Contrastive Hyperarticulation of Word-Initial Stops in a Conversational Corpus
AuthorNelson, Noah Richard
voice onset time
AdvisorWedel, Andrew B.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractIn this dissertation, I seek to clarify the nature of contrastive hyperarticulation. Unlike other kinds of hyperarticulation found in the literature, such as clear-speech hyperarticulation or the hyperarticulation of Lindblom’s (1990) Hyper-articulation and Hypo-articulation (H & H) Theory, contrastive hyperarticulation is defined not in terms of articulatory durations or gestural extents, but rather in terms of phonetic realizations that enhance the perceptual distance between the target and competing lexical forms. Despite a growing number of studies over the past decade that have investigated what might be described as contrastive hyperarticulation, the phenomenon is still not very well defined or understood. For example, some researchers have suggested that the total number of lexical-phonological neighbors to a given target word is correlated with hyperarticulation of individual phonetic cues in that target word (e.g., Fox et al., 2015). Other researchers, however, have suggested that only the existence of the specific lexical-phonological neighbor that contrasts with the target word in the measured phonetic cue will correlate with hyperarticulation of that phonetic cue (e.g., Wedel et al., 2018). In addition to uncertainty regarding the kinds of lexical competition that trigger contrastive hyperarticulation, researchers have also suggested different phonetic outcomes of this competition-driven hyperarticulation. Some researchers have suggested that, as a form of hyperarticulation, contrastive hyperarticulation can only be realized in the form of phonetic enhancement (e.g., Goldrick et al., 2013; Fricke et al., 2016). Other researchers, however, have suggested that the phonetic outcomes of contrastive hyperarticulation can include instances of phonetic reduction (such as decreased phonetic durations) in addition to more traditionally defined hyperarticulation (such as increased phonetic durations; Seyfarth et al., 2016; Wedel et al., 2018). In this dissertation I clarify both the triggers and outcomes of contrastive hyperarticulation in two studies. The first study was designed to sample a large hypothesis space regarding the kinds of lexical competitors that might trigger contrastive hyperarticulation of word-initial stop voicing contrasts in spontaneous English speech. Contrastive hyperarticulation was identified in the form of enhanced voice onset time contrasts between voiced and voiceless stops (longer voice onset times in voiceless stops, and shorter voice onset times in voiced stops). A number of possible metrics of lexical competition were evaluated as predictors of these voice onset time realizations based on the literature, including overall neighborhood density, the existence of the initial stop voicing minimal pair competitor, and a number of metrics varying in the position or type of competition between target and competitor. It was found that the existence of the initial stop voicing minimal pair competitor was a better predictor of voice onset time measurements than any other metric of lexical competition for nearly all statistical models. The outcomes of contrastive hyperarticulation were also partially addressed in this study in that voiced stops in words with voiceless stop minimal pairs were realized with shorter voice onset times than voiced stops in words without such minimal pairs. This is indicative of shorter phonetic durations that are used to enhance a lexical-phonological contrast (see also Seyfarth et al., 2016; Wedel et al., 2018). The second study was designed to further investigate the phonetic outcomes of contrastive hyperarticulation by measuring a secondary cue to the lexical-phonological contrast in question. Specifically, word-initial stop voicing in English is primarily signaled by voice onset time, but is also signaled by the fundamental frequency that follows the stop at the onset of voicing. In this study, I investigated fundamental frequency measurements following voiced and voiceless stops to see whether they are also contrastively hyperarticulated as a function of stop voicing minimal pair competitor existence. It was found that fundamental frequency correlated positively with voice onset time in voiceless stops of words with initial voiced stop minimal pairs, but this correlation did not hold for words without stop voicing minimal pairs. In other words, fundamental frequency appears to be selectively hyperarticulated alongside the primary voice onset time cue only for voiceless stops in words with voiced stop minimal pairs. This suggests that the effect is not intrinsic, but rather is selectively applied in cases where the phonetic contrast will serve a lexical-phonological contrast. The results of these studies were taken to support the notion that contrastive hyperarticulation is driven by lexical-phonological competition that is rooted in phonetic contrast. As voice onset time is a major cue to initial stop voicing in English, it is recruited contrastively in words with initial stop voicing minimal pairs. As fundamental frequency, however, is an ancillary cue, it is recruited more selectively in service of the lexical-phonological contrast. These results are considered alongside other results related to contrastive hyperarticulation, including results showing that multiple cues to a given lexical-phonological contrast can be hyperarticulated contrastively, with both cues moving in opposing directions on either side of the lexical-phonological contrast (Seyfarth et al., 2016).
Degree ProgramGraduate College