Volume 63, Issue 2-3

Pretonic schwa elision in fast speech (e.g. potato -› [pt]ato, demolish -›[dm]olish) has been studied by both phonologists and phoneticians to understandhow extralinguistic factors affect surface forms. Yet, both types of studies havemajor shortcomings. Phonological analyses attributing schwa elision to acrossthe-board segmental deletion have been based on researchers’ intuitions. Phoneticaccounts proposing that elision is best characterized as gestural overlap havebeen restricted to very few sequence types. In this study, 28 different [#CəC-]sequences are examined to define appropriate acoustic criteria for ‘elision’, toestablish whether elision is a deletion process or the endpoint of a continuum ofincreasing overlap, and to discover whether elision rates vary for individualspeakers. Results suggest that the acoustic patterns for elision are consistent withan overlap account. Individual speakers differ as to whether they increase elision

The hypothesis ‘licensing by cue’ by Steriade holds that phonological contrastsare maintained in environments that provide better acoustic cues to the contrastsand are neutralized in environments that provide poorer acoustic cues or nocues. This paper tests the hypothesis by examining the distribution of a phonologicalcontrast - the Russian plain/palatalized coronal stops /t/ and /t^j/ in various syllable-final contexts. The results of a series of acoustic and perceptual experimentspresented in this paper provide some support for the hypothesis: the relativesalience of releases in different word boundary contexts (_#k > _#n, _#s) correlatesstrongly with the general patterns of neutralization of the contrast in similarword-internal contexts (_k > _n, _s) in Russian and other related languages. At thesame time, the relative salience of VC transitions in different vowel contexts(a_ > u_ > i_) has apparently little to do with attested patterns of neutralization.The results suggest that some perceptual cues are phonologically more relevantthan others, providing evidence for interactions between phonetics and phonologymore complex than predicted by the hypothesis.

The aim of this study is to find experimental support for impressionisticclaims that there are prosodic differences between the dialects of Orkney andShetland. It was found that native listeners had no difficulty in discriminatingbetween Orkney and Shetland dialects when presented with speech fragmentscontaining only melodic information. The results of a subsequent acoustic investigationrevealed that there is a striking difference in pitch-peak location, which canbe characterised as a shift in the location of the entire rise, i.e. both the onset andthe peak. Shetland has early alignment, whereas the accent-lending rise in Orkneyoccurs late, so that in disyllabic words with initial stress the pitch peak does notcoincide with the stressed syllable, but is delayed until the post-stress syllable.Finally, the perceptual relevance of the prosodic parameters identified in theacoustic study was investigated.

This paper presents an experimental investigation of durational variation inlexical and post-lexical geminate alveolar laterals, under different stress conditions,in Cypriot Greek. Lexical geminates are found to be longer than post-lexicals,and both geminates and non-geminates are longer in word-initial position. Thedurational distinction is robust in all conditions, but particularly for word-initiallexical geminates. Post-lexical geminates and word-initial lexical geminates aresignificantly longer when pre-stress. Word-initial geminates are longer when precededby a word-final nasal (the condition for post-lexical gemination), thus creatinga kind of ‘supergeminate’ consonant and indicating that word-final nasals arenot deleted, as has previously been thought to be the case. Implications for thephonological analysis of Cypriot Greek geminates and the role of prosodic andphonotactic restrictions are considered.