Vol. X, No. 11.] The Language of the <ri/psies of Qainat. 445 



[N.S.] 



i 



suffices to debar it from being classed as a language. It is at 

 any rate acquired naturally by Zori children, as a language for 

 the home circle." 



There are many words in Gypsy language which are used 

 in the secret code of derwishes, beggars and, I believe, thieves. 

 These words are often artificial, "symbolical", even drawn 

 from several other tongues. 



The QainI idiom being the basis of local Gypsy jargon is 

 a link in the great chain of Iranian dialects, at one end of 

 which are the languages of the Pashtoo family, at the other 

 those of Kurdish. The idiom in question is closely related to 

 that of Turshiz (or Turshish) and Sebzewar, and shares many 

 words with the dialect of Tabas. This last is in a transition 

 stage towards the dialects of Biabunak, Nain etc., which prob- 

 ably belong to the same group as the dialect of Kashan. Many 

 of their phonetic rules govern also Kurdish dialects and can be 

 met, as far as I know, in the idioms of Aorami and Kendule, 

 which belong to the Tajik family, the direct heir of Zend. 



Phonetics. 



The chief peculiarities of this Gypsy idiom ; i.e. as spoken 

 in Qainat, are as follows : — 



The guttural n as in degnb. The use of cerebrals / and 

 d as in til. The r sometimes recalls the r of Hindustani, e.g. 

 in the word bard ; k and g are more velar than in Persian, 

 although not so strong as Persian g or gh. Their strong pro- 

 nunciation is retained even when they are palatalized (I mark 

 the palatalization with the sign). Dentals are usually softer 

 than in QainI. The n on the end of words is often pro- 

 nounced as a guttural n or deep Hindustani n, and very seldom 

 as Qaini m = w. 



Palatals are much softer, simiiiar to those in Hindustani ; 

 ch and j are sometimes undistinguishable. 



The sibilants are also pronounced softer. 



The labials can become a little aspirated, as bhutok, but 

 they vary very much according to individual pronunciation. 

 B and w, p are often interchangeable. Ex.: Ube = biwe, 

 faqwim = tagbin , sefid = ispid , etc . 



The liquids / and r seldom differ at all in pronunciation. 

 They very often cause the transposition (metathesis) of syllables. 

 Besides there is always some thing like a softly sounding r or ri 

 after every final vowel : goror, chetir, etc. 



The sound "a" is more guttural and deeper, than in 

 QainI, e.g. —lakar, etc. But long a is pronounced always as in 

 Q.— i.e. as 5, a sound which can never be confounded either 

 with o=# in North Persia (mostly Tehran and Azerbeyjan), 

 or with the Indian pronunciation of P. u, as in koh = P. kuh. 

 As in Q. a long d ( = o) at the end of a word becomes often eo 



