808 BELL SYSTEM TECHNICAL JOURNAL 



of the recognizabilitles of the various phonetic units. EngUsh words 

 and short sentences have also been used for testing purposes. When 

 material of this kind is used, a new element enters, namely, the 

 thought or meaning associated with the sentence or word. The 

 criterion for the correct observation of words or sentences is also 

 different from that used in the case of articulation tests. If the 

 thought or meaning of a word or sentence is correctly understood, 

 it is considered to be correctly received, even though the observer 

 may not have correctly recognized each sound that was spoken. 

 The terms "word articulation" or "sentence articulation," therefore, 

 seem inappropriate when referring to the results of such tests. The 

 term "intelligibility" has frequently been used in this sense. Since 

 it has also been used in a more general sense, the terms "discrete 

 word intelligibility" and "discrete sentence intelligibility" will be 

 used when referring to the results obtained by using disconnected 

 words or sentences for the testing material. They are defined as the 

 percentage of the total number of spoken words and sentences, 

 respectively, that are correctly interpreted according to the criterion 

 given above. 



Very early in the work of developing the telephone, words and 

 sentences which were chosen in a haphazard way were used for testing 

 purposes. Word lists of various sorts have been worked out and 

 used with some success. Even in very recent years some of these 

 word lists have been used to good advantage. The main objections 

 which have developed, to the continuous use of such lists are: (a) it 

 is hard to make the lists equally difficult without resorting to very 

 long lists of words, {b) a very large number of lists are necessary in 

 order to avoid memory effects. 



Dr. G. A. Campbell ^ was one of the first to propose a system of 

 syllabic speech sounds for testing the transmission characteristics of 

 the telephone system. These syllables had no meaning and were 

 constructed by combining the various initial consonants with the 

 vowel "ee," such as bee, fee, etc. With these lists the consonant 

 articulation was taken as a measure of the system. 



Later Dr. I. B. Crandall ^ worked out a system which used both 

 simple and compound consonant forms in a vowel-consonant and 

 consonant-vowel type of syllable. All of the common vowels were 

 used, and the combinations were formed in ways which are usually 

 found in written speech. The sounds occurred with the same fre- 

 quency as they occur in ordinary written material. As in the Camp- 

 bell lists, the articulation was based on the consonant sounds alone. 



1 "Telephonic Intelligibility," G. A. Campbell, Phil. Mag., Jan. 1910. 



2 "Composition of Speech," 1. B. Crandall, Phys. Rev., 10, p. 74, July 1917. 



