Does phonological rule of tone substitution modulate mismatch negativity?

Abstract

Abstract This study examined whether the phonological substitution rule of tone sandhi modulates tone perception in the preattentive stage. Tone sandhi is commonly present in East Asian languages. An example from Mandarin is the Tone tone 3 sandhi rule: T3 is pronounced as T2 when followed by another T3 (33\u202f→\u202f23). Previous mismatch negativity (MMN) studies in Mandarin have reported a smaller amplitude or longer latency in standard-deviant pair consisting of T2 and T3 (T2-T3) than in T1-T3. The most widely accepted explanation for this is that T2 and T3 have steeper pitch slopes than T1. This study tested an alternative account based on the phonological rule that the frequent substitution that occurs between T2 and T3 results in reduced MMN. In Experiment 1, we first tried to replicate the finding in Mandarin. In Experiment 2, using both unskilled and skilled speakers, we tested a sandhi tone pair of very different pitch slopes in Taiwanese. Delayed peak latency of sandhi pair was evident in both languages but only in skilled speakers. Our results did not support the shared-pitch-slope account and were instead consistent with the argument that a language-specific phonological rule could modulate preattentive tone processing.