Yasu Sudo: Implicature priming, inverse preference, and context adaptation
Implicature priming, inverse preference, and context adaptation
(based on joint work with Paul Marty, Jacopo Romoli and Richard Breheny)
Previous studies that use priming to investigate scalar implicatures (SIs) observe that SIs arise more often after strong primes—priming trials that force readings with SIs—than after weak primes—priming trials that force readings without SIs. To explain these results, Bott and Chemla (2016) among others claim that what is primed in implicature priming is certain aspects of the computational mechanism for SI, which is commonly assumed to involve referencing an alternative sentence and negating it. Bott and Chemla argue that referencing an alternative gives rise to a particularly robust priming effect via increasing the salience of the alternative (see also Rees & Bott 2017). The current study argues against this hypothesis, by systematically comparing the strong and weak primes to baseline conditions that involve no priming whatsoever. Our results indicate that the effects of implicature priming in this experimental paradigm are ‘inverse preference effects’ in the sense that non-default interpretations have more robust priming effects than default interpretations. As Bott and Chemla discuss, inverse preference in the general sense is compatible with their salience-based account, but we argue that it falls short of providing a full explanation of the entire pattern we observe in our data, especially regarding the large priming effects of weak primes for cases where the SI is computed by default. We propose an alternative explanation based on the idea of ‘context adaptation’.
Watch the talk via Teams:
Meeting ID: 390 137 741 961