Phases and a “finite”/“non-finite” distinction in Mandarin Chinese
Nick Huang

The absence of tense marking in Mandarin Chinese has led some linguists, eg. Hu et al. (2001), to claim that Chinese has no syntactic distinction similar to the finite/non-finite distinction observed in languages with overt tense. I present counter-evidence from Mandarin Chinese clausal complements to show that a two-way distinction exists. Adopting a Minimalist framework, I argue that, like finite and non-finite constructions, Chinese clausal complements are either strong phases or they are not. The clauses studied are the complements to what are termed “believe”-type and “persuade”-type verbs (Huang (1982)) in Mandarin Chinese. Specifically, I show that there are differences (1) in the behavior of subjects and objects of the clausal complements of “believe”-type verbs (henceforth “b- clauses”) and those of “persuade”-type verbs (“p-clauses”). This implies that there is a systematic syntactic difference between b-clauses and p-clauses.

(1) Given a sentence SM VM [ X P SE VE O], where XP is the clausal complement  of VM

Because these syntactic differences are based on subject and/or object DPs that are found in every clause with a transitive predicate, they can be used to reliably identify whether a clausal complement is a b-clause or p-clause, or whether a verb is a “believe”-type or “persuade”-type. They are thus an improvement over existing diagnostics that rely on the presence of modals, or the semantics of the matrix verbs (see Hu et al. (2001) for a summary).

More importantly, these three differences shed light on the syntax of clauses in Chinese. First, (1a) shows that the binding properties of b-clauses and p-clauses differ. I then turn to examples of raising in English (2) to explain the patterns observed in (1b) and (1c).

(2)(a) It seems [that linguists like languages].
     (b) *Linguists seem [that t like languages].
     (c) Linguists seem [t to like languages].

I observe similarities between object preposing (1b) and the distribution of the distributive particle dou (1c), and the interaction of raising verbs and the subjects of embedded clauses (2). I propose that dou in (1c) can be seen as a kind of number agreement with distributive properties, rather than an adverb, as previously proposed (cf. Cheng (1995)). Adapting arguments on the raising of embedded DPs and the Phase Impenetrability Condition (Chomsky (1999)), I show that the patterns in (1b) and (1c) are the result of b-clauses being strong phases, like the finite clausal complement in (2a), while
p-clauses are not strong phases, patterning after the infinitive clausal complement in (2c). Lastly, I also find evidence that dou adjoins to a modal-like future particle jiang “will” (6). Together, the data suggest that 1. there is a two-way distinction in Chinese clauses that parallels the finite/non- finite distinction, and 2. the differences in Chinese clauses might also be related to time and tense.


(a) Zhangsani xiangxin [ta∗i / j yuanliang-le Lisi-de cuowu].
     Zhangsan believe 3 SG forgive-P ERF Lisi’s mistake
    “Zhangsani believes he∗i / j has forgiven Lisi’s mistake.” (b-clause)

(b) Zhangsani bu-neng shuofu [zijii /*tai yuanliang Lisi-de cuowu].
     Zhangsan N EG-can convince self/*3 SG forgive Lisi’s mistake
    “Zhangsan cannot convince himself to forgive Lisi’s mistake.” (p-clause)

Preposing the object (in curly brackets) of an embedded clause (examples adapted from Ernst and Wang (1995)):             (4)
(a) Wangwu shuo [Lisi du-wan-le {na-ben xiaoshuo}].
     Wangwu say [Lisi read-finish-P ERF that-C L novel]
    “Wangwu said that Lisi finished reading that novel.” (b-clause)

(b) *Wangwu {na-ben xiaoshuo} shuo [Lisi du-wan-le t].

(c) Wangwu bi [Lisi du-wan-le {na-ben xiaoshuo}].
     Wangwu force [Lisi read-finish-P ERF {that-C L novel}]
    “Wangwu forced Lisi to finish reading that novel.” (p-clause)

(d) Wangwu {na-ben xiaoshuo} bi [Lisi du-wan-le t].

Topicalization of the object (in curly brackets) of an embedded clause and the distribution of dou “all”

(a) Wo xiangxin [Zhangsan du-wan-le {zheli-de mei-ben shu}].
      I believe Zhangsan read-P ERF here-M OD every-C L book
      “I believe Zhangsan had read every book here.” (b-clause)

(b) *?{Zheli-de mei-ben shu}, wo dou xiangxin [Zhangsan du-wan-le t]. 

(c) Wo quan [Lisi du-wan {zheli-de mei-ben shu}].
      I advise Lisi read-finish here-M OD every-C L book
     “I advised Lisi to finish reading every book here.” (p-clause)
(d) {Zheli-de mei-ben shu}, wo dou quan [Lisi du-wan t]. 

Dou “all” and future marker jiang
(6) Zheli-de mei-ben shu, Lisi dou jiang hui zai mingtian yiqian kan-wan.
     here-M OD every-C L book, Lisi all F UT will at tomorrow before read-finish
    “Lisi will finish reading every book here by tomorrow.”

Cheng, L. L.-S., 1995. On dou-quantification. Journal of East Asian Linguistics 4, 197–234.
Chomsky, N., 1999. Derivation by Phase. MIT Occasional Papers in Linguistics 18.
Ernst, T., Wang, C., 1995. Object preposing in Mandarin Chinese. Journal of East Asian Linguistics 4, 235–260.
Hu, J., Pan, H., Xu, L., 2001. Is there a finite vs. nonfinite distinction in Chinese? Linguistics 39, 1117–1148.
Huang, C.-T. J., 1982. Logical Relations in Chinese and the Theory of Grammar. MIT dissertation.


Make a Free Website with Yola.