According to Ehsan Yarshater’s informants, the Jewish community had dwindled from around 13,000 souls in 1920 to less than 1,000 by 1969, and of these about half originated from the Jewish communities of Malāyer, Tuyserkān, and various points in Kurdistan. The Jewish population lived mostly in the Darb-e Kalim-ḵāna quarter of Hamadān.




Introduction. The dialect spoken by the Jews of Hamadān (henceforth HJ) and a close variant spoken by the Jews of Tuyserkān (TJ; see Affiliations and Variantsbelow) belong to the Central Plateau Dialect (CPD) group of Northwestern Iranian languages (NWI), as opposed to Southwestern Iranian (SWI; e.g., Persian). The sources used for this description are abbreviated as follows: Abrahamian (AB), and Yarshater (YS).

Population and community. According to Ehsan Yarshater’s informants, the Jewish community had dwindled from around 13,000 souls in 1920 to less than 1,000 by 1969, and of these about half originated from the Jewish communities of Malāyer, Tuyserkān, and various points in Kurdistan. The Jewish population lived mostly in the Darb-e Kalim-ḵāna quarter of Hamadān. Haideh Sahim reports that in the mid-1970s the community numbered only about 350 (Sahim, p. 173, quoting an informant). According to members of the community that Donald Stilo encountered in 2001-02, there were only eight people from the Jewish community left in Hamadān at the time, but others can still be found in Israel, New York City, and most predominantly in Los Angeles. He was also informed that only people born before the mid-1940s were raised speaking the dialect. Even Yarshater’s informants claimed, in 1969, that use of the language in the home was dwindling. Hence it is not known how many speakers are left and whether there are any full native speakers among the generation under the age of fifty. It is likely that the Jewish community of Tuyserkān is now extinct, as Yarshater informants reported that in 1969 there were only two Jewish families left and they were at the time planning to leave.

Diachronics. The following list details the major NWI phonological developments from proto-Iranian (listed with *) present in the Jewish dialect of Hamadān, but examples typical of SWI that are also found in this dialect due to heavy borrowing from Persian are listed alongside in parentheses: *ts > s as in masar, kasar (= mas, kas + tar) “large, small” (but dah “ten”); *dz > z, as in zun/zunā “know,” heze “yesterday” (but dumād “son-in-law” del “heart”); *tsv > sp~sb as in esbid “white”; *dzv> zb~zv in zowun “tongue”; *dv- > b-, as in bar “door.” Later changes consist of: *fr- > h- in he- “preverb (< *frā)” in he-geft “take,” he-dā “give,” he-ništ “sit” (but note the SWI-like fr-, in HJ fraš/frat- “sell”); *θr > r in pir “son” (but se “three”);*w- > v- (instead of b-) in vāyom “almond,” vece “child,” vārun “rain,” vāy “wind, breeze,” ver- “preverb” (usually indicating “up”), (but bāz(-am) “again,” barre “lamb”); *w- > v- (instead of g-) in veider~ veidešd “pass” (but gorg “wolf”); *j- remains as j- in jir “down,” jande “alive,” jn~jen- “hit,” jan “woman”; *-xt > NWI -(h)t~ -(t)t, as in sot- “burned,” pet- “cooked,” ret- “poured,” vat- “said,” dot “daughter”; the development of *-ft > -(h)t~-(t)t (usually parallel to *-xt)does not seem to occur in HJ, represented only by -ft : geft- “take,” kaft- “fall,” dor-oft- “sleep,” cf. the respective Farizandi (NE subgroup of CPD) forms gat, dar-kaft, hot; *-č- > -j- as in suj- “burn,” rij- “pour” with further weakening to -y/Ø- in vā(y)- “say-pres” and note the alternates pej-~pec- “cook-pres” (but ez~az “from”). The only major developments in the HJ vowel system are (1) -i- in place of -u- in pir “boy,” ri “face,” did “smoke,” found even in Arabic loanwords, e.g., aris “bride” and (2) the -ā- vowel in mān “I,” as opposed to man-men in other Iranian languages, but this development is also common in other CPD.

Phonology. The consonantal system of HJ has the following inventory: p, b, t, d, c, j, k, g, q~γ, f, v, s, z, š, x, h, m, n, r, l, y, although Yarshater’s notes also show a pharyngeal H, especially in Hebrew and Arabic words, e.g., esHāq “Isaac” and sobāH “tomorrow,” but also in words of Iranian origin: Hamerān “I break.” No other sources show this consonant and even in Yarshater’s notes it occurs in very few words. The vowel system (i, e, a, u, o, ā, ey, ow) is also similar to other Iranian languages; ə is probably a variant of e. Stressis phonemic in HJ, e.g., such contrasts as: únā “him (direct object),” unāˊ “they.”


Number. HJ nouns have one type of plural ending for both animates and inanimates: -āˊ, e.g., šāx-ā “horns,” yehudi-ā “Jews.” There are two indefinite markers: the number ye(y) “one” (yey šarbat “a syrup”) and an unstressed (miāˊn-e bāˊγi “in a garden”), but both forms most commonly occur together (yey xiāˊl-e xabi “a good idea”).

Object marking. Definite direct objects take -(r)ā : after consonants (vecé-š-ā né-š-di “she didn’t see her child”) -rā after vowels (hamāˊ-rā bé-dān-ferāte “you have sold us”). -Rā drops when the marker for subject (Set2; see Table 1) is moved from the verb to the direct object in the past tenses (see Fronting): ow-š bəxo “he drank (the) water,” but -rā remains when this Set2 appears elsewhere: píl-ā bé-š-be “he took the money,” sér-ešān-ā hanā-šān bégeft “they hennaed their hair (lit: head).” Set2 on the verb can also express pronominal direct objects: jnút-em “he hits me,” bé-yrut-ed “that he catch you,” kāru bó-košid-eš “you must kill him,” da’vat-ešān kart-em “they invited me.”

Modifiers and the eżāfa. Modifiers follow their noun via an eżāfa connector: un taraf-e divār “that side of the wall,” kārā-ye kie “chores of the house”; adjectives: heyvún-e āqel “a smart animal,” ye xā-y emin-em “another sister of mine.” The eżāfa sometimes also drops: berā masár-em “my older brother.” Both full pronouns or Set2 may indicate possessives: gardán-e man “my neck”; buā-d “your father,” ésm-eš “his/her name”).

Demonstratives. HJ demonstratives are: in “this,” un “that,” as well as the intensives hamin/hamun “this/that very (same) one.”

Pronouns. Personal pronouns are listed in Table 1. HJ has the reflexives xo- and xoc-, which take the possessive suffixes. Both are used reflexively (be xoc-aš beš-vāt “he said to himself,” xo-š-ā beš-xost miān-e ow “threw himself in the water”), emphatically (mān xóc-am beštān “I went myself,” xo-d zuni “you yourself know”), or possessively (bāl-e xoc-aš, “his own wing,” kār-e xo-š “his own work”).

Prepositions. HJ has only prepositions: “with,” ez “from,” miān-e “in, inside,” dím-e “on,” déyr-e “around,” etc. as in vā ján-eš “with his wife,” miāˊn-e sahrā “in the field,” dím-e zamin “on the ground.” Pronoun objects may be either full forms or Set2 suffixes, e.g., berā-š “for him/her,” beš-eš “to him” (doubled Set2), lā-š “next to it.”


Verb roots. The past root in HJ is generally formed from the present root by adding: a) (zun/-ā “know,” keš/-ā “pull”), or b) -d after -n (ken/-d “dig,” xon/-d “read”), and -t after other consonants (bāf/-t “to weave,” xšār/-t “press,” sometimes with a vowel change: ker/kart “do,” etc. A third group shows either wider changes of the consonant before -t (hal/hašt “let,” veider/veidešt “pass”), root reduction (suj/sut “burn,” ferāš/ferāt “sell”), expansion (k/kaft “fall,” he-ni/ništ “sit”), or various other changes (gir/geft “take”). A fourth smaller group with no past formant drops the -n of the present root (birin/biri “cut,” j(e)n/ji “hit”). Some present and past verb roots that end in -r or -rt lose these consonants when final, e.g., vā-ker/kart, ber/bart, hamer/hamart: present root: vāˊ-ke “open (sg)!,” bebe “carry (sg)!”; past root: béšān-be “they carried,” bé-š-Hame “he broke.” These consonants are retained, however, if a suffix follows them: vek(e)r-id “open (pl)!,” bart-eš “he used to carry,” Hamœrt-em “I used to break.”

Preverbsfurther specify a root, e.g., gir/geft “catch,” he-gir/geft “take, get,” ve-gir/geft “pick up”; gard/gardā “go around,” bar-gard/gardā “return.” Many roots only occur with preverbs: he-ni/ništ “sit,” vor-os/osā “get up, rise,” vā-pars/parsā “ask,” etc. Preverbs in HJ accompany all tenses as well as all negatives: henádān “I won’t give.”

Negation is expressed by a stressed, prefixed né- ~ ná- that comes just before the verb root: néšzunā “he/she couldn’t,” nédārān “I don’t have” vānábo “it didn’t become.”

Non-finite forms include infinitives, present, and past participles. Infinitives are formed by adding -an after the (preverb-)Past Root, with a transitional -y- after a final vowel: šiyan “to go,” xordan “to eat,” vāpušāyan “to dress.” Past participles are formed by adding -e after the (preverb-)Past Root: veidešte “past,” (ne)gefte “(un)taken.”

Person endings. While there is only one type of conjugation for present forms (present, subjunctive, imperative), Table 2 shows a basic distinction between intransitive and transitive conjugations in all past forms (preterit, imperfect, perfect tenses; for the full conjugation of an intransitive and a transitive verb in the simple past tense, see Table 3). Intransitives use Set1 endings (as in the present) after the past verb stem (bé-resā-n, dar-kaft-ān “I arrived, I fell”), whereas transitives add Set2 before the stem when be-, a preverb, or a negative particle is present (bé-š-Hame, vāˊ-š-parsā, né-š-zunā “he broke, asked, couldn’t”), and after the verb stem when none of these is present, i.e., in the imperfect (Hamart-eš “he used to break”). Fronting is also a crucial process in past transitives.

Fronting. As shown in the previous paragraph, there are two sets of person endings in HJ verbs (as with most NWI languages). Since Set2 endings are somewhat unusual in comparison to English, other European languages, and Persian, we will reiterate that Set2 endings show agreement with subject onlyin the case of transitive verbs and only in the tenses of the past system, as well all tenses of the verb “to want” (an irregular verb). While the position for the Set1 endings is completely fixed and unchangeable in HJ verbs (just as with all Persian verbs, for example), Set2 endings by contrast are quite mobile. As already seen, Set2 person endings are located in different positions just in the simple past and the imperfect forms even in isolation (see bé-š-Hame, “he broke” and Hamart-eš “he used to break” in the previous paragraph). It can be said in addition, however, that there is a general tendency for Set2 to move forward, i.e., to the left, even inside the verb whenever possible (see geft-em “I would catch” vs. hé-m-e-geft “I would get,” vā-parsān “I ask” vs. vām-e-parsā “I would ask” and beri-em “I would cut” vs. n-em-e-beri “I would not cut”; see below TensesGeneral. comments).

An even further extension of this tendency of Set2 to move forward is found in the process called fronting here. This process only occurs in sentences that have other words besides the subject preceding the verb. In these cases we have the optional, but very common, process of fronting. This process moves the Set2 person endings in the past system of transitives offthe verb to a preceding word (but not to the subject), e.g., (past) har-ci-d buā “whatever you said” < b-ed-vā “you said,” mire-mān henédā “we didn’t marry her off” < he-ne-mān-dā “we didn’t give”; (imperfect) mān har ru tefilā-m exond “I would say my prayers every day” < xond-em “I would read.” Table 3 contrasts the immobility of Set1 endings in the simple past tense of intransitive verbs with the mobility of Set2 in transitive verbs in the past system by showing the optional application of the fronting process in the transitive verbs.

As shown above under “Object Marking,” fronting a Set2 verbal marker to a word that has a Set2 possessive marker is not allowed, but Set2 may remain on the verb: vā dondók-eš béš-ārt “he brought (it) with his beak.” For the effects of fronting with the verb “to want,” see the sentences híci-m nagu and har-ci-d bégu under Modals below.


General comments. The present and imperfect are formed with the prefix e- (also called the durative marker), but the latter is deleted both when it would normally occur alone in initial position, e.g., (pres) šān “I go,” zunān “I know,” (imperf.) šiāyān “I would go,” ferāˊtem “I would sell,” as well as after an ā- of a preverb:vā-parsān “I ask.” The prefix e- is retained after a consonant of a preverb (der-e-kaftān “I would fall”) or after the first element of a compound verb (xerend “they eat,” vs. šum é-xerend “they eat dinner”). It is also retained in the transitive imperfect after Set2 that moves to a preverb (geft-em “I would catch” vs. hé-m-egeft “I would get”; vā-parsān “I ask,” but vā-m-e-parsā “I would ask”), to a compound verb (gerie-šān e-ke “they were crying”), or in all negatives even if the e- normally drops in the affirmative (beri-em “I would cut,” but n-em-e-beri “I would not cut”). After a preverb ending in -e, the prefix e- is realized as -y-: pres. heygirān (< he-e-gir-ān) “I get” vs. Subj. hégirān “that I get.” The prefix e- and an initial o- of a verb root convert to ow- (eo > ew > ow), cf., Pres. dor-owsāˊn “I sleep,” and Imperf. dor-owsāˊyān “I would sleep,” vs. forms without the marker e-: Subj. dór-osān “that I sleep,” command dór-os! “sleep!,” preterit dor-oftān “I slept,” and the infinitive doroftan.

HJ has progressive forms but they appear only very occasionally in actual texts and seem to be modeled on the colloquial Persian equivalent: dārān qand hamerān “I am breaking the sugar (cone),” mān dārtem lebās-em vāpušt “I was getting dressed.”

The verbal Marker b(e)- is used in the formation of the HJ subjunctive (b-ārend “that they bring”), imperative (b-éider “pass by!), preterit (be-šān-be “they took it away”), the present perfect tense (to xorāˊket-ā be-t-xórte “you have eaten your food”), and the past perfect tense (be-šān-resenāˊye-bo “they had delivered”), but it is suppressed in verbs roots with preverbs (see Table 2), in the negative forms, and often in compound verbs: dar jnu “that he hang (him)” (< dar + bé-jn-u).

To be. Aside from the short forms of “to be” (nāxoš-ān, -i, -u, etc.“I am, you are, he is ill”), HJ has both a “to be” of existence in two forms (širini hu ~ hesu “there are sweets”) and a “to be” of location and existence-within: ke yā déru “who is here?”; yey xérsi miāˊn-e jangal déru “there is a bear in the woods”).

Modals. Modals in HJ are gu/gā “want,” zun/zunā “can” (= “know”), kāˊru ~ kār-gu “must,” vā-b/bo “be possible” (= “become”), e.g., zuni hālā béši “you can go now”; kāˊru ce kərim? "What must we do?” (Abrahamian), kār-gu bešān vājār “I must go to the bāzār” (Yarshater); vānábu mān bešān “it’s not possible for me to go.” The modal want is formed with gu/gā preceded or followed by Set2 endings, depending on the form: (pres) gum, gu-d, gu-š, gu-mān, gu-dān, gu-šān; (past) gā-m, gā-d,etc., but subjunctive: bé-m-begu, bé-d-begu, etc., past negative: né-m-e-gā, etc. Examples are: (pres) pādešāh guš xabi békru “the king wants to do good”; (past) gā-šān pādešāˊ-rā masmum kerend “they wanted to poison the king.” Fronting of Set2 also optionally occurs in all tenses of “want”: Pres: híci-m nágu “I don’t want anything,” har-ci-d bégu “whatever you want.”

To become. This verb, in HJ as in Gilaki (see GILĀN ix), Vafsi and others, has a special particle between noun or adjective and the verb: sāket-ā-bi “be quiet!,” ez masartarin reisā hasāb ā-bu “he is considered one of the biggest bosses.” This particle seems to have formally converted to a preverb, even when the verb occurs independently: (see vānábu, belowunder Modals).

Causatives, passives. The causative marker (present/past forms) is -en/-enā, or -ān/-ānā:béxandene “make (someone) laugh!” (< béxand “laugh!”). The -e of -(e)n is lost after a vowel: bédowne “make (him) run!” There are two productive ways to form the passive: either (1) with the addition of -i- to the present root (plus the past formant for the past), cf., active: darzúe/ bešdašt “he/she sews/ sewed” > passive: dœrz-i-u/bédarz-i-ā “it is sewn/was sewn,” or (2) (on the model of Persian) with the use of the past participle + vā-bi/bu “to become”: šekāfte-vā-nabu “it will not split open.” Sometimes a different past root formation is used to form the active (-t) and the passive form () of the verb: (active) sot, pet vs. (passive) pejā, sujā “cooked, burned,” respectively, but in most cases the passive formant -i- is inserted even if the roots are different: (act) hœme(rt), ret;(pass) hmeriā, rijiā “broke, spilled,” respectively.

Affiliations and variants The Jewish community of Hamadān claims to have mostly migrated there from Yazd in the 18th century, but their dialect also shows connections to the Jewish (and non-Jewish) dialects of various CPD areas (see CENTRAL DIALECTS). It would be difficult to pinpoint the exact origins of the dialect without much more research. It is probably not original to the Hamadān area and will most likely prove to stem from different CPD areas but it also has characteristics that are unique unto itself. Members of the Jewish community of Tuyserkān also spoke of their derivation as from Yazd, but they also claim a portion of them came from Isfahan, which is most likely true for Hamadān as well.

Table 4 compares a few typical features of HJ with representatives from each of the four types of CPD as laid out by Pierre Lecoq, to show the relationship of HJ to other CPD. Only some dialects in each of Lecoq’s categories can be mentioned here (and even fewer included in Table 4): northwestern CPD (Maḥallāti, Vānišāni, Ḵᵛānsāri), northeastern CPD (Kāšān Jewish, Qohrudi, Jowšaqāni, Abyānaʾi (q.v.), Farizandi, Yārāni, Meymaʾi, Abuzaydābādi (q.v.), Naṭanzi, Kešaʾi, etc.), southwestern CPD (Gazi, Eṣfahāni Jewish, Seh-Dehi), and southeastern CPD (Yazdi Zoroastrian, Kermān Jewish, Nāʾini, Zefraʾi). Features 1 through 7 in the Table, with a few sporadic exceptions, show that HJ has features that are typical of most members of all four categories of CPD. Features 8 and 9 connect with three of the groups but not with the NW group, features 10 through 12 do unite HJ with the NW group. It should be noted that while some other dialects use the same roots for either “large” or “small” (or both)—c.f., Yazdi mas, kasog, Gazi, Zefraʾi and Kešaʾi kas—only HJ and the NW group of CPD have substituted the comparative form (“larger, smaller”) for the simple forms “big, small.” Of features 13, 14, 13 unites HJ with SW (and Zefraʾi) and 14 unites HJ with Eṣfahāni Jewish (SW) and Vānišāni (NW). Features 15 through 17 are unique to HJ.

As the Jewish community of Tuyserkān was most likely derivative from Hamadān, TJ also agrees with HJ in all major grammatical points and lexical items, e.g., TJ xoc- “self,” , yānā “here,” he-gir/geft “take, get,” maser, kaser “big, small,” ferāš/feroxt “sell,” and HJ emin, TJ emi “other.” Set1 and Set2 are virtually identical in both dialects and the rules for the appearance of the durative marker e- seem to be the same as in HJ: ferāšend “they sell,” kār e-kerend “they work,” mosāferat-ešān e-ke “they used to travel.” Differences only appear in a few words, e.g., TJ pešme “sneeze,” bāyad “must” (from Table 4), and HJ , TJ xuār “sister,” HJ , vānā TJ uvā “there,” etc.



Roubène Abrahamian, Dialectologie Iranienne: dialectes des Israélites de Hamadan et d’Isphahan et dialecte de Baba Tahir, Paris, 1936.

Harold W. Bailey, “Yazdi,” BSO(A)S 8, London, 1936, pp. 335-61. 

Arthur Christensen, Contributions à la Dialectologie Iranienne: Dialecte Guiläkī de Recht, Dialectes de Färizänd, de Yaran et de Natanz, avec une supplément contenant quelques textesdans le persan vulgaire de Téhéran, Det Kgl. Danske Videnskabernes w. Selskab., Historisk-filologiske Meddelelser 17/2, Copenhagen, 1930.

Wilhelm Eilers, Westiranische Mundarten aus der Sammlung Wilhelm Eilers I: Die Mundart von Chunsar, Wiesbaden, 1976;II: Die Mund-art von Gäz, Wiesbaden, 79.

Farānak Firuzbaḵš: Barresi-e sāḵtemān-e dasturi-e guyeš-e behdinān-e šahr-e Yazd, Tehran, n.d. Irān Kalbāsi, Guyeš-e kalimiān-e Eṣfahān (yak guyeš-e irāni), Tehran, 1994.

Ann K. S. Lambton, Three Persian Dialects, London, 1938.

Pierre Lecoq, “Les dialectes du centre de l’Iran,” in Rüdiger Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 263-93.

Manfred Mayrhofer, “Vorgeschichte der iranischen Sprachen: Uriranisch,” in Rüdiger Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 4-24.

Haideh Sahim, “The Dialect of the Jews of Hamedan,” in Shaul Shaked and Amnon Netzer, eds., Irano-Judaica III, Jerusalem, 1994, pp. 171-81.

Rüdiger Schmidt, “Die altiranischen Sprachen im Überblick,” in Rüdiger Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 25-31.

Ehsan Yarshater, handwritten field notes collected in the Jewish communities of Hamadān and Tuyserkān, 1969, kindly provided to the writer.

V. A. Zhukovskiĭ, Materialy dlya izuceniya persidskikh nareciĭ, 2vols., Saint Petersburg, 1888-1922.

(Donald Stilo)

Originally Published: December 15, 2003

Last Updated: March 6, 2012

This article is available in print.
Vol. XI, Fasc. 6, pp. 623-627