Slide Slam S16 Sandbox Series
Discrimination of homophone forms – morphology, phonology or both?
Byurakn Ishkhanyan1, Mikkel Wallentin2,3, Line Burholt Kristensen1; 1Department of Nordic Studies and Linguistics, University of Copenhagen, 2School of Communication and Culture, Aarhus University, 3Department of Clinical Medicine - Center of Functionally Integrative Neuroscience, Aarhus University
In fMRI studies of language processing, morphology and phonology are mostly studied as two separate phenomena. Conditions with morphological anomalies (e.g. anomalously inflected verbs) are often contrasted with conditions with phonological anomalies (e.g. pseudohomophone spelling errors). This strand of research has shown distinct patterns for the two types of anomalies. Reading morphological anomalies is associated with increased activation of left inferior frontal gyrus (Carreiras et al. 2010; Carrieras et al. 2015), while phonological anomalies also tend to engage auditory areas (e.g. Newman & Joanisse 2011). In naturally occurring texts, the distinction between phonological and morphological anomalies is not that clearcut. In written Danish, the infinitive and the present tense of verbs are frequently confused. For verb stems with final -r, the infinitive and the present tense form are homophone – for other verb stems there is an audible difference between forms. For example there is an audible difference for overveje/overvejer (‘consider/considers’), but not for vurdere/vurderer (‘evaluate/evaluates’)In our fMRI study we investigated whether the processing of verbal inflection in Danish involves both morphological and phonological processes. Twenty-eight Danish participants (age 20-31) read 128 short stories while they performed a semantic/pragmatic coherence judgment task in the scanner. The stimuli sentences contained a critical verb presented in one of the four conditions: 1. Anomaly with homophones: Han må vurderer situationen (English: ‘He must evaluates the situation’) 2. Anomaly with non-homophones: Han må overvejer situationen (English:’He must considers the situation’) 3. Baseline for homophones: Han må vurdere situationen (English: ‘He must evaluate the situation’) 4. Baseline for non-homophones: Han må overveje situationen (English: ‘He must consider the situation’) Functional images were acquired in an EPI sequence on a Siemens Magnetom Trio 3-T MR scanner with a 32-channel head coil. The acquisition specifications were the following: TR = 1340 ms, TE = 29.6 ms, flip angle = 75°, voxel size = 1.8 x 1.8 x 1.8 mm. FWE corrected (p < 0.05) whole brain analysis showed no effect of grammar anomaly, no effect of homophony and no interaction. Four planned ROI analyses were conducted for: left IFG, primary and secondary auditory cortices, VWFA and left STG/MTG. The left pIFG, VWFA and pSTG were more activated in the non-homophone condition, compared to the homophone one. No other main effects or interactions were observed. In the coherence judgment task, there was lower accuracy for stories with anomalous inflection than those with correct inflection. Accuracy was also lower for non-homophone conditions compared to homophone conditions. Both behavioural and fMRI findings suggest that the non-homophone forms require extra processing effort, engaging extra activation in areas in the brain that are known for being involved in morphological processing (pIFG), phonological processing (pSTG) and written word recognition (VWFA). These results are an indication that processing of verbal inflections may both involve morphological and phonological processing.