857 a new epidemic started in Stockholm. On September 11, Stockholm was officially declared to be infected by cholera. From there, the disease was 128 Anders Gustavsson spread by steam ships and sailing ships to the coasts along lake Mälaren, the Vättern and Vänern regions, and to Gothenburg by way of Göta Älv, as well as to the province of Blekinge. The epidemic of 1858 was, by and large, limited to Stockholm, where 619 persons died, as well as 88 persons outside of Stockholm. In the fishing village of Grundsund in the province of Bohuslän, 24 persons died during October and November. In 1859, cholera came to five Swedish import harbours, i.e. Malmö, Gothenburg, Kalmar, Stockholm, and Uddevalla. In Gothenburg, the mortality was high in the labour quarters of Haga. Sixty-one people fell ill in the town Uddevalla, and 39 died. It was not until 1866 that cholera appeared next time in Sweden. The other Nordic countries escaped entirely. As in earlier epidemics, the harbours in Stockholm and Gothenburg were the ports of entry, the first cases appearing late June. There were a large number of cases in Gothenburg, with 638 deaths. In the archipelago of Bohuslän, the islands of Orust (the Gullholmen village), Tjörn (the Klädesholmen and Rönnäng villages), and Styrsö were hit particularly hard. The cholera also spread along the Göta Älv river, reaching the town of Karlstad on July 18. The contagion spread from Stockholm along the coast to the north as well as inland along waterways such as the Göta channel. A total of 4,706 persons died in 1866. The province of Gothenburg and Bohuslän was hit hardest, with 1,330 deaths. The last epidemic in Sweden in 1873 was limited to the province of Skåne, starting in the town of Helsingborg on July 19. Soon thereafter, the cholera spread to the town of Höganäs, some 20 kilometres away, where 356 persons fell ill and 177 died. An informant, born in Höganäs in 1859, remembers this period when he was fourteen years old and just confirmed in church. When he came back home after his first job as a sailor, he was met by a horrifying experience:

Twenty-five to thirty persons a day had died at the peak period. The wells were dry, and people had to get their water from the dirty mill channel. … The disease terrified the inhabitants. Homes were abandoned, people preferring to live in hovels. They were not allowed to bury their dead, because people were not allowed to gather in crowds. When the cholera raged, no birds were visible, and if you did see any, they did not sing (LUF M 7461:16).

After 1873, there were no cases until 1893 and 1894, and then only single, imported cases from overseas in Stockholm and Gothenburg. Related to the total Swedish population, the cholera mortality was clearly highest during the 1834 and 1853 epidemics with 425 and 239 deaths respectively per 100,000 inhabitants. In 1866 the corresponding numbers were 113 and 74 deaths respectively per 100,000 inhabitants. All ages were Nineteenth-Century Cholera Epidemics in Sweden 129 hit, from babies to old people, as can be seen in the death and burial books. As a rule, the contagion lasted just over a month and never more than two months. The severe outbreak on Gullholmen in 1866, when 36 persons died, lasted from August 7 until September 13. Summing this up, the various epidemics share the pattern that the infection has been spread across the oceans by sea vessels and then along the waterways and along the coasts of larger lakes. Big coastal towns were the hardest hit, while the interior fared better. Small coastal municipalities were hit harder than the interior. This is obvious along the Bohuslän coast (more below). In 1832, shipping outside the coastal towns was restricted through the regulation on farm transport activities. Establishing this regulation in new towns along the coast took time, however. Fishing was Fig. 2. Diagram showing the number of deaths carried out coastally until open sea in Gothenburg in 1834. A sud-den peak with fishing was developed in the 1860s. at least 80 deaths per day lasted for 11 days in August. The entire cholera outbreak, from the Until then, contacts with an outside first to the last death, lasted from July 30 until world regarding fishing were limited. Sep-tember 5. After Öberg 1988.

Barriers between Towns and the Countryside When a cholera epidemic struck, issues concerning nearby villages were raised immediately. This is consonant with the royal regulations of 1831. During the 1834 epidemic, the countryside’s trading contacts with nearby towns was essential since general stores had not yet been established outside the towns. The situation had changed, however, in the epidemics during the 1850 and 1860 decades. Free trade regulations had been established in 1846, permitting the establishment of stores in the countryside (Ejdestam 1943), and many country stores opened rapidly. Annually recurrent markets were still restricted to the towns and would be visited by large numbers of countryside people, not least the market in Gothenburg on August 10 (Skarin-Frykman 1993). This could become a dangerous source of contam- ination as was stated in recorded material (such as IFGH 3860:33) There are several stories showing the care with which town and country inhabitants were kept apart during 1834 by guards in the town outskirts. 130 Anders Gustavsson The town of Uddevalla had endured a severe cholera epidemic in the same year, but during the cholera outbreak in Gothenburg and southern Bohuslän in 1850, a tight barrier system was established, using soldiers and officers. That helped Uddevalla avoid any disease cases, a very different situation compared to 572 diseased and 270 deaths in Bohuslän up to November 8 (http://uddevallare.blogspot.com/search?q=kolera). In 1925, a man born in 1846 in the village of Ljungskile some 20 kilometres south of Uddevalla recounted his childhood memories from the 1850s:

My parents were going to Uddevalla and had received a passport from our helpful local chief of police, and I was taken along in the vehicle. When we arrived at the hill Kapellbacken close to the town we met a large troop of officers and soldiers and a big boom was lowered and an officer ordered us to stop. Show your passport. A big corporal stuck a long-shafted pair of pliers into the wagon to take the passport which he carried to a burning fire on which another soldier poured tar. The letter was duly smoked and then given to a high officer to be carefully read, after which he shouted to the soldiers to raise the boom and told my father Pass! (VFF 1109:6).

Farmers could leave their produce at custom ports in the town outskirts. Merchants needed to go there to pick it up in exchange for wares needed by the farmers. The farmers were not allowed to cross this border, although stories were traded about those who were able to when, for example, the guards fell asleep (ULMA 2293:18). That may have been caused by the large amounts of Scandinavian vodka (brännvin) which was supposed to protect the guards against the cholera. Brännvin, particularly spiced with wormwood, was given as a protective before anyone was allowed to pass the guard entering or leaving the town. “At the customs port in Kristinehamn, a guard was placed with brännvin. Everybody who entered or left were obliged to swallow some brännvin”, according to a story told by an infor mant born in 1866 in the village of Varnum in the province of Värmland (IFGH 4325:19). The ambition to keep town and countryside people apart is illustrated in a letter sent from Gothenburg to the farmer Lars Andersson on the farm Naveröd in the Röra parish on the island of Orust on September 8, 1866. His brother Nathanael informed him that:

Here in Gothenburg, the cholera is raging hard – it was said today that in the township of Haga 22 persons died in one day, the day before yesterday, I believe. Yesterday I heard that the schools are closed until further notice, or until the epidemic has eased.

For the last two weeks, the writer had planned to come home to Naveröd to eat some of the apples then available in the orchard, “but as you may fear me because I may carry the disease, I will probably stay in Gothenburg”. He also heard “that the unpleasant cholera plague has also harried the islands of Orust as well as Tjörn and Klädesholmen since some time. I therefore wish Nineteenth-Century Cholera Epidemics in Sweden 131 to inform you about how to behave during the cholera”. He had therefore sent a copy of instructions published by the Swedish Medical Association and printed by the Gothenburg Health Board. These instructions were kept among the papers left by Lars Andersson, and were titled “Instructions to the public regarding the symptoms of the cholera disease, and how to protect the healthy, and what medications should be used for the disease until a physician has arrived or when a physician is unavailable”. The instruction gave much detailed advice. Cleanliness is of the essence, “no less of body and clothing than of living quarters, vestibules, yards, and the close surroundings and places”. It “is one of the best means in preventing and curing the disease”. Floors of rooms where there are diseased should be covered with freshly chopped fir branches sprinkled with vinegar” (privately owned letter).

Barriers in the Countryside The countryside guards were careful in checking everybody arriving from neighbouring parishes. A man born in 1866 in the village Levene in the province of Västergötland related what he had heard had happened to his grandmother when a boom had been erected across the road. Somebody asked: “Have there been any travellers today?” The answer was: “Yes, if the Good Lord Himself were to come by, do not let Him pass. He must be smoked!” (ULMA Bd 01352). Smoking with tar or juniper was common in towns as well as in the countryside. Smoking was done because the cholera contagion was assumed to be carried by air, in other words a miasmatic view, as related in folklife records. It was said that you could hear the air move when the cholera was approaching. This was claimed in a recording made in 1930 from the island of Tjurkö in the archipelago of the province of Blekinge as the cholera arrived from the nearby town of Karlskrona (ULMA 2743:13). A woman born in 1858 in the village of Veinge in the province of Halland recounted in 1925 that the cholera “once flew 140 kilometres in one day” (VFF 1244:13). A woman born in 1852 in the parish of Ransäter in the province of Värmland stated in 1937 that “the air was grey from all the cholera” (IFGH 4050:38). There are several stories telling that pieces of pork were strung up to ascertain whether the cholera had arrived or was still present in the village. Many believed that the pork turned black if the cholera was in the air. Meeting a black man or woman was sometimes seen as an auspice of a coming cholera outbreak (Tallerud 1999:144f). It was also said that the birds ceased singing (as cited earlier from Höganäs 1873), which was interpreted as a sign that the air was contaminated. Along with the perceived importance of smoking, there was a broad popular perception that brännvin gave protection against cholera. “So they 132 Anders Gustavsson drank copious amounts of brännvin. They drank until they were totally drunk because they believed that the contagion then would not affect them”, a woman from the village of Råggärd in the province of Dalsland stated (ULMA 18874). According to several recordings, this had caused an over- consumption of alcohol during the cholera epidemics. It was said to be particularly true for the corpse carriers who brought the dead to the newly established cholera cemeteries (more to follow below). There are many confirmations of the fact that coastal villages which were focused on sea transportation and fishing were harder hit by cholera than the inland areas. This was particularly obvious for the village of Klädesholmen which lacked a land connection, compared to the parish of Stenkyrka further inland. In 1834, Klädesholmen was hit particularly hard because of people who were coming home from Gothenburg. The Health Board prescribed five days in quarantine for the diseased. Guards were placed around houses where there were diseased persons. Special black cholera flags had to be put in place in order to keep people away. Anybody who cheated in order to avoid the quarantine would pay fines. The people living on Klädesholmen were not permitted to visit the mainland or other islands. Food purchases were to take place along a jetty in Stenvik some distance from Klädesholmen, where suppliers and buyers had to deal outdoors. Guards were put in place to stop people from Klädesholmen from reaching the mainland. All men between 18 and 50 years of age were obliged to participate as guards (Pettersson 1979:102ff). The folklife artist Carl Gustaf Bernhardson (1915–1998) lived in the border zone between the coastal village of Grundsund on the island of Skaftö and the countryside further inland (Gustavsson 2011). He has created

Fig. 3. Carl Gustaf Bern- hardson’s painting shows how a fence prevented contacts between fishermen’s wives in Grundsund and farmers’ wives from the countryside during the 1834 cholera epidemic. Bernhardson 1978:70. Nineteenth-Century Cholera Epidemics in Sweden 133 paintings and recorded old stories for the Folklore Archives. One painting (1834) shows a fence in the outskirts of Grundsund where the fishermen’s wives placed their empty milk containers. Later, the farmers’ wives filled these containers with milk which could be picked up by the fishermen’s wives. This prevented personal meetings between the two groups, which was otherwise an established pattern well into the twentieth century. Each fisherman’s wife had her own allotted farm, which she would visit two or three times a week to pick up milk. Bernhardson named his painting Bôttre (a dialectical word for tub), and that became the name of this fence. It lived on well into the twentieth century, which has been confirmed by older people living in Grundsund. The fishing village of Edshultshall on western Orust belonging to Morlanda parish was hit severely by the 1853 epidemic. Twenty-three persons died in September, or about one third of the total population of 70. They were buried in two groups on the Sundays September 18 and 25 on the uninhabited island of Bråtö nearby (more about this later). They were transported to the island by a fisherman and his thirteen-year-old son. In three houses, everybody died. The Health Board in the Morlanda parish was activated and met weekly to report on the state of health. Edshultshall was cordoned off for six weeks after the acute cholera outbreak. In his annual report, district physician Olof Niklas Gammelin (1815–1867) stated that he visited the village four times during the month of September, and once in early October. He made sure that the village would be supplied with fresh food from the surroundings. The Edshultshall fisherman Alfred Andersson (1862–1949) described how farmers left provisions on the peninsula Anstensholmen close to the adjacent village of Hälleviksstrand. People in Edshultshall were allowed to go there by boat, the farmers having already left (Jacobsson 1989:155ff).

The Social Situation of the Diseased and Dead Investigations of cholera epidemics often point out that the disease hit particularly hard in socially deprived areas which suffered from poverty, bad hygiene, and crowdedness. Such was the case in the Gothenburg workers’ district Haga. The same pattern could be seen in the countryside. Barracks for contract workers, mill workers, and sawmill workers were a dangerous environment. During the epidemic in the province of Sörmland, contract workers with the Sörby manor in the parish of Torsåker were the only ones to be infected (Arvidsson 1972:14). Amongst the cholera-stricken in the inland farming parish of Stenkyrka on Tjörn, almost all dead belonged to the poor population, paupers and crofters (the only difference between the groups being that the former did not lease farming land) (Bosson 2012). 134 Anders Gustavsson

Fig. 4. Deceased persons living in Uddevalla during 1834. All social categories are present, from paupers to the manager of the sea side resort Gustafsberg. The grave-digger comes last, the one who will carry them all to their grave. Bohuslän Museum, Udde- valla UM 000332. https:// digitaltmuseum.se/ 011044500895/aflidna- uddevallabor-silhuett.

Regional Differences in the Countryside It has been shown that the cholera contagion reached the large towns first, and that waterways were an important factor in the spread of the disease. This was true also in the further spread to smaller towns and villages along inland waterways and large lakes. Regional differences are also visible in the countryside between coastal villages and the adjacent inland. This is of interest particularly regarding the archipelago in the province of Bohuslän. The differences are easily seen on the two islands of Orust and Tjörn. The separation in 1834 of the small island of Klädesholmen from the Tjörn hinterland has already been described. Roughly one third of the island’s 400 inhabitants died, or 135 individuals. A total of 165 persons died on Tjörn. This demonstrates the high excess mortality on Klädesholmen. The same differences between coastal and inland areas were visible on the island of Orust including the island of Skaftö in 1834. The coastal villages were situated in the western part of Orust in the Morlanda parish. The death Fig. 5. Map, showing Morlanda parish until toll was 103 died in the village of 1923. The various cholera cemeteries are in- Grundsund, 90 on Gullholmen dicated by a + sign, from the south Mollösund, Bråtö, Vedholmen, Kårehogs Hage and Åse- island, 78 in Fiskebäckskil, and 77 backe. After Jacobsson 1989:142. in Mollösund. A mere six persons Nineteenth-Century Cholera Epidemics in Sweden 135 died in the interior of the same perish. In the parish of Myckleby in the eastern part of Orust, where there are no coastal villages, only two persons died, one crofter and his nine-year-old daughter who had travelled by boat to Gothenburg. A report from the Board of Health stated that “it is likely that a wise barrier having been made easy by the patients’ isolated home on the eastern shore of the island, had contributed to the lack of a further spread of the disease” (Jacobsson 1989:153). Twenty-eight people died in the primarily farming parish of Tanum in the northern part of the Bohuslän province, 25 of whom lived in the fishing village of Grebbestad (Andersson 1998:14). The same differences between coastal and inland areas on Orust and in Tanum parish were pointed out in 1843 by the district physician Niklas Olof Gammelin in his 1858 annual report. At the same time, he pointed out that it “would be difficult to explain the underlying cause”. The 1834 and 1859 patterns on Orust and Tjörn were repeated during the last outbreak in 1866. On the island of Orust, only the fishing village of Gullholmen was hit, with 36 dead within five weeks. In the inland parish Röra where the farmer and diary keeper Jakob Jonsson (1795–1879) lived, only one person died, a 72-year-old crofter’s wife. Jonsson noted that the infection was carried to Röra by a person who came from Gullholmen. In early September, Jonsson expressed “a double sadness, the raging epidemic so widely spread that all com- merce between towns, particularly Gothenburg, and the archipelago had to be almost totally closed – and totally so in some areas”. Jonsson regarded the situation with worry and a deep concern about the consequences for the coastal popula- tions outside of Röra. He wrote: “It would appear that all these poor or pitiable coastal people will accompany each other into eternity at the same time – we can only wait and see for how long angel of death is allowed to rage and how widely he has been told to kill”. The angel of death is a concept with roots in the Old Testament’s stories about the Fig. 6. A wooden cross from the island of Israeli exodus from Egypt (www. Gullholmen in memory of the widow Inger ne.se Mordängel). Johansson, born 1816, and her son Carl Johan According to Jonsson, there was a Andersson, born 1849. They died of cholera on the same day, September 4, 1866. A winged religious path out of the difficulties. angel holds a cloth with the names and life data “A serious penance and petition for of the deceased. Photo Berth Kull-holm. 136 Anders Gustavsson mercy might serve as a reason not to punish humanity more.” On October 26 he could observe with relief that: “God be praised, the cholera disease has largely decreased in general, and particularly in towns and fishing villages where it has hit most heavily” (Jakob Jonssons dagbok 1 1991:27–30). The archipelago of the island of Tjörn was hit by its heaviest epidemic in 1866. The death toll was 116 on the island of Klädesholmen and the adjacent coastal village of Rönnäng. Eighty-six minor children lost one or both parents. In the hinterland village of Klövedal, only seven persons died. The larger inland parish of Stenkyrka was harder hit, with 23 deceased. Most of the deceased were poor (Bosson 2012). It should be noted that differences in degrees of outbreaks existed not only between coastal and inland areas, but also between different coastal villages. This is clearly seen in the reports from district physicians in the archipelago of the province of Bohuslän. The fishing village of Klädesholmen on the island of Tjörn was severely hit in 1834, Edshultshall on Orust and Kalvö on Tjörn in 1853, Hälleviksstrand on Orust and Åstol on Tjörn in 1855, Grundsund on Skaftö in 1858 and Gullholmen on Orust and Klädesholmen and Rönnäng in 1866. During the same periods, adjacent coastal villages escaped without any cases of cholera. One such village was Grundsund in 1866.

Fig. 7. Number of deaths per week in the fishing village of Klädesholmen together with the inland parish of Stenkyrka on Tjörn from the end of July to the middle of October 1866. Bosson 2012:8.

Local Experiences The cholera hit local communities very suddenly and many died within a very short time. It is not surprising that this raised feelings of terror. What might one do to protect oneself and survive, and how would one dare to Nineteenth-Century Cholera Epidemics in Sweden 137 care for the diseased? There are some stories told by people who lived with and survived cholera (compare the quotation from the village of Borgvik in the province of Värmland). The girl Anna Holmberg was fourteen years old during the great cholera outbreak on the island of Klädesholmen on Tjörn in 1834 when 68 persons died during the month of August (cf. above). She has left a devastating tale of her memories:

I remember that time with fright. People slunk about with the fear of death on their faces, scared of each other, and even more scared by death. I remember oldsters, whom you met one day, and the next day were lying on a stretcher. First, they complained about stomach pains, after a few minutes they had to lie down and screamed from their stomach cramps which then made them black in the face, and finally took their lives away from them. As soon as parents were taken ill, they would send the children out of the house. You could see these children watch fearfully through the windows. It was heart-breaking to see them stand there crying, and to know that they might lose their father or mother within the hour. And in the general chaos, nobody would care for these children. As soon as somebody showed signs of dying, they would be carried off to a barn that served as a mortuary. The old men who did the carrying of the dead faltered around dead drunk. They did not care very much about checking whether they were carrying a dead body or somebody in suspended animation. Every forenoon at eleven o’clock the church bells would toll for a long time with short breaks. They tolled the knell (Pettersson 1978:15f).

The recorded material contains stories of how the fear of contagion made people in towns and in the countryside lock themselves in their homes or outbuild- ings. The fear of touching others (reflecting a contagious view of the disease) was fundamental. Food was taken in with the help of somebody who would pass the food through a small hole. It is said that the same procedure was applied in the cholera houses which were established in towns and villages when they were infected by cholera. The folklife artist Carl Gustaf Bernhardson has visu alized this situation in a painting which was based on stories told in his own family (Brockman 2010: 175).

Fig. 8. The cholera year 1834. An infected person, who was isolated in a designated house, received food offered on a long rod through a small opening so as to avoid any physical contact with the outside. Folklife painting by Carl Gustaf Bernhardson. Bohuslän Museum, Uddevalla CGB 087. 138 Anders Gustavsson Several recordings tell how the people who were the most fearful and closeted themselves could still be infected. At times, they were among the first to die in the community. One informant, born in 1863 in the town of Filipstad, said:

There was one person who built a long tunnel leading to his home through which he drew the food. But this man was the first to die of cholera here in town (IFGH 4334:16).

In his book “My Childhood on Gullholmen” the author Olof Hansson, born in 1914, relates stories told by his grandparents. They were around thirteen at the severe outbreak of cholera in 1866. The contagion was said to have been carried by people from Gullholmen who had visited the market in Gothenburg on St. Lawrence’s day on August 10 (see Skarin Frykman 1993). This market offered an important opportunity to sell fish (Hasslöf 1949:396). Women living on Gullholmen were known to travel to Gothenburg by boat in the summer in order to sell the fish their husbands had brought in. Children obviously suffered, physically as well as mentally, when relatives were taken ill, often dying. Olof Hansson related how his grandparents: talked about how horrible life was, and about the fear felt by everybody. In the morning, no one knew what had happened during the night. There were diseased in almost every house, and 36 persons had lost their lives. Many panicked and in one house a woman locked herself in, for fear of being infected. The children who had been locked out, among them my grandmother, had heard her scream in anguish. And yet, she was one of the first to die (Hansson 1983:45).

A kind of belief in destiny is recorded at times. Those who were to die from cholera would die regardless. That made it senseless to lock oneself in. An informant in the village of Källby in the province of Västergötland told about a man in the town of Lidköping who nailed the doors to his house shut – and still died. The informant’s comment was: “Those who were to die from the disease could not escape” (IFGH 4143:37). Old folk belief could show how people were to protect themselves without having to lock themselves in. The man from the village of Borgvik in the province of Värmland referred to earlier, helped his uncle when he was suffering from cholera without being infected himself. His reflections were:

I believe that the reason I was not infected myself was that as soon as the cholera broke out, I went to the church and knocked on the door. I had been told that this would give me protection, and it does seem to have helped me (IFGH 3401:12).

Faith in God might have contributed when somebody dared to care for the diseased instead of locking oneself in. In her book “My Childhood on Tjörn”, the author Hulda Tjörne tells about the cholera in her parish Valla in 1834: Nineteenth-Century Cholera Epidemics in Sweden 139 A man, who suffered from another disease, said: “I visit the diseased in God’s name, where I find them”, and so he did. He did not catch cholera but was cured from the other disease (Tjörne 1970:119).

There are very few Christian references in the folklife records. There is only one informant, born in 1862 in the coastal village of Fjällbacka in the province of Bohuslän who has mentioned a belief in God’s punishment related to the cholera outbreak in 1859 in the village. He also spoke of “much singing of psalms, because people believed that they would escape the cholera if they turned into believers” (IFGH 4861:38f).

Handling the Dead It has already been shown that the cholera had a short and painful course with a high mortality. Many died within a very short time when a municipality was infected by this disease. Because there was a great deal of fear that the diseased person would still be contagious, the dead had to be removed to a specially designated cemetery, rather than the usual cemetery. At the peak of the disease many would die each day. Simple coffins had to be prepared very rapidly. In 1926, an informant born in 1851 in the village of Hammarö in the province of Värmland said that “there were no real coffins, they simply made unpainted boxes in which to put the dead” (IFGH 776:7). Transporting the dead was not a popular task. The Health Boards had the authority to pick people for the job. It was important to make it attractive, the risk of contagion being very high. One informant, born in 1862 in the village of Fjällbacka in the province of Bohuslän, which was severely hit by cholera in 1859, said that “everybody feared having to participate in the removal of the dead”. There was one, however, Jan, who accompanied all the transports to the lonely island of Stora Eneskär. “He would sing psalms all the way to Eneskär and by the grave”. “The bodies were covered with soil as much as possible”. The informant also said that “there are not many in Fjällbacka but they have relatives lying there” (IFGH 4861:38f). Brännvin (vodka) helped enrol carriers of the bodies. As has already been said, it was regarded as a protection against contagion. Many folklife records state that the corpse carriers were always drunk and often singing loudly when they set out with several bodies on the same wagon. It usually took place during the night. The body carrier might be the last to die during the cholera outbreak. One informant, born in 1850 in the village of Skepplanda in the province of Västergötland, related the following story which he had heard from his father:

The man who took the bodies to the cemetery on a simple cart drank and was constantly drunk. He survived until the autumn, and died just as the epidemic ended (IFGH 3387:20). 140 Anders Gustavsson When the epidemic had ceased in the village, the cholera wagon was left standing in the cemetery to rot (IFGH 4064:34, IFGH 4390:16). On the islands the dead were taken away by boat to uninhabited islands. In 1834, the island of Malmön was hit very hard by the cholera.

An old man brought the dead out on the water in a boat to prevent the contagion from spreading. When asked how he had been able to stay well, he responded: “I was drunk the whole time” (IFGH 4282:40).

The burial ceremonies became very strange when many bodies had to be buried at the same time. The informant quoted above, born in 1839 in the village of Borgvik in the province of Värmland, related about such a burial which he took place at the age of sixteen.

I remember one day when 32 bodies were buried. They had dug a big, common grave for all the coffins. Before filling the grave, they put a rod on each coffin, long enough to be seen on the surface. They put a tag on each rod that showed the name of the deceased. Our own priest was ill with cholera himself. So a strange priest buried the dead. He went from rod to rod, pulled it out and poured the three shovels of earth through the hole. A man followed him, filling the hole with chlorinated lime (IFGH 3401:9).

Very often it is said that the dead were sprinkled with lime as a sort of protection. This agreed with the royal regulation of July 9, 1831, paragraph 48, concerning cholera burials. “When a number of bodies are to be buried at the same time, a large grave should be prepared and a communal burial service should be held. The deposited coffins are to be covered with layer of unslaked lime or coal dust.” A painting by Carl Gustaf Bernhardson shows how the priest buried several bodies at the same time on an out-of-the-way cholera cemetery on the island of Skaftö. The four corpse bearers would throw themselves into the sea, clothes and all, as a protection against the contagion (Brockman 2010:176).

Fig. 9. The priest buries several bodies at the same time in an out-of-the-way cholera cemetery on the island of Skaftö. The four corpse carriers who were to throw themselves into the sea, clothes and all, are shown to the right in the painting. Painting by Carl Gustaf Bernhardson, Gothenburg City Museum GM 24869. Nineteenth-Century Cholera Epidemics in Sweden 141 Physicians as well as Health Boards warned against crowds at the burials. Those warnings were not always heeded. The district physician G. Varenius in the district of Alingsås related how the cholera was spread in the parish of Skallsjö to the east of Gothenburg in 1853. “In Svensered (part of the parish) the disease raged particularly inten- sively after a burial where the prohibition against crowds was neglected.”

Cholera Cemeteries When the cholera broke out in a municipality and many died within a short time period, the bodies had to be brought to a solitary place immediately in order to be buried there. At times, mass graves were used. Nobody wanted to go there, fearing contagion. Later, a stone wall might be erected and these cholera cemeteries were blessed by the local priest. After the cholera epidemics, they would remain deserted. A long time after the cholera outbreak, local community associations have erected a collective memorial stone. This would take place as much as one hundred years after a cholera outbreak. The parish priest would conduct an inauguration ceremony. These memorial stones would keep the memory of the nasty nineteenth-century cholera epidemics alive for later generations in the village. Not all cholera cemeteries have been preserved, however. They did not have the same legal protection as ordinary cemeteries. A record from 1926–1927 in the village of Högestad in the province of Skåne tells how a cholera cemetery was ploughed by a farmer as early as in the late nineteenth century. He put the wooden crosses on top of the fence before ploughing. When the sexton saw this, he reproached the farmer for vandalizing the burial chambers of the deceased, but the farmer just continued ploughing. The informant noted that “nowadays, practically nobody knows that Fig. 10. In 1957, the contract rector Knut Jons- people have been buried there”. A son inaugurated a memorial stone for the far- similar story is told about the cholera mer’s widow Kristina Svensdotter, who was cemetery in the town of Ystad in born in 1800. She was the only person to die of cholera during 1855 in the inland parish of Skåne. In 1942, an informant born Röra on the island of Orust. Photo privately in 1873 related: “It brings sad held. 142 Anders Gustavsson

Fig. 12. The cholera cemetery on the island of Bråtö close to the village of Edshultshall on the island of Orust, where 32 persons were buried during the severe cholera outbreak in 1853 (cf. above). The ceme tery was inaugurated in 1854 and the local association of Morlanda erected the stone in 1956. Photo Fig. 11. The cholera cemetery in the vil- taken in 1978. After Jacobsson 1989:158. lage of Kårehogen in Morlanda parish on the island of Orust was used in 1834, but was not inaugurated until 1934. The memories back to me”. They took gravel Morlanda local association erected a memorial stone with an iron cross. The from that place for various buildings here association still takes care of this cholera in town. When the railway was extended, cemetery. Photo Kristina Gustavsson. they destroyed the entire cemetery” (LUF M 9126:9).

Legends Until now, I have dealt with what happened during the various local cholera epidemics. I will now proceed to stories which might be seen as legends regarding historical events, stories which have been traded over long periods of time. The folklorist Anna Birgitta Rooth stresses that “tales are built upon a reality-related content”. It “is on the borderline between Fig. 13. It is rare to find an individual a story and supposed knowledge … a memorial stone to a person who had died from cholera during the nineteenth story’s mirror image of reality” (Rooth century in an ordinary cemetery. Outside 1978:16f). This is in line with what the the church in the village of Grundsund, German ethnologist Helge Gerndt dis- there is however, a large memorial stone to remember the inn-keeper Abraham cusses in his newly published extensive Larsson, born in 1778. He belonged to study of the character of legends in the the local elite and died together with 95 borderland between fact and fiction other inhabitants in Grundsund in the month of August 1834. Photo Kristina (Gerndt 2020:206ff). The Norwegian Gustavsson. folklorist Brynjulf Alver is of the opinion Nineteenth-Century Cholera Epidemics in Sweden 143 that in historical legends “it is the historical person or event which takes the centre stage”. You cannot expect to find “truth” in historical legends but you can expect to find the general public’s perception of what happened (Brynjulf Alver 1980). Camilla Asplund and Johanna Wassholm have studied historical legends about the 1808–1809 war between Sweden and Russia. The stories were documented one hundred years after the war. They had been recounted within the families and are locally rooted in the places where the informants lived. The study focusses on the subjectively experienced history and how it has been recounted (Asplund Ingemark & Wassholm 2009). The motifs in legend material about cholera often have links to the Black Death in both Sweden and Norway in the middle of the fourteenth century but also to the plague in 1710–1713. I will discuss issues regarding how older popular traditions can be transmitted through generations while also being adjusted to new situations. It might be ideas regarding how dying persons may have been carried away to the cholera cemeteries before being actually dead, and how living children may have been sacrificed to protect against the epidemic. My focus is on collective popular ideas, not historical truth.

Dying People Carried Away to Be Buried A widely held idea which has been recorded in several stories is that certain corpses were not really dead when they were carried off to the cholera cemetery. The question of sham death is raised. No possible cure for the severely diseased was seen and they might infect others around them. The terror was very real (cf above). In that situation the temptation was strong to carry off the diseased as soon as possible, dead or not. The corpse carriers who, it is said, were heavily under the influence of alcohol, were tasked with carrying corpses off as rapidly as possible. Let me remind you of what the fourteen-year-old girl Anna Holmberg wrote about her personal memories of the cholera outbreak on the island of Klädesholmen in 1834. “Nor did they worry about making sure whether the person they carried away was dead or just sham dead” (see the full quotation above). There are also stories about corpses having kicked in their coffin when carried away. In 1936, one informant who was born in 1850 in the village of Stora Lundby in the province of Västergötland told the following story: “Mansa-Petter (born in 1825) was severely ill. And they had already put him in a coffin when he came to. Many were put in a coffin too soon, and some would wake up and start kicking, so they were taken out again” (IFGH 3860:32). In 1937, a woman born in 1848 in the village of Önum in the province of Västergötland told how the village corpse carrier: carried off a young woman who had died from cholera. Being alone, he threw the coffin in the grave, but then the woman came to because she was, like, sham dead, 144 Anders Gustavsson and started screaming. The old man hurried back to the village and related how things were, and a couple of men accompanied him to the grave, but when they opened the coffin she really was dead (IFGH 4060:34).

The fishing village of Grundsund with a population of around 600 was hit by a cholera outbreak in October and November 1858. Forty-six persons were taken ill and 24 died. The priest Nils Lindblom and his wife Emma did not hesitate to visit the diseased and their families, something other inhabitants were reluctant to do. Before the burial, the priest was careful to check that there were no sham dead among them (Gullman 2003:65). The district physician Niklas Olof Gammelin tells an interesting story about this outbreak in 1858.

The only remarkable aspect was the Asphyetic (“not contactable” according to a physician’s handbook in the nineteenth century) 12-hour long condition of a cholera patient. He was considered dead by those around him and a coffin was prepared. The diseased man unexpectedly came to and is still alive.

This story from the district physician about an uncontactable “in between” condition gives a new dimension to popular stories about cholera corpses coming to life, kicking in the coffin or in the grave. Might mistakes have been made unintentionally regarding the actual death? This does not seem impossible. This may have caused the sham death stories found in recordings. It cannot have been easy to make all the required checks when the epidemic was at its peak. The district physician Niklas Olof Gammelin visited the severely hit fishing village of Edshultshall during the cholera outbreak of 1853 (cf. above). He has also recorded that he visited the fishing village of Grundsund seven times during the outbreak from October 15 to November 21 in 1858.

Sacrificing Live Children Another type of tale touches on the motif of sacrifice which is said to have taken place to free the village from epidemics, one of them the cholera. One or two liv- Fig. 14. Niklas Olof Gammelin was the ing children who were free of cholera district physician on the islands of Orust and Tjörn 1853–1863. Photo Carl Cur- would be sacrificed. They were said to man 1860. After Larsson 2015. have been given a sandwich to entice Nineteenth-Century Cholera Epidemics in Sweden 145 them to climb into a prepared grave. As soon as they were there, they would be covered with soil. They would cry, asking why soil was thrown on their sandwiches. They kept throwing soil on them however, until the grave was filled on top of the choking children. Almost identical stories are found in connection with the Black Death in the mid fourteenth century, and also in connection with the plague in 1710–1713. This indicates that there is a popular tradition which survived for a long time and was adapted to outbreaks of new epidemics. As a scien- tist, I am affected by this horrendous story as well as uncertain whether this might actually have taken place in the nineteenth century. The Norwegian folklorist Brynjulf Alver claims that the important aspect is that similar stories in Norway point to public belief, but cannot be determined as historical truth. Such determination lies outside the scope of a folklorist who inves- tigates the spiritual, immaterial aspects of culture in the form of sayings, popular belief including popular religion, and how they have been handed down over time. Some informants from villages in the province of Västergötland have attempted to enhance the credibility by offering concrete examples of children having been sacrificed in connection with cholera outbreaks during the nineteenth century. Some men have even been named as having shovelled soil on such a grave. They are said to have been mentally confused afterwards. In 1932, a male informant born in 1858 in the parish of St. Peder in the province of Västergötland told about the local cholera outbreak in 1856:

They buried a live child in the village of Båstorp hillside in Skepplanda (a neighbour parish). They gave the child a sandwich and enticed it to climb into a grave, and then they started shovelling soil. “Don’t put soil on my sandwich!” the child said. There were two men who did it. Both of them are said to have become insane. One was Johannes Pettersson, he was in Hökällan (a manor on the island of Hisingen, later to become the psychiatric hospital of Lillhagen) for several years. One was Kristiansson from Heden West in the village of St. Peder. This was done to stop the epidemic (IFGH 3068:18).

It may sound credible when the story is so specific, giving the names of two persons. In 1932, a woman born in 1842 stated “it happened in my village” in 1856. A child was buried alive in “Båstorp hillside in the parish of Skepplanda. It was a poor child who had been given a sandwich. … ‘Don’t throw soil on my sandwich!’ the child said” (IFGH 3068:17). Another informant born in 1854 in the same parish had been told by her father that a human being had been buried alive in 1834 (IFGH 3387:20). A third informant born in 1854 in the parish of St. Peder told about a girl having been buried alive during the Black Death in St. Peder or the neighbouring parish of Tunge. “They gave her a sandwich and made her enter the grave, then they started shovelling” (IFGH 3312:13). These are virtually the same 146 Anders Gustavsson stories told by different informants, which are linked both to the Black Death and the cholera epidemics in the nineteenth century in the very same village. This indicates that we are dealing with a tradition going all the way back to the Black Death and then being applied to new situations in the nineteenth century. This throws doubt on the issue of whether this sacrifice actually took place in the nineteenth century. It is worth noting that the medieval town of Lödöse was situated in St. Peder parish on the shore of Göta Älv river. Until Gothenburg was established in 1621, this was the only Swedish harbour with an entry to the North Sea. The town was given its name after one of the old churches in Lödöse, St. Peder. King Magnus Eriksson summoned a national session to Lödöse right at the time when the Black Death had reached Sweden from Norway in 1349. He ordered each Friday to be a fasting and prayer day. All citizens had to offer money to the Virgin Mary (Harrison 2000:391). The cholera hit the province of Västergötland very heavily (Harrison 2019:66). None of the informants in my material claims to have been present at any such sacrificial ceremony. They have only been told about such things having happened. Being a folklorist, I find it sufficient to point out that this tradition from the nineteenth century is very similar to stories told about what happened during the Black Death. I do not judge the veracity, leaving that to historians to discuss and assess. The historian Dick Harrison poses the question whether such sacrifices could have taken place during the Black Death. He appears to lean towards the conclusion that they happened. “Is it a question of migratory legends or of real events? Unfortunately, most of the evidence supports the view that they did occur. The method made sense. … Those who buried children alive tried to defend themselves against the plague, to turn it away, to demolish the whole by attacking a part” (Harrison 2019:113).

Summary My focus is on the many cholera epidemics which hit Sweden from 1834 until 1873. I have studied how cholera epidemics affected the countryside from the ethnological, folkloristic and cultural historian’s point of view. Popular practices and ideas in difficult crisis situations are the subject, rather than top-down regulations. How did the population perceive the cholera and how was the disease treated on the local level when it broke out? The Swedish folklore archives offer a rich recorded material which was collected from all parts of Sweden during the early part of the twentieth century. It contains stories about local developments during the various cholera epidemics. There is also legend material related to the cholera epidemics. I use public regulations from the nineteenth century as background material for descriptions of what happened in local parishes. The annual reports Nineteenth-Century Cholera Epidemics in Sweden 147 from district physicians give insight into the perspectives of individual physicians. The different epidemics share a common feature: the disease has been spread by shipping across the oceans and then along inland waterways and the shores of larger lakes. Large coastal towns were the hardest hit, the countryside faring much better. Small coastal villages were harder hit than the inland through their waterborne links to the larger coastal towns. A cholera epidemic outbreak immediately raised questions regarding barriers against the immediate neighbourhood. There are several accounts telling how town and countryside were separated, using guards around the towns. The fear of contacts was felt by town dwellers as well as by the inhabitants in the countryside. The guards in the countryside were careful in their checking arrivals from other parishes. Smoking with juniper or tar was used as a protection against cholera infection. This points to the opinion that the cholera contagion was airborne, namely a miasmatic view. It was even said that you could hear the cholera arriving in the air. There was a widely spread opinion that brännvin (vodka) offered protection against cholera. In towns, the disease hit the socially weak areas where poverty, bad hygiene, and overcrowding reigned. This tendency was apparent in the countryside as well. Since the cholera hit local communities suddenly and many died within a short time, it is to be expected that strong fears appeared. There are infor mants who had survived cholera and left tales of shattering memories. The great fear that the dead might be contagious made it important to carry them off to a cholera cemetery as rapidly as possible. It was not popular to be a corpse carrier. Free access to brännvin became popular. The burial ceremonies became extraordinary when many corpses were to be buried at the same time. It was important to avoid crowds, as was otherwise common at funerals. In the twentieth century, local associations have had collective memorial stones erected in cholera cemeteries. This helped keeping the memories regarding the horrendous cholera epidemics in the nineteenth century alive for future generations in the area. In many cases, the themes in legends about cholera had roots going back to the Black Death in the fourteenth century. I have stories showing how old popular traditions have been carried through generations and adapted to new situations. Such stories have told about dying people who have been carried off to the cholera cemeteries while still alive, and how live children have been sacrificed as a protection against the epidemic. Being a folklorist, I have investigated a collective popular picture of the world, not the questions regarding historical truth. 148 Anders Gustavsson Anders Gustavsson Professor emeritus Department of Culture Studies and Oriental Languages University of Oslo Box 1010 Blindern N-0315 Oslo, Norway e-mail: [email protected]

References Archival material Göteborg Department of Dialectology, Onomastics and Folklore Research, Gothenburg DAG Folklivsuppteckningar (VFF, IFGH) The Gothenburg City Museum Pictorial material

Lund Folklore Archives, University of Lund (LUF) Manuscript archive, folklife records

Stockholm The National Archives of Sweden (RA) Sundhetskollegium Archives The district physicians’ annual reports 1868–1877 E5A vol. 38–47

Strömstad Strömstad Museum Copies of Health Board and Municipality minutes

Uddevalla The Bohuslän Museum Pictorial material

Uppsala Department of Dialectology and Folklore Research, Uppsala (DFU) ULMA folklife records and copies of Health Board minutes

Privately held Letters and photographs

Internet www.digitaltmuseum.se (20 May 2020) http://www.ep.liu.se/databas/medhist.sv.asp (20 April 2017) Nineteenth-Century Cholera Epidemics in Sweden 149 www.ne.se (The National Encyclopaedia) http://uddevallare.blogspot.com/search?q=kolera (20 May 2020)

Published works Alver, Brynjulf 1980: Historiska sägner och historisk sanning. Transl. by Bengt af Klintberg. Stockholm. Arvidsson, Sven-Ove 1972: De svenska koleraepidemierna. En epidemiografisk studie. Stockholm. Asplund Ingemark, Camilla & Johanna Wassholm 2009: Historiska sägner om 1808–09 års krig. Helsingfors. Bernhardson, Carl Gustaf 1978: Bohuslänskt folkliv. Grundsund. Bosson, Emilie 2012: Koleraepidemin på Tjörn 1866. Undersökning av skill- nader i dödligheten mellan Klädesholmen och Stenkyrka. Högskolan Dalarna. Falun. Brockman, Anne-Marie 2010: C. G. Bernhardson, bohuslänsk folklivsmålare. Uddevalla. Drakman, Amelie 2018: När kroppen slöt sig och blev fast. Varför åderlåtning, miasmateori och klimatmedicin övergavs vid 1800-talets mitt. Uppsala. Ejdestam, Julius et al. 1943: Bilder ur lanthandelns historia. Västerås. Gerndt, Helge 2020: Sagen – Fakt, Fiktion oder Fake. Eine kurze Reise durch zweifelhafte Geschichten vom Mittelalter bis heute. Münster. Gullman, Sven H. 2003: Grundsunds kapellförsamling. Grundsund. Gustavsson, Anders 2011: The folk-life artist Carl Gustaf Bernhardson. Oslo. Gustavsson, Anders 2017: Folk Culture at the Interface between Emerging Pub- lic Halth Care and Older Forms of Healing. Arv.Nordic Yearbook of Folklore. Gustavsson, Karin 2014: Expeditioner i det förflutna. Etnologiska fältarbeten och försvinnande allmogekultur under 1900-talets början. Stockholm. Hagberg, Louise 1937: När döden gästar. Svenska folkseder och svensk folktro i samband med död och begravning. Stockholm. Hansson, Olof 1983: Min barndom på Gullholmen. Minnen och skärvor. Udde- valla. Harrison, Dick 2000: Stora döden. Den värsta katastrof som drabbat Europa. Stockholm. Harrison, Dick 2019: Digerdöden. Lund. Hasslöf, Olof 1949: De svenska västkustfiskarna. Uddevalla. Jacobsson, Eric 1989: Koleran i Morlanda. Morlanda hembygdsbok. 3. Ellös. Jakob Jonssons dagbok. 1–2. 1991–1997. Ed. by Tegneby hembygdsförening. Henån. Kongl. Maj:ts nådiga kungörelse om hwad iakttagas bör, i händelse farsoten cholera skulle inom riket yppas; gifwen Stockholms slott den 9 julii 1831. Stockholm. Kungl. Maj:ts nådiga kungörelse; gifwen Stockholms slott den 12 november 1831. Stockholm. Kongl. Maj:ts förnyade nådiga kungörelse om hwad tills widare iakttagas bör emot kolera-sjukdomens införande i riket. Gifwen Stockholms slott den 25 jan- uari 1833. Stockholm. 150 Anders Gustavsson Larsson, Lars Edvard 2015: Prosten Ekströms ordinationer. Skärhamn. Öberg, Lars 1988: Koleraepidemin i Göteborg 1834 – Katastrofen som vändes till hygieniska reformer. Göteborg förr och nu. Pettersson, Johan 1978: Det hände på Tjörn under 1800-talet. Malung. Pettersson, Johan 1979: Bygd och människor. Kulturminnen från Tjörn. Malung. Rooth, Anna Birgitta 1978: Saga och sägen. Uppsala. Schiøtz, Aina 2017: Viljen til liv. Medisin- og helsehistorie frå antikken til vår tid. Oslo. Skarin Frykman, Birgitta 1993: Larsmässemarknaden. En folklig karneval i 1800-talets Göteborg. Göteborg. Skott, Fredrik 2008: Folkets minnen. Traditionsinsamling i idé och praktik 1919–1964. Göteborg. Tallerud, Berndt 1999: Skräckens tid. Farsoternas kulturhistoria. Stockholm. Tallerud, Berndt 2006: Kolera. En farsots grymma framfart i Uppsala och på den uppländska landsbygden. Uppsala. Tjörne, Hulda 1970: Min barndoms Tjörn. Uddevalla. Zacke, Brita 1971: Koleraepidemien i Stockholm 1834. En socialhistorisk studie. Stockholm. Book Reviews

Presenting Writers and Their Work Chapter 1, “Authorship in museums”, Thea Aarbakke: Forfattermuseumsfunk deals with the starting points of the study sjonene. Musealiserte relasjoner mel- and presents the three institutions where lom liv og litteratur. En studie av Ham- authors’ lives and writings are exhibited sunsenteret, Bjerkebæk – Sigrid Undsets and interpreted. The empirical material hjem og Hauge-senteret. Det humanis- concerns Sigrid Undset and her home in tiske fakultet, Universitetet i Oslo 2020. Bjerkebæk, then Knut Hamsun and the 208 pp. Ill. Diss. Hamsun Centre at Hammarøy in Nord land, and finally Olav H. Hauge and Thea Aarbakke’s dissertation is all about the Hauge Centre in Ulvik. These dif- how knowledge of an author’s life and fer in both content and character. Sigrid works is presented in exhibitions in writ- Undset’s home in Bjerkebæk opened to ers’ museums and writers’ centres. The the public in 2007. It had then been in ambition is to investigate the exhibition state ownership for ten years and the space as an alternative way to understand home that had been in the family’s pos- how authors have been received and session since the author’s death had then assessed. This is accompanied by a focus been subjected to inventory, documenta- on exhibition techniques and means of tion and cataloguing. The museum was communication such as showcases, time- based on Undset’s belongings and was lines, texts, objects, and installations. restored as it was when the author lived Aarbakke’s premise is that this gives a there in the 1930s. An extension for the special understanding that differs from public, with a café, offices, shop, etc., other ways of presenting an author’s has also been built. The Hamsun Cen- work, such as biographies in book form tre opened in 2009. It is located on the or literary chronicles. The background is island of Hammarøy, Nordland, where a desire to test the relatively strong posi- Hamsun grew up with his uncle. The tion occupied by the author as subject. building was designed by the architect The dissertation has an important demar- Steven Holl and is an interpretation of cation: it considers only museums dedi- Hamsun’s literature. The centre is also cated to a single author, not museums or responsible for Hamsun’s childhood other institutions devoted to national or home, the Hamsun farm, but this is not more general literature. included in Aarbakke’s study. The cen- The dissertation consists of six chap- tre has few original artefacts and has no ters. It follows a traditional structure, collection. The Olav H. Hauge Centre, with an introductory chapter about which opened in 2014, is a documen- the subject of the study and previous tation and mediation centre located research, followed by a chapter on in Ulvik (in the old town hall) in Har- theory and method. Then come three danger. The centre has some objects empirical chapters and a concluding left by Hauge, such as private photo- chapter. The dissertation is written with graphs, original letters, and collections a sure hand, easy to follow and take in. of cuttings. It also has Hauge’s book 152 Reviews collection (owned by the Hauge Foun- an unexplored area, namely museums dation). Furthermore, the centre is part devoted to one person, a category of of the museum constellation Nynorsk museums which has become much more Kultursentrum (The Ivar Aasen Cen- numerous in recent decades. Inspired tre for Language and Written Culture), by the sociologist John Law, Aarbakke which represents the written language views her study as a contribution to Nynorsk. relational museology, which draws its Aarbakke provides a meticulous de theoretical inspiration from Actor-Net- scription of the different institutions and work-Theory (ANT). The starting point how they are organized. This is important is that reality is the result of relation- background information, which raises ships between human and non-human questions about the significance of the actors, and an important basic assump- organization for the design of the exhi- tion here is that entities are understood bitions. There is a detailed discussion of on the basis of the relationships of which what these institutions should be called: they are part. Equally important is that poet’s home, writer’s museum, or liter- everything can have materiality (or ary centre. Particular attention is paid agency). It is not only visions that create to the category of museum and whether exhibitions; materials, objects, and exhi- these three author institutions are or are bition technology are also important. not museums. Undset’s home defines Consequently, in this dissertation the itself as a museum, while the Hamsun gaze is directed towards the materiality Centre more explicitly rejects this term that can be found in showcases, time- as being too limiting. The Hauge Centre lines, artefacts, installations, and texts. sometimes calls itself a museum, but the To answer the questions outlined at question does not seem to be as impor- the start, Thea Aarbakke chose field- tant for them. I also wonder whether this work as a method, with a total of eight has any bearing on Aarbakke’s disserta- visits in the field during 2016 and 2017. tion, a question that does not get a clear The fieldwork included observations at answer. the three author institutions, semi-struc- In the first chapter, the author also tured interviews with people involved describes three research themes that in various ways in the activities at the are relevant to the development of the three selected sites, and studies of docu- study. Aarbakke was initially inspired ments about each institution. The empir- by research that problematizes the sig- ical data thus consists of documentation nificance of an author’s biography and through notes and photographs of exhi- how it can be expressed in exhibitions. bitions and guided tours, along with In addition, studies of place and identity interviews. Aarbakke has used a method creation are important for discussing how that derives its inspiration from ANT the exhibitions relate to the geographical and has searched for things that have an spaces in which they are located and effect on the exhibitions and that influ- how they help to pin literature to a spe- ence how the authors and their works are cific place. Finally, studies of authentic- understood. In ANT terms, one can say ity and presence have contributed to an that the author has studied exhibitions understanding of how the exhibitions as networks, where the aim has been to are designed. Texts about authenticity by identify the actors associated with the Camilla Mordhorst and Sian Jones have exhibition and acting in it. Given the been particularly important. knowledge she seeks to create and the In chapter 2 of the dissertation, theoretical approach, the method seems “Relational museology: About theory well chosen. and method”, Aarbakke describes her Thea Aarbakke has singled out three dissertation as a contribution to the field categories of objects that recur in differ- of museology and what she perceives as ent ways in the three exhibitions: win- Reviews 153 dow, desk, and book collection. These belonging and everyday life as much objects constitute the analytical focus of as literary inspiration. The big windows the dissertation and help her to answer on the fifth floor of the Hamsun Centre the question of how author museums and are associated with Hamsun’s broad per- literary centres select an author’s oeuvre spectives, but also with the ambivalence and biography. These three types of in the way he is perceived as a person. objects have been chosen because they This is very different from the case of appear in all the exhibitions and they Undset, where the private aspect is high- form part of the exhibition where the lighted, how she strove to shut out the works and the author’s biography meet. outside world. Hamsun is held up as a The starting point is that knowledge and world writer, whereas Bjerkebæk helps ideas about authors and their oeuvre are to demonstrate the author’s attempts to communicated materially, i.e. not only find somewhere quiet to write. in text; as objects they are involved in Chapter 4, “Desk installations”, draws socio-material processes. Windows are attention to how the working methods treated as exhibition techniques, desks of Undset, Hamsun, and Hauge are as installations, and book collections displayed in the various exhibitions are studied to see how they are used through the way their desks have been and which exhibition techniques are reconstructed and arranged. Here the employed to make them accessible. desks serve as an entrance to the authors’ Chapter 3, “Room with a view: Mak- working methods but also their lives and ing a place through windows”, exam- the people they surrounded themselves ines the significance of the windows in with. Exhibiting a desk is one way in the different exhibition spaces. It shows which an exhibition about an author can how the placing of the windows, the light materialize the actual writing process. they admit, and the view they offer also This is fascinating and thought-provok- contribute to an understanding of the ing reading. It is a way to show who the authors. Through her analysis of the sig- author was and how the author worked. nificance of windows, Aarbakke shows On this point the three exhibitions differ how the exhibitions interact with the slightly. In the exhibitions about Undset, outside world and how they help visitors her desk is on show, while in the other to understand the authors’ connection two exhibitions other forms of desk can to a place, their inspiration and every- be identified. By thinking in terms of day life. When it comes to windows as ANT, Aarbakke shows how the desks an exhibition technique, Aarbakke takes come into existence through photo- into account how the windows in the graphs and other things. The things are various exhibition rooms also help to perceived here, in line with ANT, not as constitute the exhibitions and how they static objects but as co-creators in the also add to the narrative. This leads to imaginary desks. Here the visitor also further discussion about the framing of becomes a co-creator, since the observer the exhibitions. She particularly empha- is an important part of how installations sizes how windows, together with other can be perceived depending on perspec- things, are significant for understanding tive and where in the exhibition space place and how relations between objects the viewer is located. contribute to this. With large windows, The last empirical study, chapter the world outside becomes part of the 5, “Book collections”, deals with the exhibition. But it is not a passive world books on display in the museum and outside. The relational concept of space the two writers’ centres. Approaching shows how the placement and use of win- the three book collections turns out to dows can give a new understanding of be a fruitful approach. There are great the relationships between place, author, large differences between the different and exhibition. Place here is about exhibitions. Undset’s and Hauge’s book 154 Reviews collections are huge, containing thou- are made visible, what perspectives it sands of books, while the book collec- adds to the author’s works, and how all tion at the Hamsun Centre consists of this interacts with the rest of the exhibi- 58 original editions. The accessibility of tions. Working with windows is also a the different book collections, how they way of approaching the boundary work are exhibited, and how far it is possible that takes place in the studied institu- to read them creates different networks tions. and relationships with the outside world. Aarbakke’s aims in the creation of The books in the collections were either knowledge are both reasonable and jus- owned or written by the authors. The tified. At the same time, the research study focuses on how different versions problem itself is somewhat vague – of the book collections are made acces- not in the sense that it is unclear what sible to visitors and how this is linked to is driving Aarbakke, but the problem the way the book collections are exhib- itself could have been formulated more ited. In this chapter, the author draws precisely, and I would have liked to support not only from ANT but also see more detailed discussion of the from theorists such as Michel Foucault choices made by the author and the con- and Roland Barthes. The books turn sequences of these. The text tends to out to be flexible aesthetic objects that report more than to discuss the choices connect people, ideas, and things in the of theory, method, and empirical data institutions’ buildings, bookshelves, and that were significant for the implemen- showcases. Aarbakke demonstrates how tation of the study. I would have liked museums exhibit and treat objects cre- to have read in greater depth about the ated through relationships between doc- challenges the author faced when writ- umentation managers, exhibition tech- ing the dissertation and how she justifies niques, and existing collections. In the her choice of study objects, method, and search for book collections, she leaves theory. It is not that I want a different the exhibition premises and reaches study, just a more profound considera- other parts of the museum such as librar- tion of the choice of problem, the selec- ies and databases. tion of authors and specific objects. In the final chapter 6 of the disser- The choice of institutions concerning tation, “The function of an author’s individual authors instead of institutions museum”, the results are summarized, devoted to national literature or specific with particular emphasis on how a periods seems reasonable to me, as a study of exhibition spaces can serve way to examine what relations are estab- as an alternative way to respond to the lished between literature and life. The authors’ oeuvre and biography. choice of author institutions is deter- Thea Aarbakke’s dissertation is mined by practical considerations, that it undoubtedly thought-provoking. The must be realistic economically and geo- analyses are multi-layered, demonstrat- graphically. This is perfectly reasonable, ing how an exhibition contains so much but an important question is whether this more than a set of objects presented also has analytical consequences. What with accompanying texts. The strength is the strength of a national demarca- of the dissertation lies in the interpre- tion and what difference does it make? tations of the three selected objects. What is the consequence of a practi- Here Aarbakke elucidates how both an cal demarcation? As far as the empiri- oeuvre and a biography can be under- cal foundation is concerned, it seemed stood in the respective exhibitions. She meagre to me at first glance, and the approaches windows as something other fieldwork was relatively limited, but the than just light sources, and shows how important thing, of course, is what has they interact with the exhibitions. She been collected and how this material is answers questions such as which places then used. Here Aarbakke has been cre- Reviews 155 ative and the collected material is well also thoroughly scrutinized her empiri- employed. cal data and written up her results. It is When it comes to the choice of the possible here to conceive of a great vari- three sites to investigate, Aarbakke, as ety of networks that shaped the research mentioned earlier, describes how the process and how the three objects – win- various institutions position themselves dow, desk, and book collection – have in relation to defining themselves as been written and interpreted. What does museums. Here I would have preferred the author’s contemplation mean for a discussion that takes into account that the analysis? Despite the descriptions the study is focused on exhibitions, one of the research process, there is some- or more of which are found in museums. thing capricious about Aarbakke’s own It would be interesting, then, to discuss position. Who is this dissertation writer whether the institutional framework of who interprets book collections and a museum or a writer’s centre has any museums? What is the author’s own effect on the design of the exhibition relationship to exhibition production, and, above all, the knowledge conveyed museums, authors, and literature? I am about the authors and their works. Such not calling for a more self-absorbed dis- a discussion would also have been in sertation writer, but for a framing of the line with the ANT-influenced approach foundation on which the interpretations and the discussion of the importance have been made. How does she draw the of naming as boundary work. In what boundaries around the network she stud- way does the definition as a museum ies? What is noted and ascribed agency, give materiality to the presentation of and what is ignored? An example of this an author’s biography and oeuvre? This is how the studied objects acquire their is something that could well have been status. This is relevant, for example, in taken up in the concluding discussion. the definition of what a desk is. In the The way Aarbakke has chosen to case of Hauge it is the dining table that achieve her aim is original. The three is defined as a desk, and here the author objects – window, desk, book collection could have reflected on her own naming – are well chosen and lead to interesting practice. Does the analysis collapse if it reasoning and insights. But I am curious is allowed to be a dining table instead of about how these items were selected. It being identified as a desk? There is no is possible to perceive the selection of in-depth discussion of the researcher’s study objects as being, if not haphaz- own significance for the interpretations ard, then at least arbitrary. We are told, and understandings of what the exhibi- for example, that Aarbakke was inspired tions communicate. Are there alterna- by a window that was on display at the tive interpretations here, and would they Dickens Museum in London. Did she have affected the validity of the results? consider any other items? Was there any- To continue along the same lines. One thing in these objects that also seemed to of the reasons that Aarbakke chose to use capture the more general impressions of ANT ‒ as I understand it – is that it ena- visits to the three places? How did Aar- bles more narratives, that it is possible to bakke go about searching through the think new thoughts about how to under- materialities that shaped the exhibitions? stand the authors and their works. Here Thea Aarbakke’s theoretical start- we are obliged to follow Aarbakke’s ing points raise questions about the reading. It would be interesting to see researcher’s own position in the network other possible interpretations. Are there that is studied and the consequences alternative narratives and, if so, what are this has for the results of the study. The they? Aarbakke has chosen to avoid the author has alternated between being in visitors’ experiences, which is a reason- exhibitions, searching in databases, and able demarcation based on the empirical studying the contents of books; she has data. But how are the exhibitions consti- 156 Reviews tuted in relation to visitors? What effect Sámi Media and Activism do they have? While it is certainly pos- Coppélie Cocq & Thomas A DuBois: sible to think of Aarbakke as a potential Sámi Media and Indigenous Agency visitor, a possible approach would have in the Arctic North. New Directions in been to do more detailed ethnography, Scandinavian Studies series. University i.e. describe how the exhibitions are of Washington Press, Seattle 2020. 334 actually used, in relation to Aarbakke’s pp. Ill. interpretations of them, both to test her own interpretations and to catch sight In the last fifty years or so, Sámi people of other potential perspectives. There is have been using new media formats and no reason to distrust the interpretations genres in order to build solidarity and presented in the dissertation, but the self-respect, challenging stereotypes and argument would have been reinforced if powers of societal structures, to reach there had been room for alternative nar- out to other indigenous peoples around ratives as well. the world, and to claim their presence in A discussion of the validity of the results public society. Expressive culture affords would have been one way to tackle the strong channels for expressing and per- questions raised above –for example, in forming self-conscious subjectivities. relation to the issue of the position of the This development is covered in this dissertation author and potential inter- joint project monograph by Cocq (Umeå pretations. The reason for visiting three University, now at Helsinki University), institutions was to show that there are and DuBois (University of Wisconsin, different ways of exhibiting an author’s Madison). There is a rough division of oeuvre. Is it even interesting to compare the book into chapters written largely by the different institutions, and on what DuBois focusing on music, poetry, film, basis can such comparisons be made? and pictorial art, and by Cocq on social How can the different interpretations of media and language revitalization. But desks, windows, and book collections be they have also expended a lot of effort to reconciled for the different author insti- make the book an integrated whole. Dif- tutions? Do they reinforce each other? ferent Sámi denominations for varieties Can the interpretations of these objects of snow and evocations of different geo- be said to be indicative of how the insti- graphical places are used to name and tutions treat their respective authors, or characterize the chapters, thus function- could they be perceived as some form of ing as metaphorical devices that give a counter-narratives? loose narrative structure to the outlay Overall, Thea Aarbakke’s dissertation and serve as smooth transitions. I can is a creative and original study. As I have see three themes that are recurring gen- pointed out, it raises questions about eral features through the book: Media demarcations and choices of interpre- use and media presence; The (re-)for- tation. Having said that, there is a great mation of Sámi as political actors, and deal to be gained from the dissertation. Artistic expressions as political stance. It displays playfulness as regards the Of course, using contemporary media interpretative framework and the way in order to enhance your situation is not of approaching museums in general and a new phenomenon although the extent exhibitions in particular. It demonstrates of it is restricted by the actual possibil- the potential to understand how things ity to gain any control of the means of and spaces interact in exhibitions and production. The efforts of Olaf Sirma how this affects presentations of writers in the late seventeenth century and Elsa and their work as a whole. Laula and Johan Turi at the beginning of the twentieth century to project a Sámi Lars Kaijser subjectivity to the European literati and Stockholm, Sweden political power are drawn upon to form Reviews 157 a backdrop to the more recent develop- 2013 against the plans to open a mine ment. The protests against the dam con- in a reindeer grazing area are used as struction at Álttá around 1980 are put an example. A third use of social media forward as a turning point where Sámi is as a communicative strategy along- activism took a step further, in political side, instead of, and in interaction with subjectivity as well as in communicative mass media – here again, the Gállok strategies. case shows how the communication for Two chapters follow the development a long time mostly went through social of artistic work as cultural and politi- media until the Washington Post brought cal statements by close reading of dif- the issue to international attention. ferent works of art that had significant There is one chapter that is a contri- impact on international audiences. The bution to the historicization of social selection is Nils Gaup’s film Ofelaš media, starting with the story of the (The Pathfinder, 1987–88), Nils Aslak SameNet, a platform founded in 1996 Valkeapää’s cross-art Beaivi, áhčážan as an outgrowth of a distance educa- (1989), Mari Boine’s Gula gula! music tion project, using a FirstClass server, album (1989), Aŋŋel Niedat’s music financed by two Swedish counties and album Dolla (1992), Paul-Anders Sim- the Ministry of Communications. Mak- ma’s film Oaivveskaldjut (Give us back ing e-mail contacts and chats on rele- our skeletons, 1999), Nils Gaup’s film vant topics available, it also worked as Guovdageainnu stuimmit (Kautokeino a forum for using the Sámi language; opprøret, 2008), Mari Boine’s Bas riik- another aim was the computerization kažan, a re-make in north Sámi of Ole of Sápmi. Apparently this meant a great Paus’s song Mitt lille land (2009), Sofia deal in making the internet available Jannok’s album Àhpi (2013), John Hen- for inter-Sámi connections and the use rik Fjällgren’s Eurovision entry song/ of the Sámi language on the web. How- video Manne leam friije/Jag är fri ever, the high speed of technical inno- (2015), Jörgen Stenberg’s yoik/video vations soon made it obsolete and after Vuortjis/Kråkan (2013, 2015), Katja some years of decline, when the efforts Gauriloff’s film Kuun mestän Kaisa of Sámi organizations to obtain financ- (2016), and Amanda Kernell’s film ing for technical renewal did not get Sameblod (2016). The adaptation of tra- anywhere, it was closed in 2011. Super- ditional expressions for contemporary seded by Facebook and other platforms communication, the ways of dealing for everyday contacts, the SameNet is with the new possibilities available in still remembered with nostalgia, and the new media and genre formats, the some Sámi express a loss of a digital connections established to other indige- context where Sámi could meet as Sámi. nous peoples, and the changing political This is further underlined by discussions topics are in the focus of the discussions. of contemporary social media use and One chapter deals with Sámi uses of the problems of what information to social media. Drawing upon interviews, convey on Facebook, the harassment it shows how social media complement and the internet trolls. The need for trust and intensify existing relations, such and safety is a crucial issue. as keeping up family bonds; further- The digital competence gained by more, from interviews with activists SameNet and other efforts has resulted and internet ethnography it also exam- in a distinct Sámi presence on the inter- ines how social media provide platforms net and in social media. This contempo- for establishing pan-Sámi networks, rary situation is covered by examples strengthening existing organizations and of projects reaching inwards as well building new, often occasional networks as outwards. The community-strength- around specific events and campaigns – ening aspects, and also the use of the here the protests at Gállok/Kallak from internet as a base for language policy, 158 Reviews are exemplified with the efforts to vital- ologies. For instance, the story of the ize Ume Sámi, one of the smallest Sámi SameNet gives insight into the digital languages, until recently spoken by very imagination at the turn of the millen- few and with almost no teaching mate- nium, when Sámi actually built that rials available. In 2014 the Ume Sámi kind of community-reinforcing platform organization started a course project that supposedly was one of the potential where first an Ume Sámi lexicon was functions of the worldwide web. It also constructed, using the Memrise app; shows the fragility of relying on a single later on, videos were added and made it media technology when there is a high- possible for people to hear the language speed turnover in software development. spoken in interaction. In 2016 an official The book also inspires asking new orthography established Ume Sámi as questions, and I would like to end this a written language. In 2018 the project review by proposing some themes had some 400 individual users – to be to proceed with – mostly on the art- compared with the “ten or so” speakers ist-as-activist topic but with relevance numbered in recent accounts. The inter- for questions of media strategies as well. net possibilities of enabling contacts How do the logics of the global atten- between geographically scattered peo- tion economy affect the possibilities of ple has shaped new situations where a being visible at the right times? There language can be spoken, written, heard, is always a risk of being overshadowed and read. by other urgent topics, or still worse, There are also many forms in which to getting caught on the wrong side of a reach out to others. Instagram accounts divide. There is also the sad fact that with rotating curators are used to show there is a kind of competition for atten- everyday Sámi life. Posters and videos tion with others who really would be the produced by the anonymous artist col- first allies. lective Suohpanterror are presented; The precariousness of the artist as they often comment on international a professional is hinted at. Besides the politicians’ ignorance and claim a Sámi unstable conditions for any artistic presence. career, the risks of being classed as hav- The concluding chapter returns to the ing a “niche profile” further contribute importance of the Gállok/Kallak pro- to insecurity, especially when the protests in 2013 as a moment of increased fessional status comes from interaction awareness, an impetus for activism, and with an international, mostly non-Sámi increased international indigenous sol- audience. But can the integration in idarity. The interplay between online Sámi society make for a different and and localized offline activity and the stronger base to work from? This leads inter-media aesthetics enables both an to the question of the importance of fam- international outreach and simultane- ily relations in order to be a trustworthy ously a strong connection to local con- activist – which can be problematic texts and events. when claiming a Sami identity in itself The book gives a solid presentation becomes an activist position, particu- of contemporary Sámi activism, artistic larly when family bonds have been sev- creativity, and media strategies, with a ered and Sámi identity denied for gener- historical perspective spanning all of the ations because of oppression. twentieth century and up to the present For artists there is also the balance day. As a bonus, we also implicitly get between the activist stance and the mid- a potted history of mass media devel- dle position as translator and mediator opment since the 1960s, in the shape of interacting with the outer world. This innovations that Sámi have evaluated, is a point where authenticity is made as adapted, and applied, in a dialectical well as challenged. interplay with contemporary media ide- There is also much to gain from ana- Reviews 159 lysing relations between Sámi artistic the huge extant stock. The Royal Swed- expressions and the category and genre ish Library has no comprehensive cat- systems of state cultural policies and alogue, and the broadsides are also international markets. Is the album the divided into several series. By contrast, end-point of music making? Is yoik the broadside ballads in Uppsala Uni- poetry, narrative, or music? What are versity Library were catalogued by the the artistic and political consequences historian Hanna Enefalk some time ago. of relating or not relating to the “ethnic Her catalogue, however, is tucked away slot” or the world music market? in the Alvin portal and therefore not so The history of ethnology and folklore easy to use. In her book Skillingtryck! studies reminds us of the strong politi- Historien om 1800-talets försvunna cal force potentially inherent in artistic massmedium (2013), she states that expressions of oppressed peoples. Sámi Uppsala University Library has more uses of new media technology and gen- than 11,000 broadside ballads (p. 56). res further underline the importance of Because all printing houses are obliged this perspective. to deposit copies of everything in these libraries, the stocks in the Royal Library Alf Arvidsson and in Uppsala ought to coincide, but Umeå, Sweden that is not the case, of course. Several extant broadside ballads are missing from both libraries, but can be found elsewhere. An important collection is the 2,100 prints that the English anti- The Producers of Broadsides quarian George Stephens bought during Eva Danielson: Skillingtryckarna. Skil- his years in Stockholm 1834–1851. lingtrycksproducenter under det långa Another stumbling block to research 1800-talet. (Meddelanden från Svenskt into broadside ballads has been the lack visarkiv 53.) Svenskt visarkiv/Musikver- of knowledge about the many printers ket, Stockholm 2019. 273 pp. Ill. that produced broadsides over the years – and over the whole country. That prob- Researchers today do not need to point lem has now been remedied by Eva Dan- out that the Swedish term broadside ielson’s book entitled Skillingtryckarna, ballad (skillingtryck) does not refer to “The Broadside Printers”. Eva Dan- a sentimental song from the late nine- ielson has been archivist at Svenskt teenth century, but can explain that it Visarkiv, the Centre for Swedish Folk denotes a printed genre that for several Music and Jazz Research, and as part of hundred years spread ballads and some her work she has handled a great many prose texts. Emotional ballads from the broadside ballads. She has long been past – Alpens ros, Lejonbruden, Bredvid interested in the printers who produced sin häst i gränden, and others of that the broadsides, an interest which has kind – are now so far away in time in the resulted in several previous articles. The public consciousness. Perhaps the rise present book is therefore the result of of the real broadside ballads from the decades of work, as is evident from the conceptual mist has led to these prints careful treatment of all the details. becoming visible to cultural researchers The book consists of two parts. The in several disciplines. It is clear that the first provides a kind of background to texts of the broadsides are being increas- the second. She begins with an account ingly used as source material for studies of the spread of the art of printing in with widely differing orientations. Sweden, the occupations within the A fundamental problem for anyone industry, and the economic conditions who wants to approach the broadside in which printing operated. She also ballads is the difficulty of surveying introduces the different types of printing 160 Reviews houses and the technical changes in the of small-wares could also buy a batch business. When she approaches the proof the broadsides. Private commission- duction of broadside ballads after these ers of prints had generally composed general descriptions, she first describes the words of the ballads on the broad- the broadsides as graphic products: their sides. These people travelled around and size and the number of pages, the typog- made a living by selling their printed raphy (Fraktur or Antiqua typefaces) works. The publishers were major cli- and paper grades. She then describes the ents. Svenska Visförlaget, a publisher outlets where the broadsides were sold, in Solna with Jöns Hansson Chronwall and the prices (just after 1800, a broad- as its original owner, commissioned no sheet usually cost one skilling, hence the less than 1,100 broadsides produced by name). fifteen presses in Stockholm. Malmö Printing houses not only had to sub- Visförlag – where Nils Lindström was mit copies of each printed item to the the enterprising manager for a long time Royal Library, but also had to deliver – published several hundred broadsides. detailed annual lists of everything they The second and most comprehensive produced. Eva Danielson presents – and part of the book is a detailed catalogue discusses – these “print lists”, which are of all the country’s printers who pro- important source material in her study. duced broadside ballads during the long (Some of the lists state that the press nineteenth century. (Although Sweden produced “peasant goods”, an expres- until 1809 also comprised Finland, the sion that tells us about the customer eastern half of the kingdom before that base of the broadside ballads.) Finally, year is not included.) The catalogue is she writes about the publishers that were arranged according to the place of print- responsible for issuing a large share of ing, and for each place the printers are the broadsides in the late nineteenth cen- arranged chronologically. The informa- tury and until the 1920s – the technical tion about each printing house is mainly production still took place in the print- taken from the classic Svensk boktryck- ing houses. erihistoria 1483–1883 by Klemming and The background that Danielson paints Nordin from 1883. But Eva Danielson is rich, to say the least, with abundant has also used other sources, since that terms, names, places, and dates. Among book does not cover her entire investiga- all the interesting information we learn tion period. The history of almost every that the normal print run was 1,000 printing house is complicated: changes copies (p. 62). Since each broadside of ownership, changes of name, changes was read and used by far more than of production orientation, and moves the owner, it can be estimated that it to new locations. But she still manages reached at least 5,000 people. But there to give an account of her 172 printing were broadsides that were printed in far houses (if I have counted correctly), 60 more copies, especially those with bal- of which were in Stockholm. lads about sensational current events, of After presenting the history of each which there were many. printing house, Danielson tells about From the point of view of the printing their production of broadside ballads, houses, there were two kinds of broad- based partly on the print lists. But side ballads: those initiated, produced, she constantly compares these to the and sold by the printing houses them- Uppsala University Library catalogue selves, and those produced to order on and the catalogue of George Stephens’ behalf of private individuals and pub- collection. As regards comparisons lishers. The former were always sold with the stock in the Royal Library, by the printers, but also through some she reports throughout on the results of bookshops that were often connected “sampling” in order to establish what to the printing house. Itinerant vendors Sweden’s national library has preserved Reviews 161 of the products of each printing house. side academia for four decades. Why It is safe to say that she has been very change a winning concept? It becomes ambitious and that the information she no clearer when the authors of the newly has amassed will stand for a long time written preface declare that they have to come. For the majority of the print- reworked the original text with a light ing houses she can state that the print hand. A light hand but still a completely lists do not indicate the total production revised form? Are the authors on a colli- of broadside ballads, but also that the sion course with the publisher’s wishes? print lists include broadsides that have Before we take a closer look at what not been preserved. These clarifications the revision entails in Frykman’s and will be valuable for any kind of future Löfgren’s encounter with the original research into broadside ballads. text, let me say a few words about why I can think of many different terms to Den kultiverade människan is still on describe this monument that Eva Dan- the reading list for students of ethnology ielson has erected to the production of and is still bought by educated readers. broadside ballads: book history, educa- The bourgeois culture of the Oscarian tional history, literary history, history era at the beginning of a new century of mentalities, local history, history in 1900 may seem like a study of exotic of technology, and economic history. “others”, perhaps tempting to read as I am sure there are still other ways to such, but still at a suitable distance from read it. In any case, with her outstand- the everyday life of present-day Swedes. ing research effort, Eva Danielson has But what readers gradually discover is achieved an exceedingly rich work that that the lost-looking figure on the cover many will rely on with gratitude. does not seem totally unfamiliar. It is partly about me, and that insight is both Gunnar Ternhag upsetting and fascinating. The authors Falun, Sweden take us on a journey to the world of cultural deep structures, where what we believe to be non-negotiable truths about reality are unmasked as historically specific ideas. In the art of revealing this, A Revised Classic of Cultural Den kultiverade människan is unsur- Analysis passed. It is not a matter of everyday Jonas Frykman & Orvar Löfgren: Den thoughts about existence that turn out to kultiverade människan. Gleerups Utbi- be historically conditioned in the light of ldning, Malmö 2019. 2:a omarb. uppla- what emerges from the book. It is about gan. 262 pp. Ill. destabilizing perceptions of what it is like, what is unconditionally true. It is “This book is a classic, now published the consistently implemented historici- in a completely revised form, forty zation of what is taken for granted that years after the first edition.” So says the is the attraction of Den kultiverade män- blurb on the back page of the second niskan. edition of Den kultiverade människan. Historicization cannot be done with- The promise of “a completely revised out history; this is the credo that propels form” is really the only thing that jus- the analysis. The tool used for unmasking tifies a review, a task that has fallen on the culture of the cultured human being my shoulders. The word “completely” is the method of contrasting. A culture makes me confused but at the same time only becomes visible if it is set against curious. For what can be done differ- another culture. That role is played here ently and better in a publication that is above all by the pre-industrial peasant still of great value? It is a book that has culture, which means that the reader is been a bestseller both inside and out- given an equally penetrating picture of 162 Reviews what the world looks like from the per- What knowledge? It is the showpiece of spective of peasants shaped in a feudal the Enlightenment, the scientific knowl- society. A consequence of the latter is edge that Lyotard mercilessly trans- that one sometimes becomes uncertain forms into discourse and thereby makes as to what is the main object of study. it into a narrative alongside other forms This tendency emerges in Frykman’s of knowledge that can claim to represent section “Clean and Proper”, where the the truth. This comes to mind as I read chapter on “Peasant Views of Purity the following in Den kultiverade männi- and Dirt” is given the same space as the skan 2019 about world view: “The world second main chapter, “Bourgeois Disci- view often acquires its legitimizing, sta- pline”. It may be added that workers are bilizing power with reference to forces sometimes allowed to act as objects of and powers outside man in the form of distinction, but without being granted divinity, natural conditions, common a culture of their own. It is usually in sense, or scientific logic (my emphasis) the towns and cities that the underclass – conditions that are above discussion, get to act as “the others”, often in the conditions that cannot be questioned” close proximity of the bourgeoisie in the (p. 246). A link can also be made to the form of servants in the labour-intensive first part of Lyotard’s title; “the- post homes – the ultimate source of embour- modern condition”, because Den kulti- geoisement. verade människan is undoubtedly about The theoretical influences can be a far-reaching denaturalization of the found in the tension between the human- language games, regulations, and val- ities and the social sciences, with names ues of modernity. The cultured person such as Bourdieu, Foucault, Elias, Sen- is in many respects the modern person; nett, and E. P. Thompson as sources of the disciplined individual who is able inspiration. They all use the historical to distinguish feelings and family life perspective to transform what we per- from external reality with its demands ceive as natural into cultural. for career, profit maximization, and uni- On the conceptual level, Ruth Bene- versal growth for business and society. dict’s classic words, “the lens through It is only with the postmodern gaze that which we see the world”, “culturally that human type becomes visible as a created reality”, or “world view”, are cultural being. used to designate the focus of the study. Four decades have passed. A revi- The authors point out, above all, that sion has to be undertaken. How do the “world view” is suitable for analysis, but authors handle that task? It is a delicate at the same time they acknowledge that enterprise, not without a degree of risk. it is “not a pregnant analytical term” (p. With the original from 1979, the authors 245) and there I agree with the authors. were in a secure position, with a book It feels as if they are looking for a con- that was at the scientific forefront when cept that does not yet exist to comprise it was published. Frykman and Löfgren the essence of the study object. When I cannot be blamed for overlooking what read Den kultiverade människan today, has happened since then in scholar- Jean-François Lyotard’s term language ship and society. It is only now that the games (taken from Wittgenstein) comes authors find themselves in the complex to mind. That term corresponds well to situation of having to decide what to the level of collective consciousness on keep, what to insert or omit, and what which Frykman and Löfgren focus. Lyo- to rewrite in the light of forty years of tard also uses language games to expose profound changes inside and outside deep-seated ideas that Western man has academia. The promised “completely trouble relativizing and historicizing. revised form” makes tough demands. Lyotard’s classic is called The Postmod- Judging this revised edition is also ern Condition: A Report on Knowledge. a delicate task, since a close reading Reviews 163 of each version is not possible. I have only as a category that the bourgeoisie mainly looked for new references, new observes and distances itself from. Sen- or amended chapter headings and sub- nett remains one of many “inspirers”, heads, large passages that have been but is not mentioned as a researcher of omitted and sections that have been the “outlook on humanity” of industrial added, and I have kept an eye out for workers. allusions to new scholarly perspectives This is followed by Orvar Löfgren’s and current societal problems. The form part, entitled “The Transformation of of the book is the same as before, but Emotion”, now without the subtitle from the number of pages differs: 240 pages 1979, “Time, Nature, and the Home in in the 1979 edition have swollen to 262 Bourgeois Culture”. In the chapter on in the new version. But that is not a good time, the section “The Tyranny of Time” benchmark because the new edition is has been deleted and a new one has been set in a larger typeface with wider line added under the title “The Future on spacing. My impression is that the con- Rails”. This depicts the emerging rail- tent has been compressed quite a lot in way as an important way of schooling the revised version. people in a new form of time discipline, The “Introduction” follows the main linked to the demands of industrial cap- lines of the original version, but the italism for speed and precise timing. discussion is more stringent and has an Under that heading we find parts of the accentuated critical edge aimed at the content of the section on “The Tyranny contemporary “craze for the national of Time”, as also in the concluding chap- and cultural heritage […] tendencies to ter “Time as Created by Society”. This is idyllize pre-capitalist and pre-industrial a good example of Löfgren’s technique ways of life” (p. 16). Also commenda- in the revision. Formulations are pol- ble is the excision of the text about the ished, minor clarifications are added, project “Culture and Class” and the headings disappear, but the content can homage to fellow researchers that was pop up in other sections where it may fit so common in the academic writings better, and new headings are inserted. of the time. However, there are no new In the following chapter on nature references with the exception of two we see further additions, reshuffles, and project products, Modärna tider from excisions. A fairly long new section 1985 and Försvenskningen av Sverige (7 pages) is “Our Friends in Nature”, from 1993. One difference, which is on which problematizes the cultured per- such a scale that it comes close to the son’s complex relationship to animals, description of “completely revised” and the emotional commitment to pets in which is reprehensible in my opinion, is relation to the parallel demarcation vis- the omission of the working class from à-vis the animal world. Another addition the introduction. In 1979, the industrial is “Class Encounters in the Landscape”, working class was held up as being which describes the meeting between equal, in principle, to the bourgeoisie two different ways of looking at land- by having its own “way of life and […] scape, represented by bourgeois sum- outlook on humanity” (p. 11). There is mer guests and the working population a reference to Richard Sennett, who has in a Bohuslän coastal district. The latter “analysed the confrontation between the found it difficult to understand the holi- outlook on humanity in bourgeois cul- day makers’ infatuation with the natural ture and among industrial workers” (p. scenery that was anything but captivat- 12). Under the heading “Peasants, bour- ing to them. Here workers stand out as geois, workers” we are told that a “work- representatives of their own “worlds and ing-class culture” has been created dur- lifestyles” (p. 84), which compensates ing the century (p. 13). The title is the for the loss of class perspectives in the same in 2019, but the workers now exist Introduction. One section that has been 164 Reviews deleted is “Nature as a Commodity”, tariat”. The words speak for themselves. which is regrettable because there was a These are terms that do not belong in potential here to bring in contemporary our time, they feel passé, and the fact climate and environmental problems. that they no longer occur under the The 1979 edition touched on how goods unchanged heading in 2019 can be seen are marketed by exploiting people’s long- as an indication of updating. ing for nature and all things natural, as in In the last of Löfgren’s trilogy of advertisements for herbal shampoos and chapters, “The Family Person”, a cou- corn flakes. But the commodification ple of headings have been changed. The has another side that manifests itself in section entitled “The Home” in 1979 a more brutal exploitation of nature, as is now renamed “The Home District” materialized in clear-cut forests, mining (Hembygden), and that shift is based on operations, and other industrial projects. Löfgren’s later research on the nation as Business-like approaches of that kind a home or a motel. Unfortunately, I can- also have a place in bourgeois culture. not find much in the text that justifies the This leads me to think that the cultured new heading. Although the opening is person is only one part of the bourgeoi- different, the focus is quickly shifted to sie, namely the educated Bildungsbür- the role of the home in bourgeois culture. gertum. Doctors, lawyers, accountants, The word Hembygd (Heimat) is more professors, etc. can enjoy natural scenery ideologically charged than Home, which without having to soil their hands. It is thus raises other expectations about worse for the commercial and industrial the content. One heading that has been bourgeoisie who have to make trade-offs eliminated is “A Domesticated Work- between preserving for the sake of beauty ing Class?”, and a section with much and destroying in the service of money. the same content is now headed “New The last part of the section on nature, Homes”. The reason for the change is “The Culturally Created Landscape”, possibly the same as mentioned above. is slimmed down in the new version to Finally, some comments on Jonas one page from the original two. When Frykman’s part “Clean and Proper”. counted in words, the difference is even The book from 1979 contains the sub- greater because of the change in typeface head “Body and Soul through Peasant and line spacing. This is the part that is and Bourgeois Eyes”, which has been furthest from the original in the book as removed from the 2019 version. It would a whole. And I find that surprising. These have been good if it had been kept as a were pages of fundamental importance, clear pointer to what distinguishes this giving a deeper understanding of the rela- from Löfgren’s part, namely the absence tionship between the formation of bour- of the working class as a contrast. geois culture and the profound changes in As regards the updating to cover con- working life and society that took place temporary social problems, we see one when peasant society was disintegrating example of this in the short opening chap- and industrial capitalism was making its ter, “The Cultural Basis of Physical Aver- breakthrough. One could perhaps look sion”. Briefly and succinctly, Frykman at the range of terms used to find the writes in connection with a discussion reason for the drastic revision. The fol- of people’s vulnerability and the pitiless lowing concepts are not represented in conditions in Sweden’s days of poverty: the 2019 section: “production system”, “In today’s political climate, a growing “social structure”, “production condi- number of people see this as a time when tions”, “production landscape”, “col- Sweden was still Sweden – even though onization”, “cultural revolution”, “ide- during one generation a million and a half ology”, “modern industrial capitalism”, people left their native land for America” “landscape exploitation”, “consumer (p. 149). One could wish for more sharp sphere”, “working class”, and “prole- digs like this, but given the heading one Reviews 165 should perhaps be grateful that Frykman Since the 1960s Professor Helge Gerndt sticks to the past. It is up to the gifted has mostly worked in Munich as an eth- reader to make the update. nologist. The study of legends has been With the tools I have used to assess a prominent feature of his research. It this revised version of Den kultiverade started with his doctoral dissertation in människan, there is little more to say 1966 about legendary figures of the seas, about Frykman’s contribution. The namely, the Flying Dutchman and the headings are the same, and so too is Klabautermann. In the book reviewed most of the content. The difference lies here he has compiled and elaborated on in the faster tempo due to the typeface, much of this research which he has been which has required some compression, pursuing for over fifty years. Chapter 2 a positive factor for the reading. What (pp. 17–31) deals with the question of speaks in favour of my impression that the Flying Dutchman on the ghost ship the changes are marginal is that the of that name. The oldest examples come number of new references is limited to from the late eighteenth century. Chap- one: Inger Lövkrona (p. 193). Unfortu- ter 8 (pp. 123–138) examines the Kla- nately, that publication is not included in bautermann, a guardian spirit of ships in the bibliography. the countries around the Baltic Sea. This The partially rewritten “Conclusion” supernatural being is first attested at the is a praiseworthy summary of perspec- beginning of the nineteenth century. tives, methodology, and study results, The author’s aim is to show how leg- with discussions in relation to later ends exist in the border zone between research. But it is rather too wordy and fact and fiction, between truth and contains, in my view, some irrelevant doubt. Here it is important to discuss the reasoning about the task of ethnology in meaning of the word “reality”. The con- relation to closely related research, and it nection to an alleged reality and to a spe- is therefore not as stringent and convinc- cific place distinguishes a legend from a ing as the conclusion in the 1979 version. fairy tale, which is an aesthetic product. Finally, the question of “completely In addition, there are also legends with revised” or “with a light hand”. Both supernatural elements that can tell about authors have been careful in their revi- figures in heaven or hell, for example, or sions, but Frykman more so than Löf- about spirits of nature. gren. I hope I have made it clear in what Over time, legends have been passed way. I have some critical remarks about on through processes of transmission this new edition of Den kultiverade which have led to changes to the original människan, but I would like to stress content. In this way, legends can reflect that the book in its new version is still their own time but they cannot say any- indispensable as required reading for thing about the truth of their long prior courses in ethnology. history. Legends about Duke Henry the Lion (1129–1195), a previous spe- Mats Lindqvist cial study of the author, are attested in Huddinge, Sweden medieval manuscripts, the oldest from 1180. From the sixteenth century to the nineteenth century this legend was published in new versions on handbills. In the nineteenth century the stories were Legend Research through the Ages adapted to conform to contemporary Helge Gerndt: Sagen – Fakt, Fiktion heroic poetry and the romantic view of oder Fake? Eine kurze Reise durch the soul of the people. zweifelhafte Geschichten vom Mit- In general, the notion of national telalter bis heute. Waxmann, Münster, unity became prominent in the inter- New York 2020. 244 pp. Ill. pretation of folk legends in the nine- 166 Reviews teenth century. In the twentieth century tral but purported facts that are intended this was succeeded by an interest in to confuse and frighten people. There the regional and the local, the Heimat. is not supposed to be any doubt. One National socialism could then also make example is the fifty articles in the Süd- use of folk legends. This demonstrates deutsche Zeitung between 6 October the importance of researching the con- and 6 November 2001 about letters that text of the legends. No definite truth or were sent in the USA containing anthrax reality can be attained. bacteria which caused the death of some Looking at the history of the disci- people. The texts were accompanied by pline, the author gives a detailed survey drawings. One variant illustrated how of how legends have been viewed, col- death came by mail in the letterbox lected, and explored from the Middle and how the postman wore a protective Ages until the present day. The great mask and gloves. A distinctive feature work of collecting written folk legends of modern legends is that they spread started at the beginning of the nineteenth rapidly and even globally with the aid of century through the brothers Jakob social media, but also that they are more (1785–1863) and Wilhelm (1786–1859) short-lived and fleeting than legends in Grimm. They wanted the material they older times. published to be as rich as possible. A To sum up, Gerndt’s study goes into prominent idea was the urgency of rescu- great depth, presenting collections of ing the legends from oblivion. The broth- material stretching over a long period ers Grimm published Deutsche Sagen in of time, citing previous research, and 1816 and 1818. In Germany, the publica- inserting his own reflections throughout. tion of legends has been more extensive The author conducts detailed discus- than research into legends. Before the sions of hard-to-define folkloristic con- nineteenth century, only written sources cepts such as legend, fairy tale, story, were available, while the oral tradition narrative, reality, truth, fake news etc. left no trace until the collection of folk- The book will be a standard work for lore got under way towards the end of the future research on legends. nineteenth century. The author does not content him- Anders Gustavsson self with studying legends from early Oslo University, Norway/ times, but also examines legends from Henån, Sweden the twentieth century and into the twenty-first century. He uses newspaper material as a crucial source. In Germany the term “newspaper legend” emerged in the 1930s. A modern legend of this kind A Book of Omens refers to the story of the flying cow that Sofia Gustafsson: Järtecken. Joen Petri appeared in German newspapers in the Klint och 1500-talets vidunderliga lut- late 1990s. Here the entertainment value herdom. Nordic Academic Press, Lund in the form of fiction becomes promi- 2018. 272 pp. Ill. nent, and not the question of whether it is true or false. The supernatural beings The historian Sofia Gustafsson at that occurred in some older legends have Linköping University has carried out disappeared in modern legends. Since the an extensive study of a 400-page man- 1970s, locally oriented legends have also uscript in folio format compiled by come into use in the tourism industry. Joen Petri Klint. He was vicar of Östra The author also addresses the issue Stenby in Östergötland from 1567 until of fake news in modern legends pub- his death in 1608. The entire manuscript, lished in newspapers. In this case it is which is kept in Linköping Diocesan not the entertainment value that is cen- Library under class mark N 28, is now Reviews 167 digitized and available through Alvin but far from always. Many of these ref- (www.alvin-portal.org). Most of the erences are to German written sources text was written in the 1590s. Gustafs- from Protestant contexts. Among Swed- son describes this manuscript as Klint’s ish references we find Gustav Vasa’s järteckenbok or “omen book”. Omens chronicle and Olaus Petri’s chronicle. were believed to foretell future events. There are also references to 29 almanacs This book describes unusual celestial from 1543 to 1606. Handbills, mainly in phenomena such as solar and lunar German, constitute another source, and eclipses, blood-red moons, rings around there are many references to the Bible. the moon, the northern lights, comets, Gustafsson has expended a great deal meteors, dragons, visions of armies of effort to examine many of the source fighting in the sky, and rings or arcs texts to which Klint refers. In this way, around the sun. Unusual things allegedly she has had the opportunity to study observed on the ground include earth- how he used his sources. She finds that quakes, storms in the form of thunder, he made personal selections. His book hail, or strong winds, stagnant rivers, of omens can therefore be characterized skies raining blood, sulphur, or grain. In as a subjective compilation of mainly addition, there are various kinds of mon- written information from many differ- sters in the form of malformed children ent sources, along with some of his own or animals and prophetic infants. Sev- observations and experiences. eral beautifully coloured drawings made Klint’s overall interest was in find- by Klint are reproduced in Gustafsson’s ing a connection between the unusual book. phenomena, the actual omens, and con- Klint was well informed about polit- crete events that occurred afterwards. ical events in Sweden because he was This applies not least to political events a member of the diet in the clerical and wars. Raining blood could be seen estate. The late sixteenth century was as a sign of future bloodshed. Dragons a troubled time of war and struggles were associated with influential Catho- for the throne of Sweden, and with the lics who came to Sweden during Sigis- danger of the restoration of Catholicism mund’s reign and who were perceived as through the Polish king Sigismund. He evil. The Turks were likewise regarded reigned in Sweden after the death of as monsters. Klint gives no astrolog- John III in 1592 from 1594 to 1598. ical explanations and is sparing with Wars were fought with Denmark and apocalyptic interpretations about the Russia. Elsewhere in Europe there was last day being imminent, although such great concern about the Turks and their interpretations often appear in the writ- actions and battles against Christians, to ten sources to which he refers. Neither which Klint returns several times. God nor the devil is mentioned often in Gustafsson performs a broad contex- Klint, which was otherwise common in tual analysis by placing Klint’s man- written accounts of omens. Gustafsson uscript in the contemporary Lutheran points out that Klint does not preach in world of ideas, which was fascinated his book. Instead, she sees that his inter- with omens. She views Klint as a loyal est in omens was “more scientific than Lutheran. She also studies the question theological” (p. 220). of what he himself wrote and what he In conclusion, Klint’s book of omens borrowed from previous accounts of gives many insights into the late six- omens. Overall, there are few references teenth-century world of ideas and to eyewitness reports in Klint. In a dozen beliefs concerning what unusual signs cases he himself was an eyewitness, and and events were considered to predict so too were some people in his immedi- about the near future, especially in polit- ate vicinity. ical and military respects. Gustafsson Klint sometimes cites his sources, has carried out a very thorough study of 168 Reviews Klint’s book and has done a praisewor- gle with the papacy, which had its main thy job of placing it in its contemporary stronghold in Lund. political context. The book can therefore The cathedral chapter in Lund, which be warmly recommended to any reader owned many estates in Skåne, fought with an interest in cultural history and against Lutheran preachers who were folklore. active in Malmö from 1527. These were accused of abandoning celibacy, distrib- Anders Gustavsson uting both bread and wine at commun- Oslo University, Norway/ ion, and reducing the number of sacra- Henån, Sweden ments from seven to two. In Malmö, the first reformist preacher was Claus Mortensen (1499–1575). In 1528 he published the Holy Mass in Danish. Mayor Jørgen Kock was an Implementing the Reformation in important supporter of the Reformation Denmark in Malmö. In the years immediately Reformationen i Lund – Malmö – Köpen- after 1528, Malmö was the centre for the hamn. Anders Jarlert (ed.). Bibliotheca growth of the Reformation in Denmark. Historico-Ecclesiastica Lundensis 63. The theological inspiration in Malmö, as Lund 2019. 160 pp. Ill. in other places such as Viborg in the rest of Denmark, came not only from Martin The church historian Anders Jarlert in Luther but also from the Swiss reformer Lund has edited a volume of studies about Ulrich Zwingli. the implementation of the Reformation in The Reformation led to the gradual the kingdom of Denmark, which in the dissolution of the monasteries. At a sixteenth century also included Skåne meeting in Odense in 1527, a royal dec- with the cities of Lund and Malmö. The laration was issued giving monks and book is the result of a symposium in Lund nuns the right to leave the monasteries in 2017 to mark the 500th anniversary of and marry. The crown had thus broken the Lutheran Reformation. The authors with the papacy and the monasteries are church historians and other historians were transferred to national ownership who present different aspects of the Ref- by the king. The church ordinance of ormation process. 1537 stipulated that there should be no Lund, which had been the seat of the mendicant orders in the Danish king- Catholic archbishop since 1104, was dom. most reluctant to heed the reformers’ Many churches were demolished in calls for change. Things were different connection with the Reformation. This in the merchant town of Malmö, which affected Lund very hard. Stone from the had a class of burghers who were more demolished buildings had to be deliv- sympathetic towards the reformist influ- ered for the construction of Malmöhus ences from Germany. Trade contacts Castle, which was completed in 1542. with the German Hansa were extensive. The Reformation made it easier to Two printing presses were set up, which have divorces approved and to remarry. helped to spread the writings of the Ref- Marriage was no longer a sacrament as ormation. The Malmö Book from 1530, within the Catholic Church. A divorce written by the reformer Peder Laurent- case in Malmö in 1538 between the cou- sen, is an important source of knowl- ple Inger and Niels is studied in depth edge about the implementation of the in the book. Here the reader meets com- Reformation in Malmö. Copenhagen pletely ordinary people up close and can was the centre of the royal power, which see how they were affected by the Ref- saw in the Reformation an opportunity ormation. In 1529 marital law ceased to strengthen its position in the strug- to be a concern of ecclesiastical law in Reviews 169 Denmark and became a part of secular handy little book of 104 pages. Klint- law. Inger and Niels could thus finalize berg first has a brief 13-page introduc- their divorce. tion that puts the material he presents This book provides interesting in in context. Here he discusses beliefs sights into how the Reformation was and rites, the corporeality and material- introduced in the Danish Empire. It was ity of magic, the relationship of magic a process with a hard struggle between to animals and nature, and the fact that the papacy, the reformers, and the king. magical practice often involves doing It lasted for a few years from the late something in reverse. The major part of 1520s until the church ordinance in the book presents sources concerning 1537. The book can be recommended to magical practices from the post-Refor- anyone with an interest in church history. mation era, from the seventeenth century to the twentieth century. The exam- Anders Gustavsson ples that accompany the magical rites Oslo University, Norway/ are divided thematically according to Henån, Sweden who or what they concern. Thus death, omens, or illness and injury are themes that are dealt with separately, as are topics such as wild animals and livestock, or thieves, witches, and enemies. Each Bygone Magical Rites item is accompanied by a commentary Bengt af Klintberg: Vänster hand och stating the origin of the source, who motsols. Magiska riter från förr. Eller- recorded it, and where, if anywhere, it ströms, Lund 2020. 104 pp. has previously been published. Klint- berg also comments on the symbolism, It is rare for author presentations to be as he puts the practices into broader con- superfluous as is the case with this book, texts, and makes comparisons with sim- because the name Bengt af Klintberg is ilar practices or beliefs in other parts of not only widely known in Sweden but is the country. This way of situating the also well established in the other Nordic material is a useful tool for the reader, countries. In his home country his works but it is also essential to be able to place number nearly fifty books in the genres the sources in time, space, and social of poetry, children’s literature, and pres- context. entations of Swedish folk tradition, in In the introduction to the book, addition to countless academic articles. Klintberg writes that it can be read as In the field of Swedish folk tradition he a pendang or “counterpart” to his book has dealt with ballads, legends, fairy about magic spells, Svenska trollform- tales, rumours, rhymes, rules, and other ler, which was first published in 1965 popular narrative, he has covered magic, and has since appeared in several edi- witchcraft, and folk belief, and his pub- tions. It can thus be seen as a compan- lications have become bestsellers with a ion piece that is intended to complement broad and popular readership. The fact that succinct but fact-packed survey of that urban legends are known as “Klint- the cultural history of magic. Despite bergers” in Sweden testifies to the sig- the modest size of Svenska trollform- nificance and impact of his project to ler (the 1965 edition had 139 pages), it disseminate folklore. has a surprisingly thorough and detailed With this book Bengt af Klintberg treatment of this multifaceted cultural returns to one of the core topics of history and also provides a multitude of folklore where he started his folklor- texts illustrating magical practices. In istic career, namely, magic and folk that respect, it is far from lacking exam- belief. The title means “Left hand and ples of the practical and ritual aspects of widdershins: Bygone magical rites”, a magic, although this new book attaches 170 Reviews further importance to presenting and that museums seek? These questions commenting on the historical evidence are addressed in a recent volume edited for this. by Katja Lindqvist, who has also con- With Vänster hand och motsols: Mag- tributed no fewer than six of the book’s iska riter från förr, Bengt af Klintberg fifteen chapters. In her opening article, has once again given us a concise, sim- Lindqvist points out that kompetens is a ple, and well-written book that should difficult term to translate into English, have a broad appeal among his large the dominant equivalent in research and general readership. For professionals, practice being “human resource man- the book presents no new knowledge or agement”. Her main point is that there ways of understanding the material, but is an increasing demand for something that is obviously not Klintberg’s inten- other than traditional competences tion. The target group is not really an among museum employees – in addi- academic audience, who have read all tion to knowledge about collections and this before. As a long-time communica- related historical periods and cultures, tor of Swedish culture, the main thing museums and their staff must be capa- for Klintberg is to present important ble of dealing with inclusion, digitaliza- elements of this culture, conveyed with tion, economics and cooperation, among a playful ease that entices the reader to other things. follow him on the journey rather than The rest of the volume is divided into put the book down. This ability to pop- two parts. Chapters 2 to 6 deal with the ularize scholarship is commendable and needs of museums in terms of kompe- important, and it is essential that we in tens. Sofia Dahlquist’s attempt in chap- the Nordic countries ensure the con- ter 2 to outline future competences that tinued growth of this form of popular museums will need in order to stay rel- enlightenment in the future. evant, is, however, somewhat difficult to grasp: What “hybrid competences” in Ane Ohrvik “participation, dialogue, involvement, Oslo, Norway democracy and digitalization” are, and how they are to be achieved through competence profiles such aspedagogical curator, community manager and audience advocate, among others, remains Museum Competence somewhat unclear. The contribution is Kompetens i museisektorn. Politik, prak- based on a previously published report tik och relationen till högre utbildning. by a Swedish trade union from 2017. Katja Lindqvist (ed.). Nordic Academic In chapter 3, Cecilia Bygdell and Anna Press, Lund 2019. 320 pp. Hansen present results from a 2017 survey conducted among Swedish museums Museums are complex institutions, and by the Swedish network Forskning vid they are subject to continuous change. Museer (Research in museums), about New or revised acts on museums have research in museums. Their main con- been passed or presented by the parlia- clusion is, perhaps unsurprisingly, that ments in Finland, Sweden, and Denmark there is a need to develop and stimulate in recent years, and the pronounced the capability to do research in Swed- tendency towards centralization in the ish museums through better financing, sector is also found in Norway. In these enhanced cooperation, clearer pathways circumstances, what kind of capabilities of qualification or merit, and strength- to do things – kompetens – do muse- ened research infrastructure in museums need, now and in the future? Do ums. While interesting, a discussion of universities meet these needs by edu- what kind of competence research in cating candidates with the capabilities museums requires would have strength- Reviews 171 ened the text – particularly in the light of sity is presented by Richard Petterson in the discussion on knowledge production chapter 9. Carolina Jonsson Malm and in museums in chapter 7 (see below). Bodil Petersson present the offerings at Chapters 4, 5 and 6, all written by Katja Linnaeus University, campus Kalmar, in Lindqvist, chart differing competence chapter 10. The past, present and future and developmental needs in a number of of museology in Danish universities Swedish museums. Chapter 4 presents is discussed by Ane Hejlskov Larsen a 2016 survey – conducted, somewhat (chapter 11); the Finnish case differs confusingly, not by Lindqvist, but by somewhat from the other contributions Åsa Hallén on behalf of Länsmuseer- in this part, as Sirkku Pihlman explic- nas samarbetsråd – among regional and itly discusses in chapter 12 how recent county-level museums on their capa- developments in Finnish museum policy bilities in the light of perceived needs; impact on higher education in museol- the relationship between this survey ogy. Chapter 13, by Brita Brenna and and the present volume is not expli- Kristina Skåden, offers an insight into cated. Chapter 5 is a thorough investi- how the Department of Culture Studies gation on how four different museums and Oriental Languages at the Univer- in the Skåne region deal with capabil- sity of Oslo thinks about and collabo- ity and competence development. This rates with museums and other heritage 45-page, rather detailed, investigation institutions. Wrapping up this part, Katja is supported by 15 pages of diagrams Lindqvist summarizes Nordic educa- and figures. It shows, according to Lind- tional offerings related to museums qvist, that museums aim at developing in chapter 14. Although the chapter is their capabilities according to analyses more descriptive than comparative, she of what the surrounding society expects points out that the courses offered by from them. What museums need, the institutions of higher education seldom author says, is digital competence, com- are multi-faculty, thus not meeting what munication skills and employees that are museums according to Lindqvist ask for multicompetent. Chapter 6, which wraps – truly multidisciplinary competence. up the first part of the volume, presents What academia can offer is, rather, “a a number of different scholarly takes critical perspective on the museum insti- on “collaboration”, ostensibly aiming tution and cultural heritage work” (p. at enhancing collaboration between 298). museum employees, work groups and In the concluding chapter 15, Lind- museum organizations. qvist emphasizes that even though uni- The second part of the volume pre- versities actively seek to meet the needs sents more or less current offerings in of museums through active collaboration museology or other related topics from and produce candidates with knowledge institutions of higher education in Swe- and skills relevant for museums, they den, Finland, Norway and Denmark. The still fail in providing “multicompetence” emphasis is on rather detailed descrip- which is what museums truly need. tions of examples of collaboration While informative and mostly between universities and museums in well-written, the volume as a whole is the education of candidates. In chapter 7, somewhat uneven. Some chapters are Karin Gustavsson discusses knowledge analytical, others are very descriptive; production in Helsingborg museums and the Nordic perspective which is prom- the associated collaboration with Lund ised on the blurb is not systematically University. In chapter 8, Maria Brun- applied. Swedish cases dominate, and skog and Susanna Carlsten describe a the comparative discussion is rather lim- collaboration between Uppsala Univer- ited. The opportunity to critically dis- sity and Gotland Museum, while the cuss the various visions of what muse- museology programme at Umeå Univer- ums need – or rather, what museum 172 Reviews leaders on different levels say that they the universal priesthood. People assem- need, now and in the future – is not sys- bled in specially built prayer houses, but tematically seized; rather than a critical they could also attend services in the analysis of how the needs of museums state church. Sometimes there could be are defined and met, the needs seem to conflicts with state church clergy. These be taken for granted. could concern the right of the revivalists to administer communion in their prayer Teemu Ryymin houses. Bergen, Norway Folk high schools were established by the Inner Mission, inspired by the priest and popular educator Nikolaj Fred- erik Severin Grundtvig (1783–1872) in Denmark. Increased knowledge was Religious Faith in South-Western considered just as valuable as polit- Norway ical democracy. The revivalists also Tru på Vestlandet. Birger Løvlie, Per became involved in the struggle against Halse & Kristin Hatlebrekke (eds.). alcohol and in the nineteenth-century Cappelen Damm Akademisk. Oslo 2020. efforts, especially in western Norway, 417 pp. Ill. to have Nynorsk accepted alongside the Danish-influenced national language The texts in this edited volume are mostly Riksmål. by researchers who have been active at Although the book focuses on the Volda University College on the south- nineteenth century, there are some essays west coast of Norway. This is an area dealing with conditions in the twentieth that has been characterized by pietistic century, when the revival movement religious revivals since the nineteenth within the state church was declining in century, chiefly within the Norwegian strength. This was particularly the case state church. The purpose of this book in the very last decades of the twentieth is to reflect changes in various revival century. One local study concerns the movements from the nineteenth century Nordfjord Inner Mission Association, to the present day. Sixteen researchers founded in 1879 and disbanded in 2001. in the humanities, theology, and social There were revivals as recently as 1972 sciences have each contributed a chapter and 1973 in this area. During the 1990s to the book. there was a large exodus of young peo- The revival movements gained a par- ple, which led to the dissolution of this ticularly strong foothold in coastal areas association as a separate organization. in the nineteenth century and were often A new religious development has allied with temperance movements. A arisen in south-western Norway in the distinctive feature of pietism was per- twenty-first century. This has occurred sonal repentance. The collectivist religi- because of labour immigration from osity that had previously been typical of Poland and Lithuania and through the the Norwegian state church was replaced arrival of refugees from other countries by an individualistic religiosity. The per- where Catholicism is strong. At the end sonal faith that was embraced as a result of 2017 there were 7,557 registered of repentance had specific consequences Catholics in this region, representing for the convert’s way of life. Faith and 2.8% of the total population. As a result, life had to go together. One of the essays Catholics were the largest religious in the book, inspired by sociology of minority in the region. religion, problematizes how religion This book is of interest for the study was practised in everyday life. of folk religion and its changes in the Lay people could become preachers nineteenth and twentieth centuries from with reference to Luther’s doctrine of a regional perspective. It is also a contri- Reviews 173 bution to the research on Nordic revivals detailed biographical and micro-histor- that has been pursued in the Nordveck ical material. The diaries contain both network, which has published several joyful and problematic situations. The conference volumes this century. writers report both on what happens around them and on their emotional Anders Gustavsson reactions to what happens. The reader Oslo University, Norway/ gets a good picture of life at the mission Henån, Sweden stations, with the teaching, the medical care, and much more. Several missionaries translated parts of the Bible into the native language. The missionaries came from simple Swedes in the Congo circumstances, often smallholder house- Pia Lundqvist: Ett motsägelsefullt möte. holds, in different parts of Sweden, and Svenska missionärer och bakongo i they had taken a short training course at Fristaten Kongo. Nordic Academic a missionary school in advance. A small Press. Lund 2018. 330 pp. Ill. proportion of the women were teachers, nurses, or midwives. Several of the men The historian Pia Lundqvist at the had experience of various trades such University of Gothenburg has studied as carpentry and masonry. These prac- Swedish missionaries who were sent by tical experiences came in handy at the the Swedish Missionary Society to work mission stations. The missionaries were in the Congo Free State that existed young, usually between 25 and 30 years from 1885 to 1908. This state was the old, when they went out and were then personal property of King Leopold II mostly unmarried. This was true of both of Belgium, who made himself known men and women. If they survived, they for a reign of terror with gross abuses could later marry a fellow missionary. against the local population, until 1908 Once in place, missionaries needed to when it was transferred to the Belgian hire domestic porters as servants in State. The king’s intention was to obtain order to travel between different areas maximum economic profit from the and mission stations. Congo Free State, in particular through A universalist idea of a single human the export of rubber. Some Swedes race drove the missionaries to go out to served in this state as officers, sailors, preach the gospel. They felt personally and colonial officials. The missionaries called by God. An apocalyptic view was had some contact with them. In total, the also prominent, in the sense that Christ Swedish Missionary Society sent 124 was believed to be returning at any missionaries to the Congo from 1881 time. It was therefore urgent to pursue to 1908. Of them, 59 were men and 65 missionary work among the people of were women. Lundqvist concentrates on Africa. At the same time, the missionar- the cultural encounter between the mis- ies’ individualist outlook differed from sionaries and the Bakongo population the collectivist character of Congolese who spoke the Kikongo language. She society, with the village, the kindred, also considers the missionaries’ relation- and the family at the centre. ship to the state. A major challenge in Africa was the The main material for the book con- various febrile diseases such as malaria, sists of detailed diaries kept by four dysentery, and sleeping sickness. For Swedish missionaries, two women and much of the time, the missionaries suf- two men. In addition, there is the corre- fered from fever. Their time in the field spondence with the home country. The was never long because it very often author allows the reader to share in this ended with their death from some trop- 174 Reviews ical disease. Of the 124 missionaries Selma Laman’s unpublished notes enti- from the Swedish Missionary Society tled “Levnadsteckningar” (Biography). in the Congo until 1908, as many as 51 She was born on a farm in Östergötland died there. Their white skin could cause in 1862 and spent nineteen years in the problems because it frightened the Afri- Congo from 1891 to 1919. In 1893, she cans, who believed that the whites prac- married the missionary, linguist, and tised witchcraft or that they had arisen ethnographer Karl Edvard Laman. He from the dead and had turned white. translated the entire Bible into Kikongo. The whites were also thought to be able Selma left detailed accounts of illness to “eat people”, which meant that they and death among fellow missionaries could cause death among the natives. with whom she came into contact. In Africans were also afraid of white peo- 1899, no fewer than seven of the Swed- ple because many of the government ish missionaries died. In addition to officials they met were ruthless soldiers. malaria, a common cause of death was In the latter part of the book, the four sleeping sickness, which is described in diary writers are presented, each in a detail in the notes. The colonial violence separate chapter. The first is Johan Nils- is also visible in many notes. In contrast son, who was born in 1858 and went out to Johan and Mina, however, Selma is to the Congo in 1886 after studying for sparing with expressions of her own three years at mission school. The first feelings about what she experienced. diary was started in 1886 and began with The fourth set of diary notes was kept a long look back at the author’s difficult by Ivar Johansson, who was born in childhood and adolescence. His parents 1868 in a smallholder family in Värm- were poor and died early, and Johan had land. He came to the Kibunzi mission to move between different homes. In the station in the Congo in 1896 after stud- Congo he was given responsibility for ying for three years at the missionary the school at the Mukimbungu mission school in Stockholm. In Kibunzi, which station. The diary gives good insight had existed for ten years, there were into his changing moods and his sense about ninety African members and an of loneliness. In 1887 he moved to the infirmary that received many patients newly established mission station at every day. Discipline within the con- Kibunzi. He came back to Sweden once gregation was meticulous. Many Afri- but died in 1890 of a fever attack at the cans were expelled for offences such age of 32 and was buried in Mukim- as adultery or drinking palm wine. Ivar bungu. also got involved in building a new The second diary that Lundqvist mission station in Kikenge. He worked examines in depth was written by Mina with construction, medical care, and Svensson. She was one of the first three education before he died of malaria in female missionaries sent out by the 1899, at the age of 31. Like Selma, Ivar Swedish Missionary Society in 1888. portrays colonial violence in the form of They were to work among the women punitive expeditions, looting, and forced of the Congo. Mina, who was 27 years labour for the state. At the same time, he old on her arrival, served for just two had several contacts with the Swedish years before she died of malaria in sub-lieutenant Per Glimstedt. Lundqvist 1890. Like Johan Nilsson, she had lived believes that this reveals ambivalence in foster care from the age of two after among the missionaries about the colo- her mother had died. In the Congo, she nial power. It was a delicate balancing devoted herself to health care and teach- act to avoid unnecessarily offending ing, but she also did translations. The either the despised state power or the diary gives insight into her changing subjugated indigenous population. emotional states. Lundqvist has conducted a thorough The third special study concerns empirical investigation. She places the Reviews 175 diary entries in their social context both the print runs were highest. It is based in Sweden and in the Congo. Actually, on a large body of material, partly from the four case studies could have been the Royal Library (with a collection of published in the first part of the book about 15,000 broadsides) and partly and then analysed in the second part, from George Stephens collection in instead of the other way around as is Växjö Diocesan Library (2,100 broad- now the case. These narratives are filled sides). with various difficulties undergone by The overall aim of the study is to the diary writers. This applies both to gain knowledge of a repertoire of folk their problematic upbringing in Swe- tunes over a long period of time, a task den and to the serious problems in the involving certain difficulties. The mate- Congo. Problems included the deadly rial is searchable via ballad texts and fevers and the cultural encounter with statements of the associated tunes, but the African population, which turned out not via the tunes themselves. In addi- to be more difficult than the missionar- tion, far from all broadsides indicate ies had expected. They had to struggle what tune the ballad was sung to; it is against many prejudices about white estimated that only about ten per cent people. At the same time, they could of all broadsides during the nineteenth also establish friendly relationships century state the tune. And when only a with some Africans. This book makes limited share of the material is digitized, an important contribution to research it is even more difficult to say which about diaries in earlier times and about tunes were used most and in which con- cultural encounters between Swedes and texts. However, based on samples from Africans, even though Sweden was not a different collections and comparisons of colonial power. different prints, Ramsten concludes that the broadsides show a constant influx of Anders Gustavsson tunes and a rich repertoire. Oslo University, Norway/ To further illustrate this extensive rep- Henån, Sweden ertoire, four areas have been selected for closer study. The first chapter, “Tunes in numerical notation”, presents 57 broadsides, most of them from the same printing house during the period 1844–60. The Tunes to Broadside Ballads Unusually, these broadsides have a spe- Märta Ramsten: De osynliga melo- cial Swedish form of numerical nota- dierna. Musikvärldar i 1800-talets tion, from which one can play the tune skillingtryck. Svenskt visarkiv/Statens monophonically on an instrument. All musikverk, Stockholm 2019. 216 pp. Ill. 57 tunes have been transcribed by the author; few of the tunes were previously In this book, the title of which means known in broadside contexts. Through “The invisible tunes: Musical worlds comparisons with previous research on in nineteenth-century broadsides”, large ballad collections (with particular Märta Ramsten studies a large number mention here of studies by Hanna Ene- of broadside ballads, particularly focus- falk and Jan Ling), the author concludes ing on their tunes. The broadside, “this that tunes from the eighteenth century simple little printed sheet with ballad were current until the middle of the texts”, as the definition reads on the nineteenth century, but that many new very first page of the book, has a long tunes were subsequently added, often history in Sweden, from the end of the for use with sentimental love songs, reli- sixteenth century until around 1930. The gious revival songs, and in some cases study concentrates on broadsides from also with modern dance tunes, chiefly the nineteenth century, the period when polkas. 176 Reviews The second chapter, “Ballads of sol- “tune bank”, from which tunes could diers and war, 1848–65”, describes a be borrowed. This is done according category of ballads that were dissemi- to unwritten rules, and a tune can live nated during the two wars that Denmark for several centuries by being used in waged against Germany in 1848–50 and this way. The tunes must be regarded as 1864. The ballads can be clearly linked part of the “European musical cultural to concrete political contexts, according heritage”, although some of them have to the author. The texts therefore give gradually come to be perceived as “folk an idea of what was of topical interest ballads”. in contemporary politics, exemplify- The book is well written and clearly ing how broadsides could be used for outlined. Based on a large corpus of propaganda. Chapter three, “Broadside material, it gives us additional knowl- ballads and the theatre”, highlights the edge of both the repertoire and the use importance of the theatre for the distri- of music, mainly in nineteenth-century bution of music. After a brief presenta- Sweden. It is important for our knowl- tion of Stockholm’s theatres and their edge of musical life in previous centuries audiences during the nineteenth century, that the repertoire which enjoyed a wide the author presents a number of famous spread among the broad masses is also songs and discusses the relationship given its place in the history of music. between words and tunes. From a methodological perspective, it is The object of study in chapter four, also interesting to have concrete exam- “Lampoon and satire” is a special ples of how to work with large corpuses, phrase (called a timbre) that appeared in and here we see the importance of also numerous ballads through the ages, with making material available digitally. the text “All well – you understand me For example, the initial comparisons of well”. The oldest examples are from the tunes would not have been possible if mid-eighteenth century in ballad texts the opening lines had not been searcha- by Olof von Dalin and Bellman. Dur- ble in a database. ing the nineteenth century the phrase The four studies each examine the appeared in many ballads in plays and extensive material from a different broadsides, often with a satirical edge. angle, thereby illuminating the broad- Ramsten emphasizes the political allu- side repertoire in a varied way. The sions in the ballads and believes that it transcriptions in the comprehensive first is a different type of text from what is chapter provide much new knowledge usually expected in broadsides. of the subject: on the one hand, there is In a concluding chapter, the author a clear link between words and tunes, in draws general conclusions from her a way that it is rarely possible to obtain study concerning broadsides in the nine- for broadsides; on the other hand, we see teenth century. The proportion of newer that many of the broadside ballads also ballads is larger than previously known. use new types of tunes such as dance Broadsides were also used in other con- melodies and revival songs. The stock texts than is commonly believed, for of tunes thus develops over time, as this instance in connection with political chapter clearly illustrates. events and in teaching. That particular Similarly, the fourth chapter’s focus use has hardly been noticed at all by on a particular timbre offers a less previous research, Ramsten says, but common perspective by following a it is an important part of the history of detail in a ballad. It becomes an inter- broadsides. Moreover, broadsides were esting longitudinal section over a com- an opportunity for many people to share paratively long period of time, giving in cultural expressions. In conclusion, examples of concrete transmission of the author points out that one can view a specific phrase. The satirical charac- the tunes in the broadsides as part of a ter of the texts is highlighted, and the Reviews 177 author believes that they can be seen as includes, for example, two numbers reflecting contemporary liberal views. from Mozart’s The Magic Flute, which The studies on propaganda ballads and became popular and whose texts were theatrical songs give good examples of disseminated as broadsides (p. 113). the use of broadsides and also demon- Here it is not a question of texts which strate the author’s willingness to cite could be sung to any tune that fits, but examples of broadsides that expand songs that were originally composed our understanding of the areas in which with a specific tune. The question is broadsides occurred, while still dealing not entirely simple, as is also apparent more with familiar areas than the other from the author’s conclusion that cer- two studies. tain ballads almost acquire “the status The study is essentially descriptive of works and leave the unspoken prac- throughout, with the four studies pre- tice of broadsides” (p. 177). An overall sented in a similar way. A meticulous discussion of the concept of “broadside” description of the songs with commen- would therefore have been important, tary is followed by a reflective section preferably in combination with the pre- where conclusions about the repertoire viously mentioned observations about are presented. There is no direct prob- the authors. A minor detail in this con- lematization, but the concluding chapter text is that the book is dominated by raises several general questions that also examples from the first half of the nine- serve as examples of possible topics for teenth century, even though the stated further research. Here the discussion period is the entire nineteenth century. about the stock of tunes is interesting, All in all, the book provides many giving an idea of changes in the tune interesting insights into an extensive and material over time. There was a large widespread repertoire. The broadsides influx of new ballads, which challenges can be seen as “utility music” which the previously widespread notion that occurred in many everyday contexts broadsides are just “old tradition”. Simi- and was part of many people’s musical larly, it is important to highlight the way environment, and therefore deserves the ballads were used, although the link attention. with political aims is not substantiated to any extent in the study. It is also inter- Karin Hallgren esting that the question of authorship is Växjö, Sweden raised, as regards not only the texts but also the tunes. The descriptive approach also means that the very concept of “broadside ballad” seems to be taken for granted Remembering the Deceased and not really discussed. The defini- Å minnes de døde. Døden og de døde tion given is that it is a simple “printed efter reformasjonen [Remembering the sheet with ballad texts” (p. 9). Yet it is Deceased. Death and the Deceased precisely the tunes to these texts and after the Reformation]. Tarald Rasmus- the relationship between text and tune sen (ed.) Cappelen Damm akademisk, that is the focus of the study. Especially Oslo 2019. 263 pp. Ill. in the case of ballads performed in the theatre, this relationship becomes more When the Nordic countries embraced complex than in the rest of the material. Lutheranism in the sixteenth century, The examples from the theatre show that old vessels were only partly filled with certain texts were established for certain new wine. Only gradually did “pure” tunes, and that there may be texts with Lutheranism dominate along with the specially written tunes that were also growing custom of sending students of spread as broadsides. The latter category theology abroad, above all to Witten- 178 Reviews berg, Saxony, in order to participate in The content of the book is arranged higher education at the local university. according to the social classes in Nor- This book concentrates on how the wegian society. The higher the rank ceremonies around the belief in death of people in society when they passed changed from the ruling Catholic belief away, the more detailed were the pre- in the efficiency of requiem masses, scriptions and regulations for what was indulgences, purgatory, penance, and accepted, rejected or even demanded personal actions for a successful death at their funeral. The structure of these and a good afterlife into a piety centred chapters is very clear, which makes it around individual belief only. The book easy to see the difference between the consists of several articles in which the time before and after the Reformation. authors regard the Reformation as a The correct way to remember the king cultural process. The experts represent starts the series. The king was regarded church history, art history, and cultural as representing both mundane and celes- history, particularly folkloristics. The tial power. After the royal people come investigations are based on contempo- the noble men and women, who are said rary theological literature, administra- to take the place of the Catholic saints. tive regulations, legal records, funeral In every respect they were supposed to sermons, images, sepulchral tablets, live as models for the people. Even more and grave texts. The eleven chapters closely observed was the priest, for he concern Norwegian material but also had, so to speak, one foot in the world touch on Danish sources, for until 1814 and the other one in the sacred regions. the two countries were a united mon- In Lutheran effigies the priest carried a archy. book, a link to one of Martin Luther’s It turns out that dying is not easy. main thoughts, i.e., that the word and the There were particular conditions for belief in this word should be crucial in both the dying persons and the people Christianity. In the images the priest is around them to fulfil. The ars moriendi often surrounded by his wife and chil- (art of dying) was something everybody dren, both alive and dead members. This always had to bear in mind and con- demonstrates what an important role the stantly repeat in order to live in the right family played in Luther’s theology. way and then die properly. Handbooks It is clear that the predominant differ- in how to die correctly helped humans to ence between men and women is that the see that everybody had to take the step men are said to having brought about from life into death and that they should something, whereas the women were adjust to their individual future. The “this and that”, i.e., the texts describe dance of death painted on church walls, the women’s characters whereas they also in the Nordic countries, illustrate tell about the men’s actions. It would how death reaches all social groups. probably be fruitful to analyse the However, Lutherans and Catholics, and concept of fidelity, which is central in of course other people, cannot lead their Lutheranism, in the contexts of gender. lives uninfluenced by sin, as the - arti Perhaps the concept does not mean the cles very clearly demonstrate. The only same thing for men and women. To the way to a peaceful afterlife, according women faithfulness was, without doubt, to Lutheranism, is a firm belief in God. directed to her family. No actions whatsoever are efficient. There is comparatively little evi- This was made clear by the Reforma- dence left concerning the lower classes tion. Here I miss some short paragraphs of society, such as farmers and poor from folklore about the destiny of the people. However, the authors of these deceased without peace, to demonstrate articles are extremely skilful in finding what happened to those who did not ful- details that make the reading of these fil the demands. chapters very fascinating. A separate Reviews 179 chapter is devoted to women – how they In May 2018 Åmund Norum Resløkken grieved and how others mourned for defended his doctoral dissertation at the dead women – and it is obvious that the University of Oslo. The thesis discusses particularly feminine death in childbirth “classical” folkloristic concepts such was regarded as a virtue. Probably this as folk belief and tradition, and his dis- interpretation was created to console her sertation is an ambitious and valuable husband. Fortunately, this way of think- contribution to the history of folklore ing was erased in modern times. Also, studies, using contemporary research children and people without reputation methods inspired by ANT (Actor Net- are the subject of entire and detailed work Theory) and narratology. portions in this book. It is clear that, The specific theme of the dissertation until the end of the eighteenth century, is to investigate the construction of con- Lutheran belief from the time of the cepts like “folk culture”, “folk belief” Reformation in Denmark–Norway was and “objects of tradition” as they were strict, austere, and abstract in compari- expressed and understood in the ques- son to Catholicism. tionnaires distributed to a large number Today, when the assortment of belief is of informants through Ord og sed dur- huge, even fewer people stick to Luther- ing the period 1934–1947. Ord og sed anism. Consequently, I recommend this was inspired by Wilhelm Mannhardt’s book as an important introduction to collection from 1860 of customs related Lutheran belief, particularly the belief in to farming. Resløkken demonstrates death. I do so not only because it gives how Nils Lid, professor of ethnology a good overview of the issue, but also (folkelivsgransking), formulated the the because it is a functional model for how oretical foundation of the surveys, and to compare systems of belief. Moreover, how the questionnaires in various ways Lutheran believers would also do well became an important factor in the con- to read it in order to refresh forgotten struction of the folklore and the folk knowledge, but also as a repetition of belief the researchers wanted to collect how to deal with an issue that affects all in line with their theoretical principles. of us. Norwegians in former times knew Resløkken’s purpose is to demonstrate how to show their tribulations, virtues, how “folklore” was constructed through and hope for life after death in various the ways the questions were constructed, ways. Corresponding vital conditions rather than by the answers that were col- also touch modern humans. lected. The contributors to this book are Arne The writer unites aspects of research Bugge Amundsen, Eivor Andersen Oft- history, especially developed in the first estad, Tarald Rasmussen and Kristin three chapters, with a more theoretically Bliksrud Aavitsland. founded investigation of the textual construction of categories like “folk belief”, Ulrika Wolf-Knuts and the material and immaterial objects Åbo (Turku), Finland presented in the surveys. The empirical material is principally the texts collected through the surveys and introductions and questions addressed to the informants. A major question in Resløkken’s Tradition in Questionnaires study is “what makes certain objects Åmund Norum Resløkken: “Ein lut (and not others) into ‘objects of tradi- av det nære levande livet.” Tradisjon, tion’?” tradisjonselementer og tradisjonsforsk- Further, Resløkken demonstrates that ere. En studie av spørrelistserien Ord og the theoretical and academic frames for sed 1933‒1947. Oslo 07 Media, Univer- this type of academic work were not sity of Oslo 2018. 203 pp. Ill. Diss. nationally defined in a narrow way, but 180 Reviews were based on broad ideas about “folk to restricting the perspective and dis- belief” as an evolutionistic and anthro- carding certain aspects of the empirical pological category influenced by the material. The thesis takes a “deep-dive” German professor Wilhelm Mannhardt into one part of a broader communica- and partly by the Scottish social anthro- tion process or a more comprehensive pologist and folklorist James George circulation (the questionnaires), and the Frazer. The aim of the dissertation is to reasons for this are well clarified. How- demonstrate how the concept of tradi- ever, one objection to such a perspective tion was perceived, and how “tradition is that the thesis in a way closes in on objects” were constructed in the surveys. itself. As a reader one becomes curi- Through such perspectives Resløkken ous about what the respondents really succeeds in demonstrating new aspects answered and said, and where the exam- in the history of folklore studies. ples and narratives used in the question- Resløkken’s academic skills are well naires stemmed from. There must have demonstrated in the problem formula- been some intertextual relations in these tion and research questions which are formulations that are now cut off. Yet clearly formulated and focused. The another question is to what extent this application of the empirical investiga- material has been used by other research- tion is well done, and the organization ers, since Resløkken seems to conclude of the thesis is carefully considered. The that this material was not “scalable” for analysis of the sources is rich in details other purposes: it was not usable for and insights, demonstrating that the other researchers with other theoretical writer has a comprehensive knowledge assumptions and problem formulations. of the material. The analytical focus is It is unclear what the concept scale and on how “tradition objects” are created scaling imply in the thesis. However, in and formulated in the questionnaires spite of such objections, this thesis is a conceived as “texts”. valuable contribution to modern folklore Theoretically and methodically ANT studies and discussions of central con- is an inspiration for the empirical anal- cepts within this field. yses. The objects of research in the surveys are considered as actants in a net- Torunn Selberg work, and perspectives from ANT are University of Bergen, Norway connected to concepts and perspectives in narratology. This connection is partly taken as self-evident in the thesis – explicit justification for this is missing – and so is a clarification of the function of The Cultural History of the Goldfish concepts such as “actants”, “actor”, and Anna Marie Roos: Goldfish. Reaktion “narrative”. Other concepts central in Books, London 2019. 206 pp. Ill. the thesis, but not so well-known, such as “focalization” (fokalisering), “world- In preparation for the celebration of ling” (verdensliggjøring) and “limit nowruz – the Iranian New Year at the objects” (grenseobjekter) could profit- spring equinox – aquarium shops in ably have been more clearly and peda- Sweden stock up with goldfish. They gogically explained. Their theoretical have learned that Iranians in exile in and analytical role is not always self- Sweden buy a goldfish to have on the explanatory, and in some parts it makes table that is laid on the day before the the text inaccessible. celebration. A bowl with a living gold- In spite of this critique, a consistent fish symbolizes life and movement. The theoretical, methodological and com- goldfish (Carassius auratus) is a species positional grasp gives the thesis a solid that was domesticated long ago and thus framework, but it may also contribute is under human care, but when exactly Reviews 181 this happened in China is still unclear. the eighteenth century, goldfish started When the goldfish first came to Europe to become common in Europe. The his- is also a matter of contention. In 1665 tory of the goldfish in eighteenth- and the English diarist Samuel Pepys men- nineteenth-century Europe is discussed tions a fish that is believed to have been at length in the book. Here the source a goldfish. material is all the more abundant. Keep- In the now comprehensive and reading goldfish became a custom in well- able series of animal monographs from to-do homes. The goldfish attracted the the publisher Reaktion Books, it is the attention of many people, and in the turn of the goldfish. The historian of sci- illustrated zoological literature there are ence Anna Marie Roos, a professor at numerous depictions of the seductively the University of Lincoln, has written a glistening golden fish. handy little book about the significance For a long time goldfish were kept of the goldfish in cultural history and to in beautiful porcelain bowls made in some extent in the history of science. China. They came to Europe through The book opens with Carl Linnaeus, the East Indies trade. Linnaeus, too, who was very fond of goldfish. In 1740 apparently had a bowl, as did the royal he published a description in the pro- family. Several such imported porcelain ceedings of the Royal Swedish Acad- bowls are preserved in museums and emy of Sciences, accompanied by an collections. Goldfish bowls, however, illustration. This shows that there was a soon replaced the exclusive vessels from cultivated form with distinctive tail fins, China. It was not until the latter half of which apparently could already be pro- the nineteenth century that the goldfish cured in Europe at the time. Linnaeus’ became an aquarium fish. Nowadays exemplar had come via Denmark. In this the goldfish is a global concern that is paper he not only gives a description of spread as a companion and an ornamen- the fish, but also a guide to how to care tal animal all over the world. In some for it, although he had not yet had time places it has even become a nuisance, to acquire any personal experience of an invasive species which has found its keeping the fish in captivity. However, way into watercourses and competes he would have an opportunity to do so with native fish fauna. later, and he also expended a great deal Artists have always been attracted to of effort acquiring goldfish. His first the goldfish, and Sweden is no excep- meeting with the golden-yellow fish tion. We also find goldfish in poetry, and left him enraptured. Christopher Tärn- even Bellman felt the urge to write about ström was then commissioned in 1754 a goldfish he saw (“Impromptu on See- to bring home goldfish from East Asia, ing a Chinese Goldfish at the Home of but he drowned on the way and was Capt. R.”). The goldfish can still inspire unable to complete the task. Linnaeus art and literature. Through modern Japa- later received goldfish from England. In nese animated film, the goldfish Ponyo, several letters one can follow Linnaeus’ directed by Hayao Miyazaki in 2008, interest in goldfish. has become world-famous. Roos gives a broad account of the This book about goldfish is full of early cultural history of goldfish. Espe- exquisite illustrations. It also shows that cially in Japan, one can talk about a a seemingly insignificant domesticated highly developed goldfish culture, with fish contains a multifaceted cultural his- many fascinating expressions, for exam- tory for anyone interested in humans’ ple in the form of festivals, woodcuts relationships with other species. (ukiyo-e) and textile art, not to mention the skill in breeding peculiar shapes, Ingvar Svanberg which are sought after today and can be Uppsala, Sweden very expensive. From the beginning of 182 Reviews The Devil in Folklore to back. This was called the devil’s sab- Ebbe Schön: Ängel med bockfot. Båtdo- bath. kumentationsgruppen, Skärhamn 2019. Stories of everything being turned 124 pp. Ill. upside down emerged from interroga- tions, sometime with the use of torture, The folklorist Ebbe Schön has written which were reported in courts during many popular works on Swedish folk witch trials in the seventeenth century. beliefs and folk narratives in bygone The basis for the death sentences was times. In 2019, the publisher Båtdoku- that the women had made a pact with the mentationsgruppen in Skärhamn issued devil. The author refers to the historian his book Ängel med bockfot (Angel Göran Malmstedt’s study En förtrollad with Cloven Hoof), dealing with popu- värld, 2018 (reviewed by me in Arv 2018, lar beliefs and stories about the devil and pp. 212–214). In addition, Schön has used other dangerous beings in folklore. The legends recorded and preserved in the book is richly illustrated with beautiful folklore collection at the Nordic Museum, drawings and paintings by Bengt Arne which he headed for many years. Runnerström. In the folklore records we encoun- The author goes far back in history to ter stories of two erotic female beings, pre-Christian times, both in the Nordic namely the forest nymph (skogsrå) and countries and on the European continent. the lake nymph (sjörå). They looked This is intended to reinforce his basic beautiful from the front but tried not to idea that belief in the devil first came show their hollow backs and their tails. to Sweden through Christianity about a During the seventeenth and eighteenth thousand years ago. The Church’s belief centuries there were court trials in which in the devil as a thoroughly evil invisi- men were accused of having dealings ble power did not, in the author’s view, with these female beings, including sex- catch on among the common people. ual intercourse. They could even be sen- They did not think so much about what tenced to death. Here the author refers the devil could do to people after death, to Mikael Häll’s doctoral dissertation, but rather of how he could manifest him- Skogsrået, näcken och djävulen, 2013. self to living people. He could appear in Another dangerous creature was the concrete form, for instance as a magpie, water sprite (näck), who seduced people a dog, or a cat. He was not as terrible by playing his fiddle and luring them as he was portrayed in church sermons; into the depths. people could even deceive him or scare In ancient times, sorcerers could sign him away. This representation of pop- contracts with the devil. Some of these ular perceptions of the devil is remi- contracts, known as black books, have niscent of what the folklorists Ulrika been preserved for posterity. Statements Wolf-Knuts and Ülo Valk have found in about such material were also recorded Finland and Estonia respectively. Like- in eighteenth-century trials. In folk wise, the creatures of folklore whom the belief certain priests had the ability to Church regarded as the devil’s assistants fight the devil. This was especially true were not believed to be as dangerous as if they had received their training in Wit- the priests claimed. They could be both tenberg, Germany, where there was sup- good and evil. posed to be a school of black arts. The That the devil was a male being is Word of God was one of the best tools evident in the stories of the witches’ to use against evil. Norwegian variants meeting with him on the Brocken on of black books have been studied by the night of Maundy Thursday. They the folklorist Ane Ohrvik in Medicine, had sexual intercourse and celebrated Magic and Art in Early Modern Nor- atrocious orgies with him. Everything way, 2018 (reviewed in Arv 2019, pp. was topsy-turvy. People danced back 221‒223). Reviews 183 The final chapter of Schön’s book mellan visa och verklighet. Skrifter deals with what the sociologist and phi- utgivna av Svenskt Visarkiv 47. Gid- losopher Max Weber called “disenchant- lunds förlag, Möklinta 2019. 276 pp. Ill. ment”. In connection with increased popular enlightenment based on tech- “Crime sells!” Then as now. In the old nological development and progress days there were no true-crime podcasts, in the natural sciences, old folk beliefs crime comics, or tabloids covering faded away and gradually disappeared. sensational crimes. People bought bal- According to Schön, a backlash to this lads in simple folded sheets, known as could be the fascination with occult phe- broadsides, where contemporary crim- nomena in our time, although he does inals were a popular topic. The broad- not study this. Nor does he mention sides were an information channel at a present-day Satanism. He does, how- time when there was no news journal- ever, consider what he calls the “cuti- ism. Karin Strand has studied broad- fication” of old folk beliefs as they are sides from the Royal Library’s collec- no longer associated with danger and tion of “prints about misdeeds”. These are merely cute. The forest nymph gives are about serious criminals who were only pleasure and delight. The music of sentenced to death, and they were mar- the water sprite no longer tempts listen- keted on the day of execution when the ers into the deep water. In older folk- death penalty was carried out. Of 236 lore people both respected and feared known misdeed prints, the perpetrator is the brownie or tomte, before this figure a woman in 60 cases; two-thirds of these was transformed in the twentieth cen- are named and half of them were found tury into a wholly positive equivalent guilty of infanticide. The perpetrators of Father Christmas, aimed especially at are predominantly the mothers of the children. It is also worth mentioning the victims and the murders were commit- redefinition of the term witch, which is ted within the family. no longer necessarily frightening. The book presents five case stud- Finally: Ebbe Schön is a good story- ies representing the three most com- teller. He writes on a scholarly founda- mon types of infanticide. The common tion and constantly entices the reader to denominator is that the culprit is a follow along with the text. He brings his woman who was executed and that the presentation to life by recounting several crime was the subject of a broadside personal memories of growing up in a ballad. All the cases are from Stockholm stonemason’s district in Bohuslän. Even in the period 1773–1849. Karin Strand’s a book about the devil can be easy to overall aim is to investigate how the read, or as the title reads, the “Angel with broadsides about infanticide describe Cloven Hoof”, although the author does the crimes for contemporaries – and not discuss that term. It must surely refer why. To this end, the author reconstructs to a biblical and Christian conception of the underlying event and the persons the devil as one of God’s fallen angels. involved, using legal records, newspaper articles, and other sources, in so far Anders Gustavsson as they exist, to create a micro-historical Oslo University, Norway/ context. The narrative is interpreted with Henån, Sweden the help of research on infanticide and its perpetrators, the social and religious background, the judicial system and the legislation, the role of the death penalty, press history, broadsides as a genre, Infanticide in Ballads their form, archiving principles, etc. The Karin Strand: En botfärdig synderskas author has done an impressive job, but svanesång. Barnamord i skillingtryck unfortunately the detailed reconstruc- 184 Reviews tion and contextualization takes the repent but instead protests her innocence focus away from the analysis and it is as long as possible. difficult for a reader to hold the threads The second case study, “Murdering together. The themes and digressions to get away from life”, is about what is are numerous – admittedly interesting called suicidal murder – one instance and important – and the summary of the from 1774 and one from 1849. The book, entitled “Farewell”, is too general former tells the story of a woman who to do justice to all aspects of the con- killed her nine-year-old daughter with tents. an axe, a deed that gave rise to two bal- The first case study deals with the lads, one with a moral sermon, the other most famous form of infanticide, the one formulated as an ultimate warning. The that has come to be known as the “new” narrator proceeds from what is perceived infanticide which dominated from the as the moral degeneration of society – mid-seventeenth century until well into murder, fornication, thievery, fraud. The the nineteenth century. A “new” infan- explanation for the murder in the bal- ticide was the murder of a child con- lad is that the mother was possessed by ceived and born outside wedlock, where an evil spirit, she was transformed and the woman concealed the pregnancy, dehumanized and it was interpreted as gave birth to the child alone, killed it, lack of maternal love. The second sui- and hid the body. In the contextualiza- cidal murder was committed by a maid tion and analysis, the author has been against the eight-month-old son of her able to use the fairly abundant research master by sticking a piece of bread in that exists. The legal case that is used as his throat. The woman states in the court a case study is no different from other records that the cause was an “extreme similar cases, with the horrifying ingre- weariness with life”, which is not men- dients that these often have. The woman tioned in the ballad or in the newspa- was executed in 1734 and her case has per articles generated by the murder. generated two broadside ballads. The broadsides emphasize the bestial In the ballads the woman is portrayed deed in which the victim is an innocent as a “sinner” (in the female form syn- child and the perpetrator is – somewhat derska) whose soul is in profound dis- against nature – a woman, in some cases tress, and the women talk about their the mother herself. The ballads seize on “evil deeds” and compare themselves the incompatibility of murderous vio- to wild animals, reproaching themselves lence and motherhood. The desire to die for being “cruel as a tiger” and without was one of the most common motives “a mother’s heart”. The composition of for murder in Protestant Europe from the ballad follows the dramaturgy of an the seventeenth century to the mid-nine- ideal Christian preparation for death, teenth century and is believed to be due formulated in the first person from the to the heavy taboo on suicide and the perspective of the condemned prisoner. fact that suicide could not be forgiven The infanticide mother expresses guilt by God or repented in the same way as and self-loathing, and reflects on her murder. Suicide was decriminalized in wicked life. She seeks atonement and Sweden in 1864. begs for God’s grace. Jesus receives Murder for financial reasons illus- her with open arms and she is saved trates a third type of infanticide: delib- for eternal life. Towards the end of the erate neglect or murder of a foster-child, ballad, the listeners are encouraged to known in Swedish as “angel-making”. learn from the example of the contrite The historical background can be sought sinner and not wait until it is too late in the reforms introduced at the end of to improve their lives. Corresponding the nineteenth century, such as Gustav reflections occur less frequently in court III’s decree on infanticide (1778–1779), cases where the infanticide does not which was intended to reduce the infan- Reviews 185 ticide rate. It stipulated, among other broader trend in ballads about misdeeds things, that an unmarried woman had the over time: from a spiritually coloured right to give birth to her child in another narrative in the eighteenth century and place, to remain anonymous, and to the early nineteenth century towards a refrain from naming the father. Yet the more secular narrative in several differ- desired impact on society failed to mate- ent text genres and modes. rialize, and more children were born out What is told and how it is told can be of wedlock. related to the prevailing view of crime, Two ballads with a named offender guilt, and punishment, and the broadside are analysed, one from 1827 and one ballads reflect the development of crim- from 1849. One of these depicts the cul- inal law. As it became increasingly rare prit’s preparation for death, her profound both to pronounce and to execute death pangs of conscience leading to insight sentences, the ballads about executions and repentance, and the final joy over the disappeared too. The judicial system coming of the promised bliss. The con- played down the concept of sin in favour demned woman’s farewell tells of the of a rationally organized penal system, grace that awaits everyone as the Chris- and the Church evolved towards a more tian religion preached. The other ballad individualized faith. The process of sec- likewise moralizes about the depravity ularization can be followed in the broad- of the times, in a high-flown tone and sides, where the criminal is less often aiming its critique at the obsession with described as a sinner. The view of the money. The moral of the ballad is a more crime as an offence against God gave humanitarian approach to “illegitimate” way to a social understanding. Around children. The case is framed by a social 1900 there was also a shift in the con- understanding: the exploitation of fos- tent of the ballads, to focus on the vic- ter-children and, as the balladmonger tim rather than the offender. The murder viewed it, the increased dissolution of victim is named, but not the killer, on the morals. The broader social problems title page of the broadside. that were not often discussed in the The ballads about misdeeds gradually public sphere were projected on to the espoused the public discourse on crime infanticides: the widespread poverty, the and punishment from a perspective that unregulated fostering system, the vul- was assumed to appeal to the contempo- nerable situation of unmarried mothers. rary customers. In the eighteenth-cen- At this time the death penalty was being tury ballads about misdeeds, the crim- debated, and through their glorification inal herself speaks in the first person, of death, the broadside ballads confirm and gives an edifying testimony about the legitimacy of the death penalty, the the consequences of her sin – a repent- author notes. ant subject whom the audience can hold Unlike today’s true crimes, the early up as a mirror. In the nineteenth century broadsides give no concrete details, the ballads became secular moralizing psychological explanations, or exciting tracts, portraying the offender in the descriptions of the crime. The eight- third person, as an enigmatic or sinister eenth-century ballads say hardly any- deviant, as “the Other”. In the twentieth thing about the deed, only about the century the offender is less prominent punishment. The real fates and social than the victim, and it is the tragic end circumstances of the people involved of the victim, not the offender’s punish- are only hinted at. The ballads depict ment, that is the topic of the ballads and an inner drama: the repentance of the their main selling point. sinner and her prayers for the salvation In the concluding chapter “Farewell”, of her soul, formulated from the wom- Karin Strand highlights some of the an’s own perspective. Ballads and prose many themes covered in the text, but narratives about infanticides follow the here I would have liked to see a discus- 186 Reviews sion of what is promised in the introductions, ranging from Frida Johansson of tion: an illumination of issues of “power, Arjeplog, who was nine years old and representation, and gender”. already a skilled joiker, to more experienced joikers like Niila Ribbja of Jok- Inger Lövkrona kmokk. Tirén was determined to make Lund, Sweden a cultural contribution on behalf of the Sami, and the collaboration with Peter- son-Berger is one such example; the composer’s symphony Same Ätnam is based on Tirén’s recordings. Karl Tirén as a Joik Collector Ternhag is particularly interested in Gunnar Ternhag: Jojksamlaren Karl the collection trips in 1911–1915, the Tirén. Acta Academiae Regiae Gustavi most important years for the audio doc- Adolphi CLIII. Kungl. Gustav Adolfs umentation. As Ternhag points out, the Akademien för svensk folkkultur, Upp collection work was basically completed sala 2018. 116 pp. Ill. in 1915 and one can probably see that the time up until the publication of Die Gunnar Ternhag is one of the giants in lappische Volksmusik in 1942 may be Swedish research on joik. He has pre- regarded as the years when the material viously had this work about Karl Tirén was analysed. We should be grateful published by the now defunct Dialect, that Tirén used the phonograph, which Place Name and Folklore Archive in means that we can go back and listen to Umeå (DAUM) in 2000. The present the recordings. For me as a performer book is a slightly shorter version which of joiks, that material is of tremendous has not lost any important content. Tern- significance. As a teenager I was told hag sets the focus on the committed by the old people that “we have never work of Karl Tirén, the stationmaster, joiked here” (in Jukkasjärvi parish). It violin maker, and music collector, and is all the more interesting that the fore- I like the final words, in the imagined bears of these old people, perhaps even conversation between the two women their parents, are documented in Tirén’s standing at the phonograph with Tirén material. One of Tirén’s dreams was that and Maja Wickbom: “I can’t believe he he would rescue Sami music, and I agree made us joik.” “There was something with Ternhag when he writes that Tirén special about him, kind and determined would probably be perfectly satisfied at the same time.” “Yes, that was it” (p. with the development the joik has under- 101). A way to combine fiction with gone today. It is more alive than ever scholarship? before. As regards language too, we find How to evaluate Tirén’s material? that the joik recordings are historically The collected material, in the form of priceless for documenting old Sami wax cylinders, diaries, and musical expressions. They are also valuable for notation books, differs from the articles the landscape descriptions they contain. and columns that Tirén wrote. The latter Finally, I must mention a number of are more coloured by his times. Ternhag women who have been and continue asks: “The question is whether Tirén’s to be important for anyone who delves notation should be regarded as more or into the joik material. Both Karl Tirén less useless as source material for the and Gunnar Ternhag have needed the purpose of musical analysis” (p. 95). As assistance of knowledgeable helpers, regards the musical aspect, I am not the listed here in no particular order: Maja right person to judge, but it is extremely Wickbom Karl Tirén’s assistant, Maria valuable to listen to the stories that the Persson, married name Johansson, who joik recordings provide. Tirén recorded helped Tirén to gain access to the Sami and transcribed people of all genera- community, and Inger Stenman, the Reviews 187 first to present the joik material in mod- Turi met, because the latter offered Tirén ern times. Gunnar Ternhag has written help with the Sami language. In any a lyrical review (Rig 2019) of Barbara case, Gunnar Ternhag has done a solid Sjoholm’s work about Emilie Demant job of highlighting Karl Tirén’s work. Hatt. It seems strange to me that no one, We must spend a while together up on neither Tirén nor Demant Hatt, men- Nuoljalid where Tirén is buried, with tions that they met each other, which stunning views of Čuonjávággi (Lap- they really should have noted. We know porten), so that we can joik his luohti. from Tirén’s diaries that Tirén and Johan Thank you, Gunnar!

Krister Stoor Umeå, Sweden