iii. Arabic Influences in Persian Literature
Any inquiry into the early development of Islamic Persian language and literature is faced with the same problem—the absence of contemporary material. We are not to suppose that Persian was not written down until the 3rd/9th century (from which period the earliest extant examples of Islamic Persian poetry are cited); yet we do not even know when the Pahlavi alphabet was abandoned in favor of the Arabic. References to the use of Persian and/or Arabic in administrative records may conceal evidence on this point, but it is too scanty to be of much help, and may indeed refer solely to the language in which the records were kept. The earliest surviving Persian manuscript in Arabic script (see below) dates only from the beginning of the 4th/10th century, and from this we can see that the principles underlying the adaptation of the Arabic script to the writing of Persian were already established. Arabic loanwords appear in their Arabic form without change of spelling; Persian sounds not found in Arabic are represented by their nearest equivalent, thus pāʾ by bāʾ, čīn by ǰīm, gāf by kāf (a practice that continued well into the 6th/12th century and beyond). E. G. Browne (Lit. Hist. Persia I, pp. 8-11) has plausibly suggested that the choice of script was a religious one. Whereas Zoroastrian books continued to be written in the Pahlavi script, Persian converts to Islam naturally turned to the script of the Koran. This process may very well have begun very soon after the Arab conquest, since the Pahlavi script was the preserve of a small priestly class, whereas the Muslim converts were drawn largely from the illiterate classes. For such people it would be common sense, since a new script had to be learned anyway, to choose the much simpler Arabic script; the fact that it enjoyed religious sanction was an additional bonus.
Similar problems apply to the impact of Arabic phonetics on Persian. As we shall see, a great many Arabic words found their way into the language at an early stage, and we suppose that the literati (as indeed the religious dignitaries of the present day) attempted to retain the Arabic pronunciation. Of course we have no means of knowing for certain how either Arabic or Persian was pronounced in the 1st/7th century; early grammarians (like Sībawayh—2nd/8th cent.) interested themselves in Persian sounds that did not occur in Arabic, but not in the converse. Analogy suggests that Persian, like most other languages, resisted the importation of sounds that were not native to it. Certainly that had become the case by the time European scholars and travelers began to study the language: ḍ had become ż, ṯ had become sÂ², ḏ ẕ, ṣ s, ṭ t, and so on. One might have hoped to gain some light from Persian rhyme; but unfortunately the traditional classical rhyme was based as much on the visual as on the oral form. This could mean that there remained some slight differentiation in the early stages, or it could simply be the weight of convention. Where Arabic did influence Persian phonology was in the introduction of new consonant clusters. The Arabic triliteral root permits virtually any combination of two consonants to occur, and while no work has been done on this theme, it would not be too difficult to compile a list of combinations that are peculiar to Arabic words: bṭ as in rabṭ, bh as in šebh, ḡl as in aḡlab, taḡlīz, qw as in taqwā, taqwīm, tq as in etqān, atqīāʾ, zy as in tazyīn, to mention only a handful.
But of course the most significant impact of Arabic on Persian has been in the expansion of vocabulary, a process that seems scarcely to be complete even at the present day. Once again we are handicapped in examining the origins of this development by lack of contemporary material; the earliest samples of New Persian consist of a handful of verses dating from the end of the 3rd/9th century, by which time the influx had been going on for two centuries. According to M. Bahār (Sabk-šenāsī I, p. 256) the process began with words, mainly of an administrative or religious nature, for which no Persian equivalent existed; but it must very soon have passed this stage. The four verses by Ḥanẓala of Bādḡīs (mid-3rd/9th cent.) contain, out of 54 words, three Arabic words (if we except the controversial ǰāh) ḵaṭar, ʿezz, and neʿmat; but doubts have been expressed as to the authenticity of these verses in their extant form, first cited in the 6th/12th-century Čahār maqāla. Other poets of the same period, Moḥammad b. Waṣīf and Bassām-e Kord, first cited in the 7th/13th-century Tārīḵ-e Sīstān, have a remarkably high proportion of Arabic words amounting to about a quarter of their total word stock, but it must be emphasized that available material is too small to be statistically significant. In general the percentage of Arabic words used by the early poets (as cited in Lazard, Premiers poètes) stands at about ten percent. Certain poets seem to have made an effort to avoid the use of Arabic. Rūdakī, for example, in the hundred bayts that survive of his Kalīla wa Demna, uses only about twelve words from Arabic (the approximation is necessary because, in this as in all other cases, it cannot always be determined with certainty what the origin of a particular word is—for example dīn); on the other hand, the main body of his surviving verse contains eight to nine percent. The Šāh-nāma notoriously contains very few Arabic words, though even here estimates differ widely—P. Horn offers 430, Humbert 984, Ḵānlarī 800; and these figures refer to lexical items, not to total vocabulary. Humbert estimates that 22 words are used more than 100 times, while 470 are used only once (Ḵānlarī, “Loḡathā-ye ʿarabī dar Šāh-nāma,” pp. 405-6). One might conclude from all this that the overall percentage is about two percent.
The one poet for whom, since the publication of M. N. Osmanov’s Chastotnyĭ slovar’ Unsuri (Moscow, 1970), sufficient statistical material exists to make a reasonably accurate estimate, is Abu’l-Qāsem Ḥasan ʿOnṣorī (d. 431/1039-40). Although Osmanov does not go into the question of the Arabic element in the Dīvān of ʿOnṣorī, it is possible to pick out from his complete concordance those words that may be taken to be borrowed from Arabic (though the caveat entered above continues to apply). From this it emerges that out of 4,824 different words used by ʿOnṣorī, some 1,544 are of Arabic or part-Arabic origin, or about thirty-two percent. Out of his total wordage of 46,472, 8,066 fall into this category, or about seventeen percent. From this it follows that Arabic words tend to be less frequent in use, and this is borne out by the statistics for the 2,268 words used only once, of which 777 or about thirty-four percent are of Arabic origin. The Arabic word most frequently used (139 times) turns out to be dawlat,
coming at no. 35 in the list, and preceded (apart from common particles, prepositions, etc.) only by such words as šāh, gaštan, dāštan, āmadan, dīdan.
We have of course no means of knowing how far Arabic permeated spoken Persian as opposed to the written language of the educated classes. A small clue is afforded by the dobaytīs of Bābā Ṭāher ʿOryān, who during the 5th/11th century wrote in the dialect of western Iran. It is surprising to find that Arabic words form as much as nine percent of his texts (though we have to bear in mind the possibility of later interpolations). However the figure is very similar to that arrived at by a study of a modern dialect from the same area, as exemplified by a group of folk-tales in the writer’s private collection.
In the case of prose works subject matter clearly has a greater influence than in poetry. Ḵānlarī (Tārīḵ-e zabān-e fārsī II, p. 55) has pointed out that the proportion of Arabic words is significantly less in works like Samak-e ʿAyyār or the Eskandar-nāma than in religious and philosophical treatises such as the Kašf al-maḥǰūb. The difficulty however of an objective assessment, without a detailed lexical analysis, is seen in two different estimates of the Qābūs-nāma (475/1082). Whereas Bahār (Sabk-šenāsī II, pp. 114-5) claims that the author makes definite efforts to avoid the use of Arabic words, trying to find Persian substitutes whenever he can, G. Lazard (La langue des plus anciens monuments de la prose persane, p. 103) remarks that the Qābūs-nāma is “notable for its abundance of Arabic words.” Both are indeed right, but are looking at the question from two different points of view. The earliest prose work available is an early 4th/10th-century manuscript discovered some years ago in the Shrine of Emām Reżā at Mašhad (Aḥmad ʿAlī Raǰāʾī, Pol-ī mīān-e šeʿr-e heǰāʾī wa ʿarūżī-e fārsī, Tehran, 1974). This is a Persian rhythmical translation of a long passage from the Koran. The subject matter would lead us to expect a high proportion of Arabic words, but in fact this is not the case. The overall percentage is about eleven percent, but even more surprisingly, of this figure only about three percent are repeated from the Arabic original, the remaining eight percent being words that were by this time, we must assume, sufficiently integrated into the Persian language to be intelligible to the readers for whom the manuscript was intended, that is to say, readers without a knowledge of Arabic.
It was suggested above that the primary stimulus to the importation of Arabic into Persian was utilitarian, but this clearly did not remain the case for very long. A knowledge of Arabic (which long remained the only language for religious, philosophical and scientific works) was regarded as essential equipment for an educated man, and it was obviously tempting for him to introduce Arabic words into his Persian writing. Kay Kāʾūs b. Eskandar advised his son in the Qābūs-nāma (475/1082): “Adorn your own letters with quotations, proverbs, Koranic verses and traditions of the Prophet; and if a letter is to be in Persian, do not write it in pure Persian, because that would be unacceptable.” (Kai Kāʾūs Ibn Iskandar, A Mirror for Princes, tr. R. Levy, London, 1951, p.201). Neẓāmī ʿArūżī, writing in the Čahār Maqāla in 550/1155, advised the perfect secretary to be thoroughly familiar with both the Arabic and the Persian classics. On the other hand, Neẓām-al-molk (d. 485/1092) quotes with approval in the Sīāsat-nāma a saying attributed to Ḥasan of Baṣra (d. 110/728): “The man of learning is not he who knows a great deal of Arabic and has the ability to use Arabic words and constructions. Rather he is learned who is competent in any branch of scholarship, whatever his language.” However he does add: “It is all the better if he does know Arabic; God having sent the Koran in the Arabic tongue and Mohammad having been of Arabic speech” (R. Levy, The Persian Language, London, 1951. p. 13).
Whatever the motive, the intermixture of Arabic with Persian continued apace, until by the 6th/12th century some works contained as many as eighty percent of Arabic words. The influence was not confined to the incorporation of new words (though the point was reached where any Arabic word was considered eligible for use, and dictionaries were scoured for fresh material to inflate the vocabulary of learned writers). Arabic grammatical usage also found its way into Persian, though usually only with Arabic words. Thus agreement of adjectives became usual (ṣefāt-e ḥamīda, omūr-e ḥaqīqa); Arabic plurals are used side by side with Persian (ʿolamā and ʿālemān); Arabic verbal nouns replace Arabic adjectives with Persian noun endings (boḵl for baḵīlī, ʿezzat for ʿazīzī); the Arabic dual makes its appearance in particular cases (ṭarafayn); as does the tanwīn in the accusative (samʿan wa ṭāʿatan, nasyan mansīyan). Later on complete Arabic phrases and even sentences could be incorporated (maqbūl al-qawl, ʿalā ʾayyi ḥāl ). By the post-medieval period there seemed no longer to be any distinction in the mind of the writer between Arabic and Persian words, which were used interchangeably (though it should be stressed that this did not apply to syntax except marginally (be-zadand zadan-ī saḵt = ḍaraba ḍarban šadīdan, and so on). Persian words could even acquire Arabic suffixes (dehāt) or forms (dahāqīn) and vice-versa (ṭalabīdan, raqṣīdan).
Stylistically the influence was more noticeable in prose than in poetry. Foremost was the use of saǰʿ, rhymed prose, a feature of Arabic style from very early times, as indeed of related Semitic literatures (Brockelmann, GAL I, p. 13, GAL, S., p. 22). After its use in the Koran, however, it disappeared from Arabic literature, and only gradually reappeared. The first Persian writer to have made extensive use of saǰʿ was Ḵᵛāǰa ʿAbdallāh Anṣārī (396/1006-481/1088), while the maqāma form was imitated in the Maqāmāt-e Ḥamīdī of Qāżī Ḥamīd al-dīn (d. 559/1164). By this time the use of elaborate figures of speech was a commonplace of both prose and poetry, though how far it would be true to attribute this to Arabic influence is doubtful. It was perhaps rather a feature of literature produced by a small learned class, equally fluent in both languages, and developing similar figures in both languages in a style designed to appeal to an erudite or would-be erudite class.
The same figures appear in Persian poetry from the 5th/11th century onwards, but in other respects the influence of Arabic is less obvious. The commonly held belief that the Persian poetic meters were derived from Arabic seems now finally to have been abandoned (Ḵānlarī, Wazn-e šeʿr-e fārsī, Tehran, 1958; Elwell-Sutton, The Persian Metres, Cambridge, 1976). The question of the monorhyme remains open. Certainly there is evidence of rhyme and even monorhyme in the surviving fragments of pre-Islamic Persian poetry, but scarcely enough from which to draw any conclusions. It has been argued that the monorhyme developed because of the facility provided by the structure of Arabic for the formation of rhyming words, yet the great majority of rhymes in Persian poetry are Persian. On one point there can be little argument: the qaṣīda is a conscious borrowing from Arabic, and in the hands of some poets (Manūčehrī, Moʿezzī, for example) almost slavishly follows the Arabic outline, beginning with the poet’s arrival at his beloved’s deserted encampment, even though such an incident was entirely remote from the experience of urban Persian poets. More commonly such poems open with a description of spring, praise of wine, or (in one famous poem of Farroḵī) the branding of horses. To the other verse forms found in Persian there seems less reason to attribute an Arabic origin. The idea that the ḡazal was originally a part of the qaṣīda that became detached so as to form an independent poem seems highly improbable. The stanzaic poem (mosammaṭ) is found in both languages, but whereas the earliest extant Persian ones date from the early 5th/11th century (Manūčehrī), the Arabic ones would appear to be of considerably later date. Finally the right of the robāʿī to be regarded as a native Persian growth has never seriously been challenged, though some would give it a Turkish origin in Central Asia.
The process described above continued throughout the post-medieval period, and by modern times Arabic words had became such an accepted element in Persian vocabulary that they hardly seemed to be recognized as such. Arabic words had long been treated grammatically as though they were Persian, but now Persian words began to acquire Arabic suffixes to an extent not seen before, cf. farmāyešāt, or the curious plural ending -ǰāt incorporating a long since vanished gāt: dastaǰāt, rūznāmaǰāt, and even occasionally applied to Arabic words: ʿamalaǰāt. Macaronic phrases like ḥasab al farmāyeš become commonplace.
The 13th/19th century also saw a growing trend toward simplification of the language, the so-called bāzgašt-e adabī, one of the pioneers of which was the minister Qāʾem-maqām Farāhānī (1193/1779-1251/1835). Partly under the influence of expanding literacy, partly through the example of European literature encountered for the first time by young men furthering their studies in the West, and partly under the pressure to find new words to express new concepts, writers began to move towards a simpler, less convoluted style, to use common Persian words rather than rare Arabic ones. Another leader in this was Mīrzā Malkom Khan (1249/1833-1326/1908), who already in 1273/1857 was writing pamphlets in what was practically a colloquial style. Some writers specifically composed entire treatises and books in pārsī-e sara, naḡz (pure Persian); a good example is a brief history of a few years of the Qajar dynasty written by Farhād Mīrzā, younger brother of Moḥammad Shah, in 1252/1836, while another is the Nāma-ye Ḵosrovān, also an historical work, of Jalāl-al-dīn Mīrzā, a son of Fatḥ-ʿAlī Shah, first published in 1285-88/1868-71. This movement, though it continued almost up to the present day, was always rather artificial. Aḥmad Kasrawī, who was assiduous in digging up Persian roots to form new words for new concepts, was obliged to accompany many of his books with a vocabulary; while the Farhangestān, founded in 1314 Š./1935, was able, after five years of labor, to produce only some two thousand words to replace European as well as Arabic ones, mostly of a technical, medical or scientific nature-and many of these inventions failed to stick.
It is especially significant that when the need arose for new words to describe new political ideas, particularly during the Constitutional period in the early years of the present century, politicians and journalists instinctively turned to Arabic rather than Persian. Nor did they by any means restrict themselves to the words that Arabs themselves chose. An obvious example is the word for “constitution,” Arabic dostūr (in origin a Persian word!), Persian, mašrūṭa; many others could be cited, e.g. “mandate”—Arabic entedāb, Persian qaymūmīyat; “United Nations”—Arabic al-omam al-mottaḥeda, Persian melal-e mottafeq; “charter”—Arabic mīṯāq, Persian manšūr. Here a domesticated form of Arabic exerts its influence from within the host language. Many “Arabic” words used in Persian today would scarcely be recognized in the lands of their origin: mawqeʿīyat, eʿzām, ešḡāl, monawwar al-fekr.
The tendency towards simplification of language still has some of its impetus. Novelists and fiction writers tend to restrict their use of Arabic, perhaps to as little as ten percent. Journalists accept a higher proportion, of the order of twenty percent. The highest percentage of Arabic is still to be found in the writings of religious authorities, the language of the mollās that was satirized sixty years ago by Moḥammad-ʿAlī Jamālzāda in his famous short story Fārsī šekar ast.
See also Bahār, Sabk-šenāsī, 3 vols., Tehran, 1942, 2nd ed., 1958.
C. E. Bosworth, “The Interaction of Arabic and Persian Literature and Culture in the 10th and Early 11th Centuries,” al-Abhath 27, 1978-79, pp. 59-75.
Idem, “The Persian Impact on Arabic Literature,” The Cambridge History of Arabic Literature I, Cambridge, pp. 483-96.
Browne, Lit. Hist. Persia, esp. vol. I, chap. 1, vol. II, chap. 2.
ʿA. A. Ḥekmat, Pārsī-e naḡz, Tehran, 1951.
P. Horn, “Neupersische Schriftsprache,” in Geiger and Kuhn, Grundr. Ir. Phil., pp. 2-5.
P. Humbert, “Observations sur le vocabulaire arabe du Châhnâmeh ” Mémoires de l’Université de Neuchâtel 22, 1953.
P. N. Ḵānlarī, “Loḡathā-ye ʿarabī dar Šāh-nāma,” Zabān-šenāsī wa zabān-e fārsī, Tehran, 1964, pp. 402-11.
Idem, Tārīḵ-e zabān-e fārsī, 4 vols. in 5, Tehran, 1969-78.
R. Koppe, “Statistik und Semantik der arabischen Lehnwörter in der Sprache ʿAlawi’s,” Wissenschaftliche Zeitschrift der Humboldt-Universität zu Berlin, Ser. 9, Berlin, 1959-60, pp. 585-619.
G. Lazard, “Les emprunts arabes dans la prose persane du Xe au XIIe siècle: Aperçu statistique,” Revue de l’Ecole Nationale des Langues Orientales 1, 1965, pp. 53-67.
Idem, La langue des plus anciens monuments de la prose persane, Paris, 1963.
R. Levy, The Persian Language, London, 1951.
M. Dj. Moinfar, Le vocabulaire arabe dans le Livre des rois, Wiesbaden, 1970.
W. Skalmowsky, “Ein Beitrag zur Statistik der arabischen Lehnwörter im Neupersischen,” Folia Orientalia 3, 1961, pp. 171-75.
(L. P. Elwell-Sutton)
Originally Published: December 15, 1986
Last Updated: August 10, 2011
This article is available in print.
Vol. II, Fasc. 3, pp. 233-237