Some Cues for the Distinction Between Voiced and Voiceless Stops in Initial Position

A series of experiments with synthetic speech indicated that each of the voiced stops could be made to sound like its voiceless counterpart by eliminating the beginning of the first‐formant transition. The amount of first‐formant “cutback” required to produce a voiceless stop varied according to the stop, the vowel which followed it, and the listener, but was in no case greater than 40 msec. It was also found that filling the first part of all formant transitions with noise (i.e., aspiration) contributed to the impression of voicelessness, but the importance of aspiration as a cue appeared to be considerably less than that of the first‐formant cutback. Many listeners showed great consistency in differentially indentifying the patterns as voiced or voiceless even when the stimulus difference was seemingly very small. In one experiment, for example, several listeners sorted the patterns perfectly on the basis of a 10‐msec difference in the starting time of the first formant. [This work was supported in part by the Carnegie Corporation of New York, and in part by the Department of Defense in connection with Contract DA 49‐170‐sc‐2159.]