The effect of lexical competition on vowel duration before voiced and voiceless English stops

Abstract

One source of phonetic variation is lexical competition, which has been shown to cause both hyperarticulation, especially of vowel formants (e.g., Wright 2004), and reduction, manifested as shorter segment and word durations (Kilanski 2009, Gahl et al. 2012). Studies have found that lexical competition (e.g., having a minimal pair competitor, having more phonological neighbors) causes contrastive hyperarticulation of the initial stop voicing contrast in English: greater competition makes VOT longer for voiceless stops and/or shorter for voiced stops (Baese-Berk & Goldrick 2009, Nelson & Wedel 2017). I conducted a corpus study that looked for contrastive hyperarticulation of the final stop voicing contrast in English. The cue examined was preceding vowel duration. I did not find contrastive hyperarticulation: neither minimal pair competitor existence nor higher neighborhood density caused vowels to be longer before final [d] or shorter before final [t]. Instead, both competition metrics correlated with shorter vowel durations before [t] and [d]. This result is consistent with Goldrick et al. (2013)’s failure to find contrastive hyperarticulation of the final stop voicing contrast and their hypothesis that competition affects initial and final contrasts differently. It is also consistent with Gahl et al’s (2012) finding that high neighborhood density causes reduction.