Temporal correlates of foot-level timing in English and Japanese

Abstract

Languages of the world are felt to be spoken with different kinds of rhythm. But what is the basis for this impression? Notions like ‘‘stress-timed’’ and ‘‘syllable-timed’’ have been difficult to verify experimentally because rhythm in spoken language rarely yields consistent temporal regularities. Experiments were conducted in which native English and Japanese speakers repeated phrases using a specified rhythm in time with an auditory metronome. Results demonstrate that temporal correlates exist for prosodic units approximately the size of the ‘‘foot,’’ consisting of one relatively strong syllable followed by zero or more weak syllables. Strong (foot-initial) syllables show greater temporal stability than do weak (foot-internal) syllables when the timing of syllables was perturbed by manipulating inherent segmental durations. This is found for both English and Japanese. However, in English, increasing the foot size by the addition of weak syllables is accompanied by compensatory shortening only within the foot, while in Japanese, compensation occurs both within and between adjacent feet, suggesting a weaker role for this unit. By eliciting speech in a rhythmic context, such as ‘‘speech cycling,’’ both language-general and language-specific properties of speech rhythm can be directly measured.