Early Semitic: A Diachronical Inquiry Into the Relationship of Ethiopic to the Other So-called South-East Semitic Languages 900434814X, 9789004348141

163 94 17MB

English Pages 82 Year 1967

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Early Semitic: A Diachronical Inquiry Into the Relationship of Ethiopic to the Other So-called South-East Semitic Languages
 900434814X, 9789004348141

Table of contents :
EARLY SEMITIC: A DIACHRONICAL INQUIRY INTO THE RELATIONSHIP OF ETHIOPIC TO THE OTHER SO-CALLED SOUTH-EAST SEMITIC LANGUAGES......Page 3
TABLE OF CONTENTS......Page 5
Foreword......Page 7
Introduction......Page 9
1. Consonants......Page 14
1) Gutturals......Page 15
2) Palatals......Page 16
4) Sibilants......Page 17
9) Prepalatal fricatives and affricates......Page 18
2. Vowels......Page 19
1. The personal pronoun......Page 24
2. Demonstrative pronouns......Page 35
a. Primary stem or stems......Page 39
b. Secondary stems......Page 43
a. Verbal preformatives......Page 48
b. Verbal afformatives......Page 51
c. The formation of conjugations......Page 57
1. General remarks......Page 67
2. The declension of the noun......Page 68
3. Numerals......Page 77
Conclusion......Page 80

Citation preview

EARLY SEMITIC

STUDIES IN SEMITIC LANGUAGES AND LINGUISTICS EDITED BY

G. F. PIJPER Emeritus Professor in Arabic Language and Literature in the University of Amsterdam

i A. MURTONEN

EARLY SEMITIC

LEIDEN E. J. BRILL 1967

EARLY SEMITIC A DIACHRONICAL INQUIRY INTO THE RELATIONSHIP OF ETHIOPIC TO THE OTHER SO-CALLED SOUTH-EAST SEMITIC LANGUAGES

BY

A. MURTONEN

LEIDEN E. J. BRILL 1967

Copyright 1967 by E. J. Brill, Leiden, Netherlands All rights reserved. No part of this book may be reproduced or translated in any form, by print, photoprint, microfilm or any other means without written permission from the publisher PRINTED IN THE NETHERLANDS

TABLE OF CONTENTS Page

Foreword............................................................................................... ix Introduction......................................................................................

1

Part One: Phonology......................................................................

6

1. Consonants..............................................................................

6

1) Gutturals, 7 — 2) Palatals, 8 — 3) Denti-alveolars, 9 — 4) Sibilants, 9 — 5) Labials, 10 — 6) Nasals, 10 — 7) Liquids, 10 — 8) Semi-vowels, 10 — 9) Prepalatal fricatives and affricates, 10.

2. Vowels........................................................................................11 Part Two: Morphology.................................................................... 16

A. Pronouns 1. The personal pronoun................................................................16 2. Demonstrative pronouns............................................................27 B. Verbal System 1. General remarks........................................................................31 2. Formation of verbal stems........................................................31 a. Primary stem or stems, 31 — b. Secondary stems, 35.

3. Verbal conjugation....................................................................40 a. Verbal preformatives, 40 — b. Verbal afformatives, 43 — c. The formation of conjugations, 49.

C. Noun 1. General remarks........................................................................59 2. The declension of the noun.................................................... 60 3. Numerals....................................................................................69 Conclusion

72

FOREWORD This study is a further off-shoot of my investigations of the early history of Hebrew, even if the connection may seem rather remote; its relevance to the latter, however, will become evident in Hebrew in Its West Semitic Setting now under preparation. In a sense, it is a continuation of Broken Plurals^ completed soon after the publication of this, in 1964; some adverse circumstances delayed the publication thus far. Had I written the manuscript two years later, after the in­ vestigation of the structure of the Central Australian Pintupi language, many details would have been formulated in a different way, and new insights derived not only from that study, but also from the advanced preparations for the work on Hebrew mentioned above make me see the final conclusions in a somewhat different light also, even if this is already slightly foreshadowed in the closing sentences on p. 74. This means, in the first place, that I nowadays think of the early history of Semitic languages much less as a process of their derivation from a more or less uniform Proto-Semitic mother tongue than in the terms of areal units, whether they be regarded as dialects or languages, mutually influencing each other and in the earliest times changing structurally in rapid succession as a consequence of changing their nearest contacts, as this has been and still is taking place in and around the Great Western Desert of Australia. Therefore it is not necessary to visualize, e.g., the departure of South Ethiopic from North Ethiopic, or of Tigre from Gecez, as having taken place in so remote an antiquity as some statements in this work suggest; strange as it may sound to one accustomed to think in chronological terms, it now seems to me that parts of South Ethiopic, such as Gurage or some parts of it at least, have left the Proto-Semitic stage compara­ tively recently. Therefore it is also conceivable that ancestors of Asiatic Semites were on the Arabian peninsula and perhaps had reached the Fertile Crescent during a period which still can be termed ProtoSemitic; for me, this term now means a stage in the development of Semitic languages to which the principal features of their historically attested common structure can be traced back without need to suppose great internal differences within that entity nor remarkable structural transformations in the course of the differentiation of the historically attested earliest forms of the single languages. Apart from this

X

FOREWORD

structural modification of the chronological scale, however, what I have written still seems to me essentially valid, and therefore the original manuscript is now published unchanged apart from some minor corrections and additions. After all, it is not meant to be the final word on the subject, but to elicit discussion and constructive criticism, if possible, before I finalize the manuscript on the study of Hebrew referred to above. It now appears uncertain whether Gurage is to be regarded as a language or a group of languages; this question, however, does not affect our study. In the designation of its sub-units I have followed my sources, although the name forms now seem to be largely anti­ quated. For Aymallal, Leslau writes Aymellel, and another authority on Gurage, Mr. Robert Hetzron, informs me that the most usual native designation for the whole language is Soddo, which should accordingly be adopted. Olane is probably to be identified with Wolane (and/or Walani7)\ but the present identity of Urib—if any—I do not know. Because of these uncertainties and partial discrepancies I have retained the nomenclature of my sources in this treatise still. For the correction of my English I am indebted to Mrs. Beatrice Faust, M. A. (Hons., Melb.). My thanks are due to Firma E. J. Brill (Leiden), for accepting my manuscript in its series of publications and for the careful performance of the printing work. Melbourne, October 1966

A. Murtonen

INTRODUCTION As Wolf Leslau states in the beginning of his study of “SouthEast Semitic,,1), it is generally assumed that Ethiopic and South Arabic form a dialectal unity within the Semitic languages. How far this assumption is to be considered valid, however, depends largely on the definition of the term “dialectal unity”. If the term is used in the sense — as Prof. Leslau seems to use it — that these languages are descendants of one single dialect of the postulated Proto-Semitic language, which continued its existence as — let us say — Proto-South-Arabic after the dispersion of the Proto-Semitic linguistic community, the Ethiopic languages representing a later offshoot from this parent language — then, much to my regret be­ cause of Prof. Leslau’s well established and no less well earned po­ sition as the leading authority on these languages, I must question its justification. My divergence from the general opinion may be justified partly by the fact that my recent studies have shown that many characteristics, which Prof. Leslau regarded as typical of these languages only, are in fact shared by North-West Semitic as well (see below). Because of this, the present writer cannot admit any value for those features which, in addition to Ethiopic and South Arabic, are attested in East Semitic, as witnesses for SouthEast Semitic dialectal unity; considering that the North-West Se­ mitic dialect, in which those features are attested, is of a very con­ servative character *2), and that the East Semitic documents partly date from much earlier times, it is quite reasonable to assume that other North-West Semitic dialects in earlier periods may have pos­ sessed additional features now shared by East and South-East Semitic only. After these preliminary remarks, we may study Prof. Leslau’s list point by point. 1) Mahri does not possess labio-velar consonants in the proper sense of the word, and the fact that q seems to favour back vowels is shared by gutturals, as can be seen from L. ’s own examples. Moreover Hebrew shares that same characteristic; see, e.g., the examples in Gesenius-Kautzsch27 10A In principle, the present writer also doubts x) “South-East Semitic (Ethiopic and South Arabic)”, Journal of the American Oriental Society vol. 63, New Haven 1943, p. 4-14. 2) viz., Samaritan Hebrew, on which cf. now Studia Orientalia vol. XXIX, Helsinki 1964.

2

INTRODUCTION

whether this kind of feature can be used to support a theory of dia­ chronic relationship, because of its mainly phonetic character. 2) The weakening of labial and liquid consonants, being purely phonetic in character, can hardly be regarded as a diachronic charac­ teristic, either. Moreover, the fact that it is not shared by North Ethiopic nor by Ancient South Arabic (to its full extent, at least) suggests that it is rather recent a phenomenon, dating perhaps from the times when South Arabia and Ethiopia were politically more closely connected, spread from Ethiopia (where it may have originated under Cushitic influence, as L. also admits) by soldiers or merchants etc. 3) The assimilation of a short vowel to a long one in the following syllable is common in Semitic; the development in Gurage (and Soqotri) therefore has hardly anything to do with the South Arabic umlaut tet6ber\tetiber which, moreover, may be only another example of the different development of a long a in slightly different environ­ ments, examples of which can also be found in (Northern) Arabic x) and Hebrew *2). 4) The “connecting vowel” / between nouns in plural and pro­ nominal suffixes occurs also in Samaritan Hebrew 3); moreover, it is hard to see how it could be different from the same vowel occurring after certain plural forms in Arabic and Akkadian; like all the other similar “binding vowels”, it may go back to certain kinds of case — or comparable inflectional — vowels 4).* 5) The fact that the small minority of Ethiopic languages corres­ ponds to the decisive majority of South Arabic languages as regards the consonantism of the 3rd person of the personal pronoun ought to be a clear indication that the former need not represent just one offshoot of the latter, still less bear witness to such a relationship; moreover, for the Ancient South Arabic the consonant is identical in both genders and corresponds completely to Akkadian. 5a) As L. himself states, Akkadian has no article either; and in Samaritan Hebrew the article has become (or remained?) rather an optional thing 6). x) Cf. my Broken Plurals, Leiden 1964, p. 18, 24sqq., 30, etc. 2) See Studia Orientalia XXIX, § 2\d. 8) Studia Orientalia XXIX, § 58/; and cf. the paradigms. 4) Cf., e.g., C. Brockelmann, Grundriss der vergleichenden Grammatik der semitischen Sprachen I, Berlin 1908, § 247. 6) Studia Orientalia XXIX, § 159.

INTRODUCTION

3

6) The formation of plural by vocalic change is so common in the whole South Semitic group that singling out pluralic -o cor­ responding to singular -/ seems to me somewhat arbitrary, partic­ ularly since the nominal types in which it occurs in Tigrinya are totally different from those in the South Arabic dialects. An ety­ mological connection may have to be assumed in the case of Nisba only. 7) This is only a special instance of case 2 (on which cf. above). 8) This is again shared by only a few dialects in both Ethiopic and South Arabic and therefore bears witness — if any — only to partial connection between the two groups of languages (cf. case 2 above). 9) This characteristic bears closer resemblance to genuineness, but the unfortunate lack of data from Ancient South Arabic together with the fact that even in Akkadian the consonant for 1st pers. sg. is -k renders even this uncertain. 10) Even the type of preformative conjugation with a vowel after the first radical in the primary stem is shared, apart from Akkadian, even by the Samaritan dialectx) and Palestinian tradition *2) of Hebrew; moreover, in the former — at least — the second radical is in such cases only sometimes geminated, sometimes not. 11) In my opinion, the formation of subjunctive and/or jussive in Ethiopic and South Arabic can better be derived from imperative; etymologically it is easily recognizable that the former are based on the latter, only the preformatives having been added. This is the common method for the derivation of jussive throughout the Semitic family, accordingly nothing peculiar to South-East Semitic. Where subjunc­ tive is differentiated from jussive, this takes place regularly by means of small vocalic changes; only in the formation of the indicative mood more remarkable variants are attested (cf. case 10 above) 3). 12) The use of the mere stem instead of an inflected verb in Soqotri may rather be comparable to an analogous use of the noun of action (or noun of verb) in Hebrew 4). 13) The frequentative stems may indeed be confined to SouthEast Semitic, but it is to be observed, how rudimentary their appear­ 0 Studia Orientalia XXIX, § Wef. 2) See my Materials for a Non-Masoretic Hebrew Grammar vol. I, Helsinki 1958, p. 39 etc. 3) This is another indication of the secondary formation of the indicative, where it differs notably from the jussive mood. 4) cf., e.g., my Concise Grammar for Biblical Hebrew, Melbourne 1962, § 13^.

4

INTRODUCTION

ance is even in South Arabic, so that they are almost totally confined to Ethiopic. 14) Again, the appearance of the giving up of gemination in the intensive stem in Ethiopic is so limited that it is hard to see how it could represent a continuation of a common Proto-South-Arabic element. On the other hand, the release of gemination is so common general linguistic feature that it may be more feasible to assume that it has taken place in Ethiopic due to some phonetic reason; in South Arabic, the parallel of the indicative preformative conjugation of the primary stem may have been responsible for the beginning of the development. 15) L. himself admits that the causative s is common Semitic. 16) The so-called intensive and conative stems are primarily and still to-day largely used in Ethiopic as variant forms to the primary stem, a characteristic also shared by Samaritan Hebrew x); it is there­ fore only natural that causative and causative-reflexive stems are formed from them, and etymologically in no way peculiar to SouthEast Semitic. 17) The fact that I w/j and II w/y verbs in subjunctive and jussive still mostly appear without consonantal semi-vowel is, considering the fact that voluntative use of verbs probably is the oldest one *2), an indication of the secondary nature of this consonant even in the rest of the forms; the necessary analogy for its spreading is provided by the regular verb. 18) The inability of the gutturals to be geminated in Hebrew is well known; from Akkadian we also know that having weakened considerably, they no longer prefer an a, but rather an e vowel, which is also attested in a number of cases in Hebrew; other kinds of vocalic changes caused by the gutturals are too numerous and various to be enumerated here; all in all, little — if anything — remains which is peculiar to South-East Semitic. 19) According to M. Hofner 3), the //-preformative is attested in Ancient South Arabic in triradical stems too, even if infrequently, whereas in Aramaic the simple //-stem at least is lacking 4). As to the lexical connections, when it is question of two language groups spoken in so closely adjacent areas as South Arabia and Ethiopia, they can hardly be considered as proving anything. We may therefore !) 2) 3) 4)

cf. Studia Orientalia XXIX, §§ 10/, 11**//, Yld. cf., e.g., Studia Orientalia XXIX, § 10*-*. Altsudarabische Grammatik, Leipzig 1943, § 71, cf. § 73. Brockelmann, Grundriss I p. 536. Seeing that n is generally used elsewhere

INTRODUCTION

5

pass by them, particularly since the scope of the study is meant to be grammatical1), and conclude that Leslau has not proven a his­ torical connection between Ethiopic and South Arabic languages after the common postulated Proto-Semitic period 2); therefore we are entitled to take the problem again under consideration 3).* in the Semitic family as an initial augment to form secondary roots from earlier biradical ones (cf., e.g., W. von Soden, Grundriss der akkadischen Grammatik^ Roma 1952, § 100 a; and my Concise Grammar for Biblical Hebrew § 33/?) Brockelmann’s surmise that Syr. rPhef may be a remnant of it, is quite unnecessary, particularly as Hebrew and Judeo-Aram. have the root variant^/?. The suggestion (which I am unable to trace to its source now) that the Syr. 'Ettaf'al might contain the n element has not much substance, either, because I) the reflexive stems otherwise are formed by the simple /-preformative only, 2)5 is assimilated to a preceding con­ sonant elsewhere in Syriac also (cf., e.g., Brockelmann, Grammatik, Leipzig 1951, p. 17), and 3) the assimilation of n is also irregular in Syriac (see ib. p. 19sq.). Having written this, the present writer was provided with a convenient opportunity to find some orientation in the lexical respect also, in the form of Dr. W.W. Muller’s careful study on Die Wur^eln Mediae md Tertiae Y/w im Altsiidarabischen (Tubingen 1962). According to Dr. MOller’s etymological material, supplemented by me in a few cases (mainly on Hebrew), we find that ca. 80% of all the roots mentioned are also attested in (Northern) Arabic, ca. 39% in North-Western Semitic languages, ca. 23% in the Ethiopic languages, and ca. 14% in Akkadian. The low figure for Akkadian may be partly due to the lack of a complete dictionary of that language, as Dr. MOller (p. 5) points out, and even the other groups may not have been represented as completely as Arabic; but on the other hand, the fact that these other groups are represented by several languages should compensate for part of the lack of material from individual languages, so that the relative order of the relationship of Old South Arabic to the other sub-groups of the Semitic family may be the one given above, as far as the vocabulary is concerned, and the material from the Ethiopic languages cer­ tainly cannot be so much inferior to that from Arabic and particularly from NorthWestern Semitic that it would be enough to bridge the gap; and the sample, chosen according to a purely formal principle and therefore doubtless free from any bias, is many times more than enough in size (ca. 730 without root variants, according to my calculations) to be representative. Accordingly, there does not seem to be any sufficient lexical ground for grouping Ethiopic and South Arabic languages into one dialectal unity. 2) As far as a complete unity existed even at that time, cf. particularly Gurag6 and Tigre peculiarities to be dealt with below, some of which — particularly the lack of nominal flexion in Gurage — could go back as far as to the early neolithic period, when such a type of language may have been fairly universal; this question will be dealt with in the general linguistic introduction to my forth­ coming historical grammar of Hebrew. 3) Subsequently I see (in Linguistica Semitica, Roma 1961, p. 119, in an article by A. Spitaler) that S. Moscati has dealt with Leslau’s article (cf. also W. K. Brzuski in Roc^nik Orientalistyc^ny 25: 1 p. 128 sqq.) in Le^ioni di Linguistica Semitica § 43 and Rivista degli Studi Orientali 34 p. 33 sqq. (both of which are un­ fortunately still unaccessible to me) coming in all essential to results similar to those above. Cf. also D. Cohen in Janua Linguarum Ser. Mai XII, The Hague 1964, p. 490 sqq., who finds Soqotri as closely connected with modern Arabic as with Ethiopic languages. Studies in Sem. Lang, and Ling., I

2

PART ONE

PHONOLOGY 1. Consonants

Considering the fact that both in the South Arabic and in the ear­ liest Ethiopic alphabet only consonants have been expressed by spe­ cial signs, and that, when in the latter system vowels were accorded such signs, this was made by modifications of the primary consonant signs, the separate treatment of consonants may need no further justification. Because ot the historical nature of our study, we begin with the phonology of the earliest representatives of each group of languages, and only supplement the data drawn from these by means of languages of which only documents from later periods are extant. Since both the several Ancient South Arabic dialects and Old Ethiopic are no longer used as everyday languages, and the former are known ex­ clusively from written documents, we must in each case start from the alphabet in the reconstruction of the consonant systems of these languages; some further conclusions can be drawn from the eventual inconsistencies in the use of these alphabets. However, we will re­ group the sounds, treating them in the normal phonological order rather than the more or less arbitrary alphabetical one; this is neces­ sitated even by the fact that the alphabets of the two languages place the mutually corresponding signs largely into different order. At the same time it may serve as a reminder of the fact that the alphabet — or writing system in general — is no essential part of language, a fact which all too often is forgotten in the treatises dealing with these languages, as is seen from the fact that the South Arabic origin of the Ethiopic alphabet is often used to support the arguments for the South Arabic parentage of the Ethiopic language 1).

E.g., still E. Ullendorff, The Semitic Languages of Ethiopia (London 1955), in the section, “Historical Background^ (p. 4-19), shows this attitude, cf.” . . . here it is writing and language that outshine the importance of everything else” (p. 5), “. . . the Karamatem inscription shows that language to have been of the Sabaean type” (p. 7), “The oldest Ethiopic inscription extant... is that of Matara. . . Here Ethiopic makes its first appearance as a new language” (p. 11); the last state-

CONSONANTS

7

For the sake of convenience, however, we first mention the translitteration of the alphabets in the traditional order; first the South Arabic one — because of its greater antiquity — according to Maria Hofner x) :

’, b, g, d, d, h, w, z, h, h, t, z, j, k, 1, m, n, s, ‘, g, f, s, d, q, r, s, s, t, t; then the Ethiopic one according to August Dillmann *2): h, 1, h, m, s, r, s, q, b, t, h, n, k, w, ‘, z, j, d, g, t, p, s, d, f, p. 1) Gutturals. By this term we understand all the consonants the main articulation basis of which lies beyond the velar region. According to the consensus of scholars, the South Arabic letters & h* \ and & represent sounds belonging to this group. In Old Ethiopic or Ge'e^ 3), the first five letters have correspondences in the alphabet, while the last one is lacking. In the ancient South Arabic inscriptions very few instances of the initial confusion of gutturals can be detected, all of which can be understood as dissimilatory in nature 4);* the only exception is the relatively frequent9 for ‘ which, however, is confined to the Hadramawt dialect and therefore must be regarded as a local phenomenon 6).* Since g appears throughout the history of the ancient dialects and is also present in the modern ment despite the fact that he is aware of the existence of earlier inscriptions in Ethio­ pic written with South-Arabic characters (see n. 36 ib.). The use of Aramaic in the Nabataean inscriptions does not alter the fact that the people spoke Arabic — why should the use of Sabaean only in the earliest inscriptions found in Ethio­ pia mean that no indigenous Semitic languages were spoken there from earlier times? x) Altsiidarabische Grammatik p. 5sq. 2) Table I to Ethiopic Grammar, by A. Dillmann, C. Bezold (and) J. A. Crichton, London 1907. 3) We use this conventional transcription mainly for typographical reasons, well aware of its phonetic inaccuracy as regards the modern pronunciation; incidentally, the scope of our study is primarily historical, which enables us to make a distinction between the fifth and sixth orders by means of the sign of length. 4) See Hofner, Altsiidar. Gr. §§ 18sq. In the cases of an elision, the word (or in any case, the accentual group; but CIH 532je could also be a scribal error, the initial stroke of r resembling the curve of ‘) already contains a guttural before the elided one; while ’ and the a vowel which most probably preceded it in the case of the apocopation, can be conceived as allophones of one and the same pho^ neme (cf. Studia Orientalia XXIX § 1/); the apocopation of the former at the end of a word can thus be regarded as another case of dissimilation, it being replaced by juncture. 6) This strengthens the case for regarding CIH 532 9 (cf. the preceding note) as a scribal error.

8

PHONOLOGY

dialects on the Arabian peninsula ^ its lack in Ge‘ez from the very beginning and throughout the history of the Ethiopic languages is most naturally interpreted as a distinctive characteristic for this group, particularly since the sound is lacking in most other Semitic languages as well*2). A separate group within South-East Semitic is finally formed by Soqotri in this respect; there, not only g, but h also seems originally to have been lacking, having been only rather recently introduced into it through Arabic loan words, in which it still mostly appears 3). Since the other gutturals appear to be in full force, partly even intensified, in this language 4),* the situation with it is different from that with the modern Ethiopic languages, in which general weakening and partial or total quiescization of the gutturals has taken place 6); therefore we propose to introduce So­ qotri as a third group within the larger unity of South-East Semitic, besides Ethiopic and South Arabic, even if we have only modern documents available in it. 2) Palatals, Under this heading we include those consonants the main articulation basis of which is the soft palate (including uvula), accordingly including the sounds termed by some scholars postpalatals, by some others velars. The South Arabic alphabet has signs for three such sounds, viz., g, k, and q. These are also represented in the other groups. In Ethiopic, however, the pronunciation of q differs from that in the modern South Arabic and Soqotri in a way to which we return later. Besides, in Ethiopic there is a whole group of other sounds closely related to this one, viz., the so-called labiovelars (for which this term may be most suitable, because the post­ velar h can be treated in an analogous way), in the pronunciation of which the rounding of lips is added to the basic palatal articulation; for typographical reasons we transliterate them gu, ku, qu, hu, although well aware of the phonetic inaccuracy of this method 6). However, x) Cf., e.g., B. Thomas, Four Strange Tongues from Central South Arabia — The Hadara Group (Proceedings of the British Academy vol. XXIII), London 1937, p. 13 and passim, а) Besides the Arabic group, it is attested only in Ugaritic with certainty; the Greek transcriptions of Hebrew words with Gamma for ‘Ayin probably represent a secondary guttural (cf. M. Flashar, Das Chain in der Septuaginta I, Halle a.S. 1908, particularly p. 19 sqq.). 8) Cf., e.g., W. Leslau, Lexique Soqotri (Paris 1938), p. 195 sqq. 4) See Leslau, Lexique Soqotri p. 17-23. б) Cf. Ullendorff, Semitic Languages. . . p. 35-45. •) The quality of the sound marked with u is easily recognized, anyway, because it is the only case in which the letter u in the transcriptions is followed by a vowel.

CONSONANTS

9

the labialization of these sounds appears to be relatively late, since it has still been operative during the historical period of Ge‘ez 1); therefore it may best be ascribed — at least partly; cf. below, on the sixth order vowel — to Cushitic influence and not regarded as a distinctive characteristic of Ethiopic in the diachronic sense. 3) Denti-alveolars, This group comprises the sounds produced with the tip of the tongue touching — more or less strongly — the teeth or the alveolus; which one it was in ancient South Arabic dialects and in Ge^z, is anybody’s guess and not important. Of the letters of the South Arabic alphabet, normally the following are included in this group: d, d, d, /, /, /; these can be conventionally assumed as representatives of the voiced and voiceless denti-alveolar stops, their spirants and dorsal variants; the voiced dorsal, however, seems to have been a spirant rather than stop, as it is found to have ^ as its variant 2), and also in the modern dialects, a voiced dorsal spirant is found as against a voiceless dorsal stop 3); for typographical reasons, however, we preserve the conventional transliteration. In the Ethiopic alphabet, only the stops and dorsal variants are found; the voiced dorsal again may have been a spirant, since it later on has fallen together with the dorsal sibilant (cf. the next group) in pronunciation. The lack of the simple spirants, d and / however, must again be re­ garded as a distinctive characteristic between Ethiopic and South Arabic, particularly since most other Semitic languages also lack them, among them the third partner of South-East Semitic, Soqotri 4).* 4) Sibilants. The articulation basis of this group is close to that of the preceding one, but somewhat more vague and variable, including the hard palate. The South Arabic characters representing this group are £, £, j, /, /, and /. The first four can be conventionally identified with the voiced and voiceless common sibilants with their respective dorsal variants; the fifth with the socalled sibilant proper (Engl, sh in shine etc.), while the sixth probably represents a lateral variety of j*6). In the modern dialects, £ appears to be lacking, but this need not necessarily be so, because this pronunciation was rather arbitrarily assigned to it in analogy with some Northern Arabic parallels. In­ 9 as seen particularly well in loan words and proper names of foreign origin; cf., e.g., Ullendorff, Semitic Languages. . . p. 76. 2) Hofner, Altsiidar. Gr. § 13. 3) Cf. Thomas, Four Strange Tongues p. 13 etc. 4) See, e.g., Leslau, Lexique Soqotri p. 14sq., 17. 6) Cf. Hofner, Altsiidar. Gr. p. 19 sq.; and my article on The Semitic Sibilants in the Journal of Semitic Studies vol. 11, 1966.

10

PHONOLOGY

stead, Thomas has in his list of modern South Arabic sounds x) one which could be phonetically connected with our hypothetical £, viz. dh\ but in the lack of examples we must leave the question open. The lateral quality of /, finally, appears to have spread to quite a number of other consonants 2); whether the resulting lateral variants are phonemic or allophonic, however, cannot be concluded from the few examples given. Turning to Ethiopic, we find £ and / discard­ ed in Ge^z already 3); since they are present in modern South Arabic, at least partly (even if in a different form), it is probable that here we have another distinctive characteristic between Ethiopic and South Arabic. The modern Ethiopic languages have additionally given up /, Gurage and Harari (and mostly Amharic) even /. Soqotri does not have £, but does have correspondences to the other ancient South Arabic sibilants. 5) Labials. This term includes the consonants in the pronunciation of which the role of the lips is most characteristic. In the South Arabic alphabet, only two characters seem to represent such sounds, a voiced stop £, and a voiceless, probably labiodental spirant/. In Ethiopic, the same sounds are found, but additionally two varieties of a voiceless stop, one of them of the same nature as the sounds above called dorsal; but since both seem to be introduced relatively late in Greek loan words 4),* they can be discarded in our study. In Soqotri, b and / alone are also found. 6) Nasals. These are the two common ones, bilabial m and dentialveolar //, in all the groups. 7) Liquids. These again are the common lateral / and tremulant r in all the dialects; if there are or were varieties, they must be allophonic only 6). 8) Semi-vomls. These are the common bilabial w and prepalataly, as far as we can judge; again without variations between different groups. 9) Prepalatal fricatives and affricates, where such occur in addition to those mentioned above, as well as various spirants etc., appear to be regularly of a secondary character 6) and can therefore be dis­ carded here. x) cf. n. 3, p. 9 above. 2) cf. ib. 3) See Ullendorff, Semitic Languages. . . ; how far the lack of / and / in South Ethiopic is original, cannot be found out any longer. 4) Cf., e.g. Ullendorff, Semitic Languages. . . p. 89 sq. 6) Except in Gurag£, where both sounds appear to be allophones of «; cf. Ullendorff, Semitic Languages. . . p. 126 sq. 6) cf. Ullendorff, Semitic Languages. . . p. 66sqq., 129 sqq.

CONSONANTS

11

There is, however, a group of sounds still in which Ethiopic ap­ pears to deviate, not only from South Arabic and Soqotri, but from all the other Semitic languages as well, and it is generally assumed to have preserved the original character of those sounds, viz. what above were called dorsal^ because in their pronunciation elsewhere the dorsum plays a considerable role. In Ethiopic, however, this dorsal element is replaced by a simultaneous pronunciation of \ which gives the sounds a strongly ejective character. The fact that it is uniform throughout the Ethiopic languages makes the respon­ sibility of Cushitic influence little probable, while the Cushitic pa­ rallels point to the common origin for this phenomenon in both of them; the most natural explanation is therefore that the ejective pro­ nunciation (or a tendency towards it) is an inheritance from the common Semito-Hamitic times 1). 2. Vowels

As regards the distinction of vocalic phonemes, we are in complete darkness in the ancient South Arabic dialects; the conventional method of vocalizing the inscriptions after the Northern Arabic pattern is completely arbitrary and in many instances demonstrably wrong 2); in the lack of other really reliable guides we prefer to dis­ card them altogether for the present purpose. Thus we are left with 1) Ethiopic, 2) modern South Arabic, and 3) Soqotri. In the Old Ethiopic writing, the indication of vowels was intro­ duced at a relatively early stage by means of modifications of the consonant signs after which they were pronounced. The signs being different according to the consonant signs to which they have been attached, they are divided into different “orders”. The first of these orders is normally transcribed with either a or. d\ because it normally seems to represent a Proto-Semitic short ay we use that transcription in this treatise, although d in most positions would be more accurate phonetically. Apart from it, there is only one order which mostly

1) cf. Ullendorff, Semitic Languages. . . p. 152 sqq.; on ^already Brockelmann, Grundriss I p. 121. 2) E.g., the reading of the pion. sf. hnjsn with a geminated n (e.g., Hofner, Altsiidar. Gr. p. 31, 34); that the gemination of the second consonant in the suffixes of 2nd and 3rd pers. pi. is secondary everywhere is made probable by the consider­ ations presented in Studia Orientalia XXIX § 4; and in this particular case no support is given by the other South-East Semitic languages (cf. Akkadian also).

12

PHONOLOGY

seems to represent a short vowel, and sometimes none at all, viz. the sixth one, which we again phonetically inaccurately, but typograph­ ically conveniently transcribe with £, where it appears to have been pronounced. The vowel corresponds to those in the other Semitic languages which are normally assumed to derive from Proto-Semitic / and u. The rest of the orders seem to go back to Proto Semitic long vowels and are therefore transcribed as follows: II. /7, III. /, IV. J, V. and VII.