ARTICULATION TESTING METHODS 809 



Several other lists which have not been pubHshed were proposed 

 and used, the differences being in the choice of the fundamental 

 speech sounds, in their arrangement into syllables, and in the relative 

 frequency of occurrence, both of the different syllable forms and of 

 the speech sounds in each form. There was a distinct effort to make 

 the lists as nearly like speech as possible by using the syllable forms, 

 and by using the particular combinations of fundamental sounds 

 that occur frequently in English. Difficulties were encountered in 

 testing, however, when this was carried too far in that enough different 

 syllables could not be obtained for continuous testing. On the other 

 hand, when random combinations of sounds were made, without 

 regard to the particular combinations occurring in English, syllables 

 that were very unusual and difficult to pronounce were obtained, 

 unless the combinations were restricted to the simple syllable forms 

 having only two or three sounds. In other words, testing lists must 

 be selected with two things in mind; namely, they must be repre- 

 sentative of speech and they must be suitable for making tests. The 

 experience with these various lists also indicated that the results 

 obtained with one system of lists could be calculated approximately 

 from the results obtained with other systems by properly weighting 

 the individual sound articulation values. 



This experience led to the adoption by the Laboratories of a system 

 of lists which has been used during the past ten years in studies of 

 the effects of distortion upon the recognition of speech sounds.^ 

 These lists which have been referred to in the literature as the standard 

 articulation lists were made up of only the con-vow, vow-con and 

 con-vow-con syllable forms. The various fundamental sounds of 

 English were combined at random into the syllables, such that each 

 sound occurred approximately with uniform frequency. 



During the past few years it has become evident that still further 

 simplifications in the syllable forms used in the standard articulation 

 lists might be made. Also our methods of making articulation tests 

 and interpreting the results obtained have undergone considerable 

 changes during this time. It is with these new methods that the 

 present paper is chiefly concerned. 



In order to distinguish between the old lists and the ones modified 



as described below, the prefixes "old" and "new" will be placed 



before the title "Standard Articulation Lists." When there is no 



chance for confusion, the new lists will be called simply the standard 



articulation lists, since they are the principal ones now being used in 



the work at Bell Telephone Laboratories. 



^"Nature of Speech and Its Interpretation," H. F"letcher, Journal Fra)iklin 

 Institute, June, 1922. 



