When Order Matters and When It Doesn’t: Tracking the Time Course of Temporal Implicatures in Conjunctions Using ERPs
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Conjunctive sentences in the past tense are often interpreted as describing a sequence of events occurring in order. In pragmatics, this phenomenon is known as temporal implicature. In three ERP experiments, we track the time course of how temporal implicatures are derived in conjunctions that report unrelated events. We also examine whether the contextual relevance of temporal order influences the processing of temporal order.Participants observed game events, such as turning two cards face-up in sequence, and read conjunctive sentences describing these events in either the correct or reversed order. The relevance of event order was manipulated by varying the game pay-off: in Experiment 1, the sequence affected the pay-off, while in Experiment 2, it did not. Experiment 3 further enhanced the salience of order by introducing temporal connectives ("before" and "after") as alternatives to simple conjunctions.The processing dynamics of temporal order depended on both the global experimental context and the connective used. Reversed-order sentences triggered increased P600 amplitudes at the first noun where the order violation became detectable, regardless of pay-off relevance. This effect is argued to reflect effortful combinatorial processing that is due to the violation of the expected discourse organization. In contrast, the N400 was modulated by order only in Experiments 1 and 3, where it was explicitly relevant—either through task demands or the presence of temporal connectives. This supports the hypothesis that temporal implicatures can be processed incrementally as part of enriched meaning representation when sufficiently supported by context.Additionally, an early negativity (N100-P200 time-window) was observed at the connective "and" in reversed-order sentences, suggesting increased attentiveness to connectives used in order-incongruent sentences. For "before" in reversed-order conditions, a prolonged negativity effect around the N400 time-window is observed, indicating that order incongruence with 'before' was processed as a meaning-related prediction error. In contrast, for "after" the effect is diminished, arguably due to conflict between semantic requirements for order reversal and the discourse-based expectations for iconic order. Finally, a late positivity emerged at the second noun, in reversed-order "and" sentences, indicating the involvement of late inferential processing. This effect was present in experiments 2 and 3, where the task did not explicitly emphasize the order.These findings provide new insights into the time course of temporal implicatures and the interplay between discourse structure, contextual relevance, and linguistic cues during real-time processing.