Phonetic perception.

Number 232
Year 1978
Drawer 4
Entry Date 06/03/1999
Authors Liberman, A. M., & Studdert-Kennedy, M.
Contact
Publication In R. Held, H. W. Leibowitz, & H.-L. Teuber (Eds.), Handbook of sensory physiology, Vol. VIII: Perception. New York: Springer-Verlag, 143-178.
url http://www.haskins.yale.edu/Reprints/HL0232.pdf
Abstract [Introduction] To include a chapter on phonetic perception in a handbook like this is to assume that the process is not wholly accounted for by such principles as we might find in research on the perception of nonspeech sounds. It is appropriate, then, that we here offer support for that assumption. We will not examine all relevant considerations, only those that bear most directly on the relation between the information in the acoustic signal and the listener’s perceptual response to it; in our view, those are the most pertinent. Nor will we analyze such arguments as there are for the opposite assumption-namely, that auditory mechanisms are sufficient-though we will, as is proper, refer the reader to relevant papers. Phonetic perception is what happens when, on hearing speech, a listener recovers the phonetic message. That message consists of the meaningless segments we perceive as consonants and vowels. These are ordered in strings, organized into larger units, and carried on a prosodic contour. The segments, both consonants and vowels, are called “phones”; among the larger units are syllables; the relevant aspects of prosody are stress and intonation. We must distinguish between the perceived phones and the more abstract phonologic forms that underlie them. Thus, the final segments in “cats” and “dogs” are different phones-voiceless[s] in “cats” and voiced [z] in “dogs” -yet at a more abstract phonologic (or morphophonemic) level they are the same. Our concern will be with the less abstract phones and their relation to the still less abstract sounds. Also, to keep our task within bounds, we will deal only with the segmental aspects of phonetic structure, including the organization of phones into syllables, though perception of prosody presents interesting, perhaps even similar, problems. Students of language commonly assume a complex, grammatic relation between meaning and its phonetic vehicle, but often disregard the further complications that arise in the conversion to sound. They tend rather to suppose that the phonetic segments (or their constituent features) are represented discretely in the signal, as if by an acoustic alphabet. If that were so, perceiving phones would be like perceiving any other sounds: there would be no special problem of phonetic perception and no reason for this chapter. There is evidence, however, that the sounds of speech are not an alphabet on the phones, but a complex and grammatic code. In the first section of the paper we will place that code in the larger scheme of language and identify its important characteristics. If it is true that the phones are linked to the sounds by a special code, we should suppose that extracting the phones from the sounds would require a correspondingly special decoder. In the second section we will give reasons for supposing that such a decoder may exist. There is, of course, an alternative to grappling with the problems created by the peculiar relation between sound and phonetic message: We can try to evade them. Indeed, we might suppose that phonetic perception does not occur, that the segments of the phonetic level are mere fictions, invented by linguists for their convenience, with no functional significance in language or in its psychophysiology. In that view the listener would go directly from sound to some meaningful segment (for example, word), bypassing the phonetic and phonologic structure entirely. To justify our concern with phonetic perception, we will, in the third section, argue that phonetic (and phonologic) structure plays an important role in language and is, in fact, recovered by the listener when he comprehends what is said to him.
Notes

Search Publications