Volume 72, Issue 1

In this article, we analyze the phonetic realizations of devoiced vowels from 8 fluent speakers of Southern Ute, a severely endangered Southern Numic Uto-Aztecan language spoken in Southwestern Colorado. Devoiced vowels are considered to be an important aspect of the phonology of Southern Ute, yet very little is known about the pronunciation of such segments. Our findings indicate that devoiced vowels are realized phonetically in three ways: (i) fully voiceless, (ii) partially devoiced, and (iii) fully reduced with concurrent lengthening, lower intensity and greater voicelessness of the preceding consonant. Such variable and noncategorical devoicing is seen for both high and nonhigh vowels and all consonants regardless of their manner of articulation.

We investigate the extent to which French polar questions and continuation statements, two types of utterances with similar morphosyntactic and intonational forms but different pragmatic functions, can be distinguished in conversational data based on phonetic and visual bodily information. We show that the two utterance types can be distinguished well over chance level by automatic classification models including several phonetic and visual cues. We also show that a considerable amount of relevant phonetic and visual information is present before the last portion of the utterances, potentially assisting early speech act recognition by addressees. These findings indicate that bottom-up phonetic and visual cues may play an important role during the production and recognition of speech acts alongside top-down contextual information.

The theory of relational acoustic invariance claims that there are stable acoustic properties in speech signals that correspond to a phonological feature, and that the perception system utilizes these acoustic properties for stable perception of a phoneme. The present study examines whether such an invariance exists in native listeners' perception of Japanese singleton and geminate stops despite variability in speaking rate and word length, and whether this perception corresponds to production. Native Japanese listeners identified singleton and geminate stops in continua of 3- and 4-mora words spoken at different speaking rates. Results indicated that the perception boundary is well predicted by a linear function with two variables: durations of stop closure and the (C)V(C)CV portion (with the contrasting stops underlined) of the 3- and 4-mora words. In addition, these two variables were in a consistent relationship for both perception and production of words containing 2-4 moras. The results support the relational acoustic invariance theory.