Pre-Greek

Brill Introductions toIndo-European Languages

Series Editors

Michiel de VaanAlexander Lubotsky

volume 2

The titles published in this series are listed at brill.com/iiel

Pre-Greek

Phonology, Morphology, Lexicon

By

Robert S.P. Beekes

Edited by

Stefan Norbruis

leiden | boston

Library of Congress Cataloging-in-Publication Data

Beekes, R. S. P. (Robert Stephen Paul), authorPre-Greek : phonology, morphology, lexicon / By Robert S.P. Beekes ; Edited by Stefan Norbruis.

p. cm. (Brill introductions to Indo-European languages ; Volume 2)Includes bibliographical references and index.ISBN 978-90-04-27938-4 ((pbk.) : alk. paper) ISBN 978-90-04-27944-5 (e-book)1. Greek language, ModernGrammar. 2. Greek language, ModernPhonology. 3. Greek language,

ModernMorphology. 4. Greek language, ModernSyntax. 5. Greek language, ModernLexicology. I.Norbruis, Stefan, editor. II. Title.

PA1058.B44 2014487'.1dc23

2014027870

This publication has been typeset in the multilingual Brill typeface. With over 5,100 characters coveringLatin, ipa, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For moreinformation, please see www.brill.com/brill-typeface.

issn 2214-5605isbn 978-90-04-27938-4 (paperback)isbn 978-90-04-27944-5 (e-book)

Copyright 2014 by Koninklijke Brill nv, Leiden, The Netherlands.Koninklijke Brill nv incorporates the imprints Brill, Brill Nijhoff, Global Oriental and Hotei Publishing.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 granted by Koninklijke Brill nv providedthat 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.

Contents

Preface ixAbbreviations x

1 Introduction 1

2 Phonology 41 The Phonemic System of Pre-Greek 42a Characteristic Phonemes and Phoneme Clusters 92b How to Recognize Words as Pre-Greek? 133 Prothetic Vowel 134 s-mobile 145 Consonant Variation 14

5.1 Voiceless / Voiced / Aspirated Stop 145.2 Prenasalization 145.3 Nasalization 145.4 Labial Stops / / u 155.5 Stops Interchanging with (), with Stop + / or with +

Stop 165.6 Velar / Labial / Dental Stops; Labiovelars 195.7 Dentals / Liquids 205.8 Simple / Geminate 215.9 / Zero 215.10 Velar or Dental Stop / Zero 215.11 -, - / Zero 225.12 Metathesis, Shift of Aspiration 225.13 Secondary Phonetic Developments 225.14 Other Variation 22

6 Vowel Variation 236.1 Single Vowels (Timbre) 236.2 Long / Short 256.3 Single Vowel / Diphthong 256.4 Rising Diphthongs? 266.5 Secondary Vowels (or Elision) 26

3 Morphology 271 Reduplication 272 Suffixes 27

vi contents

2.1 Introduction 272.2 Survey of the Suffixes 282.3 The Suffixes 29

3 Word End 433.1 Words Ending in a Vowel 433.2 Words Ending in - 443.3 Words with a Nom. in - or - 443.4 Words in - 443.5 Words Ending in - (-stems) 44

4 The Unity of Pre-Greek 45

5 Pre-Greek is Non-Indo-European 46

6 The Pre-Greek Lexicon 471 Landscape and Natural Phenomena 472 Minerals 513 Flora 54

3.1 Trees and Shrubs (and Their Products) 543.2 Wild and Cultivated Vegetables, Fruits and Nuts 583.3 Aromatic, Medicinal and Toxic Herbs 623.4 Other (Useful) Plants and Flowers 663.5 Fungi 70

4 Fauna 714.1 Domesticated Animals and Their Attributes 714.2 Other Mammals (Also Aquatic) 734.3 Birds 754.4 Reptiles and Amphibians 794.5 Fish 804.6 Molluscs, Crustaceans, and Other (Marine) Invertebrates 844.7 Insects and Arthropods 874.8 Worms (Also Parasites) 90

5 Agriculture 905.1 Cereal Culture 905.2 Viniculture 92

6 Prepared Food 947 Human Physiology 97

7.1 The Human Body 977.2 Affections and Diseases 1027.3 Sex 105

contents vii

8 Attire and Jewellery 1069 Equipment and Utensils 108

9.1 Furniture 1089.2 Containers 1099.3 Dishware 1139.4 Domestic and Craft Tools 1149.5 Hunting and Fishing Equipment 1179.6 Armor andWeaponry 1179.7 Horse Tack 1199.8 Means of Transport 1209.9 Other Technical Terms 121

10 Construction 12410.1 Architecture and Constructional Elements 12410.2 Infrastructure 127

11 Society 12811.1 Social Hierarchy and Administration 12811.2 Military Expressions 13011.3 Professions and Other Societal Appellatives 131

12 Culture 13312.1 Contest 13312.2 Sculpture 13412.3 Musical Instruments, Performing Arts 13512.4 Religious Festivals and Feasting 13612.5 Divine and Numinous Beings, Priests and Temples 137

13 Adjectives Marking a Certain Quality 13814 Abstract Expressions 14815 Verbs 15016 Adverbs 15817 Sounds and Interjections 15918 Theonyms, Divine Epithets, Mythical Characters 16019 Toponyms and Ethnonyms 163

Bibliography 165Index 170

Preface

One of Robert Beekes main objectives in writing his Etymological Dictionary ofGreek (edg)was to collect and analyze the Pre-Greekmaterial contained in theGreek lexicon, building especially on studies by F.B.J. Kuiper and E.J. Furne,and on his own research conducted over several decades. The present bookprovides an overviewof Beekes findings,with sections onphonology,morphol-ogy, linguistic unity and affiliation, as well as a lexicon of Greek etyma whichin Beekes view are certainly of Pre-Greek origin (those marked with pg inedg).

When working on this book, Beekes received assistance from AlexanderLubotsky and Wouter Henkelman, who compiled and classified the lexicon,respectively. Unfortunately, Beekes health did not allow him to complete thebook himself; he was unable to work on it after February 2010. Some parts,especially the lexicon, were still in a rather rudimentary condition at this point.Initial efforts to remedy thisweremadebyMichal Peyrot,who suppliedpart ofthe lexiconwith references to thepreceding chapters. Later, Iwas asked tobringthe book to completion; I am grateful to Alexander Lubotsky for entrusting mewith it.

As a starting point in the creation of the lexicon, themost important textualelements pertaining to Pre-Greek had been extracted from the correspondinglemmas in edg. Where necessary I have reestablished cohesion, added moreinformation from edg or supplemented the argumentation. The practice ofadding references has been extended to the rest of the lexicon. In the firstchapters, I have corrected some inconsistencies and typographical errors, andadapted and added a number of variations and suffixes on the basis of thelexicon. At some points in the book I have added suggestions of my own oradduced new material; such additions are introduced by [sn]. Finally, I havestandardized the format of the bibliographical references and compiled thesection on abbreviations, the bibliography and the index.

Stefan NorbruisApril 2014

Abbreviations

For book abbreviations, see the bibliography.

1 Languages

Aeol. Aeolic

Alb. Albanian

Arc. Arcadian

Arm. Classical Armenian

Att. Attic

Berb. Berber

Boeot. Boeotian

Corcyr. Corcyrean

Cret. Cretan

Cypr. Cyprian

Dor. Doric

Epid. Epidaurian

Etr. Etruscan

gloss. in glosses

Go. Gothic

Gr. Greek

Heracl. Heraclean

Hitt. Hittite

Hurr. Hurrian

ia Ionic-Attic

ie Indo-European

Ion. Ionic

Lac. Laconian

Lat. Latin

Lesb. Lesbian

Lith. Lithuanian

Locr. Locrian

Luw. Luwian

Lyc. Lycian

Maced. Macedonian

Meg. Megarian

Mess. Messenian

MGr. Middle Greek

MoE Modern English

MoGr. Modern Greek

Myc. Mycenaean

ohg Old High German

ORu. Old Russian

Pamph. Pamphylian

pg Pre-Greek

PGm. Proto-Germanic

Phoc. Phocian

pie Proto-Indo-European

Rhod. Rhodian

Syrac. Syracusan

Thess. Thessalian

2 Grammar and Text

acc. accusative

acc. to according to

adj. adjective

adv. adverb

aor. aorist

bc before Christ

c. century

cf. compare

cod. codex, manuscript

coll. collective

abbreviations xi

compar. comparative

conj. conjecture

dat. dative

du. dual

e.g. for example

en ethnonym

et al. and others

f. feminine

fut. future

gen. genitive

hn hydronym

id. idem

i.e. that is

inscr. inscription(s)

intr. intransitive

ipv. imperative

litt. literature

m mountain

m. masculine

med. middle

ms. manuscript

n. neuter

nom. nominative

opt. optative

ostr. ostracon, ostraca

p. page

p.c. personal communication

pap. papyrus, papyri

perf. perfect

pl. plural

plpf. pluperfect

pn personal name

pres. present

s.v. sub voce

sch. scholium

sg. singular

subst. substantive

tn toponym

v. verb

var. variant

vel sim. or similarly

viz. namely, to wit

v.l. varia lectio

voc. vocative

3 Authors andWorks

Only the most common authors and works are mentioned here. Please refer tolsj for a complete list.

A. Aeschylus

A. R. Apollonius Rhodius

ab Anecdota Graeca, v. i

Ael. Aelianus

Alc. Alcaeus

Alcm. Alcman

Anacr. Anacreon

ap Anthologia Graeca

Ar. Aristophanes

Archil. Archilochus

Arist. Aristoteles

Ath. Athenaeus

Call. Callimachus

Com. Comedy, Comic

Cratin. Cratinus

D. Demosthenes

D. S. Diodorus Siculus

Dsc. Dioscorides

E. Euripides

em EtymologicumMagnum

Epich. Epicharmus

Eub. Eubulus

xii abbreviations

Eup. Eupolis

Eust. Eustathius

Gal. Galenus

Gp. Geoponica

Gr. Grammaticus

gramm. grammarians

H. Hesychius

h. Hom. Hymni Homerici

Hdn. Herodianus

Hdt. Herodotus

Hell. Hellenistic

Hes. Hesiodus

Hippon. Hipponax

Hom. Homer

Hp. Hippocrates

Il. Iliad

Luc. Lucianus

lxx Septuagint

Lyc. Lycophron

lyr. lyric poetry

Lys. Lysias

Nic. Nicander or Nicias

Od. Odyssey

Opp. Oppianus

Paus. Pausanias

Ph. Philo

Pherecr. Pherecrates

Phot. Photius

Pi. Pindarus

Pl. Plato

Plb. Polybius

Plin. Pliny

Plu. Plutarchus

poet. poetica

Poll. Pollux

post-Hom. post-Homeric

S. Sophocles

Sapph. Sappho

Str. Strabo

Suid. Suidas

Tab. Heracl. Tabulae Heracleenses

Th. Thucydides

Theoc. Theocritus

Thphr. Theophrastus

trag. tragic, tragedy

X. Xenophon

koninklijke brill nv, leiden, 2014 | doi: 10.1163/9789004279445_002

chapter 1

Introduction

The substrate language of Greek will here be called Pre-Greek; this is a trans-lation of the German term das Vorgriechische. No written texts exist in thislanguage, but it is known from a considerable number of loanwords in Greek.

The study of Pre-Greek has had an unfortunate history. In the past cen-tury, it was called Pelasgian and considered a dialect of Indo-European. Thisidea fascinated scholars, and research concentrated on this proposal. But thewhole idea was clearly wrong. The latest attempt to defend it was HeubecksMinoisch-Mykenisch (discussed by Furne 1972: 5566), where the materialwas reduced to some ten words. The theory, which has done much harm, hasby now been tacitly abandoned and it is currently generally agreed upon thatthe substrate was non-Indo-European. Therefore, the term Pelasgian can nolonger be used. Frisk already had strong doubts about the Pelasgian theory, butnevertheless often mentioned the proposals of its adherents in his dictionary.Since all work following this line has turned out to be useless, I decided tomakeno mention of the theory anymore.

When Frisk completed his dictionary in 1972, Furnes book Die wichtigstenkonsonantischen Erscheinungen des Vorgriechischen, which was his disserta-tion written under the supervision of F.B.J. Kuiper, had just appeared. It was anelaboration of Kuipers 1956 study on Greek substrate words, which opened anew chapter in the research of the field. Furne rejected the Pelasgian theory,too (see especially 1972: 4055).

Furnes book met with fierce criticism and was largely neglected. In myview, this was a major mistake in Greek scholarship. True, some of his identifi-cations are improbable, and his repeated claim that certain formswere expres-sive leads nowhere. What remains, however, is that he studied a great numberof relevant forms and drew obvious conclusions from them. Pre-Greek wordsoften show a type of variation which is not found in inherited words. It is self-evident that this variation must be studied, and this is what Furne did. It hasturned out (as Kuiper had already shown) that this variation shows certainrecurrent patterns and can be used to recognize Pre-Greek elements.

Furnes book is not easy to use: every form is discussed at three or fourplaces, each time in a different context, so that it can be difficult to find outwhat his point really is. On the other hand, his treatment is very careful, andthere are hardly any obvious mistakes. I found a number of cases which he hadnot recognized (e.g. ), but this does not change the fact that his book

2 chapter 1

was the best collection at the time. Furne worked on it for twenty years, andeven now it is the only handbook on the subject. The short overview whichfollows below is based on Furnes material and on my own research of morethan thirty years.1

Furne went astray in two respects. First, he considered almost all variationto be of an expressive character, which is certainly wrong: it is evident that thevariation found is due to the adaptation of words (or phonemes) of a foreignlanguage to Greek. We shall see below that many variants can be understoodin this way. Secondly, Furne was sometimes overzealous in his search forinner-Greek correspondences. Many of Furnes discoveries are brilliant (sees.v. in 6.3.4 for an example), but sometimes he went too far: not everyalternation necessarily points to Pre-Greek origin. The author can hardly beblamed for his enthusiasm. He was exploring new ground, and it can only beexpected that he sometimes overplayed his hand.

Several scholars were baffled by Furnes proposals and hence rejected thewhole book altogether. His method, however, was sound, and I have onlyfiltered out the improbable suggestions. In many cases, of course, absolutecertainty cannot be attained, but this should not be an objection. Except for avery small number of cases, Furnes material does consist of actual Pre-Greekwords. His index contains 4,400 words, and taking into account that many ofthese words concern derivatives and variants, as well as a few Indo-Europeanwords, I estimate that Furnes book discusses some 1000 Pre-Greek etyma.2

In general, I have given only a few personal names and toponyms, and nomaterial of this kind fromoutsideGreece andAsiaMinor. The comparisonwithBasque or Caucasian languages has not been considered here, as it is beyondmy competence; it is likely that there are such connections, but I readily leavethose for other scholars to explore.

My suggested reconstructions are not essential. One may ignore them andjust consider the variation itself. These variants are often explained as inci-dental phenomena (assimilation, influence of other words, etc.), and suchexplanations may sometimes be correct. However, if the variation occurs fre-quently, a Pre-Greek origin must be considered. The etymological dictionariesbyChantraine andFrisk often seem to avoid the conclusion that aword is a sub-strate element. It is remarkable that Chantrainewas quite aware of the problem

1 Since Kuiper was my supervisor as well, I was acquainted with the book from the very

beginning (see my review in Lingua 36, 1975).

2 Note that Furne often adduces newmaterial that is not mentioned in the current etymolog-

ical dictionaries, mostly glosses from Hesychius.

introduction 3

in his Formation, but in his dictionary he often withdrew his earlier evaluation(which in my view was the correct one). It looks as if substrate elements werenot welcome.

The relationship with Anatolian languages is a separate problem. A Greekword is often called a loan from anAnatolian language, while itmay just as wellhave been borrowed from the Pre-Greek substrate. It is generally accepted, onthe basis of toponyms, that there once was a language which was spoken bothin Greece and in western Asia Minor.3 In most cases, however, it is impossibleto distinguish between substrate words and loans from Asia Minor (the latterare from a later date). A word may have been adopted through commerce,whichmust have been a regular phenomenon, or may have resulted from localborrowing in Asia Minor, from the time when Greeks settled there, probablyas early as the 15th century. From a methodological point of view, I think it isbetter to consider suchwords to be Pre-Greek, and to define themas loanwordsfrom an Anatolian language only when there is reason to do so. Still, it isclear that we may often make mistakes here. A case in point is clew,ball of wool ready for spinning. This word is clearly related to Luwian andHitt. taluppa/i- lump, clod. The Greek word is typical of Pre-Greek words:the structure CaC-up- (with a appearing as o before u) and the absence ofan Indo-European etymology (Melchert 1998 is not convincing) imply thatthe word is Pre-Greek or Pre-Anatolian. On the other hand, a word meaningclew is perhaps not easily brought from overseas; it is an everyday wordthat the speakers of Greek and Anatolian must have picked up not far fromhome. I completely agree with Furnes interpretation (1972: 3533) that theword was brought to Greece by settlers from Anatolia who spoke the languagewhich, from another perspective, we call Pre-Greek. In other words, is a loan from an Anatolian language, but this (probably non-Indo-European)language was also spoken in large parts of Greece before the Greeks (speakingan Indo-European language) arrived there.

It is essential to realize that substrate words are a frequent if not mundanephenomenon. Regret over this fact (from an Indo-Europeanist point of view, orotherwise) is irrelevant, and should not keep one from accepting the evidentexistence of Pre-Greek words. To me, it is fascinating that they allow us tocatch a glimpse of the oldest known language of Europe (including Anatolia),of which we otherwise have no evidence.

3 A point for further study is to establish how far to the east such related names can be found.

It is my impression that these names can be found as far to the southeast as Cilicia.

koninklijke brill nv, leiden, 2014 | doi: 10.1163/9789004279445_003

chapter 2

Phonology

1 The Phonemic System of Pre-Greek

Voiceless, voiced and aspirated stops may interchange in Pre-Greek words,without any apparent conditioning factors. This fact shows that voice andaspiration were not distinctive features in Pre-Greek.1 On the other hand,the Linear B signs (graphemes) for rjo, rja and tja show that palatalizationprobably was distinctive. This is confirmed by the sign pte (e.g. in ra-pte-re/hraptres/ with the agent suffix -tr-), which must go back to an earlier pye.In the Pre-Greek material, such a phonememay underlie examples like .One may wonder whether points to py > pt which was realizedwith aspiration. Further, the signs two, twe, dwo, dwe, nwa, swa, swi, point tolabialization as a distinctive feature, i.e. two, twe, dwo, dwe, nwa, swa, swi. Note thatpalatal and labial forms of graphemes are foundwith both resonants and stops,which is a phenomenon alien to Indo-European languages. The existence oflabiovelars is confirmed by qa-si-re-u = , etc. (see further Beekes 1995/6:12 f.). We may thus posit the following system:2

p py pwt ty twk ky kws sy swr ry rwl ly lwm my mwn ny nw

Of course, it is possible that one ormore of the posited phonemes did not occurin Pre-Greek (e.g.,my is a rare sound in the languages of the world).

1 Of course, it could be due to the fact that a different distinctionwas present in Pre-Greek (like

fortis / lenis, found inmost Anatolian languages), but no obvious distribution pointing in this

direction can be discerned in the material.

2 Note that I distinguish between palatals of Pre-Greek origin, which are indicated by a super-

script y (e.g. ky), and palatovelars of Indo-European origin.

phonology 5

We can now use this insight in explaining the surfacing Greek forms. Thus, / ()- can be explained from a Pre-Greek form *dakwn-.3 In theformer form, the labiovelar yields a labial stop . In the latter, it is renderedby --, with anticipation of the labial feature, while the labiovelar turns up asa velar, possibly by dissimilation from ukw. Again, note that aspiration is notphonemic in Pre-Greek. It is very important to note that we cannot predicthow a Pre-Greek form will surface in Greek: sometimes a stop turns up as anaspirate, sometimes as a voiced or voiceless stop (e.g. / , see 2.5.1below). As a consequence, it may happen that there are a large number ofvariants, but it may also be that there are no variants at all.

As a second example, we may understand / Lesb. from a pre-form *ankwn. The latter form is directly understandable, with from thelabiovelar. The first form went through *anwkn or *awnkn, giving withloss of the nasal (a development known from Armenian). Perhaps, a scenario*akwn > is also possible, with a prenasalized form *ankwn (> )besides *akwn.4 Such interpretations may be wrong in individual cases, butthis is no reason not to apply the principle as an analytical tool. Variation thatis strange from an Indo-European point of view often becomes understandablein this way, starting as we do from a limited set of assumptions.

The existence of palatalized phonemes in Pre-Greek may explain a numberof other developments. Thus, I assume that a geminate may continue Pre-Greek *ly. We know that ie *ly gave in Greek, but if a variant with single coexists, we should be warned. For example, the name has a vari-ant with single . And although the latter only occurs in Homer, thevariation points to Pre-Greek origin. The variant was preserved because it wasmetrically convenient, it was not created for metrical purposes. Of course, thefact that there was more variation at an earlier date is what we expect. As faras the other palatalized resonants are concerned, *anymay have given , *arymay have given (or , with coloring of the vowel, see section 3.2 below onthe suffixes), etc. We have --, -- but no *-- in words of Pre-Greek origin.This view is corroborated by the fact that geminate is very frequent (Furne1972: 387), whereas geminate , and aremuch less frequent, or even rare.

3 Although I assume that voice was not distinctive in Pre-Greek, I do write d- in this case,

because only - surfaces in Greek. We must avoid losing information present in the Greek

forms. Thus, my notation of Pre-Greek forms is heuristic to a certain degree, and not always

consistent with the phonemic system I tentatively reconstruct here.

4 On prenasalization, see 2.5.2 below. As an alternative, an Indo-European etymology starting

with the root *h2emgh- to tie, betroth, can be offered; see edg s.v. (although I prefer the

analysis given here).

6 chapter 2

In a similar fashion, *asy may have yielded either -- or --, cf. -, which has a v.l. . In rendering such a foreign word, the palataliza-tion may have been represented at one time, and may have been neglected atanother. This phenomenonwas themain cause of variation in Pre-Greek forms.The interpretation is further confirmed by the parallel development of labial-ized consonants. Thus, I suppose that *arw resulted in -()- (see 3.2.3.32). Inthisway,wemayunderstand / fromapre-form*kalarw-op-.Another form which shows the remarkable interchange / is /. Here one might assume a pre-form *arwask-at-. Note that the labialelement would at the same time explain the o as a variant of a in both cases.A similar mechanism must be at the basis of the etymon , , ,, which is hopeless from an Indo-European point of view. I assume thatall forms go back to Pre-Greek *alw-ak-. It gives - through anticipation,- through coloring. In this way, the first two forms, which are best attested,are directly clear. Further, / / interchange frequently, which explains and ; - is not problematic either, as both /a/s were colored to[o] by the labialized resonant. Only the Homeric accusative is problem-atic: it is the only form that has no vowel between and , and therefore maybe due to some accident of the tradition.

I do not know whether a diphthong is allowed in suffixes of the structurevc, cf. the forms in -. Structurally, one could think of *-ayw-, or even*-awy-, but such sounds are rather rare in the languages of the world. Anexample of -- due to a palatalized consonant is / / (a brilliant grouping by Furne 1972: 158, etc.), which must contain *-apy- (thepalatalization was ignored in the last form). Comparable to the developmentin is / , from *kyn- with representing palatalization,cf. Beekes (2008). Likewise, I assume that / points to *pynut-.Perhaps, wemust interpret as *syp- because of . An interestingcase is , for which I assume *lym- besides *alym-with prothetica (see 2.3 below on the prothetic vowel).

A palatalized consonant could color a to e. A good example is ,, but also , , where we have all possible variants dueto the palatalized consonant. Compare further / . Likewise,we have parallel to , where the interchange occurs after fromearlier palatalized ty. / may have had *-pty-; () parallel to() goes back to *alyap-, with the common variation a / o before a labial.A clear example is besides and , . It may beinterpreted as representing pg *lasyt-.

Kuiper (1968) already pointed out that the substrate language had labiove-lars. He especially drew attention to vs. , . I added a few

phonology 7

remarks in Beekes (1995/6: 12). From Mycenaean, we have a-to-ro-qo (-) and qe-to (), Mo-qo-so (), qi-si-pe-e (the dual of ). Furtherthere is A-i-ti-jo-qo (gen. ), with the variants and(), which cannot be explained from Indo-European. Instead of ,we would perhaps expect **. So the developments are largely as those ofinherited Greek, but not completely.

Pre-Greek probably had a /y/ and a /w/. Initial ya- presumably often lost itsy-, but it may sometimes be represented by - as in , . The ending-mayhavebeen *-uy-a (a Pre-Greek ymayhavehad adifferent developmentfrom y in inheritedwords). In the sameway, -mayderive frompg *-ay-awitha variant -, cf. . Perhaps, the y disappeared in some cases, giving besides (see 3.2.3.7 on the suffix -- / -()-).

Initial w- was often lost (), but wa- may also have been rendered by-, as in besides Cret. . The same holds for , which hasbeen considered to be identical with the root of . We find - (whichbecame -) in , Cret. . Furne (1972: 377) assumes a prothetic- in the latter word, but this seems improbable to me. Another example maybe /. The differences are probably due to the date at which the wordwas borrowed and depend on whether the Greek dialect concerned still hada at that time. Another treatment can be found in the word for truffle,for which we find , (also --), (also --), or . These areprobably all renderings of *wit-. Furne (1972: 184) again assumes a protheticvowel, - / -, which does not seem to be the right solution. He furtherassumes a variation *wit- / wut-, which also seems improbable to me, thoughthe variation / is attested. Rather, - is a form of -, with the -o- changedunder influence of the -- (cf. Lejeune 1972: 174, and note that Greek did notallow -- before consonants; of course, became in Boeotian in the 3rd c.bc; variation / is found in more Pre-Greek words). This case nicely showsthat variation in Pre-Greek words is due to different rendering of the sounds ofa foreign language, and therefore has to be taken seriously. (H.) probably attests a development *wrak- > - (as Furne 1972: 147 remarkson : Die langlufige Etymologie connecting ist wohl ohneweiteres aufzugeben.). sorb-apple (H.) continues *sorw-(cf. Lat. sorbus, Fr. sorbier, Furne 1972: 230).

It seems that there was no initial aspiration in Pre-Greek. Furne has a fewwords with -, - (one or two with -; none with -, -, -). Several of these aredoubtful; the best is (). One might conclude that the languagehad no h. This would agree with the fact that aspiration is not a distinctivefeature in the stops. However, this conclusion is remarkable for , and , which we expect to be Pre-Greek words (but note that Myc.

8 chapter 2

a-pa-i-ti-jo does not have a2-). Of course, aspiration may have been addedsecondarily in Greek in individual cases, cf. the variation in / and / , which is a variant of . However, Ruijgh pointedout to me that Mycenaean had toponyms (a2-ra-tu-wa) and personal names(a2-ku-mi-jo) with initial h-; it also occurs in inlaut (pi-a2-la, ko-ri-a2-da-na); cf.further e-ma-a2 (/Hermhs/ Hermes).

Originally, I thought that Pre-Greek had only three vowels: a, i, u. The Greekwords concerned often have and , but this would not be surprising, as thethree vowels have a wide phonetic range, and the phoneme /a/ may havesounded like [e] or [o] in many environments. The main reason for me toassume this simple three-vowel system was the fact that the system of suffixeshas a, i, u, but not e, o. We have --, --, --; prenasalized --, --, --;likewise --, --, --; and prenasalized --, --, --, but no forms with-()-, -()-, etc. The only cases I noticed are and (butas a variant of ), and with a variant ().

Recently, I have become more inclined to assume a system with the usualfive vowels, because there seems to be a distinction between the two variations / and / , on the one hand, and a stable, not interchanging , on the other.This would point to a system with a, e and o. On the other hand, it is difficultto explain why the suffixes do not show the same variation that we find in theroot vowels.

It is essential that the palatalized and labialized consonants colored anadjacent to and , respectively. On the effects of palatalized consonantssee Beekes (2008: 4655). Furne (1972: 340) has a rule > before , , (e.g. / ); this can now be understood as the o-like realization of /a/before high rounded vowels in the following syllable (see 2.5.13.2).

So, e and o originally were variants of the phoneme /a/. It is difficult toestablishwhether theyhadalreadybecome full phonemes inPre-Greek.Agoodillustration of the case is the name of Apollo. In Hittite, Appaliunas rendersApollon- (see Beekes 2003c).We know that Greek originally had-, with --arising from -a- before the palatalized ly. The -o- developed only later in Greek,but I assume that the Hittite form still shows the -a-. The Pre-Greek form was*Apalyun-.

I have long doubted (and still doubt) whether there was phonemic vowellength in Pre-Greek. Greek substrate words quite often only have a form witha long vowel. Vacillation is sometimes found, as in / (see 2.6.2),and note / , . Quite a different argument is the following: and bothmean chaff ; it is therefore probable that they containthe same suffix --; but in the first word the u is short, while it is long in thesecond.

phonology 9

Note that often represents ( / -), and as our knowledge ofthe relevant dialects is rather limited, we often simply do not know whether represents an older a or e. If we had not had Dor. , we would nothave known that it contains an old . Also, represents . Thereare well-known Pre-Greek words with < *, like .

I assume twodiphthongs,ai andau. If therewereno e ando, wedonot expectother diphthongs. A diphthong is rare (Furne 1972: 3535; I found some 12instances in the whole of Furnes material); it interchanges with . Furne(1972: 3392) calls (in mehreren Fllen) nur eine Nebenform von . Also is rather rare, and we may find more often, but mostly interchanging withother vowels (see the remark on the suffix --). See further section 2.6.1 onvowel variation.

Regarding the accentuation, I noted vacillation in: / -; /-; / -; / ; / ; / ; - / ; / ; / . Also note the almost identicalforms such as / . This does not imply that the language hadno clear stress: the Greeks who adopted a word could simply have been uncer-tain about it. The phenomenonmay, however, be important heuristically: suchvariation is very rare in inherited words.

2a Characteristic Phonemes and Phoneme Clusters

In Pre-Greek words, we find some sounds or clusters that are rare in pie words.I give the variants between brackets.

1 does occur in pie words, but only when it derives from *h2eu (mostly in ini-tial position) or *eh2u. pg examples: , , , ,, ; .

2 As is well known, *b was rare in pie. In Pre-Greek words, it seems to occurrelatively often. Examples: , , , , ,, . It is frequently found word-initially. Of course, may alsogo back to a Pre-Greek labiovelar (i.e. labialized velar), e.g. , Myc.qa-si-re-u.

10 chapter 2

3 The cluster is possible in pie words, but it is rare (see on sub 2 above).Examples: , , , , -, , ;.

4 Cf. Furne (1972: 3185). There is nothing against pie *gd, but it is infrequent. Ofcourse, the group is reminiscent of . Examples: , , (cf.), , , .

5 Example: (). On , , see 3.2.

6 The sequence is rare in ie words. Examples: , , , - (-), ; .

7 The group is regular in pie, but in Pre-Greek it is found with variants; see 2.5.5.Examples: , , .

8 The group can hardly be of ie origin; it is not frequent. I noted , ,, ; , , . The group -- is the geminate of .Cf. below on , .

9 The group is certainly possible in piewords, but it is also frequent in Pre-Greek.Examples: -, , , , , , -, , , , ; .

10 The diphthong is perfectly ie, but it is found several times in Pre-Greek. I donot think that Pre-Greek had a diphthong -ou-, but it may have arisen frome.g. -arw-, which often surfaces as --. Examples: , , ,, , , , .

phonology 11

11 The group can hardly be of pie origin, but it is rare in Pre-Greek words, too.Like in the case of , it is the geminate of . Examples: (?); ().

12 On a morpheme boundary, the group is possible in pie. Examples from Pre-Greek: , , .

13 A rare group, perhaps there is even no reason to speak of a group. Examples:, .

14 (variants , ): Examples: (--), (-), . See 3.2.3.104.

15 A occurs bothword-initially and between vowels, where it has disappeared inmost inheritedwords. Initial: , , , , ,, , . Intervocalic: , (-), (),, , , . After resonant: , , , -, (-).

16 The group is hardly known from inherited words ( is problematic).Examples: , ; . -- may continue Pre-Greek -sgw-: Myc.ti-qa-jomay stand for /thisgwaios/ .

17 Again, this group is hardly known from ie words. It may sometimes continue-tyg-, as in , (see 2.5.5). Examples: , , -, , .

18 , These groups are well known from ie, but mostly in word-initial position. Seesection 2.5.5. Examples: , , , .

12 chapter 2

19 Though the cluster contains nothing that could not be ie, it occurs more oftenin substrate words. Examples: , .

20 The group can hardly be of pie origin. In Pre-Greek, it is a variant of and (see 2.5.5). Sometimes, it is clearly the geminate of : / . Furtherexamples: , , .

21 The cluster is possible in inherited words. pg example: .

22 , Rather rare in ie; Furne (1972: 110) assumes that the nasal caused the aspira-tion. Examples: , -, .

23 Frisk gives some seventy lemmas with -. Many of these words are clearly Pre-Greek. Originally, I thought that all words with - were Pre-Greek, but thisthesis cannot be maintained. flea may be of ie origin, though ie did nothave *ps-. The ie word has *plus-: Indian plui-, Armenian lu; Balto-Slavic has*blus- (with problematic b-); for Albanian plesht Demiraj (1997) reconstructs*plous-ti. The Greek word underwentmetathesis. For I refer to the etymol-ogy in edg.

24 Of course, is perfectly ie, but it also occurs in Pre-Greek words. Examples:, , , , , , , -, , , .

25 Geminates(See also 2.5.8 on single / geminated consonants.) Indo-European had no gem-inates. Geminates did arise in Greek, but they are not very frequent. I doubtwhether Pre-Greek had geminates, but several occur in Pre-Greek words(Brixhe 1976: 95 states that there were no geminates in this language). As Pre-Greek had palatalized phonemes, I suspect that ly was (often) represented by in Greek. In a similar vein, perhaps ny might be represented as , and ry as, but this requires further investigation. For and see 2.5.5.9a. Unclearare , , , and (palatalized my is a rare sound). Some further exam-ples:

phonology 13

Stops:5 : , : , , (?): , : , , , , , ; .

Liquids: : , , , , , , : : , , , ; : , , ,

Sibilants : (), , .

2b How to RecognizeWords as Pre-Greek?

This appears to be relatively easy. A first indication is that a given word hasno Indo-European etymology. Often, there is variation which is impossible toexplain in Indo-European terms. Therefore, the discussion of these variants isessential. Then, there are numerous suffixes that are typical for Pre-Greek (seethe list below, 3.2.3). The meaning may also provide an indication. The wordsconcerned are often names of plants or animals, of equipment and utensils, orpart of viniculture (see chapter 6).

If we have some of the above features, it is quite clear that we are dealingwith a Pre-Greek word. The origin of the word is then indicated pg in edg.Inmany cases, we do not have enough data and can only suspect that the wordmight be Pre-Greek (the origin is then indicated as pg?).

3 Prothetic Vowel

Pre-Greek had a prothetic vowel, e.g. / . In most cases, thevowel is -. The numbers (Furne 1972: 368ff.) are as follows: 90, 10, 5, 3, o, 6, 2. Note that, generally speaking, may interchange with , , and. Indeed, we have cases where prothetic interchanges with , and the sameholds for (e.g. - / -, / ). Although not all other cases can beexplained away, it seems that the phenomenon originally only concerned .Examples: / ; / ; / ; /; / ; / (); / .Furne (1972: 378, 339) thinks that in the proth. vowel was lengthened.

5 We also have to recall the occurrences of , , (see above).

14 chapter 2

4 s-mobile

A large number of words show an initial - before a consonant, which is absentin practically identical variants. It occurs before a stop or m (so not beforer, l, n); the stop is mostly voiceless, sometimes aspirated; see Furne (1972:390f.). Examples: / ; (); / ; (); / -; / -; / -; () / ; /; (); (). A prothetic vowel may appear before ans-mobile (Furne 1972: 3908): / / ; / / ; / .

5 Consonant Variation

5.1 Voiceless / Voiced / Aspirated StopFurnes conclusion was that Pre-Greek was a non-Indo-European language,with no recognizable cognates. This implies that the phonemic system mayhave been different from that of Indo-European. Thus, he found that the stopsshow variation between voiced, voiceless and aspirated, so that there presum-ably was no phonemic distinction between voice and aspiration in the lan-guage. As there is no reason to assume that this is a recent phenomenon, itstrongly suggests that the languagewas non-Indo-European. For example, - belongs to a root ptk- / ptk- also seen in , -. Since such a variationis hardly understandable in Indo-European terms, thewordmust be Pre-Greek.Furnes discussion of this variation takes 86 pages (1972: 115200). Even if weallow for somemistakes, it is clear that there is abundant evidence for this phe-nomenon.

5.2 PrenasalizationBefore a stop, a nasal may be present or not in Pre-Greek words. Examples: / ; / ; / , etc. The phenomenonis extremely frequent, but its precise origin is not known (prenasalized conso-nants?).

5.3 NasalizationA consonant is replaced by a homorganic nasal: / ; - / . Furne did not discuss this phenomenon. Cf. the followingsection.

phonology 15

5.4 Labial Stops / / uThere are three interchanges: labial stop / ; labial stop / ; / .

Labial stop / (Furne 1972: 203227). Examples: / ; /; / ; / ; / ; /; / ; / .

Furne points to assimilational and dissimilational tendencies, but imme-diately concludes that this would only work for einige wenige Beispiele, and[l]eider kommt man auf diesem Weg nicht weiter (1972: 207). This way ofinterpretation then is probably not the correct way. My impression is that thephenomenon is quite general, and we must rather assume nasalization (dis-cussed above). According to this assumption, a consonant can be replaced bya nasal of the series to which it belongs. Thus, in the example / -, we find a dental nasal. In the case of labial stops, we would expect a(bi)labial nasal, and that is . Furne (1972: 210218) finds the process notablyfor / . This is in contradiction with the fact that Pre-Greek had no (separatephoneme) . So we must accept that the nasalisation occurred with the labialstops in general (or even with labials in general, in view of the inclusion of ;see below). It is of course a strong confirmation of the theory that this variationcan be explained in the same way.

Labial stop / (Furne 1972: 228242). Examples: , / ; - / ; / ; / .

I haveno explanation, despite the remarks of Furne andKuiper (1956: 215 f.).Furne (1972: 229) assumes that the variation is expressive.

/ (Furne 1972: 242247). Adifficulty here is thatGreekdidnot preserve a inmost cases, so that we often just find zero, and the can only be reconstructed.This gives rise to a certain degree of uncertainty. The evidence includes 8 or9 words in -. Examples: / ; / ; /; / (also ).

16 chapter 2

5.5 Stops Interchanging with (), with Stop + / or with + StopThis kind of variation is quite complicated. I distinguished no less than 11 (oreven 16) different types. They may be represented as follows:

a. Labials b. Velars

1. C / Ct / /

2. C / Cs / ( / )

3. C / sC ( / ) /

4. Ct / Cs / /

5. Ct / sC /

6. Cs / sC ( / ) ( / )

7. Cs / ss /

8. sC / ss /

Dentals

9. t / ss /

10. t / st /

11. ss / st /

The analysis of these variants is not easy, and I mainly present the data here.A question that needs to be explained is why exactly s or t are involved in thegiven variation.

The most complicated instance is 5b, where we find / . This type alsoyields most information, and can be solved best. Expected is a cluster with k,i.e. a consonant before or after the k. One of the two expected clusters musthave undergone metathesis. As Greek did undergo a metathesis > (andno metathesis of or ), we may assume that precisely this phenomenon wasoperative here. Thus, for an earlier stage we may reconstruct an interchange / . This interchange can be easily explained by assuming a consonant,probably unknown to Greek, which resulted either in or in . In my inter-pretation, this must have been a palatalized dental, i.e. /ty/. For instance, - / was probably *amutygala, represented first as *amusgala or*amudgala, the latter yielding *amugdala. A less clear example is Asclepios,who was called () or (). It could be that the name was*Atyklap-, giving *A(i)sklap- or *A(i)dglap-. In the latter form, metathesis didnot operate because **Agdlap- was not tolerated in Greek; the dental was then

phonology 17

simply lost. Needless to say, it often happens that only one variant is attested.The strange feature or phoneme may also be dismissed altogether, as in besides and .

One might suppose that all variants in the above group are due to a palatal-ized dental, but this is not evident, as consonant clusters are rather rare, and asthere are very few suffixes beginning with an obstruent. We may be unable todetermine what exactly happened in each case.

Type 4 is treated by Furne (1972: 2633). Since Pre-Greek did not distinguishvoice and aspiration in stops, these often vary; so if we speak of kt or , thisalso includes realization as , such as in . If we consider the variationwith labials, as in pt / ps, it is clear that we are dealing with a labial followedby a dental. The dental could also appear as s, so it is clear that the phonemeconcerned was a palatalized dental, which I note /ty/. This means that we aredealing with a group pty. In the same way, with a velar we have kty.

The example / is well-known and clear. Furne further gives (H.) besides (H.) and compares with Dor. . His example cooked besides is less evi-dent.

Among the forms with a velar, there is no problem with / .The best known example is (also ) besides on Atticvases. I have no opinion on ; it may be a Graecisized form, and, ifso, it is unimportant for Pre-Greek. See further the ethnonyms -,-, -, - and -. Other forms are less clear.

Theremay have been series with three forms, with kt / ks, pt / ps and also k orp. I can only mention / besides , and perhaps, parallelto / , the verb (togetherwith -); for both cf. Furne (1972:263).

Above, we assumed that a labial or a velar could be followed by a palatalizeddental /ty/. If this is right, we can also postulate that this consonant (labialor velar) was followed by a normal dental. This logically yielded pt and kt. Iassume that the second consonant of this group (the dental) could have beendropped, which yielded single p or k. This explains the type () (Furne1972: 50) and (with prenasalization) besides (Furne 1972:51).

I will briefly review the 11 (16) types (I call the labials 1a, etc., the velars 1b,etc.).

1. a. may represent a single phoneme py, as we saw in 2.1. Examples(Furne 1972: 315 ff.): - / - (-); / ; / ; / ; without variants: , .

18 chapter 2

1. b. is most probably explained like 5b, discussed above. Examples(Furne 1972: 319ff.): / ; / ; / -; / .

2. a. / , b. / . Furne has a whole chapter about the variation / (1972: 323329). I havedoubts about this chapter, although some instancesremaindifficult to explainotherwise. In these casesmay result from*pty,or have amorphological background. There is only little evidence for / .Examples: / (Furne 1972: 326); / (Furne 1972:327); / .

3. a. / , b. / : Both may represent *typ, *tyk. Examples: / (Furne 1972: 2922); / ; / ; / (); / ; / (Furne 1972: 295ff.).

4. a. / , b. / were discussed above and may continue *pty, *kty; theymay belong to type 2. Examples: / (Furne 1972: 2633); / (Furne 1972: 318, 324); / (Furne1972: 2633).

5. b. / was discussed above. Examples: / (Furne1972: 3012); / () (Furne 1972: 279, 319).

6. a. / , b. / . Furne (1972: 393) simply considered the interchange tobe due to metathesis, which, of course, is possible. *sp, *skmay represent*typ, *tyk. Examples (Furne 1972: 393): / ; / ; / ; / .

7. b. / . If represents *kty, the k may have disappeared in other cases(which did not give ) after which *ty became . Examples: / (Furne 1972: 13059); / (Furne 1972: 317); , / (Furne 1972: 28672); / (, );for / see 9a.

8. b. / can be explained parallel to 7b: *tyk > or, with loss of the k,*ty > . Example (Furne 1972: 300): / .

9. b. / . This is thewell-known element that yielded / . The situationis different here, as we are able to discern a distribution among the Greekdialects, and attribute the different renderings of these loanwords todialectal developments. Still, the fact remains that a foreign element wasrendered in different ways, as with all other phenomena discussed here.Furne (1972: 253) has the heading , , / (), . I think this should bereformulated as (, ), () / (), , i.e. with its usual variants , ;or the geminated (with its expected variant , which is the Greek formof geminated ), interchanging with or . If the was [sd], it does notfit in well. As to the elements interpretation, it could represent single *ty,which was rendered or , or single , (the variant would then fit in,

phonology 19

but one would also expect a variant 6). Examples (Furne 1972: 253ff.): / ; / ; / ; / ; / ; / ; / .I think that the phoneme rendered by , Att. (called the foreignphoneme or Fremdphonem) was a palatalized velar, which I write as ky,cf. Beekes (2009: 191197). This would be parallel to the development ofinherited velar + yod, which gave , Att. , as in , . Thisinterpretation is confirmed by , , where we have a variant (H.). Here we see that after the nasal (prenasalizationis well known in Pre-Greek), the palatal feature of the consonant wasdropped. This resulted in a velar (here realized as an aspirate). The variantshows that we may be dealing with a velar in cases of / . We canalso compare / , which had py; again we see thatthe palatal feature was lost after the inserted nasal.Theremay be a third representation.We know that the name of Odysseuswas -, -. This means that it probably had a palatalizedvelar, *ky. Butwe also find (Ibyc.apudDiom.Gr. p. 321 K,Hdn.Gr.,Plut.), a form which was at the basis of Latin Ulixes. This form was takenfrom a Western Greek dialect, probably Doric. A third representation ofthe foreign phoneme may therefore be --, although this may also reflect*kty, with / resulting from a simplification of this cluster (see 7b).

10. a. / may be from *tyt (cf. type 3). Examples (Furne 1972: 301 ff.): / ; / ; / ; /.

11. a. / may also represent *tyt giving or, with loss of the t, *ty > (cf.type 8b). Examples: / / .

As we saw, it is very difficult to determine what exactly happened in each case;on the other hand, it is clear that almost all variation can be understood if westart from just a few assumptions regarding Pre-Greek.

5.6 Velar / Labial / Dental Stops; LabiovelarsThere is limited evidence for variation between velar and labial, between velarand dental, and between labial and dental, and between all three classes(Furne 1972: 388ff.). We find:

6 [sn]: The absence of would be expected if the variants (), entered Greek as more

accurate renderings of pg *ty, with the usual variation in voice and aspiration, which later

developed parallel to inherited dental stop + yod (cf. Beekes 2009: 193).

20 chapter 2

/ , / , / / / / / / / / / / / /

It is remarkable that the variants mostly agree in voice / aspiration. Sinceexamples of this phenomenon are not particularly numerous, this may be anindication that the words concerned are not of Pre-Greek origin, but perhapsdue to borrowing from a different substrate. Examples:

/ : / ; / / : / / : / / : / / : / / : / / : / / : / / / : / / / / : / / .

It is tempting to assume labiovelars to explain these cases, but some cases mayhave a different origin (thus, / could be due to dissimilationin the first variant). On the existence of labiovelars in Pre-Greek, see above onthe phonemic system.

5.7 Dentals / LiquidsThere are some instances of variationbetweendentals (includingn) and liquids(l, r). This variation is incidental. Examples (Furne 1972: 387f.):

a. / : / (Furne 1972: 33027); / ; / -. Cf. Myc. gen. da-pu2-ri-to-jo (/daphurinthoio/) / ; /Myc. ka-da-mi-ta. The interchange / and the fact that Linear B has signsfor da, de, di, etc. (which Lejeune explained by assuming a specific, unusualsound ) might point to a dental fricative. / : / / : /

b. / : / / : /

phonology 21

c. / : / ; / ; / .

5.8 Simple / GeminateExcept for a few isolated cases, we find this interchange in / , but morenotably in / . On / and / see 2.5.5.9a above. Cf. Furne (1972: 386f.).Examples:

/ : (also ) / (also ); / . In this context, notethe suffix --.

/ : (); / ; / dat.pl.; / (this probably derives from pg *-alya-).

Note (), / , and the case of / / .

5.9 / ZeroWe discussed / zero before consonant under s-mobile above, section 2.4.

An s from Pre-Greek is normallymaintained. The only instances that I knowof where it may have disappeared, are (cf. Furne 1972: 241): , / (also -, -, -); / ; / Cypr. ; / . Perhaps / belongs here, too. Another instance couldbe , which is cognate with Lat. pirum, which points to *-pis-.

5.10 Velar or Dental Stop / ZeroThere are instances where a velar or a dental stop may be absent in initialposition (Furne 1972: 391 and 13159). Dentals may also be absent in inlaut.Examples:

/ zero: / ; / ; / ; /. / zero: / , but this form may be a late development. As anexplanation, one could think of a uvular q. / zero: / ; / ; / (with in lsj); / zero: / (also -).

Loss of a dental in inlaut: / ; / ; / .

22 chapter 2

5.11 -, - / Zero- and - can also be absent (Furne 1972: 391 f.): / (also -); - / ; / ; / . Perhaps it originally con-cerned palatalized ny-, ly-, varying with y-.

5.12 Metathesis, Shift of AspirationThere are instances of metathesis. It mostly concerns , sometimes . The con-sonant jumps to the other side of the vowel or the consonant: / ,; / . Cf. / ; / ; / -; / . In most cases, it cannot be determined what the originalconfiguration was. In a case like / , where may stand for (or con-tinue) , I would think that the was anticipated. It may concern an original rw.

The cases of / and / are discussed in 2.5.5 above.Shift of aspiration is found in some cases: / ; / -

. In the case of / the metathesis seems to have occurred in thelater history of Greek (Beekes 2003b).

5.13 Secondary Phonetic Developments1. We may assume secondary phonetic developments, either in Greek or per-haps already in the original language. One might consider:

- > -: / . For this case, cf. 2.5.7a / . > : / (Furne 1972: 308)- > -: / - > -: / - > -: / - > -: - / ? See 2.5.5.6a above.- > -: / ; / ; cf. , .

2. > before in the following syllable. The a was probably pronounced alittle higher before the u, and was realized as [], which resulted in . Exam-ples: > , > , *- () > , for *()-.

5.14 Other VariationThere are a few instances of isolated and puzzling variation. Imention just one,the word for night, where we have , , , . I think that insome of these cases, the solutionmay be found in a cluster. Carian, for example,allows an initial cluster kbd-. Such clusterswouldhavebeen simplified inGreek.In an inherited word, we have the parallel of Lat. pecten, Gr. , which is

phonology 23

supposed to continue *pkt-. If we assume a cluster *kdn- in our example, itmay have been reduced to kn- or, with loss of the first consonant, to dn-. Thus,the process is the same as the reduction - > -, see 2.5.13 above. Such variantsimplifications are typical for loanwords. In this way, we could connect two ofthe words; but I see no way to connect the other two.

6 Vowel Variation

6.1 Single Vowels (Timbre)The vowels show many variants. I will discuss them in the order a, e, o; withineach of these first the short vowel, then the diphthongs, then the long vowel(and the long diphthongs, but these hardly occur) are treated. Note that avariation x / y is not repeated under y.

1. the vowel .1a. / has 80 occurrences in Furnes material (1972: 347). Examples:

/ ; / ; / ; / ; / ; / ; / ; / .

1b. / . This interchange also occurs frequently. Furne (1972: 339) men-tions that he found 80 instances. Examples: / ; /; / ; / ; / ; /; / .

1c. / (Furne 1972: 336ff.). Examples: / ; /; / . The here is due to the following palatalizedconsonant.

1d. / (Furne 1972: 30237). Examples: / ; /; / . In the last example, the is probably due to thefollowing labialized phoneme lw.

1e. / : / .1f. / (Furne 1972: 3524, 3392). Examples: / ; /

; / . Both and are due to the following palatalizedconsonant.

1g. / (Furne 1972: 3535). Examples: / ; /; / .

1h. /, (Furne 1972: 30132). Examples: () / ; / ; / ; / ; / / .

1i. / (Furne 1972: 338). Examples: / ; / ; / .

1j. / . Example: () / ().

24 chapter 2

2. the vowel .2a. / : see under .2b. / (Furne 1972: 355ff.). Examples: / ; / ;

/ Myc. di-pa; / ; / ; / ; / ; / (). The ewas not phonologically distin-guished from i, and they were phonetically close.

2c. / / (Furne 1972: 35455). Example: / () /.

2d. / (Furne 1972: 115). Example: / .2e. / : see .2f. / (Furne 1972: 3392). Examples: () / ; /

.2g. / : see / .2h. / : see .2i. / (Furne 1972: 35842). Examples: / ; / -

; / ; / (); / .2j. / (Furne 1972: 171114). Examples: / ; / ;

/ .

3. the vowel .3a. / : see .3b. / (Furne 1972: 19137). Examples: / ; / ;

/ .3c. / (Furne 1972: 358ff.). Examples: / ; / -

; / -; / ; / ; /; / . and were phonetically very close, andnot distinguished phonologically (cf. on / ).

3d. / (Furne 1972: 359). Examples: / ; / (also --, --).

3e. / (Furne 1972: 279). Examples: / ; / -; / (also -); / , -; / .

3f. / (Furne 1972: 127). Example: / ().3g. / (Furne 1972: 358). Examples: / (); /

?3h. / (Furne 1972: 12029). Examples: / ; / .3i. / (Furne 1972: 133). Examples: / ; /

(Furne 1972: 148).3j. / . Example: / .3k. / (Furne 1972: 30235). Examples: / ; / ;

/ .

phonology 25

3l. / . Example: / .

4. / . There is some variation between and , but I do not know how tointerpret it. Perhaps it is due to some assimilatory and/or dissimilatoryprocess. Examples (Furne 1972: 364ff.): / ; /; / ; / ; / ; /; / ; / .

5. / . Example: / .

The behavior of the diphthongs may be summarized as follows:

/ and (vice versa) / / , / / , / , ,

All this variation is understandable in terms of adaptation from a three-vowelsystem.

6.2 Long / ShortOne may doubt whether Pre-Greek had a distinction between long and shortvowels (see 2.1). We do find and , but not very often, and the second ofthese has several variants. On the other hand, the variations / and / are not very frequent (although in this case the difference in timbre mayalso have been important, depending on the Greek dialect). Variation betweenlong and short and is frequent, especially in suffixes. Examples: /; / ; / ; / ; / ; /; / . Cf. / (cf. ); /; (-) / (-); / .

There is some evidence for short vowel + CC alternating with long vowel +C: e.g. / ; / .

6.3 Single Vowel / DiphthongThere are several instances where a diphthong alternates with a single vowel.They can be found above (2.6.1). Most frequent is / , but this is due to theeffect of a following palatalized consonant. We further find / , / , / and / . In two cases we find a diphthong alternating with a long vowel: /, / . Examples were given above.

26 chapter 2

6.4 Rising Diphthongs?Relatively frequent in Pre-Greek words are sequences of a more closed vowelfollowed by a more open one, sequences that are not found in ie. They wouldbe rising diphthongs if they formed one syllable, but in fact we may have to dowith two syllables. Examples are:

--: (-, -)--: ; ; ; ; ; . Note (-, -)--: ; (-); ; ; ; --: (); (-)

Remarkable, too, are the sequences -- in , and -- in (), .

6.5 Secondary Vowels (or Elision)Sometimes, words show a vowel that is absent in nearly identical forms. Itmostly concerns vowels between a stop and a resonant. It is often not clearwhether the presence or the absence of a vowel is secondary. See Furne(1972: 378385). Examples: / ; for *- in -; / ; / (); / ; / ; / ; / .

koninklijke brill nv, leiden, 2014 | doi: 10.1163/9789004279445_004

chapter 3

Morphology

1 Reduplication

Some forms seem to have reduplication, though we often cannot demonstratethis. Most frequent is partial reduplication, where only the first consonant anda vowel are repeated. The vowel is mostly or .

Examples: ; (); ; ; ; / (cf. ); (also -); ; (); (also -); -(?); perhaps ; (also -, -, -); . Also the names; ; ; . With prenasalization we find -, (cf. ; ).

Other reduplication vowels are found in: ; ; ; - (cf. ); ; perhaps also .

Intensive reduplication in: ; ; .More difficult to judge are / (perhaps from *--, --, but cf.

s.vv. in 6.3.2) and / (if from *--, --). Also besides (cf. ); = , also .

A completely different type is perhaps found in (cf. ), andperhaps also .

2 Suffixes

2.1 IntroductionIt appears that most suffixes have the same structure. They contain a conso-nant; if this is a stop, it can be prenasalized, i.e. -- or --, -- or --, etc. Thestop has its usual variants, like / / , etc., although mostly one of these ispredominant. The suffix usually starts with one of the vowels of the language,mostly , , (only rarely do we find or , e.g. / ). Thus, we mayfind e.g. ; , etc.

A different structure is present in suffixes containing -- (mostly followed bya vowel) directly after the root-final consonant, e.g. ; ; ;; . The groups --, --, --, -- in Pre-Greek words probablyoriginated in thisway. In the case of --, we often find a vowel again: --, --,--. The groups -- and -- are especially frequent. They are very important,as they are found in Etruscan, which for the rest shows little agreement with

28 chapter 3

Pre-Greek; -- is found as far as in Cappadocian (see Beekes 2003a: 51). Perhapsthe groups --, --, -- arose in this way, too.

Other consonants are found in suffix-initial position, too, e.g. --, --, --,rarely --. Examples: ; ; ; ; ; ;.

It is often possible to determine to which series the Pre-Greek consonantbelonged. Thus, -- could render -any-, while -aly- seems to have resulted in-- (or -- with coloring of the vowel). Likewise, -- could represent -ary-.This thesis would be nicely supported by the segment --, if this represents-arw- (e.g. / , if this form had *-arw-). Cf. 2.1 above.

Another type of suffix has followed by a dental (-); -or another stop; ; . These forms mayhave been partly adapted to Greek suffixes (-). See below on the suffix --.

A form such as -- is deviating; we do not often find a diphthong beforethe consonant. Does it stand for *-aut- from *-atw-? Cf. -aiu- in , wherewe may suspect ayw or awy (but it may be part of the root). See further section2.1.

Not seldomdowe find an alternation of a long and a short vowelwith a suffix(= consonant), e.g. , . In the case of , onemight again think of ury> uir, although ry is a rare phoneme (likemy).

2.2 Survey of the SuffixesIn principle, we find one of the three vowels of the language followed by a(prenasalized) consonant: a, i, u + (m)P, (n)T, (n)K. The nasal groups actuallyfound are (forms in brackets are rare or less frequent):

1. 2. 3. 4. 5. 6. 7. 8. 9.

aNC ()

iNC

uNC () ()

Themissing groups, then, are: 1. VN and 3. VN; 7. VN and 9. VN (except for).

In the same way, we find vowel + C. The consonant may have the normal varia-tion: plain, voiced, aspirated. A palatalized consonant could color a precedingand/or a following /a/ to [], which may also appear as . This phenomenon is

morphology 29

often seen in languages with palatalized consonants, such as Russian and Irish.Thus, we find -ary- represented as -- (-- is also possible). A palatalized -ly-may be rendered as a geminate --.

If a labialized consonant followed or preceded an , this vowel may havebeen perceived as (an allophone of) /o/. For example, -arw-may be representedas --, with anticipation of the labial element, but also as --, in which casethe was colored.

The suffixal consonant may be geminated; as there is frequent variationbetween single and geminated consonants in the language, there possibly wasno opposition.

Vowels could be either short or long; in suffixes, a long vowel was quitefrequent. A long was sometimes represented as .

2.3 The SuffixesThe examples are mostly taken from Furne, to whom I refer for details. Wordscan also be checked in edg. Variants are given in brackets. I added geographicalnames (tn) from Fick (1905), and some more material, with references.

1 --(-)(Furne 1972: 107): ; (); ; /; ;; (); ; . tn (Rhodes, Fick 1905: 47); (Caria).

2 --- (cf. Chantraine 1933: 397ff.); ; .

3 ---.

4 --.

5 --; ; -; ().

6 ---; ; ; ; ; ; , tn (Epidauros).

30 chapter 3

7 --/-()- before a VowelThere are words in - / -(), such as / (also ) / (note the hesitation in the accentuation). I suggest that the suffix was *-ay-(a),which was pronounced as [-ya] or [-eya] (we saw that often varies with). The speakers of Greek identified the suffix with Gr. -- or --, but the -y-could also be lost. In this way the three variant forms can be explained. Furtherexamples are / (); ( in H. is probably anerror); / (note the short ), besides / (these are notentirely clear to me, but cf. / ).

Furthermore, *-ay-a is likely to be the same suffix as - which makes femi-nine names, e.g. ; ; (note that inMyc. I-pe-me-de-ja, the -j- is preserved, cf. Ruijgh 1957: 1553). Of course, many place names endin -: ; ; ; ; ; , etc.

The final was often adapted to - after the dominant type, which is derivedfrom the adjectives in - (seeChantraine 1933: 91): type; cf. ;; .

We also find - used in nouns: ; ; .Nouns with -- are very rare; we find: ; ; ; ; -

(?); . It may further be found in < *-kay-an- (note the by-forms , -).

Parallel to -, -, we may expect thematic ---; we find it for examplein ; ; ; ; * (reconstructed by Furne1972: 169).

8 -()--(Furne 1972: 23322, 25532): partly from --; it is often impossible to establishwhether a form had a -- or not. See also 7 above. Examples: ; ;; ; ; (Myc. e-ra/ra3-wo); ; ; ;. tn (Fick 1905: 58).

9 ---tn (Thess.).

10 --tn (Thess.); (Arc. deme); (Thess. deme); (Arc. deme).

11 --(Furne 1972: 171117): , -; ; ; (also --);; ; ; .

morphology 31

12 --(-)represents -ary-: (also -, -, -); .

13 --(Furne 1972: 15864): ; ; ; / / ; (); (also -, -); ; ; ; , . tn, - (Lac.). Cf. --: , .

14 -.

15 -()--(Furne 1972: 25428, Beekes 2008): ; (); () (also-); . tn (Phoc. source); ; (Arc.).

16 ---(Furne 1972: 184): ; ; ; ; .

17 ---; ; . tn (Crete).

18 ---. tn () (Cos); ; (Lydia); (Kydon.); (Kydon.).

19 ---tn (hn Crete, Elis); (hn Thess.); (hn); (Fick 1905: 18).

20 --.

21 --.

22 --. tn (Fick 1905: 51); () (Pamph.);() (Fick1905: 53); .

32 chapter 3

23 -/-(Furne 1972: 19135, 21671; -- unless otherwise stated): ; (-); (but -); ; /; (Fick 1905: 69, etc.);/-; /-; ; . tn (Chios).

24 ---.

25 --(cf. --, --): ; ; (also --); tn (Crete).

26 ---(Furne 1972: 23531): ; ; . tn (Crete, Fick1905: 24).

27 -(Furne 1972: 13475), mostly neuters: ; ; ; ; ; adj.; anim. ; (gen. -; cf. Myc. da-ma / du-ma).

28 --(Furne 1972: 25736): ; ; ; ; ; -; ; . Also ? tn (Crete, Lycia); (Lycia); (Fick 1905: 75); (Crete).

29 --/-(Furne 1972: 15757): ; (); . tn (Crete); (Mess.); (Thess.).

30 ---tn (Crete); (Crete); ; (Fick 1905: 32).

31 --; . tn (Crete); (Crete, Fick 1905: 27).

32 --/-(this may continue -arw-): (); ; (); ; ; (-); ; (). tn .

33 --.

morphology 33

34 --; ; .

35 --; .

36 --perhaps .

37 --(cf. on --): ; .

38 --(cf. on -, -): ; ; .

39 --tn ; ; perhaps in .

40 --see below on --.

41 ---; ; (); . The suffix ---may continue-ery-, -ary-.

42 --/-; ; ; ; (); ; ; perhaps. Cf. the next.

43 --/-; ; ; ; . Cf. --/- above.

44 ---; ; () (if not ie); . Cf. Furne (1972: 15142).

45 --(-)-; . tn (Fick 1905: 95). Cf. Furne (1972: 15144).

34 chapter 3

46 --. Cf. . Cf. Lat. (from Etruscan) (doss-)ennus; Porsenna. It is con-ceivable that ny gave .

47 --.

48 --/-; (also -); . tn (Crete).

49 ---; ; ; ; ? tn (Att.); . Cf.Furne (1972: 1154).

50 -- as in nom. -; several pns like (); .

51 -- (-). See --/-.

52 --; . Cf. Furne (1972: 173, 1817).

53 --/-. tn ; ; (all in Lydia).

54 --(-)tn ; (Fick 1905: 67); (Pagas.); , -- (Att.).Cf. --.

55 --, --; ; ; ; ; ; / . Cf. Furne (1972:199, 24570).

56 ---; ; ; ; ; ; ; ;. Cf. Furne (1972: 1155).

morphology 35

57 -; ; (); (); ; ; (-); ;; ; ; ; perhaps . tn (Crete, Fick 1905: 25); ; (Crete). Cf. Furne (1972: 172118).

58 --.

59 -; ?

60 --; ; ; ; (-); ; ; . tn (Crete); . Cf. Furne (1972: 20410).

61 -()-/-tn (m Paros); (Boeotian, Fick 1905: 80); (Att.); (Att.). Cf. ---.

62 --(-)(); ; ; ; (Myc. te-pa). Perhaps also ()(also -, -)? tn M (Fick 1905: 71). Cf. Furne (1972: 172118).

63 --See -()-/-.

64 ---tn (Euboea); (Thracia).

65 ---; ; ; . See Chantraine (1933: 368); cf. --, --, --.

66 --/-; ; hn . On - see Furne (1972: 30339): ();. Cf. on --.

67 --.

36 chapter 3

68 ---; -.

69 -//-; ; ; (also --, --); , -; .

70 --(cf. --, Furne 1972: 3247): ; ; ; ; .

71 --(probably a combination of two suffixes, cf. on --): (cf. , --);. Cf. on --.

72 --, --(cf. --): , --; ; ; ; ; .

73 --(cf. --, Furne 1972: 226102): ; ; (); (later ); -; ; -; .

74 --; .

75 --, --; ; ; ; ; () (--) = ; .tn (Cos).

76 --/-; ; ; ; .

77 --/-(Furne 1972: 24671): (also --); ; .

78 --/-; ; (); ; ; . tn (Lemnos); (Cyclades).

79 --(-); ; ; ; ; . tn -.

morphology 37

80 --(cf. -- and --, --): ; . tn (r); (Caria).

81 --(-)(cf. --): ; ; ; . tn (Euboea); (Fick 1905: 74).

82 --; ; .

83 ---tn .

84 --/- (-); . tn ; ; (- = -, Fick 1905: 25,61).

85 ---; ; ; (and variants).

86 --(cf. --): .

87 --/-(cf. --, --, Furne 1972: 163): ; (); . tn (Crete).

88 --(cf. --): ; ; ; ; .

89 --(probably a combinationof --with apreceding consonant; see on --): ;; ; .

90 --tn (Caria); .

91 --(Furne 1972: 13265), where a preceding velar may become aspirated: ;-; ; ; ; ; ; / ; ;. tn (Cyclades).

38 chapter 3

92 --(cf. --, --, --): (also --).

93 --(Furne 1972: 107), often there is a variant with --: (); , -(-); ; ; . tn (Thess.); (Corc.).

94 --(see also 3.3.2b): (-); .

95 --, --tn (Chios); (Lydia).

96 -- (-)?

97 --(may continue -arw-): ; ; (also -, -); ;; ; . tn (Arc., the oldest town of all; Fick1905: 93).

98 -()-(Furne 1972: 19755): (also -); () (also ); . tn; (m ); .

99 --(this may rather be a suffix -- after a root in -): ; .

100 --(this suffix probably consisted of one phoneme py): ; ();.

101 --(Furne 1972: 12437, 21562): ; ; ; (= Lyc. idkre?).See also the suffixes --, -- and --.

102 -- (also --, --).

morphology 39

103 --tn (Caria).

104 --(Furne 1972: 48126, 21562): (); ; (also -);. We also find variants without --: / ; / -; / ; / . Therefore, the cluster proba-bly arose by addition of the suffix --. Note that -rn- is found in Etruscan andalready in Cappadocian (Furne 1972: 48126). See also the suffix --. tn - (Crete); ; (At.).

105 -There are several words in -: ; ; (); (); perhaps.

106 -- (--, --).

107 --(Furne 1972: 25427; in several cases this does not seem to be a suffix, but ratherthe end of a root; cf. on --, --, --): ; (also --); ; ;. tn (Crete).

108 --; ; .

109 --(cf. --): ; - (cf. Myc. te-mi-ti-ja / ti-mi-ti-ja); ; -. tn ; .

110 --(cf. --): ; (also -); (also -); ; -(); .

111 --; ; .

112 --(see 2.5.5.9a on / ): ; ; .

40 chapter 3

113 --; ; (also --); ; (also --); (--); cf. .

114 --; ; - (-). Cf. --: -.

115 --, -; ; ; ; .

116 -tn (Cos).

117 --, --; .

118 --; ; . tn (Crete, also , Fick 1905: 18, 24).

119 --; ; ; -. tn (Locris).

120 --; , -; (); .

121 --(Furne 1972: 20514): ; ; ; ; (also -); (also -).

122 --(); (also -). tn (Mess.).

123 --.

124 --; ; -. tn (Crete).

morphology 41

125 --; .

126 --(cf. Furne 1972: 24366 on -umn- in Etruscan and Cappadocian): ;. tn / (Crete); (Locr.).

127 --(see also --): ; (cf. --); ; . tn (Crete).

128 --; . Cf. on --.

129 --(cf. -/-): . tn (Rhodes).

130 -/-; ; /. tn (Fick 1905: 88); .

131 --tn (Athos).

132 -- (older ); ; ; .

133 --; (also -); ; ; ; ; . tn (Crete); (Boeotia); (Cos).

134 --; (also -); ; ; ; .

135 --(on - see -): .

136 --; . tn (Crete); (Crete).

137 -.

42 chapter 3

138 --.

139 --(). tn (Lemnos).

140 --; ; ; (); .

141 --(on --- see Chantraine 1933: 263): (also --); (also -);; .

142 --tn (Att., Fick 1905: 70).

143 -- (?); ; . tn (Cyclades).

144 --.

145 --(Furne 1972: 30339): ; ; ; ; .

146 --(a variant is --): ; -; (). tn / -(Crete); (Epirus).

147 --(Furne 1972: 21150): ; ; ; ; . tn (Chalc., Fick1905: 22).

148 --(see -()-): tn (Euboea); (Caria, Fick 1905: 26).

149 --(Furne 1972: 28383, 384132): ; -; ; ; -. tn .

morphology 43

3 Word End

Word end provides an interesting situation, as some original finals of thePre-Greek language may have been preserved. Of course, in order to arrive atthe Pre-Greek form,Greek endingsmust be removed, notably -, -. Thus, -,-mayoften continueoriginal -, -: cf.Myc.du-ni-jo /du-ni. Thewords in -have replaced almost all of those in -- (as in -).

3.1 Words Ending in a Vowela. -. A short - can only come from *-ya < *-ih2 in inherited Greek words. In allother cases, we may be dealing with a Pre-Greek ending -a that was originallyshort. It is often difficult to seewhether - is short or long; thematerial requiresfurther study. Examples: ; ; ; (?); ; ;; ; ; ; ; ; ; ; ; ;; ; ; ; ; ; (also -); ; ; ;; , etc. Note forms in -, like , and in -. Note, further,; .

For words ending in -, see 3.2.3.105 above.

b. -. ie words (neuters) in - are very rare in Greek. Examples of Pre-Greekwords in -: ; ; ; (); . We may assume thatmany words ending in -, - originally ended in -, -. Final - is frequent,too.

c. -.; ; . For -, see the foregoing. Final - is also found severaltimes: ; ; ; ; ; -; ; ; ;.

d. -. Though the ending may also be inherited from ie, in many wordsit is clearly of Pre-Greek origin, e.g. (Myc. qa-si-re-u); (). Iwithdraw my considerations in FS Kortlandt (2008: 53f.) on this point.

e. -. ; ; ; ; . The suffix alsomakes femininenames in -:; . It is usually assumed that the original inflection of all words in -derives from stems in *-oi-; I assume that Pre-Greek words secondarily joinedthis inflection. Words in - are masculine: (); ; ; .

44 chapter 3

3.2 Words Ending in -a. -. ()?; ; ; ; ; ; ; .

b. -. Examples: ; ; ; (also --); .

c. -. Examples: (Dor.); .

d. -. Examples: ; ; ; .

3.3 Words with a Nom. in - or -a. - (stem in --) is found quite often:

-: ; ; ; ; ; ; ; . hasa stem in --.-: ; ; .-: ; .-: ; ; ; ; .-: ; .-: .-: ; ; ; .

Note acc. ; acc. .

b. -: ; ; ; ; ; ; ; ; .Monosyl-labic: .

3.4 Words in -; ; ; ; ; .

3.5 Words Ending in - (-stems); ; (?); ; ; ; (); ; ;; ; ; ; .

With a stem in --: (--); (--) etc.; see 3.2.3.23.With a stem in --: ; ; ; see 3.2.3.5.

koninklijke brill nv, leiden, 2014 | doi: 10.1163/9789004279445_005

chapter 4

The Unity of Pre-Greek

The material itself shows that we are largely dealing with one language, or agroup of closely related dialects or languages. Of course, we cannot demon-strate in each and every case that the words that are non-Greek belong to thissame language. The bulk of the known non-Greek words, however, seem to fitthe general picture of the Pre-Greek substrate. For example, / - does not only show the element / , well-known from geographicalnames, but also the suffix -- with prenasalization. The pair / - also shows the element / , but has a suffix added that isalso typical for this language. The word / (-) again has thesuffix / , but also prenasalization. / has both the typical(prenasalized) suffix -- and variation / . In / we have thes-mobile and the well-known suffix, while , - has the variant withoutprenasalization, and has a different Pre-Greek suffix. In () / () we have a combination of a prothetic vowel and prenasaliza-tion.

Other languages may well have existed in the area. Thus, it is not certainthat Hieroglyphic Minoan reproduces the same language as Linear A. Further,Eteocretanhasnot yet been connectedwithother elements and seems isolated.

Another matter is that (non-Indo-European) loanwords from old Europemay have entered Greece, cf. Beekes (2000). Moreover, these may have alreadybeen adopted in Pre-Greek, as is suggested by , which has a Pre-Greeksuffix, but a root which is attested (with some variation), as a substrate word,in other European languages. Cf. also / , which shows typicallyPre-Greek features, but may ultimately be a loan from Akkadian (Kroonen2012).

However, I think that it is methodologically more sound to start from theassumption that non-Greek words are Pre-Greek. Only when there is reason todo so should we assume that they have a different origin.

koninklijke brill nv, leiden, 2014 | doi: 10.1163/9789004279445_006

chapter 5

Pre-Greek is Non-Indo-European

Our knowledge of Indo-European has expanded so much, especially in thelast thirty years (notably because of the laryngeal theory) that in some caseswe can say almost with certainty that an Indo-European reconstruction isimpossible. A good example is the word . In order to explain the -a-of this word, we need to introduce a h2. However, a preform *gnh2dh- wouldhave given Gr. *-. One might think that assuming *h2e would remedythe problem, but *gnh2edh- would yield *-. The conclusion is that noIndo-European proto-form can be reconstructed, and that the word cannot beof Indo-European origin. Another example is the word overhangingbank, for which a connection with to hang (up) used to be evident.However, we now know that most long vowels go back to a short vowel plusa laryngeal, and that long vowels cannot be postulated at random. In thisparticular case, there are simply no conceivable formations that would containa long root vowel. Thismorphological objection is strengthened by the fact thatthere is no trace of the expected root-final -- < *-h2- (as in < *kremh2-).Positively, one can say that landscape terms are frequently borrowed from asubstrate language. The inevitable conclusion is that the word is Pre-Greek.

koninklijke brill nv, leiden, 2014 | doi: 10.1163/9789004279445_007

chapter 6

The Pre-Greek Lexicon

In my edg, I marked with pg all words which, in my view, were of Pre-Greekorigin. I found 1106 words. On the criteria see 2.2b. I here present the wholeevidence, with a short indication why the words are considered Pre-Greek. Fora full discussion of their etymology, I refer the reader to edg. The material ispresented in a semantic classification.

The largest categories are Flora (178) and Fauna (180). This does not reallycome as a surprise; together with the Landscape and Natural Phenomena (37)and the Minerals (27), they form the previously unknown natural elements ofthe new home where the Greeks arrived.

An important segment refers to Agriculture (38). Many terms in this sec-tion pertain to Viniculture (20), which is one of the few categories that seemsrather unexpected, but there is nothing against the assumption that the origi-nal inhabitants knew andwidely practiced the technique of wine-making, per-haps with more advanced methods than the newcomers. Many words refer toHuman Physiology (81). To everyday-life belong the terms related to Attire andJewellery (21) and the many words referring to Equipment and Utensils (154).Well represented is Construction (34). The category Culture (51) includes termsreferring to Musical Instruments and Performing Arts (18), Religious Festivalsand Feasting (8), and Divine and Numinous Beings, Priests and Temples (12).To these wemay add the Theonyms, Divine Epithets, Mythical Characters (26).

In the edg, I have introduced a special category pg?. This category com-prises words without a good etymology, for which I nevertheless could not findpositive indications for Pre-Greek. I have not included these forms here, butthey can be easily found in the edg. It may be worthwhile to investigate thesewords (more than 780) more closely, since the indications for Pre-Greek originmay turn up at some point.

1 Landscape and Natural Phenomena

[m.] torrent (Mosch.); also a river name in Thessaly (Hes. Sc. 477) andAcarnania. No doubt, the word is non-Greek, and probably non-ie; note thesuffix --- (see 3.2.3.32).

[f.] lightning (Il.). (Il.); (Hdt.); . lightning (Paphian), (cod. -) id. and -

48 chapter 6

(H.). This wordmust be Pre-Greek because of the vocalic inter-changes. See Beekes (1987).

[n.] cleft, abyss. Variants are and , the latter probablyshortened from the former, and Arc. (representing -; cf. =). The variations - / - and / point to Pre-Greek origin (see 2.5.6and 2.6.1.1a).

[m.] depth (of the sea) (A.). bottomless, further [m.]depth of the sea. The variation / points to Pre-Greek *-ty- (see 2.5.5.9a).

[f.] earth (Il.). Dor. . Probably related to id.. Both may go back toPre-Greek *gaya- (cf. 3.2.3.7 on the suffix -ay-a-). Another variant * (with / , see 2.5.6) may be found in (Dor. -) and , but inthese cases the meaning earth cannot be ascertained.

[m.] hill (Il.), acc. to em and Orion = elevated place. Thevariants and prove Pre-Greek origin (see 2.5.1 and 2.6.1.3i).

[m.] darkness (Simon.). The group - seems to point to Pre-Greekorigin (see 2.2a.6). Cf. below.

[f.] dew, often of several fluids; pl. also young animals. The word isprobably of Pre-Greek origin; note the intervocalic -- (see 2.2a.15).

[f.] lowlands, humid pasture (Il.), also [pl.] id.. The word isprobably Pre-Greek, given the variation -- / --, which could not occurin a participle. Cf. 3.2.3.17 on the suffix ---.

[m.] straits, narrows (X., Arist.); especially the straits between Euboeaand Boeotia (h. Ap. 222, Hdt.); later also canal in general (D. H.); ventilator,fan (Gal. 10, 649) is probably a homonym, derived from in the senseblow. The word may well be Pre-Greek, cf. Ruijgh (1967a: 172374). Note thatthe long in this position is typical for Pre-Greek forms, cf. 3.2.3 s.v. --, --,--, --, --.

[f.] sea (Il.). (H.). The word, with its prenasalizedvariant, is typically Pre-Greek (see 2.5.2). The variation -- / -- / -- pointsto Pre-Greek *-ky- (see 2.5.5.9a).

[f.] wood, wooded hill (Hdt., Theoc.). As a tn , wooded hill in westernMysia (Il.) and on Crete (D. P., Paus.). A Pre-Greek word without furtheretymology.

, - [f.] rainbow (Il.), also of the halo of the moon, etc. (Arist., Thphr.,Gal.), as a plant name purple Iris, etc. (Arist., Thphr.), see Strmberg (1940:49); also name of a stone (Plin.). As a pn, , -, - daughter of Thaumasand Elektra, messenger of the gods (Il., Hes.). Furne (1972: 356) compares (H.), and concludes to Pre-Greek origin, in viewof the variation / (see 2.6.1.2b).

[m.] pit or cavern at Sparta, into which people sentenced to death (or

the pre-greek lexicon 49

their bodies) were thrown (Th. 1, 134, Paus. 4, 18,4, D. Chr. 80, 9). Also ,-,with the variations / (2.5.1), / (2.6.1.1a) and the suffix -- (3.2.3.31).A pre-form *kawyat- would probably give */-, where the is from aafter a palatalized consonant. Cf. (directly below).

[adj.] epithet of (B 581, 1; verse-final), generallytaken as full of crevices, abysses, later said of the wooden horse (Q. S. 12,314) and, by confusion with , , said of , (Nonn.). Itseems evident to connect , crevice in Sparta. Furne (1972:1806) points to the gloss . (H.), which shows that acrevice could be called . For the variation / , cf. above.

[?] only in the question , which according to some grammar-ians stands for , according to others for (Ar. Fr. 656, Pherecr.165). Also , . The prenasalization clearly shows the Pre-Greekcharacter of the words (see 2.5.2).

[n.] evening twilight, dusk, morning twilight (Il., X.). Cf. (thissection, above). The word is no doubt Pre-Greek, but the variation is notknown from other examples. See 2.5.14.

, - [m.] summit, top, pinnacle, only metaph. (Pl., Com. Adesp.,Str.), acc. to H. also = green woodpecker (i.e. , see section 4.3below) and sea fish. Also a tn, town in Ionia. This Anatoliantoponym points to substrate origin. Also note the suffix -- (see 3.2.3.145).

[m.] ice (Il.), also rock-crystal [f.] (Str., D. S.), with gender after. (H.). As Kuiper (1956: 21516) remarked, thewordis Pre-Greek because of the suffix -- (see 3.2.3.15). See Beekes (2008).

[f.] island (Il.); also (flooded) land near a river, alluvial land (Tab. Her-acl., pap.). Dor. (Rhod. ). is probably an Aegean loan.Furne (1972: 387), who points to the variation between single and gem-inate (see 2.5.8), also assumes a Pre-Greek loan.

, - [f.] rock, rocky mountain range, cliff, ridge; rock cavern, cave (Il.),boulder, stone (Hell.). No etymology. The word is probably Pre-Greek; seeFurne (1972: 272 etc.). Note the suffix -- (see 3.2.3.101).

[?] fissure in the soil, crevice (H.). See (this section,below).

, - [f.], mostly [pl.] drop ( 536 = 501, Hes., Pi.); also dust parti-cle ( . 502), spot (Opp.). [v.] to besprinkle. Variants are to be besprinkled (H.); , is besprinkled, gets wet (H.); (H., Phot.). isclearly Pre-Greek, if only because of the suffix (see 3.2.3.69).

[m.] roar (of waves, of oars), metaphorically noise in general (Hes., A.,Opp.); path, trail (Nic.; Boeot. acc. to Plu. in Hes. 13). In view of the variation

50 chapter 6

/ attested in the gloss = (sch. Nic. Th. 194, H.) and - =-, we have to assume that is a Pre-Greek word (see 2.6.1.1b).

[m.] buzzing, whistling, hissing noise, of arrows, winds (S., Ar.). As the-- in the suffix is certainly not of Indo-European origin (see 2.2a.3), theword is probably Pre-Greek. Cf. also (below).

[m.] buzzing, rushing, humming, of arrows, wings, water, etc. (epic 361, Hell.). Similar to (above). If related to it, is certainly ofPre-Greek origin, but even if it is unrelated, such an origin may still beconsidered.

[m.] turbulent movement of the sea, flushing of the waves; anchorage,roads (as opposed to a protected harbor) (S., E., Lys., Hell.), metaphoricallyof an earthquake (E. it 46), turbulent emotion (lxx, Gal., Max. Tyr.). ,- [m.] large sieve of mineworkers, also an Att. name of a potter; metallic vessel or implement (H.); [v.] to shake, id.. Already the velar suffixes, and especially the variation theydisplay, prove Pre-Greek origin for this word: --, --, -- (with / and prenasalization, see 2.5.1 and 2.5.2); cf. (see section 11.3below).

[f.] hill (Str. 8, 3, 19; 10, 2, 17); also the island names , .Withouta doubt Pre-Greek (for V-, see 2.2a.15), as argued by Fick (1905: 54 and112). Likewise, but with a different interpretation (connected to ),Alessio (1943: 121 ff.).

[n.] cavern, cave (epic Il., also Cypr. inscr.). Cf. directly below.The word is no doubt Pre-Greek (Furne 1972: 123).

[n.] cave, cavern (Pl., lxx, nt et al.). Also , - [f.] id..Furne (1972: 123) reconstructs a Gr. form *, -, from Lat. *splca.He further adduces = fissure (H., Phot.) and refers to Etr. pel(a)cave, tomb. Theword is no doubt Pre-Greek: note the s-mobile (see 2.4), theprenasalization (see 2.5.2) and the suffix -- (see 3.2.3.114).

[f.] storm, squall (Plu.). Furne (1972: 373) connects torrent (Maced.) (H.); the word is Pre-Greek in view of theprothetic vowel (see 2.3).

, - [m.] peak of a rock, of a