| Summary: | <p>The relationship between vowel duration and word-final voicing in English is well attested. In a minimal pair, contrasting in word-final obstruents, the production of a longer vowel precedes the voiced obstruent (for example [bæd] bad) relative to the voiceless obstruent (for example [bæt] bat). The research presented here focuses on the underlying representation of vowel length, and the manner in which listeners process this representation to characterise word-final voicing. A total of five experiments were run examining a range of both recorded and synthesised monophthongs and diphthongs (/æ/, /eɪ/, /əʊ/, /aɪ/), thereby assessing the representation of both underlyingly single and bimoriac vowels.</p>
<p>In two of the five experiments presented, data obtained from native and non-native English speakers are compared. The aim here is to assess how the representation and processing of this fine-grained perceptual cue differs for L2 English speakers for whom this positional contrast does not exist in the surface form of their native phonology. Here, we consider German wherein the voicing contrast is neutralised word-finally, and only voiceless word-final obstruents are permitted in the surface form. Lexical status is also considered, alongside the extent to which lexical and acoustic information interact when there a mismatch between vowel duration and voicing. Most notably, results from a series of forced choice identification tasks and reaction time data obtained from a lexical decision task with fragment priming consistently support the notion that vowel duration acts as a primary perceptual cue for word-final voicing in native English speakers. For non- native speakers, an increased sensitivity to this perceptual cue is present in speakers who have had more exposure to spoken English, though this tendency remains governed by native phonology. Specifically, where an underlying voicing contrast does exist in German, listeners are primarily guided by their L1 phonology; perceiving a majority of voiceless responses. However, where an underlying voicing contrast does not exist in German, participants are guided by L2 phonology. Here, a lexical effect is found wherein German participants who have received less exposure to spoken English do not adhere to this tendency for nonwords. In the case of production, participants adhere to their L1 phonology for nonwords containing word-final sounds in which there is no underlying voicing contrast in German, regardless of the amount of exposure to spoken English received by the speakers.</p>
|
|---|