The ogham inscriptions of Scotland: An edited corpus

HARVARD UNIVERSITY THE GRADUATE SCHOOL OP ARTS AND SCIENCES THESIS ACCEPTANCE CERTIFICATE The undersigned, appointed b...

Author: Katherine Stuart Forsyth

87 downloads 4561 Views 18MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

HARVARD UNIVERSITY THE GRADUATE SCHOOL OP ARTS AND SCIENCES

THESIS ACCEPTANCE CERTIFICATE

The undersigned, appointed by the Division Department of C e l t i c Languages and L i t e r a t u r e s Committee

have examined a thesis entitled The Ogham Inscriptions of Scotland: An Edited Corpus

presented by Katherine Stuart Forsyth candidate for the degree of Doctor of Philosophy and hereby certify that it is worthy of acceptance. Signat£™ .T. Typed name .Jqb.n. T,. JCp.ch.. Signature ... Typed name

Patrick.K,. ..F.or.d...

Signature .... Typed name

Date ....April. 9......1.95.6....

The Ogham Inscriptions of Scotland: An Edited Corpus

A thesis presented by

Katherine Stuart Forsyth to The Department of Celtic Languages and Literatures in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

in the subject of

Celtic Languages and Literatures

Harvard University Cambridge, Massachusetts

April 1996

UMI Number:

9631492

UMI Microform 9631492 Copyright 1996, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI

300 North Zeeb Road Ann Arbor, MI 48103

® 1996 by Katherine Stuart Forsyth All rights reserved

ABSTRACT

The Ogham Inscriptions of Scotland: An Edited Corpus

An illustrated catalogue of all inscriptions in the ogham script from Scotland. Thirtyseven are extant, a further three are documented but lost. They range in date from at least the sixth century AD to the tenth. Each catalogue entry contains a description of the inscribed object, its archaeological and historical context, a detailed description of the inscription, comments on the form of the script, notes towards an interpretation of the text, previous readings, and a select bibliography. It is argued that the division into 'Pictish' and 'Irish-looking' on the basis of form of script is unhelpful, and that the script of Scottish oghams is comparable to that of contemporary oghams in Ireland. The possibility of a 'palaeography' of ogham is raised. There are very few Irish oghams from the early 'post-classical' period (later seventh and eighth century), and Scottish oghams provide important evidence for this intermediate stage in the development of the script. Irish and Scottish oghamists were working within a single tradition, innovations, such as additional letters, were passed back and forth. In some instances the orthography used in Scotland diverges from that of contemporary Ireland, e.g. perpetuation of older orthographical conventions (simplex/geminate consonant distinction), innovative use of 'H'. A key difference is that post-classical ogham in Ireland had only marginal status, whereas in post-seventh century Scotland, ogham was a prestige script used on grand

public monuments such as cross-slabs. The wide range of contexts in which ogham was used there, especially informal and domestic contexts, suggests knowledge of the script was not restricted. Scottish oghams present difficulties of interpretation because many are fragmentary and weathered, the value of some supplementary characters is not certain, orthographical conventions are imperfectly understood, the language of the texts (Goidelic/Brittonic) is not always known, word-division is rarely indicated, often the kind of text to be expected is not known. More texts can be interpreted than previously thought, practical problems go a long way to explaining why others resist interpretation. Jackson's theory of a second non-Indo-European Pictish language is not upheld.

349 words

CONTENTS

TABLE OF CONTENTS

i

ACKNOWLEDGEMENTS

x

DEDICATION

xiii

PREFACE

xiv

LIST OF ILLUSTRATIONS

xvi

ABBREVIATIONS

xxvii

authorities and general (xxvii); standard abbreviations for the pre-1974 counties of Scotland (xxx); other abbreviations and conventions (xxxi)

INTRODUCTION

xxxii

THE OGHAM SCRIPT DESCRIBING OGHAM TRANSLITERATING OGHAM

xxxii iii

Transliteration of 'H' (viii); Transliteration of 'V (viii) TRANSCRIBING AN OGHAM INSCRIPTION

ix

THE PALAEOGRAPHY OF OGHAM

xi

THE TERM 'SCHOLASTIC

xviii

THE FORFEDA

xviii

A NOTE ON WORD-DIVISION

xxi

OGHAM IN SCOTLAND THE STUDY OF SCOTTISH OGHAM

Ix

THE CORPUS

lxiv

THE TEXTS

lxix

THE INTRODUCTION OF OGHAM TO SCOTLAND

lxxiii

SCHEDULE OF READINGS

lxxvi

CONTENTS

ii

GUIDE TO ENTRIES CATALOGUE OF THE OGHAM INSCRIPTIONS OF SCOTLAND ABERNETHY

lxxx I 2

discovery (2); site (2); description of object (4); description of carving (5); description of ogham inscription (7); form of script (8); interpretation of text (9); discussion (9); bibliography (10); previous readings (10)

ACKERGILL

11

discovery (11); site (11); description of object (13); description of carving (13); description of carving (symbols) (14); description of ogham inscription (15); form of script (17); interpretation of text (17); discussion (20); bibliography (22); previous readings (22)

ALTYRE

23

discovery (23); locality (25); description of object (27); description of carving decoration (28); description of carving - ogham inscription (29); form of script (35); interpretation of text (37); discussion (39); bibliography (40); previous readings (40)

AUQUHOLLIE

41

discovery (41); site (42); description of object (43); description of carving (43); description of ogham inscription (44); form of script (47); interpretation of text (48); linguistic discussion (50); general discussion (53); bibliography (54); previous readings (54)

BAC

MHIC

CONNAIN

discovery and current location (55); site (55); description of object (58); description of carving (59); form of script (61); interpretation of text (61); discussion (65); bibliography (67); previous readings (68)

55

iii BIRSAY

69

site (69); Early Christian sculpture from the Brough of Birsay (71); runic inscriptions from the Brough of Birsay (72); select bibliography (site only) (74); ogham inscriptions from the Brough of Birsay (74); discovery (74); BIRSAY 1 (76); description of object (76); carving technique (77); description of ogham inscription (77); form of script (80); interpretation of text (80); bibliography (81); previous readings (81); BIRSAY 2 (81); description of object (81); description of carving (82); rbrm of script (84); interpretation of text (85); bibliography (85); previous readings (85); BIRSAY 3 (85); description of object (86); description of carving (86); form of script (88); interpretation of text (89); bibliography (91); previous readings (91); general discussion of the Birsay oghams (91)

BLACKWATERFOOT

1-2

93

site (93); discovery (94); description of carving - OGHAM INSCRIPTION I (95); interpretation of text (96); description of carving - OGHAM INSCRIPTION 2 (97); interpretation of text (98); form of script (99); discussion (100); bibliography (100); previous readings (101)

BRANDSBUTT

102

discovery (102); site (102); description of object (103); description of carving technique (104); description of carving - decoration (105); description of carving - ogham inscription (109); form of script (112); interpretation of text (112); discussion (115); bibliography (115); previous readings (116)

BRESSAY

117

discovery (117); site (118); description of object (118); description of carving decoration (119); description of ogham inscription (122); form of script (126); interpretation of text (129); discussion (136); bibliography (137); previous readings (138)

B ROD IE discovery (139); locality (140); description of object (141); description of carving - decoration (142); description of ogham inscription (145); form of

139

iv

script (155); interpretation of text (156); discussion (157); bibliography (158); previous readings (158)

BUCKQUOY

160

discovery (160); site (160); description of object (164); carving technique (164); description of ogham inscription (167); fonn of script (174); interpretation of text (175); discussion (177); parallels (181); historical context (183); conclusion (185); bibliography (186); previous readings (186)

BURRIAN

187

discovery (187); site (187); description of object (189); description of carving (190); description of decoration (190); ogham inscription (192); fonn of script (198); interpretation of text (199); discussion (203); bibliography (204); previous readings (205)

CUNNINGSBURGH

1-3

discovery (206); site (207); runic inscriptions (208); bibliography

206 (209);

ogham-inscribed fragments-CUNNINGSBURGH 1 (209); description of object (209); description of carving (210); description of ogham (211); fonn of script (212); interpretation (212); bibliography (212); previous readings (212); oghaminscribed fragments - CUNNINGSBURGH 2 (213); description of object (213); description of carving (213); form of script (216); interpretation of text (217); bibliography (218); previous readings (219); ogham-inscribed fragments CUNNINGSBURGH 3 (219); description of object (219); description of carving (219); fonn of script (222); interpretation (223); bibliography (225); previous readings (225); the Cunningsburgh oghams - discussion (225)

DUNADD discovery (227); site (227); further indications of the use of literacy at dunadd (228); roman alphabet inscription (229); rock-carvings (230); description of ogham carving (230); form of script (236); interpretation (236); discussion (239); bibliography (242)

227

D UPPLIN

243

discovery (243); locality (244); description of object (246); description of carving - decoration (247); iconography (249); the ogham inscription (249); bibliography (252); the roman alphabet inscription (252); reading and interpretation of text (253); historical context - Custantin filius Fircus (257)

FORMASTON

261


G IG HA

1 (KILCHATTAN)

288

current location (288); history (288); site (289); description of stone (290); description of carving (291); form of script (295); interpretation of text (296); discussion (297); bibliography (298); previous readings (298)

GOLSPIE

299

discovery (299); locality (300); description of object (301); description of carving - decoration (302); description of carving - ogham inscription (307); form of script (312); interpretation of text (313); discussion (318); bibliography (319); previous readings (320)

GURNESS

321

discovery (321); site (321); description of object (323); carving technique (324); description of ogham inscription (324); form script (328); interpretation of text (329); discussion (331); bibliography (332); previous readings (332)

INCHY RA discovery (333); locality (335); description of object (336); description of carving - symbols (337); relationship of symbols and order of phases (340); relationship of symbols (to ogham inscriptions) (342); description of ogham inscriptions (344); form of script (352); interpretation of text (354); discussion (358); bibliography (359); concordance (359); previous readings (359)

333

vi LATHERON

360

discovery (360); site (360); description of object (361); description of carving decoration (362); description of carving - ogham inscription (364); form of script (367); interpretation of text (368); discussion (372); bibliography (373); previous readings (373)

LOCHGOILHEAD

374

current location (374); site (374); description of stone (374); description of carving (376); roman alphabet inscription (376); description of ogham inscription (377); form of script (379); interpretation of text (379); discussion (382); bibliography (384); previous readings (384)

LOGIE

ELPHINSTONE

385

discovery (385); site (385); historical context (386); description of object (390); carving technique (390); description of carving - symbols (392); description of ogham inscription (393); form of script (396); interpretation of text (398); discussion (399); bibliography (400); previous readings (401)

LUNNASTING

402


NEWTON

420

discovery (420); locality (422); description of object (424); carving technique (424); description of ogham inscription (426); form of script (431); interpretation of text (432); the non-ogham inscription (437); the Newton symbol stone (439); discussion (440); bibliography (441); previous readings of ogham (442)

POLTALLOCH current location (443); discovery (443); site (444); locality (447); description of object (448); description of ogham inscription (448); form of script (449);

443

CONTENTS

Vll

interpretation (450); discussion (453); bibliography (454); previous readings (455)

POOL

456 discovery (456); site (456); description of object (458); carving technique (459); description of ogham inscription (460); fonn of script (461); interpretation of text (463); discussion (465); bibliography (466); previous readings (466)

ST. N I N I A N ' S

ISLE

467

discovery (467); site (467); roman alphabet inscription (471); description of ogham-inscribed object (471); description of ogham inscription (472); fonn of script (474); interpretation of text (476); discussion (478); bibliography (479); previous readings (479)

S COON IE

480

discovery (480); site (480); description of object (480); description of carving decoration (481); description of carving - ogham inscription (483); fonn of script (486); interpretation of text (486); discussion (491); bibliography (493); previous readings (494)

WHITENESS

495

discovery (495); site (495); description of object (496); description of carving decoration (497); description of ogham inscription (498); fonn of script (500); interpretation of text (501); discussion (501); bibliography (502); previous readings (502)

APPENDIX DUBIA

ABERNETHY

503

2

discovery (504); description of object (504); description of carving (504); discussion (505); bibliography (506); previous readings (506)

504

viii

CONTENTS BROUGH

OF

BIRSAY

507

discovery (507); description of object (507); description of carving (507); discussion (508); bibliography (508)

BURRIAN

509

discovery (509); description of object (509); description of carving (509); bibliography (510)

DRUMOYNE

511

FOSHIGARRY

512

discovery (512); description of object (512); description of carving (512); bibliography (513)

GIGHA

2

514

current location (514); description of stone (514); description of carving (514); discussion (514); bibliography (515)

G U RN ESS

516

discovery (516); description of objects (516); bibliography (517)

IONA

518

L OCHNAW

519

discovery (519); description of object (519); description of carving - crosses (519); description of carving - 'ogham' (520); discussion (523); bibliography (525)

POOL

526 discovery (526); description of object (526); description of carving (526); bibliography (527)

TOLLARD

HOUSE

528

CONTENTS

ix

VAIVOE

529

OTHERS

530

BIBLIOGRAPHY OF WORKS CITED

531

Envoi

NB. Bound in two volumes with continuous pagination:

Volume I i-lxxxvi, 1-226 Prefatory Matter; Introduction; Guide to Entries; Catalogue from Abemethy to Cunningsburgh

Volume 2 227-553 Table of Contents (Vol. II), Catalogue from Dunadd to Whiteness, Appendix: Dubia, Bibliography

X

ACKNOWLEDGEMENTS

Many individuals and institutions have provided generous support to me, financial, logistical, intellectual, or emotional, during my eight years in tlie doctoral program of tlie Department of Celtic Languages and Literatures, Harvard University. It is impossible to name all those who have offered help and advice, but I am particularly indebted to tlie providers of major financial assistance, and I wish to record my gratitude to them at the outset.

Without a scholarship from tlie St. Andrews Society of tlie State of New York I could never have come to Harvard in tlie first place, and without tlie generous financial aid of tlie University itself, I could not have stayed. My first few months back in Scotland were partly financed by Harvard in the form of a Jens Aubrey Westengard Travelling Scholarship, a Radcliffe Grant for Graduate Women, and a Travelling Scholarship from the Dept. of Celtic Languages and Literatures.

The difficulty of juggling paid

employment and dissertation work in tlie following years made me all tlie more appreciative of tlie chance to devote my energies to full-time study during my tenure of tlie Sir John Rh£s Memorial Studentship in Celtic Studies, in tlie University of Oxford, and subsequently as a Julia Mann Junior Research Fellow at St. Hilda's College, Oxford. I wish also to record my gratitude for tlie smaller grants I received for specific purposes; tlie grant from tlie St. Andrew's Society of Washington D.C. which enabled me to purchase the computer on which all the work for this project was done, the grant from tlie President's Fund of tlie Edinburgh Association of University Women which helped defray the cost of photographs and other illustrations and of printing and binding tlie dissertation, and the grant from tlie Society of Antiquaries of Scotland which enabled me to travel to Orkney to conduct fieldwork.

I have lived in four countries while working on this project and have been fortunate in the support I have received from various institutions, some of which have even become for a while my home. The Department of Celtic in die University of Edinburgh allowed me to affiliate as a post-graduate worker and

ACKNOWLEDGEMENTS

xi

thus provided a lone scholar with logistical support and a stimulating intellectual and social environment. My colleagues in the Special Collections Department of the Library of the University of Edinburgh made earning my daily bread so much less of a chore. In Oxford I have indeed been fortunate, Jesus College and St. Hilda's College have given much and asked little in return. It has been a privilege and a pleasure to be a member of both institutions.

The staff of the NMR in Edinburgh were most helpful while I was conducting archival research there and I am particularly grateful to Graham Ritchie and David Easton for their assistance. The field-work for this study, which was conducted in Ireland and throughout Scotland, would have been impossible without the friendly co-operation of land-owners, church key-holders, and museum attendants. I wish to record my particular gratitude to the following curators who allowed me access to ogham inscriptions in their collections or otherwise assisted me: Colleen Batey (Kelvingrove Museum and Art Gallery), Anne Bnindle (Tankeraess House Museum), Mike King (Perth Art Gallery and Museum), Bernard Meehan (Library of Trinity College, Dublin), Raghnall 6 Floinn (National Museum of Ireland), Siobhan Ratchford and Alfred Truckell (The Observatory, Dumfries), Susan Youngs (British Museum), and especially Michael Spearman (NMS). Christopher Morris enabled me to examine the Birsay 3 ogham and I wish to thank him, John Hunter, and Anna Ritchie, for providing me with photographs and information about the oghams they excavated. Other practical assistance came from Marianna Lines, Niall Robertson, Tom Gray, and other members of the Pictish Arts Society.

The Department of Celtic Languages and Literatures has been very tolerant of a student who was, for five of her eight years, in absentia and I wish to record in particular my gratitude to my Chairman, Patrick Ford. I am ever more aware of my debt to those who taught me at Harvard - John Carey, John Koch, and Nerys Patterson - and to my fellow graduate students in the Departments of Celtic and of History. As a supervisor John Koch provided humour, support, and, above all, freedom. It was he who first encouraged

ACKNOWLEDGEMENTS

xii

me to pursue this topic and for this, as for so much else, I am very grateful to him.

Perhaps the most pleasurable aspect of working on this project is the many friendships it has spawned. It would be impossible to name all those who have sustained me through their friendship, and fostered my general intellectual development over the last eight years but I am acutely conscious of my debt to them. I have endeavoured to acknowledge all specific suggestions where appropriate in the text, but in addition [ wish to record here my profound gratitude to the following friends and colleagues who have given particularly generously of their time and expertise to offer advice and encouragement: Dauvit Broun, Thomas Charles-Edwards, Thomas Clancy, Stephen Driscoll, Joe Eska, Ian Fisher, Anthony Harvey, Isabel Henderson, John Higgitt, Jim Livesey, Damian McManus, Fionnbarr Moore, Elisabeth Okasha, Oliver Padel, Ross Samson, Richard Sharpe, Simon Taylor, Charles Thomas, Alex Woolf, and Patrick Wormald.

I have left till last my greatest debt, to my parents Moira and Alastair Forsyth. Their unerring support has been financial, logistical, intellectual, and emotional. This work could have been neither begun nor concluded without them, and it gives me the greatest pleasure to be able, at last, to dedicate it to them, with love.

K. S. F.

St. Hilda's College, Oxford March 1996

xiii

To my parents

xiv

PREFACE

My aim in this study has been, not to write a history of ogham in Scotland, but rather to provide the materials from which such a history might be written. Included are all inscriptions in the ogham alphabet from within the boundaries of modem Scotland, regardless of language of text or supposed ethnicity of carver. This marks a major departure from previous studies which have separated out supposedly Tictish' inscriptions from 'Irish-looking'. With the exception of Blackwaterfoot and Gigha, I have examined all inscriptions personally. In addition, I have visited the find-spot or original location of most of the oghams from Mainland Scotland and Orkney.

The main part of the corpus consists of individual entries, one for each site which has produced an ogham inscription. As well as providing a description of die inscribed object, a detailed account of its inscription, comments on the form of the script used, and notes towards an interpretation of the text, I have attempted to provide a broad historical and archaeological context for each ogham. I feel this too, is an innovation, one in keeping with the new topographically-aware Insular epigraphy championed by Charles Thomas. Those currently active in the field are increasingly coming to see inscribed stones, not merely as vehicles for text, but as monuments in a social and historical landscape. To view the ogham inscriptions of Scotland from this perspective engenders a radical reappraisal of their importance for our understanding of the language and culture of early medieval north Britain.

The last scholar to address this topic was Oliver Padel, whose 1972 University of Edinburgh MLitt. thesis, Inscriptions ofPialand, has remained the final word on the subject for almost a quarter of a century. The present readings were made independently of PadeFs, but were checked against his (and those of previous scholars). I found his readings uniformly accurate and sensitive, and only rarely did I find myself in disagreement. As a master's thesis, Inscriptions ofPictland was necessarily restricted in scope. Padel provided a thorough examination of each inscription, but, quite naturally, did not stray from the

PREFACE

xv

interpretive framework set out for ogham by his supervisor Kenneth Jackson. Padel did not question, for instance, the traditional model of the Dalriadic import of ogham into Scotland and the separate development of the script there, both assumptions which I would argue are no longer tenable.

Included in the present study are eight oghatns not covered by Padel. Two, Gigha and Poltalloch, which he excluded as not 'Pictish',1 and six which have come to light since he wrote: Blackwaterfoot 1-2, Birsay 3, Dupplin, Lochgoilhead, and Pool. More important, however, than new finds, are die twin revolutions in ogham scholarship and Pictish studies which have occurred over the last decade. Work done by McManus and Harvey from the mid-1980s, and more recently by Sims-Williams, has transformed our understanding of ogham. The script has been divested of its cryptic and druidic overtones and, far from being a crude and clumsy cipher, has come to be appreciated as a sophisticated indigenous response to the challenge of roman alphabet literacy. In Scotland, archaeologists, notably Leslie Alcock and his pupils, have led the way forward to a de-mystification of the Picts, who are now increasingly seen as a Celtic people, typical among their barbarians neighbours on the northern and western fringes of early medieval Europe.

My aim has been to bring to bear on Scottish ogham these new approaches, to look at the inscriptions again from a fresh perspective, and thereby to challenge the traditional orthodoxy on the Picts, Pictish, and 'Pictish' ogham. The technical complexity of the material means the work is necessarily at an interim stage. Any conclusions remain tentative. In the Introduction which follows I make some brief general observations and point to a number of ways in which future research might develop. I have not set out to produce a 'Grammar of Pictish' or to write another chapter in the history of ogham. I hope, however, that the present work has laid the ground-work which will bring such goals a little closer.

1 He included two oghams, Weeting (Norfolk) and Kirkmichael (Isle of Man), which I have not treated since they arc not in Scotland. I would argue, in any case, argue that the label 'Pictish' is inappropriate for them. I have also excluded the six roman alphabet inscriptions from Pictland mentioned by Padel.

xvi

ILLUSTRATIONS

Illustrations (unpaginated) are grouped together at tlie end of tlie relevant entry.

Unless otherwise

specified, all maps, drawings, and photographs are by the author. I am grateful to Frances Hood, John Hunter, Christopher Morris, Anna Ritchie, and especially to Michael Spearman (NMS) for providing photographs and, where appropriate, giving permission for their use. Unless otherise stated as schematic, drawings of inscriptions are from photographs and are to scale. The numbers in such diagrams correspond to numbered sections in tlie discussion of each inscription.

INTRODUCTION

Transliteration key to the ogham alphabet [after Thomas 1994 fig. 3.1] The 'Ogham-band' The Palaeography of Ogham (explanatory diagrams): Slope, profile, spacing, special characters, etc.

LOCATION MAPS: The ogham inscriptions of D41 Riada The ogham inscriptions of Orkney Tlie ogham inscriptions of the eastern Mainland (south) The ogham inscriptions of tlie North-East Caithness and Sutherland. Location of sites mentioned in text [after Batey 1991 fig. 1] The ogham inscriptions of Shetland

CATALOGUE

Location Map. The ogham inscriptions of Scotland

ILLUSTRATIONS

xvii

ABERNETHY * Conjectural map of the early medieval shire of Abernethy [Driscoll 1991 fig. 5.3] * Ogham-inscribed fragment [Butler pi.II]; (key to numbering) * Side view of recumbent grave-marker, Meigle 11 [ECMS fig.345A] * Possible reconstruction of monument

ACKERGILL * Plan of cemetery [Close-Brooks 1984 fig.5.6] * Ogham-inscribed Symbol Stone [ECMS fig. 25] * Schematic representation of the ogham (key to numbering)

ALTYRE * Location map. Moray and Easter Ross showing extent of areas of prehistoric and proto-historic settlement [Jones et al. 1993:48 fig.3.1] * Pictish Moray [after Shepherd 1993:77 fig. 4.2] * Ogham-inscribed cross and detail of inscription [Calder & Jackson 1957 figs. 14, 15] - Ogham inscription, based on impression by Marianna Lines (plant matter on cotton) [reduced]; (key to numbering)

AUQUHOLLIE * Ogham-inscribed pillar (detail) * Ogham-inscribed pillar * Ogham inscription (key to numbering)

BAC MHIC CONNAIN -

Site Location maps. Above: Bac Mhic Connain and the wheelhouse sites of Vallay Sound [Hallén

ILLUSTRATIONS

xviii

1994:190 Illus. 1]. Below: As above showing relief [Scott 1948:69 fig.7J. •

Ogham-inscribed knife handle [Hallén 1994:220 Illus. 12.5]

•

Plans of the structures at Bac Mhic Connain [Hallén 1994:191 Illus.2]

•

Ogham-inscribed knife handle [® Trustees of the National Museums of Scotland, Neg. P9557]

-

Ogham inscription (key to numbering)

BIRSAY •

Location Map. Orkney with inset Brough of Birsay and Point of Buckquoy [after Morris 1990:70 fig.3]

•

Plan of main Pictish and Norse settlement [Curie 1982 ill. 3]

•

Sculptured slab with Pictish symbols and figures [Curie 1982 ill. 4]

•

Cross-incised stones: Large [Ritchie 1986:12], small [Curie 1982 ill.45 (detail - no.607)] Birsay 1

- Rubbing of ogham inscription [Photo. Department of the Environment (Historic Scotland) No. A 2894-2, June 1971 (Crown Copyright)] • Ogham inscription [Padel 1972:56] • Schematic representation of ogham (key to numbering)

Birsay 2

• Rubbing of ogham inscription [Photo. Department of the Environment (Historic Scotland) No. A 2894-1, June 1971 (Crown Copyright)] • Schematic representation of ogham (key to numbering)

Birsay 3

* Ogham-inscribed slab, upper and lower feces [photo, supplied by Prof. C D . Morris, Dept. of Archaeology, University of Glasgow, Reg. BB80] • Ogham-inscribed slab,frontand back feces [photo, supplied by Prof. C. D. Morris, Dept. of Archaeology, University of Glasgow, Reg. BB80] • Ogham inscription • Schematic representation of ogham (key to numbering)

•

Rune-inscribed fragments, Birsay II (Barnes OR 8) [photo. Historic Scotland, Neg. A 995-1

xix

ILLUSTRATIONS (1938), Crown Copyright] •

Rune-inscribed fragment, Birsay III (Barnes OR 9) [photo. Historic Scotland, Neg. A 995-3 (1938), Crown Copyright]

•

Rune-inscribed seal's tooth, Birsay IV (Barnes OR 11) [Curie 1982:ill.37]

BLACKWATERFOOT •

King's Cave - Exterior of King's Cave [Balfour 1910 pl.XXXIII. I]; Plan of interior showing area of Balfour's excavation [Balfour 1910.214]

• Interior of cave showing position of ogham inscriptions [®RCAHMS Neg. B69005] Blackwaterfoot 1 - Ogham inscription [Jackson 1971 pi. IX]; Schematic representation of ogham inscription (key to numbering) Blackwaterfoot 2 • Ogham inscription [®RCAHMS Neg. B69011]; • Ogham inscription; Schematic representation (key to numbering)

BRANDSBUTT •

Ogham-inscribed Class I Pictish Symbol Stone [®University of Aberdeen, Anthropologcial Museum]

• Ogham-inscribed Class I Pictish Symbol Stone • Ogham inscription (key to numbering)

BRESSAY •

Ogham-inscribed cross-slab [Close-Brooks & Stevenson 1982:35]

• Ogham-inscribed cross-slab [n.a. 1865, pl.XII, based on Stuart 1856, pl.XCIIII-XCV] •

Cross-slab from Papil, Shetland [^Trustees of the National Museums of Scotland, Reg. IB 46]

• Ogham inscription on left edge [^Trustees of the National Museums of Scotland, Reg. IB 119] (key to numbering)

ILLUSTRATIONS

XX

• Ogham inscription on right edge [^Trustees of the National Museums of Scotland, Reg. IB 119] (key to numbering)

BRODIE • Ogham-inscribed cross-slab (reverse face) [Foster 1996 fig.48 (Tom Gray)] • Ogham-inscribed cross-slab showing position of inscriptions [ECMS fig. 136] • Front right: schematic representation of ogham inscription (key to numbering) • Back left: Detail of ogham (key to numbering) [After Foster 1996 fig.48] •

Back right: Detail of ogham (key to numbering) [After Foster 1996 fig.48]

BUCKQUOY •

'Figure-of-eight* house in which inscribed whorl was found [Ritchie 1977:177 fig. 3]

-

Ogham-inscribed spindle-whorl [Crown Copyright: Institute of Geological Sciences, NERC].

-


BURRIAN • Location map [MacGregor 1974:64 fig.l] •

Broch site [RCAHMS 1946 fig.88]

•

Ogham-inscribed slab [^Trustees of the National Museums of Scotland, Neg. P9554]

•

Ogham-inscribed cross-slab [MacGregor 1974:97 fig.21

-


-

Detail of cross

CUNNINGSBURGH • Location Map [Turner 1994:316 Illus.lJ Cunningsburgh 1 • Ogham-inscribed fragment [^Trustees of the National Museums of Scotland,

ILLUSTRATIONS

xxi Reg. IB 114] * Ogham-inscribed fragment, front and back; Schematic representation of the ogham (key to numbering)

Cunningsburgh 2 • Ogham-inscribed fragment [^Trustees of the National Museums of Scotland, Reg. IB 115] * Ogham-inscribed fragment; Schematic representation of the ogham, face and arris (key to numbering) Cunningsburgh 3 ' Ogham-inscribed fragment * Schematic representation of the ogham (key to numbering) * Macalister's reconstruction [Macalister 1940.216 pl.VIb]; Alternative reconstruction * Runic inscriptions. Above: Cunningsburgh 1 [ECMS fig. 14] Below: Cunningsburgh 3 [ECMS fig. 15]

DUNADD -

Plan of fort [Campbell & Lane 1993.53]. Rock carvings, including ogham inscription, at location a

* Incised rock surface with detail of boar [RCAHMS 1988:158] * Incised boar [RCAHMS 1988:157] * Incised rock surface with detail of ogham inscription [RCAHMS 1988:158 and RCAHMS 1988:159] * Ogham inscription, upper and lower lines (key to numbering) * Inscribed pebble (enlarged) [Okasha 1985 pl.VIII] * Detail of inscription

ILLUSTRATIONS

xxii

DUPPLIN * Location of the Dupplin cross in relation to Forteviot. D, Dupplin Cross; H, Haly Hill; S, scarp of the former course of the Water of May; I, site of Invermay cross. Hatching, areas of cropmarks [Alcock & Alcock 1993 illus. 12] * Free-standing cross [Alcock & Alcock 1992: illus. 13-14]. * Romilly Allen's drawing of the Dupplin Cross [ECMS fig.334]. • The base of the Dupplin Cross showing traces of ogham lettering [® The Trustees of the National Museums of Scotland |. -

Inscribed panel on the west lace of the shaft of the Dupplin cross [® The Trustees of the National Museums of Scotland].

• Text panel (cast) with lettering marked with charcoal - lines 1-2. • Text panel (cast) with lettering marked with charcoal - lines 1-3, 7. • Sculptured stones from Forteviot (not to scale). Fragment of cross-arm [Alcock & Alcock 1992 illus.5]; Front, side, and rear view of cross-slab Forteviot 1. Height 0.6m [Alcock & Alcock 1992 illus.4 (Tom Gray)]; Arch [Alcock & Alcock 1992 illus.6]

FORMASTON * Fragment of ogham-inscribed cross-slab [^Trustees of the National Museums of Scotland, Reg. IB 111] •

Reconstruction of original monument [Macalister 1940:213 pl.v]

- Cross-slab from Kinnord, Aberdeenshire. - Ogham inscription, inner and outer lines (key to numbering)

GIGHA • Ogham-inscribed pillar [RCAHMS 1971 pi. 16]

ILLUSTRATIONS

xxiii

* Comparison of previous readings: (a) Macalister 1902; (b) Diack 1926; (c) Jackson (RCAHMS) 1971; (d) Key to numbering of letters

GOLSPIE * Ogham-inscribed cross-slab [Close-Brooks 1989:cover, 9] * Ogham-inscribed cross-slab. Side view showing ogham [ECMS 48A] -


GURNESS * Plan of the Post-Broch period structures [Hedges 1987.11:64 fig.2.11] -

Plan of die 'Shamrock* and 'Annexe* [Hedges 1987.11:66 fig.2.12]

-

Carved stones recovered from die site. Above: Pictish symbol stone. Below: Cross-incised slab, unknown date. [Hedges 1987.11:125 fig.2.51, 126 fig.2.52]

* Ogham-inscribed knife handle [Hedges 1987.11:96 figs. 2.22, 118 fig.2.44, details] * Ogham inscription (key to numbering)

INCHYRA -

Symbol and ogham-incised slab [RCAHMS 1994a:95a]

-

Symbol and ogham-incised slab showing position of inscriptions [Stevenson 1959 pi.Ill]

-

'Unfinished' symbols and inscription A (i) [after Stevenson 1959 pi.IV] (key to numbering)

* Detail of damaged edge showing inscription A (ii) - key to numbering [photograph ® Perth Art Gallery and Museum] * Inscription B: Top: Edge and top of slab showing position of lettering [Stevenson 1959 pi. Ill]; Middle: Detail of inscription (key to numbering) [RCAHMS 1994a:93]; Bottom - Detail of lettering on edge [Henderson 1967 pi.24] * Detail of broad end of slab showing symbols and inscription C (key to numbering) [Stevenson

ILLUSTRATIONS

XXIV

1959 pl.IIII

LATHERON • Ogham inscribed cross-slab [Curie 1940 pi.xxii] • Ogham-inscribed cross-slab • Ogham inscription (key to numbering)

LOCHGOILHEAD Faces b and c [photograph RCAHMS Argyll.VII.194 B,C] Faces a-c [drawing RCAHMS Argyll.VII. 194 A(l)] Rubbing of face b (reduced to 60% actual size) Selection of roman letters (actual size) Rubbing of edge c (reduced to 60% actual size) Schematic representation of ogham inscription (key to numbering) Ogham inscription (actual size)

LOGIE ELPHINSTONE • Class I Pictish Symbol Stones at Logie Elphinstone: Logie Elphinstone 1 [ECMS fig. 188], Logie Elphinstone 3 [ECMS fig. 191] •

Impression by Marianna Lines (plant matter on cotton) [Lines 1989]

• Detail of ogham inscription (key to numbering)

LUNNASTING •

Ogham-inscribed slab [Goudie 1878 fig.2 (=1901:36); ECMS fig.12]

• Ogham text with key to numbering

ILLUSTRATIONS

XXV

NEWTON •

Location Map. OS First Edition, Sheet 44 (detail), showing Newton House, Bridgend of Shevack, Wood of Pitmachie, Bridge of Pitmachie, and Old Rayne. Scale - 6W to the mile.

• Class I Pictish Symbol Stone [ECMS fig. 193] •

NMS Cast of ogham-inscribed pillar [® Trustees of the National Museums of Scotland, Reg. IB 108]; (key to numbering)

•


•

Detail of roman alphabet inscription [Simpson 1940 pi. 60]

-

Roman alphabet inscription

POLTALLOCH •

Ogham-inscribed fragment [®Tmstees of the National Museums of Scotland, Neg. P9555]; (key to numbering)

POOL •

Location map [after Hunter 1990:176 ill. 10.1]

•

Symbol-inscribed stone, J. R. F. Burt [Nicoll 1995:179 fig.31]

•

Ogham-inscribed slab (detail) [photograph supplied by J.R. Hunter, Dept. of Archaeological Sciences, University of Bradford]

-


ST. NINIAN'S ISLE • Location Map [Close-Brooks 1981:3] •

Ogham-inscribed slab [^Trustees of the National Museums of Scotland, Reg. IB 112]; (key to numbering)

-

Inscribed sword chape (front and back) [Close-Brooks 1981 fig. 15]

-

Inscribed sword chape: Detail of inscription [Jackson 1973b fig.29], Form of script used [Brown

ILLUSTRATIONS 1993:246]

SCOONIE • Ogham-inscribed cross-slab [ECMS fig.360] • Ogham inscription with key to numbering - Sites associated with the cult of Ethernan

WHITENESS • Ogham-inscribed fragment • Stevenson's reconstruction of interlace pattern [1981 fig.l] • Two possible reconstructions • Schematic representation of ogham inscription (key to numbering) • Ogham-inscribed fragment [Stevenson 1947 pl.22 (4)] • Fragment of cross-slab NMS IB 248 [Tait 1938 fig.5]

APPENDIX: DUBIA

Location Map. The ogham inscriptions of Scotland: Dubia

ABERNETHY 2 • Sketch by the Earl of Southesk [1895:249]

BIRSAY • Allegedly ogham-inscribed pebble [Curie 1982:ill.45 detail]

ILLUSTRATIONS

xxvii

BURRIAN * Spindle-whorl inscribed with allegedly ogham-like markings [MacGregor 1974:91 fig 18, detail]

DRUMOYNE * Pillar with incised linear markings [Photo. A. MacLennan, Kelvingrove An Gallery & Museum] * Linear carvings

FOSHIGARRY * Bone disc inscribed with allegedly ogham-like markings [Hallgn 1994:218 Illus,5|

GIGHA 2 * Lower portion of pillar bearing ogham-like markings [Photo. Frances Hood]

GURNESS * Bone casket-lid allegedly inscribed with ogham [Hedges 1987.11:106 fig.2.32, detail]

LOCHNAW * Pillar fragment * Detail of carving with key to numbering

POOL * Bone pin with ogham-like markings [after Holder 1990:72]

ENVOI 'Hey, you've missed out theiW J.D. Moir Pictish Arts Society Newsletter 5, Spring 1990, p.5

xxviii

ABBREVIATIONS

AUTHORITIES AND GENERAL - See also bibliography

ATig

'Annals of Tigernach\ ed. Whitley Stokes, Revue celtique 16 (1895) 374-419; 17 (1896) 9-33, 119-263, 337-420; 18 (1897) 9-59, 150-97, 267-303.

AU

The Annals of Ulster (to A.D. 1131), ed. Sean Mac Airt & Gearoid Mac Niocaill, Dublin 1983.

CGH

Corpus Genealogiarum Hiberniae, Vol.1, Michael A. O'Brien, Dublin Institute for Advanced Studies, 1962.

CIIC

Corpus Inscriptionem Insularum Celticarum, 2 Vols. (1945, 1949), R. A. S. Macalister, Dublin: Stationery Office [numbering, by inscription, continuous through both volumes].

DIL

Dictionary of the Irish Language Based Mainly on Old and Middle Irish Materials: Compact Edition, Royal Irish Academy, Dublin: 1983.

DNB

The Dictionary of National Biography

ECMS

Early Christian Monuments of Scotland, James Romilly Allen & Joseph Anderson, Edinburgh: Society of Antiquaries of Scotland, 1903 [unless otherwise specified, references are to Pan III].

HE

Bedefs Ecclesiastical History of the English People, ed. and transl. Bertram Colgrave &

ABBREVIATIONS

XXIX

R. A. B. Mynors, Oxford: Clarendon, 1969 [cited by book and chapter number].

NGR

National Grid Reference

NMR

National Monuments Record of Scotland, RCAHMS, John Sinclair House, Bernard Terrace, Edinburgh

NMS

National Museums of Scotland (and bodies to which it is the successor institution, e.g.

National Museum of Antiquities of Scotland, Edinburgh)

NSA

The New Statistical Account of Scotland, 15 Vols., Edinburgh, 1845.

OD

Ordnance Datum

OS

Ordnance Survey

OSA

The Statistical Account of Scotland, 21 Vols., Edinburgh, 1791-9.

PSAS

Proceedings of the Society of Antiquaries of Scotland, Edinburgh.

RCAHMS

The Royal Commission on Ancient and Historic Monuments of Scotland.

VC

Adomndn's LifeofColumba, ed. and transl. Alan O. Anderson & Marjorie O. Anderson, Edinburgh: Thomas Nelson, 1961 [cited by book and chapter number, all translations Anderson 1991 J.

ABBREVIATIONS

XXX

STANDARD ABBREVIATIONS FOR THE PRE-1974 COUNTIES OF SCOTLAND

ABD Aberdeenshire

RNF

Renfrewshire

ANG Angus

ROS

Ross and Cromarty

ARG Argyllshire

ROX

Roxburghshire

AYR Ayrshire

SHE

Shetland

BNF

Banffshire

STL

Stirlingshire

BTE

Bute

SUT

Sutherland

BWK Berwickshire

WIG

Wigtownshire

CAI

WLO

West Lothian

Caithness

CLA Clackmannanshire DMF Dumfriesshire DNB Dunbartonshire ELO

East Lothian

FIF

Fife

INV

Inverness-shire

KCD Kincardineshire KNR Kinross-shire LAN

Lanarkshire

MLO Midlothian MOR Morayshire NAI

Nairnshire

ORK Orkney PEB

Peebles-shire

PER

Perthshire

ABBREVIATIONS

xxxi

OTHER ABBREVIATIONS AND CONVENTIONS

Languages

OW.

Old Welsh

Olr.

Old Irish

MW.

Middle Welsh

EOIr.

Early Old Irish

OC.

Old Cornish

Modlr. Modern Irish

OB.

Old Breton

ON.

Old Norse

WCB.

Welsh/Comish/Breton

L.

Latin

Symbols

developed into

*

unrecorded hypothetical form

Others

s.a.

sub anno (in Annals, under the year)

nom.

nominative

s.v.

sub verbo (under die word or heading)

gen.

genitive

s.n.

sub nomine (under a specified name)

sg./pl.

singular/plural

For conventions used in ogham transcription, see Introduction: Transcribing an ogham inscription.

xxx ii

INTRODUCTION

THE OGHAM SCRIPT

DESCRIBING OGHAM The Ogham character is an alphabetic form of writing in which ... letters ... are denoted by combinations of parallel strokes in number from one to five, set in varied positions with respect to a central stem-line' [Macalister 1945 :iv) The external form and internal structure of the ogham alphabet has been amply covered by McManus [1991:1-43), and there is no need to repeat his discussion here. What follows is a brief explanation of the methodology employed in the catalogue for describing ogham inscriptions with a note of special terminology adopted. Learned Irish texts, such as Auraicept na nÉces 'The Scholar's Primer' [Calder 1917 l.lff/2260ff; Ahlqvist 1992], De dúilibfeda naforfid (a short tract on the supplementary letters) [Calder 1917 1.5416-5463J, and In Lebor Ogaim 'The Book of Ogam' (a.k.a. 'The Ogam Tract') [Calder 1917 1.5465ff; McManus 1991:138-40J), reveal a developed vocabulary for referring to various aspects of ogham [McManus 1991:3]. However, in the interests of accessibility, I have, in general, preferred to translate the Irish terms. Thus 'stem' or 'stem-line' is used for the hypothetical base-line, not drulm ('ridge', 'edge', 'back'), and each component part of a letter is labelled a stroke, notflesc'twig', ('score' is more common in earlier literature but in order to be able to discuss the script independently of epigraphic manifestations of it, I have favoured the more neutral 'stroke', and kept 'score' for incised lines regardless of whether or not they are letter-strokes). I have, however, kept aicme (pi. aicmi, 'family', 'class', 'group') to refer to the four 'quinary groups' into which the original alphabet was divided Traditionally, each group was known by the name of its first letter, thus: Aicme Beithe (BLVSN), Aicme húatha (HDTCQ), Aicme Muine (MGGWSIR), Aicme Ailme (AOUEI). I have compromised with b-aicme, h-aicme, etc. By extension, I use 'b-strokes' to refer to letter strokes on the 'b-surface', i.e. to the right of the

xxxiii

INTRODUCTION

stem, and 'h-strokes' and 'h-surface' to the left [cf. Macalister 1945:xv|. I have not used Irish táebomnai 'consonants', ocfeda 'vowels' (or more generally 'letters'), but I have kept foifeda (sg. forfid) for the assortment of supplementary characters subsequently added to the original alphabet (see below). SimsWilliams uses a short-hand for identifying letters by their position within the aicme, thus h4 is 'the fourth member of the h-aicme, i.e. C \ m3 is 'the third member of the m-aicme, i.e. ?GW' [1993]. This system is very useful for referring to individual letters, for which purpose it is adopted below, but is too cumbersome for transcribing full texts.

Like its model the roman alphabet, ogham is read along the stem from left to right. Thus when disposed vertically, as was always standard epigraphic practice, ogham reads upwards with the reader's head inclined to the left (there is one definite case in Scotland of a top-down vertical text: still, of course, reading left-toright, but with die reader's head turned to the right). Even in the manuscripts, where ogham was habitually written on a horizontal stem, texts are described as if arranged vertically, like the orthodox inscriptions [McManus 1991:3]. I have followed this convention, regardless of the actual alignment of the stem, referring to h-strokes as being to the left (i.e. 'above' a horizontal stem) and b-strokes to the right ('below' a horizontal stem).

Similarly, the directions 'forward' and 'backward', are relative to the

direction of reading.

Native writers were not concerned with the palaeography of ogham (i.e. the style in which the characters were written) any more than they debated the form of uncials, half-uncials, or Insular minuscules employed in their manuscripts. Thus in order to discuss the finer points of the script it is necessary to generate a few new terms. In referring to different parts of a letter-stroke, I follow Macalister in using the terms 'proximal tip' and 'distal tip', to refer to the points on the stroke nearest to and furthest from the stem [1945:xv]. Long strokes which cross the stem have both right and left proximal and distal tips. Strokes which intersect with the stem at 90° are said to have no slope. The closer the angle to 45°, the greater

INTRODUCTION

xxxiv

or sharper the slope is said to be (angles lower than 45° appear to have been avoided). The direction of slope is said to be •forward* if the stroke's proximal tip is further •forward* relative to the direction of reading than its distal. If they slope at all, b-strokes habitually slope forward, and h-strokes backward. If they slope, m-strokes habitually slope forward, i.e. with their left distal tips 'forward' of the point at which they cross the stem. In some later inscriptions it was common practice to enhance legibility by linking the component strokes of a single letter with a horizontal stroke through their distal tips. This is a 'bind-stroke', and such letters are said to be 'bound'.

In order to enable discussion of the relative length of strokes of different aicmi, I have generated the term 'ogham-band' to refer to the width of the area occupied by the strokes. This is dictated in each case by the distance from the stem of the distal tips of the longest letter-strokes, usually the b- and h-strokes, but occasionally the m-strokes (the middle of the band is, of course, the stem). If a b-stroke occupies half the ogham-band, a vowel-stroke of equal length would occupy the middle two quarters of the ogham-band, one either side of the stem. The length of vowel-strokes in particular varies greatly between texts, the longest occupy the full width of the ogham-band, shorter ones might occupy less than an eighth of it. An m-stroke occupying the full width of the ogham-band is of course longer than a full-width vowel-stroke because the former is oblique and the latter perpendicular.

TRANSLITERATING OGHAM 'It is difficult to believe ... that so grotesque a word as hccwew ever had an actual place in the language of a people civilised enought to practise writing' [Macalister 1902:163] Nineteenth century authors went to the trouble of creating ogham-letter typefaces with which they were able to refer to particular ogham characters directly and unambiguously. Today, however, the inadequacies of commercial fonts, and, perhaps more importantly, the convenience of readers (!) requires that in discussion, ogham characters should be transliterated into letters of the roman alphabet. This is not as easy as it

THE OGHAM ALPHABET. TRANSLITERATION KEY [After Thomas 1994 fig.3.1]

OGHAM-BAND

Vowel-strokes occupying Va band

Vowel-strokes occupying XA band

Vowel-strokes occupying lA band

Vowel-strokes occupying full band

INTRODUCTION

XXXV

sounds. It is clear that in the later period (eighth/ninth century and later) the ogham characters were used in Ireland largely as a cipher for contemporary manuscript spelling. That the correspondence was on a graphemic level (letter to letter) is shown by the way in which it was possible to write Latin, substituting ogham letters for roman, as seen, for instance, in a margin of the St. Gall Priscian [Nigra 1872:15] and the Bodliean Annals of Innisfallen [Macalister 1910]. According to the key preserved in the manuscripts, the standard transliteration in this period was: BLFSN

HDTCQ

MGNQZR

AOUEI

Earlier scholarship took the manuscript tradition as its starting point and all previous transliterations of the Scottish oghams have been based on this key.

More recent work on the structure of the alphabet,

specifically the names of the letters, by McManus [1986, 1988, 1991] and subsequently by Sims-Williams [ 1993], has shown, however, that the above key does not accurately reflect the original values of the ogham symbols. Furthermore, McManus has been able to show that the framers of the ogham alphabet had in mind, not a letter-to-letter correspondence with Latin and the roman alphabet, but rather a letter-to-sound correspondence with the phonemic inventory of their own Irish language. In other words, that 'the values of the Ogam characters are not to be regarded as those of the Latin alphabet arranged topsy-turvy and camouflaged in a primitive Morse code, but rather as the sounds of Primitive Irish as perceived by the inventors' [McManus 1991:31]. These earlier values have been reconstructed with reasonable certainty as follows (italicized letters tentative): BLVSN

ffDTCQ

MG^JR

AOUEI

Since, in the earlier period, the correspondence was letter-sound, it would seem sensible to avoid transliteration with the roman alphabet (which introduces another layer of letter-sound correspondences), and instead, to go straight to a reconstructed phonetic/phonemic transcription enclosed within oblique strokes, //. Using the roman alphabet might, in any case, support the erroneous impression of ogham as cipher for die roman letters, rather than an alphabet in it own right. A phonetic/phonemic transcription

INTRODUCTION

xxxvi

would have the added advantage of dissimilation, reinforcing Harvey's message the Insular orthography is not to be taken at face-value [1989, 1990a, 1990b, 1991].

The draw-backs of such an approach, however, outweigh its advantages. Writing is not speaking, and all spelling is conventional, ogham included. To attempt a phonemic, let alone, a phonetic transcription would be to imply far greater precision than is attainable on present knowledge. It would be impossibly complex too, for pronunciation was dialectally variable and ever-evolving. The Scottish ogham inscriptions cover a range of dates from perhaps the sixth century to the tenth. To reflect the developing language, a new transcription scheme would be required virtually for each inscription (and in any case we have insufficient knowledge of the historical phonology of the early medieval languages of Scotland even to attempt it). So instead we must accept a system of one-to-one correspondences, aware that to do so is not to assign soundvalues, but merely to produce a conventional transliteration which requires further interpretation. That these transliterated oghams in roman letters are to be distinguished from the equivalent manuscript roman letters is indicated here by the use of capitals for the former, italicized lower case for the latter.

If all the texts in Scotland were written in a Goidelic language1 there would be little problem in transliterating them according to the revised McManus/Sims-Williams key, providing, of course, due allowance was made for the date of the inscription (thus for the appropriate stage in, for instance the delabialization of Q /kw/; V /w/ > F /f/, etc.). Certain features of the Scottish inscriptions, such as the apparent adherence to the old convention of using geminate symbols to indicated non-leiiited/nonspirantized consonants, indicate a degree of continuity with the old ogham orthography. Yet other features, such as the novel use of H, indicate that Scottish oghamists were prepared to innovate. Doubtless, this

1 One of the infelicities of established terminology is that the language common to early Ireland and the Western Highlands and Islands of Scotland is called by linguists Old Irish rather than Old Gaelic. When talking of general cultural or linguistic affinity I have used Irish' and 'Gaelic' loosely, avoiding the former in general discussion (when to use it would appear to exclude Goidclicspeaking Scotland), and favouring it in linguistic discussion (when to use 'Gaelic' in contravention of established custom would be potentially confusing). Specifically of languages, 'Old Irish' is intended to embrace the Goidelic language spoken on both sides of the North Channel.

INTRODUCTION

xxxvii

in part reflects the later date of the Scottish texts, the bulk of which date to after the main period of orthodox ogham inscriptions in Ireland. At least as important, however, was the fact that ogham was, for the first time, being used to represent a non-Goidelic language.

Although ogham was framed with the sounds of Primitive Irish in mind, since its starting point was the roman alphabet, there were inevitable orthographic deficiencies from the outset, notably the lack of sufficient symbols to differentiate the full inventory of Irish vowels. The incongruence was all the greater, however, when an attempt was made to use the twenty letters of the ogham alphabet to represent the sounds of a Brittonic language, the lack of phonemic correspondence generating both superfluous letters and uncatered-for sounds. To take just one obvious example, a Brittonic-speaking oghamist would have no need of ogham Q, yet would lack a symbol for the sound /p/. If the ogham alphabet was adopted in Scotland during the Primitive Irish period, then the value of the character Q at that time would have been /kw/. McManus has explained that its letter name, Old Irish Cert, was originally *kwerkw-, cognate with Latin quercus (ultimately from the root *peikw), cf. Welsh perth 'bush' [1991:37, cf. McManus 1986:15, 29]; the Pictish cognate is reflected in Scottish place-names in pert (Perth, Logiepert ABD) [Watson 1926:356-7]. If the borrowing was sufficiently early, it is possible that this fourth member of the h-aicme was assigned the value /p/, either because that is what the borrowers heard as the initial of *kweikw-, or because they were able to recognize the cognate word in their own language. If however, the alphabet was adopted at a date later than the sixth century, Primitive Irish /kw/ would have fallen together with /k/, for which there were now two symbols, Q and C. The correspondence between Irish mac and Welsh map was not lost on the Irish compiler of Cormac's Glossary, c.900: %Dind map Lethain i tlrib Bretan Corn ./. Dun MaicLiathain, ar is mac indnl is map isin Bremais ...' [O'Donovan 1868:111-2]. It may also have been obvious to Scottish oghamists a few centuries earlier. McManus has shown how the creation of the ogham alphabet 'was accompanied by a careful analysis of the sounds of... [the target] language1 [1991:31]. One wonders if the adoption of the alphabet by non-Goidels might be accompanied by a similarly careful

INTRODUCTION

xxxviii

analysis of the mis-match between donor alphabet and target language ? The creative input into Ogam was', he says,'quite considerable* [39]. If the oghamists of Scotland were both as linguistically aware and creative as the original framers of the alphabet, they may well have been able to take the redundant symbol 'Q' and assign to it the new and useful value /p/.

So what, then, was the value of the fourth member of the h-aicme in Scottish ogham inscriptions, /kw/, /k/, or /p/ ? The texts are surely all too late for /kw/, and /k/ is already catered for by C. Tellingly, die Q-character, appears in Scotland fcnly in the formula word MAQQ/MEQQ. Since we dius have no independent check on die value, we cannot know if in Pictland MAQQ might actually have meant /map/. Obviously, to adopt the transcription Q = /p/ would have a striking effect on our readings, rendering them suddenly a lot more 'Brittonic-looking'. Similarly with die third letter of the b-aicme, V /w/, which by the Old Irish period represented /f/, manuscript/. If this is transcribed instead by its Pictish cognate /w/, manuscript uu- (cf. Welsh gw-), the effect is equally transformative. Some Scottish oghams, however, were clearly carved in an Irish-speaking environment, where, presumably Q still meant /k/, and V meant HI. From this it can be seen diat assigning sound values to ogham graphemes is not straightforward. In transliterating the various characters, choosing one option over another can turn a Goidelic-looking text, into a Brittonic-looking one at a stroke. These difficulties must be born constantly in mind, but in an attempt to avoid prejudging die issue, I have adopted, as far as possible, a neutral transliteration, i.e. V not F or W, Q, H, and left assigning possible sound-values to die linguistic discussion of each text. I have also adopted, as far as possible, the principle of one-fbr-one transliteration (exceptions are Gw and S'). In the case of forfeda two or more letters standing for a single character are underlined, e.g. >$$$(r BRSince the letters V and H occur frequently in the inscriptions, however, dieir possible values are discussed below:

INTRODUCTION Transliteration

xxxix

of 7 / ' (See Sims-Williams 1993:162-170)

The first letter of the h-aicme does not appear in classical ogham inscriptions and we do not know its original value. The equation with h, well attested on all media from 'at least the ninth century* [SimsWilliam 1993:162], 'is no more than a cosmetic solution to the problem created by the loss of the original initial consonant' [McManus 1991:36-7]. When the older ogham convention of indicating spirants with simplex as opposed to geminate consonants had broken down, and ogham orthography was merely a cipher for contemporary manuscript spelling, H came to be used with C and T as equivalents of the manuscript digraphs ch and th [for orthographic solutions to the problem of representing the spirants Ixl and /6/ see Harvey 1989, 1990; also Sims-Williams 1992:45-51]. In the period following the Second Spirantization, before the orthographic conventions had settled down, there is clear evidence that h could be used to represent Ixl* Three roman alphabet inscriptions, two from Wales, the otlier from Ireland, preserve this usage; broho- for /broxo-/ (capitals CIIC 349, 401) and Menueh (
are well-spaced, but letters of different acimi are tightly packed, even with less space than between the component strokes of the letters in question. A notable feature of this inscription is the marked cramping of letters towards the end. The first letter is very generously space, almost over-generously, leaving only minimal space for the last few. Brandsbutt offers a fairly close comparison of script, though it has longer vowel-strokes, and b- and h-consonants which slope minimally if at all. Scoonie and St. Ninian's Isle are close too, though the former has less slope on its b- and h-aicmi and the latter has more explicitly-spaced letters.

INTERPRETATION OF TEXT The extant letters of the ogham text are clear and there is no doubt over the reading except for the final letter(s). Precedent would lead us to expect the eight or more letters to form a male personal name in the

CATALOGUE OF THE OGHAM INSCRIPTIONS OF SCOTLAND

ACKERGILL

18

genitive. The single personal name formula is a common enough sub-type among Irish ogham inscriptions [McManus 1991:52]. Neither Nehtetri nor Nehtetri- suggests an otherwise attested personal name, but depending on how the word is segmented, possible known personal name elements can be identified.

Neht- may be compared with Formaston's Nehht-y Latheron's Netfu*)- and possibly with the (I)Neit(ateon the Gurness knife. Previous authorities have usually connected these with the well-attested Pictish name Nechtany or in less Gaelicized spelling Neithon. These are from *Netíonos [O'Rahilly 1946:368]. Since the -anf-on in these names is not a diminutive suffix, but an intrinsic part of the name, it is not strictly speaking detachable. It appears, however, that a false analogy may have been made with diminutives and the -an came to be considered detachable, cf. Talorc/Talorcen, Drust/Drostan.

In Irish tradition Necht appears as the name of the divine figure Nuadu Nechu though wrongly associated with Nechtarty it is in fact an epithet, the Olr. verbal adjective meaning 'propre, pur, bland 'clean, pure, white', from the root *nikto- 'lavé', 'washed* [Vendryes 1983:s.v.]. As a noun, Olr. necht means 'niece' but it is not attested as a personal name element. There is, of course, no certainty that the -H- in the Ackergill inscription represents l%l [see Introduction for discussion of transliteration of H]. If the spirant is very weak, Ackergill's EH may be equivalent to the diphthong -e/- in the Neitano of the Peebles inscription [Steer 1969]. Neht might be comparable with the Irish Neit listed in O'Mulconry's Glossary as nomen wri, and the name of the young poet NeidelNéde in the Accallam in dá Thuraid [Stokes 1905:14, 8]. Vendryes links this name with the common Olr. noun néity 'combau lutte\ [1983:s.v.] and the name of the Irish war god Néid mentioned in Connac's Glossary, compare with the Gaulish divine name Neto [Evans 1967:370], the war god of the Accitani [Vendryes 1960: s.v. nia]. The related element Netaoccurs in composition in such names as ogham NETTA-SLOGI (CIIC 109 > NadSlúaig) [McManus 1991:109-10 §6.15], and uncompounded in Olr. Niath, Nioth, gen. Neth [Stokes & Strachan 1903:274.38, 267.39, 269.34, 273.7]. These are associated with Olr. nia 'champion', 'hero' or its homonym 'sister's


ACKERGILL

19

son, nephew'.

After the first syllable, the segmentation becomes problematic. One possibility is Neht-Etri> with the second element comparable to MW Ethri < *Ettorigos [Williams 1937:cxvi], the latter attested as Ettorigi, with probably mistaken genitive formation, on the early sixth century inscribed pillar from Llanbabo, Anglesey [CIIC 318, Nash-Williams 1950:53-4, Jackson 1955:566]. This would imply a three-element proto-form *Neta-Etto-rigi. An alternative segmentation, Nehtet-ri, brings us back to Rhps's comparison of Nditet with the Uoret of the Drosten Stone at St Vigean's, and with the Namet and Morbet of the Kinglist [Anderson 1973:246-7]. The latter two, however, appear as second names and are perhaps best interpreted as epithets rather than names proper. The syllable -ettl-edd occurs in several of the Scottish oghams, apparently on the end of personal names, but its significance has not yet been satisfactorily explained. If Nehtet is to be taken as a unit, what remains is r/-, which recalls the very common Celtic personal name element 'king, ruler'; -rix -rig etc. in Continental Celtic [Evans 1967:243-9], r(h)i- and -n in Welsh and Breton [Lloyd 1888:51]» -ri in Irish [Uhlich 1993:294]. There are numerous epigraphic examples of names containing this element [McManus 1991:104 §6.7] and the forms are discussed in detail by Jackson [1955:624-8 §180].

Is the language of the inscription Gaelic or Brittonic, and is the case nominative or genitive ? The answer depends in pan on the reading of the last two letters. If the inscriptions ends with -ri this could be Irish nominative, or Brittonic nominative or genitive. There is certainly not room for the pre-apocope < *-rigas but the two strokes necessary for a final -g, giving the correct Irish genitive rig, would just about fit. Pursuing this possibility, the Ackergill text may represent a compounded Nehtet-ri(g) or an uncompounded Nehtet Ri(g) with the second element operating perhaps as a title, 'Nehtet the king*. There are no ogham examples of uncompounded forms of rig, but Macalister has proposed that the three line text of the fifth century pillar from Llanber, Merionethshire, CIIC 413 (CAELEXTI/MONEDO/RIGI), be read as three


ACKERGILL

20

separate words meaning 'Caelextus King of the Mountains'. Jackson, however, would take MONEDORIGI as a compound personal name [ 1955:355]. Several late Insular script inscriptions from Ireland contain the word rig, usually qualified by the name of his kingdom. Four refer to 'the king' in rig and give his name, but they are all late: the famous slab at Fahan Mura (CIIC 951), a tenth century cross at Clonmacnoise (CIIC 849) and two twelfth century crosses at Tuam (CIIC 522-3). The distinguished position of the Ackergill stone in relation to the rest of the cemetery might reflect the honoured burial of the local petty king, someone with status perhaps analogous to the ri of the Irish niath, but if -ri is to be taken simply as the final element of a compound name it is to be interpreted less literally. The etymological meaning of the name was unlikely to be foremost in the mind of those who used it, Evans produces a few Gaulish examples of names compounded with -/it belonging to women [1967:244],

If the alternative reading NEHTETREB is accepted, then the final TREB immediately demands comparison with Celtic *trebo- 'inhabitation, 'settlement', cognate with Latin tribus, and English thorp. The Welsh reflex tref' homestead, hamlet' is very common as a place-name element. Its Irish cognate is, however, extremely rare as a place-name [Watson 1926:357], though common enough as a noun, treb, meaning, perhaps under influence of the Latin, 'house, farm, holding, household, tribe, stock' [DIL]. The word must also have existed in Pictish since it occurs as a place-name element in eastern Scotland [Watson 1936:357-65; Nicolaisen 1972], notably as the second element in the territorial name Moray < *mori-treb 'sea(board) settlement' [Watson 1926:115-6]. It seems unlikely that a place-name would be recorded in an inscription of this period, but the element also occurs as an ethnic name, the Atrebates of southern England, and as the name of a deity, Contrebis recorded in inscriptions from Lancaster and Overborough [Rivet & Smith 1979:259].

DISCUSSION The Ackergill ogham stone must surely be seen in the context of a sphere of cultural and perhaps political


ACKERGILL

21

influence centred on Orkney, including Caithness and much of the northern pan of the Scottish Atlantic province. The rectangle symbol has a predominantly northern distribution and the form of the script employed at Ackergill is closest to that of the chattel oghams of the Atlantic province and the older Orkney ogham stones. Were it not for the ogham inscription, Ackergill 1 would be a typical Class I symbol stone. The symbols are accomplished renditions of two 'mainstream', though not particularly common, symbols. The uneven spacing of the letters and unusual angle of the stem gives the ogham a slightly scrappy appearance in comparison with the elegant spirals of the rectangle.

Ackergill has been compared

unfavourably to Brandsbutt. Padel thought that In general confidence of execution' the former 'does not at all match up to the other stone' [107]. Ackergill is certainly more modest in scale and more lightly incised, its symbols, however, are no less accomplished than those of Brandsbutt. The contrast is primarily between the two oghams themselves. The design element in the Brandsbutt inscription is very strong, the layout and spacing is careful and the over-all impression is calligraphic. Ackergill on the otherhand, seems closer to ordinary writing and the distinction between the two is one of greater informality as against greater monumentality.

Though it cannot be proved definitively, the association between Ackergill 1 and one of the graves in the dune seems likely. It is merely an assumption that the ogham referred to the individual buried underneath. Note that, in common with the Brandsbutt stone, Ackergill does not have the typical 'X MAQQ Y' formula. If the symbol pair is taken as referring similarly to the person buried, then the implication would be that the ogham was a 'transliteration' of the message of the symbols and that both represented the name of the deceased. Ross Samson's recent suggestion that the Class I symbol pair might represent the two elements in a di-thematic personal name is of interest here [1992]. Since neither the fish nor the 'rectangle' appear on any other ogham inscribed stones (except Golspie where it is one of eight symbols) there is no control by which to check a possible correlation between fish and Neht- and a 'rectangle', whatever it is supposed to represent, and -Etri. In fact, if, like the ogham, the symbols are to be read from the bottom up, then perhaps the correlation would be rectangle/Afecfcr, fish/Em*.


ACKERGILL

22

Since the vast majority of Class I symbol stones do not bear ogham inscriptions, one might wonder why one was required at Ackergill. The bi-lingual ogham pillars of Wales spring to mind as an example of the same text conveyed to two audiences, one speaking British, the other Irish, through two scripts side by side. Might the ogham text at Ackergill have been intended for an audience who would not have understood the symbols ? If so who might these people have been ? Could this be evidence for small numbers of people of Irish origin settled amongst the indigenous population of the far north (knowledge of ogham betrays influence which is, at least ultimately, from Ireland) ? Alternatively, the ogham might be considered in some way supplementary to the message of the symbols, expressing, perhaps, something which could not be conveyed by the symbol system alone. Either way, the prominent position of the stone atop a mound by the shore next to what was probably a coastal routeway suggests the message was supposed to be read by passers-by.

BIBLIOGRAPHY

OS Record card - ND 35 SW 12

Anderson 1897:296 fig.2; Rh?s 1898:370; ECMS 28-9 ill.; RCAHMS 191 Ib: 190-1 No.587; Edwards 1926:179; Macalister 1940:214, fig.7; Diack 1944; K H Jackson 1955:141; padel 1972:107-110; CloseBrooks & Stevenson 1982:27 ill.

Site Only:

Edwards 1926, 1927; Ashmore 1980:348-9, 351-2; Close-Brooks 1980; Close-Brooks 1984:97-9 ill.; Batey 1991:49, 53-7.

PREVIOUS READINGS Rh?s, ECMS, Macalister, Padel

NEHTETRI

N

0

i 0

i........ i i

i

i

30 m«tr«i i

i i

i

'i • r

i

i

| 100

f««t

ACKERGILL - Plan of cemetery [CLOSE-BROOKS 1984 fig.5.6]

0>

i en

1 CO

"8 *c .S "T

E cd O i

C/3

S

o legal term 'coigny' [DIL]\ but diere is no reason to believe the name predates the Gaelic ascendancy of the later ninth century.

The literary evidence points to the existence in Munster in the eighth century of at least a tradition of a connection with eastern Pictland south of the Mounth. Even if one is not prepared to accept an early (sixth? seventh? century) Irish settlement in Kincardineshire, the Auquhollie ogham pillar is unequivocal evidence for some form of Irish influence. One would imagine that such influences would need to be rather strong to produce, not only the ability, but die desire on the part of Picts to erect an Irish type of monument using an Irish type of script. If one is unprepared to accept this alleged Irish settlement then must one come up with another historical mechanism by which someone living in the foothills of the eastern Mounth would have wanted to and had the means to erect an ogham pillar.

INTERPRETATION OF TEXT - Linguistic Discussion Our first expectation would be that the Auquhollie text consisted of Irish names in die usual Irish


AUQUHOLLIE

51

orthography, but even if the names were Pictish, the orthography would have to have been derived from Irish ogham, and so an Irish aspect would be inevitable. If the latter were the case the orthography might be rather experimental. Thomas has discussed the interplay of Goidelic and Brittonic on the ogham and other monuments of Dyfed and Dumnonia as the Irish settlers there Vent native', at least to a certain extent [1994].

Turning now to the ogham text, though much of this inscription is uncertain, it is clear that it lacks the central MAQI so typical of the pillar oghams of Ireland. The worn patch in the middle of the text is almost long enough for MAQI, but there is nothing in the little that remains to support such a reading. The AuquhoUie text is of fairly typical length and an interpretation 'X [*] Y', the [*] being the word for some interpersonal relationship, is highly likely. Though MAQI 'son' is by far the most common relationship expressed in the formula 'X [*] Y', a significant number have AVI 'grandson* and fewer have CELI 'client1, 'dependent', 'follower', one has NIOTTA 'nephew' [see QIC index verbomm, and McManus 1991:118-20 §6.27]. There is one example, from Eglwys Cymmin, Carmarthen (CIIC 362), of INIGENA 'daughter', the sole example of the ogham commemoration of a female. There would almost be room in the weathered central section of the AuquhoUie text for AVI, but again insufficient remains to support such a reading.

Of course, there are numerous Irish words expressing family and other relationships which might be deemed appropriate in this context, and there is, also the outside possibility that the language of the link word is Pictish. If the basic structure 'X [link] Yf is accepted, and the text is surely too long for a single name, this implies that the names on either side are something like: Vuunon [-] Tedovoni or Vuunon (l-]te) Dovoni The final -/ looks like a masculine genitive ending, if VUUNON is not complete as it stands, a genitive ending may have been lost in the worn patch:

i.e. Vuunon[- (-]te) Dovoni etc.


AUQUHOLLIE

52

An alternative interpretation is to take the letters at 7 as the possible remains of some form of NETTA or NIOTA 'sister's son* [McManus 1991:109-10 §6.15]. In Irish ogham inscriptions the word is usually the first element in a compound personal name of McManus's Type C. The most common Type C formation is that with MAQQI-, but as McManus explains, this 'does not denote a filial relationship to the second element - often a dependent genitive of a divine name or the name of a tree or a word associated with a trade - but probably originally had the meaning 'devotee* or the like* [1991:108-9 §6.14]. Names formed from NETTA, e.g. Nad~fraich, Nad Sétna, Nad-sluaig [CGH\7 had similar connotations. If, as discussed below, DOVONI is a form of the divine name DOVVINIAS, the case for a NETTA- name is strengthened.

The form in the classic ogham pillars is usually NETTA, though there are two examples with 10 vocalism, CIIC 202 and 252, in the latter NIOTTA is not pan of a name, but is the link word indicating family relationship. The relevant portion of the Auquhollie inscription is, unfortunately, obscured. There is cenainly room for 10 (NIOTE), or if I am right to suggest that there may be the remains of h-aicme strokes amongst the markings at 7, then NETTE. Final E rather than the expected A is a problem. The word may be comparable with Formaston's NEHHT-, Bressay's NAAHT-, and Gurness's -NEITA. To accept 6-9 as some form of NETTA- or NIOTTA, would necessitate separating off thefirstname VUUNO and the last name DOVONI.

WUNOIVUUNON1 Depending on how one wishes to segment the inscription, the ending could be -ONI, or simply - 0 . The former might be some form of the diminutive suffix -IGNI, -OGNI (VENDOGNI 'fair' [CIIC 422]) and especially -AGNI, which is so very common in the formation of Irish personal names [for 0 for A see McManus 1991:118 §6.26]. McManus notes a few examples of ogham -ONI for -OGNI, but these are all doubtful [1991:107-8 §6.12]. If the -I belongs to the next word (I[.]E, cf. St Vigeans/pe ?), the form may be VUUNON. McManus gives examples of early post-apocope inscriptions with -I dropped in one name but not the other [1991:94 §5.26], but the readings at Auquhollie are so uncenain that it would be unwise


AUQUHOLLIE

53

to attempt a linguistic dating on these grounds alone. Final -O might reflect the genitive of a Old Irish /- or «-stems (-0 < Primitive Irish -ós) [McManus 116 §6.25]. No satisfactory explanation of the first element, with its double vowel, is immediately apparent, though we might compare VUUNO- with ogham VOENACUNAS (recte VOINA-CUNAS [McManus 121 §6.28]), CIIC 164. In Archaic Old Irish texts vowel length may be shown by doubling. Thurneysen suggests this may be intended to show something more than mere length 'perhaps a pronunciation bordering on disyllabic in certain positions of the word in its clause or in slow speech' [Thurneysen 1946:20 §27]. Whether the third character of the b-aicme is to be transliterated /w/ or HI is a function of date and linguistic context (see Introduction). Since neither of these have been established, I will use the conventional V

DOVONI Dovoni presents fewer problems. Names in DOV-, 'black', appear in a number of inscriptions from both sides of the Irish sea (Uhlich [1989] explains why the form is consistently DOV- rather than an expected DUBU-), for instance, DOVAIDONA (> Dubáed) CIIC 503, DOVALESCI (> Duiblesc) CIIC 63, DOVATUCI (> Dubthoch) CIIC 37 [Uhlich 1993: s.v.]. The Auquhollie name may be a form of DOVAGNI (> Dubári), cf. CIIC 432, with -ONI for -OGNI/-AGNI, the diminutive suffix, as discussed above [for a discussion of ogham DOV- see McManus 1991:122 §6.29]. Of particular interest in an area with supposed Eóganacht connections is the name of the ancestor deity of the Eóganacht sept, the Corcu Duibne, which appears in ogham MUCOI DOVINIA [CIIC 175, 178, cf. 156].

DISCUSSION In every sense, Auquhollie is a classic example of the orthodox Irish ogham pillar, except, that is, for its location. It is a rough pillar, the inscription runs up the (left) corner of the stone, there is no stem-line, vowels are represented with short notches, and the consonants have similar proportions to the standard Irish type and are not elongated like many Scottish examples. The strokes are equally spaced along the stem and there are no gaps between letters. The overall length of the inscription is typical, and even if the text


AUQUHOLLIE

54

is not of the common 'X MAQI Y' variety, that is far from the only formula used, McManus list fourteen [ 1991:52 §4.6]. In archaeological terms Auquhollie can be dated most probably fifth- to seventh-centuries. A fuller interpretation of the text would be required before linguistics could offer more precision. The main significance of the inscription is to demonstrate that ogham was known in mainland Pictland at this early date. Further corroboration for early knowledge of ogham in Pictish territory is provided by the radiocarbon dating of the Pool and Birsay 3 stones. The literary evidence for a connection between Munster and the Mearns has already been discussed; see Logie Elphinstone for discussion of the re-use of prehistoric monuments in the early medieval period.

BIBLIOGRAPHY

OS Record card - NO 89 SW 10

Southesk 1884:37; Rh?s 1892:270, 1898:348; Nicholson 1896:4 (D); ECMS 203-4; Ritchie 1923:27-8 fig.5; Diack 1925, 1944; Macalister 1940:190-1 fig.3 pl.IIa; Padel 1972:52-4.

Site Only:

OS Name Book, Kincardineshire No. 10 p.93; RCAHMS 1984:19 No.95

PREVIOUS READINGS Southesk

FaDHDONANUITENn

Rh?s 1892

VINONITEDOV

Rh?s 1898

VAMUNONITEDOV

Allen

VUONON i TEDOV

Diack

AVUO ANUNAO UATE DOVENI

Macalister

VUENONITEDOVOR

Padel

VUUNAN iTEDOVOB.B

AUQUHOLLIE - Ogham-inscribed pillar

• *x/ffir:

: • W'

AUQUHOLLIE - Ogham-inscribed pillar (detail)

* ^

////

15 14 13 12

V m

11 10

9 8

5 4 3 2 1

AUQUHOLLIE - Ogham inscription (key to numbering)

55

BAC

MHIC

CONNAIN

DISCOVERY AND CURRENT LOCATION This ogham-inscribed knife-handle was discovered in 1919 in the course of Erskine Beveridge's excavation of a wheelhouse on the tidal islet of Vallay, North Uist [NGR NF 7694 7619]. The Bac Mhic Connain site is located at an altitude of 5m on a sand-covered hillock about 165m north of the bay at Saltam, in Vallay Sound, and about 400m north-west of Vallay House [see maps J. According to Armit [1990a: 108J the wheelhouse was 200m from the western shore of the now drowned machair plain of Vallay Strand which 'it is likely to have overlooked'. The variousfindsfromthe excavation are now in the NMS. The ogham handle was accessioned in December 1921 and has the catalogue number GNB 134.

SITE The six week excavation of the wheelhouse at Bac Mhic Connain was part of a larger campaign of archaeological investigation in the area carried out by Erskine Beveridge after he acquired the estate of Vallay and Griminish. These excavations were so productive that in 1948 Sir Lindsay Scott was able to assert that 'we know more of the prehistory of this estate than of any comparable area in the British Isles' [68]. By the standards of modern scientific archaeology, they were unfortunately limited, and the lack of concern for stratigraphy (there are no sections) is a serious failing [Scott 1948:69]. In the case of Bac Mhic Connain the situation was further complicated by the fact that Dr Beveridge died before publishing his results. The task of publication fell, eventually, to J.G. Callander who had only incomplete notes from which to work [Beveridge & Callander 1932].

The Bac Mhic Connain complex had been extensively reconstructed during its lengthy period of use and quarried in recent times to provide stone for the building of Vallay House, so that all that stands today are 'mutilated remains' [OS record card]. The building sequence was little understood during excavation and the lack of detail in the report, and the little that survives to the present, means that the 'plan should be


BAC MHIC CONNAIN

56

regarded as tentative' [ibid.]. It is clear, however, that the ruins at Bac Mhic Connain represent successive wheelhouse structures [Armit 1990a: 109] of the later centuries B.C. to the first century A.D. After the wheelhouse was abandoned the site continued in use for a considerable period, apparently as a metalworking workshop [Armit 1990c:63; see fig.].

A large number of artefacts were recovered from the site, including pottery, metal (iron and bronze), stone, and organic items (bone and deerhom). None of these, however, are securely connected to the construction or primary domestic use of the wheelhouse, and may relate to any phase in the post-wheelhouse sequence [ibid.]. Hallén commented that 'most of the artefacts are chronologically insensitive and could date to anywhere in this range [c.200 B.C. - eighth century A.D.]' [1994:193]. Stone casting-moulds and triangular crucibles indicate on-site metal-working which Armit would date to the Roman period [Armit 1990a. 110]. In support of this view is the small fragment of Samian ware of possibly second century A.D. date [Robertson 1970] and the whalebone mirror (NMS GNB 61) which falls into the early part of Warner's EIA style and may, therefore, date to the first or second century AD [Warner 1983:164]. The prominence of metal-working at Bac Mhic Connain may simply reflect the fact that the workshop was built on top of the former house while at other sites in the locality the industrial area was away from the habitation area and therefore not excavated [Scott 1948:79]. The preservation of bone and antler was a notable feature of the excavation of Bac Mhic Connain and neighbouring Foshigarry. The surviving finds have recently been subjected to a detailed re-assessment by Ywonne Hallén [1994].

Post-roman period activity is indicated by the bone die (NMS GNB 65) [Clarke 1970; Mackie 1971] and the ogham-inscribed knife-handle. The latter has been dated provisionally by other authorities to the sixthto eighth-centuries AD [OS record card]. This proposed range is indeed not unlikely, but typologically the Bac Mhic Connain ogham is extremely simple. There are no grounds for assigning it to any particular date except within the broadest range. The possibility that the Pool ogham may be as early as the fifthcentury means we should remain open-minded on the dating of the Uist example. At the very similar


BAC MHIC CONN AIN

57

neighbouring site of Foshigarry, the bulk of the settlement was similarly pre-Roman, though Hallén adduced pins and composite bone combs as evidence of possibly seventh or eighth century occupation [1994:193, but see Foster 1990 on the difficulties of dating such combs].

The raw materials needed for the bronze-working would have had to have been imported [Scott 1948:79]. The presence at Bac Mhic Connainof Samian ware indicates ultimately Mediterranean contacts, doubtless mediated through sites in the Irish sea province further south. Other finds imply links with Ireland, for instance the mirror handle has parallels with bronze items in Ireland [Raftery 1984:209]. An almost identical bronze version from Lochlee, Ayrshire [MacGregor 1976 No.272], complicates the picture, because on the strength of it MacGregor suggests possible influence from the western shores of northern England [MacGregor 1976.9]. Hallén comments on the possibility that such mirrors were status objects [1994:225].

The settlement at Bac Mhic Connain is fairly typical of the period in being well adapted to the local environment, and having access to both coastal and machair resources. The nature of the bone and antler finds lead Hallén to comment on the 'expert exploitation of locally available resources and considerable craft skills' of the inhabitants [1994:228]. Though the site plan shows a change from wheelhouses to cellular structures, domestic ceramics show the continuity of cultural tradition [Lane 1990]. Subsequent research [Lane 1987] has shown that there is no archaeological evidence for the Hebridean invasion of 'Scotto-Pictish' people proposed by Crawford [Crawford 1974; Crawford & Switsur 1977; see Armit 1990a:267]]. Instead, Armit has suggested that the decline of monumental 'broch' architecture and the rise of cellular structures reflects a more centralized society in the region. Though he emphasizes that power continued to be exercised at a highly localized level [1990a.276] he postulated, tentatively, the establishment of some form of centralized control based in Orkney [278]. In this context parallels between Bac Mhic Connain and the Orkney oghams become significant.


fi^C

MHIC CONN AIN

58

Nothing is known about the district in the early medieval period. At the end of the seventeenth century, Martin Martin mentions three chapels on Vallay including one dedicated to St. Ultán [1703:67]. There were several saints of this name, though dedications to him in Scotland are rare. Beveridge quotes Forbes [1872] as the source of information of St. Ultán of Kintyre, a silver shrine containing a relic of the saints arm still being in the possession of its hereditary keeper at Sanda c. 1600 [Beveridge 1911:297]. The now lost place-name of Petultin 'Ultán's share', near St Andrews [Watson 1926:409], may reflect a cult of Ultán, but the name, which means 'an Ulsterman' was also in general use in the early medieval period [Ó Corráin & Maguire 1990:174-5]. Beveridge mentions the presence of cross-incised slabs [Beveridge 1911:297-8]. The multi-period site of Coileagan an Udail nearby was occupied in the early medieval period but the results of the excavations in the 1970s remain substantially unpublished. The place-name Vallay is Norse ( < vadill 'shallow water, especially passable on horse-back) [Beveridge 1911:96], Bac Mhic Connain means the '(peat) bank of the son of Connán', the Gaelic bac being a loan-word from Norse, bakki 'bank', 'ridge' [Ordnance Survey 1968:6].

DESCRIPTION OF OBJECT A knife-handle of cetacean bone retaining the rusted tang of an iron blade, incised with ogham letters but otherwise undecorated [see fig.]. Material:

Cetacean bone and iron

Dimensions:

Handle: length 108mm, diameter 16mm; Tang: 6mm

Condition:

All of blade missing except msted tang, bone well preserved, carving clear

The bone has been planed to produce a slight ridge for one line of ogham lettering. There is no other decoration. The handle tapers slightly in a gentle curve to the butt-end where there is a conical hollow. A fracture has occurred near the blade-end, which is damaged. It can be seen from the tapering outline of the bone, however, that very little has been lost.


BAC MHIC CONNAIN

59

This is one of fourteen handles found at Bac Mhic Connain, most made of antler. Both antler and cetacean bone have 'highly desirable properties of resilience and strength' [Hallén 1994:198]. Hallén argues that it is unlikely that whales and related marine mammals were being actively hunted, instead the bones were more likely taken from stranded animals [198-9]. Since only a small part of the bone is used for each artefact and the original surface is rarely present it is usually impossible to identify species [ibid.].

DESCRIPTION OF CARVING

The inscription begins about half-way along the handle and continues for approximately 50mm, almost to the handle's end. There is no trace of anything before the first stroke, and it appears that the inscription is complete. The photograph published by Macalister is excellent, but the one published by Callander in PSAS 66 has had the strokes drawn in, not with total accuracy, and the impression given is incomplete. Marion O'Neill's drawing is misleading in that it suggests the presence of a stem-line, but is otherwise reasonably accurate, though some of the vowel-strokes are marginally longer than she suggests [Hallén 1994:Illus.l2.5]

The longest consonant-strokes are c.6mm long. The lines arefineand appear to have been cut with a sharp blade. The carving is generally clear and well-preserved, the only doubt occurs when the ridge has been worn or chipped and vowel-notches lost. The individual strokes appear to have been carefully spaced, so where notches have been lost through wear, the available space may be used to determine what is missing. The component strokes of certain letters differ somewhat in relative length. Macalister reads too much into these variations which are the result, simply, of straight lines being carved across an unstraight angle with a single knife-stroke. There is clear differentiation of vowels and consonants, and even if the outer edge of a group undulates, on the inner edge, all strokes terminate at the same distance from the ridge/stem. The very slight variation in spacing between components strokes of certain letters is negligible. Each stroke is cut with a single slash, there has been no re-cutting, which perhaps implies that the carving was


BAC MHIC CONNAIN

60

done with skill and some confidence. All strokes are perpendicular to the stem. Reading from the butt to the blade, left to right, the carving is as follows [see fig. for key to numbering]: 1

One long stroke perpendicularly across the stem - M. The upper and lower portions are clearly visible, the middle section on the ridge has been lost, but there is no doubt that it is all one stroke. There follows a small gap where the ridge has been chipped. The letters in the rest of the inscription are tightly grouped so it is unlikely that the empty space is original. Instead we should postulate a vowel of probably one, or a maximum of two, strokes - A (or possibly O).

2

Five strokes to the left of the stem - Q. The first of the group is slightly shorter than the others.

3

Three clear vowel-strokes on the ridge - U.

4

Five strokes to the right of the stem - N. The first three are clear, the last two, though barely visible, are just discernible. Their presence is confirmed by the spacing of the following letter.

5

Three strokes to the left of the stem - T. Macalister read an extra stroke at the beginning of this group (C). Though a mark is apparent in the appropriate place on Macalister's photograph, I could see no sign of the extra stroke on the handle. Padel referred to it as 'indefinite1. If it did exist it would be almost growing out of the final stroke of the preceding N. Since the rest of the inscription is so well-spaced, such extreme crowding is unlikely.

6

Four vowel-strokes on the ridge, the fourth being slightly less clear than the others - E.

7

Five strokes to the right of the stem - N. The fourth is slightly longer than the others. There is no sign of anything between this letter and the next. Padel described a group of extremely closely packed indentations. These were ignored by Callander, but interpreted as U by Macalister and the NMS. Padel himself interpreted them as 'perhaps a single vowel-stroke (A) rather closer to the N than to the ensuing letter'. Surely there is enough room for only an A.

8

Four strokes to the left of the stem - C. The distal ends of the last two are lost in the ensuing fracture. If there had been a fifth stroke its distal end would have been visible beyond the crack. Ignoring for a moment the break, the gap between the fourth stroke of the C and the first stroke of the following letter is sufficient for about two or three vowel-strokes (O or U). There is no


BAC MHIC CONN AIN

61

trace of any carving there but otherwise the space is inexplicable. The fracture is undoubtedly subsequent to the carving and even if it had been incipient at the time, it was not avoided by the oghamist: the preceding C runs right into the break and the gap is after it. 9

Three strokes to the left of the stem - T. There is no trace of anything following this letter, and the handle ends shortly after.

Reading towards the blade as above yields: M(a/o)QUNTEN( /a)CoT and in the opposite direction: VoS( /a)QEVQUN(a/o)M

Although the ogham could be read in either direction, the former may make more linguistic sense. The only available precedent is that the Gurness inscription appears to read towards the blade. The Bac Mhic Connain oghamist has been able to fit the text into the available space perfectly starting half-way along, but an experienced writer would have little difficulty in judging how much space to allow, so this need not mitigate against the proposed direction of leading.

FORM OF SCRIPT The text is written in a delicate type I form of the script comprising long, thin consonant-strokes and very short vowel-strokes. There is no stem-line. The constituent strokes of the individual letters are closely and evenly spaced, there are no gaps between letters. Twofeaturesare worth mentioning - there are no double letters in the inscription and the initial letter, M, is upright, not oblique. The lack of slope on the M-stroke is in contrast to the Gurness ogham which has a very pronounced slanting long M. In terms of script typology Bac Mhic Connain is closer to Weeting than Gurness, and, despite the gross difference in scale, has many similarities with Auquhollie.

INTERPRETATION OF TEXT No obvious interpretation of the sequence MAQUNTENaCoT immediately presents itself, but a number


g ^ C MHIC CONNAIN

62

of observations can be made. The most conspicuous of these is that the opening three letters MAQ suggest the Irish for 'son'. In the earliest Irish inscriptions this word is written MAQQI (gen.). Since palatal quality is not usually indicated in ogham texts, die post-apocope form, corresponding to MS make, is either MAQ [QIC 7, 9, 10, 55, 112, 145, 219, 220, 233, 248, 364, 409] or MAC [QIC 83, 90, 127?, 256], and die former presumably is what we have here. The same form, MAQ, appears on the Gigha pillar, and at Blackwaterfoot 1 there is the related MEQ, though die reading is difficult. These three examples of single Q contrast with occurrences of the word further north and east where the consonant is always geminate, whether MAQQ or MEQQ. In contrast to the Irish and Welsh oghams the word is always spelled in Scotland with Q(Q) rather than C(C), which, given the date of most of the examples must be some kind of fossil.

In initial position as here, MAQ is most naturally interpreted, not as a patronymic, but as the first element in an uncompounded dithematic name of the form MAQQI-N. (Mac-N.). Such names were very popular among the early Irish, especially after about 600 A.D. [McManus 1991:101, 108-9 §§ 6.2; 6.14]. The dependent genitive in Maqqi-N.-type names is frequently a theonym and such names may have had the original meaning 'devotee of N\ The difficulty with this interpretation of the Bac Mhic Connain text is that it leaves the somewhat baffling Unten(a)cot. Perhaps this is related to the name of the Romano-Celtic deity Antenociticos whose temple was at the fort of Benwell (Condercnm) on Hadrian's Wall. Remains of the temple survive along with three inscriptions [Collingwood & Wright 1965: Nos. 1327-9] and parts of a full-size statue of a youthful god [Webster 1987:pl.l6]. The name has so far defied interpretation [Webster 74] and die cult is restricted to this one site. Neither Untenacot or Untencot may be derived from the form Antenociticos as it stands, but may comefroman unattested alternative vocalization of the same root. If Unten(a)cot is in some way connected with Antenociticos then the Bac Mhic Connain text may consist of a personal name originally meaning 'devotee of Antenociticos' (though the Irish evidence indicates that similar names which presumably had their origin in pagan cults, lost their religious connotations and continued to be used well into the Christian period).


JJAC M H I C C O N N A I N

63

In addition to Maq Unten(a)coty a number of alternative segmentations are possible, for instance, Maqu Nten(a)coty Maqu n-Ten(a)cou Maq U n-Ten(a)coty or Maqun Ten(a)cot. Taking each in turn, I can find no instances of MAQU on Irish or Welsh oghams. The form MAC(C)V- for MAQ(Q)I-, indicating an intermediate stage in the delabializadon of /k w 7, appears on Roman alphabet inscriptions from the second half of the fifth century [McManus 1991:90, 98 §§ 5.18, 5.32] but this is spelled with V not U. The feet that the letter following the Q on the Gigha pillar may be U rather than I serves only to complicate matters. To interpret MAQU as accusative plural, manuscript maiccu, would require us to overlook the lack of any indication of palatalization, and also to come up with an justification of how accusative plural is appropriate in an epigraphic text of this form. To explain Bac Mhic Connain's MAQU as macu (an attested variant of macculmoccu [DH], would require some kind of confusion or conflation with the unhistorical MAQQ, since the antecedent of macculmoccu was MUCOI, i.e. /muk-/ not /mukw-/ [see Formaston for discussion of this word]. Nor is the N easily explained. We may reject the unlikely looking Ntenacot, perhaps leaving for consideration the single N indicating nasalization. But this is hard to square with the syntax of the text as far as it can be understood. The only form to cause nasalization is the genitive plural, MS mace, and this still leaves us with the U to be explained.

To get round the whole problem we might posit an alternative reading, with the five strokes of 4 as vowelrather than b-fl/c/w^-strokes. This would give I and allow the reading MAQ UI, i.e. Mac UL The formula word Ui 'descendants' is used to introduce the name of a tribe, sept, or kindred, and, in the singular, appears in several ogham inscriptions in its earlier form AVI [McManus 1991:110-2 §§6.17-18, 118-9 §6.27]. The name is usually followed by the name of the historical ancestor of the kindred in the genitive. When the name is that of a remote mythological ancestor, or deity, the name may have had religious rather than genealogical significance [McManus loc. cit.; MacNeill 1909:368-9]. The juxtaposition of the form Ul with the Q in MAQ is acceptable only if the latter is taken as a relic of the earlier spelling fossilized, as it appears to have been on all die Mainland Scottish examples, including ones presumably far later than


g ^ C MHIC CONN AIN

64

Bac Mhic Connain. In the end, however, there may be insufficient justification for reading 4 as I rather than N. The group is quite similar to 3 which is more clearly vocalic. While die strokes are slightly longer than j ' s , they are not as long as those of the other N (5). Previous commentators do not appear to have been in any doubt that 4 was to be read N.

Abandoning this tack altogether, an alternative interpretation entails taking the first five letters as a single word - MAQUN. To take this as a form of the personal name Mac-con 'son of a wolf [Ó Corráin & Maguire 1990:127] requires, not only explaining Q for c, but U for o. If we are committed to interpreting the Bac Mhic Connain text as Pictish, and are prepared to transliterate the fifth member of the h-aicme as /p/ (since p and q are cognate in Brittonic and Goidelic), then 1-5 might be taken as a form of mapan, maphan, maban, a Pictish word doubtless containing the cognate of Welsh n\ap 'son' (cf. AU 725=726 Talorcan ntaphan) [Jackson 1955:145].

Turning now to the second element, detaching in turn mac, rnacu or macun, leaves Unten(a)c(o)t, Nten(a)c(o)t, or Ten(a)c(o)t, and other permutations if further sub-segmentation is allowed. Once again, no obvious interpretation suggests itself, but a few observations may be made. The final letters C(o?)T may be compared with die Gaulish name element cot(t) (OCorn. cothy Bret, coz 'old'; from the root meaning 'living', 'lasting'). In proper names this is most common as a first element, but is attested in second position in the personal names Esanekoti and Venicotenius and the name of a (north) British tribe, the Atecotti [Evans 1967:186-7]. For the reading -TENCT, John Koch has brought to my attention the root *tantfi-, reflected in the tribal name Tencteri, in Irish Téchtae 'legal Tightness' [DH], cf. Welsh teithi. Comparison may also be made with the female personal name Tunccetaca which appears on a fifth or sixth century roman alphabet inscription at St Nicholas, Pembroke (CIIC 451). Jackson derives this from Brit. *ToncetBc5i cf. W. twng 'oath' [1953:273]. The name TENACI appears in ogham at Ballrorannig, Corca Dhuibhne, Co. Kerry (CIIC 148). The form consists of ten + the guttural suffix -acand is reflected


g ^ C MHIC C O N N AIN

65

in the sept name Ui Thenaich [McManus 1991:108 §6.13].

Since neither archaeology nor palaeography can offer much help in dating the Bac Mhic Connain ogham, we must rely solely on linguistic dating. Precision will not be possible until the text is satisfactorily explained, though the lack of anything after the final T is significant. It is not clear whether the U belongs to what precedes or follows it. McManus ascribes Irish oghams which display the results of apocope but not syncope to the middle or second half of the sixth century [1991:98-97 §5.30]. In these inscriptions, he comments, the word for 'son' tends to be written MAQ [96 §5.28]. Since, however, the formula word appears to have become fossilized in Scotland, any dating based on the form MAQ is suspect.

DISCUSSION The Bac Mhic Connain knife-handle is one of a very small group of ogham-inscribed household objects, and as such provides important evidence for the non-monumental use of the script. That the early medieval occupants of Bac Mhic Connain were engaged in metal-working implies they were people of some standing, but the site is not of especially high status. Thus we have evidence of ogham literacy at a non-church-site, on the part of people of middling social standing. One must always beware, of course, of reading too much into the location of an object as portable as a knife, but two factors combine to suggest that this example is unlikely to have travelled far. Firstly it is not a high-status object, but part of the ordinary domestic tool-kit. Secondly, it fits in closely with the other artefacts from the site. Hallén points out that cetacean bone was used at Bac Mhic Connain and Foshigarry in situations where other materials, such as antler, wood, and bronze, would have been used elsewhere [1994:198; cf. MacGregor 1985:31]. Since cetacean bone would not have been available inland, we may assume, at least, that the knife came from the Atlantic coast.

The calcareous sands of the west coast of Scotland meant that a wealth of antler and bone material was preserved at Bac Mhic Connain, artefacts which would have perished in the less favourable soil conditions


g ^ C MHIC CONNAIN

66

of the rest of the country [Hallén 1994:228]. By this chance we are given a glimpse of what might have been a far larger category of evidence - ogham on household objects. Two other ogham-inscribed knifehandles have been preserved, one at Gumess in Orkney [q.v.], the other at Weeting in Norfolk [Clarke 1952; Padel 1972:144-8; Holder 1990:58-65]. Other ogham-inscribed instrumenta domestica include the Buckquoy whorl, and the Hiberno-Norse Dublin comb [unpublished]. It is impossible to gauge how common such items might once have been, but the tiny handful which do survive are an important reminder that the current sample, so heavily weighted to public monuments, may not be a true reflection of the original extent to which ogham was used in informal contexts.

The Bac Mhic Connain knife-handle is unremarkable in all respects except for its inscription. The text appears to be a single male personal name and as such probably records the identity of the owner or maker (or donor if it was given as a gift) [see Buckquoy for discussion of the kind of texts appearing on personal belongings in the period]. Previous commentators have been united in labelling the Bac Mhic Connain ogham as 'Pictish', apparently on no more solid grounds than that it is 'unintelligible* and therefore reflects a supposedly non-Indo-European Pictish language. Wainwright took Bac Mhic Connain as evidence for continuing Pictish control of the Western Isles [1961:7]. This seems cock-eyed logic to me. Though the interpretation remains doubtful, the letters are most naturally read as Irish. Whatever the language of the text, the use of the ogham script indicates ultimately Irish influence.

Very little indeed is known about the linguistic and ethnic circumstances prevailing in the Outer Isles in the early medieval period. According to Irish tradition they were inhabited by the TuathlFir Iboth or Ibdaig, names which preserves Ptolemy's name for the archipelago, Ebudae [O'Rahilly 1946:538; Watson 1926:37-8]. The Dame appears to have non-Indo-European origins, but since the names of islands are often of great antiquity, it need not imply that the Fir Iboth spoke a non-Indo-European language (though they might have). Ptolemy's Dumna later Irish Domon, which Watson takes to refer to the Outer Hebrides [1926:40-41] is clearly Celtic. In Irish tradition the Fir Iboth are linked with the Tuath Ore and presented


BAC MHIC CONNAIN

67

as piratical raiders [O'Rahilly 1946:377; Watson 1926:60-6]. This is of great interest in the light of Armit's suggestion of growing Orcadian hegemony in the Scottish Atlantic.

Lane has contrasted the aceramic southern Hebrides (from Mull southwards) with a northern zone of Iron Age ceramics stretching from Lewis to Tiree [1990:111.7.7]. If this reflects a more general cultural divide then it pre-figures the later boundary between Pictland and Dál Ríada. It is unclear when Gaelic speech became prevalent in this northern zone, but Richard Cox's work on the place-names of Lewis demonstrates that the area was thoroughly Gaelic-speaking before the first Norse settlements in the ninth century [1991]. The language of the Bac Mhic Connain text has not yet been established definitively. It may be Irish, but even if it is not, the use of the ogham script is sufficient proof of ultimately Irish influence at the site. As noted above the mirror-handle may be further evidence of Irish contact. Such contact may have been direct from Ireland or mediated through Dál Ríada, but it cannot be dissociated from the evidence of Irish influence in Orkney. There is no positive evidence for any ecclesiastical activity at Bac Mhic Connain and the site has only ever been explained in secular terms. While all the historical evidence is, of course, for ecclesiastical links with Ireland, archaeology suggests contact was not solely restricted to the church.

BIBLIOGRAPHY

OS Record card - NF 77 NE 5

n.a. 1922: accession 9178; RCAHMS 1928:89, No. 271 [plan of earth-house, mention only of ogham]; Beveridge & Callander 1932:56, 65, fig. 11; Macalister 1940:216 pl.vi, No. 21 (sketch of inscription), fig.8 (photograph) opp.p.218; Padel 1972:133-7; Ritchie 1987:65; Holder 1990:46-52; Hallén 1994:219 ill. 12.5.

CATALOGUE OF THE OGHAM INSCRIPTIONS OF SCOTLAND Site only:

BAC MHIC CONNAIN

68

RCAHMS 1928:89, No.271; Beveridge 1931; Scott 1948:75-76, 77-80; Armit 1990a: 108-110.

Artefacts:

n.a. 1922 [bulk of Beveridge collection]; n.a. 1923 [clay pot]; n.a. 1929 [further bone items]; Tylecote 1962:36-7, 194, 199; Robenson 1970:Table 4 [Samian ware]; Clarke 1970:229, No. 4 [bone die]; Mackie 1971:70 [bone die]; MacGregor, M. 1976:9, 141, 143, 163, No.271 [mirror handle etc.]; Laing & Laing 1987:215 [hanging bowl escutcheon]; Lane 1988:55-6; Hallén 1994 [all bone and antler material].

PREVIOUS READINGS

Callender

MAQUNmDENC(o/u)T

Macalister

BELANCEN UCOTA 'Belanc's knife'

NMS label

MEQUNTENUCOT

Padel

(m/h)..QUN(t/c)EN(a/u)C..T or V..S(a/u)QE(v/s)QUN..(m/b)

Holder

AQU(l/n)(q/c)ENUC T (reading based on MacalisterTs and Callander's photographs)

ILLUS I

Site location map with other wheelhouse sites. I Foshigarry: 2 Bac Mhic Connain: 3 Garry lochdrach; 4 Sollas {Drawn by Marion O'Seil). Based upon the Ordnance Survey map 0 Crown copyright

BAC MHIC CONNAIN - Site Location maps. Above: Bac Mhic Connain and the wheelhouse sites of Vallay Sound [Hallén 1994:190 Illus. 1]. Below: As above showing relief [Scott 1948:69 fig.7|.

t

(i)Machair Leathann ; (2) Eilean Malcit; (3) Cnoc a'Comhdhalnch ; (4) Garry lochdrach ; (5) Foshigarry; (6) Bac Mhic Connain; (7) Dun Thomaidh Land over 200 feet shaded

BAC MHIC CONNAIN

i m

Plans of the structures at Bac Mhic Connain. 1 Hearth. 2 Sink. 3 Stone box. 4 Furnace. (Drawn hx Marion O'Neil (after Beveridge & Callander

l#3l~2n

Plans of the structures at Bac Mhic Connain [Hallén 1994:191 Illus.2]

BAC MHIC CONNAIN Ogham-inscribed knife handle [Hallén 1994:220 Illus. 12.5]

B AC MHIC CONNAIN - Ogham-inscribed knife handle [® Trustees of the National Museums of Scotland, Neg. P9557]

Ill II Ill MM BAC MHIC CONNAIN - Ogham inscription (key to numbering)

in 1 2

3

MM 4

5

6

7

8

9

69

BIRSAY

SITE The Brough of Birsay is a tidal islet, just over 20 hectares in area, off the north-west coast of Mainland Orkney at the head of Birsay Bay (NGR HY 239 285) [see map]. The earliest documentary reference to the site is in chapter 31 of Orkneyinga Saga wherein is recorded the decision of the mighty mid-eleventh century Earl Thorfiiiii to build his minster there [Palsson & Edwards 1978]. The exceedingly fertile soils of the hinterland coupled with the defensive, though not isolated, position of the Brough itself and the good landing-place of Birsay Bay were sufficient to ensure that the site had been occupied for many centuries before that. It is especially relevant to a consideration of the importation of ogham to note that the Bay of Birsay 'is particularly well-situated for boats setting out to the Hebrides or Ireland' [Ritchie 1983:47].

Archaeological investigation of the site began in 1934 when the then Chief Inspector of Ancient Monuments, James Richardson, initiated a programme of repair and partial restoration in Area I, the Norse church and cemetery. In 1935 the church was cleared of debris but, though work continued in this area for a couple of years, there was no actual excavation there. In 1936 Mrs Curie assisted Richardson in beginning excavation of Area II, between the cemetery and the cliffs to the east. In subsequent years the work extended to include the isolated buildings in area 111, to the west of the cemetery. All work on the site was halted in 1939 on the imminent outbreak of the Second World War [Curie 1983:67]. A brief synopsis of the major discoveries of this period were included in the 1946 RCAHMS Inventory of Orkney.

Work at the Brough was not recommenced until 1956 when Stuart Cruden and Ralegh Radford began several seasons of excavation in Area II and House C in Area 111. In 1973-4 Mrs Curie and John Hunter excavated Room 5 in Area II. The finds from the initial forty years of excavation were first published in 1982, almost fifty years after the work had first commenced [Curie 1982]. Work continued throughout the 1970s and 80s, by this time in the broader context of the 'Birsay Bay Project'. The results of the


BIRSA Y ( G E N E R A L )

70

rescue excavations of the period 1974-1982 were published by Hunter [Hunter 1986]. Morris is collecting and editing the earlier excavation archives, such as they are, for publication alongside the results of his own Birsay Bay Project excavations [pers. comm.].

In Ritchie's words 'more speculation has surrounded the archaeological and historical interpretation of the Brough of Birsay than any other Dark Age site in Scotland' [1986:5]. The problem is compounded by the non-publication of the results of the earliest excavations. It is clear that the site was important in both the Pictish and Norse periods [see fig.]. The earliest settlement remains on the Brough may date to the sixth century, but the nature of the early occupation remains controversial. Radford identified a chapel and grave-yard beneath that of the later Norse settlement and an enclosing wall. On the basis of these, the ogham inscriptions, two sculptured slabs, and a bell, he understood the site to be a Celtic monastery. Charles Thomas concurred that the Brough, with its curvilinear vallum, was a 'full monastic foundation', perhaps the principal monastery in Orkney [Thomas 1971:37], Although this interpretation was widely accepted for many years it has been called into question by more recent excavators who feel it is founded on an outdated and increasingly ill-fitting model of monastic sites. As Morris points out, Radford's monastic interpretation is based on Tintagel as a type site [Morris 1989:11], but Tintagel is no longer thought to be a religious site [12] (see Morris 1996 for a final refutation of Radford's model).

Its physical location, the substantial nature of the remains, the evidence of high-status metal-working, and the extant sculpture, lead all to agree on the 'special status' of the Brough site, but whether this 'derived from political considerations or monastic impetus is ... finely balanced according to current evidence' [Hunter 1986:171]. Morris thought the interpretation of the most recently excavated material as monastic to be 'attractive but not compulsive' [Morris 1989:15], but in his most recent publication on the site has expressed the opinion that 'the balance is now firmly tipping towards the political' [1996:61]. The small church and graveyard if they do indeed date to the Pictish period, need not be monastic [Morris 1982:81]. The bell, of a type usually diagnostic of Irish monastic influence, came, in feet, from Norse period layers

CATALOGUE OF THE OGHAM INSCRIPTIONS OF SCOTLAND [14].

BIRS A Y (GENERAL)

71

Ogham in Scotland has no obvious monastic connection.

The date and circumstances of the earliest Norse occupation of the site are not clear either, though the 'considerable number of unmistakably Pictish objects' in the Lower Norse Horizon [Curie 1983:78] implies an element of Pictish continuity or assimilation. In the Middle Norse Horizon, which lacked Pictish material, there are a number of rooms with shared walls and separate entrances which have been interpreted as monastic cells. Curie thought this unlikely because of the large number of spindle-whorls found in the area [1983:78]. It has been the traditional view that Earl Thorfinn erected a minster, 'Christ Church', the first Norse Christian Church in Orkney, on the site of a Celtic Chapel on the Brough, though the alternative site in Birsay village on the mainland has perhaps a better claim. By this time Birsay was the centre of the Norse Earldom of Orkney. In the early twelfth century there was a cathedral at Birsay, dedicated, according to tradition, to St. Peter. Though it was superseded by the erection of the cathedral in Kirkwall in the mid-twelfth century, it continued as a place of pilgrimage until the reformation.

EARLY CHRISTIAN SCULPTURE FROM THE BROUGH OF BIRSAY In addition to the ogham- and rune-inscribed stones, three slabs bearing sculpture have been recovered [see figs.]. Cross-slabs The two cross-inscribed stones are the only explicitly Christian items from the site. One, not mentioned by Curie, is a small slab incised with a simple outline cross with a circle at the intersection of the anus. It currently stands upright near the facsimile of the symbol stone, but was found near the church lying horizontal, covering a grave [Ritchie 1986:12, ill.]. The other, smaller stone, was a surface find in Area II. It is a fragment of sandstone flag, 115x220mm, dressed and smoothed on both sides. The double outline latin cross has expanded rectangular arms and a rectangular base [Curie 1982 No.607,111.45 p. 70, discussed p.92. See fig.].


BIRSAY (GENERAL)

72

Symbol Stone The famous Birsay slab was discovered in fragments in 1935 close to the side of the cemetery. It is now in the NMS, Cat. No. IB 243 (a rather poor facsimile stands on the Brough). When complete it would have been about 2.0m x 0.8 m, yet it is only 25mm thick, leading most commentators to postulate that the back, presumably carved with a cross, is missing [Curie 1982:97]. As Ritchie has explained, this is a misconception, and the famous association with the triple grave 'is the result of wishful thinking' [1986:12; 1989:53]. The extant face is incised with two symbol pairs - a 'mirror case' with crescent-and-V-rod, above a Pictísh beast with eagle - a panel containing, in false low-relief, three warriors bearing shields and spears. The slab is discussed in some detail by Curie [1982:91-2, 97-100; see also RCAHMS 1946:4-5 No.l; Close-Brooks & Stevenson 1982:30; A. Ritchie 1983:52].

RUNIC INSCRIPTIONS FROM THE BROUGH OF BIRSAY Four rune-inscribed objects, bearing between them perhaps six separate inscriptions, have been recovered from the Brough of Birsay [see figs.].

Since it cannot be ruled out that the two pre-War ogham

inscriptions date from the Norse period any discussion of them must take account of their runic counterparts.

The first inscription came to light at Birsay in 1921 when it was spotted by Hugh Marwick built into the outer face of the north chancel-wall of the Norse church about a foot from the ground [RCAHMS 1946:35 No. 120]. The runes were on its upper surface and for this reason must have been carved before the stone was incorporated in the wall. The 'longish, narrow stone, seemingly dressed' [RCAHMS 1946:35], designated Birsay 1 (Barnes OR 6), is 2'10" x 6" x 4" (approx. 0.86 x 0.15 x 0.10m). The runes occupy 25" (c.640mm) of the narrow surface and can be divided into three sections. The runes in the left-hand section are almost completely worn away, the ones in the middle are better preserved and the ones to the right are 'quite distinct' [RCAHMS loc. cit.]. From what remains it can be seen that the three groups are carved in different 'hands' which, combined with the differential wear pattern of the three sections,


BIRSAY (GENERAL)

73

suggests they are of different dates. Only the right section has been read - -filibusranru. This has been interpreted as Philippus r(eist) enn rú(nar) 'Phillippus carved (these) nines' [Olsen 1954:164], though Barnes characterized this as 'a very uncertain guess' [1993:35]. The form of the nines dates to the tenth or eleventh century [NMS label]. The text(s) on this stone are of particular interest in comparison with the Birsay oghams because like them, they are best interpreted as informal graffiti. The stone was doubtless dressed for inclusion in a building as a lintel or similar. Whatever its precise location it must have been sufficiently accessible for passers-by to carve their names. When the building became obsolete the valuable ashlar would have been put to good use in the construction of the new chapel. This scenario, which implies a degree of casual runic literacy at the Brough, is very similar to the ones put forward below for the ogham-slabs.

The two other rune-inscribed stones, which came to light in 1934, are more formal in content and form. During the clearing-out of the cemetery area by the Office of Works, fragments of runic inscriptions were discovered including two which had been split longitudinally and re-used as building material in the Norse chapel [Radford 1959:18]. In both cases only the lower pan of the inscription survives. The extant text of Birsay 2 (Barnes OR 8) consists of the lower halves of about a dozen nines and two dividers in the form of xs. Barnes would go no further than to identify one rune as U or R [13], Liestol attempted a full reconstruction in the belief that the letters were the beginning of the common formula -reisti stein henna eftir X raised (this stone after Y)' [Liestol 1984:227 fig.73]. Even less of Birsay 3 (Barnes OR 9) survives, little more than the bottom third of the ten or so characters. Again Barnes would venture no more than that one rune was 'probably R* [1993:40], while Liestol thought that 'a fairly safe reconstruction of most of the nines' yielded -Pina after w-, ie. pan of the same formula 'X raised (this stone after Y)' [Liestol 1984:226fig.72].

The fourth runic text from Birsay (Birsay 4, Barnes OR 11) is the famous nine-inscribed seal's toothpendant [Curie 1982:59-60 No.253, ills.37, 38. See fig.]. The text consists of the first six letters


BIRSAY (GENERAL)

74

of the futhark very lightly incised and, since the carving is invisible at a distance of more than a few centimetres, Curie interpreted it as amuletic rather than ornamental. This item too was recovered in the pre-War excavations conducted by the Office of Works and, according to Curie, was found in the paved passage 1 in Area II (the same context as the allegedly ogham-inscribed pebble). The form of the runes cannot be dated any more closely than 'pre-ttOO' [Curie 1982:60].

Birsay 1 was presented to the NMS by the island's proprietor shortly after its discovery and is currently on display in the museum [NMS 191]. Birsay 2 and 3 were left behind in the Birsay site museum when the other finds were moved to the Tankerness House Museum in the 1990s. Birsay 4 is on display in the

NMS.

SELECT BIBLIOGRAPHY (Site Only)

OS Record card - HY 22 NW RCAHMS 1946: Radford 1959; Cniden 1965; Morris 1981; Curie 1982; Curie 1983: Ritchie 1983; Thomson 1983; Ritchie 1985; Hunter 1986; Ritchie 1989:52-55 Morris 1991.

OGHAM INSCRIPTIONS FROM THE BROUGH OF BIRSAY DISCOVERY Three ogham-inscribed building stones have been recovered in the course of archaeological excavations on the Brough of Birsay. The most recent, and best documented of these discoveries, occurred in the summer of 1980 during excavation by the University of Durham under the supervision of Christopher Morris [Morris 1981:36]. The earlier two were discovered in the excavations before the Second World War, but, as Ritchie has pointed out [1985:192] there is some contusion over their precise find-spots. Cniden


BIRSAY (OGHAM)

75

describes one, which Padel labelled Birsay 1, as having been re-used as a building stone in Thorfínn's Palace* [1965:25] (perhaps confusing it with the rune stone found in 1921 built into the chapel wall?). In the official guide to the site [Radford 1959:5], Radford stated that one ogham came from the churchyard (i.e. the same area as Thorfinn's Palace1), but in a later publication [Radford 1962:174] refers to only one ogham from Birsay and says it was 'picked up on the beach below the Brough'. Full publication of the site was delayed for over 40 years and in her catalogue of finds Curie makes no mention of the first two ogham stones, she does however note the discovery of Birsay 3 [Curie 1982:86]. In addition to the three genuinely ogham-inscribed slabs a small pebble recovered from the Lower Norse horizon is incised with lines which Curie suggests may be ogham [1982:120, see Appendix - Dubia for details].

The existence of each of the various inscriptions has been mentioned in passing in a number of works [see bibliography for each stone] but all are otherwise unpublished. Padel covered Birsay 1 and 2 in his unpublished MLitt thesis. At the time of writing, Birsay 3 is still in the Department of Archaeology of the University of Glasgow. The current whereabouts of Birsay 1 and 2 are unknown. They were in Edinburgh in 1971, when Oliver Padel examined them in Argyle House, Lady Lawson Street, the then headquarters of the Dept. of the Environment [Padel, pers.comm.]. This is the last record of their existence. Historic Scotland, the successor to the Dept. of the Environment, has been unable to find them, or the originals of the drawing and rubbing which constitute the only extant visual record of the inscriptions, the photographs of which are reproduced here [Richard Welander, Inspector of Ancient Monuments, pers.comm. after protracted correspondence with author]. I have made extensive enquiries to museums and individuals and have been unable to gain any further information, except to rule out the National Museums of Scotland, the Tankerness House Museum, Kirkwall, and the site museum at Birsay. It is, of course, deplorable that such important items should have been lost so recently while in State care. Padel*s account, the only detailed description, is thus of the greatest importance. My discussion is based on his thesis and private notes. I am most grateful to Dr Padel for making the latter available to me.


BIRSAY

BIRSAY i

76

1

As discussed above, this stone, dubbed the 'Wall Stone' by Padel, appears to have been found in a secondary context, reused in the eleventh century structure known as 'Thorfinn's Palace', though there is some doubt over the precise circumstances of its recovery. Specifically, it is not clear if the identification of the context as 'secondary' is based on the archaeological fact of the stone's position, i.e. if the inscription was turned to the interior of the wall it would have been inaccessible, therefore the ogham would have had to have been carved before the stone was put in position, or, if it is based, solely on the assumption that an ogham must date to the Pictish phases and therefore an earlier use of the stone. Whilst the later may have been in line with then current thinking on ogham, it can no longer be supported. The bi-lingual ogham- and rune-inscribed slab from Killaloe, Tipperary [CIIC 54], has been dated to the second half of the eleventh century and there is no a priori reason to reject the possibility that Birsay 1 is also the product of a similarly mixed Celtic/Scandinavian milieu of the same period. The following account is based on Padel [1972:56-58, and private notes] and the Department of the Environment photograph A 2894-2 [see fig.].

DESCRIPTION OF OBJECT A squarish flag-stone inscribed along the narrow face ('the only really flat surface' [Padel 1972:56]) with an ogham inscription. Otherwise undecorated. Stone:

Orkney flag

Dimensions:

24tew x 12" x 4V£W (i.e. approx. 0.63 x 0.31 x 0.12m)

Condition:

Fragmentary, 'much chipped' [Padel MS notes]. According to Padel the stone has been broken and the two pieces cemented together.

This, the largest of the three Birsay ogham-stones, is a building slab, most probably carved along its only exposed surface when in situ in a wall, rather than erected specifically to bear an inscription. Unless further information comes to light it is not possible to determine whether the carving occurred while the stone was part of Thorfinn's Palace' or, rather while the stone was part of a much earlier construction.


BIRSAY 1

77

CARVING TECHNIQUE The photograph supplied by Padel is of a rubbing of the stone, rather than the surface of the stone itself [see fig.], therefore it is difficult to discuss the technique or condition of the carving in detail. Padel commented that the technique was very similar to Lunnasting and Formaston [MS notes]. The inscription starts several inches in from the left edge and runs parallel with the long edges to within an inch or two of the right edge. Thus it is more or less centred vertically but is more extensive to the right on the horizontal axis. The stem is \9}á" (c.495mm) long though the inscription occupies only Wá" (c.370mm). Padel interpreted this as indicating a lack of planning. The lettering, however, is placed centrally on the stem, so the ample length of clear stem at either end may instead be a stylistic feature. The far end of the stem is not definite, but seemed to Padel to end before the edge of the stone [MS notes]. Padel comments on the air of 'hurried casualness' of the carving of Birsay 1 [MS notes].

DESCRIPTION OF OGHAM INSCRIPTION Three features combine to give an unambiguous indication of the direction of reading - the forward slope of b- and h-aicme consonants, the hammer-head A (8) and several angled vowels. The letters are generally fairly tightly spaced, though successive letters of the same aicme are clearly differentiated, i.e. 314, 6/7, 10/1L The spacing to either side of the hammer-head A (8) is more generous, perhaps to give room for the cross-stroke. The gap after the first letter is wider than might be expected, unless Padel's 'false start' stroke is allowed. The first few inches of stem appear to lack lettering. Padel rejects two strokes either side of the cemented crack even though 'they appear to be of the same technique as the inscription' but 'cannot be fitted into any rational scheme, as far as I can see' [1972:58]. The inscription continues as follows [see fig. for key to numbering]: 1

A single stroke across the stem. The upper (left) half is perpendicular to the stem but the lower (right) slopes forward at an angle of about 45°. The two meet, not exactly on the stem, but slightly below it. The other m-aicme letters are straight across the stem so the interpretation M is not appealing. Padel suggests a 'feather-mark', i.e. a non-phonetic directional indicator —>—,


BIRS A Y 1

78

but there are no extant examples of this asymmetrical form. It could be that the carving represents an A, begun as a simple perpendicular cross-stroke but modified into an angled vowel. Without examining the stone it is impossible to offer a definitive interpretation. Padel mentions a second oblique stroke to the right of the stem just beyond this letter which he says is artificial, though not visible on the photograph. This he interprets as a possibly false start on the next letter since, in his opinion, it cannot itself be a letter. It would be hard to account for the larger than expected gap between this letter and the next unless there was some such problem. 2

Two angled vowel strokes across the stem pointing in the direction of reading - O. The second of these wavers to the right of the stem. According to Padel this part of the stroke is carved double implying the carver tried to improve it.

3

Five strokes to the right of the stem sloping forwards - N. The first of these has been recut and the second continues across the stem at a more perpendicular angle, in other words this stroke closely resembles L Padel thought the protrudence probably accidental, which does not seem to have been the case with 7. The spacing of the individual strokes suggests they are to be taken together as a single group. The strokes are fairly evenly spaced and parallel (to the right of the stem) though each is successively longer than its predecessor. If the stroke on the far side of the stem is not accidental it might have been added to rectify an earlier mistake, either to insen a missing letter (H?, A?) or to separate the five strokes into 2 + 3 .

4

As above, though not quite so clear - N. The first stroke of the group is at a slightly more acute angle than the rest, which may, as Padel suggests, be a deliberate attempt to differentiate this group from the previous one. Padel says the third stroke does not quite reach the stem though this is not clear from the rubbing, nonetheless the interpretation is not in doubt.

5

Two angled vowel-strokes across the stem pointing in the direction of reading - O. Padel described this letter as careless, since, rather than meet at a point on the stem, the two pans of the first stroke overlap in a little cross just below it. Though feint in the photograph, Padel says the letter is not in doubt.


BIRSAY 1

79

6

Five long oblique strokes across the stem, closely spaced - R.

7

As above, with the first stroke having been re-cut - R. The two letters are very clearly spaced.

8

A single stroke perpendicularly across the stem with a horizontal cross stroke towards the right distal end. Even though the cross-stroke is not at the tip of the stroke there can be no doubt that a 'hammer-head A' was intended, if carelessly carved.

9

Six strokes to the right of the stem sloping forward. The first stroke has been recut. According to Padel the third is feint and does not quite reach the stem, though again this is scarcely discemable from the rubbing. If the stroke were discounted a reading LV would be possible. Padel suggested that the carver may have begun to carve the V too close to the L, but realized his mistake before completing the stroke, hence its faintness. If one were not prepared to discount this third stroke one could observe that the third and fourth strokes are very slightly more generously spaced than the rest, which could be taken to imply 3+3. i.e. VV. A third possibility is that the oghamist has simply miscounted, that the sixth stroke is superfluous and that the letter is N. There are thus three possible readings - N, VV, LV - and it would be unwise to privilege one above the others without examining the stone. According to Padel the elongation of the fourth and fifth strokes, apparent on the rubbing, is an illusion.

10

Five long oblique strokes across the stem, sloping forwards - R. Padel read the middle stroke of this group as having 'bent tips', in other words being the character //J7/, found on Birsay 2 and Altyre. The alleged flaps are not immediately apparent on the rubbing. In his notes he describes the third stroke as 'very curved', but thought it 'very possibly not intentional'.

11

Five long strokes across the stem, sloping forwards - R. More generously spaced than previous letters but clearly a unitary group.

The stem continues for a further inch or two, apparently without any further letters. Padel identified a 'very indefinite' stroke to the left of the stem and continuing across it for a shon distance. Since there is nothing either immediately before or after it Padel was happy to reject this stroke. Which gives the following reading:

(m/h/a)ONNORRA(n/vv/lv)RR


BIRS A Y 1

80

FORM OF SCRIPT The form of the script used is a simple though slightly developed form of Type II, i.e. lib. The more developed features being that all the letters slope forward, the vowels are angled, and there is one forfid, the hammer-head A. More simple features are the lack of marked spacing between letters and the lack of bind-strokes, apart from the seriffed A. The lack of word division may not be significant if, as appears likely, the text is a single word, a personal name. The letters are 'squarish' in outline, in other words a five letter stroke has a similar vertical and horizontal length.

INTERPRETATION OF TEXT The text appears to be complete, unless the strokes Padel identifies either side of the cemented crack and at the very end turn out to be ogham letters. Since the site was inhabited for such a long time and the archaeological context of this stone is so vague, the date of this inscription could lie, potentially, any where from the early Pictish period through to the later Norse, say sixth to twelfth centuries. Because of this long date range and the known cultural affiliations of the site, the text could be in Irish, Pictish (i.e. Brittonic) or Norse.

The (m/h/a)ONNORRAvvRR

might represent

(m/h/a)onnorrqffrr or

(m/h/a)onnorrauurr. The final R(R) is reminiscent of nominative ending of a Norse male personal but value /v/ for V seems unlikely, thus the Norse male personal names Ónarr and Ávarr [Lind 1905 s.n.n.] are unlikely to offer a solution. Nonetheless, Padel suggested the text might represent Old Norse and offered ON Norrcenn 'Norse, Norwegian', though, as he himself admitted the interpretation is unlikely, not least because the adjective would not end in -r. Likewise his suggestion that 0 is ON 6 'from' is not very likely.

If the text is read (m/h/a)ONNOR RANRR or even (m/h/a)ONNO RRANRR it might be compared with the runic filibus ranru of Birsay 1, with the gemination of the R merely a hang-over from traditional ogham orthography. Regardless of whether or not Olsen's expansion of ranru to r(eist) enn rit(nar) 'carved these runes' [1954:164] is accepted (and as explained above, this is by no means certain), it is not inconceivable


BIRSAY 1

81

that the ogham text may bean attempt to render that sameformulawith ogham letters. This would leave (m/h/a)ONNO(R) as a personal name. Unfortunately, none, either Celtic or Norse, immediately suggest themselves.

DISCUSSION See general discussion of the three Birsay ogham-slabsfollowingthe entry for Birsay 3.

BIBLIOGRAPHY Radford 1959:5 (mention only); Cruden 1965:25 (mention only); Padel 1972:2, 10-11, 16, 55.

PREVIOUS READINGS Padel

BIRSAY

mONNORRA(lv/n)RR

2

As discussed above, this stone, dubbed the 'Shore Stone' by Padel, was 'picked up on the beach below the Brough' [Radford 1962:174]. Padel raises the possibility that the stone is water worn [MS notes] which may indicate that it had lain on the shore for some time before discovery. Presumably this was a secondary context and the stone had tumbled down from its place in some structure on the Brough proper. Unfortunately the precise find-spot seems to have gone unrecorded so we cannot begin to investigate from which structure it might have fallen. This leaves us without any indication as to the date of the inscription. As with the larger of the two early discoveries the ogham could lie anywhere from the sixth to twelfth centuries and relate to an Irish, Pictish (Brittonic), or Norse-speaking milieu. The following account is based on Padel [1972:56-58] and the Department of the Environment photograph A 2894-2 [see fig.].

DESCRIPTION OF OBJECT Described by Padel as a Very irregular slab', on its flat edge survives a weatheredfragmentof ogham text.


BIRS A Y 2

Stone:

Local flag-stone

Dimensions:

14#" x 9lá" x 2" (approx. 0.37 x 0.24 x 0.05 m)

Condition:

Badly weathered, text incomplete.

82

Padel does not illustrate the stone nor give any indication of the relative position of the inscription on the stone, other than to say it is on the 'flat edge1. Though he describes the inscription as very weathered, Padel does not discuss the appearance of the stone itself, nor whether or not it is intact. It is probable that, like Birsay 1 and 3, it is a building slab inscribed in situ along the only visible surface. However, until a photograph can be obtained or the stone examined in person, I cannot rule out the less likely possibility that it is the remains of a small slab, either recumbent or upright, erected specifically to carry the ogham text.

DESCRIPTION OF CARVING The photograph provided by Padel is not of the stone itself, but rather of a drawing, probably based on a rubbing. For this reason it is not clear which marks are strokes and which the result of casual damage. Also the relative weighting of strokes is unclear making it uncertain how much importance is to be attached to small additional lines etc. The stem is visible for approximately 3'4" (c.80mm), but, according to Padel, die stone is too badly weathered to indicate how much has been lost. The scale of the letters and the reading, as best as it can be discerned, imply that this is only a short section of a longer text. There is no clue however, as to where in the sequence this portion occurred. The stem wavers and appears to be more finely incised than the letter-strokes. It begins at the point of hypothetical intersection with the first stroke of 7, but it is unclear if this is a clear terminus or just where the text emerges from wear. Padel says that the stem 'seems to continue crookedly beyond lower edge?1 without specifying how far this is from the patch of wear. The direction of reading is indicated by the slope of 2. The extant portion is as follows [see fig. for key to numbering]: 1

The remains of a single oblique stroke below the stem, sloping forward. The stroke does not quite meet the stem, presumably because of wear. As it stands it should be read as be B. Padel


BIRS AY 2

83

notes, however, that the surface opposite this stroke is worn away. If the stroke originally continued at the same angle across the stem it would be M, though if produced for an equal distance on the opposite side it would come close to intersecting the first stroke of 2. Padel makes the interesting suggestion that the stroke could be the lower half of an angled A (->*). Angled vowels appear on both the other Birsay oghams and the objection that letter 3 is a straight-stroke vowel is not compelling since the two forms co-exist on Bressay, Burrian, Cunningsburgh 3, Formaston, and Lunnasting. If strokes have been lost from the beginning, it is, of course, possible that this stroke is thefinalone of any letter in either the b-, m% or angled vowel-aicmi. There follows a small nick parallel with the first stroke, also not touching the stem, which Padel thought 'probably due to weathering'. If it is the remains of a letter it could be a second stroke similar to the first. The above arguments would apply equally, though it is even less likely to be an m-aicme letter since it would intersect with the first stroke of 2 almost as soon as it crossed the stem. Though this second stroke appears to be at an angle, such a short section survives that it is just possible that it is the remains of a short straight stroke across the stem - A. This would facilitate the reading MAQI, but would strain the bounds of probability given the limited space. Padel suggests it may be a feather mark [MS notes] 2

Five strokes above the stem, sloping forwards - Q. Only the pan of the first stroke nearest the stem survives, though Padel notes that it is not as short as appears from the DOE photograph and the reading is not in doubt. The strokes are neither entirely straight nor parallel and each overshoots the stem slightly. This gives the impression of an inscription carved quickly and with confidence.

3

Five long strokes across the stem - 1 . These are not balanced on the stem but extend further below than above. The middle stroke appears slightly curved, Padel says it is carved double. It may be just a botched stroke, or alternatively the 'flapped* or 'angled S-shape' stroke which appears as a singleton on Latheron, and at Lochgoilhead and possibly Altyre in the middle of a group of five on Birsay 1. There follows a dot close to the stem. Padel thought it might be one


BIRSAY 2

84

of a pair of word-dividers, like those on Bressay, and noted that there was a mark opposite, though he conceded it could be just weathering. Another possible interpretation is that the dot is intended to differentiate between two successive letters of die same aicme. The analogy for this is Lunnasting character 16, though in die Shetland example the dot has been interpreted as having been added later to clarify two strokes carved too close together. The Birsay letters are already well-spaced, the gap between them is already about two and a half times that between 2 and 3. Without seeing the actual stone it cannot be ruled out that the dot is merely die result of casual damage. 4

A single stroke across the stem - A. Since the h-aicme strokes (2) are clearly sloped it is unlikely that m-aicme consonants would be perpendicular to the stem, hence the reading M is rejected. From the photograph it appears that three short, thin, oblique strokes are appended to the left of the lower tip of this stroke. If they are genuinely part of the inscription they are without parallel in the epigraphic ogham corpus. A suggested interpretation is that the oghamist has inserted a three stroke letter of the h-aicme (T) which was omitted in the initial carving. Padel simply says the letter is carved with several strokes, so perhaps the photograph is misleading.

5

After a gap at least as long as that between 3 and 4 there is a single oblique stroke below, but not quite touching the stem, sloping forward. Padel notes that the surface opposite is worn away and that the stroke may have continued across. Like the first stroke, it could be pan of any letter of the b-, m-, or angled-vowel-a/c/w. Clearly, some text has been lost from this end, but there is no indication of how much.

The following tentative reading is possible: -(b + /a + /m + )(a/ )QI(:/ )(t/ )A(b + /a + /m + )-

FORM OF SCRIPT Again, since the extant inscription is so short, there is little to say about the form of the script other than that it is simple Type II, broadly similar in aspect to that of Birsay 1. The flanged letter, if that is what


BIRS A Y 2

85

it is, is an obvious connection between the two, but they also share similar over-all proportions in lettershape and spacing. If the dot between 3 and 4 is accepted as a word-divider then this would imply a leter date, possibly in the Norse period (see Introduction for discussion of word-dividing dots).

INTERPRETATION OF TEXT Since the text is so shon and doubtful there is little that can be sensibly said about it. Certainly it can neither be proved or disproved to be in Irish, Pictish, or Norse. The aQI could be Olr. MAQI, and if so would be the only Scottish example of this word spelled with single Q.

DISCUSSION See general discussion of the three Birsay ogham-slabs following the entry for Birsay 3.

BIBLIOGRAPHY Radford 1962:174 (mention only); Padel 1972:55-61.

PREVIOUS READINGS Padel

BIRSAY

-(b/a)QI(:/ )A(b/a)-

3

The third ogham inscribed flag-stone from the Brough (BB 80, layer 543 ref.2761) was unearthed during the Summer of 1980 in Site IV S. The stone had been re-used in an area paved with large sandstone flags. No other finds were recovered from this context which was sealed 'by one of the rubble spreads between Sites E and S' [Morris 1981:36]. Calibrated radiocarbon dates obtained for this context were in the range 560-769, centering on c.648. Thus the secondary context of the ogham can be dated approximately to the seventh century, plus or minus 50 years. I examined the stone in the Department of Archaeology, University of Glasgow in March 1994, and most gratefully acknowledge the information and assistance


BIRS A Y 3

86

provided by Prof. Christopher Morris on that occasion. Prof. Morris also kindly supplied me with contact prints of the stone [see figs.].

DESCRIPTION OF OBJECT A parallelogram-shaped building stone, undecorated apart from a short inscription in bind oghams [see fig.]. Stone:

Local - Stromness flag series (orange brown calcereous siltstone, ferric rich)

Dimensions:

0.31 x 0.35 x 0.065 m

Condition:

Flag intact. Only middle section of inscription survives. Some individual strokes badly worn, top of third letter lost in spall.

The inscription is incised along the flattest of die narrow feces. This is not a continuous smooth surface but rather disjointed. There is a marked disparity in the level of the surface immediately after the final stroke of 4 which evidently pre-existed the carving of the inscription. Other spalls and fractures post-date the carving which has suffered much. Immediately following the single stroke of 7 the surface has flaked off taking with it the outer tip of the stroke and any subsequent letters. Also gone is the distal portion, and presumably the bind stroke of the last three strokes of 4. Surface abrasion has worn away parts of all letters except for 3 and 5, so a measure of doubt surrounds all but these.

DESCRIPTION OF CARVING The stem is visible for 190mm, but may have continued further in both directions. It runs roughly parallel with the long edge, wavering slightly. The maximum breadth of ogham-inscribed face is 50mm and the ogham stem is placed not medially, but only about a third of the way up the narrow edge. The strokes have been sharply and fairly deeply incised with a blade, the stem a little more substantially carved than the letter strokes. The direction of reading is not in doubt, since the slope of all letters but the first is diagnostic. The inscription is as follows [see fig. for key to numbering]: 1

The extant portion of the ogham begins with what appears to be a single long oblique stroke


BIRSAY 3

87

across the stem. The right hand portion is worn. There is some doubt over this stroke because die stone has suffered much damage in this area. There is nothing left of the surface immediately before this stroke and it is possible that some ogham has been lost from the beginning. Possibly M. 2

Between 1 and 3 there is a gap of 45mm. After the first 30mm there is the remains of what appears to be a stroke. The only portion to survive is the bit on the stem itself, though there may be a faint stroke above. If this is a letter it may be A. Half-way between this stroke and J is a faint stroke to the right of the stem. The gap is less than that between the other letters and this slight stroke is highly doubtful. If extended across the stem it would intersect with the first stroke of J, for this reason it cannot be interpreted as the remains of a vowel stroke. Possibly B.

3

Five strokes to left of the stem, roughly parallel, sloping backwards, bound - Q. Each stroke is slightly longer than the last (in the range 15-30mm) giving the whole a wedge-shaped appearance. The first stroke protrudes slightly beyond the bind stroke, which itself over-shoots the first stroke to form a tiny cross at the intersection. This lack of precision is reminiscent of the Burrian ogham and both it and Birsay 3 have the air of being carved slightly carelessly, but with confidence and speed.

4

Five strokes to the left of die stem. A substantial chunk of the surface has been lost at this point, taking with it between a third and a half of the distal ends of the third, fourth, and fifth strokes. On die published photograph [Morris 1983 fig.40] there appears to be a bind stroke along the tops of the last three strokes, in fact this is just the shadow cast by the broken edge and there is no bind stroke on the stone at this point. From a casual glance one gets the impression that the first two strokes are parallel with each other but not with the final three, suggesting the reading 2 + 3 DT. A closer examination, however, will reveal that the first and fifth are exactly parallel and it is the intervening strokes which are at various angles. A look back at 3 shows that the oghamist took no particular care to keep the strokes entirely straight or parallel so, though 4 is even less regular than 3 it is not unacceptably so. Since the bind-stroke continues after the second stroke


BIRSAY 3

88

as far as the break there seems no justification in not taking the five strokes as a single group Q. After the fifth stroke there is a break in the surface of the stone which clearly pre-existed the carving of the ogham. The inscriptions continues at the lower level. 5

A single stroke to the left of the stem, sloping backwards - H. This score is rather deep and not in any doubt.

6

Two pairs of strokes either side of the stem meeting to form a forward pointing arrow-head (> >•) - O. The first two strokes do not quite meet, the right hand stroke intersecting with the stem slightly beyond the left one, but the discrepancy is slight, certainly not enough to warrant interpreting them as separate letters (HB), and further suggests the rapid and facile execution of the inscription.

7

After a gap of 18mm a single stroke to the right of the stem. The surface is fractured immediately after this stroke and any subsequent carving has been lost. There is a faint shadow to the left of the stem opposite this stroke but as it stands it is insufficient to warrant explanation as the upper half of an angled vowel-stroke. This letter is best interpreted as B or any subsequent letter of the b-aicme (i.e. L, F/V, S, or N).

It is unlikely that we have the entire original inscription. Approximately a further 60mm of stone would have been available originally, sufficient for two to three letters. Even that would be surprisingly, though not improbably, short. However, if the ogham was indeed carved when the stone was in situ in a wall it is possible that the text continued across the join and onto the neighbouring stone. The tentative reading is as follows:

-m(a/ )(b/ )QQHO(b+)-

FORM OF SCRIPT Despite the brevity of the surviving inscription Birsay 3 displays a number of significant features, the bindstrokes on the consonants and the angled vowels. Both these features would normally be considered to be late in the typological sequence of ogham development. The comparatively early radio-carbon dates obtained for this item force a re-examination of the absolute dating of the typology. The Birsay 3 dates


BIRSAY 3

89

centre in the mid-seventh century, though range from the mid-sixth to mid-eighth, requiring that the typological sequence be stretched back in time. That this is necessary underlines the fact that script typology can be a guide to terminus post quern dating only.

INTERPRETATION OF TEXT With a text as fragmentary and doubtful as this one it is difficult to make clear assertions as to meaning. The two most likely readings are -mabQQHO(b+)- and -maQQHO(b+)-. The first holds out the intriguing possibility that is starts MAB, i.e. Brit, mab < map-son*. That maplmab was also the Pictish word for 'son* is suggested by the annalistic reference to the death of Talorcan maphan [AU 725]. Thurneysen took this as a form of maban with ph for pp corresponding to the cc of Olr. mace, the gemination due to affective intensification [Thurneysen 1946:92-3 § 150]. A. O. Anderson's alternative explanation is that maphan is for map-hoen 'son of Owen* [1922:222 n.7]. Either way the word attests to the use of the Common Celtic word for 'son* among the Picts. To read MAB at the beginning of Birsay 3, however, would leave the awkward cluster QQHO. In any case the reading is highly doubtful and the more obvious interpretation -maQQHO(b+)- is to be preferred. MAQQ can be taken as a fonn of Irish MAQQI > mac 'son', the same spelling of this word being attested elsewhere in Scotland at Latheron (Dunnodnatt Maqq Net-) and probably at Formaston (Maqq[?]Talluorrh). As far as I am aware, however, MAQQ is nowhere attested in Ireland, the post-apocope forms of MAQQI being MAQ and MAC not MAQQ and MACC [McManus 1991:124 §6.30]. This suggests that Birsay 3 is not a purely Irish text, like its near neighbour and younger contemporary Buckquoy, but rather is in some way assimilated to the local language (presumably Pictish).

If there is a portion of the lettering lost from the beginning, then Birsay 3 may be an example of the 'X MAQQI Y' formula so familiar from Ireland, and found in Scotland at Latheron and elsewhere. Possibilities for the Y slot, personal names beginning 0(b+)n are Ok, Olchobur, Olchu, Ollam, Ossán,


Ossine, and Onchu, to name a few listed by O'Brien [CGHisrm.].

g IRS A Y 3

90

This, however, would leave 5

unaccounted for. There is considerable uncertainty over the correct transliteration of the first letter of the h-aicme. McManus flavours a hypothetical /j/ lost before the Old Irish period [1986:26,1991:36-7], SimsWilliams favours a Latin-derived h [1993:162-70]. Though the character never appears on classical Irish ogham pillars, in the later, say ninth century and later, oghams, which are merely ciphers for standard Old Irish manuscript spelling, it is used for the letter h. This might explain the presence of H in Birsay 3, a mute letter before a vowel [Thuraeysen 1946:19-20 §25]. In Scotland, however, ogham 'H' occurs in situations, such as NEHHT- and TALUORH where one might expect a spirant, either fXJ or ///. This explanation is less easy to tally with the evidence of Birsay 3 since we would not expect lenition in this position.

One problem with interpreting Birsay 3 as a 'X MAQQI Y' type text is that there is scarcely enough room for even two additional letters before 1. For this reason, unless one is prepared to imagine the inscription beginning on a previous stone and continuing onto this one, the (X) MAQQ HOb* interpretation has to be abandoned. Rather than a patronymic phrase, we may have a solitary compound personal name of the common Irish type Mac-N. The correct transliteration of H is still as much of a problem but there is a saint of the Ui Bairrche called Mac-Onchon recorded in the genealogies of the Lagin in the Book of Leinster [CGHA6, §121 a 42] and the same name may be behind Birsay 3.

Birsay 3 displays a number of features typical of Type lib oghams. The sloping of consonants, clear spacing between letters, and, of course, bind-strokes are all widespread aids to legibility. The angling of vowels may be a further example of attempts to enhance legibility, by adding another feature to the list differentiating vowels and consonants, but it is possible that they represent nuances of vowel sounds not covered by the standard straight vowels. Nothing can be read into the lack offorfeda since the fragmentary sample is so short. The prime significance of the script of Birsay 3 is in demonstrating that bind-strokes


BIRSAY 3

91

and angled vowels were in use already by the mid-seventh century or thereabouts.

DISCUSSION See general discussion of the three Birsay ogham-slabsfollowingthis entry.

BIBLIOGRAPHY Morris 1981:36,40 (mention only); Morris 1982:81 pl.41 (ill. upside down); Curie 1982:86 (mention only); Morris 1996:54-5 (quoting author) fig. 3.8 (right way up).

PREVIOUS READINGS None

GENERAL DISCUSSION OF BIRSAY OGHAMS The size and shape of the three Birsay slabs and the position of the text on each stone suggests that the Birsay oghams were carved while the stones were in place as constructional elements in walls or other structures. The inscriptions are most easily classed as informal graffiti. Birsay 1 appears more or less complete, but Birsay 3, despite covering most of the available surface, is so shon that it is at least possible that the text began and/or continued on a neighbouring slab.

The three texts are carved in different hands and though Birsay 3 is firmly dated, there is no reason to assume that all are contemporary. Nonetheless three oghams from one site is a remarkable concentration, especially if one adds to the total the whorl from nearby Buckquoy. These four oghams constitute twothirds of the total of Orkney oghams, but whether this accumulation reflects greater Irish influence at Birsay, or higher levels of literacy, or merely the intense archaeological attention lavished on Birsay for over fifty years is unclear.


BIRS AY 3

92

Since none of the Birsay slabs were found in their primary position it is impossible to know if the oghams ran horizontally or vertically. Given the shape of the stone we might presume die former. Carving ogham horizontally represents a major departure from the classical norms of ogham inscriptions, but if the inhabitants of Birsay were familiar with non-lapidary ogham, as, for instance, the circular Buckquoy ogham or the inscribed knife-handle from Gurness, then the convention may no longer have been that strong. Lack of comparable Irish evidence prevents an evaluation of whether such features were innovative and specific to Scotland or had always been part of an Irish chattel-ogham tradition parallel with the classic pillars.

ORKNEY ISLANDS

LOCATION MAP - Orkney with inset Brough of Birsay and Point of Buckquoy [After Morris 1990:70 fig.3]

BIRSAY - Plan of main Pictish and Norse settlement [Curie 1982 ill. 3]

BIRSAY - Sculptured slab with Pictish symbols and figures [Curie 1982 ill. 4]

BIRSAY - Cross-incised stones: Large [Ritchie 1986:12] Small [Curie 1982 ill.45 (detail - no.607)]

BIRSAY - Rune-inscribed fragments, Birsay II (Barnes OR 8) [photo. Historic Scotland, Neg. A 995-1 (1938), Crown Copyright]

BIRSAY - Rune-inscribed fragment, Birsay III (Barnes OR 9) [photo. Historic Scotland, Neg. A 995-3 (1938), Crown Copyright]

0

50 mm

BIRSAY - Rune-inscribed seal's tooth, Birsay IV (Barnes OR 11) [Curie 1982:ill.37]

BIRSAY 1 - Rubbing of ogham inscription [Photo. Department of the Environment (Historic Scotland) No. A 2894-2, June 1971 (Crown Copyright)]

^Méw^'^L'W'íptiP^i BIRSAY 1 -Ogham inscription [Padel 1972:56]

-- some kind of A (see below).

31

Five short strokes to the right of the stem, sloping forward - N.

32

As above - N. Padel read another single long stroke across the stem, following this (A) though he claimed the portion of the stroke nearest the stem to the left was missing. I could see no trace of it.

Which gives a reading: CRRO( /s)SCC : NAHHTWDDAddS : DATTRR : aNN(-)

CATALOGUE OF THE OGHAM INSCRIPTIONS OF SCOTLAND Left Edge:

BRESSAY

125

Macalister inserted dots between 6/7 and 819 but I could not make these out. Padel gave

only the second pair. More of the stem is visible before the first letter but, once again, we may be sure the beginning is intact. The slope of b- and h-aicme consonants indicates the direction of reading is from the bottom up, as expected [see fig. for key to numbering]: i

A single short stroke to the right of die stem, sloping forward, rather faint, but not in doubt - B.

2

Four long strokes, perpendicularly across the stem - E.

3

Five short strokes to the right of the stem, sloping forward - N. The strokes of this letter cross the stem very slightly, but not enough to warrant reading R.

4

As above - N. This letter is very worn, but not in doubt.

5

A unique character. Five long strokes undulating across the stem, sloping slightly backwards some modified form of I ? (see below).

6

Four short strokes to the right of the stem, sloping slightly forward - S. This letter is very worn, but not in doubt.

7

Four long strokes perpendicularly across the stem - E.

8

Four short strokes to the right of the stem, sloping forward - S.

9

A single long oblique stroke across the stem, sloping forwards - M.

10

Four long strokes perpendicularly across the stem - E.

11

Five short strokes to the left of the stem, sloping backwards - Q.

12

As above - Q.

13

Two short strokes to the left of the stem, sloping forwards - D.

14

As above - D. These four successive letters of the h-aicme (11-14) are unambiguously spaced.

15

Five long oblique strokes sloping forward across the stem, overlain with five identical strokes sloping backwards - the cross-hatched HRforfid, very carefully carved (see below).

16

A diamond-shaped lozenge sitting on the stem - an angular version of ihe-frforjfd (see below).

17

A single long stroke perpendicularly across the stem - A.

18

Four short strokes to the right of the stem, sloping forward - N.


19

BRESSAY

126

As above-N. Thereafter the surface has worn away. If further letters are lost there cannot have been more than one or two. There need not have been any.

Which gives a reading: BENiSESMEQQDDRoANN(-)

FORM OF SCRIPT Both lines of the Bressay ogham are written in the same hand, a clear and regular Type lie form of the script with word-division dots and a handful of forfeda, two of which are unique. The stem is more substantial than the letter strokes probably, in pan, because it runs along the grain of the stone. The long, straight vowel-strokes take up to the full width of the ogham band. There is variation in the degree to which b- and h-aicme letters slope, some are almost perpendicular while others are sharply oblique, but all slope in the anticipated direction. Component strokes are parallel within letters and evenly spaced. All letters are clearly spaced, often with as much as a stroke's worth between them. Successive letters of the same aicme, for instance right-////2//J, are arranged to prevent ambiguity.

The letters are evenly

arranged along the entire length of the text and, like Lunnasting, there is no bunching towards the end, indicating a degree of planning absent, for instance, at Formaston, Newton, and Buckquoy. The form of right-4 suggests that an error may have been made in its initial carving.

The ratio of stroke-length to stroke-spacing gives the Bressay letters a profile closer to the mean than the rather squat letters of some of the later inscriptions like Burrian, Cunningsburgh 3, and Lunnasting. Like the three aforementioned oghams, Bressay smells of the scriptorium. The proportion of the strokes and the technique with which they were cut strengthens the manuscript-like appearance of the script (though, in contrast to Burrian, Bressay is carved on a monumental scale). The curves of letter left-5 are more suited to the pen than the chisel, and the unattached voids created by letters left-76 and, especially, left-75 are vulnerable to becoming detached and are therefore rather unsuitable for epigraphic purposes (though the carver here knew the stone - none of the enclosed lozenges have been lost). In general appearance Bressay is, perhaps, most similar to Formaston with which it shares angled vowels and the-bforfid, but


BRESSAY

127

not word division. This last feature is found at Lunnasting, which has bound letters, angled vowels, but no other developed features in common with Bressay. A little closer is Burrian which has bound letters and lacks word-division, but does share with Bressay angled vowels and the RR forfid.

There are fiveforfeda at Bressay, one of which is repeated. Three of these are paralleled on other Scottish monuments and also in Irish manuscripts, but the other two are unique to Bressay:

Character Right 4 (,| | ( ) : This is probably not a forfid but rather a corrected error in carving (see discussion above).

Character Right 20, 21: This rabbit-eared character is unique to Bressay. Since it consists of two strokes to the left of the stem previous authorities have taken it as some kind of modified D, presumably the voiced spirant /Ó/. This seems by far the most likely explanation (and if it is not we can only guess at this character's value), but implies a breakdown in the old convention DD for /d/, D for /Ó/ (presumably the initial D in the following word has the value /d/ because of the practice of avoiding geminate consonant symbols in word-inital position [Harvey 1987]).

Character Right 28 (->): Angled vowels are discussed in the entry on Formaston where it is suggested that they may indicate long vowels. Whatever its value, we may take it as some form of A, contrasting with the standard A-character occurring in the previous two words (/0, 18, 23).

Character Left 5: (12ZZZ) This unique character, five undulating strokes sloping backwards across the stem» is presumably a vowel, since it occurs between the second and third of a group of three Ns. As a fivestroke vowel we may take it as some form of I, rather than R. Angled-stroke I (?/t/) does not occur at Bressay, but since angled-stroke A does (?/5/), we may have warrant for rejecting /f/ as the value of this character. Though it is really rather different from them, it may be useful to compare úúsforfeda with


BRESSA Y

128

instances elsewhere in Scottish oghams of cross-strokes with flaps, e.g. Birsay 1-2, Latheron, Lochgoilhead, though most of these face in the opposite direction.

Character Left 15:

A character consisting of/awr-stroke cross-hatching ()$$) appears in the late

eighth or ninth century Bern ogham alphabet and syllabary under a label which has previously been read as RR. Sims-Williams has argued convincingly that instead this is SS [1992:38-39 fig.5], but rather than remove the manuscript key to the Bressay character's value, he has in fact demonstrated the precedent (with the letters BB and LL) which shows that the five-stroke character here C$$K) represents RR. SimsWilliams suggests that monograms for LL and RR were generated by 'late but keen-eared ogamists' to represent the geminate or long liquids /11/ and /rr/ [1992:72]. The context of this character at Bressay, after DD and before OA, fits Sims-Williams's interpretation, but if the forfid is /rr/, what then is meant by the doubling of the ordinary R character on the other line of text (right 213 and 26127) ? As Macalister points out, thought it may be easier to read, this character is not easier to carve, and, since there was ample space for two characters, we can only assume some nuance of phonology/orthography drove its inclusion. The only other instance of this character in practical use is at Burrian. It appears in a cross-hatched alphabet in the Book of Ballymote [No.64, Calder 1917 Aur. 6085], but there all the letters are doubled and the device has no phonetic significance.

Character Left / 6 : ( - £ - ) This character is discussed in the entry on Formaston [see also Sims-Williams 1992:58-60]. Its earliest vocalic value seems to have been /D:/, though it later came to be used for the diphthong oífóe, and ultimately the digraph oi.

Word-division: The use of pairs of dots to indicate word division is discussed in the Introduction where it is argued that, while a late feature, the device cannot be taken as diagnostic of a date in the Norse period (though we know on other grounds that Bressay certainly does date to the ninth century or later). The suggestion that word-division dots were a feature borrowed from the runic tradition was first made by


BRESSA Y

129

Graves, in connection with this stone [1855:248-9].

INTERPRETATION OF TEXT The transliteration of H and V is discussed in the Introduction. As for the forfeda, we have some notion of the likely values of right-25, left-75 and -76, but, as for right-79 and -20, and left-5 we are only guessing. For all the problems of both transliteration and interpretation, the use of these supplementary letters suggests an attempt at orthographical precision. Despite its hybrid character, the Bressay inscription is far from slap-dash. Word-division is indicated on the right edge, so we may not introduce further divisions, but, pace Macalister, it is not indicated on the left edge and must be supplied. A further question is the order in which the two sides are to be read, if indeed they are to be taken sequentially. In the earlier oghams it might have been natural to give the left side precedence, but in some later examples this old convention appears to be breaking down (cf. Brodie, Formaston, Golspie). We might take the longer line first, but the allocation of letters might simply have followed word-boundaries* in which case this is not a sure guide. It could be that the two are separate sense units and it does not matter which goes first, certainly no indication is given on the inscription itself.

The interpretation of the text is greatly helped by the word-division on the right edge: CRROSCC : NAHHTVVDDAddS : DATTRR : aNN

CRROSCC All have recognized this as meaning 'cross' (Latin < crux, see entry on Burrian). What is of particular interest is that the form here /krosk/ is virtually identical with the crosg an attested variant of the more common Modern Scottish Gaelic crasg, genitive croisg (as opposed to the Old Irish cross. Modern Irish cms). The final g has not been satisfactorily explained but is standard [Padel 1972:31]. Watson gives a sole example oicrasg meaning a carved stone cross, in a place-name recorded in the seventeenth century of a monument near Dornoch [Watson 1929:486 n. 1].


BRESSAY

130

NAHHTWDDAddS The final S of NAHHTWDDAddS looks suspiciously like a Germanic genitive case ending, doubly so since the word appears to be in syntactically genitive position - '(the) cross of Nahhtuuddadd\ Following Graves, Brash suggested [1879:355] that the second word at Bressay might commemorate the Naddodd 'of the Faroes' who, according to later Norse sources accidentally discovered Iceland when blown off course en route to Orkney in 861 [Landnamabok 4, ch.4; 26, ch.63; OlafTryggvisson's Saga ch.113 (ch.199 in Flateyiarbok 1.248); cited by Anderson 1922:338]. possibility is intriguing.

We are, of course, one syllable short, but the

The name Naddodd is presumably composed of naddr 'arrow' + oddr

'spearhead', and though not otherwise attested, is nonetheless a regularly formed Norse name [Lind 1905 s.v. Naddoddr, I owe this reference to Anne Haavaldsen, University of Bergen, and thank her for her generous advice on Norse nomenclature]. The genitive form would be Naddodds, Lind gives the spelling Naddodz which may support the reading of characters 19 and 20 as /5/.

If this name is correctly identified, the spelling appears to have become contaminated by ogham spellings of names related to Nechtani'Nation, NEHT-, NEHHT-, NETU (see entry on Ackergill). The similarity to the Norse name may, however, be fortuitous, and Bressay's NAHHTVVDDAdd(S) may be a Pictish name after all. NAHHTVV- is most closely paralleled by Fonnaston's NE4HHTV-.

In both cases a

Gaelicizrog transliteration F(F) seems less plausible than the more Pictish /w/, which renders unlikely the interpretation of DDAdd as the Old Irish diminutive ending -that, which, in any case, 'is found only with nouns denoting inanimate objects' [Thurneysen 1946:175 §274.4]. On a suggestion of Stokes [1890:40], Rh?s proposed taking the last eight characters, VVDDAddS, as Unddaóó, a less Gaelicized spelling of the name which appears elsewhere as Fothad (Fothud, Fothaid) [1892:297]. The name is well attested in Ireland as a personal name [CGH] and was borne by two Scottish bishops, Fothad son of Bran scriba and bishop of the Hebrides who died in 963, according to the Annals of the Four Masters [quoted by Anderson 1922:471-2], and Fothud ardepscop Alban, bishop of St. Andrews, who died in 1093 (AU). According


BRESSAY

131

to Irish genealogical tradition, the ancestors of the Uithne of Nonh Munster were Na Tri FotMd, the three Fothads, who were reckoned to have come from Scotland and were linked with the Tuatha

Orcandlbdig,

the tribes of Orkney and the Outer Isles [CGH 155 b 26 -156 a 4 (pp. 264-5)]. We may take it that Fothad < Uotados, the pseudo-eponym of the Uotadini > Gododdin. Bressay 's DD would suggests a non-lenited first consonant, Nait-Uudadd perhaps ?

DATTRR Previous authorities have taken this as a form of the Old Norse dottir 'daughter', while querying the likelihood of such a loan. Macalister maintained that, as a 'Bronze Age' population, the Shetlanders had no word for 'daughter' until the arrival of the Norse, and acquired mac from the Irish only when the adoption of Christianity necessitated an equivalent of Latin filias [1940:203]. This outlandish assertion, founded on a belief in non-Indo-European Picts practising matriliny, may be dismissed as a fantasy. There is no need to posit the borrowing of vocabulary for such a fundamental human relationship when it is possible that in this context die word could be a quasi-proper noun. If the person in question was habitually referred to not by their given name, but by the patronymic Naddodsdottir, then the dottir might come to be felt to be pan of their name, and thus untranslatable, rather than a common noun in a descriptive phrase. The spelling DATTRR is some way from dottir, but might reflect a Gaelic-speaker's attempt to render a word which had been heard but not seen written. Certainly, if it is not dottir. a convincing Celtic alternative is hard to find.

ANN In the light of the standard interpretation of the previous word, this has been taken as the female personal name Anna (cf. Modern Welsh Ann). Biblical names for both men and women appear to have made little impact in pre-Norman Scotland and Ireland and, if this is a Gaelic name, it is perhaps more plausibly linked to Ann the name of a female Irish deity and saint, or the more popular Áine [[Ó Corráin & Maguire


BRESSAY

132

1990:s.v.v.]. There is no evidence, however, for a final vowel [pace Diack and Padel]. But perhaps we have been misled in thinking the name must be female, An(n)y gen. Ans, was a Norse male personal name, ('a very, very old name1 [Lind 1905]), borne, for instance, by a man (and his maternal grandfather) mentioned in Landnamabok as plundering and marrying in Ireland before setting off for Iceland [Landnamabok ch. 107.p.45, cited by Anderson 1922:336]. The Old Norse female name On means 'hope' [Lind].

It is noteworthy that both lines of ogham at Bressay end in -ANN, if this is more than coincidence it may indicate that we are not, in fact, dealing with a name at all. In Modern Irish and Scottish Gaelic the adverb ann 'there' is used with the substantive verb to denote existence. No such verb is present. We could take an elided tha as read at the head of the sentence \thd) crosc Nahtudads dattrann\ but to state the obvious in this way seems an unlikely text. For reasons of syntax we may reject any form of the definite article (Goidelic or Brittonic, cf. Cornish and Old Breton unstressed definite article an [Jackson 1953:656-7].

Turning to the left line of ogham, there is not indication of word-division, which must be supplied. The sequence MEQQ in the middle immediately suggests the familiar *X mac Y? formula, i.e. BENiSES MEQQ DDROANN.

MEQQ Presumably this is Irish mac 'son'. The same form is encountered elsewhere on Shetland at St Ninian's Isle and, if not a northern dialectal variant of the MAQQ encountered further south, may be a fairly straightforward rendition of genitive meicc (with the anachronistic Q a fossilized relic of the old spelling reserved for ogham only, as standard in Scottish oghams).

Macalister's ingenious suggestion that

BENISES-MEQQ against DUNNODNNAT-MAQQ, reflects vowel affection is not refuted by St. Ninian's Isle's -ES-MEQQ, but seems highly unlikely.


BRESS A Y

133

BENiSES The precise value of 5 is not known, though we may guess at some kind of I. Brash pointed out that Naddod is recorded as having a grandson Beinir [1879:255]. Beinir means 'to straighten out', and appears in Scaldic poetry meaning 'helper, giver' [Lind 1905]. The genitive form would be Beinis, which, once again, seems one syllable short of what we have here at Bressay.

Ferguson calculated that if the

identification of this Benir were correct the inscription would have to be mid-tenth century, which need not be too late [Ferguson 1887:134-5]. Another Beinir appears as the great-great-grandson of Erp, son of Maelduin 'an earl in Scotland', given his freedom by Aud the deepminded and who sailed with her to Iceland [Landnamabok chs.82-3, Sturla's version ch. 103 p. 158, cited by Anderson 1922:384]. He cannot be the person commemorated, but further attests the use of the name. The Anglo-Saxon Chronicle of 911 records the death in a battle between Anglo-Saxons and Norse, of one Benesing [cited by Anderson 1922:401]. One looks in vain for a Celtic name of this form.

It is possible, however, that further segmentation is required, and that we should separated off the initial BEN, which could be Old Irish ben 'woman', 'wife'. Could we have 'the wife of Ises son of Droan'. Ises does not bring to mind any Celtic personal names. The name / (//) appears in a Bemitraige genealogy [CGH 111 b 50], but if Ises is the genitive, this is not Celtic. Ben- occurs in composition with other nouns to denote a female, or something pertaining to the female (the more common composition form is ban-, e.g. ban-ech 'mare', ban-liaig 'female physician' [DIL]), but what ises in ben-ises could mean is unclear. It seems that taking Ises as a form of Jesus, invariably fs(s)uffsa in Irish, may be going a bit far, tempting though it might be to interpret ben-Ises as 'wife of Christ', i.e. 'nun'.

DDRoANN(The value of character 76 remains doubtful, the correct transliteration might be D(d)ro(an(n)y D(d)róean(n)> or D(d)rdan(n)y and even the segmentation is not secure: it may be that we should separate off the final


BRESS A Y

134

ANN (D(d)roi I D(d)roe I D(d)roi an(n)). The geminate initial consonant is surprising and perhaps we are to take MEQQDDRo(ANN) as a single unit, not mac Dro(an), but Mac-Dro(an). Drón, genitive Drain, is well attested as an Irish male personal name [CGH 154 d 49, LL 316 b 59; 317 b 13; Ó Riáin 1985:704.79], cf. ogham DROGNO (CIIC 167) > Drón, gen. Dróna [McManus 1991:107], hence the tribe of Ul Dróna who gave their name to the Baronies of Idrone, Co. Carlow [McManus 1991:179 n.34]. For females there is Dron [CGH 157 42] (both names perhaps related to the common noun dron 'solid, firm, substantial, vigorous' ? [DH]). If the name were Drón or gen. Drain, why then, Bressay's A ? The ANN can scarcely be the diminutive -en (the diminutive is Drónán [Ó Riáin 1985:84, 662.48]). Unless we are to posit a rather tortuous spelling, we must look elsewhere.

If Bressay's characters 13-16 are to be transliterated Drol than we may indeed have a form of Irish dnd (later drai and droí) 'druid', 'magus', as Graves suggested, though syntactically we would expect, rather, the genitive druád [DILJ], Brash compared MEQQDDRO(i)ANN to the Ere mocu Druid7, a thief from Coll, named by Adomnán (VC I 41).

In this connection, Rhys drew parallels with ogham MAQI-

DROATA (CIIC 503, Isle of Man). He may have been closer the mark in citing the runic TRUIAN on a rune-stone from Bride, Isle of Man (D is not used in Manx runes) [Kermode 1907:169 No.92 pl.xlii]. Kermode pointed out that the Druian of the inscription may be commemorated in the name of an adjoining quarterland, Glentruan [ibid.]

CRROSCC: NAHHTWDDAddS : DATTRR : aNN | BENiSESMEQQDDRoANN Macalister put forward an elaborate explanation of the Bressay text, replete with suffixes, based on his theory that Pictish was a semi-agglutinative non-Indo-European language [1940:204-5]. This can be safely dismissed. All scholars, Macalister included, have identified MEQQ and DATTR as loanwords, from Old Irish and Old

Norse

respectively,

the identification

of

CROSSC is

also

fairly

secure.

NAHHTVVDDAdd(S) is surely a personal name, similarly (BEN)ISES (if BEN is not 'woman' /'wife') and


BRESSAY

135

DRo(ANN), and possibly the ANN of the right edge. How these words fit together, however, remains to be explained.

Previous authorities have interpreted the two lines as '(the) cross of Naddodd's daughter, Ann. (Of) Benises son of Droan'. Some or all of die proper names are Norse, but the syntax is not wholly Norse. Some or all of the linking common nouns appear to be Irish (crosg, meic, lann), but the morphology and syntax are not wholly Gaelic. As Southesk pointed out, the text conforms to neither Irish nor Norse epigraphic formulae. An example of an appropriate Norse syntax is seen in the runic text on the cross from CilleBharra (Kilbar), Barra (Outer Hebrides): Eftir Porgerdu Steinars ?dóttur es kors sjá reistr. The woman commemorated is / orgerdu Steinars dóttitr, 'Porgerdu daughter of Steinar'. If Bressay commemorates a Norse woman named Ann, then we would expect her to be referred to as ANN NAHHTVVDDAddS DATTRR, 'Ann. Naddodd's daughter'. If NAHHTVVDDAddS and ANN are names, then the syntax NAHHTVVDDAddS DATTRR ANN is more reminiscent of the Irish formula X mac Y / X ingen Y. Following this model, NAHHTVVDDAddS DATTRR ANN would be 'Nechtudad, daughter of An', an interpretation which at least has the advantage of one attested name, and it of the appropriate gender for the person bearing it. If the S is a genitive ending then it may relate to possession of the cross not the daughter, 'the cross of Nechtudad, daughter of An', but why Norse morphology in an otherwise Gaelic or Gaelicized text ? If the text on the other side is ben Ises meicDroan, it may give further information about Nechtudad, 'and wife of Ises son of Droan'. If this is rejected, Benises meqq Droann may be the name of the man who erected the cross (her husband or son, perhaps), or the man who made it. If instead it is Benises Meqq-ddroann* we might have a simple list of the names of the two men involved in the erection of the cross. Taking both sides together we might even have 'The cross of Nectudad's daughter. An, Benises, MacDroan', the honorand named solely by her relationship to a man, commemorated by three men (her sons?). Jesch has discussed the well-established practice of wealthy widows in Sweden and Norway of the eleventh century and later erecting rune-stones to the memory of their dead husbands. To give just one examples, she cites the stone at Alstad with its partly metrical text: 'Jorunn raised this stone


BRESS A Y

136

after..., who had her to wife and she brought the stone from Ringerike, out of Ulvoya | and the stone will honour them both' [1991:71]. Turning the tables at Bressay, the text might mean 'the cross of Nectudad, daughter of An (in memory of her husband) Benises son of Droan', or, if Ann is a woman, 'the cross of Nectudad's daughter An, (daughter of) Benises son of Droan', cf. the rune-stone from Dynna, Norway, in which a Christian mother commemorates her dead daughter [ibid.].

DISCUSSION It is a striking feature of early Irish inscriptions (ogham and non-ogham) that they seldom, if ever, commemorate women. Norse runic inscriptions, in marked contrast, were frequently commissioned by or for women [Jesch 1991:48-74], This is equally true of the mixed Celto-Norse milieu of Norse colonies in the British Isles. The cross from Cille-Bharra, which commemorates a woman in an entirely Norse text on a Celtic form of monument, has already been mentioned. A remarkably high proportion of the twentysix extant runic inscriptions from the Isle of Man commemorate dead women, as mothers, fostermothers, daughters, and wives [Kermode 1907; Jesch 1991:74], Of particular interest is the so-called Mai Lumkin cross, from Kirkmichael which, in addition to its runic text has two ogham inscriptions, one an alphabet, the other, though very worn and not yet satisfactorily interpreted, is probably personal names, [Kermode 1907:100-1; Macalister 1940 No.20; Padel 1972:110-111]. The runic text on this possibly Pictish-inspired cross-slab well illustrates the mixed cultural milieu of contemporary Man [Kermode 1907:195-9]: MAL : LUMKUN : RAISTI: KRUS : ThENA : EFTER : MAL : MURU : FUSTRA : SINE : TOTTR TUFKALS : KONA: i s : AThlSL: ATI x 'Mael-Lomchon raised this cross in memory of his foster mother MaelMuire, the daughter of Dufgal, the wife of Auógisl' BETRA : ES : LAIFA : FUSTRA : KUThAN : ThAN : SON : ILAN X 'It is better to leave a good foster son

than a bad son.' Not only does everyone here, except the honorand's husband, bear a Celtic name, but there are 'several glaring grammatical errors' in the Norse [Kermode 1907:198].


BRESSAY

137

As Page has said, the Manx runic inscriptions 'apply a common Norse formula of commemoration, with a common Celtic variation of wording' [Page quoted by Jesch 1991:73]. A similarly hybrid epigraphic tradition appears to be reflected in the Bressay ogham. The very fact that it commemorates a woman may be taken as a Norse symptom. Bressay affords an insight into ninth- or early tenth-century Shetland society, commemorating a family of Christians created by inter-marriage between Norse and Celtic-speaker. We see an underlying Pictish stratum, but with heavy Irish influence brought in either long ago through the church [Laing 1993b], or more recently by Hiberao-Norse Vikings. This picture contrasts with posteighth century Orkney, where the tradition of Christian Pictish sculpture ceases with the arrival of the Norse [Stevenson 1981]. It may have been in Shetland that the Norse developed an interest in ogham, the only post-ninth century epigraphic oghams survive from Celto-Norse contexts (Killaloe slab, Dublin comb, later Manx oghams).

BIBLIOGRAPHY

OS Record card - HU 54 SW 12

Charlton 1855a ill., 1855b; Graves 1855, 1856:248; 1858 ; Smart 1856:30 pl.94, 95; Wilson 1863:240 ill.; n.a. 1865 pi. 11-12; Goudie 1878:20; Brash 1879:354-8; Southesk 1884:186-98; Ferguson 1887:134-5; Rh$s 1892:296-302, 1893:411; 1898:377-80; ECMS 5-10 fig.4 a-d; Goudie 1904:35-8; Nicholson 1896 L; Macalister 1940:202-6, pl.IIIf; RCAHMS 1946:2-3, No. 1084, fig.462,476-7; Diack 1944 xiii; Jackson 1955:140-2; Padel 1972:62-7; Stevenson 1981:284-5; Close-Brooks & Stevenson 1982:35 ill.; Crawford 1987:169, 170, 171 ill. fig.60; Laing 1993:31.


BRESSAY

138

PREVIOUS READINGS

Graves

CROSC NAHDFDADS DATR ANN | BENRES MECCUDROI ANN 'the cross of Natdodd's daughter here. Benres of the sons of the Druid here*

Southesk

BERNISESSJ: MEOODDRROIANN CRRQESCC : NAHHTFFDDADDS : DATTR.ANNB['the body of Bemis, son of Dru, rests here. The body of Krusa, Natdod's daughter, rests here, wife of....'

Ferguson

CRRObSCC : NAHHTFFDDADDS : DATTRR : aNN BENNRpES : MEQQDDRROI ANN

Rhys

CRROSCC : NAHHTVVDDADDS : DATTR : ANN BENNISES : MEQQDDRROANN 'Cross of Great-Vudda88's daughter, wife of Mac-Drroanf

ECMS

CRROSCC : NAHHTVVDDADDS : DATTR : ANN BE(nn/rr)ISEF : MEQQDDRROANN

Macalister

CRROSCC : NAHHTVVDDA(dd)S : DATTRR : ANN BENNIS : ES : MEQQDDrOANN

Diack

CRROSCC NAHHTVVDDACS DATTRR ANNA BERRISESt. MEQQDDRROANN

Padel

CRROSCC : NAHHTVVDDA(dd)S : DATTRR : ANN(a) BENNISES : MEQQDDRROANN

•Back'

'Front'

Brooks & Stevenson 1982:35] .inscribed cross-slab IClose BBESSAY-Ogham

!

I

&

Iff!

i

'-LJ^Jy

BRESSA Y - Ogham-inscribed cross-slab [n.a. 1865, pl.XII, based on Stuart 1856, pl.XCIIII-XCV]

Cross-slabfromPapil, Shetland ["Trustees of the National Museums of Scotland, Reg. IB 46]

BRESS AY - Ogham inscription on right edge

32 31 30 29 28 27

26 25 24 23 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7

A fc

I

i

6 5 4 3

2 [trustees of the National Museums of Scotland, Reg. IB 119]

a

f

(key to numbering)

BRESSA Y - Ogham inscription on left edge

19 18 17 16 15

i4

12 10 9 8

q

i 1

7 6

2 1 [•^Trustees of the National Museums of Scotland, Reg. IB 119]

(key to numbering)

139

BRODIE

DISCOVERY This sculptured cross-slab, also known as 'Rodney's Stone', was uneanhed in 1781 during excavations for the foundations of a new parish church at Dyke, Moray (formerly Elginshire).

The new building was

constructed behind its pre-Reformation predecessor and the stone evidently came from the graveyard of the old church (NGR NH 990584). It was re-erected in Dyke village the following year in commemoration of Admiral Rodney's victory over the Count de Grasse at the Battle of the Saints, hence it's name. An alternative tradition concerning how die stone got its name is preserved in a note made by J. Grahame Callander in the margin of the copy of ECMS in the library of the Society of Antiquaries of Scotland in Edinburgh. Citing the authority of the Rev. John MacEwan the minister of Dyke, he states that the stone was dug up by a gravedigger 'locally known as Rotteny', and that it was after him, not the Admiral, that the stone was so called [OS Record card]. Black lists no such name in his definitive survey of Scottish surnames, the closest is Rothney an uncommon surname derived from the place in Premnay parish ABD [1946:701]. For all that it is now beyond proof, the admiral seems the more likely eponym than the gravedigger [pace Shepherd & Ralston 1979:31]. George Brydges Rodney (1719-1792) enjoyed an illustrious naval career, during which he captured Martinique, St. Lucia, and Grenada from the French and relieved the Spanish siege of Gibraltar. His victory off Dominica in 1782 was no less than a crushing of the French fleet and earned him Parliament's thanks and a peerage [DNB]. It seems entirely plausible that popular enthusiasm for the first Baron Rodney should cause his name to be attached to the stone discovered at the time of his greatest victory.

Some years before 1842 Rodney's Stone was moved to the grounds of Brodie Castle [NSA Elginshire 221], which is now the property of the National Trust for Scotland. The stone still stands there beneath a wooden shelter at the side of the castle drive (NGR NH 9842 5872) [at the time of writing plans are afoot to move the stone indoors at Brodie Castle, not before time]. As it currently stands, the broken lower edge

CATALOGUE OF OGHAM INSCRIPTIONS OF SCOTLAND

BRODIE

140

rests on a modern base and the slab is held upright by four wrought-iron struts which partially obscure sections of the inscription. At some point in its history, perhaps in the seventeenth or early eighteenth century the cross-slab was inverted and re-used as a tombstone. On the 'front' face two recessed panels were cut into the interlace designs either side of the cross-shaft, and two pairs of initials incised. The back was unharmed except for a broad recessed band cut along the top edge which carryied away the end of both lines of ogham on that side. The bottom of the slab may have been lost at this time if the stone was deliberately cut down to fit a particular space. But if the slab had already fallen or been pulled won the botom might have long since sheered off. The greater weathering of the cross-face, in comparison with the comparatively better preservation of the reverse, is to be expected if the stone lay as a grave-cover with the cross uppermost.

The ogham inscriptions on the margins of both front and back faces, though clearly visible, went unrecognized for almost a century after the discovery of the slab. Stuarts's plates give only a vague indication of horizontal and oblique strokes. Writing in 1886, Southesk records how the first intimation that these marks were ogham arrived 'some months ago' in a letter addressed to Joseph Anderson, Keeper of the National Museum of Antiquities 'by one of his more distant correspondents' [1886:20]. Southesk visited the slab and published the first account of it in the Proceedings of the Society of Antiquaries of Scotland.

LOCALITY Dyke is situated on the fertile coastal plain of the southern shore of the Moray Firth, a little west of Forres. As Ross [1983] and Jones [Jones et al. 1983] emphasize, modern topography gives a misleading impression of Dyke's original setting. It is now well back from the sea, from which it is separated by the thriving forest of Culbin. Old photographs of this 28 sq.km. area of blown sand between the mouths of


BRODIE

141

the rivers Findhorn and Nairn, show that up to the turn of the century it was a 'desert-like wilderness' with dunes up to 30m high [Ross 1983:187]. The extent of the sand has varied over the years and it is difficult to reconstruct its range in the early medieval period. Aerial photography reveals the extent to which the area immediately south of the sands was favoured for settlement in prehistory [Jones et al. 1983; see map following entry on Altyre]. Over the years a number of items, including shards of domestic pottery, beads, and jet rings, dating to the Pictish period have been found amongst the dunes, not to mention the famous bronze snake armlet of the first or second centuries A.D. [Ralston & Inglis 1984:37-8, 43, 45].

The place-name Dyke has been interpreted as simply the English 'dike' [Macauley 1976:260], but Nicolaisen comments that 'one would not expect an English name of such importance to have been recorded in the twelfth century in this pan of the country. Is it possible', he wonders, 'that an English or Scandinavian loan-word in Gaelic was employed when the name was coined ?' [1993:262 n. 18]. In the central Middle Ages Dyke was a thanage [Grant 1993:72; McNeill & Nicholson 1975 map 21], and the Anglo-Saxon influences detectable in the thanage system in Scotland are well-known [Muir 1975:26; Barrow 1973]. The names of the seventy or so thanages appear not to be new coinings, but are often very ancient (taken from specific places rather than districts [Muir 28]). Many have names of Pictish origin. There is no direct evidence for the standing of Dyke in the early medieval period, but Driscoll's emphasis on the thanage's roots in Pictish social institutions [1991:108] raises the possibility that Dyke was a centre of Pictish secular administration. Dyke church sits on a large mound within an oval enclosure, two features usually taken as diagnostic of an early medieval antecendence. The additional evidence of the cross-slab may indicate that here was the principle ecclesiastical site in the putative Pictish shire, maybe a monastery, or perhaps a minster-like foundation.

DESCRIPTION OF OBJECT A large Class II symbol-inscribed cross-slab, sculptured in relief on both sides with oghams incised up three of the vertical edges [see fig.].


BRODIE

142

Stone:

Grey sandstone

Dimensions:

Height: 1.875m (left) 1.905m (right); Width: 0.95m (top) 1.05m (bottom); Thickness: 0.1 m

Condition:

Broken at bottom, portion missing. Minor defacement due to modern re-use. Face badly weathered, carving on reverse well preserved. Ogham severely damaged and legible only in short snatches.

Rodney's Stone is a fairly typical Class II Pictish cross-slab with a relief cross on one face and symbols on the reverse. It is slightly broader and thicker at its base than its top which, paradoxically, gives the visual impression that the sides are parallel, and has the added advantage of enhancing stability. The top of the slab is weathered but substantially intact (the surface has been cut back in the area immediately below the top). The lower edge, however, is broken, and from the design on the back it is clear that a portion is missing. At least a further 100-150mm would be required to accommodate the lower arm of the Z-rod. The stone does not appear to have been deliberately defaced beyond the modifications described above made when it was re-used as a grave-slab \pace Macalister 1940:200]. A deliberate toppling at the time of the Reformation or Covenanters may account for the bottom having been snapped off, but equally the elements and straightforward neglect may be to blame. The face is considerably more weathered than the reverse which is in keeping with its having lain face up when in use as a grave-cover. The upper third is the most severely weathered on both sides, and the lower third, the least affected. The condition of the friable sandstone is not helped by the very damp conditions of its current location.

DESCRIPTION OF CARVING - DECORATION The thin Brodie slab has been carefully dressed and, apart from the incised ogham, all its carvings are in fairly high relief. Both these features point to a date later in the Pictish period [Shepherd & Ralston 1979:31]. The whole of the front face is taken up with a cross with small rounded hollows at its armpits (Romilly Allen's type No. 101a, Cramp 1984 All). This is the most popular cross-shape on Pictish slabs.


BRODIE

143

The whole of the cross is outlined in a thick roll-moulding, which is also found up the left-hand edge. On the right edge is the broad, flat, raised band which carries the ogham inscription (Allen thought there was ogham on the left also, but if there ever was, there is no trace of it now). The cross is decorated with accomplished, complex, and delicate interlacing, unfortunately obliterated in places due to lamination. The pattern consists of a single continuous band arranged in a series of knots. These have been analyzed by Romilly Allen as follows: three circular knots (No.433) arranged vertically in the centre of the cross head, above and below it; in the upper arm and again at the top of the shaft knot No. 714 repeated twice in each of two vertical rows (this knot is also found at Collieburn SUT, Glamis 2 FOR, and Gattonside ROX); the same knot, but without the circular band around it, recurs in die narrow parts of the horizontal arms; these arms have different patterns, to the left Stafford knots interwoven with extra bands placed facing in opposite directions (this knot, No.613, is found also at Nigg ROS, Glamis 2 FOR, and Meigle 4 PER), to the left is interlace composed of Nos. 265 and 431; towards the bottom of the shaft are double Stafford knots with extra bands repeated twice, above and below that, knot No.613; the whole of the shaft is arranged in two columns. It is interesting to note that Brodie should share two of its interlace patterns with Glamis 2, but it is difficult to know what to make of it. There is a certain geometric logic to such patterns and we need not assume borrowing or a common model. In each of the four background quadrants created by the arms of the cross there are interlaced beasts, in pairs and singly. The top right quadrant is almost entirely obliterated, and each of the others has suffered severe weathering, but glimpses of the original designs remain. Towards the bottom of the two lower panels squarish recesses have been cut in the modern period and the initials AC (right) and KD (left) incised, these are inverted relative to the cross.

The back is dominated by a huge pair of sea-monsters which occupy the full width of the slab and almost half its length. Below that are two massive symbols, a Pictish 'beast' and a rdouble-disc-and-Z-rod\ All are highly decorated in relief. The monsters have eyes and large jaws with rows of teeth. At their gills they have spirals and their necks and faces are decorated with triquetra knots and spirals. Their bodies are peppered with tiny bosses representing scales, a feature Stevenson diagnosed as late [1955:121]. Allen


BRODIE

144

calls these fish monsters, Stevenson refers to them as 'hippocampi [ibid.]. Since they lack both limbs and fish tails neither label is entirely accurate: Chambers' dictionary defines hippocampus as a 'fish-tailed horse-like sea-monster'. A monster of the Brodie type occurs opposite a creature with forelimbs and a fish tail on Meigle 1 PER, so we may take it that the two creatures were distinct, if related types (see Aberlemno 2 for perhaps the finest example of the hippocamp type). These here at Brodie could as easily be land serpents as water-snakes. The motif of squat, affronted, limbless creatures with large jaws and spiral tails appears on the cross-slabs from Bressay, Skinnet CAT, Mortlach BNF, on the Maiden Stone ABD, at Kettins FOR and Meigle 1. A single such creature, much debased, appears on Menmuir 1 FOR. Allen considered these creatures to be symbols, a view with which Stevenson concurred. While they may have had a symbolism analogous to the kind of symbolism supported by the stereotyped single horseman, I think it is wrong to classify this motif along with the Pictish symbols proper (the double disc, the 'flower', the serpent and Z-rod, etc.). For one thing hippocamps/sea serpents never occur on Class I stones, and they occur on stones which have no other symbols; but, crucially, they never combine with other canonical symbols to form symbol pairs and thus fail to match my definition of a 'proper' symbol [Forsyth 1996b].

The other two motifs at Brodie, the Pictish beast and the double-disc, are definitely proper symbols, in fact they are two of the three most common symbols. Their appearance here as a pair is typical. This exact combination occurs on two Class I stones from Aberdeenshire, Dyce and Kintore 2 (front), on a Class II slab (Eassie FOR), and in an anomalous triple on the St. Madoes 1 cross-slab PER. In all except the St. Madoes example, the beast is the upper symbol. The combination beast above rodless double-disc appears once on the cross-slab at Shandwick ROS. The decoration of double-discs is usually too simple to facilitate stylistic dating. The Brodie example has each disc filled with, according to Romilly Allen, eight double spirals with C-shaped connections (No. 1108), and small pellets in the spaces between them. There are further pellets in the angle of the Z-rod. The form of the beast (long thin body, small head, straight


BRODIE

145

squared 'beak') and its internal ornamentation (merely decorative knot-work - interlace spiral knots No. 1108 and a Stafford knot at the head - rather than the integral scrolls of the finer earlier examples) point unequivocally to a later date [Stevenson 1955:121]

Between the pair of opposed monsters there is an assortment of curvilinear motifs, Allen's 'curious circular objects or ornaments' [ECMS 133]. In the centre is a large spiral disc decorated with pellets, above is an object which looks a little like a penanular brooch with expanded terminals, to the right is a crescent with pellets, and to the left a circular disc decorated with a triskele, at the bottom between the monsters' tails is another small disc decorated with a triskele. Each item is distinct and individual suggesting that they are intended to represent specific items, rather than merely fill the space with abstract ornament (thought they perform this function also). If the upper item is indeed a brooch, then the crescent may also be a piece of metalwork, like the crescent-shaped silver plaque from Laws, Monifieth FOR [ECMS 280-1]. If the other items also represent metalwork then perhaps we have here two monsters guarding a hoard of treasure.

In his discussion of the development of the 'Boss Style' in Pictish an, Stevenson cites Brodie as an example of the simplification which followed the great elaboration of the later Boss Style exemplified by the great cross-slabs of Nigg, Shandwick, and Rosemarkie. Emphasizing the predominantly northern distribution of later Boss Style, Stevenson dates one of the Rosemarkie stones to the third quarter of the ninth century [1955:121]. We might have deduced from this that he would have placed the Brodie stone in the third or fourth quarters of the ninth century. He may, however, have revised his dating, in a later article he offers a date 'perhaps around 850' for Brodie [1981:287].

DESCRIPTION OF OGHAM INSCRIPTION The entire length of the right margin (front and back) and the left margin (back only) is taken up with a raised band approximately 50mm wide. A stem-line has been incised up the centre of each band with


BRODIE

146

strokes to right and left covering its full width. On the reverse face traces of ogham letters are visible along the entire extant length of the right margin. On the left margin letters are visible at least as far as the curl of the monster's tail [see fig.]. Thereafter the surface is too worn to tell and the final quarter of its length has flaked away entirely. The total extant length of the ogham text on the reverse is therefore approximately 3m, but in its complete original state it may have extended up to 3.8m on this side. On the front of the slab ogham is visible on the right margin only [see fig.]. About one third of the way up the slab the margin has laminated away and it is impossible to discern how much further, if at all, the lettering extended on this side. Ferguson describes this ogham as having a 'definite termination' at the top [1887:143], a view echoed by Macalister, but the carving is no longer clear. The arms of the cross intersect with the side margins but, though parts of the margin are still visible at the ends of the arms, these are too worn to see if they might have been carved. The portion of margin above the arm does not appear to have been inscribed, though insufficient remains to be conclusive. Even at a conservative estimate the total length of ogham inscription on this monument originally may have been as much as 4m. The left-hand edge of the front face is decorated not with a broad band but with a narrower moulding similar to that which outlines the cross. This would seem to prove that the three bands were intended to bear ogham from the outset, and thus that the inscriptions were pan of the original design, rather than added later, not that there could have been any real doubt that the three texts were contemporary with one another and with the rest of the carving.

Ferguson maintained that the front-right ogham should be read upwards, but that the other two went down. Macalister thought the oghams on the reverse read up the right and down the left. Both were mistaken. The direction of the angled-vowels and the slope of b- and h-aicme consonants indicates unequivocally that all three are to be read vertically upwards with the reader's head inclined to the left, exactly as one would expect. The three sections may be part of one continuous text, or represent independent sense units. Either way we do not know in which order they should be read. If it could have been proven that the


BRODIE

147

back-left inscription did not cover the entire length of the slab, this would suggest that the back-right was to be read first. In the earlier oghams it is usually the left-hand edge which is primary, but later Scottish oghams such as Golspie and Scoonie run up the right-hand edge. The fact that on the front it is the righthand margin which is used in preference to the left implies that at Brodie too, the old convention was breaking down (cf. Formaston for a more problematic example of a right-side ogham). If the ogham on the front is complete as it stands then either it was the final section, or it is separate from the main text. The question of the order of the three sections is, however, largely academic since the carving is so poorly preserved that only one short sequence can be interpreted with any certainty.

The nature of the ogham script is such that a letter damaged, even only partially, is virtually meaningless. Unlike the roman alphabet, it is impossible to reconstruct ogham letters from isolated strokes or pans of strokes, hence the great difficulty of extracting a usable reading from the weathered carving at Brodie. Detailed measurement and analysis might permit a partial reconstruction, but such work is impossible with the stone in its current position, out in the open, beside the road, in a dark wooded corner.

Front left This side is too damaged for certainty, but there does not appear to have been any ogham on it. I follow Padel in labelling the remaining three sides A B C as follows:

A

Front right

The inscription has already started at the fracture and it is possible that letters have been lost from the beginning. The slope of 8 confirms that the direction of reading is from the bottom up, as follows [see fig. for key to numbering]: 1

Three short strokes to the right of, and perpendicular to, the stem - V. The surface to the left of the stem is lost, so Padel's suggestion of AL is not impossible, but, given the spacing is unlikely.


BRODIE

148

This appears to be a complete letter rather than the end of a longer one, but the fracture is such that certainty is impossible. 2

Two long vowel-strokes perpendicularly across the stem - O.

3

Five short strokes to the right of, and perpendicular to, the stem - N.

4

There follows an illegible gap of 70 or 80mm, sufficient for about five strokes. Roughly a third of the way along a single long stroke is discernible, perpendicularly across the stem, suggesting either a five stroke vowel, or, since 5 is a vowel, and a double vowel is unlikely, a two or threestroke vowel (O/U) followed by a two or three stroke consonant (L/D/G, V/T).

5

Four long vowel-strokes perpendicularly across the stem - E. It is just possible that a fifth stroke has been lost in the gap - 1 . There are slight traces of carving to the right of the stem beyond this, which might comprise the three strokes of a V, but this would be rather cramped and the rest of the inscription is well-spaced.

6

Four short strokes to the left of the stem, sloping very slightly backwards - C. Macalister though he saw at this point a uniqueforfid similar to the pair of modified Ds on Bressay. No evidence for such an interpretation remains, and it is not in keeping with the general simplicity of the letter forms used elsewhere on this stone.

7

As above - C.

8

Two long strokes across the stem, angled upwards at their mid-point - O. The second of these is approximately level with the top of the recessed panel containing the modern letters AC.

9

After a short gap there is one oblique stroke to the right of the stem, sloping forwards, and then one to the left sloping backwards. These could be the remains of further strokes of the same form as 8. The gap would be sufficient to take one more to make a total of five - 1 . Several other interpretations, however, are equally possible (B+H, a second O, etc.). After this the surface has laminated away. Though the inscription probably continued at least some way beyond this point it is extremely difficult to gauge how far it reached originally. Towards the top, above the cross-


BRODIE

149

arm, sections of the right-hand border without ogham can be traced, suggesting the text did not occupy the entire length of the slab, but so little survives that the evidence is inconclusive. Which gives the following very tentative reading: -](v/v + )ON(a + /a + m + /a + b + /a + h + )(e/i)( /v)CC(i/obh/oo)[~

B

Back left

As with the other two sections, the inscription has already begun before the fracture and a section of unknown length has been lost from the beginning. From the break onwards, the ogham reads as follows [see fig. for key to numbering]: 1

Five long strokes obliquely across die stem - R. The first is feint, which, as Padel points out, accounts for Macalister's Z (he was reading down).

2

Two long strokes across the stem, sloping very slightly forward. The first is close to perpendicular (certainly less sloped than the surrounding strokes) and the second is a little faint. They seem too close to be read separately as AM. Padel preferred G, Macalister read O, really there is little to choose between them except perhaps that elsewhere in the inscription vowels tend to slope a little backwards if there is any possibility of doubt.

3

Five strokes to the right of and perpendicular to the stem. The surface of the stone immediately to the left of the stem is lost opposite the first four strokes, the fifth continues across. Since the spacing is so regular, these are unlikely to have been subdivided and therefore we ought to follow Macalister in reading I rather than SA

4

Five strokes to the right of the stem. Padel reported the surface as lost opposite die last two strokes, it has now gone opposite them all. Macalister read R suggesting there was some doubt even then - N or R.

5

Three long strokes obliquely across the stem sloping forward. Thefirststroke slopes a little less than the other two and is a little further spaced from them, whether this warrants the reading AG or MG rather than Gw is unclear. Macalister took the first stroke as a vowel (A) but its slope is


BRODIE

150

perhaps too great for that. 6

There follows a very worn section then three or four h-strokes. Since the b-surface is more or less intact we may take it that the patch of wear contained another h-stroke or two. Macalister read two groups of three, i.e. TT, Padel thought C Very likely'.

7

After a short gap which appears clear, the h-surface has laminated off for a distance sufficient for perhaps six or seven strokes. The b-surface is intact for a little over half of this, implying that the first few strokes must have been of the h-aicme.

8

Just below the metal bracket there is a group of perhaps five h-strokes - Q. Padel thought these Very dubious1.

9

Two long strokes perpendicularly across the stem - O.

10

Two (Padel) or three strokes to the left of, and perpendicular to the stem - D or T.

11

Three strokes to the left of the stem sloping slightly backwards - T. The difference in angle may be to differentiate two successive letters of the same aicme.

12

Two angled vowel-strokes pointing in the direction of reading (-^->-) - O.

13

The next section is extremely worn, any strokes are likely to have been to the right of the stem. Padel read three and then a fourth at a slightly greater distance, all perpendicular to the stem - S.

The inscription continues to a point a little higher than the base of the monster's tail, but the rest of the carving is so worn that very little can be said of it. It is over twenty years since Padel made his observations and it may well be that more of the carving was intact when he saw it. 14

Padel read a further two sloping strokes but his account of them is confused: he described them as to the left, saying the h-side (sic) was illegible, and suggesting O or L (sic). His sketch shows them to the right. It seems to me that any strokes are indeed on the left, that it is the b-side which is illegible, and that the letters are to be taken as a T.

15

Two strokes to the right of the stem. These might be the remains of angled vowel-strokes (O). Padel has the first of these perpendicular to the right of the stem but continue across it sloping


BRODIE

151

forwards, which is a peculiar arrangement. As he says, the two are surely too close to be MB. Any forward sloping projection of this first stroke is no longer visible. 16

Four strokes to the right of the stem sloping very slightly forward - S. Beyond this point the surface is entirely destroyed, the text may or may not have continued further.

And reconstruction would be so tentative that it is scarcely wonh setting our die alternatives, but the following is possible:

C

-]R(g/o)I(n/r)(ag/mg/gw)(tt/.c)(h+)( /.)qO(t/d)TOst(o/mb/g)s[-

Back right

This is the clearest of the three Brodie inscriptions, but even it is severely damaged in its middle and upper sections. The surface is more or less intact around groups 1- 4 and these are fairly clear, the top layer has, however been lost in the section that follows and groups 5 to 9 are less clear. Only traces of the bottom of the scores of 10 and 11 have survived. Groups 12 - 14 are more strongly present but thereafter very little can be discerned for certain. It appears that originally die text ran the entire length of the slab, but the top has sheered off. It is by no means certain that the text would have begun at the original ground level of die slab (at least a further 100-150mm below the current terminus), but letters may well have been lost from the beginning of die text. Starring at the ground, immediately after the fracture [see fig. for key to numbering]: 1

Four long strokes perpendicularly across the stem - E. The break in the stone crosses the ogham band at an oblique angle, but just sufficient of die surface to die left of the stem has survived between the fractured edge of die stone and the first stroke, to demonstrate that there was not originally a fifth stroke to this letter. The third stroke appears to be angled at the stem ( f | > |) to point upwards in the direction of reading. This mixing of angled and straight strokes within the one letter is unprecedented, but the spacing of the strokes does not warrant die subdivision of the group. The device may be intended to make it clear diat the letter is a vowel, not a member of the m-aicme. Though die fracture of die stone has carried off the portion of the first stroke


BRODIE

152

to the right of the stem, the reading is not in doubt. 2

Two strokes to the left of the stem, sloping backwards sharply - D.

3

As above - D. The gradient of 3 is slightly less steep than 2, but this does not affect the reading. The two pairs are widely spaced to prevent conflation as C.

4

A single long vowel-stroke perpendicularly across the stem - A.

5

Five long oblique strokes across the stem, sloping forwards - R.

6

As above - R. Letters 5 and 6 are carefully spaced.

7

Five strokes to the right of, and roughly perpendicular to, the stem - N. Padel thought the. first three slightly more sloped than the last two, but the spacing is enough to prove the reading which is not in doubt. There are undulations in the badly abraded h-surface opposite these strokes but these are much less substantial than the proper letter-strokes of the preceding groups. If the baicme strokes of 7 had continued onto the left of the stem the final two would have intersected with the first of the backward sloping strokes of 5, thus we can be sure that 7 is a member of the b- not the m-aicme.

8

Two long strokes across the stem, following on almost immediately after the final stroke of 7. There can be no doubt that these two are not perpendicular to the stem, but slope backwards, against the direction of reading. This must be to differentiate them clearly from the adjacent, forward-sloping, m-aicme letters. There is no doubt that this letter is O.

9

Five strokes to the right of, and perpendicular to, the stem - N

10

The next two groups are extremely worn. Two strokes to the right of, and perpendicular to, the stem, then, after a gap, a third stroke. The gap is sufficient for two strokes, which would bring the total to five - N. The space between the last stroke of 9 and first stroke of /0, i.e. between letters, is only marginally greater than the average space between strokes within a letter, but the division is in no real doubt. Padel complained that the sloping of strokes in these two groups was not consistent (the first are almost perpendicular, the last slope forwards. The reconstruction of


BRODIE

153

the text strongly suggests, however, that this should be taken as a second N. 11

The following section must have contained at least two letters. After a gap there appears to be two long strokes across the stem, sloping very slightly forward. If anything has been lost before these two it could be no more than a stroke or two, but the gap is not that much larger than normal inter-letter spacing. The forward slope on these strokes is so slight that they may be vowel-strokes, in which case O, or possibly U or E, but given the backward slope on the vowel 8 it may be that these strokes are pan of an m-aicme letter. Two strokes would be G, any more would be Gw or Sl but these letters are extremely rare in epigraphic oghams and we should be slow to see them here. Following this group there are four strokes to the right of the stem, again sloping very slightly forward. There is a generous gap between the last of the long strokes and die first of the short, but it is not so long that lost strokes must be posited. The four are far from certain, ponions could have been lost to die left of die stem (there are hints of such strokes but most of die surface is lost at this point). Further entire strokes immediately following it could have been lost, so the reading S or N is highly tentative. It is further possible that the four should be divided into two pairs - LL. Padel made no attempt at this section, and Macalister's claim to see a 'clear' I is dubious.

12

Following a short gap, some clarity returns. Three strokes to the left of die stem sloping backwards. Pan of die wrought iron bracket by which die stone is held upright almost entirely obscures the surface of the stone to the right of die stem, therefore it is impossible to see if the strokes continued across. The backward slope would tend to suggest T rather than U, but tliis is not certain because of the backwards slope of 8.

13

Three strokes to the left of die stem sloping backwards. The third stroke comes after the bracket and clearly does not continue to die right of the stem, confirming the reading T. The propensity for doubling ogham consonant thus supports, though does not compel, the reading of 12 as T also.

14

Five long strokes perpendicularly across die stem - 1 . The fifth is very faint and the second and


BRODIE

154

third have lost whatever was to the right of the stem. 15

The next section is severely pocked and abraded. There are hints of diagonal strokes, perhaps even forming the RR-foifid (jg$Q, but this is highly doubtful. There follows an extended section (more than 300mm) of highly degraded text in which nothing but tantalizing disjointed pans of possible letters survives. Nothing is legible.

16

Level with the top of the curl of the monster's tail there appear to be five long strokes sloping forward across the stem - R.

17

A further group of five sloping strokes - R. Letters 16 and 77 are carefully spaced.

18

The next section is badly damaged. There appear to be one or two b-strokes, perhaps B or L, but it cannot be ruled out that they are part of a larger letter, or they may even be vowels.

19

The following strokes appear to be vowel-strokes, perhaps four or five (E/I). They seem perpendicular and thus are less likely to be m-strokes, but neither this, nor the possibility of subdivision can be discounted.

20

The next section is all but gone. There does not appear to be carving to the left of the stem, so b-strokes are more plausible. Since it is not clear when 19 ends and 21 begins, 20 may have contained two, three, four, or even five strokes.

21

Level with the neck of die monster, three strokes are visible to die left of, and perpendicular to, the stem. They may have continued across die stem.

22

After an apparent gap just larger than the usual gap between strokes there are a further four or possibly five strokes similar to the preceding group. These look more like vowel-strokes - E? I?

23

It is clear that there were further letters beyond this, but they are too worn to reconstruct. There is space for another five or six letters before the top of the slab. A recessed band several centimetres wide has been removed from the top edge of the stone in recent centuries, and it is impossible to see how and where the text originally ended.

The first ten letters may be reconstructed with reasonable certainty:


EDDARRNONN[-

BRODIE

155

or -]EDDARRNONN[-

thereafter the reading is so uncertain that there is little merit in offering a reconstruction. The best I can manage is:

-(o/a+/g+)(ll/s+/si+)(t/u)]TI[(m+/n)[-

beyond which the carving is too badly damaged to attempt even something as tentative as this.

FORM OF SCRIPT Despite the extreme wear the Brodie inscription has suffered, it is still possible to analyze the form of script used. All three sections are carved in the same hand, a neatly consistent Type lib form of the script. In a couple of places the distal tips of strokes have been lost where the outer edge of the ogham band has flaked off, but even at their full extent the strokes are of no more than medium length. Straight vowelstrokes are long, occupying the entire width of the ogham band. There are occasional angled-vowels, and one case of what appears to be a vowel comprising a mix of straight and angled strokes (CI-

||>f).

In the more legible sections component strokes are strictly parallel and evenly-spaced, and individual letters are, in general, carefully spaced, especially when two letters from the same aicme are side-by-side. In most cases, b- and h-aicme strokes slope in the anticipated direction, the only exceptions are dubious through wear. Weathering cannot always be invoked to account for apparent inconsistencies of sloping between straight vowel and m-aicme letters. In a few instances the former appear to slope backwards, against the direction of reading, perhaps to differentiate them from forward-sloping m-consonants. Others are slightly further forward than upright and it is not always easy to distinguish them from m-strokes. Most of these problems occur, however, in sections which are damaged to some extent.

This is a substantial ogham, neatly andformallycarved, as befits a major public monument. In general terms the closest parallel to the script is found at St. Ninian's Isle, though the latter lacks angled vowels. Scoonie is also similar, though its letters are not as consistently sloped and it too lacks angled vowels. Birsay is closest, but it is such an informal and cursive text that its over-all appearance is quite different.


BRODIE

156

Golspie is broadly similar, and equally stately, but it has supplementary letters which Brodie lacks. The inscription is so damaged that we should be wary of arguing from the absence of letter forms, but it is noteworthy that, despite the length of the text, Brodie appears to lack both the x-forftd and the hammerhead A, the two most common forfeda in Scotland.

INTERPRETATION OF TEXT The appropriate transliteration of V is discussed in the Introduction where it is suggested that in inscriptions such as this, which clearly post-date the seventh century, V should be transliterated HI in Gaelic contexts and /w/ in Pictish. Angled vowels are also discussed in the Lunnasting entry, where it is suggested that they may have had orthographic rather than merely palaeographic significance. Their phonetic value, however, remains unclear. Perhaps the most important feature of the Brodie ogham inscription is its sheer length. Even if the three lines of text are to be taken separately as independent sense units, they comprise a long and possibly complex message. Unfortunately we have no models for epigraphic ogham texts which go much beyond the listing of names, patronymics, and gentilic affiliations. Despite the nearly total illegibility of large stretches of text, there does not appear to be a 'MAQQ' at any point, implying that the three sections of the Brodie inscription do not comprise names plus patronymics. We may instead have a list of half a dozen or more simple personal names, no more than loosely connected, but it is perhaps more plausible to imagine text incorporating more than proper nouns. The prospect that the Brodie inscription might include common nouns, prepositions, or even verbs, makes its current impoverished state of preservation all the more frustrating. Only one section can be made out with any certainty, this is the EDDARRNONN at the beginning of C (back right). This is the male personal name Ethernanlldarnon which is discussed, along with the saint who bore it, in the entry on Scoonie. Though the name could refer to an ordinary secular figure, the Brodie text may be evidence of the cult of St. Etheman in Moray. Though the name is apparently Brittonic, this spelling could be either Irish or Pictish. The other sequences are too short or doubtful to throw further light on the language of the text.


BRODIE

157

The section immediately following EDDARRNONN may be OLL, which, as it stands, could be either Brittonic or Gaelic. In Welsh oil 'all', 'wholly', cf. Irish idle,followsthe word which it modifies [Evans 1964:98 §107]. Though unattested in inscriptions, it is possible to formulate plausibly appropriate phrases containing the two words in this order, especially since text may have been lost from the beginning as well as the end of the line. Irish oil, 'great', is perhaps more easily integrated into a meaningful text, appearing in afigurativesense, as it sometimes does, following a personal name, eg. Aed ollfii andud n-áne 'Aed great at kindling of radiance' [Stokes & Strachan 1903,295.2, cited in DH]. The title ollamh (the highest grade of poet) seems to be ruled out by the extant remains of letters in thefollowingsection, but oil occurs widely as thefirstelement in compounds with substantives meaning 'great', 'vast'. If oil and what follows is a compound, this would explain thefollowingTT, since a double letter would not normally be expected in word-initial position. If the double T were intended to indicate lack of lenition then /ti(:)r/ could be tlr 'land', 'region'. Given the lack of certainty of the lettering at this point it is perhaps going tooforto suggest tiruairse 'relics', 'fragments', 'saintly relics', cf. túairse [DIL], though if the cross-slab records some major benefaction, or saintly dedication, 'great holy relics' might well be mentioned in the text. An alternative is that olltir is in some way connected with Irish ailithir (pile 'other' + tir 'land') 'pilgrim', 'stranger', which is sometimes spelled oilithir [DIL], though the loss of the middle vowel is a problem. Brittonic cognates of both these words exist, cf. Welsh ail + tir, but that leaves the problem of OLL for AIL. One 'Colman ailithir1 is commemorated in an ogham inscription from Maumanorig, Corcu Dhuibhne (CIIC 193).

DISCUSSION One of the most striking features of the Brodie ogham is its overall length. Taking the three sections together, the text must have been originally up to seventy or eighty characters long. This is considerably longer than the next two longest ogham inscriptions from Scotland (Bressay, under fifty, and Lunnasting, under forty), and certainly far longer than the longest Scottish texts in non-ogham scripts (Newton c.44,


BRODIE

158

Tarbat c.55?). This represents a substantial commitment to ogham literacy. That the script was used so prominently on a major public monuments such as this is one of the most striking examples of the high status accorded the script in post-seventh century Scotland. The precise nature of the Dyke site in the ninth century, whether small monastery, minster church, burial ground, or whatever, is not known. It is notable that the two ogham-inscribed slabs from Moray both come from sites with no other explicitly Christian early medieval sculpture (Dyke and Roseisle), and that conversely, two sites with large collections of early sculpture, Burghead and Kinneddar, both lack ogham inscriptions. This dichotomy, which is mirrored elsewhere in Scotland, may reflect a contrast in the nature of the sites involved. Crudely put, on a continuum running from small domestic habitation sites to major monasteries, the more 'ecclesiastical1 a site, the less likely one is to find ogham there. Other Moray cross-slabs from sites with no other early sculpture, Elgin and Sueno's stone, Forres, are notable for the secular themes, hunting and warfare, of their decoration. This may throw further light on possibly secular patronage at Brodie.

BIBLIOGRAPHY

OS record card - NH 5 NE 3

(Rev. M Aitken) NSA 221, 226 (stone only, no mention of ogham); Stuart 1856:9,11 pl.xxii, xxiii (no mention of ogham); Southesk 1886:14-22; Ferguson 1887:142-5; Rhys 1892:287-8,1898:367; ECMS1325; Nicholson H; Macalister 1940:199-201, pl.IIIe; Padel 1972:67-73; Shepherd & Ralston 1979:31, ill.22; Jackson 1984 ill. opp. p. 117; Foster 1996 fig.48 (Tom Gray).

PREVIOUS READINGS Ferguson

Back R (C+B) EDDARRNONNmeQIiJQeN(r/gg)RuN(urb/eel/stf etc.)URH Front

(A)

(e/st)(g/o)NN(m/a)(c)QQDIM(g/t)A(c/t)PLL(u/o)MB(a/o)(n/s)[-


Bhfs

Front R (A)

(b/n)ONaroTaTC(o/e)

Back L (B)

(r/f)(am/o)SANAGhhdOhotTOIa(l/g)S

BackR (C)

EDDARRNONn ehhTUMo[-

ECMS

BackR (C)

EDDARRNON[-]Q[-

Macalister

Front R (A)

VONSEOdTO

Back L (B)

MADE OEVV IPENN GARIOZ (reading down)

Back R (C)

EDDARRNON ILCIUDOVORR VITEOR

Front R A

(v/al)ON[-]ECCO[-

Back L B

R(g/am)iN£&chqOOTOS(o/l)mbs[-

Back R C

EDDARRNOnn[»](t/u)(t/u)IHggw[-

Padel

BRODIE

159

BRODIE - Ogham-inscribed cross-slab (reverse face) [Foster 1996 fig.48 (Tom Gray)]

'Front*

'Back'

BRODIE - Ogham-inscribed cross-slab showing position of inscriptions [ECMS fig. 1361

3 2 1

BRODIE - Front right - schematic representation of ogham inscription (key to numbering)

16 15 14 13 12 11 10 9 8

mwm

6 5 4 3 2 1

BRODIE - Back left - Detail of ogham (key to numbering) [After Foster 1996 fig.48]

17 16

14

BRODIE - Back right - Detail of ogham (key to numbering) [After Foster 1996 fig.48]

160

B UCKQUOY

[In modified form the following has been accepted for publication in PSAS 125 (forthcoming November 1996) under the title The ogham-inscribed spindle whorl from Buckquoy: evidence for the Irish language in pre-Viking Orkney ?\]

DISCOVERY The spindle-whorl was discovered by Dr Anna Ritchie in 1970 during her rescue excavation at the Point of Buckquoy, Birsay, on Mainland Orkney (NGR HY 243 282). It was described and illustrated in her excavation report [Ritchie 1977:181-2,197 fig.8 no.84, 199 pi. 13a, item 84 in the catalogue offinds]with a detailed discussion of the inscription by Kenneth Jackson [Jackson 1977] and of the geology of the stone by G. H. Collins [Collins 1977]. The inscription has been further discussed in two unpublished MLitt theses [Padel 1972:73-5; Holder 1990:66-70] and by Dr Ritchie herself [1983:62,65]. Along with all the other finds from the Buckquoy excavation, the ogham inscribed whorl is now in the Tankerness House Museum in Kirkwall, Orkney (ref.no. 1976.56).

THE SITE The threatened site, known locally as Sinclair's Brae, was a long, low mound (20m long, maximum height 0.5m), truncated at one end by coastal erosion [Ritchie 1977:174]. Ritchie estimated that, by 1970, at least half the original site had been lost over the encroaching cliff [ibid. 174]. Ten weeks of digging revealed a series of farmsteads, dating from the seventh century to the tenth, built and re-built one on top of the other [175]. The Buckquoy excavation proved seminal, for it provided the first identified example of a distinctively Pictish house-type, of cellular form [182-3; see fig.], which subsequent discoveries have shown to have been widespread [Morris 1991:72], In the absence of other clear indicators, the remains were interpreted on the basis of the contrasting and distinctive house-types. Ritchie identified two major phases (MI) of Pictish occupation, followed, after a period of disuse, by three phases (III-V) of Viking


BUCKQUOY

161

settlement. The Viking occupation appears to have been short-lived because the site was ruinous by the time it was used for an isolated burial, datable on coin evidence to the third quarter of the tenth century [1977:190-1].

The inscribed spindle-whorl was discovered immediately outside the south-west entrance to the main hall of house 4 [ibid. 181]. This structure, which has been described as having a 'distinctly anthropomorphic air' [Ritchie 1983:56], is perhaps the most sophisticated example yet uncovered of the Pictish 'figure-ofeight' house type. It differs from the older cellular type in having its cells arranged in a linear plan, not set in a circle round the central area of the house. Not enough is known about Pictish domestic structures to date this type closely [Ritchie 1977:182-3]. A similar, though simpler structure at Coileagan an Udal, Nonh Uist, was dated by its excavator between the seventh and ninth centuries [I. Crawford 1974:9]. Figure-of-eight-shaped houses made of wattle have been excavated more recently at Deer Park Farms, Glenarm, Co. Antrim [Lynn 1989], which calls into question the assumption that this type of dwelling is necessarily diagnostic of Pictish inhabitants. Most of the artefacts found in the pre-Norse layers at Buckquoy were of simple types common in various parts of north-west Europe in this period, though the painted pebble was recognized as being diagnostically Pictish [Ritchie 1972]. The ogham was also listed as a characteristically 'Pictish' item, without reference to the emphatically Irish background to the script [McManus 1991]. The possible Irish connotations of the Buckquoy inscription are discussed more fully below.

The settlement at Buckquoy was not itself of high status [Ritchie 1983:54] but its position at the nonh side of the strategically imponant Birsay Bay, very close to the tidal islet of the Brough, give it an importance beyond its size. The name derives from the ON bygg-kvt 'bere quoy', i.e. barley enclosure, and is a reminder that, traditionally, the soils of Birsay were considered the most fertile of mainland Orkney [Ritchie 1977:174]. The animal bones recovered in the excavation indicate a mixed farm with an emphasis on the pastoral, especially cattle [ibid. 191]. Ritchie has put forward the persuasive idea that Buckquoy


BUCKQUO Y

162

was 'the home farm for the inhabitants of the Brough' [1985:198]. The Bay of Birsay is one of only three sheltered bays on the west coast of the mainland of Orkney and is ideal for fishing [Ritchie 1977:174]. More importantly, perhaps, it is, as Ritchie has pointed out, particularly well-suited for boats setting out to the Hebrides or Ireland [1983:47].

Dating The discovery of an ogham inscription in the course of an archaeological excavation conducted to modern scientific standards raised the exciting possibility of the first absolute dating of an ogham inscription. Sadly, however, very little dating evidence had survived for any, bar the final, stage of activity at Buckquoy, and in the official report it was the ogham which was used to date the context, not vice-versa [Ritchie 1977:192].

The inscribed whorl was recovered from the second of the Pictish phases of

occupation [181]. Unfortunately there was nothing to permit the close dating of either phase. Only twenty-seven artefacts and some pottery sherds were found in the Pictish levels. These artefacts were, in the main, simple domestic items such as bone pins, a comb, and a spoon, none of which are closely datable [179].

The lack, until recently, of properly excavated ogham inscriptions has made it impossible to construct an accurately datable typology. On the basis of certain assumptions about the relationship between the forms of the script used in Ireland and in Scotland, and on the basis of the an historical dating of the slabs on which some inscriptions appear, it has been customary to assign most so-called 'Pictish' oghams to the eighth or ninth centuries, with simpler ones perhaps slightly earlier [Jackson 1955:139]. Since the script of Buckquoy is not typologically the most simple, Jackson's dating of the inscription to the eighth century was in keeping with general opinion at the time. Ritchie used his date to assign phase II to the eighthcentury and phase I to the seventh. While the other finds are not incompatible with this dating, they could be older. In the interim since Jackson gave his opinion, however, radio-carbon analysis of the contexts


BUCKQUOY

163

of ogham stones excavated at Pool, Sanday and the Brough of Birsay has stretched the chronology of Orkney ogham back to the sixth- and seventh-centuries, and thereby thrown open the question of dating all the Scottish oghams, including the whorl [Pool - Hunter 1990:185; Birsay - Prof. Christopher Morris pers. com.]. It would be rash, however, to place too much emphasis on these new radiocarbon dates in advance of final publication of the excavations concerned.

A terminus ante quern for the abandonment of the farmstead, and therefore the deposition of die ogham, is provided by the conventional date of the beginning of Norse settlement in Orkney, c. 800. It has been assumed that sites in and around Birsay would have been settled early on, since Birsay was an important centre in the Pictish period [Ritchie 1983:47]. If one accepts this, and allows 'a brief interval, perhaps half a century* [Ritchie 1985:194] for the ruins to develop, die conclusion is that the buildings fell out of use in the mid-eighth century. In recent years, however, authors have come to admit that there is 'as yet no firm local, historical or archaeological evidence* for the date of the landnám, and that c.800 'is not a comfortable assumption' [Bigelow 1992:10; see also Morris 1985:210-3], While raiding bases might indeed have been established at the beginning of the ninth century, permanent settlement may not have got underway for at least another generation [B. Crawford 1987:40-2], The ninth century Latin Life of Fintan paints a picture of an Orkney still Pictish in the 840s [Thomson 1986, Lówe 1986]. There is no independent dating for the Viking phases at Buckquoy, instead these are retrospectively dated from die burial (phase VI) in the ruins of the last farm, which can be securely dated to the mid- to late tenth-century [Ritchie 1977:190-1]. Thus, while Ritchie's relative chronology for the site is inherently plausible, both upper and lower limits for the Pictish occupation are fluid. Jackson's dating of the ogham may well be correct, but cannot be relied on without independent corroboration. The form of the script used is compatible with any date from the seventh to the eleventh. We fall back, then, on a chronology based on the interpretation of the history of die site. This can provide a date no more refined than 'seventh-, eighth, or early nindi-century'.


BUCKQUOY

164

DESCRIPTION OF OBJECT A chalk spindle-whorl incised on one face with ogham arranged round a circular stem [see fig.]. Material:

'A fine-grained cream coloured sandy limestone, with quartz grains up to 0.5mm in diameter' [Collins 1977:222]

Dimensions:

36mm diameter, 10mm thick

Condition:

Intact and in good condition, inscription well-preserved.

A spindle-whorl is a heavy ring, in this case a perforated stone disc, attached to the pin of the spindle to give it the necessary weight and momentum for spinning thread by hand. A total of six spindle-whorls were recoveredfromBuckquoy, all of very similar size and shape, though madefroma variety of materials [Ritchie 1977 Nos. 82-86]. None of the others are marked or decorated in any way. The two examples from phase II (Nos. 83 and 84) were both made of chalk 'presumably chosen for its creamy colour and tractable nature' [ibid. 181]. Geological analysis of the three limestone whorls showed them to be 'made from closely similar rock type, varying only in detail' [Collins 1977:222]. Collins concluded that they 'could well have been madefromchalk pebbles, obtained from local glacial deposits' [ibid. 223]. Several other limestone whorls are known from Orkney, including one from the Broch of Lingrow [NMS GE11; MacGregor 1974:92] which is made from very similar chalk to that of No. 84 [Ritchie 1977:181]. The likelihood that the whorl was made from locally obtained stone is important because it implies the ogham was carved in Orkney rather than that the whorl was, say, imported from Ireland ready-carved. Ritchie was surprised that so few spindle-whorls were recovered from Buckquoy and concluded that any sheep at the site were kept more for their meat and hides, a view supported by the evidence of sheep bones recovered. Nonetheless, the presence of the ogham inscription on No.84 would seem, she feels, 'to imply that the whorl was special to its owner' [ibid. 182].

CARVING TECHNIQUE The ogham inscription was lightly, though clearly, incised with a fine sharp blade. Such a text would be easily and quickly produced on as workable a medium as chalk and can have taken no more than a few


BUCKQUOY

165

moments to carve. The inscription is complete and well preserved. Jackson remarked that t(a)lmost every stroke is quite clearly made out when the whorl is held to a strong light at an angle and examined with a magnifying glass* [Jackson 1977:221], The carving is most readily seen on the photograph of the whorl taken under magnification by the Institute of Geological Sciences [see fig. (first published in Ritchie's excavation report, 1977 pi. 13); for further written descriptions see Jackson 1977 and Padel 1972].

A close examination of the published photograph reveals a zig-zag line traced very lightly across the surface between the perforation and the stem. It underlies the first two strokes of 12 and re-crosses under the stem just before letter 3 [see fig. for key to numbering]. This feature appears not to have been noted before, but evidently is prior to the ogham. The strokes are less substantial than those of the ogham and there is no confusion between the two. The sweep of the stem-line away from the perforation may have been to avoid this zig-zag, but in any case the stem is fairly haphazardly placed.

The stem-line, which is c.120 mm long, was incised only once and not re-cut on top of the subsequent letter-strokes. This is most clearly seen in the case of letters 5 and 11 (A and M). The individual letter strokes are c. 3-4mm in length, with 5mm for the hammer-head A and 12mm for the M, and were carved with one knife-score each. There appears to have been no re-cutting or augmentation of strokes (though see discussion of 7 and 72).

The spacing of letter 9 (N) requires some explanation. The second and third strokes are full-length, parallel, and spaced and sloped as one might expect. The first, however, is at an eccentric angle, meeting the second in a point at the stem. Even more strangely the fourth and fifth are squeezed into the inadequate space before 10, cramped to such an extent that the final stroke intersects with the following stroke well below the stem. This indicates that 10, or at least its first stroke, must have been carved before 9, or at least its last two strokes. This is puzzling since it would imply that the letters were not carved in the sequence in which they were to be read. If the carver was copying a circular model he or she might start


g UCKQUO Y

166

at any point, especially if he or she was not fluent in ogham, but surely it would have been apparent by the last stroke of 8 that space was running out and all the letters of 9 would need to be tightly spaced. That this was not done implies an alternative explanation.

The parallel second and third strokes of 9 appear firmer then the other three. They are accurately sloped and, if one ignores the first, fbunh, and fifth strokes, are nicely spaced relative to the letters on either flank. Could it be that, having completed the text, the carver realized he or she had carved two strokes where five were needed and squeezed in the first, fourth, and fifth strokes as best they could ? Perhaps closer examination of the depth of individual strokes, and therefore the pressure with which they were cut, might be instructive. Miscounting strokes in this way is the kind of slip that even an experienced oghamist might make and need not imply that the carver was following an exemplar he or she did not understand.

Several features combine to give the Buckquoy ogham a cursive appearance. These are, principally; the stem is not equidistant from the perforation all way round, the parallelism of component letter-strokes is maintained only erratically, there is disparity in the length of strokes both within and between groups, the letters are unevenly spaced (note especially the cramping of 9 and 1211). These are the kind of features often attributed to ignorant copying by an illiterate carver, yet the perceived shortcomings of the Buckquoy ogham are not blunders due to miscomprehension. Since the legibility of ogham depends entirely on layout and spacing, one can't help feeling that someone not fully comfortable with the script would have taken more care to lay it out properly. While the use of examplars is highly likely in grand monumental ogham epigraphy, such as at Branasbutt, Whiteness, or Dupplin, the texts of which are carved with special, labour-intensive carving techniques, it seems scarcely appropriate for a casual and cursive text like Buckquoy.

No special equipment or carving ability would have been required to scratch such an

inscription, anyone who could write could have carved it. While it can never be proven, I feel the features mentioned above are most naturally explained in terms of a text carved straight onto the whorl by the person who composed it.


BUCKQUOY

167

DESCRIPTION OF OGHAM INSCRIPTION Provided the curve is in keeping with the scale of the lettering there is no particular difficulty in reading a circular ogham, that is, as long as one knows where to start and in which direction to go. Indeed, a circular stem seems a very natural arrangement of a Type II ogham (i.e. one with draw-in stem-line) on a disc: much more obvious than, say, the peculiar configuration on the Ennis bead (CIIC 53). Since ogham always reads from left to right (whether horizontally or vertically), the point at issue is: does one stand, as it were, at the centre and look out, thus reading clockwise, or stand on the perimeter looking in, and read anti-clockwise. If one were keen to preserve the convention of reading vertically up the left, then one would adopt the former stance. This appears to be the direction of the Logie Elphinstone ogham, as indicated by the slope of the letter-strokes. Both Jackson and Padel preferred to read the Buckquoy ogham in this direction too, though the slope of the Buckquoy letters unequivocally indicates that is should be read owz-clockwise. Padel attempted to establish a pattern for the direction of reading circular oghams [1972:13-15], but, since the wheel oghams in the Book of Ballymote are quite different in character, the only true parallel is the circular ogham from Logie Elphinstone, which reads clockwise. If anything, the Buckquoy ogham should set the precedent for the much more inscrutable Logie ogham, but there is no need to set the authority of one against the other, since the letters of each slope in opposite directions and leave no room for doubt.

Jackson recognized that the M (11) was, as would be expected, diagonal to the stem-line, but, because he was convinced that 'die other strokes are evidently all intended to be at right-angles to it', he was not sufficiently sensitive to the pitch of the letters ~ '(m)ost are in fact more or less atright-anglesto die point where they reach the line'. The few which he conceded were 'more sloping' could be explained, in his opinion, by the distorting effect of a circular base-line [Jackson 1977:221], A close examination, however, reveals that all of the b- and h-aicme consonants are oblique. Compare, for instance, letters 4 and 5 (DA). There can be no doubt that the slope is deliberate - across any individual letter-group die stem is so short as to be, in effect, straight - but this sloping is exactly what one would expect.


BUCKQUOY

168

The sloping of b- and h-aicme consonants is a widespread feature of Type II inscriptions in Scotland. Although the gradient may vary, the incline is always in the same direction relative to the direction of reading, i.e. with the proximal end further 'forward' than the distal. Placed next to one another a correctly orientated B and H will form an arrow-head which points in the direction of reading. The pitch of maicme consonants, which slope left-to-right (up) across the stem, is not diagnostic since their outline is maintained even if the inscription is inverted, and read back-to-front. B- and h-aicme consonants, however, would, in such circumstances, point the wrong way. Even if the slope of consonants were not sufficient to indicate the direction of reading the Buckquoy text, further, unequivocal evidence is provided by die hammer-head A. Tlúsforfid appears on five other Scottish ogham inscriptions always with the cross-stroke on the right/lower distal end. There can be no doubt that Buckquoy is to be read anti-clockwise with the b-surface nearest the perimeter.

In Irish dessel, ? sun-wise\ 'right-hand-wise', means by extension 'lucky' 'favourable', 'propitious' [DH s.v.], and a similar semantic association is reflected in other Celtic languages. Given this well-attested preference for motion in a sunwise direction it might at first glance appear surprising that the Buckquoy text reads anti-clockwise, but it is all a question of perspective: though die text runs anti-clockwise, to read it one must turn die whorl clockwise. Thus as die spinner spins her yarn, and die spindle turns sunwise, the text passes legibly, if at speed, before her eyes. The presence of die pin would in any case make it exceedingly awkward to carve and read the text if it ran in the opposite direction, if, that is, one assumes the inscription was carved while the whorl was still in use for spinning.

The only remaining question is where to start. Circular texts in any script are not a common feature of Insular epigraphy. Where they occur their layout is usually a result of die shape of die objects on which they are carved, for instance, rings, coins, seals and stone fonts. The most common device for indicating the starting point of a text is the small square cross (e.g. Aldborough sun-dial [Okasha 1971:47]). Occasionally spacing is used to indicate word-division (e.g. die font from Partrishow, Cardiganshire (CIIC


BUCKQUOY

169

988)), but quite often the text is continuous and the inscriber relies on the skill of the reader in discerning where to start (e.g. the Attleborough ring [Okasha 1971:49-50]). Since circular texts are usually short, this is rarely a problem.

There are marked differences in the spacing between the various letters of the Buckquoy ogham, and it might seem logical to start after one of the larger gaps, 516 and 112, i.e. at C or N. In both case, however, the resulting text is unintelligible, and it seems more likely that the variation is merely the result of the casual way in which the text was carved.

In most cases it is obvious where an ogham texts begins, but in some Type II Irish oghams, whether manuscript (St. Gall Priscian) or epigraphic (e.g. Ballyspellan brooch (CIIC 27), Killaloe slab (54), Tullycommon bone (52), Colman Bocht slab, Clonmacnoise (749)), a non-phonetic character is used to indicate the starting point and direction of reading. Both Padel and Jackson interpreted Buckquoy's letter / ( * ) as an instance of this 'feather-symbol' or 'feather-mark'.

Yet a preliminary survey of Irish

epigraphic and manuscript oghams showed that in every case the 'feather-mark', perhaps more aptly labelled 'arrow-mark', is shaped > — or -^—, never x — or -*— [Brash 1879:322-3]. There are two Scottish examples of inscriptions beginning with x-(Lunnasting and Cunningsburgh 3), but in both cases it seems preferable to take the character as the first foifid and accord it its usual phonetic value (Id). There are no Scottish examples of a > — or - > — shaped directional indicator. It appears, therefore, that 7 is to be taken as /e/. This interpretation is based on its unambiguous shape, but has the added advantage of removing an otherwise unfeasible cluster of consonants.

The stem of the Buckquoy ogham is not a seamless ring, like that of Logie Elphinstone, but rather a looped line with two distinct ends. These overlap slightly, just above 1, and it would seem natural to take this break in the stem as the start of the text. This leaves the problem of the relationship of 1 to the rest of the text, because it is out of alignment with the other characters, and lies outside (below) rather than on


BUCKQUOY

170

die stem. Strictly speaking, the first letter after the break in the stem is 2, but a close examination shows that 1 is in fact attached to the stem. The left-hand end of the stem curves down slightly and meets the upper left arm of the forfid just above its intersection. It is as if 1 was indeed sitting on the tip of the stem, but that the whole thing had been bent down 90°. There is ample space for 1 to sit on the stem between the break and 2 and no obvious reason why it should be suspended from rather than intersected by the stem.

The precise arrangement of strokes in this portion of the text is not easy to rationalize, however one chooses to interpret it. It is clearly confused to some degree - the final stroke of 12 is hard-up against the x-forfid and partially, but not unequivocally, differentiated from the first two strokes of 12, Why the large gap before 2 ? Why the intersection of the two ends of the stem and the placing of 7 on a different alignment ? Dr Oliver Padel suggests to me that the curling-out of the stem is a deliberate device to show the stan of the inscription [pers.com.]. While this is without parallel, certainly it is a possibility.

The palaeographical conventions of ogham have to be deduced solely from observation of the extant inscriptions. Since there are relatively few post-seventh century examples and these are heterogeneous, it is sometimes difficult, in analyzing variations in individual inscriptions, to differentiate the deliberate and telling from the incidental and insignificant. This is particularly true of informal inscriptions such as Buckquoy, where the difficulty is further compounded by the virtually unique circular arrangement of the text. Generally, unambiguous mistakes are rare, but it would be foolish, either to see errors everywhere, or to deny that they ever happen - a corrected mistake is the obvious explanation of the configuration of Buckquoy 9. If one were prepared to accept the possibility of another mistake, or at least a change of plan, a possible explanation is suggested by the observation that the two ends of the stem overlap to form a little cross. Perhaps, having quickly cut the stem, the ogham-carver was struck by this x-shaped intersection and decided to press it into service as an x-forfid, but, having cut die rest of the letters, concluded that


BUCKQUOY

171

this device was not sufficiently unambiguous and added a separate x (/), putting it to the side to avoid it being interpreted as a second E.

Balancing the demands of an unforced interpretation of the carving against plausible readings of the text is the epigrapher's perennial problem. We might feel intuitively that the beginning and end of the text should coincide with the break in the stem, but given the obvious casualness with which the inscription was carved, the poor spacing of the letters and the fact that at least one mistake was made and corrected, how much confidence can be placed in such a hunch ? If an intelligible reading is provided by starting with the stroke before / to what extent should this be allowed to influence our interpretation? For the sake of convention only and without prejudice to the final reading, I take the x-forfid as the first letter and read anti-clockwise, as follows [see fig. for key to numbering]: I

A small cross below the line, consisting of two short lines intersecting more or less at their midpoint, the angle between the upper two arms being more than 90°, joined to the tip of the stem as described above. This character, the x-forfid, appears on a total of seven Scottish oghams including Buckquoy, though in initial position only on Lunnasting and Cunningsburgh 3. In both cases the letter is followed by TTE, and, even though neither text has been satisfactorily interpreted, given the apparent avoidance of double letters in word-initial position, it seems preferable to take the symbol as having its normal phonetic (vocalic) value /e/. Above theforfid, the stem curves round sharply, crosses its other end, then commences its circuit. There is a generous gap before the next letter. Jackson described it as sufficient for five strokes, although I doubt it could have taken more than two, properly spaced. Given the comfortable spacing of the next few letters, it need not be considered remarkable and is certainly no bigger than the gap after 5.

2

Five strokes to therightof the stem, sloping forward - N. The first stroke is somewhat doubtful, being faint and at a slightly greater angle than the rest (making due allowance for the curve of the stem).


3

BUCKQUOY

172

Two strokes to the left of the stem, sloping very slightly forward - D. These two are roughly parallel for most of their length but converge at their distal ends. According to Jackson, they do not join. The first stroke appears to over-shoot the stem very slightly. There is a line, like a bind-stroke, which continues well beyond the end of the first stroke. Jackson dismisses it as 'a fortuitous scratch', but it is possible that it is indeed a bind-stroke, similar to one found on the Burrian stone which has a single curved line do duty for the bind-stroke and the final letter-stroke of the group.

4

As above - D. These two are slightly closer together than the previous pair, and at a sharper angle to the stem.

5

A single long stroke almost perpendicular across the stem with a horizontal line across its lower distal end - a hammer-head A. This letter is clear and definite. Padel thought the cross-stroke unusually long. There follows a generous gap, the largest in the inscription.

6

Four strokes to the left of the stem, sloping forward with constant gradient - C. Jackson describes the second stroke as crossing the stem slightly.

7

Three strokes to the left of the stem, sloping forward with increasing gradient - T. Padel saw a very short possible score just after these but decided it was probably not part of the letter. It should be noted that there are a number of nicks and scores in the surface of the stone which are clearly just casual damage.

This first half of the text takes up two thirds of the stem, thus, although the first seven characters are very comfortably spaced, the next six are rather cramped. 8

A single stroke across the stem - A. This stroke is of very similar length to those of 10 and substantially shorter that 77, so there is no doubt that it is a vowel. It occupies the middle twothirds of the ogham band and appears to slope very slightly backwards.

9

Five strokes to the right of the stem, sloping forward at markedly different angles - N. The second and third strokes almost meet at the stem, and the final stroke intersects the first stroke of 10 and thus fails to reach the stem. The form of this letter has been discussed in detail above,


BUCKQUOY

173

there is no justification in taking it as 2+3. 10

Five strokes fairly perpendicular across the stem - 1 . Though the individual strokes are a little straggly, they are all more or less parallel, of similar length, and spaced as widely as 2, 6, and 7.

11

A single long stroke obliquely across the stem, slopingforwards- M. This is the longest stroke of all, about double the length of the preceding vowel-strokes, and very firmly sloped. There follows a generous gap, perhaps left to avoid the end of the underlying zig-zag.

12

Three strokes to the right of the stem, sloping forwards. The first two are exactly parallel and very tightly spaced, the third follows at a slightly more gentle gradient after a larger gap. There is some doubt over whether these should be taken as a single group of three strokes, in which case - V, or as two separate letters, 2+1 - LB. Elsewhere in the text the component strokes of individual letters are generally evenly spaced, but in contrast there is considerable variation in the size of the gaps between letters. The interval between the second and third strokes of 12 is comparable to that between 415, 718, and 8/9. There is no gap at all between 9110. The ogham text does not appear to have been carefully laid out, if the carver was running out of space this may have been the maximum spacing possible. There is certainly a larger gap between the second and third strokes than between the third stroke and the intersection. Furthermore, the second stroke is far longer than the other two. This may be nothing more than a slip of the blade, but, if intentional, might be intended to differentiate two closely packed though separate letters of the same aicme. On balance I would prefer to take 12 as two separate letters [cf. Holder 1990], though Jackson and Padel took it as one.

Which gives the following reading: ENDDACTANIM(v/lb)

In essentials this is in agreement with Holder's more tentative reading - E(s/n)DDACTA(n/lv)IM(v/lb) - and the possible variant passed over by Jackson and Padel - (e/ )(s/n/ )DDACTANIMV - dieir preferred reading was (e/ )TMIQAVSALL(c/q).


BUCKQUOY

174

FORM OF SCRIPT The Buckquoy whorl is inscribed in a simple Type II ogham script with the incorporation of the two forfeda most common on Scottish oghams - x (Id) and the hammer-head A J_ • The hammer-head A occurs on six other Scottish ogham stones: Birsay 1, Burrian, Lunnasting, Latheron, Golspie, and Formaston. The x-forfid, which isfoundon numerous Irish pillars inscribed with Type I ogham, is used on Burrian (three times), Newton (once), Golspie (twice), Cunningsburgh 3, and Lunnasting (once each at the beginning before TT), and at Formaston (twice). All the consonants of the Buckquoy text are sloped. The vowels, with the exception of the wojorfeda, consist of perpendicular strokes across the stem, taking up, perhaps the middle two-thirds of the ogham band, though the variation in length of individual strokes makes it difficult to be accurate on this last point. Within letters, the strokes are fairly parallel and evenly spaced. Substantial gaps are left between all letters, not just successive ones of die same aicme. If the suggested reading is correct, there is no attempt at word-division.

In general terms the script is very similar to that on the cross-slab at Latheron, which has the hammer-head A and consonants which slope, thought not as consistently as Buckquoy. The other ogham-inscribed crossslab from Sutherland, Golspie, is also similar in script to Buckquoy, but has angled vowels, and letters which are not so generously spaced. The Buckquoy script is simpler than those used on the three Birsay stones. Birsay 1 has the hammer-head A and sloping consonants, but, like Golspie, has angled vowels. Birsay 3 is written with bound letters. The Burrian ogham provides an example of the final stroke of a letter group being a continuation of the bind stroke, and it is possible that Buckquoy *s 3 is an attempt at such a letter, but equally it may be a mere slip of the blade.

The only exceptional aspect of the Buckquoy ogham is that it is carved on a circular stem-line. This, however, does not affect the aspects of the script mentioned above. Changing the shape of the stem is one of the ways to create a cryptic ogham mentioned in the Ogham tract in the Book of Ballymote [Macalister 1937:48), although, interestingly enough, a circular stem is not given as an example. While the only odier


B UCKQUOY

175

instance of a circular ogham, Logie Elphinstone, may indeed be cryptic, there is nothing enigmatic about Buckquoy. The use of a circular stem is an obvious solution to the practical problem of fitting the text to the object and is, in a sense, comparable to the 'up-top-down', or boustrophedon, arrangement on the oldest Irish pillars. There is no need to assume influence between Logie Elphinstone and Buckquoy, in either direction. The circular stem is a simple innovation which may have arisen independently in different areas. The fact that the two are read in opposite directions, and are very different in length and in context, is sufficient to suggest no direct relationship between the two.

INTERPRETATION OF TEXT For the reasons given above, eTMIQAVSALL(c/q), the favoured transliteration of Padel and Jackson, is based on an incorrect reading of the text and must be rejected. There is slight doubt over letter 12 (probably LB, but could possibly be interpreted as V), but the rest, ENDDACTANIM-, is secure. The only other doubt is over where to start. Though Jackson did in fact give ENDDACTANIMV as one of his eight alternative readings, he rejected it as 'wholly unintelligible and cannot be Celtic', so strong was his conviction about the nature of Pictish. In conclusion he stated himself to be 'content to write off this inscription as unintelligible, like all the other 'Pictish' inscriptions' [Jackson 1977:222]. It is simply not true to say that all Pictish inscriptions are unintelligible. Although a number do continue to resist interpretation, e.g. Inchyra and Logie Elphinstone, following Padel [1972] I hope to show that many, e.g. Latheron, Scoonie, and Ackergill, contain recognisable Celtic personal-names. I have argued against Jackson's hypothesis of a second, non-Indo-European Pictish language [Forsyth, forthcoming 1996c] and prefer to view the not inconsiderable difficulty of many Pictish inscriptions as an inevitable result of trying to interpret often damaged inscriptions in an otherwise only minimally-attested P-Celtic language. The proportion of doubtful or disputed Anglo-Saxon runic inscriptions, as given by Page, is a forceful reminder that epigraphic opacity is not an exclusively Pictish phenomenon [1973:14-51.


BUCKQUO Y

176

If one is prepared to begin with the stroke before 7, then die reading BENDDACTANIML suggests itself. This I would segment BENDDACT ANIM L, which I take to be Old Irish bendact anim I , 'a blessing on die soul of L \ According to diis new interpretation, BENDDACT is Old Irish bendacht, 'a blessing', from Latin benedictum [DEL and Vendryes 1959 s.v. bennachn Thurneysen 1946:450 §727] and ANIM, Old Irish animy 'soul' [Vendryes 1959 s.v. anim; D1L s.v. ainim(m); Thurneysen 1946:214 §333]. As it stands anim could be nominative, accusative, or dative singular, as attested in die Old Irish Glosses, but in the current context is most likely to be dative or accusative (see discussion below). This word is not to be confused with Old Irish ainm 'name', which occurs, in its older form ANM, on a number of classic ogham pillars in southern Ireland [McManus 1991:118 §6.27], and continued to be used in a technical sense in die Irish phrase ainm n-oguim 'funerary inscription in ogham' [ibid. 154 §8.8].

The spelling anim is relatively straightforward, diough as accusative or dative it alternates in die Glosses widi die forms anmuinlanmain [Thurneysen loccit.).

BENDDACT for bendacht is perhaps more

interesting. The gemination of consonants is a much discussed feature of ogham orthography [McManus 1991:124-6 §6.30; Harvey 1987a], and is particularly common in Scottish ogham. To die extent that it can be determined, diere is an apparent tendency for non-initial D to be doubled in Scottish ogham texts (exceptions being Blackwaterfoot 2, Latheron, Lochgoilhead) though whether this had phonetic significance is unclear.

If it were thought diat die Buckquoy text perpetuated aspects of die traditional ogham

ordiography, DD could be taken as indicating (non-lenited) /d/, on analogy widi CC and TT representing (non-lenited) /k/ and /t/ [McManus 1991:125]. Similarly, single C for /x/ could be an example of die same geminate/non-lenited--simplex/lenited distinction. However, as Dr McManus points out to me, diere are numerous parallels for such spellings in Old Irish manuscript orthography, and it is equally possible that the Buckquoy ogham is merely a cipher for die manuscript spelling. In Old Irish die spelling ct 'not infrequently represented' cht [Thurneysen 1946:21 §28], and bendact is attested in die Glosses [Wurzburg 19bl5, see Stokes and Strachan 1901-3] and on one roman alphabet inscription (CIIC 868).


BUCKQUOY

177

If, for whatever reason, the reading L were rejected, it would be possible to take 12 (LB) as a single letter consisting of three strokes, V i.e. /w/ or /f/, and still read it as Irish - not BENDDACT ANIM L, but VENDDACT ANIM or, rather, ANIM VENDDACT fthe soul of Vendact' [I owe this suggestion to Dr Paul Russell, Radley College and Jesus College, Oxford]. This would require taking Vendact as an otherwise unattested compound of the popular personal name element vendu- 'white, bright' [McManus 1991:103 §6.4] with the suffix -oct- (EOIr. acht) which was commonly used to form personal names from nouns and adjectives [McManus 1991 §6.13], or an archaic form of Findach* The third letter of the baicme may be transcribed as V or F depending on the period [McManus 1991:122; Sims-Williams 1993:140-3], but the spelling ven- is problematic. One would have expected the first syllable to show vowel affection, a change which occurred in Irish by the sixth century, certainly before syncope and apocope [McManus 1991:87, 93]. Apart from problems of absolute chronology, in terms of the relative chronology of sound changes in early Irish, vowel affection occurred before apocope (loss of final syllables), yet Buckquoy is a post-apocope text (*Vendu-oaos). Thirdly, we would expect the name to appear in the genitive case, yet it displays no genitive ending (-achta). These linguistic objections render preferable the original interpretation Bendacht anim L.

DISCUSSION The formula bendacht for anim M, 'a blessing on the soul of N. \ is well attested in the Irish epigraphic record, mainly on recumbent cross-slabs (e.g. bendacht for anmain N. (CIIC 551, 933,935(; bennachtfor anmain N. (CIIC 916, 917(). Other related forms are benfdachjt Die for anfmainj N. (CIIC 529) and bendacht arN. (CIIC 586,958). Of particular interest is the examplefromLemanaghan, Co. Offaly (CIIC 868) bendactfor anfmaijn ailbenig - with c for /x/, as appears mutatis mutandis in ogham at Buckquoy. All of these are written in the roman alphabet, but there is a single, very late, example of the use of the formula in an ogham inscription -- the eleventh century, bi-lingual, runic-Norse/ogham-Irish slab from Killaloe (CIIC 54) which has bendacht [ar] M, with the first E spelled with the x-forfid: BxNDACHT.


BUCKQUOY

178

In all these examples anim is in the accusative, governed by the preposition ar or for, the latter occasionally abbreviated (as on CIIC 529, 916, and 917), to/[Thurneysen 1946:162 §251]. The omission of the preposition from the Buckquoy text therefore demands an explanation. Epigraphers are usually quick to identify grammatical and other 'errors' in the inscriptions they study [see Macalister's attempt to 'correct' Pictísh grammar, 1940]. The Buckquoy oghamist was certainly casual, it could be that he or she was careless or ignorant as well With funher study, however, it often turns out to be the epigrapher who is ignorant, and 'bad' Irish should be invoked in this case only once all other possibilities have been exhausted. In feet there is an alternative explanation.

In his discussion of Archaic Irish, Greene confirmed that 'the elimination of unneccessary particles is one of the outstanding features of the archaic style1, which he says 'accounts for the consistent use of the independent dative where Old Irish would require a preposition1 [1977:19]. Thurneysen makes brief mention of the prepositionless dative in his Grammar of Old Irish, characterizing this nominal construction, which is attested in certain Irish legal texts and paralleled in Gaulish inscriptions, as 'archaic and extremely rare' [Thurneysen 1946:162 §251]. If Buckquoy f s anim were in the dative case, then we might have here an example of the 'independent dative'. The extreme rarity of this construction in Old Irish warrants caution, as Dr McManus has reminded me, yet archaic syntax would not be out of keeping with a seventh or eighth-century date for the Buckquoy whorl. However, as relics of an earlier period of the language, 'archaic' features continued to be used in Old Irish as a stylistic device [Breatnach 1984:458-9]. This need not, of course, rule out an early date for our text, especially since a heightened literary register is unlikely in a casual epigraphic context such as this. Nonetheless, an independent dative, if indeed that is what it is, need not necessarily indicate an early date, and thus the text cannot be dated on the basis of syntax alone.

A possible objection to the reading BENDDACT ANIM L is that the use of an initial letter, in this case L, to represent a personal name is unprecedented in a Goidelic context at this period. In the ogham


BUCKQUOY

179

tradition, as in the runic, each character was identified by a meaningful name on the acrostic principal. The letter-name for ogham L was luis, though there is some doubt whether this is luiselloise 'flame, blaze* or his 'plant, herb* [McManus 1991:36]. Later glossators identify luis as a tree, either the rowan or elm [McManus 1988:150]. Rune-writers exploited the letter-names to great effect in writing riddles in which the runic character was made to stand for its name [Page 1973:203-11], but there do not appear to be any examples of such a conceit in the early ogham corpus. Luis is not attested as a personal name, so 'L' would have to be interpreted as a straightforward abbreviation of some other personal name.

The concept of abbreviation of divine names and epithets was well known from the use of the nomina sacra but does not seem to have been extended to profane personal names. To establish this beyond doubt, however, is not easy since there are few close comparisons. Irish personal names are too varied for a comprehensive system of abbreviations such as was used in Roman epigraphy. For this reason the contexts in which it would be possible or appropriate to abbreviate a Goidelic personal name are limited. Unless there is some cryptic intent, a personal name is likely to be abbreviated only if the identity of the individual is obvious and widely known. Perhaps the most usual circumstance is the marking of ownership on a personal belonging. Most of the extant, non-monumental, insular inscriptions are on objects of great value, usually highly-accomplished pieces of secular or ecclesiastical metalwork such as brooches, weapons, and shrines. One would not expect the names of the owners, patrons, makers or honorands to be abbreviated on such formal, in some cases public, items inscribed with posterity in mind. It is instead on instmmenta domestica, transient and of little economic value, that one might expect to find the initials of an owner or maker. But of course, such objects are rarely preserved in the archaeological record. Though there are no clear examples from Celtic-speaking areas, the corpus of Anglo-Saxon runic inscriptions includes a handful of examples of objects inscribed with a single letter, or a pair of letters, most readily interpreted as owner's marks (the wooden York spoon, the Sleaford brooch, the Willoughby-on-the-Wolds bowl, mentioned by Page [1973:172]).


BUCKQUOY

180

Reasons why the alleged personal name was abbreviated in die Buckquoy text when the rest was written out might be the obvious lack of space (if die individual was well known to all the inhabitants of die farmstead, to give it in full might have been considered superfluous), or perhaps die Irish-speaker who carved it was unsure how to commit an unfamiliar native name to writing. Of course, an abbreviated message inscribed on a gift provides a shared understanding between giver and recipient thereby enhancing die intimacy of die bond between diem, and this may be sufficient explanation of itself. An appropriate analogy might by a locket engraved widi a sweetheart's intials.

On die semantic level, one might object diat 'a blessing on die soul of N.' is an essentially commemorative phrase and as such scarcely appropriate for a spindle-whorl. Certainly all the examples of bendacht or If or cited thus far occur on funerary monuments. The formula, however, was not restricted to gravestones: Macalister read bendacht ar Artgal on die Cross of Kells [CIIC 586]; in metalwork, the mid-eleventh century Stowe Missal Shrine [CIIC 932J has bendacht De ar cech anmain as a hairilliuth 'God's blessing on every soul according to its deserts'. These are commemorative or dedicatory without being funerary, as are die occurances of die formula in die scribal colophons of manuscripts. An example of die latter widi Scottish provenance is die colophon to die Gospel extracts in die Book of Deer (Cambridge UL MS.I.i.6.32). Written in die same hand as die main text it is to be dated to die ninth century. It asks of die reader a bendacht fforj anmain in truagain ro-d scribai 'his blessing on die soul of die poor wretch who has written it' [Jackson 1972:8; Stuart 1869:lx, 89].

The examples are all public texts on high status items written widi posterity in mind and thus, still radier far from the informal domestic object from Buckquoy.

Perhaps a closer parallel is provided by die

inscribed pebble found in 1822 in a grave at Temple Brecan, Na Seacht dTeampaill, Inis Mór, Aran Islands (CIIC 532) [Higgins 1987.1:23, 26-7, 137-9; 1987.2:268-9, fig.l, plate 54]. The black limestone pebble, 75mm in diameter, was lost for many years and known only from drawings. It was dien widely interpreted as a stone lamp, and thus appeared to provide a cognate example of a domestic implement inscribed widi


BUCKQUOY

181

a dedicatory text. A few years ago, however, it was re-discovered in the National Museum of Ireland, enabling the most recent writer to rule out this explanation [Higgins 1987.1:26]. Instead it belongs to a class of 'decorated pebbles' of uncertain, probably religious, use. The circular inscription in Insular minuscule reads +or(oit) arbran n-ailither, 'a prayer for Bran the pilgrim' [Macalister 1949:6,202-3 ill.]. This formula, Or(oit) ar/do, is by far the most common of Irish roman alphabet inscriptions and is found widely, though not exclusively, on funerary monuments. According to Higgins the Temple Brecan stone 'has all the signs of wear from use and handling to suggest that it had a previous function' [1987.1:23], yet the fact that it was discovered in a grave clouds the issue somewhat; it is unclear whether the inscription had been present while the item was in use, or was carved expressly for deposition in the grave. Whatever the precise interpretation, the existence of another informal, portable, personal belonging inscribed with a formula that is more familiar from grave slabs makes the Buckquoy whorl seem less exceptional.

Parallels A number of inscribed spindle-whorls have been recovered from Insular and Continental excavations. Such an intimate and personal item might be considered particularly suitable for inscribing, though the surprising frequency with which inscribed whorls are recovered may be due, perhaps, to their relative durability, coupled with how easy they were to lose, rather than because they were particularly favoured for inscriptions. Historically and cross-culturally, spinning is an activity most closely associated with women, so opening up the possibility of interest, or even skill, in literacy on the pan of women. The spindle-whorl texts vary widely in length, complexity, and meaning. Most simple are those which carry merely a personal name, presumably that of the owner. An example is the seventh or eighth century Anglo-Saxon jet spindle-whorl from Whitby Abbey inscribed with three runes, Wer [Peers & Radford 1943:74], which has been interpreted as a (male) personal name [Page 1973:171*2]. Next is the writer/maker formula for instance, another nine inscribed spindle-whorl, this time from the Northern Isles (exact provenance unknown), which reads 'Gautr wrote the runes' [NMS BE 360; Liest0l 1984:232], Personal belongings,


BUCKQUOY

182

especially tools and weapons, are sometimes inscribed with talismanic or protective inscriptions, whether pagan or Christian. Scottish examples would be runic ruthorks, such as the seal's tooth from Birsay [Liest0l 1984:232; Curie 1982:59-60 III 37], or the St Ninian's Isle chape [Small et al. 1973]. The Buckquoy whorl may tall into this category, as might the pseudo-ogham-inscribed whorl from Burrian [MacGregor 1974:91 fig. 18; see Appendix - Dubia]. A fourth category, though attested much earlier, shows that such inscriptions may have a ludic rather than strictly practical or talismanic function. These are the corpus of Gaulish-, Gallo-Latin-, and Latin-inscribed whorls from eastern Gaul [Whatmough 1949, Lambert 1994:122-5]. Meid has interpreted these texts as subtly, or more directly, erotic messages, from implied male speakers, addressed to young women. As such he characterizes them as reflecting 'spinning room amusements' [Meid 1992:52-54, ill.]. According to Lambert such objects, bearing 'des souhaits ou des compliments amoureaux\ were given 'en cadeau par des galants dont les intentions sont claires* [1994:23]. The fact that whorls were made by men and given to women, especially sweethearts, in other cultures too, is indicated by a Norse example in Rekjavik Museum, Thorá migfrá Hruna Thora owns me, from Hruni' [referred to on NMS label for spindle-whorl BE 360]. The Christian sentiment of the Buckquoy whorl is a long way from the double-entendres of the Gaulish carvings but it is not inconceivable that 'L' was a man (an Irishman ?) who inscribed the Orkney spinster's whorl in the hope of future remembrance. Page gives as an alternative interpretation of the wer of the Whitby spindle-whorl the northern form of West Saxon wcer 'token of friendship' [1973:171].

All of these comparanda are examples of inscriptions carved while the whorl was still in use for spinning. In more recent times, however, former spindle-whorls were used in the Scottish Gáidhlteachd as charms [pers.comm. Dr John Maclnnes of the School of Scottish Studies]. It is possible that die Buckquoy text relates to a time when the whorl had ceased to be perform its primary function and had taken on talismanic significance. There are no explicit early medieval references to whorl-charms, but Adomnán provides seventh century evidence of the use of pebbles as amulets. In VC he refers first to a small, unspecified benedictio, ('object that has been blessed'), which was housed in an inscribed box, and, once dipped in


BUCKQUOY

183

water, was used to cure a broken hip [VC ii.5 p. 103 n.134]; then to a piece of rock salt blessed by the saint which was hung on the wall above the bed of an invalid [ii.7 p. 105]. Most relevant to Buckquoy is the lapis candidusy 'white stone', which was taken from a Scottish river, blessed and used to work cures among the heathen Picts. Adomnán has Columba say, Signate ... hunc candidum lapidem, per quern domimts in hoc gentili populo mulms egrotonim perficiet sanitates, 'Mark this white stone. Through it the Lord will work many cures of the sick among this heathen people' [VC ii.33]. Though signate doubtless means mark with the sign of the cross, is not the Buckquoy spindle-whorl a white stone, marked with a Christian message ? Though there is considerable doubt about the interpretation of the ogham on the famous Ennis bead (CIIC 53), it should be noted that it was believed to be of assistance to women in labour, and efficacious in the treatment of eye complaints [Brash 1879:321; see Meaney 1981 for discussion of Anglo-Saxon curing stones].

Historical Context As a complete, legible but. according to the two authorities, unintelligible ogham, Buckquoy might be considered one of the stronger planks in the argument for a non-Indo-European Pictish language. If it could, however, be proved to be comprehensible, and the proportion of uninterpreted inscriptions thus reduced, the case for non-Indo-European Pictish would be weakened accordingly. Of course, if the ogham is Old Irish, this implies nothing about the nature of the Pictish language, merely that there were Irish speakers in Orkney in the eighth century. Thus, if accepted, the interpretation of the Buckquoy ogham text as a Christian phrase written in Old Irish is an important breakthrough, not only for the linguistic information it provides, but also because of what it indicates about the spread of Irish Christianity to the North.

When discussing external influences on pre-Viking Orkney, commentators have, quite rightly, tended to stress the eastern connections with mainland Pictland and Northumbria [see most recently, Lamb 1993]. The importance of the western seaboard as a conduit for influences from Ireland should not, however, be


BUCKQUOY

184

overlooked. There is documentary evidence for voyages by Irish peregrini to Orkney and beyond as early as the sixth century.

Both Brendan and Cormac Ua Lfatháin, who, according to Adomnán, visited

Columba on Hinba, came from areas well stocked with ogham stones [VC i.6, ii.42, iii. 17], a number of which have explicitly Christian associations. Of those marked with the cross, two refer to ecclesiastics Arraglen [CIIC 145] QRIMITIR 'priest', Maumanorig [CIIC 193] AILITHIR, 'pilgrim' (though the reading is problematic). Continued familiarity with the ogham script in the Irish church is suggested by the presence of ogham marginalia in manuscripts of the ninth century and later written by Irish scribes, not least by the ogham signature in the Stowe Missal, RIA MS.D.II.3 f.l lr, of one Sonid who describes himself as 'peregrimts1 [Warner 1915: xlii and plate viii].

Peter Harbison has amassed archaeological and ait historical evidence for a long tradition of pilgrimage along a western maritime arc stretching from the far south-west of Ireland to Shetland [1991:192-4, 221]. A spindle-whorl would be an unlikely piece of baggage for a monk on pilgrimage, unless, that is, it was already functioning as a relic or charm. The geological evidence, however, indicates that the Buckquoy ogham was not imported from Ireland, but instead carved locally, the implication being that there were people at Birsay in the eighth-century who understood the Irish language, either Irish settlers or bi-lingual Picts. The Orkneys were regarded by contemporaries as Pictish territory [Dumville 1976], but it is not inconceivable that there might have been an Irish colony there, as in so many other areas of along the western seaboard of Britain. The similarity of house-type at Deer Park Farms, die Udal, and Buckquoy, is just one feature worthy of further investigation in this regard.

The Christian sentiment of the Buckquoy text throws interesting light on the current debate over the relative strength of the Irish and the Nonhumbrian input in die evangelization of the Northern Isles [Morris 1990:9]. Continuing links between the Irish church and Orkney are demonstrated by the reference in die ninth-century life of St. Fintan of Rheinau [Thomson 1986, Lówe 1986] to an Orcadian bishop, presumably a Pict, who had studied in Ireland in his youth and could speak Irish to die fugitive Fintan. That he was


BUCKQUOY

185

the only person in the area who could, implies that knowledge of Irish was not widespread in Orkney in the 840s. It is impossible to evaluate the Buckquoy ogham without reference to the excavated remains on the Brough of Birsay, just a few hundred metres away. Although Buckquoy was a farmstead, not a high status site, its proximity to the Brough means it cannot be disassociated from it. Might the Buckquoy farmers have been monastic tenants of the Brough ? Perhaps, although recent interpreters have been turning away from the traditional identification of the pre-Viking structures on the Brough as a monastery [Hunter 1986:171, Morris 1990:15, 1993:287]. The ogham text implies no more than that the owner was Christian, certainly it need not imply that he or she was a monk or nun, nor that Buckquoy was an ecclesiastical site. The oghams most closely comparable to Buckquoy, those from Gurness, only ten miles away, from Pool, and from Bac Mhic Coonain, North Uist appear to have no ecclesiastical connotations. They share with the oghams from the Brough itself an informality which suggests that ogham literacy was not highly restricted, or reserved for solemn, formal purposes. It would be wrong, therefore, to assign an ecclesiastic interpretation to the Brough on the basis of the presence of ogham.

CONCLUSION In the above discussion I have attempted to explain the evidence as best I can, while frankly acknowledging the difficulties in both reading and interpretation. It is a frustrating reality of early medieval Insular epigraphy that vernacular inscriptions which do not conform to a firmly established textual formula rarely yield an utterly unambiguous reading [e.g. Williams 1949, Clancy 1993]. Even the intensely studied runic inscriptions of Anglo-Saxon England, written in a language vastly better attested than Pictish, present numerous difficulties and are much disputed [Page 1973]. In the circumstances it is perhaps unrealistic to expect such an informal ogham inscription as Buckquoy, from a little understood period in the development of the script, to be completely unequivocal. The most I have been able to show is that mine is a possible interpretation. If I have been unable to demonstrate that it must be so, then I await with interest a more convincing explanation.


BUCKQUOY

186

BIBLIOGRAPHY

OS Record card - HY 22 NE

Padel 1972:73-5; A. Ritchie 1977:181-2, 197 fig.8 no.84, 199 pi. 13a; K. H. Jackson 1977; Collins 1977; A. Ritchie 1987:62, 65; Holder 1990:1990:66-70.

PREVIOUS READINGS

Jackson Padel Holder

(e/ )TMIQAVSALL(c/q) or (e/ )(s/n/ )DDACTANIMV as above E(s/n)DDACTA(n/lv)IM(v/lb)

BUCKQUOY - 'Figure-of-eight' house in which inscribed whorl was found [Ritchie 1977:177 fig.3J

BUCKQUOY - Ogham-inscribed spindle-whorl [Crown Copyright: Institute of Geological Sciences, NERC]

I8p5

it,

10

12

BUCKQUOY ~ Ogham inscription (key to numbering)

187

BURRIAN

DISCOVERY The ogham-inscribed slab was discovered in 1870 during excavation of the Broch of Burrian on the small island of North Ronaldsay, Orkney (NGR HY 7627 5138). The excavator and proprietor of the island, Dr William Traill of Woodwick, was assisted by the antiquary Sir Henry Dryden and the excavation, which lasted from 1870-71, was comparatively rigorous for the period. Traill identified three phases of use, the primary broch occupation, a secondary occupation inside and outside the broch, and thirdly, early Christian activity, evidenced by finds only. It was in this third and final phase that the ogham slab was discovered 'towards the south side of the broch, where the wall was so low that, though the slab lay not much above the floor of the tower, it was also notforfromthe surface* [Traill 1890:346]. As early as 1871 Sir Henry Dryden had solicited Brash's views on the ogham, as is apparent from Brash's letter reproduced in 77?£ Ogam Inscribed Monuments of the Gaedhil [1879:362-3]. This is the first mention of the ogham in print and Brash's is the first illustration [see also pan of a letter from the late Brash published in Traill 1890:350-1]. Though Traill had published a report on his excavation [1890], die bulk of his finds remained unpublished until 1974 when MacGregor re-published die site, paying particular attention to die finds, including the inscribed slab. A sevendi- or eighth-century date was suggested for the early Christian phase at Burrian and theformof the incised cross has been assigned to the late eighth century on stylistic grounds [Thomas 1971:187]. In 1871, along widi odierfindsfromthe site, the ogham-inscribed slab was donated to the National Antiquities Museum where it is currently on display (NMS Cat. No. GB1).

SITE The four square miles of North Ronaldsay comprise die most northerly of the Orkiley tsles and 'surely one of the most bleak' [folacGregor 1974:63]. No-where on the island is more than 15m above sea-level. The Broch of Burrian is at the tip of a promontory in the very south-east of the island, overlooking a rocky shore and separated from the arable land by four earthen ramparts. The promontory is just to the east of


BURRI A N

188

the broad sandy 'South Bayf where the present day harbour lies [Hedges 1987.III: 105]. Some of the site has been lost to marine erosion and only the broch tower was excavated in earnest, so knowledge of the site is partial. Traill describes the 'one to two feet of rubbish' he found above the original broch floor, above this was 'an upper paved space, and various partition walls, built upon the debris that overlay and concealed the original floor* [1890:341]. These two pavements formed, in MacGregor's words, a clear stratigraphical reference point, dividing the architectural remains into primary and secondary phases. According to MacGregor, however, the separation is not as clear-cut as is often assumed. He points to discrepancies in plans and lists and wains that 'a great deal of caution is called for when using the stratigraphy in relation to the finds' [69-70].

A number of finds are diagnostically Pictish, most famously the ox phalanges inscribed with symbols (NMS GB 227), and the pebble inscribed with a pentangle and other designs (NMS GB 9). But there are other more humdrum items. Two artefacts, the cross-slab and an iron hand-bell (NMS GB 306), point to ecclesiastical activity at the site.

Radford claimed the site as a monastery or Christian hermitage

[ 1962:170], a view more recently supported by work on hand-bells of Irish type in Scotland [Bourke 1983]. Bourke has concluded that in Ireland the distribution and associations of iron bells, in possible contrast to those of bronze bells, suggests a connection with monastic churches. Iron bells of this kind were produced in Ireland from about AD 700-900 and their presence in Scotland 'can reasonably be attributed to the influence of the Irish church' [465]. At only 60mm tall, the Burrian bell is the smallest of all the extant early medieval hand-bells from the Celtic-speaking regions. Bourke entertains the possibility that it may be a cow-bell, but points to other clearly ecclesiastical bells which are only slightly larger [464]. Two other hand-bells of Irish type survive from Orkney, but both of these, from archaeological excavations on the Brough of Birsay and at Saevar Howe, were discovered in Norse contexts.

As MacGregor points out there are other examples of the foundation of monasteries on formerly defensive sites, with the ramparts then serving as a symbolic barrier between the clerics and die world [103].


BURRIAN

189

Marwick ascribed the early Christian remains at Burrian to a Ninianic mission [1952:1], but his identification of the name of the island, Rinansey or Rinarsay, with the popular form of the saints name, Rinan, has since (sic) been refuted [Taylor 1931]. The indisputably Irish background of the bell and, of course, of the ogham, renders any Ninianic, or for that matter, Northumbrian, connection unlikely. The subsequent history of the site is unknown, but impressive standing remains must have survived into the Norse period to have attracted the name 'Burrian' (< ON borg-in 'the fortress') [Marwick 1952].

DESCRIPTION OF OBJECT A small, smooth, rectangular slab, incised on one face with the outline of a cross and, parallel to the left margin, an ogham inscription. Stone:

Clay slate

Dimensions:

0.7m x 0.39-0.29m x 0.05m

Condition:

The slab is complete but its surface is severely laminated and much has been lost, including some of the carving. The ogham is, however, largely intact and clear.

The slab itself is a natural flag only roughly shaped. What survives of the carved surface is, however, very smooth. To such an extent, in fact, that Macalister suggested it had been deliberately smoothed to receive the carving. Unfortunately the stone is highly laminated and much of the friable surface has flaked off, especially in the lower and right-hand portions. According to Traill this damage was caused by the roots of plants which, on discovery, had penetrated the surface of the stone [1890:346].

From the excavation report there is no indication of how the stone lay when it was discovered, whether vertical or horizontal, or whether the excavators had an opinion as to whether it was in a primary or secondary position. From Traill's discussion [346], it appears he considered the slab an upright gravemarker. Whilst the stone bears a superficial resemblance to a recumbent grave-cover, the fact that the carving is restricted to the upper two-thirds may suggest that the stone originally stood upright in the


BURRIAN

190

ground, perhaps at the head of a grave. A cursory glance at CIIC II suggests that the carving on Irish recumbent grave-slabs tends to be more centrally placed and to cover a wider surface area. If it did stand upright, the slab would have been less than two feet tall (0.6m). Leaving aside the lost surface area, the extant carving is remarkably unweathered, which may imply that the slab was not exposed to the elements for long. Of course, the monument need not have been funerary, and the suggested interpretation of the text is non-memorial. The cross may instead have been devotional. If the position inside the broch is primary (which may account for the lack of weathering) and the site is to be interpreted as ecclesiastical, the cross may have been an item of church furniture, perhaps propped upright against the altar. In which case the comparison with the Flotta altar front gains added significance [see below].

DESCRIPTION OF CARVING Both cross and ogham are very lightly and delicately incised. The cross is a simple though pleasing blend of neatly executed straight lines and smooth curves, measured and regular, as if drawn with compass and ruler in hand. There is no trace of preliminary guide-lines or points, most likely the outline was drawn on with charcoal or traced round a template. Several writers have noted the contrastingly scrappy appearance of the ogham and commented unfavourably on its wavering stem-line and irregular lettering. The contrast between the two is bound to raise doubts over their contemporaneity, especially since the cross is slightly morefirmlyand deeply incised. But it would be wrong to make too much off this, the technique is basically the same for both, and the fact that the cross is placed a little to the right, as if to allow for the ogham, seems to prove that both are part of the same over-all design. If the suggested interpretation of the text is accepted the inscription refers to the cross which, though not incontrovertibly so, further supports the view that the two were carved at the same time.

DESCRIPTION OF DECORATION In addition to the ogham the slab is incised with the outline of an equal-armed cross (220mm broad), with circular joints and square terminals, on a narrow shaft which disappears into a patch of wear (total extant


BURRIAN

191

length c.370mm). Several writers have noted the very similar cross (though lacking the shaft) on the altar frontal from Flotta, one of the most southerly Orkney islands [Thomas 1971:186-8]. The cross form is not particularly widespread, though it does occur on the huge cross-slab at Rosemarkie ROS, again minus the shaft. Perhaps a closer parallel is found at Monymusk ABD or less precisely at Ulbster CAI, both of which share the distinctive feature of the narrow shaft which gives the Burrian example the air of a processional or free-standing altar cross. All four of the comparanda mentioned are decorated with interlace in some way, but the Burrian cross is plain. Of course, no trace of pigment survives, but if, as seems likely, the practice of painting sculpture was common, the smooth fine-grained stone of this slab would make it ideal for illumination. Thomas suggested a date in the late eighth or ninth century for the Flotta altar, based on a comparison of its interlace with that of a ninth-century cross in Cumberland [188] and on the assumption that the Burrian ogham is 'unlikely to be any older than the eighth century' [187]. To use Thomas's date for the Flotta altar to confirm the date of Burrian is thus a circular argument, based ultimately on the unfounded assumption that a developed form of ogham such as this can be no earlier than the eighth century. The probably seventh century radio-carbon dating of the bind-ogham-inscribed Birsay 3 shows that such an early date cannot be ruled out for Burrian. The Monymusk comparison may be as relevant as the Flotta, and it is doubtful if a date as late as the late eighth or ninth century could be accepted for the simple Class n Monymusk slab. No clear date for Burrian can be established on stylistic grounds alone.

The two other shafted crosses mentioned, Monymusk and Ulbster, stand on small rectangular bases, but the base of the Burrian cross, if it ever had one, is now lost. The disjointed patches of original surface which survive in the area beyond the current end of the shaft display some lightly incised lines. To the right are five or more oblique parallel lines some with mirror images below, the whole looking something like a feather; to the left are more nebulous curves. From the outset this has been referred to in print as a fish [Brash 1879:362], an idea perpetuated by no less an authority than Romilly Allen [ECMS 24]. While the carving to the right could be seen as a fish's tail (though unlike any in Pictish an), the design


BURRI AN

192

to the left is hard to rationalize as a fish-head. The original interpretation rested on no more carving than survives today, and there seems slender justification for it. The Burrian 'fish' faces left, yet with only a few exceptions (e.g. Edderton SUT which points up to the left at an angle of 45°) most Class I fish face to the right, their fins indicated, if at all, by a few slightly curving lines. There are only two Class II examples which face to the left, Golspie SUT and St. Vigeans 1 FOR, but neither has fins represented in this feather-like way.

Padel suggested a correlation between the fish symbol and ogham inscriptions and this idea has been repeated by others. In fact, of the nine ogham stones occurring with symbols, only four, Ackergill, Inchyra, Golspie and Latheron have fish. While this is higher than might be anticipated, given that the fish is not one of the more frequently occurring symbols (twelve examples on Class I, less than 10%), there is doubtful merit in performing statistical analysis on such a small sample. The marks are certainly artificial, but they are much less substantial than the cross and less substantial indeed than the ogham, so that even if one were to concede that they were intended to represent a fish, they look more like subsequent graffiti than a proper 'symbol'. There is certainly insufficient justification for classifying Burrian as a Class II slab. Macalister rejected the 'fish' and instead interpreted the carving as the lower ends of the shaft turned out and decorated with feathering. There are a variety of unusual cross bases in both Pictish and Irish an, though none of the form he describes. The NMS photograph reproduced here shows up a number of scratches and scores, some of which look more deliberate than accidental. On balance I think the 'feathering' is as likely to relate to this than to the original design.

OGHAM INSCRIPTION The ogham inscription runs vertically up the face of the slab a few centimetres in from the left margin, level with the cross. The extant length of the text is 345mm, the individual letters ranging in size from 15mm to 3mm. Thus, with 26 letters in the space of 35cm, Burrian is carved on a very fine scale. Ferguson described it accurately as 'the most minute lapidary Ogham hitherto found' [136]. The letters,


BURRI AN

193

most of which are bound, are very clearly spaced one from another though there is no explicit word division. Surface damage has rendered the beginning of die ogham indistinct. It is possible there were furdier letters below what is now visible, but no traces remain. The rest of the inscription is fairly clear diough there is slight doubt over a few groups, especially 213 and 10. As it stands, die lettering finishes level with die top of the cross. Above there is a pitted gap before the fracture line of die stone. There are no further markings apparent and the text probably ends here. Reading from die bottom up [see fig. for key to numbering]: 1

Five strokes perpendicularly across the stem -1. The left bind-stroke is visible but not die right since the surface is lost beyond die distal end of die strokes. Any letters which may have preceded have not survived.

2

The surface between here and 4 has suffered disruption, resulting in confusion and disagreement between the various audiorities. Padel and Macalister agree that first diere are two strokes to the left of the stem, bound, with die second protruding to the right of the stem. But one chose to interpret die bottle as half-full, the other as if it were half empty. Padel diought the right surface 'not too worn' and, while conceding a possible O, preferred to interpret the lines as a mistaken attempt at a D. On die other hand, Macalister saw it as an imperfectly preserved O. I diought I saw a diird stroke before the others, with the left distal end lost in a spall and continuing across the stem a similar distance to that of the third stroke. I concur with Padel that there is no obvious wear to the right of die stem at this point, but the first and third strokes certainly protrude onto that side. I am reluctant continually to postulate slips and errors so, for want of an obvious reading, proffer die following alternatives - T, D, O, U.

3

After a gap diere are two short strokes to the right of die stem, sloping forwards. Macalister interpreted these as L without further comment. The lack of a bind-stroke renders this unlikely but in any case it is clear that the second stroke, at least, continues across the stem. There is a nick at the point where die right distal end of die first stroke would lie, but no evidence of anydiing between diat and the stem. Padel interpreted these as two separate letters, BM, but was


BURRI AN

194

unhappy that they were so cramped. Neither mentions what appears to be a third stroke, sloping in the opposite direction, most obvious to the right of the stem, but continuing across to intersect with the second stroke just to the left of the stem. It is not clear if this stroke ever continued further. Given the ample spacing of the rest of the inscription I think it unlikely that there is more than a single letter here. That the first stroke is not bound to the second throws it into some doubt, since all other letters are bound. The third stroke is doubtful, but if accepted could be intended for the first forfid > > > ; each slightly larger than the previous one. The lack of bind-strokes is surprising, but nonetheless this group is probably best interpreted as an angled vowel - U. As Padel notes, the wear could be obscuring the further two or three strokes for which there is room, but there is no trace of anything further. The worn area is not panicularly wide and it would be thought that the distal ends of any lost strokes would have survived. If there are no letters missing, the apparent gap is hard to explain unless that area was already damaged when the ogham was carved.

11

Five or six strokes across the stem, bound - R. The lower tips of both bind-strokes are clearly


BURRIAN

196

visible before the first stroke (if five), the extra stroke, if present, would run along the fracture line of the laminated surface. It is curious that the first R of both pairs, 516 and 11112, may consist of six rather than five strokes. Each stroke begins progressively nearer the stem on the left and ends further away to the right with the result that the left bind-stroke slopes towards and the right bind-stroke away from the stem. 12

As above - R. The final stroke is angled at the stem with the point facing to the right. Slightly smaller and more compact than the previous letter.

13

A single stroke perpendicularly across the stem with a bind stroke on the right distal end - a hammer-head A.

14

Four short strokes to the left of the stem, sloping forward, bound - C. Each stroke is slightly longer than its predecessor with the effect that the bind-stroke slopes slightly leftwards, away from the stem.

15

Three short strokes to the left of the stem, sloping forward, bound - T. As with the previous letter the strokes get progressively longer.

16

One long stroke perpendicularly across the stem intersected by a second lying at an angle of less than 45° to the stem. The perpendicular stroke is deeper than the oblique one. Tliis has been interpreted as thefivstforfid > < and assigned a value K by Padel. It is, however, quite unlike 21 which is formed, not of straight intersecting lines, but of opposed angled lines which do not quite meet. The intersection does not occur exactly on the stem but slightly below it. If the slighter oblique stroke could be shown to over-lap the perpendicular one it might be possible to explain the carving as an attempt to correct a mis-carved M. If the initial stroke was carved perpendicularly rather than obliquely the second stroke may have been added to rectify the situation. The fact that this surprisingly long stroke runs so far to the right that it intersects with the first stroke of / 7 might suggest it was added after / 7 had been carved.

17

Four strokes across the stem, bound - E. The left bind stroke starts well before the first stroke (the way it trails off suggests it was cut with a stroke contrary to the direction of reading and


BURRI A N

197

overshot its mark). 18

Tliree strokes to the right of the stem, sloping forward, bound - V. They are well spaced and get progressively longer.

19

As above, though more tightly spaced and of equal length - V.

20

Four strokes to the left of the stem, sloping forward, bound - C. Each stroke of this tiny letter is slightly longer than its predecessor.

21

Two long angled strokes, opposing, not meeting > <EROCCS

FORM OF SCRIPT Macalister took the untidy execution as symptomatic of the 'unintelligent copying of a prepared model' by an inexperienced and possibly ignorant writer [1940:209]. He erred, however, in assuming that the oghamist's aim was necessarily to produce a neat, calligraphic and monumental text, like, for instance, Lunnasting. The Burrian script is of a quite different register. Of all the Scottish oghams this is the one most resembling ordinary 'hand-writing'. As such, it is a perfectly acceptable cursive text, which would not look out of place if it appeared in ink on the margin of a vellum manuscript. The slips and irregularities, such as over-shooting bind-strokes and variable stroke and letter size, are of the kind made by someone who knew what they were doing and understood the workings of the script. Rather than indicate ignorance of the script, I think the lax execution implies the reverse, familiarity with ogham and confidence in its use. Despite its somewhat careless execution, the Burrian text is clear and legible except where damaged. This is due to a number of features including the consistent use of bind-strokes, the varied but usually generous spacing between letters, and the general proportions of the script.

The Burrian oghamist has been criticized for not forming a proper intersecting X for the úisi forfid (27), but we are perhaps wrong to assume that this was the carver's aim. An identical character, consisting of two angled non-intersecting strokes, occurs at Formaston and possibly at Lunnasting, in the former case the two halves are even further apart than at Burrian. The characters -X-and > W. crwys, OC. crois, MB. croes, croas [Jackson 1953:535 §125]. In Irish c/ragave cross [Thurneysen 1946:575]. The Burrian spelling appears to indicate a stage when the /ks/ of the Latin was preserved, with the double C perhaps indicating a non-lenited sound, /k/ as opposed to l%l. The E in the first syllable looks like an epenthetic vowel, which as Padel points out, might arise in a borrowing if the target language didn't have initial /kr/, but this is puzzling since initial /kr/ poses no problems in Celtic. An alternative explanation is that, though 21 is the úrstforfld, it has, in this case, die consonantal value K = l%l. This would give die reading /kxroks/ which is no less inexplicable, but if the initial C were, in fact, the final letter of the previous word (the unintelligible KEVV), then the final word would read /xroks/ (i.e. /kroks/ with lenited initial). If the basic form is /kroks/, then we have a perfectly regular *CRROCCS /kroks/ from cntx /kruks/. This leaves us to explain the third word and why it would cause lenition, or rather, since we appear to be dealing with a text in a Brittonic language, spirant mutation (which entails the same initial mutation, /k/ > /%/, but for entirely different phonological reasons) [Jackson 1953:634-8]. In British, proclitics ending in a consonants gave rise, in close speech groups, to an external sandhi known as 'gemination1. In Neo-Brittonic the result of such gemination was that in the second word initial p-, t-, c- became /f/, /6/, /%/ (spelled in f, th, £h manuscripts). By analogy, spirant mutation spread to other contexts, and conversely was lost in some phonologically appropriate contexts. In Middle Welsh spirant mutation occurs only in a few specific grammatical contexts [Evans 1964:21 §24], but as Jackson comments 'there is considerable disagreement between the three languages [WCB] and it is obvious that there was much levelling out by analogy and other causes working in different ways' [1953:63].


BURRIAN

202

The Pictish word for cross may be preserved in the place-name Applecross, Apur-Crossan [ATig 722] since the first element is clearly Brittonic, i.e. Pictish. It is quite conceivable however that Adomnán wrote the name as it was known to Irish-speakers, who were able to substitute the Irish form of the very similar second element, but not the less familiar first. The word 'cross' seems to appear in another Scottish ogham, that of Bressay, though in this case the spelling, CROSSCC is very close to the Scottish Gaelic form crosg. It may be relevant to note that one of the Scottish runic inscriptions inscribed on an ultimately Celtic-inspired cross-slab at Kilbar, Barra, has a metathesized spelling of 'cross' - kors [Barnes SC 8].

The other word which may be readily identifiable is the second, URRACT. Unfonunately the initial letter of this word is damaged and doubtful. If it is indeed to be taken as UU or U then we may have here a most interesting form. John Koch suggests to me that URRACT may represent the Pictish cognate of the Old Welsh 3sg.preterite *guract 'he/she made', cf. gwreith 'I made' [Williams 1938:44.1102 (Canu Aneirin - Peis Dinogat)]. In MW the form was gwnaeth since r had become n here by analogy, though it was preserved in Cornish and Middle Breton cf. C. gwmf MB. groqff 'I do' [Evans 1989:130-2; Thurneysen 1946:111].

If the above interpretation of URRACT and CERROCCS is accepted, then the Burrian text appears to represent a sentence - 'blank he/she made blank cross'. In which case the first 'blank' is probably the name of the person who made the cross. Given that the text appears to be written in Brittonic, we might expect a Brittonic name. In addition to Birran, the Book of Llandaff, for instance, includes a number of compound names with second element bran 'raven' (Conbran/ Cinbran cf. Irish Branchu, Dibran/Dofran/Dubron, Gabran, Loitbran, Morbran, Davies 1979]. Uhlich discussed Irish compound names incorporating the element bran [1993:184-6], and there are many more, both male and female, which end in -ran(n) (e.g. Morann (m), Eórann (f) and Cochrann (f), also potentially relevant are the male Odrán, Ailerán, Sárán, Ciarán, Barrán, Láarári) [Ó


BURRIAN

203

Corráin & Maguire 1990:sw.]. The Burrian text is too corrupt for certainty, but the Irish male Idbrann [CGH LL 335 b 37] is a possible contender.

This leaves the third word (.)EVV or (.)EVVC. The doubt over the value of the initial letter is a serious problem. It is usually interpreted as the x-forfid and given the consonantal value K = /g/, but its resemblance to thisforfid is not overwhelming. Ferguson carried out an inpossibly far-fetched inversion to arrive at THETTS which he translated as 'this', which may be the right interpretation for the wrong reasons [1887:136], Comparison has been made with the HCCVVEVV of Lunnasting, which is clearly delimited by word-dividing dots, interestingly it is also the third of four words in a text. The correspondence, which is far from exact, may be purely coincidental since the personal names appear to be in different positions and Lunnasting is not a cross-slab. To accept it would rule out the word division (,)EVVC KRROCCS suggested above. Substituting the alternative interpretation of 16 produces MEW or MEVVC, neither of which is immediately suggestive of anything.

Notwithstanding the uncertainty over the third word, the above interpretation yields a sentence of SVO structure which is at variance with the expect Celtic word order VSO. The syntax of the verbal sentence in Welsh is an involved topic, complicated by the so-called 'abnormal order* of MW prose and the great variety of word order in early poetry (including SVO) [Evans 1989:179-81]. The value of the evidence of the Burrian sentence is compromised, however, by its epigraphic context. The general prominence given to personal names in inscriptions may be sufficient to account for its fronting here.

DISCUSSION At a total of twenty-six characters Burrian is a medium length text by Scottish ogham standards, its significance lies in its constituting the only extant sentence in Pictish. Not only is URRACT the only example of a Pictish verb, but in being so close to its Old Welsh cognate it provides important further evidence for the fundamentally Brittonic nature of Pictish. The text demonstrates the use of ogham to write


BURRI A N

204

a language other than Irish, not merely non-Irish personal names, but non-onomastic parts of speech. From this we can infer the adoption of ogham by the natives of Orkney, not merely its use within a circle of Irish immigrants. In being other than a commemorative inscription the Burrian text expands the range of known purposes of epigraphic texts in Scotland, in this case to record the name of the person who 'made' the cross. We should perhaps bear it in mind when we come to interpret otherwise unqualified personal names on cross-slabs before we rush to assume that they are commemorative.

As discussed above, the informality of the physical appearance of the inscription may be taken as evidence for easy familiarity with ogham on the part of the carver. It is more difficult to evaluate the implications this might have for the audience of the text. The Burrian ogham is carved on such a tiny scale that it can only be read close-up, unlike, say, the much more obviously monumental Brandsbutt ogham, which could be read at a distance of several metres or even yards. If the Burrian cross was an item of church furniture it is possible that individuals knelt before it for meditation and prayer. The oghamist may have recorded his name in the hope that he would be remembered in their prayers, though he makes no explicit request. The formula of the craftsman's request for prayers is common enough in the roman alphabet inscriptions of Ireland, a famous example being the or do Maél-ísu Mac Bratdan U Echan dorigni in gres-sa, 'a prayer for MaeMosa mac Bratdan O Echan, who made this handiwork', on one arm of the early twelfth century Cross of Cong [CIIC 552].

BIBLIOGRAPHY

OS Record card - HY 75 SE 3

Brash 1879:362 pLXLK; Southesk 1884:200; Ferguson 1887:135-7; Traill 1890:346, pl.46 (excellent); NMS 1892:231-3; Rh?s 1892:292-4; Nicholson 1896:27-8, App.15-7,72-5; Rh^s 1898:372-3; ECMS24


BURRIAN

205

fig.20; Macalister 1940:209-10, fig.6; Padel 1972:75-9; MacGregor 1976:96-7, fig.21; Sims-Williams 1992:48, 72.

Site:

RCAHMS 1946:45-7 No. 192, 51 No.201

PREVIOUS READINGS

Ferguson

-RRA(n/g)(n/g) U(u)RRACT THETTS CCRROCCS

Brash

IALELRARBANN(u/ng)£ieRRACT(ea/p)EFFX(mm/aa)RROCCS [Traill 1890:350-1]

Southesk

NAALLVORARANNIfliRRACTMHEFFCMESOCCS

Rhys

-VORRANNV(u/e)RRACTPEVVCERROCCS

ECMS

UORRANN UURRACT PEW CERROCCS

Macalister

IOLIRRBANN UTRRACT KEVV CEROCCS

Padel

IdbmlRRhANNuRRACTKEVVCERROCCS

) , bound at both ends. The lower bind-stroke clearly continues past the second stroke indicating that the letter originally contained at least three strokes - U, E or L

This yields a reading:

-]I( /d/h)R(u/e/i)[-


CUNNINGSBURGH 1

212

FORM OF SCRIPT The slight gap between the I and R is unlikely to mark word-division because where indicated, this is shown with pairs of dots on either side of the stem, not by spacing of letters. Angled vowels occur on Birsay 1, Burrian, Cunningsburgh 3, Formaston, Lunnasting, and possibly on Golspie. It is not clear whether the contrast between 'straight* and 'angled* vowels is a purely stylistic variation in script, or whether it was intended to differentiate between vowel sounds. Since both commonly appear in the same inscription, often along with further supplementary vowel letters, it may well be the latter (see Lunnasting for discussion of angled vowels).

INTERPRETATION Since so little survives, and from the middle of a possibly extensive text, it is difficult to extract linguistic information from this inscription. There is nothing about the sequence which requires or precludes interpretation as Celtic.

BIBLIOGRAPHY Goudie 1878:20; Brash 1879:366; Southesk 1884:206; Rh?s 1892:294-5; 1898:376; ECMS 17fig.lla,b; Goudie 1904:53; Nicholson 1896:30,57,85 N; Macalister 1940:215, pi.vie; RCAHMS 1946 No. 1136(1); Padel 1972:79-81.

PREVIOUS READINGS Southesk

IR-

ECMS

(e/i)R-

Macalister

IRu

Padel

IRu+


CUNNINGSBURGH 2

213

OGHAM-INSCRIBED FRAGMENTS - CUNNINGSBURGH 2

DESCRIPTION OF OBJECT A triangular fragment of grey sandstone, carved with two incomplete lines of ogham, across the face and along the arris. Stone:

Sandstone

Dimensions:

300mm max. height x 190mm max. width x 65mm thick

Condition:

Severely trimmed. Ogham on face well-preserved, but arris badly damaged.

The Cunningsburgh 2 ogham is more substantial than a piece of graffiti, but the original form of the monument is unclear. Since three of its sides, at least, must have been visible, it might have lain horizontally as a cover, or stood venically, either free-standing or as an end-slab in a composite construction. If the slab had stood upright then the two oghams would have read vertically upwards, as one would expect, but the upper surface could not then have been horizontal.

DESCRIPTION OF CARVING The only carving on the surviving portion consists of two lines of ogham letters, one written on the arris, the other across the broad face with a drawn-in stem line. On the surface, near thefracturedinside edge, there is the odd score which might be the remains of a third row of letters, but could as easily be casual damage. The short narrow face is evidently intact since the lettering continues over the edge and down the sloping side. Both sections are written in the same hand and doubtless form a single text. The slope of the consonants indicates that both are to be read in the same direction, but whether they are to be taken separately or sequentially, and if the latter, in which order, is hard to determine. It is impossible to estimate how much has been lost beyond the fracture. Reading from the fracture to the intact edge, the inscription is as follows [see fig. for key to numbering]:


CUNNINGSBURGH 2

214

Angle: 1

Two short strokes to the left of the stem. A third stroke would have occurred at the point of fracture - D, or any subsequent letter of the h-aicme (T, C or Q).

2

One short vowel stroke across the arris (A), with perhaps another two (U). Rh?s, Allen, and Macalister all read E, but four notches would be very cramped. Padel thought two (O), but was prepared to go up to three (U).

3

Three strokes to the right of the stem, sloping markedly forward, clear - V. The upper surface of the b-side abutting the arris is lost from this point to the first stroke of 6. Sufficient remains opposite 4 and 5 to indicate that no strokes have been lost there, but the missing portion is larger between 3 and 4. On the h- (i.e. narrow) side there follows a gap sufficient for a few strokes, possibly vowels. Padel thought he saw 'part of a single h-score'. Allen suggested O.

4

Two strokes to the left of the stem, sloping backwards - D. The distal tips of this pair of strokes are joined by a curved stroke, giving a 'hair-pin' impression.

5

Two short strokes to the left of, and perpendicular to, the stem - D. Slightly shorter than 4, though similarly joined at their distal tips. No doubt the difference in gradient is intended to distinguish between two successive letters of the same aicme.

6

Five long oblique strokes across the stem, sloping forwards. The portion of the first stroke to the right of the stem is lost but not in doubt - R.

7

Four short strokes to the right of the stem. The last two are badly spalled and only their tips survive - S. There may have been a fifth stroke - N. Thereafter most of the arris has been lost, but there is room for another two or three strokes. There are hints of a pair of short vowel notches, followed by a possible single vowel notch (OA, or U), but these are doubtful.

This gives a tentative reading: -]d+(o/u/e)V( /o)DDR(s/n)( /oa/u)[-

Face:

A stem-line 250mm long is carved across the broad face of the stone, parallel to the ogham-

inscribed arris. The stem-line continues over the far edge and down the narrow end face. The line is more


CUNNINGSBURGH 2

215

or less straight, but wavers slightly. The carving is generally clear and well preserved, the only doubt comes where the lettering crosses the short arris. There is an extensive spall to the right of the stem taking away the dps of the last few strokes of 1 and continuing to the edge of the first stroke of 4y but it is clear from the configuration of the extant strokes that no carving has been lost. The inscription is as follows [see fig. for key to numbering]: 1

Four strokes perpendicularly across the stem - E.

2

A single oblique stroke to the left of the stem, sloping backwards - H. The section nearest the stem is lost. The slope proves that this is an h-aicme stroke (cf. 3 and 5), not the remains of an M (cf. 8).

3

Three oblique strokes to the left of the stem, sloping backwards slightly - T.

4

Four vowel strokes - E. Macalister has the central pair bound at the right-hand edge. Padel mentions the possibility, but I could see no sign of it.

5

Four oblique strokes to the left of the stem, sloping backwards, roughly parallel with 2 rather than J-C

6

Two vowel-strokes across the stem, apparently bound with a curved 'hair-pin' stroke at the left edge only - 0.

7

Five oblique strokes to the right of the stem, sloping forwards - N.

8

A single long oblique stroke across the stem, sloping forwards - M.

9

Two vowel strokes across the stem, apparently bound with a curved 'hair-pin' stroke at the left edge only - 0.

10

Four or possibly five long oblique strokes across the stem, sloping forwards - S' or R. A stroke is visible to the right of the stem, just before the short arris, but to the left of the stem it appears to run along the arris itself, however the arris is damaged at this point. Macalister thought this final groove was less substantial than the other four and therefore doubtful, but most other authorities have accepted it. Padel mentions the possibility that this letter was 'bound or partially bound', but I could see no sign of it.


11

CUNNINGSBURGH 2

216

Padel saw four strokes to the right of the stem on the end of the stone - S. The slight doubt over the status of the possible fifth stroke of /0, raises the possibility that the final letter could have five strokes (N).

Which gives a reading: -EHTECONMORs

FORM OF SCRIPT The form of the script used is distinctive, consisting of long, thin strokes, quite different from the thick, formal, square-ish lettering of Cunningsburgh 1. The marked distinction in length between shon vowels and long consonants is noteworthy. The component strokes of letters are evenly, though closely, spaced, but the letters are very tightly packed. In a few instances, e.g. face 3/4 or (5/7, adjacent letters are closer than component strokes. The slope of m-aicme consonants is not great, but all b- and h-aicme consonants are sloped also. It is unusual to have vowels doubly distinguished from consonants, by both length and gradient.

The inscription contains noforfeda as such, but there are four instances of an otherwise unattested feature a single curved bind-stroke on the left-hand ends of a pair of strokes, giving a 'hair-pin' appearance. On the face this appears twice for a vowel (6 and 9 = 0 ) , on the arris it twice appears with consonantal force (3 and 4 = D). None of the other letters in the inscription are bound. Since the device is unprecedented there is no way of knowing whether this is merely a scribal variant of an ordinary 0 and D, or whether it is meant to convey some linguistic nuance. The transliteration of the third member of the b-aicme is discussed in the Introduction [see also Sims-Williams 1993:162-70] where it is suggested that it should be transliterated as /f/ in a Gaelic context and /w/ in a Pictish (Brittonic) context. The transliteration of H is discussed in the same entry [see also Sims-Williams loc. cit.], it may indicate a spirant. In its general proportions, the script of Cunningsburgh is comparable to that of Pool, though the resemblance ends there.


CUNNINGSBURGH 2

217

INTERPRETATION OF TEXT The arris portion is too doubtful and fragmentary to permit detailed analysis. The face, however, is more promising. In the Irish annals Latin mors, 'death', is the word which appears to be used for the violent death of churchmen [Anderson 1922:248 n.4]. Though the ogham alphabet is used to write a Latin leonine hexameter in the margin of the Bodleian Library manuscript of the Annals of Innisfallen, fol.40 v [Macalister 1910], Latin here seems rather far-fetched and a vernacular interpretation, either Gaelic or Brittonic, is to be preferred. The final -s causes some problems. It is clear and at the end of the inscription, and cannot be explained away as a segment of a following word now lost (unless the two lines are intended to be read sequentially, the face first, and the division of the two lines has not been guided by word boundaries). As Thomas Clancy has pointed out to me, the recurrence of S in what appears to be syntactically genitive position on a number of Scottish oghams (Altyre, Brandsbutt, Bressay, Gurness, Inchyra, Lunnasting) is striking and problematic, and certainly hard to explain as Celtic* it looks more Germanic. While this might be plausible at sites within a Norse ambit, including Bressay, Lunnasting, and Cunningsburgh, its recurrence in the Pictish 'heartland' makes some other explanation seem more likely. This apparently non-Celtic termination [Padel 1972:85] is all the more puzzling since the preceding letters appear to be the securely Celtic name Conmor, as identified by Rhys [Goudie 1904:54]. Welsh Conmor, attested, for instance, in the Book of Llandaff [Davies 1979], has been derived from Brit. *Cunoburros [Williams 1937:cxvii, Jackson 1953:485 (explaining the confusion between m and b)] t Cuno- being the well-known Celtic male personal name element 'hound' [Jackson 1953:413 §63, Uhlich 1993:209-20], and -burros 'puffed, bloated, big, proud', cf. W. bwr, MIr. borr, Gaulishbor- [Evans 1967:154-6]. The name is not attested in Gaelic. To interpret Cunningsburgh 2 in Irish terms would require taking ) - A.

7

Two strokes below the stem, sloping forward - L.

8

As above. These two letters slope a little more acutely than 2 and 5.

9

Two short vowel strokes perpendicularly across the stem - 0 .

10

Five oblique strokes across the stem, sloping forward - R.

11

As above - R. The b-portion of the second and third strokes is damaged but the letter is not in doubt.

12

This letter consists of two opposed angled vowel strokes (•> < with a pronounced gap is probably a separate character. What it's value might be is not clear, though in each case a vowel is more likely than a consonant, given context in each case. A character ->*>• would be 0 , -x-or >^r would probably be E, so, at a guess->— Nad-Frolch [see McManus's discussion 1991:109-10 §6.15]. The form Nia- also occurs as the first element in a number of Old Irish personal names [CGH s. v.], and when followed by a genitive or adjective Nia occurs frequently uncompounded as a sobriquet or even personal name [DEL]. We are left then, with a number of options. Given what follows, it seems unlikely that Golspie's NIA is the first element of a compound name, we may then be justified in taking it literally to mean 'champion' or 'nephew'.

McManus gives one example of ogham NIOTTA as a formula word

meaning 'nephew', CIIC 252 DUMELIMAQIGLASICONAS NIOTTA COBRANOR[... [110]. That an uncle-nephew relationship should be commemorated on an Irish monument should allay fears that Golspie's text might constitute evidence for Pictish so-called matriliny (though in the case of the latter it would be a grand-uncle-grand-son connection). In discussing the aforementioned Irish ogham, Charles-Edwards [1971:120] and Ó Cathasaigh [1986:144-5] have put forward the interesting theory that GLASICONAS is to be interpreted as cú glas, an Irish legal term for a foreigner, and that, as the son of an immigrant Dumel had to express his kindred membership through his maternal uncle [McManus 1991:110]. If Golspie was erected in a Gaelic-speaking milieu Hallorredd's Brittonic-looking name might indicate that his father was a Pictish-speaking outsider, in which case he might wish to express kin-group affiliation as 'the son of the nephew of X' (for an example of a father with a Goidelic name and his son with a Brittonic name recorded on a bi-lingual ogham-roman monument, see CIIC 488, Tavistock 3, Okasha 1993:278-81, Thomas


GOLSPIE

316

1994:265-6]). There is an alternative explanation. Ma also appears as the second element in the male personal name Macc-Ma. According to the Annals of Ulster (s.a. 1034) MaccNia Ua Uchtcun, lector of Kells, was drowned en route to Scotland. Whatever the explanation, once again, the apparent lack of a genitive is troubling, the expected form would be nioth.

VARRQJRR There is an element of doubt about the final section of the text, and several interpretations are possible. VARRC[.]RR is the most obvious reading, if transliterated FARRCARR which may be a form of the Irish male personal name Ferchar ( < fer 'man' + cha(i)r 'loving* = 'friendly') [Uhlich 1993:242,245 for the separate elements]. This name 'occurs occasionally in the early period' [Ó Corráin & Maguire 1990:96] and was panicularly popular among the Cenél Loairn. Fercar is the name of a number of figures in the genealogies, including kings, such as the famous son of Connad Cerr [see index of Anderson 1922 for further details of Fercars with a Scottish connection]. The Cenél Loairn had increasingly strong connections with Moray, which makes the use of the name by the dynasty particularly pertinent to an inscription just north of the Dornoch Firth. Is there any hope of identifying the Fercar of the inscription with a historical figure ?

The name was popular in the region (in the thirteenth century it was born by Ferchar Maccintsacarit, Earl of Ross, active in 1235 [Anderson 1922.H: 233, 404, 458, 496]) but its use in medieval Scotland was widespread [Dorward 1995:91-2], The name is still in use in the Highlands, Mod.ScG. Fearchar, anglicized Farquar, from which comes the well-known surname Farquarson. According to Watson the disused burial ground opposite Shiel school in Glenshiel ROS is called Cill Fhearchair, though no saint of that name is recorded in the Calendars [1926:304] and it is always possible that the Ferchar commemorated was a layman whose patronage endowed the ecclesiastical site.

Given the meagre

documentary material available, there is little hope of identifying the Ferchar mentioned at Golspie. Though it may be worth noting that a Ferchar m. Feradaich m. Fergusa appears in the Genelach Clainde


Lulaig in the Books of Ballymote and Leinster [CGH 162 e 11].

GOLSPIE

317

He is none other than Fercar Fota

('Fercar the long'), who was king of Lorn and, for a year, of Dál Ríada, and who died in 697 [Anderson 1973:111-2, 179]. He had sons and grandsons who appear to have been kings of Dál Ríada. One of these, Selbach, died in 730 [Anderson 1973], so a son of a nephew might be expected to last into the second half of the eighth century, about the right date (and the right status) to be honored with a monument at Golspie (after his death or during his lifetime after an act of patronage). At what date before the ninth century the Cenél Loaira first began to take an interest in Moray is not known [Crawford 1995:3], nor exactly how far north their influence extended. At the time, 'Caithness' included modern Sutherland, fairly near the southern border of which is Golspie. While we would expect a figure of considerable status to have commissioned the Golspie stone, and the ogham clearly points to ultimately Irish influence, so few Fercars made it into the historical record that it would be unreasonable to seize on one of the few who did purely because he did.

One problem with the interpretation Fercar is, of course, the A rather than E in the initial syllable. The name should appear in the genitive, but since the ante-penultimate letter is lost we cannot be sure if it does or does not. Assuming the missing letter or letters to be a vowel, there is too much room for a simple straight, or hammer-head A, though an angled A would fit in nicely. There would just be room for a straight E, though not enough for the six strokes that would be needed for AI, the anticipated genitive form. If the reading FARRC[.]RR is rejected in favour of something along the lines of FINC[.]RR, then we may have the Irish male personal name Finchar < fine + char 'loving one's fine (kindred)' cf. Venicarus [Uhlich 250]. The problem of the apparent lack of genitive case still applies.

Putting these elements together, how are we to interpret the text ? If the formula is X MAQQ Y, then we day have an Irish text with one Brittonic and one Goidelic personal name ail AloredMac Nia Fercar 'the nibnument of Alored son of the nephew of Fercar'. Though nia does appear as a formula word in one Irish ogham, the formula 'son of the nephew o f is not otherwise attested. Perhaps more likely is that Nia and


GOLSPIE

318

Fercar are pan of an otherwise unattested compound personal name NiarFercar, in which case the text is simply 'the monument of Alored son of NiaFercar'. If Mac and Nia are to be taken together as a personal name then we have a string of three ail Alored MacNia Fercar, and the text reads, either 'the monument of Alored, MacNia, and Fercar* or 'the monument of Alored: MacNia and Fercar*.

The latter

interpretation would alleviate the problem of the lack of the genitive, and calls to mind the three names on the Drosten Stone (St. Vigeans 1 FOR), Drosten/ipe Uoret/ett For/cus [Okasha 1985, Clancy 1993]. A number of Welsh monuments with Latin inscriptions carry single names, e.g. Nash-Williams 1950 Nos. 206, 159, 193, and 194. Others explain that X put up the cross for his soul and the soul of Y (and Z, A, B .„), Nash-Williams Nos. 220, 223, 231, 233. One from Merthyr Mawr, Glamorgan [No. 239] adds the name of the craftsman a me preparatus + Sciloc, and No. 222, from Llanilltud Fawr, has a main text and then four other names each in a separate panel, with no fiinher explanation of the role of the individuals concerned. The three names at Golspie could be three individuals commemorated equally, or the name of the patron (secular), patron (ecclesiastical), and craftsman, as is found on later Irish crosses such as the Cross of the Scriptures at Clonmacnoise [CIIC 849] or the Market Cross at Tuam [CIIC 523].

DISCUSSION Despite the doubt over a few of the letters, and the perennial problem of how to interpret non-standard vowels and the first character of the h-aicme, the Golspie inscription is one of the more legible and penetrable of the Scottish monumental oghams. It is unusual in having no drawn-in stem, and in being placed up the right hand edge, though the inscription on Scoonie, which does have a stem-line is similarly placed. All the ogham-inscribed Class II cross-slabs have the inscription on the reverse (i.e. non-cross) face, with the exception of Latheron which is carved on one face only. In scale and workmanship, Golspie is perhaps most closely comparable with Brodie, in both cases the ogham is separated from the rest of the carving and placed on a special band. There is no reason to doubt that these bands were intended to bear ogham from the outset and that the inscriptions relate to the original purpose of the slab. It is, of course, impossible to rule out the possibility that the Golspie text was added later, though there is no positive


GOLSPIE

319

reason to believe so; it is a monumental inscription, carefully conceived and executed, not a piece of graffiti.

Nothing conclusive may be said about any possible relationship between the inscription and the rest of the carving. Attempts to link the figure of the man to any of the people named in the text appear to me entirely misconceived. It has been argued that Pictish symbol statements are most likely to refer to personal names [Samson 1992, Forsyth 1996], and if there were three pairs to go with the three names, then it might be worth trying to match them up. But it is no more than an assumption that these are bilingual monuments such as occur in Wales and Dumnonia, it could equally be that the ogham is used to express things which could not be rendered within the native Pictish system.

BIBLIOGRAPHY OS Record card - NC 80 SW 13 Dunrobin archives - DUOCM ARC 527 Cordiner 1780:72, 1788:n.p.; Stuart 1856:12 pl.xxxiv; Brash 1879:363; Southesk 1884:193, 1886:22-30; Ferguson 1887:151-3; Rh?s 1892:288-90; Nicholson 1896; Rh?s 1898:368; ECMS 48-50; W. Baiinerman 1908:347-52; RCAHMS 191 la: 101 No.295; Curie 1940:85 pLxxxviii; Macalister 1940:206-9, pl.iv b; K. H. Jackson 1955:140, 141 pi. 12; Padel 1972:93-8; Close-Brooks 1989:14, 16 ill.; Ritchie & Fraser 1994:18 ill.; Nicoll 1995:139 fig.27 (ill. only).


GOLSPIE

PREVIOUS READINGS

Southesk

ALLDALLDQQA(a/r)DDMQQNUUFFHRRIaNN

Rh?s

ALLHHALLORrEDDMa(hc/q)(hc/q)N(o/u)UVv(h/a)Rr(eirf/iann) [='Beast and MacN's conflict']

ECMS

ALLHALLORREDDMa(qq/hchc)NUUVVHRR(e/i)(i/a)(rf/nn)

Macalister

ALLHHALLORR EDDARR NAIVALARRENN

Padel

ALLHALLORREDDM(a/e)QQN(uu/i)V(v?a)LA(hr/n)RE[-](rr/nn)

Close-Brooks

ALLHHALLORREDDMeQQNiVvhrrErr

320

c p

•I 55

i CO

"2

GOLSPIE - Ogham-inscribed cross-slab [Close-Brooks 1989:cover, 9]

GOLSPIE - Ogham-inscribed cross-slab. Side view showing ogham [ECMS 48A]

*^ ^H ~ ^*

1

fe

22 21 20 19

18

Kf 11

GOLSPIE - Ogham inscription (key to numbering)

321

GURNESS

DISCOVERY The Gumess knife was found in 1931 during excavation of the Broch of Gumess at Aikerness on the mainland of Orkney [NGR HY 3818 2685]. There was no exact record made of context or stratigraphy but it is known that the knife came from the post-broch occupation layers in the southern part of the site in the area of the dwellings known as the 'Shamrock' and annexe* [see fig.]. The knife was given a cumulative site number of GA 31.277. Along with all the other finds from the early years of the Gumess excavation the knife was sent to the National Antiquities Museum where it is now on display (NMS Cat. No. GAA). A brief description of the site was published in the form of a Ministry of Works Guide [Richardson 1948], but it was over 50 years before the Gumess finds were properly published [Hedges 1987|. Hedges pieced together his account from contemporary field-notes and records and from an examination of the extant finds. In his catalogue the ogham-inscribed knife is numbered 252 [Hedges 1987.11:96 fig.2.22 & 118 fig.2.44]

SITE The Knowe of Gumess at Aikerness was listed as a broch in Petrie's 1872 list [1890:93 No.26| but it was not explored until 1929. Over the following ten summers the grass-covered mound was excavated, first under the direction of Hewatt Craw, then, following his unexpected death in 1933, by a number of excavators under the nominal direction of J. S. Richardson [see Hedges 1987.11:1-14 for a full account of the excavation). The site is on a promontory on the shore of Eynhallow Sound which divides the Mainland from the island of Rousay. Gumess commands a hinterland of good farmland. On excavation the mound was discovered to have a 900-year history of occupation. A discussion of the broch village is outwith the scope of this entry, but it appears to have been abandoned by about AD 400 [for a brief, well-illustrated survey, see Fojut 1993:8-15]. After an interval of perhaps two centuries, the site was re-occupied and new habitations, in a distinctive style, were set into the rubble of the old structures. This post-broch settlement


GURNESS

322

coincides with the Pictish period [see fig.]. Norse activity at the site is indicated by a tenth century Viking grave.

Gurness is typical in displaying continuity of settlement locus from the Early Iron Age into the Pictish period, and continuity of artefact types between broch and post-broch layers, and between post-broch and Viking. A number of nonetheless distinctively 'Pictish' items were recovered, most noticeably a small slab (0.27 x 0.19 x 0.29m) incised with unusual geometric Pictish symbols [see fig.; Clouston 1937; J. N. G. Ritchie 1969; RCAHMS 1985:14]. In 1967 a small fragment of a second Pictish symbol stone, incised with the mirror symbol, was discovered nearby at the Sands of Evie [HY 3723 2640]. It is now in the Tankeraess House Museum [OS record card HY 32 NE 31; J. N. G. Ritchie 1969]. The lobate form of the post-broch building known as the 'Shamrock' is of a diagnostic type, the closest parallel being the cellular 'figure-of-eight' structure at Buckquoy [Hedges 42]. The 'Shamrock' appears to have been a dwelling of combined stone and turf construction, behind which there was evidence of bronze-working, including moulds for pins and brooches of seventh to eighth century type. Fojut describes it as a lowroofed, 'diminutive building', and emphasizes thefourth-andfifth-centurydecline in die status of the site [1993:10]. The precise nature of the post-broch settlement at Gumess is not clear. A small cross-marked stone implies adherence to Christianity, but it was discovered unstratified and may possibly date to the Norse period [Hedges 306, ill 126 fig 2.52].

In the official publication of the site Hedges was restrained in his criticism of die standard of the excavation, elsewhere described it as 'wincingly bad' [1990:26]. The expressed aims of the excavation were to clear out the ruins, consolidate the structure for display to the public and to recover finds. Very little attention was given to stratigraphy or proper record-keeping and it is clear that aspects of the site were only imperfectly understood at the time. For this reason we have only a general indication of where the knife was found - in the area of the 'Shamrock' and 'Annexe' - though it is clear that it came from the post-broch layers.


GURNESS

323

Hedges suggested that an unstratífied cetacean bone casket lid [Hedges 1987:207 No. 174], and a whetstoneshaped pebble [Hedges 1987:245 No.802] might be ogham-inscribed [43] but this seems unlikely [see Appendix - Dubia]. Furthermore Hedges identifies 'putative runic inscriptions' on two short orthostats [GA 77.568, Hedges 1987:245 No.802], but this identification has yet to be confirmed.

DESCRIPTION OF OBJECT A small iron knife with bone handle, inscribed with two lines of ogham but otherwise undecorated [see fig.]. Material:

Handle - bone (species unspecified). Blade - iron.

Dimensions:

Handle - 72mm long x 15mm wide x 12mm thick. Blade - 88mm long x 18mm wide x 6mm thick (max).

Condition:

Inscription well-preserved and clear except for small portion lost near haft.

Better preserved than the Bac Mhic Connain example, this knife is substantially intact, though the tip of the blade has gone and a chip in the haft has carried off a short section of the clear and otherwise complete ogham. The blade is very corroded and, according to Hedges has 'broken off flush with [the] socket' [1987:213]. Hedges described the handle as a 'slightly tapering flattened cylinder' but in fact it is somewhat triangular in section, with two long curved sides meeting at a slight point, and the third a flat narrow base. The trimming of the base has provided two well-defined ridges, but for whatever reason, these were not utilized as stem-lines. Instead the longer of the two lines of ogham is inscribed across the indistinct ridge between the two broad curved faces; and the shorter in an unusual arrangement across the flat base. Thus the two parts of the text are on opposite edges separated by the broader flanks. When the knife is held as if about to make a cut (with the width of the blade vertical), the longer of the two lines is uppermost and clearly visible.


GURNESS

324

CARVING TECHNIQUE Like that of Bac Mhic Connain, the Gurness ogham is carved on a minute scale. The longest strokes are a mere 10mm in length, and most are only a few milimetres long. They have been lightly cut with a sharp blade which has produced a fine V-section score. Several individual letter-strokes have been cut and re-cut, especially on theflattersurface of the longer text. Macalister suggests the knife is a palimpsest, but I see no reason to believe this. The re-cut strokes are clearly intended merely to improve the definition of strokes imperfectly cut with the first strike. Despite this occasional re-cutting of strokes the inscription is fairly legible. The strokes are evenly and closely spaced and of uniform length. There are clear gaps between letters and the over-all impression is neat. The cramped spacing of some strokes and the apparent need to run-on to a second line suggest a lack of planning.

DESCRIPTION OF OGHAM INSCRIPTION Though individual strokes are clear, there is doubt over certain letters because of the cramped spacing, and in some cases overlapping, of strokes. A further difficulty is in ascertaining the direction of reading of the two pans of the text and their relationship to one another. The longer section covers the full length of the blade and consists of twelve letters, with a possible thirteenth lost with the chip. There is no slope or distinctive letter-form to indicate the direction of reading, though the slight decrease in spacing towards the blade may indicate it was carved, and read, in that direction, as with the knife from Bac Mhic Connain. The shorter section opposite, consists of a further four letters squashed in hard against the socket. These are in the same 'hand' as the longer text and it has been assumed, because of their position, that they represent a 'run-on' line. The question remains, do they read in the same direction as the longer text (towards the blade) or boustrophedon (away from the blade). The latter is unlikely since boustrophedon in ogham is found only on classic Irish pillar stones when the ogham line is continuous and runs up-acrossdown the stone. At Gurness there are two quite separate lines so I think it more likely that they read in parallel. For what it is worth, a more pronounceable sequence is retrieved if the shoner text is read towards the blade. More importantly, it would be hard to account for the over-lapping of these four


GURNESS

325

letters, presumably to conserve space, were they not being carved towards the blade.

In what follows I take the longer text first and read from the butt towards the socket. All strokes, unless otherwise stated, are perpendicular to the stem [see tig. for key to numbering]. 1

Five short vowel-strokes on the curve - L These cover approximately a sixth of the width of the ogham band.

2

Five long strokes to the right of the stem ~ N.

3

Four strokes on or above the stem - E or C. It is not clear whether this letter is a member of the a- or the h-aicme since the individual strokes are longer than the vowels / and 4 but not as long as the consonants 2, 5 or 7. They are closer in length to the vowels and at a glance appear to belong to this group. A closer examination reveals that their proximal ends are on the stem, i.e. at the mid-point of the other vowel strokes, but this observation is perhaps reading too much into a cursively carved text

All previous authorities have taken this group as E and this is the

preferred reading. If not merely accidental, the disparity in the length of consecutive vowels may be deliberate, intended to clearly differentiate the two groups. 4

Five very short strokes on the curve - 1 .

5

Three long strokes to die left of the stem - T. Padel notices 'little false starts' on die first two strokes but concludes that these are too 'cramped and haphazard* to be actual letter scores.

6

A single very short stroke on the curve, If this nick is intended as a letter it is A. It has not been mentioned by any previous authority and may be no more than casual damage. Letters 5 and 7 would in any case have to be generously spaced to avoid conflation, so die gap need not indicate a missing stroke.

7

Three long strokes to the left of the stem - T. As Macalister and Padel point out, the third stroke has been re-cut.

8

Four short strokes on the curve - E.

9

A single very long oblique stroke sloping forward across the stem - M. From here to the end of


GURNESS

326

the line there is some doubt over the reading since certain strokes appear to overlap. 10

Two short strokes on die curve - O. The first stroke has been duplicated in re-cutting but the two attempts are far too close together to be taken as separate strokes.

11

A single stroke across the stem sloping forward - M. This stroke has its top end level with the previous stroke and its lower end flush with the next stroke, making it shorter by a third than the other M stroke, letter 9. Macalister had difficulty explaining this 'long score' and thought it possibly the second score of an earlier D, the first score being partially effaced. I can see no evidence in support of his palimpsest theory. Padel acknowledged the doubling of the first stroke of 70, but preferred to interpret the 'long score' as two separate strokes, one a third vowel stroke (possibly followed by a fourth), the other a sixth (mistaken) b-stroke, giving a reading EN. His doubt over the number of vowel strokes is demonstrated by his alternative readings UN and ON. There is no obvious solution to this problem. The 'long score' is continuous. It is clearly not itself a vowel, as indicated both by its length and its slight forward slope, which is demonstrable, though not as marked as 9 r s. It is clearly spaced from 12 which in any case has a full compliment of five strokes. If Padel is correct and 11 is a conflation of a vowel stroke and a b-stroke then the reading is more likely to be UB, which is almost as unhelpful for an intelligible reading as the most obvious interpretation - M. It is possible that the stroke is simply a carver's error, an elongated vowel. After a gap the size of those between the preceding strokes there is the faint trace of a line. Padel suggests it may be another vowel stroke, though it seems too insubstantial to be admitted, and, since it joins the stem between the first and second strokes of die next letter, too open to confusion.

12

Five long strokes to die right of the stem - N. Macalister alleged that five vowel points (i.e. I) were 'clearly to be traced' at the bottom ends of the scores of this letter. I could see no trace of them. So, the various possible readings of 10, 77, and 12 are, in order of increasing alleged incompetence on the part of die carver - OMN, OBN, UN, ON.

The fifth stroke is partly lost to the fracture in the surface of the bone. The portion which is missing


GURNESS

327

would have been sufficient for another letter, but too much of the surface is lost to ascenain to which aicme it would have belonged. If the shorter section is indeed a run-on of this line then it might be assumed that maximum use would have been made of this last bit of space, in which case we should indeed posit a missing letter. But if letters 13 -16 constitute a separate word, the oghamist may have preferred to leave the space unused and, since it was by now clear that he or she had to start a second line in any case, to begin it with the first letter of the next word. Padel suggested that letters 10-12 were carved overlapping because the oghamist was suddenly aware of the approaching end of the available carving area. It must surely have been apparent to the carver that it would be impossible to fit letters 13-16 in the remaining space. A possible explanation might be that the carver wanted the start of the new line to coincide with a word-boundary and was thus attempting to squeeze in 10-12. It is not certain, however, that these letters do indeed overlap.

The shorter line of ogham is squeezed into the closing millimetres of the handle with considerable overlap of the individual letters. These remain distinguishable because of the unusual three-surface arrangement across the breadth of the flat base. It is unclear why the letters should have been so cramped when the whole of that side was available. If the oghamist did not even allow sufficient room for the final four letters it does indeed imply, as Padel said 'a total lack of planning'. Rather than be written on two surfaces meeting in a notional point (the arris/stem), this short section is written across three surfaces, the two sides being the h- and b- surfaces respectively, and the broad flattened base being a band of 'stem' area. This makes it possible for letters 75 and 16 to over-lap by two and a half-strokes worth without confusion. The vowel-stroke 14 occupies the whole of the width of the base. The only irregularity in these scheme is that the m-stroke, which by analogy would be expected to begin on the b-surface flush with the distal ends of 76, cross the base, and continue on the h-surface to a point level with the distal ends of 75, in fact covers only two-thirds of this distance beginning at the lower ridge of the base (as if it were taking as the stem the arris between the h-surface and base). 13

A single long oblique stroke sloping forwards across the stem - M.


14

GURNESS

328

A single short stroke across the bottom edge - A. Padel saw a possible, very feint, second vowel stroke giving 0, but I could not see it. The A is entirely beneath the overhang of the preceding M.

15

Three strokes to the left of the stem - T. The last two strokes overlap with the first two strokes of 16. Padel points out that the first stroke of the T has been re-cut.

16

Four strokes to the right of the stem - S.

This gives a reading of: IN(e/c)IT( /a)TEM(o/om/ob/u/e/?)N( /.)MATS

or

IN(e/c)IT( /a)TEM(o/om/ob/u/e/?)N( /.)CFAM or, in the unlikely event that the two sections were read from blade to butt: CFAM( /.)Q(o/om/ob/u/e/?)MEF( /a)FI(e/s)QI

or

MATS( /.)Q(o/om/ob/u/e/?)MEF( /a)FI(e/s)QI

FORM SCRIPT No stem-line is indicated, instead the uneven ridges of the handle are used as an arris. The letter groups are differentiated by position relative to the hypothetical stem-line and by length.

All but the m-aicme

are perpendicular to the stem. The vowels are very shon strokes across the stem, the m-aicme consonants are particularly long. Thus the Gumess knife is inscribed in a form of the script which differs in nothing but scale from the Type I script of the classic ogham pillars of Munster. Among the Scottish oghams, the script of Gumess is most closely similar to that of Gigha and Poltalloch. It differs from Bac Mhic Connain in having shorter vowels and sloping m-aicme consonants, furthermore Bac Mhic Connain is carved in a far finer 'hand* than that of Gumess, or at least with a far finer blade. Gumess is unlike any of the other Orkney oghams and on typological grounds, if on no other, may be the earliest of the group. It is a moot point whether or not the simple form of the script is relevant for dating. There is evidence that simpler forms continued in use long after the introduction of more developed forms so simplicity may be more a function of the humble register of the inscription than anything else. In this regard the lack of a stem is


GURNESS

329

perhaps not very significant being dependent instead on the possibility of using a pre-existing ridge in the bone.

INTERPRETATION OF TEXT The question of the direction of reading has already been discussed. Two remaining issues are the order in which the two sections are to be taken and where word division, if any, occurs. The two most plausible readings are INEITaTEMO(m/b)N.MATS and INCITaTEMO(m/b)N.MATS.

The latter might be

segmented INCITATE MONMATS, with incitate the Latin adverb Vehemently', 'rapidly'. There should be no particular objection to the writing of Latin in the ogham script if, as appears likely from the Scottish evidence, the alphabet could be applied to the Pictish or Norse langauges. There is in fact a fourteenth century example of Latin written in ogham characters in the margin of a text bound into the Book of Fermoy [Macalister 1937:60]. Semantics is more of a problem, since incitatus is used primarily of speech, there is no obvious explanation of MONMATS, or parallel for an epigraphic message of this son. The preferred reading is, in any case, INEITaTEMO(m/b)N.MATS, and the rest of the discussion will focus on this interpretation.

When faced with an intractable text it is of assistance to state our expectations of the probability of various explanations. The text is most likely to be written in some form of the Irish language, or failing that in the local language of Orkney in the period. The latter was presumably Brittonic, but may have been partly or wholly pre-Celtic. Norse may be ruled out on chronological grounds. A text of this length may be expected to be a personal name, either with or without filiation, referring to the owner, maker or donor. Or alternatively, an invocation to the object to perform its task well, or to the owner (as, presumably, at Buckquoy). Other possibilities are talismanic texts or playful ones, such as the Dublin runic 'Hart's horn' on a piece of deer antler [Barnes IR 12]. Our working assumption is that we begin by trying to explain the text in Celtic terms and move on to other possibilities only if this attempt fails.


GURNESS

330

Working backwards we encounter an immediate problem - what is the explanation of the final -S ? Since it follows a T it cannot be an old Celtic genitive ending and there is no other obvious explanation for an S in such a position, unless it is an otherwise unattested reduced form of the Irish enclitic demonstrative panicles -so, and -sa. This explanation would imply that MAT is an indefinite genitive or adjective qualifying a previous substantive [Thumeysen 1946:299 §475]. It would be possible to interpret MAT as an adjective meaning 'good' (Olr. maith). The DIL lists two words mat: one a possibly archaic word for 'hand', the other an archaic and poetic word for 'pig' [s.w.]. Since the knife is certainly a small handknife and might perhaps be made of pig bone neither explanation is perhaps as outlandish as at first sight. Classical latin matara is listed as a Celtic javelin, pike or lance. If there is a Celtic word behind this form it is possible that its reflex is found in Pictish, meaning 'weapon' or 'knife'. Another possibility for MAT is some element of a personal name. In his dicussion of the Irish male personal name Matgen 'well/noblebora' or 'born of a bear', Uhlich gives a number of Old and Neo-Celtic examples of compounds with mat 'bear' or mat 'good' [1993:276-7]. Further Irish comparanda are Mát(a) and Mátach [CGH:22 = 118 b 6; 120 = 136 a 26; 376 = 324 e 5], from Welsh there is, for instance, Matoc and Mathus in the Book of Llandaff [Davies 1979:76b 203b 144; 143] and the famous Math of the Fourth Branch of the Mabinogi.

If MATS is indeed a separate word, then we are left with INEITTEMON- or INEITATEMON- or INEITTEMOBN- or similar (cf. Inchyra's INEHHETE-). Whether this is to be further subdivided, IN EITTEMON- or IN EITTE MON- for instance, is unclear. EITTE is reminiscent of EHTE and ETTE of Cunningsburgh 1 and 2 respectively, perhaps representing different spellings of the same word, or different stages in its phonological development, cf. the spellings of Nectan, Neitano, New. This would leave IN which might be interpreted as the Olr. definite article. Alternatively, in Olr. the prefix inis used to form adjectives with meaning 'capable of, fit for, proper for worthy of [DIL]. So, an entirely different approach would be to read IN-EITTEM ON-, taking in-eittem as a form of the Olr. adjective in-ethaim 'suitable for tillage' [DIL s.v.], a not inappropriate pronouncement on a domestic or agricultural


GURNESS

331

implement. None of the above suggestions are in any way compelling and we may have to admit defeat for the time being, in the face of a doubtful reading and the lack of any clear knowledge of the kind of text to expect.

DISCUSSION Obviously Gumess invites parallels with the other ogham-inscribed knife handles,fromBac Mhic Connain, North Uist and Weeting in Norfolk. Like the others it is an informal personal text 'almost certainly ... made during the functional life of the knife* [Holder 1990:57]. The context, date, and meaning of the Weeting inscription are a puzzle, but the Bac Mhic Connain one is slightly better understood. As far as can be discerned it appears to bear a personal name, presumably the owner, but possibly the maker or donor. The Gumess text is a few letters longer though what it says remains unclear. Until the Gumess text is understood and interpreted there is no possibility of identifying chronologically diagnostic linguistic features. Script typology points to an earlier rather than a later date, but as discussed above, is not a necessarily reliable method. Hedges dated the Gumess knife to the eighth century [1987.111:43] but only on the basis of the received wisdom on the dating of Pictish ogham. The discovery of oghamsfromPool and Birsay in radio-carbon dated contexts in the seventh and sixth centuries mean there is no reason that the Gumess knife should not be as old, if not older. The post-broch occupation at the site has been broadly dated to the fifth to eighth centuries, but on no particularly sound evidence other than a terminus ante quern for the Pictish layer provided by a Viking burial. The Gumess ogham wasfoundin very similar circumstances to those of the Buckquoy whorl - lying abandoned in or around a cellular dwelling. The Buckquoy ogham seems to datefromaround 750 or earlier though it is written in a more developed form of the script. In the absence of definitive linguistic evidence to the contrary there is nothing to rule out a date in the fifth century, so while a vague ascription to the seventh or eighth century might be more broadly acceptable a much earlier date should not be dismissed out-of-hand. Whatever its precise date, the Gumess ogham is further important evidence of the demotic use of ogham in Orkney in the Pictish period.


GURNESS

332

BIBLIOGRAPHY

OS Record card - HY 32 NE 5

Macalister 1940:218-9 pl.IVc; Padel 1972:98-100; Hedges 1987:213, Cat.No.252, ills. 96 fig.2.22, 118 fig. 2.44; A. Ritchie 1987:48 ill.; Holder 1990:52-58; Fojut 1993:11 (ill. only).

Site:

RCAHMS 1946.11:75-9 No.263; Richardson 1948; Childe 1961:46-7, 106 ill.; Hedges 1987; A. Ritchie 1987:45-48

PREVIOUS READINGS Macalister

INEITTEMON MATS-

Padel

INEITTEMEN (M(a/o)TS-) / (CV(a/o)M)

Holder

(i/u)(n/s)EITTEMON MATS (not based on personal inspection)

u X a

•'

18 24 »

44 f»t

Pitm «f post laoch-pértcxf struettm. 0 1

4

U „ ,>, - J

•

I

« 11 m * «

k»j—i—t

GURNESS - Plan of the Post-Broch period structures [Hedges 1987.11:64 fig.2.11]

GURNESS - Plan of the "Shamrock" and "Annexe" [Hedges 1987.11:66 fig.2.12]

GURNESS Carved stones recovered from the site. Above: Pictish symbol stone. Below: Cross-incised slab, unknown date. [Hedges 1987.11:125 fig.151, 126 fig.2.52]

306 (P8j

0 4 8

cms

GURNESS - Ogham-inscribed knife handle [Hedges 1987.11:96 figs. 2.22, 118 fig.2.44, details]

12 11 10 9 8 7 6 5

3 2 1

Top

Underside

GURNESS - Ogham inscription (key to numbering)

333

INCHYRA

DISCOVERY The Inchyra slab was found on 19 February 1945 during ploughing in former parkland on the Carse of Gowrie, 120m south of Inchyra House, in the parish of St. Madoes, Perthshire (NGR NO 1904 2120) [see map]. For some years previously, the field in question had been under pasture and was turned over to crops only towards the end of die War; it was deeper ploughing in the spring of 1945 which revealed the stone [Wainwright 1961:3 n.2]. On striking the stone, the farmer alened the proprietor Harold de Pass and the then Curator of Perth An Gallery and Museum, James Wood, who came to examine the site. Several weeks later 'when all had been levelled and harrowed' [Stevenson 1959:33], Thomas McLaren, the Burgh Surveyor of Penh, examined the stone and made a record of the discovery. This he read as a communication to the Society of Antiquaries of Scotland in die January of thefollowingyear [n.a. 1946]. He died, however, eighteen months later, before publishing a full account. Stevenson's account of the discovery [1959:33-4], which is followed below, is based on McLaren's typescript notes.

Unfortunately the stone was removed before the nature of the underlying finds was recognized, and it is a great pity that the site was not recorded to modern archaeological standards. Despite the incomplete nature of the record it is clear the slab was covering an inhumation burial. The slab, oriented WNW-ESE, was lying flat about a foot (0.3m) below ground level when found, resting on fony-nine water-rolled stones. These stones, which ranged in size from c.200 x 180 x 150mm to c.300 x 230 140mm, were 'above and around' some human bone: 'the upper pan of a skull (at the west end), an arm-bone and shoulder socket... identified by Mr Wood and then re-interred on the spot at the wish of the proprietor' [Stevenson 1959:34 n.l]. It is not clear which face was uppermost when die slab was uncovered. Stevenson suggests that if some 'erratic scorings* on the rougher of the two broad faces are plough-scrapes then this may have been the upper face. Nothing is now visible at the find-site. The Harold de Pass presented the stone to Perth Museum where it is now on display (Ace. No. 5/45).


INCHYRA

334

Wainwright refers to 'details collected by McLaren from persons present at the time of the discovery [which] strongly suggest that the sandstone slab was not decorated to accompany the human remains with which it was found' [1961:3]. What these details were is not elaborated. At the time the grave was discovered, virtually nothing was known about Pictish burial practice. Wainwright complained in 1952 that 'we cannot with confidence affix the label 'Pictish' to ... a single burial' [1955b:87]. In particular there were no certain examples of symbol stones associated with Pictish graves [op. cit. 96]. Our understanding of Pictish mortuary ritual was transformed in the late 1970s when a distinctively Pictish type of burial monument was first identified [Ashmore 1980] and connections between symbol stones and Pictish graves were first proved by archaeological excavation [Close-Brooks 1980]. If this knowledge had been available in the 1940s interpretation of the find might have been very different. McLaren interpreted the burial as Bronze Age and thus, in his view, the association with the Pictish stone had to have been secondary. Observing that 'the stone had obviously stood erect above the ground for some considerable time before it was buried in a horizontal position over the human remains', Wainwright concluded that the slab had been 'laid over them at a time when its original purpose had been forgotten, perhaps when the burial ... was accidentally disturbed in a more recent century' [ibid.]. This seems rather implausible. More likely is that the slab stood upright beside the grave, fell over in antiquity and subsequently became buried under accumulated deposits; or, alternatively, that having served as an upright monument for a number of years the slab was reused later in the Pictish period as a cover for the grave. It is clear from the extant carving that the stone has been reused on at least two and probably three occasions. It is unfortunate that the prevailing opinion that 'neither the nature of the burial nor its precise location are directly relevant to the problem of the stone and its original purpose' [ibid.] encouraged its discovers to neglect to record more fully die circumstances of its discovery.

When the stone was first discovered only the two bolder sets of symbols were noticed and the larger of the ogham inscriptions (B). Following McLaren's presentation to the Society of Antiquaries, Dr James Richardson re-examined the slab and recognized the third set of symbols, and three further lines of ogham.


INCHYRA

335

It was he who realized the relationship to the main slab of several fragments also discovered at the site, and thus was able to complete further sections of both ogham and symbols. Through the offices of Professor Gordon Childe, McLaren sent R. A. S. Macalister photographs, drawings and rubbings of the stone, and Macalister's letter in response was preserved in McLaren's papers deposited in the archives of the Society of Antiquaries of Scotland [Wainwright 1961:4]. The first full treatment of the stone and its ogham inscription was published by Wainwright in the French journal Ogam [1959], subsequently reprinted as a monograph by Dundee Museums [1961]. R. B. K. Stevenson discussed the stone at some length in an article in PSAS having examined it in the company of Kenneth Jackson whom he reported as being 'in general agreement' with his analysis. No interpretations of the ogham texts were offered by either author. Oliver Padel made a number of linguistic observations on the Inchyra texts in his 1972 dissertation, but offered no coherent interpretation.

LOCALITY The second element in the name Inchyra has not been explained, but the first, 'inch' (innis) refers to a haugh, or low-lying piece of ground by the bank of a river, liable to flooding. This is most appropriate to the site, lying as it does by the north bank of the river Tay. No other archaeological remains are known from the immediate vicinity except for a mound, marked 'Witch Knowe' on sheet 37 of the OS first edition, about 110m west of the burial [Stevenson 1959:33]. It has not been explored and might date to any period. More generally, this important area, in and around the mouth of the Tay, is rich in early medieval remains. There are hints of early connections with Ireland, including the famous reference in the late sixth century Irish poem Amra Choluimb Chille to Columba teaching the tuatha Toi, 'the tribes of the Tay' [Clancy & Markus 1995:104]. Nearby St. Madoes, which has given its name to the parish, may represent a dedication to Maedoc [Simon Taylor pers. comm. pace Watson 1926:327]. There are several saints of this name, but the most prominent is Maedoc of Ferns. As Dr Taylor points out to me, a possible dedication to this major Leinster saint takes on added significance in the light of traditions linking Abernethy with the greatest of Leinster saints, Brigit [see entry on Abernethy]. Abernethy is almost


INCH YRA

336

directly opposite St. Madoes on the south bank of the Tay and Dr Taylor informs me the two were still linked by a ferry as recently as the last century.

DESCRIPTION OF OBJECT A long, thin, tapering, rectangular slab incised on both broad faces, and three narrow faces with three pairs of Pictish symbols and three ogham inscriptions (arranged in a total of five lines). Stone:

Very fine grained local sandstone (mica-rich, containing both biotite and muscovite, according to Dr A. G. MacGregor, Geological Survey [quoted by Stevenson 1959:34])

Dimensions:

Length: 1.6m (5'3") Breadth: narrow end - 0.31m (12"), broad end - 0.43m (V5Hn); Thickness: 70-90mm (234-3V$")

Condition:

Narrow end intact but broad end broken, and a portion missing. Areas of the surface have flaked off, especially along the edges, taking with them pans of the carving, though some of the larger of these fragments have been recovered. The extant carving is generally well preserved. Some sections of ogham have been lost, but only A (ii) is severely damaged.

Richardson noted that one surface was smooth and undulating, die other flaky 'suggesting that it was obtained by cleavagefroma water-worn exposure' [Stevenson 1959:34]. Stevenson points out that all the narrow faces (except the broken one) are flat and smooth, their arris with the smoother broad face being rounded, and that with the rougher face, sharp [ibid.]. The Inchyra slab is entirely atypical of Class I Pictish symbol-inscribed stones in being carved on a very carefully shaped and dressed slab. It is not alone in being reused, there are several examples of that, including Logie Elphinstone 2, Kintore 2 ABD, and St. Peter's, South Ronaldsay ORK. What is exceptional is that it appears to have been reused no less than three times.


INCHYRA

337

The slab is very tall and thin, but not to the extent that it could not have stood upright. Stevenson records that 'part of one side [was] shattered at time of discovery', some of these fragments were, however, recovered, along with a substantial chunk off one edge of the narrow end. There are some horizontal grooves on the surface of the slab, some, including one right across the smooth surface, were probably made when the stone was transported to the museum. Another long groove nearer the broad end 'was unfortunately made to mark the depth the stone was to be set in a wooden stand' [Stevenson 1959:34].

DESCRIPTION OF CARVING - SYMBOLS I will follow Stevenson in designating the smooth surface the 'front', but this is merely conventional and is not meant to imply that the carving on this side is primary. It is also merely conventional to refer, as previous authorities have, to the narrow end as the 'top' and the broken broad end as the 'bottom'. Since the RCAHMS [1994a] illustrate the slab with the narrow end at the bottom, to avoid confusion I will refer instead to the 'narrow' and 'broad' ends. The slab is currently displayed in low lighting on its side in a large perspex box, conditions which made detailed examination of the carving almost impossible. P am, however, most grateful to Dr Michael King of Perth An Gallery and Museum for facilitating my examination of the stone].

'Front'/Smooth face There are two symbol 'statements' on this face of the slab, each at opposite ends of the slab, and each inverted relative to the other. At the narrow end there is one pair oriented with that end as the top, namely a notched double-disc, and a fish. Positioned very close to the top of the slab, they are the most substantial and finely drawn of all the Inchyra symbols. They have been very well carved, 'definite and accurate' to use Stevenson's phrase, using the traditional pock-and-smooth technique. Stevenson thought the smoothing might have been done 'by oblique "chiselling" with the same round-nosed tool' as was used for the pocking [1959:34]. He went on to comment that in the initial pocking 'many strokes landed outside the final quite broad outline. This is particularly clear at the rear of the dorsal fin'. This 'shrapnel' effect


INCHYRA

338

is not at all unusual, however, and there is wide variation in the extent of smoothing on different stones [Gordon 1956].

At the other end of this face, oriented with the broad end as top, is an indistinct group of symbols which were not noticed at first. The lowest is a mirror and, above that, the remains of a 'tuning-fork' symbol. These carvings have been plausibly interpreted as both incomplete and partially defaced, as if, as Wainwright suggests, the carver had a change of hean before completing the design. The carver has begun to pock the lines, getting further with the 'legs' of the tuning-fork and the upper rim of the mirror, than with the handle, as Stevenson noted, but then appears to have attempted to deface them. Stevenson thought this was aimed at making the symbols as inconspicuous as possible before the half-finished slab was re-used (certainly this is more plausible than his alternative theory that the defacement was the result of a deliberate antagonism). A possible reasonforthe abandonment of the enterprise lies in the flaky nature of the stone. Stevenson hypothesized that the slab's surface had begun to scale away under the pressure of carving and that the carver had therefore abandoned the attempt. Later, when the stone was pressed into service anew, the old half-finished symbols were pocked over to minimize their visual impact, and a different carving technique applied to the reverse. The mirror is in the expected bottom position. While unaccompanied mirrors are common, the 'L' shape surviving on one of the flakes from the left of the mirror may, as Stevenson suggests, be the remains of a very incomplete comb. We would expect a second symbol above the mirror to form a primary pair [Forsyth 1996b]. Stevenson thought one 'probable' to the left of the 'tuning-fork', but that area has largely flaked away and certainty is beyond us. Since the broad end of the slab is broken, a portion of the tuning-fork has been lost.

'Back'/Rough face Though this is the rougher of die two broad faces there has, as Stevenson noted, been some preparatory smoothing of the stone before carving. The carving consists of a single pair of symbols at the broken end of the slab oriented with die broad end to the top. The pair comprises a serpent below a fish, the upper


INCHYRA

339

pan off which has been lost at the fracture. Though somewhat less well drawn than the first pair described, these do have the appearance of being finished. Their carving technique, however, is quite exceptional for they have been cut with a chisel or knife, to produce a V-section groove, rather than pocked with a point to produce a U-section channel. They were die only example of symbols cut in this way encountered by Gordon in his survey of Pictish symbol carving technique in the eastern Mainland [1956:42],

Discussion of symbols The six symbols are organized into three pairs: notched-double disc above fish, fish above serpent, and 'tuning-fork' above mirror. The first two are classic 'symbol statements', i.e. a pair of two different symbols arranged vertically. As it stands the third group is exceptional because mirrors appear only in addition to normal symbol statement pairs. Since, however, the top of the slab and some of the surface is missing, and the mirror is in standard bottom position, there is every reason to take the extant carving as the remains of a typical symbol statement 'pair + mirror'.

Elizabeth Alcock lists twelve examples of the fish symbol on Class I monuments [1989:9], Allen has a further five on Class II, making it one of die more common of the 'Second Division' of Pictish symbols. Nonetheless to have two examples on the one monument here at Inchyra is striking, especially since, otherwise, the symbol's distribution is markedly northern [see entry on Ackergill]. The combination double-disc plus fish does not occur elsewhere, though a fish with double-disc-and-Z-Rod appears on the Class I stone from Keith Hall ABD (modified by a mirror and comb). The degree to which the rodless double-disc is to be considered separately from the otherwise identical type with Z-rod is not yet clear. The former variety is much rarer, occurring on only four other Class I stones - Drambuie 1 INV, Fyvie 2, Logie Elphinstone 1, Newton ABD - but on ten Class II - Ulbster CAI, Golspie SUT, Shandwick ROS, Monifieth 1, St. Vigeans 4, Woodwray FOR, Dunfallandy, Fowlis Wester, Meigle 3, 6 PER - and in a number of the Fife caves. The example from Newton is of particular note since it has a distinct notch


INCH YRA

340

taken out of one of its discs, the left hand, in the 'six o'clock' position. Inchyra has an identical notch out of each disc, at 'eleven o-clock' on the left-hand and at 'six o'clock' on the right-hand. Presumably, these notches modify the symbol in some way. Similar notches occur on a handful of other symbols, including the crescent [see Forsyth 1996b].

The serpent without Z-rod is similarly less common than its rodded counterpan, there being only three other examples of Class I (Kirtomy SUT (dubious), Knockando MOR, and Aberlemno 1 FOR) and three of Class II (Ulbster SUT, Glamis 1-2 FOR), excluding all serpents not in proper symbol 'statements'. The combination with fish occurs only at Glamis 2, serpent uppermost, where it is modified by a mirror. Serpents modified by Z-rods tend to be oriented right-left, but unmodified examples tend to be top-down, with the head to the top (Aberlemno being less than vertical, perhaps to fit into the constrained space at the top of the slab). The Inchyra example is conforms to the typical pattern - upper loop to the left, middle to the right and lower to the left with the tail curling round in a spiral - but with the head and neck a little straightened to fit into available space. In common with other rodless serpents, Inchyra's is undecorated, bar its head. The fish in this pair is atypical in having its tail to the right rather than left.

Only the 'legs' of the upper symbol in the third pair survive. These are longer and thinner than would be expected for a 'notched rectangle', and are more plausible interpreted as part of a 'tuning-fork' symbol (what Allen calls 'notched rectangle with curved ends' [ECMS.ii.67] though it is quite different from the 'notched rectangle'). This symbol is seen on a handful of Class I stones with predominantly northern distribution, Craigton 1, Dunrobin, Kintradwell 4 SUT, Roskeen ROS, Kintore 3 ABD, and Abernethy 1 PER, though it is possible that an acephalous symbol from Dairygreen, Collace PER is another 'tuningfork'. The symbol occurs once on Class II, on the cross-slab fragment Tarbet 1 ROS.

RELATIONSHIP OF SYMBOLS AND ORDER OF PHASES The three sets of symbols are clearly the work of at least two, probably three, hands in at least two, and


INCHYRA

341

probably three» different phases. The order and relationship of the different phases remains, however, unproven. The symbols cannot have all been visible at the same time, since if one end was embedded in the ground only the symbols at the other end would have been fully visible, and if the slab was intended to lie flat only the symbols on the upper face could have been seen. Of course, if the slab were intended from the outset to be buried, then it need not have mattered that symbols were obscured. The difference in draughtsmanship and in carving technique, and the fact that one set is inverted relative to the others does, however, suggest that there was more than one phase of carving involved.

If the slab were meant to stand upright it would be more stable with the broader edge at the bottom. This has encouraged commentators to view the narrow edge as the original top, and the fish and double-disc as the primary symbol statement. Wainwright's interpretation of the mirror and tuning-fork as a false start is appealing, and implies that the fish and serpent represent the third and final phase of use. If the slab stood upright the fish and double-disc would be embedded in the ground and thus obscure, and the imperfect symbols on the front would be unobtrusive thanks to their defacement, thus only the fish and serpent would be visible. This scenario would account for the change of technique for the third set, and the use of the rough face only as a last reson (the narrow end of the rough face is even worse than the broad which is an additional reason for not choosing the narrow end). Since the two sets of symbols on either face of the broad end are different we may take it that they relate to separate phases of use otherwise we would expect a completed tuning-fork and mirror in place of the fish and serpent. It is interesting to note that the two sets of symbols at the broad end are to the same scale, larger than the set at the narrow end, and carved at roughly the same height - the serpent's tail is level with the bottom of the mirror handle.

If the narrow end were embedded sufficiently deep in the ground there need not have been a serious problem with stability, but it may well have been that the later phases involved the slab lying horizontally as a grave cover. The Inchyra slab tapers markedly in width towards one end. If recumbent monuments


INCH YR A

342

taper they tend to have their broad end at the head. Stevenson records that the broader end of the Inchyra slab was towards the west, which was also the end where the skull was discovered. The only other factor to consider is the breakage at the broad end, which, if it occurred when the slab fell over, most plausibly did so when the narrow end was at the top (leaving the rest embedded in the soil). Since the fracture clearly post-dates the carving of the symbols at this end, it could be taken to imply that the fish and notched double-disc were the third and final phase. The slab, however, is thin and the break could as easily have occurred when the slab was lying horizontally.

RELATIONSHIP OF SYMBOLS TO OGHAMS There are ogham letters in five positions on the slab: across the rough face, along the narrow end, along the adjacent section of one long side, further along the same side, and in the equivalent position along the opposite side. Three different hands using three different carving techniques are apparent. Since there are also three pairs of symbols carved in three different hands it seems natural to associate each ogham with a different set of symbols. In two cases the connection is almost certain, in the third, it is less clear. The shortest ogham (Padel's C, Stevenson's (d), Wainwright's 3) runs without a stem-line up the broad rough face of the slab immediately to the right of the fish's back. There can be little doubt about the association, all the more so since the ogham is chisel-cut like the symbols. With the broad end (and the fish's head) to the top, this ogham reads upwards. The top of the hypothetical stem is a little further to the left than is the bottom, meaning the inscription is slightly to the left of vertical. The tapering edge of the slab makes this appear more marked than it is. Doubtless the strongly diagonal alignment of the fish has exerted a pull on the ogham's orientation. The most substantial ogham (Padel's B, Stevenson's (a), Wainwright's 1) starts midway along one long side, a little below the fish's tail, and runs 790mm (31") up to the narrow end, continuing round the corner and across the top. It thus neatly frames the fish and notched double-disc widi which it shares the same substantial pock-and-smooth technique. With the slab upright, narrow end to the top, and these symbols on the 'front', the ogham runs up the left edge and across the top - the classic ogham position. There is no real doubt about the associations of these two


INCHYRA

343

oghams, but the remaining two are not so certain. There are two lines of ogham, in exactly the same hand and carved in the same style, along the narrow edges either side of the broad end. These are Wainwright*s 2, Stevenson's (b) and (c), and Padel's A. I will follow Padel's terminology but divide them into A (i) and A (ii).

A (i) starts 45mm (Vhn) below B, runs along the narrow edge for 450mm (\llhn)

in the opposite

direction, leaving 230mm (9 n) blank at the end. On the opposite long edge, A (ii) ends in wear about 90mm (3V£") in from the broken broad edge, and can be traced back for 240mm (9lA") until it disappears in the fractured surface of the slab. Both these slight oghams read vertically upwards with the reader's head turned to the left as one faces die long edge in question. This means that A (i) is on the left when the fish and serpent are the 'front', and A (ii) is on the left when the incomplete symbols are the 'front'. If the slab were lain on its back, rough side down, A (i) would be the right way up, but A (ii) would be upside down. Padel thought these two lines of ogham very tentative and suggested they might have been 'hastily scratched' [103]. He felt they were most naturally associated with the incomplete symbols on the smooth face. Wainwright did not advance an opinion on their association, Stevenson was undecided. He thought the carving of A (i) similar to that of the snake 'and might have been made by a light application of the same tool' [38], but thought a definite association unlikely since C was already linked to that group. On the other hand, he felt 'tempted to associate ... [A(i)J with the unfinished symbols'. He concluded that it couldn't have been a preliminary setting out of a text to be carved more substantially since the strokes were too erratic and close together, but he wondered if it might have been 'a quick note, to be set out better later on'. Padel found this idea appealing, but I think it unlikely. Rather than provide an insight into the working methods of the oghamist, I think A (i) and (ii) are simply informal, casual inscriptions, perhaps even 'graffiti'. I don't think they need necessarily be associated with any particular group of symbols, though their carving in this configuration would have been most feasible when the stone was standing upright with the broad end to the top. I think it likely both were carved at the same time, but this could have been considerably after the carved stone was set up. Stevenson thought A (i) was carved


INCHYRA

344

after B since the former stans only 45mm below the first stroke of the latter, but I do not consider this conclusive (there is ample space at the far end of A, it could have begun much higher), and, for the sake of argument, it could even be taken the other way round - that B followed A (i), since B occupies all the available space between the end of A (i) and the far edge of the narrow end. A (i) starts at roughly the same level as the bottom of the mirror-handle and the lowest point on the snake's tail.

To summarize, B is associated with the fish and double-disc, C with the fish and serpent, A (i) and (ii) go together and may be associated with either of the symbol statements at the broad end, or both of them, or neither of them, but they were almost certainly carved when the stone stood upright with the broad end to the top. The evidence for the order of carving is ambiguous and the sequence remains unproven.

DESCRIPTION OF OGHAM INSCRIPTIONS

A (i)

(Stevenson (b), Wainwright 2)

This very slight inscription was first noticed by Richardson. It is casually carved with a particularly haphazard stem-line ('extremely erratic1 in Stevenson's words), but is nonetheless legible. The only doubts arise because of lack of neatness. All the consonants slope and the direction of reading is thus quite clear [see fig. for key to numbering]: i

Stevenson illustrates four short strokes across the stem. According to his drawing there is a clear gap between the start of the stem and the first letter, thus the letter can be seen to be complete. There is a very slight forward slope on these strokes, but it is not nearly as pronounced as later xorQicme letters, and is explained by the gentle curve of the stem here (they maintain perpendicularity with it). As Padel makes clear» there is no reason to read 'Z' (i.e. Si) as Wainwright did, the length of the strokes proves them to be part of a vowel - E. Unfortunately the conditions under which I examined the inscription were not ideal, but I wondered if there might in fact be two vowels here - two or possibly three strokes immediately following the crack


INCHYRA

345

in the surface, then, after a gap, three vowel strokes (UU or OU). This, however, is very uncertain. 2

Fortunately the next section is clear: Three strokes to the left of the stem sloping backwards - T.

3

As above, with longer strokes - T. Letters 2 and 3 are widely spaced.

4

Two strokes to the right of the stem sloping forwards - L.

5

Five short vowel-strokes perpendicularly across the stem - 1 . Padel comments that the third and founh strokes are slightly more widely spaced than the others, but he is quite right that subdivision (UO) would be ftoo subtle'.

6

Five short vowel-strokes perpendicularly across the stem - E. Again, these are well spaced from the previous group.

7

Three strokes to the left of the stem, sloping backwards - T.

8

Five very long oblique strokes sloping forward across the stem - R. The length of these strokes shows that all the previous cross stroke letters are vowels.

9

Four strokes across the stem - E or S. These are definitely shorter than the preceding group, but longer than earlier vowels. They are, however, perpendicular across the stem and all previous authorities have read E.

10

Four strokes to the right of the stem, sloping forward - N.

11

The next character looks like a right angled triangle sitting on its point. It consists of one vowellength stroke across the stem, followed by an angled-vowel-type stroke $>), their b-distal-tips are touching, but not the h-distal. There are at least two ways of interpreting this character, either as one half of the angled-(^character seen at Bressay (see also entry on Formaston for possible phonetic values), or an O made up of one straight and one angled stroke, cf. the similarly mixed E on the back right ogham at Brodie. Either way the character is unprecedented.

12

Five vowel-strokes more or less perpendicularly across the stem -1. That these are rather longer than, say 1 or 5/6, lends further support to the vocalic interpretation of 9. The stem has been sloping down through letters 10 and 11, is virtually invisible at 12 and starts afresh back towards


INCHYRA

346

the middle at 13. Padel's alternative suggestion of Q makes too much of this droop in the stem. 13

Two short strokes to die left of the stem, sloping backwards - D.

14

As above - D. This section (13/14) is very worn and I am reliant on Stevenson and Padel for interpretation.

15

The next portion is a little clearer. Two short vowel-strokes perpendicularly across the stem - 0 .

16

Five long oblique strokes sloping forward across the stem - R. Padel noted that the first stroke was rather widely spaced from the rest, but rightly concluded that this is not significant.

17

Three or probably four strokes to die right of the stem sloping forwards V or S. Padel commented that the first and fourth strokes were faint almost to the point of invisibility. This appears to be the end of the inscription.

Which gives the following reading: (e/uu/ou)TTL(i/uo)ETR(e/s)N(o?)IddOR(s/v) probably ETTLIETRENOIDDORS

A (ii) (Stevenson (c), Wainwright 2) This inscription is not visible as the stone is currently displayed, thus I have not been able to verify the accounts of earlier authorities by personal examination. The following is based on these previous accounts and on photographs in the possession of Perth Museum [I am grateful to Dr Michael Kingforallowing me to examine these]. Wainwright does not illustrate A (ii), nor comment on the carving, he simply lists his reading. Stevenson quotes Wainwright's reading and provides a line drawing but, according to Padel, it is inaccurate in certain details (a view supported by my examination of the photographs).

Approximately 240mm (9V£") of lettering survives, but both the beginning and end of the inscription appear to be missing and even the extant letters are damaged in several places. Stevenson says there is no trace of lettering on fragments from lower down this edge but does not give details of how far these clear


INCH YRA

347

patches are from the present beginning. Thus an unknown number of letters have been lost from the beginning. The marked difference in length between vowels and consonants and the fact that all consonants slope gives a number of clues to the interpretation of the extant carving. Unfortunately the stem-line is not particularly clear. There is a very misleading crack running the entire extant length of the lettering but the stem is just visible a few millimetres below. The slope of 2, 3 and 8 indicate the direction of reading as towards the broad end as follows [see fig. for key to numbering]. The beginning of the inscription has been lost and extant section begins in mid-letter: 1

Three short strokes perpendicularly across the stem - U, or any subsequent letter of the aaicme. Padel complains that Wainwright's drawing has these rather shorter than they really are, this is confirmed by the photograph, not that it affects the reading, of course.

2

A single oblique stroke to the left of the stem, sloping backwards - H.

3

Three oblique strokes to ihe left of the stem - T. These are roughly parallel with each other and with 2. The first is longer than the other two. There is a clear gap, more than enough for a stroke, between letters 2 and J, so there can be no doubt that they are to be read separately.

4

Three short strokes perpendicularly across the stem - U. These follow fairly hard on the end of 3. There may be a fourth stroke (E), but it is doubtful.

5

After a clear gap there is another vowel stroke perpendicularly across the stem - A. The next section is worn but at the end of it is another vowel stroke. If this is all one group it might have contained four or five vowel-strokes, judging by the space occupied by 4 - E or I. On the photograph there appears to be two strokes between the clearer ones, of a similar length but sloping forward across the stem. This seems unlikely but the section is too damaged to be sure.

6

Two long oblique strokes sloping forwards across the stem - G. From the photograph it is clear that these do cross the stem, in any case, from their slope they must be members of the m-aicme. Padel comments that there is no trace of Stevenson's dotted score before this group, certainly none is visible on the photograph, it seems doubly unlikely since, as Padel points out, if produced to the full length of an m-stroke it would intersect with 5 (in any case the third member of the m-


INCH YRA

348

aicme is extremely rare in epigraphic ogham). 7

There is ample room for the four vowel strokes read by Wainwright at this point (E), though the photograph is obscure.

8

The tips of the two h-aicme strokes read by Wainwright are clear (D), it appears as if this group is complete but shortly further on the entire surface of the stone has been lost, and the inscription could well have continued. There is a further 90mm (3 xh") before the broken end, and, of course, the slab originally continued further. By this point A (ii) is already higher than A (i), and may have gone on quite a bit more.

Which gives the following tentative reading:

B

-](u+)HTu(o/i)( /a)GED[-

(Wainwright 1, Stevenson a)

This is the boldest of the Inchyra oghams, the only one to be noticed initially. The letters are rather deeply carved with a pock and smooth technique comparable to that of the symbols at this, the narrow end. The inscription comprises approximately eighteen letters, 1-14 up the long side, 15-18 across the short end [see fig. for key to numbering]. The inscription ends there and does not continue down the other long side. All the extant carving is clear, unformnately several portions of the h-surface have flaked off and there is particular doubt over 13 and 14. The large fragment which contained the h-strokes of 15-18 has been recovered and the letters completed. The first three letters, and especially the first, are less distinct than the rest. Stevenson commented on the 'irregular punching of their margins'. His suggestion that this represents an attempt to render less distinct the portion of B visible when the narrow end was sunk into the ground seem very plausible.

Letters 77, 72, and 15-18 have bind-strokes at their b-distal tips (the h-distal tips having been lost), but, though the b-distal tips of other letters are clearly visible, they are not bound. What distinguishes 77/72 and 16/17 is that they are successive letters of the same aicme and thus in particular need of unambiguous


INCH YRA

349

grouping. The slope of letters 2, 4, 9, 16 and 77, indicate the direction of reading. The near unanimity of readings between Wainwright, Macalister (quoted by Wainwright), Stevenson (and by inference, Jackson), and Padel is an indication of the clarity of the carving. The lettering begins flush with the end of the stem-line and, occupying almost the full width of the narrow long edge, continues upwards as follows: 1

Five long straight vowel-strokes perpendicularly across the stem, quite widely spaced - 1 . The second stroke slopes a little backwards which means this could be two separate groups, 2 + 3 , OU, but all previous authorities have agreed on taking them as a single unit and the spacing is in favour of this interpretation. One would have expected successive members of the same aicme to have been more clearly differentiated.

2

Five short oblique strokes to the right of the stem, sloping forward - N. More closely spaced than the previous group.

3

Four long straight vowel-strokes perpendicularly across the stem - E.

4

One short oblique stroke to the left of the stem, sloping slightly forward - H.

5

One short stroke to the left of, and perpendicular to, the stem, - H. Presumably the differentiation of gradient is to clarify two successive letters of the same aicme.

6

Four long straight vowel-strokes perpendicularly across the stem - E. Most of the h-surface has been lost here, but there are traces of strokes on that side, especially the fourth.

7

Three short strokes to the left of the stem, sloping backwards - T. The first stroke sloped greatest and the third is almost vertical.

8

Four long straight vowel-strokes perpendicularly across the stem - E.

9

Four oblique strokes to the right of the stem, sloping backwards - S. Quite closely spaced.

From here to the end of the slab, very little of the h-surface survives. The few traces immediately above the stem must be interpreted with the help of the spacing of the extant b-strokes. 10

There are four proximal tips visible on the h-surface, the first being particularly clear, and the bsurface is intact and empty. These must therefore be h-aicme rather than vowels - C. Wainwright


INCH YRA

350

thought that, while there was space for four strokes, 'balance and spacing* was in favour of three (T). I agree with Padel, however, that C is more likely. 11

Five long straight vowel-strokes across die stem - L Only die proximal tips of the h-strokes survive. There is a very faintly pocked bind-stroke connecting the distal tips of the b-strokes.

12

As above, but only four strokes - E. The final stroke is die most fully preserved on the h-surface. Again, there is a faint pocked bind-stroke.

13

The h-surface is lost almost to the stem and only the distal tips of the strokes survive. Two are definite, there is probably a third, and possibly a fourth. Since the b-surface is intact and empty, these must be h-strokes - D, T, or C.

14

Between here and the end there is room for a few h-strokes, perhaps three, but nothing has survived the damage to die corner of the slab. It is likely that the carving went right to the very edge, since it does so on the narrow end. Wainwright took 13 and 14 as TD, DT, or Q. All are possible, though Q outside the formula word MAQQ would be exceptional.

The final four letters take up die whole of the narrow end face: 15

Five long vowel-strokes perpendicularly across the stem, bound - 1 . Padel was concerned that the fragment which completes this end did not have the h-part of the first stroke. To explain this he thought it must have been Very curved out of keeping with the other scores' [106]. He said that the b-side bind-stroke didn't meet the first stroke, and thus wondered if the latter might have been part of anodier letter. I feel, however, that all of this may be explained by the damage die corner has suffered. There is a clear gap between die end of the slab and the first stroke and the spacing is such that the group cannot be other than a single group of five cross-strokes.

16

Five oblique strokes to the right of the stem, sloping forward, bound - N. Each stroke is a little longer than its predecessor giving the group a slightly wedge-shaped appearance. The first stroke, in particular, is rather short, presumably to prevent it intersecting with the last stroke of 15.

17

As above - N.

18

Four long straight vowel-strokes perpendicularly across the stem, bound - E.


INCHYRA

351

Which gives the tentative reading: INEHHETESCIE(t/c/d)(t/.)INNE

C

(Stevenson (d), Wainwright 3)

The strokes of C are not as distinct as those of B and the inscription went unnoticed at first. Gordon thought them 'trial grooves' [1956:42], but Richardson recognized them as ogham letters cut using the same technique as the adjacent symbols. The letters are quite deep, a smooth chisel-cut groove, not pocked like the letters of B. Though markedly less substantial than B, the letters are, in feet, of a similar size, though the script used is different. The stem-line is not drawn-in but must be imagined, and the vowels are very short strokes. The slope of 1 and 3 indicates the direction of reading is from bottom up (towards the broad end), with head turned to the left, as one would expect [see fig. for key to numbering]: 1

Four long oblique strokes sloping diagonally forward across the hypothetical stem - S. Padel and Wainwright were in no doubt that the reading was S, but there is the possibility of the end of a fifth stroke before this (N).

2

Four very short vowel notches - E. Three are clearly visible, the fourth less so.

Above this

letter the surface of the stone has flaked away, but there can be no doubt that these are complete vowel-strokes. 3

Three longish oblique strokes to the left of the stem, sloping backwards, each a little longer than the previous one - T. Thefirsttwo are clearly visible, the third stroke is almost lost in a flake. Wainwright thought there might have been a fourth stroke now flaked away, but I agree with Padel [106] that this is unlikely.

4

Three short vowel notches - U. As Stevenson points out, flaking has removed the original surface of the stone and 'perhaps one or two scores' [39], as a result these three strokes are faint and worn, and portions of them are missing.

A line, which Wainwright claims post-dates the

discovery of the stone [6] has almost obliterated the third stroke. The inscription could have continued, but there is no trace of any farther strokes.


INCH YRA

352

Which gives the reading: SETU(+) or NETU(+)

FORM OF SCRIPT The three Inchyra inscriptions, A (Mi), B, C, look radically different but, the variation in scripts is not as great as might at first appear. C lacks a stem, and B has some letters bound, but setting these aside for a moment, all are written in fairly simple forms of the script with similar over-all proponions in letter size and spacing. All have b- and h-aicme letters which slope and, with the exception of character A (i) 77, there is a noticeable lack of supplementary letters and other developed forms. Nonetheless, despite these similarities, there are three distinct 'hands' visible - A (Mi), B, and C - and the differences between them or not merely those of carving technique.

A is written in an infonnal and cursive hand which, though not exactly careless, has given little attention to the finer points of length, slope, and spacing. Letters are, however, clearly distinguished, especially successive letters of the same aicme, and legibility is not compromised. The vowels are rather shon strokes across the stem and occupy a minimum of the middle third of the ogham band. The closest palaeographical comparisons are with Ackergill and Buckquoy (excepting the forfeda), but Inchyra A seems somewhat exceptional in having sloping b- and h-consonants and well-spaced letters (both features more common in later inscriptions) but having shoner vowel-strokes and lacking, with one exception, supplementary letters and other developed features, which are generally considered late.

Conversely, with B, the bind-strokes, which elsewhere are correlated with having numerous forfeda and other late features, seem a little incongruous on an otherwise very simple Type Ila ogham. It could be merely coincidental that the text was such that no supplementary letters were required, and certainly, vowel-strokes which, as here, occupy the entire ogham band, are generally later than shon-stroked ones. It is also curious that there are no m-aicme letters, but this could be similarly coincidental, the text is not long. Note the marked and consistent slope of the b- and h-aicme consonants (the only exception is 7


INCHYRA

353

whose slope diminishes with each stroke). Component strokes are generally parallel but there is some variation in how widely spaced they are, giving the letters an uneven quality. Letters, especially successive letters of different aicmi, are rather closely spaced, though they remain clearly distinguishable. Successive letters of the same aicme are more clearly spaced but by less than a stroke's wonh. In contrast to other inscriptions employing bind-strokes, which have all appropriate letters bound, bind strokes are not used for every letter at Inchyra. They appear to have been reserved for use only if there was room for doubt, though damage to the distal tips of several letters prevents a categorical assertion of this. B is a wellcarved inscription on a monumental scale compatible with that of the neighbouring symbols. It is most closely comparable the ogham from St Ninian's Isle, and to a lesser extent Scoonie and Brandsbutt (though these do not share the marked slope of b- and h-aicme letters). With one exception, all the other instances of bound letters occur in inscriptions from the Northern Isles, written in scripts markedly more complex than Inchyra's, containing rmmerousforfeda and even, in some cases, word division. The exception is, however, an important one, Abernethy. Though it has only three letters surviving, these are closely comparable with those of Inchyra B - sloping h-strokes and long, straight, vowel-strokes occupying the full ogham band. That the only two Mainland examples of bind-strokes should be within three miles of one another is striking.

The lack of a stem-line at C seems incongruous with its position across the face of the slab, but is scarcely as revolutionary as it seems. There are numerous examples of Irish oghams carved on stones so round and smooth that there is no arris, or an arris so flat as that the inscription is carved effectively across a face. Macalister lists a number of pulvinar or 'cushion-shaped' boulders from Corca Dhuibhne with such inscriptions, including Lugoagappul (CIIC 190-1), Burnham (175), Ballintaggart (155-63), and Ballineesteenig (147), but there are others elsewhere in the country, for instance, Tuckamine, Co. Carlow (CIIC 17). The incongruity of Inchyra C is merely that a stem-less ogham was carved on a slab with sharp and angular arrises. This choice may support the theory that C was the third ogham to be carved, since both narrow edges of the broad end would by then be already occupied by A (i-ii). But why leave out the


INCH YRA

354

stem-line ? Perhaps the carver was more familiar with the older conventions, certainly the very short strokes for die vowels are more in keeping with the old arris-cut oghams. But the marked slope of the band h-aicme letters, and the ample spacing of the letters have more in common with the later, manuscriptinfluenced, stem-written oghams.

The relative proportions of vowel and consonant strokes is paralleled at Auquhollie and Bac Mhic Connain (both written without a stem-line on an arris), but both have upright b- and h-strokes, similarly with Gumess. Newton has more of a slope on its b and h-consonants, and, in being split between stem-less arris and stem across the face, underlines the feet that the presence or absence of a stem may in some cases be a somewhat arbitrary indicator of date (similarly with the late but stem-less Golspie). Pool combines short vowel-strokes and sloping b- and h-strokes, and, apart from its stem-line and the exceptionally long thin profile of its letters, is perhaps the most closely comparable to Inchyra C. Previous authorities have commented on the 'Irish-looking* vowels [Padel 1972:106; Stevenson 1959:39], though this is more correctly a feature of date rather than ethnicity. The implication, however, as Stevenson noted, is that C is the earliest of the Inchyra oghams and thus that 'the order of the symbols argued above is in contradiction to the typological order of the inscriptions that seem associated with them' [ibid.]. I think it would be safer to say that the typological development of ogham, and its chronological implications, has yet to be worked out. It appears that within an essentially unified tradition, there may have been several styles used in different areas at a given time.

INTERPRETATION OF TEXT With die exception of A (ii) which is defective at both ends and in the middle, all the Inchyra oghams appear intact and, apart from a few letters which are damaged and difficult to read, are more or less legible. Nor are there many problems with transliteration, only the triangular character, A (i) 11, is obscure. The difficulty comes in knowing how to segment these undivided texts which do not conform to the standardformulas('X MAQ Y etc.). Since we do not know the functions of the monument in its


INCHYRA

355

various phases of use we have few clues as to the kind of messages to expect. We do not even know if we should be looking for text in Irish or Pictish.

Inscriptions B and C are complete and independent. Lines A (i) and (ii), however, may be pan of a single text. If they are, their order can only be guessed at, all the more so since A (ii) is missing both its beginning and end. That A (i) does not occupy the whole of the available space, and A (ii) continues higher up the slab may indicate that A (i) is the run-on, though, given ample space, word boundaries would probably have an over-riding effect on layout. Even if they are both part of the same text the order of the two sections may not have been important. The problem is largely academic, since so little of A (ii) survives, but it ought to be born in mind that A (i) need not be semantically complete or self-sufficient. The informal nature of the script and carving, may give a clue as to meaning in that the text need not express an 'official* or 'public' sentiment. Neither need it refer to the rest of the carving, it could be no more than a graffito on a standing monument.

A (i)

ETTLIETRENOIDDORS

The reading of some of the vowels is doubtful, but the only transliteration query concerns 11 the unique triangular character. If it is a mix of straight and angled vowel-strokes then we may take it as some kind of O. Even if, on the other hand, it is related to the fh/br/fc/, it is still probably some kind of O. SimsWilliams has posited /6/ for the original value of this character, replaced at a very early date by a vocalic value, perhaps / 0 : / . The diphthong value oí/óe is, in his opinion, probably no earlier than the ninth century, the digraph value oi later still [1992:58-60; see entry on Formaston for discussion].

The first problem is where to segment the text. The sequence of letters is too long for a single personal name yet no word division is indicated. The only principle we have in deciding where to insert it is the tendency to avoid geminate symbols in word-initial position. The initial ETT is paralleled in the roman alphabet inscription from St. Vigeans [Okasha 1984; Clancy 1993] where it appears between two personal


INCH YRA

356

names possibly as a variant spelling of Latin et 'and'. Macalister's ett between personal names on the roman alphabet inscription from Carew, Pembrokeshire (CIIC 1035) is a misreading [Nash-Williams 1950:184 No.303]. In ogham terms, double T may be explained as an attempt to indicate non-lenition, but it seems a rather outlandish explanation to posit a Latin word made to conform with native spelling conventions. This explanation would entail taking A (i) as following on from A (ii), being die end of a list of names et Lietren dddors.

lietren recalls the Lutrin in the Pictish king-list, which Jackson

interpreted as a Pictish reflex of Celtic *Lugutrinos. As Jackson shows, the expected Brittonic reflex of this would be *Loutrin [1955:164], which does not account for Inchyra's LIE-. According to diis interpretation OIDDORS would be some epithet of Lietren, though what this might be is unclear.

This line of enquiry may be entirely mistaken, however. Padel notes that ETT- is found as the opening sequence of Cunningsburgh 3 and Lunnasting. In die case of the latter it was very tentatively suggested that ETTE might be some form of the copula it-e, taken with the -S at the end of the section to mean, perhaps 'this is'. If the text is to be interpreted in Irish terms, then LIE may be Olr. lie 'stone' \DJL s.v. Ua], This spelling is attested for the nominative singular in the Old Irish glosses [WB 21c6, ML 131cl2, see Stokes & Strachan 1901-3]. While meaning 'stone' in a general sense, the word also has a range of more specific meanings including 'standing stone, pillar-stone, usually of a memorial stone, one marking place of death or burial, ogham-stone', later it comes to mean 'grave' or 'tomb', or a stone functioning as a landmark [DH] (this interpretation seems more likely than lie for lia 'flood', 'spate'). Irish trén is an adjective meaning 'strong', but is also used as a substantive meaning 'strong man'. It appears as the first element in compounds with nouns, adjectives, and verbs, meaning 'strongly'. Perhaps more relevant, is its use as an element in male personal names, e.g. Trénmór and Tréinfer [Uhlich 1993:302], cf. ogham TRENAGUSU (CIIC 428), TRENACCATLO (CIIC 353), TRENALUGGO (CIIC 26), MAQI-TRENI (CIIC 341). The element also appears uncompounded as an independent name, cf. ogham TRENU (CIIC 57). Inchyra's TRENO- could be the genitive of the name Trén, with LIETRENO for lie Tréno 'Trén's


INCH YRA

357

(ogham) stone'. This would leave IDDORS. The closing sequence -ORS is also found at the end of Cunningsburgh 2 but since the text is probably to be interpreted as the name Conmor + sy this may be no more than coincidence.

Alternative segmentations yield NOIDDOR, but the DD seems to preclude interpretation as Irish noithir, the present passive form oinóid 'makes known', 'spreads the feme', or some form of na(i)thir 'snake', either of which might be deemed appropriate comment on the stone, similarly, with OIDDOR for odhar present passive of oidid 'offers' 'grants' 'lends', used in the laws, in relation to the lending of property meaning 'it is granted' [DH]. To be accepted, any explanation must account for the whole of die text, so the above can remain no more than suggestions.

A (ii): -](u+)HTu(o/i)( /a)GED[This text is too short and fragmentary to be made much sense of. It may begin and end in mid-word, unknown amounts of text have been lost from both ends, and there may well be internal word division.

B:

INEHHETESCIE(t/c/d)(t/.)INNE

Unlike A (Mi), this inscription is a formal one, clearly related to a symbol statement.

We might,

therefore, expect a different kind of message. Padel comments that both sections of this text begin with IN(N)E. In Irish inne means 'wealth', or 'private possessions' [DLL], but there is no certainty that this is how the text is to be segmented.

More relevant may be the comparison between the opening

INEHHETE- and Gurness's INEITA- (unexplained). The text is complete and, apart from a few letters, easily legible. This makes our difficulties in interpreting it all the more frustrating.

C:

SETU or NETU

Like A, this is an informal text, but like B, it is clearly related to a symbol. The association with the fish may be the key to interpreting the text, since it seems to be paralleled on several other ogham stones. That


INCHYRA

358

Ackergill (NEHT-), Golspie (NET-), and Latheron (NETU) all feature fish seems more than merely coincidental. Note also the beginning of the legible section of A (ii) which may be -EHT (for NEHT-?). Netu may be a name in its own right, or a hypocoristic form of Nechtan or similar [see entries on Ackergill and Golspie for further discussion]. The extant visible carving, however, reads SETU rather than NETU. This may be a hypocoristic form of a personal name like Irish Sétnae or similar. Whichever reading is correct, the text has the air of a simple personal name.

DISCUSSION Inchyra is one of a tight little group of oghams in south-eastern Perthshire. The link with Abernethy is particularly close, but Dupplin is only 8Vi miles away. The three, which are isolated from the next nearest oghams (Scoonie to the south and Auquhollie, a considerable distance to the north), may reflect Irish influence in this part of the Pictish kingdom of Fomenn. It is unfonunate that only Inchyra is legible and doubly so, since it provides so many problems of interpretation. The Inchyra texts are in line with other Class I symbol stone oghams in not conforming to the 'X MAQIY' formula. This cannot be coincidental, especially since most of the ogham-inscribed Class II cross-slabs do hav•- 'X MAQQ Y* (of the exceptions, Scoonie has a single name, and Brodie is largely illegible). Inchyra is a galling example of how difficult it is to interpret these texts as soon as they deviate from standard formulas. Text A (i) may provide a further example of the linguistically puzzling -S in what appears to be syntactically genitive position in a Scottish ogham. Inchyra is, in some ways, an anomalous monument and one with a complex history we don't fully understand. It is impossible to date closely, but the slab's regular shape, the use of a chisel for some of the carving, and certain features of the ogham script, are more plausible interpreted at a date later rather than earlier in the sequence of Class I monuments. Why a single monument should have attracted so many inscriptions is a puzzle, but this repeated use over time, supports the notion that ogham was well-integrated into the local culture, not a rare or exotic feature. The different hands and styles used further underline the heterogeneity of the ogham tradition in Scotland.


INCHYRA

359

BIBLIOGRAPHY OS Record card - NO 12 SE 9 McLaren 1946; Stevenson 1959:33-9, pl.III-IV; Wainwright 1959 (reprinted as Wainwright 1961) (pi. 11 facing p.368); Gordon 1956:42-3; Henderson 1967:214-5, pl.24; Padel 1972:100-107; RCAHMS 1994a:92, 93, ill.; Ritchie & Fraser 1994:23.

CONCORDANCE Forsyth

A (i)

Padel

A

Stevenson

(b)

Wainwright

A(ii)

A

(c)

(2)

B

B

(a)

(1)

C

C

(d)

(3)

PREVIOUS READINGS Wainwright

1:

INEHHETES(t/c/dh)IE(td/dt/q)INNE

2:

(e/z)TTLIETRENo(i/q)DDOR(s/v) ~(u+)HTUoa(g/ng)ED

3: Stevenson

Padel

(2)

SETU (as above)

A:

ETTLIETRENOIDDORS -](u+)HTU(o/i)(a/ )GED[-

B:

INEHHETES(t/c/dh)IE(q/td/dt)[-]iNNE

C:

SETU

INCHYRA Symbol and ogham-incised slab [RCAHMS 1994a:95a]

'Rough'

AC)

INCHYRA - Symbol and ogham-incised slab showing position of inscriptions [Stevenson 1959 pl.Ill]

'^%m? INCHYRA - 'Unfinished* symbols and inscription A (i) [after Stevenson 1959 pi.IV] (key to numbering)

INCHYRA - Detail of damaged edge showing inscription A (ii) [photograph © Perth Art Gallery and Museum]

(key to numbering)

1

2

3 4 5

6

7

8

9

10

11

12

13 14

15

16

17

18

INCHYRA - Inscription B Top: Edge and top of slab showing position of lettering [Stevenson 1959 pL III]; Middle: Detail of inscription (key to numbering) [RCAHMS 1994a:931 Bottom: Detail of lettering on edge [Henderson 1967 pl.24J

5

4

.4* V. #

.

INCHYRA - Detail of broad end of slab showing symbols and inscription C (key to numbering) [Stevenson 1959 pi. Ill]

360

LATHERON

DISCOVERY This stone was found in 1903 built into the interior wall of an old byre at Latheron, Caithness [NGR ND 1981 3315]. The finder, John Nicolson of Nybster, who had previously discovered the Ackergill slab, brought it to Sir Francis Tress Barry who in turn presented it to the National Antiquities Museum (NMS Cat. No. IB 183). The previous whereabouts of the stone are unrecorded.

SITE A second fragment was found built into the same byre: a cross slab 2*3" by T P (c. 0.69 x 0.33m), on which the only carving to survive is the outline of parts of the side arms of the cross and the remains of the splayed shaft (1*7" (480mm) long, 6-8 V4" (150-220mm) wide). The angles of the cross arms are filled with T (50mm) diameter discs [RCAHMS 191 lb:82 No.298, OS record card - ND 13 SE 26]. A third fragment, part of a Class I symbol stone [OS record card - ND 13 SE 31] was found fonning the lintel of a false window in the south gable of the farmhouse of Latheron Mains [NGR ND 1990 3343]. It is incised with a 'crescent and v-rod' symbol decorated with what Stevenson has called the 'dome-and-wing' pattern [Stevenson 1959:40 pl.v.3]. Bothfragmentsremain in situ.

Apart from these three pieces of sculpture, which encompass Classes I, 11, and III, there is no other evidence for early medieval activity at Latheron. The two cross-slabs, however, might be thought indicative in themselves of some kind of ecclesiastical site. The only other pre-Norse remains in the area are afragmentof a penannular brooch from nearby Achavrole, and the possibly Pictish settlement at the Wag of Forse [Batey 1991:55,52]. It is significant that the name Latheron (or Latheron-wheel < Gaelic Latham a'phuill 'Latheron of the hole or poor, attested 1287 as Lagheryn, Laterne) is one of die few Celtic names to have survived in an area with an overwhelmingly Norse toponymy [Waugh 1993:120-1]. It is impossible to know whether the name was originally Brittonic or Goidelic since the root lath is


LATHERON

361

common to both, giving Welsh Head 'mire', Irish lathach 'puddle', Scottish Gaelic láthach [Watson 1926:122].

DESCRIPTION OF OBJECT A cross-slab of local stone carved on one face with, one above the other, the lower part of a cross in relief, an incised bird and fish, and two incised horsemen. To the left is an incised ogham inscription [see fig.]. Stone:

Caithness sandstone

Dimensions:

3'x 1 '5wx 4" (c.0.90 x 0.43 x 0.10 m)

Condition:

Left edge intact, all others fractured. Portions of unknown length lostfromtop and bottom. Carving worn, but generally clear.

The left edge of the slab is more or less intact, and is bounded tor the entire length of the carved face by a plain border. The right edge is more seriously damaged and lacks the raised margin, but, judging by the roughly central position of the carvings, it appears that only a few inches have been lost. More difficult to gauge is the size of the pieces lost from top and bottom. Several inches more would be necessary to complete the horsemen at the foot. If the rectangular design at the top is correctly interpreted as the lower part of a cross-slab than a fairly substantial piece would be needed to accommodate the missing arms and upper shaft. This would make the Latheron slab tall and thin, but not implausibly so.

In attempting a reconstruction of the original dimensions, comparison should be made with the two other Class II cross-slabs from Caithness. The Skinnet slab is a grander and more accomplished monument than Latheron, but the two have a number of features in common. Although carved on both sides, Skinnet, like Latheron, has a simple border round its margin, and incised symbols below a relief cross. On its other face, Skinnet has a relief cross above a horse, recalling Latheron's horsemen. The Skinnet slab is complete and has a height-to-width ratio of over 3:1. If we allowed an extra four inches of width for the Latheron slab then, if reconstructed in the same proportions as the Skinnet slab, it would be of the order of 5' 10" tall. If the 'missing' 2'10* were distributed 10"at the bottom and two feet at the top, there would be


LATHERON

362

ample space to complete the designs.

The tapered Ulbster stone is broader than Skinnet, more of the order of 2:1. This ratio would allow Latheron only a further 6". The Class II stone from Edderton in Easter Ross is a more sophisticated piece than Latheron, but has a number of features in common with Latheron and Skinnet. It has a simple band round its outer margin, a relief cross above symbols on one side and on the other, a relief cross above horsemen. Though not quite as slender as Skinnet, Edderton has a height-to-width ratio of 8:3, which if applied to Latheron would yield an extra 1'8\ So, if Latheron had been originally rather tall and slender, there would be no lack of precedent.

Macalister argued that, since the symbols were carved with the pock and smooth technique and the oghams were chisel cut, that the two were not contemporary. Since three different techniques were used to carve cross, symbols and horsemen, this is hardly a compelling argument. Also the lightly incised straight lines of the ogham would be most easily cut in sandstone using a blade, while the more substantial curves of the symbols might be considered more suited to the traditional punched technique. Since we would expect the ogham to be carved last in any case, that the stem appears to bend at two points to avoid the carving is of no significance. The only conclusive proof that all the carving on the slab was contemporary would be if the overall design could be seen to make allowance for the ogham inscription. While the space between the intact border and the symbols appears more generous than necessary if the carver had not had to take the ogham into consideration, it is impossible to get a clear sense of the relative position of the decorative motifs without the right margin. Nonetheless, there is no compelling reason to doubt that the ogham is intrinsic to the original design.

DESCRIPTION OF CARVING - DECORATION The carving occupies the whole of one face of the slab. At the top is a double rectangularfigurein relief which has been interpreted as the lower portion of a cross shaft. The upper rectangle is 'filled with double


LATHERON

363

spiral ornament arranged in C-shaped scrolls placed back to back' [RCAHMS 19llb:82]. The lower, narrower, rectangle is filled with an interlace pattern. If this is the remains of a cross shaft then it is of unprecedented shape. The shaft of the crosses at Ulbster, Skinnet, and Monymusk ABD, narrows, though in each case it expands again into a base [ECMS 33, 30, 192].

Below the rectangular figure there is incised a bird of prey, presumably an osprey, with its talons sunk into a fish. It is possible that these are to be interpreted as a unitary symbol rather than the typical pair, and thus in a different category to the goose and fish on the Class I slab from Easterton of Roseisle MOR [ECMS 124]. Charles Thomas believed the bird with the fish on the Class I Gairloch slab to be a goose also [1963:94], though others have interpreted it as an eagle [Stevenson 1952:110-1]. The eagle with its talons in a fish reappears on St. Vigeans 1 ANG [ECMS 238] and in two mid-ninth century Irish manuscripts, the Book of Armagh and the St. Gallen Priscian [cited by Allen ECMS 238]. Allen also mentions an Anglo-Saxon and a Continental example of the motif, which later appeared in the Bestiary tradition representing Christ and the world. Whether the Latheron creatures were intended to convey such a meaning is unclear. They may fit more easily with Pictish images of wild animals such as the Ardross wolf ROS, the Burghead bull MOR, or the Knocknagael boar INV, which surely are not open to Christian exegesis [for the Christian background to their an rather dian meaning see Hicks 1993].

The lowest register of the Latheron slab is occupied by two horsemen in profile, one behind the other, partly broken away. Pairs of horsemen are iiot uncommon in Class II and III hunting scenes, but a closer parallel for this design is found on the lowest panel of die Edderton ROS cross-slab. The Latheron 'cross' is carved in thick relief, the bird and fish deeply incised, while the horsemen are only lightly incised. This distinction in technique, in tandem with the vertical ranking of the three elements, may be indicate their relative importance within the over-all message. There are numerous parallels for a relief cross with incised symbols, for instance, Migvie ABD, Glamis ANG, and Golspie SUT, and the Caithness examples mentioned above, Ulbster and Skinnet. On an historical grounds, Curie dated die Latheron slab to the


LATHERON

364

seventh century, but this seems rather early. The horsemen sit much more happily in the context of southern Pictish art of the eighth century or even ninth.

DESCRIPTION OF CARVING - OGHAM INSCRIPTION The ogham inscription is arranged in the usual position, vertically up die left-hand side. One might have e x i t e d the stem to have been placed up the middle of the raised border as it is at Brodie, but instead it runs parallel to it on the sunken background to the cross. Contrary to Macalister's sketch, the stem runs the entire length of the extant portion of slab. At the break, the stem has already begua There is ample space before the first stroke to make it clear that the break has not occurred in mid-letter, what is not clear is whether any letters have been lost from the beginning. Although the stem of several Scottish oghams begins at the first stroke (e.g. Brandsbutt, Ackergill, and Cunningsburgh 3), there are a few examples where the beginning is intact and the stem begins a little before the first stroke (e.g. Bressay). Just over 30mm of stem is visible before the first stroke, which is proportionally longer than most other lead-ins, but, on the other hand, it is also longer than the gap between any of the other letters in the Latheron text. The mark to the right of the stem immediately after the break might be a letter-stroke, the end of any letter of the b~aicmey but it is very feint. The carving is rather worn towards the top. The upper fracture occurs part way through a letter and, from the extant text, it seem likely that the inscription continued at least a little longer. The direction of reading is confirmed by the orientation of the hammer-head A (9) and by the slope of b- and h~aicme consonants (7, \ Padel preferred to read a single vowel-stroke - A and take the subsequent strokes as two sloping h-strokes, spaced to make two separate letters, possibly HH. His explanation is not entirely convincing since he admitted the possibility of a stroke opposite the second of these h-strokes was Very dubious'. This area is worn.

17

Three short strokes to the left of, and perpendicular to, the stem - T. The third of these strokes has the last few letters of the old museum label 'Caithness1 painted over it.

18

Three long strokes across the stem sloping slightly forward, before the fracture - U or any subsequent vowel (E, 1).

Which gives a reading of: DUnNODNNATMAQQNET(u+)-


LATHERON

367

FORM OF SCRIPT The drawn-in stem line and two supplementary letters qualify Latheron as a fairly typical Type lib ogham. Aesthetically, it is not a panicularly pleasing example, but, allowing for the wear on the stone, is a clear and legible text. Slight and 'free-hand' as it is, Latheron lacks the design element of more monumental oghams such as Brandsbutt, Brodie, or Scoonie, but it is not in any way deficient as a piece of writing. Nor should it be considered as a piece of graffiti, its caiving is at least as substantial as die two horsemen at the bottom of the slab.

The stem wavers a little to fit round protruding bits of the design. The letters are well-spaced and successive letters of the same aicme are clearly separated. Padel suggested the apparent gap between 3 and 4 was meant to indicate word-division, but it is more likely to have contained strokes now lost. A more definite gap occurs before 7, but the possible b-stroke at the very edge of the fractured lower edge is very faint indeed. Since the text makes sense as it stands we should perhaps follow previous commentators in disregarding this stroke and taking the gap to be no more than a slightly longer than usual 'lead-in' on the stem (cf. the lead-in on Birsay 1 which is considerably longer). There are no bind-strokes. All the letters slope as one would expect; except for 2, the forward slope of which is not great. Though strokes are roughly parallel within letters no great attention has been paid to consistency of gradient between letters of the same aicme. Component strokes are mostly of equal length and are generally evenly spaced, however, between letters there is considerable difference between the more generously spaced (7, 5,6) and the more cramped (7, 75).

Vowels consist of long strokes occupying the whole width of the ogham-band, diere are also two vocalic forfeda. The 'hammer-head' A, which here occupies only the middle half of the ogham band, may be presumed to provide an orthographic contrast to the standard A, and is perhaps to be transliterated ai [see Formaston for fuller discussion of this character]. The supposedly angled strokes of 16 are doubtful due


LATHERON

368

to wear but the character cannot be a fully-fledged angled vowel since the first two strokes are straight. Burrian, which is even more cursive in appearance than Latheron, has two instances of the final stroke of a group being angled, but neither is a vowel. Unfortunately 12 is also doubtful due to wear. Macalister depicted it as a smooth S-shape across the stem, Padel drew a distinction between the curved lower hook and the straight-sided, angled, upper hook. There are two other instances of what is presumably the same character at Lunnasting and Lochgoilhead, though in both cases these are ordinary straight vowel-strokes with long straight serifs at their distal tips. Context suggests the Lunnasting example is a vowel, perhaps some kind of A, but the Lochgoilhead example is immediately followed by an A so perhaps some other explanation need be sought. This straight, seriffed, character appears on the first two Birsay inscriptions but in both instances as the middle of a group of five strokes, either R or I. The seriffed letter may have been meant to indicate some nuance of pronunciation, or merely to bring together visually the five strokes of the group and make the letter easier to read. The only other example of a truly curved S-shaped stroke is provided by Bressay which has one group of five, representing a vowel, presumably I. These Bressay strokes are, however, backward-sloping mirror images of the S-shaped stroke here.

Palaeographically, the Latheron ogham is, in general terms, closest, to Buckquoy, which too has a hammerhead A, and Lochgoilhead, which includes a 'seriffed' letter. Birsay 1 is also similar, though it has angled vowels. All of these inscriptions are notable for their cursive, 'free-hand1 appearance. The Buckquoy ogham is probably eighth century in date; another case of the hammer-head A, Formaston is probably later in that century or into the next at the earliest. Padel was uncomfortable with Curie's an historical dating of the cross to the seventh century, preferring, for palaeographical reasons, to place Latheron in the first half of the eight century [112]. I would be prepared to go later - even into the ninth century.

INTERPRETATION OF TEXT If 1 represents the beginning of the text then the ogham would have begun a little above the feet of the horses at the bottom of the slab. It is difficult to judge, given die current state of wear, but it seems as


LATHERON

369

if the completed design may have come quite close to the lower end of the border, in which case there may not have been room to start the ogham level with the bottom of the design. Even if a portion has been lost, it could have been little more than a letter or two. But in that case, why leave a gap before 1 ? The possible stray b-stroke is very faint, and on balance it seems far more likely to take 1 as the beginning of the text. At the other end, the surface is fractured in mid-letter and it is clear that a portion has been lost. If the upper rectangular figure is reconstructed as a full cross, then a further distance would have been available to the oghamist, though perhaps not such a great deal since the cross-arms would almost certainly have projected to the outer border.

The transliteration of the hammer-head A is not in much doubt, but the precise phonetic significance of the serifs on 12 is not clear. Other than that the only real difficulty concerns 3. Still we can be reasonably confident of the reading Dunnodnnaitma(i?)qqnet(u+)-.

Straightaway it can be seen diat Latheron

conforms to the formula X MAQI Y familiar from Ireland and seen on half-a-dozen Scottish oghams. The mainland Scottish examples are consistent in spelling mace with QQ, an unhistorical spelling which must reflect the influence of the old ogham formula. In Scotland the letter Q appears only in this formula word. Though standard in the consonants used, there is some variation in the vowel in the various examples of this word. Formaston and Altyre have the expected A, the more northerly examples, St. Ninian's Isle, Bressay, Golspie, have MEQQ (cf. ?MEQ Blackwaterfoot 1). Latheron is unique in using kforfid in this position. It could be simply an elegant variant on the standard A, but the inscription is otherwise lacking in flourishes, and elsewhere the use offorfeda seems to reflect a desire for orthographic precision [SimsWilliams 1992 passim]. If the Northern MEQQ represents a dialect difference, then perhaps theforfid in this position on the northerly Latheron, represents half-way house between die traditional 'A' and the pronunciation spelling ' E \ If a genitive is meant, the contemporary manuscript spelling would be make with the / indicating the palatal quality of the ee. The hammer-head A (representing ai ?) might be thought more appropriate in this position, but the reconstruction of its phonetic value is not beyond doubt. Still,


LATHERON

370

there can be no doubt that Latheron's 11-14 represents the Irish word 'son1.

Following it we have Netu* presumably the same element seen at Ackergill (NEHT-), Lunnasting (NEHHT-), Formaston (NEAHHT-). and Bressay (NAHHT-) [see Ackergill for a discussion of this name element].

The Lunnasting example, NEHHTONN, is the well-known Pictish male personal name

Nechtan/Neithon. Formaston's NEAHHT may be a shortened form, and Ackergill's NEHTETREB and Bressay's NAHHTVVDDAddS may be compounds using Necht- as afirstelement. Unfortunately we lack the conclusion of the Latheron text, but 18 already has too many strokes for the ending -ONN or -ANN to be reconstructed. If complete as it stands NETU may be a complete hypocoristic form of Nechton or a Necht- compound (cf. Inchyra's (s/n)ETU). One further stroke would give E, and suggest comparison with Ackergill's NEHTETREB. Latheron contrasts the other examples cited in not having an H(H) before the T. If this represents a real difference in pronunciation we may be dealing here not with a variation on the Necht theme, but instead with the Irish male personal name Neit, Néfi)d(e) [see Ackergill]. In discussing the Buckquoy ogham it was suggested that, in certain cases, Scottish oghamists followed the old convention of showing lenition with simplex consonants and non-lenition with double. Whereas T = /6/ would be a satisfactory explanation for 77, cf. Brittonic Neithon, it seems less likely for 10 where T appears to be a cipher for the manuscript t = /d/. Since t is a satisfactory explanation of 77, in line with the Gaelic Nechtan or Pictish Nation/Neiton, and we would expect an inscription to be internally consistent, it appears that Latheron is not following this old convention [Harvey 1987, where the author admits the convention was not universally applied even in classical ogham orthography]. Since Latheron is probably later than Buckquoy it is not surprising that its orthography should be more or less a caique on Olr. manuscript spelling.

Turning now to the beginning of the inscription, Anderson and Macalister read DUNNODNNAT, while Padel read DUV NODNNAT. Either form is acceptably Celtic. If the latter, then DU V may represent


LATHERON

371

a pronunciation spelling of Irish Dub /duv/. Dub 'black, dark' is a common element in (close-) compound names, either with its literal meaning, or more loosely as an intensifier [Uhlich 1993:231-4]. Dub was also in widespread use as the first element in loose-compound names in a similar way to Cú 'hound' and Mael 'devotee'. The second element was commonly a provincial or more local place-name (eg. Dub-Connacht)> or a personal name (eg. Dub-Chormacy or a common noun, often with toponymic significance (eg. Dubcluana), a parallel series wasformedon Dub-dá, 'Dub of the two . . / , (eg. Dub-dá-leithe, Dub-dá-thuatlu Dub-dá-Suibne\ cf. Cú-Ulad, Cú-Nuadau MáelSechnaill, Máel-topair [see CGH for further examples]. That such names were popular in Pictland is indicated by one Dub-Thalorg, king of Picts, and Dudabrach (for Dub-dá-brath ?), the father of the scribe of Meigle. If Dub is the preferred interpretation, then the second element NODNNAT, may, as Padel suggests, be compared with the Celtic divine figure Nodons, Nodontos, cf. Ir Nuadu W Nudd [O'Rahilly 1946:495-6], though as he admits, this is 'highly speculative'.

On balance, however, the preferred reading is not DUV NODNNAT, but DUNNODNNA(I)T, i.e. Dunodnait. Irish dun 'high place, fort' was a common element in compound male personal names, eg. Dúncath, Dúngus, Dúnlang [Uhlich 1993:235-6]. Latheron's Dunod- may be compared with Dimadach (Dúnad + adjectival ending -ach\ 'one who leads on campaigns' [Ó Corráin & Maguire 1990:80]. In this case, -NNA(I)T, could be the Irish diminutive suffix -natl-naiu commonly used to form a female version of a male personal name in -án, eg. OdránlOdornat, GobánlGobnat etc. By analogy at Latheron we may have Dúnad-nau That the name is followed by MAQQ implies that it is, however, masculine, but this need not be a problem in a Pictish context since the Pictish Gartnait is certainly male (one wonders if Irish redactors embarrassed by this apparently inappropriate name, might have emended the name of the father of the Pictish Cano, to Gartnán from Gartnait). As Thumeysen explains, though feminine in Old Irish, the suffix also appears with other genders, and was probably originally neuter becoming feminine 'as a result of its frequent use to denote females' [1946:174-5 §273.3]. Another possible example of -nait for a male Pict is the name of the father of the prehistoric king of Picts Drest, given in the P-Version of the


LATHERON

372

King-List as Munait (though Anderson compares it with a Pictish Moneit AU 728) [Anderson 1973:248]. Also, it is always possible that the parent named was the mother, cf. Nechtán son of Derile. Another possibility for DUNNOD- is a form of Irish DUNAIDONAS (< *Duna + -aidonas 'fiery1) which appears in ogham at Rathglass, Barony of Rathvilly, Co. Carlow (CIIC 16) [McManus 1991:117 §6.26]. In unstressed position the diphthong was reduced to /3/, (cf. manuscript Lugedon, Lugadon, for Lugaidon, and ogham BIVODON (CIIC 285) for BIVAIDONAS (CIIC 504) (all genitive), manuscript Béoáed [McManus 1991:121 §6.28; Uhlich 1993:177], so Dunnodnnait for Dunnáednait is perhaps possible.

The above interpretation is on the basis of Latheron being a Gaelic text. If MAQQ is taken as a purely stereotyped form without bearing on the language of the rest of the text, then it might be possible to explain the remainder as Brittonic. The element duna- is not restricted to Irish, of course. The precise cognate of Irish Dúncath ( Olr mitad. The adjective miiad means something like 'noble' or 'good' [DH] and appears as an epithet or independent name, MUAD, on the ogham-inscribed brooch from Ballyspellan, Co. Kilkenny (CIIC 27) [McManus 1991:132 §7.6, n.7onp.l81]. The element more commonly appears with the diminutive suffix -AGNI > -AN(N), i.e. Múadán, or in die feminine form Múadnat, names borne by a number of early Irish saints [Ó Corráin & Maguire 1991.139-140], A very early form, MODDAGN(I), appears on an ogham pillar from Windgap, Barony of Middle Third, Co. Waterford [CIIC 307], which McManus dates to the early fifth century [ 1991:94 § 5.24; 107 §6.12]. The element also occurs in the name of the river Moy in Connaught [ace. Mbdam in VC 1.6].

The putative second element, -MALI, is comparable with the very common personal name element MAGL- ( > Olr. mál 'prince'). Mál occurs uncombined as a personal name [DZL], but is normally found in compounds, in either first or second position [Uhlich 1993:275-6.]. The element is even more common


LOCHGOILHEAD

381

in Welsh than in Irish names, taking the form mael. In Irish it interchanges with máel 'bald, cropped* > 'devotee' [O'Brien 1973:229; O'Rahilly 1946:360 n.2]. That the element was also used by the Picts is shown by the name of the name of the father of hte Pictish king encountered by Columba, Meilochon. There is only one extant ogham example of the element, CUNAMAGLI ( > Olr. Conmál, Conmáet), Arbory, Isle of Man (CIIC 501) [McManus 1991 §§ 6.3-4, 6.10, 6.15, 6.21]. The remaimng examples are found on Roman letter inscriptionsfromWales: BROHOMAGLI ( > W. Brochmaet)fromLlandyssul, Cardigan [CIIC 349] andPentrefoelas, Denbigh (CIIC 401); CATOMAGfLI] ( > Olr. CathmáU Cathmáet) from Brawdy, Montgomery (CIIC 425); SENOMAGLIfromLlanfihangel Ioreth, Carmarthen (CIIC 370) [McManus § 5.23]; SENEMAGLI and VINNEMAGLI from Gwytherin, Denbigh (CIIC 400) [McManus 1991 § 6.26; 103 §6.4 n. 10]. Although there is no extant example of the two elements MOD/MUAD and MAGL-/MAL together, the combination is theoretically possible, perhaps with meaning 'noble prince'.

The problem with the above equation is that there is no G in the Lochgoilhead text. There may, however, be a precedent for such a spelling in the VEDOMALI of the Roman letter inscription from Margam, Glamorgan (CIIC 408) [McManus § 6.8]. The final -i suggests a genitive, but the Lochgoilhead inscription is surely far later than the accepted date of loss of final syllables (end of the sixth century), the roman alphabet inscription certainly cannot be this early. The final -I may represent the anachronistic persistence of the old epigraphic formula. A pseudo-Latin genitival ending is scarcely plausible in an ogham text, but it may be relevant to compare Artmali for Arthmail on the Latin inscription from Llantwit Major, Glamorgan, which dates from the second half of the eighth century (CIIC 1012).

Alternatively, if the fifth letter of the Lochgoilhead ogham is not to be read as M, another possibility is the element -V ALI, from the root *wal 'to be strong' (cf. Gaulish val(l), Olr. flaith 'lordship'), which is common in personal names in all branches of Celtic - Ir. CathaU Old Welsh Catgual, Gaulish Catuwalos; IT. DomnalU Old Welsh Dumngual etc. [McManus 104 §6.7, 178 n.22; Evans 1967.269-271; Pokorny


LOCHGOILHEAD

382

1959.1111-2]. There are several ogham examples of the element in first position - VALAMNI Rooves More, East Muskerry, Co. Cork (CIIC 125); VALUVI Ballyvellon, Middle Third, Co. Waterford (CIIC 302) - and in second position - TOTAVALI ( > Túathal, cf. Welsh Tudwat) from Llaosadyrnin, Carmarthen (CIIC 375) [McManus §§ 6.7, 6.28]; SUVALLOS, Ballintaggart, Corkaguiney, Co. Kerry (CIIC 158); 7SOVALINI, Drumlohan, Decies without Drum, Co. Waterford (CIIC 281); possibly CUNAVA[LI] Ballaqueeny, Isle of Man (CIIC 504) - and in a roman letter inscription, CVONOVALI (> Ir. Conall, W. Cynwal) on die stone known as Men Scryja, 'written stone1, in the parish of Madron, Cornwall (CIIC 468). Against this hypothesis is the fact that the V sound is adequately catered for within the standard ogham alphabet. The third letter of the b-aicme, transliterated V, represents /w/ > Olr. /f/, Welsh /gw/ [McManus §3.15, p.36]. Unless the oghamist was trying to represent some nuance of pronunciation, there seems no reason to invent or employ the flapped letter.

Returning to the hypothesis that letter five may represent a modified A, the reading Mudáli is possible. MU appears as an abbreviation for MUCOI (moccu) on an ogham pillar from Coumeenoole North, Corkaguiney, Co. Kerry (CIIC 178). Might MUDALI be MU(COI) DALI ? Ogham examples of the name Dála are - SIDANI MAQI DALO Rathduff Corkaguiney, Co. Kerry (CIIC 194); DALAGNI MAQI DALI Monataggart, East Muskerry, Co. Cork (CIIC 119) - and it appears as DALLUS DUMELUS on a Roman letter inscriptionfromLlanddewi Brefi, Cardigan (CIIC 35 l.II). The reading MUDALI, can be compared with the MEDALO which appears to occur on the ogham pillar CIIC 279, one of ten from a souterrain in Drumlohan, Decies without Drum, Co. Waterford, DENAVEQA MU]COI MEDALO [McManus § 6.17]. MEDALO represents the sept name Mocu-Dalon, descendent of Dála. The most famous scion of die Dál Mo Dala was St Cainnech of Aghaboe, who visited Columba on Iona [VC III. 17].

DISCUSSION The apparently bewildering number of possible interpretations of die Lochgoilhead ogham (and MOD-


LOCHGOILHEAD

383

MAGLI, MOD-VALI, MEDALI, and MU-DALI doesn't exhaust all the possibilities) should not be misconstrued. The confusion arises because of doubts over the marks at the two ends of the text, over the value of the fifth character, and the difficulty in knowing how much orthographic latitude to allow in determining a reading. The problems of exegesis should not, however, obscure the fact that the ogham is carved with some flair. The ogham letters are certainly no more problematic than the roman ones.

An understanding of how the three texts relate to one another would be of great assistance in interpreting the ogham text. The letters of the ogham inscription are more lightly and finely cut than the roman, but this may be a function of the more laboured technique required for the minuscule curves in comparison with the simple ogham strokes. The juxtaposition of ogham and roman letter inscriptions on the one stone is not unique. To the famous 'Colman bochf stone from Clonmacnoise (CIIC 749) can now be added the Dupplin Cross and, of course, the Newton Stone. Furthermore the sites of Dunadd and St. Ninian's Isle, Shetland have produced separate inscriptions in each script.

If the style of the roman letters could be closely dated it would provide, at the very least, a terminus post quern for the ogham. The pronounced horizontal back of the D on face (b) is distinctive, but the absence of a G, often chronologically diagnostic, makes it difficult to date the lettering more precisely than probably seventh or eighth centuries, or even later [see fig.].

As stated at the beginning, it is difficult to imagine the original purpose of the Lochgoilhead stone. The aforementioned explanation - a stone selected for fine work, abandoned, then found convenient for practicing lettering - is no more implausible than any other. The RCAHMS suggest either a 'didactic or liturgical purpose1 for the alphabetic inscription. Exactly what they mean by 'liturgical1 is not specified and if the carving was didactic, surely it must have been intended for teaching stone-cutters not scholars. If portable wax tablets or sheets of bark were available why wrestle with such a trying medium as this to impart the ABC, and why go only as far as N ? Such alphabets may have been intended to display


LOCHGOILHEAD

384

participation in the prestige technology of literacy. Afine,complete, latin alphabet appears in a secondary position on the possibly sixth century cross-marked pillar at Kilmalkedar, Co. Kerry (CIIC 913) [Cuppage 1986.311, fig. 184 and pl.30]. Perhaps more directly comparable are the incomplete alphabetic sequences which occurs on the stone from Llandecwyn, Merionethshire (CIIC 1029) and possibly on the Dupplin Cross.

Though there is no evidence beyond this stone and, possibly, the millstone, it seems reasonable to deduce some form of ecclesiastical activity at Lochgoilhead in the early medieval period. Though Lochgoilhead was very close indeed to the territory of the Strathclyde British, it was in the territory of the Cenél nGabráin of Dál Riada and ties are to be sought with the Gaelic church. There is nothing about the inscription which would link it directly to Iona. Most of the inscribed slabsfromIona have rather uprights Ds, the closest parallel to the Lochgoilhead horizontal D is No.47, dated by Jackson to late 8th or 9th century [RCAHMS 1982:187-6, ECMS 400]. Its D, however, is not as tightly curled, nor the 'ascender' as long or horizontal. More numerous, and far closer, parallels are to befoundat Clonmacnoise, e.g. the 'or do thuathal saer' slab (CIIC 1087).

BIBLIOGRAPHY OS record card - NN 10 SE RCAHMS 1992:194, No. 87 (1) figs, a-b

PREVIOUS READINGS None

LOCHGOILHEAD Faces * and c fphotograph RCAHMS Argyll. VII.194 B,CJ

f S ^ j V ^ i i JMiJi'i I «** *•»& ****** í^*»*^* 4 **** **«*

;.v •'•«•>' 'Ti •!">**•. *£•». S»**Vo I

The ogham inscriptions of Scotland: An edited corpus

Corpus of inscriptions of the goddess Hestia. Ph.D. dissertation

Corpus of Hieroglyphic Luwian Inscriptions: Vol. 1. Inscriptions of the Iron Age: Part 3. Plates

Corpus of Hieroglyphic Luwian Inscriptions: Vol. 2. Karatepe-Arslantaş: The Inscriptions: Facsimile Edition

Inscriptions

The History of Scotland

Ancient Hebrew Inscriptions: Volume 2: Corpus and Concordance

Scotland

English Corpus Linguistics: An Introduction

Corpus

Guardsmen Of Tomorrow Edited

Corpus

Christian Inscriptions

The Political Economy of Scotland

The Last King of Scotland

The Empire of the Nine (Omnibus edited)

The Syntax of the Boeotian Dialect Inscriptions

The Syntax of the Boeotian Dialect Inscriptions

The Last King of Scotland

Semitic Inscriptions

The Feast of Corpus Christi

A History of Scotland

A History Of Scotland

A History of Scotland

Portrait of Scotland

History of Scotland

In Search of Scotland

A History of Scotland

A History Of Scotland

History of Scotland

The ogham inscriptions of Scotland: An edited corpus

Corpus of inscriptions of the goddess Hestia. Ph.D. dissertation

Corpus of Hieroglyphic Luwian Inscriptions: Vol. 1. Inscriptions of the Iron Age: Part 3. Plates

Corpus of Hieroglyphic Luwian Inscriptions: Vol. 2. Karatepe-Arslantaş: The Inscriptions: Facsimile Edition

Inscriptions

The History of Scotland

Ancient Hebrew Inscriptions: Volume 2: Corpus and Concordance

Scotland

English Corpus Linguistics: An Introduction

Corpus

Guardsmen Of Tomorrow Edited

Corpus

Christian Inscriptions

The Political Economy of Scotland

The Last King of Scotland

The Empire of the Nine (Omnibus edited)

The Syntax of the Boeotian Dialect Inscriptions

The Syntax of the Boeotian Dialect Inscriptions

The Last King of Scotland

Semitic Inscriptions

The Feast of Corpus Christi

A History of Scotland

A History Of Scotland

A History of Scotland

Portrait of Scotland

History of Scotland

In Search of Scotland

A History of Scotland

A History Of Scotland

History of Scotland

Recommend Documents