The role of case in incremental argument interpretation: evidence from Punjabi

Abstract

Language comprehension is a continuous process of decoding multitudes of information types. The fact that the parser manages to achieve this feat at such speed in real time indicates an intricate neural architecture at play. Literature is replete with evidence that the role linguistic cues play in aiding incremental processing is non-trivial (Bornkessel & Schlesewsky, 2006). In this thesis, we explore the role of one such morphosyntactic cue, namely case markers, in a split-ergative, Indo-Aryan language Punjabi. Previous studies that have explored case markers report that the marker plays a substantial role in argument interpretations by formulating predictions about the upcoming constructions, aiding in thematic hierarchisation especially in the absence of the verb as well as in initiating structural or thematic reanalysis. Most observations in this regard however have been made using nominative-accusative languages, with the exception of only two languages namely, Basque which exhibits an ergative alignment (Zawiszewski & Friederici (2009) and D az, Sebasti an-Galles, Erdocia, Mueller & Laka (2011)) or Hindi, a split-ergative language ( Choudhary, Schlesewsky, Roehm, & Bornkessel-Schlesewsky (2009), Choudhary (2011)). Further, the evidence from this limited subset of ergative languages o ers observations di erent to those reported from accusative language types, thus highlighting the need to probe more into diverse languages. Through this thesis then, we broaden the coverage to include one such, previously unexplored language, Punjabi. This thesis embodies empirical evidence from three EEG/ERP experiments in Punjabi, a split-ergative, Indo-Aryan language. Given the variability in terms of the ERP components that have been reported across language types, these experiments were devised in order to ascertain if the underlying processing mechanism is similar at least for typologically similar languages, or do language-speci c properties modulate the processing neurophysiologically. Further, we explore experimentally, if (and how) the case markers di er in terms of their predictive capacity when used licitly and also if used illicitly. Finally, we probe to investigate if (and how) the processing mechanism di ers in case these predictions are not met. For this purpose, the fi rst experiment explored the interplay between case marking and the aspect at the position of the verb, to deduce if nominative and ergative case violations are treated the same way. The second experiment extended this further, incorporating the personbased split ergative pattern in order to explore how formal case-based mismatches are treated in comparison to the ergative and nominative case-based violations. The third experiment then explored the relation of case marking in agreement processing. Interestingly, in terms of the ERP components, all violations across the three experiments only rendered positivities. The ndings therefore suggest that while typologically similar languages do share certain processing similarities, idiosyncratic properties of the language seem to play a vital role neurophysiologically. Next, the experiments also reveal that ergative case does help in dependency formation, in comparison to the nominative case. Interestingly, this advantage is observed even when the ergative case marker is illicitly used. Furthermore, across all experiments, there seems to be a trend whereby native speakers nd the use of nominative case in the place of the ergative case acceptable but not vice-versa. Looking at these neurophysiological results in the light of language speci c features, we speculate that they are indicative of diachronic changes in the language and suggest a systematic decline in terms of reliability on case markers as a cue for incremental interpretation.