592 BELL SYSTEM TECHNICAL JOURNAL 



tory circuits take the place of the transient vibrations of the oral 

 cavities. Finally Paget (reference (9a) below) has constructed a 

 whole series of double resonators which may be excited by l)lowing 

 air into them through an "artificial larynx," and from which he has 

 obtained all of the vowel sounds. As the result of this work he has 

 given a very complete chart of the characteristic frecjuencies of the 

 vowels and he has been led to the conclusion that there are two char- 

 acteristic frequencies or regions of resonance for each vowel sound. 



From the standpoint of practical acoustics both theories have con- 

 tributed to progress, and it seems that the experimental physicist 

 would not be justified in partiality to either view. Speech is a \ariable 

 phenomenon; the cord tones are not always stable; in speaking and in 

 singing there are allowable variations in duration, intensity and fre- 

 quency of the component tones without essential change in the char- 

 acteristics of the vowel sounds. Given accurate records of the speech 

 sounds as normally pronounced by a number of speakers, we should 

 expect to arrive at nearly the same characteristic frequencies which- 

 ever mode of analysis we adopt. As pointed out by J. Q. Stewart 

 (Ref. 8i) Rayleigh has stated (.Sound, \'oi. II, p. 473) that the dis- 

 agreement between the Helmholtz-Millcr, or steady state theory of 

 vowels, and the Willis-Hcrmann-Scripturc, or transient theory is onl\- 

 apparent; to quote Stewart, "The disagreement concerns methods 

 rather than facts. Which \iewpoint should be adopted is thus a 

 matter of convenience in a given case. When the transmission of 

 speech over telephone circuits is in question, for example, the steady 

 state theory often possesses obvious mathematical achanlages. On 

 the other hand, the quantitati\e data relating to the physical nature of 

 \owels which are gi\en in D. C. Miller's well-known book "The Science 

 of Musical Sounds" expressed as ihey are in terms of the steady slate 

 theory are less compact and definite than the data of Table I (Stewart's 

 paper) which are expressed in terms of the transient theory. The 

 general agreement between the two sets of data is, of course, obvious." 



In stud\ing the behavior of \ibrating systems from the theoretical 

 standpoint, there is a tendenc\' to emphasize the intimate relations 

 that exist between transient and steady state phenomena. Both 

 depend only on the dri\ing forces and the constants of the system, 



'(a) Sir K. .-\. S. Pagcl: "The I'roduclion of .^rlifitial \owcl Sound,';." I'roc. Ko\-. 

 Soc. .A 102, Mar. 1, V)li, p. 752. 



'(!>) .A second memoir: "Tlic Nature and .Artificial Production of Consonant 

 Sounds." I'roc. Roy. S<jc. .\ 106, .Aug. 1, l')2-l, p. 150, to which reference will Ic 

 made in more detail later. 



Other papers liy Paget include: .Nature, Jan. 6, V)li, "Nature and Reproduction 

 of Speech Sounils." Klectrician, .Apr. 11, 1924. The Same Title. Proc. Land. 

 Phys. Sot. .<0 pt. i, .\pr. 15, 1924, p. 213: Discussion on Loud Speakers. 



