YAGHNOBI, term used as a noun to refer to a people and as an adjective to designate their language (yaγnobīˊ zivók “Yaghnobi language”). It is derived from the name of the Yaghnob valley and the Yaghnob river, which, according to Khromov (1987, p. 644), is a Tajik adaption of the Yaghnobi phrase ix-i nou “ice valley or ravine.” The more traditional interpretation is “ice river” (Benveniste, 1955, p. 139, note 1; Andreev and Peshchereva, 1957, p. 365). Yaghnobi is a Modern East Iranian language and the only surviving successor of a Sogdian dialect. It is not a written language, but is used for daily family communication. The speakers are Sunni Muslims. When the language was discovered in the 19th century, it was spoken in the central part of the remote high valley of the Yaghnob river. This valley is situated in Tajikistan, about 100 km north of Dushanbe. Over a distance of 120 km, the Yaghnob river runs east-west, parallel to the Zerafshan river, between the Hissar range in the south and the Zerafshan range in the north. The Yaghnob river eventually turns north and meets the Zerafshan river at Aini.

Toponomastic evidence suggests that, in the past, Yaghnobi and other Sogdian dialects were spoken in neighboring valleys as well. In 1913, Junker counted about 2,200 native speakers of Yaghnobi in 21 settlements (Junker, 1930, pp. 114-21). According to Khromov (1987, p. 644), there were about 2,500 speakers in the 1960s; 1,500 of these native speakers were living in 22 settlements within the Yaghnob valley, and roughly 900 resided outside of the valley. In the 17th century, Yaghnobi speakers migrated to the neighboring valley of the Varzob river, and in the early 1950s people also moved to the Hissar valley and to Dushanbe. In 1970, however, the Soviet authorities forced the whole population of about 3,000 people to leave the Yaghnob valley. Five hundred families migrated to Zafarobod in northern Tajikistan, 200 families to the area surrounding Dushanbe. Only about 300 people were living in the Yaghnob valley in 1990, when the Dushanbe-based Council of Ministers passed the decision to re-establish all villages whose populations had been resettled. Besides further measures such as the re-opening of schools, the Tajik Academy of Science was asked to support the preservation of the Yaghnobi language (for more details see Badenkov et al., 1994, pp. 476-84; Gunja, 1996, pp. 68-69).

The Russian scholar Alexander L. Kuhn and his Tajik companion and interpreter Mirza Mulla Abdurrakhman from Samarkand were, during their Iskanderkul expedition of 1870, the first who made recordings of the Yaghnobi language. The linguistic material collected by these and other scholars was used by C. Salemann in his unpublished Yaghnobi studies (Yagnobskie etyudy), on which W. Geiger based his description “Über das Yaghnōbī” (Geiger 1898-1901; see Oranskij, 1975, I, pp. 114-20, and Khromov, 1987, p. 648, for more details on the history of these early expeditions, including their research and publications). The scientific community’s interest in the study of Yaghnobi grew considerably when, at the beginning of the 20th century, documents in a previously unknown Eastern Middle Iranian language were found. The language in these documents was identified as Sogdian, which, as Salemann was able to show, is closely related dialectologically to Yaghnobi (see Oranskij, 1975, I, pp. 115-16). However, we know today that Yaghnobi is not a direct descendant of the linguistically rather uniform variety of Sogdian used in these texts (see Sims-Williams, 1989, p. 173), but that its origins can probably traced back to a Sogdian dialect spoken in Osrushana (Khromov, 1987, p. 645).

Due to the close linguistic connections between Yaghnobi and Sogdian, leading scholars in the exploration of Sogdian like R. Gauthiot and E. Benveniste took up the study of Yaghnobi. Thus, in 1955, Benveniste published a glossary from Salemann’s Yaghnobi Studies on the basis of a copy made by Gauthiot (Redard, 1970, p. 102). Unfortunately, the glossary is cut short in the middle of the letter k. Salemann’s Yaghnobi Studies and Geiger’s description were at R. Gauthiot and H. Junker’s disposal during their 1913 expedition to the Yaghnob Valley (see Junker, 1930, pp. 3-4, 107 with note 1). The resulting two publications by Junker (1914, 1930) represented an important step forward in the research on Yaghnobi. To date, Andreev and Peshchereva (1957) contains the largest published collection of oral folk literature and the most exhaustive glossary with references to, among other, Sogdian, Ossetic, and the Pamir languages. V. S. Sokolova’s research on the phonetics of Yaghnobi (Sokolova, 1953) is based on recordings she made in the Varzob Valley in 1949. An overall description of the language was published by Bogolyubov in 1966. The most comprehensive description of Yaghnobi is Khromov (1972). It not only contains a detailed grammar (phonetics, morphology, syntax, lexicon, and word formation) including chapters on dialectal differences, bilingualism, and Yaghnobi-Sogdian dialectal relations, but also 33 texts with Russian translations and a supplementary vocabulary to Andreev and Peshchereva (1957). A newer grammatical sketch of Yaghnobi is Khromov (1987), in which the chapter on the Yaghnobi-Sogdian dialectal relations of Khromov (1972) is replaced by a systematic description of the sound correspondences between Yaghnobi and Proto-Iranian and vice versa. Finally, a short sketch of Yaghnobi is given by the present author in the Compendium Linguarum Iranicarum (Bielmeier, 1989).

Salemann found out that Yaghnobi is divided into a western and a eastern dialect variety (see Khromov, 1972, p. 97 with note 99). This view is shared by Geiger 1898-1901 and has become the opinio communis, even though Gauthiot spoke of, and Khromov later confirmed, the existence of transitional varieties between the two poles. Generally, the western variety shows the more conservative features, e.g.: (1) western ay vs. eastern e (wayš/weš “grass”); (2) t vs. s < (met/mes “day”); (3) tr- vs. sr- < *θr- (tiráy/saráy “3”); (4) verbal ending 3rd sg. present tense -tišt vs. -či (kúntišt/kúnči “he makes”); (5) verbal ending 1st pl. preterite tense -om vs. -im (akunóm/akuním “we made”); (6) tiráš- vs. diví- “to fall,” rūn vs. rṹnak “lamb,” etc. (see Junker, 1930, pp. 123-29; Khromov, 1972, pp. 97-105; Khromov, 1987, pp. 647-48).

In the vowel system Sokolova (1953, pp. 64-70) distinguishes 8 vowels on the phonemic level, five phonetically long vowels ē, ō, ī, ū, ǖ, and three phonetically short vowels a, i, u. Therefore, vowel length is only phonemic with i vs. ī and u vs. ū (tir “go!” vs. tīr “arrow” or uxta “went down” vs. ūxta “brought”). It does not have to be noted with e, o and ü. The phonemic status of ü is unclear. It developed from ū under unclear conditions (cf. Livshits, 1962, p. 150; Bogolyubov, 1966, p. 344; Khromov, 1987, p. 655). Generally, there is quite a lot of variation in the realization of the vowels. Vowel length, for instance, is inconsistent; o and ū are in free variation under Tajik influence (kom beside kūm “which”); and ü is replaced by ī in the younger generation (xür > xīr “sun”) (Sokolova 1953, p. 63; for variation due to the speakers’ age, see also Junker, 1930, p. 124-25) in an environment where Andreev and Peshchereva (1957, p. 363) still note the more conservative form xūr “sun.”

The consonant inventory consists of p, t, k, b, d, g, v, γ, č, j, f, s, š, z, , w, y, m, n, r, l, x, , h, q, , ʿ. The consonant system is characterized by the phonemic opposition between unvoiced stops and voiced fricatives, the latter of which go back to voiced stops: p vs. v, k vs. γ, but t vs. d (< δ under Tajik influence?). On the synchronical axis, b, g, and j, which first appear in Yaghnobi mainly in loans, in word-initial position before vowels, can also be found in opposition to unvoiced stops and voiced fricatives, forming minimal pairs. Therefore, they have to be considered as phonemes as well. The phoneme l occurs mainly in loans, the phonemes h, q, , ʿ exclusively in loans. As there are no initial consonant clusters, the , which only occurs word-initially, has to be considered a phoneme (x°ar “eat!” vs. xar “donkey” or x°at “self” vs. xat(t) “ “writing” vs. wat “there”). The bilabial glide w is in phonemic opposition to v (wīr “man” vs. vīr “find!” or wov “speak!” vs. vow “come!”).

Stress is not phonemic. It usually falls on the last or on the penultimate syllable with a phonetic long vowel (cf. Sokolova, 1953, p. 64). If both syllables are long, the penultimate is stressed. If both are short, stress falls on the last syllable if it ends with a double consonant (kunánt “they may do,” farγúmč “heifer”) or if the first vowel is a Svarabhakti (epenthetic) vowel (siták “bone”). Secondarily suffixed morphemes usually are not taken into account, e.g., -išt (šawóm-išt “I go,” nīˊdom-išt “I sit down”). Tajik loans usually carry final stress, but sometimes an adaption to the pattern found in inherited words can be observed. The prefixed verbal negation is usually stressed (nánosomišt “I don“t take,” but γuš nakúntišt “he did not listen” from the Tajik calque γuš kun- “to listen”; see Khromov, 1972, p. 96). The prefixed augment is never stressed (ašáw “he went”). For stress patterns in compounds, see Khromov (1972, pp. 16, 92-93).

There is no grammatical expression of gender or of dual number. The plural is marked, in a manner similar to that of Sogdian, by a suffix -t (see Sims-Williams, 1989, p. 183), and a preceding -a changes to -o (x̌ṹta “son” > x̌ṹtot “sons”). A non-animated plural subject requires a plural predicate (yaw-t tim garíb aras-ór “the barley crops are almost fully ripe”). The case system is reduced to an unmarked absolute case and an oblique case, characterized by a suffixed unstressed -i,which is reduced to -y after vowel (eastern -ay > -e). The case marker -i follows the plural morpheme -t. There is group inflection with the noun phrase; the attribute precedes the head. The adjective remains unchanged, but used as a noun it is also formally treated like a noun. The preceding numeral ī “one” functions as indefinite article. The oblique case is mainly used to mark nouns functioning as attributes, indirect or definite direct objects, or as agents in ergative constructions with past tense verb forms (Khromov, 1987, pp. 663-64). In the comparative construction the compared member of the sentence is preceded by the preposition či, which governs the absolute or the oblique case.

The personal pronouns of 1st and 2nd person have an oblique case different from the absolute case only in the 2nd pers. sg. (tu vs. taw). The other forms man “I,” mox “we,” and šumóx “you (pl.)” are derived from earlier genitive forms and are now used for both cases. The forms of the 3rd pers. sg. ax, (obl. áwi, it) and pl. áxtit, íštit (obl. áwtiti, ítiti) are in fact forms of the demonstrative pronoun. The enclitic forms are: 1st pers. sg. -m, 2nd pers. sg. -t, 3rd pers. sg. , 1st pers. pl. -mox, 2nd and 3rd pers. pl. –šint. These are used like the full forms as possessive pronouns, to express the indirect and direct object, the agent in ergative constructions as well as the possessor in sentences with “to have” (ī γow-š ast “he has a cow”). Used as direct object, an enclitic form can be suffixed to a full pronominal form (mox-šint awénim “we saw you”). In the system of demonstrative pronouns, two degrees of proximity are distinguished: sg. , aníš (obl. it, anít), pl. íštit (obl. ítiti) “this,” ax, (obl. áw(i), a), pl. áxtit (obl. áwtiti) “that.” The interrogative pronouns, also used in indefinite function, differentiate between kax (obl. kay), pl. káxtit (obl. káytit) “who?” and čo (obl. čoy) “what, which?” In attributive function kom beside kūm (obl. kómi beside kṹmi), pl. kṹmtit (obl. kṹmtiti), or čo are used. The reflexive pronoun x°at, usually followed by an enclitic pronoun (x°át-im kunóm “I will do it myself”), is used as a noun. The reflexive pronoun xap/xep can also be used as an attribute (ax xep dásti asinóy “he washed his hands”).

The numerals are based on a decimal system and are inherited up to ten; beyond ten, they are loans from Tajiki. Beginning with “two,” the numerals are connected with the obl. sg. of the noun (tiráy γówi “three cows”). There are traces of a vigesimal system (see Bogolyubov, 1966, p. 347).

Yaghnobi has compound verbal stems consisting of noun and verb (ark kun- “to work”), like Tajiki, and simple verbal stems based on the old present tense stem (šaw- “to go”). Extended by the stressed morpheme -ón- an intransitive stem usually becomes transitive or causative (puxs- “to cook, ripen” > puxsón- “to cook, fry,” roy- “to cry” > royón- “to make [someone] cry”). Present and past tense as well as present subjunctive and imperative mood are derived from the stem by adding personal endings. Yaghnobi is the only Modern Iranian language where the augment is still kept to express past tense (šáwom-išt “I go” vs. a-šáwim “I went”). There are two basic sets of personal endings: sg. -im, -i, zero; pl. -om/-im, -ti/-si, -or, used in the simple past, and sg. -om, zero, -ot, pl. -im, -t/-s, -ant, used in the present subjunctive, where the 3rd sg. -ot is an old subjunctive form to be compared with Sogdian -āt < *-āti (for the historical development of the verbal endings in general, see Tedesco, 1923). The simple past is the usual tense in narratives. The suffix -išt (< *hišta-) adds a durative or iterative dimension. If that same suffix is added to the present subjunctive in either the 1st or the 2nd pers., the resulting form is a present tense. In the 3rd person, the endings of the present are sg. -t-išt or -či, pl. -ošt < -or-išt. In sentences with this type of predicate, the unmarked word order is SOV (man divár apénim “I opened the door”).

Periphrastic verb formations are based on the perfect participle (verbal stem + -ta or shortened -t). Together with the present tense forms of the copula (sg. īm, išt, ast(i) beside xast(i) beside -x, pl. om, ot/os, or),the perfect participle forms a resultative perfect; with the past tense forms (sg. oyim, oy(i), (x)oy, pl. iyom, iyot, iyor), a pluperfect tense; with the subjunctive of the copula, a subjunctive perfect tense; and with the perfect of the auxiliary vu-, it expresses secondhand information (Avázi xar nóta vúta-x “he apparently has taken Avaz’s donkey”). With these formations we find ergative constructions. If the predicate is based on an intransitive verb, the copula agrees with the subject; if it is based on a transitive verb, the copula agrees with the direct object and not with the agent. If the direct object is a 3rd person, the copula can be zero. The agent in the oblique case is often expressed by a noun or pronoun placed between participle and copula (púčta-t-x “you have pinched him”) or a pronoun placed between negation and verbal stem (divár na-š-péta-x “he has not opened the door yet”). Extended by the Tajik suffix -gi, the perfect participle can be used as predicate (kabü´d rūptagí “the green [grass] is cut”); without copula it can also take on attributive function (na-lakstagí odám “a person who did not get around at all”). The present participle with the copula or the auxiliary vu- functions as a predicate expressing the intention to carry out a certain action (nahíš γántum man rṹbna īm “I am going to cut this wheat,” čo ark kárna vot? “what work is he going to do?”); used with the 3rd pers. perfect forms of vu-, it includes secondhand information (peštár mórtit tim bozí kárna vútor “earlier, it is said, men played as well”). The so-called infinitive (verbal stem + -ak or simply verbal stem) in the oblique is used in final clauses (nūni pačáki wáxti-š víta-x “the time has come to bake bread”). The simple verbal stem in the oblique followed by vu- is used in ingressive clauses (bozí kári avór “they started to play”). Both infinitive forms can be followed by the modal verbs γaw- “to be necessary” or ton- “to be able” (ark kar ná-tonči “he is not able to work”).

Adverbs are not usually formally characterized. Many of them and most of the pre- and postpositions are loans from Tajiki. The enclitic conjunction -(i)k introduces not only relative clauses (har odám-k ... “every person who ...”) but also temporal, conditional, concessive, possessive, etc. clauses.

The nominal compounds comprise mostly determinative (with the first element modifying the second: Skt. Tatpuruṣa and Karmadhāraya) compounds but also adjectival, inverted possessive (inverted, Skt. Bahuvrīhi) compounds (rax-péta “whose mouth is opened” = “who stands gaping”). A kind of loose copular compound is found with enclitic -at “and” (bod-at-havo “wind and weather” = “climate”) or with the conjunction o “and” under Tajik influence (peš-o-peš “forward”). With derived nouns the unstressed suffix -(i)k is very common forming, e.g., possessive adjectives (iš kat mán-ik-x “this house is my one”). Relational adjectives are derived by -īˊna (dork “wood” > dorkīˊna “wooden”) and agent nouns by -akí (wayš “grass” > wayš-akí “he who carries grass”).

The lexicon, with the exception of verbs, is heavily influenced by Tajiki, through which also Turkish and Russian elements entered Yaghnobi. Older loans from Tajiki are phonetically adapted. The loans and calques are found even in the basic vocabulary. Sometimes we find the inherited word and the corresponding loan side by side (e.g., šowí from Tajik siyohí “darkness,” wáxin and Tajik xun “blood,” zoy and Tajik zamín “earth,” inč and Tajik zan “woman”). The so-called secret language is a cant known from certain social groups who use different words (árna instead of xar “donkey”), insert a semantically empty syllable between the syllables of the text, or change the semantics of words (see Junker, 1930, p. 125; Khromov, 1972, pp. 90-91, Khromov, 1976).


M. S. Andreev and E. M. Peshchereva, Yagnobskie Teksty (Yaghnobi Texts), Moskva and Leningrad, 1957.

J. Badenkov, A. Gunja, and P. Lindner, “Traditionelle Wirtschaftsweise und Strukturwandel in einem peripheren Gebirgsraum am Beispiel Jagnob/Tadschikistan,” in Mitteilungen der Fränkischen Geographischen Gesellschaft 41, 1994, pp. 465-87.

E. Benveniste, “Un lexique du yagnobi,” JA 243, 1955, pp. 139-62.

R. Bielmeier, “Yaghnōbī,” in R. Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 480-88.

M. N. Bogolyubov, “Yagnobskiĭ yazyk” (Yagnobi language), in Yazyki narodov SSSR I: Indoyevropeĭskie yazyki (Languages of the people of the USSR I: Indo-European languages), Moskva, 1966, pp. 342-61.

W. Geiger, “I. Die Pāmir-Dialekte, Anhang: Über das Yaghnōbī,” in Grundriss der Iranischen Philologie I, 2. Abteilung, VIII. Kleinere Dialekte und Dialektgruppen, Strassburg, 1898-1901, pp. 288-344.

A. N. Gunja, “Ressourcen, Risiken und Möglichkeiten der nachhaltigen Nutzung im Jagnobtal /Tadschikistan,” in Petermanns Geographische Mitteilungen 140, 1996/2, pp. 67-94.

H. F. Junker, “Drei Erzählungen auf Yaγnåbī,” in Sitzungsberichte der Heidelberger Akademie der Wissenschaften, phil.-hist. Klasse, 1914, 14.

Idem, Arische Forschungen, Yaghnōbī-Studien, I. Die sprachgeographische Gliederung des Yaghnōb-Tales, Leipzig, 1930.

A. L. Khromov, Yagnobskiĭ yazyk (Yagnobi language), Moskva, 1972.

Idem, “Uslovnye yazyki u iranskikh narodov” (Secret languages of the Iranian people), Vostochnaya filologiya (Eastern Philology) 4, Dushanbe 1976, pp. 3-19.

Idem, “Yagnobskiĭ yazyk” (Yagnobi language), in Osnovy iranskogo yazykoznaniya, Novoiranskie yazyki: vostochnaya gruppa (Essentials of Iranian linguistics, New Iranian languages: eastern group), Moska, 1987, pp. 644-701.

V. A. Livshits, “Yagnobskiĭ yazyk” (Yagnobi language), in Narody Sredneĭ Azii i Kazakhstana (People of Middle Asia and Kazakhstan) I, Moskva, 1962, pp. 149-52.

I. M. Oranskij, Die neuiranischen Sprachen der Sowjetunion I-II, The Hague and Paris, 1975.

G. Redard, “Other Iranian Languages,” in Current Trends in Linguistics 6, The Hague and Paris, 1970, pp. 97-135.

N. Sims-Williams, “Sogdian,” in R. Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 173-92.

V. S. Sokolova, “Yagnobskiĭ yazyk” (Yagnobi language), in V. S. Sokolova, Ocherki po fonetike iranskikh yazykov (Sketches of the phonetics of the Iranian languages), Moskva and Leningrad, 1953, pp. 59-79.

P. Tedesco, “a-Stämme und aya-Stämme im Iranischen,” Zeitschrift für Indologie und Iranistik 2, 1923, pp. 281-315.


(Roland Bielmeier)

Originally Published: August 15, 2006

Last Updated: August 15, 2006

Cite this entry:

Roland Bielmeier, “YAGHNOBI,” Encyclopædia Iranica, online edition, 2006, available at http://www.iranicaonline.org/articles/yaghnobi (accessed on 20 September 2016).