Bai Ruixue

Topic comment construction has been one of the most heated issues in the study of Chinese. This article, focusing on topics in Chinese, begins with the definition of topic. There are three major views about the relation between topic and subject in Chinese.

The traditional view takes subject-predicate as the fundamental relationship between VP and pre-verbal NP(s) and topic is not part of the sentence structure.

The second approach claims that information structures rather than syntactic structures are utilized in Chinese to convey information and the only categories grammaticalized in Chinese are topic and focus.

The third argument, the most popular one, holds that both topic and subject exist in Chinese as separate categories. However, there is no consensus on the definition of topic. A common practice is to determine whether a given pre-verbal NP is a topic or a subject according to their properties. Li & Thompson and Tsao, for instance, have listed a series of properties of topic. However, it turns out unreasonable to use properties as the definition for topic, because many properties are so general that other elements in a Chinese sentence, 昨天 in 小张我昨天见过, for instance, could have them as well. An alternative approach tries to establish a structural relation between topic and verb, but fails to account for the semantic and discourse properties. Different from all this, the author, incorporating both syntactic relation and discourse factors, defines a topic as an unmarked NP that precedes a clause and is related to a position inside the clause; it represents an entity which has been mentioned in the previous discourse and is being discussed again in the current sentence.

The advantage of the discourse consideration in this definition lies in three aspects. First, it accounts for the problem of genericity. Usually topics are claimed to be either definite or generic. The problem is that it is commonly assumed that only those genetic NPs about which the speaker and hearer share some knowledge can function as the topic. Thus it becomes an issue whether the NP in question could be considered generic. The same condition can be reformulated under the definition here to avoid the problem. If an NP represents a class of entities, it can function as a topic only when the class of entities has been mentioned in the previous discourse. For example, 狗in 狗我见过 can be interpreted as either genetic or definite with different previous discourses. Therefore the genericity becomes irrelevant to its topichood. Secondly, it has often been observed that the reason for choosing topic comment constructions is that Chinese speakers tend to first spell out the main thing they want to talk about. However, not everything a person wants to talk about, even if it is the main thing, can be presented as a topic. *一件事我想告诉妈妈 is an example in point, in which 一件事, as the main thing, cannot be spelt out first as the topic because it has not been mentioned previously and the NP remains indefinite. Thirdly, it accounts for the status of sentence initial phrases of time and location. In some analyses, all these phrases are treated as topics. Problems arise, however, when some temporal and locative phrases display unusually properties. For instance, we cannot say *一个寒冷的冬天早上啊，农夫在路上看见了一条冻僵的蛇。A related issue is the linear order of sentence initial temporal and locative phrases. It is observed that the temporal phrase precedes the locative when both are indefinite, while the locative can occur in the initial position when it has been mentioned in the previous discourse. These facts are hard to explain if all sentence initial temporal and locative phrases are considered topics, but they are the expected patterns if only the phrase related to the previous discourse is treated as the topic.

[1] [2] [3] [4] [5] 下一页  

In the second major part, the author deals with the dangling topics, or, the so-called Chinese style topics. A fundamental assumption underlying the definition for topics in this essay is that comment is an open clause containing an empty category or a resumptive pronoun which is co-referential with the topic. However, under the Chinese style topic claim, it is argued that topics in Chinese are not related to any structural position inside the comment. The relation between these so-called Chinese style topics and comments is based on “aboutness”. However, the author refutes first that the notion of aboutness has not been clearly defined syntactically and thus is too vague to explain the relation. Secondly, certain topic comment constructions are not acceptable even if the comment does say something about the topic. For instance, *这件大事我知道张校长辞职了is not well-formed although the initial NP is about 张校长辞职了. Thirdly, this claims lacks solid empirical evidence. Five types of dangling topics have been identified in the literature, and in this part of his essay, the author examines them one by one.

The first type is exemplified by 他们我看你，你看我and 他们大鱼吃小鱼 which are associated with a fixed saying. Supporters of the Chinese style topics argue that the comments, namely, 我看你，你看我and大鱼吃小鱼, are complete sentences, therefore no position is left for the initial NPs. However, the author retorts that我看你，你看我and大鱼吃小鱼do not have the interpretation of a sentence. Actually they should be treated as idiom chunks which function as VPs meaning “look at each other” and “act according to the jungle law” respectively. This is because 我and你do not refer to any antecedent NP, and the lexical items and their order are fixed. Further evidence comes from the sentence他们专门大鱼吃小鱼, in which the adverb专门can only occur between the subject and the VP. So in those sentences, there is indeed a position, namely, subject, left for the NP.

The second type is illustrated by 他们谁都不来. This one falls into the category of the so-called “double nominative sentence”, in which the two sentence initial NPs can be both understood as agent, and a possessive 的can be inserted between them which indicates their structural relation. Li & Thompson claims that since*他们的谁都不来is ill-formed, the initial NP is not related to any position inside the comment. In this connection, the author argues that the WH-word 谁 does not have the status of subject, because, for instance, 他们会谁都不来 is well-formed, in which the word会cannot precede subjects. Actually, the WH-word here behaves more like an universal quantifier which can be replaced by quantifiers like 大家or全, but not by referential NPs. In this sense, Li & Thompson’s claim is not justified.

上一页  [1] [2] [3] [4] [5] 下一页  

The third types the authors reviews is exemplified by 那所房子幸亏去年没下雪and那场火幸亏消防队员来得快, in which the initial NP seem unrelated to the comment. On this issue, the authors explains that words like 幸亏and可惜are known as connective adverbs which work in pairs. For instance, 幸亏is part of the pair 幸亏…不然 or 幸亏…才. The second adverb can sometimes be omitted together with the clause it ushers in, but the missing part is recoverable. From the recovered form 那所房子幸亏去年没下雪，不然早就塌了, we can see the NP 那所房子is structurally related to the latter part of the comment. This point is illustrated by the ill-formedness of the sentence if we take 幸亏away to get*那所房子去年没下雪.

The fourth type of the so-called Chinese style topics is characterized by 这件事情你不能光麻烦一个人. Here 你不能光麻烦一个人appears to be a complete sentence, in which there is no corresponding position to the sentence initial NP. The author argues that arguments are not the only positions for an NP to occupy; it can also function as adjunct. In many cases, the preposition in Chinese can be omitted from an adverbial PP. The above sentence actually has the same interpretation as你为这件事情不能光麻烦一个人 or 为这件事情你不能光麻烦一个人, in which the initial NP is not dangling.

The last type that has been identified can be illustrated by 这种豆子一斤30块钱, which consists of two or more NPs but has no verb to head the predicate. The author analyzes this kind of sentences as strongly context-dependent as for their interpretation. One way to interpret is to establish a complete sentence like那种豆子我一斤30块钱卖给你, in which the initial NP is obviously related to the verb. Another way is to recover the identity of the null forms from the context, so that the hearer can fill in all the missing parts to have the correct interpretation. In all cases, if the topic is analyzed as unrelated to any position in comment, there will be no predication relation between topic and comment, and there will be no interpretation for the whole sentence at all. In this section the author concludes that the so-called Chinese style topics and the so-called English style ones are essentially the same.

[1] [2] 下一页