The language of ancient Gandhāra, the area around the Peshawar Valley in the modern North-West Frontier Province of Pakistan, lying near the border of the Indian and Iranian linguistic areas.


GĀNDHĀRĪ LANGUAGE, the language of ancient Gandhāra (q.v.), the area around the Peshawar Valley in the modern North-West Frontier Province of Pakistan, lying near the border of the Indian and Iranian linguistic areas.

General. Gāndhārī belongs to the Middle Indo-Aryan (MIA) family of Indian languages and is closely related to Sanskrit, Pali, and various Prakrit dialects. Between the third century BCE and third century CE. Gāndhārī served as the literary language and lingua franca of the northwestern part of the Indian Subcontinent. Under the Kuṣāṇa Empire (first to third centuries CE; see KUSHAN DYNASTY), Gāndhārī spread into adjoining regions of India, Afghanistan, and Central Asia. Recent discoveries of large numbers of manuscripts in Gāndhārī have shown that during this period Gāndhārī was a major Buddhist literary language. The evidence of historical phonology indicates that some of the earliest Chinese translations of Buddhist texts are derived from Gāndhārī archetypes.

Gāndhārī was written in the Kharoṣṭhī script, which is derived from the Aramaic script used in the eastern parts of the Achaemenid Empire, including Gandhāra. This sets Gāndhārī off from all other Indo-Aryan languages, which are written in Brāhmī (q.v.) script and its local derivatives.

Attestation. The Gāndhārī language is attested by four classes of documents:

(1) Over five hundred inscriptions, mainly Buddhist dedicatory or donative records on stone or metal. (Catalogue of Kharoṣṭhī Inscriptions; Konow; Salomon, 1998, pp. 74-79)

(2) Buddhist manuscripts on birch bark scrolls or palm leaf poṭhīs. Only a few of these have been published, but several dozen are currently under study. (Allon; Brough; Glass; Lenz; Salomon, 1999, 2000, 2003)

(3) Legends on coins of the Indo-Greek (q.v.), Saka, Parthian, and Kuṣāṇa rulers of northwestern India, often with bilingual legends in Greek and Gāndhārī. Such coins led to the initial decipherment of Kharoṣṭhī script by James Prinsep and others in the 1830s. (Salomon, 1998, pp. 209-18)

(4) Nearly one thousand legal and administrative documents on wooden tablets from the kingdom of Kroraina (Shan-shan) in the southeastern Tarim Basin (Xinjiang, China). They are composed in a distinct dialect of Gāndhārī with innovative morphology and loan words from Greek, Iranian, and Central Asian languages. (Boyer, Rapson and Senart; Burrow, 1937; Burrow, 1940)

Structure. Gāndhārī phonology generally resembles that of the other MIA languages. The vowels ai, au, anddisappear, and aya and ava are contracted to e and o respectively. In later stages of the language, most single intervocalic consonants are voiced or elided. A feature peculiar to Gāndhārī is the development of intervocalic -th- and -dh- into -s-, probably pronounced /z/. The three sibilants of Sanskrit, ś, , and s, which merge in other MIA dialects, are mostly preserved in Gāndhārī, though some sibilant clusters undergo special developments, such as śr > ṣ and ṣy > ś (e.g., Sanskrit manuṣya- ‘human’> maṇośa-). Consonant clusters are generally simplified by assimilation as in MIA; e.g., Sanskrit sapta ‘seven’ > sata (graphic for /satta/). But some consonant clusters which are assimilated in other MIA dialects are preserved in Gāndhārī, especially those involving r and v; thus Sanskrit prasanna- ‘pleased’ > Gāndhārī prasaṇ[ṇ]a, but Pali pasanna-. Clusters with r are often subject to metathesis, as in Sanskrit durgati- ‘bad destiny’ > Gāndhārī drugadi-, vs. Pali duggati. Sanskrit kṣa, which elsewhere becomes kh or ch, is retained and represented by a special Kharoṣṭhī character.

Gāndhārī morphology is likewise similar to that of other MIA languages, but more flexible and less standardized. For example, the endings of the masculine and neuter nominative singular of noun stems in -a varies among -e, -o, -u, or -a, even within the same text, as a result of the neutralization of vowels in word final position with consequent graphic ambiguity (Fussman, pp. 460-61, 471-73). Similarly, the locative singular endings -e, -u, -o, -a, -mi, and śpi alternate freely. Among verb forms, the future stem is -iśa- (< Sanskrit -iṣya-). Preterite verbs are expressed either by derivatives of old aorists such as adhrikṣe ‘I saw’ (compare Sanskrit adrākṣam) or by periphrastic constructions with the past participle, as in aho . . . ṇidiṭhu ‘I was designated’ (= Sanskrit ahaṃ nirdiṣṭaḥ).

The lexicon of Gāndhārī is primarily Indic in origin and largely common to other Indo-Aryan languages, but it includes some words characteristic of the northwestern dialects, such as śpasa ‘sister’ instead of bhaginī and baṭa ‘stone’ instead of pāṣāṇa (Salomon, 1999, pp. 133-34). Derivatives of such regional vocabulary are sometimes found in the modern Dardic languages (see DARDESTĀN ii.), such as Torwali bāṭ ‘stone’. Besides the Indic component, Gāndhārī has some loanwords, particularly from Greek and Iranian. Greek loans involve administrative terms such as stratega ‘general, commander’ (< strategós [στρατηγóς]; Salomon, 1999, pp. 141, 148), meridarkha- ‘meridarch’ (< meridárchē [εριδαρχη]; Konow, p. 2), and sa(dera) ‘stater’ (< statḗr [στατηp]; Salomon, 1999, p. 148), and calendrical terms, especially Macedonian month names, for example avadunaka- ‘Audunaios’ (αυδυναιος; Konow, p. 154). Iranian loanwords also typically occur in the administrative and calendrical sphere; for example, kṣatrapa ‘satrap’ (< Old Persian xšaçapāvan-; Salomon, 1999, pp. 142-44), erzuna ‘prince’ (< Saka alysānai/eysānai; Konow, p. 61), and kṣuṇa ‘date’ (= Khotan Saka kṣuṇa; Konow, p. lxxiv).

Historical development. Gāndhārī developed in three stages. Early Gāndhārī is best attested in the sets of Aśoka’s major rock edicts at Shāhbāzgaṛhī and Mānsehrā. At this stage, intervocalic consonants were mostly retained as in the original Old Indo-Aryan form; for example, siyati (later siyadi)= Sanskrit syāt ‘would be’ (Shāhbāzgaṛhī XII.8). In the middle stage, found in inscriptions and manuscripts from the first century BCE to the middle of the second century CE, intervocalic consonants are voiced, elided, or modified to fricatives (Fussman, pp. 455-65). But in late Gāndhārī of the later second and early third centuries CE, the natural phonological developments are masked by extensive re-Sanskritization of the written language, whereby many consonants which had changed or disappeared in the spoken language were restored to their underlying Old Indo-Aryan form; for example, sapta ‘seven’ = Sanskrit sapta instead of earlier sat[t]a (Salomon 2001, p. 245). Some late documents written in Kharoṣṭhī script are in fact practically indistinguishable from Sanskrit (Salomon, 2001, p. 246).

Literature. Buddhist literature attested in Gāndhārī manuscripts comprises a wide range of genres, including both original Gāndhārī compositions and texts translated from other MIA languages. The best represented genre is sūtra (Allon, Salomon, 2000), including a collection of some two dozen scrolls constituting an anthology of sūtras (Salomon, 2003). Other important genres include avadānas (legends; Lenz, part 2), abhidharma (scholastic treatises), commentaries, and stotras (hymns). The Dharmapada is extant in two manuscripts (Brough; Lenz, part 1). Gāndhārī literature as known to date consists mainly of texts of “mainstream” Buddhist schools such as the Dharmaguptakas, but there is at least one instance of a Mahāyānistic text, the Bhadrakalpika-sūtra.


Bibliography (websites were accessed July 14, 2008):

Reference materials.

Bibliography of Gandhāran Studies, comp. A. Glass and S. Baums, available online (accessed 28 July 2008).

Early Buddhist Manuscripts Project, database of Gandhārī words,comp. A. Glass and S. Baums, available online (accessed 28 July 2008).

Gāndhārī language.

Gérard Fussman, “Gāndhārī écrite, Gāndhārī parlée,” in Dialectes dans les littératures indo-aryennes, ed. Colette Caillat, Paris, 1989, pp. 433-501.

Richard Salomon, “‘Gāndhārī Hybrid Sanskrit’: New Sources for the Study of the Sanskritization of Buddhist Literature,” Indo-Iranian Journal 44, 2001, pp. 241-52.

Richard Salomon, “Gāndhārī and the Other Indo-Aryan Languages in the Light of Newly-discovered Kharoṣṭhī Manuscripts,” in Indo-Iranian Languages and Peoples, ed.N. Sims-Williams, Proceedings of the British Academy116, Oxford, 2002, pp. 119-34.


Catalogue of Kharoṣṭhī Inscriptions, available online (accessed 28 July 2008).

Sten Konow, Kharoshṭhī Inscriptions with the Exception of Those of Aśoka, Corpus Inscriptionum Indicarum 2.1, Calcutta, 1929.

Richard Salomon, Indian Epigraphy: A Guide to the Study of Inscriptions in Sanskrit, Prakrit, and the Other Indo-Aryan Languages, South Asia Research, New York, 1998.

Manuscript texts.

Mark Allon, Three Gāndhārī Ekottarikāgama-Type Sūtras: British Library Kharoṣṭhī Fragments 12 and 14,Gandhāran Buddhist Texts 2, Seattle, 2001.

John Brough, The Gāndhārī Dharmapada, London, 1962.

Andrew Glass, Four Gāndhārī Saṃyuktāgama Sūtras: Senior Kharoṣṭhī Fragment 5, Gandhāran Buddhist Texts 4, Seattle, 2007.

Timothy Lenz, A New Version of the Gāndhārī Dharmapada and a Collection of Previous-Birth Stories:British Library Kharoṣṭhī Fragments 16 + 25,Gandhāran Buddhist Texts 3, Seattle, 2003.

Richard Salomon, Ancient Buddhist Scrolls from Gandhāra: The British Library Kharoṣṭhī Fragments, Seattle and London, 1999.

Idem, A Gāndhārī Version of the Rhinoceros Sūtra: British Library Kharoṣṭhī Fragment 5B, Gandhāran Buddhist Texts 1, Seattle, 2000.

Idem, “The Senior Manuscripts: Another Collection of Gandhāran Buddhist Scrolls,” Journal of the American Oriental Society 123, 2003, pp. 73-92.

Central Asian documents.

A. M. Boyer, E. J. Rapson, and E. Senart, Kharoṣṭhī Inscriptions discovered by Sir Aurel Stein in Chinese Turkestan, 3 parts, Oxford, 1920-29.

T. Burrow, The Language of the Kharoṣṭhi Documents from Chinese Turkestan, Cambridge, 1937.

Idem, Translation of the Kharoṣṭhi Documents from Chinese Turkestan, London, 1940.

July 28, 2008

(Richard Salomon)

Originally Published: July 28, 2008

Last Updated: July 28, 2008