THE WORDS AND SOUNDS OF TELEPHONE CONVERSATION 311 



heard in reasonably enunciated conversation among educated persons 

 in New York. The departures from dictionary standards are largely 

 confined to the vowels. As a result the analysis is affected to some 

 degree by the speech habits of the writers.^ It is regrettable that some 

 arbitrariness should be introduced, but this seems to be a difficulty 

 common to discussions of vowel sounds. Some of the difficulty is 

 avoided by making separate classifications for vowels for which the 

 pronunciation is indefinite, such as the vowels in unstressed positions. 

 The articles "the," "a" and "an" were omitted entirely from the 

 analysis on account of the large number of variant pronunciations to 

 which they are subject. 



The results of the analysis into speech sounds are shown in Table 

 VIII. Three divisions are given: vowels, initial consonants and final 

 consonants, based on the division into phonetic syllables. The method 

 followed was: first, to divide the words into phonetic syllables, second, 

 to assign phonetic symbols to the sounds and third, to weight each 

 sound by the total number of times the word was recorded. The 

 sounds are identified in the table, where necessary, by key words. 



No difficulties were encountered in analysis of the consonants, but a 

 few special points which arose in assigning the vowel qualities may be 

 noted. The key word "pot" is used to denote a vowel sound which is 

 pronounced differently by many natives of New England and those 

 whose habits of speech were formed elsewhere.* With these New 

 Englanders the sound tends toward the quality of the vowel in "pawn," 

 although shorter in duration. The same New Englanders make a real 

 distinction between the vowel of "pot" and the vowel of "palm." By 

 many speakers elsewhere no such distinction is made and the two are 

 lumped into a single intermediate sound which is neither the New 

 Englander's "pot" nor "palm." To avoid confusion the class denoted 

 by "pot" has been made to include "not" and many other monosylla- 

 bles of the same ending, as well as "on," "job," "stock," etc., which 

 grouping is believed to be homogeneous on either basis. The few 

 words of the class of "palm" which were encountered have been in- 

 cluded under "par." The class denoted by "par" may be subdivided 

 into: "par," 1.24; "palm," 0.07. The class denoted by "palm" 

 would be somewhat larger if the class which we may denote by " path," 

 such as "can't," "last," "ask," etc., had not been classified under 



^ For the benefit of phoneticians who may be interested it may be stated that the 

 writers are residents of Greater New York of more than six years' standing, that their 

 boyhoods were spent in Maine, Illinois and New Jersey, respectively, and their college 

 years at Maine and Princeton, Harvard and Oxford, New York University and 

 Harvard, respectively; this seems a background sufficiently varied to bring to light 

 many of the principal variants of American speech. 



* Just what the geographical lines may be, the writers do not pretend to know. A 

 phonetic map would be of interest. 



