Judith Jesch - Ships and Men in the Late Viking Age, The Vocabulary of Runic Inscriptions

Ships and Men in the Late Viking Age: The Vocabulary of Runic Inscriptions and Skaldic Verse Judith Jesch THE BOYDELL PRESS Ships and Men in the Late ...

38 downloads 401 Views 2MB Size

Ships and Men in the Late Viking Age: The Vocabulary of Runic Inscriptions and Skaldic Verse Judith Jesch

THE BOYDELL PRESS

Ships and Men in the Late Viking Age The Vocabulary of Runic Inscriptions and Skaldic Verse

Ships and Men in the Late Viking Age The Vocabulary of Runic Inscriptions and Skaldic Verse

Judith Jesch

THE BOYDELL PRESS

Disclaimer: Some images in the original version of this book are not available for inclusion in the eBook. © Judith Jesch 2001 All Rights Reserved. Except as permitted under current legislation no part of this work may be photocopied, stored in a retrieval system, published, performed in public, adapted, broadcast, transmitted, recorded or reproduced in any form or by any means, without the prior permission of the copyright owner First published 2001 The Boydell Press, Woodbridge ISBN 0 85115 826 9

The Boydell Press is an imprint of Boydell & Brewer Ltd PO Box 9, Woodbridge, Suffolk IP12 3DF, UK and of Boydell & Brewer Inc. PO Box 41026, Rochester, NY 14604–4126, USA website: http://www.boydell.co.uk A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Jesch, Judith, 1954– Ships and men in the late Viking Age : the vocabulary of runic inscriptions and skaldic verse / Judith Jesch. p. cm. Includes bibliographical references and index. ISBN 0–85115–826–9 (alk. paper) 1. Inscriptions, Runic. 2. Viking ships. 3. Sailing – Terminology. 4. Boats and boating – Terminology. 5. Old Norse language – Glossaries, vocabularies, etc. 6. Old Norse poetry – History and criticism. 7. Scalds and scaldic poetry – History and criticism. I. Title. PD2014.J47 2001 623.8'0948'09021 – dc 21 00–066735

This publication is printed on acid-free paper Printed in Great Britain by St Edmundsbury Press Ltd, Bury St Edmunds, Suffolk

Contents List of figures

ix

Acknowledgements

xi

Abbreviations

xiii

1

Introduction: Rocks and Rhymes The Karlevi stone Runic inscriptions, skaldic verse and the late Viking Age Literacy and orality The runic corpus The skaldic corpus Verse in prose contexts Reconstructing viking verse The manuscript transmission Viking verse as a historical source Semantic study of skaldic verse and runic inscriptions Skaldic vocabulary in context Runes and semantics Comparative angles Sources and conventions Ships and men in the late Viking Age

1 1 6 9 12 15 15 18 21 32 33 33 36 38 39 42

2

Viking Activities Vikings víkingr víking Death and war ‘He died’ Battles and raids The fall of warriors Trade Pilgrimage

44 44 44 54 57 57 59 62 63 66

3

Viking Destinations ‘East’ and ‘west’ The western route ‘West’ England Britain and Ireland Further west The European continent and further south Saxony and Frisia Brittany and points south Normandy and southern Italy Africa The eastern route ‘East’ The Baltic area Russia Byzantium and Jerusalem Ingvarr’s expedition Serkland Scandinavia Hedeby Denmark to Sweden Two more towns

69 69 70 70 70 77 78 80 80 83 85 89 89 89 90 95 99 102 104 107 108 113 116

4

Ships and Sailing Words for ‘ship’ skip skeið snekkja dreki kn›rr Oak and pine Miscellaneous words Summary Names of ships The ship and its parts The hull The stems Inside the hull Rudders, oars and shields Masts, sails and rigging In harbour and on land The vocabulary of sailing Description and metaphor Preparing and launching

119 120 120 123 126 127 128 132 134 136 136 137 139 144 150 154 160 166 171 171 171

The ship in the sea Shipwreck and landing

173 178

5

The Crew, the Fleet and Battles at Sea Manning a ship The owner The captain The crew The fleet and the troop lið Compounds with -lið floti leiðangr The troop Units of the fleet Summary Battles at sea Maritime warfare Place and time Preliminaries to battle Bringing the ships together Attack and defence Victory and booty Not like leeks and ale

180 180 180 181 184 187 187 190 195 195 198 202 203 203 203 206 208 210 211 213 215

6

Group and Ethos in War and Trade The group and its vocabulary drengr félagi heimþegi húskarl gildi The ideology of battle ‘He fled not’ ‘He fed eagles, ravens and wolves’ The symbolism of battle: ravens and banners Murder and betrayal Kinds of killing Treachery Loyalty Treachery and politics

216 216 216 232 235 237 239 243 243 247 252 254 254 255 258 261

7

Epilogue: Kings and Ships From vikings to kings Royal and other ships in the eleventh century

266 266 269

After the Viking Age Conclusion

270 275

Works cited

277

Appendix I: The runic corpus

295

Appendix II: The skaldic corpus

301

Index of words and names

317

General index

323

List of Figures 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.1 2.2 2.3 2.4 2.5 2.6 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11

The Karlevi stone. Öl 1. 3 The Karlevi stone. Öl 1. 4 The Karlevi stone. Öl 1. 5 Fragments of the so-called Ældste saga ‘Oldest Saga’ of 23 St Óláfr. NRA 52, fol.2r. Fragment of Fagrskinna. NRA 51, fol.1r. 24 Fragment of Kringla. Lbs frg 82, fol.1r. 25 Fragment of Egils saga. AM 162A G fol, fol.1r. 26 Fragment of Óláfs saga helga. AM 325 VII 4to, fol.2r. 27 Codex Frisianus. AM 45 fol, fol.36v. 30–31 The Fresta stone. U 260. 46 The Växjö stone. Sm 10. 47 The Härlingstorp stone. Vg 61. 55 The Råda stone. Vg 40. 60 The Mervalla stone. Sö 198. 64 Drawing by Aschaneus of the lost Stäket stone. U 605. 67 The Lingsberg stone. U 241. 71 The Yttergärde stone. U 344. 72 Detail of the Nöbbelesholm stone. Sm 101. 74 Map showing places in Britain and Ireland mentioned in the text. 75 Map showing places on the European continent mentioned in the 79 text. The Grinda stone. Sö 166. 81 Map showing places on the eastern route mentioned in the text. 91 The Frugården stone. Vg 181. 93 The Veda rock. U 209. 97 The Lundby stone. Sö 131. 105 Map showing places in Scandinavia mentioned in the text. 108

x 3.12 3.13 3.14 3.15 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 5.1 5.2 5.3 5.4 5.5 5.6 5.7 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9

Figures

Face A of the Hedeby stone. D 1. Face B of the Hedeby stone. D 1. The Forsheda stone. Sm 52. The Sigtuna stone. U 395. The Spånga stone. Sö 164. The Ekeby stone. Ög 68. The dragonhead terminal of the Oseberg ship (reconstruction). Ship graffito from Christchurch Place, Dublin. The stepped stem-post of the Skuldelev 3 ship. The tingl of the Oseberg ship? A frame from the Skuldelev 3 ship. The ship’s vane from Heggen, Buskerud, Norway. The osier ring from the Skuldelev 3 ship. The anchor and chain from the Ladby ship. The Svinnegarn stone. U 778. The Näs rock. U 348. The Kålsta stone. U 668. Detail of the Skåäng stone. Sö 33. Detail of the Karlevi stone. Öl 1. Detail of the Sylten stone. Ög 155. The Mejlby stone. D 117. The Sjörup stone. D 279. The Bjälbo stone. Ög 64. The Landeryd stone. Ög 111. The Århus stone. D 68. The Sigtuna stone. U 379. The Sigtuna stone. U 391. The Törnevalla stone. Ög MÖLM1960:230. The Gripsholm stone. Sö 179. The Braddan stone.

110 111 115 117 121 125 129 142 146 149 152 161 167 169 182 183 191 193 200 200 204 224 228 231 233 240 241 242 246 256

Acknowledgements This book has its origins in, and owes much to the vision of, the legendary ‘realia-project’, bequeathed to me by Peter Foote with the connivance of Else Roesdahl. I hope they will forgive me if it has not turned out quite as they hoped it would. Four other kindly and silver-haired professors have given encouragement, practical support and scholarly advice over the years: Michael Barnes, Anthony Faulkes, the late Christine Fell, and Raymond Page. My runic studies have benefited enormously from the privilege of being a member of the annual international workshop for field runologists. This has provided regular contact with Scandinavian runologists, and the opportunity to study many inscriptions in the company of wiser and more experienced colleagues – their numbers are not inconsiderable, so I shall have to stick to the principle of ‘ingen nevnt, ingen glemt’. I have also examined many runic inscriptions on my own, mainly on trips to Denmark and Sweden in 1993, and to Sweden in 1994, and am grateful to the innumerable ministers, sextons, farmers, landowners, museum curators and other custodians of runic monuments who, sometimes unwittingly, permitted me to study many hundreds of these original documents from the Viking Age. On the skaldic side, I am grateful to all at Det Arnamagnæanske Institut, Copenhagen, and Stofnun Árna Magnússonar, Reykjavík, for their hospitality on extended visits in 1992 and 1993, and shorter visits at other times, and to all at Den Arnamagnæanske Kommissions Ordbog for answering my lexicographical queries, especially Christopher Sanders who bore the brunt of my e-mails. The intermittent Skaldic Study Group has also provided a useful forum for the discussion of skaldic matters back home. Other persons who have kindly answered queries, provided useful information and discussed individual matters, are: Lesley Abrams, Björn Ambrosiani, Michael Angold, Ole Crumlin-Pedersen, Anton Englert, Liesbeth van Houts, Amy Lightfoot, Lena Peterson, Charlotte Roueché, Katrin Thier, Alan Vince and Doreen Waugh. Present and former colleagues in the School (previously Department) of English Studies at the University of Nottingham have all contributed to making a congenial working environment. I would in particular like to mention past and present members of the Old English Lunch Group, and the Norse and Viking Seminar: Jayne Carroll, Paul Cavill, the late Christine Fell, Kathy Holman, Richard Marsden, David Parsons, Betsy Springer, Tania Styles, Philip Tallon and Thorlac Turville-Petre.

xii

Acknowledgements

The research has been supported financially by both The Leverhulme Trust and The Arts and Humanities Research Board. I am grateful to both bodies for their generosity and to Leverhulme for the patience with which it has awaited the result. The project has also benefited indirectly from the interest in and generosity towards Viking Studies at Nottingham of Kungl. Gustav Adolfs Akademin in Uppsala. Both the British Academy and the University of Nottingham have supported my attendance at conferences, while Nottingham has been most generous in its provision of time in which to think. I wish to thank the following institutions for their kind assistance in providing me with photographs, and for granting me the permission to use these photographs in this book: Riksarkivet, Oslo (figs 1.4, 1.5); Landsbókasafn Íslands, Reykjavík (fig. 1.6); Stofnun Árna Magnússonar, Reykjavík (fig. 1.7); Det arnamagnæanske institut, Copenhagen (figs 1.8, 1.9); Antikvarisk-topografiska arkivet, Riksantikvarieämbetet, Stockholm (figs 2.3, 2.6, 3.2, 3.8, 4.2, 5.3, 6.5, 6.6, 6.8); National Museum of Denmark (figs 3.12, 3.13, 6.4, 4.10); Universitetets Oldsaksamling, Oslo (figs 4.3, 4.6, 4.8); National Museum of Ireland (fig. 4.4); Viking Ship Museum, Roskilde (figs 4.5, 4.7, 4.9). Without Tom, and his feline assistants Mayo and Tasso, I might have written this book more quickly. Yet I thank him for attempting to keep my feet on the ground and for constantly reminding me that some things are more important than Viking Studies. Nottingham July 2000

Abbreviations For the abbreviations used in citing individual runic inscriptions and skaldic stanzas, see Appendices I and II. AB AC AEW AnS ASC ASCha BACASSE BT CCS CJW CS CV DR EE Eyrb Fsk FskFJ GA GB GND Hák HANI Hkr IEW ÍO KLNM Knýtl Lat

Adam of Bremen, Gesta Hammaburgensis ecclesiae pontificum. Adémar de Chabannes, Chronique. de Vries, Altnordisches etymologisches Wörterbuch. Falk, ‘Altnordisches Seewesen’. The Anglo-Saxon Chronicle. Anglo-Saxon Charters on the World Wide Web. The British Academy Corpus of Anglo-Saxon Stone Sculpture in England. Bosworth and Toller, An Anglo-Saxon Dictionary. Caithreim Cellachain Caisil. The Victorious Career of Cellachan of Cashel or The Wars between the Irish and the Norsemen in the Middle of the 10th Century. The Chronicle of John of Worcester. Cecaumeni strategicon et incerti scriptoris de officiis regiis libellus. Cleasby et al., An Icelandic–English Dictionary. Danmarks runeindskrifter. Encomium Emmae Reginae. Eyrbyggja saga. Fagrskinna. Fagrskinna. Nóregs kononga tal. Liebermann, Die Gesetze der Angelsachsen. Lindqvist, Gotlands Bildsteine. The Gesta Normannorum Ducum of William of Jumièges, Orderic Vitalis, and Robert of Torigni. Hákonar saga Hákonarsonar. Okasha, Hand-List of Anglo-Saxon Non-Runic Inscriptions. Snorri Sturluson, Heimskringla. Alexander Jóhannesson, Isländisches etymologisches Wörterbuch. Ásgeir Blöndal Magnússon, Íslensk Orðsifjabók. Kulturhistorisk leksikon for nordisk middelalder. Knýtlinga saga. Latin

xiv LegS

Abbreviations

Olafs saga hins helga. Die ‘Legendarische Saga’ über Olaf den Heiligen (Hs. Delagard. saml. nr. 8II). LNT Künzel et al., Lexicon van nederlandse toponiemen tot 1200. LP Finnur Jónsson, Lexicon poeticum antiquæ linguæ septentrionalis. MED Middle English Dictionary. ModE Modern English ModIce Modern Icelandic ms(s) manuscript(s) Msk Morkinskinna. NAVIS NAVIS database. NDEW Falk and Torp, Norwegisch–Dänisches Etymologisches Wörterbuch. NGL Keyser et al. (eds), Norges gamle love indtil 1387. NIyR Norges innskrifter med de yngre runer. NN Kock, Notationes norrœnæ. NR Peterson, Nordiskt runnamnlexikon. Oddr Saga Óláfs Tryggvasonar af Oddr Snorrason Munk. ODu Old Dutch OE Old English OEC Old English Corpus. OEN Old East Norse OGNS Fritzner, Ordbog over det gamle norske sprog. ON Old Norse ONP Degnbol et al., Ordbog over det norrøne prosasprog. Dictionary of Old Norse Prose. Orkn Orkneyinga saga. ÓsH Saga Óláfs konungs hins helga. Den store saga on Olav den hellige. OWN Old West Norse RGA Reallexikon der Germanischen Altertumskunde. SamRun Samnordisk runtextdatabas. Skjd Finnur Jónsson, Den norsk-islandske skjaldedigtning. SnEGylf Snorri Sturluson, Edda: Prologue and Gylfaginning. SnESkskm Snorri Sturluson, Edda: Skáldskaparmál. SR Sveriges runinskrifter. SRR Peterson, Svenskt runordsregister. Sverr Sverris saga. TGT Third Grammatical Treatise. VEPN Parsons et al., The Vocabulary of English Place-Names.

1 Introduction: Rocks and Rhymes . . . language is the archives of history . . . EMERSON

The Karlevi stone One of Scandinavia’s most remarkable monuments from the Viking Age still stands today, in a field west of Karlevi, among the campsites and holiday cottages of the Swedish island of Öland. It was erected around the turn of the last millennium to honour an otherwise obscure Danish warrior and sea-captain called Sibbi. The memorial is in stone, its text is written in runes and formulated partly in verse. This rhyme on a rock is an important piece of contemporary evidence for the Viking Age ethos of masculine achievement and how it was commemorated. The Karlevi monument is a nearly rectangular block of greyish stone, with a rounded top, not quite a metre and a half high, neatly carved on three sides with a long inscription in Danish runes. The inscription falls into two parts, clearly indicated by cross- or hammer-like marks at the beginning of each. From adjacent starting points at the bottom of the stone, both texts proceed boustrophedon (‘as an ox turns in ploughing’) in different directions, the first going off to the right and occupying three lines, the second going off to the left, occupying six lines. The first text explains why the stone is where it is. Although this part of the inscription is damaged, it is possible to reconstruct what it says, the first clause with relative certainty, the second less so (Öl 1, see fig. 1.1):1 s-a... --(s)- (i)-- * satr * ai(f)tir * si(b)(a) * kuþa * sun * fultars * in hons ** liþi * sati * at * u * -ausa-(þ)-... 1

This is not the place to expand on the runological and philological problems posed by this inscription, though one important crux needs to be noted here. The runes kuþa can be interpreted in a number of ways. The word is either an oblique form of the adjective góðr ‘good’ (as suggested here and in Moltke 1985, 320), or of the noun goði ‘(pagan) priest or chieftain’. The runes have also been read as (f)(r)uþa, an oblique form of the adjective fróðr ‘wise’ (SR I, 25).

2

Ships and Men in the Late Viking Age This stone is placed in memory of Sibbi the Good, Foldarr’s son, and his retainer placed on Öland this memorial to honour the dead.

Not content with setting up a permanent memorial to his leader that still stands a millennium later, the unnamed follower also commissioned (or perhaps even composed) a memorial verse to Sibbi, and this verse forms the second part of the inscription (see fig. 1.2): fulkin : likr : hins : fulkþu : flaistr (:)* uisi : þat * maistar * taiþir : tulka þruþar : traukr : i : þaimsi * huki : munat : raiþ : uiþur : raþa : ruk : starkr * i * tanmarku --(n)tils : iarmun ** kruntar : urkrontari : lonti Folginn liggr hinns fylgðu, flestr vissi þat, mestar dæðir dolga Þrúðar draugr í þessu haugi. Munat reið-Viðurr ráða rógstarkr í Danm›rku Endils j›rmungrundar ørgrandari landi. Hidden in this mound lies one, an executor of the goddess of battles [valkyrie→warrior], whom the greatest deeds followed (most knew that). No strife-strong god of the wagon of Endill’s wide ground [sea→ship→captain] will rule land in Denmark more faultlessly.

There are also letters and symbols carved on the back of the stone, though it is not known whether these were carved by the same hand or at the same time as the runic inscription. They are two cross- or hammer-like symbols not unlike those which mark the beginnings of the two parts of the runic inscription, and the roman alphabet letters INONIN . . . EH (see fig. 1.3). The Karlevi stone displays a mixture of cultural influences from all over Scandinavia, and beyond. Its location is now Swedish, but the man being commemorated ruled in Denmark and was probably Danish. ‘Denmark’ in the Viking Age and later extended into what is today southern Sweden, yet the rune-forms and the style of the inscription are closest to those of Denmark proper, rather than those of the ‘Danish’ areas of southern Sweden. The verse adheres strictly to the skaldic metre known as dróttkvætt, associated mostly with Norwegians and Icelanders, even if occasionally composed for Danish and Swedish kings. The poetic diction has parallels in both Old Norse and other Germanic poetry (Olsen 1957; Jansson 1987, 134–6). However, the West Norse appearance of the verse might result from the accidents of literacy (it is possible that dróttkvætt stanzas were composed in Sweden and Denmark even though they were never written down), and some of the linguistic forms in the verse text point to East Norse, for instance the monophthongisation evidenced in huki (WN haugi). Jacobsen and Moltke concluded that the poet and carver (whom they considered to have been the same person) was most likely to have been an Icelander who had learned rune-carving in Denmark or Sweden, and whose

Introduction: Rocks and Rhymes

Image not available

1.1 The Karlevi stone (Öl 1), showing the prose memorial formula in the three lines on the right. Photo: Judith Jesch.

3

4

Ships and Men in the Late Viking Age

Image not available

1.2 The Karlevi stone (Öl 1), showing the six lines of the dróttkvætt stanza. Photo: Judith Jesch.

Introduction: Rocks and Rhymes

Image not available

1.3 The Karlevi stone (Öl 1), showing the roman-alphabet inscription on the reverse. Photo: Judith Jesch.

5

6

Ships and Men in the Late Viking Age

language had been influenced by his stay there (DR, 473). Other explanations are possible, beginning with the likelihood that the rune-carver and the poet were two different people.2 The roman-alphabet inscription, if it is contemporary with the runes, hints at the new literacy that came to Scandinavia with the introduction of Christianity at about the time the monument was erected. Whatever the precise ethnic, cultural and religious mix of commemorated, commissioner, carver and skald, the Karlevi stone epitomises in one monument the late Viking Age habits of commemoration, the deeds for which men were commemorated, and the vocabulary in which this commemoration was formulated, that are the subject of this book. It also brings together the two major bodies of evidence from the late Viking Age that provide the foundation for this study: memorial inscriptions that survive in their original form on rune stones (‘the runic corpus’), and memorial and historical poems that were composed and performed orally but are preserved mostly in later manuscripts (‘the skaldic corpus’). This first chapter is devoted to the definition and exploration of these two corpora, and an explanation of their uses for the study of ships and men in the late Viking Age.

Runic inscriptions, skaldic verse and the late Viking Age The aim of this work is to investigate the vocabulary and phraseology of what might be called the typically ‘viking’ aspects of the late Viking Age. These aspects are loosely defined as: ships and sailing, voyages abroad for both trading and raiding, the organisation and hierarchy of ship’s crews, and the military and social ethos which lay behind or even which resulted from such activities. The ‘late Viking Age’ is defined as c.950–c.1100. I aim to use methods which will recognise and exploit the textuality of the linguistic sources containing the relevant vocabulary and phraseology, but to combine these with a historical concern for relative chronology, regional difference and social context. In other words, this book is an attempt to write history through language. However, it is a work of neither history nor linguistics, but rather of philology in the old sense, of what has been defined as ‘the search for correct historical understanding, built upon a broad linguistic and cultural interpretation of the past’ (Haugen and Thomassen 1990, 36; my translation). It differs from history in its primary orientation towards texts, rather than events, actions or processes, and from linguistics (or ‘philology’ in the new sense) in its emphasis on language in the context of other

2

The word skald occurs in five runic inscriptions as the by-name of either the commissioner of the monument (Vg 4, N 239) or of the rune-carver (U 29, U 532, U 951). This is the usual word for ‘poet’ in OWN (Kreutzer 1977, 118–22) and might seem to suggest a connection between poetry and rune-carving. However, none of these inscriptions is in verse and it is not certain that the runic word skald means ‘poet’.

Introduction: Rocks and Rhymes

7

human activities, rather than language as a system (Haugen and Thomassen 1990, 29–31).3 To enable a ‘correct historical understanding’ of the Viking Age past, the study must be broadly-based and systematic, founded on a relatively large body of material which is considered in its entirety. Historical concerns require this body of material to be datable and localised, and not too chronologically or geographically diverse. At the same time, it is helpful to have some variety within the material to provide an inbuilt control on results and to avoid conclusions that are too simple and unilluminating. These criteria are fulfilled by the choice of two large and contemporaneous bodies of texts: runic inscriptions and skaldic verse from the mid-tenth to the turn of the twelfth century. Both of these are verbal artefacts from the Viking Age, one written, the other originally oral, though surviving in written form. The choice of dates is determined by the need to have a large body of material from a reasonably restricted period of time. Although skaldic verse was probably composed and runic inscriptions were certainly cut before the tenth century, it is only from the mid-tenth century onwards that we have a substantial amount of material in either corpus. In the case of the skaldic corpus, there is the added advantage that the poems dated to this period are easier to understand and are better preserved than the earlier poems, which are also more likely to be spurious. The cut-off point at the turn of the twelfth century coincides with the end of the Viking Age, but is also given by the material itself: there are radical changes in both skaldic verse and runic inscriptions from c.1100, although both continue to be created well after this date. These changes are related to the social, cultural and historical changes that mark the end of the Viking Age at around this time, such as the increasing influence of Christianity, the growth of manuscript literacy and the rise of monarchies. The choice of both skaldic verse and runic inscriptions, rather than just one or the other, is made for a number of reasons. It gives geographical coverage of both East and West Scandinavia, including the colonies in the west. The majority of preserved runic inscriptions from the late Viking Age originate in Sweden and, to a lesser extent, Denmark. Surviving skaldic verse from this period, however, was composed largely by Icelanders, performed mainly in Norway and the colonies, and preserved overwhelmingly in (later) Icelandic manuscripts. The two bodies of material complement each other in other ways. Semantically, runic inscriptions tend to be laconic but precise, whereas skaldic verse is slightly more verbose but possibly less precise in its choice of words. Socially, the two bodies of material represent different contexts. The skaldic poetry studied here emanates from a restricted aristocratic and male world of kings, chieftains and 3

Frank (1997) argues wittily and persuasively that ‘philology’ in the ‘old’ sense is still very much alive; she is indeed a polished practitioner of it. My own experience in Britain is however that ‘philology is frequently understood to mean linguistics, especially historical grammar and the study of past forms of languages’ (Frank 1997, 490, citing the Collins Dictionary, which recommends abandoning the term!).

8

Ships and Men in the Late Viking Age

warriors, while the runic inscriptions reflect a wider cross-section of agrarian and mercantile, as well as military, society, though a few inscriptions undoubtedly represent an aristocratic milieu. But there are also significant areas of overlap between the two bodies of material which make comparisons possible. Both skaldic verse and runic inscriptions have as one of their main functions that of commemoration, of the dead, but also of the living. Runic inscriptions are primarily memorial, in honour of the dead, but implicitly (and sometimes explicitly) also honouring the survivors. Skaldic poetry can praise either the dead or the living. In both cases, this basic verbal function of commemoration is accompanied by artistic embellishment: in skaldic verse this embellishment is poetical, while the rune stones use a mixture of monumentality, iconography and occasionally poetry. Moreover, this interest in the commemoration of (usually) male achievement ensures that there is a significant thematic overlap between the two types of sources which corresponds with the activities being investigated. Both skaldic verse and runic inscriptions are generally neglected as historical sources, sometimes with good reason. Neither can be described as straightforward narratives and their potential for historical study is not at all obvious. Thus, in a survey for historians of Old Norse literature as source material, Vésteinn Ólason (1987, 32, 34–6) mentions both runic inscriptions (very briefly) and skaldic verse (less briefly), but is rather dismissive of their potential as evidence. As vehicles of commemoration, both runic inscriptions and skaldic verse naturally give some basic facts about the activities for which the commemorated is to be remembered, or praised, but these are often hidden in a lot of verbiage in the skaldic verse and are short on telling detail in the runic inscriptions. The overall message is positive in both cases, and may not be true. In this way, both types of text are indeed inferior as sources for historical facts. However, I would argue that it is possible to tease broader historical meanings out of this recalcitrant material by the contextual study of its significant vocabulary, and this is the method I propose to use here. This vocabulary needs to be studied as a broad semantic field, or as a series of overlapping semantic fields, hence the restriction to closely related topics. Some of the words discussed below have been studied quite extensively, but never with a rigorous and systematic focus only on datable texts from a relatively short period of time. It has also not been usual to link a whole range of significant words, or to compare runic and skaldic material, yet these links and comparisons are, I believe, the key to the ‘correct historical understanding’ of this vocabulary.4 Words mean in context, and it is through the study of words in their contexts that we can plot the nuances and changes of meaning that represent changes in both material culture, and in the ideology and social structures of the Viking 4

There are signs of a welcome new interest in skaldic verse and runic inscriptions as sources for social history, though such contributions have hitherto not been sufficiently systematic or well-founded. Thus, Morris (1998) is very cavalier with the language of his sources and makes little mention of the vast scholarly literature on them.

Introduction: Rocks and Rhymes

9

Age. I agree with Page (1993, 150) that ‘in most cases only context can be a safe guide to meaning in a dead language’ and most of this work is an analysis of words in context. The context of any individual word is naturally first and foremost the other words with which it makes up a text. But the wider context in which that text was produced is also an important part of its meaning. It is here that the differences between the runic corpus and the skaldic corpus become apparent.

Literacy and orality The Karlevi stone can be classified in either the runic corpus or the skaldic corpus because it unites, in one monument, oral and literate ways of commemorating the dead. Its memorial formula in prose is paralleled in thousands of other rune stone inscriptions from the Viking Age in which the dead are remembered by their friends or family. This form of literate commemoration is known from Scandinavia both before and after the Viking Age, but the majority of the surviving stones are from the late tenth and the eleventh centuries. On the other hand, the verse part of the inscription could be said to be unique. Though there are verses on other Viking Age rune stones, this is the only one in the metrically strict, stanzaic and syllable-counting form known as dróttkvætt. Of the thousands of dróttkvætt stanzas that do survive, none (other than Karlevi) is recorded in a rune stone inscription, although a few are recorded in runes inscribed on small disposable objects which postdate the Viking Age. Other than these, the surviving dróttkvætt stanzas are recorded predominantly in Icelandic manuscripts, and in the roman alphabet, written in the thirteenth century or later. As will be discussed in more detail below, it is usually assumed that most of these stanzas had an oral prehistory before they came to be written down, and that they were remembered because they had a social and historical function of recording achievement. The Karlevi stanza is thus the only written representative of this genre from a time when it otherwise still had only an oral existence. The Karlevi stone as it stands today is paradigmatic of literacy rather than orality.5 The orality of the past (before the invention of sound recording) is not 5

There is not room here for either discussion or full documentation of the various definitions and theories of orality and literacy. However, some very useful approaches from an anthropological perspective are outlined in Finnegan 1988 (see esp. chs 6 and 8 on interactions between orality and literacy). I have explored some aspects of runic literacy in Jesch 1998a, with references to some studies of literacy in other cultures. A variety of approaches to the study of runic literacy can be found in Nyström 1997. Despite centuries of scholarship on Old Norse-Icelandic manuscripts, relatively little of this is specifically focused on their evidence for literacy. Orality has been a much more popular topic in Old Norse studies, however, such studies tend to focus on Eddic poetry (see Harris 1985, 111–26, for a survey, also Kellogg 1991 for a more recent account also exploring its literacy). Little has been written on the orality of skaldic poetry, despite clear internal evidence (Kreutzer 1977, 148–69) that it was a genre designed for public performance,

10

Ships and Men in the Late Viking Age

directly recoverable, and is only indirectly recoverable in so far as it was represented by literate means. The Karlevi stanza, and all the dróttkvætt stanzas preserved in manuscripts, may be based on oral texts which went before them, but no written text is purely a ‘representation’ of an oral text, a written form of what was once spoken. There is always the possibility that the written text is new, composed in the style of oral texts, rather than a verbatim record of any particular one. And even when the aim was to record the spoken word for posterity, any written text inevitably changes its oral source. These changes cannot be reconstructed, since we do not have the oral text for comparison, but they can be deduced. The words of the oral text may be preserved, but not the pauses, the intonation, the emphasis, the phrasing, or the speaker’s tone of voice. The written text might have features that are counterparts to some features of the oral text. Thus, it can be broken up by a variety of means into shorter segments (‘words’, or longer sequences). But the written text also acquires features not possible in an oral text. A written text can be observed (although not ‘read’) all at once, its length can be judged at a glance, whereas the oral text can only be apprehended sequentially, its length never certain until the end is reached. Most importantly, the written text has permanence in a way the oral text does not. Although oral texts can be remembered and repeated, writing allows utterances to be repeated both frequently and exactly, and to be repeated in other times and other places. Thus the inscribed stone in the field at Karlevi allows endless repetitions of its text in books about runes, or books about the Viking Age, like this one, whereas it is highly unlikely that I would still be repeating its poem if it had been preserved only in the oral tradition. However, orality and literacy are not mutually exclusive categories. While their relative strength, and the nature and extent of their interactions, varied from one period to another, and from one place to another, for most historical periods the two are closely entwined. Certainly purely oral societies once existed (and may still do so), but they are difficult enough for us to imagine in the present, and impossible to reconstruct in the past. Once literacy arrives in any society, it immediately affects oral communication by providing a means by which some aspects of orality can be preserved for posterity. There may not have been a strong desire to write down speech verbatim in the Viking Age and early Middle Ages, but it is clear that many written texts were based at least loosely on oral ones (e.g. laws), and it will be argued below that skaldic poetry was an oral genre which was preserved in a form as near-verbatim as possible, in both the oral and the written tradition. Thus, it is possible to use certain written forms to get an idea of what some oral forms were like. Writing enables the reconstruction of some of the orality of a partially-literate society.

though see the brief comments in Marold 1994 on poetry up to and including Sigvatr, Kuhn 1983, 244–7, and Gade 1995, 224–6, on metrical aspects of performance, and Gade 1995, 21–7, on the saga evidence. I attempt to reconstruct the oral contexts of some poems in Jesch 2000b and Jesch 2001c.

Introduction: Rocks and Rhymes

11

The runic inscriptions of the late Viking Age are a form of primary literacy. The rune stones are original and unique documents, specific to a particular moment in a particular place and to particular people.6 Some are signed by those who carved them, and almost all mention individuals, sometimes relatively large numbers of individuals, by name. It is possible to find traces of orality in some of the inscriptions and the minority of inscriptions which incorporate a verse suggest an earlier, or alternative, form of commemoration of the dead that was oral (Jesch 1998a, 467, 470–71). However, the very fact that these verses can be distinguished from the standard memorial formula shows that they belonged to two different genres. As a whole, the inscriptions do not replicate any particular oral genre. They are a phenomenon of literacy, concrete as well as verbal, and this is reinforced by the monumentality and decoration of the rune stones. The surviving skaldic poetry of the late Viking Age, on the other hand, is a form of secondary literacy. Poems such as that fortuitously recorded on the Karlevi stone were originally composed, delivered and remembered in an oral context. Although orality is ephemeral, human beings have always practised certain types of orality that were designed to transcend ephemerality through memory, with poetry a notable example of this. In the absence of writing to lend permanence to an utterance, the best way of ensuring the repeatability of that utterance was to structure it in such a way that it was as memorable as possible. In dróttkvætt, with its strict metrical forms, we see a full realisation of this urge to memorability. It is also, like the runic inscriptions, often specific to a particular moment in a particular place and to particular people. Skaldic poetry, especially that in the dróttkvætt metre, can thus be seen as striving towards the condition of literacy despite its origins in an oral context. Skaldic and runic memorial are united on the Karlevi stone because they are fundamentally similar. Moreover, in presenting a written version of a skaldic stanza, the Karlevi inscription is a forerunner of later developments, in which the rise of manuscript literacy in the roman alphabet provided a new method of ensuring the permanence of poetic utterances. The verbatim recollection of dróttkvætt stanzas gave way to their verbatim recording in written form. Rune stone inscriptions are primary sources for the study of the Viking Age in the fullest sense. They are contemporary documents, usually unique, that survive in their original form, many of them still in their original location. The monuments have fulfilled their function of giving permanence to utterances and have survived for a thousand years and more, abetted by the durability of stone and in particular of the granite from which so many of them were laboriously fashioned. The written versions of skaldic poems, on the other hand, are not original documents, even when they survive in a unique manuscript. To call them contemporary sources for the Viking Age depends on a process of argument. I

6

There are a few examples of more than one rune stone with the same (or nearly the same) inscription, e.g. the Jarlabanki memorials at Täby (Jansson 1987, 108), but these are exceptional.

12

Ships and Men in the Late Viking Age

believe that argument can be made, and will make it further below. But compared to the runic corpus, the skaldic corpus (with the remarkable exception of Karlevi) comprises imperfect written representations, at a remove of one or generally more centuries, of a potentially ephemeral oral genre, which has undergone both oral and written transmission before reaching the written state in which it survives. The ‘runic corpus’ and the ‘skaldic corpus’, as defined below, were created by a society in which both written and oral forms of commemoration were used, but at a time before the explosion of manuscript (and roman-alphabet) literacy from the twelfth century onwards. Each corpus can be dated to roughly the same period in the late Viking Age, and together they provide our best textual and linguistic evidence for that period.

The runic corpus There are 6,081 inscriptions currently listed in the Scandinavian Runic Database (SamRun).7 Of these, 2,540 are classed as ‘medieval’, i.e. dated to after c.1100, while 155 are from before the Viking Age and written in the older futhark.8 The remaining 3,386 inscriptions belong to the Viking Age, in that they are written in the younger futhark, and are dated to before c.1100. These runic inscriptions from the Viking Age are a various collection. Some might be called casual inscriptions, fleetingly carved on materials that happened to be to hand, sticks of wood, or pieces of bone, although the perishable nature of these materials restricts the number of such inscriptions that survive. Others are slightly less casual, but still often rather haphazardly inscribed on small, functional objects such as tools or containers of various sorts, or personal objects such as combs or brooches. Such objects are important evidence for kinds of literacy in the Viking Age, and the situations in and purposes for which writing was practised (as is also true of similar objects, even more numerous, from the medieval period). However, some of these inscriptions are fragmentary or incomprehensible, and most are minimally informative, at least for the purposes of this study. I have therefore chosen to concentrate on runic inscriptions on stone and with a monumental character, of which there are a little under 3,000 listed in the

7

The database is not a complete record of Scandinavian runic inscriptions, and probably never will be, as new discoveries continue to be made. Some geographical areas (e.g. the British Isles) or types of inscription (e.g. those in the older futhark) are not fully covered at present. This study was based on the version of the database which was made available on the web in February 1998, though I have attempted to update my comments on both the ‘runic corpus’ in general and individual inscriptions using the revised version made available in May 2000. 8 In the Scandinavian context, the word ‘medieval’ refers to the period after the Viking Age, and is normally used in that way in this work.

Introduction: Rocks and Rhymes

13

database.9 Most of these are free-standing, dressed stones which were first carved and then erected in the chosen spot. However, some 82 inscriptions have been carved directly into the living rock, or onto glacial boulders too large to move, wherever there happened to be a suitably large and smooth surface, either vertical or horizontal. While few of the erected stones have travelled far from their original site, with the large boulders immovably embedded in the ground we can be absolutely certain of their original location and orientation. These approximately 3,000 rune stones form what is called ‘the runic corpus’ in this work. The vast majority of them are memorial stones, commissioned by family or friends to honour, or at least remember, the dead, while a few have some other function, such as the Hillersjö rock (U 29), which is also a document of inheritance. Some stand in known burial grounds (e.g. U 356), or form part of a funerary monument (e.g. the early Glavendrup, D 209), while others (such as Karlevi, Öl 1, discussed above) make specific reference to a nearby burial. But, on the whole, few are gravestones in this sense, or at least we can say that most do not have either a visible or a stated connection with the site where the dead person was buried.10 In some cases it is made clear that the dead person was buried elsewhere (e.g. Sm 101). U 135 makes reference to a mound made in honour of the commemorated. However, U 136, which tells us that he died abroad, suggests that this was a cenotaph rather than a grave mound. Many rune stones are now located in or near churches, but it is likely that they were moved there subsequently, from their original places, by the roadside, at a farm boundary, on an assembly-site or at a bridge. The geographical distribution of this corpus is uneven, with the vast majority (over 2,700) located in the territory of present-day Sweden, including some 63 from Skåne. The latter are edited in the Danish corpus (DR) because in the Viking and medieval periods much of southern Sweden belonged to Denmark, and there are just 173 from the territory of modern Denmark (including a few that were found in what is today Germany). Norway has approximately 30 Viking Age inscriptions on what might be described as memorial stones. The Isle

9

The number is difficult to determine exactly because of the slightly varying methods used to identify the inscriptions in SamRun. I have counted all those which are dated to the Viking Age, and for which the database describes the material/object (‘material/föremål’) as ‘stone’, ‘rock’ or ‘glacial boulder’ (‘[run]sten’, ‘berg[vägg/häll]’, ‘flyttblock’) and arrived at a total of 2,961. However, I know for instance that some of the Norwegian memorial stones are not included in this total because the database either does not give full details, or does not use any of the above terms. At the same time, choosing those inscriptions which are described as being on stone (‘sten’) will also bring up non-monumental inscriptions on small objects such as the whetstone G 216, while not all inscriptions on rockfaces (‘berg[vägg/häll]’) are memorial inscriptions. However, the total of 3,000 is a reasonable approximation. 10 A study of earlier records of rune stones suggests that many of them (at least in Uppland) were indeed placed in Viking Age cemeteries, which have disappeared with modern agricultural methods (Gräslund 1986–7, 250–59). However, this still does not demonstrate for certain that those commemorated on these stones were also buried there.

14

Ships and Men in the Late Viking Age

of Man has 31 Viking Age rune stones, and there are smaller numbers from England (four), Scotland (up to eight) and Ireland (one or two).11 Only two of the nine inscriptions known from the Faroes are dated to the Viking Age, and only one of these (FR 1) is a rune stone. While the ‘runic corpus’ is thus overwhelmingly Swedish, much of the Swedish material is formulaic and uninformative. Stones from Denmark in particular can have an importance out of proportion to their number as they can be more informative in their inscriptions. The chronological distribution of the runic corpus is also uneven. The large number of stones from Sweden, particularly the 1,277 from Uppland, skew that distribution towards the eleventh century, especially its second half, which is the period of the great rune stone fashion in that province. Some art-historical studies have even suggested that rune stones were being erected there well into the first half of the twelfth century (Gräslund 1994), while other art historians are reluctant to stretch it quite so far: ‘the fashion of raising decorated runestones had died out certainly before c.1125 and probably before c.1100’ (Fuglesang 1998, 208). Stones from further south in Sweden tend to be somewhat earlier (though still mostly in the eleventh century), those from Denmark earlier still. Denmark in particular has a number of memorial stones from the early Viking Age, or even before the Viking Age, called by Moltke (1985, 150) the ‘transitional inscriptions’. Many of these are in fact in the younger futhark, even if they pre-date the beginning of the Viking Age c.800 (Barnes 1998, 454). But whatever the dating of these inscriptions, they have been excluded as too early for this study concerned with the late Viking Age. The bulk of Danish rune stones are from the ‘great age of runestones’, which Moltke (1985, 184) dates to the period 950–1025 (see also Stoklund 1991 on relative datings within these parameters). The Norwegian corpus, like the Swedish, seems to be concentrated in the eleventh century (see Spurkland 1995 on the problems of dating it). In the British Isles, the few surviving rune stones also seem to originate in the very late tenth or the eleventh century, in so far as they can be dated (Holman 1996, 45, 204–6, 210, 214, 278–9, 290), with the exception of the Isle of Man. Here, the traditional dating is somewhat earlier, c.930–c.1020, based on art-historical criteria. This dating has recently been questioned and a slightly later range of c.950–c.1025 suggested (Holman 1998). The most generous date ranges for the corpus of Viking Age runic memorial stones selected for this study would give a two-hundred-year period of c.930–c.1130.12 Both the earliest and the latest of these dates depend on art11

For the rune stones of England and Scotland, see Holman 1996, 28–43, 201–14, 246–7, 273–80. Most of these are fragmentary and difficult to identify with certainty as Viking Age memorial stones. For the Irish stones, see Barnes et al. 1997, 53–9. 12 The Viking Age rune stones of Bornholm (D 369–72, 374, 376–95, 397–404, 406–9) are assigned a very wide possible date range of 1050–1150 in SamRun. If they were securely dated to the end of this period, this would extend the date range of the whole corpus somewhat. The debate continues on the dating of the mainland Swedish stones, see Gräslund 1998. However, the disagreement between Fuglesang and Gräslund largely concerns the

Introduction: Rocks and Rhymes

15

historical evidence, and are drawn from comparisons with decoration on other objects which are datable from archaeological contexts, as rune stones cannot be dated by scientific methods. However, both the earliest and the latest dates have been questioned (Holman 1998, Fuglesang 1998), and a slightly narrower date range of c.950–c.1110 for the whole corpus would probably accord better with runologists’ datings on the basis of runological, linguistic and to some extent historical criteria.

The skaldic corpus Verse in prose contexts With the exception of the Karlevi stanza, and some runic verses carved on wood from the twelfth century and later (e.g. N 171, see also Liestøl 1964, 25–8; Knirk 1994), skaldic poetry is overwhelmingly recorded in Icelandic manuscripts. We do not know when skaldic verse was first recorded in manuscripts, as contemporary evidence is lacking. Magnus Olsen once suggested (1921, 166) that the last substantial poem in the corpus, Mark I, was delivered by the Icelandic bishop Jón ¡gmundarson to the Danish king Níkulás in both spoken and written form in 1105. This is not implausible, as the poem appears to survive in its entirety (or nearly so) and its contents deal with church politics as well as King Eiríkr’s battles, but I know of no evidence that would confirm this suggestion. If the earliest written records of skaldic poems were just such notations in extenso, they do not survive to tell their tale. Unlike the anonymous Eddic poems, with their mythological, legendary and gnomic matter, most of which survive in a manuscript anthology from the thirteenth century, the poems which are assigned to named poets and which we term ‘skaldic’ are not preserved for their own sake, nor in their entirety, but are recorded in a variety of prose contexts with a variety of functions, usually in the form of individual stanzas or half-stanzas, or small groups of stanzas. Thus there are the occasional verses which embellish the sagas of Icelanders, many of which, scholars agree, may not be as old as they pretend to be, or may not have been composed by the person named as their poet. More convincingly historical are the verses which form the backbone of the kings’ sagas and which are frequently, though not always, cited as authorities for their accounts by the authors of these sagas. Finally, there are the verses cited as examples by Snorri Sturluson in his Edda, particularly the section called Skáldskaparmál, an exposition of poetic diction, and in similar didactic works. Embedded in these prose works, then, is a vast corpus of verse in a variety of relative dating of inscriptions which they would both still place within the eleventh century. None of the stones Gräslund (1994, 125) lists as ‘good examples’ of her Period 5 (1100–1130) is actually used in this study, although they form a part of the corpus as a whole.

16

Ships and Men in the Late Viking Age

metres, but mainly dróttkvætt, purporting to be by a large number of poets, most of whom can be located in a particular chronological and historical context. If this chronology is taken at face value, we have a corpus ranging in date from c.800 up to the time of the writing of the prose works in which it is preserved, i.e. from c.1200 onwards.13 Such a corpus of poetry, composed and preserved in the oral tradition until it came to be written down, would, if correctly dated and therefore antedating other sources, be of unique value for studying early Scandinavia. However, there are two main questions: Are the attributions correct? And, if they are, how well do the written texts of the thirteenth century and later represent verses composed and transmitted orally, in some cases over several centuries? The answers to both of these questions presuppose much work on the whole skaldic corpus which is still being done, and they must therefore be tentative. Some points have however been clarified in a preliminary way. Thus, it is possible to make a distinction between verse that was thought to be historical by the historians and saga-authors of the thirteenth century, and verse which they did not treat as having historical authority, without necessarily implying that it was entirely fictional. This distinction is often made, but the classic statement is by Bjarni Einarsson (1972). The basis for such a distinction lies partly in explicit statements by writers such as Snorri Sturluson on the historical value of old poems (Hkr I, 7; II, 422), and partly in the way such poems were actually used in their texts. A generalisation that is a useful starting point for the study of skaldic verse is that, while some of the poems in the sagas of Icelanders may be as old as they claim to be, they are unprovably so, and any historical study should concentrate on the skaldic verse in the kings’ sagas and Snorri’s Edda (Foote 1978, 57–8). It would be possible to contest the details of such a judgement, but this is a ‘pragmatic distinction’ between verse in the kings’ sagas and the poetic treatises that is ‘on the whole more likely than not to be correctly ascribed’, and verse in sagas of Icelanders and related texts which should be left out of the historical ‘canon’ in the absence of any clear indications as to its date and circumstances of composition (Foote 1984a, 74).14 However, it is also possible to take a genre-based approach as exemplified in Fidjestøl’s Det norrøne fyrstediktet (1982). He defined a corpus of compositions in praise of kings and jarls, regardless of the source in which they are preserved, and which might therefore include sagas of Icelanders and Landnámabók, as well as kings’ sagas and the treatises. For this study, I have found it most useful to refine and combine these two 13

Skaldic verse continued to be composed in the thirteenth century and later, but such texts do not concern us here, as the disjunction between its prose context and the supposedly earlier verse is less significant or non-existent, and many of these poems are preserved as poems, rather than as quotations in prose texts. 14 The ‘poetic treatises’ include the Edda of Snorri Sturluson and the Third and Fourth Grammatical Treatises. In this study, given the further criteria for inclusion outlined below, this means in practice SnESkskm and some stanzas from TGT.

Introduction: Rocks and Rhymes

17

approaches. All stanzas from the kings’ sagas, including lausavísur ‘freestanding verses’, qualify for inclusion in my corpus, even if they are not in dróttkvætt.15 Also qualifying for inclusion are stanzas from Snorri’s Edda and other treatises as long as they are by the poets that are also cited in the kings’ sagas, or appear to be of the same genre (i.e. poetry in praise of kings and jarls in dróttkvætt). The latter distinction can be arbitrary.16 The number of stanzas of poetry in praise of kings and jarls preserved in the sagas of Icelanders is small, and sufficiently similar to the rest of this corpus to warrant inclusion.17 Elsewhere, I have discussed some references to and examples of praise poems for Icelanders and stressed their similarities to the ‘mainstream’ praise poems (Jesch 2000a), but, like Fidjestøl, I do not include them in the corpus under discussion here which is restricted to poems in praise of kings and jarls. However, I make an exception for Arn IV and Arn VII,1, which are both preserved in SnESkskm, and for that reason. The stanzas selected for the ‘skaldic corpus’ here reflect the observations of Fidjestøl (1985, 320–24) that there is very little overlap between the kings’ sagas and the sagas of Icelanders in the stanzas quoted, while Snorri’s Edda and similar treatises represent an intermediate group which overlaps with both.18 Within this selection, there is a further restriction in that only stanzas from poems dated in Skjd to between 950 and 1110 are included, to match the range of the runic corpus. In practice, the starting-point is more like 980, as the poems dated to the three previous decades include the ‘Eddic’-style praise poems such as Eiríksmál, lausavísur by Eyvindr skáldaspillir, and other poems which turn out to contain relatively little of the vocabulary studied here. Nevertheless, these 15

The ‘kings’ sagas’ are defined (in some cases with reservations) as: Ágrip, Fagrskinna, Hákonar saga Ívarssonar, Heimskringla, Jómsvíkinga saga, Knýtlinga saga, Morkinskinna, Óláfs saga helga (various versions), Óláfs saga Tryggvasonar (various versions), and Orkneyinga saga. Some of these texts are (also) included in late compilation manuscripts like Flateyjarbók and Hulda-Hrokkinskinna, which also cite quite a few stanzas not preserved elsewhere (in the case of the latter often going back to an older version of Msk, see Louis-Jensen 1977, 93). Flateyjarbók includes some stanzas which undoubtedly belong more to the sagas of Icelanders than to the kings’ saga tradition. Such stanzas are excluded from the corpus though occasionally cited (see list of ‘other poems cited’ in Appendix II). 16 Thus, two stanzas by one Snæbjr›n preserved in SnESkskm (pp. 38, 81) contain some interesting nautical vocabulary and appear to belong to the praise-poem genre but I have, with some hesitation, left them out because they really are impossible to date, nor is it clear whom they are praising. 17 Four stanzas or part-stanzas (Gunnl I; Gunnl II). I discuss these in Jesch 2001c. Egill Skallagrímsson’s praise poetry for kings and jarls does not quite make it into the corpus on the basis of the datings given in Skjd, although Egill I, at any rate, would fit into the corpus easily, Egill II somewhat less easily (though see Hines 1994–7). It is perhaps significant that the earliest manuscript of an Íslendingasaga is of Egils saga (see below). 18 Which stanzas to include from Snorri’s Edda and which to leave out is necessarily an arbitrary decision when they are by poets not otherwise known, and is based on Skjd’s equally arbitrary dating of such fragments. In any case, many of these, though included in the corpus, have not been used in this study extensively or at all, as can be seen from the index of stanzas in Appendix II.

18

Ships and Men in the Late Viking Age

earlier poems are occasionally cited below.19 The vast majority of the skaldic stanzas studied here, especially those dealing with kings of Norway, in fact fall within the period 995–1103, identified by Krag (1995, 157) as a significant period within the history of Norway. This selection process produces a corpus of 967 stanzas or part-stanzas. While this is numerically smaller than the runic corpus, it is still a substantial body of texts, and one in which the vocabulary and phraseology is much richer and more varied than in the runic corpus. Assuming, then, that these poems from the kings’ sagas and the treatises include a sufficient number of correctly-attributed stanzas to validate generalisations about the corpus, how can we tell that the recorded versions give a fair impression of these originally oral compositions? Again, we can never be sure, but many scholars have thought this possible, even likely. Skaldic verse is characterised by complex metrical rules applied within a small poetic space: any changes to the text may violate one or another of these metrical rules. This does not mean that the text is immutably fixed, only that, as Snorri himself recognised, kvæðin þykkja mér sízt ór staði fœrð, ef þau eru rétt kveðin ok skynsamliga upp tekin ‘the poems seem to me least likely to be corrupted [i.e. compared to other evidence, both oral and prose], as long as they are correctly composed and carefully interpreted’ (Hkr I, 7). While this is not a cast-iron guarantee, it does mean that our best chance of reconstructing Viking Age verse lies with skaldic verse rather than, for instance, Eddic verse with its looser structures. Not all skaldic verses are necessarily what they claim to be, or what the prose authors claim them to be. Nevertheless, it is clear that, taken as a body, they provide a useful source for the study of the Viking Age and beyond. Also, the difficulties of the verse and its preservation are not evenly distributed. While it would be a brave scholar who tried to make statements about the ninth, or even the tenth, century on the basis of verse purporting to be from that period, by the eleventh century the verse is copious and generally well-preserved, and is suitable material for comparison with the runic corpus. Reconstructing viking verse However, an essential prerequisite to the use of such poems as sources for the late Viking Age is an understanding of their textual transmission and therefore of the extent to which it is possible to reconstruct them from their recorded forms. While some small progress has been made in establishing a model for distinguishing the reliably ancient verse preserved in thirteenth-century Icelandic prose texts, there has been less progress in setting out the principles for reconstructing such verse from the available sources. Finnur Jónsson published a corpus in 1912–15 (Skjd), with both diplomatic 19

All poems assigned to the ‘skaldic corpus’ are listed in Appendix II. The index of stanzas there indicates which stanzas were actually used in this study.

Introduction: Rocks and Rhymes

19

and edited texts of all known skaldic verse dated to the period 800–1400, organised chronologically by poet. Since many of the longer praise poems are preserved in more or less dismembered form, as individual stanzas scattered among their prose contexts, Finnur had to reconstruct these longer poems, assign some unidentified stanzas to named poets and occasionally choose between poets where there were two or more attributions for a stanza. Thus, while the ‘B’ volumes of the corpus, containing Finnur’s edited (and often emended) texts with Danish translations at the bottom of the page, are obviously artificial constructs, even the ‘A’ volumes, the diplomatic texts taken from the manuscripts, show heavy editorial selection and reconstruction of poems and the work of individual poets. In the ‘A’ volumes, Finnur chose a ‘best text’ for each stanza and provided variants from other manuscripts and texts in the apparatus (though this is by no means complete, nor always consistent). Thus, any one of the longer poems in his edition can have been put together from individual stanzas preserved in a variety of texts and manuscripts. In the ‘B’ texts, Finnur does not always follow the ‘best text’ of his ‘A’ version, but appears to choose freely and eclectically from the variants, and to emend on occasion. Although his procedure is not spelled out anywhere, it is clear that he used metrical and stylistic criteria as well as following the presumed stemmatic relationships of the manuscripts. Finnur Jónsson’s corpus is thus a curious combination of editorial approaches: underlying the choice of ‘best texts’ is a set of assumptions about the manuscript relationships of the prose texts in which skaldic verse is preserved (Finnur was also an editor of several of these), but the reconstruction of individual poems and of the work of individual poets is an exercise in textual criticism that attempts to go back beyond these prose sources to the ‘original’ poetic compositions. It is exactly this relationship between prose context and verse source that is the key to the editing of skaldic verse and which needs much more discussion, not to mention theoretical underpinning. Many of Finnur’s editorial decisions were criticised by Ernst A. Kock in his monumental Notationes norroenae (NN), for reasons outlined in Kock 1938. Kock’s approach can best be described as textimmanent: any variant can be grist to his mill as long as it fits in with his metrical and stylistic criteria. Kock’s approach is fairly systematic, based on close knowledge of the whole corpus, but in the end his texts are even more idealised than Finnur’s. Kock bends all stanzas alike to his method, which ‘must’ be ‘conservative’, ‘comparative’ and ‘in accordance with linguistic common sense’ (Kock 1938, 134; my translation), while accepting all of Finnur’s reconstructions of poems and identifications of poets despite criticising the details of the edited texts. Bjarne Fidjestøl (1982) attempted a critical revaluation of Finnur’s criteria for reconstructing the praise poems for kings and chieftains (what he calls ‘fyrstedikt’), based on specific statements in the prose texts which help to identify the constituent stanzas of a poem, and the order of stanzas in the prose texts which can help to reconstruct their order in the original poem. Fidjestøl’s emphasis on the prose contexts means that his book is only the first stage in a revaluation of the corpus, and it concentrates on the poems as larger units rather than on the

20

Ships and Men in the Late Viking Age

micro-problems of reconstructing individual stanzas, though he also discusses these sporadically. Russell Poole (1991), on the other hand, approached the reconstruction of individual poems from internal criteria rather than from the external witness of the prose context. He identified stylistic devices such as the use of the present tense, or concatenation, and used them to group stanzas and to create, in some cases, previously unrecognised poems. However, without some methodological discussion about the appropriate methods for reconstructing skaldic poems independently of their prose contexts, there is the danger of a circular argument. For instance, while Poole saw the use of the present tense as an ‘accepted stylism’ that has the effect of ‘running commentary’, he assumed that the basic mode of skaldic poetry is narrative, and that those poems which use the present tense to indicate the ‘main march of events’ constitute a ‘special genre’ (1991, 24–56, 195). However, for eleventh-century praise poetry at least, I have argued that the attempt to reconstruct the original poem as far as possible, and to locate it in its original historical context, can be more fruitful an approach to explaining the poet’s use of tenses (Jesch 2000b). Individual skaldic stanzas have also been edited as a part of the prose texts in which they are preserved. Here, one might think the question of reconstruction was of less significance, since the job of the editor is presumably to arrive at the text of the poem as included in the prose source, rather than at its hypothetical ‘original’. And indeed, most editors try to avoid using variants from other text traditions in their reconstructions of the verses in their prose texts, but this is not always possible. So they turn to metrical, grammatical, lexical, stylistic, or other criteria to reconstruct the verse texts, and such criteria may indicate the choice of a variant from another prose text containing the verse in question. In Hkr, for instance, Bjarni Aðalbjarnarson generally keeps the readings of Kringla and its copies whenever they make sense, but occasionally departs from these even when others have thought it possible to construe them (for some examples, see Jesch 1994–7, 3–5). One of the few theoretical statements on the editing of skaldic texts is Poole’s study of Egill II. While less concerned with the poem performed by Egill himself in York in the tenth century than, for instance, Hines (1994–7), Poole nevertheless outlines an editorial method that would be a good basis for editing skaldic texts for those with a historical interest (even though he sees the role of the ‘theorist of textual criticism’ primarily as aiding the ‘literary interpretation of a text’). He makes two important points that future editors should take to heart (Poole 1993, 80–81). The first is that not all texts are alike, and ‘text-critical practice must accommodate itself to the particular genre and sometimes even to the particular work that it is currently scrutinizing’. Secondly, he stresses the importance of giving the reader ‘access to the full paradosis’, that is to all the variants in all the texts. To adapt his theory to the needs of historical study implies that, since not all skaldic verse is as ‘contaminated’ as the versions of Egill II, ‘the rigour of stemmatic recension’ might be appropriate for other poetry even if it is not for Egill’s ‘wild’ poem. Attention to the requirements of an

Introduction: Rocks and Rhymes

21

individual text means that it will often be possible to use textual criticism to ‘reconstruct’ a Viking Age poem, particularly if it survives in several early manuscripts, or copies of early manuscripts. Such a reconstruction of a Viking Age poem should be accompanied by full information on what has been rejected in that process of reconstruction. Knowledge of the changes that have been made to texts in the course of transmission is, paradoxically, an important aid to understanding what was, and what was not, possible, or likely, in those texts in their earlier, Viking Age, state. The manuscript transmission For any one skaldic stanza there is normally a gap of at least two hundred years between its supposed date of composition and the earliest surviving written version. The oldest surviving manuscript containing skaldic verse is AM 673 b 4to (c.1200), containing Plácitus drápa, not a part of the corpus being studied here (edited by Jonna Louis-Jensen in Tucker 1998, 89–124). The two oldest manuscripts containing verse from the ‘skaldic corpus’ studied here have been dated to c.1225.20 Ágrip (AM 325 II 4to), a brief history of the kings of Norway, contains seven stanzas or part-stanzas, three of which are not recorded elsewhere (Oddm; AnonX I,B,4; Sjórs 1). The first of these, if correctly attributed, is too early for this study, the third is similarly too late. The other four stanzas are also found in other manuscripts (Sigv XI,12; Sigv XIII,28; AnonXI Lv,11; SteigÞ), but Ágrip is their earliest surviving written record.21 The fragments of the so-called Ældste saga ‘Oldest Saga’ of St Óláfr (NRA 52) contain eight stanzas, in whole or in part (Bersi II [here attributed to Óttarr]; Þorm II,10–11,15; Ótt II,1; Ótt IV,1; Hár 2; Þorf 1; see fig. 1.4). All of these are also recorded in other sagas of St Óláfr (notably LegS, with which these fragments have a close relationship). Thus, only thirteen stanzas of the ‘skaldic corpus’ studied here are recorded as early as 1225, which is 150–200 years after they were first composed. However, a large number of manuscripts from later on in the thirteenth century contain skaldic stanzas. Dated to c.1225–50 is the ‘Legendary Saga’ of St Óláfr (DG 8) citing 63 skaldic stanzas (listed in LegS, 238–43), including some not recorded elsewhere. Several manuscripts from around the middle of the thirteenth century are only fragments of texts which survive complete in later copies or versions. Those containing skaldic verse are the earliest (Norwegian) manuscript of Fsk (NRA 51, see fig. 1.5) from c.1240–63 and the single leaf that remains of the Hkr-manuscript known as Kringla (Lbs frg 82, see fig. 1.6) from 20

All manuscript datings are taken from the first volume (Registre/Indices, 1989) of ONP. This volume can also be used to trace the manuscript sigla used below, along with editions and facsimiles of the texts. 21 Sjórs 1, if correctly attributed, would have been composed only a little over a century before the writing of the manuscript. The other stanzas are 150–200 years older than the manuscript.

22

Ships and Men in the Late Viking Age

c.1258–64.22 The earliest manuscript of an Íslendingasaga with skaldic verse is a fragment of Egils saga from c.1250 (AM 162A G fol, see fig. 1.7). From some time in the second half of the thirteenth century (c.1250–1300), are both the main manuscript (Holm perg 2 4to, containing 212 stanzas) and two fragmentary manuscripts (AM 325 VII, 4to [see fig. 1.8] and AM 325 XI 2 e 4to) of ÓsH. From the third quarter of the thirteenth century are two manuscripts of Oddr, AM 310 4to (c.1250–75) and the fragmentary DG 4–7 (c.1270), as is Msk (GKS 1009 fol., c.1275), citing 263 stanzas and source of many not known from elsewhere. From the fourth quarter of the thirteenth century (c.1275–1300) are a fragment of Egils saga (AM 162A C fol), a version of Jómsvíkinga saga (AM 291 4to) and another fragment of ÓsH (AM 325 XI 2 n 4to). There is thus a substantial amount of verse that survives in manuscripts written before 1300 (and much more, of course, in even later manuscripts). Between them, the main manuscript of ÓsH (Holm Perg 2 4to) and Msk (GKS 1009 fol.) cite 475 skaldic stanzas, with little overlap between them. Allowances have to be made for the fact that some of their stanzas are either too early or too late for the corpus, and for some overlap with the stanzas that survive in other early manuscripts. But adding on the stanzas preserved in the other early manuscripts, I would estimate conservatively that something like half the stanzas in my corpus of 967 survive in manuscripts written before 1300. Thus, although Kuhn (1983, 256) notes that there are very few thirteenth-century manuscripts containing dróttkvætt stanzas, the number of stanzas recorded in that century is not inconsiderable. However, there is no stanza in the corpus for which the gap between its supposed date of composition and the earliest surviving written version is less than 150 years. Not one of the manuscripts listed above is thought to be an archetype, that is the first written version of a text. Even Plácitus drápa is thought to have been composed up to fifty (or even more) years before the surviving manuscript (Tucker 1998, xcix–cii). All are copies of earlier manuscripts, and many are copies of copies. Thus, the gap between the composition and the recording of skaldic verse does not depend entirely on a century or two of oral transmission. It is in part a gap in the written transmission, as we simply no longer have the earliest manuscripts in which skaldic verse was recorded. This means that textcritical reconstruction can narrow the chronological gap between composition and recording, sometimes substantially. However, the precise combination of oral and written transmission involved in the preservation of skaldic verse can be difficult to reconstruct. Scholars have some idea of the relative datings of the manuscripts concerned, and various attempts have been made to reconstruct the influences of one text (if not necessarily of one manuscript) on another (see for instance the complex diagram in 22

The rest of Kringla burned in the Copenhagen fire of 1728, but can be reconstructed from the late seventeenth- and early eighteenth-century copies made of it (Stefán Karlsson 1976, 8–13).

Introduction: Rocks and Rhymes

Image not available

1.4 Fragments of the so-called Ældste saga ‘Oldest Saga’ of St Óláfr (NRA 52, fol. 2r), with Þorm II,10 beginning in the second line of the top fragments and Þorm II,11 beginning at the very end of the third line of the middle fragments. Photo: Riksarkivet, Oslo.

23

24

1.5 Fragment of Fagrskinna (NRA 51, 1r), showing ÞSkall 2 beginning in the middle of the second line from the bottom. Photo: Riksarkivet, Oslo.

Ships and Men in the Late Viking Age

Image not available

Introduction: Rocks and Rhymes

25

Image not available

1.6 Fragment of Kringla (Lbs frg 82, 1r), showing Ótt III,11 and ÞSjár III beginning in lines 13 and 18 of the lefthand column. Photo: Landsbókasafn Íslands, Reykjavík.

Image not available

1.7 Fragment of Egils saga (AM 162A G fol, 1r), with the large initials in the bottom left margin showing the beginnings of Egill VII,14–15. Photo: Stofnun Árna Magnússonar, Reykjavík.

Introduction: Rocks and Rhymes

27

Image not available

1.8 Fragment of Óláfs saga helga (AM 325 VII 4to, 2r), showing Ótt II,7–11 and Sigv I,6–12. Photo: Det Arnamagnæanske Institut, Copenhagen.

28

Ships and Men in the Late Viking Age

Fidjestøl 1982, 10, summarising earlier researches). However, it is not clear that it is possible automatically to apply any such stemma to the verses contained in a manuscript. In a brief outline of what a new edition of skaldic poetry should be like, Jón Helgason (1947, 130–31) assumed that the stemma of the prose work was more important, though not necessarily decisive. Editors, whether of the works of individual poets, or of the prose works containing skaldic stanzas, pay lip service to the fact that ‘it cannot necessarily be assumed in every case that the transmission of verses follows an identical path to that of the prose’ (Whaley 1998, 5), but it is an aspect of manuscript transmission that deserves further investigation. Scribes appear to have been more likely to ‘correct’ their exemplars in verse quotations than in the surrounding prose. But did they do this because they knew the rules of skaldic verse and attempted, rather as modern editors do, to reconstruct the verse? Or did they do it because the verses were still known in either the oral or the written tradition outside their prose contexts, and were therefore available as sources for such correction? The answer to this question is likely to have been different at different times, also in different places and for different scribes. Thus, Jonna Louis-Jensen has pointed out (1977, 152–5) that the compiler of Hulda-Hrokkinskinna attempted to improve the already corrupt texts of the stanzas in his exemplar but, as his knowledge of skaldic diction and metre was minimal, his ‘corrections’ are no more than superficial associations that sound good but are often meaningless. In other words, he was an incompetent editor of skaldic verse. The scribe of Kringla, on the other hand, has been described as ‘very perceptive’ and is suspected of ‘improving’ the verses of his exemplar extensively (Hkr III, xcv; my translation), perhaps the very model of the modern skaldic editor.23 In this case, we will not know for sure until there is a proper critical edition of Heimskringla (Vésteinn Ólason 1988, 132). Anyone who has worked with skaldic poetry in its manuscript contexts soon gets a sense that oral as well as written transmission is involved. One example will illustrate this. Codex Frisianus (AM 45 fol) is a manuscript dated to the first quarter of the fourteenth century containing a text of Hkr. On fol. 36v, the scribe started to write Hókr 8, which begins Hjalmfaldinn bar hilmi. Instead, he wrote ‘Hialm falldínn hloð alldom’. Realising his mistake, he crossed out the last two words and continued writing the correct ones ‘bar hílmí’ (see fig. 1.9). The line that first suggested itself to his mind may have been the third line of Bbreiðv 6: jarnfaldinn hlóð ›ldum, a stanza not in the corpus as it is preserved in an Íslendingasaga (Eyrb, 111). The first half of this stanza recalls Bj›rn’s youthful adventures with the Jómsvíkingar and in Sweden (compare Eyrb, ch. 29) in terms of the conventions of praise poetry (jarnfaldinn hlóð ›ldum / Eirekr í dyn geira ‘iron-enclosed [i.e. helmeted] Eiríkr [a Swedish king] piled up men [i.e. bodies]

23

Although Kringla is almost wholly lost (see previous note), it survives in a number of good copies which have not yet been fully utilised in editions (Vésteinn Ólason 1988, 136; Jesch 1998b, 112).

Introduction: Rocks and Rhymes

29

in the din of spears [battle]’), and contrasts them with his experiences on an Icelandic heath in winter, described in the second half of the stanza. This kind of pastiche depends on the standard praise poem genre being well-known. It is not possible to decide, on the basis of one line, whether the scribe of Codex Frisianus knew Bbreiðv 6 or not, but his momentary slip of the pen suggests that he had lines of skaldic verse in his head as well as in his exemplar. This suggests that at least the pastiches, and probably also the original praise poems, were still known in Iceland in the fourteenth century. The oral circulation of skaldic poetry is even more likely in the previous century, as has been demonstrated by Gísli Sigurðsson (2000) in a study of TGT, at least for the west of Iceland. He shows that Óláfr Þórðarson must have known stanzas from the oral tradition as well as from the latest written works of his time, not least because two-thirds of his examples are not known from any other source. As Fidjestøl has also shown (1985), the very fact that there is so little overlap between the stanzas cited in the poetic treatises and the kings’ sagas suggests the depth of the skaldic tradition in Iceland at this time. Authors or compilers of both types of texts had a vast treasure trove of skaldic stanzas to hand, from which they could pick and choose the stanzas they needed for their particular purposes. This treasure trove was both written and oral, although unravelling their relative proportions in any detail is difficult and few attempts are totally convincing. The question of oral versus written transmission was discussed at length by Fidjestøl, with rather uncertain results (1982, 45–60). He found a high degree of stability in the corpus, with relatively few examples that could certainly be ascribed to ‘munnleg variasjon’, variant words or lines that have entered a poem in the course of oral transmission and which result in variant readings in the written tradition when we have more than one version of a stanza. At the same time there is certainly variation that seems to have arisen in the course of written transmission, but which cannot be dismissed as mere copying errors. On the whole, it can be difficult to distinguish between ‘oral’ and ‘written’ variation, and, while the former is likely, the latter is certain, and the two often go together (Fidjestøl 1982, 51). Fidjestøl’s analysis is based on a range of manuscripts from different times, and it is hardly unexpected to find textual variation in manuscripts that we know were copies, or copies of copies. Moreover, the evidence of two runic inscriptions with poetic texts that appear to be variants of poems preserved in the Icelandic manuscript tradition suggests that the inscriptions were copied from an oral rather than a written exemplar (Seim 1986, 36; see also Knirk 1994, who is more cautious in his conclusions). At any rate, the runic texts are significantly different from the manuscript versions in both cases. Thus it remains difficult to say how accurately the poems recorded in Icelandic manuscripts of 1225 and later represent the poems that were actually composed and performed in, say, the eleventh century, short of any dramatic discoveries of lost manuscripts. There is scope for further work in the close study of the earliest manuscripts (though also of all the Hkr-manuscripts), and in the

30

Ships and Men in the Late Viking Age

Image not available

31

1.9 Codex Frisianus (AM 45 fol. 36v), showing the scribe’s correction in Hókr 8, nine lines from the bottom of the right-hand column. Photo: Det Arnamagnæanske Institut, Copenhagen.

Introduction: Rocks and Rhymes

Image not available

32

Ships and Men in the Late Viking Age

comparison of poetry in the surviving manuscripts with the stanzas or poetic quotations inscribed in runes on datable objects found in places like Bergen and Trondheim (e.g. Liestøl 1964, 19; Seim 1986; Knirk 1994). As these inscriptions are rarely earlier, and then only slightly, than our earliest manuscripts, the Karlevi stanza remains our most important link with the eleventh century. Nevertheless, it is my impression that the difficulties of reconstructing Viking Age verse can be exaggerated and that, overall, the manuscript tradition provides a fair representation of what that verse was like. One of the benefits of a global study such as this one is that it lessens the burden of proof that need be placed on any one stanza or poem, by considering it in the light of a much more extensive skaldic tradition. There is safety in numbers and where a word can be shown to have the same, or a developing, meaning across a number of texts, we need not rely on just one of those texts to demonstrate that meaning. Quantity of material can cancel out uncertainties about the evidence of any one part of that material. This is where I believe the semantic study of skaldic verse can make a contribution when the mining of the corpus for factual information has exhausted the available seams. Viking verse as a historical source The value of skaldic verse as a historical source is frequently asserted. There is also a long list of historical studies of skaldic verse, occasionally reaching somewhat negative conclusions about its value (e.g. Ashdown 1930; Bugge 1910; Campbell 1971; Johnsen 1916; Moberg 1941; Peters 1978; Poole 1987; Poole 1991, 86–115). Many of these studies, explicitly or implicitly, interpret the verse in tandem with, or heavily influenced by, their surrounding prose contexts, or with other documents, especially some well-worn English ones, that record the same events. The verses are studied with avowedly ‘historical’ aims, to squeeze information on events and actions from them. There have however been few attempts to study the whole body of skaldic verse as a source in its own right and with its own characteristics, with an open mind about the kind of information that might be gleaned from it (exceptions include Fidjestøl 1993; Foote 1978, 1984; Malmros 1985). Admittedly, the corpus, as defined above, although substantial and datable, does have its limitations. The subjects that interested the skalds are few and predictable: war, sailing and remuneration predominate, with topics like love and more mundane matters largely restricted to the possibly spurious verses outside the corpus, and not especially frequent even there. The scholar searching for enlightenment about viking activities before 1100 will find that only certain aspects of that activity are illuminated, directly or indirectly, by skaldic verse. As well as being circumscribed in subject-matter, skaldic poetry is an art-form that does not give much away, with its brilliant formalism drawing attention from the message to the medium. In any case the message is generally cliché-ridden, repetitive and often deceitful. Even where skaldic verse attempts narrativity, its main tactic is to link a series of set-pieces (a classic example is Sigv I). But it is

Introduction: Rocks and Rhymes

33

possible to make a virtue of this skaldic failing. The elaborate static quality of skaldic verse is both cause and effect of a largely nominal style, a discourse that is rich in nouns and adjectives. The vocabulary of skaldic verse, with its many synonyms and near-synonyms for both ships and men, deserves particular attention.

Semantic study of skaldic verse and runic inscriptions Skaldic vocabulary in context Some general principles are needed before the vocabulary of skaldic verse can receive this attention. The principles I suggest here arise out of my work on the vocabulary that is analysed at length in the following chapters. Some of this vocabulary has been extensively studied before, but most of these studies are etymological, comparatist and, often, decontextualised in approach. I would argue that there is much to be gained from a recontextualisation of semantic studies and from a more precise application of the comparative method. As for etymology, I tend to agree with Fell (1986, 295) that it ‘may obfuscate discussion of meaning and translation’. The importance of context in the study of older languages has been regularly stressed over the last forty or so years and, indeed, goes back to the totalising methods of the ‘old’ philology, which involved ‘scrutinizing words by using all clues given by the texts under investigation’ (Nickel 1966, 37; similarly Schabram 1970). This scrutiny, which is ‘operational’, another word for ‘contextual’, involves procedures which ‘closely study occurrences of signifiers, investigating particularly their frequency, distribution, collocability, and context’ (Nickel 1966, 37–8). While etymology is less significant, it cannot entirely be ignored, as ‘etymologically related lexemes are still semantically linked on a synchronic level’ (Nickel 1966, 40). Definition takes place ‘by means of typical conditions of usage’, and a ‘lexicographical definition which includes such typical conditions of usage can give very much more information than the standard dictionaries’ (Strauss 1985, 574, 577). I have attempted to apply these linguistic methods to the study of skaldic vocabulary, which is why I have placed so much emphasis on a global study of the corpus, and why this book is arranged as a narrative rather than as a dictionary. A dictionary also suffers from its alphabetical arrangement, which separates words of similar meaning, and sometimes even related words. A different arrangement, in which words are grouped according to semantic fields, recognises the ‘basic premise . . . that to understand lexical meaning, it is necessary to look at semantically related words, not simply at each word in isolation’ (Lehrer 1985, 283). While each word has its own history, these histories are interrelated. Even in the relatively short historical period being studied here it is possible to see the rise and fall of words relative to each other, and sometimes quite substantial changes in the meaning of individual words as other words muscle in on their semantic field.

34

Ships and Men in the Late Viking Age

I have set out the methodological requirements for a semantic study of skaldic verse elsewhere (Jesch 1993b), and here it is sufficient to summarise them. Two are especially important: the recognition of the special nature of skaldic diction, and a close attention to all possible contexts. The vocabulary of skaldic verse can operate on one of three levels. Many words, even in the most intricate skaldic stanzas, will have a straightforward denotative meaning, that is their usual meaning as in everyday speech or prose. There is very little contemporary comparative material to determine this meaning, however, although runic inscriptions are an important source. Secondly, other words will have a denotation that is limited to their use in poetry (these are usually called heiti). We can deduce this from the fact that the skalds had so many synonyms for a few things, but also because such specifically poetic usages can be demonstrated elsewhere in Old Norse, or indeed in the other Germanic languages. These first two levels can sometimes be difficult to distinguish. The word ekkja demonstrably means ‘widow’ in other Old Norse texts, but when we come across it in a skaldic verse, we need to decide whether it may not instead be used in the more generalised poetic sense of ‘woman’. The distinction may be important, for instance in the historical interpretation of Liðsm 8 (Poole 1991, 113). Similarly, when discussing words for ships in chapter 4, I attempt to distinguish between words that were normally used for particular types of ships and words that were restricted to the vocabulary of the skalds, though it is not always possible to do so. Thirdly, many words in skaldic verse acquire a new denotation when entering into the metaphorical collocations we call kennings. Thus, a ship can be called a ‘horse’ in the common kenning-type ‘horse of the sea’. It would, however, not be useful in this study to list all the instances where a ship is called a horse, because this adds more to our understanding of kennings than to our understanding of ships.24 It is of course interesting to ponder the fact that horses are hardly ever called ‘ships of the land’ in Old Norse poetry, but such speculations are beyond the scope of this study. However, kennings are occasionally referred to where one of the elements in the kenning is a word also discussed in its own right. Although the kenning rarely contributes to an understanding of that word, it is useful to note such kennings to complete the record of its uses. Synonymy can also be a problem. In dróttkvætt, with so many rules of syllable-counting, rhyme and alliteration, the skalds needed many synonyms for and variations on terms both for ships and for men, and the exigencies of skaldic metre may not have allowed them to be too fastidious about their choice of word.25 Again, I try to bring this out in the individual word-studies below, as it is 24

Such references can be traced in Simek 1982. For more on horses, see the forthcoming University of Nottingham thesis by Betsy Springer on ‘The Horse in the Viking Imagination’. 25 This is also a problem when translating, assuming the translator wishes to reflect the synonymic variation in the original. When translating Gísl I,8 in ch. 2, for instance, I translate gramr as ‘prince’ and konungr as ‘king’ which may mislead the reader, since they both refer to the same person.

Introduction: Rocks and Rhymes

35

often possible to determine when a skald is driven by the need for a synonym and when he is using a word more precisely. In theory then, any word in the skaldic corpus may or may not be used in its ‘ordinary’ meaning, and it may or may not be used precisely. While precise observation is not normally thought to be a characteristic typical of skaldic poetry, it has been noted that stanzas to do with ships and sailing are particularly naturalistic, compared for instance to stanzas about war and fighting, which are much more conventional (Fidjestøl 1982, 206–9). It is also the case that the historical imperative behind skaldic praise poetry, the urge to make a record of events and places encourages another kind of precision, making it particularly useful as a source for battles, expeditions and campaigns, and especially the places where these happened (as demonstrated in chs 2 and 4). This precision of time, place and event can also extend to the more descriptive passages associated with these events. The skalds are noted for their artifice, but many of them were also quite capable of accuracy. Despite the problems, the only way to determine the meanings available to any one word in the skaldic corpus, whether a technical term of shipbuilding or sailing, an indicator of social status, or a geographical designation, is to examine the full range of its occurrences and to describe ‘the observed meanings by words of the language of the investigator’ (Steblin-Kamenskij 1966, 24). The contexts of any word in the skaldic corpus are important to this process, as the meaning of a word is not intrinsic but determined by its relationship to other words in the linguistic system. The most immediate context might be a compound word of which the word being investigated forms one element, and these have to be considered in any discussion of that word. More important are the immediate collocations of the word, the other words and phrases with which it makes a statement, whether that is in a phrase, a clause or a sentence. In skaldic verse, with its fairly free word order and often multiple possibilities of combining lexical units, it is sometimes difficult to determine precisely which collocations the terms being investigated enter into, and the ambiguities must occasionally have been intended by the poet. Thus, recasting the stanza in prose misrepresents the way skaldic poetry works, though it is an important first step in understanding it. Sometimes it is necessary to take the helmingr or half-stanza in which the term appears as the collocational unit. Very occasionally, a statement extends right the way through all eight lines of a skaldic stanza. It is also important to consider the wider literary contexts in establishing the meaning of a word. First of these is the poem. A word may take on added significance when seen in the context of the poem as a whole, either because of repetitions throughout the poem or because of contrasts within the poem. Such analysis naturally depends on accurate reconstructions of the poems which are largely constructs of modern philological techniques, as discussed above. However well-founded these reconstructions (and I believe many of them are), there must always remain some doubt about them, and I try to signal below whenever an interpretation depends on such a reconstruction. Other poems can also form a context for an individual word. As a small,

36

Ships and Men in the Late Viking Age

professional class, most poets probably knew each other’s work well, and either borrowed from it or made use of formulaic expressions. Where identical, or even very similar, phrases or lines occur in the work of more than one poet, it is always possible that the borrowing was in spite of, rather than because of, any appropriateness of meaning. Again, the only antidote to such difficulties is to look at overall patterns in the corpus without placing too much significance on any one example. The saga or other prose text in which the verse is preserved is also an important form of context. The text-critical approach, outlined above, in which the aim is a text as close as possible to the original Viking Age composition, would seem to preclude any serious consideration of the prose context in which the verse is found. However, it is a fact that skaldic verse is mediated to us through thirteenth-century Icelandic prose texts and it is better to acknowledge this influence and make allowances for it if necessary, rather than to pretend we are able to read the verse unmediated. Again, where an interpretation of a stanza depends on its prose context, I try to signal this below. Runes and semantics On the surface, there do not seem to be as many problems in establishing the meanings of words in the runic corpus as in the skaldic corpus, and a dictionary of the Swedish Viking Age inscriptions has already been produced (SRR). There are of course many more problems of reading, interpretation and semantics in the less formulaic and more informal runic inscriptions of the medieval period (e.g. rune-sticks), or the usually brief and cryptic inscriptions in the older futhark (which seem to attract proportionally more attention from runologists than the large and relatively uncomplicated Viking Age corpus). But when hundreds, indeed thousands, of rune stones just say ‘X raised this monument in memory of Y’, the possibility of interesting variation is limited. Admittedly, there are several verbs that can be used to describe the action of erecting the monument, a surprising number of prepositions expressing ‘in memory of’, and some significant variation in the noun used to designate the monument itself (an example is the common use of kross ‘cross’ rather than steinn ‘stone’ in the Isle of Man). The inscriptions are also invaluable to onomasts as they provide a vast database of Viking Age personal names and nicknames (see e.g. NR). But most of them shed little more semantic light on Old Scandinavian linguistic usage than that. The formulaic nature of runic memorial inscriptions causes problems for a semantic analysis, as even those inscriptions that have text additional to the memorial formula provide relatively little in the way of immediate linguistic or textual context. For instance, while seventeen inscriptions in Västergötland tell us that the deceased was a góðr drengr, only four (Vg 61, Vg 113, Vg 181, Vg 184) give any further information that might help us to decide exactly what a drengr was, and none gives any real clues as to the semantic range of the adjective góðr ‘good’ (was it just a term of approbation, or did it have some deeper significance?). This relative lack of contexts means that the non-formulaic

Introduction: Rocks and Rhymes

37

vocabulary of runic inscriptions has had to be interpreted in the light of either the subsequent history of Swedish (or the other languages, mutatis mutandis), or of the much more extensive medieval literature in Old Norse-Icelandic. Oddly, though, this has rarely meant a close comparison with the full range of the exactly contemporary skaldic corpus, which is why I have written this book. I also believe it is possible to make much more of the limited contexts in the inscriptions themselves. Thus the runic vocabulary also needs to be recontextualised, as I have suggested above for skaldic vocabulary. For instance, in chapter 6 I try at length to demonstrate how such a recontextualisation can suggest the various shades of meaning which the word drengr could have, in both the runic and the skaldic corpus. Runology is sometimes seen primarily as a ‘philological’ discipline, using this adjective, I suspect, in the ‘new’ sense (the study of language) rather than the ‘old’ sense discussed above (a broad linguistic and cultural interpretation of the past). The effect of this has been to focus runologists’ attention on runic inscriptions as language and, to some extent, as text. A recent attempt to define runology (Peterson 1995, 41, 50) sees ‘språkvetenskap’ (linguistics, language study) as its core, but acknowledges its ‘stort mått av tvärrvetenskap’ (large measure of interdisciplinarity). There is no doubt that recently runologists have become more interested in runes as writing, particularly since they represent such a very different form of writing from the writing on vellum using the roman alphabet which was introduced to Scandinavia by the Christian church. Other aspects of runic monuments are still however left to other disciplines. Thus their design is the province of art historians, who sometimes ignore the texts, while their distribution in the landscape is the province of archaeologists, who may not ignore the texts, but often have a poor understanding of what they do or do not say. Historians, as already noted, are interested in recovering events and actions, and may read more into runic inscriptions than the texts warrant. Clearly, all specialists have an important contribution to make, but I think it is important for the philologically-minded runologist (such as myself) to become at least acquainted with other aspects of the monument than just its linguistic ones (see also Lerche Nielsen 1997, 49). I have stressed above that runic inscriptions are original documents from the Viking Age and as such all of their aspects deserve study, not separately in different disciplines, but together. I have argued elsewhere (Jesch 1998a) that the materiality of rune stones is also a part of their message. Sometimes the form, layout or location of the runic memorial affects the understanding of one or more words in the inscription. They certainly form a part of the context of these words, and I try to draw attention to this whenever I think it is the case in my semantic studies below (an example is the sequence uikikr in Sm 10, discussed in ch. 2).

38

Ships and Men in the Late Viking Age

Comparative angles In this study I have also occasionally found it useful to introduce comparative material from a few sources which also provide contextual clues, although the contexts might be chronological or generic, rather than linguistic and textual. While I have eschewed discussion of Eddic poetry in this study, I have made an exception for the poem known as Helgakviða Hundingsbana I (HHuI). This account of some significant episodes in the life of the legendary warrior-king Helgi has many significant parallels with skaldic praise poems, as often noticed by scholars. Neckel (1908, 365, 421, 432–6) identified reciprocal influence between the HHuI poet and various skalds, locating the poem somewhere between Þjóðólfr Arnórsson, his ‘skaldische hauptvorbild’, and Gísl Illugason, who used HHuI as a source. Jessen (1871, 59), on the other hand, saw Gísl as the author of the Eddic poem, while Bugge (1914) thought it had been composed by Arnórr Þórðarson jarlaskáld. HHuI certainly has a range of nautical and military vocabulary that begs to be compared with the skaldic corpus, and I have noted instances below where they have seemed interesting to me, although I have not attempted to do this systematically, nor to ‘solve’ the problem of HHuI in any way. I have also occasionally cited examples from Old English, especially the prose of the various versions of the Anglo-Saxon Chronicle. Though superficial comparisons are regularly made between the poems of the Chronicle and skaldic poetry, I am less concerned with this than with identifying datable vocabulary that might reflect ship technology, military organisation or warrior ideology in the eleventh century, when England was often ruled by a Scandinavian king and heavily Scandinavianised in some other ways.26 This is reflected in the ‘Scandinavian’ vocabulary of many eleventh-century Chronicle entries, which have not yet, to my knowledge, been the subject of a comprehensive study. I have also very occasionally cited relevant vocabulary items from Old English poetry, mainly though not exclusively from the datable poems such as The Battle of Maldon and The Battle of Brunanburh. On the whole, I believe that comparisons between Old English and Old Norse poetry will be more rewarding when we have better analyses of the vast body of Old Norse texts, and I have therefore resisted the temptation to stray too far down the comparative path. However, some of the lexical items discussed below cannot be discussed without some reference to Old English. I should perhaps note that I also occasionally refer to both skaldic and runic examples that are not in the respective ‘corpus’ where it has seemed illuminating to do so. I do however always try to make clear when citing such extraneous material and to use it only to support an argument, not as the main pillar of one. No doubt there are also many other sources that could have been brought to bear on 26

For a detailed comparison between the Chronicle poems and skaldic poetry, see the forthcoming University of Nottingham thesis by Jayne Carroll.

Introduction: Rocks and Rhymes

39

the topics considered in this book. Norse loan-words into languages such as Irish and French spring to mind, as particularly relevant to nautical terminology. However, I have only one lifetime and feel that my competence is already stretched to the breaking point by what I have chosen to include in this study so, however regretfully, I must leave such fascinating topics to others.

Sources and conventions The study of the runic corpus has been made much easier by the very useful aids that have emanated from Uppsala in recent years, especially the dictionary of Swedish Viking Age inscriptions (SRR) and the database of inscriptions (SamRun), already described, which has made it possible to search for all occurrences of a word. There is now also a provisional version of a dictionary of personal names in the runic corpus (NR), and extensive bibliographical aids (Owe 1995 and the annual bibliographies in Nytt om runer, both also available on the web). The study of the skaldic corpus is much less well supported at present. For a dictionary, we are still dependent on LP which, as is often pointed out, is itself heavily dependent on the interpretations of skaldic stanzas in Skjd. Some parts of the skaldic corpus have received editorial attention since Finnur Jónsson. More recent editions of texts containing large numbers of skaldic stanzas have in many cases produced radically new interpretations of those stanzas though, as I have already noted, these are more likely to be concerned with reconstructing the text known to the saga author than that which issued from the mouth of the poet. Thus, Bjarni Aðalbjarnarson is an important, and largely unsung, editor of skaldic verse in his edition of Hkr, while Anthony Faulkes’ edition of SnESkskm has clarified many aspects of the stanzas recorded in that work. On the other hand, Diana Whaley’s fine edition of the œuvre of Arnórr jarlaskáld (1998) emphasises what the poet composed rather than what saga authors thought he did. Moves are afoot to re-edit the whole of the skaldic corpus on an international and collaborative basis. This will undoubtedly revolutionise skaldic studies and perhaps even provide the basis for a new Lexicon Poeticum, on the model of ONP. But as such an edition (let alone such a dictionary) is still several years off, it is necessary to rely on Skjd as the basic edition of the skaldic corpus. This naturally has to be supplemented with consultation of the manuscripts, as Skjd A is not always accurate, nor does it give ‘access to the full paradosis’ (Poole 1993, 80), and also of any more recent editions of the poems, including critical editions of the texts containing the poems. Kock’s edition of the skaldic corpus (1946–9), though frequently cited as if it were a better edition than Skjd, is mainly of use in identifying the paragraphs in NN in which Kock discusses individual stanzas. His suggestions are often valuable, and frequently referred to below, but his posthumous edition has no value as a working text. All references to ‘the skaldic corpus’ and ‘the runic corpus’ in this book are to

40

Ships and Men in the Late Viking Age

the limited selection as outlined in this introduction. It is important to note that whenever I refer to the number of occurrences of a word, it is always in ‘the corpus’ as defined in this introduction. The ‘runic corpus’ consists of the approximately 3,000 Viking Age memorial stones as outlined above, while the inscriptions actually referred to in this study are listed in Appendix I, which also provides a key to the abbreviations and editions. Runic texts can be found in the editions listed in Appendix I, which are also the basis of the texts given in SamRun. I cite the texts from SamRun, giving them, as is conventional for Scandinavian inscriptions, in bold type. With some hesitation, I have, for the sake of consistency, chosen to follow SamRun’s practice of transliterating the fourth rune as o, even in Danish inscriptions, where it is otherwise normally rendered as =. I also follow SamRun in its simplified and word-processor-friendly system of representing the myriad forms of dividers that are found in inscriptions and the various symbols used by editors to indicate lost, damaged or otherwise uncertain runes, though not bind-runes, or those runes that need to be read twice.27 When citing only a few words of a runic inscription, I do not always include the dividers especially if they are not relevant to the argument. When citing a whole inscription, however, I do include dividers, though not any indication of line-shifts. The index in Appendix I shows which inscriptions are actually cited in this study. As indicated above, the number of inscriptions with significant, non-formulaic, vocabulary is only a proportion of the ‘runic corpus’. Quotations from runic inscriptions are unnormalised transliterations, while those from skaldic texts will normally follow the edition being cited, or will be normalised using the conventions of the Íslenzk fornrit series. The reason for not normalising runic inscriptions is that these are singular, original documents from the Viking Age and reproducing them in transliterated form recognises this fact. Skaldic texts, on the other hand, have already undergone one transformation, from the written to the oral form, and most of them survive in more than one manuscript, none of which is an archetype. There is thus no point in elevating one manuscript text to a spurious authority by reproducing that rather than a critical text (as in done, for instance, in Lange 1989). All translations of both runic and skaldic texts are my own, unless noted otherwise. In my translations of skaldic verse, I have adopted Whaley’s useful system (1998) of arrows to explain kennings in an economical fashion. All other translations are also my own, unless noted otherwise. Prose texts in ON are silently normalised when cited

27

The following should be noted in particular: round brackets () indicate a rune which is damaged or incomplete but which can nevertheless be read; square brackets [] indicate a rune or sequence of runes which can no longer be read but which are adduced from older transcriptions or drawings of the inscription; a hyphen – indicates a rune which can no longer be read, but which can still be counted; an ellipsis . . . indicates one or more lost runes which can no longer be read or counted or reconstructed from older sources. Further details on the conventions used in the database are available in SamRun.

Introduction: Rocks and Rhymes

41

from diplomatic editions, in other languages I cite texts in the orthography of the edition being used. References to ‘the skaldic corpus’ are to the selection of texts listed in Appendix II, which also provides a key to abbreviations. Again, the index indicates which stanzas are actually cited in the chapters below, which is a higher proportion of the total corpus than in the case of the runic inscriptions. Individual stanzas are referred to by poet, poem and stanza number as in Skjd. Appendix II takes some account of more recent work on the reconstruction of individual poems (especially Fidjestøl 1982, Poole 1991, and Whaley 1998) and I attempt to make clear when I accept these suggestions and when not. However, with the exception of Liðsm, I still refer to all stanzas by the numbering given them in Skjd (as Fidjestøl does), for instance I do not adopt the numbering for Arnórr’s stanzas suggested in Whaley 1998. In the text, skaldic stanzas are generally quoted in a normalised form that is most often that of the Íslenzk fornrit editions (where there is such an edition of the text containing the stanza in question). I generally choose one particular reading of a stanza (often, though not always, from an Íslenzk fornrit edition, or from SnESkskm) and cite all or part of that stanza according to that reading. I do not normally cite variants and alternative readings, except where they affect the interpretation of the vocabulary item under discussion, in which case I do try to outline the alternatives while making clear which one I prefer. I normally give a source reference in the text when I cite a whole or half stanza, but otherwise, and in any case, all stanzas can be traced in Skjd. When discussing words as words, whether in skaldic or runic texts, I use normalised OWN forms, to facilitate comparison between the two corpora, and because there is no other standard of normalisation that would accommodate Danish, Swedish, Norwegian and other runic inscriptions. This normalisation will however take into account differing declensions (e.g. pl. drengjar in runic inscriptions, drengir in skaldic verse). Names will also be given in their normalised OWN form. For better or for worse, I have followed Skjd’s spelling of the names of poets and LP’s abbreviations in Appendix II. Occasional minor differences in the spelling of poets’ names and the titles of poems in my text originate from the editions being used and should not cause any major confusion. In chapters 4–6, I attempt to document every occurrence of relevant words (though I may not have succeeded in this). In chapters 4–5 this is because the vocabulary of the skaldic and runic sources has been scantily or incorrectly utilised in the study of late Viking Age ships and men, and a survey is badly needed. Most of the words discussed in chapter 6, on the other hand, have been much discussed, but rarely with a full consideration of all the instances. In chapters 4–5, I attempt to document all words relating to ships and their equipment, their crews, and technical terms of sailing and maritime warfare, though there is a grey area between ‘sailing’ verbs and more general verbs of motion, and between warfare at sea and warfare in general. Also, there is some metaphorical vocabulary in descriptions of sailing and fighting, but less in the descriptions of ships and their parts. In chapters 2–3, I have only sampled the skaldic vocabulary, but

42

Ships and Men in the Late Viking Age

do attempt to be comprehensive in discussing the runic vocabulary, which is possible because the runic corpus, though large, is relatively limited in its vocabulary other than that found in the core memorial formula.

Ships and men in the late Viking Age The core of this book is in chapters 4 and 5, where I survey the vocabulary of ships and their crews, of fleets and sailing and battles at sea. The urge to look more closely at these lexical sets came from the realisation of the enormous potential of the skaldic corpus for the study of many different aspects of the Viking Age, along with the observation that this source was not fully exploited by Falk in his study of the language of ships and seafaring (AnS), a work still much cited by archaeologists and others (and even translated into Swedish recently), though it is nearly ninety years old. Falk’s relative neglect of skaldic poetry was in part a result of his general disregard for chronology and source criticism. To me, the skaldic corpus seemed to offer a real possibility of focusing on the Viking Age without projecting back onto this period the nautical developments that took place in later periods and which are well-instanced in the sagas that form the main basis of Falk’s study. A preliminary study of the potential of skaldic poetry for nautical history was published by Foote in 1978. This covered only a few points to do with sailing, but was very influential on my own thinking about appropriate methods and material for such a study. Since Foote, only Malmros (1985) has followed in this vein of using the whole skaldic corpus as a source for nautical history. Although I disagree with many of her conclusions, and there is a fundamental difference between her historical and my philological approach, her pioneering contribution deserves to be acknowledged. Falk certainly used skaldic sources, though sparingly, but he neglected the evidence of runic inscriptions. Thus, it seemed to me useful to supplement the work of Falk, Foote and Malmros by considering the runic evidence for nautical vocabulary. While there is not a great deal of it, it can sometimes be very significant (see e.g. the discussion of kn›rr in ch. 4, or the discussion of crew- and fleet-terms in ch. 5). In many ways, then, this book was conceived as a project to rewrite AnS for the twenty-first century, to complement the greatly increased (and continually increasing) evidence that archaeology has provided for nautical history since Falk wrote, and in particular the intensive ship-archaeological research of the last forty years or so. However, I soon realised that it was not possible to look at ships without also looking at their crews, or without looking at the reasons for which those men set out in those ships. I am not a nautical historian, and must also confess to being a landlubber of the highest order, and did not wish to write a work of purely nautical interest. It seemed to me the most useful contribution I could make would be to set the nautical vocabulary into a larger context, the more general one of ‘viking’ activity in all its aspects. So in chapters 2–3 I survey the evidence the runic and the skaldic corpus provide for what this activity

Introduction: Rocks and Rhymes

43

consisted of and where it took place, while in chapter 6 I consider some of the social and military aspects of ‘viking’ activity. I have however avoided going too far in the direction of writing a guide to ‘Viking warfare’ for two reasons. One is that it would have involved bringing in a whole new range of archaeological evidence to do with weapons, fortifications and the like, which would stretch the bounds of an already overlong book. The second, and more important one, is that I am not convinced that either the runic or the skaldic corpus provides much useful evidence for a general study of warfare (though I do consider sea-battles in ch. 5). I have already noted above that, while the skalds are refreshingly and usefully naturalistic in their descriptions of ships and sailing, they are much less so in their descriptions of fighting. There is certainly scope for a study of war in the skaldic corpus, but it is debatable how much this would tell us about the actual practice of war, as opposed to attitudes to war (see for instance my study of the ‘beasts of battle’, Jesch 2001d). War was just one aspect, though clearly an important aspect, of the late Viking Age. Much of this warfare involved the use of ships, sometimes as troop-carriers, sometimes as warships equipped for fighting at sea. Ships could also be used for trade or exploration, or sometimes all three. In the next two chapters, I survey what the skaldic and the runic corpus have to say about such ‘viking’ activities and where they took place.

2 Viking Activities hydro-heroes . . . Valhalla bound CALVERT

The modern term ‘Viking Age’ implies either an era associated with people called ‘vikings’, or one in which people engaged in an activity called ‘viking’, just as they practised chivalry in the Age of Chivalry, or enjoyed jazz in the Jazz Age. But it is not as easy to define ‘viking’ as either jazz or chivalry. This chapter is an attempt to discover what was involved in ‘viking’ activity, or rather activities, as indicated in the significant vocabulary of both the runic and the skaldic corpus.

Vikings The English word ‘Viking’ or ‘viking’ should not be confused with the ON word(s) from which it was ultimately borrowed, not least because the modern word is much commoner than its Viking Age predecessors and has developed a life of its own (Fell 1987). ON has two words, the masculine víkingr referring to a person, and the abstract feminine noun víking referring to an activity. These words have been much studied in their various manifestations. Most such studies (e.g. Askeberg 1942, 114–83; Hellberg 1980; Hødnebø 1987), range widely across languages and centuries, and are often concerned with etymology and the semantic development of the term. Fell (1986; 1987) considers, respectively, Old and Modern English cognates. Here, I am concerned only with establishing in detail the uses and nuances of the words in a limited range of sources and during a limited period of the Viking Age, as outlined in chapter 1. víkingr The noun víkingr (m., pl. víkingar), referring to a person, occurs in both runic and skaldic sources. An immediately interesting aspect of the runic evidence is that it provides examples of Víkingr used as a personal name.1 This onomastic 1

Some scholars have doubted whether this name is etymologically identical with the common noun, see von Feilitzen 1937, 405, and references.

Viking Activities

45

usage is attested in up to nineteen inscriptions. Although these examples represent a small part of the runic onomastic pool, their chronological and geographical range is wide.2 There are fifteen rune stone inscriptions in which a commissioner, the commemorated, a relative of the commemorated, or a runecarver, is called Víkingr, representing fourteen individuals in all (see fig. 2.1).3 In three further inscriptions, the name in question is fragmentary or uncertain, but is also likely to have been Víkingr (Sö 13, Sö 269, U 813).4 Finally, there is one inscription where a sequence which can be interpreted as víkingr appears in an unclear context (Sm 10, see fig. 2.2). On this stone from Växjö the memorial formula is complete and selfcontained: -u(k)i reisti * stein * e(f)tir : kunar : sun : kirims ‘Tóki raised this stone in memory of Gunnarr, son of Grímr’. The additional words uikikr and tyki appear in the centre of the design, and it is not certain in what order they should be read or how they relate to the main text. Both words run upwards from the lower part of the stone. As uikikr is to the left of tyki, it might be supposed it ought to be read first, but the use of a divider after tyki suggests that the correct sequence is actually tyki x uikikr. It could simply be a repetition of the name of the commissioner (-uki), with an added appellative, specifying him as Tóki the Viking (as it is taken in SR IV, 57). But the central text could also be interpreted as a rune-carver’s signature. Although it does not include a verb, that is not unparalleled, see Sö 266 and Sö 312. In this case it could even consist of two names, with Víkingr the name of a man who helped Tóki produce the monument. Gunnarr’s relationship to Tóki is not specified, and this makes it is less likely that he was a family member, as in the other examples discussed here, than some kind of a comrade, whether in war or trade. Thus it is just possible that the appellative víkingr is used here in the same way as some of the in-group vocabulary (e.g. drengr, félagi) discussed in chapter 6 below. However, as such terms are overwhelmingly used of the commemorated in runic inscriptions, the Växjo stone would be anomalous in using an honorific term to refer to the commissioner. Moreover, as will be seen below, there is no evidence, other than personal names, of the use of the singular form at this period. Thus, the balance of probability is that the central text tyki x uikikr represents two personal names, probably of the men who carved the monument. There are only three runic inscriptions which use víkingr as a common noun rather than as a personal name. These inscriptions are chronologically and geographically diverse, and thus attest to the use of víkingr, or rather the plural

2

3 4

There are no examples of the name Víkingr in the skaldic corpus, where personal names are relatively rare. The name is attested in Norway, Sweden and Denmark after the Viking Age, and was exported to viking colonies in England (von Feilitzen 1937, 405). Fellows Jensen (1968, 338–9) lists examples of English place-names in Lincolnshire and Yorkshire containing the personal name Víkingr. Ög 8, Sö 54, Sö 182, Sö 197, Sö 203, Sm 11, Vg 17, U 34, U 175, U 260, U 432, U 498, U 649, U 681, U 802. The same individual is referred to in both Sö 197 and Sö 203. On the oblique form uikika in Sö 269, see Otterbjörk 1983, 30.

46

Ships and Men in the Late Viking Age

Image not available

2.1 The Fresta stone (U 260), with uikinkr, the name of one of the commissioners, on the left towards the bottom. Photo: Judith Jesch.

Viking Activities

Image not available

2.2 The Växjö stone (Sm 10), showing the word or name uikikr centre left. Photo: Judith Jesch.

47

48

Ships and Men in the Late Viking Age

víkingar, in different parts of Viking Age Scandinavia, but despite this range they do not present a very clear picture of what the word might mean. The earliest example is on a stone from Tirsted on Lolland (D 216), probably from the middle of the tenth century (Moltke 1985, 301), Denmark’s second largest rune stone, and one of its most obscure. The inscription appears to have been commissioned by two men to commemorate a relative who died in Sweden and ends with the runes aliRuikikaR. This is easily enough interpreted as ‘all vikings’, but how it relates to the rest of the text is much less certain. In as far as we can decipher the context of this inscription, it appears to be military, but the relevant vocabulary is all conjectural and based on suggested corrections of the carver’s supposed errors. Thus Moltke (1985, 300, Englished by Foote) suggests the following translation of the relevant part: ‘And he was then the terror (?) of men and he met death in Sweden and was the foremost of Fregge’s host; and then: all vikings.’ Even if we accept these corrections and conjectures, it is not at all clear how the phrase ‘all vikings’ fits in. The two Swedish inscriptions which use víkingar are more easily interpreted, although the exact reference of the word is no easier to pin down. A rune stone discovered in 1988 in Hablingbo, on Gotland, which, like others on Gotland, adopts the characteristic shape of the picture stone (and can thus be dated to the eleventh century), was commissioned by two brothers to commemorate their father, of whom it is said that hn : uahR -istr : farin miþ uikikum ‘he had travelled to the west with vikings’ (G 370). Normally, travelling to the ‘west’ in Swedish runic inscriptions refers to viking raids on the British Isles, in particular the campaigns that led to the accession of Knútr in the early eleventh century. However, because of the location of Gotland in the Baltic, it is equally possible that the deceased had travelled no further west than mainland Scandinavia, or Denmark (see ch. 3, below). In either case, the journey could have been undertaken for either war or trade, and there is no way of knowing which is meant in this instance. A large and elegant rune stone from Bro in Uppland gives more context for interpreting the word, although not necessarily definitively (U 617). The inscription is long and relatively complex, and is related to others in the region which commemorate members of the same family (Jesch 1991, 58–9). The memorial was commissioned by a woman to commemorate her husband, of whom it is said that saR x uaR x uikika x uaurþr x miþ x kaeti, literally ‘he was a guard of/against vikings with Geitir’. The statement is ambiguous, but is usually interpreted to mean that ¡zurr was an officer, under Geitir, engaged in keeping watch against viking raiders, giving us ‘a glimpse of Swedish coastal defence organisation’ (Jansson 1987, 91). However, v›rðr more properly means ‘guardian’ or ‘protector’ and it would be most natural to follow those scholars (Askeberg 1942, 122; Hellberg 1980, 55) who have taken víkinga as an objective genitive, indicating that which is guarded or protected. The only runic parallel to this phrase is on a fragmentary stone found in Giberga, Södermanland in 1942, containing the sequence kiarþu skibuarþ which is interpreted as gerðu skipv›rð ‘they did ship-watch, they protected the ship’ (Sö FV1948:291), using v›rðr in the sense

Viking Activities

49

of the action rather than the actor. In skaldic usage, v›rðr is most commonly used of a person and normally collocates with one of three things: people, land, or objects such as ships (Eskál III,1; Sigv I,8; Sigv III,13; Sigv VII,7; Sigv XII,6; Mark I,3,4). In all of these examples, the v›rðr is normally the guardian or protector of, or possibly user of, these things, and the term comes to be used frequently for the king as guardian of his land and people. I have not found any examples where the genitive object refers to that which someone or something is being guarded against and thus it seems linguistically most likely that ¡zurr held some position of responsibility in a troop of vikings. These last three inscriptions are thus the shreds of contemporary evidence that people whom we might call ‘vikings’ also called themselves something similar.5 Not one is unambiguous, but the fact that these inscriptions can be read in this way suggests that there was at least a possibility of the term víkingr being used of one’s own group, apparently with approval. Presumably people in the Viking Age saw a connection between the common noun víkingr and the personal name Víkingr (whatever etymologists might say about their different origins), and so the use of the personal name provides additional evidence for the positive connotations of the noun. However, the surviving runic examples of the noun are all in the plural form, and it may be that this implies some kind of distinction, but what we do not know. Turning to skaldic verse, we find again that many of the examples are ambiguous. The unambiguous ones however suggest that, in contrast to the runic material, víkingr in the skaldic material was a pejorative term applied to one’s opponents and would not normally be used of oneself or one’s own group. The skaldic examples are like the runic examples in all being in the plural. The pejorative usage is attested right through from the late tenth to the turn of the twelfth century, with seven clear instances of the word used to refer to enemies or opponents of the king or leader being praised, both at home and abroad. The víkingar in these stanzas are: the enemies of Eiríkr sigrsæli at Fyrisvellir (ÞHjalt 2); the Jómsvíkingar at the battle of Hj›rungavágr (Þskúm; Tindr I,5); the (presumably Wendish) opponents of Eiríkr jarl during his raids on

5

Fell (1986, 313) has, I think, erroneously claimed that a comment by Adam of Bremen demonstrates that ‘the vikings called themselves wichingos’: pyratae, quos illi Wichingos appellant, nostri Ascomannos (AB, 440). The context is a description of Denmark, the wichingos are pirates who pay a tax to the Danish king in order to plunder ‘barbarians’, but who occasionally also misuse this freedom to plunder their own people. It is the Danes in general who call these pirates wichingos, not necessarily a term the latter use of themselves, thus Tschan (1959, 190) translates ‘These pirates, called Vikings by the people of Zealand, by our people, Ascomanni . . .’ The passage then goes on to describe various forms of crime (including betrayal of the king) and punishment current in Denmark. This usage seems to me to fit most closely the skaldic examples discussed below, and also from the eleventh century, where the king is shown punishing the perpetrators of internal disorder. We have to remember that the king of Denmark was one of Adam’s primary sources, and it is noteworthy that this is the only occurrence of wichingi in this substantial work, otherwise he tends to use Normanni to refer to Scandinavians of the viking persuasion.

50

Ships and Men in the Late Viking Age

Wendland (Edáð 5);6 the inhabitants of a stronghold at Hól conquered by Óláfr Haraldsson (Sigv I,10); the ‘enemy within’ in a retrospective view of Óláfr’s reign, showing how he punished miscreants and wrongdoers (Sigv XII,6); and a similar general reference to the enemies of Magnús berfœttr, probably abroad, possibly also at home in Norway (Bkrepp 3).7 There are four instances of víkingar where the interpretation of the stanza is difficult and it is hard to be certain whether the word refers to ‘them’ or ‘us’, from the point of view of the speaking poet and his patron. Here, many skaldic interpreters are influenced by their expectations that víkingar has some specific reference, possibly with implications of nationality, and also that it is generally positive and would be used by the speaker of his own group. Each of these ambiguous examples presents a different problem of interpretation. Despite Sigvatr’s unambiguous use of víkingar for óláfr’s opponents in Brittany (Sigv I,10), commentators have assumed that the word can refer to Óláfr and his men elsewhere in this poem. In an earlier stanza, víkingr appears in a very uncertain collocation, as it can be taken with a number of different words in its immediate context (Sigv I,3). The main question here is whether the víkingar are just an element of a kenning referring to the sea or ships, or whether the term is used specifically of Óláfr’s troop (and their ships).8 In the context, it does however seem most likely that the víkinga skeiðar of the stanza are the ships of Óláfr and his followers. Understanding the víkingar as Óláfr and his men in this stanza then increases the temptation to read the sixth stanza in the same way (Sigv I,6). This describes Óláfr attacking London (Hkr II, 17–18): Rétt es, at sókn en sétta, snarr þengill bauð Englum at, þars Óleifr sótti, Yggs, Lundúna bryggjur. Sverð bitu v›lsk, en v›rðu víkingar þar díki. Áttu sumt í sléttu Súðvirki lið búðir.

6

7

8

The location of these raids in Wendland depends on the prose context of the stanza, rather than on any evidence in the stanza itself. See Fidjestøl 1982, 111–14, for the reconstruction of this poem. If we accept Fidjestøl’s reconstruction of this poem (1982, 150–52), the stanza comes at the transitional point between ‘home’ and ‘abroad’, it might well have been the refrain. The alliteration of with vítt perhaps suggests enemies abroad, in a kind of summary of Magnús’ geographically wide-ranging campaigns (Jesch 1996, 120–21). With some differences in detail, the former interpretation is favoured by Jón Helgason (1935–6, 263–4) and Kock (NN, 612), while Finnur Jónsson (Skjd B I, 213), Bjarni Aðalbjarnarson (Hkr II, 11), Fell (1981, 112) and Hellberg (1980, 45–6) all favour the latter. No one manuscript has a completely satisfactory text of this stanza, and all editors print a reconstructed text using variants from different manuscript traditions.

Viking Activities

51

It is correct that the sixth battle (was) where Óláfr attacked the bridges/wharves9 of London. The effective leader offered the English Yggr’s strife [battle]. Foreign swords bit, and vikings defended the ditch there. Some of the troop had booths in level Southwark.

The first half of the stanza refers to Óláfr’s attack, and it seems logical for the second half to describe the defence, as suggested by the clear antithesis between the verbs sótti and v›rðu. However, the second half mentions both a lið ‘troop’ and some víkingar, and it is not clear whether these are the same group or not.10 For Hellberg (1980, 37), the conjunction en in l.5 is significant. Its meaning is adversative (‘but’), but it often means not much more than ‘and’.11 Hellberg prefers to take it as the latter here, so that, rather than introducing a new idea, the opponents, it introduces more information about the army in which Óláfr is fighting, who are defending themselves in Southwark. But Hellberg ignores the illogicality of the same stanza describing first Óláfr’s attack and then his defence, without any indication of what his opponents are up to in between. Also, his interpretation depends on the assumption that the word víkingar must refer to Scandinavians of some sort. I would instead suggest that the víkingar are the inhabitants/defenders of London (the ‘English’ of the first half of the stanza) who resisted Óláfr’s attack, with only the last two lines referring to the retreat to their base by Óláfr’s troop (the lið). Thus, two out of the three examples of víkingar in Sigvatr’s Víkingarvísur most likely refer to Óláfr’s opponents. Another poem describing English battles is the anonymous Liðsmannaflokkr. 9

10

11

According to Fell (1981, 115), ‘the plural form here is used for the singular’ but she does not elucidate further. She also notes that ‘the meaning is evidently adopted from the OE cognate brycg’. The only other skaldic instance of bryggja with the meaning of ‘bridge’ (and also plural) is Ótt II,7, referring to the same event and probably derivative of Sigvatr. London seems to have had only one bridge crossing the Thames at this time (see the maps in Clark 1989 and Vince 1990), and the normal Old Norse meaning of bryggja ‘pier, wharf, landing-place, jetty’ should not be dismissed (Clark 1989, 24). As a great trading centre, London would have had many such and some from this period are known archaeologically. Thus while London Bridge can be traced to the early tenth century, there were wharves at Queenhithe (upriver of London Bridge) from the late ninth century onwards. The eleventh century seems to have been just when there was a spurt of new jetty-building (Vince 1990, 33–4, 153), although the ‘busy waterfront’ may have been mainly a feature of the middle of the century (Vince 1990, 37, 106). Thus, the first jetty and revetment at St Magnus House (immediately downriver of London Bridge) are c.1020, while the first revetment at Billingsgate Lorry Park is c.1039/40, and the major revetment at Billingsgate dates to c.1055 (I am grateful to Alan Vince for explaining these matters to me). Townend (1998, 52) translates literally (‘the bridges of Lundúnir’), but asserts (1998, 73) that the stanza records ‘Óláfr’s attack on London Bridge’. Bjarni Aðalbjarnarson (Hkr II, 18) has the lið as the defenders (the víkingar who v›rðu þar díki), while Fell (1981, 115) seems to suggest that they are the attackers. Since búðir are temporary structures, it seems most likely that they would be occupied by transient warriors. However, this still leaves the status of the víkingar uncertain. The main manuscript of Heimskringla, the mostly lost Kringla, probably had enn (ms. ‘e’ or ‘eû’, judging from the copies in AM 36 fol. and AM 70 fol.), which could I suppose equally mean ‘still’.

52

Ships and Men in the Late Viking Age

Again, we have a description of a battle in which it is not entirely clear which side is which (Liðsm 4; Poole 1991, 87): Þóttut mér, es ek þátta, Þorkels liðar dvelja – sóusk eigi þeir sverða s›ng – í folk at ganga, áðr an ?hauðr? á heiði hríð víkingar kníðu – vér hlutum vápna skúrir – varð fylkt liði – harða. Þorkell’s men did not seem to me, as I saw (them), to lose time in joining battle – they did not fear the ringing of swords – before the Vikings fought a hard engagement on ?hauðr? heath; we encountered showers of weapons; the warband was in battle formation.

This stanza comes in that section of the poem which, Poole (1991, 99) has argued, is in praise of Þorkell the Tall. Thus, the víkingar, the lið, and the first-person plural verb forms would all refer to Þorkels liðar ‘Þorkell’s men’, of whom the speaker is one. However, the stanza is at the very least ambiguous, the víkingar could also be the opponents who offer Þorkell and his men battle: Poole takes the conjunction áðr at the beginning of line 5 as linking the two half-stanzas, but it could equally anticipate line 8 (giving ‘the warband was in battle formation before the Vikings offered a hard engagement on ?hauðr? heath’), and thus emphasise the statement of the first half-stanza that Þorkell’s men did not delay in joining battle.12 It would also seem logical for there to be some mention of the enemy side before Þorkell and his men ‘encountered showers of weapons’. Even Poole (1991, 108) summarises this stanza as ‘the advance of Þorkell’s army, the marshalling into battle formation, and the actual engagement with the enemy’. Thus it is quite easy to see the víkingar as the enemy here. Yet another stanza using víkingar in the context of battles in England is from Óttarr svarti’s Knútsdrápa. Again, most commentators have been misled by their assumption that the term víkingar must have ethnic significance (Ótt III,5; Knýtl, 106): Gunni lézt í grœnni, gramr, Lindisey framða. Belldu viðr, þvís vildu, víkingar þar [ms. því] ríki. Bíða léztu í breiðri borg Heminga sorgir

12

There is a similarly ambiguous example (Hfr III,17) of áðr at the beginning of the second half-stanza, which, in Skjd’s interpretation, anticipates a past tense verb later in the same half, though I suppose it could equally link back to the first half stanza.

Viking Activities

53

œstr [ms œst] fyr Úsu vestan engst folk, Svía þrengvir.13 Prince, you caused battle to be promoted in green Lindsey. Vikings deployed power there as they wished [or, Those vikings who wished resisted that power]. Oppressor of the Swedes, battle-enraged you caused the English army/people to experience sorrows in broad Hemingbrough to the west of the Ouse.

Both Finnur Jónsson (Skjd B I, 273) and Bjarni Guðnason (Knýtl, 106) understood lines 3–4 as referring to Knútr’s followers, thus the latter’s prose word order is Víkingar belldu viðr þar ríki, þvís vildu ‘vikings deployed power there as they wished’, with a conjectural emendation of því to þar (as given above). Both these scholars are forced into a rather strained interpretation of the sentence because of their assumptions about the word víkingar. In contrast, Kock (NN, 2218) attempted to understand this stanza keeping the manuscript reading in line 4, which might give a translation of lines 3–4 of ‘those vikings who wished resisted that (i.e. your) power’. In this reading, the víkingar are the enemy, the engst folk of line 8, and the því in line 4 then refers deictically back to the first two lines of the stanza, making for a (fairly) neat parallelism in the stanza: in each half there is a place-name (with adjectival or adverbial expansion), a reference to the king and what he did, and a reference to the enemy and what they did. Kock’s interpretation not only avoids emendation but also fits the meanings of bella (viðr) better, as it often means ‘to oppose, resist’, referring to aggression against, rather than on the part of, the person(s) under discussion.14 Hellberg (1980, 40) follows Finnur Jónsson in understanding bella ríki to mean ‘utöva makt, ha övermakten’, but again has to overinterpret to make the stanza mean what he wants it to. He concludes that the víkingar are Knútr’s followers, although he would not completely rule out Kock’s interpretation, but dismisses the possibility that the víkingar are English peasants. Because Hellberg’s theory is that víkingar are people from eastern Norway, he admits the possibility that in this stanza it refers to such men among Knútr’s opponents in Þorkell’s army, but not because he sees this idea of ‘opposition’ as central to the meaning of the term, as I do. Although Hellberg sees a parallel with Ótt III,2 which mentions Knútr’s Scanian and Jutish followers,15 it is made clear in that stanza that they are his followers, thus the poet says that they út fylgðu þér ‘followed you abroad’ (cf. also Ótt III,3), while each stanza from 4–8 names one or more groups of his opponents. Thus, there is a strong case for accepting Kock’s interpretation of lines 3–4 and seeing the víkingar in stanza 5 as the opponents of Knútr. 13

14 15

Cited from Knýtl, 106. It is clear from the footnote that the editor (Bjarni Guðnason) intended to print œstr in line 7, although he in fact prints œst. My translation represents his interpretation of the stanza with the alternative translation of lines 3–4 in square brackets reflecting Kock’s reading (NN, 2218). Cf. Hildr; Hfr III,21; Sigv XIII,17. The supposed counter-example in Sigv III,6 occurs in a negative context. However, he mistakenly calls it st. 4 (1980, 41).

54

Ships and Men in the Late Viking Age

There remain only two instances where there is a strong possibility that the poet is using víkingar to refer to his own group, or at least to the followers of the king he is praising, who are Haraldr harðráði and his son Óláfr kyrri. In both cases, the content of the stanza is relatively clear, but the interpretation of who the víkingar are depends on the context of that stanza in the poem, which is much less certain. A stanza by Valgarðr from a poem on Haraldr, preserved only in Snorra Edda, so that it is not certain what event it celebrates, describes an attack on a borgar virki ‘city’s fortifications’ (Valg 3). It contains the parenthetical statement víkingar brutu fíkjum ‘the vikings destroyed rapaciously’ (discussed further below). On balance, it is most likely that here the víkingar are the members of Haraldr’s army, attacking a foreign town or castle. Similarly, a stanza from Steinn Herdísarson’s Óláfsdrápa seems to record one of the (successful) battles before Stamford Bridge, where Óláfr fought alongside his father (Steinn III,3). If so, it is most likely that the vikings wading in the blood of men in that stanza (víkingar óðu bragna blóð) are the victorious Norwegian army. The skaldic evidence for the meanings of víkingr is by no means unassailable, nor does it all point the same way (as even Hellberg admits, 1980, 57). In the material presented above, there are seven clear instances where the meaning is ‘them’, two fairly clear of ‘us’, and four debatable ones. Of the debatable ones, I find only Sigv I,3 convincing as representing the ‘us’ meaning, leaving a total of ten ‘them’ and three ‘us’. On balance, then, I would argue that the plural term víkingar is used predominantly of opponents or enemies. It is certainly not the normal word used by ‘viking’ in-groups of themselves. It may however be possible to detect, in the last two examples discussed, a shift towards a romanticisation of the word in the later eleventh century which enables it to be used positively of the speaker’s own group, not unlike that which we can detect in the way it is used in later ON prose and in pastiche Viking Age poetry such as Egill VII,1. víking The evidence for the abstract noun víking (f.) is more limited than that for víkingr, but provides a useful starting-point for the consideration of what ‘viking’ activities were. The word víking in fact occurs only three times in the runic corpus and once in the skaldic corpus. The three rune stone inscriptions, two Danish (both actually from Skåne, D 330 and D 334) and one Swedish (Vg 61), commemorate men who had been (and in two cases who had died) í víkingu ‘in viking’. The meaning of this word is not at all clear, though a certain amount can be deduced from its contexts, including both other significant vocabulary in these inscriptions and the nature of the monuments themselves.16 One of the Danish inscriptions (D 16

Hellberg’s attempt (1980, 81–4) to explain these instances away as representing a phrase i vighningu referring to heathen consecration is both orthographically unlikely and unconvincing because he ignores the collocations with other ‘viking’ vocabulary (see also Salberger 1981).

Viking Activities

Image not available

2.3 The Härlingstorp stone (Vg 61), showing the words i uikiku in the bottom right of the inner band. Photo: Harald Faith-Ell, Antikvarisk-topografiska arkivet, Riksantikvarieämbetet, Stockholm.

55

56

Ships and Men in the Late Viking Age

330) needs to be partially reconstructed, so the immediate collocation of the phrase in question is conjectural: the commemorated men possibly váru víða óneisir í víkingu ‘were far and wide unafraid in ‘‘viking’’ activities’. This clause probably refers to the commissioners of the monument as well as to the commemorated, though this is not certain. The adverb suggests that ‘viking’ took place in a variety of places far from home, while the inscription also uses in-group vocabulary such as félagi and drengr (discussed in more detail in ch. 6), which can refer to both raiding and trading activities. The other Danish inscription (D 334) and the Swedish one (Vg 61) both commemorate men who died í víkingu, the former in the north, the latter in the west (see fig. 2.3). Again, the contexts could be expeditions of either raiding or trading (or both). Thus, the commissioner of D 334 (here commemorating his brother) also commissioned another stone (D 335) to commemorate a man who had owned a ship together with him, and the two rune stones along with five non-inscribed stones and a cenotaph make an impressive monument still in its original location at Västra Strö. On Vg 61 a mother remembers her son who died o uastruakum ‘on the ways west’ and this must be seen in the context of a number of inscriptions from Västergötland which commemorate men who died in the early eleventh-century Scandinavian wars in England (Vg 20, Vg 187, see also Vg 197). Whatever the activity of ‘viking’ was, it was clearly praiseworthy, at least at those times and in those places where these three monuments were made. But as it was not a term used with any frequency, we cannot really say what activities it included. The earliest occurrence of víking in skaldic poetry comes in a poem by the latest poet in the corpus, Markús Skeggjason’s Eiríksdrápa, dated to shortly after 1100 (Mark I,8a; Knýtl, 216): V›rgum eyddi Vinða fergir. Víking hepti konungr fíkjum. Þjófa hendr lét þengill stýfa. Þegnum kunni ósið hegna. The conqueror of the Wends destroyed outlaws. The king decisively stopped víking. The ruler had thieves’ hands chopped off. He curbed the rebellion of the landowners (þegnar).

Here, the activity of ‘viking’ is clearly something it was thought the Danish king Eiríkr ought to put a stop to, along with the doings of outlaws, thieves and rebellious landowners (for this pejorative use of þegn, see Jesch 1993b). In contrast to the runic inscriptions, this single skaldic example does not suggest that ‘viking’ activity was considered praiseworthy, strengthening the impression given by the ways in which víkingr was used in the skaldic corpus, which were generally not complimentary, as discussed above. Since neither víkingr nor víking helps define what exactly it was that men did abroad in the late Viking Age, it will be necessary to look at the evidence of other vocabulary in both the runic and the skaldic corpus.

Viking Activities

57

Death and war The runic inscriptions and the skaldic poems that form the basis of this study are both texts in a commemorative genre: they honour the dead and record for posterity the achievements of both the dead and the living, in both cases mostly, though not invariably, men. ‘He died’ The vast majority of the inscriptions on rune stones commemorate one or more persons who were deceased at the time the inscription was cut. Although there are some thirty-two inscriptions (out of nearly 3,000 in the corpus) in which the commissioner honours him- or herself while still alive, the majority of these also commemorate someone else, whom we can often presume, and is sometimes stated, to be dead. For example, U 133 was commissioned by a certain Guðlaug in memory of her son Holmi and herself, and it is stated that han to a lankbarþal--ti ‘he died in Langbarðaland’ (on which place-name, see ch. 3). Much of this commemoration of the dead is fairly basic. Inscriptions on rune stones are, literally, lapidary, and often do not give more than the basic information of the name(s) of the commissioner(s) and the commemorated, and most often the relationship(s) between them. But a substantial minority do give further information, mostly about the deceased, and that can include a statement of where or how they died. The standard commemorative inscription does not need to point out that the deceased is dead – that is taken as given. The majority of peaceful deaths at home in old age, or deaths from disease or fatal accidents, are presumably hidden in the thousands of inscriptions which do not refer to the manner of death. But where further information is given about the place or manner of death, it is usually because that death took place abroad, or on an expedition of some sort. There are various ways of saying ‘he died’ in the language of the runic inscriptions. One of these is the verb endas ‘meet one’s end, die’. This is used in up to twenty inscriptions, always in connection with death abroad or on an expedition. These collocations are expressed in the following ways. The verb can be modified by an adverb such as austr ‘in the east’ (Sö 216, U 518). This adverb can be made more precise by the addition of the name of a region or a group of people, e.g. austr í Grikkjum ‘in the east among the Greeks’ (Ög 81, Sö 85, Sö 148, Sö 345, Sö FV1954:22, Sm 46, U 136, U 518, Vs 1). Or the adverb can be modified by some other significant phrase, e.g. austr at þingum ‘east in battles’ (Sö 33). The location of death can be specified by a regional or place-name (Sö 40, Sö 65, Sm 27, Sm 29, U 140, U 358, G 207). The verb can be followed by an adverbial phrase referring to a particular expedition, e.g. í Ingvars helfningi ‘in Ingvarr’s troop’ (Ög 155, Sö 9).17 The verb faras ‘perish’ occurs in six runic inscriptions. In four of these (U 17

Sö 216 is fragmentary and may have had further information in addition to the adverb austr. On place- and regional names, see ch. 3. The words þing and helfningr are discussed in ch. 5.

58

Ships and Men in the Late Viking Age

201, U 349, U 363, U 1016), it occurs in the phrase faras úti, where úti implies either ‘abroad’ or ‘at sea’, and quite possibly both. The other two inscriptions (Ög 145, Sö 335) both commemorate men who died in the east, the latter (and possibly the former, too) with Ingvarr. Similarly, the verb falla ‘fall (in battle)’, collocates almost exclusively with phrases indicating that the death took place abroad or on an expedition, with the only exceptions being two fragmentary inscriptions in which it is not known how the text continued after the verb fell (U 158, U ATA4909:78). The other instances refer to death as a member of an expedition led by a named leader (Ög 8, Sö 217, U 611, U 644) or in a particular place or region (Ög 81, Sö 126, ?Sö 130, Sö 171, U 346, U 356, U 374, U 616, N 239), while U 698 specifies both of these. A full complement of indicators is given in Sö 338: han fial i urustu austr i garþum lis furugi ‘he died in battle in the east in Russia, leader of the troop’. Another way of referring to the death of someone is the phrase týna aldri ‘lose (one’s) life’ (Sm 5, Vg 187), used in both cases of men who died in England. The expression láta fj›r sitt ‘lose one’s life’, is used in Sö 174 of someone who was killed (probably in battle) in Gotland. None of these expressions makes clear that the deceased was put to death by another human being, although the collocations strongly suggest that this was the case (however, for Shepard 1982–5, 246, the verbs used in the Ingvarr stones support the assertion that most of the expedition died of disease). The verb drepa ‘to kill’ is more informative on this score. Again, those who were ‘killed’ seem to have mostly met their fate abroad, or on an expedition. The participle drepinn is frequently followed by an adverbial phrase indicating the place (invariably abroad) where the commemorated person was killed (D 380, Ög 81, Ög 104, Sö 174, Sö 333, Vg 20, Vg 135, Vg 181, U 533, U 582, U 654, U 898). The remaining inscriptions are incomplete and so it is not known how the participle was modified (Sö 348, Sö ATA6163:61, Sö FV1948:291, U 324, U 577). Only one inscription definitely does not specify the location of the killing, but since it involves an accusation of treachery and the deceased is called a drengr, it is likely that the killing took place among a band of men who had gone on expeditions together, even if they were not on an expedition at the time the killing took place (D 387, discussed in more detail in ch. 6). Similarly, where the verb drepa is used in an active form (‘X killed Y’), two of the four instances are at the hands of foreigners (U 258, G 138), a third also involves an accusation of treachery between félagar (U 954, discussed further in ch. 6), and one is too fragmentary to reconstruct (Sö 351). A more neutral verb, not necessarily implying violent death, is deyja ‘to die’. This occurs up to twenty-four times in the runic corpus. Two of these inscriptions (U 243, U 613) refer to people who died í hvítaváðum ‘in (the) white clothes (of baptism)’, and two refer to people who died young (U 29, G 111), the latter a woman (U 112 also refers to the death of a woman). The others mostly record the deaths of men abroad, or at home of men who had been abroad.18 18

Abroad: D 37, ?Sö 164, Sö 173, Sö 179, ?Sm 48, U 73, U 133, U 141, U 154, U 180, U 283, U 446, U 1048, Vs 5, G 220. At home: U 1016, G 136. Uncertain: D 68, Sö 170.

Viking Activities

59

There is a similar range of usage for the phrase verða dauðr ‘become dead, die’, the most common way of saying ‘he/she died’ in the runic corpus. This occurs fifty-three times in the corpus referring to deaths abroad, but can also be used for deaths at home (Sö 55 [of someone who had been abroad], U 170), deaths í hvítaváðum ‘in [the] white clothes [of baptism]’ (U 364, U 699, U 896, U 1036), while in three instances the manner or location of death is not specified (D 66, Ög 184, Ög HOV32). On the other hand, the adjective dauðr on its own is not linked with viking activities abroad, although in some cases this may only be because the inscription is fragmentary.19 The fact that most of those whose death is specifically referred to in their memorial inscription died abroad can be put down to the necessity of recording and broadcasting a death in which there was no body, and no funeral at home (with possible implications for inheritance). It is highly improbable that the bodies of those who died in the far-flung regions of Russia, or in England, or even nearer to home, would be brought back to Scandinavia, though see the discussion of Sm 52, below (see also Wulf 1997 on deaths by drowning). But a study of these references has also shown that ‘viking’ activity brought with it a considerable risk of unnatural, violent or premature death. Some of these deaths may well have been due to disease or accident while travelling, although the inscriptions are silent on these subjects. But an important cause of death in the course of ‘viking’ activity was the active agency of another human being, either in a brawl, or in a full-scale battle. Battles and raids The word orrosta ‘battle’ (f., pl. orrostur) occurs four times in the runic corpus, in three of which the commemorated died in the battle (D 380, Sö 338, Vg 40; possibly also in Sö 126). Vg 40 goes on to specify that this was iR bþiþus kunukaR ‘when kings fought’, which may have been in Scandinavia rather than abroad (see fig. 2.4). This could have been a naval encounter, as was the battle at Utlängan (the southernmost island off Blekinge) mentioned in D 380. The whole expression þo kunukaR barþusk ‘when kings fought’ is also used on D 66, and just the verb berjask in N 252, which may refer to the battle of Bókn between Erlingr Skjalgsson and King Óláfr, although this is impossible to prove (NIyR III, 254–8).20 Sö 338 refers to a battle in Russia and similarly, in Sö 126, two daughters commemorate their father who trauh orustu i austru[i]hi ‘did battle on the eastern route’ before his death. Battles can be implied without being actually identified as such. Thus, two inscriptions in Skåne note of the commemorated that he fló eigi at Upps›lum ‘fled not at Uppsala’ (D 279 and D 295). The former goes on to state that an ua maþ an uabn a(f)þi ‘he struck while he had weapon’, making clear that this was

19 20

D 6, D 110, Sö 15, Sö 122, Sö NOR1998:23, Sm 16, Sm 28, U 620, G 270. The extremely fragmentary M 2 also appears to contain the verb berja ‘to strike, fight’.

60

Ships and Men in the Late Viking Age

Image not available

2.4 The Råda stone (Vg 40). Photo: Judith Jesch.

a battle, though we no longer can identify the occasion. (These inscriptions are discussed further in ch. 6). Less formal raids are indicated by inscriptions such as Sö 106 and Sö 166, both of which indicate that the commemorated attacked borgir ‘towns, fortresses’ abroad. The verbs used to describe these attacks are brjóta ‘to break down, destroy’, berja ‘to strike’, and sœkja ‘seek, visit, attack’, making clear that these trips to defended sites were not tourist visits.21 The skaldic corpus is not single-minded about commemorating death, since many of the poems were in honour of living rulers, while even those in praise of the dead tend to concentrate on their achievements in life rather than their moment of death. Yet because of its preference for describing battles, the skaldic corpus also has a much richer vocabulary of death, which it would take far too long to examine in every detail. However, it is possible to draw some parallels with runic usages. Thus, the skaldic corpus has a number of examples of berjask used in phrases like the runic examples discussed above. An unidentified event is described as 21

The interpretation of the final part of Sö 106, firþ han ka(r)saR kuni alaR as ‘he knew all the journey’s fortresses’ (SamRun) is, unfortunately, impossible to substantiate.

Viking Activities

61

þars flotnar b›rðusk ‘where (sea)men fought’ (Sigv XI,3), while the battle off Rauðabj›rg is þars jarlar b›rðusk ‘where jarls fought’ (Arn V,20). This type of phrase is also used of Hákon’s battle at Fitjar (ÞSjár II,1), and occurs in HHuI,53 and in non-canonical skaldic verse (AnonXI D-o-v,3). It seems to be a formulaic way of referring to a major battle on land or at sea. The urge to record a historical event is further indicated by the fact that several of the examples also name the place where the battle happened (ÞSjár II,1; Arn V,20; AnonXI D-o-v,3). Orrosta occurs in the skaldic corpus especially in poems by Sigvatr, mainly in summarising the careers of his heroes. Thus, in keeping track of Óláfr’s youthful encounters, the poet stops at the ninth one to say nú hefk orrostur níu talðar ‘now I have enumerated nine battles’ (Sigv I,9). These battles are of various kinds, and are referred to by a variety of words and phrases throughout the rest of the poem. In summarising Óláfr’s career after his death, Sigvatr notes that he háði tjogu folkorrostur ‘held twenty major battles’ (Sigv XII,22), while the battle of Stiklestad is called an orrosta (Sigv XII,15). Óláfr’s opponent Erlingr átti fleiri orrostur ‘had more battles’ than any other lendr maðr ‘aristocrat’ (Sigv VII,10), perhaps by way of establishing him as the next most powerful man to the king who killed him in his final battle. In Sigv XI,9, the poet replies to accusations that he attempted to stop King Magnús holding a folkorrosta ‘major battle’ in Sogn. Another of Magnús’ battles is called an orrosta in ÞjóðA I,8, the only other occurrence I have found in the corpus. The word borg ‘fortified place, town’ (f., pl. borgir) also appears in the skaldic corpus for towns and other defended sites attacked by vikings. Óláfr attacked a gamla borg ‘ancient stronghold’ somewhere in southern Europe (Sigv I,13). In two stanzas by Óttarr it appears as an element in the English place-names Kantaraborg and Hemingaborg, where it is in both cases a Scandinavianisation of the OE place-name element burh (discussed in ch. 3). Both names are modified by the adjective breiðr ‘broad’, suggesting certainly that Óttarr liked an easy alliteration but possibly also that large towns were still an imposing sight to Scandinavians. Óttarr addresses his patron Óláfr and tells him tókt breiða borg Kantara ‘you captured broad Canterbury’ (Ótt II,10), while Knútr makes the English bíða sorgir í breiðri Hemingaborg ‘experience sorrows in broad Hemingaborg’ (Ótt III,5). An unnamed place, which could be anywhere, as the stanza survives only in SnESkskm and is therefore difficult to place (Fidjestøl 1982, 144), receives the attentions of Haraldr harðráði: bjartr brími sveimaði of borgar virki ‘the bright fire hovered over the defences of the town’ (Valg 3). The stanza contains the parenthetical statement víkingar brutu fíkjum ‘the vikings destroyed rapaciously’, in which we have to understand the ‘defences’ as the object. Finnur Jónsson (Skjd B I, 360) adds a pronoun to his translation to make this clear. The examples of brjóta in LP demonstrate that, although this verb can be used impersonally in a passive sense (in these instances the logical subject is the same as the grammatical object), there are no other examples in skaldic poetry of the verb used absolutely, with a subject but no object. The verb is also associated with attacks on foreign towns in Ótt II,7 (London) and, as we have seen, Sö 106

62

Ships and Men in the Late Viking Age

(where the object is similarly borg). While Faulkes (1987, 144) followed Finnur Jónsson in his translation of this stanza in Snorri’s Edda, he has since suggested an alternative reading of the stanza that would provide brutu with the object borgar virki ‘the city’s fortification’ (SnESkskm, 218). Haraldr is also celebrated for having taken eighty borgir in Serkland, the Islamic world (ÞjóðA III,2; see further on this name below), while Þorfinnr jarl attacked borgir in England (Arn V,17). A more peaceful event is when Sigvatr remembers tying up his ship in Rúðuborg, Rouen (Sigv V,1), to its arm enn vestra ‘western arm’ (discussed in ch. 4, below). Similarly, B›lverkr describes Haraldr’s approach to Constantinople in which many ships skriðu at hóum armi borgar ‘glided to the high arm of the borg’ (B›lv 2). In an eyewitness description of the burning of Hedeby, the speaker says he stood á borgar armi ‘on the arm of the borg’ (AnonXI Lv,7), one end of the semicircular rampart around the town that is still visible. In all these instances, borg refers to the defensive structures of the town, whatever form they may have taken. These quotations show that the defended towns of Europe and further away were still a magnet for viking raiders in the eleventh century, even at a time when the Scandinavians already had towns of their own. By around 1100, it can be seen that towns are significant in more than one way, they become centres of religious as well as secular and economic power, as illustrated in two contrasting stanzas by Markús Skeggjason. On his expedition against the Wends, King Eiríkr attacks them í borgum ‘in towns’ (Mark I,19), which the saga-author (Knýtl, 224) interprets as í borgum eða kast›lum ‘in towns or castles’. But on his pilgrimage to Jerusalem, Eiríkr is repeatedly honoured by and met with religious processions carrying reliquaries, treasures and crosses ór borgum stórum ‘from large towns’ on his way (Mark I,29). The fall of warriors In the skaldic corpus, falla is by far the most common verb to describe the fall of warriors in battle, and a few examples will suffice here. In a stanza eulogising Erlingr Skjalgsson, Sigvatr states simply, Erlingr fell ‘Erlingr fell’ (Sigv VII,6). In a stanza addressed to Magnús, but recalling the example of Hákon Aðalsteinsfóstri, Sigvatr identifies the latter as sás fell á Fitjum ‘he who fell at Fitjar’ (Sigv XI,4). Other heroes whose notable deaths Sigvatr describes simply, using this verb, are Bj›rn stallari and, of course, St Óláfr himself (Sigv XII,18,21). It is also the verb used when warriors declare they would rather die than yield: Magnús lézk fúss falla . . . eða eiga Danm›rk ‘declared himself eager either to fall or to own Denmark’ (Arn III,5); Haraldr’s liðsmenn kuru allir meir heldr falla of fylki an vildi grið ‘troops all chose much rather to fall around the leader than they wished (to have a) truce’ (Arn VI,15). The warriors who fall in battle can be on board ship as much as on land: herr fell á þiljur ‘the army fell on the deck-planks’ (Arn V,7). Other verbs are used far less frequently. Farask is used of the death of

Viking Activities

63

(enemy) warriors in Steinn III,4. The unmarked verb deyja, relatively common in the runic corpus, hardly occurs in the skaldic corpus. Hallfreðr uses it in summarising Óláfr Tryggvason’s viking career in Britain and Ireland, he lét eyverskan her ok Íra deyja ‘caused the army of the Isles and the Irish to die’ (Hfr II,9), undoubtedly because it provided him with a useful rhyme for eyverskr ‘of the Isles’, an important word in this stanza which ranges right across Britain and Ireland. Sigvatr uses deyja to express his lifelong loyalty to Magnús: vildak með þér lifa ok deyja ‘I would wish to live and die with you’ (Sigv XI,18). The common runic collocation, verða dauðr, also found in ON prose (OGNS) does not, as far as I have been able to discover, occur in the skaldic corpus, although both the noun dauði and the adjective dauðr do occur, in a large variety of contexts. Drepa ‘to kill’ is also not common in the skaldic corpus, perhaps because it was too bald a statement to make in a genre which preferred periphrastic ways of indicating that one person has killed another. Sigvatr seems to have used it when he wanted to make a point. Thus, both the verb drepa and the noun dráp are used of the killing of Erlingr Skjalgsson in a sea-battle against King Óláfr, in two stanzas which also refer to this event as frændsekja and ættvígi, both ‘kinslaughter’, tál ‘deceit’, and morð ‘unlawful killing’ (Sigv VII,7,8; discussed further in ch. 6). Otherwise, I have found drepa in only two stanzas, Ótt III,3 referring to Knútr’s killing of English royals on his way to the throne, and Mark I,20 describing Eiríkr’s campaign against the Wends. Finally, andask ‘breathe one’s last, die’, which has the same general meaning as, but a different etymology from, runic endas ‘meet one’s end, die’, tends to occur in non-military, even Christian contexts. Thus the newly-converted Hallfreðr is happy to die if he knows his soul is to be saved (Hfr V,28) and in Þloft III,7 it is used of the death of Óláfr, in a context emphasising the salvation of his soul. In ÞjóðA II,4 it is used of the death of Magnús, who did not die in battle, though he was on an expedition to Jutland at the time (Hkr III, 105) and, according to Adam of Bremen (AB, 338), he died on board ship (obiit in navibus).

Trade Not all viking voyages had killing as their sole, or even main, purpose, although incidental death could happen at any time. The runic corpus in particular can be squeezed to provide some evidence of expeditions with other aims, though this evidence is limited and depends to a large extent on how the texts are interpreted. A lengthy article by Düwel (1987) examines the whole body of runic inscriptions for indications of trade. In some cases, these indications are quite strong. Thus, when Sö 198 states that the commemorated uft siklt til simk(a)(l)(a) t(u)ru[m] knari um tumisnis ‘often sailed in a rich/splendid kn›rr around Domesnes to the Semigallians’ (see fig. 2.5), the word kn›rr, the naming of the destination and the route, the repetition of the action, and the use of the adjective dýrr ‘splendid’

64

Ships and Men in the Late Viking Age

Image not available

2.5 The Mervalla stone (Sö 198). Photo: Judith Jesch.

(perhaps referring to the load rather than the ship) are taken by Düwel (1987, 319) as indicators of regular trading voyages, rather than any other sort of expedition. However, none of this is decisive. The ship-term kn›rr could, in certain circumstances (and especially in runic inscriptions), be used of expeditionary ships (see ch. 4), the adjective dýrr could refer to the ship itself rather than its valuable load, and the route could be a profitable one (which would be why the commemorated kept going back to it) for raiding as well as trading. As Düwel concludes in another context, ‘[i]n der Regel ist es schwierig, zw[ischen] Kriegsfahrt und Raubwiking auf der einen und H[andel]s-Unternehmen auf der anderen Seite zu unterscheiden’ (RGA XIII, 577). The clearest reference to trade is provided by a rune stone from the very end of the Viking period from Stenkumla church on Gotland (G 207). The inscription is only partially preserved, and it is not clear who sunarla sat miþ skinum ‘in the south sat with skins’, i.e. traded furs (although it is likely to have been the father commemorated by the same three sons in G 208). This trader died at ulfshala, which has been identified with Ulvshale on the island of Møn in Denmark, on a trade route to western Europe (SR XII, 208–9). Runic inscriptions on objects other than rune stones are also relevant to the study of trade (Düwel 1987, 323–4), even though they do not form part of the corpus discussed here. Thus, a copper box from Sigtuna, which may be from the Viking Age, records that its owner fik af simskum moni skalaR þis[aR] ‘got

Viking Activities

65

these scales from a Semigallian/Samian man’ (U FV1912:8).22 Scales for weighing silver were the most important equipment of a merchant, and many are known from the Viking Age. Otherwise, the best evidence for organised trading in the runic corpus comes in the four inscriptions which mention gildar (m.pl., sg. gildi), or ‘guildbrethren’, members of a guild which could have any purpose, but most likely had trade as its aim. These inscriptions consist of two from Östergötland, in which groups of men commemorate their gildi (Ög 64, Ög MÖLM1960:230) and two from Sigtuna, in Uppland, in which groups of Frísa gildar ‘members of the Frisian guild’ commemorate one of their number (U 379, U 391). These four inscriptions have much in common, despite the geographical distance between them, and provide an insight into groups of men who organised to carry on trading activities. However, they probably provoke more questions than they answer. They will be discussed in more detail in chapter 6. The skaldic evidence for trade is slightly more substantial. Arnórr refers to a kaupf›r ‘trading voyage’ that he undertook in a knarri (Arn II,2), but does not say where they went or what kind of trading was involved (the word knarri is discussed below, in ch. 4). An obscure lausavísa by Sigvatr apparently refers to the tax or landing-fee (landaurar) paid to the Norwegian king by Icelandic traders (Sigv XIII,4). In the stanza, Sigvatr asks the generous man (the king) to give (back) halfa landaura af knerri ‘half the landing-fee from the kn›rr’ and he contrasts this request for feldar ‘sheepskins’ with his previous receipt of gold from the king. Here, kn›rr clearly refers to a trading-ship (although see further discussion of this word in ch. 4). Snorri concocts an anecdote (Hkr II, 55–6; it is not in ÓsH) to explain the stanza. Both Sveinn jarl and King Óláfr had each taken half the landaurar from the Icelandic merchants in Niðaróss who, caught in the dispute between the two rulers, ask Sigvatr to intervene on their behalf. The stanza is the poet’s request to the king. The details of the anecdote are a bit unclear, and Snorri does not say what the outcome of the request was, although in the next chapter, Sveinn drives Óláfr out of Niðaróss. Although the context is obscure, the significant vocabulary in this stanza (landaurar, feldr and kn›rr) does outline the situation of Icelandic trade in the towns of Norway which is so often described in the sagas and regulated in the medieval laws (Gelsinger 1981, 61–84). However, I would suggest that Snorri has misplaced the stanza, and that it was actually composed in connection with the agreement between King Óláfr and the Icelanders (preserved in written form in Grágás and summarised in Ari’s Íslendingabók, with slightly differing details) which included a stipulation that Icelanders should pay landaurar (NGL I, 437–8; Gelsinger 1981, 71–5). In a famous stanza describing his emotional reaction to Óláfr’s death, Sigvatr states that vask fyrr kenndr á kn›rrum ‘I was previously known on knerrir’, Sigv 22

The land of the Samians is the peninsula between Kurisches Haff and Frisisches Haff in Prussia, while Semigallia is an area south of the Gulf of Riga (see Maps 1, 3 and 4 in Christiansen 1980a).

66

Ships and Men in the Late Viking Age

XIII, 26), to explain his familiarity with the cliffs of Norway, which used to seem to him to smile, but are now much gloomier that the king is dead. Taking this in conjunction with the stanza just discussed, we can interpret kn›rr as ‘tradingship’ here, and assume that Sigvatr travelled regularly on trading-voyages, most probably coming in to Norway from Iceland. Words compounded with the element kaup- also indicate mercantile activity. Niðaróss, now Trondheim, was also known as the kaupangr m. ‘market(-place)’. It is not clear whether this word is used as a proper name or an appellative in Steinn III,10 and Mberf 6 (Skjd, for instance, capitalises the first of these but not the second), although the latter is more likely (Lockertsen 1999). Gísl I,8 describes the favourable conditions created by Magnús berfœttr for merchants (kaupmaðr m., pl. kaupmenn), who wished to sell their wares at beach-markets (Msk, 306): Gramr vann g›rvan, en glatat þjófum, kaupm›nnum frið, þanns konungr bœtti, svát í Elfi øxum hlýddi flaust fagrbúin í fj›ru skorða. The prince made peace for the merchants, which the king improved, and he flattened thieves, so that it was possible to prop the beautifullyprepared ships with axes on the shore at Götaälv.23

The compound kaupskip ‘merchant-ship’ is used by Óttarr (along with kn›rr) to describe the ships in which Óláfr Haraldsson returned to Norway from England (Ótt II,13), although it is not clear whether this was as a passenger on a trading-voyage, or whether Óláfr had acquired two ships that had previously been used for trading.

Pilgrimage It should not be forgotten that the age of raising rune stones in Sweden was also the great age of its conversion to Christianity. Although most voyages took place for military or mercantile purposes, some people travelled for religious reasons. One or two rune stones may provide evidence of pilgrimage. On U 136, a woman commemorates her husband, is suti iursalir auk antaþis ubi kirkum ‘who ‘‘sought’’ Jerusalem and died up in Greece’. The verb sœkja (lit. ‘seek’) can mean ‘attack’ (as it does in Sö 166 and N 184), or it could just mean ‘visit, travel to’, and it is difficult to decide which is meant in this case, though, given the other runic examples, the former is more likely (see also ch. 3 for a military 23

The ‘prince’ and the ‘king’ are both Magnús.

Viking Activities

Image not available

2.6 Drawing by Aschaneus of the lost Stäket stone (U 605). Photo: Antikvarisk-topografiska arkivet, Riksantikvarieämbetet, Stockholm.

67

68

Ships and Men in the Late Viking Age

expedition to Jerusalem led by Haraldr harðráði). Better evidence for pilgrimage is provided by U 605, although it is now lost and is known only from a drawing see fig. 2.6). This is one of the small group of stones commissioned by someone in their own honour, here apparently a woman who uil austr fara auk ut til iursala ‘wishes to travel east and out to Jerusalem’, and presumably arranged the erection of the rune stone in case she never came back. A woman, at any rate, would presumably only make the arduous journey to Jerusalem for religious reasons. The effects of the new Christian religion are particularly keenly felt in the poetry of the eleventh century (Paasche 1914, 11–55), showing, in particular, ‘a new concern for the inner life of humanity’ (Fidjestøl 1993, 115). The skaldic corpus provides examples of both kings and poets on pilgrimage. It is well documented in other sources that Knútr went to Rome, and this journey is alluded to in Sigvatr’s Knútsdrápa: farlystir kómu fylki hafanda staf ‘urges to travel came over the leader who had a (pilgrim’s) staff’ (Sigv X,10) and he is said to have metinn feril fetum suðr ‘measured the way south with his feet’ (Sigv X,11). Sigvatr was in Rome (presumably on pilgrimage) when Óláfr was killed at Stiklestad (Sigv XIII,25). Eiríkr Sveinsson also went on a pilgrimage to Rome and Bari, visiting helga dóma ‘holy relics’ sól at bœta ‘to cure his soul’ (Mark I,11–12). Later, he set out for Jerusalem (friði tryggða byggð Jórsala ‘the peace-secured settlement of Jerusalem’) because he fýstisk læknask en iðri sór ‘was eager to doctor his inner wounds’, at grœða sól ‘to heal his soul’, and because he vildi ›ðlask bjart líf ‘wished to achieve a bright [sinless] life’ (Mark I,28). Whether they called themselves víkingar or not, the people who engaged in the activities just outlined, of raiding, trading or even pilgrimage, travelled throughout Scandinavia and much of the then-known world. In the following chapter, I outline the destinations to which men travelled in their ships in pursuit of these aims.

3 Viking Destinations . . . afar hence Seek out a foreign fastness. POUND

There is a large number of place-names, from Scandinavia and elsewhere, in the runic and the skaldic corpus, which collectively give a useful indication of the geographical range of activity in the late Viking Age, although by no means every place vikings went to is mentioned. Here I aim to survey the material in order to provide at least a partial geographical context for the ‘viking’ activities discussed in this book.1

‘East’ and ‘west’ The range of viking activities is summed up in U 504, in which a son commemorates his father who uas uistr uk ustr ‘was west and east’. Similarly, in Sö 173, the commemorated father hafþi ystarla u(m) uaRit lenki ‘had long been in the west’, although he and his son tuu austarla meþ inkuari ‘died in the east with Ingvarr’. In Vg 197, a group of brothers commemorate two of their number, one of whom uarþ tu(þ)r uestr en anar au(s)tr ‘died in the west and the other in the east’. Most trips abroad from Scandinavia were figured as going either ‘west’ or ‘east’, although in strict geographical terms, ‘south-west’, ‘south’ and ‘southeast’ would be more appropriate. It is a commonplace of histories of the Viking Age that geography determined destination, so that Swedes went east, Danes south and Norwegians west, but the evidence, from the late Viking Age at least, shows that things were more complicated than that.

1

Some surveys of foreign place-names in Old Norse sources include Ekwall 1928; Jansson 1954b; Munch 1845–60; Metzenthin 1941; Taylor 1955. There is an extensive survey of runic evidence for ‘il tema del viaggio’ in Cucina 1989.

70

Ships and Men in the Late Viking Age

The western route ‘West’ The adverbs vestr ‘in the west’ and vestarla ‘out in the west’ occur in some twenty runic inscriptions, without any further geographical indication.2 Similarly, Vg 61, discussed above, notes that the deceased uarþ tuþr o uastruakm ‘died on the west-ways [á vestrvegum]’. In three inscriptions, the adverb vestr is followed by á Englandi ‘in England’ (Sö 166, Sm 104, Gs 8). It is likely that vestr means ‘England’ in the other inscriptions as well, since their distribution corresponds fairly closely with those which mention England (without using the adverb), discussed below. A probable exception is G 370. Gotland’s location in the Baltic means that, geographically, the whole of Scandinavia is included in the direction ‘west’, as well as the British Isles, while neither the adverb vestr nor the name England appear otherwise in any other Viking Age inscription from Gotland. The skaldic uses of vestr ‘west’ and vestan ‘from the west’ bear out the suggestion that vestr usually means England, or more generally Britain and Ireland, although this is to be expected in texts in praise of subjects who are mostly Norwegian and Danish. Thus, the skalds use vestr of Knútr’s raids on England (Ótt III,2; Hallv 2), Haraldr’s expedition of 1066 (AnonXI D-o-v,8) and Magnús berfœttr’s expedition to the British Isles in 1098 (Kali; Þham I,2). Vestan is used of kings and chieftains returning to Scandinavia from the British Isles (Sigv X,7,8; Ótt II,13; BjH 4; Steinn III,6), and once, in the less-common meaning ‘in the west’, of Óláfr’s subjection of the Northern Isles (Ótt II,19). However, the adverbs could of course also be used with a more local reference, in either a Scandinavian context (Sigv III,19; ÞjóðA I,24; ÞjóðA IV,19; Halli 1), or an English one (Ótt III,5). England Some thirty runic inscriptions mention England by name. Most commonly, these commemorate men who died in or on their way to England.3 Others state that the deceased had been to England (Sö 55, Sö 207, Vs 5, Vs 18, Gs 8), but did not necessarily die there. Some of these say hann vas farinn til Englands ‘he travelled to England’, an expression which gives rise to the by-name Englandsfari ‘England-traveller’, found in two inscriptions (U 978, U 1181). Some give

2

3

D 3, D 266, Ög 68, Ög FV1970:310, Sö 14, Sö 53, Sö 62, Sö 106, Sö 137, Sö 159, Sö 164, Sö 173, Sö 260, Sö 319, Sm 51, Vg 197, U 504, U 668, G 370. Ög 83 appears to give a place-name after the adverb vestr, but the text is not at all clear. In Sö 196, both vestr and aust- appear to have a local reference. ?D 6, Ög 104, Ög FV1950:341, Sö 46, Sö 83, Sö 160, Sm 5, Sm 27, Sm 29, Sm 77, Sm 101, Vg 20, Vg 187, U 539, U 616, U 812, Vs 9, N 184.

Viking Destinations

Image not available

3.1

The Lingsberg stone (U 241). Photo: Judith Jesch.

71

72

Ships and Men in the Late Viking Age

Image not available

3.2 The Yttergärde stone (U 344). Photo: Harald Faith-Ell, Antikvarisktopografiska arkivet, Riksantikvarieämbetet, Stockholm.

Viking Destinations

73

details of what the deceased did in England: kialti skifti ‘divided up payment’ (Sö 166); tuknuts kialt ‘took Knútr’s payment’ (U 194); hafþi . . . tuh kialtakit ‘had taken two payments’ (U 241; see fig. 3.1); hafiR . . . þru kialtakat ‘has taken three payments’ (U 344; see fig. 3.2).4 The context for these is the wars in England which eventually led to Knútr’s accession to the English throne in 1017. Knútr is even mentioned by name in some inscriptions (Ög 111, ?Sö 14, U 194, U 344, N 184), while U 344 mentions two other viking leaders, Tosti and Þorkell (on the identification of these, see Jansson 1966, 12–13).5 The names of places in England are less well represented in the runic corpus, but occur three times in the context of recording the burial place of someone who died in England and whose body was not brought home. D 337 notes of the commemorated that þeR likia i luntunum ‘they lie (buried) in London’, and it is likely the fragmentary D 6 made a similar statement about the deceased, that he a enklanti i skiu -uilis ‘rests in England at Skía’, although it is not possible to identify this place. Similarly, Sm 101 says that the deceased’s brother lagþi han i stenþr . . . a haklati i baþum ‘laid him in a stone coffin in England at Bath’, correctly using the plural form of this name (i.e. ‘Baths’; see fig. 3.3), as in OE (VEPN, s.v. bæð). The skaldic corpus has fewer references to ‘England’, but many more references to English places (see fig. 3.4). These place-names have recently been surveyed and analysed by Matthew Townend (1998). Here, it is sufficient to summarise his findings as they relate to this work, and to discuss briefly the names not covered in his book. Those names discussed by Townend which come from the skaldic corpus as defined here are mainly from stanzas listing the battles fought in England by Óláfr Haraldsson and (especially) Knútr Sveinsson. Óláfr’s campaigns took him to: Hringmaraheiðr (Sigv I,7; ÞKolb III,12; Ótt II,9), the name of a heath, the first element of which survives in ‘Ringmere Pit’, near Thetford in Norfolk; Kantaraborg (Sigv I,8; Ótt II,10), Canterbury; Lundúnir (Sigv I,6; Ótt II,7), London; Nýjamóða (Sigv I,9), a lost Newmouth on the Suffolk coast; and Súðvirki (Sigv I,6), Southwark in London.6 Knútr’s campaigns (sometimes 4 5

6

Sm 104 is too fragmentary to deduce what it says happened in England. In Ög NOR1997:28, Knútr is apparently the name of a father being commemorated and so extremely unlikely to have been the king of England. In any case, the monument (of the ‘Eskilstunakista’ type) is from a later date. The long vowel in Súð- is required by the rhyme, although the ON suðr ‘south’ normally has a short vowel, while the OE sÃð, with the same meaning, had a long vowel. Townend proposes two explanations, either that ‘Sigvatr is prepared to alter the expected form of a place-name for a purely metrical reason’ (1998, 74), or that ‘the English first element appears to have been reproduced rather than the Norse cognate substituted’ (1998, 97). As Townend notes, most of the manuscripts also lack the expected -r, except the Fsk group, which have reinterpreted the name correctly, but at the expense of the rhyme. It seems to me more likely that Sigvatr, hearing the OE form, misinterpreted the first element as equivalent to ON súð ‘planking’ and the place-name therefore as meaning something like ‘wooden, planked fortification’ rather than ‘southern fortification’. I owe this suggestion to Peter Foote. The ON word súð is discussed further in ch. 4.

74

Ships and Men in the Late Viking Age

Image not available

3.3 Detail of the Nöbbelesholm stone (Sm 101), showing the place-name baþum, ‘in Bath’. Photo: Judith Jesch.

Viking Destinations

75

Image not available

3.4 Map showing places in Britain and Ireland mentioned in the text. All placenames are given in their modern forms. Chris Lewis, Cartographic Unit, School of Geography, University of Nottingham.

76

Ships and Men in the Late Viking Age

together with other leaders such as Eiríkr jarl Hákonarson and Þorkell inn hávi) took him to: Assatúnir (Ótt III,8), which Townend identifies with Ashingdon in Essex; Brandfurða (Ótt III,7), Brentford in Middlesex; Danaskógar (Ótt III,8), the Forest of Dean; Hemingaborg (Ótt III,5), traditionally identified with Hemingbrough in East Yorkshire (an identification apparently accepted in Townend 1998, 87, but not in his discussion of the name on pp. 34–6); Lindisey (Ótt III,5), Lindsey, one of the three Ridings of Lincolnshire; Lundún(ir) (ÞKolb III,11; Liðsm 10; Knútr is called Lundúna gramr ‘prince of London’ in BjH 4), London; Norðvík (Ótt III,9), Norwich; Skorsteinn (Ótt III,6), Sherston in Wiltshire; and Steinn (Liðsm 8), Staines in Middlesex. There is also a possible place-name in Liðsm 4 (Townend 1998, 33). Also in the corpus is Jórvík (Sigv X,1), York. The corpus includes river-names: Fljót (Hallv 3), the Humber; Temps (Liðsm 5,9; Ótt III,10), the Thames; Thesa (Ótt III,6), the Tees; and Úsa (Ótt III,5; Arn VI,9; Steinn III,2), the Yorkshire Ouse. It also mentions two names of peoples, or inhabitants of particular districts: Norðimbrar (Hfr II,8; Ótt III,6), the Northumbrians; and Partar (Sigv I,8), as yet unexplained (though see Poole 1980). Townend does not discuss Hrafnseyrr (Steinn III,5), identified as the lost Ravenser, in Holderness, East Yorkshire (Smith 1970, 19). This was the place from which the remnants of Haraldr harðráði’s army, led by his son Óláfr kyrri, departed from England after losing the battle of Stamford Bridge in 1066. Although Óláfr is mentioned in ASC 1066D, the place from which he left the country is not. Townend’s survey also omits any discussion of the name ‘England’ itself, which occurs at least eight times in the skaldic corpus. Sveinn Forkbeard is celebrated as having reddened swords Englandi ‘in England’ (Þjsk II). The speaker of Liðsmannaflokkr contrasts the lot of the stay-at-home with the warriors who carry shields upp á Englandi ‘ashore in England’ (Liðsm 3). Æthelred is enn ›rva þengill Englands ‘the generous king of England’ (Gunnl I). Sigvatr calls Knútr allvaldr Englands ‘ruler of all England’ (Sigv XIII,19), while Hallvarðr states that Knútr ræðr einn Englandi ‘rules England alone’ (Hallv 6). Haraldr harðráði recalls sailing norðan fyr England sunnan ‘from the north, south of England [through the Channel]’ (Hharð 16). At Stamford Bridge, he was faced by ›flugr herr sunnan of England ‘a mighty army from the south of England’ (Arn VI,12). In Þorkell Skallason’s stanza on the betrayal of Waltheof by William the Conqueror, the poet regrets that síð mun létta . . . manndráp á Englandi ‘it will take some time for the killings in England to cease’ (ÞSkall 2). Together, the runic and skaldic instances provide useful evidence both of the concept of ‘England’ and of the disyllabic form of the name England in the eleventh century. The English sources from this period still use the compound Engla-land ‘land of the Angles/English’.7

7

According to OEC, the only occurrence of the form Englond (there are none of England) is in the lost English version of a charter of King Æthelstan which many scholars think is a fabrication (ASCha, S 391) and where it occurs alongside the form Engelond. Examples of

Viking Destinations

77

The skaldic corpus also contains many instances of Englar ‘English’, usually in generalised references to opponents (Hfr II,8; ÞKolb III,12,13; Sigv I,9; Sigv XII,19; Ótt II,9; Arn V,16; Steinn III,13), though Knútr is of course konungr Engla ‘king of the English’ (Ótt IV,2). The adjective en(g)skr ‘English’ also appears, referring to the enemy army (Ótt II,11; Ótt III,5) or the landscape (ÞKolb III,9). Britain and Ireland There are no certain references in the runic corpus to anywhere in Britain and Ireland other than England, though the Scandinavian runic inscriptions found in these islands, predominantly outside England, testify by their very existence to a range of Scandinavian activities here, in the Viking Age and after (surveyed in Holman 1996). While it has been suggested that the death of a warrior recorded in Ög 81 as a tuti means ‘in Dundee’, this seems unlikely (Jansson 1987, 90). The skaldic corpus preserves quite a few insular place-names, along with the names of the various inhabitants of Britain and Ireland, subdued or ruled by the all-conquering Scandinavians. Most frequently mentioned are the Skotar ‘Scots’ (Hfr II,8; Arn V,7,11,14; Bkrepp 6; Gísl I,9), followed by the Írar ‘Irish’ (Sindr 5; Hfr II,9; Ótt IV,2; Arn V,10) and the Bretar (probably) ‘Welsh’ (Hfr III,11), also in the adjectival forms írskr (Glúmr II,2; Arn V,14) and brezkr (Hfr II,9; Arn V,14). Þorfinnr, Earl of Orkney, has a brezk skj›ld ‘Welsh shield’ (Arn V,10). In the context (in which he is reddening points in the blood of the Irish and burning), it seems unlikely that this is because such shields were particularly valued, and more likely that it was war-booty. Liðsmannaflokkr records the sound of the sword ringing á brezkum brynjum, on the mail-coats worn by the opponents of the Scandinavian attackers (Liðsm 8). Poole (1987, 292–8) has surveyed the evidence for the meaning of brezkr and concludes that it can either refer to modern Wales, or to the Britons of the Strathclyde region. The evidence is particularly strong that the poet of Liðsmannaflokkr meant ‘Welsh’, and it seems reasonable to take this meaning for the other instances listed above. Other groups of people mentioned are kumrskar þjóðir (Hfr II,9) ‘Cumbrian peoples’, referring to the Strathclyde Britons, rather than Cumbria as we understand it today; Manverjar (Bkrepp 7), ‘dwellers in the Isle of Man’; mýlsk þjóð ‘people of Mull’ (Bkrepp 6) ; eyverskr herr (Hfr II,9), ‘army of the islanddwellers’; meyjar suðr í eyjum (Bkrepp 6), ‘girls in the southern islands (Hebrides?)’; eybúar (Ótt IV,2), ‘inhabitants of the (Northern?) Isles’; Hjaltlendingar (Ótt II,19) and Hjaltar (Arn V,10,22), both ‘Shetlanders’. The two opponents encountered by Magnús berfœttr in the Menai Strait in 1098 are valskir jarlar, in this context, ‘Norman earls’ (Gísl I,10), although Hugh, Earl of Shrewsbury, and Hugh, Earl of Chester, were well-established in England. The Engla-land and Engla-lond are however numerous. According to MED (s.v. Engelond), the disyllabic form is first recorded c.1300. See also Wormald 1999a, 371.

78

Ships and Men in the Late Viking Age

other examples of the adjective valskr refer to swords (Sigv I,6; Arn III,9; Þham I,4) or to a helmet (Sigv XIII,5). Frankish swords were particularly prized in the Viking Age, and it is likely that is what valskr means in these instances. Arnórr mentions the countries of Skotland (Arn V,9; also Glúmr I,2), and Hjaltland ‘Shetland’ (Arn V,18). The Eyjar ‘Islands’ in Arn V,24 encompass either both Shetland and Orkney or, more likely, just Orkney. Other place-names mentioned in the skaldic corpus are (roughly in geographical order, in an arc from Ireland to eastern Scotland, see fig. 3.4): Dyflinn ‘Dublin’ (Arn V,23; Mberf 6); ¡ngulsey ‘Anglesey’ (Gísl I,11) and ¡ngulssund ‘Menai Strait [i.e. the strait by Anglesey]’ (Bkrepp 9); M›n ‘Isle of Man’ (Hfr II,8; Arn V,15); Sanntíri ‘Kintyre’ (Bkrepp 7); Íl ‘Islay’ (Bkrepp 7); Sandey ‘Iona’ (Bkrepp 7); Ívist ‘North Uist’ (Gísl I,9); Tyrvist ‘Tiree’ (Bkrepp 6); Skíð ‘Skye’ (Bkrepp 6; Gísl I,9); Vatnsfj›rðr ‘Loch Vatten’ (Arn V,12); Ljóðhus ‘Lewis’ (Bkrepp 5); Ekkjall ‘River Oykell’ (Arn V,9); Torfnes ‘Tarbat Ness’ (Arn V,9).8 Arnórr jarlaskáld’s Þorfinnsdrápa further provides examples of place-names from the Northern Isles: Dýrnes ‘Deerness’ (Arn V,6); Sandvík ‘Sandwick’ (Arn V,8); Rauðabj›rg (possibly) ‘Roberry (on South Walls)’ (Arn V,20); Péttlandsfj›rðr ‘Pentland Firth’ (Arn V,21); Þursasker, probably in Shetland, though it is difficult to identify where (Arn V,23).9 Further west Iceland, whose poets composed most of the texts in the skaldic corpus, and whose historians and scribes ensured its preservation for posterity, scarcely figures in it. This is partly a result of the selection of that corpus, since the Icelandic sources about Iceland, the Sagas of Icelanders, the Bishops’ Sagas, and the Sturlunga compilation, which contain verses with such content, have not been included, for reasons explained in chapter 1. We get a few hints of the stance of the Icelandic poets, but no suggestion of viking activity any further west than Ireland. On his voyage east to Sweden, Sigvatr boasts of augun þessi íslenzk hin sv›rtu ‘these black Icelandic eyes’ (Sigv III,15), his own, naturally, which have shown him the way. His nephew Óttarr calls King Óláfr stjóri markar íslands 8

9

On the identification of the names in Bkrepp, see Fidjestøl 1982, 150–52. It has been suggested that Vatsfj›rðr might be Waterford in Ireland rather than Loch Vatten on Skye (Crawford 1987, 74, 233). On Tarbat Ness and Oykell, see Fraser 1986, 26, 31. Whaley (1998, 238) suggests that Ekkjall refers to the Dornoch Firth, rather than the River Oykell, which flows into it, presumably because Tarbat Ness is a promontory at the entrance to the firth. For the suggestion that Rauðabj›rg, the site of an important sea-battle, was Roberry on South Walls, see Taylor 1938, 364. This, although admittedly right on the Pentland Firth, is a very minor place-name, and I would prefer to think the battle took place off Roeberry on South Ronaldsay. This farm is placed atop a red headland looking out over the sheltered Widewall Bay, an ideal site for a naval battle, and it is not too far from the Pentland Firth. However, it has to be admitted that there is no shortage of red cliffs in Orkney, and Marwick (1952, 177) is skeptical about the derivation of this name.

Viking Destinations

79

Image not available

3.5 Map showing places on the European continent mentioned in the text. All place-names are given in their modern forms. Chris Lewis, Cartographic Unit, School of Geography, University of Nottingham.

(Ótt IV,1), which is generally interpreted (LP; LegS, 138) as a kenning meaning ‘ruler of the icy land of the forest [i.e. Norway]’, though there is surely at least a double entendre here. The term Íslendingar ‘Icelanders’ is famously concealed in Eyv III,14, the poet calling them álhimins lendingar ‘eel-heaven’s-landers’, in which the eel’s ‘heaven’ is ‘ice’. The runic corpus is similarly reticent on the westernmost parts of the viking world, with only the enigmatic whetstone from Timans (G 216, not a part of the corpus) recording the name islat along with other geographical designations, in a context that we can no longer reconstruct. The runic inscriptions found in the western colonies testify to the export of this script there from the Scandinavian homeland, though very few of them are from the Viking Age and none are traditional memorial stones such as those of the Scandinavian homeland. Of the nine

80

Ships and Men in the Late Viking Age

currently known runic inscriptions from the Faroes, only two are from the Viking Age, and only one of those is a rune stone (FR 1), though with a somewhat enigmatic text. The 39 inscriptions from Iceland and the 92 inscriptions from Greenland listed in SamRun are all coded ‘M’, meaning they are from after 1100.

The European continent and further south The kinds of viking activities on the European continent that are so well known from sources like the Frankish annals were concentrated in the ninth century, and had died down before the period being considered here. However, the place-names found in both the runic and the skaldic corpus demonstrate that vikings still turned their attention southwards in the eleventh century (see fig. 3.5). Saxony and Frisia One runic inscription (Sö 166; see fig. 3.6) commemorates a man who had both ‘divided up payment’ in England and ‘attacked towns’ a sahkslanti ‘in Saxland (Saxony)’, the area just south of the Danevirke, now in northern Germany and marking the southern border of Scandinavia. Eiríkr’s Danish kingdom is fyr Saxland norðan ‘to the north of Saxony’ (Mark I,25) and its border is Saxa merki ‘marker of the Saxons’ (Mark I,13). The name of Frisia, Frísland, occurs in the runic text on a silver neck-ring found on the island of Senja in northern Norway, in which the first-person speaker notes that furu- trikia frislats a uit ‘we visited Frisia in a drengr-like fashion / the drengir of Frisia’ (N 540). I have discussed this inscription at length elsewhere (Jesch 1997; see also Samplonius 1998), concluding that it could refer to either raiding or trading activities. If it records trading activities, it should be seen in the light of the Swedish inscriptions that mention Frísa gildar (discussed in ch. 6). If the speaker is boasting of having raided in Frisia, it should be seen in the light of the skaldic (and other) evidence for raids in this area in the early eleventh century (discussed further below). The incomplete sequence [f]ris on a wooden plane found in Dublin (IR 6) may also have something to do with Frisians, though the context is quite unclear (Barnes et al. 1997, 25). Neither of these inscriptions is a part of the runic corpus, not being memorial stones, though both are probably to be dated to the eleventh century. Both Saxar ‘Saxons’ and Frísir ‘Frisians’ are identified as opponents of Óláfr Tryggvason on his foreign expeditions (Hfr II,6). As Fidjestøl (1982, 215) points out, however, Hallfreðr’s use of people-names rather than country-names, let alone more specific place-names, suggests a general knowledge of viking geography rather than a precise knowledge of Óláfr’s expeditions. In the following stanza (Hfr II,7), Óláfr is depicted fighting Valkerar ‘people from Walcheren (in Holland)’ and Flæmingjar ‘Flemings’. Frisians are also mentioned as Knútr’s opponents at the battle of Brentford in Middlesex (Ótt III,7). Whether Óttarr had

Viking Destinations

Image not available

3.6

The Grinda stone (Sö 166). Photo: Judith Jesch.

81

82

Ships and Men in the Late Viking Age

any real information on this, or was more concerned to find the appropriate half-rhyme and alliteration is difficult to say, though this poem is generally rather precise in its information. Poole (1987, 274) suggests the Frisians might have been innocent traders on the Thames caught up in the fighting. There is some evidence for viking raids on Frisia in the late tenth and, particularly, the early eleventh century, although it is not often discussed (for instance Blok 1978 covers only the ninth and early tenth centuries). The fullest treatment is still that of de Vries (1923, 305–8), who summarises and discusses a number of continental chronicles which attest to such raids in the first decade of the eleventh century.10 The problem seems to be how many raids there were, and when and where they took place. The sources all refer to two raids, but give different dates, and there is also some discrepancy over whether the raids were directed at Tiel or at Utrecht, or both. Either there were just two raids in subsequent years, which are variously given the dates of either 1006 and 1007 or 1009 and 1010, or there were two sets of two raids in these years, or possibly even just two raids, not in subsequent years, in 1006–7 and 1009–10. I am not competent to solve the problems of these chronicles, and in any case they do not have the kind of detail that would enable us to link these raids to any events known from other sources, although scholars have attempted (Johnsen 1916, 9–10; de Vries 1923, 304) to make such a link with the Norse accounts of Óláfr Haraldsson’s youthful expeditions, in particular as recounted in Sigvatr’s Víkingarvísur. Óláfr Haraldsson raided extensively on the continent, and many place-names, most of them insecurely identified, or not at all, are documented in this poem about Óláfr’s youth. The most commonly accepted suggestions for identifying these places are summarised in Fell (1981), though it has to be said there is room for more work here, and there may still be information buried in continental sources unfamiliar to scholars of Scandinavian studies. The sequence of stanzas in this poem is fixed by the numbering of the battles in the poem, but the identification of the places concerned is often difficult, and not helped by the fact that the prose contexts in which the stanzas of the poem are preserved do not always agree in their identifications of the locations. Thus, Óláfr is described as fighting a battle fyr hári Kinnlimasíðu ‘off high Kinnlimasíða’ (Sigv I,5). The only prose source to locate this place is Snorri’s saga of St Óláfr (ÓsH, 41; Hkr II, 13), which says it is in Frísland ‘Frisia’. LegS (42) implies that it is in England, while Fsk (167) calls the place Kinnlimafj›rðr and does not say where it is (although Óláfr goes ‘west’ from there to England); neither of these two actually cites the stanza. Following Snorri, then, modern scholars have located the place in the Netherlands, despite the fact that it is described as ‘high’, and this is explained as either a misunderstanding or poetic license by Sigvatr who, after all, had not 10

Thus, the Annales Colonienses have an attack by piratas on Tiel in 1006. The Chronicon Tielense has an attack a piratis & Danis on Tiel in 1007. Both Anfrid and Sigebert of Gembloux note attacks by Normanni on Tiel in 1009 and on Utrecht in 1010. The Chronicon Egmondanum has Nortmanni burning Utrecht in 1010. Alpertus of Metz describes at length (but does not date) an attack by Normanni on Tiel.

Viking Destinations

83

been there. They identify it with Kennemerland, a coastal district in North Holland, just north of Amsterdam (e.g. Hkr II, 466). Samplonius (1998, 93), who rolled down the sand dunes there as a boy, testifies that they can be both steep and high. Of course, we should not take everything the poet says literally, and it is true that the impression of height is reinforced in this stanza by the description of the enemy riding ‘down’ to meet the attackers (herr reið ofan), which might well refer to the sand-dunes. But height is relative, and I find it difficult to believe that any Norwegian or Icelander would describe anything in the Netherlands as ‘high’. It is also possible to question the identification on philological grounds. Apart from the superficial similarity of the first element, it is hard to see how the early forms of this place-name, Kinhem, Kinnin, Chinheim (LNT, 204), which clearly involve the second element ODu -hem, and which ON speakers should have recognised as equivalent to their -heim, would give the Norse form Kinnlima-, even if we accept that the descriptive suffix -síða ‘stretch of coast, bank of a river’ was added by the poet or other Norse speakers. Here, I believe some diligent searching by place-name experts might well turn up some plausible alternatives, possibly even in England. The next four stanzas of Víkingarvísur describe battles fought in England, and these English sections of the poem have been much discussed by scholars, while the place-names mentioned have already been noted above. Much less work has been done on the remaining stanzas which describe a series of battles Óláfr fought elsewhere in Scandinavia and on the European continent. Most of these contain troublesome place-names, and although they have tentatively been identified, these identifications are mostly conjectural. Brittany and points south The first of the ‘European’ stanzas of the poem describes an attack í f›grum Hringsfirði ‘in lovely Hringsfj›rðr’ on a hótt ból á Hóli . . . víkingar óttu ‘high dwelling in/on Hóll, vikings owned (it)’ (Sigv I,10). The prose sources have nothing to say about the geographical location of this except that Snorri specifies that Óláfr went suðr um sjá ‘south across the sea’ from England (Hkr II, 22) to get there. While ‘lovely Hringsfj›rðr’ sounds almost Scandinavian, scholars have nevertheless concurred that this stanza refers to an attack on Dol in Brittany. The similarity (if that is what it is) between Hóll and Dol would not be sufficient grounds for this assumption,11 but here we do have some important evidence from the Gesta Normannorum Ducum by William of Jumièges. According to this, Duke Richard II of Normandy called in two Scandinavian kings, named as 11

The manuscripts of both this stanza and the equivalent prose (in texts which do not cite the stanza) differ on the form of this name (see Skjd A I, 226; FskFJ, 142; LegS, 56), though in the stanza the oblique form Hóli (rather than Hœli) is substantiated by the rhyme. Either could be understood descriptively: a hóll m. is a ‘rounded hill’, while a hœli n. is a ‘place of refuge’ (LP). The plural form in Snorri’s prose (á Hólunum, Hkr II, 22) may be influenced by the use of this name for the famous Icelandic bishopric at Hólar.

84

Ships and Men in the Late Viking Age

Olaf and Lacman, to come from abroad to help him in a feud with Count Odo of Chartres. After a nasty trick in which they trapped the Bretons by digging disguised trenches into which their horses fall, we are told that they ‘laid siege to the town of Dol, and having captured it, they set fire to it and burnt it after having killed all its inhabitants including Salomon, guardian of the place’ (GND II, 26–7). They then went to Rouen to be fêted by Richard. According to the latest editor of the text, William has here confused two events, the invasion of Brittany in 1009 and Óláfr’s visit to Normandy in 1013/14, on his way back to Norway (GND II, 24–5). The only serious, but unconvincing, attempt to question this traditional identification with Dol has been by Staffan Hellberg (1980, 37–8), who does not see it as a place-name at all. The weakness of his interpretation is that it is subordinated to an overarching theory that the word víkingr in Viking Age texts refers to a man from eastern Norway. This theory is, I think, untenable, as discussed in chapter 2. There is nothing in the stanza to prevent the identification of Hóll with Dol (which was indeed a high, fortified site, and is pictured as such in the Bayeux Tapestry, see Wilson 1985, pl. 20–21, p. 215) although it would be useful if a place-name specialist could find some explanation for the form of the name. The next stanza is more problematic. Again using the fine old principle of vague similarity, scholars identify the place name Gríslupollar (Sigv I,11) with Castropol, in Asturia, on the north coast of Spain, and Fell (1981, 119) suggests that the earl Viljálmr was invented (she does not say by whom) from the place-name Viljálmsbœr also mentioned in the stanza, which she suggests is a corruption of the nearby place-name Villamea (also Johnsen 1916, 16). Elisabeth van Houts, however, has revived (1984, 118) a suggestion originally made by Munch (see Johnsen 1916, 16) that this stanza reflects an attack on Aquitaine, then ruled by William V, as related by Adémar of Chabannes, although he does not name the viking leader (AC, 176). Adémar goes on to say that the vikings performed the same trick that William ascribed to them in Brittany, of digging trenches for their opponents’ horses to fall into. This similarity can be interpreted in several ways. I am not competent to decide whether there might be a direct literary connection between Adémar (writing before 1034) and William (writing in the 1070s). If not, it might have been a literary motif used independently by the two chroniclers, or indeed it might really have happened in both places. If so, this would suggest that the same viking troop, having tried the trick once and found it successful, used it again. This may be a small additional piece of evidence to link Sigvatr’s stanza with Aquitaine rather than Spain, though it still leaves unsolved the problem of identifying the location of Gríslupollar. The other place-names in Víkingarvísur are probably even more tenuously identified. In Sigv I,12, Óláfr’s twelfth battle at Fetlafj›rðr, it has been suggested, was in the fjord by Flavium Brigantium, now Betanzos, southeast of the Galician seaport La Coruña. As Johnsen admits (1916, 17), there is little similarity between the names. There is also no other evidence in this half-stanza to help locate the spot, although the prose sources helpfully envisage this battle as vestr ‘west’ of the previous one.

Viking Destinations

85

In Sigv I,13, Seljupollar is slightly more convincingly identified as a bay by Sil (Cilenorum aqua), now Guardia, at the mouth of the river Miñho in the northwest of Spain.12 There are records of plundering activity in this region in this period (Johnsen 1916, 17). The prose sources agree that Óláfr then went up into France, as he kom sunnan upp at Leiru ‘came from the south up to the Loire’, to Varrandi, identified as a bœr ‘settlement, town’ fjarri sjá á Peitulandi ‘far from the sea in Poitou’ (Sigv I,14). The directions are clear enough in this stanza, but there is a problem with identifying the place. The name appears to be identical to Guerrande in southern Brittany, which is unfortunately neither far from the sea nor in Poitou. It has been suggested that the mistake is Sigvatr’s (Johnsen 1916, 19), who conflated Ólafr’s activities in the Loire/Poitou area on the return journey with an otherwise unmentioned raid in Brittany, presumably on the way south.13 Óttarr also refers to Óláfr’s destruction of Peita ‘Poitou’ and, more improbably, to raids in Túskaland ‘Touraine’ (Ótt II,12). Short of as-yet-unmade discoveries in southern sources, or a closer study of these place-names by onomasts with the right linguistic background, in the end we have to rely on the saga-writers to interpret the overall course of Óláfr’s expeditions. As Fell (1981, 121) has pointed out, they all mention a battle at Karlsá, probably Cadiz (Johnsen 1916, 18), which is not recorded in the surviving stanzas of Sigvatr’s poem, and show Óláfr attempting, though not succeeding, to sail through the straits of Gibraltar, so clearly they thought Óláfr had been ‘somewhere on the west coast of the Iberian peninsula’ previously. Normandy and southern Italy Van Houts (1984, 117–19) has argued that William of Jumièges, along with some other Norman writers of the period, had access to Scandinavian sources. She noted that William has Sveinn Forkbeard visiting Richard II in Rouen in 1003 and making a pact of mutual assistance with him. William is also the only source to assert that Óláfr Haraldsson was baptised in Rouen just before his return to claim the throne of Norway. Finally, like many others, she draws attention to the fact that Óláfr’s chief poet, Sigvatr Þórðarson, visited Rouen with his friend Bergr in about 1024. She concludes that ‘[d]uring their visit they might possibly have told their version of the story of Olaf’s Viking career, a version which might have lingered on in Normandy until the time of William of 12

13

The reading Sæliuvallum in LegS, 58, can probably be explained as a copying error, influenced by the fact that this stanza describes an assault on a borg ‘fortified site’. However, LegS (almost certainly wrongly) assumes that the stanza describes two different raids, and locates the ‘ancient fortress’ in western France. Locating these events is further complicated by LegS’s curious placement of them, disconnected from the previous ones, and by the fact that the writer says Óláfr went, not to Poitou, but til Væini (p. 62). Johnsen (1916, 19) has suggested this is either a misreading of Peitu, or a form of Vilaine, a river in Brittany. The Vendée, south-east of the mouth of the Loire, has also been suggested (LegS, 62).

86

Ships and Men in the Late Viking Age

Jumièges’. This conclusion is at least relatively cautious, compared to the very wild assertion of Lauren Breese (1978, 241) that ‘in 1025 a visiting skald entertained the court of Richard II with verses not likely sung in Old French’. The stanza in question (Sigv V,1) says no such thing, though it does link in an interesting way with recent archaeological work in Rouen, as discussed in chapter 4. Snorri says that Sigvatr was in Rouen on a kaupferð ‘trading voyage’, but gives no further details (Hkr II, 271). Certainly there is nothing in the surviving half-stanza to suggest raiding. It is almost too neat to see the end of the Viking Age, at least for this part of the continent, in the contrast between Óláfr’s early raiding voyages and this trading visit by his skald. Musset (1954) propounded a theory in which the first three decades of the eleventh century show a transition from ‘Scandinavian’ to ‘French’ Normandy. This was based on a small number of finds of Norman coins of this period in which there appears to be a marked shift from finds mostly in a northerly direction (in Denmark, the British Isles, etc.) to a southerly direction (in central France, Italy and even Constantinople), reflecting the reorientation of trade routes and what Musset describes as a process by which Normandy is detached from ‘le monde nordique’ and reattached to ‘le monde latin’, though the current numismatic situation suggests things are not quite so clear-cut. However, the process is just as neatly illustrated in another section of Adémar of Chabannes’ chronicle (AC, 177–9). Adémar begins by describing an attack by Normanni supradicti ‘the previouslymentioned Northmen’, and their attack on Ireland. But later on in the chapter, the same word Normanni is used of ‘Normans’, the followers of Richard II, in the context of his attack on Apulia in 1017. These are what David Bates (1982, 241) called ‘the beginnings of a substantial exodus towards southern Europe’ which he sees as resulting from ‘the collapse of an order which these relations with the Scandinavian world had in some part upheld’. From the Norman point of view, this period may well be the end of the Scandinavian connection. But turning to southern Italy in the eleventh century, we find Scandinavians still engaging in ‘viking’ activities there, though from a different direction. Indeed, Adémar’s account of Richard’s expedition to Apulia has him come out the worse from an encounter cum gente Russorum ‘with the people of Rus’, as a result of which many of the Normans were taken away to Constantinople and imprisoned there (AC, 178). In this encounter, the two types of Normanni come together, for the ‘Rus’ here are Varangians, mercenary soldiers (often of Scandinavian origin) employed by the Byzantine emperor. Both the runic and the skaldic corpus provide evidence for this Varangian activity in southern Italy, called Langbarðaland. Although the same name as ‘Lombardy’, this runic name does not refer to the northern Italian region as we know it, but was used either as a general name for the whole of Italy or, more likely, for Longobardia, that part of southern Italy that was under Byzantine control in the eleventh century (SR VI, 199).14 14

The earliest Scandinavian evidence I have found for the use of ‘the land of the Langbarðar’ to mean ‘Italy’ is in Plácitus drápa 55, when St Eustace’s wife asks to be

87

Ships and Men in the Late Viking Age

The name occurs in three or four runic inscriptions (Sö 65, ?Sö FV1954:22, U 133, U 141) and collocates with the adverbials austarla ‘in the east’ in Sö 65, and i austruiki ‘on the eastern route’ in Sö FV1954:22 (though in this the name is fragmentary and therefore uncertain). This suggests that the men who went to Langbarðaland did so via Byzantium, as soldiers of the Emperors who made war in southern Italy on and off during the eleventh century, especially in the 1030s and 1040s, although there is evidence of Varangian activity there from at least 1009 until the Eastern Empire finally gave up its hold on Italy in 1071 (Blöndal and Benedikz 1978, 51–111). Thus, from the viking point of view, this name really belongs with the eastern European names discussed below, although because of the Norman connection, it also links with this section. It can also occasion misunderstanding, in both medieval and modern accounts. Haraldr harðráði was a well-known Varangian, and he also campaigned á landi Langbarða, according to a fragmentary and poorly-provenanced couplet (ÞjóðA III,5). These lines are preserved only in a very late saga-compilation called Hulda-Hrokkinskinna, but may go back to an earlier version of Msk (Fidjestøl 1982, 40, 134). It is useful to see this stanza together with Ill I,3, also preserved in Hulda-Hrokkinskinna, and in Flateyjarbók, in which it is said that Haraldr gekk opt á frið Frakka fyr óttu ‘often disturbed the peace of the Franks before dawn’. The saga-writers (and some modern scholars following them) understood both Langbarðaland and Frakkar in the senses they had by the thirteenth century, namely ‘Lombardy (in northern Italy)’ and ‘Franks, Frenchmen’, and assumed that Haraldr had passed through western Europe on his way to Byzantium. Although neither stanza provides much immediate context, the other stanzas of Illugi’s poem are informative: they refer to Haraldr’s austrf›r ‘journey east’ (Ill I,2), and sunnl›nd ‘southern lands’ and Mikjáll, Haraldr’s employer Emperor Michael IV (Ill I,4). Thus, ‘the Langbarðaland of the poem was the Southern Italian district which formed the Byzantine province of Longobardia, and . . . the Frakkar were the French Normans who were disputing this very Langbarðaland in Southern Italy’ (Blöndal and Benedikz 1978, 56). Bari, in Apulia, was often at the centre of this struggle but, by the end of the eleventh century, it is becoming more closely associated with Rome, as when the Danish king Eiríkr Sveinsson makes a pilgrimage til Róms ‘to Rome’, but also seeks out relics út frá Rómi ‘outside of Rome’ (Mark I,12). These relics út í Bór ‘out in Bari’ (Mark I,11) were those of St Nicholas, which the men of Bari had brought back from Myra in 1087 and re-enshrined in a new church there (Blöndal and Benedikz 1978, 111–12). The cult of St Nicholas was particularly popular with Varangians, and it is characteristic that Eiríkr’s brother and

taken home Langbarða til jarðar ‘to the ground of the Langbarðar’. This poem is preserved in a manuscript dated to around 1200, and none of the prose versions or their Latin source have this name at this point, although it is clear that Italy, or more specifically, Rome, is meant, cf. the Latin ego de terra Romanorum sum (Tucker 1998, xciv, cxxi, 58–9).

88

Ships and Men in the Late Viking Age

successor was called Níkulás.15 But Eiríkr’s journey to Bari was not by the eastern route, he seems to have walked to Rome via Venice, the balkat friðland Feneyjar ‘fenced protected land of Feney’ (Mark I,10). Similarly, the fact that Eiríkr is said to have received gifts from Frakklands stýrir ‘the ruler of France’ and a ríkr keisari ‘powerful emperor [probably the German emperor]’ (Mark I,26) suggest that he took the western route to Jerusalem (Knýtl, cxli). Sicily is mentioned four times in the skaldic corpus and Haraldr’s campaign there, on the same expedition that took him from Byzantium to Longobardia, between 1038 and 1041, was clearly a high point in his career.16 Haraldr himself remembers sailing fyr víða Sikiley ‘along wide Sicily’ (Hharð 4). As Sicily is roughly triangular, with its base along the sailing route into and out of the inner Mediterranean, this stanza describes fairly accurately Haraldr’s arrival from the east. Sicily is also described as sléttr ‘smooth, flat’ (ÞjóðA III,2), in the context of Haraldr fighting a battle there. At first sight, the poet appears ill-informed, perhaps because he had not been present (this section of the poem is in a narrative style, telling of Haraldr’s achievements in a retrospective way, thus the first half of this stanza tells of his campaigns in Serkland, linked to the second half by the conjunction áðr ‘before’). However, Sicily is mostly a plateau between 150 and 580 metres above sea level, so in a sense Þjóðólfr was right. B›lv 4 describes a battle Haraldr had fyr sunnan Sikiley, which could mean either ‘in the south of Sicily’ or, more usually, ‘to the south of Sicily’. Despite the second line b›rð renndusk at j›rðu ‘the prows ran to the ground’ which appears to describe the ships landing, the stanza is best understood as an account of a sea-battle, off the coast of Sicily. The landing could have taken place after the battle, or the line might mean the ships ran close to shore and the battle took place just off shore.17 Another poet may be referring to the same occasion when he says that suðr varð Sikiley um síðir auð ‘in the south, Sicily was eventually made empty [devastated, depopulated]’ (Valg 1).18

15

16

17 18

Cormack (1994, 137–8) surveys the Norwegian and Icelandic evidence for the veneration of St Nicholas from the mid-twelfth century. Although the name is first recorded in Iceland at that time, there is evidence for it earlier in mainland Scandinavia, for instance on the Viking Age rune stones U 347 and U 631, as well as numerous medieval inscriptions. In the summary of Haraldr’s Byzantine career in CS, 97, Sicily is identified as his first campaign there, though Kekaumenos may have been misinformed. I am grateful to Charlotte Roueché for allowing me to see her forthcoming translation of this text. This stanza is extremely difficult to construe, and none of the efforts published so far is entirely convincing (Skjd B I, 355–6; NN, 1793; Blöndal and Benedikz 1978, 68). If the adverb suðr is construed with the verb verða, as here (see also NN, 806; SnESkskm, 105), this could mean that only the south of Sicily was devastated. It is also possible to construe it with the verb halda (Skjd B I, 360; Blöndal and Benedikz 1978, 68; Faulkes 1987, 150), so that the stanza says that Haraldr took his troop south. This not only syntactically less likely (disrupting an otherwise perfectly straightforward third line), but also factually less likely, given that Haraldr was approaching Sicily from the east.

Viking Destinations

89

Africa The next stanza of B›lverkr’s poem describes Haraldr sailing in difficult conditions Blálands á vit ‘to visit Africa [lit. ‘Blue-land’, or land of ‘blue’ people]’ (B›lv 5). This name appears in twelfth- and thirteenth-century skaldic verse (LP) and, frequently, in prose (OGNS). Blöndal interprets the stanza as implying ‘that the Byzantine navy had to fight Arab pirates in the passage between Sicily and Africa’, while interpreting the statement that the j›furr Affríka ‘prince of the Africans’ could not hold his country against Haraldr’s onslaught (ÞjóðA III,3) as referring to an Arab leader in Sicily (Blöndal and Benedikz 1978, 60–61, 66). However, ‘prince of the Africans’ seems an odd way to refer to a leader of the Sicilian Arabs, and it would be equally possible to interpret the two stanzas together as indicating a raid, perhaps both brief and unofficial, by Haraldr on the coast of North Africa. Þfisk 2 notes that Haraldr once destroyed blámanna fj›rvi ‘the lives of ‘‘blue’’ people’ with arrows.

The eastern route ‘East’ The adverbials austr ‘east’, austarla ‘in the east’ and í austrvegi ‘on the eastern route’ occur in some twenty inscriptions without any further geographical indication of where the person(s) concerned died, or went.19 A further twelve inscriptions add that the individual concerned died with a named leader, mostly Ingvarr, with some inscriptions specifying a destination for this expedition.20 These adverbials can be modified by the mention of specific regions in the east, all discussed further either above or below: í Grikkjum ‘in Byzantium’ (Ög 81, Sö FV1954:20, Sm 46), á Langbarðalandi ‘in southern Italy’ (Sö 65, ?Sö FV1954:22), í Garða/G›rðum ‘in Russia’ (Sö 148, Sö 338, U 209, U 636, Vs 1).21 The woman who uil austr fara auk ut til iursala ‘wishes to travel east and out to Jórsalir (Jerusalem)’ (U 605) intended a pilgrimage to the holy city. With its largely Norwegian-Icelandic point of view, the skaldic corpus uses austr for a much broader geographical range of places than the runic corpus. Thus, Sigvatr describes a journey he made on behalf of King Óláfr austr til Svíþjóðar ‘east to Sweden’ (Sigv III,1), as á austrvega ‘on the routes east’ (Sigv III, 21). This ‘Swedish’ usage of austr is common in skaldic stanzas (e.g. Sigv IX,1; Sigv XII,8; Ótt I,2; Þham I,4), and it is also used of Óláfr’s youthful raiding 19

20

21

D 108, Ög 30, Ög 145, Sö 33, Sö 34, Sö 92, Sö 126, Sö 179, Sö 216, ?Sö 308, ?Vg 135, Vg 184, Vg 197, U 153, U 154, U 283, U 366, U 504, U 898, Vs FV1988:36. On the rather obscure Scandinavian runic inscriptions found in Eastern Europe, see Kuzmenko 1995. Ög 8 (not Ingvarr), Sö 131, Sö 173, Sö 281, Sö 320, Sö 335, U 439, U 644, U 654, U 661, U 778, U FV1992:157, Vs 19. Serkland is mentioned in Sö 131 and Sö 281, and Eistaland in U 439. In Sö 121, the text following the adverbial austr is impossible to make sense of, though it is likely to have been the name of a region, as it begins with i : ‘in’.

90

Ships and Men in the Late Viking Age

in the Baltic and Finland (Sigv I,3, Ótt II,4,6). But Norwegian kings also spent time in the ‘east’ proper, in Russia and Byzantium (Glúmr I,3,5). The phrase austr í G›rðum ‘east in Russia’ that is common in the runic inscriptions also occurs a few times in the skaldic corpus (Hfr II,2; B›lv 1; Mark I,4). Similarly, the phrase austan ór G›rðum ‘from the east, from Russia’ is used of Scandinavian kings returning from that part of the world, Magnús Óláfsson (Sigv XIII,27), Haraldr harðráði (ÞjóðA III,8; Valg 5) and Eiríkr Sveinsson (Mark I,5). Magnús’ return from there, in particular, was so well known it was sufficient to say he came austan ‘from the east’ (Arn II,5,6; Arn III,2; ÞjóðA I,1), though since he returned to Norway via Sweden, austan is actually ambiguous (Arn III,3; ÞjóðA I,3). Further east, austr is used of Haraldr’s Byzantine adventures (ÞjóðA III,6), also called austrf›r ‘journey east’ (Ill I,2). The journey west to England in 1066 is described using the adverb austan ‘from the east’ (ÞjóðA IV,27; AnonXI D-o-v,8) and, similarly, the return journey to Norway is austr ‘east’ (Steinn III,6; also in Mberf 6, returning from Ireland). An Icelandic perspective is revealed when Norwegians are called austmenn ‘easterners’ (Þfagr 8), as very commonly in ON prose (OGNS). The anonymous poet who boasted that he knew alla allvalda austr ok suðr of flausta setr ‘all the rulers east and south of the seat of ships [sea]’ (AnonXI Knútr) was referring to most of the known world beyond the viking areas of Scandinavia and the British Isles. The Baltic area The nearest regions of the eastern route were the countries on the south and east shores of the Baltic sea, and several of these are mentioned in runic inscriptions, described below in roughly geographical order from south to north (see fig. 3.7).22 The secondary inscription on the Alstad stone (N 62) records the death of a man i uitahol(m)(i) ‘in Vitaholmr’, on his way to Russia (see also below). It has been conjectured that this otherwise unidentified place-name has some connection with the Witland, on the east side of the mouth of the River Vistula (NIyR I, 155–7), that is mentioned by Wulfstan in the late ninth century (Lund 1984, 23). There may or may not be some connection between this and the place called Vindau, on the coast of Kúrland, directly opposite Gotland (SR XI, 271), which is probably mentioned in G 135, in which it is said of the commemorated that he --rþ tauþr a ui(t)au ‘died in Vindau’. The Mervalla inscription (Sö 198) has already been mentioned, commemorating a man who uft siklt til simk(a)(l)(a) . . . um tumisnis ‘often sailed to the Seimgalir . . . around Dómisnes’, probably on trading voyages. Dómisnes is the Scandinavian name for the headland projecting into the Gulf of Riga from the south, while the Seimgalir are the inhabitants of Semigallia, the plains around 22

There is a useful description of this region in Christiansen 1980a, 6–47.

Viking Destinations

Image not available

3.7 Map showing places on the eastern route mentioned in the text. All place-names are given in their modern forms. Chris Lewis, Cartographic Unit, School of Geography, University of Nottingham.

91

92

Ships and Men in the Late Viking Age

the Western Dvina as it flows towards the Gulf of Riga. This region may also have been the origin of the man who gave the Sigtuna scales-box to its owner Djarfr (see ch. 2). North of Semigallia, along the east coast of the Gulf of Riga, was Lífland, ‘land of the Livs’, later Livonia, which is mentioned in one or two runic inscriptions. U 698 commemorates a man who ut fai aliflainþi frai. . ., which has been somewhat conjecturally reconstructed as ‘died abroad in Lífland, in Freygeirr’s lið (?)’, although as the stone is lost, the reading is more than a little uncertain. In Sö 39 a man records the death of his brother, who trukn-þi [a] lflanti ‘drowned in Lífland’, perhaps on the sea-voyage there, across the Gulf of Riga, rather than in Lífland itself. The country north of Livonia is Estonia, or Eistland, ‘land of the Ests’, also mentioned in one or two inscriptions.23 In Vg 181, a father commemorates his son who uarþ trbin i estlatum ‘was killed in the Eistlands’ (see fig. 3.8). This country might also be the one meant in U 439, in which two sisters commemorate their father, who ‘steered a ship east with Ingvarr’, followed by the runes askalat-, which might be interpreted as á/í Eistaland. As the stone is lost, the reading is very uncertain, particularly since Ingvarr’s expedition is usually associated with places much further east, such as Serkland (see below). But, if the conjecture Eistaland is correct, it is possible that the father died on the other side of the Baltic and never made it any further. Shepard (1982–5, 243–4) attempts to link this inscription with the reference to an expedition to Semigallia in Yngvars saga. However, these two regions are clearly distinguished in the inscriptions, and it is preferable to see this as the first death en route to points further east. Connections with Estonia are further suggested by the name Eistfari ‘Estonia-traveller’ which occurs in one runic inscription (Sö 45). This appears to be a man’s name, since it is preceded in the inscription by another name (Guðfastr), and the following verb is in the plural, though since the two names are not joined by ok ‘and’, it is also conceivable that Eistfari is Guðfastr’s nickname, and that the plural form of the verb is a carver’s error. The runic corpus contains a few names in -fari which must originally have been nicknames, e.g. Víðfari ‘Far-traveller’ in Sö 256 and U 616, or Sæfari ‘Sea-traveller’ in U 454. But where the first element is a people- or place-name, the compound is more likely to be a nickname than a given name, compare the two examples of Englandsfari already mentioned (U 978, U 1181) and one or two of Grikkfari ‘traveller to the Greeks [i.e. Byzantines]’ (?U 270, U 956). In either case, Eistfari would have been a transparent compound at the time, and provides some onomastic evidence of Estonian links, along with names such as Eistr, Eisti and Eistulfr, also recorded in the runic corpus, although these do not necessarily indicate anything more than that the names were fashionable at the time. The northeastern part of Estonia was known as Virland (cf. Estonian Viru, 23

There is an ethnographic description of Estland and its inhabitants in the ninth century in Wulfstan’s account (Lund 1984, 23–5).

Viking Destinations

Image not available

3.8 The Frugå stone (Vg 181). Photo: Harald Faith-Ell, Antikvarisk-topografiska arkivet, Riksantikvarieämbetet, Stockholm.

93

94

Ships and Men in the Late Viking Age

Virumaa), mentioned in three runic inscriptions, though these provide only two examples, since U 346 and U 356 have texts that are almost identical, and which indicate that Ragnfríðr was commemorating her son Bj›rn, who ‘fell in Virland’ (U 346 fil a urlati, U 356 fil a uirlanti). In U 533, also commissioned by a mother commemorating her son, we are told that han uas tribin a uirlanti ‘he was killed in Virland’. The land of the Tavastians, Tafeistaland, in middle Finland, is mentioned in Gs 13, where it is said of the commemorated that h-n uarþ tauþr a tafstalonti ‘he died in Tafeistaland’. This inscription is somewhat problematic, since it is not entirely clear what the brothers were up to in Finland, and it has been claimed that the inscription includes the word leiðangr (discussed in more detail in ch. 5). But Finland was an obvious destination for an expedition from Gästrikland, one of the most northerly regions in Sweden in which runic inscriptions are found. Finnland, meaning the southwestern part of modern Finland (Jansson 1954b, 47–8), is also mentioned in U 582 as the place where the commemorated was killed. Not one of these place-names just listed is mentioned in the skaldic corpus in the same form, or even in skaldic poems outside the corpus, although Ynglingatal 25 does mention Estonians. Despite the saga account of his visit to Kúrland, the coastal region south of Dómisnes, not even Egill’s poetry provides any relevant place-names, unless we believe that the Vína in Egill VII,10 is the Dvina (cp. Glúmr II,5) rather than whatever river was nearest the battle of Vínheiðr (which may or may not have been the battle of Brunanburh). However, other names from the Baltic region do occur in the skaldic corpus. The Baltic region most often mentioned there is the land of the Wends, Denmark’s nearest neighbours to the east. The great opponents of Wends in the poems of the corpus were two Norwegian kings and a jarl, Hákon Sigurðarson, Óláfr Tryggvason and Magnús góði, and the Danish king Eiríkr Sveinsson. Thus, Vinðr are among the many peoples killed by Óláfr Tryggvason (Hfr II,4), and he is opposed by Vinða skeiðr ‘ships of Wends’ assisting Eiríkr at Sv›lðr (Hókr 7). Both Óláfr and Hákon jarl are called Vinða myrðir ‘murderer of Wends’ (Eskál III,24; Hfr III,7; Hókr 6), and Hákon is opposed by Wends at Hj›rungavágr (Eskál III,28; Tindr I,4), assisting the Danes. Magnús carried herskj›ld til Venða grundar ‘war-shield to the ground of the Wends’, where he caused Venða sorg ‘sorrow of the Wends’ (Arn II,11), something they will always remember (Arn III,8). Arnórr also mentions more precisely that Magnús raided Jóm (Arn II,12; Arn III,8), the base of the legendary Jómsvíkingar a half-century previously (and possibly the Iumne described in AB, 252). But Magnús also fought the Wends in Denmark, at Skotborgará and Heiðabýr (Arn II,13; ÞjóðA I,6–7; Þfagr 1; see also Okík I,1), discussed further below.24 A battle at Ré (Arn III,9) has been identified with the island of Rügen, though the evidence is sparse (Hkr III, 46). 24

Haraldr harðráði is also called Vinða mýgir ‘oppressor of Wends’ (Þfisk 3), but the

Viking Destinations

95

Eiríkr Sveinsson of Denmark is celebrated as an enemy of the Wends (Mark I,5,8) and Markús describes his expedition to Wendland in some detail. The cause of the expedition is that the Wends þorðu at halda veldi þats buðlungr átti ‘dared to hold the realm owned by the prince’ in contravention of a treaty (Mark I,15). Thus, Eiríkr sailed his ships in hard weather fyr Vinða g›rðum ‘off the dwellings of the Wends’ (Mark I,16). There is a great land-battle and the enemy is chased off, causing sorrowful hearts í Vinða byggðum ‘in the settlements of the Wends’, when Eiríkr burns their halls (Mark I,22). The Wendland expedition is rounded off in a stanza which makes great claims for Eiríkr’s right to rule there: the Vinðr flýðu undan ‘Wends ran away’ and Eiríkr now again rules the land that fyrr lá und Sveini ‘previously was subject to Sveinn’ (Mark I,23), not his father but his great-grandfather Sveinn Forkbeard (Knýtl, xii). Óláfr Haraldsson’s youthful viking adventures, recounted in Sigvatr’s Víkingarvísur, extended to the eastern Baltic. His second battle was í eyddri Eysýslu ‘in devastated Eysýsla’ (Sigv I,2, see also Ótt II,6), where the place-name represents the Baltic island of Ösel, just north of Dómisnes and at the entrance to the Gulf of Riga. Presumably, it was laid waste as a result of his battle, rather than before. The next stanza states clearly that Óláfr was fighting against Finnlendingar ‘Finlanders’ (Sigv I,3), so it is a reasonable assumption that the place in which he did so, Herdalar, is in Finland, though the location has not been identified. In the same stanza, the poet describes Óláfr’s ships sailing austr ‘east’: Bálagarðssíða lá brimskíðum at barði ‘Bálagarðssíða was before the prow of the sea-skis [ships]’. This is usually explained as the southwest coast of Finland, though on rather flimsy, not to say non-existent, grounds. Eiríkr Hákonarson, who raided extensively in Russia (see below), may also have turned his attention to the Baltic. Edáð 8 says that he fór herskildi of allar Sýslur ‘went with war-shield through all the Sýslur’, which Snorri (followed by modern scholars) interprets as um alla Aðalsýslu ok Eysýslu ‘through all of Aðalsýsla and Ösel’ (Hkr I, 339–40), where Aðalsýsla is the Estonian mainland (Hkr I, 373). This may just be Snorri’s conjecture. The word sýsla means ‘district’, and could be used in this stanza as an appellative in the plural. The first half of the stanza has Eiríkr fighting with Gautar ‘people from Götaland’, and it is rather a jump from Sweden to Ösel and Estonia. Russia Once across the Baltic, viking voyagers had a number of routes and destinations open to them, some near and others much further away. Þóraldr, the Norwegian who died at Vitaholmr (see above), was miþli u(i)taulms auk karþa ‘between Vitaholmr and Garðar’ (N 62), or on his way to the Garðar (m. pl.), the ‘multi-ethnic trading and handicraft centres’ (Noonan 1997, 144–5, see also map authenticity of this stanza is doubtful (see further below), and there is no other evidence that Haraldr was active in the Baltic region.

96

Ships and Men in the Late Viking Age

on p. 136) found along the Russian rivers and through which Scandinavians traded with the Islamic world. Garðar is used as a name for the whole region in which these towns were found, which is perhaps best called ‘European Russia’ and defined as ‘the entire area between the Arctic and the Black seas and between Poland and the Urals’ (Noonan 1997, 134). Although this trading activity was slowing down in the eleventh century when the bulk of our runic and skaldic evidence is from, Scandinavians were still active in this region, if only as mercenaries in either Russia or Byzantium (Shepard 1982–5, 223–31, Noonan 1997, 154–5).25 The name Garðar occurs in eight or nine runic inscriptions (Öl 58, Sö 130, Sö 148, Sö 338, U 209, U 636, ?Vs 1, G 114, N 62). Four of these indicate that the deceased died there,26 and since Sö 130 (probably) says that he ‘fell’ there, while the commemorated of Sö 338 fial i urustu austr i garþum lis furugi ‘died in battle east in Garðar, leader of the troop’, it is very likely that the men commemorated in these inscriptions were active as mercenaries for the rulers of Novgorod and Kiev, particularly Iaroslav, ‘the last great patron of the Varangians among the Rus’ (Franklin and Shephard 1996, 201). On the other hand, Þorsteinn, commemorating his son in U 209, kaubti þinsa bu auk aflaþi austr i karþum ‘bought this farm and earned the money east in Garðar’, and may have been active in trade (although presumably mercenaries were also well paid), the profits of which he ploughed back into a good life at home in Uppland. The inscription is carved in an embedded rockface and still testifies today to the financial origins of the farm at Veda (see fig. 3.9). Further testimony to Scandinavian activities, possibly of a financial nature, in Russia comes in the form of a rune stone raised in Russian territory itself. Excavations on the island of Berezan, near the mouth of the Dniepr as it flows into the Black Sea, revealed ‘the gable-stone of a coffin’ with an inscription in Scandinavian runes: krani kerþi half þisi iftir kal filaka sin ‘Grani made this sarcophagus in memory of Karl, his partner’. The implications of the term félagi for what kind of venture they were partners in will be discussed in chapter 6. But the stone is testimony to the wide-ranging activities of Scandinavians in the east, as is the now-illegible inscription on the lion from Piraeus, now in Venice (Jansson 1987, 62). Also in this southern region lay Wallachia, referred to in one runic inscription (G 134) which states that the deceased was betrayed by blakumen (possibly) ‘Vlachs, Wallachians’ on an expedition (this inscription is discussed further in ch. 6). 25

26

It may be that the Garðar were only three: Hólmgarðr (Novgorod), Kœnugarðr (Kiev, the name is recorded only in later sources) and Miklagarðr (Constantinople), ‘the beginning, the central point, and the end of the Austrvegr’ (Melnikova 1996, 15). These are the three from Södermanland, and Vs 1, if, as I think likely, the sequence karusm is a misspelling of karþum rather than a representation of the name of the very distant Khorezm in Central Asia, from which Arab merchants travelled to ‘the Bulghar markets on the middle Volga’ (Noonan 1997, 138; see also Franklin and Shephard 1996, 64).

Viking Destinations

97

Image not available

3.9 The Veda rock (U 209). Photo: Judith Jesch.

To reach the Black Sea by the river route, viking voyagers had to pass the fearsome Dniepr rapids, the Norse names of which are recorded in Constantine Porphyrogenitus’ De Administrando Imperio (Blöndal and Benedikz 1978, 9–12; Svane 1989, 21–2; Franklin and Shephard 1996, 92, 113), written c.950. One of these rapids is called Aeifor and this name appears to be recorded on a tenth-century rune stone from Gotland (G 280; see Krause 1952). Four brothers commemorate someone called Hrafn, whose relationship to them is not specified. It is said of all of them that they kuamu uit i aifur ‘got far to Eifur’. The inscription also mentions that they placed stones for Hrafn suþ fur. . . ru-s-aini ‘south of Rufsteinn’, presumably also a place in that area. In the eleventh century, Novgorod had replaced Ladoga as the main centre of power in north-western Russia, and this is reflected in the runic inscriptions, in which the only other specific place-name from this region mentioned is Holmgarðr, lit. ‘island-enclosure’, originally the Swedish name for Gorodishche, situated on an island, later transferred to the nearby ‘new town’ Novgorod (Franklin and Shephard 1996, 40, 130). Three inscriptions commemorate men who died there, including the fragmentary G 220. In Sö 171, the commemorated Sigviðr was a skaiþaR uisi ‘ship’s captain’ (see ch. 5), who [fial i h]ul(m)karþi ‘fell in Holmgarðr’, and he too may have been a mercenary. In U 687, Spjallboði uaR tauþr i hulmkarþi i olafs kriki ‘died in Holmgarðr in Óláfr’s church’, which presumably means that he was buried there, in a church dedicated to St Óláfr (d.1030) and frequented by the Scandinavians of Novgorod.

98

Ships and Men in the Late Viking Age

Garðar also occurs in the skaldic corpus with the same meaning as in the runic inscriptions, in a number of stanzas to do with the various Norwegian kings and jarls who spent time in the east as exiles, or who raided there. The rather colourful account given in the sagas of Óláfr Tryggvason’s youth, which included some time at the court of Vladimir, prince of Kiev, and raiding on his behalf, is hardly substantiated by, but referred to fleetingly in, Hfr II,2, in which Óláfr is said to have reddened weapons austr í G›rðum.27 Eiríkr Hákonarson, on the other hand, fór eyða land Valdamars ‘went to ravage Vladimir’s land’, and the poet tells him that brauzt Aldeigju . . . komzk austr í Garða ‘you destroyed Ladoga . . . you went east to Garðar’ (Edáð 6), an expedition that may be detected in the ‘traces of the violent destruction of Staraia Ladoga’s defence wall and a conflagration around the turn of the tenth and eleventh centuries’ (Franklin and Shephard 1996, 169). Óláfr Haraldsson also spent time in exile in Russia, as it was rather unkindly put in a poem in praise of his opponent Kálfr Árnason: vígmóðr bróðir Haralds varð at vitja Garða ‘the battle-bold brother of Haraldr [harðráði] had to visit Garðar’ (BjH 3).28 Óláfr’s young son Magnús went into exile in Russia with his father in 1028, and stayed there after the king’s return to Norway and death at the battle of Stiklestad in 1030. The sagas tell us that, tired of the Danish rule imposed by Knútr and carried out by his son Sveinn and his mother, the English Ælfgifu, some Norwegian nobles went to Russia to fetch Magnús back to Norway in 1034, when he was still only ten years old. Magnús’ return from Russian exile is mentioned in a number of skaldic stanzas, with the name Garðar in one (Sigv XIII,27) and the adjective gerzkr ‘from Garðar, Russian’ used of equipment and weapons on his ship in two more (Arn II,4,9). Arnórr’s prayers to God, whom he calls v›rðr Girkja ok Garða ‘guardian of the Greeks and of Garðar’ (Arn VI,19), on behalf of Haraldr harðráði, may allude to his hero’s adventures in those two regions: according to the sagas, Haraldr also went to Russia after the battle of Stiklarstaðir. Haraldr’s own gamanvísur ‘pleasurestanzas’ (Hharð 3–7) use as their refrain the couplet Þó lætr Gerðr í G›rðum / gollhrings við mér skolla ‘Yet the goddess of the gold ring [= woman] in Garðar keeps her distance from me’. This refrain has nothing to do with the stanzas it accompanies and is no more than an oblique reference to Haraldr’s sojourn in Russia. King Eiríkr Sveinsson of Denmark is also said to have been austr í Garða ‘east in Garðar’ (Mark I,4), returning ór G›rðum austan ‘west from Garðar’ in the following stanza (Mark I,5). This expedition is presented as visits to the rulers (foldar v›rðu) of the region, who give him gifts. As this theme of recognition and gifts from foreign potentates is prominent later in this poem (Mark 27 28

The occurrence of the same phrase in Hfr II,1 is almost certainly a scribal borrowing from Arn III,1 (Whaley 1998, 183–4). A possibly spurious stanza by the king himself (Ólhelg 11) refers to Garðar, in the context of the marriage of Ingigerðr, the Swedish princess Óláfr had hoped to marry, to Jarizleifr, or Iaroslav the Wise, son of Vladimir.

Viking Destinations

99

I,26,29), too, there may be a certain amount of hyperbole involved, particularly as Eiríkr is said to have become popular of Austrveg allan ‘throughout the whole of the eastern way’. The saga-author explains that Eiríkr cleared the region of heathens, so that Christians and merchants could travel through it freely (Knýtl, 212), but there is no evidence for this in the surviving stanzas and it may be the prose author’s conjecture (Knýtl, cxli). Byzantium and Jerusalem Better-known than the Scandinavian mercenaries in Russia are the mercenaries working for the Byzantine Emperor in Constantinople, known as the ‘Varangian Guard’ and made up of warriors of many nationalities, including Scandinavians. The runic corpus has many instances of the name the Scandinavians had for the Byzantine Empire: Grikkland, Grikk(j)ar ‘(land of) the Greeks’. Grikkland occurs in three inscriptions (U 112, U 374, U 540), the second and third of these both indicating that the commemorated had died out there. In U 112, Ragnvaldr, ostensibly commemorating his mother, boasted of himself that he had been a griklanti uas lis forunki ‘in Byzantium, was leader of the lið’. The meaning of the word lið is discussed in more detail in chapter 5, but it is likely that Ragnvaldr was boasting of a high position in the Varangian Guard, perhaps leading a ship-borne troop that he had brought with him from Sweden. Some twenty-four inscriptions indicate that the man being commemorated ‘was’ or ‘died’ í/með Grikk(j)um ‘among the Greeks’ or went out til Grikk(j)a ‘to the Greeks’.29 They rarely specify what it was he did out there, but many of these men are likely to have been active in Byzantium as mercenaries. It is clear from several inscriptions that a voyage to Grikkland could be extremely profitable. Thus, two inscriptions (Sö 163, Sö 165) indicate that the commemorated gulli skifti ‘divided gold’ among the Greeks. Düwel (1987, 353) has suggested that this expression is the eastern equivalent to those in which the deceased is said to have ‘taken’ or ‘divided’ a payment in England, where the word used is gjald (cp. Sö 166 kialti skifti ‘divided payment’, see also U 194, U 241, U 344). The source of these payments is less clear but, since the ‘western’ inscriptions indicate they were payments by war-leaders, it is likely that the same was the case in these ‘eastern’ examples. There were consequences for inheritance when men died abroad in Byzantium, as detailed in the inscription on the paired rune stones U 72–3. The inscription is commissioned by two brothers to commemorate the sons of their sister, and goes on to explain that hon kam þeira at arfi in þeir brþr kamu hnaa at 29

Ög 81, ?Ög 94, Sö 82, Sö 85, Sö 163, Sö 165, Sö 170, Sö 345, Sö FV1954:20, Sm 46, Vg 178, U 73, U 104, U 136, U 140, U 201, U 358, U 431, U 446, U 518, U 792, U 922, U 1016, U 1087. The nominative form krikiaR is recorded only in G 216 (not in the corpus), in an obscure context with other geographical names, from the second half of the eleventh century. On this kind of use of the name of the inhabitants of a country rather than a country-name, see Jansson 1954b, 34.

100

Ships and Men in the Late Viking Age

arfi . . . þir to i kirikium ‘she inherited from them and the [i.e. her] brothers inherited from her . . . they [her sons] died among the ‘‘Greeks’’ ’. What is not clear is whether the inheritance included any cash the younger brothers had earned out in Byzantium, perhaps brought back by their comrades, for the inheritance could just have been a matter of the farm they inherited from their father, which could not go directly to his wife while there were living children. There is however a clear example of money from the east going to the voyager’s heirs in U 792, in which a son commemorates his father thus: far aflaþi uti kri[k]um arfa sinum ‘he earned money out among the ‘‘Greeks’’ for his heir’. Trade was another possible source of money in the east. In U 1016 one of the sons being commemorated by his father died at home, but had in his lifetime ‘steered a kn›rr’ (see ch. 4 for further discussion of this ship-term) and kuam hn krikhafnir ‘he arrived in ‘‘Greek’’ harbours’. The plural form of this last word suggests a trading voyage to several markets. The element Grikk- also occurs in the nickname Grikkfari given to Ketill (probably, U 270) and Viðbj›rn (U 956). The names Gríkland and Gríkir or its metathesised variant Girkir occur in the skaldic corpus,30 particularly in stanzas relating Haraldr harðráði’s notable career in Byzantium (on which see CS, 97; Blöndal and Benedikz 1978, 54–102),31 as captain of the Varangian guard in Constantinople or Miklagarðr (the name is recorded in B›lv 2). As already noted, Arnórr’s invocation of God as v›rðr Girkja ok Garða ‘guardian of the Greeks and Russia’ (Arn VI,19) in a poem in praise of Haraldr probably alludes to his adventures in these two regions. However, God is also called gætir Gríklands ‘protector of Gríkland’ in a couplet (Þloft I), probably a refrain, which is all that survives of a lost poem on Knútr, with no similar implications about Knútr’s career. The Emperor Michael V, stólþengill Gríklands ‘throne-prince [i.e. emperor] of Byzantium’ (ÞSkegg) and both stólþengill and stillir Girkja ‘emperor’ and ‘ruler of the Byzantines’ (ÞjóðA III,6), had his eyes put out by Haraldr after the successful uprising of 1042, as these stanzas make clear. According to the former stanza, Haraldr náði enn gørr handa glóðum ‘acquired even more hand-embers [gold]’ as a result (on this event, see Blöndal and Benedikz 1978, 88–96).32 He was also apparently the agent of punishment by hanging of those Varangians who had remained on the emperor’s side, so that eru Væringjar færi ‘the Varangians are fewer’ (Valg 4). Haraldr is proleptically called Bolgara brennir ‘burner of Bulgars’ in ÞjóðA

30 31

32

On the forms of these names, see Jansson 1954b, 34–5. CS is an important near-contemporary source for Haraldr’s career, written by Kekaumenos, a retired Byzantine general writing in the 1070s. Charlotte Roueché of King’s College London is preparing an edition and translation, and there is a preliminary account of the text and its author in Roueché 2000. The reading of this stanza is that of Hkr III, 86. For some alternatives, see Blöndal and Benedikz 1978, 94. Stólþengill might be a corruption of an Old Russian term (Blöndal and Benedikz 1978, 5), but is perfectly transparent as an ON compound, with the first element stóll m. ‘throne’.

Viking Destinations

101

III,1, a stanza about his assistance to Óláfr at Stiklarstaðir. This most likely refers to the revolt of Delianus in Bulgaria in the summer of 1040 (Blöndal and Benedikz 1978, 74), which Kekaumenos (CS, 97) says Haraldr helped the Emperor to put down. Earlier in his Byzantine career, Haraldr had gone on an expedition to Jerusalem, as outlined in two stanzas by Stúfr who, according to Snorri (Hkr III, 83), had his information from the king himself. Thus, Haraldr fór ór Girkjum leggja und sik Jórsali ‘went from Byzantium to subjugate Jerusalem’, which he managed to do in such a way that the land remained óbrunnin ‘unburned’ (Stúfr 2). The expedition is figured as one of punishment for the treacheries and crimes of the inhabitants (Stúfr 3; Hkr III, 84):33 Stóðusk róð ok reiði, rann þat svikum manna, Egða grams á ýmsum orð Jórðánar borðum. Enn fyr afgørð sanna, illa gót, frá stilli þjóð fekk vísan váða. [the last line is part of a refrain] The advice and words of anger of the prince of Agder are firm on either side of the Jordan; that brought the treacheries of men to an end. The people again got certain trouble from the ruler for (their) proven transgression, (their) wicked disobedience.

The suggestion that Haraldr conquered Palestine for the Emperor seems unlikely, and it has been claimed that the poet misunderstood what the king had said about his expedition, and that what he had in fact done had been to provide a protective escort for pilgrims to Jerusalem (Blöndal and Benedikz 1978, 64–5). This might at any rate explain why the land was not burned, or otherwise devastated, and the poet’s claims could be seen as an exaggerated way of saying that Haraldr controlled routes through Palestine, including the crossing of the Jordan.34 The Danish king, Eiríkr Sveinsson, went, for quite different reasons as we have already seen, at kanna friði tryggða byggð Jórsala ‘to get to know the settlement of Jerusalem, secured by peace’ (Mark I,28). On this same journey (c.1102–3), Eiríkr continued to Constantinople, where he was welcomed and received gifts af harra sj›lfum í Miklagarði ‘from the ruler himself in Constantinople’ (Mark I,30). These gifts consisted of a large amount of gold, some royal robes and fourteen warships. The Emperor at the time was Alexius I Comnenos, a ruler of ‘prudence and financial acumen’ and it is possible that these payments

33 34

Msk and Fsk have a slightly different text, but with essentially the same message. On two runic inscriptions that mention Jerusalem, see ch. 2. The Timans whetstone (G 216) has the name iaursaliR in the context of other names, but its significance is not clear.

102

Ships and Men in the Late Viking Age

to Eiríkr were in fact by way of payment and outfitting of a fleet for those of his followers who ‘entered the Imperial service as Varangians’ (Blöndal and Benedikz 1978, 135–6). Ingvarr’s expedition An expedition to the east that stands out because it is extensively commemorated in the runic corpus (and even found in fictionalised form in a romantic Icelandic saga, Yngvars saga víðf›rla) was that led by a certain Ingvarr, to the east and to Serkland, probably ending in 1041 (Shepard 1982–5, 255–8). Whatever happened on this expedition, it seems to have been something of a disaster, since so many of its members, including its leader, never returned, and it is these members that are commemorated in a large number of inscriptions from the Mälar valley region, predominantly in Södermanland and Uppland. The inscriptions can be identified by their mention of the leader of the expedition, Ingvarr. Although a name Ingvarr occurs in some forty-four inscriptions in all in the corpus, only those in which a person of this name is neither the commemorated nor related to the commemorated (with one exception, see below), can be considered ‘Ingvarr’-stones. Thus, although in Ög 30 a certain Sigsteinn commemorates ikuar sun sin han uarþ austr tauþr ‘Ingvarr, his son, he died in the east’, the deceased just happens to have the same name as the more famous Ingvarr, and happens to have died somewhere in the east, but not necessarily on Ingvarr’s expedition. Only on the Gripsholm stone (Sö 179) does Ingvarr’s name occur in a family context (though see also discussion of Sö 279, below). The inscription was commissioned by a woman, Tóla, commemorating sun sin haralt bruþur inkuars ‘her son Haraldr, brother of Ingvarr’. From this formulation, it is likely (though cannot be proved) that the two men shared a father, but not a mother, although it is also possible that the word ‘brother’ is used metaphorically and that they were ‘brothers in arms’. The inscription goes on to describe, in an often-quoted poetic statement, what the brothers did (presented here in normalised form and set out in lines of verse): Þeir fóru drengila fjarri at gulli ok austarla erni gáfu, dóu sunnarla á Serklandi. They travelled in a drengr-like35 fashion, far for gold, and in the east gave (food) to the eagle, died in the south in Serkland.

Although rather a simple little verse, this commemorative stanza is metrically regular (equivalent to Old Icelandic fornyrðislag) and even uses the poetic 35

See ch. 6 for further discussion of this word.

Viking Destinations

103

conceit by which warriors are said to ‘feed the eagle’ or other beast of battle (see ch. 6). Some twenty-odd stones state that the commemorated travelled out or died on this expedition, most commonly using the phrase með Ingvari ‘with Ingvarr’.36 A few indicate more specifically that this was in ‘Ingvarr’s troop’: í Ingvars helfningi (Ög 155); í lið(i) Ingvars (Sö 254, U 778, ?U 837). A fragmentary inscription (Sö 277) almost certainly contains the phrase Ingvars manna ‘of Ingvarr’s men’, and is likely to belong to this group, though the context of this phrase can no longer be reconstructed. The exact number of inscriptions associated with Ingvarr’s expedition cannot be determined for certain. Above are listed twenty-one likely and two possible examples. There are a further three which may also belong in this category, because, although they are fragmentary, they are similar in phrasing to those which are better preserved. Thus, Sö 96 currently ends han uaR fa. . . and has been compared to Sö 105 and Sö 107 which both state of the deceased that hann var farinn með Ingvari ‘he went away with Ingvarr’.37 Ög 145 contains the phrase hilfnai (a)(u)str, which has been interpreted as (í) helfningi austr. This is a candidate for inclusion because the only other runic occurrence of this noun (discussed further in ch. 5) is in the nearby Ingvarr-stone Ög 155. Finally, the sadly fragmentary Sö 279 contains two significant sequences, suggesting that this might be a memorial to Ingvarr himself. The inscription commemorates . . .uni aimunt. . ., probably ‘. . . son of Eimundr’. In the Icelandic Yngvars saga, the hero’s father is indeed known by this name. The inscription concludes sunarla a se(r)kl. . . ‘in the south in Serkland’, reminiscent of the formulation of the Gripsholm inscription quoted above. If these three inscriptions are included, this gives a total of twenty-six stones now known and associated, with greater or lesser degrees of certainty, with Ingvarr’s expedition. Larsson (1990, 46–57, 107) has only twenty-five because he did not yet know of the latest Ingvarr-stone (U FV1992:157), found in 1990 during road-building at Arlanda airport (and now on display in the terminal there). No doubt further Ingvarr-stones will be discovered, and it is likely that some of those of whom it is only said that they ‘died in the east’ also perished with Ingvarr. Many attempts have been made to reconstruct what happened on Ingvarr’s expedition and when it happened, using a range of written sources, most of them not contemporary with the rune stones (for brief surveys with references to earlier studies, see e.g. Ruprecht 1958, 55–6; Larsson 1990, 106–14). Shepard 36 37

Sö 9, Sö 105, Sö 107, Sö 108, Sö 131, Sö 173, Sö 281, Sö 287, Sö 320, Sö 335, U 439, U 644, U 654, U 661, U 1143, U FV1992:157, Vs 19. It is possible that Sö 108 was also intended to end in this way. Brate reads han : uaRi : faru (SR III, 80), but the short vertical cut he reads as i looks much more like a divider, and I could not see any trace of ‘nedre punkten i det följande kolon’ which Brate thought he could find. The rune which Brate reads as u admittedly does not look much like in, but neither does it look like the carver’s other us: the curved branch may represent the carver’s attempt at a bind-rune of i+n after he realised he had missed out the n.

104

Ships and Men in the Late Viking Age

concludes from a study of the saga and the runic inscriptions that most of the participants died of disease (1982–5, 246, see also his ‘reconstruction’ of the events, pp. 272–3). Fuglesang (1998, 206) attempts an art-historical dating of the inscriptions and concludes they were carved ‘around 1000–25’, suggesting ‘that the late Icelandic texts be disregarded’. To pursue these questions would be to go far beyond the parameters of this work. But it should perhaps be stressed more than it has been hitherto that the inscriptions, taken together, refer to both an eastern (austr/austarla: Sö 131, Sö 173, Sö 179, ?Sö 281, Sö 320, Sö 335, U 439, U 644, U 654, U 661, U778, U FV1992:157, Vs 19) and a southern (sunnarla: Sö 179, Sö 279) destination for the expedition. This accords well with Shepard’s reconstruction in which the first stage of the expedition was to Russia, although the ultimate goal was ‘booty from the Moslem littoral of the Caspian Sea’ (1982–5, 271). Serkland The name Serkland is strongly associated with Ingvarr’s expedition in the runic corpus, where it occurs five times. As we have seen, it is used in the Gripsholm inscription to indicate the death-place of Ingvarr himself, as possibly also in Sö 279. Sö 131 says of the commemorated that he fur austr hiþan miþ ikuari o sirklanti likR sunR iuintaR ‘went east from here with Ingvarr, in Serkland lies the son of Eyvind’ (see fig. 3.10). Sö 281 also associates Ingvarr and Serkland, though it is fragmentary, so the exact context is unclear. U 785 commemorates someone who uarþ tuþr a srklant- ‘died in Serkland’, though there is no mention of Ingvarr on this more-or-less complete rune stone.38 The etymology of this name is uncertain, with the two most popular explanations being that it means ‘land of the Saracens (Muslims)’ or that it derives from Lat. sericum ‘silk’ and designates those regions producing silk (Shepard 1982–5, 235). Also uncertain is the geographical area it is supposed to represent. Ruprecht (1958, 55) suggested that this gradually expanded throughout the Viking Age, so that in the early Viking Age it referred to the area south of the Caspian Sea, in the late Viking Age it referred to all the Islamic areas beyond Russia, and for the saga-writers it also included North Africa, though what his evidence was for these assertions is not entirely clear. The Caspian region is favoured by most scholars as the destination of Ingvarr’s expedition, though some suggest that his band fought in Byzantine service in Syria or eastern Asia Minor, as did the Norwegian king Haraldr harðráði (Shepard 1982–5, 222–3, 234–8). Serkland has also been explained as derived from the name of the city of Sarkel and its surrounding region in the territory of the Khazars, northwest of the Caspian Sea (Jarring 1983).

38

The name also occurs (in the form serklat) on the Timans whetstone (G 216), from the second half of the eleventh century, but what the reason was for noting this and the other names in the inscription, we do not know.

Viking Destinations

Image not available

3.10 The Lundby stone (Sö 131). Photo: Judith Jesch.

105

106

Ships and Men in the Late Viking Age

Serkland and its inhabitants the Serkir also appear in the skaldic corpus, in connection with Haraldr’s Byzantine adventures. His court poet alludes to this period in a formal praise poem, Sexstefja, when he states that Haraldr was Serkjum hættr ‘dangerous to the Serkir’ and that átta tøgu borga má segja tekna á Serklandi ‘eighty towns can be said to have been taken in Serkland’ (ÞjóðA III,2). Admittedly, the description Serkjum hættr appears in the second half of the stanza, leading some commentators to surmise that the poet ‘called Sicily ‘‘Serkland’’ because it was occupied by Saracens’ (Shepard 1982–5, 236). But such phrases are frequently used of kings, often referring back to earlier episodes, and here it makes a link between the two halves of the stanza, but does not necessarily imply that Sikiley was the same as Serkland. The other three stanzas that mention Serkland occur in an anecdote preserved in Msk, 247–50, in which the king, the court poet and a Norwegian fisherman vie in the extempore composition of poetry. Haraldr remembers how long ago fjarri fóstrlandi rauðk branda í Serkja garði ‘far from my homeland I reddened swords in the enclosures of the Serkir’ (Hharð 15). Echoing the king’s stanza in the contrast between recent battles against Danes and long-ago battles further away, Þjóðólfr also notes that the king setti merki niðr á sléttu Serklandi ‘placed his banner down in flat Serkland’ (ÞjóðA IV,13). The king criticises Þjóðólfr’s stanza and praises the fisherman’s verses, especially the final stanza (perhaps because it echoes his own). In this, the fisherman recalls his fighting fyr sunnan Serkland ‘in/to the south of Serkland’ (Þfisk 3). This seems unlikely if Serkland is (in) North Africa, for there is no evidence that Haraldr went any further south than the African coast, if indeed he did go as far as that. Aside from the fact that these stanzas embellish an anecdote, and that they might be suspected of having been made up precisely to do that, there are other suggestions that these verses are spurious. Both Hharð 15 and ÞjóðA IV,13 sound as if they could be derived from ÞjóðA III,2 (cp the latter’s borga á Serklandi with í Serkja garði in Hharð 15, and its í sléttri Sikileyju with á sléttu Serklandi in ÞjóðA IV,13). As required by the anecdote, there are also similarities between Hharð 15 and both Þfisk 3 and ÞjóðA IV,13, as noted above. Probably we can only rely on the ‘official’ stanza (ÞjóðA III,2) which, as we have seen, imagines the Serkland campaign as just before the Sicilian one. This does not, however, mean they need have been geographically (or even temporally) close: this part of Þjóðólfr’s poem is summary and retrospective, and scholars have not, to my mind, placed sufficient emphasis on the conjunction áðr that divides the two halves of the stanza (e.g. Shepard 1982–5, 236), as noted above. The stanza tells of two discrete campaigns. The one to Serkland can be identified by comparison with Byzantine sources that tell of wars in Asia Minor during the early years of Emperor Michael IV. It is most likely that ÞjóðA III,2 records Haraldr’s involvement in the ‘fighting against Arabic pirates off Asia Minor, and . . . wars in the Arab sector of Asia and Syria’ (Blöndal and Benedikz 1978, 63). The most plausible explanation for Serkland in the skaldic corpus, then, is that it accords with the second of Ruprecht’s suggested meanings, the late Viking Age designation of the Islamic and Arab areas beyond Russia.

Viking Destinations

107

However, it is not incompatible with Jarring’s explanation (discussed above), which seems etymologically the most plausible, though he does not consider the skaldic evidence. But Snorri, like other saga-writers, understood Serkland to include Africa, indeed he believed it was the Varangians’ name for Africa (Affríká, er Væringjar kalla Serkland, Hkr III, 74). He (or his predecessors) deduced this from the fact that ÞjóðA III,2 links Haraldr’s expedition to Serkland with that in Sicily, and because the next stanza of this poem calls an opponent of his Affríka j›furr (discussed above). Haraldr was certainly right to describe Serkland as fjarri fóstrlandi ‘far from [my] homeland’. Wherever it was exactly, it represents the furthest eastern and southern limits of viking activity at the very end of the Viking Age.

Scandinavia Both the runic and the skaldic corpus contain large numbers of Scandinavian place-names which have not, to my knowledge, been systematically studied in their own right as a body of contemporary evidence (though see von Friesen 1930 on farm-names in Uppland). To do so here would far exceed the scope of this chapter, but at least some of these places must be mentioned, since much of the ‘viking’ activity discussed in subsequent chapters occurred in Scandinavia. ‘Viking’ activity is occasionally defined as that which took place by Scandinavians outside Scandinavia (Jesch 1991, 8), but the vocabulary studied in this work does not easily admit of such a distinction. Sailing, raiding and trading in the Scandinavian context in the late tenth and the eleventh century are recorded in just the same terms as those activities elsewhere. The most important sea-battles of the period (and the place-names associated with them) are discussed in chapter 5. Here I consider some other significant places in Scandinavia, in roughly geographical order counterclockwise from south to northwest, concentrating on a few that are interesting or problematic in some way, or that occur in a range of sources (see fig. 3.11). As the skaldic corpus is largely concerned with making war (or other viking activities) in Scandinavia, it would be tedious to list all the Scandinavian place-names occurring in it here, though a few examples are discussed below, mainly some which are also recorded in the runic corpus. It may, however, be useful to list those runic inscriptions which refer to battles or other viking activity at particular places, often obscure and not always certainly in Scandinavia, some of which are also discussed below. Otherwise, discussions of the place-names and of possible events with which they can be related can be found in the commentaries on these inscriptions in the corpus editions. The phrase varð dauðr is modified by a prepositional phrase containing a Scandinavian place-name in D1, D3, D 117, D 216, D 259, Sö 16, ?Sö FV1948:289, Sm 52, U 375, U 539, Nä 15. In D 334 (an inscription from Skåne) the same phrase is modified by the adverb nur ‘(in the) north’. Similarly, the phrase varð drepinn is modified by a phrase containing a Scandinavian place-name in D 380, Ög 81, Sö

108

Ships and Men in the Late Viking Age

Image not available

3.11 Map showing places in Scandinavia mentioned in the text. All place-names are given in their modern forms. Chris Lewis, Cartographic Unit, School of Geography, University of Nottingham.

174, Sö 333, U 582. Other significant vocabulary modified by a phrase containing a Scandinavian place-name is falla (Ög 81, N 239); endas (Sö FV1959:266, U 518); deyja (U 180); drunkna (U 214); taka gjald (U 614). Hedeby On the southern frontier of Scandinavia, the town of Hedeby was an important trading centre and its region an important link between all of Scandinavia, the Baltic, and central and western Europe (Crumlin-Pedersen 1997b, 30–48, 203–5). Its origins are Saxon and Frisian, but eventually the Danes pushed their border further south and gained control of the town, while the Swedes also had a strong interest in it because of its commanding position on trade routes. Towards

Viking Destinations

109

the end of the Viking Age, the main activities of the town were gradually transferred to the north bank of the Schlei, where the present town of Schleswig is (the relative positions of the two are clearly illustrated in a photograph in Clarke and Ambrosiani 1991, 58). The end of Hedeby is symbolically marked by Wreck 1 found in the harbour there. Only part of the ship survives, and the charred strakes show that it was burned to the waterline. Since this happened between 990 and 1010, it is unlikely to have been a pagan ship-burial, such as that described in the beginning of Beowulf. The favoured interpretation is that the ship was used as a fire-ship in an attack on the town. It was set on fire when the wind was from the right direction, and allowed to drift into the harbour and set alight any of the wooden structures on the shore with which it came into contact (Crumlin-Pedersen 1997b, 94–5). Although we cannot trace this specific event in the sources, both the runic and the skaldic corpus refer to raids on Hedeby. Two of the five runic inscriptions that mention Hedeby are from Hedeby itself. D 1 is a rune stone discussed extensively elsewhere in this book, as it has a plethora of ‘viking’ vocabulary: stýrimaðr, drengr, félagi, heimþegi (see chs 5–6). The inscription was commissioned by one of King Sveinn’s retainers in memory of one of his colleagues ias uarþ tauþr þo trekiaR satu um haiþabu ‘who died when drengjar besieged Hedeby’ (see figs 3.12 and 3.13).39 D 3 also mentions the town, and here it is the king himself who commemorates one of his retainers ias uas farin uestr ion nu uarþ tauþr at hiþabu ‘who had travelled west but now died at Hedeby’, perhaps in the same siege, since runologically the inscriptions appear to be contemporary. The King Sveinn mentioned in these inscriptions has been identified either as Sveinn Forkbeard, with the siege of Hedeby his capture of the town in 982 (Moltke 1985, 200), or as Sveinn Úlfsson (better known to Danes and others nowadays as Svend Estridsøn), who repelled an attack by Haraldr harðráði in 1051, discussed below (DR, 8–9). The latest expert runological opinion would date the two Hedeby stones to the 980s, in other words to Sveinn Forkbeard’s time (Stoklund 1991, 293). Thus, it is not possible to link the inscriptions to the fire-ship attack by Wreck 1, which was constructed around 985 and used as a ship for between five and twenty-five years before being used in the attack on the town (Crumlin-Pedersen 1997b, 94). Hedeby is also mentioned in D 63, but since so little is left of the inscription, it is impossible to say whether there is any connection between this memorial from the other end of Denmark and the events recorded in D 1 and D 3, although the inscriptions do appear to be contemporary (Stoklund 1991, 292). The town is also mentioned in a Swedish runic inscription in which a mother commemorates her son who (t)o i haiþaby ‘died in Hedeby’ (U 1048). This inscription is however from a much later period (late eleventh or even early twelfth century), as indicated partly by the poor-quality ornament on the stone. Wessén argues that

39

The spelling haiþabu on this stone (see top centre of Face B) has given rise to the modern name Haithabu, used for the Wikingermuseum Haithabu, and regularly by German archaeologists of the Viking Age town (e.g. Jankuhn 1986).

110

Ships and Men in the Late Viking Age

Image not available

3.12 Face A of the Hedeby stone (D 1). Photo: Erik Moltke, National Museum of Denmark.

Viking Destinations

111

Image not available

3.13 Face B of the Hedeby stone (D 1). Photo: Erik Moltke, National Museum of Denmark.

112

Ships and Men in the Late Viking Age

the place-name refers to the town of Schleswig rather than Hedeby, since the stone must date from the time when the former had taken over from the latter as the centre of activity (SR IX, 311). It is interesting, at any rate, that the old name was still in use at this time.40 Finally, the name also occurs in Sö 16, now sadly fragmentary, but which stated that the deceased unR tauþr ‘died’ and then later in the inscription comes the name [iþaby] (known only from early drawings of the stone). As this stone should be considerably earlier than U 1048, but later than the Danish ones, it may well represent some attack on Hedeby in the early to mid-eleventh century, but the details are too vague to make any link with attacks recorded in skaldic verse. The earliest reference to Heiðabýr in the skaldic corpus is in Hfr II,5. As usual, Hallfreðr is maddeningly vague, but he notes that Óláfr Tryggvason killed warriors í Danm›rku fyr sunnan Heiðabý ‘in Denmark, south of Hedeby’, which at least indicates that this region was considered to be Danish territory at the time. Snorri tries to identify the battle in question as one which Óláfr fought with his brother-in-law, a Wendish king (Hkr I, 262–3), but this is probably just guesswork, or creative reconstruction (Fidjestøl 1982, 107–9), as the other sagas which cite this stanza do not provide any sort of historical context (Fsk, 142; Oddr, 247). However, the association with Wends is probably correct, as Hedeby was a frontier area where many different groups mixed and, occasionally, fought. Thus, as has already been noted, Magnús góði fought the Wends in the region, apparently both fyr sunnan Heiðabý ‘south of Hedeby’ (ÞjóðA I,6) and fyr norðan Heiðabœ ‘north of Hedeby’ (Þfagr 1), unless one of these poets is confused, but these were not attacks on the town itself. Adam notes that Magnús Heidibam appulit ‘landed in Hedeby’ and that his battle was in campestribus Heidibae ‘in the plain of Hedeby’ (AB, 320). An attack on the town may have been made by Haraldr harðráði, as described in an anonymous stanza (AnonXI Lv,7) attributed to menn Haralds ‘Haraldr’s men’ (Hkr III, 114–15): Brenndr vas upp með endum allr, en þat má kalla hraustligt bragð, es hugðak, Heiðabœr af reiði. Vón es, at vinnim Sveini, vask í nótt fyr óttu, gaus hór logi ór húsum, harm, á borgar armi. All of Hedeby was burned from end to end in anger, and that can be called a bold deed, in my opinion. It is to be expected that we will

40

Adam of Bremen uses both names, Heidiba and Sliaswich/Sliaswig, suggesting that the change of name was relatively recent (AB, 228, ?236, 270, 434) in the 1070s when he was writing.

Viking Destinations

113

cause Sveinn grief; in the night before dawn I was on the arm of the fortification, a high flame gushed from the houses.

The sagas that preserve this verse (Msk, 164; Fsk, 257–8; Hkr III, 114–15) associate it with a stanza, recalling apparently the same event, from Þorleikr fagri’s poem on Sveinn Úlfsson (Þfagr 6) in which the poet, with his Danish perspective, is naturally critical of the attack. Whoever does not know getr fregna . . . hvé heiptgjarn konungr hefr árnat til Heiðabœjar ‘may ask . . . how the enmity-eager king [Haraldr] has betaken himself to Hedeby’, also called þengils bœr ‘the king’s [Sveinn’s] town’. This action is described as þarflaust ‘unnecessary’, and the year in which it happened as ár þats ón of væri ‘a year one should be without’. The archaeological evidence suggests that activity in Hedeby was dying down from around the year 1000, though there are some finds in the southern part of the harbour from the first half of the eleventh century. Crumlin-Pedersen (1997b, 43) suggests ‘that the old derelict harbour at Hedeby was only used for laid-up ships’ and that Haraldr’s attack is likely to have been on Schleswig. However, the description in AnonXI Lv,7 of the burning town viewed from borgar armi ‘the arm of the fortification’ does sound more like Hedeby than Schleswig, at least until we know more about what form (if any) Schleswig’s fortifications took in this period. Either Hedeby was still worth attacking around 1050, or this anonymous stanza has been wrongly contextualised in our medieval sources. The latter is entirely possible. The medieval authors of the kings’ sagas, faced with an anonymous stanza in which the only identifiable points were the place-name Heiðabýr and the personal name Sveinn would naturally link this stanza to the better-provenanced stanza from Þorleikr fagri’s poem on Sveinn Úlfsson, which also mentions the place-name. It is not at all certain that this link is correct, and it may very well be that AnonXI Lv,7 celebrates some other attack in which Hedeby itself was burned, perhaps even during the reign of Sveinn Forkbeard, the only other Sveinn associated with the town. And, it might be noted, if we dissociate the two stanzas, it is not at all clear that even Þfagr 6 refers to a sacking of Hedeby rather than some other action of aggressive intent. The prose sources seem to represent various attempts to make sense of a sequence of stanzas in which the actual events are not entirely clear (some of these stanzas are discussed further in ch. 5). Denmark to Sweden Both Magnús and Haraldr attacked other parts of Denmark, especially the island of Fyn, as summed up of Magnús in Arn III,18: Enn rauð hringserks lituðr frón merki á Fjóni ‘Again [or, ‘further’] the colourer of the mail-coat reddened bright banners on Fyn’ (see also ÞjóðA IV,7). Similarly, Haraldr rauð frána egg á Fjóni ‘reddened the bright blade on Fyn’ (Arn VI,1; see also ÞjóðA IV,3, Valg 7), so that the Fjónbyggvar ‘inhabitants of Fyn’ diminished. Poets gloried in listing the Danish places that received visits from these two warlike kings: Þjólarnes near

114

Ships and Men in the Late Viking Age

Denmark’s largest river, the Gudenå (Grani 2); Þjóð ‘Thy(holm)’ (Stúfr 6); Falstr ‘Falster’ (Stúfr 5); Selund ‘Sjælland’ (Valg 6); Hróiskelda ‘Roskilde’ (Valg 8); Skáney ‘Skåne’ (Arn III,16, ÞjóðA I,24, ÞjóðA IV,4,8, Valg 6) and its capital Lund (ÞjóðA IV,24). Inhabitants who suffered their raids were: the Falstrbyggvar ‘inhabitants of Falster’ (Arn III,17); Selundbyggvar ‘inhabitants of Sjælland’ (ÞjóðA III,23); m›rg Selunds mær ‘many a girl from Sjælland’ (who had to run to Hringstaðir ‘Ringsted’, ÞjóðA I,17); indeed all the Eydanir ‘island-Danes’ (Arn VI,8), as well as the Skónungar ‘inhabitants of Skåne’ (Arn III,13, ÞjóðA IV,5). According to Adam, Haraldr omnia Danorum maritima ferro vastavit et igne ‘devastated all of coastal Denmark with sword and fire’ and fought Sveinn throughout both of their lives (AB, 342). Skåne is mentioned in one runic inscription, Sm 52, in which the two sons commemorating their father give quite a lot of information about his end: es uarþ tuþr o skonu (n) karþstokum auk furþu o finhiþi ‘who died in Skáney in Garðstangir and (they) brought him (home) to Finnheiðr’ (see fig. 3.14). These places can be identified as Gårdstånga, about 10 km northeast of Lund, and Finnveden, a district in Småland, also mentioned in Sm 35 and U 130 (SR IV, 160). Although there have been attempts to link this inscription to Knútr’s wars with Óláfr Haraldsson and the Swedish king Anund Jakob in 1025–6, the evidence is really too slight, though the date of the inscription would fit well with this period (SR IV, 161). The fact that the sons brought their father home to Småland is noteworthy, though what their reasons were for undertaking this journey with their father’s body (the distance is over 200 km as the crow flies) is impossible to say. The more usual pattern is represented in the Karlevi stone (Öl 1), which states that the commissioner placed the stone at u (at ey), meaning either ‘on the island’, or ‘on Öland’ (which means ‘island-land’). This is a memorial to someone who died and was buried far from home, as both the rune-forms and the text show that the commemorated, Sibbi, was Danish. A much-discussed set of inscriptions from Skåne commemorate an otherwise-unidentified battle at ubsalum, presumably the famous political and religious centre of Uppsala in central Sweden. Their runological dating is comparable to the Hedeby stones, at least approximately (Stoklund 1991, 292), so that they can be dated to the last two decades of the tenth century. Unfortunately, we know too little about Scandinavian history in this period to pinpoint an exact event in which men travelled from Skåne to take part in a battle in Uppland. Both D 279 and D 295 commemorate, respectively, Ásbj›rn and Áskell by stating that sá fló eigi at Upps›lum ‘he fled not at Uppsalir’, and the former adds an ua maþ an uabn a(f)þi ‘he struck while he had a weapon’. Both are partially in verse (Hübler 1996, 132–4) and contain significant vocabulary which is discussed further in chapter 6.

Viking Destinations

Image not available

3.14 The Forsheda stone (Sm 52). Photo: Judith Jesch.

115

116

Ships and Men in the Late Viking Age

Two more towns A place-name in Uppland which occurs in both the runic and the skaldic corpus is Sigtuna, reflecting its growing importance as a mercantile, political and religious centre in the eleventh century (summarised in Clarke and Ambrosiani 1991, 76–9; see also AB, 474). The name is also recorded, though usually in abbreviated form (STUNE, SIHT, SIN, ST, SITUN), on some of the coins minted in the town (Malmer et al. 1991, 11, 13, 17, 21). Some of these suggest that a singular form of the place-name (Sigtún) was in use, as possibly in Valg 5 below, while the runic inscription and the other skaldic stanzas use a plural form (dat. Sigtúnum), as in the modern name. Various references to the place-name are collected and discussed in Strid 1989, though he does not discuss the apparent alternation between singular and plural forms. Although the place-name appears in LP as either n.sg. Sigtún or f.pl. Sigtúnir, strictly speaking the recorded singular forms are compatible with either n. or f. Sigtún, while the plural forms are similarly compatible with either n. Sigtún or f. Sigtúnir, since only the acc.sg. and the dat.pl. forms (which do not help to distinguish n. from f.) are recorded in the runic and skaldic sources. U 395 is unfortunately fragmentary, so we cannot really know in what context the statement -(i)m hna firþi til sihtunum ‘who brought her to Sigtún(ir)’ belongs (see fig. 3.15). The suggestion that, in this inscription, ‘a man commemorates his wife’ (B. Sawyer 1994b, 175) is possible but unprovable. The town is mentioned in three skaldic stanzas, all in connection with the return of Norwegian kings from the east. Both Arn III,2 and ÞjóðA I,2 show the young king Magnús góði arriving by ship from Russia to Sigtuna. Arnórr states that brún veðr ‘sharp weather’ (i.e. a stiff wind) carried the young king at Sigtúnum ‘to Sigtún(ir)’, while Þjóðólfr notes that his men hlóðu húnskript í Sigtúnum ‘took down the sail in Sigtún(ir)’, as they approached the town (see ch. 4). Magnús had political reasons for returning to Norway via Sweden (see e.g. Jesch 1994b), and Haraldr also seems to have needed to make friends with the Swedes on his return from Russia (see ÞjóðA III,8). Valgarðr tells Haraldr sáttu Sigtún, þás sædríf létti ‘you saw Sigtún, when the sea-spray lessened’ (Valg 5). The town of Niðaróss ‘(at the) mouth of the river Nið’, modern Trondheim, was referred to by the appellative kaupangr ‘market(-place)’, as discussed above (ch. 2). Although the town-name as such does not appear in the skaldic corpus, the river-name does, in contexts which suggest the importance of the town on that river.41 Thus, we see Óláfr sailing his fleet out of the Nið (Sigv X,3) towards Denmark, while the arrival of Knútr into the Nið with his fleet (Þloft II,6) signals his conquest of Norway. The Nið was also the starting point for Haraldr harðráði’s expeditions (Arn VI,16; Msk, 280):

41

Despite the translation ‘Trondhjem’ in Skjd B I, 300, the name Þrándheimr in Þloft III,2 refers, not to the town today called Trondheim, but to the region around it (modern Trøndelag), as indicated by the phrase ráða byggðum later in the stanza (see also Lockertsen 1999).

Viking Destinations

Image not available

3.15 The Sigtuna stone (U 395). Photo: Judith Jesch.

117

118

Ships and Men in the Late Viking Age Vítt fór v›lsungs heiti. Varð marglofaðr harða sá’s skaut ór Nið nýtla norðan herskips borði. The name of the V›lsung [prince] travelled widely; he who competently launched the plank of the war-ship southwards [lit. ‘from the north’] out of the Nið was very highly praised.42

There is a particularly attractive description of Haraldr’s fleet sailing out of the Nið in a stanza from a set of verses on this subject by Þjóðólfr (ÞjóðA IV,19; Hkr III, 142): Slyngr laugardag l›ngu lið-Baldr af sér tjaldi, út þars ekkjur líta orms súð ór bœ prúðar. Vestr réð ór Nið næsta nýri skeið at stýra ungr, en árar drengja, allvaldr, í sjá falla. On Saturday the fleet-lord throws off the long tarpaulin, where splendid widows gaze on the planking of the dragon [ship] from the town. The young ruler steered the brand-new warship west out of the Nið, and the oars of the warriors fall into the sea.

The specific vocabulary of ships and sailing that is evident here is replicated in many other stanzas in the corpus, and will be the subject of the next chapter.

42

It is, of course, necessary to sail north (and then west, as in ÞjóðA IV,19, below) to get out of the Nið, but Haraldr’s overall direction was southwards to Denmark.

4 Ships and Sailing And heaved and heaved, still unrestingly heaved the black sea. MELVILLE

Our image of the Viking Age is dominated by the viking ship. Some of the most spectacular archaeological finds of the period are of whole or partial ships, whether from burial mounds (Oseberg, Gokstad) or dredged out of harbours (Skuldelev, Hedeby). The archaeological finds symbolise the importance of ships to the viking project – Scandinavian success in raiding, trading and settlement depended on their skill in building and sailing ships. Research into the viking ship has gone beyond the recovery, preservation and reconstruction of the found ships into the recreation of viking ships using new materials but often the old methods, copying the surviving originals. Such artefactual and practical research has made use of knowledge gained from the later Scandinavian boat-building tradition, but has not neglected philological evidence. However, this philological evidence has been used eclectically, so that saga descriptions and modern terms from Icelandic or the mainland Scandinavian languages have been given equal weight (Foote 1978, 61). The post-Viking Age material is so rich, the Viking Age evidence so meagre, the continuity of boat-building traditions so strong, that it has seemed (and probably is) justified to use the evidence in this unchronological way. But to my knowledge there has not been a comprehensive study of nautical terminology closely focused on the Viking Age linguistic evidence. Falk’s classic study (AnS) ranges more widely than that in its coverage, making extensive use of saga-evidence, limited use of skaldic evidence, and none at all of runic evidence. And since he wrote in 1912, there is a wealth of new archaeological evidence and experience to which this material can be linked. This chapter will provide an exhaustive account of the linguistic evidence from skaldic poetry and, to a lesser extent, from runic inscriptions, relating to the construction and use of ships in the late Viking Age. The aim is not to deny the importance of later linguistic evidence, or evidence from other language groups, or indeed the evidence of the ships themselves over the last thousand or more years. Rather, the aim is to present as fully and as clearly as possible the nautical terminology of the Viking Age as it survives in sources that are indisputably or

120

Ships and Men in the Late Viking Age

arguably from that period (as outlined in ch. 1), and to show what deductions about that terminology and its semantic range those sources enable us or, equally importantly, do not enable us, to draw.

Words for ‘ship’ skip The most common word for a waterborne vessel of a certain size, with cognates in all the Germanic languages, is skip (n., pl. skip). This is the unmarked term used in countless sources throughout the Viking Age and up to the present day. With kn›rr, it is the main term used in the runic corpus, and it is relatively common even in the skaldic corpus, where the exigencies of rhyme and metre, as well as the preference for a more colourful vocabulary, otherwise caused the poets to use a wide range of synonyms. There is not much to be gleaned from the occurrence of the word in six, or possibly seven runic inscriptions, most of which will be discussed in other contexts below.1 In three of these (D 68, D 335 and U 778) the man being commemorated is said to have owned a ship, in two (U 439 and U 778), he is said to have ‘steered’ his ship east with Ingvarr, and in Sö 164, which is in a kind of verse, the commemorated is given a heroic cast by the statement that he stuþ trikil(a) i stafn skibi ‘he stood like a drengr in the stem of the ship’.2 This rune stone also depicts a ship, with a cross as its mast (see fig. 4.1). Although a number of rune stones have ship designs, this is the only one to make reference to a ship in the text as well as in the iconography.3 In the skaldic corpus, the word skip occurs in thirty-five stanzas.4 The instances are spread over a range of poems, and occur in a variety of contexts, describing sea-battles, warships sailing out on raids, and sailing for other 1 2 3

4

D 68, D 335, Sö 164, U 439, U 778, U FV1946:258, and possibly D EM1985:265. Hübler (1996, 110) goes against the communis opinio that this is in verse, preferring to classify it as deliberately alliterative prose. On the uses of drengr/drengila see ch. 6. However, in the late-twelfth-century N 527 the text accompanying the ship-graffito includes the word kn›rr. Rune stones from the corpus that depict a ship are: D 77, D 119, D 220, D 258, D 271, D 328, Farsø (D EM1985:253), Ög 181, Ög 224, Ög MÖLM1960:230, Sö 122, Sö 154, Sö 158, Sö 164, Sö 351, Sö 352, Vg 51, U 370, U 979, U 1052, U 1161, Vs 17. Vg 119 is too early for the corpus. Stones depicting ships which are discussed in the runic editions but which do not now have runes and may never have had runes are: Hørdum (see Moltke 1985, 275), Långtora kyrka (described in SR VIII, 402), U 1001. The distribution of ship-pictures corresponds to the distribution of rune stones, with a slight overrepresentation in Södermanland, similar to its overrepresentation of stones commemorating people who had been abroad (Varenius 1992, 98) and of stones with poetic inscriptions. For the Gotland picture stones, see GB I, 62–74, and Varenius 1992, 51–85. Sigv II,3,6,7,8; Sigv III,10; Sigv VII,1; Sigv XIII,19,23; Ótt II,4,9; Ótt III,1; J›k 1; Hallv 3; Þloft II,3; Arn II,11; Arn V,7; Hharð 18; ÞjóðA I,2,4,15,22; ÞjóðA III,16; B›lv 8; Valg 5,10; Þfagr 4,5,8; Steinn I,2; Steinn II; Steinn III,5,14; AnonXI Lv,6,16; Mark I,16.

Ships and Sailing

Image not available

4.1

The Spånga stone (Sö 164). Photo: Judith Jesch.

121

122

Ships and Men in the Late Viking Age

reasons. Thus, skip is used in stanzas on the battle of Nesjar, between Óláfr Haraldsson and Sveinn jarl Hákonarson, of Sveinn’s ships (Sigv II,3,6,7,8). The same poet uses it of Óláfr’s ships in describing the battle in which Erlingr Skjalgsson is killed (Sigv VII,1) and in a lausavísa complaining that Knútr could get more men and bigger ships than Óláfr by paying for them (Sigv XIII, 19). Óttarr uses it of the raiding ships of both Óláfr (Ótt II,4,9) and Knútr (Ótt III,1). It is used of Knútr sailing to victory in England (Hallv 3). Arnórr uses it in the context of Magnús’ expedition to Wendland (Arn II,11) and of the battle at Deerness between Þorfinnr jarl and Karl Hundason (Arn V,7). Þjóðólfr Arnórsson uses it four times in his Magnússflokkr, describing the young Magnús’ journey from Russia, his leading of a fleet to attack Denmark, and twice in his sea-battle with Sveinn Úlfsson at Helganes (ÞjóðA I,2,4,15,22). It is used of the ships on both sides in various encounters between Sveinn and Haraldr harðráði (Þfagr 4,5,8; Steinn I,2; Steinn II), in describing a battle between Sigurðr ullstrengr and Steigar-Þórir (AnonXI Lv,16) and of Eiríkr’s expedition to Wendland (Mark I,16). It is used in stanzas describing or recalling both gentle and hard sailing at sea (B›lv 8; Valg 5,10). And it is used non-specifically in a variety of contexts by a number of poets.5 This shows that, as we would expect, poets did not choose the word skip to suit a particular context of meaning. Instead, they chose it for metrical reasons. Leaving aside those stanzas which are not in dróttkvætt,6 there are twenty-one occurrences of the monosyllabic skip (sg. or pl.), and nineteen of these occur in the fourth syllable of the six-syllable line, where the poet would need a short syllable if the word occupying that position was a noun, according to ‘Craigie’s Law’ (Craigie 1900, 343–6).7 All except two of these nineteen occur in even-numbered lines; the two exceptions (Sigv III,10 and Arn V,7) are both in the first line of the stanza with the word skip alliterating. Where skip- occurs in an inflected case, so that it is disyllabic, it always occupies positions 3–4 in an even-numbered line.8 Skip also appears as an element in compound words in both corpora. In two runic instances it is the first element, of skibuarþ ‘ship-watch’ in Sö FV1948:291 (discussed in ch. 2) and of skibliþ ‘ship-troop’ in U 348 (discussed in ch. 5). In the skaldic corpus, on the other hand, skip occurs only as the second element of a compound, with the first element specifying in more detail the kind of ship meant. The most common compound is herskip ‘warship’, with ten

5 6 7

8

Sigv III,10; Sigv XIII,23; J›k 1; Hharð 18; AnonXI Lv,6. Þloft II,3; Arn II,11; Hharð 18; Mark I,16. For discussions of the metrical implications of Craigie’s Law, see Kristján Árnason 1991, 120–23, 139–43, and Gade 1995, 29–36. The stanzas with skip in position 4 are Sigv II,3,6,7,8; Sigv III,10; Sigv VII,1; Sigv XIII,19,23; Ótt II,4; Ótt III,1; Hallv 3; Arn V,7; ÞjóðA I,2; B›lv 8; Valg 5; Þfagr 4,8; Steinn III,14; AnonXI Lv,16. The exceptions are ÞjóðA I,15 and Steinn II. Ótt II,9; J›k 1; ÞjóðA I,4,22; Valg 10; Þfagr 5; Steinn I,2; Steinn III,5; AnonXI Lv,6.

Ships and Sailing

123

examples (at least half of those by Arnórr).9 Other examples include two of langskip ‘longship’ (ÞKolb III,10 and Halli 1), and one each of kaupskip ‘merchant ship’ (Ótt II,13) and the nonce-formation (and pejorative) hlœgiskip ‘ridicule-ship, wretched craft’ (Sigv III,2). Again, it can be noted that, in the eleven dróttkvætt examples, the monosyllabic element -skip invariably occurs in position 4 of the line, while the disyllabic (because inflected) -skip- invariably occurs in positions 3–4. Compound words are often difficult to fit into the lines of skaldic poetry, but just as the poets found skip(-) useful in positions 3–4, they could also use compound words with this element to fill up their line, and even to bear the alliteration. The five occurrences of compounds with -skip(-) in odd-numbered lines all participate in the alliteration.10 It is often asserted that langskip is a translation of Latin navis longa and that it means ‘warship’ (AnS, 100–101) and today we often assume that ‘longships’ are ‘viking’ ships. There are descriptive references to the length of ships (see discussion below) and, of course, some ships were long, so that the adjective langr can be found modifying a variety of ship-words: skip (ÞjóðA I,4), lung (Hfr III,14, apparently a loan-word from Old Irish long, itself from Latin navis longa, AEW), borð as a pars pro toto for ‘ship’ (Steinn III,11), and the ship-kenning sæmeiðr ‘sea-tree’ (Sigv I,1). Fell (1981, 110) says of langr sæmeiðr that it ‘is a play on langskip, longship’, but it is hardly more than a straightforward adjective + noun (like the other examples), except that here the noun is a simple kenning. The very variety of these collocations shows that any kind of ship could be long and that ‘longship’ is not a technical term. It is salutary to note how infrequently the compound word langskip is used in Viking Age sources (as opposed to the sagas, for instance), even if both the surviving examples do place the word in a military context. Nor did the word have any especially Scandinavian associations: the Anglo-Saxon Chronicle refers to the langscipu that King Alfred had built as a defence against viking ships (ongen ða æscas, ASC 897A). skeið The most frequently-used word (more common than skip) for a ship in the skaldic corpus is skeið (f., pl. skeiðr or skeiðar),11 occurring in at least forty-nine stanzas.12 The word is almost always used of ships either on the way to or in the

9

10 11 12

Hfr II,1; Hár 1; Arn II,4,17; Arn III,1; Arn VI,16; Arn VII,2; ÞjóðA IV,22; Valg 11; Mark I,30. The two lines of Hfr II,1 containing this word are identical to two lines in Arn III,1, and have probably erroneously found their way from Arnórr’s stanza into a stanza later attributed to Hallfreðr (Whaley 1998, 183–4). Hfr II,1 (if this is not the same as Arn III,1, see previous note); Sigv III,2; Arn III,1; Arn VII,2; Valg 11. Both plurals are attested and confirmed by the metre, e.g. in Sigv III,9 and Sigv VII,2 (Whaley 1998, 210). Sindr 2; Tindr I,4,5,10; Hfr III,6; Edáð 7; Hókr 2,6,7,8; ÞKolb III,1,2,4,9,13; Ólhelg 9; Sigv I,3,5; Sigv II,7,9; Sigv III,9; Sigv VII,1,2,3; Ótt II,15; Ótt III,1; Arn II,4,7,10; Arn

124

Ships and Men in the Late Viking Age

thick of battle and clearly connotes a warship. There are only two examples of skeið used in a peaceful, or at least non-military, context (Sigv III,9; Arn II,4), while an anonymous stanza makes an explicit contrast between a sea-battle on a skeið and the life of the landlubber being served leeks and ale by a maiden (AnonXI Flokkr). What is less clear is whether skeið referred to a particular type (rather than function) of ship. Falk (AnS, 104) assumes it is synonymous with langskip, though that is not much help in this context, as we have just seen.13 Etymology is not much help either. Foote and Wilson (1974, 236–7) note that skeið means ‘either ‘‘that which cuts through the water’’ or ‘‘a piece of wood long and sword-shaped’’: either description would be appropriate’ (similarly AEW), while Falk (AnS, 105) would prefer to link it to skeið ‘race(-track)’ and understand it as a navis cursoria. However, the occasional collocation can give us some idea of the characteristics of skeiðar. They are described as langar ‘long’ (Tindr I,4) and súðlangar ‘with long strakes’ (ÞKolb III,1),14 but they need not all be the same length (m›rg misl›ng skeið ‘many a skeið of varying length’, ÞKolb III,9). They can be mævar ‘slender’ (ÞKolb III,4). As warships, they are (há)brynjaðar ‘armoured (at the oarports)’ (ÞjóðA IV,22; B›lv 2).15 They can be heavily laden, hlaðnar, with armed men (B›lv 8). The word skeið is borrowed into English (and used of English ships) during the Anglo-Saxon period. Thus, King Æthelred ordered ships to be made in all of England, with each district of 310 hides to provide a scegð (ASC 1008E)16 and the word appears (also in the compound sce[i]gðmann) in a number of other sources from the late tenth or early eleventh century (Harmer 1989, 266–7; Fell 1986, 311–13). Skeið even appears two, or just possibly three, times in the runic corpus. In the Tryggevælde inscription (D 230), it refers to the monumental ship-setting of which the rune stone is a part.17 In Sö 171, the commemorated was a skeiðar vísi

13

14 15

16 17

III,13; Arn V,6,12; Arn VI,3; ÞjóðA I,12,24; ÞjóðA III,8; ÞjóðA IV,9,18,19,22; B›lv 2,5,8; Valg 1,6,9,10; Halli 2; AnonXI Flokkr. In the case of Ólhelg 9, the word occurs in most of the manuscripts which have this stanza, but the manuscripts in one branch have a form of the stanza without this word (Fidjestøl 1982, 67–8). See also n. 18, below. Whaley (1998), who otherwise translates words fairly consistently, illustrates the problem by translating skeið variously as ‘galley’ (pp. 163, 231, 272), ‘longship’ (p. 210), and ‘ship’ (pp. 149, 157, 244). Also ÞjóðA IV,18: although strictly speaking here the l›ng súð collocates with dreki, this ship is also called a skeið in this stanza. The traditional interpretation of this word envisages armouring in the form of iron plates on the prow (AnS, 38; LP); for the suggested alternative, see under the discussion of hár, below. The Latin version of the Chronicle (F) glosses this word as unam magnam nauem quæ Anglice nominatur scegþ (ASC I, 138n.). An afterthought on Tryggevælde is Falk’s only acknowledgement that some of the vocabulary in which he is interested can be found in runic inscriptions (AnS, 118). Elsewhere (AnS, 104) he claims that the word skeið is not recorded in Old Swedish or Old Danish. Twenty-odd years later, he is better informed (Shetelig and Falk 1937, 373), but Kisbye (1982, 48) still does not know of the Swedish example(s).

Ships and Sailing

Image not available

4.2 The Ekeby stone (Ög 68). Photo: Antikvarisktopografiska arkivet, Riksantikvarieämbetet, Stockholm.

125

126

Ships and Men in the Late Viking Age

‘captain of a skeið’. The collocations suggest a warship, for ‘he fell in battle (han fial) in Holmgarðr’. It has also been claimed that Ög 68 contains the word skeið (e.g. SR II, 70). The runes read han x uas x uesteR x tauþeR x i x uereks x ?ai?i ‘he died in the west in Væring’s(?) ?ai?i’. To make the last word skaiþ would involve reading s twice (across a word divider), assuming the two worn runes are k and þ, and ignoring the final rune (see fig. 4.2, an old photograph taken before 1961 when the stone was still outdoors, and in which the worn runes have been painted as k and n). From my own examination of the inscription, the two worn runes look much more like original m and n, but what word this hypothetical maini might represent is another question (SamRun gives this sequence as (k)ai-i). Two other inscriptions which say that the commemorated died on a ship (Sö 49 and U 258) both use the expression á knerri, and á would seem to be the most appropriate preposition in this context, while í would be more likely to be used with a word meaning ‘troop, army, expeditionary force’ (see the discussion of lið in ch. 5). snekkja Eight skaldic stanzas (seven of them dating to the mid-eleventh century) use the word snekkja (f., pl. snekkjur).18 The etymology of this word is uncertain, as is its connection with the clearly related OE snacc (ASC 1052C,D; 1066D,E).19 It is usually assumed (e.g. Hkr I, 159) that the main distinction between a snekkja and a skeið was that the former was smaller, but such contrasts are based on the prose information of the sagas. Using this prose information, Falk (AnS, 102) defines a snekkja as a twenty-bencher (with a bench equivalent to a pair of oars, see below). This is too precise and is contradicted by the evidence of ÞjóðA I,2 that the snekkja in which Magnús returned from Russia was a þrítøgt skip ‘ship of thirty benches’. In three stanzas (ÞKolb III,2; ÞjóðA IV,22; B›lv 2) snekkja is used in a catalogue of ships, among knerrir and skeiðar, among herskip and skeiðar, and among skeiðr and beit, respectively. In these stanzas, the poets are trying to give an impression of a large fleet composed of various types of ships and use the different words, probably not very precisely, to create this impression. Nevertheless, the fact that skeið occurs in all of these catalogues with snekkja suggests that there was a recognised distinction between the two types, and this is supported by a stanza from Arnórr’s Þorfinnsdrápa (Arn V,6).

18

19

ÞKolb III,2; Arn V,6; Arn VI,4; Arn VII,2; ÞjóðA I,2; ÞjóðA IV,22,23; B›lv 2. It also occurs in Sindr 7, from a poem which probably just qualifies for the corpus if correctly attributed: it appears to be about the living Hákon Aðalsteinsfóstri and he died in 960/61. The mss of ÓTr have skeiðum, giving a not inconceivable rhyme (Kristján Árnason 1991, 99–100). Kuhn (1983, 263) also thinks the original had skeiðum. In a forthcoming University of Münster doctoral thesis, Katrin Thier suggests the Norse form represents an original borrowing from West Germanic, with snekkja a secondary formation. The Germanic root would be *snak- ‘a sharp protuberance, nose’ (AEW).

Ships and Sailing

127

Describing Þorfinnr’s brave attack, with five ships, on Karl Hundason, with eleven ships, the difference in numbers is presumably compounded by the difference in ship-sizes: while Karl has skeiðar, Þorfinnr has only snekkjur. dreki Another image we have from the Viking Age is of the dragon ship, or the ship with a carved dragon’s head on the prow. The word dreki (m., pl. drekar) ‘dragon’ is used of a ship in only seven stanzas.20 Many writers assume dreki was a technical or semi-technical term for the ‘biggest warships of all’ (Foote and Wilson 1974, 237; Simek 1982, 39) and that it was broader than a skeið (AnS, 104; Simek 1982, 35). Falk (AnS, 107) suggests that the term was originally the name of a particular ship, and compares the use of Ormr ‘worm, dragon’ as a ship’s name. Certainly the different etymology of dreki (compared to other ship words which are originally descriptive of their shape or actions) seems significant. While it is possible that the term was used in the late Viking Age of particularly large and ornate ships, such as those belonging to Haraldr harðráði, it is worth noting that the poets also call these same ships skeið (ÞjóðA IV,18,19; Valg 6,9,10). It is more likely that calling a large warship (with or without a dragonhead prow) a ‘dragon’ is a poetic conceit rather than a terminus technicus.21 Þjóðólfr Arnórsson uses both naðr ‘snake’ and ormr of Haraldr’s dreki (ÞjóðA III,13; ÞjóðA IV,18,19,21; see also Arn VI,2).22 Other poets make the comparison without using the word dreki: sævar naðr ‘sea-snake’ (Edáð 3); læbaugs eik ‘oak of the ring of poison [serpent]’ (Hár 2). Some poets, having chosen to call a ship a ‘dragon’, develop the metaphor by referring to different parts of the ship in terms of the beast’s anatomy: Haraldr’s ormr has a fax ‘mane’ (ÞjóðA IV,18), the prow of Haraldr’s ship is an orms munnr ‘dragon’s mouth’ and it has a golden red skolpt ‘forehead’ (Valg 10), similarly Magnús berfœttr’s dreki has golden hausar (‘heads’, lit. ‘skulls’, Gísl I,16) and braut hrygg ‘broke its back’ (Gísl I,15) when sailing in rough weather. Haraldr’s naðr is ramsyndr ‘strongly swimming’ (ÞjóðA III,13). This is entirely in keeping with the poetic habit of comparing ships to various kinds of animals, primarily in kennings. Gísl’s description (I,16) of Magnús berfœttr’s dreki, using the plural form hausar ‘skulls, heads’ of its stems, led Falk (AnS, 40) to speculate whether the ship had two dragonheads, one at each end, or even whether it had a tail at one end and two heads at the other, although this seems unlikely (Simek 1982, 19). The Bergen seal from 1299 (illustrated in Varenius 1992, 124) shows a ship with a dragonhead at either end, as do several of the ships on Viking Age rune 20 21 22

Sigv X,8; ÞjóðA III,12; ÞjóðA IV,18; Valg 10; AnonXI Lv,4; Gísl I,15,16. Extreme examples of the literary use of dragons (among other fabulous creatures) in describing Scandinavian ships occur in EE, 12, 20. Both Hallfreðr and Halldórr ókristni use naðr/Naðr for Óláfr Tryggvason’s ships, both called Ormr, but here it is a play on that name, rather than a synonym for dreki (see discussion of ships’ names, below).

128

Ships and Men in the Late Viking Age

stones.23 Some dragonhead prow-ornaments are known archaeologically, the most notable example being the fragmentary spiral ornament carved with a simple dragon’s head on either side from Oseberg (Brøgger et al. 1917–28, I, 334–8, III, 24–6; see fig. 4.3). The word dreki for a ship derives from this practice of placing carved dragonheads on ships, and is part of a group of words from the same semantic field used by poets of such ships, but there is no evidence that it was a technical term for any particular type of ship. kn›rr Arguably a different type of ship altogether was the kn›rr (m., pl. knerrir), a term used by archaeologists (often in the modern Scandinavian form knarr, e.g. Bill 1997, 190) to denote the broader type of ship that could both sail across the ocean and carry a heavy cargo, such as Wreck 1 from Skuldelev 1 or Wreck 3 from Hedeby. It is significant that kn›rr, like the generic skip, appears in runic inscriptions, and it appears just as frequently, in six inscriptions, all from Södermanland and Uppland.24 The collocations in these inscriptions indicate a range of possible meanings. Two of the inscriptions refer to the death of the commemorated on board a kn›rr. Sö 49 just says han uarþ tauþr o kniri ‘he died on a/the kn›rr’ without stating what he died of, it could have been in battle, or in a shipboard accident (Düwel 1987, 326). A violent death is indicated in U 258 where it is said of the commemorated that on trabu nurminr o kniri asbiarnaR ‘Norwegians killed him on Ásbj›rn’s kn›rr’. The mention of Norwegians might suggest a sea-battle between Norwegians and Swedes (SR VI, 327, 428), but it could equally have been a merchants’ brawl on a trading voyage (Düwel 1987, 326). Trading is the more likely context for the poetic Sö 198 h[n] uft siklt til simk(a)(l)(a) t(u)ru[m] knari ‘he often sailed to Semigallia in a splendid kn›rr’. Düwel (1987, 319) assumes kn›rr is a ‘Lastschiff’ and takes the naming of the destination, the repetition of the action and the use of the adjective dýrr ‘splendid’ (referring to the load rather than the ship) as further evidence of this inscription as a ‘Zeugnis zum Handel’. Although we might note that the adjective dýrr can be used of a ship or part of a ship in skaldic verse (ÞjóðA III,9 and B›lv 8), on the whole Düwel must in this instance be right. Two further inscriptions (U 654 and U 1016) refer to the deceased’s skills in steering a kn›rr. Düwel 23

24

Ships with dragonhead stems at both ends are illustrated on D 271 and D 328. The ships on D 258 and U 979 also have or had dragonheads, but not enough of the illustration is preserved to determine whether this was true of both ends. A tenth-century (non-runic) hogback from Lowther in Westmorland depicts a ship with ‘zoomorphic stem and stern posts’ (BACASSE II, 130). Sö 49, Sö 198, U 214, U 258, U 654, U 1016. It also occurs in the twelfth-century graffito N 527, outside our corpus. U 214 is included because SamRun dates it to the Viking period, c.1100, although Düwel (RGA XIII, 578) brackets it with N 527 as twelfth-century.

Ships and Sailing

129

Image not available

4.3 The dragonhead terminal of the Oseberg ship (reconstruction). Photo: Eirik Irgens Johnsen, Universitetets Oldsaksamling, Oslo.

interprets these two inscriptions differently (1987, 319–20, 325–7). U 1016 is ‘zweifellos’ evidence of trade, as the deceased travelled regularly to the harbours of ‘Greece’, or the Byzantine Empire, but died peacefully at home, while U 654 commemorates someone who died in the east with Ingvarr, and therefore the word kn›rr must refer to a warship. However, it might just be that the rune-carver needed an alliterating word (is kuni ual knari stura) in this poetic, or at least stylised, inscription. The last runic example (U 214) is slightly later than the others and unusual in being in rhyming, rather than alliterative verse. It commemorates a man who drowned in Holms haf (either the sea around Bornholm, or the Gulf of Finland) when his kn›rr sank. Jansson (SR VI, 328) and Düwel (1987, 325) have diametrically opposed views on whether the word here means ‘warship’ or ‘trading ship’, but I cannot see that there is any evidence

130

Ships and Men in the Late Viking Age

either way. On the whole, it is safest to assume that, in Viking Age Sweden at least, a kn›rr was just another word for skip, a type of ship that could be used on either a raiding or a trading voyage, and indeed it is likely that many Swedish voyages in the Baltic and to the east at this time had elements of both. The skaldic evidence is a little more one-sided. The standard example for the meaning ‘warship’ for kn›rr is Hkv 7, which refers to Haraldr hárfagri’s ships as knerrir. This text (if correctly attributed) is too early for the corpus that forms the basis of this study, its authenticity can be doubted, and its preservation and textual status are complicated (Jón Helgason 1968, 10–15; Fidjestøl 1997, 70–71, 79–81). In the corpus, the word kn›rr occurs only occasionally and at first glance can apparently mean either an ocean-going cargo ship or a warship. One example of the latter has already been noted (ÞKolb III,2). This catalogue of Eiríkr jarl Hákonarson’s fleet uses three different words (snekkjur, knerrir and skeiðar) to suggest the variety of that fleet, and the poet could have used an unexpected ship-type to suggest the scope of Eiríkr’s operations in his attack on England. The only other unequivocal skaldic example of a kn›rr in a battle context is in a lausavísa (Vígf II) looking back on the battle of Hj›rungavágr. This stanza has a rather odd use of the word drengr to refer to the speaker’s opponents (see ch. 6), and its authenticity may also be doubtful on other grounds, so it is unwise to place any emphasis on its evidence. In the other skaldic stanzas in which kn›rr occurs the context is not a military one, and it is likely that the poet’s choice of a word with a non-military connotation was deliberate. Sigvatr uses kn›rr in two lausavísur (Sigv XIII,4,26), both referring to his personal affairs and memories and implying a context in Icelandic trading (discussed in ch. 2). Óttarr, describing Óláfr Haraldsson’s return to Norway from raiding in England in two ships, calls those ships knerrir and explicitly equates them with kaupskip ‘merchant ships’ (Ótt II,13), perhaps to obviate the impression that the king-to-be was raiding his own country.25 When incorporating this stanza into Heimskringla, Snorri obviously felt the strangeness of the successful returning warrior coming home in a cargo-ship, and explains that Óláfr switched ships (Hkr II, 35): Óláfr konungr lét þar eptir vera langskipin, en bjó þaðan kn›rru tvá ‘King Óláfr left the longships behind there, and made ready two knerrir from there’. Steinn Herdísarson (III,14) lists some of the gifts King Óláfr kyrri gave to his followers, including hábrynjuð skip ok steinda kn›rru ‘ships armoured at the oarports and painted knerrir’. There is both repetition and variation in these stanzas, and it is not certain that the poet intended these phrases to be synonymous. The emphasis is on both the range and the splendour of the king’s gifts, and he is likely to have given away different kinds of ships (compare the gifts of different kinds of armour, brynjur ok hjalma

25

That Óláfr returned in two merchant ships (cum duabus onerariis navibus) is also known to the twelfth-century synoptic historians, Theodericus monachus and the anonymous author of Historia Norwegiae (McDougall and Foote 1998, 76).

Ships and Sailing

131

‘mail-coats and helmets’, in Steinn III,16). Similarly, King Eiríkr gave his followers sverð ok kn›rru ‘swords and knerrir’ (Mark I,7). Kn›rr also occurs in kennings (Edáð 3; Sigv XIV,2; Bersi I,1) as both base-word and determinant, meaning ‘ship’, but without any further illumination of what kind of ship it was thought to be. Arnórr uses the related word knarri (m., pl. knarrar, a weak variant of the strong-declension noun kn›rr) in a quarter-stanza telling of his own journeys (Arn II,2).26 This word is a hapax legomenon, but presumably has the same meaning as kn›rr, since in the same stanza Arnórr calls his journey a kaupf›r ‘trading voyage’. It is usually assumed that this stanza belongs to Arnórr’s Magnússdrápa, with the poet recalling his own journeys before embarking on praise of King Magnús, and the oddity of this personal beginning is remarked on both in the prose anecdote which describes Arnórr’s recitation (Msk, 116) and in the Third Grammatical Treatise commentary on another fragmentary stanza from this introduction (TGT, 57). If the attribution of the two quarter-stanzas is correct, then it is clear that, in using knarri, Arnórr deliberately chose a word for his own journeys that would not by then normally be appropriate in a praise poem for a war-leader. It is impossible to discuss the meaning(s) of kn›rr without also considering its uses as a loan-word in other languages. Most notably, the two occurrences in The Battle of Brunanburh, as the simplex cnear in line 35 and in the compound nægledcnearr in line 53, used of the ships of the Norsemen fighting at Brunanburh, provide a securely-dated attestation earlier than any in the corpus (i.e. soon after 937). Although the battle took place on land, the function of these ships was to transport the Norse warriors to the battle, so at the very least, kn›rr could be used of a troop-carrier, if not an actual warship. In a recent article on the semantics of kn›rr, Sayers interprets cnear(r) as an ocean-going vessel and contends that ‘[f]or the most part only ocean-going vessels would have called on Ireland, and the requirement of seaworthiness would have favored the higher freeboard and broader hulls of the merchantman type over the sleeker fighting ships, sailed to best advantage in the coastal waters of Scandinavia’ (1996, 284–5). However, as he himself notes, the large warship from Skuldelev (Wreck 2), surely the quintessential ‘sleek fighting ship’, is now shown by dendrochronology to have been built in Ireland, and must have sailed back to Scandinavia. In any case, the journey from Ireland to Scandinavia (or vice-versa) involves only short open-sea passages. The word was also borrowed into Middle Irish, Old French and possibly even Old High German (Sayers 1996, 284–8), and at least some of the instances in these languages refer to Norse warships (e.g. CCS, 40, 42), at any rate not one of them occurs in a context of either mercantile activities or deep-sea crossings. Sayers’ article is primarily

26

The Third Grammatical Treatise explains knarri as an example of paragóge, the deliberate lengthening of a word by a syllable to make it fit the metre (TGT, 63).

132

Ships and Men in the Late Viking Age

intended to propose an etymology of kn›rr as a ‘ship whose hull was prominently marked by nail heads’ (1996, 285), and to argue that an understanding of the ‘underlying semantics’ (the etymology?) of the word is reflected in the borrowings into ‘languages of the western European seaboard’ (p. 288). Although Sayers is quite right to criticise the traditional etymology, he does not reconsider the Norse evidence and I do not find his alternative particularly convincing, but the article is useful in setting out the range and type of occurrences of the word in the borrowing languages. Loan words do not have to have the same semantic range as in their original language, but it seems that the borrowing languages all understood kn›rr as appropriate to a ship used by invading warriors, and we have to consider whether this was also true in Old Norse. The East Norse runic inscriptions have already shown that the word was used in contexts that could include both raiding and trading, and combined with the loan-word evidence, this suggests that, in early usage, the meaning was not restricted to ‘cargo-ship’ or ‘ocean-going vessel’. This semantic narrowing can however be traced in the West Norse skaldic texts from the eleventh century, where there is a clear avoidance of the word for royal warships, and is complete in the later prose sources on which most scholars base their understanding of the word. Recently, Ole Crumlin-Pedersen (1997a, 189–90) has observed that the relative beam of ships ‘changed drastically in the early phase of the process of adapting ships to carry sail, in order to provide sufficient stability at the initial stage’. This is why early warships, such as Tune and Gokstad, have a low length-to-breadth ratio, while the eleventh-century warships ‘had returned to the pre-Viking proportions’ and were long and slender ‘to maximize the effect of the rowers, now in combination with sail propulsion’. In the early period (in the ninth and most of the tenth century), then, there would not have been much difference between warships and cargo-ships in terms of their shape (although they would have been distinguished in other ways). The term kn›rr can be seen to reflect this stage, and must in its subsequent development be linked to the overall shape of the vessel. It survived as a term for cargo-ships, which continued to have low length-to-breadth ratios, even after the development of the long and narrow warships in the late tenth and eleventh centuries, the skeiðar of skaldic poetry. This explanation would even allow for the correctness of the term knerrir in Hkv. Foote urged a belief in the poet’s accuracy on this point, but the work of archaeologists since he wrote over twenty years ago suggests that we are now closer to knowing ‘what might suitably be called a kn›rr about the year 900’ (1978, 64). Oak and pine Occasionally the skalds used words which reflect the types of wood widely used in Viking Age ship construction, especially oak and pine, though this vocabulary is not especially rich. Eik ‘oak’ (f., pl. eikr) is used in three stanzas as a simple heiti for ‘ship’ (Sigv

Ships and Sailing

133

VII,1; Valg 6; AnonXII B,3),27 and once in the kenning læbaugs eik ‘oak of the ring of poison [serpent]’ (Hár 2) with the same meaning. Where a poet needed two syllables, he could use eiki (n., pl. eiki), originally a collective noun and therefore appropriate for a ship which would be made of more than one tree, and this occurs three times (Arn II,9,16; ÞjóðA I,4). Three syllables are provided by the compound eikikj›lr ‘oaken keel’ (ÞjóðA III,8). These are all warships, and the collocations suggest stately and well-made vessels. Three of these ships are also called skeið in the same stanza (Sigv VII,1; ÞjóðA III,8; Valg 6), and three of them are described using complimentary adjectives: g›fugt ‘splendid’ (Arn II,9), farligt ‘attractive, looking as if it would run well’ (Arn II,16) and vandlangt ‘long-masted’ (ÞjóðA I,4).28 Analysis of the wood used in building the five Skuldelev wrecks shows that a variety of timbers was used, but oak was the most common, and it is characteristic that wreck 2, the largest of these, built in Dublin, was made entirely of oak (Olsen and Crumlin-Pedersen 1990, 130–31), as are the Oseberg and Gokstad ships, not so much warships as royal yachts (Brøgger and Shetelig 1950, 139, 174, 196). Shipbuilding materials would depend at least in part on local conditions, and pine would have been more commonly used in western and northern Norway (Brøgger and Shetelig 1950, 255–6). With their pine-planked hulls, Skuldelev 1 and 6 are now known both to have been built in Norway (NAVIS, see also Wagner 1986 on the types of wood used in shipbuilding). However, fura ‘pine’ (f., pl. furur), occurs only twice (Arn II,10; ÞjóðA IV,11),29 in different contexts: in the former the ship is being buffeted by a storm, in the latter it is lying at anchor. Arnórr, at any rate, chose the word because he needed to alliterate on the letter f, for this is the same ship that was called eiki in the previous stanza (Arn II,9; see Malmros 1986, 102). The noun þella (f., pl. þellur), also meaning ‘pine’, occurs often enough in skaldic verse (in kennings meaning ‘woman’), but never of a ship, while the collective variant þelli (n., pl. þelli; compare eik and eiki) is once used of a ship, but outside the corpus (LP). ޛll (f., pl. þellir) ‘young pine’ is similarly used mainly in woman-kennings, but once of an oar (ÞjóðA IV,21, see below). Another tree-type occurs as a base-word in the kenning sævar hlynr ‘maple of the sea’ (Ólhelg 7). As a lausavísa attributed to a king and preserved only in Flateyjarbók, this stanza is less secure than most in the corpus, and in any case

27

28

29

In Hharð 9, I follow Hkr III, 109, in reading líneik as a woman-kenning, rather than interpreting eik as the object of the verb halda, as in Skjd B I, 330. However, if this stanza belongs with ÞjóðA IV,11 (Perkins 1982–5, 194), it might be that the poet is using two different (and contradictory) words for ‘ship’, eik and fura, in the same stanza, rather as he uses a number of words for ‘anchor’ (see below). Although this is understood as a collective noun (and translated ‘skibene’) in Skjd B I, 333, it is more likely to be a singular reference to the king’s ship Visundr than to the whole fleet of seventy ships mentioned earlier in the stanza. There appears also to have been a genuine variant with a long vowel, fúra (Whaley 1998, 163). On ÞjóðA IV,11, see n. 27, above.

134

Ships and Men in the Late Viking Age

hlyn ‘maple’ depends on Sveinbjörn Egilsson’s reading of the manuscript form hlunn (Skjd A I, 221). It could of course be argued that the archaeological evidence of the use of maple in shipbuilding (Wagner 1986, 135; Crumlin-Pedersen 1997b, 185–6) confirms the correctness of the emendation. But even so the example is not a good one, for, in a kenning, the base-word stands for a category (here, ‘tree’, compare sæmeiðr ‘sea-tree’, Sigv I,1) and does not necessarily inform us about the material from which the ship is made. Other references to the type of wood in the corpus are doubtful (see also the discussion of askr ‘ash’ below). In ÞKolb III,2 m›rg vas lind fyr landi is usually translated ‘many a lime-wood [spear/shield] defended the land’ and there was probably a similar construction in Tindr I,9 (NN, 436). Since Þórðr’s stanza is otherwise about ships, it is possible that lind ought to be translated ‘ship’ there, but even then it is more likely the poet was in need of a word that provided him with both alliteration and half-rhyme in the line, than that he intended to suggest the construction material. Lime-wood was not especially suitable for shipbuilding, although we do find it used in Skuldelev 1 (Wagner 1986, 135). Occasionally, a poet refers to a ship by the simple synecdoche viðr ‘wood’ (m., pl. viðir),30 telling us that ships were made of wood, but then we knew that already. It seems that the skaldic ideal was a ship made of oak, and that kings, at any rate, were mostly able to make that ideal a reality. Miscellaneous words Other words used by skalds to denote ships are even less informative. Most common of these miscellaneous terms is flaust (n., pl. flaust), which is etymologically related to the verb fljóta ‘to float’. It occurs ten times referring to a ship or ships and, with one exception (Arn II,15), all of these instances either alliterate (Ótt II,14; Arn II,11; Gísl I,8) or rhyme (Ótt II,4; ÞjóðA I,4; ÞjóðA III,9; Mark I,5), and two do both (Steinn III,5; Mark I,24). The word was particularly useful for poets seeking a rhyme for either austr ‘in/to the east’ or austan ‘from the east’. It also rhymes with aust- in the stanzas where it occurs as part of a kenning, for poetry (Eskál III,3) or for the sea (AnonX III,C,3; Þfagr 8; AnonXI Knútr). The fact that most viking journeys celebrated in skaldic verse went either east or west may be the explanation for the relative frequency of flaust, rather than Simek’s theory that it was ‘ein poetisches Modewort’ (1982, 105). In Gísl I,8 it is used of the ships of merchants, while the same alliterative phrase, flaust fagrbúin ‘beautifully-prepared ships’, is used of warships in HHuI,31. In Mark I,25, the poet describes how King Eiríkr built five stone churches in Denmark and, using the common metaphor of church as ship, how the rest of the tíða flaust ‘ships of services [churches]’ in his kingdom were borði merkt ‘marked by planks’, i.e. made of timber. 30

Sigv X,7; Þloft II,4; Arn III,19; Þfagr 11; AnonXII B,3. In the latter, Finnur Jónsson translates viðr as ‘plankerne’ (Skjd B I, 592), but it is clearly a pars pro toto.

Ships and Sailing

135

Another colourless, but useful, synonym for ‘ship’ is far ‘(means of) travel’ (n., pl. f›r), which as a short, monosyllabic noun was suitable for position 4 in the dróttkvætt line (Sigv III,2; ÞjóðA III,34), or when alliteration on the very common initial sound f was needed (Ótt II,14, where it alliterates with flaust). It also occurs (Arn VI,4) in the compound farskostr ‘vehicle’ (m., pl. farskostar). The closely related ferja (f., pl. ferjur), which later certainly means ‘ferry’, occurs only once in the skaldic corpus, in a sea-kenning (Glúmr I,2) which does not reveal a specific meaning. An interesting pair of cognates are beit (n., pl. beit) and bátr (m., pl. bátar). Beit, a Norse word of uncertain etymology (AEW), is used twice of Haraldr harðráði’s ships (B›lv 2; Valg 5), and once of Magnús berfœttr’s ships (AnonXII B,3).31 B›lverkr, describing Haraldr’s journey to Constantinople, calls the same ships snekkja and skeið, so they must be large warships, while at least one of Magnús’ ships has a seventy-foot mast (discussed below), so was not insignificant. Bátr, on the other hand, appears to be a loan-word, possibly from Old English (AEW, cf. modern English ‘boat’), and is used for much lower-class vessels. Sigvatr’s journey to Sweden for King Óláfr took place partly in a dodgy bátr which he also calls hlœgiskip ‘ridicule-ship, wretched craft’, so the word is not likely to connote a fine, stately craft. Similarly, in ÞjóðA IV,17, a probably spurious and certainly joky fragment, the poet questions someone whose foot is sticking out from under an upturned bátr whether he is having sex (according to the prose context, he is actually seasick). We would hardly expect to find such a word in a formal poem of praise. The craft that Sigvatr called bátr and hlœgiskip, he also called a karfi (Sigv III,2), the only occurrence of this word in the corpus. As Foote (1978, 59) has pointed out, archaeologists are over-confident when they assume they know what kind of a vessel a karfi was, and Sigvatr’s boat is unlikely to have resembled the Oseberg ship.32 It is also worth noting words which occur in other sources and might therefore be expected to occur in skaldic poetry, but which do not, or rather which occur only in kennings. Although ash was used in building ships (Wagner 1986, 132), askr ‘ship (made of ash-wood)’ (m., pl. askar) occurs only once, in the kenning Ullar asks›gn ‘crew of the askr of Ullr [shield→warriors]’ (Eskál III,2). As Falk (AnS, 87) notes, the cognates of this word are used in English and continental sources of viking ships. King Alfred’s plan to build long ships ongen ða æscas ‘against the ash-ships’ of the Danes in 897 has already been mentioned, while two decades later, the Chronicle refers to viking invaders as æscmenn (ASC 921A; the date is actually 917, see Bately 1986, 66–8) and in The Battle of Maldon, 69, the invading Scandinavian army of 991 is called æschere, possibly

31 32

See also HHuI,23. Although it has been suggested this word appears in Irish as carbh (CCS, xii), it is not at all clear that the loan was necessarily from ON.

136

Ships and Men in the Late Viking Age

from their ships, though perhaps rather from some other kind of equipment (spears or shields) they had made of ash (Fell 1986, 313–14). Calling a ship a kjóll (m., pl. kjólar) is common in Eddic poetry, the word may have been borrowed into Middle Irish (ciúil), and the Old English cognate ceol means ‘large ship’. But it is not used of ships in the skaldic corpus: it occurs once in a shield-kenning (Ullar kjóll, Eyv III,9; compare Ullar askr, above) and once in the kenning hæls hleypikjólar ‘running-ships of the heel [shoes]’ (ÞjóðA III,19). Summary Skalds used many different words to refer to ships. Certainly, the variety of this vocabulary partly reflects poetic pressures: the skald’s need for words that alliterate, rhyme or have the right number of syllables. Some words are generic (e.g. skip) and are therefore uninformative. Others are used poetically rather than descriptively (e.g. dreki). Words which describe the materials from which ships were made (e.g. eik) correspond with the factual evidence, but are not necessarily used precisely. Yet it is not fair to argue, as Malmros (1986, 107) has done, that the ships described in skaldic poetry ‘er overalt de samme’. The Skuldelev finds have shown that ships varied in size, shape and quality, and the skaldic evidence confirms that distinctions between different types of ships were important, and could be made even by the skalds (see especially the discussions of skeið, snekkja and kn›rr, above).33

Names of ships From the Mary Rose to the QEII, we are accustomed to ships named after women, as a glance at any fishing-harbour will also show. There is no evidence for this cultural stereotype in the Viking Age. Where ships’ names are preserved in contemporary sources, they are mostly grammatically masculine, and often named after animals. The list of contemporary attestations is quite short, though many more names occur in prose sources, especially those relating to the twelfth and thirteenth centuries (discussed in AnS, 32–3; see also Simek 1982, 14–27). Most famous of all is Óláfr Tryggvason’s Ormr inn langi ‘The Long Serpent’, named in six stanzas from two poems (Hfr III,13,18; Hókr 3,4,5,8). Halldórr ókristni was particularly fond of using ship-names. In his poem on Óláfr’s opponent at Sv›lðr, Eiríkr jarl Hákonarson, Halldórr always uses both the name and the epithet of Óláfr’s ship. Eiríkr’s own ship is called Barði in the skaldic sources (presumably a name derived from the part of the prow called barð, see below):

33

Even the Bayeux Tapestry shows William’s invasion fleet as consisting of ships of varying sizes (Wilson 1985, pls 36–7, 40–44, p. 227).

Ships and Sailing

137

Halldórr names it twice in stanzas in which he also names Ormr (Hókr 3,5),34 while Tindr (I,10) names it in a stanza on the battle of Hj›rungavágr. The names of two more ships belonging to Óláfr are recorded. One was called Trana ‘Crane’ (the manuscripts differ on whether this ship was called Trani m. or Trana f.), while another was also called Ormr, and known as inn skammi ‘the short’ to differentiate it from ‘The Long Serpent’. The former is named directly and the latter indirectly by Hallfreðr (III,16), describing the devastation after the battle of Sv›lðr from the point of view of a man who sá Tr›nu ok báða Naðra fljóta auða ‘saw the ‘‘Crane’’ and both ‘‘Snakes’’ floating empty’. The synonymic (?nick-)name Naðr also occurs in Hfr III,13 and Hókr 8, and Halldórr once even refers to Ormr inn langi by the name of the mythological serpent Fáfnir (Hókr 3). These variations gave the poets more latitude in their rhyming (thus Naðr twice rhymes with glaðr, and once half-rhymes and alliterates with niðr) but, as Hókr 8 shows Eiríkr taking over the defeated king’s ship, it may also be that he changed its name to suggest its change of ownership. Sigvatr once mentions a ship called Karlh›fði ‘Man-headed’, owned by Óláfr Haraldsson and used in the battle of Nesjar (Sigv II,4). From this, Snorri concocts a story that it had this name from the king’s head on the prow, carved by the king himself (Hkr II, 59), a kind of Viking Age equivalent to the female figureheads on later ships. However, Paasche’s suggestion (1914, 13) that this name recalls Óláfr’s royal ideal and model, Charlemagne, just as the naming of his son Magnús did, seems more likely. Óláfr’s better-known ship was called Visundr ‘Bison’. While ships could also be captured or inherited, Sigvatr tells us that Óláfr had this one made, in imitation of his namesake and predecessor’s Ormr (in this stanza called by the snake-kenning lyngs fiskr ‘fish of the heather’, Sigv XII,3).35 An anonymous stanza (AnonXI Lv,4) also refers to Visundr as Óláfr’s ship, and there are three references to it in the ownership of his son and successor, Magnús góði (Arn II,9,16; ÞjóðA I,4). The prose sources (e.g. Hkr II, 267) tell us that this ship had a bison-head on the prow, but the skaldic sources are not so informative, although Arnórr attempts to match the animal ship-name with an animal-based ship-kenning, elgjar œðiveðrs ‘elks of the storm’ (Arn II,16).

The ship and its parts Archaeological investigations over the last hundred years or so have given us a very clear picture of the construction of ships in the Viking Age. Quite a large number of ships and parts of ships are now known from different archaeological contexts, but the most important examples are those from ship-burials and those 34

35

This clear parallelism between the two ships suggests that Bjarni Aðalbjarnarson (Hkr I, 361) is over-cautious in doubting whether Barði is a proper noun, rather than a common noun derived from barð. On the metaphorical development of this stanza, see Hallberg 1978, 51.

138

Ships and Men in the Late Viking Age

recovered from harbours (there is a useful summary of ‘The ship as an archaeological object’ and ‘The methods employed in ship archaeology’ in Crumlin-Pedersen 1997b, 12–18). Of the ship-burials, Gokstad and Oseberg in Norway contained relatively well-preserved ships and boats (summarised in Brøgger and Shetelig 1950),36 while the ship from Ladby in Denmark left only an impression of itself (though it has now been artfully reconstructed from that impression, see Sørensen et al. 1998). The ships deliberately scuttled at Skuldelev in the Roskilde Fjord in Denmark have usefully provided a range of types of vessels from the late Viking Age (summarised in Olsen and Crumlin-Pedersen 1990), while similar finds have been made in the harbour of Hedeby (Crumlin-Pedersen 1997b, 81–104). Such finds continue to be made, most recently the discovery of no less than nine ships from the Viking Age and Middle Ages found literally at the door of the Viking Ship Museum in Roskilde in 1996–7 (Bill et al., 1998). Many of these ships, from both graves and harbours, have been reconstructed, some several times, and subjected to sailing trials, with the reconstruction process extensively documented (Andersen et al. 1997, see also Crumlin-Pedersen and Vinner 1986). It is not my intention to repeat any of the easily-available archaeological evidence, even in summary fashion. Nor do I attempt to consider the whole range of pictorial representations of ships from the Viking Age, though these are undoubtedly also contemporary sources and relatively numerous. Instead, I attempt to survey the complete range of words for the ship and its parts found in the skaldic corpus (the runic evidence is negligible in this section) and to relate this vocabulary to the archaeological and art-historical evidence where it has seemed illuminating to do so.37 Often, however, this link is hardly worth making: if a skaldic stanza mentions planks, it is hardly necessary to turn to archaeological evidence to confirm that ships were indeed made of planks. It may seem equally banal to collect such words when they are not much more informative in their skaldic contexts. Nevertheless, I feel the need to redress the omissions of Falk and others, and to give as complete a record of nautical vocabulary in the skaldic corpus as possible, and am convinced that the overall picture will be illuminating. It is only when the full range of nautical vocabulary has been considered that we can pronounce on what aspects of ships and their equipment were considered important or memorable in the Viking Age itself. The skaldic evidence will not provide a complete key to

36

37

It is important to remember that the Oseberg ship (built before the construction of the grave chamber in 834) and the Gokstad and Tune ships (both built before c.900) are all earlier than the skaldic and runic evidence under discussion here. I am aware of the inconsistency of insisting on strict chronological limits to these two textual corpora while admitting earlier (and occasionally later) evidence from the archaeological material. However, I try to do it with an awareness of the dates of the latter. On the dates of the Norwegian ships, established recently by dendrochronology, see the summary in Bonde and Christensen 1993. There is a useful glossary of modern English terms relating to the construction of ships in McGrail 1987, xviii–xx.

Ships and Sailing

139

Viking Age shipbuilding, for there are many words and concepts that never appear, and those that do are often used imprecisely. But it provides a very useful adjunct to other evidence. The hull Much of the nautical vocabulary in the skaldic corpus concerns the viking ships’ characteristic clinker-built hulls, including their keels, their planking and especially their distinctive stems, the same shape at either end. The keel (kj›lr m., pl. kilir) was the foundation of a clinker-built hull and laying it (attached to the lower part of the stems) the first stage in constructing a ship (Olsen and Crumlin-Pedersen 1967, 108). As a basic part of a ship, the word was almost bound to be used by poets to stand for the whole. Sigvatr is particularly fond of this synecdoche to describe fast-moving ships (Sigv III,9; Sigv X,4,8; see also HHuI,28), and the use of eikikj›lr ‘oaken keel’ for a ship (ÞjóðA III,8) has already been noted. More specific reference to the keel itself occurs in a late lausavísa (Eldj 1), in a description of hard sailing (þótt kj›l kosti ‘though the keel is tested’). In kennings, the word can similarly appear either as a pars pro toto (kj›lslóðir ‘ship-paths [sea]’, ÞKolb III,10) or in the primary meaning (kjalar vagn ‘wagon of the keel [ship]’, ÞjóðA I,12, in the kenning v›rðr kjalar vagna ‘guardian of the wagons of the keel [ships]’, meaning ‘captain’ and referring to Magnús góði),38 though in these the distinction is not especially clear. The sides of the hull were built up from the keel by lashing or nailing the planks together, and the words the skalds used for the hull reflect the construction process. The collective noun for this planking is súð (f., pl. súðir), etymologically related to sýja ‘sew’ and referring to the oldest method of joining planks by lashing them together (Foote and Wilson 1974, 243). CrumlinPedersen (1981, 284–5) points out that lashing frames was an outmoded technology in the Viking Age, but is found on the Norwegian royal ships precisely because ‘distinctive features of a glorious past [were] preserved as a manifestation of royalty’. The word can also stand for the whole ship (Hharð 4,16; ÞjóðA I,4; B›lv 4),39 or refer to the (usually long) sides of ships (Þloft II,6; ÞjóðA IV,18; Steinn III,5), similarly in the adjective súðlangr ‘long-sided’ describing a ship or ships (ÞKolb III,1; Hallv 1).40 In Hharð 16, the word súð stands for the

38

39

40

Some Hkr manuscripts have an alternative reading v›rðr Kjalars regna ‘guardian of the rainstorms of Óðinn [battles]’ meaning ‘warrior’, but this is less likely, see further on the collocations of v›rðr in ch. 2. In ÞjóðA I,4 it would be possible to read the plural súðir gnauðuðu as referring to the creaking movement of individual planks, rather than the movement of ships. However, the latter seems more likely as, in this carefully constructed stanza, the poet makes three references to ships in the plural (skip, flaust and súðir) before homing in on Magnús’s ship Visundr in the last two lines. This adjectival usage contradicts Falk’s view that súð is used ‘nur als pars pro toto für Schiff’ (AnS, 32).

140

Ships and Men in the Late Viking Age

whole ship but the poet hints at the construction method by modifying it with the adjective feldr ‘bevelled, scarfed’ (the techniques of overlapping planks are described in detail in Vadstrup 1997b, 96–103, see also the scarfed strakes illustrated in Crumlin-Pedersen 1997b, 226–7). This joining of strakes is called sk›r (f., pl. skarar) and a ship can be sk›rum hvélðan ‘concave because clinker-built’ (Arn II,4). I cannot find any evidence in the skaldic corpus that there was a regular distinction in the Viking Age between the two joining methods termed skarsúð and fellisúð, which Falk (AnS, 47) explains using later sources. The thinnest planks were the most supple, and so a ship can be called þunn sk›r ‘thin, narrow sk›r’ (ÞjóðA III,9). The planks in the hull of a warship were both narrow and thin (mæ borð ‘slender planks’ Ótt I,5). Olsen and Crumlin-Pedersen note that in the Skuldelev ships they ranged from 20–50cm in width (1990, 131), while wreck 2, the big warship, was built of strakes only 2–2.5cm thick (1990, 111, see also Brøgger et al. 1917–28, I, 294; Brøgger and Shetelig 1950, 170). The nailing of planks together is alluded to in ÞjóðA IV,21 where a ship (naðr ‘snake’) is described as neglðr með jarni ‘nailed with iron’. The very visible row of nails along each strake was called saumf›r, used in Arn V,21 as a synecdoche for the whole ship, as it is modified by the adjective kløkkr ‘pliant’. Although Whaley 1998, 261 sees ‘no justification for [the] assumption . . . that saumf›r . . . is a pars-pro-toto expression for ‘‘ship’’ ’, in a Viking ship it is the wood that makes the ship pliant, rather than the nails. The method of clenching round-shafted nails inside the ship over ‘roves’ is described and clearly illustrated in Crumlin-Pedersen 1991, 70–71. The word for an individual plank is the common Germanic term borð (n., pl. borð). This appears in the skaldic corpus in the collective sense, referring to the planking of the hull (singular in Arn VI,16; AnonXI Lv,16; plural in Arn VI,2; BjH 2), and in such phrases as innan borðs ‘on board [lit. ‘within the shell’]’ (Hfr III,14; Ótt II,13), fyr(ir) borð ‘overboard’ (Tindr I,10; Sigv II,7) and útan borðs ‘overboard’ (Steinn III,9). In this collective sense and meaning ‘side of a ship’ it appears in the compound hléborð ‘lee-side’ (Arn III,6 [in a ship-kenning]; ÞjóðA III,8). Large ships can be described as borðmikill ‘high-sided’ (Hókr 3), or as a borðviðr breiðr ‘broad plank-wood’ (Mark I,5). As with all words to do with the hull, borð can also appear as a pars pro toto for ‘ship’ (Sigv II,1; Arn II,11; Hharð 16; Steinn III,11).41 The word also appears as a determinant in ship-kennings: borðmarr ‘plank-horse’ (Eskál III,19), borðraukn ‘plankdraught-animal’ (Þfagr 3) and borðvigg ‘plank-steed’ (Steinn III,5).42 Borð occasionally appears in the plural, referring to the individual planks that make up the hull: in Ótt I,5 they are mjór ‘slender’ and break in the storm, in BjH 2 they are awash with blood, in Arn II,2 they are described as stinnr ‘firm, stiff’. In the

41

42

In reproducing Arn II,11, Bjarni Aðalbjarnarson (Hkr III, 39) prints borð, a variant found in a number of manuscripts, though Kringla (or rather AM 63 fol.) has b›rð (pl. of barð, a part of the stem, see below), a term that would also work as a pars pro toto. N.b. this kenning is an emendation from borðveg (Skjd A I, 410).

Ships and Sailing

141

latter example, the ship is a knarri (discussed above), and ships designed primarily for the transport of cargo were intended to be solid, rather than supple and resilient like the warships which could be beached (Olsen and Crumlin-Pedersen 1967, 109). The evidence in the skaldic corpus for specific names for individual strakes (or rows of planks) is limited, and their interpretation often dependent on later sources (AnS, 52–4). The word aurborð ‘gravel-plank’ apparently refers to the plank that rests on the ground when a ship is beached (AnS, 28, 52; see also discussion of hlunnr, below). This occurs twice in kennings (Eskál III,26; Ólhelg 7), which do not reveal the specific meaning, but as the compound is transparent, Falk’s interpretation may well be right. The keel strakes of Skuldelev 2 were ‘very worn on the outside from scraping against sand and gravel when the ship was pulled ashore’ (Olsen and Crumlin-Pedersen 1967, 117). In Modern Icelandic the fifth strake from the keel has in recent times been called a hrefna, and Falk would link this to ON hrefni (n., pl. hrefni), but as the latter occurs only once in the skaldic corpus, in a ship-kenning (ÞKolb III,3), it is not possible to confirm what specific meaning it had in the Viking Age. A word for the ‘brim’ of any vessel, not just a ship, is barmr (m., pl. barmar), which we find in a ship-kenning barms vigg ‘horse of the barmr’ (ÞKolb III,4) and in the descriptive adjective barmfagr ‘with a beautiful brim’ used of ships in B›lv 2, perhaps a reference to the decorated shields of these warships.43 In both these cases the word is chosen to rhyme and is unlikely to be intended very specifically. A more convincingly nautical term is þr›mr (m., pl. þremir), a word also meaning ‘edge, rim’, and hence ‘sheer-strake’, which appears several times (Hfr III,14; ÞjóðA I,4; B›lv 5). In the first two of these, the word is modified by an adjective indicating its pliability and suppleness, þíðr ‘pliable’ and sveigðr ‘curved’. The interpretation of þíðr as ‘pliable’ is from LP (see also Foote 1978, 62), but the basic meaning is ‘thawed’, which could be just as appropriate in this context of warm blood splashing on the þr›mr.44 A skjaldrim (f., pl. skjaldrimar) ‘shield-rail’ could be mounted on the uppermost strake, to hold the warriors’ shields (Arn V,21; B›lv 5). Falk is uncertain whether skjaldrim is the name of the topmost strake (AnS, 53) or of a special rail mounted on the topmost strake to hold the shields (AnS, 55). The small warship Skuldelev 5 had a narrow rail mounted outside the top edge of the uppermost strake, which has been interpreted as a shield-ledge (Olsen and CrumlinPedersen 1967, 137), as did the Gokstad ship (Brøgger and Shetelig 1950, 151). Also mounted on the uppermost strake was a hlýða (f., pl. hlýður) ‘washboard, washstrake’, to provide extra protection from the waves (Halli 1; Mark I,5,16), a detail that is known from visual representations of ships (Christensen

43 44

Although it is tempting to see the later meaning of barmr ‘bosom’ in this adjective, there is no evidence of this meaning before the Reformation (CV, ONP). Kock (NN, 3215) seems more concerned to attack Finnur Jónsson than to establish the meaning of the word from its range of contexts.

142

Ships and Men in the Late Viking Age

Image not available

4.4 Ship graffito from Christchurch Place, Dublin, showing washboards. Photo: National Museum of Ireland.

Ships and Sailing

143

1988, 18; see fig. 4.4), though the archaeological evidence is less than clear. The top strakes (with oar-ports) in some Viking Age ship finds from western Norway are so thin and light compared to the other strakes that they have been described as ‘hvad en maa kalde et let skvætbord’ (‘what could be called a light washboard’, Brøgger et al. 1917–28, I, 350–52), and the excavators of the Scar boat burial in Orkney thought they could identify some of the nails as derived from the fixing of a ‘washrail’ to the boat (Owen and Dalland 1999, 203). The Bayeux Tapestry consistently shows the English ships as having ‘the central portion of the gunwale plank missing’ (Wilson 1985, 226) and these ships may have had some similar arrangement. These washboards may have been optional extras: in both Halli 1 and Mark I,16 the hlýður are said to skjalfa ‘tremble’, perhaps because they are not very firmly attached, while in Mark I,5 King Eiríkr studdi borðvið breiðan hlýðu í veðr óðu ‘supported the broad plank-wood with a hlýða in foul weather’, again suggesting that it could be added when required.45 This word does not appear in any prose sources, or in OGNS or AEW, and its etymology is obscure. Several dictionaries (CV, ÍO, IEW) link it with an OE word variously spelled and meaning ‘shelter, protection, warmth’ (BT, s.v. hleowþ), but the link is a distant one, for there is no evidence that the OE word was ever used in a nautical connection. Nor is ‘washboard’ the only suggested meaning. CV suggests ‘ship’s cabin’, while ÍO tentatively suggests that hlýða might be a dissimilated variant of hlýra ‘skipskinnungur’ (see discussion of hlýr, below). However, Mark I,5 seems to indicate a part of the ship that can be easily attached and removed, and neither of these suggestions seems especially plausible. Archaeologists and others writing about viking ships frequently use the term meginhúfr for the transition strake, an especially strong, narrow, strake at the waterline (Brøgger et al. 1917–28, I, 296–7; Vadstrup 1997b, 103; Sayers 1998, 57). Falk mentions this word (AnS, 53) but does not adduce any sources for it, and both CV and OGNS give only one instance of it from a medieval prose source.46 Meginhúfr is not recorded in the skaldic corpus, although the simplex húfr (m., pl. húfar) is. Etymologically, húfr refers to something concave (AEW), while in more recent Icelandic it has developed the meaning of specific strakes, either the third and fourth, or the fourth and fifth (AnS, 53). The skaldic instances seem closer to the etymological meaning, as they mainly refer to the shape of the hull, or just to the hull as a pars pro toto (Mberf 1). Thus Arnórr refers to the húfr of a skeið as sk›rum hvélðan (Arn II,4, discussed above), but also has Magnús slicing the sea héltum húfi ‘with icy hull’ (Arn III,2). The shape is also suggested in Eldj 1 in which a ship is described as breiðhúfaðr ‘broad-hulled, 45

46

Wreck 6 from Skuldelev had an additional seventh strake added ‘some time’ after the vessel was built, to increase ‘the height of the ship’s sides’ (Olsen and Crumlin-Pedersen 1967, 152), but this seems to be a more permanent addition. The source is the later Bjarkeyjarréttr first promulgated in 1276 (NGL II, 283; see KLNM I, 659). ONP has no further citations in its archive of slips. Nevertheless, the word now appears to have entered the English, Danish and German languages (Crumlin-Pedersen 1997b, 207)!

144

Ships and Men in the Late Viking Age

beamy’. When a bay is called húfi róinn ‘hull-rowed’ (ÞjóðA III,34), the poet is presumably saying no more than that ships have passed over it. The description of a ship as húfjafn ‘smooth-hulled’ in Hfr III,18 depends on an emendation, so must remain uncertain as an example, but fits in with other references to the careful jointing of hulls.47 Valg 10 has the waves breaking und húfi ‘under the hull’. This may suggest a particular part of the hull at or near the waterline. Similarly, the example already quoted from Arnórr (Arn II,4) actually says that Magnús ‘stepped onto’ the húfr. This makes good sense if the word is being used as a pars pro toto for ‘ship’, but even better sense if it was intended to be specific, suggesting a particular strake at the level of the deck. Exactly which strake was intended depends on the number of them in a vessel, which naturally depended on its size. But the cross-beams which support the decking of loose planks are generally at the waterline level (see e.g. Brøgger and Shetelig 1950, 145, 159; Evans 1985, 71), and it may be possible to see here an early instance of this specialisation of meaning, and the basis for the more specific term meginhúfr. A more specific meaning may also be behind the ship-kenning húfs fákr ‘horse of the húfr’ (Hfr III,9), though the context cannot reveal what part is meant. There are few references to the colour of ships’ hulls in the skaldic corpus. There may be a suggestion of the tarring of ships in kolsvartir viðir ‘coal-black timbers’ (Þloft II,4), where viðr is a synecdoche for ship (though it may also be possible to take it as meaning ‘masts’, e.g. AnS, 56). The same synecdoche, but with a different colour, appears in a rauðr viðr ‘red timber’ (Þfagr 11), presumably painted. In ÞKolb III,9 the b›ru dýr ‘animals of the billow’ are blár ‘blue’ (usually a colour closer to black than to bright blue), likewise in Sigv II,2 both sides at the battle of Nesjar sail bl› borð ‘ ‘‘blue’’ ships’, while the planks of Erlingr’s ship are blakkr ‘dark’ (BjH 2).48 More colourful, perhaps, were the ships of Sveinn Úlfsson steini fagrdrifin ‘beautifully washed with paint’ (Þfagr 3; the same ships are called fagr ‘beautiful’ in Þfagr 5) and the ships which King Óláfr kyrri gave to his followers, including steinda kn›rru ‘painted knerrir’ (Steinn III,14). Wreck 1 from Hedeby, the large warship, had traces of paintwork on the inside of its planks, which currently show as a ‘strong yellow deposit’ (Crumlin-Pedersen 1997b, 86). The stems The viking ship is characterised by its symmetrical stems. It is misleading to talk of a ‘prow’ and a ‘stern’ on viking ships, since the two ends had the same shape, though the location of the rudder at the rear of the ship, the set of the sail, and other details, would indicate at a glance the difference between the fore-stem and the after-stem. Their elegant stems are a notable characteristic of viking ships, 47

48

Kock’s suggested explanation of the manuscript forms (NN, 1958) as hverjafn ‘fully smooth’ would still give the same overall meaning, though without specific reference to the hull. In HHuI,50 the ships are blásv›rt ‘blue-black’.

Ships and Sailing

145

justly celebrated in skaldic poetry. Dragonhead stems have already been discussed above, in connection with the ship-word dreki. The generic word for either end of a ship was stafn (m., pl. stafnar). This word is found in one of the two runic references to a part of a ship (Sö 164), which states of the commemorated that he stuþ trikil(a) i stafn skibi ‘stood in a drengr-like fashion in the stem of the ship’. In theory, this could either mean that he stood in the rear of the ship, steering and/or captaining it, or in the front, in the fighting position (as it is understood in SR III, 126). In the skaldic examples, the word stafn is often used in the extended meaning ‘ship’ (Ótt III,2; Arn VI,2; ÞjóðA IV,1; Halli 2; Steinn III,5) but the more specific meaning must be understood when it collocates or compounds with a word meaning ‘ship’: (langskipa stafnar Halli 1; skeiðar stafnar Sigv II,9 and AnonXI Flokkr; snekkju stafnar ÞjóðA IV,22). Similarly, in ship-kennings (Edáð 2,5), the word must stand for a part of a ship, although in the sea-kenning stafnklif ‘stafn-cliff’ (Þloft II,5), ‘ship’ works as well as ‘stem’. The compound framstafn ‘fore-stem’ occurs twice, both times, obviously, with the specific rather than the generic meaning. In Hallv 2 it collocates with a ship-kenning, while in ÞjóðA I,12 the poet refers to the leader (Magnús) fighting í fagran framstafn varar hrafni ‘in the beautiful fore-stem of the raven of the harbour [ship]’. King Haraldr’s stallari ‘marshal’ (a high-ranking official of the king) apparently declared in 1066 that the stafnrúm of the king’s ship was not the place for his ilk (Ulfr). The stafnrúm (n., pl. stafnrúm), is the space in the stem and, in this context, must refer, like framstafn in the previous example, to the occupation of this space by fighting-men. The word stafn can be modified by adjectives indicating the appearance of the stem. In ÞKolb III,3 the stems are hóvir ‘high’. The skafnir stafnar ‘shaved stems’ of ÞjóðA IV,22 refer to how they were produced.49 In describing the building of the Roar Ege, a copy of Skuldelev 3, Vadstrup (1997b, 86) notes that the final stage of carving the stepped stem-posts was time-consuming because all the surfaces (of which there are many on such a complex, sculptured object) had to be ‘plane-smooth’, something that was achieved by using planes, axes and scrapers (see fig. 4.5). Christensen (1985, 207) stresses the importance of the correct cutting of the stem and postulates both that ‘the stem was a prestigious object and that good stem-cutters were highly regarded by others’. The skalds use a number of words for the different parts of the ends (though mainly the front end) of a ship, and the precise significance of these can be difficult to determine from the skaldic evidence alone. Some of the words are metaphorical, developing the image of the ship as an animal. Where a ship bears a prow-ornament in the shape of or representing the head of a dragon or other creature, this can simply be called h›fuð ‘head’ (Edáð 5; Valg 11; Gísl I,16)50 or,

49 50

Other wooden parts said to be skafinn are the rudder (Ótt II,20) and the oars (Þór; see also HHuI,49). See also HHuI,24, langh›fðuð skip ‘long-headed ships’.

Image not available

4.5 The stepped stem-post of the Skuldelev 3 ship. Photo: Viking Ship Museum, Roskilde.

Ships and Sailing

147

more extravagantly, skolptr ‘the front part of an animal’s head’ (Valg 10), hauss ‘skull’ (Gísl I,16; Mark I,6) or even gríma ‘mask’ (Arn II,4). As we have already seen, these are often described as ‘golden’ or ‘gilded’: there is an image of the reflection in the water of the gold af roðnum hausi ‘from the reddened skull’ in Mark I,6. The skalds can home in on the detail of this ornament by focusing on its mouth, usually with reference to gilding: a ship is gollmunnuð ‘gold-mouthed’ (Þfagr 3) or gínn golli búnum munni ‘gapes with a gold-adorned mouth’ (Þfagr 11), or has gj›lnar roðnar m›lnu gulli ‘gills reddened with ground gold’ (Sigv XII,3), an extravagant image chosen because the dragon-ship is denoted by the snake-kenning lyngs fiskr ‘fish of the heather’. The head sits on the ‘neck’, which is also ‘adorned’: búinn svíri (Gísl I,16). A slightly different image is when the ship’s bows are denoted by hlýr (n., often in the pl. hlýr) ‘cheeks’, imagining the ship as a face seen head-on (Sigv I,5; Arn II,4; Þham I,2).51 The nautical meaning of hlýr is common enough for it to be used as a pars pro toto (Hharð 16; Mark I,5) and as the determinant in ship-kennings (Hfr II,1; Hfr III,18) and a sea-kenning (J›k 1). A number of more technical, and less metaphorical, words clearly refer to some of the different parts that make up each end of the ship. Brandr (m., pl. brandar), a word which may or may not be etymologically related to the homonym meaning ‘sword’ (AEW), is interpreted by Falk (AnS, 44) as the triangular, possibly decorated, piece of wood such as that reaching up from below the end of the top strake to the stem of the Oseberg ship. But if this word had such a specific meaning in the Viking Age, there is no evidence of this in the skaldic corpus, where it is used mainly as a pars pro toto for ‘ship’ (Tindr I,6; Arn III,6; Halli 1),52 or collocating with a ship-word, herskip (ÞjóðA IV,22) or snekkja (B›lv 8), suggesting a part of the ship but not which part. Foote and Wilson (1974, 234) more cautiously translate it as ‘curved gunwale fore and aft’. In B›lv 2 it is modified by a colour-adjective: svartr snekkju brandr ‘black brandr of the snekkja’, suggesting that brandar, whatever they were, would be tarred. The idea that brandar were a particularly special part of the ship may be supported by the reference to brandar glæsti gulli ‘brandar adorned with gold’ (Halli 1), although we have already seen that in the context of this stanza the word is used as a pars pro toto for ‘ship’. The word may have been borrowed into Old English, appearing as the first element of the compound brondstæfne used of a ship in the famous description of sailing in Andreas, 504. This compound would be tautologous, unless the first element referred to some special kind of part (Sandahl 1951, 39). Brooks (1961, 80) notes the tautology and rejects the ON origin of brond-, though the question must remain open, as the meaning of the ON word is not as clearly fixed as he thought. In Arn II,7, Magnús’ opponent, Sveinn 51

52

Kinnungr (m., pl. kinnungar), with the same meanings, does not appear in the skaldic corpus. There is no trace in the skaldic corpus of ON bógr, thought to be the origin of English ‘bow’, suggesting that Sandahl (1951, 37) is right to query this etymology. Tindr I,6 is a doubtful example, as the stanza has to be very heavily emended to make sense. Kock interprets brandr here as ‘sword’ (NN, 433).

148

Ships and Men in the Late Viking Age

Alfífuson, is called a skildir skeiðar brands, interpreted by Whaley (1998, 157–8) as ‘shield-provider of the ship’s prow’, while Finnur Jónsson reads skyldir skeiðarbrands and interprets it more or less as ‘captain, steersman’ (‘en som lader skibet bevæge sig’, LP s.v. skyldir). In either case, brandr is redundant. The captain causes the whole ship to move, not just the prow, while it is unlikely that shields were positioned just on the steeply-rising stem of a ship (though they may have remained there, when shields positioned elsewhere on the gunwale were removed for rowing). Falk (AnS, 55) gives some examples from prose sources where shields were hung only on the parts of the gunwales rising towards the stems. The poem from which this stanza is taken is in the hrynhent metre, which required eight syllables per line, rather than the usual six, and Arnórr is clearly engaged in a little padding (with the added benefit of providing a rhyme for landi in the same line). A word that has been linked with the brandar is the mysterious tingl (n., pl. tingl), which appears in the ship-kenning tingls marr ‘horse of the tingl’ (Sindr 2) and the shield-kenning tungl tingla tangar ‘moon of the tong of the tingl’ (Hókr 3). Falk’s explanation of the latter (AnS, 43–4) is that the ‘tongs’ are the brandar, which enclose the tingl, the triangular piece in the fore-stem where the two sides of the ship join, and which, in the Oseberg ship at least, is richly decorated with carving (see fig. 4.6). Unfortunately, the skaldic evidence is too meagre to confirm or deny this interpretation, although it is a bit odd that tingl is plural in this kenning if Falk’s suggestion is correct. But since we cannot know what a tingl really was, it is a good enough word for the object to which it has been attached.53 Another word which must mean some part of the stem, but is often used for the whole thing is barð (n., pl. b›rð). This word occurs in one runic inscription, Sö 65, where it is said of the commemorated that he austarla arþi barþi ‘in the east ploughed with his barð’. In Sigv I,3, Bálagarðssíða (a place) lá brimskíðum at barði ‘lay before the barð of the sea-skis [ships]’, and barð is used elsewhere in this context of arriving and landing (B›lv 2,4), or in more general descriptions of launching and sailing like the semi-poetic runic inscription (Arn II,11 [if the reading is b›rð rather than borð, see Hkr III, 39]; ÞjóðA IV,23; Mark I,16).54 In some of these it could equally well be a pars pro toto for ship. Slightly more specific is Sigv V,1, in which the poet describes his own arrival in Rouen, where létk b›rð fest ‘I had the b›rð fastened’, suggesting a part of the ship to which a mooring rope could be attached (if he is not using the word as a pars pro toto

53

54

Although the etymology of this word is uncertain, the cognates listed by Falk (AnS, 44) tend to support his theory. Also, in Hkv 7, not in the corpus, the tinglar are said to be grafinn ‘carved’. In Mark I,16, the vowel in b›rð is confirmed by the rhyme with g›rðum, and the word collocates with the adjective hélug, just as in Arn II,11. This, together with a Wendish context, suggest that Markús was influenced by Arnórr, and indicate b›rð rather than borð as the correct reading in Arn II,11.

Ships and Sailing

Image not available

4.6 The tingl of the Oseberg ship? Photo: Universitetets Oldsaksamling, Oslo.

149

150

Ships and Men in the Late Viking Age

again). But the evidence does not tell us any more than that barð referred to a part or whole of a ship’s stem, though it does seem to have been particularly associated with the fore-stem. The word stál (n., pl. stól) is interpreted by Falk (AnS, 36) on an etymological basis as referring to the highest, narrowest, part of the stem proper, which supported the dragonhead or other decoration, and hence as synonymous with kylfa (discussed below). Other definitions (e.g. LP, OGNS) are more cautious and see it as a part of, or even synonymous with the fore-stem in general. However, the three skaldic instances of this word, as far as they go, support a specific interpretation of ‘stem-post’, the piece (or pieces) of wood rising from the keel into which all the strakes are gathered, and which makes the hull end in a sharp profile. In BjH 3, Óláfr’s sea-journey to Russia is expressed in a metaphor of cutting: allvaldr réð rísta haf austr stáli ‘the ruler cut the sea with a stál in the east’. This metaphor works best with a part of the ship that is in contact with the water most of the time, and the main part of the stem-post rising up from the keel is most appropriate. Arnórr refers to the stól of Magnús’ ships as stirð ‘stiff’ (Arn II,10), suggesting a firm and sturdy part. A more difficult example is in B›lv 5, describing Haraldr’s journeys in the east: byrr lá á breiddu stáli ‘the tailwind was on the broadened/spread stál’. A byrr was any kind of favourable wind, but since that cannot be a headwind it must in this stanza be a tailwind on the after-stem. Kock had trouble with breiddr and suggested emending to bræddr ‘tarred’ (NN, 3090). However, if the ship had a stepped stem-post like those of the Skuldelev ships, especially wreck 3 (the carving of its replica is described in Vadstrup 1997b, 85–6), the process of hollowing out the piece of wood could be described as ‘broadening’ it. In any case, the hull of any viking ship widens quite sharply from the stem-post. This leaves kylfa (f., pl. kylfur) as the term for the highest and narrowest part of the stem, as in its only skaldic instance, in a stanza by Sigvatr on the battle of Nesjar. In an attack on Sveinn’s skeið, Óláfr orders his men to skeina harðliga svartar kylfur ‘cut strongly into the black kylfur’, and they proceed to strike across the skeiðar stafna ‘the stems of the skeið’ (Sigv II,9). The highest part of the stem was in the way and, being made of relatively slender pieces of wood, could easily be cut down to make for easier access to the warriors on the enemy ship. Such objects are not known archaeologically from the Viking Age. Later evidence comes from models from medieval Bergen, which include some ‘loose stem-tops’ (Christiansen 1985, 158–60) and graffiti, from Bergen (illustrated in Christiansen 1985, 232) and elsewhere. Inside the hull There is relatively little evidence in the skaldic corpus to tell us how ships were arranged inside the hull, although the vocabulary that is used appears to be quite specific. In building viking ships, the frames or ‘floor-timbers’, which were transverse members providing internal support, were added after the first rows of planks had been built up (see the illustrations in Crumlin-Pedersen 1986, 96,

Ships and Sailing

151

100). According to Falk (AnS, 46), these were called r›ng (f., pl. rengr) in ON.55 This word appears twice in the corpus: the part concerned is described as ristin ‘carved’ (Hharð 16) and is said to have trembled (skalf HHarð16; bifask Kali) in the currents. This is an exact description: the curved frames were normally cut out of one piece of wood to fit the hull exactly at the point at which they were inserted, and so needed skilled carving, and they were only attached to the hull at certain points, to give elasticity, and so would move independently (Vadstrup 1997b, 105–10; see also Crumlin-Pedersen 1997b, 87–8, for frames with mouldings from Hedeby; see fig. 4.7). At the battle of Sv›lðr, warriors fighting inside a ship stumbled over the þoptur as they retreated from the attack (Hókr 6). This is the only Viking Age example of þopta (f., pl. þoptur), though the noun þopti derived from it is discussed in the next chapter. In later sources þopta means ‘rowing-bench’. There do not appear to have been rowing-benches as such in viking ships, however, and rowers may have sat on chests (in which they could also have kept their belongings) placed on the loose decking (see the one from Hedeby pictured in Crumlin-Pedersen 1997b, 142–3), or on the thwarts, or cross-beams. Some ships (e.g. Gokstad) appear to have had just one kind of cross-beam, called a ‘bite’ (from ON biti ‘beam’, not recorded in the skaldic corpus) by archaeologists, one above each frame, which supported the decking, while later ships (e.g. Skuldelev 1, 3 and 5) had a system in which the ‘bites’ were set lower in the hull, and further strengthening cross-beams were inserted above some of them.56 These ‘bites’ could support a deck, while the upper beams could be used as thwarts (Olsen and Crumlin-Pedersen 1967, 140, 168; see also the diagram in Crumlin-Pedersen 1997b, 87). These upper cross-beams are thus the þoptur over which the warriors stumbled.57 The deck-planks were called þilja (f., pl. þiljur), and Sigvatr describes Óláfr mowing down the men on Erlingr Skjalgsson’s ship so that valr lá þr›ngt á þiljum ‘the dead lay tightly-packed on the deck’ (Sigv VII,2; similarly Arn V,7).58 A ship-kenning þilblakkr ‘(deck-)plank-horse’ (Sigv II,12) can be compared with ship-kennings of the same type but with the word borð as determinant (discussed above), although here the first element is þil (n.), a collective noun from the same root meaning ‘decking’. A problematic word which it has been suggested (Lindquist 1928) is synonymous with þilja is skokkr (m., pl. skokkar), twice used in a context of blood 55 56

57 58

The word was adopted into late Old English (Sandahl 1951, 114–15). ON biti appears to have been adopted into Middle English, with roughly the same meaning (Sandahl 1951, 30–31), though the Modern English ‘bitt’ now has a different meaning. It also appears in Russian, though not recorded before the seventeenth century (Svane 1989, 29–30). The Irish loan-word topta is used in a very similar context in CCS, 42 (see also pp. xii and 100). LP makes the curious suggestion that þiljur (pl.) is equivalent to the lypting (discussed below), but I cannot find any evidence for this, and AnS, 48, cited as a source, is clear that þiljur are deck-planks.

Image not available

4.7 A frame from the Skuldelev 3 ship. Photo: Viking Ship Museum, Roskilde.

Ships and Sailing

153

flowing in the ship (Arn V,21; B›lv 4). However, in two other stanzas by B›lverkr the word seems to be used as a pars pro toto for ‘ship’: in B›lv 5, it has a þr›mr ‘rim, edge, rail’, while in B›lv 8 it is dýrr ‘splendid’ and lies on the dark wave. These parallels with the uses of borð suggest that skokkr was indeed a (deck-)plank. Several stanzas refer to some kind of a structure called a lypting (f., pl. lyptingar), sometimes translated ‘poop-deck’ and interpreted as a raised place in the after-stem of the ship. The contexts give us various clues to the nature of this part of the ship. Erlingr Skjalgsson’s last stand was einn í lyptingu ‘alone in the lypting’ (Sigv VII,3), because his attackers cleared the ship of his followers from one end to the other, but perhaps also suggesting that the lypting afforded some protection for defence because he held out lengi ‘for a long time’. In a stanza summarising Magnús’ heroic career as a sea-warrior (Arn II,16), Arnórr says that Visundr carries him like a hawk innan í lypting ‘inside the lypting’, suggesting either that it was a roofed structure big enough to sit, if not actually stand, in, or that it was surrounded or enclosed in some other way. This is also suggested by Arnórr’s description, earlier in the same poem, of the ugly surf which dreif útan á lypting ‘drove onto the outside of the lypting’ (Arn II,10), again implying that it had some kind of protective function. Unfortunately, nothing resembling a structure corresponding to any of these descriptions can be traced in the archaeological record, although the decking of the Gokstad ship is higher near the rudder than elsewhere (Brøgger and Shetelig 1950, 155) and the small warship Skuldelev 5 had ‘a little horizontal seat for the helmsman resting on the top edge of the fourth strake’ (Olsen and Crumlin-Pedersen 1967, 139).59 The related word lopt (n., pl. lopt) is used in a confused kenning for ‘captain’, unnar eykjar loptbyggvir ‘lopt-dweller of the traction-beast of the wave [ship]’ (Sigv IV), which at least seems to confirm the suggestion of the lypting examples that this was the captain’s place. The word vengi (n., pl. vengi) in a ship-kenning vengis hj›rtr ‘hart of the vengi’ (Hharð 4) has been interpreted as referring to some kind of ship’s cabin (AnS, 10), but the evidence is hardly conclusive. The word vengi also appears in a ship-kenning in Ótt II,3, where it is usually interpreted as the homonym meaning ‘plain’ (vengis dreyra blakki ‘steed of the lifeblood of the plain [water]’). It is quite possible that Haraldr’s kenning is an abbreviated kenning (on which see Turville-Petre 1976, lii) of this type, with a half-kenning (‘of the plain’ rather than ‘of the lifeblood of the plain’) as the determinant of the base-word hj›rtr ‘hart’. Despite the contention of Sayers (1998, 57) that Hharð 4 confirms the interpretation of vengi as ‘a cuddy or small shelter’, the context in which this word occurs (SnESkskm, 129) suggests no more than that it was thought to be some part of the ship. This misunderstanding probably derives

59

The ‘løfting’ created for the Skuldelev 3 replica, Roar Ege (Vadstrup 1997b, 129–30) does not appear to be based on anything actually found in the ship, but is a simple seat with storage space underneath.

154

Ships and Men in the Late Viking Age

from Haraldr’s stanza itself, the relevant first half of which is quoted earlier (SnESkskm, 75). It is usually assumed that most ships had some kind of cover or awning for the open hold, called a tjald (cp. ModE ‘tilt’). This word will be discussed below, as it is difficult to separate it from the same word used to refer to sails. Rudders, oars and shields Viking ships could normally be propelled either by sail (discussed in the next section), or by manpower on the oars. In both cases, steering was done by a rudder, often called a ‘steering oar’ as its shape was not very different from a rowing oar, attached to the starboard rear side of the ship and manipulated by a tiller. There is quite a rich vocabulary for the oars, rudders and their attachments in the skaldic corpus, mainly because a few stanzas concentrate on this aspect of travelling by ship. The word for ‘oar’ is the common Germanic ár (f., pl. árar; also ór[ar]), which occurs frequently in the skaldic corpus.60 Another word for ‘oar’ is rœði (n., pl. rœði; Þór; ÞjóðA IV,20). Þjóðólfr Arnórsson twice uses the periphrastic term sæfang (n., pl. sæf›ng) ‘sea-implement’, in both cases because he needs to alliterate on s- and rhyme with -›ng- (ÞjóðA IV,20,21). The blade of an oar is called blað (n., pl. bl›ð), while hlumr (m., pl. hlumar) is used of its grip (both in ÞKolb III,3, the latter also in Þór). Óttarr refers to ships skreyttum órum ‘adorned with oars’ (Ótt II,4). This might just refer to the aesthetics of a warship and its flight of oars, but we do know that some oars were painted and carved. The Oseberg oars had both mouldings and traces of painted decoration, with the decoration designed to be on the part of the oars that was just outside the oarports (Brøgger and Shetelig 1950, 176). More workaday oars are described as ferkleyf þ›ll ‘rectangular pine’ and could be sortaðr ‘tarred [lit. blackened]’ (ÞjóðA IV,21). Oars suffered from use, which may be why the Oseberg oars were newly made for the burial (Brøgger and Shetelig 1950, 176–7). Óttarr describes Óláfr’s journey in stormy weather to raid Sweden: m›rg mj›k róin ór sleit mikla bóru und þér ‘many a much-rowed oar tore the great billow underneath you’ (Ótt II,4), and much power would be needed to counter the thrust of the waves. Thus Arnórr describes a delay caused by difficult rowing: seinkun varð, þás en ljóta bára hnikaði ór ‘there was drag, as the ugly breaker caught the oar(s)’ (Arn II,3). But good oars should last some time before falling apart, as noted by Þjóðólfr Arnórsson: ært mun, áðr sortuð sæf›ng ganga í tvau ‘there will be rowing before the tarred sea-implements split’ (ÞjóðA IV,20). This act of rowing (róðr, m., pl. róðrar, Steinn II; see also the verbs æra ‘to row’, ÞjóðA IV,20; róa ‘to row’, Þór; Arn III,13; ÞjóðA IV,21) is hard work, as 60

Sindr 1; ÞKolb III,3; Ótt II,4 (twice); Arn II,3; ÞjóðA IV,19,20,21; AnonXI Lv,4; Gísl I,3. See also Tindr I,5 (in a kenning), and HHuI,27,49.

Ships and Sailing

155

suggested by the verb slíta ‘to tear’ used of the movement of pulling the blade out of the water,61 and knýja ‘to press’ of the act of pushing down on the oars, or pulling them through the water (AnonXI Lv,4; Gísl I,3), although when a ship is leaving a sheltered harbour, the oars more easily falla í sjá ‘fall into the sea’ (ÞjóðA IV,19). The ára burð ‘movement of the oars’ is admirable when the rowers slíta rœði rétt ór verri ‘pull the oars straight out of the sea’ (ÞjóðA IV,20; the word used for ‘sea’ here, v›rr, actually means ‘stroke of an oar’, cp. ÞjóðA IV,21 and the kenning varrláð ‘land of the oarstroke [sea]’, ÞKolb III,10). From within the ship, this flight of the oars looks like the wings of an eagle: es sem líti innan arnar væng (ÞjóðA IV,21). Archaeological finds provide a number of examples of the different ways in which the movement of the oars was controlled at the point where they rest on the ship, thus some ships have oarports (holes in the top strake or sometimes lower), while smaller boats tend to have tholes, wooden pins which act as a fulcrum for the oar (various types are illustrated in Crumlin-Pedersen 1997b, 128–9). The word that later became the usual one for a tholepin in Norwegian, ON keipr (m., pl. keipar) is recorded in a stanza from the very end of our corpus (AnonXII B,3). The poet, describing sailing in storm, artfully uses the verb knýja ‘press’ (also, as we have seen, used of the actions of rowers), of the wind’s force on the pin. Otherwise, the skaldic corpus has two principal words for oarports, though both tend to be used metaphorically of the oarsman’s place at the oar, rather than technically of the device which fixes this position. One of these is hár (m., pl. háir), which is however used in a technical sense in hár heldr sjau tøgum ára til varra ‘the oarport [sg., used in a pl. sense] holds seventy oars to the stroke’ (ÞjóðA IV,21). V›rr ‘stroke of an oar’ is also used in the plural here, which has led most translators (e.g. LP; also Poole 1991, 60, though he translates it as ‘stroke’ in the previous stanza) to interpret it in the transferred sense ‘sea’. But the plural is appropriate as there are seventy oars, each ready to take a stroke, and it is likely that the primary meaning is intended here, just as hár is used technically rather than metaphorically. We thus have an image of each oar correctly positioned in its oarport, ready for the next stroke. Otherwise, the Swedish king ¡nundr brings an army to raid the Danes at há (Sigv X,4), which hardly means more than ‘by ship’, and the word appears in two kennings: hádýr ‘animal of the há [ship]’ (Þloft II,4)62 and hákesjur ‘spears of the há’ (Steinn II), which probably means ‘oars’ (Perkins 1986–9, 114). It is also found in Old English in the eleventh century. The account of Harthacnut’s tax of 8 marks on each crew member (?) uses the Norse word: viii marc æt há (ASC 1040C). The word hásæta ‘sitter at the há, oarsman’ is used in the E-version when Edward decides to change the manning of some of his fleet, both eorlas ‘earls’ and hásæton (ASC

61 62

ÞKolb III,3; Ótt II,4; ÞjóðA IV,20,21. See Gjessing 1986 on the physiology of rowing, and McGrail 1987, 207–16, on various aspects of rowing. This kenning may be implied by the mountain-name Hádýr (Hkr II, 309).

156

Ships and Men in the Late Viking Age

1052E). This suggests that ON háseti with the same meaning was current in the eleventh century, even though the word is not recorded in Scandinavian sources until later (Malmros 1986, 104). In their accounts of Harthacnut’s tax, the D- and E-versions of the Chronicle use the word hamele (ASC 1039E, 1040D), apparently derived from ON hamla ‘oar-loop, grommet’ (AEW). Large warships in the eleventh century did not use such simple devices as loops to control their oars, and in the corpus hamla (f., pl. h›mlur) always means ‘oarsman’s place’, usually in contexts describing the gathering of a troop. Magnús, before embarking on an expedition to Denmark orders his men to samnask til hverrar h›mlu ‘come together at every hamla’ (Arn II,9), and similarly, his troop bar hervæðr til h›mlu ‘carried [or ‘wore’] their war-garb to the hamla’ (Arn III,2). In a battle-context, the leader can order his men to skilda h›mlur ‘provide the h›mlur with shields’ (ÞjóðA III,13) as a defensive measure. Although Finnur Jónsson interpreted this as tying the shields to the oar-loops (Skjd B I, 342), it is more likely that it means they were placed on the shield-rim at each rowing-position. The occurrence in ASC of these two ON words, usually interpreted as referring to a ‘rowlock’ (more correctly, ‘oarport’), hence ‘oarsman’, has been fully discussed by Rodger (1995), with particular reference to the calculations historians have made, on the basis of these entries, about the number of men in each of the sixty-two ships of Harthacnut’s fleet. Noting the point that oar-loops were not the current technology for large warships, Rodger suggests that the word hamla referred, not to any part of a ship, but to ‘a unit of men which may be one man or several’ (1995, 401), in the context of a rate of taxation. The transfer of such a sense to OE would imply that some kind of fleet-organisation was already known in Scandinavia, which used this out-of-date term in an abstract rather than a literal sense. Rodger might have pointed to the provision in the leiðangr law (NGL I, 203) which stipulates that ‘Every man who had a hamla on the journey out from home shall bring it back, unless he provides the shipmaster with another man whom he is willing to accept’.63 The first half of this stipulation might make it sound as if the hamla was some concrete object (such as a rowing-bench, cp. NGL I, 202) that could be taken away and brought back, but the second makes it quite clear that it refers to a unit of men. In the equivalent sections of the Gulathing law (NGL I, 98–101), it is less clear whether hamla refers to a specific position in the ship or to a unit of men, but the latter meaning is certainly possible. However, the word is still used in a very concrete sense in other laws (NGL I, 75, of a rotten oar-loop) and the meaning ‘unit of men’ is more difficult to read into the skaldic examples of both há and hamla cited above. It is possible to imagine that the word was used to describe a unit of one or more men holding the oar, ready for the stroke (ÞjóðA IV,21), or to imagine 63

The translation is from Larson 1935, 321, though I replace his ‘rowlock’ with the original ON term hamla. Rodger (1995, 397) laments the need ‘to use twelfth- and thirteenthcentury Norwegian evidence for want of the eleventh-century Danish information we need’, but is not aware of the eleventh-century (Norwegian-Icelandic) evidence.

Ships and Sailing

157

men joining their units at the beginning of an expedition (Arn II,9; Arn III,2), although a technical meaning of ‘oarport’ is just as likely in the former, while in Arnórr’s two stanzas a concrete reference to the part of the ship occupied by the warrior-crew (the ‘room’) is also just as likely. In ÞjóðA III,13, hamla cannot mean ‘unit of men’, since the stanza describes Haraldr ordering his troop (fylking) to provide the h›mlur with shields, and makes clear that the ship is then completely encircled by these shields. Even in the ASC, it is likely that a part of the ship is meant. In ASC 1039E the expression (which occurs twice) is clearly geald . . . scipan æt ælcere hamelan viii marc ‘paid for . . . ships [at the rate of] eight marks to each hamele’. Although in effect this meant payment for a certain number of crew (although we do not know how many would be considered to be in a hamele), we cannot, I think, take the further step of asserting that either há or hamla had a meaning ‘unit of men’, in either OE or ON in the eleventh century, though it is easy to see how such a meaning could have developed later. The adjective hábrynjaðr occurs in three more or less contemporary stanzas in the corpus (ÞjóðA IV,22; Þfagr 4; Steinn III,14). LP enters it under compounds in hó- ‘high-’ and translates ‘ ‘‘höjpansret’’, hvis panser går höjt op’. For Falk (AnS, 38) the adjective brynjaðr refers to warships which have metal strips or plates covering their prow, and he compares hábrynjaðr [sic] with harðbrynjaðr. The simple adjective brynjaðr is used once of ships (B›lv 2), while harðbrynjaðr is used twice of Knútr’s ships (Ótt III,1; Hallv 3). A brynja is normally a mail-coat or other body-armour, and the other occurrence of the adjective brynjaðr in the corpus is with this meaning (Sigv XII,10; cp. HHuI,37). Using it of a ship involves a metaphorical extension of the meaning and such a metaphorical extension could occur in a number of ways. One possibility is to imagine the ships as having some sort of naval equivalent of body-armour, in the form of some kind of metal covering. But I know of no archaeological evidence for such protection for Viking Age ships.64 It seems more likely that the word brynjaðr imagines the ship as protected by its row of shields,65 and that the compound hábrynjaðr is formed with hár ‘oarport, rowing-station’, since the positioning of the shields on a warship corresponded to these positions between the frames (Brøgger and Shetelig 1950, 151; Olsen and Crumlin-Pedersen 1967, 137). A closer look at the examples suggests that this interpretation fits the contexts better. It has to be remembered that shields could not be placed on the shield-rail when the ship was under oars, only when it was being sailed or (most likely) when it was anchored in harbour. Þorleikr fagri’s stanza describes Sveinn’s large fleet intending to defend Denmark from the attack of Haraldr harðráði (Þfagr 4). 64 65

Although LP and AnS anachronistically imagine this as plate mail, chain mail would have been easier to produce and is perhaps more likely to have been used on ships as well. Bjarni Aðalbjarnarson (Hkr III, 144) translates and explains hóbrynjaðar [sic] in ÞjóðA IV,22 as ‘brynvarin (skjölduð)’, Poole (1991, 61) as ‘with their high shields’. To my knowledge, Malmros (1986, 101) was the first to make the explicit suggestion that this adjective refers to shields at the rowing-positions.

158

Ships and Men in the Late Viking Age

Þjóðólfr Arnórsson’s stanza gives the other side of the coin (though on a later occasion), Haraldr’s fleet on its way to attack Denmark (ÞjóðA IV,22). In both of these cases, the warships would naturally carry shields. Neither stanza shows the fleet actually moving and Þjóðólfr’s actually pictures the ships lurking in various sheltered spots (because of a storm, according to Snorri, Hkr III, 143–4). Steinn Herdísarson’s stanza describes the gifts Óláfr kyrri makes to his followers, which include hábrynjuð skip ok steinda kn›rru ‘hábrynjuð ships and painted knerrir’ (Steinn III,14). Here, there is certainly an intended contrast between two types of ship. The stanza particularly emphasises the uniqueness of Óláfr’s generosity, and it is likely that a particularly spectacular gift of a warship would include a set of shields, especially painted in the ship’s colours, as in the case of the Gokstad ship (Brøgger and Shetelig 1950, 152). Some decades earlier, two poets used the adjective harðbrynjaðr of Knútr’s ships. Ótt III,1 is almost certainly the first stanza of this poem, as it describes the young king setting out on an expedition for the first time (Jesch 2001a), preparing and launching his ships, and the reference to harðbrynjuð skip is followed almost immediately by a reference to red shields (Knýtl, 101–2): Hilmir, bjóttu ok hættir harðbrynjuð skip kynjum. Reiðr hafðir þú rauðar randir, Knútr, fyr landi. Chief, you prepared a hard-armoured ship and took astounding risks. Knútr, you surrounded the country with red shields, angry.

Hallv 3, on the other hand, shows the same fleet on the way to England. Even if it could carry shields on the shield-rail when under sail, this would have been an extravagant way to cross the North Sea (Brøgger and Shetelig 1950, 152). The occurrence of the adjective brynjaðr in B›lv 2 is supposedly the example that is decisive for the meaning proposed by Hjalmar Falk and Finnur Jónsson. The latter prints and translates the stanza as follows (Skjd B I, 355): Hart kníði sv›l svartan snekkju brand, fyr landi, skúr, en skrautla bóru skeiðr brynjaðar reiði; mætr hilmir sá malma Miklagarðs fyr barði; m›rg skriðu beit at borgar barmf›gr hóum armi. Den svale vind drev stærkt skibets sorte stavn frem, men de pansrede skibe bar deres prægtige udstyr langs landet; Miklegårds herlige fyrste så metalbeslagene på stavnene, mange brystskinnende skibe skred hen til borgens höje arm.

Finnur saw a reference to metal on the prow of the ship in malma fyr barði, and this confirmed his interpretation of brynjaðr as ‘pansret’. He was criticised by

Ships and Sailing

159

Kock (NN, 2035), who proposed quite different syntax in lines 5 and 6, roughly: ‘the glorious king saw the beaches of Miklagarðr before his prow’. Although Kock’s interpretation of malm- as ‘beach’ is probably wrong, his syntax is almost certainly right, as the parallels he adduces show, and the correct translation of these lines is almost certainly that in Hkr III, 71: ‘Hinn ágæti konungur (Haraldur) sá málmþök Miklagarðs fyrir stafni’, in which we have Haraldr catching sight of the copper roofs of Constantinople as he approaches it by ship. In conclusion, there is no evidence in the skaldic corpus (or the archaeological record) for any metal plating or other covering on Viking Age warships.66 When the poets use the adjective (-)brynjaðr, they are either referring in a general way to the fact that the ship is protected and that it carries shields, or, in a few cases, they may be referring more specifically to the practice of carrying shields on the shield-rim of a warship (ÞjóðA IV,22; Steinn III,14; possibly Ótt III,1). Even if the first element of the adjective in ÞjóðA IV,22, Þfagr 4 and Steinn III,14 is hó‘high-’ rather than há- ‘rowing-position’, it remains the case that the second element -brynjaðr must refer to the warship’s shields (as in harðbrynjaðr in Ótt III,1) rather than any armour-plating. Many depictions of viking ships show them displaying their shields, which protrude above the top strake.67 The word for the steering arrangement on a viking ship is stýri (n., pl. stýri), but it is not entirely clear whether this was used of both the rudder and the tiller, or just the rudder (Þór; Hfr II,1; Ótt II,20; Arn II,10; Steinn III,6).68 But when skalds describe the movements of the stýri they seem to have primarily the rudder in mind. Thus, stýri mól ‘the stýri made a grinding noise or movement’ (Hfr II,1), referring to the rudder’s turning in rough seas, and braut stóran straum of stýri ‘a great current broke over the stýri’ (Steinn III,6). The stýri was skafið ‘shaved, planed’ (Ótt II,20), and it could apparently be decorated with some kind of gold that trembled when the ship was sailing: gullit rauða um skeiðar stýri bifðisk ‘the red gold around the stýri of the skeið shook’ (Arn II,10). Although this is usually taken to mean that the ship’s hull was gilded in the general area of the rudder (e.g. Whaley 1998, 163), it may also be that the tiller was decorated in some way. The Gokstad ship, not otherwise highly decorated, had a tiller that was both carved and painted in three colours (Nicolaysen 1882, 44, pl. XI). The rudder is prominently displayed in a number of illustrations of ships.69 66

67

68

69

HHj, 13, mentions iárnborgir ‘iron fortifications’ around the prince’s floti ‘fleet’ (or ‘ship’?). What exactly is meant is not clear (though there may be some connection with skjaldborg, see ch. 5), nor can this poem be used as datable evidence. D 220, D 258, D 271, Ög 181. See also several of the Gotland picture stones (GB I, figs 79–81, 85–6, 89, 97, 105, 107, 116, 128, 142) and one of the Lowther hogbacks (BACASSE II, 130). Róðr in the sense ‘rudder’, rather than ‘act of rowing’, first appears in the twelfth-century runic graffito N 532. The meanings of what might be the same word in U 11 and U 16 are disputed (see Hjärne 1946 for a lengthy discussion). I cannot see that there is anything in these inscriptions to justify a nautical interpretation, even though Hjärne argues (1946, 111–15) that [i ruþi hakunar] on U 16 refers to a naval expedition of some sort, led by Hákon. D 77 (possibly), D 220, D 271, Ög 224, U 979, U 1052, U 1161; see also Vg 119, the

160

Ships and Men in the Late Viking Age

Masts, sails and rigging While masts and mast-fixings are relatively well represented in the archaeological finds, the details of the rigging of viking ships are harder to reconstruct and are a matter of lively debate between the specialists, who have to base their arguments on visual representations such as the Gotland picture stones, or on ethnographic evidence from more recent times (Christensen 1979). The skaldic evidence is unlikely to resolve these controversies, but does provide a range of vocabulary for many of the parts of the rig. The viking sail was a square (or rectangular or trapezoid) sail raised onto the mast and supported along its upper edge by a yard. The most common word for ‘mast’ in the skaldic corpus is v›ndr (m., pl. vendr). This word has a primary meaning of ‘rod, stick, pole’ and Falk (AnS, 56) sees its use for ‘mast’ as an instance of poetic language. The more prosaic term sigla (f., pl. siglur) occurs once in the corpus (ÞjóðA III,31), but in a kenning meaning ‘sword’.70 V›ndr occurs three times as a simplex (Arn II,4; ÞjóðA I,2; AnonXII B,3), once in the compound vandlangt ‘long-masted’ referring to Magnús’ ship Visundr (ÞjóðA I,4) and twice in the ship-kennings vandar valr ‘horse of the mast’ (Sindr 2) and vandar dýr ‘animal of the mast’ (Eskál III,20). It is likely that v›ndr was a word normally used by sailors in the late Viking Age and was not especially poetic. In Arn II,4 the masts are said to bifjask ‘shudder’, while in ÞjóðA I,2 ótt veðr vægðit sveigðum vendi ‘the bad storm did not spare the bent mast’, meaning that the wind caused the mast to bend. Both poets are referring to the same journey, Magnús góði’s return from Russia. The mast of a ship captained by Magnús berfœttr (Anon XII B,3) is also said to skjalfa ‘tremble’, but we are given the additional information that it was seventy feet long (Foote 1978, 65). As archaeologists find ever longer and longer warships (the Roskilde 6 ship, dated to 1025, was 36m long), such a long mast seems less incredible than when Foote first drew attention to this stanza. Another word for ‘mast’, which Falk brackets with v›ndr as poetical usage, is laukr ‘onion, leek’. Here, the metaphor is more obvious, and the word occurs only twice in the skaldic corpus: in B›lv 5 lofðungr reisti lauka við þungan sæ ‘the leader raised the masts in [despite] the heavy sea’, while Þfagr 3 contains the ship-kenning lauks glæsidýr ‘gleaming animal of the mast’. The top of the mast was called húnn (m., pl. húnar), a word with the primary meaning of ‘bear-cub’, but with a number of other meanings, including ‘die, playing-piece’, which apparently gave rise to the nautical meaning from the often quadrilateral strengthening of the area at the top of the mast through which the halyards passed and on which the shrouds could rest (AnS, 59; see

70

non-runic U 1001, several of the Gotland picture stones (GB I, figs 71, 79–81, 86, 89, 104, 108, 112, 139, 142), and the Stow (Lincs.) graffiti (BACASSE V, 292–3). Tré ‘tree’ is used of a mast in HHuI,26. Sigla appears to have been borrowed into Russian, and is attested there in the eleventh century (Svane 1989, 29), suggesting it was commonly used then in OEN.

Ships and Sailing

161

Image not available

4.8 The ship’s vane from Heggen, Buskerud, Norway. Photo: Universitets Oldsaksamling, Oslo.

Christensen 1979, 185, for a concise definition).71 These functions could be achieved in a variety of ways, and the skaldic instances do not give any sense of precisely what forms the húnn could have (see Andersen 1997b, 179–80, 188–91 for some possibilities). In Korm I,3, a leader is one who bindr hún beinan ‘ties the húnn straight’. Otherwise, the húnn is usually mentioned in connection with the sail, in contexts which make it clear that the sail was raised to the top of the mast (Ótt II,20; Gísl I,14), similarly the adjective hýndr ‘raised to the húnn’ (ÞjóðA I,4; Valg 6) and the sail-kenning húnskript ‘decorated cloth of the húnn’ (ÞjóðA I,2). Húnn also appears in ship-kennings, thus hreinn húnlagar ‘reindeer of the liquid of the húnn [sea]’ (Hókr 2) and hreinn húnferils ‘reindeer of the way of the húnn [sea]’ (Þfagr 4). Sigvatr has the kenning hyrsendir húna ‘sender of the fire of the húnn [gold]’ (Sigv XII,16), suggesting that mast-tops were regularly gilded. It is usually assumed that Arnórr’s description of élmars typpi ‘storm71

The rectangular object at the top of the mast on the Gosforth slab (or ‘Fishing-Stone’) has been interpreted as a húnn (BACASSE II, 108–9).

162

Ships and Men in the Late Viking Age

steed[ship]-tops’ which glóðu eldi glík ‘glowed like fire’ (Arn II,10) also refers to gilded mastheads (in poetry, it is always gold that glows like fire). However, there is another possible interpretation of Arnórr’s metaphor. Whaley (1998, 77) notes that typpi ‘can be assumed to share with its cognate toppr the senses ‘top, knob, masthead’ . . . and ‘forelock’. The latter meaning seems indicated in a kenning which compares a ship with a horse, and something on the fore-stem is a better parallel to a horse’s forelock than something on the mast-top amidships. The ship’s ‘golden forelock’ may have been a gilt copper ‘vane’, of which several are known from the eleventh century (see fig. 4.8) and which Blindheim (1982) has demonstrated were attached to or in the fore-stem of a ship, rather than to the top of the mast (although this was also done later). These had no real function beyond that of decoration, and possibly as indicators of rank and status (see Christensen 1998 for a convincing refutation of a recent suggestion that they were navigational instruments). The yard, the horizontal beam supporting the sail, was called ró (f., pl. ráar). In Sigv X,8, Knútr’s ships are said to have carried their sails við ró í byr ‘on the yard in a favourable wind’, which does not mean much more than that the ships were able to take advantage of the wind to sail, so that rowing was unnecessary. The yard was one of the parts of the ship particularly vulnerable to damage in a storm (AnS, 13), and ÞjóðA I,2 provides a dramatic picture of sailing in brisk weather: ekin ró dúðisk ‘the ‘‘driven’’ yard shook’, with the yard shuddering as it is turned to catch the wind.72 Archaeological finds of yards are scarce, but they do seem to have been typically very long and slender (McGrail 1987, 232). Several of the ‘thin round spars of fir or pine wood’ found in the Gokstad burial have been identified as yards (Nicolaysen 1882, 39, pl. IV). The common Germanic word for ‘sail’ segl (n., pl. segl) occurs five times in the corpus, commonly modified by an adjective indicating its appearance or condition.73 Sigvatr describes his pleasure in sailing, despite the hard weather: ›rðigt veðr úti á fj›rðum skóf vindblásit segl í vási ‘the high wind out on the fjords scraped the windblown sail in the difficult conditions’ (Sigv III,9). This unusual use of the verb skafa (cognate with English ‘shave’) has been interpreted in a variety of ways, but may refer to the tearing noise of the sail being blown back and forth, or perhaps even to the actual tearing of the sail. In Sigv X,8 Knútr’s ships carry bló segl ‘blue sails’ (though the colour should be imagined as something closer to black than to bright blue). In ÞjóðA I,4, as already noted, the sails are hýnd ‘raised to the húnn’ and make noises as they are blown against the forestay: rýndu við stag. In Gísl I,14 the ships’ sails, also sett við húna ‘raised to the húnns’, are sædrifin ‘sea-drenched’. Óttarr gives a clear description of Óláfr 72

73

Ró is used in HHuI,33,49. The latter stanza also provides us with the term rakki, not preserved in the skaldic corpus, in the ship-kenning rakka hj›rtr ‘deer of the mast-ring’. The mast-ring, or parrel, was a semicircular or crooked piece of wood around the mast which enabled the yard to go up and down it, see the example from Oseberg illustrated in Andersen 1997a, 170. Segl also occurs in HHuI,29.

Ships and Sailing

163

sailing to Sweden in stormy weather (addressed to the king in the second person): stundum neyttuð segls ok settuð sundvarpaði ‘occasionally you used the sail and set (it) against the strait-disturber [wind]’ (Ótt II,4), contrasting with a description of rowing that follows in the stanza, and again referring to the raising of the sail to catch the wind. The verb neyta can imply ‘use to near-destruction’ (e.g. ÞjóðA I,4; ÞjóðA IV,23), and that would be appropriate in this description of stormy sailing. But there is a range of possible words for ‘sail’. Viking Age sails could be reefed, that is shortened to lessen the area in contact with the wind, and each horizontal section by which they could be shortened was called rif (n., pl. rif; AnS, 68–9).74 The plural form of this word is used as a pars pro toto for ‘sail’ in Valg 6 (Foote 1984b, 235). Referring to the way in which the sail is produced are vefr (m., pl. vefjar) ‘weave’ (ÞjóðA III,8; Gísl I,15) and vóð (f., pl. váðir or væðr) ‘loom-width of cloth’ (Ótt III,2), of which several would go to make one sail (AnS, 63).75 The sail-kenning húnskript ‘decorated cloth of the húnn’ (ÞjóðA I,2) may indicate some kind of device on the sail. Falk (AnS, 63) suggests this device was embroidered, on the basis of a late prose source, but it is just as likely to have been painted. Or the poet may simply have meant that the stitched sections and the reef-bands of the sail made it look like a piece of worked cloth (see the drawing in Andersen 1997b, 188). Two further words need more detailed discussion. Skaut (n., pl. skaut) is used twice by Óttarr. The primary sense of this word is ‘something triangular’, hence ‘lower corner of a sail, clew’. This meaning obviously does not work in Ótt II,20, where the poet says skaut þats drósir spunnu lék við hún ‘the skaut spun by women played against the mast-top’, even if we assume the word could also refer to an upper corner of the sail. There are two possible ways of extending the meaning of skaut. As in the English cognate ‘sheet’ (Sandahl 1982, 91–2), the word for a corner of a sail could be extended to the rope attached to that corner. Further extending this to mean ‘rope’ in general would give us a reference to one of the many ropes that converged around the top of the mast. This interpretation does not work in this context because, under sail (as in this stanza), such ropes would be taut rather than playing against the mast. ‘Spun by women’ would also suggest woollen ropes, for which there is no Viking Age archaeological evidence, although there are ethnological parallels.76 More likely is that, like rif, the word skaut is used in this context as a pars pro toto for ‘sail’. We have already seen several examples of sails raised to the mast-top, and the reference to 74 75 76

The English word ‘reef’ is borrowed from ON (Sandahl 1958, 90–91). The procedure is clearly described in McGrail 1987, 239. HHuI,26 calls a sail a vefnisting ‘weave-sewing’, combining the two crafts used in producing the sail. On woollen sails, see Andersen 1995. As demonstrated in the exhibition Seilet som kvinnene spant held in Trondheim in the spring of 1999 (also in Roskilde later that year). Most Viking Age ship-ropes were made from bast (Crumlin-Pedersen 1997b, 188–90), though a full analysis is still awaited. Ohthere mentions the use of whale- (or walrus-) and seal-hides for ship-ropes in ninthcentury Norway (Lund 1984, 20, 54, 58–9).

164

Ships and Men in the Late Viking Age

spinning women provides us with an earlier stage in the production of the sail textile than the terms vefr and vóð just mentioned. In Ótt III,2 skaut- appears in the ship-kenning skauthreinn ‘reindeer of the skaut’. In theory, any of the meanings for skaut just outlined would work here, but again, ‘sail’ is probably the most likely, given that a sail is more prominent than its corner, or a rope attached to it, and that Óttarr describes the blowing of the sail, or vóð, in the same stanza. In two stanzas by Arnórr, the poet uses the word tjald (n., pl. tj›ld) of a sail (Arn II,16; Arn V,19). Normally this word is assumed to refer, either to a ship’s cover or awning, or to a land-tent, but in any case to something that provided sailors with shelter from the elements. However, I would like to suggest that this meaning could also be extended to ‘sail’. The two meanings are in any case not so easy to keep apart, as it is quite likely that sails were used to provide shelter when not in use for sailing.77 A photograph reproduced in Crumlin-Pedersen and Vinner 1984, 206, shows sails being used as tents in Trondheim at the beginning of the twentieth century, and the crew of the replica viking ship Odin’s Raven regularly used their spare sail as a tent (Binns 1980, 169, 215, 221; Ingram 1982, 37, 149), though I do not know what evidence there is that viking ships carried a spare sail. The etymology of tjald is uncertain, but both NDEW (s.v. telt) and ÍO would derive it from an Indo-European root *delâ ‘to spread out’, comparing a Greek word dólãn ‘small sail’. In ON, the word can also refer to a wall-hanging, and the idea of vertical cloth is what tents, sails and wall-hangings have in common. The Old English poem Exodus metaphorically links sails and tents in its description of God’s covering the Egyptians with a cloud: segle ofertolden ‘tented over with a sail’ (line 81; see Lucas 1977, 89–90). But far-flung parallels can only suggest, and the proposed interpretation rests more convincingly on the collocations of tjald in Arnórr’s two stanzas. In Arn V,19, it is said that Þorfinnr sleit blóu tjaldi fyr útan eyjar ‘tore the blue tjald beyond the islands [out at sea]’. This same colour-adjective was used of sails in Sigv X,8, as noted above. And since this wearing out of the tjald is said to have taken place at sea, it seems more likely to refer to a sail than to an awning or tent (compare the use of skafa in Sigv III,9, discussed above). The stanza goes on to mention snow and frost (probably in connection with the mast), and these would be just the conditions in which sails could be torn.78 In another stanza, the poet praises Magnús for a life spent mostly und drifnu tjaldi ‘under the drenched tjald’ (Arn II,16). It seems more heroic to imagine the king standing proud under the sail, however wet, than cowering under a sodden awning. It has already been noted that Gísl I,14 describes sails as sædrífin ‘sea-drenched’, and ÞjóðA III,8 pictures King Haraldr sailing und breiðum vef ‘under the broad weave’. Thus, I would like to suggest that, in these two stanzas by Arnórr, tjald specifically means ‘sail’. This may be a 77 78

Archaeologists have been unable to determine whether the textile fragments found at Oseberg were ‘sails or tents’ (Andersen 1995, 258). Whaley (1998, 303–4) translates ‘wore to shreds dark awnings’. Although she does not question the traditional translation of tjald, she recognises the oddity of these ‘enigmatic’ lines.

Ships and Sailing

165

particular poetic usage by Arnórr, and it is not to deny that, elsewhere, tjald might have the more usual meaning of ‘tent’ or ‘awning’. In ÞjóðA IV,19, King Haraldr sets the departure of his fleet from Trondheim in motion by throwing back the tjald on his ship: slyngr l›ngu tjaldi af sér ‘throws off the long tjald’. Here, the modifying adjective makes clear that this is a proper ship’s cover for the whole length of the ship when in harbour, and sails cannot be involved at this stage, since the ship moves out of harbour under oarpower. Tjald also occurs in the ship-kenning tjalda drasill ‘horse of the tjald’ (Sigv XIII,2). Kennings do not normally help us define words precisely, and the determinant here could be any word to do with ships. It may, however, be significant that the ship in this stanza is called myrkblár ‘dark blue’, a colour associated with sails, as we have seen. The vocabulary of what is normally called ‘standing and running rigging’, i.e. the ropes holding the mast and those for manipulating the sail, is not well represented in the skaldic corpus. Andersen (1997b, 183) makes the point that this distinction between two types of rigging was less relevant in ON, particularly on warships where the mast was regularly raised and lowered. The general term reiði ‘tackle’ (n., pl. reiði) makes an occasional appearance. Ótt II,15 describes Óláfr’s capture of Hákon’s ship með skreyttu reiði ‘with adorned tackle’. Presumably this refers to the careful craftsmanship of some of the wood, bone or horn blocks used to control and tension the rigging ropes (pictured in Andersen 1997b, 186–7, 199; Crumlin-Pedersen 1997b, 133–40). In Arn II,4, the reading reiði depends on an emendation to the text, although a well-founded one (Whaley 1998, 150). Again, it is modified by an adjective, this time gerzkr ‘Russian’. It is not clear whether Russian tackle was somehow special, or whether this is a simple practicality, since Magnús was just coming from there.79 According to the sagas, Magnús was brought back from Russia by some Norwegian noblemen, presumably in the same ship they went out in (called a skeið in this stanza), so it may be that they had to re-rig the ship there before they could return in it. The ‘stay’, or rope that supported the mast, was called stag (n., pl. st›g) in ON, and appears in this meaning in three stanzas (ÞjóðA I,4; Valg 6; Gísl I,15), all clearly referring to a forestay, a rope that went from the top of the mast to the fore-stem. This term has been discussed by Peter Foote (1978, 60–61; 1984, 235). As Foote points out, ‘when the yard was hoisted to the mast-head, then a high wind forced the sail against the stay’, and Þjóðólfr’s stanza, as we have already seen, even suggests the noise that this makes. The characteristic angle of the stay is alluded to in Stefnir 2, in which the poet claims he would rather stand staglútr drifinn úti ‘leaning like a stay, rain-drenched, outdoors [or ‘out at sea’]’ than warm himself in the arms of a Danish girl.80 79 80

In Arn II,9, the warriors on the ship appear to have been provided with gerzkum malmi ‘Russian metal [i.e. weapons]’. Jón Helgason 1968, 44, suggests that staglútr, drifinn is a mistake for staglút drifinn ‘sprinkled with ‘‘stay-lye’’ ’, a kind of liquid used to prepare ships’ ropes. However, this would involve an emendation, and lút f. ‘lye’ is not recorded in the skaldic corpus, whereas

166

Ships and Men in the Late Viking Age

One more rigging word is hidden away in the ship-kenning hlébarðr hanka ‘the bear of the hanki’ (Arn II,3), in which it has been suggested hanki refers to ‘a loop or other device which holds the cordage for the sails in a given position’ (Whaley 1998, 144). Osier rings for the purpose of fastening the mast shrouds, and which ‘pass through a hole in the gunwale strake and then through a cleat of beech fastened on the outside of this strake’ (Olsen and Crumlin-Pedersen 1990, 116) have been found in Wreck 3 at Skuldelev (very clearly illustrated in Andersen 1997b, 194; see fig. 4.9). It is not clear whether hanki was used of the cleat, the ring, or the whole arrangement.

In harbour and on land In the late ninth century, the north Norwegian Ohthere told King Alfred of Wessex of his journeys from his home to the market-town Sciringes heal (probably Kaupang): Þyder he cwæð þæt man ne mihte geseglian on anum monðe, gyf man on niht wicode ‘He said that a man could scarcely sail there in a month, assuming he made camp at night’, while his Anglo-Saxon counterpart Wulfstan travelled from Hedeby to Truso in the course of syfan dagum & nihtum, þæt þæt wæs ealne weg yrnende under segle ‘seven days and nights, the boat running under sail the whole way’ (Lund 1984, 21–2). Depending on circumstances, Viking Age sailors could either stop every night to camp, or sleep on board the moving ship. Clearly they would have to do the latter when setting out across the North Sea or the North Atlantic, rather than sailing along a coastal route. But even ‘camping’ at night might not necessarily imply more than anchoring in a sheltered spot, and sleeping on board ship, perhaps with the sail as a tent, as suggested above. This is what Sigvatr describes in the account of his voyage to Sweden: létum skip skolla tj›lduð við ey ‘we let the tented ships rock by the island’ (Sigv III,10). This stanza clearly refers to the covering of a ship, with either a sail or an awning, and the use of tjald in ÞjóðA IV,19, discussed above, also most likely refers to a ship-cover. There is therefore no skaldic evidence for land-tents of the kind found in the Gokstad and Oseberg ship-burials (Brøgger and Shetelig 1950, 163), perhaps because such equipment, while appropriate for a royal progress, was a luxury that took up too much space on a military or mercantile expedition. According to Falk (AnS, 78), the Scandinavians borrowed both the anchor and its name from the Romans early on (for a brief summary of finds of Viking Age anchors, see Vadstrup 1997b, 136–7). The word akkeri (n., pl. akkeri) occurs in Hharð 9, in which the king describes lying at anchor in Randersfjord to make an attack on Danish warriors who are being sung to by their wives, unawares: lótum

there are at least two other examples of compound adjectives in -lútr describing bodily attitudes (LP, s.v. lútr). There is also a mysterious ship-kenning in HHuI,29 stagstjórnmarr ‘horse of stay-steering’, whatever that might mean.

Ships and Sailing

167

Image not available

4.9 The osier ring from the Skuldelev 3 ship. Photo: Viking Ship Museum, Roskilde.

vér akkeri halda í Goðnarfirði ‘we let the anchor hold (the ship) in Goðnarfj›rðr’. In ÞjóðA IV,11, a stanza that is clearly linked to this (the medieval sources variously attribute each half-stanza or both together to Haraldr harðráði or his skald Þjóðólfr Arnórsson, see Perkins 1982–5, 194), the anchor is poetically called kaldnefr ‘cold-nosed one’, while its fluke is called fleinn ‘arrow-/spear-head’ and one of its arms krókr ‘(fish-)hook’. Especially if these two half-stanzas belong together, we have the poet ringing the changes on different words for the anchor and its parts. Another stanza with a number of anchor-words is ÞjóðA IV,23, describing the king’s ship anchored in a storm, and driven about by the wind, so that this puts a strain on the anchor ropes and the anchor itself (Hkr III, 144):

168

Ships and Men in the Late Viking Age Hléseyjar lemr hóvan hryngarð konungr barði. Neytir þá til þrautar þengill snekkju strengja. Eigi es járni bjúgu indæll skaði lindis. Gnegr af gaddi digrum grjót ok veðr en ljótu. The king strikes the high, resounding enclosure of Læsø [the sea] with the prow. The prince then makes heavy use of the ship’s anchor-ropes [sg. strengr]. The destroyer of the lime-tree [storm]81 is not pleasant to the bent iron [bjúgt járn]. Pebbles and the foul storm gnaw at the sturdy spike [gaddr].

None of these words has the appearance of being a technical term. Strengr could be used of any rope, gaddr of any nail or spike, while bjúgt járn is descriptive. But clearly Þjóðólfr enjoyed describing anchors in as many ways as possible. Another anonymous poet also had fun with anchors. Apparently on the same expedition on which he threatened the husbands of Randers, Haraldr took prisoner the daughters of a Danish chieftain, Þorkell geysa, who had taunted the Norwegians by carving anchors out of some kind of soft cheese (AnonXI Lv,6; Hkr III, 110–11): Skóru jast ór osti eybaugs Dana meyjar, þat of angraði þengil, þing, akkerishringa. Nú sér m›rg í morgin mær, hlær at því færi, ernan krók ór járni allvalds skipum halda. The maidens of the Danes carved sea-things, anchor-rings [akkerishringar] out of yeast-cheese; that distressed the prince. Now, this morning, many a maiden sees an efficient iron hook [krókr ór járni] hold the ruler’s ships; fewer laugh at that.

The story seems ludicrous, but the verse is preserved in a number of kings’ sagas, with Morkinskinna telling the fullest anecdote about how Haraldr used the girls to press a large sum of ransom money out of their father (Msk, 157–8). The anchor-rings could have been the iron rings on the anchor itself to which the anchor-ropes could be tied (AnS, 79), but one Viking Age anchor with a chain is

81

‘Destroyer of the tree’ is a normal kenning for ‘storm’, but the particular choice of tree here may also be intended to allude to the lime-bast which was the basis of much Viking Age rope, such as that found in the Oseberg burial (illustrated in Andersen 1997b, 182; see also Crumlin-Pedersen 1997b, 188).

Ships and Sailing

169

Image not available

4.10 The anchor and chain from the Ladby ship. Photo: National Museum of Denmark.

known from Ladby (Vadstrup 1997b, 138; see fig. 4.10). This formed the first part of the attachment, to help weigh down the anchor and keep it on the bottom, while the rest of the anchor-cable is likely to have been of rope. Crumlin-Pedersen (1997b, 146) makes the point that anchors are ‘the largest examples of forged ironwork’ known from the Viking Age and that, because of their cost, they were used only on the largest ships. It has been suggested that Sö 352 depicts an anchor and rope (SR III, 343) though, if so, it is rather a curious-looking one. It is perhaps more plausibly an anchor-stone, or an anchor ‘made from wood, stones and rope lashings’ (Crumlin-Pedersen 1997b, 146), as used on small ships and boats. This status value of iron anchors would explain the intense interest that the poets had in them. Þjóðólfr used the word strengr (m., pl. strengjar) for the anchor-rope. This word could in fact be used of almost any kind of rope or string including, most commonly, bowstrings (LP). Thus, when it is used in the ship-kenning strengjar jór ‘horse of the strengr’ (Sigv II,6), it may not have been intended very specifically, after all there would have been many ropes on a ship. However, the singular form may suggest an anchor- or mooring-rope rather than the many ropes of the rigging. Similarly, the ship-kenning snœris vitni ‘wolf of the snœri’ (Hfr III,16) may give us a specific term for an anchor-rope (as argued in AnS, 80), or may simply be another general word for ‘rope’. The word snœri otherwise seems to be used mostly of a leather spear-band (LP), and the kenning alludes to the well-known myth in which the gods tied the Fenriswolf up with a

170

Ships and Men in the Late Viking Age

fetter made of impossible things like the beard of a woman and the breath of a fish, and which turned out to be sléttr ok blautr sem silkirœma, en . . . traustr ok sterkr ‘smooth and soft like a silk ribbon, but reliable and strong’ (SnEGylf, 28). Clearly, ropes that were both light and strong were the ideal, but this does not give us any further clues as to its function. Mooring techniques may lie behind the ship-kenning valr krapta ‘horse of the krapti’ (Hár 2). Although the kenning by itself would not reveal this, krapti (m., pl. kraptar) is apparently the word for a bollard, the wooden protuberance on the hull of a ship (or boat) to which the mooring-rope could be attached (AnS, 24). Vadstrup (1997b, 126–7) makes the point that the bollard found on the fore-stem of the Skuldelev 3 ship is not strong enough to hold the moored or anchored ship, and that the rope would also be passed through a hole in the bulkhead. Harbours are mentioned three times in the skaldic corpus (Sigv II,3; Sigv XIII,23; Halli 1), using the common word h›fn (f., pl. hafnar; also a common place-name element), also found in one runic inscription (U 1016) in the compound grikkhafn ‘ ‘‘Greek’’ [Byzantine] harbour’. These could be natural or built-up harbours. A landing-place might also be called a st›ð (f., pl. st›ðvar) or a v›r (f., pl. varar) as in the ship-kennings st›ðvar hrafn ‘raven of the st›ð’ (Hfr I,6) and varar hrafn ‘raven of the v›r’ (ÞjóðA I,12). In both these cases, hrafn might be being used as a horse-name (originally for a raven-black horse), giving a more usual type of kenning. If not anchored, ships would be pulled up onto land on a wooden slipway laid down to make this process smoother. The ships would go back down the same way when being launched. This slipway was called a hlunnr (m., pl. hlunnar), a relatively common word in the skaldic corpus. The singular form actually means ‘roller’, but is usually used in a collective sense of the whole slipway.82 Such rollers might be evidenced by the logs found under the Oseberg ship (Brøgger et al. 1917–28, I, 174–6, pl. 18). Falk (AnS, 28) assumed these rollers were dug down into a ditch which could accommodate the keel, while the broad part of the hull rested on the sand or gravel either side of the slipway (see also discussion of aurborð, above), and it has been argued that the construction method was the same at portage sites (thus a roller-lined V-shaped portage ditch is illustrated in Ambrosiani 1991, 103, though the status and accuracy of this drawing are very uncertain). However, simple rollers seem to have been common – four were found underneath the newly-discovered Roskilde 3 ship (Bill et al. 1998, 146). The state of this slipway affected the ease with which a ship could be landed or launched, and poets comment on this regularly. In ÞKolb III,1 the launchway is dreginn ‘worn from use’. In Arn II,4 the hlunnr is harða stinnr ‘exceedingly stiff’, while in Arn II,11 it is sléttr ‘smooth’ (in this case the ship is said to be hélugr ‘icy’, which explains why it slid easily). In Þfagr 4, the ship-kenning

82

ÞKolb III,1; Arn II,4,11; ÞjóðA IV,18. Very common in ship-kennings: Edáð 8; Hfr II,5; Sigv III,14; Ótt II,20; Ótt III,10; Arn II,17; ÞjóðA IV,9; Þfagr 11.

Ships and Sailing

171

húnferils hreinn ‘reindeer of the way of the mast-top [sea]’ is modified by an appropriate adjective, hlunntamiðr ‘tamed to the slipway’. Sigv III,10 contrasts the ship rocking offshore in summer with the same ship on shore in the autumn. Once landed, a ship might be stored for the winter in a special shed or boat-house called a naust (n., pl. naust), a word which still survives in Norwegian and Icelandic, and in the language of archaeologists, but is recorded only once in our corpus, in a ship-kenning: nausta blakkr ‘horse of the nausts’ (ÞSjár II,1).83 Whether in the naust or outside, a ship would be propped up with a skorða (f., pl. skorður) ‘prop’, which occurs in the ship-kenning skorðu skíði ‘ski of the skorða’ (Arn II,9). The equivalent verb skorða ‘to prop’ is found in Gísl I,8 in a rather obscure context, but which apparently describes merchants propping up their ships (with axes!) on shore at a trading-site.84

The vocabulary of sailing Description and metaphor The discussion of the vocabulary of the ship and its parts has shown that skaldic poets could choose to use technical nautical vocabulary (mainly nominal forms), and that this vocabulary was sometimes used in a very precise way. This poetic precision could be extended, not only to accurate use of technical terminology, but to naturalistic descriptions of sailing, in which the relevant vocabulary is mainly verbal. In the praise poems, with their focus on the deeds and virtues of the leader, the metaphorical tends to take over from the naturalistic in descriptions of the loading and launching of ships, of sailing, rowing and steering them, and of landing and anchoring. Thus, these activities are frequently presented as if the leader alone was responsible for getting the ship from A to B. Nevertheless, it is possible to get some sense of the words used for these activities from the skaldic and even the runic material. This section is not exhaustive, since few of the verbs used by the skalds to describe the launching, sailing and landing of ships are restricted to nautical usage, and it is not always easy to draw a line between nautical and non-nautical instances. Preparing and launching Preparing a ship for a voyage, whether of raiding or trading, would have been a major undertaking, but the skaldic corpus gives hardly any indication of the work involved, and where it does, it contrives to suggest that the leader did it all 83

84

A study of twenty-one western Norwegian boat-houses by Myhre (1997) unfortunately includes only one example from what he calls the ‘Viking Age’ and one from the eleventh century. On beach operations, see McGrail 1987, 267–9.

172

Ships and Men in the Late Viking Age

himself. Thus, Knútr is flattered by being told bjóttu harðbrynjuð skip ‘you made ready hard-armoured ships’ (Ótt III,1), while the Danish king Eiríkr bjó ›ndurt vár veglig flaust austan ór G›rðum ‘made ready early in the spring splendid ships to go west from Russia’ (Mark I,5), both using the common verb búa ‘prepare, equip, make ready’.85 In Korm I,3, the leader is one who bindr hún beinan ‘ties the húnn straight’. More poetically, the workaday preparation of a ship for a voyage was figured as the leader ‘adorning’ or ‘embellishing’ the vessel, using the verb glæsa ‘to make shiny’. The ten-year-old Magnús glæsti herskip ór G›rðum ‘embellished (a) warship(s) (to go) from Russia’ (Arn III,1).86 Arnórr also sums up what is praiseworthy about Magnús’ career: hrósak því es hlenna dolgr glæsir herskip ‘I praise the fact that the enemy of thieves embellishes warships’ (Arn II,17). Similarly, Haraldr skreytir ›nnur unnvigg sunnan ‘adorns other wave-horses from the south’ (ÞjóðA, IV,9; see also B›lv 8). The enemy does the same: Sveinn’s ships are glæst (ÞjóðA III,16), although this does not stop Haraldr from clearing seventy of them in one instant (!). Occasionally the skalds give us glimpses of the crew loading the ship, usually with their armour and weapons. Returning from Russia, Magnús’ fim hirð bar hervæðr til h›mlu ‘athletic troop carried war-garb to the oar-loop [rowing-position]’ (Arn III,2). Again, there is a parallel with Hfr II,1, in which the leader’s men hlóðu hlýrvigg Hamðis klæðum ‘loaded the bow-horse [ship] with Hamðir’s clothes [armour]’. In a stanza describing the young Óláfr Haraldsson setting out for Sweden with warlike intent, Óttarr includes the king in this activity: bóruð á skip randir ‘you [pl.] carried shields onto the ship(s)’ (Ótt II,4).87 Just as the war-leader seems to have prepared his ships single-handed, so he is also presented as having physically launched the ships on his own, using a variety of appropriate verbs. Óttarr praised both Óláfr and Knútr to their faces for having launched ships when young: ungr hratztu vengis dreyra blakki á vit Danmarkar ‘(you were) young (when) you propelled the steed of the lifeblood of the plain [water→ship] to go to Denmark’ (Ótt II,3) and hratztu lítt gamall fram skeiðum ‘you propelled ships when not very old’ (Ótt III,1). The ten-year-old Magnús is presented as having given his ship(s) the decisive shove: vannt geyst herskip af harða stinnum hlunni ‘you caused the warship(s) to move swiftly down the very stiff slipway’ (Arn II,4). Recalling the same event, Þjóðólfr uses

85

86

87

An ambiguous instance is Bersi I,3, in which the poet, who is changing his allegiance from Sveinn Hákonarson to Óláfr Haraldsson, notes that búum í ári ólítinn Áta ›ndur til handa þér ‘this year we [probably ‘I’] (will) make ready a large ski of Áti [ship] for you’. The same expression is used of a twelve-year-old hero in a stanza of obscure origin that Finnur Jónsson wrongly attributed to Hallfreðr’s Óláfsdrápa (Hfr II,1), and in which this statement may have been scribally ‘borrowed’ from Arn III,1; the stanza is discussed in Skjd A I, 156, Fidjestøl 1982, 106–7, and Whaley 1998, 184. Both these parallels suggest to me that the correct translation of bera in Arnórr’s stanza is ‘carried’, rather than ‘wore’ as Whaley (1998, 185) has it.

Ships and Sailing

173

the verb skjóta ‘shoot’, réttu skjóta snekkju út ‘you shot the snekkja out to sea’ (ÞjóðA I,2).88 This verb is used in a number of launching scenes: Magnús’ uncle Haraldr hlaut skjóta stafni til Hallands ‘shot his prow (i.e. ship) towards Halland’ (Arn VI,2) and is celebrated as sás skaut nýtla herskips borði ‘he who skilfully shot the hull of the warship’ (Arn VI,16); he says of himself that hlautk skjóta hlýri á flœði ‘I had to shoot the bows [ship] onto the water’ (Hharð 16); returning to Norway from England in 1066, Haraldr’s son Óláfr kyrri skaut stafni þars heitir Hrafnseyrr ‘shot the prow out at the place called Ravenser’ (Steinn III,5);89 and Eiríkr skaut hlýrum ‘shot the bows’ in the east, on his way home to Denmark (Mark I,5). Sometimes pulling was required, rather than pushing, as when Arnórr reminds Magnús that drótt hélug borð af sléttu hlunni ‘you dragged icy hulls from the smooth slipway’ (Arn II,11) at the start of his Wendish campaign. In Arn II,9 the poet says that Magnús ýtti flota miklum suðr með láði ‘shoved a great fleet southwards along the coast’. Although Finnur Jónsson translated this as ‘styrede den store flåde’ (‘steered the large fleet’, Skjd A I, 308), the verb clearly refers specifically to the act of launching, as the poet goes on to describe how the ship was auðit skriðar ‘granted gliding’ (see also a further example of ýta, below). In kennings, the leader can be called a flausta ›rþeysir ‘swift-impeller of ships’ (Eskál III,3) or a hleypimeiðr hlunnviggja ‘tree [man] that causes roller-horses [ships] to run’ (Hfr II,5). Occasionally, ships are launched without the poet making clear who did the launching, again using some of the verbs discussed above, and thus still implying the physical activity involved: þryngva ‘press’ (ÞKolb III,1); hrinda ‘propel’ (ÞjóðA IV,18); ýta ‘shove’ (ÞjóðA IV,18). In Þjóðólfr’s celebrated stanzas, Haraldr gives the signal for the departure of the fleet from Niðaróss, once the ships have actually been launched into the water, by throwing back the ship’s cover: lið-Baldr slyngr l›ngu tjaldi af sér ‘the troop-lord throws off the long tjald’ (ÞjóðA IV,19, discussed above). I have not found any other such vignette in the skaldic corpus, though there is a similar one in the description of the launching of Helgi’s fleet in HHuI,26: brá stýrir stafntioldom af ‘the captain smartly pulled the stem-covers off’, followed by the raising of the sail onto the mast. The ship in the sea The skaldic corpus is rich in words describing the different aspects of the ship’s motion through the water and the captain’s and crew’s actions which influence that motion. The basic verbs sigla ‘sail’ and stýra ‘steer’ are also recorded in the runic corpus (sigla in Sö 198, stýra in U 439, U 654, U 778, U 1016). Both verbs have a ship-word as their object (in the dative), either kn›rr (Sö 198, U 654, U

88 89

The auxiliary verb ráða (réttu < réð-þú) is more or less meaningless in this kind of context, the meaning resides in the infinitive form skjóta. On this place-name, see Smith 1970, 16, 19.

174

Ships and Men in the Late Viking Age

1016) or skip (U 439; implied in U 778 as the word occurs earlier in the inscription). In U 654, the verb stýra is modified by the adverb vel ‘well’, praising the deceased for his sailing abilities. While sigla does not appear in the skaldic corpus, stýra does. The verb is ambiguous: its basic meaning is the action of holding the tiller and directing the course of the ship, but it can also refer in a more general way to the war-leader’s command of his fleet. Thus, when Sveinn Úlfsson stýrir f›grum skipum ‘steers beautiful ships’ (Þfagr 5), he cannot literally be steering all of them, and the more general meaning is intended. Where the object of the verb is singular, however, the focus tends to be on the individual’s seafaring capabilities (Hfr III,18; J›k 1). In Hfr III,13, the poet, lamenting the loss of the crew of Ormrinn langi, declares that it will be long before the ship again sees such drengir, þótt alldýrr konungr stýri ‘even if a very glorious king should be steering it’. Here, the emphasis is more on command, with the peerless captain deserving a peerless crew. In Valg 5, the verb has no explicit object, but since the poet says to Haraldr, of his stormy journey to Sigtuna, stýrðu hvatt í h›rðu glyggvi ‘you steered boldly in a hard storm’ (Valg 5) it is more likely that he means Haraldr’s way with the tiller. While the stanza refers to ‘ships’ in the plural, and Haraldr could hardly have steered all of them, there can have been no real distinction between ‘commanding’ and ‘steering’ in this context. The commander would have taken his turn at the tiller, and could only have had authority as a commander if he was a capable steersman. In ÞjóðA IV,19, Haraldr is shown at the tiller of his brand-new ship (réð at stýra næsta nýri skeið ‘steered his ‘‘nearest to new’’ skeið’) while his crew man the oars. There are just one or two technical sailing terms in the corpus. Beita ‘to beat, sail to windward’ occurs in a stanza in which the poet tells Óláfr Haraldsson that, returning from England, beittuð miðjan Nóreg ‘you reached the middle of Norway by beating’ (Ótt II,14). The meaning of the verb is made clear both in this stanza and others in the poem (esp. 13) which emphasise how hard the journey was for the ships (Fidjestøl 1992, 124). In another stanza which emphasises the physical hardships of sailing, Magnús is shown as a brave sailor who vafðir lítt (Arn II,4). In this context, the verb vefja is interpreted as a nautical idiom by Foote (1978, 63), meaning ‘take in sail, reef’, which developed from the basic meaning ‘wrap, roll’. Others (most recently Whaley 1998, 151) interpret it as ‘hesitate, waver’, a meaning which, as Foote points out, ‘seems to have little to commend it save as an extension of sailor’s idiom’. Whaley (1998, 151) notes that the nautical meaning ‘is contextually plausible but not . . . supported by usage elsewhere’, but the contextual evidence is strong, as the verb occurs in the stanza between references to reiði ‘tackle’ and vendir ‘masts’, making a connection with ‘sail’ (which could also be called by the related word vefr, see above) almost inevitable. A nautical meaning is certainly more effective in this particular stanza than a general praise of Magnús for being resolute. Once the captain has pointed his fore-stem in the right direction (vestr léztu vísat framstafni ‘you caused the prow to be directed west’, Hallv 2), the usual verb for keeping a ship on a particular course was halda ‘hold’, followed by an

Ships and Sailing

175

adverb of direction (norðan ‘from the north [south]’, Arn II,10, ÞjóðA IV,9; þangat ‘thither’, Mark I,16). This is the basis of the seafarer-kenning hald-Viðurr haffaxa ‘holding-god of the sea-horse [ship]’ (Eskál III,11). As well as the standard verbs of motion ganga ‘go’ (ÞjóðA III,8) and fara ‘travel’ (Valg 10), the stasis or progress of the ship can be indicated by descriptive verbs such as fljóta ‘float’ (Hfr III,16; Sigv XIII,23; Arn VI,4; see also HHuI,31), þyrja ‘race’ (Sigv X,4; also used of shipborne men in Eskál III,23), hníga ‘pitch’ (Valg 5) and skríða ‘glide’ (Hfr III,13; ÞKolb III,4; Sigv X,7; Arn II,16; Arn VI,2; Hharð 4; B›lv 2; see also HHuI,23). Adjectives describing the motion and angle of a ship are sælútr ‘sea-leaning, heeling’ (Arn VII,2) and hallr ‘leaning’, so Haraldr’s sodden ship is h›ll á hléborð ‘leaning to leeward’ (ÞjóðA III,8). Ships could also líða, which has a basic ON meaning of ‘glide (by land, water or air), pass’. In OE, the usage of the cognate lidan means ‘travel by sea [of a person]’ and it has been suggested (Kuhn 1991, 70) that this restriction of meaning became fashionable in Scandinavian poetry around 1025. Hofmann (1955, 92) is more sceptical and certainly the evidence is less than clear. There are eight instances of this verb in the corpus. The clearest example referring to a person travelling by sea is frægr hildar leiptra v›rðr leið of ægi ‘the renowned guardian of the flash of battle [sword→warrior] sailed across the sea’ (Hallv 3), describing Knútr’s journey to England. The other two examples of this meaning cited by Hofmann are less convincing. The stanza AnonX III,C,3 erum liðnir á leið frá láði ‘we have passed on the road from the land’ is quoted in Snorri’s Edda to demonstrate that the ‘sea’ can be called a ‘road’ (SnESksm, 93), so it is a moot point whether the verb here actually means ‘to travel by sea’, or is merely carrying through the metaphor. Þloft II,5 does indeed show the use of the verb líða of men (griðfastir friðmenn liðu), but it is flanked by two stanzas in which the same verb is used of ships (kolsvartir viðir liðu ‘coal-black timbers liðu’; svalheims valar liðu ‘horses of the cool world [sea] liðu’). In the latter, the base-word in the ship-kenning is ‘horse’, for which the native meaning of líða would also be appropriate, so that Þórarinn’s poem shows all the possible collocations for the verb: men, ships and horses. Similarly, Hallv 1 has a ship-kenning with an animal base-word (sunds dýr ‘animals of the strait’) as the subject of líða, while Sigv X,8 has a straightforward ship-word (kilir) as its subject. In Bersi I,1, the subject of the verb líða is the poet himself (in an accusative-and-infinitive construction), but there is no indication in the stanza that his travelling is to be by sea (although there is a reference to the preparation of a ship in st. 3), indeed the statement is rather general and equivalent to ‘farewell’: baðtu þenna hróðrs hagkennanda líða heilan ‘you [King Óláfr] bid this knowledgeable practitioner of poetry líða safely’. Its use in these ambiguous contexts suggests that líða was an all-purpose word for ‘travel’, and this would even cover its use in the examples in which a person travels by sea. In this instance, the influence of OE is probably a chimera. Just as the king is shown single-handedly preparing and launching his ship, so he can also be shown rowing it. An anonymous poet tells us that Óleifr knýr

176

Ships and Men in the Late Viking Age

Visund und órum ‘Óláfr presses Visund under the oars’ (AnonXI Lv,4), although it is unlikely that the king did much rowing, and certainly not by himself. But it is still the captain (and rarely his crew) that causes the ship’s motion: þeir báðir kníðu bló borð ‘they both thrust the blue ships (forward)’ (Sigv II,1), létum skeiðr eisa ‘we made the skeiðr rush’ (Sigv III,9), léztu skip dynja ‘you made the ships thunder’ (Hallv 3; similarly ÞKolb III,2), lét dreka skolla ‘(Haraldr) caused the dreki to rock’ (ÞjóðA III,12), although sometimes the ship itself is active: flaust hratt af sér br›ttum forsi ‘the ship thrust away the steep billow’ (Ótt II,14); hlýr en stinnu brutu hr›nn ‘the sturdy bows broke the wave’ (Þham I,2). The sea is kaldr ‘cold’ (Valg 11), strangr ‘strong’ (Ótt II,13; ÞjóðA IV,21), þungr ‘heavy’ (B›lv 5), sollinn ‘swollen’ (Valg 11) and skelfðr ‘turbulent’ (Arn II,16; ÞjóðA I,1). When not blár ‘blue’ (Gísl I,16), it is grár ‘grey’ (Sigv X,5; Þfagr 11) and howls (þjóta, Ótt I,5; ÞjóðA I,16; Þfagr 11). It is a mighty force around the ship (bára hristi búnar grímur ‘the bore shook the adorned masks [figureheads]’, Arn II,4; grœðir hristi hélug b›rð ‘the sea shook the icy stems’, Mark I,16) and a constant challenge to the seafarer: eigi hræddusk ægi, ér fóruð sjá stóran ‘you were not afraid of the main, you travelled the great sea’ (Ótt II,14); eigi þraut ofvægjan gram bægja við ægi ‘the formidable prince never ceased contending with the sea’ (Arn V,13). The force of a wave can be felt by the oarsmen (en ljóta bára hnikaði ór ‘the ugly breaker caught the oar[s]’, Arn II,3) and the king at the tiller (braut stóran straum of stýri konungs ‘the great stream broke over the king’s rudder’, Steinn III,6). Arnórr describes the turbulence of the sea as Magnús sails to Denmark: ljótu lauðri dreif útan á lypting ‘the ugly surf drove onto the outside of the lypting’ and álar bifðusk fyri ‘the sea-channels shook’ (Arn II,10). The very act of sailing disturbs the water: dúfu braut und húfi ‘the wave broke under the hull’ (Valg 10), and the foam is sent up over the ship: bára berr bjart lauðr of við rauðan ‘the billow carries bright surf over the red timber’ (Þfagr 11). Waves can be sylghór ‘high enough to swallow [the ship]’ (Ótt II,20) or ›rðigr ‘steep’ (ÞjóðA III,8) and, when they overwhelm the ship, bailing is needed: austr’s til hár, býðk Gíparði reiða byttu ‘the bilgewater is too high, I ask Gíparðr to pass the bailer’ (Eldj 1). Objects identified as bailers, and which might thus have been called bytta (f., pl. byttur) have been found in the Oseberg burial and at Hedeby (Crumlin-Pedersen 1997b, 144; see also Christiansen 1985, 205). Both the verb ausa ‘bail’ and the noun austr ‘bilge-water’ occur in Eskál III,3 in a complex sailing metaphor, extending over several stanzas, which actually denotes the delivery of poetry (Fidjestøl 1982, 224). The captain’s mastery of the sea is often described in metaphors. The common kenning-type ‘horse of the sea’ for ship inevitably leads poets to use verbs appropriate to horses to describe the leader’s sailing, even when such a kenning is not used in the immediate context. In Ótt II,4, the poet tells Óláfr Haraldsson that ›ttuð austr með flaustum ‘you spurred your ships east’, with a verb (etja) normally used of driving horses on. The metaphor is used more consistently in another stanza in which Eiríkr Hákonarson sails to meet an opponent; he is jarl sás atti hrefnis stóði á l›g ‘the jarl who spurred the stud of the

Ships and Sailing

177

hrefni [fleet] onto the sea’ (ÞKolb III,4). There may also be a horse-taming metaphor behind venr snekkjur sælútar úti ‘(he) accustoms the sea-leaning snekkjur to be out (i.e. at sea)’ (Arn VII,2). Unique to the runic corpus is the metaphor of ‘ploughing’ the sea with the stem of the ship (arþi barþi ‘ploughed with a barð’, Sö 65), although the comparison is obvious enough and well-known from other cultures. The ship throws up a wave on either side just as the plough throws up the soil, and the wake of the ship is like a furrow (Lewis 1994, 81–3). More originally, poets use verbs of cutting, slicing and inscribing to describe the captain’s progress through the water with his ship. The metaphor is particularly appropriate when the rudder is the object of the verb, as in skóruð bylgjur stýri ‘you scored waves with the rudder’ (Ótt II,20). More usually, it is with a larger part of the ship that the king parts the waves: hraustr þjóðkonungr skar salt héltum húfi ‘the bold king of the people scored the salt with icy hull’ (Arn III,2); léztu vatn slitna of þunnri sk›r ‘you caused the water to tear asunder around the thin planking’ (ÞjóðA III,9); reistu gœði glæstum gjalfrstóðum ‘you carved the sea with adorned sea-horses [ships]’ (B›lv 8, rather mixing his metaphors); allvaldr réð rísta haf austr stáli ‘the ruler carved the sea in the east with the prow’ (BjH 3); atseti Hleiðrar réð rísta þangs láð sunda m›rum ‘the ruler of Lejre carved the land of seaweed [sea] with strait-horses [ships]’ (Steinn I,2, again a splendidly mixed metaphor). The verb rísta ‘carve’ also appears in Valg 11, although without an instrumental ship or ship-part. This verb is one which occurs frequently in runic inscriptions, describing the action of inscribing the runes,90 and this metaphor of carving the sea implies the same kind of knowledge and mastery in the sea-captain as in the rune-carver. But it should be noted that the ship can also be described as cutting or slicing the waves, without any human agency: súð sneið ‘the hull sliced’ (Hharð 4); eikikj›lr reist ›rðigt vatn ‘the oaken keel carved the steep water’ (ÞjóðA III,8); flaust klufu flóð ‘ships split the stream’ (ÞjóðA III,9). In Sigv II,1 the reconstructed reading sund sk›ru can be understood by taking a subject from the first half of the stanza, either ‘the ships scored the sound’ (Fsk, 175), or the king (Óláfr) and the jarl (Sveinn). The latter seems preferable, since the two leaders are the subjects of all the verbs in the first half of the stanza. The nine skills in which the twelfth-century earl of Orkney, R›gnvaldr (who was born in Norway, the home of skiing), claimed expertise included rowing, knowledge of both books and runes, and gliding on skis (Rv 1). Thus, skiing, like rune-carving, could also be a metaphor for sailing. Arnórr, in a stanza on the Norwegian king Magnús, figures the ship as a ‘ski of the sea’, and its captain as an accomplished skier, using an appropriate verb and an appropriate sea-kenning: mildingr rennir sævar skíði Meita hlíðar ‘the ruler runs the ski of the sea [ship] across the slopes of Meiti [a sea-king→sea]’ (Arn II,18). Similarly,

90

The instances are too numerous to document here, but see e.g. SRR s.v. r†sta.

178

Ships and Men in the Late Viking Age

Haraldr renndi byrskíðum ‘Haraldr ran with wind-skis’ (Þfagr 6), an appropriate image of a Norwegian king sailing, and descending on the non-skiing nation of Denmark. Shipwreck and landing Not all sea-voyages reached their desired end, and runic memorials, with their factual approach, sometimes record dro