SOME SUGGESTED EXPERIMENTS FOR THE GRAPHIC 
RECORDING OF SPEECH VIBRATIONS 
ROBERT JAMES KELLOGG 
In an attempt to recast the science of phonetics on the basis 
of connected speech flow,^ I have found it necessary to criti- 
cally compare the results obtained by different methods of 
graphically recording speech vibrations. Out of this comparison 
a number of suggestions have emerged (1) for further perfecting 
certain existing types of apparatus, (2) for the application of 
apparatus already perfected to special linguistic problems, and 
(3) for the development of new types of apparatus in which the 
^Tight-lever’’ (or shifting ray of light responding to sound vibra- 
tions) partly or wholly replaces the physically vibrating parts. 
These suggestions are presented herewith in the hope that they 
may be of interest and help to other investigators, and also that, 
if a larger number can cooperate in working them out, the prac- 
tical and mechanical difficulties involved may be the sooner 
solved and the apparatus based on these principles devised and 
perfected. 
I. THE PERFECTING OF EXISTING TYPES OF APPARATUS 
It cannot be too strongly insisted that the mechanical per- 
fection of diverse types is necessary if the problems of experi- 
mental phonetics are to be fully solved. Every type of sound 
recording apparatus distorts the sound curves in its own special 
way, suppressing, damping, modifying or adding specific vibra- 
tional elements. Even in the perfected apparatus this distor- 
tion may either require complicated correction of results (as in 
the phonodeik), or it may not be fully determinable from the 
unaided standpoint of the apparatus in question (as in the case 
^ Proc. Mod. Lang. Assoc., 1915, p. xxix, title 46: About face in fonetics. 
83 
