1897 - 98 .] Mr R. J. Lloyd on Consonant- Sounds. 
239 
binations with vowels, which are not only in themselves the most 
important combinations, but which also serve, for the reason above 
indicated, to facilitate a subsequent study of those consonants 
which are always and essentially of a gliding nature. 
Limited as our problem here is, three classes of cases at once offer 
themselves for solution : (1) that of a spirate fricative beginning 
from silence and ending in a vowel, (2) that of such a fricative 
leading from a vowel to silence, (3) that of it joining two vowels. 
The last case varies greatly. If the two vowels connected are 
identical, this case is simpler than the other two : if not, it is more 
complex. Let us start from the simplest case, remembering always 
that the principle which rules the process of articulation is economy 
of adjustment — so long as that adjustment suffices to attain the 
sound desired. 
Before attacking this problem it is necessary to lay aside certain, 
prepossessions, derived from orthography, which, from an acoustic 
point of view, are more or less misleading. We are apt to think 
that the combinations asa or oso contain but three sounds each, 
a-s-a and o-s-o ; and the ear seems to confirm this impression. But 
a little consideration shows that there are really five sounds in 
each case : there are not only two vowels and an intervening con- 
sonant, but also two brief, yet inevitable, gliding sounds, the one 
connecting the first vowel to the consonant, and the other connect- 
ing the consonant to the second vowel. When the pace of articu- 
lation is sufficiently relaxed, the existence of these glides becomes 
quite evident to the ear ; and the reason of their existence becomes 
evident too. The posture of the organs in the a or o articulations 
is very different from their posture in s. In passing from a or o 
to s , two principal movements have to be carried out — a complete 
opening of the larynx, and a raising of the tongue-tip some 12 or 
20 mm. These movements are, of course, attempted simul- 
taneously. It is the duration of the longer one, therefore, which 
determines the duration of the glide. In this case the tongue- 
adjustment probably takes the longer time of the two. There is 
generally an instinctive effort to make glides as short as possible. 
They are by-products, necessary evils, which the speaker produces, 
one may say, against his will, and stifles as far as he is able. As 
a rule, he succeeds in making them so brief that they are not 
