110 THE STUDY OF SPEECH CURVES. 



does anything it registers that tone and reproduces it; even in a record 

 so bad that the vowels can not be distinguished from one another the 

 tone of the voice can be heard. In any case, the theory requires a strong 

 first partial in the results of the analysis; since such a partial usually 

 does not occur, the theory must be incorrect or inadequate. 



An inspection of the curve by the eye shows that it falls into groups 

 of small waves, the entire group corresponding to the period of one vibra- 

 tion of the glottal tone (p. 40). The e^-e notes at once that in most cases 

 a sinusoid of the length of the groups is lacking. We have here apparently 

 a demonstration directly to the eye that the glottal tone is weak or lacking. 

 Yet the glottal tone — the first partial or fundamental — is the strongest of 

 all in the sound itself; we may not be able to decide whether a voice sings 

 [o] or [a], but we certainly hear the tone on which it is sung and know 

 whether it is a bass or treble note. Since the fundamental simple sinu- 

 soid is indicated as wholly or partly lacking in the curve, the theory can 

 not be correct and there must be present some other form of vibration 

 than the simple sinusoid. 



Another reason for rejecting the overtone theory is that the har- 

 monic plots do not indicate the reinforcement of certain overtones with 

 total absence of others, but generally show that several neighboring over- 

 tones are reinforced (p. 83). Such results indicate rather that the tones 

 reinforced in the vocal cavities are not harmonic overtones, but inharmonics 

 of the fundamental; the reason that they do not appear in the plot is 

 that the analysis provides only for harmonics (p. 77). 



Still another reason for rejecting the overtone theory lies in the fact 

 that it is based on views of resonance which are not vahd for the voice. 

 Helmholtz supposed the vocal cavities to act as a series of resonators which 

 respond to definite overtones in the glottal tone. Such a supposition 

 would be appropriate if the cavities were made of metal or other hard 

 substances. The vocal cavities have, however, soft or moderately hard 

 walls Uned with moist membranes. The laws of resonance for soft cavi- 

 ties are different from those for brass resonators. The experiments on 

 resonance summarized at the close of this chapter show that cavities with 

 soft walls will respond to a range of tone which increases as the softness; 

 for example, a cavity with walls of water will respond to any tone, a cavity 

 with flesh walls to a considerable range of tone, etc. The same conclu- 

 sion can be deduced theoretically. The process of vowel production must 

 therefore differ completely from the theory that compares it to the response 

 of hard resonators to overtones. 



Wheatstone and Helmholtz were apparently led to the overtone 

 theory by the utterly erroneous suppo.sition that the glottal lips act like 

 rubber membranes with freely swinging edges. How misleading this notion 



