Assimilative phonetic parsing as perceptual grouping: The influence of attention

Phonetic context affects the perception of speech sound in two different ways, either making it more like the context or less like the context, in a dissimilative manner or an assimilative manner. Rysling et al. [AP&P 81(4), 1127–1146 (2019)] suggested that assimilation happens when listeners encounter spectral ambiguity, and then group upcoming spectral frequencies with the preceding sound. Following Murgia et al. [Vision Res. 126, 69–79 (2016)], which investigated attention in color perception, the current paper studies whether multi-stabile intentional grouping can cause assimilation while stabile unintentional grouping causes dissimilation in speech perception. Stimuli were ɛCV and CV where C is an ambiguous fricative drawn from a [s-ʃ] continuum and V is either [i] or [u]. In one condition, participants were told to focus on the ɛC portion in the ɛCV stimuli to identify the consonant, and in another, the participants were told to focus on the CV portion to identify the consonant and the vowel. In a third condition, the participants heard only CV stimuli and were told to identify the consonant and the vowel. The results indicate that intentional grouping with multi-stability triggers an assimilative parsing, which is consistent with Murgia et al.'s claim.