THE SOUXnS or SI'IHUIl 6I9 



\o\\rl as to (hf rl).ini;<.' ill llu' s|H'.ikfr, .iiul ,i >imil.ir cniuhisidii will 

 pr()l)al)ly hi' riMclu-il lor iMch of tlu- oiIht tiirru M-mi-vowrl sounds. 



I'roiii the oviiii'iici- in the records, it is dil't'iriiit to siihsrrilH- entirely 

 to a "transient" theory of these sounds, at least when they precede 

 the standard \o\vel sounils. The evidence justifies the use wiiich has 

 lieen niaiie of the steady-state itlea, and the harmonic analyses leading; 

 to a determination of characteristic frequencies. IJut there is a 

 (M)ssil)ility that the harmonic analysis does not tell the whole story. 

 These two groups of records and the acoustic spectra based on them 

 furnish outstanding examples of the niceties in\olved in speech and 

 hearing in order to achieve the miracle of articulate speech. Without 

 harmonic analysis, the most casual observer will note, for example, 

 the similarity between the corresponding records of the / and » sounds, 

 but more astonishing still is the resenil)lance between the I and ee 

 sounds shown together in Plates Nos. 107 and 108. In this latter 

 case (,/ and ee) practically the same high and low characteristic fre- 

 quencies are involved, and it would seem that the distinction, which is 

 sufficiently pronounced to the ear, must l)e based to some extent not 

 only on the relative amplitudes of these frequencies present, but also 

 on the behavior of these amplitudes during the fundamental cycle. 

 It will be noted in practically all of the records of these semi-Nowel 

 sounds that the high frequency characteristic is a transient of more 

 rapid decay than in the case of the pure vowel sounds; it is not of 

 large amplitude except at the beginning of the cycle. On the face 

 of the records this is the only explanation available for whatever dis- 

 tinctive quality these sounds, as a class, must possess. 



VI 



Si.XTEiiN Consonant Solnds 



The last two groups, X\'I and X\'II contain, respectiveK", records 

 of the "hard" and "soft" consonant sounds, each with the a sound 

 affi.\e<l, and pronounced by the two male speakers. Here the classifica- 

 tion is somewhat arbitrary; it is difficult if not impossible to arrange 

 the sounds of these two groups in any such satisfactory series as has 

 been determined for the semi-vowels of the two preceding groups. 

 The sounds dth (that) and th (thin) for example have transitional 

 characteristics that relate them to both groups; but they are placed 

 at the end of Group X\I, to emphasize their relation to the pair v/f 

 of the last group. With these reservations as to arrangement, consider 

 the general characteristics of the consonant sounds of these two 

 groups. 



