Understanding how listeners categorize highly variable speech signals into discrete units is a central question in speech perception research. Traditionally, it was assumed that listeners disregard variation and focus solely on the category itself, a concept known as categorical perception (Liberman et al., 1957). However, when faced with highly variable speech signals, such as unfamiliar accents, ignoring variation can hinder comprehension and increase processing difficulty. In this ongoing online study (expected n = 120), we investigate English-speaking adult listeners’ speech categorization patterns using a continuous measure—the Visual Analogue Scaling Task—to examine their sensitivity to spoken language variation (Kutlu etal., 2022; Apfelbaum et al., 2022). Participants’ exposure to varied accents is quantified through an extensive social network questionnaire. This questionnaire captures the extent and diversity of accents they regularly encounter (Kutlu et al., 2024). Subsequently, participants listen to sentences spoken in a variety of unfamiliar accents and transcribe them. Sentence intelligibility is assessed by scoring the accuracy of these transcriptions. Preliminary results show that sound categorization consistency predicts novel accent perception such that listeners who are more consistent in their sound categorization are more accurate in transcribing novel accents. The implications for exposure to diversity will be discussed.
Ethan Kutlu (Tue,) studied this question.