Volume 68, Issue 1-2

This experiment investigated whether fundamental frequency (f0) and duration are interdependent cues for the perceived rhythmicality of sentences, and whether this depends on the native language of listeners. The task required a judgement of which stimulus sentence had the most natural- sounding rhythm. In the stimuli, duration and f0 were manipulated on a certain syllable to test whether a deviant duration results in a less natural- sounding rhythm than a deviant f0 movement, or vice versa. The listeners were native speakers of Swiss German (SG), Swiss French (SFr) or French (Fr). SG and Fr differ from each other in terms of prosodic properties involving f0 and duration; in particular, their rhythms sound different. The results demonstrate that duration and f0 are interdependent cues for perceived rhythmicality, and that the relative significance of a non-deviant duration and a non-deviant f0 excursion in the rhythmicality judgements of listeners depends on their native language. Importantly, these findings have implications for current research on speech rhythm.

This paper takes a new look at the traditionally established divide between sounds and prosodies, viewing it as a useful heuristics in language descriptions that focus on the segmental make- up of words. It pleads for a new approach that bridges this reified compartmentalization of speech in a more global communicative perspective. Data are presented from a German perception experiment in the framework of the Semantic Differential that shows interdependence of f0 contours and the spectral characteristics of a following fricative segment, for the expression of semantic functions along the scales questioning – asserting, excited – calm, forceful – not forceful, contrary – agreeable. The results lead to the conclusion that segments shape prosodies and are shaped by them in varying ways in the coding of semantic functions. This implies that the analysis of sentence prosodies needs to integrate the manifestation of segments, just as the analysis of segments needs to consider their prosodic embedding. In communicative interaction, speakers set broad prosodic time windows of varying sizes, and listeners respond to them. So, future phonetic research needs to concentrate on speech analysis in such windows.

A theoretical framework for speech reduction is outlined in which ‘coarticulation’ and ‘articulatory control’ operate on sequences of ‘opening-closing gestures’ in linguistic and communicative settings, leading to suprasegmental properties – ‘articulatory prosodies’ – in the acoustic output. In linking this gestalt perspective in speech production to the role of phonetic detail in speech understanding, this paper reports on perception experiments that test listeners’ reactions to varying extension of an ‘articulatory prosody of palatality’ in message identification. The point of departure for the experimental design was the German utterance ich kann Ihnen das ja mal sagen ‘I can mention this to you’ from the Kiel Corpus of Spontaneous Speech, which contains the palatalized stretch [k̟^hε̈n^jn^jəs] for the sequence of function words /kan i.n(ə)n das/ kann Ihnen das. The utterance also makes sense without the personal pronoun Ihnen. Systematic experimental variation has shown that the extent of palatality has a highly significant influence on the decoding of Ihnen and that the effect of nasal consonant duration depends on the extension of palatality. These results are discussed in a plea to base future speech perception research on a paradigm that makes the traditional segment–prosody divide more permeable, and moves away from the generally practised phoneme orientation.

The present study investigates the acoustics (F1 × F2) of Catalan and Spanish mid-back vowels as produced by highly proficient, early Spanish-Catalan bilinguals residing on the island of Majorca, a bilingual speech community. Majorcan Catalan has two phonemic mid-back vowels in stressed positions (/o/ and /c/) while Spanish possesses only one (/o/). Two groups of bilinguals were recruited and asked to produce materials in both languages – one group of Spanishdominant and one of Catalan-dominant speakers. It was first found that Catalan and Spanish /o/ are virtually indistinguishable. Catalan /c/ is lower and more fronted than the other two vowels. Spanish-dominant bilinguals were found to differ from Catalan-dominant ones in that they did not produce the Catalan-specific /o/-/c/ contrast in their speech; that is, they produced a single, merged Catalan mid-back vowel. A within-subjects analysis of first- and second-language mid-back vowels further suggested, for Spanish-dominant bilinguals, that they had developed a separate vowel category to accommodate their single, merged Catalan mid-back vowel; that is, they possessed a two-category mid-back vowel system, i.e. one for their Spanish /o/ and one for their merged Catalan /o/ + /c/. Potential explanations and theoretical implications are discussed.