http://www.scribd.com/doc/2576086/fo...assical-arabic

Journal of American Oriental society 106.2



Prosody and the initial formation of Classical Arabic

العروض ودوره في تكون اللغة العربية

استرعى انتباهي ما تضمنته هذه الدراسة من رأي قد تكون له أهمية وتبعات فقررت ترجمتها وإضافة ما يخطر لي من تعليق حولها .


سألون تعليقي بالأزرق.

يقصد بالعربية في هذا المقال اللغة العربية الفصحى الكلاسيكية كما عرفت في العصر الجاهلي.

ص- 333

لعب العروض العروض العربي في تشكيل العربية ، وإجمالا فإن العلماء يتفقون على أن الفصحى كانت لغة تعم القبائلsuper-tribal ويستعملها شعراء الجاهلية. ولم تكن اللغة الدارجة لقبيلة أو مجموعة من القبائل

إن هذا القول ركيزة للموضوع، ولا أرى تلازما بين القول إن العربية كانت تعم القبائل كلغة شعرية كما يفيد السياق والقول بأنها لم تكن اللغة الدارجة لقبيلة أو مجموعة قبائل.



وهكذا فلم تكن العربية لغة طبيعية بل كانت لغة ( مهذبة – مصنّعة – مولّفة – cultivated ) تكونت لتلائم الغرض الذي وضعت لأجله وهو الشعر بمشاكله العروضية لا سيما الوزن. وكذلك لخدمة السجع.



هكذا يقرر الكاتب أن الشعر حصرا لا سواه هو سبب وجود اللغة العربية ، فكل قبيلة أو مجموعة قبائل كانت تتفاهم بلهجتها الخاصة. ولولا الشعر ما كانت هذه الفصحى لتكون أصلا



وثمة دليل على أن الوزن والقافية كانا فاعلين في صياغة البناء الصرفي للغة من خلال حملهما الشعراء على ابتكار صيغ كلامية جديدة وتعديل صيغ أخرى وبشكل عام تكوين ملامح اللغة لتخدم الوزن وأحيانا القافية. والدليل على ذلك هو نتائج عملية التنميط ( القولبة ) كما عرفناها في الشعر الجاهلي والقرآن [ الكريم ] والأعمال الأدبية الأخرى.

بقية الموضوع على الرابط:


https://sites.google.com/site/alaroo...me/arud-lughah


ككثير من المواضيع التي تتطلب وقتا رأيت أن أنشر هذاالموضوع في رابط في موقعي لأتمكن من استكماله حسب ما يتاح من وقت. ولتسهيل ما قد يتطلبه من تقويم .



Prosody and the Initial Formation of Classical Arabic Author(s): Farhat J. Ziadeh Source: Journal of
the American Oriental Society, Vol. 106, No. 2, (Apr. - Jun., 1986), pp. 333 -338 Published by: American Oriental Society Stable URL: http://www.jstor.org/stable/601598 Accessed: 10/04/2008 21:35
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPubl...lisherCode=aos. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

http://www.jstor.org

Prosodi

and the Initial Formation

of Classical Arabic

Arabic prosody had had a telling effect on the formation of the classical Arabic language. It is generally agreed among scholars that classical Arabic was a super-tribal language, a koine, that was used in pre-lslamic Arabia by poets, seers, and composers of rhymed prose, and that it never constituted the spoken vernacular of any one Arabic-speaking tribe or group. It, thus, was never a "natural" language but a "cultivated" language that developed in such a way as to fit the purpose for which it was used, namely poetry with its problems of prosody, particularly meter, and rhymed prose. Evidence presented suggests that meter and rhyme were instrumental in shaping the morphological structure of the language by forcing the poets to coin new word forms, to modify others, and generally to make many features of the language serve the requirements of meter, and sometimes rhyme. The evidence consists of the results of the molding process in the language as it is known to us in pre-lslamic poetry, the Qur'an, and later literary works.

For some time now, I have had the suspicion that Arabic prosody has had a telling effect on the formation of the classical Arabic language (hereafter c.a.l.). It is generally agreed among Western scholars that c.a.l. was a supertribal language, a koine, that was used in pre-Islamic Arabia by poets, seers and composers of rhymed prose, and that it never constituted the spoken vernacular of any one Arabic-speaking tribe or group.' Traditional Muslim scholars maintain, of course, that c.a.l. represented the language of the Quraysh tribe, the tribe of the Prophet Muhammad, inasmuch as the Qur'an employs that language, but that view is even questioned by modern Arab linguists who assert that c.a.l. developed as a distinct entity different from the vernaculars of Quraysh or Tamim or any other tribe.2 If we assume the koine theory to be true, then it follows that c.a.l. was never a "natural"language but a "cultivated"language that developed in such a way as to fit the purpose for which it was used, namely poetry with its problems of prosody, particularly meter, and rhymed prose. It will be remembered that in preIslamic Arabia literature consisted almost exclusively of poetry and rhymed prose. Scholars have already pointed out that poets, in their endeavor to find the right word for a given metrical position or a given rhyme scheme, often drew upon the diverse tribal dialects which supplied such words. Examples included equivalent forms of demonstrative pronouns, verbal nouns, and broken plurals.3 This process of borrowing from the various See Michael Zwettler, The Oral Tradition of Classical Arabic Poetry (Columbus, Ohio, 1978) p. 101 and the authorities quoted there. 2 See for instance M. T. al-Duwayk, "Al-Lughah al'ArabTyahal-Mushtarakah," Al-Sharq al-Awsat (14-2-1983) p. 12. 3 Zwettler, p. 11 and the authorities cited there. 333

dialects might help explain the rich stock of synonyms and near synonyms found in classical Arabic poetry for objects, concepts, and activities.4 This paper, however, proposes to go beyond the process of borrowing, described by these scholars, to offer some evidence that meter and rhyme were instrumental in shaping the morphological structure of the language by forcing the poets to coin new word forms, to modify others, and generally to make many features of the language serve the requirements of meter, and sometimes rhyme. That this process might not have been limited to the Arabic language can be gleaned from what M. Parry said about the Greek language. In his "The Homeric Language as the Language of an Oral Poetry" he says that the oral poet in formulating his poetry might repeat an element of diction that had been heard long ago, or modify or adapt his traditional diction to produce a new element, or adopt a word or usage from a speech group other than his own because it had no material equivalent in his own speech, or even, under the pressure of a particular metrical need, generate an otherwise nonexistent form to fill out a line.5 H. A. R. Gibb, speaking of the "un-expectedness" of preIslamic poetry, says, "At one moment Arabia seems, in a literary sense, empty and dumb ... at the next, companies of poets spring up all over Northern Arabia, reciting complex odes ... with vividness of imagination ... in an infinitely rich and highly articulated language."6 This might be true only if we understand by "moment" a sufficient period in history to allow for the develoment of an "articulated language" to fit the exacting requirements of meter and rhyme. Charles James Lyall, in his Introduction to Ancient Arabian Poetry says: Ibid. The Making of Homeric Verse: The Collected Papers of Milman Parry, ed. A. Parry (Oxford, 1971) p. 334, as quoted in Zwettler, p. 99.
5

334

Journal of the American Oriental Society 106.2 (1986) It is impossible to fix with any degree of certainty the date when the Arabs first began to practise the art of poetry. The oldest poets of whom we have any remains belong to the time ... [of] about a hundred and thirty years before the Flight [Hi'ra]. But these are spoken of, not as the inventors of the poetic art, but as the authorities for the laws of the Kasidah or ode.... What we possess of the distinguished poets to whom these laws were due is cast in forms which we cannot but suppose to be the outcome of a long education in the construction of verse. The number and complexity of the measures which they use, their established laws of quantity and rhyme, and the uniform manner in which they introduce the subjects of their poems, notwithstanding the distance which often separated one composer from another, all point to a long previous study and cultivation of the art of expression and the capacities of their language, a study of which no record now remains.7 A word of caution is necessary here. At the outset of this paper we said that we have had the suspicion that Arabic prosody influenced the formation of c.a.l. But suspicion sometimes leads to paranoia which sees evidence where none is found! If we should reach that point we hope that readers will set us aright. The first and disarmingly obvious evidence is a feature of the language for which grammarians have formulated the rule: man' ilhiqa' al-sakina 'n, (no juxtaposition of vowelless consonants) or no consonant clusters.8 It is known, though, that at least modern Arabic colloquial dialects do handle consonant clusters, and one would assume, for lack of evidence to the contrary, that ancient dialects handled them as well. Why should c.a.l. then abjure them? The answer might be found in the meter of Arabic poetry which is composed of varying series of short and long syllables made up of cv for short, and cv or cvc for long syllables. No other combination (e.g., vcc or ccv) is possible. To fit into this meter, the syllables of c.a.l. had to conform to cv, cv, and cvc patterns only. Even the Quranic cvc-cvc (e.g., radlun) becomes in poetry ci-cvcvc.9 To be sure, linguists have recognized that the most common type of syllable structure of world languages is cv or cvc. If so, then it can be said that the poets either preserved, or harked back to, that original character of the language. The other category of evidence pertains to broken plurals and their myriad forms. Why, one might ask, are there several plurals for a non-derived noun? It has already been suggested that the poet of c.a.l. often drew upon the diverse tribal dialects which supplied a broken plural to fit a metrical position or a rhyme scheme.'0 Might not such a poet use a broken plural pattern to invent a new plural of a noun, or even to modify a pattern by adding a feminine ta' or elongating a short vowel to fit his metrical position? To test the first proposition, I asked a number of my fellow Arabists what the word ashhar meant, and all answered "months" without hesitation, although they qualified their answer by saying that such a word has not been used in c.a.l. But if a poet had used it in a poetical formula and other poets followed suit, the word would have become a standard plural, as probably many others have become in a similar fashion. As an actual example of this process, one might cite the Quranic usage of 'ijaf as the plural of 'Ca/a', "lean," instead of the regular plural 'u/f because of the desire to produce an assonance or a parallelism. The reference here is to Surah XII,
8 The only quasi-,exception is in words like jaffun where the alif is considered a vowelless letter followed by another vowelless letter. But in reality the alif is a long vowel, and the rule still applies. It is instructive that such words can only be accommodated in poetry by the formjarfifun. 9 See note 8 above. 10 See note 3 above.

It is the contention of this paper that in this long period of cultivating the art of expression and the capacity of the language to deal with it certain features of the language were molded, modified, or even invented to accommodate meter, rhyme, and assonance. The result was the classical Arabic language that the Qur'dn later used-with certain modifications derived ostensibly from the dialect of Quraysh-and that has been used ever since as the literary language. Here, a question might legitimately be raised as to whether we are not putting the cart before the horse in saying that the verse meter shaped the language, whereas the natural process is for the language to shape the meter, since language comes first in point of time. This might be true in naturally spoken languages, but c.a.l., as pointed out above, was developed as an artificial, supra-tribal, language for the purpose of composing poetry to be understood in Arabia from the confines of Syria in the North to the Yemen in the South, and therefore it had to be responsive to the requirements of poetry. The base for this language, however, must have been a dialect or dialects located most probably in north and northeast Arabia from where most of the early poets hailed. The evidence that will be presented must, in the nature of things, be circumstantial; it consists in the main of the results of the molding process in the language as it is known to us in pre-Islamic poetry, the Qur'5n, and later literary works. Even some evidence of the molding process itself can be detected in those works in the attempt of the language to accommodate meter, rhyme, and assonance. Arabic Literature, Oxford, 1963, p. 13. of Ancient Arabian Poetry (Edinburgh, 1885) pp. xv-xvi.
7Translations
6

ZIADEH:

Prosody and the Initial Formation of Classical Arabic

335

Verse 43: ara sabca haqaratin sirman va'kuluhunna sahbun cij.f; "I see seven fat cows being eaten by seven lean ones," where ciijf parallels with sirman, "fat," in being of the same pattern. The Quranic commentator Baydawi says, "According to ordinary analogical formation, the word "lean"would have been C'uijsince it is the plural of 'ajfal'; but because it here its stands in rhetorical antithesis to sinman, structure has been modified to accord with the structure of sirnan."' If this can happen in the Qur'an for purposes of parallelism, it is submitted that similar modifications must have occurred in poetry to accommodate rhyme and meter. The pre-Islamic poet Shanfarah (early sixth century), in his famous ode with an L rhyme (Ldmiyyah), used the word khuyutah as a plural of khayt, "string," instead of khuyut which is of a more common form. Apparently he did so to fit a metrical position in his tawTlmeter: wa atwTCala'Ikhumsi 'Ihawdaykama 'ntawat khuyutatu mdriyyin tughdru wa tuftalu "I hold my entrails on hunger and fold them as the spinner would fold the strings that have been twisted well"

(e.g., 'umuimah and khu'ulah, "unclery"> "uncles"),l5 but this cannot be visualized for khuyutah. If ta' marbutah has been added, then, for metrical reasons, dare one ask whether it was added for identical reasons in the myriad forms of the infinitival noun of Form I verb? But more on this later, as our attention now is focused on plurals. Reference has already been made to the existence of several plurals for many non-derived nouns in c.a.l. A fruitful avenue of research might be a study of the various forms of plurals used in early Arabic poetry for each non-derived noun to determine the effect of meter in such formation. A spot check of dictionaries for words commonly used in pre-Islamic poetry, like names of animals and desert features, produced some interesting results. The Arabic word for panther, nimr or namir, has no less than eight plurals: numr, numur, anmur, anmair, nimdr, nimirah, numur, and numarah, presumably fitting all metrical possibilities! The following nouns had at least four plurals: bayt, "tent, dwelling"; thawr, "bull"; diir, "house, court"; rTh,"wind";sayf, "sword";shibl, "whelp of a lion"; shirif, "old she-camel"; shaykh, "chief of tribe"; sahib, "companion"; sakhrah, "rock";tarTq,"road";fahl, "stallion"; fawj, "collection of men"; ndqah, "she-camel" (with eleven plurals!). Another case in point is the question of plural of plurals. Why, one might ask, should there be plurals of plurals? Do they serve a function other than providing the right form or pattern for a given metrical position? To be sure, grammarians, in their attempt to give logical explanations for types of broken plurals divide these plurals into three categories: (a) plurals of paucity that include the forms af'ul, af'al, a/f''lah and fi'lah to denote entities between three and ten, (b) plurals of abundance which comprise sixteen other forms, and (c) plurals of plurals which consist of the patterns afaCil, af/aiT, fa'a'il and 'fcalTn.'6 But the differences between plurals of paucity and abundance are surrounded by exceptions and qualifications that virtually vitiate them. As to the plural of plurals, we are told that it is used "to increase the number of units that the plural comprises."'7 Since the plurals that have plurals of plurals are mostly of the "paucity"variety on the patterns of af'ul, af'Cl, and of af''lah, it was thought that the plural of plurals really had a function in increasing the number of units comprised in the plural. But, I fail to see any difference, say, between the "paucity" plural a'rah, "bedouins," and the plural of plurals a'arTh,"bedouins," save perhaps to allow poets like Abu Nuwas (d. about 810) to say metrically: 15 Notice that the word uhuilah means the "quality of a stallion" as well as "stallions." 16 See Wright, Vol. 1, p. 234, and Rash-d al-Shartuni, Mahaci ' al-A rahJiah (Beirut, 1963) pp. 89-91. '7 Shartuni, p. 90.

u--Iu---I

u--

u-u-u-

u-ul u- -I uu u

Without the additional tii' marbutah in khuyutah, the meter would not scan, especially since the ta' marbutah forms a part of an essential element of the meter known as the watid. The editors of al-Majini al-HadTthah,no doubt basing themselves on the commentaries of medieval editors, explain away the tia marbutah by saying it indicates a plural of abundance (kathrah).'2 But the plural form khuyut is itself of an abundance form since it is not of a form indicating paucity.'3 Wright, in listing khuyutah among a few others of this pattern (fu'alah) as a broken plural, characterizes this pattern as rare.14 No doubt it was listed as a plural in dictionaries because Shanfarah had used it as such. The other examples listed by Wright are for the most part collective in meaning indicating a natural grouping or class, like bu'ulah, "husbands"; 'umumah, "uncles"; numurah, "panthers";fuhulah, "stallions," etc. One can understand how a semantic shift might take place from fu'ulah as an infinitival noun (cf. muru'ah and rujalah, "manliness") to fuculah as a plural

See A. F. L. Beeston, Bai.ddwT's Commentary on Surah 12 of the Qurd'in(Oxford, 1963) p. 25. 12 Luwis Shikhu, Al-Majiini al-Hadithah (Beirut, 1946) Vol. I, p. 7. 3 See W. Wright, A Grammar of the Arabic Language (Cambridge, 1964) Vol. 1, p. 234. 14 Ibid., Vol. 1, p. 223.

336

Journal of the American Oriental Society 106.2 (1986) laysa 'la'arTbucinda'llahi min ahadi "To God, bedouins are nobody." afaTwij. Was the second pattern used to fit a zihaf, or shortening of a vowel, required by a certain meter? If so, why does the plural aqdah "featherless arrows" have only one plural of plurals, aqadih? Is it because aqadlh was never used early in c.a.l. to fit a metrical scheme? It may be noticed that the same process of shortening a vowel embraces as well the regular plural of singulars that have a syllabic structure similar to a plural (e.g., the plurals of i'sar "storm" are both a Casr and a'asir) and the shortening probably happened for metrical reasons. One final piece of evidence of a broken plural being used to fit a metrical scheme is one used by the poet al-Hutay'ah (d. 678?) who, by such usage, confused generations of commentators. In a famous verse of his he says: man yaf'ali 'Ikhayrald yu'dam jawaziyahu "He who does good would not want his rewards [or his rewarders]"

The real reason for the existence of plurals of plurals was alluded to by grammarians without comprehending the significance of that allusion. They say, "the plural of plurals assumes the pattern of the plural of a singular which has the same pattern as the plural, e.g., the plural of aklub [which itself is a plural of kalh, "dog"] is akalib in the same way that the plural of anmul, ["fingertip"] is anamil."8 They limit this statement, though, to what they call sTghat muntaha al-jumu', "the patterns of the ultimate plurals,"namely, xaxaxix and xaxaxix (where x stands for a consonant). Actually, all plural of plurals patterns are identical in form with plurals of singulars which have the same patterns as plurals, as the following chart of examples shows. It is to be noted that the plurals which have plurals of plurals are not all plurals of paucity, for the patterns fi'al and fui'ln which have p.p. are plurals of abundance. Sing. kalb, "dog" makan, "place" zahrah, "flower" jamal, "camel" 'uqab, "hawk" Plural aklub amkinah azhar jimal 'uqban P. of P. akalib amakin azahir jama'il19 Caqabmn Pattern of P. af'ul af'ilah af'al fical fu'lan

Pattern of P. of P. afa'il afacil afa'lT fa'a'il facalin

Sing. Like P. anmul, "fingertip" asirah, "tie" icsar, "storm" hizam, "belt" thu ban, "snake"

P. of S. Like P. anamil awasir a'CsTr haza'im thacabin

The inescapable conclusion is that since certain plurals have the same patterns as singulars, the language assimilates them to singulars and gives them a plural that we call plural of plurals. This phenomenon can be observed among ArabAmericans who give an Arabic plural to the English plural "shoes" thus: shuz > ashwaz, similar to the Arabic singular .iiq "market"> aswaq. It is to be noticed also that plurals which are not of the same patterns as singulars (e.g., fiu'ala, af'ila') do not have a broken plural of plurals. To the extent that c.a.l. allowed the formation of plurals of plurals, poets, one would assume, must have made full use of it and encouraged it along to fit their metrical schemes. In this regard one wonders why, for example, the plural akras "contiguous houses" and afwaj "collections of men" have two plurals of plurals each: akaris and akaris, and afawUT and

Ibid., p. 91. '9 For a sextuple plural of jamal see, M. B. Schub, "A Note on a Sextuple Plural in Arabic," Al-'Arahiivya, Vol. 15, Nos. I and 2, 1982, pp. 153-55.

18

According to Lane's Arabic-English Lexicon (under the root jaza) the word jawazin as it occurred in this verse was considered by some writers as the plural of the masdar, or infinitival noun, jaza', "reward,"while some others considered it the plural of the quasi-infinitival noun (i.e., ism masdar) jaziiah, "reward,"while still others considered it the plural of jazin, "rewarder." There is no doubt that the poet meant "reward"or "recompense,"jaza', but such a word would not fit in the metrical scheme, so, he used the plural jdwazi which vaguely suggests a plural of jaza' and left the riddle for later generations to ponder! A third category of evidence that we advance rather gingerly, because it is based more on a hunch rather than on solid grounds, is the great number of infinitival nouns of the first form of the verb. One wonders, for instance, why the first-form verb samuha, "to be generous," has six infinitival nouns: sanah, sunmuh, sam7ahah, sumuhah, samh, and simah, "generosity." Was it because of dialectal variation or metric necessity? As is well known to Arabists, the first-form verb has no less than forty-four infinitival nouns. Of these, twentyfour are formed (a) from the tri-consonantal root with a

ZIADEH: Prosody and the Initial Formation

of Classical Arabic

337

variation of vowels (e.g., fa/l, f'al,/ fu'u, il, etc.), (b) from some of these patterns with the elongation of the medial vowel etc.) (e.g.,fa'al,fuC'il,fa CTI, and (c) from the feminine (with a ta' marbutah) of these patterns. The variations that have metrical values are the existence or absence of a vowel (as in faCl compared to faCal), the elongation and consequent accentuation of a vowel (e.g., faail,hf/il, fa/lT), and the ta' marhiutah.Is it possible that these variations were the result of metrical necessity, and that this metrical necessity of ancient origin left its legacy in the language'?Alternatively, even if we assume the existence of all these patterns independently of any influence of metrical necessity, would it not be tempting for a poet to use a pattern not previously used for a given root to fit his metrical scheme, thereby enlarging the circle of infinitival nouns for that root? The fourth category of evidence embraces those instances where words are shortened apparently to fit a metrical scheme. At the top of this category is what is called by grammarians tarkhTm,"softening," the elimination of one or more letters from a noun in the vocative. Examples are afatima for afdtimatu, "O Fatimah!"; sahi for sd.hibT, "[0] friend!"; yd suci for ya suCidu, "O Su'ad!," etc.20The most famous of pre-Islamic poems, the ode of Imru'u 'l-Qays, in contains two examples of tarkhTm lines 19 (afatima) and 70 (asihi, "O friend!"). Other instances in this category are (a) the double apocopation of yakunu, "to be," in the jussive mood to yaku, taku, aku, etc., a usage found in pre-Islamic poetry (e.g., in taku in Imru' 'I-Qays, line 21) and in the Qur'an (e.g., lam aku in 19:20 and lam yaku in 19:67), (b) the elimination of the final alif of ma following a preposition (e.g., fima, "in what?," bima, "with what?," lima, "why?"instead offima, bimd, and limi), (c) the elimination of the initial vowels of huwa, "he," and hiya, "she," following the conjunctions wa and fa, thus: wahwa,fahwa, and wahya,fahya, (d) the occasional elimination of the letter mTmin baynami, "while," giving us bayna, (e) the occasional shortening of aymunu 'l-Lah, "by the oaths of God!" to aymu 'l-Lah, (f) the occasional elimination of the tiid of fifth and sixth form verbs after the ti' of the imperfect, a usage found in pre-Islamic poetry (e.g., taraqqa for tataraqqa and tasaffali for tatasaffali in Imru' 'l-Qays, line 68) and in the Qur'an (e.g., tafarraqui for tatafarraqu in 3:103),21 and the similar shortening of ittakhadhiu, "they adopted," to takhadhui,and (g) the shortening of al-ili, "the first" to al-ulia in reference to the ancient Arabs as in the expression shiyamu 'l-uld,"the character of the first (Arabs)." Although many of these shortenings can be attributed to the process known as haplology in languages, poets seemed to

have taken full advantage of such shortenings and imparted to them an acceptable status. Rhyme and assonance might have played a similar, but undoubtedly lesser, role to that of meter in the formation of c.a.l. We have already alluded to the Quranic usage of 'ijif instead of 'ujf for purposes of assonance. The Qur'an provides other examples where rhyme dictated the addition of rhyme letters to proper names that otherwise would not have had them. I am referring to sTnTn sTnii,"Sinai," and ilyiisTn for for ilyias, "Elias," in 95:2 and 37:130, respectively. But the rhyme letters have not really affected the common pronunciation of these proper names. One can also speculate as to whether the sound masculine plural form in words like calamin,22 "worlds," ar.idn, "lands," sinun, "years" were dictated by rhyme, since these words are not among the category of words that normally take the sound masculine plural. One can be certain, though, that the use of the plural form sinTn to mean "affected by drought" by al-Tirimmtah (d. early eighth century A.D.) in the following verse was necessitated by rhyme: bi-munkharaqin tahinnu r-rThu fhi hanTnal-hulbifi l-baladi s-sinTni In a gusty tract, the wind moaning therein like the moaning of goats in a country affected by drought. Now the singular word sanah means "drought" as well as "year." The broken plural form sinTnas well as the sound masculine plural form sinin are regularly used as plural of "year." But the use of sinTnas a plural of "drought" to modify the word halad, "country," is strange, and can only be explained by the requirements of rhyme. The attempts by lexicographers to explain this anomaly by saying that the word halad is a collective that can take a plural modifier are not convincing.23 The tyranny of rhyme reached ridiculous proportions in a poem entitled HadTth al-Himar, "The Conversation of a Donkey" by Bashshar ibn Burd (d. A.D. 783). In the poem a donkey describes the beauty of a she-ass, the object of his love, and says, wa-laha khaddun asTlun mithlu khaddi sh-shavfartrni She has a full and smooth cheek like the cheek of al-Shayfaran
22 In the Qur'in, the genitive/accusative form cilamTn is used seventy-three times with only one usage (25:1) not in a rhyme situation. 23 See E. W. Lane, An Arabic- English Lexicon, under entry sanah.

20 For a treatment of this subject, see Wright, Vol. II, pp. 88-89. 21 See Wright, Vol. 1, p. 65.

338

Journal of the American Oriental Society 106.2 (1986) meter, and rhymed prose. Enough evidence exists to suggest that meter and rhyme were instrumental in shaping the morphological structure of the language by forcing the poets to coin new word forms, to modify others, and generally to make many features of the language serve the requirements of meter, and sometimes rhyme or assonance.
FARHAT J. ZIADEH UNIVERSITY OF WASHINGTON

Someone asked Bashshar, "What is al-Shayfaran?" He replied, "How should I know? This must be one of the obscure words of the donkey. When you come upon him, ask him!"24 To recapitulate, c.a.l. was a "cultivated" language that developed in such a way as to fit the purpose for which it was used, namely poetry with its problems of prosody, particularly
24 Luwis Shikhu, Al-Majini al-Hadcthah, Beirut, 1946, Vol. 111,p. 23.