Periodic Reporting for period 1 - PerMSAA (Perceptual Mechanisms that Support Accent Accommodation)
Berichtszeitraum: 2023-08-01 bis 2025-07-31
To understand spoken language, listeners must map complex acoustic input onto linguistic representations in their mental lexicons. While this process often appears effortless, the acoustics of spoken language can vary dramatically between speakers. In particular, variation in how second-language (L2) speakers produce the sounds of a language (i.e. ‘foreign’ accents) can increase the difficulties of speech perception. Prior work shows that listeners are highly adaptable, and can rapidly accommodate unfamiliar variation in accented speech, but what remains unclear is how listeners rapidly adapt to accented talkers. In the European community, inter-language and inter-accent communication is increasingly pervasive.
The current project’s pathway to impact was via acquisition of knowledge that can combat societal biases. The burden of improving communication among first language (L1) and second language (L2) speakers is often placed on the L2 speakers, who are expected to “reduce” their accents. However, prior research indicates that L1 listeners can also contribute to this process via accent accommodation. To understand the mechanics of this process, we must first understand the mechanism(s) that support listeners’ adaptation to unfamiliar accents. In many international settings (e.g. university classrooms and business meetings), these tools would have enormous potential for improving communication. Given the pervasive increase in inter-language and inter-accent communication in the European community, this initiative was an important step towards improving overall communication among L1 and L2 speakers.
The scale of this project was limited, but the impact was large in terms of understanding cognition of language. The primary aim of the proposed project was to improve our theoretical understanding of the perceptual mechanisms that support accent accommodation, and that aim was met.
The current project’s pathway to impact was via acquisition of knowledge that can combat societal biases. The burden of improving communication among first language (L1) and second language (L2) speakers is often placed on the L2 speakers, who are expected to “reduce” their accents. However, prior research indicates that L1 listeners can also contribute to this process via accent accommodation. To understand the mechanics of this process, we must first understand the mechanism(s) that support listeners’ adaptation to unfamiliar accents. In many international settings (e.g. university classrooms and business meetings), these tools would have enormous potential for improving communication. Given the pervasive increase in inter-language and inter-accent communication in the European community, this initiative was an important step towards improving overall communication among L1 and L2 speakers.
The scale of this project was limited, but the impact was large in terms of understanding cognition of language. The primary aim of the proposed project was to improve our theoretical understanding of the perceptual mechanisms that support accent accommodation, and that aim was met.
Two experiments were completed. For the first experiment, I recruited 287 participants online and randomly assigned them across nine conditions to examine the research questions (i.e. how do listeners adapt to real foreign accent?). Data was prepared and analyzed with logistic mixed-effects models (and summary statistics). I then drafted the paper, presented it at my home institution for feedback and added to my analyses to further clarify the results. I submitted the manuscript to the Journal of Experimental Psychology: Human Perception and Performance, revised and resubmitted two times, and then the paper was accepted for publication. It was published in February 2025 and I have presented it another two times to scientific audiences since.
In terms of the research findings for Experiment 1, we found evidence to support the primary hypothesis, that listeners would recalibrate their phonemic category boundaries following exposure to (real) L2 accent. In some conditions this was more substantial; the bilabial trainings and trainings with a greater density of tokens of a specific place of articulation were most effective. We also found that multilingual experience (specifically, experience with English language for listening to English-accented Spanish) interacted with adaptation.
The second experiment recruited a total of 98 participants to complete the experiment onsite; here, we adapted our original design after pilot testing, determining that a between-subject manipulation was necessary (thus, 32-33 participants per each of three groups was collected). Data has been prepared and analyzed. We presented this data in June 2025 and received insights from coworkers and other scientists; we are now conducting follow-up analyses. We are also preparing the manuscript to submit in October 2025 to a journal. We are close to the proposed timeline, perhaps one to two months behind schedule. Our pupillometry results indicate a non-directional benefit of exposure to accent. We see some phonemic recalibration (directional) but also some non-directional expansion of the category consistent with criteria relaxation, as predicted
In terms of the research findings for Experiment 1, we found evidence to support the primary hypothesis, that listeners would recalibrate their phonemic category boundaries following exposure to (real) L2 accent. In some conditions this was more substantial; the bilabial trainings and trainings with a greater density of tokens of a specific place of articulation were most effective. We also found that multilingual experience (specifically, experience with English language for listening to English-accented Spanish) interacted with adaptation.
The second experiment recruited a total of 98 participants to complete the experiment onsite; here, we adapted our original design after pilot testing, determining that a between-subject manipulation was necessary (thus, 32-33 participants per each of three groups was collected). Data has been prepared and analyzed. We presented this data in June 2025 and received insights from coworkers and other scientists; we are now conducting follow-up analyses. We are also preparing the manuscript to submit in October 2025 to a journal. We are close to the proposed timeline, perhaps one to two months behind schedule. Our pupillometry results indicate a non-directional benefit of exposure to accent. We see some phonemic recalibration (directional) but also some non-directional expansion of the category consistent with criteria relaxation, as predicted
We found evidence of the hypothesized recalibration mechanism. Further research will be needed to further elucidate the bounds of the mechanism, and how it is supported by domain general working memory and attention.
To summarize: In Experiment 1, we found evidence to support the primary hypothesis, that listeners would recalibrate their phonemic category boundaries following exposure to (real) L2 accent. In some conditions this was more substantial; the bilabial trainings and trainings with a greater density of tokens of a specific place of articulation were most effective. We also found that multilingual experience (specifically, experience with English language for listening to English-accented Spanish) interacted with adaptation. In Experiment 2, we found evidence with pupillometry that indicated a non-directional benefit of exposure to accent. We see some phonemic recalibration (directional) but also some non-directional expansion of the category consistent with criteria relaxation, as predicted.
The innovative nature of our design, which used eye-tracking and pupillometry, gained new insights to the cognitive mechanism. The intellectual impact for this research area is strong.
To summarize: In Experiment 1, we found evidence to support the primary hypothesis, that listeners would recalibrate their phonemic category boundaries following exposure to (real) L2 accent. In some conditions this was more substantial; the bilabial trainings and trainings with a greater density of tokens of a specific place of articulation were most effective. We also found that multilingual experience (specifically, experience with English language for listening to English-accented Spanish) interacted with adaptation. In Experiment 2, we found evidence with pupillometry that indicated a non-directional benefit of exposure to accent. We see some phonemic recalibration (directional) but also some non-directional expansion of the category consistent with criteria relaxation, as predicted.
The innovative nature of our design, which used eye-tracking and pupillometry, gained new insights to the cognitive mechanism. The intellectual impact for this research area is strong.