Asymmetrical inference towards the antonym: disentangling the effects of dimensional and evaluative polarity on negated relative gradable adjectives
Year: 2026
Type: article
Abstract: This paper examines the polarity asymmetry in ‘inference towards the antonym’ (ITA): the tendency for negated positive adjectives (e.g. ‘not high’) to more strongly imply their antonym (‘low’) than negated negative adjectives (e.g. ‘not low’) imply their opposite (‘high’). Most prior research has operationalized polarity using a combination of evaluative valence (desirable vs. undesirable) and dimensional mapping (measuring a higher vs. lower quantity on the associated measurement scale), with these two criteria consistently aligned such that adjectives that are dimensionally positive are also evaluated more positively and dimensionally negative adjectives more negatively. In an online study, we disentangled these two operationalizations of polarity by using comparison classes that created incongruence between evaluative and dimensional polarity (e.g. ‘salary’ vs. ‘CO2-emission’ for the adjectival pair ‘high-low’: for ‘salary’, ‘high’ is both evaluatively and dimensionally positive, where for ‘CO2-emission’ it is dimensionally positive but evaluatively negative). Our results indicate that the polarity asymmetry in ITA only reliably presents when evaluative and dimensional polarity align. We discuss these findings in the context of the complexity of negative adjectives, face-threat, politeness, and the Pollyanna hypothesis.
Source: Journal of Semantics
DOI: 10.1093/jos/ffag006
Institutions: Erasmus University Rotterdam, KU Leuven
Language: English
Subfield: Language and Linguistics
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: hybrid
Estimated APC paid: $5,137
OpenAlex ID: W7204192065