Land, Sea and Home: Settlement in the Viking Period 9781905981854, 9781904350255

Table of contents :
Cover
Half Title
Title Page
Copyright Page
Table of Contents
Introduction
At Home in the Viking Period
Scandinavia and Northern Europe
Danish Coastal Landing Places and their Relation to Navigation and Trade
Viking-age Proto-Urban Centres and their Hinterlands: Some Examples from the Baltic Area
Viking-period Pre-Urban Settlements in Russia and Finds of Artefacts of Scandinavian Character
Rural Settlement and Landscape Transformations in Northern Russia, A.D. 900–1300
Early Medieval Coinage and Urban Development: A Norwegian Experience
Two Viking Hoards from the Former Island of Wieringen (The Netherlands): Viking Relations with Frisia in an Archaeological Perspective
Goldsmiths' Tools at Hedeby
The Atlantic Provinces
Settlement and Acculturation in the Irish Sea Region
Viking-age Settlement in Wales and the Evidence from Llanbedrgoch
Social and Economic Integration in Viking-age Ireland: The Evidence of the Hoards
Royal Fleets in Viking Ireland: The Evidence of Lebor na Cert, A.D. 1050–1150
Beyond War or Peace: The Study of Culture Contact in Viking-age Scotland
Spatial Analysis and Cultural Indicators: Viking Settlers at Old Scatness Broch, Shetiand?
Cille Pheadair: The Life and Times of a Norse-period Farmstead c. 1000–1300
A Find of Ringerike Art from Bornais in the Outer Hebrides
Narrative Functions of Landscape in the Old Icelandic Family Sagas
England
Jórvík: A Viking-age City
Viking-age Settlement in the North-western Countryside: Lifting the Veil?
Sedgeford: Excavations of a Rural Settlement in Norfolk
Simy Folds: Twenty Years On
Post-Roman Upland Architecture in the Craven Dales and the Dating Evidence
Timber Buildings without Earth-fast Footings in Viking-age Britain
A Push into the Margins? The Development of a Coastal Landscape in North-West Somerset During the Late Ist Millennium A.D.
Place-names and the History of Scandinavian Settlement in England
Law and Landscape
Changing Weaving Styles and Fabric Types: The Scandinavian Influence
Saxon Shoes, Viking Sheaths? Cultural Identity in Anglo-Scandinavian York
Index

Citation preview

THE SOCIETY FOR

MEDIEVAL ARCHAEOLOGY

MONOGRAPH 20 Series Editor CHRISTOPHER GERRARD

LAND, SEA AND HOME

Settlement in the Viking Period

THE SOCIETY FOR MEDIEVAL ARCHAEOLOGY

MONOGRAPHS

ISSN 0583-9106 I

2 3

G Bersu and D M Wilson (1966) Three Viking Graves in the Isle of Man F W B Charles (1967) Medieval Cruck-building and its Derivatives P A Rahtz (1969) Excavations at King John's Hunting Lodge, Writtle, Essex, 1955-57

4 5 6 7

8

9 10 II

12

13

14 IS 16 17 18 19 20 21 22 23 24

A L Meaney and S C Hawkes (1970) Two Anglo-Saxon Cemeteries at Win­ nail, Winchester, Hampshire HE J Le Patourel (1973) The Moated Sites of Yorkshire G T M Beresford (1975) The Medieval Clay-land Village: Excavations at Goltho and Barton Blount H Clarke and A Carter (1977) Excavations in King's Lynn, I963-I970 J G Hurst (general ed) Wharram. A Study of Settlement on the Yorkshire Wolds, vol I: D D Andrews and G Milne (eds) (1979) Domestic Settlement 1: Areas IO and 6 C M Mahany, A Burchard and G Simpson (1982) Excavations at Stamford, Lincolnshire, I963-69 P Mayes and K Scott (1984) Pottery Kilns at Chilvers Coton, Nuneaton J G Hurst and P A Rahtz (general eds) Wharram. A Study of Settlement on the Yorkshire Wolds, vol III: R D Bell, M W Beresford and others (1987) The Church of St Martin D Austin (1989) The Deserted Medieval Village of Thrislington, Co Durham: Excavations I97 3- I97 4 V L Yanin, E N Nosov, A S Khoroshev, A N Sorokin, E A Rybina, V L Povetkin and P G Gaidukov (1992) The Archaeology of Nov go rod, Russia K Parfitt and B Brugmann (1997) The Anglo-Saxon Cemetery on Mill Hill, Deal, Kent D Gaimster and P Stamper (eds) (1997) The Age of Transition: the Archae­ ology of English Culture 1400-1600 D A Hinton (2000) A Smith in Lindsey: the Anglo-Saxon Grave at Tattershall Thorpe, Lincolnshire S Lucy and A Reynolds (eds) (2002) Burial in Early Medieval England and Wales S T Driscoll (2002) Excavations at Glasgow Cathedral I988-I997 Philip Mayes (2002) Excavations at a Templar Preceptory: South Witham, Lincolnshire I965-67 J Hines, A Lane and M Redknap (eds) (2004) Land, Sea and Home: Settlement in the Viking Period W D Klemperer and N Boothroyd (2004) Excavations at Hulton Abbey, Staffordshire I987-I994 K Giles and C Dyer (eds) (2005) Town and Country in the Middle Ages: Contrasts, Contacts andInterconnections, IIOO-I500 S M Foster and M Cross (eds) (2005) Able Minds and Practised Hands: Scotland's Early Medieval Sculpture in the 2ISt Century A Saunders (2006) Excavations at Launceston Castle, Cornwall

LAND, SEA AND HOME

Settlement in the Viking Period

Edited by JOHN HINES, ALAN LANE

and MARK REDKNAP

Proceedings of a Conference on Viking-period Settlement, at Cardiff, July 200I

200

4

First published 2004 by Maney Publishing

2 Park Square, Milton Park, Abingdon, Oxfordshire OX14 4RN 52 VanderbiItAvenue, New York, NY 10017

Routledge is an imprint of the Taylor & Francis Group, an informa business First issued in paperback 2019 Copyright © 2004 The Society for Medieval Archaeology and Authors All rights reserved. No part of this book may be reprinted or reproduced or utilised

in any form or by any electronic, mechanical, or other means, now known or hereafter

invented, including photocopying and recording, or in any information storage or

retrieval system, without permission in writing from the publishers.

Notice:

Product or corporate names may be trademarks or registered trademarks, and are used

only for identification and explanation without intent to infringe.

Statements in the volume reflect the views of the authors, and not necessarily those

of the Society, editors or publisher.

ISBN 13: 978-1-905981-85-4 (pbk) ISBN 13: 978-1-904350-25-5 (hbk) ISSN 0583-9106

FRONT COVER

The Cliff at Penarth, Evening, Low Tide. By Alfred Sisley (r897). (© National Museum of Wales) BACK COVER

Finds from the Minino I dwelling site and Minino II burial site, c. 950-1250 (see Makarov, p. 68)

CONTENTS

PAGE

INTRODUCTION At Home in the Viking Period By JOHN HINES

I

SCANDINAVIA AND NORTHERN EUROPE Danish Coastal Landing Places and their Relation to Navigation and Trade By JENS ULRIKSEN

7

Viking-age Proto-Urban Centres and their Hinterlands: Some Examples from the Baltic Area By MICHAEL MULLER-WILLE and ASTRID TUMMUSCHEIT .

27

Viking-period Pre-Urban Settlements in Russia and Finds of Artefacts of Scandinavian Character By TAMARA PUSHKINA

41

Rural Settlement and Landscape Transformations in Northern Russia, A.D. 900-1300 By NIKOLAI MAKAROV .

55

Early Medieval Coinage and Urban Development: A Norwegian Experience By JON ANDERS RISVAAG and AXEL CHRISTOPHERSEN

75

Two Viking Hoards from the Former Island of Wieringen (The Netherlands): Viking Relations with Frisia in an Archaeological Perspective By JAN BESTEMAN .

93

Goldsmiths' Tools at Hedeby By BARBARA ARMBRUSTER

109

THE ATLANTIC PROVINCES Settlement and Acculturation in the Irish Sea Region By DAVID GRIFFITHS

125

Viking-age Settlement in Wales and the Evidence from Llanbedrgoch By MARK REDKNAP

139

Social and Economic Integration in Viking-age Ireland : The Evidence of the Hoards By JOHN SHEEHAN.

177

Royal Fleets in Viking Ireland: The Evidence of Lebor na Cert, A.D. I050-II50 By CATHERINE SWIFT

189

Beyond War or Peace: The Study of Culture Contact in Viking-age Scotland By JAMES H. BARRETT .

207

LAND, SEA AND HOME

VI

Spatial Analysis and Cultural Indicators: Viking Settlers at Old Scatness Broch,Shetiand? By AMANDA K. FORSTER, Jo THOMAS and STEVEN J. DOCKRILL

219

Cille Pheadair: The Life and Times of a Norse-period Farmstead c. 1000-1300 By MIKE PARKER PEARSON, HELEN SMITH, JACQUI MULVILLE and MARK BRENNAND

235

A Find of Ringerike Art from Bornais in the Outer Hebrides By NIALL SHARPLES

255

Narrative Functions of Landscape in the Old Icelandic Family Sagas By IAN WYATT

273

ENGLAND

J6rvik: A Viking-age City By RICHARD A. HALL

283

Viking-age Settlement in the North-western Countryside: Lifting the Veil? By NICHOLAS HIGHAM .

297

Sedgeford: Excavations of a Rural Settlement in Norfolk By SOPHIE CABOT, GARETH DAVIES and RIK HOGGETT

313

Simy Folds: Twenty Years On By DENIS COGGINSt

325

Post-Roman Upland Architecture in the Craven Dales and the Dating Evidence By ALAN KING

335

Timber Buildings without Earth-fast Footings in Viking-age Britain By MARK GARDINER

345

A Push into the Margins? The Development of a Coastal Landscape in North­ West Somerset During the Late 1st Millennium A.D. By STEPHEN RIPPON

359

Place-names and the History of Scandinavian Settlement in England By LESLEY ABRAMS and DAVID N. PARSONS

379

Law and Landscape By ANNETTE HOFF

433

Changing Weaving Styles and Fabric Types: The Scandinavian Influence By PHILIPPA A. HENRY .

443

Saxon Shoes, Viking Sheaths? Cultural Identity in Anglo-Scandinavian York By ESTHER CAMERON and QUITA MOULD

457

INDEX.

467

AT HOME IN THE VIKING PERIOD

By

JOHN HINES

The study of human settlement through archaeological and historical sources forms a broad and diverse field. It may reassure us that this field is nevertheless fundamentally coherent to think of it as an area of study to which the concept of 'home' is central. This is not least the case in respect of the Viking Period, a historical phase defined by the forceful and often violent eruptions from Scandinavia that led to new occupation in lands not previously settled by Europeans across the North Atlantic, as well as the establishment of colonies of Scandinavian cultural character in Britain and Ireland, northern France, around the Baltic Sea, and deep into what is now Russia. At a theoretical level, one might query the denial of any sort of 'settlement' history to nomadic peoples, while indeed relationships between domestic­ ated and agrarian societies and communities of nomadic herdsmen and trappers are a major aspect of life as a whole in the Arctic areas of this period. Those whom we call 'Vikings', however, inherited a culture long based in the settled, farming culture of Europe, and in this context we can maintain our focus on the concept of home with confidence. The word 'home' itself was rich in meaning in Viking-period Scandinavia. A number of grammatical variants of the word cover an informative semantic range. ':-haimaz can be reconstructed as a Common Germanic, masculine noun, which underlies Modern English home (Old English ham), German Heim, Danish hjem, etc. Old Norse had two principal variants of this noun, a neuter form heima, with the sense of a principal place of residence (and the source of the modern Scandinavian words for 'home'), and the masculine form heimr, which referred to a larger-scale area that one belonged to. Mythologically, the latter is regularly used for the imagined territories of different classes of beings: thus the Jotunheim mountains of Norway have a name that can be etymologized and translated as 'Giant-Realm'. Like the English element -ham, this form appears often to have been used in place names to denote some sort of settlement focus or nucleus, be that farm, hamlet or village. But it is also used of our world, the Earth - perhaps most famously in the title of the Icelander Snorri Sturluson's I 3th-century historical work concerned primarily with Viking-period and later Norwegian kings, Heimskringla: 'the circle of the world'. Idiomatically, the term heimr even appears in expressions where it represents human life. heimili is a further neuter variant, of some antiquity, with the geographically narrower sense of a homestead, implicitly in a particular building or building-complex. A further layer of connotations resides in the adverbial uses of Old Norse heim and heima. These are found frequently in expressions where the sense and force of the

LAND, SEA AND HOME

2

word is similar to that in the English idioms to strike home, to drive a point home. The term thus introduces the sense of some very specific target: the degree of intimacy and fitness that is implied by some person, thing or concept being judged to be absolutely in the place it is supposed to be in, or in the state one wants it to be in. The concept of 'home' could thus suffuse the whole range of perception in the explicit thought of Old Norse speakers of the Viking Period, from strictly local and specific horizons to the global or even universal. It may risk appearing irritatingly glib to glide from this review of semantic diversity to the equally variable senses of the term 'settlement' in modern academic discourse, but the connexion is not an entirely specious one. It is certainly true that using this term in the English-language title of the conference from which this present set of papers derives 1 gave a degree of imprecision and even vagueness to the scope of the meeting. This was particularly brought home to the local organizers by the need to provide Welsh-language versions of information releases on the conference, where choices had to be made between anheddiad, implying the continuing occupation and settlement of land, cyfanheddiad, implying the settlement of virgin territory, and gwladychiad, signifying the intrusive colonization of already settled land. The ability to avoid being specific by hiding behind an ambiguity is not one enjoyed only by English. In German, the terms Siedlung and Besiedlung encode the difference c1early­ the former is a static object, 'a settlement', the latter the process of settlement - but in Scandinavian the terms bebyggelse and bosetninglbostEttelse perform the same veiling trick. It takes little reflection to realize that for any non-nomadic society, unless one restricts 'settlement archaeology' strictly to a study of the material morphology of settlement sites, there is very little in the cultural life of the peoples concerned that cannot properly be associated with the study of settlement. Of course, the study of human residential practices is a fundamental element of historical settlement studies. The technological capacity of populations to construct their built environment, and the constraints that cultural tradition may have placed upon their application of those abilities, are key parameters to be identified. It is also a familiar point that the interior organization of space, and exterior relationships between houses, are potentially sensitive indicators of social relations, and indeed of the ideologies that are usually inseparable from those relations. We can hardly, however, claim that this has proved a particularly productive seam to mine in Viking-period archaeological research. 2 This is partly attributable to the relative paucity of extensively excavated Viking­ period settlement sites, and partly to the sparsity of interior definition of spaces and activity zones within buildings. But it must also to some extent be recognized as an area in which firmly targeted research is relatively underdeveloped. Another valid way in which to approach the definition of settlement archaeology is to consider what the alternatives to it are - what fields of archaeology are there that we characterize in an equivalent way but see as distinct from settlement 'Viking-period Settlement in Britain and Ireland', Cardiff, 4-7 July 200I. Perhaps the best modern study to cite as a positive example is L. Bender Jorgensen and Palle Eriksen, Trabjerg: En Vestjysk Landsby fra Vikingetiden (Aarhus, 1995)' 1

2

AT HOME IN THE VIKING PERIOD

3

archaeology? The immediate and obvious answer in the case of the Viking Period is burial archaeology - or perhaps, more subtly and appropriately, the archaeology of ritual and religion. Superficially, this distinction might be characterized as the contrastive study of the living and the dead; of the real, mundane and insistently pragmatic world in contrast to the imagined world of spirits and deities. But all such distinctions are false. Coping with death and maintaining religious observances were part of the practical life of normal communities. No one can dispute that Viking-style burial evidence is a vital component in the study of the Scandinavian settlement of the Scottish isles, Ireland and the Isle of Man; or that in many cases we are far more likely to encounter utilitarian tools - for instance, for carpentry, metalwork or agricul­ ture - in grave assemblages or in what may well have been votive hoards than as part of the retrievable inventory of an ordinary settlement site. The conclusion must be, then, that settlement history and archaeology is not a field of study that is defined by its boundaries. It is not a subject area that includes some things and excludes others. It is defined rather by its focus: by those concerns from within the whole conspectus of cultural history that it especially elevates above others. On that famous spectrum of human behaviour from the mechanical/utilitarian to the esoteric and capricious which Christopher Hawkes transformed into his scale of reliable inference,3 settlement archaeology is most concerned with the practical, technical and economic end of things. The development of sites specializing in particular forms of production and exchange was well underway both in Scandinavia and most of Continental Europe before the Viking Period, but it was still fully integrated with the settlement pattern: where people worked, people also lived, even if this may have introduced a new dimension of seasonality into occupation practices. Productive activities comprised a range from gathering and hunting activities through agriculture and pastoralism to craftwork, and were also conducted at various levels of intensity ranging from subsistence provision to centralized production for surplus and trade. All of these activities belong properly and centrally to settlement studies. A relatively recently defined adjunct to them is landscape archaeology, looking at settled, culturally modified or just culturally exploited landscapes in a broader and more comprehensive geographical perspective than was characteristic of the more atomistic traditions of earlier settlement archaeology. It is perhaps with such an imposed conceptual dichotomy between the particular settlement and the whole landscape (heima versus heimr?) in mind that we can most easily grasp the point that these aspects of archaeology cannot be detached from the historical cognitive or ideological interests of so-called symbolic or post-processual archaeology either. An exploration of contemporary perceptions of the function and character of the whole occupied environment is fundamental to as full as possible an understanding of this whole field of cultural history. Applying essentially the same style of perspective on theme and relevancy, no uncrossable boundaries of chronology or location were imposed ab initio on either the conference presentations or the papers accepted for this collection. While the .3 C. F. C. Hawkes, 'Archaeological theory and method: some suggestions from the Old World', American Anthropologist, 56 (I954), I55-68.

LAND, SEA AND HOM E

4

original conference identified Britain and Ireland as a definitive focus, and the majority of papers in the present volume continue to reflect that, it would simply be misleading to have retained any such restrictive descriptor in the book's title. Even less would it pay to be strict over chronological qualification. While it is, interestingly, the case that a majority of studies here perhaps rather unconsciously imply that the last two centuries of the 1st millennium A.D. were the beginning of a phase of considerable change in relation to what had been seen before, equally many record the importance of continuing and comparative studies of settlement beyond the Viking Period in any strict sense (and one must note that the end of the Viking Period is hardly perfectly defined anywhere): in the Western and Northern Isles of Britain, a post-Viking Norse Period is quite justifiably identified as continuing to the 13th and 15th centuries respectively.4 No apology is needed for a flexible editorial policy and diverse range of comparable studies in this volume; we are happy to recommend those as potentially one of its strengths. The organization of the book's contents could have been done in many ways, and the simple geographical groupings adopted here are merely an obvious, pragmatic solution, intended to make no significant point at all. It nevertheless provides a realistic historical and geographical perspective to begin closest to the Scandinavian homelands of Vikings and the Viking ideology and material culture, not least by looking at new research into aspects of their use of the sea, maritime communications and trade (e.g. Ulriksen, and Muller-Wille and Tummuscheit). The military character of their culture that is essential to any valid application of the term 'Viking' emerges in a number of papers across all the principal areas concerned (e.g. Swift, and Barrett). It is, however, a clear reflection of the dominant trends in recent archaeology that a majority of the papers collected in this volume revisit the central question of the intensity and character of the demographic expansion from Scandinavia in the Viking Period: the colonization - violent or unopposed - of lands in a broad arc around Scandinavia. Some papers present and discuss new discoveries and research that refines our perception of what was happening in the 9th and loth centuries at an essentially factual level (e.g. Bestemann, Redknap, and Abrams and Parsons). Several look specifically at sites or districts, revealing the level of detail at which we can explore and reconstruct the organization of practical life at that time (e.g. Forster et aI., Parker Pearson et aI., Sharples, Cabot et aI., Higham, Coggins, King, Rippon and Gardiner) - not merely, of course, as an exercise in the teasing out of facts for facts' sake, but as a rich comparative resource, and as crucial insights into the real conditions in which the less utilitarian aspects of Viking-period and later medieval cultural life were developed or maintained. Much emphasis here falls upon cultural contact between colonists and native populations, and the transformations that took place in both communities - and indeed in hybrid communities that grew from them - as a result. We can read special and welcome investigations of social and economic transformations in Russia in the east (Pushkina, and Makarov), and Ireland to the west (Sheehan). In and around Britain, meanwhile, rather more emphasis is 4

James Graham-Campbell and Colleen E. Batey, Vikings in Scotland: An Archaeological Survey (Edinburgh,

1998), esp. 1- 3.

AT HOME IN THE VIKING PERIOD

5

attracted to the impact of the Viking intrusions on perceptions and expressions of higher-level group (i.e. ethnic) identity (e.g. Griffiths). A point of particular interest is the reverse influence that Viking expansion ultimately brought to Scandinavia itself in the form, inter alia, of the wider urbanization and monetization that proceeded as the Viking Period gave way to the Christian Middle Ages there (e.g. Risvaag and Christophersen, and d. Hall). Technological studies are represented by discussions of evidence from urban sites and from the artefacts themselves (e.g. Armbruster, Henry, and Cameron and Mould), while the ideological dimension of settlement studies is most directly addressed in papers based upon historical, legal and literary sources (e.g. Wyatt, and Hoff). This collection of conference proceedings is offered as a substantial set of special studies, which reflect much current research into the concrete cultural history of the Viking Period and the centuries that immediately followed it. Such a volume could never attempt to provide a comprehensive introduction to and account of every aspect of its topic: that is the job for a text-book or a quite different style of synthetic overview. There are lacunae that we regret we could not cover - not least a report on the important recent discoveries in Dublin - but we can realistically hope, nevertheless, that readers and libraries will find in this book a wide-ranging and stimulating resource. It may be the case that the impossibility of gathering an exhaustive conspectus of studies covering everything that is known about and all the current research being done on the practical aspects of life in the fuzzy-edged zone we can refer to as the Viking world unmasks the pretensions of the initiative that led to the conference being arranged in the first place. At the same time, it demonstrates equally positively the diversity and dynamism that exists in this field of study at present. The conference was successful: informative and enjoyable. It is hoped that this publication will be a more lasting, and more widely accessible representation of that fact. The conference of which proceedings are published here was generously supported financially by the British Academy, the University of Wales Board of Celtic Studies, the Viking Society for Northern Research, the Consul-General of the Republic of Ireland in Cardiff, the Lord Mayor of Cardiff and Cardiff City Council, and Hordaland Fylkeskommune. The Centre for the Study of Medieval Society and Culture (Cardiff University) and the National Museums & Galleries of Wales were co-hosts of the conference, and together with the Norwegian Church Arts Centre in Cardiff Bay hosted memorable receptions as part of the event. The vicar, churchwar­ den and parishioners of the Church of St Mary and St David, Kilpeck, were most welcoming and hospitable to the conference delegates, and organized a lunch that contributed greatly to the success of a highly enjoyable and informative excursion. To all of these, and many individual members of staff and students, and of course to all the contributors to and participants at the conference who made it such a great success, the warmest thanks are extended. We hope that the present volume will enable them all to feel proud of their association with the event.

*

*

*

6

LAND, SEA AND HOME

DENIS COGGINS, M.A. (DUNELM.), F.S.A., F.S.A.S.

As this volume was about to go to press the editors learnt with great sadness of the passing away of one of the contributors, Denis Coggins. After military service during the Second World War, Denis had followed a career as a schoolmaster and college lecturer in the Teesdale area, before becoming the Antiquities Officer at Bowes Museum, Barnard Castle in 1981. It was in this context that his important and well­ known excavation at Simy Folds, Co. Durham, was carried out, and it is a privilege to publish here his latest thoughts on that site and its context. His career and work represent in an exemplary way the great contributions made by local experts to archaeology in Britain. Active to the end of his life, and much respected, Denis Coggins was President-elect of the Architectural and Archaeological Society of Durham and Northumberland when he died, two months before his 80th birthday. Professor John Hines, School of History and Archaeology, Cardiff University, PO Box 909, Cardiff eFIO 3xu, UK hines@cardiff·ac.uk

DANISH COASTAL LANDING PLACES

AND THEIR RELATION TO NAVIGATION

AND TRADE

By

lENS ULRIKSEN

Nowhere in Denmark is more than 65 km from the sea. In European terms, one can have no problems in regarding the geography of Denmark as belonging to the continental coastal zone, which is an area where the population has thoroughly exploited the marine resources and been widely involved in navigation. This is true enough as far as it goes, but the form of such exploitation, and its organization and intensity, have varied at different periods. In the 6th century, a change appeared in the dominant settlement pattern of Denmark. At this date a number of coastal sites were established which were both distinct in terms of their location right by the coast and also dominated by sunken huts and craft activities, unlike contemporary farms and agrarian villages. The emergence, development and decline of these sites reflects more general social change at the transition from the society of pre-Christian chiefdoms and petty kingdoms with its roots in the Roman Iron Age to the Christian kingdoms that were consolidated after the Viking Age, in the 12th century. By this time the job of the coastal sites had been done, and the majority of them disappeared almost as abruptly as they appeared. For years, Hedeby stood almost alone in representing the coastal landing places of Viking-age Denmark. However, chance finds and systematic investigations have more recently shown that such sites were common in the Danish channels, sounds and fjords. Archaeological studies of Viking-period coastal sites in Denmark have now been in train since the end of the I980s. The metal-detector, both large- and small-scale excavations, and a specific research project have combined to shed light upon these as a category of site that can be distinguished from the general settlement pattern on the basis of their coastal situation alone. Such sites vary amongst themselves, however, to a remarkable degree in terms of both buildings and the finds made there. What they share is their link with the maritime world, contributing widely differing functions such as anchorages, watch-stations, local harbours and production sites. Meanwhile they cannot be identified as true trading places. This paper will outline the relevant navigational considerations and discuss Viking-period trade in relation to a number of coastal landing places in Denmark that have been investigated in the course of the last IO to 15 years, in an attempt to understand their role in Viking-period society.

8

LAND, SEA AND HOME

CONDITIONS FOR NAVIGATION AND THE DEVELOPMENT OF SHIPS

An absolutely fundamental issue of relevance to the coastal sites is that of the sea-level. There has not been any great difference in the coastline of Denmark between the Viking Period and nowadays. Several studies have shown, however, that there have been minor fluctuations in sea-level in Danish waters. 1 The most recent study, undertaken in Roskilde harbour, has shown that the sea-level around A.D. IlOO was between 0.5 and I m higher than it is today.2 That these fluctuations were relatively small- and for the most part insignificant - is underlined by the location of Viking­ period coastal landing places, which are linked to the present coastline. The inner Danish seaways are generally marked by bays, creeks and fjords, while a large number of small islands lie off the coasts. There are thus many places of shelter for ships in passage. The sea depths are not particularly great. Typically, the sea-bed rises steadily up to the coast, and close to the shore there is normally quite shallow water. Depths of a couple of metres several hundred metres out from land are not uncommon, with the further consequences of wave action on the shore being reduced, and landing conditions improved. The Danish fjords have their own unique characteristics. They frequently have only a very narrow, twisting channel between the mouth and the inner fjord. This is equally the case in major fjord systems such as the Limfjord in Jutland and the Isefjord, of which Roskilde Fjord is part, on Sjrelland, as in the smaller fjords. Thus, even with the Viking Period's shallow-draughted craft, a degree of care must have been a constant rule of navigation, not least in unfamiliar waters. How far there may have been some form of marking of channels in the fjords, or whether one could have been piloted through the most difficult passes, we cannot tell. There is no record of sea-marks or beacons before the later Middle Ages, when shipping was quite different in character from that of the Viking Age. Some form of marking of the entrance to the Sli (Schlei) Fjord and the route to Hedeby is certainly possible. Here it was of greater importance that foreign ships reached port safely than that the sailing route was concealed. In other places, by contrast, there would have been good reasons not to show the way to foreign crews, whose intentions would not always have been peaceful. As a stranger, on the other hand, one would not have been reckless, and have risked oneself in all too unwelcoming waters without some form of protection, be that in the form of one's own strength or of guaranteed security. One possibility would have been a pilot familiar with the locality and a guarantee of safe passage

1 D. Hoffmann, 'Wasserspiegelveriinderungen an der Schleswig-holsteinischen OstseekUste im I. J ahrtausend n.Chr.', II1-16 in A. Wesse (ed.), Studien zur Archaologie des Ostseeraumes: Von der Eisenzeit zum Mittelalter. Festschrift fur Michael Muller-Wille (G6ttingen, 1998); J. Herrmann, 'Ralswiek auf RUgen . Die slawisch-wikingischen Siedlungen und deren Hinterland: Teil I - Die Hauptsiedlung', Beitrage zur Ur- und Fruhgeschichte Mecklenburg- Vorpommerns, 32 (1997), 27-31; E. Bondesen, 'Havspejl.s:Endringer og strand lin­ ier: en geologisk-ark:Eologisk oversigt med s:Erligt henblik pa Roskilde fjord', Hist. Arb. Roskilde Amt, I982, 45- 67; idem, 'Strandlinien i Roskilde', I22-30 in E. Bondes~n, R. Dahl-Nielsen and T . Jensen (eds.), Fra Amt til By: Historiske Bidrag i Anledning af Ernst Verwohlts 7.'i Arsdag I. oktober 1998 (Roskilde, 1998). 2 C. Malmros, 'Geologisk unders0gelse af et vandl0b fra vikingetid og middelalder ved Sankt Ibs Vej i Roskilde', 199-210 in T. Christensen and M. Andersen (eds.), Civitas Rosca/d - Fra Byens Begynde/se (Roskilde,2000).

DANISH COASTAL LANDING PLACES

9

from the local lord, while one might also have sold and transhipped goods at the boundary between the freer seaway and a problematic fjord. The Nordic ship-type, the Viking ship, must have a central place in the discussion of sites associated with navigation. Down to the end of the Viking Period one maritime technological event was of great importance, namely the introduction of the sail. When this happened we still cannot establish very precisely. The earliest definite evidence is the Oseberg ship which is dated to c. A.D. 820. 3 Sailing ships are depicted on several picture stones from Gotland, and although there is some uncertainty about the dating of these, which has to be based upon stylistic features, many scholars believe that they go back to the 7th century. We still lack a concrete find to prove the point, but the marked flourishing of landing places in just this period may support the view. As in the case of the sail, the date at which ship-forms diversified into different, specialized types can only roughly be established. 4 In publications of naval archae­ ology terms such as longship, warship, transport ship, cargo ship, fjord boat, fishing boat and ferry are used, quite apart from the medieval terms such as 'snekke', 'knarr', 'skejte' and so on. The original names are particularly problematic, because the sources are late or post-Viking-period and do not provide especially detailed accounts of what, for instance, 'knarr' could refer to. The transport and warship is undoubtedly the sea-going vessel, which as a type had been around for more centuries than any other. Smaller boats and jollies may, however, also belong in this category. The large Viking ships, exemplified by Oseberg, were accompanied by sailing boats that unquestionably belonged in the class of small ships. It is really only towards the end of the roth century that the first ships with cargo holds appear, and when also the sea­ going ships seem to be divisible into ocean-going craft and ships that sailed in more sheltered waters. COASTAL SITES (FIG.

r)

A 'coastal site' is defined as a site that is located no more than 500 m from the sea-edge. This defining line is, of course, artificial, and therefore not always appropriate in relation to actual circumstances. But it has proved that in many cases there is a real difference between what a coastal site looks like and the appearance of the known farms and villages that are further inland. This is the product of differences of function. Location by the shore is in all probability to be linked to seafaring in the broadest possible sense. 3

N. Bonde and A. E. Christensen, 'Dendrochronological dating of Viking ship burials', Antiquity, 67 (1993),

575- 8 3' 4

O. Crumlin-Pedersen, 'Maritime aspects of the archaeology of Roman and Migration Period Denmark' ,

385-402 in K. Hauck (ed.), Der historische Horizont der Gotterbild-Amulette aus der Obergangsepoche von

der Spiitantike zum Frahmittelalter: Bericht aber das Colloquium vom 28.II.-I.I2 . 1988 in der Werner­ Reimers-Stiftung, Bad Homburg (Goningen, 1992); idem, 'Sofart og samfund i Danmarks vikingetid', 181-208 in P. Mortensen and B. Rasmussen (eds.), Fra Starn me til Stat 2: HrilVdingsamfund og Kongemagt (Jysk Arkreol. Selskabs Skrifter, XXII:2, Arhus, 1991), with English summary; idem, Viking-Age Ships and Shipbuilding in Hedeby/Haithabu and Schleswig (Roskilde, 1997); idem and O. Olsen (eds.), The Skuldelev Ships I (Roskilde, 2002), 312-20.

10

LAND, SEA AND HOME

FIGURE I. Coastal sites and major trading places in Viking-age Denmark. I) Aggersborg, 2) Alholm, 3) Bistrup, 4) Dalgard, 5) Fyns Hoved, 6) Gedehaven, 7) Gersh0j, 8) Hedeby, 9) Jyllinge, 10) Karby, II) Lyndby, 12) Lynres, 13) Nabbe­ Kildegard, 14) Nederby, 15) Nres, 16) Ribe, 17) Roskilde, 18) Sebbersund, 19) SelS0­ Vestby, 20) Starup 0stertoft, 21) Strandby Gammeltoft, 22) Stubberup, 23) S0ndem, 24) Vedbrek Stationsvej, 25) Vester Egesborg, 26) Alborg, 27) Arhus.

DANISH COASTAL LANDING PLACES

II

The number of coastal sites of the Danish Viking Period that have been archaeologically excavated is not large. The total amount does not yet exceed 30, a number that should be compared to the more than 200 farming settlements of the Viking Period that have been investigated archaeologically in the last 15 years. Thanks to a research project that was carried out in the middle of the I990S, Roskilde Fjord on Sjxlland is a key area for the study of coastal sites. s This systematic survey has meant that such sites are known to occur quite frequently in Roskilde Fjord. Examinations of individual coastal sites in other parts of Denmark have shown, however, that this might be a relatively common type of site, which is to be found particularly in the fjords and creeks of the inner Danish navigable waters.

Functions and date-range The coastal sites served principally as cross-over points between the agrarian and maritime spheres of Viking-period society. They functioned, for instance, as local district havens, as transhipment sites or craft and trade sites, and indeed are usually multifunctional. With a few exceptions, such as Ribe and Hedeby, the analysis of the hinterland is absolutely crucial in determining the function of the coastal sites. It is necessary to consult as many sources as one can in order to be able to draw the most credible picture of the cultural landscape and settlement pattern of the Viking Period, in order then to examine the relationship between the coast and the interior. Archaeological finds are the primary evidence, but pollen diagrams and place-names are also important factors. All areas of analysis have their own pitfalls in terms of representativity, and conclusions have to be based upon a range of arguments. The coastal sites suddenly become visible in the archaeological evidence in the 6th-7th centuries A.D., which is when the trading wics and emporia of northern Europe were established in considerable numbers.6 The Danish sites continue to be used down to the lOth-IIth centuries, when their special function ceases. The site then either disappears completely or is restructured. It may re-emerge as a major farm and church as early as the 11th century,7 or it may change its character to that of a village in the course of the Middle Ages. s Only exceptionally do coastal sites continue as early Danish towns, although this is the case at Arhus on the east coast of Jutland. 9 In reality there is a marked break between the trading sites of the Viking Age, with

5

This article is based, inter alia, upon the results of a research project carried out by the author in the mid­

1990S. An interim report was published in J. Ulriksen, 'Danish sites and settlements with a maritime context, A.D. 200-1200', Antiquity, 68 (1994),797-811. A monograph with the results of the project has been published

with an English summary: idem, Anlobspladser: Besejling og Bebyggelse i Danmark mellem 200 og IIOO e.Kr. (Roskilde, 1998). A more detailed discussion of the definitions and conclusions relating to Danish landing places can be found in these two publications. 6 Ulriksen (1994), op. cit. in note 5, fig. 3. 7 Ulriksen (1998), op. cit. in note 5, 78- 9I. 8 Ibid., 42 . 9 H. H. Andersen, P. Crabb !Ind H. J. Madsen, Arhus Sondervold: En ByarklEologisk Undersogelse (.Jysk Arkreol. Selskabs Skrifter, IX, Arhus, 1971).

12

LAND, SEA AND HOME

their tight nuclear structure, and 11th-century towns such as Lund and Roskilde, which are characterized by a much greater area and more extensive settlement. 10 The Viking-period coastal sites can in the great majority of cases be categorized as specialized coastal landing places: in other words sites with a special link to the maritime world and where the agrarian element was secondary. A specialized landing place can be either built or not. In the former case, we are typically dealing with sunken huts although a few longhouses of modest dimensions may also be present. The finds consist primarily of pottery and food refuse, namely the bones of mammals, fish and fowl. Amongst the metal artefacts are found jewellery, fragments of weaponry, silver coins, hacksilver, scrap bronze, and melted bronze and lead. There are frequently signs of craftsmen, especially the weaver, bronzecaster and blacksmith, while the combmaker, gold- and silversmith, and the beadmaker are less common. One can also expect ships to have been built and repaired at these sites, although the traces of this may be slight since the wood is not preserved. One ought immediately to be able to expect that large quantities of discarded clench nails and rivet plates should be found, but this is not the case. Analyses of the iron in the clench nails from the coastal site of SelS0-Vestby in Roskilde Fjord have shown that one single nail from here contained iron from three different sources. The iron nails were obviously re­ forged and re-used. A specialized landing place would normally have served several different functions. Nevertheless it is possible to define different categories of sites, four of which will be presented in the following. The most common category is the local harbour. This would have had a settled hinterland, and have been the landing place for the rural villages and farms, which made use of the marine resources and also had a use for a harbour. Such local harbours are usually sited away from the main sailing routes, far into the fjords. Their structures typically comprise sunken huts, possibly supplemented by a few longhouses. The number of buildings is small and, viewed in relation to the long period of several centuries that the artefactual finds normally represent, one cannot imagine that more than a couple of buildings were standing at anyone time. For some years there may indeed not have been any at all. Strand by Gammeltoft in south-western Fyn can serve as an example (Fig. 2). This lies at the inner end of the south-facing Heln~s Bay. At the mouth of the bay are three small islands, associated with which the remains of a system of barriers dating to the 12th century have been observed. l1 On a coastal promontory of sand and gravel, about 3 m above sea-level, within an area of 130 m x 70, 29 sunken huts were excavated in the 1990S.12 The finds consisted mostly of undecorated, flat-bottomed domestic pottery. There were also a few sherds of Baltic pottery of the 9th or loth centuries and a single body sherd of wheel-thrown pottery. Only a very few precious metal artefacts were found, but melted drops of bronze and lead, and scorched clay 10 On Lund, see A. Andren, Den Urbana Seenen Stader oeh Samhalle i det Medeltida Danmark (Malmo, 1985) . For Roskilde, d. F. A. Birkeba:k, E. Verwohlt and M. H0j, Roskilde Bys Historie - Tiden Indti/ I)36 (Roskilde, 1992). 11 o. Crumlin-Pedersen, 'Helna:s-spa:rringen', Fynske Minder (1973), 29- 48. 12 M. B. Henriksen, 'Vikinger ved Helna:sbugten', Fynske Minder (1997), 25- 58.

DANISH COASTAL LANDING PLAC E S

FIGURE 2 .

I3

Strand by Gammeltoft. The excavation area of 1995 can be seen in the centre of the picture. (Photo: Mogens Bo Henriksen)

from an oven plate, iron slag and forging slag show that metalworking was undertaken to a minor degree. Loomweights in the sunken huts also testify to textile production. The dating of the site falls between the second half of the 7th century and the end of the lOth. About 3 km further inland is the associated rural settlement. A rather different type of landing place is found at places which are significant for sea voyages themselves. Sites of this type were not established out of regard for the hinterland but rather because they were anchorages. They are found in areas where the closest village is a long way from the coastal site. The (relatively) empty hinterland meant security for foreign crews, who might otherwise have been an obvious target for attacks from a local population with a taste for getting rich quickly. The anchorages can be difficult to find, as the temporary stays there did not produce large quantities of detritus, and because only occasionally were metal objects lost which the metal-detector can recover. Those sites which we now know permit us to hypothesize that anchorages would be found at small islands and at the very end of spits and points of land, which were normally unused spaces in the Viking Period. As an example, we can consider a site on Fyn's Hoved, the northernmost point of the island of Fyn.13 This site is situated beside an outstanding natural harbour on a primary sailing route through Danish waters, and is also a landmark for the entrance to Odense Fjord. The hinterland on the peninsula is - as far as we can tell- void of 13

M. B. Henriksen, 'Vikinger pa Fyns Haved', Fynske Minder (1994) , 181-91.

14

LAND, SEA AND HOME

traces of settlement from the Viking Period. The culture layer covered an area about 2,000 sq m in extent from 2 to 4 m above sea-level on a regular, sandy slope up from the shore. The uppermost layer included many scorched stones and large quantities of charcoal, although also found were sherds of globular pots from western Denmark, one piece of Baltic ware, clench nails, broken rivet plates, iron knives, an arrowhead, a spearhead, a copper-alloy finger ring, a lead weight, offcuts from sawn antlers, and bone from domesticated animals, game and large cod. This layer could be dated to the Viking Period. Below it was a layer with few finds but a little pottery, apparently of the Roman Iron Age, separated from the Viking-period layer by a layer of sand. Amongst the stray finds were iron axes of the loth-lIth centuries. At the westernmost point of Fyn's Head a silver hoard comprising 101 coins, a number of blanks, and a few fragments of a chopped up arm- or neckring was found. The latest coin in the hoard is from 1017-23. North-west of the landing place another silver hoard containing Kufic coins and silver rings of the Viking Period was found in the 18th century. An empty hinterland and a convenient location for seafaring are also attributes of the site of Lynxs, on the tip of the Halsnxs peninsula at the mouth of the Isefjord and the connected Roskilde Fjord. 14 The entrance to Roskilde Fjord is quite narrow, and affected by a strong current just west of Lynxs. A few kilometres to the North is the Kattegat. This site is located on a shore bank that is bounded by steep cliffs on either side. Up on the bank has been found a culture layer, between 10 and 90 cm thick. Within the culture layer and towards its base are a number of small, clay-lined pits, stone-paved hearths, and stone-lined post-holes. No buildings have been identified. The find assemblage is clearly dominated by undecorated, flat-bottomed pottery of a kind typical of eastern Danish settlements of the 6th-roth centuries. There is also a little Baltic pottery, which is a typical element in settlements of the later Viking Period. Traces of bronzecasting in the form of crucibles and fire-scorched sand were found scattered through the culture layer with a clear concentration around one large hearth. The clay-lined pits, up to a metre in diameter and about 20-30 cm deep, were probably used for the production of herring oil. The latest part of the site, dated to the 12th-13th centuries, is to the East, while the western part is from the 8th-roth centuries. The peak of activity clearly lay in the Viking Period while the later medieval period saw nothing like the same intensity. The site at Lynxs had no settled hinterland to explain its presence. The area seems to have been waste to a large extent. Finds of the Viking Period from the hinterland are distant, sparse and widely scattered, with no real settlement traces having been discoveredY Nor, indeed, do place-names along the peninsula suggest settlements of this period. The place-names provide evidence of newly founded villages that may very well have been established in virgin territory in association with the expansion of agriculture of the 12th century. Lynxs served as a watch station and probably as a departure point for fishermen. It is situated at a strategically very important position on the threshold between the 14 15

U1riksen (1998), op. cit. in note 5, 25.

Ulriksen (I998), op. cit. in note 5, fig. 2I.



DANISH COASTAL LANDING PLACES

IS

Kattegat and the branching fjord system that extends more than 30 km into the island of Sjxlland. Based at Lynxs there are good conditions for controlling the whole navigable area with a ship's crew. Lynxs may also have served as a transhipment site, not for an adjacent hinterland, but rather for the entire Roskilde Fjord. Fortified sites of the Viking Period in Denmark are few. Best known, of course, as the ring forts of the Trelleborg type, although Hedeby, Ribe and Arhus were also defended by ditches and banks. Arhus is now the second largest city in Denmark, and is referred to in a written source as early as 948, when Bishop Reginbrand is mentioned. In the centre of the town are remains of a fortified coastal site of the Viking Period. On the point of a sandy headland which rises 2-3 m above sea-level, traces of a settlement have been found which in the lOth century at least covered an area of 4-5 ha, bounded to the South by the river, to the East by the sea, and with a minor watercourse to the North and wet lowlands to the West. Excavations have been conducted at several sites within this area, revealing the presence of a settlement of sunken-featured buildings both along the shore and the river bank. 16 The initial nucleus of the town lay along the northern bank of the river, but it expanded northwards and during the loth century was enclosed with a semicircular bank, which was enlarged several times. An associated cemetery has been found beside the present cathedral in the centre of the town. Investigations along the river show, however, that even at the stage of foundation around A.D. 900 there was activity at this site. On the inner side of the wall there was a road built of sleepers which was covered over when the enclosure was expanded on one occasion. Evidence for crafts such as bronzecasting, iron preparation, textile production, wood- and bone- or antlerworking was found. The finds were dominated by a copious amount of pottery, including hemispherical vessels, flat-bottomed pots, and Baltic pottery. There were also items of Viking-period jewellery, a Kufic silver coin, weights, glass sherds, glass beads, fragments of soapstone vessels, fragments of weaponry, iron bars, basalt quernstone, and fishing equipment. The establishment of the semicircular rampart at the beginning of the loth century and a - nominal, at least - status as a bishop's see in 948, show that Arhus was of determinative regional importance, which must have been based upon something more than the agricultural economy. This economy is assumed to comprise trade and craft, which would have been encouraged by the protection of a military and administrative centre under royal control. Arhus distinguishes itself topographically from the other semicircular enclosures of Denmark by lying open to the sea, which could not simply have been fortified by a fixed barrier structure. It is therefore reasonable to imagine that the site should have had defences alongside or on the shore in addition to this rampart. Considering the much more favourable natural landing sites that can be found in various places a few leagues to the North, there must have been compelling reasons for the location of Arhus on the low sandbanks surrounded by a river and by low, wet meadowlands ,16 Andersen, Crabb and Madsen, op. cit. in note 9; H. H. Andersen and H . J. Madsen, 'Byudgravning ved Arhus Katedralskole', Kuml (1985), 35- 95; H . Sko,:, 'Udgravningerne ved Aarhus Katedralskole i 1994-95', Kuml (1995-96), 189-206; idem, 'Udgravningerne i Arhus Midtby 1994-97', Kuml (1997-98), 227-94.

16

LAND, SEA AND HOME

and marshes. Arhus is conveniently located in relation to the sailing route between the northern Kattegat and the channels around Fyn, and therefore to the route to the south-western Baltic and Hedeby.17 The hypothetical basing of warships at Arhus, and a permanent watch on the Arhus Bay, may be background to the trade and craft site being founded and enclosed with a ditch and bank by the early loth century. This was, in other words, a military stronghold which nurtured civil activities that grew to an extent that eventually made Arhus one of Denmark's earliest sees, along with Ribe. Production sites are a category which - as it appears at present - comprises coastal sites where the finds include relatively large quantities of waste from one or more crafts. One such site is Sebbersund, which is situated on the important water route of the Limfjord in Jutland. This site, which was investigated archaeologically in the 1990S, is located on a tongue of land that includes a prominent morrainic peak that rises to 31 m above sea-level, and a lower-lying, flat area about 5-6 m above sea­ level with low coastal meadows on either side. Apart from a demolished stone church of the 12th century on the top of the hill, the area of investigation remained unsettled until very recent years. The remains of activity cover an area of c. 240 m x 170, where 80 sunken huts have been excavated at the northern end of the site while to the south the location of a wooden church with a fenced churchyard with a thousand burials of the lIth century was foundY Between these was an unoccupied area which the excavators suggest was reserved for workshops and perhaps for trade. Particularly large quantities of iron slag and forging slag catch attention. Textiles were also produced here, while melted drops of gold and silver and casting crucibles for bronze show that such craftsmen also worked here. The dating of the site points to foundation in the second half of the 7th century, continuation through the 8th to 10th centuries, and a marked intensification in the quantity of finds in the lIth century, when amongst other things, Urnes brooches were produced. After this boom the site evidently came to an end around 1100, probably because the conditions for navigation in the Limfjord changed. According to the evidence of the metal artefacts, Sebbersund was in use for more than 400 years, and even though the 80 sunken huts all pre-date the richest phase of the site in the I Ith century, this is not, in itself, a mark of a large settlement. As in the case at most other sites, the buildings cannot be dated precisely enough to determine the intensity of settlement within any period less than a century. It has been suggested that trade was a decisive factor in the establishment of Sebbersund, but this is debatable. The finds are neither rich in imported goods, silver, balances or weights, nor do the crafts show the diversity that characterizes a trading site such as Hedeby. Here it is appropriate to stress the massive evidence for iron working and, presumably, its redistribution as high-quality goods, which could be an explanation of the site. This waste, however, is not dated closely enough to determine whether these activities were practised there for centuries or could possibly have been limited to a narrower 17

J. Ulriksen, 'Anl0bspladser fra Danmarks yngre jernalder pa baggrund af Roskilde Fjord-unders0gelsen',

II 6- 52 in Karm0yseminaret 1997 (Stavanger, 2000), fig. 16.

18 P. B. Christensen and E. Johansen, 'En handelsplads fra yngre jernalder og vikingetid fra Sebbersund', Aarb0g. Nordisk Oldkyndighed Hist. (1991), 199- 230.

DANISH COASTAL LANDING PLACES

17

FIGURE 3. Vester Egesborg. View over area of excavation with the culture layer in the foreground. Dybs0 Fjord can be seen behind. (Photo: Jens Ulriksen)

phase. It would not be a surprise if such a phase were to coincide with the zenith of the site in the lIth century. At Vester Egesborg in the very south of Sja:lland, there is a coastal site with dense remains of iron working and fine craftwork and a structure and size that corresponds to Sebbersund, but with a quite different location in respect of traffic, sheltered in the relatively inaccessible Dybs0 Fjord (Fig. 3). This site is still under excavation, so this report on it has to be preliminary. The site is positioned on an evenly sloping field by the shore, and the excavation season of 2001 investigated 90 sunken huts, 12 three­ aisled buildings with bowed walls, of which only one is around 20 m long, the others being between 12 and 16 m length. Compared with contemporary rural farms, these are small. A blacksmith's smithy has been excavated, and iron scrap and slag, forging slag and hammerscale from various smithing processes has been found over much of the site. Finds of copper-alloy models for making clay moulds, fragments of moulds, and melted silver and bronze reveals the production of jewellery in silver and copper alloy. A bronze patrix die for a double foil figure Igubbe) is also noteworthy. In the sunken huts has been found evidence of extensive textile production, while the combmaker and glass-beadmaker also worked here. In 2001 we also found two breaking ditches for drying flax, a discovery that shows that flax was prepared here to make linen. A special feature of the site at Vester Egesborg is the jewellery and gilt copper-alloy and silver mounts, which are of very high quality, and which are found by metal-detecting the ploughsoil and in the excavation of the sunken huts. There are

18

LAND, SEA AND HOME

also about 25 whole or fragmentary Kufic coins, and five imitation Sassanid drachmae, the latter being a coin-type that is rare in Denmark. Based on the jewellery, the mounts and the coins, the site can be dated to a period from the middle of the 6th century to the second half of the 10th, when it was abandoned. In relation to this date-span, the 90 sunken huts and 12 three-aisled buildings are not necessarily evidence of a large, permanent settlement, even though several of the buildings can be seen to have had two phases. In size and structure, Vester Egesborg is similar to the extensive settlement that was found beneath the ring fort of Aggersborg by the Limfjord. This has been regarded as a village, which was abandoned when the fort was built, although alternatively it could have been a specialized coastal site at the important crossing point between the East-West sailing route through the Limfjord and a passage way over the narrow part of the fjord for land traffic. 19 Vester Egesborg, however, is not located at a crossroads. It is hidden out of the way right inside a shallow fjord, 6-7 km from its mouth which, from the sea, is also more or less hidden by two islands. Very little is known about the hinterland around Vester Egesborg. The peninsulas on either side of the fjord seem first to have been settled as part of the agricultural expansion of the 12th-13th centuries. The nearest village is Vester Egesborg itself, which is about I km away from the coastal site. The church here was built around 1200, but a couple of newly discovered gilt copper-alloy mounts indicate that this village may have had roots in the 7th or 8th centuries. About 4 km inland is the village of Hammer, where the church dates from around lIOO and has rare examples of stone carving in the Ringerike Style. Outside this church a Kufic silver coin has been found and Viking-period pottery. Although there are still many unknown elements in the settlement of the hinterland, the coastal site at Vester Egesborg can be regarded as a combination of a local harbour and production site beside well-sheltered waters, where local chieftains kept their ships. About 10 km south of Vester Egesborg is an unusual production site, Nces by Avno Fjord, founded in the second half of the 8th century and abandoned by the year 1000.20 This was found at the mouth of a small river on a sandy promontory which rises up to 5 m above sea-level. The fjord lies to the West, while to the North and South there are wetlands. The settlement comprises 20 three-aisled buildings, two of which are proper longhouses more than 30 m long. The other three-aisled buildings were of humbler dimensions and are interpreted as ancillary buildings. There were also 69 sunken huts, which are evenly distributed across the entire site, while quite remarkably 57 wells and an artificial water channel 150 m long that flowed across some of the wells were found. Several of the wells contained wads of flax in such great quantities that there is no doubt that this site saw large-scale flax production, a conclusion that is supported by the discovery of breaking ditches in which the flax was dried. The sunken huts contained equipment that was used for textile production, such as loom-weights and seam smoothers.

Ulriksen (1998), op. cit. in note 5, 145-6. K. Meller Hansen and H . Heyer, 'Nres 59- 89. 19

20

en vikingetidsbebyggelse med herproduktion', Kuml (2000),

DANISH COASTAL LANDING PLACES

19

VIKING-PERIOD TRADE IN DENMARK

The Frankish Annals report that in the year 808 the Danish King Godfred destroyed the trading site of Reric, in what is now Germany, and moved the merchants from there to Hedeby. Merchants and their services were highly prized. In the Annals of Fulda it is noted under the year 873 that the peace between King Louis and the Danish Kings Halfdan and Sigurd was confirmed so that trade between the kingdoms could continue unhindered. This sounds practically like a regular trade pact. Other sources refer to quite different categories of goods being imported and exported. When Ottarr, a leading farmer and trader, visited Alfred the Great's court towards the end of the 9th century, his account of how he sailed from his home in northern Norway to Hedeby was written down. In Hedeby he traded his goods, which consisted of walrus teeth, hides, pelts, ropes and feathers. And even though the Arab, Ibn Fadlan, is not so detailed on the subject when describing his meeting with Scandinavians on the Volga in the loth century, he also notes pelts, and slaves as well, as important goods for the Vikings. There is no reason to believe that these reports should not be true. But all the way through they deal with things that could not - or could only in very fortunate circumstances - be found in archaeological excavations. The extent of trade in these types of goods thus cannot be more thoroughly investigated in any sort of systematic way. This is not the place to undertake a full discussion of the concept of trade in the Viking Period. 21 I wish, however, to give a very concise account of some of the problems that one must note in discussing the extent of trade in Viking-period Denmark. It is first and foremost necessary to define what one means by the term 'trade'. A view that I subscribe to is that 'trade' is to be defined as an unforced transaction between two parties, through which a given ware changes hands in return for some payment in the form of some other goods or a neutral medium of exchange such as precious metal or cloth. But the crucial problem is to discover some conclusive archaeological evidence that trade in this sense took place at a particular site. When everything is considered, Denmark is an area without natural metal resources apart from bog iron. This means that most of her iron and all of her copper, silver, gold, tin and mercury were imported, and thus are reflections of trade in some form or other. Finds typically taken as archaeological evidence of trade are the balances and weights, coins, hacksilver and imported goods already mentioned. Weights - and balances themselves, which are rarer as they are more fragile - are known from many sites of varying size and type from the 3rd century A.D. to the 12th (Fig. 4). These were not only a trader's equipment but were also used in other contexts. Scales would, for instance, have been used in connexion with the payment of fines or rewards in precious metals, and craftsmen would also have used them to weigh metals for alloying and so on. A grave with these objects thus cannot undisputably be said to contain a trader, no more than finds of the same character from a settlement allow one to conclude that trade took place there, let alone that it was a trading place. This is emphasized by the widespread distribution of 6th- to lIth-century settlement sites 21 A fuller discussion o f trade and development in the later Iron Age in Denmark, including the Viking Period, can be found in Ulriksen (1998), op. cit. in note 5, chapter 8.

2.0

LAND, S EA AND HOME

FIGURE 4.

Viking-period settlements at which weights have been found.

DANISH COASTAL LANDING PLACES

21

with weights. If this should be seen as an indicator of the presence of traders and trading sites, in my opinion the number would be improbably high compared with the mercantile level of Viking-period society. Coins and hacksilver cannot unambiguously be associated with trade any more than weighing apparatus. As with the distribution of weights, the distribution of Viking-period coins shows that they occur at a broad cross-section of sites covering the whole land. In respect of sceattas, there is a significant distribution. This type of coin is found at the Danish site of Ribe, as well as at sites on the German Baltic coast and Ahus in Skane, the southernmost part of Sweden. In other words, the sceattas are apparently restricted to sites that were participating in the North Sea and Baltic trading system. 22 In the case of hacksilver, the payment of large sums in connexion with bride­ prices, fines, tributes and so on could of course have included precious metal, and as late as the 12th and I3th centuries the sagas provide us with quite detailed information about the use of, for instance, silver rings in such a connexion. One must also remember that precious metal was also cut up, be it gold, silver or bronze, for it to be melted in the miniature crucibles in order to make jewellery and similar objects. As far as imported items go, jewellery can be argued to be generally a poor indicator of trade. It is highly likely that these de luxe artefacts came to Denmark as booty or again by association with marriage exchanges. Considered in relation to the Viking-period graves, there is nothing to indicate that there was any great interest in wearing fine jewellery from, for example, France or England. The form and style of the jewellery and mounts are predominantly Nordic. The amount of imported pottery is small, not just in absolute terms but also in comparison with other foreign artefact-types. In Jutland there is some penetration of western European pottery in the 9th and IOth centuries which is believed to have been imported for the sake of the pottery itself, but if that was the case it was not to the taste of the eastern parts of Denmark. It is also important to note that new studies of a special type of wheel-thrown pottery of the 8th century found in Ribe and its area have shown it to be locally made.23 This is surprising because the potter's wheel was not otherwise commonly used in Denmark before the I3th century. This discovery is of considerable scholarly importance, as an 'import type' has suddenly disappeared and become local. As has been implied, precious metals and other luxury or prestigious artefact­ types can be difficult to isolate in a merchant's inventory. The situation is different with the so-called consumer goods that were used in daily life, the prestige value of which was low. In this category belong basalt quernstones from Germany (Fig. 5), granite slates for quernstones, slate for whetstones and soapstone vessels from Norway (Fig. 6). It is hardly likely that these were booty, but the importation of these 22 A sceatt of the 'Continental Runic' type (c. 705-15) has also been found at Gudme on Fyn, which was a pre­ eminent central place from the 3rd century to the 6th. The subsequent development of the site from the 7th century to the loth is less clearly revealed: P. 0. Sorensen, 'Houses, farmsteads and the settlement structure in the Gudme area', 41-7 in P. O. Nielsen et al. (eds.), The Archaeology of Gudme and Lundeborg (Copenhagen,

1~94).

C. Feveile, S. Jensen and K. Lund Rasmussen, 'Produktion af drejet keramik i Ribeomriidet i sen yngre germansk jernalder', Kuml (1997-8),143-60.

22

LAND, SEA AND HOME

FIGURE

5.

Viking-period settlements at which finds of Rhenish basalt quernstones have been made.

DANISH COASTAL LANDING PLACES

FIGURE

6.

Viking-period settlements at which Norwegian soapstone has been found.

23

24

LAND, SEA AND HOME

practical items was not the consequence of some vital and otherwise unmet need. A demand for them was nonetheless there, when there was a supply. This is clearly shown by the finds from the Viking-period settlements of Denmark. But it must be emphasized that most of the dots on the distribution maps of basalt or soapstone represent one or two very small fragments, not vast numbers, let alone intact querns or vessels. Both documentary and archaeological sources identify trade as an aspect of Viking-period society. But how extensive was trade, who participated in it, and where and how were the actual transactions conducted? The need for trade and the transport of goods was not particularly high in Denmark. There were no settlements there that needed large quantities of supplies on a daily basis before the IIth century. The exceptions are trading sites with military functions such as Hedeby, Ribe and Arhus. Settlement consisted of agrarian farms and villages which were essentially self­ sufficient. It was not necessary to obtain goods from outside in order to live and survive. That one occasionally obtained high-quality iron from Sweden, a good quernstone of basalt lava from Germany, or a cooking pot of soapstone from Norway is a fact. But these were not essential as there were local alternatives. The quantity of these consumer goods at Danish settlements is very slight in spite of the fact that they are durable commodities. It seems exaggerated to use them as arguments for an extensive level of trade. The situation is the same with other imported goods, be they metal, glass or ceramic. Those who took part in trade were those who had something to sell and the means to buy. Both of these require surplus agricultural production in a country where the extraction of raw materials or hunting of fur animals, walrus or narwhal were not available options. In reality, we do not know all that much about who it was who owned the farms whose sites we have found: whether there was a large class of free farmers who could produce a surplus that could be traded, or a much smaller proportion of the population, a limited group of aristocratic families, who controlled the soil and its potential wealth. Without knowing this for certain, we can still picture to ourselves the exporting of clothing, grain, honey, beeswax, cattle on the hoof, hides and the like. In any circumstances there was a need for resources to prepare and transport agricultural products, a task that would have called upon organizational powers and privileges that extended beyond the purely local. It is hardly appropriate to believe that the individual farmer brought his wares to the shore where a ship was waiting to freight the goods on to distant lands. It is more likely that the produce was collected at a lord's seat, where it could be worked further - in some cases by specialist craftsmen - before the goods were sailed to Hedeby or some other trading site that specialized in long-distance trade. In my view, then, it was the leading families that carried out trade in Denmark. They were the heads of a society in which alliance-forming gift-giving was far more important than purchases and sales in a market. Their position and social ambitions depended very much upon their ability to share out valuable and prestigious objects. Such objects might be obtained abroad, or be produced by skilled craftsmen in the pay of a lord. He would at the same time have a manifest interest in controlling access to such objects, and would, within his own territory, have been the individual who

DANISH COASTAL LANDING PLACES

25

bargained with any foreign merchant who might pass by. Subsequently he could distribute these goods to his own family, his retainers and to those he had in his serVIce. COASTAL SITES, SEAFARING AND TRADE

Ribe and Hedeby are the best known examples of Viking-period trading sites in southern Scandinavia. That both are situated on the threshold between the North Sea and the Baltic area is no coincidence. They were established in imitation of the emporia of north-western Europe, in order to take their place within that network of sites. They were royal sites, where the king - along with his peers in other countries - demanded dues; and with fortification they also became military strongholds. A manifest degree of organization, with their small tenements, is a characteristic of these sites, and the buildings at Hedeby have no parallels amongst the now large number of building plots that have been examined at Viking-period settlements in Denmark. A further feature is minting, which did not become a regular element of Danish royal activity until the 1 Ith century, but which began as early as the 8th century in Ribe and a little later at Hedeby. The sceattas of Ribe clearly emphasize how their target area was the golden triangle of the North Sea. To the east, the Baltic was the navigational highway, a situation that is emphasized by finds from coastal sites along the German coasts. Western European pottery and sceattas have been found, for instance at Grog Stromkendorf, which was clearly the Reric that King Godfred shut down in 808, and at Ahus in eastern Skane, which was another stopping place on the route to Birka near Stockholm and the Russian river routes. It is a definite possibility that there were one or two similar sites on the Danish side of the Baltic. In that case the islands south of Sjxlland - Lolland, Falster and M0n - would have been ideal. We are told, however, in Wulfstan's account in the Alfred the Great's translation of Orosius' World History, that he sailed from Hedeby to Truso along the south coast of the Baltic, night and day. The route here has no dangerous tongues of land, islands or shallows which the Danish coasts have so many of. The coastal sites that have been discussed in this paper cannot be compared with Ribe and Hedeby. They were not trading sites or beach markets, which are characterized by a huge assemblage of objects and display a form of generally organized structure, and where foreign merchants from both near and far arrived in the course of the season. Some of them were nonetheless indirectly involved with trade, as they were production sites and landing places, established by the aristocratic families who were looking for the opportunity to trade surplus production for prestige goods. We are not faced with a general 'mercantilization' of seafaring and sites in 8th­ to 10th-century Denmark. This only happened with the drastic change of Iron-age society at the end of the loth century and especially in the lIth century, when the traditions of centuries can be seen to be breaking up. All over Denmark we can observe the abandonment of the specialized landing places in favour of coastal agrarian sites, which in some cases comprised a manor farm with a church, while other sites simply disappeared. The large trading and craft centres also moved out of

26

LAND, SEA AND HOME

their previous bounds, with the movement of the nucleus of the site either close by, as happened at Ribe and Arhus, or more markedly as at Ahus in Skane and Hedeby. The result was the urbanization of the lIth and 12th centuries. This development is apparently reflected in the finds of ships of Scandinavian character. It is in any case quite striking that wrecks from the 6th to lOth centuries are few in comparison with those from the lIth century and later. Even in the harbour at Hedeby, which was in use from the 8th century onwards, the ship finds are predominantly of the 1 Ith century, and this trend is repeated all around Denmark, not least in the light of the recent finds in Roskilde harbour. This discrepancy could of course be a product of chance, but looked at in relation to social change a different picture emerges. The number of ships quite simply seems to have increased radically in pace with the introduction and consolidation of the Christian medieval social order, a structure which included, amongst other things, coastal towns and clerical institutions that depended on the frequent supply of quite basic commodities such as food and fuel from outside. It is from the same period that we have evidence for a development in ship-types, showing a genuine specialization of craft for different purposes, a development that was spawned by a greater need for sea transport. An intensification of seafaring at this time provides food for thought in terms of the significance of sea-borne trade prior to the lIth century. The image of numerous cargo ships ploughing the waters around Denmark from the 6th to loth centuries has to be modified, and the imagined level of trade reduced in the same measure. Dr Jens Ulriksen, Roskilde Museum, Sankt Ols gade [email protected]

I5,

DK 4000 Roskilde, Denmark

VIKING-AGE PROTO-URBAN CENTRES

AND THEIR HINTERLANDS: SOME

EXAMPLES FROM THE BALTIC AREA

By

MICHAEL MULLER-WILLE

and

ASTRID TUMMUSCHEIT

This article will focus on three proto-urban centres and their hinterlands. The first two of these are Ribe and Hedeby. Both sites are situated in the southern part of Jutland: Ribe on the North Sea coast, on the western side of the peninsula, and Hedeby on the eastern side, with access to the Baltic Sea. There is a then another settlement, GroB Stromkendorf, which has only recently been investigated as part of a major research project. This site is very probably identical with the historic trading centre of Reric, which is known to be the predecessor of Hedeby. Each of the three sites dates back to the 8th century.1 While Reric was abandoned in the early 9th century, Ribe and Hedeby developed into centres of outstanding importance during this period, that is from the 9th to the IIth centuries (Fig. 1.1). By the end of the IIth century Hedeby was succeeded by Schleswig, whereas Ribe continued during the Middle Ages and into modern times until this very day. No other settlements of urban character came into existence in southern Scandinavia before the period from the IIth to the 14th centuries. 2 They are mainly distributed on the eastern side of Jutland. Only during the 15th and 16th centuries were a few towns founded on the western coast too (Fig. 1.2). We can also map the distribution of Viking-age finds: that is graves, settlements, hoards and single finds, in the area between the rivers Eider in the south and the Kongea in the north. Concentrations of sites are visible in the environs of Ribe in the north-west and Hedeby in the south-east. 3 There are further concentrations of finds on the Frisian islands, on the Eiderstedt peninsula in the west and on the Isle of Alsen in the east. Apart from the proto-urban centres of Ribe and Hedeby, rural settlements make up the large majority of settlement sites (Fig. 1.3).

1 J. Call mer, 'Urbanization in Scandinavia and the Baltic Region c. AD 700-IIOO: trading places, centres and early urban sites', 50-90 in B. Ambrosiani and H. Clarke (eds.), Developments Around the Baltic Sea in the Viking Age (Birka Studies, 3, Stockholm, 1994). 2 A. Andren, Den Urbana Seenen: Stader oeh Samhal/e i d et M edeltida Danmark (Acta Archaeologica Lundensia Series in 8°, 13, Malmo, 1985). 3 S. Eisenschmidt, 'Grabfunde der Wikingerzeit in Nord· und Siidschleswig', 55-75 in D. Meier (ed.), Syttende Tvterfaglige Vikingesymposium , Kiels Universitet I998 (Aarhus, 1998).

28

LAND, SEA AND HOME

1: Early urban centres of the 8th to lIth centuries (after Callmer, op. cit. in note 1). 2: Urban centres and towns of the Early, High and Late Middle Ages (after Andren, op. cit. in note 2). 3: Distribution of Viking-age finds between the rivers Eider and Kongea (after Eisenschmidt, op. cit. in note 3).

FIGURE 1.

PROTO-URBAN CENTRES

29

Dendrochronological dates prove that at Ribe trade and manufacturin& activities started in the early 8th century.4 Situated on the eastern banks of the Ribe A, about 5 km away from where the river flows into the North Sea, opposite the medieval town centre with its cathedral, a market site came into being (Fig. 2.1-3). An area of 200 m by 65, which ran along the banks of the Ribe A, was divided by shallow ditches or fences into well-regulated, narrow strips of land (Fig. 2.2). There is no evidence of substantial buildings in the market area except for remains of simple huts, some of them with a sunken floor. In the first half of the 9th century the market area was enclosed by a shallow ditch, which was developed into a substantial fortification in the late loth century. There were two associated cemeteries (Fig. 2.3). The following maps show the distribution of imported finds from rural settlements in the hinterland of Ribe, as recorded by Stig Jensen (Fig. 3.1-4). Objects serving everyday needs such as whetstones, quernstones, soapstone and Pingsdorf­ ware pottery are known from 30 settlements. 5 Most of these are located on the sandy soils and on the border of the fertile marshes. Investigations of some of the sites have shown that there was a great diversity of different settlement-types, including single farmsteads, small villages and chieftain's farms. Most of the settlements, 26 of them in fact, have yielded whetstones (Fig. 3.1). Nineteen of these sites have produced whetstones of Norwegian origin. Exportation from Eidsborg in Telemark, Norway, is detectable in the Ribe area from the 9th century onwards. Meanwhile only seven sites have produced fragments of soapstone vessels (Fig. 3.2). We know from the Ribe excavations, that these too belong to the 9th century. Both the whetstones and the soapstone vessels were diffused into the hinterland via Ribe. The same applies to finds from the Rhineland area. First there are quernstones of basalt lava from the Eifel region, which have so far been found at eleven sites near Ribe (Fig. 3.3); secondly there are fragments of Pingsdorf-ware pottery, which are known from six sites (Fig. 3.4). In all the cases considered here the trading centre of Ribe can be identified as the distributor of imported goods into the hinterlands. With regard to the differences of quantity one has to remember that quernstones, soapstones and Pingsdorf Ware were almost exclusively found in the course of excavations. Whetstones, on the other hand, tend to attract the attention of the average collector more easily than sherds of pottery or fragments of basalt do. It is probably due to that fact that systematic field-walking and random surface finds have added substantially to the numbers of whetstones. Another example from the Middle Ages serves to highlight the area that was closely linked to the urban centre. The distribution map of volcanic tufa imported from the Middle Rhine area, which was used for church building, defines the hinterland of Ribe very clearly (Fig. 4.1). According to Claus Feveile there is evidence 4 s. Jensen, Ribe zur Wikingerzeit (Ribe, 1991); C. Feveile, 'The latest news from Viking Age Ribe: archaeological excavations 1993', 91-9 in Ambrosiani and Clarke (eds.), op. cit. in note I; C. Feveile and S. Jensen, 'Ribe in the 8th and 9th century: a contribution to the archaeological chronology in North Western Europe', 9- 24 in S. S. Hansen and K. Randsborg (eds.), Vikings in the West (Acta Archaeologica, 71, Copenhagen,2000). S S. Jensen, 'Handel med dagligvarer i vikjngetiden', Hikuin, 16 (1990), II9- 38; S. Jensen (ed.), Marsk, Land og Bebyggelse: Ribeegnen Gennem IO.OOO Ar. (Jysk Arka!ologisk Selskabs Skrifter, 35, Aarhus, 1998).

30

LAND, SEA AND HOME

FIGURE 2. Ribe

I: Location map. 2: The market area of the 8th and 9th centuries. 3: The market area and

associated cemeteries of the 8th and 9th centuries (after Jensen, and Feveile, op. cit. in note 4).

PROTO-URBAN CENTRES

31

3. Distribution maps showing imported finds from rural settlements in the hinterland of

Ribe. I: whetstones; 2: soapstone; 3: quernstones; 4: Pingsdorf Ware (after Jensen, op. cit. in

note 5).

FIGURE

32

LAND, SEA AND HOME

FIGURE 4. I: Distribution map of volcanic tuff from the Eifel region in the Ribe area (after Feveile, op. cit. in note 6).2: Distribution of churches built of tufa in the region between Eider and Kongeii (after Ludtke, op. cit. in note 7). 3: Hedeby - the topography of the 8th century. 4: Rericl Gro~ Str6mkendorf - the topography of the 8th and early 9th centuries (after J6ns, op. cit. in note 8).

PROTO-URBAN CENTRES

33

of more than 50 churches in which tufa was used as a building material. 6 In his opinion the evidence points towards a distinct and deliberate phase of importation during the period from Il75 to 1250 which not only included tufa but also the architectural style associated with this building material in its area of origin. Compared with Ribe, the distribution of tufa in the environs of Schleswig does not show the same layout. Only a linear pattern along the Schleifjord in the east, to the port of Hollingstedt and the estuary of the River Eider in the west, is visible in the map (Fig. 4.2). Tufa was also used for burial structures, as for example is evidenced from the church that was found beneath Schleswig market square. By dendrochronol­ ogical analysis of their wooden covers, these graves could be dated to the period from 1080 to 1170.7 As at Ribe, the earliest beginnings of Hedeby, the predecessor of the town of Schleswig, date back to the 8th century (Fig. 4.3). Dendrochronological dates show that the first jetties were built around the first or the second quarter of the 8th century.8 Although there proved to be no contemporary settlement traces within the later ramparts, both the so-called SiidsiedLung or 'southern settlement' and a burial ground are close by, and these have yielded 8th-century finds. The site probably consisted almost entirely of sunken huts with only one associated longhouse. In its later phase the SiidsiedLung co-existed with a settlement further north, which was situated in the area later enclosed by the rampart. During the 9th century the SiidsiedLung was abandoned and the second settlement area developed and extended to the west. Around the middle of the loth century the site was surrounded by a semicircular rampart enclosing an area of about 24 hectares. During the 9th and loth centuries Hedeby was organized into small plots of land which were densely occupied by post-built houses. The distribution map in Figure 5.1 shows sites that date to the 8th and 9th centuries in the hinterland of Hedeby.9 There are traces of occupation on both sides of the Schleifjord and some near the Flensburg fjord, too, but graves account for most of the symbols in the map. The two settlements near Kosel, the sites at Schuby and at Winning, have been excavated. Neither of these came into existence before the 9th century and, apart from the SiidsiedLung at Hedeby, there are no traces of settlements of the 8th century in the region so far. 10 6

0

C. Feveile, 'Tufstenskirkerne i Sydvestjylland -

set i arkreologisk, handelshistorisk belysning', By, marsk

8 (1995), 31-51. 9geest, H. Liidtke, 'Die archaologischen Untersuchungen unter dem Schleswiger Rathausmarkt ', 9-84 in V. Vogel

(ed.), Kirche und Griiber(eld des I I . - I j . Jahrhunderts unter dem Rathausmarkt von Schleswig (Ausgrabungen in Schleswig: Berichte und Studien, 12, Neumiinster, 1997). 8 M. Miiller-Wille, 'Ribe-Reric-Haithabu: Zur friihen Urbanisierung im siidskandinavischen und westslaw­ ischen Gebiet', 321-37 in K. Brandt, M. Miiller-Wille and Chr. Radtke (eds.), Haithabu und die (ruhe Stadtentwicklung im nordlichen Europa (Schriften des Archaologischen Landesmuseums, 8, Neumiinster, 2002); H . Jons, 'Friihstadte. Thema: Ostseehandel', Archiiologie in Deutschland, Heft 4 (2000),22-7. 9 K.-H. Willroth, Untersuchungen zur Besiedlungsgeschichte der Landscha(ten Angeln und Schwansen von der iilteren Bronzezeit bis zum (ruhen Mittelalter: Eine Studie zur Chronologie, Chorologie und Siedlungskunde (Offa-Biicher, 72, Neumiinster, 1992); M. Miiller-Wille, 'Friihstadtische Siedlungen und)hr Umland in der Zeit von 800-1200: Einige Beispiele aus dem Ostseeraum', 27-40 in V. Kazakevicius and V. Zulkus (eds.), The Baits and their Neighbours in the Viking Age (Archaeologia Baltica, 2, Vilnius, 1997) . 10 D. Meier, 'Winning Eine wikingerzeitliche Siedlung am Ufer der inneren Schlei', 117-26 in A. Wesse (ed.), Studien zur Archiiologie des Ostseeraumes: Von der Eisenzeit zum Mittelalter (Neumiinster, 1998).

34

LAND, SEA AND HOME

FIGURE 5. The archaeological evidence in the hinterland of Hedeby.

I: 8th and 9th centuries. 2: roth century. 3: lIth century. 4: Distribution of imported finds (after

Hansen, op. cit. in note r6).

PROTO-URBAN CENTRES

35

During the loth century, when the semicircular rampart was built at Hedeby, the distribution map illustrates a spread of sites across the region of Angeln (Fig. 5.2). Most of the symbols in this area represent finds from graves and there are few hoards, too. 11 In the IIth century a different picture emergesY In contrast with the preceding centuries, there are no graves that can be dated with certainty apart from few examples at Hedeby (Fig. 5.3). This is due to the introduction of Christianity. Apart from the single site in Schleswig no churchyard burials have been excavated so far. At the same time settlements such as Kosel and Schuby continued in beingY All three maps show very clearly that Hedeby and its nearest neighbours relied on a system of waterborne communication with strong connexions to the Baltic Sea via the Schleifjord. Unlike the well-known centre of Gudme/Lundeborg on Fyn in Denmark, we do not know whether there was anything comparable in the Schlei region before the rise of Hedeby.14 Excavations at some sites in the environs of Hedeby have provided us with insights into rural settlements of the same period. In particular there are two settlements near Kosel, one to the west of the village and one to the east. 15 Both sites comprised of a large number of sunken huts and just a few associated longhouses. There were no enclosed farmsteads and in that respect the settlements at Kosel and other settlements in the hinterland of Hedeby do not display the typical layout of rural settlements of the Viking Period or of preceding times. In contrast with settlements like Vorbasse in Denmark, where farms comprised fenced plots of land, the scattering of a large number of sunken huts in association with a few longhouses is considered unusual in the rural context. The settlement site east of the village was situated on a gentle slope whereas the contemporary cemetery lay further east on a low knoll. The burials were mostly inhumations with or without a coffin and a few chamber-graves. Most of them were covered by mounds. The finds assemblages of the settlements at Kosel and of other settlements in the hinterlands of Hedeby have yielded imported material comparable with the finds from the Ribe area: there is soapstone, whetstones, quernstones made of basalt,

Willroth, op. cit. in note 9; Muller-Wille, op. cit. in note 9. Willroth, op. cit. in note 9; Muller-Wille, op. cit. in note 9. 13 D. Meier, Die wikingerzeitliche Siedlung von Kosel (Kosel-West), Kreis Rendsburg-Eckernforde (Offa­ Bucher, 76, Neumunster, 1994); D. Meier, 'Die wikingerzeitliche Siedlung mit zugeharigem Graberfeld von Kosel-Ost: Ein Beispiel aus dem Umland von Hedeby', 263-79 in L. Larsson and B. Hirdh (eds.), Centrala Platser, Centrala Fragor: Samhallsstrukturen under Jiirnaldern (Acta Archaeologica Lundensia Series in 8°,28, Stockholm, 1998) ; J. Kuhn, 'Eine Siedlung des fruhen und des hohen Mittelalters bei Schuby (Kreis Schleswig­ Flensburg)', Bericht Romisch-Germanischen Komm., 67 (1986),479-89. 14 H. jans, 'Schuby und Suderschmedeby: Zwei spatkaiserzeitliche Eisengewinnungszentren am Heerweg', Of(a, 56 (1999),67-80; A. Dobat is currently working on a thesis (Kiel University) entitled 'Die Schlei. Eine maritime Kulturlandschaft der Wikingerzeit: Eine historisch-topographische Studie', registering sites of the Iron Age and Viking Age on the shores of the Schleifjord. 15 Meier (1994), op. cit. in note 13; Meier (1998), op. cit. in note 13; M. Muller-Wille, 'Fruhstadtische Siedlungen und ihr Umland. Beispiele Haithabu und Ribe', 226-236 in H. Brachmann and j. Herrmann (eds.), Fruhgeschichte der europaischen Stadt: Voraussetzungen und Grundlagen (Berlin, 1991); M. Muller-Wille, 'Archaologische Untersuchungen landlicher Siedlungen der Wikingerzeit im Umland des fruhstadtischen Handelsplatzes Hedeby/Haithabu', Acta Praehistorica et Archaeologica, 26127 (199~1995), 39- 56. 11

12

36

LAND, SEA AND HOME

Rhenish pottery, metal objects and objects of other materials, coins, and beads of glass, carnelian and amber. I6 The distribution of soapstone vessels is almost entirely restricted to settlements near Hedeby and near the Schleifjord. There are very few single finds. As was the case in the area surrounding Ribe, whetstones are present in large numbers. In addition to the finds from the environs of Hedeby and the Schlei there are finds from the region of Angeln. Most of the whetstones originate from settlements, but some have been found in graves. The distribution of quernstones is again restricted to settlements near Hedeby and along the Schleifjord (Fig. 5.4). There is no doubt that research involving both systematic and comprehensive survey work and excavations, focused on the eastern parts of the Schleswig region, namely the hinterland of Hedeby, would change the present picture. Currently the missing sites on the banks or near the banks of the Schlei are particularly noticeable. Jens Ulriksen has shown by his investigations of the Roskilde fjord area that a number of landing places played an important role in communication between centre and hinterland. It is most likely that a comprehensive survey of the environs of the Schlei fjord, would bring the same results. I? The GrofS Stromkendorf site, situated on the shores of the Baltic Sea, on the eastern side of the Wismar Bay, south-east of the Isle of Poel, was first discovered in the I930S, when the collection of surface finds hinted at the existence of a settlement of the early Slavonic period. Throughout the rest of the 20th century more surface finds emerged and trial excavations in I950 and in the late I980s revealed more finds and settlement structures. The finds assemblages then showed a strong connexion with trade and production. After the discovery of a contemporary cemetery nearby in I993 a research project was launched and between I995 and I999 the whole of the cemetery site and about IO% of the supposed 20 hectares of the settlement site was excavated (Fig. 4.4)Y The cemetery was situated to the north of the site. A total of 258 burials, cremations and inhumations, were uncovered. The great variation of burial customs, comprising Scandinavian boat-graves and Saxon/Frisian animal graves, show different cultural influences. The settlement site at GroB Stromkendorf has also produced a finds assemblage that reflects widespread contacts, as most of the imported material could be assigned to two main areas of origin - Scandinavia and the West. Apart from the imported goods there is massive evidence of on-site manufacture that sheds light on a wide 16 The list of sites which have yielded imported material was compiled by L. Hansen in a seminar paper entitled 'Das Umland von Haithabu' (University of Kiel, summer semester 1999). 17 j. Ulriksen, Anl0bspladser: Besejling og Bebyggelse i Danmark mel/em 200 og LIOO e. Kr. En Studie af S0fartens Pladser pd Baggrund af Unders0geiser i Roskilde Fjord (Roskilde, 1998). 18 H. jons, F. Lurh and M. Muller-Wille, 'Ausgrabungen auf dem fruhgeschichtlichen Seehandelsplarz von GroB Stromkendorf, Kr. Nordwestmecklenburg', Germania, 75 (1997), 193-221; H. jons, 'Erste Ergebnisse interdisziplinarer Untersuchungen eines fruhgeschichtlichen Seehandelsplatzes mit Graberfeld bei Grog Stro!TIkendorf, Nordwestmecklenburg', 109-32 in V. Kazakevicius, C. von Carnap-Bornheim, J. Hines and V. Zulkus (eds.), The Baits and their Neighbours in the Baltic Region, 400-800: East-West Contacts (Archaeologia Baltica, 4, Vilnius, 2000); A. Tummuscheit, 'GroB Stromkendorf - a market site of the eighth century on the Baltic Sea coast', 208-20 in T. Pestell and K. Ulmschneider (eds.), Markets in Early Medieval Europe: Trading and 'Productive Sites' (Macclesfield, 2003).

PROTO-URBAN CENTRES

37

range of crafts. In general the Grog Stromkendorf site has produced a finds assemblage that can easily be paralleled on other trading sites of Early-medieval times in the Baltic Sea region and beyond. Looking at the settlement layout first, major differences are detectable between different parts of the excavated areas. Sunken huts are the predominant, if not the only house-type. Single or double rows of houses, laid out in a grid-like system, are characteristic of the southern and central parts of the site. Meanwhile the settlement traces in the immediate surroundings and within the area of the burial ground show a strikingly different layout. There is a scattered distribution of houses and other settlement structures without any signs of an underlying system. By means of dendrochronology and stratigraphy it is possible to establish a framework of the site's chronology and internal development. Within the excavated areas the first building activities are detectable around A.D. 730. The early settlement comprised a scattering of sunken huts, which were distributed not only within the area that was later to be occupied by the burial ground, but all over the excavated areas. At around A.D. 760 the internal structure of the settlement changed markedly. Then the site was characterized by dense occupation, while in certain areas rows of buildings and wells indicate a planned layout and an authority in control. During the decades after A.D. 780 building activities reached their climax, and after one last phase of minor rebuilding at the beginning of the 9th century the settlement was abandoned. So far archaeological research into the hinterland of Grog Stromkendorf has been carried out on a limited scale, and our knowledge of the interaction between centre and hinterland is comparatively limited. At present a distribution map of early Slavonic finds and archaeological sites is exclusively based on surface finds and gives only a rough impression of the settlement patterns in the surrounding area (Fig. 6.r). In contrast to the situation described in the hinterlands of Ribe and Hedeby, the archaeological record does not provide us with detailed insights into the environs of Grog Stromkendorf. It may solely be due to the state of research that the area has not yielded any imported material so far. 19 From A.D. 760 onwards the evidence of the settlement layout at Grog Stromkendorf suggests that there was a strong influence of a political leader. Apart from the evidence of the surface finds three fortresses are known in the vicinity of the trading centre. These are situated at How (two fortresses) and at Mecklenburg. In view of their proximity to the trading settlement all three fortresses could be considered as associated centres and seats of a political authority. As the dates of origin of all the fortresses are uncertain direct links are difficult to establish, but it is certain that Slavonic fortresses in the area came into existence no later than the middle of the 8th century. At the moment the archaeological evidence does not allow us to draw a coherent picture of the hinterlands of Grog Stromkendorf. It is, however, almost certain that the remains of the trading site are identical with the empori[umJ quod in oceani litore 19 R. Schmitz, 'Das Umfeld von Grog Stromkendorf, Lkr. Nordwestmecklenburg, in friihslawischer Zeit und die Grabung am Burgwall !low, Lkr. Nordwestmecklenburg', Fdpl. 2 (unpubl. M.A. thesis, Kiel University,

I 999) ·

38

LAND, SEA AND HOME

FIGURE 6. GroS Stromkendorf/Reric.

I: Distribution of finds in the hinterland (after Schmitz, op. cit. in note 19).

2: The political situation in the late 8th and 9th centuries.

PROTO-URBAN CENTRES

39

constitutum lingua Danorum Reric dicebatur ('the emporium located on the sea coast that is called Reric in the Danish tongue') which is known from the Frankish Annals. 20 Even though our knowledge of the local context of the site is limited, the written source sheds light on the wider political context in which GroB StromkendorfiReric must be seen. The account in the Frankish Annals shows that at the beginning of the 9th century Reric was at the centre of military confrontations between the Danish king and the overlord of the Slavonic Obodrites. Apart from the description of the events at the trading site itself, the document contains references to other Slavonic overlords and to the devastation of fortresses, indicating the existence of well-established political units in the vicinity of Reric. In general this was a period of political and military rivalry between the Frankish Empire, which was allied to the Obodrites, and the Danes (Fig. 6.2). The written evidence makes it quite clear that at the emporium of Reric Danish and Slavonic interests collided. As a result Godfred destroyed the trading centre and moved the merchants to Hedeby, which was probably under closer control. When the three proto-urban centres Ribe, Hedeby and GroB Stromkendorf are compared, the hinterland of Ribe stands out most clearly. This is due to the fact that a large number of settlements have been surveyed or excavated in this area. The hinterlands of Hedeby are reflected in the distribution of settlement finds, too, but a comprehensive survey, concentrating on the banks of the Schleifjord in particular, would probably increase the number of both settlements and cemetery sites considerably. Regarding the last example, RericlGroB Stromkendorf, it is now a question of future investigations to reveal more of the archaeological evidence which forms the backdrop of the central place. Yet, it is obvious that all three sites were integrated into local networks consisting of various settlement types serving different functions. 21 Prof. Dr Michael Muller-Wille, Institut fur Ur- und Fruhgeschichte der Christian-Albrechts­ Universitiit, D 24098 Kiel, Germany mmuellerwille@ufg·uni-kiel.de Astrid Tummuscheit, Archaologisches Landsmuseum, Landesamt fur Bodendenkmalpflege Mecklenburg- Vorpommern, Schloss Wiligrad, D I9069 Lubstorf, Germany [email protected]

20 The Royal Frankish Annals, s.a. 808. Quoted from R. Rau (ed.), QueLlen der Karolingischen Reichsge­ schichte, Teil 1 (Darmstadt, 1955),88. 21 Cf. H. Hamerow, 'Angles, Saxons and Anglo-Saxons: rural centres, trade and production', Studien zur Sachsenforschung, 13 (1999), 189-205.

VIKING-PERIOD PRE-URBAN

SETTLEMENTS IN RUSSIA AND FINDS OF

ARTEFACTS OF SCANDINAVIAN

CHARACTER

By

TAMARA PUSHKINA

Questions concerning the origins of the Old Russian towns have been a long­ standing problem in Russian historical scholarship. Different authors at various times have made a range of suggestions of what the immediate reasons for the emergence of the towns may have been: a good position on international trade routes; the rapid growth of the agricultural hinterland; the initiative of a ruler; the need for a market place for different tribes; the trading activity of Scandinavian merchants; and more. 1 Scholars have recently also turned their attention to the Scandinavian antiquities found in the territory of Old Russia. That there was occupation by Scandinavians is recognized in general terms, and this is attributed primarily to Scandinavian warriors, as members of a princely retinue. 2 Such a view, however, scarcely touches upon the subject of Scandinavian participation in the development of handicraft, trade, or the emergence of the towns. It is clear that the historical causes of the appearance and formation of the towns may vary greatly. None the less the majority of authorities agree with the view that a town defined as the economic centre of some large agricultural region is not to be found earlier than the beginning of the lIth century. At that time towns such as Ladoga, Novgorod, Smolensk and some others became the centres or the capitals of the principalities and territories. But in the loth century we find pairs of settlements existing side-by-side, one of which develops into a true town while the other loses its function as a trade and production centre and develops into a small feudal estate centre. Pairs of settlements of this kind are well known in Old Russia. They include

1

E. N. Nosov, 'The problem of the emergence of early urban centers in Nothern Russia', 236-56 in

J. Chapman and P. Dolukhanov (eds.), Cultural Transformation and interaction in Eastern Europe (Aldershot,

1993); B. O. KAK)Klle ceAhCKO"O HaceAeHH>! /lpellHeii PYCH', II7 in ApxeOJlOzu'tecKuu C60puUK (TpYllh' rocYllapcrneHHoro HCI'OpH4eCKol'o MY:le>!, 40, Moscow, 1966) [V. P. Levasheva, 'On the clothing of the rural population in Old Russia', from Archaeological DictionaryJ ; A . A. rOAy6ella, 'Mep>!', 80 and 361 in B. B. Celloll (ed.), cPu""O-YZpt>l U 6(lJtmh18 Jnoxy cp eihteeeKOllbR (Moscow, 1987) [L A. Golubeva, 'Merya', in V. V. Sedov, Finno-Ugrians and Baits in the Middle A gesJ.

PRE-URBAN SETTLEMENTS IN RUSSIA

43

The category of Scandinavian amulets and ritual items comprises Thor's­ hammer rings, pendants in the shape of miniature weapons, strike-a-lights, pendants in the form of human (male) faces, and valkyrie figures. Such amulets could not have been objects of trade even if the pagan religions of the Slavs, the BaIts and the Vikings were similar. Finds of everyday objects are not very numerous, with the exception of one-sided antler combs. Finds of tools such as smithing and jewellery-making tools, or clay and stone moulds for casting Scandinavian-type jewellery, are both highly interesting and very important. Finally there are some finds of runic inscriptions, rune-like signs, and possibly even magical symbols, engraved on various objects. 8 Scandinavian antiquities of the Viking Period have been found at about 150 locations in Old Russia. These sites lie over a wide area from Lake Ladoga in the North to the Lower Dnepr (Dnepr rapids) and the River Don in the South, and from Volyn' (western Ukraine) in the West to the area of the Kama River in the East (Fig. 1). The majority of the finds represent burials from the end of the 9th and the loth centuries. Known settlements of the same date are fewer in number, and only a small proportion of them have been examined scientifically. The earliest finds are from the settlement of Staraja (Old) Ladoga, the Aldeigjuborg of the Norse sagas. These date back to the second half of the 8th century. The latest finds occur in cultural layers of the IIth century at Novgorod and Suzdal. 9 This large corpus of material with such a strong representation of female objects cannot have been connected exclusively with warriors and traders. Rather it must indicate a fairly large-scale immigration of whole families from Scandinavia (mainly from central Sweden) who were strong enough to preserve their distinctive character­ istics for a couple of generations. lO The distribution of these finds is irregular. Some archaeological sites have one or two objects or burials; others have the whole series of artefacts and evidence for their local production in some cases, or groups of Scandinavian burials, or houses of a characteristically Scandinavian type. Most of the archaeological sites with Scandinavian finds are situated in the region connected to the most important routeways of eastern Europe: the Baltic­ Caspian route via the Rivers Volkhov and Volga, and the Baltic-Black Sea route via the Volkhov and the Dnepr. The majority of the hoards of Arabic coins of the 9th-loth centuries and of finds of trade equipment are from this area too. The oldest Russian towns mentioned in the Russian Primary Chronicle (Ladoga, Novgorod, Smolensk, Kiev, etc.) appeared and functioned in the same area on the banks of these rIvers. 8 A. Stalsberg, 'The Scandinavian Viking Age finds in Rus': overview and analysis', 448-71 in J. Weisbecker (ed.), Oldenburg-Wolin-Staraja Ladoga- Novgorod-Kiev: Handel und Handelsverbindungen im siidlichen und ostlichen Ostseeraum wiihrend des friihen Mittelalters (Frankfurt am Main, 1988); E. A. Melnikova, 'The eastern world of the Vikings' 81-9, tab. 2, figs. 3-5 in L. Lonnroth (ed.), Gothenburg Old Norse Studies I (Goteborg, 1996). 9 A. Stalsberg, op. cit. in note 8, 453-8; T. A. Pushkina, 'Scandinavian finds from Old Russia: a survey of the topography and chronology', 83-92. in P. Hasson (ed.), The Rural Viking in Russia and Sweden (Orebro, 1;],97). o l. Jansson, 'Warfare, trade or colonization? Some general remarks on the eastern expansion of the Scandinavians in the Viking period', in Hasson (ed.), op. cit. in note 9, at pp. 27 and 51-5.

44

LAND, SEA AND HOME

Scandinavian finds in Old Russia (e) . (I) Staraja Ladoga; (2) Ryurik Gorodishch; (3) Gnezdovo [6 sites]; (4) burials mounds of the Ladoga Lake region [24 sites]; (5) Pskov district [25 sites]; (6) Jaroslavl' district [4 sites]; (7) burial mounds of the Vladimir-Suzdal' district [10 sites]. After Stalsberg, op. cit. in note 8, with additions by the author. Plain dotted line: Baltic-Caspian route; wiggly dotted line: Baltic-Black Sea route.

FIGURE 1.

PRE-URBAN SETTLEMENTS IN RUSSIA

45

There are three clusters in the map of Scandinavian finds: distinctive Scandinav­ ian concentrations dependent upon the water roads of Old Russia. The first of these clusters is situated at the beginning of the route from the Baltic Sea to the Volga and the Dnepr. It lies on the lower River Volkhov and includes Staraja Ladoga and its hinterland. Scandinavian finds are known here from the middle of the 8th century to the loth century. The Scandinavian material culture is represented not only in jewellery and amulets but also by many everyday objects such as tools, and in some elements of the layout and structure of houses. Unfinished jewellery, antler combs and iron items indicate that some of the objects of Scandinavian character and type were produced here. According to the archaeological evidence, Scandinavians were a constant element in the population of Staraja Ladoga from its very inception. l l The second cluster is situated in the upper reaches of the River Volkhov on the shore of Lake Il'men near Novgorod the Great. This includes Ryurik Gorodishche (the hill-fort of Ryurik). Here, there are about 50 settlements and cemeteries dating from the 9th and loth centuries. This region was densely populated during the period in question. Yevgeny Nosov, the director of the investigations of Ryurik Gorodishche, has proposed that this hill-fort was a non-agrarian centre and the predecessor of Novgorod the Great as an administrative, trading and craft-production centre of the network of Slavonic rural settlements from the 9th century to the second half of the IOthY Novgorod became a town in the second half of the IOth century and later developed into the capital of the feudal republic, while Ryurik Gorodishche became a prince's residence beyond the territory of the town. In the present author's opinion, a 'Scandinavian veil' characterizes the period in which Ryurik Gorodishche flourished, when there was a princely retinue based there and collecting tribute, in addition to the Slavonic dite. The Scandinavian finds (male and female dress accessories and jewellery, amulets and everyday objects) belong mainly to the loth centuryY The other Scandinavian finds of the loth century belong to two or three sites on the banks of the River Volkhov, while only a few Scandinavian artefacts were found in Novgorod in the culture layer of the second half of the loth and first half of the IIth centuries. The third concentration of Scandinavian finds lies south of Ryurik Gorodishche in the Upper Dnepr region near Smolensk. This region presents a different picture of rural settlement in the 9th-IOth century from that in the Novgorod area (Fig. 2). Archaeological sites of this period are very sparse here. This implies that the population was low and that there was no developed agricultural hinterland. There are only a few groups of so-called long barrows that represent the local, mixed Balto­ Slavonic population, and round Slavonic mounds. Contemporary settlements are also 11 I. Jansson, op. cit. in note 10, 28-3 1; O. I. Davidan, 'Contacts between Staraja Ladoga and Scandinavia', in Hannestad et al. (eds.), op. cit. in note 5; idem, 'Kunsthandwerkliche Gegenstiinde des 8. bis IO. Jahrhunderts aus Alt-Ladoga (Die Sammlung der Staatlischen Ermitage in St. Petersburg)" Zeitschrift fur Archiiologie des Mittelalters, 20 (1992), 5-61. 12 E. N . Nosov, 'Ryurik Gorodishche and the settlements to the North of Lake limen", 5-66 in M. A. Brisbane (ed.), The Archaeology of Nov go rod, Russia (Soc. Medieval Archaeol. Mon ., 13, Lincoln, 1992), at pp. 58- 9. 13 Nosov, op. cit. in note 12,46-55 and 59; Jansson, op. cit. in note IO, 35.

46

LAND, SEA AND HOME

Archaeological sites of the 9th-13th centuries in the Smolensk region.

Black dots: 9th-IIth centuries; stippled dots: lIth-13th centuries.

FIGURE 2.

known, but these are isolated. The cemeteries usually consist of 10-30 mounds and the settlements are less than 2 ha in area. Cremation is the uniform burial rite. There are no traces of social and professional differentiation within the communities that buried in these cemeteries. They have a casual character. The quantity of finds of eastern coins and trade equipment is even lower. The majority of Scandinavian finds, including Scandinavian-type burials, are concentrated in Gnezdovo (Fig. 3). Modern Gnezdovo is a small village situated 13-15 km west of the present city of Smolensk on the banks of the River Dnepr. It is mentioned by name only twice in early documents, of the 14th and 17th centuries respectively, and nothing earlier. Smolensk is one of the earliest Russian towns, which, according to the Russian Primary Chronicle, was captured by Oleg (a Scandinavian) on his way to Kiev in A.D. 882. However, the oldest excavated layers in Smolensk are dated no earlier than

PRE-URBAN SETTLEMENTS IN RUSSIA

47

FIGURE 3. Scandinavian finds in the Upper Dnepr region (see Fig. r:(3) ). Filled circles: 9th­ century hoards; open circles: roth-century hoards; halved circles: Arabic coins; cross-in-circle: trading equipment; pendant triangles: weaponry; standing triangles: jewellery; T-shapes: amulets; rhombuses (vertical): Scandinavian burial rites; squares: evidence of local Scandinavian craft; rhombuses (horizontal): miscellaneous finds.

48

LAND, SEA AND HOME

FIGURE

4.

The Gnezdovo complex. Dotted area: the burial mounds; hatched area: open settlement zones; black square: the hill-fort.

the second half of the lIth century. Subsequently, in the 12th century, Smolensk is known as a capital of the principality.14 On the other side of the river, at Gnezdovo, there is a large area of settlement and burial mounds that relate to the Middle Viking Period. The Gnezdovo archaeological complex consists of two known settlements and about 3,000 barrows, which are organized into eight groups, or cemeteries, around the settlements. The central part of this complex of remains is governed by a hill-fort surrounded by an open settlement covering some 16 ha. Two cemeteries comprising more than 2,300 barrows lie to the West, North and East of the settlement (Fig. 4). The second part of the complex consists of a small settlement 3 km to the West of the central part, with about 400 burial mounds close by. No settlements associated with the remaining cemeteries have yet been located, but we can postulate their existence. The central settlement came into being no later than at the turn of the 9th to loth centuries. It was not very large to begin with. It was about 4 ha in area and had no 14 II. VI . ACTalll0lla, ApeB1tuii C/>lOlte1tCK do 1ta'I"Y"IKHH, 'CAaIIJIHe BOCfO'lHOH EllpOllhl HaKaHYHe ofipa:lOlIaHH>I !\pellHepYCCKoro I'Oc y/\apCl'lla', MameputlJlbl u UCCltOe06anUH no apxeOlWZUu CCCP, 152 (1968), 85-6 and II1-I2 [I. I. Ljapushkin, 'Slavonic Eastern Europe on the eve of the formation of the Old Russian state'].

PRE-URBAN SETTLEMENTS IN RUSSIA

51

pottery show that there were also central European and Baltic Slavs within the population. However, the majority of the pottery has parallels at the Slavonic sites of the Middle Dnepr region of the southern Rus' .22 The sets and typical shapes of female jewellery, the characteristic amulets, burial customs, some features of the weaponry and objects of everyday life, and certain details of handicraft practices, especially in iron working and jewellery production, prove the presence of Scandinavians here. The social and professional structure of the whole population of Gnezdovo and that of the Scandinavians there were not the same. A few large and middle-sized burials mounds stand out amongst other Scandinavian-style barrows. These have double or single cremation or inhumation burials in chamber-graves, rich and varied grave goods including eastern or Byzantine imports, ritual cauldrons, swords and armour, and drinking horns. They seem to be the burials of members of the nobility of the military retinue. Some of the burials which do not differ in their furnishings from well-known burials of the Viking Period in Scandinavia belong to freemen of middle rank. These mounds contained only characteristic brooches, or Thor's­ hammer rings, together with everyday objects such as ice-skates or antler combs. The majority of the Scandinavian finds are attributable to the 10th century, especially to its later half. Scandinavians provided permanent and quite numerous residents of Gnezdovo. They seem to have been joining the Gnezdovo community on a continuous basis - it is otherwise difficult to explain the chronological and typological resemblances between the Scandinavian finds at Gnezdovo and those in Scandinavia, especially in central Sweden. It is generally accepted that the population of Gnezdovo was both ethnically and socially heterogeneous. This fact alone distinguishes Gnezdovo from other settlements in the region in the loth century. There is also some impressive evidence of highly advanced handicraft. Numerous finds of Arabic and Byzantine coins, and scales and weights, in the burials and in the settlement show that the majority of the population here participated in trade. On one hand, silver coins were not 'dead treasure' but were rather in active use. On the other hand more than 1,100 silver coins which composed the seven Gnezdovo hoards of the 10th century were a 'treasure' of the relevant social group. As a further element, traces of Christianity appeared in Gnezdovo earlier than at other sites of the Upper Dnepr region. This region was officially converted only in 1014, but crucifix-shaped pendants, wax candles and simple oriented inhumation graves appeared at Gnezdovo around the middle or in the third quarter of the loth century. The same phenomena do not occur in rural cemeteries before the IIth century.23 The problem of interpreting Gnezdovo is basically that of the historical and archaeological context of its development. The connexions between Gnezdovo 22 E. B. KaMeHe'\Ka>" '0 HeKOTOpl>'X THllax KepaMHKH rHe:lI\Olla', COllemCKan apxe01Wlun, 1 (1988),258-62 [E. V. Kamenechkaya, 'On some types of pottery at Gnezdovo'J. 23 N. Makarov, 'Far northern parts of ancient Russia on the their way to Christianity', in M . Muller-Wille (ed.) Rom und Byzanz im Norden: Mission und Glaubenswechsel im Ostseeraum wdhrend des 8.-14. Jahrhunderts, Bd. II (Mainz-Stuttgart, 1998),269.

52

LAND, SEA AND HOME

and the rural settlements of the Upper Onepr region are unclear. 24 There are only a few individual finds of the typical Gnezdovo knives and antler combs in the hinterland of the site. Meanwhile, some evidence of agricultural activity amongst the inhabitants of Gnezdovo has been found at the central settlement, in the form of two shares for ploughing implements, and scythes. Taking into account the dispersed distribution of contemporary archaeological sites in the hinterland, we may infer that these connexions could not have been very important to Gnezdovo. This implies that there was no agricultural basis for the development of the town. The main contacts that Gnezdovo had were rather of long-distance character. It found its place in the network of early urban or pre-urban centres in Russia and the Baltic region. As we know from the written sources, especially from the Russian Primary Chronicle and the work by Constantine Porphyrogenitus De administrando Imperio, in the lOth century these centres were controlled by the princely, Kievan retinue known as the Rus', which used the Norse language even in the mid-lOth century.25 In consequence it should be no surprise that the majority of Scandinavian finds are from Ryurik Gorodishche and Gnezdovo. It is possible to suggest that Gnezdovo was a residence of the princely retinue in the lOth century. Such stations for a military elite (Rus'), collecting tribute, were called pogost in the Old Russian tradition. The stormy growth of these centres is dated from the middle of the lOth century, after Princess Olga had founded some new pogosti in the Onepr, Oesna and Novgorod regions: in fact, new settlements similar to Gnezdovo and Ryurik Gorodishche appeared on the other principal rivers of the Old Russian territory. But this system of pogosti disappeared in the period of the end of the loth century and beginning of the lIth.26 The development of agricultural settlements then became one of the most important factors in the establishment and prosperity of towns. The roots and the economic strength of these towns lay in their agricultural hinterland. Pogosti or pre-urban centres such as Gnezdovo had no such roots. The absence of real connexions between Gnezdovo and its hinterland prevented the further development of this trading and productive settlement. It did not become a town. It is noteworthy that the Scandinavian features of the blacksmith's craft disappeared with the decline of Gnezdovo. The advanced technology of multi-layered iron-welding is unkown in Smolensk - the more primitive technology of the southern Russian smiths is dominant in that town from the IIth century onwards. Gnezdovo, however, did not utterly perish at the beginning of the lIth century: it transformed into a feudal estate centre, referred to in the documents of the 14th century as a residence of the service people near Smolensk. This was how Gnezdovo co-existed with Smolensk in the I Ith century and later. 24 T. A. Pushkina, 'The Gnezdovo settlement: a Viking Age settlement near the Upper Dnieper' , Gunneria, 64 (i?9 1 ),359- 6 4. KOHCTaHTHH EarpJlHOpOI\HhIH, 06 ynpa6Jteuuu uMnepueu (meKcm, nepe/loiJ, KOMMeumapuu) (Moscow, 1989), 45 - 51, Ch. 9, commentaries 6, I I and 13 [Constantinus Porphyrogenitus, De administrando imperio (Text,

Translation, Commentary)]. 26 B. $I. nerpYXHH, fla1{QJW3mUO-K)l'tbmytmoiiucmopuu Pyw IX-X 66. (Smolensk and Moscow, 1995), 154-70 and z65-8[V. Ja. Petrukhin, The Beginning of the Ethno-cultural History of Russia in the 9th-IOth Centuries].

PRE-URBAN SETTLEMENTS IN RUSSIA

53

The network of centres whose active life depended on the presence of the princely retinue and on participation in international trade was formed in Old Russia in the lOth century. The Scandinavians - warriors, craftsmen, merchants and women ­ were a considerable part of their permanent, free population. According to the archaeological evidence we have, there was no active expression of agricultural activity by the Scandinavians here. To no small degree, the Scandinavians ensured intensive craft-production and the development of centres such as Rurik Gorodishche and Gnezdovo. When these centres were superseded, their elements of Scandinavian culture, practically unknown on rural sites, also virtually disappeared in the towns which flourished in their VICInIty. Dr Tamara Push kina, Department of Archaeology, Faculty of History, Moscow State University, B-234 Moscow II9899, Russia [email protected]

RURAL SETTLEMENT AND LANDSCAPE

TRANSFORMATIONS IN NORTHERN

RUSSIA, A.D. 900-1300

By

NIKOLAJ MAKAROV

The Viking Age in Russia is known as the period of an incipient urbanization that shaped the system of town centres on the vast territories of the Russian plain. The origin of the most important medieval urban centres dates back to this period, and primarily to the late Viking Age, as very few towns display cultural deposits of the 9th or early 10th century. However, the emergence of the network of towns was not the only innovation that occurred around the turn of the 1st to the 2nd millennium A.D. No less important, in fact, was rural colonization and the formation of new settlement patterns in the countryside throughout Russia - from the Middle Dnieper in the South to Ladoga and Beloozero in the North. Within a markedly short period the cultural landscape of Russia passed through major transformations. Many distinctive features of the national rural landscape also have their origin in this period. Rural archaeology in Russia has always been in the shadow of urban studies. Our current knowledge of Viking-age rural settlement is based upon a number of local and regional surveys and excavation projects, none of which have ever been connected in the framework of a single programme. Field research on rural sites in Russia started in 1950S, approximately at the same time as the late John Hurst launched his famous excavations at Wharram Percy. One of the most important and influential books to inspire research interest in the archaeology of the rural settlement in Russia was Sedov's monograph on the rural sites of Smolensk area published in 1960.1 The most significant and impressive research projects dealing with medieval rural settlement were carried out in the 1970S and 80S in the n'men lake area in the centre of Novgorod land,z in the areas of Rostov and Suzdal in north-eastern Russia,3 1 B. B. Ce/lOIl,CeJtbCKue nOCeJUiUUR l{eumpaJtbnblX paU01WII c..WJtenCKOU .leMJtU(VllI-XV 6/1.) (MaTepHaAhl H HCCAe/lOIl­ aHH}l 110 apxeOAorHH CCCP, N 92, Moscow, 1960) [v. V. Sedov, Rural Settlement in the Central Part of Smolensk region (VIII-XV centuries A.D.)]. 2 E. II. 1I0COII, 'ApXeOAOrHI)Kecma 13XV UI\. H "POHCXO)KI\eHHe nOAocTHOH 06U\HHhl', CO(!emcKaR apxeOJIOZUR, I (1991), 112-33 [S. Z. Chernov, 'The historical landscape of the Radonezch area: origin and semantics', Cultural Monuments: N ew Discoveries; 'Archaeological data on the internal colonization of Moscow principality in the 13th to 15th centuries and the origin of the "volost" commune', Soviet Archaeology]. 4

naMRm1orolfsson and Guoni Jon sson (eds.), Vest{iroinga sogur (islenzk fornrit VI, Reykjavik, 1943), 19. 9

10

27 6

LAND, SEA AND HOME

N

Stream (/cekr)

Srehol Roll

FIGURE 2.

Two farms separated by a stream.

this chapter of the saga. That Holl farm is shown on raised ground is probably due to its name, which in both Old and Modern Icelandic translates as hillock. However, in chapter r6 of GisLa saga Surssonar the stream is noted in the following way: 'gengr hann sioan til1rekjar pess, er fellr a milli bcejanna ok tekit var neytingarvatn af hvarumtveggja bcenum' [he [Gisli] then went to the stream, which flowed between the farms and was used for drawing water for both farms].l1 Figure 2 presents my approximation of the scene in chapter r6, in which the farms are separated by the stream. Chapter r6 of the saga does not comment on the relative positions of the farms, and chapter 5 does not note the stream. So, there are insufficient shared reference points between the descriptions in the two chapters to accurately visualize the textual stream in relation to the farms, or draw any conclusions on the possible location of an actual stream. It is therefore extremely difficult to interpret the saga's description of this part of Haukadalur as a literal representation of the physical world. However, a literary reading of the saga's geography is not bound by what may actually be present in Iceland's landscape, and so is not compelled to attempt to resolve any discrepancies that may exist between the textual and physical topo­ graphies. In literary terms, the importance of this stream is not its precise location but its narrative function. My research on saga landscapes, and that of Paul 5chach, has shown that saga landscape features often function as literary devices which the saga author may employ to direct the narrativeY This stream is a case in point. GisLa saga Ibid., 52. , Wyatt, op. cit. in note I ; Paul Schach, 'The Use of Scenery in the lslendingasogur' (unpubl. Ph.D. thesis, University of Pennsylvania, 1949). 11

12

ICELANDIC FAMILY SAGAS

277

Surssonar is an outlaw tale of tragic poignancy, which illustrates the loyalties of kinship and friendship being tested, distorted and broken. A man called Vesteinn Vesteinsson, who is a close friend and brother-in-law of the eponymous Gisli, is killed by I>orgrimr I>orsteinsson. I>orgrimr is in turn murdered by Gisli in vengeance for Vesteinn's killing. Gisli is outlawed for his action, and the story of his survival as an outlaw forms a large part of the narrative. Prior to I>orgrimr's killing the saga tells us that it has been snowing, and that Gisli leaves H6ll along a path used for drawing water from the stream, and wades along the stream. He then goes to Sceb6l along the path used there for collecting water, kills I>orgrimr and returns along the same route. So, no tracks are left in the snow to identify the killer as having travelled from H6ll. The narrative function of the stream is to conceal the identity of the killer, and in this way operates as a retardation device in order to allow further narrative developments; it is the literary function of the stream that is of fundamental importance to the structure of the saga not its actual geography.13 By concealing the identity of I>orgrimr's killer in this way the saga author creates the narrative opportunity to illustrate the tension of personal and social loyalties which lie at the core of the saga. Immediately after the murder is discovered a search party is sent out, and Gisli's farm is searched, but his brother I>orkell is a member of the search party and is able to kick the evidence of Gisli's frozen shoes out of sight before anyone else sees them. However, Gisli's brother is then reluctant to be any more involved than is necessary, because he fears for his own life. Gisli's sister 1>6rdis was I>orgrimr's wife, and when she overhears Gisli admitting to the murder in a verse she informs I>orgrimr's brother Barkr. The narrative now follows Barkr's pursuit of Gisli. Although the use of streams or rivers as retardation devices is relatively common in the sagas, their function in concealment scenes is unusual. 14 Woods are the features most commonly associated with concealment, or more specifically sk6gr. Indeed, I have found that 50 per cent of scenes in the fslendingasogur that contain a sk6gr reference employ this feature as a concealment device.15 For example, in chapter 27 of Heioarviga saga Baroi Guomundarson and his men hide in a wood prior to attacking Gisli I>orgautsson: I>a var skogr' mikill i Hvitarsi8u, sem pa var vi8a her a landi. En peir satu fyrir ofan skoginn b sex ok sa gorva ti8endi a Gullteig. Bar8i var i skoginum C ok snertu eina fra peim, par er peir sl6gu, ok peir sex samano Nu hyggr Bar8i at, hve margir menn vxri at slxttinum; nu pykkisk hann eigi vita vist, hvart kona er inn pri8i ma8rinn, er hvitt er til hofu8sins at sja, - 'e8a mun par vera Gisli?' Ganga nu ofan undan sk6ginum d hverr eptir o8rum, ok p6tti peim sva fyrst I>orgauts sonum, sem einn ma8r gengi par, ok tekr hann til or8a l>orm68r, er si8ast sl6 a teignum: 'Menn fara pama,' kva8 hann. 'En mer synisk,' segir Gisli, 'at ma8r gangi einn.' En peir gengu hart ok runnu eigi. 16 Ed. cit. in note 10, 52-3. Wyatt, op. cit. in note I, 128-73. , 15 Other woodland terms found in the lslendingasogur are: holt [copse/hillock], lundr [grove], mork [forest] and vior [forest/timber]. These terms primarily represent differing sizes of woodland, usually birch. See Wyatt, , op. cit. in note I, 85. 16 Heioarviga saga, in Sigurour Nordal and Guoni Jonsson (eds.), Borgfiroinga sogur (Islenzk fornrit III, Reykjavik, 1938),294. 13

14

27 8

LAND, SEA AND HOME

[There was a large wood in Hvira.rsioa, as there then was widely here in this country. Six of them sat above the wood and saw clearly the happenings in Gullteig. Baroi was in the wood with five others and only a short distance from those who were mowing. Now Baroi considered how many men were mowing. He was not certain whether the third person, whose head was white to the naked eye, was a man or a woman: 'or could that be Gisli there?' They now came down out of the wood, one after the other, and it seemed to the 1>orgautssons that a single man was walking there, and 1>ormoor began to speak, who was the last one to be mowing in the meadow: 'Men are travelling there,' he said. 'It looks to me,' said Gisli, 'that it is a single man walking.' And they walked briskly but did not run.]

Heioarviga saga contains five instances of sk6gr, four of which refer to this single wood (a to d), and no other woodland references at all. The opening woodland reference in this scene (a) functions to establish the location, and to historicize the locative referent by commenting on the extensive forestation of Iceland during the period when the story is set. Moreover, the author's statement that there was once a large wood at Hvltarstoa coupled with his assertion that such woods were common in Iceland at the time of the events he describes may suggest that there was not a wood at Hvirarsioa at the time he was writing. Sk6gr references (b) and (c) conceal the two parts of Baroi's company: (b) obscures the reserve party from Gisli while leaving them in a position to view the action, and (c) provides the cover necessary for Baroi to approach and watch Gisli from quite close range before attacking. Reference (d) marks a shift of narrative focus from the approach of Baroi to the response of Gisli. We are now looking towards the wood from Gtsli's viewpoint, rather than at Gisli from within the wood, but because Baroi and his men are moving in single file it appears as though only one man has emerged from the sk6gr. So, even though the narrative perspective has changed, the sk6gr continues to function as a concealing feature. At no point does the author explicitly state that Baroi and his men are hidden from view or that they are deliberately hiding - this is simply understood in context through the use of sk6gr. From a structural perspective, Baroi's victory in the battle on the heath, which follows this scene, requires a number of narrative conditions: his attack on Gisli must be successful (he must kill Gisli, but some men have to escape to raise the alarm: c and d), and his opponents should not be aware of the size of his company (b). Both conditions are fulfilled through the agency of the wood. In order to appreciate the impact of this use of sk6gr on the wider narrative we may, for a moment, imagine the scene without the wood. If Baroi were to attack down an open hillside Gisli would see his approach and could well escape. If this had been the case, Baroi might have been met by a larger force and could have lost the battle on the heath, which would have brought the saga to an early and unsatisfactory close. It seems clear, therefore, that this small detail has a greater impact on the narrative than that of simply identifying a location. Furthermore, Baroi's strike against Gisli avenges the killing of a man called Hallr, who was killed in a wood, and so the sk6gr may also be said to form a structural counterpoint to the murder of Hallr. The iF editors of Heioarviga saga cite an extract from the Arb6k Hins islenzka fornleifafeLags of 1884-5 noting that the hillside on which Baroi hides is made up of a

ICELANDIC FAMILY SAGAS

279

series of small ridges behind which it would be possible to lie undiscovered, presumably even without a wood. 1? However, Baroi would still not have been able to approach Gisli undetected without the wood. The extract further notes that there is no wood there now - meaning 1885. The Arb6k Hins islenzka fornleifafelags study is a product of the 19th-century historical reading of sagas, and it is interesting to see that elements of this same historical perspective have a prominent place 53 years later in the iF commentary on the saga's topography rather than any literary consideration of such features. By contrast, Sigurour Nordal's 1940 seminal essay argues that sagas may not be historically factual documents, and focuses on Hrafnkels saga to demonstrate his argument. Nordal is unequivocal in his views regarding the presumed historicity of this saga, concluding that it 'was the work of a single author whose purpose was not to narrate a true story but to compose a work of fiction'.18 A further conclusion of Nordal's research on Hrafnkels saga is to warn that 'much of the belief in the truthfulness of the Icelandic sagas is based on reasons which, if closely scrutinized, would turn out to be as untenable as those once used to support the veracity of Hrafnkels saga'.19 Nordal highlights a belief in the accuracy of the saga's topography as an example of 'untenable' evidence of historicity; he writes: Apparently he [the saga author] never visited Hrafnkelsdalr itself, for everything that the saga says about its topography is either wrong or based on guesswork and the vague descriptions of other men's reports ... 20

Sigurour Nordal's controversial conclusions on the topographic inaccuracy of Hrafnkels saga, and by implication other islendingasogur, had profound political ramifications: to question the place of the sagas was (in 1940) to question the very bedrock of Icelandic culture, and the importance of sagas to Icelandic nationalism. The debate between sagas as history or fiction, and the political implications of this debate is explored by Jesse Byock, who discusses the role of the islendingasogur in the development of Icelandic nationalism and illustrates the pivotal role of Sigurour Nordal's essay for the nationalist position: In this climate [Iceland's desire for liberation from Denmark], nationalism spilled over into analyses of the national treasure, the family sagas. The problem facing Icelandic intellectuals was how to lift the sagas from their status as traditions of unlettered storytellers and elevate them to the rank of world literature. 21

By defining the fslendingasogur as literary fiction rather than as records of oral history Sigurour Nordal helped to bring about a fundamental change in the perception of the sagas. However, a fierce debate ensued between the supporters of this new view and

Ibid., 294 n. 2; Arb6k Hins islenzka fornleifa(elags (Reykjavik, 1884-5), I3I. Op. cit. in note 7. English quotations from Sigurour Nordal's Hrafnkatla are taken from R. George Thomas (trans.), Hrafnkels saga Freysgooa: A Study (Cardiff, 1958), 57. 19 Thomas (trans.), op. cit. in note 18,60. 20 Ibid., 58. 21 J. Byock, 'Modern nationalism and the medieval sagas', 163-87 in Andrew Wawn (ed.), Northern Antiquity: The Post-Medieval Reception of Edda and Saga (Enfield Lock, 1994), 176. 17 18

280

LAND, SEA AND HOME

the traditionalists, who wished to continue to read the sagas as history.22 Some of the traditionalists were landowners with a personal stake in reading the sagas as history, because they lived and worked the land identified in the sagas, and may well have considered themselves as having a direct link with the past and the sagas through the land.23 The relationship between land and saga is also brought to the fore by Jon Karl Helgason in his examination of the role of the Alpingi in regulating the editing and printing of Old Icelandic texts. His discussion is primarily concerned with the parliamentary debates surrounding the state-sponsored 1944 edition of Njats saga. 24 Jon Karl notes that 'the three members of the Alpingi who were officially responsible for the proposal [a state-supported edition] all lived in the district of Rangarvallasysla, in which a substantial part of Nj!tls saga takes place'.25 Gisli Sigurosson makes a similar observation on the relationship between land, the sagas and the Icelandic sense of historical continuity: The events of many of the medieval sagas take place in Iceland, and are thus set in the same landscape as that which people still have before their eyes today. Sagas were (and still are) linked to the land in the minds of people. 26

Some scholars continue to perceive geographic accuracy in the landscape descriptions of sagas. For example, the historian Sveinbj6rn Rafnsson has studied the geography and geology of Hrafnkelsdalur, and compares the topography of the actual valley with that presented in the saga, and says that it 'has been asserted that the saga does not describe accurately the landscape of the valley and its surroundings. This is unfounded. There are no grounds for assuming that the author or the original audience were unacquainted with the scenery of the saga'.27 Perhaps the most extreme attempt to correlate the landscape of a saga with the geography of Iceland may be found in the 1991 IF edition of Baroar saga Smefellsass. 28 This saga is a biography of Baror Dumbsson, who is said to have acquired supernatural skills and near god-like status, and also relates the adventures of his son Gestr. In chapter 8 of the saga a man called Ingjaldr is tricked into rowing far out to sea by a troll-woman (Hetta) thus putting his life in danger. This scene is accompanied 22 In essence, the traditional view, the 'free-prose' school, was that the sagas were historically reliable texts based on oral sources. By contrast, the 'book-prose' view of Sigurour Nordal, and others, was that the sagas were historically unreliable works of literature (see, for example, Byock, op. cit. in note 21,180). For discussion of the impact of this debate on Icelandic approaches to archaeology see Adolf Frioriksson, op. cit. in note 6, 8- 12. 23 Byock's article (op. cit. in note 21) is a useful summary of the role of the fslendingasogur in the development of Icelandic nationalism. He also identifies the proponents of the book-prose view (e.g. Sigurour Nordal) with the growing urban population, suggesting that the book-prose/free-prose debate represents a debate between the urban and rural views of the Icelandic self-image. The article is also something of a plea for not dismissing the sagas as a potential source of social history when discounting their historicity in absolute terms. 24 Alpingi is the Icelandic parliament. Jon Karl Helgason, '''We Who Cherish Njtils saga": the Alpingi as literary patron', 143-61 in Wawn (ed.), op. cit. in note 21; Magnus Finnbogason (ed.), Njals saga (Reykjavik,

1~44).

Jon Karl Helgason, op. cit. in note 24,152. Gisli Sigurosson, 'Icelandic national identity: from romanticism to tourism', 41-75 in Perrti J. Anttonen (cd.), Making Europe in Nordic Contexts (Turku, 1996),43. 27 Sveinbjorn Rafnsson, Byggoaleifar i Hrafnkelsdal og a Bruardolum (Reykjavik, 1990), 108. , 28 Baroar saga Sntefellsass, in I>orhallur Vilmundarson and Bjarni Vilhjalmsson (eds.), Haroar saga (Islenzk fornrit XIII, Reykjavik, 1991). 26

ICELANDIC FAMILY SAGAS

FIGURE 3.

281

Bar3r's rescue of Ingjaldr. From iF XIII (op. cit. in note 28), map opp. p. 128. (Reproduced with permission of Hi3 islenzka fornritafelag)

by some supernatural elements, such as an unusually swift deterioration of the weather, and the sudden appearance to Ingjaldr - out at sea - of another figure called Grimr, who is thought to be either the god 6ainn or 1>6rr. Ingjaldr is rescued by the eponymous Barar and Grimr disappears from the scene. This scene is one of three in which Barar is shown as a positive force, coming to the aid of someone in need. The iF editors cite 6lafur Larusson's suggestion that Barar's rescue of Ingjaldr is based on a miracle story of Bishop l>orlakr. 29 Despite identifying a possible literary source for this scene, the editors include a map which makes use of a selection of the features in the actual landscape of Iceland to identify Ingjaldr's probable position at sea where these events are said to take place (Fig. 3). None of the features identified in Figure 3, such as Hrakhvammr or Blilandshofai, are actually referred to in the saga but are inferred by the editors. Having identified a literary source for this scene the editors seem to have projected the combined elements of both sources on to the real landscape of Iceland. 1>6rhallur Vilmundarson's introduction discusses textual and physical landscapes, the origins of place-names, and indicates the possible role of place-name explanations within sagas as potential sources of saga narratives:

29

Ibid., lxxviii, 126 nn . 2-3; Olafur Lirusson, op. cit. in note

I,

176.

282

LAND, SEA AND HOME

Hofundur Har8ar sogu notar ornefni sem heimildir eins og margir fornir sagnaritarar. Oft er auglj6st eoa Hklegt, ao urn se ao ra:oa alpy8legar ornefnaskyr­ ingar (folkeetymologi).30 [The author of Haroar saga, like many old saga writers, uses place-names as sources. Often it is clear, or likely, that we are dealing with popular etymologies of place-names (folk-etymologies).]

There appears to be an implicit suggestion in I>6rhallur Vilmundarson's toponymic discussion of a place-name theory of saga origins. 31 So, it would seem that even the most recent iF editors are primarily interested in the topography of the sagas as representations of the physical landscape of Iceland. As we have seen, it is not only iF editors that perceive the actual landscape of Iceland in the textual landscapes of the sagas but also historians, archaeologists and politicians. In preparation for the recent millennial celebrations the Icelandic government decided to build a visitor centre to commemorate the discovery of America by Leifr Eiriksson (the son of Eirikr the Red); this prestigious centre was to be built in Haukadalur where Leifr is believed to have been born. However, Adolf Frioriksson has argued that the identification of the site in Haukadalur had been made in the I9th century using Landnamab6k and Eiriks saga rauoa as evidence, through the use of literary analogy, and that when the site was excavated using modern archaeological techniques 'ekkert kom fram er staOfesti busetu Eiriks eoa fa:oingu sonar hans par' [nothing was found to confirm it as Eirikr's [the Red] residence nor that his son was born there].32 The visitor centre has now been built in Haukadalur. Although the islendingasogur have not been read as historically reliable documents for more than 50 years, in Iceland or elsewhere, there is still a strong tendency for commentators from a range of disciplines and nationalities to view the landscapes of the sagas as simple representations of the physical Iceland. In an earlier examination of topographic features - rivers and woods, as well as ice, fog and caves - I found that they all demonstrate clear patterns of narrative functionality.33 They are elements within the narrative grammar of the sagas, and not merely backdrops against which the actions of the sagas are set. This leads me to suggest that any saga landscape reference that may be used in an archaeological or historical context should be examined for its narrative potential before considering its possible locative function. Dr Ian Wyatt, 4 Starbold Court, Starbold Crescent, Knowle, Solihull B939LB, UK [email protected]

Ed. cit. in note 28, xxx. For furrher commenr on this possible place-name theory of saga origins see Rory McTurk, Review of iF XIII, Saga-Book, 24 (1995), 164-72. , 32 Adolf Frioriksson, 'Omenningararfur Islendinga: "Endurbygging" ba: Eiriks rauoa i Haukadal', Skirnir, 1.z2 (199 8 ),451-5, at p. 453· 3 Wyatt, op. cit. in note I. 30 31

a

/

/

JOR VIK: A VIKING-AGE CITY

By

RICHARD

A.

HALL

A wide range of data indicates that York was the main focus for Scandinavian interests in the late Anglo-Saxon Period. Documentary sources, for example, mention a sequence of political and military events, starting with the capture of York by the Viking 'great army' in 866. 1 For the next 60 years independent kings of Scandinavian ancestry were in ultimate control, until the annexation of York and Northumbria by King .tEtheistan of Wessex/England in 927. Political manoeuvrings and military confrontations in 927-54 ranged southern English kings against rivals from northern England, Scotland, Ireland and Scandinavia in the contest for control of York. King Eadred's success in 954 turned out to be definitive; there were no more independent kings of Northumbria. Yet, until the Norman Conquest and even beyond, York remained a potential bastion for rebellious northerners who drew strength, in part, from the Scandinavian element in their identity. These distinctive traits are epitomized in Earl Siward's (tI055) dedication of his church at York to St Olaf. They were acknowledged by the Wessex dynasty in their diplomatic appointments of Anglo­ Scandinavians to positions of secular and ecclesiastical authority in Northumbria, and were recognized by a series of lIth-century Danish and Norwegian kings who were encouraged to invade England via York. The written records of these events ensured that York's reputation as a centre of Scandinavian interest survived the Middle Ages, and influenced historians of the I7th century and onwards. Occasionally in the I8th century and with increasing momentum in the I9th century, antiquarians attempted to identify, to attribute to cultures and thereby to date ancient remains discovered below York. Tangible traces of a Viking presence were amongst the prizes which they pursued. The recognition of coins struck for Scandinavian kings of York initiated the study of artefacts and the discussion of street names of Scandinavian form encouraged the study of urban topography, two important and continuing strands of research. In the early 20th century, for the first time, records were made of pre-Norman timber structures uncovered during building works/ and further fragmentary building remains were excavated during the period of increased but episodic archaeological excavation in the I950S and I960s that included work by Katherine Richardson, Peter Wenham, Ian Stead and Herman Ramm. A sustained campaign of archaeological recording within and immediately around York Minster, carried out in I966-73, revealed an 1 D. W. Rollason, Sources for York History to AD noo (The Archaeology of York [a continuing series of reports published in themed volumes, edited (1972-2002) by P. V. Addyman: hereafter AYj, I, York, 1998) . G. Benson, 'Notes on excavations at 25, 26 and 27 High Ousegate, York', Yorkshire Philosophical Soc. Ann. Rep. 1902,64- 7 .

28 4

LAND, SEA AND HOME

Anglo-Scandinavian cemetery and some evidence for craft activities. 3 Studies of Anglo-Scandinavian sculpture also flourished, building upon the work of Collingwood and encouraged by the discovery of grave markers in the cemetery below York Minster. 4 Lang produced a notable corpus of the city's sculpture. s An important corpus of artefacts was brought together and published by Dudley Waterman,6 and numismatic studies flourished, with a string of papers by Christopher Blunt, Michael Dolley, Stuart Lyon, Hugh Pagan and Ian Stewart.? In 1971 Jeffrey Radley published an important paper in Medieval Archaeology, entitled 'Economic Aspects of Anglo-Danish York,.8 The following year York Archaeological Trust's excavation at 6-8 Pavement provided a more comprehensive assessment of the archaeological potential in the key area to which Radley had drawn attention. A 9-m depth of archaeological strata was identified, mostly anoxic deposits containing well-preserved organic and inorganic artefacts and palaeo-environmental data. 9 More extensive excavations of 1976-81 at 16-22 Coppergate led to a major programme of research and publication. to Other sites with Anglo-Scandinavian deposits have also been investigated by York Archaeological Trust and, since the introduction to archaeological fieldwork of competitive tendering in 1990, by some other archaeological contractors. These various archaeological activities and their popular manifestation, the JOR VIK Centre, have ensured that York remains the English city most closely identified with 'Vikings'. This paper attempts to provide an introduction to the present state of knowledge, introducing some new discoveries not yet widely known and identifying some of the major gaps in current understanding .

.1 D. Phillips and B. Heywood, Excavations at York Minster I: From Roman Fortress to Norman Cathedral (London, 1995). 4 W. G. Collingwood, 'Anglian and Anglo-Danish sculpture at York', Yorkshire Archaeol. J., 2.0 (1909),

119- 213 .

J. T. Lang, York and Eastern Yorkshire (British Academy Corpus of Anglo-Saxon Stone Sculpture, III, London, 1991). 6 D. M. Waterman, 'Late Saxon, Viking and Early Medieval finds from York', Archaeologia, 97 (1959), 59-106.

J. E. Pirie, Post-Roman Coins from York Excavations r97r-8r (AY 18/1, 1986). Radley, 'Economic aspects of Anglo-Danish York', Medieval Archaeol., 15 (1971),37-57. 9 P. V. Addyman, 'L1oyds Bank, Pavement', in P. V. Addyman and R. A. Hall, Urban Structures and Defences (AY 8/3, 1991). 10 Pirie, op. cit. in note 7; T. P. O'Connor, Bones from Anglo-Scandinavian Levels at r6-22 Coppergate (AY 15/3, 1989); P. Walton, Textile, Cordage and Raw Fibre from r6-22 Coppergate (AY 17/5, 1989); A. ]. Mainman, The Anglo-Scandinavian Pottery from r6-22 Coppergate (AY 16/5, 1990); P. J. Ottaway, Anglo­ Scandinavian Ironwork from r6-22 Coppergate (AY 17/6, 1992); J. Bayley, Non-ferrous Metalworking from r6 - 22 Coppergate (AY 17/7, 1992); H. K. Kenward and A. R. Hall, Biological Evidence from Anglo­ Scandinavian Deposits at r6-22 Coppergate (AY I-V7, 1995); P. Walton Rogers, Textile Production at r6-22 Coppergate (AY 17irl, 1997); A. MacGregor, A. ]. Mainman and N. S. H. Rogers, Bone, Antler, Ivory and Horn from Anglo-Scandinavian and Medieval York (AY 17/12, 1999); C. A. Morris, Wood and Woodworking in Anglo-Scandinavian and Medieval York (AY 17/13,2000); A. ]. Mainman and N. S. H. Rogers, Finds from Anglo-Scandinavian York (AY 17ir4, 2000); Q. Mould, E. Cameron and J. Carlisle, Leather and Leatherworking in Anglo-Scandinavian and Medieval York (AY 17/16,2003); R. A. Hall et aI., Aspects of Anglo-Scandinavian York (AY 8/4, 2004); R. A. Hall, Anglo-Scandinavian Urbanism in Coppergate: Defining a Townscape (AY 8/5, in prep.). 7

E.

S ].

,ORVIK: A VIKING-AGE CITY

28 5

From Eoforwic to Jorvlk: the 9th century in York (Fig. r) It has been suggested that the extent of pre-Viking occupation in York increased markedly in the 8th century or 9th century.ll Most emphasis has been placed on changes within and around the former Roman fortress, although everywhere within the area of the later medieval walled city has come under some scrutiny. This suggestion derives from inferences and speculations based upon several classes of evidence. Norton's starting point is his interpretation of how and when the cathedral precinct developed; both he and Tweddle discuss street names of Old English as opposed to Old Norse form, hypotheses about when some of the parish churches were founded, and the widespread distribution of artefacts whose manufacture is dated to this period on art-historical or typological criteria. Yet the reliability of much of this evidence is very often dubious; Tweddle notes its 'very flimsy and incomplete nature,.12 For example, many of the dots on the distribution maps of artefacts represent a single coin, one or two sherds of pottery or a solitary piece of decorated metalwork, devoid of any contemporary stratigraphic context. As the objects could have reached their find spots through a variety of medieval and modern mechanisms, their evidential value for the distribution of Anglian activity is diminished. Some of these items are no longer available for study, and we have to rely upon antiquarian identifications which are sometimes rather vague and, therefore, questionable. Furthermore, as Tweddle comments, some of these stylistically 'Anglian' objects have been found in contexts of certainly later date, and 'may testify to a conservative streak in the craftsmen of the Viking Age city'Y Equally, they may represent a deliberately 'English' political statement by some Viking-age craftsmen. 14 In either case, they would have no relevance to Anglian York. Similarly, in discussing the evidence for Anglian structures, Tweddle notes correctly that much of that data is 'modest, often insubstantial and chronologically ambiguous'.15 Equally, while it may be a 'reasonable perception that surviving Old English street names may represent an early survival', 16 it is worth noting how uncommon they are and that at least one of them, Aldwark (eald geweorc 'the old work') refers to the Roman fortress wall, and might have been retained into the Anglo-Scandinavian period as the same sort of linguistic legacy which elsewhere ensured the retention of names such as Pendle Hil!.l? The data for post-Roman but pre-Viking settlement at York which survive a rigorous pruning of the dubious or ill-authenticated are therefore relatively few. They suggest that the Roman intra-mural street pattern had largely disappeared except for

11 C. Norron, 'The Anglo-Saxon cathedral at York and the topography of the Anglian city', ]. British Archaeol. Assoc., 151 (1998), 1-42; D. Tweddle et aI., Anglian York: A Survey of the Evidence (AY 712,1999) . 12 Op. cit. in note II, 151. 13 Ibid., 259. 14 R. A. Hall, 'Anglo-Scandinavian attitudes: archaeological ambiguities in late ninth- to mid-eleventh-century York', 311 - 24 in D. M. Hadley and ]. D. Richards (eds.), Cultures in Contact: Scandinavian Settlement in England in the Ninth and Tenth Centuries (Turnhout, 2000). 15 Tweddle, op. cit. in note II, 190f. 16 Ibid., 160. 17 E. Ekwall, The Concise Oxford Dictionary of English Place-Names (4th ed., Oxford, 1960), 361.

286

LAND, SEA AND HOME

Outline of Roman fortress

Roman roads

Medieval walls

SI Denys' parish boundary

FIG URE I.

York. The street-plan of the centre of the modern city, showing the lines of the Roman roads and fortress and the medieval town-walls.

the main routes to the fortress gateways. The northern quadrant of the former fortress was the location of the cathedral church and other episcopal buildings; the uses to which the remainder of this enclosed area was put remain unknown, although Norton suggests that part of the western quadrant, later occupied by the medieval St

,OR ViK:

A VIKING-AGE CITY

28 7

Leonard's hospital, was already in church hands. 18 Across the River Ouse, parts at least of the waterfront within the walled area were utilized for commerce. 19 On the plateau above the river there may have been an important ecclesiastical focus in the vicinity of Holy Trinity Priory and the St Mary Bishophill churches; otherwise, this area too remains mysterious. Downstream of the Roman defended areas, close to the confluence of the River Ouse with its tributary the River Foss, excavation at 46-54 Fishergate revealed most of one relatively large and well-defined parcel of land, over r,200 m square, containing at least one timber building, and dated c. 7°°-85°.20 Associated artefacts include a range of manufacturing debris from metalworking, glassworking, etc., and goods imported from north-west Europe. 21 Yet this vicinity does not share the dense contemporary occupation seen at superficially comparable WIC sites such as London or Southampton, and it may represent a different aspect of socio-economic development. The overall interpretation proposed here is that pre-Viking York was a dispersed network of occupied areas - along the river fronts, around the cathedral, perhaps above the river valley in the Bishophill area. The Roman walls remained substantial barriers, but within them there was much relatively open ground. Some of this may have been used for growing food or raising animals, places where rubbish could be dumped - this would account for the widespread scatter of Anglian artefacts. Until traces of buildings or domestic infrastructure (wells, cess-pits, rubbish pits) are discovered, it may be doubted if occupation was any more dense than that seen at 46-54 Fishergate. A relatively small permanent population of one to two thousand people may be envisaged. The latest ceramic types from the Anglian occupation sequence at Fishergate mirror the earliest types at r6-22 Coppergate. 22 The pre-Viking coin sequence from 46-54 Fishergate ends in the 860s,23 whereas at r6-22 Coppergate activity resumed in the mid-9th century, apparently for the first time since the Roman Period. Does this solitary but reasonably closely dated example mean that a change in the settlement pattern at York can be hypothesized? Was this a change which concentrated occupation nearer to the river crossing and the defences. At a time approximately contemporary with the historically documented take over of the settlement by the Viking 'great army' in 866?

York in the Loth and IIth centuries It was in the early decades of the roth century that long, narrow property plots were established along the south-east side of Coppergate. In the 960s or 970S the number of buildings per plot and the usable space within buildings both increased, further raising the density of occupation. The chronology of urban development is Norton, op. cit. in note 11,25. A. J. Mainman, 'Imported wares in York c. AD 700-1000',189-93 in D. Piton (ed.), La Ceramique du au x·m e Sii!Cie dans L'Europe du Nord-Ouest (Nord-ouest Archaeologie Hors Serie, 1993). 20 R. L. Kemp, Anglian Settlement at 46- 5'4 Fishergate (AY 71I, 1996). 21 N . S. H. Rogers, Anglian and Other Finds from Fishergate (AY 17/9, 1993)' 22 A. J. Mainman, Pottery from 46-.'14 Fishergate (AY 16/6, 1993) ' 23 Kemp, op. cit. in note 20, 83 . 18

19

v

,

eme

288

LAND, SEA AND HOME

not so precise elsewhere in York; for example, between Coppergate and Ousegate a series of buildings similar to the later 10th-century examples from I6-22 Coppergate has been recognized,24 but they can be dated only by analogy with their near neighbours. A fragmentary but similar structure is also known from King's Square, where it was cut into the rampart of the Roman fortress near the south-eastern gateway, but its chronology is also imprecise. 25 Data from several small-scale investigations in the Coppergate/High Ousegate area, including 2 Clifford Street, 4-7 Parliament Street and 28-29 High Ousegate appear consistent with urban development taking place in the Anglo-Scandinavian rather than the pre-Viking period. 26 At York Castle a cemetery sealed beneath the Norman castle's earthworks presumably relates to a hitherto unknown pre-Norman parish church devastata in castellis by William the Conqueror. 27 At All Saints church, at the junction of Coppergate and High Ousegate, excavations in the western part of the south aisle revealed evidence for a small two-cell church built on gritstone foundations. 28 The chancel was just under 3 m long internally, with a width of more than 2.6 m (its northern wall was not uncovered); the easternmost 2.8 m of the nave was also excavated. There was no precise dating evidence for its erection, but the early 10th century would equate with the laying out of the street Coppergate and with the date of a decorated stone grave cover recovered there in I963. There have been suggestions that a pre-Viking minster church stood here but no traces of such a building were identified. Remains attributed to a succeeding, cruciform church may represent a mid-nth-century replacement of this two cell structure; perhaps this is the church with which a group of unusual, high-status polychrome relief tiles attributed to the mid-nth century was originally associated.29 Further to the north, other recent excavations are providing details of activity or occupation in the area between the Roman fortress wall and the River Foss which has not hitherto been much investigated. At 9 St Saviourgate, near the edge of the plateau above the low-lying floodplain of the River Foss, excavation uncovered structural remains of post and wattle. 30 A Viking-age pit contained a large hemispherical iron cauldron, within which was a lead vat or bucket. In close association was a group of 3I iron objects - tools, including awls, draw knives, an auger and a knife, several

24 Benson, op. cit. in note z; R. A. Hall, 'Structures at 5-7 Coppergate, with a reassessment of Benson's observation of 190Z', in Addyman and Hall, op. cit. in note 9. 25 L. P. Wenham, 'Discoveries in King's Square, York, 1963 ' , Yorkshire Archaeol.)., 4z (1968), 165-8. 26 M. Johnson, 'z Clifford Street, York: Report on an Archaeological Watching Brief and Excavation' (York Archaeo\. Trust Field Rep. 1999124, unpubl.); idem, '4- 7 Parliament Street, York: Report on an Archaeological Watching Brief' (York Archaeo\. Trust Field Rep. 1999/71, unpub\.); S. Diamond and J. McComish, 'z8-z9 High Ousegate, York: Draft Assessment Report on an Archaeological Excavation' (York Archaeo\. Trust, unpub\., Z003). 27 D. T. Evans, 'Former Female Prison, Castle Yard, York: Report on an Archaeological Evaluation' (York Archaeo\. Trust Field Rep. 1998126, unpub\.). 28 P. Ware, 'Excavations at All Saints Church, Pavement, York 1991' (MAP Archaeological Consultancy Ltd, unpub\., ZOOI); John Oxley, pers. comm. 29 L. Keen, 'Pre-conquest glazed relief tiles from All Saints Church, Pavement, York', ] . British Archaeol. Assoc., 146 (1993),67-86. 30 Anon., '9 St Saviourgate, York', Forum 1998: Annual Newsletter of Council for British Archaeology Yorkshire (1998),15 -17; John Oxley, pers. comm.

,OR VIK:

A VIKING-AGE CITY

28 9

with their wooden handles surviving, and a miscellany of iron fittings. This group may represent a smith's collection due for recycling. In the floodplain below St Saviourgate at two sites in the street Hungate where Richardson's pioneering excavation revealed what was initially interpreted as an 'Anglo-Danish embankment', excavation has, for the first time in this vicinity, located Anglo-Scandinavian deposits that look like floor levels within buildings. 3! Addition­ ally, a series of other evaluatory investigations hereabouts has indicated, contrary to some previous belief, that the Viking-age course of the River Foss ran on approxi­ mately the same line as its modern counterpart. 32 Just north of Hungate at Layerthorpe, where the medieval Layerthorpe Bridge spanned the River Foss, excavation has uncovered linear wattlework features which, it is suggested, relate to land reclamation and flood prevention in the roth and 11th centuries. Assessment of associated plant and insect remains suggests that the river's edge was used at this time for polluting activities such as flax-retting and tanning. 33 Within the Roman fortress there is hardly any evidence for pre-Norman urbanization. Few traces of Anglo-Scandinavian buildings have yet been located, and even indications of occupation or activity such as rubbish pits are scarce, although pottery distributions suggest that the southern part of the fortress was in use from the late 9th or early roth century.34 Recent dendrochronological study of wooden coffins excavated in that area, in Swinegate, near the site of St Benet's church, has dated the timbers to the Anglo-Scandinavian period, with one of them more precisely dated A.D. 975-1001.35 This suggests that St Benet's was a pre-Norman church, serving a community hereabouts. It seems likely, however, that other segments of the Roman fortress were retained for particular purposes by elite groups, and not parcelled out into urban tenement plots. Excavation at the Minster Library Extension site, some roo m north of York Minster, approximately half-way between the Minster and the northern corner of the Roman fortress, revealed virtually no evidence for Anglo­ Scandinavian activity.36 It may well be that this area was part of the pre-Norman cathedral precinct or in the vicinity of the palace of the Anglo-Saxon archbishops, as Norton has suggested. 3? Indeed, apart from evidence for burial, the only Anglo­ Scandinavian activity attested in the York Minster excavation of 1966-73 was an ironworking forge and other related evidence within the ruined shell of the former

K. M. Richardson, 'Excavations in Hungate, York', Archaeological ] ., II 6 (1959), 51-114; N. Macnab and I' (York Archaeol. Trust, unpubl., 2000). 32 M. Griffiths, pers. comm. 33 Anon., 'Layerthorpe Bridge, York. Volume I: Assessment Report' (MAP Archaeological Consultancy Ltd., unpubl., 1997); J. Carron et aI., 'Archaeological Excavations at Layerthorpe Bridge and in Peasholme Green, York: Assessment of Interpretive Potential of Biological Remains' (unpubl. rep., Environmental Archaeolob'Y Unit, University of York, 1998). 34 A. J. Mainman and N. S. H. Rogers, 'Craft and economy in Anglo-Scandinavian York', in R.A. Hall et al. (2004), op. cit. in note 10. 35 T. Bagwell and I. Tyers, ' Dendrochronological Analysis of a Coffin Assemblage from Swinegate, York, North Yorkshire' (unpubl. rep., ARCUS Dendrochronology, 2001). 36 J. Garner-La hire et aI., 'York Minster Library: Field Report' (unpubl. rep., Field Archaeology Specialists, 31

J. McComish, 'Hungate Development, York: Report on an Archaeological Evaluation Vol.

2000). 37

Norton, op. cit. in note

11.

290

LAND, SEA AND HOME

Roman Barrack 2. 38 This forge is as likely to have fulfilled the requirements of the ecclesiastical enclave here as to have produced goods for general sale in the city's markets; and the evidence from hereabouts for bone and antler working and jet/shale working at the commercial level in this period is exiguous and even less convincing. On the south-west bank of the River Ouse, within the area occupied by the former Roman civilian town Colonia Eboracensis, Anglo-Scandinavian occupation has been identified at several sites close to the river. The only, albeit indirect, evidence for the establishment of a new crossing over the Ouse, to replace the Roman bridge 160 m upstream, has come from excavations within the former Roman colonia at 1-9 Micklegate, although the site records await detailed analysis. A substantial Roman building had been levelled off at the contemporary ground surface, and a series of quite regularly spaced post-holes was gouged into at least one of its walls. The structure of which they formed a part may be associated with a small group of burials which were accompanied by knives, and may be of 7th- or 8th-century date; the structure was probably earlier than the adjacent street Micklegate, which ran at an oblique angle to the building. In contrast, at least one timber building, constructed in an Anglo-Scandinavian cellared or basement style and therefore interpreted as being of Anglo-Scandinavian date, cut across the Roman alignment. It was, however, perpendicular to Micklegate, the 'great street', that leads to Ouse Bridge. There is documentary evidence which points to a bridge at this site from at least the 12th century;39 if the provisional interpretation and dating of the excavated sequence at 1-9 Micklegate is correct, it supports the suggestion that the change to this bridging point and the creation of Micklegate occurred during the Anglo-Scandinavian period. Downstream of Ouse Bridge, excavations at 58-9 Skeldergate revealed parts of perhaps four structures. Except where one structure replaced another at the frontage, their chronological inter-relationships were unclear; and they can be dated only broadly to the 9th-lIth centuries. Their construction was notable for the variety of foundation techniques that their builders utilized, but it was not possible to link particular building methods with specific functions for the different buildings; no characteristic accumulations of craft or industrial waste were noted. Notably, they occupied not only the street frontage but also extended some 28 m back from it. Their positions indicate that the alignment of the medieval and modern plots hereabouts, and that of the nearby Carrs Lane which runs down to the River Ouse, originate in the Anglo-Scandinavian period; and they suggest that Anglo-Scandinavian occupation hereabouts was quite dense. 40 Upstream of the crossing, excavation at 16-22 Wellington Row, a site adjacent to the Roman bridge across the River Ouse, revealed some evidence for activity or occupation nearby in the Anglo-Scandinavian period. Roman walls still visible above the contemporary ground surface were dismantled to their foundations, and pits were 38 M. O. H. Carver, 'Roman to Norman at York Minster', 177-221 in D. Phillips and B. Heywood, Excavations at York Minster 1: From Roman Fortress to Norman Cathedral ((London, 1995). 39 B. M. Wilson and F. P. Mee, 'The Fairest Arch in England'. Old Ouse Bridge, York, and its Buildings: The Pictorial Evidence (AY Supp. Ser., 2, 2002). 40 S. Donaghey and R. A. Hall, 'Anglo-Scandinavian structures and features at Bishophill', 37-52 In ]. Moulden and D. Tweddle, Anglo-Scandinavian Settlement South- West of the Ouse (AY 811, 1986).

,oRviK: A VIKING-AGE CITY

29 1

dug, but no clear traces of contemporary structures or property divisions were recognized during the excavation; a more detailed understanding will come only when post-excavation analysis can be undertaken. Limited excavations nearby at 22-30 Tanner Row demonstrated that the roth century was a time of change in this area of York, with the clearance of land and the reclamation of re-usable building stone. The remains of Roman stone buildings were finally removed, and there was some evidence for contemporary activity or occupation in the vicinity. Here too, however, both site-specific and broader interpretations remain obscure. The only detailed evidence for how the waterfront itself was managed in the Anglo-Scandinavian period comes from excavations in 1993 between North Street and the Ouse, a site close to the line of the medieval lane Divelinstaynes ('Dublin Stones,).41 This name may indicate where ships from Dublin were moored; although first recorded in 1233 x 1239, it has been suggested that it was coined in the Viking Age, when links between Dublin and York were particularly strong. Excavation of a circular shaft, 6 m in diameter, uncovered evidence which suggests that a late 2nd­ century river wall had been dismantled by the end of the Roman Period. Thereafter, from perhaps the 8th century onwards, there was a series of attempts both to define the waterside and to facilitate use of the river. Stake and wattle alignments were used to revet the riverbank; wattlework hurdles were sometimes laid horizontally on the revetted surface above the river and also on the foreshore itself. In the most elaborate arrangement, of 9th- and roth-century date, a deposit of stones and pebbles was laid on the sloping bank. Five approximately parallel rows of stakes were inserted into this surface following the contours of the river bank and serving (presumably) to stabilize it. At their riverward side, horizontal wattle hurdles were held in place by larger horizontal timbers placed at right angles to the river. On the revetted platform above the river, a similar arrangement of hurdles and horizontal timbers provided a relatively firm and consolidated surface. A later phase, also dated to the loth century, was represented by a substantial wattlework hurdle laid on the foreshore; and the management process continued, in a variety of forms, until a riverside wall was built further out into the river in the 13th century. Further away from the River Ouse, higher up the pronounced slope of the Ouse valley's south-western side, archaeological deposits become thinner and soil condi­ tions less conducive to the preservation of organic materials. No structural traces of Anglo-Scandinavian secular activity have yet been found in this vicinity, although the churches of St Mary Bishophill Junior, St Mary Bishophill Senior and Holy Trinity Micklegate all testify, in one way or another, to contemporary activity. St Mary Bishophill Senior had an unusually large complement of Viking-age grave markers, a phenomenon found elsewhere in Lincolnshire and Yorkshire which Stocker equates with the existence of trading centres. On this basis, he has elaborated the hypothesis

41

R. Finlayson, 'North Street, York 1993 .1. Archive Report' (York Archaeol. Trust, unpubl., 1993)'

29 2

LAND, SEA AND HOME

that this was a deliberately planned zone in the Anglo-Scandinavian town with the suggestion that this area was an enclave for Scandinavian traders in the 10th century.42 The later medieval walled city also encompassed an area north of the River Ouse and east of the River Foss, with the street Walmgate as its spine. This area was not enclosed by walls until the 12th century, although there is a suspicion that the line of St Denys's parish boundary may represent an earlier fortification defending a bridgehead, perhaps of Anglo-Scandinavian date, to the east of the River Foss. This hypothesis awaits testing. There is, however, an accumulation of evidence that the land bordering Walmgate was sub-divided into plots of some description, and buildings were erected on at least some of them, during the Anglo-Scandinavian era. In particular, at 41-9 Walmgate, excavation in 2000-1 revealed indications that fence or wall lines perpendicular to the street had certainly been laid out by the mid- to late 10th century; the possibility that they were initiated even earlier could not be verified, as the strata containing evidence for their origin remain uninvestigated .43 There is also a possibility that the riverward edge of the Walmgate area, along the River Ouse, was being developed before the Norman Conquest; excavations at 41 Piccadilly revealed a pebble-surfaced 'hollow way' dated to the IIth century.44 This route became fossilized as what is now Dixon Lane, which runs on a line just a few metres north of this I rth-century course. From this topographical and structural evidence it is as yet impossible to comprehend how Anglo-Scandinavian York developed to become, by 1066, the second largest city in England. It could be suggested that much of the Roman fortress area was retained in ecclesiastical or royal control, and not parcelled out into urban tenement plots until relatively late. The movement of the bridging point over the Ouse from its Roman to its medieval site might be construed as not (exclusively) resulting from a necessary abandonment of the earlier bridge, but as a deliberate political and economic act to channel traffic away from the former fortress and towards the 'ness', the neck of land south-east of the Roman fortress between the rivers Ouse and Foss. It testifies to the imposition of sweeping powers to re-route what is now Micklegate across private property to the new crossing point. Whether such a move would have been to the advantage or the detriment of landowners in the former Roman fortress, such as the archbishop, or whoever controlled the upstream side of the former Roman civilian town, is debatable. However that may be, the area of Ousegate, Coppergate, Nessgate/Castlegate and Pavement, which Radley defined as the economic focus of Jorvik, contains the earliest topographical evidence for the post-Roman development of urban plots. 45 It may be that both banks of the Ouse were foci for waterborne trade and that settlement extended along them, but the extent of such settlement upslope from the 42 D. Stocker, 'Monuments and merchants: irregularities in the distribution of stone sculpture in Lincolnshire and Yorkshire in the tenth century', 179-212 in Hadley and Richards (eds.), op. cit. in note 14; D. M . Palliser, 'York's West Bank: medieval suburb or urban nucleus?', 101-8 in P. V. Addyman and V. E. Black (eds.), Archaeological Papers from York presented to M. W. Barley (York, 1984). 43 www.yorkarchaeology.co.uk/wgate 44 Anon., '41 Piccadilly, York', Forum 1998: Annual Newsletter of Council for British Archaeology Yorkshire (1.?9 8),12- I 3. 4 Radley, op. cit. in note 8.

,oRviK: A VIKING-AGE CITY

293

river's margins on its south-west side is unclear. The existence of comparatively few routes down to the river's south-west edge throughout the later medieval period, compared with the several 'water lanes' on the north-east bank, may be some indication of the relative commercial significance and population density of the two sides of the river before the Norman Conquest. ARTEFACTS, CRAFT, PRODUCTION AND EXCHANGE

Another method of charting York's development, complementary to the topographical and structural evidence addressed above, is to consider the distribution of the ceramic types which were in use during this era. Their chronology allows three main ceramic phases to be distinguished; c. 860-950, c. 950-I050, and post-I050. Preliminary and provisional results seem to indicate changes in the distribution of pottery which may represent an expansion of activity. The completed study will be published by Mainman and Rogers. 46 Further insights into York's role at this time come from a synthesis of the evidence for production. Specialist studies of the range of artefacts and manufacturing debris from 16-22 Coppergate and elsewhere have demonstrated that work was sometimes carried out on a commercial rather than a domestic scale. There are uncertainties in extrapolating from excavated debris which represents an unknown fraction of the amount that was originally generated; it may have been diminished in quantity by removal from site, by complete destruction (e.g. the burning of woodworkers' waste), or by deterioration on site. Furthermore the duration, continuity and intensity of individual 'productive episodes' cannot be measured to a fine degree of accuracy in the archaeological record. So it is debatable whether the 294 wooden cores and other lathe-turning waste from Coppergate represent full- or part-time work; whether this work was seasonal or year-round; whether it was intermittent or daily; and how many artisans were involved. These arguments could be extended to the metalworkers, leather-workers, textile workers and other craftspeople who evidently worked on the four tenement plots excavated at Coppergate, and applies equally to the potters who mass-produced standardized wares, and to other suppliers based elsewhere in York or beyond. Nevertheless, the mid-9th to mid- l Ith centuries can coarsely be categorized as a period within which large-scale production by specialist craftsmen for a mass market became the norm for many products. There was not yet the ultra-specialization of the later medieval craft guilds - it seems, for example, that 'high-temperature specialists' worked with gold, silver, lead alloys, copper alloys and iron as well as with glass. These specialist mass-producers may also have been willing to work for occasional brief periods in very different materials - this is one interpretation that would account for small quantities of some raw materials, such as amber. 47 Perhaps, alternatively, itinerant specialists made occasional forays to York. Whatever the explanation, the links between the suppliers of raw materials in the region and the craft workers in the city are in need of further probing. 46 47

Mainman and Rogers, op. cit. in note 34.

Ibid.

294

LAND, SEA AND HOME

Studies of the international contacts employed by York's Viking-age merchants and their customers have also been refined over the last decade or so. Elements in the 7th- to 9th-century pre-Viking-age trading network which were identified, for example, in the settlement excavated at 46-54 Fishergate, continued into the Viking Age. Links to the north-west European littoral and inland along the River Rhine were maintained, and the importation of Rhenish goods such as quernstones continued. But whereas in pre-Viking-age York there had been little contact with Scandinavia or the Baltic, this changed in the Anglo-Scandinavian period. The amount of amber found in York increases dramatically - although not to the point of sustaining full­ time craftsmen. An even more common Scandinavian import took the form of honestones in Norwegian schist or phyllite; and there is evidence that c1ubmoss (Diphasium complanatum) was also brought from Norway for use as a dyestuff. More exotic objects such as Byzantine silks, a cowrie shell (cypraea pantherina solander) from the Red Sea and a contemporary Arab counterfeit of a dirham from Samarkand indicate York's more distant links, but the majority of these very distant contacts were sporadic, and were centred chronologically in the first few generations of Scandinavian settlement. A HEADLINE AGENDA FOR FUTURE RESEARCH

There are key gaps in every facet of our knowledge of Viking-age York. Topographically, a fuller comprehension of how the defensive sequence evolved would assist in unravelling the pattern of urban growth. The additional ramparts which link the fortress to the Ouse and Foss are two obvious targets, as are the ramparts around the colonia area and the hypothetical line of the Walmgate bridgehead. The waterfronts - their structural form, and the information about contracts and commerce which they will contain - also deserve attention. Opportun­ ities to establish whether or not other parts of the city conform to the Coppergate model of early roth-century plot creation, and share the forms of buildings recorded there, would be welcome, particularly in the southern and eastern quadrants of the fortress and in the upper valley zones of the colonia. The extent of J6rvik's 'industrialization' and the degree of social stratification are two themes which could also be investigated in such excavations. The diagnostic material through which social stratification could be pursued - e.g. structures, artefacts, provisioning - may also assist an understanding of how ethnic and political identities were expressed in the various strata of the city's population. Important arenas for such expression include high-status secular sites, such as the royal, aristocratic and archiepiscopal residences, and also religious ones - parish churches for the well-to-do, the cathedral and its adjuncts at the highest level. The cemeteries around their sites, if relatively undisturbed by later burials, also have much to contribute on these themes, as well as allowing demographical and other study of the urban population. Comparison of that population with its rural counterparts, not just in terms of demography, palaeopathology and place of origin (using oxygen isotope analysis) but also in terms of diet and possessions, would further an understanding of the crucial relationships and interactions between Jorvik and its hinterland.

,oRviK: A VIKING-AGE CITY

295

ACKNOWLED G EMENTS

I am grateful to John Oxley, City of York Council's Archaeologist, for facilitating access to unpublished reports, and to all those who have assisted in the preparation of this essay.

Dr Richard A. Hall, York Archaeological Trust, Cromwell House, 13 Ogle{orth, York 7FG, UK [email protected]

Y01

VIKING-AGE SETTLEMENT IN THE

NORTH-WESTERN COUNTRYSIDE:

LIFTING THE VEIL?

By

NICHOLAS HIGHAM

The notion of a mass migration of Vikings into the north-west of England became something of an historical truism during the course of the 20th century. This essay sets out to explore some of the parameters of this vision of the period and suggests both that the scale of this influx may have been exaggerated and that the processes of population movement were arguably a great deal more complex than has generally been imagined. The staging posts in the development of the general notion are comparatively clear-cut. In looking back across a century, we should perhaps begin with volume two of the Victoria County History for Lancashire, published in I908, in which James Tait briefly surveyed the place-name and literary evidence and concluded that a sea­ borne settlement of 'western Scandinavians' had occurred early in the roth century.l There followed Eilert Ekwall's seminal works on the place-names of the region, published in I9I8 and I922/ which set out the case, via a detailed consideration of primarily the major (parish and township) place-names, for a large-scale Scandinavian influx resulting in a largely Norse-speaking population. This interpretation was then developed by Frederick Threlfall Wainwright, a pupil of Sir Frank Stenton, who began working in this his home region in the I930S and was still engaged in research here when he died at the tragically early age of 43 in I961. A few months after the ending of the Second World War, Wainwright published a detailed study of field-names in the Lancashire Hundred of Amounderness. In rehearsing the historical context of his work, he wrote: 'The known facts, few and easily summarized, are that hordes of Scandinavians settled in what is now Lancashire during the early years of the tenth century, that they came from Ireland not direct from Scandinavia, that they were Norsemen or Irish-Norsemen not Danes, and that they lacked the military organiza­ tion which characterized the Danish settlements in eastern England.'3 His work identified numerous field-names of Scandinavian origin, which led him to conclude 1 J. Tait, 'Political history to the end of the reign of Henry VIII', in W. Page (ed.), A History of the County of Lancashire, Vol. 2 (Victoria County History of England, London, 1908), 177-8. 2 E. Ekwall, Scandinavians and Celts in the North-West of England (Lund, 1918); idem, The Place-Names of Lancashire (Manchester, 1922). 3 F. T. Wainwright, 'Field-names of Amounderness Hundred', Trans. Hist. Soc. Lancashire Cheshire, 97 (1945-6), 71- II6; repro 229-79 in H. P. R. Finberg (ed.), Scandinavian England (Chichester, 1975), at p. 256.

29 8

LAND, SEA AND HOME

that Scandinavian had been the lingua franca of the region from the roth to the 12th centuries, so reinforcing his assumption of a large-scale influx. MAPPING THE EVIDENCE

This vision of mass immigration passed rapidly thence into the mainstream of historical writing, with Stenton remarking on 'Norse occupation' of the north-west,4 and both Peter Hunter Blair and Henry Loyn mapping 'areas of strong Norwegian settlement' across the region. s The impression of large-scale and widespread immigration is then implicit in James Graham-Campbell's mapping of Viking settlement in 1980 and Simon Keynes's map of 1998, of 'The Kingdom of the Anglo­ Saxons', both of which effectively block out the north-west as a Scandinavian space. 6 That said, these representations are not entirely consistent as regards the area of settlement, for some authors exclude southern parts of the region, which others include. These differences are not marginal: the area of Lancashire which is in question is around 84 km from north to south, and Keynes additionally included the whole of northern Cheshire plus the coastal plain of North Wales within his 'area of Norse settlement and influence'. To take the most southerly area first, both Grimston-hybrids and Norse settlement names are scarce in Cheshire, with just one small concentration on the tip of the Wirral, with an outlier nearby at Helsby on the Mersey. Norse elements in minor names and field-names are similarly clustered, with relatively few scattered elsewhere. 7 Both of these small foci may well derive from Viking settlements near Chester, associated with Ingimund's activities just after 900,8 but it is worth noting that both occur at locations that are particularly well connected with the much denser pattern of Scandinavian place-names in south-west Lancashire: the Birkenhead and Runcorn ferries were significant crossings of the Mersey in the later Middle Ages and Frodsham (very near Helsby) was additionally an out-port of Chester that was negotiable for sea-going vessels. Both these small clusters may, therefore, have derived in part at least from contact with, and small-scale population movement from, the northern banks of the Mersey. The bulk of northern Cheshire should arguably, therefore, be omitted from maps of Scandinavian place-naming, and there is little further reason to consider this an area of particular Scandinavian influence. Nor has the notion of Viking settlement in the south-east of Lancashire ever been widely accepted among regional specialists. Tait, for example, noted the presence of place-name evidence indicative of Scandinavian settlement from the Dee estuary to the Solway, but he stressed its near total absence from Lancashire south of the Ribble Sir F. M. Stenton, Anglo-Saxon England (Oxford, 1943), 33I. P. Hunter Blair, Anglo-Saxon England (Cambridge, 1956), 86; H. R. Loyn, Anglo-Saxon England and the Norman Conquest (London, 1962),56. 6 J. Graham-Campbell, The Viking World (London, 1980),66-7; S. Keynes, Anglo-Saxon History: A Select Bibliography (Kalamazoo, 1998), 136. 7 M. Gelling, 'Scandinavian settlement in Cheshire: the evidence o f place-names', 187-94 in B. E. Crawford (ed.), Scandinavian Settlement in Northern Britain (London, 1995). • Wainwright, Scandinavian England, 81-7, 131-61; J. N. Radner, Fragmentary Annals ofIreland (London, 4

5

1978).

LIFTING THE VEIL?

299

and east of Widnes. 9 Ekwall amended this picture by drawing attention to a handful of East Norse place-names around Manchester,lO but these are neither numerous nor particularly indicative of significant Scandinavian immigration. The presence of a few individuals with Viking personal names involved in the construction of several Grimston-hybrids (as Flixton, Urmston) may well derive from the widespread popularity of Norse personal names particularly in northern England in the lIth century. Rather, by far the majority of major place-names in five of the six Domesday hundreds of southern Lancashire are pre-Scandinavian, leaving only (West) Derby, stretching along the coast, with a predominance of Viking names (Figs. 1-2).11 It does appear, therefore, that the mapping process has on occasion been both overly enthusiastic, and inclined to read very particular meanings into a range of evidence which is, in practice, quite diverse, both qualitatively and quantitatively. INTERPRETING PLACE-NAMES

There is also the issue of what Scandinavian place-names actually mean in historical and archaeological terms. Tait stressed the infrequency with which, outside specific localities, later township and parish names were Scandinavian in origin, and suggested that the incomers had generally taken over previously little-used land. 12 Ekwall made precisely the same points. 13 If township names alone are taken into account, then the pattern does indeed look rather thin, with -by, -kirk and Grimston­ hybrids forming the majority. Admittedly, these are common in parts of Cumbria at the level of township and parish, but they are scarce further south, occurring predominantly in a few coastal areas, including the tip of the Wirral peninsula and the vicinity of Ormskirk and Formby.14 Even in southern Cumbria, it is noticeable just how many pre-Viking place-names survived (as Heversham, Aldingham, Pennington and Whicham) as estate centres occupying good agricultural land in 1086. Nor has the general assumption of a western Scandinavian influx survived unscathed. Gillian Fellows-Jensen has argued that the Danish and Anglo-Danish elements among regional place-names are more significant than had previously been recognized, suggesting that the area had been in part settled from the eastern Danelaw, perhaps even via the Isle of Man.15 In this interpretation, the north-west was not just an area of Norse immigration but acted as a meeting place for western and eastern Scandinavians, with immigration from the Irish Sea balanced by incomers Tait, op. cit. in note I, 177. Ekwall (1922), op. cit. in note 2, 245, followed by Wainwright, Scandinavian England, 216-17. 11 Ekwall (1922), op. cit. in note 2, 93- 125 and 250; see D. Kenyon, The Origins of Lancashire (Manchester, lZ91),. 125-35.' for the wider context. Talt, op. CIt. In note I, 178. B Ekwall (1922), op. cit. in note 2, 256. 14 The 'minimalist' picture is offered graphically by D. H. Hill, Atlas of Anglo-Saxon England (Oxford, 1981) map 68, 45. 15 G. Fellows-Jensen, 'Scandinavian settlement in the Isle of Man and Northern England : the place-name evidence', 37- 52 in C. E. Fell, P. G. Foote, J. Graham-Campbell and R. L. Thompson (eds.), The Viking Age in the Isle of Man: Select Papers from the Ninth Viking Congress (1983); eadem, Scandinavian Settlement Names in the North- West (Copenhagen, 1985); eadem, ' Scandinavian place-names of the Irish Sea province', 31-42 in J. Graham-Campbell (ed.), Viking Treasure from the North West (Liverpool, 1992). For a sympathetic critique of this work, see V. E. Watts, 'North-western place-names', Northern Hist., 23 (1987),229-30. 9

10

300

FIGURE I.

LAND, SEA AND HOME

Maximal and minimal visions of the Scandinavianization of North-West England. (Drawn by the author)

ultimately from Yorkshire or north-east Mercia, some of whom may have arrived in the 9th as well as the 10th century. Fellows-Jensen's work illustrates, additionally, one of the great difficulties of onomastic research in this region, relying as it does in many instances on manuscript sources of the 15th or 16th centuries. Indeed, this lateness of the evidence must be recognized as a serious weakness to any interpretation of settlement in the Viking Age. To this point this essay has implicitly conceded the theory that numerous place­ names formed in a particular language necessarily translate into large-scale immigra­ tion by speakers of that language. However, scholars are increasingly questioning the nature of this relationship. Edward James's map of Britain c. 900 is indicative of the

LIFTING THE VEIL?

FIGURE 2.

301

Scandinavian settlement names in northern Cheshire, Merseyside and southern

Lancashire. (Drawn by the author)

more cautious approach now being taken, illustrating 'main areas where place-names of Scandinavian origin are found' as opposed to settlement activity itself,16 so 16 E. James, Britain in the First Millennium (London, 2001), 235; see also Kenyon, op. cit. in note II, 126, whose map of Scandinavian place-names in Lancashire carries the warning 'It is not a map of Scandinavian settlement'.

302

LAND, SEA AND HOME

distinguishing interpretation from data. In this context, it is simply no longer acceptable to treat the whole of western Northumbria and the north-western corner of Mercia as if they were a single region characterized by Scandinavian settlement in the Viking Age. Rather, in-depth studies of specific parts reveal considerable differences, and it is to the micro-level that we need now to proceed. The starting point for this process was arguably John McNeal Dodgson's study of the Wirral, published in 1957,17 but new work continues to appear. For example, research around Derwentwater in the middle of the Lake District, concluded that Scandinavian naming predominates but alongside elements of Celtic and Anglian, plus Middle English and French. 1s This patterning was interpreted as suggestive of Scandinavian settlement with a degree of hybridization indicative of some survival of the pre-existing population. However, it is important to recognize that name formation with -by continued at least into the 12th century in Cumbria, resulting in names such as Johnby, while -thwaite became a dialect word in the later Middle Ages. We stumble again, therefore, against the issue of indicative Norse language elements in use in a post-Conquest environment, with a heavily Scandinavianized vernacular present far later than active Viking immigration. What is more, it is extremely difficult to imagine how contemporary communities viewed ethnicity, but it certainly cannot be argued that at any point in the Viking Age there was a sense of a single, Norse province stretching from northern Cheshire to south-west Scotland. Rather, a series of local identities emerged at latest around the end of the period. Regional names tend to be more heavily influenced by Norse the closer they lie to the coast (as in the instances of Amounderness and Copeland) but several carry quite specific and non-Scandinavian ethnic associations. Galloway recalls Irish, rather than Norse, incomers (as may several local names such as Ireby),t9 and 'Cum bra-land' the presence of Britons, perhaps associated with Welsh Strath­ clyde, even though Norse has had a far greater impact on local names in northern Cumbria than Old British; the great forest of Inglewood is 'English Wood' and Westmorland is 'the land west of the moors' - meaning the Pennines, named presumably from York, to which kingdom and then province it was apparently still attached, for example, in 927 and 1066. The wide estates of Tostig and several other prominent Yorkshire landholders in what eventually became northern Lancashire centred on places many of which had pre-Viking names (such as Whittington, Beetham, Ashton and Halton). Amounderness, probably a Viking lordship in the early 10th century,zo was bought by King lEthelstan and granted to the archdiocese of York c. 930,21 but was held by the earl in 1066. Lancashire below the Ribble emerges J. McN. Dodgson, 'The background of Brunanburh', Saga-Book 14 (1957), 303-16. D. Whaley, 'Anglo-Scandinavian problems in Cumbria, with particular reference to the Derwentwater area', Nomina, 19 (1996), 91-II3. 19 M. C. Higham, 'Scandinavian settlement in north-west England, with a special study of lreby names', 1~99-205 in Crawford (ed.), op. cit. in note 7. G. Fellows-Jensen, 'Amounderness and Holderness', 86- 94 in L. Peterson (ed.), Studia Onomastica: Festskrift till Thorsten Andersson, 23 February 1989 (Stockholm, 1989) . 21 D. Whitelock (ed.), English Historical Documents, I (2nd cd., London, 1979), 548- 51: the date of the original is 930 but it is generally amended to 934. The charter records that 'This aforesaid donation I have bought with no little money of my own', suggesting that IEtheistan had only recently acquired it, perhaps from a Scandinavian lord. 17

18

LIFTING THE VEIL?

30 3

by c. 1000 as Inter Ripam et Mersham in the will of Wulfric Spot, and was hidated and treated as if appended to Cheshire in Domesday Book. It had arguably been detached from Northumbria as early as the 920S.22 Such sub-regional naming is indicative of the failure of Northumbria to survive the Viking Age intact, but it also provides some indication of what these communities considered to be the key markers either for themselves or for their neighbours. 'Vikingness' does not seem particularly prominent among these. Nor is it clear that contemporary communities distinguished uniformly between Danes, Norse, Irish, Britons and English in the same ways and for the same reasons as modern scholars have been inclined to do. Indeed, monastic complaints concerning the willingness of King Edgar to patronize and reward heathen foreigners highlights a normal practice of the eIite,23 which can be traced back well into the 9th century, which was to reward service and promote loyal retainers irrespective of ethnicity or cultural background. A foreign presence within the households of the great and good was expected and might easily filter through into estate tenure or clerical appointment. WIRRAL AND SOUTH-WESTERN NORTHUMBRIA

On the Continent, the establishment of Viking leaders with their followers to hold disputed and often peripheral territory on behalf of the crown was a well­ established practice in 9th-century Carolingian France. We find the same practice reflected in the putative willingness of Mercia's rulers in the very early 10th century to establish a Viking leader, Ingimund, with his followers, on the very edges of their own territory in the far north-west. Supposing our admittedly late and highly rhetorical literary source to be in any sense historical,24 the incomers were established under overall English control. This was not, therefore, an uncontrolled event, likely to have been accompanied by the eviction or slaughter of the local population. This little Viking colony is reflected in the place-name evidence, with a scatter of Scandinavian and Scandinavianized place-names, plus several Irish place-names, all concentrated in the northern half, and particularly the north-western quarter, of the Domesday hundred of Wilaveston. It had, of course, its own meeting place, at Thingwall ('assembly field'), which, like many others in Cheshire, seems to have been associated with a group of burial-type mounds.25 This was apparently a multi-racial community, including English, Irish and Scandinavians of various different origins, with perhaps even some Welsh, given that Ingimund had earlier attempted to establish himself in Anglesey.26 The Grimston-hybrids (such as Thurstaston) and the -by compounds 22 N. J. Higham, 'The Cheshire burhs and the Mercian frontier to 924', Trans. Antiq. Soc. Lancashire Cheshire, 85 (1988), 193-221. 23 Anglo-Saxon Chronicle, MSS D, E and F, s.a. 959, in Whitelock (ed.), op. cit. in note 2 1,225. 24 Radner, op. cit. in note 8. 2S A. Pantos, 'Meeting-places in Wilvaston hundred, Cheshire', J. English Place-Name Soc., 31 (1998-9), 101. The place-na mes are discussed in detail in J. MeN. Dodgson, The Place-Names of Cheshire, 4 (Cambridge,

12,72 ), 166-335.

6 Annales Cambriae, s.a. 902, in E. Faral (ed.), La Legende Arthurienne, 3 (Paris, 1929),49. For the presence of Irish speakers on the Wirral, see M. C. Higham, op. cit. in note 19, 199-201; R. Coates, 'Liscard and Irish names in northern Wirral', J. English Place-Name Soc., 30 (1997-8) , 23 n. 6, detailing Liscard, Irby, Noctorum and Arrowe.

30 4

LAND, SEA AND HOME

(Greasby, Raby) may be more reflective of Anglo-Scandinavian or Danish name­ formation than of Norwegian/Irish. If Edward the Elder was responsible for the shiring of the North-west Midlands,27 then it must be significant that the Viking enclave on the Wirral was included in the local hundredal system without any apparent concession towards its special settlement history within a generation of its establishment. Indeed, not only did these settlements sit within Wilaveston hundred alongside other communities without Scandinavian place-names, but the hundredal meeting place lay well outside this vicinity and was arguably dominated by the earl's estate of Eastham and its attendant parish of Bromborough, and beyond that by Chester itself. Scandinavian place-naming may, therefore, have some potential to exaggerate the extent to which an area should be seen long-term as a Norse space, as opposed to an area where the lingua franca became Scandinavian. In terms of survival, comparatively recent name-formation has obviously been privileged over earlier examples, and processes of place-naming which were intrinsic to long-lasting changes in patterns of lordship, landownership, land-use and parochial organization likewise. More important, however, is the evidence of numerous very disparate individuals from a variety of ethnic backgrounds all living in the same locality, so with the need to communicate effectively among themselves. The interconnectedness of this corner of the Irish Sea with other places such as the Isle of Man, Dublin, Strathclyde and northern Wales, as well as Scandinavian Yorkshire, perhaps encouraged adoption of the language of the ship-men, so the Norse, at the expense of alternatives, such as Welsh or Anglo-Saxon, to the point where it became the normative means of communication, even among individuals who neither considered themselves Scandi­ navian or were considered such by others. Scandinavian and Anglo-Saxon were, of course, closely related and are likely to have been mutually comprehensible to many individuals. A further impulse to adopt the Scandinavian language may well relate to local lordship and elite patronage. With the collapse and fission of the English kingdom of Northumbria in the 860s, pre-existing tenure of estates by both the secular and religious aristocracies was disrupted and some clearly emigrated. In their absence, a new elite came into being. The take-over was characterized in some areas at least by secularization and the fission of what seem to have been pre-existing great estates. 28 For example, hog-back tombstones in northern Lancashire and Cumbria imply connexions with York and the kingdom based there, and thus integration within a system of patronage mediating estate tenure focused east of the Pennines. Their absence south of the Lune ties in with other evidence that southern Lancashire, Merseyside and Greater Manchester were never really a part of the Norse kingdom of York, having arguably been detached from Northumbria alongside Edward the

27 See discussion of the options by S. Keynes, 'Edward, King of the Anglo-Saxons', 40-66 in N. J. Higham and D. H. Hill (eds.), Edward the Elder (London, 2001),59. 2. C. D. Morris, 'Aspects of Scandinavian settlement in northern England: a review', Northern Hist., 20 (1984), 1- 22; D. Kenyon, 'Notes on Lancashire place-names 2: the later names', j. English Place-Name Soc. , 21

(1988-9),23-53.

LIFTING THE VEIL?

30 5

Elder's construction of a burh at Manchester. 29 That said, place-name evidence suggests some sort of Norse settlement in West Derby Hundred, particularly around Ormskirk, Formby and North Meols, perhaps of a kind similar to that in the Wirral, albeit the place-name evidence is both denser and more extensive. Several names in -by suggest some input from east of the Pennines, perhaps even established from Danish York. However, pre-existing English and British names are also numerous,30 so again this may reflect a re-organization of local lordship in the Viking Age, but under some over-riding political control that was not impervious to the presence of existing farming communities. SOUTHERN CUMBRIA

The coastal fringes of Cumbria and northern Lancashire provide a pertinent case study. It was from Heversham, in Kentmere, that the abbot, one Tilred, fled eastwards to buy his way into the patronage of St Cuthbert's at Chester-Ie-Street in the first 15 years of the 10th century.31 Heversham has produced pre-Viking sculpture, as have Workington, Beckermet and Brigham, which may also have been clerical estate centres. Norse place-names are numerous around Kentmere. Heversham and Kendal are unusual among pre-existing ecclesiastical centres in having produced no Viking­ period carvings, implying that real discontinuities may have occurred here, with a complete change to existing monastic tenure of local estates and the marginalization of the churches they had supported. Hiberno-Norse place-names are particularly noticeable around such prominent church sites as Heversham and Dacre, both of which may well have lost estates to Norse incomers.32 The group of Viking-period sculptures at Gosforth suggests the establishment of a new church in an area where Christianity was no longer omnipresent, presumably under the patronage of new secular lords prepared to put considerable resources into cult. Yet even the explicitly 'pagan' motifs on, for example, the magnificent Gosforth cross, were presumably the work of an existing skilled workforce which had hitherto served the clerical establishment in the region. MIGRATION AND SETTLEMENT

Immigration certainly occurred, therefore, but the displacement of the existing population was far from wholesale and seems to have varied from locale to locale. The general picture is one characterized as much by disruption as by new settlement and economic expansion. While the place-names of central Lakeland do suggest significant colonization during the Viking Age, pollen diagrams from around 29

N. ]. Higham, 'Northumbria, Mercia and the Irish Sea Norse, 893- 926', 21-30 in Graham-Campbell (ed.),

0.£. cit. in note 15.

British place-names are particularly common around the mosslands in the vicinity of Walton on the Hill, Speke, Allerton and Woolton, Ince Blundell, Maghull and Burscough: see Wainwright/Finberg (ed.), op. cit. in note 3, 26-40; P. B. Russell, 'Place-name evidence for the survival of British settlements in the West Derby hundred (Lancashire), after the Anglian invasions', Northern Hist., 28 (1992),25- 41. 31 C. Hart, The Early Charters of Northern England and the North Midlands (Leicester, 1975), doc. 158. 32 N. ] . Higham, The Kingdom of North umbria, AD .3.50 - IIOO (Stroud, 1993), 197.

306

LAND, SEA AND HOME

Morecambe Bay, and further south in Lancashire and Cheshire, show marked regeneration of woodland and other plants indicative of low levels of economic activity around the millennium. 33 This evidence suggests that some localities suffered a decline in farming and grazing at this date, and so some lowering of settlement activity. This evidence is at odds with most interpretations of the Scandinavian place­ names, which occur widely on marginal land in Cumbria and northern Lancashire, as evidence for increasingly extensive land-use during this period. This decline of land­ use may, of course, have been linked to the collapse of the Northumbrian state in the late 9th century, which created a power vacuum in the region to which local communities and immigrants together needed to react. However, the palynological evidence does cast some doubt on the vision of mass immigration so effectively popularized by Wainright. Nor were the Irish Sea Vikings obviously eager to settle in north-western England. The Irish successfully expelled the Scandinavians from Dublin around 900, and the assumption on the part of modern scholars that they fled to one or more of the estuaries or natural harbours of western England is broadly probable, with the Ribble as the favoured location. 34 This may help explain the presence of the Cuerdale hoard there/ 5 and the Ribble always offered the best route from the Irish Sea to York. However, it must be stressed that this was more an army in transit than a group of settlers in search of farmland. Of course, Viking armies had a tendency to split into their component groups. However, there is circumstantial evidence that this force retained a degree of cohesion, since they fought a series of campaigns against the rulers of Bamburgh and their Scottish allies before capturing York and ultimately regaining Dublin as well. 36 This arguably required the bulk of their manpower and all the reinforcements which they could attract. Their success opened considerable opportunities for settlement in more advantageous conditions than pertained in much of North-west England for any that stood aside. The settlement of the marginal uplands and moss-edges of North-west England is unlikely to have been high on the lists of priorities of even the most disenchanted of these warriors in the early 10th century. Indeed, the passage of Viking armies during the reign of Edward the Elder may even have encouraged (or forced) locals to join up in search of gain and adventure. In that case they may eventually have moved to the newly re-founded Scandinavian settlements in Ireland or taken advantage of improved access to patronage at Ragnald's court at York. Any notion that we are necessarily dealing with a supine local population which was incapable of adapting to new circumstances, and to whom things were just done throughout the Viking Age, is unhelpful, to say the least. We can dispense, therefore, with the monolithic vision of a mass migration of Scandinavians into North-west England. In seeking to replace this overly simplistic 33 M. Leah et aI., The Wetlands of Cheshire (North West Wetland Survey, 4, Lancaster, 1997), 154; D. Hodgkinson et aI., The Lowland Wetlands of Cumbria (North West Wetland Survey, 6, Lancaster, 2000), 41.- 8. • As first proposed by the author, op. cit. in note 29.

3S Ibid.

36 Wainwright/Finberg (ed.), op. cit. in note 3, 163- 80; N. Higham, op. cit. in note 32, 186-8.

LIFTING THE VEIL?

30 7

vision, we should postulate a model which assumes a degree of population movement both into and out of the region. However, that was only one factor affecting the development of a sub-set of localized cultures and identities along the north-western seaboard of England, and the spread also of Scandinavian as a spoken language across the region. Where 'Viking' settlement did occur - whatever precisely that meant ­ we can be reasonably sure that it did so within some sort of political framework and system of patronage, as opposed to an uncontrolled wave of immigration. Indeed, the 'Viking' factor was merely one of several concurrent pressures for change. This is particularly apparent in terms of estate and settlement organization, which seems to have been changing in various ways throughout the region. The diminution of the landholdings and political power of the greater churches is one factor in this equation, which seems to have been as prevalent in the 'English' regions of North-west Mercia as in western Northumbria. Take, for example, the group of carved stones indicative of a major and wealthy church at Sand bach around A.D. 800. By 1066, this once apparently wealthy minster had no lands and Sand bach was split into two small manors, one of 21 virgates and a slightly larger one of I hide and 11 virgates, to which the priest and church were attached. 37 The rulers of Mercia had reacted to changing circumstances early in the 10th century by building new burhs in Cheshire, Greater Manchester and perhaps Lancashire, and apparently re-organizing land, cult, labour and a share of the agricultural surplus in their support. 38 The new shire was re-ordered around the great fortress of Chester, the new royal minster of St Werburgh's and the increasingly important regional mint. However, these assets were thereafter delegated, so the crown, too, had given up all its estates in Cheshire by 1066. This left the Earl and the Bishop of Lichfield as the major landowners alongside a multiplicity of minor figures, with varying connexions with their aristocratic neighbours. In 1066 most of these had Old English names, but some few had names which might indicate a Welsh or Scandinavian background. The processes of secularization and estate-fission were common-place, therefore, not just in areas with Scandinavian place-names but equally in others, and so were presumably part of a much wider pattern of change connected with new patterns of patronage and the imposition of royal justice and the land-geld, profits from which were arguably maximized where land was held by numerous laymen. One result was the pattern of often small, local estates, which developed into the manors of the later medieval period. Many of these were apparently co-extensive with townships, which presumably came into existence during the same period. However, there is extensive evidence of townships in Cheshire divided between two or more estates by 1066, which may imply that the process of estate-fission continued past the date when the townships were established. Many manors at Domesday in Cheshire and Lancashire were little more than large single farms. So too were groups of contiguous estates/ townships held by particular individuals, whose holdings later re-emerged as the baronies of Cheshire, focused often on the same sites (as at Halton, Dunham and P. Morgan (ed.), Domesday Book: Cheshire (Chichester, 1978), I, 33; 14, 10. A. Thacker, 'Kings, saints and monasteries in pre-Viking Mercia', Midland Hist ., 10 (1985), 1-25; idem, 'Anglo-Saxon Cheshire', in B. E. Harris (ed.), Victoria History of Cheshire, I (London, 1987), 272; N. ]. Higham, op. cit. in note 22; idem, The Origins of Cheshire (Manchester, 1993), passim. 37 38

308

LAND, SEA AND HOME

Malpas). It is not too much to suggest that the pattern of aristocratic and what would later be described as gentry land-tenure in Cheshire, at least, began in the Viking Age. SETTLEMENT ARCHAEOLOGY

This period was witnessing, therefore, a significant shift in the relationships between local communities and their agricultural surpluses, the regional aristocracy and the crown, which had a considerable bearing on how land was exploited on a very local level. Rural settlements of this period have now begun to be identified, but only with great difficulty since the countryside seems to have been virtually aceramic until c. 1200. This renders it extremely difficult to distinguish occupation dates across a 700-year period. A site such as Bryant's Gill, in Kentmere, is an example of an Anglian or Viking-age settlement,39 with parallels with other upland sites such as Ribblehead and Simy Folds (this volume), as well as other so-far unexcavated settlements identified on Orton Scar. 40 Elsewhere, it has been possible to build up a picture of settlement and land-use across the medieval period by large-scale trenching, but the same problems of dating apply. At Tatton, in Cheshire, excavation centred on the remains of a deserted medieval village, which was a consequence of emparkment in the 18th century.41 This has revealed a complex history of land-use discernible intermittently from prehistory through to the present. The deserted medieval village site was first occupied in the Neolithic and saw some activity in later prehistory, but the earliest substantial remains were of two timber-framed buildings set in a saucer­ shaped, palisaded enclosure, radiocarbon-dated to the Late Roman and/or early sub­ Roman period. 42 Characteristically, there was a total absence of datable artefacts. This site was then reoccupied at a date which is ill-identified but which was apparently contemporary with the appearance of Chester Ware on site. This comprises the earliest phase of the medieval settlement. The principal feature of this primary phase of re-occupation was a timber­ framed building 14.4 m x 4.6, built with all four walls in plan bowing slightly outwards, aligned approximately NE.-SW. (Fig. 3). The long walls provided the support for the roof, with posts which were probably paired and which were presumably joined by rafters. A single post-hole in the centre of each end wall was probably intended to provide support for the ridge ends. The interior surface had been ploughed away but an area of fire-hardened subsoil arguably indicated a hearth. Adjacent ditches seem to have been developed for run-off from the roof, which would have been the more necessary since this building had been constructed within the 'Dark-age' saucer-shaped enclosure. 39 S. Dickinson, 'Bryant's Gill, Kentmere: another "Viking-period" Ribblehead?', 83- 8 in J. R. Baldwin and I. D. Whyte (eds.), The Scandinavians in Cumbria (Edinburgh, 1985). 40 N.] . Higham, 'An aerial survey ofthe upper Lune Valley', 34-5 in N.]. Higham (ed.), The Changing Past: Some Recent Work in the Archaeology orNorthern England (Manchester, 1979). 41 N. ]. Higham, 'The Tatton Park project, pa rt 2: the medieval settlements, esta tes and halls', J. Chester Archaeol. Soc., 75 (1998-9),61-133. 42 N. ]. Higham and T. Cane, 'The Tatton Park project, part 1 : prehistoric to sub-Roman settlement and land use', J. Chester Archaeol. Soc., 74 (1996- 7), 1-61.

LIFTING THE VEIL?

30 9

Tatton (Cheshire). Deserted medieval village, building F: the earliest building of the medieval reoccupation, defined by two parallel lines of post-holes, with a single post-hole at the centre of each end. The structure is cut by later medieval drainage ditches, VII, VIII, X and XI. The short ditch IX may be contemporary with the building. (Drawn by the author) FIGURE 3.

310

LAND, SEA AND HOME

It is important to note that this apparently Late Saxon or Viking-age phase reflects the reoccupation of the site following an abandonment of some centuries. It is the first phase of occupation of the principal medieval settlement at the core of the larger of the two manors into which Tatton was already divided by 1066. That division presumably conditioned reoccupation of this site. There are important links forward in terms of architectural style to other phases of medieval building, which certainly suggest that these occurred within a continuum. This building represents, therefore, the first phase of a period of medieval occupation of a manorial hamlet or a village, which was founded in the Viking Age. Similar evidence of reoccupation of earlier settlements for new medieval tofts is gradually being discovered in Merseyside and southern Lancashire,43 although elsewhere the earlier sites have been predominantly Romano-British and the medieval sites not certainly any earlier than the I3th century. In many parts of the region, such apparent continuity of land-use should be expected, if only because poor environ­ mental conditions placed a premium on the better quality habitation sites. CONCLUSIONS

To sum up, therefore, the North-west did receive incomers from surrounding areas of the British Isles during the Viking Age, although it certainly lost some members to other regions during the same period. Some immigrants might have been considered Vikings - and these were probably the largest single group - but others may well not have been, and some certainly were not. Some of the Vikings were in origin western Scandinavians and some eastern. There is little evidence of an overall population increase and some signs of a recession in some areas as regards farming activity. Major political and cultural changes in Mercia and Northumbria impacted on the region, and different local communities adapted themselves to these processes of change in various different ways. Scandinavianization was only one part of this process, albeit one with a high level of visibility today via the stonework and place­ and field-name evidence. 'Englishness' had never been so pervasive in these western margins of Northumbria and Mercia as in their core territories, so that a degree of cultural diversity may already have existed when the Viking Period began. With Northumbria's integrity collapsing during the 860s, different parts of the regional community then developed in different ways, often conditioned by the nature of political control and the elite culture in place when estate-fission was taking place. As is becoming apparent in Yorkshire,44 the pattern of medieval settlement owes a substantial debt to changes taking place in the Viking Age. Even these sub-regions, however, display numerous local differences, and a more detailed, more complex and

43 R. W. Cowell and R. A. Philpott, Prehistoric, Romano-British and Medieval Settlement in Lowland North West England (Liverpool, 2000), 213. 44 J. D. Richards, 'Identifying Anglo-Scandinavian settlements', 295-310 in D. M. Hadley and]. D. Richards (eds.), Cultures in Contact: Scandinavian Settlement in England in the Ninth and Tenth Centuries (Turnhour,

2000),30 3-5.

LIFTING THE VEIL?

3I I

more sensitive modelling is now called for as we seek to evaluate the impact of the Scandinavian world on north-western England. Dr Nicholas ]. Higham, Department of History , University of Manchester, Oxford Road, Manchester MI3 9PL, UK nick.j [email protected]

SEDGEFORD: EXCAVATIONS OF A

RURAL SETTLEMENT IN NORFOLK

By

SOPHIE CABOT, GARETH DAVIES

and

RIK HOGGETT

The Sedgeford Historical and Archaeological Research Project (SHARP) is an unfunded, volunteer staffed project which began in 1996. SHARP's remit is the study of the full extent of human settlement and land use within the parish of Sedgeford, NW. Norfolk (Fig. la-b). The project has worked on many different aspects of the study area, including surveying the parish church and excavating within the post­ Conquest settlement, but the main focus of work has remained a field to the south of the River Heacham. Known as Boneyard Field, it is here that the Middle to Late Anglo-Saxon phases of the settlement are to be found, and a contemporary Christian burial ground (Fig. IC). This is the site that forms the subject of this paper. The later medieval phases of Sedgeford cluster around the Norman parish church in the west of the settlement (Fig. Ib). Boneyard Field, a focus of Early-medieval settlement, lies on the opposite side of the river in an area that has remained undeveloped. The only later feature to have affected the site is a dam across the river, created at around the time of the Norman Conquest, which flooded the area upstream to create fishponds and reed beds, and to protect the growing settlement-area downstream. The investigation of the reasons behind this settlement shift remains one of the project's main research objectives. EARLY WORK ON BONEYARD FIELD

The earliest reference to archaeological features in Boneyard Field dates from 1913. The owner of Sedgeford Hall, Holcombe Ingleby, wrote of his discovery of eight inhumations at a location he described as being 'near my house'. He also commented that 'many skeletons have from time to time been exhumed along the side of the stream that runs through Sedgeford Valley', although he gave neither sources for these earlier discoveries or their locations. 1 When his daughter donated some of the human remains to the Royal College of Surgeons, she produced a map showing the location of the site (marked with a cross). The position given for the burials coincides with the area of the Boneyard Field, and it is assumed that they were a part of the burial group discussed below. The first assessment of the archaeological potential of the Boneyard Field was in 1954, when ploughing revealed human remains. The Norfolk Research Committee 1

H. Ingleby, The Charm of A Village (London,

I 920) .

LAND, S E A AND HOME

314

FIGURE

I.

Location map. (a) Site location; (b) Boneyard Field, Sedgeford; (c) Areas of excavation in Boneyard Field.

evaluated the site and limited excavations identified the existence of a Middle or Late Anglo-Saxon inhumation cemetery.2 A newspaper account at the time reported that the remains were 'equally spaced, [and] showed no signs of hasty burial'.3 Although it has not yet been possible to examine the material archive, there is no doubt that these inhumations should also be considered a part of the Boneyard Field skeletal assemblage. Further excavations were carried out in the Boneyard Field in 1957 by]. G. Hurst. Two rows of trial trenches were dug across the field, one running N.-S. and 2 3

MONARCH Database No. 641022.

'Ancient graves found at Sedgeford', Eastern Daily Press, 6 April 1954.

SEDGEFORD

315

the other E.-W. (Fig. IC). A number of skeletons were unearthed from the eastern side of the field and settlement evidence in the form of wattle-and-daub and midden debris was located in the western side. It was decided that more extensive excavation was required and this was conducted by Dr Peter Jewell in the summer of 1958. Jewell's excavation was never fully published, but SHARP was fortunate enough to obtain his notes and a draft excavation report before his death in 1998.4 Guided by the 1957 test pits, he excavated areas of both cemetery and settlement (Fig. IC). Although he was unable to link the two areas stratigraphically, he identified several phases of activity within each zone. He noted that several of the excavated skeletons were stratified above a horizon containing sherds of Ipswich Ware and that all of the burials were oriented W.-E. in unfurnished graves. He concluded that they were Christian burials of a Middle Anglo-Saxon date. None of the excavated burials were overlain by any discernible settlement deposits, although Jewell was confident that there 'was little doubt that the cemetery was contemporary with the habitation'.5 Jewell also detected two phases of activity in the settlement area of the site. The first phase was a series of interconnected gullies, one running E.-W. with several smaller gullies running off it, down the slope to the north. A second phase of activity which post-dated the silting up of these features involved the construction of a long timber building, half of the foundation trench of which was completely excavated. Traces of burnt daub were found within the foundation-trench and the building had a doorway which was 2 m wide, flanked by large post pits along its southern side. J ewell stated that the first settlement phase - the gullies - could be dated to the Middle Anglo-Saxon period by Ipswich Ware (produced from the early 8th to mid­ 9th centuries),6 but he was less confident about the date of the foundation trenches of the timber building. These contained more Thetford-type Ware (produced from the 9th to late lIth centuries) than Ipswich Ware, suggesting a Late Anglo-Saxon date for when they fell into disuse. THE RECENT EXCAVATION

The geology of the Boneyard Field consists of a loose glacial subsoil of sand and gravel. The subsoil grades to a fine sand towards the bottom (N.) of the valley. The site has produced large numbers of Middle Anglo-Saxon to Late Anglo-Saxon features. Interpretation of the site is complicated by a number of post-depositional processes? D. M. Wilson and]. G Hurst, 'Medieval Britain in 1958', Medieval Archaeol., 3 (1959),298. P. A. Jewell, 'The Excavation of a Middle Saxon Occupation Site and Burial Ground at Sedgeford, Norfolk' (unpub!. rep., nd.), 12-13 . 6 Wilson and Hurst, op. cit. in note 4. 7 Boneyard Field presents various problems in interpretation. The slope of the site from south-east to north­ west has resulted in colluvial build-up varying in depth from 0.4 m in the south to almost 2 m in the north. This has mixed artefacts and transported pottery downhill. It can be difficult to recognize cut features above the level of the natural sand, some of which must have been cut through earlier colluvial layers. Damming of the river has stained areas of sand a dark grey colour, making the detection of cuts especially difficult. There has also been significant bioturbation of the site by rabbits, while deep ploughing has truncated the archaeology along the top of the slope. 4

5

3 16

LAND, SEA AND HOME

The current excavation, which has uncovered an area measuring 65 m x 25, has produced evidence for a number of phases. The earliest phase is represented by large quantities of residual late Iron-age pottery (Belgic-type wares) and by one or two datable Iron-age features. 8 Following this, the site seems to be abandoned for some time. The main Early-medieval phases are: 1. Middle Anglo-Saxon Christian burial; 2. Two main phases of Middle to Late Anglo-Saxon structural and drainage features; 3. A Late Anglo-Saxon midden, and an area possibly relating to some informal industrial process.

The Middle Anglo-Saxon cemetery The earliest phase of the Anglo-Saxon site consists of 161 inhumations in varying degrees of preservation (Fig. 2a). These burials are oriented W.-E. and are without grave goods, consistent with Christian burial practice. The arm- and leg-positions of many burials suggest tight shrouding. In the east of the excavation area there are also coffin burials, which exhibit more 'bone tumble' and occasionally have iron coffin­ brackets in situ. It is not clear whether these burials were also shrouded inside the coffins. The most concentrated area of burial is to the north of the site, near to the river and on fairly flat ground. Here, a total of 61 burials has been excavated, with at least four intercutting phases, within a 5 m x 10 trench. Although the extent of the cemetery has not yet been defined, this would appear to be the most densely occupied burial area. This area was certainly reused, whereas burials further up the slope rarely intercut. None of these burials are dated by association with artefacts, but an initial radiocarbon date from a burial early in the sequence has provided a date of cal. A.D. 600-760 (GrN-25159)· A Middle Anglo-Saxon start date for the cemetery, possibly as early as the 8th century, is well supported by the artefactual evidence from the later settlement phases. The majority of the small finds date to the 8th and 9th centuries, providing a useful terminus ante quem for the burials, but it must be remembered that there is considerable post-depositional disturbance. The Middle Anglo-Saxon settlement The two phases which immediately post-date the cemetery appear to represent the margins of an area of residential or light industrial occupation (Fig. 2b). The lighter shaded features form the earlier phase, while the darker ones truncate them, but artefactually they are of very similar character. Jewell also observed two main phases of cut features in the 1950S.9 The earlier phase appears to include elements of severely truncated structures and N.-S. drainage gullies. The structural evidence has • It is envisaged that this phase will become a focus of SHARP research in the next few yea rs. Publication of the 2003 Sedgeford hoard is in hand. 9 Op. cit. in note 5.

FIGURE 2.

Features in the boneyardfieldtrench (1996-2000). (a) Burials; (b) Middle Anglo-Saxon features; (c) Late Anglo-Saxon features; (d) Saxo-Norman and post-Conquest features.

SEDGEFORD 317

3 18

LAND, SEA AND HOME

been highly disturbed but clearly falls into two concentrations. To the south of the site there is a relatively flat, possibly terraced area, where there are two E.-W. gullies, the northernmost of which has four post-depressions within it. This is associated with twenty-two small post-holes. The features to the south, which appear to have been truncated by erosion or deep ploughing, may be part of a timber 'hall', or various rebuilds of a smaller structure. The post-holes appear to be too small to represent structural timbers and may represent the remains of an internal division, the gullies themselves forming the beam slots. The surface within this structure appears to be devoid of flint, as if it has been cleared, whereas other areas contain dense flint concentrations. This structure, although severely truncated, fits well into the timber hall type and is closely paralleled at other sites in Norfolk. 10 The second structural concentration, towards the centre of the main trench, consists of a shallow pit-like feature and a number of scattered post-holes. This may represent a structural element similar to the pits associated with sunken-featured buildings, but in this case it is apparently within a larger hall structure. Alignments in the post-holes are inconclusive, and this structural area remains enigmatic. The features contain only Ipswich Ware, indicating that they are of Middle Anglo-Saxon date. l l It seems likely that the ditch complex to the east of the site has a drainage function, and the later gully is riverine in appearance (Fig. 2b). The feature provides a useful stratigraphic indicator, as it truncates earlier burials. The features from these phases have very similar artefactual assemblages, being dated by quantities of both Ipswich and Thetford-type Ware. Associated post­ cemetery layers have produced artefacts which are predominantly of 8th- and 9th­ century date, including bone-comb fragments, dress pins and an Anglian silver penny of King Eadwald (769-798). Two styli have also been recovered from these post­ cemetery layers, as well as fragments of decorated vessel glass, which may be indicative of 'higher status' use. Some features in the supposed later phase contain only the earlier Ipswich Ware, which in isolation would suggest a Middle Anglo-Saxon date. However, these features are clearly stratigraphically later than gullies which contain both Ipswich and Thetford-type Wares. In all of these largely homogeneous fills the proportions of the Ipswich to Thetford-type Wares is 1:4 which suggests a date for some of the fills later in the Anglo-Saxon Period. However, the overall Ipswich-ware assemblage is unusually large (around 2,000 sherds), suggesting an earlier date. Further excavation of the upslope area may show that much of the Thetford-type Ware is derived from a later settlement nucleus in this location. In the absence of further information, it is not yet possible to date these phases more closelyY

10 A. Rogerson, A Late Neolithic, Saxon and Medieval site at Middle Harling, Norfolk (East Anglian Archaeol., 74, Gressenhall, I995). 11 Ipswich Ware is now dated c. 720-850 (Paul Blinkhorn, pers. comm.). 12 This date-range will probably only be further clarified by a systematic programme of radiocarbon dating, which is being undertaken as funding permits.

SEDGEFORD

31 9

Late Saxon settlement (Fig. 2C) After the Middle Anglo-Saxon to Late Anglo-Saxon phase there is a small amount of colluvial build-up and the site seems to have changed character and function, possibly becoming a marginal area at the edge of a Late Anglo-Saxon settlement. Deposits associated with this phase contain vast quantities of animal bone (mostly butchered), and mixed Late Anglo-Saxon pottery, forming an archetypal midden. The most extensive feature in this area consists of layers of cobbling (possibly to consolidate an area that was becoming increasingly waterlogged) and pits containing large amounts of burnt clay which may be oven linings or rake-out. These pits may represent an informal industrial process and it is interesting to note that our only samples of bread wheat come from this phase. The nucleus of activity would appear to have shifted during this last settlement phase to somewhere beyond our current area of excavation. One tantalizing glimpse of the scale of the settlement activity is a ditch-terminal, which is about 1.5 m wide and the same measure in depth. The ditch has fourteen re-cuts and clearly relates to a much-used, well-kept part of the Late Anglo-Saxon settlement. Saxo- N orman/H igh-Medieval drainage Finally, there are a number of Saxo-Norman and later Medieval features (Fig. 2d). A large N.-S. ditch truncates all of the other features, but may still in theory be of Late Anglo-Saxon date (stratigraphically being very late). The boundary of the marshy area, used for the cultivation of reeds after the 13th century, is represented by an E.-W. cut and re-cut at the northern extent of the trench which truncates burials. The line to the south of this represents a large natural 'cut' caused by erosion due to marshy conditions or flood action after the Anglo-Saxon Period. This may well truncate features that might otherwise have given more detail to the Middle Anglo­ Saxon to Late Anglo-Saxon structural features. SEDGEFORD IN CONTEXT

A number of sites which have produced evidence through excavation or field walking appear to be similar to the Boneyard Field site. Of those where excavation has taken place, the best known examples are Brandon in Suffolk and Flixborough in HumbersideY Other sites include Cottam and West Heslerton in Yorkshire/ 4 and Bawsey in Norfolk (where a Time Team investigation recorded six styli and four series R sceattas}Y Brandon and Flixborough, which provide the closest parallels, are both large excavations, many times larger than the Boneyard Field trenches. Although the \3 R. D. Carr, A. Tester and P. Murphy, 'The Middle Saxon settlement at Staunch Meadow, Brandon', Antiquity, 62 (1988), 371-7; C. P. Loveluck, 'A high status Anglo-Saxon settlement at Flixborough, Lincolnshire', Antiquity, 72 (1998), 146-6I. \4 J. D. Richards, 'Cottam : An Anglian and Anglo-Scandinavian settlement on the Yorkshire Wolds', Archaeol. } ., 156 (1999), l-lIO; D. Powlesland, 'Archaeological excavations 1987-1990: an interim report on the Anglo­ Saxon village at West Heslerton, North Yorkshire', Medieval Settlement Res. Group Ann. Rep., 5 (r990),

37-40 . 15

Time Team/Channel 4, 1999 Series Report (London, 1999) .

320

LAND, SEA AND HOME

quantity of evidence available from the various sites is very different, this may represent the different extent to which they have been investigated. Boneyard Field has a lower number of features than the larger excavations, but the feature density in many areas could be about the same. The same is true of the finds assemblage: excavations at Brandon produced three styli, while Boneyard Field has produced two. However, Brandon also produced 234 bronze pins, 8 sceattas, 360 glass fragments and various other 'high-status' finds. Boneyard Field has produced 23 pins and I sceatt. The areas of excavation are 12,000 sq m and 1,000 sq m respectively. The Flixborough excavations did not uncover as extensive an area as those at Brandon, but the types of find are once again similar, and the structural features on some parts of the site are not unlike what we have found on Boneyard Field. This is particularly true if the current Boneyard Field trench is interpreted as a peripheral area of one of these sites. The similarly dated sites at Yarnton in Oxfordshire and Burrow Hill in Suffolk also share similarities with the Boneyard Field such as features and finds. 16 Sedgeford appears to belong to a wider group of Middle Anglo-Saxon sites which is starting to be recognized from excavation and finds recording, particularly in East Anglia. This raises questions about their nature, frequency and character. The sites of this type which have been known for longest have traditionally been interpreted as high-status, mostly on the basis that the styli which they all have in common were thought to be related to monastic literacy. Since styli have now been found on a wide range of sites, this interpretation can no longer be sustained, and the status of this type of site is still unresolved. I? Sites like Cottam were interpreted as high status until they were excavated, but the structural evidence found no longer supports such an interpretation. Yarnton also seems to fall short of what one could call high-status in terms of the building evidence, whereas both Brandon and Flixborough are on a grander scale. Within Norfolk most of the sites that might, like Sedgeford, fit into the emerging class of 'productive' sites, are unexcavated. The term is problematic, but now chiefly indicates concentrations of metal-detecting finds. Improved recording of metal findspots in recent years has increased the available data-set dramatically, and much more is known about such sites in NW. Norfolk. Work on surface find-spots in Norfolk has identified at least seven or eight relevant places: West Walton, Congham, Rudham, Crimplesham (one stylus fragment) and Wormegay (two styli), as well as the potential complex of sites at the Burnhams. 18 The metalwork and pottery found on all these sites is similar to that found at Sedgeford. At present such sites cannot be placed into a category of status based on the structural evidence, and on the basis of the metal finds they tend to be described as 'important' or 'high-status' centres. Only 16 'Yarnton Saxon and Medieval Settlement and Landscape: Results of Excavations 1990-96' (unpubl. rep., Oxford Archaeol. Unit, 1999) ; V. Fenwick, 'Insula de Burgh : excavations at Burrow Hill, Butley, Suffolk 1678- 1981', Anglo-Saxon Stud. Archaeol. Hist., 3 (1984), 35-54. 7 The identification and interpretation of styli is reviewed and discussed in detail in T. Pestell, Landscapes of Monastic Foundation: The Establishment of Religious Houses in East Anglia c. 6.50-1200 (Woodbridge, forthcoming) . 18 P. Andrews, 'Middle Saxon Norfolk: evidence for settlement, 650-850', The Annual (Norfolk Archaeol. Hist. Res. Group, 1992).

SEDGEFORD

321

much more extensive excavations than are currently possible would show us how well the structures on these sites fitted the metalwork assemblages. At Sedgeford it appears that such a find-assemblage can be associated with a comparatively 'ordinary' structural sequence, in a location which is not historically documented. SEDGEFORD IN THE VIKING PERIOD

During the period of the settlement at the Boneyard Field, East Anglia experienced two major periods of Viking conquest, and was effectively under a strong Scandinavian influence from the late 9th century onwards. How do we recognize these events archaeologically and what effect, if any, did they have upon the settlement at Sedgeford? The Anglo-Saxon Chronicle tells us that in 865 ' ... a great heathen raiding-army came to the land of the English and took winter-quarters from the East Anglians, and were provided with horses there, and they made peace with them'.19 Although the exact location of this incursion is not specified, it is generally thought to have been at least in part in Norfolk, and the record that' ... the raiding-army went from East Anglia over the mouth of the Humber to York city .. .' in 866 suggests that their route could have included the North Norfolk coast and The Wash. The Danes returned to the region in 869, and were still there in 879 when, under the terms of the Treaty of Wedmore, ' ... the raiding-army went from Cirencester into East Anglia and settled that land, and divided it Up,.20 The eastern Danelaw remained an independent kingdom ruled by Guthrum and his successors until it was reconquered by Edward the Elder in 917. Given everything which has been said above about the Boneyard Field, it seems that the settlement at Sedgeford ought to have a Viking phase, showing significant evidence of having been occupied through the period of Scandinavian influence in East Anglia. That none has been recognized is a pattern familiar to excavators of other East Anglian sites but it remains a fact that the settlement was, for a period of about 40 years, within the Danelaw. Alternatively, one might expect to find evidence for a break in occupation. It has been suggested within the project team that the end of cemetery use may represent the start of (at least partially pagan) Viking occupation. However, the edges of the burial area have yet to be identified and there is no reliable end-date for Christian burial on the site. The available evidence points to continuous settlement at Sedgeford from the 8th century, with the activity represented on Boneyard Field gradually shifting location towards the present church. Late Anglo-Saxon evidence (Ioth- to I Ith­ century date on the basis of the Thetford Ware) has periodically been uncovered alongside the church and between the two sites by SHARP and others.21 The site appears to drift westwards along the valley without a clear hiatus at the time of the Danish or Norman conquests. M . Swanton (trans.), The Anglo-Saxon Chronicles (London, 2000), 68. Ibid., 76 . 21 N. Cooke, A. N. Gardener and G. Thomas, 'Report of excavations at Sedgeford, Norfolk 1996', Papers. Inst. Archaeol., 8 (1997), 17-37. 19

20

322

LAND, SEA AND HOME

FIGURE 3. Borre-style brooch from Sedgeford. Scale 1:1. (Drawn by R. Ludford)

East Anglia has a large number of Scandinavian place-names, which Margeson has argued are the most concrete testimony of the Vikings' presence in Norfolk. However, a study of Scandinavian place-names has failed to find even significant field­ names in Sedgeford that can be shown to be Scandinavian in origin.22 If there was a large-scale immigration of Scandinavian settlers, founding and naming settlements, then the material evidence from sites in such areas might be expected to reflect this. While Margeson has shown that the quantity of Scandinavian­ style metalwork finds recovered from Norfolk reflects a far more densely populated county than has previously been supposed, her distribution maps indicate that there is very little correlation between the place-name evidence and these finds.23 Some objects found in the parish are clearly of Scandinavian style, but these are few in number, and they have not been found in a settlement context. A Borre-style brooch, (now the project logo) was found during fieldwalking in an area to the south of the site which test pitting had shown to be barren of settlement (Fig. 3). An interesting Urnes-style mount, discovered by metal-detecting some years ago is also thought not to come from the settlement area. The Borre-style brooch is an object-type common in East Anglia and thought to be produced locally, possibly at Thetford. The mount, although rarer, is also thought to have been made in England rather than to be an import. 24 Evidence for Scandinavian building- and burial-types is lacking throughout East Anglia, in stark contrast to other areas of known Scandinavian immigration. Pestell has summed up the situation accordingly: 'These diverse elements all tend towards the Viking involvement in East Anglia being one concentrated on the upper echelons of society in which mass immigration did not occur.'25 The prevailing view, based upon the available evidence, must be that despite an undeniable Scandinavian presence in East Anglia, its impact should not be exaggerated and continuity in Christian Anglo-Saxon society should be expected. Pestell has also argued that the similarity between the Middle Anglo-Saxon secular and religious estate centres (as far as the two can be differentiated) would have made them both vulnerable to changes in the upper levels of society, and that a change at the top might involve an administrative rather than a material change - something difficult to identify in the archaeological record. This would appear to be the case at 22 H. C. P. Willcox, 'A Dictionary and Analysis of the Fieldnames of the Parish of Sedgeford' (unpubl. M .A. diss., University of Nottingham, 1997). 23 S. Margeson, The Vikings in Norfolk (Gressenhall, 1996), figs. 3-4. 24 O. Owen and R. Trett, 'A Viking Urnes style mount from Sedgeford', Norfolk Archaeol., 37 (1980), 353-5. 25 Pestell, op. cit. in note 17.

SEDGEFORD

32 3

Sedgeford, where there is a continuous occupation sequence and no readily discernible change in the character of the settlement. While there is evidence of Scandinavian influence in the area, and the two surface finds mentioned above, from a material perspective the settlement appears to continue with very little disruption. There may well have been wider administrative changes within the estate associated with the settlement, but the reconstruction of this is made difficult by the lack of pre-Domesday documentary material. Sedgeford lay in the Domesday hundred of Smethdon, and it has been suggested that Sedgeford was the manorial centre of the hundred and that ' ... at the time of Domesday ... their jurisdiction was nearly always in royal hands'.26 In the case of Sedgeford, the royal hands at this time were those of Earl Gyrth, and Hart concluded that most local jurisdiction was permanently in royal hands from 917 onwards. If true, this suggests that Sedgeford may have been a royal estate from the early 10th century. However, this is entirely conjectural, and there is no definite evidence, historical or archaeological, for this. ACKNOWLEDGEMENTS

While this article is the combined effort of three authors, the SHARP Project is the combined effort of many more. In particular, we would like to thank our co-directors Anj Cox, Neil Faulkner, Chris Mackie, Pat Reid and Keith Robinson. None of the work reported here would have been possible without the several hundred volunteers who have taken part in the last seven years, or without the co-operation of the landowner, Professor Bernard Campbell, and the community of Sedgeford.

Sophie Cabot, 4 Thorpe Mews, Yarmouth Road, Norwich, Norfolk NR7 ORH, UK [email protected] Gareth Davies, 4 Thorpe Mews, Yarmouth Road, Norwich, Norfolk NR7 ORH, UK [email protected] Rik Hoggett, Centre of East Anglian Studies, University of East Anglia, Norwich [email protected]

26

C. Hart, The Danelaw (Cambridge, 1992), tab. 2.2b.

NR4 7T/,

UK

SIMY FOLDS: TWENTY YEARS ON

By

DENIS COGGINSt

The settlement site of Simy Folds (NY 888277) lies at a height of 350 m ( II 50 ft) above sea-level on Holwick Fell in Upper Teesdale, County Durham (Fig. 1). Three farmsteads with similar plans occupy a limestone terrace some 100 m broad and 400 m in length though their associated fields extend over a much larger area (Fig. 2). Excavations on the site were begun in 1976 and continued in 1979 by the author and Dr Kenneth J. Fairless as part of a lo-year programme of archaeological investigation in Upper Teesdale. A third season of work was carried out in 1981 directed jointly by the author, Dr Fairless and Dr Coleen Batey. An account of the excavations was published in MedievaL ArchaeoLogy in 1983.1 The elements common to each of the three farmsteads are: (a) a long narrow rectangular building with rounded corners, aligned east to west and with an entrance in the east gable, (b) a smaller less regular building set at right angles to the first so that the two buildings form two sides of a small enclosure yard. Two of the farmsteads (nos. 2 and 3 below) have, in addition, a third building, small in size, placed only a short distance away from the other two. Farmstead 1, at the eastern end of the terrace, had only two buildings; their internal measurements were 9 m x 4 and 8 m x 4 respectively. Both these buildings were excavated. A strip of paving ran along the centre of the larger building giving on to a paved area common to both. The walling of both buildings was largely composed of whinstone boulders derived from the glacial drift which covers much of the area. Pieces of iron slag had also been incorporated in the walls of the smaller building. A hearth against the south wall of the larger building had been used for smithing as well as presumably for domestic purposes. A second hearth was found cut into the shale bedrock of the smaller structure. At the second farmstead the main building measured internally 13.5 m x 3.54 and once again there was a central strip of paving. The other two structures were connected by paved paths that may originally have been covered in. All these buildings were excavated and hearths were found in each. The third farmstead at the western end of the site was much the same as the second example but here the plan was complicated by the presence of an enclosure 1 D. Coggins, K. ] . Fairl ess and C. Batey, 'An Early Medieval settlement site in Upper Teesdale, Co. Durham', Medieval Archaeol., 27 (1983), 1-26.

326

LAND, SEA AND HOME

Location map of Simy Folds, Upper Teesdale.

FIGURE I.

composed of large whinstone boulders apparently overlying the farmstead. Surface inspection alone, however, could not ascertain whether this was actually the case or whether the farmstead had made use of a pre-existing enclosure. None of the three larger buildings was excavated but a fourth small, apparently circular, structure attached to the exterior of the south wall of the main enclosure was examined. Upon excavation, this proved in fact to be rectangular, measuring internally 6 m x 4 and with an entrance to the east. The excavations produced very few small finds - as usual on upland sites. Associated with the buildings were a stone spindle whorl, several hones, two fragments of a tufa disc quem, a much corroded iron ferrule and several indeterminate fragments of iron. There were also quantities of iron slag of various types. Remnants

SIMY FOLDS

FIGURE 2. I: Air photograph showing the three farmsteads and part of the fields systems at Simy Folds in Upper Teesdale. (Photo Blaise Vyner) 2: Plan of Simy Folds area.

32 7

328

LAND, SEA AND HOME

of earlier occupation of the area included flints, the butt of a stone axe, a fragment of a turned shale vessel and a few sherds of prehistoric pottery. These last were recovered from a post-hole found beneath the west gable-wall of the larger building of the eastern farmstead. While the sherds are small and abraded they resemble some of the pottery from the Bronze-age site of Bracken Rigg, also in Upper Teesdale. 2 Excavation of the junctions of two of the E.-W. field boundaries with the lesser ones running N.-S. produced no dating evidence though it was concluded that the long, contouring E.-W. features were primary. One of these at least can be traced for several kilometres and it can be suggested that these may well represent a process of land-division dating to the Middle Bronze Age. In addition, a peaty area immediately north-west of the site proper was examined by Mrs Alison Donaldson of Durham University and a pollen diagram constructed from the results. An early clearance phase, possibly neolithic, was recognized but not radiocarbon-dated. A second, more extensive phase was dated to cal. 1960-1570 B.C. (HAR 5070, 3400 ± 200 BP) while a third, permanent clearance seems to have begun about cal. 800-390 B.C. (HAR 3791, 2440± 160 BP). Mrs Donaldson considered that cereal pollen, present in the undated upper levels of this phase, was probably associated with the occupation of the three farmsteads. The deposition of cereal pollen had ended by cal. A.D. 1240-1400 (HAR 5069, 690± 120 BP).3 Charcoal from the excavated buildings was also sent for radiocarbon determina­ tion. Each of the two buildings in farmstead I had contained a hearth: charcoal from that in the larger building gave a date of cal. A.D. 660-1000 (HAR 4034, 1210± 160 BP) while that from the smaller building and lying directly on the bedrock was much earlier, cal. 770-170 B.C. (HAR 4035, 2330±200 BP). Two charcoal samples from farmstead 2 were also sent for determination, one from a hearth in building 2 and the other from a hearth in building 3. The results were respectively cal. A.D. 670-I020 (HAR 1898, II70± 140 BP) and cal. A.D. 1020-1320 (HAR 5348, 800±200 BP). At the third farmstead charcoal was found on the paved floor of the 'annexe' building attached to the enclosure wall. This was dated to cal. 1734-1721 and 1700-1260 B.C. (HAR 5347, 3200± IOO BP).4 The small heap of iron slag overlying the boundary wall of the settlement was excavated and charcoal sent for radiocarbon determination, resulting in a date of cal. A.D. 1020-1290 (HAR 4506, 820± 140 BP). Samples of slag were subjected to thermoluminescent examination but the results were inconclusive. The radiocarbon evidence indicates prehistoric occupation of the Simy Folds area during the Late Iron Age, the Middle Bronze Age and possibly even earlier in the Neolithic. Small finds support this scheme. The surviving house-foundations, however, are almost certainly Early-medieval in date. This occupation probably ended not long after the Norman Conquest or alternatively, even earlier, with the later dates showing a brief re-occupation of the site. No excavation has been carried out since 1981 though regular fieldwalking has taken place. With the exception of one or two flints no stray finds have been made. 2 D. Coggins and K. J. Fairless, 'The Bronze Age settlement site of Bracken Rigg in Upper Teesdale, Co. Durham', Durham Archaeol. ]., I (I984), 5- 22. 3 Two of these results (HAR 5070 and HAR 5069) were not available for inclusion in the original report. 4 This result was not available for inclusion in the original report.

SIMY FOLDS

32 9

However, it has become evident that the field systems, cairns and slag heaps present a much more complex picture that that presented by the settlement plan published in I983. A completely new survey is highly desirable. While it is considered that the original excavation report does not require any major revision, several aspects deserve fuller consideration. When the report was written only one other site similar to Simy Folds had been published. This was the settlement at Ribblehead in North Yorkshire excavated by Mr Alan King. Since then, other similar, if not identical, sites have been discovered. In Teesdale itself there are at least four more within a few miles of Simy Folds: Willy Brig Sike (NY 9I6254), White Earth (NY 906263-909260), Stonehouses (NY 907282) and Holwick Castles (NY 89927I). The first three of these occupy similar terraces of limestone grassland on Holwick Fell while the fourth, Holwick Castles, is situated in a sheltered nook of the Whin Sill cliffs but with access to better land. It may perhaps be considered as an outlying extension of Simy Folds. Stray finds from Willy Brig and White Earth suggest that they too had been previously occupied during the Bronze Age. Further afield, the sites of Dickey Edge in Swaledale, discovered and surveyed by Mr Tim Laurie, and Braida Garth in Kingsdale, discovered and surveyed by King, are settlements of a very similar type. The present writer has noted and discussed other sites in the Pennines, especially in the Malham area, and there is little doubt that more remain to be identified. s To return to Simy Folds itself, a feature of the site that requires explanation is the successive occupation of the area over some thousands of years. This is often the case on upland sites. It is usually ascribed to climatic change: amelioration of conditions leading to the expansion of settlement into the hills and deterioration resulting in a reversal of the situation, with subsequent return to lower altitudes. 6 This, in itself, seems to be an over-simplistic explanation. It may be that a general model for such a sequence of change cannot be formulated. As Roberts has said, ' ... landscape evidence must not be forced into the Procrustean bed of generalized explanatory sequences based on historical records .. .'.7 Each area must be considered individually. For Simy Folds an important geographical feature must be taken into account when trying to explain repeated occupation. The upper reaches of the River Tees and its tributary, the Maize Beck, provide a relatively short and easy route between Upper Teesdale and the valley of the River Eden in Cumbria. This is still regularly used as it forms a section of the Pennine Way footpath but there is every reason to suppose that it has been in use since the Mesolithic. The route follows the south bank of the Tees, well above the valley floor with no steep hills and no major river tributaries to be crossed. At any period, therefore, it would have been natural that settlements would be established along the route. Another consideration may have influenced settlers in Teesdale. There are two features in particular which early people are likely to have viewed with awe: the two 5 D. Coggins, 'Aspects of early medieval settlement in the North Pennine dales', Medieval Europe 1992. Pre­ printed Papers, 8: Rural Settlement (York, 1992),95-7. 6 T. Darvill, Ancient Monuments in the Countryside (London, 1987), 148-50. 7 B. K. Roberts, 'Sequent occupance, cataclysm and continuity: modelling settlement reality' (unpubl. paper presented at conference Medieval Europe 1992, York).

33 0

LAND, SEA AND HOME

waterfalls of High Force and Cauldron Snout, both on the River Tees. The idea that natural features as well as monuments could be the objects of religious veneration is now accepted by most archaeologists and it may well be that no distinction was made in the past between the religious and the secular landscape. s Is it not conceivable that the power of these two waterfalls which attracts visitors today also attracted worshippers in the past? From the Iron Age onwards there was another important reason for the existence of settlement at Simy Folds. It will be recalled that iron slag from bloomeries and smithies, as well as fragments of furnace lining, were found especially in the excavation of the eastern farmstead and that a heap of slag was noted over the north boundary wall. Other slag heaps occur in the vicinity. A string of bell pits, known as Ore Pit Holes, extends for about a kilometre along the fell-side to the south-west of the site, while to the west lies Ore Carr, an expanse of marshy ground. Within about 4 km of Simy Folds over 40 heaps of slag are known. None of these has been securely dated, but it is probable that they range from pre-Roman times to the High Medieval period. 9 Some, at least, must be connected with the settlement site at Simy Folds. Fieldwork carried out recently by Dr Tom Gledhill and Ms Ros Nichol has resulted in the discovery of more slag heaps and also of many associated charcoal-burning sites. As recently as I995 Peter Crew could write that, ' ... in some parts of Europe charcoal was made in small pits adjacent to the smelting sites but as yet there is no evidence of this from Britain ... '.10 Such 'small pits' are exactly what has been found in Upper Teesdale and they are characteristic of Early-medieval iron-working sites in southern Norway. Lars Erik Narmo has shown that in Norway the process was in the hands of self-sufficient peasant farmers until about A.D. 950, after which iron­ production became increasingly specialized and geared towards a market. l l Until more dates become available we cannot say whether the ironworking at Simy Folds falls into the former or the latter category. Dr Gledhill has prepared a programme of investigation into the ironworking and charcoal-burning sites which is being supported by the Society for Medieval Archaeology and others. Unfortunately, the outbreak of 'foot and mouth' disease during 200I delayed the start of this project, but during 2002 an area of 7 sq km of moorland and juniper scrub was surveyed. This resulted in the discovery of 376 charcoal pits of which 77 produced surface samples of charcoal which were identified as to species. Birch was by far the most common species, forming 77.I% of the total. As yet, no HC dates are availableY We must now turn to the question of the identity of the Early-medieval inhabitants of Simy Folds and admit that we do not yet know exactly who they were or where they came from. The problems of trying to reconcile archaeological evidence R. Bradley, An Archaeology of Natural Places (London, 2000). K. J. Fairless and D. Coggins, 'Excavations at the early settlement site of Forcegarth Pasture North', Trans. Architect. Archaeol. Soc. Durham Northumberland, NS 5 (1980), 31-8; idem, 'Excavations at the early settlement site of Forcegarth Pasture South', Durham Archaeol. J., 2 (1986),25-39. \0 P. Crew, 'Bloomery iron smelting slags and other residues', Historical Metallurgy Soc. Archaeological Data Sheet, 5 (1995) · 11 L. E. Narmo, 'Iron production and self-sufficient farmers in medieval South Norway', in P. and S. Crew (eds.), Early lronworking in Europe: Archaeology and Experiment (Blaenau Ffestiniog, 1997). 12 T. Gledhill, pers. comm. 8

9

SIMY FOLDS

33 1

with historical data are well known and need not be rehearsed here. There are two radiocarbon dates, cal. A.D. 660-1000 (Har 4034, 1210 ± 160 BP) and cal. A.D. 670-1020 (Har 1898, II70± 140 BP), from hearths in farmsteads I and 2 respectively, which indicate Early-medieval occupation. In each case the difference between the earliest and latest probable dates, 340 and 350 radiocarbon years, mean that it is not possible to associate this occupation with any specific historical event. It also means that it cannot be shown with certainly that the two buildings were occupied at the same time though this seems likely to have been the case. Strictly speaking, the settlers could have been British, English or Scandinavian with regard to ethnic origin; in respect of religion the choice is between pagan and Christian. Can the morphology of the buildings themselves provide a clue to the people who built them? Each of the three farmsteads has one long narrow building with wide wall foundations of boulders and turf. Two of these buildings were excavated and each was found to have a narrow strip of paving running along its length from the entrance in the east gable. That at site I also had stone benches along the south and west walls and probably the north wall as well, though this had been partly destroyed (Fig. 3). Similar benches seem to have existed in the building at site 2, made not of stone but turf and probably timber. The features described above are characteristic of houses of Norse colonies in Scotland, Faroe and elsewhere.13 On the other hand, the arrangement of two buildings, placed at right angles so as to form two sides of a yard as at Simy Folds seems too occur in the Pennines but not so far as I know elsewhere (Fig. 4). In the absence of any diagnostic artefacts, whether English or Scandinavian, the evidence from site morphology must be regarded as inconclusive. The traditional date for the beginning of Scandinavian settlement in northern England is A.D. 876 when the 'Great Army' dispersed and shared out some of the land of Northumbria. The earliest settlement of Iceland is said to have begun at about the same time. The Scandinavian settlements in Cumbria are thought to have begun about A.D. 920, though the extent of these and the origins of the settlers is the subject of much debate, especially among place-name scholars. Such dates, however, are subject to considerable doubt. Perhaps more to the point is the suggestion by Professor Christopher Morris that the traditional dates for Norse settlement overseas derived from written sources represent ' ... a stage of formalization following a period of informal settlement ... '. The latter would of course have been unrecorded.14 Despite all the uncertainties there seems little doubt that Scandinavian settlement in Cumbria and the Eden valley in particular was substantial. This has relevance with regard to settlement in Teesdale. It has been pointed out above that close connexions between the Upper Tees Valley and the valley of the Upper Eden have probably existed for thousands of years, and it is not improbable that settlers at Simy Folds had arrived there from across the Pennines. One of the few finds from the site adds support to this assertion: the stone spindle whorl found in the main building of farmstead 2. This has recently been thin-sectioned and shown to be very similar to the red sandstone rocks

J. Graham-Campbell et al. (eds.), Cultural Atlas of the Viking World (Oxford, 1994). 14 C. D. Morris, 'Comments on the early settlement of Iceland', Norwegian Archaeol. Rev., 24 (1991), 18-20.

13

33 2

LAND, SEA AND HOME

FIGURE 3. Simy Folds, site 1. The main building looking east. Note the massive walls, central paving and benches in the interior along the south and west walls.

of the Penrith area of Cumbria strongly suggesting that it was brought from that area.15 Writing of Upper Swaledale to the south and considering its dialect, place-names, archaeology and farming practice, Victoria Airey concluded that the area could be described as 'a Norse ecological niche' .16 The same description for similar reasons could be applied with even more justification to Upper Teesdale. Place-name evidence 15 The spindle whorl was examined in December 2001 by Dr T. J. Morse of the Department of Geological Sciences, University of Durham. A summary of his report is given as an appendix to this paper. The full geological report has been deposited with the site archive in the Bowes Museum, Barnard Castle, Co. Durham. 16 Victoria Airey Skans, 'Backcans and Burdens: Upper Swaledale, North Yorkshire. A Norse ecological niche?' (unpubl. thesis, University of Stockholm, 1989).

SIMY FOLDS

FIGURE 4. I: Simy Folds, site I, looking north, with excavation completed. The building in the foreground is that shown in Figure 3 with the paving removed. 2: Final excavation plan of site I.

333

334

LAND, SEA AND HOME

of Scandinavian settlement in the dale is provided by names such as Ettersgill, Crossthwaite, Baldersdale and Briscoe. Many dialect words also are of Norse origin and two may be cited as examples: 'stee' for ladder and 'steg' for gander. More substantial evidence is provided by the Anglo-Scandinavian stone sculpture found at Gainford, Wycliffe and Cotherstone. In the absence of proof to the contrary, it is not unreasonable to believe that Simy Folds and similar sites in Teesdale mark the homes of Norse farmers who arrived in the dale probably from the Eden Valley. In Teesdale they lived a basic and self-sufficient existence, growing a little corn, pasturing their cattle and producing enough iron for their own use. Wherever the truth may lie, we may be confident that further work at and around Simy Folds, including the use of techniques and scientific instruments not available in the 1970s, will provide unrivalled information on upland settlement, particularly with regard to the Early Medieval Period. APPENDIX: SUMMARY OF GEOLOGICAL REPORT ON THE SPINDLE WHORL

Initial examination of the spindle whorl confirmed that it was made of a fine grained red sandstone containing flakes of mica and that it was probably derived from a flaggy, laminar horizon. A section was cut avoiding the central hole and the off-cut was impregnated with resin to bind the grains of sandstone together. Under microscopic examination (magnification x80) haematite coating of the grains was evident. Over 95 % of the grains were of quartz with less than 5 % of mica. The quartz grains were fine (less than 0.2 mm), well sorted and predominantly angular, showing that they had not travelled far from their source. The nearest similar sandstones are to be found in the red Permian and Triassic lithologies of the Eden Valley. Thinly bedded shales and sandstones are exposed in Hilton Beck (NY 72432°43), Dobby Hole Gill (NY 75601875) and the River Belah (NY 80181222). Though it cannot be finally proved that the spindle whorl originated from this area there is no reason why it could not ha ve done SO.17 ACKNOWLEDGEMENTS

Dr T. Gledhill kindly drew my attention to the papers by Peter Crew and Lars Eric Narmo regarding bloomeries and charcoal burning. Thanks are due to Dr T. ]. Morse for his geological report on the Simy Folds spindle whorl. Finally, I am fortunate in having been able to discuss this paper with my colleague Dr K.]. Fairless and I am, as always, grateful for his expert advice.

\7

75 ·

I. C. Burgess and D. W. Holliday, Geology of the Country around Brough under Stainmore (London, I979),

POST -ROMAN UPLAND ARCHITECTURE

IN THE CRAVEN DALES AND THE

DATING EVIDENCE

By

ALAN KING

In upland Britain the field evidence of Early-medieval settlement survives today largely as free-standing monuments, yet our understanding of this resource is minimal. The three Viking-period sites which feature on distribution maps of northern England, Simy Folds (Co. Durham), Bryant's Gill (Cumbria) and Gauber High Pasture (N. Yorks.),! were each excavated as part of long-term multi-period research projects directed by locally based archaeological enthusiasts using their own resources, and supported by a network of volunteers backed by specialist help from universities and museums. These sites are respectively 366 m (1200 ft), 290 m (950 ft) and 335 m (IlOO ft) AOD, on marginal land. At the Eighth Viking Conference in Arhus, Christopher Morris echoed the general consensus admitting our failure to uncover evidence of Scandinavian cultural influence at either Simy Folds or Ribblehead, but he was satisfied that these sites were 'the nearest we will get to seeing Viking period farmsteads in northern England'. Morris concluded that so little was known of the post-Roman building tradition in upland Britain, that identifying subsequent innovations was made more difficult. 2 This was clearly an understatement. On the limestone terraces and valley sides of the area drained by the headwaters of the Rivers Wharfe, Aire - locally Malhamdale - and the Ribble, early Roman­ period Brigantian farms and hamlets comprise stone built round houses and ancillary buildings (Fig. raj. Post-holes are impossible to identify among the buried clints and grykes of the carboniferous limestone and only one round, timber-built palisade trench for a house has been excavated, at Horse Close, Skipton on a gritstone hilltop.3 Elsewhere the widespread drystone round houses exhibit double-faced stone walls, 1 D. Coggins, K. ]. Fairless and C. Batey, 'Simy Folds, an early medieval settlement site in Upper Teesdale', Medieval Archaeol., 27 (1983 ), 1-27; D. Coggins, this vol.; S. Dickinson, 'Bryant's Gill, Kentmere: another "Viking-period" Ribblehead?', 83-5 in ]. R. Baldwin and I. D. White (eds.), The Scandinavians in Cumbria (Edinburgh, 1985); A. King, 'Gauber High Pasture, Ribblehead - an interim report', 21-5 in R. A. Hall (ed.), Viking Age York and the North (CBA Res. Rep., 27, London, 1978). 2 C. D. Morris, 'Viking and native in northern England: a case-study', Proceedings of the Eighth Viking Congress (Odense, 1981),223-44' 3 The evidence came from three seasons of a joint University of Leeds and Yorkshire Archaeological Society summer training excavation (1966- 8) under the direction initially of P. Mayes and later of Alan Aberg, who in 1968 was kind enough to supply the detailed information.

33 6

LAND, SEA AND HOME

FIGURE I.

a: Craven and North Lancashire; b: Malham Moor. (Drawn by the author)

ARCHITECTURE IN THE CRAVEN DALES

337

which, from the evidence of one of the least accessible and unrobbed sites, reach above 2 m in height; alternatively, on some sites, sockets for roof-support timbers are found at intervals around building walls. By the late 2nd century A.D., rectangular houses had been assimilated into the vernacular building repertoire though circular forms continue on the same sites, possibly housing both stock and stores. Individual circular houses range from 5 to 10 m in internal diameter. Their doorways invariably face south-east, and from the excavated ceramic evidence their useful lives extended over a hundred years. The occupants of one such farmstead at Helwith Bridge (SO 817693) burned coal on their open hearth. The rectangular houses have their doorways at one end of a long wall and not in either gable end. 4 Only on Lea Green, Grassington, at the northern end of a hundred hectare spread of Celtic fields, are there any buildings - in this case rectangular - which have yielded both Roman and post-Roman artefacts, thereby giving at least a terminus ad quem for the field-system. The writer's excavation of a rectangular building on Smearsett (SO 806677) produced no datable evidence, but the associated long, un­ lynchetted strip-fields did produce 4th-century pottery. A late 8th- or 9th-century strap-end, inlaid with niello in the Trewhiddle style, was an unrelated surface-find picked up around 1910. It became part of the Revd Crofton's collection of antiquities previously housed in Settle and is now in the British Museum (Fig. 5d).5 Dr Arthur Raistrick researched many facets of dales geology, glaciation, prehistory, and history for well over 50 years. As early as 1924 he undertook field­ survey aided by aerial photography and followed up that work with palaeobotanical analysis on Linton mires in the late 1930S. Annually through the 1950S he directed training excavations on Malham Moor. One summer's work focused on the massively solid, though small, rectangular building in a sheltered hollow 300 m north of Malham Tarn, at an altitude of 455 m AOO which he called the Priest's House (Figs. 1b and 4a).6 The inner wall faces of the stone-packed and -faced wall of this building are better finished than the outer face, where large limestone blocks have been far less easily handled. The 1.5 m-thick walls enclose a floor space 4.35 m x 2.5 that is subdivided, by a 0.6 m-thick drystone wall, into two rooms, of which the outer room contained the hearth and the smaller inner room features a kerb that may have edged a short bed. Finds included a copper-alloy annular brooch 20 mm in diameter, a 14 mm-wide tinned copper-alloy buckle, pieces of copper-alloy strip interpreted as book edgings, a bookmarker tag, and a cast, circular disc ornamented with pierced interlace (Fig. 5a)? Its main pattern elements are described as having been gilded. It has a bossed roundel central to four quadrants of pierced interlace, all encircled by a raised rim; two opposing holes occur close to its edge with a segment torn away or badly corroded. Previously published as a damaged pin, this is probably the middle element of a three-pin suite similar to the River Witham set though the decoration of this 4 S

6

These results reflect on-going research by the writer. British Museum 1968,4-2, 1. A. Raistrick, 'The archaeology of Malham Moor', Field Studies,

1~- 1 6.

Raistrick, op. cit. in note 6, 89-90.

I,

NO.4 (1962),72- 100, at p. 90, figs.

33 8

LAND, SEA AND HOME

northern find has more Northumbrian and Irish parallels. Both finds and building should now be dated to the late 8th or early 9th century. Considering that so few Early-medieval buildings have been identified in the northern hills, let alone on Malham Moor, it is remarkable that another example was excavated by Raistrick in his Malham series on Prior Rakes barely 2 km south-east of the Priest's House (SD 9°4648). Excavation of the Bolton Priory Sheep house was undertaken in 1956 following the discovery of archival details of a sheep's milk cheese ptoduction unit sited on 'Prior Rakes'. At some stage the trainees extended their work westwards 30 m, ignoring the remains of a number of substantial buildings, to a smaller rectangular building situated on a limestone terrace 2 m above the level of the Sheephouse and its enclosures (Fig. 4b). By a series of chances the writer discovered a large unsealed plastic bag containing two Northumbrian styccas of Eanred while teaching at Malham Tarn Field Studies Centre in the summer of 2000. These are the total remaining finds of those made in the 1950S excavations. Subsequently a building plan showing a large totary quernstone and a 'Saxon coin' on the building floor was recognized in the Raistrick Collection housed in Bradford University Library, the massive limestone orthostats of the inner wall face helping identification. Days later, in the Ironbridge Gorge Museum Trust's Raistrick Archive, a note identifying one of these coins as a stycca of Eanred by the moneyer Monne from the small building on Prior Rakes, drew together the evidence for another 9th-century building. The other coin had been struck by Aldates. The valley's mother church at Kirkby Malham has always suggested a Christian element in this local community during the period under discussion, but the church and churchyard are devoid of any Early-medieval evidence. Unfortunately a portion of decorated cross-shaft discovered at Long Preston, in neighbouring Ribblesdale, by Tot Lord in 1950 was taken out of the district unrecorded, other than gaining a mention in a Raistrick letter to the finder dated 17 May 1950.8 The final site to be considered is the Ribblehead farmstead, situated on the topmost, denuded white limestone terrace overlooking hectares of Romano-British pastoral enclosures, the trans-Pennine Roman road from Lancaster to Wensley, and an unquarried settlement which an earlier air photographs shows to have once been much more extensive, featuring both round and rectangular stone-based building forms (Fig. 2a-b).9 • A portion of the letter to 'My Dear Tot' on University of Durham notepaper, but with Raistrick's home address typed beneath reads (paragraph 3): I have looked up the Anglo-Danish crosses and there is no doubt whatever that the fragment of a shaft which you have found is hitherto unknown, and it will rank as a find of great importance. It would be erected in a churchyard either to mark a grave or a preaching cross, and establishes (Long) Preston as an ancient and pre­ Conquest church. As far as I remember the ornament it appears to belong to class B or possibly AB of Collingwood and so would date about 900-950 AD ... I will get over to Settle some Monday soon, and will draw the cross, the Long Preston implements etc and will get them published almost at once so that you can have a stock of reprints for the museum. 9 Gauber High Pasture is the name of the field in which the excavated farmstead is situated. However, over a period of time it has now become known as Ribblehead which is the locality description and also the name of the railway station 600 m north of the site; University of Cambridge air photo W F 25 (Ingleton parish 90/ 765787) shows the quarry expanding into a Romano-British hamlet.

ARCHI T ECTUR E IN THE CRAVEN DALES

FIGURE 2.

a: Ribblehead; b: air photograph, Cauber High Pasture. (Both by the author)

339

340

LAND, SEA AND HOME

Two of the farmstead buildings, the workshop and kitchen, respectively measure internally 8.2 m x 3.7 and 6.1 m x 3.7, the latter narrowing to 3.35 m at its western end. Each has its doorway at the end of a long wall with the kitchen door opening into a paved, double-walled, 4 m long passage running directly towards the northern doorway of the farmhouse (Fig. 3). The farmhouse has its long axis orientated just a few degrees off N.-S., with a floor area measuring more than 19 m x 4. The asymmetrically sectioned walls have coursed stones on their inner faces but single small boulders, 300-400 mm in diameter, mark their periphery. They have an earth and rubble core. Though externally larger than the timber-built Saxon palace at Cheddar, this house has few parallels in pre-Conquest northern Britain, and then only with distant buildings such as those at Jarlshof on Shetland. to Unlike the other secular rectangular buildings in the study area, an innovative timber-frame structure seems to have been utilized here, with the roof supported by posts positioned on stone pads along the inner wall-face. This produced an uninterrupted floor- space which unfortunately had suffered 200-300 mm of denudation before excavation, which precluded chemical analysis of the soil. A timber partition cross-wall divided the interior roughly into half, with a centrally placed hearth warming the southern part, although the only stone bench or bed kerb was found against the side wall in the northern end. Either doorway may still be entered between the arcing masonry of the gable ends, the southern example being a particularly solid 3 m thick with an opening of only 650 mm. An aceramic collection of small finds resulted from three seasons of excavation. These included angle-backed iron knives, an iron spearhead 400 mm long and a tinned copper-alloy clapper bell. Both upper (450 mm diameter) and lower (460 mm diameter) millstone grit quernstones were found while deturfing the kitchen. As at Prior Rakes, we are dependent on coins for our dating: four were found, one of which was totally abraded. Only one of the remaining three styccas was an authorized issue: this, of Archbishop Wulfhere minted by Wulfred at a date close to 850. One of the remaining two irregular coins attempt to give the name Odilo, issued for Aethelred II; Miss Pirie dates this coin to within five years either side of the Wulfhere coin. The last coin is a nonsense piece though probably contemporary with the others, so that all are datable pre-855 Y In this region, then, we have a house which undertakes in stone and timber what elsewhere, i.e. in lowland England, was an earthfast timber form. This farmhouse and its associated traditional ancillary buildings opened into a shared farmyard, but there is no evidence whatsoever for the housing of stock. Analysis of the animal remains suggests that any cattle over three years old were either elsewhere, or at least were not slaughtered or eaten at this farm. Today many farms at Ribblehead overwinter their cattle and sheep on the fells: these form a hardy breeding stock sought after by lowland and upland farmers alike and not in milk over the winter. Thus while we 10 P. Rahtz, 'The Saxon and medieval palaces at Cheddar, Somerset an Interim Report of Excavations inI96o- 62', Medieval Archaeol., 6-7 (1962-3), 53-66. 11 E. J. E. Pirie, Coins of the Kingdom of North umbria, c. 700- 867 (L1anfyllin , 1996).

ARCHITECTURE IN THE CRAVEN DAL E S

FIGUR E

3.

Cauber High Pasture. View from the kitchen to the northern gable-end of the farmhouse. (© The author)

HI

342

LAND, SEA AND HOME

FIGURE 4.

a: Priest's House, Malham Moor; b: Prior Rakes house, Malham Moor. (After

A. Raistrick)

ARCHITECTURE IN THE CRAVEN DAL E S

343

5. a: roundel, Priest's House; b: iron weight, Clapham; c: barrel padlock key, Clapham; d: strap-end, Smearsett. Scale approx. r :1. (Drawn by C. M . King, (b) courtesy of the British Museum)

FIGURE

have a considerable amount of information from the Ribblehead site, it remains an anomaly. Was this an extended family farm with outbarns, or was it of greater significance locally? Its position so close to the line of the still well-defined Roman road suggests it may have been a trans-Pennine staging post linking Catterick and the west coast. Though our understanding of Viking-period vernacular architecture in upland Britain badly needs detail, almost all the 'archaeological digging' in the Craven area and across the 3S km westwards to Morecambe Bay is being undertaken by metal detectorists. One man working between Clapham and Austwick (an area centring on SO 7S4688) has unearthed a cheese-shaped copper alloy-coated iron weight of 98.S6g of Viking-age type. Both the shape and the punched-dot decoration encircling the upper surface are visually similar to two weights from the Coppergate excavations in York, although the central infill differs, the basic design element being created using punched semicircular lines. Nearby but not associated with this was a later, rather anthropomorphic, copper-alloy barrel padlock key 86 mm in length (Fig. Sb-c).12 12 A. J. Mainma n and N. S. H . Rogers, Finds from Anglo-Scandinavian York (The Small Finds, 171r4, York, 2000). The two comparable finds are the weights numbered 10356 and 10357 featured on pp. 2562-3.

344

LAND, SEA AND HOME

Further west, at Carnforth (Lanes.), a small loth-century hoard of Viking hacksilver and Arab dirhams has been discovered and subsequently acquired by Lancaster City Museum,I3 while from Arnside on the southern side of the Kent estuary we have an 8th-century copper-alloy Irish bowl-mount featuring a stylized human face above a square panel exhibiting some evidence for red and yellow enamelling in straight-sided geometric champleve fields. 14 One of the better parallels for this Viking booty comes from a 9th-century male Viking grave at Myklebostad, Eid, Sogn og Fjordane, Norway.15 Ribblehead is 9 km from the village of Ingleton - also the parish name - and a further 4 km from the hamlet of Ireby. At some date the English and Irish were minor elements in the population. A synthesis of place-name elements, sculptured monu­ ments and Early-medieval architecture would be of considerable help to future scholarship. The Ingleborough massif and the western portion of Yorkshire have to be considered as part of north-western England when this county boundary, at its closest, is less than 20 km from the sea. We need to follow up the information which is now becoming available from the field-walking detectorists, and in particular to provenance their finds more accurately and to amplify their work though detailed ground and resistivity surveys. ACKNOWLEDGEMENTS

I am extremely grateful to all who helped launch my understanding of post-Roman archaeology for their continued support and friendship. Special thanks are due to Mrs Leslie Webster for drawing my attention to the British Museum's acquisition of the Smearsett strap­ end and to Elizabeth Pirie who has identified the coins and continued to update my knowledge. Mr John Brassie very kindly sent me the iron weight and barrel padlock key he had found for recording, while Patrick Ottaway drew my attention to the York parallels. However, without my wife's continued art-work and general support, especially over drystone walls(!), there would have been far less progress.

Alan King, The Tannery, Settle, Yorkshire [email protected]

BDZ49HL,

UK

13 Following a Coroner's Inquest in February 1999 the hoard was found to be treasure and bought by Lancaster City Museum. Accession number LM99.47. Andrew White, pers. comm. 14 S. Denison, 'Detectorists report thousands of new finds to archaeologists', British Archaeol., 62 (December

2001),4· 15

s. Youngs (ed.), The Work of Angels (London, 1989), 61, Exhibit 51.

TIMBER BUILDINGS WITHOUT

EARTH-FAST FOOTINGS IN VIKING-AGE BRITAIN

By

MARK GARDINER

For an earlier generation of archaeologists the pattern of development of timber buildings was clear. Before the 13th century buildings were constructed with their posts set in the ground - in individual post-holes or trenches. The posts tended to rot and therefore the buildings were crudely constructed, and perhaps only expected to last a generation or so. But, in the late 12th century, and particularly during the course of the 13th century, a major change in building construction occurred. Timber structures were increasingly constructed on sill walls. The timbers, removed from contact with the ground, could be kept drier and therefore lasted longer. More permanent buildings could then be constructed and it was worth building them to a higher standard. 1 The still-standing vernacular timber-framed buildings of the 14th century, or more typically from the 15th or 16th centuries, are examples of these houses built in a more permanent manner. That interpretation has been challenged in the last IO or 15 years. It is no longer clear that there was a sharp division between the crude construction of 'impermanent' earth-fast structures and the thicker timbers of 'permanent' houses built on sill walls. Wrathmell has identified a number of 13th-century buildings which initially were constructed with earth-fast timbers.2 The timbers were later cut off or rotted at ground level and were then under-built in stone. Such buildings have demonstrated that, when built, though they derived some of their structural stability from the post­ holes, they had an essential rigidity which was not dependent upon ground-fast construction. Carpenters in the 13th century were increasingly confident that they could construct stable buildings without setting the timbers in the ground. The change of the 13th century was evolutionary rather than revolutionary. The second reason for questioning the earlier interpretation of building development is the growing evidence of a long period of development which preceded the changes in the 13th century. Various methods were used from c. 800 onwards to 1 That view is reRected in many excavation reports, but is most cogently stated in J. G. Hurst, 'A review of archaeological research (to 1968)', 93-5 in M . W. Beresford and J. G . Hurst (eds.), Deserted Medieval Villages: Studies (London, 1971). 2 S. Wrathmell, Domestic Settlement 2: Medieval Peasant Farmsteads. Wharram: A Study of Settlement on the Yorkshire Wolds, VI (York, 1989), 12; idem, 'Medieval rural settlements', 178-94 in B. Vyner (ed.), Building on the Past: Papers Celebrating I';O Years of the Royal Archaeological Institute (London, 1994).

34 6

LAND, SEA AND HOME

try to increase the longevity of the timbers, either raising by them above the ground, or by protecting them in other ways from moisture. It is useful at the outset to define the terms used here. Timber buildings without earth-fast posts which did not rely for their structural stability upon posts buried in the ground are described as 'self-supporting,.3 Their posts were ground-set, that is placed on the ground surface, or on a low stone footings. Few of the buildings discussed here are likely to have been framed: timber-framing was a development of the period after c. lI80 when buildings were constructed using pinned mortise-and­ tenon joints. 4 Richard Harris has described such buildings with a graphic image: 'a giant could pick up a timber-framed house, turn it upside down and put it back on the ground still in one piece,.5 Most of the buildings discussed here were dependent to some degree on gravity to stay together and would have fallen apart if picked up by Harris's hypothetical giant. The present paper draws together the growing evidence for buildings with timbers set at ground-level, not in the ground. Two different types are identified. The first of these are buildings with stone or gravel footings which date from around 800 and are only found on high-status sites. Such buildings were first recognized by Rosemary Cramp and further examples have since been added to the small corpus. 6 The second group comprise mainly urban buildings constructed from the loth century onwards. Both these groups were truly constructed without any earth-fast timbers. One further category is also considered and these are an intermediate form of building in which the exterior walls were sometimes placed at ground-level, and earth-fast posts were used internally and therefore kept dry. HIGH-STATUS SITES AROUND 800

Whithorn Excavations on the ecclesiastical site of Whithorn provided evidence for a variety of construction techniques, including the possible use of cob and wicker. 7 Most of the buildings had earth-fast footings, but the problem of that method of construction is evident from the decay of the church. By about 760 the timbers of the church were 30 years old and had begun to rot at the base. Rather than rebuild the structure, trenches were dug from the exterior and stones slid in to support the timbers (Fig. la). The walls were then pulled outwards on to the stone sill, producing a structure which was slightly splayed. These measures allowed the church to continue in use for another 70 or 80 years until it was destroyed by fire. The works on the walls of the church cannot be separated from the construction of the open arcades built parallel to the long walls 3 The term 'self-supporting building' is used by J. T. Smith, 'Norwegian stave churches: a review of some recent publications', J. Brit. Archaeol. Ass., 131 (1978), II8-25. 4 G. Milne, Timber Building Techniques in London c. 900-I400 (London Middlesex Archaeol. Soc. Special Pap., IS, London, 1992),99-105, 5 R. Harris, Discovering Timber-Framed Buildings (Princes Risborough, 1978), 5. 6 R. Cramp, 'A reconsideration of the monastic site of Whitby', 64-73 in R. M. Spearman and J. Higgitt (eds.), The Age of Migrating Ideas: Early Medieval Art in Northern Britain and Ireland (Edinburgh, 1993). 7 P. Hill, Whithorn and St Ninian: The Excavation of a Monastic Town, I984-91 (Stroud, 1997), 139 and 16 5- 6.

TIMBER BUILDINGS

FIGURE 1.

347

a: Whithom, the Phase 4 church; b: Whithom hall 9b (after Hill, op. cit. in note 7). Post-pipes are indicated in solid black.

of the church. These may not have been intended to have a structural function when built, as the excavators have argued, but, as the wall timbers rotted, they will have provided some structural support. They will certainly have protected the walls of the church from rain and reduced the decay of the timbers. The church was also supported internally by an added series of timbers indicated by slight depressions. These cannot have been cruck forks as the published report suggests: it would have been impossible to rear the crucks within a standing building and there is no evidence for the use of that form of construction at such an early date. s After a fire in c. 845, the church was rebuilt with broad footings of small stones. These survived up to seven courses high, but the excavator did not consider the walls to have been built of 8 J. L. Walker, 'Late-twelfth and early-thirteenth-century aisled buildings: a comparison', Vernacular Architect., 30 (1999), 21-53.

34 8

LAND, SEA AND HOME

masonry to the height of the eaves: the surviving parts of the walls were not sufficiently substantial to support the weight of a full-height stone structure. They probably served instead as a sill wall for a timber superstructure. 9 A similar sequence of building construction and repair has been identified in Hall 9. It too begun in the later 8th century as an earth-fast structure, but when it was repaired in the early 9th century some of the posts, particularly those around the doorway, were under-built with a stone plinth (Fig. rb). The work of under-building in this instance, unlike the church, was not thorough and only some of the posts appear to have been cut away at ground level.lO

H artlepool and Dunbar The buildings without earth-fast footings at another contemporary site, Hartle­ pool, can be divided into three variants. l l The first are buildings which were originally constructed with posts set into postholes, but were later under-built in stone. The posts, like those of the church at Whithorn, had evidently begun to rot near ground­ level. They were cut off and the postholes and post-trenches were capped with gravel or with limestone slabs. Three buildings (VI, XI and XX) belong to this first variant. A fourth building (VIII) was also under-built, and the opportunity was taken to extend it to the north. The extension had a stone sill wall which measured 0.4 m wide (Fig. 2a-b). The second variant was those structures built with stone and gravel from the outset. They are likely to have been similar to the under-built structures and had footings 0.45 m wide or less. Two buildings (XXVII and XXVIII) certainly belong to this category and structure XXIX, of which only part was recorded, may also be of the same type (Fig. 2C). The final variant at Hartlepool is represented by a single building, VII, which had broader footings, at least 0.7 m wide, within which were the impressions of a number of posts (Fig. 2d). It is uncertain whether the building had a similar structure to those of the second variant - a series of posts set on gravel or stone footings - or whether it was built of stone with timber reinforcement. Any interpretation must also take into account a similar 9th-century structure at Dunbar in ScotlandY Dunbar Building 6 had broad footings into which socketed stones were set to hold timber posts (Fig. 2e). The use of stone sockets is exceptional and might suggest an attempt to translate the familiar use of packed postholes into the medium of stone. The footings, comprising two outer lines of larger stones and an inner core of rubble, survived to their full width, about I m wide, on the east side. They may have been a similar width on the other sides where apparently only part has been preserved. Footings of that width suggest that they supported masonry walls, and indeed mortar was noted during excavation, although it was dismissed by the excavator as possibly intrusive. 9

10 11

Ibid., 186-8. Ibid., 177- 9. R. Daniels, 'The Anglo-Saxon monastery at Church Close, Hartlepool, Cleveland', Archaeo/.

1J8- 210.

J., 145 (1988),

2 D. R. Perry, Castle Park, Dunbar: Two Thousand Years on a Fortified Headland (Soc. Antiq. Scotland Mon., 16, Edinburgh, 2000), 64- 77.

TIMBER BUILDINGS

349

FIGURE 2_ a: Hartlepool Building VIII; b: Hartlepool Building VI; c: Hartlepool Building

XXVII; d: Hartlepool Building VII (after Daniels, op_ cit_in note II); e: Dunbar Building 6; f:

Dunbar Building 9; g: Dunbar Building 7 (after Perry, op_ cit. in note 12).

35 0

LAND, SEA AND HOME

It is more satisfactory to interpret Hartlepool VII and Dunbar 6 as masonry buildings with timber reinforcement. Timber reinforcement is unusual, particularly at this period. 13 The hybrid method of construction might suggest that the masons were unfamiliar with techniques of stone building and retained the timbers upon which they knew they could rely. The structural properties of stone were often under­ utilized even in later masonry buildings and craftsmen continued to employ it and work it as if it was timber.14 Other buildings at Dunbar suggest a similar ambivalence about masonry construction. The south side of Building 9 had footings formed of two outer faces of larger blocks and a core of smaller stones. It was about I m in width and therefore seems to have supported a masonry superstructure. Footings for the two other surviving walls were 0.3 m wide and formed from only a single row (Fig. 2f). The difference cannot be attributed to later stone-robbing. The narrow wall on the north side was delimited on the outside by an adjoining drip-gulley and it is unlikely that it was any wider on the inside if the band of paving within the building was centrally placed. It seems probable that the building comprised one masonry wall and at least two timber beams set on stone sills. Building 7 appears to have been similar. It also had a broad wall on the south and narrower footings on the other three sides (Fig. 2g). The use of timber and stone for adjoining walls in Buildings 7 and 9, and the use of timber reinforcement in Building 6 at Dunbar and Hartlepool VII suggests an experimental approach to building construction. The builders were attempting to use materials in new ways. It is not clear whether they were using the stone to provide rigidity and support timber elements, or whether they felt uncertain about the stability of the stones walls and were unwilling to commit themselves to the extensive usage of an unfamiliar material.

Other sites The evidence of buildings from other contemporary sites is less informative, but excavations at Flixborough, Whitby and Sheppey have all uncovered structures which are comparable to those already discussed. Building I at Flixborough, constructed in the early or middle 8th century, had stone and gravel footings. It has been tentatively identified as a church or mortuary chapel on the evidence of four burials inside the building and two outside, but the presence of a hearth and the occupation debris found on the floor surface suggests that later it may have been changed to a domestic use. The published details do not allow the nature of the building to be firmly established. It seems to have had stone footings for a ground-level sill beam which supported the walls, and exterior earth-fast posts for the structural timbers.1s

R. P. Wilcox, Timber and Iron Reinforcement in Early Buildings (London, 1981). W. R. Rodwell, 'Anglo-Saxon church building: aspects of design and construction', 156- 75 in L. A. S. Butler and R. K. Morris (eds.), The Anglo-Saxon Church: Papers on History, Architecture, and Archaeology in Honour of Dr H. M. Taylor (CBA Res. Rep. 60, London, 1986). 15 C. Loveluck, 'A high-status Anglo-Saxon settlement at Flixborough, Lincolnshire', Antiquity, 72 (1998), 146- 6 1. \3

14

TIMBER BUILDINGS

351

Rosemary Cramp and Philip Rahtz have reinterpreted the excavations of the 1920S at Whitby Abbey in the light of more recent discoveries. The original report had suggested that the clay-packed stone footings of the Anglo-Saxon buildings supported a stone structure. Rahtz has noted both the absence of stone and mortar, and the presence of numerous fragments of daub. He has suggested that the footings are more likely to have supported buildings with a timber structure and daub-and­ wattle infill. Doubt remains about even this interpretation, because more recent excavation has suggested that there was considerable redeposition of material. 16 The last building in this group was discovered during excavations in 1999 at Minster in Sheppey (Kent). Earlier work by the local archaeological society in a garden 65 m to the north-east of the church had identified a ditch containing Ipswich Ware. The later excavations showed that the ditch terminated at a contemporary earth-fast building within which was a coin of Offa of the period 792-6.17 The stratigraphy of the building had been partially truncated by horticultural work, but narrow stone footings were recorded on the south side where a greater depth remained. The footings were only 0.35 m wide and a single course deep, but probably supported a timber building. The footings could not be directly dated, but the coincidence of position and alignment with the earlier earth-fast structure suggests that they may have directly succeeded it, perhaps in the early 9th century (Fig. p). It is notable that these buildings all occur on high-status sites of the later 8th and earlier 9th centuries. Some of these have been convincingly identified as monastic ­ Whithom, Hartlepool, Whitby and Minster in Sheppey. The evidence is less clear for the sites of Dunbar and Flixborough. The former may have been a royal centre, while the latter is considered by Loveluck to have been a high-status secular site during the period of Building 1. 18 At all of these sites the builders were apparently trying out new methods of construction to provide greater longevity, either for existing buildings or for new ones. It is possible to interpret the desire for more permanent buildings as a wish to avoid the costs of reconstruction, but that would be too reductive. The occupants of high-status sites are unlikely to have been primarily motivated by the matter of the cost of materials and labour. A more plausible explanation is the desire to build structures which through their longevity projected an image of the permanence of sacred or secular power. However, the degree to which builders had truly mastered the methods of timber building to allow them to erect enduring structures remains unclear. URBAN SITES

The second group of self-supporting timber buildings are distinguished by their location on urban sites and by their loth-century or later date. The majority of timber buildings continued to be constructed with earth-fast posts and it remains unclear why a small number were constructed in other ways. Evidently, the advantages of 16 P. Rahtz, 'The building plan of the Anglo-Saxon monastery at Whitby Abbey', 459-62 in D. M. Wilson (ed.), The Archaeology of Anglo-Saxon England (Cambridge, 1976); Cramp, op. cit. in note 6, 65-6. 17 Unpublished excavations by Queen's University Belfast. IS Loveluck, op. cit. in note 15, II5-17.

35 2

LAND, SEA AND HOME

FIGURE 3. a: Minster-in-Sheppey earth-fast building and its successor with stone footings; b: London, Pudding Lane, PDN3 (after Horsman et aI., op. cit. in note 19); c: York, Coppergate Building A, late lIth-century horizon (after Hall, op. cit. in note 20, modified); d: Lincoln, Flaxengate Structure 9 (after Perring, op. cit. in note 21). Post-pipes are indicated by dark tone.

TIMBER BUILDINGS

353

self-supporting buildings were not considered sufficient to lead to the general adoption of stone footings. For example, four structures were attributed to the mid-roth­ century phase (Period III) at Flaxengate in Lincoln, but only one of these was built with stone footings, and the method of construction was not widely adopted in the city until the 13th century (Fig. 3d). The published buildings from London show an interesting diversity of plan and structure. A number had sunken floors, but with timber sills which were supported on stone footings. These should be probably categorized as earth-fast, although in structural terms they were little different from similar buildings set on the ground surface. The simple distinction between ground-fast and self-supporting is therefore not very useful in this case. The sequence of buildings at Pudding Lane illustrates the point. The first two buildings on the site (PON 1,2) each had a sunken floor and walls set down below ground-level supported on a timber sill set on a course of ragstone and Roman tile. In the early or middle lIth century a ground-level building (PON 3) was constructed on the site of the earlier structures with footings of ragstone, tile and lava quem which supported a grooved sill 0.24 m wide (Fig. 3b). The sill evidently held vertical staves for the walls. The authors suggest that PON 3 was provided with stone footings, not to protect the timbers against the damp, but to provide structural stability where the walls were built over the sunken area of the preceding building, PON2. The method of construction allowed the building to survive for a considerable period (about a century) during which it was much modified and extended (PON 3 to 5), and it seems improbable that the advantages of stone footings would have gone unrecognized. 19 The survival of extensive evidence from London allows the excavated remains of comparable buildings from York and Lincoln to be interpreted. Remains of two timber buildings were recorded in the mid- to late lIth-century levels at Coppergate (York). The first (A) had a ground sill about 0.2 m wide which was under-pinned by timber set at right-angles. The transverse timbers were to help distribute the weight of the building, but may also have been to level up the sill (Fig. 3C). In the second York building (B) the sills were laid directly on the ground and supported what appears to have been a plank floor, although the possibility that the planks were from a stave wall deserves further consideration in light of the discovery of that form of building in London. 20 Structure 9 at Flaxengate, Lincoln, dated to mid-roth century, had footings of tightly packed limestone rubble faced with larger stones and therefore may have resembled the buildings in London. The footings were 0.8 m wide on one wall, 1.05 m wide on the other and only one course deep (Fig. 3d). These are wider than the London footings and their width seems quite excessive for a timber sill. It is possible that they provided the base for a stone building, particularly as the floor surface within the building included crushed mortar. However, the excavators 19 V. Horsm an, C. Milne and G. Milne, Aspects of Saxon-Norman London: I. Building and Street Development near Billingsgate and Cheapside (London Middlesex Archaeol. Soc. Special Pap., II, London, 1988), 37-47, 7 I-4 and I02-IO; D. M. Goodburn, 'London's early medieval timber buildings: little known traditions of construction', 249-57 in G. de Boe and F. Verhaeghe (eds.), Urbanism in Europe: Papers of the 'Medieval Europe Brugge I997' Conference (Zellik, I997). 20 R. A. Hall, Excavations at York: The Viking Dig (London, I984), 117-18; P. V. Addyman, 'Vernacular buildings below ground', Archaeol. )., 136 (1979), 69- 75'

354

LAND, SEA AND HOME

considered it may have supported a timber structure because the footings were not bonded with mortar and the surface was even and suitable for a sil1. 21 All the buildings of the second type discussed above have been found on urban sites. That distribution may reflect the better survival of remains within the deeper stratigraphy of urban sites. One other building of the same type has, however, been found on a rural settlement, suggesting that the method of construction may have had wider currency. The sides of an early 12th-century building at Eynesbury (Cambs., formerly Hunts.) were marked by a clay sill wall standing up to 0.15 m high. Charcoal staining on the surface of the sill wall suggested that it may have supported a timber superstructure. There was no evidence for earth-fast posts, suggesting that it was a self-supporting structure. 22 EARTH-FAST POSTS AND GROUND-LEVEL WALLS

Buildings of the second group suggest that by the loth century builders were striving towards a more permanent form of construction which required the protection of structural posts from the damp. Until framed construction was developed in the later 12th or 13th centuries, there remained considerable problems in making self-supporting buildings. The third group of buildings to be examined here adopted a partial solution by which the main structural posts were earth-fast, but were placed within the building to protect them from the weather and reduce their exposure to damp in the soil. The posts might be set just within the walls or, better still, were placed well inside the building to form broader aisles. The external timbers which were most prone to rot might be placed at ground-level, or on a low sill wall to protect them from damp within the soil, or might be earth-fast. That form of building allowed the combination of the rigidity of earth-fast construction and a long-lasting structure by keeping the aisle posts dry. The lIth-century building at Hunter's Walk in Chester was of that form. The earth-fast structural timbers were set 0.2 m inside the line of the wall which had a ground sill. The timber sill was not underpinned by stone footings, but was laid directly on the sandstone paving which formed the ground surface (Fig. 4a). The line of the sill could only be traced where it crossed a Roman wall and stones had been removed. The presence of a hearth suggests the building had a domestic function. 23 The Chester house might be described as an aisled building, though the 'aisles' were very narrow in comparison to 13th-century examples. Narrow aisles are found in a number of Ilth- and 12th-century buildings where the need to protect the main structural posts from exposure to the weather may have been more important than the extra width gained within the building. Some of the buildings had ground-set exterior walls, which presumably could be replaced when D. Perring, Early Medieval Occupation at Flaxengate Lincoln (Archaeology of Lincoln 9/r, London, 1981), and 40. 22 C. F. Tebbutr, 'An early twelfth-century building at Eynesbury, Huntingdonshire', Proc . Cambridge. Antiq. Soc., 54 (1961 ), 85-9· 23 S. W. Ward, Excavations at Chester. Saxon Occupation within the Roman Fortress: Sites Excavated 1971-1981 (Chester, 1994),45-9. 21

II

TIMBER BUILDINGS

FIGURE 4. a: Chester, Hunter's Walk (after Ward, op. cit. in note 23); b: Faccombe

Netherton Building I I (after Fairbrother, op. cit. in note 24); c: Portchester Castle

Structure 16 (after Cunliffe). Post-pipes are indicated in solid black.

355

35 6

LAND, SEA AND HOME

they rotted without rebuilding the whole structure. Building II at Faccombe Netherton (Hants.), dated to the period c. 1070 to II80, had a pair of unequal aisles. The posts of the smaller stood only 0.3 m from the wall formed by timbers set within a trench (Fig. 4b).24 Not all the buildings of the period had narrow aisles. Some had aisles of a width comparable to those from the High Middle Ages. For example, Structure 16 at Portchester Castle (Hants.), dated to the 10th century, had three pairs of aisle posts with a return aisle at one end indicating an aisle of 2 m width. Only a fragment of the aisle trench survived, but stake-holes in the base suggested that the walls may have been of wattle. Timber posts may also have been present, although they were not recorded in the surviving length of trench. The aisle posts appear to have been the main weight-supporting timbers (Fig. 4C). There is a clear line of development from buildings from such structures as Portchester Structure 16, through the aisled halls at Cheddar palace - East Halls I and II - to some of the earliest standing houses. The sequence of East Halls demonstrates the rapidly changing forms of construction in the 12th century. The earlier building had earth-fast wall timbers and aisle posts, and dates to around 1100. It was replaced by a smaller building (East Hall II) with earth-fast aisle posts and un­ mortared stone footings for the walls. The stone footings for the walls were about 0.76 m wide, but there is no evidence that they supported a masonry structure, and J. T. Smith has suggested that they may have held a cob (mud) wal1. 25 The aisle posts would have probably been the main load-bearing elements. The date of construction is uncertain, though the destruction was dated by the excavator to the first half of the 13th century from pottery found among the backfill of the posts. Revision of the dating of Ham Green pottery, in particular, allows a rather earlier date for the destruction of c. II50-70 to be given. 26 Fyfield Hall in Essex is a standing building with timbers dated by dendrochronology to II67-85. The aisle posts were earth-fast, which seems to imply an uncertainty about structural rigidity of the timbers alone, although the building was properly framed. 27 The tradition of earth-fast construction persisted well into the 13th century even though it was no longer a structural necessity for framed buildings. The Manor House, Medbourne (Leics.), for example, was constructed in c. 1238 and had aisle posts which extended 0.9 m into the ground below the floor of the building. 28 These buildings provide a link between the earth-fast structures recorded in excavation and standing houses of the later Middle Ages. It is notable that the earliest surviving standing houses are all of aisled construction. 29 The builders had been able 24 J. R. Fairbrother, Faccombe Netherton: Excavations of a Saxon and Medieval Manorial Complex (British Museum Occ. Pap., 74, London, 1990), 114-18. 2S P. A. Rahtz, The Saxon and Medieval Palaces at Cheddar (BAR British Ser., 65, Oxford, 1979), 170-87. 26 M. Ponsford, 'Dendrochronological dates from Dundas Wharf, Bristol, and the dating of Ham Green and other medieval pottery', 8I-I03 in E. Lewis (ed.), Custom and Ceramics: Essays Presented to Kenneth Barton (Wickham, 1991). I am indebted for advice on the dating of pottery from the site from Dr Alejandra Gutierrez and Dr Chris Gerrard. 27 J. L. Walker, 'Fyfield Hall; a late twelfth-century aisled hall rebuilt c. 1400 in archaic style', Archaeol. J.,

1;,6 (1999), II2-42.

• N. Hill, 'The Manor House, Medbourne: the development of Leicestershire's earliest manor house', Trans . Leicestershire Archaeol. Hist. Soc., 75 (2001),36-61. 29 Walker, op. cit. in note 8, 21.

TIMBER BUILDINGS

357

to devise a method of construction that allowed the use of earth-fast posts which gave structural rigidity but also enabled the timbers to be kept dry and prevented them indefinitely from rotting. This had been the intention of earlier builders but, until the potentialities of the aisled plan were appreciated, it was not possible to make structures which would last. CONCLUSION

The examination of the third category of buildings has strayed well beyond the Viking Period. It has been necessary, however, to trace the changes to their conclusion with the appearance of framed construction in the late 12th and early 13th centuries which obviated the need for earth-fast footings. It is not suggested that the buildings considered here constitute a tradition of construction, nor do they represent a simple developmental sequence. Indeed, it would be more true to say that many of the buildings examined stood outside traditional forms of construction. They demon­ strate various attempts to deal with the problems of constructing timber buildings which would not be susceptible to decay. The absence of a single line of technological development can be emphasized by noting the pegged mortise-and-tusk tenon joints used in the wheelhouse of the second mill at Tamworth as early as the middle of the 9th century. In that building they served a necessary structural function to secure the timbers which were liable to considerable stress during the operation of the mill. There is little evidence for the wider use of pegged mortise-and-tusk tenon joints until their appearance in London and elsewhere in the lIth century.30 Though the Tamworth mill demonstrates that the ability to cut such joints existed, the need to use that form of construction and the structural possibilities it offered were not exploited in other 9th-century buildings. It is apparent, therefore, that the adoption of different methods of construction was not a simple consequence of technological innovation. Instead, it reflects changing social demands of the performance requirements of buildings. It has already been suggested that the buildings of the first group, those on high-status sites, did not need to be self-supporting for strictly reasons of economy or function. They were built in that manner to project an impression of longevity for social or spiritual purposes. Their methods of construction were not subsequently adopted on a wider range of sites, because there was not a wider perception in the 9th century of a need for enduring buildings. This makes it all the more interesting that from the loth century there was renewed interest in methods for constructing buildings with a greater life. The reasons for that need to be considered in the context of wider changes in society. From this time onwards villages became established on permanent sites, and there were more closely defined boundaries and rigid field systems. These changes mark a fundamental shift in the permanence of the landscape. 31 Settlements were being 30 P. Rahtz and R. Meeson, An Anglo-Saxon Watermiff at Tamworth: Excavations in the Bolebridge Street area of Tamworth, Staffordshire in 1971 and 1978 (CBA Res. Rep. 83, London, 1992), 128-31; Milne op. cit. in note 4,102-5; J. Blair, 'Archaeological discoveries at Woodeaton church' , Oxoniensia, 63 (1998),221 - 37. 31 D. Hooke, 'Anglo-Saxon estates in the Vale of the White Horse', Oxoniensia, 52 (1987), 142- 53; D. H . Hall, 'The Late Saxon countryside: villages and their fields' , 99-122 in D. Hooke (ed.), Anglo-Saxon Settlements (Oxford, 1988).

35 8

LAND, SEA AND HOME

established which were intended to last. The construction of buildings with greater longevity both reflected, and contributed to, the developing appreciation of an increasingly permanent and tightly structured environment. The view of the previous generation of archaeologists which drew a simple contrast between impermanent ground-fast buildings and permanent self-supporting structures has been shown to require revision. It has been possible to outline a more complex interpretation and one which places forms of building construction within a wider social context. However, the distinction between impermanent structures and those which were intended to have a longer life remains important, and it allows insights into the approaches and aims of those who commissioned and erected buildings. ACKNOWLEDGEMENTS

I am grateful to Dr Chris Loveluck for showing me his article on Flixborough before its publication,32 to Dr Alejandra Gutierrez and Dr Chris Gerrard for advising me on the dating of pottery from Cheddar, to Prof. Barry Cunliffe for information on the revised dating of a building at Manor Farm, Chalton and to the Arts and Humanities Research Board for a grant for excavation at Minster in Sheppey.

Dr Mark Gardiner, Department of Archaeology, The Queen's University, Belfast BT7 INN, UK m [email protected]

32 C. Loveluck, 'Wealth, waste and conspicuous consumption: Flixborough and its importance for Middle and Late Saxon rural settlement studies', 78-r30 in H. Hamerow and A. MacGregor (eds.), Image and Power in the Archaeology of Early Medieval Britain: Essays in Honour of Rosemary Cramp (Oxford, 200r) .

A PUSH INTO THE MARGINS? THE

DEVELOPMENT OF A COASTAL

LANDSCAPE IN NOR TH- WEST

SOMERSET DURING THE LATE rST

MILLENNIUM A.D.

By

STEPHEN RIPPON

The study of settlement and landscape in Viking-age Britain has been dominated by a number major themes, including the extent of Scandinavian colonization in eastern and northern Britain, the excavation of a relatively small number of sites that had possibly been occupied by Scandinavian settlers (notably in the uplands and western coastal regions of northern Britain), and the extensive reorganization of rural landscapes in central England that led to the creation of nucleated villages and open fields. What is less clear is how landscapes elsewhere were evolving at this time, and, indeed, what the English landscape was like before the creation of open field-based systems of agricultural production. There has been much debate over the origins of villages and open fields, though their creation clearly reflects the perceived need to exploit the landscape more effectively. This pressure on agrarian resources is also reflected in the expansion of settlement seen at this time in both upland areas (e.g. Ribblehead and Simy Folds) and coastal lowlands (notably the Somerset Levels, Romney Marsh, the North Kent Marshes and Fenland).l The colonization of coastal marshlands at this time was an important process in that being created from an entirely natural environment ­ saltmarshes - a cultural landscape emerged that was free from the influence of earlier periods. Clearly, particular conditions within these wetland environments constrained human behaviour to a certain extent, but the landscapes that were created will also reflect the pattern of settlement- and field-systems that was current within society at that time. The study area that is the subject of this paper has another advantage when considering the origins and development of medieval landscapes: there is some debate over the significance of variations in the physical environment in determining when and why open field landscapes emerged, but within this study area such factors were not important as the soils and relief were uniform. Should adjacent areas of marshland 1

S. Rippon, The Transformation of Coastal Wetlands (London, 2000) .

360

LAND, SEA AND HOM E

FIGURE 1. The Somerset Levels, including places mentioned in the text. (Drawn by the author)

develop nucleated and dispersed settlement patterns this cannot be due to environ­ mental factors. The area selected for this study lies in the north-west of Somerset, a county that by the medieval period had a remarkable diversity of settlement patterns (Fig. 1).2 In the east and south, the landscape was dominated by Midland-style villages and open fields, in the west there was a largely dispersed settlement pattern typical of South­ West England, whereas in the north was a more varied landscape with areas of both nucleation and dispersion. One place with nucleated settlement that has seen particularly intensive study has been the parish of Shapwick, on the Polden Hills between Bridgwater and Glastonbury (Fig. I). The work of Mick Aston and his team has shown how the present planned village replaced a landscape of more dispersed settlement probably during the loth century, and that this was part of wider landscape 2 A. Ellison, Medieval Villages in South·East Somerset (Bristol , I98 3); S. J. Rippon, The Severn Estuary: Landscape Evolution and Wetland Reclamation (London, I997), fig. 35; B. K. Roberts, The Making of the English Village (Harlow, I987), fig. 9.6; B. K. Roberts and S. Wrathmell, An Atlas of Rural Settlement (London ,

2000).

NORTH-WEST SOMERSET

36I

replanning as the Glastonbury Abbey estate of Pouholt, that covered most of the Polden Hills, was sub-divided into a series of smaller units. 3 A similar fragmentation appears to have occurred on the nearby estate at Wedmore and further north around the Gordano Valley (Fig. I),4 and in each case it would appear that ancient 'multiple estates' were being sub-divided and each new unit provided with a village and open field system. Two major themes, therefore, within the landscape of Viking-age Somerset were the expansion of settlement into the coastal marshes, and the fragmentation and subsequent reorganization of multiple estates. In North-West Somerset both occurred in the same area, and this study is about how a medieval landscape was created as a result.

A marginal landscape? Coastal wetlands are by their very nature not ideally suited for settled arable­ based agriculture, and as such would be regarded as a classic 'marginal' landscape. It is now appreciated that the concept of marginality is a complex one, and that the simplistic 'population-resource' model of Postan is inadequate to describe how human communities interacted with different physical environments. In particular, many landscapes that have poor potential for arable-based farming will have a high potential for other resource-exploitation strategies: coastal marshes, for example, can offer rich pastures and the opportunity for salt production. Despite their vulnerability to flooding, modern experiments and palaeo-environmental evidence from sites in Britain and mainland Europe have also shown that the cultivation of certain crops is possible on high intertidal saltmarshes, while once reclaimed, the agricultural productivity and as a result land values, could be exceptionally high. s There were, therefore, three broad options open to human communities with an area of coastal marshland within their territory. First, they could simply exploit the rich natural resources of the marsh, without changing its character in any way. Secondly, they could modify the marsh, for example through the construction of 'summer banks' which were designed to keep the relatively low summer high tides off areas of cultivation, but which made no attempt to control the higher winter floods. Thus, the whole marsh remained an intertidal environment but conditions in one 3 M . Aston and C. Gerrard, "'Unique, traditional and charming": the Shapwick project, Somerset', Antiq. j., 79, I-58. For recent studies of the 'great replanning' see C. Lewis, P. Mitchell-Fox and C. Dyer, Village, Hamlet and Field: Changing Medieval Settlements in Central England (Manchester, 1997) ; T . Brown and G. Foard, 'The Saxon landscape: a regional perspective', 67-94 in P. Everson and T. Williamson (eds.) The Archaeology of Landscape: Studies Presented to Christopher Taylor (Manchester, 1998) ; T. Williamson, Shaping Medieval Landscapes: Settlement, Society and Environment (Macclesfield, 2003). • Rippon, op. cit. in note 2,162-5' 5 W. Van Zeist, 'Palaeobotanical studies of settlement sites in the coastal area of The Netherlands', Palaeohistoria, 16 (1974), 223-371; idem, T. C. Van Hoorn, S. Bottema and H . Woldring, 'An agricultural experiment in the unprotected saltmarsh', Palaeohistoria, 18 (1976), II I -53; S. Bottema, T. C. Van Hoorn, H. Woldring and W. H. E. Gremmen, 'An agricultural experiment in the unprotected saltmarsh, part II', Palaeohistoria, 22 (1980), 127- 40; K.-E. Behre and S. Jacomet, 'The ecological interpretation of archaeobotan­ ical data', 8I- I08 in W. Van Zeist, K. Wasylikowa and K.-E . Behre (eds.) Progress in Old World Palaeoethnobotany (Rotterdam, 1991); J. D. Rhoades, A. Kandiah and A. M. Mashali, The Use of Saline Waters for Crop Production (F.A.O. Irrigation and Drainage papers, 48, 1992); A. Crowson, T. Lane and .J. Reeve, Fenland Management Project Excavations 199.3-199.5 (Sleaford, 2000).

362

LAND, SEA AND HOME

small area were adapted to make it more conducive for agriculture. The third strategy was the full-scale transformation of the landscape through reclamation: the construc­ tion of a sea wall along the coast and creation of a drainage system for the whole area now protected from tidal inundation. The colonization of coastal wetlands in this way was a high cost, high risk, but high return strategy towards landscape utilization: high cost in terms of the initial capital investment and subsequent recurrent cost of maintaining flood defences and drainage systems; high risk in terms of the constant threat of both tidal and freshwater flooding; but high return in terms of the rich pastures, meadows and high arable yields. Though difficult environments, wetlands could also be highly rewarding.

The North Somerset LeveLs The North Somerset Levels are some 100 sq km of reclaimed estuarine alluvium (former saltmarsh and mudflats) beside the Severn Estuary near Weston-super-Mare, some 15 km south west of Bristol (Fig. I). They were reclaimed during the Roman period, though the flood defences failed and the area reverted to an intertidal marsh. 6 By Domesday the Levels were re-occupied, though it is difficult to gauge how extensive settlement was for most of the wetland area fell within estates that also extended across the surrounding upland areas, making it impossible to determine where the recorded populations lived. Just one wholly marshland community is recorded, that of Kingston Seymour where there were 43 listed tenants/slaves and at least 21 ploughs. There has been some consideration of the process whereby reclaimed landscapes like the North Somerset Levels were created. It was assumed that the first stage must have been the construction of a sea wall along the coast, behind which communities then created settlements and fields? Evidence from the Netherlands, however, suggests that this need not have been the case if small individual areas of marsh were enclosed by 'summer banks'. A number of such low embankments that probably provided only seasonal protection have been recorded the Netherlands dating to the Roman period, including Broekpolder in the west,S and sealed beneath terpen mounds at Donjum-Heringa, Peins-Oost and Wijnaldum-Tjitsum in the north. 9 Such a landscape was still subject to winter flooding and so could not have been permanently settled (there is no evidence for continental-style raised settlement mounds or terpen on the North Somerset Levels). It would not have been economical to increase the height of many small embankments to the point at which they could protect an area

6 The first phase of the North Somerset Levels Project, investigating the Romano-British landscape, has been published in S. Rippon, The Romano-British exploitation of coastal wetlands: survey and excavation on the North Somerset Levels, 1993-7', Britannia, 31 (2000),69-200. The research presented here is derived from a second phase examining the origins and development of the medieval landscape, to be published as Landscape and Community: Medieval Wetland Landscapes in the North Somerset (CBA Res. Rep., forthcoming). 7 Rippon, op. cit. in note 2, fig. 7. 8 Linda Therkorn, pers. comm. 9 ]. Bazelmans, D. Gerrets,]. de Koning and P. Vos, 'Zoden aan de dijk: Kleinschalige dijkbouw in de late prehistorie en protohistorie van noordelijk Westergo', De Vrije Fries, 17 (1999),7-74;]. c. Besteman,]. M . Bos, D. A. Gerretts, H. A. Heidinga and ] . de Koning, The Excavations at Wijnaldum: Reports on Frisia in Roman and Medieval Times 1 (Rotterdam, 1999).

NORTH-WEST SOMERSET

FIGURE 2.

36 3

Aerial view of Puxton, showing the oval-shaped 'infield' enclosure and church. (Photo: author)

from year-round flooding, and so it may only have been when there was pressure to settle on the marshes all year round that a coastal sea wall was constructed.

The earliest marshland colonization The nature of the Domesday entry for Kingston Seymour suggests that by the mid-lIth century it lay within a reclaimed landscape, implying that the construction of a coastal sea wall, and any preceding phase of landscape modification ('summer banks', etc) had already occurred. In order to establish when this was, it is important to establish how the landscape subsequently evolved so that through retrogressive analysis the earliest landscape features can be identified and dated. What emerges from this analysis is that the earliest landscape elements, in marshland landscapes on both sides of the Severn Estuary, appear to be a series a roughly oval-shaped enclosures that clearly pre-date the surrounding pattern of fields and roads (e.g. Fig. 2). Referred to as 'infields' they have certain features in common (all of which are demonstrated at Puxton: Figs. 2-4):10 they are restricted to the higher, coastal areas of the Levels (that were the first to be settled, enclosed and drained), either in clusters or individually; 10 S. Rippon, 'Medieval wetland reclamation in Somerset', 239-53 in M. Aston and C. Lewis (eds.) The Medieval Landscape of Wessex (Oxford, 1994); idem, op. cit. in note 2, 172-3; idem, 'Infield and outfield: the early stages of marshland colonisation and the evolution of medieval field systems', 54-70 in T. Lane (ed.), Through Wet and Dry: Essays in Honour of David Hall (Sleaford, 2002).

36 4

LAND, SEA AND HOME

roads and droveways run towards the enclosures but then pass around them (suggesting that the enclosures are stratigraphically early in the formation of the historic landscape); their shape is generally oval (also suggesting that they were created relatively early, in a landscape that was not too cluttered with other features that would otherwise have constrained their shape); occasionally there is evidence for a bank running around the enclosures' perimeter; their size is typically c. 5-19 ha (12-47 acres; average 13 ha, 32 acres); extant farms are almost always located on the edge or just outside the enclosed area (suggesting that they represent areas of agricultural land, not an enclosed settlement, which has been confirmed by survey and excavation at Puxton: see below); a number are associated with churches or chapels (suggesting some pre­ eminence in the settlement pattern, again possibly indicating their relatively early origins). One of the main character-defining features of these enclosures is their oval shape which coring at a number of sites has established cannot be accounted for through the presence of underlying bedrock islands. l l They are far too numerous to be related to the morphologically-similar early Christian sites seen in Cornwall and Wales, suggesting that their distinctive form simply results from their having adopted the most economical shape for the first areas to be enclosed within a previously unsettled landscape (the same reason why some upland intakes and woodland assarts also assume an oval or sub-rectangular shape). One of these early 'infields', at Puxton, has been subject to a programme of survey and excavation (Figs. 2-3).12 The 6.3 ha (15.5 acres) oval-shaped enclosure is surrounded by a bank about 13 m wide, which, though damaged by modern ploughing, survives to a height of about 0.5 m, with an internal ditch (any external ditch is likely to have been retained as the present field boundary which, as a still functioning watercourse, could not be excavated!). A key question is the function of this bank: was it to seasonally protect an area from tidal flooding (i.e. a 'summer bank'?), or was it a precursor to the 'fen-banks' in protecting reclaimed areas from freshwater flooding due to run-off from the adjacent uplands? The geographical distribution of 'infield' enclosures rules out the latter: most lie on the higher coastal areas where freshwater flooding is unlikely to have been a problem. The morphology of the bank at Puxton may also suggest it functioned as a sea wall: such a broad but low bank would have suffered less from erosion than a steeper-sided one. 11 P. Gilbert, 'The pre-Conquest landscape at Kingston Seymour on the North Somerset Levels: report on survey 1996', Archaeol. Severn Estuary, 7 (1996), 53-7; S. Rippon, 'Roman and medieval settlement on the North Somerset Levels: survey and excavation at Banwell and Puxton, 1996', ibid., 39-52. 12 Rippon, op. cit. in note II; idem, 'Roman and medieval settlement on the North Somerset Levels: the second season of survey and excavation at Banwell and Puxton, 1997', Archaeol. Severn Estuary, 8 (1997), 41 - 54; idem, 'Medieval settlement on the North Somerset Levels: the third season of survey and excavation at Puxton, 1998', Archaeol. Severn Estuary, 9 (1998),69- 78; idem, 'Medieval settlement on the North Somerset Levels: the fourth season of survey and excavation at Puxton, 1999', Archaeol. Severn Estuary, 10 (1999),

65-73·

NORTH-WEST SOMERSET

36 5

For the Severn Estuary enclosures to have functioned as sea walls, they must have lain below the contemporary highest astronomical tides. High water level in the medieval sally port at Bristol Castle lay between II.8 and 12.4 m AOD, some 4 m above the modern level (6.95 m), though the value of this observation is questionable as sluice gates may have been used to retain tidal and river-waters within the ditch for defensive reasonsY More accurate measures of medieval Mean High Water Spring Tide (MHWST) have been gained from the height of accreted marsh deposits at the Bristol waterfronts at Dundas Wharf (about 6.4 m AOD) and Canynges House (about 6.6-6.7 m AOD), some 0.3-0.6 m below that of today at this point on the river (6.95m AOD).14It is difficult to scale this back to the open Estuary but the difference between medieval and modern MHWST in Bristol is likely to have been more rather than less due to the attenuation/weakening of tidal waves in the River Avon.15 In the Inner Severn Estuary, Allen has compared the elevations of still actively accreting saltmarshes with those that were reclaimed during the medieval period, and suggests that MHWST around 1300 was about 0.9 m lower than today, revising an earlier estimate of about 1.0 m;16 the height of MHWST in the outer Estuary adjacent to the North Somerset Levels is also likely to have been higher. Most of the North Somerset 'infield' enclosures lie at about 5.5-5.8 m AOD, the surface of these marshes reflecting the height up to which they had built up immediately before reclamation occurred. This was probably around MHWST which is now 6.1 m AOD, with the Highest Astronomical Tides (HAT) around 1.2 m higher. Assuming that the medieval figures were around 1 m lower (MHWST 5.1 m, HAT 6.3 m), it would appear that North Somerset infields were about 0.4-0.7 m above their contemporary MHWST, but about 0.5-0.8 m below the HAT. The enclosure at Puxton appears to date to around the IIth century or before. The bank itself yielded some residual Romano-British pottery and one sherd of mid­ IIth- to 13th-century ware (Shapwick fabric Ul), though fieldwalking had produced large amounts of medieval pottery (mainly 12thh3th century) from the area in which the bank was excavated. If the bank was constructed after the IIth century, it might be expected that some of this 12th- to 13th-century material would have become incorporated into its fabric. Excavations within the interior of the enclosure produced relatively large amounts of pottery dated to the late IIth or 12th centuries in stratified deposits. There are, however, sherds of an earlier ware dated to the late loth/early IIth century (Shapwick AA1): a handmade fabric, usually with a grey core, grey margins and buff to grey surfaces, with abundant glassy well sorted quartz < 2 mm and moderate well sorted limestone < 3 mm tempering, and moderate rounded voids < 3 mm. At Puxton the AAl sherds are always associated with late lIth- to 12th­ century material, and it is impossible to say whether they are simply residual, and \3 M. W. Ponsford, 'Bristol', 103-4 in G. Milne and B. Hobley (eds.), Waterfront Archaeology in Britain and Northern Europe (London, 1981) . 14 R. H . Jones, 'Industry and environment in medieval Bristol', 19-26 in G. L. Wood, R. H . Jones and M. W. Ponsford (eds.), Waterfront Archaeology (London, 1991). 15 J . R. L. Allen, pers. comm. 16 J. R. L. Allen and J. Rae, 'Vertical saltmarsh accretion since the Roman period in the Severn Estuary, South West Britain', Marine Ceol., 83 (1988),225-35 ; J. R. L. Allen, 'Saltmarsh accretion and sea-level movement in the Inner Severn Estuary: the archaeological and historical contribution',]. Ceol. Soc., 148 (1991),485-94­

366

LAND, SEA AND HOME

FIGURE 3, The shrunken medieval settlement at Puxton. Earthwork survey, along with field walking and soil chemistry (d. Fig, 4) show that only a small part of the oval-shaped 'Church Field' was ever occupied by settlement, the rest being small paddocks and enclosures. The other major area of deserted settlement earthworks at Mays Lane were occupied from around the IOthlr rth to the 17th century. (Drawn by the author)

NORTH-WEST SOMERSET

36 7

represent evidence for roth-century occupation on the site, or that at Puxton they date to the end of its range. 1? As pottery in Somerset is very scarce before the roth century,

this simply indicates that the site at Puxton was occupied by the roth/early lIth century or earlier. Smaller amounts of roth- or rIth-century pottery (including fabrics AAr and Ur) were also recovered from further excavations of shrunken settlement earthworks some 300 m to the north of the 'infield' at Mays Lane (Fig. 3).18 These earthworks are suggestive of an episode of secondary settlement expansion north­ wards from the primary core next to the 'infield', and if this had also occurred by the roth/lIth centuries, the initial creation of the 'infield' might be expected to have occurred somewhat earlier. So what was the function of the 'infield'? There is no evidence that it was an enclosed settlement. Earthwork, soil chemistry, and field walking surveys at Puxton all indicate that occupation was restricted to a series of slightly raised platforms in the north-eastern corner of the enclosure by the church, with the remaining area being occupied by small fields or paddocks, which the light scatter of mostly small and abraded sherds of medieval pottery suggest were lightly manured (Fig. 4). The enclosure appears, therefore, to have been an area of agricultural land with the associated settlement tucked to one side. Analysis of the landscape surrounding the Puxton 'infield' suggests that there was at least one further bank that may represent either a contemporary but less intensively used area of land (perhaps an 'outfield'?), or second phase of enclosure, still created in a landscape with no other features to constrain its shape (Fig. 7). The surviving field boundaries within this secondary intake are long and narrow in shape, and the post-medieval pattern of land ownership was highly fragmented (far more so than in the rest of the parish), suggesting some form of common field farming.

The tenurial context of reclamation The landscape of Puxton is typical of the coastal marshes of the North Somerset Levels, in that it was clearly created in a gradual and piecemeal fashion. During this early phase of reclamation there was no planning or co-ordination, in contrast to later periods when lower-lying areas were enclosed and drained in a systematic fashion, leading to carefully laid out planned landscapes. 19 The communities who created the landscape around Puxton, and indeed on much of the North Somerset Levels, were clearly allowed to act on an individual basis, and the tenurial context within which this occurred appears to have been a large estate based at Congresbury (Fig. 5). The reconstruction of this estate's former extent, and its subsequent fragmenta­ tion requires some complex back-projection of later documentary material. Until r772 Puxton was a chapel of Banwell to the south (Fig. r)/o which led some to assume A. Guitierrez, 'Pottery from excavations at Puxton (1996-99)' (unpub!. post-excavation report). Rippon (1998), op. cit. in note 12. . 19 Rippon (1997), op. cit. in note 2, fig. 7; idem, op. cit. in note I, fig. 51. 20 The Register of Ralph of Shrewsbury, Bishop of Bath and Wells, 1,29-136" ed. T. S. Holmes (Somerset Rec. Soc., 9-10, 1896); The Register of Robert Stillington, Bishop of Bath and Wells, 1466-I49I, ed. H. C. Maxwell-Lyte (Somerset Rec. Soc., 53, 1937); The Register of Thomas Bekynton, Bishop of Bath and Wells, 1443-1465, ed. H. C. Maxwell-Lyte and M. C. B. Dawes (Somerset Rec. Soc., 49, 1934); The Register of William Knyght, Bishop of Bath and Wells, 1)4I-I547, ed. H. C. Maxwell-Lyte (Somerset Rec. Soc., 55, 1940). 17

18

368

LAND, SEA AND HOME

FIGURE 4. Puxton Church Field: the results of fieldwalking (pottery distribution) and soil chemistry. The density of Romano-British material is suggestive of a manure scatter, and test­ pitting in adjacent fields suggest a settlement focus lies to the east. The distribution of medieval material is focused on a raised platform south-east of the church. The results of phosphate and heavy metal analysis shows a strong correlation with the distribution of medieval material. (Drawn by the author)

NORTH-WEST SOMERSET

369

that it was part of the royal (and later episcopal) estate of Banwell. 21 A range of evidence, however, suggests that Puxton was actually carved out of the neighbouring royal estate of Congresbury. The parish boundary of Puxton and Congresbury contains a series of intermingled parcels, which clearly suggest the former was carved out of the latter. The configuration of parish boundaries also suggests that a block of land known as Hewish was carved out of the far western end of Congresbury and transferred to Yatton to the north (Figs. 5.D and 7; and see below). Puxton, Congresbury, and Wick St Lawrence (a parish to the west of Puxton) also shared rights in the common meadows known as the Dolmoors. In (?) 1215 the parson of Congresbury confessed he had no right in the chapel of 'Pokereleston' and sought a pardon from the Prior of Bruton. 22 This may suggest that Puxton was once part of the parish of Congresbury, but that it was now separate and that the new boundary was starting to crystallize. A series of 14th- and 15th-century deeds (notably those of the lands acquired by Merton College, Oxford) refer to Puxton as lying in the manor of Congresbury, but parish of Banwell.23 It would also appear that the tithing of Puxton was part of Congresbury Hundred (which was later absorbed into Winterstoke Hundred: Fig. 5F), though the story is a complex one and involves piecing together fragments of documentary evidence from archives as far afield as Bristol, Taunton, Oxford and London. The 1327 and 1334 Lay Subsidies list Congresbury and Banwell, but there is no reference to Puxton. 24 The Congresbury Hundred court roll for November 1342 lists its tithings that included Yatton, Cleeve,25 Wick St Lawrence, Kenn, and Claverham/6 Puxton is not mentioned, but may be illegible. 27 The court roll for Nov. 1351 includes the same list, and a reference to the Libera decenna (free tithing).28 The roll for May 1379 lists the same tithings as in 1351, but preceded by Libera Decenna: Will. Ruyssworth, Will Greve, [?J, John Stretend: assize of ale; cert' redd' 6s.8d. 29 William Rushworth held the lands in Puxton later owned by Merton College. The 'cert redd' money was brought to court by each tithing, to meet the court's expenses which started as Id. per head - the tithingpenny - and became a fixed amount, often called the 'common fine'. The Puxton Account Rolls for 1472-3, 1474-5 and 1477-8 all include a payment of 4d. to the 'tithingman of the freemen's tithing' for the common fine (along with a common fine of G. B. Grundy, 'The Saxon charters of Somerset', Somerset Archaeol. Nat. Hist., 78 (1932), Appendix. Two Cartularies of the Augustinian Priory of Bruton and the Cluniac Priory of Montacute in the County of Somerset, ed. E. Hobhouse (Somerset Rec. Soc., 8, 1894), No. 135. 23 Merton College Archives 1204-II, 1214, 1217, 1220-1 and 1230-4. 24 F. H. Dickinson, Kirby's Quest For Somerset (Somerset Rec. Soc., 3, 1889); R. E. Glasscock, The 13.34 Lay Subsidy (London, 1975), 259. 25 In the parish of Yatton. 26 In the parish of Yatton. 27 Lambeth Palace Library ED II76, f. 22. 28 Somerset Records Office BAI DDISAS CI795. 29 Lambeth Palace Library ED 351, f. I. 21

22

370

LAND, SEA AND HOME

FIGURE 5. Possible extent of the 'greater Congresbury' estate, and its fragmentation. A: the original estate. B: detachment of Kingston Seymour pre-1066. C: division of the remaining estate between Congresbury and Yatton. Wemberham lies to the north of the Yeo, but south of the postulated early boundary between Congresbury and Yatton. D: transfer of one hide (Hew ish) from Congresbury to Yatton shortly after 1066; Wick remained a detached part of Congresbury. E: detachment of Puxton. F: arrangement of Hundreds. (Drawn by the author)

NORTH-WEST SOMERSET

37 1

3d. to the Hundred of Frowardeshill (Winterstoke).30 The court roll for Congresbury Hundred in May 1560 includes eight named men for the Libera decenn de Puxton ('free tithing' of Puxton),31 followed by those for the tithings (not described as 'free') of Kenn, Cleeve, Claverham, Yatton and Wick. In September 1560 the court rolls include under the heading Libera Decenn the tithingman Richard Cotton (not a Puxton name) presenting ten named Puxton men for not appearing in court, when summoned to pay suit there. In the 1622 Lay Subsidy, and 1664/5 Hearth Tax returns, Puxton was part of the tithing of 'Congresbury, Weeke [St Lawrence] and Puxton,.32 Congresbury Court Rolls for 1653 record that Edmund Cooke was appointed constable for Wick and Puxton. 33 It would appear, therefore, that Puxton along with Congresbury's chapelry at Wick St Lawrence to the west, and the parishes of Kenn and Yatton to the north, were once part of a 'greater Congresbury' estate (Fig. 5). The early history of Congresbury is obscure. A now lost charter of King Ine, dated 688/726, granted seven hides at Predian [Priddy] and 20 hides at Conbusburie [Congresbury] to the church at Sherborne. 34 Congresbury was still assessed as 20 hides in 1066. Sometime before 1033, King Canute granted Banwell and Congresbury to Dudoc, Bishop of Wells. Dudoc left the estates in his will to the See, though they were seized by King Harold in 1060. On Harold's death in 1066, Banwell and Congresbury passed to King William. However, Giso, Bishop of Wells petitioned the king and a charter of 1068 confirms the restoration of 30 hides at Banwell including Compton Bishop.35 The king retained Congresbury but granted Yatton to the bishop instead, along with one hide of land in Congresbury. Domesday records two hides as having been removed from Congresbury after the Conquest. 36 One, held by Giso, Bishop of Wells, appears to have been Hewish, an extension of Yatton parish to the south of the Congresbury Yeo river (Fig. 5D). The Domesday entry for Yatton includes reference to 'A pasture called Waimora [Wemberham?] ... which before 1066 belonged to Congresbury,.37 The Exon Domesday elaborates: 'Of this manor's [Yatton] land Fastrad holds five hides from the Bishop, Hildebert four hides. Of the four hides which Hildebert holds, a woman, Aethelrun, had one hide jointly in 1066. With this hide, which Aethelrun held, lies a pasture called Wemberham.>38 Therefore, this hide, with the pasture at Wemberham, was transferred from the king's manor of Congresbury to the Bishop's manor of Yatton between 1066 and 1086. Although the area currently known as Wemberham Puxton account rolls 1472.-8. Somerset Records Office DD/WY/84. Bristol Records Office 042. 35. 32 Somerset Records Office DD/SASci2.75 BK81; E. Dwelly, National Records, Vol. I: Hearth Tax for Somerset 1664-; (Fleet, 1916); E. H. Bates (ed.), Quarter Session Records for the County of Somerset: Vol. 1 James 1 (Som Rec. Soc., 2.3),158 and 170. 33 Bristol Records Office BMC/~37b. 34 H. P. R. Finberg, The Early Charters of Wessex (Leicester, 1964), 372.. 35 W. H. B. Bird, Calendar of the Manuscripts of the Dean and Chapter of Wells, Vol. 1 (London, 19°7),431; P. H. Sawyer, Anglo-Saxon Charters: An Annotated List and Bibliography (London, 1968), No. 1042.. 36 Domesday Book, Somerset, ed. C. Thorn and F. Thorn (Chichester, 1980), No. I.2.I . 37 Op. cit. in note 36, No. 6.14. 3R Op. cit. in note 36, 317. 30 31

LAND, SEA AND HOME

37 2

lies north of the Congresbury Yeo this results from the river having changed course (Fig. 5C), and the hide in question is logically the extension of Yatton to the south of the river (Fig. 50). The configuration of the parish boundaries of this area certainly suggest that this extension of Yatton was carved out of Congresbury, and its transfer would have served to link the Bishops, two manors of Yatton and Banwell. The other hide removed from Congresbury after the Conquest was held by Serlo of Burcy and Gilbert son of Thorold. Gilbert held Kewstoke,39 while Serlo held Woodspring for a short time after the Conquest. 40 Logically, the hide removed from Congresbury was Wick St Lawrence, which lay adjacent to Kewstoke and Woodspring and does not otherwise appear in Domesday. Wick St Lawrence was subsequently returned to Congresbury, with which it had strong links throughout the rest of the medieval period. It has, therefore, been established that the royal estate of 'greater Congresbury' encompassed Puxton, Wick St Lawrence, Yatton and Kenn. The final element is logically Kingston Seymour whose place-name suggests it was once royal property.41 The Domesday entry for Kingston Seymour is also very curious. The two manors were assessed as follows: -

I hide, land for I7 ploughs, and valued at £6 in I066 and 1086;42

- 4t hides, land for 7 ploughs, and valued at 60S in I066 and I086. 43 For the second manor, Domesday adds, 'Before I066 this manor did not pay tax except for I hide', and the huge difference between the one hide and seventeen ploughlands in the first manor also suggests tax exemption (though this is rare in Somerset).44 Kingston Seymour was also later a detached part of Chewton Hundred (Fig. 5F), itself a royal manor with a recorded tax exemption in I086. 45 It can, therefore, be postulated that a large part of the North Somerset Levels lay within a 'greater Congresbury' royal estate. So did the colonization of the North Somerset Levels occur in the context of this estate, or following its fragmentation? The fragmentation appears to have occurred in several stages: Kingston Seymour was detached at an unknown point before I066 (Fig. 5B), as was Yatton (Fig. 5C). Hewish was removed from Congresbury and added to Yatton shortly after I066 (Fig. 50). Kenn had been detached at some point before 1086. The date when Puxton was removed is unclear as it cannot be identified in Domesday, though it may have been one of three unnamed hides within Congresbury that were held by Alfward, Ordric and Ordwulf.46 If we return to an analysis of the historic landscape, we can at least suggest the relative date at which certain key events occurred. The starting point is Puxton. Its location, some 10 km from the coast is far from the most favourable place on the 39

40 41 42 43 44 45 46

Op. cit. in note 36, No. 42.. I. Op. cit. in note 36, No. 2.7.3. M. Gelling, Signposts to the Past (Chichester, 1978), 184­ Op. cit. in note 36, No. 5.63. Op. cit. in note 36, No. 5.64. M. Costen, The Origins of Somerset (Manchester, 1992.), 12.3. Op. cit. in note 36, No. 1.2.9. Op. cit. in note 36, No. I. 2.I.

NORTH-WEST SOMERSET

373

Levels. The highest ground actually lies towards the coast, as it is here that tidal waters dumped most of their sediments as they flooded what was a saltmarsh environment (the same phenomenon explains the formation of natural levee banks along major rivers). Thus, the area around Kingston Seymour lies at about 5.8-6.r m AOD, whereas the ground around Puxton lies at c. 5.5 m AOD. Logically, therefore, we would expect the higher, coastal areas to be settled first, and only later were people forced towards the lower-lying backfen areas such as Puxton (Fig. 6, left). This simple expansion of settlement from the coast inland is seen most graphically in the Norfolk Marshland where extensive field walking has allowed the steady spread of settlement to be mapped century by century.47 As Puxton was occupied by the IIth century (and possibly earlier, such is the lack of datable pottery), it might seem logical that the Kingston Seymour area was settled somewhat earlier than this. The model for marshland colonization outlined above - of simple expansion from a primary location in Kingston Seymour implies an intact 'greater Congresbury' estate within which settlement expansion was a relatively simple process whereby colonists left the dryland, occupied the coastal area, and that subsequent movement into the backfens only occurred as the more favourable areas were fully occupied. Another strand of evidence, however, suggests that it was not that simple. As described above, shortly after ro66, the far western end of Congresbury (Wemberham) was transferred to Yatton. The western and south-western boundaries of Wemberham are marked by naturally meandering watercourses, but its eastern boundary is very straight and clearly artificial. However, this boundary does not cut through other landscape features such as fields or roads: the historic landscape clearly post-dates the laying out of this boundary, suggesting that this area of the Levels was not enclosed and drained at that time (in contrast to the area around Hewish and to the south around Puxton which had a landscape of 'infields' similar to that of Kingston Seymour). This in turn suggests that there may have been at least two primary foci of marshland colonization: one in Kingston Seymour from where the higher coastal areas were settled, and the other in that part of Congresbury itself which lay on the marshes (Hewish) (Fig. 6, right): the colonization of Puxton was, therefore, a 'push into the margins' - the colonization of lower-lying ground to the south of Hewish - but within the more limited context of Congresbury parish. The area in between was largely unoccupied at this time. This proposed model, that the Puxton area was colonized from Congresbury, rather than simply as an extension of settlement expansion from the coast, has further implications (though this entails yet further uncertainties). If the people of Puxton were simply the marshland community of Congresbury, then they cannot have been responsible for building the sea wall along the coast. If the interpretation of the 'infield' enclosures as 'summer banks' is correct, then when the Puxton area was colonized there must still have been a risk of tidal flooding. That implies that the sea wall along the coast was constructed after the fragmentation of the 'greater 47 R. J. Silvester, The Fenland Project, Number .3: Norfolk Survey, Marshland and the Nar Valley (East Anglian Archaeol., 45, 1988); idem, '''The addition of more-or-the undifferentiated dots to a distribution map"? The Fenland Project in retrospect', 24-39 in J. Gardiner (ed.), Flatlands and Wetlands: Current Themes in East Anglian Archaeology (East Anglian Archaeol., 50, 1993); Rippon, op. cit. in note 1,208-11.

374

LAND, SEA AND HOME

FIGURE 6. Two models for the colonization of the North Somerset Levels during the late 1st millennium A.D. Left: initial settlement on the higher coastal marshes, with a gradual expansion eastwards into the lower-lying backfens. Right: initial settlement in Kingston Seymouron the higher coastal marshes, and in the western part of the Congresbury Marshes, with subsequent expansion from both primary foci. (Drawn by the author)

NORTH-WEST SOMERSET

375

Congresbury' estate and the detachment of Kingston Seymour. The creation of coastal sea defences must, therefore, have been a collaborative effort between several manors/ communities.

Landscape evoLution on a 'clean sheet' The marshland landscape of the North Somerset Levels was, therefore, colonized in the late 1st millennium A.D. The cultural landscape was created on a 'blank sheet': a relatively featureless marsh with only the occasional former creek to influence how the pattern of fields, roads and settlements would evolve. This was the period when, on the Polden Hills to the south, a landscape of open fields and planned, nucleated villages was being created, and similar episodes of replanning appear to have occurred nearby on Wedmore and around the Gordano Valley.48 At first sight it appears that no such 'Midland'-style landscapes, however, were created on any of the medieval reclamations on the Somerset Levels despite their being created at the same time: these areas were characterized by predominantly dispersed settlement patterns, of isolated farmsteads and small hamlets, and a number of small common-fields amongst otherwise enclosed field systems. The area around Puxton illustrates the diversity present within these more irregular landscapes. Figure 7 shows selected facets of the historic landscape in Puxton, that part of Congresbury that extended onto the Levels (Congresbury Marsh) and Hewish (the hide removed from Congresbury and added to Yatton to the north). The settlement patterns in these areas are remarkably different (as mapped in the 19th century and which a combination of archaeological survey, standing building recording and documentary research has shown can be traced back to at least the 16th century). The eastern part of Puxton was occupied by a single nucleated settlement, while the western area was farmed from a loose green-side hamlet south of Ashfield. The whole of Hewish was settled by two small hamlets (East and West Hewish), in contrast to Congresbury Marsh where isolated farmsteads predominated. These local differences in settlement pattern are also reflected in the pattern of landholding as mapped in the Tithe Survey, and which through earlier documentary material can be traced back to at least the late 15thh6th centuries (Fig. 7A). In eastern Puxton individual tenements, typically around 15 acres, consisted of widely scattered parcels of land, as was also the case in the immediate vicinity of Ashfield. A map of c. 1770, and an examination of the field-names, field-boundary pattern and earthworks allows former open fields to be reconstructed in both these areas, that to the south of Puxton surrounded by a substantial bank (Fig. 7B). The rest of western Puxton, along with Congresbury Marsh and Hewish, was divided between largely nucleated landholdings of typically 40 acres or larger; there is no evidence for open fields in these areas. The dating of these two different landscape types has been discussed above. The primary settlement core at Puxton (the 'infield' enclosure) was in existence by the late loth-Ilth centuries, when it appears that the new landscape was created with a tendency towards communality. Hewish was carved out of Congresbury sometime between 1066 and 1086 and this occurred before that area was colonized; the pattern 48

Aston and Gerrard, op. cit. in note 3; Rippon, op. cit. in note

2,162-5 .

376

LAND, SEA AND HOM E

FIGURE 7. Settlement, landholding and field systems in Puxton, Congresbury Marsh and Hewish. A: distribution of landholdings c. 1600 showing the highly scattered pattern in Puxton and more nucleated tenements in Congresbury Marsh and Hewish. B: the primary settlement foci ('infields') and evidence for open fields in Puxton. (Drawn by the author)

NORTH-WEST SOMERSET

377

of dispersed settlement and enclosed fields must, therefore post-date this period. For whatever reason, communal landscapes were no longer created as part of the process of settlement expansion in the rump of the 'greater Congresbury' estate.

Conclusions To a considerable extent, this paper is highly speculative, although it demon­ strates two things, one methodological and the other thematic. First, we can study landscape through archaeological or documentary evidence, but their value is greatly enhanced when they are integrated with the richest historical record of all: the historic landscape. The morphology of landscape can be unravelled through historic landscape characterization and retrogressive analysis, to suggest the way(s) in which it evolved. These models then need to be verified through fieldwork, as has begun at Puxton. This paper has also outlined how there was a major phase of settlement expansion into a physically marginal environment starting around the lOth/lIth centuries (or possibly earlier). The landscape that was created was relatively unrestrained by existing features, and as such might reflect a relatively 'clean' example of a landscape dating to this period. This marshland colonization appears to have occurred within the context of a major estate that had already started to fragment. The diverse settlement pattern and a mixture of open and enclosed fields was evolving at the same time as Glastonbury Abbey and other major landowners in Somerset were reorganizing their estates with the creation of nucleated villages and open fields. Puxton itself presents evidence for this approach to landscape management being created around the loth/lIth centuries within the former 'greater Congresbury' estate, other parts of which, however, had a highly dispersed settlement pattern. There has been considerable debate recently about the relative significance of cultural and environmental factors in determining where open-field based landscapes emerged,49 but in the case of Puxton that such marked local variation in landscape character emerged, on what was a physically uniform environment, can only have been due to the different strategies lords and their tenants adopted towards estate management. The decision whether or not to adopt the new fashion for villages must have depended upon its perceived advantages and disadvantages, and the freedom of lordsltenants to undertake such a restructuring of their agrarian resources. Dr Stephen Rippon, Archaeology Department, University of Exeter, Laver Building, North Park Road, Exeter EX4 4QE, UK [email protected]

49 E.g. Lewis et al. and Williamson, 0pp. cit. in note 3.

PLACE-NAMES AND THE HISTORY OF

SCANDINAVIAN SETTLEMENT IN

ENGLAND::­

By

LESLEY ABRAMS

and

DAVID

N.

PARSONS

'The study of place-names is not a respectable occupation for a scholar.,t

Students of the Viking Age have an especially blank canvas on which to sketch out their hypotheses about settlement and society in Anglo-Scandinavian England. 2 The written sources of the late 9th century in particular, though in fact more substantial than for many other periods of Anglo-Saxon history, tell us about some things in detail, but on others they are almost silent. The thoughts of King Alfred (as prepared for public consumption), or the West Saxon view of history, or even the Viking wars, are relatively well documented. If, however, you are interested in the kingdoms created once Viking armies 'shared out the land ... and proceeded to plough and to support themselves', as the Anglo-Saxon Chronicle famously reports,3 or in the settlements established when Vikings were driven from Ireland to North­ West Britain soon after 900, written sources of any date offer precious little - and contemporary written sources even less. Anglo-Scandinavian society of the later loth and lIth centuries is only slightly less obscure. When the going gets tough in this way, historians resort, not just to the usual speculation, but to other disciplines, foremost among which is place-name studies. Engaged with or competent in or sympathetic to these disciplines to varying degrees, historians have focused naturally enough on their own historical questions, and they have for quite some time been badgering specialist practitioners for answers to them. However, as Eric Christiansen has remarked, 'what actually happened in Northumbria, Mercia, and East Anglia during the first phase of Nordic occupation is unknown, and it has been the misfortune of placename students that so many historians have expected them to find out,.4 Dawn Hadley has lately characterized past efforts to wring historical information from place-names as 'tired' ,'- This long paper departs from the editorial conventions of the present volume in two respects: (I) it uses an author-date system of reference, a concession to the amount of secondary literature reviewed; (2) following usual practice in place-name studies, it makes reference to the historic English counties before local government reform in 1974. 1 An anonymous archaeologist, quoted in Wainwright 1962,98. 2 For recent studies, see Hadley and Richards 2000 and Graham-Campbell et al. 2001. 3 ASC s.a. 876. For Chronicle references, unless directed elsewhere, see Whitelock 1961. We follow her revised dates throughout. 4 Christiansen 2002, 229.

380

LAND, SEA AND HOME

and 'unsophisticated,.5 On the other hand, F. T. Wainwright declared that 'without archaeology and place-names, ... the significance of the Scandinavian impact on the British Isles would scarcely be comprehended at all',6 and a place-name distribution map is reprinted to illustrate almost every discussion of the topic. From the point of view of the historian, every interpretation since Frank Stenton set his out in the 1920S7 has had an essential logic, despite dramatically contradicting others anchored on the same information: lack of linguistic expertise has disqualified most historians from confronting these contradictions at source. Place-name scholars, on the other hand, have constructed various narrative accounts of the progress of Scandinavian settlement which, while they can be made to fit the pathetically few historical 'facts', have failed to find a consensus of agreement. Have we perhaps reached a dead end?8 Although there is in this area such a weight of scholarly tradition that everything seems to have been said, and firmly objected to, before, the authors of this paper ­ an historian and a place-name scholar - have undertaken the exercise of reviewing the evidence and the way it has been used by their respective disciplines, with a view to understanding how conclusions have been drawn and deciding how far further co­ operation could be productive. What follows is an account of past historical scholarship and a fresh assessment - inevitably personal - of the value of name­ evidence to an interpretation of Scandinavian settlement. There are other strands to the linguistic evidence/ but this paper will concentrate principally on place-names. Its focus will also largely be restricted to eastern England: Yorkshire, the East Midlands and East Anglia. Students of the Scandinavian settlement of England have repeatedly posed the same basic questions. When (and in how many stages) did the Scandinavian settlement take place? Where did Scandinavians settle? How many of them were there? Where had they come from? What kind of land was taken - prime estates previously in English hands or uncolonized areas? Attention has also focused on the nature of the Scandinavians' exploitation of the land, their status and their relations with the local inhabitants. What kind of immigrants were they: an dite ruling class, or demobilized military men of all ranks, or mainly peasant farmers? Were they all men? How much of the native population survived, and where, and in what capacity? Further associated questions concern the nature of the society the immigrants established. Was Scandinavian England freer, less subject to ties of dependency and lordship, than the rest of the country, as the much later evidence of Domesday Book has been taken to suggest? Did the foreigners import social structures from home or assimilate to (or adapt) native ones? Finally, there has been much debate on the degree to which their arrival affected the settled areas. How much did it change the native way of life? Did Hadley 2000a, 329-30 and 21-2. Wainwright 1962, 3. 7 In, for example, Stenton 1927 ('The Danes in England'). Note that in this paper we, like Stenton, sometimes use the terms 'Danes' and 'Danish' as generics, without necessarily implying an origin in Denmark (as opposed to elsewhere in the Scandinavian homelands). • '[EJarlier attempts to use place-name distributions to chart the movements of armies, to indicate a rea s of colonization, to identify the precise locations where groups of incomers settled and to assess the scale of Scandinavian settlement simply have to be abandoned': Hadley 2002, 56. 9 For recent reviews, see Townend 2000 and Parsons 2001. 5

6

PLACE-NAMES AND HISTORY

38r

it create, stimulate, or disguise regional distinctions? Were the settlers, whether a mass migration or elite takeover, responsible for roth-century developments such as the reorganization of landholding, the transformation of the landscape, the boom in the economy? In the almost complete absence of written sources which bear on these aspects of the Scandinavian settlement, historians have looked to place-names for solutions. First among the answers hoped for has been that of date, but, from the historian's point of view, this has proved disappointingly complex; judgements have varied on the date of particular types of place-name, such as those ending in Old Norse -by or -thorp, or the hybrid English-Norse forms, including those which involve Scandinav­ ian personal names. The proliferation of place-names with Scandinavian (or Scandinavianized) elements in documents of the later medieval period has complicated the issue of the date of formation of those recorded in such profusion in Domesday Book (and in rather less profusion before). Questions about status and function ­ could types of name indicate types of settlement, such as primary centres or subordinate units? - have also been posed in the search for clues to the way the land thus named was held and/or utilized. Historians have also hoped for assistance in determining the origins of the settlers (Danes, Norwegians, Hiberno-Scandinavians?), in tracing language use, and in assessing the social and political context of the names' formation and transmission. This last question did not much trouble the first generation of 20th-century historians who considered the place-name evidence of Scandinavian England. To them a Scandinavian place-name meant a Scandinavian settlement. This kind of clarity has since been challenged by, on the one hand, more subtle interpretations of ethnicity,lO and, on the other, increasing awareness of the complexities involved in naming. Who gives a place its name - the inhabitants, the neighbours, or even the landlord or the authority that first recorded it in writing? What circumstances provoke renaming? How do new names become fixed? What circumstances affect survival? How can we allow for language-contact that may have preceded the coining of the names? Habits of name-giving react to changing political, social, and economic conditions, but they might also reflect linguistic or dialectal 'fashions' which are not necessarily easily related to historical circumstances. It will not be possible to attempt a full consideration of this complex field here - to do justice to the subject would require comparative material from far beyond Anglo-Scandinavian England - but it will be necessary in due course to give some consideration to the problematic nature of place-name evidence, and to reassess the kind of historical information that it might possibly be able to supply. PREVIOUS WORK

Before that, however, we shall review some of the conclusions that historians have derived from the place-name evidence. First, the issue of chronology and sequence of settlement. The Viking wars escalated when the Great Army arrived in 10

Innes 2000; Hadley

2002.

382

LAND, SEA AND HOME

England from the Continent in 865. Parts of this army defeated three of the four Anglo-Saxon kingdoms, and the Anglo-Saxon Chronicle testifies to settlement in Northumbria, eastern Mercia, and East Anglia in 876, 877, and 880, respectively. A. H. Smith's distribution map of parish names (Fig. I) seems to offer an exquisitely clear delineation of the lands involved. l l That, at least, has been the assumption of many historians, but the complication of course is one of date. Do these names derive from that conquest phase of the 870S and 880s, or are they the product of a more established society, created perhaps by later arrivals, or by the extension of Scandinavian activity over time, or by other factors? Answers would usefully support the routine historical task of sequencing developments, but dates matter even more because the historical situations of the 870S (Viking conquest), of the 920S and 950S (when Scandinavian territories north and south of the Humber, respectively, had in turn been conquered by the English), or even of the I030S (the heyday of Anglo­ Danish rule under King Cnut) differed enormously - as would the interpretation of the names' significance at these different dates. Similarly, the narrative of events in different parts of England, if we but knew it, would have varied dramatically. The North-West, the kingdom of Northumbria, the Five Boroughs, the South-East Midlands, East Anglia - all had different histories with respect to Scandinavian activity. Many have favoured the association of Scandinavian place-names with the original Viking conquest: 'The large groups in which [names in -by] tend to be concentrated suggest the conditions of an age when the Danish settlers in England still felt themselves strangers in a hostile land', argued Stenton. 12 Between I965 and I97I Kenneth Cameron examined in detail the place-names of the North-East Midlands and proposed that three stages of settlement could be discerned: the takeover of English settlements by the victorious army, marked by the 'Grimston­ hybrid' type names, a secondary migration of Danish colonizers who came in 'behind the protection provided for at least two generations [i.e. 877-c. 9I8] by the armies of the Five Boroughs', evidenced by the by-names, and a third phase of activity at 'a time when a greater integration had taken place between Dane and English', indicated by the names in (-)thorpY A secondary migration is not mentioned in the sources, but for those who had accepted Peter Sawyer's contention that the Viking armies were small, it provided a solution for the difficulty that their demobilization could not account for the large number of Scandinavian place-names. 14 In I97I Sawyer himself Originally published in Smith 1956, Map 10. Stenton 1942,305. Stenton claimed that, although they were not all of the same date, the names in -by 'give the best illustration that can now be found of the way in which the Danish settlement of eastern England was carried out'. 13 For thorp-names see Cameron 1970 (quotation from p. 143). For by-names see Cameron 1965 (quotation from p. 121) . For 'Grimston-hybrids' see Cameron 1971, where it was suggested that the settlements represented early Scandinavian takeovers of English villages; the further proposal that they might contain the names of leaders of the Great Army was added in Cameron 1976,22. Note that we have decided to stick to the traditional designation of the hybrids, though instances of the place-name Grimston itself may be anomalous (d. Gelling 1988,233-4); in the light of this, Cameron later adopted 'Toton-hybrids' as an alternative designation. 14 Sawyer 1957- 8; see also Sawyer 1962, chs 6 and 7. Nicholas Brooks's rebuttal (1979) has been widely accepted; see also Wormald 1982C, esp. 134- 7. Niels Lund (1969) initially expounded the idea of a secondary migration 'not long after the conquest of England', but came to revise his view (Lund 1981). 11

12

PLACE-NAMES AND HISTORY

FIGURE I.

38 3

A. H. Smith's 1956 map of the Scandinavian settlement of England. (Reproduced by permission of the English Place-Name Society)

38 4

LAND, SEA AND HOME

suggested a chronological progression from hybrid names to by-names, but with settlement contained within two phases: 876 x 880 and 896.15 The by-names, he argued, marked an early, specifically Scandinavian, expansion of settlement, the beginning of a widespread process of 'internal colonization' in England which continued through the Middle Ages. 16 More recently, Sawyer has proposed that 'the main period of Scandinavian name production was in the early years of the tenth century', the outcome of a series of Scandinavian military reverses. I? A similar date has been suggested by Gillian Fellows-Jensen: 'It was probably not until after 900 that intensification of settlement and cultivation obliged the Danes to begin to coin new names.'18 Both scholars have argued that by-names combined with personal names are especially indicative of a loth-century date; Fellows-Jensen added the refinement that by-names combined with other elements, in contrast, might be earlier, representing 'the first phase of Danish settlement in England' .19 Alternative chronolo­ gies, such as that proposed by Hadley, include the proliferation of Norse names during the period of Anglo-Scandinavian cultural assimilation in the loth and early IIth centuries. 20 Others have seen the coining of names as an ongoing process, continuing not just throughout the Anglo-Saxon period but beyond, into the post­ Conquest era. 21 Hadley's recent examinations of naming-processes have moved away from the traditional attribution of Scandinavian place-names to the language practices of Norse-speaking populations and instead stressed the role of fashion in their formation. She has in fact pursued the idea that Scandinavian names were coin able even by non­ Norse-speakers to the point of urging that we set the distribution map aside, arguing that it cannot act as an index of settlement but represents instead 'the product of the conscious and unconscious decisions made by the inhabitants of the ... Danelaw' of both Scandinavian and Anglo-Saxon background. 22 Since Norse place-names need not imply Scandinavians, she argued, there is no question of using them to identify locations or to date phases of settlement: 'in other branches of early medieval scholarship', she has claimed, 'few would now attempt to use place-name evidence to establish a chronology of settlement'.23 Issues of date and distribution inevitably raise the fundamental and thorny question of numbers. On the basis of place-names and field-names (the latter not on Smith's map but surviving in quantity, largely in later medieval documents), Stenton judged that Scandinavian numbers had been considerable. He described the place­ names as the 'most permanent memorial' of the settlement, which had 'the dimensions 15 Sawyer 1971, 166- 7 and 173. In Sawyer 1982, 102, he added a third phase of settlement (in the North, involving the Hiberno-Scandinavian takeover of York before 920). 16 Sawyer 1971, 174-5; a later phase of Anglo-Scandinavian expansion is said to be marked by the thorps on the Yorkshire Wolds (p. 167). 17 Sawyer 1982, 102- 7. The quotation is from p. 103. Sawyer's argument is set out in greater detail below, pp. 390-1 and 404-5. 18 Fellows-jensen 2OOla, 280. 19 Fellows-jensen 1999,49. Fellows-jensen's arguments are also considered in more detail below, pp. 406- 7. 20 Hadley 1997,69-75; 2oooa, 17- 22 and 329-35; 2ooob, 122-8; 2001,13-14; 2002, 56-62. 21 For example, Sawyer 1957-8, 8- 12, and Keynes 1997,64- 7. 22 Hadley 2000a, 334; 2002, 56-7. Cf. Sawyer 1957- 8, 10-14. 23 Hadley 2oooa, 21.

PLACE-NAMES AND HISTORY

38 5

of a migration' .24 Sawyer's reinterpretation of the sources dismissed the connexion between density of names and density of settlement, but Cameron's secondary migration, while accepting Sawyer's argument for small armies, continued to envisage significant numbers of settiers. 2S Niels Lund widened the debate by pointing out that although linguistic influence was clearly considerable, and not just on place-names, there was no formula for converting influence into numbers of speakers, or numbers of speakers into numbers of names. Since the linguists could not agree, he complained, the linguistic impact could not be used as a firm indicator of a massive immigration. 26 There would have been other reasons for renaming (which we shall come to) which could account for the density of place-names. Hadley's emphasis on names being the effect of interaction and resulting from contact between cultures has similarly thrown doubt on past assumptions about numbers.2? According to one way of thinking, therefore, lots of names do not necessarily mean lots of Scandinavians. This conclusion, though adopted by some, has left others uneasy. If Scandinavian numbers were few, how might the many Scandinavian place­ names be explained? One answer offered has been the status of the immigrants. A politically and socially dominant elite, it has been argued, will influence naming practices to such an extent that the effect will be disproportionate, and quantities will therefore mislead. 28 Patrick Wormald has rejected this claim: 'a mere change of landlords will not account for all the evidence' of Scandinavian impact. 29 Lund argued that historical sources show that Viking armies were made up of the warrior elite, not lesser men, and that place-names therefore 'are not evidence of the settlement of armies', but must indicate the presence of a peasant class. 30 Pauline Stafford similarly has found the place-name evidence inconsistent with 'settlement consisting of no more than aristocratic takeover', but was uncomfortable with the idea of a mass migration. Despite the large numbers of names, she concluded that the Scandinavians in the East Midlands did not make a deep impression on settlement patterns because, though a 'sizeable and vigorous' community, the 'existing population was too dense, the existing cultural, social, and economic structures too coherent, the new settlements too small to produce major change' .31 She concluded that although they altered the face of the area as new lords, the Scandinavians were absorbed as farmers and settlers. This raises questions about the nature and status of the land taken for settlement. Generally speaking, Scandinavian place-names are not attached to large or significant places (even J6rvik is only influenced by, not formed in, the Norse language). Stenton explained the absence of wholly Norse names from principal places in class terms: the leaders of the army took possession of the sites that had been the seats of great men, while the bulk of the soldiers made do with the country around. English names 24 Stenton 1927, 140; 1942, 312 (d. also I971, 52I); 'Nothing but a settlement on a scale which amounts to colonization could have introduced a mass of alien words and phrases into this familiar [farming and togographicalJ vocabulary' (Stenton 1942, 303). Sawyer 1957- 8, 8; 1998, 104-6; Cameron 1965. See also Wormald 1982C, 134- 7' 26 Lund 1981, esp. 167-8; see also Sawyer et al. I969, 204-5. 27 Works cited in n. 20 above. 2l! Sawyer 1957-8, 17; I998, 104; d. Lund I98I, 168-9; Williamson I993, 107-9. 29 Wormald 1982a, 147; see also Wormald 1982b; I982C, I 34-7; Gelling 1988, 220-1. 30 Lund I969, I99-200. 31 Stafford I985, 120-1. She contrasted this with the Isle of Man, where transformation was 'almost total'.

386

LAND, SEA AND HOME

survived in the greater estates, he suggested, because they were still farmed by English peasants. 32 According to Sawyer's first account of the settlement, however, this uneven distribution of Scandinavian place-names showed on the contrary that the Danish settlers did not take over existing English estates, but 'settled where they could, most often on land that the English had not yet occupied'. The English 'continued to live in their villages, paying rent or tribute to new lords, but otherwise little disturbed by the invaders.'33 Thus Sawyer explained not just why most major names remained English, but why Scandinavian place-names were frequently found in areas of less good land. In this interpretation (which he later revised), Sawyer characterized Scandinavians as colonizers not conquerors, and their place-names as representing land newly taken into use, not seized from native holders.34 He criticized Eilert Ekwall's suggestion that Scandinavian names replaced English ones in Lincolnshire because it was 'to some extent ... made necessary by the belief that the Danish colonization was a violent affair,.35 Others (Wormald, for example) have rejected the essentially peaceful settlement that underpinned Sawyer's original reconstruction and explained the place-name pattern quite differently, by the dynamics of conquest and the pressure of numbers - the Scandinavians were so many and the pressure on resources so great that the immigrants took good land and poorer land alike, but only the latter acquired new names. 36 The bulk of Scandinavian place-names, according to this view, did not mark new settlements but were renamings of previously occupied places (though Wormald allowed for the possibility that the less good land was unoccupied).37 Cameron's analysis of distribution and drift geology in the East Midlands suggested a hierarchy of Norse-influenced place-names, within which Grimston­ hybrids occupied attractive sites, typically in major river-valleys, by-names less attractive sites, typically along tributaries and smaller streams, and thorp-names more marginal land stil1. 38 Fellows-Jensen's study of these groups in both Yorkshire and the East Midlands encompassed a range of other factors that might be indicative of their early status (including the level of assessment in Domesday Book, the proportion of places that became parishes, and the proportion of names and settlements that survived the medieval period) and showed that by almost every test this hierarchy ­ hybrid tUn-names above by-names above thorp-names - was confirmed. 39 Originally both scholars interpreted the pattern in much the same way: the hybrids represented thriving English settlements taken over by the Vikings; the by-names implied the 32 Stenton 1927, 141; 1971, 525. 'The most remarkable feature of the English evidence [in contrast with Normandy] is the variety, which places the popular nature of the Danish settlement beyond doubt' and reveals 'the wholesale incorporation of Scandinavian loan-words into the local agricultural vocabulary' (Stenton 1945,

33~).

Sawyer 1957-8, 15-16. Sawyer later revised his view of the extent of land use at the period and substantially changed his interpretation, to encompass the alternative possibility 'that the Vikings were conquerors rather than colonists' (1978b, 7). See below, p. 387. Chris Morris favoured Sawyer's original model of settlement infilling, explicitly by a peasant population: see Morris 1977,98-9; 1981,228; 1984, 8-9. 35 Sawyer 1957-8, 15; Ekwall 1924,83-5. 36 Wormald 1982b, 163. 37 Lund 1981, 155-6; Wormald 1982b, 163_ 38 Above, n. 13. The separate studies were usefully summarized in Cameron 1985. 39 SSNY, 222- 36; SSNEM, 33 2 - 63­ 3

34

PLACE-NAMES AND HISTOR Y

387

exploitation of virgin or vacant land; the thorp-names marked a later phase of settlement expansion to still less attractive sites. Fellows-Jensen has modified her interpretation significantly over the years; in particular, her view of the by-names changed when she accepted arguments that many of them were on land unlikely to have been left untouched by the Anglo-Saxons. 40 The idea that Scandinavians spread peacefully into parts of England that the English had not reached depended, of course, on the view that a significant amount of land had indeed remained uncolonized in the last quarter of the 9th century, when the armies settled, or in the loth and I Ith, when, through a process of 'internal colonization', their descendants extended into areas of secondary settlement. 41 Glanville Jones, in his influential 1965 study of territorial organization in the North (and Sawyer, revising his previous interpretation), challenged this view. 'Many apparently irreconcilable elements in the arguments about Scandinavian settlement', Jones claimed, ' ... disappear if Scandinavian settlement is envisaged not as a widespread occupation of virgin territory or depopulated land, but rather as an adaptation of a pre-existing, and in large measure surviving, territorial organization. ,42 'There is no reason to suppose that there was much vacant land suitable for settlement when the Scandinavians arrived.'43 Finds of pottery and metalwork have been cited to support the existence of pre-Viking occupation in the vicinity of places with Scandinavian names. 44 Jones also pursued the observation that Scandinavian names seemed not to attach themselves to the centres of large estates, but only to their dependent parts or to small centres; this pattern arose, he argued, because the most powerful men (who would have had little contact with the property and its inhabitants) took the best estates, while lesser Scandinavians (Norse-speaking farmers) themselves lived and worked on the lesser lands, shaping their names in the process. 45 There would therefore have been even more Scandinavians on the land than the place-names represent, and the status of the holder, not merely his holding, could be reflected in the place-names. Hadley's observation that Scandinavian place-names in Derbyshire, Nottinghamshire, Lincolnshire, and Yorkshire were most commonly attached to lesser places, however, attributed this phenomenon less to settlement there by Scandinavians than to naming fashion and cultural change. She interpreted the retention of English names for significant places as potentially indicating a failure of the prestige which led otherwise to so many other aspects of Scandinavian identity

40 Compare her summary in SSNY (1972),250, with that in SSNEM (1978), 368-9. For further developments in Fellows-Jensen's thinking, on Grimsron-hybrids as well as by-names, see below, passim. 41 Sawyer et al. 1969, 171-2 and 203-7. 42 Jones 1965,71. Jones saw the smaller places as hamlets already in existence but 'slightly enlarged by the advent of new masters' (83). See also Jones 1990, 58-64. 43 Sawyer 1998, III; see also Sawyer 1978a, 161-3; d. Stafford 1985, 120-1 and Hadley 2000a, 23. 44 Sawyer, 1978a, 162. More recently, Kevin Leahy has noted that several settlements in Lincolnshire with place-names in -by had cemeteries of the 6th century; they were not necessarily continuously occupied until the 9th, however: see Leahy and Paterson 2001,183. Excavation at the Lincolnshire deserted village of Kettleby has uncovered substantial quantities of pottery ranging from Early Anglo-Saxon to late-medieval; see Everson et al. 1991,71 (see also 8-9). 45 Jones 1965, 83.

388

LAND, SEA AND HOME

being taken on by the English (personal names, linguistic habits, art-styles), or, alternatively, to the Scandinavian habit of adaptation and assimilation. 46 Behind all these questions lie the native inhabitants and the issue of their survival and relationship with the incomers. Although decades of war would have taken their toll, and many Anglo-Saxons would have fled or been killed, substantial numbers of Englishmen and women must have remained in regions which were subjected to Scandinavian authority. Interpretation of their situation in the past has generally depended on numbers - were the English inhabitants completely swamped, or not?47 While the linguistic evidence indicates significant contact between speakers of English and Norse, it is less forthcoming on the kind of society established in Scandinavian­ held territories. The two views that have dominated the debate have been based not on the nature of political relationships (which are overwhelmingly obscure) but on circumstances of landholding. Some, like Jones, have argued in favour of the continuing evolution of existing territorial organization in a landscape already divided into large estates with dependent parts, while others, such as Stenton, have concluded that the landscape was significantly refashioned by the Scandinavian settlers. Stenton was highly influenced by the picture of tenure he discovered in Domesday Book, where many areas which had been under Scandinavian authority in the 9th and roth centuries exhibited in the IIth a vastly higher number of free peasants (fiberi homines and sochemanni) than non-Scandinavian areas. 48 He consequently saw these peasants as the descendants of the armies of the 870S, who, he argued, had originally possessed and continued to enjoy a greater degree of freedom than their Anglo-Saxon contemporaries. Place-names which combined a Scandinavian personal name with the element by could therefore, in Stenton's view, represent these free peasants with their individual small parcels of land, and Allen Mawer similarly judged that Carlton­ names in Yorkshire 'happily illustrated' 'the qualities of freedom, equality, and independence which characterized the Danelaw,.49 Comparison with Normandy, where this type of name is absent or very rare, but where there are many place-names formed by Scandinavian personal names plus -ville, a local naming element (equivalent to personal name plus Old English -tUn), convinced Stenton that Normandy's settlement was 'largely aristocratic' and 'more individualistic' than England's, and that the native Frankish peasantry (unlike England's) was not supplemented by

Hadley 20oob, 124-5. For opposing views, see, for example, Wainwright 1962, 82, and Sawyer 1957-8, 16; later Sawyer (1998, 105-6) had modified his views to allow 'a substantial influx of Scandinavians, probably spread over several decades' (though this did not amount to 'immigration on a massive scale'). 48 Stenton 1910; 1969; 1971,515-19; see also Dodwell 1941. Liberi homines were mainly recorded in Norfolk and Suffolk's Domesday Book entries, while sochemanni were found in great quantity in those counties and Lincolnshire, as well as in Leicestershire, Nottinghamshire, Bedfordshire and Cambridgeshire. For a rough indication of distributions, see Darby and Maxwell 1962, esp. map p. 436; Darby and Terrett 1971, esp. map p. 436; Darby 1971, esp. map p. 361; and Darby and Campbell 1962, 591-5. As Darby explained (Darby and Campbell 1962,591-5), it is difficult to count sokemen in 1066 (as opposed to those recorded for 1086), but it is clear that there were even more at the earlier date. To complicate matters, David Roffe (1992, 19- 20) has pointed out that different counties' recording practices would have affected the totals. For the condition of sokemen in the lIth century, see Maitland 1897,66-79; Roffe 1990b; and Hart 1992, esp. 231-79. 49 Stenton 1910, esp. 45 (n.) and 91; 1927, 143- 6; 1942, 305-8; 1971,524; Mawer 1931- 2,16. 46 47

PLACE-NAMES AND HISTORY

389

further Scandinavian immigration. 50 Even Stenton, however, the champion of the free Scandinavian, allowed that by-names with personal names implied 'a primitive superiority' of the men named ('the leader rather than the lord of [the] inhabitants'), though he stopped short of awarding them ownership of the village lands. 51 Anne Kristensen, on the other hand, argued that many of the settlers were originally servile; she saw soke (and the names in -by) as relics of a system of military colonization ­ the centena - originally developed for the settlement of Germanic groups within the late Roman Empire. 52 In England, Stenton's free Danelaw peasants have been successively undermined by subsequent studies. In 1954, R. H. C. Davis sounded a warning note in his study of Bury St Edmunds, where sokelands proliferated in the IIth and 12th centuries: 'it is unwise to ascribe to [Danish] influence the origins of every institution that we cannot understand.'53 Support for the antiquity of soke-like estate organization included Jones's research on the institution of the multiple estate (arguing for a British origin), Sawyer's criticism of Stenton, and Geoffrey Barrow's, W. E. Kapelle's, and Hadley's respective studies of Scotland, the North of England, and the northern Danelaw, all of which criticized the attribution of sokes to outside influence. 54 The apparent restriction of sokeland to Scandinavian areas in England 'proves absolutely nothing', Kapelle has asserted, because these regions might have been distinctive before the arrival of the Danes. 55 Parallels to northern England can in fact be found in Kent, Wales, and Scotland as well as East Anglia. Furthermore, sokemen are not found in some areas with dense concentrations of Scandinavian place-names, such as Yorkshire, while areas well supplied with sokemen, such as East Anglia, have fewer such place­ names. Eric John suggested that the concentration of free peasants in the Danelaw could have derived from conditions prevailing in areas that had not been part of the great Mercian and West Saxon hegemonies of the 8th and 9th centuries, rather than from any greater Scandinavian respect for the individual. 56 Ros Faith has observed that 'the terms on which Scandinavian landholders took over land seems, in some way not yet understood, to have preserved many aspects of a comparatively free peasantry and an institutional form, the soke, which the Domesday commissioners recognized,.57 Hadley's study of Derbyshire, Lincolnshire, Nottinghamshire, and 50 Stenton 1945, 339- 44. He supported the social contrast with Normandy by citing the absence of Norse influence on native farming vocabulary . Current research in Normandy, however, appears to reveal a more substantial influence on agricultural, fishing, and shipping terminology in some parts of the region : see Ridel 2000 and Ride! forthcoming. 51 Stenton 1942,307; 1910, 91; d. 1971, 524-5. 52 Kristensen 1975. She offered three possible explanations for the appearance of the system in eastern England: it could have been indigenous, imported by the Scandinavian settlers, or imposed by the West Saxon conquest of the loth century. 53 Davis 1954, xliii; see also xxxii-xlvii, esp. xliii-xlvii, and Davis 1955. Cf. Maitland's famous warning, 'but in truth we must be careful how we use our Dane' (1897, 139). 54 Jones 1965; Sawyer 1957-8; Barrow 1973, ro-60; Kapelle 1979,62-85; Hadley 2000a, 122-40. Cf. Maitland 1897,67: 'we should be rash were we to find anything characteristically Scandinavian in the sokemen'. Fellows­ Jensen 2000, a review of Old English socn, similarly concluded that Scandinavian influence is far from proven. 55 Kapelle 1979,65 . Hadley, while supporting the importance of regional distinctions, has also warned against uncritical assumptions of continuity (20ooa, 26, 84- 93, and 140- 64); see also Hadley 1996. Unwin 1988 also questioned the continuity model. 56 John 1982, 164. 57 Faith 1997, 122.

390

LAND, SEA AND HOME

Yorkshire led her to conclude that the fragmentation of estates undermined seigneurial domination (both lay and ecclesiastical), contributing to the survival of a large number of free peasants, whose changing rights were reflected in the coining of place­ names with personal names. 58 Tom Williamson, on the other hand, has explained the distribution of sokemen in terms of climate and the logic of agricultural production. 59 As Stenton's free Danish peasants faded from the scene, a new tenurial picture was identified on the basis of place-names formed with Scandinavian personal names plus by. Jones associated these names with places occupied by the followers of the Danish army-leaders, who had received the dependent parts of the larger estates taken by their superiors. These men (later joined by their wives and families) were 'a small number of privileged newcomers who were allowed by their leader to impose a degree of intermediate authority over old established hamlets' .60 Regional studies, such as Harold Fox's investigation of the wolds in Lincolnshire, Leicestershire, Nottingham­ shire, and Yorkshire, have helped to sustain this view. Fox has shown how upland areas which had been used as pasture by lowland settlements became independent units some time between the 9th and the I Ith century.61 Developments of this sort, according to Jones, would have affected local nomenclature: 'The renaming of settlements is frequently likely to have taken place when the bonds linking the components of any multiple estate to its caput were loosened or severed.'62 Sawyer has since developed and elaborated the idea that place-names with personal names related to the break-up of large estates into smaller and independent units, each with its own lord - a man (or woman) whose name was then attached to by. He interpreted these names as established settlements which had once been, but ceased to be, dependent on estate centres; these newly fragmented parts were taken over by Scandinavians, the personal names identifying the first individual to have had full rights of ownership.63 Because Danes in Denmark rarely combined by with personal names in this way, and since this type of name is less common in parts of England that were held independently by Scandinavians for the shortest time - East Anglia and the South-East Midlands - Sawyer concluded that the 'manorialization' process that created the name-type could not have been early.64 Instead, he proposed that the many deaths sustained by Vikings in 10th-century battles against the West Saxon kings 'must have weakened the authority of the aristocracy in the Danelaw and Hadley 2oooa, esp. 180-96. Williamson 1993, 125: he argued that areas with lower summer rainfall were more reliable in terms of agricultural yield, which meant that farmers were less likely to fall into debt and 'abject dependence on local lords with superior resources'. 60 Jones 1965, 77; 'In return for military services and obligations, these followers were endowed with intermediate rights over the appendant hamlets of discrete estates' (p. 83); d. Stenton 1971,524-5. 61 Fox 1989, 90- 6. Fox, however (unlike Jones), saw many of the settlements in these upland areas as new sites (pp. 90-4)' 62 J ones 1976, 40. 63 Sawyer 1978a, 151-6 and 162-4; 1982, 104-6; 1994, 15-17; 1998, lI2-13. 64 Sawyer 1982, 104. Note that the term ' manorialization' can have a range of meanings; in some cases it is used to refer simply to the break-up of estates into smaller units with more independent holders, while in others it acts as shorthand for further extensive change involving settlement-nucleation and communal agricultural systems as well as the development of 'feudal' services (and thus decreased independence of tenants), over a period extending well beyond the Norman Conquest. See further below, esp. pp. 409-IO. For the concept of 'manorial' place-names, see Gelling 1988,180-6 and 230, and Fellows-Jensen 1994, 137-40. 58

59

PLACE-NAMES AND HISTOR Y

39 1

so given many small landowners a chance to escape from at least some of their obligations, and to claim fuller rights of ownership over their holdings'.65 According to Sawyer, therefore, the break-up of estates north and east of Watling Street was facilitated by the English advance, while the collapse of Scandinavian rule and the consequent redistribution of land by the English king further contributed to the process. 66 'The Scandinavian influence on English place-names may therefore be evidence not of expanding settlement but of changing ownership.,67 Hadley also concentrated on changes in land-holding: the arrival of the colonists was 'an important factor, but not because it introduced a stratum of free peasants, but because of the impact that it had on lordship and estate organization in the region';68 'the Scandinavian invasions disrupted prevailing patterns of lordship, spawning a very active land market in which estate organization became increasingly complex as did the stratification of the rural population' .69 Stafford considered that the number of Scandinavian names was 'artificially inflated' by this change in landscape organiza­ tion, especially in fertile lowlands where the break-up first took place. The chronology of this change would explain why there were so many by-names in Leicestershire's Wreake valley but fewer in Derbyshire's uplands ('manorialized' at a later date)?O Faith, on the other hand, argued that multiple estates survived later in the Danelaw, and that the greater independence of many Danelaw peasants in 1086 shows that manorialization, with its greater subjection of tenants to local lords' demands, was slower to develop there.71 Just as there have been many ways of interpreting the presence of Scandinavian place-names in some areas, so have there been many ways of explaining their absence from others. We shall return to the question of the absence of Scandinavian place­ names from areas where they might have been expected; but to conclude our review of research we offer the following summary of the range of explanations encountered. The areas in question were uninhabited. They were areas of dense occupation (where newcomers would have had less impact than in more sparsely settled regions). They were not settled by Scandinavians. They were not settled by enough Scandinavians. They were not settled for long. They were not settled for long before recovery by the English and the end of Scandinavian political independence. They were settled early, before Scandinavian naming-practices took hold. They were settled late, when Scandinavian speech was no longer so influential. They were settled by the Scandinavian elite and not by Scandinavian peasants (the latter, being close to the soil, would have had more influence on its naming). They were settled by Scandinavian Sawyer 1982, 104-6. Sawyer 1981, 128-9; 1998, II2- 15. 67 Sawyer 1982, 106-7; the quotation is from Sawyer 1978b, 7; d. Stafford 1985, 120-I. 68 Hadley 2000a, 26 and 167-215 (esp. 196, where she has pointed out that many place-names formed with personal names were not always attached to places that retained 'manorial' status). The quotation is from p. 21 5· 69 Hadley 1996, 12; d. Finberg 1974, 162-3. 70 Stafford 1985, II9. 71 Faith 1997, 121-3. Cf. Roffe 1990a, II, where it is pointed out that in northern Derbyshire, multiple estates persisted into the post-Conquest period, whereas in the south of the county fragmentation was 'rapid and all but complete' by 1086. Roffe has argued that fragmentation in upland zones was an especially protracted process (p. (4). 65 66

39 2

LAND, SEA AND HOME

peasants who assimilated more quickly than the rest. Areas without Scandinavian place-names were settled in the earliest period, before estates had broken up and acquired new names. They were settled later in the period, but in parts of the country where estate break-up was slow, and where changes provoking new names occurred late, with detached parts being independently identified only after the period when Scandinavian names were current. They represent parts of the country where the Norse language was given up quickly, or where recording of names was somehow retarded or slight. They belonged to the Church. They belonged to the king. A RECONSIDERATION OF THE PLACE-NAMES

In an effort to find order in what may seem to have become a free-for-all, we turn now to a re-assessment of the place-name evidence, especially the key group of by­ names. First, however, it will be helpful briefly to discuss some principles of place­ name study, principles which can on occasion be overlooked in the heat of historical hypothesizing. Above all, it should be stressed that place-name evidence is essentially linguistic: a 'Scandinavian name' is a name that shows the influence of Scandinavian language. This influence will not be restricted to villages in which Vikings settled, nor will it always come directly from the mouths of Scandinavians. Many Old Norse words entered English, and were, and are, used by English people. Thus Smith's famous map of Scandinavian parish-names is not, as it is labelled, an illustration of 'Scandinavian settlement'; rather it is fundamentally an index of the influence that Scandinavian language has had within England. This is related to the history of settlement, no doubt - after all, many names explicitly denote habitations, or cleared or cultivated land - but any simple equation of Norse linguistic traces with the homes of settlers from across the sea must always be resisted. Archaeologists have long since recognized that possession of a typically 'Anglo-Saxon' ornament proves nothing about an individual's ethnic or linguistic background; the appearance on the map of an etymologically Old Norse term is similarly limited in its implications. Secondly, we must acknowledge that place-name evidence is late. This is a general truth, since written records of place-names almost inevitably lag behind their coining; but in the east and north of Anglo-Saxon England the problem is particularly acute, because so very few documents of the pre-Conquest period survive. Again, this is clearly of great importance to an understanding of the material. Smith's map becomes a map, not just of linguistic influence, but of linguistic influence that has taken hold at some point before ro86, when most parish-names outside the far north of England are first recorded.72 If place-name evidence is to suggest anything about 72 A preliminary survey of the Norse-influenced place-names in pre-Conquest sources suggests that there are fewer than fifty (cf. SSNY, 236-7; SSNEM, 292-4; SSNNW, 336; our list excludes some uncertain cases from these studies, including a number of thorps, and adds a good number of East Anglian examples). Most of these are found only in post-Conquest copies of pre-Conquest documents. If one took the sceptical, but not unreasonable, line that scribes copying earlier documents, even documents genuine in terms of the transactions described, may have updated the place-name forms, that would leave fewer than twenty from lIth-century documents, and there seem to be only three from strictly contemporary loth-century sources - Derby, first on 'Circumscription Cross' coins of IEtheistan, 927-39 (Blunt 1974, nos 161-72 and P. 94) ; Conington, Huntingdonshire, in an original charter dated 957 (5. 649); and Melton Mowbray, Leicestershire, newly identified on 'First Hand' coins of IEtheired, c. 979- 85 (Blackburn 2000).

PLACE-NAMES AND HISTORY

393

the 9th and roth centuries, it must be recognized that interpretation will playa very large part: there are no more incontrovertible facts here than in the historical sources. Two centuries separate the initial settlements of the Great Army and the first appearance of most of the records: on the one hand, this means that the Scandinavian names found in ro86 might, in theory, have been coined throughout this period; on the other, it should be observed that the map represents their successful transmission and survival. Stenton realized that surviving names were not preserved by inertia,73 but subsequent neglect of the complexities of transmission has on occasion led historians astray. Seeking to discover political events in a sparsely documented era, they can forget that ephemeral activity will probably not influence place-names: Viking military occupation that is not succeeded by the settlement of Norse-speakers need leave no trace in the local toponymy, for example. And, as we shall discuss below, there are other reasons for names not being given or not surviving. Matthew Townend has recently pointed out that coining a place-name is not like registering a patent: its survival is not guaranteed and is subject to many factors. Since the Scandinavians have left us nothing in writing except a few inscriptions (and, later, some skaldic verse), place-names were recorded only in English sources, which were unlikely to have recorded Scandinavian names for places unless they had passed into usage in English - but there may well have been many names current in Norse­ speaking communities which were not preserved because they dropped out of use or failed to make it into writing at the right time. 74 Finally, and perhaps more controversially, it is tempting to propose as a principle, or rule of thumb, that no single place-name etymology should ever be trusted. This apparently radical statement is made in recognition of some of the fundamental problems that constantly attend place-name research. Our records of early spellings of names are never 'complete', any more than is our understanding of ancient stages of language; 'certainty' over an interpretation is inevitably relative, therefore, especially when we take into account the nature of names, with their potential for alteration and adaptation. York is the best example here: 75 ]6rvik, to a Norwegian who had never been there, might conjure up a picture of pigs gathered around a coastal creek; while Eoforwfc, from which J6rvik is directly derived, would certainly have sounded to an Anglo-Saxon like a combination of Old English eofor 'wild boar' and wfc 'farm'. Yet behind J6rvik and Eoforwfc, we happen to know, because we happen to have very early written records, lies Eburacon, a British name that may have meant 'place abounding in yew-trees'. And who knows what may lie behind the British name? Place-names are given to refashioning and reinterpretation?6 Thus any name, however transparent it may seem, needs to be treated with caution. Take a thoroughly Scandinavian name such as Aslackby, Lincolnshire. As it stands this is a compound of the Old Norse personal name Aslakr with Old Norse by Stenton 1927, 137. Townend 2000, 100-1. See also Townend 1998. 75 For a review of the history of the name, see Fellows-Jensen 1998. 76 Derby may be a similar instance: Old Norse djura-by(r) 'deer-settlement, deer-park' a recurring name in Britain - may well have been suggested by the ancient river-name Derwent (Old English Dearwente) running past the place: ON djurr would readily have been identified with English dear (cf. von Feilitzen 1937,68, §42). Fellows-Jensen (1991b, 347) made a similar suggestion. 73

74

394

LAND, SEA AND HOME

'settlement'. It does not take much philological imagination, however, to suppose that this might conceivably represent the remodelling of an Old English "·Qslae( es)­ byrig, containing the cognate Old English personal name Qslae, compounded with byrig, oblique form of Old English burh 'stronghold'. Though there is no hint of such an English original in the early spellings of Aslackby, there is evidence elsewhere for both types of substitution - of a Norse personal name-form for the English equivalent, and of by for the similar sounding byrig.77 There is no doubt that the etymology of this place-name can only be construed, on the basis of available evidence, as 'Aslak's settlement', yet it would nonetheless be inadvisable for an historian to base an argument on the 'fact' that a man called Aslakr lived in or owned the place. And, in principle, this type of argument might apply to any place-name, whether we can guess at a suitable underlying form or not. What prevents anarchy in place-name study is context. Sometimes topography provides solid context for a linguistic interpretation; very occasionally documentary evidence or archaeological discoveries offer clear support. Most commonly, however, context comes in the form of other place-names. Aslackby is one of more than seventy names in Lincolnshire that combine Old Norse personal names with by - presumably these are mostly 'genuine' examples, because it would be difficult to explain the kind of assimilation described above without assuming that the place-name type was locally established as a familiar model. It follows that the historian can be more confident in building arguments on the prevalence of this type of name in Lincolnshire than on the etymology of any individual example. In turning now to a reconsideration of the name-evidence, we shall therefore proceed on the premise that, while any single name may mislead, a strong pattern of names can be convincing. Of the three principal categories of place-name that have featured in historical discussions, the largest, and potentially the most significant for Scandinavian settlement, is surely the names in -by. The thorp-names have always been problematic, both because of uncertainty about the extent of influence from an Old English equivalent, and because - almost everyone has agreed - thorp denotes a secondary, dependent settlement, which by implication takes us at least one step away from the initial impact of the Scandinavian arrival. 78 The Grimston-hybrids, on the other hand, have certainly been linked with primary Viking land-taking, though recently some place-name scholars have begun to back away from an early dating of the whole class, and to point to examples that may derive from as late as the lIth century?9 The group does deserve re-examination, though preferably this will be undertaken in the context of an extensive survey of personal names combined with -tun in England, which is beyond the scope of this paper. 80 Meanwhile, we aim to review some of the fundamental 'facts' about the by-names in the east of the country, and to decide 77 An example of the former is Gumley, Leicestershire (with Old Norse Guomundr replacing Old English Godmund, attested in the 8th century); see SSNEM, 217. An example of the latter is Badby, Northamptonshire, ultimately an instance of the recurring English place-name (ret) Baddan byrig; see SSNEM, 293. Note, incidentally, that compound personal names in English place-names can be found with or without genitival ending, as shown by Fellows-Jensen 1975,449-50. 78 Cameron 1970; Lund 1976; Fellows-Jensen 1991-2. 79 Insley 1999,55-6; Fellows-Jensen 200la, 285-6. 80 Cf. also the comments below, p. 423.

PLACE-NAMES AND HISTORY

395

whether they offer, in our opinion - and despite all the disagreement and difficulty set out above - any unambiguous information to the historian. By is an Old Norse term for 'settlement', meaning - as far as we can see ­ anything from a farm to a village;81 it is rarely possible to deduce from the place­ names themselves much about the precise nature of the settlement-sites at the time the names were coined. 82 Nonetheless the very large number of by-names in parts of the country provides a substantial dataset for linguistic and geographical analysis (d. Fig. 2). Both Cameron and Fellows-Jensen have presented major studies of the East Midlands; Fellows-Jensen has also investigated Yorkshire. Table I presents a summary of their linguistic analyses and quantification of the by-names. Though the two scholars have based quite different historical interpretations on their material ­ interpretations that may be open to question - there is no doubt that their linguistic surveys are detailed and philologically sound: they will give a secure basis to a discussion of the name-type. It will be clear from the table, however, that there are some marked differences in the East Midlands figures produced by the two scholars. Happily, these discrepancies are easily explained. 83 To a philologist examining the first elements of the names, some are clearly Old Norse in form (e.g. austr 'east' in Austby), some clearly Old English (e.g. dryge 'dry' in Driby), and a rather large number could belong to either, since Old Norse and Old English were closely related languages. When faced with this ambiguity, Cameron has tended to choose the Old Norse alternative, on the not unreasonable grounds that - however the counting is conducted - the weight of evidence tends to support a presumption of 'Norseness'. Fellows-Jensen, however, has remained more circumspect, allowing a larger number to stand in the ambiguous category. Hence Cameron's significantly higher overall proportion of Old Norse elements (Table I, A). Table I, B shows an even larger discrepancy in the percentage of East Midlands by-names thought to involve a personal name as first element: Cameron has 68%, Fellows-Jensen just 40%. This is also readily explained. Many personal names are practically identical in form with common nouns. ON skegg 'beard', for instance, can be the base of a nickname Skeggr or Skeggi, 'the bearded one'; alternatively it might be applied to 'something jutting out, a promontory,.84 Topography is sometimes, but not always, decisive. Faced with a choice, Cameron generally accepted the personal names, while Fellows-Jensen let the

For general discussion of the implication of by in England and Scandinavia, see Parsons and Styles 2000, IO-II, and SSNNW, IO-II. Note that by is properly the Old Danish form of the word; it is here preferred over the 'standard' Old Norse, or Old West Norse, byr, which had a nominative inflexional -r that is never represented in English place-names. 82 Interesting cases include instances of Owmby and Aunsby, in Lincolnshire, and a lost Ounesbi in the North Riding of Yorkshire: all may have meant 'deserted farm' (Old Norse auon 'waste'); see SSNEM, 33- 4, though the interpretation was not accepted by Cameron 1998, 7, 95-6. Note also repeated instances of Kir(k)by, generally interpreted as an (existing, English) 'settlement with a church': see below, n. 205. 83 One simple difference, which accounts for the difference in overall totals, is that Cameron's 'territory of the Five Boroughs' included only Nottinghamshire, Derbyshire, Leicestershire and Lincolnshire, while Fellows­ Jensen's East Midlands added Rutland and Northamptonshire (and also considered a handful of Scandinavian names from eastern Staffordshire and Warwickshire) . 8' SSNEM, 69. 81

104-5; SSNEM,

396

LAND, SEA AND HOME

FIGURE 2.

The distribution of by-names, recorded by

1086,

in England.

uncertainty stand. 85 Her 40% is a minimum figure, as if none of the ambiguous cases were personal names. She calculated the maximum, if they were all taken as personal names, as 69%, very like Cameron's figure. The truth, she observed, probably lies somewhere in between. 86 So we can account for the varying figures in the East Midlands. For our purposes Fellows-Jensen's lower numbers are perhaps rather more objective. For Yorkshire we have only Fellows-Jensen's statistics, but it should be noted that she did this survey before the East Midland one, and at that time she counted rather more like Cameron, S5 She discussed the problem at length in Fellows-Jensen I975. Cf. the summary review in Fellows-Jensen I995, I79, conceding that '[tJhese revised interpretations have not all won general acceptance'. S6 SSNEM, 27.

PLACE-NAMES AND HISTORY

397

TABLE I

ANALYSIS OF THE FIRST ELEMENTS IN THE BY-NAMES OF EASTERN ENGLAND

East Midlands Cameron

Yorkshire

FellowsJensen

Fellows­ Jensen

303

333

210

% combined with ON % combined with OE

87 10

65 16

78 7

Ratio of ON : OE first elements

90: 10

81 : 19

92: 8

68

40

57

93 7

83 11

91 6

93: 7

88: 12

94: 6

A Total of by-names

B % of total involving personal names Of these: % of ON personal names % of OE personal names Ratio of ON : OE personal names

at least with regard to the personal names: she has since recalculated her 'minimum percentage' of personal names in Yorkshire as 45%,87 which would bring a corresponding slight downward shift in the overall percentages and ratios. Yet such a shift would clearly not in the least affect the general thrust of the figures, which can be summarized in two principal observations: first, in both Yorkshire and the East Midlands, place-names in -by are over four times as often combined with identifiably Old Norse elements as with identifiably Old English ones; second, very roughly half of the by-names contain personal names (usually Old Norse personal names) as first element. These are the striking patterns that emerge from the linguistic analysis of by­ names in these regions. The question now is whether they have any significance for an understanding of Scandinavian settlement in the Danelaw. To begin with the point that, where the languages can be distinguished, Norse appears in combination with by far more frequently than does English. The obvious inference to be drawn from this - and we are far from the first to draw it - is that .7 Fellows-Jensen 1975,447. She gave 42 % in Fellows-Jensen 1991b, 348. In any case, most of the alternative interpretations involve Old Norse nouns, and so the balance between languages would be little affected.

39 8

LAND, SEA AND HOME

by-names arose in a predominantly Norse-speaking environment. 88 This is of interest given the suggestion that Scandinavian place-names may be a symptom less of immigrant Scandinavians than of a mixed 'Anglo-Scandinavian' society, in which the local English language has already been heavily marked by Norse loan-words such as by. These figures appear strongly to imply otherwise: by-names in general would seem to be given by speakers of a Scandinavian language; a minimum ratio of 4:r Old Norse to Old English first elements does not suggest Old English as the base language. 89 Various indications tend to support this observation. In a small proportion of the names, typically Scandinavian inflexional endings are fossilized in the formations, coming near to proof that they were given by Norse-speakers. 90 And if the overall ratio of Norse to English elements is impressive, the ratio of Norse to English personal names is overwhelming. It might, of course, be objected here that Danelaw personal names are not simple markers either of ethnicity or of language: there are some well documented cases of children English by birth given Scandinavian names, under the influence, presumably, of a Scandinavian cultural ascendancy.91 Yet comparison with the names of pre-Conquest landholders in Domesday Book suggests that in the by­ names we are dealing with something of a different order. The ratio of Norse to English names among Lincolnshire landholders in the mid-lIth century is around 62:38; in Yorkshire it is even higher, at 70:30. Yet this is still far below the 94:6 ratio detected by both Cameron and Fellows-Jensen amongst the by-names. In the East Midland counties the average, in Domesday Book, is around 50% for each language, as against the 90% Norse of the by-names. Even given the uncertainties of fashion, it seems unlikely that the native Anglo-Saxon name-stock amongst landholders would fall to around ro% during an Anglo-Scandinavian roth century, but rise again to 50% by the mid-lIth. It therefore does not seem likely that by-names arose among an average, mixed local population open to the influence of Norse personal-naming­ habits. Instead the conclusion, again, would seem to be that in general by-names are distinctively Scandinavian, and were given in a Scandinavian, rather than an Anglo­ Scandinavian, linguistic context. 92 It is important, of course, to stress 'in general'. Several instances of the name Denby - 'the by of the Danes', with the stressed vowel of Old English Dene rather

88 Cameron I965, no; SSNY, I2, 242 and SSNEM, 278. Cf. Sawyer I97I, I73: 'The expansion of settlement marked by the names in by was undertaken at an early stage before there had been much assimilation between the colonists and their English neighbours.' 89 If as Cameron not unreasonably assumed (see above) - most of the ambiguous cases were counted as Old Norse rather than Old English, the Old Norse ratio would of course be still higher. Note, incidentally, that there is no doubt that English and Norse did in effect mix together in many local dialects (for one interpretation of how this process may have come about see Hines I99I); our argument is that the by-names do not generally illustrate, and are presumably earlier than, this development. 90 For the details, see Cameron I965, II9; SSNY, 239-40 and SSNEM, 27I-4. A sensible and cautious summary of the evidence (privileging the -ar genitives above the more questionable traces of Norse -s rather than English -es) is given by Fellows-Jensen I994, IH. The significance of this material has been accepted by Sawyer I998, 109. 91 See, e.g., Page 197I, I92. 92 For the details of this analysis of Domesday personal names see Parsons 2002, esp. 44-5. Fellows-Jensen reached the same conclusion, on the same grounds, in her study of Yorkshire: SSNY, 243.

PLACE-NAMES AND HISTOR Y

399

than Old Norse Danir - suggest that by was in some areas adopted into English. 93 Yet the meaningful compounding of an English word with by in this way is rare, perhaps extremely rare. Fellows-Jensen and Cameron have both suggested that by­ names with English first elements are usually alterations of originally English place­ names. 94 If such remodelling was indeed widespread, the low figures we have for English first elements - around 10% - would be a distinct overestimate as a guide to the language spoken by the coiners of the names. 95 Similarly, while it is true that some names wholly Scandinavian in form might represent the reshaping of Old English predecessors, it would be far-fetched to suppose that many did so, for (as has already been suggested above) such names presumably only arose by assimilation to a Scandinavian name-type that had already become locally established. Moreover, any such assimilation can also be interpreted as evidence for the influence of Scandinavian­ speakers. Thus we must conclude that the figures for the by-names in Yorkshire and the East Midlands establish a clear pattern, a pattern which implies that the names do not merely owe their ultimate inspiration to Scandinavian language, but were very predominantly given in Scandinavian language. The second notable point about our statistics was the high number of by­ names - around half of the total- with personal names as first element. It may be significant that the ratio is higher than the proportion of personal names combined with Old English -tUn in place-names in most parts of England, and, as has been noted, very much higher than the proportion of personal names in by-names in Denmark. 96 What it must imply in the present context, taken with our previous conclusion, is that a considerable amount of land in England had become associated with individuals in Old Norse-speaking areas. Given, moreover, that these lands are mappable, this is valuable information, though it leaves unanswered many of the historian's more subtle questions - why, how and when were lands passing to Norse­ speaking tenants or owners? On the question of date, if the argument so far is sound, we can place most of the by-names in a period when Scandinavian language was being spoken in England. Unfortunately it is not easy to arrive at a date for this period either. In Yorkshire a small number of by-names involve Continental-type personal names, suggesting that the term might still have been productive as late as the second half of the 11th 93 SSNY, 25; SSNEM, 44. Evidence for the word's adoption is also found in three Middle English texts of Midland or Northern provenance (Kurath et al. 1954-2001, s.v. bl n.). It may be attested also in the 10th­ century Old English gloss to the Lindisfarne Gospels, where hus vel lytelo by 'house or little by' glosses domicilium 'dwelling'. Technically this instance of by could represent a native Old English cognate of the Norse term, though the chaotic possibility that it could be this Old English word in many of the place-names is practically excluded by the enormous percentage of Old Norse first elements. Cf. Hofmann 1955, §247; Hines 1991,412-13; Parsons and Styles 2000, 104. 94 Cameron 1985, 134; Fellows-Jensen 1984, 33; d. 'For the East Midlands as a whole it can be said that the topographical and geological evidence suggests that hybrid names are normally borne by English vilis that were taken over a nd partly renamed by the Viking settlers' (SSNEM, 327). 95 It might also be noted that a r emarkable proportion of the English first elements appears in the two commonly recurring names Willoughby and Ashby. Whatever the explanation for the popularity of these hybrids (d. the discussion in Fellows-Jensen 1984, 33), their suspicious frequency suggests that it would be wise not to count them all as instances of the free combination of English elements with by. This would again tend to lower the percentage of significant English first elements still further. 96 For preliminary figures for -tun in many counties see SSNEM, 282- 5. Fellows-Jensen (1975,446) reported a figure of 10% personal names in by-names in Denmark.

400

LAND, SEA AND HOME

century.97 Whether this implies that Old Norse was being spoken at this date east of the Pennines (as it probably was west of them 98 ), whether by had by then entered local English dialect, or whether the element had merely survived in the onomasticon, so that names were coined by analogy with existing, older by-names, is hard to say (though see further below). In the East Midlands, however, there are reasons to think that by did not generally remain productive as late as the Norman Conquest. 99 Indeed, the conquest of this area by the English before 920 might be taken to suggest that Scandinavian-speakers would not have been in a position to claim farm-land here much after this time. This is a doubtful line of reasoning, however, because political control of an area and the length of time that a language survives there are not necessarily linked - there will, for instance, have been Scandinavians who came to amicable agreements with the West Saxons. lOO On the other hand, the most promising linguistic criterion for dating the by-names that has so far been identified does tend to suggest a relatively early date - albeit a very imprecise one - for the whole of eastern England, Yorkshire included. It can be shown that the short personal names based on the name-element Thor-, such as Toki, Toli and Tosti, are extremely common in lIth-century and later records from Lincolnshire and Yorkshire. Such personal names are very rarely compounded with by, however, which tends to suggest that most of the by-names were coined before they were in fashion. lOl Other internal evidence for dating is hard to find. One datable linguistic feature appears to be the reduction of the personal name element -ketilt to -kelt after c. IOOO, and Fellows-Jensen has observed that place-names involving a contracted form (e.g. Asselby, East Riding of Yorkshire, with Askelt rather than Asketilt) presumably did not take their recorded form until the IIth century. She suggested, however, that to date the coining of a by-name to the IIth century on this basis 'would be to jump to too hasty a conclusion', arguing that the name of a later tenant may have replaced that of an earlier, or that a personal name in -ketilt in an early place-name might have been reduced to -kelt when the contracted forms became prevalent. 102 This evidence does not, therefore, seem sufficient to prove the late formation of these by-names/o3 though it may offer an important indication that Norse language was still alive in England into the IIth century - especially in Yorkshire, where contracted forms seem to be particularly common. 104 97 This argument should not be pressed too far, since a good many (non-Scandinavian) Continental names are arrested in Anglo-Saxon England: d. von Feilitzen 1937,39; Smart 1986, 174-7. 9. Parsons 2001, 30 S. 99 Only two of 333 Domesday by-names involve Continental-type personal names (SSNEM, 18), and of the twenty-one by-names first recorded between lISO and ISOO (place-names which could - though need not­ have been coined after 1086) none do (SSNEM, 29). 100 See below, pp. 412 and 420. 101 SSNY, 241-2 and SSNEM, 277-8; d. Hald 1969, 186. This evidence was accepted by Sawyer 1998, 109-10. 102 Fellows-Jensen I99Ia, I 12-2I; 1994, 13S-6. The quotation is from p. 136 of the latter. 103 Pace Sawyer 1994, 17-18, who suggested that this demonstrated that 'the process of fragmentation [of estates] continued long after the original settlement' (p. 18). 104 Fellows-Jensen (1994, 136) noted that in settlement names four of them by-names - the -kell forms seem to be restricted to Yorkshire, though the precise significance of this for the survival of the language is not entirely clear, since in independent use the contracted forms are found elsewhere - for numerous Lincolnshire examples of Thorkell, for example, see Fellows-Jensen 1968, 309-10. In East Anglia a notable scarcity of contracted forms might suggest that Norse did not survive there until the lIth century - d. Insley 1979, S6-7.

PLACE-NAMES AND HISTORY

401

It is sometimes observed that the wide range of Norse personal names found combined with by is striking, and that many of them are absent from the extensive record of Scandinavian personal names in Domesday Book. It has been argued that the names in Domesday Book represent lIth-century fashions and that the by-names are significantly different, and presumably therefore earlier. lOs This argument is difficult to maintain, however, when one considers the huge range of personal names borne by 12th- and 13th-century inhabitants of Yorkshire and Lincolnshire. No fewer than half of these names are not found in Domesday Book either, demonstrating that the pool of Norse personal names was much deeper than Domesday Book alone can indicate. I06 Absence from it is not a guarantee of early date. I O? One non-linguistic pattern that has borne repeated scrutiny is the attachment of by-names to relatively low-status places, less substantial in general than local places with English names in -tUn, and situated in relatively marginal areas, away from the major centres of power. lOS This is an important finding for two reasons. First, the identification of shared non-linguistic features tends to support the linguistic indication that the by-names in Yorkshire and the East Midlands form a more or less coherent group; it tends to confirm that the potentially chaotic effects of linguistic analogy, discussed above, have not been great. Second, the observation that by-names often occupy rather inferior sites suggests that they are unlikely on the whole to represent either the takeover of thriving English villages or the spoils seized by members of a (small, elite) conquering army. So far this re-examination has not raised the question of numbers. The principal conclusions drawn - that by-names belong to a Norse, rather than an Anglo­ Scandinavian, linguistic context, and that they imply a large element of individual land-holding - involve in themselves no assumptions about numbers. One way, the conventional way, of interpreting name evidence would suggest that place-names in the early medieval period are descriptive labels that arise among the local population in the normal course of conversation. This would tend to imply a good number of Norse-speakers in areas where by-names are common. Another possibility, however, is that names are in some way imposed by a governing class, as the result, perhaps, of a bureaucratic requirement. According to this model, relatively few, high-status, Norse-speakers might dictate the naming of settlements: clusters of by-names might then indicate their areas of influence, but need not suggest hordes of Scandinavian E.g. Cameron 1971, 160-L This argument is set out in more detail in Parsons 2002, 31-3 . As observed there (33, n. 14), it is not excluded that a more complex method than simple comparison with Domesday Book may yet produce valuable indications along these lines. In his 1940S Reading Ph.D. thesis, which we have not seen, F. W. Wainwright carried out an 'examination of the personal names combined with by . .. designed to test their antiquity in so far as this is revealed by their occurrence independently in later times and by their occurrence with other and presumably later place-name elements' (Wainwright 1962,78-9). The details of this examination do not seem to have been published, and we have not been able to evaluate Wainwright's conclusion that the names generally belong to the 'first phase of settlement and that most of them arose at a date which cannot be far removed from A.D. 900' (ibid., 79). 107 This was a conclusion reached also by Fellows-Jensen: SSNY, 241-3; 1984, 32- 3; cf. also Sawyer et al. 1969,205. Elsewhere, Fellows-Jensen (1995, 174-5; SSNEM, 277) has, however, allowed some force to the old arliiument, especially in connexion with Grimston-hybrid names. 10 See above, p. 386. Exceptions to the 'rule' are discussed below, pp. 421-2. 105 106

402

LAND, SEA AND HOME

settlers. In outline, therefore, the pattern of Norse village-names is not in itself decisive on the question of numbers. The point made above, that the by-names do not much look like the work of a small conquering elite, does nevertheless begin to raise the issue of numbers: if the hundreds of Norse-named men in the by-names are not the conquerors, then who are they? Those who remain sceptical of large numbers of settlers from Scandinavia will probably turn here to the argument that people with Norse names could be ethnically English. We have already, in effect, argued against this view: the overall association of Norse names and words with the element by suggests Norse-speakers. Besides, if the by-names themselves are not admissible, then it becomes difficult to find any evidence, in the 'marginal' areas where they cluster, for the Norse-speakers who are apparently exerting such a powerful influence on local nomenclature and language. There is some reason, therefore, to suspect on the basis of the by-names that there were Scandinavian settlers beyond a small military elite. It is two other types of name-evidence that make this case most urgently, however. Field-names are sometimes marginalized in this discussion, for it is claimed that they have little to offer on the question of settlement. 109 They tend to be first recorded much later than the 'major' names, usually no earlier than the 12th century. And unlike town- and village-names, most of which we know to have survived for a thousand years, the minor names - not just of fields, but of streams, marshes, copses and banks - tend on the whole to have much shorter lifetimes, of perhaps a few generations. Yet, just like major names, field-names do offer very good evidence for linguistic influences. Indeed, because they tend to be short-lived, that evidence is rather more closely datable than the major names. Thus we can deduce that in some parts of Scandinavian England - most strikingly, in the north of Lincolnshire - Norse influence on the local vocabulary of the 12th to 15th centuries was huge. 110 And it gains something of a special significance for being recorded amongst field-names, for they presumably reflect the local vocabulary of the farming classes. In northern Lincolnshire in particular, the quantity and range of this Norse vocabulary surely indicates that somewhere behind the Anglo-Scandinavian mixture of 12th-century records there existed an earlier Norse-speaking farming population. Minority elite influence might conceivably introduce to the indigenous lower classes legal and administrative terminology (including the word by) and new fashions in personal names, but it is hard to imagine how - at least without Norse-language education, literacy, bureaucracy or Church - it would entice local farmers to use Norse words in

Lund 1981, 156- 69; Sawyer 1982, 102; cf. Fellows-Jensen 1974, 54. Cameron 1996. Fellows-Jensen 1974 analysed a 14th-century document containing names from nine midland counties; the article also surveyed, and gave references to, earlier work on Lincolnshire, Leicestershire and Lancashire. See also Insley 1985, and cf. below, nn. 159,173. 109 110

PLACE-NAMES AND HISTORY

40 3

describing furlongs, streams, hollows and hillocks.1l1 No parish in the country shows a comparable effect achieved by any French-speaking lord of the manor, and this is presumably indicative of a quite different sociolinguistic situation. ll2 Where this field­ name evidence is found in the same area as a concentration of by-names, as it is in northern Lincolnshire, it seems likely that the Norse-speaking community implied by the field-names is directly related to the Norse-speaking community implied by the by-names. As Cameron has stated, such a weight of linguistic influence 'cannot have come from thin air' .113 Personal names also point firmly in the direction of large numbers. Norse personal names became popular all over England in the I nh century, and they remained so in parts of the Danelaw for centuries. In itself this is inconclusive: personal names introduced at the Norman Conquest came utterly to dominate from around the 13th century onwards, and we know that they were introduced by an elite minority. The difference is the extraordinary variety and vitality of Norse naming in England. Although nearly everyone came to bear new names introduced from the Continent, these names are derived from a very small stock: William, John, Robert; Alice, Christine, Maud, etcY4 Likewise, there were some common and fashionable Norse names (some of them (re-)introduced by the Normans); examples include Sveinn, Thorkell and Asgeirr. But in addition we know of hundreds of linguistically Norse names attested in the Danelaw, many of them found only once or twice, some of them paralleled in Scandinavia, some of them probably Norse-language coinages this side of the North Sea. 11S Again, it is difficult to see how a model of elite dominance could explain this situation. English peasants might honour the names of their Scandinavian leaders, but how would they develop a vital, distinctively Norse, naming tradition? Communities of native Norse-speakers again seem a preferable solution. HISTORICAL DISCUSSION: (I) THE SIGNIFICANCE OF THE BY-NAMES

In our opinion, therefore, any historical assessment of the Scandinavian settlement should accept (or produce new arguments to counter) the following points, on the basis of the place-name and associated linguistic evidence: 111 Many of Lund's criticisms (1981, 156-69) were telling, though he understated the range of Scandinavian vocabulary in some concentrated areas (not that all of them had been studied when he was writing); cf. below n. I I 3. He was quite right to stress the impossibility of calculating precise numbers from the linguistic evidence. See also Sawyer's cogent comments in Sawyer et al. 1969, 204-5. Lund's model of not-so-e1ite dominance (Scandinavian leaders who had much more to do with the farming peasantry than did their Norman successors), and his hypothetical total of some 5000 settlers, are intriguing. Whether the 'large' number required by the names could equate to a huge a rmy (plus hangers-on?) settling as 'gentlemen farmers', whether it could reAect a larger, second generation of Norse-speakers colonizing from within, or whether it implies a secondary migration of peasants, is not easy to say on the basis of linguistic evidence alone. 112 Gelling 1988, 23 8. 1\3 Cameron 1996,13. The quantity and range of vocabulary in some localized areas ought to put paid to the notion that a general dialectal drift across eastern England can account for all Norse linguistic material in the Middle Ages (cf. Lund 1981, 167). To demonstrate the unevenness of spread most effectively, it would be desirable - as far as the rather uneven records allow - to extend the kind of study pursued by Cameron in part of northern Lincolnshire across a much wider area, such as the whole of the county. For comparable localized 'hot-spots' of Norse linguistic inAuence in the personal name record, see Parsons 2002, esp. 46-51' 114 Reaney and Wilson 1995, xxii-xxiii; Clark 1992, 551-63, esp. 553. 115 Fellows-Jensen 1968; Insley 1994.

40 4

LAND, SEA AND HOME

I) the by-names were in general coined by speakers of Old Norse; 2) there were sizeable communities of Norse-speakers in parts of eastern England; 3) the by-names imply a significant role for Scandinavian-speaking individuals in the holding of land in parts of England; 4) many of these land-holdings can be characterized as relatively marginal or low in status; 5) there is some reason to suspect that the majority of by-names were coined before the I nh century. 116

This information seems to us to be the 'bottom line', produced by a cautious analysis of the names. The points made here are all reflected somewhere among the opinions surveyed above, and many of the detailed historical reconstructions can sit comfortably alongside them; they do, after all, allow considerable leeway. Both colonization (the exploitation of undeveloped land) and manorialization (the fragmentation of old multiple estates into more independent parts) are models that could be compatible with the place-name evidence for settlements associated with individual named Scandinavians; if - as it presently seems - there are good reasons to prefer estate-fragmentation to colonization, then this is established by evidence independent of the place-names themselves. On the other hand, the present conclusions conflict directly with some of the historical interpretations reviewed above - especially concerning the number of settlers or their linguistic (and in that sense, presumably, 'ethnic') identity. One noteworthy argument that has been made by both Sawyer and Fellows­ Jensen goes beyond our conclusions. As recorded above, the two scholars have both ascribed the by-names with Old Norse personal names to a model of estate­ fragmentation; they have in addition felt able to date this development to the 10th rather than the 9th century. Their dating argument is based principally on place-name distributions which on inspection do not appear to support their case. Sawyer first proposed in 1981 and 1982 that it was possible to date the fragmentation of Danelaw estates to the loth century, observing that a large Derbyshire estate (Hope and Ashford) recorded as sold to an Englishman some time between 899 and 9II (and thus presumably in Scandinavian hands for no more than 25 or 30 years) later showed little Norse influence on the names of its many parts. 117 Furthermore, he stated that 'Scandinavian settlement names are rare in the parts of the Danelaw that were recovered by the English soon after 900' and claimed in particular that Cambridgeshire lacked Scandinavian place-names because in 903 the

116 To the argument from the short form of Thor- names, made above p. 400, might tentatively be added the observation that five by-names, apart from Derby, are recorded in potentially loth-century sources (SSNY, 237, and SSNEM, 293; there would be more if the Historia de sancto Cuthberto were included - see below, p. 408 and nn. 130, 160) . Although these are all preserved in later copies, and any individual spelling might be open to question (above, n. 72), cumulatively they do at least tend to confirm that by-names are not as a class an I Ith-century type. 117 Sawyer 1981, 129; 1982, 103 (with reference to S. 397 [A.D. 926J, JEtheistan's confirmation of the land; the sub-units of the estate are first recorded in Domesday Book).

PLACE-NAMES AND HISTORY

40 5

Danes 'were driven from' land they had occupied there. lls In 1998 he repeated this argument and added a further point: since places named after individuals were 'almost unknown in Denmark at that time' it was to be 'doubted whether the Scandinavian settlers named places after individuals in the early days of their colonization,.119 This last observation seems dubious: it is true that by-names are relatively rarely combined with personal names in Denmark; however, the new circumstances of a conquered and colonized country could surely have allowed new models of land-holding, and of place-naming, from the beginning - we can hardly assume that the newcomers would not have been able to think of this type of name during the first few decades of their residence. More promising is the first point, that land quickly recaptured by the English has very few Norse names compared to land that remained in Scandinavian hands for longer. It is unfortunate, however, that Sawyer gave as a principal example the expulsion of the Danes from Cambridgeshire in 903, for although the Anglo­ Saxon Chronicle tells us that Edward the Elder harried Danish-held land there in that year, it also makes clear that the battle at the end of the incursion was won by the Danes, who appear to have remained in control of the area. The Danish army of Cambridge is said to have eventually surrendered to Edward in 917, at the same time as those of Northampton and East Anglia, and only a year before Mercia is said, by the same source, to have submitted to Edward's rule. 120 This does not leave much time for a substantial difference to emerge between Cambridgeshire, with very few Norse names, and areas such as Leicestershire and central Northamptonshire, where by-names are common. In its own terms, therefore, a major part of Sawyer's argument is strained. More worryingly, the whole case rests squarely on the assumption that time is the principal variable in the equation. Sawyer did not in this context consider the possibility, for instance, that the difference between areas might owe something to the density or nature of the initial Scandinavian settlements. His dating argument could only work if one assumed that, c. 900, before any English victories, Scandinavian settlement was spread evenly across the Danelaw. If, on the other hand, it were possible that - while the whole territory was under the political control of Danish armies - some areas had received many settlers and other areas few or none, this also might explain the difference in place-name patterns. If Cambridgeshire had been regarded as a border region, it might have been less attractive to settlers than the areas further behind the front lines. Similarly, the Scandinavians' willingness to sell the Derbyshire estate of Hope and Ashford might indicate that the land was surplus to the settlers'

118 Sawyer 1982, 103; d. 1981, 128- 9. In Sawyer et al. 1969, 171-2, he had suggested that an extension of settlement in Cambridgeshire might have taken place at a relatively late date, when Old Norse was no longer much spoken. 119 Sawyer 1998, II I. For completeness note also a further point made ibid. : 'if a significant number of names had been given in the ninth century, when Scandinavian speech was still vigorous, many more examples of the use of Scandinavian inflections might have been expected, despite the fact that almost all names were first recorded in late texts written by scribes who naturally tended co "modernize" them'. The second part of the sentence effectively counters the first. Moreover, since most Old Norse and Old English inflections were similar, it might be considered remarkable that so much evidence of distinctively Norse forms survives. 120 ASC s.a. 917, 918. The submission of Leicester co J£thelfla:d is noted under 918 in the Mercian Register.

406

LAND, SEA AND HOME

requirements. We do not know the specific circumstances of these places. 121 But to assume that - if left alone for another generation - they would inevitably have developed concentrations of by-names just as parts of Leicestershire and Lincolnshire did is clearly unsafe, yet this assumption is at the heart of Sawyer's argument. Fellows-jensen echoed Sawyer's argument regarding the early capture of some areas, but added to it some observations on the distribution of different types of by­ names, both in eastern England and in the North-West. She suggested that the names incorporating English terms, common nouns ('appellatives'), adjectives and adverbs were likely to belong to a slightly earlier date than those incorporating personal namesY2 She supported this argument partly by the observation that the earliest settlers from Denmark were not used to the personal-name type, but largely by an appeal to the distribution map: in eastern England the former type is said to be relatively common in 'Suffolk, Derbyshire, the West Riding of Yorkshire, Nor­ thamptonshire, Nottinghamshire and Leicestershire', 123 with the implication that several of these counties were among those most quickly and effectively conquered by Edward the Elder or iEthelfl::ed. Although it should be noted that there are numerous by-names combined with Norse personal names in both Leicestershire and the West Riding of Yorkshire, which rather complicate the picture, Fellows-jensen's claims for distribution are nonetheless of great interest, especially the suggestion that the different types of name tend to imply different types of settlement (she favoured the takeover of English settlements for the appellatives and the break-up of estates for the personal names).124 Yet, once more, it is not clear that the distribution pattern must have chronological implications: the circumstances of settlement, of land-taking and land-holding, may have varied from area to area; could it not be this, rather than date, that is principally reflected on the map? Moving to the west of the Pennines, Fellows-jensen identified a marked and interesting distribution pattern along the same lines. In Cumberland and Dumfries­ shire by-names frequently incorporate personal names, as in much of eastern England; to the south, in Lancashire, Cheshire and Westmorland, however, by-names are very rarely of this type. Again, she has chosen to explain this chronologically - since, she argued, Lancashire and Cheshire were swiftly 'recovered' by the English, and Westmorland by the Britons of Strathclyde, the fragmentation of estates implied by by-names involving personal names had not yet taken place in these counties; further north, where the Scandinavians remained in control for longer, estates were

121 It is notable that Fellows-Jensen chose these two regions as examples when observing that topographical considerations (height on the one hand, and marsh on the other) might help 'to account for the relative absence of Scandinavian names from north-western Derbyshire and much of south-eastern Lincolnshire and Cambridgeshire. These areas must all have been unattractive for settlement in the Viking period' (199Ib, 338). 122 Initially she placed the earlier type in the last quarter of the 9th century (1983a, 54-6), but she later preferred the beginning of the 10th century (1984, 35-6); d. also the clear summary statement putting all the names after 900 in Fellows-Jensen 1994, 134. Note, incidentally, that in SSNNW, 24, she proposed that by­ names with adjectives, in the North-West at least, were later, not earlier, than the personal-name type. 123 Fellows-Jensen 1984,36. 124 It might be observed that her attribution of all the by-names to after 900 (see above, n. 122) seems to leave a notable, and apparently unexplained, gap between the arrival of the Vikings and the coining of any names, even for the English places taken over. Cf. below, pp. 408- 9.

PLACE-NAMES AND HISTORY

40 7

fragmented as in the east. 12S Once more, while this may have been the case, it does not seem to follow inevitably from the distribution. The north-western area is clearly different, in aspects of its place-names and its history, from the East, and direct comparison between them may be dangerous. Scandinavian activity in the North­ West is almost unrelentingly undocumented; but the sketchy historical record of a Scandinavian settlement in the Wirral peninsula, for instance, can be contrasted on several grounds with the sketchy record of the Great Army's divisions of land: it was later (after 900), involved a grant of land from the English authorities (in the person of lEthelfla:d) who remained in power, and the Scandinavian settlers came not from the homelands or the Continent but from Ireland. 126 Any of these factors might have played a part in differentiating the names here from those elsewhere. Similarly, the situation in Cumberland and Dumfriesshire is hardly straightforward: despite the high proportion of personal names in by-names, very few of those personal names are Old Norse, and it is evident that many of the place-names were coined (or altered) in the post-Conquest period. 127 Fellows-J ensen has built a cogent chronological structure on the by-name distributions in the British Isles, and her interpretations deserve careful consideration at every point. However, there are a number of uncertainties in her extended hypothesis - which also involves an unrecorded early migration of Danes across the Pennines, thence to the Isle of Man, and then back to take part in the colonization of the Wirral. 128 More fundamentally, there are obvious problems in using a single model to explain circumstances in a range of geographically and historically distinct regions. It is doubtful whether the distribution of names in the North-West can help date the by-names in eastern England. In our opinion, therefore, a loth-century date for the coining of by-names is not required by the place-name evidence. Sawyer proposed that the instabilities of the early 10th century had precipitated major upheavals in landholding which were reflected in the place-names, but that period of turbulence and disturbance had of course been preceded by another. As has been emphasized, the initial phase of Scandinavian conquest is extremely obscure, and the circumstances of the distribution of land in the 870S and 880s are undocumented. The Anglo-Saxon Chronicle nevertheless refers to sharing out (gedd!ian) , ploughing (erian) , cultivating (tiiian) , and settling (gesettan). King Alfred's biographer Asser adds little more: in 876 'Halfdan, king of one part of the Northumbrians, shared out the whole province between himself and his men, and together with his army cultivated the land'.129 No SSNNW, 21, 413-14; Fellows·Jensen 1983a, 54-5. Radner 1978, 166-73; Wainwright 1975, 131-61. 127 Other scholars have disagreed with Fellows-jensen's interpretation of these place-names as reflecting Viking-age Scandinavian coinages with the names of later owners substituted (1983a, 56- 8; SSNNW, 22-4) : d . Insley 1986, 172; Roberts 1989- 90, 34-5; Phythian-Adams 1996, 149-50. 128 For explanation and elaboration of this hypothesis, see Fellows-jensen 1983b; 1983a, 54-9; 1997, 82-4; SSNNW, 310,411-14. Margaret Gelling has taken a quite different view of the place-name evidence for the Norse settlement of Man, and a slightly different view of the evidence for Cheshire (Gelling 1991, 149-50; 1992,132-4; 1995; d. also jesch 2000, 3-4). Note that Fellows-jensen (200Ib) has recently radically revised her interpretation of much of the Manx evidence, suggesting that a number of the names may be 11th-century imports from England. 12 'Halfdene, rex illius partis Northanhymbrorum, totam regionem sibimet et suis diuisit et illam cum suo exercitu coluit': Stevenson 1959, ch . 50, p. 38; also in Keynes and Lapidge 1983,83. For the Old English texts of the Anglo-Saxon Chronicle see Dumville et al. 1983- . 125

126

408

LAND, SEA AND HOME

contemporary account is more expansive. Most informative is the Historia de sancto Cuthberto, a 12th-century narrative which preserves older traditions about the fate of the lands held by the episcopal community of St Cuthbert, originally resident at Lindisfarne and, later, at Durham. The Historia's account might support the assumption that seizure of land was the first step of the successful invader, though in this case the invaders do in fact belong to the early loth century, because the context is not Halfdan's activities but the takeover of the northern kingdom by Ragnall and his Hiberno-Scandinavian army c. 918. The Historia is hagiographical, retrospective (the relevant sections were written in the mid-roth or the IIth century),130 and unreliable. But while its motives may have more to do with demonstrating the community's rights to the estates at the time of writing, there is nothing implausible about its description of a seizure of the community's land by Ragnall, who is said to have granted two large tracts to his followers Onalafball and Scula; after the battle of Corbridge (in which he was victorious), Ragnall seized two further estates that the community had leased to a certain Eadred, who had died in the battle. 131 Furthermore, the Historia also claims that some decades before, the community had bought land from Guthred, the Danish king who had been raised to power with their assistance c. 883, 'and from the Danish army under him which had divided the land among themselves' .132 The land purchased appears to have previously belonged to other ecclesiastical communities (such as Jarrow, Wearmouth, and Hexham), and was presumably appropriated from them by Guthred or, before him, by Halfdan and his followers, before being sold on to the only surviving ecclesiastical power in the region.133 Loss of ecclesiastical endowments at this time presumably helps to account for the circumstance that, compared with the rest of England, by the time of Domesday Book a much higher proportion of land in Danelaw regions (with the exception of East Anglia) was in secular, not ecclesiastical, hands.134 Land held by laymen is likely to have suffered a similar fate: the Historia claims that Ragnall seized the land of Ealdred, of the native Bamburgh dynasty, and drove him into exile.135 It is probable, therefore, that 'settlement', as recorded in the Chronicle, meant more or less immediate seizure of land. What happened thereafter must have varied, some estates being regranted or sold to native holders and the rest divided, as individual followers of the successful army-leaders were rewarded and settled. Sawyer proposed, however, that the by-names, created by the fragmentation of estates and establishment of 'private ownership', 136 did not occur until after 900. Any suggestion 130 Craster 1954, 177- 8; Simpson 1989; Johnson South 2002, 27-36. Craster and Simpson favoured a date of c. 945, whereas more recently Johnson South has raised the possibility of an lIth-century context for this material. 131 Historia, chs 23-4 (Johnson South 2002, 60-3). 132 'Ethred supradictus abbas emit a praefato rege Guthred et a Danorum exercitu qui sibi sub eo terram diuiserant has uillas': Historia, ch. 19a (Johnson South 2002, 58-9). 133 There were further purchases (from unspecified sellers) by Bishop Cutheard (A.D. 899-c. 915): Historia, ch. 21 (Johnson South 2002, 58-61). 134 Fleming 1985, 249-50; see also Dumville 1992, 37-8 and 53. 135 Historia, ch. 22 (Johnson South 2002, 60-1). Morris (1981, 228) has suggested that the Scandinavian place­ names around the Skerne Valley in county Durham, where such names are otherwise uncommon, may represent an area of royal land taken over en bloc and later corresponding to the (Norse-named) wapentake of Sadberge. 136 Sawyer 1998,112- 13.

PLACE-NAMES AND HISTORY

40 9

of a pause between conquest and break-up of estates, and between this fragmentation and the beginnings of 'manorialization' or 'private ownership', needs to offer reasons for such a delay. The origins of manorialization - the more systematic exploitation of the landscape, with the development of more regular renders and labour services - are too complex for full discussion hereY7 The fragmentation of 'multiple' estates into separate manors is linked to the intensive exploitation of seigneurial or 'manorial' rights and the development of nucleated settlements and associated field-systems. Smaller land-units which had once belonged to large complex estates and were tied to their centres by services and payments increasingly came to be held by individuals who had acquired the kinds of rights over them originally only enjoyed by the larger lords. To prosper, the holders of these smaller units had to reorganize both physical and human resources significantly. Cause and effect are difficult to disentangle, and the relationship of the changes to the Scandinavian settlement remains unresolved.138 The date of the break-up of large estates has been much debated, but it was clearly a continuing process, and not simply linked to political takeover in regions settled by Scandinavians, as manorialization occurred throughout England (and much of Europe).139 While Sawyer recognized that the process had begun before the Vikings arrived, he emphasized that their settlement 'had facilitated' it and that 'the main fragmentation of estates occurred during and after the period of Danish rule' .140 He believed that disbanding army members were given land that initially remained tied to their leaders. The transfer of seigneurial rights to the lesser men followed, but mostly without the consent of those in power, who held the large estates; they relinquished seigneurial rights only because they were weakened by deaths and other setbacks delivered by their English opponents in the South-East Midlands after 900. The West Saxon conquest of Lincolnshire had the same effect, he argued, allowing lesser landholders to acquire greater rights: and it was only when lesser holders began to acquire these rights that the 'manorial' names in -by - identifying the newly independent holder - developed. Estates that resisted fragmentation - such as the endowment of the abbey of Bardney - retained their English names even under Scandinavian secular lordship. HI The conditions of tenure in Scandinavian-held territory at this time are deeply obscure, however, and Sawyer's reconstruction remains surmise. There is, for instance, no evidence that any free Danes were unable to hold land in their own right in the 9th century - indeed, the treaty between Alfred and Guthrum implies that a substantial proportion of Danes were (at least in West Saxon eyes) of equivalent status to West Saxon thegns and therefore potentially eligible to be independent

\37 See above, n. 64, on the various uses of the term. The literature on the subject is vast. A classic study is Aston 1958. Recent relevant publications include Dyer 2.002., 2.6-35, and Hooke 1998, 105-38. 138 See Brown and Foard 1998,82.-92., for an interpretation of Northamptonshire in this light. 139 A variety of da tes have been proposed for manorialization in England, ranging from the 7th to the end of the lIth century and beyond; see Unwin 1988,78-80, for a brief summary of recent arguments. I~ Sawyer 1994, 17; 1982., 106-7; 1998, 112.-13. 141 Sawyer 1998, 98- 100.

410

LAND, SEA AND HOME

holders of land.142 Moreover, as Sawyer readily admitted, entrepreneurial fragmenta­ tion from below was not the only way these smaller estates could be created: outside the Danelaw royal grants performed the same function. It has been argued that the deliberate organization of land-holding is first evident on royal vills, and that it was connected with the evolution of state structures. 143 The arrival of Scandinavian settlers - especially where numbers were great ­ must have produced pressure on existing arrangements. Continuing ecclesiastical and royal ownership could have had the effect of retarding the process of manorialization on some large estates, but this cannot be assumed to have been significant - or consistent - across the conquered regions. l44 In particular, we cannot tell whether Scandinavian leaders had the means (or the desire) to retain large blocks of territory intact; it seems more likely that, at the initial stages of settlement, when church and royal land became available, it would have been distributed among their followers. It is likely that this liberation of ecclesiastical endowments in conquered areas (especially south of the Humber) and the removal of native royal controls had a significant and invigorating effect on the land market. 145 Hadley's study has stressed the importance of an active land market in disrupting prevailing patterns of lordship, and especially in contributing to the closer association of individuals with smaller estates. l46 As settlement characterized by smaller units became entrenched, it doubtless generated further change; alienation, sale, and partition for inheritance would have accelerated the process of fragmentation further. In our view there is at least as good a historical case to be made for the start of this process in the last quarter of the 9th century as in the first third of the 10th. 147 Finally, in discussing fragmentation, Sawyer drew attention to major estates ('sokes'), later in royal hands, that seem to have kept control over their constituent parts for a longer period, such as Horncastle and Caistor in Lincolnshire.148 Although there are numerous by-names on these estates, almost none are formed with personal names. In contrast, other royal sokes, such as Grantham, exhibit a quite different pattern, having numerous place-names with personal names. Sawyer cited Horncastle and Caistor as examples of unfragmented estates that resisted manorialization; the contrast with Grantham emphasizes, however, that the Scandinavian impact on the Danelaw is likely to have varied from place to place and estate to estate, even within 142 The passage states that the wergild of a Danish freedman was equivalent to that of an Anglo-Saxon ceorl, implying that the rest of the Danes had the higher (thegnly) wergild: Liebermann 1903-16, I, 126; Keynes and Lapidge 1983, 171. See Dumville 1992, 1-27, for a revision of the traditional date of 886 x 890 for the treaty. 14 Hooke 1998, 115; Faith 1997, 153-77. 144 For documented ecclesiastical survivals, restricted to Northumbria, see below, P. 413; on the loss of eccIesiasticalland, see Dumville 1992, 32-3. 145 Sawyer 1971,174; Hart 1992, 146-7. 146 Hadley 1996; 200oa, 158; cf. Finberg 1974, 162- 3. The land market could also operate on a large scale: King iEthelstan is said to have bought the region of Agemundesnes in 934 and granted it to the archbishop of York (S. 407, for Amounderness, Lancashire). The personal name Agmundr suggests a Scandinavian seller (just possibly a known Danish hold of the early loth century: Fellows-Jensen 1990, esp. 24-5). 147 Not that we claim to have 'solved' the problem. There is certainly room for other hypotheses, such as that the imposition of West Saxon rule, with its literate bureaucracy and efficient taxation, might have provided the impetus for the coining of new names for these new units. In that case we would have to wait until after 918 in the southern Danelaw and the reign of iEtheistan (or Edmund, or Eadred, or Edgar) for the creation of these names in the North. 148 Sawyer 1994, 18; d. Bardney, above, p. 409.

PLACE-NAMES AND HISTORY

4I I

the same county. Whether or not an estate remained intact might have depended on a range of factors: the nature of previous ownership, the quality of the land, the survival and degree of co-operation of former English holders and the conditions of their surrender, the use made of the land by the conquering West Saxon kings, the speed of recovery of the Church and the subsequent accumulation of endowments by new or revived ecclesiastical establishments. 149 Any overall interpretation of the place-names needs to take into account the potential variety of reactions to both Scandinavian colonization and West Saxon conquest. Sawyer's observations on these estates should be pursued, but the variety they reveal runs contrary to the assumption of an even spread of Scandinavian influence on which his dating argument is based. HISTORICAL DISCUSSION: (II) THE SIGNIFICANCE OF ABSENCE

In the course of the discussion so far, attention has several times been drawn to arguments based on the absence of Scandinavian place-names from an area. This topic is a surprisingly important one. One of our own principal arguments has been that, though precise dates and circumstances are elusive, groups of by-names on the map of eastern England should be taken as positive evidence for the presence of Old Norse-speakers in the Viking Age. It remains to question whether the absence of by­ names (or other Norse names) from other areas should be held to indicate an absence of Scandinavians. The short answer is that it should not. As has been noted, the coining of names, and their transmission and survival, involves a series of variables. Combined with the lateness of the records and the problems with chronology it is always difficult to interpret silences in the evidence. Scandinavians may have settled densely in some areas, given up their language early, and not affected the place-names. On the West Saxon conquest some land-holders may have been driven from their homes, and their newly-coined names may not have survived. Scandinavian place­ names may conceivably have been resisted by some local populations and under­ recorded by some local bureaucracies. In some parts Scandinavian settlers, though present, may not have been as successful at acquiring ownership of land as their fellows elsewhere in the Danelaw. And so on. We have criticized Sawyer's dating argument for by-names because it assumed that time was the principal variable on the distribution map; for argument's sake, we suggested instead that there may have been more settlers where there are more names. It should be stressed, however, that this is merely one of numerous possibilities, and that our criticism is focused on the assumption that the distribution should be interpreted in a single way. Absence of evidence on the place-name map is certainly not evidence of absence. That is not to say that the absences do not merit closer consideration. Take, for example, the 'embarrassing tendency' of burhs and other main settlements in Scandinavian areas to have English names.150 Derby, a rare exception to this rule, 149 See below, pp.412- 13. Note also the case of Rutland: Cox (1989) argued that there was a significant absence of Scandinavian settlement names from the county, and that this was an ancient land unit which for some reason was able to resist colonization by Danish incomers. 150 Kapelle 1979,66. Not all the Scandinavian burhs are attested as places of importance before 900, however: see Stafford 1985, III and 114.

412

LAND, SEA AND HOME

nonetheless lacks Scandinavian names in the immediate VICllllty. Fellows-Jensen explained this by suggesting that the Danish occupation was limited to the burh and did not extend to the countryside nearby. She suggested that the Vikings of Derby kept together for protection, 'instead of spreading out throughout the country as their compatriots did in Lincolnshire' .151 Sawyer hypothesized that a similar preponderance of English names around Lincoln indicated that the district was kept intact by the Scandinavian lord controlling the burh.152 There are potentially other explanations. Although central places such as these are most likely to have seen serious disruption when native kings and ealdormen were killed and power structures overturned, some form of continuity of administration might have favoured the retention of names there both after the Scandinavian takeover and upon conquest (and the revival of recording) by the West Saxons. So might have the continuing presence of a substantial English population: pressure to retain names in bilingual areas might have been greater than elsewhere. Similarly, the best lands - and the largest estates - were most likely to be caught up in the instabilities produced by the sequence of conquests. This would have been most acute in the North, where political power continued to see-saw throughout the 920S to 950S, as the West Saxons gained and lost and regained control. West Saxon conquest did not necessarily mean dispossession of all landholders, however. Continuities may be implied by the Anglo-Saxon Chronicle's statement that when Edward the Elder repaired the burh at Nottingham in 918, he manned it with both Englishmen and Danes. A legal dispute recorded in an early 12th-century history of the abbey of Ely suggests that some Danes had been allowed to keep their lands after submitting to Edward. ls3 No charters from this period of Edward's reign have been preserved, but the Scandinavian duces witnessing those of his successor iEtheistan from 928 to 935 could have included some survivors from the once independent regimes of Southumbria; however, no ministri (thegns) with Scandinav­ ian names are recorded as having been in attendance on the king. Edward and his successors required a reliable aristocracy for the newly conquered territories, and the installation of English officials, especially ealdormen, would have demanded signific­ ant landed resources - acquired at the expense of existing holders. Hart has suggested that properties surrounding the military base at Colchester, for example, were sequestered from their Danish holders by Edward and used to endow the ealdordom of Essex. 1S4 David Dumville has compared Edward with the better documented King Edmund, who in the 940S rewarded his followers with grants of Fellows-Jensen 198 1, 141. Sawyer 1998, 102. Leahy and Patterson 2001,185, suggested that a concentration of early Old English ham­ names around Lincoln contrasted with the rest of the county, where 'these early place-names are likely to have been lost in the re-naming of fragmented estates'. 153 The case implies that the king would have confirmed the tenure of those who submitted to him when, as the Chronicle tells us, the armies of Northampton and Cambridge 'chose him especially as [their] lord and protector' (ASC s.a. 917): Blake 1962, 98-9. Hart 1992, 16, contrasted estates in Northamptonshire and Cambridgeshire with those in Huntingdonshire and Bedfordshire, attributing the more fragmented state of the former to continuity of Scandinavian ownership, but there is no evident distinction between the Chronicle's record of the surrender of the armies of Bedford, Northampton and Cambridge, and of 'all the people of that district [of Huntingdon] who had survived' (ASC s.a. 914, 915 and 917). 154 Hart I992, 134- 5. 151

152

PLACE-NAMES AND HISTORY

41 3

land recently recovered from the Vikings of YorkYs Prime estates would probably have been most vulnerable, and Scandinavian place-names least likely to have been preserved there. On the other hand, possession of poorer land by settlers and their descendants should have been more secure, and tenurial continuity of this sort could account for a better survival rate for the names of lesser estates. The only estates that are known to have remained in English hands during Scandinavian rule were held by the Church, and it has been argued that continuing ecclesiastical ownership retarded or impeded the formation of new names. Unfortu­ nately, our near-total ignorance of ecclesiastical business from the mid-9th to the mid­ 10th century makes it difficult to test the idea that ecclesiastical lands resisted Scandinavian impact - and Scandinavian settlers - most successfully. Many (some might say most) of the substantial churches in the Danelaw evidently went out of business, some forever, some only temporarily. As all dioceses south of the Humber ceased to function for almost a century (and monastic churches either followed suit or went silent), the evidence is limited to the North, where the community of St Cuthbert and the archbishopric of York survived and made their peace with the new powers in York. Yet even here the absence of sources is debilitating to any argument: the history of the archbishop's endowment cannot be written, as its composition before, during, and after the Viking wars is unknown. Sawyer made the interesting observation that in Domesday Book the archiepiscopal estates in Yorkshire's North Riding had a lower percentage of Scandinavian names than their neighboursY6 However, as charters show, the archbishop acquired many estates in the course of the loth century, and Domesday Book is an unreliable guide to lands held before the end of Scandinavian rule in 954. The context of his sokes and berewicks named in charters and in Domesday Book - and of their place-names - is therefore uncertain: they could have been in secular hands for 80 years or more before coming into ecclesiastical possession. In any event, almost all the names we have in loth- and lIth-century charters are at some remove from their original contexts, appearing as they do in late medieval manuscripts. 157 Under these circumstances, absence of Scandinavian names cannot necessarily be taken to suggest the survival of ecclesiastical lordship. Thanks to the Historia de sancto Cuthberto, as we have seen, we are better informed about the Viking-age endowment of the community of St Cuthbert. 158 It lay mainly in county Durham, an area characterized by relatively few Scandinavian place­ names. The lands identified in the Historia as having been caught up in the takeover 155 Dumville 1992, 38 and 152- 3; Sawyer (1975, 34-9) argued that grants in Derbyshire to Wulfsige the Black by King Edmund (S. 479, 484, and 1606; A.D. 942) represented 'a stage in the recovery' of the Five Boroughs and 'one of the means whereby Edmund reasserted English royal authority in central Mercia' after Olaf Guthfrithsson's successes there; cf. also Whitelock 1959. 156 Sawyer 1982, 103-4. This contrasts with the many Scandinavian place-names in Amounderness (Lancashire), granted to the archbishop of York by King 1Ethelstan in 934. If the archbishop ever took possession of this very large (and distant) estate, however, his church had lost it by 1066. See Sawyer 1981, 129; Wainwright 1975, 181-279. 157 Such as Southwell's boundary-clauses and extensive list of dependencies in S. 659 (A.D. 956), preserved in York's Magnum registrum album (s.xiv); surviving pre-Conquest texts include the archbishop of York's statement regarding church lands in Northumbria (S. 1453; A. D. 972 X 992), which occurs in a manuscript of C.1030, and a survey of archiepiscopal estates in Yorkshire added at the end of the York Gospels (S. 1461a; c. A.D. 1020). 158 Craster 1954; Morris 1977; 1981; Aird 1998,9-59; Johnson South 2002.

414

LAND, SEA AND HOME

of power and seized by incoming Vikings generally bear English names, as does the rest of the known endowment; only a handful of by-names or Grimston-hybrids exist among the latter. 159 On the other hand, it may be significant that two smaller estates - which might have been in secular lordship until the late 10th or early IIth century when they were granted to the church by laymen, Styr, son of Ulf, and Snaculf, son of Cytel- each included a by_name. 16o Another absence of names that has caused much discussion concerns more tangible remains. The lack of correlation noted by W. G. Collingwood (among others) between Scandinavian place-names and so-called Anglo-Scandinavian sculp­ ture - stone crosses with strong Scandinavian affinities in their design and iconography - was taken up by Sawyer in defence of his settlement chronology.161 Sawyer argued that the sculpture, which is mainly found in centres with English names, was to be identified with sites of early conquest and settlement, before Scandinavian naming patterns were established. It followed that the absence of sculpture from villages with by-names confirmed their later date of foundation. In 1977, however, Chris Morris warned that dates for the monuments were far from secure, and that there was a danger of circularity: the identification of places with such sculpture as existing villages taken over by Scandinavians in the last quarter of the 9th century 'is based on a discussion which in fact uses as one premise the belief that such sculpture represents the work of the men who took over English villages in the ninth century!,162 Richard Bailey likewise queried the early date attributed to the sculpture in Northumbria and further undermined its direct association with an immigrant population (of whatever date) by pointing out that it did not necessarily represent actual Scandinavian patrons, but rather, 'the tastes of the new political masters of York', a point picked up by Hadley.163 Whoever commissioned and built the sculpture, Morris argued that it would have been erected at pre-existing churches. l64 Bailey pointed out that these monuments would naturally concentrate in centres previously settled (and named) by the English, as these - unlike the majority of places with Scandinavian names - would have been on good land which had produced (and continued to produce) the wealth required to erect them. It follows from this that villages with by-names, whatever the ethnic composition or cultural inclinations of their inhabitants, lacked the wealth to produce sculpture. 165 The cultural complexity of the sculpture was discussed by Morris (and further pursued by

Watts 1988-9. A recent study of two townships near Hartlepool has indicated a correspondingly restricted of Old Norse influence on minor names (Watts 2002). 1 Historia, chs 29-30 (Johnson South 2002, 66-7). It is, however, also notable that Girsby (Grisebi) is located across the boundary in the North Riding of Yorkshire, where by-names are much more plentiful. The other name, Northmannabi, is unlocated, but apparently in the Darlington area (Johnson South 2002, 136) and possibly therefore also across the southern boundary of county Durham . 161 Collingwood 1908, 120-1; also observed by Alan Binns 1956, 5; Sawyer 1971, 163-6; Sawyer et al. 1969, 20l- 6; Fellows-Jensen, SSNY, II8- 19 and 218-21. 16 Morris 1977,99-100. Cf. Wilson 1976, 399-400. 163 Bailey 1980, 210-II; d. Hadley 1997,91-3; 2oooa, 315- 18; 2ooob, 131-3; 2001, 19-20; 2002, 62- 5' 164 Morris 1977, 99- 100. 165 Bailey 1980,213-14. 159

d~ree

PLACE-NAMES AND HISTORY

41 5

Hadley); he argued that it indicated 'an integration of Scandinavian taste with the native tradition, and not an overwhelming of one by the other'.166 Clearly there are obstacles to the easy application of this body of evidence to place-name issues (among others). Non-experts, unable to distinguish between 9th­ and roth-century sculpture for themselves, have had to face the problem that art historians can assign different dates to the same object, as demonstrated, for example, by the controversy surrounding the Middleton Cross.167 Much Anglo-Scandinavian sculpture is currently attributed to the loth century, in some regions even to the second half.168 But art historians' date-bands are often too broad (viz., 'roth-century') to be historically helpful, or are overly dependent on cross-references to a few (possibly irrelevant) dates from the written sources. While chronologies based on art styles can be refined (aided by dendrochronological dates, especially in Scandina­ via),169 Jim Lang warned that 'it is unjust to the sculpture [in Yorkshire] to see the carvings as an insular reflection of mainstream Scandinavian art. The closer the scrutiny of the stones, the less applicable become the Viking art style labels of Borre, Mammen, and Jellinge,.17o This will clearly complicate any effort to date the English material. Most helpfully, recent studies have observed and identified regional groupings and have emphasized that 'Anglo-Scandinavian sculpture' is not a single entity to which one, consistent, explanation could apply, but consists potentially of a variety of local habits with different dates and meanings.l7l Therefore, any attempt to correlate sculpture with place-names needs to be done on a local basis. THE CASE OF FLEGG • . . AND SOME CONCLUSIONS

Finally, we should like to think about an area where absence is crucial: East Anglia, where the place-names exhibit a somewhat different pattern from those of other areas settled by Scandinavians (Fig. 3). In brief, there are fewer names, more spread out, compared with the denser concentrations elsewhere. In Suffolk they are quite thin on the ground: most are hybrids and there are only three or four names in -by (in contrast to the 234 in Lincolnshire). Norfolk preserves rather more Scandinavian place-names than its neighbour, and these are generally scattered across the county, though there is one noteworthy cluster of place-names in -by on the 'island' of Flegg in the south-east of the county. Overall, compared with the rest of the Danelaw, the place-names of East Anglia have been interpreted as indicating less intense Scandinavian settlement, generally limited to inferior land, possibly topped up by later immigration once the armies of the 890S disbanded.172 Other evidence, however, appears to counter this impression. In Norfolk, for example, Scandinavian personal names and field-names occur in greater numbers and Morris 1981,233; see also Morris 1984,9-13, and Hadley (works cited in n. 163). Bailey 1980, 209-13 (see also p. 53); Morris 1977, 100. 168 See the various volumes of the British Academy's Corpus of Anglo-Saxon Stone Sculpture (Cramp 1984- ). 169 Bailey 1980, 54-8; Roesdahl 1991, 173-7. \70 Lang 197 8 , 20. \ 7 \ See, for example, Stocker 2000 and Sidebottom 2000. 172 Ekwall 1924, 81-3; Davis 1955, 30; Stenton 1945,343; Williamson 1993, 107; Insley 1999, 53 suggested that 'the Scandinavian settlement in East Anglia seems to have been largely aristocratic and seignorial'. 166 167

416

LAND, SEA AND HOME

Key names in -by

Norse personal name + -thorp

other wholly Norse names

'Grimston-hybrid'

other hybrid names

Norse-influenced hundred names

English name denoting Danes

FIGURE

3.

Scandinavian place-names in East Anglia, aiming to show all names recorded by plus further parish and hundred names recorded before c. 1300.

1086,

PLACE-NAMES AND HISTOR Y

41 7

over a wider area than might be expected from the indications of the major place­ names. 173 That there was at the least a major political and institutional disturbance is indicated by a hiatus in monastic life and the disappearance of the kingdom's two dioceses.174 Both the Church's collapse and the tenurial particularities which distinguish the region have in the past been attributed to Scandinavian influence, though not without criticism. 175 More recently, significant amounts of 10th-century metalwork of modest quality - either imported from or imitating styles current in the Scandinavian homelands - have been found in Norfolk, thanks to co-operation with metal-detectorists. Although native English men and women could of course have owned these objects, the new finds may support the presence of ordinary people with Scandinavian connexions, widely distributed across the county.176 Contemporary written sources shed little light on East Anglia's Scandinavian period, beyond the bare record in the Anglo-Saxon Chronicle of its defeat in 869, settlement in 880, and submission to Edward the Elder in 917. Chosen as a winter refuge by the Great Army in 866, it is frequently named in the Anglo-Saxon Chronicle as the source of armies that ravaged across England from the 880s to 917.177 However, it is not unlikely that the coasts of East Anglia had seen Viking fleets, at least in transit if not in action, at an earlier date, once raids began in the 790S.178 No East Anglian sources survive, and we are dependent on Wessex, preoccupied with its own kings and battles, for our historical narrative. But the absence of recorded activity in East Anglia in no sense implies that there was none. Its 9th and 10th centuries are dark indeed: East Anglia's kings from 794 to c. 855, for example, are unrecorded in written sources, and we know nothing of the stages which led to the defeat of the kingdom in 869 and the sharing out of the land in 880. Any assessment of the kingdom's reaction to the Viking wars is, therefore, conjectural. Aggression from the powerful neighbour to the west in the 820S is strongly implied by the deaths of several Mercian kings on campaign in East Anglia. 179 If these assaults were matched by Viking activity along the coast, the kingdom may have been hard pressed. In similar circumstances, when external raids and internal conflict threatened them on several fronts, Frankish rulers adopted a practical policy of turning poachers into gamekeepers. 18o In 841, for example, King Lothar ceded the island of Walcheren, in Frisia, to the Viking Harald, 'to secure the services' of his army in protecting the coasts that he had himself been raiding for some time. l8l While the takeover of native centres as military bases by Vikings is well attested in England, the only English evidence for an occupation with royal approval- in the Wirral- is late and unreliable. 182 Nevertheless, the geography and toponymy of the isle of Flegg led James Campbell to propose that it could have been a similar Viking base, and that 173 174 175 176 177 17 8 179

180 181 182

Insley 1994; Sandred 1979; Insley 1998, 112-14; cf. Hart 1992,28-9.

Whitelock 1937-45; Abrams 2001.

See above, pp. 388- 90 and nn. 48, 53 on East Anglia's free peasants.

Margeson 1996,48; 1997,41 (7 for map).

ASC s.a. 866, 885, 893, 896, 903, 917.

Vikings are recorded in East Anglia only in 841, however.

ASC s.a. 825 and 827.

Coupland 199 8.

Nelson 1991,51 (s.a. 841); see also Besteman, this vol.

Radner 1978, vii-xxxiv. See above, p. 407 and n. 126.

4 18

LAND, SEA AND HOME

Flegg's thirteen names in -by 'may stand for an unknown major event in the history of East Anglia,.183 The Roman fort at Caister-on-Sea on the southern edge of the 'island' could have served as a fortified centre. 184 Flegg's unusual concentration of by-names has often attracted attention. 18S Fellows-Jensen considered that they represented Danish communities which ­ apparently in contrast with most other such settlements elsewhere in East Anglia ­ survived Edward's conquest, due to the density of occupation on Flegg and its isolation. 186 Ekwall judged Scandinavian occupation of Flegg to have been late, as the army that left Cirencester in 880 to divide up East Anglia 'would not march right through Norfolk and settle on the lower Waveney' (i.e. on the east coast of the county) .187 A late date was supported by the view that the water table had been higher in the 9th century, and that the land only became available in the loth or lIth. 188 More recent investigation, however, has suggested that the by-villages are located in areas that may already have been settled in the pre-Viking period; 189 and attention has been drawn to the island's excellent soils and array of resources such as reeds, salt, and above all, fish. l90 By 1086 these settlements were thriving and heavily populated. As Hart observed, '[this is] not the picture one might expect from a group of communities newly settled on inferior sites,.191 That thirteen of twenty-three major names in this small area are by-names demands some kind of explanation. Military occupation alone need leave no trace in place-names; but might a Walcheren-type establishment have been supported by something more permanent, represented by the settlements in -by? Fleets need feeding, and the resources of the island could have been exploited to that end. We know next to nothing about trade in this area; but if (as now seems likely) Norwich was occupied before the Scandinavian settlement, Caister could have had a role in protecting commercial traffic en route to the interior (as it may have done in the Roman period).I92 Its location, remote perhaps from the perspective of the south and west, 183 Campbell 2001, 18-21; d. Abrams forthcoming. Sandred 1986-7,25, observed: 'perhaps what happened in this marginal fenland area did not arrract much notice in centres where documents were produced.' 184 Darling and Gurney et al. 1993; the fort was reused in the Anglo-Saxon period by a high-status community. See further, Abrams forthcoming. 185 For detailed modern comment on the names of Flegg, see Sandred 1986-7; 1996, 1- 83; Insley 1998, 114- 15; Fellows-Jensen 1999, 48-52. For the first elements of the by-names, Sandred identified eleven Scandinavian personal names, one Old Norse noun and one existing place-name; Insley's interpretation was similar, though he replaced the Norse noun (in Filby) with an Old English personal name; Fellows-Jensen saw six Scandinavian personal names, two Old Norse nouns, one Old Norse adjective, three older place-names, and one name (Ashby) which she implied was coined as a hybrid, with Old English i1!SC as first element. 186 Fellows-Jensen 1999,49. 187 Ekwall 1924, 82. 188 Hart 1992, 113. Sandred 1986-7,7 and 23, suggested that the island was exploited but not inhabited; he proposed that an absence of early archaeological evidence, combined with his interpretation of pre-existing English place-names - *Stokes, with OE stoc 'outlying pasture', and Martham and Runham, both with OE hamm in the sense 'river-meadow' - indicated that the area was only used by the Anglo-Saxons for 'producing winter fodder and for grazing in the summer' (p. 23). 189 Small quantities of earlier Anglo-Saxon pottery and metalwork have been found in most of the parishes on Flegg. Lesley Abrams would like to thank Tim Pestell, Andrew Rogerson, Dan Smith and Tom Williamson for helpful discussions on these and other aspects of Flegg. 190 Hart 1992, 113; Campbell 2002, II - 15. 191 Hart 1992, 113. 192 Ayers 2003, 23-34; d . Campbell 1975, 1-8. On Roman Caister, see Gurney 1996.

PLACE-NAMES AND HISTORY

41 9

was in reality highly strategic, at the mouth of major rivers leading into the kingdom and facing Continental markets. Sceattas found in the fort (and as single finds elsewhere on Flegg) attest to commercial engagement in the 8th century.193 More thought-provoking is a coin of Ecgberht (in circulation from the 820S to the 850S) found under the head of a burial within the fort; 9th-century burials with coins are rare and, it could be argued, are generally found in contexts where a Scandinavian interpretation would be appropriate. 194 The potentially startling hypothesis that the place-names on Flegg might derive from an early or mid-9th-century Scandinavian settlement provides a useful point from which to review some of the arguments of our paper, and to explore some possibilities for further work. First of all, it is particularly clear that this concentrated group of by-names - in which as many as twelve of the thirteen may involve Old Norse first elements (at least half of them personal names) - reflects the presence, and presumably the settlement, of Scandinavian-speakers in Norfolk. Elsewhere in eastern England, we have argued, the implications of the by-names are similar: they generally seem to indicate Norse-speakers in concentrated (if not quite so concentrated or clearly delimited) areas. Secondly, there is nothing in the Flegg names themselves that allows them to be dated precisely, though the range of personal names hints at an earlier rather than a later date. 195 Again this seems to be generally true for eastern England as a whole. We are not suggesting, however, that all by-names are therefore likely to belong to the 'pre-settlement' period. Instead, we argue that the interpretation of local circumstances is a matter for historical analysis, informed by the place-name pattern. In Flegg, it has been suggested that settlement might be associated with an early, strategic placement of Vikings; elsewhere the names might be better linked ­ more or less directly - to the historically recorded divisions of land by the army, or to subsequent, unrecorded migrations in their wake. The name-evidence suggests a framework for historical discussion; in our view that framework is a flexible one, especially when it comes to questions of date. Several objections to the Flegg hypothesis can be anticipated, and first among them is date. The Anglo-Saxon Chronicle records a Scandinavian settlement in East Anglia in 880, and it may seem reckless to propose that the clearest concentration of Norse place-names in the region might relate to earlier, wholly unrecorded, activity. Our defence - that historical details are so scarce as to leave open all sorts of possibilities - has already been made. 196 This argument also has much wider implications. There is, quite understandably, a tendency for historians (and name­ scholars) to cling to the few fixed points in the surviving narrative accounts of the Scandinavian incursions and settlements. Yet inevitably these accounts are partial­ many other events must have taken place, and sometimes, no doubt, details of those that are recorded will be inaccurate. Even where details appear trustworthy, they Darling and Gurney er al. 1993,68- 9. The coin was a Canrerbury issue of Ecgberht (Darling and Gurney er al. 1993,69-71). Thanks are owed to Mark Blackburn for his help with the numismatic evidence. On coins in Viking-age burials see Biddle et al. 193

194

1986,25-32. (95 196

Below, pp. 420 - I. Above p. 417; d. Campbell 1975,6-7.

420

LAND, SEA AND HOME

need to be used with due circumspection. It was suggested above that land-holding among Norse-speakers need not have ceased as soon as political control passed to the West Saxons. 197 In this connexion it is worth noting the evidence brought forward by Mark Blackburn suggesting Scandinavian survival in Leicestershire beyond the submission of Leicester's Viking army in 918.198 A small coin-hoard of 'very strong' Scandinavian character, including two dirham pieces used simply as bullion, was deposited in the mid-92oS at Thurcaston, near Leicester, a parish (with a Grimston­ hybrid name) which has also produced a silver Thor's hammer. It is tempting to consider this numismatic evidence for apparent and unexpectedly late Scandinavians in Leicestershire beside the numismatic evidence that could support their unexpectedly early presence at Caister; and to conclude that the dates in documentary sources should be starting-points for negotiation, not the cardinal points that have to orientate every interpretation. Another objection to the hypothesis will be that it is unwise to interpret one group of by-names differently from others. In Flegg as in other parts of eastern England the names are characteristically combined with a high proportion of Norse personal names - in this regard the eastern English by-names form a group together, apparently contrasting with the names in Denmark. As we have seen, it has been suggested that the by-names in England should therefore be taken together as generally indicative of a particular sequence of events. There is undoubtedly some appeal in this argument, and it might be right. Yet it is not at all clear that it has to be so. Land may have fallen into the hands of Scandinavian settlers in England in a variety of ways (most of them rarely occurring in the less volatile conditions of the homeland), and the response might often have involved the coining of place-names incorporating the name of the new owner. Indeed, this is perhaps a more cautious position to adopt than one that assumes too much similarity in the historical circumstances of, for instance, Norfolk and North Yorkshire in the early loth century. Flegg doubtless belonged to somebody before it was settled by Scandinavians, and the by-names may well, therefore, illustrate the break-up of an older estate. That this break-up resulted from a series of English victories which weakened seigneurial control in eastern Norfolk, in just the same way as it did across the East Midlands and Yorkshire, produces a neat single hypothesis, but it is not certain that a single explanation is required. Scandinavian conquest and settlement must have caused numerous local disruptions in the 9th and lOth centuries: we cannot be certain that the coining of the by-names closely reflects any single set of circumstances. One sign that, quite apart from the vagaries of recording, there may be significant local variety underlying the coining of by-names in various parts of the country comes from the linguistic analysis of their first elements. Interesting patterns, which may warrant closer inspection, have been noted in the language and type of these elements. 199 Similarly, Sandred identified eleven Norse personal names in the Flegg by-names, of which only that in Rollesby (recorded as RothoLfuesby in Domesday 197 198 199

p. 400 •

Blackburn

2001,

Above, p. 406 .

137.

PLACE-NAMES AND HISTORY

421

Book) is of compound type. He concluded that this ratio (9%) is significantly lower than in Yorkshire or the East Midlands (both calculated at over 30% by Fellows­ Jensen), and suggested that the difference might be due to the social status of the settlers: the by-names on Flegg would by his reasoning point to a lower class of Norseman than in the by-names further north. 2OO Alternatively one might suggest that a significant difference in the make-up of the personal names might reflect a difference in date. We should of course be cautious in using figures derived from such a small sample as the thirteen by-names on Flegg. It is noteworthy, however, that the variant interpretations offered by Insley and Fellows-Jensen would also produce relatively low figures for Norse compound names: Insley had two of eleven (18%) and Fellows­ Jensen one of just six (17%).201 There is clearly further detailed work to be done in defining and comparing the patterns made by the first elements in by-names and assessing their significance. Another indication that it might be unwise to seek a single explanation for by­ names comes from the (sometimes contradictory) evidence for the quality and status of their sites. It appears that the community on Flegg may have enjoyed a number of advantages,2°2 which would tend to set this group of names apart from the generally less well-off by-names further north. In fact, it would be interesting to see more multidisciplinary work on the status of the East Midland and Yorkshire settlements with by-names. Cameron and Fellows-Jensen have shown that, on the whole, their status is not high, and the average statistics are sufficient to support the general conclusion, drawn above, that the class seems unlikely to represent elite military takeover. The same scholars have suggested numerous exceptions, however, on grounds of geographical situation or name-type,203 and it would be interesting to see how far such exceptions can be tested, and how far they may fall into local patterns. In many areas it would be desirable to re-examine the documentary, archaeological and landscape evidence at a very local level ,204 and with reference to the distribution of the various classes of by-names (hybrid, anthroponymic, appellative).205 It is not to 200 Sandred J986-7, 8 and 24. Note that he is talking about names of traditional, 'heroic' compound type: Hr6th-ulfr, As-geirr, etc. Sometimes nicknames or bynames also have a compound form, as for instance ':'Bithi/-Aki, 'wooer Aki', which he favours for Billockby on Flegg. On the classification of Norse personal names see further Fellows-Jensen 1968, XXIX-XXXI. 201 See n. 185 above for references to these discussions. Insley favoured a compound *Bith-lcikr for the difficult Billockby. (Fellows-Jensen thought it might contain an older English place-name, *Bit-lacu.) 202 See above, p. 4 18. 203 Sites in Lincolnshire that have been considered to suggest older English settlements taken over (because the quality of their land looks higher than average) include Ulceby in north-east Lindsey (Cameron 1965, 125), and Cole by and Navenby, south-west of Lincoln (SSNEM, 311). The first elements in all three of these names appear to be Norse personal names. For the suggestion that by-names with English first elements are often partially renamed English settlements, see above, p. 399 and n. 94. 204 Archaeological material, in particular, has multiplied since Cameron and Fellows-Jensen published their detailed analyses of sites in the East Midlands; cf., e.g., Leahy and Paterson 2001. 205 For example, Fellows-Jensen usefully separated out kirk;u-by(r) source of the many Kir(k)bys - from other by-names when calculating the percentage of places that became medieval parishes (SSNY, 227-8; SSNEM, 353-5). Her finding that a much higher proportion of kirku-by(r)s than of by-names in general became parishes supports the suggestion that these were old and established sites, whether with churches or simply in ecclesiastical ownership. Recent work in Yorkshire by Tom Pickles argues the case for pre-Viking ecclesiastical ownership of many Kir(k)bys. We are grateful for the opportunity to read his work in draft form. See further Fellows-Jensen 1987, 298-9. Cameron 1985, 134, agreed that kirku-by(r) as a class probably represented takeovers.

422

LAND, SEA AND HOME

be denied that there are many difficulties in this work: assessing the quality of a site is not straightforward/o6 whilst even agreeing on a definition of previously 'uncolonized' land will prove problematic - is it restricted to virgin land, or does it encompass land that may once have been settled but has long been deserted, or even land exploited for agriculture (perhaps only seasonally) but without an immediately adjoining habita­ tion? One suspects that the settlers may have encountered all these situations; similarly, 'taking over' existing settlements could have involved a range of diverse circumstances. Our understanding of these events will perhaps best be advanced by a series of micro-studies which treat small regions separately and co-ordinate the toponymic, archaeological, environmental, and landscape evidence. 207 We began this paper by questioning whether place-names were useful to the historian of the Scandinavian settlements. We hope to have reaffirmed that they offer rich source material, though like all sources they need care, and are limited in their implications. A review of the by-names of eastern England has suggested that they should be treated, in general, as evidence for Norse-speaking communities and a relatively large number of settlers. We have argued that it might be attractive to associate the names with the disruptive events of the late 9th century, though later dates cannot be ruled out (with the interesting corollary that the later one dates the coining of the names, the longer distinctively Scandinavian communities would seem to have maintained their linguistic identity). We have cautioned against the assumption that absence of place-names necessarily indicates a lack of Scandinavian settlers. Finally, we have suggested - in more than one connexion - that patterns of by-names might be helpfully studied and interpreted in a closely local context: whilst there seem to be some general truths about the class, it may be dangerous to impose single interpretations on the whole of Britain, or England, or even eastern England. The historical value of place-name evidence is not limited to by-names, though this paper has largely concentrated upon them. Some further types of name are clearly of use to historians - those which illustrate the Hiberno-Norse presence in northern England, for instance208 - and others could profitably be considered. Scandinavian administrative names - the names of divisions, hundreds and assembly places ­ have not received the attention accorded to settlement names, and yet careful study could perhaps delineate areas subject to local assemblies and tell us about social and legal institutions before they emerged (faintly) into documentation in the late roth century and beyond.209 Of particular interest, perhaps, are administrative names ­ like Thinghoe ('assembly mound') hundred in Suffolk or Dacorum hundred ('hundred of the Danes') in Hertfordshire - which survived in areas without many settlement

206 While Sawyer (1957-8, 14-15, citing A. H. Smith) initially characterized the many by-names on the Yorkshire Wolds as attached to minor villages and undesirable land, for example, he came to re-evaluate the status of these estates as central, not marginal - at least economically (in terms of sheep-farming), if not in terms of location (Sawyer 1978a, 162). 207 We are not, of course, the first to call for such interdisciplinary work: see, e.g., Hadley 2000a, 34I. 208 Summarized in SSNNW, 319-20 (arguing that there is good reason to suppose that some of the Gaelic­ srceaking 'Hiberno-Norse' probably came from Scotland). 09 Specialised comment on such names is not lacking: see, e.g. Anderson 1934-9; Cameron 1965, Il7-18; Fellows-Jensen 1990; 1993; Meaney 1997; Pantos 1998-9; Sandred 1994.

PLACE-NAMES AND HISTORY

42 3

names. 210 Should these be regarded as evidence for administrative control without substantial settlement, or might things be more complicated? There also remains room for more work - again, preferably multidisciplinary and collaborative - on another of the major classes of Scandinavian place-names, the Grimston-hybrids. Recent study of the personal names of lIth-century landholders indicates that the ratio of Norse to English in the name-stock is identical (33% to 67%) in the adjacent counties of Suffolk and Essex. 211 How, then, should we explain the fact that there are around 25 Grimston-hybrids (combining Norse-named landholders with Old English -tun) in Suffolk, and no certain instances amongst the major names of Essex?12 This might confirm that Suffolk was part of the East Anglian region that was divided up between the Danish army in 880, that Essex effectively lay outside the Danelaw,213 and that Grimston-hybrids (despite recent doubts) are indeed generally to be dated to the settlement period. A further factor needs to be taken into consideration, however, for Domesday Book reveals a marked distinction in lIth-century social organization between Suffolk and Essex - the former (like Norfolk) abounds in free peasants, and the latter does not. 214 Perhaps it is this distinction, the Scandinavian origin of which has been disputed,215 which has determined the place-name pattern. Much thought still needs to be devoted to the relationship, if any, between free peasants, Grimston­ hybrids and Scandinavian settlement - it might be hoped that a multi-disciplinary approach to studying the names, sites and status of settlements either side of the Suffolk/Essex boundary would shed some light on the subject. In conclusion we would say that, despite the enormous range of conflicting opinions that have gone before, place-names carefully treated remain a valuable historical source. Further research into them - especially in a collaborative, interdisciplinary context - might yet improve our understanding of one of the most obscure periods in the history of England. BIBLIOGRAPHY AND ABBREVIATIONS

Abrams, L., 2001, 'The conversion of the Danelaw', 31-44 in Graham-Campbell et al. 200I. Abrams, L., forthcoming, 'Scandinavian place-names and settlement history: Fiegg, Norfolk', Proceedings of the Fourteenth Viking Congress. Aird, W. M., 1998, St Cuthbert and the Normans. The Church of Durham, l071-UB (Woodbridge) . Anderson, O. S., 1934-9, The English Hundred-Names, 3 vols (Lund).

210

For Thinghoe, see Ekwall 1924, 81, and Anderson 1934-9, I, 95- 6; for Dacorum, see Gover et at. 1938,

2z1-6, and Anderson 1934-9, III, 28-30.

Parsons 2002, 42. For a provisional list of the Suffolk instances see Insley 1999, 53-4. In Essex, intriguingly, there are numerous compounds of personal name plus -tun amongst field-names and 'nearly all of these are in the north of the county'. They include Grimston-types, as Thorcetelestone, first recorded in 1275. See Reaney 1935, 5:Z1-2. Clearly these Essex names are crying out for further attention. 13 Dumville 1992, 8-17; Williams 1996. 214 Darby 1971, 168-70 and 224-5. 215 See above, pp. 388-90 and 4 17. 11

212

424

LAND, SEA AND HOME

Aston, T. H., 1958, 'The origins of the manor in England', Trans. of the Royal Historical Soc., 5th ser. 8, 59-83; reprinted with postscript at 1-43 in T. H. Aston et al. (eds.), Social Relations and Ideas. Essays in Honour of R. H. Hilton (Cambridge, 1983) [cited from the reprint]. Ayers, B., 2003, Norwich: 'A Fine City' (London). Bailey, R. N., 1980, Viking Age Sculpture in Northern England (London). Barrow, G. W. S., 1973, The Kingdom of the Scots. Government, Church and Society from the Eleventh to the Fourteenth Century (London). Biddle, M. et aI., 1986, 'Coins of the Anglo-Saxon period from Repton, Derbyshire: 11', British Numismatic }., 56, 16-34. Binns, A., 1956, 'Tenth century carvings from Yorkshire and the Jellinge style', Universitet i Bergen Arbok 1956, Historisk-antikvarisk rekke, 2, 1-29. Blackburn, M. A. S., 2000, 'Metheltun not Medeshamstede: an Anglo-Saxon mint at Melton Mowbray rather than Peterborough Abbey', British Numismatic }., 70,143-5. Blackburn, M. A. S., 2001, 'Expansion and control: aspects of Anglo-Scandinavian minting south of the Humber', 125-42, in Graham-Campbell et al. 2001. Blake, E. O. (ed.), 1962, Liber Eliensis (Camden 3rd ser. 92, London). Blunt, C. E., 1974, 'The coinage of Athelstan, 924-939: a survey', British Numismatic }., 42, 35- 160. Brooks, N. P., 1979, 'England in the ninth century: the crucible of defeat', Trans. of the Royal Historical Soc., 5th ser. 29, 1-20. Brown, T., and G. Foard, 1998, 'The Saxon landscape: a regional perspective', 67-94 in P. Everson and T. Williamson (eds.), The Archaeology of Landscape. Studies Presented to Christopher Taylor (Manchester). Cameron, K., 1965, 'Scandinavian settlement in the territory of the Five Boroughs: the place-name evidence', first published as an inaugural lecture (Nottingham); reprinted at II5-38 in K. Cameron (ed.), Place-Name Evidence for the Anglo-Saxon Invasion and Scandinavian Settlements (Nottingham, 1975) [cited from the reprint]. Cameron, K., 1970, 'Scandinavian settlement in the territory of the Five Boroughs: the place-name evidence. Part II: place-names in thorp', Mediaeval Scandinavia, 3, 35-49; reprinted at 139-56 in K. Cameron (ed.), Place-Name Evidence for the Anglo-Saxon Invasion and Scandinavian Settlements (Nottingham, 1975) [cited from the reprint]. Cameron, K., 1971, 'Scandinavian settlement in the territory of the Five Boroughs. Part III: the Grimston hybrids', 147-63 in P. Clemoes and K. Hughes (eds.), England Before the Conquest (Cambridge); reprinted with postscript at 157-71 in K. Cameron (ed.), Place-Name Evidence for the Anglo-Saxon Invasion and Scandinavian Settlements (Nottingham, 1975) [cited from the reprint]. Cameron, K., 1976, The Significance of English Place-Names (Sir Israel Gollancz Memorial Lecture, Oxford). Cameron, K., 1985, 'Viking settlement in the East Midlands: the place-name evidence', 129-53 in R. Schiitzeichel (ed.), Gief5ener Flurnamen-Kolloquium r. bis 4. Oktober 1984 (Heidelberg). Cameron, K., 1996, 'The Scandinavian element in minor names and field-names in north-east Lincolnshire', Nomina, 19,5-27. Cameron, K., 1998, A Dictionary of Lincolnshire Place-Names (Nottingham). Campbell, J., 1975, Norwich (Historic Towns, 2, London). Campbell, J., 2001, 'What is not known about the reign of Edward the Elder', 12-24 in N. J. Higham and D. H. Hill (eds.), Edward the Elder 899-924 (London and New York). Campbell, J., 2002, 'Domesday herrings', 5-17 in C. Harper-Bill et al. (eds.), East Anglia's History. Studies in honour of Norman Scarfe (Woodbridge). Christiansen, E., 2002, The Norsemen in the Viking Age (Oxford).

PLACE-NAMES AND HISTOR Y

42 5

Clark, c., 1992, 'Onomastics', 542-606 in N. Blake (ed.), The Cambridge History of the English Language. Volume II: 1066-1476 (Cambridge). Collingwood, W. G., 1908, Scandinavian Britain (London). Coupland, S., 1998, 'From poachers to gamekeepers: Scandinavian warlords and Carolingian kings', Early Medieval Europe, 7, 85-II4. Cox, B., 1989, 'Rutland and the Scandinavian settlements: the place-name evidence', Anglo-Saxon England, 18, 135-48. Cramp, R. (gen. ed.), 1984- , Corpus of Anglo-Saxon Stone Sculpture: vol. 1, R. Cramp (ed.), County Durham and Northumberland (Oxford, 1984); vol. 2, R. N. Bailey and R. Cramp (eds.), Cumberland, Westmorland, Lancashire North-of-the-Sands (Oxford, 1988); vol. 3, J. Lang (ed.), York and Eastern Yorkshire (Oxford, 1991); vol. 5, P. Everson and D. Stocker (eds.), Lincolnshire (Oxford, 1999); vol. 6, J. Lang (ed.), Northern Yorkshire (Oxford, 2001). Craster, E., 1954, 'The patrimony of St. Cuthbert', English Historical Review, 69,177-99. Crawford, B. E. (ed.), 1995, Scandinavian Settlement in Northern Britain. Thirteen Studies of Place­ Names in their Historical Context (London and New York) . Darby, H. c., 1971, The Domesday Geography of Eastern England, 3rd ed. (Cambridge). Darby, H. c., and E. M. J. Campbell, 1962, The Domesday Geography of South-East England (Cambridge). Darby, H. c., and I. S. Maxwell, 1962, The Domesday Geography of Northern England (Cambridge). Darby, H. c., and I. B. Terrett, 1971, The Domesday Geography of Midland England, 2nd ed. (Cambridge). Darling, M. J., and D. Gurney et al. 1993, Caister-on-Sea Excavations by Charles Green, 1951-55 (East Anglian Archaeology, 60, Gressenhall). Davis, R. H. C. (ed.), 1954, The Kalendar of Abbot Samson of Bury St Edmunds and Related Documents, Camden soc., 3rd ser. 84. Davis, R. H. c., 1955, 'East Anglia and the Danelaw', Trans. of the Royal Historical Soc., 5th ser. 5, 23-37; reprinted at 15-32 in his From Alfred the Great to Stephen (London, 1991) [cited from the reprint]. Dodwell, B., 1941, 'The free peasantry of East Anglia in Domesday', Norfolk Archaeology, 27, 145-57· Dumville, D. N ., 1992, Wessex and England from Alfred to Edgar (Woodbridge). Dumville, D. N. et al. (eds.), 1983- , The Anglo-Saxon Chronicle. A Collaborative Edition (Cambridge). Dyer, c., 2002, Making a Living in the Middle Ages. The People of Britain 850-1520 (New Haven and London). Ekwall, E., 1924, 'The Scandinavian element', 55-92 in A. Mawer and F. M. Stenton (eds.), Introduction to the Survey of English Place-Names (Cambridge). Everson, P. L. et al. (eds.), 1991, Change and Continuity. Rural Settlement in North-West Lincolnshire (London). Faith, R., 1997, The English Peasantry and the Growth of Lordship (London and Washington). von Feilitzen, 0.,1937, The Pre-Conquest Personal Names of Domesday Book (Uppsala). Fellows-Jensen, G., 1968, Scandinavian Personal Names in Lincolnshire and Yorkshire (Navne­ studier, 7, Copenhagen). Fellows-Jensen, G., 1972, Scandinavian Settlement Names in Yorkshire (Navnestudier, II, Copenhagen) [abbreviated SSNY]. Fellows-Jensen, G., 1974, 'English field-names and the Danish settlement', 45-55 in P. Andersen et al. (eds.), Festskrift til Kristian Hald (Copenhagen). Fellows-Jensen, G., 1975, 'Personal name or appellative? A new look at some Danelaw place­ names', Onoma, 19,445-58.

426

LAND, SEA AND HOME

Fellows-Jensen, G., 1978, Scandinavian Settlement Names in the East Midlands (Navnestudier, 16, Copenhagen) [abbreviated SSNEM]. Fellows-Jensen, G., 1981, 'Scandinavian settlement in the Danelaw in the light of the place-names of Denmark', 133-45 in H. Bekker-Nielsen et al. (eds.), Proceedings of the Eighth Viking Congress (Odense). Fellows-Jensen, G., I983a, 'Anthroponymical specifics in place-names in -by in the British Isles', Studia Anthroponymica Scandinavica, 1,45-60. Fellows-Jensen, G., I983b, 'Scandinavian settlement in the Isle of Man and north-west England: the place-name evidence', 37-52 in C. Fell et al. (eds.), The Viking Age in the Isle of Man. Select Papers from The Ninth Viking Congress (London). Fellows-Jensen, G., 1984, 'Place-names and settlements: some problems of dating as exemplified by place-names in -bY', Nomina, 8,29-39. Fellows-Jensen, G., 1985, Scandinavian Settlement Names in the North-West (Navnestudier, 25, Copenhagen) [abbreviated SSNNW]. Fellows-Jensen, G., 1987, 'The Vikings' relationship with Christianity in the British Isles: the evidence of place-names containing the element kirkja', 295-307 in J. E. Knirk (ed.), Proceedings of the Tenth Viking Congress (Oslo). Fellows-Jensen, G., 1990, 'Amounderness and Holderness', Namn och Bygd, 78, 23-30. Fellows-Jensen, G., I99Ia, 'Of Danes - and thanes - and Domesday Book', II2-2I in I. Wood and N. Lund (eds.), People and places in Northern Europe 500-r600. Essays in Honour of Peter Hayes Sawyer (Woodbridge). Fellows-Jensen, G., I99Ib, 'Scandinavian influence on the place-names of England', 337-54 in P. S. Ureland and G. Broderick (eds.), Language Contact in the British Isles (Tiibingen). Fellows-Jensen, G., 1991-2, 'Place-names in -fJOrp: in retrospect and in turmoil', Nomina, IS, 35-51. Fellows-Jensen, G., 1993, 'Tingwall, Dingwall and Thingwall', North-West European Language Evolution, 2112, 53-67. Fellows-Jensen, G., 1994, 'Danish place-names and personal names in England: the influence of Cnut?', 125-40 in A. R. Rumble (ed.), The Reign of Cnut, King of England, Denmark and Norway (London). Fellows-Jensen, G., 1995, 'Scandinavian settlement in Yorkshire - through the rear-view mirror', 170-86 in B. E. Crawford (ed.), Scandinavian Settlement in Northern Britain. Thirteen Studies of Place-Names in their Historical Context (London and New York). Fellows-Jensen, G., 1997, 'Scandinavians in Cheshire: a reassessment of the onomastic evidence', 77-92 in A. R. Rumble and A. D. Mills (eds.), Names, Places and People. An Onomastic Miscellany in Memory of John McNeal Dodgson (Stamford). Fellows-Jensen, G., 1998, 'The origin and development of the name York', 226-37 ·in D. W. Rollason, Sources for York History to AD 1I00 (York). Fellows-Jensen, G., 1999, 'Scandinavian settlement names in East Anglia: some problems', Nomina, 22,45-60. Fellows-Jensen, G., 2000, 'Old English socn "soke" and the parish in Scandinavia', Namn och Bygd, 88, 89-106. Fellows-Jensen, G., 200Ia, 'In the steps of the Vikings', 279-88 in Graham-Campbell et al. 2001. Fellows-Jensen, G., 200Ib, 'The mystery of the by-names in Man', Nomina, 24,33-46. Finberg, H. P. R., 1974, The Formation of England 550-r042 (London). Fleming, R., 1985, 'Monastic lands and England's defence in the Viking Age', English Historical Review, 100,247-65. Fox, H. S. A., 1989, 'The people of the Wolds in English settlement history', 77-101 in M. Aston et al. (eds.), The Rural Settlements of Medieval England (Oxford).

PLACE-NAMES AND HISTORY

42 7

Gelling, M., 1988, Signposts to the Past. Place-Names and the History of England (Chichester; 2nd ed. 1988; 3rd ed. [substantially unchanged] 1997). Gelling, M., 1991, 'The place-names of the Isle of Man', 141-55 in P. S. Ureland and G. Broderick (eds.), Language Contact in the British Isles (Tiibingen). Gelling, M., 1992, The West Midlands in the Early Middle Ages (Leicester). Gelling, M., 1995, 'Scandinavian settlement in Cheshire: the evidence of place-names', 187-94 in B. E. Crawford (ed.), Scandinavian Settlement in Northern Britain. Thirteen Studies of Place­ Names in their Historical Context (London and New York). Gover, J. E. B., A. Mawer and F. M. Stenton, 1938, The Place-Names of Hertfordshire (EPNS, 15, Cambridge). Graham-Campbell, J., R. Hall, J. Jesch and D. N. Parsons (eds.), 2001, Vikings and the Danelaw. Select Papers from the Proceedings of the Thirteenth Viking Congress (Oxford). Gurney, D., 1996, 'The "Saxon Shore" in Norfolk', 30-9 in S. Margeson et al. (ed.), A Festival of Norfolk Archaeology (Norwich). Hadley, D. M., 1996, 'Multiple estates and the origins of the manorial structure of the northern Danelaw',}. of Historical Geography, 22,3-15. Hadley, D. M., 1997, '''And they proceeded to plough and to support themselves": the Scandinavian settlement of England', Anglo-Norman Studies, 19,69-96. Hadley, D. M., 2000a, The Northern Danelaw.lts Social Structure, c. 800-lIOO (London). Hadley, D. M., 2000b, '''Cockle amongst the wheat": the Scandinavian settlement of England', III-35 in W. O. Frazer and A. Tyrrell (eds.), Social Identity in Early Medieval Britain (London). Hadley, D. M., 2001, 'In search of the vikings: the problems and the possibilities of interdisciplinary approaches', 13-30 in Graham-Campbell et al. 200I. Hadley, D. M., 2002, 'Viking and native: re-thinking identity in the Danelaw', Early Medieval Europe, II, 45-70. Hadley, D. M., and J. D. Richards (eds.), 2000, Cultures in Contact. Scandinavian Settlement in England in the Ninth and Tenth Centuries (Turnhout). Hald, K., 1969, 'A reply to Peter Sawyer', 185-7 in P. H. Sawyer et aI., 'The two Viking Ages of Britain: a discussion', Mediaeval Scandinavia, 2,163-207. Hart, C. R., 1992, The Danelaw (London). Hines,J., 1991, 'Scandinavian English: a creole in context', 403-27 in P. S. Ureland and G. Broderick (eds.), Language Contact in the British Isles (Tiibingen). Hofmann, D., 1955, Nordisch-Englische Lehnheziehungen der Wikingerzeit (Copenhagen). Hooke, D., 1998, The Landscape of Anglo-Saxon England (London). Innes, M., 2000, 'Danelaw identities: ethnicity, regionalism, and political allegiance', 65-88 in Hadley and Richards 2000. Insley, J., 1979, 'Regional variation in Scandinavian personal nomenclature in England, Nomina, 3,52-60. Insley, J., 1985, 'Field names and the Scandinavian settlement of England: a comparative study of the evidence provided by the English Place-Name Society's Survey of Northamptonshire', 113-28 in R. Schiitzeichel (ed.), Gief5ener Flurnamen-Kolloquium 1. his 4. Oktoher 1984 (Heidel berg). Insley, J., 1986, 'Toponymy and settlement in the North-West', Nomina, 10, 169-76. Insley, J., 1994, Scandinavian Personal Names in Norfolk. A Survey Based on Medieval Records and Place-Names (Uppsala). Insley, J., 1998, 'The place-names of Norfolk: a Danelaw project in progress', Namn och Bygd, 86, 105- 1 7.

428

LAND, SEA AND HOME

Insley, J., 1999, 'Grimston-hybrids', 49-56 in H. Beck et al. (eds.), Reallexikon der Germanischen Altertumskunde von Johannes Hoops, 2nd ed., XIII (Berlin and New York). Jesch, J., 2000, 'Scandinavian Wirral', 1-10 in P. Cavill et al. (eds.), Wirral and its Viking Heritage, EPNS (Nottingham). John, E., 1982, 'The age of Edgar', 160-91 in J. Campbell, E. John and P. Wormald, The Anglo­ Saxons (Oxford). Johnson South, T. (ed. and trans.), 2002, Historia de sancto Cuthberto. A History of Saint Cuthbert and a Record of his Patrimony (Woodbridge). Jones, G. R. J., 1965, 'Early territorial organization in northern England and its bearing on the Scandinavian settlement', 67-84 in A. Small (ed.), [Proceedings of] the Fourth Viking Congress (Edinburgh). Jones, G. R. J., 1976, 'Multiple estates and early settlement', 15-40 in P. H. Sawyer (ed.), Medieval Settlement: Continuity and Change (London). Jones, G. R. J., 1990, 'Celts, Saxons and Scandinavians', 45-68 in R. A. Dodgshon and R. A. Butlin (eds.), An Historical Geography of England and Wales, 2nd ed. (London). Kapelle, W. E., 1979, The Norman Conquest of the North. The Region and its Transformation looO-II35 (London). Keynes, S., 1997, 'The Vikings in England, c. 790-1016', 48-82 in P. H. Sawyer (ed.), The Oxford Illustrated History of the Vikings (Oxford). Keynes, S., and M. Lapidge, 1983, Alfred the Great (Harmondsworth). Kristensen, A. K. G., 1975, 'Danelaw institutions and Danish society in the Viking Age', Mediaeval Scandinavia, 8, 27-85. Kurath, H. et al. (eds.), 1954-2001, Middle English Dictionary (Ann Arbor). Lang, J. T., 1978, 'Anglo-Scandinavian Sculpture in Yorkshire', 11-20 in R. A. Hall (ed.), Viking Age York and the North (CBA Research Report, 27, London). Leahy, K., and C. Paterson, 2001, 'New light on the Viking presence in Lincolnshire: the artefactual evidence', 181-202 in Graham-Campbell et al. 200I. Liebermann, F. (ed.), 1903-16, Die Gesetze der Angelsachsen, 3 vols (Halle). Lund, N., 1969, 'The secondary migration', 196-201 in P. H. Sawyer et aI., 'The two Viking Ages of Britain: a discussion', Mediaeval Scandinavia, 2, 163-207. Lund, N., 1976, 'Thorp-names', 223-5 in P. H . Sawyer (ed.), Medieval Settlement: Continuity and Change (London). Lund, N., 1981, 'The settlers: where do we get them from - and do we need them?', 147-71 in H. Bekker-Nielsen et al. (eds.), Proceedings of the Eighth Viking Congress (Odense). Maitland, F. W., 1897, Domesday Book and Beyond (Cambridge). Margeson, S., 1996, 'Viking settlement in Norfolk: a study of new evidence', 47-57 in S. Margeson et al. (eds.), A Festival of Norfolk Archaeology (Norwich). Margeson, S., 1997, The Vikings in Norfolk (Norwich). Mawer, A., 1931-2, 'The Scandinavian settlements in England as reflected in English place-names', Acta Philologica Scandinavica, 1-30. Meaney, A. L., 1997, 'Hundred meeting-places in the Cambridge region', 195-240 in A. R. Rumble and A. D. Mills (eds.), Names, Places and People. An Onomastic Miscellany in Memory of John McNeal Dodgson (Stamford). Morris, C. D., 1977, 'Northumbria and the Viking settlement: the evidence for land-holding', Archaeologia Aeliana, 5th ser. 5, 81-102. Morris, C. D., 1981, 'Viking and native in northern England: a case-study', 223-44 in H. Bekker­ Nielsen et al. (eds.), Proceedings of the Eighth Viking Congress (Odense). Morris, C. D., 1984, 'Aspects of Scandinavian settlement in northern England: a review', Northern History, 10, 1-22.

PLACE-NAMES AND HISTORY

42 9

Nelson, J. L. (trans.) 1991, Annals of St Bertin (Manchester). Page, R. I., 1971, 'How long did the Scandinavian language survive in England? The epigraphical evidence', 165-81 in P. Clemoes and K. Hughes (eds.), England Before the Conquest (Cambridge); reprinted with postscript at 181-96 in his Runes and Runic Inscriptions. Collected Essays on Anglo-Saxon and Viking Runes (Woodbridge, 1995) [cited from the reprint]. Pantos, A., 1998-9, 'Meeting-places in Wilvaston hundred, Cheshire', J. of the English Place-Name Soc., 31, 91-II2. Parsons, D. N., 2001, 'How long did the Scandinavian language survive in England? Again', 299-312 in Graham-Campbell et al. 2001. Parsons, D. N., 2002, 'Anna, Dot, Thorir . . . counting Domesday personal names', Nomina, 25, 29-52. Parsons, D. N., and T. Styles, 2000, The Vocabulary of English Place-Names: Brace-Ccester (Nottingham) . Phythian-Adams, c., 1996, Land of the Cumbrians. A Study in British Provincial Origins A.D. 400-II20 (Aldershot). Radner, J. N. (ed. and trans.) 1978, Fragmentary Annals of Ireland (Dublin). Reaney, P. H., 1935, The Place-Names of Essex (EPNS, 12, Cambridge) . Reaney, P. H., and R. M . Wilson, 1995, A Dictionary of English Surnames, 3rd ed. (Oxford). Ridel, E., 2000, 'Viking maritime heritage in Normandy from a British Isles perspective', Northern Studies, 35, 79-93· Ridel, E., forthcoming, 'The linguistic heritage of the Scandinavians in Normandy', in J. Adams and K. Holman (eds.), Scandinavians and Europe 800-1350. Contact, Conflict and Co­ existence. Roberts, B. K., 1989-90, 'Late -by names in the Eden Valley, Cumberland', Nomina, 13,25-40. Roesdahl, E., 1991, T he Vikings (Harmondsworth). Roffe, D. R., 1990a, 'An introduction to the Derbyshire Domesday', The Derbyshire Domesday (London), 1-27. Roffe, D. R., 1990b, 'From thegnage to barony: sake and soke, title, and tenants-in-chief', Anglo­ Norman Studies, 12, 157-76. Roffe, D. R., 1992, 'An introduction to the Lincolnshire Domesday', The Lincolnshire Domesday (London), 1-31. S. + number - see Sawyer 1968. Sandred, K. I., 1979, 'Scandinavian place-names and appellatives in Norfolk: a study of the medieval field-names of Flitcham', Namn och Bygd, 67, 98-122. Sandred, K. l., 1986-7, 'The Scandinavians in Norfolk: some observations on the place-names in ­ by',}. of the English Place-Name Soc., 19, 5-28. Sandred, K. l., 1994, 'Viking administration in the Danelaw: a look at Scandinavian and English hundred-names in Norfolk', 269-76 in B. Ambrosiani and H. Clarke (eds.), Developments around the Baltic and the North Sea in the Viking Age [The Twelfth Viking Congress] (Birka Studies, 3, Stockholm). Sandred, K. l., 1996, The Place-Names of Norfolk. Part Two: The Hundreds of East and West Flegg, Happing and Tunstead (EPNS, 72, Nottingham) . Sawyer, P. H., 1957-8, 'The density of the Danish settlement in England', University of Birmingham Historical }., 6,1-17. Sawyer, P. H., 1962, The Age of the Vikings, 1st ed. (London) . Sawyer, P. H ., 1968, Anglo-Saxon Charters. An Annotated List and Bibliography (London) [abbreviated as S.] . Sawyer, P. H., 1971, T he Age of the Vikings, 2nd ed. (London).

43 0

LAND, SEA AND HOME

Sawyer, P. H., 1975, 'The charters of Burton Abbey and the unification of England', Northern History, 10,28-39. Sawyer, P. H., 1978a, From Roman Britain to Norman England (London; 1st ed. 1978; 2nd ed. [substantially unchanged) 1998). Sawyer, P. H., 1978b, 'Some sources for the history of Viking Northumbria', 3-7 in R. A. Hall (ed.), Viking Age York and the North (CBA Research Report, 27, London). Sawyer, P. H., 1981, 'Conquest and colonization: Scandinavians in the Danelaw and in Normandy', 123-31 in H. Bekker-Nielsen et al. (eds.), Proceedings of the Eighth Viking Congress (Odense). Sawyer, P. H., 1982, Kings and Vikings. Scandinavia and Europe AD 700-1100 (London). Sawyer, P. H., 1994, Scandinavians and the English in the Viking Age (H. M. Chadwick Memorial Lecture, 5, Cambridge) . Sawyer, P. H., 1998, Anglo-Saxon Lincolnshire (Lincoln) . Sawyer, P. H., et al. 1969, 'The two Viking Ages of Britain: a discussion', Mediaeval Scandinavia, 2, 163-207. Sidebottom, P., 2000, 'Viking age stone monuments and social identity in Derbyshire', 213-35 in Hadley and Richards 2000. Simpson, L., 1989, 'The King Alfred/St Cuthbert episode in the Historia de sancto Cuthberto: its significance for mid-tenth-century English history', 397-411 in G. Bonner et al. (eds.), St Cuthbert, his Cult and his Community to A.D. 1200 (Woodbridge). Smart, V ., 1986, 'Scandinavians, Celts, and Germans in Anglo-Saxon England: the evidence of moneyers' names', 171-84 in M. A. S. Blackburn (ed.), Anglo-Saxon Monetary History. Essays in Memory of Michael Dolley (Leicester). Smith, A. H., 1956, English Place-Name Elements, 2 vols (EPNS, 25-6, Cambridge). SSNEM, see Fellows-Jensen 1978. SSNNW, see Fellows-Jensen 1985. SSNY, see Fellows-Jensen 1972. Stafford, P., 1985, The East Midlands in the Early Middle Ages (Leicester). Stenton, F. M., 1910, Types of Manorial Structure in the Northern Danelaw (Oxford). Stenton, F. M., 1927, 'The Danes in England', Proc. of the British Academy, 8,203-46; reprinted at 136-65 in D. M. Stenton (ed.), Preparatory to Anglo-Saxon England (Oxford, 1970) [cited from the reprint). Stenton, F. M., 1942, 'The Danish settlement of eastern England', Trans. of the Royal Historical Soc., 4th ser. 24, 1-24; reprinted at 298-313 in D. M. Stenton (ed.), Preparatory to Anglo­ Saxon England (Oxford, 1970) [cited from the reprint). Stenton, F. M., 1945, 'The Scandinavian colonies in England and Normandy', Trans. of the Royal Historical Soc., 4th ser. 27, 1-12; reprinted at 335-45 in D. M. Stenton (ed.), Preparatory to Anglo-Saxon England (Oxford, 1970) [cited from the reprint). Stenton, F. M., 1969, The Free Peasantry of the Northern Danelaw (Oxford) . Stenton, F. M., 1971, Anglo-Saxon England, 3rd ed. (Oxford). Stevenson, W. H. (ed.), 1959, Asser's Life of King Alfred (Oxford). Stocker, D., 2000, 'Monuments and merchants: irregularities in the distribution of stone sculpture in Lincolnshire and Yorkshire in the tenth century', 179-212 in Hadley and Richards 2000. Townend, M., 1998, English Place-Names in Skaldic Verse (Nottingham). Townend, M., 2000, 'Viking age England as a bilingual society', 89-105 in Hadley and Richards 2000. Unwin, T., 1988, 'Towards a model of Anglo-Scandinavian rural settlement in England', 77-98 in D. Hooke (ed.), Anglo-Saxon Settlements (Oxford). Wainwright, F. T., 1962, Archaeology and Place-Names and History. An Essay on Problems of Co­ ordination (London).

PLACE-NAMES AND HISTORY

43 1

Wainwright, F. T., 1975, Scandinavian England (Chichester). Watts, V., 1988-9, 'Scandinavian settlement-names in county Durham', Nomina, 12, 17-63. Watts, V., 2002, 'Medieval field-names in two South Durham townships', Nomina, 25, 53-64. Whitelock, D., 1937-45, 'The conversion of the eastern Danelaw', Saga-Book of the Viking Soc., 12,159-76. Whitelock, D., 1959, 'The dealings of the kings of England with Northumbria in the tenth and eleventh centuries', 70-88 in P. Clemoes (ed.), The Anglo-Saxons. Studies in some Aspects of their History and Culture (London). Whitelock, D. (trans.), 1961, The Anglo-Saxon Chronicle (London). Williams, A., 1996, 'The Vikings in Essex, 871-917', Essex Archaeology and History, 27, 92-10I. Williamson, T., 1993, The Origins of Norfolk (Manchester). Wilson, D. M., 1976, 'The Scandinavians in England', 393-403 in D. M. Wilson (ed.), The Archaeology of Anglo-Saxon England (Cambridge). Wormald, C. P., 1982a, 'The ninth century', 132-57 in j. Campbell et aI., The Anglo-Saxons (Oxford). Wormald, C. P., 1982b, 'Scandinavian settlement', 162-3 in j. Campbell et aI., The Anglo-Saxons (Oxford). Wormald, C. P., 1982C, 'Viking studies: whence and whither?', 128-53 in R. T. Farrell (ed.), The Vikings (London).

Dr Lesley Abrams, Balliol College, Oxford OXI }B/, UK lesley [email protected] Dr David Parsons, School of English Studies, University of Nottingham, Nottingham UK [email protected]

NG7 ZRD,

LAW AND LANDSCAPE

By

ANNETTE HOFF

The principle objective of my research project 'Law and Landscape' was to use the agricultural information in the medieval Danish provincial laws to produce an overall view of the organization and conditions of agriculture in the late Viking Period and Early-medieval times. 1 The four Danish provincial laws were written down by clerics in the period from II50 to 1250. Detailed analysis of them, however, shows that they contain strata of earlier, orally transmitted legal material from the late Viking Period of around the year lOOO. The law codes regulate many of the practical affairs of the agricultural communities, and consequently are able to provide extensive insights into the actual reality that lay behind their creation. By analysing the procedures in the Danish law codes it has proved possible to distinguish between earlier and later layers in the legal texts. The earliest procedure was based upon customary right, meaning that what had always been done was the law. This procedure meant in practice that, in the event of a crime, with sworn oaths of denial from three, six or twelve men, one could deny responsibility for a misdeed committed; the ordeal was used as evidence of one's innocence; and collective family responsibility lived on in the system of feud, by which everybody in a family was responsible for avenging the killing of a family member. The introduction of Christianity into Denmark was accompanied by a quite different, well-developed legal system - one that was based upon equity, personal responsibility, and the use of direct evidence. Canon Law was much superior to the local Danish customary law in its structure, and the clerics, who took the trouble to have the old customary law written down for the first time in the form of the provincial laws, contributed to the insinuation of the principles of Canon Law into the solution of many of the secular cases in the provincial laws. By separating the earlier customary law's procedures from those of the newly introduced Canon Law, it is possible to identify the earlier strata in the legal code. Since both earlier and later layers can also be found in the regulation of agricultural matters such as fencing, the maintenance of roadways, the use of forests and grazing areas, etc., it is possible to trace a chronological development in the organization of agricultural practices. Within the research project Law and Landscape, it has thus been possible to compare the earlier and later layers of agrarian law with the evidence of archaeology and agricultural history from the period 900-1200 with the aim of producing a 1 Annerte Hoff, Lov og Landskab:o Landskabslovenes Bidrag til ForstdeLsen af Landbrugs- og Landskabsud­ viklingen i Danmark ca. 900- I2,)O (Arhus, 1997).

434

LAND, SEA AND HOME

comprehensive understanding of the agrarian provisions of the provincial laws. This can contribute, in turn, to a fuller view of the development of agriculture and the landscape in the late Viking Period and early medieval times in Denmark. Comparative legislation from Sweden, England, Ireland and the early Frankish law codes has also been consulted. It has proved possible to demonstrate a close relationship between the Danish and Swedish sources, while it has not been possible to show any such relationship in the remainder of northern Europe. I shall return to the comparison of English and Danish laws later. Altogether, then, the objective was to reveal the actual reality that lay behind the production of the various laws, and subsequently to find out how agricultural conditions developed during this period. I have analysed all of the sections of the law that regulate the various elements of the culturally controlled landscape of the time: the village with its farms, the cultivated land with its fields and meadows, and the uncultivated land with its pastures, forests and tracks. In connexion with the analysis of the clauses on farm buildings, it has been possible, for instance, to demonstrate an agreement between the description of dwelling houses in the law codes and the houses known from archaeological excavations. I have, however, also been able to provide new explanations of certain key archaeological building features - in Old Danish the so-called la!garth (sheltered enclosures) and fa!garth (paddocks). It has also been possible to achieve a new perception of the use of the farm toft - the area within which the farm buildings were enclosed - as analyses of the legal provisions concerning the toft demonstrated that in the Viking Period the toft was so large that it functioned as a home field for the farm. This is not a revelation for English specialists, as it has long been known that the large tofts in the Danelaw were used as home fields. Another topic which the research project was able to improve understanding of was a road-term that is known in both England and Denmark ­ via regia or ha!rstrata! - the use of which in Denmark has also been revealed. To select just one example of the research method here, I shall focus on a particular detail within the investigation of the cultivation systems in the Danish provincial laws.

The cultivation systems in the Danish law codes A very pertinent example of the chronological disjunction between the earlier and the later legal procedures was revealed by the study of the provisions concerning fencing duties in the Jutish Law of 1241. This code contains two clauses on the fencing of the fields, III 57 and III 58, and the first thing that one notices is that the duty of fencing is apportioned to the individual village residents in accordance with two different land-valuation systems that could not both have been in use in the same village at the same date. One could hypothetically entertain the idea that these two clauses were intended to cover geographically distinct areas, each of which had its own system of land-valuation. In one clause, III 57, where the village land and fence duty is assigned according to the hide (Latin mansus), the code states that every man should enclose his own property in a straight line: i.e. measured out from the system of hidage Hwar man seal siin garth ga!rtha!, swo sum egha!r alla! giua! ia til, oc swo sum han a i by, oc i bol

LA W AND LANDSCAPE

435

eefteer reet reeep.2 This has to be understood as meaning that each individual farmer of the village was to fence his fields separately, and there is no reference in the code to the fields being part of a larger common property. One fenced one's own fields, and thus was oneself responsible for impounding other farmers' animals that wandered on to the fields. 3 The following clause, III 58, has the fence duty assigned according to the 'gold value'. Uvangs garth geertheer hwar man swo sum han haueer gullz withning. 4 Here it is a matter of fences that the farmers as a group placed around all of the fields of the village, which were divided into strips. We know that the gold value as the fixed measure of the size of a property was introduced in western Denmark during the mid­ I2th century, where it superseded the old system of hidation. The property valuation in this clause therefore leads one to suspect that this is a later legislative element. If one did not fulfil one's fence duty, and neglected to maintain one's own share of the common fence, the law explicitly denies one the right to defend oneself against having broken the law by the old oath of denial: /En of eengi man keennees with theen garth theer scathee wllee, tha sculee markee meen repee garthee oc wittee hweem theen garth h0rthee til oc theen theer reeeet reep falleer a, han hauee eei wald til at bivthee logh num gialdee swo witee sum f0rree war witheer lauth. s It is directly stipulated that one may not use the old process, but one should pay one's fines, which were agreed by all. The old, formal method of giving evidence was disallowed in favour of a more regular process: the payment of fines if one had not fulfilled the fencing duty.6 For various reasons, these two clauses cannot have represented current law in the same village at the same time. To begin with, the fence duty was assessed according to different land-valuation systems in the two separate clauses, and it was not practically possible to have two different valuation systems in the same settlement at the same time as such systems were also the basis for the apportioning of various rights and duties within the working area of the village. Furthermore, the fencing procedures in the two clauses represent two different systems of cultivation that were not directly compatible with one another. The fencing procedure of the first clause, which was based upon the old land­ valuation system, reflects an individual system of cultivation, under which each separate farmer in the village enclosed and protected his own particular fields - we are told that each man should fence his farm: in other words should himself take responsibility for the enclosure of his own fields. The other clause, with the newer legal procedure and the fence duty distributed according to the newer system of land­ valuation, implies, by contrast, that all the farmers together were to place common fences around all of the fields of the village. 2 'Every man must fence his yard in such a way that all owners can agree to, and according to what he owns in the village or homestead, along a straight line': Johannes Brondum·Nielsen et aI., Danmarks gamle Landskabslove med Kirkelovene, 8 vols. (Copenhagen, 1932-61), II, 473 . 1. Hoff, op. cit. in note I, 176. 4 'Every man will fence the outfield according to his gold value': ed. cit. in note 2, II, 476. 5 'But if no man will admit responsibility for the fence which is the cause of the damage, the owner of the field must rope the fence and so determine whose fence it is; and the man who is identified by the straight line will have no recourse to swear by oath, but must pay the compensation that was previously determined and make good the loss.' 6 Hoff, op. cit. in note I , 191.

43 6

LAND, SEA AND HOME

In the two clauses of the Jutish Law dealing with fencing duties we have thus identified two chronologically distinct layers within the law code that were not both current law in the same village at the same date but which must reflect a historical development in the organization of the cultivation systems. However, we could well imagine that in simple geographical terms there were certain parts of Denmark that continued with the earlier system if the natural conditions made it preferable. Closer inspection of the other clauses in the Danish provincial laws dealing with fence duty and cultivation reinforce the view of a chronological development from a simpler, individual system of cultivation, to a gradually more organized farming community represented by the open-field system. Furthermore, investigations of the legal provisions concerning the use of the farm toft in the late Viking Period have proved to be associable with the accounts of the systems of cultivation, in that it is possible to show that the toft was originally the home field of the farm, a man's personal property, which, we have to suppose, he himself was to protect by fencing. Only in the early post-Viking Middle Ages do we find, in the laws, that the utltendejord (out-field) of the village - the land lying outside of the village itself ­ was cultivated collectively, probably as the first embryonic sign of communal farming. The earliest evidence of open-field agriculture in the law codes can be dated to around the beginning of the 12th century, although in its incipient stages this system was quite simple in organization and the demands of communal cultivation were still only vague. Not until around the middle of the 13th century do we see the classical features of the fully developed open-field system: collective fencing duties in the form of the common fences that enclosed all of the fields of the village, communal grazing of the stubble in all of the fields of the village after harvest, compulsory cultivation in specific fields (in the sense that only spring crops could be sown in one field, winter crops in another, while the third lay fallow), fixed sowing and harvesting times, and finally a collective right and responsibility to catch one another's beasts in all parts of the open field if they broke in through the fences and started to eat the corn? Cultivation systems in the English law codes It is quite remarkable in this respect, that one of the English legal sources, Ine's Law of 688-94, also includes two clauses dealing with fencing duties that appear to be reminiscent of the Danish examples. Chapters 40 and 42 of Ine's Law, like the two significantly later clauses in the Jutish Law, provide information about the fencing duties of the farmers. This is unusual in itself, as the English law codes concern themselves with practical aspects of agriculture only to a very minor degree. Ine's Law 40 states that a farmer's property, woroig, is to be fenced both summer and winter. If it is not fenced, and his neighbour's beast gets in through an opening in the fence which he himself has left there, he has no claim upon the animal but must himself bear the damage the beast may cause: Ceorles woroig sceal beon wintres (and)

7

Hoff, op. cit. in note

I,

186.

LAW AND LANDSCAPE

437

sumeres betyned. Cif he bio untyned, (and) reco his neahgebures ceap in on his agen geat, nah he a!t pam ceape nan wuht . .. 8 These details are very similar to what Jutish Law III 57 stipulates concerning fencing duty. There has been some discussion of what this clause really means in England, with Fox, for instance, proposing the theory that woroig might refer to an individually fenced farm, implying that each separate farmer fenced in the fields of his own property. Chapter 42 of Ine's Law deals with the situation that farmers might have a common meadow or some other partible land (gedalland), and that it could be the case that some had constructed their share of the fence and others had not: Cif ceorlas gcerstun hcebben gemcenne oooe oper gedalland to tynanne (and) hcebben sume getyned hiora da!l, sume ncebben ... 9 There is no evidence that common land was to be cultivated in a strip system, and both Fox and Finberg advocate caution over the interpretation of this provision. How far Ine's chapter 42 might be a later stage of legislation one cannot immediately determine, but there cannot be any doubt that the fence duty reflects some form of communal cultivation or other. The degree to which there might also be a direct relationship between the legal provisions concerning fencing duties in the early English laws of Ine of the late 7th century and the much later Jutish Law of 1241 is something that after more detailed reflection one must be very doubtful about. Both laws were produced in an agricultural society, where it was vitally important to fence one's fields against the intrusion of animals. This is the reason why the law codes have regulations for fencing duties. To demonstrate a direct relationship between these two codes would require a higher level of agreement between them, whereas in fact all that they have in common is that they both deal with the construction of fences, in relation to what we have to assume were two different systems of cultivation. The diction and the substance of the laws are so different, that there are no grounds to infer that they are directly related. THE RELATIONSHIP BETWEEN THE ENGLISH AND THE DANISH LAWS

If one is to look more closely at the relationship between the English and Danish law codes one has first to recognize that the English sources are different from the Danish and Norwegian provincial laws in several crucial ways. Firstly, they were written down centuries before the Scandinavian law books. The Kentish laws of King JEthelberht date from the period between 587 and 616, and thus were first committed to writing down about six centuries before the Scandinavian law books. The second critical difference is that what we have from England are primarily royal decrees and stipulations that were promulgated as the individual kings assumed the throne rather than regular collections of law. Thirdly, the length of the so-called lawcode varies hugely, and the copying of the laws of one king by the next is frequent. to 8 'A ceorl's land is to be enclosed summer and winter. It is unfenced, (and) his neighbour's beast comes through the opening he is responsible for, he has no claim on that beast': F. L. Attenborough, The Laws of the Earliest English Kings (Cambridge, 1922), 49. 9 'If ceorls have a common meadow or some other shareland to fence, (and) some have fenced their share and som e have not .. .' : ed. cit. in note 8, ibid. \0 Hoff, op. cit. in note I, 31.

43 8

LAND, SEA AND HOME

All of the Anglo-Saxon laws are linked to the name of a specific king, and thus probably were only valid law during his reign irrespective of how short it was. This means that we are faced with inherited Anglo-Saxon law, albeit in many cases expanded with ecclesiastical ordinances. Since England consisted of a number of separate kingdoms down to the loth century, the fragments of law that survive were not valid law over the whole of England, but only where the king in question was ruling, e.g. in Kent (iEthelberht, Hlothere or Wihtred's Laws) or Wessex (Ine or Alfred's Laws). The Danish kings, by contrast, were not directly involved in the writing of the law codes as the king did not function as the law-giver in the Early Middle Ages. It was the good old law, the law transmitted at thing-moots in oral form, which, with a lot of corrections from Canon Law, was written down by clerics from the middle of the 12th century onwards. Not before the 13th century do we encounter the king as la w-giver in particular cases - for instance the peace legislation, attempting to extinguish the old family feuds. l l Finally the English law codes have diverse provisions concerning payments and taxation, and even the earliest laws reveal a manifestly hierarchical society with great differences in terms of rights and responsibilities, in respect, for instance, of the king and the Church. All of the law codes, even the earliest ones from the 7th century, bear the stamp of the introduction of Christianity, and some of the principles of Canon Law can be seen to be at work, in relation, for instance, to the peace legislation. The English law codes thus in many ways reflect a more developed legal situation in key areas, despite their great age, than the Danish ones. 12 Yet another crucial difference from the Danish law books is that the English laws include provisions regulating the general daily requirements involving animals, fields, grazing, roads and bridges to a less marked degree than the Danish ones. Only a few of the English laws stand out in this way by including provisions of general agricultural character, such as the conditions of cultivation, animal husbandry or use of woodland, although Ine's Law of c. 690 should be specially noted in this regard, as it - as already noted - does contain remarkably detailed information about practical agricultural matters. If, however, one goes in and makes a thorough analysis of the contents of the code and compares it with what we have from Denmark, it is nevertheless clear that there are great differences between Danish and English law. I have been through all of the English laws for their agrarian material, and there are specific cases where comparison with the Danish provincial laws shows a certain agreement in content. Three selected examples will be given here. The first example concerns trading in domesticated animals. Ine's Law chapter 35,1 states that when cattle were traded there could be a guarantor present Cif mon to pam men feoh geteme, oe his weer ooswaren heefdeY If, during a subsequent legal case about a stolen beast, this man denies that a trade took place he must pay a fine corresponding to the value of the animal(s}. The Danish Erik's Sjxlland Law of 1250, in chapter III 39, also states that there ought to be a guarantor present during trading of farm 11 12 I.l

Hoff, op. cit. in note I, 325.

Hoff, op. cit. in note I, 3I.

'If one make payment to a man who has confirmed his pledge': cd. cit. in note 8,47.

LAW AND LANDSCAPE

439

animals, although the case from the English law, anticipating that the guarantor may deny that the trade rook place, is unknown in the Danish legislation Kan swa wamE at noktEr man stEL andrtE tEnti hors tELLtEr mot, oc hautEr hin win tiL oc witntE, oc wiL hin sithtEn StEcttE han um tEr han fic af, tha gangtE the samtE men tEr oc hin tEr hans withtEr wartE oc witnd! thet at han k0pttE thet af hanum, oc hin tEr hans win war, gangtE na!sttE man hanum at swtErid! mtEth toLf men, at thet tEr hans k0p.14 The second example concerns the women's right (and obligation) to carry the keys of the farm. In the Danish King Valdemar's Sjxlland Law I sr, we learn that the woman who was unable to carry the keys of the house should be regarded as ill The h0sfrw callum wi siwka wara ther ey ma ganga mz sinum nikLum oc retha j synum heskepo.15 There is more information in the later Sjxlland Law of Erik, that not only the husband but also the housewife and the bride could have locked chests or coffers that no one else was allowed to open. The English law II Canute also has information about the woman as key-bearer, as chapter II 76,I describes in detail which keys the woman carried, because it is her duty to guard the keys of the following - her storeroom and her chest and her cupboard buton hit under OtES wifes ctEgLocan gebroht wtEre, sy heo cLtEne. 16 That women of the Viking Period carried keys at the girdle is also known to us from archaeological excavations in both Denmark and England, and the various legal provisions thus reflect this practice. It does not seem necessary for there to have been any relationship between the English and the Danish laws about the woman and her keys, since all they have in common is the fact that it was primarily the woman who kept the keys of the house. The actual contents of the provisions in the various laws do not immediately reveal any connexion between them. Finally, as the third example, we have the fencing provisions already described in the English Laws of Ine and the Jutish Law from Denmark. Here again, the actual contents do not argue for any connexion between Danish and English law; merely that both law codes were concerned with the fundamental issue for agriculture, that farmers had to fence their fields in order to protect them from intruding animals. This on its own does not indicate, as far as I can tell, that there must have been any relationship between the two law codes, which otherwise were written down six centuries apart. We might find a different basis for comparing English and Danish law in the laws of Cnut the Great from the early decades of the IIth century. Cnut was king of both England and Denmark, and the surviving laws of his reign were of course also in force in the colonized area of the Danelaw. It is consequently tempting to assume that these law codes had their ultimate origins in Danish law. It would be reasonable to expect that in this area of England, where place-names and many linguistic features bear the 14 'It may be that some man sell another a horse or some cattle, and will bring a friend and witness along, and the other will subsequently dispute with him about what he received, then the same men and the one who was his surety will go and testify that the other bought this of him, and the one who was his friend should stand beside him and swear as one of twelve that this is what was purchased': ed. cit. in note 2, Y, 295 . 15 'The housewife we call sick who cannot carry her keys and maintain her housekeeping': ed. cit. in note 2., III, 38. 16 'If it had not been subjected to the woman's lock and key, she is innocent': A. J. Robertson, The Laws of the Kings of England from Edmund to Henry I(Cambridge, 1925),215 .

440

LAND, SEA AND HOME

marks of Danish overlordship, one would also find reflections of an originally Danish legal order. The Danish professor of legal history, Ole Fenger, has examined Cnut the Great's laws in order to see if there were relicts of Danish or Scandinavian law to be found in the English legislation. After going through all the clauses of I Canute and II Canute, Fenger was able to conclude that, with the exception of two paragraphs, all of the provisions in Cnut's laws had predecessors or models in earlier Anglo-Saxon law, while Danish law had not had any influence at allY Furthermore Patrick Wormald has shown, that Cnut's laws were probably written down by a single author, Archbishop Wulfstan of York, who also wrote most of .lEtheirced's legislation at the end of the loth century and beginning of the 1 nh. 18 There is, however, one more legal source that could be of relevance for a comparison of the two country's laws. There is a relatively short law text for the retinue of the Danish king, known as Vederlov (the reciprocal law), which was probably written down around II82. 19 According to the evidence of the text itself, this is apparently a record of the old court law from the period of Cnut the Great, and it is supposed to have been in force both in England and Denmark. If, however, we examine the contents of the text in more detail, we see that it is a mixture of a law code and a historical account of the creation of the text. It thus relates that the text is supposed to have been issued by Cnut the Great for his retinue - called the tingLith­ and that the king sought the help of two men from the Danish island of Sjcelland, 0pe Snialle and his son .lEskil, to produce it. Earlier Danish scholars assumed that the Vederlov self-evidently was a medieval copy of previous kings' court laws going back to the Viking Period, but more recent research, by Niels Lund for instance, has raised a large question mark over this.20 I have attempted to make a direct comparison of the two law codes, Cnut the Great's public law II Canute and the short Danish Vederlov. The Danish Vederlov really contains only six clauses, while II Canute has 84. By examining some of the paragraphs here one can make a substantive comparison of the two collections. The first clause of the Danish law code describes how the king and other powerful men who have a retinue should be loyal and kind to their men, who reciprocally should give fidelity and service to their lords, and be prepared for all his commands. This formulation cannot be found in the English law code, but §31 of II Canute states that every lord shall be personally responsible as surety for the men of his own household. 21 The clause in Cnut's law is not directed specifically at the men of the retinue, but probably refers to all in a lord's service, and we do not find, as we do in the Danish text, that these are to give their lord their loyalty and service. Both provisions deal with subordinates' relationship with their lord, but the actual contents have nothing to do with one another. Ole Fenger, 'The Danelaw and the Danish Law', Scandinavian Studs . Law, 16 (1972), 83-96, at p. 96. Patrick Wormald, The Making of English Law, King Alfred to the Twelfth Century. Vol. 1: Legislation and its Limits (Oxford, 1999),346. 19 Ludvig Holberg, Dansk Rigslovgivning: Forholdet mel/em Vederlagslov og Rigslov. Rigslovene i Perioden 1241-1282 (Copenhagen, 1889),250. 20 Niels Lund, De H.xrger og Br.xnder: Danmark og England i Vikingetiden (Copenhagen, 1993), 183 . 21 '7 h.xbbe .xlc hlaford his hiredmen on his agenum borhge': ed. cit. in note 16, 193. 17

18

LAW AND LANDSCAPE

441

Paragraph 2 in the Vederlov deals with the situation in which a member of the retinue is a traitor, and betrays his lord with evil counsel, through which he condemns himself and all that he owns. 22 Cnut's law also comprises a chapter about treachery, but it is a somewhat different voice we encounter here: ... Be oam fJe fliho fram his Laforde. And oe man oe fleo fram his hLaforde oooe fram his geferan for his yrhoe, sy hit on scypfyrde, sy hit on Landfyrde, ooLie he ealles oees oe he age 7 his agenes feores, 7 fo se hLaford to oam eethon 7 to his Lande oe he him rer seaLde. 23 While the Danish law code deals with treachery involving evil counsel, in other words a planned betrayal of one's lord, the English text talks only of a man abandoning his duty. Both offences are punished by the loss of life and property, but the actual matter once again is different. We should finally mention §5 in the Danish Vederlov, which states that a member of the retinue who strikes or wounds other members of the retinue is to be driven out of the king's court in disgrace, and quit all the lands which Cnut was king over. 24 II Canute §61 deals with a similar situation: Griobryce. Gyf hwa on fyrde griobryce fuLwyrce, ooLie liues oooon weregyLdes. Gyf he samwyrce, bete be oam oe seo deed sy.25 While the Danish law code thus decrees lose of freedom for the king's man who either strikes or wounds any other members of the retinue, Cnut's law makes the punishment of a comparable offence committed by a member of the army proportion­ ate to the violence of the misdeed. If it was only a minor offence, it could be compensated for according to its actual character, but the Danish retainer forfeited his life irrespective of whether it were a minor or a major act. Once again, the two law texts diverge. Altogether, we can conclude that the two law texts both deal with the relationship between a lord and his subordinates - in the Danish case, specifically that between a king and his retainers. To specify the law between these two parties was entirely natural in a society in which the loyalty between a lord and his subordinates was one of the basic foundations upon which their relationship was based. We can therefore also expect that in a social structure like this, there will be legislation that regulates their mutual relations in close detail. I am therefore of the opinion that these two law codes emerged independently of one another, in two societies which had the same fundamental need to regulate the relationship between a lord and his subordinates, and a critical reading of the texts likewise provides no grounds to believe that there was any real relationship between them. As a result, we have to come to the conclusion that as far as English and Danish legislation in the late Viking Period and the beginning of the High Middle Ages was 22 'Of anncen hendir awotha oc uska?pite troswikere at wortha oc judaswerk at winne meth ilt rath gen herrce sinum, tha hawer han sik sielwan forgiort oc alt thet han a': J. L. A. Kolderup·Rosenvinge, Danske Gaardsretter og Stadsretter med Indledende Bema?rkninger (Copenhagen, 1827),4. 23 'Concerning the man who deserts his lord. And the man who, through cowardice, deserts his lord or his comrades on an expedition, either by sea or by land, shall lose all that he possesses and his own life, and the lord will take those goods and his land, which he formerly gave him': ed. cit. in note 16,215. 24 'Off annar brydir i laghit meth hog ellir meth saar, tha skal han wrakas aff konungs garthe meth nitiings orth oc fly al the land ther knut war konung iwer': ed. and loc. cit. in note 22. 2S 'Breach of the peace. If anyone is guilty of a capital deed of violence while serving in the army, he shall lose his life or his wergild. If he is guilty of a minor deed of violence, he shall make amends according to the nature of the deed': ed. cit. in note 16,205.

442

LAND, SEA AND HOME

concerned, one cannot demonstrate any connexion between the legal records of the two countries. They were based upon two fundamentally different legal systems, although both were nonetheless strongly influenced by Canon Law. The fact that the languages and material culture of the two countries nevertheless had much in common does, however, bear witness to the fact that there was a connexion between these two countries in the later decades of the Viking Period. Dr Annette Hoff, Horsens Museum, Sundvej rA, DK 8700 Horsens, Denmark [email protected]

CHANGING WEAVING STYLES AND

FABRIC TYPES: THE SCANDINAVIAN

INFLUENCE

By

PHILIPPA

A.

HENRY

Increasingly, over the last two decades, textile studies have been utilized in archaeology to assess cultural affinities and settlement patterns. This is particularly pertinent to the Viking Period in Britain and Ireland, where excavated textiles reveal several changes in weaving styles and fabric type, as well as differences between regions. Analysis indicates that Scandinavian influence is evident in Scotland and to some extent Ireland, but the evidence from England is not so clear-cut as the textiles appear to parallel those of the Frisian region of continental Europe and of earlier Anglo-Saxon fabrics. The discussion following this overview of the textiles from Britain and Ireland aims to place them in their cultural context. Although linen, wool and silk fabrics, encompassing a wide range of weave types, have been excavated from Viking-period sites, the main concentration in this article is on wool twills as this group shows the greatest variation in distribution, based on fleece type, spin direction, weave type and quality. The emphasis is on the textiles from Scotland and England where the majority of the textiles have been published. The textiles from Ireland await publication; the information regarding these fabrics is thus less detailed as it is of necessity based on general discussion documents. In order to further illustrate the Scandinavian influence on the textiles from Britain and Ireland, two specialist wool fabric types will be discussed, shaggy pile woven textiles and nalebinding (a form of knotless netting) fabrics. In addition, an important collection of silk head-dresses from England and Ireland are examined. Before presenting the evidence a brief outline of the technical considerations necessary for the accurate description of the textiles is presented. This is followed by an overview of the textiles, and the report concludes with a discussion on the cultural implications of the evidence.

TECHNICAL CONSIDERATIONS

Any discussion relating to the fabric structure requires some technical explana­ tions based on spin direction and weave type; these two elements, along with fleece type, form the base of this article. Wool textiles are produced from either Z-spun or

444

LAND, SEA AND HOME

Spin direction. (a) Z-spin: clockwise spin; (b) S-spin: anticlockwise spin; (c) Z/Z-spun warp and weft; (d) Z/S spin: Z-spun warp, S-spun weft; (e) SIS spin: S-spun warp and weft. Warp = vertical yarn, Weft = horizontal yarn. FIGURE 1.

~

"1

S-spun yarn which is occasionally plied, and in the case of silk yarn can be left unspun. Z-spun yarn is spun in a clockwise direction and S-spun yarn in a counter-clockwise one (Fig. la-b). The terminology refers to the direction of the twist, i.e. Z-spun yarn corresponds to the central stroke of the letter Z, and S-spun yarn to the central stroke of the letter S. Spin direction is an essential aspect of the characterization of textiles as the direction of spin in the yarns utilized produces different weave effects, including spin-patterned weaves, which create fabrics of varying appearance. The wool textiles discussed below feature ZZ- and ZS-spun yarns (Fig. I ), with the silk samples displaying a combination of Z- and S-spun yarns and fibre that is left unspun. Different effects in wool textiles can also be achieved by the way the wool is prepared and the tightness of the spin, which is very evident in the textiles examined in this paper. There are two principal weave types in Viking-period textiles: tabby and twill (Fig. 2), tabby being the simplest and the most widespread. Twills are subdivided into 212 and 2iI which relates to the number of warp threads the weft thread passes over and under, giving distinctive patterns. Twills with a 212 binding, also known as four­ shed twills, are produced by passing the weft thread over and under two warp threads, with each row stepped to one side of the row above. The same principle is employed for the production of twills with a 2iI binding, three-shed twills, only here the weft thread is passed over two and under one warp thread. Twills found in Viking-period contexts, known as plain, diamond (also referred to as lozenge) and chevron (also known as herringbone), produced fabrics with diagonal lines, diamonds and chevrons across the width and length of the cloth. Diamond twill is produced by reversing the pattern in the warp and weft, and chevron twill is achieved by reversing the order in which the warp threads are attached to the loom heddle rods. Both diamond and chevron twills can have accurate or inaccurate meetings (even or broken twills) where

WEAVING STYLES AND FABRIC TYPES

445

Weave types. (a) Plain tabby; (b) 212 plain twill; (c) 2iI plain twill; (d) 2/2 broken chevron twill; (e) 212 chevron twill with accurate meetings; (f) 2iI chevron twill with accurate meetings; (g) 2iI diamond twill with accurate meetings; (h) 212 diamond twill with inaccurate meetings. Black squares indicate where weft threads pass over warp threads (after Wild).

FIGURE 2.

the pattern unit (that which produces each diamond or chevron) either meets the adjacent unit accurately or is stepped (Fig. 2). Textiles of differing appearance can also be created where the weave is balanced or unbalanced, both of which are represented in the cloth samples discussed below. A balanced weave is achieved where the number of warp and weft threads per centimetre is equal, and an unbalanced weave occurs where there are more warp threads per centimetre than weft threads or vice versa. As well as fabric structure, fleece type also has to be considered. Ryder has categorized the main fleece types of Viking-period Britain and Ireland as 'Hairy', 'Hairy Medium', 'Generalized Medium' and 'Medium', based on fibre diameter measurements. In Scotland, Ireland and York the main fleece types are 'Hairy' and 'Hairy Medium', while in the rest of England, 'Hairy Medium' and 'Generalized Medium', are the norm. In addition, the fleece from Scotland and Ireland has a higher incidence of pigmented fibre than that of England. The full range of fleece colours has been identified in Scotland - black, brown, grey and white; in Ireland shades of brown are the norm and in England the wool in the main has less pigment, resulting in lighter coloured, and white wool, which is often dyed. 1

J M. L. Ryder, 'Fleece evolution in domestic sheep', Nature, 204 (I964), 555-9; idem, 'The evolution of Scottish breeds of sheep', Scottish Stud., I2 (I968), I27-67; idem, 'Coat colour inheritance in Soay, Orkney and Shetland sheep', j. Zool., I73 (I974), 477-85; idem, 'Fleece changes in sheep', 2I5-29 in M. Jones and G. Dimbleby (eds.), The Environment and Man: The Iron Age to the Anglo-Saxon Period (BAR Brit. Ser., 87, Oxford, I98I); idem, 'Animal fibres and fleece types', I27-8 in A. MacGregor, Anglo-Scandinavian Finds from Lloyds Bank, Pavement and Other Sites (London, I983) ; idem, Sheep and Man (London, I983); P. Walton, Textiles, Cordage and Raw Fibre from I6-22 Coppergate (London, I989).

446

LAND, SEA AND HOME

THE TEXTILES

Twills Three distinct groups of twills can be identified from the fabrics excavated from Britain and Ireland: ZZ-spun 212 plain and diamond twills, ZS-spun broken diamond and chevron twills (212 and 2/r), and ZS-spun twills (mainly 212) where the tightness of spin and colour of the yarn playa major role in the appearance of the fabric. In addition, there are those twill textiles that have different elements from each group but do not sit comfortably in any of them. Group 1 There are two forms of ZZ-spun four-shed twills in particular that stand out as specific fabric types. The first is a slightly unbalanced plain twill of medium quality produced from combed tightly spun yarn for the warp and loosely spun yarn for the weft. The warp thread is regularly dyed blue and is darker than the weft, forming a pronounced diagonal (Fig. F)' This fabric is known as the Veka-type twill, after one of the sites on which it was found in western Norway.2 The second fabric type is a good quality broken diamond twill with a slightly unbalanced weave produced from tightly spun combed yarn, producing a fabric with a glossy sheen. The pattern unit is uniform in all the textiles of this type and when combined with thread count, creates a slightly elongated diamond. This distinctive fabric is known as the Birka-type twill, the name referring to a group of the textiles excavated from the cemetery at Birka where A. Geijer first recognized it as a specific fabric type (Fig. 3b-c).3 It would appear that the Veka-type twill excavated from a Viking-period grave at Kildonnan in the Isle of Eigg, and the Birka-type twills from Greenigoe, Orphir in Mainland, Orkney, and Kneep, Uig, Isle of Lewis, unknown in Scotland prior to the 9th century, represent Scandinavian fabric types (Figs. 3 and 4).4 Neither have been identified in the English textiles, although two ZZ-spun diamond twills from York are quite similar. 5 It is possible that amongst the ZZ-spun 212 plain and diamond twills from Dublin and Waterford at least some may fit into the group 1 category.6 Group 2 ZS-spun diamond and chevron twills with inaccurate meetings form the second group of textiles; the majority have 212 bindings, with a small number of 2/1 twills. They occur mainly in England, notably from York and London (Fig. 3d), forming the 2 L. Bender jorgensen, North European Textiles until AD 1000 (Aarhus, 1992). -' A. Geijer, 'The textile finds from Birka', Acta Archaeol., 50 (1980), 209- 22. 4 L. Bender jorgensen, 'Textile Remains: analysed at the National Museum of Antiquities of Scotland' (unpub!. rep., 1983); eadem, Prehistoric Scandinavian Textiles (Copenhagen, 1986); eadem, 'The textile remains', 165-8 in R. D. E. Welander, C. Batey and T. G. Cowie, 'A Viking burial from Kneep, Uig, Isle of Lewis', Proc. Soc. Antiq. Scotland, 177 (1987), 149-74; G. M. Crowfoot, 'Textiles from a Viking grave at Kildonnan, on the Isle of Eigg', Proc. Soc. Antiq . Scotland, 83 (1948- 9), 24-8. 5 Walton, op. cit. in note I, 333; J. Hedges, 'Textiles', 102-27 in MacGregor, op. cit. in note I, 104. 6 F. A. Pritchard, 'Aspects of the wool textiles from Viking Age Dublin', 93-104 in L. Bender jorgensen and E. Munksgaard (eds.), Archaeological Textiles in Northern Europe (Copenhagen, 1992); E. Wincott Heckett, 'Medieval textiles from Waterford City', 148-56 in G. jaacks and K. Tidow (eds.), Archaeological Textiles (Neumiinster,1994)·

WEAVING STYLES AND FABRIC TYPES

447

FIGURE 3. Fabric types. (a) Veka-type twill from Kildonnan; (b) Birka-type twill from Greenigoe; (c) Birka-type twill from Kneep; (d) Z/S diamond twill with inaccurate meetings from York. ( (a-c) courtesy of the National Museum of Scotland; (d) courtesy of York Archaeological Trust)

largest group for the wool patterned twills excavated from the Anglo-Scandinavian and Late Saxon levels of these two sites? Of the 37 patterned twills from York, 21 (57%) are broken twills produced from ZS-spun yarn. From London there are 15 7 Walton, op. cir. in nore I, 324-32; Hedges, op. cir. in nore 5, 106-13; A. Prirchard, 'Lare Saxon rexriles from rhe Ciry of London', Medieval Archaeol., 28 (1984),46-76.

44 8

LAND, SEA AND HOME

recorded patterned twills, 13 (87%) of which are broken twills woven with ZS-spun yarn. Lincoln and Winchester each have one recorded sample of this weave and spin combination dated to the Late Saxon Period (Fig. 4). However, the number of patterned weaves from both sites is small, and the paucity of this fabric type should thus not be seen as significant. A small group of this textile type has also been excavated in Ireland, but to date no ZS-spun patterned weaves have been excavated from Viking-period Scotland. These fabrics parallel the textiles from the Frisian areas of The Netherlands and Germany, and earlier Anglo-Saxon fabrics. 8 Group 3 ZS-Spun twills (mainly 212) with a tightly spun combed warp and soft lightly spun uncombed weft form the final discreet group. The yarn and fleece type play an essential role in the appearance of this fabric type, where the warp, which is dark brown, is nearly always darker than the weft and produced from a coarser fleece type; the warp is 'Hairy' and the weft 'Hairy Medium'. The tightly spun darker warp threads of this fulled fabric give a distinctive diagonal pattern to the twill, which with the softer weft yarn creates a slightly matted, semi-water-repellent fabric (Fig. sa). Samples of this type of fabric, known after contemporary Icelandic documentary sources as wadmal, have been excavated from the Biggings in Papa Stour, Shetland and York (d. Fig. 4). A similar fabric has been excavated from Waterford and may represent another sample of this fabric type. 9 These textiles have definite parallels throughout the North Atlantic region and continued to be produced into the 18th century. SPECIALIST FABRICS

Piled Cloth A total of ten piled fabrics have been excavated from England, Ireland, the Isle of Man and Scotland.1O Five fabrics with inserted pile have been excavated from settlements, four from York and one from Dublin. The remaining pile textiles have been excavated from Viking graves, three from a grave at Cronk Moar, Jurby, on the Isle of Man, and two from the grave at Kildonnan, Eigg (Fig. 4). Tabby forms the ground weave in three samples from York, and the fabrics from Kildonnan and Cronk Moar, with one sample from York and the Dublin textile having twill ground weaves. The form of ground weave is not vital, however, as it does not show on the outside face; both tabby and twill are common. The pile in all the samples is produced from loosely S-spun yarn and is woven in as the weft threads are inserted, in the textiles from Dublin, Kildonnan, and two of the fabrics from Cronk Moar, which is the normal way to produce piled fabric. In contrast, the York fabrics and one of the Walton, op. cit. in note 1,415. Walton, op. cit. in note I, 329; Wincott Heckett, op. cit. in note 6, 149; P. Walton Rogers, 'Textile, yarn and fibre from the Biggings', 194-202 in B. E. Crawford, The Biggings, Papa Stour, Shetland: the History and Excavation of a Royal Norwegian Farm (Edinburgh, 1999). 10 Walton, op. cit. in note I, 335-6; Bender j0rgensen, op. cit. in note 4, 3; Henshall, op. cit. in note 4, 15; Hedges, op. cit. in note 5, II3-I4; Pritchard, op. cit. in note 6,95-8; G. M. Crowfoot, 'The textiles', 80-3 in G. Bersu and D. M. Wilson, Three Viking Graves in the Isle of Man (London, 1966). 8

9

449

WEAVING STYLES AND FABRIC TYPES

Legend Veka-type twill Birka-type twill ZJS pattern twills Piled fabrics NAiebinding

Wadmal

Silk head-dresses

FIGURE 4.

Distribution of fabric types referred to in the text.

45 0

LAND, SEA AND HOME

FIGURE 5. Fabric types: (a) Wadmal from York; (b) pile fabric from Cronk Moar;

(c) ndlebinding sock from York; (d) silk head-dress from York. ( (a, c and d) courtesy of

York Archaeological Trust; (b) courtesy of Manx National Heritage)

WEAVING STYLES AND FABRIC TYPES

451

Insertion of pile and ndlebinding technique: (a) techniques of inserting pile: I. Woven in; 2. Darned in; 3. Darned in plitting weft thread (after Crowfoot). (b) Technique of ndlebinding (knotless netting) (after Bender Jorgensen).

FIGURE 6.

textiles from Cronk Moar have darned in pile (Fig. 6a). The darned in pile from the Cronk Moar sample was probably a repair as it is combined with woven in pile. The samples from York, though, show a definite departure from the norm. The pile in all of the textiles is the same colour and fleece type as the weft thread. Piled textiles are known throughout North-West Europe from the Migration Period through to the later Middle Ages, but in the Viking Period they were most prevalent in the North Atlantic region. NdLebinding Two examples of ndLebinding, a form of knotless netting which produces an elastic fabric (Fig. 6b), have been excavated from Britain and Ireland, one from York and one from Dublin (Fig. 4)Y As ndLebinding is a Scandinavian technique, virtually unknown outside the region, these two finds are significant. The sample from York has been interpreted as a sock (Fig. 5c); it is not possible to assign a function to the Dublin textile as the sample is too small.

Silk head-dresses Finally, good quality silk tabby fabric, with specific technical details and dimensions, has been excavated from three sites, York, Lincoln and Dublin. I2 The dimensions of these textiles and the complete example from York has led to the conclusion that the silk was used to create silk head-dresses, either joined at the back or worn as a scarf (Fig. 5d). York has three samples, Lincoln one and there are three from Dublin. In addition, Dublin has a number of lightweight wool textiles interpreted as the remains of caps. These head-dresses seem to be peculiar to these sites with no apparent parallels in North-West Europe. It is possible that the head­ dresses form a Hiberno-Norsel Anglo-Scandinavian tradition, although there is earlier and later evidence of head-dresses of a different style from Scandinavia, particularly Denmark. The technical details of the silk textiles are very similar and suggest a Walton, op. cit. in note I, 341- 5; Pritchard, op. cit. in note 6,102. Walton, op. cit. in note I, 360-7; A. Muthesius, in MacGregor, op. cit. in note I, 132-6; E. Heckert, 'Some Hiberno-Norse head coverings from Fishamble Street and St John's Lane, Dublin', Textile Hist., 18 (1987), 159-73; E. Heckert, 'Some silk and wool head·coverings from Viking Dublin: uses and origins - an enquiry', 85-96 in P. Walton and]. P. Wild, Textiles in Northern Archaeology (London, 1990); F. A. Pritchard, 'Silk braids and textiles of the Viking Age from Dublin', 141-61 in L. Bender Jergensen, B. Magnus and E. Munksgaard, Archaeological Textiles (Copenhagen, 1988). II

12

45 2

LAND, SEA AND HOME

common source, possibly from the silk manufacturing workshops of the Eastern Mediterranean region. Indeed, the silk from NO.5 Coppergate, York and that from Lincoln are so similar in technical detail that Muthesius has suggested that the two silks may have been cut from the same bolt of fabric.13 CULTURAL IMPLICATIONS

What then are the cultural implications of this group of textiles, and how do they add to our knowledge of Viking-period settlement in Britain and Ireland? The distribution of the Veka-type twill throughout western Norway, and its paucity elsewhere points to a Norwegian product. At the time of writing, there is no evidence for this twill outside Norway, except for the sample from Kildonnan.14 Much debate surrounds the original place of manufacture of the Birka-type twill. It has variously been associated with the pallium fresonicum, and with an origin in Syria, based on its similarities to a twill fabric type from Palmyra.15 There are, however, differences in the technical details of the Frisian cloth and the Palmyra twills, and the Birka-type twill: the former utilizes ZS-spun yarn and the latter yarn is spun in a ZZ direction. Bender J 0rgensen has rigorously examined the distribution of find-spots of this fabric type and has come to the conclusion that this too is a western Norwegian product. As well as the distribution of the Birka-type twill, Bender J0rgensen has based her conclusions on her findings that western Norway is the only region where all three qualities of this high quality fabric have been found. The best two qualities have been excavated from trading centres including Birka and Kaupang, and high status graves, such as Oseberg, Gokstad and at Birka, whereas the lower quality fabric is rarely found outside western Norway.16 It appears, therefore, that the Veka- and Birka-type twills from Kildonnan, Greenigoe and Kneep have their origin in Norway. As all three sites are graves, it is possible that the fabrics were manufactured in western Norway and were brought in with the settlers. There is also a less convincing possibility that the textiles were produced locally by Scandinavian settlers following the traditions of their homelands, although the Birka-type twill from Greenigoe is of a quality rarely seen outside western Norway. The Kneep example is an impression on a belt buckle indicating that it was used during the manufacturing process. If the buckle was produced in Norway, it further supports a Norwegian origin for the Birka-type twill. The ZS-spun broken diamond and chevron twills from England have parallels in early Anglo-Saxon graves. In addition, ZS 212 broken diamond twill is known from the Iron Age onwards in England, continuing through the Roman Period when ZS 212 broken chevron twills were added to the repertoire. During the Viking/Late Saxon Period this fabric type continued to be produced, with 21I examples becoming more

Muthesius, op. cit. in note 12, 132. Bender Jorgensen, op. cit. in note 2,138. 15 Bender Jorgensen, ibid.; Geijer, op. cit. in note 3, 212; A. Geijer, A History of Textile Art (London, 1979); M. Hoffmann, The Warp Weighted Loom (Oslo, 1974) . 16 Bender Jorgensen, op. cit. in note 2, 139- 40. \3

14

WEAVING STYLES AND FABRIC TYPES

453

common during the latter part of the period. I? A sample from Waterford parallels one from York which suggests that either this fabric type was also being produced in Ireland, or that similar fabrics were being traded in both centres. In addition, the similarities of the technical details between some of the York and London samples further suggests trade links. Distribution maps of find-spots reveal that this fabric-group was a common type of cloth in the Frisian areas of Germany and The Netherlands, where it is also known from earlier periods. The ZS-spun broken diamond and chevron twills of the Viking! Late Saxon period therefore seem to be continuing an earlier tradition that is replicated in Frisia. The predominance of 'Hairy' and 'Hairy Medium' fleece in the York textiles of this type, and the wool textiles in general, is, though, at variance with the textiles from other English sites. The wool available to the inhabitants of the town was therefore more akin to that of Scotland and Ireland, although with far more white fleece available for use, the wool resembling that of the modern Swaledale. Is The ZS-spun 212 twill termed wadmaL is known to have been produced in Iceland from the lIth century onwards and has been excavated from settlement sites from as far apart as Greenland, notably Garden under Sandet, York and Hedeby. This heavy 212 tweed-like fabric continued to be produced into the r8th century in Iceland where it was still woven on the warp-weighted loom. During the High and Late Middle Ages wadmaL was Iceland's main export, and it was also produced in Norway, the Faeroes, Orkney and Shetland. In Shetland it was used as a form of tax, as well as being imported into Norway.I9 This fabric therefore, appears to be a distinct cloth type of the North Atlantic islands. Its presence on Papa Stour is thus to be expected. The existence of a sample in York is more unusual, but was found with the ndLebinding sock, suggesting that it may represent the remains of clothes worn by a Scandinavian settler in York, or a traded commodity. The Hedeby sample almost certainly arrived in the town via trade. The piled textile from Dublin is not unusual in that Ireland and Iceland were known for their piled cloaks; the Dublin fabric has parallels from Heynes, Iceland. Similar piled fabric has also been excavated from Scandinavian contexts, including Birka, Kaupang, Lund and Hedeby.20 The finds from Cronk Moar and Kildonnan are harder to interpret as they show some differences in their method of manufacture, with both the textiles having areas where there is little or no pile. Analysis and reconstruction of the sample from Cronk Moar shows that the effect was created by varying the thickness of the pile yarn. 21 This may have been a deliberate design element, as the areas of sparse pile are too uniform to be accidental or the result of wear. The incidence of some darned in pile, only paralleled at York, was more likely to be the result of repairs, particularly as some of the darned in pile splits the warp thread (Fig. 6a). The Kildonnan textile is Walton, op. cit. in note 1,414-16. Walton, op. cit. in note I, 305; Ryder, in MacGregor op. cit. in note I, 127- 32. 19 Walton, op. cit. in note I, 329; Walton Rogers, op. cit. in note 9, 201; P. Walton Rogers, 'The raw materials of textiles from GUS - with a note on fragments of fleece and animal pelts from the same site', 66- 73 in J. Arnborg and H . C. Gull0v (eds.), Man, Culture and Environment in Ancient Greenland (Copenhagen, I998). 20 Walton, op. cit. in note I, 336; Hedges, op. cit. in note 5, I 14; Pritchard, op. cit. in note 6, 98. 21 Crowfoot, op. cit. in note 4, 24-8; Crowfoot, op. cit. in note IO, 80-3 . 17

18

454

LAND, SEA AND HOME

more difficult to interpret, and has no known parallels in Scandinavia. It is not possible to ascertain if the areas of sparse or no pile are part of the original design or if they result from heavy wear. The zone of the textile where the lack of pile occurs does not though show heavy matting, which is a feature of wear and thus would be an indicator of the pile having been worn away in antiquity. The difference in production techniques of these two textiles make it difficult to establish a place of origin. They may form parts of blankets or cloaks brought over from the Scandinavian homeland, traded goods from Ireland, or it is possible that they are of local manufacture. Wherever their place of origin, the use of piled fabric continues a tradition known from earlier periods in Scandinavia, Britain and Ireland. The York textiles are also interesting in their method of manufacture. Although piled fabric is known from earlier periods in England, the pile was normally longer and was woven in. The darning in of the pile in the York samples, which is interestingly more akin to Danish Bronze-age samples, may be a local tradition, or it may be a case of trying to replicate the piled fabrics of Ireland and not quite getting it right. 22 All four samples from York have darned in pile, suggesting that they were made in the same area, and may therefore be York products. The ndlebinding textiles from York and Dublin were almost certainly produced by a Scandinavian, either in the homelands or by a settler. The Dublin sample is too small to ascertain its function, but the sock from York leaves us in no doubt as to its use.23 The proximity of the sock to the piece of wadmal fabric indicates that the owner of both followed a strong Scandinavian tradition in style and preference of dress. Finally, silk head-dresses (instead of wool) appear to be an everyday fashion peculiar to the Scandinavian settlers in Dublin, York and Lincoln. This strengthens the evidence for the close contacts between Dublin and York, and highlights the link between York and Lincoln. Earlier and later examples of head-dresses, albeit of a slightly different style, from Scandinavian contexts, particularly from Denmark, also point to Scandinavian influence, adding to this hypothesis. 24 Whatever the cultural affinities and origin of the fabric, the occurrence of these textiles provides strong evidence for the expansion of trade networks instigated by Viking-period Scandinavian traders. Silk was of course known in England and Ireland prior to the Viking Period but this was normally high-quality patterned silk used by the upper echelons of society and in ecclesiastical contexts, and normally excavated from high status graves or ecclesiastical centres. The appearance of the plain tabby silk, and thus a more basic fabric, in and around town tenements points to a greater availability of silk to a wider social group. It would appear that the opening up of trade routes to the eastern Mediterranean created this opening and thus reinforces the influence of the Vikings on the fabric types being introduced into Britain and Ireland. When comparing the textiles from England, Ireland, the Isle of Man and Scotland, it has to be borne in mind that the evidence from Scotland and the Isle of 22 23 24

Walton, op. cit. in note I, 336.

Walton, op. cit. in note I, 341-5; Pritchard, op. cit. in note 6, 102.

Walton, op. cit. in note 1,375-6.

WEAVING STYLES AND FABRIC TYPES

455

Man is based on grave finds, and that from England and Ireland from settlements. In addition, much of the material excavated from Scandinavia is representative of grave finds, while that from North-West Europe is from settlements. When making comparisons of this nature it is important to take this into account as there may be differences in the textile used for burial and those which are used for everyday wear from settlement sites. The evidence presented above appears to be borne out, however, by the fact that the few textiles excavated from settlement sites in Scandinavia parallel the evidence from graves. The presence of ZS-spun broken twills in earlier Anglo-Saxon graves and Irish contexts suggests that these textiles were a common fabric in England and Ireland prior to the Viking Period.

CONCLUSIONS

It can be seen that through the careful analysis of excavated textiles from Viking­ period Britain and Ireland, together with an assessment of comparative material, cultural implications can be drawn, even with differing burial environments. The fabrics discussed above illustrate that the islands of the North Atlantic appear to have strong links to Norway and each other. England and, to a slightly lesser extent, Ireland appear to have continued to follow the traditions of their earlier textile types, and those produced in the Frisian areas of continental Europe. A small number of the textiles from Dublin, Waterford and York show Scandinavian influence, but perhaps not as much as would be expected. The fleece types prevalent in York are nearer to those of the North Atlantic region and Ireland. This suggests that there are differences in the availability of fleece type, or differing preferences between the Scandinavian­ influenced North and the Saxon-dominated southern regions of England. Preference may be a factor as the fleece types evident in the excavated textiles from Late Saxon Durham are closer to the rest of England in both type and proportion of excavated fabrics than to the York fleece types. Finally, the silk head-dresses illustrate a strong connexion in dress styles, and probably trade, between Hiberno-Norse Dublin and Anglo-Scandinavian York and Lincoln. It is thus hoped that the evidence presented here has succeeded in illustrating the importance of textiles as a resource to cultural studies in archaeology, and how, through careful analysis, they can make a valuable contribution to the study of Viking-period settlement in Britain and Ireland.

ACKNOWLEDG E MENTS

The author would like to acknowledge the generosity of Penelope Walton Rogers of Textile Research Associates for responding to my queries so willingly. In addition, without the work of Professor Lisa Bender j0fgensen, Grace Crowfoot, Frances Pritchard, Dr Michael Ryder, Penelope Walton Rogers and Elizabeth Wincott Heckett, this article would not have been possible. Permission to use textile photographs was granted by the National Museums of Scotland (Fig. p-c), Manx National Heritage (Fig. Sb), and the York Archaeological Trust

4S 6

LAND, SEA AND HOME

(Figs. 3d and sa, c-d). Finally, thanks to my sister, Diana Henry, for her assistance unravelling my original illustrations and presenting them in an understandable format.

Philippa A . Henry, Wolfson College, Oxford oxz 6UD, UK [email protected]

In

SAXON SHOES, VIKING SHEATHS?

CULTURAL IDENTITY IN ANGLO­

SCANDINAVIAN YORK

By

ESTHER CAMERON

and

QUITA MOULD

The largest quantity of Anglo-Scandinavian leatherwork to have been found in England was recovered from excavations at York during the 1970s. Over 20,000 fragments were found at Coppergate alone, of which 2,340 were fully recorded as finds, forming a remarkable collection now in process of publication by the York Archaeological Trust. Prior to these discoveries, leather of Late Saxon date had been found in relatively small caches and its study was therefore of limited scope. The quality and quantity of leatherwork from roth- and IIth-century contexts in Anglo­ Scandinavian York offers an unparalleled opportunity for research. 1 This has revealed evidence of change in craft technique and style among common artefact types ­ notably footwear and knife sheaths, the subject of this paper.

Sheaths of knives and seaxes (Fig. 1) Forty-two knife sheaths found in build-up deposits to the rear of the Coppergate tenements were found in contexts ranging in date from 930 to the earlyfmid-1Ith century, the majority of which were dated to the mid-loth century. In keeping with most knife sheaths of this period, all were of a type that encloses the handle as well as the blade. With few exceptions the evidence suggests that two types of knife sheath were available in roth-century York and that one of these, type B (Fig. Ib), dominated the market. This relatively new and developing type was characterized by a seam at the back using binding stitch (also called whipped stitch) and an accentuated profile. The type B sheath was already dominant in York by 930 and it is interesting to note that a partially preserved mineralized sheath from a 9th-froth-century Viking grave at Ballateare, Jurby, Isle of Man was also described as having a seam at the back. 2 Whether the origins of the type B sheath lay in Scandinavia or elsewhere, its fashion may have been spread through Viking influence as similar sheaths have been found in roth- and IIth-century contexts in London, Lincoln and Dublin.

1 Q. Mould, I. Ca rli sle and E. Cameron, Leather and Leather-working in Anglo-Scandinavian and Medieval York (The Archaeology of York 171J6, York, 2003) . 2 G. Bersu and D. M. Wilson, Three Viking Graves in the Isle of Man (Soc. Medi eval Archaeol. Mon . 5er., I, London, 1966),61.

45 8

LAND, SEA AND HOME

Sheaths of knives from Anglo­ Scandinavian York: (a) type A; (b) type B; (c) sheath from Parliament St, York (after Tweddle, op. cit. in note 4). (Copyright Institute of Archaeology, Oxford) FIGURE I.

The position of the seam is significant because knife sheaths had not previously been made in this way in the areas of early Anglo-Saxon settlement where sheaths were seamed along the edge of the blade using a different type of stitch. A few sheaths made according to Anglo-Saxon methods were also found at York and have been designated type A (Fig. ra). Type-A sheaths are characterized by:

A seam in tunnel-stitch along the cutting edge of the knife;

A suspension flap, full- or half-length;

A V-shaped notch at the mouth;

Suspension by a thong through a slit in the flap;

Decoration sub-divided into fields;

Profile of the blade outlined in the decoration.

SAXON SHOES, VIKING SHEATHS?

459

Only five type-A sheaths were found in loth-century York, but archaeological evidence from elsewhere in England shows that this type had a pedigree reaching back to the 7th century at least. A particularly fine example, although incomplete at the mouth end, from a 7th-century Saxon burial at Broomfield (Essex) displays all but two of the characteristics listed above. 3 Knife sheaths of the 8th and 9th centuries are astonishingly rare with a unique find from Parliament Street, York, thought to be of 9th-century date. 4 Despite the moulding of its spine and prolific decoration all the essential features of its construction link it with the Broomfield sheath. It is possible that the Broomfield, Parliament Street and type-A sheaths from Coppergate represent a continuous Saxon tradition of sheath-making spanning three centuries. The five 'A types' found in roth-century York, however, represent the tail end of that tradition among the sheaths of small knives, for they disappear very rapidly after this date and could be residual finds. Interestingly, this disappearance is not reflected immediately among the sheaths of large knives (sea xes) which, despite their more spectacular size and elaborate metal fittings, are essentially type-A sheaths. Their presence in eastern England in the 7th century, in burial assemblages such as those published by Lethbridge and Matthews, may have resulted from the impact of Rhineland trade. s Their popularity continued into the loth and lIth centuries, and decorated sheaths of seaxes from York include one which, in stylistic terms, could be relatively late (Fig. lC).6 Type-A sheaths from roth-fIlth-century contexts therefore cannot be residual in their entirety and it appears that large type-A sheaths continued to be made for some time after the adoption of the new style (type B) for the majority of smaller ones. Why might this be? An 8th-century metal fitting for a seax, found in the Thames,? and the presence of rivet holes and undecorated zones at the edges of the large sheaths, confirm that they were designed to support elaborate metal fittings. At a time when the blades of swords had undergone technical development and pattern-welding was in sharp decline, the blades of sea xes continued to be decorated, not only with pattern-welding but also with inlays of copper wire to an extent that made some of them unusable. s Moreover, while the sheaths of other knives had adapted to change, the sheaths of seaxes continued to be made in the old way, demonstrating a strong adherence to Saxon tradition. Two stone depictions of Saxon warriors (Repton and Middleton) show the sea x as a weapon linked with status and it seems probable, therefore, that the special

British Museum 1894.12-16.8. D. Tweddle, Finds from Parliament St and Other Sites in the City Centre (The Archaeology of York, 1]14, London, 1986),240-1, fig. 107. 5 T. C. Lethbridge, A Cemetery at Shudy Camps, Cambridgeshire (Cambridge, 1936); c. L. Matthews, 'The Anglo-Saxon cemetery at Marina Drive, Dunstable', Bedfordshire Archaeol. }., 1 (1962),25- 47. 6 Tweddle, op. cit. in note 4, 240. 7 British Museum 1869.6-10.1. • B. J. Gilmour, 'The Technology of Anglo-Saxon Edged Weapons', 2 vols. (unpub!. Ph.D. thesis, University of London, 1991); V. I. Evison, 'A decorated seax from the Thames at Keen Edge Ferry', Berkshire Archaeol. 3

4

}., 61 (1963-4),28-36.

460

LAND, SEA AND HOME

Construction 1 tunnel stitched sole

Construction 2 edge/flesh stitched sole

FIGURE 2.

(a) construction

Shoe-sole constructions found at Anglo-Scandinavian York:

using tunnel stitch; (b) construction 2 using an edge/flesh stitch.

(Drawn by Quita Mould)

I

symbolism which resulted in its long popularity could also explain the anachronism of its sheaths. 9

Shoes (Figs. 2-4) Until the excavations at Coppergate, the amount of shoe-leather of pre-Conquest date from England was limited. Later 8th- and 9th-century shoes had been found at 9 The Repton Stone is of 8th-century date: M. Biddle and B. Kjf21lbye-Biddle, The Repton Stone', Anglo­ Saxon England, 14 (r985), 233-92; The Middleton Cross is roth-century: J. T. Lang, Anglo-Saxon and Viking Age Sculpture (BAR Brit. Ser., 49, Oxford, r978), r8r, pis. 9.2 and 9.4, fig. 9.3.

SAXON SHOES, VIKING SHEATHS?

461

Oxford and Gloucester,lO while later 9th-mid-IIth-century shoes have been found at London,!1 Winchester,12 Norwich,13 Durham 14 and other excavations at York. ls For the most part, the number of shoes in each assemblage was small and they were often fragmentary and in poor condition. Despite this, differences are apparent in the cutting patterns of the shoes and the seams they used. Study of the large assemblage of shoes from well-dated deposits at Coppergate, York, has provided an opportunity to re-evaluate differences first recognized in the smaller groups recovered elsewhere. Ninety-nine per cent of shoes from Anglo-Scandinavian Coppergate were made with separate soles and uppers and the seam joining the sole to the upper was stitched in two different ways. In the first (construction I), the sole is sewn with a tunnel stitch: i.e. the stitching medium passes through the flesh surface of the leather and emerges out of the same side, as though passing through a tunnel (Fig. 2a). In the other (construction 2), the thread passes through the side of the sole leather and out of the flesh surface in what is known as an edgelflesh stitch (Fig.2b). These two types of seam were both in use throughout the Anglo-Scandinavian period at York. The tunnel-stitched method (construction I) was most popular in the loth and early IIth centuries but does not survive into the medieval period. The edge/flesh method (construction 2) came to dominate in the lIth century and was used almost exclusively throughout the later Middle Ages. This same general pattern can be seen in contemporary shoe assemblages throughout Europe. Study of the shoes from Coppergate, however, has shown that at York the choice of stitch does not appear to have been governed either by the style of shoe being made or necessarily by when it was made. This may also prove to be true elsewhere. Two early loth-century shoe soles from Winchester have sole seams that use both edge/flesh stitching and tunnel stitching. 16 There is no obvious technological reason for using two different stitches on the sole seam, unless one was the original seam stitch, the other a later repair. 10 ] . H. Thornton, 'Leatherwork', 155-60 in B. Durham, 'Archaeological investigations in St. Aldates, Oxford', Oxoniensia, 42 (1977); C. E. Gouge, 'The leather', 193-6 in C. M. Heighway, A. P. Garrod and A. G. Vince, 'Excavations at 1 Westgate Street, Gloucester, Medieval Archaeol., 23 (1979). 11 F. Pritchard, 'Footwear', 213-18 in A. Vince (ed.), Finds and Environmental Evidence. Aspects of Saxo­ Norman London: 2 (London, 1991). More recently excavations at The Poultry, Bull Wharf and The Guildhall in the City of London have produced large shoe assemblages that are currently undergoing study (J. Keily, pers. comm .). 12 ]. H. Thornton, 'Shoes, boots, and shoe repairs', 591-617 in M. Biddle, Object and Economy in Medieval Winchester (Oxford, 1990). 13 B. Ayres and P. Murphy, Waterfront Excavation and Thetford Ware Production, Norwich (E. Anglian Archaeol., 17, 1983),23-9. 14 M. O. H. Carver, 'Three Saxo-Norman tenements in Durham city', Medieval Archaeol., 23 (1979), I-8o. 15 G . Benson, 'Notes on excavations at 25, 26, and 27 High Ousegate, York', Ann. Rep. Yorkshire Phil. Soc., 1902 (1903),64-7; l. M. Stead, 'Excavations at the south corner tower of the Roman fortress at York, 1956', Yorkshire Archaeol. j ., 39 (1958), 515-38; ]. Dyer and P. Wenham, 'Excavations and discoveries in a cellar in Mssrs. Chas. Hart's premises, Feasegate, York, 1956', Yorkshire Archaeol. j., 39 (1958), 419-25; K. M. Richardson, 'Excavations in Hungate, York', Yorkshire Archaeol. j., Il6 (1959), 5I-Il4; L. P. Wenham, 'Excavations in Low Petergate, York, 1957-8', Yorkshire Archaeol. j., 44 (1972), 65-II3; A. MacGregor, Anglo-Scandinavian Finds from Lloyds Bank, Pavement, and Other Sites (The Archaeology of York, 17/3, York, 1982), 138-42; Tweddle, op. cit. in note 4, 237-56. 16 Thornton, op. cit. in note 12, 597- 9 and fig. 159, nos. 1859 and 1861.

462

LAND, SEA AND HOME

3. Shoe styles found at Anglo-Scandinavian York: (a) low cut slip-on shoes; (b) flap­ and toggle-fastening shoes. (Drawn by Quita Mould) FIGURE

Three shoe styles predominated at Anglo-Scandinavian Coppergate and the two construction seams described above were used to make all three. All of these styles were found in the earliest deposits, dating to the mid-9th to late 9th/early roth centuries, and continued in use to the end of the period. The first style of shoe is a low cut slip-on shoe (Fig. p) that looks rather like a modern day ballet slipper. The sole has a V-shaped heel extension that is folded up the back of the foot. The uppers are cut from a single piece of leather with a seam at the back to join to the heel extension of the sole. Some had a wide top band to raise the height of the top edge. Shoes of this style have been found in large numbers at Coppergate, where well over a hundred have been recognized, and others have been found elsewhere in York, London and Dublin.

SAXON SHOES, VIKING SHEATHS?

46 3

The second shoe style is the flap and toggle-fastening shoe (Fig. 3b). The upper is also made from a single piece of leather, this time wrapping around the heel and joining at the side of the foot with a single seam. The upper extends in a flap over the instep and fastens with a toggle passing through a loop attached low down on the quarters. Again, the sole has a V -shaped heel extension that this time fits into a corresponding V-shaped notch cut out of the uppers. Examples fastening with a single flap and toggle and a double flap and toggle are found, as well as shoes that have the flaps made from separate strips of leather attached to the top edge of the uppers. The manner of attaching the fastening loops and the toggles varies from shoe to shoe apparently at the whim of the individual shoemakers. The remains of nearly ninety shoes of this general flap and toggle style have been found at Coppergate. Shoes with flap and toggle fastenings have also been found elsewhere in York, London, Norwich and Oxford.17 It is the third style of shoe found at Anglo-Scandinavian York that is potentially the most interesting (Fig. 4). Like the previous style, the upper is made from a single piece of leather that wraps around the heel and is joined by a single seam at the side of the foot. In its simplest form the shoe is of slip-on type with no fastening, but variations with drawstring fastenings are also common. It is the most commonly found pre-Conquest shoe style occurring in all the principal assemblages. At York shoes of this basic style are found with a number of different methods of fastening or minor additions (a selection of which are illustrated), but they are all of the same basic design, a one-piece upper with a single side seam. At Coppergate this third shoe style was made using two distinctly different cutting patterns. The first (Fig. 4a) has a sole with a V-shaped heel extension that fitted into a V-shaped notch in the upper just like the flap and toggle fastening shoe style (Fig. 3b). The second (Fig. 4b) has a sole with a rounded seat. Here, the upper wraps around the heel of the foot but there is no notch; instead the back of the heel area is supported and strengthened by the attachment of a separate piece on the inside, known as a heel stiffener. Shoes of the same basic upper style were being made not only with two different ways of sewing the sole seam, but in two distinctive cutting patterns: one with a V-back, the other with a round back. Over 100 examples of this general style of shoe were found at Coppergate. While nearly 90% (88.5%) of these had round backs, V-shaped versions of the same shoes were also being produced in small numbers. The round-backed shoes are found in mid- to late 9th-/eady loth­ century contexts, but both types of shoe, those with round backs and V-backs, occur in deposits dating from the mid-loth century onwards. They appear to have been in contemporary use as they were certainly being thrown away in the same deposits in the middle years of the loth century. Interestingly, another style of shoe was being made in both V-backed and round-backed versions at York. Although, as we have seen, all the flap and toggle shoes from Coppergate had V-backed soles (Fig. 3b), a

17 MacGregor, op. cit. in note 15, 138, fig. 72, no. 627; Richardson, op. cit. in note 15, fig. 22, nos. 7 and 9- 10; Pritchard, op. cit. in note II, 219, fig. 3.104; Ayres and Murphy, op. cit. in note 13,24, fig. 21, no. 5 and fig. 23, no. 7; Thornton, op. cit. in note 10, fig. 35, nos. 10- 11 .

464

LAND, SEA AND HOME SPATIAL ANALYSIS AND CULTURAL INDICATORS

223

4- Shoe styles found at Anglo-Scandinavian York: (a) V­ backed shoes with a single side seam; (b) round-backed shoes with a single side seam. (Drawn by Quita Mould) FIGURE

single example with a sole with a rounded seat was found in earlier excavations at Hungate. 18 The few Anglo-Saxon shoes from England of late 8th- and earlier 9th-century date have soles with rounded seats, and appear to have one-piece uppers joining with a single side seam. The earliest appearance of the V-shaped heel extension found so far in England, comes from St Aldates, Oxford where two flap- and toggle-style shoes with V-shaped heel extensions were found in early to mid-9th-century contexts. 19 On a simplistic level, current evidence suggests that shoes with round backs were worn in Anglo-Saxon England. The wearing of shoes with V-backs, first seen in the middle 18 19

Richardson, op. cit. in note 15, fig. 22, no. 7.

Thornton, op. cit. in note 10, fig. 35, nos. 10-11.



SAXON SHOES, VIKING SHEATHS?

46 5

years of the 9th century, quickly spread across the country, apparently under Scandinavian influence. They fell out of use in the lIth century, both in York and throughout the rest of the country, as Scandinavian influence declined. Similarly, the tunnel-stitched sole seam passed out of use rapidly during the middle years of the 1 Ith century in favour of the edge/flesh seam. At York from the later 9th century onwards we see the contemporaneous use of two different heel shapes within the same shoe styles. 20 Why would shoemakers in York use fundamentally different techniques to make what would appear to be the same product? Could it reflect the merging of local, Anglian shoemaking traditions with those of the incoming Scandinavians? Craftsmen are by nature conservative. Traditional methods of working, handed down from one generation to the next, are often closely guarded, and new techniques only adopted if they bring obvious financial benefits or are necessary to accommodate the use of new materials. The change of sewing medium from leather thong to a woollen thread and eventually flax, may account for the final replacement of the tunnel-stitched sole seam by the edge/flesh seam by the middle years of the lIth century. At York, however, these two types of seam had been in use, side by side, for the best part of the previous 200 years. Similarly, the round-backed and the V-backed soles were used concurrently, neither one inherently better nor substantially quicker to manufacture than the other. Technological advances do not appear to have governed the differences we see in the York shoes. One may speculate on the significance of the round-seated shoe and its more fashionable V-backed counterpart, but on present evidence we cannot say who was making the latter, or for whom, in roth-century York. It can be proposed with greater certainty, however, that leatherworkers at York adopted Scandinavian fashions while retaining certain traditions of their own. Further evidence of native adaptability at York may be seen in bone-working and textile production. 21 At the beginning of this paper we described how an earlier Anglo-Saxon sheath-making tradition was visible within the Anglo-Scandinavian assemblage from Coppergate. It appears that elements of Anglo-Saxon tradition were also present in a substantial proportion of the footwear made and worn in York. ACKNOWLEDGEMENTS

The authors thank staff of the York Archaeological Trust, in particular Dave Hooley for cataloguing the bulk of the Coppergate leather, Ian Carlisle for completing this work and essential data gathering, and Richard Hall, Director of the Trust, for permission to publish this paper. We also thank the staff of the Museum of London Archaeology Service for providing contextual information on excavated leather from The Poultry, Bull Wharf and The 2063 % of the total assemblage of shoes from Anglo-Scandinavian deposits at Coppergate had V-backs, 37% round backs. 21 A. MacGregor, A. J. Mainman and N. S. Rogers, Craft, Industry and Everyday Life: Bone, Antler, Ivory and Horn from Anglo-Scandinavian and Medieval York (The Archaeology of York, 17/12., York, 1999); locally produced imitations of the piled weaves imported by the Scandinavians have been found . The locally made York copies were worked with needles while the Scandinavian imported textiles were woven: P. Walton­ Rogers, Textile Production at I6-22 Coppergate (The Archaeology of York, I71!, York, 1997), 1826.

466

LAND, SEA AND HOME

Guildhall before any detailed analysis had been undertaken. We are also grateful to Dr Pat Wallace and his colleagues at the National Museum, Dublin, for kindly allowing access to unpublished material in their collections. The ideas expressed here are the responsibility of the authors.

Dr Esther Cameron, Oxfordshire Museums Resource Centre, Cotswold Dene, Witney Road, Standlake, Oxon oX297QG, UK [email protected] Quita Mould, Barbican Research Associates, Eastmoor Manor, Eastmoor, King's Lynn, Norfolk PE339Pz, UK [email protected]

INDEX Bold = refers to figure on page Aberffraw (Wales) 146, 164 Aggersborg (Denmark) 10, 18 Ahus, Skane (Sweden) 21,25,26 Ailech, over-kingdom of (Ireland) 193, 197 kingof 199 Airgialla, over-kingdom of (Ireland) 193,197 Alborg (Denmark) 10 Aldingham (England) 299 Alholm (Denmark) 10 Alsen (Isle of) 27 altar 203 amber 294 Amounderness (England) 129,297,301,302 amulets (general) 45,47,5°-1,69 Cruciform 258 Thor's-hammer rings 43,50 Thor's hammer 420 anchorages see harbour Andozero-Novoeozero 60 Angeln (region) 35 Anglesey, Isle of (Wales) 139-75,145,303 Anglians 126 Anglo-Saxon Chronicle 144, 321 , 379, 382,4°5, 4°7,4 12 ,4 17,4 19 animals, raising of 287 Annagassan 126 Annals of fulda 19, 100 Annals of lnnisfallen 190 Annals ofIreland, Three fragments 140 Annals of Tigernach 198 Annals of Ulster 189,2°3,205 antler see bone and antler, pick anvil IIO, III archaeomagnetic dates 144,152-3 Arfon (Gwynedd, Wales) 140 Argyll (Scotland) 207,214-16 Arhus (Denmark) 10, II, 15, 16, 24, 26 Arklow (Ireland) 199 armour 51 armrings (silver) 14,42,95,96,98, 136, 140, 146, 157,173,179 glass 67 soapstone 223 Arnes (Norway) 269 Arran (Scotland) 216 arrowheads (iron) 14,49-50,67,248,249,265 Ashton (Lancashire, England) 302 Aslackby (England) 393

Aspatria (England) 127 Asselby (England) 400 Asser 140, 407 Ath Cliath (Dublin), king of 194 Atiscros 137 augers 288 Austby (England) 395 awls 158,250,288 axes, battle-axes, axe-heads 42,49,136,265 stone 328 Azatskoe-Rodionovskoe (Russia) 60 bakehouses 250 bakestones 229 balances/scales 19,85 Baldersdale (England) 334 Balladoole (Isle of Man) 127,131,158,161,168 Ballateare (Isle of Man) 127,131,161,457 Baits 43,5° Baltic region 27-39 Baltic objects 42 Bamburgh (England) 306,408 Bangor (Wales) 140, 143, 146 Bantry Bay (Ireland) 196 Banwell (Somerset, England) 367,369,371,372. Bardney (abbey of, England) 409 Bardsey (Wales) 143 barley 66,159 barley cakes 2.51 barrows 48, 50, 51 bathhouses 78 Bawsey (Norfolk, England) 319 beads (general) 50,2.29 amber 36,68

bead makers 109

carnelian 36

glass 66,67,68,144,156,251,2.58

jet 130

Beckermet (England) 305 Beetham (England) 302. bells (hexagonal, copper-alloy) 135, 161, 163, 164,175,34 0 Beloozero ('White Lake') region (Russia) 55,56, 58,60,62.-4,66,69 belt fittings 69 Benllech (Wales) 136,140 Benllech Bay 164 berewicks 413

468

LAND, SEA AND HOME

billhook 250 birch 330 Birka (Sweden) 25,109,446,452-3 Birkenhead (ferry point, Mersey, England) 298 Bistrup (Denmark) 10 blankets 454 boats (fjord boat, fishing boat, ferry) 9,34,236, 25 2

boat nausts 133, 134

clenched nails from (see also nails) 246

Bolton Priory Sheephouse (England) 338 bone and antler animal bone 14,69,151,235,243,261,262 antler tine, decorated 263 bone and antler working/offcuts 14, 15, 156, 15 8,29 0 ,46 5 objects (see also combs, cylinder, pins, spindle whorls, toggles) 243,262 bones (human) see burials Book of Ballymote 189 book edgings 337 bookmarker tags 337 Book ofU; Maine 189, 192 Borg (Norway) 89 Borgeby (Skane, Sweden) 109, II5 Borgund (Norway) 224 Bornais (South Uist, Scotland) 161, 236, 253, 255-72,256,257 Borre Style II6, 160,322,415 bowls 252 bowl mount (copper-alloy) 344 box beds 246 box handle (iron) 250 bracelets 189, 196,229 Bracken Rigg (Bronze-age, Upper Teesdale, Eng­ land) 328 Bramford (Essex, England) 169 Brandon (Suffolk, England) 319, 320 Bray Head, near Dublin (Ireland) 199 bread oven see oven Bride (parish, Isle of Man) 130 bridge 290 Bridgewater (Somerset, England) 360 bridle mount (iron) 163, 175 Brigham (England) 305 Briscoe (England) 334 Bristol (England) 252,369 Canynges House 365

Castle 365

Dundas Wharf 365

Britons 126 Brenhinedd y Saesson 141 Broekpolder (The Netherlands) 362 Bromborough (Wirral, England) 304

bronze see copper alloy brooches (general) 42,49, 51

annular 337

Borre-style 322

gold filigree 112

Hunterston brooch 215

oval 42,69,161,163,167,175

penannular 69,156

pseudo-penannular 158

Urnes 16

Westness brooch 214

Broomfield (England) 459

Brough of Birsay (Orkney) 210

Brut y Tywysogyon 140 Bryn Eryr (Wales) 165

Bryant's Gill (Kentmere, England) 308,335

Brycheiniog 144 bucket or vat, lead 288 bucket handles (iron) 265,266 buckles 153,160,161,174,250,337 Buckquoy (Orkney) 210,214 buildings see structures BUlandshofOi (Iceland) 28 1 Bullion economy 180-8 Burg (The Netherlands) 106 burials (see also graves, barrows, boat graves, cemeteries, chamber graves, inhumations, long barrows) 16, 34, 35, 47, 66, 68, 13 1, 140, 154-5, 161, 207, 238, 290, 294, 313-16, 419, 459 burials in coffins 316 burial customs 5 I burial mounds 49,51,58, 127, 13 1, 133, 134, 137,3 0 3

Burrow Hill (England) 320

Bury St Edmunds (England) 389

Buttington (Powys, Wales) 145

Bysbie (Scotland) 130

Byzantine imports 51

Caer Gybi (Wales) 141

Caerwent (Wales) 161

cairns 329

Caister-on-Sea (England) 418,420

Caistor (Lincolnshire, England) 410

Caithness (Scotland) 212

Callantsoog (The Netherlands) 103

Cambridgeshire (England) 404-5

Canon law 433,438,442

cantref 169 Cardiff (Wales) 139

Carlingford Lough (Ireland) 198-9

Carlisle (England) 170

Carn Fadrun (Wales) 165

INDEX

Carnforth (England) 344 Carrick (Co. Westmeath, Ireland) 185 Castell, Porth Trefadog (Wales) 144, 166 Caihreim Chellachtiin Chaisil 193,201,202 Catterick (England) 343 cattle 168,252,340 cauldrons 51,248,251 iron 243, (copper-coated) 246,288 cemeteries see also burials, graves 29,46, 50,66, 284,288,316,321,446 Cenet nEOgain (kingdom of Ailech, Ireland) 203 cereal cultivation/grains (see also barley, rye, wheat, oats) 64,287 Ceredigion (Wales) 140 cetaceans 252 chamber-graves 50,51 charcoal pits 330 Cheddar (England) 340,356 cheese 338 Chernigov (Russia) 42 Cheshire 298,303,307,406 Chester (England) 135-6, 140-1, 158, 160-1, 168,298,301 ,304 Hunter's Walk 153,354,355 St Werburgh's 307 Chester-Ie-Street (St Cuthbert's) 305 chests/boxes 246,251,439 chisel 158,243,248,249 Christian symbols 109 The Church 417 churches/chapels 16, 18, 33, 200, 287, 288, 294, 338,346,350,364,367,371,413 Cille Pheadair (Kilpheder, South Uist, Scotland) 166,235-54,23 8 ,27 2 Cirencester (England) 321,418 Clapham (Craven Dales, England) 343 Claughton Hall (England) 127 Claverham (England) 369, 370 clay moulds 17 Cleeve (England) 369, 370 Clew Bay (Ireland) 197 cloaks 454 Clontarf, battle of (Ireland) 197 clothing 194 Cloughermore cave (Co. Kerry, Ireland) 177 coal 337 coats of mail 189, 196 coffins, wooden 289 coffin brackets (iron) 316 Cogadh Gaedhel re Gallaibh 192, 194 coin hoards 14, 140 coins (see also coin hoards, hoards) Anglo-Saxon 135, 146, 150,156,178,3 18,338 Arabic 43,46,47,50-1,93,95,98,102

469

Byzantine 51,156 Carolingian 95-6, 170, 178 dirhams 95-6,146, (counterfeit) 294,344,420 English 251 Kufic 14,15,17-18,66,156 Oriental 67 production/minting 109, 180, 187 Sassanian/Sassanid drachmae 18,95 sceattas 21,25,319,320,419 silver (general) 12, 19, 21, 36, 42, 50, 62, 75-91, 96, 97, I F, 140, 147, 155-6, 168, 172,177-88,248,253,283,287,420

solidi 95,!O3

styccas (Northum brian) 156,338,340

Western European 66,67,69

coin ornaments 95 coin brooches 95,98 Colchester (England) 412 colonization 4 combs (general) 215, 243, 247-8, 251, 253, 262-3,264,269,271 antler 43,45,49,51-2, 2II , 24 1, 243, 245, 24 8 bone 318 comb-makers/making 17,109,253 iron, for wool 246,248 com mote (cwmwd) 169 Compton Bishop (England) 371 Conaille of Muirthemne 203 Congham (England) 320 Congresbury (England) 367, 369, 370, 371-5, 376 Connacht, over-kingdom of (Ireland) 193, 197, 20 3 cookhouses 78,200 Copeland (England) 129, 302 copper 19 copper alloys 21,69 'bronzesmithing' 156,157-8 copper alloy working see also crucibles 15, 293 objects see book edgings, bookmarker tags, brooches, buckles, disc, finger rings, mounts, pins, etc. 66,243,262 scrap, waste, molten residues 12, 17 sheet 265 Corcu Baiscinn (Leim) (of South Clare, Ireland) 196-7,202 Corcu Dhuibne (of Kerry, Ireland) 196,202,204 Corcu Laigde (of Cork, Ireland) 197 Corcu Mruad (Corcomroe, Co. Clare, Ireland) 201 King of 201, Cork (Ireland) 131,196 Cornwall (England) 364

47 0

LAND, SEA AND HOME

Cotherstone (England) 334 Cottam (England) 162, 168,319,320 Co. Durham 413 court site 134 cowrie shells 294 cows 203 craft (bone- and antler-working see bone and antler; copper alloy working see copper alloys; iron-working see iron; lead-working see lead; kilns; leather and leather-working; metal­ working; silver-working see silver; weaving see textiles) general 11,17,41,45,49,78,284,316 manufacturing debris 293 Craven Dales (England) 335-44,336 cremations 5 I Crimplesham (Norfolk, England) 320 Cronk Mooar (Isle of Man) 131,448,451,453 Crossthwaite (England) 334 crozier-shaped ornament 250 crucibles (general) 49, 109 (for copper alloy) 16 Cualu, king of (Ireland) 199 Cuerdale (England) 146, 306 Cumberland 406,407 Cumbria 130,302,304-5,331 Cunningsburgh 223 cupboard 439 cups 252 Cushalogurt (Co. Mayo, Ireland) 186 cylinder, bone 255,265-9,268,269 Dacorum hundred (Hertfordshire, England) 422 Dacre (England) 305 dagger sheath 269 D!Jirfhine of Ireland 196 D!JI Riata (kingdom of, Scotland) 198,215 Dalgard (Denmark) 10 Daliburgh (Dalabrog, Ireland) 237 Danelaw 129,299,4°8,413,434 daub (see also structures) 315 De administrando Imperio (by Constantine Por­ phyrogenitus) 52 Dee Estuary (NW. England) 298 Degannwy (Wales) 143 Deisi (of Waterford, Ireland) 196-7,204 Denby (England) 398-9 dendrochronological dates 37,79, 145,289 Denmark 7-26 Deer 265 Derby (England) 4II Derbyshire 387,389,391,404,406 Derry (Ireland) 204 Derwentwater (England) 302

Deventer (The Netherlands) 105 dialect 332 dies 109, I I I , 113-15, 116, 117, 120, 121, 122 cruciform II6 for leaf-shaped pendants 119 for rectangular and trapezoidal pendants 119 Dinas Emrys (Wales) 143 Dinas Powys (Wales) 165 Din L1igwy (Wales) 165, 167 Dinorben (Wales) 144 dioceses 413 disclroundel, copper-alloy 250,337,343 ditches 147,318-19 DilJelinstaynes ('Dublin Stones', York) 291 DNA 208-10 dogs 189 dog coprolites 241 Domburg (The Netherlands) 94, 106 Domesday Book/ hundreds 137,323, 362-3,371, 380-1,386,388,389,398,4°1,4°8,413,420 Donjum-Heringa (The Netherlands) 362 Dorestad (The Netherlands) 94 Down (Ireland) 199 drains (stone-lined) 153 draw knives 288 draw-plate IIO-II dress accessories 45, 67 Driby (England) 395 Drimore (South Uist, Scotland) 260 drinking horns 51,265 Dronningens gate hoard, Trondheim (Norway) 87 droveways 364 Dublin (Ireland) 126,127,130-1,136,140,146, 159, 160-1, 167-8, 180, 185, 187, 194, 196-8, 202, 205, 235, 29 1, 3 0 4, 306 , 44 6 , 44 8 , 451, 453-5,457,4 62 Dubthach moccu Lugair 191 duces 412 Dumfriesshire (Scotland) 131,406-7 Dunadd (Scotland) 217 Dunbar (Scotland) 348,349, 350-1 Dunham (England) 307 Dunmore cave (Co. Kilkenny, Ireland) 177,180 Durham (England) 4°8,455,461 dwelling sites 58,59,60,67 Dybs", Fjord (Denmark) 17 Dysart Island (Co. Westmeath, Ireland) 186 Dyserth (Wales) 137,140 Early Christian sites 364 Eastham (Wirral, England) 304 East Anglia (England) 379-80, 382, 389, 390, 40 5,4°8,4 15,4 17

INDEX

East Midlands (England) 380,385-6,395,397, 399,401,420,421 eggs 252 Eiderstedt peninsula 27 Eidsborg, Telemark (Norway) 29 Eigg (Scotland) 215 Ely (England) 412 enamelwork 161 enclosure walls (stone) 150,165 enclosure yards 325 Eoganacht Locha Uin (Killarney Lakes, Ireland) 196 Erikstorp brooch I I 3 Essex (England) 423 estates/estate centres 41, 52, 109, 164, 304-5, 307,322,367,368,374,387,390-1,404 ecclesiastical estates 105, 305, 322, 361, 369, 392

multiple 389,409

royal estates 105,369,372

ethnicity 208-12,214,236-7 Ettersgill (England) 334 Eynesbury (England) 354 Faccombe Netherton (England) 355 Fair of Carmun 193 Falster (Denmark) 25 farming 69,306,332-3,434 farmsteads, farm buildings 35,325-34,337-44, 434 Faroe Islands 223,331 Feast of Tara 193 Fedeland 223 Fenland (England) 359 ferrule for a stick or cane 251 fields, meadows 434 field-names 297,384,402,415 field systems 329 filigree lIO, 120 finger rings (copper-alloy) 14,161,163,175 (silver) 248 Finno-Ugric objects 42 Fir Manach (the men of Fermanagh, Ireland) 197 fish (general) 64,235,252,418 cod 14,252 herring 252,271 ling 252 saith 252 fishing equipment 15 fishponds 313 flax 17,18,222,252-3,465 breaking ditches for drying flax 18 flax-retting 289

471

Flegg, island of (Norfolk, England) 415,417-18, 4 20 flints 328 Flixborough (England) 168,319-20,350-1 floor surfaces 220 flotation 237 Formby (England) 299, 305 fortified sites/fortificationslfortresses (see also ring forts) 15,37,39,109, lI8, 29 2 , 4 19 foundation deposits 248 Fragmentary Annals see also Annals or Ireland 12 7 France (Carolingian) 303 Frankish Annals 19,39 Frankish kings 417 Frankish law codes 434 free peasantslliberi homines 388 Freswick (Scotland) 269 Frisia/Frisian islands/Frisians 27,93,101-3, 104, 443,44 8 ,45 2-3,455 furnace stones lIO furs 24, 69, 72 Fyfield Hall (Essex, England) 356 Fyns Hoved/Head (Denmark) 10, 13 Fyrkat (Denmark) 109, lI8

Gall-Ghaedhil 127,210,216 Gainford (England) 334 Galloway (Scotland) 130,198,215-16,302 games Fidchell sets 196 Gamlebyen royal manor, Oslo (Norway) 87 gander ('steg') 334 Gird under Sandet (Greenland) 453 Garn Boduan (Wales) 165 Gauber High Pasture (England) 335,339,341 Gedehaven (Denmark) 10 Genemain Aeda Slane 193 geochemistry 237 Gershoj (Denmark) 10 gift-giving 24 Gilla Mo-cholm6c's, ruler of 204 glass see also beads glass production (vessel, beads) 12,17

beads 15,36

buttons 68

vessels/sherds 15,68,318,320

Glastonbury (England) 360,361,377 Gloucester (England) 461 Gnezdovo (Russia) 42,44,46,48, 50-3,66 Gokstad (Norway) 452 gold 19,21,194,196,2°3 gold ingots lIO

gold ornaments 109

472

LAND, SEA AND HOME

gold strip 246

goldsmith's tools 109-23

gold working (waste, droplets) 16,293

goldworking furnace 110

(objects see bracelets)

Goltho (England) 162 Goritsy-Shuklino (Russia) 64 Gosforth Cross/sculpture (Cumbria, England) 133, 305 gouge 248,250 Gower (Wales) 139 grain 241 Grantham (England) 410 granulation 110, 120 Grassington, Lea Green (England) 337 graves 19,27,13°, 136,214,344,452,455,457 boat graves 36,42,5° grave-goods 211 grave covers/markers 284,288,291 Greasby 304 'Great Army' 283,287,331,381,393,4°7,417 Greater Manchester (England) 304-5,307 Greenigoe (Orkney) 446,452 Grog StrOmkendorf (Germany) see also Reric 25,27,3 2 ,3 6 -7,3 8 Gudme/Lundeborg (Denmark) 35 gullies/drip-gullies 315-16,318,350 Gwynedd (Wales) 140 hacksilver see silver Halton (Lancashire, England) 302, 307 Hamar (Norway) 89 Hammer (Denmark) 18 hammer 110, III, 158 harbours, anchorages 7-26 Hartlepool (England) 348-51,349 head-dresses 443,450,451-2,455 hearths (fireplaces) domestic 14,78, 152,228,230,232,246,258, 259,3°8,328,331,337,342,35° see also ironworking Hebrides (Scotland) 189,2°7 Inner 214-16 Outer 214-15,235-53,255-72 Hedeby (Haithabu, Germany) 7, 10, II, 15-16, 19,24-7,3 2 ,33,34,35-7,39, 1°9- 2 3,166,453 Heggen weather vane 269 Heimskringla I Helsby (England) 298,301 Helwith Bridge (England) 337 Hen Domen (Wales) 144 Hen Gastell, near Swansea (Wales) 144 herring oil 14 Hesket in the Forest (England) 127

Heversham (England) 299, 305 Hewish (Somerset, England) 369,371,375,376 Hexham (England) 408 Heynes (Iceland) 453 hidage/hidation 434-5 Hiddensee 118 Hiddensee style II3,122 Hiddensee treasure 113 hill-forts 48, 49 Historia de sancto Cuthberto 408,413 Historia Gru{{udd ap Cynan 142 Historia Norwegiae 210 Hjorring (Denmark) 114 hoards 27,34,43,46,51,58,76,86,93-106,14°, 146,148,157,170,177-88,182,207,306,344, 420 of dies III H6ll farm, Iceland 275,276 Hollingstedt (Germany) 33 Holwick Castles (England) 329 hones tones see whetstones honey 172 hook 67 hooked tags 135,161 Hope and Ashford estate (England) 404-5 Horncastle (England) 410 horns 196 horses 194,196,200 hostages 173 Hrakhvammr (Iceland) 281 Hvitarsioa (Iceland) 278

Iceland 223,235,273-82,331 Ilauntannig (Ireland) 165 Il'insky pogost (Russia) 72 Illerup (Denmark) III Ilow (fortresses, Germany) 37 In Aird (baronies of Ards, Down, Ireland) 198 Inber, king of (near Arklow, Ireland) 199 Ind Airthir (baronies of Orior, Armagh, Ireland) 19 8 In Dubthrian (barony of Dufferin, Down, Ire­ land) 198 'infield' enclosures 364 Inglewood (England) 302 inhumations 50,51,155,313-14 Inishmurray (Ireland) 165 Inishowen (Ireland) 204 insect remains 289 lona (Scotland) 214 Ireby (England) 302, 344 Ireland 4, 130, 156, 177-88, 189-205, 195,407, 445- 6 ,454

INDEX

Iron 19,24 bars 15,158 'blacksmithing' 43,49,52,156,158 hammerscale 17 ironworking, smithing slag 12, 13, 15-17,49, 293,325-6,328-30 ironworking forge/hearth 289,325 objects/ironwork (see also arrowhead, clenched nails, coffin brackets, knives, weapons) 45,243,246,262 Irish Sea region 125-38 Isdjord (Denmark) 8, 14 Isle of Man 129, 131, 133, 140, 144, 146, 162, 19 8 ,235,25 2 ,299,3 0 4,4 0 7,454 Jarlshof (Shetland) 230,263,269,340 Jaroslavl' district (Russia) 44 J arrow (England) 408 Jellinge style 116,415 jet/shale working 290 jewellery (general) 12, 15, 17,21,45,47,51,66, 69,113-22 gold 109 manufacture 43,49 Moravian 50 silver 109 Johnby (England) 302 J6rvik see York Jotunheim mountains (Norway) I Jurby (parish, Isle of Man) 130 Jyllinge (Denmark) 10 Karby (Denmark) 10 Kattegat (Denmark) 15 Kaupang (Norway) 89, 109,252,452,453 Kebister (Shetland, Scotland) 223 Kendal (England) 305 Kenn (England) 369, 370 Kent (England) 389 Kentmere (England) 305 Kewstoke (Somerset, England) 372 keys 439 barrel padlock key 343 Kiev (Russia) 43,46 Kildonnan (Isle of Eigg, Scotland) 446, 448, 45 2 - 3 Kilmainham/lslandbridge (Dublin, Ireland) 126, 137,158,168 Kilmorie Cross 130 Kiloran Bay (Colonsay, Scotland) 158 Kincora (Co. Clare, Ireland) 196,204 Kingston Seymour (England) 362-3,372-3,375 Kirillov monastery (Russia) 70 Kirkby Malham (England) 338

473

Kirkcolm (Galloway, Scotland) 130 Kisnema (Russia) 60,71 kitchens 340 Kneep (Isle of Lewis, Scotland) 446,452 knives (iron) 14, 49, 52, 67, 161, 168, 246, 248, 250,288,290,265,340,459 Kosel (Germany) 33,35 Krutik (Russia) 66 ladder ('stee') 334 Ladoga (Russia) 41,55 Lagore (Ireland) 159 Laigin, over-kingdom of (Ireland) 193,199 king of 199 Lakes Almere (The Netherlands) 105 Beloe (Russia) 58,59,60,61,65,70,71,72, 73 Il'men (Russia) 45,55 Ladoga (Russia) 43,44 Llangorse 144 Lamby (Wales) 143 lamps 230 Lancashire 302,304-5,307,406 Landntlmab6k 282 language 209,216-17,236,304 latrines 78 law, law codes (general) 433-42 regulations concerning animals 438

bridges 438

cultivation 436,438

fencing 433-4,439

grazing 433,438

maintenance of roadways 433-4, 438

use of forests 433

laws English Law II Canute 439 Erik's Sju:lland Law of 1250 438 Jutish Law 434,436-7 Kentish laws of king IEtheiberht 437 King Valdemar's Sju:lland Law I 439 Laws of IEtheiberht 438 Laws of Alfred 438 Laws of Cnut the Great 439-41 Laws of Hlothere 438 Laws of Hywel Dda 165 Laws of Ine 436,438 Laws of Wihtred 438 Norwegian, medieval 76 Royal laws 191 Vederlov 440-1 lead clenched nail 246 cruciform amulet 258

474

LAND, SEA AND HOME

lead-working 12,293 objects (see also spindle whorls, weights) 69, 243,262 pendants 160, 163 plates/sheet used in goldsmithing 120, 121, 122

sheet 265

trial piece 157, 158

vat or bucket 288

waste 12

leather/leatherwork (see also sheaths, shoes) 457- 6 5 tanning 289 leather-working 159,293,465 Lebar Glinne Da Locha 190 Lebor na Cert 189-205 Leicestershire 390,405-6,420 Leinster, kingdom of (Ireland) 193,196-7 Leth Cathail (baronies of Lecale, Down, Ireland) 19 8 king of 198 Liamuin, king of 194 lids (whalebone) 265 Limerick (Ireland) 131,197 Limfjord (Denmark) 8, 16 Lincoln (England) 162,412,448,451,454-5,457 Flaxengate 352,353 Lincolnshire 387, 389, 390, 393-4, 398, 400-3, 406,409,412,415 Lindisfarne (England) 408 linen 17,443-55 literacy }20 Little Langdale (England) 131 Llanbadarn Fawr (Wales) 142 Llanbedrgoch (Wales) 136,139-75,149,151,152 Llanfaes (Wales) 169 Llanfairpwllgwyngyll (Wales) 161 Llan-gors crannog (Wales) 144-5 Llantwit Major (Wales) 142 Lolland (Denmark) 25 London 357,369,446-7,457,461-3 Pudding Lane 153,352,353 St Paul's 269 long barrows 45 longphuirt 126, 130 Long Preston (England) 3}8 Lough Erne (Ireland) 198 loom heddle-rods 444 loomweights 13,18 Lotharingen (Lorraine) 94 Lough Derg (Ireland) 200 Lough Ri (Ireland) 200 Louth (Ireland) 199 Lozskoe (Russia) 60

Lund (Sweden) II, 85-7, 109, II3, 453 Lyn;£s (Denmark) 10,14-15 Lyndby (Denmark) 10

Mabinogi 169 Mac Carthaig (provincial kingship) 200 madder 172 Maeksa-Belozersk (Russia) 60 Maeksa II 67 Maen Achwyfan (cross) 137 maenol (large estate) 169 Mag Coba (baronies of Iveagh, Down, Ireland) 19 8

battle of 203-4

kingof 201

Mag Muirthemne (Ireland) 204 magnetic susceptibility 237 Malham, Malham Moor (England) 329, 336, 337- 8 ,342 Malpas (England) 308 Mammen Style 269,415 Manchester (England) 299, 305 manorialisation/manors 409-10 mantles 189 maritime communications 4 Mastermyr tool chest 122 Mecklenburg (fortress, Germany) 37 Medbourne Manor House (England) 356 Meic Lochlainn of Ailech 204 Meliden (Wales) 137 Meols (England) 135-6, 156, 161-2, 166, 168,

30 5

Mercia 129,3°3,3°7,310,379,382,389,4°5 mercury 19 metal vessels 67 metalwork 235,387,417 metalworkers/metalworking 109-23,293 metalworkers' tools 109-23 Middelburg (The Netherlands) 106 midden (deposits) 149,250,315-16 Middleton Cross 415,459 Mide (Ireland) 199,203 kingof 199 Midlands (of Ireland) 185, 187 migration 215-16,297,305 military camps 69 mill-houses 201 milk 338 Minchin Hole (Wales) 170 Minino (Russia) 66 MininoI 68 Minino II 68 ministri 412 Momri (Wales) 143

INDEX

Mlm (Denmark) 25 Moscow region 56 moulds 43,49, 109 ingot moulds IIO mounts from caskets 263 copper-alloy 17,258 silver 17 Urnes-style 322 Muirthemne (barony of Lower Dundalk, Louth, Ireland) 198 king of 201 Mullaghbogen (Ireland) 170, 172 Munster (Ireland) 158,193-4,198 king of 195-7,202 Myklebostad (Sogn og Fjordane, Norway) 344 Nabbe-Kildegard (Denmark) 10 Nres by Avno Fjord (Denmark) 10,18 nails 246,248,250 clench/clenched nails (ship nails) and rivet plates 12,14,161,246,248,252 narwhal 24 neckrings (silver) 14,95,96,98 Nederby (Denmark) 10 needles 248,250 Newtownlow (Ireland) 170 Nickolskoe (Russia) 60,66 Nickolsky pogost (Russia) 72 Nidarnes farm 87 niello 337 Norfolk 415,419-20 Normandy 388 Norra Asarp (Vastergotland, Sweden) 269-70 Northern Isles (Britain) 4,2II Northamptonshire (England) 405-6 North Kent Marshes (England) 359 North Leinster (Ireland) 185, 187 Northumbria 129, 156, 303-5, 310, 331, 379, 382 ,4 1 4 Norwich (England) 418,461,463 Nottingham (England) 412 Nottinghamshire 387,389,390,406 Novgorod (Russia) 41-3,45,55,62,66,69 Novokemsky (Russia) 60 oats 66,252 Obodrites (Slavonic people) 39 Odense Fjord 13 Oenach Macha (Ireland) 203 Orkney 211,216,235 earldom of 210 Ormes Head (Wales) 143 Ormside (England) 127

475

Ormskirk (England) 299, 305 ornaments 62,177,236,249-50,262 Orton Scar (England) 308 Oseberg ship (Norway) 9,452 Oslo (Norway) 85-6,89 Osmeliaun/Maes Ros Meilion (Wales) 141,146 Ossory, kingdom of (Ireland) 193 Ouinn (god) 281 out-field 436 outhouse/outbarns 239,248,343 oven bread oven r65

oven linings 319

oven plate 13

Oxford 369,461,463 Merton College 369 St Aldates 464 oxygen isotope analysis 294

pallium fresonicum 452 Palmyra (Syria) 452 Pant-y-saer (Anglesey, Wales) 165,167 parish names/ parochial organization 70-2, 304, 382 pathway/paving 248,325 patrix (copper-alloy) for gold-foil figure (gubbe) 17 Peel (Isle of Man) 127,136 Pagan Lady of 131 Peins-Oost (The Netherlands) 362 pelts 19 Pembrokeshire (Wales) 139 pendants (general) 42,49 made from comb sideplate 264 cruciform 51,67,68 (amber), II3, 1I8, 160, 163,175,248

gold II6

human face form 43

Terslev-type II2

valkyrie 43

Penmon (Wales) 136, 140-1, 143, 160, 172 Pennington (England) 299 people (names) Abbot of Armagh 204 ~skil 440 ~thelberht(king) 438 lEthelfired of Mercia 129,4°6-7 lEtheired II 213,340,440 lEtheirun 371 lEthelstan 129,168,283,302,412 Alfred the Great (king) 19,25,213,379,407, 40 9

Alfward 372

Archbishop Hincmar of Rheims 101

47 6

LAND, SEA AND HOME

Archbishop Wigmund 156 Archbishop Wulfhere 340 Archbishop Wulfstan 440 Asgeirr 403 Askel (moneyer) 89 Baroi Guomundarson 277 Baror Dumbsson 280-1 Benen mac Sescnen 190 Bishop of Lichfield 307 Bishop I>orlakr 281 Bishop of Utrecht 101 Bishop Reginbrand 15 Borkr 277 Brian Boruma/Boru 194, 197 Cellachan (king of Munster) 202 Charlemagne 94, 101 Charles the Bold 105, 140 Chelachan (king of Munster) 193 Cinaed mac Alpin (king) 211 Cnut/Canute (king) 248,371,382,439,440 Cormac mac Cuilenain 190 Cuthbert (Saint) 408,413 Cyngen 140 Deiniol (Saint) 146 Diarmit Ui Briain, king of Dublin 142 Domnall Ua Lochlainn 203 Donnchad Hua F1aind (king of Eoganacht Locha Lein) 196 Donnchad mac Flainn (king of Tara) 193,203 Donnchad son of Tairdelbach 203 Donnsleibhe Ua hEochada 203-4 Dudoc, Bishop of Wells 371 Eadred (king) 283 Eadred (leaseholder) 408 Eadwald (king) 318 Ealdred 408 Eamed 338 Ecgberht(king) 419 Edgar (king) 303 Edmund (king) 150,412 Edmund Cooke 371 Edward the Elder (king) 144, 146, 304-6, 321, 405-6,412,417-18 Edwin (of Northumbria) 156 Eirikr the Red 282 Erik 438 Fastrad 371 Gerulf (Frisian) 106 Gestr 280 Gilbert son of Thorold 372 Gisela (princess, daughter of Lothar II) 101 Giso, Bishop of Wells 371 Godfred (king) 19,25,39 Godfred 93-4, 100

Grimr 280 Gruffudd ap Cynan 139,142-4 Gruffudd ap L1ywelyn 142 Gunnar (moneyer) 89 Guthred 408 Guthroth Haraldsson 172 Guthrum 213,321,409 Gyrth (earl) 323 Halfdan (king) 19,407-8 HaIld0rr Snorrasson 88 HaIlr 278 Harald (Rorik's brother) 94,105,417 Harald Fairhair (king) 236 Harald (king) 371 Harald Hardriide (king) 76,79, 84, 87-8, 90 Hemming (Rorik's uncle) 101 Hildebert 371 Hlothere (king) 438 Hywel ap CadeIl/Hywel Dda 141 Ibn Fadlan 19 Ine(king) 371,436,438 Ingimund 127,129,135,140-1,146,147,169, 29 8 ,3 0 3 Ingjaldr 280-1 John, king of England 251 Lefric (moneyer) 89 Leifr Eiriksson 282 Leofric of Mercia 142 Lothar(king) 94,100-1,105,417 Louis (king) 19 Louis the Pious (king) 94, 101 Mael Sechnaill (king) 213 Magnus Haraldsson 172 Magnus the Lawmender 76 Magnus Berrf0tt/Magnus of Norway/Magnus Barelegs (king) 88,173,198,236 Merfyn ap Rhodri 141 Muirchertach Ua Briain 190,203 Olaf (Olav, Olafr) of Dublin 140, 166 Olav Haraldsson (king, saint) 87-8,283 Olav Kyrre (king) 79, 88 Olafr Sihtricsson 142 Olav Tryggvasson (king) 87 Oleg 46 0pe Snialle 440 Ordric 372 OrdwuIf 372 Orm 140 Ottarr/Ohthere (trader) 19,213 Odilo (moneyer) 340 Offa 351 Olga (princess) 52 Onalafball 408 Patrick 190-1,193, 196

INDEX

Prior of Bruton 369 Ragnald 306 Ragnall 408 Rhodri ap Merfyn (Rhodri Mawr) 140, 166, 170 Richard Cotton 371 Rorik 93-4, 100-1 Rurik 58 Scula 408 Sedulius Scottus 140 Serlo of Burcy 372 Sigurd (king) 19 Sineus 58 Sitriuc (son of Olaf) 198 Siward (earl) 283 Snaculf, son of Cytel 414 Styr, son of Ulf 414 Sveinn 403 Tairdelbach Ua Briain 196-7,2°3,2°5 I>orgrimr I>orsteinsson 277 Thorkell 403 I>6rdis 277 Tilred 305 Tostig 302 Ua Beoain 203 Ua Conchobair (king of the Ciarraige) 203 Ulf (moneyer) 84 Ulfcel (moneyer) 89 Valdemar (king) 439 Vesteinn Vesteinsson 277 Wihtred 438 William the Conqueror 288, 371 William the Lion 212 William Rushworth 369 Wulfred (moneyer) 340 Wulfric Spot 303 Wulfstan 25 Perejaslavl Zalessky (Russia) 62 personal names (general) 403,415 -ketill, -kell name elements 400 Th6r-, T6ki-, T6Ii-, T6sti name elements 400 pick, antler 259,264 Pictish houses 236,241,251 picture stones (Gotland) 9 pins 135,15°,168,215,243,247,248,318 bone 241,243,246,247,248,251,264 copper-alloy 246,247,25°,264,265,32°,337 iron 248 pits 148,243,258,291,318 cess-pits 287 rubbish pits 287 place-names II, 14, 127, 128, 129-31, 136, 139, 140, 143, 146, 199, 207, 212, 274, 281-2, 297-311,322,332,344,379-423,439

477

-by names 381 , 3 84, 386-92, 394, 395, 396, 397,399,400,401-8,414,416,419-22 'Grimston-hybrid' names 382, 386, 394, 416, 4 20,4 2 3

Pictish names 216

-thorp names 381,387,394,416

-tun names 386, 388, 394, 399, 401

-ville names 388

plant remains/plants 289, 306 plaques (whalebone) 267 ploughshares 52 pollen 11,64,95,305,328 cereal 328 Pool 210,214 pork 252 post-holes/post-pits 14,148,241,258,308,315, 3 18 ,32 8 ,335,34 8 Portchester Castle (Hampshire, England) 355, 35 6 pottery general 18,21,25,5°,79, 172,214,230,235, 243, 244, 249, 250, 252-53, 260, 261, 262, 265,269,271,287,289,293,373,387

Badorf ware 95,96

Baltic ware (general) 12,14-15

Belgic-type wares 316

Chester Ware 156,158,168,308

cooking pots 249,251

globular 14

Ham Green 356

hand-made 62

Hebridean wares 266

hemispherical 15

Ipswich Ware 315,318,351

medieval 365,367,368

Minety Ware 248,250-51

Pingsdorf Ware 29,31

Platter Ware 241,243,249-51

post-medieval 368

Rhenish 36

Romano-British 337,365,368

Shapwick fabric AAl 365,367

Shapwick fabric UI 365,367

South-west English 266

Thetford-type Ware 315,321

undecorated, flat-bottomed 14, 15

Western Slavonic 50

wheel-thrown 12,21,62

potter's wheel 21 Pouholt, estate of Glastonbury Abbey 361 Powys (Wales) 140 pressed-sheet applique III, II5 crosses 122 pressed sheet, methods of production 120-1

47 8

LAND, SEA AND HOME

Privileges of Arfon 169 production sites 7,16 Psalter of Cashel 190 Pskov (district) (Russia) 44,62 punch 121 purse 85 Puxton (Somerset, England) 360, 363, 364, 366, 367,368,369-75,376 quays 84-5 quernstones general 29,36,159,253,338,342 basalt (from Eifel) 15, 21-4, 29, 31, 34, 35, 294,353

granite 21

Millstone Grit 340

tufa 326

Raby (England) 304 radiocarbon dating/dates 79, 106, 147-50, 153, 155, 174,228,25 1,269,328 Rampside (England) 127 Raphoe (Co. Donegal, Ireland) 186 Rathlin Island (Ireland) 198 Red Wharf Bay (Wales) 140, 146-7, 148, 164, 173 Red Wharf Bay arm-rings 140, 146 reed beds/reed 313,319,418 religion, religious veneration 212,330 Repton (England) 459 Reric (Germany) see also Grog Stromkendorf 19,25,3 2 ,3 8-9 Rhineland, trade with 459 Rhosyr (cantref, Wales) 169 Rhuddlan (Cledemutha, Wales) 137,143-4,170 Ribe (Denmark) 10, II , 15-16, 21, 24-30, 33, 36-7,39, 1°9 Ribblehead (North Yorkshire, England) 308, 329,335,33 8 ,339,343,359 ring 196 ring forts (cashels) 166 ringed pins 130,150,160,162,168,174-5 iron ringed pins 243 Ringerike Style 248,251,265-9,268,270,271 Rivers Aire (England) 335

Barrow (Ireland) 196

Blackwater (Ireland) 196

Congresbury Yeo 371-2

Dee (Kirkudbright, Scotland) 130

Don (Russia) 43

Dnepr, Dnieper 43,45,55

Heacham 313

Humber 321

Kama 43

Kema 71

()use 287,29°,292

Rhine 294

Ribble 298,306,335

Sheksna 70

Schlei 109

Solway 298

Tees 329

Thames 459

Volga 19,60

Volkhov 45

Volyn' 43

Wharfe (Egland) 335

Witham 337

roads 364 Rollesby (England) 420-1 Romney Marsh (England) 359 Rooskey (Co. Donegal, Ireland) 181 Roskilde Fjord (Denmark) 8, 10, 11-12, 14-15, 36 , 18 9 harbour 26 Rostov (Russia) 55,62 Rostov-Suzdal principality 58 Rudham (Norfolk, England) 320 Runcorn (ferry point, Mersey, England) 298 runes 43 Rus' 2,56,58,69,72 Russia (ancient), Russia 4,41-53,55-73,57 Russian Primary Chronicle 42-3,46,52,58 rye 66,252 Ryurik Gorodishche (Russia) 42,44,45,52-3 Sxb61 farm, Iceland 275,276 sagas Baroar saga Snxfellsass 280 Eiriks saga rauoa 282 Gisla saga Surssonar 273,276-7 Heioarviga saga 277-9 Hrafnkels saga 279 islendingasogur 273-82 Njals saga 273,280 of ()Iav Tryggvasson 87 ()rkneyinga Saga 210 salt/salt production 172,361,418 Sand bach (England) 307 saw 158 scales 50-I, 89, II I Scalloway (Shetland, Scotland) 223 Scatness Brough (Shetland, Scotland) 221 ()Id Scatness 220 Schlei (region) 35 Schleifjord 33,36

219-33,

INDEX

Schleswig 27,33, III Schuby (Germany) 33,35 Scotland 126,207-17,235-54,331,389,445 scythes 49, 52 sea birds 252 sea-level 8 seam-smoothers 18 sea walls 361,364 Sebbersund (Denmark) 10, 16 Sedgeford (Norfolk, England) 313-23,314,317 Selso-Vestby (Denmark) 10,12 settlement mounds 235-54,255-7,256 settlement patterns (see also towns) Denmark 7-26

Baltic area 27-39

England 379-423,383

Irish Sea area 132

rural 308-10

Somerset Levels 369-77

Severn Estuary (England, Wales) 362-3 shale vessels 328 Shannon estuary (Ireland) 196 Shapwick (polden Hills, Somerset, England) 360 sheaths 457-60,458 sheep 168,252,338,340 Sheppey (Kent, England) 350 Minster 351,352 Sherborne (England) 371 Shestovica (Russia) 42 Shetland 235 shields 196, 202 shield-boss 136 ships (longship, warship, transport ship, cargo ship) 9,24,26,85,101,105,107,189-205,291 (see also clenched nails) construction 12 donations 190-205 fleet 194, 197-8 longa 200 repaIr 12 ship-men 304 shoes 79,457,460-5,460,462,464 Shuklino (Russia) 71 Sigtuna (Sweden) 84, 109 silk 443-55 silver (general) 19,21,67, 194 hacksilver 12, 14, 19,21,93,97,98,146, ISO, 157,158,164,168,173,177-88,344 ingots 95,96,97,98, IIO, 146, ISO, 157, 158, 177-88 objects (see arm-rings, mounts, neck rings, strap-ends, etc.) rings 2J

479

silver-working (waste, droplets, 'silversmiths') 12,16,17,156,157-8,293 Simy Folds (England) 308,325-34,327,332,333, 335,359 Skaill (Orkney) 223 Skipton, Horse Close (England) 335 Skuldelev 2 (ship) 189 slag 49 Slavs 43,50 ~aves 19,168,173, 362 Slavonic objects 42 Sli (Schlei) Fjord (Denmark) 8 Smethdon (England) 323 Smolensk (Russia) 41-3,46,48, 52 soapstone/steatite (general) 271 soapstone-working/quarrying 223 armnngs 213,224 artefacts (general) 227 bakeplates/bakestones 223-4,225 furnace stones lIO lamps 223-4 spindle whorls 247 vessels 15,22-4 (Norwegian), 29, 31, 34, 35, 36,219-28,225,226,243 sock 450,454 sokes/sokeland 389,410,413 Somerset Levels (England) 359-77,370,374 Sondero (Denmark) 10 Sorbie (Scotland) 130 Sou burg (The Netherlands) 106 South-East Midlands (of England) 390 Southern Ui Neill 188 Southumbria 412 spatial analysis 219-33 spearhead, spears (iron) 14,42,136,340 spindle whorls 229,262,265 bone 247,267

lead 265

steatite 247

stone 267,331,334

spring reservoir/pool 150,151 St David's (Wales) 142, 161 St Olavsvoll, Sarpsborg, 0stfold hoard (Norway) 87 standards 202 Staraja Ladoga/ Aldeigjuborg (Russia) 43,44,45, 50 Starup 0stertoft (Denmark) 10 stews 251 stock breeding 64 stone (see also crosses, hogbacks, runestones, sculpture, spindle whorls, whetstones)

crosses 130,133,140,338,415

decorated in Ringerike Style IS, 269

480

LAND, SEA AND HOME

hogbacks 133,304 runestones 34 sculpture 133-4,136,143,269,284,3°5,3°7, 334,344,4 15,459

slate, worked 241

spindle whorls 67,326

stonework 310

tools 262

volcanic tufa (from Middle Rhineland}29, 32,33 Stonehouses (England) 329 storehouses 78,239,249 storerooms 78,246,248,439 Stora Ek (Vastergotland, Sweden) 269,27° Strandby Gammeltoft (Denmark) 10,12,13 Strangford Lough (Ireland) 198, 199 strap-ends (general) 135,161,343 iron 163, 175

copper-alloy 250, 251

silver 95,96,98,160,174,337

strap-slides (iron) 163, 175 Strathclyde 140, 304, 406 street names 283-5 strike-a-lights 43 structures (general) 151,152,153,283-94 aisled 17,18 benching {low} 152,166,340 'earth-fast' (see also post-holes) 345 halls 148,253,318 houses 45,237,246,257,258-60 huts 29 longhouses 12,18,35, 137,236-43,239,24°, 24 1 ,249 post-built 33 rectangular buildings 325-6,337-40 round-house 148 sheds 249 sill-beam/ sill wall construction 151, 166, 345-5 6 stake and wattle/post and wattle/wattlework/ wattle and daub/cob and wicker 288, 291, 315,346 ,35 1 sunken-featured/floored 12-13,16-18,29,35, 37,15 1-3,3 18 ,353 timber 148,290,3°8,315,318,34°,345-58 turf walls 246,259 wheelhouse 219-30 Stubberup (Denmark) 10 Suffolk (England) 4°6,415,423 summer banks (to modify marshland) 361 Sutherland (Scotland) 207,212 Suzdal (Russia) 43,55 styli 318-20 swage block 1I0-1I

swords (general) 42.,51,13°,189,196,2.36 pattern welded 42.,459 Talacre (Wales) 136,139 talley sticks 89 Tamworth (England) 357 Tara (Ireland) 193 Tatton (England) 308,309 Taunton (England) 369 Teesdale (England) 331 Tegeingl (Wales) 137, 140 Temair, over kingdom of (Ireland) 193,2.02. king of 199 Terpen 362. Terslev Type 112-13, 1I6, 122. Terslev-type dies 117 textiles see also flax, loomweights, seam smoothers 443-56 distribution in Britain 449 linen 2.53,443 ndlebinding 443,450,451,453-4 piled cloth 448-51,45° shaggy pile 443 silks 2.94,443,450,455 tabby 444,445 textile production, weaving 12, 15-18, 109, 15 6 ,2.93,46 5

twills 443,445,446-8,447,452.

wadmal 448,450,453-4

wool 2.53,443-53

The Biggings (Papa Stour, Shetland) 448 Thetford (England) 32.2 Thing/thing siteiAlpingi/thing-moots 86, 131, 133,135,2.80,438 Thinghoe hundred (England) 42.2. Thingwall (England) 303 Porr (god) 2.81 Thor's-hammer rings 51 Thurcaston (England) 420 Thurstaston (England) 303 Tiel (The Netherlands) 107 tiles 353 polychrome relief type 288 Roman 353 Timerevo (Russia) 66 tin 19,150 Tisso (Denmark) 114, 1I5 tofts 31°,434,436 toggles bone 2.41 Tolstrup (Denmark) 1I8 tongs 1I0 tools 45,49,62.,248,2.65 'toothed socketed tools' 158

INDEX

torque II3 touchstones I I I trade and exchange 4, 7-26, 29-, 45, 66, 126, 196 ,200,293-94,43 8 ,453

trading places 69, 164,291-2

wics and emporia I I

trapping 64,69,72 Treaty of Verdun 105 Treaty of Wed more 321 Trelleborg (Denmark) 15,109,120 Trewhiddle style 160,337 Trondheim (Norway) 75-91,77 Truso 25 Tullstorp (Skiine, Sweden) 269,270

Ua Bresaii Macha (Ireland) 203 Ua Ruaire (Ireland) 203 Udal, The (North Uist, Scotland) 251 Ui Briain kings of Munster 193, 196-8,204 Ui Ceinnseiaig/Chennseiaig (of South Leinster, Ireland) 194, 199,204 Ui Echach (of Munster, Ireland) 195 Ui Echach Arda 198 Ui Enechgiais, king of 204 Uiaid, over-kingdom of (Ireland) 189, 193, 194, 197,198,200,201,202,203,204 king of 195,199 Ulster, king of 195 Unst 223 Uppiikra (Sweden) 114-15 Upper Sheksna region (Russia) 61,64 urban centres, towns (general) 28, 41, 75-91, 109,351-4 Urnes Style 322

Varangians 42 Vedb