The Historical Role of Turkish in Relation to Persian of Iran

John Perry
Iran & the Caucasus, Vol. 5. (2001), pp. 193-200.

1. Introduction
Speakers of Turkic and Iranian languages have been in contact since pre-Islamic times. Buddhist literature of Inner Asia is found in both Old Turkic (Uighur) and Sogdian, and Manichaean literature in Old Turkic, Sogdian, Parthian and Middle Persian, though the extent of early oral contacts and bilingualism cannot be documented. It has long been noted, however, that several of the salient features that distinguish later Middle Persian and New Persian structurally from older forms of Iranian are shared with Turkish  notably the absence of grammatical gender and adjectival agreement, and the use of a singular noun after numerals (Doerfer 1967,pp. 57-59). Linguistic convergence during the Islamic period is immediately noticeable in the form of Persian influence on Turkish, i.e., in the vocabulary and syntax of literary Ottoman and Chagatay and their successor languages, modem Turkish and Uzbek. The converse of this - Turkish lexical and structural influence on Persian - is most striking and best documented at the vernacular level in present-day Central Asia, among speakers of northern Tajik Persian dialects (see Doerfer 1967 and 1992). Apart from lexical borrowing, the sound systems of Tajik and Uzbek over most of the shared dialect area have converged almost completely, and syntactic features characteristic of Turkic are widely shared.

Turkish linguistic influence on the spoken Persian of Iran is not so flagrant, and has mostly been ignored except for some indiscriminate listing of loanwords. Phonological convergence is too complex a phenomenon to pursue here; but, for an instance, it may be argued historically that the collapse of I Q l (the uvular plosive of Arabic and Turkish, not native to Persian) and l Ğ l (the uvular fricative common to Arabic, Turkish and Persian), which resulted in the identical pronunciation of qaf and ğayn in most of western Iran today, was accomplished by the fifteenth century under the influence of Turkish (Pisowicz, pp. 112-4, 117). As for Turkish lexical and syntactic influence on Persian, for all its subtlety it has nevertheless been significant and pervasive in some domains.
We should distinguish two complementary ways in which the advent of the Turks affected the language map of Iran. First, since the Turkish-speaking rulers of most Iranian polities from the Ghaznavids and Seljuks onward were already iranized and patronized Persian literature in their domains, the expansion of Turk-ruled empires served to expand the territcrial domain of written Persian into the conquered areas, notably Anatolia and Central and South Asia. Secondly, the influx of massive Turkish-speaking populations (culminating with the rank and file of the Mongol armies) and their settlement in large areas of Iran (particularly in Azerbaijan and the northwest), progressively turkicized local speakers of Persian, Kurdish and other Iranian languages. Although it is mainly the results of this latter process which will be illustrated here, it should be remembered that these developments were contemporaneous and complementary.

2. General Effects of the Safavid Accession
Both these processes peaked with the accession of the Safavid Shah Esma'il in 1501 C.E. He and his successors were Turkish-speakers. While they accepted and promoted written Persian as the established language of bureaucracy and literature, the fact that they and their tribal supporters habitually spoke Turkish in court and camp lent this vernacular an unprecedented prestige. This expanded its domains of usage, and brought it into competition with Persian as a badge of ethnic and social identity: chronicles repeatedly testify to the perceived functional specialization of Turk and Tajik (the general term for Persian-speakers) as, respectively, khan and mirza, men of the sword and men of the pen, and to their rivalry for status and power in the imperial structure. Foreign visitors (Chardin, Pietro della Valle, Olearius and others) noted that spoken Turkish was so common among all classes in Iran as to be the lingua franca. Upwardly-mobile Persians, according to Kaempfer, actively learned Turkish: "From the court it spread to the leading families of the Persians to such an extent that it is now almost shameful in Persia for a man of distinction to be ignorant of Turkish" (Gandjei 1991,pp. 311-13, 315). Officers of the court and army, and evidently lower ranks, "switched codes" with a facility that passed over the head of naive visitors such as Sir Thomas Herbert (1628): many of the words and phrases in his "Persian" glossary are pure Turkish, e.g., the personal pronouns Man, San,'I Thou He'; Ecbar Tanghy, i.e. the "Turkish ezafa" form of Tang-e Allaho akbar, the well-known gorge near Shiraz; choggee 'Healthie' (Tk. çoq iyi 'very well'). When he adds the hybrid, Na-choggea 'Sick' (i.e., the Persian negative prefix added to Turkish çoq iyi!) one suspects that his obliging bilingual escort might be ribbing him (Perry 1996, esp. pp. 275, 277).
The complementary processes of persianization of literary genres and turkicization of spoken Persian were active not only in the political but also in the religious sphere. If Arabic, the language of the evangelists of Islam, had such a profound effect on Persian, it is hardly surprising that Turkish, the language of the evangelists of state Shi'ism, exerted a direct influence on the status, lexicon and grammar of Persian. Safavid religious policy tended to promote Persian prose at the expense of Arabic on a literary level, in that Shi'i theologians such as Mohammad Baqer Majlesi sought to propagate the official cult by writing in a language that all literate Iranians could understand (Browne, pp. 416-17). Works of popular hagiography, notably the Rawzat al-safa and Fotowat-name-ye soltani of Hosayn Va'ez Kaşefi, further established Persian as the primary language of Iranian and international Shi'ism. Conversely, this propaganda introduced into vernacular Persian some popular Turkish cultic terms as established in Qizilbash usage. One such was tuq, the Turco-Mongol term for a military standard hung with yaks' or horses' tails, that was applied to certain types of 'alam or processional standard used in the Moharram mourning procession. (From this is derived the modern Persian patuq 'hang-out, local meeting-place', <  pa-ye tug 'at (the foot of) the tuq'.) Another was ta'zie, the synonym for P. şabih-xani, which has become the international technical term for the Moharram "passion play."
The latter is of course Arabic in origin, but its morphological and semantic evolution show it to be of Turkish provenance (see Fig. 1). The etymon was incorporated into Persian as ta'ziat 'mourning; condolence', as seen in a variety of compounds and idioms such as ta'ziat-name 'letter of condolence' and ta'ziat goftan 'to express condolences'; in this context it has been superseded in both Persian and Turkish by forms of tasliat and tasalli. The shifted form ta'zie , and the meaning 'funerary procession', are not found in written Persian before the late 19th century (see Loğatname-ye Dehxoda, s.v.; and Perry1991, pp. 182-83). This is the usual form, however, in all Turkic dialects, and the senses 'obsequies, funeral' and 'Moharram commemorative rites' are found in Azerbaijani Turkish. We may thus infer that the earlier form and meaning of the word entered through Persian into Turkish, where it lost final -t and (among the Qizilbash) gained its specialized cultic sense, then was re-acquired in Persian to designate a popular Shi'i practice that developed under Safavid tutelage.

3. Lexical Borrowing
There are hundreds of loanwords from Turkish of all periods listed in Gerhard Doerfer's works (he designates about 1.200 of these "Azeri", - see Doerfer 1988, p. 246), and a variety of Turkish words and phrases are attested in Persian verse from pre-Mongol times onward, as identified in several articles by Tourkhan Gandjei. Both phenomena attest to the continuing symbiosis of Turks and Persians, especially in the realms of power politics and popular culture. Neither of these corpora, however, can be used to document the lasting effects of Turkish on Persian. Much, probably most, of the Turco-Mongol vocabulary in Classical Persian histories and the like is ephemeral, i.e., it comprises obsolete military and administrative terms such as daruğe and soyurğal. Similarly, most Turkish words showcased in the Persian poetry of such as Nezami, Khaqani, Suzani and Rumi are less than ephemeral - they have never been incorporated, even temporarily, into Persian; the verb forms and phrases, in particular (e.g., oltur 'sit down', qonaq gerek 'do you want a guest?'), were not even candidates for lexical borrowing. Like Abu Nuwas in his macaronic fahlawiyyat, the poets are being cute and showing off.
Literary references do sometimes demonstrate (and even initiate) the incorporation of vocabulary, phraseology and syntactic calques. Mawlana's dictum mahi az sar gande gardad, ney ze dom in the third book of his Masnavi (cf. Homa'i; Gandjei 1998) corresponds to a still current Turkish proverb, balik baştan kokar 'a fish rots from the head down' (i.e., corruption sets in first in high places). As a consciously vernacular poet, reveling in popular sayings and jokes, Rumi was more prone than most other "Classical" writers to include elements of Turkish speech, even if in disguise. This proverb, however, is still more at home in Anatolia than Iran. An Azerbaijani Turkish-language proverb long current in central and southern Iran, amongst a majority of Persian-speakers (see Anjavi-Shirazi, pp. 32-34), was popularized by the late Mohammad-Ali Jamalzadeh as the title of a satirical piece in his famous collection Farsi şekar ast (1926). Jamalzadeh did not need to translate bile dig, bile Çoğondar (lit. 'such a (big) pot (for) such a (big) beet' or, roughly, 'pull the other one' - a colorful Isfahani comment on an unlikely tale); and such was the story's success that the catchphrase is accepted nationally in its foreign dress, much like, e.g., "plus ça change..." in English. If the principal domain of lasting Turkish influence is the vernacular, we must search for it in everyday Persian speech, or at least in writing that attempts to preserve vernacular phraseology.
Rather than list the dozens of everyday Turkish loanwords still current in Persian (such as otaq, ojaq, otu, ordek, qeyçi, qandaq , qişlaq, yaylaq, qaçaq, tasme, toman, tutun, tup, etc.), I will illustrate in Figs. 2-4 a few borrowings that are (or were) embedded structurally in the morphological, syntactic or social patterns of the Persian lexicon. We will see that Turkish has infiltrated Persian in subtle, covert ways that  have not even been identified as turkicisms.

Fig. 2(a) is a historical curiosity. The Danish traveller Carsten Niebuhr lists a series of terms for officers of provincial musketeers (tofangçi) campaigning on the Gulf coast in the 1760s. They are alternately Turkish borrowings and Persian-Turkish hybrids: modeled on the Turco-Mongol decimal system, the ranks nevertheless express some of the numerical qualifiers in Persian, as min-başi 'head of 1.000' yuz-başi 'head of 100', but pansad-başi, panjah-başi, dah-başi (see Perry 1996, p. 277). A similar lexical-structural mix - which has survived to the present - involves the productive use in Persian of the Turkish denominal agentive suffix -çi. Such derivatives have been formed on native Persian words (or assimilated loans), independently of Turkish lexical models: Şekar-çi 'hunter', taqlid-çi 'mimic', post-çi 'postman' (via French; cf. Tk. posta-ji; via Italian), tutun-çi 'tobacconist' (cf. Tk. tütün-jü, Azeri tütün-çü).
Fig. 2(b) shows the tendency of Turkish loans to occupy a lower sociostylistic register than their Persian counterparts, from which they may then diverge semantically to a degree. When the Normans conquered England in 1066 (soon after the Seljuks occupied Baghdad), they introduced a layer of French vocabulary into English. Some of the resulting synonyms were neatly distributed, with appropriate semantic specialization, in accordance with the social class and typical function of the original speaker, with (Old) English as the vernacular and (Norman) French as the elite language; so that when the Anglo-Saxon farmer delivered his English swine, kine, calves or sheep to the Norman lord of the manor, they would be served up at table as pork, beef veal or mutton. A similar selection of synonyms by sociolinguistic register may be observed in some sectors of Persian (as the elite language) and Turco-Persian (as the vernacular). Thus, while Persian kard is the hero's blade or the neutral kitchen implement, Turkish çaqu tends to connote the ruffian's switchblade; Persian banu is the literary, high-status title, xanom the everyday term. Of the words for 'horse', Persian asb is the riding steed, Turkish yabu is the pack nag (we may compare this with Latin equus versus Vulgar Latin caballus, the etymon of French cheval); Persian xar is the literary wild ass (gur xar) or the metaphorical idiot, Turkish olağ(deva1ued from its original meaning of 'post horse, relay pony') is the working donkey. Of the synonyms for 'gently, slowly', aheste is the formal or literary word (as calligraphed on warning notices), yavaş the usual spoken word.

4. Syntactic Influence
A more subtle sociolinguistic manifestation of turkicism in Persian can be seen in the syntax (as defined by word order) of certain onomastic noun phrases. In Fig. 3, the reversal in word order in royal titulature from the type with a preposed title (Şah Esma'il, Şah Abbas, Şah Solayman) to the type with a postposed title (Nader Şah,Fath'ali ŞahReza Şah),which is standard in Iran from the end of the Safavid dynasty, is simply the unconscious adoption of Turkish syntax (i.e., modifier before head, as in Mehmet aqaAli qoli xan, Süleyman padşah) in what had previously been a Persian noun phrase (head before modifier; the head noun is the constant term, i.e., the rank or title, while the modifier is the variable, i.e., the respective personal name).

This was merely a belated stage of a process that had been under way since Timurid times or earlier. In, e.g., Tahmasb Mirza and Iraj Mirza an etymologically Persian title is postposed to designate a prince of the ruling (Turkish) house, whether Safavid or Qajar, imitating the postposing of Turkish titles such as xan and aqa. (In Tajik, imitating Uzbek, the Persian title xoja/ xuja was postposed too, which did not happen in Persian.) The title soltan (of Arabic origin), when by early Qajar times it had been devalued to the rank of a village headman, was also postposed, as recorded orally of "Rustam Sultoon," headman of Khesht in Fars (Malcolm, pp. 123- 25; Perry 1990, esp. pp. 221-23).
These particular turkicisms originally expressed the political dominance of Turkish military dynasts over the Iranian world. Again we have an analogy in early English, where Norman political dominion introduced military and legal terms ordered in accordance with French syntax, which are still with us: attorney general, sergeant major, court martial. In Iran (see Fig. 3(a)), the syntactic shift left the way clear for the devalued titles şah, mir and mirza (lit. 'born of an emir') to be preposed, as per Persian syntax, for metaphorical use in the titulature of Sufi leaders and of bureaucrats and writers, representing social niches in which Persians retained cultural and quasi-political prominence: hence Şah Ne'matollah Vali, Şah Mir Hamza, Mir Haydar; Mirza [Mohammad Sadeq Nami , Mirza [Abo'1-Qasem] 'Aref. Note that "Reza Shah" is the king, even though no longer ethnically or linguistically Turkish, whereas "Shah Reza" (şah reza) is the saintly incumbent of the shrine at Qomisheh.
The traditional word order of Turkish titles borrowed into Persian followed Turkish syntax, i.e., came after the name (Hasan Aqa, etc.; Fig. 3(b), right column). When the European-style surname was adopted under Reza Shah, the Turkish aqa and xanom were persianized, i.e., adapted to the ezafe noun phrase (left column).
A number of collocations for culinary delicacies such as barre kabab, juje kabab, six kabab (a translation of Turkish şiş kabab), and perhaps sabzi polow, etc., are frozen Turkish syntagms from this same period, when the Timurid and Safavid court chefs were evolving a Turco-Persian cuisine fit for a hybrid royal culture. The Persian ezafe construction is occasionally substituted, as, e.g., kabab-e barre, and the khoreshts are syntactically Persian (as are the dolmehs, though the word is etymologically Turkish) (Perry 1990,pp. 226-27).
The lexical influence of Safavid court and camp was not, of course, confined to Persian. A new stratum of Turco-Persian loanwords (some ephemeral, some lasting) was superimposed on the earlier Iranian vocabulary already assimilated into the languages of the Caucasus. One such is the Georgian supra, denoting a celebratory banquet involving a series of toasts and the concomitant consumption of much wine and vodka - cognate with Persian sofre (sufra 'napkin, table(cloth)'), which currently refers to a pious votive repast where alcohol would be highly inappropriate. The lost link is the royal banquet (sofra) of early Safavid rulers, at which the Shah and his officers would honor Turco-Mongol steppe tradition by drinking each other under the table (see Perry 1996, p. 277).

5. Semantic shift and Rückwanderer
Perhaps the most subtle (and speculative) sphere of Turkish influence is that of semantic shift, where an etymologically Persian or Perso-Arabic term acquires a specialized meaning in a turcophone context and in this sense is passed back to Persian (known etymologically as a Rückwanderer). I have already mentioned this process as exemplified in ta'zie (Fig. 1; Fig. 4).

A similar case of an Arabic loanword in Persian which has probably been incubated in Turkish and returned to Persian (though this is harder to prove) is mosafer (i) 'traveller, passenger'; (ii) 'transient, (hotel) guest'. Both meanings are present in Persian and in Turkish of Turkey and Azerbaijan, but whereas in (literary) Persian the original sense of 'traveller' is primary, in Turkish the derived sense (backformed from mosafer-xane 'inn, transient hotel') is primary. I believe that a frequency-count of occurrences in spoken Persian would show the derived meaning 'guest at cheap hotel' to be primary here as well, arguing that the immediate provenance of Persian mosafer 'hotel guest' (as of ta'zie 'Moharram passion play') is not medieval Arabic but Turkish in recent centuries (Fig. 4). Plausible reasons for this process include the continuing influx of Turkish-speakers into the metropolis (especially Qajar Tehran, from the early 19th century) to occupy important socioeconomic niches in urban life.
Similar processes have been at work in several of the Arabicate political, technical, commercial and administrative terms that entered Persian in the course of the 19th century. Words such as melliyat 'nationality', aqalliyat 'minority', bahriye 'navy', tayyare 'aircraft', mozarebe 'limited partnership' were coined first in Ottoman Turkey and spread quickly with the burgeoning press into Persian and even Arabic, where their provenance was assumed to be wholly Arabic (cf. Farshidvard, pp. 60-61). A term still in active service which is most likely a calque on French introduced through the medium of Turkish about the turn of the 20th century is jabhe '(military, political) front'. Certainly the word Jabhat, later Jabhe 'brow, forehead' is a genuine Arabic loanword found also in Classical Persian and Turkish (and with a figurative sense 'chief ), but the military metaphor 'front line, battle front' is not recorded before the 20th century, even in Arabic.

6. Ways and Means
The high status that Turkish attained as the language of the ruling elite has diminished in Iran with the assimilation and ultimate demise of the turcophone dynasties of recent centuries. Turkish vocabulary and turkicisms, never as pervasive in Persian as the Arabic element, have been seamlessly assimilated and, where there is any sociolinguistic difference, are to be found in the lower social registers of present-day Persian speech. This very seamlessness and vernacularity are the reason why Turkish influences do not stand out with the high profile of Arabicisms in literary Persian, and the processes of incorporation of Turkish borrowings are obscured. Even the effects of translation from Turkish into Persian - the major context of linguistic influence - is disguised, for two reasons:
(1) Translation is initially oral, as in the case of a folkloric corpus such as Mulla Nasreddin jokes, and even that of folktales translated from Turkish into Persian by a bilingual Azerbaijani writer such as Samad Behrangi, in his Afsaneha-ye Azerbaijani. The original "text" is not readily available for comparison, so the traces left in calques of Turkish semantemes and speech patterns are easy to miss and hard to prove.
(2) Translation is unconscious, consisting in original composition by bilingual writers such as Mirza Abd ol-rahim Talebuf (Talibov; ca. 1844-1910) of Tabriz, or the poet Shahryar, or articles in the satirical journal Molla Nasreddin.
In conclusion, I would like to suggest that this partnership with Turkish is one reason why Persian never became flagrantly diglossic, to the extent of Arabic or of Turkish before the language reform of the 1930s-40s. Islamic New Persian evolved in symbiosis and competition with Arabic. Content for the first three or four centuries to resign to Arabic the status of "high" language (of scripture, liturgy and scholarship), written Persian and spoken Persian thus developed side by side: socio-linguistic dichotomy there was, but by way of bilingualism with Arabic, not diglossia within Persian. The Iranian intelligentsia wrote in Arabic while speaking in Persian (and writing popular essays and poetry in Persian). This state of hierarchical bilingualism was reversed after the Mongol invasions: Arabic throughout the Eastern Islamic world suffered a reduction in its intellectual domain - became, in effect, a dead, classical language - while Persian expanded into philosophy and the religious and secular sciences, written not only by Iranians but Turks, Indians, etc. Conversely and contemporaneously, spoken Persian was challenged for hegemony as the lingua franca of the same region by other vernaculars, the most important of which were several varieties of Turkish. Once again, Persian was in a hierarchical bilingual (not diglossic) partnership, this time as "high" language to Turkish's "low" - a situation that has lasted in many parts of the Iranian world up until the present time.
Obviously, bilingualism with Arabic and/ or Turkish is a sufficiently polarized sociolinguistic dilemma, without complicating things by going diglossic. When native speakers of Turkish or, say, Kurdish, Baluchi or Pashto needed to reach a wider audience, they would use written or spoken Persian; and their style, insofar as they were educated, tended toward the universal, or more formal rather than less formal, standard. When the native Persian elite wished to express their national identity vis-a-vis the Turks or Arabs in their midst, they would use a Persian comprehensible to, and valued by, their colinguals of whatever socio-economic station. The pressures for a fissure of this language of solidarity along diglossic lines were simply not there.

No comments: