Abstract

Thick terms like “courageous,” “smart,” and “tasty” combine description and evaluation, contrasting with purely evaluative terms like “good” and “bad,” and descriptive terms like “Italian” and “green.” Thick terms intuitively constitute a special class of evaluative language; but we currently do not know whether the psycholinguistic effects of these terms are reducible to known semantic dimensions. Here, we start to systematically explore this question by comparing the behavior of thick terms and non‐thick descriptive terms with similar affective valence, which is a strong candidate semantic dimension to account for differences between evaluative and non‐evaluative language. We study thick terms from English, Dutch, and Italian, combining behavioral data from the cancellability task, Cloze task, and free association networks, with natural language processing methods and psycholinguistic ratings of word valence. We find that thick and non‐thick descriptive terms are associated with different psycholinguistic effects, even when carefully matched for valence, suggesting that valence is insufficient to account for the difference between thick and non‐thick terms. Instead, we find no reliable difference between positive and negative thick terms, and between moral, epistemic, and aesthetic thick terms. Our findings indicate that thick terms form a homogeneous class of evaluative language whose psycholingusitic effects cannot be explained only in terms of affective valence.