1867.] Proceedings of the Asiatic Society. 41 



they are, they must have been taken from the Tamilians. But the ma- 

 jor premise in this argument is not tenable. The cerebral letters used 

 in the Sanscrit are r, r, sh, t, th, d, dh, and rju Of these, r and sh are 

 common to all the Arian languages, and that is enough to shew 

 that the general premise is founded on a mistake, and the deduction 

 from it consequently cannot be accepted as true. It is possible some 

 may tell me that by cerebrals Messrs. Caldwell, Norris and Thomas 

 allude to t th d dh and n, and not to all the letters of that class. 

 This shifting of the ground would scarcely be fair in argument, but 

 accepting the premises on this narrow basis, I think there is not proof 

 sufficient to support it. We know not whether the old fire-worship- 

 pers pronounced their t as T> and not ^, nor do we know the sound 

 that letter had among the Greeks and Romans, for the Greek as 

 pronounced now is not the Greek that was, and were old Homer to 

 appear among the dons of Oxford or Cambridge, he would be almost as 

 unintelligible to the Porsons of our day, as he would be to the people 

 of this country. Leaving the Zend, the Greek and the Latin as 

 uncertain, if we turn to the Teutonic and the Sclavonic, we find the 

 cerebral consonants by no means unknown. The Low German along 

 the shore of the Baltic has them, and they are dominant in the 

 Scandinavian, the Russian and the Lithuanian. In the English 

 the ^ is unknown, and, notwithstanding the dictum of grammarians 

 that the English t was a dental, it is rarely that an Englishman can 

 pronounce the sound of 3>. With him T> is the only letter known, and 

 he uses it both for T> and ^5. Mr. Norris in his paper on the 

 " Scythic Tablets" of Behistun, accounts for the presence of t (l?) in 

 the Scandinavian and the Icelandic, by supposing it to have been 

 borrowed from the Lapp — a Tartar language ; but I imagine he will 

 not try to assign to the same cause the origin of the English t. Were 

 he to do so, he would have to prove, in the first place, that nations can 

 borrow sounds, and secondly, that the Anglo-Saxons really did so. 

 It is well known that physical and social causes may lead to the 

 loss of certain sounds in a language. The Brahminic Arian original- 

 ly had a guttural q, which the enervating influence of India soon 

 softened down to the modern 3>. In our own day, the Persians and 

 Moghals in Bengal lose the guttural <3 in the course of a single gene- 

 ration. Aspirates and compound consonants are being constantly 



