A Rhythmic Model of English

Abstract

The subjective impression that English tends to be spoken rhythmically-that is, with recurrent alternations of stronger and weaker elements-is apparently due chiefly to the rhythmic recession of accents: potentially accentable prenuclear heavy syllables tend to be realized as unaccented if flanked closely by neighboring accented syllables. Accent, though it usually augments the duration of the affected syllable, is primarily manifested by a pitch obtrusion. Despite the myth of isochronism, it takes longer to get from one accented syllable to the next if several unaccented syllables intervene, even if all are light (hence, cet, par., short), than if only one or none does. The durational factors needed, e.g., to synthesize natural-sounding speech (leaving aside such paralinguistic effects as pauses, hesitations, and abrupt changes of speech rate) are (a) inherent syllable duration, depending on phonematic makeup; syllable weight (heavy ones are longer, cet. par.); accent (accented syllables are longer, cet. par.); and location in sense group (terminal syllables are drawled). Rhythmic can be used loosely to refer to differences in any of these factors, but in the strict sense English is better described as arrhythmic.