or : mah. 1858-59.] for the Indian characters. 261 



ticable, they are ready to admit all the advantages that have been 

 attributed to it, and to abandon all the objections that have been 

 hitherto mentioned as invalid ; but they are persuaded of the in- 

 adequacy of the Roman character to the expression of the Indian 

 sounds, and on this ground they reject it as impracticable. 



Until lately I held this view myself, but, like most other people, 

 I presume, without examination. A few years ago, however, I 

 found it necessary, in preparing for the Press a Comparative Gram- 

 mar of the Dravidian languages, to represent a great many words 

 belonging to those languages and some Sanskrit ones in the Ro- 

 man character, for the sake of facilitating comparison and keeping 

 down expense. I had no intention of experimenting upon the 

 adequacy of the Roman character for the purpose in view, and 

 some of the details of the system which I pursued were not 

 thoroughly satisfactory ; but I did the best I could under the cir- 

 cumstances, and as I went on I was gratified to find that the 

 use of the Roman character was much more feasible than I had ♦ 

 supposed. I met with no difficulty that appeared to be insuper- 

 able, and such difficulties as I did meet with were wholly owing 

 to the absence of an authorised system. 



Let us see then whether the alleged impracticability of the 

 scheme will hold its ground when carefully examined. I will en- 

 deavour to show, first, that the majority of the Indian sounds can 

 be expressed by the Roman characters without any change what- 

 ever, and then, that the Roman character can be enabled to ex- 

 press the remainder by the help of a very small number of accents 

 and points. In doing so, I will endeavour to help forward the 

 definitive settlement of details, a point of great present import- 

 ance, by giving the particulars of the plan which now commends 

 itself most to my own mind. 



(1.) The number of letters in the Indian languages which can- 

 not be transliterated, just as they stand, without the addition of 

 any accents or points, is in reality very small, and the number of 

 such letters is smaller in some languages than in others. 



The following vowels and consonants of the Deva-Nagari alpha- 

 bet, and of the alphabets which follow the same arrangement, can 

 Vol, xx. o. s. Vol. v. n. s. 



