236 On writing Oriental words [No. 10, new series, 



some) objection. But when we desire to frame a romanised al- 

 phabet to include Tamil and Telugu, it is absolutely necessary to 

 distinguish the long e and o from the short, as in the above men- , 

 tioned two languages there is a short e, and a short o. I think 

 the error must have arisen from a mis-print, and should at once 

 be rectified. I made some observations on this point in my 

 " Memo." (p. 33 of the printed Pamphlet.) 



Secondly. I object to the same letter and symbol, namely " % s," 

 being used to represent the Sanskrit SJ, and the Arabic ^ fC . The 

 Report in p. 7 quotes Sir W. Jones, but he only speaks of a pro- 

 bable resemblance, and I think no Oriental scholar of the piesent 

 day would consider the two letters similar in sound. The San- 

 skrit letter has a sound of " sh" in it, which the Arabic letter has 



not. It is rendered in Hindustani words by , * (vide Shak- 

 en m v . 



spear's Dictionary and Grammar) never by It is rendered in 



Telugu by in which a sound of " sh" is recognized. It is the 

 first letter of such words as " Shiva," " Shudra," " Shrotriyam," 

 " Shaster," thus exhibiting a sound quite different from I 

 would therefore retain 's for ^ as the Report proposes, but ren- 

 der W\ by 's according to the scheme of Sir W. Jones, and most 

 modern Orientalists. I would next point out what I think is an 

 omission in page 11 of the Report. It is admitted that in- 

 Tamil must be rendered, not with two ds but as tt; also pp as 

 t't; but no provision is made for which is decidedly — ch, 

 and /E/65=ng, ^^^nj, £_^=tch. I would propose to add these. 



I quite admit that theoretically it is a good rule to disallow a 

 second Roman character for the same Oriental letter, and that it is 

 better to explain in the scheme that such and such a letter is 

 sounded differently in certain positions, the same as a learner is 

 taught when studying the original language ; but still in a Glos- 

 sary for instance, written in Roman character every one is not 

 supposed to know what the original word is. Thus in the Te- 

 lugu word " pampu" (a field,) the reader would probably pro- 

 nounce the last p the same as the first, and rightly ; but in 

 the Tamil word " pampu" (a snake), he would not be aware, 

 (unless he knew in what language the word was in the original, 



