Search

• We examined neural responses to sentences with double violations. • Objects were incongruent on lexico-semantic grounds, as well as referential interpretation. • A P200-N400 effect was observed (not a P600 effect, as expected). • ERP effects confirm that semantic anomaly can be derived from independent sources. A key aspect of linguistic communication involves semantic reference to objects. Presently, we investigate neural responses at objects when reference is disrupted, e.g., "The connoisseur tasted *that wine "... vs. " ...* that roof ..." Without any previous linguistic context or visual gesture, use of the demonstrative determiner "that" renders interpretation at the noun as incoherent. This incoherence is not based on knowledge of how the world plausibly works but instead is based on grammatical rules of reference. Whereas Event-Related Potential (ERP) responses to sentences such as "The connoisseur tasted the wine ..." vs. "the roof " would result in an N400 effect, it is unclear what to expect for doubly incoherent " ...* that roof ...". Results revealed an N400 effect, as expected, preceded by a P200 component (instead of predicted P600 effect). These independent ERP components at the doubly violated condition support the notion that semantic interpretation can be partitioned into grammatical vs. contextual constructs. [ABSTRACT FROM AUTHOR]

We report two EEG/ERP experiments that examined processing of repeated name (e.g., Bill; Experiment 1) and pronoun (e.g., he; Experiment 2) subject anaphors to single antecedents (e.g., Bill) and to antecedents embedded in a conjunction (e.g., Bill and Mary) within sentences and discourses. Experiment 1 replicated previous reports of repeated references to single antecedents eliciting greater N400 negativity than repeated references to conjoined antecedents within sentences, and extended these results to cross-sentence (discourse) references. Experiment 2 found that pronouns also elicited greater N400 negativity following single than conjoined antecedents. In both experiments, references to conjoined antecedents elicited greater frontal negativity than references to single antecedents in both sentences and discourse. Our results indicate that, in processing subject anaphors, the N400 is an index of reference predictability rather than a marker of the fit between antecedent salience and reference form, and that frontal negativity marks referential ambiguity elicited by conjoined phrases. [ABSTRACT FROM AUTHOR]