Publications

The influence of a talker's first language (L1) on the production of their second (L2) can complicate the task of word recognition. Nevertheless, previous research has shown that listeners demonstrate remarkable speed and plasticity in adapting to L2-accented talkers. Moreover, listeners are able to generalize this learning to novel L2-accented talkers. There are two competing explanations for these talker-independent adaptation effects: one argues that increasing variation during exposure facilitates generalization to a novel test talker (exposure-to-variability hypothesis), while the other argues that increasing acoustic-phonetic similarity between the exposure and test talkers facilitates generalization (similarity-based hypothesis). We conducted three experiments to compare the effects of exposure variability and exposure-test similarity on generalization. During exposure, monolingual English participants performed an auditory lexical decision task with multisyllabic English tokens spoken by three Spanish-accented talkers. After exposure, we tested performance on a novel Spanish-accented talker with a cross-modal matching task (Experiments 1 and 2) or a primed cross-modal lexical decision task (Experiment 3). Exposure variability increased both identity and competitor priming, consistent with a relaxation of phonetic categorization criteria. Exposure-test similarity decreased competitor priming without changing identity priming, while exposure uniformity increased identity priming without changing competitor priming, providing some evidence of phonetic recalibration in these conditions. Our results provide equivocal support for the similarity-based hypothesis and against the exposure-to-variability hypothesis.

Listeners use their knowledge about a talker to guide speech perception. In the present study, we manipulated both the familiarity and predictability of this talker-specific knowledge. Twenty-two listeners from western Pennsylvania completed an audio-visual lexical decision task while EEG was recorded. Critically, listeners were introduced to the talkers beforehand, with short videos establishing the talker's U.S. English accent identity: Mainstream, more similar to the participants themselves; Southern, a relatively less familiar variety; and Unpredictable, which switched between Mainstream and Southern accents. During the task, listeners watched these talkers produce tokens with accents that either aligned with or violated their accent identity. Behaviorally, alignment between talker and token accent influenced performance on Southern-accented tokens, while accuracy on Mainstream-accented tokens was nearly at ceiling regardless of talker accent identity. Neurally, accent familiarity had the largest effect pre-speech (in response to the visual presentation of the talker), while both predictability and familiarity impacted processing of the speech signal itself. Overall, our results suggest that listeners use information about a talker's accent, even when it is unfamiliar or unpredictable, to alter their listening strategies for more successful speech recognition.

Listeners vary in their sensitivity to acoustic-phonetic variation. In this study, we explored different ways of measuring such sensitivity. Study 1 used a visual analogue scale (VAS) task that elicited ratings for a set of four phonetic continua. Study 2 used a two-alternative forced choice task with eye tracking, which provided both binary response data and fine-grained fixation data for the same phonetic continua as study 1. In both studies, we asked how patterns of phonetic perception changed over the course of the adult lifespan by comparing a younger adult sample (ages 18-25) to an older adult sample (age 50+). In addition, we conducted several assessments to examine how language skill, hearing acuity, and speech-in-noise perception were associated with individual differences in phonetic categorization. We found that older adults displayed more categorical (as opposed to more gradient) patterns of perception than younger adults, particularly in the VAS task, where age-related differences persisted even after controlling for hearing loss and speech-in-noise perception. There was also evidence of more categorical patterns of perception among older adults in the eye tracking measures. Overall, these data suggest that VAS tasks may be particularly sensitive to both age-related factors and differences in language skill.

Cross-linguistic similarity is a term so broad and multifaceted that it is not easily defined. The degree of overlap between languages is known to affect lexical competition during online processing and production, and its relevance for second language acquisition has also been established. Nevertheless, determining what makes two languages similar (or not) increases in complexity when multiple levels of the linguistic hierarchy (e.g., phonology, syntax) are considered at once. How can we feasibly account for the patterns of convergence and divergence at each level of representation, as well as the interactions between them? The growing field of network science brings new methodologies to bear on this longstanding question. Below, we summarize current network science approaches to modeling language structure and discuss implications for understanding various linguistic processes. Critically, we stress the particular value of multilayer techniques, unique and powerful in their ability to simultaneously accommodate an array of node-to-node relationships.

We investigated online auditory comprehension of dialectal variation in English syntax with event-related potential (ERP) analysis of electroencephalographic data. The syntactic variant under investigation was the double modal, comprising two consecutive auxiliary verbs (e.g., might could). This construction appears across subregional dialects of Southern United States English and expresses indirectness or uncertainty. We compared processing of sentences with attested double modals and single modals in two groups of young adult participants: listeners who were either familiar (Southern) or unfamiliar (Unmarked) with double modal constructions. Both Southern and Unmarked listeners engaged rapid error detection (early anterior negativity) and sentence-level reanalysis (P600) in response to attested double modals, relative to single modals. Offline acceptability and intelligibility judgments reflected dialect familiarity, contrary to the ERP data. We interpret these findings in relation to usage-based and socially weighted theories of language processing, which together capture the effects of frequency and standard language ideology.

Does gesturing help speakers find the right words? According to several theories of speech-gesture relationships, iconic gestures should facilitate speech production, but beat gestures should not. Here we tested the effects of gesturing on word production in two experiments. Participants produced low-frequency words from their definitions while instructed to perform beat gestures, iconic gestures, or while not given any instructions about gesturing (baseline condition). Compared to baseline, participants were faster to produce the target words while performing beat gestures, bimanually or with their left hand alone, but they were slower to produce the target words when instructed to perform iconic gestures. Results provide the first evidence that beat gestures can help speakers produce words. This benefit may arise from the fact that gestures are motor actions, rather than from any special properties of gestures, per se.