LOTERĀʾI (Lutrāʾi, and phonic variants thereof), term used by Iranian Jews for speech characterized by local Judeo-Iranian grammar with a special exotic substitutive vocabulary which is employed in the presence of gentiles to prevent them from understanding. I will try here to show the early Achaemenid origin of Loterāʾi and to demonstrate how Loterāʾi became a chief source for a medieval Muslim Persian underworld argot, whose vocabulary in turn entered argots of Gypsies of Iran and Central Asia, still richly attested at least in Tajikistan.

Yarshater’s study and the issues it raises. Yarshater’s pioneering scientific study of Loterāʾi (Yarshater, 1977; see also JUDEO-PERSIAN COMMUNITIES OF IRAN x. Judeo-Persian Jargon), based on fieldwork in various Jewish communities of Iran, in effect entails the following questions: (1) Is the term Loterāʾi (etc.) indeed, as believed by Iranian Jews, etymologically from Hebrew “Non-Toraic”? (2) What would be the precise feature to which the latter etymology, if correct, refers? (3) How can the Jewish etymology be reconciled with the fact that (a) the earliest allusion to Loterāʾi (lwtrʾ) in the 10th-century Ḥodud al-ʿĀlam, stating that lwtrʾ is one of the two languages of Astarābād, does not mention Jews; and (b) allusions to this speech in early Persian lexica and poetry of the 12th century cite words not known from Jewish Loterāʾi (henceforth JLtr) and, again, do not mention Jews? Note also that the use of the term lotar(ā)/lutar(ā) for an exclusionary language, found in the old dictionaries of Classical Persian (as detailed by Yarshater), and the present-day use of the term outside of Jewish communities, is used generically, without reference to Jews. (4) What is the relative chronology of the heavily predominant Hebrew component and the smaller Aramaic component, of JLtr, and (how) is this relevant to the foregoing questions? (5) What of the verbs (etc.) of hitherto unidentified origin?

The non-Jewish data. The medieval Muslim (Shiʿi) Persian argot is attested by the Ketāb-e sāsiān ba-kamāl (q.v.; henceforth KS) The Book of Accomplished Grifters,” extant inter alia (as a probably late copy) in a 14th-century anonymous Persian manuscript in Tashkent, within five pages of marginalia, of which I have obtained photo scans. A series of argot words, with Persian glosses, is given in the opening section of Ketāb-e sāsiān in nine thematic chapters. Many of the words reoccur at the end of KS in verses with some Persian annotations.  Previously a brief list of some of the words was given (with occasional misreading) by Ivanow (1922), and another brief list was given by Troitskaya, this almost exclusively for comparison with words from her Uzbek argot data. It should be noted that there are no Indic words in the KS, although Indic material abounds in the Gypsy, mendicant, and musician argots of Iran and Central Asia.

Linguistic sources and relevant abbreviations. Jewish Loterāʾi (JLtr.) of regions in Iran: Bor(ujerd) (Yarshater, 1989); Isf(ahan), Gol(pāyagān), Kash(an), Kerm(ān), Khom(eyn), Mash(had), Shir(az), Teh(ran), and Yazd (Yarshater, 1977). Some JLtr words and phrases from Iran are given without provenience in Mizrahi, pp. 121-23. Her(at) argot (found in Zarubin) is a form of JLtr (see Lazard, p. 252). In this article, all JLtr words from Iranian are from Yarshater’s two articles, unless otherwise indicated. 

20th-century non-Jewish argots:

PG = Gypsies of Eastern Iran (Ivanow, 1914; idem, 1920), and 

PD = mendicant dervishes of Eastern Iran (Ivanow, 1927).

Djougi (i.e. Jugi), the term used by de Morgan, pp. 304-6, for the speech of Gypsies of Astarābād.

AG = Ḡorbati of Arāk, a Gypsy argot whose word-formation and lexicon shares features with PG, Djougi, and the Jugi of Tajikistan (Moḡdam, pp. 142-52).

Mus = Argot of traditional musicians of Iran, specified as to performance (Sāzanda, Mehtar), Gypsy provenience (Luti, Tošmāl), or locality (Čāli, Torbat-e Jām) (Amanolahi and Norbeck, pp. 283-286).

LG = Argot of Gypsies of various localities of Iran (Fārs, Kermān, Osof) (Amanolahi and Norbeck, pp. 283-86).

Abd(oltili) “language of itinerants” designates the argots of Uzbek-speaking artisans and musicians, preachers and qalandars (Troitskaya). 

Argots of Tajik-speaking Gypsies of the Ḥeṣār Valley are: Ar(abča) = Tashkent (Luli), Chist(oni), Ju(gi), Kav(oli), Sam(arkand) L(uli), and Sogut(arosh) (Oranskiĭ), Mag(ati), an argot of Persian-speaking Gypsies (Pstrusińska, pp. 71-73, cf. pp. 105-7 for comparisons of Mag with other Gypsy argots of Central Asia.

AD = argot in the 10th-century Qaṣida Sāsāniya of Abu Dolaf Yanbuʿi (Ḵazraji) (edition and commentaries in Bosworth, 1976).

SD = argot in the 14th century Qaṣida Sāsāniya of Ṣafi-al-Din Ḥelli (edition and commentaries in Bosworth, 1976).

LuJ = Luter-e (lwtr) Jāberi, argot (with some lexical affinities to Gypsy argots of Iran and Central Asia) of itinerant Kurdish-speaking traders of the Jāber community in the village of Badra, Ilām province, southwestern Persia (Karimi). Cf. Paul.

Other abbreviations: Aram = Babylonian Jewish Aramaic; Heb = Hebrew; OIr = Old Iranian; J = Jewish; Pers = Persian.

The Aramaic component and Achaemenid origins of the jargon. I begin with Question (4) above. JLtr nouns and adjectives of Aramaic origin in Yarshater’s data are Bor yumā “day”; Gol libbā “heart”; Khom rešā “head”; Shir rakka “thin”; Gol pasilā “gentile” < pəsīl “unqualified” (also “idol”); and, integrated with conspicuous phonological change, Shir yārtak “boy” < yaldā, and a few other words noted below. More significant are the JLtr verbs of Aramaic origin. Clear examples are: Gol nazq-un- Kash nask-en- “to kill, hit” (Aram nazq- “to injure”); Shir za(w)n-, Teh dev-, Isf dam- “to sell” < zabben, Shir deyl- “to fear,” denominative from deḥel “fear, fright”; and Mash meštā-, Her meštō- “to say” < Aram meštaʿʿē “saying,” whose agreement is chronologically noteworthy. Other verbs of Aramaic origin will be noted below.

Particularly in view of “fear” and “sell,” Yarshater wondered whether, at the time of the inception of Loterāʾi, the Jews of Persia were not Aramaic-speaking and if only later in the course of time increasing Hebrew elements were imported (Yarshater, 1977, pp. 4-5). The latter situation is confirmed by (1) evidence I shall present that JLtr originated in the Achaemenid period; (2) the Aramaic origin of words cited as lwtrʾy in a 12th-century poem by Suzani of Samarkand; and (3) data of the KS and later argots of Iran and Central Asia.

We may start with the fact that Shir bika “egg” must go back to Old Aramaic form of the early Achaemenid period in which the antecedent of later Aramaic and Syriac bē(y)ʿā, Heb bēṣāh, and Arabic bayḍa(t-) “egg” had for the second consonant a sound which is spelled as qōf in other words which show the same correspondences―e.g., Old Aram ʾrqʾ (cf. Pahl. Aramaeogram ALKA), later Aram and Syriac ʾarʿā, Heb ʾereṣ, Arabic ʾarḍ, the change to the sound/letter ʿayin taking place in the late 6th century BCE.

Another Mash-Her agreement, nund- “to give,” is also found in KS, nwndydn “to give” (Pers gloss dādan), bnwnd “give!” (Pers be-deh), cf. Her be-nond “id.” (stem nund-), which I take as denominative from Old Aram *nudn- “(dowry-)gift” from Late Babylonia (9th century BCE) nudnū, nudni “dowry,” cf. Neo-Assyrian (9th-7th century) nundunnû > nudannu > Aram nədunyā “dowry,” all < Akkad n-d-n “to give.” For semantics, cf. English endow < Old French endouer, Lat dotāre “give dowry” < dōs dōtis “dowry” = *“gift par excellence,” Latin “give.”

Matching bika and nund- as early Achaemenid residue in JLtr are what I explain as fossils from Old Iranian. Some examples: 

(1) Shir āj- = Her huj- (*hoj-) “to come” < OIr *hāčaya- mid. *“to lead oneself,” cf. Avestan hācaiia- (Y. 5.18, etc.) “to lead, direct, persuade” (contra Lazard, p. 253, Her huj- is not related to Neh hez-, etc., for which see s.v. KS hzyδn, below).

(2) Neh viāj- “sell, finish with” < OIr *abi-hāčaya- or *api-hāčaya-, cf. Avestan upaŋhācaiia- (< *upa-hāčaya-) “to come to agreement with someone” (Mash velāj- = Neh viāj-, or < OIr. *upari-hāčaya-?).

(3) Shir čed- “know, understand”; Kash če(-V)-, čā(-C)- “know, see”; Isf, Yazd, Kerm čer- “see, know, understand, recognize”; KS jhstn “to see” (Pers didan), bjh “see!” (Pers be-bin), < OIr č(a)it/θ- “recognize, perceive” (Old Avestan cōiθa, acistā, čikōitərəš). The KS forms (with past stem like that of Pers dānestan “to know”) represent *č(a)iθ- > čVH- > Kash čā-/če-, whereas OIr *č(a)it- > čēd, with West Iranian development, and Isf etc. čer, with “Tatic” development; cf. Judeo-Yazdi šerin “I went,” with šer- < šuta-.

(4) Bor ča:n “good, pretty,” KS jhn = *čahn “good, beautiful” (Pers niku) < OIr *čaxnuwāh perfect participle “delighting”; cf. Old Avestan cāxnārə̄ “they delight” (-xn- > -hn-, cf. OIr tauxmā > Middle Pers tōhm > -tōm “family” in Pahl mltwm = Man. Middle Pers mrdwhm “mankind.”

(5) Mash z(ə)vā- “to say” < OIr zvā- (Old Avestan zbā- = */zuā-/) “to call, invoke.” KS bwzfʾ (with f representing , a bilabial fricative fully written as the Arabic letter f with three dots instead of one superscript dot) “say!” (Pers be-guy), wzf[ʾ?]ydn “to say” (Pers goftan) < OIr (abi-)zvā- “to call” (cf. Pahl āzbāy-); the JLtr may have labial dissimilation. 

Other JLtr/argot words of Iranian origin cannot be certainly dated to Old Iranian: Mash and Her ruj-, Gol rej- etc. “see, know” < root rauč- “(be) illumine(d)”; Shir kelows- “to laugh” < root xraus- “to shout”; Shir and Teh margun- “to hit” < “make dead”; cf. Shir. kod “hit, kill” < Aram q-t-l or q-ṭ-l “to kill.” KS mlk “man, male” (Pers mard) (scribal error, probably based on KS mlkʾ “ruler” [Pers amir] < Aram malkā “king”) may be an error for *mʾk, and early spelling for /mak(k)/; cf. Ju, Chist, AG mak(k), Mus Čāli makak, LG Osof makk “man, male.” JLtr has the comparable form mak(k)eyhū = Modern Heb ha-iš “the man” (Mizrahi, p. 123; -eyhū < Aram demonstrative?). I take mak(k) from *martk < *martaka- “man.”

Aramaic material in the non-Jewish sources. Expectably, early JLtr had Aramaic forms which have disappeared from modern JLtr; thus may be explained the designation lwtrʾy for <dx> and <zyf> in Suzani’s poem. I take these from Aram daxyā, dəxē “clear, pure, (ritually) correct” and zayif “false.” KS collocates dx (Pers nik “good”) and zyf (Pers bad “bad”). Not only is dax found for “good” in PG, PD, Djougi, and all the argots of Central Asia, but PD, Abd, Ju, Chist, and Ar keep the original Aram meanings “clean, pure, right, correct” in addition. Remarkably, AG has dax “good, right” and daxiyā “pure.”

As shall be seen, the KS and related sources show a clear predominance of Aramaic over Hebrew etyma, and many of these words have correlations in 20th-century JLtr. Accordingly, for Loterāʾi, indeed < *Loʾ-Tōrāʾī “Non-Toraic,” the reference was “having a vocabulary not from Hebrew, but from Aramaic.” Indeed, our materials cover the span of Aramaic spoken by Iranian Jews, from the early Achaemenid period (see bika and nund-, above) to forms paralleling Jewish Neo-Aramaic of Iraq: KS nʾšy “common people” (Pers ʿavām), cf. Turkish Yürük nomads naš “people,” Abd noši “a copper coin” (Troitskaya, p. 264), all reflecting late Aram nāšē (as in JNeo-Aram) < ʾənāšē “people” (the Abd form refers to the coin’s bearing the image of a human face; similarly SD mard “dirham” < Pers mard “man,” Bosworth, 1976, II, p. 303). Similarly, KS ʾmʾ “a hundred” (Pers ṣad) < late Aram ʾəmmā (as in Neo-Aram) < Aram məʾā “a hundred.”

Early preponderance of Aramaic over Hebrew in KS. The preponderance of Aramaic over Hebrew in the earlier phase of JLtr is evidenced by the fact that the verbs in KS which are of Semitic origin are from Aramaic. The following correspond to JLtr verbs:

(1) kʾlydn = *gʾlydn (Pers raftan) “to go.” Shir gāl- “go,” past stem gāled- < Aram g-l-y, ptc. galy- “to go out.”

(2) hzyδn “to go” (Pers mesloho “like that” [i.e., like the preceding word, = Pers raftan “to go”]), bhz “go!” (Pers borow), kh hznd “that they go.” The synonymy of kʾlydn and hzyδn is also indicated by Pers va “and” before hzyδn. Neh hez- “go,” behez- “go!”; cf. Gol. etc. ez- “id.” < Aram. ʾ-z-l.

(3) tknydn “to make, to do” (Pers kardan), ntkn “don’t make/do!” (Pers makon), also aux. tkyn-, tykyn-. Mash teken (not tek-en-), Her tikin- “to fix, make,” Aram taqqen, “to establish, fix” (also > Shir ta:n- “id.”).

(4) hʾlmdn “to sleep” (Pers ḵospidan), hʾlmwth “asleep” (Pers ḵofta), Aram h-l-m, ptc. ḥalm- “to dream.” The same semantic relationship in Ham etc. dar halum- “to go to sleep”: Heb ḥalōm a dream” points to a replacement (probably widespread) within JLtr of Aramaic by the Hebrew cognate. Cf., e.g., ʾʾxʾlydn below. The rendering of Aram and Heb by h is consistent throughout JLtr, from whose medieval form KS and thence the argots of Iran and Central Asia also have h; this contrasts with ḥ > x in Neo-Aramaic.

(5) KS dhlydn “to fear” (Pers tarsidan), mydhlm “I fear” (Pers mitarsam), midahlad “he fears,” ne bedahl “don’t be afraid” (Mizrāhi, p. 123). Isf dalan = tarsan “to fear” (both words are given together with čandan and glossed by Pers tarsidan in a general [i.e., non-Loterā’i] vocabulary of Judeo-Isfahani; Ebrāhimi, p. 15). Ju medahlum “I fear,” Chist dal- “to fear,” PG mīdella “fears,” and PG dōl(iden) to fear” < Aram d-ḥ-l “to fear,” ptc. (participle) daḥl-. Cf. Shir deyl- “to fear” <Aram noun deḥel “fear.”

(6) KS ʾʾxʾlydn “to eat” (Pers ḵᵛordan), cf. Ju oxolīdan id.,” KS byʾxʾl eat!” (Pers beḵᵛor), cf. Chist bioxol “id.,” PG okhōl- to eat” (sic Ivanow, 1920, p. 291: not okyōl); < Aram āxal “ate.” Abd axlamoq “to eat” may derive via medieval JLtr from Aram ptc. ʾaxl-, cf. Shir ōxel- to eat” < Heb ptc. ʾōxel-

Other Aramaic verbs shared by KS and the JLtr and gentile argots of the 20th century are:

KS br kym “get up! arise!” (Pers bar-ḵiz-). PG, Kav, Ju kim- “get up”; Ar barkim; Ju bur kim, dar kim “get going.” Gol kām- stand up, exist”; Neh causative kāmun- put in a state, cause”; Her kem-, Neh dar kām- “sit down”; Mus Luti kemeed, Mus Tošmāl keemed, LG Osof, LG Fars kemeed “went” < Aram q-w-m, with forms qām and qīm to arise, put in place.” (Lazard’s [p. 254] speculative derivation of kām- from the Indic Gypsy word seen in Ju kam “work” [< Sanskrit karman-] must now be discarded.)

KS *brsydn (unpointed) for prsydn “to eat” (Pers ḵᵛordan). Ham, Gol, Isf peris-, piris-, pris- to eat.” In the Kurdish argot LuJ, pirüs “food, eating” (= Pers ḵᵛoreš, Kurd xwârešt), notably a noun like the etymon, Aram ptc. pārīs (f. pərīsā) “broken (bread) for distribution or blessing.” LuJ words found in Gypsy argots of northern Iran and Central Asia also have düywen “ghee” (Pers rowḡan/ruḡan, Kurd řun) < *dūhan < Aram dōḥan oil (other than olive oil),” and zâyra barley” (Pers jow, Kurd jüywa), cf. KS sʾryʾ… (last letters illegible), < Aram səʿāryā (śəʿāryā) barley” (cf. Gol saʾur, Khom saʾuri < Heb śəʿōrā, pl. śəʿōrīm “barley”), PG zabul, zaul “barley” may have u for ō < ā, cf. PG mezūl below.

Other traces of Aramaic not evidenced in later JLtr are:

KS hʾzydn to show” (Pers nemudan) < Aram ḥazzey “to make see, to show.” From the Aram root ḥ-z-y; Mag has mi-azi “you see,” bi-az “look!” (Pstrusińska, p. 71). For the loss of *h-, cf. Sheikh Momedi (Mag) ādur “begging” below.

KS hrʾšydn “play a flute” (Pers ney zadan) < Aram ḥ-r-š “to enchant, hypnotize,” in reference to snake-charming, or < Aram ḥ-r-š, in view of the following brklh ṭrwšydn “to play (lit. “strike”) a lute” (Pers barbaṭ zadan, see below), bṭrwš “beat! strike!” (Pers bezan) < Aram ṭ-r-š to batter” (KS <ṭ-> probably via Arabic ṭ-r-š to deafen,” which also occurs in Aramaic).

Jewish culture in the gentile argot vocabulary. The Jewish cultural background of the Aramaic components of the KS is dramatically clear from the outset of the latter, the first line of which, beginning the chapter on names pertinent to Islam, has rhmʾnʾ “God” (Pers ḵodā) < Aram raḥmānā The Merciful One,” usual for “God” in Jewish Aramaic texts; kmʾr “John the Baptist” (Pers Yaḥyā) < Aram kəmārā priest of a pagan temple”; and hwyʾkʾr (Pers Musā), lit. “snake-handler” from Aram ḥiwyā “snake” + Pers -kār (see below) in allusion to Moses’ curative brazen/copper serpent and/or the contest with Pharaoh’s sorcerers.

Talmudic Aramaic usage is reflected in KS nhwr “eye” (Pers čašm) and separately nhwr “blind” (Pers kur). Ju nuhur and SamL nuhůr are again both “eye” and “blind”; in PG and PD nuhur (AG nhūr), Mus Luti, Mus Tošmāl, LG Osof, LG Fars, LG Kermān, náhur means “eye,” versus Ar, Mag, Sogut, and Abd nuhur “blind.” From argot came Pers nohur and Tajik dialectal nuhūr sight, eye.” The Talmud has nəhōrā light (of the eyes),” and, as euphemism for “blind,” saggī nəhōrā having much ‘light’” (Syriac saggī nuhrā, with phonological differences). The Talmudic phrase survives in Yeshiva Yiddish. KS also has nhwr tykynh “victory,” lit. “(day)light-making,” a kind of etymological calque for the additional glosses Arabic fajr al-manṣur, Pers piruzi “the victor’s dawn, victory.” Earlier explanations of the argot word as from Arab nūr light” are wrong.

Aram gālūt (the Jewish) Exile” > Gol gālut “(in) misery, miserable” (gālut-and “the are in misery”) point to KS kʾlwt as *gālūt (still a noun, obj. of nwnd- “give” in an unglossed verse) as confirmed by Djougi galout “bad” (see below), SamL gohlut, Abd gaulud, golud, Ju and Ar gohlud “ill, sick.”

The aforementioned KS hwyʾkʾr “Moses” separately reoccurs glossed as Pers mārān-gir “grasper of snakes” in the section on professions, the first of which is *knʾw (ms. kyʾw) = *gnʾw “thief” (Pers dozd), cf. Chist ganav, Ju ginop, Abd genou, Ar ginau, ginop, PG genew < Aram (and Heb) gannāβ thief,” whence Gol qannō id.

Our hwyʾkʾr represents a compound of ḥiwyā “snake” plus Pers -kār doer”; cf. the rhyming Pers daryākār seaman”; cf. with -kard “made, done,” another Aramaeo-Persian compound, dmʾkrd red” (Pers sorḵ) < “made of blood” (Aram dammā = Heb dam “blood,” whence Bor dam “red”). The uncompounded hwyʾ “snake” occurs in the section on animal names, after another Aramaeo-Persian form, dhbʾbʾ “scorpion” (Pers gazdom) *“golden-legged,” with dhbʾ from Aram dəhaβā gold” and -bʾ = -pā foot, leg”; this describes Iran’s most conspicuous scorpion, Orthobuthus doriae.

KS mylh (Pers harza “idle talk, gossip”) < Aram milleh in Talmudic usage both “his word”; cf. millē “words” and “gossip.” For -h = -ēh *“his,” see below.

Other words with Aramaic background in the KS section on animals are klbʾ and tnʾγwl. The first, klbʾ “dog” (Pers sag), cf. Ju kalpak, PG kalpik etc. “id.,” is equatable with Aram kalbā dog,” which, via comparison with Arabic kalb dog,” gave rise to an argotic attached to Arabic words: KS qlbʾ “heart,” bṭnʾ “belly,” etc.; thus KS forms like ydʾ hand” are ambiguous in origin.

KS tnʾγwl (γ unpointed) “hen, chicken” (Pers morḡ), cf. Abd and Ar tanoγul, tanaγul “id. < Aram. tarnəgōl [tarnəγōl] “rooster,” whence Bor tarnegul “id.; the notable *tanā- of the KS and Gypsy forms vs. Aram tarnə- will be discussed below.

For “lion” (Pers šir) KS has klbʾ mlkʾn “dog-render,” and for “hawk” (Pers bāz) KS has tnʾγwl mlkʾn “chicken decapitator,” structurally Persian participles, i.e., Persian compound participles based on Aram m-l-q, ptc. malq- “to lop the head off a bird, tear apart with the claws”; see further, below.

Other early Aramaisms preserved in gentile argot. KS has many other nouns from Aramaic which must have been extant in medieval JLtr, although few survive in modern JLtr. Examples are:

KS *sʾwth (ms. sʾwnh) “old (man)” (Pers pir), cf. Ju sovut, PG sobut, Ar sout “old man” < Aram *sāβūt replacing sēβūt old age” via sāβā, sāβtā “old woman”; for -ūt- Bor hevalut “bad” < Heb hevalūt vanity.”

KS rhmʾ lover” (Pers ʿāšeq) < Aram raḥamā id.

KS šydʾ insane” (Pers divāna, from div “demon”) > Pers šeydā “crazy” < Aram šēdā “demon”; cf. Shir šed(d)- “to catch disease.”

KS škrʾ “lie” (Pers doruḡ) < Aram šiqrā “id.” Cf. Heb šeqer lie,” which entered various Jewish languages, e.g., Judeo-Isfahani and Yiddish.

KS mʾhwz city” (Pers šahr), cf. SamL muhůz, Ju muγuz “town” < Aram māḥōzā city.” 

KS also uses mʾhwz in a series of argot terms for various cities of Iran and Central Asia. Very interestingly, Herat (hryw) appears to be called mʾhwz hr[ʾ?]t lsʾnk “the city of Herat speech”; *lsʾnʾ elsewhere < Arabic lisān “tongue” (Pers zabān). Assuming the apparent hrt *lsʾnk has -k = diminutive -ak, then “Herat *argot.” Alternatively, the -k could be a copyist’s error for ʾ of lsʾnʾ. In either event, the phrase thus refers to the (*Jewish?) argot of Herat or that city’s distinctive Persian dialect.

KS dkh (sic twice; not dlh = Ju dela, below) “house” (Pers ḵāna). Cf. Ju dak “locality, community,” cf. Ju dak-i mo one of our people”; Ju indak *“this place; here”; undak “that place; there.” Cf. also Ju gdok “where?” < *kudok “what place?”; Shir dāqim “place” (*“my place,” cf. Shir qutim myself” [below], or conceivably *“this place,” cf. Mid. Pers im “this”), idāqim “here.” In addition, Kav and Mag have duka “house.” Toward an etymological solution, note further Her indof “here,” Mash kondāf where” (< kudōf via *indāf). These may derive from Aram daf “framing plank, column on page” > *“locus” (cf. German Rahmen “frame, milieu, scope”). With Ju and Mag duka < early JLtr *dūq < Aram dūx place,” one can see contaminations *daf/*dāf and dūx (> *dūq) giving rise to *dax > *daq > *dak, and *dāx > *dāq > *dāk (> dāq). The phonetic affinity between x and f would have had a role; cf. German Luft < *luxt; Pers joft “paired” < yuxta-.

The change of /x/ to early JLtr /q/ to KS k is also reflected by Pers ḵᵛod “oneself,” Shir qutim “myself,” qutit “thyself,” qutiš himself,” cf. KS kwdʾwndm myself” (Pers man “I”), *kwdʾwndt “thyself” (Pers to “thou”), and kwdʾwndš himself.” Pers ḵodāvand “lord, possessor, authority,” suggests the possibility that “self” in “thyself, himself” underwent an argot expansion to “thy lordship, his lordship,” whence “myself” = “milordship,” giving inflated forms for “I, thou, he.” For */x/ > */q/, see also below on KS ʾbyk, Shir abeq < Heb ʾaβīx(ā).

Forms from Aramaic possessive nouns. Like the still-extant JLtr, KS reflects Aramaic nouns taken over independently (like Middle Iranian Aramaeograms) as non-possessive forms with what were originally various personal possessive endings. This applies in KS particularly to body parts (cf. Khom ragle “foot”) < Aram ragleh “his foot,” like Pahlavi LGLH = “foot”). Along forms with additional -h < Aram -eh “his” are KS hrh “rump, behind” (Pers kun) < Aram ḥor; lkth “finger” (Pers angošt), cf. Aram l-q-ṭ “pick up”; and (reflecting Aram my”) KS dkny “beard” (Pers riš), cf. PG dagnā, degño mouth, beard, lips, teeth,” AG daqnā mouth.”

Without reflection of Aram possessive suffix, KS hʾr (Pers gu [sic for guh]), Central Asiatic argots hor < Aram ḥārē “feces”; and kʾkʾ “tooth” (Pers dandān), cf. Ju kokon “face,” PG kōkīdan to laugh” < Aram kākkā (molar) tooth.” (Troitskaya, p. 254, wrongly cites a Syriac “kokha,” which led others to attribute Syriac as a component of our Gypsy argot vocabulary.)

KS rjyʾjh nose” (Pers bini) is probably < Aram rēḥā breath” plus Persian suffix -ča. Cf. ryhʾny “fragrance, odor” (Pers bu) < Aram rēḥānē fragrance.” KS hwtrʾ < Aram ḥuṭrā “stick” glossed Pers dār-e vey its beam,” referring to the preceding entry myʾn tnk (?) probably = Pers miān-tang “having a narrow middle” (glossed Pers sollam “ladder”). 

As Ivanow (1922, p. 378) suspected, KS br is both associative and privative. The associative meaning is found in Aram bar “son, someone or something exemplifying a part/implement.” KS brkʾlʾ “lute” (Pers barbaṭ or barboṭ), spelled brkʾlh before ṭrwš- “beat, play” (see above). Probably < Aram bar qālā *(implement) having a voice”; for formation, cf. Aram bar ṭawāy “utensil under a roast,” bar lōʿā “board securing an animal’s jaw,” and for the meaning, cf. Flamenco Caló argot sonanta guitar.” The next entry has brkʾlʾ dyγʾ “false (or Turkish) lute,” glossed as Pers jnk, i.e., čang “harp.”

KS br < Aram bar “without” (cf. Pahlavi Aramaeogram BRH = without”) is found with yet another Aram noun in brmyʾ thirsty” (Pers tešna, miscopied as fetna “sedition”), where -myʾ compares with PG, PD moi, Ju mayō, mayob, etc., Chist mai “water,” cf. PG mionew, AG mianu “water,” Mus Luti, Tošmāl meyow, Mus Čāli meyab, LG Fars meyow, all from Aram mayyā water.” Preceding brmyʾ is br hrsyt “hungry” (Pers gorosna), with KS hrsyt “bread” (Pers nān) = Central Asiatic Gypsy argots harsīt (< Arab harīsa?).

A rare instance of a correspondence of a KS and JLtr noun: KS klʾh “stone” for *glʾlh correlates with JLtr glʾlh (= Heb even, Mizrahi, p. 123) “stone, rock” < Aram gəlālā, gəlāltā stone, rock.” Despite the gloss, the KS copyist may have read kolāh hat” under influence of words for headdresses and garments some lines earlier.

The Gypsy etc. argots of Iran and Central Asia also preserve traces of Aramaic verbs not in KS, some of which have equivalents in JLtr: for “give,” Ju hob-/how- and SamL hob-/hov- straightforwardly correlate with Shir av- and Bor ab- (with h-loss) from Aram (also Heb) hav “give.”

In the same semantic field, Kav zamon- to give” compares with Shir za(w)n-, Isf dam-, Teh dev- < Aram zabben (caus. of zəvan buy”), all “to sell.” Formally the comparison is enhanced by the possibility (noted by Yarshater, 1977) that the JLtr final merged with the Iranian causative marker (-ān- > -on-/-un-). For semantics, cf. Russian davatʾ “to give”: pro-davatʾ to sell,” and English “I’ll give it to you for five dollars.”

Ju ošin- “to take (up)” is closest to Her ošin-, cf. Gol (dar) ašun- to take,” Kash der ašan-, Gol dar āšne “take!” Neh be-m-āš-i “I brought, took.” Gol mi-āšun-am (stem āšun-) “I put right, set, fix, prepare, make, render” from āš-un to put right, set, fix, prepare, make, render” points to the stem āš- < Aram ʾ-š-š: ʾiššeš “founded, made firm,” cf. nitʾošeš was confirmed”; uššā “fortification.” Here note Ju ošišt-, past stem to ošin- (ošišt- assimilated < āš-ist-, past stem -ist-?).

Words of Aramaic origin not in KS or JLtr still preserved in the Pers argots are PG tub-, Chist tup-, Ju tuvok “to sit,” tavol- to seat” from Aram tūβ sit!”

To PG sak-, seg-, sig-, Ju sak- “to observe” < Aram sakkē observing” may be added be-sok “observe (carefully)” in the mainstream general colloquial of Isfahan (datum from Habib Borjian), a trace of the earlier Isfahani Loterāʾi.

PG tubur- “to break,” Ju tarb-, tarv- “to beat” (influence of Arabic ḍ-r-b “beat”?) < Aram tabber “to smash” find correlation in Abu Dolaf”s Arabic tatbīr (< Aram) (cf. Bosworth, 1976, II, p. 308).

Ju umor-, umošt- (SamL ůmor-, ůmošt-) “to say” (constructed like Tajik gumor-, gumošt- “arrange”) < *omor- ; AG bī-āmār “say” < Aram ʾāmar said” (vocalism like Ju oxol- “eat” from Aram āxal; cf. below Ju otor < otar “bazaar”).

Alongside the above verbs are attested nouns of ultimate JLtr Aram provenance; some interesting for their semantic development from Talmudic usage are Chist parzal knife” < Aram parzel “iron, iron implement of tool,” Ju givor = Tajik bosmačī (Russian basmač guerilla brigand fighter against early Russian Soviets in Central Asia”) < Aram gibbārā hero.” Cf. Mus Mehtar geevar ( = *givār) “man,” which would be a cross of the Aram word for “hero” with the word seen in Ju γavrik, havrik “man” < Aram gaβrā man”; cf. Mus Sāzanda gaveh “man.” Ju havrik, variant of γavrik, may be due to the influence of the word represented by Aram/Heb ḥaβer friend, associate.”

For PG “mezūl(?)” “fortune-telling” < *mezōl (Ivanow, 1914, p. 452, and, for the vowel of the second syllable, pp. 445-46) is < Aram mazzāla (Bab Amoraic Mishnaic mazzāl) “fortune.” (Note Ivanow, 1920, p. 282, on Persian Gypsies not being fortunetellers; Jews were known until recent times in the Near East as practitioners of occult sciences.)

Words of everyday economy include Abd otar, otor, Ju otor “bazaar, town” < Aram ʾātar place, town, market”; PG, PD parak “cow” < Aram par(ā) bovine, cow”; and Chist turun(k) bull,” AG tirang “bovine” (Pers gāv); Ju ozaxtor calf” (*“young bull,” cf. LuJ and PD āzak “child” [< Middle Pers zahāg?] < Aram tōr, pl. tōrān bull.”

Chist katuna clothing” < Aram kittūnā shirt, clothing” (KS tnwdh = Pers pirāhan shirt” may be miscopying of *ktwnh).

Old integration of Hebrew words from Jewish culture. The above large number of Aramaisms in the gentile argots, with and without correspondences in 20th-century JLtr, confirms the etymology of Loterāʾi etc. = non-Toraic = Aramaic, in reference to the chief source of the vocabulary. However, Loterāʾi had, even in its medieval phase, a number of Hebrew words as Jewish Kulturwörter.

For 20th-century JLtr, the agreement in Hebraisms shown between the JLtr of Iran and Herat (on the medieval distinction of which speech, cf. below on the numerals) in Shir melāxā : Her maloxo “work, action, affair” and Go lāšun : Her lošun speech” (< “tongue”) suggests a fairly early common origin; thus also Mash, Yazd lex- : Her le(y)x- “to go,” cf., in the exclusionary Jewish Neo-Aram of northwestern Iraq, līx leave quickly!, scram!”

Bor noma, Isf “nouma (nwmʾ), nummâ “moon” (respectively Ebrāhimi, p. 51 and Kalbāsi, p. 229, the latter glossed as Pers māh [borj], i.e., “moon as month”; neither entry designates the word as Loterāʾi) may be explained from Heb ləβānāh, via well-paralleled assimilations and contractions, whose result points to early adaptation.

Gol tanāim, Bor tanāyim (sic) “hen” is explainable as a metaphoric development, in reference to the hen’s putative reference to the rooster, from Heb tənāyīm “wedding arrangements,” an early sacral-ritual entry. The words in fact have Yiddish equivalents: məloxə, lošn, and leyx-ləxo (a topos from Genesis 1.12, and name of a Sabbath lection, Genesis 1.12-17.27) and tnoyim (“marriage contract”). Other Hebraisms may have been part of JLtr before the desuetude of Aramaic speech among Iranian Jews (cf. the occurrence of Hebrew in early Jewish documents of Afghanistan). Note also forms like Judeo-Kashani melōx-e hamōvet, Judeo-Isfahani melax movat = Yiddish malax ha-moves “Angel of Death.”

KS attests the Hebraisms bysh “egg” (Pers ḵāya-ye morḡ) < Heb bēṣāh id.; nʾr (miswritten nʾz) boy” (Pers ḡolām) < Heb naʿar “id.; and hzʾn caller, reciter, singer” (Pers vānda [for *vānanda]) < Heb ḥazzān (synagogal) announcer, precentor, cantor.”

Interestingly, KS kymwlw “camel” (Pers oštor) also in br kymwlw “camel driver” (with br < associative Aram bar, cf. e.g. bar ḥaylā soldier”) and kymwlw mlkʾn *“camel-render” (Aram malq-, see above) = “elephant” (Pers pil). Whereas Aram gaml- is relevant here, kymwlw = */gimōlō/ < Heb gəmallō “his camel,” cf. Gol gamelli < Heb gəmallī my camel,” with a different possessive suffix. For *gym-, cf. Early Judeo-Persian nymʾz < namāz prayer” and nymyk < name(h)k “salt” (in a tafsir of Ezekiel).

KS ʾbyk father” compares with Shir abeq “father” < Heb ʾaβīx(ā) thy father”; vs. Khom ābi father” < Heb “my father”; cf. the variation underlying possessive suffixes in the word for “camel.”

SamL dela “house, door,” Ju, Abd, Ar dela, dila, Luli dila “house, tent,” AG dila “room, tent” may derive from Heb delet “door,” as does Bor delét “door.” Possibly via reanalysis as dela-t *“thy door,” or, if from early Heb pronunciation, deleθ became *deleh when Old Pers θ went to Mid. Pers h, resulting finally in dela. For semantics, cf. Pers and Tajik dar “door, gate, court”?

PG nidu, nodo, nedeo, LG Osof and Kermān nedow “woman, wife” seem to presuppose *nidō, which could derive from Heb (> Aram) niddāh menstruation, menstruant,” if KS and manuscripts of Asadi’s Loḡāt-e fors dnh “woman” (Pers zan), Abd dana, PG/PD danew, denew, dinki, and LuJ dānu, Mus Mehtar danow, Mus Sāzanda duneh, Mus Torbat-e Jām danow, LG Fars danow represent a taboo metathesis < nidā, *nidō. Influence of Arabic ḏanab tail” is possible (cf. Vulgar English slang “tail” for “woman as sexual object,” e.g., “chase tail”), cf. Abd danap, PG/PD deneb, Ju danam, Kav danap, etc. “woman.” 

Both Aramaic and Hebrew were alternative sources of older JLtr, as reflected by the names of numerals. Aram ḥad one” gave KS hʾdk “one,” PG hōt “a unit”; counting suffix -hōd, -hōt (= Pers ); so also AG (h)āt, (h)od, (h)ot. The latter forms compare with *-hat- < Aram ḥad in Her təreynatak “two” < Aram tərey(n) “two” + *hat + -ak; the segmentation tərey-natak gave the Her numerative -natak. For “two” Heb šənayim, šəney is reflected by KS šym < *šnym, SD šann (Bosworth, 1976, II, p. 313); cf. Shir. šane, Bor šené. KS has slws = *šlwš < Heb šālōš “three,” cf. Shir šalošā, but KS has štʾ from Aram šittā six,” and ʾmʾ < late Aram ʾəmmā thousand.”

Evidence for a Jewish underworld component in the old gentile argots. The actual involvement of a Jewish underworld in the early formation of the Muslim argot is confirmed by the data of the 10th-century Banu Sāsān qaṣida (AD), together with later data. Aram hādōr “circle” and hādōrā peddler, beggar” (both from the root h-d-r “to go around”) are involved in the first example: AD hʾδwr/hʾdwr “the circle [of fortunetellers and their shills operating in a street assemblage] about which people congregate” (see Bosworth, II, pp. 240-41, with a different Semitic etymology). KS hʾdwr “job” (Pers kār) may refer to this charlatanry, or professional begging, like PG khōdur, SamL, Ar, Abd hodur, etc. “beggar,” SamL hodůri begging”; cf. Mag ādur “peddling,” the chief occupation of the Sheikh Momadi (Moḥammadi) itinerants of Afghanistan (Petrusińska, p. 48, with literature), with loss of h- as in Mag az- look, see” (above, s.v. KS hʾzydn).

AD barkakk is “street dentist” (Bosworth, 1976, I, pp. 90, 146, 148, 161) < Aram bar (see above on KS br kymwlw) + kākkā (molar) tooth,” whence KS kʾkʾ tooth” (see above). Privative bar (cf. KS brmyʾ above) is seen in AD brkwš one pretending deafness” (idem, pp. 161, 175) = KS br kwš, where kwš = Pers guš “ear.”

AD maysarāni (verb maysara) “someone who begs, pretending to have fought the infidel on the frontier” (Bosworth, 1976, I, p. 175; II, pp. 194, 224), KS mysr “fighter for the faith” (Pers ḡāzi), < Aram mēyṣar “border,” mēyṣarānā pertaining to the frontier.”

In the expected field of obscenities are hurr “rump,” KS hrh (Pers kun) and kyδ penis” (Bosworth, 1976, II, pp. 192, 218-19), KS kyt = *gyt *“penis” (Pers dnd dand rib,” probably misreading of *kyr = kīr penis,” near other words pertaining to the abdomen and genitalia; cf. Ju git “penis, male,” Aram gīd vein, tendon, penis”). 

The foregoing correlations of the AD with KS again show the Jewish Aramaic of the Muslim argot, already in the 10th century.

Note also Bosworth, 1976, II, p. 210, on AD verse 187, where travelers who go from place to place to lay out rugs are equated with al-mašāṭiḥ; this is derivable from Aram mišṭāḥā land where something (e.g., a fishing-net) is spread out to dry”; cf. Bosworth, 1976, II, p. 279, with a comparable Hebrew root and derivative. The Arabic indicates a direct borrowing from the Aramaic, as against the h of hurr “rump,” whose h is due to Persian intermediation.

Bosworth (1974) tried to make a case for the presence of Jews in the Banu Sāsān, arguing from both a priori considerations and scant textual testimonia which he notes are uncertain as evidence. The correctness of his overall case is now confirmed by the linguistic evidence presented here.

Tenth-century Astarābād and the early diffusion of Jewish jargon: evidence from early twentieth century “Djougi.” It is in the 10th century (the period of Abu Dolaf), we recall, that the term lwtrʾ first occurs with regard to Astarābād, without Jewish reference. That this area was indeed instrumental in the entry of the Jewish exclusionary vocabulary into gentile argot is shown by de Morgan’s (pp. 304-6) linguistic material of “Djougi” Gypsies of Astarābād, the city in which Lō/uterā(ʾi) is first attested by the Ḥodud al-ʿālam in the 10th century. It is likely that, at least in part, this “Djougi” group described by de Morgan as impoverished migratory tent dwellers, who seasonally migrate outwards from Astarābād and return there, at least in part represent the group from which are descended the identically-named Jugi Gypsies of Tajikistan, as indicated by such unique correspondences as Djougi “homoachtan,” Jugi umoštan (ůmoštan), umor- (ůmor-), cf. AG āmār- “to say” < Aram āmar-

The vocabulary of Djougi, as given by de Morgan, has close equivalents to other Gypsy argots of Iran and Central Asia in general:

From Indic: môness “man”; djévéd “woman, wife”; “iron”; pounó “water”; bohót- “big”; vagal “goat”; gôrá horse”; gérà donkey”; bedjalonen “to light something up”; -khez, -khiz “kinship suffix”; etc.

From Arabic: khashpouk stick of wood”; ghèlil small,” cf. kölèl child”; nárák fire,” etc.

From Iranian: dakhlodj girl”; takhnoï knife”; süthaï “charcoal” (probably a Caspian reflex of suxta- “burned,” cf. Djougi southa = Tajik argots suta “black,” but note KS swdʾ < Arabic); and hedjonddan to make” (= PG ajon-, Ivanow, 1914, p. 454 with pp. 447-48 on the suffix: cf. above on Shir aj-, Her huj-).

From Aramaic/Hebrew: dakhana good”; modakhî bad” (mo- < Pers “not”); nouhour “eye”; hakhaliden “to eat”; miokholî “thou eatest”; bekimin to go”; hamoachtan “to speak, say” (see above); galout “bad”; daghno mouth”; -hot/-hod numerative suffix; and dela “house.”

Now, in addition to these forms ultimately from Aram/Hebrew which are found throughout the Central Asiatic argots, Djougi has words from Aram/Hebrew not found in the other gentile argots. These are: pichto “easy,” cf. Aram pəšiṭā id.; moda ana ewe” (Pers māda “female” + Aram ʿānā sheep”); nomárát night” (< (a)rāt night, evening” < Indic + disambiguating noma = Bor, Isf noma, nu(m)mā < Hebrew ləβānā, discussed above); and further forms:

Djougi thünoï hen” (cf. tunoï “egg,” probably parallel to Pers toḵm-e morḡ seed of chicken” = “egg,” cf. Pers toḵm “seed; egg; testicle”). This t(h)ünoï would represent something like /tənoʾī/ < */tanāʾī/, which, with the common denasalization after long vowels, compares with Gol tanāim, Bor tanā(y)īm (sic; confirmed by Prof. Yarshater in a letter of 2011). Hen,” whose Hebrew origin is discussed below, KS tnʾγwl (Pers morḡ), Abd and Ar tanaγul chicken,” emerges as a cross of *tarneγul, cf. Bor tarnegul < Aram tarnəgōl [tarnəγōl] “rooster,” and tanā(y)i(m) hen,” whose antiquity in gentile argot is reflected by Djougi.

Of the four unique Jewish Semitic words in Djougi, two have correspondences in JLtr. With the addition of hedjonndan with correspondences in PG of northern Iran (with and without causative -on-), cf. Shir āj- and Her huj- from fossilized Old Iranian, we have confirmation for the antiquity of the argot in Astarābād, for which the term lwtrʾ(y) is attested in the 10th century. These data also confirm the Jewish origin of lwtr(ʾ)(y) as both term and speech, referred to in early Persian sources.

The role of Deylam in the early diffusion. As for the passage of the exclusionary Judeo-Iranian speech into non-Jewish argot, specifically in Astarābād in the 10th century, the latter city was variously connected with Deylam (see Deylamites ii) during this unstable period, which gave rise to two dynasties of Deylami origin, the Ziyarids and more importantly the Shiʿi Buyids, whose power soon extended over most of Iran, setting the scene for peregrinations of Abu Dolaf al-Yanbuʿi and his patronage among the Buyids.

In this context it is relevant that KS refers to Deylam as mlkyʾ = Aram malkayyē “the kings,” cf. KS mlkʾ, glossed as Pers amir. Furthermore, a Deylami locus of the spread of the argot may be confirmed by KS dʾ = Pers deh “village,” if this represents the Māzandarāni pronunciation (of which Habib Borjian has informed me). KS mʾhwz mlkʾn “city of amirs” = “Ray” seems also to reflect Buyid realia. The role of a Jewish underworld―noted above for the 10th-century period of Abu Dolaf on linguistic grounds―may well have been furthered in post-Buyid Deylam via the activities of Ḥasan Ṣabbāḥ, among whose fighters were three communities of tough, rapacious Jews (thus Benjamin of Tudela). It could be expected that such groups settled in cities as a prestigious part of the Iranian underworld.

Jews and argots, East and West. The sociological scenario for the beginnings of gentilization of Loterāʾi would have been the emergence of a large Iranian underworld culminating in the 10th century in the Caspian area, of which Jews were a part. Whereas such Jews inherited a vocabulary meant originally to exclude non-Jews (the core of which speech persisted in this function among respectable Jews), for Jews of the Iranian lower class (including poor tradesmen and laborers), this vocabulary was intended to exclude the solid citizens, police, gentry, and rulers. For the non-Jewish underworld, their Jewish colleagues offered a readymade exclusionary vocabulary, whose acquisition was useful. This vocabulary was then transmitted to other marginal groups, particularly Gypsies, and is still partially attested. A very similar development took place independently in the 15th-16th century rise of Gaunersprache/Rotwelsch (Rotwälsch), with its large Jewish vocabulary, among gentile beggars and scoundrels of the German-speaking area. Indeed, Oranskiǐ (pp. 44-45) notes a brief series of gentile Iranian argot words of Jewish etymology, and on p. 46 with fn. 37 adduces Ju oxol-: Rotwelsch acheln “to eat.” Note further such correlations with early 16th century R(otwelsch) as Chist ganavidan, etc., R genffen “to steal”; Ju γavrik, etc., R gaver man”; and KS bysh “egg,” R betzam eggs.”

Early Jewish Loterāʾi reflected in the scope of the Persian vocabulary. Early Persian poetry and lexica contain a number of words of Jewish Aramaic origin via our Muslim argot. These include the aforementioned lwtrʾ words in Suzani’s poems, dx “fair,” zyf “vile”; words noted as used by the *sāsiān (mss. ʾsyʾn): sʾbwth, ṣʾbwth old woman,” dnh “woman”; and kākā “tooth”; words redolent of the underworld: hār feces” and hrh “rump”; the important word hʾdwry “member of a class of intrepid beggars”; and words which still survive: nhwr (old vocalism nuhōr, cf. Aram nəhōr(ā)) and Tajik nuhūr sight, eye”; and šeydā crazy, wild, infatuated, lovesick, enamored” (in Persian literature).

The Aramaic provenance of such words seems indicated in the two tags “suri” for the argot verses at in the final portion of the KS. This term is to be taken as the equivalent of suryāni (*“Syrian”), which was the usual term for the Aramaic language among Arabic-speaking Jews and others. Already in the Hellenistic period the similar Greek words Syriakḗ and syristí were used by Hellenistic Jews for "Aramaic" (the latter word > *swrysṭyn in the Palestinian Talmud, 4th-5th centuries CE). Syriac does not come into consideration; although Syriac is similar to the Babylonian Jewish Aramaic  source of the argot, it is in fact distinguishable by words like Syriac nuhrā vs. Aram nəhōrā “eyesight,” Syriac gīd tendon, vein,” but Aram also “penis,” a Jewish usage reflected in the Muslim argot. 

Conclusion. Loterāʾi may now be seen as not only a long-lived Judeo-Iranian speech, but a speech which exerted a remarkable influence on the course of Persian sociolects and Persian in general, with a history which illuminates the account of interethnic relationships and class in the Middle East and Central Asia.

(I thank Mahmoud Omidsalar for his invaluable remarks on KS, and, for helpful provision of suggestions and bibliographic materials toward the preparation of this article, Habib Borjian, Ken Blady, Agnes Korn, Tatiana Oranskaia, Ludwig Paul, Nahid Pirnazar, Houman Sarshar, and Ehsan Yarshater.)



Sekandar Amanolahi and Edward Norbeck, “A Note on the Secret Language of the Traditional Musicians of Iran,” Journal of the Gypsy Lore Society, Ser. 4, 1/4, 1978, pp. 283-86.

C. E. Bosworth, “Jewish Elements in the Banū Sāsān,” Bibliotheca Orientalis 33, 1974, pp. 289-94.

Idem, The Mediaeval Islamic Underworld, the Banū Sāsān in Arabic Society and Literature, 2 vols., Leiden, 1976.

Ayub Ebrāhimi, Eṣfahān neṣf-e Jahān. Farhang-e vāžahā wa eṣṭelāḥāt-e maḥalli-e Eṣfahān, 2nd ed., Los Angeles, 2006.

W. Ivanow, “On the Language of the Gypsies of Qainat (in Eastern Persia),” in Journal [and Proceedings] of the Asiatic Society of Bengal, N.S. 10, 1914, pp. 438-53.

Idem, “Further Notes on Gypsies in Persia,” Journal [and Proceedings] of the Asiatic Society of Bengal, N.S. 16, 1920, pp. 281-91.

Idem, “An Old Gypsy-Darwish Jargon,” Journal [and Proceedings] of the Asiatic Society of Bengal, N.S. 18, 1922 [1923], pp. 375-83.

Idem, “Jargon of Persian Mendicant Darwishes,” Journal [and Proceedings] of the Asiatic Society of Bengal, N.S. 23, 1927, pp. 243-45.

Irān Kalbāsi, Guyeš-e kalimiān-e Eṣfahān, Tehran, 1994.

Ḡolām-Ḥosayn Karimi, “Lwtr-e Jāberi,” Majalla-ye zabānšenāsi 7/2, 1990, pp. 64-68.

Gilbert Lazard, “Note sur le jargon des juifs d’Iran,” JA 266, 1978, pp. 251-55.

Ḥanina Mizrahi, Yehudey Paras [Jews of Persia], Tel Aviv, 1959.

M. Moḡdam (Moqaddam), Guyešhā-ye Vafs o Āštiān o Tafreš, Tehran (Irānkuda 11), ca. 1960.

J. de Morgan, Mission scientifique en Perse V. Études linguistiques, Paris, 1904.

I. M. Oranskiĭ, Tadzhikoyazychnie etnograficheskie gruppy Gissarskoĭ doliny Srednaya Aziy, Etnolingvisticheskoe issledovanie, Moscow, 1983.

Ludwig Paul, “Geheimsprache von Kahak (Tafreš),” Orientalia Suecana48, 1999, pp. 105-114.

Jadwiga Pstrusińska, O tajnych językach Afganistanu i ich użytownikach [On the secret languages of Afghanistan], Cracow, 2004.

Houman Sarshar, ed., Jewish Communities of Iran. Entries on Judeo-Persian Communities Published by the Encyclopædia Iranica, New York, 2011.

Anna Leonidovna Troitskaya, “Abdoltili―tsekha artistov i muzykantov Sredneĭ Azii,” Sovetskoe Vostokovedenie 5, 1948, pp. 251-54.

Gernot Windfuhr, “Gypsy ii. Gypsy Dialects,” Encyclopædia Iranica XI/4, 2002, pp. 415-21.

Ehsan Yarshater, “The Hybrid Language of the Jewish Community of Persia,” JAOS 97/1, 1977, pp. 1-7; abridged in Encyclopædia Iranica XV/2, pp. 156-60.

Idem, “The Dialect of Borujerd Jews,” in L. Meyer and E. Haerinck, eds., Miscellanea in Honorem Louis Vanden Berghe, Ghent, 1989, pp. 1029-46.

I. I. Zarubin, “O yazyke geratskikh evreev,” Doklady Ross. Akad. Nauk, 1924, issue no. 4, pp. 181-85.

(Martin Schwartz)

Originally Published: September 20, 2012

Last Updated: October 8, 2012