Initial and Intervocalic Cues for the Perception of Liquids and Semivowels
Louis J. Gerstman, Leigh Lisker, J. D. O'Connor, Alvin M. Liberman, Pierre C. Delattre, F. S. Cooper · The Journal of the Acoustical Society of America · 1957
Experiments with CV syllables synthesized on the pattern playback have demonstrated that the liquids r, l, and semivowels w, j, are distinguished from one another by variations in the direction and extent of the second- and third-formant transitions. For each phone, the onsets of the transitions are marked by brief steady states whose frequencies remain fairly constant from vowel to vowel: the second-formant onset is low for w (ca 600 cps), high for j (ca 2400 cps), and intermediate for r or l (900–1800 cps); the third formant for l is practically straight, but for r its onset lies just above the second-formant onset. In VCV combinations, the steady states connect the transitions from the preceding and following vowels. As in initial position, the second-formant termination distinguishes w from r-l from j and the third formant separates r from l. When the preceding and following vowels are identical, the two sets of transitions are mirror images of each other, but when the two vowels (and thus the transitions) are different, identifications are governed by the transitions to the following vowel. Despite the doubling of the transition cues, w, r, l, and j are little more intelligible intervocalically than they are initially. (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 DA49-170-sc-1642.)