Fig 3 - available via license: CC BY
Content may be subject to copyright.
Boxplots for the distribution of the Integrated Voicing Index for the initial stops in the unstressed syllable across 11 speakers: (a) the difference in the IVI between the voiced and voiceless stops; (b) the difference in the IVI between the IP-initial and IP-medial positions; and (c) the difference in the IVI between the accented (focused) and unaccented (unforced) conditions. The IVI was defined as a combined sum of VOT (as a positive value) and Voicing-in-Closure (as a negative value).
Source publication
The data reported in this article contain eleven (6 female and 5 male) individual speaker's speech production patterns for the word-initial voiced and voiceless stops (/p,t/ and /b,d/) in American English. The production patterns are documented in the acoustic parameter: the Integrated Voicing Index (IVI) obtained from Voice Onset Time (VOT) and vo...
Contexts in source publication
Context 1
... (the voiceless vs. voiced stops) in the stressed syllable across 11 speakers as a function of (a) the boundary conditions (IP-initial vs. IP-medial) and (b) the prominence conditions (accented/focused vs. unaccented/unfocused). The IVI was defined as a combined sum of VOT (as a positive value) and Voicing-in-Closure (as a negative value). Fig. 3 illustrates how the eleven individual speakers produced the voiceless vs. voiced stops in the initial unstressed syllables in iambic words in American English. This provides the information of how the phonological voicing contrast in the unstressed initial syllable is phonetically implemented along the phonetic voicing dimension: IVI, ...
Context 2
... eleven individual speakers produced the voiceless vs. voiced stops in the initial unstressed syllables in iambic words in American English. This provides the information of how the phonological voicing contrast in the unstressed initial syllable is phonetically implemented along the phonetic voicing dimension: IVI, the Integrated Voicing Index (Fig. 3a). Fig. 3 also illustrates how the individual speakers produced the stops (the voiceless and voiced stops combined) in various prosodic contexts: in the two prosodic boundary conditions, IP-initial vs. IP-medial position (Fig. 3b) and in the two prominence conditions, accented (focused) vs. unaccented (unfocused) (Fig. 3c). The ...
Context 3
... speakers produced the voiceless vs. voiced stops in the initial unstressed syllables in iambic words in American English. This provides the information of how the phonological voicing contrast in the unstressed initial syllable is phonetically implemented along the phonetic voicing dimension: IVI, the Integrated Voicing Index (Fig. 3a). Fig. 3 also illustrates how the individual speakers produced the stops (the voiceless and voiced stops combined) in various prosodic contexts: in the two prosodic boundary conditions, IP-initial vs. IP-medial position (Fig. 3b) and in the two prominence conditions, accented (focused) vs. unaccented (unfocused) (Fig. 3c). The horizontal axes ...
Context 4
... initial syllable is phonetically implemented along the phonetic voicing dimension: IVI, the Integrated Voicing Index (Fig. 3a). Fig. 3 also illustrates how the individual speakers produced the stops (the voiceless and voiced stops combined) in various prosodic contexts: in the two prosodic boundary conditions, IP-initial vs. IP-medial position (Fig. 3b) and in the two prominence conditions, accented (focused) vs. unaccented (unfocused) (Fig. 3c). The horizontal axes in the figure refer to speaker ID number along with the gender information. Fig. 4 illustrates how the individual speakers produced the voiceless vs. voiced stops (i.e., the voicing contrast) as a function of two prosodic ...
Context 5
... Integrated Voicing Index (Fig. 3a). Fig. 3 also illustrates how the individual speakers produced the stops (the voiceless and voiced stops combined) in various prosodic contexts: in the two prosodic boundary conditions, IP-initial vs. IP-medial position (Fig. 3b) and in the two prominence conditions, accented (focused) vs. unaccented (unfocused) (Fig. 3c). The horizontal axes in the figure refer to speaker ID number along with the gender information. Fig. 4 illustrates how the individual speakers produced the voiceless vs. voiced stops (i.e., the voicing contrast) as a function of two prosodic factors: Boundary (IP-initial vs. IP-medial) (Fig. 4a) and Prominence (accented vs. ...
Citations
This study explores the phonetic nature of phonological stop voicing contrast in American English by investigating how phonetic implementation of the voicing contrast is modulated by the prosodic structure along the continuum of phonetic voicing. In particular, the present study examines (1) the effects of two kinds of prosodic strengthening that can arise with prosodic structuring, a boundary-related domain-initial strengthening (DIS) and a prominence-induced strengthening, and (2) the possible enhancement types of linguistic contrasts that can underlie prosodic strengthening. The phonetic voicing was estimated using the Integrated Voicing Index (IVI), taking into account both the voicing lag (positive VOT) and the voiced interval during the closure. Results obtained with initial stops in both trochaic and iambic words are encapsulated as follows. Under the influence of DIS, both voiced and voiceless stops were produced with an increase in voicelessness, showing an enhancement of structurally motivated syntagmatic (CV) contrast. The effect size was larger for voiced stops, yielding a boundary-induced phonetic reduction of voicing contrast. Under the influence of prominence (focus), both voiced and voiceless stops showed an increase in voicelessness only in trochaic words, but this time, it was voiceless stops that showed a far greater effect, resulting in a maximization of voicing contrast—i.e., an enhancement of paradigmatic contrast. Moreover, the reduced voicing for voiced stops under prominence even in the medial position indicates that voiced stops are realized in reference to the phonetic feature {vl. unaspirated} rather than {voiced}. These findings imply that seemingly non-contrastive low-level variation is indeed systematically modulated by the prosodic structure in reference to phonetic representations that regulate the phonetic implementation of phonological contrast in a given language. An alternative account is also discussed in the framework of Articulatory Phonology.
