The Jewish Communities of Fārs
The southern Persian province of Fārs has accommodated a sizable Jewish community at least since medieval times. According to the twelfth century travelogue of Benjamin of Tudela, there were 10,000 Jewish residents in Shiraz. In the late thirteenth century, when the Ilkhanid Arghun Khan appointed the Persian Jew Saʿd-al-Dowla the grand vizier, the latter appointed Persian Jews as governors of several cities, Shiraz included.Footnote 1 It was in Shiraz that Šāhin, the most prominent poet of Judeo-Persian literature, flourished in the late fourteenth century. Under the Safavids, Jewish populations were reported in the Fārs towns of Jahrom, Lār, and ShirazFootnote 2 (see Figure 1).
Figure 1. The historical Fārs province.
Source: The base map is from Google Earth.
Historical sources from subsequent centuries reveal that the Jewish colony of Shiraz, with all its ebbs and flows, remained one of the strongest and most stable in Persia. In modern times the community was by far the largest (up to 14,000 individuals) in Persia, approaching at times a fifth of Shiraz’s population, before mass emigration in recent decades.Footnote 3 This population increase must have partially been due to the resettlement in Shiraz of the Jewish communities from various districts of Fārs, especially Lār and Jahrom, that took place several times in modern history, to the extent that no other sizable Jewish community was left in the province by the mid-twentieth century.Footnote 4 These historical facts have a bearing on the historical phonology of Judeo-Shirazi, as discussed in the sections to follow.
Up until the twentieth century Shiraz Jewry was secluded in the “Maale,” a colloquial pronunciation of Persian maḥalla “quarter.” The community opened up during the social reforms of Reza Shah’s administration (1925–41), when many families moved out of Maale to live in affluent neighborhoods of the city among gentiles (goy(i)m). Continuous migration to Tehran was already underway by the middle of the century, with a parallel mass emigration of Jews to Israel. Like other Persian Jews, sizable Shirazi Jewish communities resettled in North America after the Islamic Revolution of 1979.
This article is based on data collected from the Shirazi Jewish diaspora living in and around New York City. Their population is estimated at around 4,000 individuals, mostly residing in Brooklyn, where they have seven synagogues exclusive to Shirazis. More than a thousand Shirazi Jews who live in the suburban Great Neck share synagogues with other Persian Jews. The language’s endonyms are ǰudi “Jewish” and maaley “belonging to the [Jewish] quarter.” In New York, I found Judeo-Shirazi to be a moribund language, spoken only among the elderly and only in the most intimate situations.
The Language
Judeo-Shirazi belongs to Southwest Iranian, or Perside, group—an extensive family embracing Old, Middle, and New Persian, in addition to numerous living vernaculars spoken in southern Iran, particularly in the historical provinces of Fārs and Kermān. Judeo-Shirazi should have its pedigree, as discussed below (in §2), in what I designate here as Medieval Shirazi, a language which generated sizable corpora in verses;Footnote 5 its written tradition died out circa fifteenth century, probably along with a disappearing native vernacular of Shiraz among Muslims. One thus expects Judeo-Shirazi to be an insular vestige of Medieval Shirazi, having survived Persian in the isolated Jewish quarter of the city.
Judeo-Shirazi shows significant areal affinity in grammar and vocabulary to the rural vernaculars spoken around Shiraz and Kāzerun (Figure 2). These varieties form a rather heterogeneous subgroup of Southwest Iranian, which distinguishes itself by weighty lexical and grammatical isoglosses from the Lārestāni group to its southeast,Footnote 6 and by extension from the Garmsiri languages of Kerman.Footnote 7 Matters get complicated, however, by the fact that Shiraz absorbed Jewish immigrants from all over Fārs in recent times. The most sizable population integrated into Shiraz Jewry, as my informants told me, came from Lārestān, the southeastern district of Fārs. The fusion of Fārs Jewry in Shiraz may offer clues to explain some of the “impurities” found in Judeo-Shirazi.
Figure 2. Places in Shiraz-Kāzerun area where native Perside languages are spoken.
Source: The base map is from Google Earth.
Judeo-Shirazi has received little scholarly attention. W. Ivanow gives a list of personal pronouns and verb endings.Footnote 8 Georg Morgenstierne provides brief grammatical notes, but more importantly elucidated on the provenance of the lexeme teš “louse,”Footnote 9 leading to the notion of an early phonological split in Fārs (see §2 below), which was later elaborated by Hassan Rezai-Baghbidi.Footnote 10 Ehsan Yarshater published a single short text,Footnote 11 based on which Gernot WindfuhrFootnote 12 included Judeo-Shirazi in his typological study of the “Fārs dialects.”
Historical-Comparative Phonology
1. Old Iranian stage
Historical phonology places Judeo-Shirazi squarely within the Perside family, whose extinct members are Old Persian, Middle Persian, and Medieval Shirazi.
The oldest divergences of this family from proto-Iranian, *ts, *dz, *tr > Perside h, d, s, have these reflexes in Judeo-Shirazi: da: (< dah) “ten,”Footnote 13 rubâ “fox” (< rōbāh < Old Ir. *raupasah-) (see also §2, below); dinka “yesterday” (cf. Lori dinyâ, Mid. Pers. dīg; Northwest Ir. Kešaʾi heze), dišna “last night,” doma, dumâ “son-in-law” (cf. Mid. Pers. dāmād,Footnote 14 Av. zāmātar-Footnote 15), but δemeθθu “winter,”Footnote 16 cf. Parthian zmg, Khārgi demesto “winter,” Manichaean Mid. Pers. dmystʾnFootnote 17; poθ “son” (< *pus = Mid. Pers. < Old Pers. puça-Footnote 18 < *puθra-, Sanskrit putrá-), veθθo “pregnant” (< *aveston, cf. Mid. Pers. ābus(tan), Av. a-puθra-Footnote 19). See also §8.5.
To the oldest stratum of sound changes, we may add the development *št > st, whence Judeo-Shirazi θ(θ) that appears in angoθir “finger-ring” (< angost; cf. Mid. Pers. angust < OIr. *angušta-Footnote 20 “finger”).
2. Proto-Shirazi mutation
During the Old Iranian stage, a group of dialects of inner Fārs in the Shiraz area experienced sound changes that distinguish them from the rest of Southwest Iranian languages. This group, which is labeled as “Ancient Shirazi” by Rezai-Baghbidi,Footnote 21 but I prefer to call it “proto-Shirazi,” survives in medieval poems from Shiraz and Kāzerun and in Judeo-Shirazi and some of the rural vernaculars currently spoken in the area.
The phonological feature that characterizes proto-Shirazi is the merger of proto-Iranian *ts and *tsw into θ, and later also to t, whereas Persian, in its all historical stages, kept the two prehistoric sounds apart.Footnote 22 These sound shifts are further illustrated in Table 1 for major Iranian language families.
Table 1. Development of proto-Indo-European palatals
There are at least three Judeo-Shirazi lexemes that reflect the proto-Shirazi sound changes:Footnote 23
-
2.1. Judeo-Shirazi present stem toδ- (< toz-) “burn,” attested in Medieval Shirazi texts as toz- and θoz-, corresponding to Old Pers. θauc- (emended readingFootnote 24), Avestan saoc-, Sanskrit śoc (with ś < *ḱ). Based on the Medieval Shirazi data, Rezai-BaghbidiFootnote 25 deduces the Iranian root *tsauk, while Cheung, who does not cite Shirazi, gives the root as *sauc.Footnote 26 Note that all other Perside varieties, including Middle and New Persian and the rural dialects of Fārs, have sōz- or the like, not with initial h- or t-, as anticipated from diachrony.Footnote 27
-
2.2. Judeo-Shirazi teš “louse,” also attested in Medieval Shirazi and a number of dialects spoken to the south, southeast, and east of Shiraz, is rooted in proto-Iranian *tswiš(ā)-, which gives Avestan spiš.Footnote 28 There are two more outcomes of this etymon in Fārs (Figure 3):
-
(a) šeš, occasionally šoš, in southwestern Fārs, from Kāzerun down to the Persian Gulf and the littoral band running from Bušehr down to the Strait of Hormuz, as well as in eastern Fārs. This form probably developed from *seš, an expected outcome of *tswiš(ā)-, along the development line of Middle Persian shown in Table 1. Note that Middle Pers. spiš and New Pers. šipiš, šepeš (with assimilation of sibilants) must have been borrowed from a northwestern (Median, Parthian, etc.) form.
-
(b) heš in Lārestān, in southeastern Fārs, presents a fundamentally different change, that is, Proto-Indo-European *ḱu̯ > proto-Iranian *tsw > *proto-Lārestāni *θ > Lārestāni h, or, as Reżaʾi-Bāḡbidi proposes, from proto-Shirazi *θ. More data, particularly on the Lārestāni outcome of *ḱ, would shed light on this issue.
-
-
2.3. Judeo-Shirazi tanǰ- “drink” must correspond with Medieval Shirazi tanz-,Footnote 29 which is defined by the cognate Persian word sanǰ-,Footnote 30 a polysemous word with meanings “weigh, pull, draw, smoke”. The semantic shifts “pull, draw” > “draw water” > “irrigate, water, sprinkle” > “drink” is established by Donald Stilo for Judeo-Isfahani tanǰ- “drink.”Footnote 31 The Iranian root, θanǰ,Footnote 32 has multiple outcomes in New West Iranian, with initials t, h, and s, for both meanings “pull” and “drink.”Footnote 33 I could not find cognate words in other New Perside varieties spoken in Fārs in the available data.
-
2.4. Besides the only two certain instances presented above, there might be more Judeo-Shirazi words reflecting the development *ts > θ. For instance, Judeo-Shirazi θâl “year” may carry an ancient θ on the grounds of Medieval Shirazi/Kāzeruni θal “id” (< Old Pers. θard-, cf. Av. sarəd-); equally comparable is Judeo-Shirazi θar “head” with the same form in Medieval Kāzeruni. However, these instances remain unsubstantiated due to another sound chance, fairly recent in all likelihood, i.e. that of systematic interdental articulation of original sibilants, including s > θ (see §11). This latter shift conceals all possible occurrences of *ts > θ; for instance, it is impossible to determine whether Judeo-Shirazi θâl “year” is a continuation of Medieval ثل or else a relatively recent interdentalization of Pers. sâl.
It should be noted that regional dialects in Shiraz-Kāzerun area have the aforementioned items in §2.4 with a word-initial s-. These dialects are Māsarmi, Pāpuni, Samḡāni, Buringāni,Footnote 34 Kuzarki, Davāni, and Dahlaʾi.Footnote 35
Figure 3. The isomap for “louse” in Fārs.
Source: The base map is from Google Earth.
3. Middle Iranian stage
Here Judeo-Shirazi finds its place on the southern side of the northwest–southwest binary division, on the grounds of these sound shifts: *ǰ > z in δan (< zan) “woman”; *-č- > z in reδ (< rōz < *raučah-) “day,” a-δer “under” (˚δer < zēr < *hačā˚); *dw- > d- in dar “door”; *y- > ǰ in ǰow “barley.”
A notable split that may have occurred in this period is Old Ir. *g > w (contrasting Parthian γ) with the Judeo-Shirazi outcomes murv “hen” (< *mr̥ga-), dorow “lie” (< *drauga-), šolom “turnip,” but not deḡ/deq “buttermilk.”
4. *w
The development of Middle West Iranian initial *w- to b occurs systematically in Judeo-Shirazi, e.g. in bâd “wind,” baro “rain,” barg “leaf,” badom “almond,” beδ (cf. MPers. wabz) “bee,” biǰišk (cf. MPers. winǰišk) “sparrow.” Note that *w- > b- occurs in all attested vernaculars of Fārs for the basic glosses “wind” (bâd, bâδ, bâ), “rain” (bâru(n)), and “snow” (usually barf, but also bafr, etc.).Footnote 36 This sound change therefore must be deep-rooted in Fārs, quite possibly within the Middle Iranian period. It forms a sharp isogloss within New Southwest Iranian, bisecting the Garmsiri languages of Kermān and Fārs.Footnote 37 In Judeo-Shirazi the sound change *wi- > go is attested in a closed set, including gorâδ “boar” and gošna “hungry,” gorg “wolf,” likewise in Persian.
5. Lenition
An opposite effect, the softening of b to v, is prevalent in Judeo-Shirazi: va “with” (< abāg), verd- (< burd-) “carry,” toweθθu “summer,” veθθo “pregnant,” ow “water,” owr “cloud,” θowδ “green,” among other words.
6. Consonant clusters
The medial cluster *-xt- remains in θâxt- “make,” rext- “pour,” etc. However, *-xt- gives -ft in doft “daughter, girl,” likewise in Kondāzi duft,Footnote 38 as opposed to doht, doxt, and do:t in the rest of the varieties spoken in Fārs. Another reflex of the same sound change may be sought in toft-, the past stem of the verb “burn,” which is also attested in Medieval Shirazi texts,Footnote 39 but is absent in other languages of Fārs. In all likelihood the past stem pairs etymologically with its present stem, i.e. Medieval Shirazi toz-/θoz- and Judeo-Shirazi toδ- (see §2 above), stemming from the proto-Shirazi form *θauxt-. The root *tap (with Persian outcome tāb-: tāft- “shine, burn”),Footnote 40 while offering another possibility, leads to the problem of a mismatch between the stem pair in Shirazi.
Other consonant groups of interest are as follows. The group *ft is retained systematically in past stems: goft- “say,” xâft- “sleep,” gereft- “seize,” roft- “sweep,” oft- “fall,” bâft- “weave.” The reduction of *-xr- and *fr- are attested in ta:l “bitter” (cf. Mid. Pers. tahl < *taxra-) and reš- “sell” (< frōš-). *st > ss > θ(θ) occurs categorically, as in oθθoxun “bone,” the past stem šaθ(θ)- “sit,” the past participle marker (< -ist) in vâ-géšt-eθ-en “I have returned,” em=xard-eθ-â “I had eaten,” but not in the religious name Eθter “Esther.” Old Iranian *sp has developed as variously as aθb/aθp “horse,” eθbolak “spleen,” guθfand “sheep.”
7. Consonant elision
Elision of the final consonant is the norm: nasals: donu “tooth,” baro “rain,” δimi “earth,” and in person markers in verbs; stops: doma, dumâ “son-in-law,” etc. Loss of initial h- occurs in amum “bath,” avu “co-wife.”
8. Fronting of back vowels
A remarkable vocal development of the language is fronting of original back vowels, a prevalent feature in Fārs language varieties.Footnote 41 The generalized patterns *ū > i, *ō > e, *aw > ē, with exceptions, are inferred from the data. The outcomes are as follows:
-
8.1. i < *ū is attested in dir (< dūr < *dū-ra-) “far,” did (< dūd < *dūta-) “smoke,” bi- (< būd-) “was,” δâni/δânu (< zānūg) “knee,” xin (< xūn, cf. Av. vohunī) “blood,” θirâx (< sūrāx) “hole,” ǰiǰa (<? ǰūǰaFootnote 42) “chick,” angir (< angūr) “grapes,” and probably kiča “alley,” tit “mulberry”; in Arabic loans: ariθ “bride,” haθid “jealous”; in Hebraisms: Kipir “Kippur,” Pirim “Purim,” Miše “Moses.”
-
8.2. i <ō is found in mi (< mōy) “hair,” ri˚ (< rōy < *rauda-) “on, face” in ribun “roof,” -ši (< šōy or šūyFootnote 43) “husband,” miriča/mur (< mōr) “ant,” giθâla (< gōsāla) “calf, ǰiš- (< ǰōš-) “boil.”
-
8.3. ē̆ < ō, aw occurs in ker (< kōr) “blind,” re(δ) (< rōz) “day,” šer (< šōr) “salty,” reda (< rōda) “intestine,” reš- (< frōš-) “sell,” ǰeḡ (cf. Pers. ǰōy) “brook, stream,” del (∴ < *dōl < dalw) “bucket,” nē (< nō < *nawa-) “new,” nemâ (<? nō māh “new moon”) “month,” me:d (< Hebrew מועד moʿeḏ or Arabic موعد mawʿid) “feast, holiday,” Terâ (< Tōrā) “Tora”; note also θekâ (< suká) “sukkah.”Footnote 44
-
8.4. Note also i, e < *ŭ in xedâ/xodâ "God," ǰeḡd “owl,” šemâ “you”; ate “in, into”Footnote 45 (<? a + tū).
-
8.5. The aforementioned frontings are not universal in Judeo-Shirazi. The following items show no fronting of the back vowels: gō̆š (< gōš < *gauša-) “ear,” toδ- (< *tōz- < *tsauča-) “burn,” ḡora < γōra) “unripe grape,” ko “mountain”; moš (< mūš) “mouse,” xoroθ (< *xraus) “rooster”; to (< *tava-) “you,” avu (cf. Pers. havū) “co-wife,” abru (< brūg) “brow,” bâδu (< bāzūg) “arm,” gulu “throat,” âlu “potato,” paθtu “backroom,” angoθir “finger-ring,” rubâ “fox”; towr (< ṭawr) “manner.”
A striking attestation to the incomplete process of fronting in Judeo-Shirazi can be demonstrated in the cognates poθ (< *puθra-) “son” and veθθo (< *ā-puθra-tanū- “body with a son”) “pregnant” (see §1). Interestingly, modern Persian, which shows little tendency to front the back vowels,Footnote 46 has both words in fronted forms: pesar and âbestan. Fronting in pesar is explained by Hübschmann and Nyberg as an influence from pedar “father.”Footnote 47 Fronting of the back vowel in âbestan can be influenced by other common words such as dânestan and tavânestan in modern Persian.
9. Front maǰhul
The historical vowel ē keeps its quality in δer (< zēr) “under,” but is shifted up, as in formal Persian, in bid “willow; moat,” bini “nose,” biva “widow,” rig “pebble,” θiv “apple.”
10. Final -a and -e
The Middle Iranian word-final -ak is normally realized as -a in Judeo-Shirazi, e.g. adina “Friday,” biva “widow,” čakka “drop,” dikna “yesterday,” dišna “last night,” giθâla “calf,” ḡora “unripe grape,” gorba “cat,” gošna “hungry,” hamθâya “neighbor,” ǰiǰa “chick,” ǰoma “shirt,” kiča “alley,” ko:na “old,” mera “husband,” meriča “ant,” nava “grandchild,” penǰira “window,” reda “intestine,” vaša “sneeze,” θetâra “star,” dada “sister,” among many other words. To this list one may add non-Iranian words, such as amma “paternal aunt,” qatra “drip,” dokma “button.”
Nevertheless, contrary to the norm in Iranian languages, Persian varieties included, Judeo-Shirazi has word-final -e as well. The following words are utterances from the same speaker: bâxâǰe “grandfather,” taše “lung,” owle “smallpox,” gorǰe (Pers. gowǰe, âluče), ḡonče “bud,” solfe “cough,” ime “firewood.” It is quite possible that this dichotomy is at least partly due to amalgamation of the various dialects of Fārs that came into contact in Shiraz’s Jewish quarter during the consolidation of Fārs Jewry in the near past. Note also the doublets bača/bače “child” and ḡolba/qolve “stone,” by different speakers, in the data.
11. Interdentalization
The interdental articulation (θ δ) of original sibilants (s z) is a systematic sound change in Judeo-Shirazi. This shift can be a merger of three processes:
-
1. Historical: the original phoneme /θ/ in proto-Shirazi, as explained in §2, has been generalized.
-
2. Areal: similar trends exist in Davāni [δ] (an atypical Fārs variety language spoken near Kāzerun), if only postvocalically,Footnote 48 as well as in a few other vernaculars of Fārs, such as Galedāri,Footnote 49 but none immediately around Shiraz.
-
3. Ethnolinguistic: Although not exactly the same way as in Judeo-Shirazi, interdentalization also occurs in Judeo-Isfahani and Judeo-Kashani,Footnote 50 singling them out areally from their kindred dialects spoken among Muslims. This peculiarity suggests a phonological spread among the Jewish communities of central Iran, with evidence of strong historical ties and extensive migration among urban Jewry in the past.Footnote 51 On the other hand, Shiraz Jewry is described as utterly conservative, seldom practicing exogamy.Footnote 52 The Jewish communities of Fārs were separated from other Jewish communities of Iran by a great distance. The closest stable Jewish community, that of Isfahan, was a three-week-long caravan journey from Shiraz. The only known major coreligionist contact between the two cities, as my informants told me, was the annual pilgrimage made by the Shirazis to the Pir-e Bakrān mausoleum near Isfahan.
Conclusion
This historical-comparative analysis, particularly the outcomes of the phonological mutations from proto-Iranian *ts and *tsw, makes a strong case that Judeo-Shirazi descends directly from the medieval dialect of Shiraz/Kāzerun. This pedigree can further be substantiated from morphosyntax as well as certain lexical items, such as the present stems ǰ- “give” and nis- “put,” which are also attested in medieval manuscripts.
The author would like to thank the Persian Heritage Foundation for supporting this project.
