WORD

ISSN: 0043-7956 (Print) 2373-5112 (Online) Journal homepage: http://www.tandfonline.com/loi/rwrd20

The Phonemic System of Damascus Arabic Jean Cantineau To cite this article: Jean Cantineau (1956) The Phonemic System of Damascus Arabic, WORD, 12:1, 116-124, DOI: 10.1080/00437956.1956.11659595 To link to this article: http://dx.doi.org/10.1080/00437956.1956.11659595

Published online: 04 Dec 2015.

Submit your article to this journal

Article views: 71

View related articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=rwrd20 Download by: [37.44.207.42]

Date: 22 January 2017, At: 23:33

THE PHONEMIC SYSTEM OF DAMASCUS ARABIC JEAN CANTINEAU

The following remarks concerning the phonemic system ot Damascus Arabic have been suggested by an extensive review by Charles A. Ferguson 1 of the Manuel etemenlaire d' arabe oriental ( parler de Damas) that I wrote in collaboration with Youssef Helbaoui. Although there is, in Ferguson's appraisal of that system and mine, a wide measure of agreement, several of his criticisms reveal differences of approach and theoretical divergences which I have deemed worth while presenting and discussing. Ferguson criticizes the consonantal system I have set up on two points: the "foreign" sounds and the emphatics. I still do not believe that the Damascus consonantal system has a v and a p phoneme. There are certain words where v and p can be heard, but they are foreign, borrowed words, and they are felt as foreign precisely on account of these foreign sounds. These sounds do not seem to form distinctive oppositions with the other labial consonants. At any rate, I do not. know of any minimal pairs where they are the sole differentiating element, and Ferguson does not mention any such pairs. Furthermore, these sounds vary freely with other labial consonants: alongside kriive 'tie' there is krawiila (Bergstrasser, 61, 27), alongside narvaz 'he upset (someone)' there is narfaz (Helbaoui), alongside bravo 'bravo' there is ' In Language 30 ( 1954 ). 561-570. Ferguson thinks the book lacks "an introduction setting forth the language situation in the Syrian area and a complete glossary of Arabic word used in the text". As for the glossary, he is right, and we are setting up such a glossary: it will be published with the second- or third-year manuals we are planning. But an introduction treating the linguistic gE>ography of the area would be not at all useful; our manual is an elementary one, for the use of first-year students; these students come to the School of Oriental Languages without knowing either Arabic, or Syria ;the linguistic geography or Syrian Arabic would not be understood. Nevertheless such an introduction will have its place at the beginning of the second- or third- year manual. I shall, in another connection, revert to the morphological points touched upon by Ferguson.

THE PHONEMIC SYSTEM OF DAMASCUS ARABIC

117

briibo (Helbaoui), alongside 'awruppa, 'oroppa-'Europe' there is 'orobba (Bergstrasser, 55, 34). The same can be said for the

o

vocalic sounds 6 and which Ferguson mentions: they vary freely with on and o. All these sounds are not phonemes; they are optional stylistic variants of phonemes or groups of phonemes; they underline a certain modern style of speech. Damascus Arabic has at least four emphatic or velarized consonant phonemes:/, q, ~' ?· There may be others. I have mentioned r on account of the minimal pair 'adre 'pot', 'adra (' lliihiyye) '(divine) power', but is the r really emphatic? Isn't the timbre of the final vowel the true differentiating element? My {notation was merely phonetic, and I do not know of any minimal pairs where l and I would be the sole distinctive elements. Ferguson thinks there are three more emphatic phonemes: ~' q1, ~ but the minimal pairs he adduces are not wholly satisfactory: biiba 'her door' is distinguished from ~a 'daddy, Pope' not only by the emphasis, but also by the fact that the suffix of the third person feminine singular is not simply -a, but something like -(h)a (Helbaoui); and, as a matter of fact, one says byaqarbo 'he strikes', he will strike him', but byaqraba 'he strikes, he will strike her' exactly as if -a did not open the syllable and as if the former -h were still in existence. Mayy, opposed to qwyy 'water', is a proper name, and the use of proper names in minimal pairs is generally avoided in phonological discussions. As for niiyek, it is a very improper word, and, on that account, people carefully avoid using ~iiyek 'your (f.) flute' when speaking to a woman, the two words being too similar; ~ay tabaeek is used instead. Among other words cited by Ferguson, I have heard, without emphasis in Youssef Helbaoui's speech, biiy 'Bey (of Tunis)', veranda 'balcony', lokanda 'restaurant' and also the first m of miiqw 'mommy'. I reject the principle stated by Ferguson that "it is sufficient to establish a phoneme that a sound occur in surroundings in which it cannot reasonably be identified with any already established phoneme in the language". I consider the phoneme as a minimal distinctive unit, and nothing else. A minimal sound not having distinctive function is not a phoneme: it is a combinatory or optional variant of a phoneme. However, the phonemes are not equal among themselves. Their importance is a result of their functional yield; if a phoneme forms with other, substantially similar, phonemes some very frequent oppositions, the phoneme is an important one; if, on the contrary, the oppositions it enters into are used only in a very, small number of minimal pairs, the

118

JEAN CANTINEAU

phoneme is unimportant. For instance, in modern French, the nasal vocalic phoneme re forms an opposition re~e with the similar phoneme e, but this opposition enters into very few minimal pairs; for my part I know of only brun~brin the other pairs : un~Ain, alun~Alain, Aulun~haulain, De Mun=--.odemain, etc. contain proper names ; therefore, the phoneme ce is unimportant; many speakers replace re by e, and they are understood; it would be possible to substitute everywhere e for re without any serious distortion of the language, and if the minimal pair brun=--.o brin were unknown, nothing would prevent us from considering til as an optional individual variant of e. On the contrary, the nasal vocalic phonemes e, a, o are very important and the numerous series of words, such as pain, pan, pont; bain, bane, bon; leinl, temps, ion; daim, dent, don, etc., show it would be impossible to substitute one of these French nasal vowels for another. On account of these considerations, we must be very prudent when setting up the list of phonemes: a short list is often better than a long one. As regards the vocalic system of Damascus Arabic, my critic's reactions indicate that my stand needs clarification. I do not, in the least, distinguish among three degrees of vocalic length: long, medium, and short. This is acceptable neither phonetically nor phonemically. Phonetically speaking, the length of the vowels is extremely variable: there are not three degrees of length; their number is, in fact, indefinite. If we want to simplify things, we can distinguish five degrees: very short, short, medium, long and very long vowels. Phonemically speaking, I do not know of any Arabic dialect having more than two degrees of length: phonemically short and phonemically long vowels. Generally, the phonetically very short vowels are phonemically zero-vowels, appearing in the realization of a consonant cluster; the medium vowels are realizations either of stressed, lengthened short vocalic phonemes, or of unstressed, shortened long vocalic phonemes, or of semi-vocalic y, w in vocalic position; the very long vowels are realizations of stressed, lengthened long vocalic phonemes (see my Analyse phonologique du parler arabe d'El-Hamma de Gabes, BSL 47 (1951. 90-92). I must say that I now count only three long vocalic phonemes: ii, l, and ii. I think that e and o are realizations of the groups of phonemes ay and aw. I was led to this conclusion by morphological considerations: series of words such as names of colours with their feminines:

THE PHONEMIC SYSTEM OF DAMASCUS ARABIC

119

'a!zmar 'red' fern. !zamra, 'af!4ar 'green' fern .. !!arf.ra, 'azra' 'blue' fern. zar'a, 'abyarf. 'white' fern. ber/.a, 'as wad 'black' fern. soda, or substantives with their plurals: ia!zS 'donkey' plur. i!ztjs, bag[ 'mule' plur. bgiil, !zabl 'rope' plur. !zbiil, for 'bull' plur. lwiir, kel 'measure' plur. kyiil, prove that the interpretation of o as aw and of e as ay is deeply rooted in the system of this language. And if people criticize me for introducing morphological facts into the phonemic description, I shall answer that a language is a whole, that there is no iron curtain between phonemics and morphology, and that a phonemic description which does not take into account morphological as well as lexical facts is a bad one. It is on the subject of final vowels that the disagreement between Ferguson and myself is most clear-cut. From my experience there are no short vowels in final position in the Arabic dialects (they disappeared centuries ago). In 1936, I made some kymographic measurements in Syria (see my Etudes sur quelques parlers de nomades arabes d'Orient, II, 145-147, in Annates de l'Inslilul d'Etudes Orientales de la Faculle des Letfres d'Alger III, 1937; my Les parlers arabes du Horan, 144-146; others measurements on the Palmyra dialect and the Algerian dialects are still unpublished; compare Marguerite Durand, Voyelles tongues el voyelles breves, 60-73). The analysis of these measurements has shown that, in the Horan, for instance, the final vowels have a length of between 9 and 17 hundredths of a second: in other words, they are medium or long vowels, and not short ones. The same conclusion is valid for the Lebanese text of which Miss Durand has measured the vowels. I have not taken any measurements for the Damascus dialect, but there is no reason for its final vowels to be shorter than the final vowels of the other Arabic dialects. Ferguson analyses the forms in which a pronominal suffix of the third person masculine singular follows a short vowel as actually characterized by a lengthening of that vowel. This means that gada 'lunch', karsi 'chair', rf.arabu 'they struck' end in a SHORT vowel, and gada 'his lunch', kars[ 'his chair', rf.araba 'they struck

120

JEAN CANTINEAU

him' end in a LONG vowel and nothing else. But since·a, l, u are equivalent to aa, ii, uu, it should follow from this that (1) the pronominal suffix of the third person masculine singular is a vocalic one (just as after words ending in a consonant), (2) the timbre of that vocalic suffix is conditioned by the vocalic final of the word. This would be odd, unexpected, and in total disagreement with Arabic morphology. Ferguson is undoubtedly right on two points, namely when he states that (1) the h of the suffix is no longer audible, (2) the final vowel of the form with suffix is notably longer than that of the form without suffix. The Reverend Father H. Fleisch has proposed another interpretation in his article : Notes sur le dialecte arabe de Zahle, Melanges de l'Universite St. Joseph, 27 (1947-48). 73-116). The -h of the pronominal suffix of the 3rd person masc. sing. is no more heard in this dialect than at Damascus after words ending in a vowel, but H. Fleisch has noted an important fact, missed by Ferguson: in words without a suffix, the final vowel is unstressed and medium; in words with the pronominal suffix of the 3rd person masc. sing., the final vowel is stressed and long. This is true also for Damascus: gada, k~rsi, cjarabu without suffix, but gadd, karst, f/,arabu with suffix. It is easily understood that the stressed vowel is longer than the unstressed one. H. Fleisch thinks this displacement of stress expresses the pronominal suffix. But it is difficult to accept such an interpretation: word-stress has in the Syrian Arabic dialects (and more particularly in the Damascus dialect) no distinctive function, neither lexical, nor morphological; it is hardly credible that it should actually carry there the value of a pronominal morpheme. I will present another interpretation which takes into account what is accurate in the two preceding ones, and adds certain morphological considerations. If we examine a word ending in a vowel, like f/,arabu "they struck," and we add the pronominal suffix, we obtain the following series: f/,arabuni "they struck me" f/,arabuk "they struck you" f/,arabuki "they struck you (fern.)" rjarabu "they struck him", etc. It is clear that the stress and the length of the final vowel of the word are connected with the presence after it of a pronominal suffix. But in the case of f/,arabd "they struck him", where is the suffix? It is not audible. Shall we say that there is a zero-

THE PHONEMIC SYSTEM OF DAMASCUS ARABIC

121

suffix? This would not be accurate: there is no longer a suffix, but its effects remain: the length and the stress of the final vowel, persist as if the -h were still there. Such a suffix I propose to call a virtual suffix and its unheard h a t•irlual realization of the h phoneme. We shall say the same for the h of the pronominal suffix -ha of the 3rd person fern. sing. (see above p. 117). This virtual h can be written in parentheses: biib(h)a, f/.arabii(h). Inside the word, five degrees of phonetic length can be found as against three in final position, (medium, long and very long). In particular, there are, inside the word, vowels of medium length which have a three-fold origin: 1° realizations of unstressed and shortened long vocalic phonemes; 2° realizations of stressed and lengthened short vocalic phonemes; 3° realizations of semi-vocalic phonemes y, w in vocalic position. A phonemic transcription must note the phonemes of which the vowels are realizations; therefore I write blrii/:t, mafall/:t, and I would write nmsylo regularly) "she forgot him" (comp. rJk 3 blo), 'ahwlo "his coffee" (comp. kal3 blo). As for loanwords from Classical Arabic, the rules of the dialect's phonetics and phonemics are in a large measure, imposed on them. In unstressed open syllables short i and u of Classical Arabic cannot persist: they must either disappear or be lengthened; we have bnaye, mdlr (comp. Barthelemy, Diclionnaire, 65, 257), or, phonetically, binaye, mudlr (with medium vowels), phonemically, blnaye, miidlr. In stressed open syllables, short i and u of Classical Arabic become regularly "• for instance, '