QUALITATIVE ANALYSIS. 49 



is now being made to determine if this double diphthongization is a 

 regular characteristic of the pronoun "I." 



So much has been said of the complexity and the variability of the 

 speech curves that the impression may have been produced that they 

 are hopelessly irregular. This is not true. They are as irregular as the 

 leaves of the trees; no two are exactly alike, yet the individuals of a variety 

 resemble one another and differ from other varieties. The resemblance 

 between two cases of [i] and that between [a] and the first portions of [au] 

 and [ai] in the Depew plate has been noticed. The conformity to type, 

 with difference in each case, can be clearly seen in the several cases of [ai] 

 on the Cock Robin plate and in the extended special study of [ai] above 

 referred to. Moreover, resemblances are found between the corresponding 

 sounds on the two plates, whereby it must be remembered that the hori- 

 zontal magnification for the Depew plate is 1mm. == . 0007s. and for the 

 Cock Robin plate 1mm. = 0.0016s. 



The curves in plate x show waves from various vowels spoken by 

 Joseph Jefferson in " Rip Van Winkle's Toast." Each line contains only 

 a few waves out of the curve for a vowel. The first line is from the begin- 

 ning of the vowel [a] in "Come"; we note that the amplitude increases 

 steadily; in the latter portion of the vowel (not shown) it falls again. 

 It is a general law for American vowels that their amphtudes are convex, 

 or crescendo-diminuendo. Applying the dividers to the wave-groups, we 

 find that the pitch of the glottal tone steadily rises. It is a general law 

 that the pitch in an American vowel steadily rises and then falls, but 

 this is modified greatly by the adjacent sounds. The form of the wave- 

 group steadily changes, and we have here an illustration of a third funda- 

 mental law also, according to which the American vowel changes its sound 

 constantly. These three laws might, perhaps, he deduced from one, 

 namely, that every factor of muscular adjustment — respiratory pressure, 

 glottal tension, and vowel configuration — is continually changing in an 

 American vowel. The curve in the second line is from the first portion 

 of [i] in "Rip"; it likewise illustrates the three laws. The waves in 

 the tliird fine are from the latter half of the vowel in "what"; the amph- 

 tude does not fall gradually according to the first of the above laws, but 

 diminishes suddenly beyond the piece here shown in three waves as the 

 sound changes to [d], the words being spoken as [hwadaja]. As the reason for 

 this exception to the law of convex ampUtude I can only suggest that pos- 

 sibly the following [d] (although a consonant) is treated by the speaker — 

 unconsciously, of course — as unified with the vowel just as the latter 

 portion^of a diphthong to the first part; we have already seen how several 

 sounds may be unified into one vowel stretch; and this may indicate the 



