Are there segmental and tonal effects on syntactic encoding? Evidence from structural priming in Mandarin

Numerous studies have established that speakers tend to form utterances by reusing previously experienced sentence structures (i.e., structural priming). It was also frequently found that the repetition of lexical items enhances structural priming (i.e., lexical boost). This facilitation effect occurs not only when there is a full overlap of verbs, but also when one level of the lexical representation (semantic or phonological representation) overlaps between the prime and the target. In the current study, we further scrutinize the phonological overlap effect on structural priming. We asked whether the phonological effect is independent of orthographic overlap, and whether it is driven by overlap of segments, tone, or both. In five structural priming experiments (three labbased, two web-based experiments), native Mandarin speakers were instructed to describe transitive pictures after receiving SVO or SOV "ba" prime sentences. In Experiment 1, prime and target verbs had lexical overlap (e. g., (sic)ttuo1, to take off]-(sic)ttuo1]), semantic overlap (e.g., (sic)txie4, to remove]-(sic)ttuo1]), phonological overlap (e. g., (sic)[ttuo1, to mop]-(sic)ttuo1]), or no overlap (e.g., (sic)[tda3, to beat]-(sic)ttuo1]) while similarities at other levels were carefully avoided. There were structural priming and lexical boost effects, but semantic or phonological overlap did not boost priming. In two further lab-based experiments and their large-scale online replications, verbs in prime and target had full phonological overlap (segmental + tonal, e.g., (sic)[ttuo1]-(sic)ttuo1]), syllabic overlap only (e.g., ttuo2, to carry]-(sic)ttuo1]), tonal overlap only ((sic)tcheng1, to weigh]-(sic)ttuo1]), or no overlap. All four experiments showed structural priming, which was boosted by full phonological overlap. The syllabic overlap exerted a significant facilitation effect on structural priming, whereas no tonal effect was found. Together, these results indicate that processing at the phonological level feeds back to syntactic encoding in sentence production, which further supports an interactive view of language production.