Characteristic Time Intervals in Telephonic Conversation 



By A. C. NORWINE and O. J. MURPHY 



Two-way conversation is arbitrarily defined in terms of vocal 

 intervals and the pauses between them. These quantities, as 

 determined by the presence or absence of speech energy, have been 

 measured from continuous oscillograms of calls on a New York- 

 Chicago telephone circuit used for Bell System business, and the 

 results of statistical analyses of these data are presented. 



Introduction 



THE time pattern of a conversation may be described in terms of 

 the periods during which speech energy is issuing from the lips of 

 each talker, the pauses with which each intersperses his speech, and 

 the periods after the termination of a talker's speech during which the 

 listener prepares to reply. On a telephone circuit this can be deter- 

 mined by the presence or absence of speech energy within the circuit, 

 measured by an appropriate recording instrument. It is with observa- 

 tions of this type and the measurement of time intervals in conversation 

 obtained in this manner with which the present paper is concerned. 



It will be well to keep in mind that the fundamental basis of these 

 measurements is the presence or absence of speech energy. Many 

 of the pauses recorded in this study are of the type which are known to 

 occur within sentences, phrases, or even within words. Some of these 

 are insufficient in duration to interrupt the continuity of the flow of 

 speech, and some are too short to be noticed by a listener. The 

 intervals as defined in this paper probably do not, therefore, exactly 

 correspond to those which would be observed by a person listening to 

 the conversations. 



The study and measurement of these intervals were originally 

 undertaken to furnish information needed in the application of proba- 

 bility theory to the occurrence of lockouts on toll telephone circuits 

 equipped with tandem voice-operated devices. This problem is treated 

 in a companion paper by Mr. A. W. Horton, Jr.* Since that time, how- 

 ever, parts of the data have been used in various other technical appli- 

 cations, and it is therefore thought that the results of the study may 

 have some general interest. 



* "The Occurrence and Effect of Lockout in Telephone Connections Involving 

 Two Echo Suppressors," Arthur W. Horton, Jr., this issue of the Bell System Technical 

 Journal. 



281 



