Phonetic cue weighting in perception and production

Abstract

Speech sound contrasts differ along multiple phonetic dimensions. During speech perception, listeners must decide which cues are relevant, and determine the relative importance of each cue, while also integrating other, signal-external cues. The comparison of cue weighting in perception and production bears on a range of theoretical issues including the processes underlying sound change, the time course of learning, the nature of cues, and the perception-production interface. Research examining the relative alignment of cue weighting across the modalities, on both a community and individual level, has revealed both parallels and asymmetries between the modalities. The extraordinarily wide range of ways that have been used to conceptualize and quantify cue weights reflects the inherent theoretical, methodological, and analytical differences between the two modalities. More consideration of the choices of analytical metrics, explicit discussion of the theoretical assumptions that underlie them, and systematic investigations of different types of cues will lead to more generalizable findings that can be incorporated into computational implementable models of speech processing. This article is categorized under:Linguistics > Language in Mind and Braincue weighting, phonetics, speech perception, speech production | INTRODUCTIONSpeech sound contrasts differ on multiple phonetic dimensions: for example, the English sounds /b/ and /p/ differ systematically in voice onset time (VOT, the amount of time between the stop release and onset of voicing), but also differ, albeit less reliably, in other dimensions including the duration of the stop closure and the fundamental frequency (f0, corresponding to perceived pitch) at the onset of voicing (Lisker, 1986). Phonetic cue weighting, or the relative use of these acoustic "cues," can be conceptualized and quantified in the context of both production and perception. For example, as shown in Figure 1, English speakers' productions of /b/ and /p/ show large and consistent differences in VOT, such that /b/ vs. /p/ category membership can by well-predicted using this cue alone. On the other hand, while /b/ is followed by a lower f0 than /p/ on average, this difference is much smaller and less consistent, such that f0 is only weakly predictive of category membership. This asymmetry between the two cues is reflected in perception: when asked to categorize sounds varying in aspiration duration and f0, listeners' responses are mainly determined by aspiration (the primary cue), with the value of f0 playing a secondary, albeit still detectable, role (Abramson & Lisker, 1985).

The Reliability of Individual Differences in VOT Imitation

“…That stability does not tell us whether some individuals can be classified as “imitators” in general: we do not know if the participants who strongly imitated these particular talkers’ lengthened VOT would also strongly imitate other phonetic features or other talkers, or if imitation varies based on various factors such as the feature’s salience or an individual’s personal experience with or sensitivity to a particular variable. In fact, there is some evidence that individuals do vary in their sensitivity to different phonetic features (Sanker, 2015; Schertz & Clare, 2019). However, having greater confidence in the reliability of any particular task does suggest that a lack of cross-feature or cross-task correlations in other studies reflects genuine differences in the individual differences utilized by different features or tasks.…”

“…While aggregate patterns of convergence are widely observed, it is also known to be the case that individuals sometimes differ substantially in whether or how much they converge toward an interlocutor or experimental model talker, a point which has attracted increasing attention in recent work (Babel, 2012; Pardo et al, 2018; Sonderegger et al, 2017; Zellou, 2017). Imitation is far from the only phenomenon for which such inter-individual variability can be observed; Yu and Zellou (2019) and Schertz and Clare (2019) provide valuable overviews of the many dimensions along which individuals may differ in their speech perception and production. While individual differences are frequently observed in the literature, there is still some question as to whether such differences observed on any given occasion represent stable, characteristic properties of those individuals and their linguistic systems (Kingston et al, 2015).…”

“…This suggests that it may not be possible to classify an individual as an across-the-board imitator, although it could plausibly be the case that individual propensities toward imitation are feature-specific but nonetheless stable. Relatedly, Schertz and Clare (2019) survey a range of both positive and null results in studies testing individual-level perception–production correlations for various linguistic features. Again, this may or may not undermine our confidence in the stability of individual differences; as Schertz and Clare (2019) discuss in detail, there are many reasons why perception and production might dissociate.…”

“…Relatedly, Schertz and Clare (2019) survey a range of both positive and null results in studies testing individual-level perception–production correlations for various linguistic features. Again, this may or may not undermine our confidence in the stability of individual differences; as Schertz and Clare (2019) discuss in detail, there are many reasons why perception and production might dissociate.…”

Natural Leaders: Some Interlocutors Elicit Greater Convergence Across Conversations and Across Characteristics

“…Failing to find a correlation does not necessarily mean that it does not exist. Schertz and Clare (2020) note that some of the varied results across studies may be due to differences in task design and analytical choices.…”