You are viewing the SNL 2018 Archive Website. For the latest information, see the Current Website.


Poster D14, Friday, August 17, 4:45 – 6:30 pm, Room 2000AB

Phonotactic Rule-Learning Without Semantics: An EEG study

Enes Avcu1, Ryan Rhodes1;1University of Delaware

Artificial grammar learning (AGL) studies have been widely used for testing the learnability of phonological agreement patterns. It has been clearly shown that learners can extract adjacent and non-adjacent dependencies with relatively short training at the behavioral level. Less is known about how these patterns are encoded at the neurophysiological level. Domahs et al. (2009) and Moore-Cantwell et al. (forthcoming) reported a higher amplitude Late Positive Component (LPC) to novel words that violated a learned phonotactic constraint than novel words that satisfied it. The LPC has been reported in response to violations of many kinds of long-distance sequencing rules, both syntactic and phonotactic. However, previous AGL studies have conflated long-distance rules with semantic information. The aim of the current study is to observe an LPC to a phonotactic rule violation in the absence of semantic information, using a sibilant harmony rule which is an attested long-distance harmonic pattern. Our finding is that once syntactic and semantic components are excluded from training, the LPC is not elicited in response to novel words violating the phonotactic pattern. 15 monolingual English speakers participated in an AGL experiment in which ERPs were recorded. All training and test stimuli had three syllables of the form of CV.CV.CVC, with sibilants ([s, ʃ]) as the first, third, and last consonant. All words were either “harmonic” (all sibilants identical) or “disharmonic” (mixed [s] and [ʃ]). During training, participants listened to and repeated sibilant harmonic words. During testing, participants rated the likelihood that each word followed the rule they had learned. One-third of the test words were present during training, one-third of the words were novel and sibilant-harmonic, and the last one-third were novel and disharmonic. The total duration of both training and testing was about 75 minutes. Behavioral results show that there was a significant mean rating difference between novel-harmonic and disharmonic words (t(14)=3.685, p=0.002, d=0.951, 1-β=0.968). The stimuli elicited a clear auditory evoked potential (AEP). However, we observed no LPC following the stimulus or response, and there was no significant difference in ERP response to harmonic and disharmonic words. These results indicate that the LPC is not purely a response to violations of non-local sequencing rules. We observed a difference in behavioral response to harmonic and disharmonic words, indicating that the rule was learned. However, we saw no difference in brain response between the two conditions. In fact, a late positive component was entirely absent from the data. Although the harmonic pattern was learnable in the absence of semantic information, violations of the rule in this paradigm did not elicit the expected LPC. This suggests that the LPC, an index of syntactic and phonotactic rule violations, is also dependent on semantic integration. References Domahs, Ulrike, W. Kehrein, J. Kraus, R. Wiese, and M. Schlesewsky. (2009). Event-related potentials reflecting the processing of phonological constraint violations. Language and Speech 52: 415–435. Moore-Cantwell, Claire, J. Pater, R. Staubs, B. Zobel and L. Sanders. (forthcoming). Event-related potential evidence of abstract phonological learning in the laboratory.

Topic Area: Phonology and Phonological Working Memory

Back