Poster D67, Monday, March 27, 8:00 – 10:00 am, Pacific Concourse
Phonological rules affect natural speech processing
Miriam Munoz1, Michael Key2, Ahren B. Fitzroy1, Lisa D. Sanders1; 1University of Massachusetts, 2University of Maryland
There is ample evidence that listeners have knowledge of phonological rules. However, the extent to which these rules affect typical language processing is unknown. Event-related potentials (ERPs) have been pivotal in investigating the distinct neural systems supporting the use of syntactic (LAN and P600) and lexical/semantic (N400) knowledge during natural speech processing. To date, there is no such tool to investigate the use of phonological rules during natural language processing. In this study, we measured behavioral (experiment 1) and ERP (experiment 2) responses to phonological violations in stories presented as connected, natural speech. The phonological violations were incorrect allomorphs (e.g., dog-/s/ and walk-/d/) included in a small proportion of sentences. Canonical (e.g., dog-/z/ and walk-/t/) and violation versions were constructed by splicing single phonemes from permissible contexts. When asked to press a button in response to “anything abnormal” about the stories, adults (N=20) failed to report the phonological violations, even though they detected missing phonemes that resulted in syntactic violations (e.g., She call every morning). In the ERP experiment, adults (N=20) listening to the stories for comprehension showed an early, central positivity in response to the phonological violations. Specifically, cluster-based permutation analysis, a data-driven approach designed to correct for multiple comparisons, revealed a significant cluster 112-130 ms after the onset of an incorrect allomorph over central, right and medial regions. This early phonological positivity (EPP) is entirely different from ERPs elicited by violations of syntax (LAN and P600) and lexical/semantics (N400). Phonological rules influence natural speech processing early and automatically.
Topic Area: LANGUAGE: Other