Testing ‘constructionhood’ and the cognitive plausibility of mid-level abstractions in English possessive constructions

Abstract

This study investigates the processing relevance of mid-level constructions, partially schematic form–meaning pairings with semantically constrained slots, in English possessives. Drawing on usage-based Construction Grammar, we ask whether patterns such as [ my N _kinship ] or [ her N _bodypart ] give rise to processing effects not fully explained by semantic category membership or broader schemas like [POSS N] or [ my / her N]. We introduce mid-level chunk priming as a real-time diagnostic of constructional influence. In two primed lexical decision experiments (Stimulus Onset Asynchrony [SOA] = 200/300 ms, N  = 646), primes varied by frame (bare noun, my -N, her -N) and semantic field (kinship, body part, furniture, abstract quality), followed by related or unrelated targets. We tested whether primes instantiating a hypothesized mid-level construction (e.g., my niece ) facilitated recognition of targets from the same construction-internal category (e.g., aunt ) beyond semantic priming or a general expectation for an upcoming noun. At 300 ms, [ my N _kinship ] and [ her N _bodypart ] yielded selective facilitation; no comparable effects appeared for furniture or abstract quality nouns, nor at 200 ms. These findings suggest that such mid-level patterns differ in processing relevance and that their effects take time to unfold during prime–target processing, consistent with a role for certain construction-specific generalizations in real-time use.