Structured covariation of acoustic cues to coda voicing throughout the syllable

Abstract

In English, multiple acoustic cues signal the voiced-voiceless distinction in stop codas. The strongest cue to this contrast is preceding vowel duration: voiced codas tend to be associated with longer vowels and voiceless codas with shorter vowels. Recent research has found that speakers may use multiple cues to stop voicing, and they do so in a manner that is highly structured. Building on our prior work with female speakers (Jacewicz et al., 2021), we investigated the covariation of temporal cues to coda voicing in males (n=45, N=2,506 productions). We found that, in running speech, the coda voicing contrast was cued not only by the preceding vowel but as early as the onset of a monosyllabic word. The specific long-distance temporal relationships between onset stop (affecting closure duration, voicing during the closure and VOT) and coda consonants (stops and fricatives) indicate a syllable-wide distribution of cues to coda voicing. The temporal patterns in males were generally consistent with females. There were also systematic prosodic and regional dialect effects on the nature of stop closure voicing, with higher prevalence of prevoicing (i.e., negative VOT) in Southern males than Southern females.