Short-term, not long-term, average spectra of preceding sentences bias consonant categorization

Abstract

Phoneme perception is influenced by spectral properties of surrounding sounds. For example, listeners perceive /g/ (lower F3 onset) more often after sentences filtered to emphasize high-F3 frequencies, and perceive /d/ (higher F3 onset) more often after sentences filtered to emphasize low-F3 frequencies. These biases are known as spectral contrast effects (SCEs). Much of this work examined differences between long-term average spectra (LTAS) of preceding sounds and target phonemes. Stilp and Assgari (2018 ASA) revealed that spectra of the last 500 ms of precursor sentences, not the entire LTAS, predicted biases in consonant categorization. Here, the influences of Early (before the last 500 ms) versus Late (last 500 ms) portions of precursor sentences on subsequent consonant categorization were compared. Sentences emphasized different frequency regions in each temporal window (Early = low-F3 emphasis, Late = high-F3 emphasis, and vice versa) naturally or via filtering. Entire-sentence LTASes predicted that SCEs would not bias consonant categorization, but instead responses were biased by the spectrum of the Late window. This was replicated when the Early window did not emphasize either frequency region but the Late window did. Results endorse closer consideration of patterns of spectral energy over time in preceding sounds, not just their LTAS.