Schluter, K. 1 , Politzer-Ahles, S. 2 & Almeida, D. 1
1 New York University Abu Dhabi
2 University of Oxford
Eulitz & Lahiri (2004) observed a larger mismatch negativity (MMN) is elicited in an oddball paradigm when the standard sound has a fully specified-instead of underspecified-representation. We use this asymmetry to test the laryngeal realist hypothesis (Iverson and Salmons 1995, among others) that laryngeal features are language-specific and phonetically based. Since Russian is argued to contrast /t/ and /d/ based on a [Voice] feature (Petrova, et al. 2006) rather than [Aspiration] as in English, we expect to find opposite MMN patterns in Russian and English. English shows greater MMNs for underspecified [d] deviants embedded in specified /t/ standards than for [t] deviants in /d/ standards (e.g., Hestvik & Durvasula 2016); Russian should show the opposite pattern.
We tested the MMN response to three different voiced/voiceless contrasts in Russian: stops (de~te), coronal fricatives (s~z) and multiple fricatives (fs?~vz?). Preliminary results from 7 participants suggest a trend towards the same asymmetric effect in each contrast: the voiced deviant embedded in voiceless standards appears to evoke a stronger response than the voiceless deviant in voiced standards.
As Russian seems to show the same MMN pattern as English for the voicing contrasts, the results may be incompatible with the laryngeal realist hypothesis.