No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Perception of Assimilation of Voice as a Function of Segmental Duration and Linguistic Context
Phonetica
Abstract
The perception of assimilation of ‘voice’ in Dutch two-obstruent clusters across word boundaries was measured as a function of cluster duration. In an experiment employing synthetic speech stimuli the duration of the silent interval of the plosives and of the noise part of the fricatives was varied. The clusters were embedded in three types of linguistic context: sentences, words, and nonsense words. Longer durations gave rise to more ‘voiceless’ percepts, resulting in the perception of more ‘progressive assimilation’, at the expense of ‘regressive assimilation’ and ‘no assimilation’. Nonwords behaved slightly differently from sentences and words. The different response pattern for nonwords is ascribed to a different phonetic context rather than to lexical or phonological factors.
... The phenomenon is said to be a merger or near-merger, in which the sound change leads to the collapse of the phonemic contrast, so that the two previously distinct phonemes merge into a single phoneme (see Labov, 1994). This devoicing is a well-researched phenomenon from the point of view of phonetics (Debrock, 1977(Debrock, , 1978Slis & Cohen, 1969;Van den Berg & Slis, 1985;Van den Berg, 1989), dialectology (Van Reenen & Wattel, 1992), and language change in progress ( Van de Velde, Gerritsen & van Hout, 1996;Kissine, Van de Velde & van Hout, 2003). The functional load of labiodental fricatives is very low. ...
This study aims at testing whether there are regional differences in the perception of the labiodental fricative contrast in Dutch. Previous production studies have shown that the devoicing of initial labiodental fricatives is a change in progress in the Dutch language area. We present the results of a speeded identification task in which fricative stimuli were systematically varied for two phonetic cues, voicing and duration. Listeners (n=100) were regionally stratified, and the regions (k=5) reflect different stages of this sound change in progress. Voicing turned out to be the strongest categorization cue in all regions; duration only played a minor role. Regional differences showed up in the perception of the consonantal contrast that matched regional differences in production reported in previous studies. The addition of random slopes in the mixed model regression showed the importance of within-regional variation.
... Vocal fold vibration has been established as another major cue to the fortis-lenis distinction in fricatives (e.g., Fischer-Jørgensen 1963 for German; Slis & Cohen 1969a, 1969bvan den Berg & Slis 1985, Kissine et al. 2003 for Dutch, but see Jessen 1998 for a description of [voice] as a feature different from fortis/lenis). In general, /v, z/ are produced with vocal fold vibration, whereas /f, s/ are not. ...
This study addressed prosodic effects on the duration of and amount of glottal vibration in German word-initial fricatives /f, v, z/ in assimilatory and non-assimilatory devoicing contexts. Fricatives following /ə/ (non-assimilation context) were longer and were produced with less glottal vibration after higher prosodic boundaries, reflecting domain-initial prosodic strengthening. After /t/ (assimilation context), lenis fricatives (/v, z/) were produced with less glottal vibration than after /ə/, due to assimilatory devoicing. This devoicing was especially strong across lower prosodic boundaries, showing the influence of prosodic structure on sandhi processes. Reduction in glottal vibration made lenis fricatives more fortis-like (/f, s/). Importantly, fricative duration, another major cue to the fortis-lenis distinction, was affected by initial lengthening, but not by assimilation. Hence, at smaller boundaries, fricatives were more devoiced (more fortis-like), but also shorter (more lenis-like). As a consequence, the fortis and lenis fricatives remained acoustically distinct in all prosodic and segmental contexts. Overall, /z/ was devoiced to a greater extent than /v/. Since /z/ does not have a fortis counterpart in word-initial position, these findings suggest that phonotactic restrictions constrain phonetic processes. The present study illuminates a complex interaction of prosody, sandhi processes, and phonotactics, yielding systematic phonetic cues to prosodic structure and phonological distinctions.
Verskillende faktore, linguisties en non-l!nguisties, speel waarskynlik 'n rol by die toepassing van regressiewe stemassrmilasie in enige taal. 'n Aantal hiervan IS reeds geindentifiseer vir veral Nederlands, en op grond dnarvan en persoonlik waarnemings vir Afrikaans gehipotetiseer. In Wissing (1 990a:56) is dit gestel dat impressionuties-begronde uitsprnke deur nkoesties-eksperirnenteb ondersoeke opgevolg behoort te word. In hierdie artikel word die resultate van 'n ondersoek wat ten opsigte van die assimilasie van Is/tot [z] uitgevoer is, gerapporteer. Ses hipoteses is gestel, waardeur die rol van klemtoon, ti e tan vorm (afleidings, samestellings of frases), tipe leesvorm (gewone ortogra f e of ideogramme), aard van die segment wat op /sf volg, geslag van die s rekers en eil van opvoeding verskillenL toetsstimu% gemaak Die resultate van die ondersoek toon dat veel meer eassimileer word wanneer die kbmtoon nli Is/Val, wanneer mans aan die woorf is, en wanneer die proefpersone studente is (en nie dosente nie). Assimilasie van Is/tot [z kom meer voor vo'o'r lvl as voor /b, dl; die tipe leesvorm afleidings, samestellings of frases nie. Die resultate word ten slotte in h e r linguistiese konteks geplaas en geinterpreteer.
Hoewel assimilasie 'n bekende klankverskynsel in Afrikaans is, en assimilasie van stem, kortweg stemassimilasie genoem, nie onbekend aan Afrikaans is nie, bestaan daar nie veel teoretiese beskrywings in Afrikaans nie. Daar word op regressiewe stemassimilasie gefokus, en meer in die besonder op die wisseling /s/ > [z] in Afrikaans. Die groot hoeveelheid beskrywings van verskillende aspekte van stemassimilasie in Nederlands word behandel. Verskillende opvallende verskille tussen Afrikaans en Nederlands kom aan die lig. Daarna word eers die beskikbare Afrikaanse beskrywings kortliks aan die orde gestel. 'n Hoeveelheid opgetekende frases waarin hierdie tipe assimilasie optree, word gegee. Vervolgens word drie soorte faktore (linguistiese, sosiolinguistie-se en non-linguistiese) wat moontlik 'n rol speel by, of stemassimilasie vergemaklik, bespreek. Dit word waar moontlik aan die hand van die opgetekende voorbeelde en verwysende na die bestaande lite-ratuur gedoen. 'n Multidimensionele beskrywingsmodel vir stemassimilasie word aan die hand gedoen. Daar-in word veral gekonsentreer op die interafhanklikheid van verskillende faktore en subfaktore. Bowendien word 'n hele aantal moontlike en bewese implikasierelasies hierin uitgewys. Ten slotte word 'n aantal terreine waar-op navorsing gedoen behoort te word, aangedui.
The effects of voice onset time (VOT), voice tail (the time between oral closure and voice offset), duration and range of formant transitions, and frication noise intensity on the perception of voicing in Dutch two-obstruent sequences (C1C2) was tested in four separate experiments, employing synthetic stimuli. The results showed that VOT and voice tail were strong cues, and that frication noise intensity was a weaker cue to the perception of voicing in C1C2 sequences. Formant transitions did not have a significant effect. Moreover, the VOT and voice tail data showed that cues from temporally rather distant portions of the acoustic signal are integrated into a perceptual unit, and that a particular cue has a similar effect on the perceived voicing status of both component consonants.
The abstract for this document is available on CSA Illumina.To view the Abstract, click the Abstract button above the document title.
ResearchGate has not been able to resolve any references for this publication.