OF ARTICULATE SOUNDS. 467 



615. It is not very difficult to produce a tolerably good artificial imitation 

 of the Vowel sounds. This was accomplished l>y Kempelen, by means of an 

 India-rubber ball, with an orifice at each end, of which the lower one was 

 attached to a reed; by modifying the form of the ball, the different vowels 

 could be sounded during the action of the reed. He also employed a short 

 funnel-like tube, and obtained the different sounds by covering its wide open- 

 ing to a greater or less extent. This last experiment has been repeated by 

 Mr. Willis; who has also found that the vowel sounds might be imitated, by 

 drawing out a long straight tube from the reed. In this experiment he arrived 

 at a curious result: with a tube of a certain length, the series of vowels, i, e, 

 a, o, u, was obtained, by gradually drawing it out; but, if the length was in- 

 creased to a certain point, a further gradual increase would produce the same 

 sequence in an inverted order, u, o, a, e, i; a still further increase would pro- 

 duce a return to the first scale, and so on. When the pitch of the reed was 

 high, and the pipe short, it was found that the vowels o and u could not be 

 distinctly formed, the proper tone being injured by the elongation of the 

 pipe necessary to produce them ; and this, Mr. Willis remarks, is exactly the 

 case in the Human voice, most singers being unable to pronounce u and o 

 upon their highest notes. 



616. The most natural primary division of the Consonants is into those 

 which require a total stoppage of the breath at the moment previous to their 

 being pronounced, and which, therefore, cannot be prolonged ; and those in 

 pronouncing which the interruption is partial, and which can, like the vowel 

 sounds, be prolonged ad libitum. The former have received the designation 

 of explosive; and the latter of continuous. In pronouncing the explosive 

 consonants, the posterior nares are completely closed, so that the exit of air 

 through the nose is altogether prevented; and the current may be checked in 

 the mouth in three ways, by the approximation of the lips, by the approxi- 

 mation of the point of the tongue to the front of the palate, and by the ap- 

 proximation of the middle of the tongue to the arch of the palate. In the first 

 of these modes, we pronounce the letters b, and p; in the second, d and t; 

 in the third, the hard g, and k. The difference between b, d, and g, on the 

 one hand, and p, t, and k,* on the other, seems to depend on this ; that in 

 the former group the approximating surfaces are larger, and the breath is sent 

 through them more strongly at the moment of opening, than in the latter. 

 The continuous consonants may be again subdivided, according to the degree 

 of freedom with which the air is allowed to make its exit, and the compression 

 which it consequently experiences, i. The first class includes those, in which 

 no passage of air takes place through the nose, and in which the parts of the 

 mouth that produce the sound are nearly approximated together, so that the 

 compression is considerable. This is the case with v andy, which are pro- 

 duced by approximating the upper incisors to the lower lip; and which stand 

 in nearly the same relation to each other, as that which exists between d and 

 t, or b and p. The sibilant sounds z and s, stand in nearly the same relation 

 to each other; they are produced by the passage of air between the point of 

 the tongue and the front of the palate, the teeth being at the same time nearly 

 closed. The simple sound sh is formed, by narrowing the channel between 

 the dorsum of the tongue and the palate; the former being elevated towards 

 the latter, through a considerable part of its length. If, in sounding s, 

 we raise the point of the tongue a very little, so as to touch the palate, the 

 sound of / is evolved; and in the same manner d is produced from z. This 

 class also includes the th ; which, being a perfectly simple sound, ought to be 

 expressed by a single letter, as in Greek; instead of by two, of which the 



* For the sake of proper comparison, this letter should be sounded not as kay but as key. 



