Muralikrishnan, R. & Bornkessel-Schlesewsky, I.
University of Marburg
It has been proposed that the language processing mechanism strives to keep the arguments of the verb distinct and that it uses several cross-linguistically motivated cues with language-specific weightings to accomplish this (Bornkessel-Schlesewsky & Schlesewsky, 2009). Mechanisms to distinguish the arguments, such as case-marking, may indicate several roles for the argument depending upon the verb. For instance, a Tamil dative noun can signify an indirect object, a subject or a destination. Dative subjects occur with dative-stative verbs, which show no person/number/gender agreement with the dative subject but only default-agreement (3rd-sg-neuter). Crucially, these verbs also have an alternative literal meaning if the subject is nominative, with no dative noun in the sentence. We investigated the processing differences between ditransitive (DI) and dative-stative verbs (DS) in an ERP study on Tamil in which identical auditory stimuli (DI vs. DS) were presented in four contexts: CQ: Correct; question priming the target verb; NQ: Neutral; no verb-specific information; VQ: Verb mismatch; same verb-class; WQ: Verb mismatch; wrong verb-class. We found an early positivity in CQ and a later positivity in VQ and WQ in both sentence types, which is in line with previous findings of early and late P300 effects when stimuli are categorised as meeting or violating expectations (Roehm et al., 2007). An additional P600 was observed for dative-statives regardless of context. We argue that this effect stems from a bottom-up agreement mechanism by which the processing system seeks to identify an agreeing argument. Since there is no potential agreeing nominative, however, the stative meaning is derived via an enrichment process (see Burkhardt, 2006, for enrichment-based P600 effects). These findings thus provide an initial indication of how default agreement is processed in sentences with non-nominative subjects.