Voiced versus voiceless inaspirates: Acoustic measurements of English stops

Abstract

Attention is focused on intervocalic stops in trochaic words. Embedded in semantically plausible sentences, tokens of six pairs of words like robing and roping were recorded by three speakers of American English. These included only places of articulation, labial and dorsovelar, for which the voicing opposition is reliable for all dialects in natural speech. Four acoustic properties were examined because they are considered relevant and they are amenable to editing for perceptual experiments. They are closure pulsing, closure duration, stressed‐vowel duration, and release‐burst amplitude. The only entirely reliable correlate is presence versus absence of closure pulsing. A fairly good correlate is vowel duration. Vowels before voiced stops are somewhat longer, although there is overlap between the two ranges. Another fairly good correlate is closure duration. Voiceless closures are somewhat longer, although here, too, there is overlap. The amplitude of the burst is a poor correlate, even though voiceless bursts tend to be more intense. These data furnished a baseline for our paper on perception. [Work supported by NIH Grant HD‐01994.]