Judeo-Kashani: a Central Iranian Plateau dialect.

• Author: Borjian, Habib
• Position: Report

The city of Kashan (Kagan) in central Iran saw the quick disappearance of its Jewish community in the mid-twentieth century. The Jewish population of the city dwindled from thousands in the 1940s (Razmara, 3: 223; cf. Yeroushalmi, 68, 72, 82) to a mere eleven families or seventy-one souls in the 1970s (Yarshater, 466). Their dialect, Judeo-Kashani, survives only among the older generation of Kashani Jewish immigrant communities in Israel and North America and faces potential extinction. Deserted by its longstanding Jewry, Kashan is now an entirely Persian-speaking Muslim city.

Judeo-Kashani belongs to the group of Central Plateau Dialects (CPDs) spoken in the provinces of Kashan, Isfahan, and Yazd, all in central Iran (Lecoq; Windfuhr; Borjian 2008). Together with the Tatic group of dialects, which extends from the outskirts of Tehran westward to Azerbaijan and Talesh, and the Caspian language groups of Mazandarani and Gilaki, CPDs form the Northwest (NW) Iranian languages. These are distinct from the Southwest (SW) or Perside language group, of which Persian is the most prominent member.

Judeo-Kashani shares the closest affinity with the rest of the CPDs spoken in Kashan province. These vernaculars survive only in a few rural communities that have outlived the rapid process of Persianization since the mid-twentieth century, and their condition is moribund (see Borjian 2009). Notwithstanding geographic ties, Judeo-Kashani shows striking similarities to the dialects of Jewish communities in other cities, such as Hamadan and Isfahan, whose gentile population is Persian-speaking. This can be explained by the strong historical ties among the Jewish communities of central Iran, with evidence of significant migration among the towns in the past few centuries (cf. Yeroushalmi, 63ff., 327). These Jewish dialects spoken in an urban milieu show more Persian influence than do similar dialects spoken in the villages, and this has led to a simpler morphology and syntax. It is interesting to note that the extant Judeo-Persian literature from Kashan, namely, that of the seventeenth-century Baba'i ben Lotf and of the eighteenth-century Baba'i ben Farhad, have no trace of Judeo-Kashani (cf. Moreen, 6). (1)

The following study of Judeo-Kashani is based mainly on documentation collected by Valentin Zukovskij during his travels to central Persia in the mid-1880s. It consists of several texts in Russian transcription (Zukovskij, 2: 390-94) and a glossary shared with the dialect of Tajrish (ibid., 399-432). The texts, totaling 1,426 words, are translated from Persian. They consist of a list of forty-four short sentences (Text I, below), a short piece (Text II), twelve tales (Text III), (2) and two letters written by local Jewish residents (Texts IV and V). Although Zukovskij remains silent on his informants, the texts themselves suggest different informants who spoke varieties of the same dialect. This may be seen in adise (in Zukovskij's glossary) vs. aske (Text 1.18) ('sneeze') at the lexical level, and in tez vs. td ('sharp') (see [section]D7) at the phonological level. Another peculiar difference is between xudd (Texts I, III, IV) and the far more frequent xuz,cd (Texts III, IV), both meaning 'God'; the latter must be a loan word from Judeo-Isfahani (see [section]D12). The fact that we only find a grammatical gender distinction in the third singular personal pronoun (see [section]2.2.1) in Text I suggests that this is the last remaining trace of this morphological trait in the nineteenth century.

The texts are reproduced here in Roman transcription with due punctuation, hyphenation, and paragraphing. They transcription is modified by introducing the sounds,g [y], q, and occasionally h, which Zukovskij did not distinguish from g, k, and x, respectively, and by reducing his palatalized k and g and the y to their obvious allophones. Stress marks are reproduced when they are not predictably word final. Other necessary phonological adjustments are ignored in the reproduced Texts, but are made--based on the other sources of the dialect (see below)--when words and phrases are quoted in the Grammar and Glossary. These include ignoring the vowel length (see [section] 1.2.1) and the stress marks, which sometimes lack distinction between morphological and syntactical accentuation, and changing 6 to the more realistic a. The English translations are made not from the Persian source texts or the Russian translations provided by Zukovskij, but directly from the Judeo-Kashani texts. It should be noted that because these texts are translated from written Persian documents, they sometimes lack the quality of natural speech expected from professional dialect documentation.

Besides Valentin Zukovskij (henceforth VZ), material on Judeo-Kashani can be found in three additional sources. (1) Yacqub Tabari (1985, henceforth YT), apparently a native speaker of the dialect, published a short text (Text VI), which adds 138 words to our corpus. This text is in Persian script together with a Roman transcription of such low quality that it causes more confusion than elucidation. (2) Ehsan Yarshater (1974, henceforth EY), in a survey of various Jewish dialects of Persia, published six sentences (replicated as Text VII) and a few words from the Jewish community of Kashan. Notwithstanding its limited quantity, the data allow an objective comparison of the phonology against Zukovskij's transcription. (3) In another comparative study, Haideh Sahim (1996, henceforth HS) conjugates the substantive verb in the present tense (see Table 3) and the verbs 'do', 'fall', and 'sell' in the preterit (see Glossary).

These four sources are in general agreement but not without obvious differences, most of which, such as the preverb der-, ver-(in VZ) da:, va:-(in EY), result from chronology or variation of language and pronunciation among speakers. A notable mismatch is the verb ending for the first person singular: -om in VZ vs. -an in EY and HS (YT lacks this morpheme); these two forms are indeed so different that they mark a thick isoglottic boundary among the CPDs. Even so, such heterogeneity in documentation is not rare in other dialects akin to Judeo-Kashani (see, i.a., Borjian 2009).

PHONOLOGY

[section]1.1. Consonants are /p b t d e j k g g~q f v s z g x h m n r l y 'l, very similar to standard Persian. The fricative g and plosive q are likely to be interchangeable allophones of a single glottal phoneme. z must be foreign to the dialect; its sole occurrence is in td, which is also rendered as tez ('sharp').

[section] 1.2. The vowel inventory of the dialect is probably as simple as /i e a u o 'V, much the same as modern Persian. The phoneme /A/ represents EY a (Text VII) and VZ o (Texts I to V), which I will normalize to the more realistic round a. The numerous vowel symbols used by Zukovskij (a a o o u u i i y l e e) call for the following explanations.

[section] 1.2.1. In transcribing his dialect materials Zukovskij was apparently misled, as were some other contemporary philologists, by the traditional rules of transliterating Arabic-based scripts (Borjian 2006). Thus, he renders the kasra mark sometimes as i, following the orthographic rules of classical Persian, and sometimes as e, according to the actual pronunciation he heard from his informants. The same argument holds true for u vs. o. Zukovskij probably normalized his dialect texts to show the long vowels of classical Persian in the words he could establish an etymology for. He also exaggerated the roundness of the back low vowel a by documenting it as a. Thus, while reproducing Zukovskij's text in Texts Ito V without normalization, I slightly altered his transcription in the Grammar and Glossary to conform with other documentations at hand, e.g., dubore to dubare ('again'). Nevertheless, I keep such inconsistencies as VZ xur-or Are-= EY xor-('eat' ), VZ beguden = YT begodan ('they went' ).

[section] 1.2.2. In spite of the consistent use of a (adjusted here as a), 'Zukovskij used a in words that were not likely to be in common use among his informants (e.g., velayati, mirza, ez-in-qarar, kern-iltifati). The only native words with an a in Zukovskij's documentation are the derivatives of the stems (y)a-: timed-('come').

[section] 1.2.3. The Russian letter hi (i.e., [y] or I) is incorporated by Zukovskij in a few words: xIn ('blood'), ita Yirde Ca bit'), Wiz ('letter'), xIyal ('imagination'), dash ('inside'), yaqin ('certainty'), xatir-jam ('confident'). This sound must be an allophone of /i/ on the ground of diachronic fronting of the back vowels ([section]D14).

[section] 1.3. The diphthongs are probably /aw ow Aw eu ay ey/, as in kaug ('shoe'), nou ('new'), xliu ('sleep'), emjeu ('tonighe), xayli ('many'), pey ('after'). Note the split of the diphthong when suffixed by a vowel: xau ('sleep') vs. xati-i, in an etafa construction.

[section]1.4. The dialect tends toward vowel harmony in verbs: dar-k-om ('that I fall') -der-e-kom CI fall'), dar-ket-om (I fell') -der-e-ket-om CI would fall, I was falling'), dar-vaz-om ('that I lose') -der-em-vaza CI lost'), bar yarn ('that I come out') -ber,,e-yar-om CI bring out'). (3)

[section]1.5. Verb stems may trail off in the final position: beber ('carry!') be-ber-id ('carry ye!'); ba-m-berg CI carried') -ba-m-berd-e CI have carried'); ni4-e-xurd ('he used not to eat') -bi-g-xurd-e ('he has eaten'); .f-e-val ('he was saying') -be-g-va(t) ('he said').

[section]1.6. A stem following a stressed prefix may lose its initial consonant: da-gir-om ('that I seize') -m-e-giret CI used to seize'); be-"in-om ('that I see') -vin-e ('you see').

[section] 1.7. Stress is only documented by Zukovskij, but it is perhaps a mixture of morphological and emphatic types of stress. As in other CPDs the stress appears to be word final in nouns and word initial in verbs.

NOUN MORPHOLOGY AND SYNTAX

[section]2.1. Nouns. (1) The plural ending is -(h)a, as in milla ('hairs'). (2) Indefinite...