Are There Dangling Topics in Mandarin Chinese?
*Ray Rui-heng Huang
National Taiwan Normal University
Jen Ting
National Taiwan Normal University
This paper compares two opposing views as to whether dangling topics exist in Mandarin Chinese. The non-dangling-topic analysis considers all Chinese topics syntactically licensed by chain formation or A’-movement, while the dangling-topic analysis treats some Chinese topics (i.e. dangling topics) as semantically rather than syntactically licensed. Rejecting the semantic licensing of topics, this paper argues in favor of the non-dangling-topic approach but meanwhile fixes some problems that may arise under this approach.
Key words: topic construction, dangling topic, subject in Chinese, PP-reduced form
1. Introduction
Linguists working on topic-comment constructions in Mandarin Chinese have generally identified two kinds of grammatical topics: dangling vs. non-dangling. Non-dangling topics are structurally characterized in the sense that we can always find structural dependence between the topic NP and its comment clause. Consider (1) and (2) (cited from Shi 2000).
(1) GaoQiangi na, ZhouHua weile tai mei lai zheng shengqi ne. GaoQiang PART ZhouHua because he not come just mad PART
‘As for GaoQiang, ZhouHua is being mad because he did not come.’
(2) Zhexie shii wo juede ta shuo ∅i bu heshi. these things I think he say not proper
‘These things, I think it’s not proper for him to say.’
The topic is coreferential with a resumptive pronoun in (1) and co-indexed with a null form in (2). In both cases, there is a structural position inside the comment clause co-indexed with the topic. This structural dependence relationship between the topic and the comment clause can be explained by seeking a parallelism with predication
relation as proposed by Williams (1980). Shi (2000) thus claims that Chinese topics
*
We would like to express our gratitude to the two anonymous Concentric reviewers for providing us with constructive comments. Mistakes are exclusively our own.
must be related to a syntactic position inside the comment clause, denying the existence of dangling topics in Mandarin Chinese.
While rejected by Shi, dangling topics are posited by many linguists in terms of an
aboutness relation (Chao 1968, Li and Thompson 1976. cf. Huang 1982, Tang 1990,
Qu 1994). Consider (3).
(3) Na-hui da-huo, xingkui xiaofang-dui lai-de-zao. that-CL big-fire fortunately fire-brigade come-DE-early
‘As for that big fire, fortunately the fire brigade came early; (otherwise)…’
On the surface, there appears to be no position available inside the comment clause for the sentence-initial NP nahui dahuo ‘that big fire’. Semantically, however, such gapless topic construction is claimed to be interpreted according to the aboutness relation: the comment clause says something about the topic, e.g. by Huang (1994:162). Dangling topics like nahui dahuo ‘that big fire’ in (3) are specifically termed Chinese-style topics in contrast to English-style topics (Chafe 1976, Li and Thompson 1976, Xu and Langendoen 1985, Huang 1994); they are not selected or subcategorized by the verb of the comment clause and are unrelated to any structural position inside that clause (Li and Thompson 1981, LaPolla 1990, Tsao 1990, Ning 1993).
The aboutness account, however, is criticized by Shi as being vague and not clearly defined in a syntactic fashion. In particular, “certain topic-comment constructions are not acceptable even when the comment does say something about the topic” (Shi 2000:389). Consider the following pair from Shi.
(4) *Zhe-jian da-shi wo zhidao Zhang-Xiaozhang cizhi le. this-CL big-issue I know Zhang-Principal resign ASP ‘As for this big issue, I know that Principal Zhang has resigned.’
(5) Zhe-jian da-shi jiu shi Zhang-Xiaozhang cizhi le. this-CL big-issue exactly be Zhang-Principal resign ASP ‘This big issue is that Principal Zhang has resigned.’
Shi argues that if the topic NP zhejian dashi ‘this big issue’ in (4) equates with the principal’s resignation as shown in (5), (4) should be a well-formed topic-comment construction licensed by aboutness, since the comment clause says something about the topic. But this prediction is not borne out, suggesting that the aboutness account is not on the right track.
As seen above, Shi (2000) adopts the structural approach exclusively to analyzing topic-comment constructions in Mandarin Chinese. He lists six types of so-called dangling topics that have been identified in the literature and argues against their topichood. That is, in Shi’s view, no dangling topics exist in Mandarin Chinese. Contrary to Shi’s non-dangling-topic analysis, Pan and Hu (2002:2) propose that “topics in Chinese can be licensed not only by a syntactic gap or resumptive pronoun, but also by a semantic variable which does not have a corresponding syntactic position”. In their view, dangling topics do exist in Mandarin Chinese and they are all licensed semantically. Given the two opposing views, in this paper we will argue in favor of Shi’s analysis that so-called dangling topics are actually subjects, NP topics, NP adverbials, or PP-reduced forms. In what follows, Section 2 compares the dangling-topic and non-dangling-topic analysis and Section 3 concludes this paper.
2. Dangling-topic vs. non-dangling-topic analysis
In this section, we will examine each type of dangling topic discussed by Shi (2000) and by Pan and Hu (2002) and evaluate accounts from both sides. For a comparative preview of dangling-topic and non-dangling-topic analysis, consider the following table.
Table 1. Dangling-topic vs. non-dangling-topic analysis of six types of initial NPs
Six Types Non-dangling-topic
Analysis by Shi (2000)
Dangling-topic Analysis by Pan and Hu (2002)
1 [Tamen] da-yu chi xiao-yu. Subject Dangling topic
2 [Tamen] shei dou bu lai. Subject Dangling topic
[Na-chang huo] xingkui xiaofang-dui
lai-de-kuai. Topic Dangling topic
3
[Na-chang huo] xiaofang-dui lai-de-kuai. Sentential adverbial Dangling topic [Zhe-jian shiqing] ni bu neng guang mafan
yi-ge ren. PP-reduced form Dangling topic
[Shengwu-lunlixue] wo shi men-wai-han. PP-reduced form Dangling topic 4
[Xihongshi] wo chao le ji-dan. Not mentioned Dangling topic
5 [Na-zhong douzi] yi-jin san-shi-kuai qian. Subject Dangling topic
6 [Wu-jia] Niuyue zui gui. Subject Dangling topic
All the six types of sentence-initial NPs are treated as dangling topics by Pan and Hu, but as subjects, (non-dangling) topics, sentential adverbials, or PP-reduced forms by Shi. In the following discussion, we will examine which analysis attains more explanatory adequacy for empirical facts.
2.1 Fourth type1
Following Lu’s (1986) preposition-dropping analysis, Shi argues that some of the so-called dangling topics in Mandarin Chinese are actually generated by dropping the preposition (sometimes plus a locative particle such as fangmian ‘side’, zhizhong ‘midst’, etc.) from adjunct adverbial PPs without altering the sentence meaning. Below are Shi’s examples.
(6) a. Wei zhe-jian shiqing ni bu neng guang mafan yi-ge ren. for this-CL matter you not can only bother one-CL person ‘For this matter, you cannot just bother one person.’
b. Zhe-jian shiqing ni bu neng guang mafan yi-ge ren. this-CL matter you not can only bother one-CL person ‘This matter (topic), you can’t just bother one person.’
(7) a. Zai shengwu-lunlixue fangmian wo shi men-wai-han.
at biology-ethics side I be door-out-man
‘With regard to bioethics, I am a layman.’
b. Shengwu-lunlixue fangmian wo shi men-wai-han.
biology-ethics side I be door-out-man
‘With regard to bioethics, I am a layman.’ c. Shengwu-lunlixue wo shi men-wai-han.
biology-ethics I be door-out-man
‘With regard to bioethics, I am a layman.’
Shi claims that in both (6b) and (7c) the sentence-initial NP is not related to any structural position in the following clause, and to account for these two sentences, he adopts the preposition-dropping analysis, thus ruling out the dangling-topic analysis.
Pan and Hu object to Shi’s preposition-dropping analysis by arguing that it fails to explain why the preposition wei ‘for’ in (8) cannot be omitted, given the otherwise ill-formedness of (9).
(8) Wei zhe-jian shiqing Zhangsan dajia le. for this-CL matter Zhangsan fight ASP ‘For this matter, Zhangsan fought.’
1
The fourth type is discussed first because the preposition-dropping analysis introduced in this subsection will partially form the basis of our analysis of the remaining types.
(9) *Zhe-jian shiqing Zhangsan dajia le. this-CL matter Zhangsan fight ASP ‘For this matter, Zhangsan fought.’
In Pan and Hu’s view, the asymmetry between (6b) and (9) lies in the difference in subcategorization. They argue that the sentence-initial NP zhejian shiqing ‘this matter’ in (6b), though not mapped to an argument position in syntax,2 is subcategorized by the predicate mafan ‘bother’ in the thematic structure. That is, they take this subcategorized NP as a dangling topic and suggest that it is semantically licensed by a syntactically unrealized theta-role of Instrument or Material.3 In contrast, the NP zhejian shiqing ‘this matter’ in (9) is not subcategorized by the one-place predicate dajia ‘fight’ and thus cannot be semantically licensed as a dangling topic. Pan and Hu (2002:10) additionally point out that “an NP bearing the Instrument theta-role can appear either before or after the subject”, as shown below.
(10) Xihongshi wo chao le ji-dan. tomato I fry ASP chicken-egg ‘I fried eggs with tomatoes.’
(11) Wo xihongshi chao le ji-dan. I tomato fry ASP chicken-egg ‘I fried eggs with tomatoes.’
Under their analysis, the NP xihongshi ‘tomato’ in both sentences is assigned the Instrument theta-role. The above claims provided by Pan and Hu concerning dangling topics and the Instrument theta-role will be refuted immediately below.
We will argue that (6b), (7c), and (10) have different derivations and cannot be analyzed in the same way. Shi’s account based on preposition-dropping will be suggested to hold for the subtype represented by sentences like (7c) but not for (6b) and (10). The latter two cases, however, can be explained by movement in Shi’s system. We shall start with the derivation of the subtype like (6b). This subtype arguably involves a double-object construction, in light of sentences like (12).4
2
This NP will be argued to be syntactically licensed by being treated as the direct object of the predicate mafan ‘bother’.
3
Since Pan and Hu do not distinguish between Instrument and Material in their paper, we treat them as the same. In this paper, we adopt the former term, i.e. Instrument.
4
(12) Wo xiang mafan ni yi-jian shi. I want bother you one-CL matter ‘I want to bother you with something.’
The fact that the verb mafan ‘bother’ in the above example is a three-place predicate suggests that the topic NP zhejian shiqing ‘this matter’ in (6b) can similarly be analyzed as moving from the direct-object position subcategorized by the verb and therefore is not a dangling topic.
Likewise, the subtype like (10) can be argued to involve movement as well and thus does not contain a dangling topic. We propose that the initial NP at issue is syntactically licensed by being the complement of a complex predicate, e.g. chao
jidan ‘fry eggs’ in (10). This is how a so-called retained object is licensed as first
proposed by Thompson (1973) and followed by Huang (1982) and many others. Another example given by Huang (1982) concerns (13a), whose underlying structure is represented in (13b).
(13) a. Ta [ba zhi-men] ti le yi-ge dong. he [BA paper-door] kick ASP one-CL hole ‘He kicked a hole in the paper-door.’
b. S NP VP V’ NP ta ‘he’ V NP zhi-men ‘paper-door’ ti yi-ge dong
‘kick’ ‘a hole’
According to Huang, the verb ti ‘kick’ takes the inner object yige dong ‘a hole’ as its complement; the verb-object combination in turn forms a complex predicate and takes the outer object zhimen ‘paper-door’ as a complement. In terms of (10), however,
the outer object xihongshi ‘tomato’ has undergone topicalization rather than the
ba-derivation as in (13).5 Since the initial NP xihongshi ‘tomato’ in (10) is moved
5
from the following clause, it is thus ruled out as a dangling topic. To recap, the initial NP in both (6b) and (10) is proposed to be derived by movement rather than being generated as a (possible) PP-reduced form in Shi’s analysis.6
On the other hand, there is no way for the sentence-initial NP zhejian shiqing ‘this matter’ in Pan and Hu’s ill-formed example of (9), repeated below as (14), to be derived by movement because it is inconceivable to treat the NP at issue as a logical object of the one-place predicate dajia ‘fight’; the unacceptability thus follows.
(14) *Zhe-jian shiqing Zhangsan dajia le.
this-CL matter Zhangsan fight ASP
‘For this matter, Zhangsan fought.’
In this way, Shi’s non-dangling-topic approach to the subtypes of the fourth type, represented by sentences like (6b) and by sentences like (10), can be maintained despite the failure of his preposition-dropping analysis of these subtypes.
We shall now turn to another subtype as represented by sentences like (7c), repeated below as (15).
(15) Shengwu-lunlixue wo shi men-wai-han. biology-ethics I be door-out-man ‘With regard to bioethics, I am a layman.’
To account for such sentences, Pan and Hu’s analysis would have to assume the sentence-initial NP shengwu-lunlixue ‘bioethics’ to be semantically subcategorized by the copular verb shi ‘be’, which is contrary to fact. In fact, their Instrument account is also problematic for the initial NP shengwu-lunlixue ‘bioethics’ in (15), because in the thematic structure, the NP is in no sense an instrument and thus cannot be semantically licensed as a dangling topic as claimed by Pan and Hu.
We are in favor of Shi’s preposition-dropping analysis of the subtype like (7c). The reduction process seems fairly productive in Mandarin Chinese and does not
subtype like (6b) and the subtype like (10). That is, as illustrated by the contrast between (i) and (ii), only affected objects like xihongshi ‘tomato’ in (10) and zhimen ‘paper-door’ in (13) can be marked by the co-verb ba; the NP zhejian shiqing ‘this matter’ in the construction involving mafan ‘bother’ as in (6b), however, cannot.
(i) Wo ba xihongshi chao le ji-dan.
I BA tomato fry ASP chicken-egg
‘I fried eggs with tomatoes.’
(ii) *Ni bu neng ba zhe-jian shiqing guang mafan yi-ge ren. you not can BA this-CL matter only bother one-CL person ‘You cannot just bother one person with this matter.’
6
result in a significant change in semantics.7 More examples are given below.
(16) (Zai) jiaoxue (shang), ta congbu toulan. (at) teaching (top) he never lazy ‘In teaching, he is never lazy.’
(17) (Zai) zhe-qun ren (zhizhong), Zhangsan biaoxian zui hao. (at) this-CL person (midst) Zhangsan behave most good ‘Among these people, Zhangsan behaves best.’
Shi’s preposition-dropping analysis thus draws empirical support from this subtype of constructions. Following Shi’s analysis, we conclude that the sentence-initial NP
shengwu-lunlixue ‘bioethics’ in (7c) is not a dangling topic, but a constituent inside a
reduced PP adverbial (hereafter PP-reduced form).
In brief, in this subsection we have argued that the so-called dangling topics in sentences like (6b) and (10) are licensed by movement and that those in sentences like (7c) involve a preposition-dropping NP.
2.2 First type
Dangling topics classified as the first type by Shi are associated with idiomatic expressions.
(18) Tamen da-yu chi xiao-yu. they big-fish eat small-fish
‘They act according to the law of the jungle.’
(19) Tamen, wo kan ni, ni kan wo.
they I look you you look me
‘They look at each other.’
Shi claims that both strings dayu chi xiaoyu ‘to act according to the law of the jungle’ and wo kan ni, ni kan wo ‘I look at you and you look at me’ do not yield a literal but nonliteral reading, and thus should be treated as idiom-chunk predicates rather than as
7
Notice that not all PP adverbials in sentences like (7c) can undergo reduction, though. A good example can be found in (9). In this paper, only the reduction of zai (fangmian/zhizhong) ‘in (the respect)’ is pointed out; we leave other candidates open for discussion. Admittedly, more research is required for this preposition-dropping operation.
comment clauses. To show that syntactically saturated idioms8 like dayu chi xiaoyu ‘to act according to the law of the jungle’ may function as predicates, Shi proposes a test based on the adverb zhuanmen ‘specifically’.
(20) Tamen zhuanmen qifu wo. they specifically bully me ‘They are bullying me exclusively.’
(21) *Zhuanmen tamen qifu wo.
specifically they bully me
‘They are bullying me exclusively.’
(22) Tamen zhuanmen ta-yu chi xiao-yu.
they specifically big-fish eat small-fish
‘They are doing nothing but acting according to the law of the jungle.’
The contrast between (20) and (21) indicates that the adverb zhuanmen ‘specifically’ only occurs between the subject and predicate. In other words, this adverb can be used to test the subject-predicate structure; given the acceptability of (22), the predicatehood of the idiom dayu chi xiaoyu ‘to act according to the law of the jungle’ is thus confirmed.
Shi’s idiom-chunk analysis and the zhuanmen ‘specifically’ test, however, are both rejected by Pan and Hu. They argue that the idiom-chunk analysis runs into problems with the following example.
(23) Tamen, ni zhize wo bu dui, wo baoyuan ni bu hao. they you blame I not right I complain you not good ‘They blame each other and complain about each other.’
Such sentences obviously do not involve idiomatic meanings, contrary to the expectations under Shi’s theory.9 Furthermore, the zhuanmen ‘specifically’ test is questionable with regards to the following example.
8
By ‘saturated’, we mean the filled or realized subject slot, e.g. dayu ‘big fish’ in dayu chi xiaoyu ‘(literal) big fish eat small fish’.
9
In his paper, Shi proposes that an idiom chunk in constructions like (18) and (19) can function as a predicate. Notice that logically, this proposal does not imply or additionally mean that all strings following the initial NP can only be idiom chunks. Here Pan and Hu seem to have misinterpreted Shi’s proposal by showing (23). Even so, (23) can still be accounted for by means of Shi’s preposition-dropping analysis. See the discussion below.
(24) Tamen liang, zhuanmen XiaoWang xian lai, XiaoLi hou dao. they two specifically XiaoWang first come XiaoLi after arrive ‘As for them two, XiaoLi always comes after XiaoWang.’
What follow the adverb zhuanmen ‘specifically’ are two full sentences without any idiomatic reading, but Shi’s analysis would wrongly predict them as predicates. This
zhuanmen ‘specifically’ test is concluded to be unreliable by Pan and Hu. Here we
concur with their argument.
Having abandoned Shi’s idiom-chunk analysis and the zhuanmen ‘specifically’ test, Pan and Hu then propose a semantic account to license dangling topics of this first type. “It is the set-member relation that ties the topic and comment together, and thus licenses the topic” (Pan and Hu 2002:8). That is, the sentence-initial NP in both (18) and (19) is a dangling topic licensed semantically by the set-member relation. Each dangling topic is a set including NPs or pronouns inside the comment clause(s) as its members.
Appealing as this set-member analysis appears, we argue that it does not hold for sentences like (25).
(25) Zhangsan da-yu chi xiao-yu.
Zhangsan big-fish eat small-fish
‘Zhangsan bullies the weaker.’
In this case, the initial NP Zhangsan is not a plural set containing either dayu ‘big fish’ or xiaoyu ‘small fish’ as its member. Thus, Pan and Hu’s set-member account fails to be a unified one.
As elaboration of Shi’s analysis that treats idioms as predicates, we propose that this type of unusual predicate10 should have undergone a certain kind of reanalysis as an un-analyzable structure, viewed as a complex-V node in the hierarchical representation as in (26).
10
We consider that other idiom chunks of this type may include zhi bao bu zhu huo ‘plots or evil deeds will be exposed sooner or later’, jiu zhan que chao ‘to usurp what is another’s’, mang ren muo xiang ‘to draw a conclusion from incomplete data’, Kongrong rang li ‘to give priority to the elders’, etc.
(26) Subject-Predicate Structure IP NP I’ Infl VP tamen ‘they’ NP V’ V
dayu chi xiaoyu
‘act according to the law of the jungle’
The predicatehood of such idiom chunks can be supported by the fact that they may be preceded by aspect markers such as progressive zai.
(27) Tamen yizhi zai da-yu chi xiao-yu.11 they continually ASP big-fish eat small-fish ‘They have been bullying the weaker.’
(28) Tamen zheng zai wo kan ni, ni kan wo. they right ASP I look you you look me ‘They are looking at each other.’
In these examples, the idiom chunks behave just like regular predicates.
Regarding the apparent counterexamples of (23) and (24) provided by Pan and Hu, we propose that the so-called dangling topic is in fact a PP-reduced form as in (29) and (30) and thus can be subsumed in Shi’s system.
(29) (Zai) tamen (zhizhong), ni zhize wo bu dui, (at) they (midst) you blame I not right wo baoyuan ni bu hao.
I complain you not good
‘Among them, one blames and complains about another.’
11
(30) (Zai) tamen liang (zhizhong), zhuanmen XiaoWang xian lai, (at) they two (midst) specifically XiaoWang first come XiaoLi hou dao.
XiaoLi after arrive
‘Between them two, XiaoLi always comes after XiaoWang.’
Given the comparison between the approach proposed by Shi and by Pan and Hu in this subsection, we thus conclude that Shi’s system fares better in giving a unified account of the facts. In other words, concurring with Shi, we deny the existence of this first type of dangling topic in Mandarin Chinese.
2.3 Second type
The second type of dangling topic is concerned with a subtype of the so-called double-nominative constructions (Teng 1974).
(31) Tamen shei dou bu lai.
they who all not come
‘They (topic), none of them are coming.’
Following Li’s (1992) account, Shi indicates that when a wh-word occurs to the left of an emphatic morpheme dou ‘all’, this wh-word yields an interpretation equivalent to that of a universal quantifier. Under this analysis, the wh-word shei ‘who’ in (31) should be read as ‘anyone’ instead of ‘who/which person’. Shi further points out that this universal quantifier12 is optional and can be omitted as in (32) without significantly changing the sentence meaning.
(32) Tamen dou bu lai
they all not come
‘None of them are coming.’
Given that shei ‘who’ in (31) is a universal quantifier, Shi concludes that the sentence-initial NP tamen ‘they’ is a grammatical subject, but not a dangling topic. To justify this point, Shi (2000:390) proposes that “a salient property of subject in
12
According to Shi (2000:391), the universal quantifier shei ‘who’ in (31) is likely to function as a quantificational adverbial equivalent to the adverb quan ‘all’ as in the following sentence.
(i) Tamen quan dou bu lai. they all all not come ‘None of them are coming.’
Chinese is that it must precede certain modal verbs such as keyi ‘may’, hui ‘will’ and
neng ‘can’”. Consider his examples.
(33) a. Tamen keyi bu lai. they may not come ‘They may not come.’ b. *Keyi tamen bu lai.
may they not come ‘They may not come.’
c. Tamen keyi shei dou bu lai. they may who all not come
‘None of them may come.’
The comparison of (33a) and (33b) suggests that the modal verb keyi ‘may’ only occurs between the subject and predicate. As with the adverb zhuanmen ‘specifically’, Shi takes modals as another syntactic test for the subject-predicate structure. Since (33c) can pass this modal-verb test, the subjecthood of the initial NP
tamen ‘they’ is confirmed.
However, Pan and Hu reject both Shi’s QP analysis and the modal-verb test. With respect to their critique against Shi’s QP analysis, consider (34).
(34) Tamen shei hui lai. they who will come ‘Who of them will come?’
Pan and Hu point out that since the wh-word shei ‘who’ in (34) does not precede any emphatic morpheme, it does not yield the ‘anyone’ reading but the ‘who/which person’ reading and thus cannot be analyzed as a quantificational adverbial. Rather, they argue that the wh-word shei ‘who’ at issue is the subject, which follows the dangling topic tamen ‘they’. In parallel, for Pan and Hu, (31) is also a topic-comment construction in which the wh-word shei ‘who’ is not a QP but the subject of the comment clause.
Furthermore, Pan and Hu argue against Shi’s modal-verb test for subjecthood, based on sentences like (35).
(35) Zhe-wan mifan, keyi liang-ge ren chi. this-bowl rice may two-CL person eat ‘This bowl of rice may serve two people.’
Shi’s modal-verb test would show the sequence liangge ren chi ‘two people eat’ as a predicate, but this is implausible because liangge ren chi ‘two people eat’ is a full sentence. Pan and Hu thus conclude that Shi’s modal-verb test is unreliable. To license dangling topics of this second type, Pan and Hu again propose the set-member relation. To take (31) as an example, the dangling topic tamen ‘they’ is a set and the
wh-word shei ‘who’ as the subject of the comment clause is a member within the set.
While concurring with Pan and Hu that the modal-verb test for subjecthood proposed by Shi is not reliable, we will argue in favor of Shi’s non-dangling-topic approach rather than Pan and Hu’s dangling-topic approach.
First of all, the modal-verb test is descriptively incorrect. As indicated by Lin and Tang (1995), epistemic modals and some deontic modals like yinggai ‘should’ and keyi ‘may’ may occur in a pre-subject position as in (36).
(36) a. Yinggai tamen xian qu. should they first go ‘They should go first.’ b. Keyi tamen xian qu.
may they first go ‘They may go first.’
Regarding Pan and Hu’s critique against Shi’ QP analysis, we do not consider their attack fatal to Shi’s theory. That is, although they are right that in (34) the wh-word
shei ‘who’ is not a QP, in Shi’s system, this sentence may involve a PP-reduced form
as the sentence-initial NP as in (37).
(37) (Zai) tamen (zhizhong), shei hui lai? (at) they (midst) who will come ‘Of them, who will come?’
A more crucial problem with Pan and Hu’s analysis concerns the example in (39).
(38) Tamen shei dou bu lai.
they who all not come
‘They, whoever, are not coming.’
(39) Ta sheme dou hen youxiu. he what all very outstanding ‘He, in whatever ways, is outstanding.’
Under their approach, the two sentences are expected to share a parallel structure. Since (38) is a topic-comment construction in which the dangling topic tamen ‘they’ and the wh-word shei ‘who’ exhibit a membership relation, this relation should also hold true of (39), in which the wh-word sheme ‘what’ should serve as a member of the dangling topic ta ‘he’ as the set. Notice, however, that in (39) ta ‘he’ is human but
sheme ‘what’ is nonhuman. It is doubtful that membership could exist between two
nominals bearing different features or attributes. Hence, (39) poses a problem for Pan and Hu’s set-member analysis.13 Under Shi’s approach, the wh-word sheme ‘what’ in sentences like (39) can be analyzed as an adverbial QP on a par with the
wh-word shei ‘who’ in (38). Since Shi’s approach explains facts better, we therefore
conclude that the first nominal in this subtype of double-nominative constructions is not a dangling topic, but a subject.
2.4 Third type
The third type of dangling topic is related to connectives such as xingkui ‘fortunately’, kexi ‘pitifully’, yaobushi ‘otherwise’, etc. The following sentence is well known and often cited.
(40) Na-chang huo, xingkui xiaofang-dui lai-de-kuai. that-CL fire fortunately fire-brigade come-DE-fast
‘As for that fire, fortunately the fire brigade came quickly; (otherwise)…’
Shi indicates that (40) reflects only part of the discourse. The complete discourse for it would be like (41) or (42).
13
Alternatively, some people might argue that in this subtype of double-nominative constructions, the first nominal like tajia ‘his house’ in (i) is actually generated from tajiade ‘of his house’ in (ii) by reducing the genitive marker de without changing semantics.
(i) Ta-jia sheme dongxi dou hen xin. he-house what things all very new
‘Whatever things in his house are new.’
(ii) Ta-jia-de sheme dongxi dou hen xin. he-house-GEN what things all very new
‘Whatever things in his house are new.’
Under this analysis, the sentence-initial NP tajia ‘his house’ in (i) is a genitive specifier rather than a dangling topic or grammatical subject. From a generative viewpoint, however, these two sentences are not structurally identical. Some scholars (Huang 1982, Shi 1992, Qu 1994) argue that (i) is derived from (ii); the genitive phrase in (ii) has been preposed from the specifier position within IP to higher TopicP. If we adopt this transformational analysis, the sentence-initial NP tajia ‘his house’ in (i) is not a dangling topic but a non-dangling one, since it has undergone movement and left a gap behind.
(41) Na-chang huoi, xingkui xiaofang-dui lai-de-kuai, that-CL fire fortunately fire-brigade come-DE-fast
buran ∅i jiu hui shao-si bu-shao ren. otherwise really will burn-die not-few person
‘As for that fire, fortunately the fire brigade came quickly, or (it) would have killed many people.’
(42) Na-chang huoi, xingkui xiaofang-dui lai-de-kuai, that-CL fire fortunately fire-brigade come-DE-fast buran na-cii women dou hui shao-si.
otherwise that-time we all will burn-die
‘As for that fire, fortunately the fire brigade came quickly; otherwise we would all have been burnt to death at that time.’
In both sentences, the sentence-initial NP nachang huo ‘that fire’ is related to a position in the recovered main clause, so nachang huo ‘that fire’ is a non-dangling topic. Shi further suggests that if we take away the connective xingkui ‘fortunately’ in (40), the sentence becomes unrecoverable but is still well-formed.
(43) Na-chang huo, xiaofang-dui lai-de-kuai. that-CL fire fire-brigade come-DE-fast
‘At the time of that fire, the fire brigade came quickly.’
In this case, the initial NP nachang huo ‘that fire’ is unrelated to any position inside the following clause and thus cannot be licensed as a topic. Rather, it is claimed by Shi to be a sentential adverbial.
Rejecting Shi’s analysis based on the recoverable discourse, Pan and Hu argue that even if the recovered main clause has a gap for the sentence-initial NP, this gap does not guarantee the availability of the un-recovered sentence. Consider their examples below.
(44) Zhe-jian da shii, xingkui Zhang-Xiaozhang lai le,
this-CL big issue fortunately Zhang-Principal come ASP
yaoburan wo hai bu zhidao ruhe chuli∅i . otherwise I still not know how deal-with
‘As for this big issue, fortunately Principal Zhang has come; otherwise I do not know how to deal with it.’
(45) *Zhe-jian da shi, xingkui Zhang-Xiaozhang lai le.
this-CL big issue fortunately Zhang-Principal come ASP
‘As for this big issue, fortunately Principal Zhang has come.’
Contrary to Shi’s analysis, Pan and Hu propose a cause-effect relation to license dangling topics of this third type. They (2002:10) claim that “the semantics of
xingkui ‘fortunately’ is that if the S introduced by xingkui ‘fortunately’ is not true,
then some consequence S’ which is not given in the sentence in question will follow. The consequence S’ is caused by the topic. So the relationship between the topic and S’ is a cause-effect relation”. In brief, it is the topic that causes the undesirable effect to be true. If we apply this cause-effect account to (40), the fire is claimed to be the cause, and the undesirable effect to be people’s death.
We share the view that the recovering analysis does not hold, given sentences like the following.
(46) Na-chang huo, xingkui xiaofang-dui lai-de-kuai, that-CL fire fortunately fire-brigade come-DE-fast yaoburan ta zao jiu si le.
otherwise he early really die ASP
‘As for the fire, fortunately the fire brigade came quickly; otherwise he would have been dead.’
Assuming that there is no adjunct pro that can be co-indexed with the sentence-initial NP, there is no position in the recovered main clause that can license the topic status of the initial NP nachang huo ‘that fire’. Without syntactic dependency, nachang
huo ‘that fire’ cannot be licensed as a topic in Shi’s theory. This prediction, however,
is not borne out and thus casts doubt on the recovering account provided by Shi. Although we concur with Pan and Hu’s attack on Shi’s recovering analysis of sentences containing connectives like xingkui ‘fortunately’ as in (41) and (42), their approach based on the cause-effect relation to license dangling topics, however, has its share of problems as can be seen with examples in (47).
(47) a. Zhe-jian qiangan, xingkui jingcha lai-de-zao.
this-CL robbery fortunately police come-DE-early ‘As for this robbery, fortunately the police came early.’
b. Na-chang Wenbudun wang-qiu sai, xingkui da-yu ting le. that-CL Wimbledon net-ball match fortunately big-rain stop ASP ‘As for that Wimbledon match, fortunately the heavy rain stopped.’
c. Taizhong-xian Songhe-buluo, xingkui tu-shi-liu Taichung-county Songhe-village fortunately soil-stone-slide yi tingzhi fanlan.
already stop overflow
‘As for the Songhe village in Taichung County, fortunately the landslide already stopped spreading.’
It makes little sense to consider the sentence-initial NP, say, nachang Wenbudun
wangqiu sai ‘that Wimbledon match’ in (47b), as a cause. In this case, if the implied
effect is the delay of the match, the reasonable cause for this delay should be the heavy rain as expressed in the subordinate, connective clause, rather than the dangling topic as claimed by Pan and Hu. For this reason, we reject Pan and Hu’s cause-effect analysis of this third type of dangling topic.
In fact, there is an account under Shi’s approach that can give a unified analysis of all the facts that have been considered so far in this subsection. That is, with or without a connective in the ‘comment’ clause, the sentence-initial NP can be treated as an adverbial.14 In support of this proposal, consider the contrast between (48) and (49).
(48) a. Zhe-jian qiangan, (xingkui) jingcha lai-de-zao. this-CL robbery (fortunately) police come-DE-early ‘During this robbery, (fortunately) the police came early.’ b. Taizhong-xian Songhe-buluo, (xingkui) tu-shi-liu Taichung-county Songhe-village (fortunately) soil-stone-slide
yi tingzhi fanlan. already stop overflow
‘At the Songhe village in Taichung County, (fortunately) the landslide already stopped spreading.’
c. Na-chang Wenbudun wang-qiu sai, (xingkui) da-yu ting le. that-CL Wimbledon net-ball match (fortunately) big-rain stop ASP ‘During that Wimbledon match, (fortunately) the heavy rain stopped.’
14
NPs such as yesterday, that way, etc. have been observed to behave as an adverbial since Bresnan and Grimshaw (1978). Varied terminology appears in the literature, e.g. “bare-NP adverb” by Larson (1985), “adverbial NP” by McCawley (1988) and “adverbial” by Stroik (1992). In this paper, we use the term “NP adverbial” due to the purpose to stress the adverbial status of the NPs in question.
(49) a. *Zhe-ge xiaotou, (xingkui) jingcha lai-de-zao.
this-CL thief (fortunately) police come-DE-early
‘As for this thief, (fortunately) the police came early.’ b. *Zhe-zuo shan-wu, (xingkui) tu-shi-liu this-CL mountain-house (fortunately) soil-stone-flow yi tingzhi fanlan.
already stop overflow
‘As for this mountain cottage, (fortunately) the landslide already stopped spreading.’
c. *Zhege dan-gao, (xingkui) wo jingguo chufang.
this-CL egg-pastry (fortunately) I pass by kitchen
‘As for this cake, (fortunately) I passed by the kitchen.’
In both sets of examples, sentence-initial NPs are not subcategorized by the verb. But why is there the acceptability contrast? One natural account is that the sentence-initial NPs in (48) are NP adverbials whereas those in (49) are argument NPs that need to be subcategorized by the predicate as shown by the acceptability of (50).
(50) a. Zhe-ge xiaotoui, xingkui jingcha lai-de-zao, this-CL thief fortunately police come-DE-early yaoburan zao jiu bei tai tao-pao le.
otherwise early really BEI he flee-run ASP
‘As for this thief, fortunately the police came early; otherwise s/he would have fled.’
b. Zhe-zuo shan-wui, xingkui tu-shi-liu yi tingzhi this-CL mountain-house fortunately soil-stone-flow already stop
fanlan, yaoburan ∅i zao jiu hui le.
overflow otherwise early really destroy ASP
‘As for this mountain cottage, fortunately the landslide already stopped spreading, or (it) would have been destroyed.’
c. Zhe-ge dan-gaoi, xingkui wo jingguo chufang, this-CL egg-pastry fortunately I pass by kitchen yaoburan ∅i zao jiu bei ta chi-diao le.
otherwise early really BEI he eat-drop ASP
‘As for this cake, fortunately I passed by the kitchen, or (it) would have been eaten by him.’
categorized as the third type are actually NP adverbials.
2.5 Fifth type
The fifth type of dangling topic is concerned with nominal predicates. Shi cites the following example from Li and Thompson (1976).
(51) Na-zhong douzi yi-jin san-shi-kuai qian. that-CL bean one-catty three-ten-CL money ‘That kind of bean (topic), one catty is thirty dollars.’
This is analyzed as a subject-predicate structure by Shi. The initial NP nazhong
douzi ‘that kind of bean’ is the subject, and the following string is the nominal
predicate with yijin ‘one catty’ as a QP. In support of this analysis, Shi proposes two tests: the modal-verb test and the shi ‘be’ test. The modal-verb test has been shown to be invalid. Regarding the shi ‘be’ test, Shi indicates that “the emphatic marker shi ‘be’ can appear right before the VP, the adverbial or the subject of a sentence, but it cannot occur in front of the topic or after the verb (Teng 1979, Huang 1991, Shi 1994)”.
(52) Shi na-zhong douzi yi-jin san-shi-kuai qian. be that-CL bean one-catty three-ten-CL money ‘It is that kind of bean that is worth thirty dollars a catty.’
Given that the emphatic marker shi ‘be’ occurs before the subject but not the topic, the subjecthood of nazhong douzi ‘that kind of bean’ in (52) is thus confirmed.
Concerning the shi ‘be’ test, Pan and Hu argue against it based on examples like
(53).
(53) Shi zhe-zhong douzii wo mei mai ∅i .
be this-CL bean I not buy
‘It is this kind of bean that I did not buy.’
Here, shi ‘be’ precedes the topic NP zhezhong douzi ‘this kind of bean’, contrary to what is expected under Shi’s theory. Given this fact, Pan and Hu thus conclude that the shi ‘be’ test is unreliable. We agree with this critique of Shi’s shi ‘be’ test for distinguishing topic and subject, but we do not consider their argument valid for treating the sentence-initial NP nazhong douzi ‘that kind of bean’ in (51) as a topic.
A criterion often invoked to distinguish topic and subject is based on the definiteness of the NP at issue: a topic must be definite but a subject need not be (Li and Thompson 1976/1981, Tsao 1990, Tan 1991, Shyu 1995). Given the obligatory definiteness of the NP as shown by the unacceptability of (54), Pan and Hu argue that this clearly indicates that the NP at issue is a topic rather than a subject.
(54) *Yi-zhong douzi yi-jin san-shi-kuai qian. one-CL bean one-catty three-ten-CL money ‘One kind of bean (topic), one catty is thirty dollars.’
To license dangling topics of this fifth type, again Pan and Hu propose the set-member account. In their view, the dangling topic nazhong douzi ‘that kind of bean’ in (51) is a set among which yijin ‘one catty’ expressed in the comment clause is delimited as the member of the dangling topic.
Their analysis, however, is not tenable. First, note that there is a strong tendency for clause-initial NPs in Chinese to be definite possibly due to discourse factors. The definiteness constraint on a subject can only be relaxed in a limited number of contexts. For example, when the nominal at issue denotes a quantity (see Li 1998) as in (55a), expresses a unit reading (see Shi 2000) as in (55b), or occurs with stage-level predicates in a root context (see Shyu 1995) as in (55c), they can occur in the grammatical subject position.
(55) a. San-ge ren chi liang-wan fan. three-CL person eat two-bowl rice ‘Three people eat two bowls of rice.’ b. Yi-zhi qingwa si-tiao tui.
one-CL frog four-CL leg
‘A frog (has) four legs.’
c. Yi-wei yi-sheng xiang wo jieshao ta-de bing-ren. one-CL doctor toward I introduce he-GEN sick-person ‘A doctor introduced me to his patients.’
Consider the unacceptable example in (54). If the sentence-initial NP is indeed a subject, the sentence may be ruled out simply because the NP yizhong douzi ‘one kind of bean’ does not meet the requirement for an indefinite subject for any reason. This analysis is supported by the unacceptability of (56) without the alleged QP yijin ‘one catty’.
(56) *Yi-zhong douzi san-shi-kuai qian. one-CL bean three-ten-CL money ‘One kind of bean is thirty dollars.’
Pan and Hu’s set-member analysis is further rendered questionable if we consider the example in (57), which may be heard at a car rental store.
(57) Yi-liang che yi-tian san-qian. one-CL car one-day three-thousand ‘One car is three thousand dollars a day.’
It is not clear to us how the initial NP yiliang che ‘one car’ in this example can qualify as a set with either yitian ‘one day’ or sanqian ‘three thousand (dollars)’ included as a member of the set.
In brief, in this subsection we have shown that even though Shi’s test for distinguishing subject and topic based on shi ‘be’ does not hold, his theory fares better in accounting for empirical facts than the one proposed by Pan and Hu.
2.6 Sixth type
With regard to the last type of dangling topic, consider the following example from Chen (1996).
(58) Wu-jia Niuyue zui gui
thing-price New York most expensive
‘Speaking of the price of things, New York is the most expensive.’
Shi argues that it is more natural to think of the price of things, instead of the place/location like New York, as the highest in value and thus that the subject should be the initial NP wujia ‘the price of things’. Under this analysis, Niuyue ‘New York’ is treated as a locative adverbial. In support of the initial NP wujia ‘the price of things’ in (58) not as a dangling topic but a subject, Shi resorts to a test based on adverb placement.
(59) Wu-jia yi-ding Niuyue zui gui.
thing-price certainly New York most expensive ‘The price of things is certainly the highest in New York.’
Shi claims that adverbs like yiding ‘certainly’ and jingchang ‘often’ can occur only between the subject and predicate. Since (59) can pass this adverb test, the initial NP
wujia ‘the price of things’ must be a subject.
As with Shi’s zhuanmen ‘specifically’ and modal-verb tests, Pan and Hu argue against this adverb test based on the following example.
(60) Niuyue yiding wu-jia zui gui.
New York certainly thing-price most expensive ‘The price of things is certainly the highest in New York.’
Shi’s adverb test would predict that the post-adverb string wujia zui gui ‘the price of things is the highest’ is the predicate. However, this prediction is undesirable because the string is a full sentence. We agree with Pan and Hu’s critique that adverbs cannot be used to test the subject-predicate structure.
To account for this sixth type of dangling topic like wujia ‘the price of things’ in (58), Pan and Hu again propose the set-member account.
(61) Wu-jia, Niuyue (-de wu-jia) zui gui.
thing-price New York (-GEN thing-price) most expensive ‘As for the price of things, New York(’s price of things) is the highest.’
It is claimed that wujia ‘the price of things’ is the dangling topic, and that Niuyue (de
wujia) ‘New York (’s price of things)’ is the subject of the comment clause. As
indicated by the parentheses, the subject Niuyue ‘New York’ contains an empty head noun wujia ‘the price of things’. Under this analysis, the sentence is interpreted as meaning that when commodity prices all over the world are considered, the price in New York is the highest. That is, the initial NP wujia ‘the price of things’ is a set representing the world’s commodity prices, among which the price in New York is a member.
Promising as this set-member account may appear, it runs into a problem when we try to license the dangling topic the same way as shown in (62).
(62) *Mama, Zhangsan (-de Mama) zui qinqie. mother Zhangsan (-GEN mother) most kind ‘As for mothers, Zhangsan(’s mother) is the kindest.’
Here, (62) shares a parallel structure with (61); its unacceptability, however, is unexpected under the set-member account. We thus conclude that Shi’s
non-dangling-topic approach fares better than Pan and Hu’s dangling-topic approach.
3. Conclusion
This paper has compared and contrasted two different views as to whether dangling topics exist in Mandarin Chinese. Shi’s approach denies such a possibility. According to his analysis, all topics in Chinese must have a structural position, either occupied by a resumptive pronoun or an empty category, inside the comment clause. In other words, all Chinese topics are syntactically licensed. He attributes so-called dangling Chinese-style topics to be subjects, (non-dangling) topics, PP-reduced forms, or sentential adverbials. In contrast, Pan and Hu’s approach acknowledges the existence of dangling topics in Mandarin Chinese. Under their analysis, Chinese topics can be either syntactically or semantically licensed. Semantically licensed topics are dangling topics. They can be licensed in terms of a thematic relation, set-member relation or cause-effect relation.
In this paper, we have argued against Pan and Hu’s dangling-topic approach but in favor of Shi’s non-dangling-topic approach. Throughout the discussion, we have carefully examined the counterarguments and alternative proposals provided by Pan and Hu. The results of this study show that although some tests (e.g. based on the placement of modal verbs, shi ‘be’ and adverbs) used by Shi do not go through, the problems pointed out by Pan and Hu can be solved within Shi’s system. In other words, we have modified some analyses given by Shi in order to take the counterexamples observed by Pan and Hu into consideration. The comparison among the analyses provided by Shi, Pan and Hu and the present authors is shown in Table 2.
Table 2. Three views on status of sentence-initial NPs
Six Types Pan and Hu Shi This paper
1 [Tamen] da-yu chi xiao-yu. Dg. topic Subject Subject
2 [Tamen] shei dou bu lai. Dg. topic Subject Subject
[Na-chang huo], xingkui xiaofang-dui
lai-de-kuai. Dg. topic NP topic NP adverbials
3
[Na-chang huo] xiaofang-dui lai-de-kuai. Dg. topic NP adverbials NP adverbials [Zhe-jian shiqing] ni bu neng guang
mafan yi-ge ren. Dg. topic
PP-reduced
form NP topic
[Shengwu-lunlixue] wo shi men-wai-han. Dg. topic PP-reduced form
PP-reduced form 4
[Xihongshi] wo chao le ji-dan. Dg. topic Not mentioned NP topic 5 [Na-zhong douzi] yi-jin san-shi-kuai qian. Dg. topic Subject Subject
6 [Wu-jia] Niuyue zui gui. Dg. topic Subject Subject
As indicated in Table (2), we have taken the approach advocated by Shi but departed from his analysis in analyzing several subtypes of the alleged “dangling-topic” constructions. More specifically, what differentiates Shi’s analysis and ours concerns what we do in fixing problems that arise in Shi’s analysis. As should be clear from the discussion in Section 2, in this paper we have evaluated two opposing views toward the issue of dangling topics and the result of our investigation favors Shi’s approach. Despite adopting his approach, we have fixed some problems with his analyses that have been pointed out by Pan and Hu. Of the two approaches, Shi’s is not well recognized15 and Pan and Hu’s, in fact, is a more traditional one. Our paper concludes that the problems pointed out by Pan and Hu for Shi’s approach are only apparent – even though they are problems for Shi’s ANALYSIS, they can be solved in Shi’s SYSTEM. We believe that our analysis has shown that Shi’s approach is a workable and feasible one.
In summary, this paper rejects Pan and Hu’s semantic licensing of dangling topics and supports Shi’s analysis that so-called dangling topics are actually subjects, NP topics, NP adverbials, or PP-reduced forms.
15
Shi’s approach to “dangling-topic” constructions is not even mentioned by Huang, Li and Li (2004) in their review of previous analyses of such constructions.
References
Bresnan, J. and J. Grimshaw. 1978. The syntax of free relatives in English. Linguistic
Inquiry 9:331-391.
Chafe, Wallace. 1976. Givenness, contrastiveness, definiteness, subjects, topics and point of view. Subject and Topic, ed. by Charles Li, 27-55. New York: Academic Press.
Chao, Yuen-Ren. 1968. A Grammar of Spoken Chinese. Berkeley: University of California Press.
Chen, Ping. 1996. Pragmatic interpretation of structural topic and relativization in Chinese. Journal of Pragmatics 26:389-406.
Huang, C.-T. James. 1982. Logical Relations in Chinese and the Theory of Grammar. Cambridge: MIT dissertation.
Huang, C.-T. James. 1991. On be and have in Chinese. Essays in Memory of Professor
Fangkwei Li, ed. by Institute of History and Philology, Taipei, 43-62.
Huang, C.-T. James, Y.-H. Audrey Li and Yafei Li. 2004. Syntax of Chinese. To be published by Cambridge University Press. Downloadable at
http://my.harvard.edu/icb/icb.do?course=fas-ling171andpageid=tk.page.ling171. dir.b7fe4f2f58914cb51ea04ae77dbeb4a2
Huang, Yan. 1994. The Syntax and Pragmatics of Anaphora: With Special Reference
to Chinese. Cambridge: Cambridge University Press.
LaPolla, Randy. 1990. Grammatical Relations in Chinese: Synchronic and Diachronic
Considerations. Berkeley: University of California dissertation.
Larson, Richard K. 1985. Bare-NP adverbs. Linguistic Inquiry 16.4:595-621. Li, Charles, and Sandra Thompson. 1976. Subject and topic: A new typology of
language. Subject and Topic, ed. by Charles Li, 457-489. New York: Academic Press.
Li, Charles, and Sandra Thompson. 1981. Mandarin Chinese: A Functional Reference
Grammar. Berkeley: University of California Press.
Li, Y-H. Audrey. 1992. Indefinite WH in Mandarin Chinese. Journal of East Asian
Linguistics 1:125-156.
Li, Y-H. Audrey. 1998. Argument determiner phrases and number phrases. Linguistic
Inquiry 29.4:693-702.
Lin, Jo-Wang, and C.-C. Jane Tang. 1995. Modals as verbs in Chinese: A GB perspective. Bulletin of the Institute of History and Philology 66:53-105. Lu, Shuxiang. 1986. The flexibility of Chinese syntax. Zhongguo Yuwen 1. [In
McCawley, James D. 1988. Adverbial NPs: Bare or clad in see-through garb?
Language 64.3:583-590.
Ning, Chunyan. 1993. The Overt Syntax of Relativization and Topicalization. Irvine: University of California dissertation.
Pan, Haihua, and Jianhua Hu. 2002. Licensing dangling topics in Chinese. Paper Presented at the 2002 LSA Annual Meeting in San Francisco, CA, USA.
Qu, Yanfeng. 1994. Object Noun Phrase Dislocation in Mandarin Chinese. Vancouver: University of British Columbia dissertation.
Shi, Dingxu. 1992. The Nature of Topic Comment Constructions and Topic Chains. Los Angeles: University of Southern California dissertation.
Shi, Dingxu. 1994. The nature of Chinese emphatic sentences. Journal of East Asian
Linguistics 3:81-100.
Shi, Dingxu. 2000. Topic and topic-comment constructions in Mandarin Chinese.
Language 76.2:383-408.
Shyu, Shu-ing. 1995. The Syntax of Focus and Topic in Mandarin Chinese. Los Angeles: University of Southern California dissertation.
Stroik, Thomas. 1992. On the distribution of temporal and locative NP adverbials. The
Linguistic Review 9.3:267-284.
Tan, Fu. 1991. Notion of Subjects in Chinese. Stanford: Stanford University dissertation.
Tang, C.-C. Jane. 1990. Chinese Phrase Structure and the Extended X’-Theory. Ithaca: Cornell University dissertation.
Teng, Shou-hsin. 1974. Double nominatives in Chinese. Language 50:455-473. Teng, Shou-hsin. 1979. Remarks on cleft sentences in Chinese. Journal of Chinese
Linguistics 7:101-114.
Thompson, Sandra A. 1973. Transitivity and some problems with the ba construction in Mandarin Chinese. Journal of Chinese Linguistics 1:208-221.
Tsao, Feng-fu. 1990. Sentence and Clause Structure in Chinese: A Functional
Perspective. Taipei: Student Book Co.
Williams, Edwin. 1980. Predication. Linguistic Inquiry 11:203-238.
Xu, Liejiong, and D. Terrence Langendoen. 1985. Topic structures in Chinese.
Language 61:1-27.
[Received 3 October 2005; revised 20 November 2005; accepted 22 November 2005] Department of English
National Taiwan Normal University Taipei, TAIWAN
Rui-heng Huang: Qrayhuang@yahoo.com.tw Jen Ting: ting@ntnu.edu.tw
