Answers are grounded in this paper and cite their sources.
Do probabilistic grammars revealed by corpus-based studies capture speakers’ mental representations? The present study addresses this question empirically by examining how closely speakers’ intuitions align with corpus-based probabilistic patterns. Special attention is given to cross-varietal effects of semantic prosody in Chinese long passive constructions marked by bèi , gěi and ràng. A ranking task experiment was conducted with 115 participants from Mainland China and Taiwan, and the data were analyzed using Bayesian multivariate cumulative regression. Results revealed three main findings. First, participants’ judgments were primarily driven by relative frequency ( bèi > ràng > gěi ). Second, the cross-varietal interaction was validated only for bèi , suggesting stronger entrenchment of probabilistic patterns for prototypical variants. Third, significant effects of variety and age emerged, reflecting transferable preferences and sociolinguistic layering in Mandarin passive usage. We conclude that frequency information, prototypicality and sociolectal conditioning jointly shape passive variation in Mandarin Chinese – factors that corpus-based probabilistic grammars alone may not fully capture.
No takes yet. Share an insight, caveat, or question.
Sumin Guan (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: