The Ecology of Metre

Home , Fourteener (poetry), Metre (hymn), Metre (poetry), Shaft bow, Sikling, Tetrameter

The Retrospective Methods Network

Newsletter

A special issue of RMN Newsletter

Edited by Ilya Sverdlov and Frog

№ 11

Winter 2015/2016

RMN Newsletter is edited by Frog Helen F. Leslie-Jacobsen and Joseph S. Hopkins

Published by Folklore Studies / Dept. of Philosophy, History, Culture and Art Studies University of Helsinki, Helsinki

RMN Newsletter is a medium of contact and communication for members of the Retrospective Methods Network (RMN). The RMN is an open network which can include anyone who wishes to share in its focus. It is united by an interest in the problems, approaches, strategies and limitations related to considering some aspect of culture in one period through evidence from another, later period. Such comparisons range from investigating historical relationships to the utility of analogical parallels, and from comparisons across centuries to developing working models for the more immediate traditions behind limited sources. RMN Newsletter sets out to provide a venue and emergent discourse space in which individual scholars can discuss and engage in vital cross- disciplinary dialogue, present reports and announcements of their own current activities, and where information about events, projects and institutions is made available.

RMN Newsletter is edited by Frog, Helen F. Leslie-Jacobsen and Joseph S. Hopkins, published by Folklore Studies / Department of Philosophy, History, Culture and Art Studies University of Helsinki PO Box 59 (Unioninkatu 38 A) 00014 University of Helsinki Finland

The open-access electronic edition of this publication is available on-line at: http://www.helsinki.fi/folkloristiikka/English/RMN/

The Ecology of Metre is a special thematic issue of the journal edited by Ilya Sverdlov and Frog.

ISSN 2324-0636 (print)

ISSN 1799-4497 (electronic)

All contributions to The Ecology of Metre have been subject to double-blind peer review by two external reviewers.

Contents

Editor’s Note ...... 5

THE ECOLOGY OF METRE The Ecology of Metre: An Introduction ...... 6 Frog and Ilya Sverdlov Description of Poetic Form as a Tool for Stylistic Analysis of a Traditional Song Performance: A Case of a Western Nenets Narrative Song ...... 17 Jarkko Niemi The Metre of Yeats’s “Lake Isle of Innisfree” in Context ...... 32 Michael Cade-Stewart Formulaic Language in Minimal Metrical Requirements: The Case of Post-Medieval Icelandic þulur...... 49 Yelena Sesselja Helgadóttir Linguistic Multiforms in Kalevalaic Epic: Toward a Typology ...... 61 Frog Narratological, Metrical, and Syntactic Emphasis in the Old English Genesis A ...... 99 Ilona Paulis

REVIEW ARTICLES AND RESEARCH REPORTS Word Constellations as Tools in Skaldic Composition: Arnórr jarlaskáld and Óttarr svarti as a Case of Assimilation...... 111 Cole Nyquist

PROJECTS Personal Name Systems in Finnic and Beyond: Reconstructing the Concepts of Name Giving in Cultural Layers of Prehistory ...... 121 Terhi Ainiala

CONFERENCES AND EVENTS Old Norse Mythology Conference 2015: “Myth, Materiality and Lived Religion” ...... 124 Klas af Edholm “The Ontology of Supernatural Encounters”: The 4th Symposium of the Old Norse Folklorists Network ...... 126 Tommy Kuusela Austmarr V: “No One Is an Island” ...... 129 Kendra Willson

PUBLICATIONS Edited Books New Focus on Retrospective Methods: Resuming Methodological Discussions: Case Studies from Northern Europe ...... 133 Eldar Heide and Karen Bek-Pedersen The Viking Age in Åland: Insights into Identity and Remnants of Culture ...... 135 Joonas Ahola, Frog and Jenni Lucenius

RMN Newsletter submissions ...... 140

Editor’s Note

Since the founding of RMN Newsletter, oral the wide ranges of areas of potential interest poetry traditions and poetic texts as sources here. We hope that the concentrated attention have been addressed from different angles in given to metre and poetics in this special issue every issue of the journal. Fundamental to will stimulate attention to and discussion of poetry is, of course, metrics and the workings these topics, and the advance and refinement of language through which it manifests. of new approaches that are being developed. Formal aspects of poetics and how a poetic The articles collected in this volume system works are methodologically relevant present a number of new methods and to interpreting features of any particular methodological strategies as well as poetic source. Understanding the organizing theoretical perspectives that will be of interest principles and structures of a poetic system to a variety of scholars working with and its use in practice is a fundamental step retrospective methods. The Retrospective toward the distinction of meaningful features Methods Network (RMN) has also shown and variations. Without this perspective, a more general activity. The Austmarr Network researcher might easily project meanings and held its fifth annual meeting in Visby, interpretations on what, from the view of the Gotland (15th–16th October 2015), on the poet and his or her intended audience, might theme “No One Is an Island”, concentrating be nothing more than organic outcomes of on insular cultures in the Baltic Sea region. entextualization at the level of verbal texture The Old Norse Folklorists Network in its turn or an outcome of regular conventions held its fourth annual meeting in Tartu, independent of a poet’s intentions. The Estonia (10th–12th December 2015), this time Ecology of Metre has been organized as a taking a methodological emphasis on “The special issue in order to bring into focus the Ontology of Supernatural Encounters”. dynamics of such workings at the level of Reports on both events are to be found in the form. current issue. Plans are already in motion for Organizing a special issue on metrics and future events both organized by these poetics was proposed in 2014 by Ilya branches of the RMN and also independently Sverdlov, who advocated that this topic is as by the RMN’s members. significant as it is subtle. Sverdlov observed The positive attention to RMN Newsletter that although poetics may be addressed in itself and discussion of work published every issue of RMN Newsletter, it more often therein has also increased, for which we are than not remains in the shadows. thankful and appreciative, and do our best to Introductions to metrics seem earlier to have meet the interests and standards of our been more widely taught and even required in readership. We work to remain a venue for some institutions’ programs for folklore information and discussions that are carried studies and for some areas of philology. by many voices representing a variety of Although this seems to have waned from fields. That variety and the acuity, different areas of education, there is in fact a innovativeness and quality of work behind lively international discourse surrounding those voices is what gives life and richness to metrics and poetics. There also seems to be a this venue, and we look forward to the current rise in interest in the subtle interplays continued discussion by your voices into the of formal aspects of poetry with language in future. the communication of meanings and in the Frog structuring of images and motifs through the University of Helsinki conventions poetic representation. The Ecology of Metre only brushes the surface of

THE ECOLOGY OF METRE

The Ecology of Metre: An Introduction Frog, University of Helsinki, and Ilya Sverdlov, Helsinki

Metre is a living thing. It exists in and is system, lest their perspective become skewed. shaped by an environment that includes The present introduction reviews some of the language, genres and practices as well as oral central themes that form the cohesive web and/or written identity-bearing texts, and uniting this collection in order to make the these are reciprocally affected and shaped by dialogues between the diverse studies more metres with which they are connected and salient. When the complementary perspectives potentially also by those metres to which they comprising this volume are viewed together, stand in resonant, or perhaps contrastive the insights offered by each article is enriched relations. All of these features furthermore by the perspectives offered by others. exist and operate within broader environments of societies and cultures, linking them to The Ecology Model social roles, identities, classes and situations Applying a biological metaphor to expressive that shape how they are perceived, valued and forms affects how we think about them, and apprehended as meaningful. Taken together, thus is not neutral (Hafstein 2001). However, all of these factors, with their complex and the metaphor of ‘ecology’ has proven a often subtle relations constitute what we powerful and effective tool for conceptual- would like to call the ecology of metre. izing the dynamic interconnectedness of The articles assembled in the present individual aspects of culture, traditions and collection bring the ecology of metre into expressive media within a society and how focus, addressing it from a variety of they simultaneously become organized in complementary angles and disciplinary relation to one another. A paradigm known as perspectives with a range of methodologies. ‘ecolinguistics’ has evolved from the Discussions treat poetic forms from periods as approach to language as situated in a social wide-ranging as the Viking Age to some still and natural environment or ‘ecology of found today, and that occur in a variety of language’ (e.g. Haugen 2001 [1972]; see Fill languages including Old Norse and Icelandic & Mühlhäuser 2001). Research on folklore of different periods, both Old English and developed a corresponding concept of Modern English, Nenets, and a trans-lingual ‘tradition ecology’, with a more nuanced view Finno-Karelian poetic idiom. The genres on practices of which linguistic behaviour addressed extend from vernacular may only be one part (e.g. Honko 1981; 1985; mythological epics to Christian theological see also Kamppinen 1989: 37–46). Metre is poems and from skaldic praise poetry of royal here considered to exist and operate within courts to popular literary ballads. The strength just such ecologies, where it is a creative tool. of this volume lies precisely in the diversity Within that environment, metre simul- of these contributions that are simultaneously taneously shapes and is shaped by language, brought into dialogue with one another tradition, and textual environments with other through a number of common thematic topics co-existing metres. The ecology extends from and shared concerns. the worlds of texts and discourse to the social These works are united by addressing the environments in which these manifest. Texts metre of their sources as part of an overall and utterances are produced and received by system in which it lived. They underscore that people of different statuses, roles, ages, scholars should be consciously aware of this genders, and affiliations with all sorts of 6 groups and categories that can potentially granted with literary poetry, which is become associated with a metre and its produced for consumption from the printed situated uses. Ultimately, we can say that the page. In research on oral cultures, however, people who use, abuse, transmit, and recreate changing technologies increased the range of texts composed in a particular metre create a aspects of performance that could be relationship with that metre. That relationship practically documented. The performance- participates in how other people understand or oriented turn then brought the situated value the metre, while the use of the metre interactive aspect of verbal art into focus with simultaneously affects how the individual is the full spectrum of meaningful resources that viewed in society through alignments and a performer might employ, from changes in disalignments with that metre’s users and tempo and pitch to posture and positioning in uses. the performance space (e.g. Bauman 1984). Attending to the ecology of metre means Partly in response to the earlier norm of viewing metre as a phenomenon situated in reducing performances to text-scripts, new such an ecology. As a consequence, metre is ways were sought to graphically textualise the not viewed as an abstract and ideal metrical spectrum of meaningful features and template aloof in the ether; it is instead treated structures involved in expressions of verbal as inextricably engaged in relations with art (e.g. Hymes 1981; Fine 1984; Tedlock language, traditions and their uses in society. 1986). In the present collection, Jarkko Niemi Metre is seen as a living thing within an of the University of Tampere grapples ecology, where it interacts with and evolves precisely with this question of textualisation, in relation to other features and forms of but he also takes the issue a step further to expression that can equally be viewed as co- address the reciprocal potential of the participating in that ecology and its societal representation strategy as an analytical tool. and empirical landscapes. From this Niemi builds on the work of Lauri Honko perspective, recognizing and understanding (1998) with a perspective from ethno- this ecology is essential for the study of any musicology. He seeks to develop a dynamic metre, but also for the study of other model of textualisation that can offer a more inhabitants of that ecology that interact with holistic view of performance presented in a and evolve in relation to metres. written and graphic (print-publishable) form. This discussion is built on the analysis of a Textualisation and Analysis Nenets oral epic performance that he recorded Despite ongoing advances in technologies, at the end of the 1990s. (The Nenets are a textual and graphic representation in visual Uralic language group and shamanic culture media, whether printed or electronic, remain of Northern Russia.) The questions Niemi central to the study of poetry and poetics in addresses are not restricted to formal verbal art. This is somewhat ironic representation per se and their degree of considering that the poetics of verbal art are ‘exactitude’. Karl Reichl (2012: 9) observes aurally and temporally perceived. It is that “[o]ral poetry is as a rule sung poetry, nevertheless warranted in research. Such and yet few ethnographic transcriptions (and visual media give simultaneous access to even fewer medieval texts) record the music.” protracted sequences of text, the ability to Niemi, in contrast, gives attention precisely to move through a long text very rapidly and this acoustic domain. He is particularly with great precision, as well as provide the concerned with rendering variation in pitch potential to review that text at the rate desired across series of verses and the annotation of by the researcher rather than that of the performed time in a written form. performer. However, the textualisation of Performed time structures the rhythm of a poetry customarily reduces it to a verbal poetic metre in durationally-controlled time. transcript or text-script: out of the spectrum of The analytical grid on which Niemi maps features entailed in an individual’s sung verses reciprocally becomes a tool that performance, often only the lexical material is makes the metrical structure of the verses presented. This problem tends to be taken for graphically transparent. This method of

7 representation reveals the elocutionary have been transcribed by hand from the interface of language and melody or dictation of verses, excluding the euphonic intonational structure in time. Together, these verbal elements that are non-semantic yet form the sung metre of performance. The vital for the completion of metrical positions analysis also makes salient the role of in the rhythms of sung performance. Oral euphonic syllables that complete positions in poetry will customarily exhibit formal the performed rhythm, actualizing the differences when it is presented in a mode of metrically well-formed verses. The analytical dictation rather than in its conventional mode mapping of different features can then be of performance (see e.g. Ready 2015 and applied as a tool for stylistic analyses at works there cited). Differences like the multiple levels of realizing oral-poetic text in omission of non-semantic elements, as in the performance. The representation of verbal art dictation of Nenets epic, belong to the is itself an important tool for accessing metre, ideology of the text of the culture. The same which is not a phenomenon of language only. is true of the omission of corresponding By bringing the nuances of performance into “strictly performance-based features” (Foley focus, language becomes contextualized as 2000: 83) of sung performance by a one part of an integrated speech event that is transcriber raised within the tradition. Were it manifested by embodied individuals in time, not for the analysis of recordings of sung space, and an empirical situation. performance, the features of Nenets epic metre analyzed by Niemi and the relationship The Problem of Text-Scripts of words to metrical stress would probably Metre is a tool that manifests a perceivable remain a mystery. This has methodological quality of text – a quality dependent on implications for oral poetry that can only be articulation. This may be a vocal accessed through text-scripts. performance, which inevitably reduces all Even where text-scripts may be accurate to possible variations to a single – if transient – a sung metre, these alone may not reveal the acoustic form, or what Reuven Tsur (1992: metre per se. For example, a study by Jaan ch. 2) describes as a ‘mental performance’, Ross and Ilse Lehiste (2001: 116) led them to reading silently to oneself or going over lines argue that the metre of the Finnic alliterative of verse in one’s mind without necessarily tetrameter might never be accurately resolving all possible variations. Metres of understood without audio evidence. poetry known only through written text- Simplifying somewhat, this is because the scripts have been modelled on the basis of relationship between lexical stress and text corpora for more than a century. metrical stress simply does not ‘make sense’ Frameworks for such metrical imaginings from the perspective of Western poetries. The have in many cases been built on quantitative number of syllabic positions in a line and the analyses of lines of verse that seek to identify use of alliteration would be readily identified. patterns in linguistic features from which However, the metre has a dual parameter metre can be extrapolated. Eduard Sievers’ governing word placement according to a) 1893 study of Old Germanic alliterative lexically stressed syllables and b) the length metres, remarkable in its scope, is a classic of those syllables: long lexically stressed example. Yet Sievers also attempted to use syllables align with metrical stress while short musical notation as another means to stressed syllables contrast with it (see also understand Old Germanic metre (1912). This Frog, this volume). A systematic contrast of is because poetic metres do not exist in a lexical stress with metrical stress is counter- mathematical vacuum, and even knowing the intuitive to Western poetics and the dual number of syllables, feet, and so forth per line parameter (lexical stress + quantity) seems to rarely equals knowing what a given metre is be generally unusual (cf. Fabb 2009: 163). and how it works. The principles relating language to metrical The case in Nenets epic metre is interesting stress are so ‘foreign’ that it would quite because the majority of sources through possibly prove impossible to reverse-engineer which Nenets epic has been studied seem to them unless they were assumed as a

8 possibility in advance (and the role of culture. From this perspective, the accurate alliteration linked to lexical stress without representation of the oral metrical form in the regulation by the metrical template would transcribed text can itself be interpreted as likely only confuse investigators). reflecting a vernacular text ideology. Such an The problem of assumptions interfering ideology can be assumed to be behind the with the reception of the text-script of a poem writing out of rather than the omission of can even affect interpretations of modern expletive particles in Old Norse poems where literary poetry today. Michael Cade-Stewart these non-semantic particles are necessary for of King’s College London tackles the issue of the completion of metrical positions (Kuhn the metre of William Butler Yeats’s “The 1928). These textualisation choices may Lake of Innisfree”, which has troubled equally reflect ideologies of written text and metricists for decades. Like Niemi, Cade- how written texts were expected to be used Stewart turns to a sound recording as the key (cf. also Saarinen 2013). The investigation of to unravelling the mystery. Relying on a rare relationships between text-scripts and metre recording of Yeats reading this poem aloud, itself opens into questions of ideas and Cade-Stewart is able to demonstrate that the understandings behind the production of such problems in the metrical interpretation of the texts. poem are due to reading it in the wrong The independent durability of written text literary context. Relying on literary-critical allows text-script–metre relations to evolve history and an extensive electronic empirical over time as the same script is interpreted, corpus of 19th century English poetry, Cade- reinterpreted and poetic forms are reused and Stewart also restores the context to the poem reinvented in dialogue with the circulation of in question, which reveals a background in a written poetry. This is relevant not only to different metrical form than has generally modern poetry like that of Yeats which has been presumed. This background then been produced for circulation via the printed informs our understanding of its metre, an page. It is also relevant to revivals of understanding not accessible without moving traditions where there as been a discontinuity beyond the assumptions of a 20th-century in interpersonal transmission and people seek perspective. to reconstitute it from looking through text- The relationship between metre and text- scripts and perhaps consulting archival script is a vital site of research and recordings where these are available, as has exploration. The discussions of Niemi and been done with Karelian lament in Finland Cade-Stewart warn us of the limitations of (e.g. Wilce & Fenigsen 2015). No less retrospective methods concerning what can relevant is to consider aural traditions (see with certainty be reconstructed from corpora e.g. Coleman 1996). For example, Old Norse of text-scripts. They remind us that there may and Old English poetry seem to have had a be alternatives which have not yet been fully lively circulation in written form. Viking Age considered. The potential for the relation of and medieval Old Norse poetry seems to have text-scripts to metre to be subtle and even still received animated attention in Iceland deceptive does not, however, mean that a lack more or less until modernism encroached on of audio recordings is always and inevitably the society. Although the skaldic verse of an insurmountable obstacle. There is no sagas seems to have passed from the oral reason to assume that the Old English poem tradition, a tradition of being read aloud and Beowulf, for example, necessarily does not heard seems to have persisted much longer, accurately reflect the oral form of the metre, raising interesting questions about the aurality and therefore that its organizational principles of such verse as something experienced and necessarily cannot be inductively identified learned through performance, potentially with through analysis of the text. continuities going back to (and evolving A crucial point not often considered is the from) a time when such poetry was indeed an relationship between the representation of oral tradition. poetry and the understanding of what the The ecology of metre is a temporally ‘text’ is – i.e. the ideology of ‘text’ in the bound phenomenon, a phenomenon that is

9 continuously changing. Whole poetic systems 2015). They may also evolve highly idiomatic are in ongoing processes of evolution, quite and metaphorical systems of reference, as in often in an endless dialogue between Old Norse skaldic poetry (e.g. Meissner 1921) conservatism, innovation, and inevitable or Karelian laments (Stepanova A 2012; organic change. Poetic texts and the metrical Stepanova E 2015). The evolution of forms with which they are encoded may vocabulary is often linked to conventions of become suspended in writing, but also in an poetic form: poetic systems with an oral tradition, while language and societies organizing principle of alliteration are change around them. When addressing a text inclined to evolve equivalence vocabulary which has become suspended in this way, it is that allows ‘the same thing’ to be said with relevant to consider both the ecology of its different alliteration (Roper 2012); systems entextualization on the one hand and that of with an organizing principle of semantic its reception and reproduction on the other, parallelism may simply require ‘the same which could be quite different in terms of the thing’ to be said in at least two different ways meanings and associations with which they (Fox 2014). Nevertheless, the evolution of inform the particular work. diction remains concentrated within the subject domains and contexts of the poetry’s Metre, Poetics and Language use, leading the register of a poetic system to John Miles Foley (e.g. 1996) discusses oral be better equipped to say some things rather metre and language as existing and evolving than others in response to the social aspect of in a ‘symbiotic’ relationship. Asking ‘which its ecology in practice (Frog 2015). came first’ – metre or language – is a chicken- The structuring mechanisms of a metre and-the-egg problem that underestimates how also lead certain word combinations or both oral metre and language emerge and phraseology to become regular and evolve within an existing ecology. The predictable (reciprocally providing a model perceivable qualities of oral metres in for expression). This leads the lexicons of particular are rooted in and dependent on many oral poetries to be rich in fixed idioms features of a language’s phonology that can used in certain metrical positions be perceived and distinguished. However, as corresponding to the formulae investigated by those features of phonology change, so does Milman Parry (e.g. 1928a–b), the work at the the metre, and thus syllabic differentiations foundation of what evolved into Oral- made by poets in the Viking Age and much of Formulaic Theory (see Lord 1960; Foley the Middle Ages were no longer perceived in 1988; Foley & Ramey 2012). The quite th 18 century Iceland, and thus those restrictive view of the formula proposed by distinctions no longer operated in metres Parry at the end of the 1920s is now historically descended from that earlier considered far too narrow. Today, a formula period. Phonology within poetic diction may in poetry is more commonly regarded as a also shift, marking difference from verbal unit with a distinct meaning and/or aesthetically unmarked speech, as in Nenets package of associations, implications and so epic (Niemi, this volume). It is even possible forth (e.g. Foley & Ramey 2012; Frog 2015). for the symbiosis between metre and the In this volume, such language is discussed in diction of language within it to lead to the different forms by both Frog of the University maintenance of archaic inflectional forms and of Helsinki and Yelena Sesselja Helgadóttir whole grammatical cases that have dropped of the University of Iceland. Whereas Yelena out of other forms of speech (Nikolaev 2011). Sesselja Helgadóttir looks at formulae in Part of this symbiosis entails the evolution relation to variable metrical environments and of a type of language as the ‘way of speaking’ their relationships across poetries, Frog looks in the poetic form, which Foley (1995) refers at the systemic interaction of formulae in the to as its register (see also Agha 2007). Poetic formation of individual lines of verse and registers may be differentiated to varying sequences of verses. However, certain word degrees as part of their process of evolution, combinations can also exhibit recurrent use making the idiom more distinctive (Frog without reflecting a coherent unit of meaning;

10 they can instead have a functional role to meet across different genres (e.g. Turville-Petre metrical demands of alliteration and/or 1976: xvii; Frog & Stepanova 2011: 202; rhyme, as discussed in the contribution of Roper 2011a). Cole Nyqvist of the University of Oslo. Metre At least some of these differences seem to thus shapes use of language in the register, be connected with the mode or modes of while the register reciprocally evolves to performance. The performance mode may equip its users to meet the requirements of the yield a full actualization of the metre in metre – at least for the types of things that performed time, whereby the relationship users of the poetic form tend to say. between metrical and lexical stresses become Where formulae, verses and whole poems salient, as in Kalevala-metre. The become established in the tradition, historical performance mode may also organize the changes in the language may also affect their syllabic rhythm of a metrical line to align form. This may require a phrase to alter in metrical stress and lexical stress in performed order to remain metrically well-formed (Frog, time, producing syllabically longer lines with this volume). It may also be compensated for a distinct metrical structure as in the Nenets in performance (e.g. Lauerma 2004); for epic tradition discussed by Niemi, and yet it is example, reductions caused by syncope or only in performance that metrical stress seems apocope in a language may allow one syllable to operate in the poetry. The same also to be performed as two, in which case becomes prominent in some regions of the phonetically similar words may be treated as Finnic tetrameter (of which Kalevala-metre is metrically variable (Sarv 2015: 10; cf. Snorri the North Finnic form). In the region of Sturluson’s 13th-century description of certain Ingria, different modes of singing can syllables as ‘slow’ in a line that looks in its organize verses of the metre in a variety of textualisation to be one metrical position ways in performance. The Kalevala-metre short: Faulkes 1999: 7; cf. Helgi Skúli then moves to a deep structure as an Kjartansson 2009: 250–252 and works there organizing principle of language that becomes cited). The verbal frameworks of such verse a basis for the production of the rhythms and and poems can be capitalized on as resources structures of the various modes of for the formation of new compositions, as performance (Kallio 2013: 136–149). These Nyqvist discusses in the case of Old Norse cases can also be compared to the observable skaldic verse. On the other hand, these links difference between, for example, the poetic between language and metre are part of a metre of the libretto (read ‘text-script’) of much more dynamic process in performance some operas and its rhythmic-metrical form practice, as underscored by Niemi (cf. also when the verses are realized in performance. Lord 1960; Honko 1998). A fundamental When viewed in a broad perspective, the feature of all oral-poetic traditions is that they metre of linguistic text and the metre of its develop both linguistic resources and entextualization in a mode of performance can conventions for the production and be viewed as interfaced even as they operate reproduction of verse, and those resources and in parallel at different levels of the text. conventions evolve over time. The ecology of metre can have an extensive relationship to poetic features more A Metre or Metres? generally, or what Lauri Harvilahti (2003: 90– Metres are often treated as singular without 115) has discussed as an ethnocultural considering their relationships to other metres substratum of resources that provide the basis co-existing in the same ecology. This way of for traditional expression. The structuring of thinking is easy to slip into where a single this aspect of the ecology of metre also basic metre is used across a variety of genres, extends to cultural aesthetics. Such aesthetics as was the case with the Old Germanic seem to be responsible for the apparent long alliterative metre and the Finnic alliterative resistance of Finnic language groups to the tetrameter. However, even within these assimilation of rhymed poetic forms (Kuusi et examples there seems to have been some al. 1977: 56–57; Asplund 1994: 33–34, 801– difference in how metrical features operated 802). The same phenomenon seems to

11 underlie the addition of alliteration to metres Metre and Meanings assimilated into the metrical ecology of Metre is often viewed as a phenomenon of Iceland (Kristján Árnason 2011). The factors form only. However, Dell Hymes has affecting metres, their evolution or even reminded us that “how something is said is exclusion may be linked to quite broad part of what is said” (1986: 59): form can patterns in cultural poetics. itself be meaningful. For example, performing There often appear to be multiple poetic a story in an epic poetic form will lead an systems and metrical forms coexisting and audience familiar with that form and its uses interacting in a single metrical environment. to receive the story as being (or intended to This was clearly the case in Iceland, where be) an ‘epic’; or perhaps in relation to the already in the Viking Age and in medieval tradition of epics as a parody – but in any poetry there are not only a number of metres case, form is fundamental to meaning (Frog linked to different genres but even poems that 2016). The potential significance of form appear composed in multiple metres. In this comes to the fore in Cade-Stewart’s volume, Yelena Sesselja Helgadóttir restoration of the context of Yeats’ poem in addresses a post-medieval poetic tradition of 19th century English poetry. That frame of what are called þulur (sg. þula) that is rooted reference shows that the poem emerged and in a well-attested variety of Old Norse poetry. existed against a rich background of similar However, the evolution of this poetry has ballads ranging in content from serious to moved away from the more regular forms that folkloric and even parodic. That frame of appear linked to recognizable Old Norse reference shapes the understanding of Yeats’s metres in the 13th century. Rather than a work and the significance of its metre. regularly recurrent metrical form, performers Different Old Norse metres seem to have been could draw on a variety of possible metrical associated with different genres of practice templates as well as simply mark lines with (Gunnell 1995: 182–357). The metre known poetic features. The study of language in this as ljóðaháttr appears to have been poetry highlights the symbiosis of language characteristically used for direct speech so and metre: the regular co-occurrence of that the metre itself was reciprocally metrical form with formulaic expression associated with that mode of discourse (Quinn reveals an interrelationship of the choice of 1992; Smirnitskaya 1993). Metre can thus language with poetic form in the performance itself be a meaningful indicator of how text of the poetry. Links between this poetry’s should be interpreted and understood. formulaic expressions and those of other Variation in metre also holds potential for poetries reveal how the idiom and its potential marking text in meaningful ways. Variations metrical forms evolved in relation to other and embellishments that stand outside of the poetic traditions within the same ecology (see metrical structure of verse lines are also Haukur Þorgeirsson 2012). At the same commonplace in many poetries (e.g. Lord time, these connections, once established, 1960: 50–57). These need not have any direct seem to have remained distinct within the use relation to the content and may simply rather of each system; variation between them augment the acoustic texture and aesthetic of appears stable. In contrast, Nyqvist’s performance. In other cases, variations may discussion of skaldic diction suggests that Old be mobilized for emphasis, as in the Old Norse skaldic poets, who could compose in a English Genesis discussed by Ilona Paulis of variety of metres and switch between them, Radboud University Nijmegen. Paulis transferred word combinations for alliteration considers the creative use of metrical and/or rhyme across those metres more structures in Genesis A that allow its fluidly (see also Sverdlov 2011). These anonymous author to bring key points of discussions open new directions into looking certain episodes into focus. This sort of at language across metres in relation to the concentration of a metrical variation for practices in which they are used and the emphasis in Old English is also found with people who use them. so-called hypermetric lines. In the case of hypermetric lines, these patterns appear linked

12 to genre, since this type of use is found in regular metre. In his contribution to this narrative poetry, whereas hypermetric verses volume, Frog shows that the use of this in gnomic wisdom poetry are used with metrical form leads traditional units of greater frequency and flexibility but with less narration to develop verbal systems as serial concentration (Hartman 2014). frameworks for the production of short series However, it is not clear that the metrical of verses in performance. The analysis of structures discussed by Paulis serve any these units reveals them to be complex particular emphatic purpose by themselves in resources available to poets at a structural the corpus of Old English poetry, even where level above the formula, but smaller in scope they may coincide with this or that formula. than the unit of narrative content described as In Genesis A, on the other hand, the author a ‘theme’ in Oral-Formulaic Theory. applies them as a poetic tool by clustering Multi-verse sequences do not all carry them around certain narrative points. The poet equal weight or value within a larger oral text; thus capitalizes on potential metrical variation they may be hierarchically organized and conscripts it into service for a meaningful (Saarinen 1994: 183–184), and some function within the particular text. This sort of sequences may need to be made more example is particularly relevant within the prominent than others, as in the case ecology of a metre. The majority of discussed by Paulis. These units may also innovations applied by individual poets have play different roles in the structuring of a negligible or no impact on a tradition, others larger text and may be fundamental to can be taken up by contemporary and understanding how the larger text varies at subsequent poets, advancing to a feature of different levels in reproduction, both social tradition (see e.g. Stepanova E 2012: generally and in relation to specific social 280), which in this case could have been the situations and interaction (Frog, this volume). meaningfulness of a metrical variation. When considering prolonged performances of oral-poetic discourse, metre may come into From Verse Lines to Units of Larger Scope focus in a particular analysis, but it is The meaningfulness of a metrical feature imperative to consider metrical form as well discussed by Paulis relies on its concentrated and its features in relation to structures and recurrence across a series of verses. The patterns of larger scope (Paulis, this volume). resulting density of that feature across a This has of course been a fundamental basis sequence of text becomes a perceivable in the development of Oral-Formulaic Theory variation in verse texture. Other traditions of (e.g. ‘formula’, ‘theme’ and ‘song’ in Lord poetry involve formally structured units of 1960), an approach to verbal art that is larger scope than the line. For example, the significantly complemented by studies in the dróttkvætt metre is composed in couplets present collection. The contribution of Niemi linked by a phonic feature of alliteration, two brings in new tools and perspectives from couplets form a syntactic unit called a ethnomusicology. Frog introduces a helming, and two helmings form an eight-line compositional unit of intermediate scope to stanza. In the post-medieval þulur discussed the formula and theme. Paulis suggests a new, by Yelena Sesselja Helgadóttir, the choice of integrated methodological approach to or inclination to a metrical form or metrical observe the simultaneous interaction and feature can also be used to mark lines as systemic functioning of narratological, forming groups. Such groups of lines are not metrical and syntactic emphasis in poetic prescribed by the poetic form like stanzas, nor communication. Together, these contributions are they necessarily regular; the recurrent use highlight different aspects and varieties of of a metrical form or feature can simply be relevance of larger compositional units to used to mark verses as belonging together, for understanding a metre and how it operates. example as presenting a unit of narrative content. On the other hand, kalevalaic epic poetry is a stichic verse form (i.e. not composed in couplets or stanzas) with a

The Creative Individual in a Collective This sort of potential comes to the fore in Tradition revivalist uses of poetic forms which illustrate Metre is sometimes imagined as a sort of the social relevance of poetries from which cookie-cutter for language used to stamp out oral continuities are broken. Even medieval verses. Within this collection, in contrast, poetries may be filled with new value in later metre is regarded by the authors as a creative contexts, much as the medieval Arabic genre tool. It equips poets to perform within a of hija’ for cursing enemies recently tradition and simultaneously allows them to underwent a revival in conflicts in the Middle apply it strategically, as in Genesis A (Paulis, East for exchanges that included insults to this volume). The very fact that metre is Saddam Hussein (Bell 1997: 155). Of course, regular and predictable within a poetry allows such a revival necessarily entails new variations to be perceived as meaningful or meanings when the earlier tradition has aesthetically interesting. In the case of the become displaced from a living reality to þulur discussed by Yelena Sesselja become ‘heritage’. The potential for the Helgadóttir, metre becomes a choice and tradition’s meanings will also inevitably be device that continuously changes in the interpreted through the revivalist’s own evolving series of verses of an emergent contemporary frame of reference and ideology performance. In the more formalized (e.g. Wilce & Fenigsen 2015: 198–203). structures of skaldic composition discussed by Nyqvist, the verse form was a choice of the Perspectives poet – and presumably a strategic choice since The articles gathered in this collection vary in one of the aims of a Norse court poet was to emphasis from a focus on the relations of a be rewarded by a lord for his poem. Simply metre within its ecology to an emphasis on put, people ‘do things’ with metres not unlike other features which have evolved in relation they do things with words. to metre. On the one hand, all of these articles John Miles Foley (1995: passim) discusses might be said to concentrate on the tradition as an ‘enabling referent’ in relation ‘hardware’ of a poetic system – on its lexicon, to individual situated performance as an metre, syntax and structures. These are ‘enabling event’. Individual creativity matters often viewed as mundane and becomes perceived in the enabling event of technical, yet the discussions brought together performance in relation to the tradition as a here demonstrate that these ‘low level’ referent or frame of reference. ‘Classic’ Oral- structural elements are of key importance for Formulaic Theory (especially as outlined in a proper understanding of the larger purpose, Lord 1960) originally drew a great deal of message and meaning potential of poetic criticism for seeming to reduce oral-poetic expressions. These building blocks of poetry composition to a mechanical process that require framing in their ecology in order to be erased the poet’s potential for individual fully appreciated and understood. Moreover, creativity. This criticism was not always they and the texts produced with them have unwarranted, but it was largely a question of social lives that link them to people and roles orientation: much of the early research was in society both in the present and also in the oriented to work on structural elements of past, through which a metre can become a different scope that has only gradually connection. These papers illustrate that metre developed the rich infrastructures for looking is not at all limited to the number of syllables at formulaic units in different poetries. or lifts per line, but is part of the whole of Attention then shifted from looking at versification practice. Consequently, in order formulae as reducing the creativity of an to discuss a poetic text’s impact and import, individual’s word choice to viewing them as one should take into account – in the true ‘words’ of tradition with the potential to be spirit of interdisciplinarity – the metre and its used creatively. The same can be said of ecology, which contains valuable literary and metrical forms and their variations: they are cultural information unobtainable elsewhere. primed for the creative potential of individual One aim of this collection is not only to users. bring together a variety of complementary

14 perspectives on the ecology of metre, but to Semiotic Model”. In Genre – Text – Interpretation: stimulate attention to and discussion of this Perspectives from Folklore and Beyond. Ed. Kaarina Koski & Frog with Ulla Savolainen. Studia important topic, with the potential for further Fennica Folkloristica. Helsinki: Finnish Literature research to open the many questions that still Society. remain and lead research in new, fruitful Frog & Eila Stepanova. 2011. “Alliteration in (Balto-) directions. Finnic Languages”. In Roper 2011a: 195–218. Fill, Alwin, & Peter Mühlhäuser. 2001. The Works Cited Ecolinguistics Reader: Language, Ecology and Environment. London: Continuum. Agha, Asif. 2007. Language and Social Relations. Gunnell, Terry. 1995. The Origins of Scandinavian Cambridge: Cambridge University Press. Drama. Woodbridge: D.S. Brewer. Agha, Asif, & Frog (eds.). 2015. Registers of Hafstein, Valdimar. 2001. “Biological Metaphors in Communication. Studia Fennica Lingistica 18. Folklore Theory: An Essay in the History of Ideas”. Helsinki: Finnish Literature Society. Arv 57: 7–32. Asplund, Anneli. 1994. Balladeja ja arkkiveisuja: Hartman, Megan E. 2014. “The Form and Style of Suomalaisia kertomalauluja – Ballads and Gnomic Hypermetrics”. Studia Metrica et Poetica Broadsides: Finnish Narrative Popular Songs. 1(1): 68–99. Helsinki: Suomalaisen Kirjallisuuden Seura. Harvilahti, Lauri. 2003. The Holy Mountain: Studies on Bauman, Richard. 1984. Verbal Art as Performance. Upper Altay Oral Poetry. In collaboration with Prospect Heights: Waveland Press. Zoja S. Kazagačeva. FF Communications 282. Bell, Catherine. 1997. Ritual: Perspectives and Helsinki: Academia Scientiarum Fennica. Dimensions. Oxford: Oxford University Press. Haugen, Einar. 2001 [1972]. “The Ecology of Coleman, Joyce. 1996. Public Reading and the Language”. In Fill & Mühlhäuser 2001: 57–66. Reading Public in Late Medieval England and Haukur Þorgeirsson. 2012. “Poetic Formulas in Late France. Cambridge: Cambridge University Press. Medieval Icelandic Folk Poetry: The Case of Dewey, Tonya Kim, & Frog (eds.). 2009. Versatility in Vambarljóð”. RMN Newsletter 4: 181–196. Versification: Multidisciplinary Approaches to Helgi Skúli Kjartansson. 2009. “No Royal Road: The Metrics. Berkeley Insights in Linguistics and Extremes of dróttkvætt lines in Snorri´s Háttatal”. Semiotics 74. New York: Peter Lang. In Tonya Kim Dewey & Frog. Versatility in Fabb, Nigel. 2009. “Formal Interactions in Poetic Versification: Multidisciplinary Approaches to Meter”. In Dewey & Frog 2009: 147–166. Metrics. New York: Peter Lang. Pp. 247–257. Faulkes, Anthony (ed.). 1999. Snorri Sturluson, Edda: Honko, Lauri. 1981b. “Traditionsekologi: En Háttatal. London: Viking Society for Northern Introduktion”. In Tradition och miljö: Ett Research. kulturekologiskt perspektiv. Ed. Lauri Honko & Fine, Elizabeth Calvert. 1984. The Folklore Text: From Orvar Löfgren. NIF Publications 11. Lund: Liber Performance to Print. Bloomington: Indiana Läromedel. Pp. 9–63. University Press. Honko, Lauri. 1985. Rethinking Tradition Ecology. Foley, John Miles. 1988. The Theory of Oral Temenos: Studies in Comparative Religion 21: 55– Composition: History and Methodology. 82. Bloomington: Indiana University Press. Honko, Lauri. 1998. Textualizing the Siri Epic. FF Foley, John Miles. 1995. The Singer of Tales in Communications 264. Helsinki: Academia Performance. Bloomington: Indiana University Scientiarum Fennica. Press. Hymes, Dell. 1981. “In Vain I Tried to Tell You”: Foley, John Miles. 1996. “Guslar and Aoidos: Essays in Native American Ethnopoetics. Traditional Register in South Slavic and Homeric Philadelphia: University of Pennsylvania Press. Epic”. Transactions of the American Philological Hymes, Dell. 1986. “Models of the Interaction of Association 126: 11–41. Language and Social Life”. In Directions in Foley, John Miles. 2000. “The Textualization of South Sociolinguistics: The Ethnography of Slavic Oral Epic and Its Implications for Oral- Communication. Eds. John J. Gumperz & Dell Derived Epic”. In Textualization of Oral Epics. Ed. Hymes. Oxford: Blackwell. Pp. 35–71. Lauri Honko. Trends in Linguistics 128. Berlin: Kallio, Kati. 2013. Laulamisen tapoja: Esitysareena, Mouton de Gruyter, 2000. Pp. 71–87. rekisteri ja paikallinen laji länsi-inkeriläisessä Foley, John Miles, & Peter Ramey. 2012. “Oral Theory kalevalamittaisessa runossa. Tampere: and Medieval Studies”. In Medieval Oral Tammiprintti. Literature. Ed. Karl Reichl. Berlin: de Gruyter. Pp. Kamppinen, Matti. 1989. Cognitive Systems and 71–102. Cultural Models of Illness: A Study of Two Mestizo Fox, James J. 2014. Explorations in Semantic Peasant Communities of the Peruvian Amazon. FF Parallelism. Canberra: Australian University Press. Communications 244. Helsinki: Academia Frog. 2015. “Registers of Oral Poetry”. In Agha & Scientiarum Fennica. Frog 2015: 77–104. Kristján Árnason. 2011. “Alliteration in Iceland: From Frog. 2016 (in press). “‘Genres, Genres Everywhere, the Edda to Modern Verse and Pop Lyrics”. In but Who Knows What to Think?’: Toward a Roper 2011a: 123–140.

Kuhn, Hans. 1929. Das Füllwort of–um im & Dmitry S. Nikolayev. Moscow: Probel. Pp. 116– Altwestnordischen: Eine Untersuchung zur 121. Geschichte der germanischen Präfixe: Ein Beitrag Roper, Jonathan. 2012. “Synonymy and Rank in zur altgermanischen Metrik. Ergänzungshefte zur Alliterative Poetry”. Sign Systems Studies 40(1/2): Zeitschrift für vergleichende Sprachforschung auf 82–93. dem Gebiet der indogermanischen Sprachen 8. Ross, Jaan, & Ilse Lehiste. 2001. The Temporal Göttingen: Vanderhoeck & Ruprecht. Structure of Estonian Runic Songs. Phonology and Kuusi, Matti, Keith Bosley & Michael Branch (eds. & Phonetics 1. Berlin: Mouton de Gruyter. trans.). 1977. Finnish Folk Poetry – Epic: An Saarinen, Jukka. 1994. “The Päivölä Song of Miihkali Anthology in Finnish and English. Helsinki: Perttunen”. In Songs Beyond the Kalevala: Finnish Literature Society. Transformations of Oral Poetry. Ed. Anna-Leena Lauerma, Petri. 2004. Larin Parasken epiikan Siikala & Sinikka Vakimo. Helsinki: Finnish kielellisestä variaatiosta. Helsinki: Suomalaisen Literature Society. Pp. 180–196. Kirjallisuuden Seura. Saarinen, Jukka. 2013. “Behind the Text: Lord, Albert B. 1960. The Singer of Tales. Harvard Reconstructing the Voice of a Singer”. RMN Studies in Comparative Literature 24. Cambridge: Newsletter 7: 34–43. Harvard University Press. Sarv, Mari. 2015. “Regional Variation in Folkloric Meissner, Rudolf. 1921. Die Kenningar der Skalden: Meter: The Case of Estonian Runosong”. RMN Ein Beitrag zur skaldischen Poetik. Bonn: Newsletter 9: 6–17. Schroeder. Sievers, Eduard. 1893. Altgermanische Metrik. Nikolaev, Ilya. 2011. “Izhorian Epic Songs as a Source Sammlung kurzer Grammatiken germanischer of Linguistic Data: Synchronic and Diachronic Dialekte Erzänzungsreihe 2. Halle: Niemeyer. Perspective”. Paper presented at the Fifth Sievers, Eduard. 1912. Rhythmisch-melodische International Symposium on Finno-Ugric Studien: Vorträge und Aufsätze. Heidelberg: Languages in Groningen, the Netherlands, 7th–9th Winter. June 2011. Summary published: Nikolaev, Ilya. Smirnitskaya 1993 = Смирницкая, О.А. 1993. “‘Речи 2011. “Database of Izhorian Epic Songs Гримнира’ на эддической сцене’”. In От мифа к Morphology”. RMN Newsletter 2: 69. литературе: Сборник в честь 75-летия Parry, Milman. 1928a. L’Epithete traditionnelle dans Е.М.Мелетинского. Moscow. Pp. 257–274. Homere: Essai sur unprobleme de style homerique. Stepanova, Aleksandra. 2012. Karjalaisen Paris: Société d’Éditions “Les Belles Lettres”. itkuvirsikielen sanakirja. Helsinki: Suomalaisen Parry, Milman. 1928b. Les Formules et la metrique Kirjallisuuden Seura. d’Homere. Paris: Société d’Éditions “Les Belles Stepanova, Eila. 2012. “Mythic Elements of Karelian Lettres”. Laments: The Case of syndyzet and spuassuzet”. In Quinn, Judy. 1992. “Verseform and Voice in Eddic Mythic Discourses. Ed. Frog, Anna-Leena Siikala Poems: The Discourses of Fáfnismál”. Arkiv för & Eila Stepanova. Studia Fennica Folkloristica 20. Nordisk Filologi 107: 100–130. Helsinki: Finnish Literature Society. Pp. 257–287. Ready, Jonathan L. 2015. “The Textualization of Stepanova, Eila. 2015. “The Register of Karelian Homeric Epic by Means of Dictation”. TAPA 145: Lamenters”. In Agha & Frog 2015: 258–274. 1–75. Sverdlov, Ilya V. 2011. “Ok dulða ek þann inn Reichl, Karl. 2012. “Plotting the Map of Medieval Oral alsvinna jǫtun: Some Linguistic and Metrical Literature”. In Medieval Oral Literature. Ed. Karl Aspects of Óðinn’s Win over Vafþrúðnir”. Saga- Reichl. Berlin: de Gruyter. Pp. 3–67. Book 35: 39–72. Roper, Jonathan (ed.). 2011a. Alliteration in Culture. Tedlock, Dennis. 1983. The Spoken Word and the Houndmills: Palgrave MacMillan. Work of Interpretation. Philadelphia: University of Roper, Jonathan. 2011b. “The Metre of the Old English Pennsylvania Press. ‘Metrical’ Charms”. In Oral Charms in Structural Tsur, Reuven. 1992. Toward a Theory of Cognitive and Comparative Light – Заговорные тексты в Poetics. Amsterdam: North-Holland. структурном и сравнительном освещении: Turville-Petre, E.O.G. 1976. Scaldic Poetry. Oxford: Proceedings of the Conference of the International Clarendon Press. Society for Folk Narrative Research’s (ISFNR) Wilce, James M., & Janina Fenigsen. 2015. “Mourning Committee on Charms, Charmers and Charming, and Honor: Register in Karelian Lament”. In Agha 27–29th October 2011, Moscow. Ed. Tatyana A. & Frog 2015: 187–209. Mikhailova, Jonathan Roper, Andrey L. Toporkov

Description of Poetic Form as a Tool for Stylistic Analysis of a Traditional Song Performance: A Case of a Western Nenets Narrative Song Jarkko Niemi, University of Tampere, Finland

This paper discusses difficulties in capturing the entirety of an oral performance in a textual form. Proceeding from a sound recording of a performance of a Nenets epic, the author problematizes the issues of notating sung words, sung time, pitches, changes in pitch, pauses etc., and the issues of how such techniques affect the scholarly notions of text and metre. This paper addresses the problem of their traditions. This kind of researcher representing expressions or performances of position resembles that of an ethnolinguist, in auditory culture in textual and graphic form. the sense that the primary research data Of particular interest here is, first, the consists of recorded linguistic utterances and problem of the textualisation and structural speech acts (here: song acts). Learning a analysis of cultural expression: how does one culture and its ethnographic subtleties is, of identify and represent the elements which can course, indispensable for initiating and be interpreted as meaningful or prominent from sustaining the encounters with the the standpoint of tradition, or at least from the connoisseurs of local culture in the first place, perspective gained following the analysis of but the extensive ‘laboratory’ work is done the distribution of elements in the (recorded) with the data consisting of recorded and performance. Second, this article presents discussed song performances. some of these prominent features particularly from the standpoint of metrical analysis of the Problems of Textualisation and Structural verse-form text. Third, an attempt is made to Analysis describe graphically the totality of a lengthy From the point of view of the structural performance in order to highlight some of the analysis of sung performances documented in macro-level structural elements. audio recordings, the problem of Furthermore, as the framework here stems textualisation seems to be of a rather practical from ethnomusicology, the issue of kind – and, by no means, an unprecedented simultaneous representation of performed and one. The practical side of this concern is to intertwined language and sound structure is discuss the principles that provide a basis for relevant. In turn, this central issue generates how we define our level of exactitude in the several corollaries, as for example: how to process of the retextualisation of a cultural formulate principles of exactitude when performance when we are aiming to represent writing about performed sound phenomena, it in another semiotic form. In the academic and how to characterise conceptual world, this means, of course, writing and the borderlines between phenomena similar to use of graphic signs and symbols. This paper ‘phonemes’ and ‘sounds’ in language. discusses this issue using an example of a sung The example presented here is a recorded performance presenting a fully oral tradition performance of a Western Nenets narrative that has been documented through audio song, identified by the performer as a yarabts, recording. The performance in question belongs a song type about the hardships of the life of to a singing style in which the song forms the hero. The recordings were made in contain a full and dominant lexical content collaboration with the present author in 1994. (i.e. consisting of and dominated by linguistic For the present author, the Nenets culture words and not, for example, of euphonic, represents a remote one, although he has been metrical syllables, as in North Sami yoik), and doing fieldwork in various Tundra and Forest as a consequence it is reasonable to examine Nenets regions since the beginning of the the sung performance also from a linguistic 1990s, having become acquainted with the perspective. The crucial point here, however, ethnographic particularities, languages and is to take into account the performed time and dialects and having had long-term to try to figure out effective ways for informational relationships with several describing and understanding it. Nenets representatives and connoisseurs of

The formality and ‘grammaticality’ of this In modern cultural studies, the focus of kind of approach can also be thought of as a understanding human performance has shifted kind of methodological encouragement for the from the text-centred approaches of decades construction of a structural interpretation of past; and an emphasis on structure, text or any cultural data that is remote for the analyst. In linguistic homologies of cultural performance this kind of approach, an attempt is made to now requires a broader and more inclusive find the basis and evidence for a consistent view. We can argue to what extent structural interpretation by observing qualities performance exists in the heads of performers of elements of utterances. Analogously, it is, I in forms or modes translatable to textual believe, possible to observe, identify and even shapes and through particular kinds of action. understand the grammatical elements of a At this point, Dell Hymes’ words still language even if the observer is not fully resonate for me as a scholarly maxim familiar with the subtleties of semantics or the (reflecting his studies of oral folklore of cultural context of the language. In addition, Native Americans of the North Pacific coast): if we are aiming to interpret the structural When things were said or sung within the logic of performed cultural elements with native culture, explicit analysis – a detailed which we are not familiar, the possibility of a meta-language for dealing with form – was fruitful analysis may be derived from the not needed. Performer and audience shared possibility of observing how these cultural an implicit knowledge of language and ways elements function. If we consider ‘function’ of speaking. For us, there is no alternative to simply as a kind of interdependence ‘in explicit analysis. As with the grammar of action’, and if we can isolate our observations these languages, so with the verbal art: concerning the functioning elements (and, of underlying relationships, taken for granted course, their working together), we may be by their users, must be brought to light by able to arrive at an analytical conclusion conscious effort. Once brought to light, they can enable us to understand the creativity about their structural logic. and cogency of the discourse in which they Ultimately, this proposition is not occur. (Hymes 1981: 6.) surprising and is, in fact, quite similar to the work and ways of analytical reasoning In more recent writings, attention is required in, say, solving crossword puzzles. sometimes drawn to cases in which tradition- Although solving crosswords usually requires bearers have to adopt textual mechanisms in a full command of the language and its topical order to reutilise their own traditions when the expressions, the alternatives of reasoning and natural or traditional oral transmission is no extra- or interpolative clues can themselves be longer possible. These kinds of situations thought of as an analogy here. Thus, even the have been approached with the concept of more demanding types of crossword puzzles entextualisation (see Siikala & Siikala 2005, with minimal clues are solvable, if the analyst 58 about Polynesian cases), where the has intellectual resources for constructing possibility or inevitability of shapes of enough hypotheses concerning its elements – cultural performances are reconsidered. and particularly concerning their inter- In the structural understanding of a sung dependence. performance, the options for observable Combined with a paradigmatic method for elements are, perhaps quite obviously, language identifying and analytically presenting and structured time. While it is natural that all arranged elements of performance by their physically uttered human linguistic expressions identifiable, consistently recurring functional happen in time, sung performances usually appearances, this can, in my experience, yield consist of linguistic elements embedded in results which are helpful for making structured time. Moreover, while structures generalisations about the metrical structure of often evoke borders and repetition for entities a song.1 In some cases, this method reveals within the reach of human senses, the quite effectively some structural fundamentals possibilities of fruitful identification, concept- that may be central for the identification of ualisation, measurement and comparison also stylistic boundaries for cultural expression. increase. If this phase of analysis can be conducted reliably, we will be better equipped 18 to advocate for defining styles of sung The Study of Western Nenets Narrative Song folklore in ethnically or regionally definable Traditions: Historical Prerequisites local traditions somewhat analogically to The work to record Western Nenets sung isoglosses in areal linguistics. traditions was largely initiated by two Finnish The case study presented here concerns the linguists, M.A. Castrén and Toivo Lehtisalo. Nenets narrative song My Three Older Brothers Castrén started his first journey into the as it was performed by an elderly Nenets Nenets areas from the city of Archangel in woman Ekaterina Semënovna Kanyukova and July 1842. He gradually moved eastwards recorded by myself in the village of Nel’min- along the tundra villages in Kanin, Timan, Nos in August 1994. This was one of the last and the Malaya and Bol’shaya Zemlya tundra field recordings that I started analysing and regions. He worked with Western Nenets and this was in 2013. I then decided to write a Komi informants until he left the Western preliminary transcription of the song text Nenets tundras and arrived in the city of myself in order to anticipate the work-hours Obdorsk (now Salekhard) in November 1843. needed for the checked transcription and Thus, he worked roughly for one year among translation of the linguistic text. I asked the Western Nenets and wrote down their oral Anastasia Timofeevna Lapsui to assist me in folklore, songs and tales. There are altogether this work: although Lapsui is a speaker of the 19 text transcriptions in this Western Nenets more eastern Ob Gulf dialect of Nenets, we collection, and Castrén’s arrangement of the have worked fruitfully together in the past texts into metrical lines implies that they were with other song materials from the Malaya indeed performed in sung form. Some of the Zemlya Nenets. There is a difference between transcriptions contain a few hundred text lines; the Malaya Zemlya and Ob Gulf dialects, but some exceed 1,000 lines. (See Castrén 1940.) it has not been an insurmountable obstacle in Most probably, however, the process of our work. The analysis of the complete transcription included a phase during which performance (ca. 85 minutes) is ongoing, but Castrén worked with his Nenets informants we have now gone through almost half of it. after the sung form performance. It is, of However, the work completed thus far course, possible that he transcribed some of nevertheless illuminates the characteristics of the text lines during the actual performance the song as a verse-form, sung performance. and checked the missing parts later with the On the other hand, it is possible to present performer. This is the method that the Nenets some elements of the whole of the collectors use today, if they have no recording performance here, not at the level of the equipment available other than pen and paper. language, but at the level of the performed This, however, requires the transcriber to have sound and the quantitative features of the a good command of the language in order to verse-form text. be able to write lines down quickly. On the The opportunities to record long other hand, the Nenets performance style may performances of an oral tradition like this one also help the manual transcriber, because the among the Nenets have become quite rare. It Nenets trimetrical (hexasyllabic) scheme is is, unfortunately, possible that a major part of quite short and the meaningful text syllables the Nenets performers’ art is likely to be lost of a line are often concentrated at the forever, because the tradition and skill of the beginning of the line, whereas the endings performers will inevitably wane. This concerns may contain only the last few syllables and especially the survival of oral traditions among they may be given even longer durations of the Western Nenets. Examples like the one time (as, for example, in E.S. Kanyukova’s presented here seem to be solitary, surviving performance here). This gives the transcriber pieces, which do not help us to understand the valuable seconds to write down most, if not dimension of repertoires, at different levels of all of the lexical syllables of the text line. It is social and regional distribution of oral then quite easy to ask the performer for the traditions: the idiolects, dialects and words of any missing line endings afterwards languages of traditions, as outlined by Honko and it is thus possible to restore the actual (1998: 62; following Foley 1990). order of the performed lines and eventually

19 reach quite an exact transcription of the westwards from Pechora, he met an performance with a manual transcription. anonymous male narrator in the White Sea Furthermore, a teltanggoda was a person coastal area in the village of Nes and an assisting the singer by quickly repeating in elderly woman in the village of Mezen’ at the speech form the lines sung by the performer, end of the year. The anonymous man from and thus the message was enhanced by Nes sang him a lengthy heroic syudbabts of redundancy within the tradition. Pushkarëva more than 1,200 lines (perhaps a bit less, due (2000) has an illuminating simultaneous text to Lehtisalo’s convention of counting some transcription of a singer and teltanggoda’s separate line-initial, seemingly extra-metrical performance – although only as a written text, speech words as lines). Other performers also without any symbols of duration of time. In did well, singing songs labelled according to her examples, the places for the teltanggoda’s various Nenets genres (syudbabts, yarabts, repetition were more or less the same: the line khynts, yabye”ma) in quite extensive forms endings with just a few lexical – or only the reaching some 100–450 lines. non-lexical exclamatory – syllables. On the Lehtisalo in fact published one more other hand, in manual transcription it is very narrative song (of 323 lines) from a Western easy to lose the placement of the metrical Nenets performer, Katerina Vyucheyskaya, in code of the performed song, if the exact his last publication on Nenets folklore (1960: places and forms of the exclamatory syllables 374). Lehtisalo recorded the text in the village and the transformed sung syllables are not of Ust’-Tsyl’ma, although he reports that identified accurately enough. In addition, Vyucheyskaya was a speaker of Bol’shaya even if the metrical syllabic forms are shown, Zemlya dialect. it can be very difficult to understand their Lehtisalo was a pioneer with new function in performed, durational, rhythmical technology: he had a phonograph with him time, because the conventional symbols for during this fieldwork period of 1911–1912. linguistic text do not contain any indications Unfortunately, it seems the resulting for durations of time. phonograms were, to him, just an additional, Toivo Lehtisalo’s work during the last half experimental way of recording culture and his of 1912 to collect narrative songs of the real data consisted of manual transcription. Western Nenets parallels Castrén’s work, One reason was, no doubt, the uneconomical both in quality and in quantity. Perhaps even technical quality of the phonograph: it was more so in quality, because Lehtisalo could possible to record only a few minutes of rely on the progress made after Castrén’s sound on one wax-cylinder, which forced him times in Samoyedology and because he was to think of the phonograms as samples of more informed about the Nenets language performance sounds and not as a way of than Castrén was during his time. Lehtisalo recording a complete performance. Therefore, had already lived among the Ob Gulf Nenets it is almost tragic that alongside Lehtisalo’s for a year and by 1912 was already returning fantastic text transcriptions (1947), there are to Finland through the Western Nenets over 15 wax-cylinders recorded in the Western territories. Castrén had done the opposite in Nenets areas – but hardly any matches 1842, beginning his acquaintance with the between them. Some of the phonograms can Nenets and their language from the Archangel be identified as having been recorded in the region and moving eastwards to Siberia same places and having the same performers through the Western Nenets territories. as the published transcriptions, but due to the Lehtisalo (1947) recorded songs and tales poor documentation and sound quality of the from four informants. During the summer of phonograms, there are great difficulties in 1912, he met Timofey Taleëv and Anna utilising them to recreate the performance Taleëva in the villages of Oksino and Ust’e in styles of Lehtisalo’s informants. (On attempts the Pechora Delta area, in the neighbourhood to identify Lehtisalo’s materials, see Niemi of the then town of Pustozersk (now the with Lapsui 1995.) district capital Nar’yan-Mar) (for his travel After the consolidation of the Soviet account, see Lehtisalo 1956). Moving regime in Russia, the remote northern

20 territories became practically inaccessible for still hundreds of kilometres of stretches of foreign researchers. Russian researchers and tundra inhabited by people with their own collectors, such as Georgiy Prokof’ev, songs and stories, rich traditions that lived Grigoriy Verbov and Anton Savel’ev, started their oral life until the end. Sadly, however, doing fieldwork among the Nenets during the when the time came for a more encompassing 1930s. Unfortunately, these researchers all project to record and analyse the oral died during the war and most of their traditions of the Western Nenets at the end of materials are said to have been destroyed. the 20th century, most of the traditions had After the war, this work was continued by a already vanished. new generation of researchers, among them, Nataliya Tereshchenko, Lyudmila Khomich From Cultural Performance to Transcribed and Zinaida Kupriyanova, who collected and (Cultural) Text published Nenets folk traditions, including We naturally represent forms of cultural narrative songs, also among the Western performance through verbal texts, as this Nenets. In this context, the most famous method is so deeply rooted in our European publication is the collection published by literary-based culture. From this point of Kupriyanova (1965). In this collection, there view, it is very natural that throughout the are texts of 25 songs, all collected from history of folkloristics and the study of oral Western Nenets performers between 1939 and culture, the performances recorded and 1959. All the texts are published in prose transcribed as verbal texts have gained form, but it can be presumed that at least part autonomy and their face value almost of the performances were sung. However, the automatically (cf. Finnegan 1992: 17). songs reflect the classical themes of heroic Critically, since the ephemeral oral and fantastic syudbabtses and dramatic performance disappears into the past as yarabtses. They seem to be performed in full quickly as it does, we have to cling to the form and, at a rather rough estimate, the 309 trace it has left in transcription, however pages of Nenets song text could total about problematic, uncontrolled, biased or even 20,000–30,000 lines of verse-form text. Even misleading the working phases from more astonishing is the fact that the majority performance to verbal text may have been. of the songs recorded in Kupriyanova’s The most fruitful way to make a publication were performed by only one representation of a cultural text would person: Anastasiya Egorovna Taleeva, in the obviously be to observe and record the totality village of Nel’min-Nos in 1959.2 of the cultural cognition manifested in However, save for the traditional, generally performance by all the participants with all recurring themes in Nenets old narrative, such their resources for performance and reception, as the quest for a wife, tribal warfare, blood- including social, political, ritual etc. contexts. feud or mistreatment of an orphan in a rich However, as this seems to be a strategy not family, there are no closer thematic matches accessible for every situation, recorder, with respect to the subject matter of the earlier participant or researcher, it seems reasonable recorded narratives in Castrén’s, Lehtisalo’s to identify at least some aspects of this totality or Kupriyanova’s collections. Only some of recording cultural performance for closer surnames found only in these narrative examination. traditions (e.g. Tasyinyi) have survived in the During the 1990s, the Finnish folklorist tales of the last performers of our times. I Lauri Honko participated in a project aiming believe that, especially here, the reason is to achieve the total recording of a long epic in simply the lack of materials. As Lehtisalo the Dravidan Tulu language area of the (1947: xi–xii) observed, the time for a more Karnataka district of southwestern India. general folkloristic comparative analysis had Honko published this epic of almost 16,000 not yet come because the number of collected lines of text with an extended discussion about Nenets narratives was still so limited – understanding epic as cultural performance notwithstanding Lehtisalo’s own unique and text (1998). There are several conceptual, collection and that of Castrén’s. There were theoretical and methodological points in his

21 text that resonate with my experience and text can be ‘polished’ to achieve more clarity. understanding of the art of the indigenous However, without the participation in the storytellers of the Nenets North, although the transcription work by a performer or a person performances of the Tulu epic performer of her/his generation and a speaker of the Gopala Naika and the nearly ten years of particular dialect of the performance, it is work by Honko’s Finnish–Indian group of possible that we will never be able to clarify researchers to record, transcribe, translate – every single word of the text, in transcription, textualise – and contextualise this massive in translation, in denotation and in epic is, naturally, of far greater dimension connotation. There may also be a problem than the example of textualisation presented with transcribing dialectal or idiosyncratic here. As such, however, E.S. Kanyukova’s pronunciation. I have always preferred to skill and heritage seems to belong to the preserve the characteristics of the heritage of the great Malaya Zemlya pronunciation of the performer and not to storytellers like A.E. Taleeva, mentioned correct, for example, dialectal forms to above, although exemplified only with this conform with standard language. “Transcribe performance and the corresponding, as yet what you hear” was also the maxim of unfinished, process of its textualisation. Honko’s group (1998: 164). The entire text transcribed here is presented The English translation may seem a bit with the exactitude of orthography, not of awkward to read, but I have preferred, as phonetics, because phonetic transcription much as reasonably possible, to keep the would have required dramatically more detailed syntax of the translation as close to the syntax work. A motivation for an examination of the of the original in all my publications with phonetic level of sung language would have original language texts and their translations. been, for example, a more detailed study of Translating from one Uralic language into the possible correspondence of vowels with another (like from Nenets into Finnish, for different sonorities and varying pitch levels of example) presents fewer problems with the the melody. The reason for choosing language syntax than translating from Nenets orthographic exactitude corresponds to the into an Indo-European language (like English emphasis on working with syllabic rather than or Russian). This is also an approximation of phonetic elements in this kind of structural Honko’s (1998: 165) working rule: “as literal analysis, which aims at describing the as possible, as readable as necessary.” interrelationship of language and metrically organised melody at verse-level. In fact, this My Three Older Brothers transcription shows a bit more detail, because The central character in this song is a girl or those sounds or syllables that are put in young woman who first appears in a parentheses show, where possible, the syllable subordinate role in the household of her level phenomena that occur only in sung brothers. However, this girl becomes the language (additional euphonic, metrically central figure and heroine of the song. motivated syllables or reduced vowels). The Nevertheless, she is not identified in the song standard Nenets orthography reflects the with a personal name when she is introduced; spoken language quite well, but there are she is only identified as the one who has the some phenomena in pronunciation that the three brothers mentioned in the title of the orthography does not show. These are mainly song My Three Older Brothers. E.S. associated with sandhisation (merging of Kanyukova herself labelled this song as phonemes) or allophonic variation of yaràts – the Malaya Zemlya dialectal form of phonemes (as, for example, varying affricates the Nenets generic category yarabts [lit. /ts/–/ch/). In this particular text, these ‘weeping song; song about hardships of life’]. discrepancies are, however, quite marginal. Indeed by its content, this song could be seen The transcription and translation of the as reminiscent of a yarabts, the thematic original text can be seen as a continuum in dimensions of which are heavily inclined clarity: with rigorous re-reading and re- towards depicting the tough life and hardships checking, most of the unclear places in the of an orphan mistreated by a hostile family –

22 although the hero usually miraculously achieves he re-enters the scene of events and is alive victory and takes revenge on his or her enemies. again. The story is continued with a complex In the beginning of the song, the subordinate cycle of years of fighting, victory and slavery. girl is sent by her brothers to collect firewood The girl even ends up as a slave in the family from a nearby wood. She finds a suitable tree of the syudbya, who are giants living in a cave to chop down, but hears warnings from a red underneath a hilltop. sable and a crow who promise revenge if the The story culminates in a final and decisive tree is hacked down. These figures, however, fight between the families of Yabta Salya are not developed further during this song. Yerw and Medna Teta. In this fight, Medna After this introductory episode, the camp Teta is slain. It was at about this point when prepares for a wedding. Only at this point is Ekaterina Semënovna Kanyukova stopped her the name of the chief of the girl’s kin performance – at the 813th line. However, the mentioned, Yabta Salya Yerw. The girl story was not over. After almost 90 minutes understands that a rich chief of another family of singing, Ekaterina Semënovna Kanyukova of reindeer-herders, Medna Teta, is coming to was somewhat exhausted and, after she marry her. After the wedding feast, however, stopped singing, she explained that the end of the girl dresses as a man, ready for war and the story was near. She then finished the story action. Instead of agreeing to be the wife of in prose form: after the kin of Yabta Salya Medna Teta, the girl ends up in a confron- Yerw had beaten Medna Teta and his men, tation with him and succeeds in killing him. the three brothers and the girl with their From this point, the story adopts families ended up living the peaceful and supernatural elements. It turns out that the prosperous life of a reindeer-herding kin confrontation with Medna Teta is not over, as group.

Part of the Song Text of with Its Translation Nr. 93. Yaràts Nyakhar’ nyinyekaw – My Three Older Brothers. Perf. by Ekaterina Semënovna Kanyukova (née Wyucheyskaya), b. 1929 in the Malaya Zemlya tundra. Rec. by J.N., Village of Nel’min-Nos, 12.8.1994. Basic text transcription by Anastasia Lapsui (8; 12/2013); preliminary text transcription by J.N. (1/2013). The duration of the performance is approximately 85 minutes (including short breaks in the recording). (Introductory discussion between E.S.K. and J.N. in the beginning, omitted here. Short interpolatory dialogues between E.S.K. and the hostess of the house, Kseniya Filippovna Yawtisyi, during the performance are also omitted. Later during the performance, most of the spoken interpolations are single sentences by E.S.K. repeating her last sung text line, much in the manner of a teltanggoda, the traditional repeater-assistant of the performer in Nenets narrative performance.) Only the initial part of the recording (lines 1–41) is presented in this article. The entire text transcription with the translation is to be published later. Nyakhar(ey(ya))’ nyinyow(a)kawey, Three my brothers, nyakhar(ngey(ya))’ tetow(am)boyey... ng, three my rich ones, ter(i)(ngey(ya))khey(ya)wadow, reindeer they have got yu’ yoney(ya)ryey(yey)... ten thousand...

(Spoken interpolation.)

5. Tas yurey(ya) mya”mow, Altogether hundred huts we have, ew(ngey(ya))nandow(a)ngodey, that’s how it is, arka (nyey(ya) nya(kha)wowey... m, ey, the eldest brother, nyen Tawow(a)siyey, a woman Tawis nyudya yona... nyen Tawow(a)siyey, as his wife, a Tawis-woman, 10. syidya(ngey(ya)) yakhey(ya) nyudow, they have twins, syidya(ngey(ya)) naney(ya) nyude... two unmarried young ones... Tamna tanyow(a)nyiwey, Then there is also nyedkodow(a) nyudow, their daughter, nyud(a) tanyow(a)nyiw..., their small one they have...,

15. yud(a)m potow(a)kotsyo... a ten-year-old one...

(Spoken interpolation.)

Tyikakhad syani(yey(ya)) ebatow... ow, And after that, on one day nyakhar(ngey(ya)) nyinow(a)kawey, three my eldest brothers, nyakhar(ngey(ya)) tetow(a)mboyey, three my rich ones, pido” ng... ob(a)(ngey(yang)) ganey(ya) nyingey, they, in one (men’s riding) sledge 20. pido” ngamandiy(ya)wedey, they, it seems, sat, pido” ma... mow(a)now(a)ndow... n...: they speak (I hear): – Yalya(n) tyuku... tyukow(a) yalyow... m, – On this day (today), nye papow(a)kowow, our younger sister, poyormanz’ khantanow(a)kiyey she sets off to get firewood 25. chyon(a)tow(a) pya...(?) for all the huts...(?)

(Spoken interpolation.)

Tyikakhawow(a)khanow, Now pido” mamow(a)nondow”: they speak: – Tyun(a)” taryakhambyey’(?), – If the marks on the wood are correct, pida lus yan’ khantanow(a)kiyey..., she is seemingly to go to the twisted cape...,

(Spoken interpolation.)

30. tyun(a) taryakhangey(?), if the marks on the wood are correct(?), pida lus yan’ khantanow(a)kiyey... n. she is seemingly to go to the twisted cape. Tyikakhad ma”nyiw’... – Ti podyer(a)makey, Then she said (thought)...: – They harnessed the reindeer, tu... ti tow(a)kiyey... they already drove the reindeer...

(Spoken interpolation.)

Tyikakhawow(a)khanow, Now (then) 35. khor(a)tsam(a) nyudya nyinyiw(a)kawey, the youngest brother,

(Spoken interpolation.)

nyudya nyinyow(a)kawey, the youngest brother nyum tyaduw(a)lawow, swatted the doorway fur-hide (open), pyirtsya li(ya) malaney, at all his tall height, ta li mala... laney yes, at the height of his size 40. mya” tyuney... and went inside the hut...

(Spoken interpolation.)

Tyikakhawow(a)khanow... Now (then)...

(Spoken interpolation.) ------(break in the recording) ------Charting Some Structural Dimensions to show in an abstract way the peculiar feature Figure 1 represents the entire song of rising pitch levels during this 85-minute performance as a succession of lines, as song performance. shown in the box below. In this box, the lines The tonal range of the basic pitch level of a are represented with both black stripes and the melodic line in the song, ranging from d1 to h1 white spaces between them. This is a (again, aided by both black and white technical solution with which it was possible symbols, each characterising a musical to create a string of 813 somehow visible interval of a half-step) is shown in the left symbols fitting in one line of this page. The side of the box in Figure 1. The basic pitch stripes and spaces are small in this picture, level is identified as corresponding with the and the zoomed image above shows a section basic tonal level of the melodic line, because of them in more detail; above that, their it is the one which dominates, more or less, correspondence with the numbered the whole melodic structure. This basic pitch (melodic/text) lines is shown with musical level acts as a kind of ground level for the transcription. The aim of this representation is melody, from which the melody leaps 24

Figure 1. The graphic scheme of the entire performance of E.S. Kanyukova. upwards and to which it eventually descends. with the actual original pitch level of the There is nevertheless an ornamented final audio document of the performance. motif at the end of each line, where the However, during the performance, this melodic motif descends, actually, below this basic pitch level starts to rise gradually, due to level (see c#1 in the music transcription), but the physical and mental strain of the for a very short time. The basic tonal level of performer. This is shown with the graph lines the melody is exemplified by the final tone of in the box in Figure 1 as a kind of function the fourth line in the music transcription, between the number of the song line and the which shows the first four lines of the song. corresponding actual pitch level of the basic Thus, during these four lines, the basic tonal tonal level. Thus, the whole melody rises, as level keeps to e1. The term ‘pitch level’ refers if almost unnoticed, also raising all of the here to a physical fact that can be measured as other tones of the melody. In this kind of free, the actual vibration of sound (in Hertz), soloistic and unaccompanied musical whereas ‘tonal level’ refers to the abstraction expression, where the singer does not have to of the melody represented by the symbols in synchronise or calibrate her expression with musical transcription. These symbols are other voices or instruments, she has the full tonal steps (proportionate with each other), freedom to represent the elements of the like the aforementioned initial tonal level e1, melody in the form and exactitude that she which is chosen here to correspond roughly has become accustomed to. This freedom

Figure 2. Two melodic lines before and after the last transposition of the basic tone in the performance of E.S. Kanyukova. creates an intriguing phenomenon: with the that the raising of the pitch line during a song gradual ascension of the basic pitch level of performance could be associated with the the recurring melodic lines, the leaps of the anxiety of the performer and that this, in turn melodic line also tend to expand. Thus, the could be explained by the possible ritual upward leaps, shown in the first four lines as connections of this kind of performances in leaps of a major third (e1–g#1) or a fourth (e1– Nenets culture. Thus, she suggests that the a1), expand – embedded within the overall ‘super periods’ may be reflections of a kind of ascension of the pitch level – to a fifth performance trance. (Gomon 1990: 83, 85.) (proportionate e1–h1). Thus, as the overall This leads to the question of how to pitch level and the basic level of the melody construct a visual and symbolic representation rises, there comes a point when the singer from this kind of fact of the performance. One approaches the upper limit of her own solution is to make musical transcriptions physical tonal range. When she feels she corresponding to the physical rise in pitch and cannot sing the highest tones of the melody to show this as exactly as possible with any more, when the pitch level of the melody musical symbols. This is what I have has risen to its limits, she brings the pitch demonstrated in the musical transcription in level down again. As the graph shows, this Figure 2, presenting two melodic lines before melodic transposition happens six times and after the last transposition (between lines during this performance. 715 and 716), without any spoken The Russian musicologist Alla Gomon also interpolations, boundaries between episodes discerned this kind of phenomenon in song or sentences in the story – or any other break recordings that she collected in the 1970s by the singer. The first two lines of the (incidentally, also from the Malaya Zemlya transcription show how this singer has Nenets), calling these passages ‘super reached a1 as the pitch level of the basic tone periods’. However, she made her conclusions and thus the upper limits of her tonal range mainly using musicological argumentation, (of good quality of sound) and how the with seemingly less ethnographic evidence of upward leaps of the melody have also reached discussing their possible cultural functions or the interval of a fifth (a1–e2 in the meanings in the totality of the song transcription), even showing signs of further performance. However, she also noticed that expansion (a1–f#2). After the 715th line, she these structures seemed to happen transposes the pitch level of the basic tone independently of the structure of the content down to d1, giving her, once again, periodic of the performance. She concluded further relief for her vocal cords.

Another possible solution is to regard the sense, because it has only one type of melodic rising pitch level as a constant and to present line throughout the song. In other songs, there only a proportional, sketchy structural may be a recurring combination of two or transcription of an ‘average’ form of the three different melodic lines. We could call melodic line (as the four melodic lines in this kind of structural or schematic musical Figure 2). This is, I think, possible if we have transcription a ‘metrical’ one. Here, for reason to maintain that the average example, we can try to grasp the overall idea transcription fulfils the needs of representing of the basic metre of this song and proceed to the entire song. Whether this is really so observe how the song text fits this basic depends, naturally, on the research question structure, as illustrated in Figure 3. or the motivation for presenting musical In the above scheme, the musical symbols expression in this kind of graphic form. represent approximate (and proportionally The gradual rise of the basic pitch level is recurring) durations of musical tones with only one perspective for understanding this which the melody is sung. It suffices here to kind of song performance. Alongside this, understand the opposition of short durations, however, it is possible to explore the marked with the symbol for 1/8 note ( )  technique of performance. We can ask, for against long durations, marked with symbols example, whether the lowering of the basic for 1/4 ( ), 3/8 ( .) and 4/4 ( ) notes. The    pitch level in any way corresponds with paradigmatic representation of the syllables longer breaks during the recording of the on a metrical grid is designed here to show performance or with the singer’s frequent the tendency of the syllabic grouping within interpolations of speech during the the line, but also to give an overall performance, most of which were single proportional idea of the distribution of the sentences repeating her last sung text line, respective durations of these syllables, much in the manner of a teltanggoda, the performed in musical, metrical and therefore traditional repeater-assistant. In Figure 1, I durationally controlled time. The three have marked vertical lines showing the four vertical lines delineate the three basic sections long breaks in the recording. From these, it is of the text metre, each containing two lexical possible to deduce whether, for example, a syllables and thus forming the hexasyllabic break in the recording regularly corresponds basis of the text line. The metrically with the transposition of the pitch level. Here, additional, separate euphonic syllabic it seems that only the last of the four clearly elements appear in parentheses. It is this co-occurs with the transposition of the pitch totality of the hexasyllabic basic text line, level. It is not the purpose here to give a embedded into the sung metre of the definite answer to the dynamics of this particular song, with which the sung and phenomenon, but according to experience performed Nenets metre can be described. with other Nenets sung narrative The hexasyllabic basic text can have a wide performances, the phenomenon of the gradual range of realisations in terms of the sung rise of the overall or basic pitch level in these metre. Some of the songs only repeat the styles seems to be more like an autonomic, hexasyllabic basic text line, while some are physically motivated phenomenon: it just decorated with complex formations of seems to happen, regardless of other breaks in supplementary metrical elements. However, the performance, or perhaps even regardless in the Nenets song, the final syllables of of the structure of the episodes of the story. words are also subjected to euphonic All in all, the average transcription is used transformation in sung forms, for example: -a here as an example, with which it is possible → -ey; -a → -ow; -ya → -yow (see Hajdú to understand the distribution of durational 1978). However, I have not bracketed them units, together with the fundamental dimensions of the melodic lines, which are because they are not additional in terms of systematically present in every line of the syllabic count. This paradigmatic arrangement song. Ekaterina Semënovna’s performance presents the functioning of the text line within adopts an especially ‘basic’ form in this musical and metrical time, and also the

Figure 3. A schematised illustration of the metrical basic structure of E.S. Kanyukova’s song performance, with an example of the text lines from the beginning of the song to the first break in the recording (lines 1–41). Key: = 1/8 note; = 1/4 note; =    3/8 note; = 4/4 note. Note that in the presentation of this scheme, the syllable duration simply continues to the next syllable and the time values can easily be read from the line of the note symbols. The breaks of the syllable duration are shown with three dots (...).

                             1. Nya - kha - r(ey - (ya))’ nyi - nyo - w(a) - ka - wey, 2. nya - khar - (ngey- (ya))’ te - to - w(am) - bo - wey... ng, 3. te - r(i) - (ngey- (ya)) - (te) - khey - (ya) - wa - dow, 4. yu’ yo - ney - (ya) - ryey - (yey)... 5. Tas yu - rey - (ya) mya” - mow, 6. ew - (ngey- (ya)) - nan - do - w(a) - ngo - dey, 7. ar - ka (nyey- (ya) nya - (kha) - wo - wey... m, ey, 8. nyen Ta - wo - w(a) - si - yey, 9. nyu - dya yo - na... nyen Ta - wo - w(a) - si - yey, 10. syi - dya - (ngey- (ya)) ya - khey - (ya) nyu - dow, 11. syi - dya - (ngey- (ya)) na - ney - (ya) nyu - do... 12. Tam - na ta - nyo - w(a) - nyi - wey, 13. nye - d(a) - ko - do - w(a) nyu - dow, 14. nyu - d(a) ta - nyo - w(a) - nyiw..., 15. yu - d(a)m po - to - w(a) - ko - tsyo... 16. Tyi - ka - (nge) sya - ni - (yey - (ya)) e - ba - tow... ow, 17. nya - khar - (ngey- (ya)) nyi - nyo - w(a) - ka - wey, 18. nya - khar - (ngey- (ya)) te - to - w(a)m - bo - yey, 19. pi - do” (ng... o - b(a) - (ngey- (yang)) ga - ney - (ya) nyi - ngey, 20. pi - do” nga - man - diy - (ya) - we - dey, 21. pi - do” ma - m... ma - mo - w(a) - no - w(a)n - dow... n...:

22. – Ya - lya(n) tyu - ku... tyu - ko - w(a) ya - lyow... m, 23. nye pa - po - w(a) - ko - wow, 24. po - yor - manz’ khan - ta - no - w(a) - ki - yey 25. chyo - n(a) - to - w(a) pya... 26. Tyi - ka - kha - wo - w(a) - kha - now, 27. pi - do” ma - mo - w(a) - non - dow”: 28. – Tyu - n(a)” ta - rya - kham - byey’, 29. pi - da lus yan’ khan - ta - no - w(a) - ki - yey..., 30. tyu - n(a) ta - rya - kha - ngey, 31. pi - da lus yan’ khan - ta - no - w(a) - ki - yey... n. 32. Tyikakhad ma”nyiw’... – Ti po - dye - r(a) - ma - key, 33. tu... ti to - w(a) - ki - yey... 34. Tyi - ka - kha - wo - w(a) - kha - now, 35. kho - r(a) - tsa - m(a) nyu - dya nyi - nyo - w(a) - ka - wey, 36. nyu - dya nyi - nyo - w(a) - ka - wey, 37. nyum tya - du - w(a) - la - wow, 38. pyir - tsya li - (ya) ma - la - ney, 39. ta li ma - la... la - ney 40. mya” tyu - ney... 41. Tyi - ka - kha - wo - w(a) - kha - now...

28 tendency for the non-lexical elements to regularly only of the euphonic (non- group into certain parts of the line and their additional) transformation of the final syllable tendency to form additional syllabic units (left without brackets, as explained above). within the text line. There is one more peculiar point, which It is possible to grasp the idea of the may give an additional clue in the organisation of the syllabic elements of the interpretation of the relationship of this song text within the basic metrical structure proposed metrical basic structure and its with just a quick, general look at the syllabic seemingly great variation, namely the paradigms in this scheme. Having said that, possible hierarchical role of different word- another glance nevertheless reveals many boundary types in the sung text lines. This exceptions and variations in this basic means that, as in traditions of the Finnic organisation. However, even a quick alliterative tetrameter (Kalevala-metre in the analytical look at these 41 lines of this north and regilaul in the south), there also performance reveals the tendency of the first seems to be a kind of hierarchical or metrical section of the text to contain four oppositional relationship between sung text syllabic elements. This is also reflected in the lines, depending on the role and effect of the rhythmical organisation of this section, distribution of metrical stress on the initial represented in the scheme with four 1/8 notes syllables of words. The prosody of both the ( ). Of these, the first two Finnic and Samoyedic languages is     contain the first two lexical syllables of the characterised by word-stress on the initial text line and the second two a place for syllable, although its consequences in metrical, additional syllables (see the metrical performances seem to differ brackets, especially the first 20 lines). somewhat. The Finnic tetrameter is basically Actually this place is ‘reserved’ for a separate trochaic: although there is some flexibility in metrical (mono)syllabic ngey. This particle the number of syllables in certain positions derives, most probably, from the verb nges’ varying by region, it customarily yields an [‘to be’],4 and one of the great ‘universals’ in eight-syllable line, especially in Finland and Nenets metrical language in all regional Karelia. When lines are composed only of styles. Although it is a monosyllabic word, words with even numbers of syllables, such as ngey is pronounced with a final reductional four two-syllable words (2 2 2 2) or two two- vocal filling, breaking it into two syllabic syllable words and one four-syllable word (2 (and rhythmical) elements, resulting in the 2 4), then the first syllable of each word falls form ngey-(ya). This is not, however, a in a metrically stressed position and metrical mechanical phenomenon (see this position in stress and lexical stress correspond. However, lines where this word does not occur), lines can also be formed with words having although a very typical one, especially in an odd number of syllables, such as a two- syllable word and two three-syllable words (2 Western Nenets narrative song (see, for 3 3) or a three-syllable word and a five- example, Dobrovol’skiy 1965). syllable word (3 5). In this case, metrical The second metrical section contains (also stress contrasts with lexical stress, which falls with exceptions) three syllabic elements, of between metrically stressed positions and thus which the rhythmical manifestation is: becomes unstressed when performed in time . In this section, the tendency and    and in metre. This type of line has commonly space is reserved for two lexical syllables, been called a ‘broken line’ (on analogy to the also added by a reductional vocal filling, for Germanic tetrameter where a word-break example ya-khey-(ya) (> yakha” [‘one of two would be mandatory between the second and twins’]) in line 10. third stressed positions of a line), and creates The third metrical section tends to contain a different sound impression. The relation two final lexical syllables of the text line, between these types of lines is customarily which has longer, line-final cadence viewed as a hierarchical relationship of durations: . Here the variation caused    ‘normal’ lines to which ‘broken’ lines are a by euphonic or reductional transformation variation that creates (aesthetic) tension with seems to be more modest and consists quite 29 the regular rhythm of the basically trochaic type seems prominent, producing lines of the tetrameter’s environment. (See further form 2 (X) 2 2 (where ‘X’ represents the Sadeniemi 1951; Leino 1986; Gomon 1990.) additional metrical euphonic element This potential for tension between metrical ngey(ya)). After the first 20 lines of the song, and natural lexical stress is similar to a the singer seems to change to a form placing phenomenon in Nenets sung metre. The two text syllables in the position of the Finnic tetrameter can be considered dipodic, ngey(ya) and stretching the two last syllables or organised in two pairs of metrical feet (i.e. of the line with long durations, which shifts the two trochees of positions 1–4 and the two the line-form while remaining within the of positions 5–8). basic strategy of forming the line from words The Nenets hexasyllabic basic text line is with an even number of syllables. This is also dipodic, but trimetric, if thought of as a partially reminiscent of the appearance in this metrical form. This basic text line, with many song of the line type formed from words with of the sung metrical forms produced from it, an odd number of syllables, like line 16 seems to point to a fundamental opposition of (where, however, the basic text line is not lines with words of even and uneven numbers only formed using words with an odd number of syllables. Whereas Finnic singers produce of syllables, but the line also appears to be a tensions between metrical and lexical stress, syllable short (five syllables rather than six): Nenets singers never do so. Instead, they aim (...) sya-ni-(yey-(ya)) e-ba-tow, metrically (...) to synchronise word-stress with metrical 2 (X) 3. In short, the contrastive character of stress and thus a word with an odd number of the lines of even and odd type is not syllables is compensated within the scheme of particularly distinguishable in this song5. In the basic trimeter by completing the other songs, with other metrical basic remaining positions in the segment of the line schemes, this metrical contrast is more with supplementary verbal elements. This pronounced. This kind of general, economical allows lexically stressed syllables to be and all-purpose metrical basis seems to be delayed until there are appropriate metrically more present in narrative songs, whereas in stressed positions in the basic scheme. shorter personal songs the opposition of the Revealing this principle in Nenets metrical paired and unpaired line types is often more word-border types also evokes the question of prominent (see Niemi & Lapsui 2004). a hierarchy of line-types. As I have pointed Although this interpretational scheme out elsewhere with more extensive materials seems to point at the remarkable uniformity of (Niemi 1998), there is a discernible contrast the manifestation of the basic metre, it cannot, in Nenets sung trimeter between the varieties of course, provide final or universal of the lines with content words containing explanations for the great variation within this either an even (e.g. 2 2 2; 2 4) or odd (e.g. 3 scheme. Such variation could be related to a 3) number of syllables and in the ways these singer’s competence and also be idiolectal. It varieties activate metrical processes with could also be a fundamental, and thus socially metrical or supplementary syllables in acceptable, freedom of a singer to create this performance. The variety of even-syllable type of variation in performance, capitalising words in a line seems to be more common on these contrastive basic forms. Determining than its odd-syllable counterpart (which is, the sources for variation within this scheme however, also fairly well represented). This would require other methods of inquiry, such statistical difference also gives grounds to as long-term ethnographic work with the presume a hierarchy between the basic and performers in order to understand the contrastive principle in the Nenets sung text cognition behind a successful performance. line, even if the contrast in this poetry is Unfortunately, such research approaches between conformity and contrast of word- should have been conducted during the breaks to dipod divisions rather than between 1980s–1990s – at least among the Western metrical and lexical stress. Nenets – when more competent singers could Thus, in the present song, the principle of still be found. organising words into a regular metrical line

Conclusion Notes This paper has presented the methodological 1. See the presentation of the method in music analysis approach of using a descriptive analysis of in Niemi & Lapsui 2004; Niemi & Jouste 2013. 2. For further details about Kupriyanova’s publication, poetic form as a tool that can be reciprocally see Niemi 1997. applied for stylistic analysis of a particular 3. Running number of the 1994 Malaya Zemlya field traditional song performance, illustrated recording collection, audio files archived in through the case of a Western Nenets Folkways Archive of the University of Tampere, narrative song. If we agree on the suitability Finland as Y10675–Y10676 (DAT) and Y10679– Y10680 (cass.). and adjustability of analytical methods of a 4. See, however, Hajdú’s (1978: 361–362) critique of structural kind, the approach proposed here this view. could be developed to chart the textual 5. For more extensive examples of this line-type dimensions of the song, such as the opposition, see Niemi & Lapsui 2004. organisation of informational content at different levels of scope, from images, Works Cited Castrén, M.A. 1940. Samojedische Volksdichtung. Ed. dialogues and themes or scenes up to the level T. Lehtisalo. Mémoires de la Société Finno- of episodes in the complete plot. It is natural Ougrienne 83. Helsinki: Société Finno-Ougrienne. that most of the observations only apply to the Dobrovol’skii, B.M. 1965. “O napevakh nenetskikh single performance of this case study. Perhaps epicheskikh pesen”. In Epicheskiye pesni nentsev. only the metrical code of this song could Ed. E.M. Meletinskij Moskva: “Nauka”. Pp. 757– 780. survive in other performances, at least by the Finnegan, Ruth 1992. Oral Traditions and the Verbal same performer. A rigorous analysis of the Arts: A Guide to Research Practices. London/New kind presented in this paper is fundamentally York: Routledge. quantitative and therefore much more data is Foley, John Miles. 1990. Traditional Oral Epic: The required for reliability. The function of this Odyssey, Beowulf, and the Serbo-Croatian Return Song. Berkeley/Los Angeles/London: University of example is just to point to a possible direction California Press. for a kind of structural analysis, which, I Hajdú, Péter. 1978. “The Nenets Shaman Song and Its believe, can be designed to be reflective enough Text”. In Shamanism in Siberia. Ed. Mihály to allow discussion and criticism, since the Hoppál. Budapest: Akadémiai Kiadó. Pp. 355–372. paradigmatic layout of the analysis purports Hajdú, Péter. 1982 [1968]. Chrestomathia Samoiedica. Budapest: Tankönyvkiadó. to trace the constellation of the recurring Honko, Lauri. 1998. Textualising the Siri Epic. FF textual phenomena as embedded in the flow Communications 264. Helsinki: Academia of the performed time. Naturally, this kind of Scientiarum Fennica. approach has to be further articulated with a Hymes, Dell. 1981. “In Vain I Tried to Tell You”: culture- and data-sensitive research design. At Essays in Native American Ethnopoetics. Studies in Native American Literature 1. Philadelphia: this stage, the preliminary structural analysis University of Pennsylvania Press. of this text has yielded the fruitful results Lehtisalo, Toivo. 1947. Juraksamojedische presented here. The sequel to this paper will Volksdichtung: Gesammelt und herausgegeben von follow as research progresses. For the moment, T. Lehtisalo. Mémoires de la Société Finno- the other aspects of Ekaterina Semënovna Ougrienne 90. Helsinki: Société Finno-Ougrienne. Lehtisalo Toivo. 1956. Juraksamojedisches Kanyukova’s performance still remain Wörterbuch. Lexica Societatis Fenno-Ugricae 23. stronger than the efforts of the analysts. Helsinki: Société Finno-Ougrienne. Lehtisalo, Toivo (ed.). 1960. Samojedische Jarkko Niemi (jarkko.niemi[at]uta.fi), School of Sprachmaterialen gesammelt von M.A. Castrén und Communication, Media and Theatre, 33014 University T. Lehtisalo. Mémoires de la Société Finno- of Tampere, Finland. Ougrienne 122. Helsinki: Société Finno-Ougrienne. Acknowledgements: The author wishes to express his Niemi, Jarkko, with Anastasia Lapsuli. 1995. (Niemi) gratitude to the Scottish Gaelic traditional artist and “Transcriptions of Nenets Songs in A.O. researcher Margaret A. Callan (University of Väisänen’s ‘Samojedische Melodien’ Re- Edinburgh) for her insightful remarks and comments Evaluated”; (Niemi & Lapsui) “Appendix 1: Note on an earlier version of this text, which appeared in the transcriptions of Nenets Songs in A.O. Väisänen’s conference proceedings Song and Emergent Poetics – ‘Samojedische Melodien’”; (Niemi & Lapsui) Laulu ja runo – Песня и видоизменяющаяся “Appendix 2: Texts of Nenets Songs in A.O. поэтика, edited by Pekka Huttu-Hiltunen et al. Väisänen’s ‘Samojedische Melodien’”. Journal de (Kuhmo: Juminkeko, 2015).

la Société Finno-Ougrienne 86: 129–146, 147–161, Niemi, Jarkko, & Anastasia Lapsui. 2004. Network of 162–166. Songs: Individual Songs of the Ob’ Gulf Nenets. Niemi, Jarkko. 1997. “Music Transcriptions in Z.N. Music and Local History as Sung by Maria Kupriyanova’s Collection of Nenets Epic Songs”. Maksimovna Lapsui. Mémoires de la Société Journal de la Société Finno-Ougrienne 87: 161–224. Finno-Ougrienne 248. Helsinki: Société Finno- Niemi, Jarkko. 1998. The Nenets Songs: A Structural Ougrienne. Analysis of Text and Melody. Acta Universitatis Pushkarëva, E.T. 2000. Nenetskie pesni-khynabtsy: Tamperensis 591. Tampere: University of Tampere. Syuzhetika, semantika i poetika. Ed. B.N. Putilov. Niemi, Jarkko, & Marko Jouste. 2013. “Musiikin Moskva: Vostochnaya literatura RAN. paradigmaattinen analyysi”. In Musiikki kulttuurina. Siikala, Anna-Leena, & Jukka Siikala. 2005. Return to Ed. Pirkko Moisala & Elina Seye. Suomen Culture: Oral Tradition and Society in the Southern Etnomusikologisen Seuran Julkaisuja 21. Helsinki: Cook Islands. Helsinki: Academia Scientiarum Suomen Etnomusikologinen Seura. Pp. 173–200. Fennica.

The Metre of Yeats’s “Lake Isle of Innisfree” in Context Michael Cade-Stewart, King’s College London

This article argues that Yeats’s “Lake Isle of Innisfree” is to be scanned as isochronic, with the first three lines of a stanza having seven beats and a caesura after the 4th beat, here called ‘Sevens’ as opposed to fourteeners. The form is not mentioned in any history of English verse, but a digital search of ca. 2,500,000 lines of poetry reveals it to be representative of a neglected but significant strand of English poetry. Is it more frustrating for a poet to have his or he intimated to Robert Bridges, “I confess I her early work valued higher than later, grow not a little jealous of the ‘Lake Isle’ greater achievements, or for that work to be which has put the noses of all my other received in the wrong way? For Yeats, these children out of joint.” Yeats did not let this two problems coincide in “The Lake Isle of ‘jealousy’ cloud his business sense, however, Innisfree”, which he wrote in his early and in a letter of 23rd December 1920 he twenties. This coincidence has a bearing on resignedly suggested to his publisher the communication of metre from poet to Macmillan that the best sales of his poetry reader, and on the history of a largely would probably come from a volume forgotten poetic form. It is by attending to containing “the popular poems I wrote before Yeats’s own intentions for the rhythm of “The I knew better.” “Most people seem to want Lake Isle” that we can see how a number of whatever volume contains ‘Innisfree’,” he his other poems can best be appreciated. reflected. (InteLex I: 404; III: 90; #3834.) Moreover, we can situate these poems in a Yeats was apt to call those early poems wider context that adds meaning to the ‘popular’, since they were both intended to be poems, while also bringing light to a for and of ‘the people’, and were more neglected strand of canonical poetry. commercially successful. Today, it is those Yeats’s poem was published in The early poems, like “The Stolen Child”, which National Observer on 13th December 1890, one sees most frequently displayed on the when Yeats was twenty-five. It was then walls of public houses. included in the anthology of The Book of the In the BBC radio broadcast of 8th Rhymer’s Club (1892), and in his second September 1931, Yeats effectively announced collection of poems, The Countess Kathleen his ambivalent feelings about his early poem and Other Legends and Lyrics (1892). In an to the entire world. He would read it, he emotional trajectory that cannot be unusual, announced, “because if you know anything Yeats’s initial satisfaction with his poem gave about me, you will expect me to begin with it. way to irritation and even embarrassment, It is the only poem of mine that is very widely before grudging acceptance. Writing to known.” (TSWP, LAR: 224.) His second Robert Louis Stevenson in 1894, he sentence somewhat softens the first, but the confessed, “I need hardly tell you that your implication remains: Yeats felt obliged to praise of ‘The Lake Isle of Innisfree’ has read this poem, which had somehow eclipsed given me great pleasure.” Some seven years all his other poetry. The popularity of the later his feelings had cooled considerably, and poem may in part explain why we have extant 32 recordings of his reading of it. While most in fact isochronic. As will be seen, its formal recordings of Yeats performing his poetry features are far from unique in the English have been lost or destroyed, two exist for that canon. poem. Recorded in 1931 and 1937, the On the page, “Innisfree” is made up of similarities of these two performances reveal four-line stanzas comprising three long lines Yeats’s own intentions for the rhythm of the followed by a shorter one (VP: 117). If we poem. These recordings divide opinion. Some were to take an accentual-syllabic approach, listeners love them, but others object to the long lines have either six or seven loosely Yeats’s delivery as being too far from normal alternating beats, while the shorter lines have speech, too belaboured – it is closer to chant. less. Taken in isolation, one might conclude Yeats opened his 1931 broadcast with a that the opening two lines of the poem are statement about poetic rhythm, and the proper written in a looser form of hexameter with delivery and reception of verse: caesura, with six beats per line, and a ‘slack’ I am going to read my poems with great unstressed syllable before each caesura. The emphasis upon the rhythm, and that may reading would go something like this seem strange if you are not used to it. I (stressed syllables are underlined; numbers in remember the great English poet William square brackets indicate the number of beats Morris coming in a rage out of some lecture in a particular scansion): hall where somebody had recited a passage I will arise and go now,| and go to Innisfree, [6] out of his Sigurd the Volsung. ‘It gave me a And a small cabin build there,| of clay and devil of a lot of trouble,’ said Morris, ‘to get wattles made: [6] that thing into verse.’ It gave me a devil of a lot of trouble to get into verse the poems that This interpretation must underly the I am going to read, and that is why I will not impression that the poem is in hexameter, but read them as if they were prose. it is contradicted by Yeats’s own performance Along with “The Lake Isle of Innisfree”, the of the poem. In both extant recordings he surviving extract of Yeats’s broadcast actually articulates two consecutive beats near includes his performance of “The Fiddler of the middle of each of these lines: go now and Dooney”, which he had also written in his build there. These are followed by a marked early twenties. The two deliveries have more caesura, indicated in the text by a comma. than a little in common, despite their very This performance embodies what Bruce P. different appearances on the page. Both are Hayes and Margaret MacEachern (1998) term printed in quatrains, but most of the lines in the ‘Green-O’ cadence in folk poetry, where “The Lake Isle of Innisfree” are twice as long. two successive monosyllables are stressed in The aural affinity is instructive. performance at the end of a line (Green O). In Despite the recordings, very few people Yeats’s poem, though, these cadences occur seem to grasp the metre of the “The Lake mid-line – before the pronounced pause of a Isle”, or of Yeats’s other poems written in a caesura (‘g’ indicates a fourth beat of a similar form. Unfortunately, Yeatsian Green-O cadence): scholarship does not provide much guidance. I will arise and go now, | and go to Innisfree, Richard Ellmann declared in The Identity of [3g|3] Yeats that its “metre is not new”, but did not And a small cabin build there, | of clay and go so far as to set out what he thought it to be wattles made [3g|3] (1967: 124). He may well have assumed it Yeats then delivers the third line thus: self-evident, yet on the basis of descriptions of the poem on cribsheet-style websites, it is Nine bean-rows will I have there, | a hive for not inevitable that readers arrive at the same the honey-bee [4|3] conclusion. One common interpretation, This decisively contradicts the interpretation online and off, is that the poem is written in of the line as hexameter with caesura: in the “vaguely iambic hexameter and tetrameter” first hemistich he clearly articulates four (Shmoop 2008). It is my view that this is beats, taking the total for the line up to seven. wrong, and that the metre of Yeats’s poem is Later in the poem, we see another type of

33 cadence before the caesura: the familiar Homer’s Iliad (1616), and sections of William version with a (so-called) ‘feminine ending’, Blake’s “Prophetic Books” as examples of or ending the line with an unstressed syllable this rather different form (Greene 2012). of a polysyllabic word (‘f’ indicates a There are some descriptive terms in English feminine ending), as in: versification that seem to defy improvement – I hear lake water lapping | with low sounds ‘feminine endings/cadences’ would be a case by the shore; [3f|3] in point – but here a correction in terminology While I stand on the roadway, | or on the seems relatively straightforward. The term I pavements grey [3f|3] propose for these lines of seven beats divided 4|3 is ‘Sevens’, since they occupy seven In common with all of these lines, Yeats beats. In what follows, I will refer to lines that articulates three beats after the caesura. exhibit the ballad-metre cadences as ‘Ballad Before it, there are either: a) four alternating Sevens’, and the same form without the ballad beats; b) four beats effected by stressing two cadences as ‘4–3 Sevens’. This would leave consecutive syllables before the caesura; or c) the term ‘fourteener’ for poems with fourteen three beats with a feminine ending before the syllables or seven-stress lines that are not caesura. These three types of line endings, or broken into 4 and 3 stresses, though ‘cadences’ are attested by critics of folk forms heptameter might serve equally well. such as Hayes and MacEachern (1998). If we One apparent disadvantage of the term were to split each of these long lines at the ‘Sevens’ is that the actual rhythmic structure caesura, the poem would take the form of is in fours. As Derek Attridge argues, there is stanzas of common ballad metre (4.3.4.3), considerable textual and experimental although arranged in long stanzas of six lines evidence of what he calls an ‘unrealised beat’ rather than four. Yeats’s delivery indicates at the end of a 3-beat ballad line (1982: 84– that he expects the poem to be performed in 96). As he argues, the alternating 4.3.4.3 of an isochronic manner: the intervals between common hymn metre and ballad metre has an the stressed syllables are conspicuously underlying rhythmic structure of 4.4.4.4, or 4x4. regulated. As this clearly implies, Attridge’s conception Objective measurements of the recordings of the metre of this form is isochronic. show that these beats are not separated by From this perspective, the relation of this exact intervals, but they sound approximately ‘unrealised beat’ to this form is clear: the regular. This quality was appreciated by the division of the quatrain (4.3.4.3) into two composer, pianist, and conductor, Igor pairs is: Stravinsky who observed that: emphasised by the absence of a realised beat Articulation is mainly separation, and I can at the end of the second line, a structure give no better example of what I mean by it which is reflected graphically when the than to refer the reader to W.B. Yeats’s verse is set out as a fourteener couplet recording of three of his poems. Yeats pauses (Attridge 1982: 88). at the end of each line, he dwells a precise time on and in each word – one could as As Attridge goes on to observe, however, easily notate his verses in musical rhythm as such lines will always fall into two halves scan them in poetic metres. (Stravinsky (after the first four beats) – effectively falling 1 1959: 136. ) back into the ballad form. Any advantages These long lines would normally be that the long line might have afforded considered ‘fourteeners’, but this term is narrative or meditative verse are lost: “the somewhat unhelpful. For one, it implies that units of composition remain short, and the the lines are regulated by number of syllables, 4x4 structure, with its associations of song- when in fact the regulating feature is stress- form, remains prominent” (Attridge 1982: timing. Secondly, the term ‘fourteener’ is also 88–89). This does not ring true for “The Lake used for a different kind of poem with lines Isle”, but this speaks more to the exceptional that do not fall in half – the Princeton nature of Yeats’s poem than anything else. Encyclopaedia of Poetry and Poetics entry for In spite of the underlying 4x4 structure, the fourteeners cites Chapman’s translation of term ‘Sevens’ seems more descriptive of the 34 material than ‘eights’ or ‘sixteens’. In a which I take to be composed entirely in pleasing parallel, the term appears in a dance- Ballad Sevens. Indeed, the poem has a step used in traditional 4-time reel dances in number of features that make it more typical Ireland. The name ‘seven’ is given to a of that form than “The Lake Isle”. All of these movement of a rapid succession of seven are visible in the opening two lines: the first steps sideways across two bars of music, with has a feminine ending to its initial hemistich the weight always on the back foot, followed with a caesura encouraged by the comma, by a switching of the weight to the other foot. while the second line has seven clear beats. (The step is frequently taught with the This neatly rules out hexameter in the opening instructions: ‘1-2-3-4-5-6-7-and...’, with the lines of the poem, while the couplet rhyme ‘and’ marking the beat that is not stepped on.) scheme also places us firmly in the company The metrical theorist most famously of most other Ballad Sevens: associated with the principle of isochrony in Must we part, Von Hügel, | though much poetry is Coventry Patmore. Making his case alike, for we [3f|3] for the marking of a “caesural, or middle Accept the miracles of the saints | and pause, in some kinds of verse” in his Essay on honour sanctity? [4|3] English Metrical Law (1961 [1857]), Patmore Despite these textual signposts, the form has turns to a couplet attributed to Henry Howard, baffled recent commentators. No less a critic Earl of Surrey (1517–1547): than Helen Vendler has pronounced the form And some I see again sit still, and say but as being a mix of “fourteeners and small. hexameters”, positing that its “refusal to be That can do ten times more than they that say consistent is I suppose appropriate for its they can do all. conversational jog” (2013: 166). Yet this These lines are in a form usually identified as monologue in rhyming couplets is surely ‘Poulters’ measure’: rhymed couplets with 12 some distance from what most of us think of syllables in the first line and 14 in the second. as conversational. When considered as a Patmore instructs the reader to place a pause Ballad Seven instead, the poem becomes after again, and for this to take up a whole metrically consistent. As we will see, its beat, so that the lines have the same number rhyme scheme, and to a lesser extent its of beats. The couplet thereby serves to subject matter, place it squarely in the illustrate two points for Patmore: first, the tradition of such poems. Attridge posits that need for marking a ‘middle pause’; second, the 4x4 rhythmic structure is the strongest and the aesthetic pleasure to be derived from a most easily-perceived rhythm in English, with mismatch between the metrical pause and a considerable justification, but this rhythm is break in syntax (Patmore 1961 [1857]: 23– not always felt, even by experienced readers 24). In this way, these lines in Poulters’ of poetry. I will return to this issue in my measure are regulated in the same way as the conclusion. Sevens: the performer needs to deliver both Another distinguished critic who scans lines so that they seem to take the same poems in the form differently is Meredith amount of time. Patmore recommends Martin. In her impressive social history of inserting a pause; in other lines, the same poetic metre, The Rise and Fall of Metre: effect can be achieved by the over-stressing of Poetry and English National Culture, 1860– consecutive monosyllables (the Green-O 1930 (2012), she scans the fourteener lines of cadence) or of a polysyllable ending with an Newbolt’s “Drake’s Drum” as something unstressed syllable (the feminine cadence).2 else. For Martin, the metre of “Drake’s By such means, the lines each occupy seven Drum” changes stanza by stanza. The opening beats. is scanned as four ‘paonic’ feet (Martin 2012: Despite this example used by Patmore, the 125). Presumably this would look something principle is not usually applied to Yeats’s like “Drake he’s in his hammock an’ a poetry, even where it makes an otherwise thousand mile away.” In her account, the irregular poem fit a metre. One such poem, is second stanza then “shifts the meter from first section eight of “Vacillation” (VP: 503), paeons to tighter trochees,” but she scans the

35 penultimate line of this stanza as a kind of labelling of the poem as a ‘song’ is an irregular hexameter with caesura (and no incorrect assertion made by the compilers of initial trochee): the collection intending to make a distinction If the Dons sight Devon, I’ll quit the port between such ‘songs’ and real poems. (Martin o’Heaven. 2012: 114–115.) While the editors have taken many liberties with the poem, this is not I scan the poem differently in that I interpret among them: Davis wrote many poems he all of the longer lines as being consistently titled as songs and ballads, and this poem structured as Ballad Sevens. The line from the appears in the sections “Miscellaneous Songs stanza of trochees comes out thus: and Ballads” in John Mitchel’s edition of the If the Dons sight Devon, | I’ll quit the port The Poems of Thomas Davis – and rightly so. o’Heaven [3f|3f] The poem can easily be envisaged as a This interpretation is compatible with ‘traditional’ song, and be scanned with similar Martin’s decision to stress the three ease as a 4x4 isochronic poem. It opens with a consecutive words Dons sight Devon, though: quatrain of three Green-O cadences followed the important distinction is the isochronic by a three-stress line (which can treatment of the line. One advantage of my accommodate an unrealised beat at its end). interpretation is that all of the longer lines can The next quatrain keeps the same structure, be seen to observe this metre, so there is no but patterns it differently, using the feminine need to posit an unmarked change in metre ballad cadence in place of the Green-O (Davis mid-poem. ca. 1800: 78–79; notice that line 2.3 can Martin’s interpretation of many of these alternatively be 3g): lines as paonic, with large gaps of unstressed 1. syllables, might be otherwise explained by the She is a rich and rare land; [3g] theory of dipodism: that the stresses are Oh! she’s a fresh and fair land; [3g] structurally arranged in pairs, and that one She is a dear and rare land – [3g] This native land of mine. [3] stress is stronger than the other. So it might be scanned thus (the strongest beats are double- 2. underlined): No men than her’s are braver – [3f] Her women’s hearts ne’er waver; [3f] Drake he’s in his hammock an’ a thousand I’d freely die to save her, [3f] mile away. And think my lot divine. [3] Many poems in Sevens can comply with a The rhyme scheme is strictly maintained dipodic interpretation (with a missing final throughout the poem, and alternates with the stress), but Yeats’s recorded performances of cadences. The indented closing line of each “The Lake Isle of Innisfree” do not: the quatrain rhymes with mine, while the initial stresses do not alternate regularly in intensity. three lines of each stanza follow a scheme of Elsewhere in the same chapter of her book, double rhyme that alternates between rhyming Martin draws our attention to poems with both syllables of fair land, or save her. conscripted to schoolbooks intended for the This strictly regulated sound-patterning, education of children. One example is coupled with the heavy use of repetition, Thomas Davis’s poem “My Land”, which is gives the impression that sound is more included in Book V of Chambers’s Narrative important than sense in the poem, as is more Series of Standard Reading Books, Specially typical of song. These quatrains are wholly Adapted to the Requirements of the Revised consistent with Attridge’s (1982) theory of Code Including Those in Writing, Arithmetic, the ubiquity of 4x4 constructions in folk verse and Dictation (1863). Her scansion of this in English, and with the findings of Hayes & poem has similarities to that of Newbolt. MacEachern (1998). As Martin observes, Davis’s poem is given While it is possible to scan Davis’s song as without an author under the title “My Native a three-beat poem, this removes it from the Land”, taken from the poem itself. She reads wider stream of four-beat isochronic folk the poem as ‘three-beat’, and posits that the verse, and the rich history of such forms. The

36 same is true for poems with lines in Ballad Are wronging your image that blossoms | a Sevens like section eight of “Vacillation”, and rose in the deeps of my heart. (VP: 142.) “Drake’s Drum”: scanning them isochronic- And the latter: ally transforms them from idiosyncratic and irregular oddities into members of a much I would that we were, my beloved, | white more significant strand of poetry that has, in a birds on the foam of the sea! We tire of the flame of the meteor, | before it very literal sense, been hidden in plain sight. can fade and flee (VP: 121.) Today, then, poems like these by Yeats and others are received inconsistently even among These happily sit within the Ballad Seven distinguished scholars of poetic form. This form, with each line taking the form 3f|3. fact raises questions as to why Yeats chose Equally, though, one might read the poems as this form for quite different poems – loose hexameters with a regularly applied especially when the 4.3.4.3 ballad form was medial caesura. The poem “The Salley available to him – and how common the form Gardens” could be scanned in this way, but its was among his predecessors and contempor- origin in folk-song encourages us to treat it as aries of the 19th century. Ballad Seven. The poem’s source, “The Rambling Boys of Pleasure”, is structured as The Ballad Seven in Yeats a 4.3.4.3 ballad, typically printed in long This history can be seen by starting with stanzas of eight lines.3 “The Salley Gardens” Yeats and looking backwards. Throughout his is still sung today, and Yeats believed it to career, Yeats wrote a number of poems have been used as a marching tune for the comprised of Ballad Sevens, on subjects army of the Irish Free State.4 stretching from metaphysics, through Yeats used the form less ambiguously in meditations on love, to art criticism. These his later poems, and we might relate this poems are “The Lake Isle of Innisfree”, “His change to his movement away from his earlier Phoenix”, “Owen Aherne and His Dancers”, style at the turn of the century. After section 8 of “Vacillation”, “The Three considerable tumult in his private life in the Movements”, and “The Curse of Cromwell”. early 1900s, he embarked in a dramatically What makes it clear that these poems must be new direction in his poetry. This was to interpreted in the same way as “The Lake Isle manifest itself most clearly in his collection In of Innisfree” is the presence of lines with one the Seven Woods (1903). His earlier poetry, of the ballad cadences before the caesura, much like his earlier criticism collected in alongside lines with 7 clear beats – meaning Ideas of Good and Evil (1903) was “too that the poem cannot be comfortably read as lyrical, too full of aspirations after remote hexameter. things, too full of desires”, in a manner he no In the 1890s, Yeats and fellow poets longer approved of (InteLex III: 372). In a deliberately employed the form in poems that letter to John Quinn he went on to reflect that might best be described as lyrics. In this they “I have always felt that the soul has two were writing against the prevailing tendencies movements primarily, one to transcend forms, of the form. In other early poems, Yeats and the other to create forms;” his earlier blurred the distinction between hexameter and poetry had striven to transcend form; he the Sevens form, in writing poems that can be would now create or reinforce it: substituting consistently scanned as either. “The Lover Apollonian order for “that wild God Tells of the Rose in His Heart”, and “The Dionysius” in his new Nietzschean White Birds” are examples of such poems. formulation (InteLex III: 372). The opening of the former is: This is certainly true for his next employment of the Ballad Seven form; in All things uncomely and broken, | all things worn out and old, “His Phoenix” composed in January 1915. The cry of a child by the roadway, | the creak This poem opens with relatively clear textual of a lumbering cart, signs as to its metre. Just as with section 8 of The heavy steps of the ploughman, | “Vacillation”, the opening line has six clear splashing the wintry mould, beats, but a mid-line feminine cadence:

There is a queen in China, | or maybe it’s in Yeats’s desire to consolidate poetic form did Spain not manifest in a desire to increase the But in the poem as a whole, roughly two regularity or patterning of such features. thirds of the lines have seven clear beats It might seem somewhat perverse that (arranged 4|3), as in the third line: Yeats did not use the form for any of his early ballads, preferring the ‘opened out’ 4.3.4.3 Of her unblemished lineaments, | a form that readers are more familiar with. whiteness with no stain Instead, he only employed it in ballad form in While the metre of the poem can be in no his final decade, in “The Curse of Cromwell” doubt, the second line is challenging in that it (VP: 580). Shortly after writing the poem in seems to only have six beats. One easy way to early 1937, Yeats was keen to hear it sung, make it match up with the others is to treat the and arranged for this to happen in a BBC word birthdays as a Green-O cadence, even radio broadcast, and in more social settings. though it is mid-hemistich: To his pride and delight, the singer was asked And birth days and holidays | such praises to sing it twice by the audience at the formal can be heard dinner of the Irish Academy of Letters on 16 May 1937 (InteLex: #6953). Much like This matches the sound patterning of Patmore’s example from Surrey’s poetry, birthdays and holidays – with the initial and some lines contain only 6 stressed syllables, final syllables stressed. This treatment of and ballad cadences do not seem to be word-stress is entirely permissible under encouraged by the text: stress-timed performance, though an alternative is to follow Patmore in positing a You ask what I have found | and far and wide longer middle pause, to ensure the even I go, [3|3] Nothing but Cromwell’s house | and spacing of the beats. Cromwell’s murderous crew, [3|3] The 1:2 ratio of lines with feminine The lovers and the dancers | are beaten into cadences to those with seven clear beats is the clay, [3f|3] repeated in the first poem of “Owen Aherne And the tall men and the swordsmen | and the and His Dancers”, and section 8 of horsemen where are they? [3f|3] “Vacillation” (VP: 449, 503). In all of these And there is an old beggar | wandering in his poems we see tantalising suggestions of an pride [3f|3] emerging pattern to these cadences, a little His fathers served their fathers before | Christ like the alternating pattern of Poulters’ was crucified. [4|3] measure, but it is always broken in O what of that, O what of that [4] subsequent stanzas. This is clearly shown in What is there left to say? [3] the opening two stanzas of “Owen Aherne Other lines have the Green-O cadence, and His Dancers” from 1932: arranged 3g|3, as in line 14: A strange thing surely that my Heart, | when What can they know that we know | that love had come unsought [4|3] know the time to die? Upon the Norman upland | or in that poplar shade, [3f|3] A few have seven alternating beats: Should find no burden but itself | and yet Can pay the poet for a verse and hear the should be worn out. [4|3] fiddle sound, (l.21) It could not bear that burden | and therefore it went mad. [3f|3] Unfortunately we do not have a recording of the poem sung or even spoken, but it seems The south wind brought it longing, | and the clear that the combination of these features east wind despair, [3f|3] should alert us to the proper performance of The west wind made it pitiful, | and the north that text: namely, an isochronic delivery. In a wind afraid. [4|3] It feared to give its love a hurt | with all the song, or chant, it is facile to deliver these tempest there; [4|3] varied lines in a satisfyingly regular way. It It feared the hurt that she could give | and would seem, in fact, that far greater aesthetic therefore it went mad. [4|3] pleasure is granted by the stretching or

38 compressing the syllables to match the Hurrah for revolution | and more cannon musical score, compared with when the shot; syllables match up exactly. As a consequence, A beggar upon horseback| lashes a beggar certain circumscribed examples of textual upon foot; irregularity seem to imply affinity with song. Hurrah for revolution | and cannon come again, That Yeats should come back to his early The beggars have changed places | but the influences in his final decade was in part the lash goes on. result of his aged mind returning to the themes and ideas of his youth, but it was also Other scansions may suggest themselves to owing to the particularly Irish aspect of his new readers, but an isochronic approach seems the poem. Yeats was feeling increasingly embattled most appropriate. in his native Ireland, and what better form th with which to critique the values of the Free The 19 -Century Ballad Seven in the State than one that had been strongly Literary Canon associated with the Irish literary renaissance? As we have seen, the Ballad Seven form is a What better way to circumvent the prevailing minor but significant presence in Yeats’s censorship of the written word in Ireland than oeuvre, appearing about the same number of by making something enter the aural times in his Collected Poems as sonnets. But tradition? (See Cade-Stewart 2013: 221–234.) how common was the form among canonical Yeats also drew on the aphoristic potential authors, and for what subjects was it used? One way to investigate is to examine the of the form in two other poems from the th 1930s. “The Three Movements” is structured works of the poets active in the 19 century 4|3, but the second hemistich of each line is who are included in the Cambridge acephalous, a feature somewhat at odds with Bibliography of English Literature. Such the syntax of the final line, since it requires lie poets effectively constitute the broader canon and gasping to both be stressed, where one – few are as central to our conception of the would normally stress the adverb (VP: 485): period as the likes of Shelley and Keats, but they are all considered important enough to be Shakespearean fish swam the sea, | far away featured in the Cambridge Bibliography. from land; Examining their work is a more manageable Romantic fish swam in nets | coming to the task than one might think, since these poems hand; are included in electronic form in Chadwyck- What are all those fish that lie | gasping on the strand? Healey’s 1995 CD-ROM of The English Poetry Full-Text Database. Discs 4 and 5 With the pause of the caesura, however, such cover the period 1800–1900, meaning that lines are easy to deliver rhythmically. As will they include the poets who wrote or published now be evident, my interpretation of these most of their work between these dates.5 lines requires successive stresses – clashes if Together, these two discs include five million you will – in the middle of some hemistichs. lines of verse, however – far too many to be This seems to violate the normal alternating checked manually. My solution was to use expectation of English metrical verse, but digital tools to locate possible candidates, and Yeats was relatively fond of such clashes, as then sift through these by hand. As such, it recordings of his performances attest. In “The represents a mode of computer-assisted Lake Isle of Innisfree”, we can clearly hear criticism. him articulating successive stressed syllables Poems in Ballad Sevens are particularly in hemistichs such as: tricky to find using digital tools. In terms of And a small cabin build there”, and “with syllables, the lines may be as short as this one low sounds by the shore from Rudyard Kipling’s “Jobson’s Amen” (l.7):6 Yeats clearly enjoyed the effect. Following similar principles, the otherwise Conches in a temple, oil-lamps in a dome uneven quatrain of “The Great Day” also This structure: XoXoXf | XoXoX, is only comes out as a Ballad Seven (VP: 590): eleven syllables long, a length not uncommon

39 in iambic pentameter of the period, let alone I then performed the most laborious part of hexameter. Even this was not the shortest I the process: manual scansion of these poems. came across, however, owing to the To qualify as a Seven the poem needed to isochronic nature of the form. Ernest Jones’s consist of seven-beat lines that would split “The Cry of the Russian Serf to the Czar” after the fourth beat, or to have a stanza- opens with the line: pattern that included clear examples of such Labour! Labour! Labour! – Toil! Toil! Toil! lines. The poem was classed as a Ballad Seven if these lines included one or more This line could be read as a mere nine lines of six beats with a ballad cadence at the syllables – a particularly high ratio of beats to end of its first hemistich. If no such cadences syllables (7:9), though there is theoretically were present, the poem was classed as a 4–3 nothing to stop that ratio from becoming 1:1 Seven. A very small number of poems in the (Jones 1855: 102). Nevertheless, syllables are sample were entirely composed of six-beat the only way to narrow down the material: lines with feminine endings before the stanza lengths and rhyme schemes varied in caesura. I did not count such poems as the poems I had seen. Sevens. Although some poems structured in To reduce the poetry to a manageable this way are most likely Ballad Sevens (e.g. volume, I took a sample of half of this Yeats’s “Down by the Salley Gardens”) many collection (poets with names starting with A- others might also be classified as hexameter. K, on Disc 4 of The English Poetry Full-Text This manual checking removed false Database). This sample of roughly 2.5 positives from the list of possible candidates million lines was still very large, and likely to that I had produced by digital filtering. One be representative of the whole. To optimize consequence of this method is that there must the time-consuming process of syllable be more poems in the form of Sevens, even in counting, I used a simple letter-counting this sample, than I have identified. My figures algorithm to only consider lines of poetry that therefore represent the minimum frequency of were of about the right length (an average of such poems, but this minimum is no between 52 and 70 characters). Suitable lines insignificant number. were then posted to a text-to-speech server and converted into syllables. I used the open- A Survey of the Form source software MARY Text-To-Speech for Table 1. Poets in the sample with six or more Ballad this part of the process. MARY was a good Sevens. tool for this job, as I had considerably augmented the user-dictionary on my local Kipling, R. 39 Hopper, N. 20 server, and written Python scripts that correct Joyce, R. D. 17 its output. Bennett, W. C. 16 Where a line had at least twelve syllables, Cook, E. 10 it was added to a list. I chose twelve, rather Holmes, E. G. A. 9 than eleven, for pragmatic reasons: few Bayly, N. T. H. 9 poems will be written entirely in short lines, Allingham, W. 8 and far too many of the poems matched will Howitt, M. 8 in fact be pentameter or hexameter. Barlow, G. 8 The result of this process was a list of well Havergal, F. R. 8 over 100,000 lines. To narrow it still further, I Bickersteth, E. H. 8 only considered poems with entire stanzas, or Hickey, E. H. 7 O'Neil, M. 7 a majority of lines, in Ballad Sevens. As I was Browning, R. 6 interested in the distribution of these lines across poetic forms, I did not distinguish I found that poems in Ballad Sevens were between poems that were or were not divided included in the published work of 73 of the into regular stanzas. To find poems in Ballad 233 poets in the sample – roughly a third. By Sevens, I focused on sequences of broadly this I mean that the works of these 73 poets sequential lines near the beginning of poems. contained one or more poems with either entire stanzas or a majority of lines composed

40 in Ballad Sevens. About a third of these (28 Till they come to where the rowan trees | in of 73 poets) only included one poem in this lonely beauty grow [4|3] form. Though there were exceptions, most Beside the Fairy Hawthorn grey. [4] poets in the sample used it sparingly. Only It is worth observing at this point that the fifteen used it six times or more (see Table 1). poem is not in Poulters’ measure, although Since these poems have never been the first couplet might seem to be. For one considered together, to my knowledge, I will thing, the rhyme scheme is abab rather than in now outline the contours of the themes and couplets, but more importantly, the number of th genres that manifest in the form in 19 - syllables is not regulated in the poem. The century literary verse in English. Before I do apparent proximity to that form reveals their so, I must observe that this outline will not kinship, such that both Sevens and Poulters’ encompass folk or popular ballads of the measure can be fitted to a 4x4 structure, with period. an unrealised beat after the initial hemistich In the sample, the form was most popular where necessary, and at the end of each line. with Rudyard Kipling, but it was This is exactly what happens in the first line disproportionately used by Irish poets, with quoted, above. In a stress-timed delivery, the the likes of Nora Hopper, R.D. Joyce and caesura of that line needs only be extended a William Allington in the top six for little. In song, of course, a word can happily frequency. Emily Hickey and Moira O’Neil occupy far more beats than it has syllables, so (listed in the database under her pen name the vowel of hand could be stretched to ha- “N. Higginson”) were not far down the list, and if sung. and A.P. Graves only a little further still. This The opening lines of Ferguson’s ballad preponderance is likely to stem from the could be read as 3|3 with an unrealised beat, popularity of ballad-poetry among Irish poets but there is also potential for realising the beat and audiences. What almost all of these poets by stressing successive syllables (double- have in common, Irish or otherwise, is their underlined below): status as popular poets; the form was strongly, and rightly, associated with the demotic. Get up, our Anna dear, from the weary spinning-wheel; The form was used by Yeats’s chosen For your father’s on the hill, and your mother literary precursors: Thomas Davis, Samuel is asleep: Ferguson, and James Clarence Mangan, Come up above the crags, and we’ll dance a whose names are conspicuously enshrined in highland-reel his early poem “To Ireland in the Coming Around the fairy thorn on the steep. Times”. Here, he asserted: “Nor may I less be This gives a wrench to the stresses of natural counted one / With Davis, Mangan, syntax, and seems to be contradicted by the Ferguson” (VP: 18). Mangan used the form punctuation, but it would ‘realise’ the beats in for lyrical and introspective love poems, the same way as a Green-O cadence. Yeats serving as an example to Yeats of the may have concluded that successive syllables potential of the form. But it was from should be stressed in this way when Ferguson’s “The Fairy Thorn: An Ulster performing Ferguson’s lines, but he went Ballad” that he took the stanza shape and further, in placing successive stressed rhyme-scheme for his “The Lake Isle of syllables elsewhere in the line. As I Innisfree”. The form used by an Irish patriot demonstrated earlier, he did so in the opening was particularly suitable for his poem of of the second line: longing for idyllic rural Ireland. The fourth stanza of Ferguson’s poem And a small cabin build there exhibits most of the features of the form: The Green-O ending of build there is And linking hand in hand, | and singing as consistent with the cadences permitted in they go, [3|3] ballad form, but the clashing stresses of small The maids along the hill-side | have ta’en and cab-in is not typical of Ballad Seven their fearless way, [3f|3] poems in the sample. Such clashes were not in Ferguson, and were distinctly unusual in

41 poems of this type. While refrains and My heart is bright with Easter light, | my face opening lines might feature consecutive beat is fair to see. [4|3] syllables (as in Jones’s Toil! Toil! Toil! Because God’s risen, and out of prison | the quoted above), it was unusual for lines in the whole round world goes free. [4|3] main body of stanzas to display the feature in Hopper used the form for twenty of her arbitrary positions. Collected Poems (she was the second most In contrast, the feature is conspicuous in prominent user of the form in my sample). poems in this form that were written by poets Yeats’s influence was not limited to fellow influenced by Yeats himself. First among Irish writers, though. Ernest Dowson was his these was Nora Hopper, who was something companion at the writers’ meetings in the of a literary disciple of Yeats, and almost Cheshire Cheese pub; like Hopper, he was to certainly writing under his influence in this die young. He did not share her enthusiasm regard. In her volume Aquamarines (London, for the form, but he wrote one poem that 1902), she collected her poem on “Kathaleen utilised the Ballad Seven cadence. To draw Ny-Houlahan”, inspired no doubt by Yeats’s even more attention to the cadence he used a poems and plays on the Irish heroine; the feature of poetry that was becoming third stanza of her poem reads: increasingly anachronistic: the articulation of O Kathaleen Ny-Houlahan, | it’s old I am and -ed word-endings as an extra syllable. In the gray, example below, unreined needs to be I see the dead leaves blown about | the articulated unrein-èd if the poem is to be closing of my day; metrically regular. While not unfamiliar to The dead leaves, the red leaves, | are rotting Victorian poets, this feature came to look too in my way, much like a ‘poeticism’ by the turn of the O Kathaleen Ny-Houlahan, to-day. century: Her practise is rather more song-like than When the mad winds are unreined, wilt thou Yeats’s, owing to the repeated words, internal not storm, my sea? [3f|3] rhymes, and the simple end-rhyme: here, as Dowson’s “The Sea-Change” was collected in elsewhere in her poems, sound seems more Decorations in Verse and Prose (1899). We important than sense, in a manner that is often can see the influence of Yeats’s use of the appealing. She was a popular poet, and form in Dowson’s successive stresses of mad printed copies of her first volume of poetry, and wind. Like Yeats, his employment of such Under Quicken Boughs (1896), are available clashes constitutes a complication to the today in various impressions and bindings. simple alternating rhythms typical of poems She used the form to riddling effect in her in this form in the nineteenth century. It luminous poem “Easter”, from Songs of the emblematizes the difference between Morning (1900): Ferguson’s ballad and Yeats’s “Innisfree”, in I am the dream of April, | I am the soul of the latter’s use of the form for lyric subject May; [3f|3] matter. Unlike Ferguson’s “Fairy Thorn”, The sallows scatter, the sallows splatter | Yeats’s poem is one of lyrical introspection. their gold upon my way: [4|3] Dowson followed Yeats’s lead in this respect, The gorses swing censers of spring | to employing a fairly complex rhythm more honour Easter Day. [4|3] common to poems in other metres. This I am the baby April, | the woman May will rhythmical innovation may go part of the way be: [3f|3] to explain why modern critics have I set the berry and hang the cherry | on briar misunderstood Yeats’s poem, but it is not the and cornel-tree; [4|3] only reason. Mine ’s the shut rose, the apple-blows, | the The form was not only used in the service rainbow on the sea. [4|3] of Ireland: the most frequent user of the form My tears are all of April, | my laughter is of in my sample was Rudyard Kipling (1865– May, [3f|3] 1936), whose sympathies lay elsewhere. My sorrow’s all a cowslip-ball, | so light to Kipling employed the form, like many other toss away: [4|3]

42 writers, for comic and frequently satirical Kipling’s “The Legends of Evil” opens with poems. In “The Puzzler” he used the form to the alternating lines of common ballad metre lampoon the Irish along with the other Celtic (4.3.4.3) but switches to Ballad Sevens for the peoples: second part of the poem. Once the opening The Celt in all his variants | from Builth to section sets the rhythm, the long lines of the Ballyhoo, [4|3] Sevens are introduced to save paper. As we His mental processes are plain | – one knows will see, there are even clearer examples of what he will do, [4|3] this practice. A great many of the poems in my sample The main subject of his poem is a valorization were narratives of military engagements and of the ‘puzzling’ minds of ‘the English’. In other heroic struggles against the odds on land this, it is representative of the jingoistic aspect and sea. The affinity with narrative verse of so many of his poems in the form. arises from the roots of the form in traditional Kipling also used the form for satirising ballads, which also gives rise to its aspects of contemporary society, as in “The manifestations in other types of popular song: Gods of the Copybook Headings”, which sea-shanties, Barrack-Room Ballads and opens: works for the music hall. A representative of As I pass through my incarnations | in every the former that is still sung today is “A age and race, [3f|3] Thousand Leagues Away: A Sea Song” by I make my proper prostrations to | the Gods W.C. (William Cox) Bennett (1820–1895): of the Market-Place. [4|3] The wind is blowing fresh, Kate, | the boat His predominant use of the form was for rocks there for me; [3g|3] narrative verse, most strikingly in the One kiss and I’m away, Kate, | for two long concentrated summary poems that open his years to sea; [3g|3] stories of historical fantasy in Rewards and For two long years to think of you | – dream Fairies (1910). In “King Henry VII. and the of you night and day – [4|3] Shipwrights (A.D. 1487)” (“The Wrong To long for you across the sea | – a thousand Thing”), he used it for a historical pastiche, leagues away, [4|3] written in pseudo-Middle English: A thousand leagues away, dear Kate, [4] A thousand leagues away, [3] Harry, our King in England, | from London While round the Pole we toss and roll, [4] town is gone, [3f|3] A thousand leagues away. [3] And comen to Hamull on the Hoke | in the Countie of Suthampton. [4|3] Kipling serves us with the best examples of For there lay the Mary of the Tower, | his literary poems that aspire to be sung in ship of war so strong, [4|3] barracks, such as his “Cells”, which opens: And he would discover, certaynely, | if his I’ve a head like a concertina, | I’ve a tongue shipwrights did him wrong. [4|3] like a button-stick, [3f|3] The Ballad Seven is germane for works of I’ve a mouth like an old potato, | and I’m historical fantasy, since it seems to gesture more than a little sick, [3f|3] back to stress-timed song-poems such as “The But I’ve had my fun o’ the Corp’ral’s Guard; Cuckoo Song”, one of the earliest surviving | I’ve made the cinders fly, [4|3] And I’m here in the Clink for a thundering Middle English lyrics (Ferguet al. 2005: 15). drink | and blacking the Corporal’s eye. [4|3] It is intriguing to note that this early song- poem is written out in prose – to save He probably hoped that such songs would valuable space on the page, no doubt. But like enter into that oral tradition. the Ballad Sevens, it displays a confidence on In a fusion of the barrack-room and the behalf of the writer that the reader will lyrical “Lake Isle of Innisfree”, the English recognize the proper form. Tommy, William Oliphant penned a parody The matter of economy of paper brings us of Yeats’s poem during World War I (Allison to a reason the Ballad Seven should seem 2006: 207): suited for narrative verse in general – in its potential for turning two lines into one.

A Picardy Parody (W.B.Y … ts) There’s many a black black eye, they say, | but none so bright as mine; [4|3] I will arise and go now, and go to Picardy, There’s Margaret and Mary, | there’s Kate And a new trench-line hold there, of clay and and Caroline: [3f|3] shell-holes made, But none so fair as little Alice | in all the land No dugouts shall I have there, nor a hive for they say, [4|3] the Lewis G., So I’m to be Queen o’ the May, mother, | I’m But live on top in the b. loud glade. to be Queen o’ the May.[4|3] And I may cease to be there, for peace comes We see a less innocent side of the speaker’s dropping slow, character when she heartlessly taunts a young Dropping from the mouth of the Minnie to where the sentry sings; man, Robin, who has indicated that he loves There noon is high explosive, and night a her: gunfire glow, They call me cruel-hearted, but I care not And evening full of torpedoes’ wings. what they say,

I will arise and go now, though always night For I’m to be Queen o’ the May, mother, I’m to be Queen o’ the May. and day I’ll feel dark waters lapping with low sounds They say he’s dying all for love, but that can by the store, never be: Where all our bombs grow rusty and They say his heart is breaking, mother – countless S.A.A.; what is that to me? I’ll feel it in the trench-feet sore. There’s many a bolder lad ’ill woo me any Oliphant’s parody preserves the rhythm and summer day, rhyme scheme of Yeats’s poem extremely And I’m to be Queen o’ the May, mother, I’m to be Queen o’ the May. closely. It is unlikely that this English soldier would have heard Yeats read the poem The proud and callous young lady is set up for himself: instead, the rhythms were clearly a fall. By the “New-Year’s Eve” of the next conveyed on the printed page, as must have poem, calamity has struck: tragically, the been true for many men of his generation. young speaker is now facing premature death, As we have seen, the form was far more longing only “to live till the snowdrops come associated with satire than with lyricism; again.” The “Conclusion” gives her a voice Oliphant’s parody exploited that association beyond the grave, with which she reflects on to comic effect. Would this parodic potential her life, and atones for her slight have been so evident, or so effective, if transgression, telling her mother to “say to Yeats’s poem had not already been something Robin a kind word, and tell him not to fret” of an anomaly? and to offer the consolation that she would In the poems of the sample, we can see likely “have been his wife” had she lived. another parody working in an analogous way, Aytoun’s parody undoes the tragedy and this time with the poem sequence of “The the sentimentalism of Tennyson’s ballad. In May Queen”, by Alfred Lord Tennyson. In “The Biter Bit”, the young female speaker has this sequence, Tennyson drew on the prevaricated in accepting a young man – “He associations of the Ballad Seven form with said I kept him off and on, in hopes of higher folk verse to deliver a sentimental narrative game” – but he has given up on her, and through the voice of a young rustic maid: married another. The “Biter” has been “bit”, and her unworthy complaint, and wheedling You must wake and call me early, | call me early, mother dear; [4|3] demands on her mother, undoes our sympathy To-morrow ’ill be the happiest time | of all at the bathetic climax of the poem: the glad New-year; [4|3] I did not know my heart, mother, – I know it Of all the glad New-year, mother, | the now too late; maddest merriest day; [4|3] I thought that I without a pang could wed For I’m to be Queen o’ the May, mother, | some nobler mate; I’m to be Queen o’ the May. [4|3] But no nobler suitor sought me, – and he has

taken wing,

And my heart is gone, and I am left a lone And take a lesson from this tale, | of the and blighted thing. Spider and the Fly. [4|3]

You may lay me in my bed, mother, – my The form appears to rather lighter purpose in head is throbbing sore; the work of Lewis Carroll (1832–1898), in the And, mother, prithee let the sheets be duly well-known poem “The Mock Turtle’s Song”, aired before; which famously begins: And, if you’d please, my mother dear, your poor desponding child, “Will you walk a little faster?” | said a Draw me a pot of beer, mother, and, mother, whiting to a snail, draw it mild! “There’s a porpoise close behind us, | and he’s treading on my tail.” As with Oliphant’s parody of Yeats, Aytoun’s works by exploiting the association of the The opening lines from both poems appear to form with satire; exposing moral failings accommodate interpretation as 8-beat lines, without sentimentalism. with trochaic openings (“Will you walk a little Tennyson may have thought the form to be faster?” | said a whiting to a snail). fitting for the voice of an uncultured young Subsequent lines in both poems are not lady because of its associations with folk amenable to this interpretation, however, such verse, but also with verse intended to be for, as this stanza from Carroll’s song: or from the perspective of children. The “What matters it how far we go?” | his scaly kinship between Ballad Sevens and isochronic friend replied. [4|3] nursery rhymes may lie behind this tendency, “The further off from England | the nearer is and offers a potential explanation as to why to France. [3f|3] some poems are considered suitable for There is another shore, you know, | upon the children.7 The Ballad Seven poems intended other side. [4|3] for children had regular rhythms and rhyme Then turn not pale, beloved snail, | but come and join the dance. [4|3] schemes that were simple to perform and to Will you, won’t you, will you, won’t you, | appreciate. Unsurprisingly, the features were will you join the dance? [4|3] utilized for instructional narratives intended to Will you, won’t you, will you, won’t you, | be remembered. Perhaps the most famous will you join the dance” [4|3] example of such didactic verse is “The Spider and the Fly” by Mary Howitt (1799–1888): The explanation once more lies in the underlying 4x4 nature of the metre: in the “Will you walk into my parlour?” | said the opening lines it is simply possible to realize Spider to the Fly, [3f|3] all of the beats, rather than leave them as rests “’Tis the prettiest little parlour | that ever you (though this is not to be seen as required). did spy; [4|3] Inevitably, other comic poems in my The way into my parlour | is up a winding stair, [3f|3] sample parodied traditional nursery rhyme to And I’ve a many curious things | to shew comic effect, once more exploiting the when you are there.” [4|3] satirical associations of the form. One example in the sample comes from the priest At the climax of this cautionary tale the fly Richard Harris Barham (1788–1845): allows herself to be persuaded to approach the beguiling spider, who promptly seizes her and Mother Goose’s Tale; Or, Nursery Rhymes for Christmas, 1828. drags her off to be devoured in his “dismal den”. The conclusion of the poem makes the I sing a song of “sixpence”, a balance all moral similarly inescapable: awry, My “friend and benefactor” a piping of his And now dear little children, | who may this eye! story read, [3f|3] When his mouth was opened the clerk began To idle, silly flattering words, | I pray you to sing, ne’er give heed: [4|3] “Is not this a pretty dish? suppose we both Unto an evil counsellor, | close heart and ear take wing! and eye, [4|3] The folks are in the counting-house, counting out the money, 45

Five bankers in the parlour are looking rather The poems that lack the ballad cadences do funny!” not require an isochronic reading in order to The clerk then in the carpet-bag was packing be regular – they will accommodate an up his clothes, accentual-syllabic approach (having seven “Send out and call a ‘Jarvey’, ere somebody accented syllables in each line, albeit loosely ’turn nose!” arranged). As with all such classifications, In a juxtaposition of the sacred with the there will be poems that seem to straddle the profane, the form was also used for categories. Kipling’s “Snarleyow” serves us devotional texts. In fact, it seems to have been with examples of lines where the a rather more common use of the form. An classification as a Ballad Seven seems example of this type of poetry can be seen in somewhat indeterminate, insofar they might “To the Unknown God” by Edmond Gore or might not be read with a feminine ballad Alexander Holmes (1850–1936): cadence in the initial hemistich: O God! O Father of all things! | O Lord and 1. The Driver ’umped ’is shoulder, | for the Giver of life! [4|3] wheels was goin’ round, [3f|3] O fountain of peace and blessing! | O centre 2. The Driver ’umped ’is shoulder, for | the of storm and strife! [3f|3] wheels was goin’ round, [4|3] The waves of thy will roll onward: | I stand alone on thy shore: [3f|3] In this example, the comma would seem to I veil mine eyes in thy presence: | I seal my encourage the placement of the caesura as in lips, – and adore. [3f|3] (1), but the alternative is to stress the conjunction for, as in (2). I have elected not to Ballad metre is broadly analogous to hymn- consider such lines as exhibiting ballad metre in that both exhibit the same cadences, cadences; they do not, therefore, qualify the which arise from similar isochronic poem as Ballad Seven, despite the possibility principles. It is therefore unsurprising that the of this interpretation. Nevertheless, the fuzzy Ballad Seven form should be used for poems nature of this division between the two forms of religious fervour, such as this one by of Sevens is instructive. In terms of subject Holmes. matter, and other formal features such as This association may have informed stanza length and rhyme scheme, the 4–3 Yeats’s choice of the form for the somewhat Sevens were broadly indistinguishable from light-hearted theological discussion of the Ballad Sevens. The slight exception to this “Vacillation” section 8 (“Must we part Von being that the 4–3 Sevens included more long Hügel”), where Yeats reflects on his narrative poems, such as Samuel Ferguson’s metaphysical disagreement with the Christian epic “Conal”. apologist, Baron (Friedrich) von Hügel. In the One such poem, Dora Greenwell’s “A case of “The Lake Isle of Innisfree”, however, Story of Olden Time”, graphically reinforces the affinity with the common associations of the kinship between the 4–3 Seven form and the form is considerably fainter. common ballad form (4.3.) by opening with twenty lines of the latter before switching into The 4–3 Sevens 4–3 Sevens for the remainder. In her Poems We now return to the question of (1861), this shift coincides with the turn of the classification. As I have explained, I have page, making it look like a printing classified poems of the Sevens form that do convention (Greenwell 1861: 26–27). In not display ballad cadences at the end of the subsequent editions the lineation was first hemistich as 4–3 Sevens. At least 71 preserved, even when the shift did not poets in my sample used the 4–3 Seven form. coincide with a page break. The common This was two less than used Ballad Sevens, ballad form is typically interpreted but this smaller number of poets actually used isochronically, with an unrealised beat after the form slightly more frequently (65% used the 3-beat lines. If we accept that the 4–3 it more than once, compared to 62% that used Sevens are structurally equivalent, then they Ballad Sevens more than once), and for are similarly isochronic. considerably longer poems.

Given the somewhat arbitrary division cradle-verse and march-making verse.” The between the categories, it makes sense to also latter was bound by “time measures”; the consider the total number of poems in Sevens former, she argued, was not. In this, she was in either form found in the sample by my opposing her friend Coventry Patmore, for method. By my count, just over half of the suggesting that it should be, in his Essay on poets in this sample used a form of Sevens for English Metrical Law: at least one poem (117 of 223). It is important If Mr. Patmore really intends that his Odes to remember at this point that I have only shall be read with minim, or crotchet, or found a minimum number of poems: more quaver rests, to fill up a measure of beaten may yet be found by alternative methods. time, we are free to hold that he rather Assuming that the sample is representative for arbitrarily applies to liberal verse the laws of the whole corpus, we can conclude that more verse set for use – cradle verse and march- than half of the poets in the broader marking verse (we are, of course, not nineteenth-century canon employed this form. considering verse set to music, and thus compelled into the musical time). Liberal Conclusion verse, dramatic, narrative, meditative, can In spite of this popularity, and the distinctive surely be bound by no time measures – if for genealogy of this form, it goes unremarked in no other reason, for this: that to prescribe pauses is also to forbid any pauses most histories of form. Yeats took the form unprescribed. (Meynell 1905: 94–95.) for “The Lake Isle of Innisfree” from one of his celebrated Irish precursors, yet even this Her feelings about the plurality of types of has not been recognized. Instead, that poem, verse and metres ring true today: there are few and his others written in Ballad Sevens, have advocates for treatment of all poetry as 8 been interpreted entirely differently: as isochronic. But so unpopular is this approach irregular idiosyncrasies. In fact, they are part to prosody that even scholars of poetic form of a much broader and significant strand of cannot recognize an isochronic poem in poetry that seems to reach back to the very Ballad Sevens, even though we have audio roots of the English language. recordings of deliveries of this type. This It is a powerful reminder for all critics that becomes a particular problem when a poet has our perception is contingent upon vantage- used the form for verse that is not self- point, and our conceptions of metre are no evidently ‘for use’ – like “The Lake Isle of exception. Today, isochronic performances of Innisfree”, or indeed any of Yeats’s poems in poetry are common only in Slam Poetry the form. events, not in readings of ‘serious’ literary Although some listeners object to Yeats’s work. This division in cultural esteem seems particular delivery of “The Lake Isle”, all of to have been in place since at least the 19th the poems I have explored in this article century, and probably stretches back much benefit from a performance that is in some further. The poems I have identified in this sense isochronic. With that, I return to form include a disproportionate number that Patmore’s Essay, and his observation would be excluded from the category of regarding the couplet from Surrey: that the ‘serious’ poetry: folk verse, in many forms; recognition of a median caesura makes for a hymns; children’s verse; parodies. The status far more satisfying aesthetic experience than of these poems in the database is secured by the alternative, even (or especially) when it the poets’ other works that are considered comes at the cost of normal syntax. The serious and ‘literary’. failure to recognize the metre is not unique to The proper division between ‘serious’ modern readers, but it does seem to be a poetry and isochronic verse was evidently felt particularly acute problem today. Patmore by the poet Alice Meynell (1847–1922). In an reflected on a means of correcting such essay first published in 1893 on “Mr Coventry oversights: Patmore’s Odes”, she argued that what she It is very questionable, indeed, whether termed “liberal verse” followed different English verse has gained by the entire disuse “laws” than utilitarian “verse set for use – of the caesural dot, which was always

employed, until the middle of the fifteenth Works Cited century, to indicate the position of the Primary Sources caesura in those kinds of verse of which a A Children’s Book of Verse. 1987. Selected by marked caesura was an essential quality. Marjorie Rogers. Newmarket: Brimax. (Patmore 1961 [1857]: 24.) Davis, Thomas. ca. 1800. The Poems of Thomas Davis: With Notes, Historical Illustrations, etc. and an Some may think that it is time to reconsider Introduction by John Mitchel. New York: Kennedy. marking the caesura in some way in printed The English Poetry Full-Text Database CD-ROM. poetry – as we do for Anglo-Saxon poetry. 1995. Chadwyck-Healey. When it comes to the majority of Ballad Ferguson, Margaret, Mary Jo Salter, Jon Stallworthy (eds.). 2005. The Norton Anthology of Poetry. 5th Sevens, however, there are clear textual edn. New York / London: Norton. features that may serve as signposts: we Greenwell, Dora. 1861. Poems. Edinburgh: Alexander simply need to attend to them. Strahan. Hopper [Chesson], Nora. 1900. Songs of the Morning. Michael Cade-Stewart (michael.cade-stewart[at] London: Grant Richards. kcl.ac.uk) King’s College London, 26–29 Drury Lane, InteLex – The Collected Letters of W.B. Yeats. 2002. Strand Campus, London WC2B 5RL, U.K. InteLex Electronic Edition. Gen. ed. John Kelly. Jones, Ernest [Charles]. 1855. The Battle-Day and Notes Other Poems. London: Routledge. 1. Ronald Schuchard (2008: 311) confirms that LAR – Yeats, W.B. 2000. The Collected Works of W.B. Stravinsky refers to the recording of three poems: Yeats, Vol. X: Later Articles and Reviews. Ed. “The Lake Isle of Innisfree”, “The Song of the Old Colton Johnson. New York: Scribner. Mother”, “Coole Park and Ballylee, 1931”. Meynell, Alice. 1905. The Rhythm of Life and Other 2. For an example of the latter, see Yeats’s delivery of Essays. 9th edn. London / New York: John Lane line 6: Dropping from the veils of the morning. He Bodley Head. conspicuously extends the first syllable of morn-ing TSWP – The Spoken Word: Poets. Historic Recordings (TSWP). of Poets Born in the Nineteenth-Century. 2003. 3. See e.g. Henderson 1937: 15; Clinton-Baddeley British Library CD. 1941: 65. VP – Yeats, W.B. 1987. The Variorum Edition of the th 4. See his letter to Dorothy Wellesley, 25 September Poems of W.B. Yeats. Ed. Peter Allt & Russell [1935] (InteLex: #6363): “The work of Irish poets, Alspach. New York: Macmillan. quite deliberately put into circulation with its music thirty & more years ago, is now all over the Literature country. The Free State Army march to a tune Allison, Jonathan. 2006. “War, Passive Suffering, and called ‘Down by the Salley Garden’ without the Poet”. The Sewanee Review 114: 207–219. knowing that the march was first published with Attridge, Derek. 1982. The Rhythms of English Poetry. words of mine, words that are now folklore.” Harlow: Longman. 5. By this periodization, the earliest poet on the discs is Cade-Stewart, Michael. 2013. “Mask and Robe”. Yeats first Baron Thomas Erskine (1750–1823), whose Annual 19: 221–234. poems were published in the year of his death. His Clinton-Baddeley, V.-C. 1941. Words for Music. contemporary, William Blake (1757–1825) is Cambridge: Cambridge University Press. included on the previous disc, however, as the Ellmann, R. 1967. The Identity of Yeats. London: Faber. majority of his poetry was written and published Greene, Roland, & Stephen Cushman (eds.). 2012. prior to 1800. By the same principles, G.M. th Princeton Encyclopedia of Poetry and Poetics. 4 Hopkins (1844–1889) is included as a 19th-century edn. Princeton: Princeton University Press. poet, despite the fact that most of his poems were Hayes, Bruce P., & Margaret MacEachern. 1998. first published in 1918, long after his death. “Quatrain Form in English Folk Verse”. Language 6. For Kipling’s poems, and all other poems without a 74: 473–507. print reference, please see The English Poetry Full- Henderson, W. 1937. Victorian Street Ballads: A Text Database CD-ROM. Selection of the Popular Ballads Sold in the Street 7. That the form has long been associated with verse in the Nineteenth Century. London: Country Life. for children may go part of the way to explain why Martin, Meredith. 2012. The Rise and Fall of Meter: Alfred Noyes’s “The Highwayman” is included in Poetry and English National Culture, 1860–1930. modern anthologies of verse for children, in spite of Princeton / Oxford: Princeton University Press. its grisly and bloody account of the suicide of a Patmore, Coventry. 1961 [1857]. Essay on English young lady – a sinister glamour that seems atypical Metrical Law: A Critical Edition with a of works considered suitable for children today. See Commentary. 1961. Ed. Sister Mary Augustine Roth. A Children’s Book of Verse (1987). Washington DC: Catholic University of America. 8. Joseph Phelan (2012) has recently explored Phelan, Joseph. 2012. The Music of Verse: Metrical Coventry Patmore’s endorsement of the isochronic Experiment in Nineteenth-Century Poetry. approach to poetry. Basingstoke: Macmillan.

Schuchard, Ronald. 2008. The Last Minstrels: Yeats Stravinsky, Igor. 1959. Conversations with Igor and the Revival of the Bardic Arts. New York: Stravinsky. Garden City, NY: Doubleday. Oxford University Press. Vendler, Helen. 2013. “Vacillation: Between What and Shmoop Editorial Team. 2008. “The Lake Isle of What?” Yeats Annual 18: 151–168. Innisfree Tough-o-Meter”. Available at: http://www.shmoop.com/lake-isle- innisfree/difficulty.html (last accessed 29.11.2015).

Formulaic Language in Minimal Metrical Requirements: The Case of Post-Medieval Icelandic þulur Yelena Sesselja Helgadóttir, University of Iceland

This paper discusses the formulaic language in post-medieval Icelandic þulur, a poetic genre with minimal and inconsistent metrical requirements. It argues that the metrical properties of formulae, in different applications, are one tool to maintain the metricality of the genre and hence distinguish it from prose. In this paper, I aim to briefly discuss the intersecting, and variable.2 These texts are problem of formulaic language in post- predominantly uttered or recited, or else medieval Icelandic þulur, a poetic genre with chanted. minimal and inconsistent, yet essential, The rhythm of PMÞ is so irregular that metrical requirements. I will also demonstrate they are on the borderline between verse and some strategies with regard to how, on the prose. PMÞ are not divided into strophes, but one hand, the formulaic language of þulur is rather – if at all – into units of narration: adapted to metrically unstable conditions and, sections that consist of one or more episode on the other hand, how this formulaic or/and list, possibly accompanied by some language creates and maintains metrical space free-standing motifs.3 They make use of around itself. As this paper presents ongoing metrical devices that are characteristic of most research, the findings introduced here are Icelandic post-medieval poetic metres, such as preliminary. However, the potential alliteration and rhyme – both end rhyme and significance of these findings means it is internal rhyme (less frequently, and chiefly appropriate to bring them into academic within the same line), full rhyme and discussion in the context of this volume. consonance (and even occasional end assonances, which is unusual in Icelandic Post-Medieval Icelandic þulur poetry) – but never consistently. The lines of Post-medieval Icelandic þulur (PMÞ) are lists PMÞ are of unequal length (whether of names, sequences of short motifs and/or measured in stresses, syllables, or morae), and longer narrative episodes (that also consist of their rhythm lends itself to both accentual and names and motifs) in a very irregular metrical accentual-syllabic analysis,4 but with equal form. The narration in PMÞ is fragmentary reluctance to each. Some lines are not woven and seldom entirely coherent. PMÞ are deeply into the text by any formal metrical means. rooted in the Old Icelandic þulur of PMÞ are thus not ‘metrical’ in the sense that mythological names and poetic synonyms, but there is no single metre or metrical scheme they also draw on the material of a peculiar going through a PMÞ text. However, one or genre of post-medieval Scandinavian poetry more metrical means (as opposed to syntactic 1 known as ‘rigmaroles’ or ‘rhymes’ – and or semantic means), i.e. rhyme and/or supposedly on medieval traditions of oral alliteration and/or tact, are involved in the poetry that may have thrived alongside the great majority of PMÞ lines and exercise more learned Old Icelandic þulur mentioned metrical function: they provide both above. PMÞ are folklore texts: predominantly demarcation from and cohesion with their oral (the collection of PMÞ texts in the mid- adjacent lines. These metrical means can be th to-late 19 century and subsequent limited less strict than is considered proper in the publication apparently did not have a serious majority of post-medieval Icelandic meters impact on the oral tradition), fragmentary, (cf. e.g. the unaccented rhyme in example (0),

49 ll. 9–10). They may in this sense be regarded stroke me with a stick onto my (nape of the) as compensatory strategies to maintain the neck, rhythm of PMÞ in the lines where more the stack started burning, proper metrical means are absent. I got (/was allowed) to run Nonetheless, these metrical devices and to the bishop’s land. The bishop had a good farm, strategies apparently suffice to make the lines [he] gave me a bull and a cow, fairly ‘well-formed’ (Frog 2014: 68–70) and 5 the bull started growing, their metre socially recognizable. (In a the cow started milking, similar way, a patchwork blanket is readily the Virgin Mary gave me a ram, recognizable as a blanket even when the it turned out to be big wealth for me, patches are of irregular size, shape, and etc. colour; even when some of them overlap, In spite of this loose metrical structure, PMÞ while gaps are left elsewhere.) Only in are perceived as verse in their milieu, as exceptional cases (those comparable with indicated by, for example, the fact that PMÞ gaps in the metrical patchwork) does it prove are laid out in the same way as metrical necessary to appeal to non-metrical poetic poetry in many informants’ transcriptions (cf. features to demonstrate cohesion of a certain example (1d) below). line with its adjacent lines. The terms ‘metre’ and ‘metrical’ as referring to PMÞ in this Formulaic Language paper should be viewed in the light of this 6 PMÞ use formulaic language as part of their preliminary discussion. oral compositional strategies. Discussions of The following introductory example (0) is th formulaic language in oral poetries have from the late 17 century. It does not have evolved especially through Oral-Formulaic any loose lines, even though it is one of the Theory, beginning from the seminal work of first PMÞ to be written down. Alliteration is Milman Parry (1928) and then spreading underlined, rhyming words are indicated with internationally especially through the double-underlining, and other repetitions are publication of Albert Lord’s The Singer of indicated with dash-underlining. Note that Tales in 1960. Parry defined an oral-poetic alliteration and rhyme are only at times used formula specifically in terms of the metrical in a way that fully complies with Icelandic positions in which it appears (Parry 1928: 16; metrical rules; therefore, for the purpose of repeated in Lord 1960: 4). Research on this study, alliteration is marked in all Icelandic poetries has in general continued to relatively strong metrical positions (including view formulaic language through the lens of those where the alliteration would not be Parry’s definition (although cf. Acker 1998; acceptable in, for example, Icelandic rímur Frog 2011; also Smirnitskaya 1994: esp. 211, poetry), rhyme is marked even if it is not 221–225).8 From this view, discussing strict, etc. formulae in PMÞ would be problematic (0) Holm papp 64 fol, 99v owing to their loose metrical form, where Eg sat under Fiſkahlada faudur mÿnz, furthermore listing seemingly prevails over menn komu ad mier, narration as the main compositional principle räku staf i hnacka mier, (Yelena Sesselja Helgadóttir 2001). However, Hladinn tök ad brenna, Oral-Formulaic Theory research has advanced eg fieck ad renna considerably since 1960 (see e.g. Foley & uppa biſkupz land, Ramey 2012) and formulaic language has Biſkup ätti Gott bv, gaf mier vxa og kv, been clearly demonstrated as a compositional Vxinn tök ad vaxa7, strategy in poetry that lacks regular metre and kyrinn tök ad miölka, even in oral forms customarily described as 9 Sancti Maria gaf mier saud, prose. Listing/enumerating formulae are þad vard mier ad miklumm aud, known from other genres, both foreign and etc. Icelandic, all the way back to Homer (Powell I sat under my father’s fish stack, 1978; Edwards 1980; cf. Ziogas 2013). It is men came to me, therefore neither theoretically nor

50 methodologically problematic to approach Little Tóa mounted her white horse […] formulaic language in PMÞ, where formula She rode the most of all women will be understood as a fixed or flexible (1b) ÍF I: 224, Kvæði af Ribbaldi og Gullbrúnu verbal unit in poetic composition that has Hann sette hana ä hvijtan hest … identifiable metrical features and carries allra kvenna reid hun best distinguishable traditional meanings or associations.10 Two types of poetic formulae He put her on a white horse […] in PMÞ will be analysed here to approach this She rode the best of all women question through three case studies. The first (1c) ÍF III, 226, Vallara kvæði two cases address formulae that can plausibly Hann tok sinn hinn hvïta hest, [...] be considered to have been adapted into PMÞ rejd hann sem ad kunni best from other Icelandic poetic genres. PMÞ often He took the white horse of his, [...] incorporate material from other contemporary He rode as well as he was able genres, and Cases 1 and 2 will consider formulae that seem to have originated in the This formula appears to have eventually been genre of sagnadansar (the Icelandic form of incorporated into PMÞ, yet its context is quite Scandinavian popular ballads), which are different from that in sagnadansar. The ballad among the most formulaic post-medieval narration is absent. However, the formula in Icelandic genres. Case 3 will then turn to question settles down in PMÞ in a short (up to formulae that are likely to be intrinsic to PMÞ. approx. ten lines) narrative sequence where the formula introduces the onset of the hero’s Case 1: The hest–mest Formula journey, exactly as in the ballads above – In a number of sagnadansar, we find the except for the fact that narration is in the first word sequence that will here be referred to as person, thus the pronoun hann/hún [‘he/she’] the hest–mest [‘horse–most’] formula and is exchanged for ég [‘I’] with associated which describes the onset of a hero’s journey morphological variation of the verbs. (or the whole journey): ‘X mounts a white In terms of the rhythm, the formula also horse (variant: X puts Y on a white horse) / falls into a suitable metrical environment. he/she rode the most/best of all men/women’. Icelandic ballad meters are not strict, but they This formula often occurs in: are still regular – compare the description above, where line length (four stresses) and couplets with four-stress lines rhymed aa accompanied by a refrain, usually split (the rhyme scheme are specified for the formula in first half of which is inserted between the question; alliteration is not mandatory, as lines of the couplet) (Vésteinn Ólason 1982: ballad metres are presumably not of Icelandic 15). origin (Vésteinn Ólason 1982: 15). Unlike sagnadansar, PMÞ generally have two The refrain generally consists of 2–3 lines and (around 55%) or three (around 30%) primary has its own rhyme scheme, which, combined stresses per line, and only seldom four with the rhyme of the couplet, produces (around 13%) or more (NB: these statistics different rhyme patterns, such as abab and remain preliminary). However, in the ababb. The content of the refrain – which is fragment where the formula usually occurs, it often an aphorism or gnomic wisdom – is often preceded by lines with four or even correlates with the ballad narration on the five stressed positions. In addition, the whole, but it does not interfere with the considerable metrical similarities between narration in the couplet, whose lines form a PMÞ and ballads ease the formula’s transition coherent clause or clauses and whose syntax between the two genres. The end-rhyme, is not impacted or complicated by that of the when present in PMÞ, is chiefly couplet refrain. In the following examples (1a–c) the rhyme, aa or AA (rather than alternating or refrain is therefore omitted and its first part chiastic rhyme), which is precisely one of the marked as ‘[…]’: main metrical cores of the formula in (1a) ÍF I: 8, Kvæði af Tófu og Suffaralín question. Alliteration is not indispensable in Toa litla stie ä hvijtan hest … either PMÞ or ballads, and when it is present, allra kvenna reid hun mest

51 it does not need to follow the strict scheme the rhyme – is sufficient to integrate the lines familiar from other Icelandic poetry with one into the acoustic texture of the poetry. or two alliterating phonemes in the first line In addition, it is noteworthy that from the of the couplet and one in the first strong very beginning of its life in PMÞ, the formula position of the following line. Internally in question seems to have had two alliterating lines, as resulting from the phrase manifestations (unless they are considered as “hvijtan hest” in the hest–mest formula, are two related formulae). The form that is not unusual. Put simply, the formula could attested slightly more frequently in the corpus enter PMÞ unchanged. (especially in younger texts) is: ‘I had a(n) In spite of this possibility, the hest–mest [attribute] horse, I rode the most of all men’, formula in PMÞ differs slightly from what it as seen in example (1d) (notice that full-stops is in the ballads. Both lines are one stress that appear in 1d are used to separate metrical shorter, which could be a part of the way in lines, as is the custom in this and many other which the formula is adapted to the rhythm of Icelandic manuscripts): þulur, where two and three stresses per line (1d) Lbs 587 4to III, 14r prevail. Furthermore, the formula apparently eg átti brúnann hest. undergoes some changes in alliteration, allra manna reið eg mest. although it is difficult to tell whether those changes are induced by the aforementioned I had a brown horse[,] I rode the most of all men difference in the lines’ length – as the new metrical environment can in turn reconfigure The attributes vary from the horse’s colours to the formula’s potential flexibility for variation “vakran” [‘pacing.ACC’] (SÁM 30a, 119r & – or whether they are related to the length 242r), “sprangan” [‘strutting.ACC’] (SÁM 25f, change in some other way (if at all). 29r), and even the slightly surprising “haltan” Surprisingly, the horse in the PMÞ texts is [‘halting.ACC’] mentioned above. Much more never white, which means that the alliteration surprising is, however, the other, reverse in the hest–mest formula of sagnadansar is modification of this formula: ‘I ran the most lost.11 Only in a couple of cases is this of all men, I had no horse’, illustrated in compensated by finding another /h/- example (1e): alliterating slot-filler: e.g. “haltan” (1e) JS 289 8vo I, 28v [‘halting.ACC’] (Lbs 587 4to I, p. 135 hlióp eg allra manna mest, according to the ms. pagination), or just átti eg mer aungvan hest “hafði” [‘had’] (e.g. SÁM 30 [unregistered 12 I ran the most of all men, material]). In quite a number of cases, I had no horse especially from the 20th century (when people begin to care less about alliteration), This version also appears with a slightly alliteration in the former line remains different word order in the latter line of the 14 uncompensated. On the other hand, allra couplet, as in example (1f): kvenna [‘of all women’] in the latter line of (1f) SÁM 30 [unregistered material12] the ballad formula turns into allra manna [‘of Stökk eg manna mest, all men’] in PMÞ, where manna alliterates engan hafði eg hest with mest [‘most’]. In two cases where mest is I trotted/leapt the most of all men, for some reason substituted with another word I had no horse – best [‘best’] (HJ)13 and, strikingly again, verst [‘worst’] (SÁM 85/289 EF, # 2574) – In spite of the inversion, the formula is readily the manna–mest alliteration is not recognizable: both the rhyming elements compensated by an alternative and thus in (mest–hest) and the overall structure are the these cases neither of the two lines of the same, even though the meaning of the slot- couplet alliterates. However, this does not filler almost turns the formula inside out seem to affect either the formula itself nor its semantically. The function of the formula has context, since lack of alliteration is common nonetheless been preserved, since even in this in PMÞ: another of the metrical means – here, modification it introduces the onset of the hero’s journey. No metrical problems follow 52 from this inversion, since the rhyme, which vpp j hæga lofftinu holds the formula together, is in its place. On süafu þau bædj saman the contrary, this inversion often enriches Lord Jón and Ragnfríður, alliteration, as engan [‘no-one.ACC’] alliter- They spoke many a merry word, ates with átti [‘had’] in a number of cases In the cosy loft (e.g. DFS 67 E, 188r). On the whole, both They slept together lines that comprise the formula alliterate internally in about 2/3 of cases – which means (2b) ÍF I: 131, Ásu dans Þar var bædj gledj og gaman […] that, in those cases, the formula in PMÞ is herra Petur og Äsa gefinn voru samann metrically richer than in the ballads, even though it is not required to be so, since the It was both joy and fun […] metrics of PMÞ are formally more relaxed, or Lord Pétur and Ása got married have a low density of metrical requirements. Surprisingly, another modification of this When the formula finds itself in this low- formula is used as a closing construction density metrical atmosphere, it apparently when the denouement of a love story is tragic: seeks to enrich and maintain the metre in and (2c) ÍF I: 153, Kvæði af Jóni og Ragnfríði around itself, and thus to compensate for this Þar var meire grätur low metrical density. enn þar var gamann, tuó foru þau lÿkin Case 2: The gaman–saman Formula j eina steynþro saman Another example of formulaic language that PMÞ plausibly adopted from Icelandic ballads There was more weeping is the word sequence that here will be referred Than there was fun, Two dead bodies to as the gaman–saman [‘fun–together’] Went together into the sarcophagus formula. In sagnadansar, this formula is found in couplets like the hest–mest formula (2d) ÍF I: 229, Kvæði af Ribbaldi og Gullbrúnu of Case 1, and also in quatrains with Þar var meire gratur enn gaman, [...] alternating rhyme (where the refrain is not þriu foru lyk j steinþrö samann split but follows upon each strophe). In the There was more weeping than joy, […] latter cases, the formula is distributed over a Three dead bodies went together into the greater number of narrative lines than in sarcophagus ballad couplets (where the lines containing a It is the first modification of the formula formula are only separated by the split refrain, considered here that eventually makes its way not interfering with the narrative in the into PMÞ. We can observe it on its way into couplet, as mentioned above). The syntax of a PMÞ in a comic verse that is sometimes quatrain containing this formula is more attached to different PMÞ texts (e.g. ÍF VII: complicated than one of a couplet with the 95; cf. ÍF VIII: 114), where the formula is same formula. In a quatrain, the formula in used in the same introductory function even question can, for example, interact with other though the verse’s plot is rather a parody of a formulaic expressions such as gleði og gaman love story: [‘joy and fun’] as seen in example (2b) below. The essence of the gaman–saman formula (2e) ÍF VII: 95, Kvæði af pilti og stúlku is precisely what its designation by the Vinnumadur og vinnukona rhyme-pair gaman–saman conveys: ‘it is töludu sier til gamans pleasant to be (talk, sleep...) together’. hvurnin ætlardu ad fæda mig þegar vid tökum samann However, it is more specifically used with reference to a couple in a romantic situation: A workman and a workwoman, the formula introduces a love story or They spoke many a merry word, concludes it with a happy ending: How are you going to feed me When we start living together[?] (2a) ÍF I: 149, Kvæði af Jóni og Ragnfríði Herra Jön og Ragnfrÿdur In PMÞ, however, the formula only occurs in tóludu þau sier gaman, couplets with lines of the same length as in

53 ballad couplets, i.e. four stresses per line, alliterates with the auxiliary verb even though ballad quatrain lines are often myndi/mundi [‘would be’] – or, occasionally, shorter and thus fit better to the rhythm of with words like mörgum [‘to many (people)’] PMÞ. For example: (“Mörgum þætti meira gaman...” [‘To many (2f) DFS 67 E, 430r [people], it would seem more fun…’] (DFS Manstu nokkuð meira gaman, 67 E, 431v)). The gaman–saman formula, as er við lágum bæði saman? well as the hest–mest formula, thus exhibits more metrical stability than is present in these Do you happen to remember more fun ballad formulae, and considerably more than When we laid two together? PMÞ metrically require. However, taking a (2g) AM 247 8vo, 5r step away from the metrical stability in favour Ekki kann eg meira gaman of variation is also clearly acceptable, as enn þau láu bæði til samans example (2g) above shows: neither of the two I don’t know more fun lines of that example alliterates properly. This Than [when] they laid two together is, nonetheless, both permitted in PMÞ and suitable for the formula, since it is the rhyme Thus, line shortening was not part of the (gaman : saman) – not alliteration – which is adaptation strategy of the formula to the new crucial for holding the formula together. In metrical environment. In fact, hardly any this sense, we can say that variation prevails adaptation was needed: the ballad formula over alliteration. Moreover, the end rhyme in was able to enter into PMÞ unchanged (the this particular example is not perfect (gaman : same possibility that the hest–mest formula til samans), but this apparently does not affect had, yet did not use). Furthermore, metrical either the rhythm of this PMÞ text or the variation in the number of stressed positions formula, which is evidently viable even when in a line remains quite minimal, even though the metrical requirements are minimal. PMÞ leave a lot of room for such variation. This movement of the formula between Alliteration, however, behaves differently: two poetic systems with different metrical while it was only occasional in the ballad requirements – sagnadansar and PMÞ – could formula, the PMÞ formula develops more be compared with the ‘metrical migration’ of stable alliteration. This can be seen in the some eddic formulae between different types modification shown in examples (2f–g) of metrical lines in fornyrðislag and above: ‘Do/Don’t you remember (variant: I ljóðaháttr described by Ilya Sverdlov (2011). don’t know) more fun when/than/though From his discussion, it appears that the we/they both lay/lie together.’ It is also potential of formulae to shift between observable in the other modification, which is metrical models was considerable already in found in a greater number of examples, eddic verse – as was the potential of shifting especially in younger texts: ‘It would be/seem between metrical models (such as short line more fun if we/they all run/come together,’ as and full line, or – operating with larger units – illustrated in examples (2h–i): such as stanzas in ljóðaháttr and galdralag) (2h) DFS 67 E, 420r within a text, without violating its metrical myndi þykja meira gaman, integrity. This is of importance for PMÞ ef þeir kæmu allir saman because such shifting between different Would it seem more fun metrical models within a single text, taken to If they would come all [MASC] together[?] the extreme, can in turn result in the patchwork of numerous different metrical 15 (2i) Lbs 587 4to II, 69v models, which to a great extent accounts for Mundi verða meira gaman, the rhythmic diversity of PMÞ. ef við hlypum allar saman In spite of the apparent potential for the Would it become more fun ‘metrical migration’ of some formulae If we run all [FEM] together[?] between the poetic systems of sagnadansar The word meira [‘more’] becomes a virtually and PMÞ, there seem to be some restrictions invariable part of the formula in PMÞ and on such movement. The mismatch in the

colour of the horse in the hest–mest formula L S X1-gerður E1 in sagnadansar, with their predominantly ‘another woman’ X2-gerður E2 white horses, and PMÞ, where white occurs L = location: ‘out(side) / in(side) very seldom, despite this word seeming to be S = a verb of situation (rides/stands/sits) in metrically preferable as it would have third person singular present or past tense provided an alliterating phoneme (however, X = a monosyllable compounded (almost see note 11) is a probable indicator that, once without exception) with -gerður forming the formula became established in PMÞ, its a three-syllable woman’s name in the use and development remained separate from nominative case the corresponding sagnadansar formula. If E = a two-syllable derogatory descriptive the diction of these genres had been open to epithet / byname of the woman named in the line one another in relatively free or even moderately fluid interaction, then we would In spite of the number of open slots in this expect the horse to be either occasionally formula, the slots are semantically and/or white in PMÞ or more frequently to be a metrically conditioned or conventionalized colour other than white in sagnadansar. In a beyond the requirements of the metre or similar way, we would expect that more than syntax. The formula is also part of a larger one modification of the hest–mest formula syntactic structure in which a third woman is and/or of the gaman–saman formula would almost invariably mentioned in the following migrate from sagnadansar to PMÞ – and that line. The lines normally rhyme with each some PMÞ-modifications of the formulae other as in example (3a), but this is not a would travel back to sagnadansar. The lack necessity and there are exceptions as of free movement of formulaic diction illustrated in example (3b): between Icelandic poetic systems has been (3a) AM 148 8vo, 254r observed by Haukur Þorgeirsson (2012: 193– ut rydur Arngierdur hnuka 194) in an earlier volume of this journal Ønnur kona Þorgierdur s[l]uka17 (comparing genres of sagnakvæði and rímur). Out she rides, Arngerður Crouchy[,] This observation was then carried further by Another woman, Þorgerður Gulpy Frog (2012), who theorized that oral poets do not, without conscious effort, have cognitive (3b) AM 154 8vo XII, 1v access to formulaic diction of one type of utj situr Hallgerdur hlun[ch]a 18 poetry while composing or performing in onnur kona Þorgerdur Sluka another. Although the difference in the colour Out she sits, Hallgerður Dumpy[,] of the horse in the formula of the Another woman, Þorgerður Gulpy sagnadansar and PMÞ may seem to be a This formula is different from the two trivial detail in itself, it may have significant formulae considered above in that the two implications for how formulaic language rhyming words (the bynames) do not form the works in these poetries and for the historical fixed backbone of the formula, but are its relationships between them. variable elements instead. Thus, the rhyme is Case 3: An Enumeration Formula not an indispensable part of the metrical core The third example considered here is an of this formula. This notwithstanding, the two enumeration formula that is more likely to be lines containing the formula are fairly stable intrinsic to PMÞ insofar as listing – and thus metrically: both have relatively stable length enumeration – can be considered the main (especially the latter line, where variation compositional principle of this type of poem. chiefly occurs in the woman’s byname (E2)) This formula is almost exclusively preserved and are quite rich in alliteration. in Þornaldarþula,16 of which there are, To begin with, the onsets of each line often however, about 100 attested texts from the alliterate with one another. The first word of 17th–20th century. The formula can be the latter lines is – almost invariably – önnur summarized as follows: [‘another’], while the first word of the former line is usually either inn(i) [‘in(side)’] or út(i) [‘out(side)’]. Both inn(i) and út(i) are quite

55 strongly stressed and alliterate with önnur, less frequently Valgerður (cf. Lbs 1057 4to although the distance between the alliterating p3, 159r) and Þorgerður. This choice is phonemes would be too great to consider this obviously related to the fact that the byname proper alliteration in the Icelandic metres that (E1) begins with /h/ in roughly half of the are more strict than the metre of PMÞ. In examples – varying from the most common those exceptional cases when the first word of hlunka [‘Dumpy’]: see “hlun[ch]a” in the former line is different (e.g. heima [‘at example (3b); cf. “hlúnka” (587 4to IX, 41v, home’]: Lbs 1057 4to p2, 157r), this 46r) and húka [‘Crouchy’] (Lbs 587 4to IX, alliteration is always compensated by another 52r), its close synonym hnúka (see “hnuka” in alliteration, e.g. the first word will then example (3a)), hnúta [‘big bone, condyle’] alliterate with the name of the woman in the (AM 960 4to 8, 3r), etc. When the name (X1- same line (X1-gerður). In addition, the name gerður) also begins with /h/, this provides an of the woman in the opening line (X1-gerður) independent alliteration within the line. As often alliterates – either with inn(i)/út(i) (and, mentioned above, this alliteration was in most occasionally, heima etc.) in the beginning of cases merely additional to the basic 1+1 the line or with the woman’s byname (E1). In alliteration – or replaced it. This alliteration the former case, alliteration follows the 2+1 was by no means required by the metre of pattern – i.e. two syllables in the first line PMÞ, and quite a number of texts did well alliterate with the first stressed syllable in the without its addition when using this formula. following line of a couplet (e.g. “ut rydur However, this additional alliteration could Arngierdur… : Ønnur” in example (3a)), still be regarded as a kind of compensation for which is mandatory for most Icelandic poetry the loss of the 2+1 alliteration in this two-line with three or more metrical stresses per line. system. In other words, it provided an PMÞ lines often have three or more stresses alternative medium to create and maintain a and might thus be expected to follow the 2+1 metrically appropriate space in and around the pattern, but in practice they are an exception line (on alternative strategies for integrating to this rule (as are sagnadansar, whose metres lines lacking alliteration into the acoustic are of foreign origin). In PMÞ, the pattern 1+1 texture of Kalevala-metre poems, see Frog & (such as “utj” : “onnur” in example (3b)) is Stepanova 2011: 201). Another literary factor, widely accepted. In the latter case, the possibly related to the choice of the name independent alliteration within the line Hallgerður, is the apparent avoidance of this (“Hallgerdur hlun[ch]a”) is usually additional name in the Icelandic society in the 19th to the 1+1 alliteration, which connects the two century (Guðrún Kvaran 1991: 19). This lines; it may even replace this alliteration, avoidance could be related to the growing since these two substantives, which stand popularity of Njáls saga and thus people’s close together, produce a more pronounced knowledge that Hallgerður is a cruel and alliteration than the basic 1+1 pattern. malignant vixen: the name is hardly suitable In earlier PMÞ texts – i.e. texts that are for newborn children, but it is perfect for use first attested before the 19th century – the with a derogatory byname in the context of names appearing in the first line (X1-gerður) this þula formula. predominantly begin with a vowel (Arngerður The fact that one name out of many – and later, once, Ásgerður: Lbs 587 4to IX, becomes dominant in a certain position 36r) and thus alliterate with inn(i)/út(i) and demonstrates a tendency to non-variation in önnur in the openings of the two lines. terms of both slot-fillers and different However, later the use of such names recedes, modifications of one formula. This tendency leading to the loss of the more generally is also noticeable in the development of the proper alliteration (2+1) in these two lines. other formulae considered here. Where the The name Hallgerður, already attested in the formulae have two modifications, one earlier texts (cf. example (3a) above), becomes dominant (examples (1d), (2h–i)) becomes dominant in younger sources, with while the other virtually disappears (examples Hlaðgerður occasionally found as well (cf. (1e–f), (2f–g)), thus decreasing both semantic Lbs 587 4to IX, 41v, 46r; SÁM 38, 39v), and and metrical diversity. In some cases, this is

56 directly related to the tendency of maintaining narrative fragmentariness. Each fragment that the metrical stability of the formula (e.g. in a PMÞ text consists of can be relatively the case of the name Hallgerður), while other metrical in itself (often in its own way), but cases demonstrate steps sideways in favour of this metricality does not hold for the whole semantic variation (as in case 2). Hallgerður text. On the other hand, complete metrical established itself as the dominant X1-name irregularity in a PMÞ text is not possible not least because it started with /h/ and thus either. In all three cases considered above, we carried alliteration with the customary see examples of variations that reduce phonic following byname. Once this name became resonance in one respect being compensated more regular, the E1-byname starts – quite by metrical mechanisms that do not always unexpectedly – showing increasing variation, perfectly repair the situation from the formal indeed departing from the initial /h/ and thus point of view, but nonetheless add some from the additional independent alliteration degree of metrical (or para-metrical) cohesion within the line. In the texts that are first to the lines and enhance their integration into th attested before the 19 century, the E1- the PMÞ as a poetic text. Poetic formulae bynames with an initial /h/ account for function exactly on the level of these significantly more than half of the data, fragments, and thus do not have to face shifting to approximately one half in the texts metrical irregularity to the full extent. from the 19th century; and less than half in the Poetic formulae are used in those 20th century. In a similar way, the attribute fragments that can accommodate them – that brúnskjóttan [‘pied-brown.ACC’] (e.g. HJ; is in narrative fragments where a formula can HÖE,19 of a horse in the hest–mest formula) is be used as in any other narrative, or in listing not apparent in this formula before the 20th fragments. Within a fragment, poetic century (though already quite prominent in formulae have leeway both to adapt some other PMÞ texts), yet in the course of themselves to the rhythm of the fragment (e.g. the 20th century it gains popularity and making their own lines slightly shorter, as in supersedes most other attributes. Unlike some Case 1 above, and/or less regular) and to of the attributes, whose place the word create some metrical space for themselves – brúnskjóttan takes in the line, it does not in a way similar to the creation of specific alliterate (or rhyme), thus preventing the metrical space for – and by – heiti (i.e. poetic whole line from having alliteration. In synonyms or, more accurately, poetically addition, this development has the side-effect equivalent terms) in Old Icelandic þulur. of extending the line by one rather heavy There, extended alliteration (which spans a syllable, replacing disyllabic words like half or a whole strophe), additional (not brúnan [‘brown.ACC’] (cf. Lbs 587 4to III, systematic) internal rhymes and end rhymes, 14r) with a trisyllabic compound containing morphological and syntactic parallels between two heavy syllables. Variation here the names and heiti, and other similar devices apparently prevails over both alliteration and were used for added metrical cohesion, which tact. However, it is questionable whether was appropriate for the high social status of variation prevails over the rhyme in PMÞ; the early Old Icelandic þulur (Gurevich 1984: cases considered here do not give clear 19–20 and works there cited; Gurevich 1992a: examples of this. 69; Gurevich 2012: 195–196). PMÞ, with their fragmentariness and metrical irregularity, Preliminary Conclusions utilize those and other compensating The three cases considered above show that strategies, such as non-strict rhyme, there are good grounds to apply the concept occasional complementary distribution of of poetic formula to PMÞ, despite their loose alliteration and rhyme (where alliteration metrical form. The metrical irregularity of binds non-rhyming lines and rhyme connects PMÞ does not lie in maintaining irregularity those lines that do not alliterate), etc. These in each and every line, but rather in devices are not used systematically in PMÞ systematic inconsistency that goes hand in (any more than other metrical means that hand with a high degree of both metrical and structure metres in other Icelandic poetry,

57 such as systematic alliteration); however, they 3. Frog & Stepanova (2011: 200) mention a similar are readily available – inter alia, for poetic non-strophic structuring in kalevalaic poetry. Comparison of the rhythm of PMÞ with kalevalaic formulae that use these and other devices so metres can be illuminating. Both PMÞ and actively that at times they prove metrically kalevalaic poetry are non-stanzaic and are richer than the same formulae in other genres organized on the basis of some mandatory metrical with stricter metrical requirements. These are features (including alliteration), which are, compensating strategies for the formulae to however, not required in every line and whose application is governed by convention. This survive in the irregular metrical environment apparently results in a very flexible metrical of PMÞ – strategies that prove to be effective; template (ibid.: 199–200), which is also the case in i.e. poetic formulae are viable in PMÞ. PMÞ, and in the use of other, compensating Furthermore, formulae in PMÞ function metrical devices or strategies – including e.g. almost as if they (still) were in metrical extended alliteration and interlacing alliteration (ibid.: 200–201), which have their analogues in poetry. The formulae that come from other PMÞ. genres continue to maintain certain 4. This latter is likely to be chiefly due to the rather metricality in and around themselves (even regular trochaic-based stress pattern in Icelandic though they come into a much more loosely language, be it poetry or prose (cf. Sigurður organized metrical environment), while those Kristófer Pétursson 1924 and, more recently, Anton Karl Ingason 2007). formulae that appear to originate within PMÞ 5. Сf. Frog 2014–2015 and works there cited; similar emerge and function – at least to a certain ideas about social recognizability of a ‘scant’ metre extent – as metrical units. Thus, to a great are developed in Sverdlov 1998. extent, the formulae in PMÞ conform more 6. PMÞ may also be viewed as organized on principles closely with Parry’s original definition of of poetic features (namely constellations of poetic features without clear formal constraints on these formula than was expected. constellations) including, but not limited to, some Poetic formulae prove to be so viable in metrical devices, yet without regular metrical PMÞ that they can tolerate considerable structuring. The rhythmic organization of PMÞ variation – both semantic and metrical, might then be compared not to Kalevala-metre including steps away from metrical stability – poetry, but to Karelian laments, whose poetic system is “not bound to a regularly repeating as all the cases considered above show. The metre” (Frog & Stepanova 2011: 205, see also 204– use of poetic formulae in PMÞ is thus not 209), but is largely built on para-metrical strategies, only dynamic but also dialectic – and through which the text becomes poetic. This definitely deserves further consideration and approach can be quite productive, as it puts PMÞ – discussion. and other Icelandic oral poetic genres, medieval as well as post-medieval – better into the context of Yelena Sesselja Helgadóttir (sesselja[at]hi.is) Árni oral-poetic metres (cf. Frog 2014). One’s standpoint Magnússon Institute for Icelandic Studies, Árnagarður in analysing the rhythmic organization of PMÞ – v. Suðurgötu, 101 Reykjavík, Iceland whether they are rather built on metrical or para- metrical principles – largely depends on one’s Acknowledgements: I am grateful to Frog for his understanding of metrical regularity. If metrical encouragement and useful advice at all stages of this regularity is understood as uniform application of a metrical and formulaic undertaking, as well as to complex of phonological devices and strategies in Bjarki Karlsson and Rósa Þórsteinsdóttir for their every unit (line or stanza), then the view of PMÞ as manifold advice, to Vésteinn Ólason and Ilya Sverdlov being organized on principles of poetic features – for their comments on an earlier version of this paper, rather than on metrical principles – undoubtedly to Emily Lethbridge for her help on certain stylistic describes PMÞ better. If metrical regularity is points, and last but not least to the anonymous peer- interpreted as systematic – yet not rigorously reviewers. uniform – application of a complex of phonological devices and strategies, as is done in this paper, then Notes analysing PMÞ in terms of their ‘metre’ provides 1. Danish and Norwegian remser, Swedish ramsor; us, at the very least, with a good frame of reference. also rim in all the aforesaid languages. (Not to be 7. In this and the following line, the text exhibits a confused with Icelandic rímur.) For more details, weak rhyme in the last syllable of the line (which is see Yelena Sesselja Helgadóttir 2014. normally unstressed, but obtains some stress due to 2. For a slightly more extensive description of PMÞ its final position in the line). To visualize the and my project on PMÞ, see, respectively, Yelena difference between these two rhyme types, only the Sesselja Helgadóttir 2014 and 2011; on Old rhyming vowel is underlined in this case. Icelandic þulur, see Gurevich 1992a; 1992b; 2012; 8. For reviews of earlier scholarship on formulaic and particularly, 1984. language in Old Norse poetries, see Acker 1998:

85–110; Frog 2011: 19–28; cf. Frog 2014 I: 110– study. Note that in all the three cases – but in no 112; on formulaic language in rímur poetry, see undisputed PMÞ text considered here – the horse in Kuhn 1990–1993; in sagnakvæði (vs. rímur), see the formula is of grey colour, as it is in some texts Haukur Þorgeirsson 2010; 2012; 2013; also Frog of the ballad (see Systra kvæði in ÍF VI: 83, the text 2012b; and in sagandansar (Icelandic popular that was published as early as 1852). ballads), see Vésteinn Ólason 1982: 25–29; (and, 15. The motif of coming/running together is also found on a more popular note 1979: 43–47); cf. Conroy in poetry related to dancing. 1980 on Faroese ballads. No attempt to explore 16. On Þornaldarþula, see Yelena Sesselja Helgadóttir formulaic language of PMÞ has previously been 2015. undertaken. 17. The second character of the last word is not readily 9. On formulaic language in poetries with ‘unfixed’ legible: it could also be e.g. k, or reflect an attempt metres, see Sverdlov 1998; Frog & Stepanova to change l into k or vice versa. 2011: 204–205; cf. Stepanova 2015a: 139–140; 18. The characters ch (in particular h) in “hlun[ch]a” 2015b. On formulaic language in Icelandic prose are not perfectly visible, partly because of repairs. (sagas), see Scholes & Kellogg 1966: 43–51 (cf. However, they were plausibly better legible when Lamb 2015 and works there cited). For further Jón Samsonarson, whose transcription (in JS) I scholarly discussion, see the continuously updated have consulted, read the manuscript, and the Summative Bibliography of Oral Tradition. reading is thus almost secure. 10. Cf. the definition of John Miles Foley & Peter 19. HJ: see note 13; HÖE: Hallfreður Örn Eiríksson, Ramey (2012: 80) as “an integer of traditional unpublished research materials preserved at the meaning”, and Frog’s (2011: 28–34) approach as a Árni Magnússon Institute for Icelandic Studies, morpheme-equivalent unit. With regard to metrical light-blue notebook from 1977, 11r. features, cf. John C. Ford’s (2002: 225) approach as “a metrically useful item” (as ‘metrically useful’ is Works Cited a more flexible formulation than strictly metrical). Sources 11. Except once: alhvítan [‘all-white.ACC’] (AM 247 AM 148 8vo; AM 154 8vo; AM 247 8vo; AM 960 4to. 8vo, 6v), which alliterates with átti [‘had’] rather Manuscripts at the Árni Magnússon Institute for than with hestur. It is possible that hvítur [‘white’] Icelandic Studies. is avoided in this formula in PMÞ because of the DFS 67 E. Manuscript at the Royal Library, National specific use of colour adjectives for horses in Library of Denmark and Copenhagen University Icelandic. The traditional Icelandic attribute for a Library. white horse is not hvítur but grár [‘grey’]; the use HJ = Helga Jóhannsdóttir, unpublished research of hvítur about horses is thus less natural in an materials preserved at the Árni Magnússon Institute Icelandic þula than in a ballad of foreign origin. for Icelandic Studies. However, this specificity of Icelandic colour HÖE = Hallfreður Örn Eiríksson, unpublished research adjectives can hardly be the only explanation for materials preserved at the Árni Magnússon Institute the absence of hvítur in the PMÞ formula, since for Icelandic Studies. grár is not used in this formula in any undisputed Holm papp 64 fol. Manuscript at the National Library PMÞ text either (cf. note 14). of Sweden. 12. Two manuscripts by Ástríður Thorarensen from ÍF = Íslenzk fornkvæði: Islandske folkeviser I–VIII. Ed. Hallfreður Örn Eiríksson’s collection are among the Jón Helgason. Editiones Arnamagnæanæ B 10–17. material which is marked as SÁM 30 on the København: Munksgaard, Reitzel, 1962–1981. containing folder but is not currently registered JS = Jóns Samsonarson, unpublished research materials under SÁM 30. The first ms. is a photocopy (14 preserved at the Árni Magnússon Institute for fols.) marked “Ástríður Thorarensen, skráði um Icelandic Studies. 1920–1930” [‘Ástríður Thorarensen recorded JS 289 8vo. Manuscript at the National and University around 1920–1930’]; see 5v there. The second ms. Library of Iceland. is a school notebook with Ástríður’s records from Lbs 587 4to; Lbs 1057 4to. Manuscripts at the National 1970–1971 (according to Hallfreður’s markup on and University Library of Iceland. the cover’s inside); cf. 7r there. SÁM 25f; SÁM 30 [unregistered material]; SÁM 30a; 13. Helga Jóhannsdóttir, unpublished research SÁM 30b; SÁM 38. Manuscripts at the Árni materials preserved at the Árni Magnússon Institute Magnússon Institute for Icelandic Studies. for Icelandic Studies, transcriptions that come along SÁM 85/289 EF. Audio tape in the Folklore Collection with Sólveig Indriðadóttir’s letter from 1968, IV 5r. of the Árni Magnússon Institute for Icelandic 14. The reverse modification of the formula is Studies. Available digitally at: particularly common in older (19th-century) texts. http://www.ismus.is/. This could point to a later reintroduction of the Summative Bibliography of Oral Tradition. Offered by ballad formula into PMÞ (in its non-reversed the Center for Studies in Oral Tradition. Available modification); cf. in this context that Vallara kvæði at: http://www.oraltradition.org/bibliography/. (ÍF III: 224–227) is recorded – without its refrain – as a þula at least three times throughout the 20th century (e.g. SÁM 30b, 150r–151r). This possible reintroduction requires, however, a more thorough

Literature Haukur Þorgeirsson. 2013. “How Can You Tell Who’s Acker, Paul. 1998. Revising Oral Theory: Formulaic Talking? – Transitions between Direct Speech and Composition in Old English and Old Icelandic Narration in Vambarljóð”. RMN Newsletter 6: 20–25. Verse. Garland Studies in Medieval Literature 16. Kuhn, Hans. 1990–1993. “The Rímur Poet and His New York, London: Garland. Audience”. Saga Book 23: 454–468. Anton Karl Ingason. 2007. “Tvíliðahneigð”. Available Lamb, William. 2015. “Verbal Formulas in Gaelic at: http://www.linguist.is/skjol/tvilidahneigd.pdf Traditional Narrative: Some Aspects of Their Form (accessed 11.11.2015). and Function. In Registers of Communication. Ed. Conroy, Patricia. 1980. “Oral Composition in Faroese Asif Agha & Frog. Studia Fennica Linguistica 18. Ballads”. Jahrbuch für Volksliedforschung 25: 34–50. Helsinki: Finnish Literature Society. Pp. 225–246. Edwards, Mark W. 1980. “The Structure of Homeric Lord, Albert Bates. 1960. The Singer of Tales. Harvard Catalogues”. Transactions of the American Studies in Comparative Literature 24. Cambridge: Philological Association 110: 81–105. Harvard University Press. Foley, John Miles, & Peter Ramey. 2012. “Oral Theory Parry, Milman. 1928. L’épithète traditionnelle dans and Medieval Studies”. In Medieval Oral Homère. Paris: Société d’Éditions “Les Belles Literature. Ed. Karl Reichl. Berlin: de Gruyter. Pp. Lettres”. 71–102. Powell, Barry B. 1978. “Word Patterns in the Frog. 2011. “Alvíssmál and Orality I: Formula, Catalogue of Ships (B 494–709): A Structural Alliteration and Categories of Mythic Being”. Arkiv Analysis of Homeric Language”. Hermes 106 (2): för nordisk filologi 126: 17–72. 255–264. Frog. 2012. “On the Case of Vambarljóð II: Register Sigurður Kristófer Pétursson. 1924. Hrynjandi and Mode from Skaldic Verse to sagnakvæði”. íslenzkrar tungu. Reykjavík: Steindór Gunnarsson. RMN Newsletter 5: 49–61. Scholes, Robert, & Robert Kellogg. 1966. The Nature Frog. 2014. “Degrees of Well-Formedness: The of Narrative. London: Oxford University Press. Formula Principle in the Analysis of Oral-Poetic Smirnitskaya, Olga A. 1994 = Смирницкая, О.А. Meters”. RMN Newsletter 8: 68–70. 1994. Стих и язык древнегерманской поэзии. Frog. 2014–2015. “Mythological Names in dróttkvætt [‘The Verse and Language of Old Germanic Formulae I–III”. Studia Metrica et Poetica 1(1): Poetry’]. Москва: МГУ им. М.В. Ломоносова, 100–139; 1(2): 39–70; 2(1): 7–33. Филологический факультет, Филология. Frog & Eila Stepanova. 2011. “Alliteration in (Balto-) Stepanova, Eila. 2015a. “Registers of Seesjärvi Finnic Languages”. In Alliteration and Culture. Ed. Karelian Lamenters: Research on the Uses, Themes Jonathan Roper. Houndmills: Palgrave MacMillan. and Conceptions of Lamentation”. RMN Newsletter Pp. 195–218. 9: 137–143. Guðrún Kvaran. 1991. “Inngangur”. In Guðrún Kvaran Stepanova, Eila. 2015b. “The Register of Karelian & Sigurður Jónsson frá Arnarvatni. Nöfn Lamenters”. In Registers of Communication. Ed. Íslendinga. Reykjavík: Heimskringla, Asif Agha & Frog. Studia Fennica Linguistica 18. Háskólaforlag Máls og Menningar. Pp. 9–86. Helsinki: Finnish Literature Society. Pp. 258–274. Gurevich, Elena A. 1984 = Гуревич, Е.А. 1984 Sverdlov, Ilya V. 1998. = Свердлов, И.В. 1998 (manuscript). Древнеисландская поэтическая (manuscript). Парадигматика и синтагматика синонимика. Традиция и ее ученое осмысление в древнеисландских квидлингов как формы XII-XIII вв. [‘Systems of Old Norse Poetic речевой деятельности. [‘Paradigmatic and Synonyms: Tradition and Its Learned Reception in Syntagmatic Features of Old Norse kviðlingar as a 12–13th Centuries’]. Диссертация на соискание Type of Speech Activity’]. Дипломная работа. ученой степени кандидата филологических наук. Московский государственный университет им. Московский государственный университет им. М.В.Ломоносова. Филологический факультет. М.В. Ломоносова. Филологический факультет. Москва. Москва. Sverdlov, Ilya V. 2011. “Ok dulða ek þann inn Gurevich, Elena A. 1992a = Gurevič, Elena A. 1992a. alsvinna jǫtun: Some Linguistic and Metrical “Zur Genealogie der þula”. Alvíssmál 1: 65–98. Aspects of Óðinn’s Win over Vafþrúðnir”. Saga- Gurevich, Elena A. 1992b. “Þulur in Skáldskaparmál: Book 35: 39–72. An Attempt at Scaldic Lexicology”. Arkiv för Vésteinn Ólason. 1979. “Inngangur”. In Sagnadansar. Nordisk Filologi 107: 35–52. Ed. Vésteinn Ólason. Íslensk rit 5. Reykjavík: Gurevich, Elena A. 2012 = Гуревич, Е.А. “К вопросу Rannsóknastofnun í bókmenntafræði, Menningar- о значении наименования ‘тула’ в sjóður. Pp. 7–88. древнеисландском” [‘On the Meaning of the Vésteinn Ólason. 1982. The Traditional Ballads of Poetic Term ‘þula’ in Old Norse’]. Атлантика: Iceland: Historical Studies. Rit 22. Reykjavík: Записки по исторической поэтике 10: 183–199. Stofnun Árna Magnússonar. Haukur Þorgeirsson. 2010. Gullkársljóð og Yelena Sesselja Helgadóttir 2001 = Ершова, Елена. Hrafnagaldur: Framlag til sögu fornyrðislags. 2001. “Исландские поздние тулы – эволюция Gripla 21: 299–334. жанра” [‘Post-Medieval Icelandic þulur: Evolution Haukur Þorgeirsson. 2012. “Poetic Formulas in Late of the Genre’]. Атлантика: Записки по Medieval Icelandic Folk Poetry: The Case of исторической поэтике 5: 158–178. Vambarljóð”. RMN Newsletter 4: 181–196. 60

Yelena Sesselja Helgadóttir. 2011. “Icelandic Post- Yelena Sesselja Helgadóttir. 2015. “Þuluheimar á Medieval þulur”. RMN Newsletter 3: 91–94. tímamótum: Þornaldarþula og Grettisfærsla”. In Yelena Sesselja Helgadóttir. 2014. “Retrospective Gamanleikir Terentíusar: settir upp fyrir Terry Methods in Dating Post-Medieval Rigmarole- Gunnell sextugan 7. júlí 2015. [Ed. Rósa Verses from the North Atlantic”. In New Focus on Þorsteinsdóttir et al.] Reykjavík: Menningar- og Retrospective Methods: Resuming Methodological Minningarsjóður Mette Magnussen. Pp. 82–85. Discussions: Case Studies from Northern Europe. Ziogas, Ioannis. 2013. Ovid and Hesiod: The Ed. Eldar Heide & Karen Bek-Pedersen. FF Metamorphosis of the Catalogue of Women. Communications 307. Helsinki: Academia Cambridge / New York: Cambridge University Scientiarum Fennica. Pp. 98–119. Press.

Linguistic Multiforms in Kalevalaic Epic: Toward a Typology Frog, University of Helsinki

Abstract: This article examines the interaction of language in relation to metrical form in kalevalaic epic poetry. It presents and elaborates the concept of a linguistic multiform as a compositional framework and unit of expression that is distinguishable from an oral-poetic formula and operates according to different parameters of variation. Multiform theory provides a basis for addressing the conservatism and variation of kalevalaic epics. The present article discusses the interface Lord in The Singer of Tales (1960) has not between meter, language, and units of offered convincing accounts of verbally information mediated by language in conservative Finno-Karelian kalevalaic epic kalevalaic epic poetry. It explores how this tradition (Harvilahti 1992a). The synthesis of tripartite relationship combines with social a refined form of multiform theory with conventions to enable the remarkably current OFT offers a new way of looking at conservative reproduction of kalevalaic epics this poetry, a way which seems equally as socially stable verbal texts. Analysis builds relevant to similar short epic forms such as on the framework of Oral-Formulaic Theory1 North Russian bylina-epics and Old Norse (also called ‘Oral Theory’;2 OFT hereafter) eddic poetry. A crucial issue accounted for by augmented with a refined theory of linguistic multiform theory is what might be described ‘multiforms’. The concept of ‘multiform’ as as the verbal ‘chunking’ of units of text an analytical term was introduced by Lauri several lines in length that appear to operate and Anneli Honko (1995; 1998) to refer to a as a sort of ‘macro-formulae’. Once verbal integer and framework that operates at multiforms come into focus, it is possible to the level of verbal texture above the structural distinguish them according to formal types level of the formula, but which is not relevant to considering variation, as is done necessarily equivalent in scope to OFT’s unit here for multiforms in the kalevalaic epic called a ‘theme’. Multiform theory was tradition. The broad aim of this article is to originally developed within a strategic aim of develop a framework that helps better account accounting for the flexibility in the length of for stability and variation in verbally long epics, which had remained conservative poetries within the broader undertheorized in OFT research (cf. Zumthor approach of OFT. This framework is 1990: 92). The present study shows that developed on an empirical basis of corpus- multiform theory offers a valuable approach based analysis of kalevalaic epic poetry. to other poetic forms, and particularly to types This study of multiforms builds on of poetry that have remained peripheral to quantitative analysis, which has been greatly OFT research. facilitated by the digital edition of Suomen Focus here is on a short epic form in which Kansan Vanhat Runot [‘Old Songs of the linguistic multiforms are readily observable as Finnish People’] (www.skvr.fi; SKVR a social reality. Both OFT and the theory of hereafter), currently consisting of 89,247 multiforms were developed in relation to the items of poetry.3 Quantitatively oriented dynamic flexibility of the long epic form. studies have a long history in the research on ‘Classic’ OFT in the form outlined by Albert this poetry, which was initially studied within

61 the framework of the Historical-Geographic system of words and formulae that becomes Method.4 Quantitative analysis has also been established in the mind of an individual central to early research on the poetry’s performer and that provides a sort of metrics and poetics.5 However, the first collapsible or expandable latticework for quantitative analysis focused on stability and verbalizing a sequence of text (see also variation in the singing tradition was Lauri Honko 1995; 1996: 9–10; 1998: 100–116; Harvilahti’s Kertovan runon keinot: 2003: 113–122). The term had been used in a Inkeriläisen runoepiikan tuottamisesta variety of ways in OFT,8 but only [‘Devices of Narrative Poetry: On the contemporarily with the Honkos did it begin Production of Ingrian Epic Poetry’] (1992a), to be introduced as a term per se for ‘that which is focused on the region where which has multiple forms’ (Foley 1995: 2; cf. variation was the most pronounced and Foley 1990: 6) – which is basically the salient. Current emphasis in research on dictionary definition of its use as a noun (cf. kalevalaic poetry has been on how the poetry OED, s.v. ‘multiform’). The Honkos’ is used and adapted by individuals, its definition was narrower, referring specifically potential for meanings and social impact in to a formal type of verbal system for which no communication, as well as its flexibility in term had been established.9 Their definition relation to situational factors, interaction and might be compared to using ‘formula’ only to the processual nature of performance (e.g. refer to linguistic formulae in OFT (as Timonen 2004; Tarkka 2005; 2013; Kallio opposed to ‘story formula’ etc.). The Honkos’ 2013). The present study is centrally definition is also subtly refined by reference concerned with basic mechanics of the to that which is consistent as the framework production of kalevalaic epic poetry at the for what can be realized in multiple forms. It level of multiform units. Generalizations also designates specifically the linguistic about these multiforms may be genre- system operating at the level of verbal texture, dependent.6 The relationship of multiforms to stressing that this can potentially be structures in epic narration is introduced, but polysemic – i.e. a multiform is not bound to focus is on how multiforms operate as verbal specific contexts or meanings. The distinction systems. Like any study, certain features and of this type of phenomenon was quite phenomena are brought into focus as opposed innovative. to others, which in this case may produce an The identifyication of these variable verbal impression of a rather mechanistic model of frameworks through lexical and formulaic epic production (cf. Foley 1988: 8). However, recurrence in the Honkos’ pioneering work the methodological principle here is that we led multiforms to be defined in those terms. need first to develop an understanding of how This initial approach has since been multiforms work in order to look at how significantly refined through the investigation people were using them creatively. Otherwise, of the phenomenon of multiforms in a much we have no frame of reference for considering more diverse range of poetry in addition to nuances of specific variations or even the long epic form.10 Looking at the assessing which variations are meaningful.7 phenomenon across different types of poetry The study of the mechanics of multiforms is has made it possible to sort through tradition- here seen as a platform for considering dependent features while complementing the variations related to meanings, aesthetics, initial model with: equivalence classes, or interactional or processual aspects of interchangeable sets of verbal elements rather performance, or that may be connected to than attending to the most regularly recurring memory – as well as those that may just be an elements only; metrical and/or semantic organic outcome of how language operates in conditioning, which structure or determine the the tradition. form and/or meaning of elements in variation; and entanglement of syntax. On this basis, I What is a Linguistic Multiform? define a linguistic multiform as: The Honkos (1995; 1998) advanced a collocative system made up of specific ‘multiform’ as a formalized term for a verbal verbal units (lexemes, formulae) and/or sets

of these within function-specific (semantically, art – not words in our orthographic sense, but syntactically and/or metrically conditioned) ‘words’ in the oral sense of verbal units of equivalence classes that together form a utterance carrying a distinct unit of coherent indexical system in the memory of meaning.13 Formulaic ‘words’ can be an individual. conceived as each having a unique entry in Key terms and concepts in OFT are ‘formula’ the mental lexicon (Wray 2002; 2008). and ‘theme’ as well as ‘narrative pattern’ or Classic OFT qualified formulae according what Lord referred to as a ‘song’ (the to use in different contexts. This part of the framework on which whole song-types are definition linked to the statistical methods for entextualized). Although multiform theory identifying formulae on the one hand and to a was advanced as an alternative to OFT (see strategic orientation to show that texts were also Hakamies 2013), ‘multiform’ designates not ‘memorized’ and instead relied on a a phenomenon complementary to OFT’s generative use of traditional phraseology. ‘formula’ and ‘theme’ with the potential to However, frequency is an unreliable criterion help mediate the relation between them. for formula identification precisely because These concepts will be briefly introduced here some formulaic sequences are only used in in order to discuss multiform in relationship very specific contexts (Wray 2002: 25–31). to them. Kalevalaic epic poetry maintains a large body Classic OFT defined ‘formula’ narrowly as of expressions that were used in quite specific a (phrasal) linguistic unit recurring in the contexts.14 The corpus of ca. 150,000 items of same metrical positions (Parry 1928: 16; Lord kalevalaic poetry in a variety of genres 1960: 4). Incorporating the relation to meter nevertheless makes it clear that many of these into this definition was circularly derived phrasal units could move between contexts from the quantitative method of identifying among different groups of singers or in formulae according to the metrical positions different regions; even those that appear the in which they would appear. This was found most contextually bound exhibit at least a few not only to be too narrow for the oral poetries variations in use if they are attested in more Classic OFT initially sought to describe (e.g. than 50 or 100 examples (cf. Frog 2010a: Hainsworth 1968); it simply breaks down 365–376, 400–405; also Lord 1995: 62). when considering forms of verbal art that lack These units will be treated here as formulae or fixed meter (cf. Lamb 2015; Stepanova 2015; emic ‘words’ and considered to have unique Yelena Sessilja Helgadóttir, this volume). A entries in the mental lexicon of those fluent in formula is now more flexibly viewed as a the register. linguistic unit of the register11 of verbal art A multiform is formally distinguished from (i.e. language as used in the poetry) that a formula as a system of emic ‘words’ that operates as “a morpheme-equivalent unit” provide a framework for entextualizing (Wray 2008) or “an integer of traditional expressions of verbal art. The Honkos meaning” (Foley & Ramey 2012: 80). A developed their approach with the explicit formula is customarily distinguished from aim of accounting for flexibility in length in other lexemes by bundling elements or the long epic form. They focused on quite features, whether it is a lexical string, a loose and variable verbal frameworks lexeme and certain metrical positions, but also identified through lexical and formulaic potentially an intonational structure, a recurrence. The outcome was that their grammatical construction, gesture, etc., that description is quite formally dependent on make its regular use distinctive. It is thus oral poetry of a certain type and multiforms possible for a single orthographic word to be are described almost exclusively in terms of the only lexical part of a formula.12 Following lexical stability. They also advanced Foley (e.g. 1995), integer will be used here as multiforms as a phenomenon at the level of a term to address elements of a register that the individual performer’s idiolect only. operate as unitary signs. John Miles Foley Anna-Leena Siikala (1984: 85–93) had (1995; 1999; 2002) has emphasized viewing identified the process of ‘crystallization’ as these as emic ‘words’ of the register of verbal the formation of verbal systems in the

63 memory of narrators for representing remain formally distinguishable: formulae are especially semantically central units of minimal integers or ‘words’ within the narrative content (see also Kaivola-Bregenhøj registral lexicon, whereas multiforms are 1988: 305–313). The Honkos’ ‘multiform’ is comprised of associative systems of such essentially a designation for that system. By ‘words’. This formal difference means that turning attention from the realization of macro-formula multiforms are subject to narrative content through language to the different types of variation, such as in the multiform as a verbal system, the verbal organization or syntactic relations of their system could then be observed as it is used in constituent ‘words’, or in which ‘words’ do or relation to other content. In kalevalaic epic, do not appear, as will be illustrated below. these systems of emic ‘words’ are much ‘Theme’ developed from what is also denser: rather than a loose verbal latticework called a ‘typical scene’ or ‘type scene’. The that is filled out in performance, these term describes a traditional ‘group of ideas’ systems often manifest as much more solid which makes up a conventionalized unit of chunks of text. They were recognized in epic narration that is realized through the earlier scholarship as, for example, säejaksot idiom of traditional formulae (see e.g. Arend [‘line-series’].15 Corresponding units were 1933; Lord 1960: 68–98; Frye 1968; Foley described already by A.F. Gil’ferding (1894: 1990: 240–245, 279–284, 329–335). Like 24) in the more flexible North Russian bylina- Siikala’s work on crystallization, OFT’s epic poetry, where he observed that these research on themes took units of content as its crystallized units were repeated ‘word-for- basis for looking for recurrent patterns in word’ the same, while the connecting tissue verbalization. The term multiform designates between them would vary as would their a system of verbal material that may be organization within an epic. Such units were associated with a theme, but which may also also recognized by Lord even in South Slavic be used in other contexts. Like the Honkos’ epic, described as ‘runs’ (1960: 58–60) or pioneering work on multiforms, OFT’s work ‘blocks of lines’ (1995: 22–62), but his on themes has tended to focus on lexical discussion remained limited to showing that recurrence only. Like the term ‘motif’ in the the texts were not ‘memorized’ rather than Classic Historical-Geographic Method, OFT’s attending to how these units operate per se term ‘theme’ has generally been used (1981; 1995). Although multiforms, like practically, without formal definition as a unit formulae, crystallize in the minds of in relation to other types of formally defined performers, they can be observed in the units.17 I take a more structured approach that corpus as socially transmitted. It is precisely distinguishes minimal symbolic integers of this that is the phenomenon under scrutiny expression in terms of images (corresponding here: multiforms that circulate socially.16 to nominal units) and motifs (constructions Certain types of multiform units in entailing the equivalent of a verb and kalevalaic epic have a conventional use for involving one or more images). Themes are rendering specific units of narration. There here seen as integers at the next level of also seems to have been some emic complexity as a constellation or series of perception of sequences of text produced images and motifs; narrative patterns and through such multiforms as units, which plots are at still higher formal levels of appears to be what the famous Ižorian singer complexity. (See Frog 2015a: 38–40.) From Larin Paraske referred referred to as this view, a theme is a formal counterpart to a pätkäniekkoja [‘chunks’] (Timonen 2000: multiform; it simply operates among 642, 652). The regularity of use leads these linguistically mediated signs at the level of multiforms to function as complex signifiers complexity just above minimal integers of the mediating a coherent meaningful narrative symbolic lexicon. In the illustrative example integer. As such, they operate like formulae – of a kalevalaic epic below, for instance, the i.e. morpheme-equivalent units – which can victim’s offer of ransom for his release can be be described as macro-formulae. Nonetheless, seen as a motif. Each cycle of dialogue can be formulae and macro-formula multiforms seen as a theme, being repeated in series with

64 variation in the outcome of whether the rather than documenting multiple variants proposed ransom is refused or accepted.18 In from a single performer. This has resulted in a kalevalaic epic, multiforms tend to crystallize corpus that is well equipped to address social around images and motifs, although even one patterns in the tradition but not for individual of these may be elaborated through more than idiolects and repertoires. At the same time, one multiform. However, the categories of the corpus presents many examples of image, motif and theme are not essential to multiple texts of different genres recorded the following discussion and are presented from individual performers. It also presents here as a frame of reference for relating many examples of two or more variants of a multiforms to other traditions. particular epic recorded from individual Primary focus in this article is on performers, often by different collectors distinctions between multiforms of different across a period of years. This corpus provides formal types according to their constitution by empirical data for considering continuity and language and how they vary in social variation in multiforms as a socially transmission (as opposed to in use by transmitted phenomenon. individual performers). This formal ontology The datasets for multiforms discussed is distinct from whether a multiform operates below are developed from variants of the as a macro-formula, which is defined epics published in SKVR. This data excludes according to the relationship between a variants of the same multiforms that may be multiform and what it is customarily used to found outside of materials indexed with the signify, communicate or ‘do’, irrespective of epic texts, for example in incantations. how it is constituted and varies at the formal, Additional unpublished items in the Finnish verbal level. Literature Society’s Archive have also been consulted for The Song of Lemminkäinen, but Kalevalaic Epic as Material for Analysis the methods of analysis used here relied on 19 Kalevalaic epic was a short epic form (the digitized text, which became a technical songs sometimes called ‘lays’ rather than constraint when forming the datasets. Not ‘epics’). Epics circulated as socially every multiform is attested in every variant recognizable verbal texts or songs that varied and fragment of the relevant epic. Multiforms in narrative complexity but epic performances for the same or equivalent unit of narration would generally remain about 75–300 lines in could also alternate within and across length. The tradition and its basic idiom were different regions. The number of examples of trans-lingual, shared across dialects of each multiform analyzed therefore vary Finnish, Karelian and Ižorian. The epic considerably even though they may be tradition disappeared from the majority of associated with the same epic. Lutheran Finland before being documented to Variations in the singing tradition that are any degree. The traditions survived longer in established at a social level are approached Orthodox areas on the Russian side of the here in terms of dialects. Such variations that Finnish border in territories extending from occur at the verbal level of performance, at the White Sea to the Gulf of Finland. the level of language, rhythms and melodies, Individual epics were documented in are here described as dialects of singing (Frog anywhere from a few examples and fragments 2010b: 99–100). It is fairly common to to several hundred. The vast majority of this collapse narrative content with verbalization. th material was hand-transcribed in the 19 and However, mythology is here approached as a th 20 centuries, sometimes only in minimal distinct symbolic system that may interface notes and in shorthand, normally with little or with language and vary with it but that may no information about the collection context, also vary independently of it (see further Frog informant, or even whether the epic was 2015a). For example, Elias Lönnrot’s epic dictated or sung. Collection paradigms of the Kalevala (1835; fundamentally reorganized time valorized collecting variants from as and significantly expanded 1849), built from many informants as possible and especially collected oral poetry, impacted many oral epic variations that had not yet been documented singers, but in some cases this impact might

65 only shape epic plots and connections women’s tradition (Virtanen 1987: 18; Siikala between them without clear impact on the 1990; 2002b; Harvilahti 1992a: 14). Although verbal level of the dialect of singing. The individual singers could engage with local mythological content of songs is thus here conventions differently, potentially handling described in terms of dialects of mythology the idiom and mythological subjects with (Siikala 2012: 15). Although dialects of great freedom (Harvilahti 1992b), the epic singing and dialects of mythology are inclined tradition was generally characterized by an to interface in this tradition, there are ‘inclination to non-variation’ (Frog 2011b: nevertheless cases in which it is important to 53–55; cf. 2011a: 48–50) – i.e. with an ideal be able to distinguish between them. of ‘word-for-word’ reproduction (see also Individual examples of epics or fragments Kiparsky 1976: 96–97; Timonen 2000: 642, and summaries of them will be referred to as 653), although it nevertheless evolved in variants. Groups of variants of a particular flexibility in southern areas (cf. Harvilahti epic more similar to one another than to other 1992a). groups of variants will be referred to as redactions. Identifying a group of variants as The Singing Competition: An Example of a belonging to redactions does not presume that Kalevalaic Epic each group of variants can be collectively Whereas the Honkos analyzed multiforms for reconstructed to a common, historical their role in the flexibility of epic length, redaction-archetype (see also Siikala 2012: multiforms offer a corresponding perspective 15). Generally speaking, dialects of kalevalaic on textual stability in kalevalaic epic mythology are characterized by the redactions transmission. The ideals of this tradition’s of mythological epic material that they entail. conservatism are illustrated here through an Differences between dialects of mythology example of an epic from which multiforms manifest verbally as potential indicators of will be discussed, The Singing Competition. difference between dialects of singing, but This epic is here represented in five variants neither of these will necessarily correspond in performed by three singers representing two distribution with dialects of spoken language. generations of the Malinen family from Dialects or sub-dialects of singing and Vuonninen, Viena Karelia. The epic has been mythology may also occur between much chosen because of its simple plot, which leads smaller groups than spoken language dialects. to shorter variants. The example also has the In kalevalaic poetry, the conservatism of the added interest that the material recorded from verse form allows otherwise quite subtle Ontrei Malinen by Elias Lönnrot was among lexical variation in formulaic lines and their the most significant in the development of his organization in multiforms to become Kalevala. perceivable as linked with differences The conservatism of the poetry is between one way of singing and others (e.g. highlighted by violating the custom of Frog 2010b: 99–100; 2011b: 53–55). providing equal representation to each variant Conventions of conservatism and variation of an oral poem. Instead, all five variants are differed on a regional basis. The tradition was laid out in a manner customary for variant 20 most conservative in the north, in the region manuscripts. In (1), the earliest variant is of Viena or White Sea Karelia with increasing presented as the base text and differences flexibility to the south. This pattern followed between that exemplar and other variants are the axis of the progression into Finnic areas of presented in the form of notes. Of course, this Christian religious authority and secular sort of representation subordinates and effaces social impacts that displaced epic from a all variants to the base text, obfuscating their central role in the ritual life of the community. individual characteristics by reducing them to These same processes had the result that fragmentary notation so that they cannot be women were increasingly significant as ‘read’ without an archaeological excavation bearers of the tradition to the south; on the and reconstitution from the notes. The point Karelian Isthmus and in Ingria, the singing of here is not to imply a ‘best text’ or anything 21 kalevalaic poetry had shifted to become a about ‘reconstruction’. The purpose is a) to

66 make available an example of a kalevalaic indicating the end or beginning of a sequence epic, b) to make the multiforms in the text of text or narration). visible, and c) to enable some perspective of As a sigla for each variant, performers’ variation in multiforms within a local dialect initials are used so that the relations are more of singing, which can both be viewed in salient in the apparatus: dialogue with the variety of examples below OM1 – The base text, recorded in 1825 by as well as used to gain perspective on A.J. Sjögren from Ontrei Malinen (SKVR variation in the epic whole. (Interested readers I1 185). can access the original texts online at www.skvr.fi, where they can be found by OM2 – Recorded in 1833 by Elias Lönnrot from Ontrei Malinen (SKVR I1 185a). selecting part I1 under ‘osa’, and items 185, 185a, 186, 186a and 187 under ‘numero’.) JO1 – Recorded in 1871 by Axel Borenius The fact that closely related variants can be from Jyrki Ontreinen, son of Ontrei reasonably laid out in this way is a testament (SKVR I1 186). to conservatism in the tradition: variation is at JO2 – Recorded in 1877 by Axel Borenius the level of the inclusion and omission of from Jyrki Ontreinen (SKVR I1 186a). lines, alternative lines or variant wording, the This variant is attested through Borenius’ order of lines within a multiform, and the addition of lines and notations to the multiforms labelled M5 and M8 are absent 1871 text (his normal practice; he is from some variants. This layout is possible known for his care and precision in collection); lines are assumed to be the owing to multiform by multiform same where notations are lacking. correspondence, the regularity of those multiforms and theit order in preformance. VO – Recorded in 1872 by Axel Borenius The practicality of this layout rapidly breaks from Vassilei Ontreinen, son of Ontrei down if it is done with variants from different and elder brother of Jyrki (SKVR I1 187). dialects, as the variation between multiforms In order to understand these texts, it warrants increases, alternate multiforms may appear, pointing out that the tradition of singing and the organization of multiforms may changed rapidly across the period of differ. collection in relation to various changes in In order to make the interplay of continuity society. Borenius commented that Vassilei, and variation accessible, parallel translation of then an old man, had not sung these songs the master text is accompanied by offering since his youth; when Vassilei was finally translations of variant words and lines in the convinced to try to sing, Borenius felt that he apparatus. This is not a critical text. had once known a lot but his memory was Abbreviations in the original transcripts are already fading (Niemi 1905: 483). This seems not indicated and punctuation has been to be in the background of Vassilei’s use of removed. Minor variations in orthography the name of a different hero (i.e. (e.g. ‘ü’ for ‘y’) are not especially indicated Lemminkäinen for Joukahainen) throughout, but the original orthography of variants is as well as related to some other more subtle otherwise retained in the apparatus. Use of the variations in his rendition (e.g. in line 57; see verb form on [‘is’] as an expletive particle is Frog 2010b: 99–100; 2011b: 54). Jyrki indicated in the translation by placing it in similarly began the epic in his 1871 rendition parentheses. Where no translation with reference to characters from a different accompanies a variant in the apparatus, the (non-epic) song. Already in Ontrei’s 1825 variation can be considered non-semantic. rendition, a similar issue of memory may be Although the epic’s entextualized whole is reflected in the presentation of line 9 between more than 100 lines, it is constituted of only lines 5 and 6, although this could equally be a 16 macro-formula multiforms (most of which consequence of the slow and interruptive group in structural pairings), and with a few dictation process. These types of variations quotatives and boundary markers (a quotative present methodological issues for analyzing is an introduction to direct speech; a boundary multiforms, their variation and association marker is a line, couplet or multiform

(1) The Singing Competition within the Malinen family. Key: Spaces are placed between multiforms (not counting quotative frames). Line numbers are marked along the left (lines deviating from the lineation of OM1 are marked with asterisks); scenes, themes, macro-formula multiforms (M), quotative frames (Q) and boundary markers (BM) are marked in the right column along with their constituent elements as relevant. Punctuation has been removed. In the notes, ‘½’ is used to number additional lines not in the main text (following the practice of Borenius and others with these poems).

Yks’ on nuori Joukahainen The one (is) young Joukahainen M1 – Scene 1 Toinen vanha Väinämöinen The other old Väinämöinen Theme 1 Ajoi tiellä vastaksuten Drove against one another on the road

Rahis puuttu rahkehesen Trace jammed to trace M2a 5 Vemmel vempelen nenäh Shaft-bow to shaft-bow’s end Vesa kasvo vempelestä A sapling grew from the shaft-bow Haavat aisoista yleni Aspens from the shafts rose M2b Paju pehko rahkehista A willow-clump from the trace *Siitä siinä seisottih” *Then they were standing there.

10 Sanoi nuori Joukahainen Said young Joukahainen Q Theme 2 *Ken on tiiolta parempi* Whose knowledge is better M3 *Sen on tiellä seisominen* He on the road will stand, Ken on tiiolta pahempi Whose knowledge is worse Sen on tieltä siirtyminen He from the road will move.

15 Sano vanha Väinämöinen Said old Väinämöinen: Q Sanos tarkkoja tosia Say you exact truths M4 Valehia muinosia ancient lies.

Sano nuori Joukahainen Said young Joukahainen: Q Theme 3 Tiiän kuitengin vähäsen I know however a little M5a 20 Ennemmäiset ymmärtelen A few more things understand

M1: 1–2. Poika tuhman Tuurituisen / Tütär Keiretüisen / Ne on teitäñsä käveli / Matkojañsa meäreäil’övi [‘Lad of stupid Tuuriruinen / Daughter of Keiretyinen / They (is) go their way / Measure their journey’] JO1; Tuo vanha Väinämöini / Tai on lieto Lemminkäini [‘That old Väinämöinen / Or (is) loose Lemminkäinen’] (tai on corrected from tahi [‘or’],+ ja, sekä [‘and – and’]) VO 1. Joukahainen] Joukavainen OM1 1½. Se on teitänsä käveli [‘He (is) goes his way’] OM2 2½. Matkoaan mittelöövi [‘Measured his journey’] OM2 3. vastaksuten] vassakkahan OM2 JO1 JO2; Sopi ne tiellä vassaksutta [‘They agreed against one another on the road’] VO M2: 3½. Aisa puuttu aisam peähä [‘shaft jammed to shaft’s end’] JO1 JO2 4. first wrote “rahkesesen” OM1 5. Aisa aisahan takeltu [‘shaft to shaft got stuck’] VO 6–8. absent VO 7. Haavat] Koivut [‘birches’] JO1 JO2 Q: 10. Joukahainen] Joukavainen OM1; nuori Joukahainen] lieto Lemminkäini [‘loose Lemminkäinen’] VO M3: 10½. Oi on vanha Väinämöini [‘Oh (is) old Väinämäinen’] JO2 11–12. absent OM1 11. Ken] Kump’ [‘which(ever)’] JO1 JO2 VO 13–14. absent OM2 VO 13. Ken] Kump’ [‘which(ever)’] JO1 JO2 14. siirtyminen] seisomini [‘standing’] OM1 JO1 JO2 M4: 15½ Oi on nuori Joukahainen OM2 / Kun ollet tiiolta parempi OM2 JO1 JO2 VO / Sano korvin kuullakseni OM2 [‘Oh (is) young Joukahainen / If you are better in knowledge / Say to be heard by me with the ear’] i 17. absent VO; muinosia] viime sie [‘last’] JO1 JO2 Q: 18. nuori Joukahainen] lieto Lemminkäini [‘loose Lemminkäinen’] VO M5a–b: absent JO1 JO2 VO

Tiiän linnukset tiasen I know the tit is a bird M5b Kiiskisen veen kalaksi, The ruff is a fish of the water Pajun puita vanhimmaksi The willow is oldest of trees Tiiän kolkot kuokituksi I know dark places were scooped M6b 25 Vuoret luovuksi kokohon Mountains heaped together Kala hauat kaivetuksi Fish-hollows became dug Siverret syvennetyksi Troughs became deepened.

Sanoi siitä Väinämöinen Said of this Väinämöinen: Q Lapsen on mieli, vaimon tunti Child’s (is) mind, woman’s lore M7a 30 Ei oo partasuun urohon Not for a bearded hero Omat on kolkot kuokkimani Mine (is) the scooping of dark places M6b Vuoret luomani kokoh Mountains together my creation Kala hauat kaivamani Fish-hollows the digging

Olin miekin miessä siellä Indeed I as a man was there M8 35 Urohona kolmantena As a third hero Seitschemäntenä urossa As a seventh hero Kaarta taivon kantaissa Bearing the arch of heaven Pieltä ilmon pistäissä Sticking the post of the sky Taivoista tähittäissä Starring the heavens 40 Otavaa ojentamassa Straightening the Ursa Major

Siitä suuttu Väinämöinen Then grew angry Väinämöinen BM – Scene 2 Päälle nuoren Joukahaisen Toward young Joukahainen Theme 4

Laulo nuoren Joukahaisen [He] sang young Joukahainen M9a Laulo suohon suoni vöistä Sang into a swamp to his belt 45 Niittyhyn nisu lihoista Into a meadow to his waist Kainalosta kangahasen Into a heath up to his armpit Laulo koiran Joukahaisen Sang Joukahainen’s dog’s M9b Kynsin kylmäh kivehen Claws into a cold stone Hampahin vesi hakoh Teeth into a water-log

22–23. ordered 23 22 OM2 23½. Tiiän kuitenki vähäsen [‘I know however a little’] OM2 M6b: Identified as M6b without M6a on the basis of common use in an additional cycle of dialogue. 24–27. ordered 24 26 27 25 JO1 JO2; ordered 24 26 25 27 VO 25½. Mäet mulleroitetuksi [‘hills raised’] OM2 Q: 28. siitä] vanha [‘old’] JO1 JO2; Sanoi siitä] Sitt’ on vanha [‘Then (is) old’] VO M7: 28½. Oi nuori Joukahainen [‘Oh young Joukahainen’] OM2; Kusen konna kulkkuhusi [‘A can of piss in your throat’] JO1 JO2 29–30. absent JO1 JO2 29. on] absent OM2 31–33. ordered 31 33 32 33½ JO1 JO2; ordered 31 33½ 32 (33 absent) VO 31. on] absent OM2 32½. Mäet mulleroittamani [‘hills raised’] OM2 33. “etc.” OM2 33½. Süveret süventämäni [‘the depths my deepening’] JO1 JO2 VO M8: absent JO1 JO2 VO 34. Olin miessä 6:tena [‘I was the sixth man’] OM2 37. Kaarta taivon] Taivan kaarta [‘Heaven’s arch’] OM2 38. ilmon pistäissä] ilman pistämässä OM2 39. tähittäissä] tähyttämässä OM2 BM: 41. Siitä suuttu] Siitä vanha [‘Then old’] JO1 JO2; Sittäpä vanha [‘Then indeed old’] VO 42. absent OM2 JO1 JO2 VO M9: 43–54. ordered 50 51 52 53 47 44 46 45 (lacks 43, 48–49, 54) JO1 JO2; ordered 43 50 50½ 47 48 43 45 46 (lacks 44 49 51–54) VO 43. absent JO1 JO2; Joukahaisen] Joukaisen OM1; nuoren Joukahaisen] lieto Lemminkäisen [‘loose Lemminkäinen’] VO 44. Laulo] absent OM1

50 Laulo jousen Joukahaisen Sang Joukahainen’s bow M9c Kaariksi vesien päälle Into a (rain)bow on the water Laulo nuolen Joukahaisen Sang Joukahainen’s arrow M9d Haukaksi kiitäväksi Into a streaking hawk Ylähäksi taivosella High in the heavens.

55 Se siitä hyvin häpesi He then felt very ashamed. BM

Sanoi siitä Joukahainen Said then Joukahainen Q Theme 5I Myössytäs pyhät sanasi Make your holy words harmless M10aI Perävytä lauhiesi Turn back your sentences Annan kultia kyperän I will give a helmet(ful) of gold M10bI 60 Oman pääni päästimeksi For liberating my own head Itscheni lunastimeksi As a ransom for myself

Sanoi vanha Väinämöinen Said old Väinämöinen Q En huoli hopeistasi I don’t care about your silver M11aI On kultia itschellänikin Have indeed myself gold M11bI 65 Kahta kolmea paremmat Twice, three times better

Sanoi siitä Joukahainen Said of this Joukahainen Q Theme 5II Myössytäs pyhät sanasi Make your holy words harmless M10aII Perävytä lauhiesi Turn back your sentences Annan ainuan orihin I will give my only stallion M10bII 70 Oman pääni päästimeksi For liberating my own head Itscheni lunastimeksi As a ransom for myself

En huoli heposistas I don’t care about your horses M11aII Ilkiä orihistas Mean one, your stallions

50. jousen] jouset (plural) JO1 JO2 50½. Harakakse lentämähe [‘into a magpie flying’] VO 51. kaariksi] koiriksi [‘into dogs’] JO1 52. Joukahainen] Joukavaisen OM1 53. Haukaksi] Havu-puiksi [‘into wood of pine twigs’] JO1; Havukoikse [‘into hawks’ or ‘into a group of pine twigs’] JO2 54. absent OM2 BM: 54½. Se nuori Joukahainen [‘that young Joukahainen’] OM2 55. häpesi] pahastu [‘felt bad’] OM2 Q: 56. absent OM2 VO; first wrote Joukavaisen OM1; siitä] nuori [‘young’] JO1 JO2; Itkije t’ihustelovi [‘the weeper snivelled’] VO M10I: 56½. Oi on vanha Väinämöini [‘Oh (is) old Väinämöinen’] JO1 JO2; Ohoh vanha Väinämöini [‘Oh old Väinämöinen’] VO 57. Myössytäs] Myssytä OM2 JO1 JO2 Püörrütä [‘reel back’] VO 58. Perävytä] Peräytä OM2; Peru₍uta laupiesi JO1 JO2; Peruhuta lausiesi VO 59½. Hopehia huovan täyen OM2; Hopeita huovan täüven [‘a (?) full of silver’] JO1 JO2 VO 60–61. absent VO M11I: 62½. Kusen konna kultihisi JO1 JO2 VO [‘a can of piss on your gold’] 63. absent JO1 JO2; Ilkije hope₍ihisi [‘wicked one, on your silver’] (parallel line to 62½) VO 64–65. absent VO 64. itschellänikin] itelläini OM2 65. absent JO1 JO2 Q: 66. absent VO; siitä] nuori [‘young’] OM2 JO1 JO2 M10II: 66½. Oi on vanha Väinämöini [‘Oh (is) old Väinämöinen’] OM2 JO1 JO2 67–68. “et[c.]” OM2 (couplet absent) 68. Peru₍uta laupiesi JO1 JO2; Peruhuta lausiesi VO 70–71. absent VO Q: 71½. Sano vanha Väinämöinen [‘Said old Väinämöinen’] OM2 JO1 JO2 VO M11II: 71¾. Kusen konna oriihisi [‘A can of piss on your stallion’] JO1 JO2 VO 72–73. absent JO1 JO2 72. absent VO 73. orihistas] orisestasi OM2 hevosillasi [‘on your horses’] (parallel line to 71¾) VO

On orih itschelläniki Have indeed myself a stallion M11bII 75 Kahta on kolmea parempi Twice (is), three times better Viittä, kuutta virkeämpi Five, six times more brisk Kuutta, seitentä somempi Six, seven times nicer

Sano nuori Joukahaine Said young Joukahainen Q Theme 5III Myössytäs pyhät sanasi Make your holy words harmless M10aIII 80 Perävytä lauhiesi Turn back your sentences Annan ainoan sisäreni I will give my only sister M10bIII Oman pääni päästimeksi For liberating my own head Itscheni lunastimeksi As a ransom for myself

Siitä vanha Väinämöinen Then old Väinämöinen M10aIV – BF 85 Mössytti pyhät sanansa Made his holy words harmless Peräytti lauhiesa Turned back his sentences

Siitä pääsi Joukahaine Then got out Joukahainen BM – Scene 3 Mäni itkien kotih Went home weeping M12 Theme 6 Kallotellen kartanolle Wailing to the farm

90 Isä pääty ikkunassa Father got to the window M13 Emo aittojen vajolle Mother toward the storehouse

Emo ennätti kysyä Mother hastened to ask Q Mitäs itet poikuoni Why do you weep my lad M14 Ei ole syytä itkiäsi There is no reason for your weeping

95 Sano nuori Joukahainen Said young Joukahainen Q Ompa syytä itkeäni Is indeed reason for my weeping M15a Vaivoa valittavani Trouble to complain

74. on hepo itelläini [‘have myself a horse’] OM2 75. on] absent OM2 JO1 JO2 76–77. absent VO 76. kuutta] absent OM2 77. absent OM2 Q: 78. absent VO M10III: 78½. Oi on vanha Väinämöini [‘Oh (is) old Väinämöinen’] JO1 JO2; Ohoh vanha Väinämöini [‘Oh old Väinämöinen’] VO 79–80. absent OM2 80. Peru₍uta laupiesi JO1 JO2; Peruhuta lausiesi VO 81. sisäreni] sisaren (lacks possessive affix ‘-ni’) OM2 82–83. absent OM2 VO M10IV: 84. absent VO 86. lauhiesa] lauhiensa (underlined) OM2; Peru₍utti laupiensa JO1 JO2; Peruhutti lausiesa VO 22 BM: 87. pääsi] nuori [‘young’] OM2 JO1 JO2; Sitt’ on lieto Lemminkäini [‘Then is loose Lemminkäinen’] VO M13; 90–91. absent JO1 JO2 Q: absent VO M14: 92½. Ohoh silma poikuoni [‘oh my lad/son’] VO 93. Mitäs] Mitä OM2 JO1 JO2; Mit’ on süütä itkiesi [‘what (is) the reason for your weeping’] VO 93½. Nuorra saatuni valitat [‘my one begotten when young, complain’] OM2 94. absent VO; Ei ole] Onko [‘is there?’] JO1 JO2 94½. Vaivoja valittoasi [‘difficulties for your complaining’] OM2 JO1 JO2 Q: 95. absent VO M15–M16: 98–105. absent OM2; OM2 concludes the poem with the variation concludes the poem with the variation in line 97½ M15: 96–97. absent VO 97. valittavani] valittoani [‘my complaining’] JO1 JO2

Annon ainoan sisären I gave my only sister M15b=10b Oman pääni päästimeksi For liberating my own head 100 Itscheni lunastimeksi As a ransom for myself

Emo varsin vastaili Mother certainly answered Q Tuota toivon tuon ikäni For this I hoped for that my life M16 Vävyxeni Väinämöistä For Väinämöinen for my son-in-law Suukseni miestä suurta A great man for my kin 105 Laulajata langokseni A singer for my brother-in-law

98½. Lainasin emoni lapsen [‘loaned my mother’s child’] OM2 (concludes poem); Vanhalla Väinämöisellä [‘to old Väinämöinen’] JO2 VO Q: 101. vastoapi [‘answers’] JO1 JO2 M16: 101½. Elä itke poikuoni [‘don’t cry my lad/son’] JO2 102–105. ordered 102 103 105 104 JO1; ordered 102 102½ 104 103 105 JO2; ordered 102 104 105 103 VO

100½ Halki polveni halasin [‘across my generation wanted’] JO2

with different dialects. These issues are taken in English, in which longer words tend to into consideration in the discussion below. appear at the end of the line, and lines never 24 Owing to limitations of space, however, end in a monosyllable discussion of variations linked to memory and  Verses were stichic, avoiding enjambment relevant indicators of such variation are For most of the language areas, a line would 22 reserved for a later article. be 2–4 words, lexical choice would be stabilised by alliteration, and the position of Kalevala-Meter elements in a line would be stabilized by Understanding the crystallization of where long and short syllables can appear in multiforms and their different types of the meter. Finnic languages are heavily variation is directly related to their metrical inflected, normally using case endings where form. The so-called Kalevala-meter is named one would expect prepositions in an Indo- for Lönnrot’s Kalevala. It is the North Finnic European language. Metrical conventions that form of the common Finnic alliterative inhibit the movement of words within a line tetrameter, whereas the forms in Estonia are also operate as constraints on morphological often called regilaul (see also Sarv 2015: 6– variation – i.e. morphological variation will 7). The meter was used across a remarkable not change the number of syllabic positions number of genres, which allowed the different filled by a word. These conditions made it genres to be interpenetrating and easily quite easy for whole lines to crystallize as transposed (see Tarkka 2013). Kalevala-meter formulaic units. is characterized by the following essential 23 Although also not required of every line, features: parallelism was fundamental to this poetry  An essentially trochaic rhythm, yielding an (see Steinitz 1934; Saarinen 2014). Use of eight-syllable line with alternating strong parallelism varied by genre. In narrative and weak positions, although the number of poetry, semantic parallelism and its looser syllables in the first foot was flexible form of analogical parallelism were generally  Placement of long lexically stressed characteristic of semantically central elements (=initial) syllables in metrically stressed of narration. This often took the form of a positions and short lexically stressed crystallized couplet, although an extended syllables in metrically unstressed positions, series of parallel lines was also possible. In not manifested in every line epic poems, parallel lines can be generally  Alliteration, although not required in every described as crystallized and socially line  Parallelism, not required of every line circulating formulae rather than uniquely  A general inclination to a form of ‘right produced in the performance situation. justification’ (Foley 1993 [1990]: 96–106, Couplets were also formed of complementary 178–196), often referred to as ‘winnowing’ semantic units that could be organized in a

72 series of structurally and analogically parallel Within the present framework, other types of couplets. phenomena grouped as ‘systems’ by Parry are It is customary to conceive of poetic meter considered in terms of classes of emic in terms of an absolute paradigm. However, ‘words’ that group according to semantic, the Kalevala-meter is better conceived of in grammatical, functional and/or metrical terms of conventions with varying flexibility; equivalence. Parry sees these classes not verses were perceived in terms of degrees of simply as formally and/or semantically well-formedness rather than according to a equivalent but as forming a group of potential simple binary of ‘metrical’ versus ‘non- alternatives within the mind of a performer metrical’ (see also Frog 2014a). Most regular (cf. Parry 1971 [1930]: 274 [85]). Such an was the basically syllabic structure of the line, indexically associated group established in the bound up with the rhythm of performance, mind of a performer is here considered an followed by the placement of stressed equivalence set of emic ‘words’ which will syllables. Alliteration could be compensated not necessarily include all members of the for by other types of phonic patterning within etically described equivalence class in the and across lines in the continuous flow of register. A formula-system is here defined as a performed discourse. Various devices of constellation of complementary equivalence parallelism and lexical repetition equally sets that seem to be associated in the minds of provided strategies of integrating lines into performers and thereby enable the generation the acoustic texture of a poem. of metrically well-formed lines for relevant units of content. In kalevalaic epic, a formula- From Formulae to Formula-Systems system will be an open slot formula or Before advancing to macro-formula multiforms metrical-functional equivalence set of open and variation within them, it is necessary to slot formulae (e.g. where the fixed element is introduce briefly some aspects of formulaic in the same metrical positions) and a language in this poetry and a different type of corresponding equivalence set of formulae multiform that operates in the formation of a that are suitable as slot-fillers to complete a line or couplet. Although technically a line (e.g. as the grammatical subject). multiform, this will be referred to as formula- Open-slot formulae are formulae in which system here for easy differentiation from stable verbal material is part of a grammatical multiforms realizing multi-line units of construction that is completed with a verbal flexible length discussed below. However, slot-filler in use (see also Acker 1998: 39– before advancing to formula-system 43). An example can be taken from the text multiforms, it is necessary to note that the above (ll. 57, 67, 79), comparing the three terms ‘formula system’, ‘formulaic system’ examples of anna [‘I will give’] formulae in and ‘system of formulae’ get used in different which the line is completed by two three- ways in connection with OFT. In his syllable lexemes, one of which is consistent foundational works, Milman Parry used across two of the three uses (in curly brackets; ‘system’ to refer to: a) the operation of slot-fillers underlined): formulae and the poetic idiom generally (e.g. (2) Annan {ainoan} X 1971 [1928]: 19 [23]); b) to formulae with the Annan kultia kyperän same basic meaning completing different Annan {ainoan} orihin metrical positions (1971 [1928]: 19 [23]); c) Annan {ainoan} sisäreni to formulae of the same syntactic category and inflection (e.g. a nominative singular I will give {my only} X I will give a helmetful of gold noun phrase) in the same metrical position but I will give {my only} stallion with different meanings; as well as to d) what I will give {my only} sister are here considered open-slot formulae with variations (1971 [1930]: 274–279 [84–88]). I Within this poem, the open slot of this prefer to reserve ‘formulaic system’ for formula is syntactically, semantically and formulaic language in a register on the whole metrically conditioned. It is syntactically and how that formulaic language operates. conditioned as the object of the verb. It is semantically conditioned as something of 73 value or significance. It is metrically slot formula with a variation anna ainoan X, conditioned with allowance for alternative which can be described: anna {ainoan} X. forms. The metrical line-type (2 3 3) generally Nevertheless, all three examples of anna appears integrated into the formula and {ainoan} X in the epic are crystallized whole- conditions the slot-filler. The formula has a line manifestations of the formula associated form with the three-syllable ainoa [‘only’] with particular contexts. The line annan when used with a single concrete object. In ainoan sisäreni [‘I will give my only sister’] this case, the slot-filler must have a short is thus simultaneously a variation of the stressed syllable and must either be three annan {ainoan} X formula and also a discrete syllables when inflected or two syllables plus formula which is specifically associated with the possessive suffix -ni (to complete the third a particular event in a particular epic. In this position of the open slot). When kulta [‘gold’] tradition, the specificity of the use of annan is used to refer to abstract wealth, the meter ainoan sisäreni will implicitly reference this resists its placement in the last three positions epic event in any other context no less than To of the line because it has a long stressed be or not to be will reference Shakespeare’s syllable (kul-). It thus follows the verb and a Hamlet for us today. In contrast, sano X [‘said term for a container indicating quantity X’] is an open-slot formula completed by a follows it. Within the epic, each of the three six-syllable nominative noun phrase as the variations crystallize into stable formulae in subject of the verb (see also Steinitz 1934: their own right. Nevertheless, the annan X 56–57). Uses of this formula in particular formula is attested in a variety of contexts and epics also clearly crystallize in the memory of across genres (e.g. SKVR I2 1247, 42, 47). the individual singer into whole-line formulae In Classic OFT, formulae were seen as such as sano vanha Väinämöinen [‘said old invariable. Annan X and annan ainoan X Väinämöinen’]. Nevertheless, the sano X would thus be distinguished as separate formula is used with so many epic characters formulae and then grouped into one of Parry’s and in so many contexts that it lacks ‘systems’. On the other hand, this sort of markedness in itself. telescoping formula was characteristic of Classic OFT maintained a theory of thrift Karelian lament poetry, where a formula that Parry developed in his research on might collapse to a single one of our words in Homeric epic (1971 [1930]). ‘Thrift’ was the the orthographic sense or expand to as many principle that a poetic register would ideally as ten (Stepanova 2015: 256). The potential maintain only one formula capable of for the metrical expansion of formulae is realizing a particular idea in particular found in a number of contexts in kalevalaic metrical positions. This principle was not poetry. For example, names of kalevalaic epic equally applicable to all poetries.25 The heroes tend not to be viewed themselves as situation becomes particularly complicated formulae per se. However, prominent epic where metrical requirements of alliteration names tend to have a four-syllable form with and/or rhyme have shaped the register (see a long first (stressed) syllable that is used at also Lord 1960: 50–57). The requirements at the end of a line (Väinämöinen, Ilmarinen, the level of phonic texture of verse become a Kaukomoinen / Kaukomieli, Joukahainen, determinant on word choice that motivates the Lemminkäinen). These seem to be metrically development of a rich equivalence vocabulary conditioned variants of shorter forms with to ‘say the same thing’ within the variety of which they may be in variation (Väinö, possible sound patterns (Roper 2012; for an Ilmari, Kauko, Jouko, and cf. Lemmin poika extreme case, see Frog 2016). Moreover, [‘Lad of Lempi’]). The four-syllable names Parry’s principle of thrift was not developed also receive two-syllable epithets with a long to consider poetries with conventions of first (stressed) syllable. Together, these fill semantic parallelism, which can in fact the last six of eight positions in a line (vanha require ‘saying the same thing’ with different Väinämöinen [‘old Väinämöinen’], etc.). words in the same metrical positions in Rather than forming a system of discrete parallel lines (Fox 2014; see also Frog formulae, anna X can be viewed as an open- 2015b).

In kalevalaic epic, thrift does not operate at movement formulae (e.g. lähti [‘left’], ajo the level of ‘words’ of the poetic register [‘drove’]). Parry would have approached except in some special cases such as these as a ‘system’ that becomes associated in formulaic heroic epithets. In some cases, the mind of a performer. He would have equivalents are alternants that are full viewed formulae capable of completing the synonyms or even allomorphs that might be open slot as a parallel system. When talking unmarked in their interchangeability. In about such indexically linked ‘systems’ in the others, the difference between alternatives mind of a performer, an etic description of all may be as exclusive markers of dialect ‘words’ of an equivalence class within the difference (Frog 2010b: 99–100). For register narrows to the ‘words’ that are example, a variety of formulaic expressions established as an equivalence set in practice. are used as an open-slot formula in the There are methodological challenges to opening two positions of a line when assessing equivalence sets in the empirical establishing a grammatical subject for data: as a psychological reality, they can be subsequent verses. This is a formula like tuo assumed to vary from individual to individual on [‘that is’], se on [‘it is’], oli [‘was’], itse and equivalence sets will not necessarily be [‘him-/herself’], and so forth, followed by a uniform across all dialects. Nevertheless, six-syllable heroic epithet–name formula. ‘words’ of equivalence sets can be inductively This will here be called an identity-formula. deduced from social patterns of language use The meaning-bearing part of such lines is the in the corpus. Equivalence sets in the mind of six-syllable noun phrase (e.g. vanha a performer may remain inaccessible, but at Väinämöinen) identifying the grammatical least the core of such sets can be inferred, subject. It would be an oversimplification to allowing actual equivalence sets to be state that identity-formulae are merely a discussed as hypothetical abstractions evolved means of completing a metrical line, but they around such cores. Neither the set of open- remain semantically very light. Such slot formulae nor the corresponding set of semantically light elements seem inclined to noun-phrase formulae is viewed as a system minor variations. Many such variations in per se here: the formula-system operates in light elements like these seem to have been the linkage between these equivalence sets largely invisible or only textural in the that forms a paradigmatic network in the mind continuous flow of performed discourse. of a performer. The formula-system may also Interestingly, these functionally common but include more specific open-slot formulae of a semantically marginal formulae seem to relevant metrical-functional type (e.g. iski maintain variety rather than exhibiting thrift. [‘struck’]): Open-slot identity formulae filling the first (3) tuo on ┐┌ vanha Väinämöinen foot are completed by a grammatical-metrical sitten ││ nuori Joukahainen equivalence class of noun phrases that can oi on ├┤ Ahti saarelainen operate as slot-fillers completing the sano ││ Tiera Lieran poika remaining six positions of the line. Those ajo ┘└ Pohjolan emäntä slot-fillers are syntactically conditioned by the etc. etc. formula so that they can only appear in the that is ┐┌ old Väinämöinen nominative. Open-slot formulae also group then ││ young Joukahainen into metrically conditioned grammatical- oh (is) ├┤ Ahti island-dweller functional equivalence classes. For example, said ││ Tiera, son of Worm there is a whole set of formulae filling the drove ┘└ Mistress of Pohjola open foot of a verse followed by a six-syllable etc. etc. noun phrase as the grammatical subject. A formula-system qualifies as a multiform These include identity formulae, conjunctive because it advances from individual emic formulae (e.g. sitten [‘then’], siitä [‘following ‘words’ of the registral lexicon and their from that’]), vocative formulae (e.g. oi on alternatives (i.e. in equivalence classes) to an [‘oh (is)’], oi sie [‘oh you’]), inquit formulae associative system in the mind of a performer (e.g. sano [‘said’], virkko [‘uttered’]) and for combining those ‘words’ into larger units.

Within the context of individual epics, these circulates as an ordered series of ‘words’ that combinations crystallize to operate as whole- (normally) account for all of the elements of line formulae. Such crystallization reinforces the multiform in entextualization. Crystallized the association between constituent formulae sequence multiforms seem to be a basic type and equivalence sets within the network that for most narration. These and other multi-line is the multiform. As a consequence, individual (epic) multiforms that operate as macro- performers develop preferred although not formulae appear to resist interpenetration in exclusive combinations within their idiolects. almost all dialects of singing: they seem to For a performer with native-like fluency in have functioned as atomic ‘chunks’ of text in the idiom, expressing a desired semantic unit the progression of narrative performance (see within the field of such a multiform could also Timonen 2000: 642, 652). Although potentially be more or less automatic. The several different types of multi-line paradigm can also be extended to new multiforms are discussed in terms of their metrically and functionally appropriate flexibility and variability at a social level, express-ions in either position. Such a system these show a tendency to evolve into can also be inferred behind variations such as crystallized series multiforms in the minds of Vassilei’s use of lieto Lemminkäinen [‘loose individual singers. This tendency stresses the Lemminkäinen’] in the place of the need to maintain a perspective on social metrically-grammatically equivalent nuori variation within and across dialects on the one Joukahainen [‘young Joukahainen’] in the hand and how the same multiform may example epic above (ll. 2, 10, 16, 41, etc.). In operate for an individual singer. addition, a single open-slot formula such as Dialectal variation becomes apparent very sano X or annan {ainoan} X can be discussed quickly in the corpus, as do exceptional as a formulaic integer or ‘word’. However, examples that are most likely either unique to when discussion advances from the formula the singer’s idiolect or variations resulting per se to its relationship with an indexical set from the documentation situation. The of potential slot-fillers associated with it in multiform that conventionally opens The the memory of fluent users, focus is no longer Singing Competition offers useful illustrations on a single ‘word’ but on a system of ‘words’ of some issues of variation in this type. A and how they interrelate – i.e. a formula- survey of 161 examples and fragments of the system, which is not itself a ‘word’ of the multiform26 shows that it has a core three-line lexicon, but a multiform. structure, as in lines 1–3 of several of the The system formed by an open-slot Malinens’ performances above (cf. OM2 and formula and its conventional slot-fillers tend JO1–2). This multiform constitutes a clause: a to be the most observable formula-systems in couplet presents the identities of two heroes the data. This type of multiform accounts for followed by usually only a single line the systemic operation, observable in the presenting the predicate: they drive into one corpus, of complementary types of formulae. another on the road. The opening couplet is However, the psychological reality of such often a two-line open-slot formula yksi (on) X formula-systems remains theoretical. Their / toinen Y [‘one (is) X / the other/second Y’] extent of development and precise forms can completed by the identities of the two heroes, be assumed to vary considerably according to Väinämöinen and Joukahainen. This is a the competence, experience and interests of manifestation of a couplet formula-system. individual performers. What must be stressed, Väinämöinen is most often named first, however, is that the isolation of formulae that although this varies; the formula naming the participate in this system from one another second hero sometimes varies with a metrical may be misleading and produce an inaccurate and functional equivalent (e.g. lieto view of variation in the poetry. Lemminkäinen). The main variations of the open-slot formula are that yksi [‘one’] varies Crystallized Series Multiforms socially with other metrically and functionally In kalevalaic epic, a crystallized series equivalent identity formulae, and that toinen multiform is a type of macro-formula which [‘other/second’] varies with a conjunction

(sekä [‘along with’], ja myös [‘and also’]) or through an additional couplet in a dialect is can be replaced by a metrically equivalent generally uncommon. Its expansion with identity formula. The verb in the following related couplets is an indicator of a line is most often a derivative of ajaa [‘to connection between the redactions in these drive’] along with the lexeme vasta- dialects or of the two dialects of singing more [‘against’]. For example: generally. (4) Tuop’ oli vanha Väinämöini To the south in Ingria, where kalevalaic Toini on nuori Joukahaini poetry shifted to primarily a women’s Ajettihpa vassakkaha tradition, variation in the poems is more (SKVR I1 163.1–3, punctuation removed.) pronounced (see also Harvilahti 1992a). In Central Ingria, another multiform became That was indeed old Väinämöinen The other (is) young Joukahainen connected with the beginning of the song. Were driven indeed against one another This additional multiform would vary in order with the more common multiform or even be This multiform is specific to The Singing embedded in it or divide it with repetitions.29 Competition: it indexes the epic poem so that This sort of variation is particularly striking those familiar with it recognize immediately because it contrasts to the atomic chunking of which poem is being sung. The multiform’s macro-formula multiforms in most other verbal framework remains recognizable even regions. In the Soikkola region of western when exceptional slot-fillers (underlined) Ingria, the common three-line form developed complete the unit: the third line into a couplet and followed it (5) Yksi kirkos tuloopi with an additional couplet which is unique to Toinen kirkkohe menööpi this redaction. In analysis, it is ambiguous Vastatusten vastavuttiin whether this couplet is an extension of the end (SKVR VII1 140.1–3, punctuation removed.) of this multiform, an extension of the One comes from the church beginning of the following one, or an 30 The other goes to the church independent element linked to the theme: Met those that met (6) Yks oli vanha Väinämöin From the White Sea down to the Ladoga Toin oli nuori Joukamoin region and Karelian Isthmus, the three-line Puuttuit tiellä vassatuksin paradigm remains fairly consistent. The data- Ritahuivat rinnatuksin set reveals that the stable base of this three- Yks ei noista tietä anna Toin ei tieltä pois pakkeene line structure can be expanded through (SKVR III 1176.1–6, punctuation removed.) parallelism and/or less frequently through 3 additional couplets. Such expansions could be One was old Väionämöinen done at the level of individuals but could also The other was young Joukahainen be established at the level of dialects. There is They jammed against one another on the road They got tangled abreast of one another also a case where the multiform is expanded One would not give the road with equivalent couplets in two The other would not flee from the road geographically remote local dialects. In a dialect associated with the Vuokkiniemi In practice, the flow of performance would be singing area in Viena Karelia in the north, the continuous: there would not be breaks relevant couplet was Selvällä meren selällä / between multiforms as units. The ambiguity Ulavalla aukiella [‘On the clear back of the here is increased by the fact that the grouping sea / On the wide open (sea)’].27 In a of elements into units may not have been distinctive redaction of the multiform in the consistent or even clear at the subjective level Shemeikka kin-group’s dialect of singing in of individual singers. To borrow the words of the region of Border or Ladoga Karelia, the Ferdinand de Saussure (1916: 146): “only relevant couplet is Meren selvällä (or: habit and attention allow us to distinguish Selvällä meren) selällä / Lagialla lainehella particular elements” of language. In most [‘On the sea’s clear back / On the open regions, multiforms become apparent through wave’].28 The expansion of the multiform variation in the organization and omission at

77 the level of ‘chunks’ of text in variants of an The broader the scope of the corpus from epic. In the case of the couplet in (6), when it which generalizations are developed (e.g. is used, it is in this position, including in two ‘kalevalaic epic multiforms are, as a rule, not instances where the subsequent multiform of interpenetrating’), the more important it colliding sleighs is absent (SKVR III1 1175; becomes to remain sensitive to the fact that 1180). There is only one exception in which it such generalizations may not necessarily follows the subsequent multiform (SKVR III1 accurately represent every local dialect. 1182). The onsets of crystallized series multiforms generally tend to be the most Syntactic Structuring and Lexical Variation stable (see also Siikala 1986: 198–199). The Crystallized series multiforms tend to exhibit onset might vary in uses of parallelism, but an internal hierarchical structure. They are adding a couplet as a new onset to a quite often built from a dyadic pairing of two crystallized series multiform would be elements, each of which may be rendered exceptional, making it seem more reasonable through a parallel couplet or a series of lines to consider the couplet in (6) to result from or couplets. The elements of such dyadic pairs extending the end of this multiform. are often images or motifs that together The problems of interpreting this example represent a symbolic unit of narrative content. are of interest in part because, even if this Meter is fundamental in structuring lines and couplet emerged as an extension of the end of their crystallization into formulaic units. one multiform, “habit and attention” on a Focus on language at the level of lines may scale of generations could potentially lead it make poetic syntax seem to have relatively to be viewed as the onset of the following few organizational constraints. The language one. In (6), the collector placed a period after is heavily inflected, which allows great the fourth line and treated lines 5–6 as the flexibility in word order within a line: beginning of the following multiform. Rather although there were syntactic constraints, the than the following multiforms being primary ordering principles were in general completely absent in the two instances just the metrical placement of stressed syllables mentioned, if the couplet were its and the inclination to place longer words at conventional opening, these cases could then the end of a line irrespective of word class or reflect a variation in which a multiform is inflection. In this poetry, however, syntax is a collapsed to present its onset only (from primary structuring principle in the which the rest can be inferred by a competent organization of lines in series (see also Kuusi audience). On the other hand, the singing 1952), which reinforces the crystallization of conventions in Ingria were also more flexible that series in a multiform unit, as also seen in and multiforms could interpenetrate. As a the multiform in (4–6) above. Syntax thus consequence, it is not always clear where the operates at multiple levels in this poetry. flexibility of a multiform in this region may These levels include syntax within the line, have evolved so that its parts became separate interlinear syntax through which lines relate multiforms, or whether some multiforms to one another, and also at the level of dissolved within the dialect from textual relationships between symbolic elements or chunking into independent lines and couplets information communicated through clauses or associated with a theme. The operation of sentences in verse. multiforms in Ingria requires much broader Poetic syntax dictates that the nominal study in order to develop a better grasp of subject or subjects will appear in the first line their variation.31 Generalizing distinctions or series of lines (as above). The verb will between multiforms at a regional or supra- either be in the line with the subject or in the regional level is an abstraction that provides a immediately following line. The grammatical frame of reference for considering the object will always appear in the same line empirical data and variation observed within with its verb. Additional information that data. At the same time, multiforms may concerning the how, where, when and why of not have been differentiated in the same way the action can then appear as full-line noun in the mind of every singer within a region. phrases, prepositional or postpositional

78 phrases. Within this basic syntactic fragments of this multiform offer the basis for framework, individual lines may be extended generalizations offered here. This description through semantic parallelism. As John W. Du is most commonly formed by a dyadically Bois (2014: 392–400) discusses, syntactic structured multiform of two couplets, parallelism can itself be viewed as a syntactic especially in the more northern regions: phenomenon at the level of “the formal (7) Puuttui vemmel vempelehe relation of signs to one another” (Morris 1971 Rahet rahkehe nenähän [1938]: 22). In kalevalaic poetry, each Veri juoksi vempelestä syntactic component of a semantically parallel Rasva rahkehen nenästä line should match a component in the (SKVR I1 163a.4–7, punctuation removed.) preceding line while other components of that Stuck shaft-bow to shaft-bow line may be inferred (Kuusi et al. 1977: 66). Traces to trace’s end Ellipsis of syntactic components in parallel Blood ran from the shaft-bow lines thus subordinates lines in an ordered Fat from the trace’s end hierarchy: the line with all of the components One dyadic constituent of a multiform of this must appear first. Semantically parallel lines type may also reduce to a single line, manifest are possible as adjacent repetitions of each as a series of lines or even as a series of semantic unit of a clause. The structuring of couplets, as in examples (11–12) below. Such parallel lines thus occurs within the syntactic multiforms may also be truncated, for structuring of a series of lines, where it is example presenting only the first element of a subordinate to the interlinear syntax of the dyadic pair,32 or the multiform could be clause. Interlinear syntax can thus be seen as a elaborated with one or more couplets in the primary organizing principle for ‘words’ in same manner as example (6) above. The multi-line multiforms. dyadic structure of this multiform family (i.e. Multiforms may be constituted of multiple of its redactions) nevertheless remains clauses or sentences (see also Honko 1998: generally consistent across dialects from the 112). The dyadic (or triadic) structuring of Gulf of Finland to the White Sea. The first multiforms also engages poetic syntax at the element represents the sleighs stuck together level of relationships between clauses and the through different metonyms for sleigh information or symbols that they mediate. The organized in parallel lines, and the second conventional juxtaposition of such units is refers to liquid running from the parts of a generally indicative of: a) the first element sleigh as an expression of the strain of the standing in a causal relation to the second as horses, again in parallel lines. its outcome, as in (7–11); or b) of the second The line-order within each couplet in (7) is element being a characterization of the first, fixed by the poetic syntax: ellipsis of the verb possibly including specific relationships to is only possible in the second line of a parallel other characters or groups connecting it with group. The order of the two-part series is the plot, as in (15). The stability of correspondingly stabilized by the implied crystallized sequence multiforms as an causal relation between conventionally ordered series of ‘words’ is often directly juxtaposed clauses – i.e. liquid runs from the connected with poetic syntax, which places horses because the sleighs have collided constraints on the organization of those rather than vice versa. Lines in this multiform ‘words’ and symbols or information that they family tend to subscribe to basic schematic communicate. patterns, in this case as open-slot formulae: Dyadically structured multiforms often manifest as paired couplets. The multiform (8) Verb1 Noun1-NOM Noun1-ILL that conventionally followed the opening of Noun2-NOM Noun2-GEN nenä-ILL* The Singing Competition describes the Noun3-NOM Verb2 Noun1-ELA heroes’ sleighs locking together when they Noun4-NOM Noun2-GEN nenä-ELA * nenä is a postposition ‘end’, lit. ‘nose’. have collided. It is useful for illustrating dyadic structuring and its potential for In this case, the lines of each couplet are variation. A survey of 135 examples and semantically parallel while the couplets

(9) Puuttui vemmel vempele- hän / Rahet rahkehe- n nenä- hän juoksi *Veri vempele- stä / Rasva rahkehe- n nenä- stä

Stuck shaft.bow shaft.bow to / Traces trace ’s end to ran *Blood shaft.bow from / Fat trace ’s end from * This word has been moved from its position in the verse to make the parallelism more visible. themselves are linked by lexical repetition rahis as opposed to rahkis): the verb and the and structural parallelism, illustrated in the nominative noun may exchange positions. In diagraph (Du Bois 2014) in (9). However, the second couplet, word order is regular. rather than structural parallelism being regular Most words for liquid such as vesi [‘water’], between couplets, the line-types could repeat veri [‘blood’], hiki [‘sweat’], and so on have a within a dyadic element, and another line type short initial syllable and thus are metrically with a different syllabic structure was also avoided in the second foot. Rasva [‘fat’] common (e.g. aisa aisahan takeltu [‘shaft to would be metrically appropriate in this shaft was jammed’]). position but is used primarily in the second The lexemes completing the slots in these line-type of the couplet (52 instances in rasva schematic templates are semantically, rahkehen nenästä; 20 instances in 10 other syntactically and metrically conditioned, verbal lines). Word order with rasva is only which significantly limits their variation. For inverted in one instance (SKVR VII1 140.5) lines of the types in the example above, nouns and this is with an unconventional verb consist of two core groups of equivalence otherwise only encountered once in examples classes. The first are semantically conditioned of the multiform (lähteä [‘leave, depart’]). as words for parts of a sleigh that form a Actually, the whole variant in which this use functional equivalence class of terms that can of rasva appears is generally unusual: it is the make a metonymic reference to the sleigh source for the variant of church-goers in itself (Noun1–2). They are metrically example (5) above. conditioned as having a long initial syllable Additional line-types evolved for use with and an additional syllable on their inflected two-syllable words with a long initial syllable stem (rahis or rahkis / rahkehe- [‘trace’], that could be used as a metonym for sleigh vemmel / vempele- [‘shaft-bow’]). The second but which did not receive an additional set of nouns (Noun3–4) is of two-syllable syllable when inflected, such as aisa / aisa- words for fluids which can alliterate with a [‘shaft’]. These were used with a set of three- respective metonym for sleigh (rasva–rahis syllable verbs with a short stressed syllable [‘fat–trace’], veri/vesi–vemmel [‘blood/water– (takeltua [‘to get jammed’], taveta [‘to shaft-bow’]). There are two corresponding meet’]). This example also adds perspective sets of equivalence classes for verbs. Both on variation when verbal correspondences to couplets of the type presented here take two- the preceding example in (7–9) are syllable verbs with a long stressed syllable, underlined: making them metrically appropriate for the (10) Tartu [sic] vemmel vempelesen second foot. The verb in the first couplet is Aisa aisahan takeltu semantically conditioned to refer to the Vesi vuoti vempelestä interlocking sleighs (Verb1: e.g. puuttua [‘to Usva uuen aisan päästä be jammed’], tarttua [‘to become stuck’], Rasva rahkehen nenästä koskea [‘to touch’], kerttia [‘to pile up’]). The (SKVR I1 163a.4–7, punctuation removed.) verb in the second couplet is correspondingly Stuck shaft-bow to shaft-bow conditioned to refer to movement of liquid Shaft to shaft was jammed (Verb2: e.g. tippua [‘to drip’], kiehua [‘to Water flowed from the shaft-bow boil’], vuotaa [‘to leak, run’], juosta [‘to Fog from the new shaft’s head run’]). Metrically, this allows limited variable Fat from the trace’s end word order in the first line where the nominal The lines of this multiform tend to conform to form of the noun has a long stressed syllable syntactic templates interfaced with the meter, (i.e. for vemmel but not for the Karelian form which provided formulaic templates for 80 generating alternative lines. Thus lines like where such growth can represent both the rahis rahkeeseen takeltu (SKVR I1 158.4) passage of time and/or the expression of [‘trace to trace was jammed’] emerged where supernatural power. This redaction illustrates changes in spoken language dialects affected the potential of a central constituent of a articulation in versification; this reduced the multiform to change while another major inflected form of rah(k)is (rahkeheseen) by a constituent maintains continuity (i.e. half of a syllable and required it to be treated as dyadic pair or equivalent part of a more analogous to the line with aisa in (10). complex multiform). This can be viewed as To the south, in Border or Ladoga Karelia an innovation ultimately attributable to one or and in Ingria, this multiform tended to be more individuals, but it should be stressed longer. It became exceptionally long in West that emphasis here is on the social outcome of Ingria, where repeated use of formulaic the innovation being assimilated as a valid templates is more apparent and where model by other performers so that it becomes couplets would be generated for as many as a marker of both a local dialect of singing and three different liquids in a parallel series, as in a local dialect of mythology (see also example (11): Stepanova 2012: 280). (11) Aisa aisaa tappais Although the narrative unit of the colliding Rahe tarttui rahkehee sleighs was generally expressed through a Rinnus tarttui rinnuksee distinct multiform at the broad social level of Vemmel tarttui vempelee the tradition, that does not mean that every Rikki kiehui rinnuksesta singer necessarily internalized it that way. Rikki rinnuksen nenästä There are some cases in the data where this Rasva kiehui rahkehesta unit appears only as a couplet as in (12), and Rasva rahkehen nenästä that couplet may potentially have been Vesi kiehui vempelestä internalized as a component of the opening Vesi vempelen nenästä multiform discussed above: (SKVR III1 1182.20–29, punctuation removed.) (12) Yks ol vanha Väinämöine Shaft met shaft Toinen nuori Jougavoine Trace stuck to trace Ajettih on vassakkah Strap stuck to strap Läks_on vezi vembelestä Shaft-bow stuck to shaft-bow Rasva tippu rahkehesta Sulphur boiled from the strap (SKVR II 37.1–5, punctuation removed.) Sulpher from the strap’s end Fat boiled from the trace One was old Väinämöinen Fat from the trace’s end The other young Joukahainen Water boiled from the shaft-bow Were driven (is) against one another Water from the shaft-bow’s end Left (is) water from the shaftbow Fat dripped from the trace On a quite local basis, the mythic images of this motif underwent renewal in the dialect of In this case, the singer performed this opening mythology – i.e. at the level of mythic sequence the same way eleven years later symbols and narrative structures being (SKVR II 37a), but it is impossible to resolve mediated through verse. The interface of these how she perceived the relation between the images with their verbal representation – the units or how they were established in her linguistic encoding of mythic symbols memory. In another case, a singer who through verbal art – thus produced a distinct performed a similar five-line sequence quite local variation in the dialect of singing. The exceptionally opens the fourth line with the second dyadic element of the multiform was conjunction ja [‘and’] (SKVR II 32.4). The replaced by an image of trees growing from conjunction invites interpreting a strong link the parts of the sleigh, as reflected in the that could indicate a coherent multiform, but example of the Malinen family illustrated even this remains ambiguous since this singer above (multiform M2b; see also SKVR I1 265, used ja at the beginning of several lines (a 278–279a, 284a, 293). This image is adapted feature lacking in her performance of another from the symbolic system of the mythology, epic for the same collector: SKVR II 217).

This emphasizes that general perspectives looked at in overview, most variation seems about the social level at which multiforms are to be the result of exchanging metrically and observed provide frames of reference for semantically equivalent words in the locally considering specific cases and individual circulating formulae. performers, but it cannot necessarily resolve all questions surrounding the specific case. Collocative System Multiforms When looking across the many examples Crystallized series multiforms, characterized of this multiform collectively, variation may by the ordered sequence of elements, are appear chaotic. The local and regional forms paralleled by what are here called collocative found from the Gulf of Finland to the White system multiforms. These are multi-line Sea may seem quite diverse, but they exhibit multiforms in which the constitutive ‘words’ local regularity and stability. Moreover, the are not bound to a particular organization variety in the corpus may appear much greater although they are associated with one another than it actually was in practice. The and manifest as co-occurring verbal resources. organization of the two parts of the dyad is Technically speaking, all multiforms are invariable, although only the first of these collocative systems of ‘words’, but using this may appear. The 135 examples surveyed here as a categorical term refers to multi-line include 534 lines representing the symbolic multiforms that do not exhibit the structuring elements of the interlocking sleighs and principles that distinguish other formal types. liquids running from them (i.e. not including These also do not necessarily operate as lines of the alternate dyadic element in the macro-formulae, and may instead operate as a Malinens’ dialect or e.g. couplets with other cluster of lines, couplets and/or larger macro- types of specifying information). These 534 formula units so that the uses and lines represent a total of 70 verbal types arrangements of these co-occuring parts are identified by lexical elements and word more open to polysemy of the whole. order33 (not including uses of the particle – Some of the multiforms discussed by the pa/pä or phonological variations related to Honkos (1998; Honko 2003) probably belong dialect or dictated versus sung performance). to this broad class insofar as the verbal Of these, 55 verbal line types were framework can operate without following a encountered three times or less, accounting regular sequence. In kalevalaic epic, ‘words’ for 80 lines or 15% of the dataset; 15 types of these multiforms exhibit indexical binding account for 454 of these lines or 85% of the to each other on the level of verbal texture but dataset.34 Among these 15 types, there are also to a sequence of narrative content like a lines which are clearly regionally established ‘theme’ of OFT. This often makes it difficult and others which have potentially been to discern whether the ‘words’ are independently generated according to the independently associated with that content basic schema. Verbs exhibit more variation rather than with each other per se. However, across the corpus with clear regional framing the elements as either indexically preferences, yet these patterns of variation are associated with one another or with the unit clearly rooted in the indexical matrix of the of narration as a binary opposition would multiform as a verbal and compositional probably in most cases be a misleading system. Looking at all of the examples of this oversimplification of a potentially quite multiform as an unorganized mass of dynamic relation. Nevertheless, the more examples may give the inaccurate impression flexible a multiform of this type appears, the that it was a highly generative system – which more difficult it is to recognize with any such a multiform might be in another certainty. tradition. However, the multiform was a Because of the inclination to non-variation relatively stable component of local dialects in kalevalaic epic, variation in this type of of singing and crystallized in the minds of multiform seems for the most part to occur in individual performers. The syntactic-metrical transmission. In other words, the multiform’s templates for producing lines and couplets manifestations vary between individual had the potential to operate generatively, but, performers but it seems to evolve into (or at

82 least toward) a crystallized sequence interpenetration (outside of Ingria), the multiform in the mind of each performer. couplets of this constellation can be When looking at the corpus, it may therefore interspersed with and linked to other appear quite variable, whereas it will appear multiforms. The potential for dispersal across far more regular or even highly crystallized a larger text sequence involving other when looking at multiple performances by the multiforms suggests that the elements can same individual. In some cases, the crystallize into separate multiforms at the crystallized sequence may also be transmitted level of idiolect. This makes it reasonable to within a dialect of singing and variation only question whether the proposed collocative becomes evident when looking across system was a coherent multiform in social dialects. An initial impression that variation is circulation. At the same time, this sort of quite free might thus prove misleading when variation parallels variation observable in local dialects come under scrutiny. some multiforms in Ingria and may be linked An illustrative example can be taken from to the variation in kalevalaic multiforms in a system of couplets that co-occur in the epic more flexible genres such as lyric – a The Song of Lemminkäinen. This system of possibility which requires future testing in couplets occurs in conjunction with the hero analysis. filtering the poison from drink he has been Collocative system multiforms may also served and his consumption of that beer. The form much more cohesive units. In the epic three couplets are proverb-like units normally The Song of Creation, for example, a attributed to the hero as direct speech: cosmogonic egg breaks and the world (13) Toppa moaha luodanehe emerges from its parts or is created from it by Ruoga suuhu süödänehe the demiurge Väinämöinen. The relation of Tuopin tuoja Tuonelaha parts of the egg to especially celestial Kannan kandaja Manalla phenomena is manifested by a multiform Joi oluon onnekšehe often formulated as a series of couplets as in Meem mussam mielekšehe example (14): (SKVR I 716.209–214, punctuation removed.) 2 (14) Mi munassa valkieta Let waste be cast into the earth Se päiväkse paistamahe Let food be put into the mouth Mi munassa ruskieta The mug’s bringer to Tuonela [the realm of Se kuukse kumottamahe the dead] Se tähekse taivosella The can’s carrier to Mana Mi munass’ ülistä puolta [He] drank beer to his fortune Vaskisekse taivosekse Black mead to his satisfaction Mi munass alaista puolta A dataset of 115 instantiations of this Rautasekse moaemäkse (SKVR I 74.119–127, punctuation removed.) (potential) multiform show that the order of 1 the couplets varies and not all couplets What [is] white in the egg necessarily appear (see Frog 2010a: 365– That into the sun to shine 371). Individual couplets may be expanded or What [is] brown in the egg [= the yolk] reduced to a single line; they may be That into the moon to glow That into a star in heaven interspersed with additional lines or couplets What in the egg [is] the upper part and form repeating series. The individual Into the copper heaven couplets may also be attributed to different What in the egg [is] the lower part speakers as direct speech – a type of variation Into the iron mother earth that belongs to multiforms’ potential for polysemy (Frog 2010b: 102–103). The This multiform circulated as a stable unit. continued flexibility of the uses of these Although organization was variable, it was couplets appears related to the semantic core also hierarchically structured. Each couplet of each unit remaining distinct in would normally be structured according to a transmission. Although kalevalaic epic formulaic template and the pairing of the part multiforms generally exhibit resistance to of the egg with what is created from it was highly conventional. Unlike this example, 83 however, the stars and/or clouds would often from 113 variant poems and fragments. be rendered as a separate couplet describing Within the epic, the multiform may be used in what was created from flecks in the egg or a dialogue where the hero’s mother mentions fragments of shell (see further Kuusi 1949: the dangers on his road, on the hero’s journey 168–170). The couplets concerning the itself, or both, yielding two manifestations of white–brown and upper–lower parts of the the multiform in a single performance. Of the egg were conventionally paired. The whole 113 variants used here, 45 exhibit such multiform was ideally constituted of three repetition within a single epic performance. units, the white–brown couplet pair, the The stability of the first half of the dyadic upper–lower couplet pair, and the stars multiform is attributable to anaphora as an couplet. The units would form an unbroken organizing principle in a series of series but their order was not socially stable grammatically parallel lines. In other words, (although the stars couplet would not be first) the last word of each line becomes the first and not all of them necessarily appeared. word of the next in a progressive series. The Preferred organizations of these elements anaphora begins with a formula indicating the were associated with dialect, but most encounter with a ‘fiery X’ followed by a significant for the present context is that repeating open-slot formula X-INE(/-ADE) variation in the ordering of elements was not tulinen Y [‘in the X a fiery Y’]. The series exclusive of structuring mechanisms such as varies in length from three to six elements but interlinear syntax operating within the the order is invariably conditioned as a multiform. progression of smaller objects concluding with the eagle or its talons. The second half of Intermediate and Hybrid Types the multiform is normally comprised of either Categories of crystallized sequence and two couplets characterizing the eagle of the collocative system multiforms are abstract first half or of one such couplet followed by a and ideal while the reality could be much line expressing the eagle’s readiness for hero. more fluid. For example, the multiform of the Its particular form may be stable within local colliding sleighs is extremely stable in its dialects, but it appears that singers were form of paired couplets. However, as each frequently aware of multiple alternatives and couplet expands, the internal organization of could combine these into longer units, as in each part can open to different sorts of example (15): variation. The long redaction from West (15) Tuloop on tulińi joki Ingria illustrated in example (11), for Jovess’ on tulińi koski instance, allows variation in the order of Kosess’ on tulińi luoto elements in the first part as well as in the Luuvoss’ on tulińi koivu second, and even whether lines or couplets are Koivuss’ on tuliset oksat used for each liquid in the second part. In that Oksiss’ on tulińi kokko case, the multiform maintains an overall Kokko kynsiäh hioo dyadic structure while each dyadic component Hampahia hiiskuttau may vary from dialect to dialect between a Se on syönyt saa’an urosta crystallized sequence and a collocative Tuhonnun 1000 miestä system. Sepä on syömäh Lemminkäistä Multiforms may also circulate as hybrid Lemminkäisen peän varalla types. In The Song of Lemminkäinen, for (SKVR I2 754.70–81, punctuation removed.) example, one of the dangers passed by the hero on his journey is a fiery eagle. This Comes (is) indeed a fiery river monster is described in a dyadically In the river (is) a fiery rapids structured multiform of which the first part is In the rapids (is) a fiery skerry In the skerry (is) a fiery birch a crystallized sequence while the second is In the birch (is) fiery branches constituted of a set of elements that can vary In the branches (is) a fiery eagle in organization. This multiform is discussed here on the basis of a dataset of examples The eagle grinds its talons Makes its teeth whisper

It has eaten a hundred heroes position for a formulaic ‘word’ that will Destroyed a thousand men complete it as a unit. This slot is frequently It indeed (is) to eat Lemminkäinen part of an open-slot formula, such as the anna On hand for Lemminkäinen’s head {ainoan} X formula mentioned above. The Variation in the first element of this dyadic semantic unit represented by the ‘word’ pair occurs in the formulaic opening of the completing any such open slot may then be first line (i.e. tuloop’ on above) and in both reduplicated through parallelism. The the number and specific terms in the multiform with the anna {ainoan} X formula progression of nouns (normally a minimum of in the Malinens’ dialect of the epic above koski [‘rapids’] or sometimes joki [‘river’] (multiform M10) might be abstracted as followed by koivu [‘birch’] and kokko shown in (16). The open slot of the formula is [‘eagle’]). Variation in the second half is in here indicated as X1 while its equivalent in a the selection and organization of couplets and parallel line is indicated as Y1 with their lines. Such variation may be by convention of grammatical inflections marked; curly dialect or strategic expansion. Some of the brackets indicate verbal elements that vary variation in the dataset is likely attributable to across instantiations in variants collected from processes related to memory and/or the the Malinens: collection context. This could lead a singer to (16){Oi on vanha Väinämöini} mix relevant units used in different dialects. Myössytäs pyhät sanasi The eagle is also one of normally three Perävytä lauhiesi dangers encountered by the hero and some Annan {ainoan} X1-PART/-GEN couplets are used to characterize multiple {Y1-PART} dangers while they also sometimes seem to {Oman pääni päästimeksi get juxtaposed across them (although this is Itscheni lunastimeksi} too complex a topic to elaborate here). When {Oh (is) old Väinämöinen} reference to the eagle’s threat to the hero is Make your holy words harmless included, this almost always concludes the Turn back your sentences 35 multiform. It is most often preceded by a I will give {my only} X1 characterization of the eagle as threatening in {Y1} one or more couplets. These couplets are {For liberating my own head formed through semantic parallelism, any of as a ransom for myself} which may be reduced to a single line, but An open-slot multiform may also entail they remain stable as couplets rather than multiple slots. This occurs in Väinämöinen’s inclining to expansion.36 Although the response that pairs with this plea in the couplets vary in specific semantic content, the Malinens’ redaction. This multiform is paired use of two or more of these can be considered with the preceding one within the theme, a form of symbolic or analogical parallelism which might be abstracted as in (17) (asterisk engaging a single underlying symbolic indicates potential to vary in order): referent (Frog 2014c: 188–193). In any case, (17) *En huoli Y -ELA* the inclination to non-variation stabilizes the 2 {Kusen konna X2-ILL} lines or couplets that form the second portion {Ilkije X3-ELA/Y3-ILL} of the fiery-eagle multiform in the idiolect of {On X4-NOM itschelläniki} a singer and in most dialects of singing. The {Kahta kolmea parempi} main point here is that variation can operate [If X = horse:{Viittä, kuutta virkeämpi} quite differently in the verbalization of each {Kuutta, seitentä somempi}] of the dyadic elements of a single multiform. *I don’t care about your Y2* {A can of piss on your X2} Open-Slot Multiforms {mean one, your X3/Y3} An open-slot multiform is normally a {have indeed myself a X4} crystallized sequence multiform comparable {twice, three times better} to an open-slot formula: it includes a [If X = horse: {five, six times more brisk} semantically and syntactically conditioned {six, seven times nicer}]

Cohesion between the two multiforms is Lemminkäinen and its variation describing his established through semantic equivalence of interlocutor’s compliance, bringing him the the slot-fillers – i.e. both units of speech are armour (see also (21) below). The core about the same thing. The slot-fillers are couplet of the multiform is tuo/toi mulle metrically conditioned by their position in the sotisomani / kanna/kanto vainovoattieni lines. The embeddednes of the slot in an [‘bring/brought me my war-shirt (armour) / open-slot formula is connected with poetic carry/carried my persecution-garment’]. The syntax: these examples involve variation in interlocutor/actor of the multiform is the objects of verbs only; in dialects where invariably female and normally the hero’s the verb also varies, the whole line operates as mother. Within the epic plot, the demand is an open slot within the multiform. In practice, integrated into the hero’s preparations for a however, these completed open-slot multiforms dangerous journey to a wedding in the do not seem to vary their slot-fillers in repeat otherworld. It prompts the interlocutor to performances. They thus appear to operate as respond with a prohibition and elaborate each crystallized series multiforms in the memory one of the normally three ‘deaths’ on the road of individual performers and even in until she complies and prepares the hero for intergenerational transmission. his journey. This multiform varies by dialect with functionally equivalent couplet-core Line-Core and Couplet-Core Multiforms multiforms. It is addressed here through a Whereas an open-slot multiform is dataset of examples in 143 variant poems and customarily a more or less complete verbal fragments. In each of these items, the framework with a few variable positions, a multiform is used between one and seven line-core or couplet-core multiform is a multi- times. The couplet may be presented alone, line multiform in which the semantic or but it is customarily preceded by a vocative functional core of the unit is stable and a address to the interlocutor when used as a surrounding tissue of lines in which it is used demand (at least in the first instance), or by an remains open. These multiforms are identity-formula presenting the nominal organized on a semantic hierarchy. The core subject when narrating the action. The is formed of a minimal verbal unit that vocative address may be extended through contains a central unit of information, parallelism. The couplet itself forms a represents the basic symbolic unit of content, conventionally ordered line pair although it or fulfils the unit’s pragmatic role in does appear inverted in some variants. Lexical narration. This core becomes surrounded by variation mainly occurs in the semantically what Jukka Saarinen describes as light positions 2–3 of the first line (mulle [‘to “subordinate textual elements” that “describe, me’] above; sie nyt [‘you now’] in (18) modify, argue, specify, etc. other parts of the below).38 The core couplet may also be text” but which “cannot appear without the expanded by additional parallel lines or form 37 part [they] specify” (1994: 183). In a series with parallel demands (e.g. for a kalevalaic epic, the core may be a single line sword). The most frequent elements to follow or couplet which can potentially be realized the core couplet are references to the armour without any surrounding tissue, but the with present participles of verbs and the surrounding tissue cannot realize the location where the hero intends to wear this multiform without the core. That tissue varies garment, as in the following example: within a schema or abstract template in which (18) Oi emoni kantajani the core element exhibits stability in social Eli vaisten vanhempani transmission. Some positions in the schema Tuo sie nyt sotisomani will be semantically or functionally conditioned Kanna vainovoattieni while others may only be subject to general Häissä häilyteltäväni constraints of poetic syntax. Pivoissa pijeltäväni A particularly well-attested couplet-core Lähen Päivylän pitoih multiform is found in the hero’s demand for Salajoukon juominkih his war-shirt (armour) in The Song of (SKVR I1 847.73–80, punctuation removed.)

Oh my mother one who carried me distinguished by concentration of its semantic Or my calming parent core in a minimal verbal element of a line or Bring now my war-shirt couplet. Although the example in (18) is eight Carry my persecution-garment lines in length, the core couplet can be all that My glistening one at the wedding is conventional to present in some dialects One containing me at the feast and idiolects. I am leaving for the feast of Päivölä For the feast of the secret band Multiform Frames? The introduction of a new verb and clause Although kalevalaic epic generally exhibits into the couplet of location might be resistance to the interpenetration of interpreted as an underlying dyadic structure multiforms, a special case is what can be of the multiform. However, the couplet of called a multiform frame. This is a multiform location seems to vary with other elements that operates like an open-slot formula. The following the core of the multiform rather integer completing an open-slot formula is a than a two-element structure being stable. lexeme or formulaic ‘word’ in the lexicon that This couplet is actually an open-slot formula might be extended through semantic naming the location; it recurs through the epic parallelism. In contrast, the slot of a to identify the destination of the hero’s multiform frame receives a symbolic integer journey and later becomes the center of of narration mediated through its own action.39 The first foot of this formula is a multiform. It must immediately be stressed two-position open slot completed by a verb that this phenomenon seems to be uncommon with corresponding morphological variation for the epic tradition generally: it seems to in the parallel terms for ‘feast’ (i.e. locative manifest on a local level in some dialects of cases equating to English ‘to’, ‘from’ and ‘at’). singing, although a comprehensive survey of In this case, the appearance of the verb lähteä multiforms in the epic tradition has not been [‘to leave’] in the couplet can be considered to completed. The example in (19) is taken from be related to the operation of the formulaic a variant of The Song of Lemminkäinen, couplet rather than to the structuring of the presenting each of the mother’s warnings to multiform in sentences per se. her son about the three ‘deaths’ he will Like other multiforms with capacity for encounter on his journey. variation, couplet-core multiforms are This example illustrates the issue with this inclined to crystallize in the minds of type of multiform: rather than a normal performers. This variety of multiform is phenomenon of the register, it may have

(19) Three repetitions of a possible multiform frame from an example of The Song of Lemminkäinen. The hero’s intermediate responses are omitted; varying elements between repetitions are in italic font. Kun määt matkoja vähäsen Mänet matkoja vähäsen Mänet matkoja vähäsen Kulet teitä pikkaraisen Astut teitä pikkaraisen Kulet teitä pikkaraisen Tulopa tulinen aita Tulopa tulinen hauta Tulopa tulinen koski Moasta soakka taivahase Moasta soakka taivahase Kosessa tulinen luoto Tšiliskoin[sic] vitšastettu Siinä kuumia kiviä Luuvossa tulinen koivu Moan mavoin seivästetty Palavia poateroita Koivussa tulinen kokko Tulen suustaha puhuu Siin_on surma ensimäinen Siin_on surma keskimäinen Siin_on surma jälgimäinen (SKVR II 222.13–19.) (SKVR II 222.20–26.) (SKVR II 222.30–37.)

When you go on the journey a little You go on the journey a little You go on the journey a little Travel the road a bit Step the road a bit Travel the road a bit Will come indeed a fiery fence Will come indeed a fiery grave Will come indeed a fiery rapids From the earth to heaven From the earth to heaven In the rapids a fiery skerry Lizards woven into a wicket In it hot stones In the skerry a fiery birch With worms of the earth staked Burning slabs of rock In the birch a fiery eagle [It] speaks fire from its mouth In that is the first death In that is the middle death In that is the last death

87 evolved from an open-slot multiform within ‘words’ of the poetry necessarily entails a the particular context. Variation in the schema that requires and conditions the use of opening couplet may be idiolectal.40 The additional ‘words’ indexed by the schema, references to the first two ‘deaths’ may be this can be approached as a schema-based viewed as an open-slot multiform in which multiform. the two-syllable word for the ‘death’ An inquit formula in kalevalaic epic may completes the formula of the third line and a be an open-slot formula that forms a single second slot is a full-line attribute with a line with the grammatical subject like sano X semantically parallel line. The repetition of [‘said X’] discussed in connection with the the line moasta soakka taivahase [‘from the formula-system discussed in relatin to (3) earth to heaven’] is unusual and incongruous above. It may also be a complete line for the hauta [‘grave’] ‘death’. It may be a preceded by the subject as in (20), although slip of the singer; if so, it is an indicator that the subject might also remain implicit. In this line was indexed in her memory as an general, these formulae tend not to exhibit element of this sequence, which would parallelism (Steinitz 1934: 63–61), although it suggest that both examples manifest too may even be expanded into a series of variations on a single multiform. The third lines: example begins with the same three lines, but (20) Tuopa lieto Lemminkäińi then shifts into a distinct multiform. Linking Tahi seppo Ilmorińi multiform units syntactically into a chain is Itte noin sanoiksi virkko found in the case of quotative frames below. Sanan virkki, noin nimesi What makes this group of potential multiform (SKVR I2 772.104–107, punctuation removed.) frames exceptional is the final line, which That indeed loose Lemminkäinen functions as a type of boundary marker that Or smith Ilmarinen concludes the sequence. It is comparable to Himself thus put into words concluding multiforms representing ‘deaths’ A word uttered, thus named with a line like päen varalla Lemminkäisen Although quotative frames were not usually [‘ready for the head of Lemminkäinen’] in elaborated to this extent, example (20) example (15) above. However, that line illustrates that the quotative frame is more operates as part of the multiform representing than simply a formulaic ‘word’ of the poetry: the ‘death’. In this case, the unit of the it has the potential to be elaborated into a mother’s speech can be seen as embedding a 41 syntactically complex unit of multiple emic distinct multiform into a 3–4 line frame. A ‘words’. An alternative heroic identity comprehensive survey of multiforms in all formula is used here in semantic parallelism. epics is necessary to place this in perspective. There is, however, no ambiguity regarding Quotative Frames who is speaking in the epic, which brings Many kalevalaic epics are built around forward an operative point relevant to many dialogic interaction, which means that many multiforms. Predicate formulae may be used multiforms are instantiated in dialogue. A without a preceding formula introducing the quotative frame is a verbal framework that grammatical subject. On the other hand, attributes direct speech without being part of especially the use of formulae such as full- the direct speech itself. At a general level, line inquits can be inferred to index a quotative frames in this poetry are schemas or metrical-grammatical equivalence set of abstract templates with open slots. They formulae for grammatical subjects. This is mandate morphological variation in a basically the same sort of indexical network multiform (e.g. inflecting verbs to the first or of equivalence sets discussed for the formula- second person) depending on the type of system of (3) above, but operating interlinearly. communication they present. These schemas Quotative frames are also invariably a are not themselves multiforms, although they minimum of one full line in length: what the are entangled with inquit formulae – formulae speaker says will only begin in the following introducing direct speech. Where the use of line. They are exclusively initial, never medial or final. In other words, the quotative always

88 introduces a unit of direct speech before it flow of discourse that are not indespensible. begins; it never interrupts it, or indicates that How quotatives and vocatives are used the unit of speech has concluded. This appears to be connected in part with dialect. limitation seems to be connected with poetic However, a variety of equivalent formulae syntax and connected in some way with the circulated in the tradition and quotatives and resistance of kalevalaic multiforms to vocatives vary easily between singers even interpenetration. Consecutive multiforms of within close singing communities, which direct speech attributed to one speaker may suggests that variation between them was not nevertheless mark the onset of each multiform necessarily perceived as a relevant indicator with its own quotative frame (a marking that of dialect. is, incidentally, an indicator of a perception of multiform units). Flexibility of Length The schema of a quotative frame can be The Honkos “contend that an oral epic conceived in terms of three slots, not all of breathes through multiforms” in the sense that which are necessarily instantiated. The first is multiforms enable the flexibility in length the quotative itself, constituted of an agent + characteristic of long epic (1998: 35). inquit formula. Direct speech in kalevalaic Detecting use of multiforms as a resource for epic always entails an addressee, and thus the varying length in kalevalaic epic is rather second element is a vocative address that is difficult to assess because the corpus in most often formulaic (e.g. oi siä X [‘oh you general tends to only preserve a single variant X’]). The third component is the speech unit from each singer. As already noted, however, itself, which undergoes morphological kalevalaic multiforms generally seem to be variation as relevant. The unit of direct speech manifested very consistently both across may be a line, couplet or full multiform; it performances, where such data is available, may also be a series of multiforms (cf. M7–8 and when the same multiform recurs within a above). This unit is subject to appropriate performance. Repetitions within an epic in morphological variation. Where the speech is particular have potential to offer perspective a multiform that otherwise opens with an on flexibility in length, although especially identity-formula of a grammatical subject, earlier collectors did not necessarily write morphological variation may convert this to a down all lines of a repetition (as in ll. 31, 65 vocative address (e.g. se on vanha of variant OM2 above). The corpus therefore Väinämöinen / oi siä vanha Väinämöinen). It does make it possible to draw some inferences is not infrequent for dialogues to anticipate on flexibility. something which will occur or to refer to In the dataset of the fiery eagle multiform, something which has occurred. In such cases only 45 variants indicate that the multiform the multiform or its relevant component will was presented first in the dialogue with the vary accordingly in tense, mood and person. hero’s mother and then repeated in the 42 Direct speech is widely unmarked and account of the journey itself. Some of these quotatives are commonly omitted, particularly only record the first line followed by ‘etc.’ or 43 following the initiation of the dialogue. When its equivalent. In two variants recorded from the quotative is omitted, the vocative phrase the same singer by different collectors nearly may still mark the onset of direct speech, a decade apart, the second element is absent although it too may be omitted and direct from the repetition (SKVR II 217–217a). This speech may only be grammatically indicated suggests that shortening in repetition was and contextually implied. integrated into the performer’s way of Quotative frames crystallize in epic singing. Similarly, a localized cluster of reproduction and become linked to the similar truncation in repetition probably multiforms of direct speech that they reflects what was happening in these introduce. On the other hand, they belong to performances (SKVR I2 746, 748, 752). There the connecting tissue between semantically are a few additional cases that remain central elements of a poem (cf. Gil’ferding ambiguous where a line or couplet is not 44 1894: 24): they are functional markers in the transcribed in the repetition. A single

performance transcribed by two collectors koivu – oksat – kokko – koprat [‘river – rapids – presents an interesting case that can with skerry/ shoal/island – birch – branches – eagle – confidence also be identified with truncation talons’]) or as few as three (koski – koivu – in repetition. Here, the eight-line variant of kokko [‘rapids – birch – eagle’]). Nonetheless, the multiform reduces in the repetition to variation in this element’s repetition is either tuloo tulińi joki / jovess’ on tulińi luoto the simple omission of later lines in the series [‘came a fiery river / in the river (is) a fiery or, in one case, the already-mentioned skerry’]; the first element of the multiform is expansion of the series from koski – koivu – left incomplete, skipping directly to the hero’s kokko to koski – korko – koivu – kokko. The case response to the ‘death’ (SKVR I2 824–824a, of exapansion, however, is quite possibly counted as a single variant in the dataset). owing to remembering the fourth feature in This example is striking because variation of the process of performance. In the dataset, 19 this type could be in the background of other of the 45 variants with indications that the variants where only a line or two appears in multiform was repeated exhibit line for line the transcription of the repetition (SKVR correspondence;47 12 scattered examples seem 45 48 I2 713, 738, 805; SKVR VII1 773). There are to exhibit truncation in the repetition; while indications that multiforms could be reduced 10 remain ambiguous or exhibit variations of in length in repetition, and even a performer other, inconsistent types.49 ‘Word’ for ‘word’ who seems to have felt that once the repetition – or formula for formula repetition multiform was recognized he could skip on to – seems generally to predominate in social the next. On the other hand, there are a few practice for this multiform, although variants in which the second element of the truncation in repetition seems also to have multiform or part of it is simply different in been commonly practiced as well. the repetition (SKVR I2 726, 729, 767); one in Line-core and couplet-core multiforms which korko [‘shoal’] is added to the seem especially suited to the ‘breathing’ of anaphoric series of the first element in the epic length. They also exhibit truncation repetition (SKVR I2 828; cf. SKVR I2 722– through the omission of ‘words’ or lines and 722a, in which luoto [‘skerry’] is added in a couplets in repetition, yet these variations are later performance46); and one in which the not striking in the corpus at the level of second element of the multiform is only individual performers. ‘Breathing’ of length recorded in the repetition (SKVR VII1 791). in kalevalaic epic seems to operate between These latter examples are here considered to singers and dialects of singing in the be most likely related to memory in the expansion and contraction of these course of performance. In this respect, they multiforms, which tend to crystallize in the are comparable to how Jyrki Malinen opens minds of individual performers, as is reflected his first recorded variant with the wrong in their use in repetitions within and across protagonist (multiform M1) and corrected this performances. The ‘breathing’ of length in in performance (multiforms M3ff.), and also this tradition is more apparent in the inclusion how Vassilei Malinen uses a verb associated and omission of multiforms rather than with other dialects in line 55 and thereafter variation in their length as they are used. For uses the form conventional to his kin group example, the hero’s dialogue with his mother (ll. 65, 77, and 83). However, discussion of is a stable episode in The Song of this type of variation will be left aside for Lemminkäinen. Within this dialogue, the more developed treatment elsewhere. couplet-core multiform of the war-shirt Flexibility is clearly present in use of the demand may be used only once to open the fiery-eagle multiform, but that flexibility is dialogue; it may be used in the opening of the concentrated in the second element, often by dialogue and at its conclusion, as the simple omission. The first element has woman’s compliance with the request; it may potential for flexibility: in social practice, the be repeated with every cycle of the hero’s series of elements may be as many as six of demand and the mother’s warning; or it may the seven used (joki – koski – luoto/korko/soari – be omitted entirely.

(21) The seven repetitions of Onuhrie Lesonen’s use of the the war-shirt multiform in his 1877 performance of The Song of Lemminkäinen. Numbers in the table refer to lines of the quotative frame and multiform according to the lineation in SKVR I2 791a; lines in the center of the table indicate uses in dialogue while numbers in the right column indicate use in narrative action; morphological variations between these types of uses are indicated in italic font. Verse in dialogue Line number Verse in action Line Quotative Sano Ahti Šoarelaińi 49 – – – – – [Not applicable] Vocative Oi emońi vanhempańi 50 76 91 – 123 139 Jo emonsa vanhempanša 144 Core Tuo šie miun sot’išomańi 51 77 92 111* 124 140 Jo tuopi hänen šot’isomansa 145 couplet Kanna vainovoattieńi 52 78 93 112 125 141 Kanto vainovoattiša 146 Character- Pitoloissa piettäväńi 53 79 94 113 126 142 Pitoloissa piettävänšä 147 ization Häissä häilyteltäväńi 54 80 95 114 127 143 Häissä häilyteltävänsä 148 Location Lähen Päivöl’än pitoihe 55 81 96 115 – – ––– – Hyvän joukon juominkihi 56 82 97 116 – – ––– –

Verse in dialogue Line number Verse in action Line Quotative Said Ahti Saarilainen 49 – – – – – [Not applicable] Vocative Oh my mother my elder 50 76 91 – 123 139 Already his mother his elder 144 Core Bring my war-shirt 51 77 92 111* 124 140 Brought indeed his war-shirt 145 couplet Carry my persecution-garb 52 78 93 112 125 141 Carries his persecution-garb 146 Character- My holding-one at the feast 53 79 94 113 126 142 His holding-one at the feast 147 ization Glistening one at the wedding 54 80 95 114 127 143 Glistening one at the wedding 148 Location I leave to Päivölä’s feast of 55 81 96 115 – – ––– – The good group’s drinking-feast 56 82 97 116 – – ––– – * Line 111 presents the variation tuoppas miun sotišomańi; semantically, this variation would be so subtle that it would likely only be perceived as a textural variation in the flow of performance. This variation can be viewed in the context of the omission of the vocative phrase which would otherwise have preceded the line. An illustration of the role of multiforms in his reduction of the seven-line multiform length variation can be taken from Onuhrie remains limited to the omission of one couplet Lesonen’s performances of The Song of in repetitions, and that omission only begins Lemminkäinen. Onuhrie performed the epic in with the fourth repetition, as shown in (21). 1871 (SKVR I2 791), and then again for the Kalevalaic epic does not vary like a long epic same collector in 1877 (SKVR I2 791a). The form, nevertheless multiforms appear to be war-shirt multiform was completely absent crucial to the ways that epics vary in length in from the first variant, where the dialogue was reproduction. organized as the mother’s prohibition, followed by her warnings, each coupled with Overview a multiform of the hero’s response to the Kalevalaic epic exhibits a number of danger. The entire dialogue of this variant is distinguishable formal types of multiforms, 31 lines (ll. 4–34). Onuhrie’s second variant is including formula-systems, crystallized series, in total about twice as long and the dialogue open-slot, couplet-core and collocative system alone is 105 lines (ll. 49–153). The war-shirt multiforms. The variety identified as a multiform is used to open each cycle of formula-system operates within a line and is dialogue and the mother’s warning is characterized by semantic diversity. Other followed by another direct-speech multiform types introduced here tend to operate as multi- in which the hero tells his mother to inform line macro-formulae for particular units of him about a ‘death’ on the road – a multiform content, with the exception of collocative that may also be coupled with a repetition of system multiforms that may be distributed the war-shirt multiform. As a consequence, across a cluster of semantic elements. Both the war-shirt multiform is entextualized six varieties nevertheless remain differentiated times in three cycles of dialogue; this is from formulaic ‘words’ of the registral followed by a seventh expressing the lexicon as indexical systems of those ‘words’ mother’s compliance. Onuhrie tends to omit that are then brought into combination in the the quotative frame of the hero’s speech, but formulation of utterances. This formal

91 difference is associated with differences provides a more solid basis for considering between how multiforms vary in contrast to verbal stability and variation in kalevalaic the types of variation exhibited by formulaic epic poems in a way that OFT could not do ‘words’ of the register. effectively without the concept of multiforms. Formula-system multiforms may appear on The present discussion has not opened the the surface to be fundamentally different from topic of how multiforms are organized into the multi-line multiforms discussed. the performance of a whole epic, but it is However, this is merely a function of how rapidly apparent from the example epic above meter and syntax condition the phenomenon that they are linked as pairs and in groupings at different levels of scope. Formula-system that are interfaced with the units and multiforms are organized within a metrically sequences of the plot. The plot in question is defined environment of a line or couplet. constituted of less than twenty of these units. Meter becomes a constraint within that Whereas multiform theory makes the stability environment. Equivalence sets that participate of units of text more salient, the regular in the multiform are associated with certain organization of these units into an organized metrical positions as opposed to others, and whole also becomes more understandable. the particular metrical positions are This approach thus has potential to offer complementary. These sets make up potential insight into other verbally conservative combinations of ‘words’ to produce a poetries. It is also significant that multiforms metrically well-formed line. When ‘words’ operate on a smaller scope than a theme, are at the level of lines without a formal which invites cautious reassessment of structuring constraint on their organization relationships between the lexical surface of like forming regular stanzas, then texts and narrative themes in earlier OFT corresponding equivalence sets do not evolve: research. the ‘word-lines’ are simply organized through The present review has repeatedly stressed inter-linear syntax. This does not mean that that multiforms crystallize in the idiolect of there are no extremely dynamic multi-line individuals and that their social transmission multiforms in kalevalaic poetry, only that they involves the repetition of this process of have not been identified. On the other hand, crystallization as new singers internalize the many formula-systems are comparable to the tradition. Variation established as redactions quite flexible quotative frames: the on a social level is associated with socially associative network is constituted of semantic accepted innovations. Corresponding variations equivalence sets of inquit formulae of can also be observed at the level of individual particular types (speaking, complaining, etc.), variants and idiolects. Developing an under- while others have stable verbal elements like standing of variation in specific multiforms multi-line multiforms. In any case, whether a constructs a framework against which multiform operates at the level of integers individual variants van be assessed. At the organized within metrical lines or the same time, gaining a more general perspective organization of metrical lines and groups of on how multiforms operate and vary in a lines, conventions of non-variation in this epic tradition develops a frame of reference for tradition incline the units to crystallize in the looking at many types of variation at a formal memory of a performer and become stable in level, whether this is a slip in memory or reproduction. conscious and strategic choice. The analysis of multiforms and their The crystallization of multiforms in operation is not incompatible with OFT. memory should not be considered exclusive Instead, multiform analysis complements it by of their potential for strategic variation. Any turning attention from the formulaic ‘words’ element of such a multiform has the potential of a register to how those ‘words’ crystallize to be treated as a slot-filler exchanging into indexical constellations in the mind of a equivalent elements as in an open-slot performer as a framework for expression. multiform – they can be consciously and Attending to these units with their potential strategically varied no less than a proverb or complexity and capacity for crystallization other fixed idiom. This may be, for example,

92 a six-syllable noun–epithet formula that production and reception of registers of discourse, connects the multiform to a different although it extends to larger structural units of linguistically mediated content called ‘themes’ and character, or a whole couplet or larger ‘story patterns’ (or equivalent scripts) (cf. Foley component. Variation that appears established 1999). Some scholars have treated ‘Oral-Formulaic as a redaction of a multiform can be assumed Theory’ as what is here called ‘Classic OFT’ (i.e. of to have begun at the level of individuals and Lord 1960) and distinguished ‘Oral Theory’ as the small groups as in precisely these types of more refined models of OFT used more recently (e.g. Acker 1998). However, this differentiation is adaptation – exchanging one ‘word’ for not standard; and it is merely a rhetorical distancing another. However, recognizing and under- from earlier forms of OFT and the criticisms that it standing these processes is dependent on received which led it to be refined. developing perspectives on multiforms and 3. This statistic is based on information available at how they operate. www.skvr.fi (accessed 22.4.2016). 4. Good examples of this type of study and its The significance of introducing multiforms evolution include the classic studies by Franssila into discussion is that it creates an (1901), Hautala (1945) and Kuusi (1949); on the intermediate frame of reference between Historical-Geographic Method, see Krohn 1926; OFT’s formulae at the level of lines of verse Frog 2013. and themes at the level of more complex units 5. For example on metrics (Sadeniemi 1951), alliteration (Kuusi 1953), parallelism (Steinitz of narrative content. In verbalization, it seems 1934), and more subtle interlinear features (Kuusi to be precisely at the level of multiforms that 1952). variation is most dynamic. This suggests that 6. As Jukka Saarinen has stressed: “Each genre must the level of multiforms equally presents a be studied separately in detail, trying to extract potentially crucial site for considering how a from its texts (and contexts) its rules for reproduction” (1994: 181). tradition is used by different individuals and 7. The mechanics of multiforms investigated here how such resources might be capitalized on enable meaningful situational variation but they for subtle or explicit meaning-making. also produce incidental and accidental variation. Lauri Honko (1981: 24) distinguished ‘major Frog (mr.frog[at]helsinki.fi), PL 59 (Unioninkatu 38 variation’ as variations that are or become stable A), 00014 University of Helsinki, Finland. between performances from ‘minor variation’ as functional, situation-dependent variations that do Acknowledgements: The work presented here has a not endure beyond a single performance. He long history of development and has benefitted from subscribed to a school of thought that does not comments and discussion with a large number of acknowledge the possibility of incidental variation: individuals, to whom I am grateful. This work has been “Pure mishaps excluded, variation is always completed within the framework of the Academy of meaningful” (Honko 1998: 68). I view language as Finland project “Oral Poetry, Mythic Knowledge, and far too flexible a tool to presume that, in the flow of Vernacular Imagination” of Folklore Studies, performance, all potential alternatives are linked to University of Helsinki. nuances of meaning or aesthetics. Such a view would seem inconsistent with my own experience Notes of language use. 1. OFT is commonly conceived through its formalized 8. In The Singer of Tales, Lord uses ‘multiform’ in presentation in Albert Lord´s Singer of Tales three ways without introducing it or defining it as (1960), a form of OFT which will be distinguished an analytical term. He uses it both a) as an adjective here as ‘Classic OFT’. The explosion of research in and b) as a noun following the dictionary definition this area led OFT research to develop into a field of for that which has multiple forms, but he also uses its own that has advanced considerably across more it c) as a noun to refer to any single instantiation of than a half-century since Lord’s seminal work. On something with multiple forms (“the single the background and development of OFT up to performance is a multiform”: Lord 1960: 133). This around the end of the 1980s, see Foley 1988. third use seems to have been to avoid the term Around 1990, there was a turn from emphasis on ‘variant’ (cf. Harvilahti 1992a: 34). The term formal aspects of the operation of verbal art to ‘variant’ was associated with the comparative meanings. For a discussion introducing these, see studies of the Classic Historical-Geographic e.g. Foley & Ramey 2012. Method, which tended to reduce traditions to ‘texts’ 2. Although the term ‘Oral Theory’ has an economy and treat them in isolation from social and of fewer syllables, I am critical of this performance practices (for discussion, see Frog simplification because the theory is not concerned 2013). From here, ‘multiform’ spread with the with oral discourse generally. I prefer ‘Oral- vocabulary of OFT, although its use generally Formulaic Theory’ because the theory centrally remained as an adjective and common noun. concerns the operation of formulaic units in the

9. Although cf. Drout (2011: 447), who follows Foley 18. The recurrence of each of these units then in taking ‘multiform’ as an analytical term for constitutes a form of parallelism (Frog 2014c) and anything with multiple forms in a tradition and at the verbal level what has been described as quotes the Honkos’ description of a linguistic ‘macro-parallelism’ (Urban 1986). multiform as the definition for this broader concept. 19. The category described as kalevalaic ‘epic’ is not 10. These include Old Norse skaldic or court poetry in entirely consistent in English language scholarship. the dróttkvætt meter (Frog 2009; 2014–2015); This is because ‘epic’ is sometimes used to translate kalevalaic epic (Frog 2010a; 2010b); and Icelandic the Finnish term kertova runous [‘narrative sagnakvæði (Frog 2012). poetry’], which also includes what are essentially 11. Register is a term developed in social linguistics to ballads and fairytales in the Kalevala-meter (e.g. describe socially situated variations in language Kuusi et al. 1977). There was no clear distinction (esp. Halliday 1978). The term has gradually been between ‘god’ and ‘hero’ in this tradition. ‘Epic’ developed and extended to describe any socially here refers to mythological plots, although in some distinguishable variety of behaviour (esp. Agha regions such as in Ingria in the south, the 2007). Register theory became established in mythological plots became secularized to varying research on oral poetries and connected with OFT degrees and their fantastic elements converged with through the work of John Miles Foley (esp. 1995; those of fairytales or were excluded (see e.g. Siikala see also Frog 2015b). 1990; 2002b). Differences in how the poetry 12. Current approaches to linguistic formulae have evolved on a regional basis also impacted been inclined to be inclusive and the boundary conventions of conservatism and variation in between a ‘word’ and a ‘formula’ easily seem to production, also relevant here. blur. This is in part because we grow accustomed to 20. This began as an experiment to illustrate to thinking of language orthographically, so we view medieval philologists aspects of the conservatism of English crosswalk as a compound word, improper kalevalaic epic. It was one tool for discussing as a word with an affix, and lose sight of the variation. On the one hand, it provided an embodied interactive aspects of language in alternative to viewing variation in oral epic poetry discourse where a variety of expressivel and centrally through the lens of the highly variable functional features may become bundled while not South Slavic traditions, which has become a all of these will necessarily be realized in every use primary frame of reference for many medievalists. (e.g. Pawley 2009: 6–7). On the other hand, this frame of reference provides 13. Foley’s stance is based especially on metalinguistic an alternative to many scholars’ ideas of comments and discussions by South Slavic epic ‘memorization’ as a pristinely invariable process of singers, but the same phenomenon is found across reproduction of an ideal text (as though there were traditions of verbal art around the world and a written model performers could consult in the through history, which is unsurprising when the wake of variation). terms in question tend to be derived from verbs for 21. This presentation of transcriptions of oral poems speaking and thus refer to utterance rather than can also be read as a comment on practices of orthography (Frog 2014b: 282–283n.3). editing distinct variants of medieval oral poetry in 14. This can be compared to what Lauri Harvilahti this way, such as the reconstructive representation (1985: 112) describes in bylina-epics as a of the Old Norse poem Vǫluspá in the Neckel & ‘particular cliché’ (erityisklišee) in contrast to a Kuhn edition (1963: 1–16) as opposed to the quite ‘general cliché’ (yleiklišee) as a formula used different approach of Sophus Bugge (1967: 1–42). across different epics and contexts. In the register 22. This line formally corresponds to the boundary of Karelian laments, Eila Stepanova (2015: 260) marker formula in 85 through its use of a relevant describes poetic ‘words’ corresponding to equivalent six-position epithet–name formula. Harvilahti’s particular clichés collectively as a However, the boundary marker is formed with a ‘situation-specific lexicon’ in contrast to a ‘core verb in an open-slot formula indicating the hero’s lexicon’ of ‘words’ that are used across all contexts release; here the epithet–name formula is combined and in which every performer gains competence. with an open-slot identity formula which makes it 15. See e.g. Krohn 1903–1910; 1918; 1924–1928; see the grammatical subject of the following couplet. also Siikala 1986; Harvilahti 1992a; 1992b; The introduction of a line with the grammatical Rahimova 2002. subject is a ‘natural’ variation of a couplet-core 16. Cf. Senni Timonen’s observation concerning Larin multiform (introduced below) such as M12 in Paraske’s learning of kalevalaic songs: “The poems which the central verbal element is a predicate. On were not simply contained in her mind, instead they the formal relationship between this line and line existed within the community, among minds” 85, see the discussion of ‘formula-systems’ below. (2000: 638). 23. The classic study of the meter is Sadeniemi 1951; 17. OFT’s ‘themes’ seem simply to be tradition- in English, see e.g. Kuusi et al. 1977; Leino 1986; dependent ‘motifs’ of the Classic Historical- Frog & Stepanova 2011; on the placement of Geographic Method. The term appears to have been stressed syllables, ‘winnowing’ and syntax, see also chosen to avoid implications of cross-cultural Kuusi 1952; on alliteration, see also Kuusi 1953; comparability of the units addressed that would be Frog & Stepanova 2011; on parallelism, see also carried by the term ‘motif’. Steinitz 1934; Kuusi 1952; Saarinen 2014.

24. There are cases where the final syllable is an 34. Or 13 types accounting for almost 87% of the expletive or vocable (e.g. SKVR I2 845.31–32, dataset if variation in word order is not taken into where variation has shortened the lines of the account. couplet by a syllable) but never a lexeme. Vocables 35. Exceptions: SKVR I2 716, 717, 722a (expanded by used in this way are exceptional in epic. two lines from 722), the second use in 767 (where 25. See e.g. the concentrated discussion in Foley 1996: the earlier last line varies and an additional line 17–19; see also Honko 1998: 100–116; Frog 2016; follows); in Aunus / Olonec Karelian and Border / for a parallel concept developed in connection with Ladoga Karelia a line rauta-nokka/-nenä, kalevalaic poetry, see Kuusi 1975: 59 and cf. rautakynsi [‘iron snout/nose, iron talon/claw’] Harvilahti 1992b; Frog 2010a:198–199. appears as an apparently independent single line 26. Partly because the basic multiform has such a (SKVR II 186, 193a–193b; cf. SKVR I2 718) or used simple structure, it is ambiguous in a few cases as a grammatical subject for a following couplet whether an item in the data should be seen as a which is otherwise used independent of it (SKVR II reflection of the multiform per se or may instead 204, 220a–220b). represent an independently generated expression 36. An exception is SKVR I2 716 and cf. use of vielähä representing the unit of content. še šiunki šüöpi [‘moreover it will eat you’] in SKVR 27. SKVR I1 175.4–5, 178.4–5, 179/179a.3–4, 182.4–5, I2 766.75. 183.3–4, cf. SKVR I1 144.4, 165.26. 37 Saarinen applied this model of hierarchical 28. SKVR VII1 157.5–6, 157a.4–5, 160.5–6 and cf. 8–9, structuring to larger units of narration, but the 161.4–5, 164a.5–6, 164b.5–6 and 8–9, 164c.5–6, principles are equally valid at the level of 165.5–6, 168.5–6 and cf. 11–12, 171, 2–3 and 5–6. multiforms. It was in fact Saarinen who stressed to The multiform in this dialect is interesting because me the importance of considering hierarchies in the it incorporates a line about driving (sleighs) that organization of multiforms already in a seminar in gets inflected in the first person: Lähemmä 2010. ajelemaha [‘Let’s go out driving’]. 38. A variation of the second line as kannas vaimo 29. SKVR IV2 1854.1–6 and 12–13, 1855.1–10, 25–34 voattieni (e.g. SKVR I2 741.31) [‘carry, wife, my and 53–62, 1856.1–6 and cf. 12, 1857.1–6 and 12– clothes’] seems to have resulted as an interpretation 13, 1860.1–10, IV3 4032.1–15, 4146, 4046.1–7 and of the much more common vaino- [‘persecution’], 14–15; cf. IV1 1436.1–4, 1500.1–3, IV2 1859.1–2 which in spoken Karelian language was a form of and 12–13, 2260.16–18, IV3 3005.1, 4020.1–3. vainu [‘scent, trace; gossip’] (KKS, s.v. ‘vaino’, 30. SKVR III1 314.1–5, 596.15–18, 1175.1–6, 1176.1– ‘vainu’). The variation is syntactically interesting 6, 1177.8–11, 1178.1–6, 1179.1–3. 1180.1–6, because it places a vocative in the same line with 1181.1–11 and 25–28, 1182.1–4 and 30–31, the imperative predicate. 1236.1–6, III2 1947.16–19, 1968.1–3, 2296.1–5; cf. 39. The name of the location, normally generalized as the variant from Narvusi III1 76.8–14, which Päivölä [‘Sun-Place’], varies according to the incorporates a couplet conventional to central dialect. Ingria. 40. The variation between kun määt [‘when you go’] 31. Valuable foundational work for such research has and mänet [‘you go’] could be an idiolectal been done by Harvilahti (1992a), which could be discourse marker indicating the beginning of the advanced by bringing potential multiforms into series; variation in the following verb is a bit concentrated focus with particular attention to unusual but it is impossible on the basis of one questions of the interpenetration of multiforms and variant to determine whether it reflects personal the proximate co-occurrence of ‘words’ of the singing style, simply a slip, or is an outcome of the poetic idiom suggestive of being linked in the process of collection. minds of performers. 41. This can be compared to variant SKVR II 209, in 32. It should be noted that, in some cases, the absence which the multiform first appears embedded in this of part of a multiform from a variant may be type of a frame in the dialogue and then is repeated attributable to the collector who, once the on the journey with line-for-line correspondence. multiform was recognized and appeared to be The only variations in the second use are in the without noteworthy variation, could move on to the tense of the verb and one superficial inflectional next multiform. Some of the 19th-century collectors ending of luo-ssa [‘in a skerry’], which is likely heard literally hundreds of variants of the grammatically parallel to other words in the series, same epic and, when transcribing verse by hand, compared to the semantically more natural luo-lla frequently only recorded what was ‘different’ in [‘on a skerry’]. vocabulary or structure, not what they considered 42. This number is based on lines from the multiform ‘the same’ and already documented. itself, which could also be excluded and only the 33. Variation in word order accounts for only 7 of these hero’s response to it presented. types, 5 of which are infrequent, and one of which 43. SKVR I2 812, 826, 830 and 850; in SKVR I2 828, when combined with its variation remains an only the first line is marked in the dialogue context infrequent type (the noun–verb combination is only followed by an indicator of line omission (by the found twice in the data). singer?) and a ten-line variant of the multiform appears in the presentation of the journey; SKVR I2 812 presents Tuleepa tulinen koski / j.n.e. / kokko

[‘Comes indeed a fiery rapids / etc. / eagle’] with Foley, John Miles. 1990 [1993]. Traditional Oral Epic: no indication that the second element (first The Odyssey, Beowulf, and the Serbo-Croation presented with four lines) was performed in the Return Song. Los Angeles: University of California repetition; the transcription of SKVR I2 850 presents Press. j.n.e. [‘etc.’] after the first line in the repetition and Foley, John Miles. 1995. The Singer of Tales in later adds the multiform, possibly copying it Performance. Bloomington: Indiana University Press. directly from the preceding documentation in the Foley, John Miles. 1996. “Guslar and Aoidos: dialogue context. Traditional Register in South Slavic and Homeric 44. In the second element of the multiform in SKVR Epic”. Transactions of the American Philological I2 717, 725, 811, 841; in a dialect lacking the Association 126: 11–41. second element this is also found quite Foley, John Miles. 1999. Homer’s Traditional Art. exceptionally in the reduction of the first element University Park: Pennsylvania University Press. from the series koski–soari–koivu–kokko [‘rapids– Foley, John Miles. 2002. How to Read an Oral Poem. island–birch–eagle’] to simply koski–kokko (SKVR Urbana: University of Illinois Press. VII1 799a), although no such truncation occurred in Foley, John Miles, & Peter Ramey. 2012. “Oral Theory a different variant by the same performer (SKVR and Medieval Studies”. In Medieval Oral VII1 799). Literature. Ed. Karl Reichl. Berlin: de Gruyter. Pp. 45. This is also found in SKVR I2 834, recorded by 71–102. Elias Lönnrot, where it is most likely attributable to Fox, James J. 2014. Explorations in Semantic his piecemeal documentation of the variant. Parallelism. Canberra: Australian University Press. 46. In SKVR I2 744 & 744b, luoto is lost between Franssila, K.A. 1900. Kansanrunouden tutkimuksia I: performances. However, the first of these was Iso tammi liitteineen. Helsinki: Suomalaisen produced in a situation with two informants while Kirjallisuuden Seura. only one of these was present in the second. Some Frog. 2009. “Speech-Acts in Skaldic Verse: Genre, variations between the two transcripts of the poem Formula and Improvisation”. In Versatility in may be connected to the absence of one informant. Versification. Ed. Tonya Kim Dewey & Frog. New 47. SKVR I1 719, 742, 754, 777, 781, 791a, 793a, 821, York: Peter Lang. Pp. 223–246. 840, 845, 847, 855; II 207, 209, 220a–220b; Frog. 2010a. Baldr and Lemminkäinen: Approaching VII1 799, 800; XII1 103. the Evolution of Mythological Narrative through 48. SKVR I1 717, 725, 746, 748, 752, 805 (potentially the Activating Power of Expression. UCL Eprints. equivocal), 811, 824/824a, 841; II 217–217a; London: University College London. VII1 799a. Frog. 2010b. “Multiformit kalevalamittaisessa 49. Variations in lines in the second element: SKVR I1 epiikassa”. In Kalevalamittaisen runon tulkintoja. 726, 729, 767; expansions: SKVR I1 828; VII1 791; Ed. Seppo Knuuttila, Ulla Piela & Lotte Tarkka. potential shortening in repetition: SKVR I1 713, Helsinki: Suomen Kirjallisuuden Seura. Pp. 91–113. 738, 812; potential repetition in full: SKVR I1 826; Frog. 2011a. “Alvíssmál and Orality I”. Arkiv för generally ambiguous: SKVR I1 766b, 830, 834, 850; Nordisk Filologi 126: 17–71. VII1 773. Frog. 2011b. “Multiforms and Meaning: Playing with Referentiality in Kalevalaic Epic”. In Laulu Works Cited kulttuurisena kommunikaationa. Ed. Pekka Huttu- Sources Hiltunen et al. Kuhmo: Juminkeko. Pp. 49–63. Frog. 2012. “On the Case of Vambarljóð I: Comments SKVR = Suomen Kansan Vanhat Runot I–XV. on Formulaicity in the sagnakvæði”. RMN Helsinki: SKS, 1908–1997. Newsletter 5: 22–38. Frog. 2013. “Revisiting the Historical-Geographic Literature Method(s)”. RMN Newsletter 7: 18–34. Acker, Paul. 1998. Revising Oral Theory. New York: Frog. 2014a. “Degrees of Well-Formedness: The Garland. Formula Principle in the Analysis of Oral-Poetic Agha, Asif. 2007. Language and Social Relations. Meters”. RMN Newsletter 8: 68–70. Cambridge: Cambridge University Press. Frog. 2014b. “Oral Poetry as Language Practice: A Arend, Walter. 1933. Die typischen Scenen bei Homer. Perspective on Old Norse dróttkvætt Composition”. Berlin: Weidmannsche Buchhandlung. In Song and Emergent Poetics – Laulu ja runo – Bugge, Sophus (ed.). 1867. Sæmundar Edda hins Песня и видоизменяющаяся поэтика. Ed. P. Huttu- fróða. Christiania. Hiltunen et al. Kuhmo: Juminkeko. Pp. 279–307. Drout, Michael D. 2011. “Variation within Limits: An Frog. 2014c. “Parallelism, Mode, Medium and Orders Evolutionary Approach to the Structure and of Representation”. In Parallelism in Verbal Art Dynamics of the Multiform”. Oral Tradition 26(2): and Performance: Pre-Print Papers of the Seminar- 447–474. Workshop, 26th–27th May 2014. Ed. Frog. Du Bois, John W. 2014. “Towards a Dialogic Syntax”. Folkloristiikan toimite 21. Helsinki: Folklore Cognitive Linguistics 25(3): 359–410. Studies, University of Helsinki. Pp. 185–207. Foley, John Miles. 1988. The Theory of Oral Frog. 2014–2015. “Mythological Names in dróttkvætt Composition: History and Methodology. Formulae I–III”. Studia Metrica et Poetica 1(1): Bloomington: Indiana University Press. 100–139 (I); 1(2): 39–70 (II); 2(1): 7–33 (III).

Frog. 2015a. “Mythology in Cultural Practice: A Papers. Mysore: Central Institute of Indian Methodological Framework for Historical Languages. Vol. II, pp. 207–240. Analysis”. RMN Newsletter 10: 33–57. Honko, Lauri, & Anneli Honko. 1998. “Multiforms in Frog. 2015b. “Registers of Oral Poetry”. In Registers of Epic Composition”. In The Epic. Oral and Written. Communication. Ed. Asif Agha & Frog. Studia Ed. Lauri Honko, Jawaharlal Handoo & John Miles Fennica Linguistica 18. Helsinki: Finnish Literature Foley. Mysore: Central Institute of Indian Society. Pp. 77–104. Languages. Pp. 31–79. Frog. 2016 (in press). “Metrical Entanglement and Kaivola-Bregenhøj, Annikki. 1988. Kertomus ja dróttkvætt Composition – A Pilot Study on Battle- kerronta. Helsinki: Suomalaisen Kirjallisuuden Seura. Kennings”. In Approaches to Nordic and Germanic Kallio, Kati. 2013. Laulamisen tapoja: Esitysareena, Poetry. Ed. Kristján Árnason, Tonya Kim Dewey, rekisteri ja paikallinen laji lansi-inkerilaisessa Stephen Carey, Haukur Þorgeirsson, Ragnar Ingi kalevalamittaisessa runossa. Tampere: Tammiprintti. Aðalsteinsson & Þórhallur Eyþórsson. Reykjavík: Kiparsky, Paul. 1976. “Oral Poetry: Some Linguistic University of Iceland Press. and Typological Considerations”. In Oral Frog & Eila Stepanova. 2011. “Alliteration in (Balto-) Literature and the Formula. Ed. B. A. Stolz & R. S. Finnic Languages”. In Alliteration and Culture. Ed. Shannon III. Ann Arbor: Center for the Jonathan Roper. Houndmills: Palgrave MacMillan. Coördination of Ancient and Modern Studies, Pp. 195–218. University of Michigan. Pp. 73–106. Frye, Donald K. 1968. “Old English Formulaic Themes Krohn, Kaarle. 1903–1910. Kalevalan runojen historia. and Type-Scenes”. Neophilologus 52(1): 48–54. Helsinki: Suomalaisen Kirjallisuuden Seura. Gil’ferding, A.F. 1894. Онежския былины, записанныя Krohn, Kaarle. 1918. Kalevalankysymyksia I–II. Александром Федоровичем Гильфердингом Suomalais-Ugrilaisen Seuran Aikakauskirja 35–36. летом 1871 года.1. Санкт-Петербург. Helsinki: Suomalais-Urgilaisen Seura. Hainsworth, J.B. 1968. The Flexibility of the Homeric Krohn, Kaarle. 1924–1928. Kalevalastudien I–VI. FF Formula. Oxford: Clarendon Press. Communications, 53, 67, 71–72, 75–76. Helsinki: Hakamies, Pekka. 2013. “Issues in Folkloristics”. In Suomalainen Tiedeakatemia. Matti Kamppinen & Pekka Hakamies. The Theory Kuusi, Matti. 1949 Sampo-eepos: Typologinen of Culture of Folklorist Lauri Honko 1932–2002: analyysi. Suomalais-Ugrilaisen Seuran Tutkimuksia The Ecology of Tradition. Lewiston: Mellen Press. 96. Helsinki: Suomalais-Ugrilainen Seura. Pp. 45–62. Kuusi, Matti. 1952. “Kalevalaisen säkeen, säeryhmän Halliday, M.A.K. 1978. Language as Social Semiotic. ja runon painavoituvuudesta”. Virittäjä 56(4): 241– London: Edward Arnold. 261, 351–352. Harvilahti, Lauri. 1985. Bylinat: Venäläistä kertoma- Kuusi, Matti. 1953. “Kalevalaisen runon runoutta. Helsinki: Suomalaisen Kirjallisuuden Seura. alkusointuisuudesta”. Virittaja 57: 198–207. Harvilahti, Lauri. 1992a. Kertovan runon keinot: Kuusi, Matti. 1975. “‘The Bridge and the Church’: An Inkeriläisen runoepiikan tuottamisesta. Helsinki: Anti-Church Legend”. In Finnish Folkloristics 2. Suomalaisen Kirjallisuuden Seura. Ed. Pentti Leino with Annikki Kaivola-Bregenhøj Harvilahti, Lauri. 1992b. “The Production of Finnish & Urpo Vento. Studia Fennica 18. Helsinki: Epic Poetry”. Oral Tradition 7: 87–101. Suomalaisen Kirjallisuuden Seura. Pp. 37–75. Hautala, Jouko. 1945. Lauri Lappalaisen runo: Kuusi, Matti, Keith Bosley & Michael Branch (eds. & Vertaileva kansanrunoudentutkimus. Helsinki: trans.). 1977. Finnish Folk Poetry – Epic: An Suomalaisen Kirjallisuuden Seura. Anthology in Finnish and English. Helsinki: Honko, Lauri. 1981. “Four Forms of Adaptation of Finnish Literature Society. Tradition”. In Adaptation, Change, and Decline in Lamb, William. 2015. “Verbal Formulas in Gaelic Oral Literature. Ed. Lauri Honko & V. Voigt. Traditional Narrative: Some Aspects of Their Form Studia Fennica 26. Helsinki: Suomalaisen and Function”. In Registers of Communication. Ed. Kirjallisuuden Seura. Pp. 19–33. Asif Agha & Frog. Studia Fennica Linguistica 18. Honko, Lauri. 1995. “Multiformit ja pitkän eepoksen Helsinki: Finnish Literature Society. Pp. 225–246. arvoitus”. Sananjalka 37: 117–145. Lönnrot, Elias. 1835. Kalewala, taikka wanhoja Honko, Lauri. 1996. “Epics along the Silk Roads: Karjalan runoja Suomen kansan muinoista ajoista. Mental Text, Performance, and Written Helsinki: Suomalaisen Kirjallisuuden Seura. Codification”. Oral Tradition 11(1): 1–17. Lönnrot, Elias. 1849. Kalevala. Helsinki: Suomalaisen Honko, Lauri. 1998. Textualizing the Siri Epic. Kirjallisuuden Seura. Helsinki: Suomalainen Tiedeakatemia. Lord, Albert Bates. 1960. The Singer of Tales. Honko, Lauri. 2003. The Maiden’s Death Song and Cambridge: Harvard University Press. The Great Wedding: Anne Vabarna’s Oral Twin Lord, Albert B. 1981. “Memory, Fixity, and Genre in Epic Written down by A.O. Väisnen. Helsinki: Oral Traditional Poetry”. In Oral Traditional Suomalainen Tiedeakatemia. Literature: A Festschrift for Albert Bates Lord. Ed. Honko, Lauri, & Anneli Honko. 1995. “Multiforms in John Miles Foley. Columbus: Slavica Publishers. Epic Composition”. In XIthCongress of the Pp. 451–461. International Society for Folk-Narrative Research Lord, Albert Bates. 1995. The Singer Resumes the (ISFNR), January 6–12, 1995, Mysore, India. Tale. Ithaca: Cornell University Press.

Morris, Charles W. 1971 [1938]. “Foundations of a Comparative Research”. Journal of Folklore Theory of Signs”. In Charles W. Morris, Writings Research 23: 187–204. on the General Theory of Signs. Ed. Thomas A. Siikala, Anna-Leena. 1990. “Runolaulun käytäntö ja Sebeok. Approaches to Semiotics 16. The Hauge: runoston kehitys”. In Runo, alue, merkitys: Mouton. Pp. 17–72. Kirjoituksia vanhan kansanrunon alueellisesta Neckel, Gustav, & Hans Kuhn (eds.). 1963. Edda: Die muotoutumisesta. Ed. Pekka Hakamies. Joensuu: Lieder des Codex Regius nebst vewandten Joensuun Yliopiston Monistuskeskus. Pp. 7–28. Denkmälern I: Text. 4th edn. Heidelberg: Carl Siikala, Anna-Leena. 2002. Mythic Images and Winters Universitätsbuchhandlung. Shamanism: A Perspective on Kalevala Poetry. FF Niemi, A.R. (ed.). 1905. Runonkerääjiemme Communications 280. Helsinki: Academia matkakertomuksia 1830-luvulta 1880-luvulle. Scientiarum Fennica. Suomalaisen Kirjallisuuden Seuran Toimituksia Siikala, Anna-Leena. 2002b. “The Singer Ideal and the 109. Helsinki: Suomalaisen Kirjallisuuden Seura. Enrichment of Poetic Culture: Why Did the Parry, Milman. 1971. The Making of Homeric Verse: Ingredients for the Kalevala Come from Viena The Collected Papers of Milman Parry. Ed. Adam Karelia?”. In The Kalevala and the World’s Parry. Oxford: Clarendon Press. Traditional Epics. Ed. Lauri Honko. Studia Fennica Parry, Milman. 1971 [1928]. The Traditional Epithet in Folkloristica 12. Helsinki: Finnish Literature Homer. In Parry 1971: 1–190. Society. Pp. 26–43. Parry, Milman. 1971 [1930]. “Studies in the Epic Siikala, Anna-Leena. 2012. Itämerensuomalisten Technique of Oral Verse-Making, 1: Homer and mytologia. Helsinki: SKS. Homeric Style”. In Parry: 1971: 266–324. Steinitz, Wolfgang. 1934. Der Parallelismus in der Pawley, Andrew. 2009. “Grammarians’ Languages finnisch-karelischen Volksdichtung untersucht an versus Humanists’ Languages and the Place of den Liedern des karelischen Sangers Arhippa Speech Act Formulas in Models of Linguistic Perttunen. Helsinki: Suomalainen Tiedeakatemia. Competence”. In Formulaic Language I–II. Ed. Stepanova, Eila. 2015. “The Register of Karelian Corrigan et al. Typological Studies in Language Lamenters”. In Registers of Communication. Ed. 82–83. Amsterdam: John Benjamins. Vol. I, pp. 3–26. Asif Agha & Frog. Studia Fennica Linguistica 18. Rahimova, Elina. 2002. “Variation of the Crystallized Helsinki: Suomalaisen Kirjallisuuden Seura. Pp. Imagery in the Oral Poems in the Kalevala metre”. 258–274. In The Kalevala and the World’s Traditional Epics. Tarkka, Lotte. 2005. Rajarahvaan laulu: Tutkimus Ed. Lauri Honko. Studia Fennica Folkloristica 12. Vuokkiniemen kalevalamittaisesta runokulttuurista Pp. 388–406. 1821–1921. Helsinki: Suomalaisen Kirjallisuuden Roper, Jonathan 2012. Synonymy and Rank in Seura. Alliterative Poetry. Sign Systems Studies 40(1/2): Tarkka, Lotte. 2013. Songs of the Border People. 82–93. Helsinki: Academia Scientiarum Fennica. Saarinen, Jukka. 1994. “The Päivölä Song of Miihkali Timonen, Senni. 2000. “Thick Corpus and a Singer’s Perttunen”. In Songs Beyond the Kalevala: Poetics”. In Thick Corpus, Organic Variation and Transformations of Oral Poetry. Ed. Anna-Leena Textuality in Oral Tradition. Ed. Lauri Honko. Siikala & Sinikka Vakimo. Studia Fennica Studia Fennica Folkloristica 7. Helsinki: Finnish Folkloristica 2. Helsinki: Suomalaisen Literature Society. Pp. 627–659. Kirjallisuuden Seura. Pp. 180–196. Timonen, Senni. 2004. Minä, tila, tunne: Näkökulmia Saarinen, Jukka. 2014. “‘Said a Word, Uttered Thus’: kalevavamittaiseen kansanlyriikaan. Structures and Functions of Parallelism in Arhippa Helsinki: Suomalaisen Kirjallisuuden Seura. Perttunen’s Poems”. In Parallelism in Verbal Art Urban, Greg. 1986. “The Semiotic Functions of Macro- and Performance. Ed. Frog. Folkloristiikan toimite Parallelism in the Sholkeng Origin Myth”. In 21. Helsinki: Folklore Studies, University of Native South American Discourse. Ed. Joel Sherzer Helsinki. Pp. 106–119. & Greg Urban. Berlin: Mouton de Gruyter. Pp. 15– Sadeniemi, Matti. 1951. Die Metrik des Kalevala- 57. Verses. FF Communications 139. Helsinki: Virtanen, Leea. 1987. Suomalaisen ja virolaisen Suomalainen Tiedeakatemia. kansanrunouden suhteista. In Viron veräjät: Sarv, Mari. 2015. “Regional Variation in Folkloric Näkökulmia folkloreen. Ed. Leea Virtanen. Meter: The Case of Estonian Runosong”. RMN Helsinki: Suomalaisen Kirjallisuuden Seura. Pp. Newsletter 9: 6–17. 14–34. Saussure, Ferdinand de 1916 [1967]. Cours de Wray, Alison. 2002. Formulaic Language and the linguistique générale. Paris: Éditions Payot & Lexicon. Cambridge: Cambridge University Press. Rivages. Wray, Alison. 2008. Formulaic Language: Pushing the Siikala, Anna-Leena. 1984. Tarina ja tulkinta. Boundaries. Oxford: Oxford University Press. Helsinki: Suomalaisen Kirjallisuuden Seura. Zumthor, Paul. 1990 [1983]. Oral Poetry: An Siikala, Anna-Leena. 1986. “Variation in the Introduction. Trans. Kathryn Murphy-Judy. Theory Incantation and Mythical Thinking: The Scope of and History of Literature 70. Minneapolis: University of Minnesota Press.

Narratological, Metrical, and Syntactic Emphasis in the Old English Genesis A Ilona Paulis, Radboud University Nijmegen

This paper examines the poetic tools available for an Anglo-Saxon poet to bring narrative details into focus. On the basis of Genesis A, the paper aims to show that clustering of certain metrical types of half-lines, accompanied by use of certain syntactic patterns at key narrative points, creates an enhanced focusing effect. What means did Anglo-Saxon poets employ climaxes make Genesis A most suitable for a to bring their narrative structure into relief? close reading of how the poet has shaped Questions such as this have sparked various integrated emphasis in these episodes. approaches to investigating compositional techniques. Many philological approaches Emphasis thus far have zoomed in on one particular My study presents an integrated analysis of phenomenon or effect. I would like to present a narratological emphasis, metrical emphasis new integrated analysis of three complementary and syntactic emphasis. I retrieved the data on linguistic strategies in Old English narrative narratological emphasis first for reasons of verse which have so far mostly been objectivity. If one starts to analyse poetry considered separately: narratological emphasis, metrically first, one might be tempted to read metrical emphasis and syntactic emphasis. narratological significance into passages that Poetic emphasis can be achieved through are emphasised metrically. The calculations various linguistic strategies, which can be for first metrical and then syntactic emphasis analysed through both separate and integrated were also automated where possible. analyses. One advantage of analysing each Subsequent interpretation was done jointly to strategy for emphasis separately would be that view the interplay between the three strategies phenomena such as narrative progression or for emphasis. metrical patterns are then perhaps more easily Narratological analysis offers a range of measured and quantified. Integrated analysis, approaches, from investigating large-scale however, can reveal the synergy between narrative structures to the minute level of various strategies for emphasis. That synergy creative wording in, for instance, speeches. may in turn reveal unforeseen patterns that Old English narrative poetry is often oral- invite new understanding and interpretation of traditional or at least oral-residual, meaning certain passages in the poetry. The purpose of its written works retain characteristics of the this study, then, is to explore how an oral style. In such poetry we find episodes integrative approach to complementary marked off by “a change in location, or time, linguistic strategies for emphasis may provide or participants” (Fabb 1997: 166), and Nigel new readings in Old English poetry. I will Fabb observes that these changes can be discuss a number of passages from one of the marked by specific linguistic form (1997: religious poems in Oxford, Bodleian Library, 193). Episode boundaries themselves carry MS Junius 11: Genesis A. emphasis by providing turns in the plot, and Genesis as a whole is a narrative based on the specific linguistic form Fabb mentions can selections from the first 22 chapters of the prove fertile ground for other linguistic biblical Genesis (commonly called Genesis A) strategies for emphasis. To avoid subjective and a translation of the Old Saxon Genesis partitioning, I have used the division of major (commonly called Genesis B). Genesis A is a and minor episode boundaries in Genesis A work in epic style containing many iconic from Paul Remley’s (1996) analysis of story arcs such as Lucifer’s fall from Heaven, narrative structure in Genesis. Further Adam and Eve’s expulsion from Paradise, narratological emphasis can be found in Noah’s ark and Abraham’s travails. These speeches, which abound in texts of the Junius story arcs are in turn broken down into 11 manuscript. The speeches in Genesis A are sequences of episodes with their own mostly direct speech and easily detected emphasis. The narrative progression and through verbs that indicate speaking, such as

99 cweðan [‘to speak’], maðelian [‘to speak’] or combined thus make up roughly a quarter of andswarian [‘to answer’]. Since episode verses on average. The aforementioned boundaries and speeches might have their average that about one in four verses would own patterns of emphasis, both are have adjacent stresses might seem to suggest incorporated in the narratological share of the that they would tend to occur about three integrated analysis. verses apart. If instead such spondee-like Metrical emphasis may be found in the verses recur in direct sequence, or with one or aural effect certain verse patterns have. Old two verses in-between, then the repetition of English metre is a strong-stress metre in adjacent stresses would arguably take effect which the poetic line is divided into two half- as metrical emphasis. lines called verses. Most Old English metrical For this study, I have set the requirement patterns have alternating stressed and that verses with adjacent stresses are unstressed syllables, but three are spondee- considered part of a cluster of metrical like in the sense that they contain adjacent emphasis if they occur no more than two main stresses: verse types C, D and F. These verses removed from any other verse with types come in various forms but each contain adjacent stresses. After all, a lone verse with adjacent main stresses ( / / or / \ ): adjacent stresses does not produce as strong C: x x / / x an aural effect as a repetition of adjacent Ūs is riht micel (GenA 1a ) stresses. The metrical emphasis for the it behoves us a lot integrated analysis is thus found in clusters of verse types C, D and F. The metrical data for x x / \ x this study was retrieved through the scansion oð þæt drihtweras (GenA 1798a) so that the retainers program under development by my Radboud University colleague Monique Tangelder. Her D: / \ x x program scans Old English poetry by Gāstcyninge (GenA 2884a) assigning natural linguistic stress to the most the King of the Spirit relevant heavy mora. The scansion of Genesis / / \ x A was further processed with Excel to indicate nacod nīedw dla (GenA 929a) clusters of metrical emphasis for later naked beggar comparison with narratological and syntactic emphasis. / / x \ wer wintrum geong (GenA 2889a) Syntactic emphasis can commonly be a man young in winters found around certain constructions with the discourse marker þā. Not all instances of þā F: x x / / function as discourse markers; apart from on geweald dōn (GenA 1789b) meaning ‘then’, þā can also mean ‘when’, to give [you] as property ‘while’ or ‘because’, and it can be used as a In her article “Stress Felt, Stroke Dealt”, demonstrative pronoun. My study describes Catherine Addison argues that only those cases of þā in the sense of ‘then’. This sense of þā is “in a natural class with nu spondees, in a line of poetry, crowd prosodic ‘now’, eac ‘also’, la ‘lo’” and others, and they stresses together, contradicting the impulse toward fluency. Their turgid effect is often are “often used as rhetorical devices” (van indeed stressful, projecting and expressing Kemenade & Los 2006: 233). The rhetorical emotional stress. (Addison 2005: 161.) device þā occurs in various constructions that can have different functions. Clause-initial þā The Old English verses with adjacent stresses Vf (i.e. þā followed by a finite verb), arguably have this same aural effect, generally marks “the division of narrative especially when such verses occur close to discourse into narrative units” (Enkvist 1986: one another. In most Old English poetry, 301) and: types C and D combined total about 23% or 24% of verses (Green 1971: 92; Bliss 1967: continuation of (or return to) the main-line at 122–123). Type F makes up about 1% of different levels: the level of the temporal verses in Junius 11. Types C, D and F sequentiality of the story line (next, after

100

that), the level of participant continuity the instances of emphasis around Remley’s (return to main participant), and the level of episode boundary on line 1051b where the text structuring (next episode or next lineage of Cain begins. The þā constructions substory). (Wårvik 2011.) that induce syntactic emphasis are in italic Clause-initial þā Vf does not induce syntactic font, and the metrical emphasis is underlined. emphasis, then, but clause-initial þā on its own does. Both clause-initial þā without Passage 1. Genesis A 1043b–1054. verbs and clause-internal þā provide syntactic … / Hine Waldend on, emphasis because they induce focus on the tīrfæst Metod, / tācen sette, words directly following the þā. The þā and its 1045 freoðobēacen, Frēa / þӯ læs hine fēonda adverbial kin “are best analysed as discourse hwilc operators, separating the topic/presuppositional mid gūðþræce / grētan dorste from the focus/new information area of the feorran oððe nēan. / Hēht þā from hweorfan clause” (van Kemenade & Milicev 2005: 5). mēder and māgum / mānscyldigne, Van Kemenade et al. (2008) visualised this cnōsle sīnum. / Him þā Cain gewāt relationship between þā and new information 1050 gongan gēomormōd / Gode of gesyhðe, as follows: winelēas wrecca. / And him þā wīc gecēas ēastlandum on, / ēðelstōwe [previous discourse] → [utterance presupposition þā/þonne focus] fædergeardum feor / þǣr him frēolēcu mǣg, ides æfter æðelum / eaforan fēdde. The right-hand focus area for new information is a site for poets to induce syntactic … / On him God, the Lord of glory, / set a sign, emphasis. The þā constructions preceding this 1045 a peace mark, the Lord, / lest him any foe focus area can take several forms: with warlike intent / dared greet, þā-initial (GenA 2880): far or near. / He then bade him go forth Þā Abraham spræc tō his ombihtum from mother and kin, / guilty of woe, then Abraham said to his servants from his tribe. / Then Cain left, 1050 moved sad of heart, / out of God’s sight, V þā (GenA 2897): a friendless wretch. / And for himself Gestāh þā stīðhȳdig stēape dūne he then chose a place he climbed then resolutely the steep hill in the eastern lands, / a homeland V Pro þā (GenA 965): far from his father’s lands / where a fair Ongunnon hīe þā be Godes hǣse lady, they started then, at God’s behest a woman after his kind / gave him heirs. Pro þā (GenA 941): Three syntactic units in this sample start Hīe þā wuldres Weard wǣdum gyrede neatly with a þā construction, but the cluster then Lord of Glory covered them with clothes of metrical emphasis is not confined within him þā (GenA 882): just one unit. Indeed, metrical emphasis can Him þā Ādām eft answarode apparently string such units together. There then Adam replied to him again are two different þā constructions in this passage, and the episode boundary does not In Genesis A there are notable cases of clause- necessarily contain a þā-initial construction internal þā following pronouns and verbs. The (i.e. þā is clause-initial). Instead, each þā construction him þā appears to have a special induces focus on the theme of the passage position in the poem due to its frequent displayed in the narratological emphasis of appearance around speeches. the episode boundary: Cain’s departure and Syntactic emphasis can interact with the resettlement. The instance of him þā in 1051b narratological and metrical emphasis is followed by the syntactic emphasis for the described earlier. When various strategies for resettlement in wīc gecēas where the shift in emphasis coincide, the emotive expression of location already generates narratological each creates integrated emphasis overall. Let emphasis on the episode boundary. The us review a first sample of emphasis in sample above is one of several highlights in Genesis A 1043b–1054. This example shows

101 the narrative where strategies for emphasis promised land. coincide. Passage 2 contains four clusters of metrical emphasis, of which two precede the episode Results boundary and two bracket the first clause of I will discuss in further detail three examples the new episode. The passage is not enriched of integrated emphasis in Genesis A. Each with metrical emphasis alone, but rather with sample contains one of the dramatic climaxes a combination of strategies for emphasis. in the biblical narrative: the Fall of Adam and There are six instances of syntactic emphasis Eve, Abraham settling in Canaan and in this passage. The three instances of clause- Abraham’s sacrifice of Isaac. I will first initial þā in 1784b, 1790b and 1791b are discuss the Abraham narrative since there are followed by key words describing the two fitting samples from this same story arc. exchange between Abraham and God. The Lines 1767–1804 of Genesis A contain an last two of these instances coincide with a episode of key importance to the Abraham cluster of metrical emphasis spanning God’s narrative. It is here that Abraham travels to speech and the next two clauses, creating a and settles in Canaan. Remley’s episode further link between Abraham and his Lord. boundary falls on line 1805, the line at which The other three instances of þā in lines 1793b, Abraham begins a new part of his life. 1803b and 1805a follow pronouns or proper Passage 2 presents the episode boundary’s nouns referring to Abraham. These three environment in which three linguistic cases of þā also precede a focus part of the strategies of emphasis coincide: lines 1784b– clause with new information containing the 1810. It is here that God reveals Himself to essence of the passage. In 1793b, the new Abraham to give His faithful servant a information following the þā is that Abraham

Passage 2. Genesis A 1784b–1810. … / Þā hine Cyning engla …/ Then to him the King of angels, 1785 Abrahame / īewde Selfa, 1785 to Abraham, / revealed Himself, dōmfæst Wereda / and Drihten cwæð: the just Lord, / and God said: “Þis is sēo eorðe / þe Ic ælgrēne “This is the earth, / all green, that I tūdre þīnum, / torhte, wille, to your progeny, / beautiful, wæstmum gewlō, / on geweald dōn, rich in crops, / will give to rule, 1790 rūme rīce.” / Þā sē rinc Gode 1790 this roomy land.” / Then that man to God wībed worhte / and þā Waldende wrought an altar / and then to the Lord, līfes Lēohtfruman, / lāc onsægde, life’s Light source, / a sacrifice brought, gāsta Helme. / Him þā gyt gewāt to the Protector of souls. / He then again left, Abraham ēastan / ēagum wlitan Abraham, eastward, / his eyes to lay 1795 on landa cyst. / Lisse gemunde 1795 on the choicest of lands / mindful of the promise, Heofonweardes gehāt / þā him þurh hālig the Heavenly Warden’s / that to him through His word holy word, sigora self Cyning / sōð gecӯðde the Lord of victory Himself, / truly had said – oð þæt drihtweras / duguþum gefōran until the lord’s men / with the retainers arrived þǣr is botlwela / Bethlem hāten. to where the village / is called Bethel. 1800 Beorn blīðemōd / and his brōðor sunu 1800 The man blithe of heart / and his brother’s son [Lot] forð oferfōran / folcmǣro land forth travelled / through famed lands ēastan mid ǣhtum, / ǣfæste men, to the east with their belongings, / men firm of faith, weallstēapan hleoðu / and him þā wīc over wall-steep hills, / and for themselves then a curon spot chose þǣr him wlitebeorhte / wongas geþūhton. where to them fair / the fields seemed. 1805 Abraham þā / ōðere sīðe 1805 Abraham then / a second time wībed worhte. / Hē þǣr wordum God wrought an altar. / He then with his words to God torhtum cīgde, / tiber onsægde fairly called, / offered sacrifice his Līffrēan. / Him þæs lēan āgeaf to his Life-lord. / For that God gave him a reward, nalles hnēawlīce / þurh His hand metend not at all stingily, / by His hand meted out 1810 on þām glēdstyde / gumcystum til. 1810 on that altar, / a gracious boon.

102 heads east, which is reflected in the cluster of emphasised verses in various ways to suit the metrical emphasis in 1800a–1801a where surrounding passage. Abraham’s departure is reiterated. The new So far we have seen examples of Cain and information in 1803b, which heads another Abraham’s departures and arrivals in lines cluster of metrical emphasis, is that Abraham 1043b–1054 and 1784–1810. These examples chooses a dwelling-place. The new essential contain integrated emphasis around episode information mentioned at the episode boundaries. In both cases, the syntactic boundary in 1805b is that Abraham raises a emphasis in the focus area of the clause relays second altar and this information returns at concepts that are key to the passage. Metrical 1810a in the cluster of metrical emphasis. The emphasis is not confined to one syntactic unit three instances of syntactic emphasis closest but can bridge these units. Syntactic and to the episode boundary in 1805b are thus metrical emphasis also coincide in various connected to metrical emphasis. Moreover, instances, such as 1047b, 1790–1791 and the instances of him þā connect 1803b. A further observation is that metrical narratologically across the episode boundary emphasis can form a climax just before the through the emphasis on the altar. In the him transition to the next episode. The Genesis A þā instance in 1793b, Abraham leaves after poet thus applies all three linguistic strategies building the altar in 1791a; and in the him þā for emphasis around the showcased episode instance in 1803b, Abraham chooses a boundaries. There are further patterns to be dwelling and subsequently builds an altar found in sequences of speeches, which will be ōðere sīðe [‘a second time’] in 1806a. These the focus of the next two samples. events span the episode boundary, connecting Passage 3 contains lines 2877b–2908a, a the episodes through narratological and plot highlight just before the end of Genesis syntactic emphasis combined. Verse 1810a is A. This passage forms the prelude to the epic part of the metrical climax at the moment climax of the narrative poem: the sacrifice of where Abraham receives his rewards from Isaac. There is no episode boundary in this God. The synergy of narratological, metrical passage, but rather a peculiar interplay of and syntactic emphasis generates a narrative narratological emphasis on three speeches, highlight for this Abrahamic episode. metrical emphasis partial to certain lines and The instances of him þā in lines 1784b– syntactic emphasis in various þā constructions. 1810 are unlikely to be poetic uses of This instalment in the Abraham narrative pronouns just to suit the metre. Not only are continues a series of journeys to pay worship there ties between instances of him þā around to God. The speech Abraham initiates in line the episode border in line 1805, in lines 2881 contains the first cluster of metrical 1043b–1054 there are also passages that hark emphasis in this sample. Indeed, the metrical back to Cain’s departure and arrival. There emphasis occurs within and between the are two constructions with him þā that occur speeches sequenced from line 2881 to 2896, in both passages 1043b–1054 and 1784b– and the very first cluster seamlessly spans a 1810. The construction Him þā … gewāt [‘He speech and the three syntactic units of plain then... went’] occurs under similar metrical narration until the next speech. The poet has circumstances in lines 1049b and 1793b, but thus linked the speeches together through the poet’s inflection of the chosen verb metrical emphasis on the plain narration in cēosan [‘choose’] in lines 1051b and 1803b between them. The narrative is sequenced yields different verse patterns. The further with V þā constructions (i.e. where þā constructions of and him þā wīc curon [‘and is immediately preceded by the verb) for themselves then a spot chose’] in line followed by key words. In Gewāt him þā sē 1803b is part of a cluster of metrical æðeling [‘They departed then, the nobleman’] emphasis, whereas and him þā wīc gecēas (GenA 2885a), the þā is followed by a [‘And for himself he then chose a place’] in reminder that Abraham is a prince of the 1051b is not. Poets had a range of verbs to promised land. It is no small matter to ask a choose from and could thus construct nobleman to sacrifice his only heir. This is reflected in the metrical emphasis in lines

103

Passage 3. Genesis A 2877b–2908a. …/ Þā sē ēadega wer …/ Then that blessed man geseah hlifigan / hēa dūne saw towering / the high hills swā him sægde ǣr / swegles Aldor. as ere had told him / the Lord of Heaven. 2880 Ðā Abraham spræc / tō his ombihtum: 2880 Then Abraham spoke / to his entourage: “Rincas mīne / restað incit “My warriors, / rest yourselves hēr on þissum wīcum. / Wīt eft cumað here in this area. / We two will return siððan wīt ǣrende / uncer twēga after we have, / the two of us, Gāstcyninge / āgifen habbað.” to the King of souls / paid worship.” 2885 Gewāt him þā sē æðeling / and his 2885 They departed then, the nobleman / and his own āgen sunu son, tō þæs gemearces / þe him Metod tǣhte to the marked place / that the Lord had shown him, wadan ofer wealdas. / Wudu bær sunu, to go over the wealds. / The son bore the wood, fæder fȳr and sweord. / Ðā þæs the father fire and a sword. / Then he started to ask, fricgean ongan wer wintrum geong / wordum Abraham: the man young of winters, / of Abraham with words: 2890 “Wīt hēr fȳr and sweord, / frēa mīn, habbað. 2890 “We two here fire and sword, / my lord, have. Hwær is þæt tiber / þæt þū torht Gode Where is that sacrifice / fair that you to God tō þām brynegielde / bringan þencest?” through burning offering / think to give?” Abraham maðelode. / Hæfde on ān gehogod Abraham spoke. / He had one thing determined, þæt hē gedǣde swā / hine Drihten hēt: that he should do as / the Lord had bidden him: 2895 “Him þæt Sōðcyning / Sylfa findeð, 2895 “The True King / will find that for Himself, moncynnes Weard, / swā Him gemet þinceð.” the Guardian of mankind, / what He thinks proper.” Gestāh þā stīðhȳdig / stēape dūne He climbed then stout-hearted / the steep hills up mid his eaforan / swā him sē Eca bebēad up with his heir / as the Lord had bidden, þæt hē on hrōfe gestōd / hēan landes so that he stood on the top / of the high land 2900 on þǣre stōwe / þe him sē stranga tō, 2900 on that spot / that the firm one, wǣrfæst Metod, / wordum tǣhte. the faithful Lord, / had shown him with words. Ongan þā ād hladan, / ǣled weccan, He began then to build a pyre, / the fire to wake, and gefeterode / fēt and honda and fettered / the feet and hands bearne sīnum / and þā on bǣl āhōf of his son, / and then hoisted him on the pyre, 2905 Isaac geongne / and þā ǣdre gegrāp 2905 young Isaac, / and then quickly grabbed sweord bē gehiltum. / Wolde his sunu cwellan the sword by the hilt. / He wanted to slay his son folmum sīnum, / fȳre sencan with his own hands, / drench the fire mǣges drēore. / with kinsman’s blood. /

2885–2886, where Abraham’s title and God’s describes Abraham climbing the hills as orders carry the cluster of metrical emphasis. Gestah þā stīðhȳdig stēape dūne [‘He climbed The next two clusters of metrical emphasis then stout-hearted / the steep hills’] and in occur in Isaac and Abraham’s dialogue, where line 2902 Abraham lights a pyre to sacrifice both speeches contain a cluster of emphasis. his son: Ongan þā ād hladan, ǣled weccan The last cluster occurs in lines 2895–2896 on [‘He began then to build a pyre, / the fire to Abraham’s mysterious reply that God will wake,’]. This new information marks take the sacrifice He deems proper. The dramatic steps in the plot. tension that this speech and metrical emphasis The story arc surrounding the figure create prepares us for the climax of the Abraham contains two good examples of episode and poem. After these clusters, the integrated emphasis. The latter example poet begins a different part of the episode shows patterns of integrated emphasis without any metrical emphasis. There are two markedly different from the first. Although further instances of V þā in 2897a and 2902a, both passages contain metrical emphasis and neither is part of a cluster of metrical spanning syntactic and narratological units, emphasis. Lines 2897 and 2902 do, however, the metrical emphasis in the second sample contain words key to the passage. In line appears to be partial to the dramatic speech 2897, the new information following the þā sequence rather than episode boundaries. The

104 second sample also contains V þā constructions emphasis we already hear that God wants to that induce syntactic emphasis through key see His children, whom He had created words that highlight the plot progression. The earlier. This shifts the audience’s attention poet has indeed made an effort to translate the from God to the children, who then become biblical narrative Genesis into the Old English the subject of the next syntactic unit. The epic style with a variety of narrative syntactic emphasis in line 858 is on gangan highlights. gēomermōde [‘go sorrow-hearted’], a smooth Another long story arc in Genesis A is the transition from God’s wishes to Adam and speech-laden Fall of Adam and Eve in lines Eve’s feelings toward God’s presence. This is 852 through 966. Between lines 852 and 960 a suitable opening for an episode that will lie the episodes on God’s return to Paradise contain many speeches back and forth and the expulsion of Adam and Eve from between God, His children and a certain Paradise. The coda in lines 961–966 evildoer. All in all, there are ten passages of appropriately transitions to mankind’s life on direct and indirect speech in lines 852–966. Earth, a topic that forms the new episode Eight of these are strung together with þā starting at line 967. Since the sequence of constructions, notably constructions with a events in lines 852–966 relies heavily on pronoun followed by þā: speeches, one would expect to find unique (i) GenA 865b–866b: patterns of integrated emphasis. …/ Him þā sylfa oncwæð The new episode at line 852 resumes hēan hlēoðrade / hrægles þearfa Genesis A after the interpolation of Genesis B. Genesis A opens with God’s stroll into …/ To Him (God), he (Adam) then said, the poor one spoke, / lacking clothes Paradise. The þā Vf construction in this opening contains neither syntactic nor (ii) GenA 872: metrical emphasis. The first syntactic Him ðā ǣdre God / andswarede emphasis in the new episode is induced in line To him (Adam) then straight away God / 858a by a V Pro þā construction (i.e. the verb answered precedes þā with additional lexical material (iii) GenA 882: between them) and is accompanied by a Him þā Ādām eft / andswarode cluster of metrical emphasis: To Him (God) then Adam back / answered

Passage 4. Genesis A 852–859. (iv) GenA 887: Ðā ðæs Euan frægn / ælmihtig God Þā cōm fēran / Frēa ælmihtig Then of Eve He asked, / God almighty ofer midne dæg, / mǣre Þēoden on neorxnawang / nēode Sīne. (v) GenA 895b–896b: 855 Wolde nēosian / Nergend usser, …/ Him þā frēolecu mǣg, bilwit Fæder, / hwæt His bearn dyde. ides ǣwiscmōd / andswarode Wiste forworhte / þām Hē ǣr wlite sealde. Gewitan him þā gangan / gēomermōde …/ To Him (God) then the lovely lady, under bēamsceade, / blǣde berēafod. the woman shamed at heart, / answered Then came / the Lord almighty (vi) GenA 903–905: during midday, / the famous Ruler, Þā nædran scēop / Nergend usser, into Paradise / by His own desire. Frēa ælmihtig / fāgum wyrme 855 He wanted to find out, / our Saviour, wīde sīðas / and þā worde cwæð merciful Father, / what His children were Then for the serpent shaped / our Saviour, doing. Lord almighty, / for the guilty snake, He knew to be sinful / those that He wide wanderings, / and then spoke with words earlier had given shape. They retreated then / sorrow-hearted (vii) GenA 918: under the tree’s shadow, / of gladness bereft. Ðā tō Euan God / yrringa spræc The sudden shift in subject from God to Then to Eve God / angrily spoke Adam and Eve in line 858 is skilfully woven into the story. In the cluster of metrical

105

(viii) GenA 941–944: emphasis. The addressees are known at this Hīe þā wuldres Weard / wǣdum gyrede, point and no further Pro þās occur before the Scyppend ūser, / hēt heora sceome þeccan indirect speech. The Genesis A poet has thus Frēa frumhrægle. / Hēt hīe from hweorfan clearly indicated the speaker and the listener neorxnawange / on nearore līf. in his sequence of ten speeches within lines Them (Adam and Eve) then the Lord of glory / 852–966. Eight of these speeches are headed outfitted with garments, by a þā construction with, where appropriate, our Creator, / ordered them to cover their shame a pronoun to indicate an already known with the first clothes. / He ordered them to listener. The þā-initial construction could depart apparently provide the audience with ample from Paradise / to a poorer life. guidance as to who would be addressed next. The instances in examples (i), (ii), (iii), (v) The overall metrical circumstances of the and (viii) contain a construction of þā and a Fall of Adam and Eve differ sharply from pronoun, followed by the announcement of a those in the samples from the Abraham new agent. Whereas þā constructions in other narrative. The instances of þā in lines 852– passages around episode boundaries or 966 occur mainly around the ten passages of speeches mostly induce syntactic emphasis on direct and indirect speech. It appears that in key concepts, in these eight examples the new the area where these speeches occur, the information to the right of þā is the name of metrical emphasis is bound to the speeches the person who is about to speak. This that also receive narratological emphasis. appears to be a useful tool for poets to Whereas in previous sample passages the new emphasise the new speaker in a sequence of information after þā might coincide with speeches when the addressee is already metrical emphasis or return in later clusters of known. It should be of little surprise that, metrical emphasis, in lines 852–966 the new where we see both a new addressee and a new information serves as more of a guide to who speaker introduced, we find clause-initial þā. is speaking next. The subsequent speeches For instance, in line 887 the addressee Eve is even seem to have their own pattern of announced first and then the new speaker metrical emphasis. God: Ðā ðæs Euan frægn ælmihtig God Adam’s words to God in lines 867–871 [‘Then to Eve God angrily spoke’]. The and God’s reply in lines 873–881 are highly announcement of a reply from Eve to God in emphasised metrically; in both speeches about line 895b contains another him þā one in three verses is part of a cluster of construction because the addressee is known. metrical emphasis. The third and fourth direct In lines 903–905 God prepares to curse the speeches are Adam’s reply in lines 883–886, serpent, a new addressee. Since the serpent which contains no metrical emphasis, and had not yet been involved in the sequence of God’s questioning of Eve in lines 888–895a, speeches, a pronoun without antecedent which contains little metrical emphasis (two would have been inappropriate. Lines 903 and of fifteen verses emphasised). The next four 918 thus both contain a clause-initial rather speeches form the narratological and than clause-internal þā due to a new apparently also metrical climax of the speech addressee. Within lines 852–966 there are two sequence: Eve confesses her failure and each instances of (in)direct speech, which despite wrongdoer receives their curse from God. In their lack of þā still contain the same pattern lines 897–902, Eve, ides ǣwiscmōd of addressee first and speaker next. Line 925 [‘ashamed woman’], sorrowfully replies how announces the listener first and speaker last: the serpent misled her. Her speech is greatly Ābēad ēac Ādāme ēce Drihten [‘He (God) emphasised; one in three verses is part of a bade also to Adam, eternal God’]. In lines cluster of metrical emphasis. After Eve’s 941–944 the Hīe þā construction is used to confession, God subsequently curses the convey that Adam and Eve receive their first serpent, then Eve, and lastly Adam. The clothes from God. God gives them His orders density of metrical emphasis in God’s curses through the word hēt followed by indirect is perhaps best shown rather than described. speech coinciding with a cluster of metrical

106

Passage 5 is an overview of God’s three direct ordinary, it is curious that all the metrical speeches (distinguished by different font). emphasis in this speech is concentrated in one Each of the three speeches in this sample is long cluster. God’s cursing of the serpent prefaced by an introduction of the addressee does have a lower ratio of emphasis than the first, followed by the speaker. The order of cursing of Eve and Adam: these each have a God’s speeches seems to follow the path of ratio of one in three. Though the metrical the original sin: from the serpent to Eve and emphasis appears more spread out in the latter then to Adam. God’s curse for the serpent two curses, these curses contain a higher rate contains one large cluster of metrical of metrical emphasis overall. A further emphasis that brings the total ratio for that curiosity is the repetition of the metrically speech to one in four verses emphasised. emphasised þū lāðlīce wrōhte onstealdest While the ratio of one in four may seem [‘you have loathsomely brought sin into the

Passage 5. Genesis A 903–938. Þā nædran scēop / Nergend ūsser, Then for the serpent shaped / our Saviour, Frēa ælmihtig, / fāgum wyrme Lord almighty, / for the guilty snake, 905 wīde sīðas / and þā worde cwæð: wide wanderings, / and then spoke with words: “Þū scealt wīdeferhð, / wērig, þīnum brēostum “You shall forever, / fatigued, on your breast, bearme, tredan / brādre eorðan, your bosom, slither / in the broad world, faran fēðelēas / þenden þē feorh wunað, go footless / while in you your soul lives, gāst on innan. / Þū scealt grēot etan the spirit within. / You shall eat dirt 910 þīne līfdagas / swā þū lāðlīce in your life-days / because you have loathsomely wrōhte onstealdest. / Þē þæt wīf fēoð, brought sin into the world. / You the woman shall fight, hātað under heofnum / and þīn hēafod tredeð hate under the heavens, / and tread your head, fāh mid fōtum sīnum. / Þū scealt fiersna sǣtan the cursed thing, with her feet. / You shall waylay the heels tohtan nīwre. / Tuddor bið gemǣne in ever-new battles. / War shall there be 915 incrum orlegnīð / ā þenden standeð between your offspring / and hers as long as there is woruld under wolcnum. / Nū þū wāst and canst, a world under the welkin. / Now you know, lāð lēodsceaða, / þū lifian scealt.” loathed harmer of men, / how you shall live.” Ðā tō Euan / God yrringa spræc: Then to Eve / God angrily spoke: “Wend þē from wynne. / Þū scealt wǣpnedmen “Turn away from joy. / You shall by men 920 wesan on gewealde / mid weres egsan be controlled / with fear of men, hearde genearwad, / hēan þrōwian harshly constrained, / endure pain þīnra dǣda gedwild, / dēaðes bīdan by your own deeds, / await death, and þurh wōp and hēaf / on woruld cennan and through weeping and heaving / bring into the world, þurh sār micel / sunu and dōhtor.” through great pain, / sons and daughters.” 925 Ābēad ēac Ādāme / ēce Drihten, He bade also Adam, / eternal God, līfes Lēohtfruma, / lāð ǣrende: life’s Light source, / grave tidings: “Þū scealt ōðerne / ēðel sēcean, “You shall another / homeland seek, wynlēasran wīc, / and on wræc hweorfan a more joyless place, / and wander in misery nacod nīedwǣdla / neorxnawanges, naked and needy, / of Paradise’s 930 dugeðum bedǣled. / Þē is gedāl witod virtues bereft. / For you is separation ordained līces and sāwle. / Hwæt, þū lāðlīce of body and soul. / Lo, you have loathsomely wrōhte onstealdest. / Forþon þū winnan scealt brought sin into the world. / For that you shall struggle and on eorðan þē / þīne andlifne and on earth / your whole life selfa gerǣcan, / wegan swātighlēor for yourself obtain food, / by the sweat of your brow 935 þīnne hlāf etan / þenden þū hēr lēofast your bread eat / while you here live, oðþæt þē tō heortan / hearde gripeð until you to heart / harshly strikes ādl unlīðe / þe þū on æple ǣr the disease ungentle / that you with the apple ere selfa forswulge. / Forþon þū sweltan scealt.” swallowed yourself, / for which you shall die.”

107 world’] in lines 910b–911a (cursing of the They inhabited then after their sin / a more serpent) and 931b–932a (cursing of Adam). sorrowful land, Briefly but noticeably, the serpent and Adam a home and country/ more barren are similarly addressed. The combination of of any good things / than their original verbal repetition in two curses and different dwelling was that they after their deed / had been driven ratios of metrical emphasis in the three curses from. begs the question of what degrees of blame 965 They started then, / at God’s behest, were assigned to the serpent, Eve and Adam to beget children / as the Lord has bidden. in Genesis A. Moreover, was each of them then punished appropriately? This is just one The first þā occurs in line 961: Gesǣton þā example of the questions an integrated æfter synne sorgfulre land [‘They inhabited analysis of emphasis can trigger. then after their sin a more sorrowful land’]. There is still more content in lines 852–966 This instance has no connection with metrical that deserves a new reading. After the epic emphasis but the constituent following þā climax of the speeches in the Fall of Adam does repeat the key concept that it was sin that and Eve comes the result of those speeches, saw Adam and Eve expelled from Paradise. In namely the expulsion from Paradise. Adam lines 965–966, Adam and Eve’s settlement is and Eve’s departure is described in three complete and the story arc is thus rounded off. phases. First, in lines 939 and 940, the poet Remley’s episode boundary at 967 interjects one of the morals of the story: corresponds with the transition from source Hwæt, wē nū gehӯrað hwær ūs hearmstafas material from Genesis 3 to Genesis 4, where wrāðe onwōcan and woruldyrmðo [‘Lo, we the lineage of Adam is introduced (1996: now have heard from whence our pain came 102). The transition to this next biblical and mortal misery’]. The compounds episode contains a cluster of emphasised hearmstafas [‘harm, trouble, affliction’] and verses bracketing the verse with þā. In this woruldyrmðo [lit. ‘world-misery’] contain small cluster the expulsion and the mention of adjacent stresses and bear the metrical God’s command carry the emphasis. emphasis in the interjection. God next sends Overall, the poet seems to have used the the sinners off in lines 941–944, and this integrated emphasis in lines 852–966 for a passage starts with Hīe þā. From the plural specific purpose. In the Fall of Adam and and context one can infer that Hīe refers to Eve, the many spoken exchanges between the Adam and Eve, but, unlike in the earlier figures seem to require the syntactic emphasis sequence of speeches, there is now no direct to be mainly on the names of new speakers antecedent in God’s speech. There are two whereas metrical emphasis appears largely verses with metrical emphasis in (viii); these reserved for the actual speeches that carry the occur in God’s indirect speech at that moment plot. The narrative climaxes correspond with when He orders Adam and Eve to cover their the increase in metrical emphasis during the shame and leave. Again it seems that in this speeches as well, since Eve’s lament and story arc metrical emphasis is more often God’s three curses are highly emphasised. reserved for speech itself rather than moments The poet then prepares us for a transition to bridging speeches. Once Adam and Eve have the next episode by way of a short summary departed from Paradise, a different þā of key concepts coinciding with syntactic construction returns: the V Pro þā emphasis. The poet’s use of emphasis in lines construction. There are two instances of these 852–966 thus differs markedly from that in in lines 961–966. the Abraham story arc. The integrated analysis of linguistic strategies for emphasis Passage 6. Genesis A 961–966. can help us detect what the poet has Gesǣton þā æfter synne / sorgfulre land, foregrounded in the various passages of his eard and ēðyl, / unspēdigran work. fremena gehwilcre / þonne sē frumstōl wæs þe hīe æfter dǣde / of ādrifen wurdon. Conclusions 965 Ongunnon hīe þā / bē Godes hǣse The Genesis A poet has skilfully interwoven bearn āstrīenan / swā him Metod bebēad. narratological, metrical and syntactic 108 emphasis in a number of his passages. The insight into that particular structure? Further present integrated analysis of emphasis can syntactic emphasis may be found in verb- shed new light on the dynamics of those initial constructions indicative of lively passages. Narratological emphasis lies in the narrative; perhaps these constructions interact dynamism of plot turns at episode boundaries in some way with þā constructions or other and sequences of ceremonial speeches, and strategies for emphasis. The class of rhetorical serves to guide the audience through the devices that includes þā contains other narrative. Metrical emphasis tends to occur on members that could be studied in relation to key concepts in the passages above, but the surrounding narratological or metrical during sequences of speeches the poet lends environment. It seems worthwhile to explore the aural force that metrical emphasis further linguistic strategies that an Anglo- provides to the speakers in the poem. Saxon poet could use to bring his or her Syntactic emphasis induced by þā narrative into relief. There is still much more constructions directs the listeners’ attention to to be discovered within Old English poetry, the new information to the right of þā, and hopefully we can gain fruitful new whether that new information be a key readings of intricate passages through concept or an indication of who is speaking integrated analyses of emphasis. next. The poet has applied all three strategies to make his biblical narrative more lively and Ilona Paulis (i.paulis[at]let.ru.nl) Radboud University Nijmegen, CLS/Dept. of English Language and dramatic. Culture, Postbus 9103, 6500HD Nijmegen, The Studying each strategy for emphasis Netherlands. separately might cause one to overlook the synergy between the strategies. For instance, Works Cited concepts following a clause-internal þā often Sources return in the clusters of metrical emphasis, but Genesis. Ed. Gregory Ray Hidley. Anglo-Saxon Poetic do not always need to do so when the Records. Retrieved from University of Oxford Text constituents following clause-internal þā are Archive: http://ota.ox.ac.uk/desc/3009 (last accessed 10 September 2014). [Text adapted for used to emphasise the new speaker. In such automated scansion programme.] cases, metrical emphasis is often reserved for the speeches themselves. In the expulsion Literature from Paradise sequence in Genesis A lines Addison, Catherine. 2005. “Stress Felt, Stroke Dealt”. 852–966, main episodes are indicated with the Style 39: 153–174. undramatic clause-initial þā Vf. Within Bliss, Alan J. 1967. The Metre of Beowulf. Oxford: Blackwell Publishing. episodes, him þā apparently serves to Enkvist, Nils Erik. 1986. “More about the Textual emphasise whom the next speaker to be Functions of the Old English Adverbial þa”. In announced is addressing, whereas the content Linguistics across Historical and Geographical of the new speaker’s words carry the metrical Boundaries: In Honour of Jacek Fisiak I: Linguistic emphasis. Theory. Ed. Dieter Kastovsky & Aleksander Szwedek. Berlin: Mouton de Gruyter. Pp. 301–310. The current observations beg the question Fabb, Nigel. 1997. Linguistics and Literature. Oxford: of to what extent similar and other patterns of Blackwell. emphasis could be found elsewhere in Green, Donald C. 1971. “Formulas and Syntax in Old Genesis A and in other poems. I myself will English Poetry: A Computer Study”. Computers turn the integrated analysis to Beowulf next to and the Humanities 6: 85–93. van Kemenade, Ans, & Bettelou Los. 2006. “Discourse explore to what extent that particular work Adverbs and Clausal Syntax in Old and Middle contains patterns of integrated emphasis. For English”. In The Handbook of the History of instance, to what extent could metrical English. Ed. Ans van Kemenade & Bettelou Los. emphasis be present in major speaking roles Maldon: Blackwell Publishing. Pp. 224–248. versus minor speaking roles, and under what van Kemenade, Ans, & Tanja Milicev. 2005. “Syntax and Discourse in Old and Middle English Word (syntactic) circumstances? Could metrical Order”. In Articles from the 8th Diachronic emphasis appear in passages with Generative Syntax Conference. Ed. S. Anderson & narratological emphasis such as ring structure D. Jonas. Oxford: Oxford University Press. and if so, to what extent would it provide new van Kemenade, Ans, Tanja Milicev & R. Harald Baayen. 2008. “The Balance between Syntax and

109

Discourse in Old English”. In English Historical Wårvik, Brita. 2011. “Connective or ‘Disconnective’ Linguistics 2006: Selected Papers from the Discourse Marker? – Old English þa, Fourteenth International Conference on English Multifunctionality and Narrative Structuring”. In: Historical Linguistics. Ed. Maurizio Gotti, Marina Connectives in Synchrony and Diachrony in Dossena & Richard Dury. Amsterdam / European Languages (Varieng 8). Ed. Anneli Philadelphia: John Benjamins. Meurman-Solin & Ursula Lenker. Available at: Remley, Paul G. 1996. Old English Biblical Verse: www.helsinki.fi/varieng/journal/volumes/08/warvik Studies in Genesis, Exodus and Daniel. Cambridge: (last accessed 9.4.2015). Cambridge University Press.

110

REVIEW ARTICLES AND RESEARCH REPORTS

Word Constellations as Tools in Skaldic Composition: Arnórr jarlaskáld and Óttarr svarti as a Case of Assimilation Cole Nyquist, University of Oslo

The purpose of this article is to examine article presents a case study illustrating the connections between the works of Arnórr strategy of a poet adapting metrically relevant Þórðarson jarlaskáld [‘skald of jarls’] and constellations of words or phraseology from Óttarr svarti [‘the black’], specifically within the works of another poet. This specific case the hending and alliterative word constel- study of these two poets has been selected lations which appear in the works of each from the pilot study of my master’s thesis, in skald, with the aim to shed more light on the which I surveyed such word constellations in compositional techniques in the dróttkvætt a sample corpus of 1,486 couplets by 68 meter and related meters. The metrical different poets. The following discussion complexity of this poetry was recently should be viewed on the backdrop of that described as “exceptional among the oral pilot study. Generalizations and inferences poetries of Europe” (Frog 2014: 283, Kuhn related to that context are, of course, 1983: 15). Verses are composed in six- dependent on the degree to which the sample position lines, in which each position equates corpus is generally representative. more or less to one syllable (with some rules Rather than focusing on the poetic for variation). Two stressed syllables (but not equivalence vocabulary called heiti or the any following unstressed syllables) in each prominent type of nominal circumlocution line should rhyme, and in each couplet (i.e. a called a kenning as tools for oral composition, pair of odd line and even line, sometimes – I chose instead to analyze what is called the rarely – called fjórðungr) two stressed hending1, or the internal rhyme characteristic syllables in the first line should alliterate with of the skaldic meters called dróttkvætt, as well the first (stressed) syllable in the following as the constellations of words that are used to line, and the meter also had additional rules meet metrical requirements of alliteration regarding syllabic quantity (Gade 1995: 3–7). (Gurevich 1994; Frog 2009). Hending word- If only alliteration and rhyme are considered, pairs include both odd line skothending, i.e. this means that there are 14 phonic stressed syllables rhymed with same final requirements for the 24 metrical positions of a consonant(s) but different vowels, and even four-line helming or half-stanza. These line aðalhending, which are full rhymes with requirements form a baseline for poets to the same final consonant(s) and vowels. develop new experimental meters, such as These word constellations were then hrynhent [‘flowing rhyme’], in which lines compared between the poetry of different were extended from six to eight syllables. skalds in order to find instances where either Skilled poets must have had techniques to the alliterative word combinations, hending meet these requirements. word combinations, or both recurred in the During my master’s research at the works of different poets. The works of these University of Oslo, under the guidance of my two 11th-century skalds contain clear supervisor Mikael Males, I developed a new correspondences that appeared particularly approach for studying oral composition of pronounced within the broader survey of my skaldic poetry (Nyquist 2014). The present master’s thesis (Nyquist 2014). The examples

111

in this article are gathered from skaldic poetic Ransom’], composed ca. 1023 (Skj B1: 268– editions of Finnur Jónsson’s Den Norsk- 272; SPSMA I: 755). It may thus be stated in Islandske Skjaldedigtning (Skj), and the more advance that, if we find evidence of direct recent Poetry from the Kings’ Sagas (SPSMA influence, the non-specific dating of Arnórr’s I–II), edited by Diana Whaley and Kari Ellen poetry makes it more difficult to determine Gade. Citations are indicated by stanza and which poet has a greater chance of being line number; where the numbering of stanzas influenced by the other, but as Hǫfuðlausn is differs between editions, the numbering of from ca. 1023 and Arnórr composed at least Finnur Jónsson’s edition will appear first and one poem in the mid to late 11th century, it is the numbering of the more recent edition will perhaps more likely that Arnórr was appear following a slash (e.g. Hǫfuðlausn influenced by Óttarr. 14/15). There are some discrepancies between In order to present the hending and Finnur’s earlier editions of some of the alliteration data, I have created tables which couplets in this study and the more recent list the relevant words from each example editions: I have noted where the more recent couplet. The title of each table presents titles editions differ, observing that these may be and verses of the poems compared and their more accurate. The goal here is to present authors’ names. From left to right, each table some of the data which I have collected and includes the poet’s name, the stanza number discuss some of the possible explanations for from which the example couplet comes, the correspondence in sets of words between three alliterating words and finally the compositions attributed to Arnórr (Whaley rhyming words from each couplet. The data is 2009) and Óttarr (Townend 2012). displayed side by side for ease of comparison and relevant similarities for each example are Arnórr jarlaskáld and Óttarr svarti presented in boldfaced font. The first case of In this section, I will present nine different similarities is shown in Table 1: it is a instances where a poem by Arnórr Þórðarson comparison between Arnórr’s Hrynhenda, jarlaskáld shares word constellations in Magnúsdrápa 7.1–2, a drápa (i.e. a poem common with with Óttarr svarti’s Hǫfuðlausn with a refrain) about king Magnús inn góði (Nyquist 2014: 34–40). Both Arnórr and Óláfsson, and Óttarr’s Hǫfuðlausn 3.1–2. In th Óttarr were active during the 11 century and this case, both skalds use the same two words, it is therefore possible that each poet was ungr [‘young’] and eyðir [‘destroyer’] in the aware of the other’s works. The particular first line for vowel alliteration. Both examples fragment to be reviewed here comes from in Table 1 are from the first couplet of a Arnórr Þórðarson’s Hrynhenda, which is helming in their respective poems and are named for the poetic meter in which it was similarly addressing a young warrior or king, composed and can only be dated to sometime but share little in common apart from their th in the 11 century (Skj B1: 307; SPSMA I: theme. In this case, the ungr–eyðir 191), and Óttarr’s Hǫfuðlausn [‘Head- combination seems only to be found in these

Table 1. Hrynhenda, Magnúsdrápa 7.1–2 by Arnórr jarlaskáld and Hǫfuðlausn 12.1–2 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld Ungan eyðir ulfa Ungan þrøngva gráðar ráði Óttarr svarti ungr eyða ógnteitr Nǫ ðuð eyða ógnteitr Peitu Source Nyquist 2014: 34. Arnórr jarlaskáld Óttarr svarti Hǫfuðlausn 12.1–4 (Skj B 270). Hrynhenda, Magnúsdrápa 7.1–4 (Skj B1: 307). 1: Ungan frák þik, eyðir, þrøngva, Nôðuð ungr at eyða, ulfa gráðar, þeira ráði; ógnteitr jǫfurr, Peitu; skildir stǫkk með skœðan þokka reynduð, ræsir, steinda skeiðar brands fyr þér ór landi. rǫnd á Túskalandi.

112

Table 2. Þórfinnsdrápa 13/15.3–4 by Arnórr jarlaskáld and Hǫfuðlausn 14/15.1–2 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld eigi ægi ofvægjan eigi ægi ofvægjan bægja Óttarr svarti Eigi ægi ér Eigi ægi fóruð stóran Source Nyquist 2014: 35. Arnórr jarlaskáld Óttarr svarti Þórfinnsdrápa 13/15.1–4 (Skj B1: 318). Hǫfuðlausn 14/15.1–4 (Skj B1: 271). Nemi drótt, hvé sæ sótti Eigi hrædduzk ægi; snarlyndr konungr jarla; ér fóruð sæ stóran; eigi þraut við ægi allvaldr of getr aldar ofvægjan gram bægja. engi nýtri drengi. two poems in the skaldic corpus.2 This first parallel is merely coincidental are reduced. table shows a rather minor similarity between The couplets compared in Table 3 are two couplets which share only the first two found once again in Arnórr’s Hrynhenda words of the vowel alliteration. 12.5–6 and Óttarr’s Hǫfuðlausn 6/7.3–4. The example in Table 2 is between Unlike the previous two comparisons, these Hǫfuðlausn 14/15.1–2 by Óttarr and couplets have no common alliteration, but do Þórfinnsdrápa 13/15.3–4 by Arnórr. These have similar first line skothending based on couplets share a word pair of eigi [‘not’] and the verbs þora [‘to dare’] and varða [‘to ægir [‘sea’] which participates in the first line defend’] to form a rhyme pair that seems not vowel alliterations and the same pair also to have been used outside of these two comprise the first line’s hending word pair. In poems.4 this case, each corresponding word appears in The example in Table 4 includes line the same metrical position and form within similarities between three couplets by Arnórr their respective couplets’ first lines, with eigi and one from Óttarr. The couplets are from as the first word and ægir as the last. This Óttarr’s Hǫfuðlausn 10.1–2 and from alliterative combination can also be found in a Arnórr’s Haraldsdrápa 7/5.1–2 and 12/10.1– few additional instances in the corpus.3 It thus 2 and Magnúsdrápa 16.1–2. The second might also reflect a more general collocation alliterating term found in the first line of all used by poets in composition, or at least a four examples, Yngvi, is a personal name, the word-pair that was readapted by multiple name of god. The recurrence of this name and poets from different circulating poems. etymologically related verb and noun ganga Nevertheless, when lexical correspondence is [‘to walk, go; move’] characterizes these complemented by its completion of both examples. Irrespective of its inflectional form, alliteration and rhyme, the chances that the ganga consistently forms a skothending Table 3. Hrynhenda, Magnúsdrápa 12.5–6 by Arnórr jarlaskáld and Hǫfuðlausn 6/7.3–4 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld hvergi hallir heiðit þorði varða heiðit breiðu Óttarr svarti þorðut þér þjóðlǫnd þorðut varða þjóðlǫnd rǫndu Source Nyquist 2014: 35. Arnórr jarlaskáld Óttarr svarti Hrynhenda, Magnúsdrápa 12.5–8 (Skj B1 309). Hǫfuðlausn 6/7.1–4 (Skj B1 269). Hvergi þorði hallir varða Gildir, komt at gjaldi heiðit folk í virki breiðu; gotneskum her, flotna; buðlungr, unnuð borgarmǫnnum þorðut þér at varða bjǫrtum eldi stalldræp hjǫrtu. þjóðlǫnd firar rǫndu.

113

Table 4. Haraldsdrápa 7/5.1–2 and 12/10.1–2 and Magnúsdrápa 16.1–2 by Arnórr jarlaskáld and Hǫfuðlausn 10.1–2 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld Upp- Upp- ætt yngvi Gengr yngvi brendi Haraldsdr. 7/5.1–2 lendinga lendinga Arnórr jarlaskáld Haraldsdr. 12/10.1–2 Uppgǫngu yngvi ítr Uppgǫngu yngvi ítr lítinn Arnórr jarlaskáld Auð- Auð- Uppgǫngu yngvi Uppgǫngu yngvi gnógan Magnúsdr. 16.1–2 lógandi* lógandi Óttarr svarti Atgǫngu Yngvi ætt Atgǫngu Yngvi siklinga mikla Source Nyquist 2014: 36. * The SPSMA II edition has “arflógandi gnógan”. Arnórr jarlaskáld Arnórr jarlaskáld Haraldsdrápa 7/5.1–4 (Skj B1: 323). Magnúsdrápa 16.1–4 (Skj B1: 315). Gengr í ætt, þats yngvi Uppgǫngu vann yngvi Upplendinga brenndi arflógandi gnóga; (þjóð galt ræsis reiði,) gerði hilmir Hǫrða rǫnn (þess’s fremstr vas manna). hjǫrþey á Skáneyju.

Arnórr jarlaskáld Óttarr svarti Hǫfuðlausn 10.1–4 (Skj 270). Haraldsdrápa 12/10.1–4 (Skj B1: 324). B1: Uppgǫngu bauð yngvi Atgǫngu vannt, yngvi, ítr með helming lítinn, ætt siklinga mikla; sás á sinni ævi blíðr hilmir, rautt breiða sásk aldrigi háska. borg Kantara of morgin. rhyme with Yngvi in the first line. The fact borrowing from his own potentially earlier that Arnórr uses this constellation three times work, it seems that if either poet repurposed and in two different poems makes him seem another’s alliterating or rhyming word pair in especially taken with it, and I have only found these examples, it was more likely done by the combination of Yngvi with a derivative of Arnórr. It is also worth noting that, in the ganga in one other poem, also a examples in Table 4, Arnórr’s verse 7/5 first Magnússdrápa dedicated to king Magnús alliteration ætt [‘direction, family, berfœttr.5 Yngvi occurs in the same metrical generation’] is the same as Óttarr’s verse 10’s positions and in the nominal inflection in all third alliteration. This means that all three four examples. Óttarr’s use with atganga alliterating words of these couplets are shared [‘attack’] also forms a compound in the between the two poets, but not in all three of opening positions of the line structurally and Arnórr’s stanzas and two of his poems. The metrically paralleling Arnórr’s uses of word ætt is in a different position in uppganga [‘advance’]. These features Haraldsdrápa than in Hǫfuðlausn, but both increase the markedness of parallels in this appear alongside yngvi and Haraldsdrápa constellation of lexical elements, increasing also includes a -gǫngu compound alongside the likelihood that there is a relationship yngvi in a later couplet. If this is indeed between these examples. evidence that Arnórr has assimilated these Haraldsdrápa [‘A Poem with a Refrain rhyming and alliterating words from Óttarr’s about King Haraldr harðráði’] is dated to ca. work, consciously or subconsciously, he has 1067 (Skj B1: 322), and thus was composed done so in two poems which are potentially more than four decades after Hǫfuðlausn (ca. separated by decades. 1023). Magnúsdrápa was “presumably Curiously, this particular couplet by Óttarr composed not long after Magnús’s death in shares another such similarity with another 1047” (SPSMA II: 225). Unless Arnórr is couplet of Arnórr’s, Hrynhenda 9.1–2 in the

114

Table 5. Hrynhenda, Magnúsdrápa 9.1–2 by Arnórr jarlaskáld and Hǫfuðlausn 10.1–2 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld Síðan suðr siklingr Síðan suðr/láði siklingr miklum Óttarr svarti Atgǫngu Yngvi ætt Atgǫngu Yngvi siklinga mikla Source: Nyquist 2014: 37.

Arnórr jarlaskáld Óttarr svarti Hrynhenda, Magnúsdrápa 9.1–4 (Skj B1: 308). Hǫfuðlausn 10.1–4 (Skj B1: 270). Síðan vas, þats sunnr með láði Atgǫngu vannt, yngvi, siklingr ýtti flota miklum; ætt siklinga mikla; skíði vas þá skriðar of auðit blíðr hilmir, rautt breiða skorðu; renndi Visundr norðan. borg Kantara of morgin.

7 rhyme-pair siklingr [‘ruler’] and mikill example of all three. The second example [‘great’] shown in Table 5. While this is not shown in Table 7 has fewer words in as extensive a comparison as in the previous common, but the similar words þengill example, the pairing of these words for a [‘prince, king’] and lengi [‘long, for a long rhyme seems exclusive to these two stanzas in time’] participate in the second line the corpus.6 The recurrence of Óttarr’s aðalhending and both have the same form in couplet Hǫfuðlausn 10.1–2 sharing a each poem. This rhyme combination again similarity with one of Arnórr’s poems is also seems to be specific to these two generally worth noting. compositions in the corpus.8 Hǫfuðlausn and Magnúsdrápa also share The example in Table 8 has multiple similar word pairs in the examples in Tables 6 recurring words between Arnórr’s poems and and 7. The first comparison shown here in Óttarr’s. The relevant couplets for this Table 6 reveals that the poets share engr example are from Hǫfuðlausn 9.5–6 and [‘meadow’] and áðr [‘already, before, Arnórr’s Magnúsdrápa 6.7–8 and 9.3–4 previously’] in each couplet’s respective first alongside Þórfinnsdrápa 18/12.1–2. Here, it and third alliteration. It is important to note is the second line aðalhending which is most that both alliterations take part in their similar between these poems in the rhyme respective line’s hending as the first word of combination land [‘land’] and brandr each line’s rhyming pair and that each poem [‘firebrand, ship’s beak, sword blade’]. The shares the same second line, second hending land–brand pairing is found far more widely láði [‘land’]. Although the corpus reveals than in works of these two poets only and is a some examples of two of these three items common enough word pair that it may have appearing together, this seems to be the only been a traditional rhyme collocation.9 The

Table 6. Magnúsdrápa 7.5–6 by Arnórr jarlaskáld and Hǫfuðlausn 18/19.7–8 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld engr annarr áðr engr þengill áðr láði Óttarr svarti engr elda áðr engr þrøngvir áðr láði Source Nyquist 2014: 38. Arnórr jarlaskáld Óttarr svarti Magnúsdrápa 7.5–8 (Skj B1: 313). Hǫfuðlausn 18/19.5–8 (Skj B1: 272). Engr hefr annarr þengill Breið eru austr til Eiða áðr svá gnógu láði ættlǫnd und þér; Gǫndlar – bráskat bragnings þroski – engr sat elda þrøngvir barnungr und sik þrungit. áðr at slíku láði.

115

Table 7. Magnúsdrápa 4.5–6 by Arnórr jarlaskáld and Hǫfuðlausn 1.7–8 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld Nóregs njóta nýtr létat njóta þengill lengi Óttarr svarti dǫglings dýrka dýr verk dýrka þengill lengi Source Nyquist 2014: 38. Arnórr jarlaskáld Óttarr svarti Magnúsdrápa 4.5–8 (Skj B1: 312). Hǫfuðlausn 1.5–8 (Skj B1: 268). Létat Nóregs njóta Þat telk, garms ok Gauta nýtr þengill gram lengi; glaðnistanda misstak, hann rak Svein af sínum dǫglings verk at dýrka, sókndjarfr fǫðurarfi. dýrr þengill, mitt lengi.

alliterating words, even though they are part vary; being in position 1 in Óttarr’s verse, but of different compounds, include the word Arnórr seems to use this consistently in a land-. The word is self-standing lands different way and varies between its use in [‘lang.GEN’] in Óttarr’s couplet or as the latter position 2 only or inflected to extend into part of a compound in the works of Arnórr. position 3 which could be indicative of the This could be a rather unique way to development of idiolectal features which assimilate a common or convenient hending occur across his compositions. It is also into different poetic and geographic possible that due to the somewhat contexts.10 conventional nature of the land–brandr word- There are a few things worth mentioning pair and Arnórr’s use of compounds rather about this example. The first is that the than simplex forms could indicate the metrical positions filled by the element land correspondence here is incidental and instead

Table 8. Magnúsdrápa 6.7–8, 9.3–4 and Þórfinnsdrápa 18/12.1–2 by Arnórr jarlaskáld and Hǫfuðlausn 9.5–6 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld Magnúsdrápa 6.7–8 ǫld* œsi* Jótlandi fús* œsi* Jótlandi branda Arnórr jarlaskáld Magnúsdrápa 9.3–4 valska víðu Vestlandi rauð víðu Vestlandi branda Arnórr jarlaskáld Þórfinnsd. 18/12.1–2 Hjarri hverri Hjaltlands Harri hverri Hjaltlands branda Óttarr svarti laut létti lands laut létti lands branda Source Nyquist 2014: 36. * œsi is a reconstruction † Matthew Townend’s edition (2012) has “landfolk” rather than “lands”. Arnórr jarlaskáld Arnórr jarlaskáld Magnúsdrápa 6.5–8 (Skj B1: 312). Þórfinnsdrápa, 18/12.1–4 (Skj B1: 320). Setti bjóðr at breiðu Harri fekk í hverri brynþings – fetilstinga Hjaltlands þrumu branda fús tók ǫld við œsi – – greppr vill grams dýrð yppa – Jótlandi gramr branda. gagn, sás hæstr vas bragna.

Arnórr jarlaskáld Óttarr svarti Magnúsdrápa 9.1–4 (Skj B1: 313). Hǫfuðlausn 9.5–8 (Skj B1: 270). Fúss lét á Ré ræsir Laut fyr yðr, áðr létti, rammþing háit Glamma; lands folk í gný branda, valska rauð fyr víðu Engla ferð, at jǫrðu Vestlandi gramr branda. ótt, en mǫrg á flótta.

116

Table 9. Hrynhenda, Magnúsdrápa 4.3–4 by Arnórr jarlaskáld and Hǫfuðlausn 15/16.3–4 by Óttarr svarti. 1st Allit. 2nd Allit. 3rd Allit. 1st hend. 2nd hend. 1st hend. 2nd hend. Poet Word Word Word Word Word Word Word 1st line 1st line 2nd line 1st line 1st line 2nd line 2nd line Arnórr jarlaskáld skjǫldungr skǫrum skeiðar skjǫldungr hvélðan skeiðar reiði Óttarr svarti skatti skreyttu skeið skatti skreyttu skeið reiði Source Nyquist 2014: 39. Arnórr jarlaskáld Óttarr svarti Hrynhenda, Magnúsdrápa 4.1–4 (Skj B1: 306). Hǫfuðlausn 15/16.1–4 (Skj B1: 268). Herskip vannt af harða stinnum Blágjóða, tókt, bræðir hlunni geyst í Salt it Eystra; bengjalfrs, ok þá sjalfa, skjǫldungr, stétt á skǫrum hvéldan skatti gnœgðr, með skreyttu skeiðar húf með girzku reiði. skeið Hôkonar reiði.

this example may be more indicative of composing his own poetry, and which may Arnórr re-using the same rhyme-pairs as seen have influenced his diction more generally. in Table 4 as well. It cannot be ruled out, Thus the assimilation of useful or well-liked however, that Arnórr’s inspiration for this alliterative or hending word combinations particular example, common though it would may have been remembered and used by have been, was not from Hǫfuðlausn. Arnórr to aide in the composition of his own The last example in Table 9 comes from oral poetry. couplets from Hǫfuðlausn 15/16.3–4 and It may be that Arnórr intentionally Hrynhenda 4.3–4. In these final couplets, it is assimilated word constellations from Óttarr the second line alliteration and hending that into his own work, but it is also possible that are shared between these two poets through Arnórr simply internalized these lexical pairs the word pair skeið [‘ship’] and reiði and subconsciously incorporated these [‘rigging’]. This aðalhending rhyme is a constellations into his own work without commonplace collocation that looks to have necessarily calling Óttarr´s Hǫfuðlausn to been well established in skaldic diction.11 It is mind in each instance. The connection unremarkable except as another example of between Arnórr and Óttarr is also supported similar word constellations between the two by the origins and locations of these poets. poets. Both were Icelanders who spent time at These nine instances of similar hending various courts throughout Scandinavia and alliterative word constellations recurring (although not necessarily at the same time) between compositions of Arnórr Þórðarson and they may have been exposed to the same jarlaskáld and Óttarr svarti indicate that some popular verses circulating throughout these relationship between the two skalds and their locales. It is also likely that they would have works may exist beyond what can be had the opportunity to be exposed to one or explained by simple coincidence, although an the other’s poetry as their own verses examination of the entire corpus would be circulated orally around the social sphere of necessary to make this possibility into a Iceland and Scandinavia. The dating of their certainty. It is also interesting that all of the works nevertheless makes it more likely that instances of similarity which I have found Arnórr was influenced by Óttarr than vice between these two poets involve only Óttarr’s versa. There is, of course, the possibility of Hǫfuðlausn rather than other poems he another explanation for these similarities, composed (of which two are relatively well including that they are purely the result of preserved, along with a few so-called coincidence, but within the context of my lausavísur, or independent stanzas or larger research project, there is a real helmings). This could indicate Hǫfuðlausn possibility that such word constellations may became something of an exemplar for Arnórr, have been used to aide in the composition of a work which he called to mind to for skaldic poetry. inspiration and compositional aid when

117

Assimilation in Context Kormákr Ǫgmundarson, Jórunn skáldmær, Including the examples presented in this Þórhallr veiðimaðr, and Bersi Skáldtorfuson paper, there are a total of nine different (Nyquist 2014: 40–46). Each of these instances within my research where a poem instances of inter-poetic similarity are by Arnórr shares common word constellations possible or probable examples of Óttarr svarti with Óttarr’s Hǫfuðlausn. In addition to these, assimilating lexical pairs and constellations there are six similarities in the dataset of my from those poets, just as Arnórr Þórðarson research project between Arnórr’s verses and jarlaskáld may have assimilated useful word those attributed to Glúmr Geirason and five combinations from him. with verses attributed to other poets. These last include one with verses attributed to Conclusions Eyvindr Finnsson skáldaspillir, one with the While it is possible that not every one of the anonymous poem Oddmjór, another with couplet comparisons which appear in this verses attributed Þorleifr jarlsskáld paper is the result of one skald assimilating a Rauðfeldarson and two more with Þórbjǫrn useful rhyme or alliteration into a new hornklofi’s Glymdrápa, all of which are likely composition and context, there are enough earlier sources from which Arnórr could such similarities that a connection between at conceivably borrow (Nyquist 2014: 22–40). least these poets can be inferred with some There is always the possibility that such confidence. In their turn, a number of other parallels are coincidental or reflect skalds not mentioned in this article also show commonplace word-pairings and signs of having similarly assimilated lexical constellations that were simply part of the sets and phraseology from others and of tradition rather than the results of one poet having their own expressions adapted more drawing directly on the phraseology of widely. My data, while by no means covering another’s poem that he knew (cf. Frog 2009; all relevant verses, shows that even in this Frog 2014). It is also possible that the incomplete sample there are a great number of attribution of authorship of some verses is cases where the subtle adaptation of one incorrect or that the lexical parallels have skald’s hending and alliterative word resulted from mixing verses, rephrasing or constellations into a new context could have completing missing/forgotten verses in oral been a tool for the composition of skaldic circulation and by scribes in a scriptorium. poetry (Nyquist 2014). The case of Arnórr However, the links between Arnórr’s verses and Óttarr is illustrative of this pattern in and those of other poets are clearly weighted compositional practice. Arnórr Þórðarson to particular compositions rather than being jarlaskáld and Óttarr svarti may also serve as random. It seems incredibly unlikely that it is examples of a potentially more widespread purely accidental that, of 20 parallels between use of hending and alliterative assimilation Arnórr’s verses and those of other poets, 15 of throughout the wider skaldic corpus. It is a these would be with only two other poets, and real possibility that some skalds used their 9 (or roughly half) would be with a single knowledge of socially circulating verses to poem. subtly incorporate useful rhyming and In the sample corpus of this study, Óttarr alliterating words into their own works as svarti’s poetry, just from the poems from compositional aides for the difficult metrical which I have collected data and analyzed in requirements of skaldic verse. my master’s thesis, contains similarities in Whether or not a widespread trend is hending and alliterative word constellations likely, it certainly seems that using hending with earlier poets’ works in eight instances in and alliteration word constellations to addition to those reviewed above. The poets compare the works of Arnórr jarlaskáld and involved in these possible assimilations by Óttarr svarti reveals evidence of a direct Óttarr include verses by Goþþormr sindri, connection between their works. From my Egill Skallagrímsson, Bjǫrn Arngeirsson research, I believe that there is a real hítdœlakappi and Torf-Einarr jarl (sharing the possibility that some skalds, like Arnórr in the same similarity with two earlier works), examples presented in this article, assimilated

118 useful word constellations into their own 10. Frog (2014: 299) discusses a similar rhyme-pair compositions, which may in turn have with ey in place names and þeyr as the second element in a preceding battle kenning. developed into an integrated part of the poet’s 11. Skeið-/reið- word combinations appear to be more own idiolect. Although it cannot be said with common in Skj B1 than even land/branda, complete certainty that this type of appearing frequently together in different assimilation was widely used by Old Norse compounds and line positions (in at least 16 poets, the evidence merits a closer look at additional instances). The frequency with which these word pairs appear in the corpus most likely skaldic composition using hending and implies a different explanation for their recurrence alliterative assimilation as a possible tool for in the works of different skalds, but it is still worth oral composition. noting these similarities between Arnórr and Óttarr’s works. It is interesting to note however that Acknowledgements: The pilot study on which the skeið- usually takes the second hending position in present work is based was undertaken for the the other examples and reið- the first, but in these completion of a Master of Arts degree at the two instances their respective positions are Department of Linguistics and Scandinavian Studies at switched. the University of Oslo. I would like to thank my advisor Mikael Males for his advice and his recommendation to approach the compositional methods of skaldic Works Cited poetry by examining hending and alliteration which Sources formed the basis for my thesis. Arnórr Þórðarson jarlaskáld, 1973. Haraldsdrápa. In Skj B1: 322–326; SPSMA II1: 260–280. Notes Arnórr Þórðarson jarlaskáld, Hrynhenda, 1. Ilya Sverdlov (p.c.) points out that hendings are Magnúsdrápa. In Skj B1: 306–311; SPSMA II1: often overlooked in skaldic studies, yet are, in fact, 181–206. a key element in the structure of skaldic verse, and Arnórr Þórðarson jarlaskáld, Magnúsdrápa. In Skj B1: scholars such as Olga Smirnitskaya (1994: 345, 311–315; SPSMA II1: 206–229. 349–356, 359–377) base their original metrical Arnórr Þórðarson jarlaskáld, Þórfinnsdrápa. In Skj B1: analysis of dróttkvætt on them, in a significant 316–321; SPSMA II1: 229–260. departure from the popular system (Kuhn 1983; Ottarr svarti, Hǫfuðlausn. In Skj B1: 268–272; SPSMA Gade 1995) based on Sievers’ five types. II2: 738–767. 2. This view is based on the search engine for Skj B of Skj = Den Norsk-Islandske Skjaldedigtning B. 1973 the Skaldic Database. I have only found [1912–1915]. Ed. Finnur Jónsson. København: Ungan/eyðir and ungr/eyða together in these Rosenkilde & Bagger. couplets. SPSMA I = Poetry from the Kings’ Sagas 1: From 3. The other instances where eigi/ægir appear together Mythical Times to c. 1035. 2012. Ed. Diana in couplets are Rǫgnvaldr jarl Kali Kolsson, Whaley. Skaldic Poetry of the Scandinavian Middle lausavísa 13.5 (Eigi/œgir) and lausavísa 35.1 Ages 1. Turnhout, Belgium: Brepols Publishers, (Eigi/œgir), Skarpheðinn Njálsson, lausavísa 2.1; 2012. (Eigi/ægi), and Ketilríðr Hólmkelsdóttir, lausavísa SPSMA II = Poetry from the Kings’ Sagas 2: From c. 1.1 (Eigi/ægi). In addition, frægr–ægir [‘famous’– 1035 to c. 1300. 2009. Ed. Kari Ellen Gade. Skaldic ‘sea’]also appears as a common word pair. Poetry of the Scandinavian Middle Ages 2. 4. According to the Skj B search, the stem þorð- and Turnhout, Belgium: Brepols Publishers 2009. verb varða only appear together in these examples. 5. In Skj B, Gengr,- gǫngu/yngvi, Yngvi combinations Literature are found in these examples along with only one Frog. 2009. “Speech-Acts in Skaldic Verse: Genre, additional work by Þorkell hamarskáld Compositional Strategies and Improvisation”. In (Magnússdrápa 5.1, specifically “Uppgǫngu réð Versatility in Versification: Multidisciplinary yngvi ítr með helming lítinn; Skj B1: 408). Approaches to Metrics. Ed. Tonya Kim Dewey & 6. According to my searches of Skj B1, sikling- and Frog. New York: Peter Lang. Pp. 223–246. mikl- combinations are only found together in these Frog. 2014. “Oral Poetry as Language Practice: A examples. Perspective on Old Norse dróttkvætt Composition”. 7. This is based on the my searches of Skj B1. In Song and Emergent Poetics – Laulu ja runo – 8. These are the only instances where I have found Песня и видоизменяющаяся поэтика. Ed. Pekka þengill/lengi together in Skj B1. Huttu-Hilttunen et al. Runolaulu-Akatemian 9. In addition to these four land-/branda examples, I Julkaisuja 18. Kuhmo: Juminkeko. 279–307. have found nine other instances where poets use Gade, Kari Ellen. 1995. The Structure of Old Norse this word combination, including an additional Dróttkvætt Poetry. Islandica 49. Ithaca: Cornell example from Hǫfuðlausn 17.2 by Óttarr svarti University Press (lands rôðǫndum, branda; Skj B1: 271). Gurevich, Elena A. 1994. “The System of Kennings”. Nordica Bergensia 3: 139–156.

119

Kuhn, Hans. 1983. Das Dróttkvætt. Heidelberg: Carl Townend, Matthew (ed.). 2012. “Ottarr svarti”. In Winter. Poetry from the Kings’ Sagas 1: From Mythical Nyquist, Cole E. 2014. “Compositional Techniques and Times to C. 1025 Part 2. Ed. Diana Whaley. Perceptions of Authorship in Skaldic Poetry: Volume I of Skaldic Poetry of the Scandinavian Alliterative and Hending Word Constellations”. Middle Ages. Turnhout, Belgium: Brepols. Pp. 738– Master’s Thesis. http://www.duo.uio.no/ Oslo: 789. Reprosentralen, Universitet i Oslo. Whaley, Diana (ed.). 2009. “Arnórr jarlaskáld Smirnitskaya, Olga A. 1994. Стих и язык Þórðarson”. In Poetry from the Kings’ Sagas 2: древнегерманской поэзии [‘The Verse and From C. 1035 to C. 1300 Part 1. Ed. Kari Ellen Language of Old Germanic Poetry’]. Moscow: Gade. Volume II of Skaldic Poetry of the Moscow State University. (Extensive review in Scandinavian Middle Ages. Turnhout, Belgium: English by Anatoly Liberman from 1998 is Brepols Publishers n.v. Pp. 177–281. available in Scandinavian Studies 70(1): 87– 108.

120

PROJECTS

Personal Name Systems in Finnic and Beyond: Reconstructing the Concepts of Name Giving in Cultural Layers of Prehistory Terhi Ainiala, University of Helsinki

Personal name systems are under scrutiny in a surname systems in European languages. current research project at the University of These systems are culture-specific; some Helsinki “Personal Name Systems in Finnic languages use several first names, others user and Beyond”. The present article is an patronyms, some prefer names passed on overview of the project and contains an within the family, while some employ new introduction to the project’s goals and names for every generation, etc. methods. Led by Terhi Ainiala (University of In all cultures, giving a child a name means Helsinki), funding for the project is provided that he or she is accepted as a member of the by the Academy of Finland (2015–2019). community. Richard Alford (1988) has The project focuses on the study of demonstrated that personal names express the anthroponymic (i.e. personal name) systems identity of a person in two ways: In the first in Uralic languages with a special emphasis place, personal names tell the other members on reconstructing their historical strata and of the community who the individual in cultural contacts reflected in personal names. question is. Secondly, personal names tell the The project aims to reconstruct, as far as community who he or she is or is expected to possible, Pre-Christian or early Christian be. Personal names thus have a significant anthroponymic systems on the basis of role in building a person’s individual and archive materials, fieldwork data, and by way social identity and constituting links between of employment of material from modern generations, families, clans and tribes. They surnames and toponyms. A special emphasis express different religious, political, ethnic is laid on the study of Finnic languages and and other values associated with groups of Mari (a Uralic language spoken primarily in people. In this connection they also reflect the Mari El Republic of the Russian political, religious or value change and build Federation), but also the other Finno-Ugrian historically multi-layered systems that can be languages are taken into consideration. It is implemented in the cultural reconstruction of the research group’s aim to develop a Uralic past world views. anthroponymic database that would enable the In addition to identification, a name also systematic collection of the personal name has a function of classification. Personal material from Finno-Ugrian languages and names do not only identify a person but also provide basic guidelines for its structural and make him or her a member of a group and semantic analysis. provide him or her with a social identity, an identity which is often explicitly gendered. Personal Naming in Cultures Personal names powerfully signify speakers’ Cultural information – such as a people’s sense of local belonging, cultural identity and religious, linguistic and ethnic identity – is system of values. While the basic always reflected in personal names. As a identification function of personal names is result, historical researchers, cultural the same everywhere, their secondary societal researchers and sociologists continuously and group linked functions may vary quite a regard anthroponyms as subjects of great bit, depending on cultural context. As interest. Personal names are not isolated elements inherited from the past, names often elements in any language or its nomenclature reflect more archaic linguistic and cultural but represent a system with different sub- relations, concepts and value systems than the systems, such as first name systems and present language use, and are thus of great

121 significance for the investigation of past only in expanding place name collections but conceptual realms, inter-group relations, also in the field of comparative mythology cultural identities and beliefs. The study of and folklore, where new and significant anthroponyms is thus an inherently results enable us to reconstruct historic Finnic interdisciplinary field of investigation with cultural layers reflected in oral poetry and implications beyond pure onomastics. (pre-Christian) religious practices (cf. Siikala Notwithstanding the central nature of 2012). anthropomorphic systems as identity bearers and as source of cultural information, Finno- Objectives and Expected Results Ugrian personal name systems have been The primary aim of the project is to relatively little studied. The only reconstruct, etymologically describe and monographic treatment regarding the old contextually study Pre-Christian and Early anthoponymic system of the Finnic-speaking Christian Uralic anthroponymic systems. A people (Stoebke 1964) is already a half a special emphasis is placed on Old Finnic and century old and reflects the point of view of Mari personal name systems. In addition, early structural linguistics. Some reconstruction of anthroponymic systems of anthroponymic handbooks have been other Uralic languages will be undertaken. published, especially on surnames (cf. Material used for the purposes of Mikkonen & Paikkala 2000; Rajandi 2005; reconstruction of old personal names is Černyh 1995) and a handful of modern and primarily the variants of anthroponyms, well-founded investigations have seen mainly surnames, attested historically or in publication, particularly works on the topic of the onomastic material collected by Udmurt tribal (or Vorshud) names (Atamanov fieldwork, and also the toponyms derived 1996; 2001). However, most of the naming from anthroponyms. Similar methodologies motivations of the Uralic speaking people have earlier been employed by Karlova remain little studied. The toponymic (2007) and Saarikivi (2007) and these investigations by Mullonen (1994, 2002), approaches will be developed further in the Karlova (2004) and Saarikivi (2003; 2006) course of the research. Self-evidently, the shed light on some aspects of the old proposed reconstruction will not be complete. anthroponyms, particularly those in the Reconstruction conducted in the framework historical area of the eastern Finnic languages, of this project can only be considered the first as reconstructed on the basis of toponymic step of such an undertaking. Materials studied material. The same group of names is also in the project derive from existing databases, investigated in light of material from publications and archives, and are Novgorod birch bark letters in an article by supplemented by fieldwork conducted during Saarikivi (2007). It is obvious that a similar the project. methodology could also be implemented in Besides this primary goal, the project has search of old anthroponyms in other contexts. important implications for the cultural Yet the general impression is that while the reconstruction of the past of Uralic-speaking toponymic research in Finland developed peoples. Pre-Christian or Early Christian entirely new paradigms starting from the anthroponyms reflect the world view of the 1970s and reached many fruitful results early communities of the Northern Eurasian regarding name typology, distribution of hemisphere. It is estimated that research name models and layering of toponyms (see involving anthroponymic material will reveal e.g. Ainiala & Saarelma & Sjöblom 2012: 47– mythological concepts and denominations of 60), the historical study of anthroponyms has neighbour ethnicities as well as culturally largely come to a standstill. No attempt at an relevant naming practices related to the etymologically relevant reconstruction of the ethnolinguistic world view of past old Finnic – or other Finno-Ugrian personal populations. In this interdisciplinary name systems – has been published. This is framework, the project will employ, as far as the more regrettable given that new material possible, folkloristics, comparative religion, for such an enterprise is now available not and comparative mythology.

122

The aforementioned research will shed dissertation. Petrozavodsk: University of light on many problems central to Finno- Petrozavodsk. Karlova, Olga. 2007. “Das karelische Personennamen- Ugrian onomastics and – most notably – will systeme”. In Europäische Personennamensysteme. assist in deciphering a large number of Ein Handbuch von Abasisch bis Zentralladinisch. modern surnames, toponyms (especially Ed. Andrea Brendler & Silvio Brendler. Pp. 363– settlement names) and layers of nomenclature 371. Hamburg. systems. Further, this research is likely to Mikkonen, Pirjo & Sirkka Paikkala. 2000. Sukunimet. Helsinki: Otava. bring into focus different naming motivations Mullonen, Irma. 1994. O erki vepsskoj toponimii. deriving from various cultural periods, for Sankt-Peterburg: Nauka. instance, hunter-gatherer vs. agricultural, Mullonen, Irma. 2002. Toponimi Prisvir’ : problemy theophoric vs. characterizing of an individual ètno ykovogo kontaktirovani . Petrozavodsk: and inherited vs. borrowed motivations. In Petrozavodskij gos. Pöyhönen, Juhani. 1998. Suomalainen this respect, it is likely that project results will sukunimikartasto. Helsinki: SKS. be implemented beyond onomastic science, Rajandi, Edgar. 2005. Raamat nimedest. 2. täiendatud most notably contributing to our trükk. Tammerraamat. understanding of early history and prehistory, Saarikivi, Janne. 2003. Taipaleentakaisten tšuudien folklore, comparative religion and mythology. paikannimet: Pinegan piirikunnan suomalais- ugrilaisen substraattinimistön perusosaa For more information, please visut our vastaavassa asemassa esiintyvät elementit. website at http://blogs.helsinki.fi/personal- Unpublished licentiate thesis. Helsinki: University name-systems/. of Helsinki. Saarikivi, Janne. 2006. Substrata Uralica: studies on Key Literature Finno-Ugrian substrate in Northern Russian Ainiala, Terhi. Saarelma, Minna. Sjöblom, Paula. 2012. dialects. Helsinki: University of Helsinki. Names in Focus. An Introduction to Finnish Saarikivi, Janne. 2007. “Finnic Personal Names on Onomastics. Studia Fennica Linguistica 17. Novgorod Birchbark Documents”. In Topics on the Helsinki: Finnish Literature Society. Ethnic, Linguistic and Cultural Making of the Alford, Richard D. 1988. Naming and identity: a cross- Russian North. Ed. Nuorluoto, J. Slavica cultural study of personal naming practices. New Helsingiensia 32. Helsinki: University of Helsinki. Haven: HRAF Press. Pp.196–246. Atamanov. M.G. 1996. Istori dmurtii v Siikala, Anna-Leena. 2012. Itämerensuomalaisten geografi eskih na vani h. Iževsk: Udmurti mytologia. Helsinki: SKS. Atamanov, M.G. 2001: Po sledam udmurtskih Stoebke, Detlef-Eckhard. 1964. Die alten voršudov. Iževsk: Udmurti . ostseefinnischen Personennamen im Rahmen eines Černyh, S. J. 2005. Slovar marijskih li nyh imen. urfinnischen Namensystems. Nord- und Joškar-Ola. osteuropäische Geschichtsstudien. Band IV. Karlova, Olga. 2004. Venäjän Karjalan -l(V)-loppuisen Hamburg: Leibnitz-Verlag. nimimallin paikannimistö. Unpublished

123

CONFERENCES AND EVENTS

Old Norse Mythology Conference 2015: “Myth, Materiality and Lived Religion” 4th-5th November 2015, Stockholm, Sweden Klas af Edholm, Stockholm University

Stockholm University’s “Myth, Materiality better off looking more to the context, hence and Lived Religion” conference was held at the interpretation as depictions of the City Conference Center in Stockholm warlords/gods. The paper’s respondent was from the 4th to 5th November. Over the course Maria Pettersson (RAÄ Linköping). of its two days, the conference presented a Sigmund Oehrl (Georg-August-Universität- variety of perspectives and angles on the Göttingen) was the third speaker, presenting topics of Norse Mythology, materiality, and “Picture Stones from Gotland Analyzed with the concept of lived religion. The program the Reflectance Transformation Imaging was arranged with a speaker and a respondent, (RTI) Method”. Oehrl presented a new who were both provided time to present method of scanning picture stones, a method before discussion was open to the audience. which yields astonishing results. New On Wednesday, the conference started up iconographic material becomes visible and with a common lunch for all speakers and old interpretations can be revisited. The respondents. Afterward, Olof Sundqvist, paper’s respondent was Anne-Sofie Gräslund Bengt Norén and Annika Olsson (all (Uppsala University). representatives of the University of The second session of the day was Myths Stockholm) welcomed everyone in the in Texts I and featured Peter Jackson conference room. During the first session, (Stockholm University) in the chair. The with focus on Myths in Pictures, Olof session’s first speaker was Steve Mitchell Sundqvist had the chair. (Harvard University) with “Níðhoggr and his Anders Hultgård (Uppsala University) was Kin”. Mitchell discussed dragons, serpents the first speaker with “Myth on Stone and and other terrifying – but also protective – Tapestry: Ragnarök in Pictures?”. Hultgård creatures of Norse Mythology. The paper’s presented (with a few examples) problems respondent was Judy Quinn (University of one must deal with when interpreting Cambridge). iconographical material. Carved pictures on The second speaker of the session was rune stones and the Överhogdal tapestry can Rudolf Simek (University of Bonn) with not easily be interpreted as featuring motifs “Demons, Illnesses and Amulets”, discussing from Norse Mythology, but neither can the amulet artefacts and inscriptions related to motifs be interpreted as clearly Christian. illness. How can myth be more materialized John Lindow (University of California, than in an amulet? The chants and protective Berkeley), the respondent noted how the words of medieval amulets reflect much older material forces us to examine a context of concepts. The paper’s respondent was Pernille religions merging and interacting. Herrmann (Aarhus University). The second speaker was Margrethe Watt The third speaker of the day was Stefan (Museum of Bornholm) with “Do the Gold Brink (University of Aberdeen) with “Law Foil Figures (gullgubbar) Reflect Norse and Myth”. Brink presented reflections of Mythology?”. Watt’s presentation focused on Norse myth and cult practice in medieval law a very interesting groups of motifs that are codes. Some of these sources provide a distinguishable in the iconography of the gold glimpse at the lived religion. The paper’s foil figures. Instead of trying to identify respondent was Jonas Wellendorf (University characters appearing on the gold foil, we are of California, Berkeley).

124

That evening all speakers and respondents and her killing as the start of a primordial were invited to a reception and dinner at battle between gods and giants. The paper’s Stockholm City Hall, with the City of respondent was Eldar Heide (University of Stockholm as their host. Bergen). Beginning Thursday 5th November, the After a lunch break, the next session third session focused on the theme Myth in started with focus on Myth, Animals and Texts II and featured Catharina Raudvere as Sacrifices. Torun Zachrisson held the chair session chair. position. The first speaker of the session was The session’s first speaker was Eldar Christina Fredengren (Stockholm University) Heide (University of Bergen) with “Old with “Finitude: Human and Animal Sacrifice Norse Vision Literature and the Term leizla in a Norse Setting”. Fredengren presented ‘Guiding’: An Introduced Genre and Its research from an ongoing project involving Indigenous Background”. Heide discussed analyzing human and animal sacrifices in wet how the Old Norse term leizla might provide areas in Sweden. Recent excavations have a clue for the prominence of indigenous yieled new material that may potentially be vision literature, as contrasted with the the remains of human sacrifices and.old Christian vision literature tradition. The excavations are therefore being revisited. The paper’s respondent was Carolyne Larrington paper’s respondent was Catharina Raudvere (Oxford University). (Copenhagen University). The second speaker of the session was The second speaker of the session was Ola Frederik Wallenstein (Stockholm University) Magnell (RAÄ, Lund) with “Animals of with “What Does Odin Do to the túnriðor? – Sacrifice: Animals and blót in the Old Norse An Attempt at Contextualizing Hávamál Sources in Relation to Faunal Remains from 155”. Discussing the interpretations of the Settlements and Ritual Depositions”. Magnell stanza in the Old Norse poem Hávamál, provided examples of (and statistics for) Wallenstein highlighted the usefulness of correspondence between mythological and comparative perspectives and the narrative descriptions of animal sacrifices and contextualization of the phenomenon osteological remains from settlement finds described. The paper’s respondent was Terry and ritual deposits. Regional differences in Gunnell (University of Iceland). fauna and differences between settlement The day featured a second session. This finds and ritual deposits were clear. The session focused on the theme Myth and Social paper’s respondent was Kristin Armstrong Practice. Anders Hultgård held the chair Oma (University of Stavanger). position for the session and the first speaker The final session of the conference focused of the session was Frog (University of on the theme of Myths in Texts III and Helsinki) with “Understanding Embodiment featured Anders Andrén as the chair. through Lived Religion: Authority, Social The first speaker of the session was Jens Practice and Imaginal Experience”. Frog Peter Schjødt (Aarhus University) with pointed at how mythological themes and “Mercury – Wotan – Óðinn: One or Many?”. motifs are given different expressions and Schjødt discussed problems (and possible characteristics depending on context, genre confusion) when speaking of potentially and practice. The paper’s respondent was different gods as one and the same. There is a Margaret Clunies Ross (University of need to define in what manner one might Sydney). speak of the same god – or several different The second speaker of the session was gods. Depending on one’s perspective, the Tommy Kuusela (Stockholm University) with answer to Schjødt’s question is both ‘yes’ and “Halls, Gods and Giants: The Enigma of ‘no’. The paper’s respondent was Peter Gullveig in Odin’s Hall”. Kuusela presented a Jackson. new interpretation of the mythological motif The final speaker of the session was of the arrival and killing of Gullveig in Odin’s Merrill Kaplan (Ohio State University) with hall. A contextualization might speak in favor “Gruesome Gold in Sigurðarkviða in skamma of an interpretation of Gullveig as a giantess 49–50”. Kaplan explicated the many

125 dimensions of two stanzas of Sigurðarkviða respondent was Agneta Ney (Uppsala in skamma with their subtleties of potential University). meanings. An underlying correspondence and Peter Jackson, Andreas Nordberg reflection between gold and the color red is to (Stockholm University) and Olof Sundqvist be found in stanzas 49–50 of the Old Norse provided the final words of the evening and poem Sigurðarkviða in skamma. For example, the conference wrapped with a dinner for a red golden necklace could also mean a speakers and respondents at the Stockholm gruesome bleeding death. The paper’s restaurant Ulla Winbladh.

“The Ontology of Supernatural Encounters”: The 4th Symposium of the Old Norse Folklorists Network 10th–12th December 2015, Tartu, Estonia Tommy Kuusela, Stockholm University

The fourth symposium of the Old Norse supernatural incidents of the haugbrot episode Folklorists Network took place in the lovely within the saga from an archaeological point town of Tartu at the beginning of December. of view, and also mentioned grave mounds The weather gods approved of this and their inhabitants in later, mainly international gathering of scholars interested Norwegian, folklore. The next speaker was in the supernatural, smiling upon us and myself, with a paper that I titled “Is the Year blessing the days with mild and sunny Walk (Sw. årsgång) Ritual a Remnant from weather. Organized by Daniel Sävborg Pre-Christian Traditions?” With the (University of Tartu), Mart Kuldkepp presentation, I tried to refute the often (University College London) and Karen Bek- repeated notion that a Swedish folk tradition Pedersen (University of Southern Denmark), known as the ‘year walk’ is an ancient this time the topic was dedicated to the practice, traceable to a heathen past. After my question of the ontology of supernatural talk, Agneta Ney (Uppsala University) encounters in Old Norse-Icelandic literature, presented “Ghost Riders as Motif in Heroic myth and legend, as well as the ontology of Poetry and Ballads”. She compared the motif later Scandinavian folklore – a theme that was of the ghost riders, known from heroic poetry pondered and illuminated during the many in the Poetic Edda, with medieval ballads presentations and subsequent discussions from Scandinavia, and focused on the throughout the conference. dialogues and what social implications they The first presentation was by a keynote mediated for a contemporary audience. speaker, John Lindow (University of The other session was more focused on a California, Berkeley), who gave a most single topic, namely the dead in folklore and interesting and stimulating interpretation of burial mounds in Old Norse literature. Over “The Essential Other: Continuity and lunch at the Vilde Restaurant, word was that it Discontinuity in World View”. He very had been a most stimulating session. The convincingly showed how notions of world presenters were Jan Kozák (Charles views both changed in the past and still linger University in Prague) with “Burial Mound as yet in Scandinavian myth and folklore. a Projection Screen”, Gerður Halldóra Lindow immediately brought folklorists and Sigurðardóttir (University of Iceland) with scholars of Old Norse religion and literature “Managing the Dead: Performances of Death together, as the topic he discussed was one of in Old Norse Literature and Scandinavian interdisciplinary importance. The first two Folklore”, and Sara Duppils (Gavle parallel sessions were Dead and Buried and University) and Sandra Lantz (University of Monsters and Ghosts. As I was one of the Aberdeen) with “The Living Dead in Folk presenters in the latter, the choice of which Religion – A New Perspective”. The next two session to attend was decided for me. Arnhild sessions were Fornaldarsögur and Bodies Big Mindrebø started with her paper “The and Small. It was a tough call, but I decided Haugbrot in Harðar saga”. She discussed the to attend the first of these, which opened with 126

Helen Leslie-Jacobsen’s (University of they at the same time can potentially blur the Bergen) “Coming Face to Face with Sigurd line between natural and supernatural Fafnesbane in Medieval Norway”. Leslie- encounters. Frog was followed by Andreas Jacobsen considered the fact that while the Schmidt (Ludwig-Maximilians University, archangel Michael is biblical, and a legitimate Munich) who presented “Goddess, Fylgja or object of belief for medieval churchgoers, this Fiction? – Þorgerðr Hörðabrúðr, the Vanir is less immediately obvious of Sigurður. From and the Question of Mythic Structures”. this, she moved on to discuss how Sigurður Schmidt investigated how the chieftain was conceptualised and understood in Sigmundr Brestisson in Færeyinga saga medieval Norway, particularly in regards to undergoes an initiation to Óðinn and Þorgerðr his physicality. Next up was J.Y.H. Hui when he becomes the ruler of the Faroe (University of Cambridge) with “The Making Islands. He then moved on to an examination of Paranormal Weapons: Pointed Messages in of Þorgerðr and her nature as a whole. He Örvar Odds saga”. He discussed two sets of argued that a mythic substructure, with supernatural arrows in Ǫrvar-Odds saga. The emphasis on the Vanir, can be found within second set is given to the hero by Óðinn (or at the saga itself. In the other session Csete least an Odinic character) to defeat his Katona (University of Debrecen) considered enemies, who are worshippers of Freyr and aspects of Njáls saga with “Weak Magic? – resilient to his first set of arrows. He argued Some Notes on the Allegedly Supernatural that it can be read in conjunction with, rather Nature of Gunnar’s Halberd in Njáls saga” than in contrast to, the important subtext of and Fredrik Charpentier Ljungqvist Oddr’s Christian faith. In the parallel session (Stockholm University) discussed perceptions (which I could not attend) Eva Þordís of the world in his talk “At the Edge of the Ebenezersdóttir (University of Iceland) Known World: Supernaturality and the Old presented “(Super)Natural Bodies”, followed Norse Geographical Perceptions of the by Scott T. Shell (University of California, World”. These papers concluded the sessions Berkeley) who presented his understanding of of the opening day. Before dinner, “Hidden Children as Represented in participants were given the choice of a tour of Germanic Folklore”. The conference was in Tartu with either a Swedish- or English- full swing at this time and the subsequent speaking guide. For my part, I went back to coffee break became a field for much fruitful my hotel and, before getting some rest, I discussion on the supernatural combined with considered that the atmosphere of the appreciation for the presentations so far. symposium had been scholarly, merry and After the break, two new parallel sessions friendly – a great first day. started, entitled What is Natural? and The second day started, like the first, with Ontology. As earlier, it was a tough choice to a keynote speaker, this time Mikael Häll pick a session, but feeling in the mood for (Lund University). Häll provided a brilliant some ontology, I went for the second one. survey of fascinating material and results The session started with Frog (University of from his doctoral dissertation, composed in Helsinki) who reflected on the complicated Swedish in 2013. It was a good occasion to topic “Ontologizing Integers of Supernatural introduce his impressive study to an Encounters: Approaching Meanings in international audience. Häll’s presentation Mythic Discourse”. Frog focused on symbolic bore the long title “Carnal Spirits of Nature elements. With this he meant mythic images, and Demonic Bodies of Vapour: The motifs and plots that engage imagination and Ontology and Sexuality of Supernatural are expressed through language. He then Beings in Early Modern Swedish Folklore postulate a formal ontology as a tool for and Demonology”. In his eloquent discussing descriptions of supernatural presentation, Häll considered how intimate encounters as use and variation of these encounters and sexual intercourse with nature symbols in mythic discourse. He argued that spirits became a crime against the natural recombinations of these elements construct order in 18th and early 19th century Sweden. the meaningfulness of narrated events, while For this crime, some people were sentenced to

127 death via the secular and ecclesiastical laws of that session had also been highly stimulating. Sweden. Häll demonstrated how popular After yet another merry lunch at Vilde conceptions were translated into the language Restaurant, we readied ourselves for the final of demonology in court cases during this sessions: Saga Monstrosities and Thresholds period. Supernatural beings, like the Neck and of Christianity. I found the name of the first the forest nymph, were attributed to the Devil of the three parallel sessions most interesting and therefore became associated with evil and decided to attend it. The session opened spirits and a crime against nature. with Arngrímur Vidalín (University of After the keynote, two parallel sessions Iceland) who presented “Paranormal started, labelled Reality? and Otherworld Encounters in the Sagas”. He asked whether Beings. Once more, the choice was hard to medieval audiences could believe in make, but inspired by the last keynote, I narratives about monsters and strange settled for the second session. The first creatures or if they rather would have simply speaker was Julian Goodare (University of thought those accounts to be believable. He Edinburgh) who spoke of “Visionaries and had kindly sent us an article in advance from Nature Spirits in Scotland”. He described how a soon to be published anthology, where he visionaries in early modern Scotland reported argues for the term paranormal over the term encountering nature spirits, generally fairies, supernatural in saga literature. and that these experiences were often Rebecca Merkelbach (University of traumatic and felt to be ‘real’. Goodare went Cambridge) was up next with “Meeting the on to discuss how the visionaries themselves Monster: The Social Implications of Berserk understood their relationship with these Encounters in Íslendingasögur”. Merkelbach encounters. After Goodare, Ugnius addressed human encounters with different Mikučionis (University of Bergen) went up to kinds of berserkers in the Family Sagas, the stage and presented his paper considered the reactions they provoke in “Recognizing a Dwarf – Myths and Facts”. members of society and what they tell us He rightly disputed the often mentioned about society’s perception of the berserker as notion, at least in popular culture, that dwarfs monster. Sadly, the next speaker, Kirsi were considered a short, bearded, and male- Kanerva (University of Turku) with only race. Mikučionis concentrated his focus “Wonders, Apparitions, Delusions? – on several Old Norse texts that provide a Supernatural Phenomena in Sagas”, never more or less detailed description of the made it to the conference due to an illness. In physical appearances of the Nordic dwarfs. In the other session, Ülo Valk (University of addition, he argued that the Norse dwarfs Tartu) presented “The Devil and Nature were not considered to have looked the same Spirits between Christianization and Nation and did not fit a general description. The last Building” and Mart Kuldkepp presented “The speaker in this session was supposed to be Distress of Bárðr: Community and Sander Mändoja with “Presentation of PhD Christianity in Bárðar saga Snæfellsáss”. Project: Water Spirits in Swedish Folk It was decided that the last speaker for Legends” but unfortunately she was ill and “Thresholds of Christianity” should be could not attend, so we decided to use the last addressed to all of the participants. Everyone minutes for further deliberations on the topic gathered in anticipation for the final of Norse dwarfology. presentation: Maths Bertell (Mid Sweden The presentations for the parallel session University) with “I bumped into Thor the consisted of Daniel Sävborg with “Folk Belief Other Day; He Told Me to Say Hi – The Idea and Fantastic Fiction in Íslendingasögur: of the Holy Pre- and Post-Conversion”. What is What?”, Pasi Enges (University of Bertell asked what it means to meet the Turku) with “Approaching Reality and divine, the holy, in the perspective of a Fantasy through Belief Narratives”, and religious person. With examples from Kaarina Koski with “The Moral Grounds of comparative religion, he argued that, rather Reality in Finnish Folklore and Vernacular than searching for Christian influences in the World View”. I heard that the presentations in Icelandic texts, we should look elsewhere and

128 try to see our own material in a new light. In a The friendly quality of the reception – and way, Maths’ presentation worked as a keynote throughout the conference – was encouraging and was positively received. Like John and favourable for networking. Lindow’s opening, it was a wide-reaching My overall impressions of the conference topic that addressed everyone in the lecture is positive and optimistic, not only were the hall. Only one thing remained on the schedule presentations interesting and respectful of – a special trollological reception at the Tartu time frames, but I also feel that an City Museum, meaning that one of the interdisciplinary approach for discussing the organizers, Daniel Sävborg, had invited us to ontology of supernatural encounters is both participate in a celebration of his 50th important and bridges a gap that is all too birthday. The reception was wonderful and a often apparent between different academic proper celebration of Sävborg, who had disciplines. worked hard to make the conference possible.

Austmarr V: “No One Is an Island” 15th–16th October 2015, Visby, Gotland Kendra Willson, University of Turku

The fifth annual meeting of the Austmarr th th Islands do not merely reproduce on a network was held on 15 –16 October 2015, manageable scale the dynamics and at the Gotland Museum in Visby, Sweden. processes that exist elsewhere. Islandness is The Austmarr network is an international, an intervening variable that does not interdisciplinary network of scholars determine, but contours and conditions interested in cultural contacts and physical and social events in distinct, and developments in the Circum-Baltic region in distinctly relevant ways. (Baldacchino 2004: prehistoric and early historic times. The 278.) Network aims to overcome the traditional The island theme brought together the barriers among different disciplines interested different contributions in a discussion of what in reconstructing the past (including history, is ‘special’ about the island communities in history of religion, archaeology, folklore, the Baltic Sea. The conference had a stronger linguistics and philology), and between the archaeological emphasis than previous national traditions in each of those disciplines Austmarr seminars and a center of gravity in in re-examining the crucial role of contacts in Sweden, with particular focus on the island of shaping the modern ethnic identities Gotland. (Germanic, Baltic, Slavic, Finnic, and Sámi) Early arrivals met on the evening of found around the Baltic Sea. The 2015 event Wednesday, October 14th, for socialization at was sponsored by the National Board of the Black Sheep Arms, a pub and restaurant in Antiquities and the Gotland Museum, with central Visby. On Thursday morning, funding from Riksbankens Jubileumsfond. following a welcome address by Laila Kitzler The main organizers of the event were Laila Åhfeldt and Per Widerström, the seminar Kitzler Åhfeldt (Swedish National Heritage began with Alexander Podossinov’s (Russian Board, Visby) and Per Widerström (Gotland Academy of Sciences, Moscow) report on Museum, Visby). “Islands in the Baltic as Reflected in Ancient The theme of the meeting was “No One Is Literature from Homer to Jordanes”. Ancient an Island: Islands in the Baltic Sea 500–1500 writings mention a number of islands in the AD: Characteristics and Networks in an Northern Ocean, of which the Baltic was Interdisciplinary Perspective”. As Michael viewed as a part. According to the Odessey, Meichsner (University of Greifswald) put it in Odysseus passes three islands on his voyage his talk, this perspective involves “taking the in the Northern Ocean, featuring giant man- island thing seriously”; he cited eaters and an entrance to Hades. The island of Baldacchino’s (2004: 278) observation that Thule, described by Pytheas in the 3rd century

129

BC, was identified with various northern (Hedeby, Ribe, Birka, and Kaupang) that are locations by medieval scholars. Hecataeus known to have developed in Viking-Age from Abdera, writing in the 3rd century BC, Scandinavia. He emphasized that towns were mentions the islands Elixoia [‘Beyond highly exceptional in the Viking Age; most of Celtica’] and Scanza. The belief that society was rural and agrarian. These few Scandinavia was an island, rather than a Viking Age towns functioned as conceptual peninsula, persisted until Adam of Bremen in ‘islands’ where different rules applied than the 11th century AD. Numerous other islands elsewhere. Market areas were physically in the Baltic are mentioned by ancient Roman differentiated from the surrounding territory writers. and were protected by rulers so that Tatjana Jackson (Russian Academy of merchants would not be subject to the dangers Sciences, Moscow) spoke on “Danish Islands which attended most strangers. in Knýtlinga saga”, explaining why the saga Merili Metsvahi (University of Tartu) refers to the journey from Jutland to Funen discussed “The Legend about the Brother- and Zealand as travel north rather than east. Sister-Marriage Connected with the Lake The rotated frame of reference places the Kaali in Saaremaa”. This lake in the middle of regions of Denmark in the Norse worldview, the Estonian island Saaremaa was caused by a a worldview which divided the world into meteorite landing around 400 BCE; it has regions; directional terms refer to travel been speculated that the event led to later toward those regions. This ‘rotated’ frame of stories about flying dragons. According to the reference is not seen in other texts such as legend discussed by Metsvahi, after a Heimskringla which were among the sources wedding between siblings was performed at a for Knýtlinga saga. However, it is seen in manor house, the house sank into the ground King Alfred’s additions to his translation of and the lake appeared in its place. The same Orosius and may be reflected in other Anglo- legend is associated with other Estonian lakes, Saxon texts. The ‘shifted’ orientation scheme including Valgjärv, where Viking Age appears to be more original and is preserved remains can be seen under its shallow water in Knýtlinga saga from oral tradition. on calm days. The legend reflects the Maths Bertell (Mid Sweden University, interface between the pre-Christian Sundsvall) discussed “Viking Age Föglö: A matrilineal structure of Estonian society and Travel Route and Its Population”. Föglö is a the patrilineal Baltic Germans. Sibling collection of islands in shallow water in the marriage is a stronger taboo in matrilineal outer Åland archipelago. It is among the first societies, whereas in patrilineal ones, father- place names in Åland mentioned in medieval daughter incest is a more prevalent concern. sources (in Valdemar Sejr’s 13th century The Baltic Germans who comprised the ruling itinerary), although it was not a power center class had a stronger taboo against marrying and it lacks significant early landmarks, the lower-class peasants than their own unlike Lemböte and Kökar (locations also relatives; because they were a small mentioned in the itinerary). The area lies community, sometimes cousin marriages along travel routes but would be difficult to occurred. Metsvahi argues that this legend navigate without a local guide. This part of showing supernatural punishment for the Åland archipelago has been little studied violation of incest taboos reflects the Estonian in relation to the Viking Age, although peasants’ view of the Baltic Germans. remains from the Bronze Age and Early Iron Michael Meichsner’s “A Force from the Age have been discovered. Åland in general Outside: Gotland and the Political Networks suffers from a lack of funding for Viking Age in the Baltic Sea Region in the 15th Century” archaeology. discusses political changes in Gotland at the Sven Kalmring’s (Centre for Baltic and end of the Middle Ages. Whereas islands are Scandinavian Archaeology, Schleswig) paper now viewed as peripheries of remote centers, “Islands in the Rural Sea: Viking-Age Towns in medieval times Gotland connected and between Cosmopolitanism and Delimitation” constituted the maritime space. It was discussed the status of the few towns strategically important in the Kalmar Union.

130

Gotland in the Middle Ages had had neither museum,. The Friday morning session native nobility nor local elites. In 1407–1408, focused on Gotland. Early risers were given a the Teutonic Knights handed over Gotland to tour of Gotland Museum exhibitions by Per Eric of Pomerania, who visited the island Widerström. Widerström then presented on personally and introduced external controls of “Archaeology on Gotland: Odin and the Mask a type that had not been known there before. from Hellvi”. A 2nd century Roman cavalry Visby was allowed to maintain its traditional mask helmet representing Alexander the privileges but entered an obligation for Great was reworked as a mask featuring one mutual support with the king. Peasants (i.e. bright reflective eye and one dark eye, rich farmers) were allowed to trade outside apparently part of a pattern of ‘Odinic’ one- Visby and had an obligation to supply the eyed masks and figures seen inter alia on the king and the castle. New taxes were Sutton Hoo helmet (Price & Mortimer 2014). introduced. Five noblemen controlled Visborg The altered mask appears to have been and Gotland in the 15th century. That century mounted on a post over a cupboard full of was a time of centralization and increased drinking equipment in a Viking Age house at administrative control. By the end of the 15th Hellvi in Gotland. century, Jens Holgersen controlled the Antje Wendt (National Historical Museum, harbors and trade and Gotland became a more Stockholm) spoke about “Viking Age Gold ‘normal’ part of the integrated kingdom. Rings on Gotland”. While silver hoards The first day concluded with a runology represented ‘bank accounts’, accumulations of session. Lisbeth Imer (National Museum of means of payment from trade or plunder that Denmark, Copenhagen) discussed “The Rune could end up buried for religious reasons (or Stone Tradition on Bornholm”. Roughly 260 because they were forgotten after the death of rune stones are known from the island of the owner), gold hoards were specific-purpose Bornholm, dating from the 8th through the 11th money without commercial functions and centuries. Rune stones persist later there than which pattern quite differently in elsewhere in Denmark, with 40 stones dating archaeological finds. Rings had religious from the 11th and 12th centuries, while no functions related to ‘Odin’s law’, temple Bornholm rune stones have been found hoards, and oaths. Gold rings functioned as a contemporary with the “post-conversion rune symbol of the relationship between a chieftain stone boom” (970–1020 AD) in Jutland and and his retainers or between allies in a gift Scania. Swedish practices may have provided society. Gold rings appear to have functioned a model for ornamentation on Bornholm frequently as gifts – as seen in literary runestones. Runic coins and lead amulets are sources. Their design followed that of silver also a prominent part of the Bornholm runic rings. tradition. Ny Björn Gustafsson (Visby) spoke about Continuing from Imer’s presentation, “In Sight or out of Reach? – On the Magnus Källström’s (Swedish National Production of Gotlandic and Non-Gotlandic Heritage Board, Visby) paper “Alike or Not Dress Jewellery at Stora Karlsö in the Viking Alike? – The Writing Traditions on Viking Period”. The site on Stora Karlsö, off the Age Rune-Stones from Öland, Gotland and coast of Gotland, called Stora Förvar [‘The Bornholm” presented insular runic traditions Big Holding’], a cave partially excavated in in a comparative perspective. Traditionally the early 20th century, contains rich cultural these runic traditions have been compared to deposits dating over a 6000–7000 year period. those of neighboring mainland areas rather This is the only location on Gotland where than to each other. A comparison does not molds were found for making both reveal strong patterns shared among all the characteristic Gotlandic and non-Gotlandic islands vis-à-vis the mainland, but there are jewelry. The ‘special’ status of Stora Karlsö some pairwise similarities in which two of the (an island off the coast of an island) as not islands pattern together. belonging to either the Swedish or Gotlandic Thursday evening concluded with a dinner mainland made types of production geared at the restaurant Packhuskällaren, near the toward both markets possible.

131

The final session of the conference was Laila Kitzler Åhfeldt presented “Relations devoted to ongoing and future projects. between the Islands of Austmarr – A Charlotte Hedenstierna-Jonson (Stockholm Research Proposal”. The project would University) described her involvement in a compare the runic traditions of Gotland, large interdisciplinary project in a Öland, and Bornholm. This project would presentation entitled “Introducing the include “interdisciplinary study of runic ATLAS-Project: Late Iron Age and Early inscriptions,” involving archaeological Medieval Mobility around the Baltic perspectives and new forms of technical Reflected through Archaeology, aDNA and analysis as well as text-based runology. Isotopes”. Rapid advances in genetic Several of the contributors had recently taken sequencing techniques have enabled a field trip to examine the Bornholm sequencing of ancient genetic material inscriptions. The presentations by Imer and (aDNA) which was not possible a few years Källström at this seminar represented a pilot ago. In this collaboration, the archaeologists study for the proposed project. help to choose materials to sample (e.g. Finally, Maths Bertell and Kendra Willson individuals to sequence) and ask why they are (University of Turku) discussed “The Future interesting. Is there a match or mismatch of Austmarr Network”, tracing the history and between genetic heritage and displayed goals of the network. The network has held ‘ethnic’ identity in a burial find? Between conferences annually since 2011. An edited biological sex and displayed gender? Genetic volume based on selected contributions to the and osteological analysis may, for instance, first three seminars is well underway, as are help to resolve whether the individuals in plans for larger projects and further books. anomalous inhumation burials may have been For further information about the network and foreigners or unfree. Genes reveal people’s its activities, please see the Austmarr website heritage, population migrations, and sex. at www.austmarr.org and contact Maths Isotope analysis of bones reveals “how you Bertell (maths.bertell[at]miun.se) or Kendra use your body” – diet, geographical context, Willson (kenwil[at]utu.fi). stress, illness, injury, mobility. Funerary archaeology tells what happened when a Works Cited person died. None of these types of data by Baldacchino, Godfrey. 2004. “The Coming of Age of Island-Studies”. Tijdschrift voor Economische en themselves provides ‘the answer’ to the Sociale Geografie 95(3): 272–283. identities of ancient people, but combining Price, Neil & Paul Mortimer 2014. “An Eye for Odin? different techniques and new technologies – Divine Role-Playing in the Age of Sutton Hoo”. makes it possible to explore different pieces European Journal of Archaeology 17(3): 517–538. of the puzzle.

132

PUBLICATIONS Edited Books

New Focus on Retrospective Methods: Resuming Methodological Discussions: Case Studies from Northern Europe Eldar Heide, Bergen University College, and Karen Bek-Pedersen, University of Southern Denmark

A collection of scientific articles edited by Eldar Heide and Karen Bek-Pedersen published as volume 307 in the series Folklore Fellows’ Communications by the Academia Scientiarum Fennica (Helsinki 2014, 230 pages). New Focus on Retrospective Methods focuses specifically on the question of how and to what extent it is possible to use late source material to shed light on earlier periods – for example, the use of folklore recorded during the latter centuries in the study of pre- Christian religion. Nine different case studies presented here explore this question each in their context and each establishing thoughtful and positive, albeit cautious, responses to the central question. This book is one of the results of the first conference of the Retrospective Methods Network held in Bergen in 2010. It is edited by Eldar Heide and Karen Bek-Pedersen, both of whom have extensive backgrounds in the field of Old Norse studies and are also well versed in folklore studies. Contributions to the volume will be briefly outlined here. In “Nordic Folk Legends, Folk Traditions and Grave Mounds”, Terry Gunnell combines theoretical considerations with a practical study of Norwegian traditions from the 19th and 20th centuries. His main point is twofold: first, these late-recorded traditions have a concrete value in the study of pre-Christian Religion”, focuses on the source value of beliefs, and second, those pre-Christian post-medieval folklore for the study of pre- beliefs were themselves a form of folklore – Christian religion. Schjødt’s approach is an observation which effectively reduces the extremely potent: He proposes that all of the gap between early and late traditions to one of material at our disposal is problematic, chronology, not of nature. The implication of whether it is early or late, and therefore Gunnell’s compelling argument is that a objections to late folklore sources are not knowledge of folklore is, in fact, vital to the methodologically sound. In addition, Schjødt understanding of pre-Christian beliefs in the highlights that the early material we have at first place. our disposal, pre-Christian and medieval, is Jens Peter Schjødt, in “Folkloristic problematic and we must therefore maintain a Material and Pre-Christian Scandinavian

133 critical attitude regardless of the period of Her central argument is that it is possible to origin of any given source. anchor post-medieval Icelandic þulur in Old Eldar Heide sets out to outline a new Icelandic literature of the 13th century and at method for reconstructing parts of ancient the same time to posit the existence of an Old beliefs and ideas on the basis of later material Icelandic tradition of popular rigmarole verses in “The Semantic Side of Etymology”. His (of which almost nothing was recorded in examples are linguistic and Heide’s central writing). This sort of research calls for a argument is that the semantic side of meticulous methodology if it is to result in linguistic elements are weightier than the convincing arguments and the article fully etymological side – for the simple reason that lives up to those demands. the semantic aspects represent the active Frog attempts the reconstruction of a meaning of a word at the time when it is used. specific narrative tradition – a tale known in He furthermore argues that reconstructions the Aarne–Thompson–Uther (ATU) folktale that can be supported by independent data are classification system as tale type ATU 1148B more convincing. – in “Germanic Traditions of the Theft of the In “Scandinavian Folk Legends and Thunder-Instrument (ATU 1148B)”. He Icelandic Sagas”, Daniel Sävborg provides an compares the oldest Norse version of the tale, excellent example of how elements from the Þrymskviða, to more recent traditions from Sagas of Icelanders are in some cases much the Finno-Ugric and Baltic language areas, as more easily understood when considered in well as to a much earlier Greek version that relation to folklore material from the 19th and establishing a significant time-depth for the 20th centuries rather than in relation to other narrative. The article is enlightening in terms medieval literary sources. Tradition can, in of its approach to the material and its fact, remain very stable through centuries. In implications for the Old Norse version of the this case, motifs indisputably remain the same narrative in particular. in the 19th and 20th century Norwegian and With “Retrospection of Medieval Landscape Swedish sources as they were in medieval Change in Mid-Sweden”, Hans Antonson Icelandic sources. Sävborg persuasively combines theoretical considerations with two argues that late variants must be descended concrete case studies. The source situation from the same source tradition rather than within the study of agrarian historical based on medieval variants of written saga geography is not to be envied as there are no texts. contemporary sources from the Early and Rudolf Simek, in “Guldgubber and High Middle Ages or earlier. In a practical Retrospective Methods”, suggests an sense, this is obviously problematic to interpretation of the so-called guldgubber landscape studies for these periods, yet it is from the Scandinavian Migration Age based also enlightening in a methodological sense: it on later, medieval iconography. Since there forces researchers to extract what information are no contemporary sources that tell us the they can from the sources available. Here, function of the guldgubber of the pre- Antonson demonstrates exactly how this may Christian era, we must look elsewhere if we be done with the aid of old and modern maps want to understand them. In this article, combined with archaeology and thus once Simek excellently illustrates how evidence again highlighting the value of interdisciplinary from later periods can shed light on what approaches. would otherwise remain impenetrable Janne Saarikivi, in “Reconstruction of questions. Culture and Ethnicity”, employs historical- Yelena Sesselja Helgadóttir, in comparative linguistics to showcase how “Retrospective Methods in Dating Post- reconstruction of cultural and religious history Medieval Rigmarole-Verses from the North can be done in settings where written sources Atlantic”, handles the relationship between from ancient times are entirely lacking, as is post-medieval Icelandic þulur recorded in the the case in Finland. The reconstructive/ 19th century and rigmaroles from across the retrospective methodology described here has North Atlantic and mainland Scandinavia. – out of necessity – become particularly

134 highly developed within the Finnish academic engage rigorous methodological practices tradition, but is in many ways also highly when studying aspects of pre-Christian relevant to Old Norse studies of pre-Christian culture and to consider all relevant sources, be beliefs and traditions. they early or late, with the same critical eye. The book as a whole thus offers a diverse Further information can be found on the range of examples of how retrospective publisher’s website at: reasoning can work and the kinds of results it http://www.folklorefellows.fi/?page_id=2884. can yield. It serves as an invitation to both

The Viking Age in Åland: Insights into Identity and Remnants of Culture Joonas Ahola and Frog, University of Helsinki, and Jenni Lucenius, Åland Board of Antiquities

A collection of scientific articles edited by Joonas Ahola, Frog and Jenni Lucenius published as volume 372 in the series Academiæ Scientiarum Fennicæ, Humaniora by the Academia Scientiarum Fennica (Helsinki 2014, 427 pages). Åland has been a border area or a contact zone between Finnic and Scandinavian cultural and linguistic groups for at least two thousand years. This position has always had a significant impact on how the Ålanders have been seen – and still are seen – by themselves and by others. Nevertheless, numerous cultural features also inevitably emerge that distinguish the Ålandic people from their eastern and western neighbors across the centuries. The chapters that comprise The Viking Age in Åland represent several different scientific disciplines and in many cases are results of collaborative, interdisciplinary efforts. They illustrate and discuss different factors that contribute to the historical identity of the Ålanders and show how multifaceted a concept such as ‘identity’ may be. Major challenges to understanding this fascinating period are, on the one hand, the limitations of the data through which it can be observed and understood, and on the other hand, the challenge of overcoming images that have been constructed of Viking Age Åland by earlier scholarship. Much of that discourse has been inclined to approach thematically central, the chapters of this the Åland Islands in relation to the volume together offer a dynamic and contemporary cultures of Sweden or Finland, multifaceted arrangement of perspectives on perhaps inevitably engaging in the political Åland in the Viking Age. discussions of each scholar’s own times. This volume is one of the outcomes of the Scholars contributing to the present volume Viking Age in Finland project that was attempt to shift attention to Åland as a developed in collaboration with the Åland cultural space in its own right. Even though Board of Antiquities. In its background is a focusing on the Viking Age in Åland, the seminar-workshop held in Mariehamn, Åland, chapters of this volume discuss the concept in the autumn of 2012, but it is not a volume ‘identity’ as a research tool for scrutinizing of conference proceedings in any the prehistoric era more generally. Although conventional sense. Rather than simply the emphasis on ‘identity’ often remains 135 reproducing papers given at that event, those prominent on the Eastern Route, than their initial discussions have led in many cases to eastern neighbours on the Finnish mainland; cooperations and syntheses of research that Ålandis society/-ies seem to have been as was discussed, new explorations for which the affluent as areas of Sweden in this respect. On discussions at the workshop provided a the other hand, the situation in the islands platform, and also the involvement of scholars seems to have changed across the end of the who did not participate in that original event. Viking Age. The paucity of finds in the The resulting contributions to The Viking Age archaeological record combined with the lack in Åland approach questions of how the of continuity in place names gave rise in the Ålanders identified themselves and how they 20th century to a theory that that permanent were identified by the others through diverse inhabitants dispersed or otherwise research materials and a number of disappeared and the islands were largely or disciplines, including archaeology, folklore wholly depopulated. This has been a great studies. geopolitics, history, linguistics, mystery in the islands’ history as well as a osteology, palaeoecology and philology. An significant factor in the discourse surrounding aim of this volume has been to elaborate the it. The apparent discontinuity in language understandings of Åland in the Viking Age between the Viking Age and the medieval from the perspective of different disciplines period leaves open many questions about the and advancing toward a synthesis of such earlier language and culture. The essence of perspectives. Several of the chapters are thus Åland’s Viking Age heritage has thus been the results of close international and obscure and, as such, grounds for speculation interdisciplinary collaborations. In order to especially when Åland’s position and identity interweave the multifaceted discussions have been negotiated in political and brought together in these pages, the authors nationalistic contexts. This situation requires have also cross-referenced other chapters to continuous reassessment of earlier research produce dialogues that resonate throughout and the development of new synthetic and the volume. multidisciplinary approaches with the potential to open some of these questions. Åland and the Viking Age A historical period must be considered in The Viking Age in Åland is intriguing in relation to what preceded and followed it. numerous respects. On the one hand, it was Whereas the changes that followed the Viking situated directly along the so-called Eastern Age seem to have largely erased continuity of Route that connected Scandinavia to the evidence for considering earlier Ålandic inland water routes leading through what is culture and society, the period before the now Russia to the Middle East, and indeed a Viking Age seems to have entailed a radical major sailing route from Birka in fact passed shift in culture and restructuring of society. through the islands where the waters were This is associated with a transformative influx more sheltered. An active presence of of culture from Sweden that began from the Ålanders in those trade networks is reflected second half of the 6th century, presumably in the wealth of finds in the archaeological connected to some process of immigration. record that originate from outside of Åland The cultural transformation of this period from as far as the Orient as well as from definitely contributed to the cultural image of western continental Europe and from trade to Åland, leading Ålandic culture often to be the north. Ålandic ‘export’ is not visible in viewed as a mere extension of Scandinavian the archaeological record, yet cultural transfer culture from that time on. However, as the in clear frmo the spread of a distinctively discussions in this book reveal, Åland appears Ålandic burial rite to an area along the Volga to have developed a distinctive synthetic then inhabited by the Meryans, a Uralic culture that may have been quite versatile. culture. Although Åland did not form a center Identity and identification are complex of trade, the number of Viking Age coin finds concepts and research objects, especially so implies that Ålanders were much more active when inspected in connection with the ancient in the lively trade in silver, especially past. When looking at cultures and their

136 constructions of both their own identities and perspectives are brought together, the result is identities of others, it is essential to take a a fascinating dialogue that can now be carried dynamic view. into the future through new research that Livelihoods in the Åland Islands during the engages with it. Viking Age were predominantly maritime, including fishing, seal hunting and also bird The Organization of the Volume hunting. Farming and animal husbandry were A general introduction to the Viking Age in also practiced, and it is interesting that these Åland by Jan-Erik Tomtlund opens the appear from mortuary evidence to have been volume and the subsequent eleven chapters strongly linked to images that Ålanders are organized in three broad thematic constructed of themselves – they saw sections. Each accompanied by a brief themselves as an agrarian society. editorial introduction to help orient the reader Palaeobotanical evidence in fact reveals to its contents. Tomtlund’s concentrated continuity in agricultural practices on the overview of Åland in the Viking Age is island. This evidence contradicts theories of intended to be easily accessible to non- depopulation and raises significant questions specialist readers, leading them through about what in fact happened with the different thematic areas of culture, drawing population inhabiting the islands across that attention to many of Åland’s peculiarities that period and its identity, and how the lack of are referred to in following chapters. evidence of language continuity should be The first section of the volume, interpreted – or whether perhaps the Interpreting Evidence of the Past, then helps continuity in agriculture reflects long-distance to orient the reader in the history of discourse farming by people who had emigrated but on the Viking Age in Åland, the maintained use of the land and natural interpretations that have been put forward in resources. research and the importance of developing The centrality of maritime activities for the well-founded, interdisciplinary perspectives. sustainability in Åland meant that the means This section opens with Jenni Lucenius’ “In of transportation, which made connections Search of Identities: A Look at Interpretations with neighbouring areas possible, were of the Viking Age on the Åland Islands”. practically available for everyone, a factor Lucenius discusses how interpretations of relevant to consider when looking at Ålanders Åland’s Viking Age have been treated in in trade. Perspectives on material culture politically motivated discussions surrounding reveal contact histories and relations to both the Ålanders’ national identity since the east and west, and although the language or nineteenth century. She highlights the fact languages of the Ålanders of the Viking Age that even scientific discussion has been cannot be identified through direct empirical colored by national idealism. In “The Other evidence, it can also be discussed in relation Island: Kalevalaic Epic and History”, Joonas to these contact histories and the maintenance Ahola then takes up one such interpretation in of communication networks spanning the Finnish research history, identifying Åland Baltic Sea. Indicators of mythology and with a mythic location Saari [‘Island’] of the religion of this culture also open to kalevalaic epic tradition. Ahola deconstructs investigation with the possibility of new this politically charged interpretation and insights. Questions are also raised in this explores the problems of earlier uses of epic volume concerning how the population of as sources for historical events. He then Åland was itself perceived by the populations carries discussion a step further to explore the and polities to both its east and west. potential significations of saari in the oral These internal and external perspectives on poetry tradition and consider whether Åland identities are further complemented by would have been viewed through this mythic situating the islands and their polity or polities image as a frame for understanding the ‘other’ in a broader geopolitical perspective that in from the perspective of groups on the turn reveals both new insights and opens mainland. Per Olof Sjöstrand then makes a different sorts of questions. When all of these powerful and detailed investigation of

137 historical and toponymic research associated of a variety of other types of evidence for with the Viking Age in Åland in “History language in this cultural area. Gone Wrong: Interpretations of the Transition The closing section, Contexts, Contacts from the Viking Age to the Medieval Period and Perceptions, draws together different in Åland”. This tour de force seeks to reassess views on Åland in the Viking Age in order to a number of earlier approaches to questions of contextualize this cultural area in that period cultural continuity on the islands and its of history from different perspectives and alternative depopulation model with a rich explore and elucidate its relationships to other review of scholarship and acute analysis. cultures. The section opens with “Toponymy The second section, Between Sources and and Seafaring: Indications and Implications of Their Lack, leads the reader through Navigation along the Åland Islands”, in which discussions of different types of source Johan Schalin in collaboration with Frog materials as well as the problem that sources discuss the place names that can be are lacking for certain areas of knowledge. considered to have been established already in Rudolf Gustavsson, Jan-Erik Tomtlund, the Viking Age. They argue that the Josefina Kennebjörk and Jan Storå open the continuity of these names may be directly discussion with an impressive review of connected with their association with sailing archaeological evidence linked to a variety of routes and cannot be considered to necessarily practices in “Identities in Transition in Viking be representative of the language(s) of people Age Åland?”. They discuss the dynamics of inhabiting the islands. Several etymologies contacts and historical change in the culture are discussed in detail. Mikko Heikkilä then on the islands. Although Åland is often continues with the discussion of toponymy in conceived as a uniform and coherent space in “The Changing Language Situation in the modern discourse, this contribution highlights Åland Islands and Southwest Finland during the diversity that was present across the the Late Iron Age”. Heikkilä discusses a different islands. Teija Alenius then follows number of etymologies and identifies one with new perspectives from palaeoecology in potential local place name within Åland that “Viking Age Landscapes and Livelihoods in may date to the Viking Age as well as looking the Åland Islands: A Review of the Pollen at place names that seem to be of Evidence”. Alenius shows that there is clear Scandinavian origin within the archipelago continuity of agriculture on the islands that that would seem to date from the same period. has yet to be reconciled with the lack of Lassi Heininen, Jan Storå, Frog and Joonas archaeological evidence of clear settlement Ahola turn attention to the situation of Åland continuity and general lack of continuity in along routes of mobility and in relation to local place names. In “The Finnar in Old other polities in “Geopolitical Perspectives on Norse Sources”, Sirpa Aalto returns to the Åland in the Viking Age”. This discussion topic of perceptions of Åland and Ålanders returns to the question of cultural diversity from other cultures, this time from the with the islands and among the different perspective of Sweden. No Old Norse saga points it addresses is the possibility that Åland literature is known to have referred to Åland was at least at some stage organized into two or the Ålanders, but Aalto develops an distinct polities. The section and the argument which proposes that the Ålanders collection is then brought to a close by Frog may have been perceived from the perspective in “From Mythology to Identity and Imaginal of groups in Sweden as Finnar alongside Experience: An Exploratory Approach to the inhabitants of Finland. The section is brought Symbolic Matrix in Viking Age Åland”. This to a close with “Language(s) of Viking Age article introduces a methodological framework Åland: An Irresolvable Riddle?”, in which for approaching mythology within a historical Joonas Ahola, Frog and Johan Schalin discuss arena like Åland and illustrates this through the problem of evidence of language from the the case study of a distinctively Ålandic ritual perspective of linguistic evidence and of placing a clay model of an animal paw with advance to an exploration of the implications the cremated remains of the deceased.

138

Evolving Perspectives Viking Age meant increasing prosperity that The present volume introduces the topic of seems to have been quite directly attached to Åland in the Viking Age and discusses it from the important trade route that connected the perspectives of a number of different Central Sweden to the river routes in Russia. disciplines with emphasis on questions of This era then concluded with a rapid decline identities. It is difficult to define exactly when in economy and in population when this trade the Viking Age occurred in Åland, the period route lost importance. This period of Åland’s is generally considered to comprise the 9th and rise and fall, discussed in this volume from a 10th centuries AD and to have ended wide array of perspectives, almost coincides sometime during the 11th century. The end of with the beginning and end of the Viking the Viking Age is considered to coincide with Age. The chapters of this volume review the transition from paganism to Christianity. earlier interpretations, present current views, Åland may not have been a significant and also offer exploratory investigations that location overall in the turmoil of the Viking will stimulate future discussions and will Age, yet it was quite central in the framework certainly be of interest to specialist and non- of the northern Baltic Sea. Moreover, the specialist alike. Viking Age was a very significant period in For more information, please visit the Åland’s history. It was preceded by publisher’s site at: significant immigration and followed by http://www.tiedekirja.fi/default/viking-age-in- radical cultural change, but while it lasted, the aland.html.

139

CALLS FOR PAPERS

Would You Like to Submit to RMN Newsletter?

RMN Newsletter in an open-access biannual  PhD project summary publication that sets out to construct an – 2–7 pages body text informational resource and discourse space  MA project summary for researchers of diverse and intersecting – 1–2 page body text disciplines. Its thematic center is the – Places discussion and investigation of cultural  Outline of programmes, projects and other phenomena of different eras and the research activities or research associated with an institution, organization or network of tools and strategies relevant to retrospective organizations methods. Retrospective methods consider – preferred length, 1–5 pages some aspect of culture in one period through – Calls for Papers evidence from another, later period. Such – preferred length, 1–2 pages comparisons range from investigating historical relationships to the utility of The orientation of RMN Newsletter is toward analogical parallels, and from comparisons presenting information about events, people, across centuries to developing working activities, developments and technologies, models for the more immediate traditions and research which is ongoing or has been behind limited sources. RMN Newsletter recently completed. Rather than presenting welcomes and encourages its readership to conclusive findings, short-article engage in this discourse space and it also contributions for the Comments and promotes an awareness that participation will Communications section are generally support, maintain and also shape this oriented to discussion and/or engaging in emergent venue. discourse opened in earlier issued of RMN The publication is organized according to Newsletter or in other publications. four broad sections: Comments and The success of this publication as both a Communications, People, Places and Calls resource and discourse space is dependent on for Papers: the participation of its readership. We also recognize the necessity of opening contact – Comments and Communications with and being aware of the emerging  Short-article (discussion oriented) generation of scholars and welcome – preferred length, 5–15 pages body text summaries of on-going and recently (plus images, tables, list of works cited) completed MA and PhD research projects.

 Conference report / announcement If you are interested in making – preferred length, 2–5 pages information about your own work available  Project announcements – preferred length, 1–5 pages or participating in discussion through – Review Articles and Research Reports comments, responses or short-article  Short article contributions, please send your contributions – 2–10 pages body text in *.doc, *docx or *.rtf format to Frog at – Research and Publication Announcement [email protected].  Published article announcement (abstract / For more information and access to earlier summary of paper or unpublished research) issues of RMN Newsletter, please visit our – 1 page web-page at www.helsinki.fi/folkloristiikka/  Edited volume or monograph summary English/RMN/. – 1–5 pages body text

Published by Folklore Studies of the Department of Philosophy, History, Culture and Art Studies, University of Helsinki 140