Phonetic variability leads to gradient speech perception.

Area of Science:

• Psycholinguistics
• Auditory Perception
• Speech Processing

Background:

• Speech perception relies on categories but also accounts for acoustic variation within categories.
• This gradient system aids flexibility, ambiguity resolution, and plasticity in listeners.
• Prior research links phonetic variability to shallower categorization slopes, but the exact mechanism remains unclear.

Purpose of the Study:

• To investigate phonetic variability as a causal factor in speech perception gradiency.
• To clarify the relationship between acoustic input variance and perceptual categorization using a distributional learning paradigm.
• To resolve ambiguities in previous tasks by employing the Visual Analog Scale (VAS) task.

Main Methods:

• Three experiments utilized a distributional learning paradigm with varying voice onset time (VOT) distributions (high vs. low variance).
• The Visual Analog Scale (VAS) task was employed to assess perceptual gradiency.
• Participants (total n=337) were trained on specific VOT distributions, and their categorization slopes were measured.

Main Results:

• Experiment 1 indicated that variance increased trial-by-trial inconsistency rather than gradiency.
• Experiment 2, with more stimuli, provided robust evidence that higher input variability enhances perceptual gradiency.
• Experiment 3, including a no-learning baseline, ruled out adaptation to low variance as an explanation.

Conclusions:

• Variable phonetic input causally promotes increased perceptual gradiency in speech sound categorization.
• The findings support the role of statistical learning in shaping auditory perception.
• This research clarifies how exposure to acoustic variability influences the flexibility and sensitivity of speech processing.