Celtic from the West 3: Atlantic Europe in the Metal Ages ― Questions of Shared Language 1785702270, 9781785702273

Table of contents :
Cover
Title
Copyright
Contents
Maps, Illustrations, and Tables
Introduction: BARRY CUNLIFFE & JOHN T. KOCH
PART I: ARCHAEOLOGY
1. Behind the Warriors: Bell Beakers and identities in Atlantic Europe (3rd millennium BC): LAURE SALANOVA
2. The Lost Cultures of the Halberd Bearers: a non-Beaker ideology in later 3rd millennium Atlantic Europe: STUART NEEDHAM
3. Closed for Business or Cultural Change? Tracing the re-use and final blocking of megalithic tombs during the Beaker period: CATRIONA D. GIBSON
4. Copper mining, Prospection, and the Beaker Phenomenon in Wales—the significance of the Banc Tynddol gold disc: SIMON TIMBERLAKE
5. Burial Practices in Ireland during the Late 3rd Millennium BC—connecting new ideologies with local expressions: KERRI CLEARY
6. Stelae, Funerary Practice, and Group Identities in the Bronze and Iron Ages of SW Iberia: a moyenne durée perspective: DIRK BRANDHERM
7. Language Shift and Political Context in Bronze Age Ireland: some implications of hillfort chronology: WILLIAM O’BRIEN
8. Metal, Metalwork, and Specialization: the chemical composition of British Bronze Age swords in context: PETER BRAY
9. Emerging Settlement Monumentality in North Wales during the Late Bronze and Iron Age: the case of Meillionydd: RAIMUND KARL
10. Ephemeral Abundance at Llanmaes: Exploring the residues and resonances of an Earliest Iron Age midden and its associated archaeological context in the Vale of Glamorgan: ADAM GWILT, MARK LODWICK, JODY DEACON, NICHOLAS WELLS, RICHARD MADGWICK, & TIM YOUNG
PART II: GENETICS
11. The Genetic Structure of the British Populations and their Surnames: BRUCE J. WINNEY & WALTER F. BODMER
12. Archaeogenetic and Palaeogenetic Evidence for Metal Age Mobility in Europe: MARIA PALA, PEDRO SOARES, & MARTIN B. RICHARDS
PART III: LINGUISTICS
13. Archaeology and Language Shift in Atlantic Europe: J. P. MALLORY
14. The Question of a Hamito-Semitic Substratum in Insular Celtic and Celtic from the West: STEVE HEWITT
15. Phoenicians in the West and the Break-up of the Atlantic BronzeAge and Proto-Celtic: JOHN T. KOCH
16. Celtic ‘Dogs’ in the Iberian Peninsula: FERNANDO FERNÁNDEZ PALACIOS
17. ANCILLARY STUDY: Sound Change, the Italo-Celtic Linguistic Unity, and the Italian Homeland of Celtic: PETER SCHRIJVER
18. ANCILLARY STUDY: Celtic as Vasconized Indo-European? Three structural arguments: THEO VENNEMANN
Index

Citation preview

CELTIC FROM THE WEST 3

FRONT MATTER CW3.indd 1

31/05/2016 16:08:46

              

celtic studies publications        series editor: John T. Koch

           

celtic studies publications i The Celtic Heroic Age: Literary Sources for Ancient Celtic Europe and Early Ireland and Wales, ed. John T. Koch with John Carey (Four th Edition, revised and expanded, 2003) Pp. x + 440 isbn 1–891271–09–1 celtic studies publications iii A Single Ray of the Sun: Religious Speculation in Early Ireland, John Carey (Second Edition, 2011) Pp. x + 123 isbn 978–1–891271–18–2 celtic studies publications iv Ildánach Ildírech. A Festschrift for Proinsias Mac Cana, ed. John Carey, John T. Koch, and Pier re-Yves Lamber t (1999) Pp. xvii + 312 isbn 1–891271–01–6 celtic studies publications vii Yr Hen Iaith: Studies in Early Welsh, ed. Paul Russell (2003) Pp. viii + 224

isbn 1–891271–10–5

celtic studies publications viiI Landscape Perception in Early Celtic Literature, Francesco Benozzo (2004) Pp. xvi + 272

isbn 1–891271–11–3

celtic studies publications IX Cín Chille Cúile—Texts, Saints and Places: Essays in Honour of Pádraig Ó Riain, ed. John Carey, Máire Herbert, and Kevin Murray (2004) Pp. xxiv + 405 isbn 1–891271–13–X celtic studies publications X Archæologia Britannica: Texts and Translations, Edward Lhwyd, ed. Dewi W. Evans and Brynley F. Roberts (2009) Pp. xii + 262 isbn 978–1–891271–14–4 celtic studies publications Xi Ireland and the Grail, John Carey (2007) Pp. xxii + 421 isbn 978–1–891271–15–1 celtic studies publications XIII Tartessian: Celtic in the South-west at the Dawn of History, John T. Koch (second edition 2013) Pp. ix + 332 isbn 978–1–891271–17–5 celtic studies publications XIV Moment of Earth: Poems & Essays in Honour of Jeremy Hooker, ed. Christopher Meredith (2007) Pp. xvi + 313 isbn 978–1–891271–19–9 celtic studies publications XV Celtic from the West: Alternative Approaches from Archaeology, Genetics, Language and Literature, ed. Barry Cunliffe and John T. Koch (2010; 2012) Pp. xii + 383 isbn 978–1 –84217–475–3 celtic studies publications XVi Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, ed. John T. Koch and Barry Cunliffe (2013) Pp. viii + 237 isbn 978–1 –84217–529–3 celtic studies publications XVii Memory, Myth and Long-term Landscape Inhabitation, ed. Adrian M. Chadwick and Caitriona D. Gibson (2013), Pp. xi + 347 isbn 978–1 –78297–393–5 celtic studies publications XViii The End and Beyond: Medieval Irish Eschatology, ed. John Carey, Emma Nic Cár thaigh and Caitríona Ó Dochar taigh (2014) 2 vols, Pp. xi + 944 isbn 978–1 –891271–20–5

FRONT MATTER CW3.indd 2

31/05/2016 16:08:46

Celtic from the West 3 Atlantic Europe in the Metal Ages: questions of shared language edited by

John T. Koch and Barry Cunliffe in collaboration with

Kerri Cleary and Catriona D. Gibson OXBOW BOOKS OXFORD

2016

FRONT MATTER CW3.indd 3

31/05/2016 16:08:46

Published by Oxbow Books, Oxford, UK © Oxbow Books, John T. Koch, Barry Cunliffe, Kerri Cleary, Catriona D. Gibson, and the individual authors 2016 Hardback Edition: ISBN 978–1–78570–227–3 Digital Edition: ISBN 978–1–78570–228–0 All rights reserved. No part of this publication may be reproduced in any form or by any means, without permission from the Publisher. This book is available direct from Oxbow Books, Oxford (Phone: 01865-241249; Fax: 01865-794449) and Casemate Academic 1950 Lawrence Road, Havertown, PA 19083, USA (Phone: +1 610 853-9131; Fax: +1 610 853-9146) or from our website www.oxbowbooks.com A CIP record for this book is available from the British ­Library. Library of Congress Cataloging-in-Publication Data Names: Koch, John T. | Cunliffe, Barry W. Title: Celtic from the West 3 : Atlantic Europe in the Metal Ages : questions of shared language / edited by John T. Koch and Barry Cunliffe. Other titles: Atlantic Europe in the Metal Ages | Celtic from the West three | Celtic from the West III Description: Oxford : Oxbow Books, 2016. | Includes bibliographical references and index. Identifiers: LCCN 2016004318 (print) | LCCN 2016005273 (ebook) | ISBN 9781785702273 (hardback) | ISBN 9781785702280 (digital) | ISBN 9781785702303 (pdf) | ISBN 9781785702280 (epub) | ISBN 9781785702297 (mobi) Subjects: LCSH: Civilization, Celtic. | Celts--Atlantic Coast Region (Europe)--Antiquities. | Bronze age--Atlantic Coast Region (Europe) | Iron age--Atlantic Coast Region (Europe) | Ethnoarchaeology--Atlantic Coast Region (Europe) | Group identity--Atlantic Coast Region (Europe)--History--To 1500. | Celtic languages--Atlantic Coast Region (Europe)-History--To 1500. | Language and culture--Atlantic Coast Region (Europe)--History--To 1500. | Human population genetics--Atlantic Coast Region (Europe)--History--To 1500. | Atlantic Coast Region (Europe)--Antiquities, Celtic. Classification: LCC CB206 .C445 2016 (print) | LCC CB206 (ebook) | DDC 940/.04916--dc23 LC record available at http://lccn.loc.gov/2016004318

Text typeset in the Cynrhan type family by CSP-Cymru Cyf. Cover design by CSP-Cymru Cyf.

FRONT MATTER CW3.indd 4

31/05/2016 16:08:46

CONTENTS   Introduction   BARRY CUNLIFFE & JOHN T. KOCH       

1

part i: archaeology

1. Behind the Warriors: Bell Beakers and identities in Atlantic Europe (3rd millennium BC)   Laure Salanova     13 2. The Lost Cultures of the Halberd Bearers: a non-Beaker ideology in later 3rd millennium Atlantic Europe   stuart needham   40 3.  Closed for Business or Cultural Change? Tracing the re-use and final blocking of megalithic tombs during the Beaker period   catriona d. Gibson     83 4.  Copper mining, Prospection, and the Beaker Phenomenon in Wales—the significance of the Banc Tynddol gold disc   Simon Timberlake     111 5.  Burial Practices in Ireland during the Late 3rd Millennium BC— connecting new ideologies with local expressions   139   kerri cleary   6.  Stelae, Funerary Practice, and Group Identities in the Bronze and Iron Ages of SW Iberia: a moyenne durée perspective   Dirk brandherm     179 7.  Language Shift and Political Context in Bronze Age Ireland: some implications of hillfort chronology   William O’Brien     219 8.  Metal, Metalwork, and Specialization: the chemical composition of British Bronze Age swords in context   Peter Bray     247 9.  Emerging Settlement Monumentality in North Wales during the Late Bronze and Iron Age: the case of Meillionydd   Raimund Karl     265 10. Ephemeral Abundance at Llanmaes: Exploring the residues and resonances of an Earliest Iron Age midden and its associated archaeological context in the Vale of Glamorgan   Adam Gwilt, Mark Lodwick, Jody Deacon, Nicholas Wells, Richard Madgwick, & Tim Young   294

  [v]

FRONT MATTER CW3.indd 5

31/05/2016 16:08:47

part ii: genetics

11. The Genetic Structure of the British Populations and their Surnames   Bruce J. Winney & Walter F. Bodmer   333 12. Archaeogenetic and Palaeogenetic Evidence for Metal Age Mobility in Europe    Maria Pala, Pedro Soares, & Martin B. Richards 351 part iii: linguistics

13.  Archaeology and Language Shift in Atlantic Europe   J. P. Mallory   387 14. The Question of a Hamito-Semitic Substratum in Insular Celtic and Celtic from the West   407   Steve Hewitt 15.  Phoenicians in the West and the Break-up of the Atlantic Bronze Age and Proto-Celtic   JOHN T. KOCH   431 16. Celtic ‘Dogs’ in the Iberian Peninsula   Fernando Fernández Palacios   477 17. ancillary study: Sound Change, the Italo-Celtic Linguistic Unity, and the Italian Homeland of Celtic   Peter Schrijver   489 18. ancillary study: Celtic as Vasconized Indo-European? Three structural arguments   Theo Vennemann   503   Index  

533

  [ vi ]

FRONT MATTER CW3.indd 6

31/05/2016 16:08:47

maps, illustrations, and tables 1.1 Distribution of the Bell Beaker graves in the northern half of France   16 1.2 Inhumation layer from the Bury gallery grave (Picardie, France)   18 1.3 Family cells from the La Chaussée-Tirancourt gallery grave (Picardie, France)   19 1.4 Anthropomorphic stelae from the 5th and 4th millennium BC   21 1.5 The Bell Beaker set from the Wallers individual grave (Nord, France)   23 1.6 High quality and clumsy Bell Beakers   24 1.7 Maritime beakers from Portugal, Galicia, and Brittany   25 1.8 Maritime beakers from the Netherlands, Czech Republic, and Eastern Spain   26 1.9 The three AOO graves from northern France, linked to the Grand-Pressigny flint workshops and the Lower Rhine Valley   27 1.10 Distribution of the main regional Bell Beaker pottery styles   28 1.11 Bell Beaker settlement from La Noue SaintMarcel (Burgundy, France). Plan of the site and associated pottery   30 1.12 Settlement patterns   32 2.1 A range of typical early Beaker material from graves in Atlantic zones of Europe   40 2.2 Representative examples of the ‘Atlantic’ family of halberds 42–3 2.3 Suggested chronology of European metalheaded halberds   45 2.4 The possible thin typological thread connecting early halberds in north-western Europe     45 2.5 Maps of recovery for a) copper halberds, and b) copper daggers and knives from Britain and Ireland, c. 2450–2150 BC   47 2.6 Map of recovery for copper and bronze halberds in Britain and graves datable to the Chalcolithic and transition period, c. 2450–2100 BC. Late types of halberd are distinguished and

comprise those of bronze along with copper or unanalysed examples of the same types, c. 2200–2050 BC.   48 2.7 The early Beaker context at South Hill, Talbenny, Pembrokeshire and the Beaker pot 49 2.8 Interpreted territories for halberd-bearing and Beaker groups during the Chalcolithic to earliest Bronze Age in Britain   50 2.9 Map of recovery for halberds from Ireland with outline enclosures for concentrations   51 2.10 Map of recovery for Beaker-yielding sites in Ireland with outline enclosures for concentrations   51 2.11 Interpreted territories for halberd-bearing and Beaker groups during the Chalcolithic and earliest Bronze Age in Ireland, c. 2450–2050 BC   52 2.12 Map of recovery for burials or human skeletal remains most securely attributable to the Chalcolithic period in Ireland in relation to the hypothesized halberd/Beaker boundary   53 2.13 Areas with more than occasional evidence for formal burial in Ireland during the Early Bronze Age, c. 2150–1500 BC   54 2.14 Map of recovery for Atlantic halberds and Beaker sites of all phases in France   55 2.15 Map of recovery for halberds and Palmela points in France, also showing the changing regional balance between the two types   58 2.16 Map of recovery for Atlantic halberds and Beaker sites of all phases in Iberia   60 2.17 The hoards from Vélez Blanco (?), Almería, and Finca de la Paloma, Toledo   61 2.18 Schematic interpretation of the strength of typological links between regional Atlantic halberd groups and contemporary halberd groups to the east   66 2.19 Graves and probable grave groups datable to the Chalcolithic and transition period in Britain   68 2.20 The locations of Beaker-yielding sites in relation to the motorway network in Ireland 70

  [ vii ]

FRONT MATTER CW3.indd 7

31/05/2016 16:08:47

Table 3.1  The principal regional traditions of 4.6 Photographs of mine entrance and excavated megalithic architecture in Atlantic Europe   85 Early Bronze Age alder wood launder lying in situ. within Comet Lode Opencast 116 3.1 The three themes discussed regarding the re-use of megalithic monuments during the Beaker 4.7 Survey map of Copa Hill, Cwmystwyth, period 87 showing position of EBA mine, cairns, and site of disc find 117 3.2 Complete Beaker pots placed in the blocking deposits at the entrance of Herdade das Casas 4.8 Banc Tynddol disc detail 119 do Canal, Estremoz, Portugal   90 4.9 Banc Tynddol gold disc 119 3.3 Re-use of orthostats to block the entrance 4.10 Excavation plan and section of the Banc to the gallery grave of Méréaucourt, Somme, ­Tynddol grave 120 France   92 4.11 Pair of gold discs from Cabeceiras de Basto, 3.4 Blocking and covering of the passage graves at Portugal 123 L’île Cairn, Finistère, France   93 4.12 Gold disc from Kirk Andrews, Isle of Man 123 3.5 Beaker re-use of the angled grave of Goërem, 4.13 Gold disc from Cobham, Kent 124 Morhiban, France   93 4.14 Distribution of gold disc and basket ornament 3.6 Final deposition of disarticulated Beaker burials finds across Britain and Ireland 128 in dolmen of Monte Abraão, Sintra, Portugal, followed by final closure   94 4.15 Finds of gold lunulae in Britain and Ireland 129 3.7 Decommisioning of stelae during closure of 4.16 Alluvial and primary gold deposits in Britain the tomb of Lagunita, Cáceres, Spain   97 and Ireland 129 14 3.8 Re-use of menhir in the blocking deposit 4.17 Bayesian modelled C dates for the UK 131 at the entrance to the tomb of Llano de la 4.18 Early Bronze Age monuments and finds in the Belleza, Huelva, Spain   97 Plynlimon area, including early mines and flint 3.9 Distribution of megalithic monuments in ­arrowhead finds 134 Atlantic Europe with evidence of re-use in 5.1 Irish prehistoric boats discussed in the text 141 the Beaker period   98 Table 5.1  Published AMS results referenced in text 142 3.10 Some of the radiocarbon date sequences for Table 5.2  New AMS results referenced in text 143 Iberian megalithic monuments with evidence for re-use in the Beaker period 100 5.2 Distribution of Beaker pottery in Ireland, including discoveries in megalithic 3.11 Some of the radiocarbon date sequences for monuments 146 Iberian megalithic monuments with evidence for re-use in the Beaker period 101 5.3 Distribution of potential Chalcolithic/Beaker period burials in Ireland 148 4.1 Cobble stone mining tools and Bronze Age mines in Britain and Ireland 112 5.4 Disarticulated inhumation of adult female in cist-like end chamber of the wedge tomb at 4.2 Prehistoric mines and mineral veins of Labbacallee, Co. Cork 152 Central Wales 112 5.5 Cist-like compartments in wedge tombs: 4.3 Photograph taken in the 1990s of the Comet (a) Labacallee, Co. Cork; (b) Largantea, Co. Lode Opencast and prehistoric mine tips on Derry; (c) Loughash/Cashelbane, Co. Tyrone; Copa hill, Cwmystwyth 114 (d) Ballyedmonduff, Co. Dublin; (e) Lough 4.4 Roof of mine gallery worked with stone Gur, Co. Limerick; (f) Baurnadomeeny, Co. tools within the Comet Lode Opencast, Tipperary; (g) Ballybriest, Co. Derry 154–5 Copa Hill 114 5.6 Sections through megalithic cists and cairns 4.5 Archaeological section cut through peaty excavated at (a) Coolnatullagh, Co. Clare and sediments within the Copa Hill opencast 115 (b) Moneen, Co. Cork 157

  [ viii ]

FRONT MATTER CW3.indd 8

31/05/2016 16:08:47

5.7 Possible Chalcolithic/Beaker period burials in pits at (a) Treanbaun, Co. Galway, (b) Harlockstown, Co. Meath, and (c) Mell, Co. Louth 160 5.8 Early Bronze Age Cists at (a) Carrickinab, Co. Down, (b) Kinkit, Co. Tyrone, (c) Ballybrennan, Co. Westmeath, and (d) Straid, Co. Derry 166 6.1 Map of southern Iberia with toponyms mentioned in the text 180 6.2 Periodization of the Bronze and Early Iron Ages in SW Iberia 181 6.3 Distribution of Middle Bronze Age funerary sites in SW Iberia 182 6.4 Atalaia (Ourique, Baixo Alentejo), Middle Bronze Age grave system IV 184 6.5 Alfarrobeira (São Bartolomeu de Messines, Algarve), reconstruction of agglomerated Middle Bronze Age cist cemetery  185 6.6 FBA (Group II) slabs and stelae: 1 Granja de Céspedes; 3 Brozas; 4 Solana de Cabañas; 5 Magacela; 6 Ategua 186 6.7 Atlantic and Urnfield elements in the Final Bronze Age of southern and south-western Iberia: 1–3 Type Grigny winged axes and singlelooped palstave from Arroyomolinos, Jaén; 4 & 5 RSFO style decorated urn and ProtoVillanova razor from near Beja, Baixo Alentejo; 6 Type Herzsprung shield and Group C I helmet on stela from Santa Ana de Trujillo, Cáceres; 7 Type Rosnoën sword on stela from Fóios, Sabugal, Beira Alta; 8 Type Rixheim sword from the Ría de Huelva, Huelva 188 6.8 Examples of FBA and EIA urn types: 1 & 2 biconical urns; 3 & 4 á chardon type urns; 5 & 6 Cruz del Negro type urns 189 6.9 Distribution of Final Bronze Age funerary sites in SW Iberia 191 6.10 Pego da Sobreira (Ourique, Baixo Alentejo), Early Iron Age funerary structure 192 6.11 Vaga da Cascalheira (Ourique, Baixo Alentejo), Early Iron Age agglomerated dry-stone cist burials 193 6.12 Distribution of Early Iron Age funerary sites in SW Iberia 194 6.13 MBA (Group I) slabs and stelae: 1 Mombeja I; 2 Assento; 3 Gomes Aires; Ervidel I 197

6.14 EIA (Group III) slabs and stelae: Mealha Nova I; Bastos; Fonte Velha VI; Vale dos Vermelhos III 197 6.15 Chronology of objects represented on Group I–III monuments 198 6.16 Distribution of Group I monuments and related anthropomorphic stelae 200 6.17 Distribution of Group II monuments, related anthropomorphic stelae and rock-art panels 200 6.18 Distribution of Group III monuments 202 7.1 Distribution of prehistoric hillforts (Class 1 and 2) in Ireland, showing location of excavated sites (this paper) 221 7.2 Clashanimud hillfort, Co. Cork 225 7.3 Defences of inner enclosure, Clashanimud hillfort, Co. Cork 226 Table 7.1  Radiocarbon dates from recent hillfort excavations in Ireland 227 7.4 Calibration of radiocarbon dates, Clashanimud hillfort, Co. Cork 228 7.5 Magnetic gradiometry survey of Glanbane hillfort, Co. Kerry, showing location of excavation trenches across inner and outer enclosing elements 229 7.6 Partial reconstruction of defences of inner enclosure, Glanbane hillfort, Co. Kerry 230 7.7 Lidar image of Ballylin hillfort, Co. Limerick 231 7.8 Entrance to inner enclosure, Ballylin hillfort, 232 Co. Limerick 7.9 Plan of Formoyle hillfort, Co. Clare 233 7.10 Excavation of bank and ditch, outer enclosure of Formoyle hillfort, Co. Clare 234 7.11 Magnetic gradiometry survey of Toor More hillfort, Co. Kilkenny 236 7.12 Partial reconstruction of defences of inner enclosure, Toor More hillfort, Co. Kilkenny 237 7.13 Excavation of burnt palisade trench, outer enclosure, Toor More hillfort, Co. Kilkenny 237 7.14 Lidar image of Rathnagree hillfort, Co. Wicklow 239 7.15 Postholes of defensive line of inner enclosure, Rathnagree hillfort, Co. Wicklow 239

  [ ix ]

FRONT MATTER CW3.indd 9

31/05/2016 16:08:47

7.16 Summary presentation of radiocarbon dates for six hillforts presented in this paper 243 8.1 Copper Group (CG) Definitions. Presence Absence groups based upon Arsenic // Antimony // Silver // Nickel 251 8.2 European copper axes, daggers, and halberds: end of the 3rd millennium BC 252 8.3 The arsenic level of Early Bronze Age Copper Group 12 axes (c. 2200 to 2000 BC) 253 8.4 Copper Group 1: Early Bronze Age, ubiquity in each region 254 Table 8.1  Metal signatures for Late Bronze Age English, Welsh, and French artefact assemblages         255 8.5 Copper Group 16: Early Bronze Age, ubiquity in each region 256 Table 8.2  Analysed objects: a Waterden, Norfolk; b Duddingston Loch, Edinburgh; c Andover, Hampshire; d Stogursey, Somerset; e Guilsfield, Powys; f  Blackmoor, Hampshire; g Bradley Fen, Cambridgeshire; h Isleham, Cambridgeshire 257–8 8.6 Tin content of swords and socketed axes: all analysed British artefacts 259 8.7 Lead content of swords and socketed axes: all analysed British artefacts 259 8.8 Distribution of tin content in the Waterden Hoard, by artefact set 260 8.9 Distribution of tin content in the Bradley Fen Hoard, by artefact set 260 9.1 Late Bronze Age–Early Medieval settlements on the Ll½n peninsula 266 9.2 A view of Penll½n from the summit of Mynydd Rhiw with the site of Meillionydd in the foreground 267 9.3 Interpretation of the magnetometer survey of Meillionydd conducted by Gwynedd Archaeological Trust 268 9.4 Complete GPR interpretation, combining the 269 interpreted data of all depth slices 9.5 Interpretative sketch of the features excavated in trenches 1–5 during excavation seasons 2010–14 at Meillionydd 270 9.6 A modern field wall, next to the site, constructed in the same technique as the banks enclosing the site 272

9.7 Tentative 3D visualization of phase 1 of the Meillionydd settlement, with contemporary buildings at Castell Odo just about visible as dots in the background 277 9.8. Llys Rhosyr, Anglesey 284 9.9. Development of Cefn Graeanog II 285 10.1 Location of Llanmaes in the Vale of Glamorgan and the excavation sites 294 10.2 Geophysical survey coverage by Tim Young (2003–10) for the Llanmaes Research Project 296 10.3 Midden Field Excavation Focus, with surviving midden deposit, radiocarbon dates and the Romano-British pit 298 10.4 Excavation of midden and underlying settlement in 2006 (looking west) 298 10.5 Midden Field Excavation Focus, showing settlement features and location of 300 radiocarbon dated contexts 10.6 The Middle Bronze Age Building and Pit Cremation 302 10.7 Excavated area of the Early Bronze Age outlier burial monument 306 10.8 Percentage of the Number of Identified Specimens (NISP) of pig limb bones (including scapulae and pelves) from each quarter of the carcass 308 10.9 The range of diagnostic cauldron, bowl, and cup fragments from Llanmaes (Inset: one of 310 the Llyn Fawr cauldrons) 10.10 T he range of Armorican and Sompting axes and mouth fragments and items of earlier metalwork 312 10.11 The range of personal ornaments, early iron tools, bone, antler and stone artefacts 313 10.12 T he Pit Cremation pottery group and beads and decorated Trevisker vessels 314 10.13 Illustrations of the diagnostic vessels of Earliest Iron Age date in the pit assemblages 310 10.14 Performance artist Dylan Adams, leading a ceremony around cauldron offerings, in 2008 324 10.15 Still from artist Sean Harris’s animation, showing sacrificed pig in cauldron with shark’s tooth. Creative commission for the Origins: In Search of Early Wales archaeology gallery at the National Museum Cardiff (2007–2014) 325

[x]

FRONT MATTER CW3.indd 10

31/05/2016 16:08:47

Table 11.1  Basic information on numbers, gender, and the age distribution of the samples collected as part of the People of the British Isles is given in the top part of the table. The lower part of the table gives the median and 25% and 75% quartiles of the mean distance between grandparental places of birth for volunteers who gave information for all four grandparents 334 11.1 Distribution of mean geographic position (MGP) of grandparental places of birth for 3,646 volunteers for whom there was information for all four grandparents 335 11.2 Distribution of two surnames throughout Great Britain in 1881 and 1998 337 11.3 Location Quotient (LQ) distributions for a selection of surnames 338 11.4 Flow diagram for method used to classify a surname as either ‘local’ or ‘non-local’ 339 Table 11.2  Proportion of surnames classified as local depending on different Location Quotient (LQ) exclusion criterion. 339 11.5 Graph of the Log(maximum LQ) of the RD with the highest LQ for each surname (y-axis) against Log(surname population size) in the 1881 census (x-axis) 340 Table 11.3  Maximum likelihood admixture estimates using a limited set of markers for the most stringent criteria used to define ‘local’ and ‘non-local’ surnames. The contributions of the putative ancestral populations (East, West and Norse) to the putative admixed popula­tion (Central (CN) or Orkney (OR)) were estimated for either the ‘local’ surnames (L) alone or only the ‘non-local’ (N) surnames. 341 11.6 The effect of linkage disequilibrium (LD) 342 11.7 UK map showing the fineSTRUCTURE clustering of the PoBI samples 344 11.8 Analysis of potential sources of genetic variation due to migration 345 11.9 Renormalization using only the three groups implied by the correlation analysis 346 Table 11.4  Correlations between European groups’ contributions to the UK ancestry profiles 347 12.1 Genome-wide SNP ADMIXTURE analysis of global populations. ADMIXTURE is a ‘STRUCTURE-like’ model-based clustering

algorithm that essentially identifies subgroups of genotypes within a sample set with distinctive allele frequencies 359 12.2 PC analysis performed with XLSTAT, depicting the relationships between mtDNA haplogroup frequency variation in modern-day West Eurasian populations and samples from Early (LBK, Linear Pottery, n = 102; Rössen, RSC, n = 12; Schöningen, SCG, n = 33), Middle (Baalberge, BAC, n = 19; Salzmünde, SMC, n = 29; Bernburg, BEC, n = 17) and Late (Corded Ware, CWC, n = 44; Bell Beaker, BBC, n = 29) Neolithic, and Early Bronze Age (Unetice, UC, 367 n = 94) burials in Germany 13.1 Island of monolingual speakers of Red and its various social domains 390 13.2 Blue speakers settle on the island and establish their language in both their settlements and in some external domains. 390 13.3 Red speakers begin to become bilingual and utilize the Blue language in some nondomestic social domains. 390 13.4 Societal subtractive bilingualism emerges among former Red speakers who widely employ the Blue language in most external domains and some utilize Blue in their own households. 390 13.5 Language death. The Red language has ceased to be employed in all domains and dies with its last elderly speaker. 390 13.6 False problem: how could a small number of Blue speakers achieve language shift over a much larger Red-speaking population? 394 13.7 Blue-speaking immigrants initially interface with a restricted number of Red speakers. 394 13.8 A migration stream continually augments the number of Blue speaking immigrants who can sustain their language. 394 Table 13.1  Hypothetical confrontation between two languages in terms of social domains 401 13.9 A map of some potential domains based on Grogan’s (2013) model of the lower strata of Late Bronze Age social relations in Ireland 402 13.10 A map of some potential domains based on Grogan’s (2013) model of the higher strata of Late Bronze Age social relations in Ireland 402

[ xi ]

FRONT MATTER CW3.indd 11

31/05/2016 16:08:47

Table 14.1  Similar features in Insular Celtic and Hamito-Semitic according to author 414 Table 14.2  Similar features in Insular Celtic and Hamito-Semitic according to language 415 Table 14.3  Verbal noun or infinitive? 418 14.1 Celtic peoples during the second Iron Age 425 15.1 Distribution of Atlantic Late Bronze Age metalwork and the survival of non-IndoEuropean languages around the western Mediterranean 435 15.2 Distribution of Gündlingen swords 436 15.3 The ‘brig\ zone’ of the Iberian Peninsula 438 15.4–6  Bows and arrows as depicted on Iberian LBA warrior stelae 440 15.7 Detail of the bow and arrow from the LBA ‘warrior’ stela from Montemolín, Sevilla superimposed on a photograph of the prominent hill at Monreal de Ariza (Arcobriga), Zaragoza, Spain, viewed from the ruined Roman town 441 15.8 The south-western Iberian Peninsula in later prehistoric and protohistoric periods, showing the locations of carved stelae 451 15.9 Ancient Cyprus and the north-east Mediterranean, places mentioned in the text and sea route to the west 453 15.10–12  Two views of one of the cast-bronze lions’ heads adorning the funerary vehicle from ‘royal’ Tomb 79 of Salamis, Cyprus and reconstructed funerary chariot, showing castbronze lion’s-head hubcap from Tomb no. 17 of  the La Joya necropolis in Huelva 454 Table 15.1  Proposed antecedents for the SW pseudo‑semisyllabary 455 15.13 The SW inscribed stone from Madroñero, Cáceres, Spain (J.56.1) 460

Table 15.2  SW texts used for the statistics 462–3 15.14 Drawing of the SW inscribed stone from Alcalá del Río, Sevilla, Spain (J.53.1) 464 Table 16.1  Palaeohispanic personal names probably derived from Celtic ‘dog’ 478 Table 16.2  Celtiberian genitive plural family names, probably derived from Celtic ‘dog’ 478 16.1 kon- personal names / kun- personal and family names. 479 16.2 koun- personal and family names. From west to east: Counedioq(um) (Segovia), Couneancus (Cluniensis), and burikounikum and kounesikum (Botorrita, Zaragoza) 481 Table 16.3  Names attested in the central and eastern Peninsula, probably based on the root ‘dog’, represented with ou 481 Table 16.4  Palaeohispanic group names in Kunand Kon- from Greek and Roman sources 483 Table 16.5  Ancient town names of the Iberian Peninsula with Koni- 484 499 17.1 Proposed Italo-Celtic family tree 18.1 The Vasconic languages 505

Front cover: Early Bronze Age ‘sun-disc’ from Banc Tynddol, Cwmystwyth, Ceredigion, Wales: ©Amgueddfa Cymru – The National Museum of Wales. Back cover: reconstruction drawing of the Banc Tynddol disc by Tony Daly: ©Amgueddfa Cymru – The National Museum of Wales. The editors gratefully acknowledge the Museum’s diligence and generosity for providing these images. [cover design by J. T. Koch]

[ xii ]

FRONT MATTER CW3.indd 12

31/05/2016 16:08:47

INTRODUCTION The AEMA project

T research project, ‘Atlantic Europe in the Metal Ages (AEMA): questions of shared language’, funded

his book is an output of the inter­disciplinary

from 2013–16 by the UK’s Arts and Humanities Research Council. The AEMA project has two main lines of inquiry: (1) gathering and interpreting archaeo­logical evidence for inter-regional connections in Atlantic Europe (Ireland, Britain, Greater Armorica, and the western Iberian Peninsula) during the Copper and Bronze Ages c. 2900–800 BC; (2) gathering Ancient Celtic written evidence relevant to the recon­ struction of Proto-Celtic, paying full attention to the Palaeohispanic data and the archaeological background of inscriptional finds, such as re-use within multiperiod contexts.   The ultimate question motivating the project is how, where, and when the Proto-Celtic branch emerged from Proto-Indo-European. In other words, it seeks an archaeological context for a known historical linguistic process of later prehistory. As clearly envisioned by Mallory (Chapter 13), such a goal faces a daunting obstacle in that these are questions about language in a time and place that by definition lack written records. The methodology is to pursue an in-depth and big-data amplification of two observations that were possible on the level of a synthetic overview and led to the ‘Celtic from the West’ idea several years ago (see Cunliffe 2001, 293–7; Cunliffe & Koch 2010). First, a detailed survey of the cultures of Atlantic Europe (Ireland, Britain, and the ocean-facing sides of Gaul and the Iberian Peninsula) comprises numerous unmistakable examples of the peoples of the region generally ‘doing the same thing at the same time’— sharing materials, values, ideologies, technologies and other practices—from the megalithic age down to the end of Late Bronze Age, with a special intensity in the latter period. As opposed to a portable object that can be copied, a complex idea—such as a new religion or the entire metallurgy package—usually requires words

1_12 CW3 Introduction 18 v 16.indd 1

for it to be shared. Thus, for the complex idea to spread between communities with different lan­guages bilingualism is required. A complex idea is not expected to jump over a gap where there is no shared lan­guage. In a world without writing, both a shared language or language(s) and mobility are required; complex information must be communicated face-to-face.   Second, at the next stage, when there is writing, Ancient Celtic languages are found in all parts of Celtic Europe: ogamic Primitive Irish, Ancient Brythonic, Gaulish, Celtiberian, and other forms of HispanoCeltic in the western Iberian Peninsula, including the Celtic now widely recognized in the South-western (SW) or ‘Tartessian’ inscriptions of the Early Iron Age. When they first appear, these languages are still similar, but they are no longer the undifferentiated Proto-Celtic (PC) that linguists reconstruct; instead they are a family of separated languages. In the archaeo­logical record of the Iron Age, we also find fragmentation of the Atlantic cultural continuum. Western Iberia leaves the Atlantic Bronze Age when Phoenician influence appears strongly in Huelva c.  900 BC (cf. Chapter 15). Ireland appears isolated after the end of the Dowris phase c. 700 BC. Taken together, the combined Bronze Age, Iron Age, and linguistic evidence suggests that the shared language of the Atlantic Bronze Age commonality had been the reconstructed PC ancestor of the languages later attested there amongst the cultural fragmentation of the Iron Age.   We did not go into the AEMA project expecting a ‘eureka moment’, that a specific item of evidence would be discovered to resolve whether Proto-Celtic had been the lingua franca of Atlantic Europe before the Bronze–Iron Transition. Rather, the strength and explanatory value of the circumstantial case could be better evaluated. Exhaustive collections of high-quality data—recording numerous parameters of context and subtype—render objective and quantifiable the

31/05/2016 16:26:44

[2]

celtic from the west 3]]

degrees to which people across the region were ‘doing the same things’. The inclusion of hundreds of highquality AMS (accelerator mass spectrometry) dates clarifies how precisely at the same time these ‘same things’ were done.   With the language evidence, large collections of reassessed Ancient Celtic place-names and inscriptions (on not easily moved stones) containing Ancient Celtic forms allow refined geographical distributions. These can be juxtaposed with the mapped archaeological finds to reconsider how precisely these correlate. It is also possible to take into account the archaeology of the linguistic evidence—the site to which the name refers or where the inscription was found and that site’s history and its neighbourhood’s. Does this context imply anything about continuity from later prehistory? Or is there evidence for discontinuity or cultural inflow in the Iron Age? An illustration of cultural revival over the threshold of literacy is provided by Brandherm (Chapter 6): a Middle Bronze Age tradition of cist burials and associated stelae under circular pavements in southern Portugal was revived in the Early Iron Age, at which time the stelae had writing on them, now widely recognized as recording Celtic names. Though the writing itself is a result of Phoenician impact, everything else in the package speaks of a backwardlooking nativism.   The principal institutional collaborators in the AEMA project are the University of Wales Centre for Advanced Welsh and Celtic Studies, the School of Archaeology at the University of Oxford, the Department of Digital Humanities at King’s College London (KCL), Bangor University, Amgueddfa Cymru – National Museum of Wales, and the National Library of Wales. Through a work plan spanning the years 2013–2016, the AEMA research team have analysed and made comparable archaeological and linguistic evidence across Atlantic Europe. The data is entered together on a GIS (Geographic Information Systems) enabled database, developed in collaboration with Neil Jakeman at KCL. On 6 April 2016 the database was launched as an open-access website: www.aemap.ac.uk .   Users can now interrogate this data by selecting one of the principal categories: burials, individuals, pots, other grave goods, metalwork, stelae, inscriptions,

1_12 CW3 Introduction 18 v 16.indd 2

toponyms, or settlements. Several levels of parameters can be selected for any of these. After choosing parameters, occurrences of overlapping categories of data can searched for positively and exclusively ( = both X and Y), inclusively ( = X or Y), or negatively () for the defined feature’s absence. The requested data can be displayed as lists of sites, graphs showing comparative frequency of multiple parameters, or distribution maps (exportable as GIS files in .json format). Dots on the maps can also be individually selected to open the full record. This multidisciplinary resource is intended for researchers in various fields with various aims, not only for users focused on questions of language in prehistory. For this core question, discoveries in archaeogenetics (now expected to follow Haak et al. 2015; Allentoft et al. 2015; Cassidy et al. 2015) should soon contribute significantly, by showing to what extent those prehistoric groups in Atlantic Europe sharing cultural phenomena also shared common ancestry, with each other and with later communities whose languages are known.

Celtic from the West 2016: older thinking, newer thinking, and archaeogenetic breakthroughs The studies in this book are relevant—some directly, some tangentially—to the ‘Celtic from the West’ (CW) idea. That idea has attained some currency since first formulated in Facing the Ocean (Cunliffe 2001, 293–7), as well as the first two volumes in this series, Celtic form the West (2010) and Celtic from the West 2 (2013; see also Cunliffe 2013, 242–9; cf. Gibson & Wodtko 2013). There are now various understandings of CW. Some of these diverge from what we intended, and some can be quickly ruled out as impossibilities, though even so useful in making obvious unexamined modes of thought contributing to the longevity of the concept of a Celtic homeland equated with the earliest Iron Age in west-central Europe.   One recurrent misunderstanding can be epitomized as: since Celtic is an Indo-European (IE) language and the IE homeland was east of Atlantic Europe, how could Celtic possibly have spread from west to east? The impeding concept seems to be that communities

31/05/2016 16:26:44

introduction

and their descendants only move in a constant outward direction away from where their ancestors had lived, not unlike lifeless matter propelled by the Big Bang. The consideration does not enter that the bestdocumented and greatest migration of the Celts in ancient times flowed eastward down the Danube, into the Balkans, then Delphi,1 and onwards to central Anatolia, that is, to more or less the same place as the source zone of European agriculture and one of the debated IE homelands.   Archaeologists frequently find themselves at cross purposes with linguists, the latter understanding ‘Celts’ as shorthand for ‘speakers of Celtic languages’, in which case Celts before Celtic languages is nonsense. Without this understanding, the languages might be thought of as another attribute (like La Tène art) later acquired by a group that had existed previously. With so much interest in Celtic identity, Celtic archaeology, Celtic art, Celtic mythology, and so on—and there is nothing inherently wrong with any of that (cf. Karl 2010)—it is easy to lose sight of the fact that the modern concept of what is Celtic depends on the discovery of the family of languages and its naming as ‘Celtic’ by Edward Lhuyd in 1707. To think of Celts as a group existing before the PC branch separated from the PIE trunk is to allow the concept to break free from its scientific definition.   Therefore, for us and for the definition of CW , Celts will mean ‘speakers of Celtic languages’, as in the following formulation: The PIE parent language reached Atlantic Europe as PIE and then evolved into Celtic there. It did not undergo the linguistic innovations defining Celtic (such as the weakening of *p) in some secondary homeland (such as central Europe) and then move west in a secondary migration.   A second basic thread of the CW idea is, as stated above: ‘Proto-Celtic had been the lingua franca of Atlantic Europe before the Bronze–Iron Transition’. 1 According to Timagenes (as quoted by Strabo 1.12–13) some of the treasure captured at Delphi in 278 BC was deposited in the sacred pools of the Volcae Tectosages in Toulouse in south-west Gaul, near the Pyrenees. The Greek historians understood this episode as the Celtic raiders returning home.

1_12 CW3 Introduction 18 v 16.indd 3

[3]

  Can we be more specific about the Celtic homeland (that is, where PIE evolved into PC) in the CW hypothesis? It should be mentioned at this point that many historical linguists have proposed that PC evolved from a post-PIE Italo-Celtic node, rather than directly from PIE (see Schrijver Chapter 17). So, where did PIE or Proto-Italo-Celtic evolve into PC? In the original formulation of C, Europe’s Atlantic façade as a whole was viewed as a suitable setting for the emergence of Celtic as the region’s lingua franca (see Cunliffe 2013, fig. 7.5). Although this idea came out of archaeology, there is no inherent linguistic reason that a shared language could not have developed in a large area from a continuum of mutually intelligible dialects in a situation of intense and regular contact, while also overwhelming any other languages in the area. In fact, this is what is happening with most successful majority languages today: dialects (and minority languages) are declining in favour of a popular form of the standard language due to intensification of regular longdistance contacts.   A sentence in Celtic from the West raised the possibility of a smaller homeland and has found some subsequent resonance: No one has taken the possibility of Celtic coming from Hispania to the other Celtic countries seriously since we stopped taking Lebar Gabála Érenn (the 11th-century Irish ‘Book of Invasions’) seriously, but is now at least worth pausing to review what it is we think we know that makes that impossible. (Koch 2010, 295) It should be emphasized that this remark was not meant as an understatement. An Iberian homeland for Celtic is thinkable, but it had not been thought about recently. However, ‘Celtic out of Iberia’ is not another name for CW , but a narrow possibility within the scope of the general proposition. Even so, is the Atlantic side of the Iberian Peninsula (a.k.a. the IndoEuropean side or ‘briga zone’; Figure 15.3) closer to an appropriate size than all or most of Atlantic Europe? The former is nearer in size to that of the cradle of the Celts commonly appearing on maps where this is equated with the Hallstatt C1a along the upper Danube.   With that older model, parallel expansive processes occurred twice: first, PIE spread from a compact

31/05/2016 16:26:44

[4]

celtic from the west 3]]

homeland, then its daughter PC evolved in its compact homeland, followed by a second expansion into new lands and differentiation into grand-daughters— Goidelic, Brythonic, Gaulish, &c. In fact, essentially the same scenario occurs several times in this traditional mode of thought. With several secondary homelands, the cradles for Proto-Greek and Proto-Indo-Iranian, for example, the coming of the Celts to the Atlantic west parallels the coming of the Greeks to the Aegean and of the Aryans to Iran and India.   Underlying these ideas is the attraction of a model in which the history of IE-speaking peoples resembles the family tree of the IE languages superimposed onto a map of western Eurasia. The tree fans out to represent linguistic distance. This easily slips into an assumption of peoples steadily radiating outward from the homeland. Protolanguages that are represented as point-like nodes on the tree suit homelands that are geographically small. Favouring the same result, the best-known fully documented example of a protolanguage diversifying into a language family is the Romance family traceable to the expansion of Latin with the Roman Empire, all beginning at the city of Rome and its immediate vicinity, accurately enough represented on a map of Europe with a single dot.   As an alternative, Koch (2013) adopts the idea of Garrett (2006) concerning the formation of the IndoEuropean branches from a shallowly-differentiated Late PIE dialect continuum. Such a model is consistent with a rapid and extensive PIE expansion followed by the formation of a geographically large PC branch from contiguous dialects. Developments along these lines are also consistent with recent archaeogeneticbased conclusions (Haak et al. 2015; Allentoft et al. 2015; Cassidy et al. 2015). These studies recognize two major migration events contributing to the genetic structure of western Eurasia: a renewal of population emanating from the Near East with the beginning of farming, followed by an expansion from the Pontic– Caspian steppe at the transition of the Neolithic to the Bronze Age. Cassidy et al. find both transformations in Ireland. That study also finds discontinuity between the genetics of the Neolithic woman from Co. Down and those of the three Early-to-Middle Bronze Age men from Rathlin Island. In contrast, there is

1_12 CW3 Introduction 18 v 16.indd 4

significant continuity between the Rathlin samples and the modern population of Ireland (and to a considerable degree also those of Scotland and Wales). The Rathlin DNA contains the ‘steppe component’2 at a high percentage, and the Neolithic woman does not.   With the limited ancient DNA (aDNA) evidence available so far, the ‘dog that didn’t bark’ is the apparent lack of comparably transformative events datable to the Late Bronze Age (c. 1250–800 BC) or the Iron Age (from c. 800 BC), the periods at which the traditional model envisions speakers of Ancient Celtic migrating westward from Urnfield/Hallstatt/La Tène west-central Europe. Without genetic evidence for a major migration from central Europe to the Atlantic after c.  1250 BC, a sociolinguistic mechanism of élite dominance (that is, a new language taking hold with few new people) is needed to sustain the traditional model. Involving a second homeland and a second age of expansion, this old model is less economical than a hypothesis in which the IE branches evolved in situ from a dialect continuum after a single sustained PIE expansion event, which some now think most probably occurred during the Copper Age.   While the most usual mechanism of language transmission is from parent to child, élite dominance can occur. Thus, many people who spoke Latin as their first language in Late Antiquity had no ancestors from Rome. Many people living outside Europe today speak English, Spanish, Portuguese, or French as their first language, but have no ancestry from the European homeland of their language. However, these counterexamples result from the formation and break-up of metropolitan empires. As the opening pages to general books about the Celts often explain, they never built a great empire.   It is of course early days for collecting genomewide aDNA. We can look forward to a tenfold increase in the near future. However, as Mallory (2013a) has recently cautioned, it is easier to construct a big picture connecting dots when there are fewer dots. The genetic samples yet to come may not all fall along the lines already drawn. Variables remain in areas pivotal for the Indo-European homeland 2

termed ‘South-west Asian’ in Chapter 12.

31/05/2016 16:26:44

introduction

question, such as north India, the Iranian plateau, southern Europe, and central Anatolia. It is also less than certain that the Afanasievo culture of the Altai of the early 3rd millennium BC must be ancestral to the Tocharian speakers who left manuscripts 1000 km to the south in the 1st millennium AD, or that their language was the second Indo-European branch to split off after Anatolian. Although the recent aDNA research has been recognized thus far as consistent with the traditional model of a PIE homeland on the Pontic–Caspian steppe at Copper Age date, Renfrew’s (1987; 2013) alternative Anatolian farming hypothesis did anticipate the picture emerging to date in two important respects: first, in predicting a major renewal of population across Europe from the Near East at the beginning of agriculture; and secondly, in seeing PIE evolving into Celtic in situ across western Europe, without a secondary homeland and secondary age of expansion during the Iron Age.   As both general editors envision it, the principal artery of communication, for which PC came to function as the lingua franca, was the Atlantic coastal route itself. The navigable rivers leading into the interior were secondary. ‘Primary’ and ‘secondary’ in this formulation imply a relative, but not an absolute, chronology. The linguistic implications will differ depending on how much real time was involved. If the spread up the rivers took place gradually over a millennium or more, then the mutual intelligibility between dialects of Late PIE would probably have broken down. So one IE language (Celtic) replaced another, such as Proto-Italic, Proto-Germanic, or the lost dialect or late PIE stage sometimes identified with the alteuropäisch river names (Krahe 1964; Kitson 1996; Bichlmeier 2013). For such a case of language replacement, we would look for some compelling advantage for the Atlantic language, an imbalance favouring élite dominance.   On the other hand, if the coastal network spread up the rivers not so long after PIE had reached the Atlantic façade, say 300–500 years afterwards, or less, there would still have been a high degree of mutual intelligibility. At this point, the linguistic innovations defining the Celtic branch might have spread through a chain of communities in contact speaking their own first language, Late PIE (or Proto-Italo-Celtic).

1_12 CW3 Introduction 18 v 16.indd 5

[5]

Such a situation would be more analogous to that of the children of speakers of dialects of southern rural English who sound more like Londoners and not to that of children of Welsh speakers who speak only English. A sociolinguistic process of dialect convergence is less stark than one of language replacement, requiring neither mass migration nor a precipitous élite dominance.

The Chapters Chapters 1–16 reflect presentations given at three events: the forum on ‘Ancient Britons, Wales, and Europe—New Research in Genetics, Archaeology, and Linguistics’ held in the Reardon Smith Lecture Theatre, National Museum Cardiff, in June 2011; the forum on ‘New Light on the Ancient West—Recent Work in Archaeology, Genetics, and Linguistics on Later Prehistory and Protohistory’ held at the National Library of Wales in Aberystwyth in November 2012; and the workshop and forum called ‘Atlantic Europe in the Metal Ages—questions of shared language’ held at Cardiff University in April 2014. Chapter 1 focuses on Beaker burials in northern France in an investigation of cultural identity. Although other researchers have seen the Beaker phenomenon as a whole as a possible manifestation of the Proto-Indo-European or Proto-Celtic language, Laure Salanova finds such sweeping conclusions unconvincing. Within Bronze Age Europe, the rise of warrior phenomenon—for which the Beaker burials do indeed provide clear evidence—is too widespread to be confidently equated with a particular language. Within the Beaker region, there are significant differences between central Europe and the Atlantic region. On the other hand, she finds credible the possibility of ‘supra-regional languages’ in use among traders and craftsmen for the exchange of specialist information along the Atlantic corridor. Chapter 2 presents new data for the halberds of the Beaker period and transition to the Early Bronze Age in Britain, Ireland, France, and the Iberian Peninsula. A re-examination of the geographical distribution and

31/05/2016 16:26:44

[6]

celtic from the west 3]]

dating evidence shows the halberds to have inverse distributions to those of characteristically ‘Beaker’ material, including Beaker burials and Palmela points, the latter only in France and Iberia. This inverse pattern is clearest in Britain and Ireland, though demonstrable throughout Atlantic Europe. Stuart Needham argues for an ideology of the halberd bearers developing in opposition to that linked to the Beaker package. A negative attribute associated with the former is a lack of burials in the halberd-bearers territory, which is suggestive in view of how this same negative characteristic is seen as significant in the Atlantic Bronze Age. The continued use or re-use of earlier megalithic monuments for Beaker burials, discussed also in Chapters 1 and 5, is the main focus for Catriona Gibson in Chapter 3. The evidence considered is mostly from the Iberian Peninsula, where numerous recently-obtained, high-quality AMS dates now reveal a recurrent sequence implying an ideological shift marked by a conspicuous and intentional transformation of the ritual landscape. In many examples, formal closure of megalithic tombs coincides with the abandonment of enclosed settlements in the same area at the time of intensifying exchange networks during the later 3rd millennium BC, about 400 years after the earliest Beakers on the lower Tagus. Chapter 4 concerns the arrival of metallurgy and the Beaker package in western Britain. Among several early mining sites in Ceredigion, Simon Timberlake gives special attention to the Copa Hill workings at Cwmystwyth. New evidence, including AMS dates with Bayesian modelling, favour an ‘out of Ireland’ scenario for the incoming technology and new people, in the sequence ‘Ireland > Central Wales > North Wales > NW England’. Copper prospecting began at Cwmystwyth in the period 2200–1900 BC, before activity had ceased at Ross Island, Co. Kerry, and before full-scale production began at Great Orme near Llandudno on Wales’ north coast. This account is consistent with the testimony of the 19th-century find, near Cwmystwyth in Pontrhydygroes, of a halberd of A1 copper of Ross Island provenance. The Banc Tynddol gold sun-disc (featured as cover images for this book)

1_12 CW3 Introduction 18 v 16.indd 6

was found near the Copa Hill site in a probable grave. There are Atlantic parallels of the late 3rd millennium BC for this object, especially looking towards Armorica and Iberia. The disc can be seen as evidence for mobile metal-using newcomers arriving around the time that mining began there. In Chapter 5 Kerri Cleary’s examination of burial practices in Ireland during the late 3rd millennium BC throws new light on two phenomena that arrived, evidently together, c. 2450/2400 BC: copper metallurgy and the Beaker complex. There are numerous instances over the next centuries of megalithic traditions re­inter­preted in response to the incoming ideology, such as burials ‘individualized’ in cists or ‘cist-like compartments’. The inclusion of 18 new and 63 previously-published AMS radiocarbon dates clarifies the transformations occurring during this period, notably the emergence of new ceramic types c. 220/2150 BC. Ireland was clearly connected to the world overseas during the late 3rd millennium BC, and evidence for ocean-going vessels is also considered. The overview of ground transport includes what are now the earliest directly dated horse remains in Ireland (in light of horse from Newgrange once assigned to the Beaker period now having a Late Iron Age date). These belong to the Middle Bronze Age, though other horse bones associated with earlier dated material are discussed. In Chapter 6 Dirk Brandherm builds a case for an Early Iron Age revival of a Middle Bronze Age burial rite in southern Portugal. This development is interpreted as a nativist reaction to the Southwestern Orientalizing Complex, as found ascendant in neighbouring regions east of the Guadiana. Those Iron Age necropolises most resembling their forerunners of 500–1000 years before include find spots of SW stelae with writing. The nativist interpretation of these burials thus carries implications beyond archaeology concerning the origins of the earliest of the Palaeo­ hispanic scripts, as employed on the SW stelae, and the names recorded in the inscriptions, and their language as a whole. Brandherm also argues against the sometimes contradictory use of the term Tartes(s)os for phenomena in the southern Iberian Peninsula of the

31/05/2016 16:26:44

introduction

Late Bronze Age and Early Iron Age.3 Chapter 7 takes as its keynote a recent proposal by Mallory (2013b, 278) that the rise of the hillforts could reflect centres of power through which foreign influence introduced a new language to multiple social domains in Ireland c.  1000 BC. William O’Brien summarizes the investigation of six large, multipleenclosure hillforts in the southern half of Ireland, carried out from 2004–11. These produced numerous 14 C dates, consistent with beginnings in the Middle Bronze Age (late Bishopsland phase ~ British Penard, c.  1400–1150 BC). Irish hillforts have correctly been linked with the arrival and early proliferation of the sword c. 1300–1100 BC. The sites continued to develop until an ‘apparent collapse of hillfort culture’ in the 8th century BC (during the Dowris phase). ‘In the absence of firm evidence for a significant intrusive population’ during this period, O’Brien withholds support from the theory of the hillfort as the social vector for the Celticization or Indo-Europeanization of Ireland. Chapter 8 presents evidence for the chemical compo­ sition of British Late Bronze Age swords gathered by Peter Bray at the Oxford Archaeology Laboratory as a member of the AEMA team. Bray begins by discussing the pervasive idea that the sword was an object of special cultural significance and the resulting expectation that the production of swords had differed from that of other metal objects, being the work of craftsmen with special social status, who were either more wide ranging than other metal workers or less so, being permanently attached to particular chieftains. Despite these expectations, an analysis of the chemistry—taking into account four frequent elemental ‘impurities’ and the effects of remelting and mixing—shows that axes and swords had the same chemistry and metal histories, thus failing to support a unique role for the sword smith. The central focus of Chapter 9 is the occupation history of Iron Age hillfort of Meillionydd in north Wales, including a discussion of its environs on the 3 Readers with a background in Celtic studies may find this dispute reminiscent of the prolonged and ongoing dis-ease over the term ‘Celtic’.

1_12 CW3 Introduction 18 v 16.indd 7

[7]

Ll½n Peninsula. From this evidence, Raimund Karl develops a theory of the social significance of enclosures, as these become significant features in landscape from the Late Bronze Age. Crossing disciplines, he analyses the Welsh word llys ‘enclosure, court’ and its Celtic cognates. Evidence from the Welsh laws shows that the meaning ‘place of judgement and authority’ predates the semantic influence of AngloNorman and English court.4 In Chapter 10, Adam Gwilt, Mark Lodwick, Jody Deacon, Nicholas Wells, Richard Madgwick, and Tim Young present finds from a major excavation and public engagement project undertaken between 2003 and 2010 by Amgueddfa Cymru – National Museum Wales and supported by the Portable Antiquities Scheme in Wales. The site at Llanmaes in the western Vale of Glamorgan comprised a midden of Earliest Iron Age date overlying an open settlement that began in the Middle Bronze Age. The wide range of finds from the midden include varied metalwork of the Llyn Fawr period (c. 800–600 BC). There is a remarkable preponderance of pig amongst the faunal domesticates, at 71%, reflecting hundreds of animals of all ages, carrying implications for feasting practices. Chapter 11 by Bruce Winney and Walter Bodmer presents results and interpretations based on the high-quality and extensive genetic data collected by the Oxford-based People of the British Isles (PoBI) project, ongoing since 2004. Aiming to investigate UK population structure at a finer level, the project is an important milestone and demonstrates discernible differences between the historically Anglo-Saxon and various Celtic regions reflected in the genetics of the present-day inhabitants. The PoBI data set will be of continuing significance as ancient DNA becomes more available and diagonal comparisons with modern evidence will aid essentially in its analysis. 4 This interpretation can also be supported by documents concerning sites called lis (plural lisiou) in early medieval Brittany, as residences where rulers with the title machtiern made judgements and conducted business ­relating to communities of the status designated by the term plebes in Breton Latin (Davies 1988, 138–42).

31/05/2016 16:26:44

[8]

celtic from the west 3]]

The three disciplines involved in the studies here— archaeology, genetics, and linguistics—are progressing on the levels of primary data and methodology, but not at the same rate. Maria Pala, Pedro Soares, and Martin Richards convey the rapid, and now rapidly accelerating, advances in the genetic study of the human past. Chapter 12 provides an accessible overview of progress in the field of archaeogenetics, from its beginnings in the later 20th century. We are now poised on an important threshold. Yet fundamental questions remain concerning the population associated with the Beaker phenomenon—at its earliest stages in the Iberian Peninsula and after it had spread widely in the second half of the 3rd millennium BC—and what this might imply for the linguistic prehistory of Atlantic Europe. On familiar interdisciplinary terrain, J. P. Mallory begins Chapter 13 restating the challenge facing any attempt to map linguistics onto prehistoric archaeology. From the premise that some language other than Celtic, or the Indo-European that became Celtic, had once been spoken in Atlantic Europe, he concludes that language shift had occurred. This was more probably a process of élite dominance than a turnover of demography/subsistence or, least probably, a system collapse. He notes that the branch of linguistics that customarily deals with later prehistory is habitually separate from those dealing with the sociolinguistics of bilingualism and second language acquisition. He sets up a model of four or more generations, leading from a situation in which a new language is introduced by a minority of newcomers, is then gradually taken up in some social domains by their indigenous neighbours, subsequently spreading to more domains and communities, then taken up as the household language by bilinguals, followed finally by the death of the indigenous language (corresponding to the pre-Celtic language or languages of Atlantic Europe). For more than a century, leading scholars of the Celtic languages have observed, and sought to explain, parallels in syntactic features of the medieval and modern Goidelic and Brythonic languages with several non-Indo-European languages of the Middle East and

1_12 CW3 Introduction 18 v 16.indd 8

north Africa. These same features are generally absent from the other Indo-European languages. Repeatedly, these similarities have been seen as a substratum effect, the basis for a hypothesis that the Insular Celtic languages overlie a language or languages typologically similar to the Hamito-Semitic. One peculiarity of the published research on this ‘Hamito-Semitic hypothesis’ is that its fullest presentation is the 1993 doctoral dissertation of Gensler, which remains unpublished. Therefore, Steve Hewitt’s account in Chapter 14 is particularly useful, in providing a history of the idea and systematic listing of the diagnostic Hamito-Semitic/Insular Celtic features. The re-examination is also timely, as old assumptions are now questioned about where and when the attested Celtic languages evolved from Proto-Indo-European. Hewitt’s comparative examples are drawn mostly from Breton and Arabic. Since formulated in Cunliffe (2001), the Celtic from the West hypothesis has envisioned that Proto-Celtic was in use as the lingua franca behind observable interactions defining the Atlantic Late Bronze Age, if not earlier still in Atlantic Europe. In Chapter 15 John  T. Koch considers a range of linguistic evidence that sits more easily with this idea than with the traditional model. (1) The primary dialect cleavage in the Celtic languages between Hispano-Celtic and Gaulish-Brythonic-Goidelic can be correlated with the precocious departure of the Iberian Peninsula from the Atlantic Bronze Age catalysed by the arrival of Phoenicians by c. 900 BC. (2) Of the western IndoEuropean branches (Celtic, Italic, and Germanic) the consonant system of Celtic has evolved in a direction closest to that of the ancient non-Indo-European languages of south-west Europe, Iberian and PalaeoBasque. (3) The common Hispano-Celtic personal name Arquius ‘archer’ is derived from a western Indo-European root with no Goidelic or Brythonic counterpart; bows and arrows died out in the Middle Bronze Age in Britain and Ireland, but continued as a high status weapon throughout the Iberian Bronze Age. (4) The Celtic content of the SW inscriptions can be defended on the basis of scholarly consensus and the linguistic evidence. Whether this corpus is seen as containing Celtic elements or simply being

31/05/2016 16:26:44

[9]

introduction

written in a Celtic language, this testimony is more consistent with the Celtic from the West model than with a central European homeland. In Chapter 16 Fernando Fernández Palacios draws on extensive research on Palaeohispanic names with both the Hesperia project in Madrid and the AEMA project in Wales. He finds indigenous personal, family, and place-names to be plausibly derived from the Celtic word for ‘dog, wolf ’. A challenge in identifying these forms is that the inherited paradigm had three vowel grades: *k~, *kun-, *kon-, and probably also a derived form with full-grade *koun-. This is not the only set of Palaeohispanic names with closer parallels in Brythonic and Goidelic than in Gaulish, which raises the interesting question of whether they represent peripheral survivals of usages that had declined in Gaul or evidence for direct contact by sea. One long-standing unresolved problem of IndoEuropean linguistics is whether the Celtic and Italic branches (the latter being the ancestral group of Latin) descend from a single protolanguage, ProtoItalo-Celtic. If this was the case, the implications would be incompatible with the idea that Celtic was an eastern Indo-European language as favoured by Isaac (2010) and the late K. H. Schmidt (2012). In Chapter 17 Peter Schrijver presents a linguistic case for an Italo-Celtic protolanguage. He finds the closest connections of Celtic with Venetic and Sabellian. On this basis he seeks the Proto-Celtic homeland in Italy and argues for an identification with the Canegrate Urnfield culture of the Italian Late Bronze Age c. 1300–1100 BC, situated north of the upper Po, thus overlapping the territory of the Lepontic inscriptions. This model implies subsequent expansions of Celtic ‘out of Italy’ to central Europe, Gaul, the Iberian Peninsula, Britain, and Ireland.5 5

The general editors note that the linguistic argument for an Italo-Celtic protolanguage is not otherwise seen as demanding a Celtic homeland in Italy. The Celtic from the West hypothesis also predicts shared innovations with Italic, as a neighbouring IE branch in south-western Europe, whether or not the branches had begun as a unified post-IE node.

1_12 CW3 Introduction 18 v 16.indd 9

Theo Vennemann is well known for the ‘Vasconic theory’, which he restates with new arguments in Chapter 18. This theory is essentially that a language family whose only surviving member is Basque in the western Pyrenees was once widespread over western and central Europe before the arrival of Indo-European. There is an inherent plausibility in that IE is rarely regarded as native to western Europe,6 and there is no decisive evidence that Basque came to the region from somewhere else after IE. Vennemann follows and extends the case introduced in Chapter 15 for the evolution of Proto-Indo-European into Proto-Celtic reflecting contact with the p-less Palaeo-Basque and/ or Iberian language. Vennemann continues to favour a central European homeland for Proto-Celtic.7

6 7

Barry Cunliffe, Oxford John T. Koch, Aberystwyth April 2016

Note, however, the so-called ‘Palaeolithic Continuity Paradigm’ (see Renfrew 2013, 212). However, given the historical locations of the Basque, Aquitanian, Iberian, and Celtic languages, the general editors note that this theory of phonetic influence suits the Celtic from the West model as well or better. Apart from Lepontic, the influence of the non-IndoEuropean languages of southern central Europe, Etruscan and Rhaetic, is apparently absent from the Celtic languages.

31/05/2016 16:26:44

[ 10 ]

celtic from the west 3]]

BIBLIOGRAPHY Allentoft, M. E., M. Sikora, K.-G. Sjögren, S. Rasmussen, M. Rasmussen, J. Stenderup, P. B. Damgaard, H. Schroeder, T. Ahlström, L. Vinner, A.-S. Malaspinas, A. Margaryan, T. Higham, D. Chivall, N. Lynnerup, L. Harvig, J. Baron, Ph. Della Casa, P. Dąbrowski, P. R. Duffy, A. V. Ebel, A. Epimakhov, K. Frei, M. Furmanek, T. Gralak, K. Kristiansen 2015 ‘Population genomics of Bronze Age Eurasia’, Nature 522, 167–72. Bichlmeier, H. 2013 ‘Analyse und Bewertung der sprachwissenschaft­lichen Standards aktueller Forschungen traditioneller Art zur “alteuropäischen Hydronymie” aus der Perspektive der heutigen Indogermanistik’, Namenkundliche Informationen 101/102, 397–437. (online unter: www.namenkundliche-informationen.de) Bosch Gimpera, P. 1942 The Two Celtic Waves in Spain, Sir John Rh½s Memorial Lecture 1939, Proceedings of the British Academy 26. London. Cassidy, L. M., R. Martiniano, E. M. Murphy, M. D. Teasdale, J. Mallory, Barrie Hartwell, & D. G. Bradley 2015 ‘Neolithic and Bronze Age migration to Ireland and establishment of the insular Atlantic genome’, Proceedings of the National Academy of Sciences USA 113/2, 368–73. Cunliffe, B. 2001 Facing the Ocean: The Atlantic and its Peoples 8000 BC–AD 1500. Oxford, Oxford University Press. Cunliffe, B. 2013 Britain Begins. Oxford, Oxford University Press. Cunliffe, B. & J. T. Koch (eds) 2010 Celtic from the West: Alternative Perspectives from Archaeology, Genetics, Language, and Literature. Oxford, Oxbow Books. Davies, W. 1988 Small Worlds: The Village Community in Early Medieval Brittany. London, Duckworth. Garrett, A. 2006 ‘Convergence in the Formation of IndoEuropean Subgroups: Phylogeny and Chronology’, Phylogenetic Methods and the Prehistory of Languages, ed. P. Forster & C. Renfrew, McDonald Institute Monographs, 139–151. Cambridge, McDonald Institute for Archaeological Research. Gensler, O. D. 1993 ‘A typological evaluation of Celtic/ Hamito-Semitic syntactic parallels’, unpublished Ph.D. dissertation, University of California, Berkeley, available from University Microfilms International, Ann Arbor, Michigan, No. 9407967. Extracts printed in Karl & Stifter 2007, Vol. IV, 151–229. Gibson, C. D. & D. S. Wodtko 2013 The Background of the Celtic languages: Theories from archaeology and linguistics. Aberystwyth, University of Wales Centre for Advanced Welsh and Celtic Studies. Haak, W., I. Lazaridis, N. Patterson, N. Rohland, S. Mallick, B. Llamas, G. Brandt, S. Nordenfelt, E. Harney, K. Stewardson, Q. Fu, A. Mittnik, E. Bánffy, C. Economou,

1_12 CW3 Introduction 18 v 16.indd 10

M. Francken, S. Friederich, R. Garrido Pena, F. Hallgren, V. Khartanovich, A. Khokhlov, M. Kunst, P. Kuznetsov, H. Meller, O. Mochalov, V. Moiseyev, N. Nicklisch, S.L. Pichler, R. Risch, M. A. Rojo Guerra, C. Roth, A. Szécsényi-Nagy, J. Wahl, M. Meyer, J. Krause, D. Brown, D. Anthon, A. Cooper, K. W. Alt, & D. Reich 2015 ‘Massive migration from the steppe was a source for IndoEuropean languages in Europe’, Nature 14317. Isaac, G. R. 2010 ‘The Origins of the Celtic Languages: Language Spread from East to West’, Celtic from the West. Alternative Perspectives from Archae­o­logy, Genetics, Language, and Literature, eds B. Cunliffe & J. T. Koch, 153–67. Oxford, Oxbow Books. Karl, R. 2010 ‘The Celts from Everywhere and Nowhere: A re-evaluation of the Origins of the Celts and the Emergence of Celtic Cultures’, Celtic from the West: Alternative Perspectives from Archaeology, Genetics, Language and Literature, Celtic Studies Publications 15, eds. B. Cunliffe & J. T. Koch, 39–64. Oxford, Oxbow Books. Kitson, P. R. 1996 ‘British and European River-Names’, Transactions of the Philological Society 94/2, 73–118. Koch, J. T., & B. Cunliffe (eds) 2013 Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe. Oxford, Oxbow Books. Krahe, H. 1964 Unsere ältesten Flußnamen, Wiesbaden, Otto Harrassowitz. Lhuyd, E. 1707 Archaeologia Britannica, Giving some Account of the Languages, Histories and Customs of the Original Inhabitants of Great Britain: from Collections and Observations in Travels through Wales, Cornwal, Bas-Bretagne, Ireland and Scotland. Vol. I. Glossography. Oxford, The Theater. Mallory, J. P. 2013a ‘Twenty-First Century Clouds over IndoEuropean Homelands’, Journal of Language Relationship 9, 145–54. Mallory, J. P. 2013b The Origins of the Irish. London, Thames & Hudson. Renfrew, A. C. 1987 Archaeology and Language: The Puzzle of Indo-European Origins. London, Cape. Renfrew, A. C. 2013 ‘Early Celtic in the West: The IndoEuropean Context’, Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, Celtic Studies Publications 16, ed. John T. Koch & Barry Cunliffe, 207–217. Oxford, Oxbow Books. Schmidt, K. H. 2012 ‘Armenian and Celtic: Towards a New Classification of Early Indo-European Dialects’, Ireland and Armenia: Studies in Language, History and Narrative, Journal of Indo-European Studies Monograph Series 61, eds M. Fomin, A. Jivanyan & S. Mac Mathúna, 51–63. Washington D.C., Institute of the Study of Man.

31/05/2016 16:26:44

part one ARCHAEOLOGY

1_12 CW3 Introduction 18 v 16.indd 11

31/05/2016 16:26:44

1_12 CW3 Introduction 18 v 16.indd 12

31/05/2016 16:26:44

chapter one

Behind the warriors: Bell Beakers and identities in Atlantic Europe (third millennium B.C.) Laure Salanova

1. Introduction

T

he definition of identities, both geographically and chronologically, constitutes a founding principle of the historical disciplines, especially for periods without written resources. In archaeology, this question is generally approached from a burial context, in which the differences between individuals are the most visible. The architecture of the grave, its visibility in the landscape and the time and effort involved in its building, in addition to the quality and quantity of grave goods deposited with the dead, allow different groups to be distinguished among the population. For the oldest periods of history, it is clearly more difficult to detect these differences in domestic contexts, as the houses forming a dwelling generally indicate uniformity in the way of life of their inhabitants. This gap between the observations made in burial and domestic contexts demonstrates that our understanding of the diversity of social categories composing past societies is far from complete (Hofman & Bickle 2011). This situation likely explains the difficulties in defining some cultural patterns from archaeological remains. Hence, this contribution proposes decomposition of the various social and cultural identities recognized throughout the Bell Beaker phenomenon, which marked European countries with a uniform material culture (beakers, copper daggers, wrist guards, arrowheads) from the Atlantic shores to Poland and from North Africa to Scotland and Norway for several centuries of the 3rd millennium BC (Salanova 2005).   Because of the wide distribution of the Bell Beaker ‘phenomenon’, an extensive range of publications has focused on the political and social aspects of Europe during the 3rd millennium BC. These publications generally take into account the most common artefacts in Europe and the most visible: the objects connected with a male and warrior ‘ideology’ (defined by Strahm 1995). However, many questions remain regarding the rest of the population. For example, did this ideology concern all categories of the society, and how did the contemporaries of the warriors consider their identities?

Salanova.indd 13

31/05/2016 15:57:44

[ 14 ]

i. behind the warriors

]]

2. From archaeological remains to identities The ambiguity of the term identity, referring both to individual behaviours and to collective norms, has been highlighted many times, as have the difficulties in differentiating all scales of identity in archaeological contexts (Díaz-Andreu & Lucy 2005; Fowler 2010).   Since the advent of post-processual archaeology, material culture has been analysed regarding the social and symbolic meaning it could have represented for the population who produced and used it (Hodder 1982). Nevertheless, some other components of archaeological remains, even in a funerary context, reveal another aspect of identity, perhaps the truest part, because they are impossible to transform. 2.1 Interdisciplinarity and the quest for identity The anthropological analyses currently being undertaken on the dead (analyses of aDNA, stable isotopes, non-metric traits, and activity stress markers) allow reconstruction of the characteristics of each individual and their relationships with the other persons in a group, often demonstrating multi-identities rather than uniformity among a community. For instance, the existence of a low frequency of migrants among graves from central European cemeteries, despite the homogeneity of the material culture that defines an Eastern Bell Beaker Province, has been demonstrated using several methods of physical anthropology (Price et al. 1998; 2004; Desideri 2011).   These interdisciplinary approaches allow archaeologists to be distinguish which part of the identity was biological and which part was influenced by social and cultural circumstances, necessarily linked to a contextual framework (chronology and environment). In parallel with intellectual progress, these approaches make the analyses of identity more complex. Determination of the interactions between all of the components identified, which are all essential for understanding the organization of a society, could become impossible without a hierarchy of information levels.   This confusion regarding the various scales of the identity (individual, social and cultural) likely explains the different interpretations proposed for the Bell Beaker phenomenon, a term that will be used in this contribution when referring to the global situation of Europe during the second half of the 3rd millennium BC and the most common elements unifying a wide area under a standardized social pattern. Some scholars consider the Bell Beakers to represent a material cultural phenomenon whose widespread presence provides evidence of ‘a loose sense of ethnic boundaries and identities’ (Fowler 2010, 363). However, deconstruction of this phenomenon is necessary to indicate the different levels of the societies from the 3rd millennium BC that were affected. The results reveal an opposite pattern to a loose sense of identities. Compared with previous periods, the Bell Beakers reflect a more pronounced expression of social and cultural diversity, laying the basis for Bronze Age societies and even more recent ones (Salanova 2007).

Salanova.indd 14

31/05/2016 15:57:44

Salanova

[ 15 ]

2.2 Social and cultural identities How can we avoid this frequent confusion when approaching identity between social and cultural dimensions? Concerning the social dimension, the Nouvelle Histoire trend, and the New Archaeology one, have provided methods for approaching past societies with a focus on social and intellectual progress rather than political events. Among the different publications that could provide inspiration, the book written by Le Goff (1989) is probably the best reference for a list of all of the categories that should have existed in preindustrial societies, whatever the region of Europe. These categories refer to fundamental components we can imagine to be universal, for example, persons in charge of religion, warriors, farmers, traders, craftsmen and artists. It is probably a utopian expectation to hope to obtain such a variety for a period that did not deliver written sources. Even for historians, some parts of the population are inaccessible when resources are overly scarce. Nevertheless, we can attempt to go beyond traditional social analyses performed in archaeological contexts, which are generally focused on one or two of the categories that should be detected. This restriction is increasingly being denounced in publications. For instance, in his analysis of the Nordic Bronze Age, Kristiansen distinguished three different social identities according to the goods deposited in graves, mainly based on the types of swords that are present (Kristiansen 2011, 202–3). • The ritual chief, buried with a special package of objects, characterized by exclusive use of spiral decorations, interpreted as symbols of the sun cult and of Nordic identity; • The warrior chief, buried with an undecorated sword of international type, often resharpened, to be used for combat and, hence, associated with professional war; • A group of migrant smiths linked to trade, characterized by swords of German origins. Despite this great effort to detect coexisting forms of identities, the author hopes for a deeper decoding of the Bronze Age society, as ‘Hundreds of other objects types from prehistory are waiting for a similar contextual interpretation of their social and institutional meaning’ (op. cit., 204). This is a difficult wish to grant, as these objects likely did not only fulfil a social function. The question of cultural identity has progressively been abandoned, and it is often considered to be an old-fashion way to consider archaeological remains for reconstructing the history of a population (Roberts & Vander Linden 2011). However, in my view, cultural identity, which I define as a set of common values and ways of life, is more difficult to elucidate than social identity. This task supposes the need to take in to account an extended range of data on the organization of societies and their eventual repetitive forms, not only comparing several different geographical areas, but also analysing continuities and discontinuities through different periods to ultimately reveal patterns. Several scholars have attempted to carry out this exercise with Bell Beaker data (see for instance Besse 2003; Vander Linden 2004), taking in to account typological elements. This contribution proposes another way of defining cultural patterns during the Bell Beaker period.

Salanova.indd 15

31/05/2016 15:57:44

[ 16 ]

i. behind the warriors

]]

3. Bell Beakers, social identities and burial practices The societies from the 3rd millennium BC are still mainly known through their graves. The distribution of burial structures divides Europe into two zones. In the eastern and northern parts of Europe, individual burials are the rule. In the Atlantic and Mediterranean part of Europe, the burial tradition did not change from previous periods: collective graves were used until the end of the 3rd millennium BC. France, similar to England and Scotland, is located in the middle of these two traditions. Recent discoveries have highlighted numerous individual burials from the Bell Beaker period, mainly in the continental part of the French territory (Salanova & Tchérémissinoff 2011). These graves provide the most precise data on the dead and their status in the society. 3.1. Biological identity and social status In France, the most numerous individual graves have been found in the northern half of the country, where 44 burials are known (Salanova 2011). Their distribution shows a concentration eastwards from the Seine valley (Fig. 1.1). Radiocarbon dates indicate an extended period for these burials, between 2600 and 2000 cal BC.   When the preservation state of skeletons allows precise anthropological analyses, it is possible to observe a balance between the sexes of the dead. Among the 35 burials for which this information is known, 15 men, 11 women and 9 children were buried with Bell Beakers (Salanova 2011, tab. 3).

1.1 Distribution of the Bell Beaker graves in the northern half of France. Red dots: collective graves with Bell Beakers. Black rectangles: individuals Bell Beaker graves. From Salanova 2011

Salanova.indd 16

31/05/2016 15:57:46

Salanova

[ 17 ]

Sexual distinction The balance between men and women in Bell Beaker graves is not only characteristic of the north of France. The same proportion has been observed in Spain, Britain and the Netherlands (Salanova 2007, 217). Although the orientations of the bodies are more diverse than in central Europe, the position of the dead follows the same rules: men are placed on the left side and women on the right side. Sexual distinction therefore appears to be an essential criterion in these individual graves. This is not the case in contemporary burials within collective graves, where kinship relationships and chronological factors are more important in explaining the spatial distribution or the position of the dead, rather than their sex (see below).   The grave goods deposited with adults constitute another difference between the collective and individual contexts. During the 4th and the 3rd millennia BC, some men, some women and even some children could have received special treatment among the dead deposited in collective graves, although it is generally impossible to understand the reason for these particularities, which may take the form of a specific place inside the monument, away from others, or a profusion of ornaments, despite the usual scarcity of grave goods in this kind of burial (Chambon 2003; Salanova & Sohn 2007). Among the individual Bell Beaker graves from western Europe, the richest graves are always male burials (Salanova 2007). Women are generally buried with beakers and sometimes ornaments, but their graves never contain weapons, which are elements of the classical Bell Beaker package. The status of children The graves of children point to a particular situation. According to data from northern France, children represent 16% of individual Bell Beaker burials. This rate is considered to be low according to reference curves for pre-industrial societies, where the mortality rate for young people should have been at least 50% in the population (Ledermann 1969). For instance, this is the case among burials in collective graves from the 4th millennium BC and the beginning of the 3rd, which show such a normal mortality curve: 46% mortality among children under 12 years old for the first period of burials (3700–3600 cal BC) at Hazleton North (Meadows et al. 2007); approximately 50% for the first period of burials (3300–2800 cal BC) at La Chaussée-Tirancourt in northern France (Leclerc & Masset 2006); and 54% for the earliest phase of burials at Les Mournouards in north-eastern France (Blin 2012). Thus, not all of the children from the second half of the 3rd millennium were buried with Bell Beakers; instead, these children were selected, most likely from a specific social group.   In western Europe, Bell Beaker children were never buried with the package characteristic of warriors or with grave goods, which are normally reserved for adults. They did not appear to have the same status as in central Europe, where children are sometimes associated with the package of the warriors, and some objects are even disproportionate compared with their size (Turek 2000; Heyd 2007, 354). Despite this fact, the graves of Bell Beaker children from the west often contain specific wares. They are generally associated with nothing more than beakers. However, in England, it has been suggested that children were buried with smaller pottery than adults (Case 1995). In France, a technological analysis of Bell Beakers deposited in individual graves, taking in account the quality of the vessels,

Salanova.indd 17

31/05/2016 15:57:46

[ 18 ]

i. behind the warriors

]]

also revealed their particular careful manufacture when linked to children (Salanova 2012).   The special treatment of children in Bell Beaker burial practices, some of whom were clearly selected from among the population, and the specific wares deposited in their graves indicate that the social position of these Bell Beaker children was acquired by filiation, calling into question the link between social identities and kinship networks (Brodie 2001). 3.2. Kinship structures The family character is another difference that differentiates the social meaning of the Bell Beaker graves from central and western Europe. Small cemeteries found in the Upper Danube, containing approximately twenty graves, are assumed to represent family-based social units (Heyd 2007, 351–2). The same situation is observed in the Czech Republic, where the spatial distribution in Bell Beaker cemeteries involves groups of graves organized in separated areas, which are interpreted as family groupings (Matéjíčková & Dvořák 2012).

Bell Beakers and families In western Europe, family groupings of individual Bell Beaker graves do not exist. The graves are generally isolated, containing only one individual. The scarce cases of double

1.2 Inhumation layer from the Bury gallery grave (Picardie, France). On the left side, the wall effect of the decomposed wooden chamber. Photo L. Salanova

Salanova.indd 18

31/05/2016 15:57:50

1.3 Family cells from the La Chaussée-Tirancourt gallery grave (Picardie, France). From Leclerc & Masset 2006: fig. 19

or triple graves involve adults (mainly women) associated with young children, including new-borns (Salanova 2007, 216). This situation contrasts with the now well-established family groupings of the dead in collective graves, which were still in use during the Bell Beaker period. For instance, the collective grave found at Bury (Picardie, northern France) was constructed around the second half of the 4th millennium BC, like all of the collective graves from the Paris Basin. Its architecture is characterized by a long wooden chamber (20 m long) bordered by dry stone walls and an axial entrance marked by a perforated slab (‘trou d’homme’ ). Approximately 300 individuals and several dogs were deposited in a long-term sequence, which carried on until the end of the 3rd millennium BC (Fig. 1.2). The radiocarbon dates of most of the dead have revealed that the use of this grave was discontinuous, with an interval without burials being observed at the end of the 4th millennium BC and the beginning of the 3rd millennium BC. However, the good preservation of bones and the absence of collapse observed in the chamber demonstrated that the community maintained the architecture, probably for ceremonial activities (Salanova & Sohn 2007). Biological analyses of these skeletons show that some non-metric traits and activity markers distinguish certain individuals deposited in the monument at the back end of the chamber throughout the entire sequence (Salanova & Chambon et al. in press). This finding demonstrates that the location of the dead in the grave was defined by biological filiation, which simultaneously determined their activity and role in the society.   The spatial distribution of the burials found at Bury finds echoes that in other collective graves, such as La Chaussée-Tirancourt (Picardie, northern France), which is located in the same region as Bury. In this grave, the burials from the second period (3rd millennium BC) are organized in separated cells containing individuals with similar non-metric traits (Leclerc & Masset 2006). Consequently, each cell could have represented a family vault (Fig. 1.3). The family characteristics of the burial practices are not specific to northern France. The megalithic necropolis found at Sion Le Petit-Chasseur in Switzerland provided the same results concerning the social meaning of these monumental constructions.   The necropolis of Sion exhibits an interesting sequence of burial practices, from the

Salanova.indd 19

31/05/2016 15:57:51

[ 20 ]

i. behind the warriors

]]

Middle Neolithic (approximately 4500 cal BC) until the Early Bronze Age (first half of the 2nd millennium BC). Five collective graves (dolmens MI, MV, MVI, MXI and MXII) date from the Late Neolithic and the Bell Beaker period (Gallay 1995, 170–1). The oldest constructions (MXII and MVI) are characterized by small stone chambers (approximately 3 m2), settled in a triangular base consisting of dry stones, measuring approximately ten meters in length. Several stone cists were subsequently built around the dolmens, mainly during the second half of the 3rd millennium BC, but also during the Bronze Age. Some of these cists contain several individuals, while others contain only one, but children are frequently buried in these small constructions (Bocksberger 1976). The Sion necropolis is also famous for the anthropomorphic stelae found near the entrance of its megalithic tombs or reused in the construction of monuments during the numerous reorganizations of the site. These stelae, the tallest of which are approximately 2 m high, represent male figures with weapons and ornaments (Gallay 1995). Two stylistic types have been defined according to the design pattern of their carved decorations. Type A, dating from the Late Neolithic, is characterized by spiral ornaments and daggers similar to the Italian culture of Remedello. Type B is assigned to the Bell Beaker period because of the archery symbols (bow) and the geometric designs of their clothes similar to the pottery decoration. The Sion necropolis is not an isolated case. The same architecture and stelae have been found in northern Italy, in the Aosta Valley, but the anthropological analysis of these findings is still unpublished (Mezzena 1997).   At Sion, application of the anthropological approach to the 90 individuals deposited in dolmen MXII between 3200 and 2500 cal BC has demonstrated restricted access of the monument and selection of the buried population on the basis of kinship criteria (Mariéthoz 2011, 189). These findings could explain the differences in cranial and dental features observed between the dead buried in dolmen MVI and dolmen MXI throughout the sequence in the 3rd millennium BC, including the Bell Beaker period (Desideri & Eades 2004). Each monument could have been dedicated to one family that settled in the region. The anthropomorphic stelae are assumed to express this lineage structure, representing the mythical ancestor of the family vault (Gallay 2011, 236). A long-term perspective The observed family distribution among monuments does not differ significantly from previous periods, in which stone cists of the Chamblandes type (4000–3600 cal BC) were grouped in small cemeteries, sometimes in a settlement context, containing all of the age groups of a population (Mezzena 1997; Von Tobel 2011). Even the stelae linked to each of the monuments from Sion Le Petit-Chasseur are known to mark the location of some Chamblandes cists as early as the Middle Neolithic (Fig. 1.4). Once again, this characteristic is not specific to this region, as the erection of stelae near the entrance of megalithic tombs has also been recorded in the Atlantic world since the 5th millennium BC (Rodríguez Casal 1988). This phenomenon is comparable to the small anthropomorphic schist plaques deposited in the hundreds in the megalithic tombs at Alentejo (Portugal), which date from the 5th–4th millennia BC (Gonçalves 1999). This finding shows that throughout western Europe, the tradition of anthropomorphic representations associated

Salanova.indd 20

31/05/2016 15:57:51

Salanova

[ 21 ]

with a reference to a common ancestor began quite early, during the 5th millennium BC, with megalithic constructions. The graphic style of the stelae evolved with the introduction of some representations of weapons, from the second half of the 4th millennium BC in the Iberian Peninsula, linking the symbol of the ancestor with symbols of power and the warrior ideology, which remained important in the iconography until the Late Bronze Age (Bueno Ramírez et al. 2005).   This tradition of marking the entrance of the burials with an anthropomorphic figure appears to be a common trend in Europe that is quite old, for which it is difficult to point to an origin. Its signification refers more to the kinship structures of the population, which affect the social organization over time. Construction of social identities from the 3rd millennium BC The results of the analyses conducted in burial contexts from the 3rd millennium BC have important consequences for the approach applied for the social construction of identities, especially during the Bell Beaker period.   First, the Bell Beaker package could have been used for accelerating a process of social distinction, but it occurred in relative continuity compared with the previous periods in western Europe, where the old tradition of collective graves was maintained. In 90% of cases, the Bell Beaker grave goods from France, Spain, and Portugal were found in collective monuments built one or two millennia earlier, integrating a long-term tradition of burial practices and extending the use of family vaults all along the Atlantic coast, and even on the Mediterranean coast (Salanova 2007).   Second, anthropological analyses have allowed us to understand the criteria on which the identity of 1.4 Anthropomorphic stelae from the 5th and 4th millennium BC. 1: stelae found near the entrance of the Parxubeira passage grave (Galicia, Spain). 2: Marking stelae of Chamblandes cists in Switzerland. 3: Schist plaque from the Anta 2 da Comenda (Alentejo, Portugal). From Rodríguez Casal 1988; Moinat & Studer 2007 ; Gonçalves 1999

Salanova.indd 21

31/05/2016 15:58:00

[ 22 ]

i. behind the warriors

]]

each social group is based. The distribution of the buried population, either inside or by the monument, is strongly determined by biological and family relationships. Biological filiation appears to have simultaneously determined the activity and the role of each person in the society, as demonstrated in the collective grave at Bury (northern France), where the locations of the dead in some areas are linked both to biological and social (or professional) factors.   Third, in collective graves that were still being used during the second half of the 3rd millennium BC, burials deposited with some elements of the Bell Beaker package are associated with burials without grave goods. The radiocarbon dates of these skeletons demonstrate a strict contemporaneity of these two practices (Salanova 2007; 2011). It should be noted that before the Bell Beaker phenomenon, the identity of the dead, related to biological and social status, was discreet and was not generally marked by specific grave goods or other material markers in collective graves. This situation changed with the advent of the Bell Beaker package, especially in some inland regions in contact with central European groups, such as eastern France, where the return to individual burials appears to take place earlier.   The social organization observed in the collective graves calls into question the status of the persons buried in individual Bell Beaker graves. In some French regions located near the Seine Valley, these graves are occasionally situated close to collective graves that were still in use (Salanova 2007). The individual graves from western Europe are never organized in cemeteries, and they are generally isolated. They contain individuals who clearly had no access to the tombs with a long-term tradition representing collective graves. While some burials found in a collective context that are radiocarbon dated to the second half of the 3rd millennium BC contain no or few grave goods, still showing relatively egalitarian treatment of the dead, similar to previous periods, the contemporary individual Bell Beaker graves are always associated with items referring to professional activities that are not significantly different from those described for the Bronze Age (Kristiansen 2011). 3.3. Professional groups and supra-regional identities Because of the great diversity of the Bell Beaker graves in western Europe, it is not easy to identify recurrences that could point to clear groups, the most obvious of which is represented by copper dagger graves (Salanova 1998). These graves, which always contain male skeletons, have a special status among the dead, probably ranking somewhere at the top of the hierarchy that defines the society from the 3rd millennium BC. The warriors The association of a copper dagger with wrist guards and flint arrowheads defines the classical Bell Beaker package (Fig. 1.5). However, this type of association is scarce: only a few graves from western Europe contain the full package (Wallers, northern France; Lunteren, Netherlands; Fuente Olmedo, Spain; Amesbury, England). The same situation is observed in central Europe: the graves with the weapon set represent only a small percentage of the

Salanova.indd 22

31/05/2016 15:58:00

1.5 The Bell Beaker set from the Wallers individual grave (Nord, France). Photos L. Salanova

discoveries made to date (Brodie 2001; Matéjíčková & Dvořák 2012).   In the west, these ‘warrior’ graves are not only distinguished by their grave goods. The copper dagger is correlated with strict funerary rules: male skeletons are oriented from north–south, with the head to the north, in exactly the same way as in the male graves from central Europe (Salanova 2003). This is the only group among the Bell Beaker graves for which there are similar burial practices between western and central Europe. The dagger graves are also the richest, sometimes containing a great many items, as in Amesbury.   Were these men buried with a copper dagger professional warriors? The copper dagger found in the Bell Beaker individual graves from Wallers (Nord, France) has been used and sharpened several times (Salanova et al. 2011). Moreover, it has been demonstrated for previous periods that the social status expressed in the grave goods can represent the actual activity of the dead during his life. For example, in the Cerny burials from the Paris Basin

Salanova.indd 23

31/05/2016 15:58:04

1.6 High quality (1) and clumsy Bell Beakers (2). From Salanova 2012

(French Middle Neolithic, 5th millennium BC), some dead associated with arrowheads present activity markers that are characteristic of archery (Chambon & Thomas 2010). It is therefore plausible that the men buried with copper daggers in a Bell Beaker context were effectively warriors during their lives.   Turek (2011) highlighted an interesting point concerning the appearance of new warrior symbols with the Bell Beaker package. During the first half of the 3rd millennium BC, the Corded Ware Cultures and their westward extensions (Single Grave Cultures) are well known for the battle-axes and mace heads associated with male burials. These weapons suggest a form of man-to-man combat based on physical strength. The archery symbols characteristic of the Bell Beaker graves (wrist guards and flint arrowheads) refer to another type of war, implying some distance between combatants and recognition based more on accuracy and skills than strength.   These graves certainly prefigure the professional warrior described by K. Kristiansen (2011) for the Bronze Age, despite the fact that it is not possible to detect a hierarchy among the Bell Beaker burials, such as the difference between ritual and warrior chiefs. However, in the same way as in the Bronze Age, the Bell Beaker warrior graves refer to foreign traditions of burial practices and grave goods, clearly indicating their affiliation with a supra-regional identity. Specialist craftsmen In addition to the warrior identity, some graves refer to other professional groups. Unsurprisingly, the only detectable activities (probably because they were important enough to be expressed through grave goods) concern novelties from the 3rd millennium

Salanova.indd 24

31/05/2016 15:58:05

Salanova

[ 25 ]

BC: the production of metal and Bell Beaker pottery. However, in western Europe, these categories of individuals in the population are not visible in graves, but through the items they produced, which are associated with the richest burials. According to the reference model described by Costin (1991), I have recently proposed that this situation should be interpreted as the existence of part-time specialist craftsmen working with and closely linked to the chiefs of the 3rd millennium societies.   The pottery deposited in Bell Beaker graves reveals the existence of this type of specialist. The differences in the manufacturing process of the Bell Beaker pottery found in domestic contexts compared with that found in burials have been highlighted by several authors (for instance Clarke 1976, 464; Salanova 2000a, 173–6). In burials, the styles and shapes of the Bell Beakers are more standardized. However, their technical quality shows great variability, with some pottery being perfectly shaped and finely decorated, while some reflect clumsy production (Fig. 1.6). There is no correlation between the style of the pots and their quality, nor between the style and the sexes of the dead (Salanova 2012). However, the frequency of twin beakers in graves, i.e., two pots presenting the same typology and the same mode of production, and the fact that the highest quality vessels are always found in burial contexts suggest the possible existence of certain kinds of production dedicated to the dead and, hence, of a form of craft specialization (Salanova 2002). The most carefully made beakers, with regular profiles, proportions and decorations, are generally associated with the richest graves, particularly the ‘warriors’’ graves (Salanova 2012). This correlation between the skill of potters and the identity of the dead demonstrates the existence of a category of specialized potters who precisely controlled the mode of production for the Bell Beakers and who could have been called on a case-by-case basis in specific contexts. If the entire production process does not reflect the existence of specialists, as imagined by Clarke (1976), one component of this production, which is easy to detect according to quality markers, points to a category of highly skilled potters. These high-quality products are scarce, but a coherent group among the Maritime type has been recognized from Brittany to Portugal (Salanova 2000b; 2003; Prieto-Martínez & Salanova 2009; 2011). They share not only manufacturing quality, but also some technical criteria, such as shell-impressed

1.7 Maritime beakers from Portugal (on the left), Galicia (in the middle), and Brittany (on the right). Drawings and photo from Leisner & Leisner (1959), Salanova (2000a), Prieto Martínez & Salanova 2011

Salanova.indd 25

31/05/2016 15:58:06

[ 26 ]

i. behind the warriors

]]

decorations using some Donax species and cockles (Fig. 1.7). As the archaeometric analyses did not reveal circulation of pots along the Atlantic coast, the circulation of highly skilled potters is more likely to explain the similarities observed at typological and technical levels. At present, this network involves only the Atlantic region. Indeed, despite the typological proximity of some beakers in other parts of Europe, the associated manufacturing process clearly distinguishes the Atlantic Bell Beakers from those found elsewhere, particularly those classified as belonging to the Maritime or International type in the Netherlands and in central Europe (Fig. 1.8). This Atlantic network does not only involve pottery. Gold metallurgy is another novelty of the 3rd millennium BC, particularly from the Bell Beaker period (Armbruster & Comendador 2015). Specific types of gold ornaments found in Bell Beaker graves are common to the Atlantic regions, from southern Portugal to the north of Scotland. These areas share not only the typology of objects, but also a similar know-how, which could reflect the circulation of qualified craftsmen. The situation is different for copper production. Despite the fact that Bell Beaker copper items are among the first metal objects found in several regions in the north-west of Europe, the graves of metalworkers are quite scarce, which is opposite the situation observed in central Europe, where graves of craftsmen linked to copper production are relatively frequent (Heyd 2007, 360). In the west, such graves have also been recognized in the Netherlands (Lunteren: Lanting & Waals 1976). The Amesbury Archer is probably the most western example of stone metalworking tools deposited in a Bell Beaker grave, with two others being recorded in Portuguese collective graves (Fitzpatrick 2011; Brandherm 2011).   As compositional analyses of copper daggers from France and Spain demonstrate an incredible diversity of the raw materials as well as the techniques used for their production (Figueroa Larre 2005), metalworking reflects another social category in the society: itinerant traders who were in charge of the circulation of these prestigious items.

1.8 Maritime beakers from the Netherlands (1), Czech Republic (2), and Eastern Spain (3). Photos L. Salanova

Salanova.indd 26

31/05/2016 15:58:09

Salanova

1.9 The three AOO graves from northern France, linked to the Grand-Pressigny flint workshops and the Lower Rhine Valley. From Salanova 2002

The traders Three graves from northern France give rise to the question of the existence of traders, which could explain the long-distance exchanges recorded during the 3rd millennium BC (Fig. 1.9). These three burials share some particularities that distinguish them from other Bell Beaker graves. They contain men or women buried in oval or rectangular pits in an extended position, oriented according to an east–west axis, similar to what is observed in the Corded Wares and Single Grave Cultures. Surrounding one of these graves, located near Poitiers at La Folie, a circular ditch was found, in which postholes indicated the existence of a wooden palisade, similar to examples from the Lower Rhine Valley (Tchérémissinoff et al. 2011). In these three burials, the grave goods are invariably composed of All-OverOrnamented (AOO) beakers, with slender profiles and cord- or spatula-impressed decorations, and of blades or daggers composed of yellow flint from the Grand-Pressigny region (Indre-et-Loire, France). The two AOO graves from the Paris Basin (Jablines and Ciry-Salsogne) are dated from 2570–2450 cal BC (Salanova 2011). All of the characteristics of these graves refer to foreign burial practices, from their architecture to their grave goods, which find comparisons in the Netherlands. According to archaeometric analyses, the AOO Beakers were all produced locally, despite their typology indicative of an exogenous affinity. These graves are geographically located on the road that linked the Grand-Pressigny flint workshops to the Lower Rhine Valley, where daggers and blades imported from the GrandPressigny region have been recorded and were frequently included as grave goods associated with AOO Beakers (Lanting & Waals 1976; Delcourt-Vlaeminck 2004). These three graves could therefore reflect an ethnic identity, including foreign traders in charge of another exchange network, linking the Atlantic coast to the Rhine Valley. This network did not remain thereafter; importations stopped at approximately 2400 cal BC, probably being replaced by exchanges of copper daggers.

Salanova.indd 27

31/05/2016 15:58:10

1.10 Distribution of the main regional Bell Beaker pottery styles.

From Salanova 2005, modified

According to the observed burial practices, it is therefore possible to detect several levels of the expression of social identities. Collective and individual graves appear to define the first level, as the functions of and rules for distinguishing these two kinds of burials are significantly different. The collective graves refer to an old tradition and show relative continuity of the social organization before and during the Bell Beaker phenomenon. In contrast, the individual graves mark a break in this continuity, introducing new codes of distinction that are more distinctly based on sexual factors and professional groups. Among the observable categories, some refer to war and power, while others are linked to activities observable in importance during the 3rd millennium BC: long-distance exchange networks and specialized crafts. It must be considered that the most visible type of status, i.e., the most clearly expressed, and the most widely distributed at a supra-regional level should have been the most important for the population. These categories refer to common social identities that are shared at a European scale. They provide less information about the cultural and regional preferences associated with the Bell Beaker phenomenon, which are, however, clearly highlighted in the distribution of typological elements (Besse 2003; Vander Linden 2004). Beyond the visible social categories described previously, which ultimately represent a small portion of the population from the end of the 3rd millennium BC, how was the rest of the population structured?

Salanova.indd 28

31/05/2016 15:58:11

Salanova

[ 29 ]

4. Identity, daily life and cultural norms Studies on the Bell Beaker phenomenon have mainly focused on the funerary sphere, and the analyses conducted in this specific context allow the reconstruction of a clear picture of the most visible social components of the communities from the 3rd millennium BC. Nevertheless, as highlighted previously, the representativeness of these dead individuals compared with the whole society is a question that remains to be answered. Which proportion of the population was buried under such conditions that they can be taken in to account in current studies? To answer this question, data from the Bell Beaker settlements are essential. The poor preservation of these settlements could make this research less fruitful. However, the results demonstrate that domestic contexts highlight another facet of the society, ultimately involving a greater percentage of the living population than the Bell Beaker graves. 4.1. The Bell Beaker settlements Few data are available on settlements from the 3rd millennium BC, not only because of their poor preservation, but also because of their scarcity. This situation is common to all regions of Europe, though the western region most likely provides the most complete examples. Some environmental analyses indicate that this period coincides with a cooling of temperatures in western Europe (Fábregas Valcarce et al. 2003; Magny 2004). In parallel, the increasing human pressure recorded as beginning at the end of the 4th millennium BC appears to have intensified during the Bell Beaker period. This intensification of land appropriation and deforestation, even in high-altitude regions where no evidence of farming activities was previously recorded, characterizes the second half of the 3rd millennium BC, in addition to a definitive change towards a developed farming economy and stratified societies, with the appearance of more permanent settlements in the most northern and eastern parts of Europe (Galop 2005; Leroyer 2004; Castiglioni et al. 2008; Prescott & Glørstad 2011; Czebreszuk & Szmyt 2008).   On a broad European scale, the end of the 3rd millennium BC must be considered to represent the completion of the Neolithization process that began several millennia before (Salanova 2000a). At a regional scale, the economic systems nevertheless appear to be different, possibly explaining the regional variability of the Bell Beaker phenomenon, which declined like a musical standard according to regional traditions. 4.2. Some norms in settlement patterns As the preservation state of the settlements does not allow reconstruction of the details of the way of life of the population, we must work around this problem by analysing the way of life of the population through their pots (Salanova & Prieto Martínez et al. in press). We have combined typology and technology with the analysis of paste components to identify technical groups and to understand the mechanisms underlying the circulation of different Bell Beaker styles, which define coherent regional entities (Fig. 1.10). To which

Salanova.indd 29

31/05/2016 15:58:11

1.9. The extent of the Bell-Beaker Complex 2700–2200 BC

1.11 Bell Beaker settlement from La Noue Saint-Marcel (Burgundy, France). Plan of the site and associated pottery. From Salanova & Ducreux 2005

level of identity do these stylistic differences refer? The synthetic view of the analyses conducted on one thousand pots from nine areas in western Europe highlights the fact that these stylistic differences had real significance for the population, as their distributions are correlated with mobility and settlement patterns.   Some Bell Beaker styles are common to all of Europe, representing the standard Bell Beakers, which refers to the classical style, decorated with hatched bands in alternating opposite directions. At a European scale, this style is predominantly found on the Atlantic seaboard, especially in Brittany and central Portugal. The Atlantic regions are also characterized by a version derived from the standard style, showing the same type of vessels with hatched bands, associated with other designs, such as triangles or a lattice pattern. (‘Epimaritime style’). This group is often contrasted with the Mediterranean type, in which vessels with more diverse shapes are mainly decorated with incisions

Salanova.indd 30

31/05/2016 15:58:13

Salanova

[ 31 ]

and radial structures. Central Europe exhibits another type of Bell Beaker decoration with metope designs. In addition to decorated beakers, the domestic assemblages also provide undecorated beakers and common wares, among which some types refer to local traditions from the Late Neolithic, while others are specific to the Bell Beaker package and are more widely distributed (Besse 2003). Within the framework of our analysis, we have considered all of these categories of pottery associated with settlements (decorated beakers, undecorated beakers and common wares) to compare the circulation mechanisms of the different categories of pottery products. In all of the areas studied, the clays used for manufacturing Bell Beakers were mainly collected around the sites in the local environment, i.e., less than 7 km around the sites, corresponding to the territory the inhabitants of a village could have crossed for daily activities (Arnold 1985; Gosselain & Livingstone Smith 2005). Nevertheless, the combination of the analyses on pottery and soil samples makes it possible to be more precise, revealing a degree of regional circulation (7 km to 30 km around the sites) and, more rarely, extra-regional circulation (further than 30 km). The observed types of pottery circulation point towards very different situations.   In eastern France and western Switzerland, where the best-preserved Bell Beaker settlements have been found, a significant number of imported pots (decorated beakers as well as domestic wares) have been recorded. At La Noue Saint-Marcel, in Burgundy, almost the entire pottery assemblage was manufactured outside of the site (Fig. 1.11). Instead, most of the pottery was likely produced on the other side of the Saône Valley, where the geology is compatible (Salanova & Ducreux 2005). Taking into account the other characteristics of the dwelling, which is located in a wet zone and shows a large number of scrapers in the flint assemblage (possibly for working with skins), this settlement is expected to have been occupied for seasonal and specialized activities. Its inhabitants brought pottery produced in more permanent dwellings from upland areas. This residential system, involving both permanent and seasonal sites, could also be proposed for the French Jura, where none of the caves contained imported vessels, in contrast to the only open-air dwelling, DerrièreLe-Château (Géovreissiat), which contained a high fraction of imported wares (Convertini 2001). This goes against the general idea that the population rarely circulated with pottery during prehistory. Our case study shows that even at the dawn of the Bronze Age, the population travelled regularly, at least on a regional scale, together with their material assets.   Compared with dwellings from the Atlantic coast and the Mediterranean, the analysed cases reveal that Bell Beakers were mainly produced and used locally. The population appears to have been less mobile than in continental regions. Along Atlantic shores (Brittany and Galicia), some pottery products circulated, but only in specific cases, such as among ceremonial sites and in burial contexts (Prieto-Martínez & Salanova 2011).   These results are corroborated by the distribution of burial structures and dwelling types, which appear to outline a well-marked frontier: in the west, collective tombs and isolated domestic stone architecture are found; in the east, rectangular wattle and daub houses grouped in small villages and isolated individual burials with wooden architectures

Salanova.indd 31

31/05/2016 15:58:13

1.12 Settlement patterns. Surrounded zones correspond to the studied areas. In blue: the western pattern with isolated farms and collective graves. In orange: dwellings with rows of houses and individual graves.

are observed (Salanova in press). Based on clay processing and the archaeological data, it is therefore possible to identify two different cultural patterns expressed in major differences in the way of life (Fig. 1.12). In the eastern part of the area, groups of villages are associated with seasonal journeys and isolated graves; in the western part, isolated farmers are associated with collective gatherings for ritual ceremonies.   This residential pattern of isolated houses appears to be characteristic of Atlantic areas, at least in France. It is difficult to precisely date its origin, but it existed from the beginning to the end of the 3rd millennium BC in northern France and along the Atlantic coast from Brittany to Dordogne (Tinevez 2004; Pailler et al. 2010). 4.3. The Atlantic Bell Beaker identity Finally, the results of Bell Beaker deconstruction allow us to regard more precisely the criteria used to define the Bell Beaker identity. Beyond the warrior ideology, which is well established in the burial context and involved a small fraction of the population from the 3rd millennium BC, several other categories can be recognized, even in domestic contexts, where social and cultural norms are generally the most difficult to observe.   The earlier portion of this contribution was mainly focused on recognition of the social categories from the Bell Beaker period compared with previous periods. The examination of domestic contexts has revealed different cultural norms in the way of life, highlighting some regional differences. How were the two sides of the Bell Beaker

Salanova.indd 32

31/05/2016 15:58:17

Salanova

[ 33 ]

phenomenon connected? Beyond the common social trend that affected the 3rd millennium BC in Europe, cultural practices in burial and domestic contexts point to separate histories depending on regions. A common social trend in Europe The warrior ideology, which is easily recognized in Bell Beaker graves and iconography, can ultimately be considered to represent a unique link between the different provinces described for the second half of the 3rd millennium BC in Europe. The different categories of specialists (craftsmen and traders) linked to this ‘warrior’ élite also belong to this common social rank, which makes the Bell Beaker phenomenon a real material entity. This division into clearly identifiable social and professional groups, together with the previously described diversity, is probably the newest aspect that appeared with the Bell Beaker package. This aspect explains why the Bell Beaker phenomenon has often been compared with other phenomena from more recent periods (Lewthwaite 1987; Lemercier 2004). It confers a kind of modernity on the societies of the 3rd millennium BC, in that it allows us to imagine various categories composing a Neolithic society and their hierarchy. This is not the case for the older periods of the Neolithic, for which it is generally possible to detect only a few categories among the dead (see for instance for the Middle Neolithic Chambon & Thomas 2010). The regional expression of these social changes points to a continuity of some cultural norms. Regional expression Overall, a number of differences have been noted among the Bell Beakers that divide the western and eastern parts of Europe. They essentially concern the sexes and age status in graves, which are not expressed in the same way and, above all, the collective burial practice, which refers to different social and kinship structures compared with individual graves.   Within this broad geographic division, the regions along the Atlantic shores share several common characteristics, which define a coherent entity that was already perceptible during the previous periods of the Neolithic and persisted until at least the Late Bronze Age (Brun 1991; Bueno Ramírez et al. 2005; Quilliec 2007).   As observed for the Bronze Age, the Atlantic areas from the 3rd millennium BC are linked by common categories of metal objects and pottery, which form a spatial material group. According to recent technological analyses, this uniformity at typological and technical levels is most likely indicative of the circulation of specialists along Atlantic shores. This Atlantic entity is also defined by a common settlement pattern and a tradition of regional gathering for ritual or burial activities, which took place until the end of the 3rd millennium BC, mainly in collective monuments, extending the long-term sequence of ancestral vaults.   This Atlantic tradition detected from the Neolithic until the Late Bronze Age could certainly be explained by the specific conditions of the environment and, of course, the omnipresence of the ocean, which is expected to have had a strong impact on the construction of this identity. The ocean, with the implied possibility of navigation, is essential for understanding the frequent circulation of ideas, objects and craftsmen along

Salanova.indd 33

31/05/2016 15:58:17

[ 34 ]

i. behind the warriors

]]

the Atlantic shores of Europe (Salanova 2000b; Prieto-Martínez & Salanova 2009). For instance, the large similarities between the Maritime beakers found in southern Brittany, northern Galicia and central Portugal could not be explained without a rapid mode of communication between these three regions, excluding step-by-step circulation (Salanova 2008). The coastal distribution of some Spanish influences in some French Bell Beakers (concerning certain bowl shapes and incised decorations) points to the same phenomenon (Salanova et al. 2011). These vessels have been found only in specific areas, in the southwest part of Brittany, around the Loire estuary and along the Vendée coast, including some islands in the central part of France (île de Ré). This distribution provides some indication of the maritime roads that linked the different regions. Conclusion In conclusion, the social categories of the 3rd millennium and those that are more visible from the Bell Beaker period define several identities, which reflect a society structured into different corporative groups. All of these categories making up the society certainly did not have the same way of considering their identity. Although it could appear to be unreasonably ambitious to reconstruct what these people may have thought, the archaeological data suggest some responses to the wide question of their identity. With the exception of this social trend, which was common to all of Europe, cultural norms, which could define another level of identity corresponding to regional expression, have been taken into account to a lesser extent. Beyond the typological differences in material culture, economic, settlement and mobility patterns should be a more common subject of research, which could indicate another facet of the population and allow reconstruction of the societies with their total diversity. Along the Atlantic façade of Europe, a coherent Bell Beaker entity has been recognized, defining a regional identity that prefigures the Atlantic Bronze Age Complex.   Some scholars have attempted to link the Bell Beaker package with the emergence of Proto-Indo-European or Proto-Celtic languages (for instance: Gallay 2001; Brun 2006). Due to the differences between western and central Europe and the diversity of the cultural and social categories reflected in the Bell Beaker phenomenon, it is difficult to demonstrate the existence of a common language based only on the material culture. However, the similarities described along the Atlantic coast, which involve the circulation of know-how and of craftsmen, could effectively demonstrate the existence of supra-regional languages, shared at least by this group of specialists.

Salanova.indd 34

31/05/2016 15:58:17

Salanova

[ 35 ]

BIBLIOGRAPHY Armbruster, B. & B. Comendador 2015 ‘Early gold technology as an indicator of circulation processes in Atlantic Europe’, The Bell Beaker Transition in Europe. Mobility and local evolution during the 3rd millennium BC, ed. P. PrietoMartínez & L. Salanova, 140–9. Oxford, Oxbow Books. Arnold, D. E. 1985 Ceramic theory and cultural process. Cambridge, Cambridge University Press. Besse, M. 2003 L’Europe du 3e millénaire avant notre ère : les céramiques communes au Campaniforme. Lausanne, Cahiers d’archéologie romande 94. Blin, A. 2012 ‘Une nouvelle analyse de l’hypogée néolithique des Mournouards II au Mesnil-surOger (Marne)’, Revue Archéologique de l’Est 61, 35–54. Bocksberger, O.-J. 1976 Le site préhistorique du Petit-Chasseur (Sion, Valais). 1 – Le dolmen MVI. Lausanne, Cahiers d’archéologie romande 6. Brandherm, D. 2011 ‘Frühe “Metallurgengräber” von der Iberischen Halbinsel. Aussagemöglichkeiten zum sozialen Kontext’, Siedlung und Handwerk. Studien zu sozialen Kontexten in der Bronzezeit. Beiträge zu den Sitzungen der Arbeitsgemeinschaft Bronzezeit auf der Jahrestagung des Nordwestdeutschen Verbandes für Altertumsforschung in Schleswig 2007 und auf dem Deutschen Archäologenkongress in Mannheim 2008, eds. B. Horejs & T. Kienlin, 317–30. Bonn. Brodie, N. 2001 ‘Technological frontiers and the emergence of the Beaker culture’, Bell Beakers Today, Pottery, people, culture, symbols in prehistoric Europe, ed. F. Nicolis, 487–96. Trento, Proceedings of the International Colloquium Riva del Garda (Trento, Italy), 11–16 may 1998. Brun, P. 1991 ‘Le Bronze Atlantique et ses subdivisions culturelles : essai de definition’, L’âge du Bronze Atlantique, ses faciès, de l’Ecosse à l’Andalousie et leurs relations avec le Bronze Continental et la Méditerranée, eds. C. Chevillot & A. Coffyn, 11–24. Beynac, Actes du 1er Colloque du Parc de Beynac. Brun, P. 2006 ‘L’origine des Celtes. Communautés linguistiques et réseaux sociaux’, Celtes et Gaulois, l’Archéologie face à l’Histoire. 2 La Préhistoire des Celtes, ed. D. Vitali, 29–44. Glux-en-Glenne, Bibracte. Bueno Ramírez, P., R. de Balbín Behrmann &

Salanova.indd 35

R. Barroso Bermejo 2005 ‘Hiérarchisation et métallurgie : statues armées dans la Péninsule Ibérique’, L’anthropologie 19, 577–640. Case, H. 1995 ‘Beakers : loosening a stereotype’, Unbaked Urns of Rudely Shape : essays on British and Irish Pottery for Ian Longworth, 55–67. Oxford, Oxbow Monograph 55. Castiglioni, E., M. Cottini & M. Rottoli 2008 ‘Aspects of Bell Beaker and Chalcolithic economy in Northern Italy’, Bell Beaker in everyday life, eds. M. Baioni, V. Leonini, D. Lo Vetro, F. Martini, R. Poggiani Keller & L. Sarti, 147–50. Firenze, Museo Fiorentino di Preistoria Paolo Graziosi, Millenni Studi Archaeologia Preistorica 6. Chambon, P. 2003 Les morts dans les sépultures collectives néolithiques en France : du cadavre aux restes ultimes. Paris, XXVe supplément à Gallia Préhistoire. Chambon, P. & A. Thomas 2010 ‘The first monumental cemeteries of Western Europe: the ‘Passy type’ necropolis in the Paris basin around 4500 BC’, Megaliths and Identities. Frühe Monumentalität und Soziale Differenzierung 1, eds. F. Furholt, F. Lüth & J. Müller, 249–60. Bonn, Dr Rudolf Habelt GmbH. Clarke, D. L. 1976 ‘The beaker network. Social and economic models’, Glockenbecher Symposion (Oberried), 460–77. Convertini, F. 2001 ‘Production de la céramique campaniforme dans l’axe Rhin-Rhône’, Bell Beakers Today, Pottery, people, culture, symbols in prehistoric Europe, ed. F. Nicolis, 547–60. Trento, Proceedings of the International Colloquium Riva del Garda (Trento, Italy), 11–16 may 1998. Costin, C. L. 1991 ‘Craft Specialization: Issues in Defining, Documenting, and Explaining the Organization of Production’, Archaeological Method and Theory, vol. 1, ed. M. B. Schiffer, 1–56. Tucson, University of Arizona Press. Czebreszuk, J. & M. Szmyt 2008 ‘Bell Beakers and their role in a settlement evolution during the Neolithic–Bronze interstage on the Polish Lowland’, Bell Beaker in everyday life, eds. M. Baioni, V. Leonini, D. Lo Vetro, F. Martini, R. Poggiani Keller & L. Sarti, 221–33. Firenze, Museo Fiorentino di Preistoria Paolo Graziosi,

31/05/2016 15:58:17

[ 36 ]

i. behind the warriors

Millenni Studi Archaeologia Preistorica 6. Delcourt-Vlaeminck, M. 2004
‘Les exportations du silex du Grand Pressigny et du matériau tertiaire dans le nord-ouest de l’Europe au Néolithique final/Chalcolithique’, Anthropologica et Praehistorica 115, 139–54. Desideri, J. & S. Eades 2004 ‘Le peuplement campaniforme en Suisse. Nouveaux apports de la morphologie crânienne et dentaire’, Graves and Funerary Rituals during the Late Neolithic and the Early Bronze Age in Europe (2700–2000 BC), eds. M. Besse & J. Desideri, 99–109. Oxford, British Archaeological Report, International Series 1284. Desideri, J. 2011 When Beakers met Bell Beakers. An analysis of dental remains. Oxford, British Archaeological Report, International Series 2292. Díaz-Andreu, M. & S. Lucy 2005 ‘Introduction’, The Archaeology of Identity. Approaches to gender, age, status, ethnicity and religion, eds. M. Díaz-Andreu, S. Lucy, S. Babić ans D.N. Edwards, 1–12. New York, Routledge. Fábregas Valcarce, R., A. Martínez Cortizas, R. Blanco Chao & W. Chesworth 2003 ‘Environmental change and social dynamics in the second-third millennium BC in NW Iberia’, Journal of Archaeological Science 30, 859–71. Figueroa Larre, V. 2005 Métallurgie en contexte campaniforme : nord de la péninsule Ibérique et en France, Unpublished Master thesis, Université de Paris I. Fitzpatrick, A. P. 2011 The Amesbury Archer and the Boscombe Bowmen. Bell Beaker burials at Boscombe Down, Amesbury, Wiltshire, Great Britain. Excavations at Boscombe Down volume 1. Salisbury, Wessex Archaeology Report 27. Fowler, C. 2010 ‘From identity and material culture to personhood and materiality’, The Oxford Handbook of Material Culture studies, eds. D. Hicks & M. C. Beaudry, 352–85. Oxford, Oxford University Press. Gallay A. 2011 ‘Postface. Quelques axes de recherches possibles à partir des observations effectuées’, Le site préhistorique du Petit-Chasseur (Sion, Valais). 9 – Dolmens MXII et MXIII. Approche des différents niveaux préhistoriques, eds. S. Favre & M. Mothet, 234–37. Lausanne, Cahiers d’archéologie romande 123. Gallay, A. 1995 ‘Les stèles anthropomorphes du site mégalithique du Petit-Chasseur à Sion’, Notizie Archeologiche Bergomensi 3, 167–94.

Salanova.indd 36

]]

Gallay, A. 2001 ‘L’énigme campaniforme’, Bell Beakers Today. Pottery, people, culture, symbols in prehistoric Europe, Vol. 1, ed. F. Nicolis, 41–7. Trento, Proceedings of the International Colloquium Riva del Garda 1998. Galop, D. 2005 ‘La conquête de la montagne Pyrénéenne au Néolithique. Chronologie, rythmes et transformations des paysages à partir des données polliniques’. Populations néolithiques et envrionnements, Séminaires du Collège de France, ed. J. Guilaine, 279–95. Paris, Errance. Gonçalves, V. S. 1999 Reguengos de Monsaraz, territories megalíticos. Lisboa, Museu Nacional de Arqueologia. Gosselain, O. P. & A. Livingstone Smith 2005 ‘The source. Clay selection and processing practices in sub-Saharian Africa’, Pottery manufacturing processes: reconstruction and interpretation, eds. A. Livingstone Smith, D. Bosquet, & R. Martineau, 33–48. Oxford, British Archaeological Report, International Series 1349. Heyd V. 2007 ‘Families, Prestige Goods, Warriors and Complex Societies: Beaker Groups of the 3rd Millennium cal BC along the Upper and Middle Danube’, Proceedings of the Prehistoric Society 73, 321–70. Hodder, I. 1982 Symbols in Action. Cambridge, Cambridge University Press. Hofman, D. & P. Bickle 2011 ‘Culture, Tradition and the Settlement Burials of the Linearbandkeramik (LBK) Culture’, Investigating Archaeological Cultures. Material Culture, Variability, and Transmission, eds. B. W. Roberts & M. Vander Linden, 183–200. New York, Springer. Kristiansen, K. 2011 ‘Constructing Social and Cultural Identities in the Bronze Age’, Investigating Archaeological Cultures. Material Culture, Variability, and Transmission, eds. B. W. Roberts & M. Vander Linden, 201–10. New York, Springer. Lanting, J. N. & J. D. van der Waals 1976 ‘Beaker Culture relations in the Lower Rhine Basin’, Glockenbecher Symposion (Oberried, 1974), 1–80. Le Goff, J. 1989 L’homme médiéval. Paris, Seuil, collection Points Histoire. Leclerc, J. & C. Masset 2006 ‘L’évolution de la pratique funéraire dans la sépulture collective de la Chaussée-Tirancourt (Somme)’, Bulletin de la Société Préhistorique Française 103–1, 87–116. Ledermann, S. 1969 Nouvelles tables-types de mortalité. Paris, INED, Travaux et documents cahier 23.

31/05/2016 15:58:17

Salanova

Leisner G. & V. Leisner 1959 Die Megalithgräber der Iberischen Halbinsel. Der Westen. Madrider Forschungen Bd. 1, 2. Berlin, Walter de Gruyter & Co. Lemercier, O. 2004 Les Campaniformes dans le sud-est de la France. Lattes, Monographies d’Archéologie Méditerrannéenne 18. Leroyer, C. 2004 ‘Environnement végétal des structures funéraires et anthropisation du milieu durant le Néolithique récent/final dans le Bassin parisien’, Revue Archeologique de Picardie, n° spécial 21, 83–92. Lewthwaite, J. G. 1987 ‘The braudelian Beakers : a chalcolithic conjoncture in Western Mediterranean Prehistory’, Bell Beakers of the western Mediterranean, eds. W. H. Waldren & R. C. Kennard, 31–60. Oxford, British Archaeological Report 331. Magny, M. 2004 ‘Holocene climatic variability as reflected by mid-European lake-level fluctuations, and its probable impact on prehistoric human settlements’, Quaternary International 113, 65–80. Mariéthoz, F. 2011 ‘La chambre sépulcrale du dolmen MXII: une approche des pratiques funéraires et de la population inhumée’, Le site préhistorique du Petit-Chasseur (Sion, Valais). 9 – Dolmens MXII et MXIII. Approche des différents niveaux préhistoriques, ed. S. Favre & M. Mothet, 169–90. Lausanne, Cahiers d’archéologie romande 123. Matéjíčková, A. & P. Dvořák 2012 Funerary areas of the Bell Beaker period on the D1 Vyškov-Mořice Motorway. Brno, Pravěk, Supplementum 24. Meadows, J., A. Barclay, & A. Bayliss 2007 ‘A short passage of time: the dating of the Hazleton long cairn revisited’, Cambridge Archaeological Journal 17–1, 45–64. Mezzena, F. 1997 ‘La valle d’Aosta nel Neolitico e nell’Eneolitico’, La valle d’Aosta nel quadro della Preistoria e Protostoria dell’Arco alpino centrooccidentale, Atti della XXXI Riunione scientifica, 17– 138. Firenze, Istituto di Preistoria e Protostoria. Moinat, P. & J. Studer J. 2007 ‘Cistes en pierre et coffres en bois, inhumations simples et dépôts complexes : un bilan des pratiques funéraires à Vidy (Lausanne, Vaud) et à Chamblandes (Pully, Vaud)’, Les cistes Chamblandes et la place des coffres dans les pratiques funéraires du Néolithique moyen occidental, Actes du colloque de Lausanne (12–13 mai 2006), eds. P. Moinat & P. Chambon, 195–220. Lausanne/Paris, Cahiers d’archéologie

Salanova.indd 37

[ 37 ]

romande & Mémoire XLIII de la Société Préhistorique française. Pailler, Y., H. Gandois, E. Ihuel, C. Nicolas & Y. Sparfel 2010 ‘Le bâtiment en pierres sèches de Beg ar Loued, Ile Molène (Finistère) : évolution d’une construction du Campaniforme au Bronze ancien’, Les premiers néolithiques de l’Ouest, Actes du colloque interrégional sur le Néolithique (Le Havre, 2007), 425–49. Rennes, supplément Revue Archéologique de l’Ouest, Presses Universitaires de Rennes. Prescott, C. & H. Glørstad 2011 Becoming European. The transformation of third millennium Europe and the trajectory into the second millennium BC, 1–11. Oxford, Oxbow Books. Price, T. D., C. Knipper, G. Grupe & V. Smrcka 2004 ‘Strontium isotopes and prehistoric human migration: the Bell Beaker period in central Europe’, European Journal of Archaeology 7, 9–40. Price, T. D., G. Grupe & P. Schröter 1998 ‘Migration in the Bell Beaker period of Central Europe’, Antiquity 72–276, 405–11. Prieto Martínez, M. P. & L. Salanova 2009 ‘Coquilles et Campaniforme en Galice et en Bretagne: mécanismes de circulation et stratégies identitaires’, Bulletin de la Société Préhistorique Française 106, 73–93. Prieto Martínez, P. & L. Salanova 2011 Las Comunidades Campaniformes en Galicia. Cambios sociales en el III y II milenios BC en el NW de la Península Ibérica. Pontevedra, Diputación de Pontevedra. Quilliec, B. 2007 L’épée atlantique : échanges et prestige au Bronze final. Paris, Mémoire XLII de la Société Préhistorique Française. Roberts, B.W. & M. Vander Linden 2011 Investigating Archaeological Cultures: material culture, variability, and transmission. New York, Springer. Rodríguez Casal, A.A. 1988 La necropolis megalítica de Parxubeira (San Fins de Eirón, Galicia). Campañas arqueológicas de 1977 a 1984. A Coruña, Monografías Urxentes do Museu, Museoo Arqueologixo e Historico Provincial « Castelo de San Anton ». Salanova, L. 1998 ‘Le statut des assemblages campaniformes en contexte funéraire : la notion de “bien de prestige”’, Bulletin de la Société Préhistorique Française 95–3, 315–26. Salanova, L. 2000a La question du Campaniforme en France et dans les Iles anglo-normandes : productions,

31/05/2016 15:58:17

[ 38 ]

i. behind the warriors

chronologie et rôles d’un standard céramique. Paris, Société Préhistorique Française and Comité des Travaux Historiques et Scientifiques. Salanova, L. 2000b ‘Mécanismes de diffusion des vases campaniformes : les liens franco-portugais’, 3e Congresso de Arqueologia Peninsular (Vila Real, Portugal, setembro de 1999), vol. 4: Pré-História recente da Península ibérica, ed. V. O. Jorge, 399–409. Porto, ADECAP. Salanova, L. 2002 ‘Fabrication et circulation des céramiques campaniformes’, Matériaux, productions, circulations du Néolithique à l’Age du Bronze, séminaires du Collège de France, ed. J. Guilaine 151– 66. Paris, Errance, coll. Des Hespérides. Salanova, L. 2003 ‘Heads North: analysis of Bell Beaker graves in Western Europe’, Journal of Iberian Archaeology, 5, 163–9. Salanova, L. 2005 ‘Los orígenes del Campaniforme : descomponer, analizar, cartografiar / The origines of the Bell Beakers phenomenon : breakdown, analysis, mapping’, El Campaniforme en la Península Ibérica y su contexto europeo, eds. M. A. Rojo-Guerra, R. Garrido-Pena & I. GarcíaMartínez de Lagrán 7–27. Valladolid, Universidad de Valladolid & Junta de Castilla y León. Salanova, L. 2007 ‘Les sépultures campaniformes : lecture sociale’, Le Chalcolithique et la construction des inégalités, t. I : Le continent européen, séminaires du Collège de France, ed. J. Guilaine, 213–28. Paris, éd. Errance, coll. Des Hespérides. Salanova, L. 2008 ‘Le temps d’une diffusion : la céramique campaniforme en Europe’, Construire le temps. Histoire et méthodes des chronologies et calendriers des derniers millénaires avant notre ère en Europe occidentale, ed. A. Lehöerff, 135–49. Glux-enGlenne, Bibracte. Salanova, L. 2011 ‘Chronologie et facteurs d’évolution des sépultures individuelles campaniformes dans le Nord de la France’, Les sépultures individuelles campaniformes en France, eds. L. Salanova & Y. Tchérémissinoff, 125–42. Paris, XLIe sup. à Gallia Préhistoire. Salanova, L. 2012 ‘Productions domestiques, productions spécialisées… et le reste ? Les différents types de productions céramiques néolithiques’, Bulletin de la Société Préhistorique Française 109–200, 221–9. Salanova, L. In press The Bell Beaker Complex in France, Prehistoric France, eds. L. McFadyen & C. Marcigny. Cambridge, Cambridge University

Salanova.indd 38

]]

Press. Salanova, L. & F. Ducreux 2005 ‘L’habitat campaniforme de La Noue à Saint-Marcel (Saône-et-Loire). Eléments de définition du groupe bourguignon-jurassien’, Gallia Préhistoire 47, 33–146. Salanova, L. & M. Sohn 2007 ‘Mobilier funéraire et modes de différenciation des morts à la fin du Néolithique en Europe occidentale’, Pratiques funéraires et Sociétés. Nouvelles approches en archéologie et en anthropologie sociale, eds. L. Baray, P. Brun & A. Testart, 77–90. Dijon, Editions Universitaires de Dijon, coll. Art, Archéologie et Patrimoine. Salanova L. & Y. Tchérémissinoff 2011 Les sépultures individuelles campaniformes en France, XLIe supplément à Gallia Préhistoire, CNRS éditions Salanova, L., A. Cormenier & V. Ard 2011 ‘Un nouveau témoin campaniforme dans le CentreOuest’, Puyraveau à Saint-Léger-de-Montbrun (Deux-Sèvres). Le dolmen II. Un monument au mobilier exceptionnel de la fin du Néolithique dans le Centre-Ouest de la France, ed. V. Ard, 491–512. Chauvigny, Mémoire XLI de l’Association des Publications Chauvinoises. Salanova, L., C. Renard & B. Mille 2011 ‘Réexamen du mobilier de la sépulture campaniforme d’Arenberg (Wallers, Nord)’, Les sépultures individuelles campaniformes en France, eds. L. Salanova & Y. Tchérémissinoff, 79–9. Paris, XLIe sup. à Gallia Préhistoire. Salanova, L., P. Chambon, J.-G. Pariat, A.-S. Marçais, F. Valentin, F. Healy, A. Whittle & A. Bayliss in press ‘Cultural evolution and burial discontinuities in collective graves: Bury (Northern France, 4th–3rd mil. cal BC’, Current Approaches to Collective Burials in the Late European Prehistory. Burgos, UISPP Congress. Salanova L., P. Prieto Martínez, X. Clop García, F. Convertini, O. Lantes-Suárez & MartínezCortizas in press ‘What are large-scale archaeometric programmes for? Bell Beaker pottery and societies from the 3rd millennium BC in Western Europe’, Archaeometry. Strahm, C. 1995 Das Glockenbecher-Phänomen. Ein Seminar. Freiburg, Freiburger Archäologische Studien 2, Institut für Ur- und Frühgeschichte der Universität Freiburg I. Br. Tchérémissinoff, Y., F. Convertini, P. Fouéré & L. Salanova 2011 ‘La sépulture campaniforme de la Folie, Poitiers (Vienne)’, Les sépultures individuelles

31/05/2016 15:58:17

Salanova

campaniformes en France, eds. L. Salanova & Y. Tchérémissinoff, 11–19. Paris, XLIe supplément à Gallia Préhistoire. Tinevez, J.-Y. 2004 Le site de la Hersonnais à Pléchâtel (Ille-et-Vilaine) : un ensemble de bâtiments collectifs du Néolithique final. Paris, Travaux 5 de la Société Préhistorique Française. Turek, J. 2000 ‘Děti v Období Sňůrové Keramiky a Zvoncovitých Pohárů: Postavení Dětí ve Společnosti Pozdního Eneolitu [Being a Beaker Child: The Position of Children in Late Eneolithic society]’, In Memoriam Jan Rulf, 424–438. Prague, Památky Archeologické – Supplementum 13. Turek, J. 2011 ‘Age and Gender Identities and Social Differentiations in the Central European Copper Age’, Identity Crisis: Archaeological Perspectives on

Salanova.indd 39

[ 39 ]

Social Identity, Proceedings of the 42nd (2010) Annual Chacmool Archaeology Conference, eds. L. AmundsenMeyer, N. Engel & S. Pickering, 49–61. Calgary, Chacmool Archaeological Association, University of Calgary. Vander Linden, M. 2004 ‘Polythetic networks, coherent people: a new historical hypothesis for the Bell Beaker phenomenon’, Similar but Different. Bell Beakers in Europe, ed. J. Czebreszuk, 35–62. Poznań, Adam Mickiewicz University. Von Tobel, C. 2011 ‘Les sept sépultures du Néolithique moyen au Petit-Chasseur à Sion’, Le site préhistorique du Petit-Chasseur (Sion, Valais). 10 – Un hameau du Néolithique moyen, eds. M. Besse & M. Piguet, 207–15. Lausanne, Cahiers d’archéologie romande.

31/05/2016 15:58:17

[ 40 ]

ii. lost cultures of the halberd bearers

2

1

]]

3

4 6 7

10 5

9

8

11

12

13

14

15

16

2.1 A range of typical early Beaker material from graves in Atlantic zones of Europe: 1-3 early Beakers; 4–5 bone dress fittings; 6 small beads; 7 V-bored button; 8–9 small gold ornaments; 10–11 tanged copper knife and dagger; 12 antler (or bone) rod; 13–14 stone wristguards; 15 copper Palmela point; 16 flint arrowhead. Varied scales; see Appendix 2 for details of examples shown.

Needham CW3 FINAL 21 iv 16.indd 40

27/05/2016 12:45:16

chapter two

The lost cultures of the halberd bearers: a non-Beaker ideology in later third millennium Atlantic Europe

Stuart Needham

T

is essentially a study of comparative distributions or, rather, maps of recovery with all the attendant dangers of differential patterns of survival and recovery. These problems are especially acute in a situation, such as here, where evidence of quite different character is being compared—on the one hand a specific type of metal object potentially largely associated with a singular context type, on the other a suite of material that goes together as a cultural package and comes from a more rounded mix of context types including occupation sites (Figure 2.1). Nevertheless, crucial questions sometimes have to be addressed by making incongruous comparisons even if this limits the degree of certainty over any resulting interpretation.   The two unlike categories of my study are halberds and the Beaker culture (or Bell-Beaker culture in Continental parlance), both recurrent across Atlantic Europe in their various regional manifestations. Dirk Brandherm has previously made some observations on the relationship between these two categories in Iberia and Britain noting that although there is spatial overlap the association between them is poor (2004, 325–6). The current paper builds on these initial observations by looking at all the Atlantic regions in more detail. It also makes use of a revised chronology for Atlantic halberds (Needham 2015a; see further below & Appendix 1). Atlantic halberds have stood in an uncomfortable position in relation to relevant regional chronologies and thus to broader cultural phenomena, owing to the paucity of associations with other diagnostic types. They have frequently been taken to belong to the Early Bronze Age (beginning c. 2200/2150 BC) rather than the Chalcolithic and only in Ireland and Britain has there been a longer tradition of accepting them as primarily a feature of the copper-using phase preceding that of bronze. Relative down-dating has perhaps in part been encouraged by the sophistication of the manufacturing techniques required to make halberds. Moreover, their genuinely later dating in two regional hotspots—Aunjetitz and Argar—where halberds occur in a number of Early Bronze Age graves and hoards undoubtedly gave rise to the assumption that less well contexted halberds elsewhere should be broadly contemporary.   This was never a satisfactory assumption for Britain and Ireland (see, for example, Burgess 1979, Needham 1996) even given how little dating evidence was available and recent radiocarbon dates for the haft remains of two insular halberds have recently given independent support to the early dating (Bell 2014; Needham 2015a). But what of the halberds made of copper in Continental Atlantic zones which have traditionally been dated later? his

Needham CW3 FINAL 21 iv 16.indd 41

27/05/2016 12:45:16

[ 42 ]

ii. lost cultures of the halberd bearers

]]

3

1 2

2.2 Representative examples of the ‘Atlantic’ family of halberds, scale 20%; for details on halberds represented see Appendix 2; dotted rivet holes are interpreted as belonging to a secondary phase of hafting.

20 cm 8 4

5 9

7

6

a) 1–4, three-rivet halberds with deep haftingplates; 5–20, Continental three-rivet series— more triangular butts to left, arched butts to right

15

10 16

b) Irish three-rivet halberds: 21–4, Type Roscrea; 25–7, Type Lough Gur; 28, Type Lough Gur narrow-midrib variant; 29, Type Lough Gur with beaded midrib; 30–32, Type Hill of Allen; 33, Hillswood series ungrouped, with trilobate butt; 34–5, Type Cloonymorris; 36–8, Type Corlurgan

11 17

12 18

13 19

14

c) British, French and Iberian three-rivet halberd types: 39–40, Type Crossraguel; 41, Type Pontrhydygroes; 42–43, Type Lambeth; 44–7, Type Baútas; 48, unique squat halberd, cf Type Baútas; 49–51, Iberian halberds with stepped butts; 52–5, Type Carrapatas

20

a

d) Irish, British and French four-rivet halberds types with (sub-)trapezoidal hafting-plates: 56–8, Type Clonard; 59–60, Type Ballygawley; 61–3, Type Derrinsallagh; 64–71, Type Pistill Dewy/Etigny; 72, cf Type Pistill Dewy/ Etigny; 73–4, Type Glomel

b 22 24

21

e) Other British four-rivet halberds and late Irish and British types: 75–8, Type Sluie; 79, Type Sluie (1st phase), Type Auchingoul (2nd phase); 80, cf Type Sluie; 81–3, Type Auchingoul; 84–5, Type Breaghwy; 86–7, proto-Breaghwy halberds, 88–9, Breaghwy series ungrouped; 90, Type Eweford; 91–2, Type Corrib

23 25

29

26

28

27

30 34

31 35

32 36

37 33

38

20 cm

Needham CW3 FINAL 21 iv 16.indd 42

27/05/2016 12:45:17

[ 43 ]

Needham

d 59

57

56

58 60

62

61 63

64

70

c 39

71

65

41

66 67

40

42

72 68 48

69

73

43 47 44

74

20 cm

46

45

52

e

49

75

53

50

77

76 78

54

51

79

80

55

20 cm 81 83

82 88 84 89

90 85

91

92 86

87

20 cm

Needham CW3 FINAL 21 iv 16.indd 43

27/05/2016 12:45:18

[ 44 ]

ii. lost cultures of the halberd bearers

]]

  Standing back to look at the broader continental scale, various authors have picked up on Lawrence Barfield’s suggestion that some Italian halberds might be the earliest in Europe (Barfield 1969; Burgess 1979, 213; Brandherm 2004, 323–6; 2011, 23), and those in Remedello/ Rinaldino contexts have been suggested to date to the later 4th millennium BC (Horn 2014, 84–7; cf Needham 2015a). This may need to be revised downwards. The only radiocarbon dated context with an early halberd is the grave from Casanuova di San Biagio della Valle, where a human bone gave 4396 + 60 BP (LTL-1783A; Manfredini et al. 2009, 162 fig 15). Calibration suggests a most probable date either side of 3000 cal BC. Meanwhile, Bayesian analysis of dated contexts with Type Remedello and Spilamberto daggers places them in the earlier 3rd millennium BC, rather than earlier (Valzolgher 2014). Meanwhile a group of blades from the Carpathian Basin which have been regarded as halberds are dated even earlier, to the mid-4th millennium BC (Horn 2014, 59) on the strength of a single association in a hoard from Velehrad-Rákoš (ibid, no. 2.7). In fact there are only two close parallels for this associated example (ibid, nos 1.2, 1.4) and it might be questioned whether they have to be interpreted as halberds rather than broad daggers. Nevertheless, if the concept and technology of metalheaded halberds existed in at least one region by about 3000 BC, then it certainly opens up the possibility of transfer to other regions during the earlier 3rd millennium. This brushes aside any a priori difficulties with a pre-Early Bronze Age date for the ‘Atlantic’ styles of halberd in Continental Europe (Figure 2.2; see definitions in Appendix 1), as was hinted at long ago by Burgess (1979, 213). The clear existence of a style continuum and the evidence of occasional associations argue for a good degree of contemporaneity with the span of insular halberds, even if some of these continental cousins may continue a little later. This conclusion places the Atlantic family of halberds very much within the time-frame of the European Bell-Beaker culture (Figure 2.3), and at a general level they share much the same geographical canvas.   The invention of the metal-headed halberd at an early date further east in Europe may resolve one issue, but it creates its own problem—that of the thin representation of the type over a now protracted sequence (Figure 2.3). How this sparse representation may be understood is dealt with in more detail elsewhere (Needham 2015a), where a model has been developed which distinguishes between the circulation of halberds as an ‘undercurrent’, leaving only intermittent evidence in the archaeological record, and the vigorous and non-synchronous ‘uptake’ of halberds in certain regions. It is noteworthy for example, that even in northern Italy, where halberds start early, there are only a modest number of finds throughout the twelve centuries or more of use. The model developed also accepts that we may not find specific initiating imports in any given region. Since it now seems likely that the halberd idea was abroad for a long period of time before its vigorous and episodic uptake in some regions, it would not necessarily be represented initially by any rigidly defined types but instead by a basic concept of how a halberd should be distinguished from any other object class, especially daggers, axes and battle-axes. In such a model, early halberds may be thinly spread, not necessarily all of the same material (see below) and, furthermore, difficult to recognize collectively as a ‘type’. It is suggested that the best candidates for an ‘undercurrent’ of early metal-headed halberds in north-west Europe lies amongst a somewhat eclectic group of rather small copper blades (Needham 2015a; 2015b). Although not similar enough to one another to define a type, there are sufficient elements of comparability from one region to the next to admit the possibility of a thread of spatio-temporal continuity (Figure 2.4).

Needham CW3 FINAL 21 iv 16.indd 44

27/05/2016 12:45:18

Carpathians

3000 BC

?

Italy

France

Iberia

?

?

Ireland

Britain

Northern Europe

?

Remedello halberd horizon

2500 Atlantic

horizon

Chalcolithic

2000 Argaric/Aunjetitz horizon

1500

2.3 Suggested chronology of European metal-headed halberds

circa 3000 BC

Italy

France

2.4 The possible thin typological thread connecting early halberds in north-western Europe

Britain

France & Germany

Ireland

circa 2300 BC

Needham CW3 FINAL 21 iv 16.indd 45

27/05/2016 12:45:18

[ 46 ]

ii. lost cultures of the halberd bearers

]]

  The more archaeologically visible regional phenomena, namely the consistent presence of metal-headed halberds in definable styles, represent the stage at which particular societies and their metalworkers incorporated halberds in the core of their ideological fabric. In this model, vigorous uptake in a region is not so much the result of transplantation, more that of a ‘spark’ within the undercurrent of the halberd idea; that spark represents the point at which the potential of the weapon type to serve necessary social goals was realized. Those social goals may largely have related to establishing cultural identities. This would not be a decision to be made lightly since in many of the regions concerned the particular metalworking and hafting skills would need to be attracted in from outside.   Something should be said briefly about halberds in other materials. Certain flint and stone blades, mainly with notched butts, have been ventured from time to time to have been hafted halberd fashion. While it is possible that in some parts of Europe halberds of lithic materials played a part in the infusion of the halberd idea within a given cultural group (Ó Ríordáin 1937; Brandherm 2004), this is not in itself grounds for seeing a wholly indigenous development into metal-headed forms in those areas. Every society that used antlers as ‘picks’ had a potential halberd prototype, as was recognized by Coffey (1908–9, 109), but the production and mounting of metal-headed halberds could not be achieved with commonly held skills. Most metalworkers of the copper-working period in Atlantic Europe only produced flat-cast forms in univalve moulds, whereas halberds required skilful bivalve casting, probably in clay/sand moulds. In some cases neat furrowed mouldings were formed, probably cast into the blade surface and subsequently finished by grinding and polishing. Hafting would also have been quite an art, involving getting a close fit between the thin hafting-plates and the split wooden haft, creating co-aligned rivet holes through the sandwich, being able to judge sizes for the rivet blanks, and then having the knowhow to clench them. A clenched rivet needed to be tight, but not so tight as to split the wood, to be compressed without over-stressing the metal, and to end up leaving respectable heads for visual effect. Once this combination of skills had come together, they could be transmitted as a knowledge package from one halberd producing group to another; independent invention a second time over seems much less likely.   Comparison between the distributions of halberds and Beakers will proceed on a regional basis, since not only might there be different spatial patterns with their own implications, but there is also the issue of the relevant classes of evidence having variant chronologies (see further in Appendix 1). Britain Britain is taken as the starting point simply because it was here that some kind of ‘opposition’ between halberd-bearing groups and Beaker-affiliated groups became evident to the writer. The chronologies of the two phenomena are not identical (Appendix 1) and good geographical complementarity only emerges clearly when early Beaker material is isolated from the later. Beaker settlement data has not been comprehensively studied in recent years so it is uncertain how it should be phased relative to the better studied grave

Needham CW3 FINAL 21 iv 16.indd 46

27/05/2016 12:45:18

Needham

[ 47 ]

sequence and it is thus excluded here. Nevertheless, the funerary evidence is abundant and will hopefully be broadly representative of the spatial extent of Beaker culture in its earlier phase, down to about 2100 cal BC.   Various approaches can be taken to the inter-comparison. A minimalist approach is to compare just copper halberds with what appears to be an equivalent status ‘weapon’ in the Beaker zone, the tanged copper dagger or knife (Figure 2.5). This produces a strong geographical separation between west and east, although a few finds of either category stray into the ‘opposing’ zone.   A more comprehensive approach takes in all the halberds, a minority of which fall very early in the Early Bronze Age. Contemporary graves can then be defined as those datable to the Chalcolithic and the transition to the Early Bronze Age, based on evidence of either associated artefacts or high-integrity radiocarbon dates, or both (Appendix 1). It is still unclear how long the copper-to-bronze transition lasted in historical terms, not least because of the nature of the wiggles in the calibration curve during the critical period. A cut-off of 3740 BP is taken here as providing a useful statistical separation between calibrated dates before about 2100 cal BC and those that may be later. Despite some undoubted fuzziness at this terminus, this nevertheless provides better definition

2.5 Maps of recovery for a) copper halberds, and b) copper daggers and knives from Britain and Ireland, c. 2450–2150 BC. The north-south line largely separates halberds from daggers and knives in funerary contexts.

Copper halberds

Copper halberds

Halberd find Halberd find Uncertain provenance Uncertain provenance or in later context or in later context In hoard In hoard

2.8. Interpreted territories for halberdTanged Tanged bearing and Beaker groups during the daggers & knives daggers & knives Chalcolithic to earliest Bronze Age in

InIngrave probably) grave (or (orBritain. probably) The only individual symbols shown Singlefind find or Single orunknown unknown hoard (or (orrepresent possibly) InInhoard possibly) either finds out of keeping with

the defined territory, or isolated finds (see Fig 6 for key).

Needham CW3 FINAL 21 iv 16.indd 47

27/05/2016 12:45:31

ŽŶĞŚĂůďĞƌĚ ƚǁŽŽƌŵŽƌĞŚĂůďĞƌĚƐŝŶŚŽĂƌĚ ǀĂŐƵĞƉƌŽǀĞŶĂŶĐĞŽƌŚĂůďĞƌĚŝŶ ůĂƚĞƌĐŽŶƚĞdžƚ ĐŽƉƉĞƌŚĂůďĞƌĚƐ ůĂƚĞŚĂůďĞƌĚƐ;ŵĂŝŶůLJďƌŽŶnjĞͿ ƐƚŽŶĞŵŽƵůĚĨŽƌŚĂůďĞƌĚƐ;ůĂƚĞͿ ĞĂŬĞƌĨƵŶĞƌĂƌLJĐŽŶƚĞdžƚ

2.6 Map of recovery for copper and bronze halberds in Britain and graves datable to the Chalcolithic and transition period, c. 2450–2100 BC. Late types of halberd are distinguished and comprise those of bronze along with copper or unanalysed examples of the same types, c. 2200–2050 BC. Numerals indicate the number of contexts in dense clusters of Beaker graves (encircled)

than is possible at present for many other regions. Obviously, there are all sorts of caveats about the overall extent and very variable density of the graves plotted (shown separately in Figure 2.19); some dense concentrations are still the product of intensive fieldwork and monitoring over decades and centuries—Wessex, the Upper Thames Valley, the Yorkshire Wolds and Aberdeenshire—but interestingly the Peak District has only a modest representation of early-dated graves despite a long tradition of excavation in the region.   The resulting grave distribution is not greatly dissimilar to the more specialized one for dagger/knife-containing graves, but the greater amount of data inevitably complicates the picture. In particular, the east-west divide is less neat and instead more ‘inter-penetration’ is evident (Figure 2.6). Despite this, it is clear that most halberds come from areas devoid of, or very poor in early Beaker graves and it is possible to define regions of varying shapes and sizes that are dominated either by halberds or early Beaker graves. The main exceptions appear to be in Aberdeenshire and Lothian, and it is noteworthy that the halberds within these Beaker enclaves are all of late styles dating to around or soon after the metallurgical transition. It is possible therefore that there was a break-down over time in the pattern of mutual exclusion.   Two less clear cases are of single halberds from Sancton, East Yorkshire and Faversham, north Kent, both of which lie close to Beaker grave concentrations and might, with future discoveries, prove to fall firmly within early Beaker enclaves. Nevertheless, it is intriguing that the Sancton halberd is currently in a part of the Wolds away from the dense

Needham CW3 FINAL 21 iv 16.indd 48

27/05/2016 12:45:32

Needham

[ 49 ]

2.7 The early Beaker context at South Hill, Talbenny, Pembrokeshire (Fox 1959, 45 fig 31B) and the Beaker pot (Savory 1980, 200 fig 52)

concentration of Beaker graves to the north. More detailed scrutiny at the regional level would be necessary to confirm this as significant. It may also be of note that these halberds are two of just four in Britain which have been attributed to the Continental 3-rivet series (Figure 2.2 nos 13, 17 & 19; Needham 2015a, appendix S1, 21–2).   Setting aside regions yielding only one or two securely provenanced halberds (Great Glen; Tweed Valley; eastern Fens; Lower Thames Valley), we are left with five halberd enclaves, two apparently large ones in the west (much of Wales; south-west Scotland with Cumbria) and three small ones in eastern Scotland (east-coast Sutherland; Nairnshire (centred on); Fife). The enclaves delineated have yielded very few early Beaker graves and two from Wales merit comment. The most westerly Beaker find plotted in Wales is a Low-Carinated vessel from Talbenny, Pembrokeshire. Even allowing for the possibility of wholly decayed bones, the feature it came from is not a classic grave, as Fox recognized; he ventured that the deeper part of the feature ‘seemed to symbolize a grave rather than be intended for a human body’ (Fox 1959, 46). Although about 1.8m long, it was deep and narrow—only about 0.3m wide at the bottom (Figure 2.7). This trench-like feature runs diametrically (NNE–SSW) across the base of a circular pit, 1.8m in diameter with steep sides and a flat bottom 0.3m deep. The pit seems originally to have been packed with stones (size range 0.1–0.3m) although these only survived thus in the north-western part (just under half the feature), where they then rose out of the pit to form a mound about 0.4m above the surrounding surface. The larger half, including the fill of the trench below the middle, had a fill of yellow clay with only sparse stones. The Beaker (or part of it) was embedded amongst the stone blocks/cobbles just inside the northern edge and a flint

Needham CW3 FINAL 21 iv 16.indd 49

27/05/2016 12:45:33

[ 50 ]

ii. lost cultures of the halberd bearers

]]

implement—unfinished arrowhead or a point-cum-knife—lay a little to the west. If the pit and trench were of two phases, a trench having been cut through the small stone mound and its underlying pit (as could be read from Fox 1959, 45 fig 31B), then the Beaker was incorporated in the primary phase. A fire-reddened patch of soil lay immediately to the SSW of the features and was possibly truncated by them.   Another possible early Welsh Beaker, Low- or Mid-Carinated, was recovered at Penderyn, at the western end of the Brecon Beacons. Unfortunately, it and a stone disc pendant (or spindlewhorl) were found during roadworks and no context is known (Savory 1980, 138 no. 366; Needham 2012, appendix 1).   The detailed geography of halberd finds in relation to early Beaker graves suggests more localized patterning than the simple west-east separation suggested by comparison with dagger-/knife-graves alone. This is perhaps not surprising; if these two key attributes do represent a fundamental ideological opposition, as I have argued elsewhere (e.g. Needham 2015a), then amongst small-scale societies this is unlikely to have played out in any neat pattern at the scale of the whole of Britain. While the small-scale complementarities noted could conceivably be due in some cases to the deposition of halberds by Beaker groups at or beyond the margins of their territories, for example in unoccupied land, the near universal absence of copper halberds from any zone densely or even modestly occupied by ŽŶĞĚŽŵŝŶĂƚĞĚďLJĞĂƌůLJĞĂŬĞƌŐƌĂǀĞƐ ŽŶĞǁŝƚŚƐƉĂƌƐĞƌĞĂŬĞƌŐƌĂǀĞƐ ŽŶĞĚŽŵŝŶĂƚĞĚďLJŚĂůďĞƌĚĨŝŶĚƐ ŽŶĞǁŝƚŚƐƉĂƌƐĞŽƌĨĞǁŚĂůďĞƌĚĨŝŶĚƐ

2.8 Interpreted territories for halberd-bearing and Beaker groups during the Chalcolithic to earliest Bronze Age in Britain. The only individual symbols shown represent either finds out of keeping with the defined territory, or isolated finds (see Figure 2.6 for key).

Needham CW3 FINAL 21 iv 16.indd 50

27/05/2016 12:45:35

[ 51 ]

Needham

early Beaker graves suggests otherwise. In summary it is concluded that halberds tend to occupy different zones from early Beakers, this is summarized schematically in Figure 2.8. It is clear also that the maps condense a diachronic situation. The latest British halberds are not only confined to northern Britain (Needham 2015a, appendix S1 fig S30), but also drift eastwards, counterintuitively therefore towards the early Beaker enclaves. Ireland The quantity of halberd finds in Ireland along with their fairly widespread distribution across the island has always given the superficial impression that they were in ubiquitous use there. More detailed scrutiny now suggests otherwise. Unlike for Britain, bronze halberds do not seem to show a different distribution from their copper forerunners; although there is a slight north-westwards drift over time in the overall weight of halberd distribution (Needham 2015a, appendix S1 figs S23–25), most bronze halberd locations merely consolidate pre-existing zones of occurrence.   Taking halberds in isolation, several zones of deposition might be suggested; they are scattered the full length of the island but leave many ‘empty’ zones (Figure 2.9). Some of these empty zones have in fact proved to be densely occupied by Beaker-yielding sites which have been systematically collated by Neil Carlin (2011) and, again, empirical enclaves can be ventured (Figure 2.10); recovery biases affecting the sudden upsurge in Beaker discoveries in 2.9 Map of recovery for halberds from Ireland with outline enclosures for concentrations (re-plotted from data in Harbison 1969 &

O’Flaherty 2002, with additions; less precisely provenanced examples are named in the margins)

Needham CW3 FINAL 21 iv 16.indd 51

2.10 Map of recovery for Beaker-yielding sites in Ireland with outline enclosures for concentrations (after Carlin 2011)

27/05/2016 12:45:40

[ 52 ]

ii. lost cultures of the halberd bearers

]]

2.11 Interpreted territories for halberd-bearing and Beaker groups during the Chalcolithic and earliest Bronze Age in Ireland, c. 2450–2050 BC. The outlines in Figures 2.9 & 2.10 are shown here as bold tones and have been trimmed to take account of suggested boundaries between halberd-bearers and Beakerusing groups. The purple lines indicate boundaries separating proximate clusters of the respective types. Lighter toned areas of red and blue indicate hypothetical, as-yet undocumented extensions of halberd and Beaker groups respectively. Areas of highest uncertainty are left untoned. The symbols shown here represent halberd finds or Beaker sites which appear to lie in ‘opposing’ territory or are very close to the interpreted boundaries.

Ireland are discussed in Appendix 1. These enclaves will undoubtedly be subject to revision as more discoveries come to light, but for the purposes of the present analysis, it seemed advisable to draw the enclave boundaries tightly around known findspots. In the case of the halberds in particular, but perhaps also Beaker sites, this allows for the possibility of boundary behaviour, where the focus of deposition might well be towards boundaries with social groups with opposing ideologies. By overlapping the two sets of hypothesized zones it can be seen that there is rather little overlap between the two categories of evidence and it is possible to link up local boundaries between them into a continuous, convoluted halberd zone running along a south-west/north-east spine through the island (Figure 2.11). Only a handful of Beaker sites occur within the ‘halberd zone’ and most of these are in very marginal locations. Similarly, only a few halberd finds are any significant distance into Beaker zones. Some of these cross-over finds could well be due to shifts in allegiance during the course of the period covered. Both within and outside the halberd zone there are several significant blanks with respect to the given classes of evidence and only time will tell whether these conform to either, or instead to some entirely other kind of cultural idiom which adopted neither Beaker material nor halberds.   Despite the sudden plethora of Beaker discoveries in Ireland, there are still no classic Beaker ‘single-graves’ (cf. Cleary this volume). For a long time it was unclear whether there were any formal burials truly belonging to the Chalcolithic period. Newer finds, added to

Needham CW3 FINAL 21 iv 16.indd 52

27/05/2016 12:45:45

Needham

[ 53 ]

2.12 Map of recovery for burials or human skeletal remains most securely attributable to the Chalcolithic period in Ireland in relation to the hypothesized halberd/Beaker boundary. Solid symbols are those more securely dated to the Chalcolithic. Compiled from selected data in Carlin 2011, Mount 2012.

a few old ones, now suggest there were some burials, but even so firmly dated examples are extremely sparse by comparison with the wealth of burials in the succeeding Food Vessel era, the start of which is dated to c. 2150 cal BC by Brindley (2007, 328). Nevertheless, it is instructive to compare the distribution of the small number most confidently attributable to the Chalcolithic with the hypothesized halberd/Beaker boundary; as would be expected given the aversion of halberds and perhaps therefore also halberd users to tangible funerary rites, the Chalcolithic burials do all come from outside the halberd zone and, moreover, from established Beaker enclaves (Figure 2.12).   This background is seen to be perpetuated, albeit with some shifts, in the succeeding Food Vessel/Urn period of the Early Bronze Age. The Beaker enclaves in the east of Ireland are apparently the springboard for the adoption of single crouched inhumations in cists and pits, but this phenomenon also encroaches on limited parts of the former halberd zone as well as some areas to the west. The picture becomes yet more complicated if we also take into consideration Wedge Tombs, the construction and use of which starts in the Chalcolithic. Wedge Tombs are largely complementary geographically to cist and pit graves (Waddell 1990, 36 fig 7; 1998, 93 fig 42) and again encroach on parts of the former halberd zone. Nevertheless, some noteworthy gaps between the different burial traditions from eastern County Limerick in the south-west to Lough Neagh in the north-east roughly correspond to former halberd concentrations and suggest a zone, albeit somewhat

Needham CW3 FINAL 21 iv 16.indd 53

27/05/2016 12:45:50

[ 54 ]

ii. lost cultures of the halberd bearers

ŽŶĞƐǁŝƚŚĨŽƌŵĂůďƵƌŝĂůƚƌĂĚŝƚŝŽŶƐ ŽŶĞƐǁŝƚŚŽŶůLJŽĐĐĂƐŝŽŶĂůďƵƌŝĂůƐƌĞĐŽƌĚĞĚ ,ĂůďĞƌĚͲďĞĂƌŝŶŐnjŽŶĞŽĨƚŚĞŚĂůĐŽůŝƚŚŝĐ

]]

2.13 Areas with more than occasional evidence for formal burial in Ireland during the Early Bronze Age, c. 2150–1500 BC; the darker toned area is the conjunction of zones with Wedge Tombs and those with cist and pit burials. Much of the areas lacking any tangible burials correlate well with parts of the earlier halberd zone (cf. Figure 2.11).

contracted, still resisting formal burial rites of any kind (Figure 2.13). There are grounds for seeing a long-term historical trajectory in which the fortunes of the different cultural groups waxed and waned, but in which there was some cultural legacy of the halberd bearers in parts of central Ireland during the Early Bronze Age. France Having found evidence in Britain, and then again in Ireland, for substantial spatial separation between halberd deposits and contemporary Beaker material, it was natural to enquire whether the same might be true of other parts of Atlantic Europe. At first, on consulting a range of published syntheses, it seemed that there might well be a reasonable separation in France. Developer-led archaeology in France, as in Ireland, has led to a deluge of new discoveries; fortunately, those up until 2013 have been collated and analysed by Raphaël Gadbois-Langevin (2013; many earlier sources were also consulted and further up-to-date information was generously provided by Cyril Marcigny, Marc Talon, and Henri Gandois). The gross Beaker distribution (Figure 2.14) is still patchy across the country but, given that significant additions have been made in the recent past, it is difficult to know how many of the remaining gaps represent genuine lacunae. There are, though, still substantial areas with extremely few Beaker finds, some of which may be difficult to dismiss as being due to

Needham CW3 FINAL 21 iv 16.indd 54

27/05/2016 12:46:00

[ 55 ]

Needham

poor visibility of the evidence. Moreover, Gadbois-Langevin’s map for the early phase of Beaker culture (campaniforme ancien) shows its presence in rather more restricted zones; the implication is that there was a real historical expansion of the Beaker culture, regardless of the process or processes at play.   It is difficult to determine whether the French assemblage of halberds was mainly contemporary with the early Beaker phase. Certainly the possibility exists that they cover a longish time-span (Appendix 1); occasional examples could potentially have been derived from north Italian halberds in pre-Beaker times, while a few others could belong to the close of the 3rd millennium BC. Despite this lack of precision relative to the chronological 2.14 Map of recovery for Atlantic halberds and Beaker sites of all phases in France (halberds after Gallay 1981 with additions; Beaker sites primarily after Gadbois-Langevin 2013)

njŽŶĞŽĨĚĞŶƐĞĞĂŬĞƌƐŝƚĞƐ njŽŶĞŽĨŵŽĚĞƌĂƚĞůLJĚĞŶƐĞƐŝƚĞƐ njŽŶĞŽĨůŝŐŚƚƚŽƐƉĂƌƐĞƐŝƚĞƐ ŝƐŽůĂƚĞĚĞĂŬĞƌͲLJŝĞůĚŝŶŐƐŝƚĞ

^ƋƵĂƚŚĂůďĞƌĚƐ dLJƉĞƚŝŐŶLJ ŽŶƚŝŶĞŶƚĂůϯͲƌŝǀĞƚ KƚŚĞƌƚůĂŶƚŝĐŚĂůďĞƌĚƐ DĞĚŝƚĞƌƌĂŶĞĂŶϯͲƌŝǀĞƚ ͚^ŽŵŵĞďĂƐŝŶ͛ ͚^ŽŵŵĞǀĂůůĞLJ͛ &ĞƌƌŝğƌĞƐͲ,ĂƵƚͲůŽĐŚĞƌ ĐŚƵƌĐŚƚƌĞĂƐƵƌLJ

͍ ͚^ĂƵŵƵƌĂƌĞĂ͛

͚ĞĂƵŶĞĂƌĞĂ͛ ͚^ĂƀŶĞ ǀĂůůĞLJ͛

ŶŐŽƵůġŵĞ DƵƐĞƵŵ

ZĂLJŵŽŶĚ DƵƐĞƵŵ͕dŽƵůŽƵƐĞ

ƐŝŶŐůĞŚĂůďĞƌĚ ƚǁŽŽƌŵŽƌĞŚĂůďĞƌĚƐ ŝŶĂƐƐŽĐŝĂƚŝŽŶ

Needham CW3 FINAL 21 iv 16.indd 55

27/05/2016 12:46:01

[ 56 ]

ii. lost cultures of the halberd bearers

]]

precision possible for Britain and Ireland, the spatial relationships between halberds and Beakers are still noteworthy; they are best reviewed on a regional basis. North France and Belgium Two halberds are attributed only to the ‘Somme Valley’, another lies close to the Belgian border (Armentières) and a fourth is from Wijcheln, well into Belgium (Antwerp province). This is a zone which historically has not been well endowed with Beaker material and although there are some significant additions in France, most of these fall away from the halberd finds. Despite a reasonable amount of developer-led archaeology in the region, Beaker material continues to be fairly sparse (Marc Talon, pers. comm.). North-eastern France Just two halberds are known and, even with the increase in Beaker discoveries, they still do not occur close to obvious concentrations of Beaker activity. There could of course be some recovery biases favouring the discovery of Beaker material in the well-occupied valley bottoms relative to the flanking hills. However, it is striking that even now no Beakers are known anywhere close to the Consenvoye find (Meuse), despite the plethora of finds to the east in the Moselle Valley. The case for the Euffigneix find (Haute-Marne) is a similar but less pronounced. Lower and Middle Seine Valley Five halberd finds are strung out fairly evenly along the Seine Valley from Les Andelys in the north-west to Etigny in the south-east, one find comprising an apparent hoard of four, now lost, from Épône (there are uncertainties as to how it related to a Romano-Gallic grave found at a the same time; Gallay 1981, 126; Mohen 1977, 36). A sixth may come from land to the west around the village of Ferrières-Haut-Clocher. It is possible that these relate to a sparse string of Beaker sites along the Seine and some of its tributaries and the most northerly of the halberd finds lies close to a marked concentration of Beaker sites immediately south-east of Rouen (Billard et al. 1998, fig 1). However, despite the extent of modern disturbance and archaeological excavation in and around Paris, Beaker material in this part of the valley is not frequent; moreover, little of it is defined as early-phase. Most of the halberds could alternatively be seen to be marginal to Beaker zones. Middle Loire Valley This area has yielded as many as seven or eight halberds, all from the river or its banks, and until recently seemed to be largely bereft of Beaker material. This has changed to some degree, but Beaker findspots along the valley itself are still patchy, two clusters being to some extent mutually exclusive of halberd locations; only one halberd lies within a Beaker cluster even now. The main area that has suddenly filled out with finds of Beaker material is the interfluve between the Loire and the Seine—the Loiret/Yvelines/Seine-et-Marne borders. This falls between the two halberd yielding rivers of the Seine and the Loire. Very little early-phase Beaker material has been recognized either on the interfluve or along the middle Loire.

Needham CW3 FINAL 21 iv 16.indd 56

27/05/2016 12:46:02

Needham

[ 57 ]

North-western France and the Channel Islands Ten halberds are known from nine findspots, a hoard of two coming from Château l’Etoc, Alderney. A group of two or three from the Loire estuary are now seen to lie close to some Beaker sites, the main concentration lying on the coastward side. This cluster does include some early-phase Beaker material on the coast, and it is noteworthy that all three halberds are of squat forms which are also potentially early (Figure 2.2 no. 9; Needham 2015a).   The four halberds in the interior of Brittany are now interspersed with a scatter of Beaker sites, including the important site at Saint-Nicolas-du-Pélem (Le Provost et al. 1972). However, none of these sites is regarded as being of early Beaker phase, finds of which are confined to the south and north coasts. The dense Beaker concentration along the southern littoral has yielded just a single halberd and it is intriguing that this is from the island of Gavrinis in the Gulf of Morbihan, a potentially liminal or boundary position.   Even more striking is the fact that the pair of halberds from the Channel Islands comes, not from Jersey, Guernsey, or Herm with their abundance of Beaker finds (e.g. Salanova 2000), but instead from Alderney which is so far bereft of such material. South-western France Five halberds with the possible addition of two unprovenanced ones in local museums come from this region which has traditionally been poor in Beaker finds. The growth in documented finds has produced some minor concentrations, notably around CharenteMaritime and the Gironde estuary, along parts of the Pyrenean foothills, and in Lot and Tarn on the south-western fringes of the Massif Central. Current research suggests that early-phase material is scattered, but includes a string of finds along the northern flank of the Pyrenees and a group occupying the centre-west of France. Of the five provenanced halberds, four lie outside Beaker clusters as currently known, although that from SaintSavinien is very close to the eastern periphery of the Charente-Maritime Beaker distribution. On the other hand, the fifth find does overlap the Lot/Tarn Beaker cluster, although early Beaker material is sparse here. Saône Valley & Alps There may be up to four halberds from the Saône Valley, but two are poorly provenanced. The two better provenanced pieces come from the river at Chalons-sur-Saône and Mâcon, a reach of the valley lacking in any early Beaker material. The northerly find does, however, overlap the edge of a strong Beaker cluster which develops to the north. A halberd from Fontaine-les-Puits in the western Alps is rather isolated and may conceivably be a preBeaker introduction from northern Italy (Figure 2.4 top right; Horn 2014, 112).   Clearly there are a number of difficulties about drawing firm conclusions on the spatial relationships between halberds and Beaker material in France, not least of which is the potential for new zones of the latter to be recognized with more systematic searching in developer-led archaeology. If one were to assume that the halberd finds are actually a part of Beaker culture, then it would certainly become necessary to see their deposition as being guided by very particular circumstances that did not affect associated elements of material culture. It is noteworthy, for example, that not a single halberd is known from

Needham CW3 FINAL 21 iv 16.indd 57

27/05/2016 12:46:02

[ 58 ]

ii. lost cultures of the halberd bearers

]]

ƐŝŶŐůĞŚĂůďĞƌĚ ƚǁŽŽƌŵŽƌĞŚĂůďĞƌĚƐŝŶ ĂƐƐŽĐŝĂƚŝŽŶ ŚĂůďĞƌĚĨƌŽŵƵŶŬŶŽǁŶŽƌ ǀĂŐƵĞƉƌŽǀĞŶĂŶĐĞ WĂůŵĞůĂ ƉŽŝŶƚ

ϲ͗Ϯ

ϳ͗Ϯ ϲ͗ϭ

ϰ͗ϯ

͍ ϯ͗ϰ

Ϯ͗ϭϮ

ϯ͗ϲ

Ϭ͗ϭϲ

2.15 Map of recovery for halberds and Palmela points in France, also showing the changing regional balance between the two types

the massive concentration of Beaker finds in the Midi and Languedoc. Again, most other known Beaker foci have yielded either no halberds or only one and that often in a marginal positon; this is the case for the small dense cluster along the Moselle, and for the less dense scatters across the Oise/Aisne/Marne catchment, Normandy, the southern Breton littoral, the Loiret/Yvelines interfluve, the centre-west, and the Saône to upper Rhône complex. The wet-context deposition of the majority of French halberds could potentially provide an explanation for special geographical/topographical locations, but it would still need to be asked why so few have substantial evidence for Beaker settlement in their immediate hinterlands.

Needham CW3 FINAL 21 iv 16.indd 58

27/05/2016 12:46:03

Needham

[ 59 ]

  It is instructive to compare distributions of halberds and other copper metalwork. Copper axes pose problems because it cannot be assumed that they were specific to Beaker culture. Palmela points, on the other hand, are generally regarded as being an integral part of the Beaker package in Iberia and France, although associations between Atlantic style halberds and Palmela points should be noted at Finca de la Paloma, Toledo, and possibly Vélez Blanco, Almería, Spain, in both cases also with tanged Beaker daggers (Figure 2.17; e.g. Horn 2014, Tfn. 80 & 86). In France, the Palmela point distribution strongly echoes two concentrations of Beaker material, one along the Mediterranean coast, the other along the Atlantic coast from around the Gironde estuary to the Gulf of Morbihan (Briard & Roussot-Larroque 2002; Labaune 2010; Nicolas 2013, 394 fig 206). The scatter of finds in between and spreading up north-eastwards as far as the Moselle in some cases also occur within or close to areas of known Beaker activity. Although there is some overlap along the Atlantic coast, halberds become more frequent northwards as Palmela points diminish in frequency (Figure 2.15). Again, in central to northern France halberds predominate, while those in the far south-west lie on the fringe of the fairly dense Palmela point distribution. As was recognized by Senna-Martinez (1994, 167 figs 4 & 5, 176), the deposition of these two types does not seem to be governed by similar processes and at the very least the changing balance of representation between them suggests a broad geographical change in material culture and/or depositional practice. Iberia The situation in Iberia is complicated by both the long currency of halberds, extending well into the Iberian Early Bronze Age and even into the Middle Bronze Age (Brandherm 2003; Horn 2014) and the early emergence of Beaker cultures (e.g. Harrison & Mederos Martín 2001; Mataloto et al. 2013). The various halberd types defined by previous researchers are discussed in Appendix 1, where it is concluded that the only group securely contemporary with the local Beaker sequence, and not necessarily all of it, are those attributable to the Atlantic family (Figure 2.2 nos 4, 44–7, 49–54, 71). The limited dating evidence of the Iberian-Atlantic series makes it difficult to be sure of their overall chronological span, but they probably did not overlap the early phases of Beakers in the peninsula.   The distribution of Beaker sites in Iberia has been compiled from various sources, notably various papers in Rojo-Guerra et al. 2005 and an unpublished collation of Beaker sites in northern and western Iberia by Catriona Gibson. As is the case for France and Ireland, recorded Beaker sites have increased dramatically in recent decades (as compared, for example, with Harrison 1977) and the picture presented here may already be incomplete. It has long seemed to be the case that densely filled Beaker zones contrast with large voids, often in immediate juxtaposition. While this uneven distribution is perpetuated in the new plot (Figure 2.16), it still remains possible that some of the void areas are simply bereft of adequate fieldwork. This is a question that can only be effectively addressed by local research.   The halberds of the Iberian-Atlantic series are not numerous and the relationship to established Beaker distributions seems not to be uniform. Brandherm has previously noted

Needham CW3 FINAL 21 iv 16.indd 59

27/05/2016 12:46:03

8W

njŽŶĞŽĨĚĞŶƐĞĞĂŬĞƌƐŝƚĞƐ njŽŶĞŽĨŵŽĚĞƌĂƚĞůLJĚĞŶƐĞƐŝƚĞƐ njŽŶĞŽĨůŝŐŚƚƚŽƐƉĂƌƐĞƐŝƚĞƐ ŝƐŽůĂƚĞĚĞĂŬĞƌͲLJŝĞůĚŝŶŐƐŝƚĞ

͚ŶŽƌƚŚͲǁĞƐƚ^ƉĂŝŶ͛

{

150 km

40N

40N 0EW

1.12. Settlement patterns. Surrounded zones correspond to the studied͍ areas. In blue: the western pattern with isolated farms and collective graves. In orange: dwellings with rows of houses and individual graves. dLJƉĞĂƌƌĂƉĂƚĂƐ dLJƉĞĂƷƚĂƐ ,ĂůďĞƌĚƐǁŝƚŚƐƚĞƉƉĞĚďƵƚƚƐ dLJƉĞƚŝŐŶLJ ŽŶƚŝŶĞŶƚĂůϯͲƌŝǀĞƚ KƚŚĞƌƚůĂŶƚŝĐŚĂůďĞƌĚƐ

͚:ĂĞŶĂƌĞĂ͛

4W

͚:ĂĞŶƉƌŽǀŝŶĐĞ͛

ƐŝŶŐůĞŚĂůďĞƌĚ ƚǁŽŽƌŵŽƌĞŚĂůďĞƌĚƐŝŶ ĂƐƐŽĐŝĂƚŝŽŶ ƐŝŶŐůĞŚĂůďĞƌĚĐĂƌǀŝŶŐ ŵƵůƚŝƉůĞŚĂůďĞƌĚĐĂƌǀŝŶŐƐ

2.16 Map of recovery for Atlantic halberds and Beaker sites of all phases in Iberia (halberds re-plotted from data in Brandherm 2003 & Horn 2014; Beaker sites after various sources—see Appendix 3)

that Iberian halberds do not generally occur in association with Beakers (2004, 325), but this deserves closer scrutiny. Rock-carvings of halberds are also plotted on the map; these need not necessarily be representations of Atlantic series halberds, but the geographical association with actual halberds in the north-west of the peninsula encourages the notion that there was a relationship.   The strongest case for a pattern of complementarity lies in the north-west. In the province of Trás-os-Montes e Alto Douro, north-east Portugal, there is a cluster of five halberd findspots including three hoards (Figure 2.16), giving a total yield of ten halberds. Rock carvings of halberds, although not certainly contemporary, may add to the evidence here: two sites lie not far from the three hoards, another is further south-east and a fourth lies to the north. Together these halberd manifestations occupy a zone that is currently poor in Beaker finds, although a thin scatter is documented running east–west just south of the halberd finds.   More numerous carved depictions occur in western Galicia and are important in potentially consolidating a halberd zone here, for only two actual halberds are known

Needham CW3 FINAL 21 iv 16.indd 60

27/05/2016 12:46:05

Needham

[ 61 ]

(another is described only as from ‘north-west Spain’). Brandherm documents ten sites with carvings at the majority of which more than one halberd depiction is present; sometimes there are several (2003, nos 1715–57). Beaker sites are few within the limits of this halberd enclave but a small number occur immediately to the north and others are interleaved with the southern zone of halberd depictions.   Another cluster of rock-carved halberds occurs in the south-western corner of Portugal, where few Beaker sites are known. Three of the rock-carved sites lie hard up against the western edge of a Beaker-site concentration, part of a more general distribution of Beaker sites stretching north through the Alentejo. While south-west Portugal has not yielded any Atlantic style halberds, there are examples of Type Cano and possible examples in a much earlier grave at Alcalar. If accepted as halberds, these other types could suggest that the region had a long-standing association with the weapon type, even though actual halberd finds are lacking for the later 3rd millennium BC.   In contrast to the regions discussed so far, some isolated halberds were deposited within or very close to major Beaker-site concentrations: in the Upper Duero basin, Estremadura, western Andulacía, southern Valencia and Cataluña. This could suggest that they ‘belonged’ with the local Beaker cultures, although in the last two regions the halberd finds are intriguingly located between apparent Beaker enclaves.   Although there is a growing and fairly even distribution of Beaker sites across eastern Andalucía, no contemporary halberds have been reported from the core of this large zone. As many as five lie closer to what currently appears to be a northern limit to the Beaker territory, although sadly three of these finds are vaguely or uncertainly provenanced.   Two individual halberd locations are currently in Beaker-void areas, one on 1 2 the southern flank of the Ebro Valley, the other to the north of the major focus of Beaker sites around the Upper Tajo (Tagus) Valley south of Madrid. Two more halberd finds occur on the edge of this substantial Beaker concentration, one being a hoard from Finca de la Paloma (Brandherm 2003) containing two halberds associated, appropriately perhaps, with Beaker types—a tanged dagger, three Palmela points and a fragment of goldstrip ornament (Fig 2.17). Intriguingly, beyond these two ‘border finds’ to the north-west no Beaker finds are recorded for 100 kilometres. 2.17 The hoards from Vélez Blanco (?), Almería, and Finca de la Paloma, Toledo (Brandherm 2003)

Needham CW3 FINAL 21 iv 16.indd 61

27/05/2016 12:46:05

[ 62 ]

ii. lost cultures of the halberd bearers

]]

  Three and possibly all four of the halberds from this central Meseta zone are of Type Baútas (Figure 2.2 nos 44–7), the type-site for which lies within the Estramaduran Beaker zone at the Tagus estuary. Three other examples of the type are associated with the Andalucían Beaker zone, two being in marginal locations as already noted. This does raise the possibility that Type Baútas is a specifically Beaker type of halberd and the association at Finca da Paloma could be taken to support this hypothesis. That hoard may not be the only association between Atlantic style halberds and metalwork regarded as of Beaker type. A group of five tanged daggers and a halberd of Type Carrapatas were evidently found together during earth-moving at Monte Lioira, La Coruña (Brandherm 2003, nos 160, 164–6, 257 & 1366), while a halberd of Type Etigny is said to be part of an association uncertainly from Vélez Blanco, Almería (ibid, no. 147; Figure 2.17). In both cases the recorded locations are close to the borders of Beaker enclaves.   The overall picture in Iberia is thus of a degree of mutual exclusion between halberds and Beaker culture material, but also some intriguing spatial overlaps and points of contact in occasional hoards. Local studies may well help clarify the nature of these relationships, but it is worth noting that the gross distribution pattern in Iberia for tanged daggers of the broadly Beaker style could not be more different from that of Atlantic halberds; they are not only numerous, but also respect fairly closely the concentrations of other Beaker material (Brandherm 2003, most of types F–O, Tfn. 128–34). Discussion I have been careful to emphasize throughout this paper that the background to the adoption and deposition of halberds should not be expected to be universal throughout Atlantic Europe. Apart from the potential for regional differences of receptivity to externally introduced material culture, interpretation of the maps of recovery needs to allow for taphonomic vagaries, temporal shifts and the possibility of some accommodation between opposed social groups over time in some localities. Nevertheless, it is clear that it is difficult to reconcile the occurrence of later 3rd millennium BC halberds with zones of Beaker settlement and activity across a surprisingly broad area. The map of recovery of halberds simply does not correspond with that of Beaker culture sites, many large and dense Beaker enclaves being devoid of halberd finds, or virtually so. If halberds were in fact an integral part of Beaker culture, then we would need to explain why, unlike tanged copper daggers, knives, and Palmela points, so many were deposited away from cores, at best in marginal positons and sometimes at some remove. They represent a highly specialized type with potentially specialized symbolic connotations. The very symbolic power of the halberd as an unusual and, for its time, technologically advanced weapon would have lent itself to serving as an icon for a distinct, non-Beaker cultural persuasion.   At the more detailed regional level it has been seen that the spatial relationship of halberd finds to Beaker sites is not entirely uniform—a degree of mutual exclusion may be frequent, but it is not universal. It is difficult to explain local variation other than by speculation because of the dearth of contextual evidence for halberds, but it is nevertheless

Needham CW3 FINAL 21 iv 16.indd 62

27/05/2016 12:46:05

Needham

[ 63 ]

of interest that the evidence has to be scrutinized at a fairly small geographical scale in order to pick up patterns of potential mutual exclusion. There may be several reasons for the impingement of one type of evidence upon the other even within the overarching two-culture model being proposed. The first has already been noted—that some overlaps could easily be due to diachronic changes in boundaries over three or more centuries of co-existence. This would hardly be surprising in a situation where ideologically opposed culture groups existed side by side. It may be significant that at least four of the hoards combining halberds with Beaker metalwork types may come from boundaries—Whitespots in Ireland, Castell Coch in Wales, and Finca de la Paloma and Vélez Blanco in Spain; Monte Liori possibly also—and these could be interpreted, for example, as boundary-pact deposits. This leads on to the second possible reason, the potential for the amelioration of inter-group rivalries over time and the associated weakening of the symbolic charge of halberds such that they came to circulate to some degree amongst Beaker societies as well. Yet another possibility is that some halberds could have been carried into Beaker zones as curiosities or trophies to be buried securely at the heart of the territory.   Setting aside the particular causes of impingement at the sub-regional level, it is worth reflecting on possible differences in receptivity and separation at a broader scale. The fact that Beaker/halberd opposition seems to be particularly strong in Britain and Ireland may well suggest that the ideological background to that opposition was already entrenched before either phenomenon reached these islands. Inevitably, we have to backtrack to the Continent. It is probable that northern and western France and the Lower Rhine were the springboard for most of the cultural influences affecting the islands at this time. As has been seen, the picture in this zone is not so clear-cut, but the preconditions are nevertheless present. The land-mass between the Gironde and the Rhine has a number of strong Beaker enclaves documented, while halberd finds are mainly from zones with light to sparse Beaker presence.   The chronologies of Beaker and halberd inceptions in France were different from those in the islands, and those in Iberia possibly different again. In both continental regions there could have been a significant time lapse in the arrival of the two cultural phenomena being discussed, although these may have been in opposite directions. Halberds may have begun to circulate as an ‘undercurrent’ in France prior to the first Beaker introductions, although ‘uptake’ is unlikely to have happened until after. The Iberian situation is complicated by uncertainty over the classification of the early Type Alcalar blades as halberds, but these are unlikely anyway to be the progenitors of the Iberian-Atlantic series which the present author sees as stemming from France; the current slender dating evidence suggests that this series emerged somewhat later than the earliest Beaker culture in the peninsula. Of course, there are further complications depending on the particular timing of adoption of either manifestation at the more local level. Nevertheless, whatever the blur-factor on the Continent, the implications of the discreteness of halberds and early Beaker material in Britain and Ireland are profound and the clarity of the pattern may simply have resulted from the joint circumstances of, firstly, the background ethos’s being at variance with one another and, secondly, their arrival at more-or-less the same time. At the Atlantic Europe scale, regional historical differences in timing could well account for variant ensuing

Needham CW3 FINAL 21 iv 16.indd 63

27/05/2016 12:46:05

[ 64 ]

ii. lost cultures of the halberd bearers

]]

trajectories of opposition and/or assimilation.   Aside from the overly conservative chronologies that were applied to halberds in the past, another factor which may have militated against the recognition of any discrete halberd-bearing groups hitherto lies in the modest numbers of weapons found. With the exception of Ireland, Atlantic halberds finds are never particularly dense on the ground even where they are clearly concentrated and geographically exclusive of Beaker enclaves. Is it admissible to consider that just a handful of finds actually represent wholly different cultural groups occupying reasonable tracts of land? The economics of metal circulation help to put this into perspective. Atlantic halberds are not only normally large bodies of metal, but they are also invested with considerably more technical skill than contemporary metalwork types. It is clear from damage and use-wear patterns that many were heavily used and kept in circulation for long periods, sometimes having been re-hafted after substantial damage to the original hafting-plate (O’Flaherty 2007, 231; Brandherm 2011; Horn 2014, 182–220). A longer average use-cycle will give rise to a correspondingly smaller number of products for the same artefact population in circulation. A further and possibly even more influential factor would be the retention of the metal in circulation through regular recycling. The Galician evidence may be instructive here since the number of rock carvings far exceeds that of recovered halberds, suggesting that halberds were much more frequent in use than the finds themselves might imply. The decision to deposit a halberd in perpetuity may not have been taken lightly and perhaps one hint of the kinds of situation for which this act was deemed necessary comes from the frequency with which halberds were deposited close to interpreted territorial boundaries. This suspected behaviour creates another problem, that of the deposition pattern not being representative of the territory at large, for we cannot even presume that all boundaries would have been subject to the same pressures and depositional responses. Neither can we necessarily assume that halberdusing groups had access to their own metal supply. Certainly there is much overlap in the impurity signatures found in halberds and Beaker metalwork and both may have been dependent on the same range of sources with the possibility of there having been regional variation in relative access.   Whatever the variations in its origins and development, ‘Beaker culture’ can be seen to be a definable identity adopted by social groups as a whole (at whatever scale those were defined in the time and place concerned). The evidence for the geographical expansion of Beaker identity from an early phase to later phases almost anywhere one looks adds weight to the suggestion of there being some kind of territorial definition, however strongly or diffusely defined. The logical corollary of such a model is that other regions beyond those boundaries, held other identities, ones that either rejected the Beaker way of life or saw no gain in affiliating with it.   Characterizing these ‘other’ social groupings has not been a priority for this period. Research has been largely preoccupied with the highly distinctive and highly visible elements of the Beaker culture, distinctiveness and visibility being enhanced by the contrast with pre-existing material culture and practices throughout the Atlantic world. In many cases it must be suspected that gaps within the Beaker pattern were filled by groups that largely or partly continued indigenous traditions. If little changed in archaeological attributes, then

Needham CW3 FINAL 21 iv 16.indd 64

27/05/2016 12:46:05

Needham

[ 65 ]

only good independent dating will demonstrate continuation into a Beaker-contemporary era. But of course, this was a period of great cultural flux and it is equally likely that societies still broadly adhering to indigenous values were also reacting to the changing world around them. If so, many of those changes might be quite localized, thus needing to be identified through appropriate regional research (assuming they are archaeologically visible). If it is correct to deduce that halberds were being taken up for socio-political reasons by some of those indigenous social groups, it does mean that we cannot envisage the existence of a halberd-bearing culture in the mode of the Beaker culture (irrespective of how the latter was formulated and spread in detail). Halberds seem more likely to have been opportunistically grafted on to existing cultural bases and their unity lay only perhaps in their desire to stand against a collectively strong cultural entity. One element of the behaviour package that is discernible in all Atlantic halberd-specific zones is the absence of any tangible burial practice and the deposition of valuables in other contexts instead, but the former is again a form of negative evidence, rather than an especially unifying characteristic and the latter may be an understandable act often resulting from the opposition hypothesized.   Halberds do, however, present a very tangible cultural asset and, given their general patterning across the landscape, it is far more plausible that they were one element of the other identities than that they were associated with Beaker culture, especially early Beaker culture. As a very distinctive and highly displayable weapon form, visible and readable from a distance, halberds had the potential for being an icon of ‘opposition’ to the ever successful and encroaching Beaker way of life and it may be that this was the main driving force behind their adoption across many regions of Atlantic Europe and to varying degrees. It may also have been the driving force behind much permanent deposition, one result of which was to make these often ‘hidden’ cultures partially visible to archaeology. Beaker culture may stand out prominently against the Late Neolithic cultural background it impinged upon, but it may be time to acknowledge that other significant cultural forces were coursing through western Europe at much the same time.   To round off, something should be said about halberds with respect to the theme of ‘Celtic from the West’. It seems certain now that metal-headed halberds were not an invention of the Atlantic world, despite much past debate on the subject, and there are hints of a spreading undercurrent of halberd use prior to the later 3rd millennium BC. What I do suggest, however, is that halberds were first taken up with vigour in a number of territories, one by one, throughout the Atlantic zone and this represents the first widescale employment of the weapon type as a social and cultural icon. It would be invidious to try to chart the detailed chronology of uptake in the various regions given the sparsity of tight dating evidence, but it is likely that the process began in France and spread outwards from there. Although no uniform inception horizon may be identified, unlike the case for the primary Beaker package, nevertheless the Atlantic family of halberds has sufficient linkages to suggest that it too represented some kind of network of affiliation through much of Atlantic Europe during the last centuries of the 3rd millennium BC. Just as with the Beaker culture, regionalization kicks in rapidly such that halberd styles change

Needham CW3 FINAL 21 iv 16.indd 65

27/05/2016 12:46:05

[ 66 ]

ii. lost cultures of the halberd bearers

/ƌŝƐŚͲ ƚůĂŶƚŝĐ ƐĞƌŝĞƐ

ƌŝƚŝƐŚͲ ƚůĂŶƚŝĐ ƐĞƌŝĞƐ

EŽƌƚŚ ƵƌŽƉĞĂŶ ƚůĂŶƚŝĐͲ ƐƚLJůĞ ŚĂůďĞƌĚƐ ƵŶũĞƚŝƚnj ƐĞƌŝĞƐ

&ƌĞŶĐŚͬ>Žǁ ŽƵŶƚƌŝĞƐͲ ƚůĂŶƚŝĐƐĞƌŝĞƐ

,ŽƌŶ͛Ɛ dLJƉĞϲ EŽƌƚŚ /ƚĂůŝĂŶ ŚĂůďĞƌĚ ƐĞƌŝĞƐ

/ďĞƌŝĂŶͲ ƚůĂŶƚŝĐ ƐĞƌŝĞƐ ƌŐĂƌŝĐ ƐĞƌŝĞƐ

]]

2.18 Schematic interpretation of the strength of typological links between regional Atlantic halberd groups and contemporary halberd groups to the east. The links expressed are a combination of possible transferred objects and stylistic similarities. Essentially later halberd series are shown in pale grey.

dLJƉĞ ŽƚƌŽŶĞŝ

significantly from one end of the zone to the other (Figure 2.18) and it should certainly not be assumed that there was continuous long-term interconnectivity throughout the network; interconnectivity may have been much more partial and fluid over time being dependent on fluctuating peer-group needs. Unsurprisingly, judgements on the strength of regional connections based on typological similarities and potential imports/exports merely reinforces geographical location. There may have been another key difference from the Beaker network. The latter, in its pioneering stages, linked groups that seem to have transmitted or adopted not only an ideology, but also a material package. Halberds may instead have been more of a symbolic transmission: the form and attached meaning could have been fairly faithfully transmitted, based on their primarily oppositional role; however, there is no evidence as yet that this form-plus-concept accompanied a broader cultural package and instead it must be supposed that it was grafted onto pre-existing and varied cultural foundations. This deserves closer research at the regional level. Acknowledgements This paper proved to have a more ambitious scope than had been anticipated. The pace of discoveries of Beaker sites is such that their regional distributions are developing at an alarming rate. Fortunately, I was bailed out by various individuals with more specialized knowledge of the Continental and Irish material and hopefully what has been presented is therefore not wildly out of date. Raphaël Gadbois-Langevin’s thesis, assimilating all sites yielding Beaker material, came to my rescue for France; I am very indebted to Raphaël for encouraging me to make use of his data. Other French scholars too kindly provided information: Cyril Marcigny, Marc Talon and Henri

Needham CW3 FINAL 21 iv 16.indd 66

27/05/2016 12:46:06

Needham

[ 67 ]

Gandois. The dramatically expanded Irish evidence could not have been appreciated properly without Neil Carlin’s collation of it; again, I am grateful to Neil for access to his thesis and his willingness for me to use it in an unconventional way. Catriona Gibson and Bob Chapman pointed me to the latest summations of Beaker material in the various regions of Iberia, but responsibility for transcribing this complex evidence remains mine. I am also grateful to Marta Díaz-Guardamino Uribe for sight of her thesis on the carved stelae and statue-menhirs of Iberia, to Rui Boaventura for advice on Iberian Beaker chronologies, and especially to Erio Valzolgher for bringing my attention to the latest evidence for the dating of certain Remedello culture graves.   In relation to halberds, my main debt goes to those curators responsible for major collections in Britain and Ireland: Mary Cahill (National Museum of Ireland), Trevor Cowie (National Museums Scotland), Adam Gwilt (Amgueddfa Cymru—National Museum Wales), Neil Wilkin (British Museum) and Alison Roberts (Ashmolean Museum). Joanne Turner (Dumfries Museum) and Caroline McDonald (Museum of London) are also thanked for access to examples in their care. Greer Ramsey, Brendan O’Connor and Henri Gandois supplied additional information on halberds from Northern Ireland, Scotland and France respectively.

Appendix 1: Definition of relevant material and chronologies The main aim of this paper obviously requires that, as far as possible, coeval cultural phenomena are being compared with one another. This cannot be done with precision and, moreover, definition necessarily varies from region to region within Atlantic Europe. Not only are the chronologies of regional halberd series rarely fully understood, but there are also problems in the temporal bracketing of Beaker-culture phenomena. This appendix discusses the issues from region to region briefly.   A start date is given by the first presence of Beakers and halberds locally, and this varies with geography. Beaker-assemblage origins shift earlier in general as one moves southwards towards the Tagus estuary of western Iberia. The emergence of the metal-headed halberd may be more complicated and is currently poorly understood. Suffice it to say here that the earliest contexted examples lie well to the east of the Atlantic zone and occur well before the emergence of Beaker culture (Horn 2014; see also accompanying paper). There is no evidence for any significant uptake of halberds in any Atlantic zone of Europe prior to the mid-3rd millennium BC and thus prior to the emergence of the Beaker culture. The approximate end date used here is the turn of the 3rd/2nd millennia BC, although the tail ends of both Beaker and halberd sequences might cross this threshold in some Atlantic regions. Halberds in two major cultural complexes outside the Atlantic realm—the Aunjetitz of northern middle Europe and the Argaric of south-east Spain—have a quite different chronological span. Although both complexes are seen to start around the 22nd century BC, the majority of halberds in these two regional groups are in contexts datable to post-2000 BC. The Aunjetitz lies outside the scope of this study, but something is said about Argaric halberds below. Thus, overall, the time-frame under consideration is the second half of the 3rd millennium BC and potentially starting a little earlier towards the south.   The term Atlantic halberds is not intended to be closely definable in morphological terms, although many types can be seen to be interlinked by certain common features (Figure 2.2). Its

Needham CW3 FINAL 21 iv 16.indd 67

27/05/2016 12:46:06

[ 68 ]

primary purpose is to refer collectively to the great majority of halberds distributed through the Atlantic regions of Europe that have no close comparisons with the late Aunjetitz and Argaric series, with any other specialized late types (such as Horn’s Type 6), or with specifically Mediterranean styles (of wide-ranging date). Britain Contemporaneity between halberds and at least some Beaker material in Britain has long been clear, even though key associations are few. The swift and early transition from copper to bronze in the 22nd century BC, apparently for all classes of metalwork (Needham 2015a), immediately implies that the 90% of insular halberds made of copper are essentially of Chalcolithic date. There is no reason to doubt that most or all of the remaining 10% made in bronze followed without hiatus at the very beginning of the Early Bronze Age (Needham 2015a, appendix 4). The impurity patterns dom­ inant amongst halberds also support this early chronology in relation to insular metallurgy (ibid ). The current corpus of halberds may be found in Needham 2015a, appendix 2; this rejects three listed by 2.19 Graves and probable grave groups datable to the Horn, one being a dagger and the others Chalcolithic and transition period in Britain. Numerals inadequately documented (2014, 317–24 indicate the number of contexts in dense clusters of Beaker nos 14, 15, 42). The great majority of graves (encircled) British halberd finds can be referred to as a British-Atlantic tradition, a few instead being Irish-Atlantic or with more Continental affinities.   British Beaker contexts that should be broadly coeval with insular halberds are those belonging to the Chalcolithic or just after. Both the end of the Chalcolithic (c. 2150 BC) and the final use of halberds (c. 2050 BC) fall somewhere in the middle of the whole Beaker sequence and although clear phases can be defined at a gross level (most recently, Needham 2005; 2012), many individual graves cannot be firmly attributed to a phase due to the absence of diagnostic grave goods or secure radiocarbon dating. Graves plotted in Figure 2.19 have good evidence from either associations or independent dating, or both, that they belong to the Chalcolithic or shortly after; this includes a possible period of transition between copper and bronze working. In broad terms

Needham CW3 FINAL 21 iv 16.indd 68

27/05/2016 12:46:10

Needham

[ 69 ]

this should have isolated all graves for which there is good evidence that they date to before about 2100 BC and it is hoped these will be representative of the overall distribution for that date bracket.   Specific artefact types accepted as being essentially Chalcolithic are those defined in Needham 2012. This includes all Low-Carinated Beakers, even though a minority might continue later. Other pot forms are more problematic. Short-Necked Beakers certainly start during the Chalcolithic and may not continue a great deal later than my cut-off date of about 2100 BC (Sheridan 2007, fig 11.2). Likewise, there is increasing evidence that LowBellied Beakers (previously termed Low-Bellied S-Profile) are early and some are actually well dated to the Chalcolithic. These two forms are included where there is associated evidence (from artefacts or radiocarbon) to satisfy the above criteria, but not otherwise. Long-Necked Beakers are likewise included in the few cases where dating evidence suggests they are very early examples within that form lineage; none are likely to be earlier than transitional in terms of the metallurgical switch. Radiocarbon dates taken to be indicative of a date in the Chalcolithic or transitional period are those earlier than 3740 BP (uncalibrated). Graves were initially categorized according to the following criteria: A Both diagnostic grave goods and radiocarbon measurement are in keeping with an essentially Chalcolithic date (Σ = 33) B Diagnostic grave goods argue for essentially Chalcolithic to transitional date, but radiocarbon measurement looks too late (6) C Diagnostic grave goods argue for essentially Chalcolithic to transitional date; no radiocarbon measurement (124) D Diagnostic grave goods, supported by radiocarbon measurement where present, argue for a specifically transitional date (22) E Radiocarbon measurement argues for essentially Chalcolithic to transitional date; grave goods, if present, are not diagnostic of the Chalcolithic, but are nevertheless not out of keeping with it (53) Finds within categories A, B and E (all having radiocarbon dates) inevitably concentrate in zones that have been well provisioned by radiocarbon dating, notably southern-central England and north-east Scotland. Category C burials are much more widespread and include other concentrations in Yorkshire, East Anglia and Kent. With three exceptions, category D burials are southern, from Yorkshire to the south coast, but since this period is covered by plenty of radiocarbon dates in eastern Scotland for graves lacking diagnostic ‘transitional’ grave goods, this is not thought to be significant in terms of the extent of Beaker burial practice at this stage. All categories are therefore presented together in Figures 2.6 & 2.19. Ireland The corpus of halberds from Ireland has not changed substantially from that provided by Harbison (1969a). The few newer finds are from Ramsey et al. 1991–2; O’Flaherty 2002, Bell 2014 and unpublished finds—all are listed in Needham 2015a, appendix 3. The issues regarding the chronology of Irish halberds are essentially the same as for Britain, 90% being of copper and therefore presumptively dating to the copper-working stage (c. 2450–2150 BC). The

Needham CW3 FINAL 21 iv 16.indd 69

27/05/2016 12:46:10

[ 70 ]

ii. lost cultures of the halberd bearers

M1

M6 M50

M7

]]

2.20 The locations of Beakeryielding sites in relation to the motorway network in Ireland. Many of the sites were discovered in advance of construction of the M1–M4, M6–M9, M11, M50 and N25 between 1990 and 2012. The interpreted halberd zones (Figure 2.11) are shown toned.

N25

minority, made in bronze, are certainly mainly of the succeeding Killaha metalworking stage (c. 2200–2050 BC), although the possibility of occasional later examples cannot definitively be ruled out.   In complete contrast, Beaker discoveries have expanded exponentially during the years of economic boom after Ireland joined the European Union (see below), thus overturning past assumptions that receptivity to the Beaker material package was poor in Ireland. A high proportion of the finds came up in the course of developer-led archaeology and few are published. Fortunately, Neil Carlin has pulled together a comprehensive corpus for his doctoral thesis (2011) and it is probable that relatively few finds have been made since. Carlin’s evaluation of highest integrity radiocarbon dates associated with Beaker material in Ireland gave a date range for Irish Beakers as c. 2550/2500–2100 cal BC. If available dates are representative of the full span, then Beaker material does not outlive halberds, unlike the case in Britain, especially southern Britain. Factors affecting the distributions of halberds and Beakers in Ireland A high proportion of the newer Beaker discoveries has come to light as a result of large roadbuilding schemes in the Republic of Ireland. Meanwhile, casual finds of metalwork by people manually cutting peat, once commonplace, are now at an all-time low and this is significant for halberds, many of which come from wetland locations. Immediately, therefore, there is the potential for strong biases caused by two utterly different backgrounds to recovery; they may

Needham CW3 FINAL 21 iv 16.indd 70

27/05/2016 12:46:12

Needham

[ 71 ]

be giving radically different snapshots of real geographical representation.   A plot of the Beaker sites against the major road schemes opened in the Republic over the past quarter century shows an obvious correlation (Figure 2.20). Several strings of sites, often closely spaced, follow the road lines. This correlation may put limits on our understanding of overall Beaker site distribution across the landscape, but it is at the same time of considerable value. Roads can offer archaeologically impartial transects across the land and give an important window on relative presences and absences (assuming consistent surveillance). It is noteworthy that although most of the motorway routes yielded strings of Beaker sites, these are far from uniformly distributed (Figure 2.20). The dense to moderate distributions along the M1, M3, M8, M9, and M11 would seem to be representative of fairly systematic Beaker presence. However, the picture presented by the M4, M6 and M7, the westward routes from Dublin, is quite different. The M4/M6 has yielded virtually no Beaker finds and two of these are near Dublin itself. The M7 likewise has finds near to Dublin and again towards its western end—Limerick and the lower Shannon—but very little in between. These absences of Beaker material (other than occasional) coincide well with the heartland of halberd deposition.   The possibility that halberd distributions have been differentially enhanced by past handcutting of peat, particularly in the Irish Midlands, does not stand up well to scrutiny. It can never be known for certain whether finds from some environments have been recovered more frequently than from others, so the important comparison to make is with like material which has similar prospects for discovery. Copper and Early Bronze Age axes are by far the most abundant metalwork finds and these are widespread across Ireland despite the fact that many are from wet contexts (Harbison 1969b, pl 79). Taking a more specific group contemporary with halberds, Type Cappeen thick-butted axes (as defined in Needham et al. 2015, appendix 2) have a distribution that follows its own rubric; the majority of finds come from the southern half of the island and these are split between the halberd-yielding and Beaker-yielding zones. Copper daggers and knives, on the other hand, are almost all from the northern half of Ireland (Figure 2.5) and these are mostly from the halberd zone despite the fact that they are considered to be an archetypal ‘Beaker’ type. Given this variety of distributional emphases amongst contemporary metalwork types, the relative concentration of halberd finds in certain specific zones is likely to be significant in terms of original patterns of deposition, even if a little distorted by differential recovery patterns. France, the Channel Islands and the Low Countries The stock publication for halberds from France and the Channel Islands is that of Gallay, 1981, while the three examples from the Low Countries are covered by Butler 1963. A few finds have been made since (Cantet 1996; Cordier 2009; Despriée 1983; Le Roux 1985; Thévenot 1985). France, the Channel Islands and the Low Countries have yielded a good number of halberds; approaching 50 examples can be confidently ascribed to the class. They represent a fairly eclectic group in morphological terms, but most halberds can be described as falling within the Atlantic family and, given regional trends, may be termed French-Atlantic. Although dating evidence is sparse, these probably belong mainly to the latter half of the 3rd millennium BC.   Most numerous is the Continental 3-rivet series with between 17 and 20 examples (see examples in Figure 2.2, nos 5–9, 11–12, 14–15, 18). The series is discussed in Needham 2015a,

Needham CW3 FINAL 21 iv 16.indd 71

27/05/2016 12:46:12

[ 72 ]

ii. lost cultures of the halberd bearers

]]

appendix 1, and little need be added here. French hoards, such as those from Épône and Château l’Etoc, do not help with relative dating. The series is not really a feature of the Aunjetitz zone, but two occur in the large Dieskau II hoard (Wüstemann 1995, nos 137 & 173). The dating of Dieskau II and comparable Saxo-Thuringian hoards and grave groups is still highly debated (Horn 2014, 130–9) but deposition around the 20th century BC would certainly be in keeping with the British axe it contains. One of the Dieskau halberds plus occasional Scandinavian examples are made of bronze (e.g. Horn 2014, Tab.1 nos 13.13, 13.18 & 15.34), likewise implying a fairly late currency not earlier than Reinecke A1. However, given the lack of datable contexts in France and the typological threads linking Continental 3-rivet halberds to Italian halberds in one direction and insular halberds in the other, it is probable that they had earlier origins, particularly in the west. The possible association from Faversham, Kent, would allow the series to be current by at least the 22nd century BC (Needham 2012). The three-rivet halberd from Avoine compares well with Iberian Type Carrapatas (Figure 2.2 no. 55) and is thus excluded form the Continental 3-rivet series.   Eight squat halberds have been treated as a separate group in recent discussion (Needham 2015b). Although somewhat disparate in detailed form, they again mostly have three rivets (six examples) and one or two of these do resemble the larger halberds just dealt with. The halberd in the Wageningen hoard, Holland, probably dates to around the 21st century BC on the basis of its associations, but it is proposed that most squat halberds in this region are relatively early examples of the weapon class (ibid).   Between six and eight halberds have four rivets set within a sub-trapezoidal hafting-plate and most of these are similar enough to one another to be grouped as a type—Type Etigny (Figure 2.2 nos 65, 68–70). These have strong morphological links to British Type Pistill Dewy (ibid nos 64, 66–67). A further four-rivet halberd, from La-Grande-Paroisse, Pincevent (ibid no 80), instead recalls British Type Sluie. A rather individual style is seen in a two-rivet weapon from Saint-Lizier-du-Planté (Cantet 1996), but its general blade proportions and fairly broad midrib fit comfortably within the Atlantic suite. Others included with variable confidence are the halberds from Gavrinis, Châlons-sur-Saône, Vallée de la Saône and no provenance (Le Roux 1985; Thévenot 1985; Gallay 1981, nos 500 & 534A).   The remaining halberds plotted in Figure 2.14 probably belong instead with various Mediterranean styles. Two have three rivets set in a shallow arc within a low-arched butt, and are joined in this feature by one of the squat halberds (Gallay 1981, nos 506 & 530; Cordier 2009, fig 4 no. 7); this rivet format is present on halberds from Italy and the southeastern half of Iberia and also appears in central to northern Europe (e.g. at HofkirchenUnterschöllnach on the German/Austrian border; Horn 2014, 136–8). The three-rivet halberd from Mâcon (Gallay 1981, no. 529) has the more usual deeper triangle of rivets, but in other respects this does not look close to Atlantic styles and may again have more in common with certain Mediterranean halberds, notably Horn’s Type 8 (2014, Tf. 8; Bianco Peroni 1994, 30–2—Type Cotronei). Horn favours a date in the last three centuries of the 3rd millennium BC for this type (2014, 90–1). A halberd provenanced only to the Somme Basin (Gallay 1981, 533) has been placed by Horn with a group of halberds mainly from Italy (his Type 1; 2014, Tf. 1), a group which partially overlaps Bianco Peroni’s Types Calvatone and Montemerano (1994, 15–16, 35–6). These are not well dated and could possibly belong to a pre-Beaker horizon.   Two halberds could be marginal in terms of the defined date bracket (Figure 2.2 nos 73–4).

Needham CW3 FINAL 21 iv 16.indd 72

27/05/2016 12:46:12

Needham

[ 73 ]

That from Glomel (no. 73; Gallay 1981, no. 504) seems to have been associated with a range of daggers and an axe, perhaps deriving from a tomb assemblage (though not necessarily a singlephase assemblage). All but one of the daggers are of Gallay’s Type Rumédon and these could therefore fall either before or after 2000 BC; the exception has a midrib, placing it in Type Trévérec, currently thought not to have been developed until after the turn of the millennium. There is therefore scope for the halberd to have been deposited prior to 2000 BC, especially if the Trévérec dagger was the latest in a sequence of burial deposits. The rivet formation of the Glomel halberd may be linked to that of Type Etigny, but it is noteworthy that its shoulders are exceptionally ‘beaked’ and this links it instead to a halberd from Armentières which only has three rivets (Figure 2.2 no. 74; Gallay 1981, no. 505). The latter example has the strongly splayed midrib which is mainly characteristic of the Aunjetitz halberd series, but is also present on Chalcolithic Type Pontrhydygroes in Wales (Figure 2.2 no. 41). Excluded halberds One French halberd has been definitively excluded as being too late. The very long, decorated example from Montreuil-sur-Mer (Gallay 1981, no. 531) belongs to Horn’s Type 6 (2014, 23– 6), contemporary with the classical Aunjetitiz phase. This example may, however, show that halberds were of long-standing importance in northernmost France.   Five implements that Gallay classified as halberds are questionable (Gallay 1981, nos 517, 529, 532, 534, 536). No. 517 is a small blade with no obvious asymmetry and can be paralleled amongst daggers and knives elsewhere on the Continent. No. 532 could perhaps have been from a halberd, but it is again small and fairly slender with very thin rivet holes; it is not clear whether these are secondary after damage. The drawn outline of the lost and apparently corroded fragment No. 534 suggests a dagger blade because of its acutely angled tip. The tworivet, broad-proportioned blade from Mâcon (Gallay 1981, no. 529) is only a halberd by virtue of comparisons with implements in the west Mediterranean zone that have been regarded as halberds, including examples in stone or flint (e.g. Horn 2014, Tf. 45e, g & p, Tf. 50q & s). Finally, No. 536 is a lost and incomplete object which could very readily be identified as a long form of dagger based on the dimensions given. Iberia A very detailed corpus of Iberian halberds is owed to Brandherm (2003). There are two major halberd series in Iberia—the Iberian-Atlantic and the Argaric—and although there may be some temporal overlap, the present author believes it to have been rather little. The Iberian-Atlantic suite, comprising 27 halberds, has sufficient stylistic similarities to other ‘Atlantic’ types further north to argue for broadly comparable chronology in the late centuries of the 3rd millennium BC (Figure 2.2 nos 3–4, 44–7, 49–54, 71). Moreover, there are a few associations with developed tanged daggers of Beaker type which best equate with the Ciempozuelos Beaker pottery stage and the ‘Montelevar horizon’ in the last two centuries of the 3rd millennium BC (Harrison 1980, 140–6; Garrido Pena 1997; 1999, 334). Horn considered his Variant 12a, some members of which constitute a key ‘Atlantic’ type (Brandherm’s Type Baútas), to date to 2400–2000 BC (2014, 104–5), but he is less specific about other Atlantic types. Nevertheless, it is probable that the great majority of Iberian-Atlantic halberds pre-date the turn of the 3rd to 2nd millennia BC.

Needham CW3 FINAL 21 iv 16.indd 73

27/05/2016 12:46:12

[ 74 ]

ii. lost cultures of the halberd bearers

]]

  Halberds in Argaric contexts, very predominantly grave contexts, are of styles quite distinct from the Iberian-Atlantic series; instead of the steadily tapered blade of the latter, giving a broad-blade character, Argaric halberds contract rapidly below the shoulders or in some cases are narrower throughout (Types Monteagudo, Barranquete, Las Laderas and San Antón; Brandherm 2003). Halberds are found in both Argaric ‘A’ and ‘B’ graves, but not from the beginning of ‘A’ (c. 2200/2150 BC). The earliest radiocarbon date directly associated with a halberd-bearing grave is that from Herrerías, Grube Iberia, and while the determination itself (2141–1950 cal BC; 68% confidence) would allow a pre-2000 date, Horn believes the associations—a pottery vessel of Type 6b and a silver spiral ring—place it at the beginning of the 2nd millennium BC (Horn 2014, 100). On the evidence of associated pottery forms, Horn suggests that other early-looking halberd contexts are no earlier (ibid ). Grave 75 at Fuente Álamo contained a female skeleton dated 2121–1932 cal BC (68%), but also a male dated somewhat later, to 1961–1771 cal BC (68%); these may be successive burials (ibid , 99) and, if so, the halberd is perhaps most likely to go with the later insertion after 2000 BC. Graves with halberds are certainly well represented in Argaric ‘B’, c. 1900–1750/1700 BC, but may be primarily a feature of the earlier part since they do not occur in pithos graves (ibid , 100–1).   At a general level, therefore, it is relatively easy to remove the Argaric halberd series from the current enquiry. Outside these two main halberd series there are a variety of other forms, usually occurring in small numbers and sometimes singly. There can be problems with both dating and function, and not all can be discussed here. Other excluded halberds Potentially the earliest Iberian halberds are the blades of Horn’s Type 17 which only occur in datable contexts at Alcalar grave 3, Algarve, Portugal, and Los Millares grave 57, Almería, Spain, dated to between 3200 and 2600 BC by Horn (2014, 103), with an additional example from El Moral, Málaga, Spain. These are slender, midribbed blades with a notch either side of the tang or hafting-plate. The main argument for them being halberds rests on the slight skewing of haft-lines preserved in the surface of some examples. These implements are broadly earlier than Beaker culture, so even if acceptable as halberds they are not relevant to the current enquiry. The three sites occur strung out along the southern coastlands of Iberia from Almería to the Algarve.   Horn recognizes some very small, one- or two-riveted blades as halberds because of the survival of transverse wood-grain on the hafting-plate of some examples (2014, Tf. 44k, 44r, 44s & 45b). This does imply halberd style hafting and they may be diminutives; Brandherm notes that one was associated with a juvenile burial (2011, 36). They seem to be part of the Argaric funerary halberd phenomenon.   Again with two rivets are the possible halberds defined as Type Cano, mainly from the south-west of Iberia (Brandherm 2003, 238–42). These implements are slender throughout their length and although originally defined as only those with two rivets, Horn has recognized in his Variant 14a that some three-rivet examples are extremely similar in proportions (2014, 51–2). It is possible that the type began before the close of the 3rd millennium BC, but their main currency was probably later and, indeed, three examples have been shown to be bronze and thus probably not dating before the Iberian Middle Bronze Age (Senna-Martínez 2013, 14). There is the added problem that some Type Cano weapons have omega hilt-marks

Needham CW3 FINAL 21 iv 16.indd 74

27/05/2016 12:46:12

Needham

[ 75 ]

indicative of having been hilted dagger fashion. There is a complication in understanding how these blades functioned in that six of the longer blades within the group and two smaller blades have residues of both a skewed straight haft-line appropriate to a halberd and an orthogonal omega hilt-line appropriate to a dagger (Brandherm 2003, 241); others have only the omega hilt-line. It is possible that functional ambiguity and switching between functions were acceptable within this class (Brandherm 2011, 25–7), but there may be other explanations. One example (Brandherm 2003, no. 773) has a two-phase hafting represented by an omega hilt-line and a double-omega line, both of which would be expected to go with daggers. Where straight lines occur with omega lines, they are mainly found at the base of, or a little below the omega and might alternatively therefore record the top of a sheath rather than the line of hilting or hafting. Something similar is shown, for example, on a dagger from San Antón (ibid, no. 585). The examples with the straight line above the base of the omega amongst Brandherm’s Type Carvalhal (ibid, nos 792, 798) could indeed be due to re-hafting, indications of which are recurrent on Iberian daggers. With all these difficulties and a clear pattern of asymmetry in some unequivocal dagger hilts in Iberia, it may be safer to accept a default identification of Types Cano and Carvalhal as daggers.   There are a handful of further three-rivet implements categorized as halberds which tend to have fairly broad butts, but on which the rivets and butt-line form a very shallow arc. As well as the examples Brandherm grouped together as Type Arrayanes (2003, 378–80; see also Horn’s Variant 14c– 2014, 55), there are occasional similar looking blades placed elsewhere (e.g. Brandherm 2003, nos 1410, 1414, 1417). Although this type can occur in late contexts, shallow-arched three-rivet halberds were a long-running feature of the western Mediterranean zone, some in Italy being amongst the earliest halberds in Europe. When the general form came to be adopted in Iberia is unclear on current evidence, but Horn sees them as potentially of pre-Argaric adoption (2014, 55).   The individual halberd possibly from Valdepeñas (Brandherm 2003, no. 1384) is quite out of keeping with peninsular halberds due to its very deep hafting-plate and its trapezoidal arrangement of four rivet-holes. It also has an elongate triangular blade and Horn links it to a scatter of halberds with similar blades occurring much further north-east in Continental Europe, his Type 6 (2014, 23–6).   Type Montejícar (Horn’s Type 4) is a very distinctive type believed to date much later than 2000 BC, in the younger part of the Iberian Middle Bronze Age (Brandherm 2003, 253–5). Only a few are known and attribution to halberds rather than daggers rests again on haft-lines being occasionally skewed; in other cases they appear not to be (Brandherm 2003, Tf. 52; Horn 2014, Tf. 5a–h).

Needham CW3 FINAL 21 iv 16.indd 75

27/05/2016 12:46:12

[ 76 ]

ii. lost cultures of the halberd bearers

]]

Appendix 2: Sources for illustrations of Atlantic Beaker material and halberds in Figures 2.1 and 2.2 [for Figure 2.1] No Type

Provenance

Source

1

low-bellied Beaker

Harrison 1977, fig 68 no 1228

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

low-bellied Beaker low-carinated Beaker bone tube-toggle bone belt-ring shale disc beads amber V-bored button gold ‘basket’ ornament gold ‘sun-disc’ copper tanged knife copper tanged dagger antler rod stone 2-hole wristguard stone 4-hole wristguard copper Palmela point flint barbed –and-tanged arrowhead

Veiga dos Mouros tumulus 242, La Coruña, Spain Petit-Rosmeur, Finistère, France Chilbolton, Hampshire, England Sewell, Bedfordshire, England Sittingbourne, Kent, England Chilbolton, Hampshire, England Driffield, East Yorkshire, England Chilbolton, Hampshire, England Farleigh Wick, Wiltshire, England Chilbolton, Hampshire, England Sutton Courtenay, Oxfordshire, England Chilbolton, Hampshire, England Newlands, Aberdeenshire, Scotland Sewell, Bedfordshire, England Los Pasos, Zamora, Spain Roundway G8, Wiltshire, England

Salanova 2000, 230 fig Br 84 Russel 1990, 161 fig 5.1 Kinnes 1985, A9 no 4 Kinnes 1985, A13 no 3 Russel 1990, 159 fig 4 Kinnes 1985, A11 no 4 Russel 1990, 165 fig 7.14 Grinsell 1968, 51 fig 5 no 43b Russel 1990, 162 fig 6.1 Kinnes 1994, A18 no 1 Russel 1990, 162 fig 6.2 Woodward & Hunter 2011, 153 ID 91 Kinnes 1985, A9 no 3 Harrison 1977, fig 68 Annable & Simpson 1964, 88 fig 60

[for Figure 2.2] With rare exceptions, the halberds chosen for illustration in Figure 2.2 are in good condition with little corrosion loss and only minor damage. Outlines have been ‘restored’ where there is obvious damage in order that it does not detract from the likely original form. Where rivets survive, their heads have been centred on their holes (many are drawn or photographed in skewed positions). The resulting depictions gloss over nuances of midrib section, blade wing features and rivet form, but the main purpose is to facilitate ready inter-comparison amongst the full repertoire of halberd morphologies within the ‘Atlantic’ series. Occasional unique halberds have been left out (see Needham 2015b for a fuller range of squat halberds of Types Roscrea and Clonard and miscellaneous Continental examples). Occasional examples of very similar halberds from beyond the Atlantic zone are also shown.   For British and Irish finds numbers in square parentheses after the provenance refer to catalogue numbers in appendices 2 and 3 respectively in Needham 2015a. No Type

Provenance

Three-rivet halberds with deep hafting-plates 1 ?River Thames, England [14] 2 ?Baile-nan-Coille hoard (3), Strath Brora, Highland, Scotland [15] 3 Beluso, Boiro, La Coruña, Spain 4 Calle de Juan Margall, Gerona, Spain Continental three-rivet series 5 with triangular Saint-Denis-en-Val (3), Loiret, France butt 6 Ferrières-Haut-Clocher area (?), Eure, France 7 Euffigneix, Haute-Marne, France 8 with sinuous butt Luynes, River Loire, Indre-et-Loire, France 9 squat halberd Paimboeuf, River Loire, Loire-Atlantique, with arched butt France

Needham CW3 FINAL 21 iv 16.indd 76

Illustration based on; comments personal study Coles 1968-9, 38 fig 32.4 & personal study Brandherm 2003, no 1390 Brandherm 2003, no 1376 Despriée 1983, 259 fig 11 Gallay 1981, no 515 Gallay 1981, no 508 Gallay 1981, no 494 Gallay 1981, no 497

27/05/2016 12:46:13

Highland, Scotland [15] 3 Beluso, Boiro, La Coruña, Spain Brandherm 2003, no 1390 Continental three-rivet series 4 with triangular Calle de Juan Margall, Gerona, Spain Brandherm 2003, no 1376 butt 5 Saint-Denis-en-Val (3), Loiret, France Despriée 1983, 259 fig 11 6 Ferrières-Haut-Clocher area (?), Eure, France Gallay 1981, no 515 7 Euffigneix, Haute-Marne, France Gallay 1981, no 508 8 with sinuous butt Luynes, River Loire, Indre-et-Loire, France Gallay 1981, no 494 Needham 9 squat halberd Paimboeuf, River Loire, Loire-Atlantique, Gallay 1981, no 497 with arched butt France 10 with subRislev, Praestø, Denmark Vandkilde 1996, 198 fig 193 no 582 triangular butt 11 Saumur area, Maine-et-Loire, France Gallay 1981, no 519 12 Roermond, Limburg, Holland Butler 1963, 18 fig 2 13 Maryport, Cumbria, England [11] personal study 14 Château l’Etoc hoard (1), Alderney, Channel Gallay 1981, no 524: two rivet holes may be Islands secondary additions 15 Sost(-en-Barousse), Hérèchede, HautesGallay 1981, no 510 Pyrénées 16 with arched butt Dieskau 2 hoard, Saalkreis, Sachsen-Anhalt, Wüstemann 1995, no 137 Germany 17 Faversham (?hoard), Kent, England [13] personal study 18 Château l’Etoc hoard (2), Alderney, Channel O’Connor et al 2010 Islands 19 Sancton, East Yorkshire, England [12] Manby et al 2003, 60 fig 23.1 20 Pederstrup, Ballerup, København, Denmark Vandkilde 1996, 198 fig 193 no 579 Irish three-rivet halberds Hillswood series 21 Type Roscrea Roscrea, Co Tipperary, Ireland [21] personal study 22 No provenance (Ireland) [34] Harbison 1969a, no 183 & personal study 23 No provenance (Ireland) [35] personal study 24 near Mallow, Co Cork, Ireland [23] Harbison 1969a, no 158 & personal study 25 Type Lough Gur Hillswood hoard (2), Co Galway, Ireland [39] personal study 26 Lough Gur, Co Limerick, Ireland [37] personal study 27 Castell Coch hoard (2), South Glamorgan, Needham et al 1985, A1 Wales [2] 28 Inchigeelagh, Co Cork, Ireland [81] personal study; variant with beaded midrib 29 No provenance (Ireland) [65] Harbison 1969a, no 170 & personal study; variant with narrow midrib 30 Type Hill of Allen Hill of Allen, Co Kildare, Ireland [101] Harbison 1969a, no 216 & personal study; notches for two further rivets in shoulder angles 31 Slievecoilta Commons, Co Wexford, Ireland personal study [109] 32 Greagnafarna, Co Leitrim, Ireland [108] personal study 33 Ungrouped Co Meath, Ireland [113] Harbison 1969a, no 228 & personal study Hillswood series Corlurgan series 34 Type Tullamore, Co Offaly, Ireland [118] personal study Cloonymorris 35 No provenance (Ireland) [123] personal study 36 Type Corlurgan Corlurgan, Co Cavan, Ireland [126] personal study 37 No provenance (Ireland) [142] personal study 38 Shanmullagh or Ballycullen, River Blackwater, Ramsey et al 1991–2, 141 fig 2.3 Co Armagh, Ireland [133] British, French & Iberian three-rivet halberd types 39 Type Crossraguel Moss-side, Crossraguel, Argyll & Bute, personal study; NB. the type can have 3 or 4 Scotland [27] rivets 40 West Row Fen, Mildenhall, Suffolk, England Pendleton 1999, fig 44 no 321; blade wings are [28] heavily damaged and the restoration may be minimal 41 Type Pontrhydygroes, Yspytty Ystwyth, Ceredigion, Savory 1980, fig 29 no 100 & personal study Pontrhydygroes Wales [25] 42 Type Lambeth River Thames, Lambeth, Greater London, personal study England [22] 43 Shotton, Flintshire, Wales [23] personal study; lower blade is missing 44 Type Baútas Baútas, Mina, Amadora, Estramadura, Brandherm 2003, no 1380 Portugal 45 Cerro Benzala, Villadombardo, Jaén, Spain Brandherm 2003, no 1386 46 ‘Manzanarestal’, Madrid, Spain Brandherm 2003, no 1387 47 Finca de la Paloma hoard (1), Pantoja, Brandherm 2003, no 1381 Toledo, Spain 48 squat halberd Eysines, Gironde, France Gallay 1981, no 495; sub-trapezoid hafting-plate, cf Type Baútas 49 halberds with Paradilla, Autilla del Pino, Palencia, Spain Brandherm 2003, no 1382 stepped butts 50 Jaén area, Jaén, Spain Brandherm 2003, no 1378 51 Ibi, Alicante, Spain Brandherm 2003, no 1377 52 Type Carrapatas El Arribanzo, Fariza, Zamora, Spain Brandherm 2003, no 1362 53 Monte do Castro hoard (1), Carrapatas, Brandherm 2003, no 1364 Macedo de Cavaleiros, Trás-os-Montes e Alto Duoro, Portugal 54 ‘North-west Spain’ Brandherm 2003, no 1368 55 Avoine, Némans, Indre-et-Loire, France Cordier 2009

Needham CW3 FINAL 21 iv 16.indd 77

[ 77 ]

27/05/2016 12:46:13

[ 78 ]

ii. lost cultures of the halberd bearers

]]

Irish, British & French four-rivet halberd types with (sub-) trapezoidal hafting-plates Clonard series 56 Type Clonard near Clonard, Co Meath, Ireland [1] Harbison 1969a, no 280 & personal study 57 Derrycassan Lough, Co Cavan, Ireland [3] Harbison 1969a, no 281 & personal study 58 Hillquarter, Co Westmeath, Ireland [5] personal study 59 Type Ballygawley No provenance (Ireland) [11] Harbison 1969a, no 286 & personal study 60 Ballygawley, Co Tyrone, Ireland [9] personal study; three rivet-holes in triangle plus a fourth are thought to be from secondary hafting 61 Type Derrinsallagh ?hoard (1), Co Laoighis, Ireland personal study Derrinsallagh [14] 62 No provenance (Ireland) [19] Harbison 1969a, no 293 & personal study 63 No provenance (Ireland) [15] Harbison 1969a, no 249 & personal study Pistill Dewy/Etigny series 64 Type Pistill Pistill Dewy Hill, Towy, Llanarthney, Ó Ríordáin 1937, fig 57.7 Dewy/Etigny Carmarthenshire, Wales [17] 65 Etigny, River Yonne, Yonne, France Gallay 1981, no 502 66 Hundleton, Pembrokeshire, Wales [20] photograph supplied by A. Gwilt & personal study; variant with beaded midrib 67 Culgower Hill, Helmsdale, Highland, Scotland photograph supplied by T. Cowie [18] 68 Ris-Orangis, Essonne, France Gallay 1981, no 501 69 Saint-Denis-en-Val (2) (?hoard), Loiret, Gallay 1981, no 499 France 70 Amboise (1), River Loire, Indre-et-Loire, Gallay 1981, no 498 France 71 Vélez Blanco(?) (?hoard), Almería, Spain Brandherm 2003, no 1385 72 cf Type Pistill Foiano della Chiana, Arrezzo, Toscana, Italy Bianco Peroni 1994, no 132; unique form for Italy Dewy/Etigny Halberds with beaked shoulders 73 Type Glomel Glomel (?association), Côtes-d’Armor, France Gallay 1981, no 504 74 Armentières, Nord, France Gallay 1981, no 505 Other British four-rivet halberds (Type Sluie) & late Irish and British types 75 Type Sluie Sluie, Edenkillie, Moray, Scotland [30] personal study 76 Shropshire, England [34] personal study 77 Tonfannau Quarry hoard (1), Towyn, personal study; beaded midrib variant Llangelynin, Gwynedd, Wales [41] 78 Manea Fen, Cambridgeshire, England [40] Fox 1923, pl VI no 10; beaded midrib variant 79 Type Sluie → Culloden, Highland, Scotland [35 & 54] personal study; interpreted as having been Type Auchingoul converted into Type Auchingoul in second hafting phase (5 slender rivets) 80 cf Type Sluie La-Grande-Paroisse, Pincevent, Seine-etGallay 1981, no 525 Marne, France 81 Type Auchingoul Auchingoul, Inverkeithny, Moray, Scotland personal study [49] 82 Assich, Croy & Dalross, Highland, Scotland Coles 1968-9, 38 fig 32.1 & personal study [52] 83 Portmoak Moss, Loch Leven, Perthshire & personal study Kinross, Scotland [53] Breaghwy series 84 Type Breaghwy Breaghwy, Co Mayo, Ireland [153] personal study 85 Rough Island, Co Derry, Ireland [159] Harbison 1969a, no 304 & personal study 86 proto-Breaghwy Culmore, River Bann (near Kilrea), Co Derry, Harbison 1969a, no 307 halberd – Type Ireland [149] Kilcolagh 87 proto-Breaghwy No provenance (Ireland) [147] personal study halberd 88 ungrouped Kells, Co Meath, Ireland [151] personal study Breaghwy series 89 Ardlagheen More (or Highlake), Co Harbison 1969a, no 308 & personal study Roscommon, Ireland [150] Other late types 90 Type Eweford No provenance (Ireland) [163] personal study 91 Type Corrib Rockforest, Co Tipperary, Ireland [165] Harbison 1969a, no 310 & personal study 92 near Cavan, Co Cavan, Ireland [166] Harbison 1969a, no 309; ?unfinished halberd

Needham CW3 FINAL 21 iv 16.indd 78

27/05/2016 12:46:13

Needham

[ 79 ]

BIBLIOGRAPHY Annable, F. K. & D. D A. Simpson 1964 Guide Catalogue of the Neolithic and Bronze Age collections in Devizes Museum. Devizes, Wiltshire Archaeological and Natural History Society. Barfield, L. H. 1969 ‘Two Italian halberds and the question of the earliest European halberds’, Origini 3, 67–83. Bell, D. 2014 ‘A radiocarbon date for an Irish Bronze Age halberd’, PAST no 76, 15–16. Bianco Peroni, V. 1994 I Pugnali nell’Italia Continentale. Prähistorische Bronzefunde VI, 10. Stuttgart, Franz Steiner. Billard, C., G. Querré, L. & Salanova 1998 ‘Le phénomène campaniforme dans la basse vallée de la Seine: chronologie et relation habitatssépultures’, Bulletin de la Société Préhistorique Française 95, 351–63. Brandherm, D. 2003. Die Dolche und Stabdolche der Steinkupfer- und der älteren Bronzezeit auf der Iberischen Halbinsel, Prähistorische Bronzefunde VI, 12. Stuttgart, Franz Steiner. Brandherm, D. 2004. ‘Porteurs de hallebardes? Überlegungen zur Herkunft, Entwicklung und Funktion der bronzezeitlichen Stabklingen’, Varia Neolithica 3, 279–334 Brandherm, D. 2011 ‘Use-wear on Bronze Age halberds: the case of Iberia’, Bronze Age Warfare: Manufacture and Use of Weaponry, ed. M. Uckelmann & M. Mödlinger, 23–38. Oxford. British Archaeological Reports, International Series 2255. Briard, J., & J. Roussot-Larroque 2002 ‘Les Débuts de la Métallurgie dans la France Atlantique’, Die Anfänge des Metallurgie in der Alten Welt / The Beginnings of Metallurgy in the Old World, Forschungen zur Archäometrie und Altertumswissenschaft, Band 1, ed. M. Bartelheim, E. Pernicka, & R. Krause, 135–60. Rahden, Westfalen, Leidorf. Brindley, A. L. 2007. The Dating of Food Vessels and Urns in Ireland. Galway: National University of Ireland, Department of Archaeology; Bronze Age Studies 7 Burgess, C. 1979 ‘The background of early metalworking in Ireland and Britain’, The Origins of Metallurgy in Atlantic Europe, ed. M. Ryan, 207– 14. Dublin, Stationery Office. Butler, J. J. 1963 ‘Bronze Age connections across

Needham CW3 FINAL 21 iv 16.indd 79

the North Sea: a study in prehistoric trade and industrial relations between the British Isles, the Netherlands, north Germany and Scandinavia, c. 1700–700 BC’, Palaeohistoria 9. Cantet, J.-P. 1996 ‘L’Âge du Bronze dans le Saves et le Site du Goute à Seysses-Savès (Gers)’, Bulletin Prèhistoire du Sud-Ouest, Nouvelles Études 3, 87–90. Carlin, N. 2011 ‘A proper place for everything: the character and context of Beaker depositional practice in Ireland’ Unpublished doctoral thesis (2 vols). University College Dublin. Coffey, G. 1908–9 ‘Irish copper halberds’, Proceedings of the Royal Irish Academy 27C, 94–114. Coles, J. M. 1968–9 ‘Scottish Early Bronze Age metalwork’, Proceedings of the Society of Antiquaries of Scotland 101, 1–110. Cordier, G. 2009 ‘Une nouvelle hallebarde du Bronze ancien sur la Loire (Avoine, Indre-et-Loire)’, Bulletin Préhistoire du Sud-Ouest 17, 243–5. Despriée, J. 1983 ‘Circonscription du Centre’, Gallia Préhistoire 26, 249–81. Fox, C. 1923 The Archaeology of the Cambridge Region. Cambridge, Cambridge University Press. Fox, C. 1959 Life and Death in the Bronze Age. London, Routledge & Kegan Paul. Gadbois-Langevin, R. 2013 ‘Le Campaniforme en France: Études spatiale de l’évolution d’un territoire. Un ensemble divisé?’ Unpublished Mémoire de Master 2 (2 vols), Dijon, Université de Bourgogne/ Gallay, G. 1981 Die Kupfer- und Altbronzezeitlichen Dolche und Stabdolche in Frankreich. Prähistorische Bronzefunde VI, 5. Munich, Beck’sche. Garrido Pena, R. 1997 Bell Beakers in the southern Meseta of the Iberian Peninsula: socioeconomic context and new data. Oxford Journal of Archaeology 16, 187–207. Garrido Pena, R. 1999 El Campaniforme en la Meseta: Análisis de su contexto social, económico y ritual. Madrid, Universidad Complutense, Departamento de Prehistoria. Grinsell, L.V. 1968 Guide Catalogue to the South Western British Prehistoric Collections. Bristol, City Museum Harbison, P. 1969a The Daggers and Halberds of the Early Bronze Age in Ireland, Prähistorische Bronzefunde VI, 1. Munich, Beck’sche. Harbison, P. 1969b ‘Catalogue of Irish Early Bronze

27/05/2016 12:46:13

[ 80 ]

ii. lost cultures of the halberd bearers

Age associated finds containing copper or bronze’, Proceedings of the Royal Irish Academy 67C, 35–91. Harrison, R. J. 1977 The Bell Beaker Cultures of Spain and Portugal, Harvard University, Peabody Museum of Archaeology and Ethnology, Bulletin 35. Cambridge, Massachusetts. Harrison, R.J. 1980. The Beaker Folk: Copper Age Archaeology in Western Europe. London, Thames & Hudson. Harrison, R. J. & A. Mederos Martín 2001 ‘Bell Beakers and social complexity in central Spain’, Bell Beakrs Today: Pottery, People, Culture, Symbols in Prheistoric Europe, 111–24. Provinzia Autonoma di Trento, Servizio Beni Culturali, Ufficio Beni Archeologici. Horn, C. 2014 Studien zu den Europäischen Stabdolche, Universitätsforschungen zur Prähistorischen Archäologie, Band 246 (2 vols). Berlin, Institut für Prähistorische Archäologie der Freien Universität. Kinnes, I. A. 1985 Beaker and Early Bronze Age Grave Groups, British Bronze Age Metalwork, Associated Finds Series, A7–16. London, British Museum Publications. Kinnes, I. A. 1994 Beaker and Early Bronze Age Grave Groups, British Bronze Age Metalwork, Associated Finds Series, A17–30. London, British Museum Publication.s Labaune, M. 2010 Le Métal et la Métallurgie Campaniforme en France au IIIe millénaire av. n. è. Unpublished Mémoire de Master 2 (2 vols), Dijon, Université de Bourgogne/ Le Provost, F., P. R. Giot, & Y. Onnee 1972 ‘Prospections sur les collines de SaintNicolas-du-Pélem (Côtes d’Armor) du Chalcolithique à la Protohistoire’, Annales de Bretagne 79, 39–48. Le Roux, C. T. 1985 ‘La hallebarde de Gavrinis’, Paléometallurgie de la France Atlantique; Âge du Bronze (2), Travaux du Laboratoire Anthropologie-Préhistoire-ProtohistoireQuarternaire Armoricains, ed. J. Briard, 105–6. Rennes. Manby, T. G., A. King, & B. E. Vyner 2003 ‘The Neolithic and Bronze Age: a time of early agriculture’, The Archaeology of Yorkshire: an assessment at the beginning of the 21st century, Yorkshire Archaeological Society Occasional

Needham CW3 FINAL 21 iv 16.indd 80

]]

Paper no. 3, ed. T. G. Manby, S. Moorhouse & P. Ottaway, 35–116. York. Manfredini A., M. A. Fugazzola Delpino, L. Sarti, M. Silvestrini, F. Martini, C. Conati Barbaro, I. M. Muntoni, G. Pizziolo, & N. Volante 2009 ‘Adriatico e Tirreno a confronto: analisi dell’occupazione territoriale tra il Neolitico finale e l’età del Rame in alcune aree campione dell’Italia centrale’, Rivista di Scienze Preistoriche 59, 115–80 Mataloto, R., J. M. Matos Martins, & A.M. Monge Soares 2013 ‘Cronologia absoluta para o Bronze do Sudoeste. Periodizacao, base de dados, tratamento estatistico’, Estudos Arqueologicos de Oeiras 20, 303–38 Mohen, J.-P. 1977 L’Âge du Bronze dans la Région de Paris. Paris, Musées Nationaux. Mount, C. 2012 A note on some Beaker period pit burials in Ireland. Journal of Irish Archaeology 21, 1–6. Needham, S. 1996 ‘Chronology and periodisation in the British Bronze Age’, Absolute Chronology: Archaeological Europe 2500–500 BC. Acta Archaeologica 67,ed. K. Randsborg, 121–40. Needham, S. 2005 ‘Transforming Beaker culture in north-west Europe; processes of fusion and fission’, Proceedings of the Prehistoric Society 71, 171–217. Needham, S. 2012 ‘Case and place for the British Chalcolithic’, Is there a British Chalcolithic? People, Place and Polity in the Later Third Millennium, Prehistoric Society Research Paper 4, ed. M. J. Allen, J. Gardiner, & A. Sheridan, 1–26. Oxford, Oxbow. Needham, S. (with contributions by Mary Davis, Adam Gwilt, Mark Lodwick, Phil Parkes, & Peter Reavill) 2015a ‘A hafted halberd excavated at Trecastell, Powys: from undercurrent to uptake— the emergence and contextualisation of halberds in Wales and north-west Europe’, Proceedings of the Prehistoric Society 81. Needham, S. (with contributions from M. Davis, R. Gale, & P. Parkes) 2015b ‘The halberd and its haft remains; Trecastell and the halberds of Britain, Ireland and the near Continent’. http://www. cofiadurcahcymru.org.uk/ Needham, S., Gwilt, A., Lodwick, M., Northover, P. and Reavill, P. 2015 ‘Halberds and early metalwork in Wales and the Marches’ http://www. cofiadurcahcymru.org.uk/ Needham, S. P., Lawson, A. J. and Green, H. S. 1985 Early Bronze Age Hoards, British Bronze Age

27/05/2016 12:46:13

Needham

Metalwork, Associated Finds Series A1‑6. London, British Museum Publications. Nicolas, C. 2013 ‘Symboles de Pouvoir au Temps de Stonehenge: les Productions d’Armatures de Prestige de la Bretagne au Danemark (2500–1700 av. J.-C.)’, Unpublished doctoral thesis (2 vols);, Université Paris I—Panthéon Sorbonne. O’Connor, B., T. Cowie, & C. Horn, 2010 ‘Une trouvaille oubliée de l’ancien duché de Normandie et sa redécourverte’, Bulletin de l’Association pour la Promotion des Recherches sur l’Âge du Bronze 7, 4–5. O’Flaherty, R. 2002 ‘A Consideration of the Early Bronze Age Halberd in Ireland. Unpublished doctoral thesis’, University College Dublin. O’Flaherty, R. 2007 ‘A weapon of choice— experiments with a replica Irish Early Bronze Age halberd. Antiquity 81, 423–34 Ó Ríordáin, S. 1937 ‘The halberd in prehistoric Europe: a study in prehistoric origins, evolution, distribution, and chronology’, Archaeologia 86, 195–321. Pendleton, C. F. 1999 Bronze Age Metalwork in Northern East Anglia: a Study of its Distribution and Interpretation, British Archaeological Reports, British Series 279. Oxford. Ramsey, G., C, Bourke, & D. Crone 1991–2 ‘Antiquities from the River Blackwater: I, Bronze Age metalwork’, Ulster Journal of Archaeology 54–5, 138–49. Rojo-Guerra, M. A., R. Garrido-Pena, & I. GarciaMartinez de Lagran, (eds) 2005 El Campaniforme en la Peninsula Iberica y su contexto Europeo. Universidad de Valladolid Russel, A. D. 1990 ‘Two Beaker burials from Chilbolton, Hampshire’, Proceedings of the Prehistoric Society 56, 153–72 Salanova, L. 2000 La Question du Campaniforme en France et dans les Îles Anglo-Normandes: production,

Needham CW3 FINAL 21 iv 16.indd 81

[ 81 ]

chronologie et rôles d’un standard céramique. Paris, Société Préhistorique Française, Comité des Travaux Historiques et Scientifiques. Savory, H. N. 1980 Guide Catalogue of the Bronze Age Collections. Cardiff, National Museum of Wales. Senna-Martínez, J. C. 1994 ‘Subsídios para o estudo do Bronze Pleno na Estremadura atlântica: a alabarda de tipo ‘Atlântico’ do habitat das Baútas (Amadora)’m Zephyrus 46, 161–82 Senna-Martínez, J. C. 2013 ‘Metals, technique and society. The Iberian Peninsula between the first peasant societies with metallurgy and the ‘urban revolution’ A Ouriversaria Pré-Histórica do Ocidente Penisnular Atlántico. Compreender para Preservar, ed. M. F. Guerra & I. Tissot, 11–20. Lisbon, Projecto AuCORRE. Sheridan. A. 2007 ‘Dating the Scottish Bronze Age: “There is clearly much that the material can still tell us”’, Beyond Stonehenge: Essays on the Bronze Age in Honour of Colin Burgess, ed. C. Burgess, P. Topping & F. Lynch, 162–85. Oxford, Oxbow Books. Thévenot, J.-P. 1985 ‘Circonscription de Bourgogne’, Gallia Préhistoire 28, 171–210. Vandkilde, H. 1996 From Stone to Bronze: the Metalwork of the Late Neolithic and Earliest Bronze Age in Denmark. Aarhus, Jutland Archaeological Society. Waddell, J. 1990 The Bronze Age Burials of Ireland. Galway, Galway University Press. Waddell, J. 1998 The Prehistoric Archaeology of Ireland. Galway, Galway University Press. Woodward, A., & J. Hunter, J. 2011 An Examination of Prehistoric Stone Bracers from Britain. Oxford, Oxbow Books. Wüstemann, H. 1995 Die Dolche und Stabdolche in Ostdeutschland. Prähistorische Bronzefunde VI, 8. Stuttgart, Franz Steiner.

27/05/2016 12:46:13

This page intentionally left blank

chapter three

Closed for business or cultural change? Tracing the re-use and final blocking of megalithic tombs during the Beaker period Catriona D. Gibson

Introduction: Megaliths and Beakers

M

monuments and the Bell Beaker complex form two of the major defining phenomena of later prehistoric Europe, yet in the past they have often been studied in separate research spheres. This chapter will explore the suggestion that aspects of these traditions were inter-related, and a better appreciation of both can result if some of their elements are considered together. In Atlantic Europe, the construction of megalithic monuments dates predominantly from the 5th millennium BC onwards, although a few early 6th millennium examples exist (e.g. Bueno, Barroso & Balbín 2007; Rojo & Garrido 2012). They increase in number, size and complexity throughout the 4th and 3rd millennia (Delibes, Alonso & Rojo 1987, 183–5; Bueno, Barroso & Balbín 2004, 89–94; García Sanjuán 2006, 155), with construction of new monuments declining between c. 3400–3000 BC. This process is assumed to be either a symptom or a result of other changes, paving the way for new burial practices and associated social and ideological transformations.   The significance of changes in mortuary practices from predominantly collective funerary rites within megalithic monuments to single burials, accompanied by their own personal grave goods, is still hotly debated (Thomas 2000, 653–4), and has been since the later 19th century (e.g. Fergusson 1872; Thurnam 1872). Thus, the emergence and floruit of the Bell Beaker phenomenon during the 3rd millennium BC continues to be perceived as one of the most intriguing cultural complexes in prehistoric Europe, due to the widespread adoption of Beaker traditions that seemingly signal fundamental social and cultural changes (e.g. Barrett 1988; 1990; 1994; Fokkens & Nicolis 2012). The innovations associated with the Beaker package have encouraged archaeologists to pay more consideration to the elements of transformation that accompany it, such as the adoption of pioneering technologies including metal and new ideologies. Yet by concentrating on distinctions with the preceding period, important elements of over-riding continuity, perhaps even

Gibson_CW3.indd 83

egalithic

27/05/2016 13:18:47

[ 84 ]

iii. Closed for business or cultural change?

]]

conservatism, or deliberate resistance to change, have been unwittingly neglected. It is one of these aspects—the persistence of the megalithic tradition—that this chapter will focus on. Redressing the dichotomies The transition from the Late Neolithic to the Chalcolithic/Beaker period in Europe can no longer be neatly tied to the abandonment of collective funerary rites and the switch to individual burial. Collective or multiple burial continued into the Chalcolithic and beyond (e.g. Boaventura 2009; Woodward 2000). Examples include a recently excavated grave within a barrow in Amesbury, near Stonehenge in Wiltshire, which contained the remains of a minimum of six predominantly disarticulated individuals accompanied by Beakers; known as the ‘Boscombe Bowmen’ (Fitzpatrick 2011, 8–13). Seven articulated and a number of disarticulated burials were placed in the central grave of a round barrow at South Dumpton Down in Kent, along with a Beaker (Perkins 2004). Multiple Beaker burials are also known from other parts of Britain, such as Bee Low, in Derbyshire (Marsden 1970), and Broomend of Crichie, in Aberdeenshire (Sheridan 2007). Numerous examples are also known from Iberia, and several of these are outlined later in this chapter.   Furthermore individual burial was already being practiced during the Neolithic in parts of Atlantic Europe, including Poitou and Vendée in France; Yorkshire, Oxfordshire, and Hampshire in England; the Alentejo and Algarve in Portugal; and Cantabria in Spain. In England, Neolithic single burials, sometimes accompanied by individual grave goods including maceheads, axes and knives, were recovered from the oval or long barrows of Alfriton in Sussex (Drewett 1975), Duggleby Howe in Yorkshire (Loveday 2002), Hazelton in Gloucestershire (Saville 1990), and Radley in Oxfordshire (Barclay & Halpin 1998). The earlier Neolithic (4500–3800 cal BC) cemetery of Castelo Belinho, in the Algarve, comprised a series of individual burials (Gomes 2012). The mid-4th millennium BC pit burials from the cemeteries of La Garma and Urtao II, both in Cantabria, comprised single flexed interments, accompanied by personal grave goods such as pots and flint arrowheads (Arias 2012, 16). Thus, during the 3rd millennium BC, the changes that occurred throughout Europe were neither so clear-cut, sudden, nor so all-pervasive as once perceived (e.g. Gibson 2007, 47–9). The megalithic stone tradition ‘Megalithic monument’ is an umbrella term that covers a diverse range of monumental architecture, with variation due to both regional and chronological factors (Table 3.1). The sets of radiocarbon dates generated for many of these monuments throughout Atlantic Europe demonstrate that they are not indicative of a short-lived tradition, but rather that they were built, used and re-used over an extensive time-frame, predominantly between c. 5000–2000 BC (e.g. Cardoso & Soares 1990–92; Cardoso 1997; Cardoso, Carlos & Henrique 2000, Gonçalves 2003; 2005; 2008; Senna-Martíne & Ventura 2008; Cardoso 2014; Cassen 2014;). Even perhaps more significant is the realization that these monuments

Gibson_CW3.indd 84

27/05/2016 13:18:47

[ 85 ]

Gibson

were not constructed in a single event, and remained unchanged from that moment on, but that their architectural forms were subject to manipulation, modification and re-adaptation at distinct times. Many of these monuments had complex biographies and the episodic cycles of use, re-activation, augmentation, abandonment and final closure witnessed are indicative of the changing roles they performed (Scarre 2010, 17–20; Thomas 2012, 64–6; Boaventura & Mataloto 2013, 84–8, 97–8; Aranda & Lozano 2014, 372–6). Iberia

France

Chambered Allées tomb couverte Gallery grave Angled grave Hypogaeum/ rock cut tomb Megalithic cist Passage tomb Portal dolmen Rock-cut tomb Tholos

Gallery grave

Passage grave

Ireland Court cairn Megalithic cist Passage grave

Western Southern Britain/Scotland Britain Clyde tomb Chambered tomb Cotswold Severn Earthen tomb long barrow Court cairn

Portal tomb Passage grave Wedge tomb

Portal dolmen

Table 3.1 The principal regional traditions of megalithic architecture in Atlantic Europe

  In the past, the revisitation of earlier monuments, as evidenced through modification, displacement of bodies, or re-use for later burial, was often seen to represent acts of robbing or deliberate destruction. In particular, re-use of megalithic monuments in the Beaker period has been described as an aggressive activity, and interpreted as intrusion or violation by Beaker groups. The presence of Beaker material implied rupture, thus representing a break in the use of the tomb, and signifying that control of rites to ancestry and territory had now been claimed by an exogenous political group (Gallay 1976; L’Helgouach 1976; Andrés 2000, 66; Weiss-Krejci 2005). These interpretations are now being challenged (e.g. Salanova 2003, 386; 2007). Indeed, more often than not, the evidence for re-use does not imply destruction or obliteration, but rather respect and consideration for the earlier monuments and the individuals housed within them. In some cases, new radiocarbon dates indicate that, where archaeologists previously perceived rupture, elements of continuity with existing practices may be witnessed. This realization will be outlined in more detail below, supported by a number of case-studies.

Gibson_CW3.indd 85

27/05/2016 13:18:47

[ 86 ]

iii. Closed for business or cultural change?

]]

Beaker burial traditions In Britain, early Beaker burial (c. 2420–2300 cal BC; Barclay, Marshall & Higham 2011, 181, table 32) tends to be characterized by single crouched or flexed interments in flat graves or cists, or in pits under barrows/cairns, with individual grave goods. The accompanying artefacts may range from a single Beaker pot (e.g. Donside Field in Aberdeenshire (Shepherd 1989, 89) and Drem Farm in Lothian (Smith 1882, 299–300)), to more richly furnished grave good sets that include arrowheads, bracers, knives, daggers and ornaments alongside the Beaker vessel (e.g. the Amesbury Archer in Wiltshire (Fitzpatrick 2011, 67– 71), Driffield in North Humberside (Kinnes 1985, 19–22), and Ferry Fryston in Yorkshire (Brown et al. 2007, 30)). There are of course exceptions to this rule, some of which have been summarized above.   In western France and western Iberia, however, evidence for interment in single graves is markedly rare, at least during the early Beaker period (c. 2800–2400/2350 BC). The majority of burials accompanied by Beaker pots and associated Beaker-related grave sets come from megalithic tombs, although they were sometimes interred in segregated spaces/ niches within them. These were returned to again and again for successive acts of burial, and associated funerary or ritual activities, suggesting continuity with later Neolithic practices. As a result, many of these monuments have complicated and confusing stratigraphies, often further exacerbated by later disturbance. The more prominent megalithic monuments have suffered from unwarranted attention, including tomb robbing and unrecorded antiquarian interventions. Thus the presence of subsequent Beaker material has not always been appreciated, and its potential significance has been neglected.   This is an opportune time to re-explore these connections with more conviction, since a large number of excavations at megalithic sites have occurred over the last two decades, especially in Iberia, where the bulk of this chapter will focus. The ability to trace the complex narratives for these monuments has been vastly enriched by the generation of many new wellcontextualized radiocarbon determinations, including multiple AMS dates for individual sites. The re-use of older tombs during Beaker and later periods had been seriously underestimated until these dates became available. These chronological sequences now provide a clearer appreciation of the history and evolution of individual monuments, enabling specific phases of use, architectural elaboration and re-use to be charted more accurately1. An attempt to tie these rhythms of episodic use and closure with broader sets of regional and supra-regional developments occurring throughout Atlantic Europe can now be made.   Certain patterns began to emerge during the data collation for evidence of re-use of megalithic monuments during the Beaker period, and these can be grouped under three main themes (Figure 3.1). The first trend noted is that Beaker burials are often placed in peripheral parts of monuments or in new structures added during subsequent architectural elaboration. The second pattern identified is that, following Beaker re-use, some megalithic monuments were formally closed either through a deliberate blocking of the entrances or 1 All radiocarbon dates cited in the text are fully referenced with laboratory numbers and raw dates in Appendix; also see Figures 3.10 & 3.11 for some of the specific sequences. All 14C dates are calibrated in OxCal 4.2.4 (Bronk Ramsey 2013; Reimer et al. 2013) and presented at 2 sigma (95.4%).

Gibson_CW3.indd 86

27/05/2016 13:18:47

[ 87 ]

Gibson

a covering of the monuments with mounds, or both. Thirdly, megalithic monuments with Beaker burials are not ubiquitous throughout Atlantic Europe but form distinct regional zones, and many of the new dates indicate synchronicity in their final closure.

Beaker burial in peripheral part of monument or in additional chamber Additional chamber

1a

Blocking and covering with mounds

Location of megaliths with Beaker re-use Specific geographic distributions

Blocking

Covering with mound

Beaker burial almost immediately followed by deliberate closure, sealing or covering of monuments 1b

1c

3.1 The three themes discussed regarding the re-use of megalithic monuments during the Beaker period

Gibson_CW3.indd 87

27/05/2016 13:18:49

[ 88 ]

iii. Closed for business or cultural change?

]]

Re-use of peripheral zones and restructuring of monuments in Iberia Spain When megalithic tombs were restructured during the Beaker period, later burials or sets of activity often occurred in the peripheral zones of the monuments, or in new structures added as extensions. The majority of examples where this has been documented come from Iberia, and include the Late Neolithic chambered tomb of La Mina in Alava (Navarte 2007, 158). Its entrance was later remodelled, through the building of a cist and antechamber attached to the original polygonal chamber (Marqués de Loriana 1953, 191). Maritime Beaker sherds and a fragmentary bronze blade were found in both these structures, although they were not mentioned in the original publication, and only noted later (Navarte 2007, 150). Beaker architectural elaboration was also identified at the corridor tomb of Tres Montes, in Navarre. Here a cist was added to the front of the corridor in the earlier 3rd millennium BC (2814–2675 cal BC), containing crouched interments accompanied by Maritime Beakers and arrowheads (Andrés, García & Sesma 2001, 315). The burials were subsequently heataffected during an intensive burning event that occurred shortly afterwards (2704–2555 cal BC), marking the tomb’s closure (Andrés, García & Sesma 2002, 198–200).   A similar fire ritual was noted at the dolmen of La Peña de La Abuela in Soria, after the deposition of several interments between 3800–3700 cal BC (Rojo, Garrido & García 2010, 254–5). Nearly 1500 years later, the dolmen was covered by a large mound and several Bell Beaker inhumations were inserted within it (ibid., 256–7). A stela was also placed on top of the mound at this time (Rojo et al. 2005). The nearby complex tomb of La Sima was subject to three separate periods of funerary activity (Rojo, Kunst & Palomino 2002). The Neolithic phase comprised several burials placed in a lime-kiln tomb that was completely razed to the ground between 3700–3600 cal BC (Rojo, Morán & Kunst 2003). A new tomb was constructed directly on top of the burnt remains of the earlier structure, and this was further elaborated through the addition of a stone corridor around 2400 cal BC. Several inhumations accompanied by Beaker pots and metal objects were buried in this corridor, and the previous stone mound surrounding the tomb was also rebuilt at this time (ibid., 259–60).   The collective tumulus of La Atalayuela, in La Rioja, contained over 70 inhumations, predominantly articulated. They were buried successively in five separate layers, and care was seemingly taken so that later interments did not disturb any of the earlier remains (Barandiarán 1973, 96–7; Marcos 1973, 44–9). The latest burials dated to the earlier 3rd millennium BC (2878–2564 cal BC and 2765–2470 cal BC), and were accompanied by broken Beaker vessels, arrowheads and V-perforated buttons (Harrison 1988, 465; Andrés & Barandiarán 2004, 94, 101). At the Cueva de El Vaquero hypogaeum2, Alcalá de Guadaíra, in Seville, three Beaker burials were inserted in the later tholos3 built against the monument’s entrance; a fourth was placed in the earth mound that subsequently covered it (Lazarich & Sánchez 2000, 331). They were dated to the mid-3rd millennium BC (García-Sanjuán 2005, 2 A hypogaeum is an underground burial chamber 3 A tholos is a circular tomb with a high, often beehive-shaped, corbelled roof. These drystone monuments have narrow entrances and a long corridor, and tend to be partially excavated into the bedrock.

Gibson_CW3.indd 88

27/05/2016 13:18:50

Gibson

[ 89 ]

90), and the monument was abandoned immediately after this. Nearby, re-use of the tholos of El Gandul, also in Alcalá de Guadaira, is indicated by the burial of four inhumations accompanied by Beaker grave goods, again of mid-3rd millennium BC date (Pellicer & Hurtado 1987). Two of these were placed towards the end of the corridor, and the other two were buried directly in front of the entrance (Hurtado & Amores 1984, 156–7).   Another Sevillian example is the dolmen of Montelirio, which forms part of the large megalithic cemetery adjacent to the Chalcolithic settlement of Valencina de la Concepción. Recent excavations were complemented by an extensive suite of radiocarbon dates, allowing the identification of two distinct phases of activity (Fernández & Aycart 2013; Mederos 2013). In the early 3rd millennium BC, 18 articulated or partially articulated individuals were interred in the large chamber, mainly adult females. A peripheral antechamber was constructed slightly later, and two crouched, fully articulated individuals were buried inside it. They were of an adult male and adult female, and were accompanied by a panoply of grave goods that included ivory combs, a flint dagger with amber pommel, a gold strip and fragments of north African ostrich egg shell. Portugal: Tagus estuary Recent dating of several of the Neolithic megalithic monuments around the Tagus estuary, in Baixa Extremadura, has demonstrated that many of these were subject to punctuated episodes of re-use during the early to mid-3rd millennium BC (Boaventura 2009). The anta4 of Estria contained up to 60 individuals (Hillier 2008a; Ribeiro 1871–75; 1880), with associated grave goods including schist plaques, idols, a lunula, and a possible halberd. The latest burials, dated between 2850 and 2570 cal BC, were related to the alteration of the tomb and a rebuilding of its entrance on a different orientation (Gonçalves 1999; Hoskin 2001).   At the nearby hypogaeum of Verdelha dos Ruivos, a number of semi-articulated and disarticulated burials were interred during the earlier 3rd millennium BC (2880–2560 cal BC and 2620–2460 cal BC). Several complete Beaker pots were placed near the tomb’s entrance between 2580–2340 cal BC (Cardoso & Soares 1990–92; Leitão et al. 1984). No burials were associated with this deposition of vessels and it may be that it was not of a strictly funerary nature.   The tholos tomb of Carrascal, near Lisbon, was constructed and used during the early to mid-4th millennium BC (3650–3350 cal BC; Boaventura 2009, 73; Hillier 2008b). In the mid-3rd millennium BC, the tomb was re-activated, and the final disarticulated interments were placed towards the end of the corridor, accompanied by Beaker pots and a bracer (Ribeiro 1880; Hillier 2008b; Boaventura 2009, 77). Similarly, at the dolmen of Trigache 2, also in Lisbon, later re-use of the Neolithic monument in the mid-3rd millennium BC is represented by disarticulated human remains in the entrance area, accompanied by copper beads, stone axes and fragments of several Beaker vessels (Boaventura 2009, 88). The adjacent dolmen of Trigache 3 also witnessed re-activation in the Beaker period—here two Beaker vessels and a copper pin were deposited within the blocking material at the 4 An anta is the Portuguese term for a dolmen. It literally describes the pillars on either side of the entrance.

Gibson_CW3.indd 89

27/05/2016 13:18:50

[ 90 ]

iii. Closed for business or cultural change?

]]

front of the tomb (Leisner & Ferreira 1961; Leisner 1965, 18). Two complete Beaker pots were also placed upright in the blocking deposits at the entrance of the anta of Herdade das Casas do Canal, in Estremoz (Salanova 2003, fig. 4) (Figure 3.2).   A large number of articulated inhumation burials were found in the chamber of the dolmen of Pedras da Granja, in Sintra, predominantly of late 4th millennium BC date (Zbyszweski et al., 1977, 224, figs. 3, 6 & 7). Most of these bodies were in situ and the latest burials, dating to 2700–2470 cal BC, were placed in the upper layers of the monument (Boaventura & Cardoso 2011, 147). As at La Atatuela, this seems to indicate continued respect for the dead, as care was taken not to disturb the previous interments.

N

Key 0

2m

Orthostats Blocking material

3.2 Complete Beaker pots placed in the blocking deposits at the entrance of Herdade das Casas do Canal, Estremoz, Portugal (Source: redrawn after Salanova 2003, fig 4)

Gibson_CW3.indd 90

27/05/2016 13:18:50

Gibson

[ 91 ]

Other Atlantic regions In some Atlantic regions, later funerary activity involved an element of clearance of existing deposits within the monument, albeit in a seemingly reverential manner. In the passage grave of Le Déhus, Guernsey, for instance, three inhumations accompanied by three Beakers and a tanged copper dagger were buried in the main chamber, but only after the Middle Neolithic funerary remains had been carefully moved and sealed off in side chambers (Kendrick 1928, 143). Subsequently, the chamber was infilled and the entrance sealed with a thick layer of limpet shells. Discussion There are numerous other examples of megalithic monuments with evidence of re-use during Beaker times in Iberia and other Atlantic regions. In many cases they have only been identified recently, as a result of new radiocarbon dates being made available. In all the examples cited, however, the later episodes of activity do not suggest deliberate violation or destruction to the monument, as has been inferred in the past. As the tombs filled up with bones over successive episodes of use, rather than clearing these human remains out to make space for the new, extensions were often built on to the existing monument or Beaker interments were squeezed into the peripheral areas in the corridors, or close to the entrances. When this did not occur, as at Le Déhus, La Atalayuela, or Pedras da Granja, the previous burials were carefully moved into the adjacent chambers, or thick layers of soil were used to cover earlier interments and prevent disturbance by new insertions. None of this indicates a deliberate attempt by new groups to stake their own claims and legitimacy on to existing monuments, through erasing earlier memories and destroying those already buried there. Rather it suggests that alternative narratives should be put forward to explain why there was evident respect for Neolithic interments, and furthermore, continuity in burial traditions and associated practices, at many of these monuments. Closure, blocking and abandonment France New radiocarbon schemes provide the chronological detail necessary to chart episodes of re-use, and subsequent, often quite dramatic abandonment processes that megalithic monuments were subject to during the Beaker period. In France, there is abundant evidence for deliberate closure during the 3rd millennium BC, and in the majority of cases, it was focused on gallery graves (Masset 1992; 1998). Virtually all the Armorican megaliths were abandoned during this time, and this was generally undertaken through sealing the monuments. Closure was accompanied by final funerary or ritual deposits such as a complete Bell Beaker from Kerlescan in Carnac and a bracer from Lestrigou in Finistère (L’Helgouach 1965).   Some of the Breton and Paris Basin gallery graves were covered by sterile layers of rubble or soil, as at Chaussée-Tirancourt in the Somme, where the entrance was blocked and then set on fire towards the end of the Beaker period (Leclerc & Masset 1980).

Gibson_CW3.indd 91

27/05/2016 13:18:50

[ 92 ]

iii. Closed for business or cultural change?

]]

N

Chaussée-Tirancourt is one of the best-conserved megalithic tombs in the Paris Basin, with a detailed stratigraphic sequence complemented by well-preserved human remains (Leclerc & Masset 2006; Sohn 2006). A minimum of 300 individuals were buried within this monument, spanning an extensive time-frame from c. 3400–2200 cal BC, and associated with three distinct phases of funerary activity (Leclerc & Masset 2006, 86). During the first phase (c. 3400–2850 cal BC), the dead were placed in the centre of the burial chamber, predominantly in a disarticulated state, whereas in the second phase (c. 2850–2400 cal BC) they were interred as articulated skeletons within individual boxes or cists. Most grave goods from this phase—including broken pots, antler picks, bone awls and copper beads—were focused near the entrance to the tomb. The final phase (c. 2400–2200 cal BC) comprised a series of disarticulated human remains deposited in an ossuary layer by the entrance. Following this, the interior of the tomb was infilled with soil, and the monument was then capped by a large mound, accompanied by a series of fires signalling its final closure (Sohn 2006, 118). The combined stratigraphic and radiocarbon sequence suggest this mortuary activity was not continuous but occurred at distinct punctuated episodes.   Similarly, the nearby gallery grave of Méréaucourt, also in the Somme, was active for a lengthy time-frame (Guy & Masset 1991; 1995). The majority of burials interred were already in a partially or fully defleshed condition; for instance, there was a marked under-representation of the smallest bones (Masset et al. 2013, 8). The recently published radiocarbon dates indicate that the monument was used for c. 600–800 years between 3300/3100–2500 cal BC, before being formally closed (ibid., 78). The original capstones were dismantled and used to block the entrance. This was described by the excavators as a stone ‘cork’ (Guy & Masset 1995, 266), as holes noted on the orthostats may have once held wooden pegs to keep the blocking in place and support it from behind (ibid., 267). Therefore the final location of these capstones Blocking and plug (each weighing over 800 kg) did not result re-using capstones from natural collapse; they were inten­tionally repositioned (Figure 3.3). A human tibia found under one of these drilled orthostats dated to 2450–2200 cal BC, implying that the entrance was blocked around or just after this time. Like Chaussée-Tirancourt, the tomb was covered by a huge mound of soil around 2200–2100 cal BC, transforming it into a tumulus (Masset et al. 2013, 75). Key Orthostats Capstones (some re-used as blocking material) Mound

0

Gibson_CW3.indd 92

3.3 Re-use of orthostats to block the entrance to the gallery grave of Méréaucourt, Somme, France (Source: redrawn and adapted after Guy & Masset 1995, fig. 2)

2m

27/05/2016 13:18:50

[ 93 ]

Gibson

  Three contemporaneous passage graves at L’île Cairn, in Finistère, were used for burial during the early 3rd millennium BC (3030–2425 cal BC). They were all positioned next to each other, and after 2400 cal BC, the passage to the central tomb was closed with rubble and a massive trapezoidal cairn was constructed over all three monuments, thus blocking further access into them. This was originally described as an act of condemnation (Giot, L’Helgouach & Monnier 1979, 169; Figure 3.4). The four-chambered angled grave of Goërem, in Morbihan, was built in the later 4th millennium BC (3360–2920 cal BC; L’Helgouach 1976). After a possible lengthy break in burial activity (unfortunately human bone did not survive in the acidic soil), two Beaker pots were placed alongside a copper awl, two barbed and tanged arrowheads, and two gold fasteners in the end chamber, probably as grave goods (Figure 3.5). These marked the final deposits within this tomb; it was subsequently infilled with a thick layer of sterile marine sand, and abandoned (Patton 1993, 161). N Massive blocking cairn

3.4 Blocking and covering of the passage graves at L’île Cairn, Finistère, France

(Source: redrawn after Giot, L’Helgouach & Monnier 1979, fig. 2 and Patton 1993, fig. 7.3)

Blocking deposits

2m

0

0

5m

Key Beaker pots and arrowheads Mound N

3.5 Beaker re-use of the angled grave of Goërem, Morhiban, France (Source: redrawn and adapted after L’Helgouach 1976, fig. 3)

0

Gibson_CW3.indd 93

5m

27/05/2016 13:18:51

[ 94 ]

iii. Closed for business or cultural change?

]]

Portugal Many of the megalithic monuments in the Iberian Peninsula were also formally closed during the early to mid-3rd millennium BC, particularly in the Tagus estuary, and these include sites already mentioned in the previous section. After 2600 cal BC, the entrance to the tomb of Estria was blocked (Gonçalves 1999; Hoskin 2001). Similarly, after the deposition of Beaker burials in the nearby anta of Trigache 2, the entrance to the corridor was dismantled and immediately blocked (Ferreira 1963). In the early to mid-3rd millennium BC (2900–2630 cal BC and 2840–2460 cal BC) a series of disarticulated burials were placed closed to the entrance at the dolmen of Monte Abraão, in Sintra (Figure 3.6). These were accompanied by broken Beaker pots, schist plaques, barbed-and-tanged arrowheads, flint daggers and V-perforated buttons (Boaventura 2009, 58, 64). After these internments, around 2600–2400 cal BC, the entrance to the monument was blocked with large quantities of rubble.

3.6 Final deposition of disarticulated Beaker burials in dolmen of Monte Abraão, Sintra, Portugal, followed by final closure

N 0

5m

(Source: redrawn after V. Leisner 1965, fig. 36)

Key Orthostats Blocking material Disarticulated burials

S

N 0

1m

  The corridor tomb of Conchadas, in the Tagus estuary, was also subject to deliberate closure, and this too was originally described as an act of violation (Leisner & Ferreira 1961, 314). Prior to the blocking, in the mid-3rd millennium BC, an adult male was buried in the vestibule of the tomb, accompanied by two Maritime Beakers and seven V-perforated bone buttons (Boaventura 2009, 207). Part of the burial lay under the blocking stones, which re-used several of the original orthostats and capstones. Re-excavation of the nearby chambered tomb of Pedras Grandes has provided new detail concerning its final

Gibson_CW3.indd 94

27/05/2016 13:18:51

Gibson

[ 95 ]

phase of use (Era-Arqueologia 2001), missed during earlier excavations. After use in the later 4th millennium BC (3370–3090 cal BC), the entrance orthostats were deliberately toppled to close the chamber, and a drystone wall was constructed in front of the entrance (Boaventura 2009, 122). In the mid-3rd millennium BC, a large tumulus was erected to cover the stone framework of the tomb (ibid., 124). Likewise, after an extension to the corridor of the dolmen of Monte da Romea, in Pontevedra, in the early 3rd millennium BC (3018–2679 cal BC), the monument was covered with a thick soil and rubble layer containing Beaker pottery (Prieto 2007, 104–6, 113). As with some of the French examples, the raising of mounds over the megalithic monuments of Pedras Grandes and Monte da Romea implies significant and deliberate transformations from stone structures to earthen tumuli or barrows.   Many other megalithic monuments were also finally closed through large stone blocking and capping deposits, but often these were disturbed by antiquarian excavations and not recognized as such. In other cases, these closure deposits have been interpreted as natural rubble collapse when they were clearly intentional, as at Méréaucourt and Pedras Grandas; another example of deliberate blocking recently noted is Los Cameros in La Rioja (López de Calle & Ilarraza 1997). Not all these deposits of Beaker-related material culture in the blocking horizons need have been grave goods associated with burials. They may have represented the residues of feasts or other types of ritual activity associated with abandonment processes. Perhaps they were potent symbols deliberately attached to the closure of these tombs; emblems that were bound up with such transformations.   To summarize, many of the megalithic monuments along the Atlantic façade were closed through the erection of stone walls or rubble plugs at the entrance and/or the building of massive cairns or mounds to enclose and cover these monuments. Similar practices may be evident in other regions, and instances of blocking or capping deposits associated with Beaker burials are known in southern Britain. One obvious example is West Kennet long barrow, the chamber of which was completely infilled with soil and rubble after the deposition of Beaker burials. Even more spectacular was the erection of three massive sarsen blocking-stones across the entrance of this tomb, creating a new façade (Piggott 1962). Capping deposits have also been identified over the top of Beaker burials, with examples from Shrewton in Wiltshire (Green & Rollo-Smith 1984) and Monktonup-Wimborne in Dorset (Green 2000, 82–3), amongst others. These capping layers have been thought to serve prosaic functions, such as forming small flint cairns or protecting those buried beneath from scavenging animals. They are not obviously marker cairns as they usually form flat, not mounded, deposits. Furthermore, most of the burials with these flint seals are in substantial pits, often greater than a metre in depth; way beyond the reach of burrowing dogs or foxes. In the majority of cases where these capping layers were present, they did not relate to the primary burial, but were deposited after later burials or revisitations. Instead, perhaps these should be considered in a similar way to the blocking deposits of megalithic monuments; these flint caps were both physically and symbolically marking a formal end to burial activity at these sites.   It is anticipated that future investigations will determine that such closure events are more common than currently realized, and thus its wider significance may be under-

Gibson_CW3.indd 95

27/05/2016 13:18:51

[ 96 ]

iii. Closed for business or cultural change?

]]

appreciated. The raising of large earthen mounds over some of these tombs transformed them into barrow monuments, perhaps intentionally, in keeping with the new architectural types emerging throughout other parts of Atlantic Europe. In northern Portugal and Galicia, other burial traditions were introduced at the beginning of the Beaker period, including the construction of mamoas5, flat cists, and the use of natural and artificial caves (e.g. Prieto & Salanova 2011; Suárez 2011). Mamoas exhibit similarities with the Beaker and Early Bronze Age barrow/cairn traditions in western France and Britain. Furthermore, some of the earliest Beaker burials in Scotland and England were interred in flat graves or cists (e.g. Sheridan 2007; Fitzpatrick 2011), and perhaps indicate stronger interaction networks developing between these Atlantic regions during this time.   Where radiocarbon dates are available, the closure events seem to have occurred between 2700 and 2250 cal BC, with a specific concentration around 2450–2250 cal BC. Often these final blocking events were accompanied by fire rituals—perhaps to cleanse, symbolically seal and physically alter the monument through burning. Metaphorically these may have been linked to a wider social and symbolic transformation. One example, already mentioned above, is that of Chaussée-Tirancourt, and La Hoguette, in Calvados, was also subject to combustion by quick-lime between 2500–2200 cal BC (Caillaud & Lagnel 1972, 25–67; Chambon 2003, 218–19). In Portugal, the dolmen of Picoto do Vasco, in Beira Alta, was closed and then burnt c. 2600–2400 cal BC—this was fire ritual on an extreme scale. The entrance to the monument was completely blocked and then totally vitrified (Abrunhosa, Gonçalves & Cruz 1995, 683–4), indicating that the fire had reached very high temperatures (above 1230o Celsius), and had burned for longer than 12 hours. This alludes to a protracted ritual event that was intended to create a visually dramatic and memorable closure process for large groups of people to witness, resulting in the total physical transformation of the monument. Stelae and megaliths There is also an association between the closure of megaliths and the deposition of both menhirs and anthropomorphic stelae with engraved weapons in Iberia and France. These stelae were sometimes placed in the chambers of monuments during the 3rd millennium BC to enhance existing funerary arenas. When the entrance to the tomb of Lagunita, in Cáceres, was sealed with rubble during the Beaker period, two stelae placed in the corridor acted as closure devices (Bueno et al. 2011, 144–6). As one lay next to a large stone pit, whose dimensions fitted exactly with the base of the stela, it was likely that it had originally stood erect at the chamber entrance (ibid., fig. 5), and was then subsequently decommissioned (Figure 3.7).   At the necropolis of Alconétar, also in Cáceres, the final burials were accompanied by Beaker grave goods including bracers, copper tools, and a fragmentary trapezoidal copper dagger (Leisner & Leisner 1959, fig. 55). An anthroporphic stela, depicted with a halberd and dagger, was then placed decorated side face down on a pile of stones in the burial 5 

Gibson_CW3.indd 96

Mamoa is Portuguese for an earthen tumulus.

27/05/2016 13:18:51

[ 97 ]

Gibson

monument. Similarly, at the nearby dolmen of Cañamero a stela with engraved weapons lay on top of a cairn immediately adjacent to the dolmen (Bueno, Balbín & Barroso. 2005; Bueno, Barroso & Balbín 2010). The dolmen of Llano de la Belleza, in Huelva, was blocked with rubble that included broken menhir fragments, incorporated into the entrance of a megalithic monument. (García Sanjuán, Rivera & Wheatley 2004, 190; Figure 3.8). 2

1

N

0

2m

1

3.7 Decommisioning of stelae during closure of the tomb of Lagunita, Cáceres, Spain

(Source: redrawn after Bueno, Balbín & Barroso 2007, fig. 33)

2

0

3.8 Re-use of menhir in the blocking deposit at the entrance to the tomb of Llano de la Belleza, Huelva, Spain (Source: redrawn after García Sanjuán, Rivera & Wheatley 2004, fig. 12)

0.5m

Key Menhir fragment Blocking material E

W 0

2m

  This re-use of menhirs in the blocking deposits of megalithic monuments during the 3rd millennium BC has been noted in other regions, particularly in Brittany. For instance, in the earlier 3rd millennium BC, the material used to close the passage tomb of Gavrinis in Morhiban included broken decorated menhir fragments (Patton 1993, 56), accompanied by several placed deposits of polished stone axes, animal bone, and pots, perhaps indicative of sets of blocking rituals (Le Roux 1984, 240–5). Stelae and menhirs were imbued with symbolism that was seemingly re-ordered or even rendered obsolete through their dismantling or breaking and use as monument closure devices in the Chalcolithic. Perhaps the stelae represented outstanding social figures or images of the ancestors turned to stone, and their use in the decommissioning of megalithic monuments may signal the end of these particular ideologies. After the mid-3rd millennium BC new belief systems now began to take their place.

Gibson_CW3.indd 97

27/05/2016 13:18:52

[ 98 ]

iii. Closed for business or cultural change?

]]

8. Geographic distribution, regional re-activation and dating closure The new dating sequences indicate that megalithic monuments con­tinue to be used and occasionally even con­ structed during the 3rd mil­len­nium BC. Evidence for Beaker re-use is to some extent restricted to certain regions of Atlantic Europe. In Iberia, there is a significant focus in the Tagus estuary around Lisbon, along the Atlantic coast, and the navigable river arteries of the Tagus and Guadi­ana (Ferreira 1959; Boaventura 2009). Although the Alentejo region has one of the highest densities of mega­lithic monuments in Europe, few of these were re-used in the Beaker period. Salanova (2003, 388– 9, fig. 2) calculated that less than two per cent had evidence for later Beaker activ­ity, but newly generated AMS dates from some of these monuments has increased that number to just under five per cent (Boaventura 2009; table 2). Comparatively fewer megalithic monuments were re-used in northern Portugal and Galicia, perhaps due to the introduction of new types of burial monuments during the Beaker period, such as barrows.   In France, re-used megaliths are con­cen­trated in the north-west, par­ tic­ularly Brittany, where one in five monu­ments contained Beaker materials (Salanova 2003, 387; 2007). Only two examples are known from the south of France out of a corpus of over 200

{ 300 km


Gibson_CW3.indd 98








3.9 Distribution of megalithic monuments in Atlantic Europe with evidence of re-use in the Beaker period (Source: author)


 10/06/2016 17:00:28

Gibson

[ 99 ]

(Gutherz & Hugues 1980), although more are known in the Paris and Seine Basins (Guy & Masset 1991; 1995; Chambon & Salanova 1996). In Britain, they mainly cluster in the south, particularly the south-west such as Cornwall and the Isles of Scilly, or near the coast; Anglesey for instance (Figure 3.9). Thus megalithic monuments that were re-used during the Beaker period predominantly hug or lie close to the western coastlines or the principal fluvial routes. Perhaps these practices highlight particular sets of long-distance maritime exchanges and reflect other shared traditions across these Atlantic regions throughout this time. Beaker burials in megalithic tombs indicate a fusion of both existing and new practices, signifying different ways in which the ‘Beaker package’ was adopted and interpreted in these areas. Here the concept of collective or communal burial spaces endured for longer, and preceding traditions continued to be respected. In addition, although Beaker inhumations were sometimes interred as complete crouched bodies with individual grave goods, often they were deposited in partially or fully disarticulated states. At Carrascal and Monte Abraão, for instance, the presence of femurs and skulls (Hillier 2008 a & b) implies bodies were excarnated prior to deposition, a practice that is more in keeping with pre-existing later Neolithic funerary traditions.   Intriguing patterns emerge when the sequences of radiocarbon dates now available for the Iberian megalithic monuments are compared. In some cases the dates indicate punctuated episodes of activity spanning the 4th and 3rd millennia BC, separated by breaks in monument use that may last up to several hundred years. It is unfortunate, however, that many megalithic monuments were constructed on acidic soils, and furthermore the assemblages of human bones from antiquarian excavations were not always kept; thus biases in both the selection and availability of organic material for dating must be acknowledged. The Beaker phase of use is quite long lived—spanning over 500 years (2800–2300/2250 cal BC). The tables modelling these dates highlight a clear re-activation phase at the beginning of the Beaker period and another quite pronounced flurry of use around 2450/2400–2250 cal BC, the latter often coinciding with their dramatic closure.   These modelled are detailed in Figures 3.10–11 and Appendix, but some may be summarized here. The earliest Beaker activity at the hypogaeum of São Paulo, in Setúbal, can be dated to c. 2780–2560 cal BC, with the final Beaker burials placed c. 2470–2200 cal BC (Barros & Santo 1997). The latest burials in the anta of Cebolinos, in the Alentejo, were interred around 2490–2270 cal BC (Gonçalves 2003), while those in the dolmen of Los Gabrieles, in Huelva, were buried around 2470–2280 cal BC (García Sanjuán, 2006). The anta of Cabeçuda, in Beira, was blocked between 2400 and 2290 cal BC (Oliveira 1997), while that of Santa Margarida, in Reguengos de Monsaraz, was closed between 2350 and 2130 cal BC (Gonçalves 2003). Charting these closure events suggests many are roughly simultaneous, and took place over a relatively short time-span. Contrary to ideas of violation and usurpation of earlier monuments by Beaker-using groups in an effort to claim new legitimacy of existing ritual and funerary centres, some archaeologists suggest that the re-use of Neolithic monuments in the Beaker period was a reaction to the assumed dramatic changes that accompanied the adoption of elements of the Beaker package (e.g. García-Sanjuán 2006; Gibson 2004; 2007; Harrison & Heyd 2007). Perhaps the re-use of megalithic monuments provided one way to mediate

Gibson_CW3.indd 99

27/05/2016 13:19:02

[ 100 ]

iii. Closed for business or cultural change?

]]

Pedra dos Mouras, Sintra

7

Mamoa do Castelo, Trás-os-Montes

Burning/closure of corridor

Chã da Parada, Porto Meninas do Crasto 4, Porto

Closure Construction of mound/cairn

Dombate, Galicia Closure

Coto dos Mouros, Beira El Collado del Mallo, La Rioja Valle de las Higueras, Toledo

Montelirio, Seville

Verdelha dos Ruivos, Tagus estuary

Beaker burials near entrance Beaker burials in antechamber

Later burial in added chamber

Later burial in added chamber Beaker pots by entrance

Carcavelos, Lisbon

3.10 Some of the radiocarbon date sequences for Iberian megalithic monuments with evidence for re-use in the Beaker period. Sites: Pedra dos Mouras (Sintra, Portugal), Mamoa do Castelo (Trás-os-Montes, Portugal), Chã da Parada (Porto, Portugal), Meninas do Crasto 4 (Baião, Portugal), Dombate (A Coruña, Galicia, Spain), Coto dos Mouros (Beira, Portugal), El Collado dzel Mallo (La Rioja, Spain), Valle de las Higueras (Toledo, Spain), Montelirio (Seville, Spain), Verdelha dos Ruivios (Tagus estuary, Portugal), and Carcavelos (Lisbon, Portugal) (Source: author)

Gibson_CW3.indd 100

27/05/2016 13:19:03

[ 101 ]

Gibson Vale do Rodrigo II, Évora

Blocking and closure

Horta, Évora Estanque, Alentejo Santa Margarida, Reguengos de Monsaraz Closure Cebolinos, Évora Cabeçuda, Beira

Blocking

Olival da Pega, Alentejo La Pijotilla, Badajoz

Construction and use

Huerta Montero, Badajoz Monte Abraão, Sintra

Beaker burials at entrance

Estria, Estremadura

Alteration of tomb and latest Beaker burials

Pedras da Granja, Sintra

Burials in upper layers

Alapraia, Estremadura

São Pedro do Estoril, Estremadura

Beaker burials (one with gold spiral)

Lapa da Furada, Estremadura Anta da Joaninha, Beira Anta da Cabeçuda, Beira

Late cist burial Closure

Poço Velho, Estremadura

3.11 Some of the radiocarbon date sequences for Iberian megalithic monuments with evidence for re-use in the Beaker period. Sites: Vale do Rodrigo II (Évora, Portugal), Horta (Évora, Portugal), Estanque (Alentejo, Portugal), Santa Margarida (Reguengos de Monsaraz), Cebolinos (Évora, Portugal), Cabeçuda (Beira, Portugal), Olival da Pega (Alentejo, Portugal), La Pijotilla (Badajoz, Spain), Huerta Montero (Badajoz, Spain), Monte Abraão (Sintra, Portugal), Estria (Estremadura, Portugal), Pedras da Granja (Sintra, Portugal), Alapraia (Estremadura, Portugal), São Pedro do Estoril (Estremadura, Portugal), Lapa da Furada (Estremadura, Portugal), Anta da Joaninha (Beira, Portugal), Anta da Cabeçuda (Beira, Portugal) and Poço Velho (Estremadura, Portugal) (Source: author)

Gibson_CW3.indd 101

27/05/2016 13:19:03

[ 102 ]

iii. Closed for business or cultural change?

]]

these developments, through embedding the new practices within pre-existing ones, thus creating unbroken social memory. New traditions were superimposed over, absorbed and adapted into old orders to ensure stability. This idea may be more relevant in explanations for isolated episodes of re-use of megalithic monuments in southern England, western Wales, and Scotland, such as at the Pont-y-saer cromlech, in Anglesey (Savory 1956, 230), Capel Garmon, in Denbighshire (ibid., 235), or Bosiliack, in Cornwall (Thomas 1984). Re-use at these sites is generally limited to a single occasion. Re-use of Iberian and French megaliths in the Beaker period, however, indicates more complicated and protracted sets of interactions with these monuments than evident in the British Isles. The new data outlined above suggest episodic, and perhaps even relatively uninterrupted use of many megalithic monuments throughout the 4th and 3rd millennia, with activity declining or formal closure occurring around 2450–2200 cal BC (see Figures 3.10–11). The apparent continuity between Neolithic, pre-Campaniforme Chalcolithic, and Beaker burial practices in Iberia and western France cannot be ignored. Rather than significant changes marking the beginning of the Beaker period, here we may be witnessing more important transformations occurring at a later stage. Beaker ceramics and other elements of material culture were being made, used and deposited, but in Iberia it is possible to trace continuity between Beaker ceramics and preceding traditions, rather than evoking cultural change. Similarities can be drawn between the shapes of Late Neolithic Vila Nova de São Pedro and Iberian Maritime Beaker vessels (Kunst 1995b; 1996); in addition the geometric motifs on the Late Neolithic schist plaques have affinities with decoration on Beaker pots. Metalworking technologies had already been introduced in the pre-Campaniforme Chalcolithic in some Iberian and French regions, and thus was not an innovation of the Beaker period (Blas 2005, 200–2; Montero 2005, 190–1; Needham 2007, 42). In Iberia, unlike the British Isles, there is well-documented evidence for Chalcolithic settlements, represented by both small and large walled or ditched enclosures, with dense concentrations around the Tagus estuary. In recent years, further clusters have been identified in the Alentejo, Meseta, and Extremadura (Garrido, Rojo, & García 2005; Mataloto 2005; Valera 2003; 2010). Recent excavations include Zambujal (Kunst 1995a; Kunst & Uerpmann 2002), Leceia (Cardoso 1994; 2000), and Perdigões (Valera, Silva, & Marques 2014). Monte da Tumba (Silva & Soares 1987) and São Pedro (Mataloto, Estrela, & Alves 2010), among many others, have demonstrated continuity in settlement from the Late Neolithic onwards until the later Beaker period (c. 3200–2300 cal BC). The fact that the closure of the megalithic monuments roughly coincides with the abandonment of many of the large enclosed settlements is significant and quite likely related. In Iberia at least, it would appear that more momentous transformations occur roughly 400 years after the initial uptake of Beakers. There is now increasing evidence that it is really only after c. 2400/2350 BC that there was a marked increase in interaction between Iberia, western France, and other parts of Europe, and it is perhaps this intensification of social and exchange networks that triggered or were accompanied by other changes (Brodie 1997; 2001; Needham 2005; 2007). One indicator that new ideologies and burial rites were adopted and adapted as part of this wider interaction may be that megalithic monuments were no longer the appropriate

Gibson_CW3.indd 102

27/05/2016 13:19:03

Gibson

[ 103 ]

domain to place the dead. These old funerary arenas were finally abandoned—they no longer held the relevance they once had, and thus the tombs were formally sealed and closed. Covering these monuments with large mounds of soil was an important element of this process—their old structures were hidden from view as they were transformed into new barrow forms—a potent symbolic transition, indicative of changing cosmologies. At the same time, the ditched and walled enclosure sites ( povoadas fortificadas) were also deserted, and settlement evidence becomes more elusive. This implies major social and economic changes, and a process of fissioning and fragmentation in both funerary and domestic spheres. While these developments may have been triggered by environmental deterioration to some extent, including soil exhaustion in areas adjacent to large settlements, other factors undoubtedly played a part. The synchronicity of the closure of megalithic tombs and abandonment of enclosed settlements with the intensification of exchange networks throughout Europe in the later 3rd millennium BC may not be coincidental, and this demands further investigation. It is still unclear whether the wider suite of changes evident in the Iberian archaeological record during this time, perhaps including turbulent transformation, rupture and displacement of social groups, formed the catalysts for or were the repercussions of wider and more complex supra-regional interaction. The impressive corpus of radiocarbon dates that have only become available in the last decade should act as a starting point for providing greater precision to the Iberian Beaker sequences. It is hardly surprising that changes can be observed throughout such an extensive time-frame; in some parts of Iberia, Beaker material culture was employed for 800 years or more. Yet the existing literature still tends to focus on the beginning and end of this period, and thus implicitly implies stability and stasis throughout much of the intermediate time. Like Needham’s 2005 tripartite model for Beaker development in the British Isles, we can begin to divide the Iberian Beaker period into several horizons, and thus build more informative and detailed schemes. Some of the early diagnostic Beaker elements (e.g. Maritime vessels and copper alloy Palmela projectile points), were not exotic, but rather represent internal progression from Late Neolithic material practices. Continuity in both settlement and burial practices during the early Beaker period (c. 2800– 2450/2400 cal BC) suggest that not only objects but other cultural traditions show steady evolution during this time. By the mid-3rd millennium BC, incipient diversity in material culture emerges, accompanied by changes in funerary and settlement traditions. The panoply of Beakerrelated objects increases (e.g. the addition of V-perforated buttons and bracers), and the abandonment of megalithic monuments is counterbalanced by new burial types. The introduction of small barrows, cists and flat pit graves highlight greater diversity in funerary traditions; in addition cremation now becomes more common (Valera, Silva & Romero 2014, 22–3). From this time onwards (c. 2450/2400–2200 cal BC), previous continuity and coherence breaks down, replaced by emergent fluidity and hybridity, as new traditions were adopted, fused with and superimposed upon existing practices. Charting the temporality and complexity of these wider exchange networks between Iberia and other Atlantic regions can now be realized, and this will help tease out and refine our understanding of the levels, directions and implications of such interaction.

Gibson_CW3.indd 103

27/05/2016 13:19:03

[ 104 ]

iii. Closed for business or cultural change?

]]

Acknowledgements This research forms part of the collaborative project ‘Atlantic Europe in the Metal Ages’, and I am most grateful to the other members of this project for their input and ideas. In particular I wish to thank Kerri Cleary for providing me with lots of suggestions and advice during the paper’s construction, and Fernando Fernández for checking my Iberian references. I am also grateful to Professor John T. Koch and Professor Sir Barry Cunliffe for undertaking the onerous editorial work. I want to thank Chad for offering valuable comments on this paper. Needless to say, all omissions and inaccuracies remain the responsibility of this author. Finally I wish to dedicate this paper to the memory of Colin Burgess. It was through several inspirational conversations with him that sparked the ideas that led to this paper’s creation, and I am sorry it was finished too late for him to read it.

BIBLIOGRAPHY Abrunhosa, M. J., A. A. Gonçalves, & D. J. da Cruz 1995 ‘Ocorrência de rochas vitrificadas no dolmen do ‘Picoto do Vasco’ (Vila Nova de Pavia), Viseu’, Estudós Pré-Historicos 3, 167–85. Andrés Rupérez, M. T. 2000 ‘El espacio funerario dolménico: abandono y clausura’, Saldvie 50, 59–76. Andrés Rupérez, M. T., M. L. García, & J. Sesma 2001 ‘El sepulcro campaniforme de Tres Montes (Bardenas Reales, Navarra). Intervención de urgencia de 1991 y campañas de 1996 y 1997, Trabajos de Arqueologia Navarra 15, 315–22. Andrés Rupérez, M. T., M. L. García, & J. Sesma 2002 ‘Una tumba destruida por el fuego: el sepulcro campaniforme de Tres Montes, en Las Bardenas Reales (Navarra)’, Sobre el Significado del Fuego en los Rituales Funerarios del Neolítico, eds M. A. RojoGuerra & M. Kunst, 191–218. Valladolid, Studia Archaeologica 91. Andrés Rupérez, M. T. & I. Barandiarán Maestu 2004 La tumba Calcolítica de La Atayuela, treinta y cinco años después, Saldvie 4, 85–124. Apellániz, J. M. 1973 ‘Corpus de materiales de las cultural prehistóricos con cerámica de la población de cavernas de País Vasco Meridional’, Munibe, Suplemento 1, 1–366. Aranda Jiménez, G. & A. Lozano Medina 2014 ‘The chronology of megalithic funerary practices: a Bayesian approach to grave 11 at El Barranquete necropolis (Almería, Spain)’, Journal of Archaeological Science 50, 369–82. Arias, P. 2012 ‘Funerary practices in Cantabrian Spain (9000–3000BC)’, Funerary Practices in the Iberian Peninsula from the Mesolithic to the Chalcolithic, eds J. F. Gibaja, A. F. Carvalho, & P. Chambon, 7–20. BAR

Gibson_CW3.indd 104

International Series S 2417. Oxford, Archaeopress. Barandiarán Maestu, I. 1973 ‘Nota preliminar sobre el enterramiento colectivo de La Atayuela en Agoncillo (Logroño)’, Miscelánea de Arqueología Riojana, 79–99. Barclay, A. & C. Halpin 1998 Excavations at Barrow Hills, Radley, Oxfordshire. Volume I: The Neolithic and Bronze Age Monument Complex. Oxford, Oxford Archaeological Unit. Barclay, A., P. Marshall & T. F. G. Higham 2011 ‘Chronology and the radiocarbon dating’ , The Amesbury Archer and the Boscombe Bowmen, Bell Beaker Burials ar Boscpmbe Down, Amesbury, Wiltshire, ed. A. P. Fitzpatrick, 167–84. Wessex Archaeological Reports 27. Salisbury, Wessex Archaeology. Barrett, J. C. 1988 ‘The living, the dead, and the ancestors: Neolithic and Early Bronze Age mortuary practices’, The Archaeology of Context in the Neolithic and Bronze Age, eds J. C. Barrett & I. A. Kinnes, 30–41. Sheffield, Department of Archaeology and Prehistory. Barrett, J. C. 1990 ‘The monumentality of death: the character of Early Bronze Age mortuary mounds in southern Britain’, World Archaeology 22, 179–89. Barrett, J. C. 1994 Fragments from Antiquity. Oxford, Blackwell. Barros, L. & P. E. Santo 1997 ‘Gruta artificial de São Paulo’, Setúbal Arqueológica, 11–12, 217–220. Blas Cortina, M. 2005 ‘Un témoignage probant de l’exploitation préhistorique du cuivre dans le nord de la Péninsule Ibérique: le complexe minier d’El Aramo (Asturias)’, La première métallurgie en France et dans les pays limitrophes, 195–206, eds P. Ambert & J. Vaquer. Mémoires de la Société Préhistorique Française 37.

27/05/2016 13:19:03

Gibson

Boaventura, R. J. 2009 ‘As antas e o Megalitismo da região de Lisboa’. Unpublished doctoral thesis, University of Lisbon (2 vols). Boaventura, R. & J. L. Cardoso 2011 ‘Revendo is artefactos lascados da anta de Pedras da Granja, Sintra’, Estudos Arqueológicos de Oeiras 18, 175–99. Boaventura, R. & R. Mataloto 2013 ‘Entre mortos e vivos: nótulas acerca da cronologia absoluta do Megalitismo do Sul de Portugal’, Revista Portuguesa de Arqueologia 16, 81–101. Brodie, N. 1997 ‘New perspectives on the Bell Beaker culture’, Oxford Journal of Archaeology 16(3), 297–314. Brodie, N. 2001 ‘Technological frontiers and the emergence of the Beaker culture’, Bell Beakers Today: Pottery, People, Culture, Symbols in Prehistoric Europe. Proceedings of the International Colloquium Riva del Garda (Trento, Italy), 11–16 May 1998, ed. F. Nicolis, 487–96. Trento, Ufficio Beni Culturali. Bronk Ramsey, C. 2013 OxCal v. 4.2.2, r.2 (https:// c14.arch.ox.ac.uk/oxcal/OxCal.html). Radiocarbon Accelerator Unit, University of Oxford. Brown, F., C. Howard-Davis, & M. Brennand 2007 ‘The early prehistoric landscape’, The Archaeology of the A1 (M) Darrington to Dishforth Road Scheme, eds F. Brown, C. Howard-Davies, M. Brennand, A. Boyle, T. Evans, S. O’Connor, A. Spence, R. Heawood, & A. Lupton, 17–119. Lancaster, Imprints. Bueno Ramírez, P. R. de Balbín Behrmann, & R. Barroso Bermejo 2005 ‘Hierarchization et métallurgie: statues armées dans la Péninsule Ibérique’, L’ Anthropologie 109, 577–640. Bueno Ramírez, P., R. Barroso Bermejo, & R. de Balbín Behrmann 2004 ‘Construcciones megalíticas avanzadas de la cuenca interior del Tajo’, SPAL 13, 83–112. Bueno Ramírez, P., R. Barroso Bermejo, & R. de Balbín Behrmann 2007 ‘Campaniforme en las construcciones hipogeas del Megalitismo reciente al interior de la Península Ibérica’, Veleia 24–5, 771–90. Bueno Ramírez, P., R. Barroso Bermejo, & R. de Balbín Behrmann 2010 ‘Metal and the symbols of ancestors in Northern Iberia’, Conceptualising Space and Place: On the role of Agency, Memory and Identity in the Construction of Space from the Upper Palaeolithic to the Iron Age in Europe. Proceedings of the XV UISPP World Congress (Lisbon, 4-9 September 2006). Vol. 41, eds A. M. S. Bettencourt, M. Jesus Sanches, L. B. Alves, & R. Fábregas Valcarce, 71–87. BAR International Series S 2058. Oxford, Archaeopress. Bueno Ramírez, P., R. Barroso Bermejo, R. de Balbín Behrmann, E. Cerrillo Cuenca, A. Gonzalez Cordero, & A. Prada Gallardo 2011 ‘Megaliths and stelae in the inner basin of the Tagus river: Santago

Gibson_CW3.indd 105

[ 105 ]

de Alcántara, Alconétar and Cañamero (Cáceres, Spain)’, From the Origins: the Prehistory of the Inner Tagus Region, eds P. Bueno-Ramírez, E. Cerrillo Cuenca, & A. Gonzalez Cordero, 143–60. BAR International Series S 2219. Oxford, Archaeopress. Caillaud, R. & É. Lagnel 1972 ‘Le cairn et le crématoire néolithiques de La Hoguette à Fontenay-le-Marmion’, Gallia Préhistoire 15, 137–97. Cardoso, J. L. 1994 ‘Leceia 1983-1993. Escavações do povoado fortificado pré-histórico’, Estudos Arqueológicos de Oeiras. Oeiras, Câmara Municipal, número especial. Cardoso, J. L. 1997 ‘A cronologia absoluta do depósito arqueológico da Lapa da Furada—Azoia, Sesimbra: Seu significado e incidências rituais e culturais’, Sesimbra Cultural 6, 10–15. Cardoso, J. L. 2000 ‘The fortified site of Leceia (Oeiras) in the context of the Chalcolithic in Portuguese Estremadura’, Oxford Journal of Archaeology 19(1), 37–55. Cardoso, J. L. 2014 ‘Absolute chronology of the Beaker phenomenon North of the Tagus estuary: demographic and social implications’, Trabajos de Prehistoria 71(1), 56–75. Cardoso, J. L. & A. M. Soares 1990-1992 ‘Cronologia absoluta para o campaniforme da Estremadura e do Sudoeste de Portugal’, O Arqueólogo Português Série IV(10), 203–28. Cardoso, J.L., J. Carlos Caninas, & F. Henrique 2000 ‘Arquitectura, espólio e rituais de dois monumentos megalíticos da Beira Interior: estudo comparado’, Muita gente, poucas antas? Espaços e contextos do Megalitismos, Actas do II Colóquio Internacional sobre Megalitismo Trabajos de Arqueologia 16, 195–215. Cassen, S. 2014 ‘The first radiocarbon dates for the construction and use of the interior of the monument at Gavrinis (Larmor-Baden, France)’, PAST, The Newsletter of the Prehistoric Society 77, 1–4. Chambon, P. 2003 Les morts dans les sépultures collectives néolithiques en France. Supplement to Gallia Préhistoire. Paris, CNRS. Chambon, P. & L. Salanova 1996 ‘Chronologie des sépultures du IIIe millénaire dans le bassin de la Seine’, Bulletin de la Société Préhistorique Française 93(1), 103–18. Cruz, D. J., J. S. López Saez, A. L. Canha, S. L. Mendes, A. Valinho, & M. A. Vieira, 2003 Projecto “O Alto Paiva: Sociedade e estratégias de povoamento desde a Pré-história Recente à Alta Idade Média”: Relatório final (1998-2002). Coimbra, Policopiado. Acessível no arquivo do IGESPAR, Proc. 98/1(762). Delibes de Castro, G., M. Alonso Diez & M. RojoGuerra 1987 Los sepulcros colectivos del Duero

27/05/2016 13:19:03

[ 106 ]

iii. Closed for business or cultural change?

Medio y las Loras y su conexión com el foco dolménico riojano, El Megalitismo en la Peninsula Ibérica, ed. G. Delibes de Castro, 181–97. Madrid, Ministerio de Cultura. Drewett, P. 1975 ‘The excavation of an oval burial mound of the third millennium BC at Alfriston, East Sussex’, Proceedings of the Prehistoric Society 41, 119–52. Era-Arqueologia 2001 ‘Valorização do património arqueológico do município de Odivelas: Anta das Pedras Grandes’, Caneças: Relatório dos trabalhos arqueológicos. Lisbon, Policopiado. Fergusson, J. 1872 Rude Stone Monuments in all Countries: their Ages and Uses. London, Murray. Fernández Florez, A. & V. Aycart 2013 ‘Montelirio. Un sepulcro clave para la comprehensión del registro de los grandes monument megalíticos de Valencina de la Concepción-Castilleja de Guzmán (Sevilla)’, El asentamiento prehistórico de Valencina de la Concepción (Sevilla): investigación y tutela en el 150 aniversario del descubrimiento de La Pastora, eds J. M. Vargas, V. Hurtado, T. Ruiz Moreno, & R. Cruz-Auñón, 233–59. Seville, University of Seville. Ferreira, O. V. 1959 ‘Inventário dos monumentos megalíticos dos arredores de Lisboa’, Actas e Memórias do 1º Congresso Nacional de Arqueologia, Lisboa, 15 a 20 Dezembro de 1958, vol. 1, 215–33. Ferreira, O. V. 1963 ‘Notícia de algumas estações préhistóricas e objectos isolados inéditos ou pouco conhecidos’, Boletim da Junta Distrital de Lisboa 2ª série, 59–60. Fitzpatrick, A.P. 2011 The Amesbury Archer and the Boscombe Bowmen. Bell Beaker Burials at Boscombe Down, Amesbury, Wiltshire. Wessex Archaeological Reports 27. Salisbury, Wessex Archaeology. Fokkens, H. & F. Nicolis eds 2012 Background to Beakers. Inquiries into Regional Cultural Backgrounds of the Bell Beaker Complex. Leiden, Sidestone Press. Gallay, A. 1976 ‘The position of the Bell Beaker civilization in the chronological sequence of the Petit-Chasseur (Sion, Valais, Switzerland), Glockenbecher Symposion, (Oberried, 18-23 Marz 1974), eds J. N. Lanting & J. D. van der Waals, 279–306. Haarlem, Bussum. García Sanjuán, L. 2005 ‘Las piedras de la memoria. La permanencia del megalitismo en el suroeste de la Península Ibérica durante el II y I milenios a.n.e’, Trabajos de Prehistoria 62(1), 85–109. García Sanjuán, L. 2006 ‘Funerary ideology and social inequality in the Late Prehistory of the Iberian South-West (c. 3300-850 cal BC)’, Social Inequality in Iberian Late Prehistory, eds P. Díaz del Río, & L. García Sanjuán, 149–70. BAR International Series

Gibson_CW3.indd 106

]]

S1525. Oxford, Archaeopress García Sanjuán, L., T. Rivera Jiménez, & D. W. Wheatley 2004 ‘Prospección de superficie y documentación gráfica en el Dolmen del Llano de la Belleza (Aroche, Huelva)’, Anuario Arqueológico de Andalucía 2, 181–192. Gibson, A. M. 2004. Burials and Beakers: Seeing Beneath the Veneer in Late Neolithic Britain, Similar but Different: Bell Beakers in Europe, ed. J. Czebreszuk, 173–91. Poznań, Adam Mickiewicz University. Gibson, A. M. 2007 ‘A Beaker veneer? Some evidence from the burial record’, From Stonehenge to the Baltic: Living with Cultural Diversity in the 3rd Millennium BC, eds M. Larsson & M. Parker Pearson, 47–64. BAR International Series S 1692. Oxford, Archaeopress. Giot, P. R., J. L’Helgouach, & J.-L. Monnier 1979 Préhistoire de la Bretagne. Rennes, Ouest-France. Gomes, M.V. 2012 ‘Early Neolithic practices in Castelo Belinho’s village (Western Algarve, Portugal)’, Funerary Practices in the Iberian Peninsula from the Mesolithic to the Chalcolithic, eds J. F. Gibaja, A. F. Carvalho & P. Chambon, 113–23. BAR International Series S 2417. Oxford, Archaeopress. Gonçalves, V. S. 1999 ‘Time, landscape and burials. 1. Megalithic rites of ancient peasant societies in central and southern Portugal’, Journal of Iberian Archaeology 1, 83–91. Gonçalves, V. S. 2003 ‘A anta 2 da Herdade dos Cebolinhos (Reguengos de Monsaraz, Évora): Sinopse das intervenções de 1996–97 e duas datações de radiocarbono para a última utilização da câmara ortostática’, Revista Portuguesa de Arqueologia 6(2), 143–66. Gonçalves, V. S. 2005 ‘Espaços construídos, símbolos e ritos da morte das antigas sociedades camponesas no Extremo Sul de Portugal: algumas reflexões sobre a forma de sete qmf ’, Mainaké 36, 89–113. Gonçalves, V. S. 2008 ‘Na primeira metade do 3º milénio a.n.e., dois subsistemas mágico-religiosos no Centro e Sul de Portugal’, Actas del 4º Congreso del Neolítico Peninsular, eds M. Hernández-Pérez, J. Soler-Díaz, J., & J. López Padilla, 112–20. Alicante, Museo Arqueológico de Alicante. Green, C. and Rollo-Smith, S. 1984 ‘The excavation of eighteen round barrows near Shrewton, Wilt­ shire’, Proceedings of the Prehistoric Society 50, 225–318. Green, M. 2000 A Landscape Revealed. 10, 000 Years on a Chalkland Farm. Stroud, The History Press. Gutherz, X. & C. Hugues 1980 ‘La culture des vases campaniformes dans le département du Gard (France)’, École Antique de Nîmes 15, 5–26. Guy, H. & C. Masset 1991 ‘Procédure de

27/05/2016 13:19:04

Gibson

condemnation d’une allée couverte Seine-OiseMarne (Méréaucourt, Somme)’, Bulletin de la Société Préhistorique Française 88(9), 282–8. Guy, H. & C. Masset 1995 ‘Le dispositive de fermeture de l’ allée couverte de Méréaucourt, (Somme)’, Bulletin de la Société Préhistorique Française 92(2), 266–8. Harrison, R. J. 1988 ‘Bell Beakers in Spain and Portugal: working with radiocarbon dates in the 3rd millennium BC’, Antiquity 62, 464–72. Harrison, R. J. & V. Heyd 2007 ‘The transformation of Europe in the third millennium BC: The example of ‘Le Petit Chasseur I+III’ (Sion, Valais, Switzerland)’, Praehistorische Zeitschrift 82/2, 129–214. Hillier, M. 2008a Osteological report—The dolmen of Estria. Unpublished report prepared for Rui Boaventura PortAnta: Archaeological Opportunities in Portugal. Hillier, M. 2008b Osteological report—The dolmen of Carrascal. Unpublished report prepared for Rui Boaventura âmbito do programa Megaosteology, 2007, Anexo 6. Hoskin, M. 2001 Tombs, Temples and Their Orientations: A New Perspective on Mediterranean Prehistory. Oxford, Ocarina Books. Hurtado-Pérez, V. & A. Amores 1984 ‘El tholos deLas Canteras y los enterramientos del Bronce en la necropolis de El Gandul (Alcalá de Guadaira, Sevilla)’, Cuadernos de Prehistoria de la Universidad de Granada 9, 147–74 Kendrick, T. D. 1928 The Archaeology of the Channel Islands—Volume 1, Guernsey. London, Methuen. Kinnes, I. 1985 Beaker and Early Bronze Age Grave Groups, British Bronze Age Metalwork Associated Finds Series A7-16. London, British Museum. Kunst, M. 1995a ‘Central places and social complexity in the Iberian Copper Age’, The Origin of Complex Societies in Late Prehistoric Iberia, ed. K. Lillios, 32–43. Ann Arbor, Archaeological Series 8, International Monographs in Prehistory. Kunst, M. 1995b ‘Cerâmica do Zambujal—novos resultados para a cronologia da cerâmica Calcolítica’, Origens, estruturas e relações das culturas Calcolíticas da Península Ibérica, ed. M. Kunst, 21–9. Trabalhos de Arqueologia 7, Lisboa, IPPAR. Kunst, M. 1996 ‘As ceramicas decoradas do Zambujal e o faseamento do Calcolitico da Estremadura portuguesa’, Estudos Arqueológicos de Oeiras 6, 257–87. Kunst, M., & H.-P. Uerpmann 2002 ‘Zambujal (Torres Vedras, Lisboa): relatório das escavações de 1994 e 1995, Revista Portuguesa de Arqueologia 5(1), 67–120. Larsson, L. 2000 ‘Symbols in stone—ritual activities and petrified traditions’, Actas do 3º Congresso de Arqueologia Peninsular, Porto, vol. 3, 445–58.

Gibson_CW3.indd 107

[ 107 ]

Lazarich González, M. & M. Sánchez Andreu 2000 ‘Los enteramientos campaniformes en supultres’, Monuments and Landscape in Atlantic Europe. Perception and Society during the Neolithic and Bronze Age, ed. C. Scarre, 152–76. London, Routledge. Leclerc, J. & C. Masset 1980 ‘Construction, remaniements et condemnation d’une sépultre collective néolithique: La Chaussée-Tirancourt (Somme)’, Bulletin de la Société Préhistorique Française 77(2), 57–64. Leclerc, J. & C. Masset 2006. ‘L’évolution de la pratique funéraire dans la sépulture collective néolithique de La Chaussée-Tirancourt (Somme)’, Bulletin de la Société Préhistorique Française 103(1), 87–116. Leisner, G. & V. Leisner 1959 Die Megalithgräber der Iberischen Halbinsel. Der Westen. Vol. 2. Berlin, Walter de Gruyter & Co. Leisner, V. 1965 Die Megalithgräber der Iberischen Halbinsel. Der Westen. Vol. 3. Berlin, Walter de Gruyter & Co. Leisner, V. & O. V. Ferreira 1961 ‘Monumentos megalíticos de Trigache e Beja: II: Monumentos megalíticos’, Comunicações dos Serviços Geológicos de Portugal 40, 300–37. Leitão, M., C. T. North, J. Norton, O. da Ferreira, & G. Zbyszweski 1984 ‘The prehistoric burial cave at Verdelha dos Ruivos (Vialonga), Portugal’, L’Âge du Cuivre européen: Civilizations à vases campaniformes, ed. J. Guilaine, 221–39. Paris, CNRS. Le Roux, C.-T. 1984 ‘A propos des fouilles de Gavrinis (Morbihan): nouvelles données sur l’art mégalithique Armorican’, Bulletin de la Société Préhistorique Francaise 81, 240–5. L’Helgouach, J. 1965 Les sépultures mégalithiques en Amorique. Dolmens à couloir et allées couvertes. Unpublished Doctoral thesis, Université de Bretagne, Rennes.  L’Helgouach, J. 1976 ‘Les relations entre le groupe des vases campaniformes et les groupes néolithiques dans l’Ouest de la France’, Glockenbecher Symposion, (Oberried, 18-23 Marz 1974), eds J. N. Lanting & J. D. van der Waals, 439–51. Haarlem, Bussum. López de Calle, C. & J. A. Ilarraza 1997 ‘Fases antiguas del Megalitismo de Cameros (La Rioja): Caracterización y cronología’, Neolitico Atlântico e as orixes do Megalitismo: Actas do Coloquio Internacional, Abril 1996, ed. A. Rodriguez Casal, 415–43. Santiago de Compostela, Universidade de Santiago de Compostela. Loveday, R. 2002 ‘Duggleby Howe revisited’, Oxford Journal of Archaeology 21(2), 135–46. Marcos, A. 1973 ‘Trabajos del seminario de

27/05/2016 13:19:04

[ 108 ]

iii. Closed for business or cultural change?

Arqueología de la Universidad de Navarra en la provincia de Lograño durante los años 1965 y 1966’, Miscelánea de Arqueología Riojana 9–52. Marqués de Loriana 1953 ‘Dos nuevos hallazgos megalíticos en la provincia de Molinilla’, Noticiario Arqueológico Hispánico 1, 191. Marsden B.M. 1970 ‘The excavation of the Bee Low round cairn, Youlgreave, Derbyshire’, Antiquaries Journal 50, 186–215. Masset, C. 1992 ‘Un solution pour le mystère d’Hérouval’, Bulletin de la Société Préhistorique Française. 89(8), 227–8. Masset, C. 1998 ‘Préface’, La France des dolmens el des sépultres collectives (4500-200 avant J.C.), ed. P. Soulier. Paris, Éditions Errance. Masset, C., J. Pelegrin, H. Plisson, A. Blin, F. Mazhoud, & M. Girard, 2013 ‘L’allée couverte du bois d’Archemont à Méréaucourt (Somme)’, Gallia Préhistoire 55, 79–179. Mataloto, R. 2005 ‘Meio Mundo 2: a fortificação calcolítico do Alto de São Gens (Redondo/ Estremoz, Alentejo Central)’, Revista Portuguesa de Arqueología 8(2), 115–28. Mataloto, R., S. Estrela, & C. Alves, C. 2010 ‘O IV o/IIIo Milénio a. C. no povoado de São Pedro (Redondo, Alentejo Central): Fortifiçacão e Povoamento na Planicie Centro Alentejana’, Transformação e Mudança no Centro e Sul de Portugal 3500 a 2000 a.n.e., ed. V. S. Goncalves, 263–95. Cascais, Camara Municipal. Mederos Martín, A. 2013 ‘La cronología de Montelirio (Castilleja de Guzmán, Seville), VI Encuentro de Arqueología del Suroeste Peninsular, 2597–612. Montero Ruiz, I. 2005 ‘Métallurgie ancienne dans la Péninsule Ibérique’, La première métallurgie en France et dans les Pays limitrophes, eds P. Ambert & J. Vaquer, 187–94. Paris, Mémoire de la Société Préhistorique Française 37. Morán, E. & R. Parreira 2004 Alcalar 7: Estudo e reabilitação de um monumento megalítico. Lisboa, IPPAR. Navarte Sanz, N. 2007 ‘Revisión del dolmen de la Mina (Molinilla, Álava). Aspectos arquitectónicos y nuevas datos materiales’, Cæsaraugusta 78, 145–58. Needham, S. P. 2005 ‘Transforming Beaker culture in North-West Europe; processes of fusion and fission’, Proceedings of the Prehistoric Society 71, 171–205. Needham, S. P. 2007 ‘Isotopic aliens: Beaker movement and cultural transmission’, From Stonehenge to the Baltic: Living with Cultural Diversity in the Third Millennium BC, eds M. Larsson & M. Parker Pearson, 41–6. BAR International Series S 1692. Oxford, Archaeopress. Oliveira, J. 1997 ‘Datas absolutas de monumentos

Gibson_CW3.indd 108

]]

megalíticos da bacia hidrográfica do rio Sever’, II Congreso de Arqueologia Peninsular: Neolítico, Calcolítico y Bronce. Zamora, 24 a 27 de Septiembre de 1996, eds R. de Balbín Behrmann & P. Bueno Ramírez, vol. 2, 229–39. Zamora, Fundación Rei Afonso Henriques. Patton, M. 1993 Statements in Stone. Monuments and Society in Neolithic Brittany. London, Routledge. Pellicer Catalán, M. & V. Hurtado-Pérez 1987 ‘Excavaciones en La Mesa del Gandul (Alcalá de Guadaira, Sevilla)’, Anuario Arqueológico de Andalucía 2, 338–41. Perkins, D. R. J. 2004 ‘Oval barrows in Thanet’, Towards a New Stone Age: Aspects of the Neolithic in South-East England, eds J. Cotton & D. Field, 76–81. London, CBA Research Report 137. Piggott, S. 1962. The West Kennet Long Barrow. Excavation 1955–6. London, HMSO. Prieto Martínez, M.a P. 2007 ‘Volviendo a un mismo lugar: recipientes y espacios en un monumento megalítico gallego (NW de España)’, Revista Portuguesa de Arqueología 2, 101–25. Prieto Martínez, M.a P. & L. Salanova 2011 Las comunidades Campaniformes en Galicia. Cambios sociales en el III y II milenios BC en el NW de la Península Ibérica. Pontevedra, Deputación Pontevedra. Reimer, P. J., E. Bard, A. Bayliss, J. W. Beck, P. G. Blackwell, C. Bronk Ramsey, P. M. Grootes, T. P. Guilderson, H. Haflidason, I. Hajdas, C. Hattž, T. J. Heaton, D. L. Hoffmann, A. G. Hogg, K. A. Hughen, K. F. Kaiser, B. Kromer, S. W. Manning, M. Niu, R. W. Reimer, D. A. Richards, E. M. Scott, J. R. Southon, R. A. Staff, C. S. M. Turney, & J. van der Plicht 2013 ‘IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0–50,000 Years cal BP’, Radiocarbon 55(4), 1869–87. Ribeiro, C. 1871–1875 Caderno de Campo 22–VI a 10–VIII–1871 até Jan. e Fev. 1875. (Not catalogued). Unpublished site notebooks, kept in the Arquivo Histórico, Geológico e Mineiro. Laboratório Nacional de Energia e Geologia. Ribeiro, C. 1880 Estudos Prehistoricos em Portugal: Noticia de algumas estações e monumentos prehistoricos. II - Monumentos megalithicos das visinhanças de Bellas. Lisboa, Typographia da Academia. Rojo-Guerra, M. A., R. Garrido-Pena, & I. GarcíaMartínez de Lagrán 2010 ‘Tombs for the dead, monuments to eternity: the deliberate destruction of megalithic graves by fire in the interior highlands of Iberia (Soria province, Spain)’, Oxford Journal of Archaeology 29(3), 253–75. Rojo-Guerra, M. A., G. Morán Dauchez, & M. Kunst 2003 ‘Un défi à l’éternité: genèse et réutilizations du Tumulus de La Sima (Miño de Medinaceli, Soria,

27/05/2016 13:19:04

Gibson

Espagne)’, Revue Archéologique de Picardie No. Spécial 21, 173–84. Rojo-Guerra, M. A. & R. Garrido-Pena 2012 ‘From pits to megaliths: Neolithic burials in the interior of Iberia’, Funerary Practices in the Iberian Peninsula from the Mesolithic to the Chalcolithic, eds J. F. Gibaja, A. F. Carvalho & P. Chambon, 21–8. BAR International Series S 2417. Oxford, Archaeopress. Rojo-Guerra, M. A., M. Kunst, & A. L. PalominoLázaro 2002 ‘El fuego como procedimiento de clausura en tres tumbas monumentales de la Submeseta Norte’, Sobre el Significado del Fuego en los Rituales Funerarios del Neolítico, eds M. A. RojoGuerra & M. Kunst, 21–38. Valladolid, Studia Archaeologica 91. Rojo-Guerra, M.A., R. Garrido-Pena, M. Kunst, R. Garrido-Pena, I. García-Martínez de Lagrán, & G. Morán Dauchez 2005 Un desafío a la eternidad. Tumbas monumentales del Valle de Ambrona (Valladolid). Valladolid, Junta de Castilla y León. Salanova, L. 2003 ‘Les sépultures mégalithiques et le phénoméne campaniforme’, Muita gente, poucas antas? Espaços e contextos do Megalitismos, Actas do II Colóquio Internacional sobre Megalitismo, Trabajos de Arqueologia 16, 385–93. Salanova, L. 2007 ‘Les sépultures campaniformes: lecture sociale’, Le Chalcolithique et la Construction des Inégalités, ed. J. Guilaine, 213–28. Paris, Éditions Errance. Saville, A. 1990 Hazelton North. The Excavation of a Neolithic Long cairn of the Cotswold-Severn Group. London, English Heritage. Savory, H. N. 1956 ‘Archaeological section. Report for 1955-6’, Archaeologia Cambrensis 105, 220–42. Scarre, C. 2010 ‘Rock of ages: tempo and time in megalithic monuments’, European Journal of Archaeology 13(2), 175–93. Shepherd, I. 1989 ‘Notes on a Beaker cist at Donside Field, North East Manar, Inverurie, Aberdeenshire’, Discovery and Excavation in Scotland 1989, 89–90. Senna-Martinez J. C. & J. M. Ventura 2008 ‘Do mundo das sombras ao mundo dos vivos: Octávio da Veiga Ferreira e o Megalitismo da Beira Alta, meio séculos depois’, Octávio da Veiga Ferreira Homenagem ao homem, ao arqueólogo e ao professor, ed. J. L. Cardoso, 317–50. Estudos Arqueológicos de Oeiras, Oeiras, Câmara Municipal, 16. Sheridan, A. 2007 ‘Scottish Beaker dates: the good, the bad and the ugly’, From Stonehenge to the Baltic: Living with Cultural Diversity in the 3rd Millennium BC, eds M. Larsson & M. Parker Pearson, 91–203. BAR International Series S 1692. Oxford, Archaeopress. Silva, C.T. & J. Soares 1987 ‘O povoado fortificado

Gibson_CW3.indd 109

[ 109 ]

Calcolítico do Monte da Tumba. I—Escavações de 1982–86 (resultados preliminares)’, Setúbal Arqueológica 8, 29–79. Smith, J A 1882 ‘Notice of a short stone cist, containing a human skeleton and a “drinking cup” urn, found at Drem, East Lothian’, Proceedings of the Society of Antiquaries of Scotland 16, 299–300. Sohn, M. 2006 ‘Le mobilier de la sépulture collective de La Chaussé Tirancourt (Somme) dans l’espace et dans le temps: de la fin du IVe à la fin du IIIe millénaire av. J.-C.’, Bulletin de la Société Préhistorique Française. 103(1), 117–31. Suárez Otero, J. 2011 ‘Descubrimento y caracterisación del Campaniforme Gallego: la Península de O Morrazo como paradigma’, Las comunidades Campaniformes en Galicia. Cambios sociales en el III y II milenios BC en el NW de la Península Ibérica, eds M. P. Prieto Martínez & L. Salanova, 285–95. Pontevedra, Deputación Pontevedra. Thomas, C. 1984. Archaeological excavations at Bosiliack, Madron, Cornwall, Cornwall Council, Report No. 2012RO17. Unpublished report. Thomas, J. 2000 ‘Death, identity and the body in Neolithic Britain’, Journal of the Royal; Anthropological Institute (N.S.) 6, 653–68. Thomas, J. 2012 ‘Los monumentos megalíticos de Europa’, El neolítico en la Península Ibérica y su contexto europeo, eds M. A. Rojo Guerra, R. Garrido Pena, & I. García-Martínez de Lagrán, 55–70. Madrid, Catedra. Thurnam, J. 1872 ‘On ancient British barrows, especially those of Wiltshire and the adjoining counties (part II, round barrows)’, Archaeologia 43, 285–552. Valera, A. C. 2003 ‘O abandono de povoados fortificados calcolíticos no Ocidente Peninsular’, ERA-Arqueologia 5, Lisboa, Colibri, 126–48. Valera, A. C. 2010 ‘Construção da temporalidade dos Perdigões: contextos neolíticos da área centra’, Apontamentos de Arqueologia e Património 5, 19–26. Valera, A. C., A. M. Silva, & J. E. Márques Romero 2014 ‘The temporality of Perdigões enclosures: absolute chronology of the structures and social practices, SPAL 23 11–26. Weiss-Krejci, E. 2005 ‘Formation processes of deposits with burned human remains in Neolithic and Chalcolithic Portugal’, Journal of Iberian Archaeology 7, 37–73. Woodward, A. 2000 British Barrows: a Matter of Life or Death. Stroud, Tempus. Zbyszweski, G., O. da V. Ferreira, M. Leitão, C. T. North, & J. Norton 1977 ‘Le monument de “Pedras da Granja” ou de “Pedras Altas” dans la “Várzea de Sintra”’, Ciências da Terra 3, 197–239.

27/05/2016 13:19:04

Appendix. All radiocarbon dates cited in text associated with re-use of megalithic monuments in the Iberian Peninsula. Dates are listed in alphabetical order In material: Ch–Charcoal; HB–Human Bone Site Name

Lab ref

Context

Abrão, Queluz, Sintra

Beta–228580 Beta–228579

Alcalar 7, Portimão, Algarve

Beta–180980

GrA–9577 GrN–23198 GrN–9255 BM–2366 BM–2367 BM–2365

Chamber; disarticulated body In blocking material near entrance; partially articulated body Hearth 4 in layer under stone constructed in front of façade/blocking Burning at base of chamber associated with human bones Lower part of chamber Chamber Corridor Middle layer, female burial Middle later, male burial Upper layer, male burial

Uncal Date BP

Cal date BC 92 sigma)

Reference

HB HB

4180 ± 40 4110 ± 40

2900–2630 2880–2570

Boaventura 2009: Table 22

Ch – pistarcia lenticus Ch

3860 ± 40

2470–2200

Morán & Parreira 2004

4160 ± 50

2890–2580

HB Ch HB HB HB HB

3740 ± 50 3710 ± 60 3650 ± 50 4120 ± 70 4110 ± 60 4060 ± 60

2300–2010 2290–1940 2150–1890 2880–2566 2878–2564 2765–2470

ICEN–979 ICEN–977 Beta–228577 Beta–226167 Beta–176899 Beta–177471 GrA–21405 GrA–21403 GrA–21402 GrA–2140 Gif–62593 GrA–21403 GrA–21401 Gif–6529 Gif–1372 Beta–208950

Beaker Unknown Chamber Chamber Chamber Chamber Base of chamber Layer V–1 Base layer of chamber Upper layer (layer III–1) Layer I (top layer) Base of chamber Middle part of chamber Upper part of chamber Layer I (upper part) Blocking deposit of tomb

HB Ch HB HB HB HB Ch Ch Ch HB HB HB HB Ch Ch HB

3720 ± 45 3650 ± 110 4770 ± 40 4640 ± 40 3900 ± 40 3840 ± 40 4465 ± 45 4145 ± 45 4195 ± 45 4095 ± 45 3790 ± 70 4145 ± 45 4095 ± 45 3790 ± 70 3700 ± 120 4180 ± 50

2200–2030 2200–1830 3610–3520 34320–3260 2490–2270 2470–2190 3365–2960 2950–2575 2910–2635 2875–2505 2450–1900 2950–2575 2876–2505 2450–1900 2450–1770 2900–2620

Beta–228578

Blocking deposit of tomb

HB

4110 ± 50

2880–2570

Beta–185649 Beta–185648 Gif–5766 CNA–585

Later burial Later burial Blocking of front of monument Main chamber upper level, female Main chamber lower level, female main chamber basal level, layer below burials

Ch Ch Ch HB

3920 ± 50 3850 ± 40 4470 ± 80 4250 ± 35

2500–2280 2470–2200 3025–2425 2913–2708

HB

4165 ± 30

2819–2659

Ch

4440 ± 60

3195–2923

Ua–40801 Ua–40802 Beta–225171

Small chamber, male Small chamber, female Latest interments in upper layer

HB HB HB

4180 ± 30 4002 ± 31 4050 ± 40

2817–2665 2579–2467 2700–2470

Beta–205946 Beta–234136 KIA–4781

Earliest burials Earliest burials Lowest layer of interments

HB HB Ch

4590 ± 40 4530 ± 40 5050 ± 50

3510–3100 3370–3090 3964–3707

Beta–176897

Chamber. Disturbed burial long bones & teeth Cranium associated with a schist plaque Funerary deposit Cm–5 Funerary deposit Cm–6 Funerary deposit Cm–3 funerary deposit Cm–2 One of the later interments in hypgaeum La Sima phase II. From ossuary in second tholos tomb. Above the lime floor of La Sima I La Sima phase II La Sima phase III. Beaker interment 1 La Sima phase III. Beaker interment 2 Main chamber Additional cist with burial associated with Maritime Beaker vessels Cranium at end of corridor

HB

4290 ± 40

3030–2870

Rojo, Guerra & Garcia 2005 Gonçalves 2003

HB

4270 ± 40

3020–2860

Rojo et al. 2005

HB HB HB HB HB

4170 ± 40 4100 ± 40 3780 ± 40 3770 ± 40 3870 ± 70

2890–2620 2780–2560 2350–2120 2310–2030 2500–2140

Silva 2002

HB

4919 ± 28

3770–3640

HB HB

4862 ± 27 3860 ± 30

3710–3630 2470–2270

HB

3862 ± 28

2470–2270

HB HB

4330 ± 100 4080 ± 80

2814–2675 2704–2555

HB

4340 ± 40

3020–2900

Activity prior to construction of anta Stage 4, prior to closure of entrance to corridor & associated with cluster of finds

Ch

4905 ± 70

3820–3620

Ch

3905 ± 75

2580–2190

GrA–9573 Arquinha da Moura,Tondela

La Atalayuela, Agoncillo, Logrõo Cabeçuda, Marvao, Beira Carrascal, Lisbon Cebolinos, Évora

La Chaussée–Tirancourt, Somme, Paris Basin

Estria, Amadora, Estremadura

Los Gabrieles, Valverde del Camino, Huelva Gavrinis, Morhiban

Montelirio, Castilleja de Guzman, Seville

Pedras da Granja, Vila Verde, Sintra Pedras Grandes, Tagus estuary Peña de la Abuela, Soria

Ua–40803 CNA–589

Beta–166422 Santa Margarida, Reguengos de Monsaraz

São Paulo 2, Almada, Setubal

Beta–176896 Beta–166423 Beta–166418 Beta–166417 UBAR–630 KIA–21551

La Sima, Ambrona velley, Soria

KIA–21552 KIA–17999 KIA–18000

Tres Montes, Navarre

Trigache 2, Lisbon

Beta–239755 Ua–10830

Vale do Rodrigo 2, Alentejo

Gibson_CW3.indd 110

Ua–0831

Material

Cruz et al. 2003 Andrés Rupérez & Barandiarán 2004, 101; Harrison 1998, 464 Oliveira 1997, 1998 Boaventura 2009, 73 Gonçalves 2003

Le Clerc & Masset 2006, 94, table 1

Boaventura 2009, table 22 Boaventura 2009: Table 22 Garcia Sanjuán, 2006; Linares Catela, 2006 Le Roux 1984 Mederos 2013: 2604, table 1

Boaventura & Cardoso 2011, 147 Era-Arqueologia 2001

Rojo et al. 2005

Andrés, García & Sesma 2001, 315 Boaventura 2009, table 22

Larsson 2000

27/05/2016 13:19:04

chapter four

Copper mining, prospection and the Beaker phenomenon in Wales — the significance of the Banc Tynddol gold disc Simon Timberlake

T

discovery of twelve Early Bronze Age copper mines in England and Wales is beginning to change ideas on the importance of local ores, and the role these might have played in shaping the earliest use of metal in these islands (Timberlake 2009; O’Brien 2015). Pivotal in this is the position of Wales, as it is likely to be the first metal to be mined on the British mainland and probably also the location of the earliest identifiable phase of mineral prospection which took place between 2200 and 1850 BC (Timberlake & Marshall 2013).   Most of these mines are located along the western seaboard of Britain (Figure 4.1), with a small but significant concentration on the North Wales coast, centred upon Parys Mountain in Anglesey and Great Orme’s Head near Llandudno; the latter having been worked deeper on account of the abundance of ore and well-drained karstic limestone becoming one of the largest Bronze Age copper mines in Europe (O’Brien 2014, 146). However, 60 miles to the south of here lies a larger group of much smaller mines associated with lead-zinc veins and minor outcrops of copper mineralization occurring within the Plynlimon mining field of Ceredigion and west Montgomeryshire. The final group of sites is to be found in NW-central England at Alderley Edge and Mottram St Andrew in Cheshire (Timberlake & Prag 2005), and Ecton in NW Staffordshire (Timberlake 2014).   The current paper will examine the significance of an Early Bronze Age gold disc belonging to the Primary Bell Beaker Goldwork Tradition (2500–2100 BC) which was found associated with a possible grave cut at Banc Tynddol, Cwmystwyth, Ceredigion in 2002. The site lay at the foot of a copper-bearing mineral vein which was worked during the Early Bronze Age at the point where it outcrops on the top of nearby Copa Hill—the Comet Lode Opencast site here being one of the largest as well as earliest exploited Bronze Age mines within the mid-Wales area (Timberlake 2003). The finding of this disc has been referred to within a number of academic (Timberlake 2002a; 2004a, 137–9; 2009, 102–4; Timberlake, Gwilt & Davis 2004 (Antiquity Project Gallery); Fitzpatrick 2012, 59; Needham & Sheridan 2014, 907–8) as well as popular publications (see Heath 2012, 47; Morgan 2005; Spencer 2004), yet the particular circum­stances of its discovery, its significance, and likely date, have never properly been described.

Timberlake CW3 16 v 2016.indd 111

he

31/05/2016 16:43:19

[ 112 ] iv. Copper mining and the Beaker phenomenon in Wales

4.1 Cobble stone mining tools and Bronze Age mines in Britain and Ireland (B. Craddock)

4.2 Prehistoric mines and mineral veins of Central Wales

Timberlake CW3 16 v 2016.indd 112

31/05/2016 16:43:19

Timberlake

[ 113 ]

Bronze Age mining in Central Wales—Copa Hill, Cwmystwyth, and Plynlimon The largest concentration of Early Bronze Age mines and prospecting trials on the British mainland is to be found within a 25 km radius of Plynlimon in Central Wales; all the sites being linked to surface outcrops of copper ore associated with the lead-zinc-silver mineral veins of the Central Wales orefield (Figure 4.2). Within some of these veins we find small pockets of partly-altered chalcopyrite which form discontinuous sub-economic enrichments, yet the very small number of Bronze Age mines amongst these appear as more localized sub-groups of sites within a much wider spread of copper mineralization. Most of these sites were worked at the points where small but locally enriched pockets of copper mineralization lay closest to the surface, and where the primary sulphide minerals were weathered (Francis 1874; Timberlake 2002, 183). These sub-groups are thought to be examples of ‘prehistoric prospection areas’ (Timberlake ibid.), and consist of the following Bronze Age mines and trials: (1) Dovey (Balkan Hill Aberdovey [SN 613963] (Bowen & Gresham 1963) and Panteidal [SN 661974] (Timberlake 2009b, 6) on the north side of the Dovey Estuary and Ogof Wyddon (Park Lodge) near Machynlleth [SH 760001] (Timberlake & Mason 1997)); (2) Cors Fochno which forms the eastern lowland-upland margin of Borth Bog (Llancynfelin [SN 652951] (Timberlake 2009b,7–8), Pwll Roman [SN 652951] (Timberlake 2009b,8) and Erglodd [SN 657 903] (Timberlake 2006, 82–5)); (3) the Melindwr-Nant Silo-Rheidol valleys east of Penrhyncoch and west of Devil’s Bridge (Twll y mwyn (Cwm Darren) [SN 682 833] (Timberlake 2006, 79–82), Nantyrarian (Blaendyffryn) [SN 705815] (Timberlake 2009b,8–9) and Tyn y fron [SN 724785] (Timberlake 1996, 61–63)); (4) the Upper Ystwyth (Grogwynion [SN 712723] (Thorburn 1987) and Copa Hill Cwmystwyth [SN 811752] (Timberlake 2003); and (5) Plynlimon–River Severn (i.e. Nantyreira (Snowbrook) [SN 826874] (Timberlake 1990a) and Nantyricket mines [SN 865868] located on the banks of the River Severn and its tributary, just a few kilometres to the east of its source (Timberlake 2009b, 9). Both Copa Hill and Nantyreira were true upland sites; both of these located between the 400 m and 500 m (aOD) contours at two of the most isolated mining locations in Wales. All of the sites have been identified from finds of cobble stone mining tools, and whilst most have been radiocarbon dated, some such as Balkan Hill, Panteidal, and Pwll Roman remain to be excavated. Grogwynion which has no copper and is probably a medieval-postmedieval lead mine yielded just one or two stone tools, therefore this site might have been ‘visited’ just once during the course of Bronze Age prospection along the Ystwyth Valley (Timberlake & Marshall 2013) Copa Hill (Copper Hill), Cwmystwyth

The best studied of these mid-Wales Bronze Age mines is the Comet Lode Opencast on Copa Hill (SN 81167523). In some respects this site resembles those of the stone

Timberlake CW3 16 v 2016.indd 113

31/05/2016 16:43:19

[ 114 ] iv. Copper mining and the Beaker phenomenon in Wales

4.3 Photograph taken in the 1990s of the Comet Lode Opencast and prehistoric mine tips on Copa hill, Cwmystwyth (S.Timberlake)

4.4 Roof of mine gallery worked with stone tools within the Comet Lode Opencast, Copa Hill (S.Timberlake)

Timberlake CW3 16 v 2016.indd 114

31/05/2016 16:43:22

Timberlake

axe factories of the Late Neolithic, at least in terms of its weathered outcrop quarries and peat-covered worked stone screes. This is in fact a good example of a small-medium sized Bronze Age mine, one which is also unique in terms of its near-complete survival, and lack of later disturbance. Today more than five thousand tons of prehistoric mine spoil cover the hillside below the outcrop of the over 10m-deep infilled Bronze Age opencast (Figure 4.3). Short fireset galleries with the marks of stone tools in their roofs were found on the north side of the opencast when this was first excavated in the early 1990s (Timberlake 1990b) (Figure 4.4). Within the main opencast, mining was carried out on a series of different levels, with rock benches left in between the minedout veins, and with climbing routes on wooden stemples descending the narrower worked-out fissures. Flooding and then gradual infill of this over 10m-deep mine began around 1800 BC, depositing slumped-in mine spoil and peat within its base, yet the site was only finally abandoned as a mine around 1600–1500 BC. Preserved here within the peat-filled interior are more than 8 m of sediments dating from the Early Bronze Age to the medieval, this providing an almost complete palaeo-environmental record for the mine and surrounding landscape (Figure 4.5). The waterlogged condition of the mine sediments overlying some of the rock floors permitted the survival of a variety of objects including the twisted handles made of hazel withy that once held hammerstones, fragments of the kreel-type baskets used to carry or to wash ore, ropes for lifting bags or for carrying wood, alongside the remains of antler picks and wooden mallets and wedges. When these items were discarded by the Bronze Age miners, some were also thrown onto the fires to be used as fuel to break up the rocks. However, the most sensational find from these excavations was the discovery of a hollowed-out 5m-long log

Timberlake CW3 16 v 2016.indd 115

[ 115 ]

4.5 Archaeological section cut through peaty sediments within the Copa Hill opencast (S.Timberlake 1993)

31/05/2016 16:43:23

4.6 Photographs of mine entrance and excavated Early Bronze Age alder wood launder lying in situ. within Comet Lode Opencast (S.Timberlake)

launder, perhaps one of the earliest examples of mine drainage known. This was one of three such launders excavated, this one being found in situ in its last functional position within the entrance to the mine (Figure 4.6). On the sides could be seen the marks of the metal axes used to carve it out. We now know many specific details about the way they mined, but can only guess at the type of mineral concentrate they were producing. For this reason we have only a hazy idea of the ore grade, but a much better idea of the tonnage of rock extracted (5,000 tons). If we assumed a 1–3% extraction of rich oxide/carbonate concentrate from about 500 tons of well-mineralized dressed vein material, we might then be looking at a recovery of between 2–7 tons of copper metal, though perhaps less (Timberlake 2009a). The earliest dates we have for mining at Cwmystwyth come from the spoil layers present within the Lateral Tip lying upslope of the main Comet Lode Opencast prehistoric dumps. This tip represents the rock and veinstuff removed from the original outcropping vein, formed as this cliff face was worked back with stone tools around 2100 BC (Timberlake 2004b, 139–41). However, there are one or two anomalously early dates from this part of the site which may suggest some sort of Mesolithic–Neolithic interest in pigment extraction at this mineralized outcrop (Timberlake 2003, 55). If this were the case, there may be still further examples of late 3rd millennium (pre-2100 BC) prospection activity close to this outcrop, a factor which may be of some relevance to the discovery and date of the Banc Tynddol gold disc.

Timberlake CW3 16 v 2016.indd 116

10/06/2016 17:24:32

Timberlake

[ 117 ]

The Bronze Age landscape of the Upper Ystwyth Valley This account of the Copa Hill mine is the background context for our understanding of Early Bronze Age mineral prospection and mining activity at the head of the Ystwyth Valley; there being no evidence so far of settlement, prehistoric smelting activity, and little in the way of contemporary burial monuments. The exception to this seems to be the discovery of a damaged kerb cairn and destroyed cist on the plateau of Pant Morcell (SN 80627502) which immediately overlooks Banc Tynddol; a site which lies 600 m away from the Bronze Age mine, and about 40 m below it on the opposite side of the Nant yr Onnen valley. There are also three small quartz boulder cairns of uncertain date upon one of the SW–NE spurs of Copa Hill 400 m to the north of the Comet Lode Opencast (Figure 4.7). 4.7 Survey map of Copa Hill, Cwmystwyth, showing position of EBA mine (green), cairns (red), and site of disc find (blue) (EMRG B. Craddock)

Timberlake CW3 16 v 2016.indd 117

31/05/2016 16:43:38

[ 118 ] iv. Copper mining and the Beaker phenomenon in Wales

Beyond this we are looking at a local landscape with a fairly thin distribution of known or probable Bronze Age monuments; the nearest significant cairns being several kilometres distant capping the summits of Pen y Garn to the north-east and Domen Milwyn to the south, although a group of kerb cairns have been noted some 4 km to the east at Craig y Lluest, and beyond that cairns at Cistfaen and Carnbwlchcloddiau. There are also unconfirmed reports of finds of flint arrowheads beneath the peat at Llyn Gwngu, 3 km to the south-east of the mine. It is possible therefore that the extensive blanket peat which covers the mountain plateau both to the north and south of the deeply incised head of the Ystwyth Valley may yet conceal numbers of low-lying cairns and other Bronze Age monuments. In this respect the various palaeoenvironmental studies carried out within the vicinity of the mine may prove useful in determining the record of past human activity. The Holocene peat within the watershed of the Elan Valley some 5–6 km to the east was looked at by Moore in 1968. He suggested that the fairly long (but undated) pollen record here indicated woodland clearance and regeneration during the Bronze Age, most likely an indication of transhumant agriculture. However, dated peat cores have now been recovered from the blanket peat on Copa Hill just 300–400 m to the north of the Bronze Age mine by Mighall & Chambers (1993) and Mighall et al. (2000a & b). Whilst an earlier date of 3020– 2580 cal BC was obtained from the base of the more distant one, the core with the studied pollen sequence starts at 1890–1680 cal BC during the period of Bronze Age mining. Although this doesn’t cover the prospecting period, the small-scale declines in tree pollen (particularly of Quercus and Corylus) are consistent with the exploitation of the valley side and valley floor woodland on a periodic-seasonal basis for firesetting and for small-scale wood usage within the mine. The palynomorphs, non-arboreal ruderal pollen and low levels of cereal on the other hand suggests the presence of pastoralism and a non-intensive mixed small-scale agricultural regime from at least the early 2nd millennium BC. The picture therefore is probably one of a continuing tradition of transhumance agriculture from the Late Neolithic into the Early Bronze Age, with periodic or seasonal mining taking place over a 400 year period. This evidence for the local exploitation of valley woodland (see Mighall in Timberlake 2003, 66) and for the carriage of fresh hammerstones from the coast (Timberlake 2002b, 184) suggests that the Ystwyth Valley was then open, and used as a persistent pathway into the interior. The landscape around the mine appears to have been seasonally occupied, and used for pastoralism and for hunting (Timberlake 2009a, 101). It was here, on a small plateau at the foot of a really quite dramatic mineral outcrop that the Banc Tynddol disc was deposited, almost certainly as an intentional burial. The Banc Tynddol gold disc The circumstances of its find

The discovery was made on 16th October 2002 during the course of an archaeological excavation by the Early Mines Research Group (EMRG) of a Roman–early medieval lead smelting site. The denuded traces of this former smelting site were spread across some 625 square metres of lead-contaminated grassland growing on thin soils covering the top of

Timberlake CW3 16 v 2016.indd 118

31/05/2016 16:43:39

{

4.8 Banc Tynddol disc detail, scale 30 mm (B.Craddock)

4.9 Banc Tynddol gold disc, scale 30 mm

(photo reproduced by permission of Amgeddfa Cymru – The National Museum of Wales)

an upstanding glacial outwash moraine. The site lay to the north of the valley road, and to the west of the Nant yr Onnen stream at Cwmystwyth (SN 8090 7484) (see Figure 4.7).   The Banc Tynddol disc was found at a depth of only 15 cm below the destroyed site of a 9th–11th century AD lead smelting hearth, and at the time it was uncertain as to whether this had been intentionally deposited. No earlier archaeological features had been identified, and the shallow find-spot suggested that the object may have been disturbed and re-deposited some distance downslope from its original context. It was reported to HM Coroner for Ceredigion as a Treasure find on 30th October 2002 (Timberlake et al. 2003) following its identification as a Copper Age–Early Bronze Age gold foil ‘sundisc’ (Figures 4.8–4.9), otherwise referred to as an early flat gold disc type (S. Needham pers.comm.; Needham & Sheridan 2014, 909). Given the concerns over the possibility of further artefacts remaining in the ground, re-excavation of the site took place in March 2003, with funding provided by the National Museums and Galleries of Wales.   On discovery this object was found lying upside down within a small round patch of green silt a short distance below the turf, yet on re-excavation an oval-shaped 1.75 m

Timberlake CW3 16 v 2016.indd 119

31/05/2016 16:43:46

[ 120 ] iv. Copper mining and the Beaker phenomenon in Wales

"

4.10 Excavation plan and section of the Banc Tyn­ ddol grave, position of disc indicated (B. Craddock)

"

long shallow-cut ‘grave’ was identified almost directly beneath the original findspot (Figure 4.10). Overlying and partly concealing the cut was a thin spread of small boulders originally thought by the excavators to have been gravel moraine (this was subsequently re-interpreted as being the possible remains of a denuded and collapsed bank cairn). Part of the interior of the ‘grave’ appeared to have been robbed or quarried away, an action which may have been contemporary with, but more likely pre-dated the Roman to early medieval lead smelting activity.   In addition to the ‘robbing’ there was evidence for quite extensive animal burrowing, the soil conditions also being a little too acidic for the preservation of skeletal bone. However, fragments of bone, possibly human were found, the samples yielded insufficient collagen for dating purposes, and being inconclusive in terms of phosphate. Yet, the location of the disc towards the top of the ‘grave’ fill places this item almost directly above what is most likely to have been the upper torso or chest area of an interned individual; this is assuming of course that there was formerly a body laid out in a partly flexed position, with the head looking south-west directly down valley. The cut of this feature, if indeed it was a grave, seems to have been aligned on the narrow ‘V-cut’ end to the western horizon. More perplexing though was the complete absence of any other artefactual evidence, given that pottery had been expected. Thus whilst the shape of this feature appears to be recognizable as a grave, the disturbed contents remain something of an enigma.

Timberlake CW3 16 v 2016.indd 120

31/05/2016 16:43:48

Timberlake

[ 121 ]

  Three samples of charcoal were recovered from the grave fill; one of these pieces suggesting Mesolithic activity associated with the pre-burial land surface (8210–7760 cal BC; OxA–12983: 8850+40), another a Late Roman date for the ground surface covering the edge of the grave (320–430 cal AD; OxA–12955: 1675+28), whilst the third consisted of a single large piece of oak charcoal derived from the early medieval smelting horizon which lay directly above the burial (670–840 cal AD; OxA-12956: 1264+27) (Timberlake 2004a, 137–9). There seems little doubt that the latter piece was intrusive to the grave, since this was associated with an old animal burrow which clearly pre-dated the smelting, yet contained fragments of charcoal and small pieces of metallurgical debris washed down into it. Whilst not providing us with any definitive answer as to the date of the feature, these results do at least support the likelihood of this being a prehistoric feature, with the possibility of this being a burial. Its manufacture and composition

Description of the disc In 2003 an examination of the gold foil disc (38.9 mm diameter; Central Wales > North Wales > NW England) matches quite closely the route of the ‘Beaker Folk Penetration’ mapped by Lal Chitty and Christopher Hawkes more than 60 years ago; one which links East Anglia with the Peak District and Cheshire Plain, with North Wales and Anglesey, and with the middle of Wales through a route linking the Dovey Valley and the coast of Ceredigion (Varley 1964). Of course this route was shown going the other way, and the dates are all wrong, yet there is some similarity here in the overall pattern, which emphasises the use of the same well-worn paths and sea-ways, and hints at the metallurgical/mining interests of those who brought with them the ‘Beaker package’.   Aubrey Burl back in 1976 made the link between the movement of ‘Beaker peoples’ (what we would now refer to as the ‘Beaker package’) into Wales, and the presence of stone circles in Pembrokeshire. Alongside this was seen the evidence for burials, the re-use of sites, and of course bluestone extraction in the Preselis; the physical evidence for which has now been suggested by recent excavations on Carn Menyn (Darvill et al. 2005) and at Craig Rhos y Felin (Parker Pearson 2013). Significantly Burl suggested a link between the Preselis bluestone and ‘Beaker Period’ metal prospectors travelling by boat down the coastline of Wales, and around the headlands off Pembrokeshire. Other possible links between Wessex and Wales have now been suggested by isotope studies carried out on bone from the skeletons of the Boscome Bowmen burials. The isotope data obtained from the latter suggested that these individuals may have lived for at least part of their adult lives within a landscape dominated by Lower Palaeozoic rocks—the nearest source of these being Wales (Chenery & Evans 2011, 184-187; Fitzpatrick 2012, 49). However, whether or not these and other similar individuals were hunters, begetters of bluestone or metallurgists, their presence here may well have acted as a conduit, catalyst or source of economic motivation for the earliest exploitation of copper.

Timberlake CW3 16 v 2016.indd 130

31/05/2016 16:43:56

[ 131 ]

Timberlake =England =nort Wales =mid Wales Phase Beginnings boundary end early mining Last early mining Prior end Alderley Edge Prior start Alderley Edge Prior end SQM Prior start SQM Prior end the Lumb Prior start the Lumb First England Phase England Prior end Parys Prior start Parys Prior end Great Orme? Prior start Great Orme First north Wales Phase north Wales Prior end Copa Hill Prior start Copa Hill R_Date BM-2538 After Layer (4) R_Date BM-2581 After Layer (4) Sequence Nantyreira Mine R_Date Beta-214364 After Erglodd Mine R_Date Beta-214364 After Twll-y-mwyn Mine R_Date Beta-120592 AfterTyn-y-fron Mine R_Date BM-2930 AfterNatyreira Mine R_Date BM-2916 AfterLlancynfelin Mine R_Date BM-3146 Phase Ogof Wyddon (Park Lodge) First mid Wales Phase mid Wales First early mining Phase Boundary start early mining Sequence early mining 4000

3000

2000

1000

cal BC/cal AD

Posterior Density Estimate (cal BC/cal AD)

=England =north Wales =mid Wales Phase Beginnings

2750

2500

2250

2000

1750

Posterior Density Estimate (cal BC)

4.17 Bayesian modelled 14C dates for the UK. The lower figure shows the probable phase beginnings in mid Wales, north Wales, and England. (P. Marshall)

Timberlake CW3 16 v 2016.indd 131

31/05/2016 16:43:56

[ 132 ] iv. Copper mining and the Beaker phenomenon in Wales

  Whilst this may be true, caution must still be applied to the certainty with which it is believed that these individuals were metalworkers, as has been claimed for the Amesbury Archer and more recently the individual associated with the Kirkhaugh burial (a site recently re-excavated in the northern Pennines). In both cases these interpretations were based upon accompanying finds of cushion stones (Fitzpatrick 2102, 52; 2014, 6; Fitzpatrick forthcoming). It is probably true that these were used for hammering and burnishing metal, yet anyone possessing a knife or axe might have needed to do this. The maintenance of a tool may have necessitated from time to time the hammering out of nicks on the blade, edge hardening this as part of the sharpening process, and also regular burnishing to help enhance this as a status item. Putting it another way, saddle querns and whetstones were both commonly used domestic items from the Neolithic onwards, yet the possession of these did not make the owners of these items millers or cutlers; the use of these was simply part of daily life for self-reliant people. For sure we cannot tell from the sets of artefacts accompanying these male Beaker burials whether the individuals concerned were ever involved in casting metal, or for that matter the procurement of metal through smelting, mining, or prospecting (see Fitzpatrick 2014: ‘..I think the man buried at Kirkhaugh was part of a small group that was prospecting for copper over 4000 years ago’). Our best hope in this case would be to uncover and excavate an undisturbed wellpreserved burial of this type lying close to (and associated with) a contemporary mining, smelting or metallurgical site. In this respect the only partial survival of evidence from Cwmystwyth is unfortunate, yet the find clearly does raise a number interesting questions. Copper Age/Early Bronze Age incomers to Central Wales and their links with mining

The presence of small hunting groups passing through the mountainous mineral landscapes of mid-Wales is suggested by the very considerable numbers of finds of ‘Beaker-type’ barbed-and-tanged flint arrowheads (more than 70) on the Plynlimon moorlands around Llyn Bugeilyn in Montgomeryshire (Peate 1925a; 1925b; Savory 1980; CPAT SMR 2002). All of these arrowheads turned up as chance finds and loose scatters, which suggests that the original losses were made during hunting expeditions. Many were described as being of ‘Irish affinity’, a type rare in Britain, but with parallels in counties Antrim, Down, and Mayo (Peate 1928). In fact the actual distribution of these arrowheads within the uplands is a good deal wider than this, with find spots ranging from Dylife to Cwmystwyth, with many of them being found within sheltered moorland areas surrounding some of the natural tarns, perhaps originally clearings within the wooded Bronze Age landscape where red deer congregated to drink. The presence of these animals and the hunting of them is likewise suggested by the abundant fragments of antler and several well-preserved antler tools found within the mines at Cwmystwyth (Timberlake 2003, 84–5) and Nantyreira (Timberlake 1990, 18). For example, it has been estimated that 100–300 antler tools may have been consumed at Cwmystwyth during the working life of this mine (Timberlake 2009b, 102).   The ‘Beaker’ arrival and the connection of these incomers with the early mines may have been limited to the very first wave of prospection around 2200–1900 BC. If there was a

Timberlake CW3 16 v 2016.indd 132

31/05/2016 16:43:56

Timberlake

[ 133 ]

role for the use of ‘middle-men’ or metallurgists in the initiation of mining, this is likely to have been replaced by others during the main period of copper production that took place between 1900–1600 BC.   I think it is unlikely, given what we now know of some of the better studied mines, that metallurgists were ever directly involved in the working of these sites. Instead the periodic or seasonal mining which took place was probably the role of transhumant pastoralists with traditional grazing rights in these mountain areas; marginal peoples living on marginal lands, but who nevertheless would have related and traded with the incomer hunters and metalworkers who passed through their territories (Timberlake 2002, 184). Therefore the trade in metal (or perhaps just ore concentrate) may simply have been the means by which these subsistence peoples entered into an exchange economy (see Shennan 1999; Timberlake 2003). The incomers with their ‘Beaker package’ will probably have remained on the outside of most of the established territory-holding lineal groups (Parker Pearson 1993), and for this reason their itinerant passage into the upland grazing and hunting landscapes of Wales for reasons of metal procurement would not have been that surprising. However, little in the way of associated artefacts or burial monuments survive within the area of the mines. Yet in terms of the contemporary landscape, we find a number of these sites located on the margins of tree-cleared areas in which burial cairns dominate the highest points. It is interesting to note as well that between Penrhyncoch and the upper headwaters of the River Severn there is an intermittent line of standing stones and other Bronze Age monuments which appears to define a territorial divide or routeway skirting the northern slopes of Plynlimon, passing close to a number of the earliest copper mines, such as Twll y Mwyn (Darren), Nantyreira, and Nantyrickets (Timberlake 2002, 186; Figure 4.18). Not far from this same route we also find several of the earliest Beaker burials, one associated with a Beaker and two flint arrowheads at Banc Troedrhiwseiri (Briggs 1994, 147) and another near Plas Goggerdan, Penrhyncoch. Apart from the Banc Tynddol disc, the only metal artefact of this period associated in any way with the mines was a copper halberd made of A1 (Irish or ‘Ross Island type’ metal; see O’Brien 2004; Needham 2002) found ‘near a copper-mine at Pontrhydygroes’ (near Cwmystwyth) sometime during the 19th century (Savory 1980; Briggs 1994,155). We know very little about the circumstances of this find, except that it provides us with a possible link between these early Welsh mines and the metalwork of Ireland, and most probably dates to the end of the 3rd millennium BC. Concluding comments So what can the discovery of the disc tell us? We can be pretty certain of its date and cultural context, and I think also the likelihood of it being associated in some way with the beginning of mining on Copa Hill. Perhaps the link was with an individual who was either a prospector, or who had an interest in procuring copper or lead ores for metallurgy. The association with the ‘Bell Beaker package’ however is clear here, as it was at the Ross Island mine, although most of the dates from the working of the Cwmystwyth mine are later. What we are probably looking at therefore is the likely ‘signature’ of its discoverer.

Timberlake CW3 16 v 2016.indd 133

31/05/2016 16:43:56

[ 134 ] iv. Copper mining and the Beaker phenomenon in Wales

Flint arrowhead finds

4.18 Early Bronze Age monuments and finds in the Plynlimon area, including early mines and flint ­arrowhead finds (S. Timberlake & B. Craddock)

  For some years now we have collectively accepted the ‘out of Ireland’ hypothesis for the origins of copper mining, smelting and metallurgy in these islands. In some respects this is not an unreasonable idea, given the much later dates we have for the mainland mines, alongside the circulation of A1-type Ross Island metal across Britain, followed by its re-cycling and later mixing with local copper (Bray & Pollard 2012). Yet some caution

Timberlake CW3 16 v 2016.indd 134

31/05/2016 16:44:00

Timberlake

[ 135 ]

should be applied to this model. The recent Bayesian modelling of British Bronze Age mining dates suggests that the West Wales mines were probably exploited earlier than we had previously thought, and certainly prior to the exhaustion of the Ross Island deposit. If we then take a look at the evidence for the anomalously early dates from Copa Hill which were excluded from this analysis, we see a deposit first discovered and worked at outcrop during the pre-metal age. It is hardly surprising therefore that some of the initiators of mining and metallurgy in Britain at the time of the Amesbury Archer and Boscombe Bowmen knew of these deposits, and visited them.   More importantly, the Banc Tynddol and other early gold disc types from Britain are quite different from those found in Ireland, their simple concentric linear and dot repoussé decoration being typical of the Atlantic fringe to the east, the immediate links being with Amorica and Iberia rather than central Europe, and with a Bell Beaker tradition of decoration that we see typically on basket ornaments and other small items of Chalcolithic gold. These are copied designs, yet it seems increasingly likely that the gold used was British, and obtained from local alluvial sources; some perhaps made from the gold recovered from the Cornish tin streams, and others from the gold in the rivers draining copper-rich prospecting areas of Wales.   We cannot be certain whether it was groups of prospectors or just the knowledge of ores and metalworking which travelled, yet in all probability the re-exploitation of the supergene copper ore deposits of Britain at the end of the 3rd millennium BC using stone, bone and antler tools was linked in some small way to the arrival of the Beaker cultural package from the continent.   For this reason the discovery of the disc and potentially also a grave found close to the foot of this major copper vein mined at the beginning of the Bronze Age is immensely important in helping to understand the possible role ‘Beaker’ metalworkers had in the opening-up and working of these deposits following their arrival in the metal-rich areas of western Britain at the end of the 3rd millennium BC.   Acknowledgements I would like to express my thanks to Adam Gwilt (Later Prehistorian, National Museums and Galleries of Wales (NMGW)) for allowing me to use the information he had researched on other examples of these flat gold discs for the unpublished Treasure Act report written in 2003. Some of his material has been used in this paper, but in an updated form, though any mistakes of interpretation are likely to be mine. The National Museum also permitted the reproduction of the accompanying photograph, the object itself currently residing within the collections of the Department of Archaeology and Numismatics of the NMGW. Both the NMGW and the Cambrian Archaeological Association supported the relevant fieldwork. Brenda Craddock of the Early Mines Research Group was responsible for the drawings, and I am grateful to her and the other members of the group for their skills in excavation which resulted in its discovery, and the subsequent re-examination of the site. Mary Raw of Ty Llwyd Farm, Cwmystwyth gave permission for us to excavate on her land, whilst CADW gave us permission to excavate on a scheduled ancient monument.

Timberlake CW3 16 v 2016.indd 135

31/05/2016 16:44:01

[ 136 ] iv. Copper mining and the Beaker phenomenon in Wales

BIBLIOGRAPHY Aldhouse-Green, S. & J. P. Northover 1996 ‘Note: Recent Finds of Late Bronze Age Gold from South Wales’, Antiquaries Journal 76, 223–8. Annable, F. K. & D. D. A. Simpson 1964 Guide Catalogue of the Neolithic and Bronze Age Collections in Devizes Museum. Devizes, Wiltshire Archaeological & Natural History Soc. Armstrong, E. C. R. 1920 Guide to the Collection of Irish Antiquities: Catalogue of Irish Gold Ornaments in the Collection of the Royal Irish Academy. Dublin, H.M.S.O. Bowen, E. & C. Gresham 1967 History of Merioneth Vol. 1, p. 55 (quote from R. Fenton’s ‘Tours in Wales 1804– 1813’ re-published as a Supplement in Archaeologia Cambrensis (1917)) Bray, P. J. & A. M. Pollard 2012 ‘A New Interpretative Approach to the Chemistry of Copper-Alloy Objects: Source, Recycling and Technology’, Antiquity 86, 853–67. Briggs, C. S. 1994 ‘The Bronze Age’, Cardiganshire County History; Volume I, From the Earliest Times to the Coming of the Normans, ed. J. L. Davies & D. P. Kirby. Cardiff, University of Wales Press. Burl, A. 1976 The Stone Circles of the British Isles. Newhaven & London, Yale University Press, Cahill,M. 2006 ‘John Windele’s Golden Legacy —Prehistoric and Later Gold Ornaments from Co.Cork and Co. Waterford’, Proceedings Royal Irish Academy 106C, 219–137. Case, H. 1977 ‘An Early Accession to the Ashmolean Museum’, Ancient Europe and the Mediterranean; Studies presented in honour of Hugh Hencken, ed. V. Markotic, 18–34. Warminster, Aris & Phillips. Chenery, C. A. & J. A. Evans 2011 ‘A summary of the strontium and oxygen isotope evidence for the origins of Bell Beaker individuals found near Stonehenge’, The Amesbury Archer and the Boscombe Bowmen. Bell Beaker burials at Boscombe Down, Amesbury, Wiltshire. Excavations at Boscombe Down volume 1, Wessex Archaeology Report 27, ed. A. P. Fitzpatrick. Salisbury, Wessex Archaeology. Clarke, D. L. 1970 Beaker Pottery of Great Britain and Ireland (2 vols). Cambridge, Cambridge University Press. Clarke, D. V., T. G. Cowie, & A. Foxon 1985 Symbols of Power; At the Time of Stonehenge. Edinburgh, National Museum of Antiquities of Scotland/H.M.S.O. Darvill, T., D. Evans, R. Fyfe, & G. Wainwright 2005 ‘Strumble-Preseli Ancient Communities and Environment Study (SPACES): Fourth Report

Timberlake CW3 16 v 2016.indd 136

2005’, Archaeology in Wales 45, 17–25 Eluère, C. 1982 L’âge du bronze en France –2: Les Ors Préhistoriques. Paris, Picard. Eogan, G. 1994 The Accomplished Art; Gold and GoldWorking in Britain and Ireland during the Bronze Age (c. 2300–650 BC), Oxbow Monograph 42. Oxford, Oxbow Books. Fitzpatrick, A. P. 2009 ‘In his Hands and in his Head: the Amesbury Archer as a metalworker’, Bronze Age Connections—Cultural Contact in Prehistoric Europe, ed. P. Clark, 176–87. Oxford & Oakville, Oxbow Books. Fitzpatrick, A. P. 2011 The Amesbury Archer and the Boscombe Bowmen. Bell Beaker burials at Boscombe Down, Amesbury, Wiltshire. Excavations at Boscombe Down volume 1, Wessex Archaeology Report 27. Salisbury, Wessex Archaeology. Fitzpatrick, A. P. 2012 ‘The Arrival of the Bell Beaker Set in Britain and Ireland’, Celtic from the West 2: Rethinking the Bronze Age and the arrival of Indo-European in Atlantic Europe, ed. J. T. Koch & B. Cunliffe, 41–70. Oxford, Oxbowbooks. Fitzpatrick, A. P. forthcoming ‘Stone Tools and Early Bell Beaker Metalworking in Britain and Ireland’ (2012 draft) Frances, A. 1874 History of the Cardiganshire Mines from the Earliest Date, and the Authenticated History to AD 1874, with their Present Position and Prospect. Goginan, Aberystwyth (re-printed in 1987 by Mining Facsimiles no.14, Sheffield) Gage, J. 1835–6 ‘A Letter from John Gage, Esq. F.R.S. Director, to Sir Henry Ellis, K.H. F.R.S., Secretary, accompanying a Gold British Corselet exhibited to the Society, and since purchased by the Trustees of the British Museum’, Archaeologia 26, 422–3. Gerloff, S. 1975 The Early Bronze Age Daggers in Great Britain and a Reconsideration of the Wessex Culture, Prähistorische Bronzefunde, Abt. VI.2. München, C. H. Beck’sche Verlagsbuchhandlung. Hartmann, A. 1970 Prähistorische Goldfunde aus Europa. Berlin, Gebr. Mann Verlag. Hartmann, A. 1980 ‘Appendix 3: Analyses by A. Hartmann of British Prehistoric Gold and some British Ores’, Bronze Age Goldwork of The British Isles, ed. J. J. Taylor, 138–41. Cambridge, Cambridge University Press. Heath, J. 2012 Life in Copper Age Britain. Stroud, Glos., Amberley. Hemp, W. J. 1917–8 ‘Brynkir Station Group’, Proceedings of the Society of Antiquaries of London 30 (2nd series), 177–83 & Fig. 5.

31/05/2016 16:44:01

Timberlake

Kinnes, I., A. Gibson, J. Ambers, S. Bowman, M. Leese, & R. Boast 1991 ‘Radiocarbon Dating and British Beakers: The British Museum Programme’, Scottish Archaeological Review 8, 35–68. Lehrenberger, G. 1995 ‘The Gold Deposits of Europe: An Overview of the Possible Metal Sources for Pre­his­toric Gold Objects’, Prehistoric Gold in Europe; Mines, Metallurgy and Manufacture, ed. G. Morteani & J. P. Northover, 115–44. London, Kluwer Academic Publishers. Maryon, H. 1936 Excavation of Two Bronze Age Barrows at Kirkhaugh, Northumberland, Archaeologia Aeliana 13 (4th series), 207–17. McArdle, P., J. H. Morris, & P. R. R. Gardiner 1987 ‘Gold in Ireland: the potential for mineralization’, Geological Survey of Ireland Report Series RS 87/1. Mighall, T. & F. Chambersm 1993 ‘The Environmental Impact of Prehistoric Mining at Copa Hill, Cwmystwyth, Wales’, The Holocene 3, 260–4. Mighall, T., S. Timberlake, J. Grattan, & S. Forsyth 2000a B’ronze Age Lead Mining at Copa Hill, Cwmystwyth—fact or fantasy?’, Historical Metallurgy 34 (1), 1–12 Mighall, T., S. Timberlake, J. Grattan, & S. Forsyth 2000b ‘Tracing Atmospheric Metal Mining Pollution in Blanket Peat’, Tracers in Geomorphology, ed. D. L. Foster, J. Wiley, 101–21. Moore, P. D. 1968 ‘Human Influence upon Vegetation History in North Cardiganshire’, Nature 217, 1006–9. Morgan, G. 2005 Ceredigion: A Wealth of History. Llandysul, Ceredigion Gomer Press,. Needham, S. 1996 ‘Chronology and Periodisation in the British Bronze Age’, Acta Archaeologica 67, 121–40. Needham, S. 2000 ‘The Development of Embossed Goldwork in Bronze Age Europe’, Antiquaries Journal 80, 27–65. Needham, S. 2002 ‘Analytical implications for Beaker metallurgy in North-West Europe’, The Beginnings of Metallurgy in the Old World: Forschungen zur Archaeometrie und Altertumswissenschaft 1, ed. M. Bartelheim, E. Pernicka, & R. Krause, 99–134. Rahden, Westfalia Institut fur Archaometrie, Freiberg, Verlag Marie Leidorf Gmbh, Needham, S. 2007 ‘Bronze Makes a Bronze Age? Considering the systemics of Bronze Age metal use and the implications of selective deposition’, Beyond Stonehenge: Essays in Honour of Colin Burgess, ed. C. Burgess, P. Topping & F. Lynch, 278–87. Oxford, Oxbow Books. Needham, S. 2011 ‘Gold Basket-Shaped Ornamemts from Graves 1291 (Amesbury Archer) and 1236’,

Timberlake CW3 16 v 2016.indd 137

[ 137 ]

The Amesbury Archer and Boscombe Bowmen: Bell Beaker Burials on Boscombe Down, Amesbury, Wiltshire, Wessex Archaeological Report 27, ed. A. P. Fitzpatrick, 129–38. Salisbury, Wessex Archaeology. Needham, S. & A. Sheridan 2014 ‘Chalcolithic and Early Bronze Age Goldwork from Britain: new finds and new perspectives’, Metals of power—Early gold and silver : Tagungen des Landesmuseum für Vorgeschichte Hale 2014 Band 11/1, H. H.Meller, R. Risch, & E. Pernicka, , 903–29. Halle (Saale) Northover, J. P. 1995 ‘Bronze Age Gold in Britain’, Prehistoric Gold in Europe; Mines, Metallurgy and Manufacture, ed. G. Morteani & J. P. Northover, 515–31. London, Kluwer Academic Publishers. Northover, J. P. 1999 ‘Analysis of Early Bronze Age Metalwork from Barrow Hills’, Excavations at Barrow Hills, Radley, Oxfordshire: The Neolithic and Bronze Age Monument Complex Volume I (Thames Valley Landscapes Vol. 11), ed. A. Barclay & C. Halpin (eds.), 192–5. O’Brien, W. 2004 Ross Island—Mining, Metal and Society in Early Ireland, Bronze Age Studies 6. National University of Ireland, Galway. O’Brien, W. 2015 Prehistoric Copper Mining in Europe 5500–500BC. Oxford, Oxford University Press. Parker-Pearson, M. 1993 Bronze Age Britain. London, Batsford/English Heritage. Parker-Pearson, M. 2013 ‘Researching Stonehenge: Theories Past and Present’, Archaeology International 16, 72–83. Peate, I. 1925a ‘Arrow-heads from Bugeilyn’, Archaeologia Cambrensis, 80 (1), 196–202. Peate, I. 1925b ‘Arrow-heads from Bugeilyn’, Archaeologia Cambrensis, 80 (2), 415–416. Peate, I. 1928 More arrow-heads from Bugeilyn, Archaeologia Cambrensis 83 (2), 344–345 Penhallurick, R. 1986 Tin in Antiquity. London, Institute of Metals. Piggott, S. 1958 ‘Knowes of Trotty (GB 33)’, Inventaria Archaeologica, Great Britain (5th set: GB.25–34), ed. S. Piggott & M. Stewart. Garraway, London. Powell, T. G. E. 1953 ‘The Gold Ornament from Mold, Flintshire North Wales’, Proceedings of the Prehistoric Society 19, 161–79. Russel, A. D. 1990 ‘Two Beaker Burials from Chilbolton, Hampshire’, Proceedings of the Prehistoric Society 56, 153–72. Savory, H. N. 1980 Guide Catalogue of the Bronze Age Collections. Cardiff, National Museum of Wales. Shennan, S. 1999 ‘Cost, Benefit and Value in the Organization of Early European Copper Oroduction, Antiquity 73 (280), 352–63. Sherratt, A. 1986 ‘The Radley “Earrings” Revisited’, Oxford Journal of Archaeology 5(1), 61–6.

31/05/2016 16:44:01

[ 138 ] iv. Copper mining and the Beaker phenomenon in Wales Simmonds, A. 2012 ‘A Sun-Disc from Tubney Wood, Oxfordshire’, PAST (Newsletter of the Prehistoric Society) 70, 11–12, Spencer, K. 2004 ‘Out of the Wild—Cwmystwyth’, Walking Wales Issue 2(2004), 26–9 Standish, C. D., Dhuime, C. J. Hawkesworth, & A. W. G. Pike 2014 ‘New Insights into the Source of Irish Chalcolithic and Early Bronze Age Gold through Lead Isotope Analysis’, Metals of power—Early gold and silver : Tagungen des Landesmuseum für Vorgeschichte Hale 2014 Band 11/1, ed. H. H. Meller, R. Risch, & E. Pernicka, 209–22. Halle (Saale). Taylor, J. 1980 Bronze Age Goldwork of the British Isles. Cambridge, Cambridge University Press. Taylor, J. 1994 ‘The Oliver Davis lecture: The first golden age of Europe was in Ireland and Britain (circa 2400–1400 BC)’, Ulster Journal of Archaeology 57, 37–60. Thorburn, J. 1990 ‘Stone Mining Tools and the Field Evidence for Early Mining in Mid-Wales’, Early Mining in the British Isles, Plas Tan y Bwlch Occasional Paper 1, ed. P. Crew & S. Crew, 43–6. Blaenau Ffestiniog. Timberlake, S. 1990a ‘Excavations at Parys Mountain and Nantyreira’, Early Mining in the British Isles, Plas Tan y Bwlch Occasional Paper 1, ed. P. Crew & S. Crew, 15–21. Blaenau Ffestiniog. Timberlake, S. 1990b ‘Excavations and fieldwork on Copa Hill, Cwmystwyth, Dyfed, 1989’, Early Mining in the British Isles, Plas Tan y Bwlch Occasional Paper 1, , ed. P. Crew & S. Crew, 22–9. Blaenau Ffestiniog. Timberlake, S. 1996 ‘Tyn y fron Mine’, Archaeology in Wales 36, 60–3. Timberlake, S. 2002a ‘Cwmystwyth, Banc Tynddol’, Archaeology in Wales 42, 97–98. Timberlake, S. 2002b ‘Mining and Prospection for Metals in Early Bronze Age Britain—making claims within the archaeological landscape’, Bronze Age Landscapes: Tradition and Transformation, ed. J. Brück, 179–192. Oxford, Oxbow Books. Timberlake, S. 2003 Excavations on Copa Hill, Cwmystwyth (1986–1999): An Early Bronze Age copper mine within the uplands of Central Wales, British Archaeological Reports (British Series) 348. Archaeopress, Oxford. Timberlake, S. 2004a Banc Tynddol Beaker Gold Disc, Archaeology in Wales 44, 137–138. Timberlake, S. 2004b ‘Comet Lode Opencast, Copa Hill, Cwmystwyth’, Archaeology in Wales 44, 139–42. Timberlake, S. 2006 ‘Excavations of Early Mineworkings at Twll y Mwyn (Cwm Darren) and Erglodd, Ceredigion’, Archaeology in Wales 46, 79–86. Timberlake, S. 2009a ‘Copper Mining and Metal

Timberlake CW3 16 v 2016.indd 138

Production at the Beginning of the British Bronze Age’, Bronze Age Connections, ed. P. Clark, 95–122. Oxford, Oxbow Books. Timberlake S. 2009b ‘The Origins of Metal Mining in Britain: the exploration and archaeological excavations of the Early Mines Research Group in Central Wales’, Welsh Mines and Mining no.1, ed. D. J. Linton, 3–16. Llanaber, Gwynedd. Timberlake, S. 2014 ‘Prehistoric Copper Extraction in Britain: Ecton Hill, Staffordshire’, Proceedings of the Prehistoric Society 80, 159–206. Timberlake, S., A. Gwilt, & M. Davis 2003 ‘Treasure Act 1996 (Case 02.11—Wales) A Copper Age/ Early Bronze Age gold disc from Penguelan, Cwmystwyth, Ceredigion’ (unpublished report) Timberlake, S., A. Gwilt, & M. Davis 2004 ‘A Copper Age/Early Bronze Age gold disc from Banc Tynddol (Penguelan, Cwmystwyth Mines, Ceredigion)’, Antiquity 78 (302) (December 2004) Timberlake, S. & P. Marshall 2013 ‘The Beginnings of Metal Production in Britain: A new light on the exploitation of ores and the dates of Bronze Age mines’, The Origins of Metallurgy in Europe, Historical Metallurgy 47(1),ed. P. Craddock, 93–109. Timberlake, S. & J. Mason 1997 ‘Ogof Wyddon (Machynlleth Park Copper Mine), Machynlleth’, Archaeology in Wales 37, 62–5. Timberlake, S. & J. Prag (eds.)2005 The Archaeology of Alderley Edge—Survey, Excavation and Experiment in an Ancient Mining Landscape, British Archaeological Reports 396. Oxford, John and Erica Hedges. Underwood, G. 1946 Early British Settlement at Farleigh Wick and Conkwell, Wilts., The Wiltshire Archaeological & Natural History Magazine 51, No. 185, 440–52 Underwood, G. 1947 ‘Note: Farleigh Wick’, The Wiltshire Archaeological & Natural History Magazine 52, No. 188, 270–1 Varley, W. 1964 Cheshire before the Romans, History of Cheshire series 1. Chester, Cheshire Community Council and C Tinley & Co. Varndell, G. 2001 ‘Longbridge Deverill, Wiltshire: ?Early Bronze Age gold disc’, Treasure Annual Report 1998–1999. Department Culture Media & Sport 2000, 10, Cat. 2. Warner, R. 1993 ‘Irish Prehistoric Goldwork: a provisional analysis’, Archaeomaterials 7, 101–13. Warner, R., N. Moles, R. Chapman, & M. Cahill 2010 ‘The Mournes: a source of early Bronze Age tin and gold’, Archaeology Ireland 24, 18–21.

31/05/2016 16:44:01

chapter five

Burial practices in Ireland during the late third millennium BC connecting new ideologies with local expressions

Kerri Cleary

I

long seen as on the westernmost edge of Atlantic Europe, has been occupied from at least the 8th millennium BC. Since that time, the movement of people, animals, artefacts, technologies and ideologies to and from the island have influenced the formation of the population, their landscape and the networks of contact that emerged. By exploring the archaeological evidence for such mobility throughout the ‘Metal Ages’ we can begin to map periods of connectivity and disconnectivity and create a better understanding of how Ireland’s relationships with other areas of Atlantic Europe changed over time. The importance of the island in the prehistory of western Europe has long been established, from the exploitation of porcellanite for axe manufacture in the Neolithic to the mining of copper from at least c. 2400 BC onwards and the crafting of exquisite gold objects throughout the Bronze Age; all of which placed Ireland not on the edge but bound to the other communities of the Atlantic seaboard and beyond.   Burial practices are one aspect of the archaeological record long used to identify spheres of influence, whereby evidence for interaction and integration can be inferred from changing traditions. ‘In most regions fluctuations between periods of relatively standardized burial customs and times of considerable local variation can be recognized’ (Holst 2013, 104) and by examining such practices we aim to untangle these webs of connectivity. The inclusion of ‘exotic’ artefacts in burials, for example, may hint at external connections but the biographies of those artefacts also need to be considered. As highlighted by Leary (2014, 4), when the movement of both objects and people are discussed, emphasis tends to be placed on where they came from or ended up, while the movements in between are rarely discussed because they are generally intangible. Although this is problematic from an analytical perspective, it does not preclude archaeologists from theorizing about ways in which people and objects moved across considerable distances during a time when travelling was difficult and most likely sporadic. The much debated adoption of the ‘Bell Beaker culture’, traditionally most intensively displayed through individual burials, stereotypically as high status ‘warriors’ or ‘travelling specialists’, is a perfect example and this paper will therefore explore the ways in which Ireland absorbed and added to the changing traditions underway in other parts of Atlantic Europe from the early to mid-3rd millennium BC onwards.

Cleary CW3.indd 139

reland,

27/05/2016 14:56:44

[ 140 ] v. burial practices during the late third millennium bc

Crossing Seas Before we consider what the burial evidence can tell us about long-distance connections, we must first examine the ability of individuals and groups to make these journeys to and from Ireland. Despite a lack of direct archaeological evidence for suitable seagoing vessels, the colonization of the island, surviving material culture and introduced species of animals and plants indicate that people had the ability to travel cross-channel and over considerable distances from an early date. Ireland is clearly visible across the Irish Sea from parts of Britain, notably from north Wales towards south Leinster and from south-west Scotland or northern England and the Isle of Man to east Ulster (Breen & Forsythe 2004, 28) and also looking the other way, from Fair Head in Co. Antrim eastwards across the North Channel to Kintyre and northwards to Islay on the west coast of Scotland (Brett 2009, 30). The capability to cross seas during the Bronze Age has been demonstrated by experimental voyages with replicas of the boats discovered at Ferriby and Dover on the eastern and southeastern coast of England, which suggested that timber sewn-plank vessels probably provided the main means for sea travel over several millennia (Coates 2005). Although comparable sewn-plank boats have yet to be found in Ireland, the recent discovery of a logboat with sewn-plank repairs at Lee’s Island, Lough Corrib in Co. Galway, indicates that knowledge of that boat building technique was already firmly established by the Middle Bronze Age (K. Brady 2014a, 13; 2014b, 37).   In contrast to the lack of sewn-plank boats, there are numerous logboats from Ireland and of those dated, 15 are from the Late Neolithic to the Late Bronze Age (Figure 5.1; Lanting & Brindley 1996; K. Brady 2014a; Niall Brady pers. comm.). The tradition of using these logboats on loughs and rivers is well-attested, such as that from a bog1 at Cuilmore, Co. Mayo, radiocarbon dated to 1914–1517 cal BC2 and another, found upside down in the bed of a small stream at Ballyvogham, Co. Limerick, radiocarbon dated to 1643–1504 cal BC (Table 5.1; Lanting & Brindley 1996, 87–8). Perhaps the most impressive example is from a bog (previous lake) in Lurgan, Co. Galway; it is over 15 m long and 1 m wide, hollowed from a single oak tree and radiocarbon dated to the Chalcolithic (2561–2345 cal BC; Table 5.1; ibid., 87). The survival rate of this logboat type over others, such as log- and bundle-rafts and hide-covered vessels, as well as the inevitable bias created by their contexts of abandonment or loss, also make them subject to much debate in relation to their sea-going ability. Fry (2006, 373–4) has argued that boats made partly or wholly out of a single hollowed-out tree trunk could have been used on both inshore waters and further out to sea, particularly in conjunction with sails. It might therefore be worth highlighting those specifically from river mouths and the vicinity of the sea loughs around the Irish coastline, such as two Early– Middle Neolithic fragmentary hulls from the shore of Lough Larne in Co. Antrim and a 1 The find location suggested the existence of ‘a Bronze Age landscape in which surface water was considerably more extensive than at present. … a mosaic of small lakes, rivers and expanding bog’ (Robinson et al. 1996, 13). Contexts of boat and boat fragment discoveries thus far also suggest abandonment when damaged, deposition for future retrieval and reuse as building material (Gregory 1997), rather than any obvious acts of ‘ritual deposition’, although research has been limited. 14 2 All C dates are calibrated in OxCal 4.2.4 (Bronk Ramsey 2013; Reimer et al. 2013) and presented at 2 sigma (95.4%).

Cleary CW3.indd 140

27/05/2016 14:56:44

5.1 Irish prehistoric boats discussed in the text

1. 2. 3. 4. 5. 6. 7.

Larne Lough Greyabbey Gormanston Cahore Lurgan Lee’s Island Carrowneden

larger, Middle Neolithic logboat from the bed of Strangford Lough near Greyabbey Bay (ibid., 371–2). One of two logboats3 recovered from the silted estuary at Cahore, Co. Wexford, also had a raised bow, and this may have minimized swamping from wave action (Gregory 1997, 107). In 2002 part of another oak logboat, radiocarbon dated to 1193–1013 cal BC4, was found buried c. 1 km offshore at Gormanston, Co. Meath, with perforations in the gunwales suggesting outrigger attachments for sea travel (Table 5.2; N. Brady 2007, 325–6). Robinson et al. (1999, 907) also argue that the presence of spines, ridges and paired holes on the Chalcolithic Lurgan boat could have been designed not only to handle rough-water lake conditions but also, with the addition of stabilizers, operate as a sea-going vessel, ‘confined mainly to coastal waters but venturing short open-sea crossings in favourable weather’. A similar possibility is suggested for the comparable logboat discovered less than 20 km to the north at Carrowneden, Co. Mayo, and also dated to the Late Neolithic/Early Bronze Age (2619–2049 cal BC; Table 5.1; Lanting & Brindley 1996, 88). The ability to undertake such sea voyages may have been greatly facilitated by island-hoping and might be attested by the relatively swift adoption of Irish influences on the Isle of Man during the initial Early Bronze Age, from Irish type copper and bronze axeheads to bowl and vase funerary pottery of typically Irish form and decoration (see below; Woodcock 2004; 2008).   The 1st-century BC gold Broighter boat model, found near Lough Foyle in Co. Derry, is thought to be a depiction of a hide-covered craft, and these were probably also used as sea-going vessels from an early date; indeed the modern Irish currach is one of the few 3 These boats, found before 1857, are not dated and do not survive, but are generally interpreted as prehistoric (for example Breen & Forsythe 2004, 34). 4 This date is courtesy of Dr Niall Brady and The Archaeological Diving Company Ltd.

Cleary CW3.indd 141

27/05/2016 14:56:46

Site

Lab number

C-14 result BP

95.4% calibration

Sample context

Cuilmore, Co. Mayo

Beta-83891

3410±80

1914–1517 BC

Logboat

Carrowneden, Co. Mayo

Beta-85979

3890±90

2619–2049 BC

Logboat

Ballyvogham, Co. Limerick Lurgan, Co. Galway

Edercloon, Co. Longford

Killuragh Cave, Co. Limerick Coolroe, Co. Mayo

Kilmessan, Co. Meath Poulnabrone, Co. Clare

Ballyrenan, Co. Tyrone

Paulstown, Co. Kilkenny

Drumanone, Co. Roscommon Aghanaglack, Co. Fermanagh Audleystown, Co. Down Knowth, Co. Meath

Carrowkeel, Co. Sligo Poulawack, Co. Clare Altar, Co. Cork

Labbacallee, Co. Cork Largantea, Co. Derry

Ballybriest, Co. Derry Moneen, Co. Cork

Coolnatullagh, Co. Clare Brackagh, Co. Derry

Lismullin, Co. Meath

Harlockstown, Co. Meath Treanbaun, Co. Galway Mell, Co. Louth

Cookstown, Co. Meath

Ballynacarriga, Co. Cork

Brother’s Cave, Co. Waterford Ballynamintra Cave, Co. Waterford Killuragh Cave, Co. Limerick

Pollthanacarra, Co. Fermanagh Ballybrennan, Co. Westmeath Straid, Co. Derry

Newtownstewart Castle, Co. Tyrone Moone, Co. Kildare

Cleary CW3.indd 142

GrN-18261 GrN-18565 Wk-20961 OxA-6754

GrN-27642 Wk-24268 Wk-24267

UBA-23505 UB-6706

UBA-15435 UB-6696 UB-7188 UB-7190 UB-7593 UB-7189 UB-6979 UB-6978

UBA-12683 Ua-510

OxA-3260 OxA-3262 OxA-3263 OxA-3289

GrN-11359 OxA-2759 OxA-2760

UBA-6974 UBA-6976 UBA-6977

GrA-13273 GrA-13254

GrN-11904

OxA–10530 UBA-10619 UBA-10618

SUERC-23489 SUERC-23490 Wk-18192 Wk-18182 Wk-22560 Wk-22563 Wk-22561 Wk-17463 Wk-17938 UB-14777 UB-13165

GrA21488

Beta-277398 Beta-277384 OxA-6748

OxA-25558 UBA-8154 UBA-8153 UBA-8152

GrA-24139 GrA-24179

GrN-15492 GrA-14825 UB-6784 UB-6783

SUERC-24981

3300±30 3940±25 2909±30 3020±45 3950±40 3589±30 3550±30 3822±37

3743±36 3821±26 3639±37 3608±38

3774±36 3732±35

3719±33 3713±28 3549±35

4265±24

3770±100 3830±90 3520±60 3600±65 3670±80 3805±45 3780±70

3630±70

3837±35 3828±37

3871±37 3630±50 3580±50 3755±30 3835±45 4037±21

3930±21 3905±30 3845±30 3799±33 3620±41 3515±30 3957±30 4111±30 3894±50 3986±69

3852±34 3861±23 3760±50 3470±40 3720±40 3775±40 3599±29 3804±34 3745±34

3629±37 3650±40

3660±40 3840±35 3690±40 3680±38 3897±30 3685±30

1643–1504 BC 2561–2345 BC 1206–970 BC

1402–1127 BC 2572–2307 BC 2028–1883 BC 2009–1772 BC 2457–2142 BC 2281–2033 BC 2430–2147 BC 2134–1905 BC 2128–1881 BC 2333–2041 BC 2277–2029 BC 2205–1985 BC 2199–2030 BC 2011–1771 BC 2912–2877 BC

2473–1936 BC 2562–2030 BC 2020–1692 BC 2139–1770 BC 2293–1781 BC 2458–2062 BC 2459–2031 BC 2201–1775 BC 2458–2200 BC 2458–2147 BC 2467–2210 BC 2141–1882 BC 2120–1769 BC 2260–2140 BC 2460–2140 BC 2621–2486 BC 2484–2342 BC 2471–2298 BC 2457–2205 BC 2397–2136 BC 2132–1886 BC 1921–1751 BC 2570–2347 BC 2865–2575 BC 2488–2206 BC 2853–2289 BC 2460–2206 BC 2461–2211 BC 2343–2026 BC 2270–1977 BC 2275–1980 BC 2344–2036 BC 2028–1890 BC 2430–2137 BC 2279–2036 BC 2132–1894 BC 2140–1916 BC 2190–1926 BC 2458–2202 BC 2199–1960 BC 2196–1951 BC 2476–2213 BC 2194–1972 BC

Logboat Logboat

Birch brushwood overlying block wheel Horse sacrum

Hazel charcoal from burnt mound spread overlying layer containing horse metacarpus Alder charcoal from earliest burnt mound spread, pre-dating sediment layer containing horse vertebra

Alder charcoal from latest burnt mound spread, post-dating sediment layer containing horse vertebra Non-burnt human cranium from under edge of portal tomb cairn Cremated human bone from portal tomb end chamber

Hazel charcoal from the primary fill of pit with 23 disc beads and sherds from at least 23 Beakers Cremated human skull fragment from portal tomb chamber Cremated human longbone from court tomb chamber Cremated human skull from court tomb chamber 5

Non-burnt human mandible from court tomb chamber 6 Cremated human longbone from court tomb chamber 5 Non-burnt human tooth from court tomb chamber 6

Non-burnt human mandible from court tomb chamber 6 Cremated human bone from passage tomb 15

Non-burnt human tooth from passage tomb cairn M Non-burnt human bone from cist 4

Non-burnt human bone from cist 6, SW compartment Non-burnt human bone from cist 6, NE compartment Non-burnt human tooth from wedge tomb

Non-burnt human longbone from wedge tomb cist-like chamber Non-burnt human bone from wedge tomb Non-burnt human bone from wedge tomb

Cremated human skull from wedge tomb chamber

Cremated human longbone from wedge tomb chamber

Cremated human longbone from wedge tomb chamber Cremated human bone from wedge tomb chamber

Cremated human bone from wedge tomb antechamber pit Non-burnt human bone from cist Non-burnt human bone from cist

Cremated human bone from cist 1 Cremated human bone from cist 2 Hazel charcoal from pit F2241

Hazel charcoal from pit F2284

Ash/pomoideae/alder charcoal from pit

Cattle scapula from basal fill of ring-ditch Cremated human bone from cist

Undiagnostic burnt bone from pit sealing layer Cremated human bone from pit

Non-burnt human bone from pit

Oak charcoal from pit cut into inhumation burial Cremated human bone from cist C2194 Pomoideae charcoal from cist C2161 Non-burnt human tooth from cave Non-burnt human tooth from cave

Non-burnt human metatarsal from cave

Non-burnt human mandible from pit in cave Non-burnt human radius from cave

Non-burnt human humerus from cave Non-burnt human humerus from cave Non-burnt human humerus from cave Non-burnt human bone from cist Cremated human bone from cist

Non-burnt human bone from cist Cremated human bone from cist

Cremated human bone from cist, larger compartment

Cremated human bone from cist, smaller compartment Non-burnt human rib from pit

27/05/2016 14:56:46

[ 143 ]

Cleary

European survivals of such a craft. The very nature of the construction of such boats, however, also means they are far less likely to survive in the archaeological record. While the translations are somewhat problematic (see Freeman 2000, 28–33), the 4th-century AD poem, Ora Maritima by Avienus, apparently referencing the lost 6th- or 5th-century BC Greek itinerary of the western seaways, Massaliote Periplus, can be interpreted according to McGrail (1995, 264) as recalling people from western Brittany that used hide boats to obtain tin and lead from Ireland and Britain. Short coastal journeys, that would have been possible in the boats thus far discovered, could also have formed ‘an interconnecting chain of contacts extending considerable distances along the Atlantic coasts of Europe’ (Waddell 1991/92, 29). Such a system of sporadic connection and interaction could be responsible for the broad similarities in such things as funerary practices and ceramic styles identified across large areas but may also explain the regional differences that developed in tandem. Reaching Land The Irish coastline is richly endowed with many natural inlets that can provide both shelter and safe access from the sea. A study of the fishing industry in southern Ireland in the late 1960s identified over 870 harbours and landing places, the majority along the west and south coasts (O’Connor et al. 1980, 127). When the importance of local knowledge for reaching and successfully manoeuvring many of these landing points is considered, along with the ability to control and direct the type of seagoing vessels discussed above, it is clear that both geographical and navigational expertise would have been essential to safe travel. While many argue that early seafarers relied wholly on naturally observable phenomena, it has also been suggested that the coastal siting of some Neolithic and Bronze Age Table 5.1 (facing page) Published AMS results referenced in text Table 5.2 (below) New AMS results referenced in text Site Gormanston, Co. Meath Coolroe I, Co. Mayo

Fahee South, Co. Clare

Fourknocks II, Co. Meath Loughash, Co. Tyrone

Loughash/Cashelbane, Co. Tyrone Lyles Hill, Co. Antrim Lismullin, Co. Meath

Piperstown (Site K), Co. Dublin Treanbaun, Co. Galway Killarah, Co. Cavan

Cleary CW3.indd 143

Lab number

C-14 result BP

95.4% calibration

AMS δ13C

Sample context

UBA-9717

2905±20

1193–1013

-25.9

Logboat (courtesy of Niall Brady)

UBA-29672

3780±53

2450–2033 BC

-21.4

Non-burnt human ulna from cist 2

UBA-30802 UBA-29700 UBA-29673 UBA-29676 UBA-29675 UBA-29674 UBA-29663 UBA-29662 UBA-29665 UBA-29664 UBA-29666 UBA-30815 UB-7825

UBA-29699 UBA-29698 UBA-29681

failed

2835±30 3567±32 3848±34 failed

3518±30 3458±31 failed

3888±31 4355±31 3431±31 failed

3958±37 4004±40 3455±38 4019±32

1106–911 BC

2023–1777 BC 2458–2206 BC 1926–1751 BC

-22.5 -21.4

1880–1692 BC

Non-burnt human femur from cist 4 Non-burnt human femur from cist 4 Cremated human femur from cist 4

Cremated human femur from wedge tomb chamber Cremated human femur from wedge tomb, cist B

2573–2344 BC

-27.0

2620–2470 BC

-22.5

1886–1667 BC

Cremated human tibia from cist 2

Cremated human femur from wedge tomb chamber

3058–2902 BC

2832–2457 BC

Horse incisor from fulacht fia trough

Cremated human tibia from wedge tomb chamber

2470–2286 BC 1877–1644 BC

Horse metacarpal from layer sealed by burnt mound

Cremated human femur from cist

Cremated human longbone from pit F2241

Oak charcoal from pit (courtesy of Kim Rice) Cremated human skull from pit

Cremated human longbone from stone-lined pit Non-burnt human skull from cist

27/05/2016 14:56:47

[ 144 ] v. burial practices during the late third millennium bc

monuments may have aided navigation (see Van de Noort 2011, 102–3, 139–41). The small Middle Neolithic passage tomb cemeteries of Gormanston in Co. Meath and Bremore in Co. Dublin, for example, flank the River Delvin where it flows into the sea, perhaps marking a navigable route that also led to the base of a prominent ridge a few kilometres inland at Fourknocks, and which was similarly selected as the site of a passage tomb cemetery. This area probably continued to provide sea-access into the later prehistoric period and indeed the Late Bronze Age Gormanston logboat was discovered just off this shoreline. A detailed examination of the costal siting of Neolithic and Bronze Age monuments might therefore provide additional information on navigating sea journeys, suitable landing places and travelling inland during prehistory.   As discussed above, logboats could have been used on loughs and rivers to move inland relatively quickly, while evidence for wooden trackways across large areas of wetland is wellattested, with Brindley and Lanting (1998) confirming clusters of wetland construction in the ‘drier’ periods, including at c. 2250 BC, 1600–1400 BC and 1100–850 BC. Recent excavations at Edercloon, Co. Longford also revealed a multi-period complex of wellpreserved wooden trackways and platforms as well as part of an unfinished tripartite block wheel with a directly overlying piece of birch brushwood radiocarbon dated to 1206–970 cal BC (Table 5.1; Moore & Chiriotti 2010). This provides the earliest evidence for the wheel in Ireland which had previously been dated to the Earlier Iron Age, c. 400 BC, based on radiocarbon dates from two examples found at Doogarrymore, Co. Roscommon (Lucas 1972). Furthermore, while two probable cart pieces recovered from trackways in counties Longford and Tipperary are also dated to the Iron Age (Raftery 1996, 273–5; Taylor 2008, 55), of the 16 yokes for draughting animals discovered to date, at least three have been dated to the Late Bronze Age/Early Iron Age (Stanley et al. 2003). While early trackways are therefore well-attested, the use of wheeled vehicles or animals pulling loads is more problematic with a lack of archaeological evidence thus far pre-dating the Late Bronze Age.   The use of horse for travelling across land also requires comment. Until recently the earliest evidence for the post-glacial reintroduction of horse, thought to be domesticated and by default linked with the introduction of metal, came from Newgrange, Co. Meath (see McCormick 2007). Over 150 horse bones and teeth were reportedly from contexts associated with the ‘Beaker settlement’ around the southern perimeter of the large passage tomb (van Wijngaarden-Bakker 1974; 1986). Two of the horse teeth, however, were subsequently radiocarbon dated to the Later Iron Age, throwing the Beaker period dates for all the Newgrange horses into doubt (Bendrey et al. 2013). Reanalysis of the assemblage also identified evidence for the use of metal bits on two horse teeth and suggested that these animals were significantly smaller than contemporary Beaker period horses from elsewhere in Europe. Thus, currently the earliest evidence for post-glacial horse in Ireland is from Killuragh Cave, Co. Limerick, where a horse sacrum was dated to 1402–1127 cal BC (Table 5.1; Kaagan 2000, table 4.5, 171). This horse bone, together with a human jaw, reportedly came from a shallow pit in Cave 1 that was stratified below a layer that contained microliths, reiterating the mixed nature of the cave assemblage. Elsewhere, horse has been indirectly dated to the Later Bronze Age, such as the remains recovered from ditched enclosures at Ballyveelish, Co. Tipperary and Sheephouse, Co. Meath, and from a well associated with a fulacht fia at Ballykeeffe, Co. Limerick (McCormick

Cleary CW3.indd 144

27/05/2016 14:56:47

Cleary

[ 145 ]

1987; Strid 2012; Bermingham et al. 2013, 78). Many of these remains are fragmented, implying marrow removal, and cut marks have also been observed suggesting food remains, although perhaps only chosen in periods of food shortage (McCormick 2007, 89). Horse bones and teeth have also been recovered from potentially earlier contexts, such as at Coolroe, Co. Mayo, where the distal portion of a metacarpus from an adult horse was found in a layer sealed by a burnt mound spread dated to the Chalcolithic by hazel charcoal (2572–2307 cal BC; Table 5.1; Gillespie 2012). A similar deposit was uncovered at Kilmessan, Co. Meath, where a horse vertebra was recovered from a sediment layer sealing the earliest of three burnt mound spreads; alder charcoal samples returned dates of 2028–1883 cal BC and 2009–1772 cal BC from the earliest and latest burnt mound spreads respectively (Table 5.1; Conran 2010). None of these horse bones have been directly dated5, however, and this needs to be addressed in order to ascertain the use of horse in Ireland, where they originated, what prompted their reintroduction, presumably as domesticates, and how this may have impacted the ability of individuals to travel around the island. Ireland and the Beaker culture Those who worked the earliest known copper mines at Ross Island have been described as a ‘community of occupation’, whereby their social identity may have been closely linked to their shared participation in mining (O’Brien 2004, 475–6). The archaeological evidence also suggests that the Ross Island mines were only worked seasonally or part-time, from c. 2400–1900 BC, and that the individuals probably returned to a farming community for the remainder of the year. It is plausible then that the influences experienced by these people could have equally affected the communities within which they were based, with new ideas moving through the webs of kinship and friendship both within these close social units and further afield through various networks of contact and exchange. The excavations at Ross Island have also made the individuals who formed these early mining communities more tangible and therefore a group that we can perhaps more easily envision as both travelling abroad and receiving travellers from other regions. It has been argued, for example, that the production of arsenical copper at Ross Island is indicative of considerable metallurgical expertise from the beginning and that know-how must have come from the Continent, with similar practices underway in Atlantic France, south-eastern France, Iberia and northern Italy (ibid., 560–1). These networks of knowledge transfer must have been influential and the exchange of artefact ideas during this period is well-attested, from axeheads with close morphological parallels to the Irish Lough Ravel type found in Iberia and France (Sheridan 5 Supported by AHRC grant AH/K002600/1: Atlantic Europe in the Metal Ages (AEMA): questions of shared language, the sample from Coolroe was submitted for dating but failed. The Kilmessan horse bone was not available for dating. A non-burnt horse incisor from the trough of a potentially early fulacht fia at Fahee South, Co. Clare returned a Late Bronze Age date of 1106–911 cal BC (Table 5.2); wood from this trough had previously returned a Middle Bronze Age date (Ó Drisceoil 1988, 672). The AEMA project has also funded dates detailed in this paper (Table 5.2) from the following sites: Fourknocks II, Co. Meath, Loughash and Loughash/Cashelbane, Co. Tyrone, Lyles Hill, Co. Antrim, Piperstown (Site K), Co. Dublin, Treanbaun, Co. Galway and Killarah, Co. Cavan. A date on cremated human bone from Lismullin, Co. Meath was also attempted but failed.

Cleary CW3.indd 145

27/05/2016 14:56:47

[ 146 ] v. burial practices during the late third millennium BC [ 146 ] 5.2 Distribution of Beaker pottery in Ireland, including discoveries in megalithic monuments (after Carlin 2011a, fig. 9.12)

Beaker pottery Beaker in portal tomb Beaker in court tomb Beaker in passage tomb Beaker in wedge tomb Beaker in cist

1983, 15) to tanged copper daggers manufactured in Ireland but with affinities to those from southern England and the Netherlands but also the ‘West European’ form generally associated with Maritime Beakers6 (Sheridan & Northover 1993, 64–5).   Despite a recent excavation boom in Ireland that saw a significant increase in the number 6 There is a notable absence of Maritime Beakers from Ireland (Sheridan 1983, 17), with the possible exception of sherds from two Beaker vessels with ‘Maritime-type decoration’ from the Hill of Rath, Co. Louth (Carlin 2011a, 84). These were reportedly found in a pit with burnt bone, sherds representing at least 28 other Beaker vessels and worked flint, including blades, scrapers, one partial and one complete barbedand-tanged arrowhead.

Cleary CW3.indd 146

27/05/2016 14:56:50

Cleary

[ 147 ]

of sites with Beaker pottery, the development of the Beaker Tradition and its dating still remain poorly understood (Figure 5.2). This is due to the dominance of sherd assemblages rather than complete or reconstructed vessels and an absence of both secure, short-life contexts, such as graves, and chronologically definable associations, as well as the various wiggles of the calibration curve affecting radiocarbon dates for the latter half of the 3rd millennium BC (see Brindley 2004, 334; 2007, 250–1). Despite these limitations a threefold style scheme developed by Case (1993; 1995) identified a small number of early, Continental-style Beakers of classic AOO or AOC form from c. 2500/2450 BC, a second hybridized Bell Beaker style for the majority of vessels, combining elements of the early Atlantic tradition as well as more north-western European and British influences, generally dated to 2450–2200 BC, and a few late-style Beakers that probably extended in use until c. 2050 BC and overlapped with the beginning of the (Food Vessel) Bowl Tradition, c. 2200/2150 BC (Brindley 2004, 334–6; 2007, 250; Grogan & Roche 2010, 36; Carlin 2011a, 217–24, 227–8; Bayliss & O’Sullivan 2013, 69).   While those who initially worked the mines at Ross Island also manufactured7 and utilized the relatively newly adopted Beaker style pottery, there is as yet limited evidence that they felt the need to incorporate these vessels into their burial practices or adopt the associated rite of single crouched inhumation burial that was fashionable elsewhere, such as southern Britain and central Europe. Where Beaker period burial has been identified in Ireland, collective rather than individual burial has generally been emphasized and this, along with the reuse of older monuments, has linked Ireland to an Atlantic tradition stretching from Iberia to northern France (Case 1995, 19, 25; Cunliffe 2001, 242; O’Brien 2004, 564; Mallory 2013, 116). Recent analysis of Beaker depositional practices in Ireland has confirmed that of the 219 sites with Beaker pottery identified by Carlin (2012, fig. 2), 79% could be classified as settlement and only 17% as funerary. This is in stark contrast to Brittany, where almost all of the 121 sites with Beaker pottery and associated artefacts were funerary, 81% of which reused megalithic tombs (Salanova 2004, 66), but it has perhaps better parallels with the Iberian evidence, where many elements of the standard Beaker set come from settlement contexts and the grave goods accompanying burials tend to be restricted (C. Gibson 2013, 77). While it is tempting to interpret these as distinct regional expressions it is also important to acknowledge that these divergences may be exaggerated by the excavation histories in individual countries. In Ireland, for example, only c. 6%8 of Neolithic megalithic tombs have been excavated, the majority before the 1960s. A recent construction boom, much of it focused on generally low-lying landscapes representing favourable settlement locations, has however resulted in the discovery of previously unknown sub-surface sites, including a small number of definite Beaker burials (see below).   It is pertinent to examine not only how Irish burial practices of this period fit into this Atlantic tradition of the reuse of earlier collective graves, but also the recent identification that some Bell Beaker burials in Brittany and southern Portugal were in fact ‘individualized’. 7 Petrological analysis of sherds from eight Beaker vessels suggested they were locally made (Ixer 2004) and most of the pottery has a similar form and ornament suggesting it belongs to the earlier style Beakers, dating to 2450–2200 BC (Brindley 2004, 334–6). 8 Approximately 66 out of c. 1,040 portal, court and passage tombs have been excavated or partially excavated, however this figure generally excludes various antiquarian investigations.

Cleary CW3.indd 147

27/05/2016 14:56:50

[ 148 ] v. burial practices during the late third millennium bc

This was achieved through the use of defined compartments within earlier monuments, megalithic cists and, where bone survives, evidence for separate episodes of burial (Salanova 2003; 2004; C. Gibson 2013, 77). Examples include the reuse of the end chamber of the passage grave at Gâvres in Morbihan as a ‘cist’ for Beaker deposition and the transformation of the entrance to the tomb of Casas do Canal in Estremoz, Portugal into a ‘cist’ to contain two Bell Beaker vessels probably associated with one individual (Salanova 2004, 66, 71). In addition, reassessment of the ‘classic Beaker burial’ in Britain suggests that mortuary practices there were also more complex, incorporating incomplete and disarticulated inhumations, cremations and multiple internments, and further analysis indicates that some of these burials were also ‘individualized’, specifically through sequential interments (Petersen 1972; A. Gibson 2004). 5.3 Distribution of potential Chalcolithic/Beaker period burials in Ireland

Reuse of Neolithic burial monuments Despite a widespread assumption that older megalithic tombs were reused for collective Beakerassociated burial in Ireland, later activity and disturbance at many of these sites hinders a true understanding of their chronological use and requires further targeted 14C dating projects (see Brindley et al. 2005; Brindley 2007; Kytmannov 2008; Schulting et al. 2012; Bayliss & O’Sullivan 2013; Bergh & Hensey 2013; Hensey et al. 2013; Schulting et al. forthcoming). Radiocarbon dating also has an important role to play when no diagnostic artefacts are recovered, or indeed are recovered but not directly associated with the human remains. Taking the dates detailed above as reflecting the period of Beaker use (c. 2450–2200/2150 BC), just 10 earlier megalithic tombs with potentially contemporary burial activity can be identified, although several of these are also problematic and far from definite burials (Figure 5.3).   Of the 20 or so Early Neolithic portal tombs excavated or partially excavated in Ireland, just one has definite evidence for Beaker-associated use. At Poulnabrone, Co. Clare, at least two undecorated Beaker pottery sherds, an unfinished hollow-based chert arrowhead and additional diagnostic lithics were recovered from the chamber and the dating of at least one individual to this period confirmed that the site may have been a focus for some ‘burial activity’; a fragment of human cranium in a probably disturbed context, under the edge of the

Cleary CW3.indd 148

27/05/2016 14:56:51

Cleary

[ 149 ]

cairn, returned a dated of 2457–2142 cal BC (Table 5.1; Lynch 2014, 44–6, 51–2). The portico or small cist-like feature constructed directly in front of the portal stones is also notable as it appears to be unique to Poulnabrone and may be a secondary feature. It was not possible to date due to the long timespan of deposition in this area, as indicated by finds of Neolithic and Middle Bronze Age pottery and radiocarbon results on human and animal bones ranging from the Chalcolithic to the Iron Age (ibid., 49–51, 179–80). A further two portal tombs also have potential burial contemporary with later style Beakers, in use until c. 2050 BC, but with a lack of Beaker pottery or other diagnostic artefacts these are tentative and could equally be assigned an Early Bronze Age date, despite a similar lack of associated pottery. At Ballyrenan, Co. Tyrone, cremated human bone from the end chamber was dated to 2281–2033 cal BC (Table 5.1; Kytmanov 2008, 107) and was reported to have been ‘wedged among the stones … laid down to seal the burial’ (Davies 1937, 95–6). Six disc and two fusiform stone beads (schist and shale), types generally assigned an Early Bronze Age date in Ireland (see Brindley 2007, 249; Sheridan 2007), were also recovered from the tomb but reportedly from a different chamber to the dated bone, although according to a workman who assisted at the excavation one shale disc bead (BM3) may have come from the same chamber (Davies 1937, n 10). The six disc beads appear to be graded in size, with BM3 the smallest, both in diameter and thickness. The stylistic simplicity of disc beads, however, make them difficult to accurately date, but at Paulstown, Co. Kilkenny, hazel charcoal from the primary fill of a pit with 23 disc beads and sherds from at least 23 Beakers returned a date of 2430–2147 cal BC, confirming a direct link to Beaker-associated activity (Table 5.1; Carlin 2011a, 1:84). From the chamber of a portal tomb at Drumanone, Co. Roscommon, a cremated human skull fragment returned a date of 2134–1905 cal BC (Table 5.1; Kytmanov 2008, 107) but the only notable artefact recovered from a mixed deposit was a miniature polished porcellanite axehead (Topp 1962).   Although at least 14 of the 36 court tombs excavated in Ireland had associated Beaker pottery sherds (Carlin 2011a, 1:135), just two also had potentially contemporary human bone deposits and a third had human bone dated to the Chalcolithic but no diagnostic artefacts; all three monuments are classified as dual or ‘double horned’ court tombs. At Ballybriest, Co. Derry, a stone-lined pit or cist containing the cremated remains of an adult male associated with sherds from a Beaker vessel had been dug into the cairn of the court tomb (Evans 1939, 9–10; Herity 1987, 154). At Aghanaglack, Co. Fermanagh, two Green Low type barbedand-tanged arrowheads (Green 1980) were found in Chamber I West, sherds of a possible Beaker vessel came from Chamber II West and a cremated human longbone from one of the chambers returned a date of 2128–1881 cal BC, while two dates on non-burnt bone from two individuals fell within the period 1884–1637 BC and probably represent a second phase of reuse in the Early Bronze Age (Table 5.1; Davies 1939a; Herity 1987, 154, 215; Schulting et al. 2012, table 2, 34). At Audleystown, Co. Down, both cremated and non-burnt bones were recovered and represented the deposition of some 30–33 individuals, at least two of which may represent late Beaker period burials (Collins 1954; Schulting et al. 2012, table 2, 34–5); a cremated individual9, probably from Chamber 5 (2333–2041 cal BC) and a non-burnt 9 A second cremated bone sample (longbone) from the same area was also selected for dating (2205–1985 cal BC; Table 5.1) and potentially these could belong to the same individual or a second individual contemporaneous with the non-burnt individual(s) in Chamber 6.

Cleary CW3.indd 149

27/05/2016 14:56:51

[ 150 ] v. burial practices during the late third millennium bc

adult from Chamber 6 (2277–2029 cal BC; Table 5.1). Notably, no Beaker pottery or other diagnostic artefacts were recovered from the tomb and indeed a non-burnt tooth from a child in Chamber 6, apparently associated with Bowl Tradition sherds, yielded a roughly comparable date of 2199–2030 cal BC, while a second date of 2011–1771 cal BC was returned for the non-burnt mandible of an adult from the same area (Table 5.1). Arguably, both fall within the expected range for Bowl Tradition pottery and, similar to Aghanaglack, may represent a second phase of reuse or continuous reuse from the Late Neolithic into the Early Bronze Age.   Despite various levels of excavation undertaken at approximately 23 passage tomb sites across Ireland, including several cemeteries, the only one with an often quoted Beaker burial, Tomb 15 at Knowth, Co. Meath, has now been called into question. Despite the cremated remains of an adult and child in apparent association with an undecorated Bell Beaker in the passage (Eogan 1984, 90, 308–12), a Late Neolithic date of 2912–2877 cal BC was returned for a longbone fragment from the adult (Table 5.1; Schulting et al. forthcoming). The child remains undated. In passage tomb Cairn M at Carrowkeel, Co. Sligo, a date of 2473–1936 cal BC from a human tooth, might also indicate some Beaker activity, particularly as a stone bracer was found in the vicinity of this passage tomb cemetery (Table 5.1; Harbison 1976, 24), but the large standard deviation makes this tentative, as does the recovery of Early Bronze Age funerary pottery from other passage tombs within the complex (Bergh 1995, 103; Hensey et al. 2013, 18). At Fourknocks II, Co. Meath, recent dating on non-burnt bone from Cists 2 and 4 suggests the interment of Chalcolithic remains in secondary cists inserted into the covering mound and previously interpreted as Early Bronze Age reuse (Hartnett 1971, 64–74). Cist 2 contained the skeleton of an adult female crouched on her right side with cremated bone reportedly found in the upper fill but also under the skeleton. Associated artefacts consisted of an awl, a flint knife, flake and chip. A non-burnt right ulna fragment returned a date of 2450–2033 cal BC, while a cremated tibia shaft fragment returned a later date of 2023–1777 cal BC suggesting these deaths were not contemporary (Table 5.2). The cremated bone from above and below the skeleton was not stored separately and as such it is possible that a sample from above the skeleton was dated, or that bone filtered down as a result of natural formation processes. The occurrence of what appears to be a doublepointed awl of Thomas’s Group 1B type is notable if linked with the female inhumation; it can be paralleled with other Group 1 awls that Thomas (1968; 2005) has recorded as tending to be found in Beaker graves. Clarke (1970, 448) also noted awls occurring with his later style Beakers and more recently Woodward and Hunter (2015, 89) have also recorded their association with Early Bronze Age ceramics. While the awl from Fourknocks II is referred to as bronze this has not been confirmed by metal analysis and it is notable that some Beaker awls are of unalloyed copper (Thomas 2005, 220; Woodward & Hunter 2015, 89). Cist 4 contained the poorly preserved skeleton of an adult, disturbed but apparently once placed in a crouched position, some elements of a second inhumation, also disturbed, and the cremated bones of an adult scattered among and below the skeletal remains. A Food Vessel Bowl stood upright in the north-east corner of the cist, near the skull of the crouched inhumation. A non-burnt left femur fragment from one of the skeletons returned a date of 2458–2206 cal BC, an attempted date on the same element from the second non-burnt individual failed

Cleary CW3.indd 150

27/05/2016 14:56:51

Cleary

[ 151 ]

and a cremated femur shaft fragment returned a date of 1926–1751 cal BC (Table 5.2). It is therefore tempting to interpret the original interment as a Beaker burial that was perhaps disturbed by the insertion of the Bowl (expected date of 2160–1920 as proposed by Brindley 2007) and crouched inhumation, and subsequently the cremation, but this is tentative as both inhumations were stored together and so it was not possible to determine which sample came from which individual. Further evidence for Beaker activity on the Fourknocks Ridge consists of a fragment of a stone bracer (Harbison’s Type A) and a barbed-and-tanged arrowhead (Green’s Sutton a type) recovered from topsoil during excavations approximately 550 m east of the passage tomb cemetery (King 1999; Harbison 1976; Green 1980).   At Poulawack, Co. Clare, just 1.5 km to the south of Poulnabrone portal tomb, six cists were inserted into the kerbed cairn of a Middle Neolithic Linkardstown-type megalithic cist. As with Aghanaglack and possibly Audleystown, this reuse took place in multiple phases (Hencken & Movius 1935; Brindley & Lanting 1991/92a). Cist Grave 4 on the outer edge of the cairn contained the disturbed, non-burnt remains of an adult and child, with a bone sample returning a date of 2562–2030 cal BC (Table 5.1). A compartmented cist, Grave 6, in the other quadrant of the cairn, contained the non-burnt bones of an adult male and one sherd of Beaker pottery in the south-western compartment, and the cremated remains of a young adult with the non-burnt bones of an infant less than 4 years, a child aged 5–10 years and an adolescent aged c. 15 years in the north-eastern compartment (Jessica Beckett pers. comm.). Some of the non-burnt adolescent was also recovered from the south-western compartment but probably just represents spill from the other compartment, indeed based on the taphonomy, Beckett concluded that most of the individuals were placed in as whole skeletons (or bodies) that were disturbed at a later date. The non-burnt bone from the adult male and the adolescent returned dates of 2020–1692 cal BC and 2139–1770 cal BC respectively (Table 5.1). The single Beaker sherd and the later dates for Grave 6 may indicate that the Beaker pottery was a residual inclusion (see Moone, Co. Kildare, below) or that the non-burnt adult male and adolescent were later insertions, particularly given the earlier date for Grave 4, which hints at Beaker reuse of the earlier monument. Additional dating on the remaining individuals, the infant, child and cremated young adult, might yet shed more light on the sequence of burial. The role of Wedge Tombs Given this paucity of burials in earlier monuments it is of interest that another type of Irish megalithic monument, wedge tombs, suddenly emerged in both the south and north of the island in the period 2540–2300 BC (Schulting et al. 2008). These monuments are predominately focused across the western half of Ireland but with distinct regional concentrations, and were quickly linked to the advent of metal and the use of Beaker pottery. Despite a widespread presumption of use for Beaker funerary activity, to date only about 40 of the 560 or so known wedge tombs have seen any level of investigation and Beaker pottery in an apparently primary position has been found in just 11 (28%), and of these, only in nine (23%) was the pottery loosely in association with human bone (Figure 5.3; ibid., 1,

Cleary CW3.indd 151

27/05/2016 14:56:51

[ 152 ] v. burial practices during the late third millennium bc

table 1; Carlin 2011b, 91; O’Brien 2012a, 76; Grant 2009; Delaney 2012; Ros Ó Maoldúin pers. comm.). As with the reuse of older megalithic tombs, however, it is often difficult without systematic dating to identify the Beaker component of the funerary remains as many wedge tombs have evidence for multi-period activity. At the site of Loughash (Giant’s Grave), Co. Tyrone, for example, three near complete Beaker pots and the cremated remains of at least four individuals were found in the chamber, but from mixed deposits that also included Early Bronze Age Vase Urn and Encrusted Urn pottery and Late Bronze Age ceramics (Davies 1939b). A cremated tibia fragment from a pit sealed by black-stained boulders near the back of the chamber returned a date of 5.4 Disarticulated inhumation of adult female in cist-like 1880–1692 cal BC (Table 5.2). This pit end chamber of the wedge tomb at Labbacallee, Co. Cork (after Leask & Price 1936, fig. 15) also contained a few Vase Urn sherds and a fragment of a bronze blade and therefore represents an Early Bronze Age phase of use. A second date attempted on a femur fragment from a concentration of cremated bone at the central portal in the chamber failed.   Similarly, at Kilhoyle, Co. Derry, while the cremated remains of an adult female were closely associated with Beaker sherds on the floor of the main chamber, at least two more cremated individuals (adult male and adult female) could have been contemporary, as well as a Green Low type barbed-and-tanged arrowhead from under the septal stone (Herring & May 1937/38; Ó Ríordáin & Waddell 1993, 97; Green 1980; Carlin 2011a, 1:32). The disturbed nature of the deposits however, and the recovery of at least two Food Vessel Bowls means that absolute dating is essential in order to determine the burial sequence. Antiquarian excavations at Moytirra, Co. Sligo, also uncovered the non-burnt remains of five adults and one child, with at least one adult described as ‘in a crouching posture….the skull and bones in a heap and nearby lay a long thin piece of bronze’, which may have been a rapier (Cremin Madden 1969, 156–7). While early-style Beaker pots were also recovered from this tomb, their association with the human bones remains unknown10.   Despite a significant cluster of wedge tombs in south-west Ireland, there are none in the immediate vicinity of the Ross Island mines. O’Brien (2004, 495, fig. 255) noted that the closest ‘possible wedge tomb’ to Ross Island was 11 km to the south-east in the townland 10 The current whereabouts of the human remains from Moytirra is not known and it is unclear from the reporting whether or not they were retained (see Wood-Martin 1883, 467–70).

Cleary CW3.indd 152

27/05/2016 14:56:51

Cleary

[ 153 ]

of Foiladuane, however this tomb was ‘not considered to be an antiquity’ following site inspection by the Archaeological Survey of Ireland, but another wedge tomb is recorded just over 4 km further to the south-west, in the townland of Crohane. Ó Maoldúin (2014a, n 145) has also commented on the existence of two further wedge tombs in the townlands of Gortlicka and Gortnakilla, approximately 15 km south-east of Ross Island but alongside the River Flesk; this river would have been navigable in small craft and runs into Lough Leane, the lake surrounding Ross Island. Excavations to date at eight wedge tombs in the wider Cork-Kerry region have also failed to recover any Beaker pottery (but see below re Moneen), although 14C determinations on human remains from Altar and Labbacallee might suggest Chalcolithic burial activity. At Altar, on the Mizen Peninsula, a notable spread of cremated bone was uncovered in the entrance area, although the bone was minute and difficult to identify at least some was classified as human adult and one non-burnt tooth from this spread, representing a subadult aged at least 12–13 years, returned a date of 2293–1781 cal BC (Table 5.1; O’Brien 1999, 91–140). The non-burnt remains of three individuals, an adult female, adult male and child, were recovered from the wedge tomb at Labbacallee (Leask & Price 1936). The adult female had been buried in a partially disarticulated state within the terminal cist-like chamber (Figures 5.4 & 5.5a) and an adult female skull that probably belonged to this skeleton was found within the main chamber. A longbone from this individual returned a date of 2458–2062 cal BC and a comparable date of 2459–2031 cal BC was returned for the adult male (Table 5.1). The child had a slightly later date of 2201–1775 cal BC and may represent a separate phase of burial (Table 5.1; Brindley & Lanting 1991/92b, 21). Comparable to the cist-like end chamber at Labbacallee, similar ‘compartments’ are evident at other wedge tombs and their potential importance has been highlighted by Carlin (2011b, 91). In other regions these compartments have been interpreted as a way of ‘individualizing’ Beaker-associated burials (Salanova 2003; 2004), but they may also have functioned to restrict access or block tombs’ entrances. Although many wedge tombs appear open to us today this may not have been so in the past; the construction of antechambers, door stones, septal stones, jamb- and sill-stones, along with cairns and mounds, may all have provided ways in which access was curtailed and tombs closed (Walsh 1995; O’Brien 1999, 198–9).   At Largantea, Co. Derry, two almost complete Beaker vessels and sherds from a third were deposited in a compartment inside the entrance (Figure 5.5b), while at least three cremated indi­vid­u­als in the main chamber area were dated to 2455–2208 BC (Table 5.1; Herring 1938; Schulting et al. 2008). At Loughash/Cashelbane, Co. Tyrone, two small stone compartments to the rear of the main chamber contained large amounts of cremated bone, representing at least one adult male, and Beaker pottery sherds (Figure 5.5c), while the main chamber also contained Beaker pottery, cremated bone from at least four individuals (two adult females, an adult male and a subadult) and two barbed-and-tanged arrowheads, one Sutton and one Green Low type (Davies & Mullin 1940; Green 1980). While further radiocarbon dating is necessary11, particularly as some burials may relate to the deposition of two Food Vessel Bowls, a cremated femur fragment from the chamber did return a date 11 Cremated bone from Cist B returned an anomalously early date of 3058–2902 cal BC (Table 5.2) and therefore requires further dating and analysis.

Cleary CW3.indd 153

27/05/2016 14:56:51

[ 154 ] v. burial practices during the late third millennium bc

Cleary CW3.indd 154

27/05/2016 14:57:01

Cleary

[ 155 ]

5.5 Cist-like compartments in wedge tombs: (a) Labacallee, Co. Cork (after Leask & Price 1936, plate XII); (b) Largantea, Co. Derry (after Herring 1938, fig. 1); (c) Loughash/Cashelbane, Co. Tyrone (after Davies & Mullin 1940, fig. 2); (d) Ballyedmonduff, Co. Dublin (after Ó Ríordáin & de Valera 1952, plate XIII); (e) Lough Gur, Co. Limerick (after Ó Ríordáin & Ó h-Iceadha 1955, fig. 2);

(f) Baurnadomeeny, Co. Tipperary (after O’Kelly 1960, fig. 1); and (g) Ballybriest, Co. Derry

(after Hurl & Murphy 2001, fig. 4)

Cleary CW3.indd 155

27/05/2016 14:57:08

[ 156 ] v. burial practices during the late third millennium bc

of 2470–2286 cal BC (Table 5.2). The badly disturbed wedge tomb at Ballyedmonduff, Co. Dublin, also had a large cist-like compartment to the rear of the main chamber and although Beaker pottery was recovered from this area, as well as the main and antechamber (Figure 5.5d), the only burial evidence was a small quantity of undated cremated human bone from the main chamber (Ó Ríordáin & de Valera 1952). A small cist-like feature that contained ‘probably human cremated bone and pottery’ had also been constructed in the antechamber of the disturbed wedge tomb at Lough Gur, Co. Limerick (Figure 5.5e), where early-style Beaker pottery was also recovered (Ó Ríordáin & Ó h-Iceadha 1955, 36, 46–7). It is unclear from the report if this compartment contained Beaker pottery and the cremated bone has not been dated, but five of the nine non-burnt individuals uncovered (three adults, a child and an infant) did return radiocarbon dates spanning the period 2480–1900 cal BC and some at least may be contemporary with the Beaker pottery. At least a further three individuals (two adults and an infant) probably represent slightly later Early Bronze Age use, perhaps in association with the Bowl and Vase Tradition pottery recovered (Brindley & Lanting 1991/92b, 24). At Baurnadomeeny, Co. Tipperary, a cist-like compartment in the antechamber contained the cremated bone of an adult female and sherds of Beaker pottery (Figure 5.5f; O’Kelly 1960; Carlin 2011a, 2:223), while a similar interpretation might be given to the stone-lined pit located outside the kerb encircling the chamber of a wedge tomb at Ballybriest, Co. Derry (Figure 5.5g; Hurl & Murphy 2001). Although only a flint blade was recovered from the latter pit, the chamber and antechamber contained sherds from six Beaker vessels and cremated bone representing the remains of at least seven people (including an adult male, an adult female, a child and an infant). A sample of cremated bone from the main chamber, and the same deposit that contained sherds from four late-style Beakers, returned a late date of 2141–1882 cal BC, and a second sample of cremated bone from a pit dug into the antechamber returned a comparable date of 2120–1769 cal BC (Table 5.1). This pit lacked Beaker pottery and may indicate that at least some of the interments could represent a slightly later phase of burial not associated with the deposition of the Beaker pots; more dating is required. Although as yet unpublished, the excavation of a wedge tomb at Aughrim, Co. Cavan, reportedly revealed both cremated and non-burnt bone in association with Beaker and Bowl Tradition pottery, although the latter were high up in the filling of the chamber and probably represent a later stage of deposition, while information is also currently lacking on the ‘three cist burials inset against the inner edge of the kerb’ (Channing 1993; Brindley 2007, 51). Megalithic cists and some recent discoveries The construction and use of perhaps comparable above-ground ‘megalithic cists’ (Figure 5.3) during this period is attested by an example at Furness, Co. Kildare, which may have originally been covered with a mound12 and was marked by a large standing stone. The cist reportedly contained the cremated remains of two adults (one male, one female) accompanied by plain 12 There is no mention in the report that the cist was inserted into a pit and it was noted that the top of the side stone (no capstone was recovered) was ‘just concealed beneath the surface of the earth’ and that they quickly realized they ‘were dealing not with a natural land surface but with made earth’ (Macalister et al. 1913, 353–4).

Cleary CW3.indd 156

27/05/2016 14:57:08

Cleary

[ 157 ]

5.6 Sections through megalithic cists and cairns excavated at (a) Coolnatullagh, Co. Clare (after Eogan 2002, and (b) Moneen, Co. Cork (after O’Kelly 1952, plate XXXIV)

fig. 8; photograph with permission from James Eogan)

Beaker pottery, part of a two-holed stone bracer and a possible disc bead (Macalister et al. 1913; Carlin 2011b, 92). Beaker pottery was also recovered from cairns at Moneen, Co. Cork, and Coolnatullagh, Co. Clare (Figure 5.6), where non-burnt bone from associated cists returned dates of 2260–2140 cal BC and 2460–2140 cal BC respectively (Table 5.1; Brindley et al. 1987/88, 13; Eogan 2002, 147–8). The central cist at Moneen contained the partial remains of two adult inhumations as well as cremated bone (subadult) that may have been contemporary, although it was interpreted by the excavator as secondary (O’Kelly 1952, 124– 6). Similarly, in the disturbed and only partially excavated cist at Coolnatullagh, the same fill contained non-burnt bone representing an adult and a child (scapula only) and cremated bone from an adult (Eogan 2002). At Lyles Hill, Co. Antrim, a similar cist at the centre of a cairn, built over earlier Neolithic activity, contained sherds of an ‘anomalous necked Beaker’ and the cremated remains of an adolescent aged 12–16 years13 (Evans 1953, 10; Case 1961, 202, 13 A cremated femur fragment from an individual aged over 16 years (Jonny Geber pers. comm.) returned a date of 1877–1644 cal BC (Table 5.2) and may suggest that this ‘megalithic cist’ is in fact Early Bronze Age and perhaps therefore contemporary with the three ‘secondary cists’ (Evans 1953). However, the context from which the sample was taken is not without question; it was labelled as ‘paved floor of chamber (254)’ which might represent the ‘crazy-paving’ of the central grave but without an original excavation report with corresponding context numbers this is tentative and further research, including a full reassessment of the human remains, is necessary before any conclusions can be drawn.

Cleary CW3.indd 157

27/05/2016 14:57:09

[ 158 ] v. burial practices during the late third millennium bc

224), while at Gortcorbies, Co. Derry, a cist covered by a cairn contained a small amount of cremated human bone, fragments of at least seven late-style Beaker vessels, including two almost-complete vessels, and a miniature Bowl (May 1947; Ó Ríordáin & Waddell 1993, 96; Carlin 2011b, 91). In 2008 a burial monument excavated at Brackagh, Co. Derry comprised a low cairn over a trapezoidal arrangement of 11 large post-holes surrounding a figure-ofeight-shaped stone setting around two small cists, with cremated bone representing two individuals in each cist, and to the south-west, a small rectangular arrangement of stones interpreted as an unused chamber (O’Regan et al. 2009). It was noted that the monument combined constructional elements of wedge tombs, the cairn being trapezoidal in plan with an entrance at the widest point, and Linkardstown-type cists, having concentric stone arrangements around a central cist(s), as well as other Neolithic monuments in the use of large timber posts (ibid., 32). Radiocarbon dating of the cremations suggested that the bone from Cist 1 (2621–2486 cal BC) was c. 100 years older than the bone from Cist 2 (2484–2342 cal BC), perhaps indicating continued use from the Late Neolithic into the Chalcolithic, but further dating is necessary in order to fully understand this site (Table 5.1).   Carlin (2011b, 91), developing the ideas of O’Kelly (1952, 123–4) and Cooney and Grogan (1994, 86) on megalithic cists, suggests that many of these sites can be classified as submegalithic cists and may belong to the wedge tomb tradition or at least have affinities with the cist-like compartments associated with many wedge tombs. They might also be seen in light of that which had gone before, perhaps adapted reflections of Linkardstown-type tombs. Similarities might also be drawn with the smaller wedge tombs, examples of which are common in the Burren, Co. Clare, where the difficulty of distinguishing between true megalithic tombs and long cist graves has been highlighted (de Valera and Ó Nualláin 1961, 101–2; Jones et al. 1996, 99–100), or indeed the simple, box-like structure of the wedge tombs recorded along the Mizen Peninsula, Co. Cork, although these are generally without cairns (O’Brien 1999, 84).   A small number of Chalcolithic burials in pits have also been identified as a result of recent large-scale infrastructural projects and radiocarbon dating programmes (Figure 5.3). Carlin (2011a, 1:150–3; 2011b, 88) suggested up to eight examples but also acknowledged that the majority of these were highly problematic with the association between the Beaker sherds and the bone often quite poor, the bone not dated directly or only identifiable as probably or possibly human. Subsequent analysis of pit deposits at Monadreela, Co. Tipperary, for example, revealed the bone to be animal (Richard O’Brien pers. comm.). It is only by applying new methodologies to such assemblages where bone is highly fragmented and mixed, such as collagen peptide mass fingerprinting, that it will become possible to distinguishing between human- and animal-derived bone fragments (Buckley & Kansa 2011). Mount (2012) similarly highlighted a number of pits with potential Beaker period burials, but emphasized that the burnt bone included in most of these pits tended to be very small and fragmentary, making it difficult to positively identify as human, as at Broomfield, Co. Dublin and Brownstown, Co. Kildare, where at least some of the bone was identifiable as animal. At Lismullin, Co. Meath, however, at least two pits and one undated spread or sealing layer over prehistoric pits did produce cremated bone identifiable as human (O’Connell 2013, table 2.2). One of these pits also contained pottery sherds representing two domestic Beaker vessels and hazel

Cleary CW3.indd 158

27/05/2016 14:57:09

Cleary

[ 159 ]

charcoal returned a complementary date of 2471–2298 cal BC, however, two sherds of an Early Neolithic carinated bowl and a broken polished macehead of Bush Barrow type were also present (Table 5.1; ibid., 25, 27). Furthermore, of the 242.2g of burnt bone, just 35.5% or 86g was identifiable as representing a single individual. While this might indicate a Beakerassociated burial it remains uncertain as an effort to date the human bone failed. This is particularly unfortunate in light of the mixed nature of the deposit and the general dating of Bush Barrow type maceheads to the Early Bronze Age (Simpson 1996, 69; Needham & Woodward 2008). Similarly in the second pit, just 0.03% or 0.1g of the cremated bone was identifiable as human (skull and tooth fragments) and associated hazel charcoal returned a date of 2457–2205 cal BC but the fill also contained three sherds of Middle Neolithic broad-rimmed bowl (Table 5.1; O’Connell 2013, 27). The cremation of an adult male in a pit at Harlockstown, Co. Meath (Figure 5.7b), was also dated by charcoal (ash/pomoideae/ alder) to the Chalcolithic (2397–2136 cal BC) and may have related to an adjacent post-hole containing two sherds of Beaker pottery and a broken chert blade (Table 5.1; O’Connor 2008). This pit was also enclosed by a ring-ditch, where a cattle scapula from the basal fill returned a date of 2132–1886 cal BC (Table 5.1); up-cast from this ditch may have been used to create a low stone mound into which two later crouched inhumations with Food Vessel Bowls were placed. Approximately 3 m south-east of the ring-ditch an undated deposit of cremated bone, representing an adult, could also be contemporary with the earlier burial evidence. At Corbally, Co. Kildare, a pit with evidence for in situ burning contained sherds from two Beaker vessels, a chert barbed-and-tanged arrowhead, four burnt but worked flints and 23g of burnt bone, identified as representing both human (mainly long bone) and animal (Purcell 2001).   Directly dating the bone in these types of contexts is essential when trying to untangle the complexities of the burial practices at this time. At Piperstown (Site K), Co. Dublin, for example, a pit roughly centrally located within a cairn circle contained the cremated remains of an adult male and a heat-damaged flint flake, with oak charcoal returning a Chalcolithic date of 2573–2344 cal BC (Table 5.2; Rice 2006; Kim Rice pers. comm.). A subsequent date directly on a fragment of cremated skull returned an earlier date of 2832–2457 cal BC (Table 5.2), suggesting the site most likely represents Late Neolithic burial within this upland archaeological complex (Rynne & Ó hÉailidhe 1965–6). At Treanbaun, Co. Galway, an irregular stone-lined pit contained the remains of a Beaker vessel, somewhat disturbed but interpreted as originally inverted over the cremated remains of an adult (Figure 5.7a; Muñiz-Pérez 2014; Coughlan 2009). However, a cremated long bone fragment returned the unexpectedly late date of 1886–1667 cal BC, which is more comparable to the date from a stratigraphically later cist that also contained the cremated remains of an adult, accompanied by a chert plano-convex knife (Tables 5.1 & 2; Muñiz-Pérez 2014, 133). The dated sample from the pit may therefore represent contamination and additional dating is required to determine the true association between the cremated bone and the Beaker vessel. Notably, about 700 m to the west in the same townland a multi-phased site included pits dating from the Early Neolithic to the Late Bronze Age, many of which contained small quantities of burnt bone, but only a small number of which could be positively identified as human (Muñiz-Pérez & Bermingham 2014). A date of 2570–2347 cal BC (Table 5.1) was returned for undiagnostic

Cleary CW3.indd 159

27/05/2016 14:57:09

[ 160 ] v. burial practices during the late third millennium bc

5.7 Possible Chalcolithic/Beaker period burials in pits at (a) Treanbaun, Co. Galway (after Muñiz-Pérez 2014, illus 4.8.3–4), (b) Harlockstown, Co. Meath (after O’Connor 2008, fig. 7), and (c) Mell, Co. Louth (after McQuade 2005, fig. 13)

Cleary CW3.indd 160

27/05/2016 14:57:11

Cleary

[ 161 ]

burnt bone from a pit sealing layer, but this was the only potentially Beaker period activity identified, while part of an Early Bronze Age Food Vessel Vase was recovered from another pit.   One inhumation burial where the human bone has been directly dated to the Chalcolithic was uncovered at Mell, Co. Louth, where the bone returned a date of 2488–2206 cal BC (Table 5.1; Figure 5.7c; McQuade 2005). This adult female had been placed in a partially stone-lined sub-rectangular pit, however, her body posture was somewhat unusual; she was interred in a prone flexed position, orientated east–west with the head to the west, the legs bent at the knees, the left arm flexed under the chest and a large stone across the back (ibid., 36; see Rogers 2013, 20). Although there was no associated Beaker pottery, occupation layers 60 m to the south-east contained sherds from 38 Beaker vessels, including a polypod bowl. A shallow pit at Cookstown, Co. Meath contained the unaccompanied inhumation of an older adolescent/young adult, crouched on the left side with the head to the north, presumed to be Early Bronze Age in date but with a lack of surviving collagen preventing 14C dating (Clutterbuck 2009). Notably, the grave was cut by a small pit containing two sherds of Beaker pottery, two flint flakes, a flint bladelet and oak charcoal that returned a date of 2853–2289 cal BC (Table 5.1). In light of the burial at Mell it is possible that Cookstown could represent another Chalcolithic crouched inhumation. At Ballynacarriga, Co. Cork, occupation during the 3rd millennium BC was suggested by a large assemblage of Grooved Ware associated with possible structures, a smaller quantity of Beaker pottery and diagnostic lithics, as well as associated 14C dates (Lehane et al. 2011). Approximately 53 m to the south was a series of Early Bronze Age burials, several associated with vessels of the Vase Tradition. Within this burial cluster, however, was a partially stone-lined pit/cist containing a neat pile of cremated bone representing an infant less than one year old, and a juvenile aged 3–7 years, with a bone sample returning a date of 2460–2206 cal BC (Table 5.1). A second cist, c. 4.5 m to the south-west, was empty, or the bone was not preserved, but Pomoideae charcoal returned a comparable date of 2461–2211 cal BC (Table 5.1). Directly dating some of the remaining burials without diagnostic pottery could confirm Chalcolithic burial activity, suggesting a funerary site initially associated with the nearby Beaker settlement but also used into the Early Bronze Age (but also see below the ‘old wood’ effect on cremated bone).   Recognizing Chalcolithic burial in caves is similarly difficult without extensive programmes of scientific dating but Dowd (2015, 125) has noted that many of the caves that were places of Neolithic ritual activity played a similar role in the Bronze Age, some of which included the deposition of human remains (Figure 5.3). Dowd (ibid., 62–5) also acknowledges that due to the nature of caves, complete skeletons and intact burials of any date are rare, leaving the dispersed disarticulated bones and deposits of cremated remains open to interpretation, ranging from remnants of formal burial to natural deaths, excarnation, animal scavenging and deposits carried in by moving water. At Brother’s Cave, Co. Waterford, for example, a sherd from a domestic Beaker vessel and two adult molars (possibly from the same individual) that returned dates of 2343–2026 cal BC and 2270–1977 cal BC may indicate some Chalcolithic activity, although multi-phased revisiting of the site was also evident, including a sherd from an Early Bronze Age Cordoned Urn and a piece of bronze that may represent a dirk or rapier (Table 5.1; ibid., table 6.1, 129). While the teeth are all that now remain of the assemblage, records of the early 20th-century AD excavations

Cleary CW3.indd 161

27/05/2016 14:57:11

[ 162 ] v. burial practices during the late third millennium bc

note the discovery of over 50 human bones representing at least one foetus, two children, a young adult and cremated remains of a second adult (ibid., 129). Similarly, at the nearby Oonaglour Cave a mixed assemblage included sherds from two Beaker vessels and over 500 human bones representing adults and juveniles; the surviving 74 bones have yet to be dated (ibid., 131). Also in Co. Waterford, investigations at Ballynamintra Cave in 1879 recovered 51 human bones, primarily of small skeletal elements representing at least four individuals, including an adult dated to 2275–1980 cal BC, and the nature of the anatomical remains has led to the suggestion that this site could have been a designated place for excarnation (Table 5.1; ibid., 105, 113–4, 126–9). A single amber bead was also recovered from this cave but it is not possible to date more closely as such beads occur throughout the Bronze Age into the Iron Age. Excavations at Killuragh Cave, Co. Limerick, revealed evidence for the deposition of small amounts of disarticulated human bones over a long period of time, including an adult mandible from a pit at the rear of the cave that returned a date of 2344–2036 cal BC, while a second individual represented by an adult radius was dated to 2028–1890 cal BC and may relate to sherds from three Vase Urns (Table 5.1; Woodman 1997; Dowd 2015, table 6.1, 131). At the base of a swallowhole at Pollthanacarra, Co. Fermanagh, the remains of four adult individuals were uncovered and although only 137 bones were retrieved the presence of most anatomical regions and mainly large skeletal elements indicated that these were probably intact, fleshed bodies, most likely ritually deposited (Dowd 2015, 154–6). The left humeri of three of these adults returned dates of 2430–2137 cal BC, 2279–2036 cal BC, and 2132–1894 cal BC, spanning the Chalcolithic into the Early Bronze Age (Table 5.1). Local adaptions It is clear from this review of burial between c. 2450 and 2200/2150 BC that Ireland did not adopt the ‘typical Bell Beaker funerary rite’; instead it appears to have followed regional interpretations of what was going on elsewhere, as well as continuing existing burial practices.   If we consider what went immediately before, as yet the only confirmed burial evidence directly associated with Grooved Ware pottery is limited to the reuse of passage tombs at Knowth (see above; Eogan & Roche 1997; Schulting et al. forthcoming) and an unpublished report of a pit burial at Lusk, Co. Dublin (Grogan & Roche 2010, 34). As mentioned previously, a cluster of pits at Treanbaun, Co. Galway, were also dated to the Late Neolithic, including one pit where cremated human bone was positively identified and returned a date of 2865–2575 cal BC, while another pit contained sherds from a single Grooved Ware vessel (Table 5.1; Muñiz-Pérez & Bermingham 2014, 124). Similarly, the Late Neolithic date detailed above for the cremated adult male at Piperstown Hill strengthens the connection between that burial and the nearby structures with rectangular, stone-edged hearths reminiscent of Grooved Ware buildings elsewhere (Smyth 2014, 92–3). Late Neolithic dates have also been returned for non-burnt human bone in two caves; adults in both Elderbush Cave, Co. Clare and Ballynamintra Cave, Co. Waterford (Dowd 2015, 97) and for the inhumation of an adult female in a cist at the eastern edge of a cairn at Killarah, Co. Cavan (2620–2470 cal BC; Table 5.2; Ó Ríordáin 1933). While future 14C dating programmes will no doubt

Cleary CW3.indd 162

27/05/2016 14:57:11

Cleary

[ 163 ]

reveal additional examples for this period, from the low number of confirmed burials to date it must be surmised that the majority of the dead continued to be disposed of in archaeologically invisible ways.   Placing the approximately 560 known wedge tombs to one side, from the examples detailed above it is evident that the number of confirmed burials continued to remain low until 2200/2150 BC. This is roughly comparable to Britain where the number of Beaker burials between 2500 and 2250 cal BC was small relative to later, and Beaker culture was interstitial with Final Neolithic Grooved Ware culture (Needham 2005). The general lack of nonpottery grave goods in Ireland is also notable, particularly as typical Beaker artefacts such as copper daggers, stone bracers and gold adornments are present in non-funerary contexts. This suggests that Beaker objects in Ireland did not share the same associations with the dead as they did elsewhere (Carlin 2011b, 97). In considering wedge tombs, now modelled as suddenly appearing in both the south and north of the island sometime during the period 2540–2300 BC, their initial use for burial in the period prior to 2200/2150 BC is as yet limited to just four sites (Labbacallee, Largantea, Lough Gur and Loughash/Cashelbane) where both cremated and non-cremated human bone has been dated accordingly. Although at least an additional six wedge tombs (Aughrim, Ballybriest, Ballyedmonduff, Kilnagarns Lower, Moytirra, and Loughash) contained Beaker pottery (see Schulting et al. 2008). The importance of recognizing these as monuments with open and accessible chambers that were re-visited and re-modelled into the Early Bronze Age and beyond, however, has long been demonstrated and supported by both diagnostic pottery and human remains dated to the period after 2200/2150 BC (ibid.; Brindley & Lanting 1991/92b; O’Brien 1999). While further dating on human remains and more modern excavations of these monuments is necessary to help understand their role as burial sites it is well attested that there were ‘varied rituals carried out both within, in front of, and adjacent to the monuments’, suggesting they played a significant role as ritual foci for small-scale segmentary societies attempting to legitimize their claims to status, just one element of which was through burial (Jones et al. 2015, 5; O’Brien 1999; 2004; 2012b).   Overall, burial evidence from the Late Neolithic until c. 2200/2150 BC remains low, alongside a limited use of grave goods. Could this suggest that many aspects of funerary practices elsewhere were simply ‘lost in translation’ or ‘ritually rejected’? Neither seems appropriate given the obvious willingness to adopt Beaker pottery and metalworking knowhow. Is it better, therefore, to interpret the burial evidence as maintaining pre-existing ideologies while drawing upon elements of wider traditions, perhaps because there was no immediate pressure to conform? The importance of the Irish sources of copper and possibly gold/gold-working (see Standish et al. 2015) at this time should not be underestimated and perhaps these influences meant that ‘Irelanders’14 were in a strong position to maintain their existing social practices, even embellish them through the construction of wedge tombs and megalithic cists, rather than import wholesale new ideologies from other regions.   Of those that were buried, the use of well-defined space above ground, in various forms of stone-lined compartments placed on but not in the earth, seems to have predominated. 14 This term is taken from Mallory (2013, 42).

Cleary CW3.indd 163

27/05/2016 14:57:11

[ 164 ] v. burial practices during the late third millennium bc

Architecturally, the ‘megalithic cist’ contexts compare well with the evidence of tomb reuse in other areas of Atlantic Europe and may relate to a re-envisioning of a longstanding belief system, whereby ‘open’ monuments that served a wider community were both appropriated for and replaced by ‘closed’ burials for a restricted section of the population (see Bradley 1998, chapter 9). The construction of new monuments in Ireland, both wedge tombs and large above ground cists, also seem to represent a local expression, perhaps an innovation that was more acceptable to the surrounding communities as it looked to past traditions but maintained the appearance of individualized, closed graves. While it is often complicated by a lack of scientific dating and evidence for continual reuse, it is evident from the examples cited here that the burial of just one individual or the ‘individualization’ of burials occurred at several sites, such as Mell and Labbacallee. Other sites have potentially more than the one dated burial, such as Loughash/Cashelbane and Ballybriest wedge tombs, but whether or not these represent successive single internments or collective burial remains to be proven. What is evident is that the number of individuals was never substantial; the seven cremated bodies at Ballybriest is the largest potentially collective burial, and at least one of these was ‘individualized’ through placement in a pit dug into the antechamber (and see above re dating).   As yet very few definite Beaker burials have been found in pits or caves. Of those in pits detailed above, the examples of definite human cremations with associated Beaker pottery and/or charcoal radiocarbon dates, as at Lismullin and Harlockstown, represent discrete burials of single individuals but need to be confirmed by direct dating on the human remains, as aptly demonstrated by the dates from Piperstown. The directly dated female inhumation at Mell, while not associated with any diagnostic artefacts, is comparable to the head to the west orientation of female Beaker burials previously observed in east Yorkshire and north-east Scotland (Tuckwell 1975; Shepherd 1989; 2012). The prone position is however more unusual, with perhaps the closest parallel from Barrow 36f at Wilsford in Wiltshire, where a young adult female placed on the original ground surface was ‘curiously contorted’; positioned front down with head twisted over on its right, looking eastwards, arms slightly akimbo but folded tightly upwards so forearms and hands partly beneath the body and legs closely bent at knees and lying to the west (Grimes 1964). This burial remains undated however, and the small pottery sherds from the mound of the barrow, including Peterborough ware and Group B Beaker, could be residual.   The contemporary use of both cremation and inhumation is also notable; excluding the ambiguous nature of the caves, 60% of the potential sites discussed here had cremation only, 20% inhumation only and the remaining 20% had both rites. This is in contrast to the majority of other Atlantic regions where inhumation was dominant and probably represents another important local reflection of past traditions, certainly there does not appear to be any simple, underlying demographic explanation (see O’Brien 1999, 203–7; Schulting et al. 2008, 11). No substantial difference in the way non-burnt and cremated bodies were deposited has been recognized as yet but this is perhaps hampered by the small sample size, limited radiocarbon dating and later disturbance, which leave any suggestions tentative. In the court tomb at Audleystown for example, it is notable that of the samples dated to the Chalcolithic, the cremated and non-burnt bone were deposited in different chambers but in the megalithic cists at Moneen and Coolnatullagh both seem to have been deposited

Cleary CW3.indd 164

27/05/2016 14:57:11

Cleary

[ 165 ]

together. The question of complete versus partial burial also requires further research. The inhumation of an adult female in the cist-like compartment of the wedge tomb at Labbacallee, for example, was partially disarticulated, including the probable deposition of the skull in the main chamber. The fragmented nature of most of the Beaker pottery recovered from these burial contexts is also notable, particularly in light of similar practices in parts of north-western France and Iberia (Salanova 2004; Catriona Gibson pers. comm.), although again taphonomic processes should be carefully considered. The multiple sherds representing at least six Beakers in the chamber and antechamber of the wedge tomb at Ballybriest, for example, suggested that the vessels were probably deposited whole and broken in situ, while in the wedge tomb at Largantea, although two complete Beaker vessels were deposited in the cist-like compartment in the antechamber, the chamber contained just a small number of sherds representing a further four vessels, raising the question of what happened to the rest of these vessels? A fusion of traditions: The earliest Bronze Age burial practices From c. 2200/2150 BC, as the use of Beaker pottery was greatly diminishing in Ireland, pottery of the Bowl Tradition or Food Vessel Bowls emerged, followed shortly later by vessels of the Vase Tradition (see Brindley 2007; Bayliss & O’Sullivan 2013). This new ceramic style is long argued to have been influenced by the Beaker pottery that had dominated for the previous 300 or so years (Ó Ríordáin & Waddell 1993, 39–44; Brindley 2007, 250–1). It also appears to have been an innovation that deliberately replaced Beakers while at the same time related to the adoption of single crouched inhumation burial practices, some of which incorporated aspects of the suite of grave goods indicative of Beaker use elsewhere. Three V-perforated buttons, for example, were recovered from the chamber of the Mound of Hostages passage tomb at Tara, Co. Meath, at least one of which appeared to accompany a crouched inhumation (see below; O’Sullivan 2005) and a typologically early tin-bronze dagger was associated with ‘unburnt bones’ and a tripartite Bowl in a cist at Corkey, Co. Antrim (Waddell 1990, 22, 45; Brindley 2007, 55). Like Beakers elsewhere, these new ceramic styles are predominantly recovered from funerary contexts, specifically cists and pits, but accompanying burial rites were more fluid. Collective but also ‘individualized’ burials appear to have continued and both cremation and inhumation remained popular. It has been estimated, for example, that 57% of Food Vessel Bowls were found with cremation rather than inhumation burials (Ó Ríordáin & Waddell 1993, 19). As with the inhumation burials, some cremations associated with Bowls also incorporated ‘classic Beaker grave goods’, such as an early style bronze dagger and awl (Group 1B) from a cist at Carrickinab, Co. Down and the V-perforated button from a cist at Kinkit, Co. Tyrone (Figures 5.8a & b; Collins & Evans 1968; Glover 1975). The latter was made from bone and may represent a local emulation of objects more commonly of jet and jet-like material (see Sheridan et al. 2002). Although rare, some of the British bone examples of early date are speculated to be brought in from the continent (Woodward & Hunter 2015, 155), and indeed both bone and ivory V-perforated buttons are known from Iberian Bell Beaker sites (Schuhmacher et al. 2013).   Individuals subject to different mortuary treatments also continued to be buried together.

Cleary CW3.indd 165

27/05/2016 14:57:11

5.8 Early Bronze Age Cists at (a) Carrickinab, Co. Down (after Collins & Evans 1968, figs 2–3 and plate I),

(b) Kinkit, Co. Tyrone

(after Glover 1975, figs 45–6),

(c) Ballybrennan, Co. Westmeath

(after Hartnett & Prendergast 2011, figs 3.177–8), and

(d) Straid, Co. Derry

(after Brannon et al. 1990, fig. 6 and photograph © Crown copyright, reproduced with the permission of the Controller of HMSO and Northern Ireland Environment Agency)

At Ballybrennan, Co. Westmeath, for example, a cist contained the crouched inhumation of an adult male with a Bowl close to his head and the cremated remains of an adult male carefully placed over the knees (Figure 5.8c; Hartnett & Prendergast 2011). Dating both the non-burnt (2140–1916 cal BC) and cremated bone (2190–1926 cal BC) confirmed that they could have been contemporary deaths selected for different burial rites (Table 5.1). Similarly, in a cist at Straid, Co. Derry, cremated bones from two adult females were found beneath the left elbow and the right leg bones of the flexed inhumation of an adult male with a Bowl near the head (Figure 5.8d; Brannon et al. 1990; Brindley 2007, 56). Resembling the discoveries in Cists 2 and 4 at Fourknocks II (see above), the non-burnt individual returned a Chalcolithic date of 2458–2202 cal BC but the cremated remains (individual not specified) returned a substantially later date of 2199–1960 cal BP (Table 5.1). This difference

Cleary CW3.indd 166

27/05/2016 14:57:12

Cleary

[ 167 ]

between the older than expected collagen date and the younger carbonate date was, however, interpreted as probably due to the effect of freshwater fish in the diet (see Lanting & van der Plicht 1998, 160; Brindley 2007, 56). This is something that needs to be considered when dated samples are interpreted, although in other examples, such as those detailed above for Fourknocks II, there may be no indicators in the stable isotope measurements that a reservoir effect has altered the ages. This aside, the location of cremated human remains, and indeed disarticulated non-burnt bone, in locations where one might normally expect to find grave goods, such as by the head, hands and feet, also raises questions about the distinction between people and objects in these contexts. Is it possible that this interplay between contrasting, perhaps even competing, burial practices was another means of reflecting the wider context of social and cultural changes?   These funerary practices therefore appear to represent an Irish version of the contemporary later Beaker practices in Britain. Parallels can be drawn to east-central Scotland, where single burial practices did not occur in great numbers until after 2200/2150 BC and only then in association with a greater range of funerary traditions, most notably Food Vessels, bronze daggers, jet-like ornaments, the adoption of cremation, a wide range of funerary monuments and the reuse of older monuments (Curtis & Wilkin 2012, 249). Comparably, on the Isle of Man the evidence for Beaker burial is limited to a single late-style Beaker from a cist at Baroose in Kirk Lonan on the east side of the island, but Food Vessel Bowls and Vases were adopted alongside both crouched inhumation and cremation burial rites and older monuments were reused (Woodcock 2008). Continued reuse of earlier burial monuments As with the centuries before c. 2200/2150 BC, there was not a complete abandonment of previous funerary practices, but again an adaptation of that which had gone before. The reuse of earlier monuments continued, with the megalithic tombs, mounds and cairns of at least 53 sites15 revisited in the Early Bronze Age, including several wedge tombs (Mount 2013, table 1). A variety of pottery styles and burial rites are recorded and the details require further analysis and chronological reassessment outside the scope of the current paper, but some examples, in relation to the ‘transition’ between Beaker and Bowl Tradition practices are worth highlighting. The continued and indeed intensified ‘individualization’ of funerary deposits is notable and many, placed in cists and pits, were inserted into the mounds and cairns, as opposed to simply being deposited in the chambers where they could have replaced earlier burials. As discussed by Gibson in Chapter Three, there is a question of respecting the earlier burials but also accessibility and the deliberate closing of monuments. This requires further work, but as yet there is little evidence for this practice in Ireland (see O’Brien 1999, 199) and indeed the interiors of many tombs show evidence of later visitors 15 Notably, ‘the most popular type of old monument reused was the passage tomb, with fifteen examples’ (Mount 2013, 185). This contrasts with the Chalcolithic where, as detailed above, three possible sites were reused, however the number of Beaker burials generally is so small that it is not advisable to draw any conclusions on the changing suitability of passage tombs as repositories for the dead.

Cleary CW3.indd 167

27/05/2016 14:57:12

[ 168 ] v. burial practices during the late third millennium bc

and disturbance, which may equally obscure the true sequence of events.   At the passage tomb of Moylehid on Belmore Mountain in Co. Fermanagh, for example, not only was a cist containing a Food Vessel Bowl and cremation burial inserted into the cairn, but a second Bowl with cremated bone was placed on a stone slab in one of the chamber recesses, possibly on top of the covering stone and thereby respecting the earlier burials (Coffey 1889–1901, 663; Herity 1974, 231). Based on Bayesian modelling the radiocarbon dates from the passage tomb at Tara, Co. Meath, it has been argued that the first phase of Early Bronze Age reuse consisted of deposits in the chamber commencing 2120–1980 cal BC, followed by a sequence of burials placed into the covering mound (Mount 2013). Furthermore, despite the large quantity of bone recovered from the chamber of this tomb, at least four later burials appear to have been further ‘individualized’. Crouched inhumation burial 18, associated with a Bowl and at least one V-perforated button, was placed in a pit dug into the chamber fill, above this and separated from it by a paving of stones on a ‘clean deposit of yellow clay’ was a second crouched inhumation burial 19, also associated with one, possibly two, Bowls (O’Sullivan 2005, 107–12). Cremation burial 25, associated with a Food Vessel Vase, was ‘nestled in a cist-like arrangement of stones’ and burial 24, consisting of an Encrusted Urn inverted over a cremation burial with an accompanying Vase, was similarly ‘protected by a cist-like arrangement of slabs’ (ibid., 91, 105–7). Food Vessel Bowls have been recovered from the chambers of at least five wedge tombs that also witnessed Beaker associated burial/activity; Loughash/Cashelbane, Aughrim, Kilhoyle, Lough Gur and Largantea. These pottery deposits most likely represent burials but data on the specific contexts is often lacking and further radiocarbon dating is needed to confirm contemporary burials (see Brindley & Lanting 1991/92b, 24 for potentially compatible dates on human remains at Lough Gur). At Largantea, for example, the Bowl was deposited in the antechamber and although some cremated bone was recovered from this area it was interpreted as disturbed from the main bone deposit in ‘Chamber II’ and remains undated, as do four of the remaining eight individuals identified in the bone assemblage (see Schulting et al. 2008). Coexisting burial strategies From the earliest stage of Bowl Tradition pottery in Ireland burials were also being placed in both cists and pits with no earlier monument associations (Brindley 2007, 250). In contrast to the majority of the known Chalcolithic burials and indeed the Early Bronze Age burials inserted into older monuments, these individuals were placed directly in the ground. Some were within stone settings and others were not; these details may represent different stages of adoption or local adaption of burial ideologies or possibly even competing ideologies. As observed by Quinn and Kuijt (2013, 200) in relation to the Early Bronze Age reuse of the Mound of the Hostages passage tomb, ‘coexisting strategies may indicate numerous things, including variation in lineage membership, community use of the cemetery, social roles, political competition, identities and cultural mechanisms for regulating mortuary treatment. The periods where one strategy persists may indicate stability in the same variables’. Some of these tempos of burial practice therefore require further exploration beyond the scope

Cleary CW3.indd 168

27/05/2016 14:57:12

Cleary

[ 169 ]

of this paper, particularly in relation to grave good selections and exclusions. For example, does the inclusion of traditional ‘Beaker-associated artefacts’ with Food Vessel Bowls, both inhumations and cremations, represent deliberate burial strategies?   An interesting example to conclude on is a compartmented cist excavated at Newtown­ stewart Castle, Co. Tyrone (Ó Baoill 2005). The larger compartment contained the cremated remains of an adult female with a Bowl placed immediately beside the pile of bone, a sample of which returned a compatible date of 2196–1951 cal BC placing it in the earliest stage of Bowl use (Table 5.1; Brindley 2007). The smaller compartment contained a second Bowl placed on top of the cremated remains of an adolescent with an associated hollowbased flint arrowhead; a sample of this bone returned the surprisingly early date of 2476– 2213 cal BC (Table 5.1). Although additional dating to confirm the latter date is advisable, particularly in light of the ‘old wood’ effect whereby 14C dating of cremated bones can reflect the burning atmosphere of the cremation fire (see Hüls et al. 2010; Snoeck et al. 2014), the site does offer an intriguing possibility. Could the cremated remains, and probably the arrowhead, have been removed from a pre-existing Chalcolithic burial for inclusion with a burial associated with the ‘new style’ of pottery? A need to create this link is not unimaginable given the evidence for Beaker burials and Food Vessel burials in some of the same funerary contexts, such as wedge tombs. In a similar fashion, the presence of Beaker sherds in association with Bowl Tradition burials indicates a relationship between the two styles, with the Beakers perhaps used as heirlooms or relics (q.v. Woodward 2000). Such a practice might be seen at Moone, Co. Kildare, where a single Beaker sherd was found in the grave of a crouched inhumation dated to 2194–1972 cal BC and accompanied by a Bowl, and a further two Beaker sherds, from two different vessels, were found with a cremation placed in an inverted Vase Urn (Table 5.1; Hackett 2010). Adapting and adopting: Some conclusions on the choices made Despite the as yet limited evidence for seagoing vessels, various other facets of the Irish archaeological record imply that people were clearly engaged in long distance travel from the earliest occupation of the island. This contact, exchange of artefacts, technologies, and ideologies appears to have waxed and waned during early prehistory but one period of intensification was probably from around 2400 BC when the earliest copper mining began at an advanced stage of knowledge and the vogue Beaker pottery was quickly adopted as an allpurpose vessel. In contrast to changing burial practices in many other Beaker-using regions of Europe, the majority of Beaker contemporary burials in Ireland appear to have followed a rite that leaves a limited archaeological record and perhaps reflects an initial continuation of existing practices. This is of course excluding wedge tombs, which thus far have limited evidence for burial pre-2150 BC and may, initially at least, have had other dominant roles, such as indicating social status and as the foci for non-burial ceremonial displays. Where Chalcolithic burial has been discovered, its closest parallels are in the Atlantic tradition of reusing earlier collective graves but also ‘individualizing’ the burials through successive single internments and placement in cist-like compartments and megalithic cists. These types of

Cleary CW3.indd 169

27/05/2016 14:57:12

[ 170 ] v. burial practices during the late third millennium bc

‘individualization’ could therefore be seen as an adaption of the Beaker-associated practice of individual burial, which intensified as one moved away from the Atlantic coastline and into areas where single graves were more common. The inclusion of Beaker pottery and associated artefacts with these burials also varies across the region. As outlined above, definitively identifying the inclusion of Beaker artefacts as grave goods in many of the Irish burial contexts is problematic, particularly given the frequent evidence for later reuse and disturbance, not to mention the high percentage of unprovenanced Beaker artefacts.16   Based on current evidence therefore, the lack of Beaker grave goods and when present, the fragmentary nature of the pots, certainly seems to have more in common with particular parts of Iberia and north-western France, such as the Tagus Estuary, Alentejo, and the Paris Basin (Salanova 2004). In contrast to Ireland, inhumation is the predominant burial rite in other parts of Atlantic Europe, particularly Iberia and north-western France, but a more varied and comparable tradition has been recorded in parts of Britain where cremation and inhumation, both articulated and disarticulated, have been identified (Petersen 1972; A. Gibson 2004). Notably, many partially or fully disarticulated inhumations are also known from Iberia (see Gibson, Chapter Three). In both Ireland and Britain, the two burial rites extend back into the Neolithic and so the practices adopted during the Chalcolithic may reflect the need to amalgamate local traditions with new influences. This argument may also relate to the reuse of older monuments, which were probably seen as a convenient way to connect to the past.   The construction and use of wedge tombs during this period in Ireland may represent a regionalized expression of this need to relate to and even echo earlier constructions. The dating of wedge tombs suggest they are contemporary with the advent of metallurgy and so external influences have long been sought. The insular development of these monuments as a ‘reinvention of a megalithic tradition’ should not be overlooked however (Carlin & Brück 2012, 197; Cooney & Grogan 1994, 84), particularly if one considers the importance of above ground burial and enclosure in stone settings or compartments during this period, while others favour a combination of indigenous development with external influences (O’Brien 2012b). In many earlier megalithic tombs, be they in Ireland or northwestern France, human remains were placed in closed, stone-defined chambers and recesses, sometimes with overlying cairns, and the wedge tomb architecture clearly imitates these traditions, particularly the divisions of chambers and antechambers, the insertion of cistlike compartments and the addition of cairns at some sites. Another form of burial at this time is in megalithic cists, often with associated cairns and covering mounds, and these may represent ‘hybrid’ sites. They are clearly a reflection of earlier monuments, particularly Linkardstown-type burials, but Carlin (2011b, 91) argues that some, such as those excavated at Gortcobies and Moneen, may also belong to the wedge tomb tradition due to their affinities with cists within or at the termini of wedge tombs. During the later 3rd millennium BC, c. 2450–2200/2150 BC, it therefore appears that Ireland was content to maintain preexisting funerary practices but in a reinvented, even embellished manner, some aspects of 16 The majority of stone bracers, for example, are unprovenanced finds without any associations or archaeological contexts (Harbison 1976, 7).

Cleary CW3.indd 170

27/05/2016 14:57:12

Cleary

[ 171 ]

which also drew on new ideologies. This was not because of a lack of interaction with other areas but perhaps because it was an important metal-producing region and the emphasis on conforming was less urgent. Control of copper and possibly gold supplies/gold-working by local groups may have greatly facilitated this ability to selectively incorporate that which was useful or desirable within these emerging beliefs and social changes.   It was not until c. 2200/2150 BC that the ‘classic Beaker’ single crouched inhumation with a suite of distinctive grave goods, including a new pot style17, was finally adopted in Ireland. Although the Bowl Tradition that emerged was probably stylistically influenced by the preceding Beakers they were also a local innovation used to mimic the Beaker burials seen elsewhere. This burial rite was also a syncretized version that incorporated some of the newer funerary trends with elements of the old traditions, such as the use of collective burial, cremation and inhumation. Could this increased willingness to adopt more aspects of the burial practices in use elsewhere represent a move towards more intensive and perhaps extensive interactions and was this at least partially as a result of tin-bronze? The debate on when and indeed how long the transition to bronze took place is still under examination with some suggesting it may have occurred in Ireland around c. 2000 BC (O’Brien 2012b, 220). The introduction of tin-bronze manufacture to north-east Scotland has been placed at c. 2200 BC (Needham 2004) and indeed others have emphasised how subsequent metal recycling may be hiding the existence of very early tin-bronzes (Bray 2012). Perhaps an incremental increase in the need for tin and other resources resulted in more mutualistic relationships developing and the gradual adoption of more external influences reflected in the adapted Beaker burial with the local Food Vessel Bowl. The close connections with Scotland, previously evident in the Middle–Late Neolithic, certainly seem to have been renewed around this time or at least become more archaeologically visible. Some Food Vessels exhibiting clear Irish affinities have been discovered in Scotland, and indeed it is proposed that some contemporary Beakers were also emulating Irish Food Vessels while others were reacting against their profile and form, perhaps suggesting competing identities (Cressey & Sheridan 2003; Wilkin 2010). Current evidence also suggests that exploitation of the Welsh copper mines began around this time, a few decades before or after 2100 cal BC (Burrow 2012, 175). This may have forced Ireland to increase interaction networks, perhaps resulting in a growing need to conform, while simultaneously incorporating expressions of local identity.   The archaeologically visible variations in funerary practices during the later 3rd and into the early 2nd millennium BC can tell us a lot about Ireland’s connections with the wider Atlantic area. These range from a relatively low number of Chalcolithic burials in a predominantly Irish tradition but with an Atlantic twist to a significant increase in funerary activity coinciding with the local development of a new pottery style and an interpretation of Beaker burial ideologies prominent in parts of Britain at the same time. These insular 17 In contrast to Beakers, it seems that initially this new Bowl Tradition pottery was almost exclusively for funerary contexts, although it is proposed that a Bowl Tradition domestic pottery did also develop, perhaps alongside a late form of Beaker pottery (Brindley 2007, 251). At French Furze, Co. Kildare, for example, Bowl pottery from a domestic context was suggested to belong to Brindley’s Stage 3 (Ó Maoldúin 2014b). Therefore, without further refining the dating of these phenomena, the argument remains that the Bowl was initially created specifically for funerary rituals before being adapted for domestic contexts.

Cleary CW3.indd 171

27/05/2016 14:57:12

[ 172 ] v. burial practices during the late third millennium BC

adaptions of ‘correct’ burial rites may reflect the fluctuations in, and perhaps even dependence on, supra-regional interactions. By refining mining dates, for example, we may gain some insights into whether or not the opening and closing of copper mines or the exportation and importation of ore had wider social implications that were ultimately reflected in burial practices. Consequently, the requirement of multiple burial strategies may be indicative of related, complex social structures and multiple identities as and when those contacts intensified or abated. Perhaps further analysis will also uncover more refined regional and temporal distinctions that can highlight specific spheres of influence. There is little doubt that Ireland’s resources and geographical positioning on important routes of travel along the Atlantic façade ensured that the island was linked to diverse communities as the ‘Metal Ages’ began. It seems then that the adoption of wider fashions was deliberately selective and perhaps linked to the supply of metal and gold-work, ultimately resulting in the coexistence of inter-related funerary practices that fused new ideologies with local expressions. Acknowledgements The author would like to acknowledge the assistance of Karl Brady, Dr Niall Brady (The Archaeological Diving Company Ltd.), Dr Kim Rice and Dr Jessica Beckett, whom provided valuable information for various aspects of the paper. Thanks are also due to the various people that assisted in relation to acquiring new 14C dates; Mary Cahill and Maeve Sikora of the National Museum of Ireland, Dr Greer Ramsey of the National Museums Northern Ireland, Laureen Buckley, Dr Jonny Geber, Dr Fiona Beglane, Dr Alan Hawkes, Donald Murphy, Dr Barra Ó Donnabháin and Richard Gillespie. Final appreciation is to Dr Ros Ó Maoldúin, Dr Neil Carlin and Dr Catriona Gibson for useful discussions and guidance and to the other co-editors, Professors Barry Cunliffe and John T. Koch.

BIBLIOGRAPHY Bayliss, A. & M. O’Sullivan 2013 ‘Interpreting chronologies for the Mound of the Hostages, Tara, and its contemporary contexts in Neolithic and Bronze Age Ireland’, Tara from the past to the future, eds. M. O’Sullivan, C. Scarre, & M. Doyle, 26–104. Dublin, Wordwell Ltd. in association with the UCD School of Archaeology. Bendrey, R., N. Thorpe, A. Outram, & L. van Wijngaarden-Bakker 2013 ‘The Origins of Domestic Horses in North-west Europe: new Direct Dates on the Horses of Newgrange, Ireland’, Proceedings of the Prehistoric Society 79, 1–13. Bergh, S. 1995 Landscape of the monuments: a study of the passage tombs in the Cúil Irra region, Co. Sligo. Stockholm, Riksantikvarieämbetet Arkeologiska Undersökningar. Bergh, S. & R. Hensey 2013 ‘Unpicking the chronology of Carrowmore’, Oxford Journal of Archaeology 32(4), 343–66. Bermingham, N., with E. Grogan, L. Lynch, S.

Cleary CW3.indd 172

Mandal, C. Mount, J. Nolan, & H. Roche 2013 ‘Chapter 2: Prehistoric Life and Death’ River Road: The archaeology of the Limerick Southern Ring Road, N. Bermingham, F. Coyne, G. Hull, F. Reilly, & K. Taylor, 29–82. NRA Scheme Monographs 14. Dublin, National Roads Authority. Bradley, R. 1998 The Significance of Monuments: On the shaping of human experience in Neolithic and Bronze Age Europe. London, Routledge. Brady, K. 2014a ‘The logboats in the lake: Bronze Age wrecks and Viking-style battleaxes from Lough Corrib, Ireland’, Current Archaeology 292, 10–15. Brady, K. 2014b ‘Secrets of the Lake: the Lough Corrib logboats’, Archaeology Ireland 28(4), 34–38. Brady, N. 2007 ‘1462. Gormanston Beach’, The bronze Age landscape of the Pipeline to the West: An integrated archaeological and environmental assessment, E. Grogan, L. O’Donnell, & P. Johnston, 325–6. Bray, Wordwell Ltd. Brannon, N. F., B. B. Williams, & J. L. Wilkinson 1990 ‘The salvage excavation of Bronze Age Cists, Straid

27/05/2016 14:57:12

Cleary

townland, County Londonderry’, Ulster Journal of Archaeology 53, 29–39. Bray, P. 2012 ‘Before ₂₉CU became Copper: tracing the recognition and invention of metalleity in Britain and Ireland during the 3rd millennium BC’, Is there a British Chalcolithic? People, place and polity in the late 3rd millennium, eds. M. J. Allen, J. Gardiner, & A. Sheridan, 56–70. Prehistoric Society Research Paper No. 4. Oxford, Oxbow Books and The Prehistoric Society. Breen, C. & W. Forsythe 2004 Boats & Shipwrecks of Ireland. Gloucestershire, Tempus Publishing Ltd. Brett, D. 2009 A Book Around the Irish Sea: History without nations. Dublin, Wordwell Ltd. Brindley, A. L. 2004 ‘Prehistoric Pottery’, Ross Island: Mining, Metal and Society in Early Ireland, W. O’Brien, 316–38. Bronze Age Studies 6. Galway, Department of Archaeology, National University of Ireland. Brindley, A. L. 2007 The Dating of Food Vessels and Urns in Ireland. Bronze Age Studies 7. Galway, Department of Archaeology, National University of Ireland. Brindley, A. L. & J. N. Lanting 1991/92a ‘Radiocarbon dates from the Cemetery at Poulwack, Co. Clare’, Journal of Irish Archaeology 6, 13–17. Brindley, A. L. & J. N. Lanting 1991/92b ‘Radiocarbon dates from wedge tombs’, Journal of Irish Archaeology 6, 19–26. Brindley, A. L. & J. N. Lanting 1998 ‘Radiocarbon Dates for Irish Trackways’, Journal of Irish Archaeology 9, 45–67. Brindley, A. L., J. N. Lanting, & W. G. Mook 1987/88 ‘Radiocarbon Dates from Moneen and Labbacallee, County Cork’, Journal of Irish Archaeology 4, 13–20. Brindley, A. L., J. N. Lanting, & J. van der Plicht 2005 ‘Radiocarbon-dated samples from the Mound of the Hostages’, Duma na nGiall. The Mound of the Hostages, Tara, M. O’Sullivan, 281–96. Wicklow, Wordwell Ltd. in association with the UCD School of Archaeology. Bronk Ramsey, C. 2013 OxCal v. 4.2.2, r.2 (https:// c14.arch.ox.ac.uk/oxcal/OxCal.html). Radiocarbon Accelerator Unit, University of Oxford. Buckley, M. & S. W. Kansa 2011 ‘Collagen finger­ printing of archaeological bone and teeth remains from Domuztepe, South Eastern Turkey’, Archaeological and Anthropological Sciences 3(3), 271–80. Burrow, S. 2012 ‘A Date with the Chalcolithic in Wales: a review of radiocarbon measurements for 2450– 2100 cal BC’, Is there a British Chalcolithic? People, places and polity in the late 3rd millennium, eds. M. J. Allen, J. Gardiner, & A. Sheridan, 172–87. Prehistoric Society Research Paper No. 4. Oxford, Oxbow Books and

Cleary CW3.indd 173

[ 173 ]

The Prehistoric Society. Carlin, N. 2011a A proper place for everything: the character and context of beaker depositional practice in Ireland (2 vols). Unpublished PhD thesis, School of Archaeology, University College Dublin. Carlin, N. 2011b ‘Into the west: placing Beakers within their Irish contexts’, Beyond the Core: Reflections on Regionality in Prehistory, eds. A. M. Jones & G. Kirkham, 87–100. Oxford, Oxbow Books. Carlin, N. 2012 ‘Keep going, sure it’s grand: understanding the Irish Late Neolithic–Early Bronze Age’ Proceedings of the IAI Autumn 2012 Conference: The Archaeology of Disaster and Recovery, 18–27. http://iai. ie/index.php/iai-conferences/ Carlin, N. & J. Brück 2012 ‘Searching for the Chalcolithic: continuity and change in the Irish Final Neolithic/Early Bronze Age’, Is there a British Chalcolithic? People, place and polity in the late 3rd millennium, eds. M. J. Allen, J. Gardiner, & A. Sheridan, 193–210. Prehistoric Society Research Paper No. 4. Oxford, Oxbow Books and The Prehistoric Society. Case, H. 1961 ‘Irish Neolithic Pottery: Distribution and Sequence’, Proceedings of the Prehistoric Society 27, 174–233. Case, H. 1993 ‘Beakers: Deconstruction and After’, Proceedings of the Prehistoric Society 59, 241–68. Case, H. 1995 ‘Irish Beakers in their European Context’, Ireland in the Bronze Age: Proceedings of the Dublin Conference, April 1995, eds. J. Waddell & E. Shee Twohig, 14–29. Dublin, The Stationery Office. Channing, J. 1993 ‘Aughrim’, Excavations 1992, ed. I. Bennett, 4. Wicklow, Wordwell Ltd. Clarke, D.L. 1970 Beaker Pottery of Great Britain and Ireland. London, Cambridge University Press. Clutterbuck, R. 2009 Report on Archaeological Excavation of Site 25, Cookstown, County Meath. Unpublished report for Meath County Council and National Roads Authority/ Cultural Resource Development Services Ltd. Coates, J. 2005 ‘Early seafaring in northwest Europe: could planked vessels have played a significant part?’, The Mariner’s Mirror 91(4), 517–30. Coffey, G. 1889–1901 ‘On a Cairn Excavated by Thomas Plunkett, M.R.I.A., on Belmore Mountain, Co. Fermanagh’, Proceedings of the Royal Irish Academy 4C, 659–66. Collins, A. E. P. 1954 ‘The Excavation of a Double Horned Cairn at Audleystown, Co. Down’, Ulster Journal of Archaeology 17, 7–56. Collins, A. E. P. & E. E. Evans 1968 ‘A Cist Burial at Carrickinab, Co. Down’, Ulster Journal of Archaeology 31, 16–24.

27/05/2016 14:57:12

[ 174 ] v. burial practices during the late third millennium BC Conran, S. 2010 CRDS Job no. 410 Dunshaughlin to Castletown Tara Sewerage Scheme: Kilmessan 2 (05E0689), County Meath. Unpublished excavation report for Nicholas O’Dwyer Consulting Engineers. Cooney, G. & E. Grogan 1994 Irish prehistory: a social perspective. Bray, Wordwell Ltd. Coughlan, J. 2009 ‘Appendix 12: Report on Burnt Human Bone’, N6 Galway to East Ballinasloe, PPP Scheme, Archaeological Contract 3, Phase 2, Final Report E2123, Treanbaun, Co. Galway: Bronze Age Site and Early Medieval Burial and Enclosure, M. Muñiz-Pérez & F. O’Carroll, 222–33. Unpublished excavation report for Galway County Council. Cremin Madden, A. 1969 ‘The Beaker Wedge Tomb at Moytirra, Co. Sligo’, Journal of the Royal Society of Antiquaries of Ireland 99, 151–9. Cressey, M. & A. Sheridan 2003 ‘The excavation of a Bronze Age cemetery at Seafield West, near Inverness, Highland’, Proceedings of the Society of Antiquaries of Scotland 133, 47–84. Cunliffe, B. 2001 Facing the Ocean: The Atlantic and its Peoples 8000 BC–AD 1500. Oxford, Oxford University Press. Curtis, N. G. W. & N. C. A. Wilkin 2012 ‘The regionality of Beakers and Bodies in the Chalcolithic of North-East Scotland’, Is there a British Chalcolithic? People, places and polity in the late 3rd millennium, eds. M. J. Allen, J. Gardiner, & A. Sheridan, 237–56. Prehistoric Society Research Paper No. 4. Oxford, Oxbow Books and The Prehistoric Society. Davies, O. 1937 ‘Excavation at Ballyrenan, Co. Tyrone’, Journal of the Royal Society of Antiquaries of Ireland 7, 89–100. Davies, O. 1939a ‘Excavation of a Horned Cairn at Aghanaglack, Co. Fermanagh’, Journal of the Royal Society of Antiquaries of Ireland 9, 21–38. Davies, O. 1939b ‘Excavations at the Giant’s Grave, Loughash’, Ulster Journal of Archaeology 2, 254–68. Davies, O. & J. B. Mullin 1940 ‘Excavation of Cashelbane cairn, Loughash, Co. Tyrone’, Journal of the Royal Society of Antiquaries of Ireland 70, 143–63. Delaney, S. 2012 ‘Change and adaption in the Barnesmore Gap, Co. Donegal’, Proceedings of the IAI Autumn 2012 Conference: The Archaeology of Disaster and Recovery, Conference Proceedings Issue 1, 12–7. http:// iai.ie/index.php/iai-conferences/ de Valera, R. & S. Ó Nualláin 1961 Survey of the Megalithic Tombs of Ireland, Vol. I, Co. Clare. Dublin, The Stationery Office. Dowd, M. 2015 The Archaeology of Caves in Ireland. Oxford, Oxbow Books. Eogan, G. 1984 Excavations at Knowth 1: Smaller Passage Tombs, Neolithic Occupation and Beaker Activity. Dublin,

Cleary CW3.indd 174

Royal Irish Academy. Eogan, G., and H. Roche 1997 Excavations at Knowth 2: Settlement and Ritual Sites of the Fourth and Third Millennia BC. Dublin, Royal Irish Academy. Eogan, J. 2002 ‘Excavations at a cairn at Coolnatullagh townland, County Clare’, North Munster Antiquarian Journal 42, 113–50. Evans, E. E. 1939 ‘Excavations at Carnanbane, County Londonderry: A Double Horned Cairn’, Proceedings of the Royal Irish Academy 45C, 1–12. Evans, E. E. 1953 Lyles Hill: A Late Neolithic site in County Antrim. Archaeological Research Publications (Northern Ireland) No. 2. Belfast, Her Majesty’s Stationery Office. Freeman, P. 2000. Ireland and the Classical World. Texas, University of Texas Press. Fry, M. 2006 ‘A Creature of the Deep: The Dugout Boat in sea-going mode. Evidence from the shores of the North of Ireland’, The Modern Traveller to our Past. Festschrift in honour of Ann Hamlin, ed. M. Meek, 370–76. Belfast, DPK Publishers. Gibson, A. 2004 ‘Burials and Beakers: Seeing beneath the veneer in Late Neolithic Britain’, Similar but Different: Bell Beakers in Europe, ed. J. Czebreszuk, 173–92. Poznań, Adam Mickiewicz University. Gibson, C. 2013 ‘Beakers into Bronze: Tracing connections between western Iberia and the British Isles 2800–800 BC’, Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European into Atlantic Europe, eds. J. T. Koch & B. Cunliffe, 71–99. Oxford, Oxbow Books. Gillespie, R. 2012 Claremorris Sewerage Scheme: Excavation at Coolroe, County Mayo. Unpublished excavation report for Mayo County Council. Glover, W. 1975 ‘Segmented Cist Grave in Kinkit Townland, County Tyrone’, Journal of the Royal Society of Antiquaries of Ireland 105, 150–55. Grant, C. 2009 ‘Early Bronze Age date for Burren wedge tomb’, Archaeology Ireland 23(4), 5. Green, H.S. 1980 The Flint Arrowheads of the British Isles. BAR British Series 75. Oxford, British Archaeological Reports. Gregory, N. T. N. 1997 A comparative study of Irish and Scottish logboats. Unpublished PhD thesis, University of Edinburgh. Grimes, W.F. 1964 ‘Excavations in the Lake Group of Barrows, Wilsford, Wiltshire, in 1959’, University of London Bulletin of the Institute of Archaeology 4, 89–121. Grogan, E. & H. Roche 2010 ‘Clay and fire: the develop­ment and distribution of pottery traditions in prehistoric Ireland’, Creative Minds: production, manufacturing and invention in ancient Ireland, eds. M. Stanley, E. Danaher, & J. Eogan, 27–44. NRA

27/05/2016 14:57:12

Cleary

Monograph Series No. 7. Dublin, National Roads Authority. Hackett, L. 2010 Final Report on archaeological investigations at Site E2980, in the townland of Moone, Co. Kildare. Unpublished report for Kildare County Council and National Roads Authority/ Headland Archaeology Ltd. Harbison, P. 1976 Bracers and V-perforated Buttons in the Beaker and Food Vessel Cultures of Ireland. Archaeologica Atlantica Research Report 1. Bad Bramstedt, Moreland. Hartnett, P.J. 1971 ‘The Excavation of Two Tumuli at Fourknocks (Sites II and III), Co. Meath’, Proceedings of the Royal Irish Academy 71C, 35–89. Hartnett, P. J. & E. Prendergast 2011 ‘Ballybrennan, Co. Westmeath, E1162’, Breaking Ground, Finding Graves— reports on the excavations of burials by the National Museum of Ireland, 1927–2006 Volume 1, eds. M. Cahill & M. Sikora, 484–501. Dublin, Wordwell Ltd. in association with the National Museum of Ireland. Hencken, H. O’N. & H. L. Movius 1935 ‘A Cairn at Poulawack, County Clare: With a Report on the Human Remains’, Journal of the Royal Society of Antiquaries of Ireland 5, 191–222. Hensey, R., P. Meehan, M. Dowd, & S. Moore 2013 ‘A century of archaeology—historical excavation and modern research at the Carrowkeel passage tombs, County Sligo’, Proceedings of the Royal Irish Academy 114C, 1–31. Herity, M. 1974 Irish Passage Graves: Neolithic TombBuilders in Ireland and Britain 2500 B.C. Dublin, Irish University Press. Herity, M. 1987 ‘The Finds from Irish Court Tombs’, Proceedings of the Royal Irish Academy 87C, 103–281. Herring, I. 1938 ‘The cairn excavation at Well Glass Spring, Largantea, Co. Londonderry’, Ulster Journal of Archaeology 1, 164–88. Herring, I. & A. May 1937/38 ‘The Giants Grave, Kilhoyle, Co. Londonderry’, Proceedings of the Belfast Natural History and Philosophical Society 1(3), 34–48. Holst, M. K. 2013 ‘Burials’, The Oxford Handbook of The European Bronze Age, eds. H. Fokkens & A. Harding, 102–20. Oxford, Oxford University Press. Hüls, C.M., H. Erlenkeuser, M-J. Nadeau, P.M. Grootes & N. Andersen 2010 ‘Experimental study on the ori­gin of cremated bone apatite carbon’, Radiocarbon 52(2), 587–99. Hurl, D. & E. Murphy 2001 ‘The excavation of a wedge tomb at Ballybriest, County Londonderry’, Ulster Journal of Archaeology 60, 9–31. Ixer, R. A. 2004 ‘Appendix 7: The petrography of prehistoric pottery from Ross Island mine’, Ross Island: Mining, Metal and Society in Early Ireland, W.

Cleary CW3.indd 175

[ 175 ]

O’Brien, 643–9. Bronze Age Studies 6. Galway, Department of Archaeology, National University of Ireland. Jones, C., P. Walsh, & P. Ó Cearbhaill 1996 ‘Recent Discoveries on Roughaun Hill, County Clare’, The Journal of the Royal Society of Antiquaries of Ireland 126, 86–107. Jones, C., T. McVeigh, & R. Ó Maoldúin 2015 ‘Monuments, landscape and identity in Chalcolithic Ireland’, Landscape and Identity: Archaeology and Human Geography, ed. K. D. Springs, 3–25. BAR International Series 2709. Oxford, Archaeopress. Kaagan, L.M. 2000 The Horse in Late Pleistocene and Holocene Britain. Unpublished PhD thesis, Department of Biology, University College London. King, H.A. 1999 ‘Excavation on the Fourknocks Ridge, Co. Meath’, Proceedings of the Royal Irish Academy 99C, 157–198. Kytmannow, T. 2008 Portal Tombs in the Landscape: The chronology, morphology and landscape setting of the portal tombs of Ireland, Wales and Cornwall. BAR British Series 455. Oxford, Archaeopress. Lanting, J. N. & A. L. Brindley 1996 ‘Irish logboats and their European context’, Journal of Irish Archaeology 7, 85–95. Lanting, J. N. & J. van der Plicht 1998 ‘Reservoir effects and apparent 14C-ages’, Journal of Irish Archaeology 9, 151–65. Leary, J. 2014 ‘Past Mobility: An Introduction’, Past Mobilities: Archaeological Approaches to Movement and Mobility, ed. J. Leary, 1–19. Surrey, Ashgate Publishing Ltd. Leask, H. G. & L. Price 1936 ‘The Labbacallee megalith, Co. Cork’, Proceedings of the Royal Irish Academy 43C, 77–101. Lehane, J., P. Johnston, & D. Leigh 2011 Archaeological Excavation Report E2412 - Ballynacarriga 3, Co. Cork: Prehistoric site with enclosure, structures, two ring ditches and associated cist burials. Unpublished report for Cork County Council and National Roads Authority/Eachtra Archaeological Projects. Lucas, A. T. 1972 ‘Prehistoric Block-Wheels from Doogarymore, Co. Roscommon, and Timahoe East, Co. Kildare’, Journal of the Royal Society of Antiquaries of Ireland 102, 19–48. Lynch, A. 2014 Poulnabrone: An early Neolithic portal tomb in Ireland. Archaeological Monograph Series: 9. Dublin, The Stationery Office. Macalister, R. A. S., E. C. R. Armstrong, & R. Lloyd Praeger 1913 ‘On a Bronze-Age Interment with Associated Standing-Stone and Earthen Ring, near Naas, Co. Kildare’, Proceedings of the Royal Irish Academy 30C, 351–60.

27/05/2016 14:57:13

[ 176 ] v. burial practices during the late third millennium BC Mallory, J. P. 2013 The Origins of the Irish. London, Thames & Hudson Ltd. May, A. McL. 1947 ‘Burial mound, circles and cairn, Gortcorbies, Co. Londonderry’, Journal of the Royal Society of Antiquaries of Ireland 77, 5–22. McCormick, F. 1987 ‘Appendix 1: The Animal Bones’, Archaeological Excavations on the Cork–Dublin Gas Pipeline (1981–82), eds. R. M. Cleary, M. F. Hurley, & E. A. Twohig, 26–29. Cork, Department of Archaeology, University College Cork. McCormick, F. 2007 ‘The horse in early Ireland’, Anthropozoologica 42, 85–104. McGrail, S. 1995 ‘Celtic Seafaring and Transport’, The Celtic World, ed. M. J. Green, 254–81. London, Routledge. McQuade, M. 2005 ‘Archaeological Excavation of a Multi-Period Prehistoric Settlement at Waterunder, Mell, Co. Louth’, County Louth Archaeological and Historical Journal 26, 31–66. Moore, C. & C. Chiriotti 2010 ‘Reinventing the wheel: new evidence from Edercloon, Co. Longford’, Creative Minds: production, manufacturing and invention in ancient Ireland, eds. M. Stanley, E. Danaher, & J. Eogan, 57–68. NRA Monograph Series No. 7. Dublin, National Roads Authority. Mount, C. 2012 ‘A note on some Beaker period pit burials in Ireland’, Journal of Irish Archaeology 21, 1–6. Mount, C. 2013 ‘The context of the Early Bronze Age cemetery in the Mound of the Hostages, Tara’, Tara—from the past to the future. Towards a new research agenda, eds. M. O’Sullivan, C. Scarre, & M. Doyle, 184–95. Dublin, Wordwell Ltd. in association with the UCD School of Archaeology. Muñiz-Pérez, M. 2014 ‘Prehistoric lead mine and cists at Treanbaun’, The Quiet Landscape: Archaeological investigations on the M6 Galway to Ballinasloe national road scheme, eds. J. McKeon & J. O’Sullivan, 130–4. NRA Scheme Monographs 15. Dublin, The National Roads Authority. Muñiz-Pérez, M. & N. Bermingham. 2014. ‘Cremation cemetery and palisaded enclosure at Treanbaun’, The Quiet Landscape: Archaeological investigations on the M6 Galway to Ballinasloe national road scheme, eds. J. McKeon & J. O’Sullivan, 122–9. NRA Scheme Monographs 15. Dublin, The National Roads Authority. Needham, S. 2004 ‘Migdale-Marnoch: sunburnt of Scottish metallurgy’, Scotland in Ancient Europe: the Neolithic and Early Bronze Age of Scotland in their European context, eds. I. A. G. Shepherd & G. J. Barclay, 217–45. Edinburgh, Society of Antiquaries of Scotland. Needham, S. 2005 ‘Transforming Beaker culture in north-west Europe; processes of fusion and fission’,

Cleary CW3.indd 176

Proceedings of the Prehistoric Society 71, 171–217. Needham, S. & A. Woodward. 2008 ‘The Clandon Barrow finery: a synopsis of success in an Early Bronze Age world’, Proceedings of the Prehistoric Society 74, 1–52. Ó Baoill, R. 2005 ‘The Excavation of a segmented Bronze Age cist at Newtownstewart Castle, County Tyrone’, Ulster Journal of Archaeology 64, 26–42. O’Brien, W. 1999 Sacred Ground: Megalithic Tombs in Coastal South-West Ireland. Bronze Age Studies 4. Galway, Department of Archaeology, National University of Ireland. O’Brien, W. 2004 Ross Island: Mining, Metal and Society in Early Ireland. Bronze Age Studies 6. Galway, Department of Archaeology, National University of Ireland. O’Brien, W. 2012a Iverni: A Prehistory of Cork. Cork, The Collins Press. O’Brien, W. 2012b ‘The Chalcolithic in Ireland: a chronological and cultural framework’, Is there a British Chalcolithic? People, places and polity in the late 3rd millennium, eds. M. J. Allen, J. Gardiner, & A. Sheridan, 211–25. Prehistoric Society Research Paper No. 4. Oxford, Oxbow Books and The Prehistoric Society. O’Connell, A. 2013 Harvesting the Stars: A pagan temple at Lismullin, Co. Meath. NRA Scheme Monographs 11. Dublin, National Roads Authority. O’Connor, D.J. 2008 N2 Finglas-Ashbourne Road Scheme: Report on Archaeological Excavation of Site 19, Harlockstown, County Meath. Unpublished excavation report for Meath County Council. O’Connor, R., J. A. Crutchfield, B. J. Whelan, & K. E. Mellon 1980 Development of the Irish Sea Fishing Industry and its Regional Implications. Dublin, The Economic and Social Research Institute. Ó Drisceoil, D.A. 1988 ‘Burnt mounds: cooking or bathing?’, Antiquity 62, 671–80. O’Kelly, M. J. 1952 ‘Excavation of a Cairn at Moneen, Co. Cork’, Proceedings of the Royal Irish Academy 54C, 121–59. O’Kelly, M. J. 1960 ‘A wedge-shaped gallery grave at Baurnadomeeny, Co. Tipperary’, Journal of the Cork Historical and Archaeological Society 65, 85–115. Ó Maoldúin, R. 2014a Exchange in Chalcolithic and Early Bronze Age (EBA) Ireland: Connecting people, objects and ideas, Unpublished PhD thesis, Department of Archaeology, School of Geography and Archaeology, National University of Ireland, Galway. Ó Maoldúin, R. 2014b ‘The people behind the pots: considering the Early Bronze Age remains from French Furze, Tully East, Co. Kildare’, Fragments of lives past: archaeological objects from Irish road schemes, eds. B. Kelly, N. Roycroft & M. Stanley, 27–38.

27/05/2016 14:57:13

Cleary

Dublin, National Roads Authority. O’Regan, C., S. Jones & C. Randall 2009 ‘A unique burial monument at Brackagh, County Londonderry’, Ulster Journal of Archaeology 68, 27–39. Ó Ríordáin, B. & J. Waddell 1993 The Funerary Bowls and Vases of the Irish Bronze Age. Galway, Galway University Press for the National Museum of Ireland. Ó Ríordáin, S. P. 1933 ‘Discovery of an ancient burial in a cairn near Ballyconnell, Co. Cavan’, Journal of the Royal Society of Antiquaries of Ireland 3(2), 167–71. Ó Ríordáin, S. P. & R. de Valera 1952 ‘Excavation of a megalithic tomb at Ballyedmonduff, Co. Dublin’, Proceedings of the Royal Irish Academy 55C, 61–81. Ó Ríordáin, S. P. & G. Ó h-Iceadha 1955 ‘Lough Gur Excavations: The Megalithic Tomb’, Journal of the Royal Society of Antiquaries of Ireland 85, 34–50. O’Sullivan, M. 2005 Duma na nGiall. The Mound of the Hostages, Tara. Wicklow, Wordwell Ltd. in association with the UCD School of Archaeology. Petersen, F. 1972 ‘Traditions of multiple burial in Later Neolithic and Early Bronze Age England’, The Archaeology Journal 129, 22–55. Purcell, A. 2001 Excavation and Specialist Reports, Corbally, Kilcullen, Co. Kildare, Licence Nos 97E0449 & 98E0094. Unpublished report for Kilsaran Concrete/Margaret Gowen & Co. Ltd. Quinn, C. P. & I. Kuijt 2013 ‘The tempo of life and death during the Early Bronze Age at the Mound of the Hostages, Tara’, Tara—from the past to the future. Towards a new research agenda, eds. M. O’Sullivan, C. Scarre, & M. Doyle, 196–206. Dublin, Wordwell Ltd. in association with the UCD School of Archaeology. Raftery, B. 1996 Trackway excavations in the Mountdillon Bogs, Co. Longford, 1985–1991. Irish Archaeological Wetland Unit Transactions 3. Dublin, Crannog Publications. Reimer, P. J., E. Bard, A. Bayliss, J. W. Beck, P. G. Blackwell, C. Bronk Ramsey, P. M. Grootes, T. P. Guilderson, H. Haflidason, I. Hajdas, C. Hattž, T. J. Heaton, D. L. Hoffmann, A. G. Hogg, K. A. Hughen, K. F. Kaiser, B. Kromer, S. W. Manning, M. Niu, R. W. Reimer, D. A. Richards, E. M. Scott, J. R. Southon, R. A. Staff, C. S. M. Turney & J. van der Plicht 2013 ‘IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0-50,000 Years cal BP’, Radiocarbon 55(4), 1869–87. Rice, K. 2006. The prehistory of Piperstown. A reassessment of an upland landscape. Unpublished MA thesis, School of Archaeology,

Cleary CW3.indd 177

[ 177 ]

University College Dublin. Robinson, M., D. Shimwell, & G. Cribbin 1996 ‘Boating in the Bronze Age: Two Logboats from Co. Mayo’, Archaeology Ireland 10(1), 12–13. Robinson, M. E., D. W. Shimwell, & G. Cribbin 1999 ‘Re-assessing the logboat from Lurgan Townland, Co. Galway, Ireland’, Antiquity 73, 903–8. Rogers, A. 2013 Female Burial Traditions of the Chalcolithic and Early Bronze Age: a pilot study based on modern excavations. BAR British Series 581. Oxford, Archaeopress. Rynne, E. & P. Ó hÉailidhe 1965–6 ‘A group of prehistoric sites at Piperstown, Co. Dublin’, Proceedings of the Royal Irish Academy 64C, 61–84. Salanova, L. 2003 ‘Heads North: Analysis of Bell Beaker Graves in Western Europe’, Journal of Iberian Archaeology 5, 163–9. Salanova, L. 2004 ‘The Frontiers Inside the Western Bell Beaker Block’, Similar but Different: Bell Beakers in Europe, ed. J. Czebreszuk, 63–75. Poznań, Adam Mickiewicz University. Schuhmacher, T. X., A. Banerjee, W. Dindorf, C. Sastri, & T. Sauvage 2013 ‘The use of sperm whale ivory in Chalcolithic Portugal’, Trabajos de Prehistoria 70(1), 185–203. Schulting, R., C. Bronk Ramsey, P. Reimer, G. Eogan, K. Cleary, G. Cooney, & A. Sheridan forthcoming ‘Dating Neolithic human remains at Knowth’, Excavations at Knowth 6: The Great Mound at Knowth (Tomb 1) and its passage tomb archaeology, G. Eogan & K. Cleary. Dublin, Royal Irish Academy. Schulting, R., E. Murphy, C. Jones, & G. Warren 2012 ‘New dates from the north and a proposed chronology for Irish court tombs’, Proceedings of the Royal Irish Academy 112C, 1–60. Schulting, R., A. Sheridan, S. Clarke, & C. Bronk Ramsey 2008 ‘Largantea and the Dating of Irish Wedge Tombs’, Journal of Irish Archaeology 17, 1–17. Shepherd, A.N. 1989 ‘A note on the orientation of beaker burials in north-east Scotland’ in Greig, M.K., C. Greig, A.N. Shepherd & I.A.G. Shepherd, ‘A beaker cist from Chapelden, Tore of Troup, Aberdour, Banff and Buchan District, with a note on the orientation of beaker burials in north-east Scotland’, Proceedings of the Society of Antiquaries of Scotland 119, 73–81. Shepherd, A. 2012 ‘Stepping Out Together: men, women, and their Beakers in time and space’, Is there a British Chalcolithic? People, place and polity in the late 3rd millennium, eds. M. J. Allen, J. Gardiner, & A. Sheridan, 257–80. Prehistoric Society Research Paper No. 4. Oxford, Oxbow Books and The Prehistoric Society.

27/05/2016 14:57:13

[ 178 ] v. burial practices during the late third millennium BC Sheridan, A. 1983 ‘A Reconsideration of the Origins of Irish Metallurgy’, Journal of Irish Archaeology 1, 11–19. Sheridan, A. 2007 ‘The beads from the Bronze Age graves at Caltragh’, Monumental Beginnings: The archaeology of the N4 Sligo Inner Relief Road, E. Danaher, CD-ROM. NRA Scheme Monographs 1. Dublin, National Roads Authority. Sheridan, A., M. Davis, I. Clark, & H. Redvers-Jones 2002 ‘Investigating jet and jet-like artefacts from prehistoric Scotland: the National Museums of Scotland project’, Antiquity 76, 812–25. Sheridan, A. & P. Northover 1993 ‘A Beaker Period Copper Dagger Blade from the Sillees River near Ross Lough, Co. Fermanagh’, Ulster Journal of Archaeology 56, 61–9. Simpson, D.D.A. 1996 ‘Irish perforated stone imple­ments in context’, Journal of Irish Archaeology 7, 65–76. Smyth, J. 2014 Settlement in the Irish Neolithic: New discoveries at the edge of Europe. Prehistoric Society Research Paper 6 in association with the Heritage Council, Ireland. Oxford, Oxbow Books. Snoeck, C., F. Brock & R.J. Schulting 2014 ‘Carbon exchanges between bone apatite and fuels during cremation: impact on radiocarbon dates’, Radiocarbon 56(2), 591–602. Standish, C.D., B. Dhuime, C.J. Hawkesworth & A.W.G. Pike 2015 ‘A Non-local Source of Irish Chalcolithic and Early Bronze Age Gold’, Proceedings of the Prehistoric Society 81, 1–29. Stanley, M., C. McDermott, C. Moore, & C. Murray 2003 ‘Throwing off the Yoke’, Archaeology Ireland 17(2), 6–8. Strid, L. 2012 Appendix 2.6: Faunal Assemblage Report, 00E0811: Sheephouse 3 Final Report, D. Nelis, cix–cxxi. Unpublished excavation report for Meath County Council/Irish Archaeological Consultancy Ltd. Taylor, K. 2008 ‘At home and on the road: two Iron Age sites in County Tipperary’, Seanda 3, 54–5. Thomas, N. 1968 ‘Appendix 1: Note on the Carrickinab awl’, A Cist Burial at Carrickinab, Co. Down, A. E. P. Collins & E. E. Evans, 23–4. Ulster Journal of Archaeology 31, 16–24. Thomas, N., with E.C. Ellwood 2005 ‘Early Bronze Age copper-alloy awls from Sites I and II, with metal analysis and classification, Snail Down, Wiltshire: The Bronze Age Barrow Cemetery and Related Earthworks, in the parishes of Collingbourne Ducis and Collingbourne Kingston Excavations, 1953, 1955 and 1957, N. Thomas, 219–22. Devizes, Wiltshire Archaeological and Natural History Society. Topp, C. 1962 ‘The portal dolmen of Dromanone –

Cleary CW3.indd 178

Co. Roscommon’, Bulletin of the Institute of Archaeology 3, 38–46. Tuckwell, A.N. 1975 ‘Patterns of burial orientation in the round barrows of East Yorkshire’, Bulletin of the Institute of Archaeology, University of London 12, 95–123. Van de Noort, R. 2011 North Sea Archaeologies: A Maritime Biography, 10,000 BC – AD 1500. Oxford, Oxford University Press. van Wijngaarden-Bakker, L. H. 1974 ‘The animal remains from the Beaker settlement at Newgrange, Co. Meath–first report’, Proceedings of the Royal Irish Academy 74C, 313–85. van Wijngaarden-Bakker, L. H. 1986 ‘The animal remains from the Beaker settlement at Newgrange, Co. Meath: final report’, Proceedings of the Royal Irish Academy 86C, 17–111. Waddell, J. 1990 The Bronze Age Burials of Ireland. Galway, Galway University Press. Waddell, J. 1991/92 ‘The Irish Sea in Prehistory’, The Journal of Irish Archaeology 6, 29–40. Walsh, P. 1995 ‘Structure and Deposition in Irish Wedge Tombs: An Open and Shut Case?’, Ireland in the Bronze Age: Proceedings of the Dublin Conference, April 1995, eds. J. Waddell & E. Shee Twohig, 113–27. Dublin, The Stationery Office. Wilkin, N. 2010 ‘Soul of the age? New light on Roseisle’s ‘big man’ and the Early Bronze Age funerary practices of the Moray Firth region’, Beakers, Bones & Birnie – Archaeological works in progress associated with Elgin Museum, Moray Society Annual Conference 25th-26th April, 2009, eds. J. Trythall & B. Dalgarno, 45–66. Elgin, Elgin Museum. Woodcock, J. 2004 ‘The early Bronze Age on the Isle of Man: back into the mainstream?’, The Neolithic of the Irish Sea: Materiality and traditions of practice, eds. V. Cummings & C. Fowler, 214–23. Oxford, Oxbow Books. Woodcock, J. 2008 The Bronze Age Pottery of the Isle of Man: Evidence for cultural movement around the Irish Sea basin. BAR British Series 475. Oxford, Archaeopress. Woodman, P. C. 1997 ‘Killuragh Cave’, Excavations 1996, ed. I. Bennett, 67–8. Wicklow, Wordwell Ltd. Wood-Martin, W. G. 1883 ‘The Battle-ground and Ancient Monuments of Northern Moytirra’, Journal of the Royal Historical and Archaeological Association of Ireland 6(60), 442–70. Woodward, A. 2000 ‘Beads and Beakers: heirlooms and relics in the British Early Bronze Age’, Antiquity 76 (294), 1040–47. Woodward, A. & J. Hunter 2015 Ritual in Early Bronze Age Grave Goods. Oxford, Oxbow Books.

27/05/2016 14:57:13

chapter six

Stelae, funerary practice, and group identities in the Bronze and Iron Ages of SW Iberia: a moyenne durée perspective

Dirk Brandherm Introduction

A

working in south-western Iberia have long toiled hard to find material evidence to substantiate Herodotus’ account of the semi-mythical polity of Tartessos1, or even to support its identification with the biblical Taršiš2 ever since the early years of the 20th century, when Schulten first attempted to locate its archaeological remains in the wetlands of the Coto de Doñana.3   Despite Schulten’s lack of success in this attempt, which saw him conducting extensive fieldwork at the Cerro del Trigo (Almonte, Huelva), most scholars today would agree that Herodotus’ Tartessos is to be sought in the area situated between the lower reaches of the Guadalquivir and the rivers Tinto and Odiel (Figure 6.1). While dissenting voices have favoured a location in the Spanish Levant4 or Sardinia5, these alternative proposals tend to suffer from a limited or biased reading of the available textual sources.6 Where they draw mainly on texts mentioning the name ‘Taršiš’ rather than on the Greek ‘Tartessos’ tradition, they also implicitly assume that both labels do not apply to the same geographical region, as none of the relevant alternative proposals is easily reconciled with Herodotus’ reference to rchaeologiists

1 Herodotus I.163. For references to Tartessos by earlier Greek authors, only surviving in secondary sources, and for the phonetic reconstruction of the original indigenous toponym see Sieglin (1934), with additional comments by M. Koch (1984, 111f) and Hoz (2010, 217–22). As the semantic distinction between ‘Tartesos’—almost certainly closer to the likely self-designation of the indigenous trt/ trs-population—and the Hellenized form ‘Tartessos’ is of limited relevance to the issues discussed in the present contribution, this text will employ the form ‘Tartes(s)os’ unless explicit reference is made to a specific author’s usage. 2 For a detailed discussion of the relevant sources see M. Koch (1984, ix); supplementary comments in Hoz (2010, 242–4). 3 Schulten 1922, 81–90; 1950, 155–76. 4 Sureda 1979, 20–25. 5 Cross 1972, 16; Thompson & Skaggs 2013, 3.2. 6 cf. M. Koch 2005, 28–33; Rollinger 2008, 690.

Brandherm CW3 CS4 19 i 2016.indd 179

27/05/2016 15:42:17

6.1 Map of southern Iberia with toponyms mentioned in the text, other than sites mapped in Figures 6.3, 6.9, 6.12, 6.16, 6.17, & 6.18

ethnic groups bearing Celtic names in the immediate vicinity of Tartessos.7 In the present pages we shall follow M. Koch’s reasoning, who so far has presented the most compelling case for a localization of Taršiš/Tartes(s)os in south-western Iberia.8   This is not to say that these two terms should be used interchangeably. The texts employing one or the other differ widely in date and refer to very different historical circumstances. While the original biblical Taršiš references reflect an early stage of regular Levantine contacts with the Far West, Herodotus refers to a completely different situation some three to five centuries later9, following not only the formation of a western Phoenician koiné and its socio-economic impact on indigenous communities, but also the profound crisis this new system suffered over the course of the 6th century BC and its subsequent reconfiguration.10   The above discrepancy is also borne out in the long-running debate between two different schools of thought in Iberian Iron Age archaeology. A majority of scholars who have taken sides in this debate would limit the use of the ‘Tartes(s)os’ label to the material remains of the Orientalizing period. In their perspective, the formation of ‘Tartes(s)os’ follows from 7 Arguing along different, but hardly more convincing lines, a localization of the original biblical Taršiš in the same geographical region as Herodotus’ Tartessos has also been vehemently denied by Aubet (1994, 180–3). 8 M. Koch 1984, 139–43; 2005, 38–42. 9 While Herodotus’ (IV.152) account seems to indicate a date for Kolaios’ s Tartes(s)ian venture between 640–630 BC, it is not at all clear if his ethnographic map of south-western Iberia is meant to depict the state of affairs at that time, if it reflects a later situation, or if indeed it captures chronologically heterogeneous information, projected into the writer’s ethnographic present. 10 Fernández 2009, 152–6; M. Koch 2010, 569–71; Ordóñez 2011, 343–8.

Brandherm CW3 CS4 19 i 2016.indd 180

27/05/2016 15:42:23

Brandherm

[ 181 ]

a process of intense acculturation, triggered by the exposure of indigenous communities centred on the lower Guadalquivir basin to imports of Eastern Mediterranean goods and ideas that reached the area after the Phoenicians first established a permanent presence on the shores of southern Iberia.11   The other school of thought to which a significant number of scholars have rallied when it comes to applying the ‘Tartes(s)os’ nomenclature to the archaeological record is based on the likely indigenous root of the term. This would advocate a broader use of this label, to refer to the material culture of the native population inhabiting south-western Iberia during the Early Iron Age at large, and by extension also during the Final Bronze Age (Figure 6.2).12 This is despite the fact that we may at best have an educated guess at what this term might have referred to in an indigenous context, and no real criteria to assess its time depth prior to the biblical ‘Taršiš’ references of the 10th century BC, whose relationship to the much later ‘Tartes(s)os’ terminology itself remains an object of debate.13

6.2 Periodization of the Bronze and Early Iron Ages in SW Iberia (after Parreira 1995, 132; modified after Roberts et al. 2013, fig. 2.5)

  As a consequence, there also is no agreement among scholars about the geographic extent to which this label may be meaningfully applied to the indigenous population inhabiting the hinterland of Andalusia’s Atlantic coastline during the Final Bronze and Early Iron Ages.14 Already in Antiquity, some considerable variation in the geographic scope of its application can 11 Abad Casal 1979, 176f n 18; Celestino 2008, 97f. 12 Albuquerque 2013a, 637. For an outline of the chronological framework and terminology employed in the present contribution, see Roberts et al. 2013, 33–7. In contrast to the usual convention in Englishlanguage literature dealing with the Bronze Age in Atlantic Europe, the final centuries of the 2nd millennium and the beginning of the 1st millennium BC here are referred to as ‘Final Bronze Age’ rather than ‘Late Bronze Age’, to reflect the terminological distinction Spanish—but generally not Portuguese—archaeology makes between ‘Bronce Tardío’ and ‘Bronce Final’ (cf. Molina 1978; Parreira 1995). 13 M. Koch 2005, 23f; Álvarez 2013, 228–40. 14 Rodríguez & Enríquez 2001, 35–45; Torres 2002, 11–15.

Brandherm CW3 CS4 19 i 2016.indd 181

27/05/2016 15:42:28

Legend   Agglomerated cist cemetery   Megalithic monument   Cave   Cist   Flat grave   Hypogaeum   Silo burial

6.3 Distribution of Middle Bronze Age funerary sites in SW Iberia (data from AEMA database)

be observed. Even during the final centuries of the 1st millennium BC the ‘Tartes(s)os’ concept was extended from an original core area in the modern-day provinces of Huelva and Seville (‘tartesius ager’) to include all of Andalusia as well as parts of Murcia (‘terminus tartesiorum’).15   Despite this later extension, there is nothing in the textual sources to justify the notion of Tartes(s)os as a polity whose territory would have encompassed not only parts of Andalusia, but also large swaths of Spain’s neighbouring Extremadura. This notion is entirely a legacy of Schulten’s16 imaginative writing which has influenced several generations of scholars17, and the same may be said of attempts to extend the chronological range of the ‘Tartes(s)os’ label and apply it to the archaeological record of the later 2nd millennium BC.18 It is precisely this legacy, with the suggestive power of its ‘Tartes(s)os’ concept—aided by a dearth of detail in the original textual sources—and the resulting tendency to try and expand the geographic and chronological scope of what may be captured under this term, which in many instances can be held responsible for hampering unbiased analysis of the archaeological record and hindering objective interpretation.19   This problem not only affects the disciplines of ancient history and archaeology. As Schulten also labelled the writing system found on Early Iron Age stelae in southern Portugal ‘Tartessian’, historical linguistics has been drawn into the same quagmire.20 The issue has 15 M. Koch 1984, 109–26; Torres 2002, 12f. 16 Schulten 1950, 110–32. 17 e.g. Pereira 1984; Maluquer 1985, 66–71; on the wider intellectual background cf. Álvarez 2005, 119–61; Ferrer & Prados 2013, 396–403. 18 e.g. Pellicer 1982, 42; 1989, 206; Bendala 2013, 123–9. 19 cf. Álvarez 2005, 185–204; Brandherm 2008, 484; Ferrer & Prados 2013, 403–6. 20 Schulten (1950, 148) justifies this choice of label with a reference to Strabo, according to whom the

Brandherm CW3 CS4 19 i 2016.indd 182

10/06/2016 17:57:25

Brandherm

[ 183 ]

recently gained additional relevance with the case put forward by J.  Koch for identifying the language of these inscriptions as Celtic, thus adding ‘Tartessian’ to the family tree of Indo-European languages.21 The choice of this particular label would hardly merit further discussion, were it not for the implicit connotations that come with it. As it is, the question needs to be asked to what extent the archaeological record actually supports this labelling, not only of the stelae themselves and the writing system, but also of the language of these inscriptions as ‘Tartes(s)ian’.   Besides other considerations, the present contribution will attempt to answer this question. In order to do so, it will explore the expression of group identities in the Bronze Age and Early Iron Age of south-western Iberia. Because of the apparent funerary character of the stelae and their inscriptions, the main focus of the analysis is on the funerary evidence. The following section will outline and compare the various different traditions in burial customs and funerary architecture found in south-western Iberia, both in terms of their temporal and spatial dimensions. The subsequent section will discuss the use of stelae and inscribed slabs as funerary monuments in the same geographical area. Where necessary, the discussion will also draw on information from neighbouring regions. Despite the obvious limitations of such an approach when it comes to answering the question of prehistoric group identities, we are confident that it will open up new perspectives on this old problem. Funerary architecture and burial rites The funerary record of the Middle Bronze Age in south-western Iberia is characterized by inhumation practices and the widespread occurrence of cist burials. In many parts of southern Portugal, except for coastal areas and inland stretches of the southern Baixo Alentejo and Algarve, where the general pattern of small cist cemeteries prevails, the latter are complemented—and sometimes substituted—by hypogaea, rock-cut tombs, and silo burials (Figure 6.3).22 A limited number of inhumation burials attributable to the Middle Bronze Age have also been recorded from earlier megalithic tombs.23   Apart from the fairly ubiquitous occurrence of simple cists arranged in loose scatters, in some areas of the Baixo Alentejo and Algarve we also find cists and simple pits set into either Tartes(s)ians would have used a script of their own. This claim by Schulten is most misleading. The relevant passage in Strabo (139, 21–4) explicitly refers to the Turdetanians, which Schulten equates with the Tartes(s)ians without further comment. None of the original sources supports such a simple equation, and Strabo (139, 17–21) himself clearly preserves the memory of an older, more diverse ethnographic map of the Iberian south-west (cf. Cruz Andreotti 2010, 20–6). 21 J. Koch 2011; 2013. The presence of Celtic proper names in the surviving SW texts has been acknowledged for some time, and the possibility that the matrix language in which they occur might also be Celtic has been discussed previously (cf. Correa 1989; 1996, 72f; Untermann 1995, 253–5), but it is J. Koch’s recent works that have put this notion at the centre of the current linguistic debate (cf. Eska 2014; J. Koch 2014a; 2014b; Prósper 2014; Valério 2014). 22 Ferreira & Almeida 1971, 119–22; Schubart 1975, 19–24; Soares 1994, 180–2; Soares et al. 2009, 437–45; Alves et al. 2010, 135–46; Porfírio 2014, 253–64; Serra 2014, 274–7; Vilaça 2014, 109–12. 23 Schubart 1973; Kalb 1994; Mataloto 2005, 121–7; Rocha & Duarte 2009, 772 n 6.

Brandherm CW3 CS4 19 i 2016.indd 183

27/05/2016 15:42:35

[ 184 ]

vi. Stelae, funerary practice, and group identities

]]

6.4 Atalaia (Ourique, Baixo Alentejo), Middle Bronze Age grave system IV: top – topsoil stripped; centre – paving removed; bottom – cross section (after Schubart 1975, inserts 6 & 7) 6.5 Facing page: Alfarrobeira (São Bartolomeu de Messines, Algarve), reconstruction of agglomerated Middle Bronze Age cist cemetery (after Varela Gomes 1994, fig. 48)

Brandherm CW3 CS4 19 i 2016.indd 184

27/05/2016 15:42:41

Brandherm

[ 185 ]

small cairns or paved and sometimes raised platforms.24 While drawing a sharp distinction between actual cairns and raised platforms may not always be feasible, the former tend to have a more northerly distribution than the latter, which remain largely limited to the southern Baixo Alentejo and the Algarve.25 In the south-western Baixo Alentejo, between the upper reaches of the Mira and Sado rivers, paved and kerbed platforms have been stacked horizontally to form agglomerated honeycomb-like structures (Figure 6.4 & 6.5). In the Algarve, the tradition of embedding tombs in paved platforms harks back at least to the 3rd millennium BC, and here the Anta de Malhão (Afonso Vicente, Alcoutim) also provides the first example of a stela erected over such a platform.26   The horizontal stacking of such structures, however, does not seem to occur before the Middle Bronze Age, and the same holds true for most types of tombs mentioned above, other than simple cists. On the other hand, evidence for chronological differentiation between the various types of funerary architecture over the course of the Middle Bronze Age is limited. Regional differences notwithstanding, the available radiocarbon dates seem to indicate that the diverse traditions of hypogaea, rock-cut tombs, silo burials, and cist cemeteries coexist over various centuries prior to the Middle Bronze Age/Final Bronze Age transition.27 24 Varela Gomes 1995, 141f; Parreira 1998, 270–2; Porfírio 2014, 261f. 25 Viana 1959, 28–36; 1960, 188–95; Schubart 1975, 11–16; Varela Gomes et al. 1986, 63–72; Parreira & Barros 2007, 94. 26 Cardoso & Gradim, 2003, 170–2; 2010, 60f. 27 García 1998, 166–74; Mataloto et al. 2013, tab. 2; Porfírio 2014, fig. 10.

Brandherm CW3 CS4 19 i 2016.indd 185

27/05/2016 15:42:43

[ 186 ]

6.6 FBA (Group II) slabs and stelae: 1 Granja de Céspedes (after Harrison 2004, fig. p. 276); 2 Baraçal (ibid. fig. p. 193); 3 Brozas (ibid. fig. p. 206); 4 Solana de Cabañas (ibid. fig. p. 219); 5 Magacela (ibid. fig. p. 255); 6 Ategua (ibid. fig. p. 299)

  With the beginning of the Final Bronze Age, from c. 1300/1250 BC, most of those funerary architectures seemingly quickly disappear. Human remains are now generally disposed of in a manner leaving no discernible trace in the archaeological record. The only notable exception is the continued currency of silo or pit burials in parts mostly of the Alto Altentejo, where they appear to linger on until the end of the 2nd millennium BC.28   Other than the latter, only a handful of inhumation burials are known from the Final Bronze Age in south-western Iberia. Based on the associated pottery, a singular inhumation from a cave in the Peña de Arias Montano (Alájar, Huelva) can be dated to the earlier part of 28 Soares et al. 2009, 448; Antunes et al. 2012, 279; 282f; Porfírio 2014, 264.

Brandherm CW3 CS4 19 i 2016.indd 186

27/05/2016 15:42:47

Brandherm

[ 187 ]

the Final Bronze Age.29 No grave goods are known from an inhumation burial at Granja de Céspedes near Badajoz, and there are no radiocarbon dates in this case, but the iconography on the grave slab covering this burial suggests an attribution to the same period (Figure 6.6, 1).30 For a burial from Solana de Cabañas (Logrosán, Cáceres) the situation is similar. Here also, human remains were observed in association with a decorated stela or grave slab of Final Bronze Age date, but in this case the information provided in the original report is insufficient to determine if we are dealing with an inhumation or a cremation burial (Figure 6.6, 4).31 However, it may not be a coincidence that the monuments from Granja de Céspedes and Solana de Cabañas feature a more archaic iconography than the stelae or grave slabs found with burials securely identified as cremations (see below).   A small number of inhumations possibly dating to the Final Bronze Age has also been identified among the human remains from much earlier megalithic structures. At Zambujeiro 4 (Montemor-o-Novo, Alto Alentejo), Centirã 2 (Serpa, Baixo Alentejo), and La Venta (Zalamea la Real, Huelva) these were not accompanied by any grave goods, but radiocarbon dates obtained from the bones suggest dates in the later 2nd and early 1st millennia BC.32 The only other inhumation burials from south-western Iberia which based on their context can be securely dated to the Final Bronze Age are known from La Arboleda (San Juan del Puerto, Huelva)33 and Vega de Santa Lucía (Palma del Río, Córdoba).34 The latter of these displays a number of singular features. It stands out not only for being located within a coeval settlement, but also for the rather unusual representations of water fowl depicted on a clay tablet retrieved from the grave pit, which bear a close resemblance to symbolic depictions of similar birds in the Urnfield culture of central Europe.35   This is not the only indication that at least some of the profound changes in funerary practice at the Middle to Final Bronze Age transition may be linked to the adoption of new ideas about death and the afterlife originating from other parts of the Bronze Age world. Rather, the general disappearance of archaeologically visible burial rites broadly coincides with the incorporation of a range of new items into the material culture of south-western Iberia. Whereas the representation of Rosnoën swords and Herzsprung shields on some of the warrior stelae of the Final Bronze Age36 seems to indicate links with the Atlantic Late Bronze Age of north-western Europe, representations of crested helmets, an imported Rixheim sword from the Ría de Huelva deposit, winged axes of Type Grigny and local copies of Urnfield pottery point at connections between southern Iberia and the area of the 29 Gómez et al. 1992. 30 Almagro 1966, 105–7. 31 ibid. 27–9. 32 Henriques et al. 2013, tab. 3 (Centirã 2: 1265–1031 cal. BC [Sac-2789: 2950 ± 80 BP]); Mataloto et al. 2013, tab. 2 (Zambujeiro 4: 1400–1190 cal. BC [Beta-196093: 3040 ± 40 BP]; La Venta: 1015–917 cal. BC [Beta-150152: 2820 ± 40 BP]). All radiocarbon dates in the present contribution have been calibrated to 1 sigma with the Calib 7.1 software, based on the IntCal13 calibration curve (Reimer et al. 2013). 33 Gómez 1997, 206. 34 Torres 2004, 428. 35 Murillo 1994, 127–31, plate 4.3. 36 Brandherm 2013a, 134f.

Brandherm CW3 CS4 19 i 2016.indd 187

27/05/2016 15:42:47

[ 188 ]

vi. Stelae, funerary practice, and group identities

]]

6.7 Atlantic and Urnfield elements in the Final Bronze Age of southern and south-western Iberia: 1–3 Type Grigny winged axes and single-looped palstave from Arroyomolinos, Jaén (after Siret 1913, fig. 131); 4 & 5 RSFO style decorated urn and Proto-Villanova razor from near Beja, Baixo Alentejo (after Vilaça 2009, figs 8 & 9); 6 Type Herzsprung shield and Group C I helmet on stela from Santa Ana de Trujillo, Cáceres (after Harrison 2004, fig. p. 212); 7 Type Rosnoën sword on stela from Fóios, Sabugal, Beira Alta (ibid. fig. p. 194); 8 Type Rixheim sword from the Ría de Huelva, Huelva (after Brandherm 2007, plate 1, 6), not to scale

western Urnfield culture at this time (Figure 6.7).37   While it cannot be ruled out that Urnfield influence is also making itself felt in the sporadic occurrence of cremation burials dating to the Final Bronze Age, these remain few in number and are also limited in their geographic distribution. Some authors have questioned the adoption of urn burials prior to the beginning of the Early Iron Age.38 However, based on their association with stelae displaying an iconography clearly attributable to the Final Bronze Age at Setefilla (Lora del Río, Sevilla)39 and Cortijo de la Reina (Guadalcázar, Córdoba)40, it seems more than likely that the practice of depositing cremated human remains in pottery urns on the lower Guadalquivir was adopted well before the transition to the Iron Age. An early date for the introduction of this type of burial is also supported by 37 Lucas 1995, 117f; Brandherm 2013b, 148–50; Vilaça 2014, 112 fig. 6. 38 Belén 2001, 44. 39 Aubet 1997, 163–5. 40 Murillo et al. 2005, 25–32.

Brandherm CW3 CS4 19 i 2016.indd 188

27/05/2016 15:42:54

Brandherm

[ 189 ]

6.8 Examples of FBA and EIA urn types: 1 & 2 biconical urns (after Aubet 1978, figs 15, 3; 22, 2); 3 &

4 á chardon type urns (ibid. figs 19, 1. 4); 5 & 6 Cruz del Negro type urns (ibid. 1976– 1978, fig. 1, 3. 4)

{ 0

15 cm

radiocarbon evidence from Tanchoal dos Patudos (Alpiarça, Ribatejo)41, although this site is located north of the Tagus river, where the development of burial traditions would appear to follow a slightly different trajectory.42   As a general rule of thumb, in south-western Iberia the cremation burials of the Final Bronze Age are distinguished from those of the Early Iron Age by a much more limited range of urn types. With very few exceptions, in the Final Bronze Age we are dealing with vessels of biconical shape, and while similar pots are still occasionally used as urns in the Early Iron Age, the vast majority of urn vessels following the Final Bronze Age/Early Iron Age transition are either of á chardon or Cruz del Negro type (Figure 6.8).43 Other characteristic features of Final Bronze Age cremation burials are a general lack of grave goods and the minute amount of human remains deposited. These are true token burials, the most extensive evidence for which so far has been identified at Setefilla and Alpiarça.44 Other sites that have produced cremation burials of a similar date and nature, though generally less well documented, are Rabadanes (Las Cabezas de San Juan, Sevilla), Los Praditos (Aroche, Huelva), La Nicoba (San Juan del Puerto, Huelva), Las Cumbres (Puerto de Santa María, 41 Vilaça et al. 1999, 14 (1048–917 cal. BC [GrA-9270: 2830 ± 50 BP]; 1007–856 cal. BC [GrA-9572: 2790 ± 50 BP]). 42 cf. Spindler 1981, 192–5; Vilaça & Arruda 2004, 17, 28f. Also, in contrast to the situation in southern Iberia, the adoption of cremation burials in the Final Bronze Age of central Portugal may hardly be linked to Eastern Mediterranean influence. Rather, the range of pottery types encountered at sites like Tanchoal dos Patudos (conc. Alpiarça) is clearly indicative of Urnfield connections (cf. Kalb 1995, figs. pp. 91f). 43 Torres 1999, 175f. 44 Aubet 1975, 78–123 figs. 12, 38, 50; ibid. 1978, 184–204, figs. 15, 20, 22, 28; Kalb & Höck 1980, 101–4, figs. 8, 9.

Brandherm CW3 CS4 19 i 2016.indd 189

27/05/2016 15:42:58

[ 190 ]

vi. Stelae, funerary practice, and group identities

]]

prov. Cádiz) and Mesas de Asta (Jerez de la Frontera, Cádiz).45   Apart from the above mix of flat cemeteries and barrows, like in other parts of southern Iberia, a small number of cremation burials dating to the Final Bronze Age have been recovered from much earlier megalithic structures. Their ongoing, if occasional use for funerary purposes is one of the few continuities spanning the entire duration of the Bronze Age in south-western Iberia, continuing even into the Early Iron Age.46 For cremated human remains from the tholos tomb of Nora Velha (Ourique, Baixo Alentejo)47 and the Dolmen de Palacio III (Almadén de la Plata, prov. Sevilla)48 a Final Bronze Age date seems certain. For similar cremation burials from other megalithic tombs in south-western Iberia, an attribution to the Early Iron Age would appear equally plausible.49   Other than Tanchoal dos Patudos and the Dolmen de Palacio  III, none of the aforementioned sites have so far produced any scientific dating evidence for Final Bronze Age cremation burials. However, following conventional typo-chronologies, both the iconography of the stelae from Setefilla and Cortijo de la Reina as well as the pottery types used as urns seem to indicate a date between the 11th and 9th centuries BC for this earliest horizon of cremation burials in south-western Iberia.50 This would render them coeval with similar burials in the southern Levant and in eastern Andalusia.51   Despite this sporadic occurrence of both inhumation and cremation burials in the Final Bronze Age of south-western Iberia, there is no denying that the transition from the Middle to the Final Bronze Age heralds a wholesale abandonment of most archaeologically visible burial traditions across the region. First and foremost this holds true for the cist burials that dominate much of the funerary record of the Middle Bronze Age, be it in the form of loose clusters in the Algarve and parts of westernmost Andalusia, or in the from of agglomerated honeycomb-like structures in the southern Baixo Alentejo. From the last quarter of the 2nd millennium and the first two centuries of the 1st millennium BC, little more than a handful of inhumation burials is known from south-western Iberia, whereas neighbouring regions have produced a significantly larger quantity, albeit still very limited in terms of 45 Torres 2004, 426f; Pellicer & Escacena 2007, 11–13. 46 Schubart 1975, 23. 47 Cardoso 2004, 206. 48 García 2005, 598f. No grave goods accompanying this burial have been identified, but the cremated human remains have produced a 14C date of 940–760 cal. BC (Beta-165552: 2660 ± 90 BP). 49 Cardoso 2004, 203–9. Apart from the instances mentioned by this author, fragments of a Type Monte Sa Idda sword recovered from the megalithic tomb of Las Alcobainas (Jerez de la Frontera, prov. Cádiz) likely indicate the presence of a Final Bronze Age/Early Iron Age burial also at this site (Esteve 1954/55). 50 The traditional low chronology that would situate these remains between the 9th and 7th centuries BC can no longer be upheld (cf. Murillo 1994, 65–126; Ruiz Mata 1995, 266–71; Murillo et al. 2005, 28–31). It is hoped that research currently underway as part of a project conducted at Adam Mickiewicz University in Poznań and Queen’s University Belfast will address the dearth of 14C dates for funerary assemblages from the Final Bronze and Early Iron Ages of south-western Iberia (Krueger & Brandherm forthcoming). However, the low crystallization indices of cremated human bone material from most contexts that have been sampled to date as part of this project continues to pose considerable difficulties when it comes to establishing a 14C-based chronological framework along the lines recently advocated by Balsera et al. 2015, 144–6. 51 cf. Lorrio 2008, 324–57; 2009/10, 139–43.

Brandherm CW3 CS4 19 i 2016.indd 190

27/05/2016 15:42:58

Legend Brandherm

               

Barrow Cave [ 191 ] Cist Flat grave Hypogaeum Megalithic monument Silo burial Cremation burial

6.9 Distribution of Final Bronze Age funerary sites in SW Iberia (data from AEMA database)

their overall number.52 While within our main region of interest coeval cremation burials occur somewhat more numerously, they remain largely limited to the lower Guadalquivir basin. Despite the general similarity between the latter and the Final Bronze Age cremation burials from the lower Tagus region, they are but completely lacking from Portugal south of the Sado river (Figure 6.9).   This pattern becomes even more pronounced during the Early Iron Age, when the number of burials across the south-western quadrant of the Iberian Peninsula experiences a dramatic rise. In the lower Guadalquivir basin the quota of inhumations among the overall burial population now drops below ten percent. This figure includes a clear social bias, as more than a third among the central graves of Early Iron Age barrows contain inhumation burials.53 Among the very richest category of burials they even form the majority.54   In any case, the overall picture of burial customs in the lower Guadalquivir basin during the opening stages of the Iron Age presents itself as clearly different from the situation seen in the Algarve and the Baixo Alentejo, where we now observe a general resurgence of inhumation burials, contrasted by an almost total lack of cremations throughout the Early Iron Age. Particularly for the western Baixo Alentejo, the area between the upper Mira and Sado rivers, an extensive programme of surveys and excavations conducted by Caetano de 52 The vast majority of the relevant finds is from eastern Andalusia (Lorrio 2008, 379–81); for Portugal only the double inhumation burial from Roça do Casal do Meio near Sesimbra can be named here (Spindler & Ferreira 1973). 53 Beba 2008, 17 fig. 3. 54 ibid. 127f.

Brandherm CW3 CS4 19 i 2016.indd 191

10/06/2016 17:57:53

[ 192 ]

]]

6.10 Pego da Sobreira (Ourique, Baixo Alentejo), Early Iron Age funerary structure

(after Correia & Parreira 2002, fig. p. 56)

Mello Beirão55 throughout the 1970s and ’80s was able to demonstrate that the funerary architecture of the opening phase of the Iron Age, usually consisting of cist graves at the centre of paved and kerbed circular platforms or cairns, closely mimics the funerary structures built in the area during the Middle Bronze Age, and that these generally contained inhumation burials (Figure 6.10). In other parts of the Baixo Alentejo and in the Algarve we observe a similar return to Middle Bronze Age funerary practices, with the re-emergence of inhumation burials in dispersed cist cemeteries. In the upper Mira-Sado area, it is only with the appearance of rectangular grave structures during the later part of the Early Iron Age, that cremation burials start to occur (Figure 6.11), and it is not until the beginning of the later Iron Age, from the mid-1st millennium BC, that inhumation burials largely disappear and the practice of depositing cremated remains in cinerary urns is generally adopted throughout the wider region.56 55 Beirão 1986, 49. 56 Dias & Coelho 1983, 197; Correia 1993, 360; Rodríguez 2002, 86f; Barros et al. 2013, 1171f. Due to the close resemblance between simple cists dating to the Middle Bronze and Early Iron Ages, where diagnostic grave goods are lacking, it can be difficult to distinguish between the two in individual instances. As a general rule, however, Iron Age cists tend to have somewhat more rectangular proportions than their

Brandherm CW3 CS4 19 i 2016.indd 192

27/05/2016 15:43:13

Brandherm

[ 193 ]

6.11 Vaga da Cascalheira (Ourique, Baixo Alentejo), Early Iron Age agglomerated drystone cist burials (after Correia & Parreira 2002, fig. p. 58)

  Thus, the development of funerary customs in the Algarve and Baixo Alentejo stands in stark contrast to the treatment of the dead in the Guadalquivir basin and neighbouring parts of Extremadura, where cremation burials and the use of cinerary urns had become very much the norm already in the Final Bronze Age.57 The only exceptions from this general rule Middle Bronze Age predecessors. Particularly in the Algarve, Iron Age cists frequently also feature slightly trapezoidal rather than rectangular cross sections (cf. Rocha 1896; Schubart 1975, 20–3; Varela Gomes et al. 1986; Frías & Pérez 1989; Varela Gomes 1994; Enríquez & Carrasco 1995; Silva & Soares 2009). 57 This comes with a number of caveats. First of all, an accurate assessment of the development of funerary customs in the Algarve and Baixo Alentejo is hindered by the fact that the vast majority of Early Iron Age graves in this region were systematically emptied of any grave goods soon after the burial took place. To what extent this represents systematic looting or may form an integral part of funerary practice is difficult to determine (Correia 1993, 353f; Cardoso & Gradim 2006, 121f ). Also, any interpretation of the relevant archaeological evidence depends on the rather summary publication of Beirão’s extensive fieldwork activities (Beirão 1986; 1990). Because of these difficulties, some authors have suggested that already from the start of the Iron Age, inhumation and cremation burials would have existed side by side across all of southern Portugal (Correia 1993, 355f; Arruda 2000, 103f; 2004, 471). However, having excavated the vast majority of relevant sites in the southern Alentejo, Beirão himself was quite adamant that funerary structures with a circular ground plan, i.e. those graves dating to the earlier part of the Early Iron Age, invariably contained inhumation burials (Beirão 1986, 49). Brandherm CW3 CS4 19 i 2016.indd 193

27/05/2016 15:43:22

Legend

6.12 Distribution of Early Iron Age funerary sites in SW Iberia (data from Beirão 1986; Torres Ortiz 1999; Beba 2008)

are the necropolises directly linked to the Phoenician presence in the south-eastern Algarve (e.g. at Balsa, modern-day Tavira)58 and some isolated urn burials found on the western bank of the Guadiana (e.g. Palhais, Beja).59 However, except for the SW Orientalizing Complex (SWOC)60 enclave at Mértola, the latter constitute only a short-lived intrusion, and from the middle of the 7th century BC until the end of the Early Iron Age inhumation burials take over again in this part of southern Portugal.61   It is not only in the treatment of the dead that the Algarve and Baixo Alentejo follow very much a different trajectory from the regions east of the Guadiana. Also the funerary architecture of the Early Iron Age here is very different, with a deliberate return to Middle Bronze Age traditions. After a hiatus of several centuries, in the Early Iron Age of this area we witness a revival both of dispersed cist cemeteries and of agglomerated cist burials under paved and kerbed platforms.62 On the other hand, and in stark contrast to what can be observed in neighbouring regions, any referencing of Final Bronze Age funerary traditions appears to be purposefully avoided (Figure 6.12). 58 Arruda et al. 2008. 59 Santos et al. 2010, 777–83. 60 This label is introduced here as a substitute for the misleading ‘Tartes(s)os’ nomenclature when referring to archaeological sources (Spanish: COSO = Complejo Orientalizante del Suroeste). This is not to deny that he term ‘Orientalizing’ also has its problems. However, while in principle accepting the validity of Said’s (1978) criticism levelled at the underlying concept, we would maintain that most of his arguments bear no particular relevance to Mediterranean Iron Age archaeology (contra Purcell 2006). In any case, in the present context this term turns out to be certainly much less problematic than ‘Tartes(s)ian’. 61 Barros 2010, 428f. 62 Correia 1993, 356; 2014, 79; Correia & Parreira 2002, 50–61; Cardoso & Gradim 2006, 210; 2008, 109f.

Brandherm CW3 CS4 19 i 2016.indd 194

10/06/2016 17:58:21

Brandherm

[ 195 ]

  The absolute chronology of the respective funerary structures continues to cause some controversy among scholars. The only two radiocarbon dates currently available for cist burials from an Early Iron Age paved and kerbed ‘tumulus’ pertain to Structure VIII at Nora Velha 2 (Ourique, Baixo Alentejo).63 If taken at face value, they would confirm a re-emergence of this type of burial in the 8th century BC, as previously argued by Virgilio  Correia.64 However, other authors have suggested that these dates may suffer from an old-wood effect, and that structures of this type do not occur until the following century,65 or even that the Early Iron Age necropolises of the upper Mira-Sado region in their entirety do not predate the 5th century BC.66   Another important distinction in the development of funerary practice that sets the main distribution area of Early Iron Age cist graves apart from neighbouring regions only emerges following the ‘crisis of the 6th century’. At that stage, in those parts of southwestern Iberia which since the early 1st millennium BC had been characterized by a clear dominance of cremation burials, archaeologically visible burial rites once again largely disappear.67 However, this does not affect the Algarve and Baixo Alentejo, where at the beginning of the Iron Age we observed a re-emergence of funerary traditions harking back to the Middle Bronze Age.68   It can hardly be coincidental that we are seeing such a diametrically opposed behaviour in terms of funerary practice between what is generally taken to be the core area of the SWOC in the lower Guadalquivir basin on the one hand and the lands to the west of the Guadiana on the other. Rather, it seems difficult to escape the conclusion that this mutually anti-cyclical development, starting in the Final Bronze Age and stretching into the later Iron Age, draws on deep-rooted notions of communal identity.69 In the following, we shall refer to the archaeological culture of southern Portugal identified by a return to Middle Bronze Age funerary practice in the Early Iron Age as the Algarve-Baixo Alentejo Complex (ABAC).70 Stelae and grave slabs In the Algarve and southern Baixo Alentejo, the return during the Early Iron Age to forms of funerary practice that had been current throughout this region already in the Middle Bronze Age, but apparently abandoned during the centuries in between, is not limited to the re-emergence of inhumation burials and specific templates for funerary architecture. Exactly 63 Vilhena 2008, 380–3 tab. 3 (Grave 8A: 905–820 cal. BC [ICEN-1102: 2720 ± 50 BP]; Grave 8B: 802–541 cal. BC [ICEN-1103: 2540 ± 90 BP]). 64 Correia 1993, 360. 65 Soares & Martins 2013, 665. 66 Jiménez 2004, 107f. 67 Belén 2001, 68; Neville 2007, 163–70. Within those parts of south-western Iberia which from the early 1st millennium BC are characterized by a dominance of cremation burials, it is only in the coastal enclaves under direct Phoenician control that this continues unabated beyond the 6th century BC. 68 Barros et al. 2003, 46–9. 69 cf. Arruda 2009, 517; Albuquerque 2013b, 48f. 70 Portuguese: CABA = Complexo do Algave-Baixo Alentejo.

Brandherm CW3 CS4 19 i 2016.indd 195

27/05/2016 15:43:31

[ 196 ]

vi. Stelae, funerary practice, and group identities

]]

the same phenomenon can also be observed with the reappearance of decorated stelae and grave slabs at the start of the Iron Age.71   Over the last few decades various classificatory systems have been proposed for the non-anthropomorphic stelae and grave slabs of south-western Iberia.72 While occasionally applying different labels, most of these have followed Almagro73 in distinguishing two basic groups, with his Group I comprising the decorated slabs and stelae of the Middle Bronze Age and his Group II including the so-called ‘warrior stelae’ and related grave slabs of the Final Bronze Age (Figure 6.13).74   For the latter, Pingel75 proposed a further distinction based on iconographic criteria, leading to his definition of Groups IIa, IIb, and IIc. The first of these is characterized by the representation of sword, shield and spear only (Figures 6.6, 1 & 2), the second sees the addition of other object types to this canonical set (Figure 6.6, 3), while the third is defined by the appearance of human figures (Figures 6.6, 4, 5, & 6).76 Some likely temporal overlap between these various subgroups notwithstanding, Pingel’s scheme for the first time facilitated a chronological development to be recognized among Group  II monuments.77 Within his Group IIc, subsequently Varela Gomes and Monteiro78 distinguished between compositions in which the shield continues to occupy a central position—as it had already done with the monuments of Groups IIa and IIb—and iconographies in which the human figure takes centre-stage. Only for the former do they continue to use Pingel’s Group IIc label (Figure 6.6, 4), while the latter are referred to as Group IId (Figures 6.6, 5 & 6). Later authors, in trying to develop terminologies applicable also to other, mostly anthropomorphic categories of stelae, have proposed alternative nomenclatures.79 However, the present contribution will continue to use the classificatory criteria and terminology originally outlined by Almagro, and further developed first by Pingel and then by Varela Gomes and Monteiro, as these continue to provide the most robust framework for understanding this particular group of 71 For a majority of specimens, an original function as stelae can be inferred from a clear division between the decorated or inscribed part of their surface and a plain section that would have been inserted into the ground to maintain the monument in an upright position (cf. Valério 2008, 108). However, this does not hold true in all cases, and particularly among the monuments of Groups I and III we find a significant number of comparatively thin slabs lacking such a partition. For the latter, other dispositions would seem more likely. For a comprehensive overview of Chalcolithic and Bronze Age stelae traditions in Iberia see Díaz-Guardamino 2010; for the respective Iron Age monuments also see Quesada 2011. 72 Celestino & Salgado 2011, 420–31. 73 Almagro 1966. 74 For a current inventory of Group I monuments see Díaz-Guardamino 2010, fig. 187, with additional examples documented in Banha et al. 2009, and Serra 2014, tab. 3; for the most recent listing of Group II monuments see Celestino & Salgado 2011, 438–42; with additional specimens in Alves & Reis 2011, 201–4 plate 8; Reboreda & Nieto 2012; Bettencourt 2013, 167 fig. 10 75 Pingel 1974. 76 In a number of instances, Group II stelae were modified from their original state by the later addition of iconographic elements. Attribution to individual subgroups in our Figure 6.17 is based on the composition of elements in their finale state. 77 cf. Brandherm 2007, 21–5. 78 Varela Gomes & Monteiro 1977, 185–8. 79 e.g. Celestino 2001, 91–7; Díaz-Guardamino 2010, 334–40; Celestino & Salgado 2011, 423–31.

Brandherm CW3 CS4 19 i 2016.indd 196

27/05/2016 15:43:31

6.13 MBA (Group I) slabs and stelae: 1 Mombeja I (after Almagro 1966, plate 8); 2 Assento (ibid. plate 28); 3 Gomes Aires (ibid. plate 36); Ervidel I (after Varela Gomes & Monteiro 1977, plate 5)

6.14 EIA (Group III) slabs and stelae: Mealha Nova I

(after Koch 2013, fig. p. 94); Bastos (ibid. fig. p. 101); Fonte Velha VI (ibid. fig. p. 30); Vale dos Vermelhos III (ibid. fig. p. 53)

monuments.80   Building on this framework, in the following we will be referring to the inscribed stelae and slabs of the Early Iron Age as Group III of the non-anthropomorphic stelae tradition of south-western Iberia.81 Based mostly on palaeographic criteria, two well-defined subgroups can be distinguished within this particular category of monuments. The stelae and slabs of Group IIIa are characterized by a relatively plain execution of their inscriptions (Figure 6.14, 1 & 2). Their distribution remains largely limited to the Baixo Alentejo. In contrast, the characters found on Group IIIb monuments display more baroque features, and the relevant 80 No further subdivision of Group  I monuments based on iconographic criteria is attempted here. Previously, Varela Gomes (1994, 119 fig. 71) proposed a distinction between three different subgroups (Ia: central position of the ‘anchor-shaped object’; Ib: more complex compositions in which the ‘anchorshaped object’ has lost its central role; Ic: central position of the sword). However, this approach does not produce the same kind of clear results as is the case with Group  II. In particular, based on the monuments currently known, the chronological reading of these various subgroups along the lines proposed by Varela Gomes remains unconvincing. The methodological shortcomings of an earlier attempt by Almagro (1966, 197) to divide his Group I into two subgroups using iconographic criteria have already been highlighted by Schubart (1975, 108 n 581). 81 Brandherm 2013a, 141.

Brandherm CW3 CS4 19 i 2016.indd 197

27/05/2016 15:43:46

[ 198 ]

vi. Stelae, funerary practice, and group identities

]]

6.15 Chronology of objects represented on Group I–III monuments (after Brandherm 2013a, pate 11)

inscriptions also include a wider range of different formulae (Figure 6.14, 3 & 4).82 It is interesting to note that the distinction between these two subgroups appears to coincide with differences in Early Iron Age funerary architecture within the ABAC area, in as far as the distribution of Group III  a monuments largely corresponds to that of agglomerated cist burials in the south-western Baixo Alentejo, whereas the distribution of Group III b monuments generally matches that of the more loosely structured cist cemeteries of the Algarve.   Turning to the question of the chronology of the three main groups of nonanthropomorphic stelae and grave slabs found in south-western Iberia, the typology of the objects represented on Group I monuments—sword and halberd types in particular—leaves 82 Correia 1997, 274f.

Brandherm CW3 CS4 19 i 2016.indd 198

27/05/2016 15:43:49

Brandherm

[ 199 ]

little doubt that in their majority they date to the earlier part of the Middle Bronze Age.83 However, a number of specimens can clearly be attributed to the later part of this period (Figure 6.15). A sword depicted on slab  I from Mombeja currently represents the latest object type shown on any Group I monument (Figure 6.13, 1). Its hilt shape clearly betrays Eastern Mediterranean influence, and its immediate prototypes are unlikely to date much later than c. 1300 BC.84   A general synchroneity between Group  I monuments and the Middle Bronze Age cist cemeteries of the southern Alentejo and Algarve is also supported by a number of instances in which Group I slabs were used as cist covers. This seems to have been the case at Santa Vitória, Trigaches, Mombeja, and possibly Marmelete, though these are old finds and the available documentation in all of these instances remains somewhat unsatisfactory.85 Also, at least in some of these cases, there is evidence to suggest that their use as cist covers may not correspond to the original purpose of the relevant slabs.86 These uncertainties notwithstanding, it seems safe to assume that even where we are dealing with instances of secondary use, the original purpose of these monuments would have been closely linked to the funerary sphere. This is clearly attested by grave 2 from the cist cemetery at Alfarrobeira, where we have no reason to doubt that the use of an associated Group I stelae as a grave marker represents its original function.87 Finally, grave system V at Atalaia has produced a miniature stela whose technical execution betrays clear links to typical Group I monuments, even if its iconography is very different.88 An intimate connection of these monuments to the funerary sphere is also suggested by the representation of a pair of footprints on the slabs from Gomes Aires (Almodôvar, Baixo Alentejo)89 and Ervidel I (Aljustrel, Baixo Alentejo)90 (Figures 6.13, 3 & 4). Very similar footprint motifs have been identified as a symbol of physical departure or death—with a possible inverse reading as an indicator of divine presence in some instances—in Scandinavian rock art of the same period.91   The geographic distribution of the slabs and stelae of Group I largely coincides with the distribution area of Middle Bronze Age cist cemeteries in the Baixo Alentejo and Algarve (Figure 6.16). On the other hand, their distribution stands in sharp contrast to that of the later Group  II monuments (Figure 6.17). While there can be little doubt that the characterization amongst the latter of different subgroups by Pingel92 and Varela Gomes 83 cf. Barceló 1991, 21f; Varela Gomes 2006, 47f; Díaz-Guardamino 2010, 312–15. 84 Brandherm 2013a, 132. 85 Díaz-Guardamino 2010, fig. 183. 86 ibid. 313f. 87 Gomes 1994, 25 figs 8–11. 88 Schubart 1975, 234 fig. 29. 89 Almagro 1966, 120 fig. 41. 90 Coelho 1975. 91 Hauptman Wahlgren 2002, 74. Like the aforementioned representation of an Eastern Mediterranean sword-hilt type on a Group I slab from Mombeja, this evidence for the sharing between south-western Iberia and western Scandinavia of symbolic elements not otherwise attested in the Atlantic Bronze Age raises a number of interesting questions which are, however, beyond the scope of the present contribution (cf. Celestino 2001, 125–35; 177–81; 224–9; Vilaça 2007; Brandherm 2013a). 92 Pingel 1974; 1993.

Brandherm CW3 CS4 19 i 2016.indd 199

27/05/2016 15:43:49

Legend

[ 200 ]

i. behind the warriors

]]

  Group I   Re-use of earlier   stela or menhir

6.16 Distribution of Group I monuments and related anthropomorphic stelae (data from AEMA database)

6.17 Distribution of Group II monuments, related anthropomorphic stelae and rock-art panels (data from AEMA database)

Legend   Group II a   Group II b   Group II c   Group II d   Group II unclassified   Re-use of earlier stela   or menhir   Stela-style rock panel

Brandherm CW3 CS4 19 i 2016.indd 200

27/05/2016 15:44:01

Brandherm

[ 201 ]

and Monteiro93 to some extent reflects a chronological trajectory, the main point to note here is that the temporal range of the object types represented on Group  II monuments spans more or less the entire duration of the Final Bronze Age. The lack of typological detail particularly among the representations found on monuments of Groups IIc and IId makes it difficult to ascertain if any of the relevant objects might be of Iron Age date.94 The currency of Group II then would seem to cover at least the period from c. 1200 to 900 BC. However, a start date already in the 13th century and an end date only around 800 BC can by no means be ruled out (Figure 6.15).   In the vast majority of cases we have no information on the original context of Group II monuments. Most have not been found in their original locations, and even in those few cases where there is reason to believe that we are not dealing with secondary contexts, detailed documentation of find circumstances is usually lacking. Their occasional association with human remains has already been mentioned above. Given the development of funerary practice in the Final Bronze Age of south-western Iberia, it may be significant that definite inhumation burials so far are only attested in conjunction with Group II  a monuments, whereas cremated remains have only been found with slabs or stelae of Groups IIc and IId.95 Disregarding the human remains discovered with the Group IIc slab or stelae from Solana de Cabañas, which may or may not have been cremated, it is also unsurprising that all cases in which Group II monuments were found in association with cremation burials cluster in the lower Guadalquivir basin, where the new funerary practice first appears to have gained a foothold.   Considering the clear discrepancy between the main distribution areas of Group I and II slabs and stelae, a direct continuity between these two groups of monuments seems rather questionable. While so far only very few examples of Group I monuments have been identified outside their core distribution area in the Baixo Alentejo and Algarve, Group II specimens remain almost entirely absent from this particular region (Figures 6.16 & 6.17). What is more, between the spread of the latter on the one hand—ranging from the middle Tagus, across the middle Guadiana down to the lower Guadalquivir—and the distribution area of Group I monuments on the other, a marked gap emerges which has not produced monuments of either group. Only very few outliers cross the divide of this ‘no man’s land’. Also, for the only Group  II monuments from southern Portugal—the specimens from Ervidel96 and Figueira97—an early date within the development of this group is ruled out by 93 Varela Gomes & Monteiro 1977. 94 Díaz-Guardamino 2010, 349–61; 2012, 406–9 ; Mederos 2012, 425–45; Brandherm 2013a, 134. 95 Apart from the aforementioned Group II a specimen from Granja de Cespedes near Badajoz, which served as a covering slab for an inhumation burial, the poorly documented Group II a or II b stela from Toya (Peal de Becerro, prov. Jaén) has to be mentioned here. According to the excavator (Mergelina 1943/44, 27–30) it served as a marker for the collective inhumation of several individuals. The practice of using a single funerary space for multiple inhumation burials so far does not have any close parallels in the Iberian south-west, and in that respect this instance from the upper Guadalquivir basin might be more closely related to funerary practices current in the Final Bronze Age of south-eastern Iberia (cf. Lorrio 2008, 382f). 96 Díaz-Guardamino 2010, no. 298. 97 ibid. no. 304.

Brandherm CW3 CS4 19 i 2016.indd 201

27/05/2016 15:44:01

Legend   Group III a   Group III b   Group III unclassified

6.18 Distribution of Group III monuments (data from Correia 1996; J. Koch 2013)

their iconography. In turn, the northernmost examples of Group I-related monuments, while stretching to the lower reaches of the middle Tagus basin and beyond, without exception take the form of anthropomorphic stelae which are very different from the vast majority of both Group I and II a specimens, so that it is difficult to envisage them as a connecting link between the two groups.98   Given the lack of a direct link between Groups I and II, the clear parallels that exist between the monuments of Groups I and III are all the more striking. This close relationship was highlighted originally by Schubart.99 Despite the recognition since, of a chronological hiatus of several centuries between the two groups, his assessment still holds true. While based on palaeographic criteria the SW script is unlikely to have emerged before the late 9th century BC, providing a terminus post quem for the inscribed slabs and stelae of Group III,100 both their distribution area and the types of context in which they occur closely match those of their Middle Bronze Age predecessors. Like these, Group  III monuments are found mainly in the Baixo Alentejo and the Algarve (Figure 6.18), and as in their case, the majority of Group  III specimens with recorded find circumstances come from cist cemeteries.101 Notwithstanding the lack of detailed documentation in most cases, we have every reason to assume that, as was the case with the exponents of Group I, the majority of 98 Banha et al. 2009. Likewise, most examples of Group II-related stelae found to the north and east of the Tagus basin feature anthropomorphic or otherwise atypical features that set them apart from the majority of both Group I and II monuments (cf. Celestino 2001, 449–54 nos.90–3; Reboreda & Nieto 2012). 99 Schubart 1975, 108f map 34. 100 Rodríguez 2002, 87 n 2. 101 Dias et al. 1970, 191–3; Coelho 1971, 167f; 179; Beirão & Varela Gomes 1988, 118–20.

Brandherm CW3 CS4 19 i 2016.indd 202

27/05/2016 15:44:10

Brandherm

[ 203 ]

Group III specimens served as covering slabs for cist burials or as grave markers.102 There can be no doubt that they represent a deliberate attempt at emulating the funerary practice and invoking the sepulchral-related culture of memory of the Middle Bronze Age. As the funerary architecture of the Early Iron Age in southern Portugal as a whole attempts to mimic Middle Bronze Age traditions, the similarities between the sculpted slabs of the Bronze Age and the inscribed specimens of the Iron Age do not come entirely unexpected.   In comparison, the occurrence of similar inscriptions on two Group  II stelae from Capote (Higuera la Real)103 and Cabeza del Buey IV104, both in Badajoz province, very much remains a peripheral phenomenon. Like the Roman epitaph on the stela from Ibahernando (Cáceres),105 these isolated instances of SW inscriptions on Group II monuments almost certainly represent later additions and do not provide any missing link between Groups II and III.106 In the absence of a direct connection between Group II and III monuments, and the obvious parallels between Groups I and III, one has to ask for the reasons behind the clearly intentional revival of funerary practices, architecture, and monuments after several centuries of nearly complete abstinence by local communities from practising any archaeologically visible burial rites or erecting funerary monuments.107 Bridging the moyenne durée The fact that in the Early Iron Age of southern Portugal we witness a deliberate return to earlier forms of funerary practice should not come as a great surprise. In other parts of Europe, starting from the 9th century BC, very similar phenomena can be observed. The burial customs of the Hallstatt culture reference the funerary practice of the Tumulus and early Urnfield cultures in very much the same way as the cist burials of the Early Iron Age in south-western Iberia mimic their Middle Bronze Age predecessors. However, within the Iberian Peninsula at this time such a deliberate revival of earlier funerary practice seems quite unique. 102 As many Group III slabs lack an exact provenance or have been retrieved from what are clearly secondary contexts, the proportion of specimens recovered from cist cemeteries or their immediate environs amounts to only 40% of the total population. Based on this figure, their funerary function was recently disputed by Vilhena (2008, 375–9), who suggests that all Group III monuments found at or near funerary sites either predate the relevant necropolises or were brought there from their original locations as building material in a secondary capacity. 103 J. Koch 2013, 14f, 116. 104 ibid. 124. 105 Almagro 1966, 94. 106 It may be tempting to relate their presence to the intrusion of ABAC-inspired funerary architecture in the middle Guadiana basin from the 5th century BC (see note 129). However, we are currently lacking reliable criteria for establishing their chronological position in relation to similar inscriptions found on Group III monuments. 107 Parreira (1998, 269–73) has rightly stressed that with the transition from the Middle to the Final Bronze Age in southern Portugal, a shift from funerary to settlement structures as the main elements for structuring space and asserting control over parts of the landscape can be observed. The reverse appears to hold true for the transition from the Final Bronze Age to the Early Iron Age. However, the reasons behind these shifts still remain poorly understood (cf. Serra 2014, 289f).

Brandherm CW3 CS4 19 i 2016.indd 203

27/05/2016 15:44:10

[ 204 ]

vi. Stelae, funerary practice, and group identities

]]

  In contrast to Continental western and central Europe, where the return to much earlier practices in the burial customs of the late Urnfield and Hallstatt cultures constitutes a widespread phenomenon, affecting most regions between eastern France and the Carpathian basin. In contrast, by mimicking Middle Bronze Age funerary practices the Early Iron Age communities of the ABAC region set themselves decidedly apart from populations inhabiting the surrounding areas between the Tagus basin and the lower Guadalquivir. Not only does the return to Middle Bronze Age forms of expression constitute a sharp break with the practices prevailing in south-western Iberia throughout the Final Bronze Age, it also serves Early Iron Age communities in the southern Baixo Alentejo and Algarve to distance themselves from developments characterizing the beginning of the Iron Age east of the Guadiana, where urn burials are now widely adopted.   The exact motivations behind this deliberate distancing, given the nature of the available sources, are of course difficult to understand, but one can hardly escape the conclusion that the significant socio-economic changes which gripped indigenous communities following the foundation of Phoenician settlements along the coastline of Andalusia and southern Portugal, ultimately leading to the formation of the SW Orientalizing Complex of the Early Iron Age, had to play an important part in this development. Initially these changes would have made themselves felt most strongly in the area between the lower Guadiana and Guadalquivir rivers, the future ‘tartesius ager’, where the beginning of large-scale silver mining for export to the Eastern Mediterranean fundamentally transformed the economic basis of indigenous communities.108 It is probably safe to assume that the consequences of this transformation would not have been greeted with unfettered enthusiasm by all sectors of Early Iron Age society.109   Against this backdrop, the deliberate attempt by the Early Iron Age population of southern Portugal to distance themselves from their SWOC neighbours by returning to traditional funerary practices is best understood as the expression of a nativist movement rejecting the profound socio-economic changes that affected south-western Iberia at the start of the Iron Age. If one accepts this assessment, it may also help to shed new light on other phenomena characterizing the beginning of the Iron Age west of the Guadiana.   This includes the writing system found on Group  III monuments, as functional considerations alone would not seem to provide a satisfactory rationale for the development of a new semi-syllabic script in this instance. Assuming that J. Koch’s110 reconstruction of the phonetic system for the language of these inscriptions is by and large correct, the SW script would have offered little or no advantage for writing that language when compared to the Phoenician aleph-beth, from whom its characters are ultimately derived. It is also interesting to note that its system of stop consonant + vowel combinations appears to have been adapted from the Cypriot syllabary rather than from the Phoenician writing system.111 The existence of Cypriot elements in the SW script is perhaps unsurprising, given the prominent presence of Cypriot influence in the material culture of the Final Bronze and 108 Torres 2002, 379–85; Bartelheim 2007, 143–50; Neville 2007, 140–58. 109 cf. Barceló 1992, 263f; 1995, 561–4; González 2013, 351f. 110 J. Koch 2011, 113–63. 111 ibid. 168.

Brandherm CW3 CS4 19 i 2016.indd 204

27/05/2016 15:44:10

Brandherm

[ 205 ]

Early Iron Ages of south-western Iberia.112 However, one has to ask why elements of different Eastern Mediterranean writing systems were combined to create a new indigenous script in the first place, if the purpose of this creation was not to facilitate a better adaptation to the phonetic structure of the target language. One might of course hypothesize that originally the SW semi-syllabary may have been developed for a different language than that found on the stelae and slabs of our Group III, but there is little evidence to support such reasoning.113   This suggests that the development of the SW script may represent a conscious attempt by parts of the indigenous population to distance themselves culturally from users of the Phoenician aleph-beth. The latter group might not only have included new arrivals from the Eastern Mediterranean living in the coastal emporiae, but also members of the SWOC communities inhabiting their hinterland.114 Such a scenario challenges the notion that the new writing system would necessarily have been developed in the lower Guadalquivir basin. That the development of the new script should have taken place in an area intensely exposed to Phoenician influence from an early stage is no doubt a plausible proposition.115 However, there are other possibilities, and when looking for a Cypriot connection, it is striking that most of the relevant material is actually found elsewhere.116   That the introduction to SW Iberia of new ideas originating from the Eastern Mediterranean would have led to cultural conflicts and ideological struggles may also be borne out by the Early Iron Age sequence at the Castro dos Ratinhos (Moura, Baixo Alentejo), a hillfort occupying a strategic position on the western bank of the Guadiana. Between the late 9th and early 8th centuries BC we witness the construction of a building with religious functions at this site whose layout closely resembles that of Phoenician temples, which then perishes in a conflagration only a few decades later. In contrast to similar structures east of the Guadiana, which continued in use until the onset of the ‘crisis of the 6th century’, in this case no attempt was made at rebuilding.117 Given that one of the most distinctive features of SWOC society is the religious acculturation of its élites118, it is tempting to read this sequence of events as indicative of an ultimately unsuccessful attempt to extend the sway of the new ideology to communities on the western fringe of the original ‘tartesius ager’. 112 cf. Mederos 1996; Jiménez 2000; Schattner 2011. 113 It should be stressed, however, that if the origin of the new writing system was indeed to be sought in the region centred on the lower Guadalquivir, the available—but generally much later—onomastic evidence might be taken as indicative of a linguistic divide between that region and the main distribution area of SW inscriptions. Whereas early onomastic material from southern Portugal, but also from the Spanish Extremadura is mostly Indo-European, the same is clearly not the case with the lower Guadalquivir basin (Hoz 2010, 455–62 map 2.11). 114 Mederos & Ruiz 2001, 103–8. It should be remembered here that one of the likely reasons why the Cypriot syllabary continued in use on its native island alongside Phoenician and later Greek alphabetic scripts until the 3rd century BC—despite its much greater complexity in comparison to these—resides with its role as a cultural expression of a distinct Cypriot identity. 115 Hoz 2005; 2010, 485f. 116 Mederos 1996; Schattner 2011. 117 Prados 2010, 274–6; for comparator cases see Celestino 2008, 240–56; Fernández 2009, 153–6; more specifically Fernández & Rodríguez 2010 (for El Carambolo) and García & Blázquez 1996 (for Cástulo). 118 cf. Beba 2008, 127–35.

Brandherm CW3 CS4 19 i 2016.indd 205

27/05/2016 15:44:10

[ 206 ]

vi. Stelae, funerary practice, and group identities

]]

  Other readings of course are equally possible, and it must be stressed that our knowledge of the Early Iron Age settlement west of the Guadiana, despite considerable advances in recent years, remains far too limited to draw further conclusions based on this isolated case.119 Not only is more fieldwork required here. Also, a systematic and critical approach will be needed to escape the temptation of reading any odd idiosyncrasy in the settlement record as an expression of nativist rejectionism.   However, when it comes to the funerary record of the Early Iron Age in southern Portugal, the emerging pattern in terms of prevailing inhumation practice, funerary architecture, as well as inscribed slabs and stelae is quite unequivocal. The Early Iron Age population in the Alentejo and the Algarve deliberately seeks to distance itself from its SWOC neighbours and to showcase a group identity that provides self-reassertion by emulating the Monumenta Antichi of the region. Only some coastal enclaves remain exempt from this phenomenon, and instead undergo the same process of acculturation that generally characterizes the lands east of the Guadiana.120   What is still very much unclear, however, is the extent to which this emulation of the past draws on either real or imagined continuities. The largely invisible funerary practices of the Final Bronze Age do constitute a significant break with earlier traditions, and it does seem unquestionable that in this case, too, an influx of new ideas and apparently also people had a significant role to play.121 The re-emergence of much earlier types of funerary architecture and monuments in the Early Iron Age of southern Portugal by itself does not provide any more evidence for unbroken ethnic or linguistic continuity than, for example, the widespread revival of long disused architectural formulae across much of Europe during the Renaissance. It is likely, therefore, that the jury will have to remain out on this issue for some time to come. Of phantoms and unhelpful labels—concluding remarks In the present author’s view there is little point in trying to match the group identities behind the phenomena outlined above to ethnic labels recorded by Greek and Roman authors. While exercises of this type have long been a popular pursuit, not only in Iberian Iron Age archaeology, and while there can be little doubt that much ink will continue to be spilled on this subject, it contributes very little to a better understanding of the historical process and its socio-economic underpinnings. As previously pointed out by Correia122, identifying the factors driving economic and societal change in the Early Iron Age of south-western Iberia is considerably more important than attaching specific names to the protagonists of the process. This holds true in particular for the ‘Tartes(s)os’ nomenclature, whose exact 119 cf. Pérez 1996; Arruda 2007a; Antunes et al. 2012; Mataloto 2013; Serra 2014. 120 How to interpret the formation of these enclaves is the subject of an ongoing discussion; cf. Gamito 1992, 332f; Torres 2005, 205–8; Almagro & Torres 2009, 119–29; Arruda 2007b, 118–20; 2013, 215–18. 121 Brandherm 2013b, 148–53. 122 Correia 1999, 709f.

Brandherm CW3 CS4 19 i 2016.indd 206

27/05/2016 15:44:10

Brandherm

[ 207 ]

meaning and usage at the time we have little means of fathoming.123   Another example that illustrates the problematic nature of simplistic equations between specific group identities gleaned from the archaeological record and ethnic labels extracted from textual sources is the case of the Kynetes, first mentioned by Herodotus.124 Ever since Schlichthorst125 published his seminal work on the subject more than two centuries ago, the general consensus among scholars has been to locate their lands in the Algarve and southern Alentejo.126 An alternative proposal recently put forward by Alarcão127 and since embraced by Almagro Gorbea128 favours their location in an area centred on the middle Guadiana basin. Under the first scenario, trying to align the archaeological evidence with the sparse information provided by the textual sources will lead to an image of the Kynetes as a population actively rejecting SWOC forms of cultural expression, whereas accepting Alarcão’s alternative reading as a premise, the same exercise will have them as one of its main exponents.129   The situation is further complicated by the fact that Iberian Iron Age archaeology is still struggling to rid itself of the legacy of Schulten’s both mistaken and misleading perception of the Algarve and southern Alentejo as a constituent part of his entirely fictitious ‘Tartes(s)ian Empire’.130 Ironically, Schulten’s construct was based on the premiss that the Kynetes, as the Iron Age people inhabiting these parts, used the ‘Tartes(s)ian’ language and writing system.131 As argued above, the exact opposite might have been the case, in that the SW script may well have been conceived as an explicit expression of a non-SWOC identity.   The unfortunate designation of Group  III monuments, the writing system associated with them, and the language of the relevant inscriptions as ‘Tartes(s)ian’ has recently gained further prominence as a consequence of J. Koch’s132 linguistic work. It is certainly beyond 123 cf. Álvarez 2009, 92f. 124 Herodotus II.33; IV.49. 125 Schlichthorst 1793. 126 Hübner 1901; Gamito, 1992, 330f; 1993, 129–32; J. Koch 2013, 14. 127 Alarcão 2001, 335–8. 128 Almagro 2008, 93f. 129 It should be noted that the sources lending support to the scenario favoured by Alarc ão and Almagro refer to the situation during the Roman military campaigns of the 2nd century BC. They are thus separated from the state of affairs Herodotus describes by three to five centuries of dynamic development. In particular, the sudden appearance during the 5th and 4th centuries BC, on the middle Guadiana of a funerary architecture strongly reminiscent of that found in the Early Iron Age cemeteries of the upper Mira-Sado region suggests a potential shift in group identities and/or influx of new population elements from the ABAC area into the middle Guadiana basin, following the disintegration of the SWOC during the 6th century BC (cf. Jiménez 2004; 2008; 2012; Zarzalejos & López 2005, 826–30). The most likely explanation for the apparent inconsistency between Schlichthorst’s ­localization of the Kynetes and the scenario proposed by Alarcão then would appear to be that the latter reflects changes in the political and ethnographic landscape of south-western Iberia occurring only in the wake of the ‘crisis of the 6th century’. Previous authors have attributed the reconfiguration of settlement patterns that affects the middle to lower Guadiana during the 5th century to an influx of new population elements from the Meseta (Pérez 1996, 107–10), but the process may well have been more complex than this. 130 cf. Untermann 1995, 245f n 10; Celestino 2008, 93f. 131 Schulten 1950, 123. 132 J. Koch 2011; 2013.

Brandherm CW3 CS4 19 i 2016.indd 207

27/05/2016 15:44:10

[ 208 ]

vi. Stelae, funerary practice, and group identities

]]

the scope of the present contribution—and would definitely stretch the present author’s expertise—to try and engage directly with the linguistic argument, but regardless of its merits or potential fallacies we would still maintain that a designation of the relevant monuments, of the writing system documented on them, and of the language codified in that writing system as ‘Tartes(s)ian’ carries an undue burden of preconceived ideas, born from the wishful thinking of previous generations of scholars.133   The continued obsession among many scholars with all things ‘Tartes(s)ian’, and the resulting tendency to try and apply the preconceived concepts attached to this terminology to the archaeological record have led to strange and myopic distortions. This holds particularly true in cases where the scope of the ‘Tartes(s)os’ label has been extended to include pre-Orientalizing indigenous communities of the Final Bronze and Early Iron Ages. Stepping back from this fixation, and taking an unbiased look at the big picture of the Early Iron Age in southern Iberia, what we observe is a fairly uniform indigenous material culture—regional differences in some pottery types notwithstanding—that stretches from the lower Guadalquivir to the eastern extents of the southern Meseta134, and beyond to the lower Segura.135 This uniformity is particularly pervasive in the funerary record, where regional differences are mostly due to the absence in the east of ‘élite’ burials similar to those found at some of the south-western sites. Rather than a priori ethnic or linguistic boundaries, it is the increasing economic and social divides within Early Iron Age society in the lower Guadalquivir basin, visible particularly in the appearance of these ‘élite’ burials, which separate that region from more easterly parts of this original Early Iron Age koiné.   The latter’s final disintegration only comes with the process of restructuring that grips most of southern Iberia during the 6th century BC, leading to the formation of what we have become used to calling the ‘Iberian’ culture in the East, and the post-orientalizing social formations conventionally labelled ‘Turdetanian’ in the south-west.136 In any case, most of the lands west of the Guadiana never formed part of that original southern Early Iron Age koiné to begin with, and it is not without irony that under the influence of a very specific— and by no means unassailable—reading of the patchy textual sources, archaeologists have chosen to lump together the western parts of that koiné and coeval—but rather distinct— phenomena in southern Portugal and the Spanish Extremadura under the ‘Tartes(s)os’ label. The result of all this is a pronounced dichotomy between what has been inferred from— mostly much later—textual sources, and what the archaeological evidence tells us when 133 The highly questionable nature of the ‘Tartes(s)os’ nomenclature for the inscribed slabs and stelae of our Group III and for the corresponding writing system has previously been highlighted by M. Koch (2003, 702f). The underlying problem is now also recognized by J. Koch (2013, 18f ), who paradoxically still sticks to this terminology. Needless to say, occasional attempts to extend the same nomenclature to the Group II monuments of the Final Bronze Age cause even more serious problems. The occurrence of some late exponents of this latter group in the lower Guadalquivir basin at the time of the earliest biblical ‘Taršiš’ references—even granting that those might indeed refer to the general area—can hardly justify the application of this label to an entire class of monuments, most of which are considerably earlier and have a much more northerly distribution (cf. Brandherm 2008, 484). 134 Ros Sala 1990; Hernández & Gil 2001/02; 2004; Zarzalejos & López 2005, 816–36. 135 González 2002. 136 cf. the parallel transformation from ‘Phoenician’ to ‘Punic’.

Brandherm CW3 CS4 19 i 2016.indd 208

27/05/2016 15:44:10

Brandherm

[ 209 ]

taken at face value.   Other authors have previously pointed out that most surviving SW inscriptions come from an area whose material culture shows little affinity to that of the so-called ‘Tartes(s)ian’ lands of the lower Guadalquivir basin.137 The point highlighted here is that the milieu of these inscriptions should be perceived not simply as ‘non-Tartes(s)ian’, but—if that term is to retain any meaning at all—as outright ‘anti-Tartes(s)ian’ in nature. At best then, ‘Tartes(s)ian’ might be used as a term to label the specific symbolic language of a social ‘élite’ that benefited from the conversion of what during the Final Bronze Age was still very much a subsistence economy, to an export-oriented economic system based on large-scale silver mining in the Early Iron Age.138 At worst, it is employed as a pseudo-ethnic label with linguistic connotations that does nothing but muddy the waters of Iberian Iron Age archaeology.   Clearly, we are profoundly stuck in a terminological mess, and the view taken in this paper is that the most sensible way to cut this Gordian knot would be to abandon the ‘Tartes(s)os’ label altogether, and instead resort to a more neutral terminology, be it ‘SW Iron Age I’ (SWIA I)139 to refer to the relevant chronological period, or ‘SW Orientalizing Complex’ (SWOC) when referring to the material culture of communities or sectors of Early Iron Age society that following the establishment of a permanent Phoenician presence in southern Iberia underwent a pronounced process of acculturation. These terms of course have far less ring and much less imaginative appeal to them than the more colourful ‘Tartes(s) os’ or ‘Tartes(s)ian’, but this is precisely why they are much better suited for the purposes of scholarly inquiry.140   One might argue that the names attached to specific archaeological phenomena should be of no importance, and indeed Karl141, when discussing the suitability of the ‘Celtic’ label in an earlier volume of this series, adopted very much a laissez-faire approach, calling us to give up terminological skirmishes and instead focus on achieving a better understanding of the historical process. While the author would wholeheartedly agree with his call to focus more on structures and processes, the problem is that skewed terminologies such as that discussed here very much get in the way of precisely this, forcing the expert into a permanent state of doublethink, and perpetually misleading the uninitiated. Its only true beneficiaries then would appear to be writers of historical fiction.142   Adding a linguistic component to what is already an unfortunate mess can only make matters worse, particularly since the main language used by the SWOC communities of the 137 Hoz 2010, 354–6; Correia 2014, 84; Valério 2014, 441 n 5. 138 cf. Celestino 2008, 141–4. 139 Spanish: Hierro del Sudoeste I; cf. Ferrer & Prados 2013, 404. 140 Despite being less loaded with this particular baggage, but not necessarily more correct for that, other labels carrying connotations of ethnicity—‘Kynetian’ might sound particularly tempting in this case— offer little advantage here. This is not only because of the competing hypotheses concerning their localization put forward by Schlichthorst (1793) and Alarcão (2001). As the recent debate concerning the use of the term ‘Celtic’ has amply demonstrated (Sims-Williams 1998; James 1999; Collis 2003), such labels always come with their particular set of problems, and rarely have anything to contribute to a better understanding of historic processes in what essentially remains a prehistoric setting. 141 Karl 2010, 62f. 142 cf. Alvar 2010.

Brandherm CW3 CS4 19 i 2016.indd 209

27/05/2016 15:44:10

[ 210 ]

vi. Stelae, funerary practice, and group identities

]]

lower Guadalquivir basin—the ‘Tartes(s)ian’ core area of conventional parlance—might still turn out to bear no relationship whatsoever to the language of the inscriptions found on the Early Iron Age stelae of southern Portugal. Acknowledgements The author would like to thank John Koch for his persevering patience while waiting for this contribution to materialize, to Catriona Gibson for creating endless draft distribution maps from the AEMA database, and to both Virgílio Correia and Daniel Neumann for their kind help in chasing down hard-to-come-by literature.

BIBLIOGRAPHY Abad Casal, L. 1979 ‘Consideraciones en torno a Tartessos y el orígen de la cultura ibérica’, Archivo Español de Arqueología 52, 175–93. Alarcão, J. de 2001 ‘Novas perspectivas sobre os Lusitanos (e outros mundos)’, Revista Portuguesa de Arqueologia 4 (2), 293–349. Albuquerque, P. 2013a ‘Tartessos e tartéssios, de Estesícoro a Éforo’, Arqueologia em Portugal—150 Anos, ed. J. Morais Arnaud, A. Martins, & C. Neves, 633–9. Lisboa, Associação dos Arqueólogos Portugueses. Albuquerque, P. 2013b ‘Alguns pontos de interrogação sobre identidade(s) e território(s) em Tartessos’, Spal 22, 47–60. Almagro, M. 1966 Las estelas decoradas del Suroeste peninsular. Biblioteca Praehistorica Hispana 8. Madrid, CSIC—Universidad de Madrid. Almagro Gorbea, M. 2008 ‘Medellín-Conisturgis: reinterpretación geográfica del Suroeste de Iberia’, Boletim da Sociedade de Geografia de Lisboa 126, 84–115. Almagro Gorbea, M. & M. Torres Ortiz 2009 ‘La colonización de la costa atlántica de Portugal:¿fenicios o tartesios?’, Palaeohispánica 9, 113–42. Alvar Ezquerra, J. 2010 Tartesos, un reino soñado. Madrid, La Esfera del Libro. Álvarez Martí-Aguilar, M. 2005 Tarteso. La construcción de un mito en la historiografía española. Málaga, Centro de Ediciones de la Diputación de Málaga. Álvarez Martí-Aguilar, M. 2009 ‘Identidad y etnia en Tartesos’, Arqueología Espacial 27, 79–111. Álvarez Martí-Aguilar, M. 2013 ‘Defindiendo Tarteso:

Brandherm CW3 CS4 19 i 2016.indd 210

indígenas y fenicios’, Tarteso. El emporio del metal, eds J. M. Campos & J. Alvar, 223–46. Córdoba, Almuzara. Alves, C., C. Costeira, S. Estrela, E. Porfírio, M. Serra, A. M. M. Soares, & M. Moreno-Garcí 2010 ‘Hipogeos funerarios del Bronce Pleno en Torre Velha 3 (Serpa, Portugal). ¡¿El Sureste en el Suroeste?!’, Zephyrus 66, 133–53. Alves, L. B. & M. Reis 2011 ‘Memoriais de pedra, símbolos de identidade. Duas novas peças excultóricas de Cervos (Montalegre, Vila Real)’, Estelas e estátuas-menires da Pré à Proto-história. Actas das IV Jornadas Raianas (Sabugal, 2009), ed. R. Vilaça, 187–216. Sabugal, Fundação para a Ciência e a Tecnologia. Antunes, A. S., M. de Deus, A. M. M. Soares, F. Santos, L. Arêz, L., J. Dewulf, L. Baptista, & L. Oliveira, 2012 ‘Povoados abertos do Bronze Final no médio Guadiana’, Sidereum Ana II. El río Guadiana en el Bronce Final, ed. J. Jiménez Ávila, 277–308. Anejos de Archivo Español de Arqueología 62. Mérida, CSIC. Arruda, A. M. 2000 ‘Práticas e rituais funerários no Sul de Portugal durante a Proto-História’, Actas do 3o Congresso de Arqueología Peninsular V. Proto-história da Península Ibérica, ed. V. Oliveira Jorge, 101–8. Porto, ADECAP. Arruda, A. M. 2004 ‘Necrópoles proto-históricas do sul de Portugal. O mundo oriental e orientalizante’, El mundo funerario. Actas del III Seminario Internacional sobre Temas Fenicios, Guadamar del Segura, 3 a 5 de mayo de 2002. Homenaje al Prof. D. Manuel Pellicer Catalán,

27/05/2016 15:44:10

Brandherm

ed. A. González Prats, 457–94. Alicante, Instituto Alicantino Juan Gil-Albert. Arruda, A. M. 2007a ‘A Idade do Ferro do Sul de Portugal. Estado da investigação’, Madrider Mitteilungen 48, 114–39. Arruda, A. M. 2007b ‘A Idade do Ferro no Algarve. Velhos dados (e outros mais recentes) e novas histórias’, Xelb 7, 115–30. Arruda, A. M. 2009 ‘Os espaços funerários e a construção das novas entidades sociais e culturais do extremo ocidente europeo (1° milénio a. n. e.)’, Estudos Arqueológicos de Oeiras 17, 513–20. Arruda, A. M. 2013 ‘De que falamos quando falamos de Tartesso?’, Tarteso. El emporio del metal, eds J. M. Campos & J. Alvar, 211–22. Córdoba, Almuzara. Arruda, A. M., J. Covaneiro, & S. Cavaco 2008 ‘A necrópole da Idade do Ferro do Convento da Graça, Tavira’, Xelb 8, 117–35. Aubet, M. E. 1975 La necrópolis de Setefilla en Lora del Río, Sevilla. Barcelona, CSIC—Universidad de Barcelona. Aubet, M. E. 1976–1978 ‘La cerámica a torno de la Cruz del Negro (Carmona, Sevilla)’, Ampurias 38–40, 267–88. Aubet, M. E. 1978 La necrópolis de Setefilla en Lora del Río, Sevilla (Túmulo B). Barcelona, CSIC—Universidad de Barcelona. Aubet, M. E. 1994 Tiro y las colonias fenicias de occidente, 2nd edition. Barcelona, Crítica. Aubet, M. E. 1997 ‘A propósito de una vieja estela’, Saguntum 30, 163–72. Balsera, V., J. Bernabeu Aubán, M. Costa Caramé, P. Díaz del Río, L. García Sanjuan, & S. Pardo 2015 ‘The radiocarbon chronology of southern Spain’s late prehistory (5600–1000 cal BC): a comparative review’, Oxford Journal of Archaeology 34, 139–56. Banha, C., A. M. Veiga, & S. Ferro 2009 ‘A estátuamenir de Corgas (Donas, Fundão). Contribute para o estudo da Idade do Bronze na Beira Interior’, AÇAFA On-Line 2, 1–16. http://www.altotejo.org/ acafa/docsN2/A_estatua-menir_de_Corgas.pdf accessed: 14.03.2014 Barceló, J. A. 1991 ‘El Bronce del Suroeste y la cronología de las estelas alentejanas’, Arqueologia 21, 15–24. Barceló, J. A. 1992 ‘Una interpretación socioeconómica del Bronce Final en el Sudoeste de la Península Ibérica’, Trabajos de Prehistoria 49, 259–75. Barceló, J. A. 1995 ‘Sociedad y economía en el Bronce Final tartesico’, Tartessos 25 años después. 1968–1993, Jérez de la Frontera: Actas del Congreso Conmemorativo del V Symposium Internacional de Prehistoria Peninsular, 561–89. Jérez de la Frontera, Ayuntamiento de Jérez de la Frontera.

Brandherm CW3 CS4 19 i 2016.indd 211

[ 211 ]

Barros, P. 2010 ‘Mértola entre os séculos VI e III a. C.’, Mainake 32 (1), 417–36. Barros, P., G. Branco, C. Duarte, & J. Correia 2003 ‘A cista dos Gregórios (Silves)’, Xelb 5, 41–52. Barros, P., S. Melro, & D. Gonçalves 2013 ‘A necrópole da Idade do Ferro da Abóbada (Almodôvar)’, Actas del VI Encuentro de Arqueología del Suroeste Peninsular, eds F. J. Jiménez Ávila, M. Bustamante Alvarez & M. García Cabezas, 1157–77. Villafranca de los Barros, Ayuntamiento de Villafranca de los Barros. Bartelheim, M. 2007 Die Rolle der Metallurgie in vorgeschichtlichen Gesellschaften. Forschungen zur Archäometrie und Altertumswissenschaft 2. Rahden, Marie Leidorf. Beba, S. 2008 Die „tartessischen“ Fürstengräber in Andalusien. Bochumer Forschungen zur ur- und frühgeschichtlichen Archäologie 1. Rahden, Marie Leidorf. Beirão, C. de Mello 1986 Une civilisation protohistorique du sud du Portugal (1er Âge du Fer). Paris, Editions de Boccard. Beirão, C. de Mello 1990 ‘Epigrafia da I Idade do Ferro do sudoeste da Península Ibérica. Novos dados arqueológicos’, Estudos Orientais 1, 107–18. Beirão, C. de Mello & M. Varela Gomes 1988 ‘A estela epigrafada do Pardeiro, S. Martinho das Amoreiras (Odemira, Beja)’, Veleia 5, 115–24. Belén Deamos, M. 2001 ‘La incineración en las necrópolis tartésicas’, Arqueología funeraria: las necrópolis de incineración , eds R. García Huerta & J. Morales Hervás, 37–78. Cuenca, Ediciones de la Universidad de Castilla-La Mancha. Bendala Galán, M. 2013 ‘La génesis de Tarteso en la etapa “precolonial” del segundo milenio: notas para una discusión’, Tarteso. El emporio del metal, eds J. M. Campos & J. Alvar, 123–36. Córdoba, Almuzara. Bettencourt, A. 2013 ‘O Bronze Final no Noroeste português. Uma rede complexa de lugares’, Estudos Arqueológicos de Oeiras 20, 157–72. Brandherm, D. 2007 Las espadas del Bronce Final en la Península Ibérica y Baleares. Prähistorische Bronzefunde IV, 16. Stuttgart, Steiner. Brandherm, D. 2008 ‘The warriors’ new headgear’, Antiquity 82, 480–4. Brandherm, D. 2013a ‘Mediterranes, Atlantisches und Kontinentales in der bronze- und ältereisenzeitlichen Stelenkunst der Iberischen Halbinsel’, Petasos. Festschrift für Hans Lohmann, eds G. Kalaitzoglou & G. Lüdorf, 131–48. Paderborn, Wilhelm Fink— Ferdinand Schöningh. Brandherm, D. 2013b ‘Westward Ho? Sword-bearers and all the rest of it…’, Celtic from the West 2.

27/05/2016 15:44:10

[ 212 ]

vi. Stelae, funerary practice, and group identities

Rethinking the Bronze Age and the Arrival of IndoEuropean in Atlantic Europe, eds J. T. Koch & B. Cunliffe, 147–55. Oxford, Oxbow Books. Cardoso, J. L. 2004 ‘Uma tumulação do final do Bronze Final/inícios da Idade do Ferro no sul de Portugal: a tholos do Cerro do Malhanito (Alcoutim)’, O passado em cena: narrativas e fragmentos. Miscelánea oferecida a Jorge de Alarcão, M. C. Lopes & R. Vilaça, 193–223. Coimbra—Porto, CAUCP. Cardoso, J. L. & A. Gradim 2003 ‘A cista megalítica do Cerro do Malhão (Alcoutim)’, Revista Portuguesa de Arqueologia 6 (2), 167–79. Cardoso, J. L. & A. Gradim 2006 ‘A necrópole da I Idade do Ferro de Cabeço da Vaca 1 (Alcoutim)’, Xelb 6, 201–26. Cardoso, J. L. & A. Gradim 2008 ‘O núcleo II da necrópole da Idade do Ferro de Cabeço da Vaca 1 (Alcoutim)’, Xelb 8, 103–15. Cardoso, J. L. & A. Gradim 2010 ‘A anta do Malhão (Alcoutim) e o “Horizonte de Ferradeira”’, Xelb 10, 55–72. Celestino Pérez, S. 2001 Estelas de guerrero y estelas diademadas. La precolonización y formación del mundo tartésico. Barcelona, Bellaterra. Celestino Pérez, S. 2008 ‘Tartessos’, De Iberia a Hispania, ed. F. Gracia Alonso, 93–345. Madrid, Ariel. Celestino Pérez, S. & J. A. Salgado Carmona 2011 ‘Nuevas metodologías para la distribución espacial de las estelas del Oeste peninsular’, Estelas e estátuasmenires da Pré à Proto-história. Actas das IV Jornadas Raianas (Sabugal, 2009), ed. R. Vilaça, 417–48. Sabugal, Fundação para a Ciência e a Tecnologia. Coelho, L. 1971 ‘Inscrições da necrópole protohistórica da Herdade do Pêgo (Ourique)’, O Arqueólogo Português 5, 167–80. Coelho, L. 1975 ‘Nueva estela insculturada proveniente del Baixo Alentejo (Ervidel, Portugal)’, Trabajos de Prehistoria 32, 195–7. Collis, J. R. 2003 The Celts: Origins, Myths, Inventions. Stroud, History Press. Correa, J. A. 1989 ‘Posibles antropónimos en las inscripciones en escritura del S.O. (o tartesia)’, Veleia 6, 243–52. Correa, J. A. 1996 ‘La epigrafía del Sudoeste: estado de la cuestión’, La Hispania Prerromana. Actas del VI Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica (Coimbra, 13–15 de octubre de 1994), eds F. Villar & J. d’Encarnação, 65–75. Salamanca, Ediciones Universidad de Salamanca. Correia, V.  H. 1993 ‘As necrópoles da Idade do Ferro do Sul de Portugal: arquitectura e rituais’, Trabalhos

Brandherm CW3 CS4 19 i 2016.indd 212

]]

de Antropologia e Etnologia 33 (3/4), 351–75. Correia, V. H. 1997 ‘As necrópoles algarvías da I Idade do Ferro e a escrita do Sudoeste’, Noventa séculos entre a serra e o mar, eds M. F. Barata & R. Parreira, 265–79. Lisboa, IPPA. Correia, V. H. 1999 ‘Algumas considerações sobre os centros de poder na proto-história do Sul de Portugal’, Revista de Guimarães, Volume Especial— Actas do Congresso de Proto-História Europeia, Vol. II, 699–714. Correia, V. H. 2014 ‘A escrita do Sudoeste da Península Ibérica: velhos dados, novas teorias e a sua importância para o estudo das antigas culturas hispânicas’, Portugalia 35 (New Series), 77–93. Correia, V. H. & R. Parreira 2002 Cola. Circuito Arqueológico. Roteiros de Arqueologia Portuguesa 8. Lisboa, IPPA. Cross, F. M. 1972 ‘An interpretation of the Nora stone’, Bulletin of the American Schools of Oriental Research 208, 13–19. Cruz Andreotti, G. 2010 ‘Tarteso-Turdetania o la reconstrucción de un mito identitario’, El Carambolo. 50 años de un tesoro, eds M. L de la Bandera Romero & E. Ferrer Albelda, 17–52. Sevilla, Secretario de Publicaciones de la Universidad de Sevilla. Dias, M. M. A., C. de Mello Beirão, & L. Coelho 1970 ‘Duas necrópoles da idade do ferro no BaixoAlentejo, Ourique (notícia preliminar)’, O Arqueólogo Português 4 (3rd Series), 175–219. Dias, M. M. A. & L. Coelho 1983 ‘Objectos arqueológicos dum túmulo de incineração da necrópole proto-histórica da Herdade da Favela Nova’, O Arqueólogo Português 1 (4th Series), 197–206. Díaz-Guardamino Uribe, M. 2010 Las estelas decoradas en la prehistoria de la Península Ibérica. Madrid, Universidad Complutense. Díaz-Guardamino Uribe, M. 2012 ‘Estelas decoradas del Bronce Final en la Península Ibérica: datos para su articulación cronológica’, Sidereum Ana II. El río Guadiana en el Bronce Final, ed. J. Jiménez Ávila, 289–415. Anejos de Archivo Español de Arqueología 62. Mérida, CSIC. Enríquez Navascuéz, J. J. & M. J. Carrasco Martín 1995 ‘Las necrópolis de cistas de “Las Arquetas” (Frenegal de la Sierra, Badajoz) y otros restos de necrópolis de cistas en las estribaciones occidentales de la Sierra Morena extremeña’, SPAL 4, 101–130. Eska, J. F. 2014 ‘Comments on John T. Koch’s Tartessian-as-Celtic enterprise’, Journal of IndoEuropean Studies 42, 428–38. Esteve Guerrero, M. 1954/55 ‘Bornos’, Noticiario Arqueológico Hispánico 3/4, 253.

27/05/2016 15:44:10

Brandherm

Fernández Flores, A. 2009 ‘Sobre Tarteso y la colonización oriental. Reflexiones a partir de las nuevas intervenciones en El Carambolo’, Tendencias y aplicaciones en la investigación arqueológica. Encuentros de Jóvenes Investigadores 2006–2007, Universidad de Sevilla, eds F. J. García Fernández & O. Rodríguez Gutiérrez, 145–60. Sevilla, Fénix Editores. Fernández Flores, A. & A. Rodríguez Azogue 2010 ‘El Carambolo, secuencia cronocultural del yacimiento. Síntesis de las intervenciones 2002–2005, El Carambolo. 50 años de un tesoro, eds M. L de la Bandera Romero & E. Ferrer Albelda, 203–70. Sevilla, Secretario de Publicaciones de la Universidad de Sevilla. Ferreira, O. da Veiga & F. de Almeida 1971 ‘A necrópole do Bronze meridional português da Herdade do Peral (Évora)’, Madrider Mitteilungen 12, 115–22. Ferrer Albelda, E. & E. Prados Pérez 2013 ‘Tarteso, de ciudad a imperio (o sobre la creación de identidades ficticias)’, Tarteso. El emporio del metal, eds J. M. Campos & J. Alvar, 395–414. Córdoba, Almuzara. Frías, C. & J. A. Pérez Macías 1989 ‘La necrópolis de cistas en la Parrita, Nerva. Huelva y los indicios de la metalurgia de la plata en las minas de Riotinto’, Cuadernos del Suroeste 1, 11–21. Gamito, T. J. 1992 ‘Paleoetnologia do centro e sul de Portugal’, Complutum 2/3, 329–37. Gamito, T. J. 1993 ‘The internal and external dynamics of the development and collapse of Tartessos’, Lengua y cultura en la Hispania prerromana. Actas del V Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, Colonia, 25–28 de noviembre de 1989, eds J. Untermann & F. Villar, 127–42. Salamanca, Universidad de Salamanca. García-Gelabert, M. P. & J. M. Blázquez Martínez 1996 ‘Relación entre el proceso histórico: Tartessos / colonización fenicia y la Alta Andalucía’, Homenaje al profesor Manuel Fernández Miranda, Vol. I, eds M. A. Querol  & T. Chapa, 327–38. Complutum Extra 6. Madrid, Universidad Complutense. García Sanjuán, L. 1998 ‘La Traviesa. Análisis del registro funerario de una comunidad de la Edad del Bronce’, La Traviesa. Ritual funerario y jerarquización social en una comunidad de la Edad del Bronce de Sierra Morena occidental, ed. L. García Sanjuán, 101–89. Spal Monografías 1. Sevilla, Universidad de Sevilla. García Sanjuán, L. 2005 ‘Grandes piedras viejas, memoria y pasado. Reutilizaciones del dolmen de palacio III (Almadén de la Plata, Sevilla) durante la Edad del Hierro’, El periodo orientalizante. Actas del III Simposio Internacional de Arqueología de Mérida, Protohistoria del Mediterráneo Occidental, Vol. I, eds F. J. Jiménez Ávila

Brandherm CW3 CS4 19 i 2016.indd 213

[ 213 ]

& S. Celestino Pérez, 595–604. Anejos de Archivo Español de Arqueología 35. Mérida, CSIC. Gómez Toscano, F. 1997 El Final de la Edad del Bronce entre el Guadiana y el Guadalquivir. Huelva, Universidad de Huelva. Gómez Toscano, F. J., G. Álvarez García, & F. Borja Barrera 1992 ‘Depósito funerario del Bronce en el travertino de Alájar (Huelva). La cavidad Al-24Geos’, Cuadernos del Suroeste 3, 43–55. González Prats, A. 2002 La necrópolis de cremación de Les Moreres (Crevillente, Alicante, España) (s. IX–VII AC). Alicante, Universidad de Alicante. González Wagner, C. 2013 ‘Tartessos and the orientalizing elites’, The Prehistory of Iberia. Debating early social stratification and the state, eds M. Cruz Berrocal, L. García Sanjuán & A. Gilman, 337–56. London, Routledge. Harrison, R. 2004 Symbols and Warriors. Images of the European Bronze Age. Bristol, Western Academic & Specialist Press. Hauptman Wahlgren, K. 2002 Bilder av betydelse. Hällristningar och bronsålderslandskap i nordöstra Östergötland. Stockholm Studies in Archaeology 23. Lindome, Bricoleur Press. Henriques, F. J. R., A. M. M. Soares, T. F. A. António, F. Curate, P. Valério, & S. P. Rosa 2013 ‘O tholos Centirã 2 (Brinches, Serpa)—construtores e utilizadores; práticas funerárias e cronologias’, Actas del VI Encuentro de Arqueología del Suroeste Peninsular, eds F. J. Jiménez Ávila, M. Bustamante Alvarez & M. García Cabezas, 319–355. Villafranca de los Barros, Ayuntamiento de Villafranca de los Barros. Hernández Carrion, E. & F. Gil González 2001/02 ‘Encachados tumulares del Bronce Final / Hierro Antiguo en la necrópolis del Collado y Pinar de Santa Ana (Jumilla, Murcia)’, Anales de Prehistoria y Arqueología 16/17, 73–94. Hernández Carrion, E. & F. Gil González 2004 ‘La necrópolis del Bronce Final del Collado y Pinar de Santa Ana de Jumilla (Murcia)’, La Edad del Bronce en tierras valencianas y zonas limítrofes, eds L. Hernández Alcara & M. Hernández Pérez, 441–54. Alicante, Instituto Alicantino Juan Gil-Albert. Hoz Bravo, J. de 2005 ‘La recepción de la escritura en Hispania como fenómeno orientalizante’, El Periodo orientalizante. Actas del III Simposio Internacional de Arqueología de Mérida: Protohistória del Mediterráneo Occidental, eds S. Celestino Pérez & J. Jiménez Ávila, 363–81. Anejos de Anuario Español de Arqueología 35. Madrid, CSIC. Hoz Bravo, J. de 2010 Historia lingüistica de la Península Ibérica en la antigüedad I. Preliminares y mundo meridional

27/05/2016 15:44:10

[ 214 ]

vi. Stelae, funerary practice, and group identities

prerromano. Madrid, CSIC. Hübner, E. 1901 ‘Cynetes’, Paulys Realencyclopädie der classischen Altertumswissenschaft, Vol. IV,2, ed. G. Wissowa, 1906–1908. Stuttgart, J. B. Metzler. James, S. 1999 The Atlantic Celts: Ancient People or Modern Invention? London, British Museum Press. Jiménez Ávila, J. 2000 ‘Timiaterios chipriotas de bronce: centros de producción occidentales’, Actas del IV Congreso Internacional de Estudios Fenicios y Púnicos—Cádiz, 2 al 6 de octubre de 1995, eds M. Barthélemy & M. E. Aubet Semmler, 1581–94. Cádiz, Universidad de Cádiz. Jiménez Ávila, J. 2004 ‘La necrópolis de El Jardal (Herrera del Duque, Badajoz). Elementos para una revisión cronológica de las necrópolis de la 1ª Edad del Hierro del Sur de Portugal’, Actas do II Encontro de Arqueología do Sudoeste Peninsular, Faro 7 e 8 de Novembro de 1996. 105–14. Faro, Universidade do Algarve. Jiménez Ávila, J. 2008 ‘El final del Hierro Antiguo en el Guadiana Medio’, Sidereum Ana I. El río Guadiana en época post-orientalizante. ed. J. Jiménez Ávila, 101–34. Anejos de Archivo Español de Arqueología 46. Madrid, CSIC. Kalb, P. 1994 ‘Reflexões sobre a utilização de necrópoles megalíticas na Idade do Bronze’, Actas do Seminário “O Megalitismo no Centro de Portugal”, 415–26. Estudos Pré-Históricos 2. Viseu, Centro de Estudos Pré-Históricos da Beira-Alta. Kalb, P. 1995 ‘Alpiarça’, A Idade do Bronze em Portugal: discursos de poder, 90–3. Lisboa, IPM. Kalb, P. & M. Höck 1980 ‘Cabeço da Bruxa, Alpiarça (Distrikt Santarém). Vorbericht über die Grabung im Januar und Februar 1979’, Madrider Mitteilungen 21, 91–104. Karl, R. 2010 ‘The Celts from everywhere and nowhere: a re-evaluation of the origins of the Celts and the emergence of Celtic cultures’, Celtic from the West. Alternative perspectives from archaeology, genetics, language and literature, eds B. Cunliffe & J. T. Koch, 39–64. Oxford, Oxbow Books. Koch, J. T. 2011 Tartessian 2: The Inscription of Mesas do Castelinho, ro and the verbal complex, preliminaries to historical phonology. Aberystwyth, Canolfan Uwchefrydiau. Koch, J. T. 2013 Tartessian: Celtic in the south-west at the dawn of history, 2nd edition. Celtic Studies Publications 13. Aberystwyth, CSP-Cymru Cyf. Koch, J. T. 2014a ‘On the debate over the classification of the language of the south-western (SW) inscriptions, also known as Tartessian’, Journal of Indo-European Studies 42, 335–427.

Brandherm CW3 CS4 19 i 2016.indd 214

]]

Koch, J. T. 2014b ‘A decipherment interrupted: proceeding from Valério, Eska, and Prósper’, Journal of Indo-European Studies 42, 487–524. Koch, M. 1984 Tarschisch und Hispanien. Madrider Forschungen 14. Berlin, Walter de Gruyter. Koch, M. 2003 ‘Review of “Monumenta Linguarum Hispanicarum. Unter Mitwirkung von Dagmar Wodtko hrsg. von Jürgen Untermann. Band IV: Die tartessischen, keltiberischen und lusitanischen Inschriften”’, Gnomon 75, 701–7. Koch, M. 2005 ‘Zur spanischen Ausgabe von „Tarschisch und Hispanien“’, Madrider Mitteilungen 46, 22–45. Koch, M. 2010 ‘Von Tarschisch bis nach Indien. Die Integration der Iberischen Halbinsel in das Weltherrschaftsschema des Alten Orients und seiner Erben’, Palaeohispanica 10, 567–78. Krueger, M. & D. Brandherm forthcoming ‘Early Iron Age pottery in south-western Iberia: archaeometry and chronology’, Proceedings of the XVII UISPP World Congress—Session B34: Archaeometry approaches regarding the study of networks of trade in raw materials and technological innovations in prehistory and protohistory, eds J. C. Baptista, D. Delfino, & P. Piccardo. Oxford, Archaeopress. Lorrio, A. J. 2008 Qurénima: el Bronce Final del Sureste de la Península Ibérica. Bibliotheca Archaeologica Hispana 27—Anejo a la Revista Lucentum 17. Madrid, Real Academia de la Historia. Lorrio, A. J. 2009/10 ‘El Bronce Final en el Sureste de la Península Ibérica. Una (re)visión desde la arqueología funerária’, Anales de Prehistoria y Arqueología 25/26, 119–46. Lucas Pellicer, M. R. 1995 ‘Cerámicas con apliques de metal’, Boletín de la Asociación Española de Amigos de la Arqueología 35, 107–22. Maluquer de Motes, J. 1985 La civilisación de Tartessos. Biblioteca de Cultura Andaluza 18. Sevilla, Editoriales Andaluzas Unidas. Mataloto, R. 2005 ‘A propósito de um achado na Herdade das Casas (Redondo): megalitismo e Idade do Bronze no Alto Alentejo’, Revista Portuguesa de Arqueologia 8 (2), 115–28. Mataloto, R. 2013 ‘Do vale á montanha, da montanha ao monte. A ocupação do final da Idade do Bronze no Alentejo central’, Estudos Arqueológicos de Oeiras 20, 221–72. Mataloto, R., J. M. M. Martins, & A. M. M. Soares 2013 ‘Cronologia absoluta para o Bronze do Sudoeste. Periodização, base de dados, tratamento estatístico’, Estudos Arqueológicos de Oeiras 20, 303–38. Mederos Martín, A. 1996 ‘La conexión Levantino-

27/05/2016 15:44:10

Brandherm

Chipriota. Indicios de comercio atlántico con el Mediterráneo Oriental durante el Bronce Final (1150–950 a.C.)’, Trabajos de Prehistoria 53 (2), 95–115. Mederos Martín, A. 2012 ‘El origen de las estelas decoradas del Suroeste de la Península Ibérica en el Bronce Final II (1325–1150 a. C.)’, Sidereum Ana II. El río Guadiana en el Bronce Final. ed. J. Jiménez Ávila, 417–54. Anejos de Archivo Español de Arqueología 62. Mérida, CSIC. Mederos Martín, A. & L. A. Ruiz Cabrero 2001 ‘Los inicios de la escritura en la Península Ibérica. Grafitos en cerámicas del Bronce Final III y fenicias’, Complutum 12, 97–112. Mergelina y Luna, C. 1943/44 ‘Tugia. Reseña de unos trabajos’, Boletín del Seminario de Estudios de Arte y Arqueología 10, 13–32. Molina González, F. 1978 ‘Definición y sistematisación del Bronce Tardío y Final en el Sudeste de la Península Ibérica’, Cuadernos de Prehistoria y Arqueología de la Universidad de Granada 3, 159–232. Murillo Redondo, J. F. 1994 ‘La cultura tartésica en el Guadalquivir medio’, Ariadna 13/14, 1–496. Murillo Redondo, J. F., J. A. Morena López, & D. Ruiz Lara 2005 ‘Nuevas estelas de guerrero procedentes de las provincias de Córdoba y de Ciudad Real’, Romula 4, 7–46. Neville, A. 2007 Mountains of Silver and Rivers of Gold. The Phoenicians in Iberia. University of British Columbia Studies in the Ancient World 1. Oxford, Oxbow Books. Ordóñez Fernández, R. 2011 La crisis del siglo VI a. C. en las colonias fenicias del Occidente Mediterráneo. Contracción económica, concentración poblacional y cambio cultural. Unpublished thesis, Universidad de Oviedo. Parreira, R. 1995 ‘Aspectos da Idade do Bronze no Alentejo interior’, A Idade do Bronze em Portugal: discursos de poder, 131–5. Lisboa, IPM. Parreira, R. 1998 ‘As arquitecturas como factor de construção da paisagem na Idade do Bronze do Alentejo interior’, Existe uma Idade do Bronze Atlântico?, ed. S. O. Jorge, 267–73. Lisboa, IPA. Parreira, R. & Barros, P. 2007 ‘Necrópoles do Algarve no 2o e 1o milénio a.n.e.’, Xelb 7, 89–102. Pellicer Catalán, M. 1982 ‘Hacia una periodización del Bronce Final en Andalucía occidental’, Huelva Archaeologica 6, 41–7. Pellicer Catalán, M. 1989 ‘Observaciones sobre la problemática tartesia’, Habis 20, 205–16. Pellicer Catalán, M. & J. L. Escacena Carrasco 2007 ‘Rabadanes. Una nueva necrópolis de época tartésica en le bajo Guadalquivir’, Lucentum 26, 7–21.

Brandherm CW3 CS4 19 i 2016.indd 215

[ 215 ]

Pereira Siesgo, J. 1984 ‘Nuevos datos para la valoración del hinterland tartésico. El enterramiento de la Casa del Carpio (Belvis de la Jara, Toledo)’, Tartessos. Arqueología protohistórica el Bajo Guadalquivir, ed. M. E. Aubet, 395–409. Sabadell, AUSA. Pérez Macias, J. A. 1996 ‘La transición a la Edad del Hierro en el Suroeste peninsular. El problema de los “Celtici”’, SPAL 5, 101–14. Pingel, V. 1974 ‘Bemerkungen zu den ritzverzierten Stelen und zur beginnenden Eisenzeit im Südwesten der Iberischen Halbinsel’, Hamburger Beiträge zur Archäologie 4, 1–19. Pingel, V. 1993 ‘Bemerkungen zu den ritzverzierten Stelen im Südwesten der Iberischen Halbinsel’, Lengua y cultura en la Hispania prerromana. Actas del V Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, Colonia, 25–28 de Noviembre de 1989, eds. J. Untermann & F. Villar, 209–31. Salamanca, Universidad de Salamanca. Porfírio, E. 2014 ‘Arquitecturas e práticas funerárias da Idade do Bronze no concelho de Serpa: o caso de Torre Velha 3’, Antrope 1, 250–68. Prados Martínez, F. 2010 ‘La Arquitectura sagrada: un santuario del siglo IX A. C.’, O Castro dos Ratinhos (Barragem do Alqueva, Moura). Escavações num povoado proto-histórico do Guadiana, 2004–2007, eds L. Berrocal Rangel & A. C. Silva, 259–76. O Arqueólogo Português—Suplemento 6. Lisboa, Museu Nacional de Arqueologia. Prósper, B. M.a 2014 ‘Some observations on the classification of Tartessian as a Celtic language’, Journal of Indo-European Studies 42, 468–86. Purcell, N. 2006 ‘Orientalizing: five historical questions’, Debating Ancient Orientalization: multidisciplinary approaches to change in the ancient Mediterranean, eds. C. Riva & N. Vella, 21–30. London, Equinox. Quesada Sanz, F. 2011 ‘Approche chronologique des stèles et statues du domaine ibérique à travers les représentations de l’armement’, Stèles et statues du début de l’âge du Fer dans le Midi de la France (VIIIe–IVe s. av. J.-C.): chronologies, fonctions et comparaisons. Actes de la table ronde de Rodez, eds P. Gruat & D. Garcia, 85–98. Documents d’Archéologie Méridionale 34. Lattes, ADAM. Reboreda Carreira, A. & E. B. Nieto Muñiz 2012 A Pedra Alta de Castrelo do Val. Peza do Mes. Ourense, Museo Arqueolóxico Provincial de Ourense. http:// www.musarqourense.xunta.es/wp-content/iles_mf/ pm_2012_10_gal.pdf accessed: 14.03.2014 Reimer, P. J., E. Bard, A. Bayliss, J. W. Beck, P. G. Blackwell, C. Bronk Ramsey, C. E. Buck, H. Cheng,

27/05/2016 15:44:10

[ 216 ]

vi. Stelae, funerary practice, and group identities

R. L. Edwards, P. Friedrich, P. M. Grootes, T. P. Guilderson, H. Haflidason, P. Hajdas, C. Hatté, T. J. Heaton, D. L. Hoffmann, A. G. Hogg, K. A. Hughen, K. F. Kaiser, B. Kromer, S. W. Manning, M. Niu, R. W. Reimer, D. A. Richards, M. Scott, J. R. Southon, R. A. Staff, C. S. M. Turney, & J. van der Plicht 2013 ‘IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0–50,000 Years cal BP’, Radiocarbon 55, 1869–1887. Roberts, B., M. Uckelmann, & D. Brandherm 2013 ‘Old Father Time: the Bronze Age chronology of Western Europe’, The Oxford Handbook of the European Bronze Age, eds H. Fokkens & A. Harding, 17–41. Oxford, Oxford University Press. Rocha, A. dos S. 1896 ‘A necrópole protohistorica da Fonte Velha, em Bensafrim, no concelho de Lagos’, Revista de Sciências Naturães e Sociães 4, 129–45. Rocha, L. & Duarte, C. 2009 ‘Megalitismo funerário no Alentejo Central: os dados antropológicos das escavaçoẽs de Manuel Heleno’, Investicagiones históricomédicas sobre salud y enfermedad en el pasado. Actas del IX Congreso Nacional de Paleopatología, Morella (Castelló), 26– 29 septiembre de 2007, eds M. Polo Cerdá & E. García Prósper, 763–81. Valencia, Grupo Paleolab. Rodríguez Díaz, A. & J.-J. Enríquez Navascués 2001 Extremadura tartésica. Arqueología de un proceso histórico. Barcelona, Bellaterra. Rodríguez Ramos, J. 2002 ‘Las inscripciones sudlusitano-tartesias: su función, lengua y contexto socio-económico’, Complutum 13, 85–95. Rollinger, R. 2008 ‘Das altorientalische Weltbild und der ferne Westen in neuassyrischer Zeit. Antike Lebenswelten’, Festschrift für Ingomar Weiler zum 70. Geburtstag, eds P. Mauritsch, W. Petermandl, R. Rollinger, & C. Ulf, 683–95. Philippika— Altertumswissenschaftliche Abhandlungen 25. Wiesbaden, Harrassowitz. Ros Sala, M. M. 1990 ‘Datos para una definición del Bronce Final pleno en el altiplano Yecla-Jumilla: las cazuelas carenadas de Coimbra del Barranco Ancho’, Trabajos de Prehistoria 47, 351–62. Ruiz Mata, D. 1995 ‘Las cerámicas del Bronce Final, un soporte tipológico para delimitar el tiempo y el espacio tartésico’, Tartessos 25 años después. 1968–1993, Jérez de la Frontera: Actas del Congreso Conmemorativo del V Symposium Internacional de Prehistoria Peninsular, 265–313. Jérez de la Frontera, Ayuntamiento de Jérez de la Frontera. Said, E. W. 1978 Orientalism. London, Routledge & Kegan. Santos, F. J. C., A. S. T. Antunes, & M. de Deus 2010 ‘A necrópole da I Idade do Ferro de Palhais (Beringel,

Brandherm CW3 CS4 19 i 2016.indd 216

]]

Beja). Resultados preliminares de uma intervenção de emergência no Baixo Alentejo’, IV Encuentro de Arqueología del Suroeste Peninsular, eds J. A. Pérez Macías & E. Romero Bomba, 746–804. Huelva, Universidad de Huelva. Schattner, T. 2011 ‘Zu den Kesselwagen von Baiões’, Madrider Mitteilungen 52, 266–305. Schlichthorst, H. 1793 Ueber den Wohnsiz der Kynesier oder Kyneter—Herodot II 33. IV 49—Ein Beytrag zur Aufklärung der alten Erdbeschreibung. Göttingen, Vandenhoeck & Ruprecht. Schubart, H. 1973 ‘Tumbas megalíticas con enterramientos secundarios de la Edad del Bronce, de Colada de Monte Nuevo de Olivenza’ Crónica del XII Congreso Arqueológico Nacional (Jaén 1971), 175–90. Zaragoza, Universidad de Zaragoza. Schubart, H. 1975 Die Kultur der Bronzezeit im Südwesten der Iberischen Halbinsel. Madrider Forschungen 9. Berlin, Walter de Gruyter. Schulten, A. 1922 Tartessos. Ein Beitrag zur ältesten Geschichte des Westens. Abhandlungen aus dem Gebiet der Auslandskunde 8. Hamburg, Friederichsen. Schulten, A. 1950 Tartessos. Ein Beitrag zur ältesten Geschichte des Westens, 2nd edition. Abhandlungen aus dem Gebiet der Auslandskunde 54. Hamburg, Walter de Gruyter. Serra, M. 2014 ‘Os senhores da planície. A ocupação da Idade do Bronze nos “Barros de Beja” (Baixo Alentejo, Portugal)’, Antrope 1, 269–96. Sieglin, W. 1934 ‘Die Namensform der Stadt Tartesos’, Zeitschrift für Ortsnamenforschung 10, 253–65. Silva, C. T. de & J. Soares 2009 ‘Práticas funerárias no Bronze Pleno do litoral alentejano. O Monumento II do Pessegueiro’, Estudos Arqueológicos de Oeiras 17, 389–420. Sims-Williams, P. 1998 ‘Celtomania and Celtoscepticism’, Cambrian Mediaeval Celtic Studies 36, 1–35. Siret, L. 1913 Questions de chronologie et d’ethnographie ibériques I: de la fin du quarternaire a la fin du bronze. Paris, Geuthner. Soares, A. M. M. 1994 ‘O Bronze do Sudoeste na margem esquerda do Guadiana. As necrópoles do concelho de Serpa’, Actas das V Jornadas Arqueológicas da Associação dos Arqueólogos Portugueses (Lisboa 1993), 179–97. Lisboa, Associação dos Arqueólogos Portugueses. Soares, A. M. M., F.  J. C. Santos, J. Dewulf, M. de Deus, & A. S. Antunes 2009 ‘Práticas rituais no Bronze do Sudoeste: alguns dados’, Estudos Arqueológicos de Oeiras 17, 433–56. Soares, R. M. & A. Martins 2013 ‘A necrópole da Nora Velha 2 (Ourique). Novos dados e interpretações 20 anos após a sua escavação’, Arqueologia em

27/05/2016 15:44:10

Brandherm

Portugal—150 anos, eds J. M. Arnaud, A. Martins, & C. Neves, 661–669. Lisboa, Associação dos Arqueólogos Portugueses. Soares, J. & C. T. da Silva 1995 ‘O Alentejo litoral no contexto da Idade do Bronze do Sudoeste peninsular’, A Idade do Bronze em Portugal: discursos de poder, 136–9. Lisboa, IPM. Spindler, K. 1981 Cova da Moura. Die Besiedlung des atlantischen Küstengebietes Mittelportugals vom Neolithikum bis an das Ende der Eisenzeit. Madrider Beiträge 7. Mainz, Philipp von Zabern. Spindler, K. & O. V. Ferreira 1973 ‘Der spätbronzezeit­ liche Kuppelbau von der Roça do Casal do Meio in Portugal’, Madrider Mitteilungen 14, 60–108. Sureda Carrión, N. 1979 Las fuentes sobre Tartessos y su relación con el Sureste peninsular. Murcia, Universidad de Murcia. Thompson, C. & S. Skaggs 2013 ‘King Solomon’s silver? Southern Phoenician hacksilber hoards and the location of Tarshish’, Internet Archaeology 35 (doi:10.11141/ia.35.6) accessed: 30.03.2014 Torres Ortiz, M. 1999 Sociedad y mundo funerario en Tartessos. Bibliotheca Archaeologica Hispana 3. Madrid, Real Academia de la Historia. Torres Ortiz, M. 2002 Tartessos. Bibliotheca Archaeologica Hispana 14—Studia HispanoPhoenicia 1. Madrid, Real Academia de la Historia. Torres Ortiz, M. 2004 ‘Las necrópolis tartésicas’, El mundo funerario. Actas del III Seminario Internacional sobre Temas Fenicios, Guardamar del Segura, 3 a 5 de mayo de 2002, ed. A. Gonzáles Prats, 425–56. Alicante, Instituto Alicantino Juan Gil-Albert. Torres Ortiz, M. 2005 ‘Una colonización tartésica en el interfluvio Tajo-Sado durante la Primera Edad del Hierro?’, Revista Portuguesa de Arqueologia 8 (2), 193–213. Untermann, J. 1995 ‘Zum Stand der Deutung der „tartessischen“ Inschriften’, Hispano-Gallo-Brittonica. Essays in honour of Professor D. Ellis Evans on the occasion of his sixty-fifth birthday, eds J. F. Eska, R. Geraint & N. Jabobs, 244–59. Cardiff, University of Wales Press. Valério, M. 2008 ‘Origin and development of the Paleohispanic scripts: the orthography and phonology of the Southwestern alphabet’, Revista Portuguesa de Arqueologia 11 (2), 107–38. Valério, M. 2014 ‘The interpretative limits of the southwestern script’, Journal of Indo-European Studies 42, 439–67. Varela Gomes, M. 1994 ‘A necrópole de Alfarrobeira (S. Bartolomeu de Messines) e a Idade do Bronze no concelho de Silves’, Xelb 2, 9–162. Varela Gomes, M. 1995 ‘A Idade do Bronze no Algarve’, A Idade do Bronze em Portugal: discursos de poder, 140–3.

Brandherm CW3 CS4 19 i 2016.indd 217

[ 217 ]

Lisboa, IPM. Varela Gomes, M. 2006 ‘Estelas funerárias da Idade do Bronze Médio do Sudoeste Peninsular. A iconografia do poder’, Actas do VIII Congresso Internacional de Estelas Funerárias. Museu Nacional de Arqueologia, 16, 17, 18 de maio de 2005, 47–62. O Arqueólogo Português, Suplemento 3. Lisboa, MNA. Varela Gomes, M., R. V. Gomes, C. de M. Beirão, & J. L. de Matos J. L. de 1986 A necrópole da Vinha do Casão (Vilamoura, Algarve) no contexto da Idade do Bronze do sudoeste peninsular. Trabalhos de Arqueologia 2. Lisboa, IPAC. Varela Gomes, M. & J. P. Monteiro 1977 ‘Las estelas decoradas do Pomar (Beja, Portugal). Estudio comparado’, Trabalhos de Prehistoria 34, 165–214. Viana, A. 1959 ‘Notas históricas, arqueológicas e etnográficas do Baixo Alentejo’, Arquivo de Beja 16, 3–48. Viana, A. 1960 ‘Notas históricas, arqueológicas e etnográficas do Baixo Alentejo’, Arquivo de Beja 17, 138–231. Vilaça, R. 2007 ‘Todos os caminhos vão dar ao ocidente: trocas e contactos no Bronze Final’, Estudos Arqueológicos de Oeiras 15, 135–54. Vilaça, R. 2009 ‘Sobre rituais do corpo em finais do II milénio/inícios do I milénio a. C.: do espaço europeu ao território português’, Estudos Arqueológicos de Oeiras 17, 489–511. Vilaça, R. 2014 ‘Ensaio sobre a região de Beja em torno do ano mil a. C.: entre a tradição e a inovação’, A Idade do Bronze do Sudoeste: novas perspetivas sobre uma velha problemática, eds R. Vilaça & M. Serra, 99–123. Coimbra: Universidade de Coimbra. Vilaça, R. & A. M. Arruda 2004 ‘Ao longo do Tejo, do Bronze ao Ferro’, Conimbriga 43, 11–45. Vilaça, R., D. J. da Cruz, & A. H. B. Gonçalves 1999 ‘A Necropole de Tanchoal dos Patudos (Alpiarça, Santarém)’, Conimbriga 38, 5–29. Vilhena, J. 2008 ‘As armas e os barões assinaldos? Reflexões em torno das necrópoles monumentais do “Ferro de Ourique” (Sul de Portugal)’, Sidereum Ana I. El río Guadiana en época post-orientalizante. ed. J. Jiménez Ávila, 373–97. Anejos de Archivo Español de Arqueología 46. Madrid, CSIC. Zarzalejos Prieto, M. del M. & F. J. López Precioso, 2005 ‘Apuntes para una caracterización de los procesos orientalizantes en al Meseta Sur’, El periodo orientalizante. Actas del III Simposio Internacional de Arqueología de Mérida, Protohistoria del Mediterráneo Occidental, Vol. I, eds F. J. Jiménez Ávila & S. Celestino Pérez, 595–604. Anejos de Archivo Español de Arqueología 35. Mérida, CSIC.

27/05/2016 15:44:10

Brandherm CW3 CS4 19 i 2016.indd 218

27/05/2016 15:44:10

chapter seven

Language shift and political context in Bronze Age Ireland: some implications of hillfort chronology

William O’Brien The rise of the hillforts at c. 1000 BC poses interesting problems and possibilities. On the one hand, we have seen that treating hillforts as evidence for foreign contacts is just possible, but the evidence is hardly compelling. On the other hand, the hillforts do reflect new central places in the landscape of Ireland, just the type of venue that might have prompted a language shift if they were associated with a new language. They are the archaeological expression of potentially several social domains—religious ceremonies, exchange, social hierarchies—and so it is not difficult to imagine them as places where the local population might be enticed to adopt the new language of the people who introduced them (assuming they were introduced from abroad). (original emphasis)

T

extract from Jim Mallory’s thought-provoking work, The Origins of the Irish (2013, 278), captures the importance and the uncertainties surrounding the prehistoric hillfort in Ireland in relation to language origins. Mallory views the 1st millennium BC as the most likely ‘window’ for the arrival of a new language (Proto-Irish) to Ireland. He links this language shift to a period of significant culture change around 1000 BC, marked by the rise of a warrior élite with important connections to Britain and the Continent. Those contacts were responsible for the spread of hillfort building and sword warfare to Ireland, brought by groups speaking the new language. Across Europe the hillfort is a key site type in any understanding of the emergence of complex societies in later prehistory. This paper presents interim results of a recent study of warfare and political relations in Bronze Age Ireland, with particular reference to the significance of these sites. Fieldwork involving remote sensing survey, large-scale and sample excavation with radiocarbon dating was planned for ten hillfort sites in southern Ireland. The results from six completed investigations are presented, along with radiocarbon dates for the construction of some of the earliest hillforts in Ireland. The implications of this research for Mallory’s thesis are discussed, mindful as he is of the many problems in using archaeological sources to explore language origins in prehistory. Current information suggests there are approximately 100 sites in Ireland that might be properly termed hillforts (James O’Driscoll pers. comm.). This is based on criteria of size

OBrien.indd 219

his

27/05/2016 16:27:07

[ 220 ] vii. language shift and political context in bronze age ireland

(>1 ha in internal area), a prominent landscape position, and the imposing form of the enclosing elements. Hillforts were typically built in prominent elevated positions, where areas of 1–10 ha, and occasionally more, were enclosed by one or more stone walls, earthen banks, ditches and/or timber fencing. They have a wide distribution across the island (Figure 7.1), occurring in isolation or in small clusters, the best known being a group of eight impressive examples in the Baltinglass area of Co. Wicklow. The most widely used classification was devised by Raftery (1972), which separates univallate (Class 1) from what he termed multivallate (Class 2) and inland promontory forts (Class 3). The latter are related in form to a separate category of coastal promontory forts, most of which probably date to the historic era. Hillforts have also been classified on the basis of size, with a distinction made between those over 5 ha in area, examples in the 1–5 ha range, and those under 1 ha (Grogan et al. 1996). The latter group includes an estimated 60 sites now known as ‘hilltop enclosures’, though this number varies depending on the criteria used (134 examples listed in Grogan 2005b, tables 7.6 & 7.7). These are mostly univallate enclosures of less than 1  ha in extent, the chronology and function of which continues to be a matter of speculation. As the name suggests, most hillforts in Ireland are located in elevated positions, generally on the highest point or upper slopes of a prominent hill or ridge. The majority occur at altitudes of 100–300 m OD, with a small number in mountainous settings above 400 m OD, the highest example being the inland promontory fort at Faha in the Dingle Peninsula Co. Kerry (762 m OD). Most hillforts command panoramic views over broad expanses of lowland, from which they are visible over great distances. This is particularly true of Class 2 hillforts, including the examples considered in this paper, with most sited at the break of upland and lowland. It is generally accepted that this landscape setting had a strategic aim, designed to ‘control’ a specific territory and its resources, as well as significant routeways in the landscape (Condit & O’Sullivan 1999). Whatever the intended or acquired significance, the visual impact of a hillfort was an important consideration in its landscape setting. Though often described as multivallate, the Class 2 hillfort in Ireland is more properly termed a ‘multiple enclosure hillfort’. These are comprised of two or three (rarely four) enclosures, circular or oval in plan, where concentric lines of stone wall, earthen/stone bank, ditches and/or timber fencing are spaced 20–100 m apart, either along or across the natural contours of a hill or ridge, or on a cliff edge. Current estimates indicate around 40 Class 2 hillforts spread widely across Ireland, with a notable concentration in north Munster and south Leinster where most of the largest examples are found (Figure 7.1). Hillfort chronology As in many parts of Europe, the phenomenon of hilltop enclosure has a long history in Ireland, dating from Neolithic to medieval times. The earliest examples date from the early/mid-4th millennium BC, and include sites such as Lyles Hill, Co. Antrim (Evans 1953; Simpson & Gibson 1989), Donegore, Co. Antrim (Mallory et al. 2011), and Knocknarea,

OBrien.indd 220

27/05/2016 16:27:07

O’Brien

[ 221 ]

7.1 Distribution of prehistoric hillforts (Class 1 and 2) in Ireland, showing location of excavated sites (this paper). The distribution of ‘hilltop enclosures’ is not shown, and the only inland promontory forts depicted are those most likely to be prehistoric. (source James O’Driscoll)

OBrien.indd 221

27/05/2016 16:27:21

[ 222 ] vii. language shift and political context in bronze age ireland

Co. Sligo (Bergh 2002). Other examples include the causewayed enclosure at Magheraboy, Co. Sligo (Danaher 2007) and a palisaded Neolithic settlement at Tullahedy, Co. Tipperary (Cleary & Kelleher 2011). The latter is an example where a low hill partly surrounded by water provided a defensive location, which was also a consideration for the Neolithic settlement on Knockadoon, Lough Gur, Co. Limerick. While there are similarities with later hillforts, there are no examples of continuous occupation from the Neolithic to Late Bronze Age in these sites. A number of Irish hillforts were built on hills with older burial monuments; for example, Rathcoran, Co. Wicklow, Knocknashee, Co. Sligo and Freestone Hill, Co. Kilkenny. This might be interpreted as a referencing of ancestral origins in the landscape; alternatively, it could represent the appropriation of such connections by new population groups. In most instances the former is more likely, which could be used to argue for an indigenous background to the first hillforts. The available evidence points to the emergence of the hillfort as a new phenomenon in Ireland during the Middle Bronze Age, 1400–1100 BC. The origins of the Irish hillfort remain uncertain. Enclosure is certainly a growing element of small-scale residential settlement in the Middle and Late Bronze Age, with the use of palisades, ditches and stone walls at such sites as Aughinish, Co. Limerick; Ballybrowney, Co. Cork; Carrigillihy, Co. Cork; Ballyveelish, Co. Tipperary; Chancellorsland, Co. Tipperary; Kilsharvan, Co. Meath; Lagavooreen, Co. Louth, and Clonfinlough, Co. Offaly, to name but a few (see Waddell 2010, 216–24). While some common elements are present, it is difficult to make a direct connection between smaller enclosed settlements and the emergence of the hillfort, a type of site that represents an entirely different scale of construction and communal effort. Indigenous inputs notwithstanding, the origins of the hillfort in Ireland can be sought elsewhere. The phenomenon as a whole can be traced back to developments in central and eastern Europe during the Early Bronze Age, at sites such as Nitriansky Hrádok in Slovakia and others. The construction of hillforts increased greatly during the Middle and Late Bronze Age, when heavily fortified settlements were built in many parts of Urnfield and Atlantic Europe (see Harding 2006 for review). This was connected to new forms of warfare and territorial ambition, a militarism that is visible in the mass production of offensive weaponry and the rise of a warrior élite. These influences from the Continent reached Britain by 1300 BC and from there spread quickly to Ireland. British hillforts have long been associated with the Iron Age, with such iconic monuments as Maiden Castle and Danebury. This perception has been revised in recent years with the identification of an early hillfort horizon dating to the Middle/Late Bronze Age transition. Of particular interest is that many of the earliest British sites were palisaded enclosures, broadly similar and contemporary with Irish examples considered in this paper. They include Rams Hill in Berkshire with its Middle Bronze Age (Penard phase) origin, and Taplow in Buckinghamshire dated to the 11th century BC, among others. At some sites these fenced enclosures were eventually replaced during the Late Bronze age by larger hillforts with earthen ramparts. This sequence is recorded in the Welsh Marches at sites such as the Breidden, Old Oswestry, and Llwyn Bryn-dinas (see Brown 2009, 28–33 and Harding 2012, 154–6 for summaries). Turning to Ireland, where the different categories of hillfort are widely regarded

OBrien.indd 222

27/05/2016 16:27:21

O’Brien

[ 223 ]

as having a chronological basis, with the multiple enclosure (Class 2) examples widely regarded as Late Bronze Age in date, as opposed to a later date for the univallate (Class 1) examples (though some of these have a Bronze Age background). These assumptions are based on a small number of excavated sites, which includes some prominent examples. In most cases the dates for hillfort construction, use, and abandonment are not well established, with only a small number of well-established occupation records. Prior to this project, a small number of Class 2 (multiple enclosure) hillforts had been excavated in Ireland. The first of these was the 1969–79 excavation project at Rathgall, Co. Wicklow (Raftery 1972; 1976). This site was instrumental in changing perceptions of Irish hillforts as Iron Age and related to the well-known British examples. Rathgall is located on a prominent ridge (137 m OD) in a lowlying setting in south-west Wicklow. It extends over 7.3 ha and comprises four concentric enclosures, the innermost being a stone cashel of early medieval date (Becker 2010; Waddell 2010, 285–6). Excavation in the central area uncovered a high-status occupation with specialized metalworking and an associated funerary ritual complex. A series of radiocarbon dates centred on 1200–1000 BC is consistent with the discovery of items of Roscommon phase metalwork. Though the hillfort defences are not closely dated, this Bronze Age occupation is likely to be contemporary with the middle and possibly the outer ramparts. Further consideration of the Bronze Age hillfort at Rathgall must await final publication of those excavations (Becker in preparation). The Rathgall study was followed in 1987–95 by investigations at Haughey’s Fort, located on a prominent hill 1 km west of the Iron Age ceremonial enclosure at Navan Fort, Co. Armagh. This hillfort is mostly levelled, but originally comprised three concentric bank and ditch enclosures, with an overall size of 340 m by 310 m. Excavation revealed a long duration of occupation, lasting from 1300–900 BC, with evidence of activity pre-dating the hillfort, as well as some Iron Age occupation (Mallory 1995; Mallory et al. 1996). Radiocarbon dates for the infilling of the inner ditch suggest the defences may have been established in the 12th century BC (see also Waddell 2010, 224–7). This is supported by a dendro-date of 1166 BC for wood from the base of that ditch (Baillie & Brown 1998). The next phase of Class 2 hillfort research commenced in the early 1990s with excavations at Mooghaun, Co. Clare and Dún Aonghasa on Inishmore, Aran Islands. The former is an impressive hillfort comprised of three concentric stone walls enclosing an area of 11 ha. The excavator argued that the hillfort was built as a single design, most probably in the period 915–905 cal BC, preceded by activity around 925–915 cal BC with continued occupation to possibly 815 cal BC (Grogan 2005a, 240, fig. 7.10). An earlier date is possible as the excavation did return dates of 1125–918 cal BC for the middle enclosure, while charcoal from under the outer bank is dated 1255–917 cal BC. The possibility that some of the enclosing walls date to the Iron Age has been raised (Henderson 2007, 182–3). The chronology for Dún Aonghasa on Inishmore, the largest of the Aran Islands off Galway, is even more complex (Cotter 2012, fig. 12.3). The innermost enclosure is an impressive citadel, spectacularly sited on an 87 m high vertical sea cliff. This, together with the outer walls and the formidable chevaux-de-frise in stone, creates several layers of defence over an enclosed area of 5.7 ha. Radiocarbon results indicate settlement on this cliff edge as early as 1300 cal BC, with the earliest phase of wall building dating around 1100 cal BC.

OBrien.indd 223

27/05/2016 16:27:21

[ 224 ] vii. language shift and political context in bronze age ireland

The site was occupied to 800 cal BC, with further settlement well into the Iron Age when modifications may have been made to the enclosure walls (see Warner in Cotter 2012, 212–24; also Henderson 2007, 179–82). The investigations at Dún Aonghasa and Mooghaun highlighted the difficulty of dating the phases of stone wall construction at large multi-period hillforts. This can be especially problematic where there was earlier site occupation, and where the hillfort occupation was of some duration. Despite the upsurge in archaeological excavation in Ireland in recent years, there have been relatively few hillfort investigations to provide new data. A large Class 2 hillfort at Rahally, Co. Galway, was excavated in 2004–5 in advance of road construction (Mullins 2008). Radiocarbon dates for charcoal from the base of each of the three enclosing ditches returned dates of 994–827 cal BC, 790–520 cal BC, and 1090–900 cal BC respectively. Pottery from the interior confirms that the site was occupied during the Late Bronze Age. Limited excavation carried out at a double enclosure site at Knockanacuig, Co. Kerry, returned Iron Age dates that date neither the construction nor primary occupation of this hillfort (Dunne & Bartlett 2009). There is also growing evidence for Late Bronze Age activity at other hillfort types in Ireland. These include the excavation of a triple-ramparted inland promontory fort at Knockhu, Co. Antrim, where radiocarbon results indicate Late Bronze Age occupation (McNeary 2014). This raises the possibility that other inland promontory forts (Raftery Class 3) may date to the Late Bronze Age. The same is possible for some of the Class I (univallate hillforts) in Ireland. Excavation has uncovered Late Bronze Age pottery at Freestone Hill, Co. Kilkenny (Raftery 1969; Ó Floinn 2000); Downpatrick (Cathedral Hill), Co. Down (Proudfoot 1954; Raftery 1976); and Clogher, Co. Tyrone (Warner 2009). In each instance the hillfort defences cannot be closely dated, and there is evidence of significant occupation from Iron Age to medieval times. There continues to be uncertainty as to the date range of the class of ‘hillfort enclosures’ identified by Grogan (2005b). This category includes a range of univallate sites of varying sizes under 1 ha, from large ringforts of the early medieval period to examples that are probably of prehistoric date. Separate from those sites are the internally ditched enclosures occurring at ceremonial sites of the developed Iron Age. These include such well-known examples as Rath na Riogh on the Hill of Tara, Co. Meath, and Navan Fort, Co. Armagh. In summary, the results from Dún Aonghasa, Mooghaun, Haughey’s Fort and Rathgall suggest that the building and earliest occupation of Class 2 hillforts commenced in the Late Bronze Age, and lasted for several centuries (1200–800 BC). These excavations highlight the problem of obtaining secure dates for the initial construction, modification and abandonment of hillfort defences, particularly where these are built in stone. The available dates come from either pre-rampart or ditch contexts and at best provide terminus ante/ post quem statements on the construction of these hillforts. The problems are considerable, and it should be noted that at least two hillfort excavations in recent times were unable to obtain any secure dating evidence, namely the investigation of univallate enclosures on Knockacarriggeen Hill, Co. Galway (Carey 2002) and at Clenagh, Co. Clare (Grogan & Daly 2005).

OBrien.indd 224

27/05/2016 16:27:21

O’Brien

[ 225 ]

7.2 Clashanimud hillfort, Co. Cork

This Project In 2011 the author initiated a project to address the question of hillfort chronology in Ireland, with a specific focus on Class 2 (multiple enclosure) examples. This is part of a wider study of warfare and society in later Bronze Age Ireland. The project developed out of large-scale excavation undertaken in 2004–6 at Clashanimud, Co. Cork. The second phase of research targeted nine other Class 2 hillforts for sample excavation and radiocarbon dating, five of which are now completed. The results of these six hillfort investigations are summarized here (Figure 7.1), along with a general presentation of the radiocarbon evidence. Clashanimud, Co. Cork This hillfort is sited on a spur off a prominent ridge 14 km to the west of Cork city. The location is strategic in respect of a broad valley to the immediate south, with panoramic views over a wide area of the mid and west Cork lowlands. The hillfort comprises two concentric oval enclosures, spaced c. 48 m apart, with an overall size of 350 m × 250 m over 8.8 ha (Figure 7.2). There are no prehistoric features visible in the interior, and a possible cairn or small enclosure recorded in 1901 near the summit is no longer extant. The inner enclosure is surrounded by a low bank, with an external ditch and a small counterscarp bank. The main bank was built using soil and broken bedrock from the ditch, and was finished with an external stone facing and an internal wooden revetment. Excavation revealed large postholes, spaced 1.5 m apart, following the central axis of the bank. These held upright posts, possibly 4–5 m in height, which were secured by packing stones and the enclosing bank. An estimated 360 posts were used in this fashion along a 0.8 km perimeter

OBrien.indd 225

10/06/2016 18:16:54

of the inner enclosure. Charcoal analysis indicates the use of oak roundwood posts with a diameter of 0.2–0.3 m. The digging of a shallow trench between the main posts, to allow the insertion of smaller posts and possibly wattlework, completed the palisade. The adjacent ditch cut through subsoil into bedrock, and had a U-shaped profile, measuring 2.4 m wide at the top and 0.65–1.4 m wide at the base, with a central depth of 1.7–1.8 m. Taken together, these features represent a 6–7 m high barrier (from base of ditch to estimated top of palisade) protecting the inner enclosure (Figure 7.3). This points to a strong defensive purpose rather than a symbolic meaning or animal stockade. An original entrance to the inner enclosure was identified on the western side of the site. This comprises a 6 m wide causeway across the rock-cut ditch, leading towards a 2.2 m wide opening in the enclosure bank, the sides of which were faced with stone 7.3 Defences of inner enclosure, Clashanimud walling. Rock-cut postholes on either side hillfort, Co. Cork of this entrance were part of a wooden gate connected to the bank palisade. Excavation confirmed that both this entrance and the entire palisade of the inner enclosure were burned down soon after the hillfort was built, in what must have been a deliberate and hostile act. Evidence for this comes from charcoal deposits and burnt timbers found within the collapsed bank. The bank surface was also intensely fire-reddened, with a large number of vitrified stones in the entrance area. The outer hillfort enclosure has a perimeter of 1.04 km, and was originally protected by a low earthen bank, measuring 4 m wide at the base with a surviving height of 0.6– 0.8 m. Excavation of the truncated bank surface revealed two rows of opposing stakeholes indicative of stake fencing. This was much lighter than the palisade of the inner enclosure and does not seem to have been burnt down. The external ditch is now infilled along the entire enclosure perimeter, but originally had a V-section profile, measuring 2.85 m wide at the top with a central depth of 1.7 m. A total of twelve radiocarbon dates are available for Clashanimud hillfort (Table 7.1). Eleven of these relate to the enclosing elements, with the remaining result connected to pre-hillfort activity. Nine dates come from the burnt palisade at three separate points of excavation along of the inner enclosure (Figure 7.4). The first five of these (GrA-

OBrien.indd 226

27/05/2016 16:27:23

[ 227 ]

O’Brien Hillfort Clashanimud, Co. Cork

Lab number

C-14

95.4%

result BP

calibration

GrA-42679

3105±35

1442–1273 BC

Palisade post charcoal, entrance to inner enclosure

GrN-29046

3050±20

1396–1231 BC

Palisade post charcoal, bank of inner enclosure

GrN-29043 GrN-29044 GrN-29045 GrA-42683

GrN-29042 GrA-42684 GrA-42682 GrA-42678 GrA-42677 Glanbane, Co. Limerick

2920±50 2915±35 2910±35 3160±40 2985±35

1260–1116 BC 1258–1027 BC 1263–976 BC

1216–1008 BC 1214–1006 BC 1511–1303 BC 1378–1089 BC

Charcoal layer within bank of inner enclosure Palisade charcoal, bank of inner enclosure

Charcoal at base of ditch; TAQ for inner enclosure Palisade charcoal, entrance to inner enclosure Palisade charcoal, entrance to inner enclosure Pre-bank charcoal; TPQ for outer enclosure

Charcoal at base of ditch; TAQ for outer enclosure Charcoal at base of ditch; TAQ for inner enclosure

GrA-54510

2975±30

1368–1059 BC

Charcoal from gate posthole, inner enclosure

GrA-54616

2915±45

1258–979 BC

Charcoal from entrance passage, inner enclosure

2850±30 2930±50

1111–927 BC 1279–980 BC

Charcoal at base of ditch; TAQ for inner enclosure Charcoal from pre-bank slip surface, inner enclosure

GrA-54958

3125±35

1494–1289 BC

Pre-bank charcoal; TPQ for inner enclosure

GrA-54956

2840±30

1108–917 BC

Charcoal from inside bank of inner enclosure

GrA-54964 GrA-54961 GrA-54965 GrA-54963

2890±30 2880±30 2880±30 2920±35 2900±30

1195–978 BC 1192–939 BC 1192–939 BC

1218–1011 BC 1207–1004 BC

Charcoal at base of ditch; TAQ for inner enclosure Charcoal from stake fence, outer enclosure bank Charcoal from stake fence, outer enclosure bank Pre-bank charcoal; TPQ for outer enclosure

Charcoal from stake fence, outer enclosure bank

GrA-58713

2975±35

1369–1056 BC

Burnt bone from palisade trench; inner enclosure

GrA-58574

2945±30

1258–1049 BC

Twig at base of ditch, inner enclosure

GrA-58572 GrA-58575 GrA-58614 GrA-58577 GrA-58576 Rathnagree, Co. Wicklow

2940±35

Palisade timber charcoal, bank of inner enclosure

1118–929 BC

GrA-54957

Toor More, Co. Kilkenny

2965±20

1371–1130 BC

Palisade post charcoal, bank of inner enclosure

2860±30

GrA-54599 Formoyle, Co. Clare

3000±20

1409–1280 BC

GrA-54509 GrA-54508

Ballylin, Co. Limerick

3075±20

Sample context

2970±30 2935±35 2845±35 3000±30 2920±35

1281–1058 BC 1257–1019 BC 1114–917 BC

1377–1126 BC 1218–1011 BC

Palisade trench charcoal, inner enclosure Wood at base of ditch, inner enclosure

Charcoal at base of ditch; TAQ for inner enclosure Charcoal from burnt spread, outer enclosure Palisade trench charcoal, outer enclosure

GrA-60927

3055±40

1417–1213 BC

Palisade post charcoal, inner enclosure

GrA-60929

3050±40

1417–1208 BC

Charcoal spread, middle enclosure (Period 1)

GrA-60925 GrA-60930 GrA-60931 GrA-60932

3045±40 3015±40 3020±40 3035±40

1414–1135 BC 1395–1126 BC 1397–1128 BC 1411–1131 BC

Charcoal from palisade timber, inner enclosure Charcoal from burnt stake, middle enclosure (Period 1) Charcoal from palisade timber, outer enclosure Palisade post charcoal, outer enclosure

Table 7.1 Radiocarbon dates from recent hillfort excavations in Ireland (this project). Calibration based on OxCal v.4.2.4. Radiocarbon dates for peat samples from Ballylin and Formoyle not included.

OBrien.indd 227

27/05/2016 16:27:24

[ 228 ] vii. language shift and political context in bronze age ireland

7.4 Calibration of radiocarbon dates, Clashanimud hillfort, Co. Cork (based on OxCal v4.2.4)

29042–6) come from mature oakwood, and span the period c. 1400–1000 BC. The dating of four samples of oak sapwood allows this date range to be narrowed to the 12th and 11th centuries BC (GrA-42682–4), with one outlier (GrA-42679). There are no palisade samples available for the outer hillfort enclosure, the building of which should lie between 1500–1100 BC (GrA-42677–8). In summary, the hillfort at Clashanimud was built some time c. 1250–1050 BC, during the Middle/Late Bronze Age transition in Ireland. Extensive excavation of both the inner and outer enclosures did not uncover any evidence of prehistoric occupation. This must be connected to the deliberate destruction of the hillfort soon after it was built, either as a result of a direct attack or following conflict in the vicinity. Whatever the circumstances, the destruction of this hillfort was a deliberate slighting of the political power that it represented. There is no evidence that this stronghold was re-built or re-occupied after the events that led to its destruction.

OBrien.indd 228

27/05/2016 16:27:26

O’Brien

[ 229 ]

7.5 Magnetic gradiometry survey of Glanbane hillfort, Co. Kerry, showing location of excavation trenches across inner and outer enclosing elements (red). Note circular post structures in centre of hillfort (not excavated) (source: James O’Driscoll)

Glanbane hillfort, Co. Kerry This hillfort is located on the highest point (89 m OD) of a ridge at the eastern end of the Sliabh Mis mountains in west Kerry. This is a prominent location, with panoramic views of the Castleisland and Killarney lowlands. The site was levelled by agriculture in the late 1960s, with only low-relief indications of the enclosing elements visible today. Aerial photographs indicate two concentric enclosures, spaced 20–25 m apart, each comprising a bank with external ditch (e.g. Raftery 1994, late 12). This is confirmed by magnetic gradiomery survey, which shows the hillfort was originally sub-circular in plan, measuring 290 m × 280 m, over 6.58 ha (Figure 7.5). The geophysical survey indicates possible entrances on the north-east side of the hillfort, as well as several arcuate features in the central area including an ovoid enclosure and what appear to be post-built structures. In 2012 two trenches were excavated at Glanbane hillfort, across the eastern side of both enclosures respectively. While there are no occupation-related finds, some dating information was recovered for the defences of the inner enclosure. The inner enclosure measures 240 m × 230 m and extends over an area of 4.7 ha. Excavation uncovered a U-shaped ditch, the upper part of which cut through a thin layer of B-horizon subsoil into

OBrien.indd 229

27/05/2016 16:27:36

[ 230 ] vii. language shift and political context in bronze age ireland

7.6 Partial reconstruction of defences of inner enclosure, Glanbane hillfort, Co. Kerry. These include a rockcut ditch and rebuilt stone facing (original stones) of now destroyed inner bank (scale: 20 cm divisions)

bedrock. The ditch measured 2.4–3.1 m in upper width, with a depth of 1.35 m on the inner side and 1.05 m on the exterior. Excavation uncovered a large number of blocky stones in the lower ditch fill. These had fallen from a wall originally built along the inner edge of the ditch to retain an earth and stone bank (Figure 7.6). The outer enclosure was surrounded by a similar bank-and-ditch arrangement. Excavation uncovered a U-shaped ditch cut through a thin layer of subsoil into bedrock. The ditch measured 1.8–2.05 m in width with a maximum depth of 0.8 m. The discovery of blocky stones in the lower ditch fill indicates a stone facing to an inner bank, which collapsed gradually over a period of time. The surviving portion of the bank was levelled in the modern era. There are three radiocarbon dates available from the Glanbane excavation, two of which relate to the hillfort phase. Both are charcoal samples obtained close to the bottom of the inner ditch, which provide a broad terminus ante quem of 1111–929 cal BC for construction of the inner enclosure (see Table 7.1). There are no reliable dates for the outer enclosing ditch, where residual charcoal of Mesolithic date was found. Though limited in scale, the excavation at Glanbane did yield important information on the original hillfort defences. Shallow ditches were dug through thin subsoil into bedrock, and the soil and stone extracted was piled on the inner side to create a stone-

OBrien.indd 230

27/05/2016 16:27:40

O’Brien

[ 231 ]

faced bank, a metre or so in height. There is no evidence of a timber palisade or stake fence accompanying either bank. The overall defensive height of the inner enclosure (base of ditch to top of bank) is estimated at 2.5–3 m, while that of the outer enclosure is possibly 2 m. Though somewhat slight, the scale of construction is impressive over the entire extent of the hillfort. The perimeter of the inner enclosure is 790 m, while that of the outer enclosure in 932 m, equalling almost 1.7 km of artificial defences. The total estimate of rock extraction for both enclosure ditches at Glanbane is 2643.9 tonnes, with an estimated 33,000 stones required to face the associated banks. Ballylin, Co. Limerick This hillfort is located 2 km north-west of Ardagh village, on the summit (243 m OD) of a prominent ridge off the eastern margin of the Mullaghareirk mountain range. The location has panoramic views of the lowlands of central Limerick and the lower Shannon estuary to the north. Aerial survey and fieldwalking confirm that the hill is surrounded by two concentric enclosures over an area of 21 ha (Cody 1981), making this one of the largest hillforts in Ireland. The two enclosures are defined by bank and ditch earthworks, visible either as low-relief features or vegetation patterns. This is confirmed by recent geophysical survey and lidar imaging, which indicates the enclosing lines comprise of an inner and outer bank with an intervening ditch (Figure 7.7). These surveys also identified an unrecorded opening on the 7.7 Lidar image of Ballylin hillfort, Co. Limerick. The location of the two excavation trenches is shown (red), the inner example located at the original entrance to the inner enclosure. (source: this project)

OBrien.indd 231

27/05/2016 16:27:57

[ 232 ] vii. language shift and political context in bronze age ireland

eastern side of the inner enclosure. This original entrance was excavated in 2012, as was the line of the outer enclosure. The inner hillfort enclosure measures 300 m × 315 m over an area of 7.5 ha. Excavation in the entrance area exposed a 3.5 m wide by 1 m deep ditch with a flat base. Soil and broken bedrock from the ditch had been piled on the inner side to form a low bank. This measured c. 2 m wide and 0.35 m high in its surviving form. There is no evidence for a bank palisade, though two large post-holes with packing stones, placed 2.9 m apart on either side of the passageway, signifies some kind of wooden gate (Figure 7.8). There is no indication that the posts decayed or burnt in situ, and their removal cannot be linked to any event in the history of the hillfort. Excavation did reveal small concentrations of charcoal along the entrance passageway, but not enough to suggest any deliberate destruction by fire of the entrance ‘gate’. The outer hillfort enclosure measures 500 m × 525 m over an area of 13.5 ha. The enclosing element comprises a shallow soil-cut ditch with a low bank on either side. The absence of a wooden fence, together with the low height of the bank and ditch combined ( 

 ~, ?  < Ar.  ~ vow. red.  < Ar.    (11+) 

?? ?          

           





Table 14.2 Similar features in Insular Celtic and Hamito-Semitic according to language

27/05/2016 19:00:13

[ 416 ] xiv. a hamito-semitic substratum and celtic from the west

]

  4. Relative clause copying, not gapping: the bed that I slept in it. Yes, Breton ar gwele a meus kousked ennañ [the bed.m aff I.have slept in.it.m]; Arabic as-sar¼r al-ladh¼ nimt f¼-h [the-bed.m rel.m I.slept in‑it.m] ‘the bed I slept in/in which I slept’. The World Atlas of Language Structures Online (Feature/Chapter 123: Relativization on Obliques) does show a heavy concentration of this strategy, apart from in Irish and Scottish Gaelic, in Semitic (Hebrew and Arabic) and in numerous languages of central sub-Saharan Africa, but also in Persian, Eastern Kayah Li (Thailand, Myanmar), Paamese (Vanuatu, South Pacific), and Guaraní (Paraguay, Brazil). Berber never has resumptive pronouns, only movement of the bare preposition; Old Irish is similar, but the order of relator and preposition is the reverse: Berber rel+prep; Old Irish prep+rel.   5. Special relative tensed verb form. This is present in Irish (-as vs. -aidh, -ann, &c.) and apparently in Egyptian and Berber; in Brythonic, the only modern trace is in the present relative form of the copula Welsh sy(dd), Breton so (Modern Breton zo); apart from Akkadian, this is unknown in Semitic; the Berber ‘relative form’ is commonly called a ‘participle’.   6. Subject and object marking in verb. Yes, for both Insular Celtic and Semitic; this concerns most strongly Old Irish, Berber, and Egyptian. Object pronouns are traditionally proclitic in Celtic and postclitic in Semitic; this is not the same, however, as the true subjectand-object-marking verbal morphology of Georgian. The cliticization of object pronouns on the verb is hardly a rare trait.   7. Object marker: preverb-infix-V/V-suffix. This concerns especially Old Irish and Berber. It should be noted more generaly that pre- versus post-cliticization of object pronouns concerns many languages, cf. Romance, Serbo-Croat, &c.   8. Genitive construction: def. art. on dependent only: house the-man. Breton ti ar roue [house the king], Arabic bait al‑malik [house the‑king] ‘the king’s house’. Known as the ‘construct state’ cs among Semiticists, this is not necessarily due to substratal influence. In the typology of genitive constructions, there is a limited number of parameters: (1) the order of head and dependent: h d or d h (vso normally implies h d order); (2) the presence or absence of an article (on d only; on both d and h; no examples of the article on h only; all Insular Celtic and most, but not all, Hamito-Semitic languages have a definite article); (3) the relation marking may be on either h or d; and finally (4) a limited number of relator mechanisms (one or more are possible): (a) simple adjacency (h  d, as in all Insular Celtic and Hamito-Semitic languages except Amharic, or d h); (b) phonetic modification of either h or d : phonetic cs marking of h in Hebrew: bayit ‘house’, but b{t ham‑m ɛl ɛk [house. cs the‑king]; d\b\r ‘word’, but də̆ bar ham‑mɛlɛk ‘the king’s word’; or phonetic modification of d in Berber: agellid ‘king’, but axxam (n) ugellid [house (of) king. cs ]; (c) case: gen , dat , obl , &c.; (d) possessive poss: Turkish d-gen + h-poss: kral-ın ev-i [king‑gen house‑his]; (e) link particle lnk: in both Hindi-Urdu and Swahili the link agrees with h: Hindi-Urdu la²k\ ‘boy.nom’, la²k{ k\ ghar [boy.obl lnkx housex] ‘the boy’s house’; Swahili: nyumba ya mfalme [house.cl9 lnk.cl9 king.cl1] ‘the king’s house’; (f) adposition: preposition, postposition, cf. English the door of the house.   In Insular Celtic and Hamito-Semitic languages which no longer have case, the construct state is defined solely by the adjacency of the head and dependent, and the restriction of the article to the dependent. However, in both families this is probably the result of independent

Hewitt CW3 3 iii 16.indd 416

27/05/2016 19:00:13

Hewitt

[ 417 ]

evolution. While Breton, Welsh, Hebrew, and colloquial Arabic have no cases: Breton dor an ti, Welsh drßs y t½, Hebrew dɛlɛt hab-bayit, colloquial Arabic b\b al‑bait [door the‑house] ‘the door of the house’, Classical and formal Modern Standard Arabic and Irish (for some items at least) conserve case endings: Arabic b\b-u l-bait-i; Irish doras an tí [door.nom the‑house. gen], and these help to define the genitive relation. It is only with the loss of the case endings that the [h [the‑d]] structure becomes crucial to defining the genitive construction.   Germanic has both a compact genitive construction the king’s house, with genitive case and only one article possible, on d, and a periphrastic construction the house of the king with two articles and the genitive relation expressed by the preposition of. While it is not obvious to ordinary English-speakers which element the article the applies to in the king’s house, other Germanic languages provide a clue: German des Königs Haus [the.gen king.gen house. nom ] or Swedish: kungens hus [king.the. gen house]. It therefore seems logical to bracket the phrase as follows: [[the king’s] house], which is simply the reverse of the order of the two main constituents h and d in Breton [ti [ar roue]] or Arabic [bait [al‑malik]]. Indeed, in one Germanic language which has lost all genitive case-marking, the [[the‑d]  h] order actually defines the genitive relation: the highly evolved form of English found in Jamaican Creole: [[di king] hoos]. Seen in this light, the Insular Celtic genitive structure is rather less exotic than it might appear at first sight; there is little need to appeal to Hamito-Semitic for a source.   Insular Celtic and Hamito-Semitic differ with regard to adjective placement in genitive constructions. While neither Celtic nor Germanic has any problem in attaching adjectives to either or both h and d: Breton ti bihan ar roue bras [house little the king big] ‘the big king’s little house’, Semitic cannot do this; any adjectives go obligatorily after the genitive construct, which is more akin to a compound noun ‘king‑house’, so bait al‑malik al‑kab¼r [house the‑king the‑big] can in principle mean either ‘the king’s big house’ or ‘the big king’s house’. Only in formal Arabic is it possible to tell which the adjective applies to, from the case-marking. The usual way of applying adjectives to both h and d is to use a longer construction with two articles and a preposition, structurally similar to the periphrastic construction the house of the king: al‑bait aṣ‑ṣagh¼r li‑l‑malik al‑kab¼r [the‑house the‑little to‑the‑king the‑big] ‘the big king’s little house’. If the genitive construction in Insular Celtic really had its origins in a Hamito-Semitic substratum, it is difficult to understand why this major structural constraint prohibiting the insertion of adjectives between h and d would not also apply in Celtic.   9. Non-agreement of verb with plural noun subject. This is a striking parallel, strongest in Welsh, Breton, Egyptian, Classical Arabic and to some extent Biblical Hebrew, in the latter two with VSO order only; in Breton also with SVO order in the affirmative, but not in the negative; not in Berber. Non-agreement is fairly common with VS order worldwide, cf. Greenberg (1966), Universal 33: ‘When number agreement between the noun and verb is suspended and the rule is based on order, the case is always one in which the verb precedes and the verb is in the singular.’ Number non-agreement with plural post-verbal subjects is lost in spoken Arabic, and was lost in Hebrew from the Mishnaic period, so those two languages have moved away from non-agreement. There was usually agreement with postverbal plural subjects in Old Welsh and Old Breton, so non-agreement appears to have come in since those periods. Non-agreement appears to be even more recent in Gaelic; indeed, in

Hewitt CW3 3 iii 16.indd 417

27/05/2016 19:00:13

[ 418 ] xiv. a hamito-semitic substratum and celtic from the west

]

Gender Article possible Pron. object: possessive Lexical object: genitive

   

   ()

  > ()>

 ‒ ? ?

Table 14.3 Verbal noun or infinitive?

‒ ‒  ()

   

Hebrew

Arabic

Egyptia n

Berber

Breton

Welsh

Nominal features

Irish

some dialects, such as Munster, there is often still agreement. These are exceedingly long times for some putative substratal non-agreement to have filtered through.   10. Verbal noun, not infinitive (object in genitive, not accusative). There appears to be more of a cline than a sharp distinction between the abstract verbal noun (Arabic, Georgian maṣdar) and the infinitive (see Table 14.3). The criterion for distinguishing between the two is whether objects are in the genitive (verbal noun) or accusative (infinitive). With the development in Breton since the 18th century (with the exception of the SE Gwened/ Vannes dialect) of true ‘accusative’ object pronouns (etymologically ‘of + pronoun’): ma gweled [my seeing] > gweled ahanon [see.inf/vn of.me] ‘to see me’, little now distinguishes the Breton verbal noun from the French infinitive. In Insular Celtic, only the Irish verbal noun seems truly maṣ dar-like.

? ?  ()

  11. Predicative particle: he is in a farmer. Especially in Welsh and Egyptian, if Welsh predicative yn really is ‘in’, and this has recently been challenged by Gensler (2002). The construction is ‘In his farmer’ in Irish. This feature is very limited in Breton, and is marginal in Hebrew and Arabic.   12. Prepositional periphrastic: he is at singing. As Comrie points out (1976, 100–102), apart from Insular Celtic languages, which all have prepositional periphrastic constructions, copular locative phrases expressing the progressive are found in numerous other languages: Chinese, Georgian, Yoruba, Shona, Igbo, Kpelle, other Nigero-Congolese languages, Hindi/Urdu, Punjabi, North American Indian, &c., not to mention Icelandic, various German dialects, and Continental Scandinavian. However, no Semitic languages do this, although there is an increasing use of active participles. In Egyptian ³r ‘on’, r ‘towards’, and m ‘in’ are all used with verbal nouns to express a progressive. Paradoxically, the Breton progressive is much closer in force to the English progressive than the Welsh or Scottish Gaelic periphrastic constructions, which have become a general cursive (imperfective) which freely allows statives (cf. Hewitt 1986; 1990)). Indeed, there appears to be a general tendency in many languages for a parallel evolution of simple tense > specialized uses, and progressive > general imperfective.   13. Periphrastic do: he does singing. This conflates at least three distinct uses: (1) activity do with dynamic (non-stative) VPs: (Middle) Welsh, Breton; (2)  do with NPs: numerous languages; (3)  do as an empty auxiliary: North Welsh; Breton to avert V-1 in the affirmative (also English do with negative, interrogative). This is not typical of Semitic.

Hewitt CW3 3 iii 16.indd 418

27/05/2016 19:00:13

Hewitt

[ 419 ]

  14. Circumstantial clause and s pred (subordinating and ). This feature is typical of both Insular Celtic and Hamito-Semitic (in Berber it is possibly borrowed from Arabic): Breton gweled neus ahanon ha me o tond er-maes [seen he.has me.obj and I prog come. inf out], Arabic laqad ra’\-n¼ wa-’an\ ṭ\li‘ [ pfv he.saw-me and I coming.out] ‘he saw me as I was coming out’. The construction is syntactically coordinate (in both Celtic and Semitic, the order after ‘and’ is always SVO), but semantically subordinate; an adversive ‘although’ connotation is possible.   15. Nonfinite possible instead of finite main-clause verb. This is particularly prevalent in Welsh, followed by Hebrew (infinitive absolute), but not in Arabic. There are sporadic examples in Irish and Breton.   16. Word-initial phonetic changes (mutations), various syntactic functions. The highly grammaticalized Insular Celtic initial consonant mutations are hardly comparable to the Berber ‘construct state’ initial changes argaz > urgaz (w@rgaz) ‘man’, tamɣart > tmɣart ‘town’, which appear to be more akin to vowel contraction, cf. the Hebrew construct state forms described under feature 8. As formulated, this is rather too abstract a feature to be confidently attributed to substratal influence, and there are numerous instances worldwide of results of phonetic changes acquiring a grammatical function.   17. Idiomatic genitive kinship constructions: son of X. This is very productive in Semitic, cf. Iraqi Arabic abu cheg\yir [father.cs cigarettes] ‘(street) cigarette seller’. It is not typical of Brythonic; the few examples in Insular Celtic are in Irish: mac tíre [son land. gen ] ‘wolf ’.   18. Nominal clause (absence of copula). There is no copula in the present tense in Semitic, only in non-present (future, past) tenses. This is not the same as ellipsis of the copula in Insular Celtic, especially Welsh, and to a lesser extent Breton, in gnomic expressions.   19. Amplification of negative by noun after verb: French pas. Arabic dialects (Palestine and westwards) have developed a French-like circumfix m\ V-sh ( aarkuui [arkw(u̯ )uːi], in which the palatal glide *i̯ has been assimilated to articulation of the flanking labials, a development generally consistent with the observable phonetic tendencies of Tartessian.

¶outside the briga-Zone: Statvivs Arqvio 8E

(CIL II, 2990; Castillo et al. 1981, 27 — Monteagudo, Navarra). 12E

16

27/05/2016 19:31:00

[ 439 ]

Koch

In the absence of a compelling case for a different etymology, the well attested series Arcivs, &c., may be considered a delabialized variant of ArQVivs, including the following examples: ¶Celtiberian region. Arcea Alticon Aleonei filia

(Abásolo 1974, 48 — Lara de los Infantes, Burgos);

Arcea [---] Ambati f. (Abásolo 1974, 188 — Lara de los Infantes, Burgos); Arcea [---]avca Ambati Terenti f. (EE, VIII 150; Abásolo 1974, 160 — Lara de los Infantes, Burgos); Arceae Desiiae Cadaeci f. (Abásolo 1974a, 95 — Lara de los Infantes, Burgos); [A]rceae Dessicae Pat[er]ni f. (AE, 1983, 600; HEp, 4, 198 — Lara de los Infantes, Burgos); Arcea Elanioca Paterni f. (Abásolo 1974, 146 — Lara de los Infantes, Burgos); Arceae Longinae C. f. (matri) (CIL II, 5799; Abásolo 1974, 154 — Lara de los Infantes, Burgos); Arcea Plandica Maticvla(e) f. (CIL II, 2860; Abásolo 1974, 176 — Lara de los

Infantes, Burgos).

(HEp, 4, 1023; ERRBragança, 22; HEp, 12, 599 — Meixedo, Bragança); Arcivs Epeici f. Bracarvs (HAE, 992; HEp, 11, 647; HEp, 13, 647 — Vila da Feira, Aveiro); Tavrvs Arci f. Ammicivs (AE, 1971, 146; HEp, 7, 168 — Villar del Rey, Badajoz); Anivs Arci (ILER, 974 — Idanha-a-Velha, Idanha-a-Nova, Castelo Branco); Arcivs (AE, 1967, 157 — Idanha-a-Velha, Idanhaa-Nova, Castelo Branco); [Cae]nivs Arci (Almeida 1956, 154, nº 20 — Idanha-a-Velha, Idanha-a-Nova, Castelo Branco); L. Svlla Arci f. (AE, 1967, 157 — Idanha-a-Velha, Idanha-a-Nova, Castelo Branco); Tertvla Arci f. (HAE, 1172; Almeida 1956, 133 — Idanha-a-Velha, Idanha-a-Nova, Castelo Branco); Tovtoni Arci f. (AE, 1967, 144; HEp, 2, 770; HEp, 5, 989 ¶find spot: Idanha-a-Velha, Idanha-a-Nova, Castelo Branco); Arcivs (HEp, 5, 1055 —Ponte da Barca, Viana do Castelo); Arcissvs Arenier(i) f. (CIL II, 733; CPILC, 116 — Cáceres); Medamvs Arcisi f. caste[l]lo Meidvnio (CIL II, 2520; IRG IV, 130 — Cadones, Celanova, Ourense). ¶Western Peninsula. Arciae Mo[---]

Welsh arffed meant ‘lap’ as early as Trawsganu Cynan Garwyn, an archaic panegyric praising a chieftain of the 6th century AD, which includes the line cant armell ym arffet ‘a hundred arm rings in my lap’. Arffed may be compared with Latin arcu\tus ‘bent like a bow’, the source of Spanish arqueado ‘curved’, which can be used to describe bowed legs. So it seems reasonable, along the lines Rh½s proposed, to derive arffed from Proto-Celtic (PC) *arkweto- ‘(human body) flexed like a bow’.   The root in question, PIE *Haérkwos, means ‘bow and/or arrow’ and is otherwise limited to Italic and Germanic, cf. Gothic arƕazna ‘arrow’, Old English earh. PC *arkwo- ‘bow’ does not, except for arffed, survive in Goidelic or Brythonic. Middle Irish boga and Welsh bwa are derived from Old Norse. Old Irish saiget ‘arrow’, Welsh saeth, and Middle Breton saez are from Latin sagitta.   A different inherited word for ‘bow’ is seen in Middle Welsh gwarac ‘bow, crooked stick, yoke, loop, handle, half circle, curve’ corresponding to Middle Breton goarec ‘bow, arch’,3 and Middle Cornish gwarak ‘bow’, cf. Old Cornish guarac gl. ‘diploma’ < ?‘something coiled up’ (GPC sn. gwarag). Within Breton, goarec appears secondary to the adjective goar ‘bent, curved’. However, if Breton goar is the cognate of Welsh gßyr ‘bent, crooked, curved’ (as according to Jackson 1967, §623), Middle Welsh gwarac could not be a straightforward suffixed derivative of gßyr. In any event, the etymology is uncertain and the breadth of comparative evidence takes us back only as far as Common Brythonic, thus falling short of indicating a second word for ‘bow (and/or arrow)’ as old as Proto-Celtic let alone Proto-Indo-European. A second Breton word (in addition to saez) for ‘arrow’ is bir, which is probably related to ber ‘roasting spit’, cognate with Welsh bêr ‘dart, spear, slender shaft, spit, skewer’, Old Irish biur ‘stake, spit, point, spear’, Latin ver~ ‘spit, dart, javelin’, Avestan grava ‘staff ’ < PIE *gwéru ‘spear, spit’. Arrows and archer’s wrist guards were characteristic components of the Beaker package of Copper Age western Europe. After the Early Bronze Age, there is little trace of archery 3

I’m grateful to Steve Hewitt and Guto Rhys for drawing this word to my attention.

Koch CW3 16 iii 16.indd 439

27/05/2016 19:31:00

[ 440 ]

xv. phoenicians in the west

]]

15.4–6 Bows and arrows as depicted on Iberian LBA warrior stelae. The bottom image (15.6) includes all known examples. (from Harrison 2004)

C 76. ECIJA III, Sevilla

C 79. MONTEMOLÍN, Sevilla

Koch CW3 16 iii 16.indd 440

in later prehistoric Britain or Ireland (likewise most of northern and eastern Gaul). So it is not surprising that later borrowings and coinings should prevail in the vocabulary for archery in the attested Insular Celtic languages. As Parker Pearson (2005, 28) has argued, this gap can be explained as an effect of the rise of the ‘pan-European warrior cult’ (during the Middle Bronze Age), which valued hand-to-hand engagements between high-status combatants and equipment suitable for such activity: leather shields, leather armour, spears, rapiers, and later swords. Thus exceptionally, in the Iberian Peninsula, bows and arrows are one of the recurrent motifs of the Late Bronze Age warrior stelae, where they occur together with other items of the highstatus warrior’s panoply: shields, spears, swords, helmets, chariots, brooches, mirrors, combs (Figures 15.4–6). There are eleven published examples of images of bows (always shown with arrows) from the corpus of Late Bronze Age warrior stelae of the south-western Peninsula (Harrison 2004, 144, 146). Díaz-Guardamino (2010) records two bow representations on the Early to Middle Bronze Age alentejano stelae, and she views these as survivals of indigenous archery in the Peninsula

27/05/2016 19:31:09

Koch

[ 441 ]

15.7 Detail of the bow and arrow from the LBA ‘warrior’ stela from Montemolín, Sevilla (see 15.5) superimposed on a photograph of the prominent hill at Monreal de Ariza (Arcobriga), Zaragoza, Spain, viewed from the ruined Roman town (Source: http:// aeternitasnumismatics.blogspot. co.uk/2012/03/ la-ciudad-celtibera-dearcobriga.html)

during the Copper and Early Bronze Ages rather than as an accoutrement reintroduced from elsewhere. Arrowheads in Bronze Age contexts are common finds in all regions of the Peninsula (Kaiser 2003), including, for example, the bronze weaponry of the Ría de Huelva deposition of 1052–898 cal BC (Ruiz-Gálvez 1995b). Burgess (2012, 145f ) argued that a completely different style of fighting prevailed in Late Bronze Age Iberia and Aquitania, in which the mounted archer dominated the battlefield. The Hispano-Celtic place-name Arcobriga probably also contains Western PIE *Haérkwos. The second element of Arcobriga is the very common, in fact diagnostic, Celtic form meaning ‘hill, hillfort’. The first element has sometimes been equated with PIE *H2ŕ̥tḱos ‘bear’, attested in Celtic as Welsh arth, the Old Irish personal name Art, and the Gaulish goddess name DEAE ARTIONI (dative). As the Celtic forms otherwise point to PC *arto-, Hispano-Celtic arcoseems an unlikely development, especially in view of Artivs (CIRG II, 87; HEp, 6, 765 — Achacan, Pontevedra), a man’s name, probably meaning ‘bear-like’ or ‘bear-hunter’.   There were three places in the ancient Iberian Peninsula called Arcobriga. The location of only one is certain, Monreal de Ariza in the province of Zaragoza, within the ancient Celtiberian region. An impressive hill overlooks the ruined Roman town there, although the expectation that this hill had been the site of a pre-Roman hillfort has not been borne out. Viewed from the site of the ancient town, the silhouette of the hill on the horizon, especially when this is compared with the images of bows from the LBA stelae, leaves little doubt that the place-name means ‘bow-shaped hill’ and has nothing to do with bears (Figure 15.7). Phonologically, what is involved in the development of PC *arkwo- > arco-, as in Latin, is that the distinction between the labiovelar [kw] and simple velar [k] became neutralized before the round vowel. This is a very common and unremarkable phonological change, for which there is abundant evidence in the Palaeohispanic corpus.

Koch CW3 16 iii 16.indd 441

27/05/2016 19:31:09

[ 442 ]

xv. phoenicians in the west

]]

  In sum then, taking the linguistic and archaeological evidence together, we can see that a disparity that developed in the post-Beaker cultures of Atlantic Europe during the 2nd millennium BC can be correlated with a disparity in the survival of Western PIE vocabulary across the Celtic dialects. The old word for ‘bow’ was kept where the object itself survived and where it also retained high status. Elsewhere, the culture changed in this regard and the language had no choice but change accordingly.   It appears, therefore, that some at least of the substantial differences in the onomastic stock of the Ancient Celtic languages can be traced back this far. In other words, both the cultures and languages of Atlantic Europe had become different in some details at the period when they were still in close contact and sharing cultural innovations, before the arrival of Phoenicians and the western Peninsula’s departure from the Atlantic Bronze Age c. 900 BC. The Bronze–Iron Transition and protohistory In the western Iberian Peninsula the Bronze–Iron Transition can be understood as a change from an Atlantic Bronze Age to a Mediterranean Iron Age or, noting regional disparities, Orientalization in the south and isolation in the north (Armada 2011, 176). As both a microcosm and a centre of far-reaching cultural and economic influence, evidence from Huelva witnesses the transition taking place between the Ría de Huelva phase (1050– 950/900 BC)4 and the Baiões–Plaza de las Monjas phase (950/900–825 BC)5, adopting the framework of Torres (2008). The heavy Tyrian presence, whose arrival defines the latter phase, signals a decisive cultural reorientation away from ALBA. The case set out here is that this shift can be correlated on the sociolinguistic plane with a discontinuity in the lingua franca of the Atlantic zone. The written evidence that appears at the threshold of literacy in the Iron Age implies that this lingua franca corresponds to the reconstructed stage usually termed Proto-Celtic. On the other hand, the name Arquius and the evidence for Bronze Age archery discussed in the previous section imply that we should not imagine that there had been sociocultural uniformity across the entire Atlantic zone during any part of the Bronze Age nor that its lingua franca remained devoid of regional variation.   The cultural and linguistic situations in Atlantic Europe at about 1000–800 BC have been clarified by new information in the following areas: 1. Huelva That the key events of the Iberian Bronze–Iron Transition occurred earlier than previously recognized is now apparent from evidence from the Plaza de las Monjas site in Huelva town. This material can be dated by both ceramic typology and absolute chronology.6 In 4 5

6

Ría de Huelva (the estuary site) — Atlantic Bronze Age élite ritual deposition site: 1052–898 cal BC (95.4% probability) Plaza de las Monjas (the Huelva town site) — Tyrian emporium: 2755 + 15 BP; 930–839 cal BC (95.4% probability) Results obtained averaging the layer: 2755 + 15 BP; 930–839 cal BC (95.4% probability). Mederos (2006) allows a higher chronology, beginning in the earlier 10th or late 11th century BC. Nuñez (2008) and Gilboa (2013) fully review the evidence and cautiously favour lower chronologies.

Koch CW3 16 iii 16.indd 442

27/05/2016 19:31:09

Koch

[ 443 ]

the revised scheme devised by Torres (but also consistent with González de Canales et al. 2004; 2006; 2008; 2010; Mederos 2006; Nijboer & van der Plicht 2006; 2008; Aubet 2008), three chronological phases are recognized for the western Peninsula, distinguishable by the type and degree of influence from the eastern Mediterranean: ‘Ría de Huelva (1050– 950/900 BC), Baiões–Plaza de las Monjas (950/900–825 BC) and [Phoenician] Colonial (c. 825 BC onwards). ‘Each of them shows an increasing articulation of western Iberia in the Mediterranean trading networks spreading from the Levant to the Western Mediterranean’ (Torres 2008, 135). For present purposes, it is useful to insert a ‘phase zero’ at the head of this scheme, a sub-Mycenaean hiatus (1200/1150–1050 BC), during which there was little recognizable contact between the eastern Mediterranean and Atlantic Hispania. The Ría de Huelva phase (named for the Bronce Final IIIA deposition in the Odiel estuary) is still predominantly Atlantic Bronze Age in character (Ruiz-Gálvez 1995b); however, following the hiatus between the fall of Mycenae, c. 1200 BC and c. 1050 BC, eastern objects are in phase 1 once again entering the Atlantic networks with Cyprus as the most prominent eastern terminus. In the following Baiões–Plaza de las Monjas phase, the Phoenician city state of Tyre becomes dominant as the source of eastern influence. The Plaza de las Monjas material has mostly been interpreted as reflecting a coastal emporium, rather than a colony. The period has traditionally been counted as the last of the Bronze Age—in alignment with the LBA chronologies of Gaul, central Europe, and the British Isles—but is more accurately understood as Bronze–Iron Transition or Early Iron Age, as recognized indirectly by Burgess who wrote recently about the earlier Ría de Huelva find: ‘For this writer the [Ría de] Huelva find, whatever its nature, is a one-off, like Blackmoor, but rather than a snapshot of what was to come in relationship to what had been, it is indicative of what was coming to an end, the last gasp of the Atlantic Late Bronze Age in Iberia’ (Burgess 2012, 148).

Full-blown Phoenician colonies on Iberia’s Atlantic coast (most notably Gadir, Cádiz today) are characteristic of the following phase 3 from the late 9th century BC. In much of the new work stimulated by the Plaza de las Monjas finds, there is general acceptance of the likelihood that Huelva was the core, or an important part of the core, of the polity called Ταρτησσος7 by the Greeks. The new discoveries have also brought fresh credibility to the long-debated identification of Ταρτησσος with Phoenician tršš on the inscribed stela from Nora, Sardinia, c. 825 BC and the biblical Tarshish. The new dates for Phoenician material in Huelva town bring matters intriguingly close to the traditional dates of Hiram I of Tyre 971–939 BC and Solomon of Israel 961–922 BC, whose joint venture is described in connection with luxury-laden ‘ships of Tarshish’ in the Old Testament (González de Canales et al. 2004; 2006; 2008; 2010; Mederos 2006; passages in Freeman 2010). Note that López-Ruiz (2009) also cautiously favours Ταρτησσος=Tarshish along a different line of argument.

7

With foreign names recorded in Greek script, the accents are omitted here, as they are not likely to reflect the nature or position of the accent in the source language.

Koch CW3 16 iii 16.indd 443

27/05/2016 19:31:09

[ 444 ]

xv. phoenicians in the west

]]

2. Reassessments of the Atlantic Bronze Age phenomenon The material culture, especially metalwork, of the Atlantic region (Ireland, Britain, Atlantic Gaul, and the western Iberian Peninsula) has lately been reconsidered in three ways relevant to the present subject. First, after a period of critical scepticism, it has now been generally reaffirmed that ALBA represents a meaningful cultural region in the period c. 1250–800 BC (e.g. Burgess & O’Connor 2008; Gerloff 2010; Milcent 2012). Second, the Huelva swords are now recognized as a distinct type, earlier by a century than the best known Carp’s Tongue types found in northern Gaul and southern Britain (Brandherm 2007). Third, partly on the basis of the reinterpretation of the Huelva swords, but also other metalwork types, it has come to be recognized that ALBA ended earlier in the Iberian Peninsula than in Atlantic Gaul, Britain, and Ireland and that this was due to a precocious Early Iron Age stimulated by contacts with the eastern Mediterranean, as summed up by Burgess and O’Connor: LBA4 in Britain, Llyn Fawr, 850/800–700 [BC], Hallstatt C in central-European terms, Hallstatt 1 or First Iron Age in France, has no equivalent in Iberia, where the beginning of Phoenician colonisation changed everything. (2008, 41; cf. Cunliffe 2008, 299–300)

We can now modify this statement to note that everything had begun changing already at the time of the Tyrian emporium in Huelva, during the Baiões–Plaza de las Monjas phase of 950/900–825 BC.   The Ría de Huelva or estuary hoard provides a microcosmic contrast of cultural reorientation at one important centre during successive stages. We go from exchange dominated by long-distance maritime contacts northwards along the Atlantic Façade to an age dominated by long-distance contacts with the eastern Mediterranean. The end result, that Phoenician influence had the effect of pushing the Atlantic Bronze Age out of the Iberian Peninsula, is clear enough. Why this was the result is less obvious. It may have been a simple matter of superior ships and navigational technology, so that the indigenous mariners of Atlantic Europe simply lost control of the seaways. The economic dimension—including huge increases in the extraction, production, and overseas exchange of metals—can hardly be over-estimated. More elusive factors of cultural preference and identity may also have been at work, a dimension that will be considered below. 3. Linguistic phylogeny The Celtic languages that have survived till today are uncontroversially subdivided into Goidelic and Brythonic branches. However, when the extinct Celtic languages of Ancient Europe and Anatolia are brought into consideration, the picture becomes more complicated, and many different configurations have been proposed for the Celtic family tree. The intractability of this problem can be taken as a broad indication that much of the early history of the Celtic peoples was spent as speakers of mutually intelligible dialects in contact. In fact, what we know of the history and archaeology of early Gaul, Britain, Ireland, and northern Italy is enough to confirm this.   Nonetheless, in recent work, there is an emerging consensus regarding a primary divergence within the Celtic languages, one that looks more like a true split caused by a break-down in contact, rather than a pattern produced by overlapping waves in a continuum.

Koch CW3 16 iii 16.indd 444

27/05/2016 19:31:09

Koch

[ 445 ]

This primary division is between Celtiberian/Hispano-Celtic on the one side, with Gaulish, Brythonic, and Goidelic forming a commonality on the other (McCone 1994, 65; 2008, 38; Watkins 1999, 540; Isaac 2005, 198; De Bernardo-Stempel 2006, 44–7).8 The earlier view (e.g. Lejeune 1972; Schmidt 1986)—that Celtiberian, Gaulish, and Brythonic were more closely related to each other than to Goidelic—was based on the idea that the reflex of the Indo-European syllabic nasals *ṇ and *ṃ had been Goidelic *en and *em, rather than an and am, but this is no longer widely believed (McCone 1991; 1994, 70, 77–8; Isaac 2007a, 63).   To summarize, a decisive turning away from the Atlantic world and towards the eastern Mediterranean occurred between the Huelva estuary find (1050–950/900 BC) and the town find (950/900–825 BC), so around 950/900 BC. This change of cultural orientation near the protohistoric horizon is reflected in both the archaeological and historical linguistic evidence sets, the departure of the western Iberian Peninsula from the Atlantic Bronze Age and the fragmentation of Proto-Celtic into two separate languages. 4. The SW (Tartessian) inscriptions and Western Hispano-Celtic In the recent work mentioned above, which agrees in where it places the first split in Celtic phylogeny, different terms are used. For example, McCone (2008) calls the GaulishBrythonic-Goidelic commonality ‘Gallo-Insular Celtic’. Most refer to the other sibling simply as ‘Celtiberian’. However, in the model of De Bernardo Stempel (2008) it is ‘Celtic in the Iberian Peninsula’, which then gives rise to Celtiberian and other unnamed descendants. This is better. Celtiberia and the Κελτιβηρες belong only to the eastern Meseta, which is also where the Celtiberian language is found, both as onomastics and written continuously in inscriptions. The Celtiberian inscriptions were produced in a version of the northern Palaeohispanic script in the last centuries BC, with later usage tending to show a switchover to Roman script. However, Ancient Celtic onomastics are found densely distributed, running northwards and westwards from Celtiberia all the way to the Atlantic coast—from south Portugal to Galicia in the north-west and almost to the Pyrenees in the north-east. It is useful, therefore, to distinguish between Celtiberian, meaning the Ancient Celtic language of eastern Meseta, and Hispano-Celtic, as including Celtiberian, but also the Celtic found in other parts of the Peninsula.   There is no geographic discontinuity between Celtiberian and the greater Hispano-Celtic zone spreading north and west. The dialectal unity of Hispano-Celtic cannot simply be taken for granted. However, the frequent recurrence of names and name elements within Celtiberia and outside of it to the north and west—for example, but by no means only, the place-names in *-brig\—contributes to the general impression that this is a dialect continuum, rather than the remains of fully distinct languages with separate histories. Future detailed work will illuminate further this copious and fascinating material, work that is complicated by challenges in differentiating between Western Hispano-Celtic and Lusitanian as isolated onomastic items 8 The model of Schrijver (2012) is somewhat different. There is again a commonality of Gaulish, British, and Goidelic, but the other side of the split is not Hispano-Celtic, but a Southern Celtic that includes both (Early) Lepontic and Celtiberian. In this model, Late Lepontic comes under the influence of Gaulish and so shifts allegience, which is not an unlikely scenario in view of the archaeological and historical evidence for migrations crossing the Alps north to south from the 5th century BC.

Koch CW3 16 iii 16.indd 445

27/05/2016 19:31:09

[ 446 ]

xv. phoenicians in the west

]]

in short epigraphic texts (cf. Luján 2007; Prósper 2008; 2010; Wodtko 2009; 2010).   It is probable that the early change of PC *-χt- > -t- represents an innovation shared by all Hispano-Celtic, distinguishing this group as a family from the other Ancient Celtic languages. However, some ambiguity has been seen in the attestations. One of the most common Palaeohispanic personal names Ambatos, together with its numerous variants, is probably an example of this feature. Instances are widely attested from Galicia across the northern central Peninsula and western central Peninsula to the margins of Celtiberia (Vallejo 2005, 134–40). Examples include the following: ¶CELTIBERIAN REGION. Ambata

(Abásolo 1974, 99; Albertos 1975a, — Lara de los Infantes, Burgos);

Ambatae [---] Segei f. (Abásolo 1974, 194 — Quintanilla de las Viñas, Burgos); Ambatae Aioncae T[---]ti f. (Abásolo 1974, 155 — Lara de los Infantes, Burgos); Ambatae Aioncae Lovgei f. (Abásolo 1974, 185 — Lara de los Infantes, Burgos); Ambata Albeavca ? Segovetis f. (CIL II, 2855; Abásolo 1974, 18 — Iglesia Pinta, Burgos); Ambata Betvca Ambati f. (Abásolo 1974, 60 — Lara de los Infantes, Burgos); Ambata Caelica Cai f. (Abásolo 1974, 24 — Iglesia Pinta, Burgos); Ambata Cor(---) (HEp, 10, 88 — Belorado, Burgos); Ambatae [D]essic[a]e Rvfi [f.] (socerae) (AE, 1983, 600; HEp, 4, 198 — Lara de los Infantes, Burgos); Ambatae Medicae Verati f. (HEp, 10, 81 — Belorado, Burgos); Ambatae Medicae Placidi f. (Abásolo 1974, 81; HEp, 4, 199 — Lara de los Infantes, Burgos); Ambata Paesica Argamonica Ambati vxor (CIL II, 2856; Abásolo 1974, 177 — Lara de los Infantes, Burgos); Ambata(e) Peditage Ambati (Reyes 2000, 24; HEp, 10, 87 —Belorado, Burgos); Ambatae Plandidae (EE, VIII 172 — Pancorbo, Burgos); Amb[a]tae Veniaenae Valeri Crescenti[s] f. (CIL II, 2878 = CIL II, 2882; Abásolo 1974, 214; HEp, 5, 153; HEp, 6, 172 — San Pedro de Arlanza, Hortigüela, Burgos); [Ca]lpvrniae Ambatae Lovgei f. (AE, 1980, 587 — Lara de los Infantes, Burgos); Semproniae Ambatae Celtiberi (Abásolo 1974, 209 — San Millán de Lara, Burgos); Ambatae Terentiae Severi f. (CIL II, 2857; Abásolo 1974, 212 — San Pedro de Arlanza, Hortigüela; Burgos); Valeria Ambadae (CIL II, 2909 — Villafranca-Montes de Oca, Burgos); [---] Ambati l. (CIL II, 2884; Abásolo 1974, 141 — Lara de los Infantes, Burgos); [A] mbatvs (CIL II, 2790; Palol-Vilella 1987, 219 — Peñalba de Castro, Burgos); [A]mbato Alebbio [B] odani f. (Reyes 2000, 5 — Belorado, Burgos); Ambato Bvrgae Segili f. (HEp, 10, 84 — Belorado, Burgos); Ambatvs Vemenvs Ati f. (Abásolo 1974, 55 — Lara de los Infantes, Burgos); Ambato Virovarco (HEp, 9, 246 — Ubierna, Burgos); Arcea [---] Ambati f. (Abásolo 1974, 188 — Lara de los Infantes, Burgos); Arcea [---]avca Ambati Terenti f. (EE, VIII 150; Abásolo 1974, 160 — Lara de los Infantes, Burgos); Cabedvs Seggves Ambati f. (CIL II, 2863; AE, 1977, 447 — Carazo, Burgos); Madicenvs Calaetvs Ambati f. (CIL II, 2869; EE, VIII 154; Abásolo 1974, 108 — Lara de los Infantes, Burgos); Secontio Ebvren[i]q(vm) Ambati f. (Reyes 2000, 18 — Belorado, Burgos); Segilo Aespanco(n) Ambata[e] filio (HEp, 10, 83 — Belorado, Burgos); Talavs Caesarivs Ambati f. (Abásolo 1974, 13 — Hontoria de la Cantera, Burgos); Metelio Rebvrro Ambati f. (HEp, 10,

102 — Belorado, Burgos). ¶CENTRAL REGION. Ambat[o] (HEp, 4, 103; ERAv, 30 — Ávila); Ambato (HEp, 4, 72; ERAv, 11 — Ávila); Ata Ambaticorvm Hirni f. (HEp, 10, 8; ERAv, 142 — Candeleda, Ávila); Vernacvlvs Ambatic(vm) Modesti f.[ ---] (HEp, 1, 79; HEp, 9, 83; ERAv, 143 — Candeleda, Ávila); Acceti Cariqo Ambati f. (HEp, 2, 618; ERSg, 5 — Coca, Segovia); Ambat(a) (CIL II, 94*/5320 — Talavera de la Reina, Toledo). ¶WESTERN PENINSULA. Fvsci Cabedi Ambati f. Vadiniensis (CIL II, 2709; ERAsturias, 51 — Corao, Cangas de Onís. Asturias); Macer Ambati f. Obisoq(vm (Roso de Luna 1904, 127 — Casas de Don Pedro, Badajoz); [---] Ambati f. (HEp, 1, 668; ERRBragança, 95; HEp, 12, 587 — Donai, Bragança); Ambatvs (CIL II, 738, 739; CPILC, 44 = CPILC, 45; HEp, 9, 248 — Arroyo de la Luz, Cáceres); Ambatvs (CPILC, 50; CILCC I, 75 — Arroyo de la Luz, Cáceres); Ambatvs Pe[l]li (CIL II, 853; CPILC, 392 — Plasencia, Cáceres); A[n]dercia Ambati f. (AE, 1978, 393; AE, 2006, 625; HEp, 15, 92 — Monroy, Cáceres); Arc[o]ni Ambati f. Camalicvm (CPILC, 660 = CPILC, 803 — Villar del Pedroso, Cáceres); Camira Ambati (CIL II, 623; CPILC, 527 — Trujillo, Cáceres); Coria Ambat(i) f. (CPILC, 146 — Cáceres); Irinevs Ambati f. (CPILC, 367 — Pedroso de Acim, Cáceres); Ambatvs (ERCan, 8 — Luriezo, Cantabria); Ambati Pentovieci Ambatiq. Pentovi f. (ERCan, 8 — Luriezo, Cantabria); Tillegvs Ambati f. Svsarrvs Ɔ Aiobaigiaeco (IRLugo, 55; HEp, 8, 334 — Esperante, Folgoso do Caurel, Lugo); Ambati Bvrili Tvroli f. (HAE, 1367 — Yecla de Yeltes, Salamanca); Ambatvs Divli f. (HEp, 4, 962 — Hinojosa de Duero, Salamanca); Cavrvnivs Ambati Cavrvnicvm

Koch CW3 16 iii 16.indd 446

27/05/2016 19:31:09

Koch

[ 447 ]

(Albertos 1975a, 18. nº 196 — Yecla de Yeltes, Salamanca); [A]mbatvs (AE, 1972, 287 — Salamanca); Ambatvs Pintovi (HAE, 1327 — Saldeana, Salamanca); Ambatvs Tancinili f. (HEp, 2, 617; HEp, 5, 677 — San Martín del Castañar, Salamanca); Clovti[a] Ambati filia (HAE, 1265 — Hinojosa de Duero, Salamanca); Ianva Ambati (HAE, 1253 — Cerralbo, Salamanca); Mentina Ambati f.(CIL II, 5036; HEp, 10, 513 — Yecla de Yeltes, Salamanca); Ambati Arqvici (HEp, 11, 361 — Barruecopardo; Salamanca); Ambato Arqvi f. (ERZamora, 114; CIRPZ, 241 — Villalcampo, Zamora); Avelco Ambati f. (HAE, 920; CIRPZ, 246; ERZamora, 29 — Villalcampo, Zamora); Pintovio Ambati (ILER, 2333; ERZamora, 210; CIRPZ, 271 — Villalcampo, Zamora; Ambato (HEp, 18, 486 — Villardiegua de la Ribera, Zamora); Ambatia (HEp, 18, 488 —Villardiegua de la Ribera, Zamora); ¶S.W. inscriptions. ]anbaatiia iobaa[ (J.16.2 — San Salvador, Ourique, Beja) can be provisionally interpreted as nominative |Amba(χ)tiā i̯ōamā| ‘the youngest daughter of Amba(χ)tos’ or more generally ‘the youngest kinswoman or female descendant of Amba(χ)tos’. ¶outside the briga-Zone. Ambata Appae f. (CIL II, 2950 — Contrasta, Álava); Ambato (HAE, 2522 — Angostina, Álava); Ambatvs Serme f (CIL II, 2951 — Contrasta, Álava); Amba[t]vs Plendi f. (CIL II, 2948 — Eguilaz, Álava); [A]mbatvs [A]ravi f. (HAE, 2571; HEp, 4, 1 — Urabáin, Álava); [---] cvs Ambati f (HAE, 2563; HEp, 4, 11 — San Román de San Millán, Álava); Elanvs Tvraesamicio Ambati f(ilivs) (CIL II, 5819; Albertos 1975a, 13. nº 74 — Iruña, Álava); Segontivs Ambati Vecti f. (CIL II, 2956 — Contrasta, Álava); Ambata (Castillo et al. 1981, 48 — Gastiáin, Navarra); Doitena Ambati Celti f. (EE, VIII 167; Castillo et al. 1981, 53 — Marañón, Navarra); Doiterv[s ---] Ambati f. (Castillo et al. 1981, 55; HEp, 5, 623 — Marañón, Navarra); Ivnia Ambata Viro[ni] f. (CIL II, 5827; Castillo et al. 1981, 45 — Gastiáin, Navarra); Porcia Ambata Segonti filia (CIL II, 5829; Fita 1913b, 565, nº — Gastiáin, Navarra); Ambatv[s] (HAE, 185; Alföldy 1975, 337 — Tarragona); L. Postvmivs Ambatvs (CIL II, 4024 — Villar del Arzobispo, Valencia).

  The standard etymology is to see this name as the cognate Gaulish AMBACTVS, which appears on the coins of the Mediomatrici (Allen 1980, 206); the word is used by Caesar for clients of Gaulish chieftains (Bello Gallico 6, 15); it is cognate with Welsh amaeth ‘ploughman, &c.’, deriving from *ambi-aχtos ‘one who is sent around’ (cf. GPC sn. amaeth; Matasović 2009, 32).9 The recent discovery of piχte in the Gaulish inscription of Rezé/Ratiatum, where it clearly means in context ‘as a fifth’ (Lambert & Stifter 2012), throws new light on - χt- > -t- as a diagnostically Hispano-Celtic innovation. It had previously been thought that the numerous Palaeohispanic personal names based on Pinto-, Pento- and similar (e.g. Ambatvs Pintovi (Saldeana, Salamanca; HAE, 1327) were Indo-European, but not Celtic, because Gaulish pinpetos, OW pimpet, OIr. coíced all reflect a notional IE *penkwetos, not the more basic *penkwtos (Villar 1994). However, in light of the discovery of Gaulish piχte, Prósper (2015, 25) concludes: . . . the names *p¼χto- and *pinto- could be the outcome of the same structure in different Celtic 9

However, Vallejo (2005, 140) proposes that Hispanic Ambatus, &c., possibly has a different etymology, with the same suffix found in the Hispanic names Boutati, Celtiatus, Tongati, Veniati, and Viriatus. But these forms are probably not comparable with Ambatus, as they are suffixed nouns, whereas *ambi- is a preposition. Vallejo also mentions the variation retukenos in Celtiberian Palaeohispanic script, contrasting with the RECTVGENVS found twice in Roman script in the Celtiberian region (Raybould & Sims-Williams 2007, 69–70), which he sees as implying that PC χt was sometimes retained in Celtiberian. However, this spelling was probably influenced by the correct perception that the first element of the name was related to the common Latin word rectus ‘correct, &c.’ and had a similar meaning. Note that Hispanic RETVGENVS probably also occurs, though the reading is in doubt (ibid. 2007, 70). Incidentally, Raybould & Sims-Williams list and map the latter name as being from Almodôvar, Portugal; it should be Almodóvar del Río, near Córdoba, Spain (ancient Carbula; Hernando 2007). RETVGENVS is thus one of the south-easternmost Celtic personal names attested in the Iberian Peninsula. The inscription DVATIVS APINIS F. BANDI TATIBIIAICVI VOCTO (sic) SOLVI (sic) (HAE 980; Encarnação 1975, 136; Blázquez 1962, 534; AE 1961, 341; Vallejo 2013 — Queiriz, Fornos de Algodres, Guarda), with VOCTO for VOTO, shows that CT could be pronounced [t] in parts of the Peninsula.

Koch CW3 16 iii 16.indd 447

27/05/2016 19:31:09

[ 448 ]

xv. phoenicians in the west

]]

dialects . . . The Gaulish result with compensatory lengthening is otherwise attested in Insular Celtic at least for words containing an original velar: OIr. técht ‘frozen’ < PCelt. *tanχto-, etc.

In other words, it now appears likely that the following Palaeohispanic personal names are examples of Proto-Celtic *kwɪnχto- ‘5th’ > Hispano-Celtic *pɪnto-, thus illustrating a HispanoCeltic innovation distinct from attested developments in Gaulish, Goidelic, and Brythonic.10 The South-Western inscriptions interpreted as nativism

If we call Italy ‘south-central Europe’, then the South-Western (SW) inscriptions represent the earliest written indigenous language in western Europe. Form and context indicate that many of them, at least, are funerary stelae.11 The densest concentration is in south Portugal with a thinner scatter over south-west Spain. There are 90–100 examples known today (Correia 2009). Most of the SW inscriptions are on stone, but on the map (Fig 15.8), some examples of inscriptions in SW script on ceramics are included: Abul, Moura, Niebla/ Ilipula (Correa & Toscano 2014), and Doña Blanca.   Recent work in connection with the Early Iron Age necropolis at Medellín confirms that there was a shared name for the literate group occupying a vast area: about 300 km from Sagres point to the upper Guadiana (Almagro-Gorbea et al. 2008). Konisturgis, meaning ‘town of the Kονιοι’, was most probably the pre-Roman name of Medellín. According to the Greek geographer Strabo, Konisturgis was the most famous town of the Κελτικοί, i.e. the ‘Celts’ of the south-western Iberian Peninsula (Alarcão 2001). Kονιοι and its variant Κυνητες Kun{tes are probably Celtic names, related to Welsh cßn ‘dogs’ (De Bernardo Stempel 2008). That word was common as an epithet for heroes, like the Irish Cú Chulainn, the legendary ‘Hound of Ulster’. Cynwydion occurs as the name of the war-band of Dark Age Strathclyde, essentially the same name as Κυνητες (Koch 2013a, 190–92; cf. Charles-Edwards 1978, 66–8). According to Herodotus, writing in the mid 5th century BC, the Κυνητες were the westernmost people of Europe and neighbours of the Κελτοί. §2.33. For the Nile does, after all, flow from Libya—indeed it bisects it—so my own conjecture, based on analogy between what is apparent and is unknown, would be that its length is equivalent to that of the Ister [Ἴστρος ‘Danube’]. That river, which splits Europe in two, has its source at the city of Pyrene [Πυρηνη], in the land of the Celts [Κελτοί], who dwell beyond the Pillars of Heracles, and are the neighbours of the Cynetes [Κυνησιοι], the people who live closer to the setting sun than any other in Europe. (trans. Holland 2013, 121; cf. Koch 2014a) 10 For further discussion of the Palaeohispanic names in Pent- and Pint- together with an extensive collection of attestations based on Vallejo (2005) and the Hesperia database, see Koch (2015). 11 Ruiz-Gálvez (2013) argues that the SW stelae were probably not primarily funerary in their original function. Rather, like the Late Bronze Age ‘warrior’ stelae that preceded them, they were meant to express the immortalization and divinization of heroes. This theory is, incidentally, consistent with what I have suggested is the meaning of the Tartessian epigraphic formula, which appears in its fullest form as NAMES OF DECEASED+uarbaan tee-ro-baare baa naŕkeentii ‘. . . has carried away [the named deceased] to the highest place/being/state (feminine) so they now lie under [together]’ (Koch 2013b).

Koch CW3 16 iii 16.indd 448

27/05/2016 19:31:09

Koch

[ 449 ]

§4.49. . . . the Ister [[Ἴστρος ‘Danube’] flows through the whole of Europe: its source lies in the land of the Celts [Κελτοί], a people who live further from us and closer to the setting sun than any other except for the Cynetes [Κυνητες], and only after the Ister has flowed through the whole of Europe, and reached the very edge of Scythia, does it empty into the sea. (trans. Holland 2013, 281; cf. Koch 2014a)

  The SW inscriptions are notoriously hard to date. But this too has changed as a result of the Medellín excavations. The site produced a contextualized reused SW stela datable to 650– 625 BC (stela J.57.1 = Medellín T1, 86H/EN12–112). The stela not only uses SW script, but is clearly in the same language as the stelae of the less-urbanized Algarve 200+ kilometres to the west. J.57.1 reads: ]lokoon keeloia naŕkee[. . . The text thus shares items of vocabulary (lokoon and naŕkee) with, for example, the ‘Fonte Velha 6’ (J.1.1) inscription from near the end of the mainland in Sagres.13 Inscriptions from both sites show a standardized form of the SW script, the a-be-ka-tu in right-to-left orientation, in use from the western Algarve to the upper Guadiana. Medellín’s excavators argued that this stela belonged to a mature and widespread second stage of literacy. In other words, Medellín reflects the middle—not the beginning—of the tradition. There was also Tartessian graffiti on pottery from Medellín of the late 7th and 6th centuries BC.14 These show further developments in the writing system (Almagro-Gorbea 2004, 14–16; 2008, 766–71).   As to the beginning of indigenous literacy, de Hoz (2010, 516) concludes that the invention of the SW script cannot be later than the first half of the 7th century BC. This would be hard to dispute. Current thinking has generally moved away from the idea that the Greek alphabet contributed to the initial formation of the script (cf. Hoz 2010, 495–500), thus eliminating one constraining factor on the earliest possible date for the script’s invention. This milestone could hardly have been achieved during the hiatus or near-hiatus in contact between the Iberian Peninsula and eastern Mediterranean of the Sub-Mycenaean phase 1200/1150–1050 BC. But even after eliminating this phase—and almost surely any period earlier than that—we are still left with a wide range of possibilities. There was contact with literate maritime cultures of the east (particularly those of Cyprus) during the Ría de Huelva phase (1050–950/900 BC), and it is not farfetched to suppose that Phoenicians and/or Canaanites were literally on board by this time. The following Baiões–Plaza de las Monjas phase (950/900–825 BC) is defined by the establishment of a Tyrian emporium at Huelva, which of course saw mariners, artisans, and merchants with knowledge of the alephat (alphabet lacking vowels). landing in the west. By c. 825 BC onwards there were fullblown Phoenician colonies founded in the southern Peninsula, including Gadir/Cádiz no 12 This date range corresponds to the second generation using the necropolis in the demographic scheme of Almagro-Gorbea (2010). Stela J.57.1 is an instance of re-use in this necropolis of the Early Iron Age, a practice attested elsewhere for both the stelae with SW inscriptions and the so-called ‘warrior stelae’ of the Late Bronze Age. The burial in which the broken stela was re-used dates to 525–500 BC. The period of the stela’s primary use as an upright garve marker was worked out within the overall chronology of the Medellín necropolis. 13 If Indo-European, lokoon in the Medellín and Fonte Velha texts would formally suit an o-stem accusative singular or nominative/accusative neuter. In funerary contexts—and Fonte Velha like Medellín is a necropolis—we may compare Cisalpine Gaulish LOKAN ‘interment, funerary urn’, corresponding to VRNVM in the parallel Latin text at Todi (1st century BC), thus possibly as an item of Common Celtic funerary vocabulary. 14 Medellín T2, 86H/13–1 c. 625–600 BC; T3, 86/TP–1 c. 550–500 BC.

Koch CW3 16 iii 16.indd 449

27/05/2016 19:31:09

[ 450 ]

xv. phoenicians in the west

]]

later than c. 775 BC.   Similarities in form and distribution link the SW inscriptions with the corpus of c. 150 ‘warrior stelae’ of the Late Bronze Age, about 1250–900/800 BC, found widely across the south and west of the Peninsula (Celestino 2001; Harrison 2004; Díaz-Guardamino 2010; 2012). There are also instances of reuse and a transitional example (J.12.1 — Gomes Aires, Almodôvar, Beja). In the recent work of Díaz-Guardamino, the warrior stelae of Capote (J.54.1) and Cabeza del Buey IV (Majada Honda, J.110) are amongst the latest in her chronological scheme and both also have SW inscriptions. She proposes a date of the 8th or 7th century BC, with this date depending on the beginning of SW literacy (DíazGuardamino 2012, 408–9). Therefore, it seems to be only the presence of writing that suggests that warrior stelae could be as late as the 7th century. The writing on these two has an orientation that is upside-down from the perspective of the images. However, as RuizGálvez (2013) emphasizes, the position of writing respects that of the warrior emblems and works together with them to form an overall design. Therefore, the relationship of the writing and images on these stones do not imply a prolonged time lag or complete discontinuity between the traditions. In the case of the photogenic ‘guerreiro’ (J.12.1), features of the representation of the warrior are unlike those of the other warrior stelae.15 The writing and image are integral to a tightly organized overall design and so were almost surely executed at the same time.   The Late Bronze Age warrior stelae are distributed widely across the south and west of the Peninsula with a few outliers in the east and north-west. They show eclectic warrior’s gear of the Atlantic Bronze Age—swords, spears, and shields—as well as Mediterranean chariots, lyres, mirrors, combs, brooches; and central European helmets. According to Brandherm none of the artefact types depicted in the warrior stelae are certainly later than the 10th century BC (2013b and this volume). Therefore, that series seems to end—or at least its subject matter stops developing—near the Bronze–Iron Transition, near the margin of the Ría de Huelva and Baiões–Plaza de las Monjas phases, and somewhat before the foundation of the Phoenician colonies on Spain’s south coast, in the second half of the 9th century.   The Ría de Huelva deposition in the Odiel estuary shows that Atlantic forms still predominated dureing the 10th century BC, though co-existing with eastern Mediterranean materials and some central European types. Following the scheme of Burgess and O’Connor (2008), the great re-orientation occurred at the next stage: Bronce Final 3 synchronized with the British Ewart Park phase at 950–850/800 BC. It is at this period that the inventory of the western Peninsula shifts away from ALBA and towards the Mediterranean Iron Age.   In their geographical distribution and associated burial practices there are especially close links between the SW stelae with writing and the ‘Alentejo’ stelae of the Early to Middle Bronze Age (Brandherm 2013b and this volume). Both groups are concentrated in south Portugal (see Figure 15.8). The alentejanas display carvings of halberds, dirks, and other weapons datable to the span 1800–1300 BC (Díaz-Guardamino 2010). Both the EBA/MBA 15

In representing what seems to be body armour and the position of the feet and legs, the main figure of the well-known warrior stela of Ategua (Córdoba) shows similarities to ‘o guerreiro’ (J.12.1). Díaz-Guardamino regards Ategua as ‘posiblemente uno de los más reciente de la serie’ (2010, 432) dating from the late 9th century or during the 8th (cf. Díaz-Guardamino 2012, 404–5, 408).

Koch CW3 16 iii 16.indd 450

27/05/2016 19:31:10

{

Koch

[ 451 ]

î

SALACIA/ ?beuibon SALACIA % SALACIA

îî î î î

Cabeza del Buey 55 Siruela

î î îî

H

IULIPA

UAMA

Sal/Abul

î

KONIOI

î

H

4W

î î îî îîî î î î î î î îîî îî î î î î îîî î î î î

Lisbon % OLISIPO % Quinta do Alcácer do Almaraz

î î

î

57 Medellín

B

î

î

î

î

î

î î

î

î

î

B

MBA Alentejo stelae (c. 1800–1300 BC.)

the -briga line place-names with IP(P)O SW (Tartessian) inscriptions (c. 750–400 BC) LBA warrior stelae (c. 1250–750 BC) % pre-colonial Phoenician trading posts (c. 950–800 BC) % Phoenician colonies (c. 800–500 BC)

H

HIPSCA

îî î

î

î

î

LIPPOSH

AESURIS %

î î

H

IPTUCI H

IPONUBA ILIPA H HIPPO NOVA Niebla 53 Alcalá H MAGNA H IPAGRuM IPAGRUM ILIPULA del Río IPOCOBVLCOLA IPOCOBULCOLA H H ORIPPOH 51 Puente Genil % 52 VillaH VENTIPPO HHOSTIPPO BASILIPPO ILIPULA ONOBA manrique H Huelva MINOR DIPPOS H Mezquitilla SERIPPO H

%

î

Sacrum Promontorium

HIPORCA % Río Tinto

KUNETES % LACCOBRIGA Lagos

HEPORA?

î

î

IPSES

54 Capote

TUROBRIGA

B

B B B B î B

B B B B B B B B î B B B B B

Castelo de Moura

î

H KALLIPOUS CANTIPO

40N

LAEPIA

î î

H LACIPEA

KONISTURGIS

15.8 The south-western Iberian Peninsula in later prehistoric and protohistoric periods, showing the locations of carved stelae

H

56 Almoroquí Almoroqui

î

B

SCALLABIS Alcáçova de Santorém %

î

î î î

8W

î

H COLLIPO

î î î î

î î î

150 km

î

% KONIMBRIGA 40N

îîî î î î î

%

40N

î

î î î

LIPPOS H

GADIR Cádiz

%

Doña Blanca %

ACINIPPO H

H SAEPO

MAINAKE

MALAKA Malaga % Cerro del Villar %

C I L B ILACIPPO CENI H

% Alcorrín

Chorreras

%% % %

MAENUBA Toscanos

%

%

SEXI SEXI Almuñécar Almuñécar

HBAESIPPO FRETUM TARTESSIUM

and EIA stelae are connected with burials. In contrast, few of the intermediate warrior stelae group are clearly connected with burials. There was apparently a hiatus in ritual practice of a few centuries. As Brandherm explains (this volume), the MBA and EIA burials associated with stelae are enough alike to be hard to differentiate without scientific dating. Neither type has a mound, but instead a circular or sub-circular pavement on the surface with a cist situated beneath the middle. The pavements of burials at the same site can overlap.   My thinking has changed in recent years about the cultural message of the SW stelae with writing. When I began working on the corpus in 2007, I saw them as straightforward reflections of the Orientalization, that the makers of the inscriptions were embracing, unambiguously, cultural inflow from the east. Their values and attitude, even their identity, had been positively affected by the Phoenicians.   However, the more we have learned about the stelae and the Iberian Bronze–Iron Transition, the less this seems to be the correct interpretation. First of all, as more inscriptions have come to light, the distribution persistently shuns the areas with the heaviest eastern Mediterranean influence and new wealth. They do not occur inside the Phoenician colonies. No stelae with SW writing have yet been found in Huelva, which is often considered the core of the orientalized native polity called Ταρτησσος by the Greeks. As mentioned above, the identification with the Tarshish of the Old Testament is today viewed as likely. Only two stelae with writing were found along the Guadalquivir, another area of intense Orientalization. One of these is the complete, but now lost, inscription of Alcalá del Río, which is remarkable in several respects and is reconsidered below.   A picture of the Κονιοι or Κυνητες as a coherent group distinct from—though in contact with—Phoenicians and Tartessians (i.e. the peoples of the Orientalizing core zone) is now better defined. The Κυνητες and their inscriptions were peripheral to Ταρτησσος and the

Koch CW3 16 iii 16.indd 451

27/05/2016 19:31:12

[ 452 ]

xv. phoenicians in the west

]]

primary zone of eastern Mediterranean influence.   There are further key details in the writing system itself. The SW script derives mainly from a West Semitic alephat. In some cases, the forms of the SW signs look more like the earlier Canaanite alephat than its Phoenician descendant. For example, the recently discovered 10th-century inscription on a pithos from Jerusalem shows similar letter forms. These archaic antecedents of the signs have been an important strand in the arguments of Ruiz-Gálvez (2000; 2008; 2013) in favour of dating the SW stelae with writing as beginning before the end of the 9th century BC.   The SW script is the earliest of four closely related Palaeohispanic scripts. Though all four of these reflect their affinities with their Phoenician or Canaanite source, they also share one feature unlike the alephat. They are all hybrids—neither purely alphabets (like Phoenician, Greek, and Latin), in which one sign represents one phoneme, nor are they purely syllabaries (like Linear A, Linear B, Cypro-Minoan, and the Cypriot syllabary), in which signs for consonants also represent the following vowel. The Palaeohispanic scripts are semisyllabaries: some signs represent just a vowel or just a consonant, others a consonant plus a vowel. The consonants for which syllabic signs were used are those that phoneticians call the stops: [b, k g, t d]. The earliest Palaeohispanic writing systems did not distinguish between the voiceless stops [k] and [t] and their voiced counterparts [g] and [d]. I omit [p] here, because pre-Roman languages of the Iberian Peninsula that used versions of Palaeohispanic script (Iberian and Celtic) apparently did not possess the phoneme /p/ as a segment, at least as a common one, in their phonological systems.16 The SW script, in common with the other Palaeohispanic scripts, represents five vowels and does not distinguish length: thus, a e i o u, which could represent [a(ː) e(ː) i(ː) o(ː) u(ː)]. Thus, as a basic structural feature underlying all the Palaeohispanic writing systems, there are 15 syllabograms used to represent the stop consonants and following vowels, transliterated as ba be bi bo bu, ka ke ki ko ku, ta te ti to tu. In the case of the SW inscriptions, we should refer to the script as a ‘pseudo-semisyllabary’ with 15 ‘pseudo-syllabograms’, because the usual practice was what is termed ‘redundancy’, that is, to write the consonant sign that implied the following vowel and to write the vowel sign afterwards also, thus in transliteration: baa bee bii boo buu, kaa kee kii koo kuu, taa tee tii too tuu.

  After puzzling over the origin of this hybrid system for years, I now believe that the creator or creators of the prototype Palaeohispanic script were not merely trying to improve the alephat—as the Greeks did so well in devising the alphabet—but they were also trying to make it different, to create a national script that would be obviously distinct when displayed publicly in memorial inscriptions. Similar motives have been seen behind the preservation of indigenous syllabic writing in Cyprus alongside the Phoenician alephat down to Hellenistic times. The Phoenicians came to the Iberian Peninsula mostly or entirely by way of Cyprus (cf. Figure 15.9). At the time there were at least three languages and two 16 The situation is somewhat complicated in Celtiberian, but there are numerous examples of loss of PIE *p, which was the regular development in most phonetic environments. It is probable also that some Celtic (as opposed to Lusitanian) dialects in the Peninsula show /p/ from earlier *kw, as in Brythonic, Gaulish, and Lepontic. This change may also occur in Lusitanian, PVPPID in inscription L.1.1 being a likely example, but that language also shows P retained from PIE *p as in PORCOM ‘pig’.

Koch CW3 16 iii 16.indd 452

27/05/2016 19:31:12

[ 453 ]



Koch £ £

£

{

35 N 35 E

KITION

60 km

AMATHUS



15.9 Ancient Cyprus and the north-east Mediterranean, places mentioned in the text and sea route to the west

PALAEPAPHOS

£

SALAMIS

£

SIDON

TYRE

£



writing systems in use on the island: the alephat used for the language of the Phoenicians and the Cypriot syllabary used to write Greek and the undeciphered Eteocypriot language. Unlike the rest of the Greeks (who lost the Linear B syllabary and literacy altogether when Mycenae collapsed in the 12th century BC), the Greeks of Cyprus continued writing with a syllabic script, related to Linear B, Linear A, and, most closely, to the Cypro-Minoan syllabic script of Bronze Age Cyprus, from which the Iron Age Cypriot syllabary clearly derives. It is now generally recognized that a form of Cypriot syllabic script was used to write Greek on the bronze obelos (skewer) of c. 1050–950 BC from a tomb from near Palaepaphos inscribed with the name o‑pe‑le‑ta‑u = Οφέλταυ (Olivier 2013). As a system for writing the Greek language, compared to the Greek alphabet (which is an adaptation of the alephat), the Cypriot syllabary is inefficient. It has 55 or so hard to differentiate signs and represents the Greek sound system poorly. But they kept it until after the time of Alexander. Why? I am not alone in concluding that nativist cultural pride was a factor (Sherratt 2003). It was clearly different from the alephat, more different than the Greek alphabet. It was derived from scripts that had been used in Cyprus and the Aegean since the Bronze Age. It was thus connected with the age of the Homeric heroes whom some Cypriot kingdoms claimed as founders.   The Cypriot syllabary has a striking structural similarity with the Palaeohispanic scripts. Like them—and at odds with the Greek sound system—it writes 15 signs combining stop consonants and vowels: pa pe pi po pu, ka ke ki ko ku, ta te ti to tu. The corresponding series have 25 signs in Linear B. But the Palaeohispanic scripts represent precisely the same 15 combinations of stop+vowel as the Cypriot syllabary.   As mentioned above, a Cypriot link with Atlantic Iberia became prominent when contact with eastern Mediterranean resumed during the Ría de Huelva phase of 1050–950/900 BC (Mederos 1996; Almagro-Gorbea 2001; Blázquez 2011). A Cypriot origin dating to this period has been proposed for the Berzocana bronze bowl (Mederos 1996, 104–7; Torres 2008, 138). The exchange was then two-way, as reflected in a fibula of Ría de Huelva type and an Atlantic rotary spit placed in Tomb 523 in Amathus, south Cyprus, c.  1000 BC (Karageorghis & Lo Schiavo 1989), one of two spits of this type found in Cyprus (Burgess & O’Connor 2004; Armada et al. 2008, 483–4). Even as direct influence from Tyre became

Koch CW3 16 iii 16.indd 453

10/06/2016 18:49:45

15.10–12 Two views of one of the cast-bronze lions’ heads adorning the funerary vehicle from ‘royal’ Tomb 79 of Salamis, Cyprus (by permission of the Department of Antiquities, Cyprus) and reconstructed funerary chariot, showing cast-bronze lion’shead hubcap from Tomb no. 17 of the La Joya necropolis in Huelva

more prominent in the following Baiões–Plaza de las Monjas phase (950/900–825 BC), the influence from Cyprus continued. Cypriot ceramics occurred together with Tyrian wares in the eponymous find in Huelva town (Gonzáles de Canales et al. 2004; 2006), and the elaborate bronze ritual wheeled stands from Nossa Senhora da Guia (Baiões) have close parallels from Cyprus at a similar absolute date, c. 900 BC (Torres 2008, 138). The Cypriot connection remains evident into the Orientalizing period (750–550 BC). For example, there is a striking similarity between the five cast-bronze lions’ heads adorning the funerary vehicle found in the ‘royal’ Tomb 79 of Salamis, Cyprus, dating to c. 700 BC (Karageorghis 1969, 76–98; Figures 15.10–11) and the cast-bronze lions’ heads used as hub caps (Figure 15.12) on the funerary chariot of the lavish Tomb no. 17 of the La Joya necropolis in Huelva, probably dating to the 7th century BC (Garrido 1983). The closely similar lion’s-head protome found in Huelva town had probably originally been deposited in Sector A of the La Joya necropolis, of the late 8th to late 7th century BC (González de Canales et al. 2012).   In the light of the foregoing, I suggest that when literate peoples came into contact with the natives of the Iberian Peninsula, Cypriots had a leading role in this enterprise and contributed some ideas about writing as well as writing itself: • First: the indigenous languages of a country should be written with a script distinct from the alephat of the Phoenicians. • Second: the script for the native language should be structurally distinctive, by usefully including vowels and not so usefully 15 syllabic signs for stop+vowel. • Third: the inability to distinguish voiced and voiceless stop consonants was the price to be paid for a script that could (unlike the alephat) represent vowels. • Fourth: the native script was to be used to make indigenous languages visible where they could be associated with claims of legitimacy and descent from the native heroes and founders of the Bronze Age.   In the case of Cyprus the fourth idea is true: the syllabary is a relic of the Aegean/ Cypriot Bronze Age. For the Palaeohispanic script, this is a nativist myth or a cognitive

Koch CW3 16 iii 16.indd 454

27/05/2016 19:31:25

[ 455 ]

Koch

over-simplification concerning a newer writing system devised to make visible the earlier language. The peculiar structure of the Palaeohispanic semisyllabaries (including the SW pseudo-semisyllabary) makes more sense—and becomes a more impressive intellectual achievement—if we realize that it was probably inspired by travellers who did not know the Greek alphabet and possibly before the Greek alphabet itself existed. A literate person or group who knew only two writing systems, the alephat and Cypriot syllabic writing, faced with the challenge of writing a new language, could have taken this opportunity to combine the better features, and remedy the defects, of both systems. In Cyprus itself, the two scripts were traditional and inalterable within established sociocultural domains.   The alephat had the practical advantage of a relatively small number of signs that were easily learned and easily written and distinguished on a wide variety of media. It is for this reason, I think, that most of the Palaeohispanic signs are derived from West Semitic signs, some showing specifically archaic forms in use before c. 800 BC. To these, a few new simple geometric signs were added. SW sign & transcription aa

graphic model for sign

model for phonetic value

ii

Phoenician ayin

Phoenician yodh

Cypriot 𐠁𐠁 ε η

ee oo uu b ba

Table 15.1 Proposed antecedents for the SW pseudo-semisyllabary

B be U bi p bo P bu k ka K ke q ki g ko Q ku t ta h te 0 ti D to d tu ll m…m nn rr Rŕ ss Sś A h (?)

Koch CW3 16 iii 16.indd 455

Phoenician aleph

(Cypriot 𐠞𐠞 πα βα φα) ?

Phoenician waw

Phoenician mem Phoenician beth

(Cypriot 𐠯𐠯 τι δι θι)?

Phoenician gimel Phoenician kaph

Phoenician qoph Phoenician heth Phoenician taw

Phoenician heth Phoenician tet

Phoenician daleth

Phoenician daleth Phoenician lamedh Phoenician mem Phoenician nun Phoenician resh

Phoenician samekh Phoenician sin Phoenician he

Cypriot 𐠀𐠀 α Cypriot 𐠂𐠂 ι

Cypriot 𐠃𐠃 ο ω Cypriot 𐠄𐠄 υ

Cypriot 𐠞𐠞 πα βα φα Cypriot 𐠟𐠟 πε βε φε

πη βη φη

Cypriot 𐠠𐠠 πι βι φι

Cypriot 𐠡𐠡 πο βο φο πω βω φω Cypriot 𐠢𐠢 πυ βυ φυ Cypriot 𐠊𐠊 κα γα χα

Cypriot 𐠋𐠋 κε γε χε κη γη χη Cypriot 𐠌𐠌 κι γι χι

Cypriot 𐠍𐠍 κο γο χο κω γω χω Cypriot 𐠎𐠎 κυ γυ χυ Cypriot 𐠭𐠭 τα δα θα

Cypriot 𐠮𐠮 τε γε θε τη δη θη Cypriot 𐠯𐠯 τι δι θι

Cypriot 𐠰𐠰 το δο θο Cypriot 𐠱𐠱 τυ δυ θυ Phoenician lamedh Phoenician mem Phoenician nun Phoenician resh

τω δω θω

Phoenician samekh Phoenician sin Phoenician he (?)

27/05/2016 19:31:25

[ 456 ]

xv. phoenicians in the west

]]

  On the other hand, the syllabaries could represent vowels, which was a great advantage over a writing system lacking them. An inventor ignorant of the Greek alphabet, and any later alphabet with vowels derived from the Greek alphabet, might easily have been conditioned to think that a script representing vowels was incompatible with the alphabetic principle (one sign = one phoneme). This would explain why structural features of syllabic writing, as well as vowel signs, have been incorporated into the Palaeohispanic scripts.   What is proposed here for the genesis of Palaeohispanic writing is analogous to the creation, more than millennium later, of the Irish Ogam script. In reality Ogam most probably derives from Romano-British literacy and influence (McManus 1991, 19–41). As with Linear B, the Cypriot syllabary as used for Greek, and the northern Palaeohispanic script as applied to Celtiberian, Ogam’s system of five vowels not distinguishing vowel length replicates its source and misrepresents the language for which it was used, in this analogous case, Primitive Irish. Despite most probably being derived from the Latin alphabet, the Ogam script was intentionally made to look different from Roman letters. It occurs most densely farthest from Roman Britain (McManus 1991, 44–9; Koch et al. 2007, §20a, §§20.2, pp. 170–82) and was claimed to be as old as the Tower of Babel (Ahlqvist 1982). This, then, is my working hypothesis about the cultural message of the SW inscriptions: they do not embrace the Orientalization, but represent a secondary reaction against it on its periphery. In their archaeological context, the inscriptions—their script and language—can be seen as elements within a multi-faceted nativist package. The Κυνητες were making a statement that they were the people who had lived in this country for a thousand years. Of course, even clear statements can be lies or honest mistakes, and one can write, ‘We were here before the Saxons’ in English. Nevertheless, the adoption of a nativist cultural package in contrast to the Orientalization of their neighbours would make sense as the response of a group on the margins of eastern Mediterranean influence who spoke as their first language what had been the Atlantic lingua franca before the arrival of the Phoenicians. The Iron Age cist burials of the south-west are not a new style of grave inserted into prehistoric monuments that were conspicuous in the landscape. Rather they replicated the architecture of Middle Bronze Age cist burials, including features inconspicuously below the ground (Brandherm, this volume). That looks more like genuine continuity. Later on, the Greek and Roman writers record a few traditions from the former lands of the Tartessians and Κυνητες. Some of these have a nativist spirit reminiscent of the legendary history of Early Christian Ireland. Justin’s Epitome of Trogus Pompeius tells us that in the days before farming and towns a king named Gargoris ruled the Cun{tes, dwelling in ‘the forests of the Tartessians.’17 Gargoris repeatedly sought to kill the grandson Habis shamefully begotten on his own daughter. The boy eventually overthrew and succeeded the old king and then, rejecting the wild life and diet he had known, introduced cultivated food and urban life to his people, as the founder of a long-enduring Tartessian dynasty.18 The Greek geographer Strabo (III, 1, 6) wrote of the Turdetani, a people living in his time (c. 64 BC–post AD 24) in what had been part of the Tartessian kingdom, and whose 17 Cun{tes is the most likely reading for Curetes in the extant manuscripts. 18 See the translation and edited Latin text of Freeman (2010, 316–18).

Koch CW3 16 iii 16.indd 456

27/05/2016 19:31:26

Koch

[ 457 ]

name possibly derived from the older place-name: ‘They are the most civilized of the Iberian peoples, having books and poetry and laws in verse going back 6,000 years’19 (cf. AlmagroGorbea 2005). Between Justin and Strabo, we have both relative and absolute chronology for the beginning of Tartessian civilization—the First Neolithic and c. 6000 BC. The chronological correspondences are probably lucky and nothing to get too excited about. Yet it is a strongly nativist claim, intended to exceed the priority of every other civilization of the ancient world.   To sumarize, the SW inscriptions show literacy in far western Europe in the Early Iron Age. That the language of this corpus is partly or wholly Celtic is discussed below. There is no doubt that Palaeohispanic literacy began as a result of West Semitic stimulus. But its earliest manifestation in the lands of the Κυνητες is peripheral to the Orientalization and looks back to the regional Bronze Age as part of a nativist package. Whether or not it is unique to Europe’s Atlantic façade, we find strongly asserted beliefs there in extreme indigenous antiquity. We can be certain that the written languages—such as Ogam and Palaeohispanic—do not go back several thousand years. But the claims for spoken languages, cultural identities, and genetic ancestries with ancient roots in the west are not so easily falsified. We should continue to investigate these claims. On the Celticity of the SW inscriptions The language of the SW inscriptions, also called ‘Tartessian’, is found classified as Celtic sometimes. As well as my own recent work (e.g. Koch 2013a), Tartessian is listed amongst the Celtic languages in the updated Indo-European family tree of Hamp 2013. Jordán Cólera (2005, 8; 2007, 751) assigns the language to the Indo-European macro-family and Celtic family with queries. Lorrio and Ruiz Zapatero (2005, 18) also note this classification as possible but uncertain. Broderick (2010, 304–6) supports the possibility that Tartessian is a Celtic language with a detailed collection of evidence (mostly based on, and attributed to, Koch 2009). In a substantial collaborative publication, Villar, Prósper, Jordán, and Fernández Álvarez conclude: ‘Más reciemente J. T. Koch ha proporcionado argumentos lingüísticos de mayor enjundia en favor de la tesis de la filiación celta, de manera que en la actualidad conviene retirar, al menos provisionalmente la lengua de las inscripciones del suroeste como miembro del listado no indoeuropeo’ (2011, 100). Maier (2012, 211) proposes that Tartessian is either a Celtic language or that the corpus contains a borrowed Celtic element. Tartessian is classified unambiguously as Celtic by Kaufman in a study of the full corpus (2015) where he also confirms that this is the view of Hamp, including specific points of Hamp’s analysis. Werner Nahm also argued that the SW language was Celtic in a seminar given at the Dublin Institute for Advanced Studies in October 2015.   The position of Correa in his pioneering publications of the 1980s and early 1990s was that Tartessian was a Celtic language (e.g. Correa 1989; 1992). Untermann (1995 & 19 νυνὶ δ᾽ ἐν αὐτοῖς οὐδεὶς φαίνεται διορισμός. σοφώτατοι δ᾽ ἐξετάζονται τῶν Ἰβήρων οὗτοι καὶ γραμματικῇ χρῶνται καὶ τῆς παλαιᾶς μνήμης ἔχουσι συγγράμματα καὶ ποιήματα καὶ νόμους ἐμμέτρους ἑξακισχιλίων ἐπῶν, ὥς φασι: καὶ οἱ ἄλλοι δ᾽ Ἴβηρες χρῶνται γραμματικῇ, οὐ μιᾷ δ᾽ ἰδέᾳ

Koch CW3 16 iii 16.indd 457

27/05/2016 19:31:26

[ 458 ]

xv. phoenicians in the west

]]

MLH IV) augmented Correa’s case with further examples, but concluded that it was not yet possible to prove conclusively.20   One also finds the somewhat different idea that the SW corpus contains a substantial component of Indo-European/Celtic21 names, but the main language of the inscriptions, i.e. ‘the matrix language,’ is something else, possibly or probably non-Indo-European. This is Correa’s second position (e.g. Correa 1996) and also that of Villar (2004). Fortson (2009, 313) sounds more like this group, but he does not explicitly exclude a straightforward Celtic classification for the language: ‘On the southernmost tip of the Iberian Peninsula was spoken another poorly known language called Tartessian, which has some Celtic linguistic material in its personal names.’   Whether lined up with the first theory (the SW language is Indo-European/Celtic) or the second (the SW corpus is comprised of Indo-European/Celtic names in a different matrix language), or not comitted in favour of either one of these alternatives explicitly, several researchers have published lists of Celtic-looking names, sometimes together with other Celtic-looking elements, from the corpus. The list of Correa (1992, 98–101) includes akoosioś, ]tuurkaaio[, bootiieana, taalainon, lokooboo, niiraboo, aibuuris, tiirtoos.22 Untermann (MLH IV, 166–8) listed liirnestaakuun, ]taarnekuun, lokooboo, niiraboo, koobeeliboo, rinoeboo, lokoon, akoosioś, aalaein, aibuuris, albooroi; ] anbaatiia, ‑uarbuui, aarkuui, bootiieana, ooŕoir, sarune(ea), taalainon, tiirtoos, ] tuurkaaio[, uursaar; also seeing as Indo-European and not regarded as other than Celtic, naŕkeentii and baarentii (probably being verbs) and sabooi (possibly a common noun, if not a name).23 The list of Villar (2004, 262–6) is lokooboo, niiraboo, rinoeboo, koobeeliboo, aibuuris, uarbaanubuu, lokoon. Villar also mentions in the same section 20 As well as additions to Correa’s list of Celtic onomastics, Untermann pointed to forms, such as teee-baarentii (J.23.1) that appear to be inflected as verbs with Indo-European primary endings and forms with these same stems with ro prefixed to them, e.g. tee-ro-baare (J.1.1, J.12.1). The preverb ro < PIE *pro is one of the most common and characteristic features of the verbal systems of the early Celtic languages, especially, but not only, Old Irish. Therefore, in the light of his own detailed appraisal of the corpus, Untermann’s conclusion (‘the formulaic segments do not completely exclude the possibility that the language of the corpus is Celtic’) is hyper-cautious. He provided no reasonable alternative as to how a language with such earmarks could be something other than Celtic. That Untermann had once intended to take a stronger position is apparent: he wrote that he had been tempted to call MLH IV Die Inschriften in altkeltischen Sprachen (MLH IV, 89), instead of the title used, Die tartessischen, keltiberischen und lusitanischen Inschriften. 21 The term Indo-European/Celtic is used because, once we consider the possibility the SW corpus contains more than one language and therefore break the texts down into names, words, and other kinds of elements, we will find some segments that look Indo-European (and are not inconsistent with a classification as Celtic), but lack distinctively Celtic diagnostic criteria. Although it is possible that the SW corpus would contain a mixture of linguistic material that was Celtic together with material that was Indo-European, but not Celtic, as is generally thought to be the case with the five longer Lusitanian inscriptions, the term Indo-European/ Celtic as used here is not meant to imply this. 22 uarbaan and uarbaanubuu are also interpreted as Celtic in this paper of Correa’s, but not seen as names. For the sake of ease of comparison, the lists of forms are all given using the conventions of Romanized transcription used in Untermann’s MLH IV. Where the proposed word divisions vary between the writers, I have used my own (Koch 2013a). The segmentation of the forms is not the point at issue presently. 23 Several of the forms identified by Untermann as having Celtic and/or Palaeohispanic Indo-European comparanda are attributed by him to publications of Correa of the 1980s.

Koch CW3 16 iii 16.indd 458

27/05/2016 19:31:26

Koch

[ 459 ]

of his 2004 article the names of kings of Tartessos, known from Greek and Roman sources, ’Αργανθωνιος and Gargoris. The latter is recognized as probably a compound name with Celtic *‑r¼χs ‘king, &c.’24 The list of Ballester (2004, 119–20; 2012, 10–11) is ’Αργανθωνιος, anbaatiia, akoosioś, aibuuris, aarkuui, bootiieana, tiirtoos, eertaaune (which is probably not a name), lokooboo, mutuuirea, sarune(ea), liirnestaakuun. AlmagroGorbea et al. (2008, 1050) list the following forms as evidence for the Celtic character of the Κυνητες, also known as Κονιοι, of south Portugal and the Guadiana basin: akoosioś, aalaein, aarkuui, aibuuris, albooroi; anbaatiia, bootiieana, sarune(ea), taalainon, tiirtoos, ]tuurkaaio[. They also mention ’Αργανθωνιος as Celtic (ibid., 1051).   In the catalogue in Koch (2013a), there are 88 SW inscriptions. Of these, 16 are short, badly damaged, or for various reasons offer little basis for segmenting the text into words and names that could be compared with the rest of the corpus or with other languages.25 The 72 remaining inscriptions contain 1752 graphemic signs (see below). The sequences of signs that I have provisionally identified as names all have Indo-European or Palaeohispanic parallels, usually both. Most often these forms have specifically Celtic affinities, including case endings that are consistent with a classification as Celtic. This onomastic subset comprises 590 signs or 33.7% of the corpus.26 Earlier writers have generally not attempted exhaustive lists. Therefore, my list is longer. In any event, the figures offered here will serve to indicate a general order of magnitude.   For the question of the time and place of the emergence of Celtic from Indo-European and that of the whereabouts of Celtic speakers in the Early Iron Age, the presence of Celtic elements in the corpus is the key point. The classification of the matrix language is of secondary importance for non-linguists. The evidence of the corpus can stand as direct testimony for Celtic in the far south-west of Europe by the mid 7th century BC, probably by the 8th, and possibly before that.   Of the personal names from the SW corpus cited above and the corpus list below, analogues can be found for several of them in Celtiberian and other Ancient Celtic languages. But the corresponding forms are to be found most frequently amongst the pre-Roman names of the western Peninsula. This generalization includes aarkuui, aibuuris, albooroi; anbaatiia, bootiieana, ebuuŕoi, taalainon, tiirtoos, soloir, lokooboo, niiraboo, tiilekuurkuu, and tuurea. Particularly rich sources for onomastic comparanda are the collections of Albertos (1985) and Vallejo (2005). Thus, what has appeared by applying the Correa/Untermann phonetic key to the SW corpus is not just a particular language, but— time and again—the pre-Roman Indo-European of the western Iberian Peninsula. 24 In the light of Old Irish garg ‘savage’, Gargoris can be etymologized as ‘savage king’, which perfectly suits the character in the Tartessian foundation myth above. 25 The inscriptions excluded from the statistics are: J.1.6, J.4.2, J.6.2, J.6.3, J.7.3, J.7.4, J.7.7, J.8.1, J.12.2, J.12.5, J.16.4, J.18.3, J.28.1, Medellín T3 86H/1TP–1, Folha do Ranjão, Salacia. 26 In compiling these statistics, signs that could be read with relative certainty were all counted. Signs that could be read, but were damaged or the reading was for some other reason doubtful, were also counted. Signs where the reading was totally uncertain were represented in the transcriptions with * not counted. Signs that were missing due to a break in the stone or were simply not carved, but their value can be inferred because of the usual orthography of the recurrent formula, are supplied within square brackets [ ] in the catalogue, but they were not counted in compiling these totals.

Koch CW3 16 iii 16.indd 459

27/05/2016 19:31:26

15.13 The SW inscribed stone from Madroñero, Cáceres, Spain (J.56.1)

  Although the SW inscriptions are written in scriptio continua, the uncertainty of the word divisions can be over-estimated. As argued previously (Koch 2011, 37–42), there is no doubt about the segmentation of most forms in the longer and better-preserved texts in the SW corpus, due to the following factors: 1. the beginning and/or ending of the text preserved intact; 2. the agreement of ‘redundancy’ between the signs for the stop consonants and following vowels; 3. the recurrent formula uarbaan tee-ro-baare baa-naŕkeentii with elements of it sometimes rearranged; 4. other recurrent stems, word endings (some of these showing case agreement, such as lokooboo niiraboo), and prefixes, which retain their recognizable form when rearranged in the syntax.   Although a conclusion similar to Correa’s second (two-language) theory has been presented by Villar, in the same important study he argued that the Indo-European/Celticlooking forms were not limited to the onomastics. A syntagm like akoosioś naŕkeetii (Untermann 1997, J.56.1 [Figure 15.13]) seems undoubtedly to be a funerary formula from an Indo-European language with a thematic nominative singular anthroponym followed by a third person singular verb, also with thematic inflexion. (Villar 2004, 264)27

As Villar saw, the reading akoosioś naŕkeetii28 (clearly carved and leaving plenty of space on all sides had more signs been wanted) implies that the decipherment of the onomastics as Indo-European is correct, in which case the decipherment of the formulaic naŕkeetii should also be correct as appearing to be in the same language using the same transcription system. This observation implicates much of the matrix language, because what Villar identifies as an Indo-European verb is a variant of the most common word in the corpus and the most essential element of the Tartessian epigraphic formula (Villar 2004, 264): naŕkeenii (J.2.1, J.21.1), n[aŕke]enii (J.6.1), naŕkeentii (J.12.1, J.16.1, J.17.2, J.18.1), [n]aŕkeentii (J.1.5), na]ŕkeentii[ (J.4.3), n(a)ŕkeenii (J.11.1), n]aŕkeenii (J.11.3), na]ŕkeeni (Corte Pinheiro) naŕ]keenii (J.19.1), naŕrkee:n: (J.23.1), n[a]ŕkeen (Cabeza del Buey IV), na[ŕ]keen (Monte Gordo), naŕkeenai (J.7.1, J.55.1), ]naŕkeeuu[ (Corte do Freixo 2). 27 I understand from personal communication that Villar has subsequently come to the view that the language of the SW inscriptions as a whole is Celtic (30 May 2009). 28 ?Cf. κεῖται Πάτροκλος ‘[here] lies Patroklos’ (Iliad 23.210; cf. Wikander 1966; Mallory & Adams 2006, 296).

Koch CW3 16 iii 16.indd 460

27/05/2016 19:31:26

Koch

[ 461 ]

naŕkeentii and these variants make up 279 of the 1752 signs of the 72 most readable SW inscriptions mentioned above (see Table 15.2), thus 15.9% of the corpus. Moreover, beyond that 15.9%, the SW corpus contains several other forms that appear to be inflected the same way: tee-baantii (Mesas do Castelinho), teee-baarentii (J.23.1), lakeentii (J.53.1), &c.   Also relevant in this connection is Lusitanian DOENTI , on the rock-cut inscription L.2.1 of Lamas de Moledo, Viseu, Portugal. DOENTI is usually interpreted as a third person plural present tense verb and translated ‘they give.’ If the SW matrix language is non-Indo-European, we then must believe that by coincidence this pre-Roman language of Portugal makes use of a two-syllable, four-phoneme ending phonetically identical with one easily understood as a verbal ending in a second, completely unrelated pre-Roman Indo-European language of Portugal.   My current working hypothesis is that the formula is written in a Celtic language. A more accurate phonetic transliteration of it is |u̯ara man de-ro-b\re ma-narkenti|. I propose that it means approximately ‘. . . has carried away [the named deceased] to the highest one (< *u(p)era m\m); so they now lie below [this stone]’ (Koch 2013b). This formula counts as part of the matrix language, rather than the onomastics. Variations of the epigraphic formula comprise 33.2% of the 72 inscriptions examined statistically here: 581 signs of the total of 1752 or 50.3% of the matrix language, i.e. excluding the names (counted as 590 signs or 33.7% of the corpus; see Table 15.2 below). Therefore, determining what language the formula is in is tantamount to determining the classification of the SW matrix language (Koch 2013b). That means that in order to maintain the non-IE matrix language theory one must defend the proposition that naŕkeentii, &c., resembles Indo-European verbs by coincidence and that tee and ro correspond in their form and syntax to two of the most common Old Irish preverbs, de and ro (McCone 2006, 177–224) also by coincidence.29   One reason to associate the Ταρτησσος of history with Celtic, at least with the use of Celtic names, is ’Αργανθωνιος, the name of the king said by Herodotus (1, 163–5) to have held power in Ταρτησσος for an improbably long reign c. 625–c. 545 BC. Several researchers have recently emphasized the unambiguous Celticity of ’Αργανθωνιος so as to clearly reject alternative interprtetations of the name (e.g. Villar 2004, 264; Ballester 2004, 119; 2012, 14; De BernardoStempel 2006, 47; Almagro-Gorbea et al. 2008, 1051). The name is based on a well-attested Celtic word of Indo-European origin, meaning ‘silver’: Celtiberian arkanta, Arganta, Old Irish arggat, airget ‘silver’, Middle Welsh aryant ‘silver’, Breton arc’hant, argant ‘silver’, Latin argentum < PIE *H2erĝṇtom ‘silver’ < *H2erĝ- ‘white, bright’ (cf. Falileyev 2010, 55–6), Gaulish ARGANTODANNOS/ARCANTO­DANNOS ‘moneyer, magistrate of silver and/or coinage’ on the coins of the Lexovii and Meldi (Hoz 2007, 192–3). The funerary inscription from Alconétar, Cáceres, which impresses Ballester as a recurrence of ’Αργανθωνιος, reads FLACCVS ARGANTONI [FILIVS] MAGILANICVM MIROBRIGENSIS (Sánchez Moreno 1996, 127; Vallejo 2005, 186–7; cf. Luján 2007, 253). This site is on the periphery of the historical Ταρτησσος and the geographical distribution of the SW inscriptions. Argantonius at Alconétar is wholly consistent with Celtic naming patterns found in the Peninsula and beyond: for example, ARGANTO MEDVTICA MELMANIQ[VM] (Riba de Saelices [Vallejo 2005, 29 For an extended debate on the classification of the language of the SW inscriptions, see Journal of IndoEuropean Studies 42 (Eska 2014; Koch 2014b; 2014c; Prósper 2014; Valerio 2014).

Koch CW3 16 iii 16.indd 461

27/05/2016 19:31:26

[ 462 ]

xv. phoenicians in the west

]]

186–7]), the group name LVGGONI ARGANTICAENI (Villaviciosa, Oviedo [Búa 2000, 274]; the name reappears as that of the Breton noblewoman Argantken in the 12th century AD [Evans 1988, 549]), the family name of [T]OVTONI ARGANTIOQ[VM] AMBATI F[ILIVS] (Palencia [González Rodríguez 1986, 123; Vallejo 2005, 186–7]); cf. Cisalpine Gaulish (Vercelli) a r kato ko〈k〉mat e r e ko s/ARGANTO­COMATERECVS . Silver was the principal high-value export of Ταρτησσος to the eastern Mediterranean. Table 15.2 SW texts used for the statistics bold italic type

= forms counted as onomastics (elements of naming phrases)

= formula words (including some examples inflected like absent formula words taking their place in the syntax)

bold Roman type

bold type anditalic arguably

bold Roman type

normal weight italic type = other forms that are relatively strongly suspected of being IndoEuropean, but not counted in either of the preceding categories and thus accounted within the normal weight italic unassigned 33% type of the 1752 signs. %

The order of figures in square brackets after each text is as follows: total number of readable signs followed by readable signs counted as onomastics followed by number of signs in formula words (or equivalent).

J.1.1 J.1.2 J.1.3 J.1.4 J.1.5 J.2.1 J.3.1 J.4.1

loko oboo≡niirabo o to o aŕaiai kaaltee loko on ane naŕke e ka aki iśiinko olobo o ii te e-ro-ba are (be )e te asiioonii [74, 27, 13] k o o-b e elib o o na-k i i-b u u oira uarb a an t i irt oos ne-b a a-naŕk e eni [40, 14, 15] [27, 0, 13] ]ŕak u urś t e e-ba are naŕk e eni ak a a**ir-ion a[ ]seku ui uurke e ote erka a ŕ*[ ]aehaeoleaala[ [27, 9, 0] mut u uirea b a ar[e n]aŕk e ent ii a(a/m)musok e eonii [27, 8, 10] ]b o oara naŕk e enii [13, 0, 13] aib u uris[ ]a k i inb a aib i i ro-la?a uarb a an ub u[u]i [31, 15, 8] ?ib ooi ion asune≡ uarb aan ≡ek uuŕine ob aar b aara*******t aa oret oo [39, 12, 10]

J.4.3 ]*r t e eaion(k a )a[ . . . na]ŕk e ent i i[ [14, 7, 6] J.4.4 ]*reonuu[ ]u[a]rb a a[an . . . naŕk e ]enii[ [14, 0, 8] J.5.1 sabooi : istaaib oo rinoeb oo anak eenake:eib oo iib aan b aareii [43, 31, 8] J.6.1 te]ea-ba are n[aŕke]enii [11, 0, 10] J.7.1 aśtaa boo(ti)ir naŕkeenai aśtaa na-boolon [27, 0, 8] J.7.2 bootoo?ar[ ]*aa kaaŕner-ion ire [21, 4, 0] J.7.5 uarbo on i[ | naŕke en [13, 0, 12] J.7.6 aarkuui oriou〈ti〉bea:i :elurear[ ]nii [27, 6, 3] J.7.8 * *k e e≡uuak e e*[ ]eb o o t ee-ba ere naŕke en emun t u urea≡iub a a [35, 17, 12,] J.7.9 iru b a arua-ion b a a[ [13, 3, 7] J.7.10 ]****naŕkeenii raśen ba are [17, 12] J.9.1 ] aanan uarb a an eb e e naŕ[k e e . . . [17, 5, 9] J.10.1 ariariśe : oni?ak a atii-śe : o?er-beeŕi : leoine ar-ba arie(?n)i?ensere / (?t a )au [55, 14, 15] J.11.1 k i ielaoe:≡ oiśaua ≡b a ane≡ rob a ae n(a)ŕk e enii [29, 22, 7] J.11.2 ]on linb o oire anb a a[ [14, 12, 0] J.11.3 soloir uarba an[ ]ina o*[ | n]aŕkeenii [23, 5, 13] J.11.4 aiooŕorainn b a aanon**[ | ea ro-n-ba aren naŕke enii aliśne śti aś*ta a*ta a [49, 12, 16] J.11.5 a]nbeiki[i ]arsk eeirn*[ naŕkee]ntii [15, 0, 3] J.12.1 iru≡alku u sie: naŕke enti i muba a tee-ro-ba are hata aneate e [40, 20, 16] J.12.3 (])uult iina ar-beieŕit uu la[. . . [18, 7, 9] J.12.4 salsaloi ?i [ | ]b a e b a a-lak i in ?i i [20, 8, 9]

J.14.1 J.15.1 J.15.2 J.15.3 J.16.1 J.16.2 J.16.3 J.16.5 J.16.6 J.17.1 J.17.2 J.17.3 J.17.4 Koch CW3 16 iii 16.indd 462 J.18.1

t a alainon t u uŕek u ui or[ | [ ] i [ ] | ]noś t a ae-b a are naŕk e en [34, 15, 13] hait u ura meleśae≡: :≡b a aenae *(*)n [23, 20, 0] ]n / aŕ[k e e(n) . . . [3, 0, 3] aalaein ŕe[ nar´]k eeni [13, 7, 4] uursaau *arb a an t e e-bar[e] b a a-naŕk eent ii [27, 6, 21] ( )omuŕika a[ ]anba ati ia≡ioba a[ ]*e ba a-[na]ŕke e keeo-ion[ [30, 18, 5] itiiabe ŕebe anaka a | ro-ba are ba a-naŕke [e]nti i [28, 5, 15] uab a an : ne[ [8, 0, 5] ba]are[ [3, 0, 3] kuui k aaosa naŕ[k e e]n(t i i) [12, 0, 4] k u ui arairb u b u [u | ]b a are naŕke ent i i [22, 0, 12] baar]e naŕk ee[ [6, 0, 6] : k e enila (*) rin≡ | ≡b e e:lin enb e ·k a arne : [24, 15, 0] bo oti ieana≡ ke erto o ≡roba a te e-ba are ba a-naŕke enti i [33, 17, 16] e

a

e

27/05/2016 19:31:27

J.12.3

J.12.4 J.14.1 J.15.1 J.15.2 J.15.3 J.16.1 J.16.2 J.16.3 J.16.5 J.16.6 J.17.1 J.17.2 J.17.3 J.17.4 J.18.1 J.18.2 J.19.1 J.19.2 J.20.1 J.21.1 J.22.1 J.22.2 J.23.1

(])uult iina ar-beieŕit uu la[. . . [18, 7, 9] salsaloi ?i [ | ]b a e b a a-lak i in ?i i [20, 8, 9] t a alainon t u uŕek u ui or[ | [ ] i [ ] | ]noś t a ae-b a are naŕk e en [34, 15, 13] hait u ura meleśae≡: :≡b a aenae *(*)n [23, 20, 0] ]n / aŕ[k e e(n) . . . [3, 0, 3] aalaein ŕe[ nar´]keeni [13, 7, 4] uursaau *arb a an t e e-bar[e] b a a-naŕk eent ii [27, 6, 21] e keeo-ion[ [30, 18, 5] ( )omuŕika a[ ]anba ati ia≡ioba a[ ]*e ba a-[na]ŕkeKoch itiiabe ŕebe anaka a | ro-ba are ba a-naŕke [e]nti i [28, 5, 15] uab a an : ne[ [8, 0, 5] ba]are[ [3, 0, 3] kuui k aaosa naŕ[k e e]n(t i i) [12, 0, 4] k u ui arairb u b u [u | ]b a are naŕke ent i i [22, 0, 12] baar]e naŕk ee[ [6, 0, 6] : k e enila (*) rin≡ | ≡b e e:lin enb e ·k a arne : [24, 15, 0] bo oti ieana≡ ke erto o ≡roba a te e-ba are ba a-naŕke enti i [33, 17, 16] Note. Inscription J.18.2: in this position ]an is very ]an teee-ro-baare na[ŕk e e(. . .) [13, 0, 11] probably a fragment of uarbaan, in which case the ]liirnest a ak u un b a ane≡ooŕoire b a a[re naŕ]k e enii [30, 23, 7] score would be [13, 0, 13]. ooŕoir naŕk e enb i i [14, 6, 8] ]uŕni be eliśon uarn|ba an e* ba ar(e)n naŕke en[ . .[27, 6, 17] . . .]uarb a an t ee[(e)-ro-b a]are naŕk eenii [19, 0, 19]

[ 463 ]

uarb o oiir sarune ea b a are naŕk e enii [28, 14, 12] ]sarune ea oar[ [11, 6, 0] beet iisai t eee-b aarent ii iru≡|(u)arb uu i el naŕrk e e:n: uśnee [40, 19, 18] ]la***b a[ ]*e albooroi[ ] isakaaoeaŕte[ ]atoore/]kaaae[ [30, 7, 0]

J.24.1 J.26.1 ]t a arnek u 〈k u 〉un≡b a ane |[ro-]b a are naŕk e [e . . [20, 12, 8] J.27.1 ]ukee śaen baare naŕkee* [ ] beeś**n*[ [20, 0, 9] J.51.1 ]kiu | ]**baa | ]tuurkaaio[ [11, 7, 0] J.52.1 ]ire ab are la[ read ]ire ba are la[k [9, 0, 4] J.53.1 ko -tu -ua-ratee tunbi ite sba an orba a≡seta a lake entii raha≡ka aśeta ana || bo be ko oŕbe o ba arlete [57, 17, 0] J.54.1 ]*ik eei**[ ]*uosor ert aau[ne [14, 0, 0] J.55.1 ro- k o olion eert a aune t a arielnon : liŕniene naŕke enai [42, 15, 10] J.56.1 ak o osioś naŕke et i i [14, 7, 7] J.57.1 ]lok o on k eeloia nar´k e [e . . | li[ | b a a[re [19, 6, 6] Medellín T2 86H/13-1 tetunae śnoror [11, 5, 0] Monte Novo do Castelinho ]k ooloion : k o oloar[ ]ŕ[.]s[?k e]nti i [19, 7, 4] Corte Freixo 2 naŕk ee( )uu[ [7, 0, 5] São Martinho b a ast e eb u uŕoi onunaio t e ?e […]i[…]o*reiar*nio eb u u alak i imuŕbo ? a naŕk e e b a a* | ean b a ara | b o ? [60, 10, 11] Cabeza del Buey IV ]ki iu [---] ke eilau ke e iśa n[a]ŕke en [19, 6, 5] Mesas do Castelinho :tt i ilek u urk u u≡ark a ast a amu t e e-ba anti i leb o oiire ro-b a are naŕk e [e(n)-o]lak iiuu lii*eianiit a a ea nira-k a alt e e t aao b e e saru[?n]an [82, 41, 17] Castelo do Moura ]*nab aora[ [6, 0, 0] Corte Pinheiro b e eu*[ ]ae* b a [a]re [na]ŕk e eni [13, 0, 8] Vale de Águia ]*******b a t e e-b a are na[ŕk e e---] [9, 0, 8] Monte Gordo uuŕerkaar ua[rbaa]n kiikee≡arkaare ro-n-baare na[ŕ]keen taa-bee anoŕ-ion [44, 18, 15] Abul t u ur ? n[ [5, 5]

totals: readable signs probable onomastics formula words naŕkeent i i, &c.

1752 100% 590 33.7% 581 33.2% 279 15.9%

The inscription of Alcalá del Río, Sevilla (J.53.1): a ‘smoking gun’ for the Hamito-Semitic hypothesis? The Hamito-Semitic substratum hypothesis of the Insular Celtic languages, as usually framed (e.g. Gensler 1993; Jongeling 2000; Isaac 2007b arguing the contrary; Hewitt this volume), supposes that Celtic, when introduced to Britain and Ireland, spread amongst a population that had previously spoken a language related to Semitic, Berber, or Egyptian. It is as a result of such a non-Indo-European substratum that the syntax of attested Brythonic and Goidelic shows features otherwise unusual in Indo-European, but characteristic of the Afro-Asiatic languages, the macro-family which is also known as Hamito-Semitic. The

Koch CW3 16 iii 16.indd 463

27/05/2016 19:31:27

[ 464 ]

xv. phoenicians in the west

]]

various versions of the hypothesis differ somewhat in the lists of syntactic features that they see as indicative of linguistic contact (see especially Hewitt this volume). However, the predominance of the verb-initial sentence type has always been viewed as especially significant, as well as genitives and adjectives regularly following the noun they modify.   Formulating the hypothesis in this way, the arrival of the Phoenicians on the Atlantic by c. 950/900 BC, followed by a millennium of contact of Phoenician and Punic with Hispano-Celtic, is not the mechanism of contact usually envisioned. Vennemann can be seen as exceptional in this regard, in characterizing the influence, ‘The European Atlantic Littoral was, at the dawn of history, explored and colonized by Mediterranean, probably Palaeo-Phoenician seafarers’ (2001, 351).   There are reasons that the well-documented interaction of Semitic and Celtic speakers in the Iberian Peninsula has most often been overlooked in work on the Hamito-Semitic Hypothesis. First, the suspected effects of the contact have been found entirely in Goidelic and Brythonic, far away from the geographical area of documented contact between the Ancient Celtic and West Semitic languages. Furthermore, the sentential syntax attested for Gaulish, Lepontic, and Celtiberian—although extremely limited and almost all of it far from areas with evidence for contact with the Phoenicians and/or Carthaginians—have been seen as enough to prove that syntactic features under consideration were entirely absent from all varieties and stages of Continental Celtic.   If the SW corpus is accepted—as I argue it should be—as Celtic overall and not just containing a sizeable proportion of Celtic names, this would not alter this picture greatly. Although, as can be seen in the collection above, the SW texts vary a great deal, the most characteristic syntactic arrangement in the corpus is: #naming phrase uarbaan tee-robaare baa-naŕkeentii#. Therefore, the sequence of signs that looks most clearly like an Indo-European verb characteristically appears at the end of the statement. The formula words tee-ro-baare that characteristically precede naŕkee- look like another verb—particularly in view of teee-baarentii (J.23.1)—this one 15.14 Drawing of the SW inscribed stone from with preverbs. uarbaan preceding tee-ro-baare Alcalá del Río, Sevilla, Spain (J.53.1) conforms for­mal­ly to an accusative singular. Thus, the stereo­typical SW epigraphic statement appears not to exemplify the Insular Celtic/Hamito-Semitic syntactic type, but is rather in accord with the usual early Indo-European verb-final type.   Of the variations within the SW corpus, a particularly striking exception is the complete but now lost inscription of Alcalá del Río, Sevilla (J.53.1; Figure 15.14), where the main text reads: ko‑tu-uaratee tunbiitesbaan orbaa setaa lakeentii raha kaaśetaana. As Werner Nahm points out to me, the name raha is very probably Semitic, as is made clear by comparing Palaeohispanic examples of the Roman Period, RAPPA attested in the territory of the south-western Celtici and RAPETIGVS MEDICVS

Koch CW3 16 iii 16.indd 464

27/05/2016 19:31:28

Koch

[ 465 ]

commemorated in Rome. West Semitic rapa means ‘heal, healer, healing’ and is attested in Aramaic, Phoenician, and Hebrew, including many instances in the Hebrew Bible (Hoftijzer & Jongeling 1995, sn. rapa). I am again indebted to Werner Nahm for noting the significance of the detail that the [p] of Semitic rapa is represented at Alcalá del Río by a sign derived from the Phoenician/Canaanite letter with the phonetic value [h]. We see from this that the name was borrowed into a Palaeohispanic language that subsequently weakened [p] to [h], or it was borrowed into a language that already lacked [p], so that [h] was substituted for the foreign sound.30 The later comparanda RAPPA and RAPETIGVS date from a period when the Palaeohispanic languages had acquired [p]. Students of the history of the Irish language might thus understand raha as analogous to a loanword of the earlier p-less stratum (or so-called Cothraige words), and RAPPA analogous to the later Pátraic stratum.   Along with the Semitic name, the Alcalá del Río inscription has several further unusual features; it is unlikely that they all co-occur as a matter of coincidence: 1. Unlike the bulk of the SW corpus, J.53.1 was found near the lower Guadalquivir, in the core region of the Orientalization of the Early Iron Age, i.e. Ταρτησσος proper. 2. The name raha is found alongside and in apparent case concord with kaaśetaana, which can be understood as a feminine form equivalent to the Gaulish title cassidannos, probably meaning ‘overseer of tin’ or ‘of bronze’ (De Bernardo Stempel 1998; Delamarre 2003), cf. Gaulish A RG A NTODA NNOS ‘overseer of silver’. 3. None of the words of the epigraphic formula is used. However, words resembling the morphology of the formula words take their place and can all be supplied with Indo-European or specifically Celtic etymologies carrying a similar meaning. Thus, tee-ro-baare (verb third sg. perfect and preverbs ‘has carried away’) is omitted, but there is ko‑tu-ua-ratee (verb third sg. perfect and preverbs ‘has delivered to’). uarbaan (superlative feminine accusative singular of destination ‘highest [one, place]’) is absent, but there is tunbiitesbaan, which can be interpreted as |tumitesa man| (superlative feminine accusative singular of destination ‘to the greatest tumulus’), and there is also orbaa, possibly nominative |o̯ora m\| < *u(p)era m\, a phonological byform of accusative uarbaan < *u(p)era m\m. The standard closing with a form of naŕkeentii is absent, but there is lakeentii ‘(they) now lie down’ with the same verbal ending. 4. The forms resembling verbs stand at the beginning of their clauses rather than ending them: #ko‑tu-ua-ratee tunbiitesbaan orbaa setaa with the suggested translation ‘the highest throne has delivered [the deceased] to the greatest tumulus’ and #lakeentii raha kaaśetaana with the suggested translation ‘Raha the Bronze Minister now lies down’. 5. There are six lapses of the usual redundancy between the semisyllabogram representing CIVIS HISPANIS

30 The sign derived from Phoenician/Canaanite he is not common in the SW corpus, but the other examples are also likely to belong to forms with an earlier *[p]. I derive hataaneatee (‘Abóboda 1’ J.12.1 — Gomes Aires, Almodôvar, Beja) < PC *(p)atani̯atei ‘for the winged one’ (likewise Kaufman 2015). The feminine name haituura (‘Pardieiro 1’ J.15.1 — São Martinho das Amoreiras, Odemira, Beja) can be compared with Aetvrae Arqvi f. (CIL II, 2465 — Valença, Viana do Castelo) and the place-name Baeturia, all possibly reflecting an original *Paitur(-) with weakening of *p.

Koch CW3 16 iii 16.indd 465

27/05/2016 19:31:28

[ 466 ]

xv. phoenicians in the west

]]

a stop consonant plus a vowel after which the vowel itself is also written in most SW inscriptions. 6. The text is written left-to-right rather than the more common right-to-left. This constellation of features is consistently different from the norm in the lands of the Κυνητες west of Guadiana. Given the interpretation by Brandherm in this volume and my own above that the tradition of SW inscriptions represents a nativist reaction against Orientalization, the Alcalá del Río inscription can be interpreted as the other side of the coin. A cultural differentness and possible antipathy is articulated from both sides.   We must remember nevertheless that this stela from the Orientalizing zone is not in Phoenician, but written in the SW script and language. Some centuries earlier, pioneering Semitic seafarers and the indigenous people of the Atlantic Iberia had confronted each other as uncomprehending representatives of different worlds. But by the era of the SW corpus a once vast gap of culture and communication had grown inward towards sectarianism—nativism and anti-nativism articulating competing claims within an overlapping cultural space.   When the six features listed above are weighed together, they are closer to Vennemann’s epitome of the Hamito-Semitic hypothesis. Raha is identified as of Semitic background. Like her forbears, she is engaged in the metal trade. This is the basis of her special social status. She is commemorated in the core lands of the Orientalization, where she and her Phoenician or Phoenicianized ancestors had probably lived and operated. Her memorial does not use the standard SW formula, for understandable reasons. Owing to her cultural background, she probably worshipped different gods and had afterlife beliefs differing from those of the Κυνητες. Her epitaph is in Celtic, but with a verb-initial syntax like Semitic. Presumably this was the variety of Celtic as naturalized in her community.   We know that, when the Phoenicians became established on the Atlantic at the emporium of Huelva, this occasioned an economically-driven cultural reorientation. A great demand for metals in the eastern Mediterranean was met. There was also a technological leap forward: Iberia abruptly entered the Iron Age. As well as revolutionizing the production and exchange of metals, the arrival of Phoenicians brought new and superior techniques of shipbuilding and navigation.   What were the linguistic implications? When the Phoenicians reached the Atlantic, they did not find an undeveloped wilderness. There was a long-established and vigorous bronze industry, and an existing maritime network across which cultural values and complex information were efficiently exchanged and rapidly updated. The intensity and complexity of the connections of the Atlantic Bronze Age amount to a circumstantial case for a lingua franca, which had possibly evolved in situ in the region since the beginnings of the Bronze Age. For numerous reasons, Proto-Celtic would be the likely candidate for this shared language.   When the Phoenicians encountered this Atlantic system, they did not come in a mass migration or conquest. They were specialists whose success depended on building symbiotic relations with indigenous peoples. They fostered good will with hereditary landholders and military aristocracies who controlled territories with mineral resources and could supply necessities. In this situation, the Phoenicians were in no position to impose their language on their new partners, though they excelled in some socio-economic domains. Circumstances

Koch CW3 16 iii 16.indd 466

27/05/2016 19:31:28

Koch

[ 467 ]

dictated that they should learn to a functional level the pre-existing lingua franca of the Atlantic system.   As explained above, when adults must learn a new language well enough to operate in a foreign society, what usually happens is that they learn the vocabulary and morphology completely, or nearly so. They can ‘get by’ with that while retaining much of the syntax and phonetics of their first language, which are less of a bar to intelligibility. This tolerance of the imperfections of the second-language speakers is characteristic of lingua francas, as their users expect these languages to be spoken by people having different first languages and divergent regional dialects. Once the Phoenicians had mastered the Atlantic lingua franca to sufficient intelligibility and had secured an advantageous economic niche in the exchange networks, there might have been no reason for them to erase the linguistic peculiarities of the founder generation. Groups whose identities are not stigmatized but advantageous will be in no hurry to change the way they speak.   If this is truly the background of Hamito-Semitic affinities of the medieval and modern Celtic languages, why should it surface most prominently in Insular Celtic? The impression that this is definitely the case may change as more Continental Celtic evidence becomes known. However, such an effect is not so surprising in the context of the picture sketched here for the simple reason that Britain and Ireland are islands separated from the mainland and each other by the formidable north Atlantic. Their exposure to the lingua franca had always been brought by master seafarers. More often in the lifetimes of most Irish and British people during the Bronze Age the greater world was encountered as small groups of specialists pulling ashore in impressive ocean-going vessels, rather than as invading armies or tribal communities relocating en masse. When the most accomplished sailers and peripatetic metal workers came to speak the lingua franca in new ways, this will have had particular impact for the reception and renewal of a supra-regional language in the more remote insular circuits of the network.   As a sociolinguistic phenomenon, we may think of the metropolitan version of a lingua franca, even though embodying subtstratum features from the speech of recent immigrants, transforming varieties of the lingua franca spoken in cultural backwaters, even when these were the product of continuous native-speaker transmission over many generations. The relationship of the English of New York City versus dialects of Appalachia can be described in this way. This modern parallel of course differs in many respects from the hypothetical ancient situation posed here, one being, as the Huelva town find has lately revealed, that the transformative incomers reaching Atlantic Europe at the Bronze–Iron Transition had come overwhelmingly from one particular place and spoke one urban dialect of one foreign language.   Beyond this interpretation of the single inscription from Alcalá del Río, the more general suggestion offered is that proponents of the Hamito-Semitic Hypothesis should not overlook the period between about 1000 BC and the 1st century AD when we know Semitic and Celtic were in contact in the Iberian Peninsula. However, there is an obvious obstacle to explaining the syntactic features of the Insular Celtic languages with this prolonged episode. As has been the main focus of this chapter, archaeological evidence shows a significant cultural disconnection at the time of the Peninsula’s Bronze–Iron Transition c. 900 BC,

Koch CW3 16 iii 16.indd 467

27/05/2016 19:31:28

[ 468 ]

xv. phoenicians in the west

]]

which corresponds in time and place to the primary split within the Celtic branch of IndoEuropean. The Phoenician influence catalysed this break by being unevenly strong in the Iberian Peninsula, but weak in the territories of Gaulish-Brythonic-Goidelic subgroup. Therefore, if linguistic influence emenating from the Phoenicians actually penetrated further north along the Atlantic Façade, a more plausible context for this occurred specifically at the Ría de Huelva phase (1050–950/900 BC) preceding the Baiões–Plaza de las Monjas phase (950/900–825 BC). At the earlier stage, prestige imports from the eastern Mediterranean were finding their way into the still vigorous and integrated Atlantic Bronze Age networks, and likewise into the iconography of the Iberian warrior stelae (e.g. lyres, combs, and mirrors, alongside spears, shields, and swords of Atlantic type). We may think of a prolonged phase of freelance exchange, preceding the large-scale enterprise initiated by the rulers of Tyre. In other words, towards the turn of the 2nd and 1st millennia BC, there was a century or so of dynamic cultural and economic devlopment during which the Atlantic and Mediterranean networks overlapped and interacted. It is possible that the Celtic spoken in the Atlantic Zone retained some imprint from this phase into the Iron Age and afterwards. Sketch summary Before the Phoenicians reached south-western Europe, Indo-European had come into contact with a different non-Indo-European language or languages there, speech similar to the attested Iberian and Aquitanian/Palaeo-Basque. Some of the phonological innovations that define Celtic—such as the weakening and loss of *p—can be understood as the result of this contact. The sociolinguistic background for the emergence of Proto-Celtic was that a relatively large number of speakers of non-Indo-European learned Indo-European as a second language in this region.   The common Hispano-Celtic name Arquius, which means ‘archer’, has no counterpart in the other Celtic languages. This disparity can be correlated with the fact that archery survived vigorously from the Beaker period in the Iberian Peninsula and south-west France, but died out during the Middle Bronze Age in other parts of Gaul and Britain and Ireland. It seems, therefore, that there were regional differences within Celtic as early as this, parallel to sociocultural differences observable across Atlantic Europe.   The arrival of the Phoenicians on the Atlantic by c. 900 BC can be correlated with three developments in Celtic-speaking areas. First, their presence transformed the economy and culture of the Iberian Peninsula to such an extent as to break up the cultural continuum of the Atlantic Bronze Age. This probably accelerated the drift of Hispano-Celtic away from the Celtic of Gaul, Britain, and Ireland. Second, Phoenician cultural influence catalysed— partly as imitation and partly as nativist reaction—the earliest writing of an indigenous language in the west. These SW inscriptions contain a recognized Celtic element and are probably mainly in a Celtic language. Third, the SW inscription of Alcalá del Río underscores the possibility that the resemblance of the syntax of Insular Celtic to that of Afro-Asiatic—West Semitic especially—owes something to influential Phoenician mariners who had learned Celtic as a second language.

Koch CW3 16 iii 16.indd 468

27/05/2016 19:31:28

Koch

[ 469 ]

  Where the Phoenicians set up full-scale colonies—such as Kition on Cyprus, Carthage, Gadir (Cádiz) and elsewhere along the Peninsula’s south coast—their society and culture as a whole were transplanted. This included language, religion, political and economic structures. Further outward in Atlantic Europe their impact was more complex and uneven, rippling through as both actions and reactions. In many situations, their influence can be seen as hastening the fragmentation of the longstanding but fragile unities of the Bronze Age. The Atlantic Bronze Age exchange network was one casualty of the Bronze–Iron Transition. And Proto-Celtic as a continuum of dialects (with a high degree of mutual intelligibility) gave way to a patchwork of fully separated languages.

BIBLIOGRAPHY CIL II = Hüber, E. 1869–92 Corpus Inscriptionum Latinarum II. Inscriptiones Hispaniae Latinae. Berlin. Inscriptionum Latinarum Supplementum. Berlin. CIL VI = Henzen, G., I. B. De Rossi, E. Borman, C. Huelsen, M. Bang 1876–1989 Corpus Inscriptionum Latinarum VI. Inscriptiones Urbis Romae Latinae. Partes I–VII. Berlin. CIL II 2/5 = Stylow, A. U., R. Atencia, J. González, C. González, M. Pastor, P. Rodríguez 1998 Corpus Inscriptionum Latinarum II. Inscriptiones Hispaniae Latinae. Editio Altera. Pars V. Conventus Astigitanus. Berlin / New York. CIL 2/14 = Alföldy, G., M. Clauss, M. Mayer 1995 Corpus Inscriptionum Latinarum II. Inscriptiones Hispaniae Latinae. Editio Altera. Pars XIV, fasiculis I. Pars meriodinialis conventus Tarraconenis. Berlin / New York. CPILC = Hurtado de San Antonio, R. 1977 Corpus provincial de inscripciones latinas (Cáceres). Cáceres, Diputación Provincial de Cáceres. EE = Ephemeris epigraphica. ERAsturias = Diego Santos, F. 1985 Epigrafía romana de Asturias. Oviedo. ERAv = Hernando Sobrino, M. R. 2005 Epigrafía romana de Ávila, Petrae Hispaniarum 4. Burdeos. ERCAN = Iglesias, J. M., & A. Ruiz 1998 Epigrafía romana de Cantabria. Bordeaux / Santander. ERPL = Rabanal Alonso, M. A., & S. M.ª García Martínez 2001 Epigrafía romana de la provincia de León: revisión y actualizaciones. León. ERRBragança = Redentor, A. 2002 Epigrafia romana da região de Bragança, Trabalhos de Arqueologia 24. Lisboa. ERZamora = Bragado Toranzo, J. M.a 1991 Fuentes

Koch CW3 16 iii 16.indd 469

literarias y epigráficas de la provincial de Zamora y su relación con las vías romanas de la Cuenca del Duero. Universidad de León. FE = Ficheiro Epigráfico. Suplemento de Conimbriga. Coimbra. GPC = Geiriadur Prifysgol Cymru, A Dictionary of the Welsh Language. Caerdydd, Gwasg Prifysgol Cymru, 1950–2002. HAE = Hispania Antiqua Epigraphica, Supplemento annual de Archivo Español de Arqueología. Madrid, Consejo Superior de Investigaciones Científicas. HEp = Hispania Epigraphica, Archivo Epigráfico de Hispania. Madrid, Universidad Complutense. ILER = Vives, J. 1971–2 Inscripciones latinas de la España romana I–II. Barcelona. IRLugo = Arias, F., P. Le Roux & A. Tranoy 1979 Inscriptions romaines de la province de Lugo. Paris, Diffusion de Boccard. IRPL = Diego Santos, F. 1986 Inscripciones romanas de la provincia de León. León, Diputación Provincial de León. IRPP = Hernández Guerra, L. 1994 Inscripciones romanas en la provincia de Palencia. Valladolid. LIV = Rix, H. 2001 Lexikon der indogermanischen Verben. Wiesbaden, Ludwig Reichert. MLH IV = Untermann, J. (mit D. S. Wodtko) 1997 Monumenta Linguarum Hispanicarum. Band IV. Die tartessischen, keltiberischen und lusitanischen Inschriften. Wiesbaden. Abásolo, J. A. 1974 Epigrafía romana de la región de Lara de los Infantes. Burgos, Diputación Provincial. Abascal Palazón, J. M. 1994 ‘Inscripciones romanas y celtibéricas en los manuscritos de Fidel Fita en la Real Academia de la Historia’, Archivo de Prehistoria

27/05/2016 19:31:28

[ 470 ]

xv. phoenicians in the west

Levantina 21, 367–390. Valencia. Ahlqvist, A., 1982 The Early Irish Linguist, Commentationes Humanarum Litterarum 73, Helsinki. Alarcão, J. de 2001 ‘Novas perspectivas sobre os Lusitanos (e outros mundos)’, Revista Portuguesa de Arquelogia 4/2, 293–349. Albertos Firmat, M.a L. 1985 ‘La onomástica personal indígena del noroeste peninsular (astures y galaicos)’, Actas del III Coloquio sobre Lenguas y Culturas Paleohispánicas, ed. J. de Hoz, 255–310. Salamanca, Universidad de Salamanca. Alföldy, G. 1975 Die römische Inschriften von Tarraco I–II. Berlin. Allen, D. F. 1980 The Coins of the Ancient Celts, ed. D. Nash. Edinburgh, Edinburgh University Press. Allentoft, M. E., M. Sikora, K.-G. Sjögren, S. Rasmussen, M. Rasmussen, J. Stenderup, P. B. Damgaard, H. Schroeder, T. Ahlström, L. Vinner, A.S. Malaspinas, A. Margaryan, T. Higham, D. Chivall, N. Lynnerup, L. Harvig, J. Baron, Ph. Della Casa, P. Dąbrowski, P. R. Duffy, A. V. Ebel, A. Epimakhov, K. Frei, M.Furmanek, T. Gralak 2015 ‘Population genomics of Bronze Age Eurasia’, Nature 522, 167–72. Almagro-Gorbea, M. 2001 ‘Cyprus, Phoenicia and Iberia: from “Precolonization” to colonization in the “Far West”’, Italy and Cyprus in Antiquity 1500– 450 BC, ed. L. Bonfante & V. Karageorghis, 239–70. Nicosia, Severis Foundation. Almagro-Gorbea, M. 2004 ‘Inscripciones y grafitos tartesicos de la necrópolis orientalizante de Medellín’, Palaeohispanica 4, 13–44. Almagro-Gorbea, M. 2005 ‘La literatura tartésica. ­Fuentes históricas e iconográficas’, Gerión 23/1, 39–80. Almagro-Gorbea, M. 2008 ‘Inscripciones tartésicas y grafitos’, La necrópolis de Medellín II: estudio de los hallazgos (Bibliotheca Archaeologica Hispanica 26–2/Studia Hispano-Phoenica 5–2), ed. M. Almagro-Gorbea, 751–71. Madrid, Real Academia de la Historia. Almagro-Gorbea, M. 2010 ‘Paleodemografía de la necrópolis tartesia de Medellín’, Revista de Demografía Histórica, XXVIII, I, 2010, segunda época, 33–70. Almagro-Gorbea, M., A. J. Lorrio, A. Mederos, & M. Torres 2008 La necrópolis de Medellín III: estudios analíticos; IV: interpretacíon de la necrópolis; V: el marco histórico de Medellín-Conisturgis (Bibliotheca Archaeologica Hispanica 26–3/Studia HispanoPhoenica 5–3). Madrid, Real Academia de la Historia. Almagro-Gorbea, M., & M. Torres Ortiz 2009 ‘La colonización de la costa atlántica de Portugal: ¿Fenicios o Tartesios?’ Palaeohispanica 9, 113–42. Almeida, D. F. de 1956 Egitânia. História e arqueologia.

Koch CW3 16 iii 16.indd 470

]]

Lisboa, Universidade de Lisboa. Armada, X.-L. 2011 ‘Feasting Metals and the Ideology of Power in the Late Bronze Age of Atlantic Iberia’, Guess Who’s Coming to Dinner? Feasting Rituals in the Prehistoric Societies of Europe and the Near East, ed. G. Aranda Jiménez, S. Montón-Subías, & M. Sánchez Romero, 158–83. Oxford, Oxbow Books. Armada, X.-L., N. Rafel, & I. Montero 2008 ‘Contactos precoloniales, actividad metalúrgica y biografás de objetos de bronce en la Península Ibéria’, Contacto cultural entre el Mediterráneo y el Atlántica (siglos XII–VIII ANE). La precolonización a debate, eds S. Celestino, N. Rafel, & X.-L. Armada, 465–508. Madrid, Escuela Española de Historia y Arqueolgía en Roma-CSIC. Aubet, M. E. 2001 The Phoenicians and the West: Politics, Colonies, and Trade, trans. M. Turton. 2nd edn. Cambridge, Cambridge University Press. (1st English edition, 1993) Aubet, M. E. 2008 ‘Political and Economic Implica­ tions of the New Phoenician Chronologies’, Beyond the Homeland: Mariners in Phoenician Chronology, Ancient Near Eastern Studies Supplement 28, ed. C. Sagona, 247–60. Leuven, Peeters. Ballester, X. 2004 ‘Hablas indoeuropeas y anindoeuropeas en la Hispania prerromana’, Real Academia de Cultura Valenciana, sección de estudios ibéricos. Estudios de lenguas y epigrafía antiguas – ELEA 6, 107–38. Ballester, X. 2012 Falas Indo-Europeias e Anindo-Europeias na Hispânia Pré-Romana. Lisboa, Apenas Livros. Blázquez, J. M.a 1962 Religiones primitivas de Hispania. Roma, CSIC Delegación de Roma. Blázquez, J. M.a 2011 ‘Chipre y la Península Ibérica’. Fenicios en Tartessos: nuevas perspectivas, BAR International Series 2245, ed. M. Álvarez MartíAguilar, 7–31. Oxford, Archaeopress. Brandherm, D. 2007 Las Espadas del Bronce Final en la Península Ibérica y Baleares, Prähistorische Bronzefunde IV, 16. Stuttgart, Franz Steiner Verlag. Brandherm, D. 2013a ‘Westward Ho? SwordBearers and All the Rest of it . . .’, Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, Celtic Studies Publications 16, eds J. T. Koch & B. Cunliffe, 147– 156. Oxford, Oxbow Books. Brandherm, D. 2013b ‘Mediterranes, Atlantisches und Kontinentales in der bronze- und ältereisenzeitlichen Stelenkunst der Iberischen Halbinsel’, Petasos: Festschrift für Hans Lohmann zugeeignet von sein Schülern, Freunden und Kollegen su seinem 65. Geburtstag, ed. G. Kalaitzoglou & G. Lüdorf, 131–48. Mittelmeerstudien 2, Paderborn, Fink & Schöningh.

27/05/2016 19:31:28

Koch

Bochum, Ruhr-Universität. Broderick, G. 2010 ‘Die vorrömischen Sprachen auf der iberischen Halbinsel’, Handbuch der Eurolinguistik, ed. U. Hinrichs, 287–320. Wiesbaden, Otto Harrassowitz Verlag. Brotherton, P., W. Haak, J. Templeton, C. J. Adler, G. Brandt, R. Ganslmeier, S. Richards, J. Soubrier, M. van Oven, R. Kenyon, M. van der Hoek, J. Korlach, M. Gardner, J. Metcalf, S. Ho, L. Quintana-Murci, D. M. Behar, H. Meller, K. W. Alt, & A. Cooper 2013 ‘Neolithic mitochondrial haplogroup H genomes and the genetic origins of Europeans’, Nature Communications 4, 1764. Brun, P., B. Chaume, L. Dhennequin, & B. Quilliec 2009 ‘Le passage de l’âge du Bronze à l’âge du Fer … au fil de l’épée’, De l’âge du Bronze à l’âge du Fer en France et en Europe occidentale (Xe–VIIe siècle av. J.-C.). La moyenne vallée du Rhône aux âges du Fer. Actes du XXXe colloque international de l’A.F.E.A.F., co-organisé avec l’A.P.R.A.B. (Saint-Romain-en-Gal, 26–28 mai 2006)(27e suppl. à la R.A.E.), 477–85. Búa Carballo, J. C. 2000 ‘Estudio lingüístico de la teonima lusitano-gallega’, Tesis Doctoral, Salamanca. Burgess, C. 2012 ‘Alignments: Revising the Atlantic Bronze Age Sequence’, Archaeological Journal 169, 127–58. Burgess, C., & B. O’Connor 2004 ‘Bronze Age Rotary Spits: Finds Old and New, Some False, Some True’, From Megaliths to Metals: Essays in Honour of George Eogan, eds. H. Roche, E. Grogan, J. Bradley, J. Coles, & B. Raftery, 184–99. Oxford, Oxbow Books. Burgess, C., & B. O’Connor 2008 ‘Iberia, the Atlantic Bronze Age and the Mediterranean’, Contacto cultural entre el Mediterráneo y el Atlántico (siglos XII–VIII ANE), La precolonización a debate, Serie Arqueológica 11, eds. S. Celestino, N. Rafel & X-L. Armada, 41–58. Madrid, Escuela Española de Historia y Arqueología en Roma-CSIC. Cassidy, L. M., R. Martiniano, E. M. Murphy, M. D. Teasdale, J. Mallory, B. Hartwell, & D. G. Bradley 2015 ‘Neolithic and Bronze Age migration to Ireland and establishment of the insular Atlantic genome’, Proceedings of the National Academy of Sciences USA 113/2, 368–73. Castillo, C., J. Gómez-Pantoja & M.a D. Mauleón 1981 Inscripciones romanas del Museo de Navarra. Pamplona, Museo de Navarra. Celestino Pérez, S. 2001 Estelas de guerrero y stelas diademadas: La precolonización y formación del mundo tartésico. Barcelona, Edicions Bellaterra. Charles-Edwards, T. M. 1978 ‘The Authenticity of the Gododdin: An Historian’s View’, Astudiaethau ar yr Hengerdd: Studies in Old Welsh Poetry, eds R. Bromwich &

Koch CW3 16 iii 16.indd 471

[ 471 ]

R. B. Jones, 44–71. Caerdydd, Gwasg Prifysgol Cymru. Correa Rodríguez, J. A. 1989 ‘Posibles antropónimos en las inscripciones en escritura del SO. (o tartesia)’, Veleia 6.243–52. Correa Rodríguez, J. A. 1992 ‘La epigrafía tartesia’, Andalusien zwischen Vorgeschichte und Mittelalter, eds. D. Hertel & J. Untermann, 75–114. Cologne, Böhlau. Correa Rodríguez, J. A. 1993 ‘El signario de Espanca (Castro Verde) y la escritura tartesia’, Lengua y cultura en la Hispania prerromana, Actas del V Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, eds. J. Untermann & F. Villar, 521–62. Salamanca, Ediciones Universidad de Salamanca. Correa Rodríguez, J. A. 1995 ‘Reflexiones sobre la epigrafía paleohispánica de la Península Ibérica’, Tartessos 25 años después: Congreso Conmemorativo del V Symposium Internacional de Prehistoria Peninsular, 609–17. Ayuntamiento de Jérez de la Frontera. Correa Rodríguez, J. A. 1996 ‘La epigrafía del Sudoeste: Estado de la cuestión’, La Hispania Prerromana, Actas del VI Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, eds. F. Villar & J. d’Encarnação, 65–75. Salamanca, Ediciones Universidad de Salamanca. Correa Rodríguez, J. A. & C. Toscano-Pérez 2014 ‘Grafitos tartesios hallados en Niebla (Huelva) y su contexto arqueológico’, Revista Onoba 2, 45–54. Correia, V. H. 2009 ‘A escrita do sudoeste: uma visão retrospectiva e prospectiva’, Acta Palaeohispanica X / Palaeohispanica 9, 309–21. Cunliffe, B. 2001 Facing the Ocean: The Atlantic and its Peoples 8000 BC–AD 1500. Oxford, Oxford University Press. Cunliffe, B. 2008 Europe between the Oceans, 9000 BC– AD 1000. New Haven, Yale University Press. Cunliffe, B. 2013. Britain Begins. Oxford, Oxford University Press. De Bernardo Stempel, P. 1998 ‘Minima Celtica zwischen Sprach- und Kultur-geschichte’, Man and the Animal World: Studies . . . in memoriam Sándor Bökönyi, eds. P. Anreiter, L. Bartosiewucz, E . Jerem, & W. Meid, 601–10. Budapest, Archaeolingua. De Bernardo Stempel, P. 2006 ‘Language and Historiography of Celtic-speaking Peoples’, Celtes et Gaulois, l’Archéologie face à l’Histoire: Celtes et Gaulois dans l’Histoire, l’historiographie et l’idéologie moderne, ed. S. Rieckhoff, Bibracte 12/1, 33–56. De Bernardo Stempel, P. 2008 ‘Linguistically Celtic Ethnonyms: Towards a Classification’, Celtic and Other Languages in Ancient Europe, ed. J. L. García Alonso, 101–18. Salamanca, Ediciones Universidad de Salamanca.

27/05/2016 19:31:28

[ 472 ]

xv. phoenicians in the west

Delamarre, X. 2003 Dictionnaire de la langue gauloise: une approche linguistique du vieux-celtique continental. Collection des Hespérides. Paris, Errance. First published, 2001. Díaz-Guardamino Uribe, M. 2010 ‘Las estelas decoradas en la Prehistoria de la Península Ibérica’, PhD dissertation, Madrid, Universidad Complutense de Madrid. Díaz-Guardamino Uribe, M. 2012 ‘Estelas decoradas del Bronce Final en la Península Ibérica: datos para su articulación cronológica’, Sidereum Ana II: El Río Guadiana en el Bronce Final, Anejos de Archivo Español de Arqueología 62, ed. J. Jimenez Avila, 389–415. Merida, España, Consejo Superior de Investigaciones Científicas. Egurtzegi, A. K. 2013 ‘4. Phonetics and Phonology’, Basque and Proto-Basque: Language-Internal and Typological Approaches to Linguistic Reconstruction, Minority Language Studies vol 5, ed. M. MartínezAreta, 119–72. Frankfurt, Peter Lang. Encarnação, J. d’ 1975 Divindades indigenas sob o dominio romano em Portugal. Lisboa, Imprensa Nacional. Eska, J. F. 2014 ‘Comments on John T. Koch’s Tartessian-as-Celtic enterprise’, Journal of IndoEuropean Studies 42, 428–38. Evans, C. 1988 ‘Women’s Names in Early Brittany’, Proceedings of the First North American Congress of Celtic Studies, Ottawa 1986, ed. G. W. MacLennan, 545–53. Ottawa, Chair of Celtic Studies. Falileyev, A., with A. E. Gohil & N. Ward 2010 Dictionary of Continental Celtic Place-Names: A Celtic Companion to the Barrington Atlas of the Greek and Roman World. Aberystwyth, CMCS. Ferrer i Jané, J. 2009 ‘El sistema de numerales ibérico: avances en su conocimiento’, Acta Palaeohispanica X / Palaeohispanica 9, 451–79. Fortson, B. W. 2009 Indo-European Language and Culture: An Introduction. Oxford, Wiley-Blackwell. Freeman, P. M. 2010 ‘Ancient References to Tartessos’, Celtic from the West: Alternative Perspectives from Archaeology, Genetics, Language and Literature, Celtic Studies Publications 15, eds B. Cunliffe & J. T. Koch, 303–34. Oxford, Oxbow Books. Garrido Roiz, J. P. 1983 ‘Presencia fenicia en el área atlántica adnadulza: La necropolis Orientalizante de Huelva (La Joya)’, Atti del I Congreso Internazionale di Studi Fenici e Punicia III, 857–63. Gensler, O. D. 1993 ‘A Typological Evaluation of Celtic/Hamito-Semitic Syntactic Parallels’. Ph.D. thesis, University of California at Berkeley. Gerloff, S. 2004 ‘Hallstatt Fascination: “Hallstatt” Buckets, Swords and Chapes from Britain and

Koch CW3 16 iii 16.indd 472

]]

Ireland’, From Megaliths to Metals: Essays in Honour of George Eogan, eds H. Roche, E. Grogan, J. Bradley, J. Coles, & B. Raftery, 124–54. Oxford, Oxbow Books. Gerloff, S. (with J. P. Northover) 2010 Atlantic Cauldrons and Buckets of the Late Bronze and Early Iron Ages in Western Europe: With a Review of Comparable Vessels from Central Europe and Italy, Prähistorische Bronzefunde II, 18. Stuttgart, Franz Steiner. Gimeno, H., & A. U. Stylow 1993 ‘Juan Pérez Holguín y la epigrafía trujillana’, Veleia 10, 117–78. González de Canales, F., L. Serrano, & J. Llompart 2004 El emporio fenicio precolonial de Huelva (ca. 900770 a.C.). Madrid, Editorial Biblioteca Nueva. González de Canales, F., L. Serrano, & J. Llompart 2006 ‘The precolonial Phoenician emporium of Huelva, ca. 900–770 BC: A résumé’, BAbesch (Bulletin Antieke Beschaving, Annual Papers on Classical Archaeology) 81, 13–29. González de Canales, F., L. Serrano, & J. Llompart 2008 ‘The Emporium of Huelva and Phoenician Chronology: Present and Future Possibilities’, Beyond the Homeland: Mariners in Phoenician Chronology. Ancient Near Eastern Studies Supplement 28, ed. C. Sagon, 631–655. Leuven, Peeters. González de Canales, F., L. Serrano, & J. Llompart 2010 ‘Tarshish and the United Monarchy of Israel’, ANES 47, 136–63. González de Canales, F., A. Montano, L. Serrano, & J. Llompart 2012 ‘Phoenician Bronze Lion’s Head Protome Found in Huelva’, Actas do V Encontro de Arqueologia do Sudoeste Peninsular (Almodôvar, Portugal, 18-20 November 2010), 289–99. Almodôvar, Digital Edition. González Rodríguez, M.a C. 1986 Las unidades organizativas indígenas del área indoeuropea de Hispania, Veleia Anejo 2. Vitoria/Gasteiz. Gorrochategui Churruca, J. 2013 ‘Linguistisque et peuplement en Aquitania’, L’âge du Fer en Aquitaine et sur ses marges. Mobilité des hommes, diffusion des idées, circulation des biens dans l’espace européenà l’âge du Fer. Actes du 35e Colloque international de l’AFEAF (Bordeaux, 2–5 juin 2011), Aquitania Supplément 30, dir. A. Colin, F. Verdin, 17–32. Hamp, E. P. (with D. Q. Adams) 2013 The Expansion of the Indo-European Languages: An Indo-Europeanist’s Evolving View, Sino-Platonic Papers 239, Philadelphia, Department of East Asian Languages and Civilizations, University of Pennsylvania. Haak, W., I. Lazaridis, N. Patterson, N. Rohland, S. Mallick, B. Llamas, G. Brandt, S. Nordenfelt, E. Harney, K. Stewardson, Q. Fu, A. Mittnik, E. Bánffy, C. Economou, M. Francken, S. Friederich,

27/05/2016 19:31:28

Koch

R. Garrido Pena, F. Hallgren, V. Khartanovich, A. Khokhlov, M. Kunst, P. Kuznetsov, H. Meller, O. Mochalov, V. Moiseyev, N. Nicklisch, S.L. Pichler, R. Risch, M. A. Rojo Guerra, C. Roth, A. SzécsényiNagy, J. Wahl, M. Meyer, J. Krause, D. Brown, D. Anthon, A. Cooper, K. W. Alt, & D. Reich 2015 ‘Massive migration from the steppe was a source for Indo-European languages in Europe’, Nature 14317. Harrison, R. J. 2004 Symbols and Warriors: Images of the European Bronze Age. Bristol, Western Academic & Specialist Press. Hernando Sobrino, M.a del R. 2007 ‘El Padre Andrés Marcos Burriel y la procedencia de CIL II 2323 Y CIL II 2324’, Gerión 25/1, 489–500. Herodotus, Herodotus: The Histories, trans. T. Holland, 2013. London, Penguin. Hoftijzer, J. & K. Jongeling 1995 Dictionary of the NorthWest Semitic Inscriptions. Leiden/New York/Köln, E. J. Brill. Hoz, J. de 2007 ‘The Institutional Vocabulary of the Continental Celts’, Gaulois et Celtique continental, eds. P.-Y. Lambert & G.-J. Pinault, 189–214. Genève, Librairie Droz. Hoz, J. de 2010 Historia lingüística de la Península Ibérica en la antigüedad. I. Preliminarties y mundo meridional prerromana, Manuales y Anejos de “Emerita”. Madrid, Consejo Superior de Investigaciones Científicas. Hoz, J. de 2011 Historia lingüística de la Península Ibérica en la antigüedad. II. El mundo ibérico prerromano y la indoeuropeización, Manuales y Anejos de “Emerita”. Madrid, Consejo Seperior de Investigaciones Científicas. Isaac, G. R. 2004 ‘The Nature and Origins of the Celtic Languages: Atlantic Seaways, Italo-Celtic and Other Paralinguistic Misapprehensions’, Studia Celtica 38.49–58. Isaac, G. R. 2005 ‘Insular Celtic vs Gallo-Brittonic: an Empirical and Methodological Question’, Celtic Connections: Papers from the Tenth International Congress of Celtic Studies, Edinburgh, 1995, 2. Archaeology, Numismatics, Historical Linguistics, ed. W. Gillies & D. W. Harding, 190–202. University of Edinburgh Archaeology Monograph Series 2. Isaac, G. R. 2007a Studies in Celtic Sound Changes and their Chronology. Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Isaac, G. R. 2007b ‘Celtic and Afro-Asiatic’. The Celtic Languages in Contact: Papers from the Workshop within the Framework of the XIII International Conference of Celtic Studies. Bonn 26–27 July, 2007, ed. H. L. C. Tristram, 25–80. Potsdam, Potsdam University Press.

Koch CW3 16 iii 16.indd 473

[ 473 ]

Jackson, K. H. 1967 A Historical Phonology of Breton. Dublin, Dublin Institute for Advanced Studies. Jongeling, K. 2000 Comparing Welsh and Hebrew. CNWS Publications 81. Leiden, Research School of Asian, African and Amerindian Studies, Universiteit Leiden. Jordán Cólera, C. 2005 Celtibérico. Zaragoza, Ediciones del Departamento de Ciencias de la Antigüedad. Jordán Cólera, C. 2006 ‘[K.3.3]: Crónica de un teicidio anunciado’, Real Academia de Cultura Valenciana, sección de Estudios Ibéricos “D. Fletcher Valls”, Estudios de lenguas y epigrafía antiguas – ELEA 7, 2006, 37–72. Jordán Cólera, C. 2007 ‘Celtiberian’, e-Keltoi 6: The Celts in the Iberian Peninsula, 749–850. Kaiser, J. M.a 2003 ‘Puntas de flecha de la Edad del Bronce en la Península Ibérica. Producción, circulación y cronología’, Complutum 14, 73–106. Karageorghis, V. 1969 Salamis in Cyprus: Homeric, Hellenistic and Roman. London, Thames and Hudson. Karageorghis, V. & F. Lo Schiavo 1989 ‘A West Mediterranean Obelos from Amathus’, Rivista di Studi Fenici 17, 15–29. Kaufman, T. 2015 Notes on the Decipherment of Tartessian as Celtic, Journal of Indo-European Monograph Series 62. Washington DC, Institute of the Study of Man. Koch, J. T. 1995 ‘Further Thoughts on Indo-European gwh in Celtic’, Hispano-Gallo-Brittonica: Essays in honour of D. E. Evans on his sixty-fifth birthday, eds J. F. Eska, R. G. Gruffydd, & N. Jacobs, 79–95. Cardiff & Dublin, University of Wales Press. Koch, J. T. 1991 ‘Ériu, Alba, and Letha: When was a Language Ancestral to Gaelic First Spoken in Ireland?’ Emania 9, 17–27. Koch, J. T. 2011 Tartessian 2: The Inscription of Mesas do Castelinho, ro and the Verbal Complex, Preliminaries to Historical Phonology. Aberystwyth, University of Wales Centre for Advanced Welsh and Celtic Studies. Koch, J. T. 2013a Tartessian: Celtic in the South-west at the Dawn of History, Celtic Studies Publications 13, Aberystwyth. Revised & expanded edition. First edition published 2009. Koch, J. T. 2013b ‘La fórmula epigráfica tartesia a la luz de los descubrimientos de la necrópolis de Medellín’, Acta Palaeohispanica XI/Palaeohispanica 13, 347–57. Koch, J. T. 2014a ‘Once again, Herodotus, the Κελτοί, the Source of the Danube, and the Pillars of Hercules’, Celtic Art in Europe: making connections. Essays in honour of Vincent Megaw on his 80th birthday, eds C. Gosden, S. Crawford, & K. Ulmschneider, 6–18. Oxford, Oxbow Books. Koch, J. T. 2014b ‘On the Debate over the Classification of the Language of the South-Western (SW)

27/05/2016 19:31:28

[ 474 ]

xv. phoenicians in the west

Inscriptions, also known as Tartessian’, Journal of Indo-European Studies 42/3–4, 1–91. Koch, J. T. 2014c ‘A Decipherment Interrupted: Proceeding from Valério, Eska, and Prósper’, Journal of Indo-European Studies 42/3–4. Koch, J. T. 2015 ‘Some Palaeohispanic Implications of the Gaulish Inscription of Rezé (Ratiatum)’, Mélanges en l’honneur de Pierre-Yves Lambert, eds G. Oudaer, G. Hily, & H. Le Bihan, 333–46. Rennes, Université de Rennes 2. Koch, J. T., with R. Karl, A. Minard, & S. Ó Faoláin 2007 An Atlas for Celtic Studies: Archaeology and Names in Ancient Europe and Early Medieval Ireland, Britain, and Brittany. Celtic Studies Publications 12. Oxford, Oxbow Books. Lambert, P.-Y. & D. Stifter 2012 ‘Le plomb gaulois de Rezé’, Études Celtiques 38, 141–64. Lejeune, M. 1972 ‘Celtibère et lépontique’, Homenaje a Antonio Tovar ofrecidos por sus discipulos, colegas y amigos, 265–71. Madrid, Gredos. López-Ruiz, C. 2009 ‘Tarshish and Tartessos Revisited: Textual Problems and Historical Implications’, Colonial Encounters in Ancient Iberia: Phoenician, Greek, and Indigenous Relations, eds M. Dietler & C. LópezRuiz, 255–80. Chicago, University of Chicago Press. Lorrio, A. J., & G. Ruiz Zapatero 2005b ‘The Celts in Iberia: An Overview’, e-Keltoi 6, 167–254. Luján Martínez, E. R. 2007 ‘L’onomastique des Vettons: analyse linguistique’, Gaulois et celtique continental, eds P.-Y. Lambert & G.-J. Pinault, 245–75. Genève, Librairie Droz. McCone, K. R. 1991 ‘The PIE Stops and Syllabic Nasals in Celtic’, Studia Celtica Japonica NS 4, 37–69. McCone, K. R. 1994 ‘An tSean-Ghaeilge agus a Réamh­ stair’, Stair na Gaeilge in ómós do Pádraig Ó Fiannachta, eds K. McCone, D. McManus, Cathal Ó Háinle, N. Williams, & L. Breatnach, 61–219. Roinn na SeanGhaeilge, Coláiste Phádraig, Maigh Nuad. McCone, K. R. 1996 Towards a Relative Chronology of Ancient and Medieval Celtic Sound Change. Maynooth Studies in Celtic Linguistics I. Maynooth, Depart­ ment of Old and Middle Irish, St Patrick’s College. McCone, K. R. 2006 The Origins and Development of the Insular Celtic Verbal Complex, Maynooth Studies in Celtic Linguistics 6. Maynooth, The Department of Old Irish, National University of Ireland. McCone, K. R. 2008 The Celtic Question: Modern Constructs and Ancient Realities, Myles Dillon Memorial Lecture. Dublin, Dublin Institute for Advanced Studies. McManus, D. 1991 A Guide to Ogam. Maynooth Monographs 4. Maynooth, An Sagart.

Koch CW3 16 iii 16.indd 474

]]

Maier, B. 2012 Geschichte und Kultur der Kelten. München, Verlag C. H. Beck. Mallory, J. P., & D. Q. Adams 2006 The Oxford Introduction to Proto-Indo-European and the Proto-IndoEuropean World. Oxford, Oxford University Press. Matasović, R. 2009 Etymological Dictionary of ProtoCeltic, Leiden Indo-European Etymological Dictionary Series 9. Leiden/Boston, Brill. Maurin, L., J.-P. Bost, & J.-M. Roddaz (dir.) 1992 Les racines de l’Aquitaine. Bordeaux, Centre Charles Higounet-Centre Pierre Paris, Université de Bordeaux III. Mederos Martín, A. 1996 ‘La conexión levantinochipriota. Indicios de comercio atlántico con el Mediterráneo oriental durante el Bronce Final (1150– 950 AC)’, Trabajos de Prehistoria 53/2, 95–115. Mederos Martín, A. 2006 Fenicios en Huelva, en el siglo X a.C., durante el reinado de Hîrãm I de Tiro. Spal 15, 167–188. Michelena, L. 1977 Fonética histórica vasca, 2nd edn., Donostia–San Sebastián, Diputación de Guipúzcoa. (first published 1961) Milcent, P.-Y. 2009. ‘Le passage de l’âge du Bronze à l’âge du Fer en Gaule au miroir des élites sociales : une crise au VIIIe siècle av. J.‑C. ?’, De l’Âge du Bronze à l’Âge du Fer en France et en Europe occidentale (XeVIIe siècle av. J.-C.); la moyenne vallée du Rhône aux âges du Fer, Actes du XXXème colloque international de l’AFEAF, co-organisé avec l’APRAB, Saint-Romain-enGal, 26–28 mai 2006, Revue Archéologique de de l’Est, supplément 27, eds. M. J. Roulière-Lambert, A. Daubigney, P. -Y. Milcent, M. Talon Marc, & J. Vital Joël, 453–76. Milcent, P.-Y. 2012 Le temps des élites en Gaule atlantique: chronologie des mobiliers et rythmes de constitution des dépots métalliques dans le contexte européen (XIIIe–VIIe av. J.C.). Rennes, Presses Universitaires de Rennes. Nijboer, A. J. & J. van der Plicht 2006 ‘An interpretation of the Radiocarbon Determinations of the Oldest Indigenous-Phoenician Stratum, thus far, Excavated at Huelva, Tartessos (Southwestern Spain)’, BABesch (Bulletin Antieke Beschavin, Annual Papers on Mediterranean Archaeology) 81, 31–36. Nijboer, A. J. & J. van der Plicht 2008 ‘The Iron Age in the Mediterranean: Recent Radiocarbon Research at the University of Groningen’ A New Dawn for the Dark Age? Shifting Paradigms in Mediterranean Iron Age Chronology. Proceedings of the XV Congress of the UISPP Colloquium C53, Lisboa 4–9 September 2006, BAR International Series 1871, eds D. Brandherm & M. Trachselm, 103–108. Oxford, Archaeopress. O’Connor, B. 2007 ‘Llyn Fawr Metalwork in Britain:

27/05/2016 19:31:28

Koch

A Review’, The Early Iron Age in Britain and the Near Continent, eds C. Haselgrove & R. Pope, 64–79. Oxford, Oxbow Books. Olivier, J.-P. 2013 ‘The Development of the Cypriot Syllabries, from Enkomi to Kafizin’, Syllabic Writing on Cyprus and its Context, ed. P. M. Steele, 7–26. Cambridge University Press. Palol, P. de & J. Vilella 1987 Clunia II. La epigrafía de Clunia [EAE 150]. Madrid. Parker Pearson, M. 2005, ‘Warfare, violence and slavery in later prehistory: an introduction’, Warfare, Violence and Slavery in Prehistory, BAR International Series 1374, eds M. Parker Pearson & I. J. N. Thorpe, 19–33. Oxford, Archaeopress. Prósper, B. M.a 2008 ‘Lusitanian: A Non-Celtic IndoEuropean Language of Western Hispania’, Celtic and Other Languages in Ancient Europe, ed. J. L. García Alonso, 53–64. Salamanca, Ediciones Universidad de Salamanca. Prósper, B. M.a 2010 ‘La lengua lusitana en el marco de las lenguas indoeuropeas occidentales y su relación con las lengas itálicas’, El Mediterráneo antiguo. Lenguas y escrituras, eds Carrasco Serrano, G. & J. C. Oliva Mompeán, 361–92. Cuenca, Ediciones de la Universidad de Castilla-La Mancha / Madrid, Centro de Estudios del Próximo Oriente. Prósper, B. M.a 2014 ‘Some observations on the classification of Tartessian as a Celtic language’, Journal of Indo-European Studies 42, 468–486. Prósper, B. M.a 2015 ‘The Indo–European ordinal numerals “fourth” and “fifth” and the reconstruction of the Celtic and Italic numeral systems’, Die Sprach 51, 1–50. Raybould, M. E. & P. Sims-Williams 2007 The Geography of Celtic Personal Names in the Latin Inscriptions of the Roman Empire. Aberystwyth, CMCS. Rh½s, J. 1891–94 ‘The Celts and the other Aryans of the P and Q groups’, Transactions of the Philological Society, 104–31. Ruiz-Gálvez Priego, M. (ed.) 1995a Ritos de paso y puntos de paso. La Ría de Huelva en el mundo del Bronce Final europeo. Madrid, Editorial Complutense. Ruiz-Gálvez Priego, M. 1995b ‘Cronología de la Ría de Huelva en el marco del Bronce Final de Europa occidental’, Ritos de paso y puntos de paso. La Ría de Huelva en el mundo del Bronce Final europeo, ed. M. Ruiz-Gálvez, 79–83. Madrid, Editorial Complutense. Ruiz-Gálvez Priego, M. 2000 ‘Weight Systems and Exchange Networks in Bronze Age Europe’, Metals Make the World Go Round: Supply and Circulation of Metals in Bronze Age Europe, ed. C. F. E. Pare, 267–79. Oxford, Oxbow Books.

Koch CW3 16 iii 16.indd 475

[ 475 ]

Ruiz-Gálvez Priego, M. 2008 ‘Writing, Counting, SelfAwareness, Experiencing Distant Worlds. Identity Processes and Free-lance Trade in the Bronze Age/Iron Age Transition’, Contacto cultural entre el Mediterráneo y el Atlántico (siglos XII–VII ane). La precolonización a debate, eds S. Celestino, N. Rafel, & X.-L. Armada, 27–40. Madrid, Escuela Española de Historia y Arqueología en Roma. Ruiz-Gálvez Priego, M. 2013 Con el fenicio en los talones: Los inicios de la Edad del hierro en la cuenca del Mediterráneo. Barcelona, Bellaterra. Sánchez Moreno, E. 1996 ‘A proposito de las gentilitates: los grupos familiares del área vetona y su adecuación para la interpretación de la organización social prerromana’, Veleia 13, 115–42. Schmidt, K. H. 1986 ‘The Celtic Languages in their European Context’, Proceddings of the 7th International Congress of Celtic Studies, Oxford, 10–15 July 1983, ed. D. Ellis Evans, J. G. Griffith, & E. M. Jope, 199–221. Oxford, D. Ellis Evans. Schrijver, P. 2012 ‘The Origin of Celtic: How, When, Where?’ The Anders Ahlqvist Lecture 12/6/2012 (handout), Helsinki. Schrijver, P. 2015 ‘Pruners and Trainers of the Celtic Family Tree: the Rise and Development of Celtic in the Light of Language Contact’, Proc. XIV International Congress of Celtic Studies, Maynooth 2011, eds L. Breatnach, R. Ó hUiginn, D. McMnaus, & K. Simms, 191–219. Dublin, Dublin Institute for Advanced Studies. Sherratt, S. 2003 ‘Visible Writing: Questions of Script and Identity in Early Iron Age Greece and Cyprus’, Oxford Journal of Archaeology 23(3), 225–42. Torres Ortiz, M. 2008 ‘The Chronology of the Late Bronze Age in Western Iberia and the Beginning of the Phoenician Colonization in the Western Mediterranean’, A New Dawn for the Dark Age? ­Shifting Paradigms in Mediterranean Iron Age Chronology / L’âge obscur se fait-il jour de nouveau? Les paradigmes changeants de la chronologie de l’âge du Fer en Mediterranée, British Archaeological Reports International Series, vol. 1871, eds D. Brandherm & M. Trachsel, 135–47. Oxford, Archaeopress. Trask, R. L. 1997 The History of Basque. London, Routledge. Untermann, J. 1985–86 ‘Lusitanisch, Keltiberisch, Keltisch’, Studia Palaeohispanica. Actas del IV Coloquio sobre lenguas y culturas paleohispánicas, Vitoria/Gasteiz 1985 = Veleia 2–3, 57–76. Untermann, J. 1995 ‘Zum Stand der Deutung der “tartessischen” Inschriften’, Hispano-Gallo-Brittonica: Essays in Honour of Professor D. Ellis Evans on the

27/05/2016 19:31:28

[ 476 ]

xv. phoenicians in the west

Occasion of his Sixty-Fifth Birthday, eds J. F. Eska, R. G. Gruffydd, & N. Jacobs, 244–59. Cardiff, University of Wales Press. Valério, M. 2014 ‘The interpretative limits of the southwestern script’, Journal of Indo-European Studies 42, 439–67. Vallejo Ruiz, J. M.a 2005 Antroponimia indígena de la Lusitania romana. Vitoria/Gasteiz, Anejos de Veleia, Series Minor 23. Vallejo Ruiz, J. M.a 2013 ‘Hacia una definición del Lusitano’, Acta Palaeohispanica XI/Palaeohispanica 13, 273–91. Vennemann, Th. 2001 ‘Atlantis Semitica: Structural contact features in Celtic and English’, Historical Linguistics 1999: Selected Papers from the 14th International Conference on Historical Linguistics, ed. L. Brinton, 351–69. Amsterdam, John Benjamins. Vennemann, Th. 2002 ‘Semitic > Celtic > English: The transitivity of language contact’, The Celtic Roots of English, eds M. Filppula, J. Klemola, & H. Pitkänen, 295–330. Joensuu, University of Joensuu. Villar, F. 1994 ‘Los antropónimos en Pent-, Pint- y las lenguas indoeuropeas prerromanas de la Península Ibérica’, Indo-Germanica et Caucasica: Festschrift für Karl Horst Schmidt zum 65. Geburstag, eds R. Bilemeier, R. Stempel, & R. Lanszweert, 234–64.

Koch CW3 16 iii 16.indd 476

]]

Berlin, De Gruyter. Villar, F. 2004 ‘The Celtic Language of the Iberian Peninsula’, Studies in Baltic and Indo-European Linguistics in Honor of William R. Schmalstieg, eds P. Baldi & P. U. Dini, 243–74. Amsterdam, John Benjamins. Villar, F., B. M.a Prósper, C. Jordán, & M.a P. Fernández Álvarez 2011 Lenguas, genes y culturas en la prehistoria de Europa y Asia suroccidental. Salamanca, Ediciones Universidad de Salamanca Watkins, C. W. 1999 ‘Two Celtic Notes’, Studia Celtica et Indogermanica: Festschrift für Wolfgang Meid, ed. P. Anreiter & E. Jerem, 539–43. Budapest, Archeolingua. Wikander, S. 1966 ‘Sur la langue des inscriptions sudhispaniques’, Studia linguistica 20, 1–8. Wodtko, D. S. 2009 ‘Language Contact in Lusitania’, International Journal of Diachronic Linguistics and Linguistic Reconstruction 6, 1–48. Wodtko, D. S. 2010 ‘The Problem of Lusitanian’, Celtic from the West. Alternative Perspectives from Archaeology, Genetics, Language, and Literature, eds B. Cunliffe & J. T. Koch, 335–67. Oxford, Oxbow Books. Wodtko, D. S. 2013 ‘Models of Language Spread and Language Development in Prehistoric Europe’, Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, eds J. T. Koch & B. Cunliffe, 185–206. Oxford, Oxbow Books.

27/05/2016 19:31:28

chapter sixteen

Celtic ‘Dogs’ in the Iberian Peninsula Fernando Fernández Palacios

1. Personal and family names

T

identification of the personal name Quno from Forua (Busturialdea, Bizkaia, in the Spanish Basque Country) as containing Celtic *kun- ‘dog’1 led me to a revision of the examples from the Iberian Peninsula which can be ascribed to the same root. In the Iberian Peninsula there are several probable instances of *kun- ‘dog’: the personal name Conia (CIL II 2589, from Lugo, Galicia),2 Coni (gen.) (HEp 6: 833, from Hinojosa de Duero, Salamanca), the personal name Suconia (CIL II 1267, from Villalba del Alcor, Huelva), Talaconius (AE 1999: 868, from Sortelha),3 the family name konikum (MLH IV, K.1.3, II-49, III-26), and the family name kunikum (MLH IV, K.1.3, II-27).4 Another example is Conicodius (CIL II, 6330, from Terena, Alandroal, Évora), usually interpreted as a sort of compound name.5 Be that as it may, the possible etymological relationship to the group name Conii in the south-west of the Peninsula has been pointed out for this example.6 It is possible that a form derived from the same root also occurs in Conia which appears in the longer sequence Coniagellietar (rock inscription from La Senia, Almatret, Lérida).7 It is possible that in this last example Conia- is part of an Ibericized form, like uiŕe adapted from Celtic *uiros on a bronze from Monteró (Camarasa, la Noguera, Lérida; Ferrer et al. 2009, 127a). 1 2 3 4

5 6 7

he

Matasović 2009 sn. *kwon- ‘dog’; see also Fernández Palacios (forthcoming), where I give more details about the Common Celtic form and its derivatives. D(is) M(anibus) s(acrum) / Coniae / Puscin(a)e / an(n)orum / LII. Explained by Delamarre (2010–12, 119) as ‘un “guerrier de soutien”, un “auxilliaire” (*Tal\-con-io-s, avec *c~/con- “chien” > “loup” > “guerrier”)’. Untermann compared konikum with the personal name Conia which was related to the family name Conium, but surprisingly he did not venture any etymology (Untermann 1996, 143b, 144b). Moreover, Conium (CIL II 5866, from Ávila) is not a family name, but stands in the inscription for origo Conium[brig(ensis)]. Cf. Codius (CIL VI 15946 and Schulze, 1991, 557) for the second part and Conisovinus and Coniletus for the first one (Moralejo 2007, 191). Encarnação (1984, 574 (no. 493)), with references. Pita Mercé (1958, 232–3); IRC II 39–40 (no. 12); Siles (1985, 172 (no. 684)); Mayer (2005, 268).

Fernandez CW3 16 iii 16.indd 477

27/05/2016 19:49:44

[ 478 ]

xvi. celtic ‘dogs’ in the iberian peninsula

]]

  All but the family name kunikum and the personal name Quno from Forua (Bizkaia) show -u- > -o-, which can be explained in any one of several possible ways. The forms spelled with o could reflect a simplified diphthong ou (i.e., }), or the old o of the accusative, as explained by Lionel Joseph (1990), or a lowering u to o as in Primitive Irish8 and sometimes in Gaulish before nasals,9 or even the phonetic convergence of Celtic ŏ and ŭ in this environment in varieties of Hispano-Celtic.   Conia, Suconia, konikum and kunikum are easily explained: the first two and the last examples would represent a feminine theme in -ia, with the prefix Su- ‘good’ in the case of Suconia. konikum and kunikum are adjectives in -iko- in genitive plural with an alternating vowel o/u in the first syllable10 (see Tables 16.1–2, Figure 16.1).

CONI (gen.) CONIA CONIA CONICODIUS SUCONIA TALACONIUS QVNO

kon-/ kun- PERSONAL NAMES Hinojosa de Duero (Salamanca) Lugo Almatret (Lérida) Terena, Alandroal, Évora Villalba del Alcor (Huelva) Sortelha Forua (Bizkaia)

HEp, 6, 833 CIL II, 2589 IRC II, 12 CIL II, 6330 CIL II, 1267 AE 1999, 868 Fernández Palacios 2004, 483

Table 16.1 Palaeohispanic personal names probably derived from Celtic ‘dog’

koni- / kuni- FAMILY NAMES

konikum

Botorrita (Zaragoza)

MLH IV, K.1.3, II–49, III–26

kunikum

Botorrita (Zaragoza)

MLH IV, K.1.3, II–27

Table 16.2 Celtiberian genitive plural family names, probably derived from Celtic ‘dog’

  McCone proposed to interpret the personal name uiroku from the inscription of Botorrita III (I–5, I–51, III–26) as a compound of uiro and k~ with the meaning ‘werewolf ’ (cf. Old Irish Ferchú, Old Welsh Gurci, modern Gwrgi, Old Breton Gurki, modern Gourgi).11 8

In Old Irish one of the most typical alternations in open stressed syllables is that of [u – o], cf. Jaskuła (2006, 174–81). 9 Cf., for example, the variation in the composition vowels in Gaulish place-names Virdumarus/Virdomarus, Ratumagus/Ratomagus (Dottin 1920, 59). 10 kunikum is the only example along with Quno from Forua (Bizkaia) who shows -u- instead of -o- or -ou-. 11 McCone (1987) and McCone (2005, 483–4). It is given as a fact in Lucht (2007, 113) and in Hyllested (2010, 111). It is worth mentioning, although it is possibly no more than a coincidence, that in III–26 it occurs alongside uiroku: uiroku : konikum : statulos.

Fernandez CW3 16 iii 16.indd 478

27/05/2016 19:49:44

[ 479 ]

Fernández

%

# CONIA

QV N O

kunikum %

# CO N I A

# CONI # TALACONIVS

# CONICODIVS

# SVCONIA

{ 300 km

16.1 kon- # personal names / kun- ∎ personal and family names. From north to south and west to east: Conia (Lugo), Coni (gen.) (Hinojosa de Duero, Salamanca), Talaconius (Sortelha), Conicodius (Terena, Alandroal, Évora), Suconia (Villalba del Alcor, Huelva), Quno (Forua, Bizkaia), kunikum (Botorrita) and Conia (Almatret, Lérida)

Wodtko (1999, 738–9) responded that she doubted the existence of an inherited compound because the attested forms cannot be derived regularly from Common Celtic *Wirok~. McCone replied, proposing that it could have been very easy to create a new nominative Ferchú by analogy.12 In the light of the issue raised by Wodtko (1999), it seems likely that uiroku could be merely another example, this time of *uiros (Old Irish fer ‘man’) or *u īro(Old Irish fír ‘true’) (Sims-Williams 2013, 46), of the well-attested and productive Peninsular derivative in -oko-.13   I think that it is very likely that we also have the *kun- in Couneancus (a personal name 12 McCone (2001, 484). Wodtko (2003, 8) remained sceptical. John T. Koch (pers. comm.) thinks that McCone is right: so long as the etymology was understood, the compound would be renewed so as to remain clearly two words, ‘man’ and ‘dog/wolf ’, in the contemporary language. 13 For similar explanations, see Rubio Orecilla (2001, 584) and Vallejo (2005a, 575a), to name but a few.

Fernandez CW3 16 iii 16.indd 479

27/05/2016 19:49:47

[ 480 ]

xvi. celtic ‘dogs’ in the iberian peninsula

]]

from Clunia in Burgos), the attestation of which was discovered in Villa Pouca de Aguiar, CIL II 2390),14 the family names Couneidoq(um) from Segovia (CIL II 5779),15 kounesikum (from Botorrita I, MLH IV, K.1.1, B–1) and burikounikum (from Botorrita, MLH IV, K.1.3). In the four cases it is possible to isolate coun(e)-. In the first example the word ends with the known sequence -anko-, very productive in Celtiberian personal names,16 unless there has been a process of syncopation -niko- > -nko-;17 in the second it occurs as a derivative in -eid‑, as in Protaeidi f(ilius) (cf. Vallejo 2005a, 581a). The third one, kounesikum, has been the subject of a long-standing debate. An etymology *kom-ness-iko- was proposed by Eichner (1989, 35, 44), comparing the Celtiberian with an Old Irish comnesach ‘neighbour’; this proposal was well received (for example, Villar & Prósper 2005, 180), but Wodtko showed that no such Old Irish word existed (Wodtko 2000, 202–3; 2003, 6). In the same way, the parallels offered here make us suspect that we are not dealing with a defective representation of /kloun-/, as de Bernardo proposed (cf. Wodtko 2000, 203). I see no problem in deriving kounesikum from *koun- plus a derivational suffix in ‑es- (on which see Vallejo 2005a, 656–7) plus a second suffix in ‑iko-. We do not have a full-grade form anywhere in the principal paradigm for the word meaning just ‘dog’, but Old Irish cúan and Welsh cun, with the meanings ‘litter, pack of dogs’, mentioned further below, show that there was most probably a Celtic *koun\ or*kouno- derived from the same root *kun-.18   Sometimes *kouno- has been interpreted as a Celtic element, although signifying ‘beautiful’19 or ‘shining’,20 but apart from the dubious convenience of such an etymology, I see no problem in the derivation from the root *kun-. The spellings with a diphthong could represent either the old diphthong preserved as ou in forms derived from the e-grade or o-grade of the root or orthographic hesitation in representing a round open sound [u] where original short u and short o had become indistinct before nasals. It is attested as Old Irish cúan and Welsh cun, both meaning ‘litter, pack of dogs’, and this shows that Celtic must have had a form *koun- which was used to describe things connected with dogs (the older, Indo-European form, could have been something like *k´(w)eun-, as the old diphthongs eu and ou fell together as ou in Common Celtic). For the attested forms and their geographical distribution, see Table 16.3 and Figure 16.2. 14 C(ornelius?) Coune/ancus / Fusci [f(ilius)] Clu(niensis) / [a]n(norum) XL / LA CIV / [- - - - - -] / VS C / XXX h(ic) s(itus) e(st), no. 8207 in Hispania Epigraphica (http://eda-bea.es/pub/record_card_2. php?refpage=%2Fpub%2Fsearch_select.php&quicksearch=vila+pouca+de+aguiar&rec=8207, consulted 19th March 2014). 15 No. 12085 of Hispania Epigraphica. Cf. Albertos (1979, 141). It was found by Fidel Fita embedded in the Alcázar: Allae / Couneidoq(um) . Cipoll / us uxori / s . t . t . l. 16 See now Jordán (2013), where Couneancus is not mentioned. For toponyms in ‑anco- see, for example, Moralejo Laso (1979, 14); these are also present on the northern coast of Spain. 17 In fact Vallejo (2010, 638) explains Couneancus as a name in -icus, and it is not impossible that all the Celtiberian forms have -nko- from earlier -niko-. 18 Matasović (2009 sn. *kown\) is uncertain about this, but the semantic connection with ‘dog’ is so strong in both Insular Celtic subfamilies, that the conclusion is hard to avoid. 19 Thus Carnoy (1907, 22). 20 Vallejo (2005b, 116).

Fernandez CW3 16 iii 16.indd 480

27/05/2016 19:49:47

[ 481 ]

Fernández

COVNEA N C VS

# #

COVNEDIOQ

burikounikum kounesikum

#

{ 300 km

16.2 koun- personal and family names. From west to east: Counedioq(um) (Segovia), Couneancus (Cluniensis), and burikounikum and kounesikum (Botorrita, Zaragoza)

COUNEANCUS COUNEDIOQ(UM) burikounikum kounesikum

koun- PERSONAL AND FAMILY NAMES Villa Pouca de Aguiar (Portugal), CIL II 2390 but Cluniensis Segovia CIL II 5779 Botorrita (Zaragoza) MLH IV, K.1.3, II–53 Botorrita (Zaragoza) MLH IV, K.1.1, B–1

Table 16.3 Names attested in the central and eastern Peninsula, probably based on the root ‘dog’, represented with ou (plotted Figure 16.2)

Fernandez CW3 16 iii 16.indd 481

27/05/2016 19:49:50

[ 482 ]

xvi. celtic ‘dogs’ in the iberian peninsula

]]

2. The geography of the personal and family names Figures 16.1 and 16.2 reveal interesting distribution patterns for the different forms attested in the personal and family names. Except for those two with kon-, all occur along an almost vertical north–south line which begins in central Galicia and ends in the south, not far away from the mouth of the Guadalquivir (Figure 16.1). The two examples of kun-, on the contrary, seem to be related to the Celtiberian world: one occurs in Botorrita and the other one on the Basque coast in Forua, Bizkaia (see Fernández Palacios in press). Lastly, the kounexamples are clearly Celtiberian or Celtiberian-related: two come from Botorrita, one from Segovia and one is attested as Cluniensis (Figure 16.2). Several interpretations may be offered to explain this distribution, but with respect to the very different distribution of kon- on the one hand and kun- and koun- on the other, it is tempting to interpret the data as reflecting a dialectal difference between the Celtiberian and Celtiberian-related zones that preserve the original vowel sounds -u-, -ou-, whereas the rest of the Peninsula innovates through the change -u- > -o-. This is only one of the several possible explanations that can be given for such distribution. The difference could be either a matter of phonology or spelling or of the various Hispanic Indo-European languages having generalized different vowel grades from the inherited ablauting paradigm. 3. Population names Celtic *kun- is probably also present in the Kun{síoi (2, 33)21 and Kun{tes (4, 49)22 of Herodotus, which seem to be names for the same people as called Koníoi by Polybius (ἐν τοῖς kονίοις, 10, 7, 5), 23 who lived in the south-western part of the Peninsula and perhaps were even mentioned by Hecateus c. 500 BC (cf. de Hoz 2010, 252–4, 310–7). The notice of Herodotus 2, 33, 3 supposedly would be ‘algo anterior a mediados del siglo V’ 24 (see de Hoz 2011, 490–1). If Herodotus’s Kun{tes has a short u, as seems probable (like for example Welsh Maelgwn), then it could be the exact cognate of the Welsh river- and place-name Cynwyd and 21

22

Ἴστρος τε γὰρ ποταμὸς ἀρξάμενος ἐκ Κελτῶν καὶ Πυρήνης πόλιος ῥέει μέσην σχίζων τὴν Εὐρώπην· οἱ δὲ Κελτοὶ εἰσὶ ἔξω Ἡρακλέων στηλέων, ὁμουρέουσι δὲ Κυνησίοισι, οἳ ἔσχατοι πρὸς δυσμέων οἰκέουσι τῶν ἐν τῇ Εὐρώπῃ κατοικημένων (for the river Ister begins from the Keltoi and the city of Pyrene and

so runs that it divides Europe in the midst (now the Keltoi are outside the Pillars of Heracles and border upon the Kynesians, who dwell furthest towards the sunset of all those who have their dwelling in Europe) (translation from Macaulay 1890). Maybe the harenae litoris Cynetici of Avienus’s Ora Maritima, 566 are referring to the same Cynetes about which Herodotus speaks (see de Hoz 2011, 125).

ῥέει γὰρ δὴ διὰ πάσης τῆς Εὐρώπης ὁ Ἴστρος, ἀρξάμενος ἐκ Κελτῶν, οἳ ἔσχατοι πρὸς ἡλίου δυσμέων μετὰ Κύνητας οἰκέουσι τῶν ἐν τῇ Εὐρώπη· ῥέων δὲ διὰ πάσης τῆς Εὐρώπης ἐς τὰ πλάγια τῆς Σκυθίης ἐσβάλλει (for the Ister flows in fact through the whole of Europe, beginning in the land of the Keltoi,

who after the Kynesians dwell furthest towards the setting sun of all the peoples of Europe; and thus flowing through all Europe it falls into the sea by the side of Scythia) (translation from Macaulay 1890). 23 Interestingly enough, the normal variant seems to be cunei, koúneoi in Greek, with -u- and ‑ou‑, although it is only found referring to the populus in Appian, Iberike 57–8, 68, which is not dependent on Polybius for this passage (de Hoz 2010, 253). Justin 44, 4, 1, according to Gascó (1985), speaks of curetes, not cunetes and, if that is retained as the preferred reading, it would have nothing to do with our populus. 24 ‘somewhat before the middle of the 5th century [BC].’

Fernandez CW3 16 iii 16.indd 482

27/05/2016 19:49:51

[ 483 ]

Fernández

Romano-British place-name Cun{tio (Rivet & Smith 1979, 328–9).   Verging on the lands of the Iberians in the east and south of the Peninsula, we find the Celtici mentioned in Roman times (Pliny 3, 3, 13–4). They lived in the western part of Baeturia, a region located between the rivers Guadiana and Guadalquivir. A continuous identity for this Celtic group could go back at least to the 4th century BC, as implied by continuity in the archaeological record, although Pliny the Elder merely speaks of Celticos a Celtiberis ex Lusitania advenisse manifestum est sacris, lingua, oppidorum vocabulis, quae cognominibus in Baetica distinguntur (Naturalis Historia 3, 13–14), which is usually taken as referring to events in the more recent past dating to the 2nd century and earliest 1st century BC.25   We also have the people from Cantabria called Konískoi (toîs konískois, Strabo 3, 4, 12) and Koniakoi (Strabo 3, 3, 8),26 which are more easily explained through our Celtic term than appealing to a Pre-Indo-European language (on which see de Hoz 2011, 547–8), and also the Concani < *Conicani mentioned by Horace (et laetum equino sanguine concanum, Carm. 3, 4, 34) and Silius Italicus (nec qui Massageten monstrans feritate parentem / cornipedis fusa satiaris, concane, vena, Pun. 3, 360–1).27 For the several population names discussed in this section, see Table 16.4. Concani

kon- / kun- POPULI Cantabria

Koniakoi Koníoi Konískoi Kunēsíoi Kunētes

Cantabria South-west of the Peninsula Cantabria South-west of the Peninsula South-west of the Peninsula

Horace, Carm. 3, 4, 34 Silius Italicus, Pun. 3, 360–1 Strabo 3, 3, 8 Polybius 10, 7, 5 Strabo 3, 4, 12 Herodotus 2, 33, 3 Herodotus 4, 49

Table 16.4 Palaeohispanic group names in Kun- and Kon- from Greek and Roman sources: this table does not present an exhaustive list of these names as attested in classical literature.

4. Place-names It is possible that Celtic *kun- is not only present in the name of populi, family and personal names in the Iberian Peninsula but also in the names of at least two ancient places. The first 25 Cf., for example, Berrocal-Rangel (1995; 2005). John Koch points out to me that the explanation given above is the standard interpretation of this passage, but that he doesn’t think it is clear when Pliny thought this happened, and it sounds like he was guessing (not unreasonably) on the basis of the language and the names of the populi. 26 Taken as probably the same people as the koniskoi by Schulten (FHA VI, 220, 247) and placed by Strabo on the sources of the Ebro river. 27 Schulten distinguished them from the koniskoi/koniakoi (FHA VI, 247). I leave aside the modern toponymic evidence, for example Cuénabres (Burón, León) < *Cŏnăbris, maybe *Conĭbr¼s (Corominas 1976, 92). In the north of Britannia Ptolemy 2, 3, 9 mentioned the Venicones, translated by Delamarre (2007, 200) as ‘loups du clan’, -cones having been generally interpreted as containing *kun-, although Watson (1993, 23) suggested comparison with Irish vennicnioi ‘descendants of Vennos’, cf. Venico in RIB no. 1543, from Carrawburgh on the Hadrian’s Wall, and Venuli (gen.) from Sablon (CIL XIII 4411).

Fernandez CW3 16 iii 16.indd 483

27/05/2016 19:49:51

[ 484 ]

xvi. celtic ‘dogs’ in the iberian peninsula

]]

one is Conimbriga, an important city from the Roman province of Lusitania identified with the remains of present-day Condeixa-a-Velha in Coimbra, at the mouth of the Mondego river. The city was built on top of an Iron Age oppidum, but there are even some objects from the site that are attributed to the Orientalizing period, at least as early as the 8th century BC (de Hoz 2010, 292–3).The place-name has been compared to that of the Conii mentioned earlier. Isaac (2004, appendix on the Antonine Itinerary) divides the place-name in Con-im-briga, taking con- and im- as two prefixes, although he contemplates the possibility of another, more reasonable, division: conim- and briga-. If Conimbriga represents Conium brig\, that could be a purely Celtic ‘hillfort of the Konioi’. Joaquín Gorrochategui (1997) thinks that it is not certain that Conim- is Celtic, and de Hoz takes it as a mixed placename, the second part evidently Celtic.28 Moralejo (2007, 190–5) has analysed extensively the problems with respect to the etymology of Conimbriga and concluded that the only possible segmentation is *Coni-mbriga, with -m(b)riga < *-mrg-eH2, from IEW 738 mereg‘Rand, Grenze’ (cf. OIr. mruig) plus epenthetic -b-.   The second place-name is Conistorgis, which played a leading role in several episodes during the Lusitanian wars of the 2nd century BC.29 A segmentation Conis- and -torgis can be proposed,30 although an alternative division of the compound as Coni- and -storgis it is not altogether impossible.31 Appian (Iberica 56–7) placed the city among the Cunei or Cynetes, and Strabo 3, 141 among the Celtici of the south-west (see Table 16.5). CONI- PLACE-NAMES Conimbriga

Condeixa-a Velha-(Portugal)

Several epigraphic and literary sources

Conistorgis

South-west of the Peninsula

Appian, Ib. 56–7, Strabo 3, 141

Table 16.5 Ancient town names of the Iberian Peninsula with Koni-

28 de Hoz (2011, 547). De Hoz argues that the first element is non-Indo-European. On some early proposals of detecting a pre-Indo-European root *cun- (*gon-) with the meaning ‘hill, height, promontory’ in several regions of the west Mediterranean (see Bertoldi 1947, 140–4). 29 FHA IV, 96–7. It was mentioned also by Sallustius, Hist. 1, 119 in the context of the war between Metellus and Sertorius (FHA IV, 175). 30 See de Hoz (2010, 471) and also FHA VI, p. 157. For the etymology, see Schmoll (1959, 32), ‘Zusammenrückung *Coni(o)s *Torgis > Conistorgis’, with -torgis meaning ‘market, forum’, from Indo-European *trg-. For more Indo-European explanations, see Moralejo (2007, 178–9, n. 137). Schulten pointed out that it was seemingly placed in the Algarve and that it was the main centre of the conii, which according to him lived in the Algarve as well as further north (in what is today central Portugal), thus making the same group the owners of Conimbriga too (FHA IV, 98). 31 See Moralejo (2007, 179, n. 137), *(s)ter-, *(s)ter-h-, *(s)tr-eh1- ‘starr, steif sein, starrer, fester Gegenstand...’, with a guttural lengthening *(s)ter-g-, *(s)tr-eg-. There is also the possibility of deriving the form from Indo-European *strig- ‘to brush, strip’, cf. Latin striga ‘furrow, swath’. For the Indo-European etymology, see de Vaan (2008, 592), and for the meaning ‘furrow’, see Lewis & Short (1879, s.v. strĭga, II).

Fernandez CW3 16 iii 16.indd 484

27/05/2016 19:49:51

Fernández

[ 485 ]

5. ‘Rabbit’ < ‘dog’? Finally, the Celtic word *k~, genitive *kunos, could have been the origin of the Spanish word conejo ‘rabbit’ (Latin cŭn¼cŭlus). As Xavier Ballester has argued,32 there is no problem in reconstructing the Celtic stem *kun- plus adjectival suffix (or even better diminutive suffix) -ik- plus a second diminutive suffix [-kl-] most probably from Latin,33 with a meaning ‘little doggie’.34 According to Ballester, in ancient times it was thought that Latin cŭn¼cŭlus came originally from a similar word used in Hispania,35 and so ‘Roman legionnaires and other travellers of the time would have spread the use of this word to other languages in Western Europe’.36 The derivation of the meaning ‘rabbit’ from ‘dog’ could have happened because ‘prehistoric man might well have seen a rabbit for the first time and described the new animal using some reference or comparison to a “dog”’.37 I agree with Ballester´s opinion but would see the origin of the word as not necessarily from Celtiberian in particular but from the wider spectrum of Hispano-Celtic, as is implied in Ballester’s second article on the subject (Ballester & Quinn 2002). 6. Conclusions The present examination of the numerous forms from the Iberian Peninsula that can be related to Celtic *kun- ‘dog’ has several interesting implications. Rather than merely identifying similar forms in the extant records, we find three distinct forms of the stem used for naming families and people: kon-, kun-, and koun-, variants which could be related to dialectal differences. Moreover, the root *kun- ‘dog’ that is behind all the HispanoCeltic examples leads to a coherent explanation for names of ancient populi of the Iberian Peninsula including Kun{tes, Koníoi, Konískoi, Koniakoi and Concani, as well as allowing us to identify the probable root of the first element of the important place-names Conimbriga and Conistorgis and the probable derivation of the Spanish word conejo.

32 Ballester (2000) (in Spanish) and Ballester & Quin (2002) (in English). 33 It can be compared to the place-names Segesamunclo (Itinerary of Antonine 394, 3), between Libia and Virovesca, versus Segisamum (Sasamón, Burgos). 34 Nevertheless, de Hoz (2010, 192–3) thinks that the ending of cunicula is non-Indo-European and favours the possibility of linking it to place-names in -(c)ula, -(c)uleia which are mainly scattered in Turdetanian areas of the Peninsula. See also de Hoz (2010, 216). With respect to the root, he does not say anything explicitly on cunicula, but seems to be thinking also in non-Indo-European terms if we judge from his opinion about Conimbriga and Conistorgis. 35 The references to the sources can be found in de Hoz (2010, 192–3). 36 Ballester & Quinn (2002, 125). It has been accepted by Simkin (2012, 88). 37 Ballester & Quinn (2002, 128). Cuniculus was also a term used in mining in the north-west of the Iberian Peninsula in Roman times (Pliny the Elder, Naturalis Historia 33, 70), but as de Hoz (2010, 169) states, its use ‘es posiblemente secundario, y en todo caso la palabra había sido asimilada en latín mucho antes de que Plinio visitase el NO’.

Fernandez CW3 16 iii 16.indd 485

27/05/2016 19:49:51

[ 486 ]

xvi. celtic ‘dogs’ in the iberian peninsula

]]

Acknowledgements I would like to thank John T. Koch (CAWCS, Aberystwyth, Wales) and Xaverio Ballester (Universidad de Valencia, Spain) for their commentaries on a draft of the present paper, and to Professor Koch also for revising and correcting my English. Thanks also are due to the Hesperia Project (Spain), of which I was a project member for many years, for permission to use its database.

BIBLIOGRAPHY AE = L’Année Épigraphique (Paris). CIL = Corpus Inscriptionum Latinarum. FHA IV = Schulten, A. (ed.) 1937 Fontes Hispaniae Antiquae. Fascículo IV. Las guerras de 154–72 a. de J. C. Barcelona, Barcinone. FHA VI = Schulten, A. (ed.) 1952 Fontes Hispaniae Antiquae. Fascículo VI. Estrabón. Geografía de Iberia. Barcelona, Barcinone. HEp = Hispania Epigraphica. Madrid. IEW = Pokorny, J. 1959 Indogermanisches etymologisches Wörterbuch. Bern, Francke. IRC = Inscriptions romaines de Catalogne, 5 vols. Paris, 1984–2002. MLH IV = Untermann, J. 1997 Monumenta Linguarum Hispanicarum. Band IV. Die tartessischen, keltiberischen und lusitanischen Inschriften. Wiesbaden, Reichert.

Berrocal-Rangel, L. 2005 ‘The Celts of the Southwestern Iberian Peninsula’, E-Keltoi. Journal of Interdisciplinary Celtic Studies, 6. The Celts in the Iberian Peninsula, 481–96. Bertoldi, V. 1947 ‘La Iberia en el substrato étnicolingüístico del Mediterráneo occidental’, Nueva Revista de Filología Hispánica, 1, 128–47. Carnoy, A. 1907 ‘Éléments celtiques dans les noms de personnes des inscriptions d´Espagne’, Le Muséon, 8, 1–40. Corominas, J. 1976 ‘Elementos prelatinos en las lenguas romances hispánicas’, Actas del I Coloquio sobre lenguas y culturas prerromanas de la Península Ibérica (Salamanca, 1974) eds F. Jordá, J. de Hoz, & L. Michelena, 87–164. Salamanca, Ediciones Universidad de Salamanca. Delamarre, X. 2007 Noms de personnes celtiques dans l´épigraphie classique. Paris, Errance. Delamarre, X. 2010–2012 ‘Notes d’onomastique Albertos Firmat, M. L. 1979 ‘La onomástica de la vieille-celtique’, Keltische Forschungen, 5, 99–137. Celtiberia’, Actas del II Coloquio sobre lenguas y culturas Dottin, G. 1920 La langue gauloise. Paris, Kliencksieck. prerromanas de la Península Ibérica (Tübingen, 1976), eds Eichner, H. 1989 ‘Damals und heute: Probleme der Tovar, A., M. Faust, F. Fischer & M. Koch, 131–67. Erschließung des Altkeltischen zu Zeußens Zeit und in der Gegenwart’, Erlanger Gedenkfeier für Johann Salamanca, Universidad de Salamanca. Kaspar Zeuß. ed. B. Forssman, 9–56. Forssman, B. Ballester, X. 2000 ‘Conejo, Étimo Celtibérico (y (ed.) 1989 Erlanger Gedenkfeier für Johann Kaspar Zeuß. Can, Étimo Transcontinental)’, Boletín de la Sociedad Nürnberg, Erlangen. Castellonense de Cultura, 75, 455–66. Ballester, X., & R. Quinn 2002 ‘Cuniculus “rabbit”— A d’Encarnação, J. 1984 Inscrições romanas do Conventus Pacensis: subsídios para o estudo da romanização. Celtic Etymology’, World Rabbit Science 10/3, 125–9. Coimbra, Universidade de Coimbra. Beltrán, F., J. de Hoz, & J. Untermann 1996, Aragon. Fernández Palacios, F. forthcoming ‘Quno, a Celtic 1996 El tercer bronce de Botorrita (Contrebia Belaisca), personal name in a Roman inscription from Forua Colección Arquelogía 19. Zaragoza, Aragon. (Busturialdea, Bizkaia, Basque Country) and Celtic Berrocal-Rangel, L. 1995 ‘Etnogénesis y territorio: personal names in Bizkaia (Spain)’. jefaturas, estatalización y moneda entre los pueblos Ferrer, J., I. Garcés, J. R. González, J. Principal, & J. betúricos’, in García-Bellido and Centeno (eds) I. Rodríguez et al. 2009 ‘Els materials arqueològics (1995), 117–28.

Fernandez CW3 16 iii 16.indd 486

27/05/2016 19:49:51

Fernández

I epigràfics de Monteró (Camarasa, la Noguera, Lleida). Troballes anteriors a les excavacions de l´any 2002’, Cuadernos de Prehistoria y Arqueología Castellonenses 27, 109–54. García-Bellido, M. P., & C. Centeno (eds) 1995 La Moneda Hispánica. Ciudad y Territorio. Madrid, Centro de Estudios Históricos. Gascó, F. 1985 ‘¿Curetes o cunetes? Justino XLIV, 4, 1’, Gerión 5, 183–94. Gorrochategui, J. 1997 ‘Gallaecia e as linguas prerromanas da Península Ibérica’, O feito diferencial galego na Historia. Vol. 1. Galicia fai dous mil anos, ed. G. Pereira Menaut, 15–49. Santiago de Compostela, Museo do Pobo Galego. de Hoz, J. 2010 Historia lingüística de la Península Ibérica en la Antigüedad. I. Preliminares y mundo meridional prerromano. Madrid, Consejo Superior de Investigaciones Científicas. de Hoz, J. 2011 Historia lingüística de la Península Ibérica en la Antigüedad. II. El mundo ibérico prerromano y la indoeuropeización. Madrid, Consejo Superior de Investigaciones Científicas. Hyllested, A. 2010 ‘The Precursors of Celtic and Germanic’, Proceedings of the 21st Annual UCLA Indo-European Conference, eds S. W. Jamison, H. C. Melchert, & B. Vine, 107–28. Bremen, Hempen. Isaac, G. R. 2004 Place-Names in Ptolemy’s Geography. Aberystwyth, Department of Welsh (CD). Jaskuła, K. 2006 Ancient Sound Changes and Old Irish Phonology. Lublin, Wydawnictwo KUL. Jordá, F. et al. (eds.) 1976. Actas del I Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica (Salamanca, 27–31 de mayo de 1974). Salamanca, Ediciones Universidad de Salamanca. Jordán, C. 2013 ‘The -anko- Sequence in Celtiberian Anthroponymy’, Continental Celtic Word Formation. The Onomastic Data, ed. J. L. García Alonso, 165–73. Salamanca, Ediciones Universidad de Salamanca. Joseph, L. S. 1990 ‘Old Irish cú: A Naïve Reinter­pret­ ation’, Celtic Language, Celtic Culture: A Festschrift for Eric P. Hamp, eds A. T. E. Matonis & D. F. Melia, 110–30. Van Nuys Calif., Ford & Bailie. Lewis, C. T., & C. Short 1879 A Latin Dictionary. Oxford, Oxford University Press. Lucht, M. 2007 Der Grundwortschatz des Altirischen. Bonn, Inaugural-Dissertation zur Erlangung der Doktorwürde der Philosophischen Fakultät der Rheinischen Friedrich-Wilhelms-Universität zu Bonn. Macaulay, G. C. (transl.) 1890 The History of Herodotus.

Fernandez CW3 16 iii 16.indd 487

[ 487 ]

London-New York. McCone, K. 1987 ‘Hund, Wolf und Krieger bei den Indogermanen’, Studien zum indogermanischen Wortschatz, ed. W. Meid, 101–54. Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. McCone, K. 2001 ‘Celtibérico, celta continental y celta común’, Religión, lengua y cultura prerromanas de Hispania (VIII Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, Salamanca, 1999), eds F. Villar & M. P. Fernández Álvarez, 483–94. Salamanca, Universidad de Salamanca. McCone, K. 2005 ‘Mögliche nicht-indogermanische Elemente in der keltischen Sprachen’, Sprachkontakt und Sprachwandel: Akten der XI. Fachtagung der Indogermanischen Gesellschaft, 17.–23. September 2000, Halle an der Saale, eds. G. Meiser, G. & O. Hackstein, 395–435. Wiesbaden, Reichert. Matasović, R. 2009 Etymological Dictionary of ProtoCeltic, Leiden Indo-European Etymological Dictionary Series 9. Leiden/Boston, Brill. Mayer, M. 2005 ‘La onomástica indígena en la zona norte del Conventus Tarraconensis’, Acta Palaeohispanica IX/Palaeohispanica 5, 259–72. Moralejo, J. J. 2007 Callaica Nomina. Estudios de Onomástica Gallega. A Coruña, Biblioteca Filolóxica Galega. Moralejo Laso, A. 1979 ‘Ojeada a los topónimos hispánicos y especialmente a los gallegos de origen prelatino de J. Corominas (Continuación)’, Verba 6, 13–16. Pita Mercé, R. 1958 ‘Localizaciones arqueológicas en el Bajo Cinca’, Argensola 35, 229–48. Prósper, B. 1997 ‘El nombre de la diosa lusitana Nabia y el problema del betacismo en las lenguas indígenas del Occidente Peninsular’, ‘Ilu 2, 141–9. Rivet, A. L. F., & C. Smith 1979 The Place-Names of Roman Britain. London, Batsford. Rubio Orecilla, F. J. 2001 ‘Las formas secundarias en -ko- del celtibérico’, Religión, lengua y cultura prerromanas de Hispania (VIII Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, Salamanca, 1999), eds F. Villar & M. P. Fernández Álvarez, 581–94. Salamanca, Universidad de Salamanca. Schmoll, U. 1959 Die Sprachen der vorkeltischen Indogermanen Hispaniens und das Keltiberische. Wiesbaden, Reichert. Schulze, W. 1991 Zur Geschichte lateinischer Eigennamen. Zurich-Hildesheim, Weidmann. Siles, J. 1985 Léxico de inscripciones ibéricas. Madrid, Ministerio de Cultura.

27/05/2016 19:49:51

[ 488 ]

xvi. celtic ‘dogs’ in the iberian peninsula

Simkin, O. 2012 ‘Language contact in the pre-Roman and Roman Iberian peninsula. Direct and indirect evidence’, Multilingualism in the Graeco-Roman Worlds, eds A. Mullen & P. James, 77–105. Cambridge, Cambridge University Press. Sims-Williams, P. 2013 ‘The Celtic Composition Vowels -o- and -io-’, Continental Celtic Word Formation. The Onomastic Data, ed. J. L. García Alonso, 37–50. Salamanca, Ediciones Universidad de Salamanca. Untermann, J. 1996 ‘Onomástica’, El tercer bronce de Botorrita (Contrebia Belaisca), Colección Arquelogía 19, eds F. Beltrán, J. de Hoz, J. Untermann, 109–66. Zaragoza, Aragon. de Vaan, M. 2008 Etymological Dictionary of Latin and the other Italic Languages. Leiden-Boston, Brill. Vallejo, J. M. 2005a Antroponimia indígena de la Lusitania romana. Vitoria/Gasteiz, Anejos de Veleia. Vallejo, J. M. 2005b ‘La composición en la antroponimia antigua de la Península Ibérica’, Acta Palaeohispanica IX/Palaeohispanica 5, 99–134. Vallejo, J. M. 2010 ‘Los celtas y la onomástica. El caso

Fernandez CW3 16 iii 16.indd 488

]]

hispano’, Serta Palaeohispanica J. de Hoz. Palaeohispanica 10, 629–47. Villar, F. & F. Beltrán (eds) 1999 Pueblos, Lenguas y Escrituras en la Hispania Prerromana. Salamanca Ediciones Universidad de Salamanca. Villar, F. & B. M.a Prósper 2005 Vascos, celtas e indoeuropeos. Genes y lenguas. Salamanca, Ediciones Universidad de Salamanca. Watson, W. J. 1993 The History of the Celtic Place-Names of Scotland. Edinburgh, Birlinn. Wodtko, D. S. 1999 ‘Remarks on Celtiberian Etymology’, Pueblos, Lenguas y Escrituras en la Hispania Prerromana, eds F. Villar & F. Beltrán, 733–44. Salamanca Ediciones Universidad de Salamanca. Wodtko, D. S. 2000 Monumenta Linguarum Hispanicarum. Band V.1. Wörterbuch der keltiberischen Inschriften. Wiesbaden, Reichert. Wodtko, D. S. 2003 An outline of Celtiberian grammar. Freiburg. http://www.freidok.uni-freiburg.de/ volltexte/747/

27/05/2016 19:49:51

chapter seventeen — ancillary study

Sound change, the Italo-Celtic linguistic unity, and the Italian homeland of Celtic Peter Schrijver

1. Introduction

F

a notion that has been dealt various death blows over the 150 year of its existence, the idea that there was a common Italo-Celtic node on the Indo-European family tree has proved to be remarkably like the Hydra of Lerna: pruning its head has led to increased vigour. First conceived in 1861 by Carl Lottner and flourishing for almost seventy years, its first burial was at the hands of Carl Marstrander (1929) and Giacomo Devoto (1929). Later burials followed, most influentially by Calvert Watkins (1966) and most recently by Graham Isaac (2007, 75–95). The most persuasive evidence in favour of a common Italo-Celtic stage has always been morphological: e.g. Cowgill (1970) on superlatives, Jasanoff (1994) on the \-subjunctive, Jasanoff (1997) on the medio-passive, Schrijver (2006) and Kortlandt (2007) on the verbal system, Schrijver (1997, 14–16) on generalization of *s- in the PIE *so/to- demonstrative pronoun, and Schrijver (2007) on the expansion of the o-stem ablative in *-d to all nominal stem classes in all of Celtic. It seems fair to say that those who nowadays support the existence of an Italo-Celtic unity do so largely on the basis of a few morphological innovations (in the vein of Cowgill 1970) and stress the ephemeral nature of the phenomenon. Early sound changes hardly ever make it onto the Italo-Celtic list (e.g. Ringe 1988; Zair 2012, 268–71).   This article specifically addresses sound change, providing a discussion of phonological innovations that are shared by Italic and Celtic (section 2).1 In the light of this evidence for an Italo-Celtic unity, similarities in the early development of obstruents (section 3) and of long vowels (section 4) are discussed, and implications for Italo-Celtic are discussed (section 5). Finally, the question where the Celtic language family originated is addressed (section 6).

1

Schrijver CW3.indd 489

or

The list is based on the lists in Schrijver (2006) and Schrijver (2015, 196–7) but contains more discussion, most of which is in reaction to Zair (2012), which appeared too late to be taken into full account in Schrijver (ibid.), which was written in 2011 and 2012.

27/05/2016 19:52:58

[ 490 ]

xvii. italo-celtic linguistic unit y

]]

2. Italo-Celtic sound changes Two related languages can be assumed to share a common prehistoric node on the family tree if they underwent a series of identical innovations that other related languages did not, and if they underwent those changes in an identical relative order. Positing that those changes occurred only once, at a stage when the two languages were still one, is preferable because this requires fewer assumptions than positing that the changes occurred twice, independently within each language. If both conditions are fulfilled (identical innovation and identical relative order), the risk of promoting chance resemblances to the status of common innovations is reduced.   This procedure is applicable to all related languages, irrespective of whether they are closely or distantly related. In the case of Italic and Celtic, the relationship is distant, both in time (approximately the first half or middle of the 2nd millennium BC would be a reasonably conservative guess) and in terms of the amount of intervening changes: between the attested Italic languages, such as Latin and Umbrian, and their Celtic counterparts, such as Old Irish and Celtiberian, lie two proto-languages and multitudes of sound laws and morphological and syntactic innovations. An additional complication is that the ProtoItalic and Proto-Celtic languages have not yet been reconstructed to their full potential. The more distant a relationship, the less fine-grained is the reconstruction of the common ancestor, which affects what can realistically be expected to constitute good evidence for common innovations. So it is in the case of Italo-Celtic. The procedure that I adopt in order to argue for a common Italo-Celtic node is a specific version of the one outlined above. With respect to what constitutes an identical innovation, I assume that if a phonological development in Celtic is similar enough to a development in Italic, one should attempt to formulate them as an identical sound law, and if that can be done successfully, the sound law presents evidence for Italo-Celtic. With respect to what constitutes an identical relative order, I assume that identical innovations in Italic and Celtic occurred in the same order if there is no reason to think that they did not. This procedure is evidently weaker than one which prescribes, first, that two innovations are identical only if they cannot possibly be formulated differently, and, second, that they occurred in the same order only if a relative chronology which confirms this has been explicitly worked out. Yet the version of the procedure adopted here has two advantages: it is, in my opinion, all we can aspire to in the case of distant relationships, while still being open to falsification.   Here follows a list of arguably Italo-Celtic sound changes: (1) Dybo’s rule: long vowels are shortened if they precede a resonant plus a vowel that was stressed in Proto-Indo-European (PIE). See Schrijver (1991, 334–57) for an extensive discussion.2 Example: PIE *wiH-ró- > post-PIE *w¼ró- (Sanskrit v¼rá- ‘hero’) > Italo-Celtic *wiro- > Old Irish fer, Latin vir, Old English wer ‘man’; PIE *HoHmó- > post-PIE *}mó(Sanskrit \má- ‘raw’) > Italo-Celtic *omo- > Old Irish om, Middle Welsh of ‘raw’. This rule affects Italic and Celtic. It also affects Germanic but as part of a different relative chronology of sound changes: in Germanic, Dybo’s shortening follows the development of 2

Schrijver CW3.indd 490

See also Meiser (1998, 75). The rule is based on earlier work by Kortlandt (1981) and Dybo (1961).

27/05/2016 19:52:58

Schrijver

[ 491 ]

pretonic *CHIC to *C ĪC, so that this new C ĪC is affected by it. In Italic and Celtic, pretonic *CHIC always yielded short *C ĬC, irrespective of whether the final C was a resonant (see development (2) below).   Isaac (2007, 21–59) only discusses Celtic material and presents a whole host of sound laws that involve laryngeal loss rather than shortening. The sheer number of sound changes and the fact that they are phonetically baffling (e.g. *h1 > ø /CR̥ _Ci, *h2 > ø /u_C) inspires little confidence in the correctness of this alternative to Dybo’s rule. In spite of Isaac’s formulation of numerous micro-conditions, Zair (2012, 146) managed to find a small number of reliable counterexamples.   Matasović (2009, 6–7) all too briefly touched upon Dybo’s rule, preferring the formulation that pretonic vowels (from PIE *V̄ and from PIE *VH) were shortened irrespective of the following consonant. Since there is no proper discussion, this opinion must stand as just that.   Zair (2012, 132–50) provides an extensive discussion of the relevant Celtic material. He rejects the formulation that I had adopted by quoting ‘reliable counterevidence’ (ibid. 2012, 145). His position can be explained in part by the fact that we disagree on rule (2) below: Old Irish sith- ‘long’ < *sh1 i-ti/tu- has *CHIC, which according to me regularly developed into short CĬC in pretonic position, a development which has nothing to do with Dybo’s rule (Schrijver 1991, 527), but Zair rejects the CHIC rule and is therefore forced to introduce sith under the heading of Dybo’s rule (2012, 145). For the same reason we disagree on Latin pŭtus ‘clean (?)’ < *ph1 u-tó- and lucrum ‘profit’ < *lh2 u-tló- (Schrijver 1991, 240–1; Zair 2012, 131, 144). Another source of disagreement concerns the problem of establishing the position of the stress in PIE. Zair’s rule of thumb is formulated as follows (2012, 113): [I]t is safe to assume that all zero-grade adjectives with the sufixes *-ro-, *-no-, *-to- and *-mo- were stressed on the suffix ... These adjectives could subsequently be substantivised, so zero grade nouns with these suffixes are included. Since it seems likely that nominalization tended to lead to accent retraction, these forms should, however, be treated with care.

The problem is that it is not safe to assume that adjectives with the aforementioned suffixes were always stressed on the suffix. Vedic adjectives such as vípra- ‘trembling’, túmra- ‘strong’, śvítna- ‘whitish’ are clear exceptions. It is correct, however, to state that verbal adjectives with *-no- and *-to- as well as all adjectives in *-mo- are always oxytone in Vedic and by extension in PIE.3 Hence without comparative evidence for oxytonesis in exact counterparts, neither Middle Welsh blin ‘weary’ < *ml¼no- (?) (Zair 2012, 114), nor Old Irish crín ‘withered’ < *kr¼no- (?) (ibid., 125–6), nor Old Irish úr ‘fresh’ < *p~ro- (ibid., 120–1), nor Old Norse súrr ‘sour’ < *s~ro- constitute counterevidence to my formulation of Dybo’s rule. That still leaves three counterexamples: Latin f~mus < *dhuH-mó- ‘smoke’ (Sanskrit dh~má- ‘smoke’, ibid., 144), Latin d~rus ‘hard’ (Sanskrit d~rá- ‘far’, ibid.) and Old Norse stúrr ‘big, strong’ (Sanskrit sth~rá- ‘strong’, ibid.). Since f~mus formed part of an etymological family of words in which *~ was represented (verbal *dh~-je- in Latin suf-f¼re ‘to fumigate’, nominal *dh~‑lVin Latin f~l¼g} ‘soot’, cf. also Lithuanian d~́lis ‘smoke’), I proposed that on the basis of these 3

Schrijver CW3.indd 491

See in general Lubotsky (1988).

27/05/2016 19:52:58

[ 492 ]

xvii. italo-celtic linguistic unit y

]]

forms *~ was restored in f~mus (Schrijver 1991, 342).4 With respect to Old Norse stúrr I remarked that this adjective may originally have been barytone and that Vedic sth~rá- shows oxytonesis because that became productive in adjectives with a suffix *-ro- in Vedic (ibid., 355–6). The same suggestion may be made in order to account for Latin d~rus, Vedic d~rá-. None of these three suggestions can be proved in any way, and they may even seem arbitrary. But such is the cost of saving a simple rule that accounts for the short vowel in a number of hitherto problematic words. Its plausibility is not to be measured in comparison with the ideal sound law (whatever that is) but by its simplicity and completeness in comparison with the simplicity and completeness of competing accounts. Zair himself does not present such an alternative, and his discussion ends in a prolonged question mark (2012, 147–50). (2) Pretonic CHIC > C ĬC. This development is found only in Italic and Celtic (Schrijver 1991, 225–48, 525–34, based on Kortlandt 1981). Example: PIE *bhHu-tó, -tí- > Italo-Celtic *bhŭ-tV- > Old Irish -both ‘one was’, buith ‘being’, Latin fut~rus ‘future’ (cf. Sanskrit bh~tá- ‘become’, bh~tí- ‘well-being’); PIE *gwHi-tú- (no independent evidence for the stress position) > Italo-Celtic *gwĭtu- > Old Irish bith, Middle Welsh byd ‘world’; PIE *ĝhh2u-tu- (no independent evidence for the stress position) > Italo-Celtic *ghŭ-tu- > Old Irish guth ‘voice’.   Twenty years of non-acceptance by Indo-Europeanist not belonging to the ‘Leiden school’ have shown this to be a highly controversial sound change. While there is no doubt that sequences of the type *CHIC existed beside *CIHC in PIE and therefore could in theory develop a different outcome in the daughter languages, what has proved to be a bone of contention is the solid identification of individual roots that contained a sequence *HI rather than *IH, especially in as far as this is tied up with Balto-Slavic accentuation (Schrijver 1991, 228–9). Additionally, short vowel reflexes of roots that originally contained *HI have been explained by analogy rather than sound law. For instance, it is assumed that *gwHi- ‘live’ had a stressed variant *gwiH-, which would regularly develop into *gwĭ- before a vowel. This allomorph was allegedly generalized in other positions, so that it ended up before consonants, too, in for example *gwi-tu-, which develops into Old Irish bith ‘world’. While Zair provides an extensive discussion of short vowel reflexes *HI in Celtic and concludes that there is no good evidence for them, he does not commit himself to an explanation of the problematic short vowels in words like Old Irish bith, guth, buith (2012, 121–5, 128–32). Therefore, the situation is comparable to Dybo’s rule discussed under (1): a relatively simple rule accounts for Italic and Celtic instances of short *i and *u if one assumes that they derive from PIE pretonic *HI (an assumption that is based on an analysis of material outside Celtic and Italic). The merit of this rule can only be judged in relative terms (is there a simpler, more correct and more complete account?) rather than in absolute terms (is the evidence for the sound change by itself compelling?). 4

Schrijver CW3.indd 492

Zair states that the vowel in *dh~-jé- should have been shortened according to Dybo’s rule since verbs with the suffix *je and a zero grade root had the stress on the suffix (2012, 144). That ignores the existence of the barytone Vedic type d¼́vyati ‘plays’, however. His remark that analogy to f~l¼g} is unlikely because that is a rare word in Latin confuses synchrony with diachrony.

27/05/2016 19:52:58

Schrijver

[ 493 ]

(3) The development of PIE mobile stress into word-initial stress is common to prehistoric Italic (e.g. Weiss 2009, 109–10 with n 16) and, probably, prehistoric Celtic (Schrijver 1995, 16–22). Given its typological triviality, this could be an independent development (as in Germanic, where fixing of initial stress occurred after a specifically Germanic development, Verner’s Law; or in Czech and Latvian, where it replaces inherited Balto-Slavic mobility). A reason for linking the change in Italic and Celtic can be formulated negatively: there is no reason for thinking it was not a common innovation (whereas there is a reason in the case of Germanic, Czech, and Latvian). If this is an Italo-Celtic development, it postdates (1) and (2), both of which presuppose the presence of PIE mobile accent. (4) PIE *CRHjV- became Italo-Celtic *CR¼jV-, possibly by a resyllabification rule: *CRHjV > *CRHijV > *CRiHjV- > *CR¼jV-. This rule combines the Italic forms that traditionally fall under the Proto-Italic ‘pius rule’, which states that long *~ became long *¼ before *j, with a number of problematic Celtic forms that show *¼ (Schrijver 2003, 75–8). Italic examples comprise *puH-jo- > *pw¼jo- > Latin p¼us ‘pious’; PIE *bhuH-joH > *bhw¼j} > Latin f¼} ‘I become’. Celtic examples include Old Irish biid, Welsh bydd ‘is wont to be’ < *bhw¼jeti < PIE *bhuH-jeti; Old Irish do-gní ‘does, makes’ < *gn¼jeti < PIE *ĝnh1-jeti. Nicholas Zair (see also Zair 2009) drew my attention to the problematic form, Latin inciens ‘pregnant’, which according to the rule should have become *inquiens < *en‑kw¼jent- < *enkuH-jent-. Since the rule produces an incorrect result, Zair rejects it and, for Italic, returns to the traditional formulation of the pius rule. A possible escape route is that inciens reflects a form with non-syllabic -j-, *enkjent- < enkwjent- (compare socius ‘ally’ < *sokwjos for the loss of *w in this position), and that this in turn reflects earlier *enkwĭent- < *enkw¼jent-, cf. (Meiser 1998, 76) on early shortening of *~, *¼ before vowel in non-initial syllable: f~it atque is pr}fŭit in Plautus (Captivi 555).5   Acceptance of this rule comes as a package deal: verbal forms such as Latin parĕre ‘to bring forth’ and Old Irish ‑gainethar ‘is born’ cannot reflect *prh3-je-, *ĝnh1-je-, as is usually assumed, but go back to athematic i-presents *prh3-(e)i-, *ĝnh1-(e)i- (Schrijver 2003, 78), a verbal category whose existence is not accepted by most Indo-Europeanists. (5) Between two consonants, a laryngeal yields short *ă in Italic and Celtic. It does so irrespective of whether the laryngeal stands in initial or medial syllable: *h2enh1m- > *anam- > Proto-Celtic *anam~ > Middle Breton eneff ‘soul’; Proto-Italic *anamo- > Latin animus ‘soul, spirit’ (Schrijver 1991, 415). The treatment was different in Germanic, Armenian, and possibly also in Albanian, where zero rather than *ă resulted in medial syllables (pace Zair 2012, 269; see Beekes 1988, 77, 96, 103). Tocharian does show the same treatment as Italic and Celtic (Beekes 1988, 86–7), but this probably represents an independent development, since Tocharian on the one hand, and Italic and Celtic on the other, do not share any other developments that can be translated into a common node on the family tree. 5

Schrijver CW3.indd 493

Zair (2012, 171) cites the Latin form as inc¼ens, with long -¼-, which is incorrect, first of all because the form does not occur in metrical texts, and secondly because of the antevocalic shortening rule in non-initial syllables.

27/05/2016 19:52:58

[ 494 ]

xvii. italo-celtic linguistic unit y

]]

(6) R̥T [voiced unaspirated]C > RaTC. Between a syllabic resonant to its left and a PIE voiced unaspirated plosive followed by a consonant to its right, a short *a develops in Italic and in Celtic. For example, *m̥g-lo- > Proto-Celtic *maglo- > Old Irish mál, Middle Welsh mael ‘prince’; *m̥g-no- > Proto-Italic *magno- > Latin magnus (Schrijver 1991, 415–16, 477–85; Meiser 1998, 64–5). Zair (2012, 65, 71 n 20) rejects the rule for Celtic because of a supposed lack of good evidence, although he retains the rule in his discussion of Old Irish mál (ibid., 190). (7) *mj > *nj. This change affects Italic and Celtic: *dhĝhom-jo- > Proto-Celtic *xdonjo- > Old Irish duine, Middle Welsh dyn ‘person’; *gwm̥ -j} > Latin veni} ‘I come’. The same rule also occurred in prehistoric Greek (*gwm̥ -j} > *bamj} > *banj} > βαίνω ‘I step’), but since Greek does not share other phonological innovations exclusively with Italic and/or Celtic, the rule probably occurred independently in Greek. (8) *p...kw > *kw...kw. The rule affects Italic and Celtic: PIE *penkwe > Italo-Celtic *kwenkwe > Middle Welsh pymp, Old Irish cóic, Latin quinque, Oscan púmp- ‘five’ (see McCone 1996, 44, also on presumably Celtic Hercynia). (9) A short vowel was syncopated if it stood (1) in a medial syllable and (2) between *s(T) and *s : *-sVs- > *-ss-; *-sTVs- > *-ts-. This highly specific rule affected Italic and Celtic superlatives (Cowgill 1970, 130–31): *nesd-is-m̥ Ho- > *nesdsm̥ Ho ‘nearest, next’ > Proto-Italic *netsemo- (> ultimately Oscan ness-imas ‘closest’, with unexplained -i-); Proto-Celtic *netsamo- > Gaulish neđđamon [ts], Middle Welsh nessaf ‘next, closest’. (10) *-mw- > *-w-: shared by Italic (Sommer & Pfister 1977, 195) and Celtic (McCone 1996, 49): *kom-wir-jo/\ > Proto-Italic *kowirjo/\ > Volscan couehriu, Latin c~ria ‘meetingplace’; *kom-wari(o)- > Proto-Celtic *kowari(o)- > Old Irish coäir, Middle Welsh kyweir ‘proper’. (11) PIE *sr-, *-sr- > Italo-Celtic *θr-, *-ðr-. In Latin, these joined the Italic reflexes of PIE *dh and yielded Latin fr-, -br- (*sr¼gos > fr¼gus ‘cold’, *kerasrom > cerebrum ‘brain’). In Celtic, word-initial *θr- became fr- in British Celtic and in Gaulish, while Irish apparently retained sr-: *sruto- > Welsh ffrwd, Old Irish sruth, ‘stream’. However, the Ogam alphabet had a separate letter for the reflex of PIE *s before *r (viz Ogam sraiph), which indicates that in Primitive Irish the sound was not *s (which is written with the Ogam letter sail ) but a different rather similar sound, such as *ts or *θ. After vowels, Celtic *-ðr- was initially retained (Gaulish nominative plural tidres ‘three’) but on its way into Insular Celtic *ð it was lost with compensation: accusative plural *tisr\s > Old Irish téora, Middle Welsh teir ‘three’ (feminine; cf. Cowgill 1957, Schrijver 1995, 440–52, pace McCone 1993). This concludes the list of identical sound changes, which form the phonological evidence for an Italo-Celtic linguistic unity. There are a number of early developments in Italic and Celtic

Schrijver CW3.indd 494

27/05/2016 19:52:58

Schrijver

[ 495 ]

that are not identical but do resemble one another strikingly. They probably belong to the stage at which the Italo-Celtic unity was gradually breaking up and innovations began to show small dialectal differences. It is notable that the PIE laryngeals were still retained at the time. (12) PIE *n̥, *m̥, had already begun to differentiate into Celtic *an, *am but Latin *en, *em; Sabellian Italic is intermediate: *an, *am in initial syllables, and *en, *em in other syllables (Meiser 1986, 69–70). The only secure example in the Venetic branch of Italic is the third person plural preterite donasan ‘they gave’ < *-s- n̥t (Lejeune 1974, 107), which shows that in non-initial syllables it joined Celtic rather than Latin or Sabellian.   This difference in vocalization between Celtic and the various branches of Italic is reflected in the following two sound changes, which otherwise proceeded in identical fashion. (13) *CR̥HV > *CVRV: PIE *sm̥h2-el-i- > Latin *semali- > similis ‘similar’; Proto-Celtic *samali- > Old Irish samail ‘similar’; PIE *wl̥h2-V- > Latin val-{re ‘to be strong’, Proto-Celtic *wal-o- ‘ruler’ in for example Old Irish Domn-ual > Domnall (Schrijver 1991, 220–22; Zair 2012, 169–70). (14) *HR̥C- > *VRC: PIE *h3 m̥bh-el- > *h3 embhel- > *ombhel- > Latin umbil¼cus ‘navel’; Proto-Celtic *ambel- > Old Irish imbliu ‘navel’; PIE *h2 r̥ ĝn̥to- > Latin argentum ‘silver’, Proto-Celtic *argantom > Old Irish arcat ‘silver’ (Schrijver 1991, 72–3, 419; Zair 2012, 29–38). (15) Both Italic and Celtic show a double treatment of the cluster *CR̥HC. In Italic, this yielded *CR\C unless the final C was a plosive (symbol T ) that was followed by another consonant, in which case *CRăTC resulted (e.g. Latin l\tus ‘carried’ < *tl̥h2-to- versus glăber ‘smooth’ < *ghl̥h2 dh-ro-; Schrijver 1991, 417). For Celtic, I suggested that *CRHC became *CR\C unless the final consonant C was a plosive T, in which case *CRăT ensued (e.g. Old Irish grán ‘corn’ < *ĝr̥ Hno- versus flaith ‘rule’ < *wl̥h2ti-; Schrijver 1995, 168–91; Schumacher 2004, 136–9). In order to account for a handful of difficult cases in Celtic, such as Old Irish tláith ‘weak’ < *tl(e)h2-ti-, Zair (2012, 69–89) modified the Celtic rule by suggesting that short -a- only resulted in the constellation *C [non-plosive]RHT, while TRHT yielded TR\T, as in tláith. Since Zair did not provide new considerations concerning those difficult cases, and since he and I only disagree on the weight that can be ascribed to them, it is impossible to make further headway. Whatever the exact conditions of the Celtic change, both the difference and the similarity between the double treatment in Celtic and in Latin are evident. As a first conclusion we may state that, on the basis of the evidence for Italo-Celtic, it is possible to extend the family tree of the Italic subgroup of Indo-European by adding Celtic. Celtic was probably the first to branch off, before Venetic did. The position of Venetic was clarified by Meiser (2003, 35–6, 84–5), who argued convincingly that Venetic belongs to the Italic family because it shares the development of the mediae aspiratae with Italic (rather than Celtic; see section 3 below), but it split off before the specifically Italic way of merging aorist and perfect developed, which can be found in Latino-Faliscan and Sabellian.

Schrijver CW3.indd 495

27/05/2016 19:52:58

[ 496 ]

xvii. italo-celtic linguistic unit y

]]

3. The development of the PIE voiced and aspirated plosives in Celtic and Italic In Celtic, the PIE voiced (*b, d, ĝ, g, gw) and aspirated plosives (*bh, dh, ĝh, gh, gwh) merged as voiced plosives: Proto-Celtic *b, d, g, gw (where *g < PIE *ĝ and *g, and *b < PIE *b and *gw; PIE *gwh became Proto-Celtic *gw). How and why the merger of the two PIE series took place is unclear; the development is generally taken for granted.   During the early stages of Italic, the two series remained distinct. In Proto-Italic, including Venetic, the aspirated plosives became fricatives: *bh, dh, ĝh/gh, gwh > /β, ð, γ, γw/. This development did not take place after nasals, where voiced plosives developed (Meiser 1986, 75–7; 1998, 104), nor after *s, but fricatives did occur after the liquids *r and *l.6 Still in Proto-Italic, these fricatives developed voiceless allophones [f, θ, x, xw] in word-initial position, and subsequently these allophones split off as separate phonemes when [f, θ, xw] merged into /f/, leaving [x], which became /h/. Meanwhile, the PIE voiced plosives *b, d, ĝ/g, g w remained unchanged, apart from the merger of * ĝ and *g into *g. 7   The considerable differences between the developments of the plosives in Italic and Celtic leave little room for assuming common Italo-Celtic innovations. Essentially, it seems, the Italo-Celtic plosive system was identical to the PIE plosive system. As I shall argue, however, this may not be accurate: if we assume that Celtic took part in the first innovation that affected the aspirates in Italic, viz the fricativization of *bh, dh, ĝh/gh, gwh to *β, ð, γ, γw, we may be able to understand why the PIE voiced and aspirated plosives merged in ProtoCeltic. The argument runs as follows.   The first stage is the PIE stage, which is illustrated here on the basis of the dental series only (mutatis mutandis, the labial and three velar series behave identically): A: PIE

word-initially

after *V, *r/l

/t/

t-

t

/d/

d-

d

/dh/

dh-

dh

after *N, *s

word-finally

t

-

d

-d8

dh



-

Let us suppose that the initial stage of the Italic development, when /dh/ became a fricative /ð/ except after nasal and *s, belonged to the Italo-Celtic period:

6

7 8

Schrijver CW3.indd 496

A voiced fricative probably did not develop after *s, judging by forms like Latin vastus ‘vast’ < *wasdho-, hasta ‘spear’ < *ghasdh\ as opposed to n¼dus ‘nest’ < *nisdo- (cf. Meiser 1998, 101, 119). These forms suggest either that *sdh became either *sθ > st, or by loss of aspiration and voice but without an intermediate fricative stage,*st. After *r and *l, a voiced fricative did develop: Umbrian ALFU < *albh\ ‘white’, UERFALE ‘templum’ (Meiser 1986, 78, 1998, 104). For example Meiser (1998, 101–5; 2003, 31.) See also in general Stuart-Smith (2004) with differences of detail concerning intermediate stages. PIE lacked an opposition between the three phonation series in word-final position, cf. Szemerényi (1973); Cowgill (1975, 52).

27/05/2016 19:52:58

[ 497 ]

Schrijver

B: (It-Cl.)

word-initially

after *V, *r/l

after *N, *s word-finally

/t/

t-

Vt

Ct

/d/

d-

Vd

Cd

/ð/

ð-

Vð

Cd( )



-d h

-

The next stage C is Proto-Celtic and involves the first stage of Celtic lenition, which because it affected all Celtic languages including Celtiberian was probably Proto-Celtic: voiced plosives after vowels became fricatives (McCone 1996, 85–7): C: (PCl.)



word-initially after *V

after *r/l after *N, *s word-finally

/t/

t-

t

t

t

-

/d/

d- ~ -V ð-

ð

d

d

-ð

/ð/

ð-

ð

ð

d

-

On the assumption that this lenition was effective across word boundaries, as later Celtic lenitions were, word-initial *d obtained an allophone [ð] if the preceding word ended in a vowel. This first lenition seriously weakened the contrastiveness of the phonemes /d/ and /ð/, which now remained in force word-initially, but only if the preceding word did not end in a vowel, and word-internally after *r, *l but not after other consonants. In all other contexts, there was no opposition between /d/ and /ð/. What Celtic did in reaction was to simplify the system by accompanying the lenition rule ‘*d > ð after a vowel’ with a complemetary rule ‘*ð > d after a consonant’, thus giving up the contrast between the /d/ and /ð/ series altogether: D: (PCelt.) word-initially after vowel

after consonant

word-finally

/t/

t-

Vt

Ct

-

/d/

d- ~ -V ð-

Vð

Cd

-ð

In summary, if we assume that Celtic shared the fricativization of the PIE aspirated plosives with Italic, it is possible to understand how Celtic lost the opposition between the PIE voiced and aspirated plosives as soon as it underwent the first lenition of voiced plosives. This is preferable to the traditional ‘just so’ assumption.   It is also possible to understand why Celtic did not take part in subsequent Italic developments of the PIE aspirated plosives: the merger of the word-initial devoiced allo­ phones [ f, θ, xw] of the phonemes /β, ð, γw/ into the new phoneme /f/. In Proto-Celtic, PIE *p had developed into a voiceless labial fricative /f/ or /φ/. Phonologically this was an insignificant development, but the presence of this phoneme in Proto-Celtic may well have entailed a much more significant consequence: it could have blocked the rise of new /f/ from [f, θ, xw] and thus helped to force Celtic away from the rest of Italo-Celtic.   The position of Venetic is intriguing. On the one hand, it sided with Italic in retaining PIE /p/ and as a result went along with Italic in the development of word-initial /β, ð, γw/ into /f/. On the other hand, it sided with Celtic with regard to the development of

Schrijver CW3.indd 497

27/05/2016 19:52:58

[ 498 ]

xvii. italo-celtic linguistic unit y

]]

the syllabic nasals to *aN (on the sole evidence of Venetic donasan, see section 2, rule (12)). Venetic also sided with Celtic in abandoning the phonological distinction between /d, b, g/ (< PIE *d, *b, *ĝ/g) and /ð, β, γ/ (< PIE *dh, *bh, *ĝh/gh). As Helmut Rix has argued in a brilliant article (1997), the Venetic phonemes /d, b, g/ had plosive and fricative allophones: plosive [d, b, g] in word-initial position (which had developed from PIE *d, *b, *ĝ/g); and fricative [ð, β, γ] post-vocalically and possibly word-internally in general (which had developed from PIE *d, *b, *ĝ/g, as well as *dh, *bh, *ĝh/gh). In other words, Venetic adopted the equivalent of the Celtic lenition rule, and like Celtic underwent the spirantization of postvocalic *d, *b, *g and their merger with the Proto-Italic phonemes *ð, *β, *γ. The only securely known difference between the Celtic and Venetic treatment of the plosives concerns the PIE aspirated plosives in word-initial position, which Venetic turned into *f and h, while Celtic merged them with *d, *b, *g. As I argued earlier, this difference is a consequence of the development of *p > *f/φ in Celtic. 4. The Proto-Celtic long vowel system As a result of two Proto-Celtic developments, the Italo-Celtic system of six long vowels was reduced to four. Italo-Celtic *} split into Proto-Celtic *~ in final syllables and *\ elsewhere. Italo-Celtic *{ became Proto-Celtic *¼. It did so after the loss of intervocalic *j (*eje > *ee > *{ > *¼; McCone 1996, 49). Between the split of *} > *~/*\ and the raising of *{ > *¼, another Proto-Celtic sound change took place: long vowels before word-final nasal were shortened (Old Irish accusative singular dé ‘day’ < Proto-Celtic *djĕm < PIE *dj{m). The loss of intervocalic *j also affected Proto-Italic (Meiser 1998, 91; 2003, 31) and there is no reason why this development could not be pushed back to Italo-Celtic. Shortening of long vowels before word-final nasal also affected Latin, but this change cannot belong to ProtoItalic, let alone Italo-Celtic, because Oscan paam (= Latin quăm) preserves vowel length (Meiser 1998, 77).   Hence the Italo-Celtic long vowel system (A) became the Proto-Celtic system (B) after it had passed through at least one specifically Proto-Celtic sound change (-V̄ N > -V̆ N ):

A:



ī

ū

ē

ō



9

Schrijver CW3.indd 498

B:

ī

ū

ē

ā

ā

When we compare this development with changes that affected Italic, it is remarkable that system A was preserved in Latino-Faliscan and, probably, in Venetic (Lejeune 1974, 108– 9), while A developed into system B in Sabellian. However, the sound laws that produced system B in Sabellian were different from those that produced it in Celtic. In Sabellian, old *~ became *ȳ, then *¼, while old *} filled the resulting gap in the system by developing into *~ generally (not just in final syllables, as in Celtic).9 So the triangular long vowel system

Meiser (1986, 49–52) has *} develop into closed *ọ̄, but since this is the only close long round back vowel, it may as well be written /~/. Meiser (ibid., 52–4) hesitates accepting *~ > *¼, but the two exceptions, Paelignian saluta and Oscan fruktatiuf, may well have been influenced by Latin sal~t- and fr~ctus.

27/05/2016 19:52:58

Schrijver

[ 499 ]

evolved by different routes into the same square system of Sabellic and Proto-Celtic, while Latino-Faliscan and Venetic retained the triangular system. For that reason, the changes observed in Proto-Celtic and Sabellian do not reflect a common node on the family tree. The fact that the same long vowel system is produced in Celtic and Sabellian by different sound laws points either to a chance similarity or to contact. Since we have seen evidence that Italo-Celtic gradually broke up into its constituent members, it is a distinct possibility that the ancestors of Celtic and of Sabellian were spoken in one another’s vicinity and that this was the setting in which their long vowel systems converged. 5. Conclusion on Italo-Celtic Eleven sound laws are shared between Italic and Celtic, forming the phonological evidence for an Italo-Celtic node on the Indo-European family tree (section 2, 1–11). Four sound laws are sufficiently similar to warrant ascribing them to a stage when Italo-Celtic was beginning to break up (section 2, 12–15). During the breakup, syllabic nasals became sequences of vowel + nasal and laryngeals finally disappeared. Syllabic liquids were still retained (*r̥ , *l̥ > Italic *or, *ol, Celtic *ri/ar, *li/al). Against this backdrop, describing changes to the plosive system of Celtic within the framework of a gradually disintegrating Italo-Celtic subgroup increases our explanatory grasp of the changes involved, and is therefore preferred to describing those changes on the basis of Celtic (or Italic) alone (section 3). It seems that the development of *p to *f/φ in Celtic was instrumental in driving Celtic away from the rest of Italo-Celtic because this prevented Celtic from turning the PIE aspirated plosives into word-initial *f-. Apart from that, the similarity of Celtic and Venetic is striking. After the breakup phase, the development of the long vowel system in Celtic and Sabellian is convergent, which suggests that the two language groups were spoken in geographical proximity (section 4).   The results of the discussion can be presented in simplified form as the upper section of the following Italo-Celtic family tree (Figure 17.1): Proto-Italo-Celtic

Latino-Faliscan Latin Faliscan

Sabellian Umbrian South Picene

Schrijver CW3.indd 499

Venetic

Proto-Celtic

Oscan 17.1 Proposed ItaloCeltic family tree

27/05/2016 19:52:59

[ 500 ]

xvii. italo-celtic linguistic unit y

]]

Celtic was the first to branch off from Italo-Celtic, followed by Venetic. The similarity of early Celtic to early Venetic and the convergence of the Celtic and Sabellian long vowel systems suggest that Celtic was spoken in geographical proximity to those two Italic branches. 6. Language and prehistory The results of sections 1–5 are relevant to the question of where the speakers of ProtoCeltic may be located. The close linguistic ties between Italic, more particularly Venetic and, later, Sabellian, and the earliest stages of Celtic on the one hand, and the fact that all Italic languages were spoken within the confines of present-day Italy on the other, strongly suggest that Celtic originated somewhere in or close to northern Italy. This chimes well with the consequence that the earliest attested Celtic language, Lepontic, whose texts, dating from the 6th to the 1st century BC, are found in a wide arc around Lake Lugano, roughly between the St. Gotthard pass and Milan, had stayed close to the original homeland of the Celtic family.   The culture that produced the early Lepontic texts is the later Golasecca culture. This developed from the Canegrate culture of the 13th–12th centuries BC, which was part of the Urnfield cultural complex. The Canegrate culture was probably intrusive, showing as it did connections with Urnfield cultures in north-western Italy, the south-east of France and north-western Switzerland. By contrast, the Canegrate culture was isolated from its southern and eastern neighbours until the 10th century BC. Those western and north-western links of the Canegrate area have their roots in the Middle Bronze Age. It has been suggested that the Canegrate area and north-western Italy may have been the points of departure for cultural influences, which perhaps included population movements, across the north-western Alps rather than vice versa (De Marinis 2000).   This archaeologial scenario ties in well with the evidence from linguistics as discussed in this chapter: the Celtic homeland may be sought in north-western Italy, where Celtic split off from its Italo-Celtic brethren and gradually extended itself north-westwards via the Alpine valleys towards central Europe and, one might add, south-westwards towards Spain (Celtiberian).   This scenario enables us to solve another problem that bedevils the history of Lepontic. The linguistic differences between early Lepontic (6th–5th centuries BC) and the language of the Gauls, who entered Italy around and after 400 BC, apparently were small enough to allow complete convergence into Cisalpine Gaulish in the centuries after 400 BC. That sits uneasily with the idea that Lepontic had been separated for at least 800 years from the rest of Celtic, which would follow from the assumption that the Canegrate culture was introduced by an isolated pocket of Celtic speakers who had migrated from a Celtic homeland beyond the Alps. Less problematic would seem to be a scenario according to which ‘Canegrate’ Celtic was on the south-eastern edge of an early Celtic continuum stretching westwards across Piemonte into France and north-westwards to the Helvetian plain. Contact with its western Celtic neighbours, who may be De Marinis’ Bronze Age cultures in Piemonte (2000, 119), could in that case have slowed down the process of linguistic differentiation between Lepontic and the rest of Celtic.   The isolation of the Canegrate culture towards the east and south until the 10th century

Schrijver CW3.indd 500

27/05/2016 19:52:59

Schrijver

[ 501 ]

BC might be related to a lack of contact between speakers of Celtic and speakers of RhaetoEtruscan during that period.10 This changed after c. 900 BC, when the Lepontic Golasecca culture flourished as a result of transalpine trade between the Etruscans of the Po Valley and central Europe. As lasting tokens of this new liaison, Lepontic adopted the RhaetoEtruscan script, Etruscan grave culture and the Etruscan genitive suffix -al, which became the basis for new patronymic suffixes -al-os, -al-\.

BIBLIOGRAPHY Beekes, R. S. P. 1988 ‘Laryngeal Developments: a survey’, Die Laryngaltheorie, ed. A. Bammesberger, 59–105. Heidelberg, Winter. Cowgill, W. 1957 ‘Old Irish teoir and cetheoir’, Language 33, 341–5. Cowgill, W. 1970 ‘Italic and Celtic Superlatives and the Dialects of Indo-European’, Indo-European and Indo-Europeans, eds G. Cardona, H. Hoenigswald, & A. Senn, 113–53. Philadelphia, University of Pennsylvania Press. Cowgill, W. 1975 ‘The Origin of the Insular Celtic Conjunct and Absolute Verbal Inflexions’, Flexion und Wortbildung, ed. H. Rix, 40–70. Wiesbaden, Reichert. De Marinis, R. 2000 ‘Il Bronzo Recente nel canton Ticino e la cultura di Canegrate’, I Leponti tra mito e realtà I-II, eds. R. de Marinis & S. Biaggio Simona, 93–121. Locarno, Armando Dadò. Devoto, G. 1929 ‘Italo-greco e Italo-celto’, Silloge linguistica dedicata alla memoria di G.I. Ascoli, 200–40. Turin, Chiantore. Dybo, V. 1961 ‘Sokrashchenie dolgot v kel’to-italijskix i ego znachenie dlja balto-slavjanskoj i indoevropejskoj akcentologii’, Voprosy slavjanskogo jazykoznanija 5, 9–34. Isaac, G. R. 2007 Studies in Celtic Sound Changes and Their Chronology. Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Jasanoff, J. 1994 ‘The Brittonic Subjunctive and Future’, In Honorem Holger Pedersen, ed. J. E. Rasmussen, 199–220. Wiesbaden, Reichert. Jasanoff, J. 1997 ‘An Italo-Celtic Isogloss: the 3pl. mediopassive in *-ntro’, Festschrift for Eric Hamp, vol. 1, ed. D. Q. Adams, 146–61. Washington, D.C., Journal of Indo-European Studies. Kortlandt, F. 1981 ‘More evidence for Italo-Celtic’, Ériu 10

Schrijver CW3.indd 501

32, 1–22. Kortlandt, F. 2007 ‘Italo-Celtic’, Italo-Celtic Origins and Prehistoric Development of the Irish Language, ed. F. Kortlandt, 149–57. Amsterdam/Atlanta, Rodopi. Lejeune, M. 1974 Manuel de la langue vénète. Heidelberg, Winter. Lottner, C. 1861 ‘Celtisch-italisch’, Beiträge zur vergleichenden Sprachforschung 2, eds A. Kuhn & A. Schleicher, 309–29. Lubotsky, A. M. 1988 The System of Nominal Accentuation in Sanskrit and Proto-Indo-European. Leiden/Boston, Brill. McCone, K. R. 1993 ‘Old Irish “three” and “four”: a question of gender’, Ériu 44, 53–73. McCone, K. R. 1996 Towards a Relative Chronology of Ancient and Medieval Celtic Sound Change. Maynooth, Department of Old and Middle Irish, St. Patrick’s College. Marstrander, C. 1929 ‘De l’unité italo-celtique’, Norsk Tidskrift for Sprogvidenskap 3, 241–68. Matasović, R. 2009 Etymological Dictionary of ProtoCeltic. Leiden/Boston, Brill. Meiser, G. 1986 Lautgeschichte der umbrischen Sprache. Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Meiser, G. 1998 Historische Laut- und Formenlehre der lateinischen Sprache. Darmstadt, Wissenschaftliche Buchgesellschaft. Meiser, G. 2003 Veni Vidi Vici. Die Vorgeschichte des lateinischen Perfektsystems. München, Beck. Ringe, D. A. 1988 ‘Laryngeal isoglosses in the western Indo-European languages’, Die Laryngaltheorie, ed. A. Bammesberger, 415–41. Heidelberg, Winter. Rix, H. 1997 ‘Germanische Runen und venetische

See Schumacher (1998) on the relationship between Rhaetic and Etruscan.

27/05/2016 19:52:59

[ 502 ]

xvii. italo-celtic linguistic unit y

Phonetik’, Vergleichende germanische Philologie und Skandinavistik, eds T. Birkmann, H. Klingenberg, D. Nübling, E. Ronneberger-Sibold, 231–48. Tübingen, Max Niemeyer. Schrijver, P. 1991 The Reflexes of the Proto-Indo-European Laryngeals in Latin. Amsterdam/Atlanta, Rodopi. Schrijver, P. 1995 Studies in British Celtic Historical Phonology, Leiden Studies in Indo-European 5. Amsterdam, Rodopi. Schrijver, P. 1997 Studies in the History of Celtic Pronouns and Particles. Maynooth, Department of Old and Middle Irish, National University of Ireland Maynooth. Schrijver, P. 2003 ‘Athematic i-presents: the Italic and Celtic evidence’, Incontri Linguistici 26, 59–86. Schrijver, P. 2006 Review of Gerhard Meiser, Veni, Vidi, Vici. Die Vorgeschichte des lateinischen Perfektsystems (München 2003), Kratylos 51, 46–64. Schrijver, P. 2007 ‘Some Common Developments of Continental and Insular Celtic’, Gaulois et celtique continental, eds. P.-Y. Lambert & G.-J. Pinault, 355–71. Geneve, Droz. Schrijver, P. 2015 ‘Pruners and Trainers of the Celtic Family Tree: the Rise and Development of Celtic in the Light of Language Contact’, Proceedings of the XIV International Congress of Celtic Studies, Maynooth 2011, ed. L. Breatnach, R. Ó hUiginn, D. McManus, K. Simms, 191–219. Dublin, Dublin Institute for

Schrijver CW3.indd 502

]]

Advanced Studies. Schumacher, S. 1998 ‘Sprachliche Gemeinsamkeiten zwischen Rätisch und Etruskisch’, Der Schlern 72(2), 90–114. Schumacher, S. 2004 Die keltischen Primärverben. Innsbruck, Innsbrucker Beiträge zur Sprachwissenschaft. Sommer, F. & R. Pfister 1977 Handbuch der lateinischen Laut- und Formenlehre I: Einleitung und Lautlehre. Heidelberg, Winter. Stuart-Smith, J. 2004 Phonetics and Philology: Sound Change in Italic. Oxford, Oxford University Press. Szemerényi, O. 1973 ‘Marked–Unmarked and a Problem of Latin Diachrony’, Transactions of the Philological Society 1973, 55–74. Watkins, C. 1966 ‘Italo-Celtic Revisited’, Ancient IndoEuropean Dialects, eds H. Birnbaum & J. Puhvel, 29– 50. Berkeley/Los Angeles, University of California Press. Weiss, M. 2009 Outline of the Historical and Comparative Grammar of Latin. Ann Arbor, Beech Stave Press. Zair, N. 2009 ‘OIr. biid < *bhuH-ye/o- and ‘Hiatus’ Verbs’, Proceedings of the 20th UCLA Indo-European Conference, Los Angeles 2008, eds S. Jamison, C. Melchert, & B. Vine, 213–20. Bremen, Hempen Verlag. Zair, N. 2012 The Reflexes of the Proto-Indo-European Laryngeals in Celtic. Leiden/Boston, Brill.

27/05/2016 19:52:59

chapter eighteen — ancillary study

Celtic as Vasconized Indo-European? Three structural arguments

Theo Vennemann

Introduction: Celtic from the East or from the West?

A

to the traditional theory of the origin and spread of Celtic, as taught in the handbooks of Indo-European and of Celtic, the Celtic branch of Indo-European was brought westward to Europe from an Indo-European homeland farther east.1 Proto-Celtic acquired its distinguishing features during and after its separation from the Proto-Indo-European speech community, and in a remarkable series of expansions, emanating from weste1rn central Europe (present-day western Austria, southern Germany, northern Switzerland, and eastern France), carried them to northern Germany, Gaul, the Iberian Peninsula, the British Isles, Italy, and Asia Minor. Here are two quotations, one from an Indo-Europeanist, the other from a Celtologist: ccording

The Celts came from Central Europe, from whence they swarmed out in a number of directions. [...] For instance, the Galatians of St. Paul’s letters who lived in Central Turkey were Celts. [...] In Italy they settled in the Po valley, in the Iberian peninsula in the east already in the 6th c[entury] BC. (the Celtiberians). From France they moved to the British Isles, where their languages survived. (Beekes 2011, 27, 309) The earliest named Celts (in Greek and Latin sources) are associated with two major central European Iron Age cultures, the Hallstatt, dated to the seventh century BC, and La Tène, dated to the fifth century BC. The archaeological evidence suggests a cultural continuity backwards through the late Bronze Age Urnfield Culture with no material evidence that the Celts were newcomers to the region. During the Hallstatt and La Tène eras the Celts enjoyed a period of great power and expansion. They spread from their central domain in different migrations over the whole of Europe: east and south through the Balkans to Asia Minor (crossing through the Hellespont in 278 BC), south into Italy (Rome was captured in 390 BC), west into the Iberian Peninsula, north to the Atlantic coast and across into Britain and lreland, where they were a dominant force by the third century BC. (MacAulay 1992, 1) 1

As in previous publications I mark reconstructed forms by a raised cross (+) and incorrect forms by an asterisk (*). In cited material I respect the practice of the authors.

Vennemann CW3 19 iv 2016.indd 503

27/05/2016 19:56:00

[ 504 ]

xviii. celtic as vasconized indo-european?

Celtologists in particular seemed to be in considerable agreement on the question of Celtic origins, as shown by a comparison of the second quotation above e.g. with the very similar introductory passage in Fife (1993, 3).   However, in a series of articles in two recent books (Cunliffe & Koch 2010; Koch & Cunliffe 2013), an alternative theory is presented according to which Celtic acquired its defining features in or near the Iberian Peninsula and then spread from there, long before the Hallstatt Iron Age, to the British Isles, to France, and to central Europe, carrying its specific linguistic features to these secondarily Celticized regions. While I am not qualified to discuss the historical and archaeological evidence produced in the two books in favour of this new theory, one of the arguments used for the ‘Celtic from the West’ theory is a linguistic one. It focusses on phonological Celtic features which have counterparts in Iberian and in Basque and Aquitanian, suggesting to the authors of this new theory that Celtic may have acquired these features in contact with those languages, i.e. in the ‘West’, and taken them from there to the British Isles and central Europe.2 Celtic and Vasconic In the present paper I will contrast this new theory with the Vasconic Theory which, though originally addressed to very different problems and being more general, does have a direct application to the question of Celtic origins. The Vasconic Theory argues that the languages spoken in large parts of Europe north of the Pyrenees and the Alps before the IndoEuropeanization of most of the Continent, the Old European languages, were Vasconic languages, i.e. languages of a family whose only survivor is Basque. The Old European branch of Vasconic is extinct. A further extinct branch, suggested by the toponymic research of Hubschmid (1963)3 and confirmed by recent work of Eduardo Blasco Ferrer (2010; 2013; 2015; forthcoming, also Ingrassia & Blasco Ferrer 2009, 18–23), is one of the pre-Roman languages of Sardinia, Palaeo-Sardic. Thus three branches of the Vasconic language family4 have so far been identified, the third (or first) being the family’s namesake Basque with 2

Zeidler (2011) says, ‘In the end, the hypothesis of an Atlantic origin of the Celtic languages remains a problematic issue. […] The results from historical linguistics are so far not decisive.’ Meid (2015) rejects the idea for a variety of reasons. Neither of them mentions the major linguistic argument proposed in its favour, the loss of the voiceless labial obstruent on a Basque/Iberian substrate. In van Sluis (2014) the loss of the voiceless labial obstruent is mentioned, but the specific use made of it within the ‘Celtic from the West’ theory is neither mentioned nor discussed. 3 Cited in Vennemann (1993, 472 [2003, 65]). Some of Hubschmid’s toponyms based on is-, which he compares to Basque +iz- ‘water’, recur in Wolf ’s (1998, 38) list of toponyms: isene, iseri, isoro, isalle, isedone, iserrai (cf. also Blasco Ferrer 2015, 167, 177f). 4 The name of the language family derives from that of a pre-Roman tribe named Vascones, between the upper course of the Ebro and the Pyrenees, in modern Navarra and parts of Aragon, mentioned by several Classical authors (Strabo, Pliny, Ptolemy, Livy; cf. Cancik & Schneider 2002, XII, s.v.) and assumed to be ancestral to the Basques of mediaeval and modern history; see also the Internet site ‘Vascones’. The name Vasconic for the language family seems to have found a measure of acceptance; cf. the Wikipedia sites ‘Vasconic languages’ and ‘Vasconic substratum theory’.

Vennemann CW3 19 iv 2016.indd 504

27/05/2016 19:56:00

[ 505 ]

vennemann

closely related extinct Aquitanian.5 Thus the family can be presented graphically by the following tree (Figure 18.1)6: Vasconic

BasqueAquitanian

PalaeoSardic

Old European7

18.1 The Vasconic languages

Two further branches may have to be added, which may be named (a) Northern Hispanic and (b) Palaeo-Iberic (the latter named on the model of Palaeo-Sardic). Ad (a): Garvens (1964) has shown that the toponymy of all of northern Spain contains Basque elements, which points to a Vasconic substrate; and that Castilian developed on a Vasconic substrate is a widely accepted view and will be demonstrated with selected material below. Ad (b): Román del Cerro (1990) identifies a considerable number of toponyms with Basque roots in the Alicante region and draws the conclusion—erroneously, in my opinion—that Iberian is related to Basque; the conclusion to be drawn is, in my view, that Iberian developed on a Vasconic substrate evident in Román del Cerro’s toponyms.8   It follows that the Indo-European languages of the West all came into contact with Vasconic languages, and some developed on Vasconic substrates. They may therefore be assumed to preserve, to a greater or lesser extent, linguistic features that are non-IndoEuropean but can be traced to Vasconic, because these features either survive in Basque or may be reconstructed for older levels of Basque.   I suggested as early as 1994 that this development on a Vasconic substrate holds in particular for the Celtic languages (cf. Vennemann 1994a; 1995). In terms of the general theory of language contact, these Vasconic features of Celtic were carried into pre-Celtic by Western Central European Vascons in the process of language shifting: the Old Europeans who learned the new Indo-European superstrate languages of their territory learned them imperfectly, unable to rid themselves of the habits of their first languages, which were 5 6 7 8

Cf. Trask (1997, 398–403) for a demonstration of the close relatedness or identity of Basque and Aquitanian. The tree combines my tree for Old European (Vennemann 2003a, 320 [2012, 258]) with Blasco Ferrer’s tree for Palaeo-Sardic (2013, 31). The designation ‘Old European’ is the one I used in my studies of the Vasconic European toponymy. It may be preferable in certain contexts to say ‘Old Western Central European’. Placing the three (or five) branches of Vasconic on the same horizontal line is not to imply that they must be seen on the same time level. The Indo-European languages are represented by similar trees, yet their branches have vastly diverging chronologies. The languages on the branches are likely to have been different from each other not only in space and time but also in linguistic terms. However, considering the extremely conservative character of Basque—words attested in both Aquitanian and Basque show hardly any structural or semantic difference, despite a time gap of nearly two thousand years—I would expect the languages on the Vasconic family tree to be much more similar than those on Indo-European family trees.

Vennemann CW3 19 iv 2016.indd 505

27/05/2016 19:56:00

[ 506 ]

xviii. celtic as vasconized indo-european?

Vasconic and thus lexically and structurally different from Indo-European.   General principles of language contact as taught in the specialized literature9 warn not to expect much lexical importation from substrates into superstrates, except in the toponymy, but to investigate structural modifications instead. Nevertheless there are a number of much-cited words, typical substrate borrowings, which prove that Celtic had linguistic contact with Vasconic. I only cite the following two:   (1) The word for ‘(young) woman, lady’, Basque and Aquitanian andere, which was carried by the Celts all the way to the British Isles: Irish aindir ‘girl’, Welsh anner ‘heifer’, also Middle Breton annoer ‘heifer’ (cf. Vennemann 1998, 12 [2003b, 603]; Schrijver 2002), and which also made it into French, either directly or via Gaulish: 14th-century andre ‘woman’, andrimelle ‘woman or girl’ (argot, a. 1597), andrumelle ‘(female) negotiator, procuress’ (argot, a. 1623; cf. Vennemann 1998, 25 [2003b, 613]).   (2) The word for ‘horn’, Old Irish adarc ‘horn’, Basque adar ‘horn’ (cf. Pokorny 1923, 273; Vennemann forthcoming, §1.1). Pokorny (1927–1930, 112) wrote: ‘Das gallische adarcá “Schilfschaum” < ursprünglich “Horn” erscheint sonst nur im Irischen adarc “Horn” und ist von den irischen Kelten zweifellos aus Gallien mit herübergebracht worden. ... Das Wort gehört zu bask. adar “Horn”, das mit dem keltischen -ko-, -ka-Suffix versehen wurde.’ [Gaulish adarcá ‘adarce’ < originally ‘horn’ is otherwise only evident in Irish adarc ‘horn’ and was undoubtedly brought over from Gaul by the Irish Celts. ... The word belongs to Basque adar ‘horn’, which was provided with the Celtic -ko-, -ka- suffix.]   There occur furthermore in the western Indo-European languages a number of suffixes which have close counterparts in Basque or Aquitanian (for Celtic in particular cf. Vennemann 2010b, §§3.2, 3.4). For example, the ‘Celtic -ko-, -ka- suffix’ of Old Irish adarc (cf. the preceding paragraph) also looks very Basque; e.g. Basque adarko ‘little horn (cf. German Hörnchen)’, adarreko ‘of, at the horn’.10   The present paper, however, is exclusively addressed to the question of structural importations from Vasconic into Proto-Celtic: the word-initial accent11, the two copulas, and the loss of the voiceless labial obstruent. A further structural feature, vigesimal counting (i.e. by twenties) in Insular Celtic, is omitted here. It is undoubtedly of Vasconic origin. Welsh, Cornish, and Breton are vigesimal, suggesting that the same holds for the Common Brythonic of the 6th century AD. Yet vigesimality cannot count as Proto-Celtic because Old Irish writing was still entirely decimal (Greene 1992, 511). One may compare the following explanatory sketch of the subsequent development of counting in Irish: 9

E.g. Weinreich 1953; Thomason & Kaufman 1988; Winford 2003; Matras 2009; and summarized in Vennemann 2011, 218–25. 10 But cf. Russell (1990) on the etymology of velar suffixes in Celtic. 11 The details of the new rule vary in the case of prefixed verbs: Early Germanic places the accent on the first syllable of the simplex, i.e. the underived stem plus a sequence of derivational and flectional suffixes, possibly empty. In early Latin, the first syllable of the prefix receives the accent. The rules for Proto-Celtic are not so clear; the main rule for Old Irish is this: ‘The accent in [Old] Ir[ish] as a rule falls on the first syll[able]. […] The combination preverb (preverbs) + verb is normally accented on the first syllable of the second element (whether preverb or verb)’ (Lewis & Pedersen 1989, §99). The common denominator of the three systems is the mechanical determination of the accented syllable counting from the beginning of the word.

Vennemann CW3 19 iv 2016.indd 506

27/05/2016 19:56:00

vennemann

[ 507 ]

[In Middle Irish] the vigesimal system begins to be normal, although all the decads up to and including ‘ninety’ are still attested. [...] [In Classical Modern Irish] fiche ‘twenty’ plays an increasingly important part in the system. [...] All these shifts arise from a tension between the literary standard, which tried to preserve the decads, and the spoken language, which had undoubtedly gone over to the vigesimal system by this time; no modern Irish dialect has preserved any of the decads above ‘twenty’, nor is there any trace of them in Scotish Gaelic or Manx. (Greene 1992, 525, 530)12

The first argument: first-syllable word accent The argument I proposed in 1994a and 1995 in favour of the assumption of a Celtic shift to word-initial accent on a Vasconic substrate rests on the two observations, (1) that three and only three Indo-European sub-families—Germanic, Italic, and Celtic—changed from the inherited Proto-Indo-European system of variable accent to a system of fixed word-initial accent and (2) that renowned Vasconists have reconstructed a corresponding accent system for an earlier stage of Basque.   Ad (1): The shift to initial accent13 is best accounted for as a result of language contact. Since there is no language-internal explanation for the development of intial accent in these three languages, language contact with non-Indo-European languages has been held responsible for over a hundred years (since Feist 1913, 375). Salmons (1992) shows that prosodic properties are transferred with special ease from substrates to their superstrates, and he proposed that contact with Uralic initial-accent languages was responsible for the accentual change in the three languages. I have given arguments in Vennemann (1994a, §7.6) for the assumption that the substrate of this group of Indo-European initial-accent languages, Old European as reflected in the Old European toponymy, had initial word accent14; and in the same and several other articles, that this language was Vasconic, which 12 Cf. the brief overviews of European vigesimality in Vennemann (1998, 8-12 [2003, 598–602]; 2010a, 384f). 13 I have collected arguments for this shift once again in Vennemann (2010a, 395f). For Germanic and Italic the assumption of such a shift seems uncontested. For Celtic the case seems less clear. Not being a Celtologist, I gratefully accept the following brief caveat of John Koch (pers. comm. 25 July 2015): The clearest evidence for initial stress in Celtic is in Goidelic. Not everybody would agree with Schrijver that the whole branch went through such a stage. Isaac and I, for example, think that the PIE accent position (as in Vedic) survived for some time in Celtic. Patrizia De Bernardo also has ideas about stress position without initial stress being generalized in Celtic. It’s complicated, and of course the accent position could have changed more than once, as clearly was the case within the literate period. But initial stress generalized across Italo-Celtic would be a problem. The evidence Schrijver has for initial-stress in Gaulish and Brythonic is far less compelling and clear-cut than that for Goidelic. Even with such provisos what did happen most clearly in Goidelic may be significant in this connection, but an awareness of the less obliging Gaulish and Brythonic evidence should be acknowledged. 14 Schrijver (2015, 196) mentions the ‘Development of word-initial stress (Italic, Celtic; [...])’ in section ‘2.1. ItaloCeltic’ and writes, ‘fixing of initial stress in Germanic occurred after a specifically Germanic development, Verner’s Law, so is a separate development’. It is certainly a separate development (as may be true for the fixing of initial stress in Italic and Celtic, too). But this fact alone would not suffice to assume a different cause for one of the three essentially identical developments than for the other two. What I did not consider when suggesting an identical cause for the threefold accent shift is the possibility that we have to assume early

Vennemann CW3 19 iv 2016.indd 507

27/05/2016 19:56:00

[ 508 ]

xviii. celtic as vasconized indo-european?

is also the position taken in the present paper. Palaeo-Sardic appears to agree with Old European, for Blasco Ferrer (2015, 174)15 writes: Was den Akzent betrifft, so scheint er, aus zahlreichen Beispielen, die scheinbar ein älteres Stadium aufweisen, auf die erste Silbe der rekonstruierten Basis zu fallen: ós/ana, óv/ana, sd. órga, ír/ ul/i, s[ar]d[isch] g/úsp/inu. [As to the accent, it seems to have fallen on the first syllable of the reconstructed base, as numerous examples apparently reflecting an earlier stage show: ós/ana, óv/ana, Sardinian órga, ír/ ul/i, sd. g/úsp/inu.]

  Ad (2): As to the reconstruction of initial accent for an early stage of Basque, I cited and adopted Martinet (1964) and Michelena (1977) in Vennemann 1994a. Hualde (2003; 2007) has argued that all accentual systems observable in Basque (for which see Hualde 1999; 2003), including the initial accent systems reconstructed by Martinet and Michelena, have developed from an earlier stage at which the language lacked word accent and only possessed phrase-level prosody (Hualde 2007), a kind of system known from Modern French, though possibly with different contours. This result does not, of course, preclude the existence of a word-accentual system with initial accent at a still earlier stage. Indeed, the French system developed from the well-known Classical Latin accent rule, which itself replaced the wordinitial accent of Early Latin.   In favour of the assumption of a very early stage of word-initial accent I would like to point to a fact which I have not seen used in discussions of the prehistoric accentuation of Basque: it derives from the work of Larry Trask and has so far been used in Vennemann (2010a, 397) only in an abbreviated version. Consonant clusters of two (the maximum in traditional Basque) are extremely common, but only between the first and second syllable, not between the second and the third (cf. Trask 1997, 178; 2008, 20). In other words, trisyllabic simplex words of the structure (C)VCVCV and (C)VCCVCV are normal, e.g. +aRaNo (> arrano) ‘eagle’, itsaso ‘sea’, +anari (> ahari, aari) ‘ram’, +aiNala (> ainhara, enara etc.) ‘swallow’, +eztaRi (eztarri) ‘throat’; or they can be reconstructed, because often the last C is lost through weakening sound changes, e.g. +ardano (> ardao, ardo, ardu) ‘wine’. But there are very few examples of trisyllabic words ending with -VCCV.16 Beside the constraint that there may not be more than one consonant between the second and the third syllable, there is a second constraint disallowing this one consonant to be a plosive.17 contact of Germanic with Uralic, namely if it is true (cf. Koivulehto & Vennemann 1996) that Verner’s Law did not develop independently of Finnish Consonant Gradation. In this case it would be a plausible, though not cogent, assumption that the Germanic accent too shifted under Uralic influence. Since nothing further depends on this question in the present paper, I leave it open. Unfortunately, Schrijver himself does not talk about causes of the accent shift. 15 Cf. Blasco Ferrer (2010, 153f). 16 There are only two simplexes of more than three syllables, goroldio ‘moss’ and arerio ‘enemy’. Neither may be safely assigned to pre-Basque (Trask 2008, 20). 17 Trask (2008) mentions three exeptions to these constraints, ipurdi ‘buttocks’, izerdi ‘sweat’, and gorosti ‘holly’; we may add izurde ‘dolphin’. Trask says of his three words that the first two may be originally bimorphemic, the third, bimorphemic or non-native (ibid., 20). According to de Azkue (1984, s.v.), izurde is a compound of

Vennemann CW3 19 iv 2016.indd 508

27/05/2016 19:56:00

vennemann

[ 509 ]

  Both constraints can be easily understood if first-syllable accent is assumed for an early stage of prehistoric Basque. First, a coda consonant in the structure (C)VCCV-, i.e., with Basque syllabication, (C)VC.CV-, adds complexity to its syllable, the optimal syllable being universally an open syllable, CV; and ‘syllabic complexities are less disfavored in stressed syllables than in unstressed syllables’ (Vennemann 1988, 58). Second, strong intervocalic consonants are the less preferred the less rhythmically prominent (the less accented) their environment.   In Vennemann 1994b I reconstructed initial word accent for Old European, a view which finds support in the above reconstruction of initial word accent for an early stage of Basque; and I added in a final note: ‘Als prosodische Eigenschaft in die späteren Superstratsprachen Paläo-Italisch, Keltisch und Germanisch vererbt.’ [Bequeathed to the later superstrate languages Palaeo-Italic, Celtic, and Germanic.] Since Old European is Vasconic, it follows that the initial word accent of Celtic is a Vasconic substrate feature. The second argument: development of a second copula The second argument (proposed in Vennemann 2010a) rests on the observation that Celtic shares with Basque the grammatical property of distinguishing two copulas, i.e. two verbs translatable into present-day English as ‘to be’ but having different use conditions. The same phenomenon is also found in all western Romance languages. Verbs of this kind are now best known from Spanish where the two copulas are (the paradigms of ) ser and estar. But the grammatical property is shared to a greater or lesser extent by all western Romance languages, i.e. all Romance languages except Romanian. Celtic too preserves the feature in the British Isles. Consider for example the following original thesis: The two copular paradigms based on Latin esse and stare exist in Portuguese, Galician, Castilian, and Catalan as well as, to a lesser extent, in Italian and, even lesser, in Sicilian, not at all in Rumanian. They once existed in French. [...] The areas south of the Pyrenees, as well as north of the Pyrenees and the Alps, plus to a lesser extent northern Italy, are precisely the regions assumed to have been Vasconic before the Indo-Europeanization of Europe. The null hypothesis accounting for this distribution of the rise of a second copular paradigm—in north-western Romance and in Continental Celtic—is, on the basis of the theory of a once-Vasconic western and central Europe, that this particular way of syntactic ‘thinking’ in terms of two copulas was carried from the Vasconic substrate into those IndoEuropean superstrates in the process of langugage shifting. (Vennemann 2010a, 393)

The Irish even have formed new double-copula paradigms in their varieties of English (Irish English, Hiberno-English).18   Systematically distinguishing two copulas both of which are translated with ‘to be’ is not +iz- ‘water’ (only occurring in compounds) with urde ‘pig’, iz erdi is +iz- ‘water’ plus a derivational element + -erdi meaning ‘accouchement, delivery, birth’, hence literally ‘transudation’. 18 Cf. ‘§2.4. Two copulas in Irish English’ in Vennemann 2010a, which is based on material from Hickey (2007).

Vennemann CW3 19 iv 2016.indd 509

27/05/2016 19:56:01

[ 510 ]

xviii. celtic as vasconized indo-european?

an inherited Indo-European grammatical property.19 But it is a property of Basque, where the two copulas are nowadays (the paradigms of) izan and egon. Since Basque texts revealing such fine lexico-structural detail only date from the 16th century AD and later, it is unknown how old this feature is in Basque. But the principle of Occam’s razor teaches us that when a feature occurs both in Basque and in certain western Indo-European languages, but is not originally Indo-European, then it is originally Basque (Vasconic).20 The theory therefore says that this property was originally Vasconic and was carried (1) from Vasconic into Celtic, where it is preserved in the Celtic languages of the British Isles, and (2) from (Continental) Celtic into the (Continental) West Germanic languages, where it is residually reflected in the re-unified copular paradigm of German and Frisian, e.g. the b-forms in German (ich) bin, (du) bist vs. Gothic (ik) im, (þu) is, Old Norse (ek) em, ( þú) es(t). According to the theory, the property was also carried by the Celts to the British Isles where it was passed on to Old English (wesan vs. b{on) after the Anglo-Saxon invasions, perhaps merely strengthening the earlier West Germanic import.21 The two English copulas lived through the centuries of Old English but then merged into dialectally different single paradigms, such as Standard English be, am, are, is, was, were, and been. Here am, are, is, was, and were are reflexes of the inherited paradigm, with be and been remnants of the second paradigm which arose through Celtic influence.   The fact that traces of the second or b- paradigm are found in all West Germanic languages including the continental ones is evidence that Celtic first passed this property on to Germanic in Continental Europe. It follows that Celtic itself had possessed this feature while spoken in Continental Europe. Since the feature is not an inherited but an acquired one, the giving language being Vasconic according to the theory, it follows that very early language shifting from Vasconic to Celtic occurred in western central Europe before Germanic was carried from its northern homelands into the historical West Germanic territories, which harmonizes with the observation that Celtic is quite generally more strongly Vasconized than Germanic.22 19 Broderick (2015) cites Latin examples for the use of st\re ‘to stand’ where esse ‘to be’ would save the same semantic purpose, and suggests that Celtic developed its second copula under Latin influence. However, for a language to permit using a word for ‘stand’ with such ‘bleached’ semantics does not mean having a second copula. German too allows such abstract use of stehen ‘to stand’: unter Verdacht stehen ‘to be under suspicion’; Es steht zu befürchten for Es ist zu befürchten ‘There is (reason) to fear’; Wir stehen selbst enttäuscht und sehn betroffen / Den Vorhang zu und alle Fragen offen (Bertold Brecht, Der gute Mensch von Sezuan, with stehen for sind to avoid metrical stress clash) ‘We are disappointed ourselves and see with concern / The curtain closed and all questions open’. It does not follow from this that German has two copulas. Nor does it follow from Broderick’s (2015) Latin examples that Latin had two copulas; the two-copula syntax of Insular Celtic is not likely to be a Latin feature. I therefore see no reason to revise my 2010 Vasconic explanation cited above. 20 Occam’s razor is a principle of the philosophy of science which says: Pluralitas nunquam est ponenda sine necessitate. [No more factors may be used in an argument than are absolutely necessary.] 21 The question of whether the Continental West Germanic languages also had a fully fledged second paradigm or merely borrowed the b- forms into their single inherited copular paradigms is discussed controversially, cf. Schumacher (2007), Lutz (2009). What matters here is that both authors agree that the b- forms in the copular paradigms of all West Germanic dialects are owed to Celtic influence. A recent overview of the twocopula problem in Western Europe may be found in Trudgill (2011). 22 Cf. the section ‘Celtic from the East’ below.

Vennemann CW3 19 iv 2016.indd 510

27/05/2016 19:56:01

vennemann

[ 511 ]

  This then is a strong argument for the theory that western central Europe was Vasconic before it became Indo-European and, in particular, that Celtic was Vasconized in western central Europe. This is, in my opinion, an improvement in comparison with the ‘Celtic from the West’ theory. Both theories rest on the assumption that a number of specifically Celtic features were acquired by the Indo-European language of a group of western Indo-Europeans in contact with Vasconic (or Vasconic and Iberian). The difference is that the ‘Celtic from the West’ theory assumes this contact to have taken place in the ‘West’, the Vasco-Iberian areas around the Pyrenees, whereas my Vasconic Theory naturally assumes these contacts to have occurred in western central Europe. The ‘Celtic from the West’ theory therefore has to assume that speakers of Indo-European first moved westwards into the Vasco-Iberian areas around the Pyrenees where their language acquired the features that define it as Celtic, and that then this early Celtic spread eastward to pass on certain features, in particular the double-paradigm feature, to West Germanic.23 By contrast, the Vasconic Theory permits the assumption that the origin of Celtic by contact with Vasconic took place in western central Europe, so that contact with West Germanic resulted from the general prehistoric westward spread of Celtic. Clearly the latter account is simpler and thus suggests that the Vasconic Theory provides an improved and preferable version of the Celtic-by-Vasconization theory. The third argument: loss of the voiceless labial obstruent At the 2015 Atlantiar conference in Irun (Gipuzkoa, Basque Country)24, John Koch (cf. Koch 2015) suggested that the development of the Proto-Indo-European +p phoneme in Celtic and gaps in the inventory of labial consonants of Basque may not be independent of each other, a suggestion which also occurs in the first Celtic from the West volume: The fact that the non-Indo-European language of the Iberian inscriptions of the eastern Peninsula also lacks p raises the possibility that it was prolonged exposure to the Iberian sound system that led to the loss of Indo-European p in Celtic. Celtiberian and Tartessian are both p-less Celtic languages. They were also both situated alongside the territory of the Iberians with whom they came to share a writing system. [...] Across the Pyrenees, p-less Celtic was to find itself in contact with Italic and Germanic and other languages with p. It was here that Celtic developed, or restored, a sound system more like those of its centralEuropean neighbours: inherited k¯ came to be articulated as p, the defining criterion of the ‘P-Celtic’.   The Gaulish (and Goidelic and British) affinities of Tartessian contrast with the conservative and relatively isolated character of Celtiberian. But, if one views the Atlantic maritime routes as the primary avenue of Celticization and subsequent innovations within Celtic, this is hardly surprising. Land-locked and wedged between non-Indo-European 23

In the words of John Koch (pers. comm. 25 July 2015): ‘Some speakers of PIE (or Proto-Italo-Celtic) arrived in the west. Contact with Iberian and/or Vasconic there resulted in that PIE or Proto-Italo-Celtic losing *p (&c.) and thus becoming Proto-Celtic. That earliest Celtic then spread east to eastern Gaul and the Alpine area in later prehistory, much like Celtic continued to spread onward to the middle Danube, Thrace, and Galatia in historical times.’ 24 On 15 May at the Ficoba Exhibition Centre in Irun.

Vennemann CW3 19 iv 2016.indd 511

27/05/2016 19:56:01

[ 512 ]

xviii. celtic as vasconized indo-european?

Iberian, non-Indo-European Aquitanian/Palaeo-Basque, and the Pyrenees, Celtiberian was ideally placed to be insulated from innovations affecting the other Celtic languages. No one has taken the possibility of Celtic coming from Hispania to the other Celtic countries seriously since we stopped taking Lebar Gabála Érenn (the 11th-century Irish ‘Book of Invasions’) seriously, but it is now at least worth pausing to review what it is that we think we know that makes that impossible. (Koch 2010, 294f)

In the second Celtic from the West volume too this is a recurring theme: Tartessian and Celtiberian, both of which show numerous examples of loss of IndoEuropean *p, were spoken in territories directly adjacent to the country where Iberian inscriptions and place-names are found. (Koch 2013a, 7) The modifications of Indo-European *p—mostly weakenings and in most phonetic positions complete loss—have traditionally been given the status of signifying the branching off of Proto-Celtic from Proto-Indo-European. In several respects focusing on the fate of this particular sound is a less than ideal choice. Indo-European *p > Ø occurs also, independently, in Armenian. […] On the other hand, this traditional diagnostic holds the advantage of identifying a likely geographic context for the emergence of Proto-Celtic. McCone (1996, 43), Ballester (2004, 114–17), Koch (2009a, 21), and most recently Schrijver (2011)25 have suggested that contact with non-Indo-European languages of south-west Europe that lacked p, such as Iberian and probably also Aquitanian/Palaeo-Basque, could have affected the sound system of the Indo-European spoken in the same region, leading to the loss of Indo-European *p and hence the definitive emergence of Celtic. […] Schrijver plausibly argues that the key sociolinguistic situation was a cultural expansion, which resulted in numerous speakers of Iberian learning Indo-European as a second language. To put it in layman’s terms, Celtic is Indo-European spoken with an Iberian accent. (Koch 2013b, 123) In this period of cultural dynamism, coinciding with the Greek Dark Ages, the Iberianized dialect of Nuclear Indo-European (i.e. Proto-Celtic) was prestigious, and its distinctive features (including weakening of *p) spread within the network of the Atlantic Bronze Age (and possibly also to the western Urnfield area, ‘Rhin-Suisse-France orientale’). Contact with, and expansion into the territories of other non-Indo-European languages is likely to have been a further factor in the differentiation of Proto-Celtic—Aquitanian/Palaeo-Basque, an indigenous language of the region, and Palaeo-Berber across the Straits. (ibid., 137)

Even in the present volume we see the same view expressed: Before the Phoenicians reached south-western Europe, Indo-European had come into contact with a different non-Indo-European language or languages there, speech similar to the attested Iberian and Aquitanian/Palaeo-Basque. Some of the phonological innovations that define Celtic—such as the weakening and loss of *p—can be seen as the result of this contact. The sociolinguistic background for the emergence of Proto-Celtic was that a relatively large number speakers of non-Indo-European learned Indo-European as a second language in this region. (Koch this volume) 25

Cf. now Schrijver (2015, 200).

Vennemann CW3 19 iv 2016.indd 512

27/05/2016 19:56:01

vennemann

[ 513 ]

In a discussion the evening following the 2015 Alantiar conference26, Koch more specifically stated his belief that the Proto-Celtic developments took place when speakers of Iberian and Basque imperfectly learned to speak the language of the Indo-European newcomers.27 When I suggested that this may have occurred on ‘my’ Vasconic substrate of Europe north of the Alps, Koch agreed that he considered this a possibility.   One of the striking features of Proto-Celtic is that it does not possess a +p28, except perhaps immediately following an s.29 Several mechanisms were at work by which ProtoIndo-European +p was changed in Proto-Celtic: (1) before other plosives and before +s it was fricated and appeared as +x; (2) before liquids (+r, +l) it was voice-assimilated, +p > b; (3) before nasals it was weakened into a semivowel, w; (4) elsewhere it was homorganically fricated, +p > +f (Meier-Brügger 2010, §L 337), a change which also occurred in Germanic, there as part of the First Consonant Shift or Grimm’s Law.30 But then something unexpected 26 In the Jauregisarrea palace of Xabier Otero in Arraioz (Baztan Valley). 27 This interpretation is also found in the second Celtic from the West volume: These groups could not have talked to each other without someone learning a very different language. If circumstances arose in which many non-Indo-Europeans wanted or needed to speak to Indo-Europeans, this was more than a matter of adjusting dialect, but extensive bilingualism. In regions where that bilingualism led to language shift over time, the Indo-European spoken in its new territory would be expected to bear some stamp of substratum effects—usually phonetic and/or syntactic—from the language it replaced. (Koch 2013b, 116f.) Koch’s substrate explanation is challenged by Prósper (2014): Predictably, Koch […] puts down the Celtic fate of I[ndo-]E[uropean] /p/ to substrate effects, ‘such as unstable articulations of /p/’ among Aquitanians or Iberians who were second-language speakers of Proto-Indo-European or Italo-Celtic. But, considering these non-Indo-European languages are a sort of bottleneck, the homogeneous spread of /p/ loss all over Hispania and western Europe is quite astonishing, especially taking into account that the only substrate effect we would expect in this case is the identification of /p/ with the closest existent phoneme in Iberian or Aquitanian, which in all likelihood would be /b/. It stands to reason that we should find many more clues of linguistic interaction in this region of the continuum, that is to say more substrate effect, which in fact is nowhere to be found. (Prósper 2014, 474f ) All four arguments against Koch’s substrate explanation—(1) the ‘bottleneck’ position of Aquitanian and Iberian, (2) the presumed amazingness of ‘the homogeneous spread of /p/ loss all over Hispania and western Europe’, (3) the non-identification ‘of /p/ with the closest existent phoneme in Iberian or Aquitanian, which in all likelihood would be /b/’, and (4) the alleged scarceness of ‘clues of linguistic interaction in this region of the continuum, that is to say [of] more substrate effect’—become inapplicable in relation to the substrate explanation to be offered in the present article. 28 The chain *p > *φ (> h > zero) occurs in the recent list of defining features of Celtic in Schrijver (2015, 198), there attributed to McCone (1996, 37–65). 29 The only other Indo-European language in prehistoric times changing Proto-Indo-European +p (probably via +f) into h, which is partly lost, is Armenian (cf. Schmitt 1981, 56f., 74). Interestingly, substrate influences have been held responsible for the change (ibid., 74). Armenia shares a border with Georgia, and Georgian does not have an f (cf. the Internet site ‘Georgian language’; for Old Georgian in paticular Fähnrich 1994, 38), so that at least the loss of +f may be attributed to contact with Georgian. 30 The labial fricative is often represented as bilabial +ɸ (also written [φ]) rather than as labio-dental +f. Since this fricative is lost in Celtic, there can be no evidence for its precise phonetic nature. Prósper (2014, 468 n) writes: ‘Following the current scientific habit, I use for the P[roto-]C[eltic] result of I[ndo-]E[uropean] /p/ before aspiration and disappearance, in spite of the labiodental /f/ being less marked than the bilabial /ɸ/ and consequently universally occurring when there is only one labial fricative.’ Stephen Laker (pers.

Vennemann CW3 19 iv 2016.indd 513

27/05/2016 19:56:01

[ 514 ]

xviii. celtic as vasconized indo-european?

happened: Instead of stabilizing as f, as in Germanic where the Lautverschiebung f has remained intact for nearly two and a half thousand years, and also, for example, in Arabic, where f may be just as old, the Celtic +f was lost still within Proto-Celtic, probably with an intermediate stage of mere aspiration: +f > +h > zero.31 I cite the rules and examples from Lewis and Pedersen’s Comparative Celtic Grammar: I[ndo-]E[uropean] p (Lat[in], G[ree]k p, G[er]m[ani]c f, Arm[enian] h-, lost, -w-, S[ans] kr[it] p) was in Celt. treated differently from all other IE. explosives. It became throughout a spirant, first a bilabial f; an echo of this is found in the development of sp and in the w-diphthong resulting from op before n [...]. Before s [...] and t [...] f became /x/; elsewhere it became h, which disappeared entirely. The same applies to IE. ph. Initially: Ir[ish] athir ‘father’ G[aulish] Ateronius: Lat. pater Gk. πατήρ Goth. fadar Arm. hair Skr. pit\; MlIr. [Middle Irish?] iasc ‘fish’: Lat. piscis Goth. fisks; [...] Before consonants: W[elsh] rhyd ‘ford’ O[ld] Co[rnish] rid O[ld] Br[eton] rit G. Augusto-ritum : Lat. portus ‘harbour’ O[ld] H[igh] G[erman] furt ‘ford’; Ir. lán, lám [< +pl-, ‘full’]; Ir. lethan [< +pl-, ‘broad’]. (Lewis and Pedersen 1989, §29)32 (IE. p between vowels). -ap-: Ir. caera ‘sheep’ M[oder]n Ir. caora W. caer-iwrch ‘roebuck’: *qapero-, cf. Lat. caper ‘he-goat’ Gk. κάπρος ‘wild-boar’ [...] -ep- before the final syllable: [...] Ir. niae ‘nephew’, g[enitive] niath W. nai OCo. noi Br[eton] ni, pl[ural] nied : Lat. nep}s ‘grandson’ OHG. nefo ‘nephew’ (W. nai goes back to *nep}t-, -}- in final syll. > ū > ¼). [...]). (Lewis & Pedersen 1989, §30)

  It may be argued that p > f > h > zero, or p > ɸ > h > zero, is a natural chain of sound change. This may be true. But f is nonetheless a common and stable consonant. In a frequency hierarchy of the most common eight fricatives in the languages of the world, labiodental f occupies the third position after s and š (Maddieson 1984, 50). One hundred and thirty-five of the 317 languages in Maddieson’s sample have an f phoneme (ibid. 1984, 227), i.e. 42.6%. Languages combining—as may be assumed for early Celtic—exactly the two fricatives s and f number 25.8% of the world’s languages (ibid. 1984, 53). Thus reference to naturalness does not answer the question as to why the newly developed Proto-Celtic +f was lost, and why so quickly, whilst in other languages the same fricative holds out for millennia.33 comm. 4 July 2015) argues in favor of bilabial ɸ rather than f because of the contextual variation mentioned above and because of the greater acoustic similarity of [ɸ] to [h] which would facilitate its replacement by the latter. In what follows, I will write f as a cover symbol and leave the phonetic interpretation as either bilabial [ɸ]/[φ] or labiodental [f] to the specialists of the languages involved. According to Eska (2013), the bilabial fricative /ɸ/ was preserved intervovalically in Proto-Celtic. 31 The evidence for the intermediate stage with h does not appear to be strong. The example that everyone cites is the name of the Hercynian Forest, with +perkuni\ > herkunia (> erkunia), base +perkʷu- ‘oak’ (cp. Latin quercus ‘oak’, Old High German Fergunna ‘Erzgebirge / Ore Mountains’); cf. Pokorny (1989, s.v. perkwu-s ‘Eiche’ [‘oak’] I.822f). 32 For further examples of p in contact with consonants cf. Lewis and Pedersen (1989, §§31–3). 33 Patrizia Noel (pers. comm. 5 July 2015) has kindly pointed out that all exemplification for the loss of f (or ɸ) cited in the present article (mostly from the work of Rohlfs and Wolf, see below) is taken from Romance languages and that this is not sufficient to establish the claim that the change is a natural one. Since the position elaborated in the present section is that the loss of f in Romance is not a sound change but a case of negative importation in language contact with Vasconic languages, I have indeed not cited a single case in support of this surmise. Thus the burden of proof that the loss of f (or ɸ) in Romance is probable as an internal development is entirely on the shoulders of those Romanists who advocate that position.

Vennemann CW3 19 iv 2016.indd 514

27/05/2016 19:56:01

vennemann

[ 515 ]

  There is, however, a straightforward solution to the problem of the change of f into h. Such rapid changes are typical effects of language shifting, the acquisition of a newly introduced language by an indigenous population with a structurally different language. The indigenous population of western central Europe, into which the pre-Celtic IndoEuropeans invaded, spoke such a language, according to the Vasconic Theory. Pre-Basque, according to the reconstruction of Michelena (1977, §13.2), adopted by Trask (1997, 167), had a consonant inventory without f. The consonants which it did have, all of them subject to severe occurrence constraints, were: +/b d g, p t k; z tz, s ts; n N, l L, r R; h/.34 The Vasconic Theory says that the Old European language of the Vascons of central and western Europe was essentially the same as Pre-Basque, hence, in particular, that it had no +f but did have an +h. Therefore, in the shifting process, the Vascons learning Celtic did not acquire the Proto-Celtic +f in words containing it; the best they could do was substitute their +h, [h] being phonetically the most similar speech sound to [f] (or [ɸ]) within the Pre-Basque inventory. In positions where +h would have been uncomfortable owing to the restrictions of their own language, they may have omitted the +f of the Celtic words they were acquiring entirely, except where exingencies of understandability required a different approach.35   As can be seen, I treat the problem of the loss of the voiceless labial consonant of ProtoCeltic on the level of the fricative. This seems adequate in view of the following statement: The case of the phoneme /f/, irrespective of whether a bilabial or labiodental realization is assumed, does not cause much doubt. It is certain that its introduction into Basque is a relatively recent event. (author’s translation of Spanish original: Michelena 1990, §13.2)

But it seems worthy of consideration whether the problems might have to be studied on the level of the plosive, p. H. Schuchardt pointed out as early as 1887 the rarity of Basque p-, except in loanwords, even though his conclusions should nowadays be modified, precisely because of only dealing with the initial position. Much more important is the rarity of -p-, observed by Martinet [1964, 381, n. 18], between vowels and between r and a vowel. (author’s translation of Spanish original: Michelena 1990, §13.1)

But since there also exist Basque words with p whose great age is not in doubt (Michelena 1990, §13.1), I will not pursue this alternative here.   Thus, the replacement of f by h, with partial loss of h, in the Vasconic-to-Celtic language shifting is what the general theory of language contact would lead us to expect. But we are more fortunate than that, as shown in the following three sub-sections. There exist two widely accepted exact parallels of the above scenario, plus one less well-elaborated 34 /N L R/ represent fortis (perhaps geminate) variants of /n l r/. As to /h/, Trask (1997, 157) writes: Pre-Basque had a phonetic aspiration which was both frequent and prominent. The occurrence of that aspiration within words was subject to severe constraints which are largely, but not entirely, understood. In the Aquitanian materials at our disposal, the aspiration is regularly written (as H) in approximately the same positions in which it occurs today in those dialects of Basque which retain the aspiration. Concerning Proto-Basque +p, the matter is not so clear. See the quotations that follow above. 35 See note 41 below.

Vennemann CW3 19 iv 2016.indd 515

27/05/2016 19:56:01

[ 516 ]

xviii. celtic as vasconized indo-european?

but nonetheless suggestive case, all three of them involving an Indo-European language on a Vasconic substrate. The Indo-European language is in all three cases the Latin of the western Roman Empire. The Vasconic substrate is in the first, oft-discussed, case the ProtoBasque south of the Pyrenees, in the second case the Aquitanian Basque of Vasconia, i.e. Gascony, north of the Pyrenees, and in the third case the Vasconic substrate of a Sardinian dialect group. The resulting language is in the first case Castilian (whence modern Standard Spanish), in the second case Gascon (which some scholars see as a dialect of Occitan, others as a separate Romance language), and in the third case the Barbagian dialects of Sardinia.

f > h in Castilian The change of inherited Latin f into h, with subsequent loss (except in dialects), is one of the distinguishing features of Castilian (hence of modern Standard Spanish) among the major Romance languages. It is treated in all historical accounts of the language.   The internet site ‘Spanish phonetics and phonology: The “f > h”change’ of Prof. Christopher J. Pountain is the most accessible source today, a concise and precise scholarly account, containing extensive word lists both of ‘cases’ and of ‘exceptions’36: Latin initial /f/ is progressively weakened and eventually lost in standard Castilian in popular words, unless the /f/ is followed by a liquid consonant or a back semivowel ([w]), or in some instances by a front semivowel ([j]). Weakening also sometimes takes place intervocalically. This eventually produces a phoneme split, creating for a time /h/ in addition to /f/, until /h/ is eventually lost. Thus Lat. furnu(m) /furnu/ > Sp. horno /oɾno/ [‘oven’] Lat. faba /faba/ > Sp. haba /aba/ [‘bean’] Lat. ferru(m) /ferru/ > Sp. hierro /iero/ [‘iron’] Lat. subf~m\re /subfu:ma:re/ > Sp. sahumar /saumaR/ [‘to perfume’] Lat. d{f{nsa /de:fe:nsa/ > Sp. dehesa /deesa/ [‘pasture’]   but Lat. fr\tre(m) /fra:tre/ > Sp. fraile /fRaile/ [‘friar’] (before a liquid consonant) Lat. f}cu(m) /foku/ > Sp. fuego /fuego/ [‘fire’] (before [w]) Lat. festa /festa/ > Sp. fiesta /fiesta/ [‘party’] (before [j]) Lat. facile(m) /fakile/ > Sp. fácil /faθil/ [‘easy’] (learned borrowing)

The author adds, ‘We cannot be sure of the phonetic detail of the change. [...] A possible general scenario is: [f ] > [ɸ] > [h] > Ø’.37 He goes on to break this down into several

36 For a similar account cf. Pountain (2001, 271–4). 37 Cf. Lauberg (1956, I. §59): ‘Die stimmlose Variante [...] [ɸ] [...] ist vielleicht als Zwischenstufe im Wandel von [f] > [h] im Span[ischen] anzusetzen: fil̯iu > *[ ɸil̯iu] > a[lt]span. hijo [hižo] (n[eu]span. [iχo]).’ [The voiceless variant [ɸ] may have to be posited as an intermediate stage in the change f > h in Spanish.]

Vennemann CW3 19 iv 2016.indd 516

27/05/2016 19:56:01

vennemann

[ 517 ]

context-sensitive steps. He then asks the question ‘Why did the change occur?’ and offers several answers, first among them ‘The substrate hypothesis: Latin [f] was modified because Basque learners of Latin found difficulty in pronouncing it. [...] Basque in the early Middle Ages had no /f/ phoneme.’ He lists four counterarguments: (1) ‘/f/ in Latin loanwords in Basque is not adapted to [ɸ] or [h] but to a bilabial plosive: Lat. festa /festa/ > Basque besta /besta/, Lat. f\gu(m) /fa:gu/ > Basque pago /pago/.’ (2) ‘Later Basque speakers do not seem to have problems in borrowing Castilian words beginning with /f/.’ (3) ‘Why does /f/ survive in Castilian at all?’ (4) ‘The movement of [f] to [h] is not an unusual change: it is evidenced in a number of southern Italian dialects (Rohlfs 1966, 206–7) and is widespread in Andalusia and Latin America. It can easily be seen as an example of articulatory weakening which could have happened quite independently and so needs no appeal to substrate influence at all.’

All of these objections allow for straightforward answers: Ad (1) Borrowing is a different process than language shifting; there is no reason to expect the same results. Ad (2) After centuries of exposure (Spanish dominance and widespread bilingualism) it becomes natural to adopt words with shapes ever closer to those in the giving language.38 Ad (3) Here the author himself provides an answer: ‘French clerics brought to Castile a new way of reading Latin aloud [...] through which the labiodental [f] may have become familiar. Such circumstances might have favoured the adoption of [f] more generally, and would plausibly explain the many ‘exceptions’ to the f > h rule.’39 Ad (4) A change requires an explanation, even if it is not unusual. After all, most Italian varieties preserve f, and English shows no sign of weakening its more than two thousand year old f. Even in modern Castilian f is stable. But the author himself offers a decisive argument in favor of the substrate explanation: ‘There is a similar f > h change in Gascon, a striking coincidence (both Castile and Gascony bordered the Basque Country).’   A very thorough handbook description of the change and its distribution is contained in Lloyd (1987, 212–23). Lloyd in particular focusses the ‘Basque-Romance bilinguism’ 38 The process of gradual approximation to the phonology of another language through continued exposure can be observed in other contact situations, too. For example, northern speakers of Standard German have learned to pronounce [ʒ] in French loanwords like Journalist [ʒʊrnaˈlist] and Garage [gaˈra:ʒə], whilst speakers learning Standard German on the basis of Bavarian dialects still substitute their native [ʃ]. Broadcasters on German television again prove this point by saying [dʒʊrnaˈlist], showing off their knowledge of English. 39 The largest group of Castilian words beginning with f in Pountain’s ‘Appendix’ consists of ‘Latin learned [words]’, e.g. (‘popular doublets in brackets’): fábrica ‘factory’, fábula ‘fable’ (habla ‘speech’), facción ‘faction’ (hacer ‘to do’), faceta ‘facet’ and facial ‘facial’ (haz ‘face’), férvido ‘fervid’ (hervir ‘to boil’), fingir ‘to feign’ (heñir ‘to knead’), fortuna ‘fortune’, fundir ‘to melt’ (hundir ‘to sink, destroy’).

Vennemann CW3 19 iv 2016.indd 517

27/05/2016 19:56:01

[ 518 ]

xviii. celtic as vasconized indo-european?

explanation (ibid., 218–20) and carefully weighs the evidence for and against both this substratum approach and the internal causation kind of approach (ibid., 220–4). As to the former, he reaches the following conclusion: Even if the change were largely an internal one, we would still have to account for the fact that both early Castilian and Gascon were the Romance dialects in this area that adopted the innovation first and generalized it. We could claim that it was simply a coincidence, but after all, there is no reason to assume that internal structural developments cannot work in tandem with inter-language influences (Weinrich 1958, 104). Given that large numbers of Basque speakers participated in the organization of the emerging county of Castile in its earliest days and that they did influence the development of the voiced sibilants, as we shall see later40, it is entirely plausible that they also helped to favor the innovation of /f/ > /h/. (Lloyd 1987, 222)

  Bollée and Neumann-Holzschuh (2003, 16f.) evasively treat the change of f into h in Castilian (and in Gascon) as ‘ein Musterbeispiel für auseinander gehende Meinungen und eine kaum mehr überschaubare Diskussion’ [a model case of diverging opinions and a nearly unmanageable vast disussion]. They end their own brief account with the ‘denkbar’ [thinkable] compromise that ‘eine interne Entwicklung durch den Sprachkontakt verstärkt worden ist’ [an internal development was enforced by language contact], i.e. essentially the same conclusion as in the above quotation from Lloyd (1987), to which they refer.41 40 The reference here is to the devoicing of the voiced sibilants, a Castilian feature unique in western Romance and attributed to early contact with Basque, Basque having only voiceless sibilants, /s ś š/, i.e. exactly the same as retained in early Castilian (modern Castilian /θ ś x/); cf. Lloyd (1987, 267–73). 41 Readers who would like to know more about counter-arguments against the substrate explanation of the f > h change in Castilian are referred to Trask (1997, 425–8). Several of Trask’s arguments are not strictly linguistic but historiographic in nature. A linguistic one is argument no. 2: Though word-initial Latin [f-] became [h-] in most cases, it failed to do so before the diphthong /ue/ and before /r/, where [f-] remained: hence, for example, Latin facere yields Castilian hacer ‘do’; Latin ficu yields Castilian higo ‘fig’; and Latin fornu yields Castilian horno ‘oven’; while Latin forte yields Castilian fuerte ‘strong’; Latin fonte yields Castilian fuente ‘spring’; and Latin fronte yields Castilian frente ‘forehead’. Proponents of the substrate view are therefore seemingly obliged to argue that the Basques could pronounce [f-] without difficulty before /ue/ or /r/, but not otherwise, which is absurd. (Trask 1997, 425) But what would be Trask’s own proposal? Language internal weakening and loss of f? That would be equally absurd (to employ Trask’s favourite valuating expression): Word-initial pre-sonorant consonant weakening and loss begins in the least favoured environment, that is before the strongest sonorant, and works its way down on the strength scale (Nasals > Liquids > Semivowels > Vowels). For example, in English the loss of h- occurred before consonants in the standard language, and there plus before vowels in Cockney; nowhere was h- lost before vowels but left intact before consonants. English k- and g- were lost before the strongest consonant, n, and are under attack before l; before r and before vowels they are completely safe. The Germanic loss of f- so far occurred only before n (cf. Proto-Germanic +fneosan > German niesen (in English > sneeze, with salvatory strengthening of f- into s-) but has not yet affected f- before the weaker consonants and before vowels (cf. Lutz 1991, 226f, 234–44). In the light of these changes a language-internal weakening and loss of f- before vowels but not before consonants would be very surprising indeed. This argument leaves language contact as the only viable alternative, as long as no universals of change in language contact rule out the substrate explanation of the change and loss of f- in Castilian. In my view the correct account is the following: Basque speakers, not having an f in their own language, tended to substitute their h as the auditorily most similar speech sound of their own repertory. This worked well before vowels but not before stronger

Vennemann CW3 19 iv 2016.indd 518

27/05/2016 19:56:01

vennemann

[ 519 ]

f > h in Gascon Rohlfs (1977) writes about ‘Le passage de f en h’: One of the most characteristic traits of the Gascon area is the change of f into h: hèsto ‘fête’ [party], hìu ‘fil’ [thread], hour ‘four’ [oven], hèt ‘fait’ [fact], hàdo ‘fée’ [fairy], hàure ‘forgeron’ [blacksmith] < fabrum.42 [...] In some regions of Gascony the aspiration is quite weak (e.g. in Bordeaux) and is often lost entirely. Complete loss of the aspiration tends to occur most of all in the Luchon Valley, in the Upper Garonne Valley, in the Val d’Aran, less regularly in the Bethmale Valley and in the Upper Lez Valley: ourm¼go [...] ‘fourmi’ [ant] < formicam, awo ‘fève’ [bean] < fabam [etc.].43 (author’s translation of French original; Rohlfs 1977, §460) The change of f into h is not restricted to the initial position. We also find it in the middle of words, intervocalically and even in contact with consonants: ahamàt ‘affamé’ [starving], ahounà ‘enfoncer’ [to drive in], ahà ou ahè ‘affaire’ [matter], ahourcà ‘enfourcher’ [to mount], ahrountà ‘affronter’ [to face], dehéne ‘défendre’ [to defend] [...], gahà ‘saisir’ [to seize] < *gaffare, bouhà ‘souffler’ [to blow] < *buffare, gouhì [...] ‘mouiller’ [to wet] (cp. Modern Provençal gofe ‘tout trempé’ [soaking wet]), auherì ‘offrir’ [to offer], couhessà ‘confesser’ [to confess], couhì ‘borne de limite’ [border stone] [...] < confine, inhèr ou ihèr ‘enfer’ [hell]. Landes dialect ehlàt or ihlàt [...], Béarn dialect islàt or eslàt [...], ellàt [...] ‘enflé’ [swollen], Bigorre prehoùn [...], Landes prouhoun [...] ‘profond’ [deep]. Couflens (Ariège) is pronounced Couhlens in Gascon. (author’s translation of French original; Rohlfs 1977, §463)

Rohlfs compares the Gascon change of f into h to that in Castilian: The change of f into h is one of the most remarkable features relating Gascon to the

42

43

speech sounds where h was prohibited in Basque by the constraint *hC (i.e. no h before consonants). Therefore they made a special effort to reproduce the f- before consonants in order to remain understandable to the superstrate speakers. Learning a supertrate language is a compromise between ease and intelligibility. Within their native speech community no special precautions were necessary; there the Basques simply dropped the f- before consonants: firmu- > +frimu- → +rime > errime ‘firm, strong’, flore → lore 'flower’, flamma → lama ‘flame’ (cf. Agud & Tovar 1991; 1995.) Here Rohlfs adds a footnote: This articulation must be rather old, even though the medieval texts almost always write f. We are obviously dealing here with a traditional orthography explainable as the influence of Latin and of literary Provençal. In the Livre Noir de Dax one finds alongside fems, finar, ferir, for, forque, even faut ‘haut’ [high]. Other texts offer fala ‘halle’ [hall], fasta ‘hâte’ [haste], fuche ‘huche’ [chest]: The medieval scribes mechanically replace Gascon h by f, this h having the reputation of being especially vulgar. (author’s translation of French original) Evidently the scribes, in their desire to stay clear of this vulgarism, were applying an orthographic inverse rule ⟨h⟩ → ⟨f⟩ of the original language change f > h. It is typical for inverse rules to be overgeneralized (cf. Vennemann 1972). Here too Rohlfs adds examples of the corresponding inverse rule, zero → ⟨h⟩, abducted from the language change h > zero: The vacillation between hèsto and èsto ‘fête’ [party] sometimes has the effect that even words are aspirated that never had an h. Corominas [1931, 56] records hésca ‘amadou’ [tinder] < escam for the Val d’Aran. At Fos [...] one can hear halo ‘aile’ [wing], hòbro ‘œuvre’ [work], hescoùbo ‘balai’ [broom] < scopam. (author’s translation of French original)

Vennemann CW3 19 iv 2016.indd 519

27/05/2016 19:56:01

[ 520 ]

xviii. celtic as vasconized indo-european?

consonant system of Castilian. [...] Compared to the situation in Spain (Castilian), the Gascon change of f into h is more general. (author’s translation of French original; Rohlfs 1977, §§460, 461)

As to the reason for the change of f into h in Gascon, the ancient Vasconia, and in Castilian, Rohlfs says the following: Most scholars have spoken in favor of a pre-Roman substrate. It was entirely natural to think of an influence of Basque in a period of Basco-Roman bilingualism. Indeed Basque did not possess the consonant f, at least not in its inherited inventory. […] This hypothesis is strengthened by the fact that the letter f does not occur at all in the preRoman inscriptions of Aquitaine (Michelena 1954, 448). […] Even though we do not have any absolute proof for the view that the change of f into h is a survival or an effect of the ancient Iberian pronunciation, the fact that this articulation has come into existence in two areas which are in intimate contact with the Basque Country does lend a certain likelihood to this hypothesis. (author’s translation of French original; Rohlfs 1977, §464)

That this view is still entertained is evident from a comparison with the Internet site ‘Gascon language’: The language spoken in Gascony before Roman rule was part of the Basque dialectal continuum […]. There is a proven Basque substrate in the development of Gascon. This explains some of the major differences that exist between Gascon and other Occitan dialects. A typically Gascon feature that may arise from this substrate is the change from f to h. […]. Although some linguists deny the plausibility of the Basque substrate theory, it is widely assumed that Basque […] is the underlying language spreading around the Pyrenees onto the banks of the Garonne River, maybe as far east as the Mediterranean in Roman times.

The replacement of f by h both in Castilian and in Gascon is also discussed, together with other Basque substrate influences, by Tagliavini (1998, 109–11).

The aversion to articulating f in Barbagian Max Leopold Wagner writes in his book about the Sardinian language: The dialects of Barbagia are characterized by a number of strange phenomena: […] the aversion against the articulation of f- […; …] a dislike for the articulation of f is a characteristic trait of the Pyrenean dialects. […] Since these phenomena are found in the most archaic area of the island, which has also preserved numerous pre-Roman words, they may be considered surviving traits of Palaeo-Sardic. (author’s translation of Italian original; Wagner 1980, 314f)44

Indeed Barbagia is a typical retreat area: The area is full of hard hills, and there is little human presence. Barbagia is one of the least 44 Much information about the history of research in Sardinian linguistics, including sections on Palaeo-Sardic, may be found in Blasco Ferrer (2002).

Vennemann CW3 19 iv 2016.indd 520

27/05/2016 19:56:01

[ 521 ]

vennemann

populated areas in Europe, which has allowed Barbagia to preserve its cultural and natural treasures. (Internet site ‘Barbagia’)

It would therefore be quite natural to find structural details of a Vasconic substrate preserved in the Barbagian dialects45, even those that have been lost in all other Sardinian dialects.   The situation concerning f is succinctly described at the beginning of the chapter ‘La constrictive labio-dentale sourde [f]’ of Contini (1987a): Once again it is the central-eastern region which presents the most interesting situation. Its dialects differ from all the other dialects of the island [of Sardinia] by the absence of the voiceless fricative [f] word-initially.



central-eastern region ẹ̀mina ọ̀ku

other regions fẹ̀m(m)ina fọ̀ɣu

‘woman’ ‘fire’

This trait is characteristic of the District of Bitti, of Baronia, of the Barbagia of Ollolai and of Teti ..., in the Barbagia of Olisai (Mandrolisai), where however the labio-dental survives in certain words. (author’s translation of French original; Contini 1987a, 217)

Contini also raises the question of the dating of the change (1987a, 333–6) and presents a map for the change based on the occurrence of ìku (< fìku ‘fig’) (Contini 1987b, map 44).46 Examples for the change of f > h and for the total loss of f (probably via h) can also be found in Wolf (199 , 34, 37, 39, 45): hidza, idza, idʒa ← Lat. filia ‘daughter’ oddaʔe ← cadastral form foddache (toponym) trihodzu, triodzu, θiriodzu ← Lat. trifolium ‘clover’47

With regard to the third of these examples, Wolf (1998, 45f) writes: Evidently intervocalic -f- has changed into -h- in the Ovodda dialect and has vanished in the other dialects, exactly as at the beginning of the word. [...] We think therefore that both the initial and the intervocalic f changed into the aspirate h in Ovodda but was lost in the other Barbagian (and Baronian48) villages. (author’s translation of Italian original)

In his 1992 book, Wolf offers a one and a half page list of appellatives which begin with f- in Italian and Vulgar Latin but with h- in the dialect of Ovodda. There the change appears to be recent (see below), but the examples are nevertheless representative also for all the other Barbagian dialects having the ‘aversion’, except that the latter also drop the h-. The following is a selection: 45 Cf. the Vasconic family tree in the section ‘Celtic and Vasconic’ above. 46 There is also a map ‘F > [h] > Ø’ in Blasco Ferrer (2010, 234). 47 Paulis (1992, §119 “I nomi dei ‹trifogli›”) offers trióddzu for Bitti and Lollove, trióǧǧu for Oliena and Urzulei, and triódzu for Illorai. 48 Baronia is the region north-east of Barbagia.

Vennemann CW3 19 iv 2016.indd 521

27/05/2016 19:56:01

[ 522 ]

xviii. celtic as vasconized indo-european?

Barbagian haba háʔere halle hémina herru hesta híʔu hidzu híliʔe hoʔu hodza horru huíre humu huntána

Italian fava fare falce (donna) ferro festa fico figlio felce fuoco foglia forno fuggire fumo (sorgente)

Vulgar Latin < faba ‘bean’ < facere ‘to make’ < falce ‘scythe’ < femina ‘woman’ < ferru ‘iron’ < festa ‘party’ < ficu ‘fig’ < filiu ‘son’ < filice ‘fern’ < focu ‘fire(-place)’ < folia ‘leaves’ < fornu ‘oven’ < fugire ‘to flee’ < fumu ‘smoke’ < fontana ‘fountain’

In the following list of toponyms of Ovedda the cadastrial and map-making forms regularly show f where h is pronounced (ibid., 46). birihai ← Birifai galihai ← Galifai su hiʔu ← Sa Figu hiliddai ←Filiddai sa hiliʔossa ← Sa Filigosa hiloleri ← Filoleri hiuonele ← Finonele hoddateddi ← Foddateddi hoddis ← Fodis honadeo ← Fonadeo sos horros ← Sos Forros sos horreddos ← Sos Forreddos huntana aripa ← Funtana Aripa huntanedda ← Funtanedda brunku sa huʔilessa ← Bruncu Foghilessa huradu ← Furadu istehune ← Istefune orohole ← Orofole predas hittas ← Pedras Fittas tsolovehune ← Zolovefume

  Wolf (1998, 46) writes, ‘Per Ovodda ... il passaggio di f- a h- è recente’ [For Ovodda the change of f into h is recent], with reference to Wolf (1992, 15–26, i.e. to chapter ‘I. Mutamento fonetico a Ovodda: f > h’). This is not the place to discuss the age of the loss of f in the other Barbagian dialects. The existence of written forms with f does not mean that

Vennemann CW3 19 iv 2016.indd 522

27/05/2016 19:56:01

vennemann

[ 523 ]

the f was still pronounced as such at the time of their recording, as the above 20th century examples show. Therefore also an f in a medieval piece of writing need not reflect an actual pronunciation. Wolf continues after the above name list: [In other dialects, in which the (-)f- has disappeared completely, it is very likely that the hiatus of vowels in a certain number of toponyms is a result of the phenomenon mentioned. We have proof of this for quite a lot of toponyms, situated outside our nine communities49 but always inside the ‘area of the aversion to f’ (Wagner), thanks to medieval testimony. The following are attested in the 14th century:   Gorefa, Olefa/Ulifai, Ulosuffe/Ulusofe/Ulusufe, Unifa/Unifai/Unifay, Uniferi, Urdofe.   Half of these toponyms are still known with the names of Garofai (community of Bitti), Onifai and Oniferi according to the official spelling, but the actual pronunciation is goreai, uniai/oniai and unieri/onieri, without traces of the original, and probably etymological, -f-. I therefore consider it legitimate to reconstruct the larger part of the toponyms in -iai and -eai as terminating in -ifai and -efai, and also -oe/-ue as -ofe/-ufe, and also, namely in view of Oniferi > unieri, to take into consideration, despite the ‘Italian’ appearance of pilieri [...], an ancient *piliferi, supported by pili. (author’s translation of Italian original; Wolf 1998, 46f.)

  There is thus much and unequivocal evidence for the change f > (h) > zero in Barbagia, i.e. a constraint (an ‘aversion’) working against f. Exactly how long this constraint has been in effect may be impossible to determine. But from what is known about the persistence of phonetic constraints in other language shift situations, it is not surprising to find an ‘avversione per l’articolazione dell’ f-’ as a Vasconism in effect even in the most recent occurrence of the f > h change, as in Ovodda. Since a Vasconic substrate has been definitely established for Sardinia in the cited recent work of Eduardo Blasco Ferrer, it appears difficult not to see in the Barbagian loss of f the same kind of substrate effect as in Castilian and Gascon. This seems also to be the view of Blasco Ferrer: All that remains to be discussed is the extended coexistence of [f] and [h] in the Barbagian dialects. Wolf has clearly established that the final demise of the aspirate at Fonni, in central Barbagia, must have happened between 1904 and 1927, according to Wagner’s own notes. Moreover, the land registers prepared at the end of the 19th century still indicated an intial [f] for toponyms which today do not show any trace of it, and also the meagre medieval documentation does not seem to reveal any loss of the consonant (as, e.g., for Fonni < Fonnius, -¼, today Onne, but attested as Fonne in the 14th and 15th centuries). As in the case of Spanish, we will have to assume a situation of diglossia lasting for centuries after the advent of the Romans: The Latin [f] will have been maintained by people alphabetized together with its aspiration, next to the larger part of speakers belonging to the popular classes, and the few occasions of mediaeval or modern recording will have favoured exactly the first cultured realization, which was moreover connected to the regular maintenance of the sound in Núoro and the neighboring territories. The concept of latency of substrate effects seems, in conclusion, entirely connatural to the Barbagian and Baronian situation. (author’s translation of Italian original; Blasco Ferrer 2010, 155) 49 Cf. the title of Wolf ’s 1998 book.

Vennemann CW3 19 iv 2016.indd 523

27/05/2016 19:56:01

[ 524 ]

xviii. celtic as vasconized indo-european?

  This ‘Abweisung des Phonems f ’ [rejection of the phoneme f] in the dialects of Barbagia is also cited, with ascription to Wagner (1980), in Tagliavini (1998, 93), together with the vowel prosthesis of words beginning with r-. This constraint too is typical of Basque (and Gascon) and was likewise already exemplified (with erríu, arríu for riu, rio ‘river’) and identified as a «propensione nelle lingue pirenaiche” [tendency in the Pyrenean languages] by Wagner (1980, 315). For a recent mention of the loss of f see Mensching and Remberger (forthcoming, §16.2.2, with reference to Virdis 1988, 908): ‘In some Nuorese sub-dialects, word-initial f- is deleted in absolute initial position.’ A recent survey of Palaeo-Sardic begins its Conclusion with the following words: Palaeo-Sardic, which it is possible to reconstruct from the toponyms of Barbagia, had a consonant system characterized by remarkable structural resemblances with the consonantism of (pre-)Proto-Basque and of Iberian. (author’s translation of Italian original; Paulis 2008, 46f)

While this is welcome support for the theory of the loss of f in Barbagian dialects on a Vasconic substrate adopted in the present paper, Paulis’ continuation sounds considerably less positive and calls the expectation of an actual Vasconic substrate of Palaeo-Sardic into question: On the other hand, Basque can explain only 10% of the 120 probable lexical Palaeo-Sardic relics preserved in the modern dialects and is not of great help in the interpretation of the very rich toponomastic heritage of Palaeo-Sardic origin. From this point of view, the relationship between Palaeo-Sardic and Basque is strangely analogous to be found between Iberian and Basque: the great phonological similarity of the two languages stands in contrast with the scarcity of the lexical correspondences. (author’s translation of Italian original; ibid., 47)

As it happens, such doubts have been shown to be unfounded by the recent research of Blasco Ferrer (2015): both the compositional and agglutinative word structure of the PalaeoSardic toponyms and the lexical and grammatical elements combined in these structures prove the common descent of Palaeo-Sardic with Basque as well as with the language of the Old European Toponymy.   This genetic relatedness is most directly evident from the roots within the toponyms. For example, Blasco Ferrer’s Palaeo-Sardic roots ard(i) ‘sheep’, bid(e)- ‘way’, is- ‘(body of) water’, (i)turr(i)- ‘fountain’, mand(o)- ‘donkey’, mend(i)- ‘mountain’, ol(a)- ‘hut’, os(a)- ‘(mouth of a) river’, ur- ‘water’ (2015, §2.7) all occur also in Basque (ibid., §4.3), and in Old European (cf. Böhm 2003; Welscher 2005; Vennemann 2003b [passim]; 2010b; forthcoming), with corresponding phonological forms and similar meanings.   But there are also conspicuous correspondences among the derivational suffixes. To mention but one: The suffix -enn- is exemplified in Blasco Ferrer (2015, 164, 167, 168, 173, 180, 186) with the names ort/énn/oro, dosk/ol/énn/ero, ist/énn/ero50, lok/énn/ere, lok/énn/er/o, soru/énn/or/o, Arkénnere, Arkennui. Gorrochategui (1984, 368) lists -enn50 Blasco Ferrer’s (2015, 167) ist/énne/ero looks like a typing error.

Vennemann CW3 19 iv 2016.indd 524

27/05/2016 19:56:01

vennemann

[ 525 ]

among the suffixes occurring in the Aquitanian inscriptions of the Roman era and exemplifies it with the names BELEXENNIS (ibid., 160, genitive), LEHERENNO (ibid., 342–4, dative), and BORIENNO (ibid., 317, dative); he also refers to the numerous Pyrenean toponyms ending in -ein, -egn < +-enn-u, e.g. Andrein, Bugnein (ibid., 161, 370, summarized in Vennemann 2010b, 279f). These may be compared to the Old European name of the Ardennes, Caesar’s silva Arduenna, i.e. Ardu-enna, folk-etymological Celticization as ‘High Area’ of Vasconic +Ardi-enn-a ‘the Sheep Area’ 51; also the name of the Cevennes, Caesar’s mons Cebenna, may be explained as Ceb-enn-a < +Ce¯-enn-a ‘the Smoke Area’52, cf. Basque ke, eke, kee ‘smoke’, derivation form kei-, with normal changes ¯ > b and eu > ei in Basque (cf. Agud & Tovar 1993, s.v. ke ; Trask 1997, 152; Michelena 1977, 98).53   In his book Paleosardo of 2010, Blasco Ferrer, besides presenting roots and suffixes (in §5.4) very much as in the two preceding paragraphs54, also offers a paragraph on ‘Aspirazione e perdita di [f]-iniziale’ (§7.4), where he says:

A telling trait of the Barbagian dialects (of the Fonni-Ollolai group) and the Baronian dialects consists in the loss of the labiodental [f] in initial position: fĕmĭnam > [‘emina], facĕre > [‘akere, ‘aʔere], fĭlĭum > [‘idzu]. The intermediate stage, a faint aspirate [h], is today typical only of the community of Ovodda, although I have recorded similar realizations among the elderly at Baunei in the eighties.55 As is well known, the aspiration of initial [f] is typical of the entire Basque and Castilian area56, and the aspirate outcomes appear in the written texts as early as following the 9th century in the Hispanic area. In Basque, the Latin [f] is assimilated into a bilabial fricative [φ], which then yielded two alternative results: (1) a merger with the phoneme /p/ or the voiced /b/ ( f¼cum > [‘piku, ‘biku], fagum > [‘pago, ‘bago]); (2) aspiration with subsequent loss ( f¼lum > [‘hiru, ‘iru], f¼cum > [‘iku]). (author’s translation of Italian original; Blasco Ferrer 2010, 154f.)

Blasco Ferrer continues stressing the similarity between the Sardinian and the Basque reflexes of Latin f-: 51 52 53

54 55 56

Sheep raising was the major industry of that huge forest area in Antiquity and the Middle Ages (cf. Vennemann forthcoming). I.e. ‘the Fog Mountains’, which describes that mountain range well, cf. Vennemann Forthcoming. Patrizia Noel suggests (pers. comm. 5 July 2015) that the ‘Abweisung des Phonems f ’ should also be evident in the Old European Toponymy. I have dealt with the frequent occurrence of f- in the toponymy of western central Europe and in the names of the Rhenish Matrons in Vennemann 1994b, pointing out that these f’s could not be Germanic (no Germans living in western central Europe early enough to be responsible for those names), or Celtic or Basque (these languages not possessing an f ). My proposal then was that those f’s derived from Vasconic b’s interpreted by speakers of Italic (then still living in western central Europe) as +bh’s and taken by them through the Italic Consonant Shift, part +bh- > f-. For þ- and h- I likewise assumed an origin in Italic +dh- > þ- and +gh- > h-. I have not seen a critique of this proposal, so I let it stand. The comparison of roots is even more copious in Blasco Ferrer 2010 than in Blasco Ferrer 2015. For example, Blasco Ferrer 2010 includes toponyms reflecting the Vasconic root (h)aran ‘valley’, which plays a major role in the Old European branch of Vasconic; cf. Vennemann 2006. Footnote: ‘[…] A Baunei sono riuscito ad attestare anche una fricativa [φ] per [f], che è lo stadio anteriore all’aspirata.’ [At Baunei also a fricative [φ] for [f] could be recorded, which is the stage preceding the aspirate.] Footnote: ‘Rohlfs (1970, 145–8); Baldinger (1971, 22–5).’

Vennemann CW3 19 iv 2016.indd 525

27/05/2016 19:56:01

[ 526 ]

xviii. celtic as vasconized indo-european?

In this context the surprising parallelism between the first Basque solution and the reflex of [f] in certain Barbagian and central dialects deserves underlining, such as fŏnt\nam > puntana, pintana ‘spring, source’.   The twofold co-incidence of results between the central and eastern Sardinian dialects on one hand and the Basque development lessens the counter-probative value of occasional cases of aspiration in Neo-Latin dialects for which no Basque influence can be claimed, and also makes more costly the structuralist thesis which recognizes in the present phenomenon a simple extension of various lenition phenomena for voiceless obstruents. Moreover, what is even more surprising is the neat overlap between the area of the greatest density of PalaeoIberian toponyms and the area in which we find the phonetic-phonological phenomenon here discussed. (author’s translation of Italian original; Blasco Ferrer 2010, 155)

  In short, we have two model cases, Castilian and Gascon, for the change of f into h with subsequent loss in an Indo-European language on Vasconic substrates, plus one further case, the Barbagian dialects of Sardinia, which points in the same direction. This strengthens the thesis that the same change in Proto-Celtic likewise occurred on a Vasconic substrate.   Rohlfs (1977, §460) mentions as a surprising fact that the change of f into h is not found in Aragon, that part of Spain which is geographically closest to Gascony. However, such differences are not unusual for changes originating through language contact, since everything depends on details of the socio-linguistics of the contact.57 We find this exemplified by the changes in the preceding sections: Proto-Germanic, a neighbour of Proto-Celtic, only participates in the accent shift but not in the development of a second copula58 and not in the loss of +p (or +f ); and Italic partakes in the accent shift and shows traces of a second copula59, but keeps Proto-Indo-European +p unaltered and also preserves the f developed in the Italic Consonant Shift.   Rohlfs (ibid.) mentions in a closing footnote that the change of +f into h (and zero) is not unknown in northern Italy and even in several zones of Calabria. He attributes this to a previous change of f into bilabial ɸ, which is, of course, possible but may also be nothing more than an artefact serving the purpose of explaining the loss by internal motivation, rather than by language contact with languages lacking f. I have not studied the Italian cases, 57 The great variety of reactions to phonological contact influences is stressed in Matras (2009, §8.4.1, especially p. 224). In terms of Matras’ (ibid., 225) typology of ‘processes leading to phonological change’, Gascon may have developed through type C, ‘Convergence of systems during second-language acquistion’, and Aragonese, through type D, ‘Convergence of systems in stable, intensive bilingualism’. The two types are distinguished socio-linguistically by the categories ‘Speakers/Bilingualism’ and ‘Language attitude’ as follows: type C: ‘Emerging bilingualism; stable minority bilingualism; emergence of ethnolect or language shift’, ‘Strong group identity coupled with a need (pressure) to acquire the target language’; type D: ‘Intensive and widespread bilingualism’, “Second language is ‘prestige language”’. 58 Since Old Norse and Gothic show no sign of a second copula, the influence did not reach Proto-Germanic. Only West Germanic was affected, and only secondarily, viz. by contact with Celtic. 59 Cf. Lewis & Pedersen (1989, §476.1): ‘The paradigm of the verb ‘to be’ consists in Italo-Celtic of forms of the roots *es- and *bheu-. In Celtic a pres. stem *bhw¼-, *bhwije-, derived from *bheu-, also appears. This latter present denotes either a praesens consuetudinale or a future, a natural development from an orig[inal] meaning ‘to become’ (Lat. fi}). The same root is also used in the subjunctive. The root *es- stands only in the pres. and ipf. ind. in Celtic; in Ir[ish] it is not found in the ipf.’ In historical Latin the +bh- forms do not (no longer?) form a separate copular paradigm but have been merged with the +es- forms into a single paradigm. Cf. the chapter ‘The verb ‘be’ in Greek and Latin’ (Sihler 1995, §§491–6).

Vennemann CW3 19 iv 2016.indd 526

27/05/2016 19:56:01

vennemann

[ 527 ]

but for northern Italy I would expect contact with Vasconic and with Vasconized Celtic substrates, for southern Italy with Greek substrates (Magna Graecia); when the Greek colonies of Southern Italy were Latinized, Greek φ [ph] had not yet changed into f. 60 Celtic from the east In the continuing discussion with John Koch61 I suggested that the ‘Vasconic’ scenario of the origin of Celtic would be harmonious with the tradional ‘Hallstatt’ model of the spread of Celtic. Salt mining in the Hallstatt area began in pre-Celtic, viz. Vasconic times on the Gosau side of the Salzberg (Salt Mountain), the village Gosau preserving the Vasconic word for salt (Basque gatz ‘salt’, gaz-i ‘salty’) in its name. There occur other Vasconic names in the area as well, as I pointed out in my presentation at the 2014 Atlantiar conference.62 The oldest Hallstatt Celticity is conventionally dated as early as 800 BC. There is no apparent reason not to assume that the Celts were already Vasconized (i.e. truly Celts and not merely Indo-Europeans) at this time. Expanding deeper into western Europe they spread their Vasconized Celticity from this central European area in a westerly direction to northern Germany, the Low Countries, to Gaul (where their Vasconicity is likely to have intensified), and further to the Iberian Pensinsula and to the British Isles, taking everywhere the structural Vasconisms treated in the present article. The traditional model of the Celticization of Europe thus appears adequate. Therefore attempts to locate the formation of Celtic in the Iberian Peninsula (or at least in the vicinity of the Pyrenees) and have it spread to the rest of Europe from there, as advocated in several papers contained in the Celtic from the West volumes, appear unnecessary from a linguistic point of view. The linguistic arguments in the present article support the century-old view, which ought to be the null hypothesis even on purely geographical grounds, that the Celts, after entering western central Europe as Indo-Europeans from the east, and becoming Celts as they did, carried their Celticity further north- and westward all the way to the Iberian Peninsula and to the British Isles. The only point on which my account differs from this venerable tradition is the thesis that in the course of their westward expansion, those Indo-Europeans became Celts through their Vasconization in western central Europe. To paraphrase John Koch: Celtic is Indo-European spoken with a Vasconic accent.63 60 To my surprise I find the view that the change of f into h is likely everywhere to be owed to substrate effects already expressed in 1956: Basque linguistic habit has to be held responsible for this phenomenon [f > h in Spanish]. It is also found on the northern side of he Basque Country, in Gascon. […] The replacement of f- by h- is also found— probably owing to substrate habits—in southern Italian and Sardinian dialects. (author’s translation of German original; Lausberg 1956, II. §302) 61 Cf. the beginning of the section on the loss of the voiceless labial obstruent above. 62 The material was also presented at the 2014 conference of the Indogermanische Gesellschaft in Copenhagen. It is to appear in Vennemann (forthcoming). 63 This manner of speaking is used by Matras (2009, 223) in a discussion of phonology, but it may, of course, be employed more generally: ‘The substrate effect may lead to a group-particular ‘accent’ (e.g. in the case of an ethnolect), or in the event of language shift to the emergence of a new native variety of the target language with its own distinctive phonology. The collective substrate effect is no different in principle from what we call ‘accent’ or ‘interference’ at the level of the individual speaker.’

Vennemann CW3 19 iv 2016.indd 527

27/05/2016 19:56:01

[ 528 ]

xviii. celtic as vasconized indo-european?

  In conclusion, I would like to single out a consequence already touched upon of comparing the three western-most Indo-European language families on the basis of the three structural properties focused in this article. Germanic, Italic, and Celtic must be arranged in this particular order, if the criterion used is the three branches’ degree of Vasconization:

Initial accent

Germanic Italic Celtic

64  

Two copulas  

Loss of +p



By this criterion—and that is at the same time my answer to the question in the title of this article—Celtic is the most Vasconized branch of Indo-European. This is, of course, not surprising in view of the fact that it is also that branch of the language family that penetrated most deeply into Vasconic territories. Acknowledgements In preparing this article I have gratefully received valuable help, advice, and encouragement from the following colleagues: Eduardo Blasco Ferrer (University of Cagliari), Andreas Dufter (University of Munich), John T. Koch (University of Wales, Aberystwyth), Stephen Laker (Kyushu University, Fukuoka), Angelika Lutz (University of Erlangen), Robert Mailhammer (University of Western Sydney), Guido Mensching (University of Göttingen), Patrizia Noel (University of Bamberg), Peter Schrijver (University of Utrecht), and Iva Welscher (Munich). Antoniy Dimitrov M.A. (University of Munich) has helped with the bibliography. The responsibility for remaining mistakes and deficiencies of form or content is mine alone.

BIBLIOGRAPHY Agud, M., & A. Tovar 1990–95 Diccionario etimológico vasco. Donostia / San Sebastian, Diputación Foral de Gipuzkoa. (Only 7 vols., A – Orloi.) de Azkue, R. M. 1984 Diccionario vasco—español— francés. 1st edn. 1905, Bilbao, Euskaltzaindia. [Reprint.] Baldinger, K. 1971 2nd edn. La formación de los dominios lingüísticos en la Península Ibérica. Madrid, Gredos. Ballester Gómez, X. 2004 ‘Hablas indoeuropeas y anindoeuropeas en la Hispania prerromana’, Real Académia de Cultura Valenciana: Sección de estudios ibéricos ‘D. Fletcher Valls’: Estudios de lenguas y epigrafía

antiguas—ELEA 6, 107–38. Beekes, R. S. P. 2011 Comparative Indo-European linguistics: An introduction. 2nd edn., rev. and corr. by M. de Vaan. Amsterdam, John Benjamins. [1st edn. 1995.] Blasco Ferrer, E. 2002 Linguistica Sarda: Storia, metodi, problemi. Cagliari, Condaghes. Blasco Ferrer, E. 2010 Paleosardo: Le radici linguistiche della Sardegna neolitica. Berlin, Walter de Gruyter. Blasco Ferrer, E. 2013 ‘Palaeosardinian’, Iberia e Sardegna: Legami linguistici, archeologici e genetici dal Mesolitico all´Età del Bronzo: Proceedings of the International Congress “Gorosti U5b3” (Cagliari-

64 But see note 14.

Vennemann CW3 19 iv 2016.indd 528

27/05/2016 19:56:01

vennemann

Alghero, June 12–16, 2012), eds E. Blasco Ferrer, P. Francalacci, A. Nocentini, & G. Tanda, 7–33. Florence, Le Monnier Università. Blasco Ferrer, E. 2015 ‘Paläosardisch und Paläobaskisch’, The linguistic roots of Europe (Copenhagen Studies in Indo-European, 6), eds R. Mailhammer, Th. Vennemann, & B. A. Olsen, 155–200. Copenhagen, University of Copenhagen, Museum Tusculanum Press. Blasco Ferrer, E. (forthcoming) ‘Substrata residue, linguistic reconstruction, and linking: Methodological premises, and the case history of Palaeo-Sardinian’, Voprosy onomastiki. Böhm, A. 2003 Probleme der Deutung mitteleuropäischer Ortsnamen, mit besonderer Berücksichtigung der Toponymie des deutschsprachigen Raumes und einem Ausblick auf den appellativischen Wortschatz des Deutschen (Reihe Sprachund Literaturwissenschaften, 8). Munich, Herbert Utz. Bollée, A., & I. Neumann-Holzschuh 2003 Spanische Sprachgeschichte (Uni-Wissen Romanistik). Stuttgart, Ernst Klett. [2007 printing.] Broderick, G. 2015 ‘Latin and Celtic: The substantive verb’, Glotta 91, 3–14. Cancik, H., & H. Schneider (eds) 1996–2002 Der Neue Pauly: Enzyklopädie der Antike, 12 vols. Stuttgart, J. B. Metzler. Contini, M. 1987a Étude de géographie phonétique et de phonétique instrumentale du sarde: Texte. Alessandria, Edizioni dell’Orso. Contini, M. 1987b Étude de géographie phonétique et de phonétique instrumentale du sarde: Atlas et album phonétique. Alessandria, Edizioni dell’Orso. Corominas, J. 1931 ‘Vocabulario aranés’. Doctoral thesis. Barcelona, Imprenta de la Casa de la Caridad. Cunliffe, B., & J. T. Koch (eds) 2010 Celtic from the West: Alternative perspectives from archaeology, genetics, language and literature, Celtic Studies Publications 15. Oxford, Oxbow Books. Eska, J. F. 2013 ‘In defense of Celtic /Φ/’, Multi Nominis Grammaticus: Studies in Classical and Indo-European linguistics in honor of Alan J. Nussbaum on the occasion of his sixty-fifth birthday, eds A. I. Cooper, J. Rau & M. Weiss, 32–43. Ann Arbor, Beech Stave Press. Fähnrich, H. 1994 Grammatik der altgeorgischen Sprache. Hamburg, Helmut Buske. Feist, S. 1913 Kultur, Ausbreitung und Herkunft der Indogermanen. Berlin, Weidmann. Fife, J. 1993 ‘Introduction’, The Celtic languages (Routledge Language Family Descriptions), ed. M. J.

Vennemann CW3 19 iv 2016.indd 529

[ 529 ]

Ball with J. Fife, 3–25. London, Routledge. Garvens, F. 1964 ‘Die vorrömische Toponymie Nordspaniens’. Dissertation, University of Münster/Westphalia. Gorrochategui Churruca, J. 1984 Estudios sobre la onomástica indígena de Aquitania. Salamanca, Universidad de Salamanca, Secretariado de Publicaciones. Greene, D. 1992 ‘Celtic’, Indo-European Numerals, Trends in Lingustics 57, ed. J. Gvozdanović, 496–554. Berlin, Mouton de Gruyter. Hickey, R. 2007 Irish English: History and Present-day Forms (Studies in English Language). Cambridge, Cambridge University Press. Hualde, J. I. 1999 ‘Basque accentuation’, WordProsodic Systems in the Languages of Europe, Empirical approaches to language typology, Eurotyp 20–4, ed. H. van der Hulst, 947–93. Berlin, Walter de Gruyter. Hualde, J. I.. 2003 ‘Accentuation’, A grammar of Basque, Mouton Grammar Library 26, eds J. I. Hualde & J. Ortiz de Urbina 65–72. Berlin, Mouton de Gruyter. Hualde, J. I.. 2007 ‘Historical convergence and divergence in Basque accentuation’, Tones and Tunes, vol. 1: Typological studies in word and sentence prosody, Phonology and Phonetics 12.1, eds T. Riad & C. Gussenhoven 291–322. Berlin, Mouton de Gruyter. Hubschmid, J. 1963 ‘Paläosardische Ortsnamen’, VII Congresso Internazionale di Scienze Onomastiche: Memorie della Sezione Toponomastica, part 2, ed. C. Battisti, 145– 80. Florence, Istituto de glottologia dell´Università degli Studi. Ingrassia, G., & E. Blasco Ferrer 2009 Storia della lingua sarda: Dal Paleosardo alla musica rap: Evoluzione storico-culturale, letteraria, linguistica: Scelta di brani esemplari commentati e tradotti (I prontuari della lingua, 3). Cagliari, Cooperativa Universitaria Editrice Cagliaritana. Koch, J. T. 2009 Tartessian: Celtic in the South-west at the Dawn of History, Celtic Studies Publications 13. Aberystwyth, Celtic Studies Publications. [2nd edn. 2013.] Koch, J. T. 2010 ‘Paradigm Shift? Interpreting Tartessian as Celtic’, Celtic from the West: Alternative Perspectives from Archaeology, Genetics, Language and Literature, Celtic Studies Publications 15, eds B. Cunliffe & J. T. Koch, 185–301. Oxford, Oxbow Books. Koch, J. T. 2013a ‘Ha C1a ≠ PC (“The earliest Hallstatt Iron Age cannot equal Proto-Celtic”)’, Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, eds J. T. Koch &

27/05/2016 19:56:01

[ 530 ]

xviii. celtic as vasconized indo-european?

B. Cunliffe (Celtic Studies Publications, 16), 1–16. Oxford: Oxbow Books. Koch, J. T. 2013b ‘Out of the Flow and Ebb of the European Bronze Age: Heroes, Tartessos, and Celtic’, Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, Celtic Studies Publications 16, eds J. T. Koch & B. Cunliffe, 101–146. Oxford, Oxbow Books. Koch, J. T. 2015 ‘The emergence of the Celtic language in Atlantic Europe in later prehistory and its relationship with Proto Basque’, unpublished paper presented at the 2015 Atlantiar conference in Irun (Basque Country) on 15 May 2015. Koch, J. T., & B. Cunliffe (eds) 2013 Celtic from the West 2: Rethinking the Bronze Age and the Arrival of Indo-European in Atlantic Europe, Celtic Studies Publications, 16. Oxford, Oxbow Books. Koivulehto, J., & Th. Vennemann 1996 ‘Der finnische Stufenwechsel und das Vernersche Gesetz’, Beiträge zur Geschichte der deutschen Sprache und Literatur 118, 163–82. Lausberg, H. 1956 Romanische Sprachwissenschaft, vol. 1: Einleitung und Vokalismus, vol. 2: Konsonantismus (Sammlung Göschen 128/128a, 250). Berlin, Walter de Gruyter. Lewis, H., & H. Pedersen 1989 A Concise Comparative Celtic Grammar. (3rd edn., 2nd impr. with the supplement of 1961 by H. Lewis) Göttingen, Vandenhoeck & Ruprecht. Lloyd, P. M. 1987 From Latin to Spanish: Vol. I: Historical phonology and morphology of the Spanish language (Memoires of the American Philological Society, 173). Second printing 1989. Philadelphia, PA, American Philological Society. Lutz, A. 1991 Phonotaktisch gesteuerte Konsonantenveränderungen in der Geschichte des Englischen (Linguistische Arbeiten, 272). Tübingen, Max Niemeyer. Lutz, A. 2009 ‘Celtic influence on Old English and West Germanic’, Re-evaluating the Celtic hypothesis, eds M. Filppula & J. Klemola (Special issue of English Language and Linguistics 13), 227–49. MacAulay, D. 1992 ‘The Celtic languages: an overview’, The Celtic Languages, ed. D. MacAulay (Cambridge Language Surveys). Cambridge, Cambridge University Press. Maddieson, I. 1984 Patterns of sounds, Cambridge Studies in Speech Science and Communication. Cambridge, Cambridge University Press. [Paperback reprint 2009.] Martinet, A. 1964 Économie des changements phonétiques:

Vennemann CW3 19 iv 2016.indd 530

Traité de phonologie diachronique. Berne, A. Francke. [1st ed. 1955.] Matras, Y. 2009 Language contact, Cambridge Textbooks in Linguistics. Cambridge, Cambridge University Press. McCone, K. R. 1996 Towards a Relative Chronology of Ancient and Medieval Celtic Sound Change, Maynooth Studies in Celtic Linguistics 1. Maynooth, St Patrick’s College, Department of Old and Middle Irish. Meid, W. 2015 ‘Celtic from the West?’ Diachronie und Sprachvergleich: Beiträge aus der Arbeitsgruppe ‘historischvergleichende Sprachwissenschaft’ bei der 40. Österreichischen Linguistiktagung 2013 in Salzburg, Innsbrucker Beiträge zur Sprachwissenschaft 150, eds T. Krisch & S. Niederreiter, 27–34. Innsbruck, Institut für Sprachen und Literaturen der Universität Innsbruck, Bereich Sprachwissenschaft. Meier-Brügger, M. 2010. Indogermanische Sprachwissenschaft. (9th edn.) Berlin, Walter de Gruyter. Mensching, G., & E.-M. Remberger Forthcoming ‘Chapter 16: Sardinian’, The Oxford Guide to the Romance Languages, eds A. Ledgeway & M. Maiden. Oxford, Oxford University Press. Michelena, L. 1954 ‘De onomástica aquitana’, Pirineos 10, 409–58. Michelena, L. 1977 Fonética histórica vasca (2nd edn., 1990 reprint). San Sebastián, Diputación Foral de Gipuzkoa. [1st edn. 1961.] Paulis, G. 1992 I nomi popolari delle piante in Sardegna. Sassari, Carlo Delfino. Paulis, G. 2008 ‘Il paleosardo: retrospettiva e prospettive’, ΑΙΩΝ: Sezione linguistica 30, 11–61. Pokorny, J. 1923 ‘Keltisch-Baskisches’, Zeitschrift für celtische Philologie 14, 272–3. Pokorny, J. 1927–30 ‘Das nicht-indogermanische Substrat im Irischen’, Zeitschrift für celtische Philologie 16, 95–144, 231–66, 363–94; 17, 373–88; 18, 233–48. Pokorny, J. 1989 Indogermanisches etymologisches Wörterbuch, 2 vols. (vol. 2: Index by H. B. Partridge). 2nd edn. Berne, Francke. [1st edn. 1959.] Pountain, Ch. J. 2001 A history of the Spanish language through texts. London, Routledge. Prósper, B. M. 2014 ‘Some Observations on the Classification of Tartessian as a Celtic Language’, The Journal of Indo-European Studies 42, 468–86. Rohlfs, G. 1966–69 Grammatica storica della lingua italiana e dei suoi dialetti, 3 vols. Turin, Einaudi. Rohlfs, G. 1977 Le gascon: Études de philologie pyrénéenne, Beihefte zur Zeitschrift für romanische Philologie

27/05/2016 19:56:01

vennemann

85, 3rd augm. edn. Tübingen, Max Niemeyer. [1st edn. Halle (Saale), Max Niemeyer, 1935; 2nd edn. Tübingen, Max Niemeyer, 1970.] Román del Cerro, J. L. 1990 El desciframiento de la lengua ibérica en ‘La Ofrenda de los Pueblos’. Alicante, Aguaclara. Russell, P. 1990 Celtic Word Formation: The Velar Suffixes. Dublin, Dublin Institute for Advanced Studies. Salmons, J. C. 1992 Accentual Change and Language Contact: Comparative Survey and a Case Study of Early Northern Europe. London, Routledge. Schmitt, R. 1981 Grammatik des Klasisch-Armenischen mit sprachvergleichenden Erläuterungen (Innsbrucker Beiträge zur Sprachwissenschaft, 32). Innsbruck, Universität Innsbruck, Institut für Sprachwissenschaft. Schrijver, P. 2002 ‘Irish aindir, Welsh anner, Breton annoar, Basque andere’, Sounds and Systems: Studies in Structure and Change: A festschrift for Theo Vennemann, Trends in Linguistics: Studies and Monographs 141, eds D. Restle & D. Zaefferer, 205–19. Berlin, Mouton de Gruyter. Schrijver, P. 2011 refers to the talk and handout published as Schrijver 2015. Schrijver, P. 2015 ‘Pruners and trainers of the Celtic family tree: The rise and development of Celtic in the light of language contact’, Proceedings of the XIV[th] International Congress of Celtic Studies, held in Maynooth University, 1–5 August 2011, eds L. Breatnach, R. Ó hUiginn, D. McManus, & K. Simms, 191–219. Dublin, Dublin Institute for Advanced Studies. Available on the Internet at www.academia. edu/13749183/Pruners_and_trainers_of_the_Celtic_ family_tree_the_rise_and_development_of_Celtic_ in_the_light_of_language_contact (18 August 2015). Schuchardt, H. 1887 ‘Romano-baskisches I. P-’, Zeitschrift für romanische Philologie 11, 474–512. Schumacher, St. 2007 ‘Die Deutschen und die Nachbarstämme: Lexikalische und strukturelle Sprachkontaktphänomene entlang der keltischgermanischen Übergangszone’, Johann Kaspar Zeuß im kultur- und sprachwissenschaftlichen Kontext (19. bis 21. Jahrhundert), Kronach 21.7.–23.7. 2006, eds H. Hablitzel & D. Stifter, 167–207. Wien, Praesens Verlag. Sihler, A. L. 1995 New Comparative Grammar of Greek and Latin. Oxford, Oxford University Press. van Sluis, P. 2014 ‘The “Atlantic Fringe” hypothesis for the Celtic homeland and the Tartessian inscriptions’, Academia page https://fryske-akademy.academia. edu/PaulusvanSluis (15 September 2015).

Vennemann CW3 19 iv 2016.indd 531

[ 531 ]

Tagliavini, C. 1998 Einführung in die romanische Philologie. 2nd, improved edn. Tübingen: A. Francke. [1st edn. München: C. H. Beck, 1973. Translation by R. Meisterfeld & U. Petersen of Le origini delle lingue neolatine: Introduzione alla filologia romanza, 6th edn., Bologna, Prof. Riccardo Pàtron, 1969.] Thomason, S., & T. Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley, University of California Press. Trask, R. L. 1997 The History of Basque. London, Routledge. Trask, R. L. 2008 Etymological Dictionary of Basque, ed. for web publication by M. W. Wheeler. Brighton, University of Sussex. Cf. http://web.archive.org/ web/20110607202936/http://www.sussex.ac.uk/ linguistics/documents/lxwp23-08_edb.pdf (18 August 2015). Trudgill, P. 2011 ‘A tale of two copulas: Languagecontact speculations on first-millennium England’, NOWELE: North-Western European Language Evolution 62–3, 285–320. Vennemann, Th. 1972 ‘Rule inversion’, Lingua 29, 209–42. Vennemann, Th. 1988 Preference Laws for Syllable Structure and the Explanation of Sound Change: With Special Reference to German, Germanic, Italian and Latin. Berlin, Mouton de Gruyter. Vennemann, Th. 1993 ‘Zur Erklärung bayerischer Gewässer- und Siedlungsnamen’, Sprachwissenschaft, 425–83. [Reprinted as chapter 3 in Vennemann 2003b.] Vennemann, Th. 1994a ‘Linguistic reconstruction in the context of European prehistory’, Transactions of the Philological Society 92, 215–84. [Reprinted as chapter 6 in Vennemann 2003b.] Vennemann, Th. 1994b ‘Die mitteleuropäischen Ortsund Matronennamen mit f, þ, h und die Spätphase der Indogermania’, Früh-, Mittel-, Spätindogermanisch: Akten der IX. Fachtagung der Indogermanischen Gesellschaft vom 5. bis 9. Oktober 1992 in Zürich, eds G. E. Dunkel, G. Meyer, S. Scarlata, & Ch. Seidl, 403–26. Wiesbaden, Ludwig Reichert. [Reprinted as chapter 4 in Vennemann 2003b.] Vennemann, Th. 1995 ‘Etymologische Beziehungen im Alten Europa’, Der GinkgoBaum: Germanistisches Jahrbuch für Nordeuropa 13, 39–115. [Reprinted as chapter 7 in Vennemann 2003b.] Vennemann, Th. 1998 ‘Andromeda and the Apples of the Hesperides’, Proceedings of the Ninth Annual UCLA Indo-European Conference, Los Angeles, May 23, 24,

27/05/2016 19:56:01

[ 532 ]

xviii. celtic as vasconized indo-european?

1997, Journal of Indo-European Studies Monograph Series 28, eds K. Jones-Bley, A. Della Volpe, M. Robbins Dexter, & M. E. Huld, 1–68. Washington, D.C., Institute for the Study of Man. [Reprinted as chapter 18 in Vennemann 2003b.] Vennemann, Th. 2003a ‘Languages in prehistoric Europe north of the Alps’, Languages in Prehistoric Europe, eds A. Bammesberger & Th. Vennemann (Indogermanische Bibliothek, Dritte Reihe), 319–32. Heidelberg, Carl Winter. [Reprinted as chapter 15 in Vennemann 2012.] Vennemann, Th. 2003b Europa Vasconica—Europa Semitica, Trends in Linguistics: Studies and Monographs 138, ed. P. Noel Aziz Hanna. Berlin, Mouton de Gruyter. Vennemann, Th. 2006 ‘“A satisfactory etymology has long been available”: Notes on Vasconic names outside the Basque country: With particular reference to some British Arn- and Earn- names and to German Arnoldsweiler’, Studies in Basque and Historical Linguistics in Memory of R. L. Trask, Anuario del Seminario de Filología Vasca “Julio de Urquijo”, 40.1–2, eds J. I. Hualde & J. Lakarra, 969–92. Donostia/San Sebastián, Diputación Foral de Gipuzkoa. Vennemann, Th. 2010a ‘Contact and prehistory: The Indo-European Northwest’, The Handbook of Language Contact, ed. R. Hickey (Blackwell Handbooks in Linguistics), 380–405. Malden, MA, Wiley-Blackwell. Vennemann, Th. 2010b ‘Contact and prehistory in the the Indo-European Northwest: Lexical influences’, Sprachwissenschaft 35, 247–290. Vennemann, Th. 2011 ‘English as a contact language: Typology and comparison’, Anglia 129, 217–57. Vennemann, Th. 2012 Germania Semitica, Trends in Linguistics: Studies and Monographs 259, ed. P. Noel Aziz Hanna. Berlin, De Gruyter Mouton. Vennemann, Th. forthcoming ‘Nicht-indogermanische Spuren vorgeschichtlicher Wirtschaftsformen im toponymischen Lexikon Mitteleuropas: Seen des Salzkammerguts und Ardennen’, Paper presented at the 14th Fachtagung ‘Etymology and the European Lexicon’ of the Indogermanische Gesellschaft, University of Copenhagen, 17–22 September 2012. Also presented as ‘The Vasconic substrate in Central Europe: Toponyms reflecting economy’ at the 2014 Atlantiar Conference in Irun. Virdis, M. 1988 ‘Sardisch: Areallinguistik (Aree linguistiche)’, Lexikon der Romanistischen Linguistik (LRL), eds G. Holtus, M. Metzeltin, & Ch.

Vennemann CW3 19 iv 2016.indd 532

Schmitt, vol. 4: Italienisch, Korsisch, Sardisch, 897–913. Tübingen, Max Niemeyer. Wagner, M. L. 1980 La lingua sarda: Storia, spirito e forma (Bibliotheca Romanica, Series prima, 3). 2nd edn. Berne, A. Francke. [1st edn. 1951.] Weinreich, U. 1953 Languages in Contact: Findings and Problems (Publications of the Linguistic Circle of New York, 1). New York, Linguistic Circle of New York. [1963 and later reprints: The Hague, Mouton]. Weinrich, H. 1958 Phonologische Studien zur romanischen Sprachgeschichte, Forschungen zur romanischen Philologie 6. Münster, Westfalen Aschendorff. Welscher, I. 2005 ‘Mitteleuropa und Südosteuropa im Kontext der Alteuropäischen Toponymie’. Dissertation, Fakultät für Sprach- und Literaturwissenschaften, Universität München. Winford, D. 2003 An Introduction to Contact Linguistics. Oxford, Blackwell. Wolf, H. J. 1992 Studi Barbaricini: Miscellanea di saggi di linguistica sarda (Studi di Linguistica sarda, 2). Cagliari, Edizioni della Torre. Wolf, H. J. 1998 Toponomastica Barbaricina: I nomi di luogo dei comuni di Fonni, Gagoi, Lodine, Mamoiada, Oliena, Ollolai, Olzai, Orgòsolo, Ovodda (Glossa, 2). Nùoro, Insula. Zeidler, J. 2011 Review of Cunliffe & Koch (eds) 2010 Bryn Mawr Classical Review 2011.09.57 (www. bmcreview.org/2011/09/20110957.html, 12 September 2015).

Internet sites (all last accessed on 18 August 2015) ‘Barbagia’, en.wikipedia.org/wiki/Barbagia. ‘The “f > h change” 1’ [part of the ‘Spanish phonetics and phonology’ series of Prof. Christopher J. Pountain], people.ds.cam.ac.uk/cjp16/ learnsupp/fh.pdf. ‘Gascon language’, en.wikipedia.org/wiki/Gascon_ language. ‘Georgian language’, en.wikipedia.org/wiki/ Georgian_language. ‘Vascones’, en.wikipedia.org/wiki/Vascones. ‘Vasconic languages’, en.wikipedia.org/wiki/ Vasconic_languages. ‘Vasconic substratum theory’, en.wikipedia.org/wiki/ Vasconic_substratum_theory.

27/05/2016 19:56:01

INDEX abbreviated language names used in the index

Aquit Aquitanian

MIr

Ctb

MW Middle Welsh

Celtiberian

Middle Irish

PIE    Proto-Indo-European Sem

Semitic

Gaul Gaulish

OB

Old Breton

Skt

Sanskrit (incl. Vedic)

HC

Hispano-Celtic

OE

Old English

T

IC

(Proto-)Italo-Celtic

OHG Old High German

Ir

Irish

OIr

Old Irish

Tartessian (language of the SW inscriptions)

Lat

Latin

ON

Old Norse

W

Welsh

Lus

Lusitanian

OW

Old Welsh

MB

Middle Breton

PC

Proto-Celtic

aarkuui (T) 

438, 459 ABAC (=Algarve-Baixo Alentejo Complex)  195, 198, 203–4, 207 ABO blood groups  336 accent, word-initial  493, 506–9 -acum, Gallo-Roman placenames  436 adar ‘horn’ (Basque) 506 adarc ‘horn’ (OIr)  506 aDNA (=ancient DNA)  2, 4, 14, 333, 351–7, 360–1, 368–70, 375 AEMA project  1–2, 7, 9, 431 Afro-Asiatic languages  407, 411 aindir ‘girl’ (OIr)  506 akoosioś naŕkeetii (T)  460 -al, Etruscan suffix  501 Alcalá del Río, SW inscription  463–8 Alentejo region, Baixo Alentejo  98, 183–7, 190–207 Alentejo stelae, of MBA  196–201, 440, 450

mynegai_CW3.indd 533

Alfarrobeira MBA cist cemetery  184 Algarve region  182–5, 190–5, 198–207 *allobrogs ‘stranger, foreigner’  399 *alloto¯tios ‘stranger, foreigner’  399 alluvial gold deposits  129 Alteuropäisch  5 *altros ‘fosterage’  399 \má- ‘raw’ (Skt)  490 Ambata, Ambatos (HC)  406–7 Amesbury, Amesbury Archer  23, 26, 84, 126, 132, 135 Amgueddfa Cymru – National Museum Wales  295 AMS dates   2, 6, 142–3 Anatolian farming hypothesis  5 Ancient British fringe, ancestry  336–40, 346–8 andere ‘young woman’ (Basque & Aquit)  506 Andronovo horizon  366 Anglo-Saxon invasion  348

Anglo-Saxon population/ ancestry  7, 336, 341, 346 animus ‘soul, spirit’ (Lat)  493 anner ‘heifer’ (W)  506 annoer ‘heifer’ (MB)  506 anta  89 antimony in bronze  250–2 antler tools  132 AOC Beakers  147 AOO Beakers  27, 147 apersonal verb-marking  407 Aquitanian language  432–4, 468 Arabic language  407–9, 413, 416–24 arcat ‘silver’ (OIr)  495 archaeogenetics  2, 4, 8, 351, 431 archery  438–42, 468 Arcia, Arcius  439 Arcobriga [Monreal de Ariza, Zaragoza]  441–2 arcu\tus ‘bent like a bow’ (Lat)  439 Arduenna < Vasconic +Ardi-enn-a

31/05/2016 18:10:47

[ 534 ] ‘the Sheep Area’  525 arffed ‘lap or abdomen’  438–9 ’Αργανθωνιος  459–62 Argar, Argaric culture  41, 67–8, 72–4 argentum ‘silver’ (Lat)  495 arƕazna ‘arrow’ (Gothic)  439 Armorica  135 Armorican axes  311–12, 315, 319–22 Arquius (HC)  8, 438–9 arsenic in bronze  250–5 Atalaia, MBA grave system  184, 199 Atlantic Bronze Age / Atlantic Late Bronze Age  1, 6, 8, 34, 181, 187–8, 434–6 Atlantic buckets  435 Aunjetitz  41, 67–8, 72–3 axes, Bronze Age  253–61, 299, 307–15, 319–22 Baiões–Plaza de las Monjas phase  442–4, 449–50, 454, 468 Ballybriest, Co. Derry, wedge tomb  154–6, 165 Ballyedmonduff, Co. Dublin, wedge tomb  154–6 Ballylin hillfort  221, 231–3 Banc Tynddol gold disc  6, 111–21 Barbagian dialects of Sardinia  516, 520–6 barbed-and-tanged flint arrowheads  132–4, 152 basket ornaments  125–8 Basque language  9, 409, 419 Basque toponymy in Spain  505 Baurnadomeeny, Co. Tipperary, wedge romb  154–6 Beaker complex/package  5–6, 8, 130, 133, 353, 364, 439 Beaker dispersal from Iberia  364– 6, 369 Beaker pottery in Ireland  146 Beaker prospecting  130 BELEXENNIS (Aquit)  525 Bell Beaker culture  139 Bell Beaker Goldwork Tradition, Primary  111, 130 Bell Beaker Set  127–8, 133

mynegai_CW3.indd 534

index

Berber language(s)  408–12, 416– 21, 424–5 biid ‘is wont to be’ (OIr)  493 bilingualism  8, 389–92, 396, 403 Bishopsland phase  242–4 bismuth in prehistoric bronze  250–1 bith ‘world’ (OIr)  492 Black-Burnished Ware  300 blin ‘weary’ (MW)  491 boga ‘bow’ (MIr)  439 Bopfingen  280 BORIENNO (Aquit)  525 Boscombe Bowmen  84, 135 bow and arrow  438–40 Bowl Tradition  147, 156, 159, 165–71 bowls, ring-handled  309–11, 322 Brean Down, Somerset, pottery  316–17, 323 Breton language  407–13, 416–24 briga-zone  3, 438 Broighter gold boat  141 burikounikum (Ctb)  480–1 Bury, collective grave  18–19, 22 bwa ‘bow’(W)  439 byd ‘world’ (W)  492 bydd ‘is wont to be’ (W)  493 C, haplogroup  358 Cabeceiras de Basto, Portugal, discs  123–7 Cádiz, see Gades/Gadir Canaanite script  452, 465 Canegrate culture  500 Castell Odo hillfort  266–7, 272, 276–80, 292 Castro dos Ratinhos hillfort  205 Caucasus region, genetic traits  355, 360, 365, 372–3 cauldrons  295, 299–301, 307–12, 319–24, 435 cave burials, Irish Beaker period  161–2 Cebenna < Vasconic +Ce¯-enn-a ‘the Smoke Area’  525 Cefn Graeanog II, fortified site  283–5

Celtiberian language  445–7, 456, 459, 464, 500 Celtic family tree  444–5 Celtic from the West  1–3, 8–9, 65, 425–7, 503–4 Celtic homeland   3–5, 9, 503–4 Celtic languages  3, 387 Celtici 483–4 Centum languages  432–4 Chaussée-Tirancourt, collective grave  19–20, 91–2, 96 Childe, V. Gordon  248–9 children, in Beaker burials  17–18 cist burials  2, 6, 85, 103, 148–51, 183–5, 190–9, 202–3, 451, 456 cists, megalithic, in Ireland  6, 156–8 Clashanimud hillfort  221, 225–8, 243 clientage  400 closing of monuments, Beaker period  83–103, 167 clothing and identity  397–8, 404 Clunia (Ctb)  480 coäir ‘proper’ (OIr)  494 Cobham, Kent, disc  123–4, 128 cóic ‘5’ (OIr)  494 Collared Urns  306–7, 315 collective burials  17–22, 28, 31–3, 83–4, 88, 99, 147–8, 164–5, 169–71 Comet Lode Opencast, Wales  115– 18 comingaire (OIr)  399 Common Celtic  388 comnessam ‘nearest’ (OIr)  399 Concani (HC)  483–5 conejo ‘rabbit’ (Spanish)  485 Conia (HC)  477–9 Coniagellietar (Ibericized)  477 Conicodius (HC)  477–9 Conii (HC)   477 Conimbriga (HC)  484–5 Conistorgis, Konisturgis 448, 484–5 conjugated prepositions  407–12, 423 construct state [syntactic feature]  407–9, 416, 419–23 Continental Celtic  407, 426, 464

31/05/2016 18:10:47

index

Coolnatullagh, Co. Clare, burial  157 Copa Hill copper mine  6, 111–18, 133 copper mining, prehistoric  111–35, 163, 172 copper-to-bronze transition  47– 8, 165–7, 171 copula, second, as Vasconic influence  506, 509–10, 526 Corded Ware cultures  24, 27, 366–70, 373–4 Cornish tin, gold  135 Cornwall and Devon, genetic evidence  343 couehriu ‘meeting-place’ (Volscan)  494 Couneancus (HC)  479–81  Couneidoq(um) (HC)  480–1 court, social institution  265, 279–81, 289–91 Cowgill, W.  489 craftsmen, travelling  392 cremation and inhumation, Irish Beaker period  164–5 cremation, Iberian BA 187–95, 201 cremation, Welsh BA 303–6, 315, 318, 322 crín ‘withered’ (OIr)  491 cúan ‘litter, pack of dogs’ (OIr)  480 cun ‘litter, pack of dogs’ (W)  480 Cunei (HC)  484 Cun{tio (Brit)  483 cŭn¼cŭlus (Lat)  485 c~ria ‘meeting-place’ (Lat)  494 Cynetes (HC)  484 Cynwyd (W)  482 Cypriot syllabic writing  204–5, 452–6 Cyprus, Cypriot influence  443, 449, 452–4, 469 daggers, Beaker period  27, 47, 59–61, 71, 96 daggers, EBA  252–4 Degannwy  284, 288 Denisovans  351

mynegai_CW3.indd 535

deposition of metalwork  244, 402–3, 435 Deverel-Rimbury pottery  304, 315, 318 Devoto, G.  489 dialect continuum  434, 444–5, 469 Dinas Dinlle, enclosure  267 Dinas Emrys  284, 288 DOENTI (Lus)  461 Doggerland  348 do-gní ‘does, makes’ (OIr)  493 Domnall (OIr)  495 donasan ‘they gave’ (Venetic)  495 Dowris phase  1, 7, 242–4 duine ‘person’ (OIr)  494 Dún Aonghasa  223–4, 243–5 d~rus ‘hard’ (Lat) 491–2 Dybo’s rule  490–1 dyn ‘person’ (W)  494 Early Mines Research Group (=EMRG)  118 Eastern Bell Beaker Province  14 Egyptian language  408–13, 416– 20, 424 élite dominance  4–5, 8, 388–9, 393 enclosures  265 279, 402 enclosures, ceremonial  403–4 eneff ‘soul’ (MB)   493 -enn- suffix  524–5 Ewart Park metalwork  311, 320, 437 f > (h) > zero in Barbagia   523–4 f > h in Castilian  516–18 f > h in Gascon  519–20 *f lost within PC  514–15 Face of Britain, TV programme  333, 336 feasting  7, 295, 308–9, 317, 320–3, 402–4 fer ‘man’ (OIr)  490 Ferchú (OIr)  478–9 ffrwd ‘stream’ (W)  494 fineSTRUCTURE analysis of Europe  345–6

[ 535 ] fineSTRUCTURE clustering, UK  344–8 f¼} ‘I become’ (Lat)  493 fír ‘true’(OIr)  479 flaith ‘rule’ (OIr)  495 Flemish settlement in Wales  333, 348 Food Vessels  53, 125, 147, 150, 161, 165, 168, 306 footprint motif   199 Formoyle hillfort  221, 233–5 fosterage  399 founder’s effect  395 Fourknocks II, Co. Meath, burial  150 Freestone Hill hillfort  224 fulacht fia  244 f~mus ‘smoke’ (Lat)  491–2 G2a, haplogroup  358 Gadir/Gades [Cádiz]  426, 443, 449, 469 ‑gainethar ‘is born’ (OIr)  493 Gargoris and Habis, foundation legend  456, 459 Garn Boduan hillfort  266–7 Garonne frontier  434–6 Gaulish-Brythonic-Goidelic  8, 437, 445, 468 Gavrinis, Morhiban  97 genetic structure of Europe  333 genomics  351 Gildas  289 glacial refuge  355–6, 360 Glanbane hillfort  221, 229–31 goarec ‘bow, arch’ (MB)  439 Goërem, Morbihan  93 Golasecca culture  500–1 Gomes Aires, stelae  199 grán ‘corn’ (OIr)  495 grave slabs  183, 187, 195–208 Great Orme’s Head, near Llandudno, mine  6, 111 Greek emporia  392 Grooved Ware  162–3 Guadalquivir, basin  179–81, 188, 191–5, 201, 204–10, 451, 465 Gündlingen swords  436–7 Gurci (OW)  478

31/05/2016 18:10:47

[ 536 ] Gurki (OB)  478 guth ‘voice’ (OIr)  492 Gwynedd, kingdom  283 H, haplogroup  356–8, 363–5, 369, 374 halberds  5–6, 40–78, 96, 198, 253, 450 Hallstatt culture  3–4, 203–4, 311–15, 320, 437, 503 Hallstatt, pre-Celtic salt mining at  527 Hamito-Semitic, languages/ substratum   8, 407–11, 416–25, 463–8 haplogroups  352 Harlockstown, Co. Meath, pit burial  160–1 Haughey’s Fort, Co. Armagh  223– 4, 402 Hebrew language  408–13, 416–24 Hecateus  482 Herodotus  179–80, 207, 448–9, 482 Herzsprung shields  187–8 Hiberno-English  432 hide boats  143 hillforts, British  7, 222, 244 hillforts, Bronze Age  222–3, 242–4 hillforts, Irish  7, 219–45 hillforts, post-Roman refortification  284–5 hillforts, social significance  241, 402 Hispano-Celtic  445 historical linguistics  395 HLA (=human leukocyte antigen) gene  333, 336 horse remains, Bronze Age Ireland  6, 144–5 hospitality  399 hostel  400 Huelva swords  444 Huelva, Phoenicians in  1, 442–4 Human Genome Project  353 Hungary, aDNA  354 hunter-gatherer-forager (=HG) groups  356–67

mynegai_CW3.indd 536

index

hypogaeum burials  85, 88–9, 99, 182–3, 191 I2a, haplogroup  358 Iberian inscriptions [NE]  435 Iberian language  432–6, 504–5, 511–13, 520, 524 identity, identity shift  396–7 Île Cairn, Finistère  93 imbliu ‘navel’ (OIr)  495 inciens ‘pregnant’ (Lat)   493 individual burial  84 individualized burials  147–8, 164–7 Indo-European homeland  4–5 Indo-European languages  370–5, 407–11, 424–6 Indo-Iranian languages  366 infinitive  410, 418–19, 423–4 inhumation, Iberian BA 183, 186–7, 190–5 Insular Celtic / Hamito-Semitic question  8, 407–11, 416–25 Irish language  408–12, 417–24 Italo-Celtic  3–5, 9, 489–500 Italo-Celtic family tree  499 kaaśetaana (T)  465 Keltoi, Κελτοί 448 Kirk Andrews, Isle of Man, disc  122–7 Kirkhaugh burial  132 Knowes of Trotty, Orkney, disc  125 *komarom ‘communal ploughing’  399 *komnetsamos ‘nearest, neighbour’  399 Koniakoi (HC)  483, 485 konikum (Ctb)  477–9 Koníoi, Kονιοι (HC)  448, 459, 482, 485 Konískoi (HC)  483–5 Konisturgis, Conistorgis 448, 484–5 *kor¾os ‘warband’ (PC)  400 ko‑tu-ua-ratee (T)  465 *koun\, *kouno- (PC)  9, 480–1 kounesikum (Ctb)  480–1 *k~ ‘dog, wolf ’ (PC)  485

*k¯ennom ‘head (of household)’  392 *kun- ‘dog’ (PC)  9, 477–82, 485 Kun{síoi (HC)  482 kunikum (Ctb)  477–9 *kunos genitive (PC)  485 Kurgan cultures  366, 370 *kwenkwe ‘5’ (IC)  494 *kwon- ‘dog’ (PC)  477 Kynetes, Cynetes, Kun{tes, Κυνητες 207–9, 448–51, 457–9, 466, 482, 485 kyweir ‘proper’ (MW)  494 La Joya, Huelva, necropolis  454 La Tène culture  3–4, 503 Labbacallee, Co. Cork, wedge tomb  152–4 Lagunita, Cáceres  96–7 lakeentii (T)  465 language contact  391–5 language shift  219, 244–5, 388–97, 403–4 Largantea, Co. Derry, wedge tomb  152, 165, 168 Late Glacial expansion  355–60, 365, 369, 372 Latin and the Romance languages  388 laws, Irish  288 laws, Welsh  286–8 leaded tin bronze   255, 258 legal proceedings  281, 286–8, 290–1 LEHERENNO (Aquit)  525 Lepontic language  9, 500–1 les (OIr)  282 Lhuyd, E.  3 Linear Pottery Culture or Linearbandkeramic (=LBK)  356–68 lingua franca  1, 3, 5, 8, 392, 431–4, 437, 442, 456, 466–7 linkage disequilibrium (=LD)  341–2 liss (OB)  7, 282 lissos (Gaul)  282 Little England beyond Wales  333, 343, 348 Llanllyfni lunula  129

31/05/2016 18:10:47

index

Llanmaes, Vale of Glamorgan  7, 294–326 Llano de la Belleza, Huelva  97 Llyn Fawr, hoard / period  309, 314–15, 321, 444 Ll½n Peninsula  7, 265–7, 272, 279–83, 292 llys ‘court’  7, 281–6, 289, 292 Llys Rhosyr, Angelsey  283–6 llysaf ‘reject, refuse, &c.’ (W)  282 local surnames  338–40 location quotient (=LQ)  338–9 logboats, Bronze Age  140–4 lokoon (T) 449 long vowel system, ProtoCeltic  498–9 Longbridge Deverill, Wiltshire, disc  123, 126–7 Lottner, C.  489 Lough Gur, Co. Limerick, wedge tomb  154–6 lucrum ‘profit’ (Lat)  491 lunulae  89, 122, 128–9 Lusitanian language  445, 452, 458, 461 mael ‘prince’ (MW)  494 Maelgwn (W)  482 magnus (Lat)  494 mál ‘prince’ (OIr)  494 mamoa  96 Maritime Bell Beakers  25–6, 34, 88, 94, 102–3, 127, 146 Marstrander, C.  489 Martinet, A.  508, 515 Massaliote Periplus  143 Medellín necropolis  448–9 megalithic monuments, tombs  83– 5, 103, 146–7, 187 Meillionydd hillfort  7, 265–92 Mell, Co. Louth, pit burial  160–1 Méréaucourt, Somme  92, 94 Meridional [SE Iberian] inscriptions  435 Mesolithic  356–65, 370, 373–4 metal chemistry  7, 247–61 metalling, of road/path  273 Michelena, L.  433, 508, 515

mynegai_CW3.indd 537

middens, Llanmaes midden  7, 295–325 Middle East, first farmers’ migration from  333, 351 migration stream  394–5 Millares, Los  74 mining, prehistoric  111–35 mitochondrial DNA (=mtDNA)  352–3 Mold Cape  122, 126 molecular clock  352, 355 Moneen, Co. Cork burial  157, 170 Monte Abraão, Sintra  94, 99 Montelevar horizon  73 Mooghaun hillfort  223–4, 243–5 Morris Jones, J.  408, 412, 420, 424 multivallate/multiple-enclosure hillforts, Irish  220 Mycenae, fall of   443, 449, 453 N1a1a, haplogroup  354–8, 361–5, 368, 373 Nantyreira, mining site  113 naŕkee, naŕkeentii, &c. (T)  449, 464 nativism in EIA south Portugal  204–6, 453–7, 466 Navan Fort, Co. Armagh  223–4, 278 Neanderthals  351 Near East, first farmers’ migration from  333, 351–2, 355–65, 368, 372–3 neđđamon ‘next, closest’ (Gaul)  494 neighbour alliance  399 Neolithic burial monument, reuse  83–110, 148–9 Neolithic transition [from HG]  355–61, 365, 368, 372–5 nessaf ‘next, closest’ (W)  494 nessimas ‘closest’ (Oscan)  494 New Guinea, multilingualism  392 next-generation sequencing (=NGS)  351–3 nickel in prehistoric bronze  250–2 non-Indo-European languages  387

[ 537 ] Nordic Bronze Age  15 Norse ancestry  338, 341, 346–8 oblique relatives with pronoun copies  407–8, 413, 422–3 of ‘raw’ (W)  490 Ogam script  1, 456 Old European, as Vasconic language  504–9, 515, 524–5 Old Irish language  416, 420, 423 om ‘raw’ (OIr)  490 *omo- ‘raw’ (IC)  490 Ora Maritima  143 orbaa (T)  465 Orientalization  442, 451, 456–7, 465–6 Orientalizing period  180, 194, 208 Orkney population, ancestry  341– 4, 348, 361 -os, -osse, -oz, -otz, -ués (Aquit place-name suffix)  434–6 Ötzi, Tyrolean Iceman  358–60, 370 oxidative loss of impurities  250–1 *p, weakening/loss/absence of   3, 9, 432, 468, 511–27 Palaeo-Basque language  432–4, 468 palaeogenetics  351 Palaeohispanic scripts  445–7, 452–7 Palaeo-Sardic, pre-Roman language  504–5, 508, 520, 524 palimpsest effect, genetic  354, 361 Palmela points  6, 40, 58–61, 103 palstaves  255 Parys Mountain, Anglesey  111 peer polities  392 Pego da Sobreira, EIA funerary structure  192 Penll½n  267, 276, 279, 292 People of the British Isles (=PoBI) project  7, 334–6, 343–5 Phoenician language, script  204–5, 411, 425–6, 452–4, 465 Phoenicians, Phoenician influence  180–1, 195, 204–9, 425–6, 443–4, 449–52, 464–8

31/05/2016 18:10:47

[ 538 ] phylogenetics  352 phylogeography  352–7, 368 pidjinization  393 pig remains, Llanmaes  7, 308–9 Pinto-, Pento-, Palaeohispanic personal names  447–8 pit burials, Irish Chalcolithic  158– 60 pit burials, SW Iberia  183, 186–7 pit cremation, Welsh MBA  303, 315, 318 pius phonological rule  493 Plaza de las Monjas, Huelva  442–3 Plynlimon [Pumlumon] mining field  111 Pokorny, J.  409–12, 420, 424 polymerase chain reaction (=PCR)  352–3 Pontic–Caspian steppe  4–5, 366, 371–5 Pontrhydygroes halberd  73, 133 population structure, UK  336 Portable Antiquities Scheme in Wales  295, 312 postglacial population expansion  355–60, 365, 369, 372 post-processual archaeology  14 Pre-Basque language  508, 515 principle components analysis (=PCA)  333, 351 Proto-Celtic (=PC)  1, 3–5, 8–9, 34, 387–8, 395, 437–9, 442, 445, 466–9 Proto-Indo-European (=PIE)  1–9, 34, 431–4, 442 Proto-Irish language  219 Proto-Italo-Celtic  431–4 púmp- ‘5’ (Oscan)  494 pymp ‘5’ (MW)  494 quinque ‘5’ (Lat)  494 Quno (HC)  477–9 R1b, haplogroup  358, 369–71, 374 (T)  465 random mutations  352 rapa ‘heal’ (Sem)  465

raha

mynegai_CW3.indd 538

index RAPETIGVS MEDICVS CIVIS HISPANIS  464–5 RAPPA (Palaeohispanic)  464–5

Rathgall hillfort, Co. Wicklow  223–4, 242–5 Rathnagree hillfort, Co. Wicklow  221, 238–44 red hair [gene(s)]  336 relative constructions  413, 416, 421–3 Remedello culture  20, 44 remelting of bronze  253 Rh½s, J. 408, 438 Rhaeto-Etruscan languages & script  501 Rhesus blood group  336 Ría de Huelva deposition  187–8, 441–4, 450 Ría de Huelva fibulas  453 Ría de Huelva phase  443, 449–50, 468 Romano-British population  348 Romano-British Wales/finds in Wales  289, 298–300, 305, 309–11, 322 Roscommon phase  242–4 Rosnöen swords  187–8 Ross Island metal  133–4, 250, 253 Ross Island, mine  130, 133, 145–7, 152–3, 253 roundhouses  268–9, 272–9, 292, 297, 301–5, 310, 315–18, 322–3 Sabellian, contact with ProtoCeltic  9, 499–500 saeth ‘arrow’ (W)  439 saez ‘arrow’ (MB)  439 saiget ‘arrow’ (OIr)  439 Salamis, Cyprus, ‘royal’ tombs  454 samail ‘similar’ (OIr)  495 Samian Ware  300 Sanger sequencing  353 Sardinia, modern genetic data  360, 367–8, 370–2 Scandinavian rock art  199 Schulten, A.  179, 182–3, 207 second language acquisition  8, 395–6, 403

Seine valley/Paris Basin, Beaker sites  16, 19, 22–3, 27, 56, 91, 170 Setefilla, LBA stela  188–90 sewn-plank vessels  140 silo burials  182–6, 191 silver in prehistoric bronze  250–2 similis ‘similar’ (Lat)  495 Single Grave Culture  24, 27 Sion Le Petit-Chasseur, necropolis  19 sith- ‘long’ (OIr)  491 skin colour, genes  336 slabs as cist covers  199 smiths  15, 248, 261 SNPs (=single-nucleotide polymorphisms)  353 social domains  389–93, 398, 401 social lexicon  398 sociolinguistics  395 socius ‘ally’ (Lat)  493 socketed axes  299, 307–13, 320 Solana de Cabañas, stela & burial  186–7, 201 Sompting axes  295, 310–12, 319–22 South-west Orientalizing Complex (=SWOC)  6, 194–5, 204–9 South-western (=SW) script, inscriptions  1, 6, 8, 183, 202–9, 436, 445, 448–54, 466 spearheads, chemistry  257, 260–1 specialists, artisans  248–9 Spenser, E.  399  sruth ‘stream’ (OIr) 494 stable isotopes  14 Starčevo culture  357–8, 362–4 stelae, anthropomorphic and nonanthropomorphic  20–21, 96–7, 183–90, 195–205, 208–10, 440 sth~rá- ‘strong’ (Skt)  491 stone mining tools  112 STRs (=short-tandem repeats)  353 STRUCTURE-like clustering  353, 359, 370 stúrr ‘big, strong’ (ON)  491 substratum, linguistic  409–12, 416–21, 424–6, 431–4, 463 Suconia (HC)  477–9 sun-discs, gold  119–25 surnames and UK genetics  336–40, 348, 361

31/05/2016 18:10:47

index

súrr ‘sour’ (ON)  491 SVO order  409, 413, 417–24 swan’s-neck pins  313 swords, metal chemistry  7, 247–61 swords, social significance  7, 241–3, 247–9 syllabic nasals  495 system collapse  388 T1a, haplogroup  358, 366, 373 Tagus estuary, early Beakers  6, 67, 89, 94, 98, 102, 170 Talaconius (HC)  477–9 Talbenny, Pembrokeshire, grave with Beaker  49 Taršiš, Tarshish  179–81, 208, 443, 451  tartesius ager  182, 204–5 Tartessian language, inscriptions  1, 425–6, 445–50, 457–68 Tartessos, Ταρτησσος, ‘Tartessian’  6, 8, 179–83, 194, 204–10, 443, 451, 461–2, 465 teee-baarentii (T)  464 teir ‘3’ fem. (W)  494 téora ‘3’ fem. (OIr)  494 tholos burials  85, 88–9 tidres ‘3’ (Gaul) 494 tin bronze   171, 255 tláith ‘weak’ (OIr) 495 Toor More hillfort  221, 235–8, 243 *to¯t\ ‘tribe’  399 trackways in Bronze Age Ireland  144 trade language, Celtic as?  391–3 Trask, L.  508, 515, 518 TRB (=Funnel Beaker) culture  360, 366–7 Treanbaun, Co. Galway, pit burial  160–2 Tre’r Ceiri hillfort  266–7

mynegai_CW3.indd 539

Trevisker pottery  303–4, 314–18, 323 Tubney Wood, Oxfordshire, disc  124 tunbiitesbaan (T)  465 Tyre, Tyrians  443–4, 449, 453–4, 468 U, mtDNA haplogroup  356–7, 363, 366 uarbaan (T)  465 uarbaan tee-ro-baare baanaŕkeentii (T)  460–1, 464–5 uiŕe (Ibericized)  477 *¯īro- ‘true’ (PC)  479 uiroku (Ctb) 478–9

*¯iros ‘man, hero’ (PC)  477–9, 490 umbil¼cus ‘navel’ (Lat)  495 Únětice culture  367–8 univallate hillforts, Irish  220 úr ‘fresh’ (OIr)  491 Uralic languages  507–8 urn burials  189–94 Urnfield culture  4, 9, 187–9, 203–4, 222, 500 V-perforated buttons  165, 168 Vaga da Cascalheira, EIA cist burials  193 val{re ‘to be strong’ (Lat)  495 Vasconic family tree  505 Vasconic influence on Germanic and Italic  509–10, 528 Vasconic substratum  9, 431–7, 505–9, 513, 516–28 Vasconic Theory  9, 504, 511, 515 Venetic, similarity to Celtic  9, 500 veni} ‘I come’ (Lat)  494 verbal noun  408–11, 418, 423–4

[ 539 ] Viereckschanzen  265, 280 vigesimal counting  506–7 Viking, settlement/ancestry  333–4, 361 vir ‘man’ (Lat)  490 v¼rá- ‘hero’ (Skt)  490 voiced aspirated stops, PIE  432–3, 496–8 VSO order  407–13, 416–17, 421–4 Wagner, H.  411–12, 420, 424 Wales, population genetics  344–6 Wales, surnames of   340 warband  400 warrior burials  22–5 warrior culture/ideology  13, 22–3, 33, 241–4, 247–8, 400, 440 warrior élite  222 warrior stelae  440–1, 448–51, 468 Watkins, C.  489 Wedge Tombs  53, 85, 151–6, 167, 170 Welsh language  408–13, 417–20, 423–4, 432 Wenglish  432 wer ‘man’ (OE)  490 wheels in Bronze Age Ireland  144 *wiH-ró- (PIE)  490 Wilburton period  257–8 Wilde, Sir William  249 *wiro- (IC)  490 word-initial accent   493, 506–9 Y chromosome  336–8, 352–3 Yamnaya, culture/package  369, 373–5 yards [enclosures]  281–4, 291 yn (W)  408, 413, 418–20, 424 yokes in Bronze Age Ireland  144 Younger Dryas  348

31/05/2016 18:10:47