Structural priming offers a window on how L2 speakers retrieve and assemble grammatical knowledge, yet evidence from Mandarin existential templates remains limited. We report two production experiments with L2 learners. In Experiment 1, participants completed a picture description task targeting three existential patterns: EX-ZHE (Location +V + + NP), EX-LE (Location + V + + NP), and EX-YOU (Location + + NP). A binomial mixed-effects model of existential realization (existential vs. non-existential output) showed reliable priming, strong prime-type differences, and a Prime Type × Proficiency interaction: EX-ZHE and EX-YOU primes increased existential production, whereas EX-LE reduced it; proficiency increased existential realization overall but most clearly following EX-ZHE primes. Experiment 2 fixed EX-ZHE as the target and manipulated prime-target lexical overlap (verb, sentence-initial locative, post-verbal noun, or no overlap). Lexical boost was locus-dependent: verb overlap produced the largest facilitation, locative overlap also increased EX-ZHE realization, and post-verbal noun overlap did not reliably differ from baseline; proficiency did not modulate these effects. Together, the findings support a cost and cue-sensitive account in which constructional availability and diagnostic lexical retrieval cues jointly shape L2 alignment. Any instructional implications are treated as hypotheses from controlled tasks and require corpus- and classroom-based validation.
Qian et al. (Thu,) studied this question.