Anaphoric reference to quantified antecedents: an event-related brain potential study.

We report an event-related brain potential (ERP) study examining how readers process sentences containing anaphoric reference to quantified antecedents. Previous studies indicate that positive (e.g. many) and negative (e.g. not many) quantifiers cause readers to focus on different sets of entities. For example in Many of the fans attended the game, focus is on the fans who attended (the reference set), and subsequent pronominal reference to this set, as in, Their presence was a boost to the team, is facilitated. In contrast, if many is replaced by not many, focus shifts to the fans who did not attend (the complement set), and reference to this set, as in, Their absence was disappointing, is preferred. In the current studies, the electroencephalogram (EEG) was recorded while participants read positive or negative quantified statements followed by anaphoric reference to the reference set or complement set. Results showed that the pronoun their elicited a larger N400 following negative than positive quantifiers. There was also a larger N400 on the disambiguating word (presence/absence) for complement set reference following a positive quantifier, and for reference set reference following a negative quantifier. Findings are discussed in relation to theoretical accounts of complement anaphora.

The on-line processing of written irony.

We report an eye-tracking study in which we investigate the on-line processing of written irony. Specifically, participants' eye movements were recorded while they read sentences which were either intended ironically, or non-ironically, and subsequent text which contained pronominal reference to the ironic (or non-ironic) phrase. Results showed longer reading times for ironic comments compared to a non-ironic baseline, suggesting that additional processing was required in ironic compared to non-ironic conditions. Reading times for subsequent pronominal reference indicated that for ironic materials, both the ironic and literal interpretations of the text were equally accessible during on-line language comprehension. This finding is most in-line with predictions of the graded salience hypothesis, which, in conjunction with the retention hypothesis, states that readers represent both the literal and ironic interpretation of an ironic utterance.