Island repair effects of the Left Branch Condition in Mandarin Chinese

Abstract

This study employs the island repair effect on the Left Branch Condition (LBC) to illuminate the derivation of Mandarin sluicing. It utilizes three unique properties of Mandarin island repair related to the LBC involving (i) covert antecedents, (ii) contrastive modifiers, and (iii) multiple islands including LBC structures in order to examine two deletion-based analyses of sluicing in the literature. It is shown that these analyses fail to satisfactorily explain the properties discussed. To capture the facts, a pseudosluicing analysis is proposed which claims that sluiced clauses in Mandarin are simply composed of a subject pro, an (optional) copula shi 'be', and a wh-in-situ wh-remnant serving as a predicate. The strong redemptive ability of repairing LBC effects in Mandarin is attributed to the construal of pro instead of deletion. From a typological point of view, among East Asian languages, Mandarin sluices differ from Japanese and Korean sluices in that the pro of the former cannot be interpreted as a concealed cleft structure but instead functions as an implicit subject.(iii) multiple islands involving the LBC. The paper evaluates two deletion-based analyses, the Focus condition analysis (Merchant 2001) and the multi-movement and deletion analysis (Wang 2007), and demonstrates that these analyses cannot satisfactorily explain the three concerned discrepancies and other fundamental properties in Mandarin sluicing, such as syntactic parallelism under deletion and the distribution of the copula shi 'be'. To capture these properties and patterns, a pseudosluicing analysis of sluicing in Mandarin (Wei 2004) is proposed, which explains the redemptive effect of the LBC in Mandarin via the construal of pro rather than deletion. In this sense, Mandarin sluicing is similar to pseudosluicing in English, is a base-generated structure consisting in a simple clause [pro (shi) wh-remnant], and is different from English deletion-based canonical sluicing. Additionally, it is claimed that Mandarin sluices are not concealed cleft structures resulting from argument ellipsis, unlike Japanese and Korean sluices (Nishiyama et al. 1996;Saito 2004Saito , 2007 Takahashi 2008a, b). It is proposed that the pro in Mandarin sluices can be reinterpreted as an E-type pro (Evans 1980), which explains the generation of sloppy identity readings. Quite generally, the analysis makes use of two well-established properties of Mandarin Chinese to explain its sluicing construction: the pro-drop nature of Chinese and topic-comment predication, and the fact that no new ad hoc stipulations are required in the analysis.The organization of this paper is as follows. Section 2 compares island repair effects in LBC structures in English with Mandarin Chinese. Section 3 reviews two potential deletion-centered accounts of sluicing and argues that they are not appropriate for Chinese. Section 4 then proposes the pseudosluicing analysis. Section 5 discusses the origin of the strong island/LBC repair effect in Mandarin Chinese. Section 6 further clarifies the ...

Sluicing, Sprouting and Missing Objects

“…The optionality of shi is claimed to be captured indirectly. Later, Adams and Tomioka (2012) concur with Wei's (2004Wei's ( , 2011 In addition, as argued by Wei (2011), the construction in question is more like English pseudo-sluicing than sluicing in allowing a strongly non-D-linked whphrase as in (8-9) and allowing a wh-phrase that is related to a noun phrase in a left-branch position as in (10)(11) Zhangsan buy le one-cl car but I not know be how big de 'Zhangsan bought a car, but I don't know how big. '…”

“…Nonetheless, it has also been noted that the two constructions in the languages are not the same (cf. Adams 2004;Adams and Tomioka 2012;Wei 2004Wei , 2011Li and Wei 2014, among others). One important difference concerns the occurrence of the morpheme shi 'be', which is the copular verb in Chinese and is often used to mark the constituent following it as focused.…”

“…On the other hand, Adams (2004), Wei (2004Wei ( , 2011, Adams and Tomioka (2012), and Li and Wei (2014), among others, argue that apparent sluicing cases in Chinese are not sluicing and therefore should be analyzed as a base-generated pseudo-sluicing structure. Movement and deletion do not apply.…”

“…3 Wei (2004Wei ( , 2011 argues that shei 'who' and shenme 'what' cannot function as predicates, in contrast to other whphrases, resulting in different requirements on the occurrence of the copula shi. The requirement on the occurrence of shi in the relevant patterns is determined by the predicate status of the wh-phrase.…”

“…One involves cases of wh-phrases that are not derivable via fronting to the clausal peripheral position. For instance, Wei (2004) notes that the sluicing analysis would not be able to derive certain constructions, such as the ones containing dao-nali 'to where' in a PP [cong…] 3 Adams (2004) and Wei (2004Wei ( , 2011 have independently proposed similar analyses, according to which a Chinese sluice clause is composed of three essential elements: pro, copula, and whremnant. The main difference between the two analyses lies in the generation of shi.…”