Virtual Semantics Circle: The inference towards the antonym: a methodological contribution

Vortragende(r)Nicolas Ruytenbeek
Institution(en) Ghent University
Datum 22.05.2020, 10-12 Uhr
Uhrzeit 10:00 Uhr

External participants: please contact Stephanie Solt for an invitation to this event!


When one says that John is not tall, we can understand it as a denial that John is tall, but we can also interpret this utterance as a statement that John is short. This type of inference is called  an “inference towards the antonym” (for short, ITA). To date, the nature and the extent to which the linguistic properties of adjectives determine the strength of ITAs remain unclear. In this talk, I first summarize the major findings of Ruytenbeek et al.’s (2017) experiments, namely, a stronger ITA effect for positive vs. negative adjectives, and a higher ITA strength asymmetry for morphological antonymic pairs. I discuss how these results fit with previous work on ITA effects. I also address methodological issues having to do with the operationalization of adjective positivity/negativity, of the evaluativity inherent to some adjectival pairs, and of the ITA itself. Then, I consider an alternative explanation for the fact that the linguistic constructions used in our experiments gave rise to lower acceptability scores when they included negative adjectives. This explanation, based on the “negative adjective complexity hypothesis” (NACH), will be rejected on the basis of response time data suggesting that constructions with negative adjectives do not take longer to process than their counterparts with positive adjectives. The final part of this presentation will center on the possible processing correlates of NACH. I will present a lexical decision task probing into response times differences between the two members of adjectival antonymic pairs. This study constitutes a first step towards the operationalization of the ITA in terms of lexical activation strength.