Time reference processing in Bosnian/Croatian/Serbian:An ERP study

Previous studies on time reference reported a P600 effect and sentence final negativities for temporal violations by simple and periphrastic verb forms (Baggio 2008; Bos, Dragoy, Stowe & Bastiaanse 2013; Dragoy, Stowe, Bos & Bastiaanse 2012; Qiu & Zhou 2012). The goal of the present study was to further investigate this phenomenon by examining a group of languages for which ERP data have not yet been reported. Typical language processing of time reference was examined in three South Slavic languages: Bosnian, Croatian, and Serbian (BCS). We report ERP data on the processing of future time reference violations in which the verb forms within the periphrastic construction do not match a time frame previously set by a temporal-lexical adverb. ERP responses to auxiliaries in the present tense with either congruent past time reference or incongruent future time reference were compared in two closely related experiments. Violations by a present tense verb in a past periphrastic construction yielded a negativity in the 700-1000 ms time window in Experiment 1, while no significant effect was observed in Experiment 2. The late negativity that occurred in Experiment 1 cannot be clearly attributed to time reference processing since this kind of effect has not been reported in previous studies. Based on several studies that found a similar effect and connected it to higher processing demands, we propose that the effect might be due to an increase in memory load. Regarding the absence of the P600 effect, we propose that time-reference violations in BCS may not be processed as fully ungrammatical when measured online, despite people judging them as ungrammatical in an offline judgement task. Such discrepancy between online and offline results may reflect the participants’ prescriptivist education rather than their genuine language processin