Representation of speech variability.

Area of Science:

• Cognitive Science
• Psycholinguistics
• Sociolinguistics

Background:

• Speech signals convey both linguistic content and social-indexical information about the speaker.
• Listeners form detailed representations of speech, linking it to linguistic and social categories.
• Existing research uses various methods to study cognitive-perceptual representations of social-indexical speech information.

Purpose of the Study:

• To review current understanding of how listeners perceive social-indexical information from speech.
• To identify limitations in current research, including underrepresentation of non-binary social categories and intersectionality.
• To propose future research directions incorporating diverse scholars and examining real-world interaction impacts.

Main Methods:

• Review of existing literature utilizing forced-choice categorization, rating, and free classification methods.
• Analysis of studies on listener accuracy in identifying speaker characteristics like age, sex, gender, race, and ethnicity.
• Examination of research gaps concerning nuanced social constructs and intersecting identities.

Main Results:

• Listeners accurately identify some speaker characteristics (e.g., native language, age, sex, gender).
• Listener accuracy is lower or more variable for characteristics such as sexual orientation, dialect, race, and ethnicity.
• Current research often lacks consideration of non-binary or intersectional social identities.

Conclusions:

• Listeners extract significant social-indexical information from speech, though with varying accuracy.
• There is a need to update research to include contemporary understandings of social constructs and diverse identities.
• Future research should explore the interplay of multiple social categories and their influence on interpersonal interactions.