adiego carian language

8/20/2019 Adiego Carian Language

http://slidepdf.com/reader/full/adiego-carian-language 1/540



The Carian Language



HANDBOOK OF ORIENTAL STUDIES

SECTION ONE

THE NEAR AND MIDDLE EAST

Ancient Near East

Editor-in-Chief W. H. van Soldt

Editors

G. Beckman • C. Leitz • B. A. Levine

P. Michalowski • P. Miglus

Middle East

R. S. O’Fahey • C. H. M. Versteegh

VOLUME EIGHTY-SIX



The Carian Language

by

Ignacio J. Adiego

BRILLLEIDEN • BOSTON

2007

with an appendix by

Koray Konuk



ISSN 0169-9423ISBN-10 90 04 15281 4

ISBN-13 978 90 04 15281 6

© Copyright 2007 by Koninklijke Brill NV, Leiden, The Netherlands.Koninklijke Brill NV incorporates the imprints Brill Hotei Publishers,

IDC Publishers, Martinus Nijhoff Publishers, and VSP.

All rights reserved. No part of this publication may be reproduced, translated,stored in a retrieval system, or transmitted in any form or by any means,electronic, mechanical, photocopying, recording or otherwise, without prior

written permission from the publisher.

Authorization to photocopy items for internal or personal use is grantedby Brill provided that the appropriate fees are paid directly to

The Copyright Clearance Center, 222 Rosewood Drive, Suite 910,Danvers, MA 01923, USA.Fees are subject to change.

This book is printed on acid-free paper.

Library of Congress Cataloging-in-Publication Data

Adiego Lajara, Ignacio-Javier.The Carian language / by Ignacio J. Adiego ; with an appendix by Koray Konuk.p. cm. — (Handbook of Oriental studies. Section 1, The Near and Middle

East ; v. 86).Includes bibliographical references.ISBN-13 : 978-90-04-15281-6 (hardback)ISBN-10 : 90-04-15281-4 (hardback)1. Carian language. 2. Carian language—Writing. 3. Inscriptions, Carian—Egypt.

4. Inscriptions, Carian—Turkey—Caria. I. Title. II.

P946.A35 2006

491’.998—dc22 2006051655



Günter NeumannIn memoriam



CONTENTS

Foreword ........................................................................................ xiAcknowledgments for the Use of Illustrations ............................ xiii

Chapter One Introduction .......................................................... 1

Chapter Two The Indirect Sources .......................................... 7A. The Glosses .......................................................................... 7

1. Glosses und Pseudo-glosses .................................................... 72. Interpretation ........................................................................ 10

B. The Proper Names .............................................................. 121. The Unity and Continuity of Anatolian Onomastics .............. 122. Present Compilations of Carian and Anatolian Names ............ 143. Carian Names from Indirect Sources vs. Those from Direct

Sources ................................................................................ 15

Chapter Three The Inscriptions ................................................ 17A. Introduction .......................................................................... 17

1. The Revised System of Transcription of Carian Letters .......... 182. Vocalism .............................................................................. 183. Consonantism ...................................................................... 19

B. ‘Para-Carian’ or ‘Caroide’ Inscriptions .............................. 221. ‘Para-Carian’ Inscriptions from Caria .................................... 222. ‘Para-Carian’ Inscriptions from Other Places .......................... 233. The Ostrakon of Hou and the Naukratis Fragment ................ 264. Carian Gra fi tti from Sardis .................................................. 275. Carian Inscription from Old Smyrna .................................... 29

C. The Carian Inscriptions from Egypt .................................. 301. Sais (E.Sa) ........................................................................ 322. Memphis (E.Me) ................................................................ 343. Abydos (E.Ab) .................................................................... 794. Thebes, Tomb of Montuemhat (E.Th) .................................. 95

5. Luxor Temple (E.Lu) .......................................................... 1066. Murw àw (E.Mu) ................................................................ 1097. Silsilis (E.Si) ...................................................................... 1108. Abu Simbel (E.AS) .............................................................. 115



viii

9. Buhen (E.Bu) .................................................................. 11910. Gebel Sheik Suleiman (E.SS) ............................................ 123

11. Unknown Origin, Likely from Egypt (E.xx) ........................ 124D. The Carian Inscriptions from Caria .................................. 128

1. Tralleis (C.Tr) .................................................................. 1302. Alabanda and Surroundings (C.Al) ...................................... 1323. Euromos (C.Eu) ................................................................ 1324. Kindye (C.Kn) .................................................................. 1345. Hyllarima (C.Hy) ............................................................ 1356. Mylasa (C.My) ................................................................ 1377. Sanctuary of Sinuri near Mylasa (C.Si) .............................. 1388. Kildara (C.Ki) .................................................................. 1419. Stratonikeia (C.St) ............................................................ 142

10. Halikarnassos (C.Ha) ........................................................ 14411. Didyma (Ionia, near Milet) (C.Di) .................................... 14512. Iasos (C.Ia) ...................................................................... 14513. Keramos (C.Ke) ................................................................ 15014. Kaunos (C.Ka) .................................................................. 15115. Krya (C.Kr) ...................................................................... 158

16. Inscriptions of Unknown Origin, Presumably from Caria(C.xx) .................................................................................. 159E. The Carian Inscriptions from Greece ................................ 164

Chapter Four The History of the Decipherment .................. 166A. The ‘Semisyllabic Era’ (1887–1962) .................................. 166B. The ‘Greek Alphabetic’ Era ................................................ 176C. The ‘Egyptian Approach’ .................................................... 187

1. The First Attempts ............................................................ 1872. The Seminal Work of Ray ................................................ 191

...... 197

Chapter Five The Carian Alphabet .......................................... 205A. Alphabetic Varieties ............................................................ 205

1. Alphabetic Varieties of Caria Proper .................................... 2062. Inscriptions from Continental Greece .................................... 2193. Egyptian Alphabets ............................................................ 219

4. The Classi fi cation of the Alphabets of Caria Proper .............. 2235. The Relationship between the Alphabet from Egypt andthe Local Alphabets from Caria Proper .............................. 226

6. The Common Origin of the Carian Alphabetic Varieties ........ 228B. The Origin of the Carian Alphabet .................................. 230

3. The De fi nitive Decipherment (‘Ray-Schürr-Adiego System’)



ix

Chapter Six Phonological Features ............................................ 234A. The Phonological System .................................................... 234

1. Vowels and Semivowels ........................................................ 2342. Consonants .......................................................................... 2423. Letters of Uncertain Value .................................................... 2514. Letters of Unknown Value .................................................... 2535. Phonotactics ........................................................................ 254

B. Overview of the Historical Phonology of Carian ............ 2561. Vocalism ............................................................................ 2572. Consonants .......................................................................... 2593. Some Secondary Changes ...................................................... 262

Chapter Seven Analyzing Carian Inscriptions .......................... 264A. Basic Onomastic Formulae .................................................. 264

1. Inscriptions Consisting of Only an Individual Name .............. 2642. Inscriptions Consisting of Only a Twofold Onomastic

Formula .............................................................................. 265B. The Structure of the Stelae from Memphis ...................... 267

1. Threefold Formulae .............................................................. 267

2. Stelae for Women ................................................................ 2713. Inscriptions with Ted and En ............................................ 2734. Other More Complex Funerary Inscriptions ............................ 2755. The Rest of the Inscriptions from the Memphis Corpus .......... 2766. A First Summary ................................................................ 279

C. Analyzing Brief Inscriptions ................................................ 2801. Inscriptions on Objects .......................................................... 2812. Funerary Inscriptions of Caria and Athens ............................ 2873. The Longest Gra ffi to from Abu Simbel (E.AS 7) .................. 293

D. The Longer Inscriptions ...................................................... 2941. The Kaunos Bilingual Inscription .......................................... 2952. The Kaunos Inscription C.Ka 2 ............................................ 3013. Sinuri’s Longer Inscription (C.Si 2) ...................................... 3024. The Hyllarima Inscription (C.Hy 1) .................................... 3055. Other Inscriptions from Caria .............................................. 308

E. Summary and Some Controversial Questions .................. 310

Chapter Eight Morphological Features ...................................... 312A. Nominal Inflection .............................................................. 3121. Nominative Singular ............................................................ 3122. Accusative Singular .............................................................. 313



x

3. Genitive Singular .................................................................. 3144. The -s Ending and the Problem of Carian Datives ................ 314

5. Other Possible Datives .......................................................... 3176. Nominative Plural ................................................................ 3187. Accusative Plural .................................................................. 3188. Other Possible Case Endings ................................................ 319

B. Pronominal Inflection .......................................................... 319C. Verbal Inflection .................................................................. 321

Chapter Nine The General Vocabulary and the ProperNames .......................................................................................... 326A. General vocabulary .............................................................. 326B. Proper Names ...................................................................... 328

1. Theophores .......................................................................... 3312. Some Nominal Stems ............................................................ 3333. Verbal Stems ...................................................................... 3394. Adverbs .............................................................................. 3395. Lallnamen .......................................................................... 3406. Su ffi xes ................................................................................ 341

Chapter Ten Carian as an Indo-European AnatolianLanguage .................................................................................... 345

Chapter Eleven Carian Glossary ................................................ 348

AppendicesA. Carian Inscriptions in Transcription .................................. 443B. Carian Glosses ...................................................................... 455C. Carian Names in Greek Sources ........................................ 456D. Concordances ........................................................................ 464E. Coin Legends in Carian (by K. Konuk) ............................ 471

Abbreviations and Bibliography .................................................... 493

Table I: The Carian Alphabet .................................................... 508Table II: Carian signs in coin legends ........................................ 509

Index .............................................................................................. 511Plates .............................................................................................. 519Maps .......................................................................................... 521Coins .......................................................................................... 523



FOREWORD

Fifteen years after my doctoral dissertation on Carian (Studia Carica.Investigaciones sobre la escritura y lengua carias, y su relación con la familia lingüística anatolia indoeuropea , University of Barcelona, 1990), directed byPere J. Quetglas and supervised by Jürgen Untermann, and twelve years after my book Studia Carica (Barcelona 1993), a revised edition of my dissertation, I now off er a new work on the Carian script and lan-guage. This is not an English translation of the above titles. In fact,very little remains of the structure and content of these former works,mainly due to three, closely interrelated reasons. Firstly, both my dis-sertation and the subsequent book were conceived to defend a newproposal for deciphering Carian, beginning with the fundamental workof the British Egyptologist John D. Ray, which, in the 1980s, markeda dramatic breakthrough in the history of research on Carian. My aimat the time was to further develop the decipherment, after noting that

Ray’s system, although certainly well founded, was by no means totallysatisfactory. Now, following the discovery of the bilingual inscription of Kaunos in 1996, the system proposed in my book Studia Carica hasbeen proved correct and is accepted—with slight modifications—as thestandard tool for transcribing the Carian signs. Therefore, a new bookon Carian no longer needs to off er a ‘combative’ justification for a par-ticular decipherment system, but should rather take the definitive deci-pherment as a starting point for the analysis of Carian texts. Secondly,the corpus of Carian inscriptions has been augmented by the appear-ance of several new texts, among which the Carian-Greek bilingual of Kaunos mentioned above is undoubtedly the most relevant. In addi-tion, some other inscriptions that were already known have subsequentlybeen revised, yielding new possible interpretations. Thirdly, the newdecipherment encouraged many scholars to apply themselves to thestudy of Carian, so that although many uncertainties remain and ourknowledge of Carian language continues to be very limited, importantprogress has been made in recent years, and this must now be incor-

porated into an updated analysis of the subject.I wish to repeat here my most profound gratitude to those thathelped me when I began to work on Carian, particularly the afore-mentioned Pere J. Quetglas and J. Untermann, without whose help



xii

and encouragement I could not have undertaken a doctoral disserta-tion on a topic that was then beset by so many risks. I am also grate-

ful to Theo van den Hout, who invited me to prepare this book forBrill; to Brill for accepting this project; to Koray Konuk, for giving methe opportunity to substantially improve this book with an appendixon Carian coin legends; and to Wolfgang Blümel, H. Craig Melchertand Diether Schürr, not only for their kind assistance with some of theproblems that arose during the preparation of this work, but also forthe consistently productive and exciting interchange of ideas aboutCarian. Georg Rehrenböck (Kleinasiatische Kommission, Österreichi-sche Akademie der Wissenchaften) also deserves a special mention forhis extreme kindness and unfailing generosity in answering all of myqueries about copies of Carian inscriptions conserved in Vienna.

I am equally indebted to Peter Cottee for his accurate revision of my imperfect English. Needless to say, all the possible errors and omis-sions are of my own doing.

On a more personal level, I must express my gratitude to my wife,Anna, and to my daughters, Alba-Artemísia and Laura-Neït, for bear-ing with such patience the long period during which the book was put

together.This book is dedicated to the memory of Günter Neumann: for a

decipherment to be successful it must not only be correct, but alsocredible and convincing, and he showed me how to achieve this withhis open-minded consideration of ideas that questioned the prevailing communis opinio on Carian, his astuteness in refraining from excessivespeculation, his numerous suggestions of improvements, his discreet butvery eff ective work in favour of the new decipherment, and his adviceto wait patiently for a proposal marked with the signum veritatis (to quotehis own expression) to finally achieve general acceptance.



ACKNOWLEDGMENTS FOR THE USE OF ILLUSTRATIONS

Drawings of inscriptions come from O. Masson, Carian Inscriptions from North Saqqara and Buhen (Masson 1978), and are reproduced by kindpermission of the Egypt Exploration Society.

Drawings taken from O. Masson-J. Yoyotte Objets pharaoniques à inscrip-tion carienne (Masson-Yoyotte 1956) and O. Masson “Remarques sur lesgraffites cariens d’Abou Simbel”, Hommages à la mémoire de S.Sauneron II (Masson 1979) are reproduced with the permission of theInstitut français d’archéologie orientale.

Drawings of graffiti published in The Epigraphic Survey, Reliefs and Inscriptions at Luxor Temple II, The facade, Portals, Upper Register Scenes,Columns, Marginalia, and Statuary in the Colonnade Hall (ESS 1998) arereproduced by courtesy of the Oriental Institute of the University of Chicago.

Drawings extracted from L. Deroy “Les inscriptions cariennes de

Carie”, L’Antiquité Classique 24 (Deroy 1955) are reproduced with per-mission of the editors of the journal.



CHAPTER ONE

INTRODUCTION

It is not clear when Caria and the Carians enter into ancient History.This is dependent on equating classical Caria with the land of Karkiya/Karkisa mentioned in Hittite sources. This supposition, eminently suit-able from a purely linguistic point of view (karkº in Karkisa , Karkiyais practically identical to the Old Persian word for ‘Carian’, k º ka- ), iscomplicated by the uncertainties regarding the exact location of Karkisa/Karkiya on the map, a problem intimately bound to the complex ques-tion of Hittite geography, a topic still subject to controversy despite thegreat progress made in recent years.

In any case, no information about the language of the land of Karkiya/Karkisa can be obtained from Hittite sources, so that even if the equation could be confirmed, its value for the study of the Carian

language would be very limited. The only relevant (but overly generic)datum is that Karkiya/Karkisa is a land located in the western regionof Anatolia, an area occupied by Luwian population groups, and thusconsistent with the clear similarities between Carian and Luwian, Lycianand other Indo-European Anatolian dialects that can now, since thedecipherment of Carian, be clearly traced (see Chapter 10).

Classical Caria, the country situated in western Anatolia betweenLydia and Lycia, must therefore be the starting point of the researchon Carian language. It is during this period that we find both direct

documentation of Carian and a wealth of information about this landand its inhabitants in indirect, mostly Greek, sources.

Particularly meaningful are the consistent ties that we can establishbetween various types of records on Carian and the Carians regard-ing one of the most remarkable characteristics of Carian language doc-umentation: the fact that the greatest number of Carian inscriptionshave been found in Egypt, and not in Caria itself.

From Greek sources, we know that Carian and Ionian mercenaries

were employed by the pharaoh Psammetichus I (664–610) for consol-idating his throne (Herod. I, 151). According to Herodotus, these mer-cenaries were based in the Delta area, near Bubastis (Herod. II, 154).It is no coincidence then that the oldest datable Carian document from



2

Egypt, a bilingual inscription on a statuette of the goddess Neith (E.Sa2) from the times of Psammetichus I, can be traced to Sais, another

Egyptian city situated on the Delta.A further connection can be drawn between Carian documents and

historical facts under the reign of the grandson of Psammetichus I,Psammetichus II (595–589): a well-known Greek graffito from AbuSimbel attests the participation of foreign mercenaries in his Nubiancampaign (593/592, see below p. 31 for this dating), and this infor-mation is consistent with the existence of Carian graffiti in Abu Simbeland in other locations further to the south (Buhen, Gebel el-Sheik el-Suleiman).

Under Amasis (568–526), Ionian and Carian settlements were movedfrom the Delta to the city of Memphis (Herod. II, 154), where a ‘Carianquarter’ (KarikÒn ) and a ‘Greek quarter’ (ÑEllhnikÒn ) existed for many years (Aristagoras of Milet apud Stephan of Byzantium, s. v. ÑEllhnikÒn ).Once again this event can be linked to epigraphical evidence: the mostimportant sub-corpus of Carian inscriptions is the collection of funer-ary stelae found in Saqqâra, one of the necropoleis of Memphis. Asfor the rest of Carian inscriptions found in Egypt—mostly graffiti—

from Thebes, Abydos, Silsilis, etc.), no connections can be establishedwith historical facts, and we can only assume that they are the marksof Carian visitors, similar to Greek graffiti found in these and anotherparts of Egypt.

Caria itself does not off er such striking results. The Carian inscrip-tions found in Caria are far less numerous than those from Egypt,come from diff erent cities, and appear more heterogeneous, both incontent and in form, thus constituting a very fragmentary and incom-plete view and lacking a clear connection with historical facts.1 In fact,the sole inscription that gives any indication of a link to the history of Caria is the bilingual inscription from the temple of the god Sinuri(C.Si 2), which can confidently be interpreted as a decree enacted bythe Carian dynasts of the Hekatomnid era, Idrieus and Ada, whose joint reign is dated in the period 351/350–344/343. But not even thisinscription has any real implications for Carian history: it is simply partof a wider corpus of regulations of a local syngeneia—mostly in Greek— produced by the satrapal couple. In the case of another of the most

1 This is not the place for a history of Caria. I refer the reader to Hornblower(1982).



3

important Carian inscriptions from Caria itself, namely the bilingualproxeny decree of Kaunos (C.Ka 5), dated at the end of the IV cen-

tury BC, the two Athenian citizens honoured as proxenoi are not clearlyidentifiable with any figures found in classical sources. The new inscrip-tion from Hyllarima (C.Hy 1b), and perhaps also the inscription recentlyfound in Mylasa (C.My 1), off er lists of priests that can be linked tolocal cults. The two texts from Stratonikeia also contain lists of names,but in these cases their religious character cannot be confirmed. Nothing can be ascertained about the exact content of the inscription fromKildara (C.Ki 1) or the longest Carian inscription from Kaunos (C.Ka2), but their official character is beyond doubt. This heterogeneous cor-pus is completed by a number of funerary, and therefore private, inscrip-tions and by others about whose content and functionality we knownothing. The absence of a precise dating for most of these texts fur-ther compounds the difficulties in placing them in a definite context.

If a conclusion must be drawn exclusively from the present corpusof Carian inscriptions from Caria, we must conclude that written useof Carian was very limited and confined to certain local communities,mainly linked to religious cults and to private funerary contexts: in this

latter case there are so few examples that it would be tenuous to makea comparison with the much richer documents of neighbouring Lycia.The only notable exception to this extreme scarcity of documentationis Kaunos, where a small, but very significant corpus of inscriptionshas been established, thanks particularly to the archaeological excava-tions carried on in the last forty years. But this is an exception only if considered in relative terms compared to the rest of the documenta-tion. In absolute terms, Kaunos provides us with a very a modest col-lection of Carian inscriptions.

It is astonishing that we have at our disposal only a single, reliableexample of the use of Carian as a co-official language (alongside Greek)in the Hekatomnid period: the aforementioned decree from Sinuri, C.Si2. This is particularly surprising since it is generally believed that dur-ing the activity of the Hekatomnid dynasty, a ‘Karianization’ (to useHornblower’s word) took place, so that “Mausolus and his family, them-selves native Karians, encouraged the institutions (such as the koinã ),the cults, and the language of Karia” (Hornblower 1982:352; the empha-

sis is mine). The Carian documentation found to this day in no waysupports this view. It is true that this current interpretation of the writ-ten use of Carian must be viewed with some caution, since it is pos-sible that the lack of documentation is in fact a result of our insufficient



4

archaeological knowledge of the land, and the perishable materials onwhich many inscriptions were written. Moreover, both the hypotheses

on the origin of Carian letters envisaged here (p. 231) and the stronglydiff erentiated local alphabetic variants point clearly to a much widerand prolonged use of Carian script. However, at least in the case of the alleged importance of the Carian language as a tool for ‘Karianization’in the Hekatomnid period, there is compelling counter-evidence: thetotal disappearance of Carian letters, and their replacement by singleGreek legends, from the well-documented coinage of the Hekatomnidsfrom Hekatomnos onward,2 which would seem to disprove the theoryof the co-official status of Carian.

Carian belongs to the Indo–European family of Anatolian languages,which also comprises Hittite, Cuneiform and Hieroglyphic Luwian,Palaic, Lycian, Lydian and lesser known Sidetic and Pisidian. Moreprecisely, it forms part of the so-called ‘Luwic’3 group. This classifica-tion of Carian has already been proposed by several scholars—notablyV. V. ”evoro“kin—on the basis of indirect documentation affirmed inClassical sources, mostly Carian proper names that showed strong con-nections with Anatolian. But this is now clearly confirmed by direct

sources, now that the Carian alphabet has been deciphered, and deci-sive linguistic evidence has been obtained from Carian texts.

This book is intended to show the new look of Carian following thedecipherment of the Carian writing system. My aim is to off er as com-plete and updated a view as possible of our present knowledge of Carian, paying special attention to existing documentation, the Carianalphabet, its origin and variants, and to the still developing linguisticinterpretation of the materials and its repercussions for the classificationof Carian as an Anatolian language. In Chapter 2, a brief summaryof the indirect sources on Carian is outlined. Chapter 3 is devoted tothe presentation of the entire corpus of Carian inscriptions available to

2 On Hekatomnids’ coinage, see the decisive work of Koray Konuk (Konuk 1998a).3 From now on, I shall use the term ‘Luwic’ for this group of dialects, following

the suggestion made by Melchert, as a convenient and non-confusing form to refer toa series of Anatolian dialects that share important diff erentiating issogloses but whose

exact internal relationship—originating from a common branch, or rather the resultof an areal convergence—is still debated (see Melchert 2003:176). The former use of ‘Luwian’, both for two dialects (Hieroglyphic and Cuneiform) and for the entire group(embracing also Lycian, Carian, Sidetic and Pisidian), could lead to confusion and mustbe therefore abandoned.



5

me at the time this book was written.4 Only the coin legends are notdiscussed there, since they are the subject of a specific appendix, kindly

prepared by Koray Konuk. Chapter 3 does not constitute an epigraphicaledition. The inscriptions are simply intended to be a useful tool withwhich to tackle the following chapters and the analysis of Carian writ-ing and language in general, hence the inclusion of drawings and theobservations about reading problems.

Chapter 4 off ers a general history of the decipherment of Carian. Itis based mostly on the corresponding chapter of my Studia Carica (Adiego1993a), but has been expanded to include a succinct exposition of thedecisive progress of decipherment, accomplished during the final decadeof the 20th Century.

Chapter 5 deals with the Carian alphabet, off ering an analysis of the diff erent local alphabetic variants and some reflections on the pos-sible origins of this extremely peculiar writing system.

Chapters 6 to 11 focus on linguistic aspects. Chapter 6 off ers anoverview of Carian phonology from a synchronic and, insofar as is pos-sible, diachronic perspective. Chapter 7 introduces the possibility of analysing a great number of Carian inscriptions, from the briefest and

most transparent, to the more extensive, wherein the difficulties of inter-pretation are practically insurmountable. Chapter 8 discusses the (scarce)morphological traits that have so far been identified in Carian. Chapter9 is of a lexical nature: an inventory of all the Carian common wordsto have been identified is put together and analyzed, and the same isdone for the proper names, in this latter case in the context of Anatolianonomastics. As a means of concluding the study, chapter 10 presentsthe evidence that suggests Carian can be classified in the group of Anatolian ‘Luwic’ dialects. Finally, Chapter 11 provides a glossary of all the forms shown in Carian inscriptions, inspired by similar workssuch as the Lydisches Wörterbuch by Roberto Gusmani, or the more recent Dictionary of the Lycian Language by H. Craig Melchert. The book isaccompanied by five appendices: an editio minor of the inscriptions intranscription, the collection of glosses and a list of proper names foundin Greek sources, a table of concordances with other editions of Carian

4 The only unavailable source is a Carian inscription found in Greece and pub-lished some years ago (see G. Neumann, “Epigraphische Mitteilungen—Kleinasien” in:Kadmos 39, 2000:190). Despite my eff orts, it was impossible to obtain when prepar-ing this book.



6

inscriptions, and a special section by Koray Konuk on Carian coinswith Carian legends, as mentioned above. The bibliography that com-

pletes the book is intended to be an exhaustive collection of all publi-cations on Carian language and inscriptions from Sir Archibald Sayceto the present, together with other works not specifically devoted toCarian, but that are cited thoroughout this study.

Despite my best eff orts, it is inevitable that a study of a languagethat is only partially understood, and surrounded by uncertainties, willcontain a number of provisional results, fragile hypotheses and evenmere speculations, not necessarily shared by other scholars. This willnot come as a surprise to seasoned specialists in the study of Trümmer-sprachen, but I consider it necessary to advise scholars of related disci-plines against taking all of the interpretations included in my work tobe demonstrated or indeed definitive. Much more will have to be done,and much more new material will have to be made available beforewe are able to paint a confident picture of this language. In any case,I have tried at all times to separate the assertions presently consideredplausible by the communis opinio (for example, the overall validity of thedecipherment system Ray–Schürr–Adiego) from those interpretations

that rest on more fragile indices, and which are shared by only a hand-ful of scholars, or indeed those hypotheses that are my own creation.I have tried where possible to avoid resorting to overly hypotheticalexplanations, and to steer clear of long, confusing paragraphs of a crit-ical nature, which also explains the limited use I have made of worksby other scholars, when their results seemed to me excessively tenuousor somewhat premature. Only in the glossary, for the sake of com-prehensiveness, have I included some interpretations suggested for diff er-ent Carian words that I don’t share or that I find too speculative.

Proposed in this study is a new system of classification for Carianinscriptions, which is intended to be more rational than previous off erings,and also some modifications of the transcription system of Carian signs(the reasons for this are explained in pp. 18–20). I hope that theseinnovations will not create any difficulties for the reader. For moreclarity, Appendix C (Concordances) can be used to find the equiva-lences between the new classification system and the older ones, andTable I at the end of the book shows the transcription system of Carian

that is used here.



CHAPTER TWO

THE INDIRECT SOURCES

A. T G

“Sur le chapitre des gloses, on peut passer assez rapidement. Le matérielest d’une valeur très médiocre” (Masson 1973:190).

Olivier Masson’s categorical statement is absolutely true. There arevery few Carian glosses, and the possibilities of connecting them bothto the direct documentation of the Carian language and to the com-mon Anatolian lexicon are very limited or simply non-existent. Despitethis, we will examine this modest corpus.

1. Glosses and Pseudo-glosses

The first problem that must be addressed is the distinction betweentrue Carian glosses and merely fictitious forms. From the time of thefirst compilation of scientific interest—Sayce (1887[92]:116–120)—to themost recent—by Dorsi himself (Dorsi 1979)—the inventory of Carianglosses was artificially expanded by invalid entries until Dorsi significantlyreduced it. Sayce off ered a total of 21 glosses, which Dorsi then short-ened to six definite ones, and three doubtful, adding three relevantnames of gods. Between Sayce and Dorsi, in Brandenstein (1935a) thisartificial expansion of the number of glosses reached its peak: Brandenstein

lists more than sixty words under the title “Wörter, Glossen, usw.”I have analysed elsewhere the methodological errors that led to this

situation (Adiego 1992b). Here we need simply remember that a sourceof tremendous confusion was the lack of diff erentiation in Brandenstein(1935a) between actual glosses and a modern, semantic explanation of Carian proper names (for the most part very weakly argued). As aresult, some theories were constructed based only on a purely hypo-thetical meaning attributed to Carian names.

As has been said, Dorsi’s corpus consists only of six definite glosses,and three doubtful ones. I choose to disregard the three god names,since no meaning is off ered in ancient sources, and they therefore



8

contribute nothing to our knowledge of the Carian lexicon.1 The sixdefinite glosses are êla ‘horse’, bãnda ‘victory’, soËa( n )2 ‘tomb’, g°la

‘king’, g¤ssa ‘stone’, and ko›on ‘sheep’. Of these, the first five are linkedby the fact that they are all found in the work of Stephan of Byzantium,who quotes them when speaking about the Carian place names ÉAlã-banda, ÑUlloÊala, Souãggela and MonÒgissa (see appendix B for therelevant passages).

As for the sixth gloss, although the Carian character of the wordconcerned is expressly mentioned in the sources, there are some prob-lems concerning the exact form of the word. The sources are two scho-lia to the Iliad XIV, 255, one in Eusthatius’ scholia, the other in themanuscript T of the Scholia Vetera in Iliadem,3 both of which are relatedto the name of the island of Kos (K«w ). In T, the word is given asko›on, while Eustathius’s passage is ambiguous: he states that K«n (acc.)is spelled KÒon by some, but then immediately adds that the Cariansuse this name for sheep (fas‹ d¢ toÁw Kãraw oÏtv kale›n tå prÒbata ).4

It has traditionally been interpreted5 that the Carian word was in factidentical to the name of Kos (hence k«w ‘sheep’), but Dorsi believedthat Eustathius left out the actual Carian word simply through over-

sight, which would be ko›on as it appears in the other scholion.However, Erbse, in a work published after Dorsi’s article (Erbse 1986:

389–390), is suspicious of the form ko›on of T, off ering it inter cruces.His preference is the kÒon form of Eustathius, interpreting the passageof Eustathius’ work mentioned above as a true gloss6 and noting more-over that this latter form appears in another passage by Eustathius

1

The three god names are ÖImbramow = Hermes (St. Byz. s. v. ÖImbrow; but notethe variant reading ÖImbrasow, which Dorsi does not mention, reported by the Scholia vetera in Theogoniam v. 338, and Eustathius, Commentarii ad Iliadem XIV, 281), Mãsariw= Dionysus (St. Byz. s. v. Mãstaura ), and ÉOsog«a = Zenoposeidon (Strabo XIV,659, Pausanias VIII; 10, 4).

2 As Dorsi rightly points out, “la n finale di soËan può essere un semplice morfemadi accusativo greco (peraltro non necessario: cfr. sopra êla ), ma può anche essere statasuggerita (a torto o a ragione) dalla scomposizione del toponimo in souan-gela” (Dorsi1979:29).

3 Edited by Hartmut Erbse: Scholia Graeca in Homeri Iliadem (Scholia uetera), Berlin,1979–1984. T is the Cod. Brit. Mus. Burney 86, to be dated in 1014 or 1059 A.D.

4 Eusthatius, Commentarii ad Homeri Iliadem 983, 33 (= ad Iliad. XIV, 255). I follow

the Leipzig edition of 1827 (re-edited in Hildesheim-New York 1970).5 Cf. Sayce (1887’92]:118): K«w (k«n ) ‘a sheep’, although he adds the spelling kÒowin Eusthatius. See also Brandenstein (1935:142).

6 At least this is what I have deduced from the very terse and implicit treatmentof the problem in Erbse (1986).



9

(318,41 = ad Iliad. II, 677).7 Although Erbse does not discount thepossibility that Eusthatius could have replaced ko›on with kÒon in order

to “improve” the etymological explanation in both passages, he seeskÒon as the genuine form, opting for the simpler solution that ko›on isa graphical error.

The three glosses that Dorsi lists as dubious are: tãba ‘rock’, tou-ssÊloi ‘dwarfs, pigmies’, and tumn¤a ‘stick’, all of which are also attestedby Stephan of Byzantium. The inclusion of tumn¤a, even if it is brandedas dubious, seems rather inadequate; it is true that Stephan mentionsthis word when referring to the Carian city of TumnhsÒw, but he attrib-utes it to the language of the inhabitants of Xanthos, a Lycian city,implying that the word must be Lycian rather than Carian.8

In the case of tãba ‘rock’, the word is cited by Stephan of Byzantiumin connection with a Lydian city called Tãbai. He adds that the wordtãba, which he does not attribute specifically to any language, is trans-lated in Greek as ‘rock’. Following this, he mentions another Tãbai,in this case situated in Caria, but it seems to be merely a passing ref-erence. From his observations then, the gloss had to be interpreted asLydian, but Dorsi rightly observes (1979:29) that modern scholarship

coincides in its estimation that no Lydian city of this name existed andthat the two cities are in fact only one, situated in Caria (cf. ZgustaKON § 1277–1, Blümel KarON:179). It is therefore feasible, althoughimpossible to demonstrate, that Stephan’s mistaken belief that the placename belonged to a Lydian country could have led to an error whenattempting to establish the origin of the gloss. The problem becomeseven more intractable given the existence of Tabhnoi (pl.), the namegiven to the inhabitants of some part of Lydia, which suggests that aLydian Tabai or similar might actually have existed (see ZgustaKON:593).

As for toussÊloi, the text is ambiguous and obscure. The wordappears under the entry Kãttouza, a Thracian city inhabited by pig-mies. Reference to Carians is therefore secondary and open to variousinterpretations.

7 However, in this case, the word kÒon is not attributed expressly to the Carians.8 Jãnyioi går tØn =ãbdon tumn¤an l°gousin.

10

2. Interpretation

If the collection of Carian glosses is very small and their value mediocre,the attempts to interpret them are equally disappointing. The onlyattractive etymological interpretation is that suggested by Carruba,regarding ko›on ‘sheep’ (Carruba 1965). Carruba proposed connecting it with Cuneiform Luwian ¢àôa/ì -, Hieroglyphic Luwian ha-wa/i/- <PIE *h2e/owo- “sheep”. The new form for the gloss defended by Erbse(1986), kÒon, would support this etymological explanation, because itwould come from an intermediate form *kÒWon, a plausible Carianresult of Proto-Anatolian (henceforth PA) *H àwo- (the stem was not in

-i- originally in Anatolian, cf. Lyc. xawa-, and see Melchert CLL s. v.¢àôa/ì - ). The treatment of PA laryngeal as velar stop in Carian (as inLycian) is now clearly confirmed (see below p. 260).

As for the possibility of finding some of these glosses in the Carianinscriptions, the results are equally discouraging, although in theorywords meaning ‘tomb’ or ‘king’ are likely to appear. In fact, thanks tothe bilingual inscription of Athens (G. 1), we know that ≤ jas is proba-bly one of the Carian words used for referring to a tomb or a funer-ary monument. The word appears as ≤ as in Euromos (C.Eu 1). Thepossibility of connecting these forms with the gloss soËa( n ) was con-ceived by Meier-Brügger (Meier-Brügger 1979:81),9 but we must beaware that in order to connect all of these forms, a lot of non-trivialsound changes are needed (*/swa/- > */swa/- > */sja/ > ≤ ja - (> ≤ a - ),for instance), and in any case, -s in ≤ ( j )as ?, vs. -n or -Ø in soËa( n )would remain unexplained.

In the case of the word for ‘king’, it is commonly supposed that itcould be very similar in Carian to the corresponding word in Lycian,

x ñtawat(i)- /k–dawati/. Adiego (1994a:240) proposed that the formºk ? d ow “ (part of a word esak ? d ow “ , E.AS 7) could be the Carian wordfor ‘king’, and this hypothesis was substantiated in Adiego (1995:18–21)by the Lycian-Greek-Aramaic inscription of Xanthos (Lycia, N 320),wherein the Carian divinity “King of Kaunos” appears in Aramaic asKNDWÍ (KNDWS) KBYD”Y. This seems to imply that KNDWÍ-KNDWS could be the Carian word for ‘king’ (see below Chapter 11s. v. esak ? d ow “ for more details).

9 It is somewhat remarkable that Meier-Brügger’s proposal, which implies an ≤ (asort of sibilant) value for z, was formulated some time before this value of z was con-vincingly established by J. D. Ray in the context of a wider system of decipherment.

11

If we accept these identifications, the problem of g°la is very simi-lar to that of soËa( n ): g°la and a Carian word k d ow (“ ) have some

points in common (the velar initial, a possible correspondence -l - :-d- (<*/nd/),10 but a series of fragile hypotheses about sound changeswould be needed to bring both forms closer.

An alternative approach to these glosses has been to question theirvalidity by taking into account that some of the place names glossedby Stephan of Byzantium allow an analysis based on Anatolian ono-mastics, which is clearly at odds with the etymological explanation of the Greek source. For example, in the case of ÉAlãbanda, Brandenstein(1936) observed that the word displayed the typical Anatolian suffix-anda , frequently used for place names, thus ruling out the segmenta-tion ÉAlã-banda on which Stephan’s etymological explanation was based,and consequently the validity of his information.

However, discrediting the information provided by Stephan is hardlyreconcilable with the idea, expressed by Dorsi (1979:34), that theseCarian glosses must all be considered as qualitatively important, sincethey may have borrowed from Carian writers of the Alexandrian era.This would imply that we are dealing with first-hand evidence, going

back to authors on Carian topics who could be still fluent in Carian,and who are therefore extremely reliable sources.

In my opinion, there is no implicit contradiction in accepting boththe information provided by Stephan of Byzantium and the diff erentinterpretation of the place name from the perspective of Anatolian ono-mastics: a folk-etymological analysis of ÉAlã-banda as a compound of two authentic Carian names does not say anything about the actualorigin of this Carian place name. In other words, even if we acceptthat Stephan’s etymological explanation of the place name ÉAlã-bandawas probably flawed, this does not necessarily invalidate the informa-tion that êla means ‘horse’ and bãnda ‘victory’. A good example of this is to take just to one of the place names involved: Souãggela waslater replaced by Yeãggela, undoubtedly by a process of folk-etymo-logical hellenisation (yeÒw + êggelow ). If someone had glossed this placename as “coming from yeÒw ‘god’, and êggelow ‘messenger’” and asmeaning ‘messenger of God’ or similar, the information about these

10 In the Carian language of Thebes, the letter & d is not used, and if one acceptsthat Carian of Thebes mlane corresponds to Carian of elsewhere md ane , a sound changed > l could be imaginable.



12

latter words would be correct, despite the fact that the place name wasnot created ex nihilo with this slightly absurd meaning, but rather was

the result of an intricate process of deformation.11 A similar explana-tion could be acceptable for all the Carian glosses based on placenames.12

B. T P N

1. The Unity and Continuity of Anatolian Onomastics

Before the decipherment of Carian writing, our only source of reliableinformation on the Carian language was (besides the scarce glossesanalysed in the foregoing section) the onomastic corpus, essentially per-son and place names transmitted by indirect sources, particularly Greekones. The most notable aspects of the Carian onomastics observed dur-ing the history of the research are shown here:

(1) Carian onomastics is to a large extent inseparable from the ono-

mastics of other minorasiatic regions that have also reached us throughindirect sources.

(2) These onomastics of Asia Minor in Greek transmission, whichcan be ascribed to a period that stretches from the middle of the firstmillennium B.C. to the first centuries A.D., turn out to be a clear con-tinuation of the Hittite-Luwian onomastics of the second millenniumB.C.

(3) A great part of these onomastics, traceable from the beginning of the second millennium B.C. to the start of Christian Era, can clearlybe linguistically interpreted as belonging to the Indo-European Anatoliangroup. We are dealing therefore with onomastics that are, linguisticallyspeaking, Anatolian.

(4) In any case, Carian proper names display certain characteristicsthat make them distinguishable from other Anatolian onomastics:

11 An intermediate phase Yuagg°l /a/ (ethnic Yuagg [ °leuw ] is also documented, see

Hornblower (1992:99, n. 160).12 For some etymological proposals (all rather provisional) formulated about Carianglosses, see Adiego (1993a:22). For ÉAlãbanda as ‘rich in horses’ (my suggestion, purelyhypothetical), see Adiego (1993a:21).



13

several name types, for instance those in -vllow, -vlliw, are typicallyCarian, and this singularity can be attributed to specific phonological

and/or morphological traits of the Carian language.The unity of the Anatolian onomastics transmitted by Greek sources

was established by Paul Kretschmer in chapter X (“Die kleinasiatischeSprachen”) of his Einleitung in die Geschichte der griechischen Sprache (Kretschmer1896). Kretschmer believed that in these onomastics a Pre-Indo-Europeansubstratum common to all Asia Minor (with the exception of Phrygia)was recognizable, given that many concrete elements (lexemes as wellas suffixes) appeared in diff erent regions of the Anatolian Peninsula.

The hypothesis of a single minorasiatic group sui generis (Kretschmer1896:292) constituted in those days an innovation, because the earliertheories of other scholars (reviewed summarily in Kretschmer 1896:289– 292) had tried to establish diff erent linguistic groups and attribute adiff erent external kinship to each one (with Indo-European, Semitic orCaucasian languages). It is true that Kretschmer was wrong in classi-fying this sui generis group as non-Indo-European, but in his defense,we must bear in mind that the existence of an Indo-European Anatolianfamily was in those days difficult to imagine. But in fact, this negation

of the Indo-European character of minorasiatic languages had a posi-tive eff ect: it obliged scholars to adopt the combinatory method foranalyzing Lycian (and later Lydian) inscriptions, and to discard morefragile etymological approaches.

Moreover, Kretschmer’s seminal work already outlined some of theideas—either new or systematized by him—that in the course of the 20thCentury have become vital to the research on Anatolian: the identificationof -ss- and - nd- as suffixes and their possible connection with Greekplace names in -ss- and - nd-; the frequent appearance of the so-calledLallnamen (names whose structure seems to be characteristic of children’slanguage: CV, CVCV, VCV, etc., like Dada, Nana, Ada . . .); and theisolation of lexical items that enter in compounding or derivation, aspig -, imbr-, tarku- or -muhw, nowadays easily interpretable as Indo-European Anatolian stems.

The approach begun by Kretschmer reached its peak in the monu-mental work of Johannes Sundwall devoted to Lycian indigenous names(Sundwall 1913). Sundwall tried to establish a systematic study of Lycian

onomastics, isolating and grouping the diff erent formative elements inthe proper names. Although it contains mistaken readings and namesthat are clearly Greek incorrectly analyzed as indigenous, Sundwall’s



14

book was a fundamental work in the history of the research on Anatolianonomastics. It was only to be superseded by the work of Ladislav Zgusta,

who in two books of enormous significance produced an exhaustivecollection of the Minor Asian person and place names of Classicalsources (Zgusta 1964, 1984).

When a connection could be established between these onomasticsand those of Hittite and Luwian of the second millennium, it heraldeda dramatic change of focus in the study of Minor Asian onomasticsfrom Classical sources. Obviously, this step forward became possibleonly when these languages in Cuneiform writing were sufficiently under-stood. Friedrich (1931) had already drawn attention to the lexeme muwa-in Hittite and Luwian proper names, a lexeme identical to the elementmuwa- present in the indigenous names of the first millennium. Theidea of a linguistic continuity between the second and first millenniathat could be detected thorough the proper names was given furtherweight by new evidence in Goetze (1951), but the definitive work onthis subject is the fundamental book of Houwink Ten Cate (1961), whodemonstrated with a great number of examples that the Lycian namescontained lexical elements of Anatolian origin. This theory came as no

surprise, however; parallel to the study of onomastics, Lycian had beenclearly established as a member of the Anatolian family, first in Pedersen(1945) and later by the work of Laroche, who demonstrated that Lycianwas very similar to Luwian (Laroche 1958, 1960, 1968). More impor-tantly, given the fact (already established by Kretschmer many yearsbefore, cf. above) that Anatolian onomastics in Classical sources showeda clear unity that suggested the existence of a linguistic group, the attri-bution of Lycian to the Indo-European Anatolian family together withHittite, Luwian, and Palaic, made it very likely that the languages spo-ken in Cilicia, Isauria, Caria and in the other Minor Asian regionswhere this type of onomastics is well documented also belonged to thissame Anatolian Indo-European group. In the case of Carian, this hypoth-esis constitutes one of the basic tenets of the proposals formulated by”evoro“kin since his first works, and seems to be confirmed by thedefinitive decipherment of Carian.

2. Present Compilations of Carian and Anatolian Names

Nowadays, we have at our disposal two complete and very up-to-daterepertories of Carian place and person names respectively, both com-piled by Wolfgang Blümel (Blümel 1992, 1998a). For the other Anatolian



15

proper names from Classical sources, the books of Ladislav Zgusta(Zgusta KPN and the supplement Zgusta 1970 for the personal names,

Zgusta KON for the place names) are still invaluable, although in thecase of Lycian, there are now two updated repertories of personal namescompiled respectively by Anne-Valerie Schweyer and Nicola Cau, whichappeared after Zgusta’s works (Schweyer 2002:95–128, Cau 2003). Itis to be hoped that these updates of Zgusta’s continue to appear, sincenot only do they incorporate new forms, but also off er revised formsand correct some mistakes, inevitable in works of such dimensions asZgusta KPN and KON.

For Anatolian onomastics of the second millennium, Laroche’s reper-tories remain indispensable: Laroche (LNH) for person names, andLaroche (TA1) and (TA2) for place names.

3. Carian Names from Indirect Sources vs. Those from Direct Sources

A methodological problem arises when dealing with the Carian namesfound in indirect sources. In Adiego (1993a), these were analysed(although not exhaustively, looking only at the most significant aspects)

dispensing totally with those Carian names directly attested in Carianinscriptions. The reasoning seemed be well founded at the time; theprincipal aim of the book was to defend a new decipherment of Carianwherein the use of Carian onomastics from indirect sources played afundamental role, so it seemed preferable to avoid mixing forms directlyobtained from the decipherment with indirect forms from Greek sources,which were the key to that decipherment (In Adiego 1993:26, n. 4,”evoro“kin’s mixing of forms in his book of 1965 was severely criticized).

Nowadays the situation is clearly diff erent; this book is not conceivedas a justification of a concrete proposal of decipherment, but is meantto off er as complete a picture as possible of our current knowledge of Carian, now that the new decipherment has been universally accepted.I therefore believe that all the data available from the indirect attes-tation of Carian names must be taken into account together with theCarian onomastics from direct sources. The study of Carian onomas-tics, both from indirect and direct sources, will consequently be dealtwith in a specific section (Chapter 8). In Appendix C, a list of the

Carian personal and place names, mostly based on the compilationspublished by Wolfgang Blümel (Blümel KarPN and KarON) is pro-vided. The list is merely intended to off er a convenient compendiumof Carian indirect onomastics: for the data corresponding to each name,



16

the reader should turn to the articles written by Blümel. For extremelyuseful, updated accounts of the data obtainable using only the Carian

names from indirect sources, without discussion of the direct documen-tation, I refer the reader to Neumann (1988) and, above all, Neumann(1994).



CHAPTER THREE

THE INSCRIPTIONS

A. I

The most direct and important sources of Carian language are obvi-ously the inscriptions in Carian alphabet, although strangely the bulkof this epigraphic corpus does not come from Caria itself, but fromvarious other locations in Egypt. The historical reasons for this curi-ous circumstance have been covered in Chapter 1.

Inscriptions on funerary stelae and other objects, mainly from Memphisand Sais, and graffiti found in other parts of Egypt are the result of this long presence of Carian-speakers in Egypt. About 170 inscriptionshave been found in Egypt to date. All these texts are relatively short,given their typology (onomastic formulae in funerary texts—Carians

were somewhat laconic when writing epitaphs—and brief graffiti).1

The epigraphic material found in Caria itself is far less abundant(approximately 30 inscriptions), but it includes several texts that aremore extensive than those discovered in Egypt, particularly the fol-lowing three: a decree from Kaunos whose precise terms are stillunknown (C.Ka 2), the proxeny decree for two Athenian citizens writ-ten in Carian and Greek, also from Kaunos (C.Ka 5), and a decreeenacted by the Carian satraps Idrieus and Ada, possibly concerning asyngeneia of the temple of the god Sinuri, near Mylasa (C.Si 2). To these

three inscriptions now must be added the new inscriptions of Mylasa(C.My 1) and Hyllarima (C.Hy 1), the latter in fact a fragment thatcompletes the inscription already known.

Besides Egypt and Caria, we know of several other inscriptions foundin the bordering regions of Lydia and Lycia, as well as in Greece. Forconvenience, I will classify the texts of Tralleis and of Krya (on theGulf of Telmessos) as Carian, since we are dealing in both caseswith areas very close to Caria. It is logical to assume that there was

1 On Carians in Egypt, see Masson (1969), Masson (1977[78]) and now Vittmann(2003:155–179).



18

a linguistic continuity beyond the alleged political boundaries of Caria,and Carian was spoken in some border zones of Lydia and Lycia.2 For

the same reason I classify as Carian the graffito from Didyma, nearMilet, situated on the Carian border.

1. The Revised System of Transcription of Carian Letters

Before dealing with the Carian corpus of inscriptions, it seems appro-priate to present here the system of transcription of Carian letters thatwill be used throughout this book, since certain new conventions areintroduced here. The phonological reasons behind the new transcrip-tion procedures for several signs will be dealt with in Chapter 7, andhere I will limit myself to giving a brief justification of them.

Although the bilingual of Kaunos has confirmed the overall ‘Ray-Schürr-Adiego’ system (see Chapter 4), and by extension the deci-pherment tables in Adiego (1993a), (1994a) remain essentially valid,3 Ibelieve that this is a good opportunity to introduce slight modificationsto the system. These will allow us to adjust the transcriptions of sev-eral letters more precisely to their actual phonological value, and to

simplify other conventions in transcribing Carian signs. Moreover, sincethe publication of the works mentioned above, hypothetical sound val-ues have been proposed for some formerly undeciphered letters, whichwill also be considered in the revised system presented here.

2. Vocalism

The method of transcribing vowels used until now has been largelysuperseded by our improved understanding of the Carian vocalic sys-tem. In the case of i / j and u / v, a purely diacritical distinction

2 Tralleis was situated north of the River Maiander, which served as the traditionalboundary between Lydia and Caria, but this boundary was undoubtedly permeable tocontact between people, see Hornblower (1982:2). According to Strabo (XIV, 1, 42),Tralleis was inhabited by Lydians, Carians and Ionians, and the Carian flavour of thealternation -ll -/-l d- in Trall e›w vs. Tral de›w was already noted by Benveniste (apud Robert 1945:20, n. 2). As for the Krya inscription, it clearly belongs to the Kaunian

alphabetic variety, which is congruent with the geographical proximity of the twoplaces.3 As a exception note only the Kaunian letters T t (vs. “ in the rest of the alpha-

bets) and / “ (a specific Kaunian sign), whose value has been established from thebilingual (about these letters see here pp. 228–229).

19

(“i”/“í”, “u”/“ú”) was employed in order to reflect the alternationsdetected between each pair of letters. At present, it seems clear that,

at least originally, the alternation lay in a vowel vs. semivowel oppo-sition, so that a transcription i / j, u / w would be more adequate. Inthe Carian alphabets of Caria itself, j and v were abandoned, andi and u took up their functions (cf. Kaunian u≤ ol ≤ vs. Carian of Egyptw ≤ ol ≤ ).

As for W and w, their character of ‘u-sounds’ led to a conventionaland perhaps flawed transcription “ù” and “w”, respectively. The ideathat these letters represent a /y/ sound, already proposed in Adiego,has been confirmed, at least in the case of W, thanks to the Kaunosbilingual (see p. 237). Therefore a transcription of W by means of ymust be preferred. Consequently w, a letter that alternates with W, willbe transcribed as ÿ. It is possible that w represented the semivowel cor-responding to /y/, id est, /w/, but it is not absolutely certain, andmoreover the transcription w could lead to confusion due to its formalresemblance to y (cf. the similar problem caused in Lydian transcrip-tion between n and v ).

The letter V/W is only documented in Sinuri, Kildara and Mylasa,

where no traces of W are found. Assuming that a similar process of eliminating semivowel letters occurred, the simplest solution is to inter-pret V/W as the Sinuri-Kildara-Mylasa form for W = y, despite of the(apparently) more formal proximity of V/W to w than to W. For thisreason, V/W will be transcribed by <y>, in the same way that T, theKaunian form of t = t , is transcribed simply by <t>.

3. Consonantism

The case of Kaunian T = t and / = “ has already been mentioned;given that there are no doubts about these values, and that these let-ters are simply the forms that t and f, respectively, adopted in theKaunian alphabet, it is not necessary to resort to more complex tran-scriptions, such as t 2 —used by Marek-Frei in their works—or “ 2.

The transcription of x X merits further consideration. In Ray, andin my early works, it was transcribed by h. A new transcription wasintroduced—suggested by Neumann—in my subsequent works: x, which

has been commonly accepted. While this latter transcription is moreconsistent with the tectal and stop value of x X, discernable from theCarian-Egyptian equivalence urs xli - = 3rskr , a more in-depth analysisof Carian has allowed us to establish a more precise value for this

20

letter, which cannot be a velar aspirated stop or a velar fricative, thekind of sounds that the transcription x would imply. Rather there is

evidence that points to a palatal stop (see Chapter 7), hence the pro-posal that it should be transcribed as ∞, the grapheme used in Indo-European Linguistics for PIE voiceless palatal stop.4

The remaining consonants whose values are well known remainunchanged. Further modifications involve several letters of less certainvalue. In the case of 1 = z (instead of z ), the reason is to avoid employ-ing too many Greek letters in transcribing Carian. Insofar as z was notused at all, I think it reasonable to substitute it for z, although I rec-ognize that z is a somewhat ambiguous letter and its IPA value is cer-tainly not the same as 1 presumably had in Carian. However, I believethat this simplification is desirable, in the same way that there is a ten-dency to use x instead of x in transcribing Lycian.

As for 0 and %, it is likely that they represent several types of tec-tal, but the evidence is inconclusive, and an exact value is impossibleto determine. The possible identification ≤ u0l i - = Souagg ela points toa *ng origin for 0, in parallel to & d < *nd and Ø b < *mb, so thata corresponding g transcription is adopted, although it is true that the

option of <g> was still available. Regarding %, the only clue we haveis Schürr’s proposal to identify pr %idas with Bragx¤dai. Although henow seems to have abandoned this hypothesis, I still see it as an attrac-tive idea that should not be ruled out. The purely conventional—andfar from certain—transcription of <ã> is provisionally adopted here,although there is no strong supporting evidence.

A rather diff erent problem is posed by O, an exclusively Kaunianletter. In Adiego (2002) I off ered arguments in favour of identifying itwith the far more widespread c letter, absent in Kaunos (see here p.252). In any case, I recognize that my arguments are not particularlystrong, so I adopt, also cautiously, a transcription t2.

To avoid confusion, I shall not attach a question mark to these ratheruncertain values in transcribing Carian texts, but instead I will indi-cate such cases in the sign tables.

The following table shows the new system of transcription of Carianused in this book:

4 A likely alternative was <c>, which would coincide with the letter used in IPA for this type of sound. However, this is a very ambiguous letter in Indo-European stud-ies, so a more precise letter like ∞ seems preferable to me.



21

Nº (Masson) Letters Transcription

1(+8) a A ~ À (E ) a3 d D G d4 l l5(+41) W ù V W y6 r R r7 s 2 L l 9 q Q q10 b 5 B b11 m M m12 o o

14 t T t15 f F S / _ “17 s s18 H ?19 u U u20 ñ ñ21 x X ∞22 n N n24(+2) p p ( ) p25 z Z ≤

26(+8) I í Î Ï y ì Y I i27 e e e28 w ÿ29+30 k K k31 & d32 v Ú w33(+34) 0 8? g 35(+36) 1 9 9 z37 % ã

38 j T j

39 _ 1 ?40(+23?) c C / O? t / t2?42 6 ®43–44–45 Ø ® 4 B &? b? b46 ÿ b2?

Note finally two conventions also used throughout this book. Dagger (†)indicates a transcription of a word or letter by means of a diff erentsystem of decipherment. Double-dagger (‡) is used to mark an old read-ing of a word or letter now discarded. A combination of both signs(†‡) serves to signal both the use of a superseded deciphering systemand an erroneous reading.



22

B. ‘P -C ’ ‘C ’ I

A serious problem is caused by a heterogeneous body of inscriptions,whose only common characteristic is the use of writing systems thathave, or seem to have, certain similarities with the Carian alphabet,but which cannot be definitively classified as Carian inscriptions. A detailed analysis of this material cannot be provided here, so I merelyintend to establish the problem in precise terms and to enumeratethecorpus of these inscriptions regularly referred to as “para-Carian”or “caroide”.

For years, the temptation has existed to attribute any inscription fromAsia Minor written in an unknown or barely recognizable alphabet toCarian. In a sort of obscurum per obscurius , such materials were classedas Carian at a time when the Carian alphabet itself remained un-deci-phered. Today, we have a better understanding of the Carian alpha-bet (letter values, geographical variants, a complete inventory of signs)and we can reject the theory that these materials are Carian (canoni-cal Carian, at least).

1. “Para-Carian” Inscriptions from Caria

The “para-Carian” or “caroide” inscriptions discovered in the Carianarea can be separated geographically in five groups: (1) Khalketor, (2)Ancin (south of Alabanda), (3) Labraunda, (4) Stratonikeia (a singleinscription), (5) Aphrodisias (also a single text). Groups 2–4 are in fac-trock graffiti that are practically impossible to read. From the existing editions of the inscriptions, nothing can be acceptably identified asCarian.

The cases of Khalketor and Aphrodisias are diff erent; in Khalketor,at least one of the two inscriptions discovered is clearly readable. Butthis simply allows us to state that the alphabet of Khalketor has noth-ing to do with the Carian writing system. Concerning the language of texts from Khalketor, all attempts to interpret them using our limitedknowledge of Carian have proved fruitless. In recent years the theoryhas been suggested that we are in fact dealing with much more recentinscriptions, perhaps written in Turkish.

Finally, the Aphrodisias inscription consists of eight letters (the lastone incomplete) and does not pose problems of reading. Some lettersresemble Carian ones, but others are totally alien to the Carian alpha-bet. As a matter of fact, we cannot discard the possibility that thisinscription actually represents the Carian alphabetic variety of Aphrodisias,



23

but in this case, we ought to accept that this variety was extremelyuncharacteristic. Since further Carian epigraphic testimonies from

Aphrodisias are lacking, this text (which is also uninterpretable) will beexcluded from our corpus.

These texts had been included in the successive inventories of Carian inscrip-tions: Ancin = D 5, Labraunda = D 17 (recent edition: Meier-Brügger 1983),Estratonicea = 26* (Hanfmann-Masson 1967); Chalketor = D 4 and 27*(Neumann 1969a; revised in Blümel 1988); for the Aphrodisias text, seeInnocente (1994:107–108 [text nº 7]). Blümel (1998:168) mentions the exis-tence of several other inscriptions in Chalketor’s alphabet. The ‘Turkishapproach’ has been suggested by Hasan Malay to Blümel, see Blümel (1998:169).

I have some doubts about the true Carian character of the two identicalinscriptions from Keramos (here C.Ke 1, C.Ke 2), but I retain them, thoughin a provisory way, in the corpus insofar as all the letters that appear therecan be analysed as Carian.

The so-called ‘tegola de Iasos’, an inscription consisting of eight signs pub-lished by Lucia Innocente (Innocente 2002) could be Carian, but, as Innocenteherself observes, the very few letters that can be identified are not unequivo-cally Carian, and no valuable results can be obtained. Therefore, I excludeit from the present corpus.

For similar reasons I exclude from the corpus the alleged Carian inscrip-tion from Labraunda, recently published by Belli and Gusmani (Belli-Gusmani2001). According to Gusmani, the inscription reads (a) e E 2 e s m (b) M

U.. While part (a) has a Carian flavour, the co-occurrence of e and 2 wouldbe very strange. Also, a form such as E is rather puzzling. As for (b), asGusmani rightly observes (Belli-Gusmani 2001:41), a form M instead of thetypical Carian form s for s is surprising.

As for the alleged Carian fragment from Kaunos 51*, consisting of onlytwo letters c a (see Frei-Marek 2000:125–126), I consider it as non-Carian: theletter c is totally absent from the Kaunian alphabetic inventory, and the formof a is very diff erent from the all the variants attested for this letter in theCarian alphabets.

2. ‘Para-Carian’ Inscriptions from Other Places

The rest of the ‘para-Carian’ or ‘Caroide’ inscriptions come from otherlocations: from the islands of Cos and Rhodes, possibly part of theCarian linguistic area; from the neighbouring countries Lydia and Lycia;from Egypt, and also from Persia. Even some texts of unknown originhave been arbitrarily interpreted as “Carian”. None of these texts canbe included in a strict corpus of Carian inscriptions, but given theirpresence in former collections (particularly in Meier-Brügger 1983), Iwill include them here, accompanied by a brief note:



24

• Inscription from the area of Telmessos (Lycia; 36* = Meier-Brügger1981). It seems to be related to the Chalketor corpus, but its authen-

ticity is not absolutely conclusive.• Inscriptions of Belevi (Ephesos, Lydia; 24* = Dressler 1966–67).

Three inscriptions on stone fragments. Some letters seem distinctlyCarian, whilst others are clearly alien to the Carian alphabet.

• Inscription of Persia (32* = Pugliese Carratelli 1974[77]). An inscrip-tion on a bronze bowl, whose only alleged Carian trait is a sign, &,which recalls an identical Mylasean letter now known thanks to thenew inscription C.My 1, and also, although more remotely, the typ-ical Carian letter &. However, the presence of letters alien to theCarian signs inventory (P, K, Z) invalidates their classfication asCarian.

• Tablets Peiser-Böhl-von Grotthus (23*). A group of three tablets, thefirst two edited by Böhl (1932/33), the third by Friedrich (1965), andcarefully analyzed by Meriggi (1966). The exact origin of the tabletsis unknown, as is their dating, but they have traditionally been attrib-uted to Cappadocia. This means that the labelling of their writing system as ‘para-Carian’ is based merely on the alleged graphic sim-

ilarities between this system and the Carian alphabet.

Such similarities are indeed clear. It is particularly striking that severaltypical—and in some cases exclusive—Carian letters such as p, z, k,v, j seem to appear. But there are also Greek letters alien to theCarian alphabet (for instance K ), and a number of signs that cannotbe compared either to the Carian or to the Greek alphabet.

The main difficulty in interpreting these three tablets lies in manycases in the impossibility of distinguishing between which signs are actu-ally diff erent, and which are mere variants. For example, the thirdtablet (Friedrich 1965), contains several types of r pointing in diff erentdirections and with slight modifications of shape, and it is impossibleto know whether we are dealing with the same sign in all instances.As an indication of these difficulties, it is worth considering that whereasFriedrich identifies 217 diff erent signs in this tablet, Meriggi reducesthis number to 61, and Nahm (1974) further still to 34 (although inthis latter case, Nahm applies somewhat unconvincing criteria). As a

result, it is hard to ascertain whether the graphic system is alphabetic,syllabic, of a mixture of both. Furthermore, we are unaware of the lan-guage in which the tablets are written. All these problems make thesetexts undecipherable, and the Carian specialist must limit himself to



25

noting the rather surprising graphic affinities between the Peiser-Böhl-von Grotthus tablets and the Carian alphabet. Only a significant increase

of the documentation available would allow us to move forward intheir study.

The only eff ort at decipherment of which I am aware is that of Nahm, regard-ing the third tablet (Nahm 1974). After reducing the inventory of diff erentsigns as far as possible (cf. above), he transcribes with Greek letters those signsthat show a certain similarity to them. His observations on several recurrentsequences are of some interest, but they are conditioned by the somewhatquestionable attempt to decrease the total number of signs.

• Graffito of Cos (29* = Metzger 1973). A graffito inscribed on anamphora coming from the island of Cos. In this case, doubts arisefrom the almost complete illegibility of the inscription. ”evoro“kin(apud Metzger 1973:77) and Metzger claim to recognize some Carianletters such as j, v, or i, but a complete reading of the text isunfeasible.

• Graffito of Ialysos (Rhodes) (Innocente 1994:101–104 [text nº 1]).Innocente has defended the inclusion of this four-letter graffito, for-

merly considered to be Lycian, in the Carian corpus. It is true thatall the letters of the graffiti could be Carian, but none of the possi-ble readings of this brief inscription off ers compelling results (← u“ t u,ua t u, even ut ku /→ ut “ u, ut ku )

• “Poetto’s two para-Carian inscriptions”. Innocente (1994:108–109[texts nº 8 and 9]) includes in her collection of supposed Carian‘monstrosities’ two texts whose existence has been noted by MassimoPoetto: a cylindrical seal of unknown origin and extremely old pos-

sible dating (1000 B.C.!), and an object classified among the “Grie-chische geometrische Gemmen”, also of unidentified provenance.Each inscription contains only a few signs, and their attribution toCarian is based only on the possible resemblance of some letters tothose of the Carian alphabet. Consequently, I believe that it is prefer-able to exclude them from the Carian corpus.

• The “Oxford Para-Carian Inscription”, a tablet from Sardis first pub-lished by Sayce, and recently re-published with excellent photogra-phy by Innocente (Innocente 1995). The situation is similar to the

preceding para-Carian texts: whilst some letters essentially resemblecertain Carian signs, others are completely alien to known Carianrepertoires (for example the three-straight sigma), and no linguisticinterpretation is possible.



26

• The most extreme case of resorting to Carian script as a panacea,is the attempt of to classify as Carian a number of masons’ marks

found in very diff erent locations. Gosline (1992) proposed a Carianorigin for the masons’ marks discovered in Elephantine (Egypt); inanother paper, the same author proposes the same origin for themasons’ marks of Pasagardae (Persia) (Gosline 1998). This line of argument has also been taken up by Franklin (2001) in order toexplain masons’ marks found in North Israel (Samaria and Megiddo),dated to as far back as the IX century B.C. More recently, Avishurand Heltzer devoted a brief article to connecting these latter exam-ples from Israel with the biblical references to the k àr ì (Avishur-Heltzer 2003).

Although it remains possible that Carian masons were responsible forsome of these marks (in the case of North Israel, however, I wouldthink it unlikely),5 a careful observation of the collections of marks doesnot throw up compelling evidence of a connection with Carian: wherethe letters that can be easily identified are not specifically Carian (forexample, an inconsequential letter such as a ), the links proposed with

distinctively Carian letters (w, j ) are extremely tenuous.

3. The Ostrakon of Hou and the Naukratis Fragment

None of the para-Carian or Caroide texts mentioned above comes fromEgypt. In fact, the only examples of para-Carian inscriptions found inEgypt until now are the so-called Carian Ostrakon of Hou (DiospolisParva), and the inscribed fragment of a piece of pottery known as theNaukratis fragment.

The Ostrakon was discovered by Flinders Petrie in 1899. Sayce (apud Petrie 1901) considered it to be “Kaunian”, and later re-edited it him-self as a Carian inscription (Sayce 1905). Masson and Yoyotte excludedthe text from their edition of “pharaonic objects”, but some years later,Masson published it as a truly Carian text (Masson 1967). This view,however, seems to be too optimistic: once again, as in the para-Carian

5 Possible mentions of Carians in Biblical sources are, according to Hornblower, “aninviting but nebulous topic” (Hornblower 1982:16, n. 82). Avishur-Heltzer (2003) doesnot add any new evidence for these alleged references, and employs dangereously cir-cular reasoning: the k àr ì of the Bible are Carians precisely because the masons’ marksshow Carian letters (!).



27

texts mentioned above, a few resemblances to Carian letters are foundalongside clearly non-Carian letters, and none of the sequences of let-

ters off ers the slightest possibility of a connection with known Carianwords and sequences. It is rather surprising that Masson, generally scep-tical about the acceptance of other documents as Carian (compare forinstance his exclusion of the scarabaeum inscription, E.xx 5, much morelikely to be Carian, Masson 1959b) defended the addition of this textto the Carian corpus (Masson 1967, a view repeated in Masson 1969:32).I include here Masson’s drawing of the ostrakon:

Ostrakon of Hou (Masson 1967)

The text inscribed on a piece of pottery (the ‘Naukratis fragment’), alsofound by Flinders Petrie, consists of only three letters, whose Carian

character is unverifiable, and must therefore be excluded from a seri-ously conceived repertory of Carian texts. See Masson-Yoyotte (1956:14–15) for further remarks.

4. Carian Gra fi tti from Sardis

The Carian graffiti found in Sardis deserve closer attention. The Cariancharacter of these texts, edited by Gusmani (1975) is in my opinionbeyond doubt, since they contain some of the most typical Carian let-

ters, such as p, z or %. However, the scarcity of the inscriptions,their highly fragmentary nature, and the serious difficulties encounteredwhen analyzing and interpreting the texts (the direction of reading andthe exact identification of the letters remain unclear in most cases) mean



28

that we must exclude them, at least provisionally, from the currentCarian corpus.

A good indication of the uniqueness of and the difficulties presentedby the graffiti from Sardis can be seen in the longest example: the join-ing made by Gusmani (1990) of two fragments (C I 1 and C I 5 inGusmani 1975):

Sardis C I 1 + C I 5 (Gusmani 1990)

At first sight, many typical Carian signs can be easily recognized: v

w w ÿ, p p, j j , z ≤ , i i , etc., so that the attribution to Carian isincontestable. The first difficulties arise when one tries to establish theorientation of the reading. For example, the second line off ers contra-dictory evidence: there are two m, oppositely oriented (m M ). Even areversal of the text cannot be rejected, due to the form of a. Nearlyevery possibility of reading direction implies a diff erent interpretationof the letters (for instance B = b, but u = u ). A further complicationis that some letters are not easily identifiable: Z in the third line could

simply be a variant ofL

l , but it could also be interpreted as a formof z; the letter 8 in the left part of the second line resembles a let-ter only documented in the alphabetic variant of Kaunos, a trait thatseems suspicious. Similar problems of identification are found in othergraffiti: a letter E appears at least three times. Are we dealing, then,with an ancient form of e e (note the alternative form e in Hyllarimaand Mylasa) or rather with a quadrangular variant of i?

But the main problem with this and the other Sardis graffiti is thatnone of the diff erent possible readings yields a sequence that can be

compared to the rest of the Carian documentation. The possible con-nections are far from conclusive, and only can be based on a handfulof signs. For example, in C I 2, ]-?-mzto [, transcribed as ]-?-m≤ to [,a possible ending of a name in genitive -m≤ could be identified,



29

but the brevity of the fragment makes this interpretation simply an ad hoc one.

In conclusion, the presence of the Carian alphabet in Sardis is unde-niable, and this may have interesting implications for the chronologyand spread of Carian writing, but from a linguistic point of view, thescarcity of these materials renders them useless.

5. Carian Inscription from Old Smyrna

A similar problem to that of the graffiti from Sardis is posed by aninscription from Old Smyrna, the photography and drawing of whichwas published by Jeff ery (1964, nº 23). It contains clearly Carian let-ters (note &, e and perhaps also p, f ), but it is impossible to obtaina comprehensive and satisfactory reading of the inscription given theambiguity of many of the signs, which can be interpreted in verydiff erent ways, and the impossibility of establishing the exact directionof the writing.6

One of the most significant features of this inscription is its possibleage: the object is dated at the end of the 7th century (so it would be

more or less contemporary with Carian graffiti from Abu Simbel). Thereare no reasons to suggest that the inscription should be dated muchlater than the object: the alleged palaeographical reasons put forwardby Jeff ery are untenable from a Carian perspective (Adiego, 1993a:86).

6 Schürr (2001c) off ers a very hypothetical interpretation of the text: em-?-l / sa l pde/. . . u b r od bore “ . No clear connection with the Carian corpus can be established for anyof these sequences.

Old Smyrna ( Jeff ery 1964)



30

7 Masson (1978:6–7), Martin-Nicholls apud Masson (1978).

C. T C I E

The epigraphic materials from Egypt (about 170 inscriptions, to which50 new and still unedited inscriptions found by ”evoro“kin must beadded, cf. below) are the most important part of the direct documen-tation of Carian.

From a typological point of view, a diff erentiation can be madebetween funerary and votive inscriptions on the one hand, and graffition the other. The first group has been the best source for decipher-ing Carian, mainly due to the relative lack of reading problems, theregularity of the graphic system used, and the excellent editions of theinscriptions that subsequently appeared.

The scenario presented by the graffiti is quite diff erent. In contrastto the high degree of standardization of the alphabet used in the votiveand funerary inscriptions, the graffiti show a markedly less careful useof writing, as a result of the spontaneity involved in their execution.This problem is further complicated by the lack of modern and reli-able editions of the main body of graffiti. Of the graffiti already knownsince the times of Sayce, only the graffiti of Abu Simbel and Buhen

have been revised and edited with a sufficient degree of reliability tobe used without significant reservations. The rest of the documentationis still pending new collations, which in some cases will become unfea-sible due to the likely destruction of the graffiti.

The typological classification outlined above can easily be linked togeographical distribution: the funerary and votive inscriptions comefrom Memphis and Sais, whereas the graffiti have been found in var-ious locations further to the south. The historical reason for this isclear; the graffiti are present because of the temporary presence of Carian visitors, particularly during the military campaigns, while thefunerary and votive monuments exist due to the settlement of Carianpeople in cities like Memphis or Sais.

When considering the dating of the Egyptian corpus, it is essentialto mention the important work of Kammerzell on the Memphis Corpus.It had been assumed that this corpus should be dated after the settle-ment of Carians in Memphis by order of Amasis, i.e. around the mid-dle of the VI century (Amasis is believed to have ruled Egypt between

570 and 526 B.C.).7

But Kammerzell, after carefully analyzing the typol-



31

ogy of the funerary stelae, has established that some of these inscrip-tions can be assigned earlier dates. It is true that not all of the method-

ological procedures used by Kammerzell in order to determine thechronology of Caromemphite stelae are similarly convincing, but atleast in the case of E.Me 7, wherein the same person is mentioned inboth the Carian and the Egyptian parts, a dating before 570 seemsvery probable. A more speculative suggestion is his identification of thename pikre- of the stela E.Me 3, as corresponding to the Pigres (P¤grhw )mentioned in Polyaenus’ Stratagemata (7, 3). This figure was referred toas an adviser of Psammetichus I in the early years of his reign, so thisstela should be dated approximately between 660 and 620 B.C. But itis very dubious to base, as Kammerzell does, the chronological attri-bution of E.Me 3 and typologically similar stelae only on this indemon-strable personal identification.8

Until now, the oldest Carian inscription from Egypt is the base of a statue of the goddess Isis, which can be dated to the second half of VII century thanks to the presence of a cartouche displaying the nameof the pharaoh Psammetichus I. This document is therefore chrono-logically very close to the arrival of Carian and Ionian mercenaries in

Egypt and their subsequent settlement in the Eastern Delta (Masson1969:35–36, 1977[78]:335). As for the rest of the Carian documenta-tion from Egypt, a precise dating can be given only to the graffiti fromAbu-Simbel (E.AS); there is no doubt that these graffiti were inscribedin the course of the great Nubian campaign ordered by PsammetichusII and conducted by Potasimto, as is particularly evident in a long Greek graffito, where mention is made of this historical context. Thiscampaign has commonly been dated in 591 B.C., but Ray (1982:85)suggests revising this chronology slightly, to situate the event in 593/92B.C. The Buhen (E.Bu), Gebel Sheik Suleiman (E.SS) and Murwàw(E.Mu) graffiti are also likely to date from the same period.

The inscription on a bronze lion “de provenance égyptienne” (E.xx.7)is dated by Masson (1976) at around 500 B.C., given the Achaemenidartistic influence visible in the figure of the lion.

For the remaining inscriptions (mainly graffiti plus some texts foundinscribed on various objects), there is no certain dating. Only in thecase of the graffiti from Abydos do we have some idea: Masson has

8 See the same criticism in Masson (1995:176).



32

suggested that they may be contemporary with the oldest Greek graffitifrom the same location, so that they can be dated around the end of

the V century.

1. Sais (E.Sa)

The sub-corpus of Sais is currently constituted by two bronze votiveobjects, included in Masson-Yoyotte (1956). Both texts are bilingual andhave proved fundamental to the deciphering of Carian.

E.Sa 1 (= MY L)

E.Sa 1 (Masson-Yoyotte 1956)

“arkbiom: zidks mdane: ÿn-[?]mo | den: tumn(Egyptian text: Jtm ntr ‘3 dj ‘n¢ snb ” 3rkbym )

Inscription on a reliquary for mummified reptiles. The only problem-atic letter is the last letter of the first line. ”evoro“kin’s attempt to com-plete ÿn[s]mo (see ”evoro“kin 1965:119–120) is only based on thecomparison with ÿnsmsos in E.AS 3 (cf. also E.Mu 1). Masson-Yoyotte(1956) mention two extremely divergent proposals made by Wild andRaphäel after examining the original (K k vs. 1 z , respectively!), andleave the letter without interpretation. The photography in Masson-Yoyotte is in this case unusable. The Carian inscription is accompa-

nied by an Egyptian formula, “Atum the great god may give life andhealth to ” 3rkbym”. The non-Egyptian, and presumably Carian, char-acter of the name ” 3rkbym was already noted by Sayce (1905:124), seeMasson-Yoyotte (1956:52). Also correct are the observations in Masson-



33

Yoyotte (1956:52) regarding the vocalisation o of º ym, based on itsspelling using the Egyptian word for ‘sea’ ( ym = * yò m cf. Coptic eiom ),

and about the possibility that the name would be *Sarkebivmow in aGreek transcription (the form is not yet documented, but a form Kebivmowhas appeared in the meantime). Surprisingly, though, none of this evi-dence encouraged Masson and Yoyotte to attempt to find this namein the Carian text, and the name was not identified there until Kowalski(1975).

Photograph and drawing: Masson-Yoyotte (1956:53, pl. VI).

E.Sa 2 (= MY M)

9 Other names belonging to the genealogy of the dead are also mentioned. Aboutthese forms and the possible genealogical tree of P3-dj-Njt , see Masson-Yoyotte (1956:61).

E.Sa 2

pdnejt q ÿri≤ ∞iInscription found on the base of a statuette of the Goddess Neith. Forthis text in scriptio continua , I adopt the division suggested by Meier-Brügger (1979a:81–82), contra Masson (1978) and Meier-Brügger (1979b).In the Egyptian texts that accompany the Carian inscriptions, a P3-dj- Njt son of K3rr is mentioned.9 These two names clearly correspond tothe pdnejt son of q ÿri- in the Carian part. The use of the biliteral signk3 for a syllable /ku/ or sim. (= Carian q ÿº) has been correctly identified

by Vittmann, see Vittmann (1996).The presence of a cartouche with the name of Psammetichus I allows

us to date the object to the times of this Saite pharaoh (663–609 B.C.),which would mean that this is the oldest datable inscription of theCarian corpus from Egypt.

Photograph and drawing: Masson-Yoyotte (1956:63, pl. VIII).



34

2. Memphis (E.Me)

The inscriptions published in Masson (1978) and those of Memphiteorigin included in Masson-Yoyotte (1956) are grouped together underthis label. This corpus has been crucial for the decipherment of Carian.The excellent quality and preservation of a great number of stelae, thefact that they appear on monuments, which implies a very standard-ized use of writing, the geographical and chronological consistency of the corpus, and the fact that it includes some bilingual texts (E.Me 5,E.Me 7, E.Me 8, E.Me 9, E.Me 15) make this sub-corpus the mostimportant direct documentation of Carian.

To this sub-corpus, we must add the so-called stela of Abusir (Masson1978:91, Kammerzell 1993:138–139), although we need also to con-sider that its reading is very difficult. Very recently, Diether Schürr hastried to improve the reading of this inscription (Schürr 2003), and hiseff orts will be taken into account here. Finally, I also include the frag-ment 180* from Kammerzell (1993) (here E.Me 66), although it is alargely unusable document.

Generally, I adopt the readings given in Masson-Yoyotte (1956) andMasson (1978). The diff erences, concerning certain details of reading and, above all, the order in which some texts must be read, will beduly indicated.

E.Me 1 (= MY A)

E.Me 1 (Masson-Yoyotte 1956)

←

ttbazi[ ≤ ] | p. iub[a]Ωi≤ | aor[ ≤ ]



35

Inscription on a false-door stela. Reading according to Schürr (1996),the main diff erence being the interpretation of the first letter of the

second name: Masson’s reading was l l , which must be discarded.

Photograph: Masson (1953: pl. XII). Drawings: Masson-Yoyotte (1956:2–3).

E.Me 2 (= MY B)


←

uksmu | lkor≤ | mrsi≤

False-door stela. Unproblematic reading

Photograph: Masson (1953: pl. XIII). Drawings: Masson-Yoyotte (1956:3, 5).



36

E.Me 3 (= MY D)

E.Me 3

←

pikre≤ ue “arwljat≤ msnord≤

The fifth letter of the second line is not in fact k k , but an invertedform of l l , given the clear onomastic identification (“ ar )wljat = Uliatow.For the penultimate letter of the second line, I adopt the reading d d instead of Masson i i , according to Schürr (2001b:103), who followson from a new direct reading of ”evoro“kin: cf. also E.Me 48. Closeobservation of the photograph in Masson-Yoyotte (1956) supports thenew reading.

Photograph: Masson-Yoyotte (1956: pl. IX). Drawing: Masson-Yoyotte (1956:10),Kammerzell (1993:146).



37

E.Me 4 (= MY E)


←

ter ÿez≤ | upe | nuol ∞. [—]sarmrol ∞ yt

The reading of the text after the second division mark is very doubt-ful, and the photography in Masson-Yoyotte is not a great help. Theending in t -t is unexpected. Could it simply be a z?

Photograph: Masson-Yoyotte (1956: pl. I). Drawings: Masson-Yoyotte (1956:19).



38

E.Me 5 (= MY F)


←

psm“kwneit≤ | ue | naria≤ | ≤ugl iq | sarl?

The inscription, published in Masson-Yoyotte (1956), was re-edited inMasson (1978). Despite this new edition, many doubts remain aboutthe reading of the second, third, and final word. Particularly trouble-some is the last letter: I adopt Schürr’s reading, l l (the photographypoints to l rather than to À ).

The first word must not be divided in two, contra Masson.Bilingual stela. In the Egyptian part, the name of the dead figure is

mentioned: Psmtk-‘wj-Njt , son of W3˙-jb-r‘-[ . . . . ]. The first name cor-

responds clearly to psm“ kwneit-, but there is no connection between theCarian and Egytian patronyms.

Photograph: Masson-Yoyotte (1956: pl. II). Drawings: Masson-Yoyotte (1956:21, 25),Masson (1978:92).



39

E.Me 6 (= MY G)


←

triqo: parma≤≤ ∞iklorul ∞i

There are no problems of reading.The stela also contains an inscription in Egyptian, but the Egyptian

names (P3-dj-st , and his mother T3-dj(t)-wsir ) do not correspond to theCarian ones. Therefore, either the stela seems to have been reused, orwe must accept a double denomination—Egyptian and Carian—at leastin the case of the deceased (the second name could be the father’sname in the Carian text).

Photograph: Masson-Yoyotte (1956, pl. III). Drawing: Masson-Yoyotte (1956:28, 30).



40

E.Me 7 (= MY H)

E.Me 8a

E.Me 8b

E.Me 7

←

tamou tanai≤ qarsio[-?]

Curved stela with images and hieroghlyphical inscriptions.The reading adopted here is taken from Schürr (2001b:118). The

controversial point is the last letter: in Masson-Yoyotte (1956), it wasread as o o, but in later works, Masson preferred the reading z ≤ (seefor instance the index of words in Masson 1978). The photographypoints clearly to o, and the most likely explanation is that the wordis incomplete, since the stone is broken just after this letter. The breakin the text is not particularly big, so only one letter, or at the most,

two, have been lost. Given the typical structure of Carian onomasticformulae, an integration qarsio[ ≤ ] would be a good solution, but I pre-fer to leave the question open.

The stela provides an Egyptian inscription that also mentions thedead man T3j-Ó p-jm=w son of T3[. . .]. The correspondence to theCarian text is evident: t amou, son of t anai .

Photograph: Masson-Yoyotte (1956: pl. IVa). Drawing: Masson-Yoyotte (1956:33).

E.Me 8 (= MY K)



41

→

a. paraeym: armon ∞i

b. para!eym: sb polo

Inscription on a bronze Apis. The fourth letter b, E, has always beena source of difficulty, due to the clear alternation with a (paraeWm/ parEeWm ). It has even been considered an independent sign (Massonnº 10), and the alternation a / E has received varying explanations.In my opinion, the simplest solution is to interpret E as an a, itsstrange form perhaps being the result of a spelling error and subse-quent correction.10

The Apis also contains an Egyptian inscription: Ó 3py dj ‘n¢ Prjm p3w m “Apis may give life to Prjm the dragoman”. The non-Egyptianname, Prjm, is logically the transcription of the Carian name Paraeymmentioned twice in the Carian section. For the problems posed by theEgyptian word p3w m, apparently translated in Carian as armon, seeChapter 11, s. v. armon.

Photographs: Masson-Yoyotte (1956: Pl. Va, Pl. VIa). Drawing: Masson-Yoyotte (1956:

43, 48).

E.Me 9 (= M 1)

10 Stephen Durnford (pers. comm. to C. Melchert) has suggested an excellent andvery likely explanation for E as a spelling error: the scribe mistakenly wrote e rightafter r, omitting a. He then immediately noticed his error and repaired it by simplyturning e into a by adding the horizontal bar (E ).

E.Me 9 (Masson 1978)



42

←

arli“≤: upe: arlio-

[m≤ ] ∞i: yjas[i≤ ]

The reading proposed here is an attempt to improve Masson’s edition.I read the penultimate letter of the first line as i i , instead of Masson’sj j , according to the photography. As for the integration of the sec-ond line, both arlio[m≤ ] and yjas[ i ≤ ] are forms documented in otherSaqqâra inscriptions. The space for integration seems to be adequate,and from a syntactic point of view, one expects two genitives in -≤ (arli “ -≤ . . . arliom-≤ . . . yjasi-≤ ).

The inscription is bilingual. The Egyptian part contains the names Jr “ (3) son of Jrym3, two non-Egyptian names that are undoubtedly theadaptation of arli “ - and arliom-, respectively.

Photographs: Masson (1978, pl. I, 1; II, 1). Drawings: Masson (1978:20, pl. XXXI, 1).

E.Me 10 (= M 2)


←

[—]q. årm≤: q[—]≤ ∞i: p∂uüi≤ mno≤[mw]don≤ ∞[i —]w≤ord≤ ∞i

New reading following the work of Schürr (2002:169, n. 9 [but note

ibid . the error u≤ ord for w ≤ ord !]).

Photograph: Masson (1978, pl. I, 2). Drawings: Masson (1978:21, pl. XXXI, 2).



43

E.Me 11 (= M 3)


→

(a) wår[—]t[——]i[—]≤ | mda ÿn

(b) [—15—]a[–]i≤ | mda ÿn

The stela represents a man and a woman. According to Masson, thedrawing included in his edition and reproduced here is too optimisticwith regard to the reading of some signs, so it is preferable to makea very prudent reading. This would mean that only the respective finalwords of each line, in both cases mda ÿn, and the existence of a pre-ceding form in -≤ — also in both lines—are definite. For the icono-graphical importance of this stela, see Martin-Nicholls apud Masson

(1978:61–70).

Photograph: Masson (1978, pl. II, 2, I II). Drawings: Masson (1978:22, pl. XXXII).



44

E.Me 12 (= M 4)


→

pjabrm | w≤ol ≤ | mwdon≤ ∞ikbjom≤ | m[no≤ ]

In the stela, the prothesis of a woman is represented, which indicates afemale character for the name pjabrm.

The possible integration of the last word was suggested by Kammerzell

(1993:213).

Photograph: Masson (1978, pl. IV, 1). Drawings: Masson (1978:23, pl. XXXIII, 1).



45

E.Me 13 (= M 5)


→

“dtat≤ | upa | w | wet≤ ∞i | mwdon≤ ∞i

As in E.M 13, here the prothesis of a woman is represented.

Photograph: Masson (1978, pl. V, 1). Drawings: Masson (1978:24, pl. XXXIII, 2).



46

E.Me 14 (= M 6)


→

irow | pikarm≤ | mwdon!≤

The surprising form of the last word, mvdoUZ mwdou≤ [sic], insteadof the usual mvdoNz mwdon≤ , was convincingly argued by Masson asa simple case of an incomplete sign U u for N n.

Photograph: Masson (1978, pl. V, 2). Drawings: Masson (1978:24, pl. XXXIV, 2).



47

E.Me 15 (= M 7)


←

arli“≤urs∞le≤kidbsi≤(Egyptian text: Jr “ (3) s3 n 3rskr s3 J‘ (?)[. . .)

Bilingual inscription. The Egyptian part was interpreted by Martin andNicholls (apud Masson 1978:86) as Jr “ (3) s3 Nrskr s3 J‘ (?)[. . . “’Jresh(a)

son of Nerseker son of Ja˙(?)-. . .”. While the correspondence of eachfirst name poses no problems ( Jr “ (3) = arli “ , as in E.Me 9), the diver-gences between Nrskr ~ urs ∞le - have been debated at length. The cor-rect solution was formulated by Kammerzell (1993:12), who discardedMartin-Nicholls’s analysis and argued persuasively in favour of an inter-pretation, Jr “ (3) s3 n 3rskr s3 J‘ - ‘‘ Jr “ (3 ) son of 3rskr son of J‘ (?)- . . .”.This solution is far more suitable for the second name in the Carianpart, and is now the commonly accepted interpretation.

Photograph: Masson (1978, pl. VI). Drawings: Masson (1978:25, pl. XXXV, 1).



48

E.Me 16 (= M 8)



←

irow | p. ikra≤ ∞i

semw≤ | mno≤mwdon≤ ∞i

No reading problems, with the exception of the initial letter of the second

word, for which Masson’s intepretation, p, seems to be the best solution.

Photograph: Masson (1978, pl. VII, 1). Drawings: Masson (1978:26, pl. XXXV, 2).

E.Me 17 (= M 9)



49

←

(a) ta“ubt≤kuari≤b-ar | ≤enniqau≤ptnupi

(b) idmuon≤∞i | mdayn∞i


←

“arnai≤

upe | quq≤bem≤ ∞i md-a ÿn

No reading problems.

Photograph: Masson (1978, pl. VII, 2). Drawings: Masson (1978:27, pl. XXXV, 3).

E.Me 18 (= M 10)



50

E.Me 19

The two parts have been inscribed by diff erent hands. Whilst neithersection poses problems of reading, the overall structure of the text

remains obscure.

Photograph: Masson (1978, pl. VIII, 1). Drawings: Masson (1978:28, pl. XXXV, 4).

E.Me 19 (= M 11)

←

pnu≤ol zmu≤ ∞i

Curved stela (‘stèle cintrée’). No problems of reading.

Photograph: Masson (1978, pl. VIII, 2). Drawing: Masson (1978:29).



51

E.Me 20 (= M 12)

E.Me 20

←

uqsi | “rwli≤ ∞i | mwdon≤ ∞i

False-door stela. Unproblematic reading.

Photograph: Masson (1978, pl. IX, 1). Drawing: Masson (1978:29).

E.Me 21 (= M 13)




52

←

punw≤ol ≤: somne≤

q ÿblsi≤ ∞i

False-door stela. No reading problems.

Photograph: Masson (1978, pl. IX, 2). Drawings: Masson (1978:30, pl- XXXVI, 1).

E.Me 22 (= M 14)

E.Me 22

E.Me 23

←

artay≤: upe: [. . .

False-door stela. In the damaged part there is sufficient space for acomplete onomastic formula.

Photograph: Masson (1978, pl. X, 1). Drawing: Masson (1978:30).

E.Me 23 (= M 15)



53

→

ap[—]ws

a[rb]ikarm≤ ∞i

False-door stela. The integration of the second word was proposed byKammerzell (1993:214), who also claimed to have identified a formerinscription under the current one. Nevertheless, his reading of this pre-vious text does not lead to any connection with the rest of the Carianmaterials: (a)?p[. . .]ws # --]b[-]aubm[-].

Photograph: Masson (1978, pl. X, 2). Drawing: Masson (1978:31).

E.Me 24 (= M 16)

E.Me 24

→

tdu≤ol kbos | “amsqi[. . .?

False-door stela. The crack in the stone makes it impossible to estab-

lish whether the last word is complete.

Photograph: Masson (1978, pl. XI, 1). Drawing: Masson (1978:31).



54

E.Me 25 (= M 17)

E.Me 25

←

“ayriq | parpeym≤ ∞i yiasi

False-door stela. The order of reading adopted here is diff erent to that

used by Masson (who began from yiasi ), and is the same as in Kammerzell(1993:214). Contra Masson and Kammerzell, I believe that there is notext after yiasi . It is true that the stone is damaged, but the existing part shows evidence enough to assume that yiasi is a complete wordand that the text finishes here.

Photograph: Masson (1978, pl. XI, 2). Drawing: Masson (1978:32).



55

E.Me 26 (= M 18)

11 The inscription appears marked with an asterisk (*) in the transliteration appen-dix (Kammerzell 1993:214), which means that he controlled the text, but if he wasable to recognize any letter before ]u≤ , one would expect it to be marked as a doubt-ful reading, rather than using [ ].

E.Me 26

←

[. . .]u≤ | upe sa | triel≤ | mrsi≤

False-door stela. Kammerzell completes the first word as [ arm ]u≤ , buthe does not propose any argument for this integration.11 In any case,the space preceding u≤ is small, so it is likely that only two or threeletters have disappeared.

Photograph: Masson (1978, pl. XII, 1). Drawing: Masson (1978:32).



56

E.Me 27 (= M 19)

E.Me 27


←

irow≤: psH ÿm[-]≤pttu≤: mno≤

False-door stela.

Photograph: Masson (1978, pl. XII, 2). Drawing: Masson (1978:33).

E.Me 28 (= M 20)



57

←

sanuq≤ | ue | pntmun≤ ∞i

mwdon≤ ∞i

False-door stela. Unproblematic.

Photograph: Masson (1978, pl. XIII, 1). Drawings: Masson (1978:33, pl. XXXVI 2).

E.Me 29 (= M 21)

E.Me 29

←

s[—]et≤ | [ue] | ynemori≤ | mwdon≤

False-door stela.The second word is definitely ue (already suggested by Masson 1978).

Photograph: Masson (1978, pl. XIII, 2). Drawing: Masson (1978:34).



58

E.Me 30 (= M 22)

E.Me 31

E.Me 30

←

“aru≤ol pl eq≤ ∞i: ≤ugl i≤

False-door stela. Unproblematic.

Photograph: Masson (1978, pl. XIV, 1). Drawing: Masson (1978:34).

E.Me 31 (= M 23)



59

←

wnuti≤ | kwar≤ mHm≤ ∞i | mwdon≤ [ ∞ ]i

False-door stela. The integration of the last word, already proposed byMasson, is suitable. For the second sequence, I adopt the segmenta-tion kwar ≤ mHm≤ suggested by Schürr (apud Vittmann 2001:48, n. 40).

Photograph: Masson (1978, pl. XIV, 2). Drawing: Masson (1978:35).

E.Me 32 (= M 24)

E.Me 32

←

iturow≤ | kbjom≤ | ∞i en | mw[d]on≤ ∞i

False-door stela. Unproblematic. The integration of mw [ d ]on≤ is clear.

Photograph: Masson (1978, pl. XV, 1). Drawing: Masson (1978:35).



60

E.Me 33 (= M 25)


E.Me 33

←

(a) idmns | myre≤ ∞i | mdayn ∞i(b) idmns | myre≤ ∞i

False-door stela. As Masson (1978) observes, both (a) and (b)—in facta partial copy of (a)—seem to have been written by the same hand.

Photograph: Masson (1978, pl. XV, 2). Drawing: Masson (1978:36).

E.Me 34 (= M 26)



61

←

me®≤ | somne≤ | t®∞ata[r]≤

False-door stela. The integration of the last word, proposed by Masson,is based on the occurence of the same word in complete form in E.Me41.

Photograph: Masson (1978, pl. XVI, 1). Drawings: Masson (1978:36, pl. XXXVI, 3).

E.Me 35 (= M 27)

E.Me 35

←

ntokris | dw≤ol ≤ | mwdon≤ ∞i

False-door stela. The current solution to the uncertainty regarding thefourth letter of the first word (a form of l l or an inverted K k ?, seeMasson 1978:37) is to favour the last option, which facilitates a verygood onomastic identification (ntokris = Nitokris, a well known Egyptianfemale name, see Chapter 11, s. v. ).

Photograph: Masson (1978, pl. XVI, 2). Drawing: Masson (1978:37).



62

E.Me 36 (= M 28)

E.Me 37

E.Me 36

← wksmu≤ | wpe | lkor≤ ∞ j→

qarpsi≤

False-door stela. This order of reading was already adopted in Adiego1993a—see also Kammerzell 1993—and diff ers from Masson’s (he readqarpsis as the first word).

Photograph: Masson (1978, pl. XVII, 1). Drawing: Masson (1978:37).

E.Me 37 (= M 29)



63

←

qlal i≤ | [. . .]

tkrabi≤

False-door stela. It is not possible to calculate the extent of the letterslost after qla l i ≤ (see Masson 1978:38).

Photograph: Masson (1978, pl. XVII, 2). Drawing: Masson (1978:38).

E.Me 38 (= M 30)

E.Me 38

←

“ÿin≤ | upe | arie?≤ ∞i ted

False-door stela. This reading was made by ”evoro“kin, who has con-trolled the original. It diff ers from Masson’s interpretation in identifying e (with doubts) in arie ? ≤ and, particularly, in the interpretation of theantepenultimate letter, a clear t t , not o o as indicated by Masson.

Photograph: Masson (1978, pl. XVIII, 1). Drawing: Masson (1978:39).



64

E.Me 39 (= M 31)

E.Me 39

←

[. . .]s? | ar∞ila≤mno≤

False-door stela. Diff erent reading order (cf . also Kammerzell 1993:215):Masson’s reading began at mno≤ .

Photograph: Masson (1978, pl. XVIII, 2). Drawing: Masson (1978:39).

E.Me 40 (= M 32)

E.Me 40



65

→

plqo | pikrm≤ ∞i | mwdon≤ ∞i

False-door stela. The sign l of the second word is actually a k inverted(= /k/) (already noted in Adiego 1993a).

Photograph: Masson (1978, pl. XIX, 1). Drawing: Masson (1978:40).

E.Me 41 (= M 33)

E.Me 41

→

|? or≤ | wpe | qdar®ou≤ | t®∞atar≤

False-door stela. The initial vertical small stroke and the brevity of thefirst name, or ≤ , perplexed Masson, who thought that, ‘pour des raisonsobscures, le lapicide n’ait jamais gravé le début du premier mot.’ Infact, it seems to me most likely that the stoke is an accidental andintrusive mark. In any case, contrary to Masson, I see no problem inaccepting the existence of a name or-≤ in Carian.

Photograph: Masson (1978, pl. XIX, 2). Drawing: Masson (1978:40).



66

E.Me 42 (= M 34)


←

(1) arjom≤: ue: mwsat≤: ∞i: mwdon≤: ∞i(2) tbridbd≤: ∞i

False-door stela. Very clear reading (‘belle stèle intacte’, Masson 1978:41).

Photograph: Masson (1978, pl. XX, 1). Drawings: Masson (1978:41, pl. XXXVI, 4).

E.Me 43 (= M 35)

E.Me 43



67

←

(a) l ÿ∞si≤ | upe | “rquq≤ ∞i | ksolb≤

(b) arliom≤ | mno≤ ∞i

False-door stela. Diff erent reading order (already in Adiego 1993a,Kammerzell 1993): Masson read firstly (b), then (a). The reading of theinitial letter of the third word of (a) is also diff erent: f “ instead of Masson’s a “ .

Photograph: Masson (1978, pl. XX, 2). Drawing: Masson (1978:41).

E.Me 44 (= M 36)


←

(a) apmen “rquq≤ kojol ∞i(b) mwton≤ ∞i

False-door stela. Reading order diff erent from Masson’s (already in

Adiego 1993a, Kammerzell 1993): f “ instead of a a at the beginning of the second word of (a).

Photograph: Masson (1978, pl. XX,I 1). Drawings: Masson (1978:42, pl. XXXVII, 1).



68

E.Me 45 (= M 37)

→

[q? ]lal is[?]iam≤ ∞i

alos ∞arnos

False-door stela. The integration of the initial letter of the first wordis already in Kammerzell (1993), from the parallel form qla l i ≤ in E.Me37.

Photograph: Masson (1978, pl. XXI, 2). Drawing: Masson (1978:43).

E.Me 46 (= M 38)

E.Me 45

E.Me 46



69

←

(a) ÿasd≤ | yi≤∞biks≤ ∞i

(b) mwdon≤ ∞i

False-door stela. Diff erent reading order (Adiego 1993a, Kammerzell1993). Masson’s reading began with (b).

Masson (1978) reads a sign X ∞ between ≤ and b in the second wordof (a). A careful observation of the photography shows that this allegedsign appears placed between ≤ and b without keeping distance withthem, unlike the other letters of the inscription. As Schürr (1996b:151)aleady pointed out, this

∞seems to be ‘eine Illusion’, and I have there-

fore decided to discard it.As for the final sequence of this word, I maintain that the reading

generally accepted until now, yi ≤ biks ∞i ≤ (see Masson’s drawing in thebottom part of the illustration), must be replaced by a more natural ÿi ≤ biks ≤ ∞i (cf. ÿ≤ biks in C.xx 2): in fact, the three later letters can beread beginning at ≤ .

Photographs: Masson (1978, pl. XXII, 1, 2). Drawing: Masson (1978:43).

E.Me 47 (= M 39)


←

tqtes | paraibrel ≤ ∞i | mn[o-?]

With this inscription, a series of texts on diff erent objects and frag-ments begins (E.Me 47–64). In this case, it is a limestone plaque whose

use is unclear (Masson 1978:43).

Photograph: Masson (1978, pl. XXIII, 1). Drawings: Masson (1978:43, pl. XXXVII, 2).



70

E.Me 48 (= M 40)


←

[—]j[-]≤[-]owt≤∞i: msn-ord≤

Note our reading of d d instead of i i in the final word, cf. above,E.Me 3.

Photograph: Masson (1978, pl. XXIII, 2). Drawings: Masson (1978:44, pl. XXXVII, 3).

E.Me 49 (= M 41)




72

←

arli“≤ | psikro≤ [?

ue

Upper fragment of a curved stela (‘stèle cintrée’). The reading off eredhere is new: in my opinion, the last sign of the second word is clearlyz /≤/, and not N /n/ (see plate xxiv in Masson 1978). As for ue ,which appears just below arli “≤ , it is possible that it must be read imme-diately after this latter word (arli ““ ue | psikro≤ ), which would representa more logical structure (ue, upe , etc. always appear in a second position).

Photograph: Masson (1978, pl. XXIV, 1). Drawings: Masson (1978:46, XXXVII, 5).

E.Me 52 (= M 44)

E.Me 52

←

[. . .] ardybyr≤ | md[. . .]

Upper fragment of a false-door stela.

Photograph: Masson (1978, pl. XXIV, 2). Drawing: Masson (1978:46).



73

E.Me 53 (= M 45)

E.Me 53

←

[. . .]q≤si≤

Fragment of a false-door stela.

Photograph: Masson (1978, pl. XXIV, 3). Drawing: Masson (1978:47).

E.Me 54 (= M 45a)

E.Me 54

←

[. . .] mrsj[. . .]


Photograph: Masson (1978, pl. XXV, 1). Drawing: Masson (1978:47).



74

E.Me 55 (= M 46)

E.Me 55

←

[. . .] psma[ ≤/“k . . . ]

The integration proposed here is based on the identification of thebeginning of the typical Egyptian name, Psammetichus. Given that thisname can appear in Carian spelled either with “ or ≤ , I choose not to

discard either possibility.


E.Me 56 (= M 47)

E.Me 56

←

[. . .] “ark[bi/jom . . .?]

Fragment of a false-door stela.I follow the suggestion of Meier-Brügger (1979b) and Ray (1982b:

189) of reading f “ instead of i i (Masson 1978) at the beginning.



75

The integration is based on the form “ arkbiom, which appears in otherinscriptions.


E.Me 57 (= M 47a)

E.Me 57

←

[. . .]i≤ ∞i

Fragment of a false-door stela. Clear ending in genitive + particle ∞i .

Photograph: Masson (1978, pl. XXVI, 1). Drawing: Masson (1978:48).

E.Me 58 (= M 47b)

E.Me 58

←

[. . .]s≤ ∞i

Undetermined fragment. A typical ending in genitive + particle ∞i , asin E.Me 57.




76

E.Me 59 (= M 48)

E.Me 59

←

[. . .]utr[. . .]



E.Me 60 (= M 48a)

E.Me 60

→ (?)[ . . . ]∞≤





77

E.Me 61 (= M 48b)


[. . .]i

Undetermined fragment.

Drawing: Masson (1978:48).

E.Me 62 (= M 48c)

[ . . . ]≤[ . . . ]

Undetermined fragment.


E.Me 63 (= M 48d)


E.Me 63

←

(a) idyes≤(b) m [?



78

Fragment of a false-door stela. Although Masson indicates the presenceof a M m, he does not transcribe it in his edition of the text. I follow

Kammerzell (1993) in including the letter in the transcription.


E.Me 64 (= M 49)


←

(a) [. . . u? ]p.e : pd[ (b) [. . .]mi [. . .]

Fragment of a curved stela. I concur with Kammerzell (1993:217) inthe interpretation of the first word.


E.Me 65 (= Stela of Abusir)

E.Me 65 (Schürr 2003)

←

u[. . .]m | punm[-]≤ | mudo[n]≤

See Schürr (2003) for a new attempt at reading this extremely difficultinscription. I adopt his reading (with the integration of [n] in mudo[-]≤ )and reproduce his drawing (Schürr 2003:94). Former references: Masson(1978:91), who was the first to confirm the Carian character of the



79

inscription, already suggested by L. H. Jeff ery: Kammerzell (1993:138–139).

Photograph: Masson (1978, pl. XXX). Drawings: Masson (1978:91), Kammerzell(1993:138), Schürr (2003:94).

E.Me 66 (= Kammerzell *180) —].[. . u][. . .]p[-]n[—

Concerning this inscription on a fragment of stela, conserved in the

British Museum, see Kammerzell (1993:144, n. 101). To my knowl-edge, neither the photography nor the drawing have yet been pub-lished, so I limit myself to follow Kammerzell’s reading.

3. Abydos (E.Ab)

The Carian inscriptions from Abydos were found and copied for thefirst time by Archibald H. Sayce, who published them in his most

important work on Carian (Sayce 1887[1892]). He discovered them inthe temples of Ramses II and, more significantly, Seti I.Regrettably, there is not a definitive edition of the Abydos graffiti.

Jean Yoyotte seemed to have revised these graffiti and even to havefound some others in 1955–1956, but his work has never been pub-lished. In his index of Carian words, Masson (1978) pointed out thathe had adopted, “insofar as it was possible”, Yoyotte’s revision, andadded the most important unpublished graffiti found by Yoyotte. Meier-Brügger (1979b) followed Masson in his collection of Carian inscrip-

tions in transcription. ”evoro“kin also (1965) used Yoyotte’s new readings.My attempts to obtain more information on this sub-corpus from JeanYoyotte were unsuccessful.

As a result, our present knowledge of the Carian graffiti of Abydosis unsatisfactory. Neither Masson (1978) nor Meier-Brügger (1979b) off erany additional information on the readings adopted for each graffiti,so that in cases where Masson’s readings are not coincident with thosemade by Sayce (Sayce 1887[92] followed by Friedrich 1932), it is impos-sible to ascertain whether this is due to the revisions of Yoyotte, or itis simply a new interpretation of Sayce’s copies based on mere divina-tio. However, thanks to ”evoro“kin (1965) and to some notices scatteredthroughout the works of Masson, it is possible to establish in a greatnumber of cases that the text is in fact the result of Yoyotte’s collation.



80

As for the unpublished graffiti, Masson and Meier-Brügger take intoaccount, as already established, only the most important examples: they

off er only 8 from a total of at least 34. Moreover, given that bothauthors limit themselves to transcribing the texts, nothing can be saidabout the graphic traits of these inscriptions. The only exception isE.Ab 43 (= Ab. 34 Y)—in fact an already known graffito—publishedin Murray (1904: 10 and pl. XII) and included in ”evoro“kin (1965)(= 24 ” ), though absent in the editions of Sayce and Friedrich.

Confusion with the Abydos graffiti is even augmented by the factthat Bork (1930) re-ordered Sayce’s list of inscriptions, putting togetherunder the same reference number those inscriptions that he believedhad the same content. This decision was not only disputable from anepigraphical point of view, but was in fact revealed as altogether erro-neous, since in some cases Bork grouped together texts which were notstrictly identical. This explains the posterior use of redundant referencenumbers like “2a”, “2b” in works by Friedrich or Masson. In the newnumber system introduced here, I prefer to assign each inscription adiff erent number, even in the case of inscriptions that seem to havethe same content.

Given the lack of an up-to-date edition, I adopt in general termsMasson’s readings, but in some cases I also include new attempts atimproving these readings, mainly from Schürr (in various papers) andVittmann (2001). Since the Abydos graffiti pose such problems, in thefollowing sections, I will comment briefl y upon each inscription andwill explain the reading adopted here. For want of a better solution, Ishall limit myself to reproducing Sayce’s drawings,12 although in somecases the present readings may not coincide.

E.Ab 1 (= Ab. 1 F = Sayce 1887[92] 1)

12 For technical reasons, the source of my illustrations will be Friedrich (1932), whereSayce’s drawings are reproduced.

E.Ab 1 (Friedrich 1932)



81

←

pisiri

Revised by Yoyotte (Masson 1974:131). There are no reading problems.

E.Ab 2 (= Ab. 2a F = Sayce 1887[92] 2)


←

panejt iarja≤

Expressly mentioned as revised in Masson-Yoyotte (1956:13, 63). Masson’sreading is adopted, but with diff erent segmentation (the first name isclearly an Egyptian one whose final element is -nejt ).

E.Ab 3 (= Ab. 2b F = Sayce 1887[92] 3)


→

ptn“e | ibarsi≤

Revised by Yoyotte, according to Masson-Yoyotte (1956:63). For thesecond name, I follow Schürr’s proposal of reading i b arsi ≤ , not ‡irarsi ≤ (Masson).

E.Ab 4 (= Ab. 3b F = Sayce 1887[92] 24)




82

←

“amow ltari≤

Revised by Yoyotte (cf. Masson-Yoyotte 1956:39). Masson’s reading isadopted, but with diff erent segmentation (Masson’s reading would indi-cate ‡“ amowl tari ≤ .

E.Ab 5 (= Ab. 3c F = Sayce 1887[92] 25)


←

“amow ltari[ ≤ ]

Revised by Yoyotte (cf. Masson-Yoyotte 1956:39).

Masson groups Ab. 3b F and Ab. 3c F under a single epigraph, “Ab3bc F”. This procedure, debatable even if the content were exactly thesame, is inappropriate here, as the second word appears complete inthe case of E.Ab 4 = Ab 3b F, but incomplete in E.Ab 5 = Ab. 3cF (ltari[ ≤ ]). On segmentation, see above E.Ab 4.

E.Ab 6 (= Ab. 4 F = Sayce 1887[92] 5)


←

“aru≤ol | ÿrsbe | pdubi≤

Revised by Yoyotte (cf. Masson 1978:34).



83

E.Ab 7 (= Ab 5a F = Sayce 1887[92] 6)


←

pl at | pals≤

E.Ab 8 (= Ab 5b F = Sayce 1887[92] 9)


←

pl at pals≤

E.Ab 9 (= Ab 5c F = Sayce 1887[92] 10)


←

pl at pals≤

E.Ab 7, 8 and 9 do not pose reading problems. The three inscriptionsare grouped together by Masson.



84

E.Ab 10 = (Ab. 6 F = Sayce 1887[92] 7)

←

piubez

qurbo≤

Revised by Yoyotte (Masson-Yoyotte 1956:9), this inscription formedby two personal names still raises serious reading problems. Masson’sreading (1978) distances the first name from the very similar examplefound in E.Ab 15. Schürr’s reading seems preferable (Schürr 2000:172,n. 7), because it allows us to connect this name with the second namein E.Me 1 and with the name mentioned in E.Ab 15 (although we

cannot establish an absolutely certain link). Regarding the second name,although in Masson-Yoyotte (1956:9) the reading ‡≤ urbo≤ is preferred(thus also ”evoro“kin 1965), in Masson (1978), Sayce’s reading qurbo≤ is restored.

E.Ab 11 (= excluded by Masson [Ab 7 F = Sayce 1887[92] 8])



←

≤?

[. . .]it

Since the letters are apparently Carian, I reintroduce this very frag-mentary inscription to the corpus, even though it is of no use to us.



85

E.Ab 12 (= Ab. 8a F = Sayce 1887[92] 11)


←

untri uantrpo

E.Ab 13 (= Ab. 8b F = Sayce 1887[92] 20)


←

untri | uantrpu≤

E.Ab 14 = Ab. 9 F (= Sayce 1887[92] 12)


←

abrq∞[ . . . ?

Regarding E.Ab 12, 13 and 14, I am unsure as to whether Masson’sreadings are the result of a revision of the texts. Lacking a better read-ing, I adopt Masson’s.

E.Ab 15 = Ab. 10 F (= Sayce 1887[92] 13)




86

←

pdubez or≤

Revised by Yoyotte (Masson-Yoyotte 1956:13). The reading in Masson(1978) is not the same as in Masson-Yoyotte (1956:13): ‡ pdubtzor ≤ vs.‡ pdubtzt r ≤ (with r ≤ as a part of a diff erent graffito!) respectively. I adoptthe reading proposed by Schürr and Vittmann (see Vittmann 2001:42).

E.Ab 16 = Ab. 11 F (= Sayce 1887[92] 14)


←

nprosn≤

Revised? In any case, it does not pose reading problems.

E.Ab 17 = Ab. 12 F (= Sayce 1887[92] 15)


←

pa[-]in[-]t≤

Revised? I follow Masson (1978)

E.Ab 18 = Ab. 13a F (= Sayce 1887[92] 16)


←

tamosi | inut≤



87

E.Ab 19 = Ab. 13b F (= Sayce 1887[92] 17)

13 Vittmann (2001:43) mentions Yoyotte’s collation of E.Ab 20.


←

tamosi utnu≤

E.Ab 20 = Ab. 14 F (= Sayce 1887[92] 18)

←

ninut | tamosi≤

E.Ab 21 = Excluded by Masson [Ab. 15 F (= Sayce 1887[92] 19)]



←

to[-]a[—]l tamosi u?tnu≤?

These four graffiti seem to coincide in that they contain the same per-sonal name. Rather surprisingly, in Masson’s list, this name is read astrmosi- in E.Ab 18, E.Ab 19, and as tamosi- in E.Ab 20, while E.Ab 21is excluded. These readings were adopted in Adiego (1993a). It is quite

possible—although I cannot confirm it—that this discrepancy has arisenbecause only E.Ab 20 was revised by Yoyotte.13



88

Vittmann (2001:43) has correctly noted this incongruity, and hasoff ered a convincing reading of tamosi- everywhere, a view that is sup-

ported here.14 More problematic, and still unresolved, are the respec-tive readings of the names that accompany tamosi- in the four inscriptions.They look very similar, but it is not easy to imagine a complete iden-tity for all the forms. In the case of E.Ab 20, I continue to work with”evoro“kin’s suggestion of reading ninut instead of ‡∞inut . For the othernames found in the remaining graffiti, Masson’s readings are adopted.I include Ab 15 F, excluded by Masson for reasons that remain unknownfor me.

E.Ab 22 = Excluded by Masson [Ab. 16 F (= Sayce 1887[92] 21])

14 A theory already envisaged by Ray, see Ray (1994:205).


←

[-]untlau[-]|

This brief graffito is very difficult to read, but it seems to show trueCarian letters. For this reason, I include it in the list (as ”evoro“kindid: ”evoro“kin (1965), 34 ” ).

E.Ab 23 = Ab. 17 F (= Sayce 1887[92] 22)


←

be≤ol



89

E.Ab 24 = Ab. 18 F (= Sayce 1887[92] 23)


←

[. . .] arli“

Revised by Yoyotte (cf. Masson 1978:51). This revision identifies the

well-known Carian name arli “ (contrary to Sayce’s reading, ‡araii ).

E.Ab 25 = Ab. 19 F (= Sayce 1887[92] 26)


←

ttubazi katt ÿri≤

Revised by Yoyotte (Masson-Yoyotte 1956:4, 33). According to Masson-Yoyotte (1956:33), the sign for t here adopts the form C, as in E.Me7. The segmentation adopted here diff ers from that of Masson and isbased on Schürr’s works.

E.Ab 26 = Ab. 20 F (= Sayce 1887[92] 27)


←

[. . .]pri | ptnuq?i?

Masson (1978:28) points out that this graffito was not rediscovered byYoyotte in 1956. Note that in Adiego (1993a) the end of the (incom-plete) first word was transcribed incorrectly as †] pre , instead of ] pri (a



90

mistake noted by Vittmann 2001:44). Here I adopt one of the twointeresting corrections proposed by Vittmann for the second word

(Vittmann 2001:44).15

E.Ab 27 = Ab. 21 F (= Sayce 1887[92] 28)

15 The other possibility suggested by Vittmann is ptnuti . Both readings allow us toidentify good correspondences with Egyptian personal names (see Chapter 11, s. v. ptnuq ? i ).


←

yysmt≤oHa[

Divergences between Sayce’s and Masson’s readings lead me to thinkthat the graffito was revised by Yoyotte, but I cannot confirm this. Ifollow Masson’s reading. Note that the drawing is based on Sayce/Friedrich editions.

E.Ab 28 = Ab. 22 F (= Sayce 1887[92] 29)


←

Hosurz | srton[-]t[. . .?]

(or: →

. . . +t[-]nota

/rs | za

/rusoH/l ? Schürr)

In Masson-Yoyotte (1956:68) this graffito is referred to as “not redis-covered”. Indeed, the reading adopted by Masson seems to be basedon Sayce’s drawing. I maintain Masson’s reading, but I also consideras a plausible alternative the suggestions made in Schürr (1996a:65)that the text could be read in the opposite direction and that the let-ter H could in fact be L = l (a possibility to be ruled out if Masson’sreading is upheld, since it is very unlikely that Carian L l would appear

at the beginning of a word).



91

E.Ab 29 = Ab. 24 F (= Sayce 1887[92] 31)


←

[. . .]r[--]tnit

Graffito consisting of only five signs. The readings given by Sayce and

Masson coincide.

E.Ab 30 = Ab. 25 F (= Sayce 1874 5+6 = Sayce 1887[92] 32)


←

bid≤lemsa: “a[ru]≤ol : “a ÿdiq≤[. . .]all ia: bsis

I am unaware of whether it has been revised. In any case, Sayce’sreading is generally followed by Masson, who only improves the read-

ing of the second word in order to obtain a well-known Carian name(fa[ ru ]zoL = “ ar [ u≤ ]ol ).

E.Ab 31 = Ab. 26a F (= Sayce 1887[92] 33)


←

∞aye



92

E.Ab 32 = Ab. 26b F (= Sayce 1887[92] 34)


←

∞arr≤

E.Ab 33 = Ab. 27 F (= Sayce 1887[92] 35)


←

“arpt≤ | p[-]lu≤

For E.Ab 31–33 I follow Masson’s readings.

E.Ab 34 = Excluded by Masson [= Ab. 28 F (= Sayce 1887[92] 36)


←

dbkrm [-]kb?[

Despite the difficulties of interpretation, Schürr has been able to iden-tify a good Carian sequence, dbkrmº (Schürr 1996b:154, n. 8; see Chapter11, s. v. ). For this reason I have decided to reinstate this graffito to theAbydos sub-corpus.



93

E.Ab 35 = Ab. 29 F (= Sayce 1874 4 = 1887[92] 37)


←

u≤ol | mi∞≤≤ kdu.usi≤

For this graffito, I adopt the reading proposed by Schürr (2001b:108,

111), apparently based on the collations of Yoyotte and ”evoro“kin.E.Ab 36—E. Ab 43: “inédits de Yoyotte”, unpublished new graffitifound by Yoyotte in 1955–1956. I include only those transcribed byMasson (1978). As mentioned above, E.Ab 43 = Ab. 34 Y is not strictlyan unpublished graffito: a drawing of it appeared in Murray (1904),and it is also included in ”evoro“kin (1965) (= 24 ” ).

E.Ab 36 = Ab. 8 Y [-]ars, ∞[-]urb≤

E.Ab 37 = Ab. 9 Y “arur≤

E.Ab 38 = Ab. 15 Y piew

E.Ab 39 = Ab. 26 Y uarila[-]os≤

E.Ab 40 = Ab. 27 Y iall i | q∞blio≤

E.Ab 41 = Ab. 28 Y ttbazi kt?tri≤I follow Schürr’s reading (see Schürr 1996a:60).

E.Ab 42 = Ab. 29 Y “aru≤[ . . ?



94

16 ”evoro“kin (1965:313–314) unifies Ab 3a, b, c under a sole entry (20 ” ).

E.Ab 43 = Ab 34 Y = Murray (1904)

E.Ab 43 (Murray 1904)

?-rasAlthough Murray’s drawing points clearly to an initial k k (by exten-sion, kras ), Masson’s reading leaves the sign unread.

Graffiti excluded from our collection:Ab. 3a F (= Sayce 1887[92] 4), also excluded by Masson.16 It seems

to begin with “ am[, which explains how Bork could link it to E.Ab 4and E.Ab 5.

Ab. 23 F (= Sayce 1887[92] 30), also excluded by Masson and by

”evoro“kin. Apparently it has not been revised since Sayce (Masson-Yoyotte 1956:46: ‘non retrouvé par J. Yoyotte; Masson 1974:131 n. 43:“non revisé”). The graffito could possibly be Greek.

Ab. 30 F (= Sayce 1887[92] 38), also excluded by Masson. It is analmost illegible graffito, which cannot even be definitely identified asCarian.

The drawings made by Sayce are reproduced below:

Ab 3a F (Friedrich 1932)

Ab 23 F (Friedrich 1932)



95

4. Thebes, Tomb of Montuemhat (E.Th)

The case of the corpus of Carian inscriptions found in Thebes (tombof Montuemhat) is even worse than that of Abydos: the corpus hasnever been published correctly. Leclant, the first to find fifteen graffitiengraved on the walls of the open court of the tomb, never producedan edition of them, instead publishing only the photographs of two

graffiti (E.Th 13 and E.Th 14, Leclant (1951, tab. LXIV). ”evoro“kinincluded in his book on Carian (”evoro“kin 1965) a merely provisionaledition, with drawings of these graffiti. This edition forms the basis of Masson’s transcriptions (in Masson 1978, reproduced in Meier-Brügger1979b), although Masson aknowledges that these were improved thanksto the photographs and notes of Leclant that he had at his disposal.On the other hand, some years later, ”evoro“kin himself revised Leclant’sgraffiti and discovered some new examples, although these have not

been published either, and the new graffiti are circulated only in pri-vate copies. In the Carian conference held in Rome in 1993, Massonannounced a new enterprise in order to adequately publish the Thebancorpus (cf. Masson 1994a), but I do not know if this enterprise is infact underway.

More recently, Diether Schürr, who was able to gain access to theentire dossier of drawings, has drawn up a provisional corpus of Thebangraffiti in transcription, which he has kindly made known to me. Thiscorpus consists of four parts: (A) inscriptions found in the entrance of

the temple. I already knew of the drawings of these texts thanks to aprivate copy made by ”evoro“kin and circulated among scholars; (B)inscriptions found at the end of the corridor; (C) graffiti located in thevestibule, the drawings of which I have never seen, and from which Iknew only certain texts, occasionally cited by ”evoro“kin in recent arti-cles; and (D) graffiti from the open court, in fact already mentionedin Leclant’s corpus (with some additions).

I choose to adopt Masson’s readings for (D) (Leclant’s graffiti), intro-

ducing some corrections suggested by ”evoro“kin in published or unpub-lished works, and I include the drawings published in ”evoro“kin (1965).For (A), I provide the transcription, following Schürr and controlling ”evoro“kin’s drawings, but I refrain from reproducing the drawings,

Ab 30 F (Friedrich 1932)



96

since they are as yet unpublished; my aim is merely to contribute totheir use in linguistic research, not to actually produce an epigraphi-

cal edition, which must be made by ”evoro“kin himself. For (B) and(C), I limit myself to reproducing Schürr’s transcriptions, adapting themto the decipherment system followed in this study. According to stan-dard practice within the Carian corpus, I begin my numeration of theinscriptions with the first to be published (Schürr D), which are thenfollowed by the remaining inscriptions (A, B, C).

E.Th 1 = 47 ” (= D 1)

E.Th 1 (”evoro“kin 1965)

←

uarbe

E.Th 2 = 48 ” (= D 3)

←

dt ÿbr | kbokt≤k≤at ÿbr

E.Th 3 = 49 ” (= D 2)



pla?t



97

E.Th 4 = 50 ” (= D 5)


→ dokmmpint seqqejewsk | mqt jq← ÿpdnmwd

The second line was usually read in the opposite direction. The newreading was suggested by Schürr. I am somewhat sceptical about thelast sign, now read & d, a letter rarely found in Thebes (in fact, it onlyappears in one other graffito, still unpublished and whose drawing Ihave not seen; E. Th 14, see below). Former readings pointed ratherto B b (cf. ”evoro“kin’s drawing here reproduced). The reading of theletters preceding q t jq at the end of the first line probably comes froma new collation of the graffito, the old drawing showing a lacuna and

some illegible signs.

E.Th 5 = 51 ” (= D 6)


←

d ÿbr | t®∞atr≤The second word, which posed serious problems in ”evoro“kin (1965),was re-read in a new collation by ”evoro“kin as t ® ∞atr ≤ , undoubtedlyusing the model of the form t ® ∞atar ≤ that appears twice in Saqqâra(”evoro“kin, “Corrections to Existing Copies”, ms.). Schürr, however,does not rule out a reading of l l instead of ® .



98

E.Th 6 = 52 ” (= D 7)

→

bebnd

The last letter, d , does not appear in the old drawing by ”evoro“kin

and was thus absent in older collections of Carian inscriptions (forinstance, Adiego 1993a, where the reading bebn was given).

E.Th 7 = 53 ” (= D 10)



→

wljat

The doubts about the second and fourth letters, reflected in the draw-ing, can be dismissed due to the clear onomastic connection of theword (→ wljat ).

E.Th 8 = 54 ” (= D 11)


→

qutbe



99

E.Th 9 = 55 ” (= D12)


→

kudtubrThe initial letter is now read as k k .

E.Th 10 = 56 ” (= D 8)


→

a?q≤baq ewm ≤emot qtblo owdown[. . .]mwarudk≤o mlane

The most difficult letter to determine is the first one: read k by ”evoro“kin(1965) and in the Masson/Meier-Brügger collections. In Adiego (1993a),a reading w was introduced, based on ”evoro“kin’s observations. Schürrnow reads it as a ?. For the rest of the letters, the present reading diff ersvery little from that off ered in Adiego (1993a). Only a letter b imme-diately after owdown is not read by Schürr.

E.Th 11 = 57 ” + 58 ” (= D 12)

(”evoro“kin 1965, 57 ” )



100

→

psma≤[k] [?| nm[ mplat | o[

Schürr’s dossier puts these two lines together, formerly interpreted asdiff erent graffiti. The clearest element is the presence of the name psma ≤ k .

E.Th 12 = 59 ” (= D 13)

E.Th 11 (”evoro“kin 1965, 58 ” )


→

?-˚bjqmq ewmlane qeb≤t | u[. . .]ü≤q | qwsal | mqabaewleqo“oski.oms

This reading coincides almost exactly with that presented in Adiego(1993a), where some corrections from the work of ”evoro“kin hadalready been introduced. The sole diff erence introduced by Schürr’sreading involves the final sequence kioms , formerly read as ‡lkoms .



101

→ dbiks | kbjoms | wdwn | sb a≤b≤t← ewm

The reading corresponds precisely to that off ered in Adiego (1993a).Note that ”evoro“kin’s drawing shows sbb? instead of sba in the fourthword. This latter, corrected reading was introduced by ”evoro“kin him-self after revising the inscription (“Corrections to Existing Copies”, ms.),and was already reflected in Adiego (1993a).

E.Th 14 (= D 4) ]q[. . .]btdeo

This graffito from the open court was absent from the corpus of pub-lished inscriptions. The presence of d is surprising, as in the new read-ing of E.Th 4 (see above). Schürr observes that a vertical p p appearsunder the graffito.

E.Th 15 (= D 15)(Very uncertain reading)

Also a new graffito. Schürr mentions two very diff erent alternative read-ings, suggested by Kayser and ”evoro“kin respectively (see ”evoro“kin1994:145).17

E.Th 16 (= A 1)∂saml-?-?-o (vacat ) dy “a

Three unclear signs found under the end of A 1.

17 The readings are, respectively, [1] u? ∞a pn-?-ek . . . i / plsiwbms and [2] ub∞l pn-?-ekleai/pksiwrm≤ .

E.Th 13 = 60 ” (= D 14)




102

E.Th 17 (= A 2)ku

E.Th 18 (= A 3)t n

Very uncertain reading.

E.Th 19 (= A 4)dbikrm

k presents an inverted form l, apparently l , but the identification isbased on the good onomastic correspondence of the reading dbikrm (seeChapter 11, s. v. ).

E.Th 20 (= A 5)orbá ˚ r i“

E.Th 21 (= A 6)

mmn∞al

An inverse reading la ∞mmn is also possible.

E.Th 22 (= A 7)mwk | te

E.Th 23 (= A 8)bebi

E.Th 24 (= A 9)kow[?-?]

The drawing made by ”evoro“kin available to me points to koweq .

E.Th 25 (= A 10)ktmno

Inverted k (l ). The interpretation as k, not l, is supported by thegood onomastic connection of ktmno (see Chapter 11, s. v. ).

E.Th 26 (= A 11)brsi yri≤



103

E.Th 27 (= A 12)pnw≤ol | mlqi≤

E.Th 28 (= A 13)bejeym | tebootK

bebi.nt ken

The strange ‘diamond-sign’ also appears in E.Si 3, preceding preciselythe same word bebint , but it is not present in the other examples of thisword (E.Th 30, E.AS 7), which makes it difficult to consider it an actualletter. The exact function therefore remains unexplained (see pp. 253–254).

E.Th 29 (= A 14) ]ke

Some uncertain signs follow.

E.Th 30 (= A 15)

bebint | psrkrte | mumn“tnse-?»ßwk˚n

E.Th 31 (= A 16)(Very uncertain reading)

“Impossible to describe” (Schürr). Allegedly a sign with the form sappears twice, and a sign % appears once, both of which are alien tothe standard inventory of Carian letters (for this latter sign see alsoE.Th 34 below).

E.Th 32 (= A 17)tqlow

A reading in the opposite direction is also possible (wolqt )

E.Th 33 (= A 18)

∞lbiks≤

The second letter could also be a k (”evoro“kin’s reading).

E.Th 34 (= A 19)sl∞maewm | urt | kwri≤ | prna∞non | dm-?-n | maãtnor | qanor | uro



104

The uncertain letter in the fifth word seems to be similar to %, but”evoro“kin’s drawing in fact displays a sort of % form. Schürr ques-tions whether it could merely be o o.

E.Th 35 (= A 20)l ÿ∞se | “i“≤ | mlan[-?]

”evoro“kin’s drawing and reading point to a final letter e e , somewhatdamaged. Schürr omits it in his reading.

E.Th 36 (= A 21)\ or ≤ The first sign, apparently a \, could be the rest of a letter.

E.Th 37 (= A 22)ktmnPerhaps an incomplete form of ktmno, cf. E.Th 25 above.

E.Th 38 (= A 23) ]bewmsmnwdiq tebwnqmw

For wdiq (and the suggested segmentation involved), see below E.Th 46.

E.Th 39 (= A 24)krws | ko“m≤

E.Th 40 (= B 1)

pnu≤ol

From this graffito on, I shall simply reproduce Schürr’s readings, as Ihave not seen the drawings of this part of the corpus. I only commenton the forms that off er some parallel with known words.

E.Th 41 (= B 2)tmonks

E.Th 42 (= B 3)rdudmm»≤

E.Th 43 (= B 4)p



105

E.Th 44 (= B 5)dquq | ewmlane | tebot | g kem≤

The first word shows a good onomastic identification, see Chapter 11,s. v. The second word contains the typically Theban sequence (m)lane.The third word is the same as in E.Th 28.

E.Th 45 (= B 6)krwß

Cf. the first word of E.Th 39.

E.Th 46 (= B 7)prpwri∞ kblow≤

Apparently a sequence of two names. The first seems to belong to thefamily of yriq/ydiq names, but spelled with w , not y/ ÿ, and with ∞ (asin Euromos idyri ∞≤ ). For the first spelling, cf. perhaps the sequence ºwdiq in E.Th 38, although I do not think we should rule out that v w

may in fact be a variant of W y, so that the name could be read * prpyri ∞.

E.Th 47 (= B 8)w. dbo≤kn ewál.å»e ˚[

E.Th 48 (= B 9)brsi

A well-known Carian name, see Chapter 11, s. v. and cf. above E.Th 26.

E.Th 49 (= B 10)bal ewlane | “rb˚[-]sal|

E.Th 50 (= C 1)pn-?

E.Th 51 (= C 2)

p

E.Th 52 (= C 3)plqodse | ewm-?-?-?-? | rqemw | k-?[



106

In plqodse , perhaps the name plqo should be identified, see Chapter 11,s. v.

E.Th 53 (= C 4)dr“≤iem

5. Luxor Temple (E.Lu)

Some Carian graffiti were recently identified in the Luxor Temple.18

They are very brief, damaged texts, and only one of the seven I con-sider to be genuinely Carian19 suggests a clear connection with the restof the Carian Corpus.

E.Lu 1 (= G 19)

18 They have been published in: ESS (1998). I am very grateful to Richard Jasnowand to Theo van den Hout for the information provided about this sub corpus.

19 The graffito G 12 is excluded here: it could also be Carian, but none of its signscan be clearly identified.

E.Lu 1 (ESS 1998)

→

ds-?

E.Lu 2 (= G 21)

E.Lu 2 (ESS 1998)



107

→

rsy

suso“?rquq [. . .?

The clearest graffito of the collection. The last word is—if the reading of the first letter here proposed is accepted—the well-known Carianname “ rquq .

E.Lu 3 (= G 22)

E.Lu 3 (ESS 1998)

Very uncertain reading: o or t followed by m and another m or ratheru. The order of reading is unclear.

E.Lu 4 (= G 23)

E.Lu 4 (ESS 1998)

→

?-?-[-]ms[-]ry-?-?

Very uncertain letters at the beginning and the end of the graffito. Theletter immediately after y could be j j , and the last letter a D d .



108

E.Lu 5 (= G 24)

E.Lu 5 (ESS 1998)

←

b?s?ui∞am | oã?

Van den Hout (pers. comm.) suggests reading the apparent interpunc-tion trace .as L l . The resulting sequence ml o would have a good par-allel in C.Ka 9 [. . .]ois ? ur ? ml o. The presence of % h, not a typical letterin the Carian alphabet of Egypt, is not certain: the letter could be alsor r . But note that % also appears in the nearby corpus of Thebes.

E.Lu 6 (= G 25)

E.Lu 6 (ESS 1998)

→

| urq

The reading is far from certain. The last sign could also be ≤ .



109

E.Lu 7 (= G 26)

E.Lu 7 (ESS 1998)

tksr or rather tasr?

Neither of the two possible readings off ers a good connection with otherCarian sequences.

Summing up, although these graffiti are practically unusable (withthe exception of E.Lu 2), the presence of Carian graffiti among thegraffiti of the Luxor temple is indisputable, as demonstrated by the

presence of genuine Carian letters such as I or f.

6. Murw àw (E.Mu)

In his monumental work on the inscriptions of Lower Nubia, Zbyn^k¥ába published, in collaboration with Fritz Hintze, a Carian graffitofound in the region of Murwàw, about two kilometres north of thetemple of Dendùr (¥ába 1974[79] nº 196 & fig. 323).

This inscription does not appear either in Masson’s indexes (Masson1978) or in the transcribed corpus of Meier-Brügger (Meier-Brügger1979b). Our only source is the unsatisfactory edition of ¥ába, whichincludes a fairly poor quality photograph, an inaccurate and unreliablefacsimile (reproduced here), a transliteration to the systems of Bork-Friedrich and ”evoro“kin, and some rather confusing notes. Only thereading of two words known elsewhere is certain.



110

E.Mu 1 (= ¥ába 1971, 196)

E.Mu 1 (¥ába 1971)

→

p?owk | wljat≤ | ÿnsmsossaaw?on sa?awon

Despite the drawing, the readings wljat ≤ and ÿnsmsos are practicallyassured because they are known elsewhere: wljat ≤ (not ‡wljot ≤ !) = E.Th7 wljat , and ÿnsmsos = E.AS 3 ÿnsmsos . The remaining words pose veryserious reading problems. As well as powk , an alternative reading sowk is possible, and even the last letter of this word is far from certain: itcould also be t t , which gives sowt/powt as further alternative readings.

The alleged letter / in the second line—unexpected in an Egyptianinscription—seems in fact to be ~ a . The two words of the secondline are generally taken as almost identical (already in ¥ába, see draw-ing), but they could be diff erent, coinciding only in their final part(-on ). In this sense, the letter under s in the second word might be notan a but in fact a b b, so that a reading sb-awon would be possible. Inconclusion, only s—on is completely certain for both words of the sec-ond line. Therefore, the reading adopted here is merely provisional.

7. Silsilis (E.Si)

There is not a modern edition of the Silsilis graffiti, it should be addedat this point that in all probability producing this new edition will neverbe possible: according to Masson (1969:32), a great number of thesegraffiti must have been destroyed.

The corpus consists of 11 generally very short and impenetrablegraffiti. Only a pair of words that correspond to the rest of the Carian

documentation can be clearly interpreted: psma ≤ k ≤ , the name Psammetichus,well documented in the Carian of Egypt (E.Si 2, E.Si 7), and be ? bint (E.Si 4), also attested in Thebes. I adopt the readings of Masson (1978),with the exception of E.Si 2 and E.Si 10. In the first case, I follow



111

Schürr’s proposal of reading the text as boustrophedon, and modifying theinterpretation of some signs from the drawings of Sayce and Legrain

(Schürr 2000:172). In the case of E.Si 11, Schürr has argued con-vincingly for a new reading, based on a photograph published by HansA. Winckler in 1939 (see below).

I provisionally reinstate the inscription Si 61 F (= Sayce 1906, n. 5),excluded from Masson’s corpus. This is also present in ”evoro“kin (1965)(69 ” ), see also Schürr (1996b:149).20

E.Si 1 = Si 39 F

20 For technical reasons, the source of my illustrations will be Friedrich (1932), whereSayce’s drawings are reproduced.

E.Si 1 (Friedrich 1932)

←

∞iqud | marariso[-. . .]

E.Si 2 = Si 53 F

Sayce I

Sayce II



112

→ [—]e∞ld | wa | psma≤k≤ |← uejresi | qan | kolt | kowrn[. . . ?

I off er the three existing copies of this graffito (two by Sayce, the thirdby Legrain 1905) and adopt Schürr’s proposed reading (Schürr 2000:172).In any case, as noted above, the only clearly recognizable form is thename psma ≤ k ≤ in the first line.

E.Si 3 = Si 54 F

LegrainE.Si 2 (Friedrich 1932)


←

irasa | n[-]eakrnanb

E.Si 4 = Si 55 F

Sayce

LegrainE.Si 4 (Friedrich 1932)



113

→

[ . . . ] t bebint | sqlumidun | sqla

For the diamond-sign preceding the same word, see above E.Th 28.The reading bebint (instead of ‡b ÿbint in Adiego 1993a) now seems prefer-able, given the clearer evidence of the Theban inscriptions.

E.Si 5 = Si 56 F


→

betkrqit[—. . .]

E.Si 6 = Si 57 F


→

b ÿta“ | sursiabk | dr[-. . .]qku

E.Si 7 = Si 58 F

→

psma≤k




114

E.Si 8 = Si 59 F


→

bij≤≤pe (. . . ?)

E.Si 9 = Si 60 F


→

[. . .]rbn“a[—. . .

E.Si 10 = Si 61 F


←

∞?mpi

E.Si 11 = Si 62 F

E.Si 11 (Friedrich 1932, modified)



115

←

dmo“bqs

Reading taken from Schürr. The photograph of the inscription pub-lished by Winckler in 1939 (which I was made aware of by Schürr)quite clearly substantiates the reading proposed by Schürr. It also servesto definitively rule out the alleged presence of a letter b in the Carianalphabet of Silsilis.21

8. Abu Simbel (E.AS)

Fortunately the current situation of the Abu Simbel corpus is verydiff erent to those mentioned above; we have at our disposal a recentand very careful edition of these graffiti, thanks to the eff orts of OlivierMasson (Masson 1979), who was also responsible for the publication,some years ago and in collaboration with André Bernand, of the cor-pus of Greek graffiti from the same location (Bernand-Masson 1957).

The Carian graffiti from Abu Simbel are strongly tied to the historyof research on Carian: it was Richard Lepsius who first edited and cor-

rectly identified as Carian these inscriptions found on the legs of thetwo colossi of Abu Simbel (Lepsius 1844, Abt. 6, Bl. 98 [Kar. 1–3],and Bl. 99 [Kar. 4–7]). This theory was confirmed by Sayce (1874,1887[92]) when texts in a similar alphabet were found in Caria andbordering areas. However, from an epigraphical point of view, Saycecould not contribute to an improved edition of the text, and he hadto be satisfied with Lepsius’ copies (see Masson 1979:35–36 for details).

Masson’s edition is based on the drawings and photographs madeby André Bernand and Abd el Latif Ahmed Aly in 1956. The draw-ings had already been published provisionally (Bernand-Aly 1959?), andreproduced in ”evoro“kin (1965). The corpus presented here introducessome corrections and additions made by Diether Schürr.

21 See Schürr (1996b) where, contrastingly, the presence of a letter / is argued.Later, Schürr argued in favour of a f, which seems more likely. This latter reading is the one that I adopt here. For the alleged initial letter “my” in the inscriptions, seethe convincing arguments in Schürr (1996b) for rejecting it.



116

E.AS 1 = AS 1

E.AS 1 (Masson 1979)

←

par≤olou[. . .]oe

Graffito discovered by Bernand.

E.AS 2 = AS 2


←

“abd?aikalAlso a new graffito found by Bernand, absent from Lepsius’ corpus.

Schürr’s reading is followed here.

E.AS 3 = AS 3 = Lepsius Kar 4


←

pisma“k | “arnw≤ | ÿnsmsos



117

E.AS 4 = AS 4 + Lepsius Kar 7

E.AS 4 (Masson 1979 + Lepsius 1844)

→

a∞akowr | emsg lpn | b[. . .] pisma[ “/≤k . . . ]

The first part of the graffiti (which appears in the drawing) was rec-ognized by Bernand, while Lepsius was only able to identify some illeg-ible characters (Bernand apud Masson 1979:39, n. 39). But as Schürrnotes ( per litteras ), Lepsius Kar. 7 (left out by Masson because it wasnot re-found by Bernand) is actually the continuation of E.AS 4, andthe beginning of the Egyptian name (in its Carian adaptation) pisma “ k / pisma ≤ k is easily readable in Lepsius’ copy.

The direction of the reading off ered for AS 4 in earlier works (forinstance Adiego 1993a) was incorrect.



←

pnyri≤ru | i ÿkr≤ | “a[--]i≤b?wn

Schürr suggests that the final letters may be part of another graffito.



119

←

nid≤kusas | me ÿqak | sn≤ | ≤t≤ | ≤uni≤ | k“mmsm[. . .]

[. . .]r≤wk[-]“[

Graffiti discovered by Bernand. Masson’s reading is followed, despitethe significant problems posed by the final part.

E.AS 9 = Kar 2

E.AS 9 (Lepsius 1844)

→

ÿsm [?

This graffito is omitted without any reference in Masson (1979). It hadalready been identified as a part of a Greek graffito in Bernand-Masson(1957:38; the graffito is Bernand-Masson nº 32), but Schürr ( per litteras )maintains that it is in fact Carian. In fact, no connection seems to existbetween the clear drawing wsm of Lepsius and the Greek graffiti asreproduced in Bernand-Masson loc. cit.

9. Buhen (E.Bu)

Also in the case of the graffiti from Buhen, we can refer to an excel-

lent, recent edition, made by Masson and published together with theSaqqâra corpus (Masson 1978).In 1895 Sayce published seven inscriptions copied in the Southern

Temple of Buhen (then mentioned as “Temple of Thothmes III atWadi Halfa (Sayce 1895). He had serious doubts about the Cariannature of two of these (numbers 6 and 7). They have not been redis-covered since and must remain outside our collection.

For his edition, Masson was able to re-read three of the five clearlyCarian graffiti (E.Bu 2, 3, 6), and he also added a new and importantgraffito found during the British campaign 1962–1963 (E.Bu 1). Forthe remaining Carian graffiti (4 and 5), Masson was obliged to workonly from Sayce’s copies.



120

Despite the quality of his edition, Masson was unable to find a sat-isfactory solution to the problem posed by the same word that appears

repeatedly in diff erent inscriptions from Buhen. Following the deci-pherment of Carian, we now know that this word is the name Psam-metichus in Carian, and that the many divergent readings adopted byMasson must be brought together. This correction aff ects E.Bu 1 ( psma ≤ k ≤ ,not ‡ psma ≤ u≤ ), E.Bu 4 ( psma ≤ k , not ‡ psma ≤ m ) and E.Bu 5 (identical cor-rection). Other discrepancies with Masson’s edition aff ect the word ar ® i “ (not ‡arli “ !) in E.Bu 1 and E.Bu 2 (”evoro“kin’s correction, see ”evoro“kin1984[86]:199), and i b rsi ≤ (not ‡iyrsi ≤ !) in E.Bu 4 (Schürr’s correction:Schürr 1991–1993). These corrections were already set out in Adiego(1993a).

E.Bu 1 = M 50

E.Bu 1 (Masson 1978)

←

[—]msal | ar-

[ ® ]i“ | psma≤-k≤ | urm≤ | an-kbu“ | trelkbou≤



121

E.Bu 2 = M 51


←

euml?bna-sal | ar®i“

pdtom≤urom≤ | an-kbu“

E.Bu 3 = M 52

←

[-]tmai≤[—]




122

E.Bu 4 (= M 53)

E.Bu 4

Masson’s copy of Sayce’s drawing

Sayce (1895)

←

psma≤kibrsi≤

This constitutes a good example of a biased interpretation of a textknown only from a single copy; the inscription was not found againwhen Masson was preparing his edition, so he had to use Sayce’s draw-ing. Masson interpreted the second letter of the second line as a W,

modifying Sayce’s copy in order to approximate the traces of the let-ter to a more standard form. However, as Schürr (1991–1993:167) hasconvincingly demonstrated, this letter is actually Ø b , and not W y,since we are dealing with the typical Carian name i b (a)rsi , b rsi , attestedby a number of examples. The other problematic letter is the last oneof the first line, for which an interpretation as k k is now beyonddoubt, as the whole word is the Carian form of the Egyptian namePsammetichus.

E.Bu 5 (= M 54)


←

psma≤k

As in E.Bu 4, the last sign is now interpreted as k, in order to obtainthe well-known Egyptian name psma ≤ k .



123

E.Bu 6 (= M 55)


←

eypsalpuor≤ | aor≤ursea∞k ∞i

This inscription is extremely difficult to read, so some uncertaintiesremain. In any case, the reading aor ≤ for the third word should be pre-

ferred to ‡a ≤ r ≤ , as found in previous works, given the good parallel inE.Me 1 and the sound onomastic identification (see Chapter 11, s. v.aor ≤ ).

10. Gebel Sheik Suleiman (E.SS)

E.SS 1 (= 72 F)

This graffito, which constitutes the southernmost document of Carianto be found in Africa, is known only from two diff erent copies: Sayce(Sayce 1910 = Friedrich 1932) and A. J. Arkell (reproduced in ”evoro“kin1965) respectively. As we are informed by Masson (1978: 35, n. 1, and98), this inscription, which never was photographed, now lies beneaththe waters of the Nile. The correction introduced in Adiego (1993a) ismaintained here: the second word of the second line is read “ÿin≤ , not‡“ÿ“ n≤ , because it is undoubtedly the same word that appears in E.Me38.



124

Friedrich (1932)

←

n≤n[-]s“|“aru≤ol pneit|“ÿin≤par ÿd

∞≤

11. Unknown Origin, Likely from Egypt (E.xx)

Under this title I include seven inscriptions whose Egyptian origin seemsclear, but for which no exact location can be determined.

E.xx 1 (= MY C)

”evoro“kin (1965)

E.xx 1 (Masson-Yoyotte 1956)



125

←

qorb | isor≤ ∞i | ≤ugl i≤

This stela is known as the “stèle de Grenoble”. The attribution of anorigin from Bubastis was discarded in Masson-Yoyotte (1956:7). In anycase, the alphabet is clearly of the same kind as that used in Memphisand Sais. Unproblematic reading.

Photograph: Masson-Yoyotte (1956: pl. IV). Drawings: Masson-Yoyotte (1956:8–9).

E.xx 2 (= MY I)

E.xx 2

←

wliat

This so-called “ichneumon of Berlin”, is actually a reliquary for amummified shrew with a Carian name followed by an Egyptian votiveformula ‘X (the god incarnated by the animal) may give life’.

Photograph: Masson-Yoyotte (1956: pl. 4). Drawing: Masson-Yoyotte (1956:37). Mydrawing of the shrew is based on the photograph.

E.xx 3 (= MY a)

E.xx 3



126

→

ionel ≤

Inscription on the mount of a ring.

Photograph: Masson-Yoyotte (1956: pl. VIII). Drawing: Masson-Yoyotte (1956:11). Mydrawing of the object is based on the photograph.

E.xx 4 (= MY b)

E.xx 4

→

pduba

Word engraved on a vase.Photograph: Masson-Yoyotte (1956: pl. VIII). Drawing: Masson-Yoyotte (1956:12). Mydrawing of the object is based on the photograph.

E.xx 5 (= MY c)

E.xx 5 (Friedrich 1932)



127

←

ow∞meb≤t

This is the “scarabaeum with a Carian inscription”, first edited bySayce (1887[92]). It was briefl y referred to in Masson-Yoyotte (1956)(MY c), but neither photography nor drawing or transcription wasattempted at the time, given that neither editor knew where the objectwas. The object was rediscovered a short time after the publication of Masson-Yoyotte (1956), and Masson dealt with it in an article (Masson1959b). In the absence of more information, Masson considered it tobe “pseudo-Carian”. In Masson (1978) it was omitted.

Contrastingly, ”evoro“kin included it in his book on Carian (”evoro“kin1965: 127; 6 ” ) and off ered a very plausible reading (ow ∞meb≤ t ). Thefinal sequence ºb≤ t is a particularly attractive suggestion, because it alsoappears in two graffiti from Thebes (E.Th 12 and E.Th 13). I havetherefore included alongside the other Carian texts.

Photograph: Masson (1959: pl. 2). Drawing: Masson (1959: pl. 3).

E.xx 6 (= 4 ” )

E.xx 6

←

“arnajs | sb taqbos

Inscription on the so-called “Leningrad Isis”, published by ”evoro“kin(”evoro“kin 1964b; ”evoro“kin 1965). As in the case of E.xx 2, a Carian



128

onomastic formula is completed by an Egyptian votive phrase, here Js.t dj ‘n˙ ‘Isis may give life’.

I adopt here a new suggested reading formulated by Diether Schürr.

Photographs: ”evoro“kin (1964b:pl. I–IV). Drawing: ”evoro“kin (1964b:58–59). Mydrawing of the inscription is based on the photograph.

E.xx 7 (= Lion)

E.xx 7 (Masson 1976)

←

ntros: prãidasor“anu mdane: uksi wrm≤

Inscription on a lion of Egyptian origin, edited by Masson (Masson1976). Some years ago, I suggested to Masson the possibility of read-ing ntros ntros , instead of noros ‡noros (sic in Masson 1976), asuggestion based both on the photograph and the comparable sequence

ntro ntro in C.xx 1. Yoyotte’s revision of the inscription made thiscorrection possible, which was already included in Adiego (1993a) andis now widely accepted.

Photograph: Masson (1976: pl. 1). Drawing: Masson (1976 pl. 2).

D. T C I C

The first and last specific edition of all the Carian inscriptions fromCaria available at the time was produced by Louis Robert and LouisDeroy, in two complementary articles: Robert (1950), which includes



129

photographs and epigraphical information, and Deroy (1955), a col-lection of drawings based on the photographs published by Robert.

The numbering of the inscriptions introduced by Robert and adoptedby Deroy became the standard (hence the use of “D” or, less frequently,“R.-D.” followed by a number to refer to the Carian inscriptions fromCaria). This ordering did not follow a coherent set of criteria.22

As for the inscriptions published after the Robert-Deroy edition, inMeier-Brügger (1983) a catalogue was established (although the inscrip-tions were not edited) and the numbering system of Robert-Deroy wascontinued, but an asterisk mark was added to the new inscriptionsinstead of the “D” and “R.-D” abbreviations.23 The numbering crite-rion employed was the order of publication of the new inscriptions. A source of further confusion was the fact that both Robert-Deroy’s andMeier-Brügger’s catalogues included Para-Carian inscriptions, here omit-ted from the true Carian corpus. The inscriptions from locations otherthan Caria were also listed, so this catalogue actually comprises all thenon-Egyptian-originated epigraphical documents of Carian.

In this study, a new classification system, parallel to that introducedfor Carian inscriptions from Egypt, will be introduced. The inscriptions

are classified according to the locations in which they were found.We are clearly in a better position to analyse the corpus from Cariathan the corpus of Egyptian texts: with the sole exception of C.Tr 2,at least photographic evidence is available for all the inscriptions.Moreover, some of the texts edited by Robert and Deroy have beenrevised thanks to the eff orts of Blümel, Gusmani, Frei and Marek,whereas the texts found after Deroy’s compilation have generally beenedited satisfactorily.

22 1–3 are already published. 4–5 are inscriptions neglected by scholars prior toRobert (in fact, “Para-Carian” inscriptions from Chalketor and Ancin, south of Alabanda).6–13 are inscriptions found by Robert himself. 14 is a text discovered by G. E. Bean,15 is a Carian (Kaunian) inscription from Lycia already known and previously pub-

lished (as 1–3), and 16, off ered in addendum, is the great inscription of Kaunos, alsodiscovered by Bean. In Deroy’s article, the confusion continues: D 17 are Para-Cariangraffiti from Labraunda, D 18 Carian coins, and D 19, the Greek-Carian bilingualinscription from Athens.

23 For the continuation of the numbering after Meier-Brügger (1983), see Meier-Brügger (1994:113), Frei-Marek (1997:6, n. 10), Frei-Marek (2000:85).



130

24 I am grateful to Dr. Georg Rehrenböck (Austrian Academy of Sciences, Vienna),for your extreme kindness in sending me a reproduction of Kubitschek’s drawing andnotes concerning this inscription.

1. Tralleis (C.Tr)

C.Tr 1 (= D 1)

Deroy (1955) Kubitschek’s copy conserved in Vienna

←

sdi amt[ pau≤art mon

This inscription has not been re-found since it was seen and copied atthe end of the 19th Century, but Deroy’s drawing, based both on aphotograph of the cast and a copy conserved in the notes of Kubitschekin the Austrian Academy of Vienna (which I reproduce here)24 is quitereliable. Moreover, the inscription does not pose any particular prob-

lems of reading. The third letter of the first line is definitely Y i .I adopt the reading suggested by Schürr (cf. Schürr 2001b:109,

n. 12), but I am unsure about the segmentation of the first line.

Photograph of a cast and drawing in Deroy (1955:307 and pl. I).

C.Tr 1



131

C.Tr 2 (= D 2)

Pappakonstatinou’s drawing(Meier-Brügger 1978) Kubitschek’s copy conserved in Vienna

C.Tr 2

→

an sidi a-rtmi pau≤parãaq?

This inscription, which has since disappeared, is known only throughdrawings.

I retain the reading adopted in Adiego (1993a), based on the imagepublished by Meier-Brügger (1978:81), a direct reproduction of thedrawing edited by Pappakonstantinou in 1895. The copy conserved inVienna among Kubitschek’s travel notes (see above C.Tr 1), broughtto my knowledge by Dr. Georg Rehrenböck, is concurrent with Pappa-konstantinou’s drawing, the only surprising diff erence being the omis-

sion (accidental?) of r in the third line ( pa ãaq ? instead of par ãaq ), seethe illustration.

The only problem of reading is the last letter of the inscription:Schürr (2001b:109, n. 12) argues that it is ≤ (“das letzte Zeichen wohl≤ ”), but Pappakonstantinou’s and Kubitschek’s drawings points ratherto a Q q .

Drawings: Deroy (1955:308), Meier-Brügger (1978:81).



132

2. Alabanda and Surroundings (C.Al)

C.Al 1 (= D 13)

(Deroy 1955)

→

sdi a[-]mob[

Found by Robert in Eski Çine, south of Alabanda. To my knowledge,the inscription has not been seen again. I adopt Schürr’s reading (Schürr2001b:109, n. 12).

Photograph: (1950: pl. II 1, VIII 2, XXI 2). Drawings: Robert (1950:17), Deroy (1955:319).

3. Euromos (C.Eu)

C.Eu 1 (= D 3)

Drawing conserved in Vienna

→

≤as: ktais idyri∞≤: mn[os? ]

This inscription was copied, photographed and edited for the first andlast time, by E. Hula and E. Szanto in 1894. As Meier-Brügger hasdemonstrated (Meier-Brügger 1978:78–79), the drawing made by the

first editors is considerably better than that published in Sayce (1905)and Deroy (1955). I reproduce here a copy of the original drawing byHula and Szanto conserved in Vienna, kindly sent to me by Georg Rehrenböck.



133

The main difficulty is posed by the alleged sequence WW yy. Some years ago, ”evoro“kin (in his unpublished document “Corrections to

Existing Copies”) proposed to correct it on the basis of the direct studyof the cast: according to him, the sequence was actually a Wr yr . Thiscorrection is eminently suitable from a linguistic point of view, since asequence WW yy is hardly acceptable; moreover, a reading Wr yr yieldsa good onomastic identification (see Chapter 11, s. v. idyri ∞≤ ). However,considered from the point of view of the local alphabet of Euromos,we would expect a form R —rather than a form r —for r , as can beseen in the other inscription from Euromos (C.Eu 2).

Thanks to the kindness of Dr. Rehrenböck, I have been able to con-trol directly the two (very similar) casts of the inscription conserved inVienna. Both Dr. Rehrenböck and I agree that, at first sight, an W canbe recognised in both casts. Without totally ruling out the possibilityof an original mistake made by the engraver, one can suppose thateither the intermediate trait or the lower trait is intrusive. It is truethat the intermediate trait seems weaker, favouring thus a reading R(which would be more consistent with the testimony of C.Eu 2), but itis frankly impossible to decide definitively between a reading r, or a

reading R.The last word, incomplete, must be without any doubt the Carian

word for ‘son’, mno-. The difficulty lies in establishing whether it wasinflected. In my proposal of integration, I tentatively suggest the com-pletion mn[ os ? ], based on the assumed concordance of this word withthe name ktais , very possibly a name with an -s ending.

Photograph: Robert (1950:Pl. VI 4). Drawings: Deroy (1955:309), Meier-Brügger(1978:80).

C.Eu 2 (= D 8)

C.Eu 2 (Blümel 1988)



134

→

omob ∞i: temazi

≤dun: ≤o≤niabkolarmon qyrbmudolomanon

This inscription was re-edited almost simultaneously, and with divergentreadings, by Blümel (1988) and Gusmani (1990). Wolfgang Blümel waskind enough to allow me to see a cast of the inscription, and I canconfirm the accuracy of his edition. I reproduce Blümel’s drawing here.

The text does not pose reading problems. The most remarkable inno-vation of the current transcription is the interpretation of the letter enot as a variant (Kaunian-like type) of L l , but rather as a local formof e e , as in Hyllarima and probably also in Mylasa (see below p. 225).

Photographs: Robert (1950), pl. IV. Drawings: Deroy (1955:316),Blümel (1988:262), Gusmani (1990:49).

4. Kindye (C.Kn)

C.Kn 1 (= D 6)

C.Kn 1 (Deroy 1955)

→

pare ÿs

Extremely fragmentary inscription found by Laumonier, which has beennot recovered.

Drawing: Deroy (1955:314). Photograph: Laumonier (1933:35).

135

5. Hyllarima (C.Hy)

C.Hy 1 (= D 7 + new fragment)

C.Hy 1 (col a: Adiego-Debord-Varinlio< lu 2005; col. b: Laumonier 1934)

→

(a) “asqariod dymdamuot armotrqdosq

brsi ari“≤ brsi≤mane: u≤ol ≤rtim u≤ol ≤ pur?i≤

u≤bzol tñu≤ brsi≤pau mane≤ ybr-s≤

(b) kdu≤opizipususotmol“ msot ylarmit

(Greek text—only the oldest inscriptions that appear also in the illustration):(a) flere›ew ye«n pãntvn:ÑErm¤aw Fan°v ÑErm¤adow

(b) flereÁw ye«n pãntvn:ÑUssvllow ÉArrissiow

The stone containing the Col. b, was discovered by Laumonier in 1933.In 2004, a new bilingual inscription was found in Hyllarima, which



136

turned out to be a fragment of the same stone, the two pieces fitting together. This sensational discovery provides us with almost the com-

plete marble stela (only the lower part is missing), and the result is acomplex mixture of inscriptions from diff erent periods, which makesthe interpretation of the Carian text difficult, both internally and inconnection with the Greek texts.

The first problem is to establish whether the two first Carian linesof each column must be read as two complete lines (the first line of column (a) being followed immediately by the first line of column (b),and likewise for the two second lines), or if one must begin reading from the first two lines of (a) and then the first two lines of (b), orwhether column (a) and column (b) in fact represent two independentinscriptions.

Also problematic, and related to this discussion, is the connectionbetween Carian and Greek texts. It is clear that the majority of Greektexts seem to have been engraved long after the Carian lines,25 but thefirst four lines of column (b) could be contemporary with the Cariantexts. Whether this contemporaneity implies a connection between thetwo texts or not is impossible to decide.

The problem is compounded somewhat by the fact that in the Cariantext of (a), the last three lines, clearly separated from the preceding ones, show marked diff erences in the shape of the letters, and thusseem to have been written long after the preceding lines. A similar sit-uation can be observed in the four Greek lines mentioned, the last twoshowing divergent traits to the first. This opens the way for a numberof diff erent hypotheses about the order in which the sections of theCarian inscription and the sections of the oldest Greek inscription wereengraved.

Independent from the problems of interpretation, which will be dealtwith in pp. 305–308, the inscription of Hyllarima is an exceptionaldocument: it is the best-preserved long Carian inscription, the text being complete and with only one difficulty of reading: the antepenultimateletter of the fifth line of column (a), apparently a R r .

25 The Greek texts are the following: in col. (a), immediately after the Carian inscrip-tion, a list of Apollo’s priests at the time of the joint rulership of Antiochos and his

son, dated in 263–262. In col. (b), after the Carian text, a list of priests of all the godsfollowed by a sale of the priesthood of all the gods, and a land renting document. Inthe lateral side of col. (a), another sale of priesthood of the same date as that of col.(b), the name of the divinities implied not being readable. In the lateral side of col.(b), other land renting documents.

137

Col. b, photographs: Laumonier (1934:347), Robert (1950:pl. V), drawings:Laumonier (1934:346), Deroy (1955:315). Edition of the new frragment in

Adiego-Debord-Varinlio< lu (2005).

6. Mylasa (C.My)

C.My 1

C.My 1

→

idrayridsemdbq mol“ ty∞[ tsial tusol≤: moi m[-]sao[ banol paruos≤: p?au paryri∞≤qzali obrbi≤: tsial obrbi≤banol yrqso≤: paryri∞ psoir≤[-]bdo pnu≤o≤: myze trdy≤“arkbiom qzali≤: ≤umo kbdmu≤skdubrotoz≤: pau ∞toi≤[-]qo idyri∞≤: ksbo idu≤ol≤[-]obiokli≤: ∞toi yrqso≤

Inscription found in 2004 in Kırca< iz, near Mylasa. As with otherCarian inscriptions, this text poses serious reading problems regarding the letters z, Q and o, which in several cases can barely be distin-guished. It also seems that the engraver was not overly experienced in



138

his task—as the editors of the inscription have pointed out (Blümel-Kızıl 2004: 138)—and that some difficult readings could in fact be mere

errors.The reading adopted here follows in general terms that given in

Adiego (2005), but with the inclusion of improvements suggested morerecently by Blümel (Blümel 2005). I am aware of some reading prob-lems that still exist. In any case, this inscription is a very importantdocument both for Carian onomastics and for our knowledge of theCarian alphabetic variety of Mylasa.

Photograph: Blümel-Kızıl (2004). The drawing presented here has been made on thephotograph.

7. Sancutary of Sinuri near Mylasa (C.Si)

C.Si 1 (= D 9)

C.Si 1 (Deroy 1955)

→

adymd“: yri∞ñ: t[-]rsi: [. . .?]tbe≤(vacat) yri∞ñ: binq: sñaidl o



139

Found by Pierre Devambez in 1935, it has been not recovered since.The drawing reproduced here is from Deroy (1955). Several reading

problems therefore remain unresolved. I adopt the reading that can bededuced from Deroy’s drawing.

Photographs: Robert (1950: pl. 1). Drawing: Deroy (1955:316).

C.Si 2 (= D 10)

C.Si 2 (Deroy 1955)

→

(a) [—]ryin ∞tmño≤: sb ada ∞tmño≤eri: pisñoi mda: pñmnn≤ñ: pda-∞m≤uñ ∞i “aoyr∞ri mt∞elãñmail o mda lrHñ: stspñ vacat sm“s[—5—] sb añmsñsi mdasm[—7—]a∞e[ ∞[—8—]tuñdñ[

ñe-?-[

(b) pim[. . .]Ha?[ . . . ]



140

(Greek text:)Robert (1945), n. 75

[ ÉIdri°vw ÑEkatÒmnv][ ka‹ ÖA]daw ÑEkatÒmnv ka[[ ?S]uennitvn fler°iow [[ ]now Ponmoonnou [[ét°l]eian pãntvn ep[

[ÉId]rieÁw ÑEkatÒ[mnv ka‹ ÖAda]ÑEkatÒmnv ¶[dvkan?

ét°leia[nsunge[ntis o[oug.[

Robert (1945), n. 74 (+ C.Si 2).] ßkastow

]areihw?k]atÉ §niautÚn

?§]p‹ Nhsaiou]w

to›]w Suenn¤oiw

Bilingual inscription found by Louis Robert. The Carian part was editedby Robert (1950) and Deroy (1955). The Greek part can be found inRobert (1945). Regrettably, neither Blümel’s eff orts to locate the inscrip-tion years later nor the steps given by Schürr in order to obtain Robert’scast have been successful.

In any case, the reading of the three first lines is quite certain,with the important correction made by Schürr, consisting in reading d

d instead of F, in the name ada . Things are not so clear from the fourthline onwards: H in lr Hñ could be U (therefore lruñ ), and the t t in stspñ could

be o o (sospñ ), etc. Also problematic is the reading of the entire fifth line.Schürr’s suggestion of linking this fragment of stone with another

from Sinuri that contains two decrees in Greek by Idrieus and Ada(Robert 1945, nº 75) seems to me a very attractive theory, and it is

adopted here (see Schürr 1992:136–138; cf. Adiego 2000:134–135 fordetails; I reproduce the text Greek edited in this latter article).26

26 This edition was revised by W. Blümel.



141

This inscription can be now dated more accurately: the mention of the joint ruling of Idrieus and Ada situates it between 351/350 and

344/343 (cf. Hornblower 1982:45).

Photographs: Robert (1950:pl. II and III). Drawing: Deroy (1955:317).

8. Kildara (C.Ki)

C.Ki 1 (= D 11)

27 The Greek inscription has also been published by W. Blümel in Die Inschriften von Mylasa. II, Inschriften aus der Umgebung der Stadt (= I. K. 35), Bonn 1988.

C.Ki 1 (Blümel-Adiego 1993)

→

[ . . . . . . . ( . ) ] z ol ba∞a[..(.)] kil [ [. . .]uda[. . .] trqdimr qrds tazomd[ kil arad[-]ybzsdmHnmkda[-]aHuq[ iasoum

Greek text:¶doje KildareËsin, §kklhs¤hw genom°nhw: ÑUss[vllvi?]

Samvou eÈerg°thi genom°nvi Kildar°vn ét°[leian]doËnai ka‹ proedr¤an ka‹ §sagvgØn ka‹ §jagv[gØn]ka‹ §n efirÆnhi ésule¤ ka‹ ésponde‹ ka‹ aÈt«ika‹ §kgÒnoiw: ka‹ Kildar°aw e‰nai ín y°[lvsin?]

Discovered by Louis Robert in 1934, this inscription is followed by aGreek decree honouring Uss[ vllow ], son of Samvow. The inscriptionwas seen again by Blümel, and a new edition of the Carian text wassubsequently published in Kadmos (Blümel-Adiego 1993).27 The present



142

drawing and reading synthesize the readings of Deroy and Blümel-Adiego, both conserving the letters that were impossible to identify in

the revision, and incorporating the new readings of some letters.I still see the exact reading of the last word as problematic; the read-

ing m of the last letter implies the presence of a M oriented to theleft, contrary to all the other instances of the letter in the inscription(m ), and also the rightward direction of the writing. The reading of theremaining letters is also uncertain. I believe therefore that a reading *ßAmous “ amous , instead of ÎAsouM iasoum, cannot be altogetherdismissed. Such a reading would be compelling, insofar as it wouldallow us to identify in the Carian text the name of the father of theindividual honoured in the Greek section, and consequently to estab-lish a connection between the texts. But for want of sufficient evidence,I maintain the old reading, iasoum.

Photographs: Robert (1950:pl. VI 2, VIII 1, IX 2, X), Blümel-Adiego (1993:pl. 1).Drawings: Deroy (1955:318), Blümel-Adiego (1993:89).

9. Stratonikeia (C.St)

C.St 1 (= D 12)

C.St 1

→

]sel “ a[—]a[———]om≤ ]som[n? ]e brsi≤ ula[——]ol ]latmne≤ ≤ ysñal [



143

] ari“ maql y≤[ ]sel “ piks[

]sel “ p[

This inscription, found by Robert in 1946, has not been revised sincethen. The present reading and modified drawing are based on my owninterpretation of the photograph published by Robert (1950). The mostremarkable innovation appears in the second line. I propose that a pos-sible personal name som[ n ]e (see Chapter 11, s. v. somne/somne ≤ ) can berecognized, followed by the well-known name b rsi ≤ , in genitive: the let-ter R r seems to be certain (and not W y, as formerly read), and thesign preceding it (b ) seems to be the local form of $ 4 b , althoughneither Robert’s photograph nor Deroy’s drawing provides a clear imageof what the exact form of the letter was. Also new is the reading Q

q , not o o, of the seventh sign of the fourth line.For the first line, I adopt the reading of Schürr (2001b:106).

Photograph: Robert (1950:pl. VI). Drawing: Deroy (1955:319). The drawing off eredhere has been made on the photograph.

C.St 2 (= 36*)

C.St 2 (ahin 1980 modiWed)

→

u≤ol ≤ uodrou u[ mute≤ ymezus[ ∞diye≤ uodryia[ uliade pidaru[



144

mañ“qaraH≤rl -?-[ dar“qemorms[

Hda“qedormñs[

Inscription edited by ahin (1980).I adopt Schürr’s suggestion that l. 7, sign 5 is Q q , not o o. In fact,

some rounded signs of the inscriptions are difficult to distinguish cor-rectly (they could be o, z, or Q ).

Photograph: ahin (1980:pl. V 1). Drawing: ahin (1980:206). The drawing presentedhere is a modified version of ahin’s, in order to introduce the new readings of some

letters.

10. Halikarnassos (C.Ha)

C.Ha 1 (= 33*)

C.Ha 1

←

smd ÿbrs | psnl o | ml orkn t ÿn | snn

Inscription on a bronze phiale (6th B.C.) published by Jucker and Meier-Brügger ( Jucker-Meier-Brügger 1978). The attribution to Halikarnassosis based on the information given by the dealer to Jucker (“Karien,Bereich von Bodrum”, Jucker-Meier-Brügger 1978:104; cf. also 109:“Der Händler, der als Fundort die Gegend von Halikarnass (. . .)nannte . .”). It is true that a phiale is an easily movable object, so itcould have been made elsewhere. For this reason, the possibility thatthe alphabetic variant used here may represent that of Halikarnassos

must be viewed with caution.The text does not pose reading problems.

Photograph: Jucker-Meier-Brügger (1978:pl. 2). Drawing: Jucker-Meier-Brügger (1978:109).My drawing is based on the photograph.



145

11. Didyma (Ionia, near Milet) (C.Di)

C.Di 1 (= 21*)

C.Di 1

←

]ub“ÿ

Inscription on a broken fragment, datable to the 6th century B.C.Although this graffito is located not in Caria, but in Ionia, I includeit in the Carian section of the corpus given the clear proximity of Didyma (and Milet) to the Carian country.

This is a very difficult text: not even the direction of reading is clear.It is ascribed to the Carian corpus based on the clear presence of theletter w ÿ. The overall reading is far from certain (see remarks in Adiego1993a:80). The reading proposed here (the same as in Adiego 1993a)is merely hypothetical. Steinherr’s reading, adapted to the present deci-pherment system, would be ‡ul “ t , whilst Innocente (1994:106) prefersto read ‡∞a “ÿ, in reverse direction

Photograph: Naumann-Tuchelt (1963/64: pl. 25). Drawing: Adiego (1993a:324). Seealso Tuchelt (1970), which includes a reading and some notes made by Steinherr. Mynew drawing is based on the photograph.

12. Iasos (C.Ia)

Italian excavations in the Carian city of Iasos have brought severalCarian inscriptions to light. Their value, with the exception of C.Ia 3,is very limited: the texts are extremely short and fragmentary, and noparallels with the rest of the Carian corpus can be traced.



146

C.Ia 1 (= 20* a)

(Pugliese Carratelli 1985[86])

→

]lalimtaoa | [ om

Inscription on the neck of a vase. I have not seen any photography,and my only sources of information are the drawings and commentsin Levi-Pugliese Carratelli (1961–62 [1963]:632) and Pugliese Carratelli(1985[86]:151).

Drawings: Levi-Pugliese Carratelli (1961–62 [1963]:632), Pugliese Carratelli (1985[86]:151).

C.Ia 2 (= 20* b)


←

]ue∞l | ∞ob[

Although a reverse reading is possible, given that all the letters presentin the inscription are symmetrical, the theory is practically dismissed,insofar as it would situate L l as the initial letter of a word, which is

highly improbable.

Photograph: Levi-Pugliese Carratelli (1961–62 [1963]:632); Drawings: Levi-PuglieseCarratelli (1961–62 [1963]:632), Pugliese Carratelli (1985[86]: 151).



147

C.Ia 3 (= 38* a)

C.Ia 3

→

?] are“ | “anne ml ne | siyklo≤ | “ann | trqul e | ∞l mud [?

Inscription on a cratera. Published for the first time by Pugliese Carratelli(1985[86]:150–151). Gusmani (1988:145–149) off ered a new reading

that challenged Pugliese Carratelli regarding the identification of thetwo ~ in the second and fourth sequences: Gusmani proposed thatthey be read in both cases as /. This new reading would subsequentlybe discredited, because this letter is typical only in the Kaunian alpha-bet, where its existence is inseparable from the fact that in this alpha-bet there is no t t , and t is represented by T, so that a diacriticallydiff erentiated / (and variants) is used to represented “ , instead ofthe f F of other alphabets. In the present inscription from Iasos, botht and f do appear, which leaves the need for a letter such as /

unexplained.Careful observation of the photograph provided by Pugliese Carratelli

shows that, at least in its second appearance, the letter in question



148

clearly has a form of ~ very similar to the example in the first word(the only diff erence being the somewhat rounded upper trace). It can

therefore extremely unlikely to be a letter comparable to / and vari-ants. Therefore, Pugliese Carratelli’s reading is preferable.

Photograph: Pugliese Carratelli (1985[86]:pl. I). Drawings: Pugliese Carratelli (1985[86]:150), Gusmani (1988:146). My drawing has been made directly on the photographypublished by Pugliese Carratelli.

C.Ia 4 (= 38* b)


→

n[. . .]pr[. . .]is[. . .]

Very fragmentary inscription published by Pugliese Carratelli (1985[86]).The few remaining letters are easily readable.

Photograph and drawing: Pugliese Carratelli (1985[86]:152).

C.Ia 5 (= 47*)

C.Ia 5 (Berti-Innocente 1998)

→

baqg k[. . .]



149

C.Ia 6 (= 48*)


→

[ . . . ]b?e≤

C.Ia 7 (= Berti-Innocente 2005)


←

?]y?n“

C.Ia 5–7 are three very recently discovered graffiti on vases, edited byFede Berti and Lucia Innocente (Berti-Innocente 1998 for C.Ia 5, 6;Berti-Innocente 2005 for C.Ia 7). Note the surprising letter v in C.Ia7, apparently a variant of Mylasa W, Sinuri-Kildara V (cf. Berti-Innocente2005:21).



150

13. Keramos (C.Ke)

C.Ke 1 (= *39a)

C.Ke 1 (Varinlio[ lu 1986)

→

uso-t

C.Ke 2 (= *39b)

→

uso-t

This pair of inscriptions with the same text, first published in Varinlio[ lu(1986), were briefl y the focus of attention when Ray ingeniously sug-gested that the sole word they contained could be identified with the

well known Carian name Ussollos, u≤ ol (Ray 1988). This identificationwas based on an assumption that has since turned out to be false: thatthe letter c could be equivalent here and in Hyllarima to the letter L

l . It is now clear that c does not represent l in Hyllarima, wherethere is a particular letter for this sound (L ), and the supposedly equiv-alent sequence ºusoc in Hyllarima has nothing to do with the Carianname mentioned (even the segmentation is far from certain!). As alreadymentioned (p. 23), I even have a number of doubts concerning thetrue Carian nature of these documents, which contain an initial sign

| that makes no sense as a letter in Carian. The disposition of the let-ters is also a little strange, and their inclusion in the Carian corpusmust be accepted not without certain reservations.

Drawing: Varinlio[ lu (1986).



151

14. Kaunos (C.Ka)

C.Ka 1 (= D 14)

C.Ka 1 (Deroy 1955)

→

sñis: sdisa-s: psu≤ol ≤mal≤: mno≤

This funerary inscription, discovered by G. E. Bean and published byRobert, does not pose any reading problems.

Photographs: Robert (1950: pl. VI 3 and IX, 1); Masson (1973[75]: pl. I, 1) Drawing:Deroy (1955:320).

C.Ka 2 (= D 16)

C.Ka 2



152

→

[ui? ]oml ã qrds g rdso[-]i[

[-]r sb a∞mnnartnyr obsmns[ [-]∞arl anoã sb z“ariosã i∞[ [-]nudrma ∞ yrpai sarni“ sb u[ [-]aH punot2 otr“ bi sb a∞tmsk[m[-]d bi 1aitk ouor gdb“l aã1_i[-][-] sarni“ sb 1orsol“ sb uHbit[-]bi qrdsol “ ait 1mali H∞it[-]intnor ∞ yrapai≤ umot2 oba[-]

diurt obsmsmñ1ñ ouor mt1_ yr

[—]abrun∞ur[-]“ yn“H ynn sb vacat [——————]tbsms _1mali [ [——————]maH sb an[ [.............................]ba vacat

This inscription, the longest Carian text known to date, has recentlybeen revised by Marek and Frei, and this new reading is followed here.The segmentations are purely hypothetical, mostly based on the iden-

tification of some clear words (for example the conjunction sb ) andendings (-“ , -≤ ).

Photographs: Robert (1950: pl. XXIX, XXX, XXVIII, 2); Masson (1973[75]: pl. II,2); Frei-Marek (2000:86). Drawings: Steinherr (1950[51]:331), Deroy (1955:321), Masson(1973[75]:125). For an evaluation of these drawings, see Meier-Brügger (1978). I havemade the drawing on the basis of the excellent photograph published by Frei-Marek.

C.Ka 3 (= 28*)

C.Ka 3 (Roos 1972 modified)



153

→

“oru≤

ann ibrs≤

This funerary inscription, first published in Roos (1972), has been revisedby Schürr (see Schürr 1996c, Adiego 1996), and the new reading hashighlighted some important corrections: the first letter is not—as Roos’drawing indicated—an unexpected form of a letter pi , but in fact avariant of the typical Kaunian letter /. This new reading is nowconfirmed by the photograph published by Frei-Marek (2000). Schürr’srevision also allows us to identify the well-known Carian stem i b rs-.

Photograph: Frei-Marek (2000:126). Regrettably, this photograph shows only part of the inscription. Drawings: Roos (1972:93, and pl. 40), Schürr (1996c:158). My draw-ing is based on that of Roos, but implements the new readings made by Schürr andthe information available from Marek’s photography.

C.Ka 4 (= 30*)

C.Ka 4

→

[. . .]u≤ou≤ ibrsdr[-][. . .]a yoml n r1_i[. . .]dar1_ idym“

This fragment, published by Masson (1973[75]), has been re-read andre-published in Frei-Marek (2000). The new revision has allowed some



154

corrections of former readings and, perhaps more importantly, has leadto the rejection of Masson’s suggestion (supported by Peter Hermann)

that this was a further fragment of the long stela C.Ka 2: the thick-ness of C.Ka 4 is more than 3 cm greater than that of C.Ka 2 (Frei-Marek 2000:99).

Photographs: Masson (1973[75]: pl. I, 2 and II, 1); Frei-Marek (2000:98). Drawings:Masson (1973[75]:125), Frei-Marek (2000:97). My drawing is based on the photographpublished by Frei and Marek.

C.Ka 5 (= 44*)

C.Ka 5 (Deroy 1955)



155

→

kbidn uioml n i[—]

inis drual nik[—]lan lysiklas[-? ]otonosn sb lys[ikl]an lysikratas[-? ]otonosn sarni[ “ ]mdot2 un sb undo[—]tl “ kbdyn“ sb b2o[—]ol“ otr“ sb a∞t[ms]— kmt absims sb [—] yt2 oru sb a∞t[—]bu∞ y[——]i[——]i[—]≤ un moa[-]l boror[—]Hl ∞sasot2 orttab sb ort[-] sb Hor-ouo bi mslmnliapurmoruos mnosaitusi

Greek text:

¶doje Kaun[ ¤ ]oiw §p‹ dhmio[ u ]-rgoË ÑIpposy°nouw: Nikok-l°a Lusikl°ouw ÉAyhna›o[ n ]ka‹ Lusikl°a Lusikrãt[ ouw ][ ÉA ]yhna›on proj°nouw e[ ‰nai k-]a‹ eÈerg°taw Kaun¤v[ n auto-]Êw ka‹ §kgÒnouw ka‹ [——]

n auto›w §[ . . .

This is the now well-known bilingual inscription honouring two Athenians.The reading I adopt here is that published in Frei-Marek (1998), whichimproves the editio princeps (Frei-Marek 1997) and adds a new frag-ment of the stela. Concerning the dating, Frei-Marek (1997) proposesconnecting the proxeny decree contained in the stela with events that

happened in 322 B.C. However, Descat (1998) has argued in favourof a slightly more recent dating: 314 B.C. From a linguistic point of view this discrepancy is not particularly relevant. In both cases, the useof Carian writing at the very end of 4th century, in a post-Hekatomnidera, seems meaningful enough.

Edited by Foxit ReaderCopyright(C) by Foxit Software Company,2005-2007For Evaluation Only.



156

The segmentation of the words is based in part on my own analy-sis of the text, and must be taken as provisional.

Photographs: Frei-Marek (1997), (1998). The drawing here is mine.

C.Ka 6 (= 45*; Schmaltz 1998)

C.Ka 6

→

or

Photograph: Schmaltz (1998: Abb. 4). My drawing is based on the photograph.

C.Ka 7 (= 46*; Schmaltz 1998)

]no≤?(or better : ]noñ?)

C.Ka 6 and 7 are two very brief graffiti found on vases and publishedby Schmaltz (1998). No photograph is available for the second of thesein Schmaltz, so we must make do with the reading he off ers, leaving a doubt about the last letter (z or ñ? Schmaltz 1998:209).

C.Ka 8 (= 49* Frei-Marek 2000:116–119)

C.Ka 8

→

potko≤l ≤? aba?d?

ya



157

Inscription on two blocks from to the wall of the so-called Demeter’stemple, published for the first time in Frei-Marek (2000:116–119). The

reading off ered here diff ers slightly from that of the first editors: insteadof potko≤ l o, I prefer to read—with reservations— potko≤ l ≤ ? with a final Z

≤ , not o o: the diff erentiating trace is quite clear, and I do not believe,contra Frei-Marek, that the diff erent form of the vertical trace regard-ing the other letter ≤ can is an obstacle to this reading. Moreover, Igive (with some doubts) the reading of the last two signs of the firstline. Both are clearly visible in the photograph, and even the two edi-tors suggest these readings in their comment on the single signs of theinscriptions, although they decide against including them in their edi-tion of the text.

Photograph: Frei-Marek (2000:116). My drawing is based directly on that photograph.

C.Ka 9 (= 50* Frei Marek 2000:120–125)

C.Ka 9

→

[. . .]ois?ur?mlo

Inscription on a broken fragment, published in Frei-Marek (2000:120–125).This text poses serious problems of reading, since if the we adopt theinterpretation of signs proposed by Frei-Marek, it is not easy to estab-lish the direction in which the letters must be read: Frei-Marek sug-gests four possibilities. I believe however that Schürr (2001a) off ers thecorrect solution to the problem in reading i, not Z, for sign 2, and

s, not m, for sign 3. If the s reading is not absolutely certain (theapparent trace that would allow to read it as s could be an intrusivemark), in the case of i, Schürr’s proposal must be accepted, and thislatter reading would mean that the orientation of the text no longer



158

poses a problem: the fragment must be oriented as it appears in mydrawing, and it must be read from left to right (the photograph in

Frei-Marek 2000 must therefore be inverted; the photography is cor-rectly reproduced in Schürr 2001a:63). Some doubts remain, however,about the sign read as r r by Schürr.

Photograph: Frei-Marek (200:120). My drawing has been obtained directly from thisphotograph.

15. Krya (C.Kr)

C.Kr 1 (= D 15)

C-Kr 1

→

qot2omu sdisa-s? n≤ “odubr≤ or rather : mn≤ “odubr≤?sb mno≤ knor

noril?

amsor rather : norimams?

The single inscription from Krya (the modern Ta yaka) in the gulf of Telmessos, whose alphabet is clearly similar to the Kaunian variety, isnot free from reading problems: there are several drawings of the inscrip-tion, yet none of these coincides exactly each the others. Sayce(1887[1892]: pl. III) reproduced two divergent drawings, by von Hammer-Purgstall and by Forbes and Hoskyns. A further drawing was publishedby Cecil Smith in 1888. Kalinka’s edition in the Tituli Linguae Lyciae

off ered a new drawing and reading, made in collaboration with Heberdey(Kalinka 1901:93). Deroy also included a new drawing based on twophotographs sent by G. E. Bean (Deroy 1955). Finally, Gusmani (1990)



159

produced a revision of the inscription, following the autopsy made byRoos. Gusmani’s reading was used in Adiego (1993a), although the

drawing included there was a reproduction of Deroy’s copy (Gusmanidid not publish any drawing in his paper).

From these various readings and the photograph published by Deroy,and if we disregard several old reading errors that have since beensuperseded, it is clear that controversy still exists regarding two letters:the first letter of the second line, and the fifth letter of the last line.In the first case, Kalinka’s and Deroy’s readings, which both proposea letter M m, have been challenged by Gusmani’s suggestion that thisletter could in fact be s s . As for the second problematic letter,Gusmani, following Roos, proposes M m instead of Deroy’s l l (Hammer-Purgstall, Smith, and Kalinka’s readings B can be ruled out, as thiskind of letter is alien to the Kaunian alphabetic variety, to which thisinscription belongs).

The drawing I off er here is based directly on the photograph of thecalque published by Deroy (1950). From this photograph it is impos-sible to tell if the first letter of the second line is M, as it would seemto be, or actually a broken s, as Gusmani suggests. This latter read-

ing has the advantage that it creates the sequence s d is as , parallel tos d is as in C.Ka 1 (see p. 291 for an interpretation of this sequence).For the other problematic letter, the drawing points to l, but the

horizontal line could be an intrusive one. The most prudent solutionis to off er a reading that leaves open the alternative readings for bothletters.

Photograph: Deroy (1955:pl. II). Drawing: Deroy (1955:320).

16. Inscriptions of Unknown Origin, Presumably from Caria (C.xx)

I group together under this title diverse inscriptions whose commoncharacteristics are their unknown origins and the fact that the objecton which they are incised points to a Greek-Carian background andnot, as in the case of the “pharaonic objects”, to an Egyptian one.



160

C.xx 1 (= 34*)

←

“rquq | qtblem≤ | ÿbt | snn | orkn | ntro | pjdl

Text on a bronze phiale published by Gusmani (1978). Some doubtshave arisen about the letter À, which Gusmani interpreted as a , butwhich is in fact l (as shown by the clear onomastic identification, qtblem≤

= Kutbelhmiw ). Also concerning the last word, a reading pjdl must bepreferred over a : from Gusmani’s photographs, it is beyond doubt thatthe final letter is the same as the third letter of qtblem≤ .

Photographs: Gusmani (1978: Pl. I, II). Drawing: Gusmani (1978:69). My two draw-ings showing the part of the phiale with the inscription have been made on the photo-graphs of Gusmani.

C.xx 1



161

C.xx 2 (=35*)

C.xx 2

→

ÿ≤biks not: alosd ∞arnosd: jzpe mdane

This inscription, on a bronze dinos, was published by Gusmani togetherwith the previous text (Gusmani 1978). There are no reading prob-lems. The segmentation of the first sequence as ÿ≤ biks not is based onmy own analysis of the text (see below p. 284).

Photographs: Gusmani (1978: pl. III and IV). Drawing: Gusmani (1978:71). My draw-ing has been made by combining the two photographs of Gusmani (1978).



162

C.xx 3 (= 40*)

C.xx 3

←

akymydu ÿeryly[ vacat ]d

An inscription on a cult object conserved in Geneva, edited by Meier-Brügger (1994:112–113; photograph and transcription). I adopt Schürr’sproposed reading (in Schürr 2001c).

Photograph: Meier-Brügger (1994:112). Drawing: Schürr 2001c:118). My drawing isbased on the photograph published by Meier-Brügger.

Meier-Brügger (1994:113) suggests that C.xx1, C.xx 2, C.xx 3 and alsoC.Ha 1, all of which have surfaced in the European antiques trade inrecent years, come from the same Carian location, perhaps a sanctu-ary of the god * Natr- (or rather ntro- = Apollo according to the Lyciantrilingual, see Chapter 11, s. v. ntro for further details), from which they

may have been stolen.

C.xx 4 (= 41*)

C.xx 4 (Zalhaas-Neumann 1994)



163

→

kdu≤ol“

C. xx 5 (= 41*)

C.xx 5 (Zalhaas-Neumann 1994)

→

kdu≤ol“

Regularized form of the letters of both inscriptions:

(Zalhaas-Neumann 1994)

C.xx 4 and C.xx 5 are two identical inscriptions on two bracelets pub-

lished by Zalhaas-Neumann (1994). Although the origin of both objects,conserved in Munich’s Prähistorische Staatssamlung, is unknown, theirinscriptions are not only clearly Carian, but even display manifest con-nections with the Kaunian alphabetic variant: note the form Z of z

≤ and particularly the presence of the letter / “ , characteristic in thisalphabet (see already Zalhaas-Neumann 1994:166). Note however theuse of l l instead of l (L, Kaunos 2 ) in ºu≤ ol (vs. u≤ ol in other places).

There are no reading problems.

Photographs: Zalhaas-Neumann (1994: between 166–167). Drawings: Zalhaas-Neumann(1994:161–161, 164).



164

E. T C I G

G 1 (= D 16, Athens)

G 1

→

seÇma ma tÒde: Tur[ KarÚw toÇ SkÊl [ akow ]≤ jas: san tur[ ÉA ]ristokleÇw §p[ o¤e- ]

This is the well-known Athenian bilingual, dated ca. 525/520 B.C. Imaintain the reading adopted in Adiego (1993a), confirmed by my ownrevision of the inscription.

Photograph: Masson (1977:between 90–91). Drawing: Deroy (1955:335). My drawing has been made on my own photographs.

G 2 (= 42*, Thessaloniki)

G 2 (Tzanavari-Christidis (1995)



165

→

qlal i≤ | k?[

This inscription was found in Thessaloniki and edited by Tzanavariand Christidis (Tzanavari-Christidis 1995). It appears on a fragment of a skyphos. The vase is dated by Tzanavari to the third quarter of the5th century B.C. The Carian character of the inscription is indisputable.

Photograph: Tzanavari-Christidis (1995:pl. I). Drawing: Tzanavari-Christidis (1995:14).



CHAPTER FOUR

THE HISTORY OF THE DECIPHERMENT

The history of the research to decipher the Carian Alphabet can bedivided for convenience into three periods:1

1. The ‘semisyllabic era’2. The ‘alphabetic Greek era’3. The ‘Egyptian approach’ and the definitive decipherment of Carian

It must be said that only the third period constitutes the true historyof the Carian decipherment: until the pioneering work of Zauzich (1972),who was the first to correctly identify the solution—although his approachwas unsuccessful—practically all the previous eff orts had produced noresults. This does not mean, however, that all the work preceding

Zauzich is worthless: for example, all the studies devoted to clarifying the inventory of Carian letters—and in this sense, the roles played byOlivier Masson or Vitali ”evoro“kin should not be dismissed—must beconsidered important contributions to the decipherment of CarianIndeed, some intuitions have proved to be right, as for instance Steinherr’sinterpretation of the bilingual coin of Erbbina (see below), although itis true that cases such as this are the exception.

A. T ‘S E’ (1887–1962)

The ‘semisyllabic era’ was dominated by the figures of Sayce and Bork,and lasted until the 1960s. As a conventional date for its end, I use1962, the year ”evoro“kin’s influential article appeared in RHA (”evoro“kin

1 This periodization is diff erent from that adopted in Adiego (1993a). There, it wasmore important to separate the previous unsucessful or incomplete eff orts (including Zauzich and Ray) from the principal goal of the book: to present a complete deci-pherment of the Carian alphabet.



167

1962). It is true that the foundations of the alleged Carian semisyl-labism had already been questioned at the very beginning of the 1950s,

when Theodor B observed that the then recently discovered long inscription of Kaunos pointed to a purely alphabetic system, but it isalso true that the semisyllabic transcription was still used (indeed in aconventional way) by Masson (1959b).

To Archibald H. S we owe, among many other things, the firstpaper to off er an overall table of sound values for Carian letters (Sayce1887[92]).2

In order to decipher Carian, Sayce started with the mixed origin of the Carian writing system: he stated that, on the one hand, a greatnumber of letters were of Greek-Phoenician origin, whereas some signscame from what he called an “Old Asianic Syllabary”, traces of whichwould also remain in the Lycian alphabet and whose last vestige wouldbe the Cypriot syllabary:

It is clear at first sight that the main part of the letters is derived fromthe Phoenicio-Greek Alphabet, but that, as in the case of the Lykianalphabet, certain other characters have been added to express soundswhich were unrepresented in the Greek. Now Dr. Deecke, Dr. Isaac

Taylor and myself have pointed out that these additional charactershave in the case of Lykian been taken from the old Asianic syllabary,a local form of which continued to be used in Cyprus down to a latedate. A probability therefore arises that the additional characters in theKarian inscriptions also come from the same source (Sayce 1887[92]:128).

2 Sayce had devoted a previous article to Carian (Sayce 1874), but Sayce (1887[92])is the first (and last) complete study of the Carian alphabet by the British scholar.



168

Sayce (1887[92])

The following table reproduces that of Sayce (1887[92]):



169

Therefore, whereas a r W m, for instance, come from a Greek model,the letter & has for Sayce the same origin and value as the equivalent

sign in the Cypriot Syllabary. A dental value is assigned to 1 (dh ), onlyon the basis of its similar value in Lycian, assuming that both lettershave the same origin.

In one case, Sayce was forced to accept that a sign coming fromthat alleged syllabary was used in Carian for a sound existing in Greek:the sign p, which he transcribed as mi, me, m on the basis of the valuemi of a similar sign in the Cypriot syllabary. This anomaly is justifiedby the concurrence of the letter s with s .

Sayce opens a series of interpretations of letters that will be contin-ued by many other scholars: the diff erentiation between m and M andthe vocalic value of t are also present in ”evoro“kin’s decipherment.The vocalic character of e and w and the dental character of 1, alsodefended by ”evoro“kin, are nowadays confirmed, following the definitivedecipherment of Carian. Yet this does not mean that the argumentsthen used were as compelling as the present ones, and indeed it is clearthat Sayce’s decipherment, like all theories prior to the Egyptian approach,was wrong.

It is worth noting that Sayce’s system was not exactly semisyllabic.For Sayce, the syllabic component of Carian writing is in fact of a his-torical character: the Carian alphabet comes in part from a syllabary,but it is not actually a syllabary or semisyllabary as such. Our label“semisyllabic era” must therefore be interpreted in a broad sense. Thesemisyllabic character of Carian writing was developed by Bork.

As for the Carian language, Sayce was commendably prudent.Moreover, he was able to establish two basic and correct principles of

interpreting the texts: that the Carian inscriptions of Egypt must con-tain essentially person names, and that -z ending was of genitival char-acter. However, Sundwall (1911) was very much to the point in criticizing the failure of Sayce’s decipherment to off er forms comparable to theCarian onomastics known from the Greek sources. The possible equiv-alences are scarce and—we are now able to say—ill founded: †Ü-z-â-kho-e = Osogva (E.xx 3, now reads in the opposite direction: ione l ),†‡L-e-le-kh-ä = Lelegians (mythical inhabitants of Caria) (E.AS 7: bebint ).

The transcription system of Sayce (1887[92]) is also used in his subse-quent contributions to Carian, although there are doubts, modifications(the clearest ones regarding vocalism, in Sayce 1906) and some incongruities.



170

As for the connection of Carian to other languages, although Sayceassumes provisionally that Carian belongs to the “Aryan family” (Sayce

1887[92]:117), this hypothesis is not based on data from his decipher-ment, but merely on a combination of his own impression that Lydianwas an Indo-European Language, and the notices given by Herodot of a kinship between Lydians and Carians.

Sayce’s decipherment was only used until the appearance of Bork’sarticles (Bork 1930, 1931). Yet despite the failure of his decipherment,and the dilettantism of many of his proposals, Sayce deserves recognitionas the founder of Carian studies: he localized and edited all the textsthen available and established the basis for a scientific study of Carian.

Although they cannot be considered decipherments per se , since theyare merely contributions to the line of research begun by Sayce, someother works must be mentioned: K (1896:377–384), wherean o-value for z was proposed, and where the typical final sequence-xi, read †hä (now ∞i ) was compared with Lycian genitival ending -he ;

T (1903:43–50), wherein Sayce’s system was adopted in an attemptto find a sigmatic nominative in Carian; and S (1911), whoformulated several interesting ideas, for instance, that e was diff erent

from k, and w from K, and that k and K were mere variants of onesign. He also tried to analyze the Carian inscriptions in the wider back-ground of Minor Asian onomastics.

In Torp (1903:44) a value i for t is proposed for the first time, an idealater repeated by ”evoro“kin.

The next decipherment was proposed by F. B, who in two papers(Bork 1930, Bork 1931) laid the foundations of the semisyllabic theory(the so-called Bork-Friedrich system).

In Bork (1930) the proposed decipherment diff ered from Sayce’s intwo basic points:

a) A clear-cut distinction between alphabetic signs and syllabic signs.Moreover, the latter have the same value as the (allegedly) corre-sponding signs of the Cypriot syllabary. Needless to say, in mostcases, the similarities between Cypriot signs and Carian were extremelyquestionable, both formally and phonetically.

b) A fixation of the sound values for the Carian signs more precisethan that of Sayce, based on the assumption that the sound inven-tory of Carian must be not very diff erent from that of Lycian, giventheir common origin (according to Bork) and their geographicalproximity.



171

This latter point could have off ered interesting results—at least regard-ing the criterion of geographic proximity, since it is well known that

nearby languages tend to share phonetic traits—, but Bork had a verysui generis vision of Lycian phonetics, far removed from the communis opinio. For both languages Bork assumed the existence of sounds like/pf/ or /k’h’/ (this latter an aff ricate palatal), and in Carian he estab-lished five series of consonants (labials, dentals, palatals, gutturals andvelars), with three articulation modes for each one ( fortes , aff ricates andspirants). Bork’s transcriptions therefore appear complex, with an arrayof aspirates mixed with syllabic signs. This is his decipherment system,as appears in Bork (1930):

Bork (1930)

172

This is, v. gr. his transcription of E.Sa 1:

R-a-v-ro-l.h’(e).on | No-qh-ro-s.n R(e)-a-kh’.ja | va-kh’-?-no | re.ja-kh’| h(e)-u.vu.kh’

[ “arkbiom zidks mdane yn[-?]mo den tumn]

Bork (1931) focused on the linguistic analysis of Carian. The gram-matical elements that Bork believed to have recognized are not wor-thy of consideration, as his analyses are totally arbitrary. Similarly, themeanings he attributes to the words are capricious. Finally, the typicalCarian onomastics known from Greek sources do not appear anywherein his study. To quote a few examples, in the aforementioned inscrip-tion E.Sa 1, Bork claimed to recognize two Carian names hardly rec-oncilable with the indirect sources, Ravrol and Noqhros . Other “Carianpersonal names” were forms like Kh’asja, Ravuvoc , or Rejajape , equallyaberrant.

Bork (1931) off ers the first attempt to situate linguistically the Carianlanguage based on a particular decipherment. According to Bork, Carian

would belong to an “Old Caucasian” family that would also includeSumerian, Elamite, “Mitani” (nowadays known as Hurrian), Lycian and“Alasian” (the language of the Cypriot Minoic inscriptions). At present,none of these languages is believed to have any connection with theothers: Sumerian is an isolated language, Elamite has been linked— very tentatively, however—to the Dravidian family, Hurrian forms anindependent group together with Urartian, Lycian is an Anatolian Indo-European dialect, and “Alasian” is yet to be deciphered. Needless tosay, all these speculations, based on an invalid decipherment and anonexistent linguistic family, have been totally superseded.

The relative success of an overly unscientific decipherment must beattributed to the authority of J. F and W. B.

Friedrich used it in his important work Kleinasiatische Sprachdenkmäler (Friedrich 1932) for transcribing the Carian inscriptions. However,Friedrich himself recognized that Carian could still not be seen as com-prehensively deciphered (Friedrich 1932:91) and uses a simplified tran-scription that substitutes for <pf>, <k’> for <k’h’>, etc., which

in fact means destroying Bork’s complex speculations about the soundsystem of Carian:



173

Friedrich (1932)



174

The so-called Bork-Friedrich system was adopted in a purely instrumen-tal way by Masson in his editions of Carian texts until 1976.

Brandenstein also adopted Bork’s system—although with some modi-fications—in his article “Karische Sprache” for the Pauly-Wissowa’sRealencyclopädie (Brandenstein 1935a): he not only took on the simpli-fications of Friedrich (1932), but also recovered two values proposedby Sayce: t = ä (Bork he ) and i = e (Bork h’e , Friedrich he ). His pro-posal of attributing a value p to x X is new, and based on the allegedlabial value of X in Lydian (at present, this Lydian letter is in fact com-monly interpreted as representing a labiovelar [transcribed by q ]).3 This

proposal, put forward by Brandenstein in a former paper (Brandenstein1934b) was unsuccessful:

3 See now Gérard (2005:56–57).

Brandenstein (1934b)



175

4 “Die entscheidende Erkenntnis ist, daß das karische Alphabet keine Mischung von

Brandenstein made some eff orts to find Carian personal names in theinscriptions: in E.Ab 16, read as †k-mi-v-o-s-k-vo (now nprosn≤ ), he believed

he had identified the final element of the Carian names Se-skvw, Sa-skvw. In other cases, he included Bork’s interpretations. In general,Brandenstein’s contributions are scarce and not particularly convincing.

After Bork’s decipherment and until the end of 50s, I know of onlyone attempt to decipher Carian letters: that of A. M (Mentz 1940).Its only interest lies in being the first work to employ the Carian-Egyptian bilingual inscriptions in order to establish sound values forthe Carian signs, and in having ruled out the alleged semisyllabic char-acter of Carian writing.4 Otherwise, Mentz’s decipherment is an exer-cise in dilettantism, and has received harsh, but merited, criticism: “ganzdilettantisch in Lesung und Deutung” (Friedrich, quoted by Masson1973:207); “tentative ambitieuse (. . .) qui n’a pas convaincu” (Massonibid. ); “It is difficult to imagine that there can exist works on Carianlanguage more prone to fantasizing than Bork’s ‘researches’. Mentzshowed that it was indeed possible” (”evoro“kin 1965:51; my translation).

Mentz (1940)



176

Mentz’s use of bilingual inscriptions is very surprising, as he tries tofind Egyptian verbs and nouns in the Carian parts. The resulting Carian

is a somewhat bizarre, mixed language. In his translation, the Carianshort graffiti turn out to be of a pleading tone: “Râ, gib”, “Râ Amon, gib Leben”, “Hathor, gib”. Such demonstrations of Carian devotion toEgyptian deities may indeed be praiseworthy, but when they are basedonly on an analysis like E.Si 11 = †‡pn – obpya “Apis Ptah”, they arehighly suspect.

B. T ‘G A ’ E

In 1949, George E. Bean found an inscription of Kaunos (C.Ka 2)that remains to this day the longest Carian text conserved. Almostimmediately, Th. B (apud Steinherr 1950–51:332) rightly observedthat this long inscription, where less than 30 diff erent letters were used,pointed clearly to a pure alphabetic system: “Nach dem Zeichenbestandder Kaunos-Inschrift zu urteilen kann von einer Mischung von Alphabet-und Silbenscrhrift nicht die Rede sein. Es handelt sich um eine reine

Buchstabenschrift (. . .)”. It is true, as we have seen above, that Mentzwas in fact the first to discard the semisyllabic approach and to defendthe purely alphabetic nature of the Carian writing system, but his onlyargument was his decipherment itself, which we have already establishedas being even worse than Bork’s (if this is possible). In any case, it wasBossert’s well-reasoned contribution that put an end to the alleged semi-syllabism of Carian. In general, the proposals of decipherment of thisnew period concur in their assumption that the Carian writing systemis alphabetic, and in attributing Greek values to letters of Greek shape.5

Between Mentz’s and Stoltenberg’s (see below) decipherments, it isalso fair to mention the contributions of F. S (Steinherr 1950–51,1955). Steinherr did not off er a concrete decipherment, rather he lim-ited himself to formulating remarks on certain Carian signs. His obser-vations are generally very thought-provoking and, in some cases,surprisingly accurate. Among them, the following four must be highlighted:

Laut- und Silbenschrift ist, wie man seit Sayce allgemein annahm, sondern ein reinLautschrift.” (Mentz 1940:279)

5 The sole exception is Stoltenberg’s decipherment, where syllabic values are assignedto a few letters in a rather capricious way. But even he refuses any connection toCypriot syllabary, see below.



177

1. Steinherr (1950–51:336) proposed a liquid value for the sign z resort-ing, among other arguments, to the comparison of the “genitival”

endings in -z with the oblique case in -l of Lydian. This proposalanticipated an idea developed later by ”evoro“kin, but which hassince been superseded.

2. Steinherr suggested that j had to have a value near to i, becauseboth alternate in some inscriptions. Consequently, he proposed atranscription ê / e , respectively. This observation, for many yearsignored (a nasal-like value for j was for a long time the preferredtheory), has since proved to be correct, now that the Carian alpha-

bet has been definitively deciphered (i = i , j = j ).3. Steinherr pointed out that the sign f F (and variants) could hardlyhave an r value, because it often appears in an initial position,whereas Anatolian languages show a clear reluctance to an initial r ,a well-founded argument reiterated by Ray (1987).

4. For this latter sign, Steinherr suggested a t value. We now knowthat he was wrong, but only partially: the sign has in general an “ value, but the bilingual inscription of Kaunos has revealed that inthe Kaunian alphabet the letter T was actually used for t . Moreover,

in at least one case his analysis was right: he proposed that the let-ters i F in the bilingual coin of Erbbina, dynast of Telmessos, readt e , could correspond to the Lycian legend teleb(ehi) erbbina ‘Telmessos/ Erbbina’ in other coins (Steinherr 1955:184–192). Similar rea-soning is nowadays used to explain the Carian legend i F, by assum-ing that the alphabet used is of Kaunian type (a very plausibleassumption from a geographical point of view), see Adiego (1998b:58–60), Meier-Brügger (1998:45).

Steinherr’s work can hardly be considered a decipherment, but it doesdeserve a great deal of respect: before the “Egyptian approach”, thiswas the only moment in the history of decipherment that a correctmethod of analyzing the Carian alphabet had been envisaged, even if it was in a very limited way.

The fifties ended with another failed attempt at decipherment, byH. L. S (Stoltenberg 1958a, b, 1959).

Stoltenberg started with the automatic attribution of Greek values to

the allegedly Greek letters: a = a , o = o, m = n, p = m, and so on.As for the apparently non-Greek signs, he rejected any comparison withthe Cypriot syllabary and proposed as an alternative to resort to theMinor Asian onomastics of Greek sources, and the comparison with



178

Stoltenberg (1958a)

‘Termilian’ (= Lycian). This latter hypothesis would in principle beacceptable, were it not that that Stoltenberg, like Bork, had a very

peculiar vision of Lycian: he believed that Lycian belonged to a groupthat he called ‘Laric’ together with Etruscan, ‘Tyrsenian’ (= Lemnian,an Etruscan dialect of the Lemnos island) and the language of the“Cretan pelasgians”. Moreover, this alleged ‘Laric’ group would be con-nected with Uralic languages. This classification of Lycian is clearlywrong, and most of the meanings for Lycian words that Stoltenberg proposed are also completely incorrect (see Neumann 1969b:364).

If this approach does not promise good results, any hope of successvanishes when the values attributed to the ‘non-Greek’ letters are takeninto account: let us examine the following table that contains his deci-pherment system (Stoltenberg 1958a):



179

By attributing values such as ija, nda , and öm to some letters, it iseasy to obtain words of a falsely Anatolian flavour. For instance,

‡1wmuip ← ‡ piumz , now read piubez (E.Ab 10), transcribed as †‡m-e-u-n-uwa-nda can be easily compared with the typical Anatolian namesin -anda, -wanda . However, even resorting to these tricks Stoltenberg was unable to obtain convincing results.

It is worth remembering that during the fifties the Carian corpusunderwent a considerable improvement. It was during this period thatthe complete corpus of Carian inscriptions of Caria to date was pub-lished by Robert (1950) and Deroy (1955), Masson and Yoyotte col-lected the inscriptions on pharaonic objects (Masson-Yoyotte 1956), anda provisory edition of graffiti of Abu-Simbel appeared (Bernand-Aly1959?). Moreover, the important Greek-Carian bilingual stela of Athenswas discovered, as well as graffiti in Thebes.

At the beginning of the sixties appeared the first works of Vitali”“, one of the most prominent Carian scholars, and the authorof the first book devoted exclusively to the decipherment of Carian(”evoro“kin 1965).

”evoro“kin’s method off ers some innovative elements. In order to

separate vowels from consonants, he resorts to typology: given that thelanguages combine vowels and consonants in diff erent ways, which im-plies diff erent patterns of sequences such as CVC. VCV, etc., ”evoro“kintries to establish the patterns of Carian on the basis of personal andplace names from Greek sources. Remarkably, the results are not farremoved from the distribution of vowels and consonants in Sayce’sdecipherment. The explanation for these rather surprising results is thatSayce himself took into account the distributive properties of the signs,albeit in a rudimentary way.

After establishing a first separation between vowels and consonants,”evoro“kin moves on to the statistical comparison between Carian signsand the phonemes present in the names of Greek transmission. However,in some cases this criterion seems to be used merely to confirm a valueobtained by other means. These other methods are the formal anal-ogy with Greek letters and the identification of onomastic elements inCarian words.

The following tables reflect the system of ”evoro“kin as it appears

in ”evoro“kin (1965) and thirty years later, in ”evoro“kin 1994, respectively:



180

”evoro“kin (1965)



181

”evoro“kin 1994

182

It is clear that ”evoro“kin follows the traditional equivalence betweenCarian signs and Greek letters in a great number of cases. Regarding

the “non-Greek” signs, one also comes across values already outlinedby other scholars: t = i , suggested by Torp 1903; z = l (1965) / L (1994), for which a liquid value was defended by Steinherr. A less exactprecedent is the value ù for e (Sayce ô ) or the dental character of c,already suggested by Bork. However, other proposals are totally new:j = ñ (based undoubtedly on the Lycian N ñ ); & = t (1965) / D (1994), a dental sound, based primarily on the sequences m& that”evoro“kin identified with the typically Minor Asian consonantal groupond o ); k K = p (1965), b (1994), based on statistical criteria and on thecomparison with names from indirect sources.

Careful examination of both tables shows that the changes intro-duced by ”evoroskin during three decades of research on Carian wereof no great importance. Perhaps the most significant modification is v

= U , resulting from the observation of Memphis corpus, where it alter-nates with u = u (formerly, ”evoro“kin took v as a sort of e ). In theremaining cases, the modifications are aimed at adjusting the tran-scription to etymological explanations. This is the case, for instance, of

x X, transcribed as q (formerly h ) in order to better reflect his ety-mology of xi †qe [now ∞i ] < PIE *k w e , or of k K, previously tran-scribed as p and more recently as b.

One of the most characteristic values that ”evoro“kin defended wasz = L (formerly transcribed as l ). This was a not totally new idea (itwas previously considered by Steinherr 1950–51, see above), but in”evoro“kin’s hands it took on greater importance. Using this value, hebelieved he could recognize the Carian name Lujhw in zu0Li-, readas †Luxze- [now ≤ ug li- ], comparing the genitival ending -z = †-L withthe genitival adjective ending -li- in Lydian, and he claimed to findtypical Carian names in -vllow/-vldow in some sequences uz = †uL [now u≤ ]. Obviously, these proposals have since been proved erroneous,as the sign has been definitively identified as representing a sibilantsound.

However, the main problem with ”evoro“kin’s decipherment is thatthe typical Carian names attested by indirect sources do not appearanywhere in Carian inscriptions if we adopt his system of decipher-

ment. In this sense, a case such as Lujhw = †Luxze- is almost unique.In most cases, the only means of establishing connections between”evoro“kin’s readings and the stock of Anatolian names in Greek (andHittite) sources is to resort to partial comparisons of constituents, ety-



183

mologically reconstructed forms, complex sound changes, and so on.Indeed, such techniques can be used to explain any form as Anatolian

and/or Indo-European. However, almost none of the typical Cariannames (for instance Ussvllow and related forms) are obtained with thisdecipherment. Masson rightly pointed out this deficiency:

Contrairement à bien d’autres tentatives, celle de ”evoro“kin paraît raison-nable: les discussions philologiques sont judicieuses et les comparaisonslinguistiques sont intéressantes, sinon toujours convaincantes. Peut-on alorsparler d’un véritablement déchiff rement? Il ne le paraît pas, car l’ensembledes lectures ne donne pas une impression d’évidence, et les noms d’hommescariens (. . .) ne se retrouvent pas de manière tangible.

This inability to find recognizable Carian names could explain why, inhis works of the 80s and 90s, ”evoro“kin focused his eff orts on thesearch for nouns and verbs in Carian texts, paying far less attentionto the amount of onomastic formulae provided by the recently pub-lished corpus of Memphis inscriptions. ”evoro“kin seemed to feel morecomfortable translating long and complex texts in a rather speculativeway than analyzing the abundant personal names in the laconic funer-ary inscriptions from Saqqâra.

Nowadays, in the knowledge that ”evoro“kin’s decipherment was afailed attempt, continuing critical analysis seems unnecessary, and it ispreferable to instead note some important merits of his work. The first,although apparently negative, does in fact have a positive reading:despite his methodological weakness, due above all to the abuse ofetymological and analytical speculations, ”evoro“kin’s study remainsthe only serious eff ort to interpret the Carian alphabet as a ‘normal’Anatolian alphabet, similar to Phrygian, Lydian and Lycian writing sys-

tems, wherein letters with Greek shape have their Greek or like-Greeksound value, and other letters are added to reflect peculiar sounds notfound in Greek. After ”evoro“kin’s failure, this method became imprac-ticable, and a diff erent approach became necessary.

The other merits of his work are decidedly more positive: ”evoro“kincontributed decisively to the demonstration that Carian writing is purelyalphabetic. He showed that the seemingly large number of letters wasdue to the existence of alphabetic varieties, which he was able to sep-arate correctly. Moreover, ”evoro“kin has always defended the suppo-

sition that Carian was an Anatolian Indo-European language, ruling out the attempts of the dilettanti to connect it with other languages, andemphasizing the need to include the other Anatolian languages in ananalysis of Carian forms.




184

Reviews of ”evoro“kin 1965: Pisani 1967, Gusmani 1967 (laudatory,although he believes that it is precipitate to consider the decipherment

as done), Zgusta 1968 (totally favourable: “We can then, conclude that . . . thebasic step is done and the Carian inscriptions are really deciphered”(p. 154)).

In contrast to the predominantly serious approach adopted by ”evoro“kinthe so-called decipherment of R. S (Shafer 1965) can be takenas an example of how a language must not be deciphered. He com-bines a lack of rigor (diff erent values assigned to the same sign, absenceof a table to display sound values, senseless hypotheses about the mean-ing of the inscriptions . . .) and an arrogant dismissal of other scholars,which explains the harsh but just opinion formulated by Masson: “ten-tative vaine et très prétentieuse d’un renouvellement complet des lec-tures et des interprétations” (Masson 1973:211).

As is often the case in other attempts to decipher Carian, the begin-ning of Shafer’s work seems very promising, despite the somewhat inap-propriate tone. He rightly criticizes Bork-Friedrich’s decipherment system,the outcome of which is a language he derisively terms “Super-Hawaiian”because of the profusion of vowels and semivowels, phonologically very

distant from the Carian names in Greek sources.But the interest of his paper ends here. When trying to establishsound values for Carian letters, the solutions he resorts to are routinelybizarre: for instance, he attributes a strange value br to p, takes theclear interpunction sign | as a letter i , and imagines surprising evolu-tions of the letters, like L > R > r. As for the linguistic interpretation,in the Egyptian-Carian bilinguals, he claims to identify stereotyped for-mulae referring to pharaohs in the Carian part. In many cases, he evenreads the texts in the opposite direction to that commonly accepted.

One of his “interpretations” surpasses Mentz’s litanies:

†(q)-lolak qirmdun b(a)bu kimrda (E.Si 4)

According to Shafer, this means: “In the language of the Leleges qirm-dun is spoken kimrda ” or “The Leleges qirmdun pronounce kimrda ”. It isnot easy to imagine the reason for which a Carian mercenary wouldfeel the need to write a sentence such as this, which instead seems tobelong to a sort of Carian Appendix Probi !

This period of the history of Carian decipherment ends with anothersomewhat superficial eff ort, that of Ju. V. O“‘ (Otkup“‘ikov1966; cf also Otkup“‘ikov 1968, on the origin of the Carian alphabet).His proposal is shown below:




185

Otkup“‘ikov (1966)



186

These are the more notable elements of his decipherment:

1. Otkup“‘ikov assumes that some signs have two forms, one frequent(‘ asto ), and other more infrequent or rare (redko ). For instance, B

would be the infrequent form corresponding to the frequent a (!).2. By this logic, Otkup“‘ikov converts the Carian alphabet into a pure

Greek alphabet, wherein the apparently non-Greek letters are sub-sumed beneath the Greek ones. The values attributed are in gen-eral also the Greek ones, note for example j = ks .

3. In fact, for Otkup“‘ikov, the Carian language is merely a Greek dialect .

None of these three assumptions are convincingly argued. The group-ings are largely unjustified and absurd (for example k = u ), andOtkup“‘ikov arbitrarily assigns diff erent values to a single letter: forinstance, q is both a frequent letter with value th and a redko letterequivalent to o, and similarly in the case of B (=l / a ).

As for the supposed Greek character of Carian, the dialect result-ing from this decipherment proves to be rather curious: ue uh wouldmean ‘daughter’ and would be connected with the Greek uflÒw ‘son’(!)

(Otkup“‘ikov 1966:22); paraeWm, read SaWaieien would be an opta-tive (Otkup“‘ikov 1966:24) and so on. It is not difficult to agree withMasson’s judgement: “on aura peine à conclure autrement que par unscepticisme total devant la langue qui nous est proposée . . .” (Masson1973:193).

Reviews of Otkup“‘ikov (1966): O. Masson (1967), Heubeck (1967–68), Jordan (1968), all adverse.

Before concluding this section, it is worth mentioning the contributions

made to Carian studies by the Italian scholars Piero M andRoberto G. Their presence here can be justified by the fact that,in their approach to the language, both scholars broadly adopted the‘Greek alphabetic’ interpretation of the Carian writing system.

In the 60s, Meriggi published two papers on Carian (Meriggi 1966,1967). The first, though devoted to the study of Para-Carian tablets,contains very useful inventories of signs. The second consists of a seriesof remarks regarding several inscriptions from Caria itself and certain

Carian letters. Although Meriggi did not off er a decipherment as such,he assumed implicitly that the “Greek” letters had Greek sound val-ues. In the 70s, Meriggi dedicated two of his last works to Carian:Meriggi (1978) adds little new material to Meriggi (1967), aside from




187

his argument in favour of the Luwian character of Carian based in aprocess of elimination. More interesting is his review of Masson (1978

(Meriggi 1980), in which he leaves aside any discussion of the lettervalues and aims instead to analyze the structure of the inscriptions of Saqqâra, with some thought provoking hypotheses, such as the mean-ing ‘son’ for mno-≤ (already suggested in Meriggi 1967:223), or the pos-sibility that mwdon-≤ could be an ethnic (‘Caromemphite’).

Gusmani’s first approach to the study of Carian—other than hisreview of ”evoro“kin (1965), Gusmani 1967—was his edition of theCarian graffiti from Sardis (Gusmani 1975), wherein he adopts the con-ventions of ”evoro“kin’s system (except in a few cases, such as j = †≤ ,not †ñ ). In any case, the very limited linguistic usefulness of these graffitidoes not allow Gusmani to enter into discussions about the sound val-ues of the signs. Similarly, no linguistic discussion is present in Gusmani(1978), which contains two Carian inscriptions: Gusmani adopts Masson’ssystem of transcription, with the sole diff erence of & = †≥ (= Masson“31”).

In fact, the very first work by Gusmani devoted specifically to Cariandecipherment itself is Gusmani (1979). Here he focuses on the letters

2, t, i, e, K and j. The most valuable contribution is the observa-tion that t alternates with d, which invalidates ”evoro“kin’s theory of a vocalic value for t. Less successful is his attempt to defend a simi-lar alternation between e and o, leading to a transcription of e bymeans of †ò. Other suggestions are equally unconvincing, with the soleexception of the alternation j / i. Most of the ideas Gusmani putsforward in this article are no longer recognised, having been super-seded by the definitive decipherment.

C. T ‘E A ’

1. The First Attempts

The Egyptologist K.-Th. Z deserves to be considered as the pio-neer of the so-called ‘Egyptian approach’: his use of Egyptian-Carianbilinguals as the starting point to decipher the Carian alphabet is an

approach that has proved successful.From a historical perspective, it is interesting that for one hundred years (from Sayce’s first work to Zauzich’s monography), with the sole



188

and frustrating exception of Mentz, nobody considered the possibilityof using the bilingual inscriptions to decipher Carian, despite this being

the most customary way of focusing a decipherment.The flaw in Zauzich’s attempt is its inability to establish a clear con-

nection between Greek and Carian parts of the bilingual inscriptions.Practically the only correct value that he discovered was p = p, whichappears at the beginning of three names in both the Carian and Greekpart of three bilinguals (E.Sa 2. E.Me 5 and E.Me 8).6 Other than thisequivalence, both the methods and the results are very disappointing:he corrects Carian texts whose readings are not in doubt, off ers uncon-vincing equivalences, like Prjm = † patalem [now paraeym ], assigns thesame value to diff erent letters, as b e = l , z v = s , and, finally, con-cludes that Carian is a Greek dialect, rendering his study as defectiveas Otkup“‘ikov’s (see above). Consequently his decipherment receivedfierce criticism from serious scholars like Masson and Heubeck.

Reviews of Zauzich (1972), both very critical: Masson (1973[74]), Heubeck(1974:97 col. 1): “Die Entziff erung Z[auzich]s hat uns . . . keinen Schrittweiter gebracht”.

The way opened by Zauzich (1972) was followed by Thomas K,with more promising results (Kowalski 1975). The most striking ele-ment of Kowalski’s only paper devoted to Carian is that he sets mostof the values later established by J. D. Ray, but he does so in anunusual way, often starting from weak or even erroneous suppositions.Moreover, his article gives the impression of being a condensed ver-sion of a more extensive study on Carian that was never published:nowhere does it attempt to justify, for example, the grouping togetherof clearly diff erent signs or the separation of mere variants.

6

Zauzich also believed that a Carian name beginning with p could be connectedwith an Egyptian name with an initial p in E.Me 6, an incorrect assumption, becausethis is a case of a re-used stela, and indicates no correspondence between Egyptianand Carian epigraphs. It is also true that Zauzich was right in assigning a sibilantvalue to z, but this attribution was based on faulty interpretations.



189

Kowalski (1975)



190

I consider the main deficiencies of Kowalski’s proposal to be thefollowing:

1. Kowalski tried to obtain and correlate results from very dissimilarinscriptions, treating Egyptian-Carian and Greek-Carian inscriptionsin the same way, even though connections are far more difficult toestablish within the latter type, increasing the risk of falling into haz-ardous speculations. He also considered the inscription E.Me 6 trulybilingual, when it is in fact a reused stela and indicates no connec-tion between Carian and Egyptian names.

2. His transcription and interpretation of the frequent sequence xi as† yrº ‘son’, clearly reminiscent of Zauzich (1972),7 was not only absurd per se , but also seriously impeded the overall decipherment, becauseit implies an rº value for i, a clearly vocalic letter.

3. It has already been mentioned that Kowalski’s article lacks a detailedexplanation of the treatment of certain Carian letters as mere vari-ants, and vice versa. The above table demonstrates how inconsis-tent these groupings and separations are: Kowalski separates N (=t’ ) from ø (= h ), L (= z ) from 2 (= ), and yY (= c ) from i (=

r º ). Conversely, he puts together v and ç (= t ). The most surpris-ing case is z: two diff erent values are off ered: ¬ , and Æ . In the firstcase it is grouped with 9 and 6, and in the second, with ñ. Allthese assumptions have turned out to be false.

4. Kowalski does not apply the values obtained from the bilingualinscriptions to the rest of Carian texts, which would be the onlyway to check the validity of the decipherment, particularly regard-ing those signs whose values are dependent on the weak interpre-tation of a single inscription.

To sum up, Kowalski’s decipherment—though it far exceeds Zauzich’sattempt when Egyptian-Carian bilinguals are used, off ering correct val-ues for some important signs (p = p, k = k , r = r , m = m, b = b,f = “ )—is fatally flawed by excessive speculation and its chaotic treat-ment of signs and variants that inexplicably disregards the illuminating

7 = Zauzich ‡ir . However, Kowalski does not make reference to Zauzich’s inter-pretation. Zauzich aimed to connect this form with Greek uflÒw, an assumption unten-able from the point of view of both Greek dialectology and the structure of Carianinscriptions.



191

work of ”evoro“kin and Masson. As a result, its usefulness is extremelylimited, working only for those words where the correctly deciphered

letters mentioned above appear almost exclusively, for instance infarKbiom = †“ arkbrºom [now “ arkbiom ], a not entirely correct tran-scription—note rº— based on the Egyptian form ” 3rkbym.

The ‘Egyptian approach’ took a brief step back with Jean F

(Faucounau 1980). In this article, the Egyptian-Carian bilinguals areconsidered, but the overall attempt is rendered worthless by the author’scustomary dilettantism. Faucounau bases his work on two assumptions:that Carian local alphabets are very dissimilar to each other, and thatthey may reflect diff erent dialects or even languages.

Both assumptions are untenable, at least in the form in whichFaucounau develops them: it is true that there are local alphabetic vari-eties, with some peculiar forms in each case for certain signs, butFaucounau proposes absurdly that letters such as f or t may havediff erent values in diverse inscriptions from Egypt, where the generalunity of the Carian alphabet is undeniable. As for the existence of diff erent languages, not only is there no evidence for this assertion, butin fact the presence of similar words and endings in distant Carian

sub-corpora points in the opposite direction. The presumed existenceof diff erent alphabetic systems and languages combined with a markedlyunscientific approach result in an continuously ad hoc assignation of sound values to the letters and to an inconsistent decipherment. Faucounaualso attempts to identify proper names, but in many cases he resortsto alleged “Carian” names transcribed in Greek (without any reference)that do not appear in either Zgusta’s or other similar collections.

2. The Seminal Work of RayThe ‘Egyptian approach’, as the definitive method for deciphering Carian,begins in earnest with the fundamental work of British Egyptologist John D. R . It must be pointed out that for his decipherment Raystarts from scratch: it is clear that when he begins to work on Carian,he is still unfamiliar with Kowalski’s article. Whilst his deciphermentof Carian cannot be considered complete, the definitive deciphermentcould not have been produced—at first independently, and later jointly

by Diether Shürr and myself—without the seminal work of Ray, whichacted as its starting point.Ray’s first work on Carian is his recension of Masson (1978), received

by the editor of the Journal of Egyptian Archaeology ( JEA ) in 1980,



192

but published three years later (Ray 1983). This brief article is impor-tant for two reasons: here Ray already establishes a criterion for the

use of bilingual inscriptions, consisting in leaving aside those inscrip-tions where the individual has an Egyptian name in the Egyptian partand deciphering only those where the Egyptian part contains a non-Egyptian, presumably Carian, name. The other contribution is hishypothesis, now confirmed, that the name Psammetichus can be identifiedin the Carian sequence of signs kfa«sip ←/ pismazk →,which would suggest totally innovative values for allegedly ‘Greek’ signsas p, m or f.

It is interesting to note that the first assertion is now known to bepartially false: Egyptian names from the Egyptian part of the bilingualinscriptions do appear in the Carian part (E.Sa 2, E.Me 5). However,in somewhat paradoxical fashion, this limitation imposed on the inven-tory of bilingual inscriptions proved instrumental in triggering the firstdramatic steps towards a definitive decipherment. Apart from the twoinscriptions mentioned above, Ray also left out E.Me 6, in this casewith good reason: there is no correspondence between the names inthe Carian and Egyptian parts. So we can see that, unlike Kowalski,

Ray did not obtain incorrect sound values for his decipherment byusing an inappropriate bilingual inscription. Kowalski’s error was, aswe have seen, to include this inscription in his decipherment.

Ray (1981) off ered the first complete proposal of decipherment, basedon the inscriptions of Saqqâra. His method consisted in bringing thesound values obtained from the bilingual texts to the remaining inscrip-tions, in order to find Carian proper names comparable to those pre-served in Greek sources. This onomastic comparison as a form of confirming and developing the decipherment was not new, as the his-tory of Carian studies shows. The singularity of Ray’s approach to theonomastics of indirect sources lies rather in the cautiousness with whichhe operates: the onomastic identifications are not so numerous, butsome important similarities begin to appear. Another prudent decisionwas the initial limitation to the Saqqâra corpus: almost all the bilin-gual inscriptions used were from Saqqâra, where a very unitary andstandardized alphabet was used. This left little room for confusionbetween letters and variants, a risk that the former followers of the

‘Egyptian approach’ did not take care to avoid. It must be added thatRay was also much more rigorous than his predecessors in the use of the Carian alphabet: unlike some, he did not distort the Carian sig-nary in order to obtain a particular equivalence between forms.



193

In Ray (1982b), undoubtedly his most ambitious and important work,numerous inscriptions are transcribed and briefl y commented, not only

those from Saqqâra, but also the graffiti from Buhen, Abu-Simbel,Silsilis and Abydos, as well as other inscriptions. Ray (1982a) is an arti-cle focused rather on historical aspects and theoretical problems of deci-pherment. In Ray (1987) he responds to criticisms formulated by Gusmani(1986), by comparing the results of his own decipherment with thosegained from the traditional approach.

Ray (1982b)

The proposed sound values vary slightly from one article to another,but they can be characterized in the following way:

1) A significant number of signs had been transcribed in a similar man-ner by Kowalski. The coincidence is logical: these are the letters thatappear in the bilingual inscriptions used by both authors. However, incontrast to Kowalski, Ray does not off er arbitrary groupings or sepa-rations of signs.

Ray’s use of bilingual inscriptions is shown in the table of p. 194.There were admittedly some problems of interpretation: the value of l, that in Egyptian appears adapted by means of r , and which was



194

S o u n d v a l u e s e s t a b l i s h e d b y J o h n D .

R a y o n t h e b a s i s o f C a r i a n - E g y p t i a n b i l i n g u a l s

I n s c r i p t i o n

C a r i a n

E g y p t i a n

R a y ’ s

C u r r e n t

R e s u l t i n g s o u n d

v a l u e s

f o r m s

v e r s i o n

t r a n s c r

i p t i o n

i n t e r p r e t a t i o n

f = “

r = r

k = k

m = m

p

= p

l = d s

= s x = h c

= ‘

( R a y 1

9 8 2 b )

E . S a 1

f a r K b i o m ” 3 r k b y m

“ - a - r - k - b - e - o - m

“ a r k b i o m

+

+

+

+

E . M e 5

p s m f k . . . P s m ∆ k . . .

p - s - m - “ - k - . . .

p s m “ k . . .

+

+

+

+

E . M e 7

C a m o u

T 3 - n - j m . w

‘ - a - m - o - u

t a m o u

+

+

E . M e 7

C a n a i z

T 3 [

‘ - a - k ‘ - a - e - ≤

t a n a i ≤

+

E . M e 8

p a r a e W m

P r j m

p - a - r - a

- j - é - m

p a r a e y m

+

+

+

E . M e 9 , 1

5

a r l i f z

J r “ 3 , J w

r “ 3

a - r - d - e

- “ - ≤

a r l i “ ≤

+

+

+

E . M e 9

a r l i o

[ J r y [ m ] 3

a - r - d - ê

[ s i c ] - o -

[ m - ≤ ]

a r l i o [ m ≤ ]

+

+

+

E . M e 1 5

u r s x l e z

N r s k r [ s i c ;

u - r - s - h - d -

j - ≤

u r s ∞ l e ≤

+

+

+

+

n o w : 3 r s k r ]



195

troublesome for Ray; the only partial analysis of the name in E.Me 5;the problematic reading Nrskr (now read 3rskr ); the imprecise analysis

of vocalism, hampered by the almost purely consonantal adaptation inEgyptian. Yet most of the values proposed by Ray have in fact turnedout to be right. Note that a great number of the sound values wereestablished on the basis of more than one correspondence: m = m resultsfrom its presence at the end of two names (” 3rkbym, Prjm ), and in themiddle of a third one (Ô 3n-j.mw ), f = “ is established thanks to E.Sa1, E.M 5, E.Me 9, and E.Me 15, etc.

2) Other values are the result of graphical alternations in the inscrip-tions of Saqqâra already noted by Masson (1978) and Meier-Brügger(1979a): i / j, u / v, etc.

3) Curiously, two of the most important sound values proposed byRay were not obtained from the bilingual inscriptions: L = l (Rayused ld to transcribe it) and z = ≤ .

The first equivalence- totally new- is based on C.Ki 1(Kildara) †k-e-ld-a[ (now ki l a [ ) interpreted as the Carian form of the place nameKildara (following Kowalski’s suggestion, although he interpreted L as†z ), and the endings in -oL = †o-ld (now -ol ), that Ray correctly com-

pared with the typical Carian names in -vllow from Greek sources.The second equivalence (z = ≤ ) was already suggested in Meier-Brügger(1979a)—as Ray pointed out—but based on weak argumentation, andwithout exploiting the consequences. Ray’s basis for this equivalencewas initially only the interpretation of the name vzoL- as †ü-≤-o-ld-(now w ≤ ol - ), corresponding thus to the Carian name Ussvllow (Ray1981:161), although it was subsequently noted in Ray (1982b:189) thatmore convincing evidence could be found: the clear alternation betweenz and f = ≤ in the Egyptian name pisma ≤ k / pisma ≤ k .

4) For the remaining signs, whose values he could not ascertain fromthe bilinguals used, Ray generally kept to the traditional transcriptionused by other scholars (”evoro“kin, Meriggi, Gusmani, Masson), id est,that is based on the similarities with Greek alphabet: d = g , t = q , n= k, k’ [= kt ?], q = t, even l = d , etc. This is also valid for vocalicsigns: given the consonantal character of the majority of Egyptian tran-scriptions, Carian-Egyptian bilinguals were useless for establishing vocalicvalues, and Ray chose to keep the traditional value e for i, despite

the evidence of onomastic identifications, that pointed clearly to i .In Ray (1988), several inscriptions from Caria itself are analysed.The most relevant contribution is the comparison of the final sequence



196

in C.Hy 1, read †é-d-a-r-m-e-ld (now ylarmi t ) with the place name wherethe inscription was found, ÑUllãrima.

Two further articles by Ray previous to the definitive decipherment shouldbe cited here: Ray (1990a) is an ambitious attempt to systematize theresults of his decipherment, but there are few new ideas. The most strik-ing contribution is the discussion devoted to the sign n, in which Raymentions, and rejects, the proposal of a nasal value suggested to him pri-vately by Schürr.

Ray (1990b) off ers an interpretation of the initial line of C.Si 2, whichowes much to Schürr’s brilliant analysis, also privately communicated (seebelow). However, in trying to adapt this to his own decipherment system,

Ray seriously distorts Schürr’s ideas and the final result is unconvincing.The role played by John D. Ray in the decipherment of Carian is cru-cial: for the first time, true Carian names emerge from the inscriptions,although in some cases Ray’s transcription does not correspond exactlyto the Greek adaptation: †u-≤-o-ld = Ussvllow, †“-a-r-u-≤-o-ld =Sarussvllow, †a-r-d-e-“ = Arlissiw. Even more importantly, a greatnumber of the values involved in these forms were established fromthe independent evidence of Egyptian bilinguals and the alternation of

letters.Yet despite these merits, Ray was unable to off er a complete deci-pherment system. The values proposed for very important consonan-tal signs were either wrong (d t q n ), or inaccurate (l & X ), whileothers remained undeciphered. As for vowels, Ray’s transcriptions werealso imprecise, when not mistaken (i = †e, j = †ê, e = †j, W = †é),to the extent that the Egyptian bilinguals were rendered useless giventhe almost purely consonantal notation of Carian names. Only in thecase of a o u v and to some extent w, did his transcriptions seem

to more accurately reflect the actual value of the letters. The soundsystem that arose from his proposal was also inadequate, the mostimportant gap being the absence of a letter for n, the basic nasal con-sonant in all the sound systems of the world. As for the onomasticidentifications, although the spectacular comparisons cited above pointedclearly to the accuracy of the Egyptian approach, they were counter-balanced by the great number of personal names in the Carian inscrip-tions left without suitable interpretation. Forms such as †ü-d-e-a-q,†‡e-r-t-u-t, †‡t-q-b-a-j-m-, †“-a-ju-g-e-t-, †‡“-a-r-ü-k-ê-a-q-, †‡m-s-k’-o-r-e-,

(now wliat, “ !rquq, qtbl !em-, “ a ÿdiq-, “ arwl ! jat-, msnord- ) and many others,which are now easy to interpret from the point of view of Carian ono-mastics by using the definitive system, remained obscure in Ray’s system.



197

The first eff ort to improve Ray’s dechipherment was made byFaucounau (Faucounau 1984). But the typical dilettantism of this scholar

converts his work into a practically worthless attempt. The only inter-esting suggestion was l = l (vs. Ray †d ) later confirmed more definitively.

Much more serious is the work of Frank K , who devotedhis doctoral dissertation of 1990 to Carian (published with a Nachbemerkung in 1993: Kammerzell 1993). However, his decipherment diff ers too lit-tle from Ray’s: any diff erences—not in transcription, but in interpre-tation—are found only in the vocalic system. The consonantal valuesfor letters are coincident with those already proposed by J. D. Ray.

3. The De fi nitive Decipherment (‘Ray-Schürr-Adiego System’)

In this history of the decipherment of the Carian alphabet, I mustspeak of my own contributions to the subject. In my first work, pro-duced in 1989 and published one year later (Adiego 1990), I concen-trated on only two letters, l and n. For the first, new evidence waspresented in favour of an l value, following the suggestion made byFaucounau. As for n, I tried to demonstrate that this sign was actu-

ally the Carian letter for n, a sound surprisingly absent in Ray’s sys-tem (see above). In this later case, the evidence was exclusively basedon onomastic identifications that seemed convincing (for instance p-n-u≤ ol = pnu≤ ol = Ponussvllow ). Shortly after writing this brief article,I decided to devote my doctoral dissertation to Carian, with the aimof off ering a complete analysis of all Carian materials available (Adiego1990b). For the decipherment, the most important contribution was touse as a further clue two Carian-Egyptian bilingual inscriptions disre-garded by Ray because they contained true Egyptian names in theEgyptian part. A more attentive study of these two inscriptions (E.Sa2, E.Me 5) showed that the Carian parts did indeed reflect the sameEgyptian names present in the Egyptian ones. This lead to new val-ues for very important letters: n, whose value n was now confirmed,t, which reflected a t sound (transcribed t in Adiego 1990a, t alreadyin Adiego 1993a), and d, a d sound (d in Adiego 1990a, d already inAdiego 1993a):



198

E.Me 5 E.Sa 2

Carian form psmfkvneitz pdnejtEgyptian form Psm“ k-‘wj-Njt P3-dj-Njt Ray p-s-m-“-k-ü-k‘-j-e-q-≤ p-g-k‘-j-e-qAdiego (1993a) p-s-m-“-k-ú-n-e-i-t-≤ p-d-n-e-í-t

[now: psm“ kwneit ≤ ] [now: pdnejt ]

The t value of t allowed me to analyse the Athens bilingual inscrip-tion in a both straightforward and convincing way: read ≤ ías : san tur [ (now transcribed ≤ jas : san tur [ ), the third last letters could be connected

to the Carian name present in the Greek part, Tur[.Another important improvement concerned the vocalic letters: i =i (not †e ), j = í (given its alternation with the former, not †ê [now j ]),e = e (not † j ) and sound values close to /u/ for the letters w = w (Ray † j u, but on a weak basis; now [ ÿ ]), W = ù (Ray †é [now y ]). Allthese values are now universally accepted.

As for other letters, a labial character was proposed for $ (m [nowb ]), based on the good correspondence ( para)i $re l = Imbarhldow, anda liquid value for 6.

A brief report on this new decipherment system was published inKadmos (Adiego 1992a).

An error of decipherment, regarding q, was rectified shortly afterthe completion of the doctoral dissertation, in which a more traditionalvalue t was envisaged (which explains the use of t for transcribing t ).However, a more detailed analysis of E.Sa 2 brought me to the con-clusion that the second name in the Egyptian part, K3rr , was alsoidentifiable in the Carian part, qwri-, so that a tectal value for q

was proposed. This value was also reinforced by convincing onomas-tic identifications, such as quq = Gugow, or qtblem = Kotbelhmow, Kutbelhmiw.The new value was presented in a lecture held in Cologne in 1992and incorporated into my book Studia Carica (Adiego 1993a), whereina definitive decipherment of Carian was proposed.



199

Adiego (1993a)



200

As for the linguistic results obtainable by means of this decipher-ment, a very prudent method was adopted: both in Adiego (1990a) and

in Adiego (1993a) the attention was focused exclusively on the mostevident grammatical information recognizable in the Carian texts: theexistence of an opposition Nominative -Ø / Genitive (or ‘Possessive’)-≤ . The clear connection with Hittite, Luwian, Lycian, Sidetic, andPisidian sigmatic genitives and possessives, and the Lycian, Sidetic, andPisidan zero nominatives was seen as a clue for the consideration Carianas an Indo-European Anatolian language, more closely related to theselater dialects. This assertion was fully consistent with the indices observ-able in Carian onomastics, as ”evoro“kin and others had already showed.

During these decisive years for the elaboration of my own deci-pherment system (1989–1992), I was unaware that another scholar,Diether SÜ, had already obtained to the same results for the soundvalues of t, d, q, n, i and j. As early as 1982, also inspired by theworks of Ray, Schürr had observed the significance of the bilingualsE.Sa 2 and E.Me 6 for deciphering Carian, which explains the coin-cidence of the values proposed for these letters. However, Schürr merelycommunicated his theory privately to Ray, who rejected it, and after

this disappointing answer he became silent until the appearance of myfirst works. This explains why his first contribution appeared after Adiego(1992a) (Schürr 1992). The positive side was that the two studies hadreached very similar conclusions independently of each other, whichserved to reinforce the validity of the decipherment.

However, Schürr’s decipherment was not totally coincident with myown: the main diff erence lay in the values attributed to vocalic signs.While he agreed with i = i , and j = í [now j ], he was very scepti-cal about e = e, or about the u-like values of w and W. In any case,Schürr has since come to accept these values. He has been also scep-tical about other aspects of my decipherment, such as my interpreta-tion of the Athenes bilingual, or the identification of the name ÑUllãrimain C.Hy 1 ( ylarmi t )—an idea formulated by Ray (see above) and eas-ily integrated into my decipherment system.

Despite these discrepancies, Schürr’s proposals opened up new inter-pretations of some aspects I had disregarded. The most relevant wasundoubtedly his brilliant explanation of the Sinuri bilingual inscription

(C.Si 2), in which he was able to recognize the Carian version of theformula ‘Idrieus (son) of Hekatomnos and Ada (daughter) of Hekatomnos’that appears in Greek, in some decrees enacted by this pair of Cariansatraps. Schürr’s explanation of the letters & and $, which he con-



201

vincingly interpreted as representing the outcome of Proto-Anatolian(PA) *nd and *mb respectively, is also relevant.

In 1993, an international congress on Carian decipherment tookplace in Rome. In this congress, I presented a long list of onomasticidentifications that could be obtained with the new system (the “Ray-Schürr-Adiego”).8 The new decipherment clearly won the support (ex-pressed either in public or privately) of many relevant scholars, likeGünter Neumann and Heiner Eichner, but both ”evoro“kin and Gusmanirejected it: while the former continued to argue in favour of his ownsystem, Gusmani declared his scepticism for a decipherment that impliedextremely surprising sound values for letters of Greek shape (the so-called ‘metakharakterismós’).

I must stress here the role played by Günter N in backing the new decipherment. Although his public position was very cautious(for instance at the Rome congress), he encouraged it with meaningfulgestures, for instance writing the preface for Adiego (1993a) or off ering support and suggesting ideas in private correspondence. In Neumann(1993) he expressed his conviction that the ‘Egyptian approach’ of Rayand myself was a step in the right direction.9 In the same paper, he

suggested linking Carian sb, very likely to be a coordinative conjunc-tion, to Milyan sebe ‘and’. It is also worth mentioning the clear sup-port for the decipherment manifested by Heiner E in the RomeSymposium (see Eichner 1994).

Even more decisive in gaining wide acceptance of the new deci-pherment was H. Craig M’s contribution. In an important arti-cle published in Kadmos (Melchert 1993), Melchert not only acceptedthe new system, but also employed it to interpret Carian texts, open-ing new and thought-provoking ways of deciphering the language. Of particular note was his proposal of connecting Carian ÿbt (in C.xx 1)to Lycian ubete ‘he off ered’). As in the case of Neumann’s sb = Milyansebe , a word of common lexicon was recognized, and the genericconnection of Carian with Lycian and in general Anatolian was thus

8 Diff erent ordering of names has been given by scholars to refer to the system of decipherment that has become standard (alphabetical: Adiego-Ray-Schürr, strictly chrono-

logical regarding the published works: Ray-Adiego-Schürr). The ordering adopted hereresponds to the authentic chronological succession of the research, even if Schürr’s firstcontribution appeared after my first works.

9 ‘. . . aber nun endlich scheint doch ein erfolgreicher und nahezu vollständigerAbschluß der Entziff erung in Reichweite zu kommen’ (Neumann 1993:296).

202

reinforced. Also important was the analysis of snn and orkn as accusativesin -n (< PA *-n < PIE *-m ), an analysis equally relevant for the linguistic

position of Carian, and later confirmed by the Kaunos bilingual.Another scholar to take the Ray-Schürr-Adiego system as a starting

point for research on Carian was Michael J , who in his contri-bution to Rome Symposium off ered a series of new ideas, among themthe fascinating hypothesis (developed here) regarding the presence of ethnic names in the onomastic formulae of Saqqâra ( J 1994).Finally, Ivo H was another of the scholars quick to take up thenew decipherment, with excellent results: he devoted a paper to theCarian particle

∞i , analysing it as originally a relative pronoun from

PIE *k w i- (Hajnal 1997a). Moreover, in an independent way, Hajnaland Schürr (1996[98]) identified two further Carian words: ted ‘father’and en ‘mother’, clearly related to the corresponding Lycian and Lydiannouns.

It is not easy to briefl y sum up the amount of new ideas and hypothe-ses regarding Carian—some confirmed, others ruled out- that arose inthe first half of the 90s: apart from those quoted above, other partic-ularly sound theories were: the sound value z (< *st ) proposed in Schürr

(1996a) for the letter 1; the identification of the name of the AnatolianStorm God under the forms trqud e, trq d (Blümel-Adiego 1993) and of the name Hekatomnos in the unedited graffiti from Thebes (Adiego1995); the reconstruction of the Carian name of Kaunos as *kbid-(Adiego 1995:21), definitively confirmed a year later by the discoveryof the Kaunos bilingual; the attribution to Kaunos of the coins withthe Carian legend kb, interpreted as the beginning of the place name,discovered by K and also confirmed by the bilingual (Konuk1998);10 the satem-treatment of PIE palatals in Carian, as in Luwianand Lycian (Melchert 1993); the identification of the Carian family of names in ( i)b r- = Gr. Imb(a)r- made by Schürr (1991–1993), etc.

During the excavations of summer 1996 in Kaunos, a new bilingualinscription in Greek and Carian was found. The behaviour of the twoscholars responsible for the editing of the inscription, Peter FREI andChristian M , was exemplary: they edited the text as quickly aspossible, but also extremely accurately (Frei-Marek 1997), and suggested

10 Although this paper appeared after the publication of the Kaunos bilingual, it isclear that Konuk’s hypothesis was formulated previously (see Konuk 1998:218, n. 14,and 223 n. 51).



203

holding an international conference on the subject in the same year astheir article appeared, to which the principal scholars researching Carian

were invited (1997, in Feusisberg, Switzerland).The new inscription was the outside evidence that was needed to

remove any doubt about the validity of the Ray-Schürr-Adiego system.The presence of the place name Kaunos, and of the onomastic formulaeof two Athenian citizens in both parts of the inscription led to valuessuch as k = k , b = b, d = d , i = i , l = l , W = y, r = r , n = n, s

= s, a = a fully coincident with the values established in Adiego (1993a).The only anomaly, T = t (not = “ according to the Egyptian approach)is merely apparent, and easy to explain: there is no t (= t ) in theKaunian alphabet, so that in its writing system, a letter in the form of T represents the value t , whereas for “ a clearly diacritical form of thecommon Carian form for “ (/ ) was used in order to distinguish it fromT. This particularity aff ects only the Kaunian alphabet: in the otheralphabetic varieties, the value “ for f continues to be valid.

Shortly before the Feusisberg meeting, during the summer campaignof 1997, a new fragment of the inscription was found, which enabledthe beginning of the Greek part to be completed, showing that Frei-

Marek’s integration of the lost endings of the line, based on the Carian(!)part, was correct. Therefore, the complete edition of the inscriptioncan be found now in Frei-Marek 1998. The following table shows theGreek-Carian correspondences obtainable from the bilingual inscriptionof Kaunos:

Greek Carian Transcription Lycian evidence

Kaun[ ¤ ]oiw, k5idn, kbidn, Cf. Lyc. Xbide- ‘Kaunos’,

Kaun¤v[ n k5dWn/ kbdyn“ xbideñne/i- ‘Kaunian’Ipposy° nouw i[—]ini i [—]ini Nikokl °a ni[—]lan ni [—]lanLusikl °ouw lWsiklas[ lysiklas [ Lusikl °a lWs[—]an lys [—]anLusikrãt[ ouw ] lWsikraTas [ lysikratas [ ÉAyhna›o[ n ] oTonosn otonosn[ ÉA ]yhna›on oTonosn otonosn

Unfortunately, the bilingual inscription of Kaunos, so useful for confirming

the decipherment of the writing, has proved to be less useful in deci-phering the Carian language, given the incompleteness of the Greekpart and the possible non-literal correspondence between the two texts:

204

perhaps the most notable new discoveries are: (1) the word otr-“ , whoseprobable meaning is ‘self ’ and is possibly etymologically connected to

the Lycian atra/etli ‘self ’, (2) the ethnic suffix -yn- in kbdyn-“ ‘Kaunians’,clearly comparable to Luw. -wanni-, Mil. -wñni-, Lyc. -ñni- (with thesame function in all three languages), (3) the accusative plural ending -“ , related to Lyc. and Mil. -s < *-ns . To these discoveries we mustadd the confirmation of -n as an accusative singular ending, and of sbas a coordinative conjunction. A plausible interpretation of the remain-ing words and formants has yet to be made.



CHAPTER FIVE

THE CARIAN ALPHABET

A. A V

One of the areas of Carian studies in which our knowledge has dra-matically improved in recent years is undoubtedly the writing system.Thanks to the definitive decipherment and the appearance of newinscriptions, we can now create an overview of the diff erent alphabeticvarieties that, although still limited and with very important gaps, wasunthinkable some years ago. If we compare the table of the Carianalphabet established by Masson in the 1970s with the results that willbe uncovered here, it will be easy to acknowledge the progress madein this field. Masson was able to off er a complete inventory only forthe alphabetic varieties of Egypt and Kaunos; we now have an almost

complete inventory for the alphabets used in two other two Cariancities (Hyllarima and Mylasa). Moreover, the combination of new infor-mation obtained both from the decipherment and from these newinscriptions gives us a much better understanding of the relationshipbetween the diff erent alphabetic varieties. Certain signs that Massonhad deemed independent are now interpreted as variants, in some caseswithin an alphabetic variety, in others belonging to diff erent alphabeticinventories but functionally equivalent. Finally, the total of 44 diff erentletters in the table drawn up by Masson, the result of combining the

alphabets of Egypt and Caria could give the false impression that theCarian alphabet consisted of a surprisingly high number of diff erentletters. In fact, as we will see, the number of diff erent letters found ineach alphabetic variety barely exceeds 30 (31 in Memphis is currentlythe highest number of letters documented for a concrete alphabet).

Many questions remain unresolved, however. The new inscriptionsfrom Caria complicate to a certain extent former classifications of thediff erent alphabets of Caria proper, based on a geographical criterion

that now seems to be too rigid. Our current knowledge also makes itdifficult to establish the exact nature of the relationship between theCarian alphabet of Egypt, a very homogeneous sign inventory, and thealphabets of Caria proper, which are much more varied between



206

themselves, and are also diff erent from the Egyptian alphabet. Finally,the most intriguing question of all is yet to receive a wholly satisfac-

tory answer: what is the origin of the Carian writing system?In the following pages, we will see firstly the known inventory for

each Carian alphabetic variety, together with a succinct commentaryon the main characteristics in each case. In the subsequent sections of this chapter, the open questions mentioned above will be briefl y dealtwith.

1. Alphabetic Varieties of Caria Proper

Alphabet of Tralleis We do not yet possess sufficient Trallean materialsto reconstruct the alphabet used in this Carian-Lydian city: C.Tr 1off ers 12 diff erent letters, C.Tr 2, 12 or perhaps 13.1 Given the brevityof the texts and the coincidence of some words, the combination of both inscriptions produces an inventory of only 13–14 diff erent letters.This means that in all probability less than the half of the completeinventory of letters is attested. However, despite the very limited num-ber of letters, some interesting traits can be observed. Perhaps moststriking are the unique characteristics of C.Tr 1: it shows a right-to-left orientation, totally atypical for Carian inscriptions from Caria proper.Moreover, the letters for a and r seem to be A and A respectively.This is a clear contrast with the corresponding forms in C.Tr 2: a

and R. The form of a A is the same as that found in the Sinuri-Kildaravariety, and shows the particularity of a dextrorse orientation in a sin-istrorse context. As for A, at a glance it seems closer to Egyptian andKaunian r (r ) than to the more extended shape of the letter in the

alphabets of Caria proper (R ). In any case, it is very likely that thedefinitive form of both letters is the result of internal tendencies to sym-metry (A | A ): such tendencies to structural uniformities are well knownin the history of the writing system. Regarding the rest of the lettersfrom C.Tr 1, particularly remarkable is the form adopted by the signfor i , a right-to-left version ( Y ) of the i found in Hyllarima (y ). Alsoin this case, C.Tr 1 diverges—though to a lesser degree—from C.Tr2, where the corresponding letter is Y. Both y and Y point to an iden-tical subtype of Carian i characteristic of the Northern zone (see above),

1 The real sign that appears at the very end of C.Tr 2 is much disputed: it couldbe a Q (as assumed here) or simply a z. See p. 131 and here below.



207

but it is rather surprising that the form found in C.Tr 1 is in factcloser to that of the Hyllarimean alphabet than to its equivalent in the

other Trallean inscription.Compared to C.Tr 1, C.Tr 2 appears more as we would expect: a

for a , R for r . It is worth noting the presence of %, a letter absentfrom the alphabet of Memphis. But the most uncertain point of thisinscription is the last letter, as mentioned in p. 131: the drawings areambiguous, pointing to either Z or q. The first of these forms wouldsuggest a z ≤ of Kaunian type, but the second drawing, and the factthat in the same inscription ≤ appears as z, make it more plausible inmy opinion to interpret the letter as a q , comparable to the form thatthis sign adopts in the most local alphabets of Caria proper (Q ).

Alphabet of Hyllarima The new fragment of the sole remaining inscrip-tion of Hyllarima, added to the part of the text already known, hasallowed us to create an almost complete inventory of the letters of thealphabetic variant used in that city. They are shown below:

Alphabet of Hyllarima

a a U uD d ñ ñl l N nW y p pR r z ≤L l y iQ q e eB b K km m & d

o o 9 zt t c t S “ B bs s

The most remarkable absence is the letter for ∞ (x X in other alpha-bets). This absence is most probably a matter of chance. It is also pos-sible that other letters, less common but nonetheless found in otheralphabets of Caria proper (such as %, H ) could have belonged to theinventory of Hyllarima. In any case, these 25 letters constitute an inven-tory large enough to allow us to consider our knowledge of Hyllarima’salphabet to be fairly comprehensive.

208

While most of the letters used in Hyllarima are clearly comparableto the other Carian alphabets, the new fragment sees the appearance

of three somewhat surprising new forms of letters: L l , e e , and B b .In the case of L and B, it seems clear that we are dealing with graph-ical innovations produced within Hyllarima’s alphabet. Both letters haveundoubtedly been created from Greek letters: L is a diacritised formof B that only can be explained by a Greek lambda (B is b in Hyllarima).2

B is simply a Greek beta , used for the Carian sound represented in otheralphabets by Ø 4 (< *mb ). e, however, could be an old variantof e e assuming a common origin from Greek archaic eta E, and con-sidering its probable presence in the Carian alphabets of Euromos andMylasa (see below).

The reasons for a recent introduction of L and B are not clear. Inthe first case, it is worth mentioning the fact that some alphabetic vari-ants (Thebes, Mylasa) lack a specific letter for l , this sound being cov-ered by the sign for the ‘basic’ lateral liquid l , l. Perhaps this wasalso the original situation in Hyllarima, so a specific letter may havebeen created when the local alphabet already existed. Regarding B, thesame explanation could be envisaged, although the alternative suggested

in Adiego (2005) must not be discounted: that the original form of Ø4 was &, i.e. the form attested in Mylasa, and that the similar-ity of & to & could mean discarding the first letter in favour of a newsign directly borrowed from the Greek alphabet, B.3

Alphabet of Euromos Our knowledge of the alphabet used in the Cariancity of Euromos is more limited than that of Hyllarima. The two brief inscriptions that remain give us an inventory of 20 letters:

Alphabet of Euromos

a a s sd d U ul l x ∞W y n N nR (/ r?)4 r z ≤Q q í I i

2 However, we must not rule out an alternative origin from Carian l l .3 For a detailed discussion on these new letters, see Adiego-Debord-Varinlio[ lu (2005).4 On the doubts about the presence of the variant r for r in C.Eu 1, see p. 133.

(continued on next page )



209

B b e em m k K ko o & dt t 9 z

The absence of the signs for two very important and widespread Cariansounds ( p and “ ) is without doubt a coincidence. It is also likely that

the alphabet of Euromos had specific letters for ñ , t and b, as inHyllarima. It is particularly unfortunate that we are unable to ascer-tain whether there was a specific letter for l , or if this sound wasassumed by l l , as is the case in Thebes and Mylasa.5 The letter e,formerly interpreted as a sign for l similar to the Kaunos form of theletter (rotated 2 against L in other alphabets), must now be considereda sign for e : Kaunian 2 is a peculiarity that is hard to separate fromthe similarly rotated form Z of z ≤ , a letter that appears in Euromosin the more widespread form z. e as e in Euromos is more consistent

with the use of the sign with this value in the nearby cities of Mylasaand Hyllarima.6

Alphabet of Mylasa The newly published single inscription of Mylasa,of significant length, off ers us a very wide inventory of letters, whichundoubtedly marks a clear improvement in our knowledge of the localalphabetic varieties of Caria.

Alphabet of Mylasaa a U uD d x ∞l l N nW y p pR r z ≤Q q I iB b e e

5 The two sole examples of l (in C.Eu 2) appear in contexts without the parallelsthat would allow us to decide if they represent l or l .

6 Schürr (pers. comm.) has independently come to the same conclusion.

table (cont .)

Alphabet of Euromos




210

m m k ko o & dt t 9 zF “ & b (?)s s

The presently known inventory consists of 23 letters. None of the most

widespread Carian signs are missing here (on l , see immediately below).If we compare this inscription with that of Hyllarima, for example, theonly notable absences are c and ñ, neither of which is a particularlycommon Carian letter.

The most notable traits of the alphabet of Mylasa are the following:

• There is not a specific sign for l , the sound involved being repre-sented by means of l l : note idu≤ ol ≤ besides dw ≤ ol ≤ (E.Me 35).

• Given the use of l for l in idu≤ ol ≤ , the letter e cannot represent this

sound here, and it is more likely that we are dealing rather with aletter for e , as in Hyllarima and perhaps also Euromos. Unfortunately,the only example of this letter in E.My 1 appears in an unclear con-text.

• For y, a letter similar to that used in Sinuri and Kildara is employed:W (Sinuri-Kildara V ).

• Also very close to Sinuri and Kildara alphabetic variety is the formadopted by i in Mylasa: I (cf. Sinuri-Kildara Î ).

• It is interesting to note the angular form of some letters: besides thesign for i ( I ), the same trait is observed in 9 z (9, 1 in otheralphabets), and in & if we accept the interpretation of this lattersign proposed here (see the following point).

• The most troublesome sign in this alphabet is &. I have proposedto interpret it as the local form of the letter for b, which in otheralphabets is Ø, 4 (Adiego 2005).

From these observations, the alphabet of Mylasa turns out to be of a

very idiosyncratic nature; while it shows some clear affinities with theSinuri-Kildara variety, as could be expected (W = y, I = i ), the useof e for e (against Sinuri e ) or the lack of a specific sign for l (which

table (cont .)

Alphabet of Mylasa



211

does exist in Sinuri-Kildara) prevent us from speaking of a commonlocal alphabet for the region of Mylasa, Sinuri and Kildara.

Alphabet of Stratonikeia The two inscriptions from Stratonikeia give usan inventory of 25 letters:

Alphabet of Stratonikeia

a a U ud d ñ ñl l X ∞

W y N nR r p pL l z ≤Q q Ï im m e eo o K kt t & dF “ 9 zs s b (?) bH ?

It is somewhat strange, but perhaps merely a matter of chance, thatthere is no letter for the widespread sound b. Leaving aside this oneuncertainty, the inventory of Stratonikeia, with a number of diff erentletters identical to that Hyllarima, seems almost complete. When com-pared with Hyllarima, the only letter that might be missing is c,7 sincethe other letters that are not present here are rather idiosyncratic forsome local alphabets or for the alphabet attested in Egypt.

I off er the new suggestion of recognizing in C.St 1 the Stratonikeianletter for b, a hypothesis based on the identification of the name b rsi in the second line (see p. 143). The problem arises from the fact thatit is not easy to establish the exact form of the letter: the photographin Robert (1950) is not absolutely clear, and the drawing published byDeroy (1955) could not be totally faithful. Given that the question can-not be satisfactorily resolved, I use conventionally a form based onDeroy’s drawing (b ). In any case, this letter seems to be formally relatedto the Egyptian (Ø ) and Kaunian (4 ) signs for b.

7 About the possibility that in other alphabets of Caria proper the undeciphered let-ter H could in fact represent c t, see p. 253.



212

Apart from this new letter, the alphabet of Stratonikeia does notoff er other remarkable traits, with the exception of the form of i , Ï,

until now exclusively documented in this alphabet, and which showscertain similarities to corresponding signs in Mylasa and Sinuri-Kildara.Also comparable with Mylasa and Sinuri- Kildara is the form of F “ .Regarding the rest of the letters, we find the typical forms Q, 9 of Caria proper (against q, 1 respectively, in Egypt and Kaunos). In anycase, it is worth noting the form of ∞, X, against x in most of theCarian alphabets from Caria proper (including Kaunos).8

Alphabet of Sinuri-Kildara The combination of the letters documented inthe inscriptions of Sinuri and Kildara gives a total of 26 diff erent signs:

Alphabet of Sinuri-Kildara

Sinuri Kildara

A A ad d dl l

V V yR R rL L l

q (??) Q qB b B bm m M mo o ot t t

T F “s s s

H (?) ?

U u uñ ñx x ∞N n np pz ≤Î Î ie e

k k& & d

9z% ã?

8 The form X only appears in the inscription C.Ia 3 (but note also x in Iasos, C.Ia2) and in the inscription C.xx 2 (of unknown origin).



213

An attentive examination of the inventories of letters used in each loca-tion reveals the close affinity between the alphabets used in Sinuri and

Kildara, which allows us to speak of a single alphabetic variety. Noteparticularly the forms of a (A ), y (V ), i (Î ). The new inscription fromMylasa has served to reinforce this theory, because several letters, whichformerly seemed unimportant given their presence in other varieties,have become more representative insofar as they are absent, or diff erent,in the inventory of Mylasa. This is the case of e e (against Mylasa e )and L (absent in Mylasa).

What is alleged to be the most remarkable diff erence between Sinuriand Kildara—the use of q in Sinuri against Q in Kildara—perhapsdoes not exist: the reading of q in C.Si 1 is far from assured (seep. 138). Note that other divergences occur within the alphabet of oneor both places: b and B in Sinuri, m and M in Kildara (but about thedoubts regarding the reading of this latter sign, see p. 142).

We can suppose that this inventory of Sinuri-Kildara is practicallycomplete. If compared with the inventories of Hyllarima, Mylasa, andStratonikeia, only the letters for b and t are missing.9 Conversely, it isinteresting to note the presence of % (as in Tralleis and Kaunos).

Alphabet of Kaunos The alphabetic variety of Kaunos is the best knownof all the varieties attested in Caria proper, thanks for the most partto the long inscription C.Ka 2, and the bilingual text C.Ka 5. Theinventory of signs, which we can consider complete,10 consists of 29letters.11

Alphabet of Kaunos

a a ñ ñd d x ∞l l N nW y p pr r Z ≤2 l I iq q k k

9 This latter sound, however, could be represented by H, see the similar observa-tion on Stratonikeia above.

10 Obviously, it is not impossible that some letter has still not been documented,but in this case its functional value would be very low.

11 I assume that ÿ is the form of C.Ka 5 corresponding to 4 in C.Ka 4.




214

5 b & dM m 8 g ? (if = 0 )o o 1 zT t % ã? _/ “ 1 ?s s O t (if = c )H ? 4, ÿ bU u

The main characteristics of this alphabet, some of which have alreadybeen mentioned, are the following:

• Leaving aside the particular cases of Iasos and Halikarnassos (seebelow), this is the alphabet that shows the most resemblances to theCarian alphabet of Egypt: as the table seen above makes clear, inthis alphabet the forms for r, q and z are identical to those found

in Egyptian inscriptions (r,

q,

1against

R,

Q,

9in other alphabetsof Caria proper, respectively).

• However, despite these formal analogies with the Egyptian alpha-betic model, from a functional point of view, Kaunos behaves as atypically Carian local alphabet: unlike the Egyptian model, there areno specific letters for semivowels (j j , v w , perhaps also w ÿ ), andit also lacks 6 ® . Conversely, it does contain letters such as ñ ñ or% ã? , both of which are absent in the Memphis Alphabet.

• As for the specific traits that characterise Kaunos, the following must

be mentioned:1. A letter for e is missing. The reasons for this absence seem to be

rather of phonological nature (the Kaunian vowel system wouldlack a sound /e/).

2. Some letters have only been documented in Kaunos: 8, O, 1_.It is very likely that some, if not all of these letters have the samevalue as other letters with a diff erent shape in the rest of thealphabets: 8 seems to be the Kaunian form of 0, and for O Ihave proposed an identification with c

t(Adiego 2002; see here

p. 253). But even if these correspondences were correct, theseKaunian letters would remain unique.

table (cont .)

Alphabet of Kaunos



215

3. The letters for l and ≤ show a rotation (2, Z ) regarding the restof alphabetic varieties (L, z ).12

Other Less Documented Alphabets (Kindye, Eski Çine, Keramos) Little can besaid about the alphabet varieties of other Carian locations due to thescarcity of documents.

In Kindye, the most remarkable feature of the single, six-letter inscrip-tion is the sign w, unknown in the other alphabets of Caria proper.Merely hypothetically, we could consider the possibility that this is theform adopted by V y in this alphabetic variety. Note that in Kindye

the ‘Egyptian-like’ letter w exists side by side with R, the typical formfor r in Caria proper (against Egypt and Kaunos r ).In the case of the Eski Çine (near Alabanda) inscription (C.Al 1),

note the shape of the sign for i , apparently ì,13 which recalls the typ-ical “northern” forms (Tralleis Y, Hyllarima y ).

The Problem of the Iasos and Halikarnassos Alphabets The alphabets stud-ied to date share the characteristic that they are attested in inscriptionsfor which an origin from the location where they are found is the sim-

plest and most logical explanation. In the case of Hyllarima, Kaunos,or Sinuri-Kildara, for example, there is no doubt that each alphabeticinventory reflects the local writing used in each place; in Kaunos,Kildara and Hyllarima, evidence is given by the mention of the placename (ki l ara, kbidº ‘Kaunos’) or the ethnic name ( ylarmi t ‘Hyllarimean(s)’,cf. also kbdyn“ ‘Kaunians’). In Sinuri, the Carian text of C.Si 2 is insep-arable from the preceding Greek text, and the two together fit well inthe context of inscriptions concerning the syngeneia in charge of the

sanctuary of Sinuri. The funerary inscriptions of Tralleis and Euromos,and the long inscription of Mylasa also seem to be closely related tothe places where they were found, and the characteristics of the respec-tive alphabetic varieties are consistent with the geographical situation.

The case presented by the inscriptions from Halikarnassos and Iasosis very diff erent.

12 It is true that z can occasionally appear as Z in some Egyptian inscriptions (see

for instance E.Me 14), but this occurs as a result of the disposition of the text on thestela. The particularity of the Kaunos alphabet is that this rotated form is the canon-ical form of the letter.

13 At least this is the form that can be hypothesized both from the photograph inRobert and the drawing in Robert and Deroy, see p. 132.



216

For Halikarnassos, there is only one inscription, found on an objectthat could be moved very easily, but even its origin in Halikarnassos

cannot be assured. Here is the inventory of letters documented in C.Ha1 (note that the inscription is written right to left):

Alphabet of C.Ha 1

# rL l ª bM m

o ot ts sN np pw ÿK k& d

The possibility that this alphabet could be that used in Halikarnassoshas important repercussions for the discussion of the origin of thediff erent Carian local scripts, because although the inventory of lettersthat can be obtained is very limited, it is sufficient to draw some clearparallels with the Carian alphabet attested in Egypt: w ÿ, # r , or theright to left orientation, a very unusual practice in the alphabets of Caria proper but well known in Egypt. This would make the regionof Halikarnassos the most likely origin of the Carian alphabet fromEgypt.

The situation is similar, but more complex, in Iasos. Here, six of theseven inscriptions documented until now appear on fragments of vases,so they may be of non-Iasian origin. The sole inscription found on astone fragment, C.Ia 4, consists of only five letters, the only remark-able trait being the form R (‘not-Egyptian’) of r (against Egypt-Kaunosr ). Of the remaining inscriptions, the longest is C.Ia 3, where we find,rather surprisingly, r and R side by side. Also of note is the ‘Egyptian’or ‘Kaunian’ form for q , q (against Q in the rest of Caria, including another inscription from Iasos, C.Ia 5). Another ‘Egyptian’ trait couldbe recognized in the brief fragment C.Ia 6, where it would possible toidentify the letter Ø b, not attested in this form in the rest of the alpha-bets of Caria proper. However, this reading is not certain (see p. 149).



217

Similarly puzzling is the letter v in C.Ia 7 (= Mylasa W, Sinuri-KildaraV?).

The following table synthesizes the letters documented in the corpusof Iasian inscriptions:

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

~ ~ a ad

l l lW v(?) y

r R R rL L l

q Q qB B b

m m mo o o ot t t

ß F D “s s s

U u ux X ∞

ë ë ë np p

z ≤i i I z i

e e e ek k k& d

0 g ?Ø? b

Supposing that all the inscriptions found in Iasos reflect the local writ-ing, we obtain a total of 24/25 diff erent letters, a number very simi-lar to the almost complete inventories of signs documented in otherCarian cities. Note the presence of 0, a typical ‘Egyptian’ letter, andthe absence of H, c, ñ, %, and of a sign for z (1, 9, 9 in other alpha-bets). Such absences could be due to chance. None of the specifically“Egyptian” letters j, v, w appears (but note v!).

The Alphabets of the Inscriptions of Unknown (but Presumably Carian) OriginIt is particularly regrettable that we are ignorant of the origins of the



218

two inscriptions on vases, C.xx 1 and C.xx 2, since they show the clos-est affinities with the Carian alphabet of Egypt: in both inscriptions the

letter j j is found; moreover, in C.xx.2 the letter w appears, also typ-ically Egyptian; signs for r, q, i, z are of the Egyptian (and Kaunian)type: r, q, I, 1. In fact, both inscriptions could fit well in the Egyptiancorpus, because all the letters have a similar form to the Carian alpha-bet of Egypt, and there is no example of a letter alien to that writing variety. Here is the inventory of letters from both inscriptions:

C.xx 1 C.xx 2

a a£ dÀ l l# + rq qª bm m mo o ot t tD “s s su u

X ∞N n N np p pz ÷ ≤

i ie e ew w ÿk k k

& d1 z

j j j

C.xx 3 also shows a notable “Egyptian” trait: the coexistence of w andW in the same alphabetic inventory. As far as we are aware, in thealphabets of Caria proper there is only one letter for the sound /y/and its semivocalic form, /w/ (W, V, perhaps also w in Kindye, but the

evidence is lacking, see above). In any case, unfortunately the inscrip-tion C.xx 3 is hard to analyse, and the reading of some letters remainuncertain. This, together with the brevity of the text, prevents us using it as a valuable source on the Carian alphabet.



219

We must recall the possibility, suggested by Meier-Brügger (1994:113)that Cxx 1, C.xx 2 and C.xx 3, together with C.Ha 1, could come

from the same place (a sanctuary of the god ntro- ), see above p. 162.The affinities between the alphabets of all these inscriptions are con-sistent with this hypothesis, and a common origin from Halikarnassosis an attractive possibility.

Finally, C.xx 4 and C.xx 5, in fact two identical inscriptions con-taining only a single word, clearly belong to the Kaunian alphabet, aswas observed previously by the editors (Zalhaas-Neumann 1994) andhas been referred to here (see above p. 163). The close connection tothe Kaunian alphabet can be ascertained on the basis of the use of the letter Z (= Kaunos Z against z in the rest of the alphabets) and,above all, the letter / for “ : this letter only appears in the Kaunianalphabet, where its presence is closely related to the use of T for t (seeabove). But contrary to Kaunos, these inscriptions show the lack of aspecific letter for l , instead of which the letter l (l ) is used (as inMylasa and Thebes).

2. Inscriptions from Continental Greece

In the case of G 1, G 2 only G 1, the Athens bilingual inscription,shows some significant traits regarding the Carian alphabet: we notethe presence of the “Egyptian” letter j j , clearly used for the semi-vowel /j/ (≤ jas ). However, this “Egyptian” trait is counterbalanced bythe use of R (not ‘Egyptian’ r ) for r .

3. Egyptian Alphabets

The main characteristic of the Carian alphabets used in Egypt is theirhomogeneity, which contrasts with the pronounced local diff erences thatcan be seen in the alphabetic varieties found in Caria proper. Thishomogeneity has been demonstrated by the Saqqâra alphabet, whichoff ers us a point of reference for analysing the other Egyptian alpha-bets. It is true that the diff erent sub-corpora of graffiti can at initiallysuggest the contrary, because the forms of the letters show a high degreeof variety, but these diff erences are in reality very superficial, and can

be attributed to the less accurate writing in the graffiti. If we look atthe more pronounced diff erences for some letters in the varieties of Caria proper (for example q / Q for q , r / R for r , i / y / Y /Ï . . . f o r i ), we can observe that in all the documentation found in



220

Egypt only one form for each letter is found (q, r, i ), which impliesa common alphabetic model for the Carian of Egypt. Only in the case

of Theban graffiti can we envisage a sufficient number of remarkabletraits to speak of a specific variety, and even in this case the diff erencesare more functional (the absence of some letters whose function isfulfilled by another sign) than formal.

Alphabet of Memphis As already said, the corpus of the alphabet of Memphis (Saqqâra) shows a complete and clear inventory of letters:

Alphabet of Memphis1 a a 17 n N n2 d D d 18 p p3 l l 19 z Z ≤4 W ù y 20 I i5 r r 21 k K k6 L l 22 & d7 q q 23 0 g 8 b B b 24 1 z

9m M

m 25c

t10 o o 26 Ø b11 t t 27 e e12 f g “ 28 w ÿ13 s s 29 v Ú w14 H ? 30 j T _ j15 u U u 31 6 ®16 x X ∞

The following traits must be taken into account when making a com-parison with the best-known alphabets of Caria proper:

• The Memphis Alphabet, along with the other alphabets from Egypt,has specific letters for the semivowels /j/ /w/ and perhaps also /w/:j, v and w or W. In this latter case, it is not possible to clearlyestablish which of the two letters was originally used for the semi-vowel, and which for the vowel. In the Carian alphabets of Cariaitself, there is neither j nor v, and in the case of w/W only one of

the signs is used (V/ W, very probably an evolution of w in Sinuri-Kildara and Mylasa, W in the rest).• Also specific to Memphis, and to some other Egyptian alphabets, is

the use of the letter 6 for a special liquid sound (® ). No comparable



221

letter has been identified in any alphabet of Caria proper.14 Moreover,it is interesting to note that if the identification ar ® i “ = Arrissiw is

correct, this would mean that, at least in Hyllarima, where the nameArrissiw can be definitively identified under the form ari “ , there isnot any particular strategy for spelling a possible sound cluster -r ® -(Greek rr). A similar situation seems to exist in Stratonikeia, wherethe same name also appears (ari “ ).

• Typical shapes of certain letters separate the Egyptian alphabet frommost of the alphabets of Caria proper, as has been noted repeatedlyin the preceding pages: q q , r r , I i , 1 z (against Q, R, and diversevariants of i , respectively).

• None of the Carian alphabets of Caria proper display a simultane-ous use of H and c (the only exception would be Kaunos, and onlyif the very hypothetical identification O = c is accepted). Obviously,this does not necessarily mean that both letters must be related, butthe fact certainly merits consideration.

• Two letters documented in the alphabets of Caria proper are absentfrom the inventory of Memphis: % and ñ. The first of these lettersdoes appear in the inscription (of apparently Egyptian origin) E.xx

7 and probably also in Thebes (E.Th 34) and in the Luxor temple(E.Lu 5). In Caria, it is documented in Tralleis, Sinuri and Kaunos.More meaningful is the absence of ñ, until now not documented atall in Egypt, but very widespread in Caria proper (Hyllarima, Sinuri,Stratonikeia, Kaunos).

The Remaining Carian Inscriptions of Egypt Leaving aside the particularcase of Thebes and E.xx 7 (see immediately below), the alphabet usedin all the inscriptions from Egypt is consistent with the inventory of letters that makes up the alphabet of Memphis. Even the less commonlyattested letters in Memphis are present to some extent in the diff erentsub-corpora. This is the case, for instance, of 6 ® , present in Buhen(and also in Thebes), or Ø b, documented in Abydos and Buhen (andThebes). Allegedly ‘abnormal’ letters, such as B in Silsilis (E.Si 11), haveturned out to be the result of an incorrect reading (it must be read g

“ ). Perhaps the most significant variants found in the other Egyptianinscriptions are B for b (against Memphis b ( ), and G for d (against

14 However, it is methodologically advisable not to rule out the possibility that inKaunos the sign 1, still undeciphered, corresponds to 6 in Memphis.



222

Memphis d D ). Regarding the remaining letters, the diff erences are notso marked, and in fact seem to correspond to the more informal char-

acter of the graffiti or to varying degrees of skill of the engravers.

The Case of Thebes The alphabet of Thebes shows some singularitieswhen compared to the alphabetic model of Memphis. Unfortunately,the importance of these diff erences is hard to estimate without a com-plete epigraphical edition of the entire Theban corpus. I can only basemy observations on the partial knowledge gained from the drawings of the inscriptions, thus the conclusions that will be formulated here must

be taken as merely provisional.One singularity of the alphabet of Thebes is the absence of a letterfor l , a sound represented simply by l l , a feature also apparent inthe alphabet of Mylasa and in the two objects C.xx 4 and C.xx 5.This particularity is best attested by the clear example pnw ≤ ol , against pnu≤ ol , punw ≤ ol ≤ in Memphis.

Another trait already observed in this alphabet, namely the absenceof a letter for d (against & in the rest of the alphabets), and the useof l (and occasionally t )15 in its place, is nowadays less clear: in the cor-

pus of Theban inscriptions provisionally elaborated by Diether Schürrand followed here, the letter & does appear (in E.Th 4 and E.Th 14).Rather surprisingly, both inscriptions are not new, but were in factalready published in ”evoro“kin (1965), where the letters in questionwere read in a diff erent way. Due to a lack of further information, thequestion remains open. In any case, it must be stressed that the prob-lem posed by the presence or absence of & in Thebes transcends thetopic of the alphabetic varieties and aff ects the linguistic interpretation

of Theban inscriptions, because the identification of the Theban wordmlane with the form md ane depends on the hypothetical lack of & inThebes and its substitution by l l . Also, from a phonological point of view, this trait will be inseparable from the absence of a letter for l :in Thebes, l l would represent three diff erent but related sounds (l,l , d ) or the convergence of these three sounds in a single one.

In the unpublished inscriptions of Thebes, other rather surprising traits appear, but they must be considered with caution until a satis-factory edition of the whole corpus is edited. According to the drawings

15 This latter use would be visible in the name k ≤ at ÿbr if its equivalence with Janduberiw(attested in Lycia) were accepted (see Chapter 11, s. v.).



223

of ”evoro“kin, at least an example of % ã is attested (in E.Th 34),16

and two new letters, until now unknown in Carian, are present: %

and s. Both of these letters seem to appear in E.Th 31. The first alsoseems to be present in E.Th 34. However, in the case of E.Th 31,Schürr decides against giving a reading of the whole inscription, andin E.Th 34, he proposes an alternative reading o (see above p. 103–104).Therefore the existence of both letters seems to be highly doubtful. Asfor %, it must be pointed out that this letter can be easily mistakenfor a more standard letter x. Summing up, the alphabet of Thebescannot be used in any discussion of the Carian alphabet without the

definitive edition of all the Theban graffiti.E.xx 7, the inscription of a bronze lion, shows (together with a typ-ical Egyptian trait, the presence of v w ) the use of %, a ‘non-Egyptian’letter.

4. The Classi fi cation of the Alphabets of Caria Proper

We have seen that in contrast to the homogeneity of the alphabet usedin Egypt, a high degree of variation is found between the diff erent local

scripts of Caria proper.Former studies of the alphabets used in Caria proper have attempted

to classify them in groups on the basis of the marked diff erences inthe shape of some letters (see ”evoro“kin 1965, followed in Adiego1993a). These groups coincide roughly with the geographical situationof the local varieties, so that a Northern group (Tralleis, Hyllarima,Eski Çine [South of Alabanda]), a NW group (Euromos, Kindye), aCentral group (Stratonikeia), a Western group (Sinuri-Kildara), and a

SW group (Kaunos and the inscription of Krya) have been established.I now think that this classification should be revised, not least becauseour knowledge of the alphabets of Kindye and Eski Çine is so limitedthat we cannot possibly use them in a serious discussion of alphabeticvariants. Even in the case of Tralleis and Euromos, we cannot workwith inventories that are sufficiently extensive to give an idea of all thecharacteristics of these alphabets.

Moreover, the new inscription of Mylasa presents a strong argumentagainst such groupings; despite the clear affinities between the alpha-

bet of Mylasa and the Sinuri-Kildara variety, there are also significant

16 Note also its possible (but doubtful) presence in Luxor (E.Lu 5).



224

diff erences, some of which point rather to a connection with some“northern” and “north-western” alphabets (Hyllarima, Euromos).

In fact, if we leave aside Kindye from the north-western group andEski-Çine and Tralleis from the northern group, given the scarcity of the materials, and if we also claim that Mylasa did not have the samealphabet as the nearby sanctuary of Sinuri, we come to a surprising conclusion: with the sole exceptions of Sinuri and Kildara (two placeswith an identical alphabet), and Krya in Lycia (where a Kaunian alpha-bet was used), each of the other places where the Carian alphabet isattested seem to have had their own alphabet.

For this reason, I believe that a more suitable way of analysing theconnections between the diff erent local alphabets of Carian is to focuson the ‘isographs’, i.e. the identical or similar shapes of several lettersthat are shared by diverse alphabetic varieties.

The following table off ers a comparative overview of the most rep-resentative letters in the main alphabetic varieties of Caria. I leave outthe inscriptions of Kindye and Eski Çine for the reasons stated above,and the inscriptions of Iasos and Halikarnassos, given the particularproblems mentioned in the preceding pages. The cities are ordered

from North to South and from West to East:

Tralleis Hyllarima Euromos Mylasa Stratonikeia Sinuri- KaunosKildara

a a (A ) a a a a A a y ? W W W W V Wr R (/A!) R R (/r?) R R R rl ? L ? - L L 2q Q (?) Q Q Q Q Q q

b ? B B B ? b B 5t t t t t t t T“ ? S ? F F F /≤ z z z z z z Zi Y y I I Ï Î Ie ? e e e e e -z ? 9 9 9 9 9 1b ? B b (?) & (?) ? ? 4

The first, and most important, contrast can be established between theKaunos alphabet and the rest of the alphabetic varieties. As we havecommented previously, some (but not all!) of the particularities of the



225

Kaunian alphabet connect it to the Egyptian alphabet: q q , r r , 1 z .Other singular traits are specific to Kaunos: the rotated form of ≤ and

l , the use of T for t (and the subsequent creation of a specific letterfor “ ) and the form of b. To these we must also add the absence of a letter for e and the presence of letters not found in the rest of thevarieties (8, 1 ).

As for the remaining alphabets from Caria, there is clear evidenceof internal affinities that correspond to some extent to the geographi-cal situation of the cities. But, as we have already seen, with the excep-tion of Sinuri and Kildara, whose alphabets show sufficient similarities

to speak of a single alphabetic variety, we find diff erent isographs, notalways coincident, rather than real groupings of alphabets.In this sense, the diff erent forms of i are particularly striking; whereas

the “northern” Alphabets of Tralleis and Hyllarima show forms thatare clearly interrelated and diff erentiated from the rest (y, Y ), the formsin the “central” alphabets of the Mylasa-Stratonikeia-Sinuri/Kildara tri-angle point to a common original form.

The coinciding form for y in Mylasa and Sinuri-Kildara is also impor-tant; in this case, Stratonikeia does not share this characteristic, so we

can limit the extent of V/W to a more restricted area.The distribution of the forms for e off ers a somewhat diff erent per-

spective. If one accepts my interpretation of Euromos and Mylasa eas an e -sign, we can establish an area that includes these two alpha-bets and that of Hyllarima, in contrast with Sinuri (a letter for e is notdocumented in Kildara, probably a matter of chance) and Stratonikeia.

This distribution of e -signs can be connected with the signs for l .As in the case of e , Sinuri-Kildara and Stratonikeia coincide in the

form of the letter,L.

, while in the “e-zone” the situation is totallydiff erent: in Mylasa there is no specific sign for l , l l carrying out itsfunction, and in Hyllarima a special letter L is used, probably a newsign introduced more recently on the basis of Greek L or Carian l.Unfortunately, in the case of Euromos, the only two inscriptions tohave been found do not help us to establish whether this connectionbetween the signs for e and l also existed there.

The explanation for these ‘isographs’ is not clear: could they be theresult of a common origin for a group of alphabets, or rather the con-

sequence of a diff usion? Our knowledge of Carian writing is still tooincomplete to give a clear answer. The case of Sinuri-Kildara andMylasa is particularly puzzling; the corresponding use of a special formfor y and the form of i , both features that point to a clear common



226

origin, is counterbalanced by the treatment in Mylasa of e and l . Giving priority to certain signs to the detriment of others in order to explain

the relationship between both alphabetic variants, would be arbitrary.

5. The Relationship between the Alphabet from Egypt and the Local Alphabets from Caria Proper

An initial conclusion can be drawn from a comparison of the alpha-bet from Egypt with the local alphabets from Caria proper: none of the alphabets of Caria proper identified until now can be consideredas the source of the alphabetic variety attested in Egypt. In theory, wecan formulate two extremely divergent hypotheses in order to explainthis circumstance: (1) the Carian alphabet of Egypt is a sort of ‘Uralphabet’, prior to a diff erentiation of the original Carian alphabetinto diff erent local varieties, or; (2) the Carian alphabet of Egypt ismerely another local variety, with the particularity that we are yet tofind any Carian inscriptions in the Carian area it comes from, and weare thus unable to establish where this place was.

Certain characteristics of the Egyptian alphabet, which could be inter-

preted as archaic, could be used to support the first hypothesis, in par-ticular the existence of pairs of letters for vowels and semivowels (i/j,u/v, W/w ) against the reduction to a single series of vocalic signs inthe alphabets of Caria proper. Moreover, in the latter case, the splitobserved within the local writings, where some use W for /y/ whileothers resort to a letter that seems to come from w (V/W ), could reinforcethis hypothesis of a more archaic system in Egypt. The chronology isalso favourable to this theory (the inscriptions from Egypt belong toan earlier period) as is the homogeneity of the alphabet used in Egyptset against the clear fragmentation within local varieties in Caria proper.

However, this hypothesis encounters some difficulties: not all the vari-ants of the letters found in the local alphabets of Caria proper can beexplained by the letters used in Egypt. The most notable case is theletter for i ; ‘northern’ variants (y, Y ) can hardly come directly from aproto-form identical to the Egyptian shape of the letter (i ), but insteadseem to indicate a common form that would share the angular formof y, Y, and the horizontal trait visible in i (and also in other vari-

ants of the letter: Î, Ï, I ). The absence in Egypt of certain lettersfound throughout the scripts of Caria proper (% and particularly ñ ) isfurther evidence that the Egyptian alphabet cannot be considered simplyas the Carian ‘Uralphabet’.



227

Therefore, I believe that the second hypothesis must be preferred,although it must be qualified taking into account the possible archaic

traits mentioned above. We can base the theory that the Carian alpha-bet used in Egypt could represent a local variety on the following evidence:

1. There is a very significant gap in our epigraphical documentationof Carian in Caria proper; leaving aside the only possible excep-tion, which I shall address in the following point, there are noinscriptions from most of the Carian coast. Only in Iasos and Kaunoshave we obtained epigraphical documentation. The rest of the knownlocal writings come from the inner part of Caria (Tralleis, Hyllarima,Euromos, Mylasa, Stratonikeia, Sinuri, Kildara). It is therefore pos-sible that the Carian mercenaries who arrived in Egypt by sea camefrom coastal cities where no epigraphical documentation is available.

2. We have already commented on the ‘Egyptian’ features recognizablein a Carian inscription that could possibly come from Halikarnassos.It is true that the Halikarnassean origin of this inscription is notabsolutely certain, and that it appears on an object that could easily

be moved, thus there is no guarantee that it reflects the alphabeticvariety used in the place it was discovered. Yet despite this, a pos-sible origin from Halikarnassos would be consistent with the coastalorigin of the Carian alphabet of Egypt.

3. Despite some clear diff erences, the local writing most closely relatedto the Egyptian alphabet is that of the coastal city of Kaunos.

4. Also, in one of the inscriptions from Iasos we find important par-allels with the Egyptian alphabet (C.Ia 3). Of course, the problemmentioned for C.Ha 1 (that the object could easily have been moved)is again an issue, but in this case at least the place in which it wasdiscovered is certain.

Is therefore possible that the Carian alphabet found in Egypt comesfrom certain western, mostly coastal areas (Myndos, Halikarnassos,Keramos, perhaps also Iasos, and others), where an alphabetic varietystill not clearly attested in Caria was used. A further argument can begiven, but it must be used with a degree of caution: the strong possi-

bility that in the onomastic formulae of Memphis, some ethnic namesappear (see below p. 269–271). If we accept this hypothesis, some formsbecome very meaningful: ksolb-, from Kasolaba, ≤ ugl i-, from Suangela



228

(both cities near Halikarnassos), yjasi-, from Iasos, and kojol - from Kos.17

The possibility that alos ∞arnos may be the Carian form of the place

name Halikarnassos would also have some interesting consequences:alos ∞arnos appears in an inscription from Memphis (E.Me 45), whichwould support the idea that this place was one of the points of originof the Carians in Egypt, and also in C.xx 2, one of the inscriptions of unknown origin that contains an alphabet very close to the Egyptianalphabetic variety. However, these examples cannot be considered asconclusive, and must be used with care.

6. The Common Origin of the Carian Alphabetic Varieties The existence of a number of strongly diff erentiated alphabetic vari-eties in Caria proper, and the hypothesis that the alphabetic varietyused in Egypt is not a sort of ‘Uralphabet’ but rather another localvariety, does not preclude the possibility of an original unity behindthe divergent Carian local writing systems. There is clear evidence infavour of a common origin for these varieties: all the alphabetic varieties share the same aberrant sound values assigned to the ‘Greek-looking letters’ . In

other words, it is the unexpected use of the Greek letters that allowsus to speak of a common origin for the Carian alphabetic varieties.Had the Greek letters retained the same sound values as in Greek, a polygenesis of the diverse local writings could be conceivable. Conversely,the fact that q Q is q , b B is b, m M is m, and so on, in all the Carianalphabets can only be explained by attributing them a single origin.The (apparently) non-Greek letters also show consistent values in thediff erent local writing systems, which reinforces the hypothesis of a mono- genesis of the Carian alphabet. Even the exceptions can in general beeasily explained. Note the case of Kaunian t (T ): this letter (or othersof a similar shape) represents “ in the rest of the Carian alphabets,where t is used for t . But the Kaunian situation is not a chaotic one,and can be explained in the wider context of Carian writing: (1) T andt letters share enough formal similarities to consider a single origin;(2) in Kaunos, / is used for “ , a letter that seems to have been createdfrom T by means of a diacritic mark. The reason for the introduction

17 Cf. Adiego (2004:310).



229

of this diacritic mark is, obviously, to diff erentiate this letter fromKaunian T t . This implies that (1) in a earlier period, T or a very sim-

ilar letter was used for “ in the Kaunian alphabet, and (2) the needfor introducing a diacritic mark was caused by a sort of evolution of the letter for t , which led to an excessive similarity between the lettersfor t (> T ) and for “ (already T ). The starting point of Kaunian T t isnot attested but the abovementioned formal similarities between t andT point to a t, or at least to a form equally similar to both t and T.In conclusion, before Kaunian T and / there seems to have existed apair of letters that were very close to those present in the rest of theCarian alphabets (t, T/F/f ).18

We can now imagine a sort of Carian ‘Uralphabet’ that would includeat least all of the letters attested in all, or at least most, of the Carianlocal varieties. The resulting Carian ‘Uralphabet’ could be the following:

Carian ‘Uralphabet’

a a x X ∞d D d n N nl l p p

ù W y z Z ≤r R r I I Y y Î Ï iL 2 l e 2 eq Q q w ÿ

b B 5 b k K km M m & d

o o v Ú wt T t 0 8 g ?

f F (> / ) “ 1 9 9 zs s j T _ jH (?) c C (O ?) t

u U u $ ç Ø 4 (ÿ ? & ?) b

In this inventory only four letters, of more limited use, are missing:

18 The evolution that led to Kaunian T and / strongly recalls the process thatoccurred in Latin regarding P and R: the first letter, originally G, changed to P, andconsequently, the original P become R.



230

ñ ñ 1 (?)% ã ? 6 ®

Obviously, this ‘Uralphabet’ is only a very hypothetical assumption,and it may possibly be too rigid, insofar as some letters could havebeen created after the diff erentiation of the local varieties, and couldhave arrived to a wide number of alphabets as the result of diff usion.Also, some or all of the four letters excluded from the hypothetical‘Uralphabet’ could have existed in the original inventory of signs. Butirrespective of the exact number of letters present in the original alpha-

bet, the key fact is that the total number of diff erent letters identifiedto date is 34, and that they never appear all together in a single alpha-betic variety. The highest number of diff erent letters documented in alocal writing variety is 31 (in Memphis). This means that the Carianalphabet, although it seems highly anomalous from a formal point of view, given the strange phonetic values of a lot of signs, is much more‘normal’ from a functional perspective: both the possible Carian orig-inal alphabet and the diff erent local varieties contain a number of let-

ters that do not diff er greatly from either the Greek alphabet or fromthe non-Hellenic alphabets of Anatolia (Phrygian, Lydian, Lycian, evenSidetic). The former image of a writing system characterized by a hugenumber of letters, which led to speculations about a semi-syllabic char-acter, has now been clearly superseded thanks to our better under-standing of the local diff erences and of the values of the letters.

B. T O C A

As has been repeatedly mentioned, the most striking trait of the Carianalphabet following its decipherment is that most of the alleged Greekletters have an aberrant phonetic value in Carian. Strictly speaking,only the letters a o u s show a good correspondence between formand value from a Greek point of view. The rest of the ‘Greek’ lettershave unexpected values, and while in some cases these could be explainedphonologically (for example, the value /l/ for l ), many other forms

cannot:t

= /t/,n

= /n/,m M

= /m/,r R

= /r/, etc. At first sight,the Carian alphabet seems to have suff ered a ‘metakharakterismós’ (touse Gusmani’s expression, Gusmani 1994:120) which clearly contravenes



231

the ‘stability principle’ present in the Lycian, Lydian and Phrygianalphabets, according to which the letters for sounds common to both

Greek and the language that borrows the Greek alphabet are includedin the new alphabet, retaining their respective original forms and pho-netic values.19

How, then, do we explain this clear anomaly in the Carian alpha-bet? The hypothesis of a chaotic, arbitrary assignment of phonetic val-ues to Greek letters cannot be ruled out, given that several parallelsexist in the adoption of Western writing systems by illiterate people inmodern times. However, this is absolutely the exception in the Anatoliancontext. In addition, this hypothesis is a very frustrating one: if weacknowledge a free attribution of phonetic values, the Carian alphabetis rendered an inexplicable unicum. Moreover, the ‘chaos hypothesis’does not explain why some letters did conserve their Greek phoneticvalue (the abovementioned a o u s ), and overlooks other possiblepoints of contact between Greek, Carian and other Anatolian alpha-bets: for example the letter p for /p/, so closely linked to Lydian beta b, whose value is also /p/.

It is precisely these possible points of contact that have led me to

develop an alternative hypothesis: that the Carian alphabet was origi-nally also an alphabet of Greek origin, based on the same stability prin-ciple as Phrygian, Lydian or Lycian (Adiego 1998b). The main diff erenceis that this first Carian alphabet, or the Greek model on which it wasbased, underwent a strong process of cursivization that dramaticallyaltered the form of many letters. According to this hypothesis, Cariant (‘qoppa’-sign) would come from a H, m M from a m, n N from an and so on. My hypothesis requires—and this is perhaps its weakestpoint—that at a certain moment in the history of the Carian alpha-bet, the strongly cursivized graphical system underwent a sort of ‘restruc-turing’ to a system of ‘capital letters’, so to speak, and that to this endGreek capital letters were again used as models, although in this caseonly from a formal point of view and without consideration of theirphonetic values. A qoppa could inspire the new form for the /t/-letter,a ny for the /m/-letter, and so on. This could also explain why theletter H, a very poorly documented letter, and one whose phoneticvalue is still unknown, has the shape of one of the most common Greek

letters.

19 For this ‘stability principle’ (‘principe de stabilité’), see Boisson (1994).



232

If we follow to its logical conclusion the hypothesis that the Carianalphabet was originally totally comparable to the Phrygian, Lydian and

Lycian alphabets regarding the stability principle, we must assume that,like in Phrygian, Lydian, and Lycian alphabets, the majority of theCarian signs that represent sounds existing in Greek originate from thecorresponding Greek letters. This assumption was developed in Adiego(1998b), (2004:315–317). The following table represents, with some cor-rections and updates, the hypothesis formulated in Adiego (2004) foreach Carian letter of the ‘Uralphabet’.

Letter Value Possible origina a Greek. a /a/

d D d Greek lD /d/l l Greek B /l/

ù W y Originally a non-Greek valuePerhaps a modi fi cation of r /w/?

r R r Greek T /r/L 2 l Non-Greek value

Note in Hylllarima L , directly from Greek B /l/ plus

diacritic markq Q q Greek value? In any case, an origin from Greek tseems likely

b B 5 b Greek B (?). Note the archaic forms of beta inCrete

m M m Greek M (more exactly a M form)o o Greek O t T t Greek T

f F (> / ) “ Non-Greek values s Greek M ‘san’ or S ‘sigma’H (?) ? (most probably with a non-Greek value)

u U u Greek Ux X ∞ Non-Greek value

Modi fi cation of one of the tectal letters K /k/ q q x /kh/? n N n Greek Np p Greek B (cf. Lydian b /p/)

z Z ≤ Non-Greek valueRelated to sampi-letter 3?

I I Y y i Greek E /e/, E or EI /ej/ (> closed /e/)?Î Ïe 2 e Greek E, H /è/w ÿ Non-Greek value

A modi fi cation of Carian e? k K k Greek K /k/ or rather N /kh/?

,




233

& d Non-Greek value (at least originally) A ligature of Greek ll /dd/?

v Ú w Non-Greek value (?) A modi fi cation of r /w/?

0 8 g ? Non-Greek value1 9 9 z Non-Greek value (?)j T _ j Non-Greek value

Perhaps related to Phrygian letter for /j/, y, Yc C (O ?) t Non-Greek value

Related to sampi -letter 3?$ ç Ø 4 b Non-Greek value (at least originally)(ÿ ? & ?) A ligature of Carian bb /bb/?

Note in Hyllarima the use of B directly from Greek

It is clear that this hypothesis poses serious problems that make itdifficult to accept: it forces us to assign a very high chronology to acursivized writing system. No examples in favour of this old phase of

the Greek or Carian alphabets exist and consequently, the evolution-ary processes proposed for each Carian letter, no matter how con-vincing they may seem, are totally ad hoc.20 However, this is, togetherwith the hypothesis of a purely chaotic imitation of a Greek model,the only solution that I am able to imagine, if one intends to provethat the Carian alphabet comes from the Greek alphabet. Alternativehypotheses would have to accept the intervention of other writing sys-tems that could be responsible for the final form of the Carian alpha-bet, and such a solution turns out to be an obscurum per obscurius , since

these supposed writing systems are unknown.

20 In any case, I consider that the parallel examples in other writing systems, althoughthey are not at all useful as evidence, do off er some support several explanatory hypothe-ses on the origin of the Carian letters. This is the case of the South-Picenian lettersfor t or q , see Adiego (1998b:68).

table (cont. )

Letter Value Possible origin



CHAPTER SIX

PHONOLOGICAL FEATURES

A. T P S

In the present section, I will try to illustrate the phonological systemof Carian. This goal can be achieved only in an approximate way, by

interpreting the graphemes of Carian on a twofold basis: from the cor-respondences between Carian letters and how they are reflected in theadaptation of Carian names in Greek and Egyptian, and from the rela-tionships between graphemes within the Carian graphical system. Thedegree of precision in describing the Carian phonological system can-not be particularly high. In many cases, we can only attain a verygeneric characterization of the phonological values of the graphemes.I will try to avoid excessively Byzantine arguments. Firstly, I will attempt

to establish a descriptive table of possible Carian phonemes. A furthersection will explore the Anatolian background of these phonemes.

1. Vowels and Semivowels

The inventory of possible graphemes for vowels and semivowels variesin the diff erent sub-corpora of Carian inscriptions. In general terms,we can divide these into three areas: (a) the Egypto-Carian inventory,(b) the Carian (except Kaunos) inventory, and (c) the Kaunos inven-

tory. Despite the diff erences between these three types, I believe thatsome interesting generalizations can be traced, so that the resulting sys-tems are coherent and can be easily inter-related.

(a) The Egypto-Carian inventory is the most complete. Here we find9 diff erent graphemes for ‘vocalic’ (in a broad sense, including possi-ble semivowels) phonemes: a e i o u y j w ÿ.

(b) In the Carian inventory, the graphemes are reduced to 6: a, e,i, o, u, y (for y, the letter in most of the alphabets is W; in Sinuri-

Kildara,V

; in Mylasa, W ).(c) In the Kaunos alphabet, the inventory is limited to 5 diff erentgraphemes: a, i, o, u, y.



235

The most remarkable fact is undoubtedly that the Egypto-Carian sys-tem displays a higher number of graphemes than the other two sys-

tems. How can this discrepancy be explained? During the deciphermentprocess, it became evident that the Egypto-Carian system containsgraphic alternations: i alternates with j, u alternates with w , and y alter-nates with ÿ:

i /j u/w y/ ÿ

∞i/∞ j pnu≤ ol /pnw ≤ ol mdayn/mda ÿn(psm“ kw)neit/(pd)nejt uksmu/wksmu-≤ ( “ ar)kbiom/kbjom-≤ upe/wpe “ arnai-≤ /“ arnaj-s u≤ ol /w ≤ ol ≤ wliat/wljat

yiasi/yjas[i-≤ ]

These alternations certainly cannot be explained as morphological , becausethey appeared in diff erent positions and in some cases, although notall, no morphological diff erences between the words implied could beobserved. The simplest solution, at least in the case of i/j, u/w has

already been suggested in earlier works (see Adiego 1994a:47–48): j andw seemed to be the graphemes for the semivowels corresponding to i and u respectively. It is true that in some cases an interpretation of j and w as semivowels is not absolutely clear, but in such cases, it isplausible that the semivowels were used for their corresponding vow-els on occasions. The advantage of this hypothesis is that it allows usto explain the relationship between the Egypto-Carian and Carian sys-tems in a straightforward way; given that the Egypto-Carian system ischronologically older than the actual Carian inventories, a graphemicreduction process could take place, so that the use of specific graphemesfor the semivowels /j/ and /w/ was abandoned, and the simple vocalicletters /i/, /u/ were used in their place. This could be used to explaina form such as u≤ ol ≤ in Hyllarima vs. w ≤ ol ≤ in Egypt.

The situation is not so clear in the case of y vs. ÿ; while for i/j andu/w , an opposition vowel (i , u )/semivowel ( j, w ) can be established,both from their distribution properties and from the survival of i, u inthe reduced systems of Caria itself, neither of these criteria is valid in

establishing the diff erence between y and ÿ. The number of appear-ances is very similar, but the only clear example of alternation is notsufficient to attribute to one or the other letter a vowel vs. semivowelvalue, and in the case of the alphabets of Caria itself, while W continues

236

as a vocalic sign in many alphabets (including Kaunos), the letters V/Wof Kildara-Sinuri-Mylasa seems to be closer to w than to W. In anycase, I believe that the hypothesis of a vowel/semivowel opposition forexplaining both the existence of two graphemes in the Egypto-Cariansystem and the reduction to one grapheme in the Carian proper andthe Kaunos systems is also the simplest solution for y/ ÿ. The sole unre-solved problem would be to ascertain which of the letters representsthe vowel sound, and which the semivowel. In the transcription systemadopted here, preference is given to W, transcribed y, but this must betaken as a purely conventional solution.

According to this hypothesis, systems (a) and (b) coincide in that theyboth consist of six diff erent vocalic sounds, a, e, i, o, u, y, to whichthree graphemes for the semivowels corresponding to the closest vocalicsounds (i, u, y ) are added in (a) ( j, w, ÿ ). In the Kaunos system, wherethere are no specific graphemes for semivowels, the most notable char-acteristic is the absence of an e vowel.

As for the phonological value attributed to each vocalic letter, thetranscription system adopted here aims to off er a broad reflection of the quality of the vowels:

a is used to adapt Greek a in Lusikrãthw = lysikrata-, Ekata›ow =ktai-. Conversely, a corresponds systematically to a sound transcribedin Greek as <a>: ada = Ada; ki l ara = Kildara; par(a)- (in para-eym, par-yri ∞ ) = Para- (in Para-ussvllow, Par-ussvldow, Para-udigow; “ aru≤ ol = Sarusvllow.

e is used for Greek h in the name OÈliãdhw (Carian uliade ). In Greek,Carian e is adapted systematically by means of <h>: mane = Manhw; pl eq-≤ = Peldhkow; qtblem-≤ = Kotbelhmow, Kutbelhmiw; somne-≤ = Svmnhw.

i is used to adapt Greek i: Lusikl∞w = lysikla-; Lusikrãthw = lysikrata ;Nikokl∞w = nik [—]la-; OÈliãdhw = uliade . Note also ‘Ipposy°nhw = i [—–]ini.This latter example also demonstrates the use of Carian i to reflectGreek e, but one must note that the example comes from Kaunos,where no specific letter for e-sounds (Greek e, h ) existed.1 Conversely,Carian i/j is transcribed in Greek as i: ari “ = Ar(r)issiw; arli “ = Arlissiw;i b arsi-≤ /i b rsi-≤ = Imbarsiw, Imbras(s)iw; ( “ ar)kbiom/kbjom-≤ = Kebivmow; pikre/pikra = Pigrhw.

1 The adaptation in Kaunos of Greek e by means of Car. i vs. that of Greek h bymeans of Car. a seems therefore an attempt to reflect the diff erent sound qualities of (close) e and (open) h.

237

o is represented by Greek v: ( “ ar)kbiom/kbjom-≤ = Kebivmow; ksolb-≤ =cf . Kasvlãba; ktmno, ∞tmño = Ekatomnvw; msnord-≤ = cf . Masanvrãda;

plqo = Pel(l )ekvw; somne-≤ = Svmnhw; u≤ ol (and compounds) = Ussvllow(and compounds); yrqso-≤ = Urgosvw.

 is used to reflect Greek ou in OÈliãdhw = uliade. Conversely,Carian /<w> corresponds to Greek <ou>, , <o>: quq, dquq =Gugow, Idagugow; pun-w ≤ ol -≤ = Pon-usvllow; tñu-≤ = Tonnouw; wljat/wliat =Oliatow/ Uliatow; u≤ ol /w ≤ ol ≤ = Ussvllow.

For the phonological value of y/ ÿ, the Greek-Carian bilingual inscrip-tion of Kaunos C.Ka 5 clearly supports the argument for a /y/ (semi-vowel /w/) sound; in this inscription, W y is used in the Carian adaptationof the Greek names Lusikl∞w and Lusikrãthw (lysiklas [, lysikratas [ ). Thisis consistent with the systematic use of Greek u for the transcription of y/ ÿ in the Carian names: “ ayriq (E.Me 25) = Saurigow, yrqso (C.My 1) = Urgosvw, etc. Unlike u, no examples of Greek ou, o for Carian y, ÿ aredocumented in Carian personal names of Greek sources. Therefore, itseems that the best explanation for the apparent proliferation of “u-letters” in Egypto-Carian (a puzzling aspect of earlier works, see Adiego1993a:273–275) is to assume that in Carian there were two [+ high],

[+ rounded] vowels, /u/ and /y/, diff erentiated by their back vs. frontarticulation, respectively, and that in Egypto-Carian each one of thesesounds had a specific grapheme for the corresponding semivocalic glides(/w/, /w/).

In Kaunian the vocalic system is reduced to five vowels: /i, a, o, u, y/. There is no e , and this absence is apparent in the adaptation of Greek names, where a is used for h: both Lusikl∞w and Lusikrãthwappear as a -stems in lysikla-, lysikrata-. More obscure is the case of theethnic otonosn (accusative) for ‘Athenian’. The presence of o for Greekh (’Ay∞nai, ’Ayhna›ow ) seems inseparable from the apparent vocalicmetaphony that has converted the three vowels of the Carian form intoo. The trigger for this metaphony could be the final vowel: the Greekà-stem would have been adapted as an o-stem, and this o would havecaused the metaphony (atano- > otono-, see below p. 259). It is irrele-vant if the original form for Athens in Kaunian was in fact a Doricform with a long a instead of h or not, because both à and è wouldbe adapted as a in Kaunian, as shown above by the personal names.2

2 Blümel (1998b:172–173) has argued convincingly against a Doric model for theCarian word, pointing out, among other things, that Kaunian Greeks did not speakDoric.



238

The Defective Notation of Vowels

One of the most difficult problems posed by Carian vocalism is theapparently defective notation of vowels. This phenomenon can bedetected above all in the comparison of the Carian nouns with theiradaptation in Greek (1), but also in alternations within Carian docu-mentation (2), and by comparing several Lycian (and Milyan) parallelforms (3):

(1)artmon— ’Art°mvn

dquq— Idagugowkbjom≤ — Kebivmowksolb≤ — (place name) Kasvlabaktais— ‘Ekata›owktmno, ∞tmño≤ — Ekatomnvwmsnord ≤ — (place name) Masanvradantro—cf . Neter-bimow, Lyc. Natr-bbij e mi par ÿd ∞≤ — Paraudigow plqo— Pel(l)ekvw pl eq ≤ — Peldhkow pñmnn≤ ñ— Ponmoonnowqla l i ≤ — Kolaldiw, Kulaldiwqtblem≤ — Kutbelhmiw, Kotbelhmowq ÿblsi ≤ — (ethnic) Kubliss//ow//≤ ugl i ≤ — (place name) Souaggelatñu≤ — Tonnouw ylarm-i t — (place name) Ullarima

(2)b rsi—i b rsi ≤ —i b arsi ≤ dw ≤ ol ≤ —idu≤ ol ≤ kbidn/kbdyn“ ∞ yrpai— ∞ yrapai ≤ mdayn, mda ÿn—mwdon≤ pikrm≤ —pikarm≤ pnu≤ ol —punw ≤ ol ≤ (Ponussvllow )

psm“ k(wneit)—psma ≤ k—pisma “ k/pisma ≤ k s d i(sas)—si d i “ r(quq)— “ ar(kbiom, etc.) (Greek Sar-)



239

trq d -—trqud e cf . Lycian Trqqñt-ttbazi—ttubazi

urm≤ —urom≤

(3)sb—cf . Mil. sebe ÿbt—cf . Lyc. ubete

To these examples we may add a number of forms that display chainsof consonants without vowels: adymd ≤ , a ∞tmsk “ , ank b u[, (idrayri)dsemd ? bq,

pd a ∞m≤ uñ, snn, sñaid l o, tbridbd ≤ , etc.Some of these examples could be interpreted as the result of syllabic

liquids and nasals that in Greek must be represented with the aid of a vowel. This explanation is valid for forms such as tñu-≤ vs. Tonnouwor ñ vs. on in pñmnn-≤ ñ vs. Ponmoonnow. Moreover, several internal alter-nations such as pikrm-≤ /pikarm-≤ or “ r-/“ ar- point to a syllabic liquid,which develops an intrusive vowel in Carian (see below p. 262). Butthis explanation does not account for many other forms in which theinterpretation of liquids or nasals as syllabic nuclei is less plausible; in

the case of msnord-≤ vs. Masanvrada, an interpretation as /µ.s–.o.ºd/seems very unlikely. Also in examples like ntro, ntokris , it seems unlikelythat n can be a syllabic nasal in direct contact with t : a sequence -nt-seems to have evolved into -nd- and is systematically noted by meansof a specific letter, d (trq d , trqud , cf . CLuw. Tar ¢unt- ). In other cases,there are no liquids or nasals involved at all: sb vs. Lyc. sebe, kt º vsEkatº etc.

Yet conversely, if in several cases the inclusion of intrusive vowelsin Greek adaptations could be seen as an attempt to resolve the difficultiesof pronouncing some consonant groups in Greek (for example ºdq º →ºdag º in dquq → Idagugow ), this explanation is unfeasible when appliedto cases such as qla l i- → Ko/ulaldiw, where kl was phonotacticallyavailable in Greek. This explanation also fails to account for internalalternations within Carian ( pnº/punº, dº/id º ). Therefore, at least in aconsiderable number of cases, we must accept the actual existence inCarian of vowels that are not reflected graphically. The question of how to explain this defective notation of vowels is indeed a very difficult

one. Can one assume that Carian writing shows the vestiges of a graph-ical system wherein vowels were not noted? This hypothesis would havefascinating implications for the origin of the Carian alphabet, since itcould imply that Carian writing was borrowing from a graphical



240

system very close to the Semitic model of the Greek alphabet, that isto say from a purely consonantal system. But this hypothesis, attrac-

tive as it may seem at first sight, runs into serious obstacles. The useof vowel signs is generally very consistent; no forms such as **u≤ l oreven **≤ l besides u≤ ol or **kbjm besides kbjom are found, forms whichone would expect to find if the defective notation of vowels was apurely graphical question. Moreover, in general, at least one vowel signis found in each independent Carian word, a particularity that war-rants an explanation.3

Given that these consistencies are more important than occasionalalternations like dw ≤ ol -≤ /idu≤ ol-≤ , a linguistic basis for the defective nota-tion of vowels seems more likely. The possibility that the omitted vow-els is a sort of schwa -sound can explain some cases, particularly thosein which a vowel a or e is found in the Greek adaptations. Note par-ticularly an alternation like Ussaldvmow/ Usseldvmow. Even in someexamples of Greek o/u vs. Ø in Carian, the back character of the schwa-vowel could be attributed to a secondary backing caused by the pre-ceding consonant; in all of these cases, the sound preceding the unrecordedback vowel is q , a possible uvular sound (see below p. 244). However,

this explanation would make it difficult to explain i in dw ≤ ol -≤ vs. idu≤ ol ≤ .Perhaps then we are dealing not with schwa-like vowels , but simply withnon-stressed, short vowels (*a, *e, *i, *o/u ) vs. stressed and/or long vow-els. Finally, another factor that could account for the presence of vow-els in Greek and their absence in Carian has been conveniently notedby Tremblay (1998:119); it is possible that some Carian names wereborrowed by Greek before certain processes of vowel weakening or losstook place in Carian. Tremblay (ibid.) put forward an important argu-ment in favour of this hypothesis, suggesting that some indirect Greek

forms seem to be more archaic than the direct Carian equivalents (notefor instance the absence of original medial i in the ethnic ylarmi t vs.Greek Ullarima, if it continues the cuneiform place name Wallarima ).4

3 Exceptions like sb are undoubtedly connected to the fact that this type of wordalways appears attached to the following word (sb=polo, sb=ada , etc.), which would sug-

gest a proclitic nature. In other cases such as snn, the presence of syllabic nasals couldexplain the absence of vowels.4 Other examples given by Tremblay (ibid .) seem to me less compelling. Also the

attempt of elaborating a relative chronology (Tremblay 1998:117–118) is in my opin-ion premature.



241

The tendency to omit unstressed vowels (whether schwa-like vowelsor simply short vowels, without discarding an actual vowel loss) would

be a satisfactory explanation for the above statement that at least onevowel per word is noted in Carian writing. It could also explain theapparently morphological alternation found in the place name kbid-nvs. the ethnic form in pl. acc. kbdyn-“ , where stress movement wouldbe responsible for the i/Ø alternation: /’kbi.dn/→ /kb(i).’dy.n(V)“/.

In cases like i b rsi/b rsi, trqud e/trq d -, si d i/s d i , the notation vs. absenceof notation of the vowels seems inseparable from the type of conso-nant that immediately follows the relevant vowel: b and d are at leastoriginally *mb, *nd respectively (see below). The presence vs. absenceof a full vowel is consistent therefore with an alternation between asyllabic nasal µ , – and a V m, V n sequence (or a nasalized vowel)(/imbrsi/, /ıbrsi/ ~ //µbrsi/, /trqund-/, /trqud-/ ~ /trq–d-/, /sindi/,/sıdi/ ~ /s–di/).

When w, j appear as a seemingly syllabic nucleus, there is somedoubt about their real value. Are they simply used instead of the purelyvocalic letters (i, u, respectively) or must they be interpreted as truesemivowels, omitting the vowel that accompanies them? Many exam-

ples are ambiguous, and point to either one interpretation or the other.We must first of all note that in the case of j , the instances of an(apparent) syllabic use are extremely scarce: generally, j appears witha clear semivocalic function, immediately preceding and/or following a vowel (wljat( ≤ ), arjom≤ , pdnejt, pjabrm, kbjom≤ , kojol ). In the entire Saqqârasub-corpus, only a single example of j in a syllabic position can befound: ∞ j in E.Me 36, which seems to be merely an uncommon use of j instead of i (cf . the systematic spelling ∞i of this frequently used wordin the rest of the inscriptions of the sub-corpus). In the other Carian

inscriptions, I have found only two unclear examples in Thebes, anequally complicated form in Silsilis, and the words pjdl in C.xx1 and jzpemd ane in C.xx2. All of these forms are ambiguous.

The situation of w is markedly diff erent; the examples of this letterin a syllabic position are far more frequent. Ultimately a single, com-prehensive solution for all cases is simply impossible to find. Cases suchas wpe (E.Me 36, E.Me 41) seem to point to a simple graphical alter-nation with u (cf . the more spread form upe for this word). Conversely,

the systematic use of w in the spelling of the (possibly) ethnic wordmwdon≤ (mwton≤ 1x), widely documented in Saqqâra, supports its inter-pretation as a true semivowel followed (or preceded) by an un-notatedvowel.



242

The behaviour of the names of the u≤ ol -family in the Saqqâra sub-corpus also substantiates the use of w for w(V); in this case, the name

u≤ ol , both in its solitary and compounded forms, appears with u whenin nominative case, but with w when in genitive: pnu≤ ol (E.Me 19),tdu≤ ol (E.Me 24), “ aru≤ ol (E.Me 30) vs. w ≤ ol ≤ (E.Me 12), punw ≤ ol ≤ (E.Me13), dw ≤ ol ≤ (E.Me 35).5 A compelling explanation for this situation couldbe that there was a stress displacement, with /’u≤ol / vs. /w(V).’≤ol ≤/.In the genitives, the unstressed vowel would not be notated graphically,and w would represent a true semivowel preceding it. This hypothesishas a counter-example: in Thebes a nominative pnw ≤ ol is found. However,in the Saqqâra sub-corpus the rule functions well and the examplesseem to go some way towards supporting this explanation. It is alsocorroborated by the likely connection of u≤ ol with the family of nounsending in -uassiw (Aktauassiw, Panuassiw, Saruassiw, cf . also Oa3a3iw ),where a *wa ≤ - basis could be reconstructed.

Also in the case of the names wljat, wliat, “ arwljat-≤ and “ r-wli ≤ , theetymological connections proposed (see Chapter 11, s. v. wliat ) point tooriginal *waliat-, *wali- forms.

2. Consonants

Voiceless Stops

The following five voiceless stops can be recognized in Carian:

p t/T xX kK q/Q

p t ∞ k q

Labial p, dental t and velar k do not pose any particular problems.

The phonological value of the corresponding letters has been firmlyestablished in previous works and it should be sufficient here to pro-vide a few relevant examples: p is used to adapt Egyptian p and Greek p in proper nouns: pisma “ k

(and variants) = Psmtk, pdnejt = P3djNjt, pi d aru = P¤ndarow. Conversely,Egyptian p and Greek p reflect Carian p in Carian personal names: paraeym = Prjm, pnu≤ ol = Ponussvllow, pikre = Pigrhw, etc.

If the explanation of nproº (in nprosn≤ ) as a Carian version of theEgyptian name Nfr-˙r (Nefervw ) is accepted (Schürr 1996a:68, n. 18,

5 Adiego (1993a:273).

243

Vittmann 2001:42), it could serve as a good example of an Egyptian f adapted to p in Carian. See Chapter 11, s. v.

t is used for Egyptian t and for Greek t, and also y (/th

/): pdnejt =P3djNjt, ntokris = Nj.t-jqr (Nitvkriw ), lysikrata-s = Lusikrãthw, otono-sn =ÉAyhna-. This latter example confirms the absence of voiceless aspiratestops in Carian, the corresponding Greek ones being adapted by meansof the simple voiceless stop. For the use of Greek t for Carian t inpersonal names, cf . wljat (and variants) = Uliatow, Oliatow, tñu≤ = Tonnouw,etc.

k is used for Egyptian k in the abovementioned personal names pisma “ k (and variants), ntokris , and for Greek k in the names lysiklas [ and lysikratas [ = Lusikl∞w, Lusikrãthw. k is used for k in the adaptation of Cariannouns: kbjom≤ = Kebivmow, ki l ara- = Kildara/Killara, ksolb≤ cf . Kasvlaba.Note also the equivalence of Carian kbidn ‘Kaunos’ = Lycian X bide ‘id.’(with Lycian < x > = /k/)

The exact values of ∞ and q are less clear. That ∞ represents a tec-tal sound is easily deducible from the use of Egyptian k in the tran-scription of the Carian name urs ∞le ≤ , 3rskr. For this letter, a palatal value(palatal stop [c] or palatalized velar stop [k j ]) can be envisaged, although

the argument is based only on comparative evidence: its use in the(originally) relative pronoun ∞i , whose Anatolial proto-form is *k w is (Hitt.,CLuw. kui “ ). It is plausible to imagine that Carian has undergone aprocess similar to Milyan: the labiovelar stop has first been delabial-ized (*k w is > *ki , with loss of the final s ) and then palatalized before i (*ki > /k ji/ or /ci/, Milyan ki [ci]). In Lycian, /ci/ has eventuallybecome /ti/, completing the fronting process.6

The palatal character of ∞ is also recognizable in the striking alter-nation ∞/k, shown by the name ktmno (Thebas) vs. ∞tmño≤ (Sinuri). Here,

if k and t were in contact, t could have caused the fronting of k to ∞.Certainly, the Greek form of the name, Ekatomnvw, with a between kand t, seems contradict the existence of a contact between the twostops, but it is possible that the Greek form has undergone the ana-logical influence of the goddess’ name Hekate, so that the a vowelwould be absent in the original Carian name.7 An alternative view, fol-lowing on from Tremblay (see above) could be to assume that the

6 This explanation was formulated for the first time in Adiego (1995:29–31).7 It is interesting to recall that Ekatomnvw can plausibly be seen as a pure Greek

name, as Neumann has repeatedly pointed out, see Neumann (1994:17).

244

Greek form reflects a more archaic stage than Carian, and that in thislatter language a syncope would have taken place, permitting a con-

tact between k and t. In any case, it seems more than a mere coinci-dence that the same name appears in Lycian spelled as katamla , withk [c], not x [k].

Apart from urs ∞le ≤ and ∞tmño-≤ , there are no more clear examples of onomastic identifications where ∞ would be included: the connection of alos( d ) ∞arnos( d ) with the place name Halikarnassos remains very doubt-ful. In any case, it would only be useful for confirming the tectal valueof the sound. As for p∞simt ≤ , although its connection with the well-known Egyptian name Potasimto (P3-dj-Ó r-sm3-t3wy, shorter form P3-dj-sm3-t3wy, Greek Potasimto) suggested by Schürr (apud Ray 1994:205)is a compelling theory, the phonetic details of the Carian adaptationof the name remain obscure to me.

The tectal character of q is also beyond question; in the Greek adap-tation of Carian names it is systematically represented by k or g : qtblem≤ = Kutbelhmiw, qla l i ≤ , qla l is = Kolaldiw, Kulaldiw, quq = Gugow etc. Inthe bilingual inscription E.Sa 2, the Carian name qyri ≤ is adapted inEgyptian as K3rr , with the use of the biliteral sign <k3> for ku ~ q ÿº

(Vittmann 1996). It is also used in the possible Carian adaptation of the Egyptian name Ny-k3w , Nek«w, niqau≤ .This letter appears in several of examples followed by o, u (triqo, plqo,

quq, trqud e , etc.), and o, u are also the vowels used in Greek after kwhen this latter corresponds to a q in contexts of defective vowel nota-tion (qla l i ≤ , qla l i ≤ = Kolaldiw, qtblem≤ = Kutbelhmiw, Kotbelhmow, per-haps also qtblo = Kotobalvw, and this would seem to point to a backarticulation of q (uvular /q/?)

Occasionally, both k and q (there are no examples available for ∞ ) arerendered in Greek through a voiced velar stop: pikre = Pigrhw, yrqso≤ =

Urgosvw, quq = Gugow, dquq = Idagugow. The two first examples can beeasily categorized as allophonic; k, q would tend to become voiced incontact with the sonorant r (note the alternative spelling Pikrhw for thesecond name, which confirms the rather spontaneous and irregular voic-ing phenomenon). More delicate is the case of (Ida )-gugow, wherein thespelling of the two voiced stops is systematic, making possible the exis-tence of true tectal voiced sounds, not noted graphically in Carian.

An alternation ∞/q can be identified in the family of names in yriq/yri ∞(= Greek -urigow ), note “ ayriq vs. idyri ∞-≤ , paryri ∞(-≤ ), and perhaps also in ÿdiq/ ÿd ∞ (“ a ÿdiq-≤ vs. par ÿd ∞-≤ ) (for the possible common origin of yriq-and ÿdiq- stems, see below pp. 262–263). It would be difficult to off er

245

a clear explanation of this alternation if the rather distant sound val-ues argued here (∞ palatal vs. q uvular) are accepted.

Voiced Obstruents(?)

b B 5 d Db /b/ [ b ]? d /d/ [ ä ]?

The Greek rendering of Carian names, along with additional evidencefrom Egyptian and Lycian, confirms the point of articulation and thevoiced character of these two letters:

b is rendered by b: kbjom≤ = Kebivmow, ksolb≤ , cf . (place name) Kasvlaba,k ≤ at ÿbr — Janduberiw, qtblem≤ = Kutbelhmiw, Kotbelhmow, q ÿblsi ≤ , cf . theethnic Kublisse›w, etc.

In addition to this evidence, we must also consider the use of Carianb for Egyptian b in some personal names of Egyptian origin: ttbazi,ttubazi = T3-dj(.t)-b3st.t (Tetobastiw ), pdubez = P3-dj-b3st.t , Petobastiw,Petoubestiw. The comparison with Lycian and Milyan reinforces thisphonological value; as well as the correspondence between the Carianand the Lycian name of Kaunos, Car. kbid-, Lyc. Xbide , the lexical

equivalences Car. sb ‘and’ = Mil. sebe , Car. ÿbt = Lyc. ubete ‘off ered’are also very significant.d is adapted as d in Greek: ada = Ada, dquq = Idagugow, dw ≤ ol ≤ ,

idu≤ ol ≤ = Idussvllow, msnord ≤ cf . place name Masanvrada, ardybyr ≤ =Arduberow. Conversely, Carian d reflects Greek d in the name uliade =OÈliãdhw.

Carian d is used to render Egyptian d: pdnejt = P3djNjt .A good example of the correspondence Carian d : Lycian d is the

word ted associated with Lycian tedi ‘father’.

It seems that there is no letter for /g/ [ g ]. See above for Greek g corresponding to <k>, <q> in Carian.

It is possible that b, d were articulated as fricatives in intervocalicposition, as is the case in Lycian, but there is no direct evidence forthis type of articulation. Our suspicions are based mostly on the exis-tence of the letters & d and Ø/4/B/ b: they represented, at leastoriginally, clusters of nasal + voiced stop (*/nd/, */mb/), precisely thecontext typical for the articulation of voiced obstruents as stops vs. the

fricative realisation of these sounds in most other contexts (cf . the sim-ilar situation in Lycian). But the doubts concerning the real sound valueof <d> and in Carian (see below) are sufficient to cast doubt onthe fricative character of b, d .

&

246

It is also possible that in an initial position, the original /b/ and/d/ have become voiceless, as in Lycian. However, the Carian docu-

mentation is not as clear as the Lycian evidence. Although examplesof b and d are scarce, and some of them can be interpreted as theresult of a lack of notation of the initial vowel (for instance, dquq =Idagugow ), there is not enough evidence to support the assertion thatall the examples of initial b and d must be interpreted in this way.

As for the indirect documentation, the number of forms with initial band d is also minimal, and in a considerable part of these a contact witha sonorant r , l, n could be responsible for an allophonic voicing of a voice-less stop: Bruajiw/Bruassiw, Brvlow (personal names), Br¤oula, Bridaw(place names), and perhaps also Beryaw, Berrablviow, Dandvmow, Daru..ow,Dersvmanhw, Dersvw, Dersv . . . tiw (personal names), Bãrgasa/Bãrgaza,BargÊlia, Bolli.evn (place names). A good argument in favour of thisexplanation is the alternation P-/B- attested by the name Bãrgasã/Pãrgasa.The remaining examples of B- and D- would thus become very scarce(personal names: Boivmow, Deibow, place names: Babein, Bvnitv, BubassÒwand variants, Bvrand//a//, D°dmasa, Didassai, D¤duma —although thisis doubtful, as it could have been influenced by Greek—and DÊndason).

An alternative explanation for some of these words, to my knowl-edge not proposed until now, would be to assume that the Greek ini-tials B- and D- are in fact a reflection of Carian (< *(V)mb ) and<d> (< *(V)nd ). This explanation is plausible at least in the case of the name Dersvw, which could be compared simultaneously withAndarsvw/Androsvw and with d ar “ (if a PN). Also Brvlow, if from*(V)mbrol-, could be connected to the family of names in (i)b r- (cf . par-ticularly para-i b re l - ). The absence of an initial vowel in the Greek adap-tation would correspond precisely to the situation in Carian, where wefind vis-à-vis forms both with an initial vowel (vgr. i b rsi- ) and withoutit (vgr. b rsi- ).

Arguing in favour of a devoicing process of b in an initial position,we should finally mention a good indication found directly in Carian:the internal alternation seen in Carian between piks- and dbiks, pik(a)rmand dbikrm, where b is clearly etymological (bik- < PIE *bh è h2- ).

Voiced Stops or Nasal + Voiced Stops

Ø, 4, (Hyllarima) B, (Mylasa) &b d/mb/ or /b/? /nd/ or /d/?

&

247

As Schürr has demonstrated, practically all the interpretable examplesof the letters b and d find their origin in (respectively) *-mb- *-nd- groups

and are reflected systematically in Greek by -mb-, - nd-: (i)b rsi , i b arsi =Imbrassiw, Imbarsiw, trq d , trqud e = CLuw. Tar ¢unt-, lycian trqqñt (<nt>,<ñt> = /nd/). We are left with the problem of ascertaining what theseletters actually represent in Carian. Three main hypotheses are plau-sible:

(1) b and d denote true consonantal groups -mb-, -nd-, respectively.(2) b and d denote pre-nasalized consonants (/mb/, /nd/), a type of

stop that exists in diff erent world languages.8

(3) b and d denote only the oral voiced stops /b/ and /d/; like inmany other languages (from Lycian to Modern Greek or Spanish), /b/and /d/ would appear as stops only in certain contexts, particularlyafter nasals. In the remaining contexts, their articulation would be frica-tive and denoted by b [ b ] and d [ ä ] (see above). The graphemes and <d> would therefore be used only to represent these stops, andthe nasal that precedes them would not be noted graphically.

With our current knowledge of Carian, it is impossible to choosebetween these hypotheses. Preference could be given to (3) if we were

to find examples wherein b or d were used for a voiced stop not aris-ing from a nasal + stop sequence. Unfortunately, no concrete exam-ples of this kind have been found; the only form for which this explanationhas been suggested is pd aº in pd a ∞m≤ uñ , compared in Adiego (2000:145)with Lycian pdde ‘place’, from PIE * pedom. According to this interpre-tation, d would represent a stop articulation of the dental in direct con-tact with the stop p, as in Lycian (where a gemination has also takenplace). However, this connection is not conclusive enough to be usedas the basis of a definitive argument in favour of d = [d], and not

[nd] or [ nd].

Liquids

r R l 6 L 2 Lr l ® l /r/ /l/ /r j/? /l.l/?

8 Cf . Boisson (1994:219).



248

Four letters have been identified as representing liquids and nasals inCarian.

r corresponds to Greek r and Egyptian r in the adaptation of per-sonal names in Carian: lysikratas [ = Lusikrãthw, ntokris = Nj.t-jqr (Nitvkriw ).Conversely, both Greek r and Egyptian r reflect Carian r in Cariannames: ki l ara = Kildara/Killara, arli “ ( ≤ ) = Arlissiw, arliom≤ = Arlivmow, pikre ≤ = Pigrhw, etc. paraeym = Prjm, arli “ ( ≤ ) = Jr “ 3, “ arkbiom = ” 3rkbym,urs ∞le ≤ = 3rskr .

l renders Greek l in lysiklas [, lysikratas [ = Lusikl∞w, Lusikrãthw, uli-ade = OÈliãdhw, and l is used in Greek for Carian l: arli “ = Arlissiw,arliom≤ = Arlivmow, l ÿ∞se, l ÿ∞si ≤ = Lujhw, qla l i ≤ , qla l is = Kolaldiw etc.In Egyptian, where the graphical reflection of l was always problem-atic (see Loprieno 1995: 31, 33), we find r for Carian l in 3rskr =urs ∞le ≤ , whereas the sequence ºrl º in arli “ , arliom≤ is noted by means of a single r ( Jr “ 3, Jrym ). Although more precise conclusions about theirexact articulation (was r flapped or trilled, and was l more or less den-tal?) cannot be established, it is clear that the letters r and l representthe two basic liquids of the Carian phonological system.

The status of the two remaining liquid letters is less clear, and it is

significant that these signs are not found in all the Carian alphabets:l is absent from the Thebes and Mylasa inventories, and ® is even lesswidespread, having only been found in Egypt.

The use of ® is limited to the following words: ar ® i “ , me ®≤ , qdar ® ou≤ and t ® ∞at(a)r ≤ . Both ar ® i “ = Arrissiw and t ® ∞at(a)r ≤ maintain the theoryof ® as a palatalized r /r j/, as a result of the contact with a palatalsound (i, ∞ ).9 This explanation is less convincing in the case of qdar ® ou≤ ,whilst me ®≤ does not off er any evidence either for or against this hypoth-esis. In arri “ , an assimilation process could be behind r ® if we assume

that there is a connection with the more widespread name arli “ . Thisexplanation would also be feasible for qdar ® ou≤ if the word is related toCLuw. ¢utarl à- ‘slave’, as suggested in Adiego (1995:24–25).

l is the letter used in Carian for the sound(s) transcribed in Greekadaptations as ll and ld. Boisson (1994:216–217) off ers a detailedanalysis of earlier proposals for interpreting this letter, and formulatessome possible values. While I believe that one cannot rule out the

9 It is worth noting that the new text of Hyllarima (C.Hy 1a) off ers for the nameArrissiw the form ari “ , without any special sign for the (possibly peculiar) sound ren-dered in Greek as -rr-.



249

possibility that l represents a peculiar unitary sound, for which theoscillating Greek spelling ll /ld would be a rough translation, it seems

to me preferable to interpret l as the simple notation of a geminatesound, whose articulation was more dental than that of the Greek ll ;indeed this could explain the alternative spelling ld. That the soundwas very close to Carian l is demonstrated by the systematic use of l in those alphabets (namely Thebes and Mylasa) that do not have aspecific letter for this geminate. Also, the use in Hyllarima of a dia-critised Greek lambda for the sound (L ) is a clear indication of an l sound. The geminate nature of l (originally, at least) is supported byits distribution: l never appears in an initial position, which probablyimplies that it needed at least a vowel preceding it (/Vl.l/).

Nasals

m M n N ñ

m n ñ /m/ /n/ ?

m and n represent the labial and dental nasal stops typical for many

phonological inventories of world languages.m is used to adapt Egyptian m: pisma “ k (and variants) = Psmtk .Conversely, Carian m is reflected in Egyptian by m, and in Greek bym: paraeym = Prjm, “ arkbiom = ” 3rkbym, kbjom≤ = Kebivmow, msnord ≤ cf .Masanvrada, etc.

n reflects Egyptian n and Greek n: -nejt, -neit = -Njt, ntokri ≤ = Nj.t-jqr (Nitvkriw), niqau≤ = Ny-k3w (Nexvw ), nik [ ]la- = Nikokl∞w; and Greek ntranscribes Carian n: somne ≤ = Svmnhw , pnu≤ ol , punw ≤ ol ≤ = Ponussvllow,msnord ≤ cf . Masanvrada, etc.

Schürr correctly established that ñ ñ , a letter absent from the Carianalphabet of Egypt, also represents a kind of n, on the basis of the ono-mastic identifications ∞tmño-≤ (2× ) = Ekatomnvw and pñmnn-≤ ñ = Ponmoonnowin the Sinuri bilingual text (E.Si 2), and this has since been confirmed bythe new inscription of Hyllarima, where tñu≤ is found side by side withTonnouw. Note also the alternation ñ/n in ∞tmño≤ vs. (Thebes) ktmno.

Its absence from the Egyptian inventory is a little surprising, espe-cially considering its wide distribution in most of the Carian alphabets

of Caria itself—it appears in such distant places as Hyllarima andKaunos—and the tendency of Egypto-Carian writing to contain lettersthat the alphabets of Caria have lost (j j , v w , for example). Perhapsthe formal resemblance of ñ to z ≤ played a part in its disappearance.



250

Its exact phonetic value is difficult to determine. In most cases, itcan be attributed the value of a syllabic nasal (like Lycian ñ ): pñmnn≤ ñ,

ñmai l omd a, yri ∞ñ, tñu≤ (note the Greek adaptation Tonnouw that could bea reflection of /tn.nu/-), but its use in pd a ∞m≤ uñ weakens the validityof this interpretation. In any case, a functional diff erence between ñ and n seems more probable than an articulatory diff erence: none of the examples seems to support the interpretation of ñ as a nasal artic-ulated as palatal, velar, or the like.

Fricatives

s f F / z Zs “ ≤ /s/ /“/ /ç/?

The exact value of the three fricative sibilants of Carian, and also theirorigin (see below) is undoubtedly the phonological particularity of Carianthat has yielded most discussion.10 In this case, Greek adaptation of Carian names proves to be largely useless, due to the existence in thislanguage of a single sibilant s /s/. The three Carian sibilants are sys-

tematically transcribed as s- /-s(s )-: “ aru≤ ol = Sarussvllow, arli “ =Arlissiw, msnord-“ = Masanvrada, (i)b rsi, i b arsi = Imbarsiw, Imbras(s)iwetc. In the other direction, Greek s is adapted to s in Carian (lysiklas,lysikratas ).

More interesting, however, is the contribution of Egyptian, in whichCarian s is adapted to s and Carian “ to “ : urs ∞le-≤ = 3rskr, “ arkbiom =” 3rkbym. As for the adaptation of Egyptian names, it is particularlynoticeable that ≤ and “ alternate in rendering the sound t /t“/ of theEgyptian name Psmtk ( pisma “ k, pisma ≤ k , etc.), in contrast with the use of

Carian t for the same Egyptian sound in t amou (see below). A possi-ble use of Carian s for Egyptian s could be seen in p∞simt if the con-nection of this form with Egyptian Potasimto (P3-dj-Ó r-sm3-t3wy,P3-dj-sm3-t3wy, Potasimto ) were accepted.

From all this information, certain conclusions can be drawn: (1)Carian s most probably represents the basic voiceless dental sibilant inthe phonological systems of the world, as shown by its use for transcribing Greek s in the two Greek names quoted above; (2) “ is probably a

palato-alveolar voiceless fricative, which can be concluded from the use

10 See Hajnal (1998), Schürr (2001b), Melchert (2002).



251

of Egyptian “ in ” 3rkbym = “ arkbiom; (3) As for ≤ , its use in pisma ≤ k allowstwo possible interpretations: it could be either a sound very close to

Carian “ (perhaps a palatal /ç/, like German ch in ich ), given the alter-nation pisma “ k/pisma “ k , or an a ff ricate sound diff erent from /t“/ (for whichthe letter t already exists), corresponding roughly to the Egyptian aff ricatet . If the latter were true, /ts/ would be a logical solution. In any case,the interpretation as (palatal) fricative seems preferable; if ≤ were torepresent /ts/, we would expect to come across an occasional use of Greek z in the transcription of the u≤ ol -family of names, but in fact,ss is systematically found.

A ff ricates

c C 1 9t z/t“/ /ts/ and/or /st/?

The attribution of the phonological value of t is based exclusively onits use in the bilingual inscription E.Me 7, where it is used in theCarian adaptation of the Egyptian name T3j-jm = w [ ‘amòu] (Tamvw,

Yamvw, Samauw, Samv#w ). Further examples of this letter are extremelyscarce, and no satisfactory onomastic interpretations have been pro-posed for the possible personal names in which it appears.

The value of z was established by Schürr on the basis of Cariannames that appear to contain the name of the Egyptian goddess Bastet( B3st.t ): tt(u)bazi, piub[a]zi (see Chapter 11, ss. vv. ). Doubt remains as towhether z represents a consonantal heterosyllabic sequence s + t (ors + d or similar), or rather a phonetic result of this sequence (mostprobably an aff ricate /ts/ or /dz/). No clear evidence can be found

in the other forms that contain this letter. In the new inscription of Mylasa some possible onomastic identifications have been suggested (seeAdiego 2005) that point to both /st/ and to /ts/ as possible values of z: qzali = Kostvlliw vs. myze cf. Mouzhaw. However, both identificationsare rather tenuous.

3. Letters of Uncertain Value

0 and its possible Kaunian form 8 can represent a nasal + voiced tec-tal stop, but this assumption is based practically on a single piece of evidence: the likely connection of ≤ u0l i ≤ with the place name Souaggela.This tectal value is also suggested by the sequence qrds 8rdso[- in the

252

first line of C.Ka 2, in which a sort of fi gura etymologica formed fromthe word qrds (cf . qrds , C.Ki 1, and qrdsol “ in the same inscription C.Ka

2) seems to exist. The rather conventional transcription g is adoptedhere for these letters.As for %, a letter absent from all the Egypto-Carian alphabets with

the sole exception of Thebes (and also E.xx 7 if the inscription reallycomes from Egypt), we can draw on two rather weak sources: firstly pr %idas in E.xx 7, for which Schürr has proposed a connection withBragxidai, the name of the priests of Apollon in Didyma (near Milet),and secondly ]oml % in C.K 2, which may be related to C.K 5 uioml n.The first case would seem to point to a nasal + tectal value, as for 0,but if the second connection is reliable, perhaps % would in fact berather a type of nasal. Both contexts would then favour a nasal witha tectal mode of articulation (for example, a velar nasal /fl/. Otherexamples are not so clear, but some of them indicate a nasal ratherthan a nasal + stop: the third line of C.Ka 2 shows a construction[-]∞ar l ano% sb z ≤ arios % where % seems to be a morphological ending.If a (tectal) nasal value is accepted, the forms could be interpreted assingular accusatives in n (cf . lusikla-n in C.Ka 5) spelled with a tectal

nasal, for reasons we are unaware of. In any case, the possibility of a/nasal + tectal/ value cannot be ruled out. In such a case, the diff erencebetween 0 and % would come from the point of articulation of thetectal, in other words, both letters would be the correlate of one of thethree voiceless tectal stops that exist in Carian (∞ , k, q ).

I provisionally and conventionally adopt the transliteration <fl> for%, although this is not to say that I consider that attribution of a velarnasal value to this letter certain, or indeed even preferable.

The letter O appears exclusively in the alphabetic variant of Kaunos.

Adiego (2002) proposed that it represents the letter c C t /t“/ of otheralphabetic variants. This proposal was based principally on the dis-tributive properties of the sign, which appears in a final sequence -oO,at least in the word punoO (C.Ka 2; the segmentation is guaranteedthanks to the following word, otr “ , equally segmentable in C.Ka 5). Thissequence can be compared to the endings in -ot from Hyllarima C.Hy1, º pususot , msot , to which we can now add muot , from the new frag-ment of the inscription recently discovered.

An alternative approach, adopted by ”evoro“kin and tentatively sup-ported in recent times by Schürr and Melchert, is to view O as a sortof glide /w/, given its appearance between two o in most examples.This explanation is complicated by the fact that in Kaunos there are



253

three clear examples of ouo sequences, where it is difficult to interpretu as anything other than a /w/ (ºouorº 2×, ºorouoº).

For this sign a transcription t2 is adopted.

4. Letters of Unknown Value

Strictly speaking, there remain only two letters for which no possiblephonological value can be suggested at the present: H and 1. Of thesetwo letters, the first is undoubtedly the most important, as it appearsin Egypto-Carian alphabets (Saqqâra 2×, Abydos 2× ) as well as in inSinuri-Kildara (4× ), in Stratonikeia (2× ), and in Kaunos (7× ). No expla-nation has been found for any of the forms where H is used, not eventhe two clear personal names from Saqqâra (E.Me 28 ps H ÿm[-]≤ , E.Me31 mHm≤ ). Most likely is that it had a consonantal value, given its inter-vocalic context in forms such as ºa Huq [ (C.Ki 1), sb uHbit (C.Ka 2); thisexplanation is consistent with its use in mHm≤ , if a typical defectivevowel notation (ºm(V)H(V)º) is used.

As was noted in the chapter devoted to the Carian alphabet, theapparently complementary distribution of H and c in the alphabets of

Caria itself is curious: where c appears, there is no H (as in Hyllarima)and vice-versa (Sinuri-Kildara, Stratonikeia, Kaunos). One would betempted to interpret H as a local variant of c, a hypothesis also sup-ported by the formal proximity of the two signs, but the existence of both letters side by side in the Egypto-Carian alphabet, and the pos-sibility that the Kaunian counterpart of c may in fact be O (see above),make it difficult to argue the case further.

The situation of 1 is very diff erent: it is only present in the Kaunianalphabet, and only 9 examples (7 in C.Ka 2, 2 in C.Ka 4) can be

found. Purely for the sake of hypothesis, one could assume that it is aparticular Kaunian form for a letter existing in other alphabetic vari-ants, in which case, the only remaining possibility would be to equateit to Egypto-Carian 6 ® . However, this cannot be confirmed by anyof the instances in which 1 is used, and the possibility therefore remainsthat the letter has a phonological value specific to Kaunian.

According to the drawings of ”evoro“kin, the still unedited graffitiof Thebes seems to show two new signs: and Z. Since I do not pos-

sess a definitive and accurate epigraphical edition of this new corpus,I shall simply draw attention to this possible existence (see pp. 103–104).Finally, very little (indeed almost nothing) can be said about the

strange ‘diamond-like’ sign (K, t ) that appears on two occasions (E.Th

0 0



254

28, E.Si 4), in both cases preceding the same word (bebint ). Note thatthe other two examples of the word are not accompanied by this sign

(E.Th 30, E.AS 7), making it highly unlikely that the sign was actu-ally a letter.

5. Phonotactics

The defective notation of vowels makes it very difficult to draw anaccurate picture of Carian phonotactics, since the task of distinguish-ing whether a sequence of consonants actually represents a consonan-tal group, or whether in fact one vowel is graphically missing, is extremelycomplicated. Accuracy is also compromised when resorting to the indi-rect evidence in the Greek adaptation of Carian names, as there is arisk of identifying certain characteristics as Carian when they in factbelong to Greek adaptations of Carian onomastics. The only solution,although far from perfect, is to combine both sources, but the resultsare then incomplete and many gaps and uncertainties remain. Therefore,in the following points, I shall limit myself to pointing out certain traitsthat in principle can be definitively attributed to Carian.

1. Carian seems to share with other Anatolian languages the absenceof initial r-: there is no example of R- in Greek adaptations of Carianpersonal and place names, and the only definite example of initial r inCarian is found the name rtim, in the new inscription of Hyllarima (towhich rtmi of Tralleis can be added if a segmentation s d ia rtmi is pre-ferred to s d i artmi ). In this case, a defective vowel notation (rtmi for/artmi/) or a syllabic r produced by aphaeresis could be the cause/ºtmi/ (for this latter possibility, cf . the similar explanation given byLycian r Mmazata in Melchert 1994:297). In any case, it seems certain

that the possibilty of rV- at the beginning of a word does not exist inCarian.

2. As stated above, there is no definite example of the letter l inan initial position (Adiego 1993a:276). This restriction is consistent withthe geminate origin of the sound represented by l .

3. A similar tendency can be observed when considering d, as evi-dence of (possible) initial d is very scarce. Our glossary only containsthree forms: d ar “ qemorms [, d en, d rual . We should also note that for two

examples a connection with two etymologically related function wordshas been suggested (d en as preposition comparable to Hitt. andan; d rual ,segmented in d = rual , where d would be a preposition comparable toLycian ñte , see Chapter 11 ss. vv. for details). Also in this case—like l —

255

the origin of d from a consonant cluster -nd- is a good explanation forthis restriction: in general, this cluster is preceded by a vocalic sound

(-Vnd- ), which limits the appearance of d < -nd- in absolute initial po-sition, which is then only possible if the vowel disappears secondarilyor is not noted graphically. A similar situation can be envisagedfor b. The (also scarce) examples of an initial b, when etymologicallyclear, point to a preceding Vmb- sequence: b rsi < i b rsi (also attested) <*imbrsi-.

4. Note the following clusters in internal positions, documented bothin Carian and in Greek sources:

-bl- = -bl -: q ÿblsi- = Kublisse›w-b - = -mb-: i b (a)rsi = Imbarsiw, Imbrassiw-kr- = -gr-/-kr- = pikre / pikra- = Pigrhw, Pikrhw-∞s- = -j-: ly∞se /ly∞si- = Lujhw-ks- = j: uksmu, wksmu-≤ = Omajamoaw-mn-/mñ- = -mn-: ktmno-, ∞tmño = Ekatomnvw, somne = Svmnhw-ñm- = - nm- = pñmnn- = Ponmoonnow-rd- = -rd-: ardybyr- = Arduberow-rl- = -rl -: arli “ = Arlissiw-rm- = -rm-: armo- cf . Ermapiw-rn- = -rn-? alos ∞arnos , cf . Alikarnassow?-rq- = -rg -/-rk- : yrqso- = Urgosvw; trqud e = Tarkonda[-rt- = -rt-: artay- = Artaow, Arthumow-rs- = i b arsi = Imbarsiw-tb- = -tb-: qtblem- = Kotbelhmow, Kutbelhmiw-d - = - nd--d b- = - ndu-? ki d bsi- cf . Kinduh?

Other internal clusters seems also to have existed: for instance -rk- in“ arkbiom, or -dr- (note idrayridsemd ? bq, uodrou, uodryia [ in Carian and ÉIdrieÈwin Greek sources).

5. For initial consonant clusters, the evidence is less certain: unlikeinterior clusters, there are no clear examples appearing simultaneouslyin both direct and indirect sources, with the exception of k ≤ at ÿbr = Lyc.Janduberiw (Zgusta KPN § 1061), which points to an initial k ≤ - (cf . alsoJermedu<be>row as an example of initial k + sibilant cluster). Some evi-

dence does exist of an initial cluster such as kb- (Carian kbidn = Lycian Xbide ), and—although very limited—for clusters of the type s + stop(skdub rotoz ≤ , sqla, sqlumidun besides sp-, sk-: Skoaranow, Spareudigow ).Greek sources also contain several cases of stop + liquid beginnings



256

(br-, gl -, kr-, pl -, pr-, tr-),11 but there are no definitive examples inCarian, in the sense that practically all the examples can be alternatively

interpreted as sequences of stop plus syllabic liquid (note for example pr fl idas, prpwri ∞ , trqud e ) or as cases of defective vowel notation (qla l i- =Greek Kolaldiw, etc.).

B. O H P C

The content of the following pages must be considered purely provi-sional. Our present knowledge of Carian is such that we must be cau-tious when attempting to sketch an overview of this kind. It must benoted that most of the information that can be obtained from Cariansources comes from Carian proper names, and onomastics is not alwaysa good tool for comparative research: firstly, because the etymologicalinterpretation of proper names can only be based on formal criteria,since they lack an actual meaning; secondly, because we cannot be surethat all the personal names correspond to the actual language, as someof them could come from nearby languages, and others could betray

the conservation of old traits, modifications caused by analogical processes,and other such characteristics. In any case, both problems can beavoided to a certain extent when the volume of evidence and internalconsistencies makes it logical to interpret them as a true reflection of Carian language. But I deem it necessary to issue this warning whenone comes to evaluating the information that follows.

I adopt the reconstruction of Proto-Anatolian (PA) as it appears inthe fundamental work of Melchert (Melchert 1994), which currentlyrepresents the most useful tool for comparison.

It is impossible to give a complete account of the outcome of PA sounds in Carian, with particular difficulties being found in the vocal-ism. Evidence is extremely limited in many cases, due to the defectivevowel notation, and the complexity of the treatments of vocalic soundsfrom PIE to PA and from PA to the particular Anatolian dialects. Inother cases, the evidence is simply nonexistent. Here I shall merelyindicate the more relevant aspects that can be identified from our cur-rent knowledge of Carian.

11 There are, however, few examples.

257

1. Vocalism

(1) One of the few traits that can be established is of great relevanceto the position of Carian, since it supports the theory that Carianbelongs to the group of “Luwic” dialects of Anatolian (Luwian, Lycianand Milyan). This is the change of PA *è to /i(:)/. This is demonstratedabove all by the root pik-/-bik- < PA */be:H/- < PIE *bhè h2- ‘to shine’,present in diff erent Carian names (dbiks, pik(a)rm, pikre . . . ).

(2) A unique characteristic of Carian is the emergence of a roundedfront vowel y, ÿ /y/ and its semivocalic counterpart /w/. This is a con-ditioned change, but it is impossible to identify all the precise contexts

in which it takes place. However, at least a trigger for the fronting seems clear: the immediate contact of original /u(:)/, /w/ with /i/,/j/. We come across examples such as yiasi, yi ≤ ∞?biks ∞i , [—]ryin,uodryia [, yjas[i ≤ ], “ÿin≤ (2× ), and ∞diye ≤ , siyklo≤ , i ÿkr ≤ that point to */wi/> /wi/, * /uj/ > */yj/, */iw/ > /iw/, */ju/ > /jy/. Particularly strik-ing is the Kaunian alternation yoml n (C.Ka 4) / uioml n (C.Ka 5). It isvery likely that both spellings are attempts at representing /wi( j)o/-.The absence of a specific grapheme for the semivowel /w/ in the alpha-bet of Kaunos would explain the graphical oscillation. Note that ui isonly attested in Carian in this word, and that there are no examplesof wi, uj in the entire Carian documentation. It is also likely that otherexamples of y, ÿ before a vowel other than i should be explained in asimilar fashion, as yoml n: idyes ≤ , ∞aye, ter ÿez ≤ , ÿas d ≤ . Note particularly thelast example: it is tempting to bring it closer to yiasi, yjas [ i ≤ ] and toreconstruct a protoform */wijas(V)nd/-.

For the remaining examples of y, ÿ, the case is not so clear, but theinfluence of a near i, j can be envisaged: in the family of names in

ydiq/yriq (see below pp. 262–263), a sort of metaphony caused by thei in the following syllable could be suggested. Even in ylarmi t , one couldimagine a more distant assimilation, triggered by the final i , or ratherby the (un-notated) i present in the Greek form of the place name,

Ullãrima (*/ularim/º > /ylar(i)m/º. In the case of ethnics formed withthe Luwic suffix CLuw. -wanni , Mil. -wñni, kbdyn“ , mdayn/mda ÿn, the con-nection with [—]ryin cannot be overlooked, but the exact interpreta-tion of y/ ÿ in these forms remains unclear: is y/ ÿ here used for /wi/,like in yoml n? Or rather has /wi/ coalesced into a single vocalic sound

/y/? A third possibility would be to ascribe the fronting process tometaphony, by postulating a *-uni- > -yn(i) evolution.

258

(3) Another vowel that seems to have appeared secondarily in Carian(at least for some words) is o. It appears to come from */a:/ and also

from */a/ in accented syllable (where */a/ would become a long vowel,a change that is typical in Anatolian languages. Forms such as armo- (inarmotrq d osq , E.Hy 1a) < *armà-12 would also suggest this process. A sim-ilar explanation could be given for ntro < *n(e?)tr à (cf . perhaps for theformation Lyc. kbatra < ºtr + à ). In general, Carian names with -o couldbe explained as original stems in -à: note particularly plqo and ksbo, where-o is clearly accented. The latter form can be etymologically related tothe Lycian x ahba , ‘grandchild’, also an à-stem. Finally, Kaunian otr “ ,‘themselves’, as well as Lyc. atra-, HLuw. atra/i- also point to an accenteda > o, although the etymological origin of these words remains unclear.

This leads to an interesting explanation of -ol (Greek -vllow, -vldow )names: Assuming Melchert’s explanation that Luwian -alla- forms comefrom*-élo- with ’op’s law (*-élo > *-álla- ), in Carian a lengthening of á and subsequent change to o could have taken place *-élo- > *-állV- >*-àllV- > *-÷oll(V), spelled -ol (with l for, or the result of, geminated l ,see above p. 249). The forms with -e l ( para-i b re l ≤ , ione l ≤ ) ought to beexplained then as the result of a suffix with i -motion followed by

metaphony caused by i (see below): *-éli- > *-álli- > *-é(:)ll(i)- > *e l .In the case of -on in mwdon≤ , a monophthongization process had beensuggested (see Adiego 1994:94) *mwdawn≤ > mwdon≤ , but I now preferto look for another solution, given the form mda ÿn/mdayn, where themonophtongization process has not taken place. Perhaps o here alsorepresents * à, which in this case comes from the contraction *ºa-wan- >*ºa-an- (with loss of w ) > *ºàn- > ºon-. In mdayn/mda ÿn, the lack of con-traction can be attributed to a diff erent vocalism of the suffix (*ºa-wen- ),whether original or originated by metaphony (*ºa-wen-i- vs. *ºa-wan-

asin mwdon≤ ?). The other examples of possible monophthongi-zation cited in Adiego (1994:49) (arliom≤ < *arliya-uma , etc.) are by nomeans conclusive. It could also be argued that they come directly from* à (> o ).

12 That arma was an a -stem in PA seems the most likely explanation, although theevidence is not certain: note Lyc. N304, 5 ar Mma , wherein the isolated context of theform does not allow us to confirm that in Lycian it was also an a -stem (Melchert, DLLs. v. ).

259

13 Melchert (1994:296).14 For the probable non-Doric origin of the form (according to Blümel), see above

p. 237. The non-existence of e in Kaunian is sufficient to explain the adaptation of Greek h to a .

(4) The vocalism of the words en, ‘mother’, and ted , ‘father’, is a clearindication of an umlaut process a > e / _.Ci similar to that of Lycian:13

PA sg. nom. *anna-s (cf . Hittte anna-“ ) > Luwic *anni-s (with i -motion,cf . CLuw. anni-“ ) > carian en, like Lycian e ˜ni ; PA *dáda-s > Luwic *tádi-s (with i-motion, cf . CLuw. t àti-“ ) > Carian ted , like Lycian tedi . Cf . thesimilar explanation above for -e l .

(5) The form otonosn shows a strange ‘vocalic harmony’ in o. At leastthe second and third vowels can be explained with *à > o (*ÉAyhn*a- >*at ànà-,14 cf . Lycian atãnaze/i ) > *atono- ), but the first seems to be causedby metaphony (*atono- > otono- ).

2. Consonants

(1) The PA labial and dental voiceless stops remain unaltered inCarian:

PA */p/ > Carian p /p/: para(eym), para(i b re l ≤ ), Para- < *PA pr ò /pr ò :cf . Hittite par à ‘forth’. Cf. pun-/pnin punw ≤ ol ≤ , pnu≤ ol , perhaps also in

punot 2, < Luwic puna- ‘all’ (Lyc. punãma- ‘totality’, CLuw pùna- ‘all’).*/t/ > Carian t /t/: trqud -e, trq d -os < PA *T º H –t- ‘Storm-god’ (CLuw

Tar ¢unt-, Lyc. Trqqñt- ). -t < -te or -ti < PA pret. 3rd sg. *-to or pres.3rd sg. *-ti in ÿbt < *ubete ‘off ered’ or < *ubeti ‘off ers’, cf . Lyc. ubete .(2) Like the other Luwic dialects, Carian presents a ‘satem-like’ treat-

ment of PIE, PA *∞, as can be seen in the demonstrative pronoun sa-/sn- ‘this’ in sa, san, snn < PA *∞o- (Hitt. ka-, Luw. za- ). Perhaps also si d i,s d i ‘tomb’, if it can be connected with PIE *∞ei- ‘to lie’ (Lyc. sije- ).

(3) Luwic also deals with * · > Ø in * · emro- ‘steppe’ > *imr- >*imbr- >i b r-/b r-: (i)b r-si < *imbrV- < * · emr-, (para) -i b re l -≤ < * · emréli-.

(4) The only clear example of the treatment *k w is the (original) rel-

ative pronoun ∞i < PIE, PA *k w is (Hitt., CLuw. kui “ , Lyc. ti , Mil. ki [ci]). in this case Carian displays a process similar to Milyan: delabi-alization and fronting before i .

(5) Forms like tedi < PA *dáda/i- or pik- (in pikre-≤ , pikarm-≤ , etc.) inaddition to dbiks and dbikrm point to an ‘unvoicing’ of voiced stops ininitial position, a process that has also taken place in parallel to thisin other Anatolian languages (cf . Lycian tedi , Lydian taadas ).

260

(6) In the sequences of Nasal + voiceless stop, this latter examplehas become, as in Lycian (and probably in the rest of the Anatolian

languages), a voiced stop. The evidence is clear for dentals: *T º H –t- >trq(u)d -, with *-nd- > d, see above.(7) Traits (5) and (6) clearly indicate that the situation for Carian is

very similar to that of Lycian: voiced stops merge with voiceless stopsin initial position, become fricative in intervocalic position, and remainas voiced stops only after a nasal, where voiceless stops merge with them.

(8) Carian liquids l, r and nasals m, n come from the respective liquidand nasal sounds in PA and/or in Luwic:

r < *r: para(eym), para(i b re l ≤ ), Para- < *PA pr ò /pr ò : cf . Hitt. par à‘forth’, armo < PA *armà- ‘moon’, “ ar-, “ r- ‘upper’ (in “ ar-u≤ ol , “ r-wli-≤ ,etc.), cf . Hitt. “è r , CLuw. “ arri , Lyc. hr-i , Mil. zri-

l < *l: wljat. Cf . Hitt. walliwalli- ‘strong, powerful’m < *m: msn-ord ≤ . *msn- ‘god’: Cf . in the rest of Luwic languages:

CLuw. mà““ an(i)-, Lyc. maha(na)-, Mil. masa-, Sidetic ma ≤ ara (pl. dat.)‘gods’; *mu- (in uksmu- ) = CLuw., muwa- Mil. muwa- ‘might, power’.

n < *msn-ord ≤ < msn- ‘god’, see immediately above; -n acc. sg. end-ing: lysikla-n, ork-n, etc. < PA *-n < PIE *-m; -yn- suffix for ethnic names

(kbd-yn-“ ‘Kaunians’) = CLuw. -wanni-, Lyc. -ñni-, Mil. -wñni-.As for ® , l , ñ , see the remarks in pp. 248–250 above. l seems topoint to a geminate *-ll- (*-élo > *-állV- > -ol , see above). ® can be apalatalized r , which in some cases could come from l (qdar ® ou≤ , cf . CLuw¢utarl à- ‘slave’?). ñ comes from n at least in the ending -ñ of pñmnn-≤ ñ (acc. sg of a possessive adjective).

(9) PA ‘Laryngeal’ *H (PIE < *h2 ) appears as a tectal voiceless stopk, q in Carian. The process is therefore parallel to that of Lycian (<x>/k/ / <q>). Examples: for *H > k: k d ou- (< *Hntaw º-, cf . Lyc. x ñtawat(i),

from PIE *h2ent- ), pikre-, piks- (both containing PIE *bhè h2- > Luwic * piH- );for *H > q : trqud e, trq d - < Luwic *T º H –t- (CLuw. Tar ¢unt- Lycian,Milyan trqqñt- ). If the word quq comes from PA *HuHo- ‘grandfather’(> Lycian x uga ), the lenition process seen in Lycian ( x uga = /kug a/) isabsent, at least graphically, in Carian. But the Greek form Gugow pointsto a voiced articulation (the initial G- remains unclear).

(10) The existence of at least three fricative sibilants in Carian (s, “ ,≤ ) poses a puzzling situation regarding their respective origins. It seems

that (at least some) results of the three sibilants come ultimately from thesingle fricative voiceless dental fricative PA *s . Recent work by Melchert(see Melchert 2002) has significantly clarified the matter, and new evi-dence from the Hyllarima inscription seems to corroborate his views.

261

Concerning Carian s , Melchert (2002) has stated that it reflects aprehistoric simple *s in forms such as i [—]inis (C.Ka 5), ntros, “ arnais ,

etc., in which he claims to have identified genitive-dative forms coming from PIE *-e/oso (Melchert 2002:309). This hypothesis has essentiallybeen confirmed by the new inscription of Hyllarima, wherein an analy-sis of the ending -os in armotrq d os would suggest that it originates froma genitive ending *-Vso, or from an adjective suffix *-Vso-.15 Anotherpossible new example of Carian s from *s is the name ksbo in C.My1, if it is related to Lycian x ahba ‘grandchild’ < PIE *h2onsu secondar-ily converted to an à-stem.

As for “ , the clearest examples for a conditioned origin are “ (a)r- andacc. pl. -“ (in otr “ , kbdyn“ , sarni “ ). The first example could be interpretedas the result of a palatalisation process caused by direct contact withr (cf . perhaps also zri-, not sri- in Milyan). One must therefore startfrom *sri- (cf . Mil. zri-, Lyc. hri-, CLuw. “ arri ) and to postulate *sri- >*s º - > “ º - (palatalisation) > “ ar- (samprasàra»a):16 the second example isthe outcome of Luwic accusative plural *-ns (> -“ ). The parallelismswith the use of z in Milyan are striking: there we find *sri- > zri- (zri- gali, zriqali ) and nom.-ac. pl. -z < *-ns (masaiz, Xbadiz, x uwasaz, etc. ) vs.

Lycian hri , (pl. acc.) -s . Other instances of “ are not so clear (mol “ , k “ ow “ ).Finally, in the case of “ , Melchert proposes that Carian possessive-genitive -“ continues the PIE possessive suffix *-asso- in its form -ass ì ,i.e. with i -mutation (Melchert 2002:311). From a phonological point of view, the supposition is extremely compelling, given the likely palatalcharacter of ≤ /ç/ (see above), although some doubts do remain.17

3. Some Secondary Changes

In this brief section I off er a succinct analysis of several cases of sec-ondary changes observable in the Carian documentation. Once again,the material and results are far from conclusive, and must therefore beviewed with a certain amount of caution.

15 The exact analysis of -s and its actual value in Carian is not relevant here, seepp. 314–317 for the problems posed by the Carian s- ending.

16 I consider this interpretation of “ in “ (a)r- as preferable to Melchert’s use of Lydian

serli-/selli- (note that Lydian <s> = /“/!) for explaining the palatalization. The devel-opment of a secondary support vowel in contact with syllabic r is also visible in pikarm-≤ vs. pikrm-≤ (from pikrº, cf. pikre-≤ ) and the Milyan testimony seems then more compelling than the need for postulating that Carian “ ar- comes directly from *ser-.

17 No explanation has yet been found for the form i b (a)rsi- if it comes from an -ass ì -

262

1. The family of names b rsi/i b (a)rsi , ( para- )i b re l , Greek Imbras(s)iw,Imbarsiw, Imbarhldow, if the etymology from PA* · emro- is accepted,

would indicate a sound change *-mr- > -mbr- ( ), a kind of epentheticdevelopment well known in other languages (cf . Greek êmbrotow < *amro-tos < PIE * –-mº to-s , Spanish hombro [ombro] from Vulgar Lat. *umru(m)< Lat. umerum ).

Perhaps a similar process can be identified in the names Andarsvw,Androssvw (cf . also d ar “ ?): it is tempting to start from *naras à- (to berelated to Narasow, epithet of Zeus in Panamara, Caria)18 > *–ras ò - >*–dras ò > * –d º s ò > Andarsvw, Androssvw.

2. The examples mentioned above (1) point to a secondary charac-ter of a in i b arsi- (Greek Imbarsiw ), given that epenthetic b could onlyappear in direct contact with the following r . Cf . also Andarsvw, d ar “ if the explanation proposed here is accepted. It seems that in Cariana kind of samprasàra»a could have taken place when r became syllabic,and the entire process could therefore be the following: *imrasi- >* imbrasi - > *imbº si- > /imbarsi/º (i b arsi- ). The same rule can be appliedto alternations such as pikrm-/pikarm-, and “ ar-/“ r- (cf . above for the eff ectof this explanation on the merging of palatal “ in these latter forms).

3. It is plausible to imagine a common origin for the collection of stems ÿdiq/ ÿd ∞-/yriq-/yri ∞-/wri ∞- if one accepts a progressive dissimila-tion process. The following are all the forms that have been identifiedthus far (E = in Egypt; C = in Caria):

(perhaps re-derived by -*iye in *-assiye- ) suffix, because the phonological context is very

similar to that assumed for the origin of -≤ . The possibility that these forms contain adiff erent suffix (*-∞o- or *-tyo-, see Melchert 2002:310 n.13) cannot be ruled out, butit is a more ad hoc solution.

18 For this form a connection with CLuw. annara/i- ‘strong’ can be envisaged (Neumann1994:22; see here p. 333).



263

-d- -r-

Greek ÿdiq- ÿd ∞- Greek yriq- yri ∞- (wri ∞-?) yri ∞-ñ (?) Cidyri ∞-≤ C

Paraudigow par ÿd ∞-≤ E paryri ∞(-≤ ) C

prpwri ∞? Erather prp ÿri ∞?

“ a ÿdiq-≤ E Saurigow “ ayriq ESenurigow

Spareudigow

Semeuritow?

It is possible that the original form of the stem was * yriq-/yri ∞-,19 asthe non-compound form yri ∞-ñ suggests. The forms with d are limited to par(a)-, “ a- and Spare-compounds, and for the two first, the correspond-ing forms with r are also attested. If one assumes that “ a-/Sa- is a vari-ant of the well-known adverbial stem “ ar-/“ r-, originating from a loss of r in intervocalic position (* “ ar-yriq > “ a-yriq , cf . perhaps Saussvllow besidesSarusvllow ), the resulting forms in d in compounds with par(a)-, “ a(r)-

could be explained as a dissimilation r-r > r-d: par-yri ∞- > * par-ydi ∞ ( par- ÿd ∞- )-, * “ a(r)-yriq (cf . “ a-yriq ) > * “ a(r)-ydiq (“ a ÿdiq- ).

4. Schürr has argued in favour of a change p > Ø from some allegedexamples of alternation (Schürr 1992:141). Perhaps the most compelling example of this possible alternation would be the words for ‘stela’ (orsimilar) in Memphis: upe/ue , although it is also possible that we are infact dealing with two diff erent words. Less convincing is the name paraeym vs. parpeym, where a diff erent second element (ºeym/ºpeym ) could

also be suggested.20

No other clear examples are known of, either inCarian direct sources or in Greek ones.5. Other examples of vowel and consonantal alternations, in Carian

direct sources or in Greek indirect sources, are far more occasional,and not easy to explain. Note for example a/e in pikra/pikre, upa/upe ,which could be a morphological rather than a phonological alterna-tion. For changes detectable in Greek sources, see Neumann (1994:18–19).

19 The alternation q/∞ remains unexplained (cf. supra pp. 244–245). As for y/ ÿ, seeabove pp. 235–236.

20 A further example, qarpsi- vs. qarsi-, off ers a totally diff erent context, and cannotbe compared to the two cases mentioned above, where p would be lost between vowels.



CHAPTER SEVEN

ANALYZING CARIAN INSCRIPTIONS

Contrary to the customary order in grammars, where morphology,together with phonology, precedes syntax, in the case of Carian it seemsmethodologically more accurate to begin with the analysis of Cariantexts; our knowledge of Carian morphology depends on the way in

which the texts can be interpreted syntactically, and such an interpre-tation remains in most cases controversial, to say the least. In the fol-lowing pages I will try to analyze Carian texts from the ‘easiest’ to the‘most difficult’, beginning with those that contain only very basic ono-mastic formulae. The following step will be to analyze the inscriptionsof the Memphis sub-corpus, where we find more complex onomasticformulae, but no recognizable verbal forms are attested. This analysiswill also allow us to identify some common nouns used in the formu-

lae of these funerary texts. The third section will off er an analysis of some brief inscriptions (mainly on objects) that seem to contain formsother than onomastic formulae. Finally, a few aspects of the interpre-tation of the longer inscriptions, where there are serious difficulties of analysis, will be briefl y addressed.

A. B O F

1. Inscriptions Consisting of Only an Individual Name

The briefest Carian inscriptions consist of a single word, which in thegreat majority of cases can be confidently interpreted as an individualname. As one can easily imagine, these types of texts are found among the graffiti and also on some objects. In the graffiti, the name appearswithout an ending, in a case that we can consider the ‘nominative’, alogical explanation given that the inscription can be interpreted merely

as a type of signature: pisiri (E.Ab 1), piew (E.Ab 38), pla ?

t (E.Th 3) ,wljat (E.Th 7), psma ≤ k (E.Si 7, E.Bu 5) etc. Only occasionally does the



265

name appear with the -≤ ending, commonly interpreted as a ‘genitive’:E.Ab 32 ∞arr ≤ , E.Ab 37 “ arur ≤ .1

Also, in the so-called ‘pharaonic objects’ (excepting funerary stelae)published in Masson-Yoyotte (1956) we come across inscriptions con-sisting of only an individual name, in nominative or in genitive: wliat (E.xx 2) pduba (E.xx 4,) vs. ione l ≤ (E.xx 3).

The only possible examples of a form in another case—leaving asideforms with no clear interpretation—are provided by the two identicalinscriptions on bracelets that contain the word k d u≤ ol “ , where the final-“ could represent a diff erent case ending.

2. Inscriptions Consisting of Only a Twofold Onomastic Formula

Despite the above examples, the most typical Carian onomastic for-mula, as in other ancient Indo-European languages, seems to have beena twofold one: individual name + father’s name. This latter is system-atically expressed in Carian by the genitive ending ≤ .2 This formula canexplain a number of Carian inscriptions consisting of two words, aswell as others in which a particle ∞i appears together with the two

names.We can attempt a typological classification of twofold formulae onthe basis of the case of the first member and the presence or absenceof ∞i :

Nominative + Genitive panejt iarja≤ E.Ab 2ptn“e | ibarsi≤ E.Ab 3“amow ltari≤ E.Ab 4, 5pl at | pals≤ E.Ab 7, 8, 9

piubez qurbo≤ E.Ab 10untri | uantrpu≤ E.Ab 13pdubez or≤ E.Ab 15tamosi | inut≤ E.Ab 18tamosi utnu≤ E.Ab 19ninut | tamosi≤ E.Ab 20ttubazi katt ÿri≤ E.Ab 25iall i | q∞blio≤ E.Ab 40

1 The interpretation of the ending -z (-≤ ) as a genitive was first made by Sayce

(1887[92] = 1893:141–142).2 For the use of an ending -s- in C.Ka 5, see below p. 316.

266

ttbazi kt?tri≤ E.Ab 41d ÿbr | t®∞atr≤ E.Th 5

psma≤k ibrsi≤ E.Bu 4 Nominative + Genitive ∞ipdnejt q ÿri≤ ∞i E.Sa 1pnu≤ol zmu≤ ∞i E.Me 19(a) “enurt (b) p∞simt≤ ∞i E.Me 50pl att sla ÿ≤ ∞i E.AS 5

s–ending (?) + Genitive ∞iap[---]ws a[rb]ikarm≤ ∞i E.Me 23idmns | myre≤ ∞i E.Me 33b, cf . E.Me 33a

As the examples show, the twofold formula is typically used in graffiti,particularly those from Abydos. In practically all of the quoted exam-ples of graffiti, with the sole exception of E.AS 5, the formula consistsof the individual name in nominative followed by the father’s name ingenitive. E.AS 5 is the only one that contains the genitive accompa-nied by a postclitic ∞i . In this and the other cases where the structureN-Ø N-≤- ∞i appears (E.Sa 1, E.Me 19, E.Me 50), there is no visiblesemantic diff erence when compared to the structures without ∞i . The

role of ∞i in all of these cases seems to be merely to connect the nom-inal complement in genitive to the name it refers to. The most likelyexplanation, as Hajnal (1997a) suggests, is that this construction pre-supposes and/or comes from an elliptical word for ‘son’: pdnejt q ÿri ≤ ∞i ‘Pdnejt, the (son) of Q ÿri- = ‘that of Q ÿri-’.3 Given that from an ety-mological point of view, an origin of ∞i from a Proto-Anatolian rela-tive pronoun *k w is (< PIE *k w is ) is a convincing interpretation, it is easyto assume that behind pdnejt q ÿri ≤ ∞i and similar constructions, therewas an original meaning ‘Pdnejt, who (is the son) of Q ÿri’.

As we can see, twofold formulae are very scarce among the funer-ary stelae of Memphis, but the examples are very interesting in thatthey show another pattern of the formula; as well as N-Ø N-≤ (3 exam-ples), we can identify two instances with possible ‘s-endings’ (N-s N-≤,both with the genitive followed by ∞i ), but we should bear in mind thepossibility that we are in fact dealing with nominatives of s -stems (onthis problem, see pp. 314–317).

This twofold formula ‘individual name + father’s name in genitive’

is easily recognizable in longer inscriptions, which we will address below:

3 We can conclude that the second name represents the father’s name from theEgyptian text that accompanies the Carian one, where P3-dj-Njt is mentioned as theson of K3rr = q ÿri-.



267

in votive texts such as the inscription on a bronze lion (E.xx 7 uksi wrm≤ ), or the inscription on a phiale (C.xx 1 “ rquq qtblem≤ ), in the list

of names from Mylasa (C.My 1), and even in the mention of the twosatraps Idrieus and Ada ([--]ryin ∞tmño≤ sb ada ∞tmno≤ ‘[--]ryin of Hekatomnosand Ada of Hekatomnos’).

B. T S S M

The onomastic formulae limited to the two basic constructions men-tioned above (‘Individual name’ or ‘Individual name + father’s name’)are extremely scarce in the Memphis sub-corpus: to the examples men-tioned above (four, and only for twofold formulae), we can add onlyone other in which the individual name in the genitive is accompaniedby the father’s name and one of the formula words in Memphis, ue ,used to designate the object (‘stela’; on this word see Chapter 11, s. v. ).

arli“≤ | psikro≤ ue E.Me 51

In general, the stelae from Memphis, insofar as they are integrally (or

almost integrally) preserved, show more complex onomastic formulae.It is common to find threefold formulae, as well as another type of formulae that includes more proper names.

1. Threefold Formulae

To the ‘individual name + father’s name’ formula, a third word ingenitive can be added. The simplest interpretation of these threefoldformulae would be to take the third name as the grandfather’s name

(papponym). As we will see below, this interpretation is certain in casessuch as the threefold onomastics formulae of the new inscription of Hyllarima (C.Hy 1a), but in the Memphis sub-corpus, where this typeof formula is very frequent, (in fact, it can be considered the typicalonomastic formula of the funerary stelae) the situation is not so clear.

In Memphis we find the following three types of threefold structures:N-Ø N-≤ N-≤, N-≤ N-≤ N-≤ and N-s N-≤ N-≤ (with or without ∞i afterthe second and/or the third name). When the individual name is in

genitive, a word for ‘stela’ (ue, upe,4

wpe, upa ) can appear (cf . above the

4 In one case (E.Me 26) accompanied by a demonstrative pronoun: upe sa ‘this stela’(on sa , see Chapter 11, s. v.).



268

example of E.Me 51), so that the use of the genitive simply indicatesthat the stela belongs to the individual it mentions.

The inscriptions consisting of only these kinds of threefold formulae— I leave aside for now the inscriptions of a more complex structure— are the following:5

Nominative + Genitive + Genitive uksmu | lkor≤ | mrsi≤ E.Me 2tamou tanai≤ qarsio[-?] ? E.Me 7irow | pikarm≤ | mwdon≤ E.Me 14u≤ol | mi∞≤≤ kdûsi≤ E.Ab 35

Nominative + Genitive∞i + Genitive

“aru≤ol pl eq≤ ∞i : ≤ugl i≤ E.Me 30qorb | isor≤ ∞i | ≤ugl i≤ E.xx 1

Nominative + Genitive ∞i + Genitive ∞iuqsi | “rwli≤ ∞i | mwdon≤ ∞i E.Me 20plqo | pikrm≤ ∞i | mwdon≤ ∞i E.Me 40

Genitive + Genitive + Genitive ttbazi[ ≤ ] | piub[a]zi≤ | aor[ ≤ ] E.Me 1pikre≤ ue “arwljat≤ msnord≤ E.Me 3arli“≤ urs

∞le≤ ki

dbsi≤ E.Me 15

[. . .]u≤ | upe sa | triel≤ | mrsi≤ E.Me 26s[--]et≤ | [--] | ynemori≤ | mwdon≤ E.Me 29me®≤ | somne≤ | t®∞ata[r]≤ E.Me 34| or≤ | wpe | qdar®ou≤ | t®∞atar≤ E.Me 41

Genitive + Genitive ∞i + Genitive arli“≤ : upe : arlio[m≤ ] ∞i : yjas[i≤ ] E.Me 9wksmu≤ | wpe | lkor≤ ∞ j qarpsi≤ E.Me 36[--]j[-]≤ [ - ]owt≤ ∞i : msnord≤ E.Me 48

Genitive + Genitive + Genitive∞i

punw≤ol ≤ : somne≤ q ÿblsi≤ ∞i E.Me 21(a) ÿasd≤ | yi≤∞?biks∞i≤ (b) mwdon≤ ∞i E.Me 46

Genitive + Genitive ∞i + Genitive ∞i“dtat≤ | upa | w | wet≤ ∞i | mwdon≤ ∞i E.Me 13sanuq≤ | ue | pntmun≤ ∞i mwdon≤ ∞i E.Me 28

s–Ending + Genitive + Genitive ∞intokris | dw≤ol ≤ | mwdon≤ ∞i E.Me 35

5 I add to this list an example from Abydos (E.Ab 35), and also the inscription of unknown origin (but clearly close to the Memphis stelae) E.xx 1, which is consistentwith this type of threefold formula.

269

The possibility that the third word is not a papponym can be deducedfrom at least three factors: firstly, the very large number of examples

of the word mwdon≤ in that position, a word for which there are noparallels among the Carian personal names in Greek sources; secondly,the iteration of some words in that very position (≤ ugl i ≤ , msnord ≤ , mrsi ≤ );thirdly, the fact that these words never appear as a first name in anyonomastic formula, and even their appearance as a second name orpatronym is limited to the word t ® ∞atar ≤ , which appears in E.The 5 asa clear patronym (d ÿbr t ® ∞atr ≤ ). These distributive properties do notfavour the interpretation of most of these forms as simple personalnames functioning as papponyms.

An alternative interpretation that was envisaged some years ago isto classify them as ethnic names.6 This possibility is very clear in thecase of mwdon≤ ; there are few doubts that mwdon≤ is the genitive cor-responding to the nominative mdayn, mda ÿn also found in the Memphissub-corpus, and following the discovery of the Kaunos bilingual inscrip-tion, -yn/- ÿn has been confirmed as a suffix for the formation of eth-nic names (kbd-yn-“ ‘Kaunians’).7 The exact meaning of mwdon≤ is adiff erent question, for which I refer to the Glossary (Chapter 11), where

a discussion of the various proposals of interpretation is off ered. Fornow it is sufficient to state that it could simply mean ‘foreigner’ or, if it refers to a concrete place, that this must be the main point of originfor Carian mercenaries in Egypt, given the high number of occurrences.

For some other third-position words, the possible connections withwell-known place names have not gone unnoticed (see Janda 1994:174–176; Melchert apud Adiego 1995:20; Adiego 2004:310): ≤ ugl i ≤ <Souaggela, ki d bsi ≤ and/or k d !usi ≤ < Kinduh, yjasi[ ≤ ] < Iasow, msnord ≤ <Maosanvrada, ksolb≤ < Kasvlaba, q ÿblsi ≤ < Kubliss/ow/. There is also

the interesting possibility of recognizing certain suffixal formations: -si-in ki d bsi ≤ , k d !usi ≤ , and also probably in mrsi ≤ , qarpsi ≤ ; -i- in yjasi[ ≤ ] and

6 In Adiego (1993:212), the hypothesis formulated by Meriggi (1980) of interpreting mwdon- as an ethnic name, was already taken into consideration, but the first authorto propose that a number of third names in onomastic formulae could be interpretedas ethnic names was in fact Janda, see Janda (1994:174–176). The idea was also takenup by Melchert (apud Adiego 1995:20) and further developed in Adiego (2004:309–310).

That I did not mention Janda’s work in this latter paper is a regrettable oversight, forwhich I ask forgiveness.7 For mwdon“ as an ethnic name, see Adiego (1993:212). For integration of mdayn/

mda ÿn — mwdon“ into the same paradigm and for the identification of the Luwic ethnicsuffix, see Melchert (1993:82–83).



270

q ÿblsi ≤ . For -si-, a connection with Lycian -zi- is likely. For -i-, cf . Lyc.-i(je)- (in Tr ◊ mmili(je)- ‘Lycian’), and above all the Carian form ylarmi t ,

which seems to contain an -i- suffix attached to the place name * ylar(i)m(a)-= Hyllarima in order to express the meaning ‘Hyllarimean’. In thecase of msnord ≤ , ksolb≤ , no suffixation can be recognized, perhaps dueto the defective vowel notation. It is rather puzzling that none of theseforms, leaving aside mwdon≤ , show the typical -yn-/-on- ethnical suffix,but this could be due to chance (cf . the great variety of ethnical suffixesalso present in Lycian).

A new argument can be added to this evidence for ethnic names:there are three inscriptions in which a N-Ø N-≤ N-Ø formula is found,which means that the third name can hardly be a papponym, since itagrees in nominative case with the first name. Moreover, in two of thethree cases, the third word can be interpreted as an ethnic name:

“ayriq | parpeym≤ ∞i yiasi E.Me 25idmns | myre≤ ∞i | mdayn ∞i E.Me 33atriqo : parma≤≤ ∞i klorul ∞i E.Me 6

In the first example, read thus,8 we find the word yiasi , which cannotbe anything other than the nominative form, referred directly to theindividual name “ ayriq , of the word that appears in E.Me 9 as yjas [ i ≤ ],and which we have just connected to the place name Iasow.

E.Me 33a is equally clear: here, the third position in nominative isoccupied by mdayn, which has also already been interpreted as an eth-nic name (corresponding to genitive mwdon≤ ). In this same inscription,the onomastic formula appears iterated, but for the second time (E.Me33b) without the ethnic name (idmns myre ≤ ).9

In the case of klorul , a similar explanation can be envisaged, although

it is necessary to admit that there are no parallels in the place namesof Greek sources for a place name *k(V)loru- or similar.10

A further example of a possible ethnic name in nominative is off eredby E.Me 44:

(a) apmen “rquq≤ kojol ∞i(b) mwton≤ ∞i

8 On this reading, see p. 54.9 For the problem of idmns (s-stem nominative or rather a stem with s-ending?), see

below pp. 314–317.10 An alternative analysis would be to think of a type of title referring to triqo.



271

Leaving aside the particle ∞i , the structure of E.Me 44a is identical toE.Me 25, E.Me 6 and (with the caveat of -s in idmns ) E.Me 33a. The

last form therefore also appears in nominative. A good connection canbe established here with the name of the island of Cos. The exact par-allelism with the name of the inhabitants of this island in Greek sources,K≈Ûoi, is astonishing. For -l , note also a possible similar suffix in kloru-l .As for (b), the genitive mwton≤ must be related to the father’s name“ rqurq ≤ . This specification has been added outside of the basic formula,but connected with it by means of the agreement.11

To sum up, there seems to be good evidence for interpreting a greatnumber of third words in three-fold formulae as ethnic names. However,cases like t ® ∞atar ≤ prevent us from extending this interpretation to all the examples of three-fold formulae. As the examples from Hyllarimashow, a threefold structure consisting of name + patronym + pap-ponym also existed in Carian, and some examples from Egypt couldcorrespond to this kind of structure.

A more complicated question is which part the ethnic name mustbe attributed to: in the inscriptions with the structures N-Ø N-≤ N-Ø(ethnic name), N-Ø N-≤ N-≤ (ethnic name), or N-s N-≤ N-≤ (ethnic

name), there is no ambiguity: in the first case, the ethnic name refersto the deceased, in the second and third, to the father of the deceased.But in N-≤ N-≤ N-≤ (ethnic), the structure is ambiguous, as the last gen-itive could refer to the first one or to the second one. The existenceof both possibilities, as demonstrated by the unambiguous formulaeabove, does not help to resolve the problem. It is possible that thediff erent uses of ∞i also mark diff erences in the structure, but it is gen-erally difficult to tell which functions this particle is bearing when it isused.

2. Stelae for Women

At least in two cases, the illustration in the stela makes it clear thatthe deceased was a woman: in E.Me 12 and in E.Me 13:

pjabrm | u≤ol ≤ | mwdon≤ ∞i kbjom≤ | m[no≤ ] E.(Me 12)“dtat≤ | upa | w | wet≤ ∞i | mwdon≤ ∞i (E.Me 13)

11 Note, however, that this overall analysis of E.Me 44 raises questions that are stillunclear: the individual to which kojol is referred bears an Egyptian name (apmen ), whichis rather strange (we must accept, then, that this individual, coming from Kos, adopteda new, Egyptian, name). Moreover, this intepretation implies that while apmen camefrom Kos, his father was a mwdon-.



272

E.Me 13 does not show any special characteristics. We have alreadyincluded it in the list of Genitive—Genitive—Genitive inscriptions above.

Like other inscriptions of the same kind, the role of mwdon≤ is ambigu-ous, and could complement either the first or the second name: wet ≤ could be the father’s name, but it is also possible that it represents thehusband’s name (see below). The w after upa is unusual: is it a mistake,as Masson suggests, or does it represent a word, either complete orabbreviated?

E.Me 12 off ers a more interesting structure. It presents a four-foldonomastic formula, pjabrm- u≤ ol ≤ - mwdon≤ - kbjom≤ and a last, incompleteword, whose integration as m[ no≤ ] seems suitable using the example of another four-fold inscription, E.Me 16, which we will come to imme-diately after. The meaning ‘son’ for mno≤ has been already proposedon several occasions (see Chapter 11 s. v. for further details), and seemsto be the simplest and most logical solution, especially in the light of certain funerary inscriptions of Caria (see below p. 289 and ss.). Thereseems to be some connection between the fact that the deceased is awoman and the presence of an atypical four-fold formula. A good solu-tion would therefore be to interpret u≤ ol ≤ mwdon≤ kbjom≤ m[ no≤ ] “of U≤ol ,

the mwdon, the son of Kbjom” as the name of the husband .A comparable analysis can be proposed for E.Me 16, which presentsa structure very similar to E.Me 12, as has been mentioned above:

irow | pikra≤ ∞i semw≤ | mno≤ mwdon≤ ∞i (E.Me 16)

irow is a name of Egyptian origin, where it appears documented asboth masculine and as feminine (see Chapter 11 s. v.). If we choose tointerpret it as feminine, the structure and analysis of the text is iden-tical to E.Me 12, the sole diff erence being the position of mwdon≤ ∞i ,which in E.Me 12 precedes the filiation formula with mno≤ . Therefore,here we can translate it thus: “Irou (f ), (husband) of Pikra, the son of Semw, the mwdon-.” It is impossible to decide whether the diff erentplacing of mwdon≤ ∞i is due to the fact that here, unlike in E.Me 12,mwdon≤ refers to the father’s name, semw ≤ , and not to the husband’sname, pikra ≤ : the accumulation of genitives renders the overall struc-ture ambiguous.

The following inscription also contains the co-occurrence of the name

irow and the word mno≤ :irow≤ : psH ÿm[-]≤ pttu≤ : mno≤ (E.Me 27)



273

In this case, it is a threefold formula, and a superficial approach wouldsuggest that it be grouped with the other Genitive-Genitive-Genitive

formulae. But the presence of mno≤ after the third name seems to chal-lenge this simple interpretation. If: ps H ÿm[ - ]≤ pttu≤ was a sequence of patronym and papponym of irow– , we would instead expect the wordfor ‘son’ to follow ps H ÿm[-] ≤ , not pttu≤ . But if in this case irow ≤ is afemale name followed by the onomastic formula of the husband, theplacing of mno≤ makes sense: ‘Of Irow (f.), (the husband) of PsH ÿm[-],son of Pttu’.

3. Inscriptions with ted and enHajnal and Schürr independently proposed that the Carian words for‘father’ and ‘mother’ could be recognised in the forms ted (E.Me 38)and en (E.Me 32) respectively. The correspondences with Lycian andLuwian are striking (CLuw. t àti “ , Lyc. tedi ; CLuw. anni “ , Lyc. e ˜ ni , seeChapter 11, ss. vv. ) and the contexts in which they appear favour thisanalysis as kinship terms:

“ÿin≤ | upe | arie?≤ ∞i ted (E.Me 38)

iturow≤ | kbjom≤ | ∞i en | mw[d]on≤ ∞i (E.Me 32)

The structures of both inscriptions are very similar, and the diff erenceseasy to explain. In both cases, the name of the deceased appears ingenitive; in E.Me 38 it depends on the word for ‘stela’ (here upe ),whereas in E.Me 32, as in many other cases, the word for ‘stela’ iselliptical. The construction of the kinship noun with ∞i is identical inboth texts: N-≤ ∞i ted , N-≤ ∞i en, and the interpretation, clearly paral-lel: ‘who (is) the father of N’, ‘who (is) the mother of N’.

The construction with the particle ∞i presents very interesting char-acteristics. First of all, the interpunction in E.Me 32 seems to be morethan a mere coincidence; unlike the most widespread type of con-struction with ∞i , characterized by the postclitical position of this parti-cle (it appears systematically attached to the preceding word), in thiscase, ∞i ‘hangs’ on the following word en. This prosodic discrepancyseems to correspond to a deeper, syntactic, diff erence. It must be notedthat in both examples, ted and en must be analysed as nominatives,

which implies that the construction with proclitic ∞i constitutes a truerelative clause, despite the absence of an express verb, with ∞i andted/en in nominative (as the syntax of the relative clause demands) vs.the respective antecedents in genitive. This diff ers from the postclitic



274

uses of ∞i : in an inscription like the above-mentioned E.Me 16, repeatedhere:

irow | pikra≤ ∞i semw≤ | mno≤ mwdon≤ ∞i

Here, mwdon≤ is an ethnic name referring to semw ≤ . If the constructionof postclitic ∞i were similar to that of proclitic ∞i , we would expect*semw ≤ mdayn ∞i , i.e. both ∞i and mwdon≤ in nominative. Instead, mwdon≤ agrees with the antecedent of ∞i, semw ≤ . As Hajnal (1997a) has con-vincingly demonstrated, this construction is a further evolution of con-structions of the proclitic type; from the use of elliptical constructionsN1-(≤ ) N2-≤ ∞i (*mno ) ‘(Of ) N1, who (is the son) of N2, the position of N2-≤ was extended to receive predicates and appositions syntacticallyattracted by the antecedent.

The existence of two uses of ∞i —as proclitic, introducing what canstill be considered a relative clause, and as postclitic, with a functioncloser to a simple article—allows us to interpret the structure of E.Me17 with a certain precision:

“arnai≤ upe | quq≤ bem≤ ∞i mda ÿn (E.Me 17)

Regardless of whether “ arnai- is in this case a female name—if so, quq-could be interpreted as the husband’s name, as in E.Me 16, 32 and38—or a masculine one—which would imply that we are dealing withan individual name + patronym + papponym formula—the most rel-evant fact is that b em-≤ (gen.) ∞i mda ÿn (nom.) is exactly parallel to“ÿin-≤ . . . ∞i ted and iturow-≤ . . . ∞i = en, in contrast to the abundant exam-ples of postclitic constructions (mwdon“ = ∞i (irow pikra ≤ = ∞i semw ≤ mno≤ mwdon“ = ∞i E.Me 16, etc.).

This analysis of ∞i mdayn/mda ÿn vs. mwdon“ = ∞i encounters some prob-

lems in the interpretation of two inscriptions containing mdayn = ∞i :(a) ta“ubt≤ / kuari≤b/ar | ≤en / niqau≤ / ptnupi(b) idmuon≤ / ∞i | mdayn ∞i (E.Me 18)

(a) idmns | myre≤ ∞i | mdayn ∞i(b) idmns | myre≤ ∞i (E.Me 33)

The difficulties posed by E.Me 33 are not so serious: we can assumethat idmns is an s-stem nominative form, and that mdayn depends directly

on idmns. The case of E.Me 18 is more difficult, but in this inscriptionit is the overall sense of the text that remains unclear. In any case,idmuon≤ ∞i mdayn ∞i , added by a diff erent hand, must refer to a namein part (a), as shown by the presence of ∞i after the two words. The



275

closest name is ptnupi , in nominative form, which allows us to interpretmdayn ∞i as an ethnic name referring directly to ptnupi , in which case

the construction of the postclitic ∞i would be regular (mdayn is also anominative, like ptnupi ).It must be noted that the hypothesis envisaged here about the exis-

tence of two diff erent uses of ∞i —a postclitic one introducing not onlygenitive but also attributive and appositional words, and with attrac-tion to the case of the antecedent, and a proclitic one in a construc-tion closer to the original structure of relative clauses—is based onevidence that remains scarce, and must therefore be considered asmerely provisional.

4. Other More Complex Funerary Inscriptions

Leaving aside those already analysed (E.Me 12, E.Me 16 and E.Me44), and the very difficult stela E.Me 18, mentioned above, there areonly three other inscriptions from Memphis that include complex for-mulae with more than three names:

arjom≤ : ue : mwsat≤ : ∞i : mwdon≤ : ∞i tbridbd≤ : ∞i (E.Me 42)

(a) l ÿ∞si≤ | upe | “rquq≤ ∞i | ksolb≤ (b) arliom≤ | mno≤ ∞i (E.Me 43)[--]qarm≤ : q[---]≤ ∞i : pdubi≤ mno≤ [mw]don≤ ∞[i ---]w≤ord≤ ∞i (E.Me 10)

E.Me 42 can be compared to other typical inscriptions that presentthe structure N-Ø N-≤ N-mwdon≤ (E.Me 13, 28, 31), the sole diff erencebeing the addition of a fourth name. The simplest interpretation is toview this as the papponym, but expressed formally as a “patronym of the patronym”: ‘Of Arjom the stela, who (is the son) of Mwsat, themwdon-, the (son) of Tbridbd-.”

E.Me 43 is perhaps less complex than it appears, as one can in factspeak of two diff erent formulae. In E.Me 43a, we find a typical three-fold formula N-≤, N-≤ ∞i N-≤, where the third name seems to be anethnic, as already seen: ‘Of L ÿ∞si- the stela, the (son) of ”rquq-Kasolabean” (with the ambiguity of the exact referent of the ethnicname: ”rquq or L ÿ∞si?). E.Me 43b, arliom≤ mno≤ ∞i , literally translatedas ‘of Arliom, who (is) the son’ makes complete sense when consideredas a further addition to E.Me 43a. Therefore, the stela would be ded-icated to both L ÿ∞si and his son Arliom. An satisfactory overall trans-lation would be, “Of l ÿ∞si- the stela, the (son) of ”rquq- Kasolabean.(And) of Arliom, the son”.



276

As for E.Me 10, the lacunary character of the text makes it difficultto understand. The most remarkable fact is that this inscription appar-

ently contains a five-fold formula, which represents an exceptionallylong structure. This unusual characteristic, and the presence of mno≤ after the third name, leads us to the possible conclusion that we areonce again dealing with a stela for a woman; aside from the last name,the structure is very similar to E.Me 16, the only diff erence being thegenitive vs. nominative case for the first name, although this is largelyirrelevant:

[--]qarm≤ q[---]≤ ∞i: pdubi≤ mno≤ [mw]don≤ ∞[i] ]w≤ord≤ ∞i (E.Me 10)irow pikra≤

∞i semw≤ mno≤ mwdon“

∞i Ø (E.Me 16)

In this example, a woman, [--]qarm, would therefore be mentioned asthe wife of Q[---], the son of Pdubi≤, the mwdon-. As in the case of theirow inscription, the function of mwdon- is ambiguous, insofar as it couldrefer either to pdubis or to q [---] (see above for E.Me 16).

If this interpretation is correct, the last name [---]w ≤ ord ≤ must beeither the name of the father of pdubi- (i.e., the papponym of q[---])or, less likely, the ethnic name of pdubi- if [ mw ]don- refers to q [---].

5. The Rest of the Inscriptions from the Memphis Corpus

The formulae and structures analyzed in the preceding pages accountfor most of the inscriptions from the Saqqâra corpus. Only a few inscrip-tions have not yet been cited and/or analyzed. Several of these are sofragmentary that they cannot be included in any of the structuresalready mentioned (E.Me 22, E.Me 37, E.Me 39, E.Me 52–E.Me 66).The remaining examples will be commented on briefl y in the follow-

ing paragraphs.ter ÿez≤ | upe | nuol ∞[---]sarmrol ∞ yt (E.Me 4)

Although the beginning of the inscription points to a typical N-≤upe . . . formula, no parallels can be found for the final part. Unfortunately,the gap in the middle of the text complicates the interpretation stillfurther. The final -ol in ?]sarmrol would be consistent with an indi-vidual name in -ol (= -vllow ) or with a possible ethnic name, likekojol . This interpretation would leave the last three letters as an inde-pendent word ∞ yt . Could this be a verb? The final -t recalls the pos-sible third singular ending -t in ÿbt (see p. 281–282). This would mean



277

that a typical formula ‘Stela of Ter ÿez . . .’ would be followed by thename of the dedicant, ‘?]sarmol made/dedicated’. This is, however, a

very hypothetical solution.12

psm“kwneit≤ | ue | naria≤ | ≤ugl iq | sarl? (E.Me 5)

The characteristics of E.Me 5 are to some extent comparable to thoseof E.Me 4. The inscription begins in a strictly ordinary way (N-≤ ue N-≤ ‘Stela of Psm“kwneit, (son) of Naria-’), but the following two wordsmove away from the usual formulae; ≤ ugl i- has already been alreadymentioned, appearing in two other inscriptions as third name, and thisinterpretation as a possible ethnic form is corroborated by its resem-blance to the place name Souaggela. But the most surprising featureof the present inscription is that we would expect to see a form in gen-itive ≤ ugl i ≤ , according to the typical formula N-≤ N-≤ N-≤. With this inmind, is ≤ ugl iq merely a mistaken form intended to be ≤ ugl i ≤ (id est,q q erroneously used for z ≤ ), or is the final -q linked in some wayto the presence of the final word sarl ? , a hapax? The existence of theparticle -q , perhaps of connective nature, is now confirmed thanks tothe new inscription of Hyllarima (armotrq d os=q ), but its function in this

latter text unfortunately remains unclear.The difficult reading of the last letter (sarl ? is a proposal made bySchürr; an alternative reading is sara ) makes an analysis of the inscrip-tion even more complicated.

(a) wa r[---]t[------]i[---]≤ | mda ÿn(b) [--15--]a[-]i≤ | mda ÿn (E.Me 11)

The only reason for citing such a fragmentary, largely illegible inscrip-tion is the repeated presence of mda ÿn. As the illustration on p. 43

shows, this stela represents a male-female couple, and each line of theinscription, situated behind each member of the couple, seems to cor-respond to the individuals drawn on the stone. We can imagine thatthe structure of both onomastic formulae was identical: N-Ø N-≤ . . . mda ÿn,i.e. individual name in nominative + father’s name (and also grandfa-ther’s name?) in genitive + ethnic name mda ÿn in nominative. Theseare the only examples in which mda ÿn appears without ∞i .

12 Schürr (1992:155) tries to connect the peculiar content of this inscription to thefact that the stela in which it appears is a ‘stèle de donation’, representing the PharaohApries making an off ering to the god Ptah.



278

tdu≤ol / kbos | “amsqi[. . .? (E.Me 24)

The inscription, and consequently the formula, is almost complete: onlya few signs can be missing in the final part of the last word. The mostnotable feature of this inscription is that the individual name in nom-inative is followed by another name ending in -s . Interpreting kbo-s asa dative of dedication, which would be a good argument for -s as atrue dative ending, seems unlikely, not only because of the resulting exceptional structure (“Tdu≤ol to Kbo”), but also because of the par-ticular disposition of the text in the stela; tdu≤ ol , situated at the verytop of the stela, and inscribed in longer letters, must be the name of

the deceased. The simplest solution is to consider kbos an attribute orapposition in nominative to tdu≤ ol . It could certainly be a title accom-panying the name of the deceased, but an interesting connection alsoarises if we interpret it as an ethnic name: the coin legend kbo, whichKonuk has convincingly classified as a place name, identifying it withKeramos (Konuk 2000b). It is true that kbo may only be the initial let-ters of the place name, unknown in Greek sources, but the possibilitythat kbo is a complete form cannot be discarded: kbo-s could thereforebe an ethnic name corresponding to that place name. The suffixationcan be compared directly to (accusative) -s-n in otono-s-n ‘Athenian’,from *otono- ‘Athens’. As for samsqi [, it could be the patronym (sam-sqi [ ? . . .-≤ ].

wnuti≤ | kwar≤ mHm≤ ∞i | mwdon≤ [ ∞ ]i (E.Me 31)

The analysis of this inscription is hampered by the problematic sequencekwar ≤ mHm≤ , where it is not clear if we are dealing with a single nameor rather with two diff erent names, both in genitive. The latter solu-

tion has been adopted here (p. 59), but the resulting structure is some-what strange: it would apparently be a four-fold structure, where mwdon≤ appears as a fourth name, and not as third name, which it is moretypical.

Vittmann (2001:48–49) has off ered a plausible solution to this prob-lem and at the same time a very convincing interpretation of the firstword, wnuti ≤ . He suggests that wnuti- could be the adaptation of Egyptianwnwtj , ‘hour-observer, astronomer’ (see Chapter 11 s. v. wnuti ≤ ), so thatwnuti- would be a title referring to kwar-≤ , the name of the deceased.

Thus the structure of the onomastic formula becomes an ordinary three-fold one: ‘Of the astronomer Kwar-, who (is the son) of MHm-, who(is the) mwdon-’ (once again with the ambiguity about the exact refer-ent for mwdon- ).



279

[q? ]lal is / [?]iam≤ ∞i / alos ∞arnos (E.Me 45)

In the two initial words we can easily identify a formula, Nominativeor “s-ending” + Genitive. As in other cases, it is difficult to decidewhether the name is an s-stem in nominative or an s-ending, in whichcase the formula would have to be compared to the ntokri-s type (E.Me35).

The pair of words alos ∞arnos is far more difficult to analyse. Theresemblance to the well-known place name Halikarnassos, first sug-gested in Adiego (1990), continues to be an attractive—if problematic— possibility. The word reappears in another inscription (C.xx 2, under

the form alos d ∞arnos d ), and I refer to the analysis of that inscription(below p. 284). In this case, we cannot rule out the possibility of inter-preting it as an ethnic name or even a place name.

tqtes | paraibrel ≤ ∞i | mn[o-?] (E.Me 47)

The initial name could again be a nominative of an s -stem or ans-ending form, tqte-s . The following sequence, parai b re l ≤ ∞i mn[ o-?], wherewe find N-≤ ∞i mn[ o-?] ‘son’, recalls the structures arie ? ≤ ∞i ted (E.Me 38),kbjom≤ ∞i en (E.Me 32) mentioned above, and leads to an interpretation‘tqte(s). who (is) the son of Paraibrel ’. In this case, it is very likely thatthe word ‘son’ appeared in nominative (*mno ), a form that is still undoc-umented.

loubaw | siral | pnld≤wl (E.Me 49)

This inscription is perhaps the most unusual example found in theCarian corpus of Saqqara. Apparently a threefold formula, the secondand third names end in -l , which would be a morphological ending

without clear parallels. Moreover, no parallel can be found in the Carianonomastics for any of the three alleged names. In any case, we mustbear in mind that the overall reading of the text remains very uncer-tain (see p. 71 for remarks).

6. A First Summary

The previous pages have allowed us to familiarize ourselves with theonomastic formulae, from the most basic to the more complex. Our

main sources of information were the funerary stelae of Saqqâra, whereit is common to find more complex structures than simply the indi-vidual name, or of the individual name + father’s name. We have seen



280

the high frequency of threefold formulae, which contain a third name,possibly an ethnic name, but also in some cases the name of the grand-

father. Even more complex structures can also be found, and in someremarkable cases the increased complexity seems linked to the fact thatthe deceased is a woman; in such cases, the onomastic formula thataccompanies the female individual name seems to refer to the husbandand his genealogy. In two exceptional cases, the deceased is mentionedas the father or mother of another person (E.Me 38, E.Me 32).

Also typical in the Memphis funerary stelae is the absence of verbforms. Only in the case of E.Me 5 could this possibility be envisaged,but the fragmentary nature of the text prevents reaching a definitiveconclusion. The rest of the Saqqâra epitaphs are characterised by thedirect reference to the deceased in nominative or in an s-ending form,or to the stele in an expressed or elliptical form, in which case thename of the deceased appears in genitive, indicating possession. Noneof these cases require the use of a verb form. Indeed, not even in theconstructions with ∞i is a verb used.

The absence of verb forms is counterbalanced by the presence of certain nouns that belong to the Carian common lexicon: mno- ‘son’,

ted ‘father’, en ‘mother’, upe (and variants), ue ‘stela’, to which twopronominal forms can be added: the abovementioned relative ∞i andthe demonstrative sa (in upe sa ). As for the use of ∞i , we have identifiedtwo diff erent constructions: the first, closer to a use as relative, wherethe nominal predicate introduced by ∞i is in the nominative while theantecedent is in the genitive (type arie ? ≤ ∞i ted ), whilst the second is morecomparable to an article or connecting particle, where the complementintroduced by ∞i is attracted by the antecedent (type semw ≤ . . . mwdon“ ∞i ). Although the evidence is not sufficient to draw definitive conclu-

sions, we have been able to observe a clear correspondence betweenthe position of ∞i and each of these two functions: it seems to be pro-clitic when used as ‘relative’, and postclitic when used as ‘article’.

C. A B I

In this section, I shall analyze a selection of brief Carian inscriptions

other that those mentioned in the preceding sections. I leave out thosegraffiti containing more than simply onomastic formulae, given thedifficulties of reading present in most cases, which have not been ade-quately edited (Thebes, Silsilis, Abydos). Speculating about the struc-



281

ture and meaning of inscriptions for which we do not possess a reli-able reading would be a very risky practice. In my view it is prefer-

able to limit the research to those inscriptions we can be confident of reading correctly.

1. Inscriptions on Objects

Two ‘Pharaonic Objects’ (E.Me 8, E.xx 6) and the Use of sb ‘and’

We begin with two very brief inscriptions on so-called ‘pharaonic objects’,which serve to illustrate the use of the coordinative conjunction sb

E.Me 8, a bilingual inscription on the base of a statuette of Apis,consists of two parts. The text that appears in the first part, paraeym:armon ∞i , is now interpreted without difficulties as ‘Paraeym the inter-preter’, in direct correspondence to an identical formula that appearselsewhere in Egyptian (see p. 41 and Chapter 11 s. v. armon ). In thiscase ∞i introduces an apposition to the personal name in nominative.In the second part, two personal names in nominative are united bysb, which has been unanimously interpreted as a coordinative conjunction (‘and’), above all since the discovery of the Kaunos bilingual,where it appears repeatedly with this function. E.Me 8b paraeym sb polocan therefore be interpreted as ‘Paraeym and Polo’.

A construction similar to E.Me 8b, but in ‘s-ending’ case, is visiblein the inscription E.xx 6 on the basis of a statuette of Isis: “ arnajs sbtaqbos ‘For/of ”arnaj and Taqbo’. It is probable, but impossible todemonstrate, that these formulae of EMe 8b and E.xx 6, consisting of a pair of names, represent the names of a husband and wife.

Three Inscriptions on bowls (C.xx 1, C.Ha 1, C.xx 2)

Three inscriptions on bowls constitute a type of small sub-corpus of particular interest. In a very influential and decisive article, Melchert(1993) off ered an interpretation of one of these texts (C.xx 1), whichin my opinion remains essentially valid. We will begin thus with thisinscription, adopting in general terms the views expressed by Melchert:

“rquq | qtblem≤ | ÿbt | snn | orkn | ntro | pjdl? C.xx 1

In this inscription, an onomastic formula had already been identified:

“ rquq qtblem≤ ‘”rquq (son) of Qtblem’. Melchert’s interpretation of theremainder of the text is based on two fundamental and compelling hypotheses: firstly, that ÿbt is a verb comparable both formally and



282

semantically to Lycian ubete , ‘off ered’, and secondly that snn orkn is thedirect object of this verb, formed by a demonstrative snn (which would

belong to the same paradigm as san in the Athens bilingual inscription)and a common noun referring to the phiale, orkn. Both elements appearin the accusative singular, morphologically reflected by the ending -n(sn-n ork-n ). Melchert’s elegant interpretation of the five first words of C.xx 1 is now supported by the discovery of the bilingual inscriptionfrom Kaunos, which has confirmed the existence in Carian of anaccusative sg. ending -n and has dispelled the doubts about the valueof y/ ÿ (close to u ) , thus assuring the equivalence ÿbt = Lyc. ubete .

The last two words remain problematic. Melchert has recently revisedhis initial views on these forms. In his 1993 article he intepreted ntroas a dative of the Carian name for Apollo. For pjdl , he proposed theanalysis as an apposition to snn orkn with the meaning ‘gift’, etymo-logically related to Hitt. pài-/piya-, Luw. piya-, Lyc. pije- ‘to give’: -dl would represent a suffix*-d hlo- ). However, Melchert (2002:309–310)denies the existence of Carian datives ending in a vowel and, follow-ing a suggestion by Schürr, prefers to interpret of ntro as referring to“ rquq , with the meaning ‘priest of Apollo’: ‘”rquq (son) of Qtblem ded-

icated this bowl—the priest of Apollo as a gift’. Leaving aside for nowthe discussion about the alleged “datives” in -s , the internal syntacticreasons that Melchert adduces in refusing the interpretation of ntro asdative are not particularly convincing: he states that pjdl cannot be anapposition to the direct object if ntro is a dative, due to the separationthat this latter word introduces between snn orkn and pjdl . For this rea-son, he classifies both ntro and pjdl as two ‘epexegetic, add on phrase[s]’defining, respectively, the subject and the direct object, as his transla-tion off ered above tries to reflect. However, following this logic, I see

no reason why we should not consider only the word pjdl as an ‘epex-egetic phrase’ (= ‘”rquq (son) of Qtblem dedicated this bowl to Apollo— as a gift’). A further complication arises if we take ntro to be not theCarian name for Apollo, but rather a derivative of it: it obliges us torecognize the same derivative in E.xx 7, where an s-ending form appears(ntro-s ), which Melchert interprets as a dative. We must also assumetherefore that this latter inscription is dedicated not directly to ApolloBranchid but to an unmentioned priest of Apollo Branchid, perhaps a

less satisfactory solution (see below p. 317).In any case, despite these difficulties that depend to a great extenton the crux about datives in Carian—on which see pp. 314–317 for



283

further remarks—Melchert’s interpretation of the overall meaning andthe first part of the syntactic structure of C.xx 1 remains one of the

most brilliant contributions to the understanding of a Carian text, andthe consequences of the etymological connections it highlights are veryimportant.

The connection of C.xx 1 with another inscription on a bowl (C.Ha 1)has not gone unnoticed:

smd ÿbrs | psnl o | md orkn t ÿn | snn (C.Ha 1)

Melchert (1993:81) correctly notes the presence of snn and orkn, thesame nominal phrase that appears as a direct object in C.xx 1. Giventhat both inscriptions are engraved on similar objects, this parallelismis consistent with the interpretation of snn orkn as ‘this bowl (acc.)’. Butthe correspondence between the two inscriptions ends here, and it wouldbe very difficult to interpret the structure of C.Ha 1 using the exam-ple of C.xx 1. The only clear form (apart from snn orkn ) is smd ÿbrs ,which is undoubtedly a personal name, although this is ambiguous dueto the final -s , which could be either an inflection mark (‘s-ending’) orsimply the final consonant of an s -stem. Forms such as k ≤ t ÿbr, ardybyr-≤ ,

dt ÿbr , Arduberow could point to the first solution, but the new personalname ybrs-≤ from Hyllarima (C.Hy 1a), a true s-stem that seems to con-tain the same root ybr-/ ÿbr-, allows us to classify smd ÿbrs as a pure nom-inative form.

The presence of orkn . . . snn leads to the assumption that a transitiveverb must be also present in the inscription. Without entirely discard-ing the possibility of omitting the verb, suggested by Melchert (1993:81),it is advisable from a methodological point of view to pay careful con-sideration to whether any of the three remaining words ( psnl o, md , t ÿn )

could represent the verb of the inscription. If we look at each of thethree possibilities in turn, we will see that the choice here is not easy.It seems almost certain that t ÿn is a word in agreement with orkn snn,given that it is situated between these two words; a diff erent explana-tion would imply that the noun of the recipient and the demonstrativethat accompanies it form a discontinuous noun phrase, a hyperbatonthat is not altogether impossible, but certainly quite unlikely.13 In thecase of md , we must not overlook its resemblance to the forms md a,

13 However, the interpunction that separates md orkn t ÿn from snn suggests that thispossibility should not be ruled out altogether.

284

md ane , which have been interpreted as verbs in various studies (see forinstance Melchert 2002:308, n. 7). However, in my view, the forms md a

and md ane forms cannot be verbs, as we will see later (p. 324). Thisleaves psnl o as the form most likely to be a verb. As evidence to sup-port this hypothesis we could consider the resemblance of psnl o | md

to pisñoimd a in the Hekatomnids’ decree from Sinuri (C.Si 2). In Adiego(2000:141–143), this latter form was connected with the Anatolian verbfor ‘to give’ (Hitt. pài-/piya-, Luw. piya-, Lyc. pije-, etc.), see p. 304.

An alternative analysis of psnl o would be to take it as a personalname; this possibility would lead us to accept Melchert’s hypothesis onthe absence of an express verb, or to attribute this function to md ort ÿn, the difficulties of which have already been mentioned. If pns l o isa personal name, the resulting sequence smd ÿbrs psnl o would create anembarrassing dilemma: is smd ÿbrs the indirect object and psnl o the subject, or vice versa? The dilemma is inseparable from the problem of the Carian dative (see pp. 314–317).

Shown below is the third inscription on a bowl:

ÿ≤biks not : alosd ∞arnosd : jzpe mdane (C.xx 2)

Adiego (2000:153–155) off ered an interpretation of this inscription. Thisinterpretation hangs on two fundamental hypotheses: that ÿ≤ biksnot mustbe segmented ÿ≤ biks not and interpreted as a sequence Individual Name14

+ Verb (where not would be morphologically comparable to ÿbt in C.xx1, see above), and that the well-known word md ane is not a verb, butrather a sequence of particles and clitic pronouns. For no-t, a connec-tion to Hitt. nà(i)-, CLuw. (reduplicated stem) nana- < * PIE *neyH– ‘to bring’ was cautiously put forward. Regarding md ane , an analysisalready proposed in former works was reintroduced: in -ne , a clitic

accusative pronoun -n- + a clitic dative pronoun -e attached to a basemd -, for which no explanation was given. The form jzpe was consid-ered a proper name dative (‘to/for jzpe ’). Finally, alos-d ∞arnos-d wasinterpreted as an ablative instrumental (with -d = Luw. -ti , Lyc. -di ) of the word (in fact a noun phrase) alos ∞arnos , also present in E.Me 45and tentatively identified (in Adiego 1990a) as the Carian place nameHalikarnassos. This yielded a possible translation as ‘Ÿ≤biks brought itto Jzpe from Halikarnassos’. Obviously, this proposal was based only

14 For ÿsbiks as PN, cf . yi ≤ ∞bik ≤ (E.Me 46a).

285

on a set of hypotheses, and as such is far from conclusive. But the coreof the discussion must be retained here: not and md ane are the most

obvious choices for constituting the verb of the sentence, and eitherchoice has important repercussions for the interpretation of the otherinscriptions containing md a, md ane.

The inscription on a bronze lion (E.xx 7)

Another very interesting short inscription is found on the base of abronze statue representing a lion (E.xx 7):

ntros : prãidas / or“a / nu mdane : uksi wrm≤

ntro-s pr ãida-s shows an inflected form of ntro, the same word that appearsin C.xx 1, where it was interpreted as a god name (the Carian Apollo;cf . however the alternative explanation as a derivative, ‘priest of Apollo’).ntro-s is accompanied by pr ãida-s , whose final s seems to indicate anagreement with ntro-s . Schürr proposed that this could be identified asa Carian form connected to Greek Bragx¤dai, ‘the Branchids’—thename of the family of priests consecrated to the cult of Apollo in thesanctuary of Didyma near Milet (Schürr 1998:158)—but as attractive

as this theory is, it is dependent on the rather ad hoc assumption of a phonological value close to ºgxº for the infrequent and obscure Carianletter %. Lacking further evidence, I have chosen to adopt this assump-tion here in purely conventional fashion (which explains the use of <ã>for transliterating %, see p. 20). In any case, there is a general con-sensus that pr ãidas constitutes an attribute or an apposition to ntros , withwhich it would agree.

Melchert’s interpretation of the whole inscription (see Melchert2002:308) is as follows: “Uksi, (son) of Ur(o)m, has now given it, the

or “ a , to the priest of Apollo, the Branchid”. There are no particularproblems posed by analyzing uksi wrm≤ as a typical onomastic formula,individual name + father’s name in genitive, and uksi as a nominativeand consequently, the name of the donor. The rest of Melchert’s inter-pretation is more controversial: he claims that md ane is the verb, nu isan adverb,15 or “ a is the name of the object and finally ntros pr ãidas mustthen be the indirect object in dative case. We have already seen (above

15 No comments are made about this word in Melchert (2002), but it is self-evidentthat he assumes an etymological connection with PIE *nu- ‘now’ (Hitt. nu-, Lat. nunc ,Gr. nËn, etc.)



286

p. 282) that in this paper, Melchert abandons his earlier interpretationof ntro/ntros as a god name in favour of the meaning ‘priest of Apollo’.

This change is a consequence both of his refusal to analyze ntro asdative and of his assumption that the s-ending is a dative mark. Theresult of this new analysis is in my opinion rather unsatisfactory: E.xx.7becomes an inscription dedicated to an anonymous priest of Apollo,whereas in C.xx 1 the verb ÿbt , ‘off ered’, is not accompanied by thename of the being to whom the object is off ered. In my opinion itwould be simpler to interpret ntro/ntros as the god name (‘Apollo’), ntroas dative and to search for an alternative explanation for the s-ending in ntros pr %idas , but I leave the discussion of the forms with s-ending until the following chapter.

The interpretation of md ane as a verb is also unconfirmed, and thisword will be analysed in the following chapter. In any case, if it is notclassified as a verb, the verbal form ought to be looked for in the pre-ceding sequence of words: either nu, or “ a , or less probably or “ anu (with-out discarding further segmentations). Of all these possibilities, nu seemsthe most likely: final -u immediately preceding md ane reminds us of ew mlane (/ ew lane 1× ), a sequence that appears in some inscriptions of

Thebes, where it could also be a verb.The ‘Tarhunt-inscription’ of Iasos (C.Ia 3)

The inscription of Iasos C.Ia 3 off ers some very interesting forms butunfortunately the overall structure and meaning remain unclear:

?] are“ | “anne ml ne | siyklo≤ | “ann | trqude | ∞l mud [?

There is no question that the most significant word is trqud e , interpretedas a god name (Tarhunt, the Anatolian Storm God) in Blümel-Adiego

(1993), which has now been confirmed by the form armotrq d osq inHyllarima (C.Hy 1), where a dvandva Arma-Tarhunt (the moon-godplus the storm-god) is easy to recognize.

Given that the inscription appears on a cratera, we can assume thatthis object is off ered to the god Tarhunt, which would imply that trqud e should be considered an indirect object in dative. However, the rest of the text does not allow us to confirm this syntactic analysis; the onlyother form that can be analyzed with any confidence is siyklo≤ , a gen-

itive, probably of a personal name, which implies that it depends onanother personal name, as part of an onomastic formula. The most suit-able candidate is ?]are “ , the word that apparently begins the inscrip-tion, but the discontinuity of the onomastic formula would then be



287

somewhat surprising. Also unclear are the formal connection between“ anne and “ ann and the resemblance of ml ne to the sequence ºml n in

C.Ka 4 yoml n C.Ka 5 uioml n, two forms which should probably beinterpreted as verbs (see below pp. 298–299, 301). Finally, the analy-sis of ∞l mud [? is hampered by the uncertainty about the final part of the word. Gusmani’s proposed complete form, ∞l mud [ e ], would give apossible epithet for trqud e .

The inscription of ” arkbiom (E.Sa 1)

The text of E.Sa 1, inscribed on the base of a reliquary for threemummified reptiles, is the following:

“arkbiom : zidks mdane : ÿn[ -?]/mo | den : tumn

The text is even more impenetrable than those of the inscriptions onrecipients analyzed above. It clearly begins with a personal name innominative (“ arkbiom, also cited in the Egyptian part of this bilingualtext). This name seems to appear here as a complete onomastic for-mula, with no mention of the father’s name. The inscription again con-tains the verb or particle sequence md ane , a word ending in -o whose

complete reading is not possible ( ÿn[ -?]mo ), and a very obscure formwith no parallels, zidks . For the two last words, a tentative analysis wasmade in Adiego (1995:21–23): given that the Egyptian text contains aformula, ‘that Atum the great god may give life and health to ”úrkbym’,I suggested that the Egyptian god name Atum could be reflected inthe form tumn at the end of the Carian text. This latter form wouldtherefore be an accusative in -n of a stem tum-. The somewhat unusualpresence of an accusative instead of a more typical dedication dativewas explained by analysing d en as a type of preposition, etymologically

related to Hitt. anda, andan, Lyc. ñte .Irrespective of this latter interpretation, the overall structure and

meaning of the inscription remain unclear. If we retain the hypothesisthat md ane is a particle sequence, the search for verbs can be limitedto the words zidks and ÿn[ -?]mo, for which no parallels can be given.

2. Funerary Inscriptions of Caria and Athens

A brief sub-corpus of funerary inscriptions other than those from Egyptcan be identified. It consists of eight inscriptions, seven for diff erentCarian places (C.Tr 1, C.Tr 2, C.Al 1, C.Eu 1, C.Ka 1, C.Ka 3, C.Kr 1)and the bilingual inscription of Athens (G.1). The funerary character of

288

other inscriptions (C.Eu 2, C.Si 1) cannot be ruled out, particularly inthe first case, but they do not display the requisite lexical elements that

would allow us to include them in this typological classification.These lexical elements are the words s(i)d i , present in C.Tr, CTr 2,C.Al 1, C.Ka 1, C.Te 1, and ≤ (j)as in C.Eu 1, G.1. Although neitherof these appears in C.Ka 3, the context of the inscription, engravedon the facade of a typical Kaunian rock tomb, is undoubtedly funerary.

The only bilingual text is found in the inscription from Athens. InAdiego (1992a:32–33; 1993a:165–170) an interpretation was attempted,based on the hypothesis that the single Carian line corresponds exactlyto the first line of the Greek text:

s∞ma tÒde Tur[ ≤ jas : san tur[

According to this inscription, Carian ≤ jas would be equivalent to Greeks∞ma, whereas san would be a demonstrative pronoun corresponding to Greek tÒde. This latter equivalence indicates a very interesting ety-mological connection: san could be related to Luw. za-, Hitt. ka- ‘this’< PIE *∞o-, and would also be evidence of the ‘satem’ treatment of

PIE palatals in Carian, as in the rest of Luwic languages (cf . Adiego1995:12). Melchert (1993:79–80) refined this comparison, explaining theCarian final n of san by comparing with the Hitt. adverb k àn(i) ‘lookhere, lo!’, eni ‘this, the aforementioned’, and combining san and snnfrom the bowl inscriptions in a single paradigm.

The interpretation of ≤ jas as a word for ‘tomb’, ‘funerary monument’or similar is reinforced by its presence in C.Eu 1, where it appears— in the slightly diff erent form ≤ as— accompanying a standard twofoldonomastic formula:

“as : ktais idyri∞≤ : mn[o-?]

Both ktais and idyri ∞≤ admit good onomastic identifications: ktais recallsÑEkata›ow, a Greek name commonly attested in Caria, perhaps becauseof its resemblance to the indigenous name ktmno (adapted in Greek as

16 In former works, I suggested that ktais , analyzed as nominative, could contain ans-stem created from the Greek nominative, on which the Carian inflection of the loan-word could have been formed (cf. Lyc. zeus- in dat. zeus-i , from Greek ZeÊw ). But forthe reasons explained above, I believe that the use of nominative can be excluded inthis case.



289

ÑEkatÒmnvw ). As for idyri ∞≤ , it belongs to the same family of names asSaurigow, Senurigow (on this family, see pp. 262–263). Once again, the

final -s in ktais can seem ambiguous (s-stem16 or case ending), but inthis case the simplest solution is to interpret it as a inflectional mark,given the presence of a word referring to the monument, which rulesout the possibility of expressing the name of the deceased directly innominative. ktai-s must therefore be an ‘s-ending’ form, and, as in manyother cases, it can be translated as true dative = indirect object (‘tomb for Ktai’) as well as possessive or genitive (‘tomb of Ktai’). idyri ∞≤ doesnot pose any problems: it is the name of the father of ktai , expressedin the usual genitive in -≤ . As for mn[ o- ], it is regrettable that the finalpart of the inscription is broken. An integration mn[ os ] seems the mostlikely solution, because this word must agree with ktais . The completetext can thus be translated: ‘Funerary monument for/of Ktai, son of Idyri∞’, without ruling out, however, the possibility that mn[ os ] couldactually be a juxtaposed word to ktais , which would give the following meaning: ‘Funerary monument for/of Ktai (son) of Idyri∞ (and his) son.

Interpreting the inscriptions with s(i)d i is more problematic, sincetheir structure is more ambiguous and the segmentation of the words

in the contexts where s(i)d i appears is not always certain.We can begin with C.Tr 2, a complete and brief inscription withvirtually no problems of segmentation:17

an sidi a-/rtmi pau≤ / parãaq?

However, the interpretation of this text is seriously impeded by thedifficulties of reading posed by the last sign of the last word. As notedin p. 131, all the existing drawings of this lost inscription seem to pointto Q, but the well-known problem of distinguishing between z, Q andalso o make it unwise to automatically accept a reading as Q. Therefore,we must also consider the possibility of an alternative reading with z

( par ãa ≤ ? ),18 a theory that has been vehemently defended by Schürr.The main advantage of the reading par ãa ≤ is that it results in a clear

threefold onomastic formula, N-Ø N-≤ N-≤: rtmi pau-≤ par ãa-. Although

17 Only the segmentation si d i artmi could be contested, because one could theoreti-

cally separate si d ia rtmi (for this latter, cf . rtim C.Hy 1a), but this alternative segmen-tation does not aff ect most of the possible interpretations that will be considered here.

18 The third possibility—reading o, hence par ãao? —seems to be excluded, given thata final sequence ao in this position has no parallels and does not lead to any satis-factory analysis.



290

there are no clear parallels for a name par ãa-, the typical onomasticelement par(a)- can be easily recognized. Even a connection to pr ãidas

could be considered: if pr ãida-s corresponds to the name of the priestfamily of Didyma Branchidai, par ãa- could be the Carian name corre-sponding to Greek Brãgxow. However, the problem is to syntacticallyconnect the onomastic formula rtmi pau-≤ par ãa-≤ with the initial for-mula an si d i . This latter seems to consist of a noun that would denotethe funerary monument, si d i , preceded by what is probably a demon-strative pronoun an (cf . san in G.1): therefore, an si d i would mean ‘thistomb’, or ‘this (is) the tomb’, or similar. The presence of this formulaseems little consistent with an N-Ø (nominative) rendering of the nameof the deceased. This sentence could only make sense if we interpretan si d i as an accusative , and assuming an elliptical verb (‘made’).

An attractive alternative is to suppose that artmi is not an accusative,but rather a dative , which would mean that the overall sentence mustbe interpreted as ‘This tomb/this (is) the tomb for A., (son) of P., (grand-son) of P.’ Developing this hypothesis further, one might wonder if -i could be a true dative ending for a stem artm-. This possibility wouldallow us to integrate artm-i and the name (in nominative) rtim from

C.Hy 1a in a single paradigm (nominative (a)rtim-Ø / dative art(V)m-i ).However, the existence in Minor Asian onomastics of a large numberof names formed on the basis of the onomastic element art(e/i)m-, butwith diff erent derivations, makes this paradigmatic connection of artmi and rtim a very fragile theory.19 Consequently, it is currently impossi-ble to decide whether artmi is a nominative or a dative, since there aresignificant difficulties encountered by either solution.

The reading par ãaq would give a new perspective. If par ãa- is a per-sonal name stem, -q could recall ≤ ugl i-q in E.Me 5, but this is an obscu-

rum per obscurius solution, given that the interpretation of ≤ ugl i-q in thecontext of E.Me 5 is also very problematic (see above). A very diff erentway of analyzing par ãaq was suggested in Adiego (1993a:263) and devel-oped further by Hajnal (1995[97]:20). Taking as a starting point myproposal of connecting par ãaq with the Lycian verb prñnawa- ‘to build’,very common in Lycian funerary inscriptions (ebe ñne x upa prñnawat e X, . . . ‘X has built this tomb . . .’ and variants), Hajnal tries to connectfinal -q with the 1st singular active preterite ending in Luwic languages:

19 See Zgusta (KPN § 108) and here Chapter 11 s. v. artmi.

291

CLuw. -¢a , Lyc. - x a, -ga . The advantage of this hypothesis is that itconvincingly resolves the structural problems posed by the text; an si d i

would be therefore the accusative of a verb meaning ‘to make’ (or,with a more specialized meaning, ‘to build’), and the problem of anonomastic formula with an individual name in nominative would nolonger be relevant, as the text would read, ‘This tomb (acc.) Artmi,(son) of Pau, I made’. This is undoubtedly a very attractive hypothe-sis, but it is seriously weakened by the doubts surrounding the exactphonological value of % <ã>.

sñis : sdisa-/s : psu≤ol ≤ / mal≤ : mno≤ C.Ka 1

The main difficulty of analysing of C.Ka 1 lies in how to interpret theinitial formula sñis s d isas , where a word s d i , undoubtedly a variant formof si d i with defective vowel notation, can be identified. In psu≤ ol -≤ mal ≤ mno≤ , it is not difficult to recognize a twofold formula, with both per-sonal names in genitive ( psu≤ ol -≤ , mal ≤ ). The well-known Carian wordfor ‘son’, mno-≤ , also in genitive, could agree with the name of thedeceased and govern the other genitive, mal-≤ (‘of Psu≤ol , the son of Mal’), but, as in C.Eu 1, an asyndetic construction must not be ruled

out: ‘of Psu≤ol (son) of Mal, (and) of (his) son.’As for the initial words, sñis s d isas , I believe it very unlikely that sñis can be a personal name, although not impossible. The hypothesis thatwill be considered here argues instead that it is part of the referenceto the tomb. The main problem is how to reconcile s d isas with theform an si d i in C.Eu 1. One could attempt the segmentation s d i sas and recognize here the diff erent pronominal stem sa-/san-/snn-, but thefinal s would then remain unexplained. Without dismissing this possi-bility, I wish to propose an alternative: if we intend to find here thesame pronominal stem of an, the only possible procedure is to use thesegmentation s d is as . This gives a sequence sñi-s s d i-s a-s , which seemsto contain a common ending -s for all three words. Could this ending be a plural mark? If this were the case, sñis could be interpreted as aplural form of the pronominal stem san-/sn- (sñ-is ), and the overall senseof sñis s d isas could be ‘These (are) the burials . . .’ or ‘These burials arethose . . .’, completed by the genitives that follow (‘. . . of Psu≤ol (son)of Mal (and) the son’. If I must be honest, this latter translation is

somewhat forced, as it assumes an asyndetic construction, which is apossible, but not the only, interpretation of the whole sequence (seeabove). However, the link between the presence of s d isas and a collective



292

tomb is much clearer in the other inscription in which this form appears(C.Kr 1):

qot2omu sdisa-s?n≤ “odubr≤ or rather : mn≤ “odubr≤?sb mno≤ knornoril?ams or rather : norimams?

In this inscription, the existence of more than one burial can be deducednot only from the clearest part of the inscription, “ od ubr-≤ sb mno-≤ ,which can be translated ‘of ”odubr and the son’, but also from thefact that the tomb clearly contained three burial chambers, as pointed

out by Olivier Henry (pers. comm.). Unfortunately, the overall struc-ture of the inscription remains very unclear. It is possible that qot 2omucould be a personal name, which would allow us to recognize the ref-erence to three individuals (qot 2omu, “ od ubr-≤ and one son, mno-≤ ), butthe syntactical connection of qot 2omu with the rest of the inscription,and the sequence n≤ that appears after isolating s d is-as , remain unex-plained. Note also that the reading of the final s in s d isas and the sub-sequent segmentation are far from conclusive. If we accept (with somereservations) the reading s d isas, we could tentatively envisage that qot 2omuis not a nominative functioning as subject, but rather a nominative of appellation, designating the owner of the tomb. In this case, n≤ couldbe a resumptive pronoun referring to Qot2omu: ‘Qot2omu. These tombs(are) of him, of ”odubr, and of the son . . .’ Needless to say, this inter-pretation is more a desideratum than a fact based on solid evidence.As for knor norilams (or knor norimams ), practically nothing can be said.20

The two remaining inscriptions that contain s d i- are too fragmen-tary to attempt an interpretation. In C.Tr 1, two personal names, one

in nominative (artmon ) and the other in genitive ( pau≤ ) can be identified,but si d i appears followed by an unclear and incomplete sequence, amt [.The case of E.Al is worse still, where only s d i a[-]mob[ is legible.

The last funerary inscription of this sub-corpus, although completeand without reading problems, thanks to the recent autopsy made byDiether Schürr, is very brief and contains no specific word to desig-nate the tomb:

“oru≤ / ann ibrs≤ C.Ka 3

20 Only as a purely assonant connection can we resort to Lyc. nere/i-, a term of relationship whose exact meaning is unknown (Therefore sb mno≤ k-nor noril ? ams ‘and of the son and(?) the nor nori- relatives’?).



293

The segmentation proposed here (already suggested in Adiego 1996)seems to me the most plausible, although other possibilities exist.21

According to this segmentation, we can identify two names in genitive(“ oru≤ , i b rs ≤ ) and the word ann, probably a demonstrative pronoun thatcan be related to an (see above). The meaning of the inscription wouldtherefore be ‘This (tomb is) of ”oru, (son) of Ibrs’ or, simply, ‘That (is)of ”oru, (son) of Ibrs’.

In conclusion, the study of this sub-corpus of funerary inscriptions isseriously limited by the scarcity of materials, and the divergent for-mulae employed—despite the coincidence of several formulaic words(s(i)d i, ≤ (j)as , and some demonstrative pronouns—cannot be reduced toa single, stereotyped model of expression.

3. The Longest Gra ffi to from Abu Simbel (E.AS 7)

While a great number of the Carian graffiti from Egypt contain onlyonomastic formulae (usually N-Ø or N-Ø N-≤ ), some graffiti are moreextended and probably also include common lexicon. However, thismaterial cannot realistically be analyzed; for some important sub-cor-

pora there are no updated and reliable editions, and in the case of theAbu-Simbel and Buhen graffiti, for which we are lucky enough to haveOlivier Masson’s reliable editions, the reading difficulties are such thatin most cases they cannot be used with any degree of confidence. Forthat reason, I will limit my analysis to a long graffito with certain par-ticularities that have attracted the interest of scholars.

naz ∞i∞ | b ÿ≤ | esak?dow“ | m ÿqudem | pisma≤k | bebint | mo | ne |ps ÿ“[|? ] ai[ - ]iqom

The most striking feature of this inscription, engraved by a Carian sol-dier during the Nubian campaign in 593/592 BC, is the suspicion thatthe personal name pisma ≤ k alludes directly to the Pharaoh PsammetichusII, under whose supreme command the military campaign was under-taken; this would closely link the graffito to the well-known Greekinscription, also from Abu Simbel, in which the same Pharaoh is men-tioned.22 It is true that pisma ≤ k (and variants) is well documented as a

21 For instance, to isolate only two personal names, “ oru≤ and anni b rs ≤ . Also theoret-ically possible is the segmentation anni b rs ≤ .22 Bernand-Masson (1957), inscription nº 1: basil°ow §lyÒntow §w ÉElefant¤nan

Camat¤xo, / taËta ¶gracan to‹ sÁn Cammat¤xoi tÇoi YeoklÇow / ¶pleon, ∑lyon d¢ K°rkiowkatÊperye, uÂw ı potamÚw / én¤h: éloglÒsow dÉ ∑xe Potasimto, Afigupt¤ow d¢ ÖAmasiw:



294

name often used by the Carians in Egypt, but the appearance of theword esak ? d ow “ in the graffito is favourable to this hypothesis: k d ow º

/kndowº/ (or its phonetic result) has been convincingly compared toLycian x ñtawa- ‘to rule’ (cf . also x ñtawat(i)- ‘ruler, king’ = CLuw ¢andawat(i)-‘supreme authority, king’).23 Therefore, in esak ?d ow “ . . . pisma ≤ k , we couldattempt to find a meaning ‘king Psammetichus’ or similar. Both theinitial sequence esa- and the precise function of final -“ are problem-atic, and diff erent possibilities can be envisaged: ese- could be comparedwith the Lycian preverb ese-, for which a meaning ‘with’ has been pro-posed (see Melchert DLL s. v.). Perhaps in this instance it may havea reinforcing function. As for -“ , two hypotheses can be considered: itcould be a suffix attached to the verbal stem k d ow- ‘to rule’ in orderto create a noun (cf . Lycian -za ?), or it could be a nominative pluralending. This latter theory is less consistent with an interpretation of pisma ≤ k as the name of the Pharaoh, since he would appear at the samelevel as the other alleged commanders. Moreover, in the rest of theCarian graffiti there is no form that can be likened to the names of military officials cited in the corresponding Greek inscription.

Ultimately, these doubts cannot be resolved, as the rest of the inscrip-

tion remains unclear. It is possible that both the identification of pisma ≤ k with Psammetichus and the equation of k d owº = Lycian x ñtawa- arecorrect, but the overall interpretation of both words as ‘king Psammetichus’may still be erroneous.

D. T L I

Some of the Carian inscriptions found in Caria proper are more extended

than the rest. We may assume that their content can be broadly definedas legal: honorary decrees, regulation of cults, lists of priests, and soon. As the study of fragmentary languages shows, the longer texts arethe most difficult to analyze. A good example is our meagre knowl-edge of Milyan vis-à-vis Lycian: having at our disposal a text as long

/ ¶grafe dÉ èm¢ ÖArxon ÉAmoib¤xo ka‹ P°leqow OÍdãmo. “When the king Psammetichus

came to Elephantina, this was written by those who sailed together with Psammetichus,(son) of Theokles and who came beyond Kerkis, as far as the river permitted. Thosespeaking foreign tongues were leaded by Potasimto, the Egyptians by Amasis. We havebeen written by Arkhon (son) of Amoibikhos, and Pelekos, son of Eudamos”.

23 Adiego (1995:19–20).



295

as the Milyan inscription in the Xanthos stela does not necessarily guar-antee an understanding of the grammar and lexicon of this language,

but in fact implies quite the opposite. It is preferable to work with asizeable corpus of relatively brief texts—as is the case with Lycian— than to be restricted to a single, very long, and generally impenetra-ble text.

This is also the situation with Carian. Texts such as the Kildarainscription are currently impossible to analyse. We can identify someisolated words, but we know nothing about the context in which theyappear.

The only way of ‘getting inside’ such a text is through the existenceof a translation. Of the entire Carian sub-corpus of longer inscriptions,the only inscription that is accompanied by even a minimally legibletext, and whose content corresponds undoubtedly to the Carian text,is the bilingual inscription of Kaunos. But even in this case, the resultsthat we are able to obtain are very limited. Our analysis of longerinscriptions will begin therefore with this example.

A problem that seriously impedes the interpretation of this type of inscription is our current inability to identify verbal forms. Without

knowing which words represent verbs it is practically impossible todevise any approach to analysing the structure of the text and the func-tion of the common nouns that it contains. Even in the bilingual of Kaunos, where the existence of the Greek inscription ought to helpidentify verbal forms, no agreement has been reached on which of thewords must be identified as verbs. As we have already seen, the prob-lem of identifying verbs also aff ects shorter inscriptions, and only witha substantial increase in the Carian documentation available will webe able to resolve this great problem.

1. The Kaunos Bilingual Inscription

We will begin with the only text that can be read alongside a paral-lel Greek text. The possibility of comparing the two texts explains thespecial attention that this inscription has received since its discovery.Even the editors of the inscription, Frei and Marek, did not limit them-selves to a simple epigraphical edition, but off ered in addition a first

attempt to establish the parallel passages in both texts, paving the wayfor further research.24 Following this initial study, other scholars have

24 See Frei-Marek (1997), (1998).



296

drawn up essays of interpretation for the overall text or for concretesections.25 In the present work, I follow in general terms the views

adopted in Adiego (1998a), completing the discussion by referring tothe works cited above. No real developments have appeared since then,and the initial enthusiasm inevitably generated by the discovery of sucha unique document has clearly diminished, since the results that havebeen obtained are relatively modest.

Of the 18 lines of the Carian inscription, the current possible inter-pretations are in fact limited to the first 10, essentially for three rea-sons: (1) these 10 lines correspond roughly to the conserved part of Greek inscription, the rest of which has been lost; (2) lines 11–12 of the Carian text is unfortunately very difficult to read because of thecrack in the stone, which creates a lacune that makes it very difficultto continue the analysis of the text; (3) the last lines of the Carian textare practically impossible to interpret, since most of the forms are hapaxlegomena and display no external characteristics that could help theirunderstanding. The segmentation of words in these final lines is also avery difficult task.

The identification of proper names that appear in both texts has

proved fundamental to an analysis of the comparable parts of theinscriptions. The correct interpretation of sb as a coordinative conjunction is also relevant, as it thus becomes an important tool for estab-lishing the phrases and sentences of the Carian text. A third decisivefactor, already noted by the editors, is the existence of textual paral-lelisms with the other large text from Kaunos (C.Ka 2).

Kaunos bilingual (lines 1–10) Greek version

kbidn uioml n i [---] ¶doje Kaun[ ¤ ]oiw §p¤ dhmio[ u ]-inis drual nik[--] rgoË ÑIpposy°nouw: Nikok-lan lysiklas[-? ] l°a Lusikl°ouw ÉAyhna›o[ n ]otonosn sb lys[ikl] ka‹ Lusikl°a Lusikrãt[ ouw ]an lysikratas[-? ] [ ÉA ]yhna›on proj°nouw e[ ‰nai k-]otonosn sarni[ “ ] a‹ eÈerg°taw Kaun¤v[n auto-]mdot2 un sb undo[--] Áw ka‹ §kgÒnouw ka‹ [-----]tl “ kbdyn“ sb b2o[--] n auto›w §[ . . .ol“ otr“ sb a∞t [ms]-kmt absims sb [---]

25 See among others Hajnal (1997b), Melchert (1998), Neumann (1998).



297

A quick look is sufficient to identify the correspondence of all the propernames that appear in both texts. Even in the only case where the words

are not formally related—the reference to the Kaunians, reflected inCarian by means of kbidn, kbdyn“ —Lycian evidence ( X bide is the Lycianform for Kaunos) dispels any doubt. All these correspondences havebeen underlined in the table above.

The onomastic formulae appear clearly in accusative case (markedwith the ending -n: nik [ -- ]lan, lys [ ik ]lan ), which allowed the first editorsto correctly identify the words ending in -“ as plural accusatives . Thismeans that the words that present this ending (including sarni [ “ ]) couldrepresent the Carian equivalent of the Greek accusative plural formulaproj°nouw . . . [ k ]a‹ eÈerg°taw Kaun¤v[ n auto ]Áw ka‹ §kgÒnouw. The factthat the Carian formula contains more words that the corresponding formula in Greek made it difficult to establish the precise connectionbetween the two texts, and the attempt the first editors led to mistakenassumptions. The identification of the word otr “ with Lycian atra-/etli-,‘person, self ’, made independently by various authors, has clarified thesituation decisively; it implies that the Carian passage corresponding tothe Greek formula proj°nouw e[ ‰nai k]a‹ eÈerg°taw Kaun¤v[ n ], which

immediately precedes [ auto ]Áw, must be identified in the more extendedconstruction, sarni [ “ ] mdot 2 un sb und o[ -- ]t l “ kbdyn“ sb b 2o[ -- ]ol “ , whichin turn precedes otr “ .

The possible explanation for this extended formula in comparison tothe Greek version will be dealt with later. Before addressing this prob-lem, we must focus our attention on what is undoubtedly the mainobstacle to interpreting these first lines of the Kaunos bilingual: theidentification of the possible verbal forms. The need to first locate theverbs in the bilingual is clearly justified by the fact that some of the

forms appearing in the sequence sarni [ “ ] . . .b 2o[ -- ]ol “ have entered intothe discussion about possible verbal forms.

The Greek construction of the sentence is clear: an impersonal verb,‘it seemed good’, constructed with dative (‘to the Kaunians’) and aninfinitive clause as the subject of the impersonal verb (‘that Nikokles(. . .) and Lysikles (. . .) were public guests and benefactors of theKaunians’). As Greek syntax demands, the subject and nominal pred-icate of the infinitive clause are in the accusative case. Much discus-

sion has arisen as to whether the Carian text contains, or indeed cancontain, a construction of this kind, or whether the formula of theproxeny decree is instead expressed in Carian in a very diff erent way.The only clear indication that the Carian text off ers is the fact, mentioned



298

above, that the personal names of the honoured Athenian citizens arein accusative. This implies at least a verb that either directly or indi-

rectly governs these accusatives, and all the proposals formulated todate attempt to find verbs among the words of these first lines. Scholarshave adopted two diff ering approaches to the question; while some of them have assumed that there is only a verb in personal form, with-out the actual presence of an infinitive, others have claimed to recog-nize an infinitive form, which would indicate the further existence of a main verb or equivalent construction.

The options suggested as possible infinite or finite verbal forms areuioml n, mdot 2un, and un (2×; from segmenting mdot 2 un and un d o[ -- ]t l “ ).Although the verbal character of rual has also been considered, in the-ory there is a certain consensus to interpret it rather as a noun, withina formula equivalent to Greek §p‹ dhmio[ u ]rgoË ÑIpposy°nouw.

The first editors adopted the approach of postulating a finite verbconstruction. In fact, they claimed to have identified two finite verbs,uioml n and mdot 1un (interpreting O t 2 as a glide sound between o andu ). Both would be third person plural preterite verbs with the respec-tive meanings ‘to decide’ and ‘to make, to invest as, to establish as’

(Frei-Marek 1997:29–30), and the accusatives would be dependent onthis latter verb.Melchert (1998) suggested that the accusatives are directly constructed

with a finite verb. But unlike Frei-Marek’s proposal, he prefers to clas-sify uioml n as a noun with the meaning ‘decree’ (kbidn uioml n ‘decreeof Kaunos’, with kbidn as plural genitive of the place name),26 andalthough he agrees with the first editors in considering mdot 1un (alsointerpreted as mdo/w/un, with O t 2 as a glide) to be a finite verb gov-erning the accusatives, he analyzes it as a fi rst person plural preterite,

‘we have established, we have install (as)’. A weak point of Melchert’sinterpretation is that the analysis of this word as a first person verbdepends on the assumption that O represents a /w/ sound, as he com-pares an alleged ending -/wun/ with Hittite preterite first plural end-ing -wen and with the corresponding Lydian ending -w n .

The proposal in Adiego (1998a) is very diff erent. Here, the suggestedsegmentation is mdot 1 un (although without considerations about thephonological value of O t 2, which was studied in a subsequent paper,

26 For a place name kbid - as plurale tantum, Melchert reminds us of the similarinterpretation for the Lycian name of Kaunos, X bide (Melchert 1998:37).



299

Adiego 2002), and both this un and the un segmentable in und o[ -- ]t l “ are taken as in fi nitives depending on the finitive verb uioml n, for which

Frei-Marek’s interpretation as ‘they decided’ is accepted. As for un asinfinitive, it is compared with the Luwian infinitive ending -una (aduna ‘to eat’) and the whole form is considered as the infinitive of a verb‘to make’, by comparing Lycian a(i)-, Cuneiform Luwian à-/à ya- ‘tomake’. The remaining word mdot 2 is then interpreted as a complementof sarni [ “ ] that would therefore constitute a locution equivalent to Greekproj°nouw. This hypothesis was refined in Adiego (2002), where argu-ments were given in favour of O as the Kaunian form of the letter ct , and for the interpretation of endings in -ot , -ot 2 as plural genitives.

These hypotheses do not cover every theoretically possible interpre-tation of each of the words involved, and it is impossible with our cur-rent knowledge of Carian to choose one above the rest, or to simplydeem them all incorrect. In any case, all these analyses coincide in see-ing sarni [ “ ] (or sarni [ “ ] mdot 2 ) and und o[ -- ]t l “ (or d o[ -- ]t l “ ) kbdyn“ as theCarian equivalent to Greek proj°nouw [ k ]a‹ eÈerg°taw Kaun¤v[ n ]. Thisinterpretation seems to be one of the scarce certainties—leaving asidethe onomastic forms—that currently exist about the structure and mean-

ing of the Carian text. However, it does pose a serious problem, inso-far as between kbdyn“ and otr “ = aÈtoÊw we are left with sb b 2o[ -- ]ol “ ,which has no recognizable correspondence in the Greek part. It mustbe remembered that this sequence led the first editors to make mis-takes, because they assume that sb b 2o[ -- ]ol “ is the Carian equivalent of Greek ka‹ §kgÒnouw ‘and the descendents’, and otr “ was therefore inter-preted as a possessive (‘of them. their’). This intepretation was undoubt-edly caused by the fact that b 2o[ -- ]ol “ is linked to the former text bysb, which made it difficult to find an equivalent to aÈtoÊw, since itcould hardly be preceded by a coordinative conjunction.

The Lycian etymology for otr “ mentioned above showed that thisinterpretation was mistaken, since otr “ must be the Carian word trans-lated into Greek as aÈtoÊw. But this explanation, based on a very con-vincing etymological approach, fails to account in any way for sbb 2o[ -- ]ol “ .

A provisional solution was suggested in Adiego (1998:23): b 2o[ -- ]ol “ could be another ethnic name coordinated with kbdyn“ , referring to

ÖImbrow, the fortress near Kaunos27

that Bean (1953:22) identified with

27 Strabo XIV, 651.

300

the archeological remains on the summit of Ölemez Da< , the moun-tain north of Kaunos. This hypothesis would fit well with the initial b2

of b 2o[ -- ]ol “ , but the fact that the Carian word is incomplete makes itless certain. Moreover, it is not unthinkable that sb is in fact coordinat-ing b 2o[ -- ]ol “ not with kbdyn“ , but directly with sarni [ “ ] and und o[ - ]t l “ /d o[ -- ]t l “ , so that the Carian equivalent to the Greek twofold formulaproj°nouw [ k ]a‹ eÈerg°taw would be a threefold one, (sarni “ . . . (un)d o[--]t l “ . . .b 2o[ -- ]ol “ ).

The identification otr “ = aÈtoÊw makes it likely that the following sequence, sb a ∞t [ ms ]kmtabsims sb [, corresponds to Greek ka‹ §kgÒnouwka‹ [.28 In Adiego (1998a) an etymological connection for a ∞t [ ms ]km(t? )was proposed, based on the identification of mskmº with the Luwianstem mas ¢a- recognizable in ma “¢à¢it- ‘growth, prosperity’, and the sug-gestion that a ∞t- could be compared with Hittite katta, kattan ‘down’.a ∞t [ ms ]km(t? ) could therefore be a compound noun formally and seman-tically comparable to ‘off -spring’, German ‘Nach-kommenschaft’ andSpanish ‘de-scendencia’. As for the remaining ºtabsims or ºabsims , it wassuggested in the same study (Adiego 1998a:25) that this could be apossessive, comparable to Lycian ehbi(je) < *ebesi(je)-. All these con-

nections are phonologically sound, but they fail to explain ºt º (if it doesnot belong to the word a ∞t [ ms ]km ) and the final -ms .From absims sb on, the possibilities of interpreting the Carian text

decrease dramatically: both the sudden interruption of the Greek part,and the lacuna of almost an entire line (l. 12) make it impossible tocontinue with the analysis. We can identify further possible instancesof the conjunction sb ‘and’, but it is impossible to identify which typeof syntactic elements it is coordinating. Among the words of this finalsection of the inscription, special attention should be paid to the sequence

orouo, which recalls Lycian arawa- ‘freedom’, a term that is consistentwith the sort of privileges conferred on the proxenoi in these types of decrees. However, the segmentation is far from certain, as it leaves anisolated letter H between sb and orouo. Also of interest here are the pos-sible presence of the word mno- ‘son’ under the form mnos , and thefinal sequence aitusi , where we could identify a possible imperative aituor preterite ait of a verb corresponding to Lycian a(i)- ‘to do, to make’(cf . Lycian ait e ‘they made’).

28 The integration [ ms ] is based in the parallel sequence a ∞tmsk [ in C.Ka 1.



301

2. The Kaunos Inscription C.Ka 2

[-(?)-ui? ]oml ã qrdsg rdso[-]i[

[-?-]r sb a∞mnnartnyr obsmns[ [-?-]∞arl anoã sb z“ariosã i∞[ [-?-]nudrma ∞ yrpai sarni“ sb u[ [-?-]aH punot2 otr“bi sb a∞tmsk[m[-?-]dbi _1aitk ouor gdb“l aã1_i[-][-?-] sarni“ sb 1orsol“ sb uHbit[-?-]bi qrdsol “ ait _1mali H∞it[-?-]intnor ∞ yrapai≤ umot2 oba[-?-]_iurt obsmsmñ1ñ ouor mt1_ yr[-?-][-]abrun∞ur[-]“ yn“H ynn sb vacat

[-?-----------------]tbsms _1mali[ [-?------------------]maH 1sb an[ [-?----------------------]ba vacat

C.Ka 2, the most extensive Carian inscription, remains a virtuallyimpenetrable text. The discovery of the Kaunos bilingual has allowedus to establish some connections between the two texts, and it is cer-tainly very significant that the word sarni “ , which in the new bilingualseems to represent either totally or partially the Carian expression for

‘proxenoi’, is also present in C.Ka 2. Other important correspondencesare the word otr “ = aÈtoÊw, the possible word for ‘descendence’, a ∞tmskm(t?),and the noun or verb that appears apparently to indicate that theinscription is a decree (uioml n, [ -- ]ml n in C.Ka 2). All of these ele-ments seem to suggest that the content of C.Ka 2 and C.Ka 5 has asimilar purpose, but these correspondences are not corroborated by acomparable overall structure. In this respect, the absence of clear ono-mastic formulae, which we would expect to find if dealing with theconcesion of proxeny to more than one individual (note the plural

sarni “ ), is very puzzling.In the first line, after ]ml n, a word qrds can be segmented, recalling

a similar sequence in Kildara (C.Ki 1). In this case it seems to be fol-lowed by another word etymologically related to it, g rdso[, which pointsto a possible fi gura etymologica . Another derivative of qrds appears in line8, followed by a word ait : qrdsol “ ait . The final -“ points to an animateaccusative (and perhaps also nominative) plural, and ait recalls, asMelchert (1998:35) has pointed out, Lycian ait e ‘they made’, so that an

interpretation as ‘the qrdsol -s made’ or ‘they made (them) qrdso

l -s’ seemslikely. The presence of qrds immediately after ]ml n, the noun or verb



302

that initiates a decree, implies that qrds could be referring to the cityor to a city institution such as the assembly.29

sarni “ appears twice, in both cases followed by sb, ‘and’. In the firstinstance, it is impossible to decide whether sb is joining sarni “ to anotherplural noun, due to the crack of the stone. The second example ismuch clearer: sarni “ sb 1orsol “ is undoubtedly a sequence of two accusative(or nominative) plural nouns. The first editors of C.Ka 5 tried to com-pare this latter word with C.Ka 5, l. 8–9 b 2o[ -- ]ol “ , and proposed theintegration b 2o[ rs ]ol “ . Although this hypothesis is very attractive, it iscomplicated by a serious difficulty: ÿ b 2 seems to be formally closer to± b (C.Ka 4) than to 1.30

The only form that can be identified as a possible proper name isl. 4 ∞ yrpai (nominative?), which reappears as genitive singular in l. 9(∞ yrapai-≤ ).

The clearest syntactic parallel between C.Ka 2 and the Kaunos bilin-gual is l. 6–7, otr “ bi sb a ∞tmsk[m . . .] d bi = Bilingual otr “ sb a ∞t[ms]kmt absims ‘themselves and (their) descendants’. Unfortunately, it is impos-sible to calculate the letters missing between a ∞tmsk [ and ]d bi , but itseems likely that =bi was attached to (a form of ) the word a ∞tmskm(t ?),

so that the formula in C.Ka 2 would present a ‘X= bi and X= bi ’structure: otr “ =bi sb a ∞tmsk[m . . .] d =bi . The exact function of this = bi is unknown. It could be a particle reinforcing the coordination (‘boththem and (their) off spring’, ‘not only . . . but also . . .’), but it could alsobe a postposition (‘for’, ‘with’?). Both interpretations permit an etymo-logical connection with Lycian -ppi in hr-ppi ‘on; for’ (cf . for this formLyc. hri = CLuw. “ arri ‘up; (on) top’). Carian =bi could be a generalisa-tion of the lenited form of this particle instead of Lycian unlenited -ppi (cf. the parallel process in Lycian -be vs. CLuw. -ppa , Melchert DLL

s. v.).

3. Sinuri’s Longer Inscription (C.Si 2)

(a) [--]ryin ∞tmño≤ : sb ada ∞tmño≤eri : pisñoi mda : pñmnn≤ñ : pda-∞m≤uñ ∞i “aoyr∞ri mt∞elã

29

For a purely theoretical etymology suggested in Blümel-Adiego (1993:94) seeChapter 11, s. v.30 That 1_ and ± are diff erent letters (which rules out a triple equivalence 1_ = ±

= ÿ ) is confirmed by their co-occurrence in C.Ka 4.



303

ñmail o mda lrHñ : stspñ vacat sm“s[--5--] sb añmsñsi mda

sm[--7--]a∞e[ ∞[--8--]tuñdñ[ ñe-?-[

(b) pim[. . .]Ha?[ . . . ]

The longer inscription from Sinuri, engraved after a long Greek text,could be a true bilingual, but the Greek text is so badly damaged thatit turns out to be practically unusable as a means of understanding the

Carian one.The greatest progress in the analysis C.Si 2 has been made by DietherSchürr: he was able to identify in the first line of the text a referenceto the Carian dynasts Idrieus and Ada, sons of Hekatomnos ([ -- ]ryin∞tmño≤ : sb ada ∞tmño≤ ). Moreover, Schürr argued convincingly that itbelonged to the same stela as a fragment of a Greek inscription alsodiscovered in Sinuri. This fragment adds very limited, but nonethelessinteresting information about the content of the Greek inscription pre-ceding the Carian one, and it is certainly more valuable than the infor-

mation that can be obtained from the scarce remains of the Greek textin the bilingual fragment. The Greek unified text contains two decreesof the abovementioned dynasts, both conceding an ét°leia (tax exemp-tion). A personal name, Ponmoonnow, is also mentioned in genitive inthe first Greek decree, and Schürr also found it in the Carian text(l. 2 pñmnn≤ ñ , alternative reading ? pñmun≤ ñ ). This could be the samename that appears in later Sinuri inscriptions in the form Pormounow,as the denomination of the syngeneia devoted to the administration of the temple.

In Adiego (2000), I attempted to make some progress in the under-standing of the text. Many of the proposals outlined in that study werevery hypothetical, and I will not repeat them in any detail in the pre-sent book. Perhaps the main idea that should be retained is my insis-tence on searching for verbal forms: it is logical to assume that someverbal form must exist not very far from the mention of the Hekatomnids,expressing the action carried out by the dynasts, in the same way thatin the second Greek text the names of the dynasts were very proba-

bly followed by a verb ¶[ dvkan ] (if Robert’s very plausible integrationis accepted). In Adiego (2000) it was suggested that both pisñoimd a (l. 2)and ñmai l omd a (l. 4) were verbs. In the first case, a connection was

304

proposed with the Common Anatolian verb for ‘to give’ (Hitt. pài-/piya-CLuw. piya-, Lyc. pije , etc.); in the second, I suggested a rather kling-

klang resemblance to Lycian ◊ mmait e in the trilingual of the Letoon of Xanthos, a third person plural verb with a possible meaning ‘to install,to build’. In both alleged verbs, md a was interpreted as the same ele-ment that appears as md ane in other Carian inscriptions, which I con-sider to be a particle chain (see p. 324). As for the resulting sequences pisñoiº, ñmai l oº, o (in both forms) and i (in pisñoiº ) were interpreted asclitic pronouns (Adiego 2000:142):

Verb Direct Object Indirect Object -md a element Translation

pisñ- -o- -i- -md a ‘they gave-it-to him/tothem’

ñmai l - -o- -md a ‘they established it/him’

These postclitic pronouns, attached to the verb, were syntactically com-pared with the well-known ‘nasalized preterites’ of Lycian ( pijet e, ˜ mmait e ),where the nasalized vowel comes from the univerbation of the verbwith a clitic pronoun ( pije-t e < *-to+om ‘he gave-it’).

The rest of the ideas contained in Adiego (2002) are merely hypo-thetical. In the sequence pñmnn≤ ñ: pd a ∞m≤ uñ ∞i a construction with therelative ∞i was envisaged, and two possible interpretations were sug-gested: ‘Pñmnn- who (is) the priest’ or ‘the syngeneia of Pñmnn-’. Bothinterpretations for pd a ∞m≤ uñ were considered to be consistent with anetymological connection of pd aº with Lycian pdd en- and Hitt. peda-,‘place’. For eri , I proposed a possible word for ét°leia, on the basis of its resemblance to Lycian arawa-, ‘exemption (ét°leia ), freedom’, but

without discounting the alternative view suggested by Neumann (apud Adiego 2000:144) of eri as preverb identical to Lycian eri (cf . Hitt. ar ¢a ).Leaving aside all these speculative attempts, the most conclusive and

remarkable morphological information provided by the bilingual inscrip-tion of Sinuri is the form pñmnn≤ ñ , which must surely be an accusativus genitivi or, more correctly, an accusative formed on the basis of the (old)possessive adjective, comparable to the Lycian forms in -hñ (Lusãñtrahñ ).The same analysis is probably also valid for pd a ∞m≤uñ , although thepresence of u between ≤ and ñ obliges us not to disregard alternative

interpretations.

305

4. The Hyllarima Inscription (C.Hy 1)

The recently discovered bilingual stone in Hyllarima has proved to bea fragment that fits well alongside what was until recently the onlyknown inscription from that location. Joining the two pieces togetherhas provided us with a complete version of the Carian part of the orig-inal stela with practically no problems of reading, and has also helpedus to understand the Greek texts that accompanied the Carian inscrip-tion. Of these, the only one that seems to be contemporary or veryclose to the Carian texts is the brief list of ‘priests of all the gods’ incolumn B:

Column A Column B

“asqariod dymda kdu≤opizipususotmuot armotrqdosq mol“ msot ylarmit

brsi ari“≤ brsi≤ flere›ew ye«n pãntvn:mane : u≤ol ≤ ÑErm¤aw Fan°v ÑErm¤adow:rtim u≤ol ≤ pur?i≤ flereÁw ye«n pãntvn:

ÑUssvllow ÉArrissiow

u≤bzol tñu≤ brsi≤pau mane≤ ybr-s≤(A later Greek inscription containing (A later Greek inscription containing aa list of priests of Apollon follows) priesthood sale follows)

(Recall that other, later Greek inscriptions appear on both sides of theinscription, see p. 136, n. 25 for details).

Any attempt to analyse the Carian text is seriously hampered from

the beginning by the doubts about the exact order in which the firsttwo lines of both columns (in fact, the only part containing commonvocabulary, since the remaining lines contain only onomastic formulae)should be read. As pointed out in Adiego-Debord-Varinlio< lu (2005),three posibilities can be considered: (1) to read the first line of eachcolumn and then the second line of each column (A1 + B1 +A2 +B2); (2) to read the two first lines of column A and followed immedi-ately by the two Carian lines of column B (A1 + A2 + B1 + B2); (3)

to read all of column A before reading column B, that is, to treat thecolumns as (relatively) independent texts (A + B). The existence of a divid-ing line bewteen the columns seems to support the last solution, butgiven that the two columns are apparently lists of priests, it is strange

306

that an identical word for ‘priests’ cannot be found in each column.This option would also raise another intriguing question: why are the

names of the priests in the first column in Carian and the names inthe second column in Greek , despite the fact that Carian is used for theheadings of both columns?

Both solution (1) and solution (2) imply that the two first lines of both columns would be the common heading for the entire Carianinscription. But there are no clear arguments for or against either solu-tion: the first lines in both are for now impenetrable, and in the caseof the second lines, although they do contain words that can be inter-preted (see immediately below), their structure does not allow us toclarify whether or not they consitute a consecutive sequence.

From the two initial lines of both columns, the only forms for whichwe can find a plausible interpretation are armotrq d osq and mol “ msot ylar-mi t . The first, armotrq d osq , is undoubtedly the most relevant contributionof the new Hyllarima fragment: its interpretation is straighforward— although the syntactic connection with the rest of the text remainsunclear—and it is of great interest from both a linguistic and a cul-tural perspective. In armotrq d osq , it is easy to recognize two typically

Anatolian god names: armo-, the name of the moon good (Hitt. Arma- ), and trq d -, the storm god already documented in other Carianinscriptions (Kildara, Iasos). As noted in Adiego-Debord-Varinlio< lu(2005), the simplest explanation of armotraq d - is to see it as a dvandvacompound (“Arma (and) Tarhunt”), which would allude to worshipdevoted to a divine paredria.31 As for the morpho-syntactical interpre-tation of armotrq d osq , we can identify here an ending -os (assuming thattrq d - retained its original nt-stem), followed by q , perhaps some sort of enclitic connecting particle (in Adiego-Debord-Varinlio< lu 2005, I pro-

pose an etymological link with the enclitic coordinative conjunctionCLuw. -¢a , ‘and’), although it is by no means clear what this -q is actu-ally connecting.

Regarding mol “ msot ylarmi t in column B, in Adiego (2002:17) it wasanalysed as a formula corresponding roughly to Greek flere›ew ye«npãntvn ‘priests of all gods’, which would mean ‘priests (mol “ ) of theHyllarimean gods (msot ylarmi t )’. Although this interpretation requires

31 Other solutions seem less convincing, see Adiego-Debord-Varinlio< lu (2005:616–617)for details.

307

us to make a series of (sometimes fragile) assumptions about the mor-phology and semantics of the three words involved, at least in the case

of ylarmi t , it seems plausible to consider a connection with the placename Hyllarima, as suggested for the first time by Ray (1988:152).32

The interpretation of msot as, ‘of the gods’, is based on the possibilityof connecting it with the Luwic word for ‘god’: note particularly Milyanmasa- ‘god’, and also Lyc. maha(na)-, CLuw. mà““ an(i)-. The stem alsoappears in Carian onomastics: msnord-, Masanvrada (see Chapter 11s. v. msnord ≤ ). Less definite connections can be established for mol “ =‘priests’,33 where moreover the character of the final -“ (an ending orrather a derivative suffix) remains unclear (see p. 318), although thenew inscription from Mylasa (C.My 1) now offers another example of mol “ in the first line, preceding a long list of individuals. Even if themeaning ‘priests’ proved to be incorrect, the appearance of mol “ in theheading of two inscriptions consisting of lists of onomastic formulaewould still be worth noting.

From a morphological point of view, this interpretation of mol “ msot ylarmi t assumes that mol “ must be in plural, presumably in nominative(as in the Greek formula), although the possibility of another case can-

not be totally ruled out. Following the same interpretation, msot ylar-mi t must by contrast be a genitive plural , characterized by the ending -t .On these morphological repercussions of interpreting mol “ msot ylarmi t as ‘priests of the Hyllarimean gods’, see p. 319.

The remaining lines of column A consist exclusively of five onomasticformulae, engraved at diff erent times (see Adiego-Debord-Varinlio< lu2005). With the exception of the second formula, where only the namesof the individual and the father are recorded, the rest of the formulaeare threefold: individual name, father’s name and grandfather’s name.

It is interesting to note that the names cited in the Carian part coin-cide with several names that appear in the Greek inscriptions onthe same stone; particularly striking is the fact that in one of thepriesthood sales, the daimones of an Arrisis son of Imbrasis, and aHermias son of Arrisis, are quoted among the other divinities mentioned.

32 Schürr (1992:146) had attempted to identify a name corresponding to Greek

Eemiaw in the Greek list of priests by isolating armi t , but the value y /y/ of W, nowconfirmed by the Kaunos bilingual, makes the connection between ylarmi t and Hyllarimaclearly preferable.

33 Cf . perhaps Lyc. mle-, ‘sacrificial off ering (??)’ and related words (particularly mla-traza, mluhidaza , two priestly titles) in Melchert DLL, ss. vv.

308

Both Arrisis and Imbrasis are names that appear in the first Carianonomastic formula (b rsi ari “≤ b rsi ≤ ), and it is reasonable to assume that

we are dealing with the same individuals.34

5. Other Inscriptions from Caria

The new inscription from Mylasa (C.My 1) consists of an initial lineas the heading for a list of personal names. Therefore, only this firstline contains what seems to be common lexicon. As is the case of mostCarian texts without personal names, it remains practically impenetra-ble:

idrayridsemd ?bq mol “ ty∞[

The most interesting characteristic is the presence of the word mol “ ,also found in Hyllarima, which has been analysed as accusative or nom-inative plural with the meaning ‘priests’ (see above). The presence of a noun in plural fits well with the following list of personal names. InAdiego (2005:92–93) it was also suggested that ty∞[ could be the Carianadaptation of Greek TÊxh, the goddess of Fortune. Therefore mol “ ty∞[

could mean ‘priests of Fortune’, ‘priests for Fortune’, although this ismerely hypothetical. The preceding sequence idrayridsemd bq is very difficultto analyze. If the reading of d is certain, perhaps a word or particlechain, md (cf . md orkn ), could be segmented. In idrayriº, we can attemptto identify two stems idra- and yri- that recall, respectively, the placename Idriaw and the yri-formant in personal names such as “ ayriq, idyri ∞≤ ,Saurigow. Cf . also (without tectal suffixation) E.Th 26 yri ≤ and possiblyE.AS 5 pnyri ≤ ru.

Little can be said about the rest of Carian inscriptions from Caria.

The Euromos inscription C.Eu 2 does not seem to be a funeraryinscription, insofar as none of the typical words for ‘tomb, burial’ canbe identified, as already noted. Neither can we identify any clear ono-mastic formulae. It is possible that the first sequence omob∞i contains

34 See Adiego-Debord-Varinlio< lu (2005:614) for a rather speculative attempt to con-

nect the formula b rsi ari “≤ b rsi ≤ with Arrisis, Imbrasis and Hermias, by assuming thatthe two Arrissis are the same person (which implies a descendence Hermias, son of Arrisis, son of Imbrasis), and that the Greek individual Hermias would the same asthe Carian one b rsi , a supposition based in the Carian name of Hermes Imbrasow/Imbramow.



309

the relative pronoun (hence the segmentation adopted here). As for thefollowing words, perhaps a genitive temazi ≤ can be segmented. In this

case, the d that immediately follows it could be compared with d ini [ --- ]inis =d =rual from C.Ka 5. Another word that may be present inthe text is armon, which cannot be clearly interpreted: it could be directlycompared to armon in E.Me 8 (‘dragoman, interpreter’, although thismeaning would seem somewhat out of place here) or be analyzed asan accusative in -n of the moon-god armo (attested in C.Hy 1a). Therest of the inscription is even more difficult to fathom.

In the shorter text of Sinuri (C.Si 1) at least a possible onomasticformula can be identified:

(1) adymd“ : yri∞ñ : t[-]rsi : [. . .?]tbe≤(vacat)(2) yri∞ñ : binq : sñaidl o

t[ - ]rsi : [. . .?]tbe≤ is convincing as a N-Ø N-≤ structure, and the incom-plete second name can be compared with Thebes qutbe , Kuatbhw.

The rest of the inscription remains unclear, a situation that is fur-ther compounded by the reading difficulties. In Adiego (2000), sñaid l owas tentatively analysed as a verb, comparable to pisñoi=md a, ñmai l o=md a . ºaidº was directly linked to Lycian a(i)-, ‘to make’, and sñº wascompared to Lycian se=ññ(e). Leaving aside these more provisionalattempts, one must focus on the striking disposition of the two texts; arelatively large gap has been left between the two, and the fact thatin the first text a word adymd “ that possibly contains a plural ending (-“ , cf . mol “ ) is followed by a single onomastic formula, makes it plau-sible that the formula adymd “ : yri ∞ñ was conceived as the heading of a

list of personal names, and that this list was never completed. Theinscription from Hyllarima shows how diff erent names were added tothe list on subsequent occasions, and this could also have been alsothe initial aim of C.Si 1.

The inscription of Kildara (C.Ki 1), shown below, is even harder toanalyse than the preceding texts:

[ . . . . . . . ( . ) ] zol ba∞a[..(.)] kil [ [. . .]uda[. . .] trqdimr qrds tazomd[ kil arad[-]ybzsdmHnmkda[-]aHuq[ iasoum

The only elements to have been isolated as recognizable words are: (1)the two instances of the place name Kildara (l.1: ki l [, l.3; ki l ara or



310

ki l ara d ? -d could be a case ending or a clitic form d , comparable to-d - in i[ --- ]inis =d =rual ); (2) the god name trq d (l.2, or better trq d imr as

a compound name?); (3) the word qrds , also in l.2, undoubtedly relatedto qrds, qrdsol in C.Ka 2, and for which an ‘institutional’ meaning hasbeen proposed. The only clear evidence for the content of the Kildarainscription is negative : there is no point of contact with the Greek textinscribed immediately after it. The name of the individual honouredas proxenos and eueregetes in the Greek inscription, Uss[ vllow?] son of Samvow, does not appear,35 and there is no sequence that can be com-pared with the phraseology used in the proxeny decree of Kaunos(C.Ka 5).

As for C.Ka 4, the fragmentary inscription from Kaunos that islargely incomplete and has no clear context, it is impossible to evenattempt an analysis:

[. . .]u≤ou≤ ibrsdr[-][. . .]a yoml nr_1i[. . .]dar1_ idym“

Only two elements have been recognized here: the form i b rsº, without

any doubt related to the family of names i b rs(i)-,

b rsi , and yom

l n, avariant of the word uioml n in the first line of the Kaunos bilingual,

interpreted as a verb or noun pertaining to the semantic field of ‘decree,decision, to decree, to decide’.

The two Stratonikeia inscriptions (C.St 1, C.St 2), as far as we cancurrently tell, consist exclusively of onomastic formulae.

E. B S C Q

The sum of the preceding analysis of a great part of the Carian cor-pus can seem at first sight somewhat disappointing: the results that canessentially be considered conclusive are very scarce, and even some of these are perhaps overly optimistic, since speculation has been unavoid-able in several cases. Strictly speaking, our ability to analyze Carianinscriptions is limited to the identification of proper names and ono-mastic formulae, which allows us to understand only a very small area

35 However, one could be tempted to read ßAmous “ ! amous instead of ÎAsouMiasoum in the last line, in order to obtain the name of the father of Uss[ollos?], Samoos,mentioned in the Greek inscription, see p. 142.



311

of Carian grammar. Only in the case of the Kaunos bilingual canthe interpretation be developed to any extent, thanks to existence of

the Greek version, but even with this advantage we do not obtain theresults we would hope for. As for the rest of the texts, the inscriptionof “ rquq qtblem≤ (C.xx 1) is something of an exception, since the bril-liance of Melchert’s interpretation has not been matched for otherinscriptions.

These limited results will be systematized in the next chapter, whereI shall focus on two problems that have arisen in the course of thischapter, namely the s -ending and the alleged verbal forms md ane andvariants.



CHAPTER EIGHT

MORPHOLOGICAL FEATURES

In this chapter I summarize the morphological traits that can be obtainedfrom an analysis of the inscriptions, and interpret them from a histor-ical and comparative perspective. Needless to say, the resultant picturewill remain exasperatingly incomplete, and in many cases will be based

on extremely fragile evidence. It will however be sufficient to demon-strate the clear relationship between Carian and Indo-European Anatolianlanguages and, more specifically, Luwic dialects.

A. N I

Our present knowledge of nominal inflection is based mainly on the

information given by the personal names, to which we can also addsome common nouns. There are no examples of adjective inflection,with the exception of ethnic forms, which can be interpreted as adjec-tival complements (for instance otonosn in C.Ka 5). Only a few endingscan be established with some confidence.

1. Nominative Singular

The singular nominative ending is systematically -Ø, as appears clearly

in the onomastic formulae (u≤ ol , arli “ , ada , etc.). This also seems to bethe ending for common nouns, for instance upe, ue ‘stele’.

This zero ending can be directly compared with Lycian and Milyansg. nom. c. -Ø, and we can imagine it to have a similar origin: a PIE*-s > Proto-Anatolian (PA) *-s (Hitt., CLuw, HLuw, Pal., Lyd. -/s/)dropped in absolute final position. Words such as the abovementionedupe, ue seem to be vocalic stems of the common gender, also with lossof final -s .1

1 If they are not old eh2- stems with inherited ending -ø for the sg. nom.

313

2. Accusative Singular

An ending -n for the singular accusative has been clearly establishedthanks to the Kaunos bilingual, confirming former interpretations whichled to the same conclusion.2 The personal names nik [ — ]la-n, lys [ ikl ]a-nshow this ending for a -stems of personal names (Frei-Marek 1997:34,48). Another clear example of -n accusative is orkn (‘bowl’ or similar).The ending -n for singular accusative also appears in the pronominalflexion (see below p. 320).

This ending is also comparatively transparent: PIE *-m > PA *-n.Unlike in Lycian, where the ending in contact with a stem final vowel

a, e results in a nasalized vowel (noted by means of special letters <ã>,<e>), the Carian forms with -an could point to a conservation of thenasal, although it is not impossible that a process similar to the Lycianexample has occurred, though not noted with the same graphical pre-cision. In ork-n we find the same ending, but in this case after a con-sonant, which could represent a syllabic nasal or merely a defectivevowel notation (/orkVn/).

3. Genitive Singular The most characteristic ending for the genitive singular, confirmed onlyby personal names, is -≤ . Melchert (2002:311) has argued convincinglythat its origin lies in PA *-ass ì -, a possessive suffix with i-mutation thatserves to create adjectival forms in Luwic dialects (CLuw. -ass ì -, HLuw.-asi-, Lycian -ahi, -ehi , Milyan -asi, -esi ) that carry out the proper func-tion of a genitive (cf . Lyc. ˜ eni mahanahi ‘mother divine’ = ‘mother of the gods’). This etymological connection explains the presence of -≤ ,

most probably a palatal sibilant /ç/ resulting from the contact of s with ì , and either dropped or not graphically noted in Carian.3

It is less clear if -≤ in Carian continues to act as an adjectival suffix,agreeing in number and case with the word that complements it, or if

2 See Melchert (1993:80).3 Melchert (personal communication) informs me that his views about the origin of

Carian -≤ have changed since the publication of Melchert (2002). He now proposes

that -≤ is cognate with Hierohplyphic Luwian -(a)si , and that both forms come fromthe true PIE genitival ending, *-osyo. This interesting alternative hypothesis merits fur-ther discussion that will not be included here. In any case, it does not alter the phono-logical explanation, formulated above, of -≤ as a palatal sibilant resulting from thecontact of s with i .



314

in fact it has been reinterpreted as a pure case ending. A single form, pñmnn≤ ñ in C.Si 2, suggests that -≤ retains its adjectival value, as it shows

a further accusative ending -ñ . However, the example is too isolatedto be considered proof that the adjectival character of -≤ has been pre-served: it could merely be the result of a secondary process of agreement,as occurred in Lycian with -hñ . In any case, whatever the explanationof pñmnn≤ ñ , it is a good illustration of that fact that in Carian, like inLycian, the limits between genitive and possessive adjective constructionswere rather blurred.4

4. The -s Ending and the Problem of Carian Datives Throughout the chapter devoted to the analysis of Carian inscriptions,we have repeatedly mentioned the existence of a specific -s ending,which has been interpreted as a dative ending by some authors (Schürr,Melchert, Vittmann). Our analysis has led us to the provisional con-clusion that the solution is perhaps not so clear. We shall now attemptto analyse this ending more carefully and evaluate the plausibility of interpreting it as dative.

Any analysis of the possible examples of s-endings is hampered fromthe beginning by the difficulty of distinguishing between true s-ending forms and simple Ø-nominatives of s-stems (whose existence is undeni-able, given the presence of s-stems inflected in genitive: i b rs-≤ , idyes-≤ , pals-≤ , etc.). To quote a case already mentioned (see p. 283), a wordsuch as smd ÿbrs (C.Ha 1) could for example be analysed as an s-ending form of a stem smd ÿbr-, comparable to k “ at ÿbr, dt ÿbr ardybyr-≤ , or as anominative of an s-stem smd ÿbrs-, parallel to ybrs-≤ . The examples of formsending in -s from the sub-corpus of Memphis also help to illustrate

this. Listed below are all the Memphis inscriptions that present thesetypes of words. I exclude only E.Me 39, where the reading of -s is notcertain. The inscription is also very fragmentary and hardly usable:

ap[—]ws / a[rb]ikarm≤ ∞i (E.Me 23)tdu≤ol / kbos | “amsqi[. . .? (E.Me 24)(a) idmns | myre≤ ∞i | mdayn ∞i / (b) idmns | myre≤ ∞i (E.Me 33)ntokris | dw≤ol ≤ | mwdon≤ ∞i (E.Me 35)[q? ]lal is / [?]iam≤ ∞i / alos ∞arnos (E.Me 45)tqtes | paraibrel ≤ ∞i | mn[o-?] (E.Me 47)

4 More problematic is the intepretation of the form pd a ∞m≤ uñ , see the discussion inMelchert (2001:311).



315

From all of these examples, the most likely to have the s-ending formis undoubtedly ntokris ; Vittmann (2001:52) has argued convincingly that

this name was taken directly from Egyptian, ruling out Ray’s theory thatit can be identified as a form that arrived in Carian from the GreekNitvkriw, the only way to explain the final s as a part of the stem (byassuming it to be an s-stem arising from the Greek sigmatic nominative).Also suggestive of an s -ending, if the reconstruction of the initial letteris accepted, is [ q ?]la l is , given the existence of a name qla l i-≤ (G 2),Greek Kolaldiw, Kulaldiw.

The rest of the examples are ambiguous: ap[ — ]ws, idmns and tktes allow more than one interpretation.5 Finally, kbos is very unlikely to bea personal name in nominative of an s-stem, because it occurs pre-cisely after the nominative tdu≤ ol , i.e. the name of the deceased. However,it could also be a title or an adjective accompanying tdu≤ ol (an ethnicname? see p. 278), so that it would also be in nominative. A similarexplanation could be given for the enigmatic alos ∞arnos of E.Me 45.

For convenience, I will assume that all these examples can be inter-preted as s-ending forms. In fact, this will not aff ect the ideas to beformulated about the value of the s-ending to any great degree, since

with the exception of kbos and alos ∞arnos , the context in which we findthe rest of the examples is similar to that of the clearest ones (ntokris,[ q ? ]la l is ): as the first word of an onomastic formula—which implies thatthey represent the name of the deceased—followed by one or morepersonal names in genitive.

By analyzing the possibility that the s-ending recognizable in somewords could be a mark of a dative case, we can first of all clearly sep-arate the formal aspects from the functional ones. Formally, we canagree with the hypothesis formulated by Melchert (2002:309) that this

s-ending comes from a genitive ending PIE *-e/oso (like Lycian -a/-ehe ,cf . Adiego 1994c). An alternative view—a Luwic possessive suffix with-out i-mutation *-asso — cannot be ruled out, although Melchert (ibid .)is right in pointing out that there is no evidence to support the pro-posal of an adjectival character for -s in Carian. In any case, bothsolutions are equally attractive from a phonological point of view, asboth off er a straightforward explanation for the presence of a diff erent

5 In the case of idmns , the name idmuon-≤ could favour an analysis as idmn- if bothforms belonged to the same paradigm, but the diff erences between idmn- and idmuon-do not allow us to confim this hypothesis.

316

sibilant sound in this -s-ending (< *-e/oso or *-asso- ) vs. the Carian gen-itives in -≤ (< *-ass ì , see above).

If we accept the above arguments, the problem for Carian datives in-s can be reduced to a purely functional one, the question being whetherthere is evidence for a functional displacement of these old genitivesto the expression of a dative value. As Melchert points out, such a dis-placement would not be particularly problematic because it is well doc-umented in other languages. However, in my opinion, the evidencecurrently at our disposal is not at all convincing.

The most radical version of this hypothesis—that -s has become exclu-sively the mark for dative, contrasting with -≤ specialized as genitive/pos-sessive—is disproved, as Melchert admits, by the formula i [ — ]inis=d =rual ,where -s cannot be anything other than a genitive or possessive gov-erned by rual (= Gr. §p‹ dhmio[u]rgoË ÑIpposy°nouw ). Moreover, the ono-mastic formulae in the same inscription nik [—]lan lysiklas [-? ], lys [ ikl ]anlysikratas [-? ] suggest that -s should be interpreted as a genitive orpossessive.

If one accepts the existence of genitives/possessives in -s-, the claimof an s-dative must be substantiated by unequivocal evidence, and this

is hard to find, since all the examples of alleged datives can also beinterpreted as genitives/possessives:a) All the examples of possible datives in -s in the Memphis sub-

corpus can also easily be interpreted as genitives or possessives: if oneobserves the texts quoted above, a translation ‘for X’ or ‘of X’ soundsequally acceptable.

b) The same analysis can also be applied to other examples: an inter-pretation of “ arnajs sb taqbos as ‘of ”arnaj and of Taqbo’ is as good asa dative interpretation ‘to ”arnaj and to Taqbo’. Vittmann (2001:52–54)

has made the case for a dative value based on the fact that the Carianinscription appears integrated in an Egyptian formula, ‘give life’, whichseems to require a dative. He cites the parallel examples of certainEgyptian-Phoenician bilingual inscriptions where the same Egyptian for-mula ‘give life’ is used, and where the personal name is introduced inPhoenician by the preposition l , ‘to, for’. However, while the Phoenicianparallel may seem to provide a strong argument for seeing the Cariansequence as syntactically integrated in the Egyptian sentence, the oppo-

site may also be true, namely that the Carian and Egyptian parts aremore loosely related: the Egyptian sentence might simply be a stereotypedformula, and the Carian phrase a mere indication of possession. Eventhe Phoenician example accommodates a similar loose relationshipbetween both parts, insofar as l is usually employed in Phoenician for

317

introducing the target of a dedication. Regarding armotrq d osq in Hyllarima,the context does not allow us to ascertain whether the word is a gen-

itive or a dative, but the character of the inscription, which presentsa list of priests, seems more consistent with an interpretation as geni-tive (‘priests of the god Armotrq d -’) than as dative (although a reading ‘priests for the god Armotrq d -’ or a generic reference to the establish-ment of a cult ‘to Armotrq d -’ cannot be ruled out).

c) The -s-dative hypothesis adds an unnecessary complication to theinterpretation of ntro/ntros in their respective inscriptions (see abovepp. 285–286). The dativum dedicationis with the god name ntro, ‘Apollo’,a simple and elegant interpretation that fits well with the overall senseof the text, becomes a nominative with the meaning ‘priest of Apollo’and, conversely, ntros pr ãidas must then be an anonymous priest of Apollo to whom the object is dedicated instead of a direct referenceto the divinity. But even accepting this explanation, there is nothing to prevent us interpreting ntros pr ãidas as a possessive, insofar as theoverall structure of the inscription remains unclear.

d) Attributing a dative role to the forms in -s means that the formtrqud e from Iasos remains unexplained. It is true, as Melchert remarks,

that the overall sense of the inscription is not clear, but if trqud e issimply the name of a god—as armotrq d osq in the new inscription fromHyllarima seems to confirm—it is difficult both formally and semanticallyto assign it a value other than dative. It cannot be an accusative (aswe would expect a final -n ), a genitive (there is neither -≤ nor ‘possessive’-s) or a nominative (the subject of the entire text seems to be a personalname, perhaps ]are “ , on which the patronym in the genitive siyklo-≤ isdependent).

To summarize, we can draw two conclusions: (1) there is evidencefor the retention of the original genitive/possessive value of -s , at leastin some clear examples, and (2) there is no conclusive evidence toaffirm that -s has adopted a dative value.

5. Other Possible Datives

Note that my opinion, which does little to support the existence ofs-datives, is accompanied by an attempt to recognize other words as

datives. In this book at least three possible candidates have been men-tioned: in the previous pages I have referred to trqud e and ntro. We canadd also rtmi , but it depends on the analysis of the inscription C.Tr 2,which remains very controversial. From an etymological point of view,ntro could be explained as a dative in -o < *-à comparable with the

318

(admittedly scarce) examples of datives in -a for a-stems in Lycian (seeMelchert DLL:x; note for instance Urebillaha ). As for trqud e , we can

envisage a PIE dative singular ending *-ei , as can be expected in annt-stem such as trqud - < *t º Hnt-. Finally, rtmi could be a dative in -i of an i-stem (*artimi-?, cf . rtim nominative in C.Hy 1a).

6. Nominative Plural

No clear examples of nominative plural can be cited at present. Wecan only mention two possible forms:

• in C.Hy 1, and also in C.My 1, it has been proposed that mol “ couldbe a common plural nominative with the meaning ‘priests’, althoughother explanations are possible. If this hypothesis were true, a suit-able interpretation of mol “ could be mol-“ , with an “ -ending formallyidentical to that of the common plural accusative -“ (see below), atrait that Carian would share with Milyan (Nom. pl. -z / Ac. pl.-z ) against Lycian (-Ø / -s ). However, -“ could in this case be aderivative suffix, so that mol “ ought to be interpreted as mol “ -Ø , lead-ing to the converse situation: a nom. pl. in -Ø nearer to Lycianthan to Milyan.

• in the funerary inscriptions of Kaunos and Krya, the possibility wasenvisaged that s d is could be a nominative plural corresponding tosi d i (see p. 291). However, this interpretation is difficult to reconcilewith either of the two above hypotheses for mol “ . We could attributethe discrepancies to a diff erentiation between a Kaunian dialect anda purely Carian one, but at this point it becomes a merely ad hoc solution.

7. Accusative Plural

The Kaunos bilingual off ers good evidence of -“ as a plural accusativeending for the common gender: sarni “ , (un )d o[ — ]t l “ , kbdyn“ , b 2o[ — ]ol “ make reference to the two Athenian citizens honoured in the decree,who are mentioned in accusative. The simplest etymological explanationof -“ as an accusative plural mark can be found in Lycian -s and Milyan-z , both from Luwic*-ns (< PIE *-ns ). The appearance of the (presumably)

palatoalveolar sound “ instead of the generic, unmarked dental sibilants recalls the Milyan use of z instead of s , irrespective of the exact natureof z in Milyan. In both languages, the use of a sibilant other than s can be interpreted as the outcome of the original cluster *-ns .



319

8. Other Possible Case Endings

1) In Adiego (2002) it was argued that some words ending in -t wereplural genitives. The basis of this hypothesis is the interpretation of mol “ msot ylarmi t as ‘priests of Hyllarimean gods’, roughly corresponding tothe Greek text ‘priests of all the gods’ (see pp. 306–307). The hypoth-esis was developed from the assumption that the Carian letter O wasthe Kaunian variant for the sign t, and consequently some words end-ing in O = -t were also interpreted as genitives.

If this hypothesis is accurate, we still do not know what the originof this genitive plural ending might be. Given that -t probably repre-

sents a dental-palatal aff ricate /t“/ or similar, a direct derivation fromPIE -om (> Old Hittite -an, Lycian -e ) is impossible, which opens theway for analogical process. In this way, the palatal character of -t couldbe somehow related to the palatal ending -≤ of the genitive singular.Assuming a plural nominative *-ns (see above), an analogical propor-tion Nom. sg. *-s (>-Ø)—Genitive sg. *-si-s (>-≤ ): Nom. pl. *-ns (-“ ?)→ Gen. pl. *-ns-is (>- ) could be proposed (with a phonological process*-nsis > *-nt ≤ > -t ).

2) The contrast between alos ∞arnos (E.Me 45) and alos-d ∞arnos-d (C.xx 2) leads me to think that -d could be a true case ending in thelatter example. The (very hypothetical) interpretation of C.xx 2 out-lined in Adiego (2000:153–155; see here p. 284) pointed implicitly toan analysis of alos-d ∞arnos-d as ablative (‘from Halikarnassos’). The end-ing could then be compared to Lycian -adi/edi CLuw. -ati , HLuw.-ari-/ati , but this comparison is complicated by the fact that in mostcases of the origin or Greek transcription of Carian d that can be con-trolled, this letter reflects an original -nd- group, which would not be

the case of -d from *-Vdi (see p. 247).Other analyses are also possible: one could connect -d with the likelypreposition d (cf . Lycian ñte ) in C.Ka 5 d =rual (see p. 254). However,one ought to assume then that d became a sort of case ending, theonly possible explanation for the iteration alos-d ∞arnos-d .

B. P I

We can identify in Carian at least two demonstrative pronominal stems:s(a)- and a-.

The only clear indication of a pronominal paradigm can be stated forthe pronoun sa , on the basis of sa E.Me 27, san (G 1), snn (C.Ha 1,

320

C.xx 1): sa and san seem to be singular nominatives according to thecontext in which they appear, whereas snn is undoubtedly a singular

accusative. The difficulty is to establish the relation between the twonominatives sa and san, for which there are two possibilities: the diff erencebetween both forms could consist of a gender distinction (sa commongender vs. san neuter gender, reflecting -Ø vs. -n respectively), or alter-natively that san is a kind of extended form (this was the explanationgiven in Melchert 1993:79–80) independent of the possible gender dis-tinction. This latter hypothesis is supported by the accusative form snnthat seems to imply a s(a)n- stem to which the accusative ending -nhas been attached. However, it is also possible that in snn, both n actu-ally constitute an accusative mark,6 thus the first interpretation can bemaintained. In this sense, there is a striking parallel between Lycianinflection and the pronoun ebe-:

Nom sg. c. ebe-Ø cf . Carian sa-Ø < *-s Ac sg c. ebeñne cf . Carian snn “long” forms of original *-nNom-Ac sg. nt. ebe cf . Carian san < *-n

However, perhaps this parallelism is merely an illusion: in the case of the pronoun a-, we find ann in C.Ka 3, where it can hardly be a com-mon accusative parallel to snn, ebeñne , since it appears in a context with-out a verb. It might instead be interpreted as a (neuter?) nominative.The interpretation of an in an si d i is not clear either, because it is notcertain whether in this inscription it represents an accusative or a nom-inative. In the latter case, interpreting it as neuter seems inconsistentwith the proposed analysis of s d is as as common plurals (cf. supra). Insummary, the current inventory of pronominal forms is too incompleteand their interpretations too ambiguous for even a minimally coherent

paradigm to be traced.Besides these demonstrative pronouns, we can also refer to the (atleast originally) relative pronoun ∞i , whose etymology is unproblematic:it comes from PIE, PA *k w is (cf . Hitt, CLuw. kui “ , Lyc. ti , Mil. ki ). Asseen in Chapter 7, this pronoun became in Carian a type of connect-ing particle, in a similar way to the Old Persian haya/taya , althoughsome examples of its use can be still identified as remaining very closeto its original relative value.

6 Cf. Neumann (apud Melchert 1993:80, n. 5), who compared snn with Lycian eb eñn e .

321

C. V I

The difficulties of identifying verbal forms were outlined in our analysisof the Carian inscriptions. The only word I consider to be even a min-imally reliable verbal form is ÿbt , for which Melchert suggested a veryplausible interpretation (= Lyc. ubete ‘he off ered’). The search for otherpossible verbal forms can take one of two directions, which althoughnot totally incompatible, are in some cases difficult to reconcile. On theone hand, recognizing an ending -t , comparable to Lyc. -te (or perhapsalso -ti if the verb was a present, not a preterite), opens the search forother forms also ending in -t that can be compared with these singularendings or with other endings that in Lycian are also characterized bya -t (for example, the third plural present and preterit endings -ti/-te after diphthongs, such as ºai-ti, ºei-ti, ºai-te / ºei-te ). In such cases, theresemblance of the sound context (t after a vowel) regarding the sin-gular ending could make it likely that the outcome in Carian is also t (although this is an optimistic supposition, and disregards possible ana-logical alterations). The results of such a search are however very limited:in our analysis of Carian inscriptions (Chapter 7), only two forms have

appeared that, based on their context, can be interpreted as verbs withthese type of endings: (1) not in C.xx 2 (connected there to Hitt.nà(i)-, CLuw. nana- < * PIE *neyH-, so that a meaning ‘he brings/hebrought’—3rd sg. present or preterite—could be suggested); (2) ait inC.Ka 2, l. 8, where it appears in a sequence qrdsol “ ait: ait has beententatively connected to Lycian ait e , a 3rd pl. preterite form of the verba- ‘to made’ (see p. 301). To these two forms one can also provision-ally add the word aitusi in C.Ka 5, if segmented ait + usi , althoughproblems arise when trying to establish the meaning of usi . An alter-

native view—to segment aitu + si— is an attractive solution as it yieldsa possible imperative form (cf . -tu in Lycian), but it fails to analyse theword si .

Other possible forms ending in -t could allow a similar interpretation.A good example is 1aitk , also in C.Ka 2, if segmented 1ait =k ), but theverbal stem cannot be established as we are unsure of the phonologi-cal value of the initial letter 1.

On the other hand, scholars have searched for verbal forms by means

of an internal and combinatory analysis of the Carian texts. In this sense,there has been a certain consensus in regarding a sequence repeatedin diff erent inscriptions as a possible verbal form: the sequence md ane (Thebes mlane ). The fact that this interpretation is based on internal



322

and combinatory grounds explains why it preceded the definitive deci-pherment of Carian: it was ”evoro“kin who suggested that md ane, mlane

was a verb (see for instance ”evoro“kin 1977:124). Following the deci-pherment, authors such as Hajnal, Melchert and Schürr have followedthis line of argument, and in the case of these possible verbal formsthey have also added some etymological connections to the functionalbasis of the hypothesis. The most recent and comprehensive treatmentof the subject is an unpublished paper by Melchert (Melchert, md ane ).7

The interpretation of md ane, mlane as a verb is supported above allby evidence from three brief inscriptions found on objects, where md ane appears to be the one verb of a sentence. By contrast, most of theexamples of mlane , from Thebes, are contextually far less clear. All theexamples are given below:

md ane:“arkbiom : zidks md ane : ÿn-?-mo | den : tumn (E.Sa 1)ntros : prãidas / or“a / nu md ane : uksi wrm≤ (E.xx 7)

ÿ≤ biks not : alosd ∞arnosd : jzpe md ane (C.xx 2)

mlane:a?q≤baq ewm ≤emot / qtblo owdown[. . .]mwarudk≤o mlane (E.Th 10)

?-˚bjqmq ew mlane qeb≤t | u[. . .]ü≤q | qwsal | mqabaewleqo“oski.oms(E.Th 12)l ÿ∞se | “i“≤ | mlan[ -?] (E.Th 35)dquq | ewmlane | tebot | g kem≤ (E.Th 44)w. dbo≤kn ewá¬å»e ˚[ (E.Th 47)

(cf . also lane in: balewlane | “rbk˚[ - ]sal | (E.Th 49).

Of the three examples of md ane , the last is of particular importance,since there is a possible conflict with the word not , which I have proposedto be the verb of the sentence.8 In the other two instances, it must be

recognized that there is a reasonable chance of md ane being a verb,although there is no reason not to look for a verb in some of the otherwords that appear in both inscriptions, and that remain uninterpreted.If md ane is a verb, from the inscriptions in which it appears we couldtake it to mean, ‘to vow, to off er’ or similar (cf . Melchert, md ane ).

As stated above, Theban graffiti with the variant form mlane aremuch more obscure. Only E.Th 44, and perhaps also E.Th 35—but

7

I thank H. C. Melchert for allowing me to use this paper. Although I disagreewith Melchert’s views on md ane and related forms, I consider his usage and analysisof the forms superior to Hajnal’s (see Hajnal 1997b:151–157), who arbitrarily sepa-rates mlane from md ane and does not include md a forms in his study.

8 Indeed, there is no reason why the inscription should not contain two diff erentverbs, not and md ane

323

note the lack of a final letter—off er apparently complete and brief textsin which mlane could be a verb. In the remaining examples, it appears

after a sequence ew , very close to the form ewm, which is also typicalin the Thebes corpus. This creates the possibility of a diff erent seg-mentation of ewmlane : ewm + lane , instead of ew + mlane . The exampleE.Th 49 introduces even more confusion: the sequence ewlane , analyzedas ew + lane , supports both ew and lane as isolable words in ewmlane . A possible compromise, and perhaps the correct solution, would be toisolate three elements: ew + m + lane . This would allow diff erent com-binations: ew + m, ew + lane, m + lane and ew + m + lane.9

The difficulties of analysing md ane as a verb begin when other, lessfavourable factors become involved: besides md ane , the sequence md a appears repeatedly in C.Si 2:

. . . eri : pisñoi md a : pñmnn≤ñ : pda/∞m≤uñ ∞i “aoyr∞ri mt∞elã / ñmail omd a lrHñ : stspñ vacat / sm“s[-5-] sb añmsñsi md a (C.Si 2a)10

Moreover, in C.Ha 1, a sequence md can be easily segmented:11

smd ÿbrs | psnl o | md orkn t ÿn | snn

As I have argued (see Adiego 1994a:54–55, 2000:140), this md a/md

seems almost inseparable from md ane , a theory also expounded byMelchert in his unpublished paper. In fact, he interprets md - to be thelexical root of all these alleged verbal forms. Melchert even extendsthe presence of this root to the form ml n in C.Ka 5, unifying md , md a,md ane, ml n under a common root *mVld- (by assuming a *-ld- > -nd-process in the forms with d ), which he compares with Hitt. mald-12

‘vow, solemnly pronounce’, but also ‘dedicate, give’. As for the diff erentendings of the md -family of words, Melchert (md ane ) off ers the following analysis: ml n as a preterite third plural from *mVld-onto;13 md a as a pre-sent third singular (used as a preterite) with an ending parallel to Hittite-ài in mald ài ; -ne would be an enclitic object pronoun (following Schürr1996a:66).

9 Given the difficulties regarding the segmentaton of the (possible) elements thatmake up ewmlane , I enter this in the Glossary (Chapter 11) as a complete word.

10 Cf . also C.Hy 1a dymd a , but in this latter case I am not sure of the segmenta-tion dy md a .

11

ºmd º also appears as a sequence in other inscriptions (C.My 1 (?), C.Si 2, C.Ki1: note also in the first word of C.Ha 1), but the contexts are too obscure (C.My 1,C.Ki 1) or simply too unlikely (C.Si 2, C.Ha 1) to allow us to isolate md as a word.

12 As Melchert recalls, the connection ml n = Hitt. mald- was already suggested inHajnal (1997:152).

13 Without discounting the possible interpretation as a present (< *mVld -enti ).



324

In Adiego (2000:140), I argued against this interpretation of md /md a/md ane/mlane as a verb, and I believe that the objections for-

mulated in that study are still valid: it would be very unusual not onlythat the same verb was used in very diff erent classes of texts (dedica-tory inscriptions on objects, visitors’ graffiti, a public decree, the heading of a list of priests . . .) but also that the same verb was used at least three times in the same inscription (the decree from Sinuri, C.Si 1). Thus if we were to accept this interpretation we would be obliged to acknowl-edge both an excessively polysemic or general meaning for md -, andan unrealistic lack of lexical variety in Carian. For this reason, I chooseto recognise in md /md a/md ane/mlane a functional word or a chain of functional words. If one wants to retain the interpretation of md - as averb, the only remaining possibility would be to see it as an auxiliary,lacking any lexical meaning (cf . Adiego, ibid .). I am however moreinclined to categorise md /md a/md ane as a chain of particles; in Adiego(1994:54–55) I already introduced the idea of analyzing md ane as achain of introductory particles followed by two enclitic pronouns, -n-(accusative) and -e- (dative), although I now believe that an analysis of -ne as a unique enclitic pronoun, as suggested by Schürr and Melchert,

is perhaps more plausible. In any case, md would remain as particlechain m-d — documented tel quel in C.Ha 1—where one could tentativelyimagine a connection with the Lycian introducing particle me for initialm-. It is true, nevertheless, that this comparison is a mere desideratum,since the behaviour of Lycian me vs. Carian m- would be totally diff erentin terms of the order of constituents: whereas Lycian me always appearsin the first or second position in a clause,14 Carian m- can seeminglybe placed anywhere.15

Besides md ane (and variants), other forms have been considered pos-

sible verbs on the basis of the context in which they appear. Perhapsthe most likely of these to in fact be a verb is uioml n, documented inthe Kaunos bilingual inscription (C.Ka 5), and possibly also in C.Ka2 [ ui ?]oml ã and C.Ka yoml n, if these forms are interpreted as purevariants of uioml n. It is possible that this word is the Carian form cor-responding to the verb ¶doje in the Greek part of the bilingual, althoughthis does not necessarily mean that the Carian version had a literal

14 See Melchert (2004), s. v. me.15 The interpretation of md ane as a chain m+d a+ne , where m stands as a particle, is

consistent with the tentative analysis of the Theban form ewmlane as ew+m+lane (analy-sis based on the independent existence of ewlane and ewm, see above p. 323).

325

correspondence to the Greek one (indeed, other possible analyses of uioml n have been envisaged, see Chapter 11 s. v. ). Of the diff erent

attempts to explain uioml n, perhaps the most convincing is the pro-posal made by Hajnal and Melchert to isolate a verb, ml n, a preteritethird plural of a stem, ml - < *mVld-, so that ml n could come from*mVld-onto (cf . Melchert, md ane ).

One of the advantages of this analysis is that it can be applied toanother of the words to have been identified as a possible verb (Adiego2000:142, see here p. 304): pisñ (in the sequence pisñoi md a ) in C.Si 1a.In Adiego (2000), pisñ was also interpreted as a preterite third pluralwith the meaning ‘they gave’, formed from verbal root pi-, equivalentto Hitt., Luw. piya-, Lyc. pije- ‘to give’, followed by an iterative suffix-s- and a personal ending -ñ . The form would therefore be almost par-allel to the Lycian, tusñti , a present third plural of an iterative stem tu-s-, created from tuwe-. Both the function and the meaning proposedfor pisñ are consistent with the context: the form appears following theonomastic formula of the Hekatomnids Idrieus and Ada at the beginning of a decree, so a suitable translation might be ‘Idrieus (son) of Hekatomnosand Ada (daughter) of Hekatomnos gave . . .”.16 This would mean that

pisñ could contain a preterite third plural ending -ñ from *-onto, paral-lel to -n in ml n.Other possible verbal forms (ñmai l in C.Si 1a, cf . Adiego 2000:142;

nu in E.xx 7) have been suggested on very hypothetical grounds, andthey will not be discussed here.

To summarise, the results of the search for verbal forms in Carianare currently very uncertain. We can establish, with the utmost cau-tion, only three possible endings: -t as a preterite (or present) third sin-gular ending, and -t and -n/ñ as preterite (or present) third plural

endings (the first after the diphthong -ai , the second in other contexts,similar to those that in Lycian become -ñte/-ñti ).

16 In Carian: [—]ryin ∞tmño≤ : sb ada ∞tmño≤ eri : pisñoi md a . . . As for eri , it could bea preverb or a noun (direct object of pisñ ?), see above p. 304.



CHAPTER NINE

THE GENERAL VOCABULARY AND THE PROPER NAMES

A. G V

Our knowledge of Carian general vocabulary is unavoidably very lim-ited; we have already seen (Chapter 2) that the indirect sources pro-

vide us with a very small number of glosses, and these in turn, despitetheir seemingly reliable origins (they appear to date back to Greekauthors of Carian origin), remain uncertain insofar as none of themhas been identified in the Carian inscriptions up to now. As for thedirect documentation, the only texts that can be interpreted with anydegree of confidence are a number of brief inscriptions that containexclusively, or almost exclusively, onomastic formulae. It is true thatsome of the longer texts must contain essentially Carian common words,

but they are in general impossible to analyse. In this necessarily brief section, I will limit myself to collecting in a synthetic way the generalvocabulary that arises from an analysis of the inscriptions (see Chapter7): further details about these forms and about others, which are inter-preted more hypothetically, will be presented in the corresponding entries of the Glossary (Chapter 11).

Given the typology of a great number of Carian inscriptions, it isnot surprising that one group of the common words whose meaningswe can establish belongs to the semantic field of kinship: we know the

Carian words for ‘son’, mno-, ‘father’ ted , and ‘mother’ en, although inthe case of these two latter terms, the interpretation is based solely onthe combination of an etymological connection with other Anatoliandialects and the suitable, but unique, context in which they appear (E.Me38 and E.Me 32, respectively). No other kinship terms have been iden-tified directly in the Carian inscriptions, and we can only speculate thatthe personal names quq-Gugow and ksbo- Xasbvw might also be the com-mon words for ‘grandfather’ and ‘grandchild’, on the basis of the good

Lycian parallels x uga ‘grandfather’ and x ahba ‘grandchild’.Due to the funerary character of many Carian texts, a collection of words has been compiled which refer to the funerary stele or, moregenerically, to the tomb: upe/wpe/upa, ue, ≤ jas/≤ as, s(i)d i . However, no



327

clear etymological connections can be established for any of these words(see Chapter 11 ss. vv. ) and it is impossible to specify the exact mean-

ing in each case. The only definite word to have been identified fordiff erent objects is ork-, which appears on two bronze phiales in accusative(orkn ), hence it must mean directly ‘phiale’ or more generically ‘bowl,receptacle’. Also secure is the identification of øtr- as the Carian wordfor ‘person, self ’.

Apart from these few words, most other items of Carian generalvocabulary have been identified based on merely hypothetical analysesof the texts, which are not necessarily accepted by all the scholars. Forexample, I have argued for mso- as the Carian word for ‘god’. Theconnection is formally suitable (cf. CLuw. mà“ san(i)-, Lyc. maha(na)- andparticularly Mil. masa- ), and has good parallels in Carian onomastics(msnord- ), but it relies on my own interpretation of C.Hy 1, where it isfound. This problem is particularly evident in the case of Carian verbs;we have already seen (Chapter 8) how difficult it is to find verbal formsin Carian, so practically all the possible examples of verbs are in factdependent on a concrete interpretation of certain Carian inscriptions,which are in general very difficult to analyse. The clearest form, ÿbt

in C.xx 1, points to a verbal stem ÿb-, ‘to off er’, but this interpretationis essentially based only on the good parallel ubete ‘off ered’ in Lycian.Similar situations arise in the case of other possible verbs, such as ait ,which would seem to indicate a stem a- ‘to make’ (= CLuw. à/à ya-,Lyc. a(i)- ), or pisñ , compared (see above Chapter 8) to CLuw. piya-,Lyc. pije- ‘to give’. Needless to say, in cases like this, there is a cleardanger of resorting to circular arguments: if we take the example of thelatter form, we can establish a meaning ‘to give’ for an alleged verbalform pisñ only on the basis of an etymological connection, yet we then

believe it plausible to claim Carian has a verb pi- ‘to give’, which inturn confirms a genetic relationship between Carian and the Anatolianlanguages! Lacking more concrete evidence (bilingual texts, convincing combinatory analysis and so on), these forms—usually hapax legomenain uncertain contexts—must be used with great care, and must beincluded in our analysis only as provisory and hypothetical possibilities.

On functional words (demonstrative and relative pronouns), I referto Chapter 8. To those forms we can add the clear identification of a

copulative conjunction: sb (to be connected to Mil. sebe , ‘and’).



328

B. P N

As stated repeatedly in this work, the area that yields the greatest vol-ume of information is Carian onomastics: the sources, both direct andindirect, provide a large number of proper names, currently the mostvaluable information on Carian available.

Carian onomastics can be analysed in two diff erent, but comple-mentary and interrelated ways. Firstly, diverse stems and suffixes canbe identified through an internal analysis of the compounding and deriva-tion mechanisms that clearly intervene in their formation. Secondly,many of these names can be totally or partially compared to the rest of Anatolian onomastics, both from the first and the second millennia B.C.

Given our scarce knowledge of Carian common vocabulary, the firstapproach yields very limited results: we are only able to establish someregular patterns, both in composition and derivation, from a purelycombinatory analysis of the onomastic materials, and the meaning of the elements—lexical stems, suffixes—that we isolate cannot be established.Yet despite its limitations, this method of analysis is not altogetherworthless, insofar as it allows us to identify some recurrent procedures

in the formation of proper names. A good example is the series of stems that appear combined, giving rise to a number of very charac-teristic compound proper names, as the table of the p. 330 is intendedto show. We find here a set of first elements (i)d-, par(a)-, pun-, “ ar-,etc. that in general never appear as independent words and that areused alongside a set of second elements (quq, u≤ ol , etc.), which are forthe most part also documented in a simple, non-compounded, form.This implies a diff erent functional nature of first vs. second elements,a possibility that can be confirmed in at least those cases where an ety-

mological explanation is plausible (see below).More difficult is to establish possible derivation procedures without

the use of comparative evidence. The segmentation of suffixes is notalways easy, and it is also difficult to specify in some cases whether weare dealing with a suffix or a lexical stem in composition; for instance,the useful list of suffixes, created by Blümel using only this approach,and included in his corpus of Carian personal names (Blümel KarPN:32–33), off ers some elements that can confidently be considered (on

the basis of an etymological approach) lexical stems, and not suffixes(-biw, -muhw ).1 In any case, several suffixes can be well established from

1 It must be said, however, that Blümel, very cautiously, does not speak of ‘suffixes’:



329

directly and indirectly attested personal Carian names: -ol (Greek-vllow/-vldow, -vlliw ), -s(i)- (Greek -assiw ), -om (Greek -vmow ). As for

place names, the well known suffixes -(V)ssow and -(V) nda are abun-dantly attested.Comparison with the other Anatolian proper names off ers more

promising results. As I explained in Chapter 2, the research on Anatolianonomastics during the last century demonstrated not only the existenceof a clear linguistic unity to which Carian onomastics belongs, but alsothe Anatolian Indo-European character of this linguistic unity. In otherwords, an important part of this geographical and chronological ono-mastic continuum can be confidently analysed etymologically, since alot of lexical stems, suffixes, and morphological procedures of theAnatolian Indo-European family can be identified. Three excellent exam-ples of this type of interpretation are the decisive book of Laroche onCuneiform Anatolian personal names (Laroche: LNH), the equally influ-ential work of Houwink Ten Cate devoted to Lycian personal names(Houwink Ten Cate 1961), and finally the contribution of Zgusta (Zgusta1964b), which although cited less frequently, is nonetheless very valu-able. Comparison of Carian proper names with the rest of Anatolian

onomastics therefore not only provides us with a greater capacity forcombinatorial analysis, but also allows us to establish the meaning of many lexical and functional elements that intervene in their construction.

In the following pages, I will off er (in a non-exhaustive, but in myopinion sufficiently representative way) examples of this latter approachby collecting several lexical elements of Anatolian origin that can beidentified in the Carian name system, and which generally also appearin the known inventories of proper names from the rest of Asia Minor.This exposition is very similar to that presented in Adiego (1993a),where Houwink Ten Cate (1961) was used as a model,2 but in thiscase, the names directly attested in Carian inscriptions are also included.3

the list appears in a section entitled ‘Komposition und Wortbildung’, although truesuffixes appear mixed together with these lexical stems.

2 Names with no indication of origin must be taken as Carian.3 The reasons for this inclusion are given in p. 15.



330

i b r e l -

k b i o m -

- m u h w

q u q -

p / b i k ( a ) r m - p / b i k s -

w l i / j a t -

- u a s s i w

u ≤ o l / w ≤ o l

y d i q - / y r i q -

O t h e

r s

/ k b j o m -

/ y r i ∞ - / - y d ∞

Ø

I m b

a r h l d o w

k b j o m -

q u q -

p i k ( a ) r m

p i k s -

w l i a t -

u ≤ o l - /

w ≤ o l -

K e b i v m o w

G u g o w

/ w l j a t

U s s v

l l o w

U l i a t o w ,

O l i a t o w

A k t a -

A k t a u a s s i w

A k t a

u s s v l l o w

A k t a d h m o w

( i ) d -

d q u q -

d b i k r m ,

d b i k s -

d w ≤ o l

- , i d u ≤ o l

i d y r i ∞ -

i d m n (

- s )

I d a g u g o w

d b k r m

I d u s s v l l o w

i d m u o

n -

i d y e s -

p a r ( a ) -

p a r a i b r e l -

P a r (

a ) u s s v l l o w

p a r ÿ d ∞

p a r a e y m -

p a r y r i ∞

p a r p e y m

P a r a u d i g o w

P a n ( a ) -

P a n a m u h w

P a n u a s s i w

P a n

a b l h m i w

p ( u ) n -

p n u ≤ o l - / p u n w ≤ o l

p ñ n m n n -

- p n w ≤ o l - , p n u ≤ o l -

P o n m o o n n o w ?

P o n u

s s v l l o w

“ a -

S a u s

s v l l o w

S a u r i g o w

“ a ÿ d i q -

“ a y r i q -

“ ( a ) r -

“ a r k b i o m -

“ r q u q -

“ a r w l j a t

S a r u a s s i w

“ a r u ≤ o l -

“ r w l i -

S a r u

s ( s ) v l l o w

O t h e r s

E k a m u h w ,

a [ r b ] i k a r m

k d u ≤ o l ( - “ ) ?

p r p w r i ∞ ?

E j a m u h w

t d u ≤ o l

-

S e n u r i g o w

X e r a m u h w

p s u ≤ o l -

Y u s s

v l l o w

K a r u

s v l d o w

M a u s s v l l o w

C a r i a n C o m p o u n d

P r o p e r N a m e s

331

1. Theophores

A type of personal name very widespread in Anatolian onomastics isthe theophore: we find god names used directly as personal names,dvandva compounds in which two god names are associated¸ compoundsconsisting of a personal name and a lexical element, nouns derived bysuffixation from a god name, and so on (see the enlightening study byLaroche LNH:282–295).

§ 1. 1 (Hitt., Luw.) Arma , Carian armo-Arma is the Hittite and Luwian Moon-god (Laroche NDH: 80). This

god name is now also documented in Carian, in the dvandva formarmotrq d - from Hyllarima (C.Hy 1).4 Its use in the formation of propernames is commonly seen in the languages of the second millennium aswell as those of the first millennium.

Carian PN Ermapiw < Arma + piya (§ 3.1) = Armapiya (Laroche LNHnº 135).

Cf . Ermapiaw (Zgusta KPN § 355–20, Lycia), Armapiaw (KPN § 97–3,Lycia, Cilicia), Armapia (fem., KPN § 97–4, South of Phrygia-Lycia).

a) Other instances (not in Caria) of Arma , either alone or in composition withmuwa-, nani-, etc.: Houwink Ten Cate (1961:132–134), In Hittite and Luwian:Laroche (LNH:290)

b) The variant Erma- (besides Arma-, Arma- ) of the Carian name and others,is usually attributed to the analogical influence of the Greek divine nameÑErm∞w (already Kretschmer 1896:361).

c) The Egyptian flavour of the name Ermapiw (Sayce 1887[92]:122) was con-sidered by Masson as merely coincidental, because the name can be satis-factorily analysed from an exclusive Minor Asian point of view (Masson1959:167–170). However, Hornblower (1982:357, n. 35), though accepting

Masson’s views, wonders if in cases such as this a sort of homonymy couldinfluence the choice of the name.

§ 1. 2. Luw. Tar ¢unt-, Lyc. Trqqas , Mil. Trqqiz , Carian trq(u)d - Cf. Hitt.Tar ¢u-The Anatolian Storm god. The name has a good PIE etymology (*terh2-‘to cross’, ‘to pass’: in Anatolian ‘to overcome’, * térh2-u- from a-u-present).

4 Although I prefer to interpret armotrq d - as a theonym representing a divine pare-dra due to the context in which it appears, I do not completely rule out the possi-bility that it could be a personal name.



332

The god name is directly attested in Carian: trqud e C.Ia 3, ?/trq d /?C.Ki 1, armo-trq d - C.Hy 1. At least in Iasos and Hyllarima, it is most

likely that we are dealing with the divine name itself.Place name Tarkondar//a// deducible from the name of a syn-geneia Tarkondare›w (Mylasa).

More dubious are Konodvrkond//a// possible place name, from aname of phyle ≤ Konodvrkondevn in Mylasa, Otvrkond//a//, from aname of a phyle ÉOtvrkonde›w.

a) Internal analysis of Tarkondar//a// is not clear: Tarkond+ar(a)-? (perhapsbetter than Tarkon+dar(a ), Adiego 1993a:28).

b) In Konodvrkond(a), Otvrkond(a ), neither the internal structure nor the ori-gin of v vocal are clear (the connection with Hitt. ¢anna- ‘grandmother’,Adiego 1993a:28 is very hypothetical). For the second name, a prothetico- could be postulated (Adiego 1993a:28),

§ 1. 3. Lyc. natrº-, god name to be identified with ApolloThe existence of a god name natrº, equivalent to Apollo, can be deducedfrom the well-known ‘translation’ ÉApollÒdotow of the indigenous name Natrbbijemi in the Xanthos bilingual. The god name seems to appeardirectly in Carian in the form ntro- (see Chapter 11, s. v.), and canalso be recognized also as intervening in the formation of several per-sonal names from Greek sources. The clearest example is Neterbimow,undoubtedly the same name as Lycian Natrbbijemi . Also of this familyare the names Nvtrassiw and, less clearly, Nutar.

§ 1. 4. CLuw. màssan(i)-, Lyc. maha(na)-, Mil. masa- ‘god’, Carian mso-(?). Cf. also Sidetic masara = yeo›wA clear isogloss shared by Luwian, Lycian, Milyan and Sidetic (the pre-

sumed ‘Luwic group’) in contrast to Hittite, Palaic and Lydian, is thecommon word for ‘god’. While Hittite and Lydian have forms inheritedfrom PIE *dyew- (Hitt. “ iu(n)-, secondarily thematic forms “ iuna-, “ iuni-;Lyd. civ- ), the Luwic dialects present a form masa-, mà““ an(i)-/*masa(na)(> Lyc. maha(na)- ), masar(a/i)-, of unknown origin and with seriousformal problems (is masa- the original form, and ma ““ an(i)-, *masana-massar(a/i)- secondary formations?).

Place name Masan≈rada, msnord-, possible ethnical name from thiscity. This can be analysed as a compound name with a second ele-ment equivalent to Luwian *(a)radu-, see below.

Place name Massvn//a// (name of a demos in Mylasa).Massariw (Carian name of Dionysos), PN Massarabiw, seem to con-

tain an r-form of the stem, as in Sidetic. Massarabiw can be analysedas *masara-bi- (-bi = -piya , see below § 3. 1).



333

2. Some Nominal Stems

§ 2. 1. CLuw. annara/i- ‘forceful, virile’, ànnari- ‘forcefulness, virility’,etc. Cf. Hittite innar à-, ‘forceful’.Perhaps this stem can be seen in the place names Naras//a// andNaruandow5 and in the name (or title?) naria- (see Chapter 11. s. v. ). Cf .also Narbaw?

Cf . also the name Andarsvw, if the explanation of -nd(a)r- as a resultof *-nr-, given in p. 262, is accepted.

§ 2. 2. CLuw.(?) *(a)radu-This element can be identified in two Arzawian names, Tarhundaraduand Piyammaradu. The status of (a)radu- is unclear: it could be a godname related to the name of the Luwian stag-god, Runt(iy)a- (a morerecent form of Kurunta- ), but this connection is far from certain (cf .Laroche LNH:282, n. 6). Schürr has claimed to identify this elementin the Carian place name Masanvrada, ethnic (?) msnord-, assuming that the place name derives from a PN Masanvradow, the name of the city’s founder according to the information given by Stephan of

Byzantium (Schürr 2002). Although the validity of this latter point mustbe viewed cautiously (Masanvradow could be an artificial creation fromthe place name), Schürr’s argument is in my opinion quite persuasive.Schürr also proposes that the same element can be identified in theCarian place name Parembvrda (analysed as Paremb-vrda, Schürr2002:170), and that the possibility also exists that other Carian namesin Greek sources that include a sequence ardº, ordº belong to this fam-ily of nouns (place name Ardur//a//, PN Ordomaw ). Even the PNArduberow/ardybyr-, if analysed as Ard-uberow/ard-ybyr- (not Ar-duberow/ar-

dybyr- ) on the basis of ybrs- (C.Hy 1), could contain the same stem.6

§ 2. 3. CLuw. arpa- ‘confusion, tumult, strife’, Lyc. erbbe- ‘strife, battle’7

The stem CLuw. arpa-, Lyc. erbbe- intervenes in the formation of propernames, but it is only attested in those of the first millennium, as Houwink

5 Neumann (1998:185), by contrast, connects Naruandow to an alleged Hitt. naru-,

a type of plant, but this latter word is actually Akkadian (*narû , a type of malt, seeTischler 2001, s. v. ).6 According to Melchert, “the element -arada may be analyzed as a derivative of

the word seen in Hittite ard- ‘companion’. A Luwic *arada- would mean ‘community’,which seems appropriate for a placename. The further derived -aradu would againmean ‘companion’ or similar” (pers. comm.).

7 Not ‘defeat’ (Adiego1993a:30), see Melchert DLL.



334

Ten Cate has pointed out (1961:147–148).The basis for his association of -arba-/-arbh- with this Luwian word

is a passage in Milyan of the Xanthos stela: trqq[i]z : esetesi[k]e er[b]besi[k]e:lusasi : esene[m]la (TL 44d 12–13). Houwink Ten Cate takes the sameview as Laroche, who associated erbbe with Luw. arpa-. As for esetesi [ k ]e ,CLuwian also off ers a possible direct cognate: a ““ atta ““ i-, Gen. Adj. of a ““ atti- documented as divine epitheton (d À la “ a ““ atta ““ i “ ). It is clear there-fore that er[b]besi[k]e and esetesi[k]e are genitive adjectives used as epi-thets of god Trqq[i]z . The key to connecting erbbe-/arpa- with onomasticforms is the Cilician proper name Trokoarbasiw (KPN § 1512–22)where the same god name and one of his epithets appear together incomposition.

PN Arbhssiw, Arbhsiw. The same name in Cilicia: Arbasiw (ZgustaKPN § 85–1).

§ 2. 4. HLuw. hasu- Lyc. x ahba- ‘grandson’PN ksbo, Greek Xasbvw. Both Carian and Lycian point to a secondary-à stem (Carian -à > -o, cf . armo ) from the u-stem attested in HieroglyphicLuwian (cf. Melchert DLL). Cf . also Kasbvlliw, apparently derived

from ksbo/ Xasbvw.

§ 2. 5. CLuw. ¢à pa/i- ‘river’, *¢apài- ‘irrigate, water’It possible that the sequence kb- (Greek Kebº, Kbº) that appears in someCarian names (kbjom-Kebivmow and “ arkbiom, kbdmu-, Kbondiassiw, Kbvdhw,and the indigenous place names kbid- for Kaunos and kbo- for Keramos)is the Carian result of this Luwic stem: at least from a phonologicalpoint of view, the correspondences are appropriate, and in the case of the two place names, a denomination ‘river’, ‘irrigated (land)’ or sim-

ilar is very suitable (cf . Schürr 2001a:64 for Keramos-kbo ).

§ 2. 6. CLuw. ¢ù¢a-, Hitt. ¢u¢¢a-, Lyc. x uga-, Mil. x uga- ‘grandfather’This word intervenes in the formation of the onomastics of the secondmillennium: Huhanani (LNH nº 379), Huhhaziti (LNH nº 385), etc. It isalso documented alone in the Lycian name Kougaw (KPN § 717).

Carian: quq , Greek Gugow.dquq , Greek Idagugow. The (first element, (i)d- Ida-, is well docu-

mented as a proper name, either alone (Ida KPN § 451–1, Eida KPN§ 451–2, both in Lycia) or as the first element of compounding: Carianidu≤ ol-, du≤ ol - = Greek Idussvllow, dbiks-, dbkrm-; Lyc. Eida-ssala,Ida-zzala KPN § 451–10, cf . Salaw, Zzala KPN § 1358–1).

I now have little doubt that the name of the Lydian king GÊghw must



335

have the same origin. The problem posed by the phonetics (Lydiandoes not conserve PIE laryngeal *h2, unlike the other Anatolian dialects)

can be overcome if we imagine the name to have a Carian origin. Asa result, the long discussion in Adiego (1993a:40–41) does not need tobe repeated here.

§ 2. 7. CLuw. im(ma)ra/i- ‘open country’im(ma)ra/i- (only verified as common word under the form of Gen Adjim(ma)rassa/i- ) is the CLuw. word corresponding to Hitt. gimmara-. Itappears in the formation of a proper name already found in Cuneiformsources: Immaraziti (Laroche LNH nº 450). In the onomastics of thefirst millennium it adopts the forms imbr-, imbar-, etc., id est, with thedevelopment of an epenthetic labial. A possible form without labialcould be the place name Imrougara (KON § 374, Eastern Phrygia; cf .Zgusta KON:198–199). In Lycian direct documentation, the word seemsto appear as ipre . With this in mind, the proper name Ipresidi is veryinteresting, as it could correspond exactly to Luw. Immaraziti .

Most instances of proper names containing this element from thedocumentation of the first millennium come from Lycia and Caria.

Place name ÖImbrow (promontory near Kaunos), from the pure stem.PN para-i b re l , Imbarhldow. This corresponds formally to CLuw.im(ma)ralla-, adj. attested fragmentarily (im-ra-a [ l ) and as a place-name(URUImralla ), see Melchert CLL s. v. )

PN i b rsi-/i b arsi-/b rsi-, Imbrasiw, Imbrassiw, Imbarsiw. It correspondsto the CLuw. Gen. Adj. im(ma)ra ““ a/i- or to a further derivation of thisword (*im(ma)ra ““ iya- ).

ÖImbramow, var. ÖImbrasow, Carian name of Hermes, Steph. Byz. s. v.ÖImbrow. ÖImbrasow is also the name of a river in Samos.

Schürr’s objections8 to connecting these forms with Luw. im(ma)ra-are in my opinion not overly convincing.

§ 2. 8. Lyc. mere- ‘law’, maraza- ‘judge, arbitrator’Panamara. Cf . also PN Mareuw, Marow.

§ 2. 9. Hitt. muwa-, Luw. *mùwa-, Mil. muwa- ‘might, power’This word, common to both Hittite and Luwian, is one of the most

characteristic terms in the formation of proper names from all periods.

8 Schürr (1991[93]:171), (2001b:104–105).



336

Before the discovery of Hittite and Luwian it had already been identifiedby Kretschmer (1896:332–333). It was Friedrich (1931) who connected

the Anatolian names of the first millennium with the names in -muwa of the second millennium, and who identified muwa- as a common wordin Cuneiform texts.

Carian: MÒtulow, mythic founder of the Carian city Samylia (Steph.Byhz. s. v. Samul¤a, Zgusta KPN § 976: not in Blümel KarPN) = Muwatalli-, a name well documented in Cuneiform sources (LarocheLNH nº 837).9

With first element pan(a)-/pun(a)-: pñmnn-, Greek Ponmoonnow, Panam-uhw, Panamuaw, an inhabitant of Kos (Zgusta KPN:695), place namePounomoua.

Pormounow is most likely a more recent form of pñmnn-, since itappears as the name of a Syngeneia in the temple of Sinuri, where pñmnn-/Ponmoonnow is documented in an earlier inscription.10 The soundchange required is commonplace (nm > rm by nasal dissimilation).

Other compounds with muwa- as a second element: uksmu-/wksmu-,a Carian name corresponding to Ouajamoaw (Zgusta KPN § 1141–2,Isauria, Cilicia), Ouajamvw (Zgusta KPN § 11141–2, ibid .). Perhaps also

kbdmu-.muwa- as the first element in mwsat-, corresponding to Lydian Mousathw(Zgusta KPN § 987a), Pisidian Moushta, Moshta (all equivalent toCLuw. Muwaziti- Laroche, LNH nº 840?). Less clear is myze-, Mouzeaw.

§ 2. 10. Hitt. peda-, Lyc. pdden- ‘place’Some Carian place-names show a stem Phda-, Pida-, Peda- that canbe connected with Hitt. peda-, Lyc. pdde and, ultimately, with a Proto-Anatolian * pédon ‘place’, from PIE * pédom (> Greek p°don, etc.). The

stem also appears in the place names from Cuneiform sources, Petassa and Pitassa (See Laroche TA1, nº 49).

Phdasa, Pidasa (pl. nt.) (§ 1054–1). According to Blümel KarON,s. v., at least two places bore this name: Phdasa near Halikarnassos,and Pidasa in Grion.

Pedanass//ow//. For the reading, see Blümel (KarON:177, n. 56),not Pidº (Adiego 1993a:43)!

9

It has been suggested that MÒtulow not only corresponds to the name Muwattallibut is in fact also the name of a Hittite king, transmitted by the Greek sources. Thequestion has been raised in the more general problem of the historicity of the TrojanWar (Motylos was a king who received Paris and Hellena during their flight to Troy).

10 The equivalence of the two forms was already suggested by Robert (1945:98).



337

Pidossus § 1058–1, an island near Halikarnassos.Cf . Perhaps also the personal name Pedvldow.

In Adiego (2000), I proposed that the same stem could be seen inC.Si 2 pd a ∞m≤ uñ /?, see Chapter 11, s. v.

§ 2. 11. Luw. * pi ¢a- ‘luminosity, splendour, might’This stem appears in Luwian in this form only in personal names (SeeMelchert CLL), but various derivatives are verified in common lexicon( pi ¢amma/i-, pi ¢a ““ a/i-, pi ¢atta/i- ).11 The etymological interpretation isclear: pi ¢a- comes from PIE *bheh2- ‘shine, glow’ (OInd bhà ti ‘shines’,Greek fa¤nv ‘make visible’ etc.). The stem had already been identifiedby Houwink Ten Cate (1961:156–157) in the Anatolian onomastic sys-tem, particularly in the Luwian area (at this point, however, the mean-ing of the root was unknown).

Carian: piks-, dbiks-; Pigassvw. pikre-/pikra- = Greek Pigrhw (var. Pikrhw, Pitrhw?). This is a frequently

documented name (either in this form or through derivates, as Pigrassiw,§ 1255–5, Pamphilia), not only in Caria but also in other countries of

Asia Minor: Lycia, Pisidia, Pamphilia, etc. The only possible cognatein Cuneiform sources is Pi ¢irim, the name of a Cilician king, LarocheLNH nº 977, but this is an isolated form (cf. Adiego 1993a: 36–37).

pikarm-/pikrm-, dbikrm-/dbkrm-, cf . the Lycian name from Greek sourcesPigramiw, Pigramow (Zgusta KPN § 1255–1/2).

Pijvdarow (internal structure not clear).It is very doubtful that the two Carian place-names, Peigelasow and

Piginda, belong to this group.

§ 2. 12. CLuw. pùna- ‘all, totality’, cf . also CLuw. pùnata/i- ‘all’, Lyc. punãma- ‘totality’In Anatolian onomastics from all periods, two elements, puna- and pana-are repeatedly found. Laroche (LNH:32) considered them two variantsof a single form, the meaning of which was unknown at the time.Nowadays, a quantifier meaning ‘all, totality’ or similar seems to bewell established for Luwian pùna- (and derivative pùnata- ), and for Lycian

11 The last two only in the form of GenAdj, with -a ““ a/i- suffix attached to them(Melchert CLL, s. v. ).



338

punãma-. The same meaning can be determined for the onomastic com-ponent. Whether pana- was actually a variant of puna- is less clear ( pana-

does not appear as a common lexical item in Luwian). If this is thecase, the alternation must be very old, because both puna- and pana-are found in the Cappadocian onomastics, from the beginning of thesecond millennium.

P ùna- in Carian:PN pñ-mnn- = Pon-moonnow, Por-mounow.Place name Pouno-mouaFor these forms, and particularly for Pormounow, see above muwa-. pnu≤ ol -/pnw ≤ ol -/pnw ≤ ol- Pon-ussvllow.More doubtful: pnyri ≤ ru-.It is possible, although very hypothetical, that the common noun also

appears in Carian direct documentation in the form punot (see Chapter11 s. v. ).

Pana- in Carian:Pana-blhmiw. For the second element, cf. Lyc. -plemiw, -plemi, -pl ˜ mmi

in Sedeplemiw, Sedepl˜ mmi, Esedeplemi (KPN § 1387–1/3).Pana-muhw. For the second element, see above § 2. 9.Pan-uassiw. For the second element, cf . Akta-uassiw, Sar-uassiw.Place name Pana-mara (for -mara, see above § 2. 8).

§ 2. 13. CLuw. ura- ‘great’, HLuw. ura/i-, cf . Hitt. ura/i- ‘great’This adjective appears in the Anatolian onomastics of both the secondand the first millennia (cf . Houwink Ten Cate 1961:164–165): LarocheLNH: nº 774: Massanaura , nº 872 Nattaura , n 1431 Ura , 1437 Urawalkui ,

etc. Oraw (Lycian; Zgusta KPN § 1100–1), Ouramoutaw (Cilician; ZgustaKPN § 1169), etc.

In Carian it can be recognised in the personal names urom-, urm-,wrm-, and in the place name Urvmow (converted to Eurvmow, Eurvpowby the influence of Greek). However, the etymological connection of this latter form to ura- is challenged by the existence of variants suchas Kuròmew and huròmew (see Blümel KON s. v. Urvmow ) because theinitial k-/h- would then remain unexplained.

§ 2. 14. Hitt. *walli-, walliwalli- ‘stark, mighty’ (Cf . CLuw. wallant- ‘fit,capable’)?Some Carian names point to a stem that could be reconstructed as*wala/i-:

339

wliat/wljat , Greek Uliatow, Oliatow (documented directly in Carianas wliat /wljat )

Oaloalow (about this form, see Adiego 1993b).At least in the second case, a direct connection with the redupli-cated stem Hitt. walliwalli- seems very likely (< PIE * äuelH- ‘to be stark’).

For Uliatow, however, there is also a possible relationship with thefamily of Hittite walli-, ‘glory’, CLuw. walli(ya)-, HLuw. wa/iliya-, ‘toexalt’ (see Melchert CLL s. v. ), from PIE * äuelh1– ‘to choose, to select’(if we are not actually dealing with the same root).

3. Verbal Stems

§ 3. 1. Hitt. pài-/ piya-, Luw. piya-, Lyc. pije-, Pal. pi (sa )-, Lid. bid- [ pid- ]‘to give’All the Anatolian languages in which the verb ‘to give’ is found con-tain a similar form that has a clear common Proto-Anatolian origin.Its use in onomastics is also common to all Anatolian linguistics andvery productive, above all in the formation of theophores: Tarhundapiya (LNH nº 1267) = Tarkumbiou (gen.; KPN § 1512–13, Cilicia).

Ermapiw, cf . above arma-Massarabiw cf . above massan( ì )-As well as the simple form of the stem, we also find in Anatolian

onomastics forms that can be clearly denoted ‘Luwic’ participles in*-mo/i- (cf . CLuw. -mma/i-, Lycian -me/i- ): Laroche LNH: nº 980 *Piyama-dKAL, nº 981 Piyamaradu, Lycian Armadapimiw, Pisidan Kouadapemiw, etc.In Carian: Neter-bimow = Lyc. Natr-bbij˜ emi = ÉApollÒdotow (in the trilin-gual inscription of Xanthos).

4. Adverbs

As previously mentioned, Carian contains some characteristic compoundnames that were systematized in the table above (p. 330) only on com-binatory grounds. A comparative approach serves to confirm that, atleast in two cases ( par(a)-, “ ar- ), the first elements of these compoundnames can easily be connected to adverbial stems in other AnatolianLanguages. The use of adverbs as the first elements of compounds is

also verified by the rest of the Anatolian proper names, particularly inLycian (see Houwink Ten Cate 1961:172–175 and the forms citedbelow).



340

§ 4. 1. Hitt. par à, CLuw. par ì , HLuw. para/i ‘forth, away’, Lyc. pri ‘forth; in front’.

This adverb appears in Carian as para-/par-, Greek Para-/Par-: para-eym, para-i b re l , par- ÿd ∞- (= Greek Para-udigow ), paryri ∞-.Para-ussvllow, Par-ussvldow. Probably also in the place name

Paremborda (Neumann 1988:191, cf. also above p. 333).

§ 4. 2. Hitt. “ ara, “è r , Luw. “ arra, “ arri , Lyc. hri “ ar-/“ r-, Greek Sar- in Carian:“ ar-u≤ ol -, Greek Sar-us(s)vllow“ r-quq, “ r-wli — Sar-uassiw. For the second element, cf . Pan-uassiw.The following forms might perhaps be considered as a variant of

Sar-: “ a-, Greek Sa-:“ a-yriq- (= Greek Sa-urigow ), Sa-usvllowNote the parallel use of hri- in the Lycian personal name Hri x ttbili

(besides simple Ktibilaw, cf . Melchert DLL, s. v. Hri x ttbili ).

5. Lallnamen

As mentioned in p. 13, Lallnamen—hypocoristic names whose structurerecalls the language of children, characterised by open syllables, withor without diff erent patterns of reduplication—were identified byKretschmer (1896) as characteristic of the Anatolian onomastic system.Hittite and Luwian evidence confirms the antiquity of such formationsin Anatolian (see the exhaustive analysis in Laroche LNH:239–246).Carian is no exception: we can identify a considerable number of namesthat can be interpreted this way, although mostly in Greek sources,

since only one Lallname, ada , is directly attested. I refer here to AppendixC, where it will not be difficult for the reader to identify them in thelist of Carian names from indirect sources. Many of these Lallnamen seemto obey simple structures of the type aC(C)a (Aba, Abaw Abbaw ada-Ada,Adaw Appa ) and reduplicated CaC(C)a (Nana, Nanaw, Papaw, Tata, Tataw,perhaps also forms in -h, -o- and -v if these stem vowels are secondary:Nanh, Nannh, Nannow, Nannow, Nannv ), while others show somewhatdiff erent patterns (Minnaw, Nonnow ) or are derived from suffixes of un-

known value (Amiaw, Ammiaow, Nannixow, Papiaw, Tatarion, etc.).



341

6. Su ffi xes

Extensive evidence exists of the use of diff erent suffixes in the forma-tion of Anatolian proper names, but as Laroche (LNH:327–328) rightlypointed out, their analysis is hampered by diverse factors: the impos-sibility of explaining their actual origins and values, the risk of incor-rect segmentations (“les dangers du découpage formel” in Laroche’swords),12 and the great variety of stems to which they can be attached,which prevents us from identifying the derivational mechanisms behindthe construction of such names. These problems are compounded inthe case of the indirect documentation, insofar as the Greek adapta-

tions may be masking a more complex situation. A good example of this is provided by sigmatic suffixes; taking into account only Greekevidence, one might be tempted to consider a single suffix -ssi- bothfor Arlissiw and for Imbrassiw. Direct evidence, however, available forthese forms but not for others, indicates two diff erent Carian suffixes,-“ - for arli “ -Arlissiw and -si- for (i)b (a)rsi-Imbrassiw. Note the parallelsituation in Lycian, where -si- in Mollisiw and Triendasiw from indi-rect sources, treated as a single suffix (-zi [sic]) in Houwink Ten Cate(1961:183–184), in fact corresponds to two Lycian diff erent suffixes,-se/i- ( Mullijese/i- ) and -zi- (Trijetezi- ).13 It is because of these doubts thatI shall limit myself to listing a very reduced number of suffixes, citing only those cases where the identification of the suffix seems clear.

§ 6.1 Place name suffixes -assa-, -nda- (Cuneiform sources) = -(a)ssow,-(a)nda (Greek sources)These two suffixes intervene in the formation of a large number of Anatolian place names, documented both in the second and the first

millennia. As we have seen (cf . p. 13), it was Kretschmer (1896) who,systematizing former ideas, rightly concluded that the Anatolian placenames in -ss-, - nd- widely attested in Greek sources belonged to a sin-gle linguistic family spoken in Anatolia. Later, the (re)discovery of theAnatolian languages of the second millennium allowed us to confirm thistheory, thanks to the appearance of numerous place names in -assa-,-nda-, as well as authenticating these suffixes as purely Indo-European

12 Laroche (LNH:328, m. 25).13 It is even possible, if we accept Neumann’s analysis (Neumann 1978:64; cf . also

Melchert DLL, s. v. ) that Mullijesi- is actually a compound name from *mulli+ esi ‘shallbe strong’, parallel to Aruwãtijese/i- ‘shall be high’ (*aruwãti + esi ).



343

§ 6. 2 Luwic participle suffix (CLuw. -mma/i-, Lyc. -me/i- )Laroche categorised as original Luwian participles a number of per-

sonal names with a -mma/i- suffix in Cuneiform sources (for instance Ayami/aimi , from Luw. à/-à ya- ‘to make’, Laroche LNH:530), althoughnot all the names that contain these suffixes can or indeed should beconsidered participles.

Names in -mow, -miw are abundant in the Anatolian onomastics of Greeksources, and Carian is no diff erent. However, as in the case of the sec-ond millennium names, it cannot be automatically assumed that theseare original participles, as this would constitute an oversimplification.Rather, it is more advisable to interpret them in this way only if theirinternal structure is clear and the base to which the suffix is attachedcan be confidently identified as a verbal stem, in a similar fashion tothe abovementioned name Ayami/Aimi. In assuming this more accurateanalysis, the possibilities of recognizing participles among the Cariannames with nasal labial suffix decreases dramatically. In fact, we areleft with only one Carian name that can be interpreted as a participlewith absolute certainty: Neterbimow, thanks to its clear correspondenceto Lyc. Natrbbijemi , a transparent name formed by a theophore ( Natr- )

and a participle of the verb pije- ‘to give’ (see above § 1.3 and § 3.1).The search for other participles in Carian names yields only hypo-thetical results: it is plausible that names such as qtblem-/Kutbelhmiw/Kotbelhmow, or those ending in -om (kbiom-Kebivmow, arliom-Arlivmow,etc.) may come from participles, but without a clear identification of possible verbal stems in the sequences preceding -m-, it remains indemon-strable.

§ 6. 3. -alla/i-

Anatolian personal names characterized by a suffix -alla/i- are wellattested (Laroche LNH:329–330). As with -mma/i names, formations of diff erent origin and meaning have converged under an apparently sin-gle -alla/i- suffix, which could also be valid in the case of Carian com-parable forms.

This suffix appears in Carian with the form -ol (Greek-vll -/-vld- )and perhaps also -e l (Grek -hld- ). I refer the reader to Chapter 6,p. 258 for a more detailed phonological explanation of these Carian

forms; here it is sufficient to note that a vocalic change à > o, commonlyseen in Carian, serves to explain the very particular form -ol - takenby this suffix in this Anatolian dialect. Examples of -vll -/-vld- inGreek sources are extremely numerous (see examples in Appendix C),

16 In Adiego (1993a:44–45), ”evoro“kin’s proposal of connecting u≤ -/ Uss- to CLuw.

w à“ u- was accepted. Now I have serious doubts about this connection, insofar as itdoes not explain why the peculiar sibilant sound ≤ appears here. The explanation of Carian genitives in -≤ as arising from *-ºsi- (see p. 313) points rather to a *usi- stemas a starting point, although without discounting other possibilities. Could u≤ - < usi-to be related to CLuw. u““ a/i-, Lycian uhe/i- ‘year’?

and it is also well documented directly in Carian, particularly by thefamily of u≤ ol - Ussvllow names, which undoubtedly contain this suffix

even though the origin and meaning of the stem to which they areattached (u≤ º ) is by no means clear.16

For other possible suffixes recognizable in Carian place names (-ulia-,-um- ), and a tentative connection to Hittite-Luwian onomastics, seeNeumann (1988:389).

344



CHAPTER TEN

CARIAN AS AN INDO-EUROPEANANATOLIAN LANGUAGE

Despite the scarcity of linguistic information obtainable from the analy-sis of Carian inscriptions, several traits can be observed that clearlyplace Carian within the Indo-European Anatolian family of languages.

More precisely, some of these traits situate Carian in the group of so-called Luwic dialects, integrated by Cuneiform and Hieroglyphic Luwian,Lycian, Milyan, and probably also Sidetic and Pisidian, which share aseries of phonological and morphological features that diff erentiate themfrom other Anatolian dialects (Hittite, Palaic, Lydian). The presentchapter will be devoted to summarizing all of these traits found inCarian. For this task, it is essential to use evidence drawn from ono-mastics, in order to create a more complete picture of the Luwic char-

acter of Carian. I am aware of the risk involved in basing an argumenton the etymological interpretations of proper names, but I think thata significant number of these interpretations can be confidently usedto demonstrate the proximity of Carian to Luwian, Lycian and the restof the Luwic dialects. I refrain from off ering an exhaustive treatmentof this subject because I consider it to be more realistic, and also moreillustrative, to focus on a reduced, but very meaningful set of traits thatclearly establish the Anatolian—and particularly Luwic—character of Carian.

Beginning with phonology, a good indication that Carian belongs toAnatolian is the preservation of the PIE laryngeal *h2, a trait thatdiff erentiates Anatolian—with the exception of Lydian—from otherIndo-European languages. As shown in Chapter 6, in Carian (and alsoin Lycian and Milyan) this PIE laryngeal becomes a tectal stop (lenitedin some positions, at least in Lycian and Milyan), in contrast to Hittiteand Luwian, where it appears as a (velar?) fricative (Hitt. CLuw. ¢,HLuw. h ). The examples are taken from onomastic materials, but they

seem convincing: pikº/bikº in piks, dbiks, pikre (Pigrhw ), all from PIE *bh

è h2-vs. CLuw. pi ¢º (see p. 337), quq = Lyc. x uga , ‘grandfather’, vs. CLuw.¢ù¢a-.

Specifically Luwic is the satem treatment of the PIE palatal voice-less stop *∞ (CLuw., HLuw. z , Lyc., Myl. s , against k used elsewhere

346

in Anatolian). Evidence for this treatment in Carian (> s ) is providedby the demonstrative pronoun s(a)-: sa, san, snn, comparable to CLuw.,

HLuw. za- vs. Hitt. ka-. We can also consider as Luwic the loss of PIE* · in (i)b rº = CLuw. im(ma)ra/i- ‘open country’, against its preservationin Hitt. gim(ma)ra- (< PA * · emro- ).

In derivational morphology, a remarkable trait that once again sit-uates Carian in the Luwic subgroup is the presence of an ethnic suffix-yn-/- ÿn-, easily interpretable as the Carian counterpart of Luwian-wanni-, HLuw. -wani-, Lyc. -ñni , Mil. -wñni-.

In inflectional morphology we can also identify the ‘i-Mutation’, i.e.the insertion of -i- (probably from PIE *-ih2- ) before the ending (replac-ing the thematic vowel if this exists) in the nominative and accusativesingular and plural of the common gender nouns, a widespread phe-nomenon in the Luwic subgroup. It is true that this -i- does not appearas a result of the apparently defective vowel notation in Carian,1 butits eff ects can be detected in the umlaut displayed by the vowel of thepreceding syllable in the word ted , ‘father’, for which an original a canbe postulated, therefore *tadi- > *tedi- > ted- (cf . Lyd. taada- ‘father’ vs.Lyc. tedi- ‘id.’, also with umlaut). Another possible example of ‘i-Mutation’

can be considered for the Carian genitive -≤ , if we accept Melchert’svery plausible etymological explanation (from the possessive *-ass ì -, infact an ‘i-mutated’ form of the possessive *-asso- ).

Melchert’s explanation of -≤ also implies that Carian employed thissigmatic adjectival suffix to build nominal complements, which repre-sents another clearly Luwic feature of Carian: it is in these languages(Cuneiform and Hieroglyphic Luwian, Lycian) that we find this suffixused as a replacement for the inherited genitive endings, either in allcases or in the great majority, depending on the dialect.

Other Carian case endings are equally consistent with this theory.While the acc. sg. -n informs us merely of the IE character of Carian(-n < PIE *-m ), the common gender acc. pl. -“ , if really indicative of a *-ns origin (as the preservation of the sibilant seems to suggest, vs.its loss when it comes from *-Vs in the nom. sg.), allows us to see theproximity of Carian to Luwic: *-i-ns is the ac. pl. c. ending behindCLuw. -inz, HLuw. <-iza > (= -/inz/), Lyc. -is , Mil. -iz . The presenceof the Carian palatoalveolar (?) sibilant -“ , instead of simple dental -s ,

1 Regarding this problem, see above pp. 238–242.



- 347

suggests a secondary change of s , perhaps due to the contact with n,similar to the case of CLuw, HLuw, Mil. -z , irrespective of which sound

actually represents z in each of these dialects, a question that is thesubject of much debate.Verbal morphology cannot really be used as evidence of Carian as

an Indo-European Anatolian language, since the insufficient direct exam-ples noted in Chapter 8 ( ÿbt, ait , etc.) produce a clear risk of circu-larity: we interpret them as verbs mainly on the basis of their formalresemblance to Lycian (ubete, aite ), so it is dangerous to use them asevidence for the classification of Carian. As for indirect examples, thename Neterbimow, a Carian version of the Lycian Natrbbijemi (see abovep. 343), is a good, although isolated example that suggests the exis-tence in Carian of a participle suffix -m, comparable to CLuw. -mma/i-,Lycian -me/i-, itself another example of a Luwic trait.

Several words of the scarce vocabulary from Carian inscriptions whosemeaning can actually be confidently established also show clear paral-lels to the rest of the Anatolian languages: en ‘mother’ = Hitt. anna-,CLuw. ànna/i-, Lyc. ene/i-, Lyd. ena-; ted ‘father’ = CLuw. t àta/i-, HLuw.tata/i- Lyc. tede/i-, Lyd. taada- (vs. Hitt. atta- ); armo- ‘Moongod’ = Hitt.

arma-, Luw. * Arma- (see Melchert CLL, s. v. ), Lyc. ar ˜ mma-, cf. Lyd. armt a-‘belonging to Arma’. Specifically Luwic are mso- ‘god’ (?), cf. also ethnicmsnord- = CLuw. mà““ an(i)-, Lyc. maha(na)-, Mil. masa-, Sidetic ma ≤ ara (dat.pl.), and the form of the name of the Storm-god, Carian trq(u)d - =CLuw. Tar ¢unt-, HLuw. TONITRUS-hut- (= *Tarhu(n)t- ) Lyc. Trqqñt-vs. Hitt. Tar ¢u-. Finally, a meaningful isogloss shared by Carian andMilyan is the copulative conjunction Car. sb, Mil. sebe ‘and’ (cf . alsoLycian se ‘and’).

Therefore, although the number of Carian phonological, morpho-

logical and lexical traits that can be used for comparative purposes islimited, they are in my opinion significant and consistent enough toclassify Carian as an Indo-European Anatolian language, closely relatedto the so-called Luwic dialects (Cuneiform and Hieroglyphic Luwian,Lycian, and Sidetic). This comes as no surprise: even before decipher-ing Carian, the analysis of Carian proper names from indirect sourceshad already allowed some scholars to identify Carian as an Anatoliandialect, and to point out its affinities with Lycian and other southern

Anatolian dialects, but we can now, for the first time, actually confirmthis linguistic classification of Carian on the basis of direct testimoniesfrom the language.

CHAPTER ELEVEN

CARIAN GLOSSARY

A list of the words found in the Carian inscriptions is presented here,accompanied by a brief interpretation (whenever possible) and a shortbibliographical note. It should be taken into account that in the casesof inscriptions without interpunction, the entry is the result of my seg-

mentation, and must therefore be taken as hypothetical and viewedwith caution.

The following is the order of the letters (in transcription) adopted here:

a b b d d e g i/j k ∞ l l m n ñ ã o p q r ® s ≤ “ t t u/w y/ ÿ/ z H 1

Note that the letters for the pairs of vowel/semivowel sounds have beenput together. This allows us to clearly show the equivalences betweenforms such as w ≤ ol ≤ (Egypt) and u≤ ol ≤ (Caria proper, where a specific

letter for /w/ does not exist).I exclude from the glossary the sequences consisting of one or twoletters that appear in contexts that are very fragmentary, isolated andimpossible to interpret. The coin legends are also excluded (see theappendix by K. Konuk).

aba ? d ?

C.Ka 8

abrq ∞[ E.Ab 14

absims C.Ka 5Pronominal form?

In Adiego (1998a:25) absiº is tentatively connected to Lyc. ehbi < *ebesi ‘his’ (there-fore a ∞t [ ms ]kmt absiº = ‘his (for ‘their’?) descendents’), but this hypothesis leavesfinal ms unexplained.



349

ada C.Si 2a

Nom. sg. in the formula ]ryin ∞tmño≤ sb=ada ∞tmno≤ ‘Idrieus of Hekatomnosand Ada of Hekatomnos’.Fem. PN: Ada (the Carian queen, Hekatomnos’ daughter and Idrieus’

sister and wife): Ada in Greek sources, a typical Carian feminine name(cf. Aba, Alasta ) documented particularly in the regions of Mylasa andStratonikeia (Blümel KarPN:9).

Schürr (1992:138), Adiego (1994a: 40).

adymd “ C.Si 1Singular or plural nominative? Verbal form?

Cf. Adiego 2000:152. Final -“ could point to a nominative-accusative ending.Melchert (apud Hajnal (1995[97]:18, n. 15) suggests that it could be a verb ina reflexive construction (‘he made for himself ’): ad (‘made’, cf . Lyc. ade ) + md (particle) + “ (reflexive); y would be an anaptyctic vowel.

Cf . dymd a in Hyllarima?

ait C.Ka 2Possible verbal form: 3. plur. pret. ‘they have made’ = Lyc. aite (Car.ai- = Lyc. a(i) ‘to do, to make’?).

Melchert (1998:35). See also Adiego (1998a:25, 2000:140–141).

aitusi C.Ka 5Perhaps related to ait . A segmentation aitu could off er a good connec-tion with Anatolian imperatives (cf . Lycian tãtu ‘they must put’): there-fore aitu, ‘they must make’, ai- ‘to do, to make’ = Lyc. a(i) ‘id.’), butthe resulting final word si would remain unexplained.

Adiego (1998a:25).

ai [-]iqomE.AS 7The segmentation of the word is very doubtful.

350

Reading and segmentation according to Schürr. Ray (1994:203), starting fromthe older reading ps ÿ≤ ainiqom, proposes comparing ps ÿ≤ ainiº to the Egyptian PNP3-sb3-¢‘j-m-njw.t , Greek

Cousennhw, literally ‘the star arisen in Thebes’, although

he recognizes that no explanation can be given for the isolated three finalletters qom. Vittmann (2001:58) mentions Ray’s proposal without furthercommentaries.

akymyd u ÿeryly[vacat]d C.xx 3A complete inscription whose segmentation into words remains unclear.The meaning of the whole inscription is unknown.

On this inscription, see Schürr (2001c). Schürr points out that the abundanceof vowel signs present here is very unusual in Carian. He isolates a form aky-mud u as a possible 3rd pl. imperative, with a < */-ndu/ ending comparableto Hittite and Luwian corresponding endings. For akymy-, he proposes a con-nection with Hitt. ak(k)- ‘to die’. All these proposals are formulated within theframework of a very speculative interpretation of the possible content of thetext. In any case, a segmentation akymyd u ÿeryl [ is likely, given the unusualsequence ºu ÿeº.

The syllabic iteration y . . . y recalls the similar situation in ardybyr ≤ (as well

as dt ÿbr, k ≤ at ÿbr, smd ÿbrs ), so that the first y in these cases could in fact be sec-ondary (epenthetic?): akymyd uº < *akmyd u-.

Could this inscription actually be a sort of alphabet model (with the namesof letters: a-ky-my-d u, etc.)?

a ∞akowr E.AS 4a ∞mnnartnyr C.Ka 2

Van den Hout (1999) claims to recognize here a clitic chain a=∞=m=n, to becompared with Lycian parallel forms. Cf . his analysis of a ∞t [ ms ]kmt .

a ∞t [ ms ]kmt C.Ka 5a ∞tmsk [ m- ]d C.Ka 2

The respective integrations are dependent on each other, but given thetextual connections between C.Ka 2 and C.Ka 5, they seem to be wellfounded. Some doubts persist, however, regarding the final letters of



351

each word. Perhaps we are dealing with two diff erently inflected formsof the same stem.

In C.Ka 5, the corresponding Greek text makes the meaning “descen-dants” a plausible interpretation.

See Adiego (1998a:24) and here on p. 300 for a very speculative attempt atan etymological explanation (a ∞t- = Hitt. katta, mskm-, to be related to Luw.ma “¢a ¢it- ‘growth, prosperity’, the overall sense of the word being ‘off spring,

Nachkommenschaft ’).A radically diff erent approach is taken in van den Hout (1999): he suggests

analysing this form as a clitic-sequence, in which he underlines =ms =, inter-preted as a pl. dat. of a 3rd sing. personal pronoun. As for the rest of theelements (segmented as a=∞=t=ms=km ), he suggests some possible Lydianparallels.

a ∞t [ C.Ka 5It could be the same word as in the two preceding entries (a ∞t [ mskm . . .]).

alos

E.Me 45alos d C.xx 2Attested in both cases in agreement with the word ∞arnos: alos ∞arnos,alos d ∞arnos d . Tentatively connected with the Carian place name Halikar-nassos (ÑAlikarnassÒw ), but this raises serious formal difficulties.

Halikarnassos-identification already included in Adiego (1990:135–136). Withmore reservations: Adiego (1993a:245–246), (1994a:40).

The duplicate ending -d points clearly to two diff erent nouns (substantive+ adjective or vice versa). In Adiego (2000:154) alos-d ∞arnos-d is interpretedas an ablative singular (Carian -d = Luw. -ti ) ‘from Halikarnassos’; alos ∞arnos in E.Me would be merely the name of the city in nominative (see here pp. 279,315).

Other proposals: a salutation or wish formula (Gusmani 1979a:222), (1986:62);Schürr (1992:153) interprets -s as a dative ending, which makes an analysisof these forms as place names difficult.

amt [

C.Tr 1



352

anC.Tr 2

annC.Ka 3These forms appear in two funerary inscriptions. In C.Tr 2, an is accom-panied by si d i , a typical word found in funerary contexts. In C.Ka 3,it appears preceded and followed by two personal names in genitive(“ orus and i b rs ≤ ). The simplest interpretation is to consider an/ann ademonstrative that functions as adjective modifying si d i in C.Tr 2 (‘thistomb(?)’), and as substantive governing the personal names in C.Ka 3(‘this of ”oru (son of ) Ibrs’).

For this interpretation, see Adiego (1996:161) and here pp. 290, 320. Also,Hajnal sees here a demonstrative (Hajnal (1995[97]:20, from */eno-/). Adiego(loc. cit. ) suggests other possible forms of the pronoun in ank b u“ and añmsñsi .

Schürr (1996c:158) proposes that ann C.Ka 3 designates the tomb, but givenan si d i of C.Tr 2, the interpretation as demonstrative seems more suitable.

ank b u“ E.Bu 1, E.Bu 2

Perhaps a title or a kinship term (in nominative), given the contexts inwhich it appears.

See Schürr’s etymological attempt (Schürr 1996[98]:97–98] to connect ºk b u“ with the kinship terms kombow, kombion ‘grandson’, attested in several late Greekinscriptions from Caria, and related to Hitt. kappi- ‘little’ by Neumann (1961:61).anº would then be a sort of prefix modifying in some way the kinship term(cf . epñ-neni in Lycian, from n eni ‘brother’).

an[ C.Ka 2

añmsñsi C.Si 2aPerhaps to be segmented into añ msñsi . In such a case, a comparisonwith Luwian anni “ ma “ sana ““ i “ , Lycian [ e ]ni mahanahi ‘mother of the Gods’would be striking. In any case, the reading, based only on the Robert-Deroy edition of C.Si 2, is not absolutely certain.

Connection of msñsi to Luwian ma ““ ana ““ a/i-, Lycian maha(na)- suggested invan den Hout (1999:39).



353

aor ≤ E.Me 1 (aor[ ≤ ]), E.Bu 6

PN in genitive, Carian adaptation of an Egyptian name ( J-Ó r literally‘O(?) Horus’, [a˙òr], Greek Avw (?), see DNb:55).

Vittmann (2001:42).

apmenE.Me 44aPN in nominative, Carian adaptation of the Egyptian name Ó p-mn (lit-erally ‘Apis is perpetual’, *[ apimèn], Greek ÑApimenhw, see DNb:781).

ap[---]ws E.Me 23PN. It is not clear if it is a nominative of a s-stem, or a ‘s -case’.

a ? q ≤ baq E.Th 10

a [ rb ]ikarm≤ E.Me 23PN in genitive. If the form is correctly completed, it gives a Carianname corresponding to Lycian (in Greek sources) Arpigramow (ZgustaKPN §104–1). The name would therefore be a compound ar-bikarm-.For the second element, cf . pikarm-/pikrm-.

Completed and connected to Arpigramow by Kammerzell (1993:214), this viewis commonly accepted: see for instance Schürr (1992:139), Adiego (1994a:31)

(with further remarks on the structure of the name).

ardybyr ≤ E.Me 52PN in genitive. Corresponding to the Carian name in Greek adaptationArduberow. A name belonging to the family of names in -(d ) ybr-.

Connection to Arduberow already suggested by Ray (1982b:189), but seriouslyhampered by his decipherment system (†argébér ≤ ). See Adiego (1993a:225–226),

(1994a:40).The doubts about the exact reading raised in Schürr (2001c:119)—who sug-

gests an alternative interpretation a | rdybyr ≤ )—are not particularly convinc-ing, and in any case do not alter this correct identification.

On this family of names, see Adiego (1993a:224–227).



354

are “ C.Ia 3

Probably a PN in singular nominative (but a plural nominative cannotbe ruled out). It is possible that the beginning of the word is incomplete.

arie ? ≤ E.Me 38PN in genitive.

arjom≤ E.Me 42PN in genitive. The connection with the PN arliom- is unclear: is itperhaps a variant, parallel to ar´ ri “ , ari “ (Arrissiw ), as well as arli “ (Arlissiw ). Independently of this possible connection, compare also arj-with the Carian PNs Ariauow or Aridvliw.

Schürr (1992:134) suggests attributing the alternation arjom-/arliom to a dialec-tal variation.

ari “ C.St 1ari “≤ C.Hy 1aPN in nominative (ari “ ) and genitive (ari “≤ ). Correspondence to theGreek adaptation Arrissiw is confirmed by the evidence of the bilingualinscription C.Hy 1, where the name Arrissiw appears in the Greekpart (although the individuals mentioned are not necessarily the same).Given this identification, ari “ could be a variant spelling of ar´ ri “ , q.v.

Both ari “ -Arrissiw and ar ri “ could be variants of arli “ -Arlissiw, as arjomagainst arliom. See arjom for further remarks.

ar ∞ila ≤ E.Me 39PN in genitive. Perhaps a Carian adaptation of the Greek PN ÉArx°laow(Dor. ÉArx°law ). For the use of ∞ for a Greek velar stop, cf. urs ∞le-.

Greek origin already suggested by Ray (1994:202)]. For an interpretation asan purely Anatolian name, see Adiego (1993a:240).



355

arliom≤ E.Me 9 (arlio[m≤ ]), E.Me 43b

PN in genitive. It is the Carian name that appears as Arlivmow inGreek sources.

For the identification, see Adiego (1990:134; 1993a:231; 1994a:35).

arli “ E.Ab 24arli “≤ E.Me 9, E.Me 15, E.Me 51PN in nominative (arli “ ) and genitive (arli “≤ ), corresponding to Arlissiwin Greek sources. Cf . moreover the Carian place name Arlissow. Thestem seems to be the same as arliom-. Note also the possible variantsari “ , arri “ .

Connection to Arlissiw already stated in Faucounau (1984:236). Cf. Adiego(1993a:230; 1994a, n. 3.3).

armonE.Me 8a, C.Eu 2In E.Me 8a, noun in nominative: ‘dragoman, interpreter’, corresponding to Egyptian p3 w m ‘dragoman, interpreter’ in this bilingual inscription.

In C.Eu 2: function and meaning unknown (it would be very unlikelyfor it to have the same meaning as in the other example).

For a detailed discussion of the problems posed by the exact meaning of Egyptian p3 w m, see Vittmann (2001:50–52), who argues convincingly for the

sense ‘interpreter’, and dispels all the doubts raised in Masson-Yoyotte (1956)and subsequent literature about this interpretation. Janda (1994:180–182), starting from the alleged meaning ‘herald’, proposes

for armon- an origin from *ar (V)ma-wanni-, where *ar (V)ma- would correspondto HLuw. ataman-/adaman-/ ‘name’, assuming a semantic change, ‘name’ >‘determination, decision’. The meaning ‘dragoman’ corresponds even better tothe semantics of this explanation (perhaps ‘interpreter’: ‘who names the thingsin another language’) but the sound changes assumed are ad hoc.

For armon in C.Eu 2, see above p. 309.

armotrq d os C.Hy 1aMost probably a dvandva-like compound formed by the divine namesarmo- (Carian version of the Anatolian moon-god Arma- ) and trq d - (the

356

Anatolian Storm-God, Hitt. Tarhu-, CLuw. Tarhunt-, Lyc. Trqqñt-, seebelow, trqud e ). A (possessive?) ending -os (= Luw. -a ““ a-?) has been added

to the compound stem armo+trq d -.

On this form, and for a diff erent possible analysis, see Adiego-Debord-Varinlio< lu(2005:616–617). The possibility cannot be discounted that a secondary vocalicstem (trq d o- ) may have been formed from the original nt-stem (as in the restof Luwic languages) trqud - (<*tarHunt- ), so that the ending would be -s (like inntro/ntro-s ).

ar ® i “

E.Bu 1(ar[ ® ]i“ ), E.Bu 2PN in nominative. The identification with the Carian name from Greeksources, Arris(s)iw, depends on the exact value of the letter transcribedas ® , but this is now reinforced by the certain equivalence ari “ (q.v. )= Arrissiw in Hyllarima (where there is no letter ® ). The connectionof all of these forms to arli “ -Arlissiw is unclear (cf . also arjom- andarliom- ).

Adiego (1992a:34). See also Adiego (1994a:35).

artay≤ E.Me 22PN in genitive, to be compared with Artaow, although the Greek adap-tation as an o-stem could mean that the stem is not exactly the same.Cf . also Arthumow (Zgusta KPN § 109, Blümel KarPN:11), which canbe analyzed as a stem, *artay-/artey-, followed by a suffix, -m-. The finalpart of Arthumow recalls names such as paraeym- parpeym-. For artº/Art-

cf . also Artuassiw, Aryuassiw.Adiego (1993a:231–232; 1994a:35).

artmi C.Tr 2PN in nominative(?). It could be the Greek goddess name ÖArtemiw,directly used as personal name (cf . Lydian artimu≤ , also attested both aspersonal and god names), or a variant form of the Carian name Artimhw.

The problem posed by the relationship between the Greek divine name (of Asian origin?) Artemis, and the family of Anatolian names collected in ZgustaKPN § 108 (Arteimaw, Arteimow, Artimhw, etc.), is not at all clear, and cannot

357

be dealt with here (see remarks in Zgusta KPN:102 and Adiego-Debord-Varinlio< lu 2005:611). Cf . also the Carian name rtim.

The ambiguous structure of C.Tr 2 does not mean we can dismiss the the-ory that artmi could be a dative, see above p. 290.

artmonC.Tr 1(art mon)PN in nominative. It is the Carian adaptation of the Greek nameÉArt°mvn. Note the identical adaptation in Sidetic (artmon ).

The Greek PN ÉArt°mvn is well documented in Caria. For instance,it appears in the well-known inscription of Halikarnassos, SGDI 5727,

where it is used by individuals whose father’s name is clearly Carian(ÉArt°mvnow toË Panamuv, ÉArt°m[v]na Territò, etc.).

For the identification, see Schürr apud Adiego (1994, 3.49).

a ≤ b≤ t E.Th 13

a [ - ]mob[ C.Al 1

a [--]a [----]om≤ C.St 1The final part clearly indicates a PN in genitive (for the ending, cf .names such as arliom-, kbjom-, etc.

bal

E. Th 49PN?

baq g k [ . . . ]C.Ia 5

bebi E.Th 23It seems to be an incomplete form of bebint , see the following entry.



358

bebint E.Th 28 (bebi.nt), E.Th 30, E.Si 4 (be?bint), E.AS 7

Word of unknown function and meaning. In E.Th 28 and E.Si 4 it appearspreceded by a diamond-like sign ( t, K ) whose function is uncertain.

Former readings of some of these testimonies raised doubts about the secondletter of the word (w ÿ rather than e e ), but a unified reading, bebint , mustnow be preferred (see Schürr 2001b:108).

Schürr has argued in diff erent works in favour of an interpretation as averb (a third singular preterite form, see Schürr 1996a:65) with the meaning ‘to off er’ (Schürr 2001b:108). It is also seen as a verb by Hajnal (1995[97]:18):a 3rd sg. present ‘he sends’ of the verbal root that he also recognizes also inbinq (q.v. ).

bebnd E.Th 6Perhaps related to the preceding entry (bebint ).

bejeymE.Th 28

Perhaps a PN in nominative. As for the final part of the word, cf . thePNs paraeym, parpeym.

be ≤ ol E.Ab 23PN in nominative. Although no directly comparable form is attestedin Greek sources, it recalls other Carian names with -ol (Greek -vllow,-vldow ).

be t krqit [ - - . . . ]E.Si 5

bi C.Ka 2 (2× )C.Ka 5Apparently a conjunction or particle, as it appears in C.Ka 2 follow-ing an accusative otr “ (‘themselves’). In this same passage, it seems to

be used as a postclitic in correlation: . . . bi . . . bi . Its identification inC.Ka 5 is less certain. Cf . also another example ]bi in C.Ka 2.

On the possible diff erent interpretations of bi , see p. 302.

359

bi d ≤ lemsa E.Ab 30

Function and meaning unknown. Not a PN, because it precedes anonomastic formula in nominative. Cf . perhaps [--]msal (E.Bu 1), whichalso precedes a PN in nominative.

Hajnal (1995[97]:23) proposes the segmentation bi d ≤ lem sa , where sa would bea demonstrative (see sa ). For the first word, he suggests that ºem indicates aparticiple.

bij ≤≤ pe

E. Si 8Probably a PN in nominative.

binq C.Si 1A verb? The reading of the last letter is not absolutely certain.

Hajnal (1995[97]:18): ‘ich schenkte’, 1st sg. pret. < */ pí äiannà-¢a /, to be com-pared to Hitt, piyannài- ‘to send’. For the same verb also ºbint , → bebint as 3rdsg. present.

bsis E.Ab 30

b? s ? ui ∞amE.Lu 5

bu∞ y[-----]i [-----]i C.Ka 5

b ÿ“ E.AS 7A word that apparently agrees with the following: (esak ? d ow “ ). A demon-strative pronoun in accusative or nominative plural?

The idea of a demonstrative pronoun has been repeatedly defended by Schürr

(see Schürr 1996a:69, 2001b:98, 112): b- would come from PA *obó/i-( >CLuw. apa-, Lyc. ebe- ) with aphaeresis (as in HLuw. pa- besides apa-, or Lyd.bi- [pi-]). The ending -“ can be compared with -“ , the acc. pl. animate end-ing in C.Ka 5.



360

b ÿta “ E.Si 6

PN in nominative?

b anol C.My 1 (2× )PN in nominative. Perhaps the Carian name corresponding to Ibanvlliwin its Greek adaptation.

For this possible connection with Ibanvlliw, see Blümel-Kızıl (2004:134), Adiego(2005:85).

b em≤ E.Me 17PN in genitive.

b 2o[--]ol “ C.Ka 5Noun in plural accusative. It appears in a long sequence that corre-

sponds to the Greek formula proj°nouw e[ ‰nai k ]a‹ eÈerg°taw Kaun¤v[ n ].Frei-Marek (1997:38–39) try to bring this word closer to 1orsol “ (completing therefore [ rs ]). However, this solution implies that the initial letter of the word,ÿ (here transcribed by b 2 ), has to be likened to 1, contradicting the theorydefended in the present book, that ÿ is a letter for b (see also Adiego 1998a:23).

As for the meaning, it hardly can be ‘descendents’, as Frei-Marek suggests,given the clear correspondence otr “ = aÈtoÊw. b 2o[--]ol “ must rather be a wordat the same level as kbdyn“ (in kbdyn“ sb b 2o[--]ol “ ). In Adiego a connection of b 2o[--]ol “ with Imbros, the citadel near Kaunos, is very tentatively suggested.

See pp. 299–300.

b rsi E.Th 26, E.Th 48, C.Hy 1ab rsi ≤ C.Hy 1a (2× ), C.St 1PN in nominative (b rsi ) and genitive (b rsi ≤ ). Carian name rendered inGreek as Imbarsiw, Imbras(s)iw. Form with aphaeresis or no notationof initial vowel that coexists with the full forms i b rsi-, i b arsi-. Identificationnow assured thanks to the bilingual inscription C.Hy 1, where bothb rsi and Imbrasiw appear (although not necessarily referring to the sameindividuals).

361

Schürr (1991[93], passim ). For b rsi — Imbrasiw in Hyllarima, see Adiego-Debord-Varinlio< lu (2005).

dbikrmE.Th 19PN in nominative. Can be analysed as (i)d- + bikrm. For the first ele-ment, cf. d-quq- (Greek Idagugow ), d-w ≤ ol /id-u≤ ol - (Greek Idussvllow ),d-biks . For the second element, cf . pikrm/ pikarm (cf . also Pigramiw, Pigramowin Lycia, Zgusta KPN § 1255–1, 2).

Note below the ‘devocalised’ form dbkrm.

dbiks E.Th 13PN in nominative. A compound name formed by d- (see dbikrm above)and biks (see piks [ ).

dbkrmE.Ab 34PN in nominative. A variant form of dbikrm.

dmo“ bqs E.Si 11Perhaps a PN.

See the study dedicated to this word by Schürr (1996b). There, ºbqs is con-nected with piks-/biks- and other forms that are derived from PIE *bhè hos (see

piks- ).

dm-?-nE.Th 34It could be a noun in accusative singular, given the context ( prna ∞nondm-?-n ).

dokmmpint E.Th 4

dquq

E.Th 44PN in nominative. It corresponds to Idagugow in Greek sources, a com-pound name (i)d- (see dbikrm- ) + quq- (see quq- ).



362

For this identification, see Adiego (1993a:235; 1994a:35–36). Tentatively com-pared in Adiego (1995:27, n. 9) with Milyan dd x ug [, but the segmentation andthe interpretation of this latter form are doubtful (see Schürr 1996b:154).

dr “≤ iemE.Th 53PN in nominative?

Initial dr “ º recalls d ar “ º in d ar “ qemorms [. Note that in Thebes d is practicallyabsent, and its function could be assumed here by d .

∂ saml-?-?-$oE.Th 16

dt ÿbr E.Th 2PN in nominative. It belongs to the family of names in -(d)ybr- (andvariants). Cf . ardybyr-, dybr , etc.

On this family of names, see Adiego (1993a:224–227)

dw ≤ ol ≤ E.Me 35PN in genitive. This is a variant with aphaeresis (or no notation of theinitial vowel) of the name now attested in its complete form in Mylasa(idu≤ ol- ), which corresponds to Idussvllow. A compound name (i)d- (seedbikrm- ) + u≤ ol ≤ (cf . u≤ ol -/ Ussvllow ).

For this identification, see Adiego (1993a:241). Aphaeresis suggested by Neumannapud Adiego ibid .).Schürr (1996b:154) recalls, in addition to Idussvllow, the Carian name

Yussvllow.

d ÿbr E.Th 5PN in nominative. A name belonging to the family of names in -(d)ybr-.

On this family of names, see Adiego (1993a:224–227)

363

dymd a C.Hy 1a

Meaning and function unknown. Cf . in any case adymd “ in C.Si 1.

d ar “ qemorms [ C.St 2Part of an onomastic formula, like those that precede it in the sameinscription?

Schürr (2001c:119) proposes the segmentation d ar “ qem orm s [ and interpret-ing “ qem as a participle. For d ar “ , if a PN, compare with Andarsvw and prob-

ably also Dersvw (see p. 246).

d enE.Sa 1

Analysed as a sort of preposition (comparable to Hitt. andan ) governing tumn(= ‘for Atum’) in Adiego (1995:21–23). See here p. 287.

d rual C.Ka 5Unclear form. It must correspond in some way to Greek §p‹ dhmio[u]rgoËin the bilingual inscription C.Ka 5, but the precise analysis remainsunclear.

An attractive theory has been formulated independently by Hajnal (1997b:150)and Melchert (1998:37; cf . also 2002:308) that the entire sequence i [--]inis d rual can be analysed as ‘under the ship of Hi[pposth]enes’, with d < *nde =Lyc. ñte ‘in(to)’, i [—]inis genitive/possessive, and rual as a noun correspond-ing to the title dhmio[u]rgoË. Attempts to find an etymological explanation forrual (Hajnal 1997b:151, from */erowà-/ ‘freeedom’, cf . Lyc. arawa-; Schürr2001b:109–110, a similar explanation, but starting from a diff erent analysisof the sequence) seem somewhat rash.

ems g lpnE.AS 4



364

enE.Me 32

‘mother’ (nominative). It corresponds to Lycian eni , Lydian ena ≤ , CLuw.anni “ ‘mother’. Carian vocalism points to an ‘i-mutation’ stem (*ani- )and subsequent umlaut a > e caused by this i-suffix.

For this meaning, see Schürr (1996a:62), Hajnal (1995[97]:21–22). For i-muta-tion and umlaut, see Schürr (2001b:97) and here on p. 259.

eri C.Si 2a

Function and meaning unknown.

In Adiego (2000:143–144) it is interpreted as a noun with the meaning ‘taxexemption, ét°leia’, functioning as direct object of pisñoi md a (q.v. ). Moreover,a connection with Lycian arawa- ‘tax exemption’ is suggested. Neumann (apud Adiego 2000:144) envisages the possibility that eri could be a preverb identi-cal to Lycian eri . All of these hypotheses are dependent on a particular inter-pretation of C.Si 2 traced in Adiego (2000), cf . here on p. 304.

esak ?

d ow “ E.AS 7Very likely to be a compound noun that seems to contain the stemk d ow- (cf . k d ou≤ in E.Bu 1), usually interpreted as the Carian word for‘king’ (cf . Lycian x ñtawat (i )- ‘king’, and the form KNDWÍ/KNDWS inthe Aramaic part of the Trilingual inscription of the Letoon of Xanthos,which is perhaps a direct reflection of the Carian word).

See Adiego (1995:18–21) for a detailed discussion of this question. The most

compelling point of this interpretation is undoubtedly the coexistence in E.AS7 of esak ? d ow “ and pisma ≤ k (= Psammetichus)—separated by the word m ÿqudem— that opens the possibility of interpreting them as ‘the king . . . Psammetichus’,and of linking this inscription to the well-known long Greek graffito, also fromAbu-Simbel, in which this Pharaoh is also mentioned. But decisive formaldetails of esak ? d ow “ remain unexplained: the value of initial esa-, and the precisefunction of -“ , which seems to agree with the word preceding esak ? d ow “ , b ÿ“ and which recalls the ending of plural accusative in C.Ka 5, which would makethe interpretation ‘king Psammetichus’ very difficult to argue (see here onp. 294).



365

ewlane E.Th 49

See following entry and ewmlane .

ewmE.Th 10, E.Th 13A typical sequence in Theban graffiti: Cf . sl ∞maewm, which can perhapsbe segmented into sl ∞ma ewm, the incomplete form ewm-?-?-?-? , and thesequence ewmlane . Note also ew in the preceding entry ewlane .

See p. 323 for a possible segmentation into two elements ew+m.

ewmlane E.Th 12, E.Th 44One of the alleged ‘verbs’ in mlane, md ane, q.v. Note that the segmen-tation is not clear (ew+mlane, ewm+lane or ew+m+lane? ), see p. 323.

euml ? bnasal E.Bu 2

Analysis, function and meaning are unknown. Cf . eypsal also in Buhen,which would allow us to isolate an element sal .

Could be sal and adverb with the meaning ‘here’ (cf ., for the ending, Lyc.ebeli , and for the stem, the pronoun sa-/sn- ).

ewm-?-?-?-? E.Th 52See ewm entry.

eypsal E.Bu 6See euml ? bnasal above.

g kem≤ E.Th 44PN in genitive

g rdso[-]i [ C.Ka 2Probably related to qrds, qrdsol , especially since it appears in a sequenceqrds g rdso[ - ]i [ that could constitute a type of fi gura etymologica .

366

gd b“ l a ã1i [-]C.Ka 2

i [---]inis C.Ka 5PN, probably with s-ending. This incomplete name must be the Carianadaptation of the Greek Hipposthenes, the name of the demiurge inthe bilingual inscription C.Ka 5. Tentatively completed as i [ poz ]inis (Frei-Marek 1997:31).

For this interpretation, see d rual . However, an alternative segmentation i [---]ini s d rual (thus Schürr) cannot be discounted. As for the missing letters, Frei-Marek’s solution is quite good, but not the only possibility (the use of <z>for Greek sy is impossible to demonstrate).

ial l i E.Ab 40PN in nominative.

iarja ≤ E.Ab 2PN in genitive.

iasoumC.Ki 1

See here p. 142 on the doubts about the exact reading and a very hypo-thetical alternative reading.

i b arsi ≤ E.Ab 3i b rsi ≤ E.Bu 4PNs in genitive. Alternate forms of a name corresponding to Imbras(s)iw/Imbarsiw in the Greek source. For further remarks, see the variant formb rsi-.

i b rs d r [ - ]C.Ka 4The initial sequence is undoubtedly related to the i b arsi-/i b rsi-/b rsi-family of words, but both the precise analysis and the segmentation

367

into elements are not clear. Perhaps we must segment i b r-s d r [ andidentify here a construction similar to i [---]ni-s d rual (‘under the demi-

urge Hipposthenes’, according to the Greek translation) in C.Ka 5.

Connection to i b (a)rsi -family already noted in Schürr (1992:140). Schürr alsorecalls Imbrow, the name of a citadel near Kaunos (Zgusta KON § 373–1,Blümel KarON:168), and whilst in former works he suggested that i b rsº couldbe the ethnic name of Imbros, following the discovery of the Kaunos bilin-gual he prefers to take i b r º as the actual name of the citadel, and to inter-pret i b r-s d r [ u ] as ‘(in?) Imbros, ‘the people’ (with *s d ru- ‘people’ as the basisfor s d rual ‘demiurgos’). However, s is most likely to be an ending, both hereand in C.Ka 5.

i b rs ≤ C.Ka 3PN in genitive. The same name as i b (a)rsi-, or a name closely relatedto it.

The doubts about the analysis lie in the absence of i at the end of the stem.It could simply be the result of a defective notation, or perhaps the reflectionof a diff erent suffixation (for instance -s- < *-so- against -si- < *-siyo- ).

idmns E.Me 33aE.Me 33bPN. Case unclear (nominative of an s-stem, or rather a stem idmn- witha ‘dative’ -s ending?). It seems to be a compound whose first elementis (i)d-, cf, idu≤ ol-/dw ≤ ol -/Idussvllow, etc.

idmuon≤ E.Me 18bPN in genitive. No parallel forms in Greek sources. It could be ana-lyzed as a compound name: id + muon-. For the first element, see thepreceding entry (idmns ), whilst -muon- could belong to the family of muwa- names.

Janda (1994:176) suggests interpreting this name as a derivative of the Carianplace name Iduma by means of the ethnic suffix Luw. -wanni-: ‘aus Iduma stam-

mend’.



368

idrayridsemd bq C.My 1

Heading of the long inscription of Mylasa, followed by the words mol “ tu∞[, and a list of onomastic formulae. This sequence, undoubtedly con-stituted by more than one word, remains impossible to analyse.

See here on p. 308 for the sequence idrayriº.

idu≤ ol ≤ C.My 1PN in genitive. Carian name rendered in Greek as Idussvllow. For itsanalysis, see the variant form dw ≤ ol -.

Blümel-Kızıl (2004:137).

idyes ≤ E.Me 63aPN in genitive. Initial id º recalls the lexical element (i)d-/Id-, see dw ≤ ol -.

idym“ C.Ka 4The final -“ could point to a plural accusative (and also nominative?).

A stem idym- recalls the Carian place name Iduma.

idyri ∞≤ C.Eu 1C.My 1PN in genitive. A clear compound name id+yri ∞-, still not attested inGreek sources. For the first element, cf . (i)d-/Id- in dquq, idu≤ ol , etc. Thesecond element is the well-known stem yri ∞-/yriq- / ÿdiq-/ ÿd ∞- (Greek-urigow, -udigow ).

Cf . Blümel-Kızıl (2004:137), who reconstruct a possible Greek adaptation*Idurigow. For a unifying explanation of yri ∞-/yriq- / ÿdiq-/ ÿd ∞-, see pp. 262–263.

inut ≤ E.Ab 18PN in genitive.

369

ione l ≤ E.xx 3

PN in genitive. It could contain the same stem as ‘Ionia’, ‘Ionian’.

For this connection, see Schürr (1991[93]:173), Adiego (apud Schürr ibid. ). InAdiego (1994:49, n.15) a detailed account of the possible process is given:starting from *iona- < *iyauna- < *iyawana- ‘Ionian’ (cf . Lyc. Ijãna- ), *-wana-would be the ethnic suffix Luw. -wanni-, Lyc. -ñni , Carian -yn/-on.

irasa E.Si 3

PN in nominative? The context is very unclear.

irow E.Me 14, E.Me 16irow ≤ E.Me 27PN in nominative (irow ) and genitive (irow ≤ ). Most likely to be a femi-nine name in E.Me 16 and 27 (see pp. 272–273). As according toVittmann, a name of Egyptian origin: J.r=w (attested both as a mas-culine and feminine name), phonetically [iròw] or [ j6ròw].

Vittmann (2001:45). Egyptian origin already suggested in Ray (1994: 202). TheEgyptian interpretation for irow now seems preferable to former attempts toanalyse it as Anatolian (cf. Ray 1982b:184, Melchert apud Adiego 1995:23,Hajnal 1995[97]:27, n. 38), especially given the difficulties raised by the diff erentproposals of this type (see Adiego 1993a:247; 1995:23–24).

isor ≤

E.xx 1PN in genitive, of Egyptian origin: it is the adaptation of Ns-˙r (liter-ally ‘(s)he belongs to Horus’, phonetically reconstructed *[ 6s˙òr]; Greekperhaps ÉEsour, ÉEsouriw, etc., DemNb:685).

Vittmann (2001:50). Former attempts to connect it to the Anatolian place nameIsaura (Adiego 1993a:247) must be ruled out.



370

iturow ≤ E.Me 32

PN in genitive, whose feminine character is clear from the context (theword en ‘mother’ refers to it). Carian adaptation of the Egyptian (bothmasc. and fem.) PN Jr.t=w-r.r=w (DemNb:70) *[ j6turòw], Greek ÉIyorvw.

Ray (1994:202). See Vittmann (2001:45) for the phonetically reconstructed form.

i ÿkr ≤ E.AS 5PN in genitive

jzpe C.xx 2PN? It immediately precedes md ane in the inscription.

In Adiego (2000:154) it is analysed as a PN in dative. Schürr (1996a:65), takes Jzpe to be the name of the donor of the object (presumably in nominative).Hajnal (1997b:150) interprets the entire sequence alos d ∞arnos d jzpe as a dat-ing formula ‘zur Zeit der ?-schaft von Jzpe ’, but he does not explain the pre-cise morpho-syntactical status of jzpe .

All the attempts to find explanations for a possible PN jzpe are somewhatspeculative: Schürr (1996a:65, n. 14) compares jzpe with the Lydian PN i ≤ tube l m-, and Hajnal (1997b:150, n. 14) recalls the Persan PN Vi “ t àspa-/ÑUstãsphw.

ka t t ÿri ≤ E.Ab 25PN in genitive

kbdmu≤ C.My 1PN in genitive. Perhaps a compound name kbd+mu-. The first elementclearly recalls the place name kbid- ‘Kaunos’ (for the omission of thevowel, cf . particularly the form of the ethnic name kbd-yn-“ ), althoughit is not clear if the PN alluded directly to the place name or if it con-tained the common noun from which the place name was created. Asfor the second element, it seems to be the well-known Anatolian stem

muwa-, ‘strength, force’.

371

For the connection of kbd (+mu )- to kbid - ‘Kaunos’, see Blümel-Kızıl (2004:136),who also recall the Carian name from Greek sources Kbvdhw. For a possibleAnatolian etymology of all these names in kbº, see p. 334.

kbdyn≤ C.Ka 5‘Kaunians’ (in plural accusative). Meaning assured by the Greek partof the bilingual C.Ka 5. An ethnic noun derived from the place namekbid- ‘Kaunos’ (q.v. ) by means of a suffix -yn- that corresponds to Lycian-ñni-, Milyan -wñni-, CLuw. -wanni-.

Identification as ethnic name and comparison of -yn- with the Luwic suffixalready in Frei-Marek (1997:37, 50).

kbidnC.Ka 5Carian name of the city of Kaunos (cf . the Lycian form x bide ), or, moreimprobably, an ethnic name derived from it (cf . kbdyn“ ). Morphologicalanalysis is unclear.

Cf . the Lycian name of this city, Xbide , and the Aramaic adaptationof the god name ‘Kaunian king’, KDWÍ/KDWS KBYD”Y.

For the reading of the last letter of the word, see Frei-Marek (1998:2). It ispossible that this city name was in Carian a plurale tantum (as is presumablyLycian x bide ), according to Hajnal (1997b:149), see also Melchert (2001:310,n. 12). Both Hajnal and Melchert (loc. cit.) point to a genitive plural (-n <*-ò m ), but while Hajnal imagines a true genitival value (‘decree of Kaunos’),Melchert suggests that this genitive plural could become a dative-locative, likein Lydian (therefore kbidn ‘in Kaunos’).

For the possibility that kbidn is an ethnic name (in nominative plural, from*kbid -wen-is ), see Adiego (1998a:20), (2002:19–20). For a possible etymology of the names in kbº, see p. 334.

kbjom≤ E.Me 12, E.Me 32, E.Th 13PN in genitive. Carian name adapted in Greek as Kebivmow. kbjom- alsoappears in the compound name “ arkbiom.

Adiego (1993a:232). Hajnal (1995[97]) suggests that this name would containa participle < */ pi ì emmo/ì /- ‘given’, with an alleged change *^ m > om. For apossible etymology of the names in kbº, see p. 334.



372

kblow ≤ E.Th 46

PN in genitive.

kbokt ≤ E.Th 2PN in genitive.

See Adiego (1994b:252), where the name is analysed as a compound formedby kbo- (cf . kbo-s , perhaps also taqbo-s ) and kt (cf. ktmno ).

kbos E.Me 24Ethnic name? (‘Keramean’?). It could be formed from a place namekbo- (‘Keramos’? see Konuk 2000b) by means of an -s- suffix (cf . otono-s-n from otono- ‘Athens’).

For this interpretation, based on the peculiar structure of E.Me 24, see abovep. 278. The interpretation was already suggested by Schürr (2003:116, n. 1).For a possible etymology of the names in kbº, see p. 334.

k d ou≤ E.Bu 1Noun in genitive.

Could it be the Carian word for ‘king’? See esa ? k d ow “ .

k d u. si ≤

E.Ab 35PN in genitive. It seems to contain the lexical element k d -, as esa ? k d ow “ ,k d ou≤ , k d u≤ ol “ (= Hitt., Luw. ¢ant- ).

k d u≤ ol “ C.xx 4, C.xx 5PN? It seems to be a name of the u≤ ol -family, but the final -“ couldbe a plural ending (cf. kbdyn“ , sarni “ , etc.) Note also the element k d - (seepreceding entry).

Schürr (2001b:117) suggests that we are dealing with the plural of an adjec-tive whose meaning would be ‘belonging to the king (= the god)’ (for themeaning ‘king’, cf. esa ? k d ow “ , k d ou≤ ).



373

k d u≤ opizipususot C.Hy 1b

Sequence most probably consisting of more than one word, but impos-sible to be segment with any confidence.

Note the presence of the sequence k d u≤ º, to be related either totally or par-tially to esa ? k d ow “ , kdou“ , k d u≤ ol “ .

kenE.Th 28

ki d bsi ≤ E.Me 15Ethnic name (less probably PN) in genitive. It could be (at least orig-inally) the ethnic name derived from a place name *ki d b-, to be identifiedwith the Carian city Kinduh in Greek sources.

Hajnal (1998:90). Schürr has attempted to establish diff erent connections forthis word, which he instead considers a PN: comparison with the Carian nameKindacow (Blümel KarPN:16), name of the father of the founder of Masanvrada,according to Stephan of Byzantium (Schürr 1991[93]:170); connection to Mil. x ñtabasi , poss. adj. of x ñtaba ‘ruler’ (Schürr 1996b:152) and to the Ciliciannames Kendebhw, Kendhbhw, Kendhbaw (Schürr 2001b:105), for these latter names,see Zgusta KPN § 576).

ki l ara d [ C.Ki 1ki l [ C.Ki 1

Place name: the Carian city of Kildara/Killara (Kildara/Killara ). Inthe first example it is not clear if d belongs to the word.

The identification was made already by Kowalski (1975:79, 83), although histranscription was still very unsatisfactory (krº´ zara ).

klorul E.Me 6Ethnic name, title, or common noun in nominative. Hardly a PN, giventhe structure of the inscription in which it appears.



374

See above p. 270, where the inscription is analysed and this word is taken asan ethnic name. Schürr (1992:135) instead suggests that it is a common nounwith the meaning ‘wife’, because he interprets the entire inscription triqo parma ≤≤ ∞i klorul ∞i as ‘Triqo (f.), the wife of Parma≤’.

knor C.Kr 1

kojol E.Me 44aEthnic name, title or sim. in nominative. Hardly a PN, given the struc-

ture of the inscription in which it appears.

Above p. 271, kojol is analysed as an ethnic name and tentatively related tothe name of the island Kos. This connection is hampered by the fact that thename of the person to which kojol seems to make reference, is of Egyptianorigin. Schürr (1992:155) proposes interpreting kojol as a ‘title in -ol ’ (likenuol º, sarmrol º ), and suggests connecting it with the Carian gloss ko›on, ‘sheep’.

kolt

E.Si 2

ko“ m≤ E.Th 39PN in genitive

kowrn[ . . . ?E.Si 2

kow [?-?]E.Th 24

krws E.Th 39, E.Th 45 (krwß )PN in nominative.

ksbo

C.My 1PN in nominative. Carian name that corresponds to Xasbvw (all theexamples are from Mylasa) in Greek sources. Cf . also Kasbvlliw.




375

ksolb≤ E.Me 43a

Ethnic name (less probably PN) in genitive (see above p. 269). In eithercase, it seems to be related to the place name Kasvlaba from Greeksources.

Schürr (1992:143), Adiego (1994a:36). As an ethnic name: Janda (1994:174).

k ≤ at ÿbr E.Th 2PN in nominative. Name belonging to the -(d)ybr -family. Cf . in this casethe Lycian name Janduberiw (Zgusta KPN § 1061).

For this family of names, see d ÿbr-.

k “ mmsm[ . . . ]E.AS 8

ktais

C.Eu 1PN in ‘s -case’ (‘dative case’) or, less likely, in nominative. Carian formof the Greek name ÑEkata›ow.

For the identification: Schürr (1992:154), Schürr apud Adiego (1994a:39,1994b:252). In Adiego (1994a:39) ktais is interpreted as an s-stem created froma Greek nominative (like Lycian zeus-, from ZeÊw ), but in the present book ananalysis of -s as an ending is preferred (see pp. 288–289).

ktmnE.Th 37PN, perhaps an incomplete form of ktmno (see the following entry).

Adiego (1994b:251).

ktmnoE.Th 25PN in nominative, corresponding to Ekatomnvw (Zgusta KPN § 325–1–3,

Blümel KarPN:13) in Greek sources. Note the variant ∞tmño-.

Adiego (1994b:251–252).



376

k t ? tri ≤ E.Ab 40.

PN in genitive. Perhaps related to the PN ka t t ÿri ≤ .

kuari ≤ bar E.Me 18aAn unclear form. Perhaps it must be segmented into kuari ≤ bar , the firstword being a PN in genitive (cf . the following entry kwar ≤ ). However,bar would remain unexplained.

kwar ≤ E.Me 31PN in genitive.

kudtubr E.Th 9PN in nominative. It apparently belongs to the family of names in-(d)ybr-, but note the use of u instead of y/ ÿ (like “ od ubr-, q.v. )

kwri ≤ E.Th 34PN in genitive. Perhaps it contains the same stem as kwar-, kuari ≤ bar .

∞arnos E.Me 45∞arnos d C.xx 2See alos, alos d .

∞arr ≤ E.Ab 32PN in genitive.

∞aye E.Ab 31PN in nominative.

∞diye ≤ C.St 2PN in genitive.

377

∞i E.Sa 2: q ÿri≤ ∞i; E.Me 6 (2× ): parma≤≤ ∞i, klorul ∞i; E.Me 8b: armon

∞i, E.Me 9: arlio[m≤ ] ∞i; E.Me 10 (3× ): q[---]≤ ∞i, [mw]don≤ ∞[i], [--- ]w≤ord≤ ∞i; E.Me 12: mwdon≤ ∞i; E.Me 13 (2× ): wet≤ ∞i, mwdon≤ ∞i;E.Me 16 (2× ): pikra≤ ∞i, mwdon≤ ∞i; E.Me 17: bem≤ ∞i; E.Me 18b (2× ):idmuon≤ ∞i, mdayn ∞i; E.Me 19: zmu≤ ∞i; E.Me 20 (2× ): “rwli≤ ∞i,mwdon≤ ∞i; E.Me 21: q ÿblsi≤ ∞i; E.Me 23: a[rb]ikarm≤ ∞i; E.Me 25:parpeym≤ ∞i; E.Me 28 (2× ): pntmun≤ ∞i, mwdon≤ ∞i; E.Me 30: pl eq≤∞i; E.Me 31 (2× ): kwar≤mHm≤ ∞i, mwdon≤ [ ∞ ]i; E.Me 32 (2× ): ∞i en,mw[d]on≤ ∞i; E.Me 33a (2× ): myre≤ ∞i, mdayn ∞i; E.Me 33b: myre≤ ∞i;E.Me 35: mwdon≤ ∞i; E.Me 38: ∞i ted; E.Me 40 (2× ): pikrm≤ ∞i, mwdon≤∞i; E.Me 42 (3× ): mwsat≤ : ∞i, mwdon≤ : ∞i, tbridbd≤ : ∞i; E.Me 43a:“rquq≤ ∞i; E.Me 43b: mno≤ ∞i; E.Me 44a: kojol ∞i; E.Me 44b: mwton≤∞i; E.Me 45 (2× ): [?]iam≤ ∞i, yi≤∞biks≤ ∞i; E.Me 46b: mwdon≤ ∞i;E.Me 47: paraibrel ≤ ∞i; E.Me 48: [-]owt≤ ∞i; E.Me 50b: p∞simt≤ ∞i;E.Me 57: ]i≤ ∞i; E.Me 58: ]s≤ ∞i; E.Bu 6: ursea∞k ∞i; E.xx 1: isor≤ ∞i;C.Eu 2: omob ∞i; C.Si 2a: pda∞m≤uñ ∞i

∞ j

E.Me 36Originally a relative pronoun, turned into a particle for introducing complements. From PA *k w is < PIE *k w is (Hitt., CLuw. kui “ , Lycian ti ,Milyan ki ). Most spellings point to a postclitical usage, with the excep-tion of E.Me 32: ∞i en, where it seems to be proclitic.

See Adiego (1993a:213–216) for a brief status quaestionis and for a functionalcomparison of ∞i with the Old Persian relative constructions. Compared toLycian and Milyan relatives in Adiego (1994a:46). See Hajnal (1997a) for a

more detailed treatment and here pp. 273–275.

∞i ∞E.AS 7Meaning and function are unclear.

Schürr (2001:98) compares it to the Lycian indefinite pronoun tike (cf . alsoMilyan -kike ). While the comparison is sound from a phonological point of view, the presence of an indefinite pronoun in E.AS 7 depends on the over-

all interpretation of the inscription, a question that remains unresolved.

∞iqud E.Si 1Very probably a PN in nominative.



378

∞lbiks ≤ E.Th 33

PN in genitive. It seems to include the nominal stem -biks- (cf . piks [,dbiks, ÿ≤ biks, yi ≤ ∞biks-, but there are no clear parallels for the result-ing first element ∞l-.

∞l mud [?C.Ia 3Perhaps an epithet of the word that it follows, the GN trqud e ‘Tarhunt’,although other interpretations cannot be discounted (for instance, a ver-bal form).

The possibility of a word in agreement with trqud e would increase if we acceptGusmani’s proposal of reading a letter e after ∞l mud [ (therefore trqud e ∞l mud e ,with the same ending for both words).

∞? mpi E.Si 10

∞tmño≤

C.Si 2a (2× )PN in genitive. It is the typical Carian name Ekatomnvw (cf. the vari-ant form ktmno ), that here makes direct reference to the well-knownCarian dynast Hekatomnos, the father of Maussollos, Artemisia, Idrieusand Ada.

Schürr (1992:137). See also Adiego (1994b). Neumann has repeatedly arguedin favour of a purely Greek origin of the name (a hypochoristicon of an

*ÑEkatÒmnhstow: Neumann apud Schürr 1993:137, n. 6; Neumann apud Adiego1994b:248; Neumann 1994:17), but the existence in Carian of a noun mno-‘son’ means that we can analyse the name as a Carian compound kt+mno (forkt- cf. Akta-ussvllow, kbo-kt-, etc. see Adiego 1994b).

∞toi C.My 1∞toi ≤ PN in nominative (∞toi ) and genitive (∞toi ≤ ). Tentatively compared to

ktai-, ÑEkata›ow.See Adiego (2005:90–91). For the reading ∞toi ≤ , see Blümel (2005:188).



379

∞ yrapai ≤ C.Ka 2

∞ yrpai C.Ka 2Two forms apparently belonging to the same paradigm: genitive (∞ura- pai-≤ ) and perhaps nominative (∞ yrpai ). It is not clear, however, if weare dealing with a PN. Note the divergent vocalisation ºrpº/ºrapº.

∞[ - ]urb≤ E.Ab 36PN in genitive

∞[--8--]tuñ d ñ [ C.Si 2a

limtaoa C.Ia 1

lkor ≤

E.Me 2, E.Me 36PN in genitive

loubaw E.Me 49PN in nominative. It appears in the contentious inscription E.Me 49.

lr Hñ C.Si 2a

ltari ≤ E.Ab 4, E.Ab 5 (ltari[ ≤ ])PN in genitive

l ÿ∞se E.Th 35l ÿ∞si ≤

E.Me 43aPN in nominative and genitive. However, it is not totally clear if weare dealing with two forms belonging to the same paradigm (which



380

would imply a vowel-stem alternation ºe / ºi- ), or two diff erent stemsderived from the same root. In any case, the name seems to corre-

spond to the Carian name Lujhw in Greek sources (Zgusta KPN § 836,Blümel KarPN:18). Cf. also Lu-uk-“ u, the name of an Egypto-Carianin Borsippa.

Adiego (1990a:134). On Lu-uk-“ u, see Eilers (1935:217), (1940:192), and Waer-zeggers, Borsippa , where the Egyptian origin of the Carians of Borsippa isstated.

lys [ ikl ]an

C.Ka 5lysiklas [-? ]C.Ka 5PN in accusative ( lysikla-n ) and in ‘genitive/possessive’ (lysikla-s(-?) ).Carian adaptation of the Greek name Lusikl∞w. It is not clear if lysik-las is a complete form (of a true genitive in -s ) or if a further lettermust be added. In this latter case, the most suitable solution is lysik-las [ n ], a possessive adjective in accusative for expressing the name of the father of

Nikokl∞win the inscription.

Frei-Marek (1997).

lysikratas [-? ]C.Ka 5PN in ‘possessive’ (lysikrata-s(-?) ). Carian adaptation of the Greek nameLusikrãthw. As in the case of lysiklas [ (see the preceding entry), it isnot clear if the word is complete (representing a true ‘genitive’) or

whether it must be completed (most probably as lysikratas [ n ]) in orderto obtain a possessive adjective in accusative.

Frei-Marek (1997).

mal ≤ C.Ka 1PN in genitive.

381

mane C.Hy 1a

mane ≤ C.Hy 1aPN in nominative and genitive. Typical Carian name that appears inGreek as Manhw.

Adiego-Debord-Varinlio< lu (2005:607).

manonC.Eu 2

mañ “ qara H≤ r l -?- [ C.St 2

ma ãtnor E.Th 34

maq l y≤ [

C.St 1Perhaps a complete form. If so, it would be a PN in genitive.

marariso[ - . . . ]E.Si 1Apparently the beginning of a PN.

mdaynE.Me 18b, E.Me 33a

mda ÿnE.Me 11a, E.Me 11b, E.Me 17Ethnic name (or similar) in nominative, the genitive of which is mwdon≤ (q.v. ). Diff erent possible interpretations have been envisaged, none of them definitive: ‘foreigner’, ‘Carian’, ‘inhabitant of Myndos (a Cariancoastal city)’. In any case, the word seems to contain the suffix -yn-/- ÿn- equivalent to the Luwic suffix for ethnic names *-weno/i- (cf . inCarian kbd-yn-“ ).

For the paradigmatic relationship with mwdon≤ , see Melchert (1993:82–83). If this relationship is, as it seems, correct, former attempts to explain mdayn/mda ÿnas a word for ‘husband/wife’ (Merigi 1980:35b–36a) or ‘farewell’ (Ray 1982b:184)must be ruled out ( pace Adiego 1993a:219–220).



382

mdot 2C.Ka 5

Function and meaning are unknown. The segmentation mdot 2 un adoptedhere is not definite.

In Adiego (2002:17–18) the word is analyzed as a plural genitive correspond-ing to mdayn/mda ÿn, and a meaning ‘foreigner’ is assumed for all of theseforms. The sense of sarni “ mdot 2 in C.Ka 5 would therefore be ‘representatives(sarni-“ ) of the foreigners’, corresponding to Greek prÒjenoi ).

Melchert (1998) prefers to read mdot 2un as a complete word and, assuming a glide-value (/w/) for the letter O (here transcribed as t 2 ), analyzes the formas a preterite first plural with the meaning ‘we have established’, and withmd º to be connected with Hitt. mid à(i)- ‘fix, fasten’.

md [ . . . ]E.Me 52Perhaps to be completed as md [ ay/ ÿn ].

md

C.Ha 1

See below md ane

md a C.Si 2ª (3× )See below md ane .

md ane E.Sa 1, E.xx 7, C.xx 2Analysis of this form has been much discussed (a verb or a chain of

particles?). Cf . also the possible variant mlane in Thebes (see mlane, ewm-lane, ewlane ).

On md ane and the two preceding forms md , md a see discussion in pp. 321–324.

me ®≤ E.Me 34PN in genitive.

me ÿqak E.AS 8



383

mi ∞≤≤ E.Ab 35

PN in genitive.

mlane E.Th 10Cf . md ane , and see p. 323 on the intricate relationship between thisform and ewlane, ewmlane.

mlan[-?]E.Th 35Cf. the preceding entry.

mlqi ≤ E.Th 27PN in genitive.

ml ne C.Ia 3

Connection to md ane/mlane and/or ºml n in uioml n, yoml n is possible,but far from certain.

mmn∞al E.Th 21

mnos C.Eu 1 (mn[os? ]), C.Ka 5mno≤

E.Me 10, E.Me 12 (m[no≤ ]), E.Me 16, E.Me 27, E.Me 39, E.Me 43b,C.Ka 1, C.Kr 1mn[ o-? ]E.Me 47

Common noun in genitive (mno-≤ ) and in a ‘s-case’ (mno-s ), ‘son’. Perhapsin some way related to HLuw. nimuwiza-, ‘son’. The precise analysis of mnos in C.Ka 5 is far from certain, given the unclear context in which

it appears.

Meriggi (1967:223), (1980:35a), Ray (1982b:184–185), Gusmani (1986:63),Adiego (1993a:216–219).



384

moa [ - ]l boror C.Ka 5

moi C.My 1PN in nominative. Perhaps it corresponds to the Carian name in Greeksources, Moiw.


mol “ C.Hy 1bC.My 1Plural nominative with the meaning ‘priests’?

See above pp. 306–207 on this interpretation, based on the analysis of C.Hy1b and its possible correspondence to the Greek text that follows it.

mplat

E.Th 11PN in nominative.

mqabaewleqo“ osk $ioms E.Th 12An impenetrable sequence, undoubtedly consisting of more than oneword. Note the isolable sequence ew , to be related to ew lane, ewm, alsoin Thebes.

mq t jq E.Th 4

mrsi ≤ E.Me 2, E.Me 26Ethnic name or PN in genitive.

mrsj [ . . . ]

E.Me 54Very probably related to the preceding entry.

mslmnlia C.Ka 5



385

msnord ≤ E.Me 3, E.Me 48

Ethnic name or, less probably, PN. Clearly related to the Carian placename Masanvrada (Zgusta KON § 782, Blümel KarON:174). According to Schürr, msnord-/Masanvradº can be analyzed as msn + ord , with asecond element comparable to Luwian -aradu in Tar ¢unt-aradu (LarocheLNH n. 1268, Piyam-aradu (Laroche LNH n. 981), so that msnord- =Luw. * Ma ““ an-aradu-

Connection to Masanvrada already mentioned in Adiego (1990a:136). Notethat the word was formerly read as †msnori ≤ , which conditioned some pro-

posals of interpretation. For the corrected reading, see Schürr 2001b:103,Schürr 2002:166–168).

As for Schürr’s analysis, see Schürr (2002:165–168). He interprets msnord-as a PN and instead prefers to compare it with the name of the eponymousfounder of Masanorada, Masanvradow (Steph. Byz. s.v. Masanvrada ).

On aradu-, see above p. 333.

msot C.Hy 1b

Genitive plural (?) or another case from a stem mso- or similar = ‘god’?

See above pp. 306–307 for this interpretation, based on the search for par-allels between C.Hy 1b and the Greek texts that follow it.

mt 1 yr C.Ka 2

mt ∞el ã

C.Si 2a

mudo[ n ]≤ E.Me 65mwdon≤ E.Me 10 ([mw]don≤ ), E.Me 12, E.Me 13, E.Me 14 (mwdon!≤ ), E.Me16, E.Me 20, E.Me 28, E.Me 29, E.Me 31, E.Me 32 (mw[d]on≤ ), E.Me35, E.Me 40, E.Me 42, E.Me 46b

Genitive of the ethnic name (or sim.) mdayn/mda ÿn, q.v.



386

Interpretation as ethnic name already suggested in Meriggi (1980:35). Identificationof -on- with the Luwic ethnic suffix -wanni- and connection to the Carian placename

Mundowalready noted in Adiego (1990b:501–502). See also Adiego

(1993a:210–212), Melchert (1993:82–83), Janda (1994:173–174).

mwk E.Th 22

mumn“ tnse-? E.Th 30

muot C.Hy 1a

-ot recalls identical endings in C.Hy 1 (k d u≤ opizipususot , msot , cf . also ylarmi t ). If the analysis of this latter is accepted (see s. v. ), muot can be also a genitive plural.

mwsat ≤ E.Me 42PN in genitive. Cf. the Lydian name Mousathw (Zgusta KPN § 987a).

Perhaps both names, and also the Pisidian names Moushta, Moshta,correspond to Luwian PN Muwaziti (Laroche LNH 840), a compoundof muwa- Hitt., Luw. ‘strength, force’, and Luw. ziti ‘man’.

Adiego (1992a:32), (1993a:233), (1994a:36).

mwton≤ E.Me 44b

Variant form of mwdon≤ , q.v.mute ≤ C.St 2PN in genitive. Cf . the Cilician name Moutaw (Zgusta KPN § 989–2).Behind muº, the Anatolian stem muwa- can be identified.

Adiego (1994a:36).

m ÿqudemE.AS 7



387

myre ≤ E.Me 33a, E.Me 33b

PN in genitive.

myze C.My 1PN in nominative.

Tentatively compared in Adiego (2005:91) to the Carian PN in Greek sources,Mouzeaw.

mHm≤ E.Me 31PN in genitive.

m[-]sao[ C.My 1

naria ≤

E.Me 5PN or title in genitive. If it is a personal name, it must be the fatherof psm“ kúneit in the bilingual text E.Me 5, which implies that this manhad a double denomination, Egyptian W3˙-jb-r‘-nb-[ (in the Egyptianpart) and Carian Naria-. However, it could instead be a title of Psm“kúneit(‘general’, ‘priest’ or similar). Possibly related in some way to the fam-ily of place names Naras/a/, Narisbara, Naruandow (connected toCLuw annarai-, ‘forceful, virile’ = Hitt. innar à- ‘forceful, violent’, allderived from PIE *h2nè r ‘man’).

For these and other possible examples of this stem in Carian onomastics, seep. 333.

naz E.AS 7

ne E.AS 7

nid ≤ kusas E.AS 8

388

nik [--]lanC.Ka 5

PN in accusative, to be completed nik [ ok ]lan, nik [ uk ]lan, or similar. It isthe Carian adaptation of the Greek name Nikokl∞w.

Frei-Marek (1997).

ninut E.Ab 20PN in nominative.

niqau≤ E.Me 18aPN in genitive. Very likely to be a Carian adaptation of the Egyptianpharaonic name Nechao/Necho ( Ny-k3w , Greek Nexvw ).

Adiego (apud Schürr 1996a:63, n. 11).

noril ? ams

C.Kr 1

not C.xx 2Verbal form (‘he brought’)?

This is the interpretation proposed in Adiego (2000:153–155), where not isanalyzed as a preterite third singular verb, to be connected to Hitt. nà(i)-,CLuw. (reduplicated stem) nana- < * PIE *neyH- ‘to bring’, see above p. 284.

nprosn≤ E.Ab 16Apparently a complete PN in genitive. However, Schürr has proposeda segmentation npro + sn≤ , interpreted as a PN (nominative) + PN(father’s name in genitive). As for the assumed first name, he comparesit with the Egyptian name Nfr-˙r Nefervw (DNb:641), phonetically[nefer˙ó] according to Vittmann. For sn≤ , Schürr provides the sameform in E.AS 8 and allegedly in C.Kr 1 (read and segmented diff erentlyin this case). However, although the explanation of npro is very sound(see also Vittmann), the existence of a Carian name *sn- is doubtful.

Schürr (1996a:68, n. 18), Vittmann (2001:42).



389

»ß w nE.Th 30

n≤ C.Kr 1

See a very hypothetical explanation (as pronoun) in p. 292.

n≤ n[-]s “ E.SS 1

ntokris E.Me 35(Presumably) feminine PN in ‘s-case’, a Carian adaptation of the Egyptianfem. name Nj.t-jqr (literally ‘Neith is perfect’), Greek Nitvkriw (DNb:628).This was the name of a daughter of Psammetichus I.

Schürr (1992:152, n. 9), Ray (1994:202–203). Vittmann (2001:52–53) arguesconvincingly (against Ray loc. cit.) that ntokris arrived in Carian directly fromEgyptian, without Greek intermediaries. This implies that in ntokris , the finals must be a case ending, not a vestige of the Greek sigmatic nomimative (seeabove p. 315).

ntroC.xx 1ntros E.xx 7Carian God name, assimilated to Greek Apollo, in dative (ntro ) and in

s-case (ntro-s ) or, less probably, a priest title (in nominative ntro and ins-case ntro-s ), derived from a god name ntrº = Apollo. Independently of either interpretation, there is a general consensus that ntro- should beconnected to Lycian Natr- in the PN Natr-bbijemi , ‘translated’ in Greekas ÉApollÒ-dotow.

On these two alternative interpretations (the latter defended in Melchert 2002),see above p. 282. Carruba (2002) rejects the equation Natr-/ntro- = Apollo,and argues for a generic meaning ‘god’ and for an Egyptian loanword in

Carian (from Egypt. ntr ‘god’).



390

nuE.xx 7

Function and meaning are unknown.

Melchert’s tentative translation of E.xx 7 (see Melchert (2002:308) implies theassumption that nu is an adverb, ‘now’ (to be compared therefore with Hitt.nu- ). See above p. 286, where an intepretation as a verb is suggested.

nuol $∞[---]sarmrol ∞ yt E.Me 4

Perhaps more than one word, but the lacuna does not allow us to iso-late words in the sequence.

Schürr (1992:155) claims to identify two “titles in -ol ”, nuol and sarmrol (cf .also his similar analysis of kojol ). While the repeated ending in -ol could begood evidence for segmentation, no parallel forms, either to nuol or to sarm-rol , are attested in Carian.

Final sequence ∞ yt could point to a verbal form, see above p. 276.

n[-]eakrnanbE.Si 3

ñmai l oC.Si 2aA sequence containing a verbal form?

Interpreted as a verb ñmai l + enclitic pronoun -o- in Adiego (2000:141–142),where it is further connected to Lycian ˜ mmaite ‘they have established’), see here

on p. 304.

oba C.Ka 2

obrbi ≤ C.My 1 (2× )PN in genitive.

obsmns [ C.Ka 2



391

obsmsmñ 1ñ C.Ka 2

omobC.Eu 2

or C.Ka 6PN?

orbáE.Th 20

orknC.Ha 1, C.xx 1Noun in singular accusative, of a stem ork- ‘phiale’, ‘vessel’, or similar.

Melchert (1993:80). Neumann and Edwin Brown (apud Melchert, ibid .) suggesta comparison with Gr. Ïrxh ‘jar’, Lat. orca ‘butt, tun’. Cf . also Lat. urceus, urna ‘water-pot’ (all these words probably share a common origin).

or ≤ E.Ab 15, E.Me 41, E.Th 36PN in genitive. According to Vittmann, an adaptation of the Egyptianname Ó r ‘Horus’ (phonetically [hòr])

Vittmann (2001:42). However, cf . or in Kaunos.

or “ a E.xx 7

ort C.Ka 5ort [-]C.Ka 5

oruC.Ka 5



392

otonosnC.Ka 5 (2× )

Ethnic name in accusative, ‘Athenian’. Derived from a Carian form*otono- for ‘Athens’ by means of a -s- suffix.

Much has been discussed about the precise nature of the -s- suffix. It could bea possessive suffix (= Lyc. -h-; also Frei-Marek 1997:35), an ethnic suffix equiv-alent to Lyc. -s- (Schürr 1998:161; similarly Hajnal 1997b:160, but resorting to Lyc, -is- ) or, less probably, an ethnic suffix equivalent to Lycian -zi (cf .Hajnal 1997b:160–161, n. 32, where the difficulties of this explanation arehighlighted). All three hypotheses are envisaged in Adiego (1998a:18).

otr “ C.Ka 2, C.Ka 5Pronoun in acc. pl., corresponding to Greek aÈtoÊw in the bilingualinscription C.Ka 5. It can be connected etymologically with Lyc. atla-/atra- ‘person, self ’ (cf . also HLuw. atra/i- ‘person; image; soul’.

This interpretation and etymological connection was established independentlyby various scholars: Adiego (1998a:21), Hajnal (1997b:164; 1998:102), Melchert

(1998:34), Neumann (1998:20)—who very appositely recalls the correspondanceLycian atru: Greek •autÒn in TL 25a, 4. See here on p. 297.

owdown[ . . . ]mwarudk ≤ oE.Th 10A chain probably containing more than one word.

The initial sequence owdown hardly seems to be separated from wdwn, q.v.

ow ∞meb≤ t E.xx 5

ouor C.Ka 2 (2× )

pals ≤ E.Ab 7, E.Ab 8, E.Ab 9PN in genitive.



393

panejt E.Ab 2

PN in nominative. Carian adaptation of the Egyptian name P3-n-Nj.t ,Greek Panitiw, literally “the one of Neith” (DNb:385). Cf . the variantform pneit )

Schürr (1992:152, n. 9), Adiego (1993a:254), Ray (1994:203 and n. 19). Cf .also Vittmann (2001:58).

paraeymE.Me 8a, E.Me 8b (para!eym)

PN in nominative. The name presents the well-known adverbial stem para- as a first element of a compound (cf. para-i b re l , Para-ussvllow,etc.). As for -eym, it recalls Arthumow. Compare also parpeym-.

parai b re l ≤ E.Me 47PN in genitive. A compound name consisting of para- (cf . above paraeym )and i b re l (= Greek Imbarhldow ).

Adiego (1994:36–37). On i b re l -, a stem derived from *i b r- = CLuw. im(ma)ra/i-,see above p. 335.

pare ÿs C.Kn 1

parma ≤≤ E.Me 6

PN in genitive.

par ãaq ?

C.Tr 2

Adiego (1993a:263), Hajnal (1995[97]:20). See the discussion of the reading and possible interpretations of this word in pp. 289–291.



394

parpeym≤ E.Me 25

PN in genitive. It is not clear if we are dealing with a compound with par(a)- as a first element, or whether in fact a stem parp- should be rec-ognized. For the final part of the word, cf . paraeym-.

For parpº, see s. v. prpwri ∞.

par ≤ olouE.AS 1

paruos ≤ C.My 1PN in genitive. Cf . the Carian name (f.) Paruv.

Adiego (2005:91).

par ÿd ∞≤ E.SS 1

PN in genitive. It corresponds to the Carian name Paraudigow in Greeksources. A compound name par- (cf. para-i b re l , Para-ussvllow, etc.) + ÿd ∞-

Adiego (1994:43). For the family of names containing the stem yri ∞-/yriq-/ ÿdiq/ ÿd ∞-, see pp. 262–263.

paryri ∞C.My 1

paryri ∞≤ C.My 1PN in nominative and genitive. A compound name: par- + yri ∞-. Seethe preceding entry.

pauC.Hy 1a, C.My 1 p ? au

C.My 1



395

pau≤ C.Tr 1, C.Tr 2

PN in nominative ( pau ) and genitive ( pau≤ ) corresponding to the Carianname transcribed in Greek as Paow.

Adiego (1994:37).

pa [-]in[-]t ≤ E.Ab 17Apparently a PN in genitive.

pd [ E.Me 64aCf . the following entries.

pdnejt E.Sa 2PN in nominative. Carian adaptation of the Egyptian P3-dj-Njt , liter-ally ‘the one whom Neith has given’ (Greek Petenaiyiw, Petenhyiw,

Petenht, see DNb:316).

Adiego (1992a:29–30).

pdtom≤ E.Bu 2NP, a Carian adaptation of the Egyptian name P3-dj-Jtm, literally ‘theone whom Atum has given’, Greek Peteyumiw, Petetumiw (DNb:294).

Schürr (1992:152, n. 9) and apud Ray (1994:205); Vittmann (2001:58).

pduba E.xx 4PN in nominative.

pdubez E.Ab 15

PN in nominative. Carian adaptation of the Egyptian name P3-dj-B3st.t ,literally ‘the one whom Bastet has given’, Greek Petyubestiow, Petobastiw,Petoubastow, Petoubestiw (DNb:303).

396

Identification proposed in Schürr (1996a:60), where the word is still read as† pdub ÿz . Corrected reading given in Schürr (2000:172). See also Vittmann(2001:42). Cf . also the corresponding feminine theophoric name ttbazi, ttubazi.

pdubi ≤ E.Me 10 (p∂uüi≤ ), E.Ab 6PN in genitive. Cf . pduba . It is possible that we are dealing with twoforms of the same paradigm, cf . the similar situation in (l ÿ∞se/l ÿ∞si ≤ ).

pd a ∞m≤ uñ C.Si 2a

An accusativus genetivi in agreement with the PN pñmnn≤ ñ that precedes it?

See Adiego (2000:144–148) for this morpho-syntactic analysis and for someattemps at etymological explanations (particularly the hypothetical connectionof pd º with Lyc. pdde ‘place’, cf . also here on p. 304).

pjabrmE.Me 12PN in nominative. According to the illustration of the stela in which

it appears, the name is feminine.

pjdl ?

C.xx 1Noun in apposition to acc. sg. orkn? Other interpretations are alsopossible.

The meaning and function of this word have been much discussed. Melchert(1993:80–81) interpreted it as a word ‘gift off ering’ (from a * piyod hlom, to berelated to the stem CLuw. piya-, Lyc. pije- ‘to give’), a view followed here (seep. 282). Totally diff erent is the approach of Janda (1994:178), who preferredto see here a verb comparable to Lyd. bill /pill/ (< * pid-l ) ‘he has given’.

pi d aru[ C.St 2PN, probably to be completed pi d aru[ ≤ ] in genitive. A possible Carianadaptation of the Greek name P¤ndarow.

Adiego (1994a:39–40).



397

piew E.Ab 38

PN in nominative. Adaptation of the Egyptian name P3-n-jwjw (liter-ally ‘that of the dog’), Greek Pieuw, Pihow, Pih# (Cf. DNb:349).

See Vittmann (2001:44) for the identification and for further details on theEgyptian variants of the name.

pikarm≤ E.Me 14PN in genitive. It is equivalent to the Lycian names in Greek sources

Pigramiw, Pigramow (Zgusta KPN § 1255–1/2). The name contains thesame stem as pikra-/pikre-, q.v., to which a m-suffix has been added. Cf .also the variant form pikrm≤ , and the compund name dbikrm, dbkrm(d +bikrm = pik(a)rm- ).

Adiego (1992a:36), (1993a:233), Kammerzell (1993:19, 22).

pikra ≤

E.Me 16 pikre ≤ E.Me 3PN in genitive. It is not clear if we are dealing with a simple alternancea/e or with two diff erent stems, one in ºa- and the other in ºe . The nameappears in Greek sources as Pigrhw/Pikrhw, a very spread Anatolian name.

The name shows the stem pik- = Luw. * pi ¢a-.

Adiego (1992a:36), (1993a:228–229). On the family of names built on this

stem, see p. 337.

pikrm≤ E.Me 40PN in genitive. A variant, ‘devocalised’ form of pikarm≤ , q.v.

piks [ C.St 1A PN or part of a PN. The part conserved clearly contains the nom-inal element piks-/biks- Cf . dbiks, ÿ≤ biks-/yi ≤ ∞biks-, derived from thestem pik- = CLuw. * pi ¢a-, as pikre-/pikra-, pik(a)rm-.

On this stem, see p. 337.



398

pim[ . . . ]C.Si 2b

pisiri E.Ab 1PN in nominative. Very likely to be an adaptation of the Egyptianname P3-n-Wsjr , literally ‘the one of Osiris’ or P3j-Wsir , ‘this of Osiris’,Greek Pisiriw.

Ray (1994:203); cf. also Schürr (1996a:61–62), Zauzich (apud Schürr, ibid .),Vittmann (2001:58). This Egyptian interpretation of the name must be pre-

ferred to former attempts to connect this name with Anatolian proper names(Adiego 1993a:248 and, with many reservations, 1994:43).

pisma ≤ k E.AS 7, E.AS 3PN in nominative. Carian adaptation of the Egyptian name Psmtk ,Greek Cammhtixow. See also the devocalised form psma ≤ k .

Identification already proposed in Kowalski (1975:91).

pisma [ “ /≤ k . . . ]E.AS 4The same name as the preceding entry. Doubts about completing “ or≤ arise from the existence of psm“ k (wneit ), as well as pisma ≤ k .

pisñoi C.Si 2a

Analyzed as a verb pisñ + clitics (-o-i ), in Adiego (2000:141), where pisñ is identified as a preterite third plural of a root pi-, ‘to give’. See abovep. 304 for details.

p.iub[ a ]Ω i ≤ E.Me 1PN in genitive.Probably the Carian adaptation of an Egyptian name whose final partcontained the name of the goddess Bastet ( B3st.t ), although no exact

parallels for the whole name can be found.

Schürr (1996:62). For ‘Bastet-names’ in Carian, cf. ttubazi-, ttbazi-.



399

piubez E.Ab 10

PN in genitive. It seems to be a variant of the preceding entry.

p∞simt ≤ E.Me 50bPN in genitive. Tentatively compared by Schürr to the Egyptian nameP3-dj-Ó r-sm3-t3wy (shorter form P3-dj-sm3-t3wy ), literally ‘one whomHorus, uniter of two lands, has given’, Greek Potasimto.

Schürr (apud Ray 1994:205). Not included by Vittmann in his list of Egyptiannames in Carian sources (Vittmann 2001:58–59). Although the use of thisEgyptian name among the Carians of Egypt fits well with Egypto-Carian envi-ronment (Potasimto was the name of the commander of the Carian and Ionianmercenaires in the Nubian campaign of Psammetichus II), and the similari-ties between Carian and Greek adaptations are striking, the use of Carian ∞for Egyptian d or dj is surprising (compare the use of Carian d or t for Egyptiand or dj in other names that also include the Egyptian verb dj , ‘to give’: pdnejt,

pdtom, ptnupi , etc.).An Anatolian interpretation, by connecting p∞si º with Pija-, Pije-, Pijo-:

Lycian Pijaw (Zgusta KPN § 1263–1), Pije-darow (KPN § 1263–2), Carian

and Lycian Pijodarow ) is given in Adiego (1993a:248).

pla ? t E.Th 3PN in nominative. If the reading plat is accepted, it could be a variantof the name pl at q.v. (l instead of l is typical in Theban inscriptions).

plqo

E.Me 40PN in nominative. It appears in Greek sources as Pellekvw, Pelekvw.Note particularly Peleqow in the Greek graffiti of Abu-Simbel.

Adiego (1993a:234), Schürr (1991–1993:170). On Peleqow = Pel(l)ekvw (againstformer interpretations as a Greek name derived from p°lekuw ), see Adiego(1994a:37), Masson (1994b:140), (1995:175).

plqodse

E.Th 52PN? It seems to include the name plqo, but final ºdse remains impossibleto analyse.



400

pl at E.Ab 7, E.Ab 8, E.Ab 9

PN in nominative.

pl at t E.AS 6PN in nominative. Perhaps related to the preceding entry, but the final-t remains unexplained.

pl eq ≤ E.Me 30PN in genitive. It corresponds to the Carian name in Greek sourcesPeldhkow.

Adiego (1993a:234), (1994a:37).

pneit E.SS 1PN in nominative. Variant form of panejt (q.v. ), an Egyptian name.

Schürr (1992:152, n. 9), Adiego (1993a:254), Ray (1994:203).

pnl d ≤ wl E.Me 49

About the peculiar inscription where it appears, the reading of which is verydifficult, see p. 279.

pntmun≤ E.Me 28PN (or title?) in genitive. According to Vittmann, it could be the adap-tation of Egyptian p3 ˙m-n∆ r n Jmn *[ p˙ent 6n6mùn ], literally ‘the Prophetof Amun’. Among other possibilities, Vittmann suggests that it couldbe a title (therefore sanuq “ ue pntmun“ ∞i mwdon≤ ∞i : “Stela of sanuq , who(was) the ‘Prophet of Amun’ (a priest title), who (was) mwdon.”).

Vittmann (2001:46–47). The fact that the three names of E.Me 28 appearin genitive makes the structure of the inscription very ambiguous. See Vitmann(2001:47) for diff erent possible analyses. Given this ambiguity, I prefer to inter-pret the inscription as a N-≤ (PN)—N-≤ (father’s name)— mwdon-≤ (ethnic nameprobably referring to the father).

401

pnu≤ ol E.Th 40

pnw ≤ ol E.Th 27 pnu≤ ol E.Me 19PN in nominative. It appears in its Greek adaptation as Ponussvllow.Note the variant form for the genitive, punw ≤ ol ≤ .

A name of the u≤ ol -family (q.v. ).

Adiego (1990a:135). On the Anatolian origin of the element p(u)n-, see pp.337–338.

pnu≤ o≤ C.My 1PN in genitive. No parallel form is found in Greek sources, but it ispossible that the word should be corrected as pnu≤ o<l>≤ , so that thename would be the same as that of the preceding entry.

Adiego (2005:84).

pnyri ≤ ruE.AS 5

pñmnn≤ ñ C.Si 2aPN apparently in ‘accusativus genetivi’. The name appears in Greekas Ponmoonnow (documented also in Sinuri).

Schürr (1992:138).

poloE.Me 8bPN (?) in nominative. Coordinated with the PN in nominative paraeymby means of sb, ‘and’.

Melchert (1993:84) suggests that polo is a common noun representing a kin-ship relation. Although this possibility cannot be ruled out, this tentative pro-posal of meaning and etymology (‘son’, comparing Hitt. pulla- ‘child, son’) is,as Melchert himself recognizes, very speculative.



402

potko≤ l ≤ ?

C.Ka 8

If the reading of the final letter is accepted, it is apparently a genitive(of a PN?).

p? owk E.Mu 1PN in nominative.

prna ∞nonE.Th 34

A form in accusative sg.? It seems to agree with another word ending in -n(dm-?-n ).

pr ãidas E.xx 7A word with s-ending, in agreement with ntros ‘Apollon’, q.v. Tentativelyconnected by Schürr with Bragx¤dai ‘Branchids’, the priestly family in

charge of the sanctuary of Apollo in Didyma, near Milet.

Schürr (1998:158).

prpwri ∞E.Th 46PN in nominative. Apparently a compound name that can be seg-mented as prp + wri ∞. For the first element, cf . perhaps the Lycian namePerpenduberiw (Zgusta KON § 1242–1) or even Carian parpeym-. The

second element seems to be a variant of ºyri ∞, also present in othercompound names (idyri ∞-, paryri ∞- ).

For a possible explanation of the alleged use of w instead of y, see p. 105.

psikro≤ E.Me 51PN in genitive.



403

psma ≤ k E.Th 11 (psma≤[k]), E.Si 7, E.Bu 4, E.Bu 5

psma ≤ k ≤ E.Si 2, E.Bu 1 psma [ ≤ /“ k . . . ]E.Me 55PN in nominative ( psma ≤ k ) and genitive ( psma ≤ k-≤ ). Carian adaptationof the Egyptian name Psmtk , Cammhtixow. Cf . the variant form pisma ≤ k .

psm“ kwneit ≤ E.Me 5PN in genitive. Carian adaptation of the Egyptian name Psmtk-‘wj-Njt ,literally ‘Psammetichus in the arms of Neith’. Carian and Egyptianforms appear together in the bilingual text E.Me 5.

Adiego (1992a:29–30).

psnl oC.Ha 1

On the diff erent possible analyses of this word, see p. 284.

psoir ≤ C.My 1PN in genitive.

psrkrte E.Th 30

psu≤ ol ≤ C.Ka 1PN in genitive. It belongs to the family of names in u≤ ol / Ussvllow,but the remaining first element ( ps- ) is not clear.

ps ÿ“ [|? ]E.AS 7

Reading and segmentation (suggested by Schürr) are very doubtful.

See the entry ai [-]iqom for an alternative analysis.



404

ps H ÿm[ - ]≤ E.Me 27

PN in genitive.

ptn“ e E.Ab 3PN in nominative.

ptnupi E.Me 18a

PN in nominative (?) of Egyptian origin: P3-dj-Jnp, Greek Petenoupiw(literally ‘the one whom Anubis has given’ (DNb:27).

Schürr (1992:152, n. 9). Ray (1994:204). Ray (ibid .) also off ers an alternativeEgyptian explanation, starting from *P3-dj-nfr (literally ‘the gift of the goodone’), but this name, as he recalls, is not documented in Egyptian.

ptnuq ? i ?

E.Ab 26

PN in nominative of Egyptian origin, according to the new reading proposed by Vittmann: ptnuqi would be Egyptian P3-dj-‘nq.t (literally‘the one whom [the goddess] Anukis has given’, DNb:294), phoneti-cally interpreted as [pe†e‘anùqi].

Vittmann (2001:44). Vittmann does not rule out an alternative reading ptnuti ,which also has a good correspondence in Egyptian: P3-dj-(p3)-ntr (literally ‘theone whom the god has given’), Greek Petepnouyiw, DNb:306 (phonologicalreconstruction: [pe†e(p)nùte)].

pttu≤ E.Me 27PN in genitive. Tentatively interpreted as an adaptation of an Egyptianname, which is not in fact documented, *P3-dj-t3wy, literally ‘One whomthe two lands have given’.

Ray (1994:205); Vittmann (2001:58).



405

punm[ - ]≤ E.Me 65

PN in genitive. It seems to be a compound name with pun- as the firstelement.

See Schürr (2003a:95), who goes a little further and compares it with Lyc.Punamuwe , Ponamoaw (Zgusta KPN § 1288–1), and Punamu(W)aw (KPN § 1288–2,Pamphylia); cf . also Puna-A.A (= *Punamuwa ) in cuneiform sources (LarocheLNH: 1050). These latter comparisons depend on a restitution punm[ u ]≤ .

punot 2

C.Ka 2

See Adiego (2002) for a connection with Luwic puna-, ‘all’, and for a mor-phological interpretation as plural genitive.

punw ≤ ol ≤ E.Me 21PN in genitive. It is the same name as pnu≤ ol (and variants, equivalentto Ponussvllow ), q.v.

puor ≤ E.Bu 6PN in genitive. As according to Vittmann, an Egyptian name adaptedin Carian: P3-whr (literally ‘the dog’), Greek Povriw, Pouvriw (DNb:181)pronounced [puhór].

Vittmann (2001:41).

pur ? i ≤ C.Hy 1aPN in genitive.

purmoruos C.Ka 5

p[-]lu≤

E.Ab 33PN in genitive.



406

=q C.Hy 1

Connecting particle? It appears after armotrq d os, q.v.


qanE.Si 2

See the somewhat risky interpretation in Schürr (2000:171) as a word withthe meaning ‘dog’ and borrowed from Lyd. *kãn- (also ‘dog’), based on the

hunting scene drawn under the graffito E.Si 2.

qanor E.Th 34

Connected with the preceding entry in Schürr (2000:172).

#q #arm≤ E.Me 10PN in genitive.

qarpsi ≤ E.Me 36Ethnic name (less likely to be PN) in genitive.

On the structure of the inscription, which suggests an ethnic name, see pp. 267– 271. Note also the suffix -si- or -i- that appears in other possible ethnic names

(ki d bsi-, q ÿbsli-, yiasi- ylarmi- ). Among the possible Carian place names thatcould be connected with qarpsi- (see Appendix B), Karbasuand/a/, Karpasuand/a/ is a good candidate.

qarsio[ -?]E.Me 7

qdar ® ou≤ E.Me 41

PN in genitive.

See Adiego (1995:24–25) for an attempted etymological analysis (compared withCLuw. ¢utarl à- ‘slave, servant’; cf . the PNs in Cuneiform sources › u-da-ar-lá,› u-du-ur-lá, › u-u-tar-li, › u-ut-ra-la-(a “ ), › u-ut-ra-li-i “ (Laroche LNH n. 411).



407

qeb≤ t E.Th 12

qkuE.Si 6

q ∞blio≤ E.Ab 40PN in genitive.

qla l i ≤ E.Me 37G 2[ q ? ]la l is E.Me 45PN in genitive (qla l i-≤ ) and in ‘s-case’ ([ q ]la l i-s ). This name appears inGreek sources as Kolaldiw, Kulaldiw.

Adiego (1993a:235)

qorbE.xx 1PN in nominative.

qot 2omuC.Kr 1

qrds C.Ki 1, C.Ka 2Word with a possible ‘institutional’ meaning, given its appearance intwo legal texts. Cf . the following entry and also g rdso[-]i [.

In Blümel-Adiego (1993:94) this is tentatively compared with Hitt. gurta-, ‘castle,citadel, acropolis’ (cf . also the place-name of the Phrygian-Carian borderlandGordio/ n/). Melchert (1998:35, n. 2) suggests connecting it with Mil. kridesi , aplace-name.

qrdsol “ C.Ka 2Acc. (or nom.?) pl. of a stem qrdsol -, apparently a noun derived bymeans of the suffix -ol - from qrds : ‘belonging to the qrds ’(?).



408

Blümel-Adiego (1993:94), Melchert (1998:35). The sequence it appears in, qrd-sol “ ait , could be interpreted as ‘they have made them belonging-to-qrds ’ or‘the belonging-to-qrds ones have made.’ (cf . Melchert ibid ., Adiego 1998a:22).

qtblem≤ C.xx 1PN in genitive. Name corresponding to Kotbelhmow (Blümel KarPN:17),Kutbelhmiw (Zgusta KPN § 771, Blümel KarPN:18).

Adiego apud Schürr (1992:142), Adiego (1993a:235), Melchert (1993:78).

qtbloE.Th 10If PN (in nominative), it must be the Carian name adapted in Greekas Kotobalvw.

Schürr (apud Adiego 1994a:43). The doubts about its character as PN are theresult of the very unclear context in which the form appears (see the remarksin Adiego ibid .).

quq ≤ E.Me 17PN in genitive. It is the Carian name that appears in Greek sourcesas Gugow. Cf. also the compound names dquq, “ rquq .

Adiego (1993a:235, 1994a:37). On the possible Anatolian etymology of quq ,see p. 334.

qurbo≤ E.Ab 10PN in genitive.

qwsal E.Th 12

qutbe E.Th 8

PN corresponding to the Carian name in Greek sources, Kuatbhw.

Adiego (1993a:235), (1994a:37).



409

q ÿblsi ≤ E.Me 21

Ethnic name or, less probably, PN. If an ethnic name, it clearly recallsthe place name Kublissow (Zgusta KON § 1296, Blümel KarON:171).

See somne ≤ (name to which q ÿblsi ≤ is referred) for the curious coincidence of personal name and ethnic name in the sole example of Svmnhw in the Greeksources.

Comparison with the ethnic name (but still taking it as a PN) in Adiego(1993a:235). Analysis as an ethnic name (with the same connection) in Janda(1994:174).

qyrbmud oloC.Eu 2

q ÿri ≤ E.Sa 2PN in genitive. It appears adapted in the Egyptian part of E.Sa 2 asK3rr .

For the identification of q ÿri ≤ with K3rr : Schürr (1992:135), Adiego (1993a:161).Egyptian k3 (written by means of a biconsonantic sign) is used here to reflect

a syllabic sound /ku/, so *kur º (Vittmann 1996). Note that in Egyptian a vowel/y/ did not exist, so that the use of /u/ for /y/ seems reasonable. The dou-ble r is explained by Zauzich (apud Schürr 1996:68) as a graphic attempt toemphasize that r did sound (the final r was not pronounced in Late Egyptian).

qzali C.My 1

qzali ≤ C.My 1PN in nominative and genitive.

Connection with the Carian name of Greek sources Kostvlliw (Adiego 2005:91)is very hypothetical!

q [---]≤ E.Me 10

PN in genitive.

rdudmm»≤ E.Th 42

410

rqemw E.Th 52

rsyE.Lu 2

rtimC.Hy 1aPN in nominative. Cf . the Carian name Artimhw and, more generally,the family of Anatolian names collected in Zgusta KPN § 108): Arteimaw,Arteimianow, Arteimow, Arteimhw, Artimaw, Artimhw, etc.

Adiego-Debord-Varinlio< lu (2005:611). On the possibilty that rtim and artmi (q.v. ) could be part of a single paradigm, see above p. 290.

r 1i C.Ka 4

sa

E.Me 26Demonstrative pronoun in nominative (‘this’), which appears following the word upe ‘stela’: upe sa : ‘this stela’. From PIE *∞o-/* ∞eh2- > P A *∞o-/∞à- (> Hitt. ka-, CLuw., HLuw. za- ). See also san, snn.

Adiego (1992a:33). Hajnal (1995[97]:23) suggests that the same form can berecognized in the sequence bi d ≤ lemsa.

On these pronominal forms, see pp. 319–320.

sa ?

awonE.Mu 1saawonE.Mu 1

sanG 1Demonstrative pronoun in nominative. It corresponds etymologically toHitt. ka-, CLuw., HLuw. za- ‘this’, see above sa .

Adiego (1992a:33). On the final n, see Melchert (1993:79–80) and here onpp. 288, 320.

411

sanuq ≤ E.Me 28

PN in genitive.

sarl ?

E.Me 5

sarni “ C.Ka 2 (2× ), C.Ka 5 (sarni[ “ ])Plural accusative of a stem sarni-. This word (or the wider sequence itappears in) corresponds to Greek proj°nouw in the Kaunos bilingual(C.Ka 5).

Frei-Marek (1997:39). There are diverse etymological proposals, none of themdefinitive: connection to CLuw. “ arri ‘above; up; for (?)’ (Frei-Marek, ibid., fol-lowed by Hajnal 1997b:164, hard to accept, because *“ arni “ would be expected,as they themselves recognize); related to CLuw. zar “ iya- ‘safe-conduct, Gastrecht ’or sim. (therefore sarni-, ‘guarantor’, Neumann 1998:29). Schürr (apud Hajnal1997b:164, n. 35) compares sarni- with Lyd. saretas ‘benefactor(?)’.

In Adiego (1998a:22) the equivalence to proj°nouw is seen in the whole

sequence sarni [ “ ] mdot 2 ‘representative of the foreigners’ (with mdot 2 [gen. pl.]related to mdayn/mda ÿn/mwdon-, interpreted as ‘foreigner’).

sbE.Me 8b, E.Th 13, E.xx 6, C.Si 2a (2× ), C.Ka 2 (8× ), C.Ka 5 (8× ),C.Kr 1

Coordinative conjunction: ‘and’. When there is interpunction in thetext, it always appears attached to the following word, as a sort ofproclitic.

Cf . Lycian se and particularly, Milyan sebe (both ‘and’). From PIE*∞e (cf . Venetic ke ‘and’), plus a reinforcing particle *-be ?

For sb = ‘and’, see the explanation already off ered in Schürr apud Ray(1990b:129–130). Connection with Milyan sebe : Neumann (1993:296). For se-< PIE *∞e : Adiego (1995:31–32).



412

s d i C.Tr 1, C.Al 1

s d isas C.Ka 1s d isas ?

C.Kr 1Noun used in funerary contexts (therefore ‘tomb’, ‘stela’ or sim.) Themorphological analysis of these forms remains unclear. Cf . the variantform si d i .

Connected with PIE *∞ei- ‘to lie’ (Lyc. sije- )?

semw ≤ E.Me 16PN in genitive.

seqqejewsk E.Th 4

si d i C.Tr 2A variant form of s d i, q.v.

siral E.Me 49

siyklo≤ C.Ia 3PN (?) in genitive,

skdub rotoz ≤ C.My 1A sequence containing an onomastic formula PN-Ø PN-≤, the maindifficulty being the point of segmentation. The best solution seems tobe skdu b rotoz ≤ , but other alternatives cannot be dismissed.

sla ÿ≤ E.AS 6PN on genitive.



413

sl ∞maewmE.Th 34

smd ÿbrs C.Ha 1PN in nominative (of a s-stem) or in s-ending (if the stem is smd ÿbr- ).The name belongs to the family of nouns in -(d)ybr-/-(d) ÿbr , cf. ardy-byr, d ÿbr , etc. A comparison with the Carian name in Greek sourcesZermenduberow by Blümel (1990:81) is attractive, but the lack of r posesa problem.

sm“ s [–5–]C.Si 2a

sm[–7–]a ∞e [ C.Si 2a

snnC.Ha 1, C.xx 1

Acc. sg. of the demonstrative pronominal stem sa/san-, q.v.

Melchert (1993:79).

sn≤ E.AS 8

sñaid l oC.Si 1

Tentatively interpreted as a verb (aid- ‘they made’, cf. ait ) preceded by a pro-noun or introducing particles and followed by clitics in Adiego (2000:152),where even an exact correspondence with Lyc. señnait e is proposed.

sñis C.Ka 1

See p. 291 for a possible interpretation as a demonstrative pronoun (relatedto sa/san-, ‘this’).



414

som[ n ? ]e C.St 1

somne ≤ E.Me 21, E.Me 34PN in nominative (somne if the reading is accepted) and genitive (somne ≤ ).Directly comparable to the Carian name in Greek sources, Svmnhw.

It is undoubtedly a matter of chance, but it is curious nonetheless to note thatsomne- is followed in E.Me 21 by a possible ethnic name q ÿblsi-, while theonly example of Svmnhw in Greek sources is an individual belonging to theKublisse›w (inhabitants of Kyblissos), mentioned as witnesses in a treaty between

Mylasa and Kindya (Inscription of Mylasa, n. 11 in Blümel’s edition).

sqla E.Si 4

sqlumidunE.Si 4

srton[ - ]t [. . .?]

E.Ab 28

stspñ C.Si 2a

sursiabk E.Si 6

susoE.Lu 2

s [--]et ≤ E.Me 29PN in genitive.

≤ as C.Eu 1

A variant form of ≤ jas (‘tomb’, or similar).



415

≤ d unC.Eu 2

≤ emot E.Th 10

≤ enE.Me 18a

≤ jas G 1Noun corresponding to Greek s∞ma in the bilingual inscription G 1.

Adiego (1992a:33). For the difficulties of relating ≤ jas to the Carian gloss soËa(n),see p. 10.

≤ o≤ niabkol C.Eu 2

≤ t ≤ E.AS 8

≤ ugl iq E.Me 5≤ ugl i ≤ E.Me 30, E.xx 1Ethnic name or, less likely, PN, attested at least in genitive (≤ ugl i ≤ ).The exact analysis of ≤ ug liq is unclear. If an ethnic name, it can be

connected with the Carian place name Souaggela, although the doubtsabout the precise sound value of 0 g make this identification moredifficult.

Adiego (2004:310). Connection to Souaggela suggested to me independentlyby Pérez Orozco and Melchert (both in pers. comm.).

≤ umoC.My 1

≤ uni ≤ E.AS 8



416

≤ ysña l [ C.St 1

“ abd ? aikal E.AS 2

“ amow E.Ab 4, E.Ab 5PN in nominative. It corresponds to Samvuow, Samvow in Greek sources.

Adiego (1994a:38). Vittmann (2001:55–57) does not rule out the possibilitythat “ amow-Samv(u)ow could be a ‘Carianised’ form of the Egyptian name Ô 3j-

jm=w (directly attested in Carian as t amou q.v. ).

“ amsqi [ . . . ?E.Me 24

“ annC.Ia 3

“ anne C.Ia 3

“ aoyr ∞ri C.Si 2a

See Adiego (2000:148–149) for a very hypothetical attempt to connect it with Xrusaor¤w, the oldest name of Stratonikeia.

“ arkbiomE.Sa 1, E.Me 56 (“ark[bi/jom . . .?]), C.My 1PN in nominative. Transcribed as ” 3rkbym in the Egyptian part of thebilingual E.Sa 1. Not found in Greek sources (although the form couldbe confidently reconstructed as *Sarkebivmow ). Compound name formedby the adverbial stem “ ar- (/gr. Sar- ) (= CLuw. “ arri , Lyc. hri , Mil. zri ;cf . also Hitt. “è r ‘upon’) and the stem kbiom-, also attested as an inde-pendent name (see kbjom-, Greek Kebivmow ).



417

Curiously, the Greek reconstruction *Sarkebivmow was already given as thepossible form corresponding to Egyptian ” 3rkbym in Masson-Yoyotte (1956:52),

when both the true form “ arkbiom behind the Carian letters and the nameKebivmow were still unknown.An initial, and still imperfect transcription of the name appears in Kowalski

(1975:90). See also Ray (1981:153). For an analysis of the name: Adiego(1993a:242).

“ arnai ≤ E.Me 17“ arnajs E.xx 6PN in genitive (“ arnai-≤ ) and in ‘s-case’ (“ arnaj-s ). It is not clear if thestem can be related to the adverbial stem “ ar-. A connection to theCarian name of Greek sources Sarnow is hampered by the doubts aboutthe reading of the name (alternative reading Parnow, see Zgusta KPN:449,n. 6, Blümel KarPN:24)

Adiego (1993a:250)

“ arnw ≤ E.AS 3PN in genitive. Perhaps formed on the same stem as “ arnai-/“ arnaj- (seepreceding entry).

Adiego (1993a:250)

“ arpt ≤ E.Ab 33PN in genitive.

“ arwljat ≤ E.Me 3PN in genitive. A compound formed by “ ar- (cf . “ arkbiom, “ aru≤ ol ) andwljat- (q.v. ). No Greek adaptation of the name has been found to date(a form such as *Sar-uliatow, *Sar-oliatow is the most likely possibility).

Adiego (1993a:242–243).

“ arur ≤ E.Ab 37PN in genitive.



418

“ aru≤ ol E.Me 30, E.Ab 6, E.Ab 30 (“a[ru]≤ol ), E.SS 1

PN in nominative. Carian name that appears in Greek sources asSarus(s)vllow. A compound name formed by Sar- (cf. “ arkbiom ) andu≤ ol (q.v. ).

Ray (1981:155, 161).

“ aru≤ [ . . . ?E.Ab 42If complete, a PN in genitive. But it is more likely to be an incom-plete form of the noun “ aru≤ ol (see the preceding entry).

“ asqariod C.Hy 1a

“ a ÿdiq ≤ E.Ab 30PN in genitive. A compound name “ a- + ÿdiq , perhaps a variant of

the name in the following entry.

On this explanation, see pp. 262–263.

“ ayriq E.Me 25PN in nominative. It corresponds to the Carian name from Greeksources, Saurigow. A compound name “ a- (~ “ ar ?) + yriq . For the firstelement, cf. Sa-ussvllow. The second element is the well-known stem yri ∞-/yriq- / ÿdiq-/ ÿd ∞- (see idyri ∞-, paryri ∞-, etc.).

Adiego (1993a:250), (1994a:44). On this family of names, see pp. 262–263.

“ a [--]i ≤ b? wnE.AS 5

“ d t at ≤

E.Me 13PN in genitive.



419

“ enurt E.Me 50a

PN in nominative. It probably corresponds to the Carian name Sanortowin Greek sources.

Adiego (1993a:236), and with some doubts about the identification (1994a:43).

“ i “≤ E.Th 35PN in genitive?

“ od ubr ≤ C.Kr 1PN in genitive. It seems to belong to the family of names in -(d)ybr-/-(d) ÿbr-, but in this case u, and not y, is used.

Could this name be the Kaunian version of k ≤ at ÿbr (= Lycian Janduberiw )?For a > o, cf. otonosn and the following entry.

“ oru≤ C.Ka 3PN in genitive.

Assuming an a > o change (cf. otonosn and the preceding entry), a compari-son with the Carian name Sarow could be feasible (for the adaptation of aCarian u-stem in Greek as a thematic one, cf. pau- = Paow ).

“ rb˚[ - ]sal

E.Th 49

“ rquq E.Lu 2 (“?rquq), C.xx 1“ rquq ≤ E.Me 43a, E.Me 44aPN in nominative (“ rquq ) and in genitive (“ rquq ≤ ). A compound nameformed by “ r- = “ ar-/Sar- (cf. “ ar+u≤ ol -Sar+ussvllow, “ ar+kbiom, “ r+wli- )+ quq (= Gugow in Greek sources).

Adiego (1993a:243). A name †arquq (Adiego 1993a:241, 1994a:35) does notexist: all the alleged examples are actually misreadings of “ rquq .



420

“ rwli ≤ E.Me 20

PN in genitive. Can be analysed as a compound “ r- (cf. “ r-quq ) + wli-,very probably the same stem behind wliat/wljat (q.v. ). Moreover, wli-can be directly compared to the Isaurian name Oualiw (Zgusta KPN§ 1134–3/4). Cf. also Pisidian Oliw, Zgusta KPN § 1086–1.

See Adiego (1993a:243).

“ÿin≤ E.Me 38, E.SS 1PN in genitive.

- ÿin- recalls -yin in [--]ryin (the Carian form of the dynast name Idrieus), whereit is analyzed as a form of the ethnic suffix -yn-/- ÿn (see s. v. [--]ryin

tabC.Ka 5

tamosi E.Ab 18, E.Ab 19, E.Ab 21tamosi ≤ E.Ab 20PN in nominative (tamosi ) and genitive (tamosi ≤ ). Vittmann suggests thatwe can recognise here the Egyptian name Pt -ms (literally ‘Ptah is born’),note particularly the old Akkadian adaptation, Ta ¢ma ““ i .

Vittmann (2001:43). However, note that there is no further evidence to sup-

port pt > t in Carian adaptations of foreign names.

taqbos E.xx 6PN in s-ending, coordinated with “ arnaj-s by means of sb, ‘and’.

ta “ ubt ≤ E.Me 18aPN in genitive.

tazomd [ C.Ki 1



421

tbridbd ≤ E.Me 42

PN in genitive.

tdu≤ ol E.Me 24PN in nominative. It belongs to the u≤ ol -family of names, but the firstelement, td-, is not clear. It is perhaps to be related to ted-, ‘father’?

Schürr (2003b:69, n. 1) sugggests considering initial t as a mistake and thatinstead we should simply recognize here the same name as dw ≤ ol -.

tebot E.Th 28 [teb%ot ], E.Th 44

tebwnqmw E.Th 38

ted

E.Me 38Common noun in nominative: ‘father’. Compare CLuw. t àta/i-, Lyc. tedi-,Lyd. taada- ‘father’. Note the apparent umlaut *a > e as in Lycian, whichimplies an original stem in -i- or with i-mutation (*tadi- > *tedi- > ted- ).

Schürr (1996[98], already suggested in 1996a:68), Hajnal (1997a:210).

temazi C.Eu 2

ter ÿez ≤ E.Me 4PN in genitive.

tkra b i ≤ E.Me 37PN in genitive.

tksr E.Lu 7



422

tmonks E.Th 41

tñu≤ C.Hy 1aPN in genitive. The name appears adapted in Greek as Tonnouw in oneof the inscriptions that accompanies C.Hy 1.


tqlow E.Th 32

tqtes E.Me 47PN in ‘s-case’ (or rather a nominative of an s-stem?).

trdy≤ C.My 1

PN in genitive

triel ≤ E.Me 26PN in genitive.

triqoE.Me 6PN in nominative.

trq d imr C.Ki 1Sequence that contains the god name trq(u)d -, ‘Tarhunt’ (see following entry and armotrq d os ). Neither the internal structure (perhaps it must besegmented into two words trq d imr or trq d i mr ) nor the function andmeaning (a PN or place name, or the god name proper?) are clear inthis case.

See Blümel-Adiego (1993:94–95), where possible connections for imrº and mrºare proposed (imrº = CLuw. im(ma)ra/i- [but note that the normal correspon-dence of this latter word in Carian is (i)b r-!], ºmr = Lyc. mere- ‘laws’).



423

trqud e C.Ia 3

God name: trqud - = Hitt. Tar ¢u-, CLuw. Tar ¢unt-, Lyc. trqqñt-, theAnatolian Storm God. Unclear ending: perhaps a dative?See also trq d imr, armotrq d os.

Blümel-Adiego (1993:94). trqud e as dative: Adiego (1994a:38, 50). The simplestinterpretation is to assume that trqud e is the divine name to whom the craterawhere C.Ia 3 is inscribed was dedicated, (Blümel-Adiego 1993:95), although theoverall context of the inscription is unclear and makes it difficult to give aprecise analysis (Melchert 2002:310).

t ® ∞atar ≤ E.Me 34 (t®∞ata[r]≤ ), E.Me 41t ® ∞atr ≤ E.Th 5PN in genitive.

The attempt to compare t ® ∞at(a)r- with Tarhunt- and the Carian place nameTarkondar/a/ (Adiego 1992a:34, 1994a:43; see also Janda 1994: 175, who

interprets the word directly as an ethnic name) is hampered by the unsuit-ability of the sound correspondences, particularly if compared with trq(u)d -,the usual form of Tarhunt- in Carian. The interpretation as an ethnic name,feasible for the examples from Memphis, is questioned by the clear PN(patronym) from Thebes, see p. 269.

tsial C.My 1 (2× )PN in nominative.

ttbazi E.Ab 41ttbazi [ ≤ ]E.Me 1ttubazi E.Ab 25Feminine (?) PN in nominative (ttbazi, ttubazi ) and genitive (ttbazi [ ≤ ]).As suggested by Schürr, an adaptation of the Egyptian name T3-dj(.t)b3st.t

(literally ‘the one (fem.) whom (the goddess) Bastet has given), GreekTetobastiw.

See Schürr (1996a), for details about the reading of the inscriptions in question.



424

tumnE.Sa 1

Accusative of tum-, a Carian adaptation of the Egyptian god name Jtm(Atum)?

For this interpretation, see Adiego (1993a:255) and particularly (1995:21–23).

tur [ G 1Beginning of a PN. It corresponds exactly to the truncated name thatappears in the Greek part of the bilingual G 1 (Tur[ ).

Adiego (1992a:33).

tusol ≤ C.My 1PN in genitive. The final ol of the stem could correspond in Mylasa tothe typical ending -vll /-vld- in Carian names. However, the namehardly can belong to the family of the u≤ ol -/ Ussvllow names, given

the use of s , and not ≤ , as would be expected.

ty∞[ C.My 1

For a very hypothetical interpretation, see Adiego (2005:92–93) and here onp. 308.

t ÿn

C.Ha 1

See p. 283.

t [ - ]rsi C.Si 1

t amou

E.Me 7PN in nominative, an adaptation of the Egyptian name that appearsin the hieroglyphic part of this bilingual inscription as Ô 3j-Ó p-jm=w (literally ‘may Apis take them’ *[ ‘i˙pimòw]. But the Carian adaptationin fact corresponds to Ô 3j-jm=w or Ô 3j-n.jm=w , Greek Tamvw, Yamvw,



425

Samauw, Samv#w, in this case an abbreviated form (phonetically *[ ‘amòw])of the abovementioned name (DNb:1348–1349).

Ray (1981:58). See Vittmann (2001:55–56) for a discussion of the relationshipbetween t amou and the pure Carian name “ amow .

t anai ≤ E.Me 7PN in genitive. Ray suggests that it may be an Egyptian name: *Ô 3-n-n3-j w , literally ‘off spring of the (sacred) cows’.

Ray (1994:200). Vittmann (2001:56, n. 87) notes that the name is not attestedin Egyptian.

t rel E.Bu 1

w E.Me 13

Complete word, abbreviated form, or a simple mistake? See p. 272.uantrpoE.Ab 12uantrpu≤ E.Ab 13PN in nominative and genitive. It is very likely, but not completelycertain, that both forms belong to the same paradigm (therefore withan alternation -o / -u≤ ; the parallel with -e / -i ≤ in l ÿ∞se/l ÿ∞si ≤ is striking).

uarbe E.Th 1PN in nominative.

uarila [ - ]os ≤ E.Ab 39Apparently a PN in genitive.

w #ar [---]t [------]i [---]≤ E.Me 11aRemains of an onomastic formula N-Ø . . . N-≤.

426

%wdbo≤ knE.Th 47

wdwnE.Th 13

Cf. owdown[ (E.Th 10). Janda (1994:182–183) observes the striking resemblanceof wdwn/owdown[ to the Pisidian sequence oudoun, but this comparison can-not be taken further due to the similarly unclear contexts in which Carianand Pisidian forms appear.

ue E.Me 3, E.Me 5, E.Me 28, E.Me 29 ([ue]), E.Me 42, E.Me 51‘Funerary stela’, or similar. It seems to be similar or correspondent toupe/upa , but the precise relationship between the words (if it indeedexists) is not clear.

See below s. v. upe about upe/ue connection. As indicated there, Schürr’shypothesis of a loss of p (upe > ue ; Schürr 1992:141; 1993:172) is attractive,but ad hoc .

uejresi E.Si 2

wet ≤ E.Me 13PN in genitive.

uioml nC.Ka 5[ ui ? ]oml ã

C.Ka 2Cf . also yoml n, very likely to be a variant.Probably a verbal finite form, corresponding in some way to Greek¶doje, ‘It seemed good’, in the bilingual inscription C.Ka 5. However,the precise analysis remains unclear. An alternative view, suggested byMelchert, is to analyze it as a noun with the meaning ‘decree’.

All the analyses take as a starting point the example of the bilingual C.Ka 5,contextually more clear (note moreover that the value of the final letter of theC.Ka 2 example, here transcribed as <ã>, is far from being assured).



427

Frei-Marek (1997:30) propose a third plural person of a past tense with themeaning ‘to decide’, whose subject would be kbidn, interpreted as ‘the Kauniansdecided’. Both Hajnal (1997b:151–153) and Neumann (1998:30) suggest thatthe verb must correctly be ºml n, and try to connect the verbal stem ml - withdiff erent Hittite verbs: malai- ‘to approve’ (Neumann), mald- ‘solemnly pro-nounce, vow’. Similar analysis and etymological connection given in Adiego(2002:20).

Eichner’s interpretation (only partially deducible from references apud Tremblay1998:117, 123) concurs in isolating ml n as a verb. He translates kbidn uio ml nas ‘il plaît/plaisant à la communauté des Cauniens’ with uio interpreted as adative sg (‘communauté’).

For Melchert’s view, see Melchert (1998:37): kbidn uioml n ‘decree of Kaunos’(kbidn: place name, plurale tantum, here in genitive). In a supplementary note,Melchert suggests very tentatively a connection of uioml n with CLuw. wayam-man-, ‘cry, howl’, cf . also Hitt. wiyài-, ‘to cry’, assuming a semantic develop-ment comparable to Lat. proclamatio to ‘proclamation’. However, in Melchert(md ane ), Hajnal’s view is preferred: he isolates ml n and analyzes it as a preteritethird plural from, *mVld-onto, to be related to Hitt. mald-.

uksi E.xx 7PN in nominative (?)

uksmuE.Me 2wksmu≤ E.Me 36PN in nominative (uksmu ) and genitive (wksmu≤ ). Cf . the Anatolian namesOuajamoaw (Zgusta KPN § 1141–2, Isauria, Cilicia), Ouajamvw (ZgustaKPN § 11141–2, ibid .). This compound name therefore contains as a

second element the well-known Anatolian stem muwa-, ‘strength, force’.It is very likely that the individual alluded to could be the same inboth inscriptions.

Neumann apud Adiego (1993a:236). On the identification of both individuals(both show the same father’s name), see Masson (1976:38), Ray (1982b:187).See here p. 336.

ula [----]ol

C.St 1



428

uliade C.St 2

PN in nominative. Carian adaptation of the Greek name OÈliãdhw,very widespread in Caria, probably due to its resemblance to the purelyCarian name wliat (see the following entry).

Adiego (1994a:39–40). On the name OÈliãdhw see Masson (1988b).

wliat E.xx 2wljat E.Th 7wljat ≤ E.Mu 1PN in nominative (wliat/wljat ) and genitive (wljat ≤ ). It is the Carianname that appears in Greek sources as Uliatow or Oliatow. Note thecompound name “ arwljat-.

For the identification, see Adiego (1992a:31). The stem of the name has been

connected with Hittite walliwalli-, ‘strong, powerful’ (also on the basis of otherAnatolian names: Walawala, Walawali , and particularly Carian Oaloalow, aboutwhich see Adiego 1993b), cf . Adiego (1993a:238). See here p. 339.

umot 2C.Ka 2

unC.Ka 5 (2× )

Tentatively analyzed as an infinitive in Adiego (1998a:22), see here on p. 299.Segmentation in both cases is not guaranteed!

und o[--]t l “ C.Ka 5Acc. pl. c. of a stem und o[--]t l -. It seems to correspond to Greek eÈrg°taw‘benefactors’ in the bilingual inscription C.Ka 5.

Already proposed in Frei-Marek (1997:38), who added the important obser-vation that the sequence ºt l - could represent a suffix of a nomen agentis com-parable to Hitt. -talla-.



429

In Adiego (1998a:22) a segmentation un d o[--]t l “ is suggested, taking un asan iteration of un (infinitive: ‘to make’) that appears just before (see preced-

ing entry), so that onlyd o[--]t

l “ was the word corresponding to eÈrg°taw.

untri E.Ab 12, E.Ab 13PN in nominative.

wnuti ≤ E.Me 31Genitive. According to Vittmann, this is a title rather than a PN, given

the good correspondence to Egyptian wnwtj (*[w6nùti/e] or *[wnùti/e])‘hour-observer, horoscoper, astronomer’. Vittmann rightly observes thatthe absence of an article in the adaptation of the Egyptian title fits wellwith the Egyptian syntactical practice, consisting of omitting the arti-cle when the title precedes the noun that it qualifies (in E.Me 31, wnuti ≤ is the first word of the inscription, followed by the personal name kwar ≤ ).

Vittmann (2001:48–49). See here on p. 278. Vitmann is reasonably cautiousin suggesting this interpretation, but the correspondence Carian wnuti = Egyptian

wnwtj cannot simply be a matter of chance.

uodrouC.St 2

uodryia [ C.St 2

upa E.Me 13upe E.Me 17, E.Me 4, E.Me 9, E.Me 22, E.Me 26, E.Me 38, E.Me 43a,E.Me 64a ([. . . u? ]pe)wpe E.Me 36, E.Me 41Common noun in nominative: ‘(funerary) stela’, or ‘tomb’. Its connec-tion with ue , used in similar contexts, is not clear.

Perhaps somehow related to Lycian x upa ‘tomb’?



430

It is clear that upe/upa , independently of its precise meaning, makes referenceto the object where the inscription stands (‘funerary stela’), or to its function(‘tomb’): see Meriggi (1980:36), followed by Adiego (1993a:208). E.Me 26clearly supports this explanation, where upe appears accompanied by the demon-strative sa , ‘this’ (Adiego 1993a:209). Other interpretations (‘son’, Ray 1982b,followed by Kammerzell 1993; ‘I am’, Ray 1990a:72; a demonstrative, Rayibid .) must be ruled out.

Schürr has argued in favour of a generic meaning ‘stela’ instead of ‘tomb’,assuming that E.Me 4 is a ‘stèle de donation’, and that the Carian text doesnot seem to contain a typical onomastic formula (Schürr 1992:155). However,

pace Schürr, this stela has clearly been re-used (there is no connection betweenthe Egyptian and Carian texts), and the Carian inscription poses serious read-ing problems that do not allow us to identify the type of formula used.

For upe/wpe vs. ue , Schürr has proposed a purely phonetic explanation, byresorting to an alleged tendency in Carian to drop p in intervocalic position(Schürr 1993:172; however, the evidence adduced is not convincing).

uqsi E.Me 20PN in nominative. Perhaps a simple graphical variant of PN uksi-?

urm≤ E.Bu 1wrm≤ E.xx 7PN in genitive. Note also the “vocalized” variant urom≤ in E.Bu 2, whichpossibly alludes to the same person as urm≤ in E.Bu 1.

On the possible connection of this name to Luw. ura-, ‘great’, an adjectivethat enters into the formation of Anatolian proper names, see above p. 338.

ur #oE.Th 34

urom≤ E.Bu 2PN in genitive. See urm≤ , wrm≤ .

urq

E.Lu 6

431

ursea ∞k E.Bu 6

urs ∞le ≤ E.Me 15PN in genitive. It appears transcribed as 3rskr in the Egyptian part of the bilingual inscription E.Me 15.

Final º∞le of the stem has led us to imagine a Carian adaptation of a Greekname in -kl∞w (Neumann ers. comm. suggested ÉOrsikl∞w ). The use of palatal∞ for Greek k recalls Lyc. k (also a palatal) in Perikle < Gr. Perikl∞w.

urt E.Th 34PN in nominative?

u≤ bzol C.Hy 1aPN in nominative.

u≤ ol E.Ab 35u≤ ol ≤ C.Hy 1a (2× ), C.St 2w ≤ ol ≤ E.Me 12PN in nominative (u≤ ol ) and genitive (u≤ ol ≤ /w ≤ ol ≤ ), corresponding tothe Carian name that appears as Ussvldow, Ussvllow in Greek sources

(Zgusta KPN § 1629–7/8, Blümel KarPN:27); u≤ ol - enters in compo-sition with a series of prefixes ( pn-u≤ ol , “ ar-u≤ ol , id-u≤ ol , etc.)

For the identification u≤ ol = Ussvllow, one of the most decisive steps in theCarian decipherment, see Ray (1981:160). About the possible etymology of the name, see here p. 344, n. 16.

usot C.Ke 1, C.Ke 2

On the doubts about these inscriptions, see p. 150.



432

utnu≤ E.Ab 19, E.Ab 21 (u?tnu≤? )

PN in genitive.

uHbit C.Ka 2

u[ . . . ]ü≤ q E.Th 12

ya C.Ka 8

ÿas d ≤ E.Me 46aPN in genitive.

Initial ÿasº recalls the ethnic name (?) yiasi-, yjas[i]- ‘Iasean’ (see s. v. ); ÿ vs. yi/yj- finds a good parallel in y≤ biks vs.- yi ≤ ∞biks-. As for ºd -, it can be inter-preted as a nt-suffix.

ybrs ≤ C.Hy 1aPN in genitive. The stem ybrs- recalls the family of names in -(d)ybr-/(d) ÿbr-, particularly smd ÿbrs (see pp. 283, 314). It is possible that thisindigenous name was identified with the Greek name ÑUbr°aw (ZgustaKPN § 1624), commonly used in Caria, in a process similar to that of wliat-OÈliãdhw.


ÿbt C.xx 1Probably a verb: 3rd singular preterite or present of a verb ÿb- =Lycian ube- ‘to off er’: If preterite, it would be completely equivalent toLycian ubete , ‘he off ered’. If present, it would be from *ube-ti .

For this interpretation, see Melchert (1993:78–79). Melchert’s interpretation isfollowed by, among others, Adiego (1994a:240) and Hajnal (1995[97]). Analternative view was attempted by Janda, who prefers to consider pjdl as the

433

verb in C.xx 1; Janda (1994:179) suggests that ÿbt can be a particle chain ora noun corresponding formally to CLuw. upatit- ‘landgrant’ (< *‘donation’.The Carian word would retain this original meaning).

About the possibility that ÿbt can be a present, cf . Hajnal (1995[97]:17).

yiasi E.Me 25 yjas [ i ≤ ]E.Me 9Ethnic name (?) in nominative ( yiasi ) and genitive yjas [ i ≤ ]. Connectedwith the Carian place name Iasos (Iasow )?

Adiego (2004:310) and here p. 270.

yi ≤ ∞biks ≤ E.Me 46aPN in genitive. If the corrected reading is accepted (see p. 69), it is avariant form of ÿ≤ biks, q.v.

ylarmi t

C.Hy 1bEthnic name (in genitive plural?) referring to the Carian city of Hyllarima.

Connection of this word to the place name Hyllarima already noted in Ray(1988:152). For ylarmi t as genitive plural with the meaning ‘Hyllarimeans’, seeAdiego (2002:17).

ymezus [ C.St 2

ÿn-?-moE.Sa 1Cf . ÿnsmsos , although the integration of s between n and m is by nomeans clear.

ynemori ≤ E.Me 29

PN in genitive.



434

ÿnsmsos E.Mu 1

E.AS 3Possibly a title or adjective in nominative, where a suffix -os can beidentified (cf . kbos E.Me 24).

yoml nC.Ka 4Perhaps a variant form of uioml n, q.v.

ÿ pdnmw %d E.Th 4

yri ∞ñ C.Si 1 (2× )Final ñ (if the reading is accepted), points to a possible accusative. A possible stem yri ∞- would recall the family of names in yri ∞-/yriq- / ÿdiq-/ ÿd ∞- (see idyri ∞-, paryri ∞-, etc.).Can yri ∞ñ have any connection with the GN Sinuri?

yri ≤ E.Th 26PN in genitive.

Cf . pnyri ≤ ru. Could yri- be related to yri ∞-/yriq-?

yrqso≤ C.My 1 (2× )

PN in genitive. It corresponds to the name adapted in Greek as Urgosvw.

Adiego (2005:90)

ÿrsbe E.Ab 6Unclear word. Cannot be a PN in nominative, given the context inwhich it appears (preceded by a PN in nominative (“ aru≤ ol ) and fol-

lowed by a PN in genitive ( pdubi ≤ ). Perhaps a title?



435

ÿ≤ biks C.xx 2

PN in nominative. Compare yi ≤ ∞biks ≤ . The name is a compound whosesecond element is biks (cf . piks-, dbiks- and see p. 337 for an etymolog-ical explanation).

ÿsmE.AS 9

yysmt ≤ oHa [ E.Ab 27

zidks E.Sa 1Sequence immediately followed by md ane . Function and meaning unknown(an ‘s-case’ of a PN? A verb?).

zmu≤ E.Me 19

PN in genitive.

z “ arios ãC.Ka 2Coordinated by means of sb to another word ending in -ã ([ - ]∞ar l anoã ).

Hd “ qe d ormñs [ C.St 2

H∞it C.Ka 2

Hnmk d a [-]a Huq [ C.Ki 1

HorouoC.Ka 5

Function and meaning unknown.

In Adiego (2000:144) attention is drawn to the good parallel between ºorouoand Lycian arawa- ‘tax exemption, ét°leia’, but the connection is hamperedby the initial H (difficult to separate from orouo, given that Horouo appearsimmediately after sb, ‘and’).



436

Hosurz E.Ab 28

[41]1aitk C.Ka 2Function and meaning unknown. Perhaps k could be an enclitic ele-ment, and the resulting form 1aitº could be compared with ait in thesame inscription, for which an analysis as a third plural person verbhas been proposed (see ait ).

1mali C.Ka 2 (2× )

1orsol “ C.Ka 2Apparently an accusative plural, coordinated with sarni “ by means of sb. Meaning unknown.

?-˚bjqmqew E.Th 12Function and meaning unknown. Perhaps we are dealing with morethan one word. Note the final ew (see ewm ), and the sequence ºkbjº,which recalls the PN kbjom≤ , “ arkbiom and place name kbidn.

?-ras E.Ab 43PN in nominative (?)

A W

[ - ]a HC.Ka 2Cf. a similar ending (]ma H ) in the same inscription.

[ - ]ars E.Ab 36Apparently a PN in nominative.



437

[ - ]bd oC.My 1

PN in nominative. No parallels have been established.

[ - ]bi C.Ka 2Perhaps the same enclitic element that appears twice in the same inscrip-tion, see bi .

[ - ]d iurt C.Ka 2

[ -?]iam≤ E.Me 45PN in genitive.

[ - ]intnor C.Ka 2

[ - ]∞ar l anoãC.Ka 2Coordinated by means of sb to another word ending in -ã (z “ arios ã ).

[ - ]nudrma C.Ka 2

[ - ]obiokli ≤ C.My 1Onomastic formula consisting of a PN in nominative followed by a PNin genitive, the difficulty being the isolation of the two names.

[ - ]owt ≤ E.Me 48PN in genitive.

[ - ]qo

C.My 1PN in nominative. For the final, cf . plqo-Pellekvw.

[ - ]tmai ≤ [--]E.Bu 3



438

[ - ]untlau[ - ]E.Ab 22

[--]abrun∞ur [ - ]“ yn“ H ynnC.Ka 2A sequence of words whose segmentation is not clear.

Neumann (1998:26–27) proposes isolating a word ∞ur [-]“ yn≤ , which he inter-prets as an ethnic name in plural (with the - yn- ethnic suffix, cf. kbdyn“ ) andconnects to the Carian place name Koar(r)end/a/, Koarenz/a/ Koaranz/a/,Kvra( n )z/a/. Although the proposed segmentation is attractive (particularly

regarding the ending -yn-“ ), the relation with the mentioned place name is farless compelling; note that the Carian word cannot be read integrally: also thecorrespondence Car. “ : Greek - nd-/-( n ) z- (in the diff erent variants of the name)is not at all satisfactory.

[--]e ∞l d E.Si 2

[--] j [ - ]≤

E.Me 48Extant letters of a PN in genitive.

[--]msal E.Bu 1Last letters of an initial word in a graffito from Buhen. Final ºsal recallseuml ? bnasal, eypsal , also initial words in Buhen graffiti. See euml ? bnasal fora proposed interpretation of ºsal .

[--]ryinC.Si 2aPN in nominative. Given the context in which it appears, it must bethe Carian name of the dynast Idrieus. It is clear that the Carian formof the name was not totally coincident with the Greek one. One couldtentatively complete *[ id ]ryin and assume a PN formed on the name of the Carian city Idriaw, *idr-, by means of the suffix for ethnic names-yin (cf . kbd-yn-“ , mda-yn/mda- ÿn ). The resulting meaning, ‘inhabitat of

Idrias’, could roughly correspond to the Greek ÉIdrieÊw.Schürr (1992:137–138).



439

[--]Hl ∞sasot 2C.Ka 5

The final -ot 2 recalls similar endings in this inscription and in C.Ka 2:mdot 2, punot 2, umot 2.

[--]w ≤ ord ≤ E.Me 10Final part of a PN or ethnic name in genitive. For the ending, cf . theethnic name (or PN) msnord-

[--] yt 2C.Ka 5

]a [ - ]i ≤ E.Me 11bExtant letters of a PN in genitive.

]al l ia E.Ab 30

]bewmsmnwdiq E.Th 38A sequence that seems to contain more than one word. We could per-haps isolate a word smnwdiq , probably a PN related to the family of names in yri ∞-/yriq- / ÿdiq-/ ÿd ∞- (Greek -urigow, -udigow ) (but note herethe apparent use of w ). As for ]bewm, it recalls the element ewm pre-sent in other graffiti from Thebes (see the corresponding entry).

]bt d eoE.Th 14

]b ? e ≤ C.Ia 6Probably the final letters of a PN in genitive.

]dar 1

C.Ka 4

]i ≤ E.Me 57Final part of a PN in genitive.



440

]∞≤ E.Me 60

Very likely to be the final part of a PN in genitive.

]latmne ≤ C.St 1PN in genitive. It is possible that the name may be complete. In thiscase, cf . perhaps the Carian place name Latmow (Zgusta KON § 696,Blümel KarON:173). The segmentation ]la tmne ≤ would also be anattractive theory, as well as a comparison of this latter with the CarianPN Tumnhw (Zgusta KPN § 1615, Blümel KarPN:26).

]ma HC.Ka 2Cf . [ - ]a H in the same inscription.

]no≤ ?

C.Ka 7Last letters of a PN in genitive?

]ois ? ur ? ml oC.Ka 9

] pri E.Ab 26Probably the last letters of a PN in nominative.

]q ≤ si ≤

E.Me 53Final part of a PN or ethnic name in genitive.

]rbn“ a [ E.Si 9

]r ≤ wk [ - ]“ [ E.AS 8

]r [—]tnit E.Ab 29



441

]se l “ C.St 1 (3× )

Very likely to be the final letters of a PN in nominative.

]s ≤ E.Me 58Final letters of a PN (s-stem) in genitive.

]tbe ≤ C.Si 1It seems to be the final part of a word (perhaps a PN?) in genitive.

]tbsms C.Ka 2

]ub“ÿC.Di 1

]ud a [

C.Ki 1

]ue ∞l C.Ia 2

]u≤ E.Me 26Last letters of a PN in genitive.

]u≤ ou≤ C.Ka 4

]utr [ E.Me 59

] ybzsdmC.Ki 1

]zol ba ∞a [ . . . ]C.Ki 1



Sa = SaisMe = MemphisAb = AbydosTh = ThebesLu = LuxorMu = MurwàwSi = SilsilisAS = Abu Simbel

SS = Gebel Sheik SuleimanBu = Buhenxx = Unknown origin

E.Sa = Sais

E.Sa 1“arkbiom : zidks mdane : ÿn-?-mo | den : tumn

E.Sa 2pdnejt q ÿri≤ ∞i

E.Me = Memphis

E.Me 1ttbazi[ ≤ ] | Ñiub[a]Ωi≤ | aor[ ≤ ]

E.Me 2uksmu | lkor≤ | mrsi≤

E.Me 3pikre≤ ue“arwljat≤ msnord≤

E.Me 4ter ÿez≤ | upe | nuol +∞[---]sarmrol ∞ yt

E.Me 5psm“kwneit≤ | ue | naria≤ | ≤ugl iq |sarl?

E.Me 6triqo : parma≤≤ ∞iklorul ∞i

E.Me 7

tamou

tanai≤ qarsio[-?]

E.Me 8a. paraeym : armon ∞ib. para!eym : sb polo

E.Me 9arli“≤ : upe : arlio[m≤ ] ∞i : yjas[i≤ ]

E.Me 10œårm≤ : q[---]≤ ∞i : p∂uüi≤ mno≤[mw]don≤ ∞[i ---]w≤ord≤ ∞i

APPENDIX A

CARIAN INSCRIPTIONS IN TRANSCRIPTION

E = EC = C ( L L )G = G

E = E



446

E.Me 56[. . .] “ark[bi/jom . . .?]

E.Me 57[. . .]i≤ ∞i

E.Me 58[. . .]s≤ ∞i

E.Me 59[. . .]utr[. . .]

E.Me 60[ . . . ]∞≤

E.Me 61[. . .]i

E.Me 62[ . . . ]≤[ . . . ]

E.Me 63(a) idyes≤(b) m [?

E.Me 64(a) [. . .u? ]pe : pd[ (b) [. . .]mi

E.Me 65u[. . .]m | punm[-]≤ | mudo[n]≤

E.Me 66---].[..u][. . .]p[-]n[---

E.Ab = Abydos

E.Ab 1pisiri

E.Ab 2panejt iarja≤

E.Ab 3ptn“e | ibarsi≤

E.Ab 4“amow ltari≤

E.Ab 5“amow ltari[ ≤ ]

E.Ab 6“aru≤ol | ÿrsbe | pdubi≤

E.Ab 7pl at | pals≤

E.Ab 8pl at pals≤

E.Ab 9pl at pals≤

E.Ab 10piubezqurbo≤

E.Ab 11≤? / [. . .]it

E.Ab 12untri uantrpo

E.Ab 13untri | uantrpu≤

E.Ab 14abrq∞[ . . . ?

E.Ab 15pdubez or≤

E.Ab 16nprosn≤



447

E.Ab 17pa[-]in[-]t≤

E.Ab 18tamosi | inut≤

E.Ab 19tamosi utnu≤

E.Ab 20ninut | tamosi≤

E.Ab 21to[-]a[---]l tamosi u?tnu≤?

E.Ab 22[-]untlau[-]|

E.Ab 23be≤o

l

E.Ab 24[. . .] arli“

E.Ab 25ttubazi katt ÿri≤

E.Ab 26[. . .]pri | ptnuq?i?

E.Ab 27 yysmt≤oHa[

E.Ab 28Hosurz | srton[-]_[. . .?](or: . . . +t[-]nota/rs | za/rusoH/l ?)

E.Ab 29[. . .]r[--]tnit

E.Ab 30bid≤lemsa : “a[ru]≤ol : “a ÿdiq≤[. . .]al

l ia : bsis

E.Ab 31∞aye

E.Ab 32∞arr≤

E.Ab 33“arpt≤ | p[-]lu≤

E.Ab 34dbkrm [-]kb?[

E.Ab 35u≤ol | mi∞≤≤ kdûsi≤

E.Ab 36[-]ars,

∞[-]urb≤

E.Ab 37“arur≤

E.Ab 38piew

E.Ab 39uarila[-]os≤

E.Ab 40iall i | q∞blio≤

E.Ab 41ttbazi kt?tri≤

E.Ab 42“aru≤[..?

E.Ab 43?-ras



448

E.Th = Thebes

E.Th 1uarbe

E.Th 2dt ÿbr | kbokt≤k≤at ÿbr

E.Th 3pla?t

E.Th 4dokmmpint seqqejewsk | mqt jq

ÿpdnmwd.

E.Th 5d ÿbr | t®∞atr≤

E.Th 6bebnd

E.Th 7wljat

E.Th 8qutbe

E.Th 9kudtubr

E.Th 10a?q≤baq ewm ≤emotqtblo owdown[. . .]mwarudk≤o mlane

E.Th 11psma≤[k] [?| nm[ mplat | o[

E.Th 12?-˚bjqmq ewmlane qeb≤t | u[. . .]ü≤q| qwsal | mqabaewleqo“osk)ioms

E.Th 13dbiks | kbjoms | wdwn | sb a≤b≤tewm

E.Th 14 ]q[. . .]btdeo

E.Th 15Very uncertain reading!

E.Th 16∂saml-?-?-” (vacat ) dy “a

E.Th 17ku

E.Th 18t n

E.Th 19dbikrm

E.Th 20orbá ˚ r i“

E.Th 21mmn∞al

E.Th 22mwk | te

E.Th 23bebi

E.Th 24kow[?-?]

E.Th 25ktmno

E.Th 26brsi yri≤



452

Kn = KindyeKi = KildaraHa = HalikarnassosDi = Didyma (Ionia)Ia = IasosKe = keramosKa = KaunosKr = Krya (Lycia)xx = Unknown origin

C.Tr = Tralles (Lydia-Caria border)

C.Tr 1sdi amt[ pau≤art mon

C.Tr 2an sidi a-rtmi pau≤parãaq?

C.Al = Alabanda and surroundings

C.Al 1 (Eski Çine)sdi a[-]mob[

C.Eu = Euromos

C.Eu 1

≤as : ktais idyri∞≤ : mn[os?

]

C.Eu 2omob ∞i : temazi≤dun : ≤o≤niabkolarmon qyrbmudolomanon

C.Kn = Kindye

C.Kn 1pare ÿs

C.Hy = Hyllarima

C.Hy 1(a) “asqariod dymdamuo

tarmotrq

dosq

brsi ari“≤ brsi≤mane : u≤ol ≤rtim u≤ol ≤ pur?i≤

u≤bzol tñu≤ brsi≤pau mane≤ ybr-s≤

(b) kdu≤opizipususot

mol“ msot ylarmit

C.My = Mylasa

C.My 1idrayridsemdbq mol“ ty∞[ tsial tusol≤ : moi m[-]sao[ banol paruos≤ : p?au paryri∞≤qzali obrbi≤ : tsial obrbi≤banol yrqso≤ : paryri∞ psoir≤

[-]bdo pnu≤o≤ : myze trdy≤“arkbiom qzali≤ : ≤umo kbdmu≤skdubrotoz≤ : pau ∞toi≤[-]qo idyri∞≤ : ksbo idu≤ol≤[-]obiokli≤ : ∞toi yrqso≤

C.Si = Sancutary of Sinuri nearMylasa

C.Si 1adymd“ : yri∞ñ : t[-]rsi : [. . .?]tbe≤(vacat)

yri∞ñ : binq : sñaidl o

C.Si 2(a) [--]ryin ∞tmño≤ : sb ada ∞tmño≤eri : pisñoi mda : pñmnn≤ñ : pda-∞m≤uñ ∞i “aoyr∞ri mt∞elã

ñmail o mda lrHñ : stspñ vacat sm“s[--5--] sb añmsñsi mdasm[--7--]a∞e[ ∞[--8--]tuñdñ[ ñe-?-[



453

(b) pim[. . .]Ha?[ . . . ]

C.Ki = Kildara

C.Ki 1[ . . . . . . . ( . ) ] zol ba∞a[. . (.)] kil [ [. . .]uda[. . .] trqdimr qrds tazomd[ kil arad[-]ybzsdmHnmkda[-]aHuq[ iasoum

C.St = Stratonikeia

C.St 1 ]sel “ a[--]a[----]om≤ ]som[n? ]e brsi≤ ula[----]ol ]latmne≤ ≤ ysñal [ ] ari“ maql y≤[ ]sel “ piks[

]sel “ p[

C.St 2u≤ol ≤ uodrou u[ mute≤ ymezus[ ∞diye≤ uodryia[ uliade pidaru[ mañ“qaraH≤rl -?-[ dar“qemorms[ Hda“qedormñs[

C.Ha = Halikarnassos

C.Ha 1smd ÿbrs | psnl o | md orkn t ÿn | snn

C.Di = Didyma (Ionia)

C.Di 1 ]ub“ÿ

C.Ia = Iasos

C.Ia 1 ]la

limtaoa | [ om

C.Ia 2 ]ue∞l | ∞ob[

C.Ia 3?] are“ | “anne ml ne | siyklo≤ | “ann| trqude | ∞l mud [?

C.Ia 4n[. . .]pr[. . .]is[. . .]

C.Ia 5baqg k[. . .]

C.Ia 6[ . . . ]b?e≤

C.Ke = Keramos

C.Ke 1uso-t

C.Ke 2uso-t

C.Ka = Kaunos

C.Ka 1sñis : sdisa-s : psu≤ol ≤mal≤ : mno≤

C.Ka 2

[ui? ]oml ã qrds g rdso[-]i[ [-]r sb a∞mnnartnyr obsmns[ [-]∞arl anoã sb z“ariosã i∞[ [-]nudrma ∞ yrpai sarni“ sb u[ [-]aH punot2 otr“ bi sb a∞tmsk[m



454

[-]d bi 1aitk ouor gdb“l aã1_i[-][-] sarni“ sb1orsol“ sb uHbit

[-]bi qrdsol “ ait 1_mali H∞it[-]intnor ∞ yrapai≤ umot2 oba[-]diurt obsmsmñ1 ñ ouor mt1 yr[--]abrun∞ur[-]“ yn“H ynn sb vacat [------------------]tbsms 1mali [[-------------------]maH sb an[ [ . . . . . . . . . . . . . . . . . . . . . . . . ] b a vacat

C.Ka 3“oru≤

ann ibrs≤

C.Ka 4[. . .]u≤ou≤ ibrsdr[-][. . .]a yoml n1 r_i[. . .]dar1_ idym“

C.Ka 5kbidn uioml n i[---]

inis drual nik[--]lan lysiklas[-? ]otonosn sb lys[ikl]an lysikratas[-? ]otonosn sarni[ “ ]mdot2 un sb undo[--]tl “ kbdyn“ sb b2o[--]ol“ otr“ sb a∞t[ms]-kmt absims sb [---]

yt2 oru sb a∞t[----]bu

∞ y[-----]i[-----]i

[--]≤ un moa[-]l boror[--]Hl ∞sasot2 orttab sb ort[-] sb Hor-ouo bi mslmnliapurmoruos mnosaitusi

C.Ka 6or

C.Ka 7 ]no≤? (or rather : ]noñ?)

C.Ka 8potko≤l ≤? aba?d?

ya

C.Ka 9[. . .]ois?ur?ml o

C.Kr = Krya (Lycia)

C.Kr 1qot2omu sdisam?n≤ “o

dubr≤ or rather : sn≤ “o

dubr≤?

sb mno≤ knornoril?ams or rather : norimams?

C.xx = Unknown origin (presumablyfrom Caria)

C.xx 1“rquq | qtblem≤ | ÿbt | snn | orkn| ntro | pjdl?

C.xx 2 ÿ≤biks not : alosd ∞arnosd : jzpe mdane

C.xx 3akymydu ÿeryl[ vacat ]d

C.xx 4kdu≤ol“

C.xx 5kdu≤ol“

G = Greece

G 1 (Athens)≤ jas | san tur[

G 2 (Thessalonike)qlal i≤ | k?[

êla ‘horse’bãnda ‘victory’

Stephan of Byzantium, s. v. ÉAlãbanda:

ÉAlãbanda, pÒliw Kar¤aw (. . .) kt¤smad¢ KarÚw ∑n, épÚ toË paidÚw aÈtoËk l h y e › s a t o Ë g e n n h y ° n t o w é p ÚKallirrÒhw t∞w Maiãndrou, metå n¤khnflppomaxikÆn, ka‹ klhy°ntow ÉAlabãndou,˜ §sti katå tØn Kar«n fvnØn flppÒnikow.êla går tÚn ·ppon, bãnda1 d¢ tØn n¤khnkaloËsin.

Stephan of Byzantium, s. v. ÑUlloÊala:

ÑUlloÊala, d∞mow Kar¤aw (. . .) êlagår ofl Kçrew tÚn ·ppon ¶legon, …w ka‹prÒteron e‡rhtai.

g°la ‘king’soËa or soËan ‘tomb’

Stephan of Byzantium, s. v. Souãggela:Souãggela, pÒliw Kar¤aw, ¶nya ı

tãfow ∑n toË KarÒw, …w dhlo› ka‹toÎnoma. kaloËsi går ofl Kçrew soËantÚn tãfon, g°lan d¢ tÚn basil°a.

g¤ssa ‘stone’

Stephan of Byzantium, s. v. MonÒgisaMonÒgisa, <pÒliw Kar¤aw> (. . .)

g¤ssa går tª Kar«n fvnª l¤yow •rmh- neÊetai. ka‹ nËn toÁw plak≈deiw ka‹malak≈deiw l¤youw g¤ssa l°gousi.

kÒon ‘sheep’

Scholia ad Il. XIV, 255: tÚ d¢ prÒbatonkÒon (ms. ko›on ) ofl Kçrew Ùnomãzousin,˜yen K«w ≤ poluyr°mmvn.

Cf. Eustathius, ad. Hom. Il. XIV,

255: K«w (. . .) fas‹ d¢ toÁw KãrawoÏtv kale›n tå prÒbata, ˜yen ka¤ ı n∞sow K«w …w poluyr°mmvn.

1 Reading mãnda in the two best manuscripts of Stephan.

APPENDIX B

CARIAN GLOSSES



APPENDIX C

CARIAN NAMES IN GREEK SOURCES

A. P N (§: Z KON )

Aba § 2–1.Agan/a/ § 8.

Aggvr/a/ § 65Agorhsow § 10.Adhssow § 17–3.Ayumbra § 25.Alabanda § 37–4.Alikarnassow § 44–7.Alinda § 44–11.Alkizv § 45.Allianoi § 44–4.Allvss/ow/ § 50–2

Amnist/ow/ cf. § 820, BlümelKarON:165.Amow § 60.Amuzvn, Amuzvn/a/ § 61–1.Amunand/a/ § 61–2.Andanow § 66–3.Anyemi § 71.Apodes[ Blümel KarON:165.Arara § 85–16.Ardur/a/ § 90–6.

Arissoullh Blümel KarON:165.Arlai/a/ § 95–1.Arlissow § 95–2.Armel /a/ Blümel KarON:165.Armokodvka § 96–1.Arnaso/w/ § 97–3.Arpasa Lat. Harpasa § 98.Artoub/a/ § 100–2.Asshsow § 108–3.Babein § 122–3.

Bargasa (var. Pargasa, Bargaza )§ 135–1.Bargulia (later var. Barbulia) § 135–2.Barkok≈mh Blümel (KarON:166).Boll- § 158.

Bridaw § 173.Bubassow, Boubassow, Bubastow,

Boubastow § 177.Bvnitv § 159–2.Bvrand/a/ § 181.Gerga § 202–1.Gordio/ n/ § 215–3.D°dmasa § 250.Didassai § 263.Dundason § 281.Enn/a/ § 297.Eorma § 298.

Erezow § 302–1.Ermapilow § 305–1.Yaruai § 335.Yasyar/a/ § 336.Yembrihmow, Yembrimow § 338.Yemhs(s)ow, Yemissow § 339–1.Yigrow § 343.Yudonow § 351.Yuhssow § 352–1.Yumbria § 353–3.

Yussanouw § 355.Yvdasa Blümel (KarON:167).Iasow § 358.Idriaw § 363.Iduma § 364–1.Imbrow § 373–1.Io.d- § 378.Kaduih § 403–2.Kalbisso/w/ § 413.Kalunda § 414–2.

Kandasa § 426–2.Kandhb/a/, Kendhb/a/ § 428 (s. v.Kanduba ).

Kanhbion § 430.Kaprima § 436.



457

Karbasuand/a/, Karpasuand/a/§ 439–5.

Karoura § 452–2.Karu/a/ § 454–3.Karuanda § 454–1.Kasa § 455–2 (s. v. †Kasaio/ n/).Kasar/a/ § 455–7.Kastabo/w/ § 458–1.Kasvk/a/ § 461–1.Kasvlaba § 461–2.Kasvsso/w/ § 461–4.Ka. nar/a/ § 423–6.Kebialea § 471.Kelimara Blümel (KarON:169).Kemhsso/w/ Blümel (KarON:169).Kendhbocorow § 477.Kenendvlab/a/ § 479.Keni- § 480.Kepranow § 481.Keraskord/a/ § 486.Kecaro/w/ § 496–1.K¤dram/a/ § 501.Kildar/a/, Killar/a/ § 510.Kinduh, later var. Kunduh § 518–1.Kisariw § 522.Koarbvnd/a/ § 538–1.Koar(r)end/a/, Koarenz/a/

Koaranz/a/, Kvra( n)z/a/ § 538–2.Kodap/a/ § 541.Kodouvka § 544.Kozanata § 547.Koliorg /a/, Kolierg /a/ § 552.Koloura § 558–1 (not in Blümel!).Komurion § 658.Komvond/a/ § 568.Kond- Blümel (KarON:170).Konodvrkond/a/ § 575.Korell /a/ § 581.Kormoskvn/a/ § 583–4.Korrit/a/ § 589.Kostobalo/w/, Kvstobalo/w/ § 662.Kot- Blümel (KarON:171).Kot/a/ § 593–1.

Kourb/a/ § 607.Crusa § 631.Kruassow § 632–2.Kuarda § 635.

Kubassow § 636.Kubim/a/ § 639–1.

Kubisyih § 639–2.Kubliss/ow/, hublis/ow/ § 1396.Kullandow § 645–1.Kumniss/ow/ § 646.Kumvr/a/ § 647–2.Kuogrissiw § 649.Kuon § 652–2.Kuprand/a/ Blümel (KarON:172).Kurbasa § 651–2.Kuw § 652–1.Kushr/a/ § 653–1.Kusshliw § 653–2.Kvrai/vn/ § 659.Labara § 665.Labraunda (var. Larabiunda,

Labranda, Labrainda, Labrauunda,Labrenda, Lambraunda Labraenda )§ 666.

Lagina, Lageina § 670Lagnvk/a/ § 671.Laras/a/ § 688–2.Larb/a/ § 689.Latmow § 696.Leibo/w/ (rather a person or god name)

§ 704.Leukoid/a/ Blümel (KarON:173).Lec- (epiclesis ZeÁw Lecunow, Lecinow.

cf. also Lepsia , a Carian island) § 706.Lhcimandow (diff erent variants for the

inhabitants’ name: Lefsimãnioi,Lefs¤mandoi, LefsimandÇew,LefsumandÇew, LhcimandÇew) § 707.

Lidh § 712–1.Lobold/a/ § 716.Lor.os/a/ § 720.Lurisso/w/ § 731.Lvm/a/ § 736.Lvndarg /a/ § 737.Lvndokvmh Blümel (KarON:173).Lvr/a/ § 738.Lvruma § 739.

Lvsso/w/ § 740.Madnas/a/, M°dmasow, Medmassa/

Mednassa § 748.Mali/a/ § 756–2.



458

Masanvrada § 782.Massvn/a/ § 787.

Mastaura § 788.Maunn/a/, Maiunn/a/ § 793.Messaba § 804–1.Mhyasai § 806.Milhtow § 809.Mniesu/a/ § 819.MÒboll /a/, Mogola, Mvgla § 822.Mokold/a/ § 828.Monnara pl § 832.Monogissa § 833.Mosoun/a/ § 842–1.Mugissow § 858.Mudon- § 859.Muhss/ow/ § 863.Mulas(s)a (very late form: Milasa )

§ 861–1.Mundow § 862.Murshl /a/ § 866–1.Mursileia § 866–2.Mvss/on/ § 871.Naras/a/ § 885–3.Narisbar/a/ § 886.Naruandow § 888.NinÒh § 898–1.Nouik/a/ Blümel (KarON:175).Jerasso/w/ § 907.Ogond/a/ § 912.Ol /a/, Oul /a/ § 925.Oloss/iw/ § 930–2.Olumo/w/ § 932.Omb/a/ § 934.Ondoura § 935.Orbhla § 938–4.Oryondouvk/a/ § 941.Orsubli/a/ Blümel (KarON:176).Orsvll /a/ § 950.Otvrkond/a/ § 958.Ouasso/w/ § 966.Palgosvlda § 996.Panamara § 1000.Pandaj/a/ § 1001–2.

Parableia (Parablia?) § 1005.Parembvrd/a/ § 1007.Parkall /a/ § 1009.Parpar- § 1012.

Pasand/a/, Pasada § 1015.Passala § 1018.

Patarous/a/ § 1022–3.Pedanass/ow/, Pedanass/ow/ § 1028,1059–2.

Peig°lasow § 1031.Peldek-ì t- § 1035.Phgasa § 1053–1.Phdasa, Pidasa § 1054–1.Piginda § 1058–1.Pidvssow, Pidossus § 1059–2.Pisymoi § 1065.Pisiliw § 1066–1.Pisuh § 1066–4.Pitaon, Pitaium § 1067–3.Pladas/a/, Platas/a/ § 1068, 1072.Plamow § 1070.Plarasa § 1071.Pluar/a/ § 1080.Poluara § 1083–1.Pounomou/a/ § 1093–1.Prinassow § 1101–1.Proposs/ow/ Blümel (KarON:178).Pruondr- Blümel (KarON:178).Purindow § 1114–1.Purnow § 1114–2.Pustow § 1116.Salei/a/ § 1148–2.Salmakiw § 1150–1.Samn/h/ § 1153.Samulia § 1152–2.Saranso/w/ § 1165–2.Sasanda § 1176–2.Siana § 1261–1.Sikim/a/ Blümel (KarON:178).Sind/a/ § 1219–2.Sindhssow § 1219–4Sinuri § 1222.Solo/a/ § 1244–1.Solvn/a/ § 1244–3.Sparz/a/ § 1255.Suana § 1261–1.Suaggela, Souaggela, Sfaggela (later

Yeaggela, Theangela ) § 1261–2.Suarbeu[ § 1262.Suista § 1267.Surna § 1272.



459

Svbala § 1274.Svssow Blümel (KarON:179).

Tabai § 1277–1.Tabarniw § 1277–5.Talagr/a/ § 1284.Tapass/oi/ § 1294.Taramptow § 1295.Tarban/a/ § 1297–1.Tarbetv (?, or rather a personal name?)

Blümel (KarON:179).Tarkondar/a/ § 1299.Tarm/ow/ (?, or rather a personal

name?) § 1300.Tezhra (?, or rather a personal name?)

§ 1310.Teleseitiw § 1312.Telmhssow, Telmessow, Telmissow,

Telemessow, Telmisum § 1314.Temoesso/ow/ Blümel (KarON:180).Tendhba § 1318.Termera, Telmera, Termera § 1320–2.Terssvgass/ow/ Blümel (KarON:181).Tnussow § 1347.Traldeiw, Tralleiw, Trallis § 1361–1.Trara § 1362.Trobaliss/ow/ § 1368.Truban/a/ § 1374.Tuennesso/w/ § 1379.Tumnhssow § 1384–5.Tumnow § 1384–4.

Ualvka § 1393. Uarbesu/a/ § 1394.

Ubliss/ow/ § 1396. Ugas(s)ow 1397. Udai, Kudai § 1398–2. Udis(s)ow § 1398–4. Uyubir/a/ § 1400. Uissow § 1402. Ulim/a/ § 1404–1. Ullarima § 1404–2. Ullouala § 1404–3. Umess/ow/ § 1405. Urvmow, Kurvmow, Eurvmow, Eurvpow

§ 1412. Usarbid/a/ § 1414. Ussome[ Blümel (KarON:182). Utarm/ow/ Blümel (KarON:182). Xalkhtvr, Xalkhetorew Blümel

(KarON:182).Vlasha § 1443.Vndr/a/ § 1444.Vnzvssuaso/w/ § 1445.Vspraonno/w/ § 1447.

]akondia Blümel (KarON:182). ]eadovka § 1450. ]erra Blümel (KarON:182). ]hvka § 1457. ]kermu.ion Blümel (KarON:182). ] nirea Blümel (KarON:182).

B. P N (§; Z KPN)

Aba, Abaw, Abbaw § 1–1/2/3/5.Abersi § 5.Ada § 15–1.Ayuasiw § 128.Akarmomeldvw § 27.Aktadhmow § 38–1.Aktauassiw § 38–2.Aktaussvllow § 38–3.Alasta § 42–1.

Alganiw § 44.Alleaw § 52–3.Amiaw, Ammiaow, Ammh, Ammin, Ammeiaw,

etc. § 57.

Andarsvw § 59–4.Appa, Apfia, Apfiaw, Apfion, Apfianow,

Afia, Afion, Affion, etc. § 66.Apoukvw § 79.Arbhs(s)iw § 85–2/3.Arduberow § 86–6.Ariauow § 89–2.Aridvliw § 89–4.Arlissiw § 95–1/2.

Arlivmow § 95–3.Arris(s)iw § 106–1/2.Arshliw § 107–12.Artaow Blümel (KarPN:11).

460

Arthumow § 109.Artimhw Arteimhw § 108–4/6.

Artuassiw, Aryuassiw § 110–1/2.Aruassiw § 111.Arvsiw Blümel (KarPN:11).Atthw § 119–10.Beryaw § 162–1.Berrablviow Blümel (KarPN:11).Boivmow § 178–4.Bruajiw, Bruassiw § 196–1/2.Brvlvw § 197.Geiw § 210–3.Glouw § 224.Gugow Blümel (KarPN:12).Dandvmow § 251.Daru . . . ow § 254.Deibow § 264.Dersvmanhw Blümel (KarPN:12).Dersvw Blümel (KarPN:12).Dersv . . . tiw § 275.Ekamuhw Blümel (KarPN:12).Ekatomnvw (more recent variants:

Ekatomnvn, Ekatomnow ) § 325–1/2/3.Ejamuhw § 340–2.Ermapiw § 355–21.Zermeduberow (var. Jermedu<be>row )

Blümel (KarPN:13).Zonzolow § 390.Yekuilow § 417.Youw Blümel (KarPN:14).Yualdiw § 438.Yussow § 445–1Ibanvlliw § 450.Idagugow § 451–4.Idakow § 451–5.Idbelaw Blümel (KarPN:14).Ideghbow Blümel (KarPN:14).Idmamu . . . ow § 452–6.ÉIdrieÊw § 453.Idubl[ § 454.Idussvllow Blümel (KarPN:15).Il[.]uthw § 1678.Imbarhldow § 467.

Imbarsiw, Imbras(s)iw § 469–3/4/5.Imbrhw § 469–9.Innivn § 471–8.Indow § 473–1.

Isedum.xow § 485.Isemenda . . . ow § 486.

Iublhsiw § 494.I . . . uagow § 1679.Kay.divn § 1680.Kakraw § 509–1.Kalbalaw § 512.Karama . . . ow § 531.Karjaw § 539.Karreiw § 540–3.Karusvldow § 544.Kasballiw § 546.Kasbvlliw § 545.Kashsiw § 547–1.Kaifenh § 558.Kbondiassiw § 566.Kbvdhw § 567–1.Kebivmow Blümel (KarPN:16).Keldnassiw § 573.Kemptuw § 575.*Kendhbhw § 576–9.Ketambissiw § 593.Kindacow (KarPN:16).Kinjimow § 617.Kit.essvw (KarPN:16).Koboldvow (KarPN:16).Koibilow § 652–1.Koidvw § 653.Kolaldiw, Kulaldiw (KarPN:17).Koldobaw § 660.Kolvldow § 661.Kondalow § 676–1.Kondmalaw § 676–2.Kondo[ § 676–5.Korollow (KarPN:17).Korriw § 686–3.Koshtiow (KarPN:17).Kosinaw § 703.Kostvlliw § 705.Kotbelhmow (KarPN:17) cf. Kutbelhmiw.Kotobalvw (KarPN:17).Kouldoiw § 727.Kourvn § 737–5.

Ktouboldow § 761.Kuaremow § 764.Kuatbhw § 765.Kulaldiw (KarPN 18) cf. Kolaldiw.



461

Kutbelhmiw § 771.Kutpiw § 772.

Kvbhw § 774–1.Kvglvw § 775.Latarshw § 799.Lugdamiw § 834.Lujhw § 836.Makow § 848–1.Malosvow (KarPN:18).Manhw § 858–1.Manitaw § 864–1.Mareuw § 873–6.Marow § 873–13.Massarabiw § 880–2.Matiw § 882–6.Mausvllow § 885–1–6.Memakow Blümel (KarPN:19).Metebiw Blümel (KarPN:19).Mindrvn § 921.Minnaw, Minnion, Minniw, Minnh,

Minnivn, Minnow, etc. § 922.Miskow/Miskvw § 929–2.Mohnnow § 941.Moiw § 942.Mokollhw § 944–1.Molhw § 946–1.Monnhw § 959.Mosraiow Blümel (KarPN:19).Mouzeaw § 980–2.Naduw § 1008a.Nana, Nanh, Nanaw, Nannh, Nannion,

Nanniw, Nannixow, Nannow, Nannv, etc.Narbaw § 1013.Neterbimow Blümel (KarPN:20).Nonnh, Nonnow, etc.Nutar Blümel (KarPN:20).Nvtrassiw Blümel (KarPN:20).Oaloalow § 1134–2 (cf. Adiego 1993b).Oa3a3iw § 1145–8.Obrokaw Blümel (KarPN:20).Oletaw § 1085–1.Oliatow § 1085–2Olohtow Blümel (KarPN:20).

Opinaw § 1096–2.Ordomaw § 1104–3.Oridhumiw Blümel (KarPN:20).Ortassiw § 1114–2.

O (?sa )rthumow § 1114–3.Oseaw § 1121.

Ouvkhw § 1180.Pagadow § 1186.Paktuhw § 1193.Patkuvlliw Blümel (KarPN:21).Paktuiskow Blümel (KarPN:21).Panablhmiw § 1197–3.Panamuhw § 1197–6.Panuassiw § 1198.Paow Blümel (KarPN:21).Papaw, Papiaw, Papow, etc. § 1199.Paparivn § 1200–2.Paraskvw § 1203–3.Paraudigow § 1203–5.Paraussvllow, Paraussvldow,

Parussvldow § 1203–6/8.Pargistaw § 1205.Parmumiw Blümel (KarPN:22).Parnow § 1207.Paruv § 1212–2.Paruinna § 1212–1.Passidhrow § 1219.Pedvldow § 1232.Pelaow Blümel (KarPN:22).Peldemiw, Peldemvw

§ 1234–1.Peldhkow § 1234–2.Pel(l)ekvw § 1234–3/4.Pelkisiw § 1235.Perbilaw § 1239.Perignaw/Petignaw § 1241.Phdisaw § 1249.Pigassvw Blümel (KarPN:22).Pigrhw § 1255–6.Pijvdarow § 1263–3.Pirvmiw § 1266.Pisindhliw § 1268.Pisku[ Blümel (KarPN:23).Pis. nvw Blümel (KarPN:23).Pitakolow (not †Gitakolow, Zgusta

§ 221!) Blümel (KarPN:23).Pittaw § 1270–1.

Plouw § 1277–1.Ponmoonnow Blümel (KarPN:23).Ponusvllow § 1189.Pormounow Blümel (KarPN:23).



462

Purkeaw/PurkehwSaggotbhriw Blümel (KarPN:23).

Saggvw § 1369.Samassiw § 1361.Sampaktuhw, Sambaktuw § 1364–1.Samvow § 1367–1.Samvuow § 1367–2.Sanamvw Blümel (KarPN:24).Sanortow § 1371.Sarow § 1377.Saruassiw Blümel (KarPN:24).Sarussvllow § 1378–1/2.Saskow/Saskvw § 1381.Sassvmow § 1379–6.Saurigow Blümel (KarPN:24).Sausvllow Blümel (KarPN:24).Seikilow § 1390.Semeuritow Blümel (KarPN:25).Senurigow Blümel (KarPN:25).Seskvw § 1410–1.Sesvlhw § 1411.Sibilvw § 1416–1.Siduatow Blümel (KarPN:25).Sidulhmiw § 1422.Silbow § 1426.Skoaranow Blümel (KarPN:25).Spareudigow § 1466.Sueskurebow § 1477.S[u]s[k]h[w] (?)Blümel (KarPN:25).Suskvw § 1486.Svmnhw Blümel (KarPN:25).

Tarmow Blümel (KarPN:25).Tarv § 1515–2.

Tata, Tatarion, Tataw, Tateiw, Tath,Tatia, Tatiaw, Tation, etc. § 1517.Tausaw § 1520.Tendessiw § 1534.Territow § 1538.Tiaimow § 1533.Tobororow § 1577.Tounobow § 1592.Totoliw § 1598.Trus(s)hw § 1608.Truvlhw/Truvlow § 1609.Tumnhw/Tumnow § 1615.

Uarkelaw Blümel (KarPN:26). Uyesmaw Blümel (KarPN:26). Uyhw Blümel (KarPN:26). Uliatow § 1627. Urgaw Blümel (KarPN:26). Urgilow Blümel (KarPN:26). Urgosvw Blümel (KarPN:26). Ussaldow Blümel (KarPN:26). Ussaldvmow/ Usseldvmow § 1629–4/6. Ussiw § 1629–1. Ussisiw § 1629–2. Ussvihw/ Ussviow § 1629–3. Ussvldow, Ussvllow, Ussvlow

§ 1629–7/8. Xasbvw Blümel (KarPN:27). Xhramuhw (cf. § 1639).

Acephalic Forms (Blümel KarPN 27–28)

]alvldow ]anvrremow ]ketaw ]kokvw ]ldoudhw ]llv[do ]w]omvliw]ramow

]rgigougou

]rouessiw]ruassiw]teieow]toldiw]uassiw]ujki]vldow]vllow

]vrlemiw



463

Idumow Tischler (1977:66).Indos , (var.) Lindow (an erroneus

form?) Tischler (1977:67) cf. KON§ 375.

Kalbiw Tischler (1977:69).Kenivw (var. Kinevw ) Tischler (1977:78).Kitvn (var. Keitvn ) Tischler

(1977:80–81)Kubersow Tischler (1977:85–86), KON

§ 639–4).

Maiandrow Tischler (1977:93–94), Cf.KON § 752.

Marsuaw Tischler (1977:96–97).Morsunow Tischler (1977:102).Salmakiw (a source) Tischler (1977:128).Telmedius(?) Tischler (1977:143–144)Timelhw, Teimelhw Tischler (1977:143,

148). Cf. KON § 1338. Ubando/w/ Tischler (1977:64) Cf. KON

§ 1395.

ÖImbramow, var. ÖImbrasow (= Hermes;St. Byz. s. v. ÖImbrow, Scholia vetera inTheogoniam v. 338, and Eustathius,Commentarii ad Iliadem XIV, 281)

Mãsariw (= Dyonisus; St. Byz. s. v.Mãstaura )

ÉOsog«a (= Zenoposeidon; Strabo XIV,659, Pausanias VIII; 10, 4).

R N

G N L S



SaisE.Sa 1 MY LE.Sa 2 MY M

MemphisE.Me 1 MY A E.Me 2 MY BE.Me 3 MY DE.Me 4 MY EE.Me 5 MY F

E.Me 6 MY GE.Me 7 MY HE.Me 8 MY KE.Me 9 M 1E.Me 10 M 2E.Me 11 M 3E.Me 12 M 4E.Me 13 M 5E.Me 14 M 6E.Me 15 M 7

E.Me 16 M 8E.Me 17 M 9E.Me 18 M 10E.Me 19 M 11E.Me 20 M 12E.Me 21 M 13E.Me 22 M 14E.Me 23 M 15E.Me 24 M 16E.Me 25 M 17

E.Me 26 M 18E.Me 27 M 19E.Me 28 M 20E.Me 29 M 21E.Me 30 M 22E.Me 31 M 23



E.Me 52 M 44E.Me 53 M 45E.Me 54 M 45aE.Me 55 M 46E.Me 56 M 47E.Me 57 M 47aE.Me 58 M 47bE.Me 59 M 48E.Me 60 M 48aE.Me 61 M 48bE.Me 62 M 48cE.Me 63 M 48dE.Me 64 M 49E.Me 65 AbusirE.Me 66 Kammerzell

*180

APPENDIX D

CONCORDANCES

A. P F

EGYPT



465

AbydosE.Ab 1 Ab 1 FE.Ab 2 Ab 2a FE.Ab 3 Ab 2b FE.Ab 4 Ab 3b FE.Ab 5 Ab 3c FE.Ab 6 Ab 4 FE.Ab 7 Ab 5a FE.Ab 8 Ab 5b FE.Ab 9 Ab 5c FE.Ab 10 Ab 6 FE.Ab 11 Ab 7 FE.Ab 12 Ab 8a FE.Ab 13 Ab 8b FE.Ab 14 Ab 9 FE.Ab 15 Ab 10 FE.Ab 16 Ab 11 FE.Ab 17 Ab 12 FE.Ab 18 Ab 13a FE.Ab 19 Ab 13b FE.Ab 20 Ab 14 FE.Ab 21 Ab 15 FE.Ab 22 Ab 16 FE.Ab 23 Ab 17 FE.Ab 24 Ab 18 FE.Ab 25 Ab 19 FE.Ab 26 Ab 20 FE.Ab 20 Ab 14 FE.Ab 21 Ab 15 FE.Ab 22 Ab 16 FE.Ab 23 Ab 17 FE.Ab 27 Ab 21 FE.Ab 28 Ab 22 F

E.Ab 29 Ab 24 FE.Ab 30 Ab 25 FE.Ab 31 Ab 26a FE.Ab 32 Ab 26b FE.Ab 33 Ab 27 FE.Ab 34 Ab 28 FE.Ab 35 Ab 29 FE.Ab 36 Ab 8 YE.Ab 37 Ab 9 YE.Ab 38 Ab 15 Y

E.Ab 39 Ab 26 YE.Ab 40 Ab 27 YE.Ab 41 Ab 28 YE.Ab 42 Ab 29 YE.Ab 43 Ab 34 Y

ThebasE.Th 1 Th 47 ”E.Th 2 Th 48 ”E.Th 3 Th 49 ”E.Th 4 Th 50 ”E.Th 5 Th 51 ”E.Th 6 Th 52 ”E.Th 7 Th 53 ”E.Th 8 Th 54 ”E.Th 9 Th 55 ”E.Th 10 Th 56 ”E.Th 11 Th 57+58 ”E.Th 12 Th 59 ”E.Th 13 Th 60 ”

Luxor E.Lu 1 G 19E.Lu 2 G 21E.Lu 3 G 22E.Lu 4 G 23E.Lu 5 G 24E.Lu 6 G 25E.Lu 7 G 26

MurwàwE.Mu 1 ¥aba (1971), nº 196

SilsilisE.Si 1 Si 39 FE.Si 2 Si 53 FE.Si 3 Si 54 FE.Si 4 Si 55 F

E.Si 5 Si 56 FE.Si 6 Si 57 FE.Si 7 Si 58 FE.Si 8 Si 59 FE.Si 9 Si 60 FE.Si 10 Si 61 FE.Si 11 Si 62 F

Abu SimbelE.AS 1 AS 1

E.AS 2 AS 2E.AS 3 AS 3E.AS 4 AS 4 + Lepsius

Kar 4E.AS 5 AS 5

466

E.AS 6 AS 6E.AS 7 AS 7E.AS 8 AS 8E.AS 9 Lepsius Kar 2

BuhenE.Bu 1 M 50E.Bu 2 M 51E.Bu 3 M 52E.Bu 4 M 53E.Bu 5 M 54E.Bu 6 M 55

Gebel Sheik SuleimanE.SS 1 72 F

Unknown origin, likely to be fromEgyptE.xx 1 MY CE.xx 2 MY IE.xx 3 MY aE.xx 4 MY bE.xx 5 MY cE.xx 6 4 ”E.xx 7 Lion

CARIA

TrallesC.Tr 1 D 1C.Tr 2 D 2

Alabanda and surroundingsC.Al 1 D 13

EuromosC.Eu 1 D 3C.Eu 2 D 8

KindyeC.Kn 1 D 6

HyllarimaC.Hy 1 D 7 +

Adiego-Debord-Varinlio< lu 2005

MylasaC.My 1 Blümel-Kızıl 2004

SinuriC.Si 1 D 9C.Si 2 D 10

KildaraC.Ki 1 D 11

StratonikeiaC.St 1 D 12C.St 2 36*

HalikarnassosC.Ha 1 33*

DidymaC.Di 1 21*

IasosC.Ia 1 20a*C.Ia 2 20b*C.Ia 3 38a*C.Ia 4 38b*C.Ia 5 47*C.Ia 6 48*

C.Ia 7 Berti-Innocente 2005

KeramosC.Ke 1 39a*C.Ke 2 39b*

KaunosC.Ka 1 D 14C.Ka 2 D 16C.Ka 3 28*C.Ka 4 30*C.Ka 5 44*C.Ka 6 45*



467

C.Ka 7 46*C.Ka 8 49*C.Ka 9 50*

LyciaC.Kr 1 D15

Unknown origin, likely to be fromCariaC.xx 1 34*C.xx 2 35*

C.xx 3 40*C.xx 4 41a*C.xx 5 41b*

GreeceG.1 D 19G.2 42*

B. F P For the Egyptian inscriptions the order adopted in Meier-Brügger(1979b) is followed here.

EGYPT

Abu Simbel(Masson 1979)AS 1 E.AS 1AS 2 E.AS 2

AS 3 E.AS 3AS 4 E.AS 4 (+ LepsiusKar 4)

AS 5 E.AS 5AS 6 E.AS 6AS 7 E.AS 7AS 8 E.AS 8

Buhen(M = Masson 1978)

M 50 E.Bu 1M 51 E.Bu 2M 52 E.Bu 3M 53 E.Bu 4M 54 E.Bu 5M 55 E.Bu 6

Gebel Sheik Suleiman(F = Friedrich 1932)GSS 72 F E.SS 1

‘Pharaonic objects’(MY = Masson-Yoyotte 1956)MY A E.Me 1MY B E.Me 2

MY C E.xx 1MY D E.Me 3MY a E.xx 3MY b E.xx 4

MY c E.xx 5MY E E.Me 4MY F E.Me 5MY G E.Me 6MY H E.Me 7MY I E.xx 2MY K E.Me 8MY L E.Sa 1MY M E.Sa 2

‘Leningrad Isis’(” = ”evoro“kin 1965)4 ” E.xx 6

‘Lion’(Masson 1976)Lion E.xx 7

Memphis-Saqqara(M = Masson 1978)

M 1 E.Me 9M 2 E.Me 10M 3 E.Me 11M 4 E.Me 12M 5 E.Me 13



468

M 6 E.Me 14M 7 E.Me 15M 8 E.Me 16M 9 E.Me 17M 10 E.Me 18M 11 E.Me 19M 12 E.Me 20M 13 E.Me 21M 14 E.Me 22M 15 E.Me 23M 16 E.Me 24M 17 E.Me 25M 18 E.Me 26M 19 E.Me 27M 20 E.Me 28M 21 E.Me 29M 22 E.Me 30M 23 E.Me 31M 24 E.Me 32M 25 E.Me 33M 26 E.Me 34M 27 E.Me 35M 28 E.Me 36M 29 E.Me 37M 30 E.Me 38M 31 E.Me 39M 32 E.Me 40M 33 E.Me 41M 34 E.Me 42M 35 E.Me 43M 36 E.Me 44M 37 E.Me 45M 38 E.Me 46

M 39 E.Me 47M 40 E.Me 48M 41 E.Me 49M 42 E.Me 50M 43 E.Me 51M 44 E.Me 52M 45 E.Me 53M 45a E.Me 54M 46 E.Me 55M 47 E.Me 56

M 47a E.Me 57M 47b E.Me 58M 48 E.Me 59M 48a E.Me 60

M 48b E.Me 61M 48c E.Me 62M 48d E.Me 63M 49 E.Me 64Abusir E.Me 65Kammerzell (1993)*180 E.Me 66

Silsilis(F = Friedrich 1932)Si 39 F E.Si 1Si 53 F E.Si 2

Si 54 F E.Si 3Si 55 F E.Si 4Si 56 F E.Si 5Si 57 F E.Si 6Si 58 F E.Si 7Si 59 F E.Si 8Si 60 F E.Si 9Si 61 F E.Si 10Si 62 F E.Si 11

Thebes(” = ”evoro“kin 1965)Th 47 ” E.Th 1Th 48 ” E.Th 2Th 49 ” E.Th 3Th 50 ” E.Th 4Th 51 ” E.Th 5Th 52 ” E.Th 6Th 53 ” E.Th 7Th 54 ” E.Th 8

Th 55 ” E.Th 9Th 56 ” E.Th 10Th 57–58 ” E.Th 11Th 59 ” E.Th 12Th 60 ” E.Th 13

Abydos(F = Friedrich 1932)(Y = Yoyotte apud Meier-Brügger

1979)

Ab 1 F E.Ab 1Ab 2a F E.Ab 2Ab 2b F E.Ab 3Ab 3a F excluded



469

Ab 3b F E.Ab 4Ab 3c F E.Ab 5Ab 4 F E.Ab 6Ab 5a F E.Ab 7Ab 5b F E.Ab 8Ab 5c F E.Ab 9Ab 6 F E.Ab 10Ab 7 F E.Ab 11Ab 8a F E.Ab 12Ab 8b F E.Ab 13Ab 9 F E.Ab 14Ab 10 F E.Ab 15Ab 11 F E.Ab 16Ab 12 F E.Ab 17Ab 13a F E.Ab 18Ab 13b F E.Ab 19Ab 14 F E.Ab 20Ab 15 F E.Ab 21Ab 16 F E.Ab 22Ab 17 F E.Ab 23Ab 18 F E.Ab 24Ab 19 F E.Ab 25Ab 20 F E.Ab 26Ab 14 F E.Ab 20Ab 15 F E.Ab 21Ab 16 F E.Ab 22Ab 17 F E.Ab 23Ab 21 F E.Ab 27Ab 22 F E.Ab 28Ab 23 F excludedAb 24 F E.Ab 29Ab 25 F E.Ab 30Ab 26a F E.Ab 31

Ab 26b F E.Ab 32Ab 27 F E.Ab 33Ab 28 F E.Ab 34Ab 29 F E.Ab 35Ab 30 F excludedAb 8 Y E.Ab 36Ab 9 Y E.Ab 37Ab 15 Y E.Ab 38Ab 26 Y E.Ab 39Ab 27 Y E.Ab 40

Ab 28 Y E.Ab 41Ab 29 Y E.Ab 42Ab 34 Y E.Ab 43

Luxor (ESS 1998)G 19 E.Lu 1G 21 E.Lu 2G 22 E.Lu 3G 23 E.Lu 4G 24 E.Lu 5G 25 E.Lu 6G 26 E.Lu 7

Caria and other LocationsD 1 C.Tr 1

D 2 C.Tr 2D 3 C.Eu 1D 4 excludedD 5 excludedD 6 C.Kn 1D 7 C.Hy 1D 8 C.Eu 2D 9 C.Si 1D 10 C.Si 2D 11 C.Ki 1D 12 C.St 1D 13 C.Al 1D 14 C.Ka 1D 15 C.Kr 1D 16 C.Ka 2D 17 excludedD 18 coin legendsD 19 G.120a* C.Ia 120b* C.Ia 221* C.Di 122* excluded23* excluded24* excluded25* excluded26* excluded27* excluded28* C.Ka 329* excluded30* C.Ka 431* excluded32* excluded33* C.Ha 134* C.xx 1

470

35* C.xx 236* C.St 237* excluded38a* C.Ia 338b* C.Ia 439a* C.Ke 139b* C.Ke 240* C.xx 341a* C.xx 441b* C.xx 542* G.243* excluded44* C.Ka 545* C.Ka 646* C.Ka 747* C.Ia 5

48* C.Ia 649* C.Ka 850* C.Ka 951* excludedBelli-Gusmani 2001[rock inscriptionfrom Labraunda] excludedInnocente 2002[‘tegola di Iasos’] excludedBlümel-Kızıl 2004 C.My 1Berti-Innocente 2005 C.Ia 7Adiego-Debord-Varinlio< lu 2005 C.Hy 1[+D 7]



The following catalogue aims to include all known coins bearing let-ters in the Carian script.1 These coins form an integral part of the writ-ten record, and, in spite of their small size, they are of great importance,as they throw light on various aspects of Carian society and its lan-guage in particular. The material, collected over several years, comesfrom a variety of sources.2 While many coins are without a provenance,some have been found locally and are today housed near their findspots in museums and private collections. This may provide useful infor-mation for their attribution which in many cases remains a difficultmatter. Our purpose here, however, is to focus on the legends and

questions of attributions are only very briefl y discussed, especially whenspecific studies are available. The catalogue presents first coins which canbe ascribed to a mint, then coins whose attribution remains uncertain.

Coin legends are obviously related to inscriptions, but there aresignificant diff erences between them. Each coin issue was produced inthousands of specimens, even if today only a handful, in some casesone or two specimens, are extant. They were struck with dies whichhad to be individually engraved in negative. The engraver cut the mir-ror-like image of the design (type) and the letters on the die which,

when struck on a piece of metal, appeared in positive. Working in neg-ative could result in confusion in the direction and position of the let-ters. Carian was inscribed in either direction, though more often from

APPENDIX E

COIN LEGENDS IN CARIAN

Koray Konuk

1 Much of the discussion and many of the attributions presented here and elsewherewere first made public in a paper read at the Royal Numismatic Society in January1996 and entitled ‘Carian Coin Legends’ (hereafter ‘1996 RNS paper’), of which thisis an extended and updated version. I am very grateful to Professor Ignacio Adiego

for kindly including this appendix in his book and for his useful comments, and toRichard Ashton for improving my text.2 For nearly 70 years, Robinson 1939 remained the only comprehensive study of

coin legends in Carian (listing fewer than a dozen examples). Not only has our under-standing of Carian dramatically increased, but our documentation has also quadrupled.



472

left to right, especially on later inscriptions, and it is not always clearwhether on coins the right-to-left direction was meant or was simply

the result of the engraver’s confusion. In the description of the obverseand reverse of each coin, legends are first transcribed between brack-ets as they appear on the coin. When needed, questions pertaining tothe direction and reading of the legends are discussed.

One specimen has been cited to illustrate each variety. I have triedto select the best preserved specimen, which is not necessarily the oneillustrated in standard catalogues. The weight is followed, when known,by the die-axis. All coins are silver unless otherwise indicated.

I. M

Mylasa

For detailed discussion and attribution to Mylasa, see Konuk (forth-coming [a]). The mint of Mylasa was first suggested in my 1996 RNSpaper and appeared in print in Konuk (1998a:22–26) in which the last

two letters of M5 (my ) were read as the beginning of the ethnic of Mylasa in Carian. See also SNG Kayhan, 833–840. All silver fractionsare on the Milesian standard.

M1 M

Obv. Forepart of lion left; on its shoulder, O; below, one foreleg left.Rev. Two rectangular punches applied separately, one of which has

M1 (m ).Lydian (Persic) standard stater; c. 500 B.C.

Weber , 6448 (11.13g) = Naville 14 (1929), 378.

The Lydian-weight lion forepart issues were quite prolific and rankamong the earliest coinages of Caria. Their attribution is debated,Kaunos and Mylasa have been suggested (for an overview, seeKonuk 2000a:172 and Konuk forthcoming [a]). Several phasescan be observed which span the second half of the sixth centuryB.C. M1 comes late in the sequence of minting and is linked tothe issues which have various signs on the shoulder of the lion.

These are not letters but linear devices. M1 (m ), engraved in oneof the two rectangular punches, is attested on a few dies and,although the possibility cannot be entirely ruled out, it does not



473

appear to be a random mark. Subsequent issues attributed toMylasa carry letters only sporadically (see below), and most of its

early coinage is uninscribed. If M was intended as a Carian let-ter, it would represent the earliest occurrence of the initial of theCarian ethnic of Mylasa.

M2 M

Obv. Head of lion left; below, one foreleg left; dotted circle.Rev. Bird standing left, wings open; below left, M2 (m ); all within incuse

square.Tetartemorion; c. 500–450 B.C.Klein (1999:no 499) (0.24g) = Hauck & Aufhäuser 18 (2004), 250= Hirsch 187 (1995), 423.

This issue, of which three specimens are known to me (one is aprobable die duplicate in the Bodrum Museum of UnderwaterArchaeology, 1–20–78 [0.24g, 12H], the other is in Oxford,Ashmolean Museum [0.29g; 12H]; for a similar but anepigraphicearly example, see SNG Kayhan, 939), is to be linked to a sub-stantial series of Milesian-weight tetartemoria of the same type butlater style (e.g. SNG Kayhan, 940–948; many are housed in thearchaeological museums of Milâs and Bodrum). These have beentraditionally attributed to Miletos on the basis of the weight stan-dard and obverse type. This early issue, however, depicts the lionin a very diff erent way. Here, as on the Kayhan specimen, thelion’s head has a straightforward posture (as on M1), whereas onMilesian issues, the lion is depicted as a forepart with its headturned back and the foreleg reversed. The lion’s forepart evolved

on later issues and came to be modelled on the Milesian type. Inthis case there is no doubt that M2 (m ) represents the initial of the Carian ethnic of Mylasa.

M3 M

Obv. Forepart of roaring lion right, head turned back; below, reversedforeleg; linear outline of its back between jaws.

Rev. Facing head of a lion with its forelegs on either side, in upperright corner, M3 (m ); all within incuse square.

Tetartemorion; c. 450–420 B.C.Bodrum Museum of Underwater Archaeology, 3–26–91 (0.35g;07H).

474

Also part of an extensive series, only a handful of inscribed spec-imens is known. Among these a variety of the same early style

has with a smaller M standing next to the middle of the rightforeleg of the lion (Bodrum Museum of Underwater Archaeology,10–7–91 [0.45g; 06H]). As on the previous example, M3 (m ) standsfor the initial of the Carian ethnic of Mylasa.

M4 mObv. As last.Rev. Young male head (Apollo?) facing; in right field, M4 (m ); all within

shallow round incuse.Tetartemorion; c. 420–390 B.C.Künker 62 (2001), 129 (0.20g).

A variety features the letter in the left field (Oxford, AshmoleanMuseum (0.28g; 06H). For later issues, see M9 and M10 below.M4 is the early phase of a coinage which ends with the tetarte-moria in the name of the satrap Hekatomnos inscribed EK andEKA (see below).

M5 V mWObv. As last.Rev. Facing head of a lion with its forelegs on either side; below, M5

(w my ); all within incuse square.Hemiobol; c. 420–390 B.C.SNG von Aulock , 7807 (0.42g) = Troxell (1984:n° 1A).

The style of this coin looks later than M3 (same type). M, m andm are the same letter in the Carian alphabet (m ). The next letter

is somewhat problematic. Adiego suggests that W was V in thealphabet of Mylasa. If that is the case, the occurrence of both let-ters in the same legend calls for an explanation. Even though thenew inscription found near Mylasa (Kırca< ız) does not include theletter W, M5 proves that it was part of the city’s alphabet. Twoexplanations spring to mind: either these letters are not the sameand therefore have diff erent values, or V is not a letter but a tri-dent as suggested by Troxell (1984:250). When the trident expla-

nation was proposed (Troxell considered mW to be Greek letters),it seemed the most likely solution, as the Carian letter V wasunknown in that shape (apart from the rare occurrence of V atSinuri and Kildara) and no evidence existed at that time that it

475

was part of the alphabet of Mylasa. However, tridents are notnormally depicted on coins in such a simple way with mere strokes

(as with a letter): there is usually some ornamentation (arrow-liketips of the tines, often volutes departing from the base of the shaft)even when the size is minute. Moreover, the occurrence of Ú andV on M6–M10 below in the same position on the reverse stronglysuggests that V is a letter, for Ú is definitely a Carian letter, eventhough it is not attested on the new inscription from Kırca< ız. Iwould be tempted to suggest that Ú is a variant of V, the formerbeing perhaps an earlier form of the latter. If so, the value of Ú(w ) would represent the same value for V. Finally, what is thevalue of W which is also absent from the new inscription? Thesuggestion that W and V are the same letters cannot be main-tained on the basis of M5. Adiego gives W, a rather common let-ter, the value y. The two letters mW (my ) would thus plausiblyrepresent the first two letters of the Carian ethnic of Mylasa(Konuk 1998a:23). For a discussion of V and a possible attribu-tion to Hyssaldomos, see M9 below.

M6 VObv. As last.Rev. As last but M6 (w ); the incuse is round.

Hemiobol; c. 420–390 B.C.Pfeiler (1962:20, 2) = Konuk (1998:223, 130).

M7 Ú

Obv. As last.Rev. As last but M7 (w ).

Hemiobol; c. 420–390 B.C.Private collection (0.51g; 09H).

M8 Ú

Obv. As last but M8 (w ) on lion’s muzzle.Rev. As last but anepigraphic.

Hemiobol; c. 420–390 B.C.Muharrem Kayhan collection, MK1231 (0.50g; 03H).

The position of Ú on the lion’s muzzle reminds one of the Milesian-type tetrobols, diobols and obols of the Carian satrap Hekatomnoswhich have on the lion’s muzzle the Greek letters EKA, EK andE respectively. For a possible attribution to Hyssaldomos, see M9.



476

M9 Ú

Obv. As last.

Rev. Young male head (Apollo?) facing; in the lower right field, M9(w ); all within shallow round incuse.Tetartemorion; c. 420–390 B.C.Private collection (0.23g; 12H).

The letter is sometimes placed just below the facing head (e.g.New York, ANS, 1980.23.5 [0.22g] = Troxell [1984:no 2A], M9misdiscribed as a trident). These tetartemoria are distinguished bya diff erent reverse type. Tetartemoria of the same types but later

style were struck by Hekatomnos who put the Greek letters EKA or EK in place of Ú or V. In Konuk 1998a: 22–26, I suggestedthat Ú and (M5–M6 and M10) V (w ) might also be the initial of the name of a Carian dynast preceding Hekatomnos. His father’sname is Hyssaldomos and an attribution to him is quite likely.

M10 VObv. As last.Rev. Young male head (Apollo?) facing, turned slightly left; in the lower

left field, M10 (w ); all within round incuse.Tetartemorion; c. 420–390 B.C.New York, ANS, 1983.53.464 (0.23g; 12H)

M11 Ú

Obv. As last.Rev. Male head (Apollo?) right; in the lower left field, M11 (w ); all

within shallow round incuse.

Hemitetartemorion; c. 420–390 B.C.Private collection (0.14g; 10H).

Kasolaba?

For a likely attribution to the mint of Kasolaba, see Konuk (forth-coming [b]). Given the wide time-span during which these coins wereissued, probably over a century, the legend is bound to refer to an eth-nic rather than a dynast. M20 and M21 bear three letters: a9o (azo ),which at first glance are difficult to match with an ethnic. A large num-ber of these coins occur in the collections of the archaeological muse-ums of Milâs and Bodrum, and several find spots have been recorded,which fall in the area between Mylasa and Halikarnassos. By studying



477

the Athenian Tribute Lists and the recently discovered inscription fromSekköy, Descat (1994:66–68) demonstrates that the city of Kasolaba

ought to be located in that area. The reading azo shows a remarkablesimilarity to the Greek ethnic of Kasolaba. The omission of a gutturalinitial in the Carian legend should not be surprising since examples of ethnics like Kyromos / Hyromos / Euromos, Kydai / Hydai andKyblissos / Hyblissos in the same district testify that such variationswere frequent. Kasolaba is the Greek transcription of a Carian ethnicwhose native spelling remains uncertain. It has been suggested, how-ever, that ksolbz (ksolb≤ ) found in an inscription from Egypt (E.Me43) may have been the genitive form of Kasolaba in Carian. Whetheror not this is the case, the coins suggest that the Carian ethnic startedwith azo. It would not be far-fetched to expect that a new Greek inscrip-tion with the form Hasolaba may one day come to light. All coins areMilesian-standard hemiobols.

M12 36

Obv. Head of ram right.

Rev. Young male head left; to the left, M12 (az ); all within incusesquare.Private collection (0.43g; 09H); c. 450–400 B.C.

M12 and M13 show a distinctive archaic style; the deep incusesquare of the reverse also indicates an early issue. The initial let-ter, a on later issues, takes a rather odd shape on these earlyexamples.

M13 45Obv. As last.Rev. Young male head right; to the right, M13 (az ); all within incuse

square.SNG Keckman, 865 (0.29g; 06H) = Troxell (1984:no 8); c. 450–400 B.C.

M14 3“

Obv. As last.Rev. Young male head right; to the right, M14 (az ); all within incuse

squareHauck & Aufhäuser 15 (2000), 206 (0.52g); c. 450–400 B.C.



478

The shape of the initial letter establishes a link between the firstphase of this coinage and the later issues (M15 onwards).

Same types from the next coin onwards unless otherwise indicated:Obv. Head of ram right.Rev. Young male head right; on either side, M15 (za ); M16 (za ); M17

(za ); M18 (az ); M19 (az ); M20 (azo ); M21 (azo ); all within squareor round incuse.

M15 8 ABodrum Museum of Underwater Archaeology, 31–9–85 (0.42g;

01H).Only the incuse square variety is known. The letter on the rightalso has the shape M on a specimen in Oxford, Ashmolean Museum(0.43g; 12H). There is also a specimen with Q in the BodrumMuseum of Underwater Archaeology, 9–19–91 (0.51g; 12H).

M16 8 lPrivate collection (0.34g; 02H).

This specimen has a round incuse. The letter on the right alsohas the shape À on a specimen in SNG Keckman, 870 (0.46g; 09H;a die-duplicate is in the Ashmolean, Oxford).

M17 9 aMuharrem Kayhan collection, MK 1236 (0.49g; 06H).

The incuse on the Kayhan specimen is square; the round varietyis also attested (e.g. SNG Keckman, 869 [0.46g; 09H]). A variety

with 8 is known (SNG Keckman, 877 [0.39g; 03H]).

M18 a 9New York, ANS (0.41g; 06H) = Troxell (1984:no 9B).

On M18–M21, when the shape of the incuse can be determined,it is circular, and on some very shallow.

M19 a (obv.)

a 9 (rev.)Obv. Head of ram right; below, M19 (a ).

Private collection (0.36g; 10H).



479

M20 a 9oSNG Kayhan, 997 (0.38g; 12H).

M21 a 9o

Obv. Persian hero-king right, in running-kneeling position, holding dag-ger in the right hand and bow in the left; groundline.London (BM), CM 1999–10–7–1 (0.34g; 6H).

Keramos

For the attribution to Keramos, see Konuk (2000), based on the con-vincing attribution to Keramos by Ashton (1998) of a slightly earliercoinage of the same types bearing the Greek letters KE.

M22 kBo

Obv. Bull standing right.Rev. Dolphin leaping right; below, M22 (kbo ).

AE chalkous; c. 400 B.C.Private collection (1.05g; 04H) = Konuk (2000: no 2).

An obverse variety of this type has the forepart of the bull (Konuk2000: no 1). M22 (kbo ) is the beginning of the Carian ethnic of Keramos.

M23 _NW (obv.)luo (rev.)

Obv. Bull standing right; in front, M23 ( jy or kse ).Rev. Dolphin leaping right; underneath, M23 (kbo ).

AE chalkous; c. 400 B.C.SNG Kayhan, 804 (0.91g; 09H) = Konuk (2003:no 74) = Konuk(2000:no 5).

The reverse legend of M23 is upside-down. In the present form,the letters resemble a Greek delta, an upsilon and an omicron.The most likely explanation is a mistake made by a Greek die-cutter who, working in negative on the die, was led into error byhis mother-tongue. The variety with the obverse legend (which

might be Greek) appears to fall at the end of the series after whichCarian legend chalkoi were superseded by Greek legend chalkoi.



480

Kaunos

For a detailed discussion of this coinage, see Konuk (1998b). The definiteattribution to Kaunos was first presented in my 1996 RNS paper. Itwas based on the reading of the coin legends with the new values givento the Carian script. The initial of the ethnic on M24 and M25 (k )and the subsequent legends (M26–M28) giving a second letter (b ),resulted in kb which is the beginning of the native name of Kaunos:Kbid-, known from the Lycian version of the trilingual inscription of the Letôon. In summer 1997, the discovery of a bilingual inscriptionin Kaunos (C.Ka 5) gave for the first time the ethnic of Kaunos in

Carian and the first two letters on the new inscription were the sameas on the coin legends.The earliest coinage of Kaunos is anepigraphic and spans the period

c. 490–450 B.C. M24 (k ), the first occurrence of the beginning of theCarian ethnic of Kaunos, appears towards the middle of the fifth cen-tury B.C. The use of Carian ends with the chalkoi in c. 370 B.C.; theseare followed by chalkoi bearing the first three letters the ethnic of Kaunos in Greek.

The following coins are Aeginetic standard staters unless otherwiseindicated.

M24 k

Obv. Female deity (Iris?) with curved wings and outstretched handsfl ying left, looking right; holding a kerykeion in right hand and awreath in left.

Rev. Granulated patterns on either side of triangular baetyl; above left,M24 (k ); all within incuse square.

Paris, BN, 703 (11.76g; 09H) = Konuk (1998b:no

90a) (c. 450–430B.C.).

Variants with the granulated patterns in the shape of stylised birds(Konuk 1998b:no 95 [ c. 430–410 B.C.]) or bunches of grapes areknown (Konuk 1998b:no 96 [ c. 430–410 B.C.]).

M25 k

n

Obv. As last.Rev. Bunch of grapes on either side of triangular baetyl with M25 (n )in its centre; above left, M25 (k ); all within incuse square.



481

London, BM (11.27g; 12H) = Konuk (1998b:no 99bis x) (c. 430–410B.C.).

n (n ) is added to a die in whose original state the baetyl wasanepigraphic.

M26 k 5

Obv. As last.Rev. M26 (kb ) flanking triangular baetyl; all within incuse square.

Sotheby 27 Oct. 1993 (Zürich), 694 (11.57g) = (Konuk 1998b:no

100h) (c. 410–390 B.C.).

Variants with the granulated patterns in the shape of stylised birds(Konuk 1998b:no 95 [ c. 430–410 B.C.]) or bunches of grapes areknown (Konuk 1998b:no 96 [ c. 430–410 B.C.]).

M27 k 5

JObv. As last.Rev. M27 (kb J ) flanking triangular baetyl; all within incuse square.

Lanz 38 (1984), 272 (11.56g; 07H) = Konuk (1998b:no

110c)(c. 410–390 B.C.).

It is uncertain whether the third and final sign ( J ) is an actual let-ter. The use of this sign as a letter is not attested in Carian inscrip-tions where it is sometimes used as a dividing stroke. There ishowever another coin legend (M33) which includes the same sign.As with M33, its final position in the legend raises the possibilitythat it was used as a letter and not as a separation mark. For

further discussion, see M33 below.

M28 k 5

Obv. Head of Apollo three-quarter facing right or left, with on somedies, chlamys fastened at neck.

Rev. M28 (kb ) flanking sphinx seated left.AE chalkous.Künker 61 (2001), 67 (1,27g); Konuk (1998b:no 118) (c. 390–370B.C.).

M29–M30 have been tentatively attributed to Kaunos: see Ashton(2003:39–40).



482

M29 k

Obv. Bunch of grapes.Rev. M29 (k ) within circle of dots.

1/16th (?) Aeginetic stater; c. 400–350 B.C.Terzian collection (0.48g; 02H) = Ashton (2003:39, 1).

M30 k

Obv. Corngrain within circle of dots.Rev. M30 (k ) within circle of dots.

1/32nd Aeginetic stater; c. 400–350 B.C.

Ashton collection (0.35g; 02H) = Ashton (2003:39, 2b).

Telmessos

M31 i F

Obv. Head of Athena left in Attic helmet; in front, linear device l;dotted circle.

Rev. Heracles fighting left with club, left foot placed on rock; along the right edge, erbbinna in Lycian characters; on either side of

Herakles, M31 (i t ) all in incuse square with dotted border.Light Lycian standard stater; c. 420 B.C.Munich, Staatliche Münzsammlung (7.87g) = Babelon 1910:n°385.

This stater of the Lycian dynast Erbbina, part of his regular issuesfrom Telmessos, is the only example to bear a legend in Carian.Various readings based on the changing values given through the years to the letters have been proposed (er, ir, i “ and finally it ).

New evidence from the bilingual inscription from Kaunos has ledAdiego (1998b:58–60) and Meier-Brügger (1998:45) to give F thevalue t . They wonder whether i might not stand for the initial of Erbbina in Carian and t for the initial of Telmessos (Telebehi inLycian). But the legend may also be transliterated as ti and standfor the first two letters of the same ethnic.



483

II. U M

Mint A

M32 £c¢ñÌ Bg

Obv. Naked male figure, with wings at shoulder and heel, in the kneel-ing-running position advancing left, head and legs left, trunk frontal,left arm raised and right arm lowered; groundline.

Rev. Bull standing right; above and below, in two lines, M32 (d t a /

ñibr ); all within incuse square.Aeginetic standard stater; c. 450–400 B.C.London, BM, CM 1934–0611–4 (11.62g; 09H).

For a detailed discussion of M32, see Konuk forthcoming (c). Onlyone specimen of this type is known and, according to Robinson(1939:270), was reportedly obtained near Fethiye (Telmessos). Theorientation of letters suggests a reading from right to left. The lastletter, partly erased, is ñ (ñ) rather than z (≤ ), even though the

latter is a more common ending used for indicating the genitiveform. The transliteration should thus be a t d / rbiñ . As an inde-pendent word, rbiñ shows a striking similarity with Erbbina, theLycian dynast who minted at Telmessos c. 420–400 B.C. (see pre-vious coin).

Mint B

All the following coins are Aeginetic standard staters unless otherwise

indicated.

M33 J P sNObv. Naked male figure, with wings at shoulder and heel, in kneeling-

running position advancing left, head and legs left, right arm raisedand left arm lowered; above right wing, ; dotted groundline.

Rev. Lion standing left with head turned back; above its back, divid-ing M33 (J? ps g ); dotted groundline; all within incuse square.

London, BM (11.68g; 09H) = Robinson (1936:pl. 14, 8); c. 450B.C.



484

There is an obverse variant with the winged male figure advanc-ing right (Robinson 1936: no 11). I consider to be a linear

device, perhaps even an object rather than a monogram of theletters Uo or z. On some contemporary issues from a diff erentmint (Troxell 1979:pl. 31, 35), is depicted much like an objectdecorated with dots. J is omitted on the following examples, sug-gesting that it is the last sign of the legend which, therefore, isread from right to left (g sp J ? ). The meaning of the last sign ( J ) isproblematic. Its sole function on Carian inscriptions is to separatewords, and it is never used as a letter. The two other legends inwhich J appears are M36–M37 and M27. In the case of M36–M37,our sign may well have served its normal function of separating words, but for M27, this explanation can hardly be valid since itis the last sign of the legend, as it is in the present example. ForM27 it is tempting to consider J as the equivalent of a Greek iotabecause this is the very sound which comes after kb (for kbid ),although it must be admitted that I is used as the third letter of kbi in the bilingual inscription from Kaunos. To sum up, M27and M33 compel us to accept J as a letter even though its absence

as a letter from inscriptions remains puzzling. As the sign J onlyappears on coin legends, it may well be a direct influence of coinengravers being more familiar with engraving Greek legends (seealso M23). It is worth noting here that the letter P was knownsolely on coins until 2004, when a new inscription found atHyllarima revealed for the first time the letter B on an inscriptionfrom Caria, albeit with a diff erent value.

M34 °s0

Obv. Naked male figure, with wings at shoulder and heel, in kneeling-running position advancing right, head and legs right, left armraised and right arm lowered; above left wing, ; groundline.

Rev. Lion standing left with head turned back, right forepaw raised;above its back, M34 ( ps g ); underneath legend, small ; dottedgroundline; all within incuse square.Paris, BN (11.64g; 11H); c. 450 B.C.

The more angular shape of ° is further confirmation that this let-

ter is a variant of p with a 90° rotation. There is a reverse vari-ant which shows the lion with both forepaws standing on thegoundline (Robinson 1936:4, 9ter ).



485

M35 s0

P

Obv. Naked male figure, with wings at shoulder and heel, in kneeling-running position advancing right, head and legs right, left armraised and right arm lowered; above left wing, ; groundline.

Rev. Lion standing left with head turned back, right forepaw raised;above its back, and M35 (s g / p ) whose third letter is in thelower left corner; all within incuse square with dotted frame.New York, ANS, 67.152.457 (11.78g; 09H); c. 450 B.C.

M36? s N_ J s?d

Obv. Naked male figure, with wings at shoulder and heel, in the kneeling-running position advancing right, head and legs right, trunk frontal,left arm raised and right arm lowered; above left wing, ; dottedgroundline.

Rev. Lion standing left with head turned back, right forepaw raised;above its back, dividing M36 (bsj |sbd ); all within incuse squarewith dotted frame.Paris, BN (11.72g; 12H); c. 450 B.C.

There is an obverse variety with the male figure advancing left(Robinson 1936:pl. 14, 16). M36 links the old winged male / liontype to the new issues with lion forepart / male head type witha diff erent legend. The right-to-left direction proposed for previ-ous coins is also suggested for this coin by the orientation of thetwo instances of ?; hence M36 should be transliterated as dbs|jsb.The orientation of d does not help much since it is inconsistentbetween M36 and M37.

M37 £Hs J N_sH

Obv. Forepart of lion right, jaws open; both forelegs visible.Rev. M37 (dbs|jsb ) vertically in front of wreathed male head left; behind

neck, ; all within incuse square.Stater: New York, ANS, 63.35.1 (11.18g; 07H); triobol: SNG Kayhan,979 (2.93g; 11H); obol or corroded diobol?: Peus 212 (2000), 276(1.29g). c. 430–400 B.C.

On the New York stater, first published by Thompson (1966:8),the first and penultimate letters are clearly H, not (as on M36) ?.All recorded triobols are die-duplicates and feature H as the firstletter and ? as the penultimate. M37 must of course be the same



486

legend as M36. Perhaps the coin engraver was confused by theorientation of these letters and turned ? into H, which would make

it readable in both directions, although it remains puzzling that? was not rendered to be consistent with M36. For a similarconfusion, see M23 above.

Mint C

M38 7Obv. Head of lion three-quarter facing left, within linear circle.Rev. Head of bull left; behind neck truncation, M38 (z or symbol); all

within round incuse.Milesian standard hemiobol; c. 400–350 B.C.Muharrem Kayhan collection (0.42g; 10H).

It is uncertain whether the sign on the bull’s neck is a characteror a symbol (linear device), because a variety of the same series(Hirsch 221 [2002], 267; another specimen from diff erent dies:Hirsch 226 [2003], 1421) carries another sign ( h ) which is clearlya linear device of a type encountered on other non-Carian issues(e.g. SNG Kayhan, 744). There is also a variety of M38 withoutany letter or symbol (SNG Kayhan, 990).

put

1 LM39 z

Obv. Head of lion three-quarter facing left, within linear circle.Rev. Head of bull three-quarter facing left; above, put ( put ); behind

neck truncation, L ( ); below, z (≤ ); on the neck, 1 (z or sym-bol); all within shallow round incuse.Milesian standard hemiobol; c. 400–350 B.C.Muharrem Kayhan collection (0.49g; 03H).

All known specimens of M39 are off -flan and lack the letter onthe right edge of the coin (e.g. SNG Keckman, 862; Klein 1999:n° 503). The Kayhan specimen is the only one known to me witha full legend. The direction of the legend follows the shape of the

coin: put

≤ . The last letter (≤ ) is used in Carian to indicate thegenitive form. We may interpret the legend as (the coin of ) putl ,most probably the name of a local dynast. The reverse is flat



487

without any trace of an incuse; this would place M39 after M38in the sequence of issues.

Mint D

Types as follows unless otherwise indicated.Obv. Forepart of bull left.Rev. Forepart of bull left; below its head, M40 ( p ); M41 (s ); M42 (db );

M43 (∞ ); M44 (∞ ); all within incuse square.All coins are Milesian standard diobols, apart from M46 whichis a Milesian standard obol.

M40 pParis, BN, fonds général 3292 (2.14g) = Babelon (1910:no 1789[listed under Samos]), pl. CL, 12 = Troxell 1984:no 12A.

M41 sHirsch 55 (1967), 2175 (2.19g) = Cancio (1989:83).

M42 _Cahn 60 (1928), 858 (2.01g) = Troxell (1984:256, 12B).

_ is another example of a Carian letter used on a coin which isotherwise only attested in inscriptions from Egypt (see above, M32).The orientation of letters suggests a right-to-left reading whichwould give the transcription bd .

M43 x

Private collection (2.16g; 12H).The style of M43 is later than the preceding examples.

M44 X

Obv. Confronted foreparts of two bulls, their horns crossed.Rev. As last, M44 (∞ ).

SNG Kayhan, 958 (2.10g; 12H).



488

M45 ux

Obv. As last, but later style; heads of bulls three-quarter facing, horns

not crossed.Rev. Forepart of bull left; above its head, M45 (u∞ ); all within incuse

square.London, BM (2.12g; 08H) = Six (1890:239, 42; pl. 17, 9).

The transcription ∞u is confirmed by M43 and M44 where ∞ isthe initial. The variety with the two bull foreparts was a prolificcoinage, especially in its early phase which is anepigraphic andfeatures the bull foreparts in profile with their horns crossed as if

the letter x was meant. Since other letters are attested (see above),it is uncertain whether X, x or ux should be regarded as thefirst letters of an ethnic. But if the letter x can be construed fromthe crossed horns, a Carian ethnic beginning with ∞ and ∞u suchas Kydai / Hydai and Kyblissos / Hyblissos is an attractive pos-sibility; see Konuk (2003:n° 69).

M46 ^Obv. Forepart of bull left.Rev. Head of bull left; below, M46; all within incuse square.

Milesian standard obol.SNG Kayhan, 974 (1.14g; 03H).

It is uncertain whether ^ is a character or a linear device (sym-bol). It is otherwise unknown.

Mint E

M47 ¥

Obv. Bearded head right.Rev. Forepart of bull left; on its shoulder, M47 ( y ); all within incuse

square.Private collection (1.24g; 12H); c. 400 B.C.

M48 !

Obv. As last.

Rev. Forepart of bull left; in lower left corner, M48 ( y ); all withinincuse square.Paris, BN (1.59g; 12H) = Babelon (1910:n° 2494); c. 400 B.C.



489

M49 ÚObv. Male head right; wreathed?

Rev. Forepart of bull left; on its neck, M49 ( y ).Berlin, Staatliche Museen—Münzkabinett, gift T. Wiegand 1354/1931 (1.10g; 04H); c. 400 B.C.

Winzer’s attribution of this series to the mint of Mylasa underHekatomnos by interpreting M48 and M49 as the Greek initialof his name and taking the bearded head on the obverse as hisportrait is speculative (Winzer 2005:13.1). The diff erent orienta-tions of the letter, especially M47, suggest Carian and not Greek.

There is also an early variety with the same type without letter(private collection, 0.68g; 09H). The length of time needed forthe stylistic change from the early issue to M49 suggests a civiccoinage rather than a dynastic.

Mint F

M50 oul or luo

Obv. Forepart of lion right, head turned back; to the right, M50 (oul or luo ).

Rev. Square punch mark.Aeginetic standard stater; c. 500 B.C.Paris, BN (11.71g).

The legend on this series has been long regarded as Greek. It hasusually been read as ouB ( BMC Ionia , xxxiv) or oBu (Six 1890:223). The former reading prompted an attribution to a dynast of Miletos, the latter to the town of Olymos in Caria. Either read-ing depends upon whether the legend is meant to be read fromthe inside or the outside. As Head rightly points out in BMC Ionia ,a reading from the outside is extremely rare on archaic coins,which undermines Six’s attribution to Olymos. However, Head’sattribution to a dynast of Miletos is no longer satisfactory in viewof the recorded provenances which clearly point to a mint fur-ther south in Caria. In my 1996 RNS paper, I suggested that thelegend should be regarded as Carian. The occurrence of this series

in early hoards such as the Santorini find (IGCH 7) points to adate of c. 500 B.C. When Carian staters of Aeginetic standardbear a legend in the fifth century B.C., it is usually in the Carian



490

script. My reading of the third letter (the lowest on the coin) isl, not B; the horizontal stroke, even though a little erased, being

clearly visible. We would thus have oul for oul . This conclusionhas also been reached by Iık (2003:124–126) who proposes thesame reading. oul may be linked to personal names like Ouliades(Herodotos 5, 37 mentions one Oliatos ruler of Mylasa, c. 500 B.C.)or Oulios. When the legend is read from the outside, there is a puz-zling similarity with M22 and M23 which give the beginning of theethnic of Keramos in Carian. At this stage, an attribution to Keramoswould be premature, but this may change with future finds. Oneshould not exclude, however, a retrograde reading which is rathercommon on early Carian coins. We would have luo (luo ).

Mint G

M51 orou

Obv. Forepart of winged human-headed bull right.Rev. Female head right; behind, M51 (orou ); all within incuse square

with dotted frame.

Aeginetic standard hemidrachm (triobol) and trihemiobol; c. 450–400B.C.Oxford, Ashmolean Museum (2.87g; 06H); trihemiobol: privatecollection (1.25g).

M52 orou

Obv. Sphinx seated right.Rev. Female head right; behind, M52 (orou ); all within incuse square

with dotted frame.

Aeginetic standard; obol; c. 450–400 B.C.

For a detailed discussion of Mint G, see Konuk (forthcoming [d]).The legends of M51 and M52 are usually regarded as Lycian andare attributed to a dynast by the name of Uvug or Uwug. A Lycian origin, however, is far from certain. First, the weight stan-dard is clearly Aeginetic with half staters (same type as M51 butanepigraphic) weighing c. 5.80g. That standard was quite wide-

spread among Carian mints active in the fifth century B.C. Lycia,on the other hand, had its own weight systems and is not knownto have ever used the Aeginetic standard. Another argument againsta Lycian origin is the obverse type of a winged human-headedbull, a very unusual iconography for Lycia. On the other hand,

Oxford, Ashmolean Museum (1.00g; 06H).



491

the depiction of the sphinx on M52 is very close to that of thecontemporary coins of Kaunos, see Konuk (1998b:119). The style

and the particular care given to the striking are also features nottypical of Lycian mints. A further important piece of evidence tosupport a Carian origin is the occurrence of several of the half staters (mentioned above) in an unpublished hoard from Cariadating to the middle of the fifth century B.C. which included avariety of Carian coins (e.g. from Kaunos, Knidos, uncertain mints).Coins of Caria and Lycia did not usually circulate together andit is no surprise therefore that not a single coin from Lycia isreported in the hoard. All in all, the cumulative weight of evi-dence points to a mint in Caria, orou being the name of a Cariandynast of the second half of the fifth century B.C.

Mint H

M53 i

Obv. Persian hero-king right, in kneeling-running position, holding transverse spear and bow.

Rev. Ship’s prow left; on its rail, M53 (i ).AV Persian daric, c. 450–400 B.C.Paris, BN, collection de Luynes (8.25g) = Six (1890:pl. 17, 13) =Babelon (1910:pl. 97, 24).

For a brief discussion of this unique daric, see Konuk (2000a:179).The tentative attribution to Salmakis near Halikarnassos, first sug-gested by Six and followed by myself, appears now to be unfounded.It was based on an incorrect association with bronze coins which

turned out to be from the Cypriot mint of Salamis. According tothe typology established for darics and sigloi, the Persian hero-king holding spear and bow is type IIIb which ends c. 400 B.C.

Mint I

M54

M54Obv. Boar advancing left; above, M54; double groundline; dotted border.Rev. Triskeles ending with duck’s heads; floral ornament growing from

central ring; around, kuprlli in Lycian; all within dotted borderincuse square.



Lydian (Persic) standard stater; c. 450 B.C.Paris, BN (10.85g) = Babelon (1910:no 253).

The reverse legend is clearly Lycian and names the dynast Kuprllias the issuer. The five-character legend of the obverse is, on theother hand, quite problematic. It has been variously transcribedbut a better preserved specimen from the same pair of dies (SNG von Aulock , 4156; 10.70g) has allowed some precision in the read-ing of the legend. The right-to-left direction is suggested by theorientation of the letters. Mørkholm-Neumann (1978:no M 301a)described the obverse as Carian with a question mark, while oth-

ers have not hesitated in recognising Carian: e.g. Durnford (1991)(his reading of M54 as the Carian ethnic of Xanthos is no longertenable as some of the letter values he used are not acceptedtoday), Cau (1999). The only other instance of Carian being usedon a Lycian issue is M31 listed above. On close examination thefirst character cannot be o as previously thought but is probablyd (g ) a character form already encountered on some varieties of M33. The second letter S may be “ as seen on the alphabet of Hyllarima. The letter _ is not attested in the Carian alphabet, butit may be a form of g (r ). the letter K is quite difficult to inter-pret; it may be a form of k (k ) or l (l )? The last letter may be¢ (a). All in all, there are far too many uncertainties over M54to even describe it as Carian, and at this stage, I prefer not tospeculate on the various ways of transcribing it. The weight stan-dard appears to be Lydian (Persic), which has as its stater a dou-ble siglos weighing slightly less than 11.00g. This is exceptionalfor Kuprlli and Lycian mints in general which struck coins in

their own local standards. The decision to use the Lydian stan-dard may well have something to do with the obverse legend.

492



ABBREVIATIONS AND BIBLIOGRAPHY

A

ABSA = The Annual of the Britisch School at Athens. AArch = Acta Archeologica Acta Ant. Hung . = Acta Antiqua Academiae Scientiarum Hungaricae. AC = L’Antiquité Classique. AfO = Archiv für Orientforschung.

ANSMN = American Numismatic Society Museum Notes. ArOr = Archiv Orientální. ASBW = Archiv für Schreib- und Buchwesen. ASNP = Annali della Scuola Superiore Normale di Pisa. BB = Beiträge zur Kunde der indogermanischen Sprachen. BCH = Bulletin de Correspondance Hellenique. BiOr = Bibliotheca Orientalis. BMC Caria = B. V. Head, Catalogue of the Greek Coins in The British Museum, Greek Coins

of Caria, Cos, Rhodes, & c., London, 1897. BMC Ionia = B. V. Head, Catalogue of the Greek Coins in The British Museum, Greek Coins

of Ionia , London, 1892.

BSL = Bulletin de la Société Linguistique de Paris. BzN = Beiträge zur Namenforschung.Cahn = Adolph E. Cahn, Frankfurt a. M.CFC = Cuadernos de Filología Clásica.CNG = Classical Numismatic Group, Lancaster, Pa (USA)—London (UK).Colloquium Caricum = W. Blümel – P. Frei – C. Marek (eds.), Colloquium Caricum. Akten

der internationalen Tagung über die karisch-griechische Bilingue von Kaunos 31.10–1.11.1997 in Feusisberg bei Zürich = Kadmos 37 (1998).

Decifrazione del cario = M. E. Giannotta et alii (eds.), La decifrazione del cario, Roma, 1994.

Eothen = F. Imparati (ed.), Eothen. Studi di storia e di fi lologia anatolica dedicati a Giovanni Pugliese Carratelli , Firenze, 1988.

EpAnat = Epigraphica Anatolica. Zeitschrift für Epigraphik und historische Geographie Anatoliens .Fs. Bloesch = Zur griechischen Kunst. Hansjörg Bloesch zum 60. Geburtstag am 5. Juli 1972,Bern, 1973.

Fs. Friedrich = R. von Kienle et alii (eds.), Festschrift Johannes Friedrich zum 65. Geburtstag am 27. August 1958 gewidmet , Heidelberg, 1959.

Fs. Grumach = W. Brice (ed.), Europa: Studien zur Geschichte und Epigraphik der frühen Agais.Festschrift für Ernst Grumach, Berlin, 1968.

Fs. Neumann I = J. Tischler (ed.), Serta Indogermanica. Festschrift G. Neumann, Innsbruck,1982.

Fs. Neumann II = M. Fritz – S. Zeilfelder (eds.), Novalis Indogermanica. Festschrift fürG. Neumann zum 80. Geburtstag , Graz, 2002.

zum 70. Geburtstag gewidmet , Innsbruck, 1986.Gs. Kretschmer = H. Kronasser (ed.), MNHMHS XARIN. Gedenkschrift Paul Kretschmer , Wien,

1956–1957 (I–II).Hauck & Aufhäuser = Hauck & Aufhäuser Privatbankiers, Munich.

AGI = Archivio Glottologico Italiano.

Demotisches Namenbuch

Fs. Oberhuber = W. Meid H. Trenkwalder (eds.), Im Bannkreis des Alten Orients, K. Oberhuber –

DNb = E. Lüddeckens – H. J. Thissen (eds.), , Wiesbaden, 1980.



494

Hirsch = Münzhandlung Gerhard Hirsch Nachfolger, Munich.HS = Historische Sprachforschung.IF = Indogermanische Forschungen.

IGCH = M. Thompson – O. Mørkholm – C. M. Kraay (eds.), An Inventory of Greek Coin Hoards , New York, 1973.InL = Incontri Linguistici.IstMitt = Istanbuler Mitteilungen. JAOS = Journal of the American Oriental Society. JCS = Journal of Cuneiform Studies. JEA = Journal of Egyptian Archaeology. JHS = Journal of Hellenic Studies.KarON see Blümel (1998a).KarPN see Blümel (1992).KlF = Kleinasiatische Forschungen, Weimar, 1927–1930.KON see Zgusta (1984).KPN see Zgusta (1964).Künker = Fritz Rudolph Künker Münzenhandlung, Osnabrück.KZ = Zeitschrift für vergleichende Sprachforschung.Lanz = Numismatik Lanz, Munich.LNH see Laroche (1966). MH = Museum Helveticum.MM = Münzen und Medaillen AG, Basel. MSS = Münchener Studien zur Sprachwissenschaft.Naville = Naville & Ars Classica, Lucerne. NC = Numismatic Chronicle. NDH see Laroche (1947).

OLZ = Orientalische Literaturzeitung.Or = Orientalia. Commentarii periodici Ponti fi cii Instituti Biblici Nova Series.ÖJh = Jahreshefte des österreichischen Archäologischen Institutes.Peus = Dr. Busso Peus Nachf., Münzhandlung, Frankfurt a. M.PP = Parola del Passato.PSBA = Proceedings of the Society of Biblical Archaeology.RALinc = Atti della Accademia Nazionale dei Lincei, Rendiconti della Classe di Scienze morali,

storiche e fi lologiche, Serie VIII.RE = Paulys Real-Encyclopädie der classischen Altertumswissenschaft: neue Bearbeitung unter

Mitwirkung zahlreicher Fachgenossen; herausgegeben von Georg Wissowa , Stuttgart, 1894–1963.REA = Revue des Études Anciennes.REG = Revue des Études Grecques.

RHA = Revue Hittite et Asianique .RPh = Revue de Philologie.SEAP = Studi di Eggitologia e di Antichità Puniche.SM = Schweizer Münzblätter.SMEA = Studi Micenei ed Egeo-Anatolici.SNG Kayhan = Sylloge Nummorum Graecorum, Turkey 1, The Muharrem Kayhan Collection,

Istanbul-Bordeaux, 2002.SNG Keckman = Sylloge Nummorum Graecorum, Finland, The Erkki Keckman Collection in the

Skopbank, Helsinki, part I Karia , Helsinki, 1994.SNG von Aulock = Sylloge Nummorum Graecorum, Deutschland: Sammlung von Aulock, 18 fasci-

cules , Berlin 1957–1981.TA1 see Laroche (1957).TA2 see Laroche (1961).TSBA = Transactions of the Society of Biblical Archaeology.VJa = Voprosy Jazykoznanija.Weber = L. Forrer, Descriptive Catalogue of the Collection of Greek Coins formed by Sir Hermann

Weber M.D., vol. 3, London, 1929.

495

WO = Die Welt des Orients: wissenschaftliche Beiträge zur Kunde des Morgenlandes.ZDMG = Zeitschrift der deutschen morgenländischen Gesellschaft.ZPE = Zeitschrift für Papyrologie und Epigraphik.

B

(1990a) “Deux notes sur la langue et l’écriture cariennes”, Kadmos 29, 133–138.(1990b) Studia Carica. Investigaciones sobre la escritura y lengua carias, y su relación con la familia

lingüística anataolia indoeuropea , Doctoral dissertation, Barcelona. (published as microfilm).(1992a) “Recherches cariennes: essai d’amélioration du système de J. D. Ray”, Kadmos

31, 25–39.

Senmartí (eds.), Homenatge a Josep Alsina. Actes del X è Simposi d’Estudis Clàssics, Tarragona 28–30 de novembre del 1990 , I, Tarragona, 51–54.

(1993a) Studia Carica. Investigaciones sobre la escritura y lengua carias, Barcelona.(1993b) “Sobre OALOALON SGDI 5727.d30”, Kadmos 32 (1993), 173–174.(1994a) “Les identifications onomastiques dans le déchiff rement du carien”, Decifrazione

del cario, 27–63, “Considerazioni conclusive”, ibid. 239–240.(1994b) “El nombre cario Hecatomno”, CFC (Estudios griegos e indoeuropeos) n. s. 4,

247–256.(1994c) “Genitiu singular en lici i protoluvi”, Anuari de Filologia. Studia Graeca et Latina

17, 11–33.(1995) “Contribuciones al desciframiento del cario”, Kadmos 34, 18–34.(1996) “Comentarios a la nueva lectura de la inscripción caria 28*”, Kadmos 35, 160–163.

(1998a) “La nueva bilingüe greco-caria de Cauno y el desciframiento del cario”, Aula Orientalis 16 (1998), 5–26.(1998b) “Die neue Bilingue von Kaunos und das Problem des karischen Alphabets”,

Colloquium Caricum, 57–79.(2000) “La inscripción greco-caria de los Hecatómnidas en el santuario de Sinuri”,

Kadmos 39, 133–157.(2002) “Cario de Cauno punoO”, Aula Orientalis 20, 13–20.

S. Torallas Tovar – E. R. Luján – M. Á. Gallego (eds.), Lenguas en contacto: El testi-monio escrito, Madrid, 299–320.

Adiego, I.-J. – Debord, P. – Varinlio< lu, E.(2005) “La stèle caro-grecque d’Hyllarima”, REA 107 nº 2, 601–653.

Ashton, R.

NC 158, 46–49.(2003) “(vi) Kaunian Notes”, in: R. Ashton – P. Kinns, “Opuscula Anatolica II”, NC

Avishur, Y. – Heltzer, M.(2003) “Carians as skilled masons in Israel and mercenaries in Judah in the early I

millennium B.C.E.”, Kadmos 42, 87–90.

Babelon, E.

(1910) Traité des monnaies grecques et romaines. Deuxième partie, description historique , t. 2, Paris.Bean, G. F.(1953) “Notes and inscriptions from Caunus”, JHS 73, 10–35.(1954) “Notes and inscriptions from Caunus (continued)”, JHS 74, 85–110.

163, 36–40.

Adiego, I.-J.

(1992b) “Glosses i pseudoglosses càries en fonts gregues”, in: J. Zaragoza – A. González

(2005) “La nueva inscripción caria de Milasa”, Kadmos 44, 81–94.

(2004) “Los alfabetos epicóricos anhelénicos de Asia Menor”, in: P. Bádenas de la Peña –

(1998) “Keramos”, in: R. Ashton et alii., “Some Greek Coins in the British Museum”,



496

Belli, P. – Gusmani, R.(2001) “Una nuova iscrizione rupestre presso il Santuario di Labraunda in Caria”, PP

56, 33–41.

Bengston, H.(1954–55) “Skylax von Karyanda und Herakleides von Mylasa”, Historia 3, 302–305.

Benveniste, E.(1931) “Noms cariens”, RHA I. 2, 52–57.

Bernand, A. – Aly, A.(1959?) Abou-Simbel, inscriptions grecques, cariennes et sémitiques des statues de la façade , Le

Caire, Centre de documentation égyptologique, Collection scientifique.

Bernand, A. – Masson, O.(1957) “Les inscriptions grecques d’Abu Simbel”, REG 70, 1–46.

Berti, F. – Innocente, L.(1998) “Due nuovi graffiti in alfabeto cario da Iasos”, Colloquium Caricum, 137–142.(2005) “Graffito cario su piede di coppa attica”, Bollettino dell’Associazione Iasos di Caria ,

11, 20–21.

Bertoldi, V.(1948) “Souangela, Tomba del Re”, PP 3, 5–11.

Blümel, W.KarON: see (1998a).KarPN: see (1992).

Toplantısı, Ankara 23–27 Mayıs 1988 , 261–264.

(1990) “Zwei neue Inschriften aus Mylasa aus der Zeit des Maussollos”, EpAnat 16,29–43 Taf. 12.(1992) “Einheimische Personennamen in griechischen Inschriften aus Karien”, EpAnat

20, 7–34.

(1994) “Über die chronologische und geographische Verteilung einheimischer Personen-namen in griechischen Inschriften aus Karien”, Decifrazione del cario, 65–86.

(1996) “Epigraphische Forschungen im Westen Kariens 1994”, in: XIII Ara tırma Sonuçları Toplantısı , Ankara, 261–264.

(1998a) “Einheimische Ortsnamen in Karien”, EpAnat 30 (1998), 163–184.(1998b) “Karien, die Karer und ihre Nachbarn in Kleinasien”, Colloquium Caricum,

163–173.(2005) “Problematische Lesungen in der karischen Inschrift aus der Region von Mylasa”,

Kadmos 44, 188.

Blümel, W. – Adiego, I.-J.(1993) “Die karische Inschrift von Kildara”, Kadmos 32, 87–95.

Blümel, W. – Kızıl, A.(2004) “Eine neue karische Inschrift aus der Region von Mylasa”, Kadmos 43, 131–138.

Bockisch, G.(1969) “Die Karer und ihre Dynasten”, Klio 51, 117–175.

Bohl, F. M.

(1932–33) “Inschriften mit unbekannter Schrift aus der Leidener Sammlung”, AfO 8,173–174.

Boisson, C.

Decifrazione del cario, 207–232.

(1988) “Epigraphische Forschungen in der Region von Mylasa”, in: VI Ara tırma Sonuçları

(1993) “SGDI 5727 (Halikarnassos): eine Revision”, Kadmos 32, 1–18.

(1994) “Conséquences phonétiques de certaines hypothèses de déchiff rement du carien”,




497

Bork, F.(1930) “Die Schrift der Karer”, ASBW 4, 18–30.(1931) “Die Sprache der Karer”, AfO 7, 14–23.

Brandenstein, W.(1934a) “J F X C in den epichorischen Alphabeten Kleinasiens”, Klio 27, 69–73.(1935a) “Karische Sprache” in: Pauly-Wissowa, RE Supplementband VI, col. 140–146.(1936) “Streifzüge . . . Zwei Karische Ortsnamen”, Glotta 25, 32–35.

Brixhe, C.(1996) Review of Decifrazione del cario in: BSL 91/2, 214–221.

Cancio, L.(1989) “A New Satrapal Coin of the KIM—EKA Series”, SM 156, 83.

Carruba, O.(1998) “Zum Stand der Entziff erung des Karischen”, Colloquium Caricum, 47–56.(1999a) “Bildungen karischer Ethnika”, SMEA 41/2 (1999), 175–180.(1999b) “Ar/w/wazuma”, Kadmos 38, 50–58.(2000) “Der Name der Karer”, Athenaeum 88, 49–57.

Cau, N.(1999a) “La legenda caria su una serie monetale del dinasta Kuprlli”, Studi Ellenistici

12, 9–17.(1999b) “Una nuova lettura di alcune leggende monetali carie”, Kadmos 38, 43–49.(2003) “Nuovi antroponimi indigeni nelle iscrizioni greche della Licia di età ellenistico-

romana”, Studi Ellenistici 15, 297–340.

Deroy, L.

(1955) “Les inscriptions cariennes de Carie”, AC 24, 305–335.(1959) Review of Masson-Yoyotte (1956) in: Or 28, 101–102.

Descat, R.(1994) “La géographie dans les listes de tributs attiques: Lepsimandos et Kasôlaba en

Carie”, ZPE 104, 61–68.(1998) “La carrière d’Eupolemos, Stratège macédonien en Asie Mineure. Appendice:

Note sur une inscription caro-grecque de Caunos”, REA 100, 167–190.

Dorsi, P.(1979) “Le glosse carie”, InL 5, 27–35.

Dressler, W.(1966–67[68]) “Karoide Inschriften im Steinbruch von Belevi”, ÖJh 48, 73–76.

Durnford, S.(1991) “An instance of the Lycian name for Xanthos in Carian script”, Kadmos 30, 90–92.

Eilers, W.(1935) “Das Volk der karka in den epichorischen Alphabeten Kleinasiens”, OLZ 38,

201–213.(1940) “Kleinasiatisches”, ZDMG 94, 189–233.

Eichner, H.(1994) “Zur Entziff erung des Karischen”, Decifrazione del cario, 167–170.

Erbse, H.(1986) “Zu der Ilias-Scholien (Curae secundae II)”, Hermes 114.4, 385–398.

ESS (= Epigrahic Survey Staff ).(1988) Reliefs and Inscriptions at Luxor Temple, volume 2: The Facade, Portals, Upper Register

Scenes, Columns Marginalia, and Statuary in the Colonnade Hall (University of ChicagoOriental Institute Publications).



498

Faucounau, J.(1980) “Réflexions sur le déchiff rement des inscriptions cariennes”, Klio 62.2, 289–305.(1984) “A propos de récents progrès dans le déchiff rement de l’écriture carienne”, BSL

79/1, 229–238.

(1994) “Remarks on the Carian Alphabet of Sinuri”, Decifrazione del cario, 233–236.

Franklin, N.(2001) “Masons’ Marks from the Ninth Century B.C.E. Northern Kingdom of Israel.

Evidence of the Nascent Carian Alphabet?”, Kadmos 40, 107–116.

Frei, P. – Marek, C.(1997) “Die karisch-griechische Bilingue von Kaunos. Eine zweisprachige Staatsurkunde

(1998) “Die karisch-griechische Bilingue von Kaunos. Ein neues Textfragment”, ColloquiumCaricum, 1–18.

(2000) “Neues zu den karischen Inschriften von Kaunos”, Kadmos 39, 83–132.Friedrich, J.(1931) “Zu den kleinasiatischen Personennamen mit dem Element muwa ”, Kleinasiatische

Forschungen I, 359–378.(1932) Kleinasiatische Sprachdenkmäler , Berlin.(1952) “Karer in Numidien?”, Or 21, 231–233.(1965) “Ein wohl kleinasiatisches Tontäfelchen mit unbekannter Schrift”, Kadmos 3,

156–169.

Georgiev, V. I.(1960) “Der indoeuropäische Charakter der karischen Sprache”, ArOr 28, 607–619.

(1966) Introduzione alla storia delle lingue indoeuropee (= Incunabula graeca 9), Roma.(1975) “Ein Versuch zur Deutung der griechisch-karischen bilinguis”, Kadmos 14, 64–67.(1981) Introduction to the History of the Indo-European Languages , Sofia. [210–214].

Gérard, R.(2005) Phonétique et morphologie de la langue lydienne , Louvain-la-Neuve.

Goetze, A.

74–81.

Gosline, S. L.(1992) “Carian quarry markings on Elephantine Island”, Kadmos 31, 43–50.(1998) “Quarry, Setting and Team Marks: the Carian Connection”, Journal of Ancient

Civilizations , 13, 59–82.Gusmani, R.(1967) Review of ”evoro“kin (1965) in: AGI 52, 79–84.(1975) Neue epichorische Schriftzeugnisse aus Sardis (1958–1971) = Archaeological Exploration

of Sardis, Monograph 3, Cambridge, Mass.(1978) “Zwei neue Gefässinschriften in karischer Sprache”, Kadmos 17, 67–75.(1979a) Review of Masson (1978) in: Paideia 34, 220–223.(1979b) “Spunti per la decrittazione di segni carii”, InL 5, 193–197.(1982) “Zum Karischen”, Fs. Neumann I , 193–197.(1986) “Die Erforschung des Karischen”, Fs. Oberhuber , 55–67.(1988) “Karische Beiträge”, Kadmos 27.2, 139–149.

(1990) “Karische Beiträge II”, Kadmos 29.1, 47–53.(1994) “Kritisches und Autokritisches zu den Entziff erungsversuchen”, Decifrazione del cario, 115–120.

(1989) “A propos de la lecture des inscriptions cariennes”, Kadmos 28, 174–175.

des 4. Jh.s v. Chr.”, Kadmos 36, 1–89.

(1954) “The Linguistic Continuity of Anatolia as shown by its Proper Names”, JCS 8,



499

Hajnal, I.(1996) Die Entzi ff erung unbekannter Schriften: Drei Fallstudien—ein Szenario? , Bern.(1995[97]) “Das Vokalsystem des Karischen: Eine provisorische Bestandsaufnahme”,

Die Sprache 37, 12–30.(1997a) “«Indogermanische» Syntax in einer neuerschlossenen anatolischen Sprache:Die karische Partikel - x i ”, in: E. Crespo – J. L. García Ramón (eds.), Berthold Delbrück y la sintaxis indoeuropea hoy, Madrid-Wiesbaden, 193–217.

(1997b) “Die karisch-griechische Bilingue 44* aus Kaunos: ein erster Augenschein”,Kadmos 36, 141–166.

(1997c) “Definite nominale Determination im Indogermanischen”, IF 102, 38–73.(1998) “, Jungluwisches‘ *s und die karische Evidenz: Versucheiner dialektologischen

Klärung”, Colloquium Caricum, 80–108.

Hanfmann, G. M. A – Masson, O.(1967) “Carian Inscriptions from Sardis and Stratonikeia”, Kadmos 6, 123–134.

Hegyi, D.(1998) “The Cult of Sinuri in Caria”, Acta Ant. Hung. 38, 157–163.

Heubeck, A.(1959a) Review of Masson-Yoyotte (1956) in: Gnomon 31, 332–336.(1967–68) Review of Otkup“‘ikov (1966) in: IF 72, 331–333.(1974) Review of Zauzich (1972) in: BiOr 31, 95–97.

Hornblower, S.(1982) Mausolus , Oxford.

Houwink Ten Cate, Ph. H. J.(1961) The Luwian Population Groups of Lycia and Cilicia Aspera during the Hellenistic Period

(= Documenta et monumenta orientis antiqui 10), Leiden.de Hoz, J.(2004) “De cómo los protogriegos crearon el griego y los pregriegos lo aprendieron”,

in: P. Bádenas de la Peña – S. Torallas Tovar – E. R. Luján – M. Á. Gallego(eds.), Lenguas en contacto: El testimonio escrito, Madrid, 35–56.

Innocente, L.

(1995) “The Oxford Paracarian Inscription”, Kadmos 34, 149–154.(1997) Review of Adiego (1993) in: Or 66, 116–117.(2002) “Tegola di Iasos”, Kadmos 41, 179–180.Iık, C.(1998) “Die Ergebnisse der Ausgrabungen in Kaunos bis zur Entdeckung der Bilingue”,

Colloquium Caricum, 183–202.Ißık, E.(2003) Frühe Silberprägungen in Städten Westkleinasiens , Saarbrücken.

Janda, M.(1994) “Beiträge zum Karischen”, Decifrazione del cario, 171–190.

Jeff ery, L. H.(1961) The Local Scripts of Archaic Greece , Oxford.

(1964[65]) “Old Smyrna: Inscriptions on sherds and small objects”, ABSA 59, 39–49,pl. 5–8.

Jordan, H.(1968) Review of Otkup“‘ikov (1966) in: OLZ 63, 125–130.

(1992) “Stato delle ricerche sul cario”, Vicino Oriente 8/2, 213–222.

(1994) “Note epigrafiche”, Decifrazione del cario, 101–110.(1992a) “Concordanze delle iscrizioni carie”, SMEA 30, 25–87.



500

Jucker, H. – Meier, M.(1978) “Eine Bronzephiale mit karischer Inschrift”, MH 35, 104–115.

Kalinka, E.(1901) Tituli Asiae Minoris, vol. I Tituli Linguae Lyciae lingua lycia conscripti , Wien.

Kammenhuber, A.(1969b) “Karer, Karia, Karische Sprache”, K. Ziegler – W. Sontheimer (eds.), Der

Kleine Pauly III, Stuttgart, col. 118–121.

Kammerzell, F.(1990) Studien zu Sprache und Geschichte der Karer in Ägypten , Diss. Göttingen.(1993) Studien zu Sprache und Geschichte der Karer in Ägypten, Wiesbaden.

Klein, D.(1999) Sammlung von griechischen Kleinsilbermüzen und Bronzen, Nomismata 3, Milan.

Konuk, K.(1998a) The Coinage of the Hekatomnids of Caria (unpublished Oxford University D. Phil.Diss.).

(1998b) “The Early Coinage of Kaunos”, in: R. Ashton – S. Hurter (eds.), Studies in

(2000a) “Influences et éléments achéménides dans le monnayage de la Carie”, in:O. Casabonne (ed.), Mécanismes et innovations monétaires dans l’Anatolie achéménide. Numismatique

et histoire , Istanbul, 171–183.

(2003) Karun’dan Karia’ya, Muharrem Kayhan Koleksiyonundan Erken Anadolu Sikkeleri. FromKroisos to Karia, Early Anatolian Coins from the Muharrem Kayhan Collection, Istanbul.

(forthcoming [a]) “The Early Coinage of Mylasa”, NC .

Essays in Honour of Ioannis Touratsoglou, Athens.

Tony Hackens , Louvain-la-Neuve.(forthcoming [d]) “Orou, dynaste de Carie”, in: P. Brun (ed.) Anatolica, mélanges en l’hon-

neur de Pierre Debord , Bordeaux.

Kowalski, Th.(1975) “Lettres cariennes: essai de déchiff rement de l’écriture carienne”, Kadmos 14,

73–93.

Kretschmer, P.(1896) Einleitung in der Geschichte der griech. Sprachen, Göttingen.

160.

Laroche, E.LNH see (1966).NDH see (1947).TA1 see (1957).TA2 see (1961).(1947) Recherches sur les noms des dieux hittites , Paris.(1957) “Notes de toponymie anatolienne”, Gs. Kretschmer II, 1–7.(1961) “Etudes de toponymie anatolienne”, RHA XIX.69, 57–98.

(1966) Les noms des hittites , Paris.Laumonier, A.(1933) “Notes sur un voyage en Carie”, Revue archéologique II, 31–55.(1934) “Inscriptions de Carie”, BCH 58, 291–380.(1958) Les Cultes indigènes de Carie , Paris.

Greek Numismatics in Memory of Martin Jessop Price , London, 197–223, pl. 47–50.

(2000b) “Coin evidence for the Carian Name of Keramos”, Kadmos 39, 159–164.

(forthcoming [b]) “Kasolaba, a New Mint in Karia?”, in: S. Drougou – E. Ralli (eds.),

(forthcoming [c]) “Erbbina en Carie?”, in: F. de Callatay et alii (eds.), Liber amicorum

(1929) “Eine neue karische Inschrift”, KlF 1, 318–320.(1954) “Zu der karischen Zeile der SÁma-Inschrift aus Athen oben S. 67”, Glotta 34,



501

Leclant, J.(1951) “Fouilles et travaux en Egypte, 1950–1951”, Or 20, p. 474; pl. LXIV, 37–38.(1960) Review of Masson-Yoyotte (1956) in: RPh, 339–340.

Legrain, G.(1905) “Inscriptions from Gebel Abou Gorâb”, PSBA 27, 129.

Levi, D. – Pugliese Carratelli, G.(1961–62 [1963]) “Nuove Iscrizioni di Iasos”, Annuario della Scuola Archeologica di Atene

vol. 39–40, 573–632.

Loprieno, A.(1995) Ancient Egyptian, Cambridge.

Martin Cary, J.(1991) “Carian in Egypt: the Demotic evidence”, Kadmos 30.2, 173–174.

Masson, O.(1953) “Textes cariens d’Egypte I”, RHA XII (55) 32–38 and pl. XII–XIV.(1954) “Epigraphie asianique . . . L’épigraphie carienne”, Or 23, 439–441.(1959a) “Notes d’anthroponymie grecque et asianique”, BzN 10, 159–170.(1959b) “Documents énigmatiques à inscription pseudo-chypriote et pseudo-carienne”,

Fs. Friedrich, 315–321(1967) “L’ostrakon carien de Hou-Diospolis Parva (38 Friedrich)”, Fs. Grumach, 211–217,

pl. XX.

25–36.(1973) “Que savons-nous de l’écriture et de la langue des Cariens?”, BSL 68/1, 187–213(1973[74]) Review of Zauzich (1972) in: Kratylos 18, 38–43.

(1974) “Notes d’épigraphie carienne”, Kadmos 13, 124–132.(1973[75]) “Un nouveau fragment d’inscription carienne de Kaunos”, Anadolu 17,123–131.

o ff erts à E. Benveniste , Paris, 407–414.(1976) “Un lion de bronze de provenance égyptienne avec inscription carienne”, Kadmos

15.1, 80–83.(1977) “Notes d’épigraphie carienne III–V”, Kadmos 16, 87–94.(1977[78]) “Karer in Ägypten”, Lexikon der Ägyptologie III, col. 333–337.(1978) Carian Inscriptions from North Saqqâra and Buhen, Egypt Exploration Society, London.(1979) “Remarques sur les graf fites cariens d’Abou Simbel”, Hommages à la mémoire de

S. Sauneron II, Le Caire, 35–49.(1988a) “Noms cariens à Iasos”, in: Eothen, Firenze, 155–157.(1988b) “Le culte ionien d’ Apollon Oulios d’après des données onomastiques nouvelles”,

Journal des Savants , Juillet–Décembre 1988, 173–183.

The Assyrian and Babylonian Empires and other States of the Near East, from theEighth to the Sixth Centuries B. C., Cambridge. 666–676, Bibl.: 855–860.

del cario, 191–194.(1994b) “La grande inscription grecque d’Abou-Simbel et le nom probablement carien

Peleqos ”, SMEA 34, 137–140.(1994c) Review of Adiego (1993a) in: BSL 89/2, 185–186.

(1995) Review of Kammerzell (1993) in: Kratylos 40, 172–177.(unpublished [1994]) “Le carien: état de la question”, Communication du 18 juin 1994,Société de Linguistique de Paris.

(1975) “Le nom des Cariens dans quelques langues de l’antiquité”, Mélanges linguistiques

(1969) “Les Cariens en Egypte , Bulletin de la Société Française d’Egyptologie 56, nov. 1969,

”

”

”

(1991) “Anatolian Languages”, in: The Cambridge Ancient History 2a ed. vol III, Part 2:

(1994a) “Les inscriptions cariennes du tombeau de Montuemhat (Thèbes), Decifrazione



502

Masson, O. – Yoyotte, J.(1956) Objets pharaoniques à inscription carienne , Le Caire.

Meier-Brügger, M.(1976) “Zum karischen Namen von Kaunos”, MSS 34, 95–100.(1978) “Karika”, Kadmos 17, 76–84.(1979a) “Karika II–III”, Kadmos 18, 80–88.(1979b) “Ein Buchstabenindex zu den karischen Schriftdenkmälern aus Ägypten”, Kadmos

18, 130–177.(1980a) Review of Masson (1978) in: Gnomon 52, 383–384.(1980b) “Karisch. Eine Bestandsaufnahme”, XX. Deutscher Orientalistentag 1977 in Erlangen

(= ZDMG Suppl. IV), 88–90.(1981) “Eine weitere ,parakarische‘ Inschrift?”, Kadmos 20, 76–78, pl. I–III).(1983) “Die Karischen Inschriften”, in: Labraunda, Swedish Excavations and Researches II,

Part 4 , Stockholm.

(1994) “Ein neuer Blick nach zehn Jahren”, Decifrazione del cario, 111–114.(1998) “Zu den Münzlegenden von Kaunos”, Colloquium Caricum, 42–46.

Melchert, H. C.CLL = Melchert (1993b).DLL = Melchert (2004).(1993) “Some remarks on new readings in Carian”, Kadmos 32.2 (1993), 77–86.(1993b) Cuneiform Luvian Lexicon, Chapel Hill.(1994) Anatolian Historical Phonology, Amsterdam-Atlanta.(1998) “Carian mdoVun ‘we have established’”, Colloquium Caricum, 33–41.(2002) “Sibilants in Carian”, in: Fs. Neumann II , 305–313.(2003) Chapters 1, 2, 5 of H. C. Melchert (ed.), The Luwians , Leiden-Boston.

(2004) A Dictionary of the Lycian Language , Ann Arbor.(mdane) “Carian md ane ”, unpublished paper (1999).

Mentz, A.(1940) “Schrift und Sprache der Karer”, IF 57, 265–280.

Meriggi, P.

(1978) “Sulla scrittura caria”, ASNP , Cl. di Lett. e Fil. serie III, vol. VIII, 3, 791–803.(1980) Review of Masson (1978) in: BiOr 37, 33–37.

Metzger, I. R.

Meyer, G.(1886) “Die karier. Eine ethnographisch-linguistische untersuchung”, BB 10, 147–202.

Mørkholm, O. – Neumann, G.(1978) Die lykischen Münzlegenden, Göttingen.

Mørkholm, O. – Zahle, J.(1976) “The Coinages of the Lycian Dynasts Kheriga, Kherêi and Erbbina. A Numismatic

and Archaeological Study”, AArch 47, 47–90.

Murray, M. A.

(1904) The Osireion at Abydos , London.Nahm, W.

Naumann, R. – Tuchelt, K.

13/14, 16 ss. [pl. 25: Carian graf fiti from Didyma].

(1963) “Karisch ÉIt≈ana, hier. het. itapana-”, Kadmos 2, 73.

(1967) “Zum Karischen”, Fs Grumach, 218–228.

(1973) “Eine geometrische Amphora im Rätischen Museum in Chur”, Fs. Bloesch, 74–77

(1966) “Zur neuen ‘para-karischen’ Schrift”, Kadmos 5, 61–102.

(1969) “Neue Lesungsvorschläge zur Grotthus-Tafel”, Kadmos 8, 58–73.

(1963/1964) “Die Ausgrabung in Südwesten des Tempels von Didyma 1962”, IstMitt

503

Neumann, G.(1961) Untersuchungen zum Weiterleben hethitischen und luwischen Sprachgutes in hellenistischer und

römischer Zeit , Wiesbaden.

(1969a) “Eine neue karische Inschrift aus Chalketor”, Kadmos 8, 152–157.1 Abt., 2. Bd. 1–2 Abschn., Lief. 2,358–396.

(1978) “Spätluwische Namen”, KZ 92, 126–131.(1984) “Zum Namen des Cheramyes von Samos”, Würzburger Jahrbücher für die Altertums-

wissenschaft Neue Folge, Band 10, 41–43.(1988) “Beobachtungen an karischen Ortsnamen”, in: Eothen 183–191.(1993) “Zu den epichorischen Sprachen Kleinasiens”, in: G. Dobesch – G. Rehrenböck

(eds.), Die epigraphische und altertumskundliche Erforschung Kleinasiens: Hundert Jahre Kleinaisatische Kommission der Österreichichesn Akademie der Wissenschaften, Wien, 289–296.

(1994) “Zur Nebenlieferung des Karischen”, Decifrazione del cario, 15–25.(1998) “Sprachvergleichendes zur Kaunos-Bilingue”, Colloquium Caricum, 19–32.

Nicholls, R. V.

Ö< ün, B.(1998) “Warum Kaunos?”, Colloquium Caricum, 175–182.

Otkup“‘ikov, J. V. [ OÚÍÛÔ˘ËÍÓ‚, ˛. ‚.](1966) KapuÈcÍue Ì‡‰nucu AÙpuÍu. èpÂ‰‚‡pumeÎ¸Ì˚e pÂÁyÎmam˚ ‰e¯uÙpÓ‚Íu. [Carian

Inscriptions of Africa. Preliminary results of a decipherment], Leningrad.(1968) “O· ÓÚÌÓ¯eÌËË ÍapËÈcÍÓ„Ó aÎÙa‚ËÚa Í ÍpËÚÓ-ÏËÍeÌcÍÓÏy Ë ÍËÔpcÍÓÏy cËÎÎa·apË˛

[On the connection of the Carian alphabet with the Creto-Mycenaean and Cypriot

syllabary]”, Atti e Memorie del I Congresso Internazionale di Micenologia-1967 , Roma,426–432 and 433–437 [abstract in Italian: “Sul rapporto dell’alfabeto cario con ilsillabario cretese-miceneo e cipriota”].

(1989) “K‡pËÈcÍËÈ Ë „pe˜ÂcÍËÈ flÁ˚Í. „eÌeÚË˜ÂcÍËe Ë ÂÚÌÓ-ÍyÎ¸ÚypÌ˚Â ÓÚÌÓ¯ÂÌËfl [Carian

Paribeni, R.(1936) “Etimologie dalla lingua dei Cari (?)”, Rivista di Fiologia Classica 64, 292–293.

Pedley, J. G.(1974) “Carians in Sardis”, JHS 94, 96–99.

Petrie, W. M. F.

(1901) Diospolis Parva. The Cemeteries of Abadiyeh and Hu, 1898–1899, London.Pfeiler, B.(1962) “Zur Münzkunde von Milet”, SM 46, 20–22.

Pisani, V.(1967) Review of ”evoro“kin (1965) in: Paideia 22, 420–424.

Poetto, M.

Pugliese Carratelli, G.(1948) “Cari in Libia”, PP 3, 15–19.(1974) “Un’epigrafe caria in Persia”, Gurur à jamañjarik à (Studi in onore di Giuseppe Tucci) I,

Napoli, 163–166.(1985[86]) “Cari in Iasos”, RALinc vol. XI, fasc. 5–6, 149–155.

Ray, J. D.(1981) “An approach to the Carian script”, Kadmos 20.2, 150–162.(1982a) “The Carian Script”, Proceedings of the Cambridge Philological Society 208, 77–90.

for 1970–71, 75–76, no 26, fig. 16.

(1969b) “Lykisch”, in: Handbuch der Orientalistik

(1971) “Recent Acquisitions by the Fitzwillian Museum, Cambridge”, Archaeological Reports

and Greek language. Genetic and ethno-cultural connections]”, Klio 71, 66–69.

(1984) “Nuove monete carie”, Kadmos 23, 74–75.



504

(1982b) “The Carian inscriptions from Egypt”, JEA 68, 181–198.(1983) Review of Masson (1978) in: JEA 69, 194–195.(1985) “The Carian coins from Aphrodisias”, Kadmos 24.1, 86–88.

(1987) “The Egyptian approach to Carian”, Kadmos 26.1, 98–103.(1988) “Ussollos in Caria”, Kadmos 27.2, 150–54.(1990a) “An Outline of Carian Grammar”, Kadmos 29, 54–83.(1990b) “A Carian Text: The Longer Inscription from Sinuri”, Kadmos 29, 126–132.(1990c) “The names Psammetichus and Takheta”, JEA 76, 196–199.(1994) “New Egyptian Names in Carian”, Decifrazione del cario, 195–206.(1998) “Aegypto-Carica”, Colloquium Caricum, 125–136.

Robert, L.Première partie: les inscriptions grecques,

Paris.

VII, 59–68.I, 5–22, fig. 1–2, pl. I–VI, VII–X, XXVIII–XXX.

Hellenica VIII, section I, 23–38.(1957) “Deux inscriptions d’Iasos”, REG 70, 362–375.

Robinson, E. S. G.(1936) “A Find of Archaic Coins from South-West Asia Minor”, NC 5/16, 1–16.(1939) “Coin-Legends in Carian Script”, Anatolian Studies presented to W. H. Buckler ,

Manchester, 269–275.

Rochette, B.

(1997–98) “La langue des Cariens à propos de B 867”, Glotta 74, 227–236.Roos, P.(1972) The Rock-Tombs of Caunus, I The Architecture (= Studies in Mediterranean Archaeology

34; 1), Göteborg.(1985) Survey of rock-cut chamber-tombs in Caria, Part 1: South-eastern Caria and the Lyco-Carian

borderland, 9, Göteborg.

ahin, M. Ç.(1973[75]) “Lagina’dan (Koranza) iki yeni yazıt” / “Two New Inscriptions from Lagina

(Koranza)”, Anadolu 17, 178–185 (Turkish) / 187–195 (English).(1976) The political and religious structure in the territory of Stratonikeia in Caria , Ankara.(1980) “A Carian and three Greek inscriptions from Stratonikeia”, ZPE 39, 205–213.

Salmeri, G.(1994) “I Greci e le lingue indigene d’Asia Minore: il caso del cario”, Decifrazione del

cario, 87–99.

Sapir, E.(1936) “kÊbda a Carian Gloss”, JAOS 56, 85.

Sayce, A. H.(1874) “The Karian Inscriptions”, Transactions of the Royal Society of Literature , second

series, 10, 546–564.(1887[92] = 1893) “The Karian language and inscriptions”, TSBA 9.1, 112–154.

28,171–177

(1910) “Karian, Egyptian and Nubian-Greek inscriptions from the Sudan”, PSBA 32,261–268.

(1945) Le sanctuaire de Sinuri près de Mylasa

(1949) “Décret d’une syngeneia carienne au sanctuaire de Sinuri”, in: L. Robert, Hellenica

(1950) “Inscriptions inédites en langue carienne”, in: L. Robert, Hellenica VIII, section

(1950b) “Le carien Mys et l’oracle du Ptôon (Hérodote, VIII, 135)”, in: L. Robert,

(1895) “The Karian and Lydian inscriptions”, PSBA 17, 39–43.

(1906) “An inscription of S-ankh-ka-ra. Karian and other inscriptions”, PSBA

(1908) “Karian, Aramaic, and Greek graf fiti from Heshân”, PSBA 30, 28–29.

(1905) “Lydian and Karian inscriptions”, PSBA 27, 123–128 Lám. I–II.



505

Säflund, G.

Schweyer, A.-V.(2002) Les Lyciens et la mort. Une étude d’histoire sociale , Paris.

Schmaltz, B.(1998) “Vorhellenistische Keramikimporte in Kaunos—Versuch einer Perspektive”,

Colloquium Caricum, 203–210.

Schmitt, R.(1978[79]) Review of Masson (1978) in: Kratylos 23, 98–104.

Schürr, D.(1992) “Zur Bestimmung der Lautwerte des karischen Alphabets”, Kadmos 31, 127–156(1991–1993) “Imbr- in lykischer und karischer Schrift”, Die Sprache 35.2, 163–175.(1993) “Zu ]NAPOUKV SGDI 5727.b4”, Kadmos 32, 172–173.(1996a) “Bastet-Namen in karischen Inschriften Ägyptens”, Kadmos 35, 55–71.(1996b) “Zur karischen Felschinschrift Si. 62 F”, Kadmos 25, 149–156(1996c) “Zur karischen Felsgrabinschrift von Kaunos (28*)”, Kadmos 25, 157–159.(1996[98]) “Karisch ‘Mutter’ und ‘Vater’”, Sprache 38, 93–98.(1998) “Kaunos in lykischen Inschriften”, Colloquium Caricum, 143–162.(2000) “Lydisches III: Rund um lydisch ,Hund‘”, Kadmos 39, 165–176.(2001a) “Zur Inschrift nr. 50 von Kaunos und zum karischen Namen von Keramos”,

Kadmos 40, 61–64.(2001b) “Karische und lykische Sibilanten”, IF 106, 94–121.(2001c) “Zur karischen Inschrift auf dem Genfer Kultgegenstand”, Kadmos 40, 117–126.

(2002) “Karische Parallelen zu zwei Arzawa-Namen”, Kadmos 41, 163–167.(2003a) “Zur karischen Inschrift der Stele von Abusir”, Kadmos 42, 91–103.(2003b) “Zum Namen des Flusses Kalbis bei Kaunos in Karien”, HS 116, 69–74.

”evoro“kin, V. I. [ òÂ‚ÓÓ¯ÍËÌ, Ç. à.](1962) “ä‡ËÈÒÍËÈ ‚ÓÔÓÒ [The Carian problem]”, VJa 1962.5, 93–100.(1963) “O ıÂÚÚÓ-ÎÛ‚ËËÒÍÓÏ ı‡‡ÍÚÂÂ Í‡ËÈÒÍÓ„Ó flÁ˚Í‡ [On the Hittite-Luwian charac-

ter of the Carian Language]”, VJa 1963.3, 83–84.(1964a) “On Karian”, RHA XXII 74, 1–55.(1964b) “Aegyptisch-karische Inschrift am Sockel einer Isisstatuette (Leningrader Staats-

ermitage)”, RHA XXII 74, 57–65 and pl. I–IV.(1964c) “Zur karischen Schrift und Sprache”, Kadmos 3, 72–87.(1964d) “Karijskij jazyk; sovremennoe sostojanie de“ifrovki i izu‘enija” [La lengua caria;

el estado actual del desciframiento y la investigación] Problemy indoevropejskogo jazykoz-nanija , Moskva, 18–39.

(1965) àÒÒÎÂ‰Ó‚‡ÌËfl ÔÓ ‰Â¯ËÙÓ‚ÍÂ Í‡ËÈÒÍËı Ì‡‰ÔËÒÂÈ [Studies on the deciphermentof the Carian inscriptions], Moskva.

(1968a) “Zur Entstehung und Entwicklung der kleinasiatischen Buchstabenschriften”,Kadmos 7, 150–173.

(1968b) “Karisch und Lykisch”, Atti e Memorie del I Congresso Internazionale di Micenologia-1967 , Roma, 462–472.

(1968c) “Karisch, Lydish, Lykisch”, Klio 50, 53–69.(1969a) “Zu den ‘späthethitischen’ Sprachen”, XVII. Deutscher Orientalistentag-1968, Würzburg

= ZDMG Supplem. I. Wiesbaden, 250–271.

(1969b) “Zur Erforschung der kleinasiatischen Onomastik”, 10. Internationaler Kongress für Namenforschung II, Vienna, 341–350.(1977) “Zu einigen karischen Wörtern”, MSS 35 117–130.(1982–83) “Über den Lautwert des karischen Buchstaben y”, InL 8, 71–78.(1984[86]) “Verbesserte Lesungen von karischen Wörtern”, InL 9, 199–200.(1988) “Carian proper names”, Onomata 12, 497–505.

(1953) “Karische Inschriften aus Labranda”, Opuscula Atheniensia I, 199–205.

(1980) “Karer”, in: Reallexikon der Assyrologie und vorderasiatischen Archeologie V, 423–425.



506

(1991) “On Carian Language and Writing”, Perspectives on Indo-European Language Culture and Religion. Studies in Honor of Edgard C. Polomé. vol. I (= Journal of Indo-EuropeanStudies. Monograph Number 7). 116–135.

(1992) “On Carian Language and Writing”, Newsletter for Anatolian Studies Vol. 8/1, 3.

Shafer, R.(1961) Review of Masson-Yoyotte (1956) in: RHA XIX. 68, 39–40.(1965) “A break in the Carian Dam”, AC 34, 398–424.

Six, J. P.(1890) “Monnaies grecques inédites et incertaines”, NC , 185–259.

Snodgrass, A. M.(1964) “Carian armoures. The Growth of a Tradition”, JHS 84, 107–118.

Spiegelberg, W.(1928) “Eine Ichneumonbronze mit hieroglyphischer und karischer Inschrift”, OLZ 21,

545–548.

Steinherr, F.(1950–51) “Zu den neuen karischen Inschriften”, Jahrbuch der kleinasiatische Forschung 1,

328–336(1955) “Der karische Apollon”, WO 2, 184–192.(1957) “Der Stand der Erforschung des Karischen”, Proceedings of the twenty-second Congress

of Orientalists . . ., Istanbul, 1951, Leyde, 44–49.

Stoltenberg, H. L.(1958a) “Neue Lesung der karischen Schrift”, Die Sprache 4, 139–151.

(1958b) “Die karische Grabinschrift von Kaunos”, AC 27, 108–109.(1959) “Deutung karischer Inschriften”, ArOr 27, 1–4.

Sundwall, J.(1911) “Zu den karischen Inschriften und den darin vorkommenden Namen”, Klio 11,

464–480.(1913) Die einheimischen Namen der Lykier nebst einem Verzeichnisse kleinasiatischer Namenstämme

[= Klio 11, Beiheft], Leipzig.

Thompson, M.(1966) “Some Noteworthy Greek Accessions”, ANSMN 12, 1–18.

Tischler, J.(1977) Kleinasiatische Hydronimie. Semantische und morphologische Analyse der griechischen

Gewässernamen, Wiesbaden.(2001) Hethitisches Handwörterbuch, Innsbruck.

Torp, A.(1903) Die vorgriechische Inschrift von Lemnos = Christiania Videnskabs-Selskabs Skrifter,

hist.-fil. Kl. 1903, 4.

Tremblay, X.(1998) “Controversa Carica”, Colloquium Caricum, 109–124.

Treu, M.(1954) “Eine griechisch-karische Bilingue und ihre Bedeutung für die Geschichte der

karischen Schrift”, Glotta 34, 67–71.

Troxell, H.(1979) “Winged Carians”, in: O. Mørkholm—N. M. Waggoner (eds.), Greek Numismatics

and Archaeology, Essays in Honor of Margaret Thompson, Wetteren, 257–268.

(1994) “Carian – Three decades later”, Decifrazione del cario, 131–166.

507

Tuchelt, K.

Tzanavari, K. – CHRISTIDIS, A.-Ph.(1995) “A Carian Graf fito from the Lebet Table, Thessaloniki”, Kadmos 34, 13–17.

van den Hout, T.(1999) “-ms(-): a Carian Enclitic Pronoun?”, Kadmos 38, 31–42.

Varinlio< lu, E.(1986) Die Inschriften von Keramos , Bonn.

Vittmann, G.(1996) “Zum Gebrauch des k3-Zeichens im Demotischen”, SEAP 15, 1–12.

(2001) “Ägyptisch-Karisches”, Kadmos 40, 39–59.(2003) Ägypten und die Fremden im ersten vorchristlichen Jahrtausend , Mainz.

Waerzeggers, C.(Borsippa) “Carians in Borsippa”, Iraq (forthcoming).

Winzer, A.(2005), Antike Portraitmünzen der Perser und Greichen aus vor-hellenistischer Zeit (Zeitraum ca.

510–322 v. Chr), March-Hugstetten.

¥ába, Z.(1974[79]) The Rock Inscriptions of Lower Nubia, Czehoslovak Concession, Prague.

Zahlhaas, G. – Neumann, G.

(1994) “Zwei neue karische Inschriften aus München”, Kadmos 33, 160–168.Zauzich, K.-Th.(1972) Einige karische Inschriften aus Ägypten und Kleinasien und ihre Deutung nach der Entzi ff erung

der karischen Schrift , Wiesbaden.

Zgusta, L.KON: see (1984).KPN: see (1964a).(1964a) Kleinasiatische Personennamen, Prague.(1964b) Anatolische Personennamensippen, Prague.(1968) Review of ”evoro“kin (1965) in: ArOr 36, 153–154(1970) Neue Beiträge zur kleinasiatischen Anthroponymie , Prague.

(1984) Kleinasiatische Ortsnamen, Heidelberg.

(1984) “Carians in Miniature”, in: A. Houghton – S. Hurter (eds.), Festschrift für /Studies in honor of Leo Mildenberg , Wetteren, 249–257, pl. 40.

(1970) Die archaischen Skulpturen von Didyma (= Istanbuler Forschungen 27), Berlin.



TABLE ITHE CARIAN ALPHABET

Nº (Masson) Letters 1 Transcription Notes

1 a A ~ À (E ) a3 d D G d4 l l5(+41) W ù V V y (formerly ù/ü)6 r R r7 L 2 L l 9 q Q q

10 b 5 B b11 m M m12 o o14 t T t15 f F S òT _ “17 s s18 H ?19 u U u20 ñ ñ21 x X ∞ (formerly x )22 n N n24(+2) p p ( ) p25 z Z ≤26(+8) I í Î Ï y ì Y I i27 e e e28 w ÿ (formerly w)

29+30 k K k31 & d32 v Ú w (formerly ú)33(+34) 0 8? g 35(+36) 1 9 9 z (formerly z )37 % ∆

38 j _ j (formerly í)39 _1 ?40(+23?) c C / O? t / t2?42 6 ®

43–44–45 Ø ® 4 B &? b? b46 ÿ b2?

1 Note that, for the sake of simplicity, the letters are given in a left-to-right direc-tionality, which in some cases is not actually documented (for example, ® is onlyattested as ç in right-to-left directionality).



TABLE IICARIAN SIGNS IN COIN LEGENDS

L

6 5 [ A l (!) a M Q À a M12–21, M32B u (!) 5 H(!) b M22–23, M26–28, M32, M36–37,

M41

£ d _ d M32, M36–37, M42g r r M32, M51–52l l M50W ¥ y (or rather w?) M5, M23, M47–48L l M39M m « m M1–5o o M20–21, M50–52t F t M31, M39s s M33–37, M42u u M39, M45, M50–52

n n M25ñ ñ M32x X ∞ M43–45z ≤ M39i Ì i M31–32, M53p ° P p M33–35, M39–40k l (!) k M22–30V Ú w (or rather y?) M5–M11, M49N 0 g (?) M33–35T j M23, M36–37

3 4 8 9 z M12–21 7 1 (z or symbol?) M38–39c t M32J letter or M27, M33, M36–37

separationmark

^ letter or symbol? M46



510

L

o M1

l M31 M33–35

h M38 7 1 (or letter?) M38–39^ (or letter?) M46



INDICES1

1. A P N

Carian Proper Names in Indirect Sources

In Egyptian3rskr 19, 47, 194, 195, 243, 248, 250,

431 Jr “ 3 42, 47, 194, 248 Jrym3 42, 194, 248K3rr 33, 198, 244, 266, 409Prjm 41, 188, 194, 195, 242, 248, 249” 3rkbym 32, 191, 194, 195, 248, 250,

251, 416, 417

In AkkadianLu-uk-“ u 380

In Greek Aba 340, 349

Abaw 340Abbaw 340Ada 13, 140, 236, 245, 340, 349Adaw 340Aktadhmow 330Aktauassiw 242, 330, 338Aktaussvllow 330, 378ÉAlãbanda 8, 11, 12Alasta 349Alikarnassow 255, 351Amiaw 340Ammiaow 340

Andarsvw 246, 262, 333, 363Androsvw 246, 262Appa 340Arbhs(s )iw 334Arduberow 245, 255, 283, 333, 353Ardur/a/ 333Ariauow 354Aridvliw 354Arlissiw 196, 248, 250, 255, 341,

354, 355, 356Arlissow 355Arlivmow 248, 343, 355

Arrissiw 135, 221, 248, 305, 354, 356Artaow 255, 356Arthumow 255, 356, 393

Artimhw356, 410

Aryuassiw 356Aryuassiw 356Babein 246Bãrgasa/Bãrgaza 246BargÊlia 246Beryaw 246Berrablviow 246Bolli.evn 246Boivmow 246Br¤oula 246Bridaw 246

Bruajiw/Bruassiw 246Brvlow 246BubassÒw 246Bvnitv 246Bvrand/a/ 246Gugow 198, 237, 244, 260, 326, 330,

334, 408, 419Dandvmow 246Daru..ow 246D°dmasa 246Deibow 246Dersvmanhw 246

Dersvw 246, 363Dersv . . . tiw 246Didassai 246D¤duma 246DÊndason 246Ekamuhw, Ejamuhw 330Ekatomnvw 140, 237, 238, 243, 249,

255, 289, 375, 378Ermapiw 255, 331, 339Zermenduberow 413Yuagg°l /a/ 12Yussvllow 330, 362

1 For Carian words, see Chapter 11.

512

Iasow 270, 433Ibanvlliw 360Idagugow 237, 238, 239, 244, 245, 330,

334, 361Idriaw 308, 438Iduma 367, 368*Idurigow 368Idussvllow 245, 330, 361, 362, 367,

368Imbarhldow 198, 262, 330, 335, 393ÖImbramow (var. ÖImbrasow ) 8, 308, 335Imbarsiw 236, 247, 250, 255, 262, 335,

360, 366Imbras(s )iw 236, 247, 250, 255, 262,

335, 341, 360, 361, 366ÖImbrow 8, 299, 335, 367Karbasuand/a/ 406Karusvldow 330Kasbvlliw 334, 374Kasvlãba 237, 238, 243, 245, 375Kbondiassiw 334Kbvdhw 334, 371Kebivmow 33, 236, 237, 238, 243, 245,

249, 330, 334, 343, 371, 416, 417Kildar/a/, Killar/a/ 141, 236, 243,

248, 373

Kinduh 255, 269, 373Koar(r )end/a/, Koarenz/a/Koaranz/a/, Kvra( n)z/a/ 438

Kolaldiw 238, 239, 244, 256, 315, 407Konodvrkond/a/ 332Kostvlliw 251, 409Kotobalvw 244, 408Kotbelhmow 198, 236, 238, 244, 245,

255, 343, 408Kuatbhw 309, 408Kubliss/ow/, Kublisse›w 238, 245,

255, 269, 409

Kulaldiw 238, 239, 244, 248, 315,407Kuròmew 338Kutbelhmiw 160, 198, 236, 238, 244,

245, 255, 343, 408Latmow 440Lujhw 182, 248, 255, 380Manhw 236, 381Mareuw 335Marow 335Masanvrada 237, 238, 239, 245, 249,

250, 269, 307, 332, 333, 385Masanvradow 333, 385Massarabiw 332, 339Massariw 332Massvn/a/ 332Mãstaura 8

Maussvllow 330Minnaw 340Moiw 384

MonÒgissa 8MÒtulow 336Mouzhaw 251, 336, 387Mundow 386Nana 13, 340Nanaw 340Nannixow 340Nannow 340Nannv 340Naras/a/ 333, 387Narasow 262

Narbaw333

Narisbara 387Naruandow 333, 387Neterbimow 238, 332, 339, 343, 347Nonnow 340Nutar 332Nvtrassiw 332Jrmedu<be>row 255,Oaloalow 339, 428Oa3a3iw 242Oliatow 237, 243, 330, 339, 428Ordomaw 333

Osogva 8, 169Otvrkond/a/ 332Panablhmiw 330, 338Panamara 335, 338Panamuhw 330, 336, 338, 357Panuassiw 242, 330, 338, 340Paow 395, 419Papaw 340Papiaw 340Pãrgasa 246Par(a )ussvllow 236, 330, 340, 393,

394

Paraudigow 236, 238, 263, 330,394Parnow 417Paruv 394Parembvrda 333, 340Pedanass/ow/ 336Pedvldow 337Peigelasow 337Pel(l )ekvw, Peleqow 237, 238, 293,

399, 437Peldhkow 236, 238, 400Phdasa 336Pigassvw 337Piginda 337Pigrhw, Pikrhw 31, 236, 243, 244, 248,

255, 337, 345, 397Pidasa 336



513

Pidossus 337Pijvdarow 337, 399Ponmoonnow 140, 238, 239, 249, 255,

303, 330, 336, 338, 401Ponussvllow 197, 237, 238, 243, 249,

330, 338, 401, 405Pormounow 303, 336, 338Pounomoua 336, 338Samul¤a 336Samv(u )ow 141, 310, 416Sanortow 419*Sarkebivmow 33, 416, 417Sarnoj 417*Sar-oliatow 417

Sarow419

Saruassiw 242, 330, 338, 340*Sar-uliatow 417Sarussvllow 196, 236, 250, 263, 330,

340, 418, 419Saskvw 175Saurigow 237, 263, 289, 308, 330,

340, 418Saussvllow 263, 330, 340, 418Semeuritow 263Senurigow 263, 289, 330Seskvw 175

Sinuri 434Skoaranow 255Souãggela 8, 11, 20, 238, 251, 269,

277, 415

Spareudigow 255, 263Svmnhw 236, 237, 249, 255, 409, 414Tãbai 9Tarkonda[ 255Tarkondar/a/ 332, 423Tata 340Tatarion 340Tataw 340Territow 357Tonnouw 237, 238, 239, 243, 249, 250,

422Tralle›w, Tralde›w 18Tumnhw 440TumnhsÒw 9

Tur[ 164, 197, 288, 424 Uliatow 36, 237, 243, 330, 339,428

Ñ` Ullãrima 196, 200, 238, 240,257

ÑUlloÊala 8 Urgosvw 237, 244, 255, 434huròmew 338 Urvmow 338 Ussaldvmow 240 Usseldvmow 240 Ussvllow, Ussvldow 135, 141, 183,

195, 196, 237, 305, 310, 330, 344,362, 403, 424, 431 Xasbvw 326, 334, 374 Xeramuhw 330

In Greek Arbasiw 334Armadapimiw 339Armapia 331Armapiaw 331Arpigramow 353Arteimianow 410Arteimaw 356, 410Arteimow 356, 410Arteimhw 410Artimaw 410Artimhw 356, 410GÊghw 334Dada 13Eida 334Eidassala 334

Ermapiaw331

Ida 334ÖImbrasow 335Imrougara 335Kendebhw 373Kendhbhw 373

Kendhbaw 373Kouadapemiw 339Kougaw 334Ktibilaw 340Mollisiw 341

Mousathw 336, 386Moshta 336, 386Moushta 336, 386Moutaw 386Janduberiw 222, 245, 255, 375, 419Oliw 421Oraw 338Oualiw 421Ouajamoaw 255, 336, 427Ouajamvw 336, 427Ouramoutaw 338

Panamuaw336

Perpenduberiw 402Pigramiw 337, 361, 397Pigramow 337, 361, 397Pigrassiw 337Pijaw 399

Other Anatolian Proper Names



514

Pijedarow 399Ponamoaw 405Salaw 334Sedeplemiw 338Tabhnoi 9Tarkumbiou (gen.) 339Triendasiw 341Trokoarbasiw 334

In cuneiform sources À la- 334 Arma- 306, 331, 347 Ayami/aimi 343› u-da-ar-lá 406

› u-du-ur-lá 406› u-u-tar-li 406› u-ut-ra-la-(a “ ) 406› u-ut-ra-li-i “ 406Immaraziti 335Lukka 342 Massanaura 338 Millawa(n)da 342 Mutamutassa 342 Muwatalli 336 Muwaziti 336, 386 Nattaura 338

Petassa 336Pi ¢irim 337*Piyama-dKAL 339Piyamaradu 339, 385Pitassa 336Puna-A.A = *Punamuwa- 405Runtiya- 333Tar ¢u- 331, 347, 356, 423Tar ¢unt- 239, 247, 259, 260, 347, 356,

423Tarhundapiya 339

Tar ¢untaradu 385Ura 338Urawalkui 338

Walawala 428Walawali 428Wallarima 240

In Hieroglyphlic LuwianTarhu(n)t- (TONITRUS-hut- ) 347

In LycianErbbina 177Esedepl ‘ emi 338Hri x ttbili 340Idazzala 334Ijãna 369Ipresidi 335Katamla- 244 Mullijese/i- 341 Natr- 162, 332, 389 Natrbbij ‘ emi 238, 332, 339, 343, 347,

389Punamuwe 405Sedepl Mmi 338Teleb(ehi) 177Trij ‘ etezi- 341

Trqqñt- (nom. Trqqas ) 239, 247, 259,260, 331, 347, 356, 423Urebillaha (dat.) 318 Xbide- 203, 243, 245, 255, 297, 298,

371Zzala 334

In MilyanKridesi 407Trqqñt- (nom. Trqqiz ) 260, 331, 334 Xbadiz 261

In Greek ÉAy∞nai 237, 243, 259ÉAyhna›ow 155, 203, 237, 296ÉApollÒdotow 332, 339, 389ÉAristokl∞w 164ÖArtemiw 356ÉArt°mvn 238, 357ÉArx°laow 354Bragx¤dai 20, 251, 285, 402

Brãgxow 290ÑEkata›ow 236, 238, 288, 375, 378*ÑEkatÒmnhstow 378ÑEllhnikÒn 2ÑErm∞w 331

ÑErm¤aw 135, 305, 307ZeÊw 289, 375ÉIdrieÊw 140, 255, 438ÑIpposy°nhw 155, 203, 236, 296, 298,

316KarikÒn 2Kãttouza 9Kaun¤oi 155, 203, 296, 297, 299, 360K≈Ûoi 271

K«w8

Lusikl∞w 155, 203, 236, 237, 243,248, 296, 380

Lusikrãthw 155, 203, 236, 237, 243,248, 296, 380

Greek Proper Names



515

M¤lhtow 342Nikokl∞w 155, 203, 236, 249, 296,

380, 388OÈliãdhw 236, 245, 248, 428, 432ÉOrsikl∞w 431Perikl∞w 431P¤ndarow 242, 396SkÊlaj 164

TÊxh 308ÑUbr°aw 432Fãnhw 135, 305Crusaor¤w 416

In LycianLusãñtrahñ (gen.) 304Perikle 431

In Egyptian J‘ (?)[ 47 J.r=w 369

Jr.t=w-r.r=w 370 J-Ó r 353 Js.t 128 Jtm 32, 424W3h-jb-r‘-nb-[ 38, 387 B3st.t 251, 398P3j-Wsir 398P3-whr 405P3-n-jwjw 397P3-n-Wsjr 398P3-n-Nj.t 393P3-sb3-¢‘j-m-njw.t 350

P3-dj-Jnp 404P3-dj-Jtm 395P3-dj-‘nq.t 404P3-dj-B3st.t 245, 395P3-dj-(p3)-n∆ r 404P3-dj-Njt 33, 198, 242, 243, 245, 266,

395*P3-dj-nfr 404P3-dj-Ó r-sm3-t3wy 244, 250, 399P3-dj-sm3-t3wy 244, 250, 399P3-dj-st 39*P3-dj-t3wy 404Psm∆ k 194, 242, 249, 250, 398,

403Psm∆ k-‘wj-Njt 38, 198, 403Pt -ms 420 Nj.t 249 Nj.t-jqr 243, 248, 249, 389 Ny-k3w 244, 249, 388 Nfr-˙r 242, 388 Ns-˙r 369Ó 3py 41Ó p-mn 353

Ó r 391T3-dj(t)-wsir 39T3-dj(.t)b3st.t 245, 423Ô 3j-jm=w, Ô 3j-n.jm=w 251, 194, 195,

416, 424

Ô 3j-Ó p-jm=w 424Ô 3[ = *Ô 3-n-n3-j w 194, 425

In AkkadianTa ¢ma ““ i 420

In Greek ÖAmasiw 293ÑApimenhw 353Avw 353ÉEsour 369ÉEsouriw 369Yamvw 251, 424ÉIyorvw 370Nek«w, Nexvw 244, 249, 388

Nefervw 242, 388Nitvkriw 243, 248, 249, 315, 389Panitiw 393Peteyumiw 395Petenaiyiw 395Petenhyiw 395Petenht 395Petenoupiw 404Petepnouyiw 404Petetumiw 395Petyubestiow 395Petobastiw 245, 395Petoubastow 395Petoubestiw 245, 395Pieuw 397Pihow 397Pih# 397Pisiriw 398Potasimto 244, 250, 293, 399Pouvriw 405Povriw 405Samauw 251, 425Samw#w 251, 425

Tamvw 251, 424Tetobastiw 245, 423Cammhtixow Cam(m )atixow 293, 398,

403Cousennhw 350

Egyptians Proper Names



516

In Old Persian

Vi “ t àspa- 370

In Greek

ÑUstãsphw 370

2. L

Carian glosses in Greek êla 8, 11bãnda 8, 11g°la 8, 11g¤ssa 8ko›on, kÒon 8 10soËa( n ) 8, 10, 415

tãba 9toussÚloi 9tumn¤a 9

Cuneiform Luwian (CLuw)à/à ya- 327, 343annara/i- 262, 333, 352, 387ànnari- 333ànna/i- 347, 259, 273, 364apa- 359*arada-,* aradu- 332, 333, 385arpa- 333, 334

*a ““ att(i)-, a ““ atta ““ a/i- 334aduna (Inf.) 299-¢a 306¢ant- 372¢andawat(i)- 294¢à pa/i- 334*¢apài- 334¢ù¢a- 334, 345¢utarl à- 248, 260, 406im(ma)ra/i- 335, 346, 393, 422im(ma)ralla/i- 335

im(ma)ra ““ a/i- 335*im(ma)ra ““ iya- 335kui- 243, 259, 320, 377-mma/i- 339, 343, 347mà““ an(i)- 260, 307, 327, 332, 347mà““ ana ““ a/i- 352ma “¢a ¢it- 351muwa- 260, 335nana- 284, 321, 388-ppa 302 par ì 340 pi ¢a- 337, 345, 397

piya- 282, 284, 304, 327, 339, 396 pùna- 259, 337 pùnata/i- 337“ arra 340“ arri 260, 261, 302, 340, 411, 416

t àta/i- 259, 273, 347, 421upatit- 433ura- 338, 430u““ a/i- 344wallant- 338walli( ya)- 339-wanni- 257, 260, 346, 367, 369, 371

wayamman- 427w à“ u- 344za- 259, 288, 346, 410zar “ iya- 411

Hieroglyphic Luwian (HLuw)ataman- 355atra/i- 258, 392hasu- 334nimuwiza- 383 pa- 359 para/i 340

piya- 282, 284, 339tata/i- 347ura/i- 338wa/iliya 339-wani- 346za- 259, 288, 346, 410

Hittite (Hitt.)anda 287andan 254, 287, 363anna- 259 , 347ak(k)- 350ar ¢a 304ard- 333atta- 347¢anna- 332¢ant- 372¢u¢¢a- 334innar à- 333, 387ka- 288, 346, 410k àn(i) 259, 288kappi- 352katta 300, 351

kattan 300 gim(ma)ra- 335, 346kui- 243, 259, 260, 320, 377 gurta- 407mald- 323, 427

Old Persian Proper Names



517

malai- 427mid à(i)- 382muwa- 335, 386

nà(i)- 284, 321, 388naru- 333nu- 285, 390 pài-/piya- 282, 284, 304, 325, 339 peda- 304, 336 par à 259, 260, 340 piyannài- 359 pulla- 401“ ara 340“è r 260, 340, 416“ iu(n)- 332-talla- 428ura/i- 338walli- 339*walli- 338walliwalli- 260, 338, 339, 428wiyài- 427

Lycian (Lyc.)a(i)- 299, 300, 301, 309, 321, 347, 349arawa- 300, 363, 435ar Mma 258, 347atãnaze/i- 259

atra/etli- 204, 258, 297, 392-be 302ebe- 290, 320, 359ebeli 365ehbi 348ehbi( je)- 300‘ ene/i- 259, 273, 313, 347, 352, 364epñn‘ eni 352erbbe- 333eri- 304, 364ese- 294hri- 260, 261, 302, 340, 416

hrppi 302ipre 335kbatra- 258maha(na)- 260, 307, 327, 332, 347, 352mahanahe/i- 313, 352maraza- 335me 324-me/i- 339, 343, 347mere- 335, 422mle- 307Mmai- 304, 390n‘ eni 352nere/i- 292-ñni 204, 260, 346, 369, 371ñte 254, 287, 319, 363 pdd ‘ en- 247, 304, 336, 396

pije- 282, 284, 304, 325, 327, 339, 396-ppi 302 pri 40

prñnawa- 290 punãma- 259, 337r Mmazata 254se 347, 411s=eññe 309s ‘ eñnait ‘ e 413sije- 259, 412tãtu (imperative) 349tede/i- 245, 259, 260, 273, 346, 347,

421ti 243, 259, 320, 377tike 377tr Mmili( je)- 270tuwe-, tu-s- 325ube- 201, 239, 245, 259, 282, 321, 327,

347, 432uhe/i- 344 x ahba 258, 261, 326, 334 x awa- 10xbid ‘ eñne/i- 203 x ñtawa- 294 x ñtawat(i)- 10, 260, 294, 364 x uga- 260, 326, 334, 345

x upa- 290, 429-za 294-ze/zi- 270, 392zeus- 288, 375

Milyan (Mil.)dd x ug [ 362erbbe/i- 334erbbesi 334esetesi 334ki 243, 259, 320, 377-kike 377

masa- 260, 261, 237, 307, 332, 347muwa- 260, 335sebe 201, 239, 245, 327, 347, 411-wñni 204, 257, 260, 346, 371 x ñtaba 373 x ñtabasi 373 x ruwasaz (nom.-acc. pl.) 261 x uga- 334zri- 260, 261, 416

Lydian (Lyd.)armt a- 347artimu- 356bi- 359bil-l 396civ 332



518

‘ ena- 346, 364i ≤ tube l m- 370kãn- 406

sareta 411serli-/selli- 261taada- 260, 346, 347, 421-w n 298

Sidetic artmon 357ma ≤ ara (dat. pl.) 260, 332,

347

Pisidianoudoun 426

Proto-Anatolian (PA)*anna-s 259*armà- 258, 260*/be:H/- 257*dáda-s 259* · emro- 262, 346*H àwo- 10*HuHo- 260*∞o-/∞à- 259, 410*k w is 259, 266, 320, 377

*obó/i- 359* pédon 336 pr ò /pr ò 259, 260*T º H –t- 259

Greek êggelow 11êmbrotow 262aÊtoÊw (acc. pl.) 392•autÒn (acc.) 392¶doje (aor.) 426eÈrg°taw (acc. pl.) 428, 429

yeÒw 11kombow, kombion 352p°lekuw 399proj°nouw (acc. pl.) 382, 411s∞ma 288, 415Ïrxh 391

Latinorca 391urceus 391

urna 391

Old Persiank º ka- 1haya/taya 320

Proto-Indo-European (PIE)*bhè h2- 246, 257, 260, 337, 345, 361*dyew- 332*h2e/owo- 10*h2ent- 260*h2n¶ r 387*h2onsu- 261*∞e 411*∞ei- 259, 412*∞o-/Keh2 – 288, 410*k w e 182*k w i- 202, 259, 266, 320, 377neyH- 284, 321, 388* –-mº to- 262*nu- 285* pedo- 247, 336*terh2- 331

* 8elH- 339* 8elh1- 339

Akkadian*narû 333

Egyptian‘3 32‘n¢ 32, 41, 128 ym 33wnwtj 278, 429 p3 w m 41, 355

p3 ˙m-n∆ r n Jmn 400n 47n∆ r 32, 389s3 47snb 32dj 32, 41, 128, 399

Coptic eiom 33



PLATES



521

M a p 1 .

C a r i a n i n s c r i p t i o n s i n C a r i a .



522

Map 2. Carian inscriptions in Egypt.



523

Plate 1



524

Plate 2



525