INDO-IRANIAN FRONTIER LANGUAGES. Since the time of the Achaemenid Empire, the territory of present-day Pakistan has been under the cultural and linguistic influence of successive stages of the Persian language. The political history of Persian in South Asia and its spread as the language of literature and politics is traced in Alam (1998, 2003), while its role in South Asian education is discussed in Rahman (2002, pp. 121-60).

The northern and western parts of this region are now occupied by speakers of Khowar, Kalasha, and Shina (Indo-Aryan, “Dardic” languages: see DARDESTĀN ii), and recently Urdu (q.v. at, Indo-Aryan); Balochi (Western Iranian; see BALUCHISTAN iii), Pashto and Wakhi (Eastern Iranian); Brahui (q.v., Dravidian); and the isolate Burushaski (q.v.). Since Urdu became the national language of Pakistan in 1947 and increasingly functions as the country’s lingua franca, it has replaced Persian as a compulsory language in the curriculum. From the 1980s the presence of Persian in the educational system became negligible. Despite this, a significant influx of additional Perso-Arabic words has entered the lexicons of all the languages of Pakistan through Urdu. Several recent publications deal with such loanwords: Bukhari (2003) lists 1,003 words of Persian and Arabic origin common to Urdu and Khowar; Akbar (1992), a similar work, lists numerous words common to Urdu and Shina. Khattak et al. (1977), lists some 3,700 nominal forms common to Urdu, Sindhi, Pashto, Panjabi, and Balochi; of the 200 basic verbs listed there, most are not shared. The great majority of the shared words are of Perso-Arabic origin and, in addition to their nominal or adjectival function, provide the raw material for new verbs, since the primary mechanism for formation of new verbs in these languages is the combining of a nominal loan element (Perso-Arabic or English) with a native verbalizer, most frequently “be” or “do.”

Persian models have served as the stimulus for the beginnings of literary production in most of these languages, and Persian poetic forms are still highly influential in Khowar and Balochi. The contemporary Balochi poet Gul Khan Nasir wrote first in Persian and Urdu, then in Balochi, and his Balochi poetry employs the prosodic structures and poetic genres of classical Persian (Jahani, 1994-95). Persian has also been the principal vehicle for the transmission of Arabic vocabulary throughout the Islamic culture area (Perry, p. 3). The strength and influence of the Perso-Arabic and Islamic cultural heritage is the major reason for the decisions of most writers in these languages to retain the original spellings of their numerous Perso-Arabic loans, rather than to devise more phonologically based orthographies. The case of Balochi differs somewhat in this respect; a significant percentage of Perso-Arabic words are written with “balochified” spellings (Jahani, 1989, p. 159), e.g., tufān “storm” instead of ṭufān.

Persian has influenced the phonology, lexicon, and syntax of all these languages, both directly and indirectly; Khowar and Balochi, for example, show a pervasively direct influence. In turn, Khowar has been a channel of secondary transmission of Persian lexical items and grammatical features to Kalasha and Burushaski. Balochi has been the vector for secondary transfer of many Persian words into Brahui. Pashto too has functioned as a vehicle for Perso-Arabic words into languages spoken in areas where it is the regional contact language. For example, in Torwali, spoken in upper Swat (Inamullah), we find laṛaz-u “to tremble” (< Psht. laṛz-edal “id.” < Pers. larzīd-an “id.”); adōs “ritual cleansing before prayer” (< Psht. awdas “id.” [Bellew, p. 7]) < Pers. ābdast “id.” [Steingass, p. 6]), in which the meaning has remained constant. In some cases, meaning has shifted; e.g., Torwali barē “possibly, presumably” reflects a change from Pashto bāre “once, at last, then” (Bellew, p. 10), and Persian bāra “once, sometimes, at length, in short, at all events” (Steingass, p. 143).

Persian grammatical features, such as subordinate clauses introduced with the multi-functional conjunction ki “that, when,” and to lesser degree the eżāfa (q.v.) construction and compounds formed with unstressed, enclitic o “and,” have entered most of the languages of this region. Even Urdu relative clauses, which are formed with an inherited Indo-Aryan relative word in j-,now often exhibit the j- element supplemented with ki, e.g., jo ki, “which” instead of jo. These influences are so massive and pervasive that one can discern, in the area where Urdu is now the lingua franca, an emerging Pakistani linguistic area (sprachbund), superimposed on the earlier, Persian-induced cultural/linguistic area and the larger South Asian linguistic area (cf. Masica, 1976, passim). Of the languages in this area, Pashto, Shina, and the Kohistani languages seem to show fewer of these structures. For example, Pashto employs its native conjunctions ce and tse to construct relative clauses; Shina employs prenominal finite clauses in the relative function, and Kalam Kohistani employs either participial relatives or a finite clause including the head noun and introduced by an indefinite element kā~ (Baart, pp. 26-27).

In most of the languages under consideration, the phonemes /x/, /γ/, /f/, /q/, and /z/ are marginal, reflecting relatively recent NPers./Urdu influence, and speakers frequently replace them with /kh/, /g/, /ph/, /k/ and /j/, respectively. In Balochi, their status varies from dialect to dialect: /q/ is normally replaced by /k/ in all dialects (Jahani, 1989, pp. 81-82), while in Rakhshani, /f/, /x/, and /γ/ are usually preserved, and in Makrani they are usually replaced by /p/, /k/, and /h/. Khowar and Burushaski (see below) are exceptions in this regard.

Balochi dialects spoken in Pakistan reflect an earlier stage of Persian influence, in comparison with the more recent and ongoing direct influences of New Persian on Iranian Balochi discussed in Jahani (1999), who concludes that influence from Iranian Persian is likely to be the source of divergence between Pakistani Balochi and that of Afghanistan and Iran. Both Eastern (EB) and Western Balochi (WB) form relative clauses (RC) with ki/ke, in addition to participial constructions. For example: WB amā mard RC[ki rāh-ā rawag-ā int] manī māmā int “The man RC[walking on the road] is my uncle,” and EB hawā~ maṛd RC[ke oδā jakkiθī e~] maī~ bābā e~ “The man RC[(who is) standing there] is my uncle” (Bashir, 1991b, p. 169). Relative clauses constructed with ki are found in Brahui as well. For example: amo bandaγ RC[ke nī oṛ-tun īt karesa] kanā syāl ass “The man RC[with whom you were talking] is my relative” (Bashir, 1991a, p.153).

The Brahui lexicon reflects several strata of Persian (and other Iranian) influence, analyzed in Rossi (1979). Rossi (pp. 1-55) enumerates 418 borrowings from Balochi, of which at least 247 are ultimately of Persian origin, e.g., āṛtī/ārtī “flour used for a wedding feast” (Bal. art, Pers. ārd “flour” [p.2]); and 1,461 borrowings from Persian (and Balochi) or non-Iranian languages (mainly Urdu), e.g., swārī “riding” (Pers. suwārī, Ur. səwārī, “rider; vehicle,” Rakshani Balochi sUari [p. 270]). Some of these have Brahui forms identical to Persian and Urdu but not to Balochi, e.g., fauj “army” (Pers. fawj, Ur. fəuj, Rx. Bal. pəwj, Bal. p’əwj (Rossi, p. 191).

Khowar shows substantial influence from Iranian languages at various time depths (Morgenstierne, 1936). Among the languages of northwestern Pakistan, it shows the greatest direct influence of New Persian. In Chitral, until quite recently Persian dominated both the secular and the religious spheres. It is second only to Arabic as a language of religion, and, except for the Qurʾān, basic religious texts are in Persian. Even today in the Chitral area, the niyyat “intention (to say prayers)” is taught in Persian in many homes. Persian was the only language of written communication and government of the princely state of Chitral until 1953, when it was replaced by Urdu as the official language. Pre-1940 education was informal, conducted in the homes of the royal family and the elite, and the curriculum included five canonical Persian literary texts, the pānč kitāb. When formal education was introduced in Chitral in 1940, Persian was a required part of the curriculum until about 1950, when it became elective. During the fighting of the 1990s in Afghanistan, a large number of refugees, mainly from Panjshir and Badakhshan, were living in Chitral, and familiarity with spoken Persian was partially revived, in that a significant number of local people in Chitral town learned some Dari Persian during dealings with the refugees, and a few new words entered Khowar, e.g., taxtabandi “arrangement in hotels for sitting on a raised platform to eat.” A dialect of Persian is still spoken in the Madaglasht valley in Chitral (Lorimer, 1922).

The first preserved literature from Chitral dates back to the17th century. Poets wrote first in Persian, then in a mixture of Persian and Khowar. Important among these are: Ataleγ Muhammad Šukūr Ḡarīb (late 17th century), Muhammad Sayyar (Bābā Sayyar, late 18th- to mid-19th century), Tajammal Šāh Maḥvī, and Muazzam Xān Lāl. The Persian diwan of Bābā Sayyar, consisting of 224 ḡazāls, four ḥamd and naʿt, four moḵ-ammas and eight robāʿi, has recently been translated into Urdu by Maula Nigah Nigah, a prominent Chitrali poet (Sayyar, 2004). The poetry of Tajammal Šāh Maḥvī has been compiled and translated into Urdu by Muhammad Irfan Chitrari, another well-known poet (Maḥvī, 2003). Muhammad Šukūr Ḡarīb’s diwan, as yet untranslated, includes poetry in Persian, Khowar, and a mixed style, exemplified in this line, quoted by Inayatullah Faizi: gurzéna mo angyé hargiz zāγ o zāγdān rā “Don’t ever bring crows and owls into (my) garden,” in which gurzéna mo angyé “don’t bring into the garden!” is Khowar and hargiz zāγ o zāγdān rā “ever crows and owls” is Persian. Some contemporary poets also employ this mixed style.

Persian words and phrases have been integrated into the spoken language at various stages. Some earlier borrowings reflect phonological and/or semantic changes, e.g., Pers. palatal > Khowar retroflex sibilants and affricates: Kho. ṧéγun “liver” (cf. Pers. šugūn/šugŭn “auspicious (omen)”; Kho. niṧán “gift” (cf. Pers. nišān “sign”); Kho. daṧmán “mullah” (< Pers. dānišmand “wise”; contrast the later borrowing dúšman “enemy” from Urdu); Kho. čˊhīr “milk” (cf. Pers. šīr “id.”); Kho. čˊhoi “six” (cf. Pers. šaš, and Urdu čhe). Semantic doublets are found, e.g., xošani “festivity, marriage, happiness” (Pers. k “happy”) and šādi “marriage” (< Ur. šādī “marriage” < Pers. šādī “festivity, marriage”). Repeated borrowing of the same element at different stages also results in doublets, e.g., we-sóoru “widow” (lit. “lacking a head”), an earlier borrowing, and be-talím “uneducated,” a recent Urdu loan. Some entire phrases have entered the spoken language, e.g., anč-e-bayād “as much as necessary”: ma brār ma-sum anč-e-bayād madát areér “My brother helped me as much as was needed.” The Persian names for days of the week are used in Khowar, as in Balochi and Brahui.

Grammatical influences on Khowar include subordinate clauses introduced by ki, the eẓāfa construction, and a spreading use of the Persian (animate) plural marker -ān. Direct case plurals in -án (< Pers.), originally used with Persian words denoting animate beings, e.g., buzurg-án “elders,” are spreading to native words, e.g., Ḍaq-án “boys,” replacing the older, unmarked direct plural. Relative clauses employing ki (< Pers. and Ur.), as opposed to indigenous constructions employing sentence-initial finite clauses, participial relative clauses, or agent nouns, are increasingly used, especially in writing. One of the most common types of these relative structures is kya [Noun Phrase] ki ... “the [NP] which...”, followed by a demonstrative pronoun in the matrix clause. For example: RC[kya wólTi-áar ki gáan ošói] haté wólTi čˊhamúT-o haté hisá ḍheéir “The part of the finger on the side RC[from which the wind is (coming)] feels cold.” Kalasha, whose speakers are almost all bilingual in Khowar and many in Urdu, is acquiring ki clauses indirectly, through Khowar (and Urdu). For example: RC[kúra ki khú~ḍiái áyis] šasé bačˊáas čhu “The one RC[who called you] was a king’s daughter.”

The status of /f/, /x/, /γ/, /q/ and /z/ in Khowar, where they belong to the basic phonological inventory of the language, is unique. This may reflect an earlier stratum of Iranian/Persian influence, but may also result from substratum influence from Burushaski. Notably, they are also part of the basic Burushaski and Wakhi inventories, with the status of /q/ being marginal in Wakhi. Of these, only /z/ is shared by Kalasha, which has not acquired /f/, /x/, /γ/, or /q/ from Khowar.

Burushaski has numerous Perso-Arabic loans, acquired first through Khowar and later through Urdu. In Yasin Burushaski (YB), this process has reached the point where, in the language of some young speakers, more than 50 percent of the lexical items used are Urdu borrowings (Patry and Tiffou, p. 9). Semantic and phonological changes are observed: /ḥ/ > /q/ (> /k/), e.g., Hunza Burushaski (HB) aqmaq “bloom of youth; drunkenness” (Berger, III, p. 20; cf. Kho. aqmaq/aḥmaq “fool”; also > Kal. hakmák “fool,” < Pers. aḥmaq “fool”; /x/ > /qh/, e.g., HB qharáp “bad” (Berger, III, p. 353; < Ur. xarāb “id.”); /h/ > /x/, e.g., Kho. ambóx “much” < Pers. anbūh/anbŭh “large, full” (Steingass, p. 104), also YB amboh “much” (Lorimer, 1962, p. 14); Pers./Ur. /f/ > /ph/, e.g., YB phiryát “request, plea” (cf. Ur. faryād “id.”). Epenthetic consonants develop, e.g., YB ambrōz “kind of pear” (Lorimer, 1962, p. 14), Kho. ambróz “id.” < Pers. amrūd “pear, guava” (Steingass, p. 100). Yasin Burushaski and Khowar share some of the (early) phonological changes affecting Persian loans; for example, YB ṧūra “saline material” with the retroflex sibilant (cf. Kho. ṧóor “[too] salty” vs. Ur. šor n. “salinity,” adj. “brackish”) indicates borrowing from Khowar rather than from Urdu. Subordinating constructions with the complementizer ke “that, when,” an influence from Khowar and Urdu, are increasingly frequent. For example, [je girámar ke in] ními “[When I came to the village], he left” (Berger, I, p.191).

Wakhi, spoken in Pakistan, Tajikistan, and Afghanistan, is subject to increasing influence from Tajiki Persian in Afghanistan and Tajikistan and from Urdu in Pakistan. Pakistani Wakhi shows two strata of Persian influence, a waning presence of former direct contact, and an increasing indirect influx of Perso-Arabic words through Urdu. Until about 50 years ago, the majority of the male population of Gilgit is said to have been able to speak Persian (Reinhold, 1992), and some words and phrases used in Pakistani Wakhi still reflect the Persian rather than the Urdu pronunciation of the “long a” sound, e.g., əz sar təpâ “from head to foot.”

Wakhi has a long history of close contact with Khowar (Bashir, 2001), and they share some Eastern Iranian developments of early Persian loans, like the change of palatal sibilants to retroflexes, e.g., Kho. ṧałmú “turnip” and Wakhi ṧolm “id.” (cf. Pers. šalam/šalγam “id.” > Ur. šalγam “id.”); Kho. ṧoṧp “type of halwa” and Wakhi ṧuṧp “id.” (cf. Tajik Pers. šošp “id.”). The main native Wakhi subordination marker is tsə (pre-verbal)/tsəy (clause-final). However, Wakhi has also borrowed Tajik Persian ki, which, like tsə, introduces a wide range of subordinate clauses, including relative clauses, various adverbial clauses, and complements of verbs of cognition. For example, yem xun-i ha-yá halg-ev-en RC[kumd-ar ki sak-e ҳi δegit δetk] “This house belongs to the people [to whom we have given our daughter]” (Bashir, forthcoming).

Pashto too has incorporated a large number of Perso-Arabic borrowings, the earlier of which have undergone change of palatal sibilants to retroflex and palatal affricates to dental: e.g., duṧmán “enemy” (Pers. dušman “id”); doJák “hell” (cf. Parthian dōžak “id.”); tsarx “circle” (Pers. čarḵ- “id.”); dzigár “liver” (Pers. jigar “id.”). Later loanwords retain the palatals, e.g., jism “body” (Pers./Ur. jism “id.”; Elfenbein, pp. 757-58). Some borrowed elements have been morphologically verbalized with the indigenous infinitive ending -edal, e.g., šarmedal “to blush, be modest, etc.” (< Pers. šarm “bashfulness, modesty, shame” [Steingass, p. 742]). However, Pashto has adopted fewer Perso-Urdu grammatical features than other languages of the area. For example, it forms its relative clauses using the Pashto conjunction če, rather than the Persian/Urdu ki. The gender assignment of Arabic words reaching Pashto through Persian and/or Urdu has been restructured according to the dominant Pashto pattern, in which consonant-final nouns are masculine and vowel-final nouns are feminine. Thus e.g. zanjīr “chain” is feminine in Urdu, but dzandzir “id.” is masculine in Pashto; darwāza “door” is masculine in Urdu, but darwaza “id.” is feminine in Pashto; ḥālat “condition” is feminine in Urdu, but masculine in Pashto; iżāfa “increase” is masculine in Urdu, but feminine in Pashto.



A. H. Akbar, Urdū aur šinā ke muštarak alfāẓ, Islamabad, 1992.

M. Alam, “The Pursuit of Persian: Language in Mughal Politics,” Modern Asian Studies 32/2, 1998, pp. 317-49.

Idem, “Persian in Precolonial Hindustan,” in Literary Cultures in History: Reconstructions from South Asia, ed. Sheldon Pollock, Berkeley, 2003, pp. 131-98.

J. Baart, A Sketch of Kalam Kohistani Grammar, Islamabad, 1999.

E. Bashir, A Contrastive Analysis of Brahui and Urdu, Peshawar and Washington, D.C., 1991a.

Idem, A Contrastive Analysis of Balochi and Urdu. Peshawar and Washington, D.C., 1991b.

Idem, “Khowar-Wakhi Contact Relationships,” in Toḥfa-e-Dil. Festschrift Helmut Nespital, ed. D. W. Lönne, Reinbek, Germany, 2001.

Idem, “Wakhi,” in The Iranian Languages, ed. Gernot Windfuhr, Surrey, UK (forthcoming).

H. W. Bellew, A Dictionary of the Pukkhto or Pukshto Language, 2nd ed., Lahore, 1901; repr., Peshawar, 1982.

H. Berger, Die Burushaski-Sprache von Hunza und Nager, Teil I, Grammatik, Teil II, Texte, Teil III, Wörterbuch, Wiesbaden, 1998.

B. M. Bukhari, Urdū aur Khowār ke lisānī rawābiṭ, Islamabad, 2003.

J. Elfenbein, “Pashto Phonology,” in Phonologies of Asia and Africa (including the Caucasus), ed. A. S. Kaye, Winona Lake, Ind., 1997, II, pp. 733-60.

Inamullah, Digital Dictionary of Torwali (to appear on the website of the Digital Dictionaries of South Asia Project, at, in preparation.

C. Jahani, Standardization and Orthography in the Balochi Language, Acta Universitatis Upsaliensis, Studia Iranica Upsaliensia 1, Uppsala,1989.

Idem, “The Formal Structure of Gul Khān Nasir’s Poetry,” Orientalia Suecana 43-74, 1994-95, pp. 141-47.

Idem, “Persian Influence on Some Verbal Constructions in Iranian Balochi,” Stud. Ir. 28, 1999, pp.123-43.

Pareshan Khattak, Purdil Khattak, and M. Ishaq, ATūT lisānī rawābit, Peshawar, 1977.

D. L. R. Lorimer, The Phonology of the Bakhtiari, Badakhshani, and Madaglashti Dialects of Modern Persian, withVocabularies, London, 1922.

Idem, Werchikwar English Vocabulary, Oslo, 1962.

Tajammal Šāh Maḥvī, Kulliyāt Maḥvī, ed. and tr.Muhammad IrfanChitrari, Chitral, n.d.,

C. P. Masica, Defining a Linguistic Area: South Asia, Chicago, 1976.

G. Morgenstierne, “Iranian Elements in Khowar,” BSOAS 8, 1936, pp. 657-71.

Y.-C. Morin and L. Dagenais, “Les emprunts Ourdous en Bourouchaski,” JA 265, 1977, pp. 307-43.

R. Patry and É.Tiffou, “Les emprunts lexicaux à l’ourdou en bourouchaski du Yasin: un phénomène qui varie selon l’âge,” Communication au congrès New Wave, Québec, 1997.

J. R. Perry, Form and Meaning in Persian Vocabulary: The Arabic Feminine Ending, Costa Mesa, and New York,1991.

T. Rahman, Language, Ideology and Power: Language Learning Among the Muslims of Pakistan and North India, Karachi, 2002.

B. Reinhold, “The Incorporation of Urdu and Persian into the Wakhi of the Karakorams,” unpubl. ms., 1992.

A. V. Rossi, Iranian Lexical Elements in Brāhuī, Naples,1979.

M. (Bābā) Sayyār, Dewān Bābā Sayyar, tr. and comm.Maula Nigah, Islamabad, 2004.

F. Steingass, A Comprehensive Persian-English Dictionary, London, 1892; repr., Lahore, 1981.

(Elena Bashir)

Originally Published: November 15, 2006

Last Updated: November 15, 2006