248 carnegie institution 



Results Obtained. 

 I. The Nature of Vowels. 



According to Helmholtz a vowel consists of the reinforcement, by 

 the vocal cavities, of overtones in the tone from the vocal cords, these 

 reinforced tones lying within definite regions for each vowel. Thus, 

 for the vowel a the vocal cavity is so adjusted that it reinforces a 

 certain tone ; this tone, however, as an overtone, must stand in 

 respect to its pitch in one of the relations i, 2, 3, etc., to the tone 

 of the vocal cords ; that is, its number of vibrations must be an even 

 multiple of that of the cords, and cannot be, for example, i^ or 

 4y% times as great. According to Hermann a vowel consists of the 

 presence of a tone or tones of definite pitch, the pitch being quite 

 independent of any tone or overtone from the vocal cords — that is, 

 the vowel tone may have, for example, i^^ or 5ff, as many vibra- 

 tions, etc. The two theories agree in asserting the presence of tones 

 of a definite pitch or range of pitch for a particular vowel ; thus for 

 a there must be a tone of one pitch in the vocal sound, for a tone 

 of another pitch, etc. The two theories disagree as to the relation 

 of this tone to the cord tone. 



Helmholtz based his theory on attempts to manufacture vowels by 

 reed pipes and by forks, with results that were not satisfactory. 

 Hermann based his on curves of vowels sung by his own voice into 

 a phonograph and then traced off. The material in either case was 

 extremely limited. The results now studied under this grant com- 

 prised several hundred vowels by different speakers ; being curves 

 of really spoken vowels, thej' required new methods of treatment and 

 gave results quite different from the rather artificial vowels of song.' 



It is to be understood that this work consisted of collecting speci- 

 mens of speech of clearly evident excellence on gramophone plates, 

 of tracing off the vibrations on paper in great enlargement, and of 

 analyzing the waves of the cun^es by means of microscope measure- 

 ment and mathematical formulas. Each single vibration of a vowel, 

 for example, could be analyzed to show the set of tones from the 

 cords and the vocal cavities. Each result stands as a fixed datum 

 with which any theory has to reckon. 



The results soon showed that neither the theory of Helmholtz nor 

 that of Hermann could be applied. No explanation or new theory 

 could be found until a suggestion was received from another inves- 

 tigation, as follows: 



