Lecture | CHiLL series
How Syntactic Structure and Classifier Congruency Shape Mandarin Sentence Production: Behavioural and ERP Insights
- Date
- Wednesday 13 May 2026
- Time
- Serie
- Chinese Linguistics in Leiden (ChiLL)
- Address
-
Lipsius
Cleveringaplaats 1
2311 BD Leiden - Room
- 2.24
Abstract
How much of a sentence do speakers plan before they begin to speak, and what determines how that planning is organised? In this talk, I address this question through a picture-description study in Mandarin Chinese, which serves as a testing ground for examining how syntactic structure building and lemma-level classifier information differentially contribute to grammatical encoding during sentence planning. Syntactic structure was manipulated by varying whether the target locative sentence began with a single noun or a conjoined noun phrase, while classifier congruency was used to probe whether lemma-level information can influence sentence planning even when the classifier itself is not overtly produced.
Thirty native Mandarin speakers described displays of three objects using locative sentences such as "A 在 B 和 C 上边/下边" (A zài B hé C shàngbian/xiàbian, "A is above/below B and C") or "A 和 B 在 C 上边/下边" (A hé B zài C shàngbian/xiàbian, "A and B are above/below C"). The first two nouns in each sentence either shared or did not share the same potential dominant classifier, although no classifier was required in the spoken response. This design allowed us to examine whether pre-articulatory planning in Mandarin is shaped primarily by syntactic structure, by non-overt classifier-related information, or by both.
The results reveal a striking pattern. Syntactic structure significantly affected pre-articulatory planning: sentences beginning with a conjoined noun phrase were initiated faster than those beginning with a single noun, and structure-related differences also emerged in posterior ERP activity during the planning window. Notably, this latency pattern diverges from much of the previous sentence-production literature — based largely on English and other Indo-European languages — in which larger sentence-initial constituents are typically associated with slower speech onset. Classifier congruency, by contrast, did not yield robust effects. I will discuss what these findings suggest about the mechanism of grammatical encoding in Mandarin sentence production, and how evidence from Mandarin can inform broader theories of how speakers coordinate lexical and structural information prior to articulation.