Event-related potentials in pragmatic priming

Abstract

Priming of pragmatic enrichment has been found in behavioural studies. We extend this by examining the neural correlates of priming for two implicature categories, quantifiers and disjunctions. Participants engaged in a primed sentence-picture matching task where they were presented with a sentence (e.g., "some of the letters are Bs") followed by a picture. In prime trials the pictures were either consistent with an enriched interpretation (some but not all) or a basic interpretation (some and possibly all) of the sentence. The pictures in target trials were always consistent with the enriched interpretation. Using ERPs, we found a priming effect on the picture reflected in a reduced positivity for quantifiers when the preceding trial had an enriched interpretation, and no effect for disjunction. The pragmatic priming effect can be dissociated from expectation-based processes. It suggests that abstract derivation processes are primed during pragmatic alignment.

Some sentences prime pragmatic reasoning in the verification and evaluation of comparisons.

“…The starting point of our investigation was the finding that underinformative “some” statements have been found to elicit high rates of pragmatic responses in numerous sentence verification studies (e.g., “Some elephants have trunks” tends to be judged as FALSE), whereas Rubio-Fernandez et al (2017) observed that underinformative comparisons elicited much lower rates of pragmatic responses (see Figure 1 for a summary of this literature). Given the stark difference between the rates of pragmatic responses elicited by underinformative “some” statements and comparisons, and recent evidence of pragmatic enrichment priming (Bott & Chemla, 2016; de Carvalho et al, 2016; Rees & Bott, 2018; Rees et al, 2019), we predicted that introducing canonical “some” and “all” statements in a sentence verification and sentence evaluation task would result in an informativity bias, whereby categorization and comparison statements would be interpreted pragmatically more often.…”

“…Given the relatively high rates of pragmatic responses that have been reported in scalar implicature studies with “some” and “all” (see Figure 1), we wanted to investigate whether a sentence verification task with categorization and comparison statements would result in higher rates of pragmatic responses to the comparisons if “some” and “all” statements were intermixed in the materials. Recent scalar implicature studies have employed structural priming paradigms to investigate the nature of pragmatic enrichment, revealing that participants were more likely to derive an enriched interpretation of a scalar expression (e.g., “Some but not all of the symbols are stars”) if the previous trial required that they derived such an interpretation (Bott & Chemla, 2016; de Carvalho et al, 2016; Rees & Bott, 2018; Rees et al, 2019). The results of these studies confirm that different types of enrichment normally considered pragmatic or semantic can be primed in the same way as syntactic structures.…”

“…Adding “some” and “all” statements to categorization and comparison statements in a sentence verification task could be understood as a way of priming pragmatic enrichment. However, it must be noted that the experimental designs employed to prime pragmatic enrichment always paired the prime and target trials (see Bott & Chemla, 2016; de Carvalho et al, 2016; Rees & Bott, 2018; Rees et al, 2019), such that prime trials (requiring enrichment) immediately preceded target trials (where enrichment was optional). By contrast, our experimental design simply intermixes weak and strong statements of both kinds (i.e., canonical “some”/”all” statements and comparison/categorization statements) in a randomized order, not necessarily presenting “some” trials prior to comparison trials in order to elicit structural priming.…”