Cue restructuring in English u-fronting: The role of phonetics and phonology in feature-learning

Abstract

Back tense /u/ is fronting in English in the Northeast US, which results in cue restructuring for the high front lax, back tense contrast (/u--I/). They are no longer distinguished by F2, but the F1 distinction between them is enhanced. In this paper I investigate whether this cue restructuring applies to the phonologically minimally different mid pair /o-E/. I present results from a perception experiment testing to what extent speakers use F1, F2, F3 and duration cues to distinguish between /u--I/ and /o--E/, respectively. Results show that while speakers use the new cues that are available for the /u-I/ contrast, this cue is not useful for the /o-E/, even though these sounds also differ in F1. F3 and duration are not used for either pair. This indicates that cue restructuring stemming from u-fronting does not happen on a featural level, therefore the experiment does not find evidence for a [+/-back] feature. By for a feature in a phonetically natural but phonologically inactive set, this work also contributes to research on mental representations.