The phonological domains of final lengthening

Abstract

A syllable at the end of a phrase is considerably longer than it would be phrase‐internally. Similarly, a stressed monosyllable is generally longer than any identical syllable separated from the word's edge by one or more unstressed syllables. Two sets of experiments have been designed to determine the phonological domain of these effects. The first compared target syllables at various distances from the ends of words ([pe] in pep, pepper, peppermint; [pa] in pop opposed, poppa posed) in sentences where they either did or did not precede an obligatory intonational phase break. These experiments showed that (1) there is a smaller word‐final effect distinct from the substantial phrase‐final lengthening at intonational boundaries, and (2) the word‐final effect must also be a boundary effect rather than a result of stress timing, since it occurred even in the pop opposed‐poppa posed utterances, where the target syllable was always followed by exactly one unstressed syllable. The second set of experiments tests whether word‐final lengthening is affected by pitch‐accent and stress placement, or whether it defines a phrasal unit independent of metrics.