Volume 54, Issue 2

Fundamental frequency (F₀) correlates of the Swedish tonal word accents, ‘grave’ and ‘acute’, were studied in spontaneous speech produced by 3 male speakers of Central Standard Swedish. The grave accent was invariably marked by an F₀ fall on the primary stress syllable, while the corresponding F₀ contour in acute words seemed to be predictable from sentence-level intonation. The latter observation, which could not be tested on the basis of spontaneous speech data, was confirmed in a companion paper in which a controlled speech material was analysed. In combination, the two papers corroborate the view that an F₀ fall is a positive phonetic feature of the grave accent, while the acute accent constitutes the unmarked member of the contrast. They also illustrate the methodological advantage of using complementary spontaneous and controlled speech materials.

Across a variety of languages, phonation type and vocal-tract shape systematically covary in vowel production. Breathy phonation tends to accompany vowels produced with a raised tongue body and/or advanced tongue root. A potential explanation for this regularity, based on a hypothesized interaction between the acoustic effects of vocal-tract shape and phonation type, is evaluated. It is suggested that increased spectral tilt and first-harmonic amplitude resulting from breathy phonation interact with the lower-frequency first formant resulting from a raised tongue body to produce a perceptually ‘higher’ vowel. To test this hypothesis, breathy and modal versions of vowel series modelled after male and female productions of English vowel pairs /i/ and /i/, /u/ and /ω/, and /λ/ and /a/ were synthesized. Results indicate that for most cases, breathy voice quality led to more tokens being identified as the higher vowel (i.e. /i/, /u/, /ÊŒ/). In addition, the effect of voice quality is greater for vowels modelled after female productions. These results are consistent with a hypothesized perceptual explanation for the covariation of phonation type and tongue-root advancement in West African languages. The findings may also be relevant to gender differences in phonation type.

This paper describes the phonetic characteristics of Banawá, an endangered language spoken in Brazil. The qualities of the Banawá vowels are described in terms of their formant frequencies. The places of articulation of each consonant, the voice onset time and the manner of articulation are documented. The structure of syllables and words is delimited, and the location of stressed syllables is described and verified experimentally.