128 Mr. Latham's Facts and Observatiotis 



tween the semivowel sounds of 1/ (as in i/a) and the mute 

 sound of g (as in ga). 



The sound of 10 (as in xva) is to the sound of v (as in va) 

 in a certain degree of relation ; and, through this, it is allied 

 to b (as in ba), in a maimer determinable by the relation of 

 V and b : and, again, ?/ (as in i/a) is to g (as in ga) as xv is to 

 V. Hence it arises that the sound which shall be to^ as v is 

 to b, shall be to g and j/ as t; is to 6 and w. Expressed formally, 

 the fact stands thus : 



Gx : g and y : : v : b and xo. 



In a short tract, pubHshed in 183.5, I stated that the sound 

 in point must be on the " ^/-side of g," and that, although a 

 ludicrous mode of speaking, expressed my meaning fully. 



The conditions then o/Gx are, that it must be in a certain 

 relation to g and // ; in which relation I am satisfied that gh 

 is not. 



It may not be thought superfluous to remark, that the two 

 last assertions, that is, the assumptions of the question in hand, 

 are made irrespective of any previous reasoning. I consider 

 that they rest upon the evidence of our sense of hearing. 

 Expressed formally, the facts and inference are as follows : 



Gx : g and y :: v. b and tw . which gh is not. 



Whatg/i is, is another question. Unless the number of mutes 

 is to be reduced to four, sounds like b, j), f, v, &c. &c. must 

 be considered not as varieties of a given typical sound, but as 

 sounds specijically distinct : and, again, unless the number of 

 mutes is to be indefinitely increased, sounds like the t and 

 d cerebral of the Hindoos (I speak in the way of illustra- 

 tion) must be considered, not as sounds specifically distinct, 

 but as varieties of a given typical sound. 



To come, then, under the necessary conditions, the sound of 

 Gx must be specifically different from that of G. Now Gh 

 (as may be shown at another time) is no specijic sound, but 

 only a variety. 



The force of the words specific and variational depends 

 upon their definitions. To say that [> is a variety of t is true or 

 false, according to the definition of the term variety ; and the 

 assertion is objectionable only on the score of inconvenience. 

 To say, however, that \ is a variety of s (as is often said by 

 those who are strangers to the sound) involves something 

 more than an inconvenient definition. It is an error in re- 

 spect to the fact. 



Having (to my own satisfaction at least) inferred that Gh 



