Over the last quarter-century, there has been a surge of research on Chinese syntax. A cursory look at the programs of Chinese linguistics conferences held since 1985 shows that at least a full continuous session has been devoted to Chinese syntax throughout each day of every conference. Those who were involved in organizing such conferences can also recall the large number of syntax abstracts, routinely accounting for fifty to sixty percent of all abstracts received for review. It is also during this past quarter-century that a significant number of theoretically oriented works on Chinese syntax began to appear in major refereed academic journals published in the West. Several monographic, theoretical treatments of Chinese syntax have also appeared that distinguished themselves from earlier general descriptions or reference grammars. In the field of theoretical linguistics, more works than before make crucial reference to Chinese syntax. It is clear that research on Chinese syntax that is informed by modern linguistic theories has been productive. In turn, it is also clear that the study of Chinese syntax has played an ever-increasing role in linguists' construction of modern “mainstream” syntactic theories.

Most of these “modern syntactic theories” are in one form or another theories falling under the formal paradigm of generative grammar. Of these formal treatments, much research has been carried out in the Principles-and-Parameters (P&P) approach initiated by Chomsky and his colleagues and students around 1980, plus and minus two or three years, in its various incarnations including the so-called Government-and-Binding (GB) framework, the Barriers framework, and recent attempts at theoretical economy aimed at the ideals of the Minimalist-Program (MP).

Each category affects the grammaticality of a sentence differently. For verbs, the most conspicuous property is transitivity, which we investigate in this chapter. Following the convention in theoretical syntax, the subject and object(s) of a verb are called its arguments, and the semantic relation between a verb and any of its arguments called a thematic relation. The first section introduces the basic properties of thematic relations and demonstrates how they can help explain certain linguistic phenomena. A few recent attempts to understand the nature of thematic relations are critically reviewed in Section 2.2. An alternative theory is proposed in Section 2.3.

Arguments and theta-roles

It is obvious to any linguistically minded observer that in a typical active sentence built around a transitive action verb, such as ta chang minge ‘she sing folk.song’ or ni xie shi ‘you write poem,’ the subject argument is always the one initiating and performing the action while the object argument is always what is acted upon. This simple fact suggests the possibility that not every detail in the thematic relation between an argument and a verb matters in syntactic computation. For instance, ta ‘she’ and ni ‘you’ are subjects only because these NPs represent the “doers” of the actions; whether an action is done through singing or writing has no effect on qualifying an NP as the subject argument.

We take it as our starting point that a Chinese sentence is composed of words and that words have different behaviors in a sentence. For instance, while dayan fei ‘wild.goose fly’ is an acceptable sentence, *fei dayan ‘fly wild.goose’ is not. The most obvious reason for the contrast is that dayan ‘wild.goose’ is a noun that canonically serves as the subject of the sentence and fei ‘fly’ is a verb whose canonical function is to be the predicate occurring after the subject. This means that in order to understand the syntax of Chinese, or the syntax of any language for that matter, we minimally need to understand how the words in a language are classified and how these different classes of words are put together to form sentences. In this book, word classes are referred to as lexical categories, or just categories for short, following the terminological convention of generative syntax.

While the basic distinction between nouns and verbs is universally recognized in modern literature on Chinese syntax, scholars differ, sometimes drastically, on other categories. See Chao (1968), Li and Thompson (1981), Zhu (1982), and Xing and Ma (1992) for a few examples. The differences in opinion arise partly because linguists with different theoretical backgrounds may employ different criteria for word classification, and partly because we still lack sufficient knowledge about certain words and their properties.