Abbreviation of consonants in clusters: A condition on the temporal structure of syllables

Abstract

Consonants in clusters are abbreviated by as much as 30% [Haggard, J. Phon. (1973), Klatt, J. Acoust. Soc. Am. 59, 1208–1221 (1976)]./s/ and /k/, e.g., are shorter in mask than in mass or Mac. Although previous studies have attempted to explain this phenomenon in terms of phonetic context, they leave one basic question unanswered: is shortening an active process? (Earlier investigators considered only two‐segment clusters within a single morpheme, which may be frozen forms functioning as single consonants.) In order to address this question, a production experiment was devised in which single consonant morphemes (/t, d, s, z/) were affixed to monosyllables ending in two‐segment clusters. The results indicate that the addition of a final segment has the effect of shortening a non‐contiguous segment (e.g., the /s/ in masked is shorter than the /s/ in mask). This illustrates two points: (1) shortening is an active process; and (2) it is not simply a function of immediate phonetic context; rather it is controlled on a higher level of organization as a condition on the temporal structure of syllables. There seems to be a tendency to maintain the final consonant portion of a syllable at a fixed ratio to the total duration of the syllable [Port, J. Acoust. Soc. Am. 69, 262 (1981)]. Other processes affecting clusters (e.g., cluster reduction) can also be explained by this principle.