Abstract

While canonical gapping in English can be analyzed as either VPellipsis (e.g., Sag 1976, Pesetsky 1982, Jayaseelan 1990, Lasnik 1999a, Johnson 2000, Coppock 2001, Lin 2002, Baltin 2003, Takahashi 2004) or across-the-board V/VP-movement (Johnson 1994, 2004, 2006, 2009), certain English-like gapping constructions in Modern Mandarin are argued to be multiple sentence fragments, formed by a series of syntactic operations that involve topicalization, focus movement, and IP-deletion.

1 Introduction

Whether Modern Mandarin allows gapping or not has long been controversial in the literature. Some linguists working on Chinese (e.g., Tai 1969, M.-D. Li 1988, Tsai 1994, Tang 2001) maintain that Modern Mandarin disallows English-like canonical gapping of the type shown in (1a–b).

(1)

  • a.

    John likes apples and Mary _____ oranges.

  • b.

    John saw Mary and Bill _____ Susan.

According to these researchers, the Modern Mandarin counterparts of (1a–b) are ungrammatical, as indicated in (2)–(3).

(2)

  • *Zhangsan xihuan pingguo, Lisi _____ juzi.

  • Zhangsan like apple Lisi orange

  • ‘Zhangsan likes apples and Lisi _____ oranges.’

(3)

  • *Zhangsan kanjian-le Shufen, Lisi _____ Yaping.

  • Zhangsan see-ASP Shufen Lisi Yaping

  • ‘Zhangsan saw Shufen and Lisi _____ Yaping.’

For those who disagree with these judgments (see, e.g., Wu 2002, 2003), the controversy arises from the fact that (2) and (3) can be improved with context. For example, utterance (2) is marginally acceptable as an answer to the question Tamen dou xihuan shenme? (‘What do they all like?’). Utterance (3) can also be greatly improved if it is an answer to the question Natian zai shanshang, tamen dou kanjian-le shei? (‘That day on the mountain, who did they all see?’). This is illustrated in (4) and (5), respectively.

(4)

  • Q:

    Tamen dou xihuan shenme?

    they all like what

    ‘(Lit.) They all like what?’1

  • A:

    (?)Zhangsan xihuan pingguo. Lisi _____ juzi.

    Zhangsan like apple Lisi orange

    ‘Zhangsan likes apples and Lisi _____ oranges.’

(5)

  • Q:

    Natian zai shan-shang, tamen dou kanjian-le shei?

    that.day on mountain-above they all see-ASP who

    ‘(Lit.) That day on the mountain, they all saw whom?’

  • A:

    (?)Zhangsan kanjian-le Shufen. Lisi _____ Yaping.

    Zhangsan see-ASP Shufen Lisi Yaping

    ‘Zhangsan saw Shufen and Lisi _____ Yaping.’

These context-sensitive gapping examples, along with some other English-like canonical gapping examples that do not require a special context, such as (6) (see, e.g., Paul 1996, 1997, 1999, Y.- H. A. Li 1998, Wu 2002, 2003), point to one conclusion: Modern Mandarin (at least for some linguists working on Chinese) does have a construction reminiscent of English canonical gapping.

(6)

  • Zhangsan chi-le san-ge pingguo, Lisi _____ si-ge juzi.

  • Zhangsan eat-ASP three-CL apple Lisi four-CL orange

  • ‘Zhangsan ate three apples, and Lisi _____ four oranges.’

In this article, I will not discuss to what extent gapping is restricted in Modern Mandarin, since this has been well discussed in the literature (see, e.g., Paul 1999, Tang 2001, and Ai 2005, for the debatable generalization that it is related to the quantificational force of the object NP and nongeneric VPs). Instead, I will focus on the syntactic analysis of the aforementioned elliptical constructions that behave like English canonical gapping.

2 The Problem

Gapping in English has been traditionally analyzed as (VP-) ellipsis that applies after the relevant object NP has been extracted from the ellipsis site (e.g., Sag 1976, Pesetsky 1982, Jayaseelan 1990, Lasnik 1999a, Johnson 2000, Coppock 2001, Lin 2002, Baltin 2003, Takahashi 2004). Applied to (1a), for example, this approach can be schematized as in (7), abstracting away from details and complications.

(7)

graphic

This analysis predominated in the literature until the work of Johnson (1994, 2004, 2006, 2009), who argues extensively that canonical English gapping is not an instance of (VP-) ellipsis. On Johnson’s approach, the relevant construction actually involves across-the-board (ATB) movement of predicates. Diagram (8) illustrates this approach, again abstracting away from details and complications.

(8)

graphic

This analysis is based crucially on the following two pieces of empirical evidence: (a) gapping in English is restricted to coordinate structures (CSs), and (b) gapping in English is sensitive to island constraints (see also Donati 2003). Island sensitivity is normally a diagnostic for movement operations in the literature.2

In this article, I will assume that Johnson’s (ATB) movement analysis of gapping in English is on the right track. Paul (1999) has also attempted to analyze English-like gapping in Modern Mandarin along the same lines. However, a detailed examination of English-like gapping in Modern Mandarin reveals that such an analysis is unsupported. In the first place, gapping in Modern Mandarin is not restricted to CSs. For example, the gap can be embedded in a sentential subject position, as illustrated in (9). Furthermore, gapping in Modern Mandarin does not seem to obey typical island constraints. This is demonstrated in (10), a straightforward violation of the Complex NP Constraint (CNPC) if the (ATB) movement analysis is assumed for gapping in Modern Mandarin.

(9)

  • Zhangsan kaojin-le Tsinghua,

  • Zhangsan exam.enter-ASP Tsinghua

  • ‘Zhangsan exam-entered Tsinghua University,

  • [wo _____ Beida] ziran mei wenti. (*CSs)

  • I Beida of.course no problem

  • [that I *(exam-entered) Beijing University] is of course no problem.’

(10)

  • Zhiyao Zhangsan yitian bu kaojin Tsinghua,

  • as.long.as Zhangsan one.day not exam.enter Tsinghua

  • ‘As long as Zhangsan has not exam-entered Tsinghua University,

  • ni jiu beng ti [[wo _____ Beida] de shi]. (*CNPC)

  • you then not mention I Beida DE fact

  • you will not mention the fact that I *(have exam-entered) Beijing University.’

These two examples jointly undermine the foundation for an (ATB) movement account for English- like gapping in Modern Mandarin.

3 Tang’s (2001) Analysis

For Tang (2001), examples like (6) are not instances of gapping in Modern Mandarin.3 Instead, they are simply empty-verb sentences, similar to (11).

(11)

  • Wo _____ Zhangsan.

  • I Zhangsan

  • ‘I am Zhangsan.’

Example (11) illustrates that the copular shi ‘be’ can be omitted in Modern Mandarin. Tang (2001) argues that verb omission in Modern Mandarin is more general: not only can the copular shi be omitted, but there exists an empty verb in Modern Mandarin that can license empty-verb sentences with the structure in (12).

(12)

  • Empty-verb sentences in Modern Mandarin

  • XP [VP [V ∅] YP]

  • (Tang 2001:208, (17))

In (12), the empty verb can be interpreted as a transitive verb. This accounts for examples like (13), which, according to Tang (2001), are greatly improved compared with (2) and (3).4

(13)

  • Wo xuan-le Zhangsan, ni Lisi.

  • I elect-ASP Zhangsan you Lisi

  • ‘I voted for Zhangsan, and you Lisi.’

  • (Tang 2001:207, (27))

Thus, instead of treating the second conjunct in (13) as an instance of gapping, Tang (2001:208) argues that ‘‘the second conjunct can be treated as an empty verb sentence, in which Lisi is the object of the empty verb, interpreted as the patient of the action.’’

4 Problems with Tang’s (2001) Analysis

There are several problems with Tang’s (2001) analysis. First of all, empty-verb sentences have a very limited distribution in Modern Mandarin. They are only used in pragmatic contexts where people need to be identified, such as self-introduction (11), or in contexts where listing is demanded, such as (14) or (15).

(14) Record-keeping

  • Zhangsan _____ pingguo san-ge; Lisi _____ juzi si-ge.

  • Zhangsan apple three-CL Lisi orange four-CL

  • ‘Zhangsan [bought, ate, etc.] three apples, and Lisi, four oranges.’

(15) Juxtaposed answers to a question that demands listing, similar to (4)–(5)

  • Q:

    Ni-men dou chi-le shenme?

    you-PL all eat-ASP what

    ‘What did you all eat?’

  • A:

    Wo _____ san-wan fan, Zhangsan _____ liang-wan tang, Lisi _____ shi-ge shuijiao.

    I three-CL rice Zhangsan two-CL soup Lisi ten-CL dumpling

    ‘I ate three bowls of rice, Zhangsan drank three bowls of soup, and Lisi ate ten dumplings.’

In (14), the context calls for record-keeping. In this scenario, the verb can be reconstructed as ‘buy’, ‘eat’, or any other verb that suits the context. To keep a record for ‘Who bought what?’,for example, (14) can be appropriately uttered. In the context in (15), juxtaposed answers are needed for a question that demands listing. Interestingly, notice that in (15), the ‘‘reconstructed’’ verbs do not have to be identical. Thus, the relationship between wo ‘I’ and san-wan fan ‘three bowls of rice’, and between Lisi and shi-ge shuijiao ‘ten pieces of dumpling’, can be reconstructed as chi ‘eat’; and the relationship between Zhangsan and liang-wan tang ‘two bowls of soup’ can be reconstructed as another verb, he ‘drink’. This is because for most native speakers of Modern Mandarin, it is more customary to say he tang ‘to drink soup’ than to say chi tang ‘to eat soup’ (though the latter is acceptable in some dialects of Modern Mandarin).5 This is in sharp contrast with the aforementioned English-like gapping examples in Modern Mandarin. Take (6), for example, repeated here as (16).

(16)

  • Zhangsan chi-le san-ge pingguo, Lisi _____ si-ge juzi.

  • Zhangsan eat-ASP three-CL apple Lisi four-CL orange

  • ‘Zhangsan ate three apples, and Lisi _____ four oranges.’

Here, the reconstructed verb of the target clause (the second conjunct) must be identical with the verb in the antecedent clause. In other words, the verb in the target clause cannot be reconstructed as a verb other than chi ‘to eat’. Thus, (17a) is the only possible interpretation for (16); (17b) is excluded.

(17) Zhangsan chi-le san-ge pingguo. Lisi _____ si-ge juzi.

  • a.

    Zhangsan chi-le san-ge pingguo. Lisi chi-le si-ge juzi.

  • b.

    *Zhangsan chi-le san-ge pingguo. Lisi mai-le si-ge juzi.

    buy-ASP

Recall that ‘‘identity’’ is always the licensing condition for ellipsis (see Ai 2008 for extensive discussion). Examples like (16) must have involved elliptical constructions that require identical linguistic antecedents. This is further supported by the observation that without certain identical linguistic antecedents, English-like gapping examples in Modern Mandarin such as those in (4) and (5) become ungrammatical; compare (18a) with (4) and (18b) with (5).

(18)

  • a.

    *Lisi _____ juzi.

    Lisi oranges

  • b.

    *Lisi _____ Yaping.

    Lisi Yaping

If these were empty-verb sentences, they should be able to occur independently like other emptyverb sentences such as (11).6

5 The New Analysis

Since neither the (ATB-)movement-of-predicates analysis nor the empty-verb-sentences analysis can successfully account for the properties of English-like gapping in Modern Mandarin, an alternative analysis must be found. In what follows, I argue with Tang (2001) that English-like gapping in Modern Mandarin is not ‘‘gapping’’ per se. However, contra Tang’s (2001) emptyverb- sentences analysis, I argue that the relevant construction is nothing more than multiple sentence fragments, formed by a series of syntactic operations that involve topicalization, focus movement, and IP-deletion.

5.1 Theoretical Background

To start with, it is well-known that the (ATB-)movement-of-predicates analysis in general involves two or more coordinated conjuncts. The operation requires that we consider all conjuncts simultaneously. If gapping in Modern Mandarin is not derived via (ATB) movement, then we do not have to consider all conjuncts at the same time. In other words, we can just focus on the target clause—that is, the gapped one. Some other elliptical constructions, such as pseudogapping, have been analyzed in this fashion (Lasnik 1995, 1999a,b). Under this approach, the derivation of a typical pseudogapping construction like (19) is as shown in (20).

(19) John will kick you, like he did _____ his brother.

(20)

graphic

In this fashion, gapping and pseudogapping have been analyzed similarly in the literature before Johnson 1994, 2004, 2006, 2009. The only difference between Lasnik’s analysis of pseudogapping (20) and the (VP-) ellipsis analysis of gapping (7) is that instead of moving rightward and adjoining to VP, the seemingly ‘‘stranded’’ object NP now moves leftward and merges into Spec,AgrO. The lower VP in both gapping and pseudogapping is subsequently deleted.

If gapping in Modern Mandarin cannot be analyzed as (ATB) movement of predicates, it is natural to treat it as an instance of ellipsis, along the lines of the (VP-) ellipsis approach illustrated in (7). In this article, I will follow Lasnik in focusing on the target clause when analyzing Englishlike gapping in Modern Mandarin. I will also assume Lasnik’s leftward movement of the object NP. The modifications that I will make are that (a) the landing site for the object NP is actually higher than that of gapping (VP-adjunction) or pseudogapping (Spec,AgrO) in English; (b) English- like gapping in Modern Mandarin also involves topicalization and focus movement; and (c) the relevant constituents that are deleted in English-like gapping in Modern Mandarin are not VPs, but IPs. These proposals will be explicated in the following sections.

5.2 The First NP

Since the target clause of English-like gapping in Modern Mandarin consists of only two NPs (e.g., (4)–(6) and (9)–(10)), the analysis of the construction depends on how to account for the derivation of these two NPs.7 I argue that the first NP has been topicalized (see also Wu 2002, 2003 for a similar topicalization proposal that involves not the first NP, but the whole event).

Modern Mandarin is well-known for its robust topic-comment structure. In the literature, it is commonly assumed that a topic in Modern Mandarin can be either base-generated in Spec, TopicP as in (21) if there is no gap associated with the topic, or moved into Spec,TopicP as in (22) if there is a gap associated with the topic. (See Xu and Langendoen 1985 for a comprehensive discussion of topic-comment structure in Modern Mandarin.)

(21)

  • Topic . . . comment: aboutness

  • [TopicP Na-jian shi, [IP duo-kui ni tixing wo]]!

  • that-CL thing many-thanks you remind me

  • ‘As for that thing, I was really thankful that you reminded me!’

(22)

  • Topicalization: Ā-movement

  • [TopicP Zhangsani, [IP wo bu xihuan [ _____ ]i]].

  • Zhangsan I not like

  • ‘As for Zhangsan, I do not like *(him).’

For English-like gapping in Modern Mandarin, the first NP is already at the left periphery. It seems that the topicalization movement, if any, is vacuous. Take the answer in (5), for example. The derivation is given in (23).

(23)

graphic

The gap in Spec,IP in (23) is associated with the topic. As mentioned above, as long as there is a gap in the ‘‘comment,’’ a movement analysis is commonly assumed for the topic-comment structure in Modern Mandarin (for relevant argumentation, see e.g. Xu and Langendoen 1985, Ning 1993). In this article, I will further assume the VP-internal subject hypothesis (e.g., Zagona 1982, Fukui and Speas 1986, Kitagawa 1986, Contreras 1987, Kuroda 1988, Sportiche 1988, Koopman and Sportiche 1991, Burton and Grimshaw 1992, McNally 1992). I have already assumed this in (7) without explicitly stating it. Under this hypothesis, the topic in a typical Englishlike gapping construction in Modern Mandarin is always derived through subject movement, as shown in (24).

(24)

graphic

As discussed in section 4, ellipsis typically requires ‘‘identical’’ linguistic antecedents. I will leave it open whether the identity is morphological, syntactic, or semantic (see Ai 2008 for detailed summaries of literature on these various arguments and further discussion of the identity condition for VP-ellipsis). On the surface, the topicalization process in the target clause in (23) clearly destroys parallelism, if not total ‘‘identity’’ between the antecedent clause and the target clause. The movement analysis for topicalization can now provide a derivational stage at which the licensing parallelism/identity condition between the antecedent clause and the target clause can be met. This stage is prior to the movement of the relevant topic in the target clause—that is, at either the vP level or the IP level, as shown in (24). It is the further movement of Lisi from Spec,IP to Spec,TopicP that destroys the parallelism/identity between the antecedent clause and the target clause.

5.3 The Second NP

I suggest that the second NP undergoes focus movement. The position to which it moves is between IP and TopicP (cf. Rizzi 1997). After the movement, the relevant IP is deleted at PF, giving rise to English-like gapping in Modern Mandarin. Consider (5) again, repeated here as (25). It receives the analysis shown in (26).

(25)

  • Q:

    Natian zai shan-shang, tamen dou kanjian-le shei?

    that.day on mountain-above they all see-ASP who

    ‘(Lit.) That day on the mountain, they all saw whom?’

  • A:

    (?)Zhangsan kanjian-le Shufen. Lisi _____ Yaping.

    Zhangsan see-ASP Shufen Lisi Yaping

    ‘Zhangsan saw Shufen and Lisi _____ Yaping.’

(26)

graphic
8

As shown in (26a), before Lisi is topicalized, the second NP Yaping undergoes focus movement to Spec,FocusP, which is above IP, but below TopicP. As shown in (26b), the remnant IP [tikanjian-le tj ] is then deleted at PF, giving rise to (25). Notice that under this analysis, the notation for gaps given in (25) for English-like gapping in Modern Mandarin is not quite accurate, as the gap is not really between the first NP Lisi and the second NP Yaping—rather, it is after the second NP. Thus, a more accurate notation of English-like gapping in Modern Mandarin would be as follows:

(27)

  • (?)Zhangsan kanjian-le Shufen. Lisi, Yaping [IP _____ ].

  • Zhangsan see-ASP Shufen Lisi Yaping

  • ‘Zhangsan saw Shufen and Lisi, Yaping.’

In (27), the gap no longer indicates a position left behind by (ATB) movement of relevant predicates or (VP-) ellipsis. Instead, it indicates an IP that has been elided. Thus, what seems to be a gap between the first NP and the second NP is actually a brief intonational pause separating the topicalized phrase from the rest of the utterance. This is the way a typical instance of topicalization is uttered, as in (21)–(22).

The gapping examples involving violation of CSs and the CNPC can be analyzed in a similar fashion. This is shown in (28)–(29) (= (9)–(10)). The violation of CSs and the CNPC is not an issue, as nothing has moved out of any islands.

(28)

  • Zhangsan kaojin-le Tsinghua,

  • Zhangsan exam.enter-ASP Tsinghua

  • ‘Zhangsan exam-entered Tsinghua University,

  • [woi, Beidaj [IP tt kaojin tj]] ziran mei wenti. (✓CSs)

  • I Beida exam.enter of.course no problem

  • [that I *(exam-entered) Beijing University] is of course no problem.’

(29)

  • Zhiyao Zhangsan yitian bu kaojin Tsinghua,

  • as.long.as Zhangsan one.day not exam.enter Tsinghua

  • ‘As long as Zhangsan has not exam-entered Tsinghua University,

  • ni jiu beng ti [[woi, Beidaj [IP tt kaojin tj]] de shi]. (✓CNPC)

  • you then not mention I Beida exam.enter DE fact

  • you will not mention the fact that I *(have exam-entered) Beijing University.’

In the next section, I will argue for the existence of a focus projection and focus movement in Modern Mandarin.

5.4 Evidence for a Focus Projection and Focus Movement in Modern Mandarin

The availability of a Focus position (i.e., above IP and below TopicP) can be supported first by the focus movement of wh-words in Modern Mandarin, a language that is otherwise argued to be wh-in-situ (Huang 1982). To illustrate, I provide an example with topicalization.

(30)

  • Zhangsan, ta baba mai-le shenme? (wh-in-situ)

  • Zhangsan his father buy-ASP what

  • ‘Zhangsan, what did his father buy?’

In (30), the wh-word shenme ‘what’ is in situ and the question is well-formed in Modern Mandarin. To achieve some kinds of emphatic effects, however, the wh-word can be fronted to a position after the topic.

(31)

  • Zhangsan, SHENME ta baba mai-le?9

  • Zhangsan what his father buy-ASP

  • ‘Zhangsan, what (on earth) did his father buy?’

The moved wh-word must be stressed, indicating that it has moved into a focus position (to check the relevant focus feature).

(32)

  • Focus movement of wh-words in Modern Mandarin

  • graphic
    10

Interestingly, such focus movement of wh-words can also occur in English-like gapping in Modern Mandarin, given an appropriate context. Compare (33) with (5).

(33)

  • Q:

    Natian zai shan-shang, tamen dou kanjian-le shei?

    that.day on mountain-above they all see-ASP who

    ‘(Lit.) That day on the mountain, they all saw whom?’

  • A:

    Zhangsan kanjian-le Shufen. Lisi, SHEI . . . (bu ji-de le).11

    Zhangsan see-ASP Shufen Lisi who not remember PART

    ‘Zhangsan saw Shufen. Who did Lisi see . . . (I could no longer remember).’

Given the possibility that wh-words in Modern Mandarin can undergo focus movement within a topic-comment structure as in (32), it is quite natural to propose that English-like gapping in Modern Mandarin, with the second NP being a wh-word as in (33), also involves focus movement. The only difference between (32) and the derivation of the relevant target clause in (33) is that after the focus movement of the relevant wh-word (34a), IP-deletion subsequently applies (34b).

(34)

graphic

The next piece of evidence for a focus projection and focus movement comes from the cleft construction in Modern Mandarin, alternatively known as the shi . . . de construction. The exact analysis of this construction has been controversial (for one analysis, see Simpson and Wu 2002). Nonetheless, it is generally agreed in the literature that shi ‘be’ is a focus marker and that typically, it cannot be used to focus-mark an object (e.g., Huang 1990, Paul and Whitman 2001, Simpson and Wu 2002).

(35)

  • *Zhangsan, ta da-le shi LISI de.

  • Zhangsan he beat-ASP SHI Lisi DE

  • ‘(Intended) Zhangsan, it is LISI that he has beaten.’

However, when the object Lisi is moved into a position after the topic, shi can be inserted and a cleft construction (shi . . . de construction) can be formed.

(36)

  • Zhangsan, shi LISI ta da-le de.

  • Zhangsan SHI Lisi he beat-ASPDE

  • ‘Zhangsan, it is LISI that he has beaten.’

This indicates that the object Lisi has moved into a focus position (between TopicP and IP), where shi serves as the direct focus marker. When (36) is uttered, Lisi must be stressed.

(37)

  • Shi . . . de construction (clefts in Modern Mandarin)

  • graphic

Intriguingly, this cleft construction can readily be applied to English-like gapping in Modern Mandarin. It works perfectly when the second NP is a wh-word, as in (38) (cf. (33)).

(38)

  • Q:

    Natian zai shan-shang, tamen dou kanjian-le shei?

    that.day on mountain-above they all see-ASP who

    ‘(Lit.) That day on the mountain, they all saw whom?’

  • A:

    Zhangsan kanjian-le Shufen. Lisi, shi SHEI . . . (bu ji-de le).

    Zhangsan see-ASP Shufen Lisi SHI who not remember PART

    ‘Zhangsan saw Shufen. It is who that Lisi saw . . . (I could no longer remember).’

If English-like gapping in Modern Mandarin is derived by IP-deletion as depicted in (26), it is expected that it can also be derived from the shi . . . de construction, as long as the relevant IP is deleted. This expectation is borne out if we assume that similar focus movement in the shi . . . de construction, as depicted in (37), feeds IP-deletion.

(39)

graphic

Assuming the analysis in (39), what happens in the relevant target clause of (38) is nothing more than the IP-deletion of a shi . . . de construction. The syntactic position for the focus in the shi . . . de construction and the focus in English-like gapping in Modern Mandarin is the same.

Given a certain context, the non-wh-word counterpart for the second NP in English-like gapping in Modern Mandarin can also occur in an elliptical shi . . . de construction (see Wu 2002, 2003 for more examples of this type).

(40)

  • Q:

    Tamen dou shangjiao-le shenme?

    they all turn.in-ASP what

    ‘(Lit.) They all turned in what?’

  • A:

    ?Zhangsan shangjiao-le yi-he yan. Lisi, shi yi-ba dao.12

    Zhangsan turn.in-ASP one-CL cigarette Lisi SHI one-CL knife

    ‘Zhangsan turned in a pack of cigarettes. Lisi, (it) was a knife.’

The analysis for the relevant target clause in (40) will be identical to that in (39), replacing whwords with a regular NP.

(41)

graphic

Given the current analysis, it is also expected that the second NP in the target clause of English-like gapping in Modern Mandarin will receive stress, as it ultimately moves into a syntactic focus position. This expectation is also borne out. Take (5), for example, repeated here.

(42)

  • Q:

    Natian zai shan-shang, tamen dou kanjian-le shei?

    that.day on mountain-above they all see-ASP who

    ‘That day on the mountain, they all saw whom?’

  • A:

    (?)Zhangsan kanjian-le Shufen. Lisi, Yaping.

    Zhangsan see-ASP Shufen Lisi Yaping

    ‘Zhangsan saw Shufen and Lisi, Yaping.’

The acceptability of (42) improves significantly if the second NP in the target clause is stressed and the whole answer is read in a topic-focus pattern.

(43)

  • Q:

    Natian zai shan-shang, tamen dou kanjian-le shei?

    that.day on mountain-above they all see-ASP who

    ‘That day on the mountain, they all saw whom?’

  • A:

    Zhangsan kanjian-le SHUFEN; Lisi, YAPING!

    Zhangsan see-ASP Shufen Lisi Yaping

    ‘Zhangsan saw Shufen and Lisi, Yaping!’

This indicates that the proposed syntactic analysis is also in line with the prosody of Modern Mandarin.

6 Related Issues

In this section, I will address issues related to the current analysis of English-like gapping in Modern Mandarin. In particular, I will discuss the focus marker lian ‘even’ in Modern Mandarin, the scope of negation/modality over coordination, and the direction of NP-movement in Modern Mandarin.

6.1 The Focus Marker Lian

Besides shi ‘be’, Modern Mandarin has another focus marker, lian ‘even’ (see Badan 2007 for a more comprehensive discussion of lian-related constructions). Typically, unlike shi, lian cannot occur in English-like gapping in Modern Mandarin.

(44)

  • Zhangsan zhua-dao san-tiao yu. *Lisi, lian yi-zhi qingwa.

  • Zhangsan catch-reach three-CL fish Lisi LIAN one-CL frog

  • ‘(Intended) Zhangsan caught three fish. Lisi did not even catch one frog.’

This is because, unlike shi, which is typically treated as an adverbial modal (see, e.g., Huang 1990), lian in Modern Mandarin behaves more like either a negative polarity item (NPI) or a positive polarity item ( PPI). As a polarity item, it cannot occur by itself.

(45)

  • a.

    Lisi, lian yi-zhi qingwa *(dou) *(mei) zhua-dao. (NPI)

    Lisi LIAN one-CL frog all not catch-reach

    ‘Lisi did not catch even a frog.’

  • b.

    Lisi, lian she *(dou) (gan) chi. (PPI)

    Lisi LIAN snake all dare eat

    ‘Lisi eats/dares to eat even a snake.’

In (45a), lian must be overtly collocated with (licensed by) the quantifier dou ‘all’ and the negator mei ‘not’ as an NPI, and in (45b), with the quantifier dou and the optional modal gan ‘dare’ as a PPI. In other words, the quantifier dou must be present in both cases; without it, both (45a) and (45b) will be ungrammatical. Because of the idiosyncratic behavior of lian—that is, its being a polarity-sensitive item—it is unsurprising that it cannot occur in English-like gapping in Modern Mandarin (44), as its overtly required licenser, if any, has been deleted.

(46)

graphic

6.2 The Scope of Negation/Modality

The coordination of conjuncts has been argued to occur at the vP/VP level for canonical gapping in English, as shown in (8) (see, e.g., Coppock 2001, Johnson 2009, for extensive discussion). The crucial evidence for this argument comes from examples with modals and negation.

(47) John can’t eat caviar and Bill _____ beans.

It seems that the modal and negation embedded in the first conjunct can take scope over the coordination—that is, can take scope into the second conjunct. The whole utterance in (47) can only mean ‘It is not possible that John eats caviar while Bill eats beans’, indicating that coordination occurs at the vP/VP level, that is, below modals and negation. In discussing English-like gapping in Modern Mandarin, I have adopted a methodology that separates the target clause from the antecedent clause (section 5.1). Essentially, this means that the target clause and the antecedent clause are two independent IPs. Given that a topic-comment structure is also involved in Englishlike gapping in Modern Mandarin, the coordination, if any, must happen at the CP level. This predicts that modals and negation embedded in the antecedent clause cannot take scope over the target clause. This prediction is borne out.

(48)

  • Zhangsan, ni bu-neng chi zhe-wan fan. Lisi, NEI-wan FAN!13

  • Zhangsan you not-can eat this-CL rice Lisi that-CL rice

  • ‘(Lit.) Zhangsan, you cannot eat this bowl of rice. Lisi, that bowl of rice!’

What (48) literally means, given an appropriate context (e.g., Wangwu uttered (48) when he was tired of pointing to various dishes and telling each guest who should eat what at a banquet), is ‘Zhangsan, you cannot eat this bowl of rice. Lisi, you eat that bowl of rice!’ It is clear that the scope of negation/modality initiated in the antecedent clause is maintained within the antecedent clause. The target clause has not been affected.14

6.3 The Direction of NP-Movement

Following Lasnik (1995, 1999a,b), I have adopted leftward focus movement of the second NP for English-like gapping in Modern Mandarin. The direction of NP-movement could also be rightward, as in the traditional analysis of gapping in English (7), repeated here.

(49)

graphic

The differences between the current analysis of English-like gapping in Modern Mandarin and the traditional analysis of canonical gapping in English are threefold. (a) While the first NP in Modern Mandarin is topicalized into the CP domain and is thus an instance of Ā-movement, the first NP in canonical English gapping moves from Spec,vP to Spec,TP and is thus an instance of A-movement. (b) While the second NP in Modern Mandarin moves leftward into Spec,FocusP, the second NP in canonical English gapping moves rightward and adjoins to VP. (c) While English-like gapping in Modern Mandarin deletes the remnant IP, canonical gapping in English deletes a remnant VP (49).

Suppose there is a scenario in which the first NP undergoes leftward Ā-movement into the CP domain and at the same time the second NP moves rightward and lands slightly above IP. This is shown schematically in (50).

(50)

graphic

If IP-deletion follows, the analysis in (50) also derives English-like gapping in Modern Mandarin. The first NP, for example, is topicalized into the CP domain and the second NP undergoes rightward focus movement and adjoins to IP (followed by IP-deletion). In fact, Lasnik (2011) has explored exactly this analysis to account for multiple sluicing in English.

(51)

  • I know that in each instance one of the girls got something from one of the boys. But they didn’t tell me which from which.

  • (Lasnik 2011:5, (20))

There are two wh-related words in (51): which and from which. Lasnik suggests that while which undergoes normal wh-movement into the CP domain, from which undergoes rightward focus movement and adjoins to a position slightly higher than IP, which is then deleted as shown in (50), giving rise to multiple sluicing in English. The main argument for the rightward focus movement is that English allows heavy NP/PP shift.

(52)

Normally, PP shift is more acceptable than NP shift.

(53)

graphic

That explains why in the best examples of multiple sluicing in English, the second wh-related word is always contained in a PP (51) (e.g., Richards 1997, 2001). As far as English-like gapping in Modern Mandarin is concerned, if the analysis in (50) is applicable at all, it must be the case that Modern Mandarin also allows heavy NP/PP shift so that such a right-adjoined focus position can be empirically attested. It is clear that heavy PP shift is not allowed in Modern Mandarin.

(54)

  • a.

    Zhangsan zuotian gen ta pengyou shuohua le.

    Zhangsan yesterday with his friend speak PART

    ‘Zhangsan spoke to his friend yesterday.’

  • b.

    *Zhangsan zuotian shuohua le gen ta pengyou.15

    Zhangsan yesterday speak PART with his friend

As the adverbial (e.g., zuotian ‘yesterday’) precedes the verb in Modern Mandarin, as shown in (54), it is hard to construct examples with heavy NP shift akin to those in (52a) and (53b). The only possible relevant constructions are certain cases of object scrambling like those discussed by Soh (1998).

(55)

  • a.

    Zhangsan zuotian kanjian-le ta-de lao pengyou Lisi san-ci.

    Zhangsan yesterday see-ASP his-POSS old friend Lisi three-CL

    ‘Zhangsan saw his old friend Lisi three times yesterday.’

  • b.

    Zhangsan zuotian kanjian-le san-ci ta-de lao pengyou Lisi.

    Zhangsan yesterday see-ASP three-CL his-POSS old friend Lisi

Soh argues extensively that the scrambled object occupies a position within the relevant VP. It cannot move as high as IP. In this case, a rightward-adjoined focus position that is above IP still cannot be attested in Modern Mandarin. The leftward-adjoined focus position that is above IP, however, has been argued to exist in Modern Mandarin, as in the focus movement of wh-words and the shi . . . de construction (section 5.4).

Besides heavy NP/PP shift, Modern Mandarin also seems to be missing one consistent property of rightward movement. As Lasnik (2011) discusses, the rightward movement in English multiple sluicing is strictly clause-bound, even if no island is crossed.

(56)

  • Some students said that John was talking to some policeman . . .

  • # . . . But I don’t know which students to which policeman.

This is unacceptable because it would require the second wh-word to cross a clause boundary to escape the ellipsis site. English-like gapping in Modern Mandarin, on the other hand, seems to be immune to this requirement, with (57) or without (58) the focus marker shi (cf. (43) and (40), respectively).

(57)

  • Q:

    Wangwu shuo natian zai shan-shang, tamen dou kanjian-le shei?

    Wangwu say that.day on mountain-above they all see-ASP who

    ‘(Lit.) Wangwu said (that) that day on the mountain, they all saw whom?’

  • A:

    (?)Zhangsan, Wangwu shuo kanjian-le SHUFEN; Lisi, YAPING!16

    Zhangsan Wangwu say see-ASP Shufen Lisi Yaping

    ‘(Lit.) Zhangsani, Wangwu said that (hei) saw Shufen and Lisi, Yaping!’

(58)

  • Q:

    Wangwu shuo tamen dou shangjiao-le shenme?

    Wangwu say they all turn.in-ASP what

    ‘(Lit.) Wangwu said (that) they all turned in what?’

  • A:

    ?Zhangsan, Wangwu shuo shangjiao-le yi-he yan. Lisi, shi yi-ba dao.

    Zhangsan Wangwu say turn.in-ASP one-CL cigarette Lisi SHI one-CL knife

    ‘Zhangsani, Wangwu said that (hei) turned in a pack of cigarettes. Lisi, (it) was a knife.’

What the target clause in (57) means is that ‘As far as Lisi is concerned, Wangwu said that he (Lisi) saw Yaping’. Similarly, the target clause in (58) means that ‘As far as Lisi is concerned, Wangwu said that he (Lisi) turned in a knife’. The second NP (as well as the first NP—that is, the topic) has obviously crossed the (embedded) clause boundary to escape the ellipsis site.

It is for all the above reasons that I have adopted a leftward focus movement for the second NP in English-like gapping for Modern Mandarin.

7 Concluding Remarks

In this article, I have analyzed English-like gapping in Modern Mandarin as multiple sentence fragments that are constructed by a series of syntactic operations. These include (a) topicalization of the first NP, (b) focus movement of the second NP, and (c) subsequent deletion of the relevant IP. Thus, instead of being produced by (ATB) movement of predicates or the (VP-) ellipsis that creates canonical gapping in English, the gap in English-like gapping in Modern Mandarin is produced by more complex syntactic operations. One consequence of this approach is that it leaves the formation of English-like gapping in Modern Mandarin on a par with the derivation of sluicing (Merchant 2001) or stripping/sentence fragments in English (Merchant 2004, 2005, 2006), if the topicalization process is disregarded.

(59)

graphic

As (59) indicates, sluicing, stripping/sentence fragments, and English-like gapping in Modern Mandarin all involveĀ-movement of certain kinds of constituents (i.e., interrogative/focused whwords in sluicing and focused NPs in stripping/sentence fragments and English-like gapping in Modern Mandarin) to the CP domain. All of these movements are followed by the deletion of the relevant IP. The only difference among these constructions is the number of constituents that can be moved to the CP domain. While sluicing (excluding multiple sluicing) and stripping/ sentence fragments typically move only one constituent to the CP domain (59a–b), Englishlike gapping in Modern Mandarin moves two constituents (59c). Because of this additional NPmovement (topicalization), English-like gapping in Modern Mandarin is best described as ‘‘multiple sentence fragments.’’

Notes

Thanks are due to Cedric Boeckx, Lisa Cheng, Jeremy Hartman, C.-T. James Huang, Kyle Johnson, Andrew Nevins, and two anonymous LI reviewers for their insightful comments and suggestions. I am solely responsible for errors.

1 Most English translations in this article are literal. English wh-in-situ questions specifically disallow the pair-list reading.

2 There are some other pieces of evidence, such as scope of negation/modality. For brevity, I will only discuss two of the arguments that Johnson has made. The scope of negation/modality over coordination will be discussed in section 6.

3Tang (2001) claims that examples like (2)–(3) are ungrammatical. Also, Tang does not discuss examples like (9)–(10).

4Tang (2001) attributes the acceptability of examples like (13) to some nonlinguistic factors in the end (see his endnote 9 for more discussion). To me, (13) is as natural as (2) and (3), given certain linguistic or pragmatic context.

5 ‘‘Identity,’’ which is a requirement on ellipsis, should not be a concern here as these are so-called empty-verb sentences (whose content can be variably reconstructed on the basis of pragmatic context). The question should not be treated as a linguistic antecedent for the answers. It simply provides a context in which the empty-verb sentences are used. The free alternation of chi ‘to eat’ and he ‘to drink’ in some dialects may also suggest that identity does not need to be exact in empty-verb sentences.

6Tang (2001) and many others attribute the nonoccurrence of examples like (18) to either semantics or pragmatics.

7 I assume NPs instead of DPs in this article. The distinction between NP and DP is irrelevant to my argumentation.

8 For brevity, the vP/VP-internal trace for Lisi is not marked.

9 There seem to be two focus positions in Modern Mandarin. (31) illustrates one of them. The wh-word can also be fronted to sentence-initial position, crossing the topic.

(i)

  • (?)SHENME, Zhangsan, ta baba mai-le?

  • what Zhangsan his father buy-ASP

  • ‘(Lit.) What (on earth) did Zhangsan, his father buy?’

This position is not relevant to the discussion in this article.

10 For brevity, the topicalization process for Zhangsan is not marked. The resumptive pronoun ta indicates the position from which Zhangsan is moved. In other words, it is an ‘‘overt’’ copy of the topic, Zhangsan.

11 The phrase bu ji-de le ‘no longer remember’ adds a natural context for the relevant utterance (e.g., talking to a friend nearby). Without it, the remaining utterance is also perfectly fine (e.g., in the context of a person talking to himor herself ). The same is true for (38). Strictly speaking, the focus movement of wh-words followed by IP-deletion as in (34) is ‘‘embedded’’ in the current analysis of ‘‘gapping’’ in Mandarin. It does not argue for it.

12 The grammaticality judgment for (40) can vary from person to person. This can be partially attributed to the violation of parallelism between the antecedent clause, which does not utilize the shi . . . de construction, and the elliptical target clause, which does. The judgment improves significantly if a relatively longer pause is imposed after Lisi, indicating that the syntax of the target clause is working independently from the syntax of the antecedent clause.

13 To illustrate a possible CP conjunction, I insert a resumptive pronoun ni ‘you’ into the antecedent clause to make it a topic-comment structure that parallels the target clause.

14 With the intervention of modals and negation, it seems that the parallelism/identity condition needs to be checked even ‘‘earlier’’—for example, at the vP level. See section 5.2 for the possibility that it is checked at either the vP or the IP level. Further research with empirical evidence is needed to support this hypothesis.

15 The ungrammaticality of (54b) is based on the reading where it is treated as one simple sentence. If a full stop is added after the particle le, changing the whole utterance into a discourse with two sentences, it becomes grammatical.

(i) Zhangsan zuotian shuohua le. Gen ta pengyou.

Zhangsan yesterday speak PART with his friend

In this case, the second sentence (actually a PP phrase), Gen ta pengyou (‘With his friend’), contributes additional information to the first.

16 To simplify the matter, the antecedent clause is also topicalized to maintain parallelism with the target clause in both (57) and (58). For discussion of the grammaticality judgments for (57) and (58), see section 5.4 and footnote 12.

References

References
Ai
,
Ruixi Ressy
.
2005
.
Gapping in Chinese: A functional approach
. In
Harvard working papers in linguistics 11
, ed. by
Jeffrey
Bourns
,
Andrew
Nevins
, and
Yaroslav
Gorbachov
,
1
33
.
Cambridge, MA
:
Harvard University, Department of Linguistics
.
Ai
,
Ruixi Ressy
.
2008
.
Elliptical predicate constructions in Mandarin
.
Munich
:
LINCOM
.
Badan
,
Linda
.
2007
.
High and low periphery: A comparison between Italian and Chinese
.
Doctoral dissertation, Universita` degli Studi di Padova
.
Baltin
,
Mark
.
2003
.
The interaction of ellipsis and binding: Implications for the sequencing of Principle A
.
Natural Language and Linguistic Theory
21
:
215
246
.
Burton
,
Strang
, and
Jane
Grimshaw
.
1992
.
Coordination and VP-internal subjects
.
Linguistic Inquiry
23
:
305
313
.
Contreras
,
Heles
.
1987
.
Small clauses in Spanish and English
.
Natural Language and Linguistic Theory
5
:
225
244
.
Coppock
,
Elizabeth
.
2001
.
Gapping: In defense of deletion
. In
CLS 37-1: The Main Session
, ed. by
Mary
Andronis
,
Christopher
Ball
,
Heidi
Elston
, and
Sylvain
Neuvel
,
133
148
.
Chicago
:
University of Chicago, Chicago Linguistic Society
.
Donati
,
Caterina
.
2003
.
Merge copy
. In
The interfaces: Deriving and interpreting omitted structures
, ed. by
Kerstin
Schwabe
and
Susanne
Winkler
,
155
175
.
Amsterdam
:
John Benjamins
.
Fukui
,
Naoki
, and
Margaret
Speas
.
1986
.
Specifiers and projection
. In
Papers in theoretical linguistics
, ed. by
Naoki
Fukui
,
Tova R.
Rapoport
, and
Elizabeth
Sagey
,
128
172
.
MIT Working Papers in Linguistics 8. Cambridge, MA
:
MIT, MIT Working Papers in Linguistics
.
Huang
,
C.-T. James
.
1982
.
Logical relations in Chinese and the theory of grammar
.
Doctoral dissertation, MIT, Cambridge, MA. Published, New York: Garland (1998)
.
Huang
,
C.-T. James
.
1990
.
On ‘be’ and ‘have’ in Chinese
.
Bulletin of the Institute of History and Philology, Academia Sinica
59
:
43
64
.
Jayaseelan
,
K. A
.
1990
.
Incomplete VP deletion and gapping
.
Linguistic Analysis
20
:
64
81
.
Johnson
,
Kyle
.
1994
.
Bridging the gap
.
Ms., University of Massachusetts, Amherst
.
Johnson
,
Kyle
.
2000
.
How far will quantifiers go?
In
Step by step
, ed. by
Roger
Martin
,
David
Michaels
, and
Juan
Uriagereka
,
187
210
.
Cambridge, MA
:
MIT Press
.
Johnson
,
Kyle
.
2004
.
In search of the English middle field
.
Ms., University of Massachusetts, Amherst
.
Johnson
,
Kyle
.
2006
.
Too many an example is thought to be ellipsis, and too few, across-the-board movement
.
Invited talk given at MIT Colloquium, 17 February
.
Johnson
,
Kyle
.
2009
.
Gapping is not (VP-) ellipsis
.
Linguistic Inquiry
40
:
289
328
.
Kitagawa
,
Yoshihisa
.
1986
.
Subjects in Japanese and English
.
Doctoral dissertation, University of Massachusetts, Amherst
.
Koopman
,
Hilda
, and
Dominique
Sportiche
.
1991
.
The position of subjects
.
Lingua
85
:
211
258
.
Kuroda
,
S.-Y
.
1988
.
Whether we agree or not: A comparative syntax of English and Japanese
. In
Papers from the Second International Workshop on Japanese Syntax
, ed. by
William
Poser
,
103
143
.
Stanford, CA
:
CSLI Publications
.
Lasnik
,
Howard
.
1995
.
A note on pseudogapping
. In
Papers on Minimalist syntax
, ed. by
Rob
Pensalfini
and
Hiroyuki
Ura
,
143
163
.
MIT Working Papers in Linguistics 27
.
Cambridge, MA
:
MIT, MIT Working Papers in Linguistics
.
Lasnik
,
Howard
.
1999a
.
On feature strength: Three Minimalist approaches to overt movement
.
Linguistic Inquiry
30
:
197
217
.
Lasnik
,
Howard
.
1999b
.
Pseudo-gapping puzzles
. In
Fragments: Studies in ellipsis and gapping
, ed. by
Shalom
Lappin
and
Elabbas
Benmamoun
,
141
174
.
Oxford
:
Oxford University Press
.
Lasnik
,
Howard
.
2011
.
Multiple sluicing in English?
Ms., University of Maryland, College Park
.
Li
,
Mei-Du
.
1988
.
Anaphoric structures of Chinese
.
Taipei
:
Student Book Co
.
Li
,
Yen-hui Audrey
.
1998
.
Argument determiner phrases and number phrases
.
Linguistic Inquiry
29
:
693
702
.
Lin
,
I.-W. Vivian
.
2002
.
Coordination and sharing at the interfaces
.
Doctoral dissertation, MIT, Cambridge, MA
.
McNally
,
Louise
.
1992
.
VP coordination and the VP-internal subject hypothesis
.
Linguistic Inquiry
23
:
336
341
.
Merchant
,
Jason
.
2001
.
The syntax of silence: Sluicing, islands, and identity in ellipsis
.
Oxford
:
Oxford University Press
.
Merchant
,
Jason
.
2004
.
Fragments and ellipsis
.
Linguistics and Philosophy
27
:
661
738
.
Merchant
,
Jason
.
2005
.
Ellipsis and nonsentential speech
. In
Selected papers on theoretical and applied linguistics
, ed. by
Marina
Mattheoudakis
and
Angeliki
Psaltou-Joycey
,
175
189
.
Thessaloniki
:
Aristotle University of Thessaloniki
.
Merchant
,
Jason
.
2006
.
‘Small structures’: A sententialist perspective
. In
The syntax of nonsententials: Multidisciplinary perspectives
, ed. by
Ljiljana
Progovac
,
Kate
Paesani
,
Eugenia
Casielles
, and
Ellen
Barton
,
73
91
.
Amsterdam
:
John Benjamins
.
Ning
,
Chunyan
.
1993
.
The overt syntax of relativization and topicalization in Chinese
.
Doctoral dissertation, University of California, Irvine
.
Paul
,
Waltraud
.
1996
.
Verb raising in Mandarin Chinese
. In
Studies in Chinese linguistics
, ed. by
Chinchuan
Cheng
,
Jerome L.
Packard
, and
James Y.
Yoon
,
1:255
270
.
Studies in the Linguistic Sciences 26. Urbana-Champaign
:
University of Illinois, Department of Linguistics
.
Paul
,
Waltraud
.
1997
.
Verb raising in Chinese
. In
Proceedings of the Eighth North America Conference on Chinese Linguistics
, ed. by
Chin-chuan
Cheng
,
Jerome L.
Packard
,
James Y.
Yoon
, and
Y.-L.
Yoon
,
1:260
276
.
Los Angeles
:
University of Southern California, Graduate Students in Linguistics
.
Paul
,
Waltraud
.
1999
.
Verb gapping in Chinese: A case of verb raising
.
Lingua
107
:
207
226
.
Paul
,
Waltraud
, and
John
Whitman
.
2001
.
The complement structure of shi . . . de construction
.
Paper presented at the 10th conference of the International Association of Chinese Linguistics in conjunction with North America Conference on Chinese Linguistics 13, University of California, Irvine
.
Pesetsky
,
David
.
1982
.
Paths and categories
.
Doctoral dissertation, MIT, Cambridge, MA
.
Richards
,
Norvin
.
1997
.
What moves where when in which language
.
Doctoral dissertation, MIT, Cambridge, MA
.
Richards
,
Norvin
.
2001
.
Movement in language
.
Oxford
:
Oxford University Press
.
Rizzi
,
Luigi
.
1997
.
The fine structure of the left periphery
. In
Elements of grammar
, ed. by
Liliane
Haegeman
,
281
337
.
Dordrecht
:
Kluwer
.
Sag
,
Ivan
.
1976
.
Deletion and Logical Form
.
Doctoral dissertation, MIT, Cambridge, MA
.
Simpson
,
Andrew
, and
Zoe
Wu
.
2002
.
From D to T: Determiner incorporation and the creation of tense
.
Journal of East Asian Linguistics
11
:
169
209
.
Soh
,
Hooi Ling
.
1998
.
Object scrambling in Chinese
.
Doctoral dissertation, MIT, Cambridge, MA
.
Sportiche
,
Dominique
.
1988
.
A theory of floating quantifiers and its corollaries for constituent structure
.
Linguistic Inquiry
19
:
425
450
.
Tai
,
H. J
.
1969
.
Coordination reduction
.
Doctoral dissertation, Indiana University, Bloomington
.
Takahashi
,
Shoichi
.
2004
.
Pseudogapping and cyclic linearization
. In
NELS 34
, ed. by
Keir
Moulton
and
Matthew
Wolf
,
571
585
.
Amherst
:
University of Massachusetts, Graduate Linguistic Student Association
.
Tang
,
Sze-Wing
.
2001
.
The (non-) existence of gapping in Chinese and its implications for the theory of gapping
.
Journal of East Asian Linguistics
10
:
201
224
.
Tsai
,
W.-T. Dylan
.
1994
.
On economizing the theory of A-bar dependencies
.
Doctoral dissertation, MIT, Cambridge, MA
.
Wu
,
H.-H. Iris
.
2002
.
Gapping in classical and modern Chinese
.
Ms., National Tsinghua University
.
Wu
,
H.-H. Iris
.
2003
.
On ellipsis and gapping in Mandarin Chinese
.
Master’s thesis, National Tsinghua University
.
Xu
,
Liejiong
, and
D. Terence
Langendoen
.
1985
.
Topic structure in Chinese
.
Language
61
:
1
27
.
Zagona
,
Karen
.
1982
.
Government and proper government of verbal projections
.
Doctoral dissertation, University of Washington, Seattle
.