1. Introduction
In Mandarin Chinese, pronouns can precede [Num-Cl-NP] to form a complex nominal expression [P-Num-Cl-NP]. The sequence of [P-Num-Cl-NP] can further fall into two types of constructions. For one case, the pronoun semantically denotes the possessor of the referent(s) expressed by the [Num-Cl-NP], giving rise to a typical possessive construction. For the other case, the pronoun and the [Num-Cl-NP] are associated with the same referent(s) in the context, yielding the so-called appositive construction, with the pronoun as the anchor and the [Num-Cl-NP] as the apposition (Chao Reference Chao1968; J. Li Reference Li2013; X. Tang Reference Tang2013; Xu Reference Xu2013; W. Li Reference Li2016; Li & Chen Reference Li and Yi2016; Liu Reference Liu2016). A test that can straightforwardly distinguish the appositive case from the possessive one is the insertion of the possessive marker de: whereas de is allowed to be inserted in between the P and [Num-Cl-NP] in the possessive [P-Num-Cl-NP], this operation is prohibited from the appositive [P-Num-Cl-NP], as illustrated below:Footnote 2
Among various types of nominal phrases in Mandarin Chinese, the appositive construction has long been relatively under-researched. While it has been well noted that the [P-Num-Cl-NP] construction as shown above can be ambiguous between possessives and appositives (e.g. X. Tang Reference Tang2013), what has passed unnoticed is the discrepancies between different appositive [P-Num-Cl-NP] expressions. To fill the research niche, this paper will conduct a closer scrutiny of this construction and argue in favor of a two-way subcategorization, based on which a novel dichotomous syntactic analysis will be developed for Chinese appositives. The paper is organized as follows. Section 2 gives a critical review on the previous syntactic analyses of Chinese appositives. Section 3 examines the asymmetries exhibited by different appositive [P-Num-Cl-NP] expressions in terms of semantic and pragmatic properties, in view of which an attributional versus identificational dichotomy is proposed. Section 4 develops a dichotomous formal account for the appositive construction. Section 5 concludes the paper.
2. Previous syntactic analyses of Chinese appositives
In the limited number of studies on Chinese appositives, while some endeavors have been made to describe relevant linguistic phenomena (see Liu Reference Liu2016 and the references therein), theoretical syntactic investigations are still far from conclusive. This section will review three representative syntactic analyses proposed in the prior research.
Along the first approach, as pursued in Del Gobbo (Reference Del Gobbo1999) and Huang, Li & Li (Reference Huang, Li and Li2009), given an appositive construction in the form of [P-Num-Cl-NP], the P syntactically occupies the D head, and the [Num-Cl-NP] sequence is taken by D as its complement, as in (3). When a demonstrative intervenes to give rise to a more complex appositive [P-Dem(onstrative)-Num-Cl-NP], the P and the Dem, in accordance with Huang et al. (Reference Huang, Li and Li2009), form a complex D head that takes [Num-Cl-NP] as its complement, as in (4):
Along the second approach, as proposed by S. Tang (Reference Tang2010), appositives are subsumed under the coordination construction, and the anchor and the apposition are analyzed as conjuncts of the coordination construction. At the structural level, the anchor P and the apposition [Num-Cl-NP] are assumed to be originated as the specifier and the complement of a null-headed CoP, respectively, as given in (5):
Along the third approach, as proposed by Hong & Shi (Reference Hong and Dingxu2012), given an appositive construction like [P-(Dem)-Num-Cl-NP], the P is generated at [Spec, DP], and the [Num-Cl-NP] sequence is taken by D as its complement; the intervening Dem, if occurring, is located under D, as shown in (6):
As outlined above, all of the three approaches have attempted a unified syntactic account for Chinese appositives. An expectation implied by this line is that all appositive [P-Num-Cl-NP] expressions in Mandarin Chinese should syntactically pattern alike. This expectation is, nevertheless, failed by an intriguing observation that different appositive [P-Num-Cl-NP] expressions may exhibit nonparallel syntactic behaviors, a situation in turn closely correlated with the singular versus plural nature (subscripted as ‘sg’ and ‘pl’ hereafter) of the pronoun and the numeral involved. To illustrate, first, while [Ppl-Numpl-Cl-NP] generally permits nominal ellipsis, this is not always allowed for [Psg-Numsg-Cl-NP], as shown by the contrast between (7) and (8):
Second, while the NP in [Ppl-Numpl-Cl-NP] can always undergo topicalization, this is not the case for the NP in [Psg-Numsg-Cl-NP], as shown by (9) versus (10):
To the best of my knowledge, James Huang is the first to note, in his personal communications with Del Gobbo (Reference Del Gobbo1999: fn. 5), that there are differences between plural and singular nominal appositions. Nevertheless, so far little research has been dedicated to this issue. The present study is aimed to fill this research niche by presenting new empirical evidence to argue for the necessity to structurally disambiguate two types of appositive [P-Num-Cl-NP] in Mandarin Chinese.
3. Semantic and pragmatic asymmetries of the appositive [P-Num-Cl-NP]
3.1 Semantic asymmetries
When approaching the use of the appositive [P-Num-Cl-NP], we observe an important distinction between [Ppl-Numpl-Cl-NP] and [Psg-Numsg-Cl-NP] in terms of the semantics of the NP allowed to be used. To be concrete, the nominal occurring in the appositive [Psg-Numsg-Cl-NP] is normally one that expresses an identity regarding which certain contextually or conventionally held evaluations/judgments can be well implied in the current utterance, whereas the nominal in [Ppl-Numpl-Cl-NP] is not subject to this condition.
To begin with, note that the nominals used in the appositive [Psg-Numsg-Cl-NP] typically include (but are not limited to) those related to (i) profession/occupation, (ii) social rank/status, (iii) gender, etc., as exemplified in (11)–(13).Footnote 11
To elaborate on the evaluations/judgments implied by the nominal involved, for instance, in the context of (11a), the occupation zuojia ‘writer’ can be associated with an unexpressed common understanding that writers usually are more concerned about art than about money. That is why the speaker argues that even ‘his’ job is a writer, making money is still very important given that ‘he’ needs to make a living. In (12b), kuotaitai ‘rich wife’ may imply a series of characteristics (often highly biased though) conventionally labeled to this group of people, such as, e.g. being materialistic, living a luxury life, etc. It is due to these implicit evaluations that the speaker feels surprised that ‘she’, a rich wife, would live in a suite rather than in a villa. Likewise, nühaijia ‘girl’ in (13b) implies some conservative judgments conventionally held about girls, such as being physically weak, vulnerable, etc. That is why the speaker uses the adverb dao ‘but’ in the predicate: that ‘you’, a girl, can fix something in such a windy and rainy day is totally contrary to the speaker’s expectation.
That the nominal used in the appositive [Psg-Numsg-Cl-NP] is required to be one able to well imply contextually or conventionally held evaluations/judgments about the associated identity can be best demonstrated by the discrepancy between [Psg-Numsg-Cl-NP] and [Ppl-Numpl-Cl-NP] in terms of the (non-)licensing of the nominal contained. Consider the following examples involving the nominals that express relative appellations:
Albeit fuqin and baba both lexically mean ‘father’, the appositive [Ppl-Numpl-Cl-NP] allows for both fuqin and baba, whereas the appositive [Psg-Numsg-Cl-NP] can only accommodate fuqin, as shown above. To understand this contrast, it is worth noticing that fuqin and baba differ in genre: baba is generally used for daily appellation, while nowadays fuqin is mainly adopted in written contexts and not for face-to-face addressing. Moreover, the identity represented by fuqin tends to be interpreted as associated with social expectations or criteria with respect to ‘what a qualified father should be like’, whereas baba is a rather neutral appellation. This can evidently be manifested by (15). Example (15a) is a line excerpted from a poem; here baba and fuqin are not interchangeable, as evidenced by the awkwardness of (15b). The reason behind is that the identity of fuqin, in addition to being defined in terms of biological/legal relations, would further imply ethical, emotional connections between a man and his child(ren). Therefore, it logically makes sense to say a man is a baba yet he does not qualify as a fuqin, but not the other way around.
Example (14) shows that the appositive [Psg-Numsg-Cl-NP] construction accommodates only the ‘evaluation-flavored’ nominal of ‘father’, whereas [Ppl-Numpl-Cl-NP] is not subject to this condition. A similar example is shown in (16). Nainai is a neutral daily appellation for ‘grandma’, while the compound laonainai ‘old grandma’, due to the existence of the adjectival morpheme lao ‘old’, could imply some evaluations/judgments about the characteristics typically considered to be owned by aged people. Observe that the nominal permitted to occur in [Psg-Numsg-Cl-NP] is laonainai rather than nainai in (16a), a context in which the speaker intends to implicitly convey a judgment like ‘lacking modern science knowledge’ about the referent under discussion. By contrast, the nominal in [Ppl-Numpl-Cl-NP] in (16b) can be either laonainai or nainai.
The last piece of evidence manifesting the evaluation-/judgment-related meaning of the nominal in [Psg-Numsg-Cl-NP] can be found from the fact that, while [Ppl-Numpl-Cl-NP] can accommodate a semantically highly generic nominal like ren ‘person’, [Psg-Numsg-Cl-NP] always requires the identity information encoded by the nominal to be as specific as to properly underpin the evaluation/judgment intended by the speaker. Compare (17a) and (17b):
As shown above, for [Ppl-Numpl-Cl-NP] in (17a), both ren ‘person’ and chujiaren (‘leave-home-person’) ‘monk’ can be permitted. By contrast, for [Psg-Numsg-Cl-NP] in (17b), only chujiaren, but not ren, can be ruled in. To understand this contrast, note that chujiaren differs from ren in that the identity denoted by chujiaren can easily trigger some understandings commonly held about monks, such as ‘normally a monk would keep a low profile and stay away from people’s attention’, whereas ren is too generic to imply specific evaluations/judgments as chujiaren does. As a result, only chujiaren, but not ren, can meet the semantic requirement of the NP in [Psg-Numsg-Cl-NP] in the context of (17b).
To summarize, this subsection illustrated the discrepancy between the appositive [Psg-Numsg-Cl-NP] and [Ppl-Numpl-Cl-NP] in terms of the semantics of the NP contained. It was shown that the nominal in [Psg-Numsg-Cl-NP] is normally one that expresses an identity regarding which certain contextually or conventionally held evaluations/judgments can be well implied in the current utterance, whereas this restriction is irrelevant to the nominal in [Ppl-Numpl-Cl-NP]. In the next subsection, further comparison will be conducted between the two constructions in their pragmatic properties.
3.2 Pragmatic asymmetries
With respect to pragmatic properties, so far, previous scholarly attention has been primarily paid to the [Psg-Numsg-Cl-NP] construction. A consensus has been reached among linguists that [Psg-Numsg-Cl-NP] is used for the purpose of conveying a subjective evaluation regarding the characteristics or behaviors of the referent denoted by the P in light of the referent’s identity denoted by the NP (e.g. X. Tang Reference Tang2013; Li & Chen Reference Li and Yi2016; Liu Reference Liu2016; Jin Reference Jin2017, Reference Jin2020; Wang Reference Wang2019). Consider the following examples (see also Jin Reference Jin2020):
Both (18b) and (19b) pertain to conveying the speaker’s certain subjective evaluation about the referent concerned, as explicitly indicated by the evaluative expressions like jiu yinggai ‘exactly should…’ and hen zhengchang ‘very common’. These subjective evaluations are in turn made in light of the identity denoted by the NP. Specifically, for (18b), it is taken natural that ‘he’, due to his ‘student’ identity, should study hard; for (19b), it is regarded reasonable, in view of the ‘newcomer’ identity, that ‘you’ would have a huge workload. Different from the (b) examples, the (a) examples make objective statements about some hearsay rather than expressing speaker-oriented evaluations. As demonstrated by the hash symbols, the use of the appositive [Psg-Numsg-Cl-NP] in the (a) examples would lead to unnaturalness.
However, an important fact largely neglected in prior studies is that the above shown pragmatic condition on the use of [Psg-Numsg-Cl-NP] is absent in the case of the appositive [Ppl-Numpl-Cl-NP]. Compare (20) with the (a) examples in (18)–(19):
The examples in (20) minimally differ from the (a) examples in (18)–(19) in that the former adopt the appositive [Ppl-Numpl-Cl-NP] whereas the latter use the appositive [Psg-Numsg-Cl-NP]. The naturalness of the former, but not the latter, clearly demonstrates that the pragmatic condition on the conveying of subjective evaluations does not universally apply to all appositive [P-Num-Cl-NP] expressions in Mandarin Chinese.
Another pragmatic asymmetry rarely noticed in the literature concerns the discourse function of the apposition [Num-Cl-NP] contained. The present study observes that the [Numsg-Cl-NP] in [Psg-Numsg-Cl-NP] and the [Numpl-Cl-NP] in [Ppl-Numpl-Cl-NP] typically perform different discourse functions. Consider the hypothetical conversation below:
In (21), to respond to speaker B’s question about the candidate(s) that speaker A has in mind, speaker A can adopt either tamen ‘they’ or tamen san ge xinsheng ‘they three fresh students’ in the reply. Specifically, while a simple tamen ‘they’ is sufficient, adding san ge xinsheng ‘three fresh students’ after tamen can further draw the addressee’s attention to an identifiable quantified set of students in the given context, hence facilitating the addressee’s process of referent identification. Such a contextual referent identification function is, nevertheless, not available for yi ge xinsheng ‘a fresh student’, as demonstrated by the awkwardness brought about by the addition of yi ge xinsheng after ta ‘he’ in the reply.Footnote 25
In the meanwhile, it is observed that the [Numsg-Cl-NP] in the appositive [Psg-Numsg-Cl-NP] often precludes, whereas the [Numpl-Cl-NP] in the appositive [Ppl-Numpl-Cl-NP] allows for, co-indexation with another referring expression, as illustrated below:
In view of the above contrast in licensing co-referentiality, the present study proposes that the apposition [Numsg-Cl-NP] as shown above, which is incompatible with co-referentiality, is a property-denoting constituent, while the apposition [Numpl-Cl-NP], which is well compatible with co-referentiality, is individual-denoting.Footnote 27 This coupled with the above discussion on the discourse function of [Num-Cl-NP] leads the present study to claim that, at the level of discourse, the individual-denoting [Numpl-Cl-NP] in [Ppl-Numpl-Cl-NP] serves to help the addressee contextually anchor down the referents in question via identifying a specific quantified set of referents, and the property-denoting [Numsg-Cl-NP] in [Psg-Numsg-Cl-NP] serves to ascribe a property – one deemed essential for the speaker to make an evaluation – to the referent of the Psg.Footnote 28
The distinction between the two types of [Num-Cl-NP] appositions in discourse function can help to explain a discrepancy we observe regarding the possibility to elide [Num-Cl-NP] from the appositive [P-Num-Cl-NP]. Consider (24):
In (24a), the speaker intends to express their surprise about the fact that ‘he’, a well-established director in the film industry, would buy coffee by himself. Under the present analysis, the awkwardness brought about by the ellipsis of the apposition yi ge da daoyan can be explained in that the property-denoting [Numsg-Cl-NP], which defines Alan Lekongte’s identity as ‘a big director’, plays a key role in rationalizing the speaker’s unexpectedness about the ‘coffee-buying’ event, hence essential to the current discourse and not omittable. Differently, in (24b), the individual-denoting [Numpl-Cl-NP] ji ge chujiaren/ren serves to add information about the headcount of the referents, which is not essential to the understanding of the whole utterance but primarily used to facilitate identification of the target people. Therefore, the omission of ji ge chujiaren/ren is possible.Footnote 31
3.3 Interim summary
To summarize the discussion so far, it has been shown in the past subsections that there are discrepancies among different appositive [P-Num-Cl-NP] expressions in terms of the semantic characteristics of the nominal contained, the pragmatic properties of the whole construction, and the discourse function performed by the [Num-Cl-NP] with respect to the P. Accordingly, a two-way distinction has been drawn that, for one type of appositive [P-Num-Cl-NP], the [Num-Cl-NP] is property-denoting and the whole construction is for the purpose of conveying speaker-oriented evaluations; for the other type of appositive [P-Num-Cl-NP], the [Num-Cl-NP] is individual-denoting and the appropriate use of the construction is not subject to evaluation-making. In the rest of the paper, the two types of appositives will be called attributional and identificational appositives, respectively. Based on this dichotomy, the next section will develop a novel, non-unified syntactic analysis for the appositive [P-Num-Cl-NP].
4. A novel syntactic proposal
4.1 Syntactic behaviors of the appositive [P-Num-Cl-NP]
Before proceeding to the syntactic proposal, a more in-depth examination of the syntactic properties of the appositive [P-Num-Cl-NP] is warranted.
Corresponding to the semantic and pragmatic asymmetries as shown above, the attributional and identificational appositive [P-Num-Cl-NP] are observed to exhibit nonparallel behaviors at the syntactic level too. In addition to the aforementioned discrepancies in nominal ellipsis and topicalization (see Section 2), another syntactic difference is detected with respect to the modification of the apposition [Num-Cl-NP]. To illustrate, first, the [Numsg-Cl-NP] sequence contained in [Psg-Numsg-Cl-NP] can be directly modified by temporal adverbials, whereas such modification is not always applicable to the [Numpl-Cl-NP] in [Ppl-Numpl-Cl-NP]. As exemplified below, the [Numsg-Cl-NP] sequences in (25) can be modified by the temporal adverbials cengjing ‘once’ and xianzai ‘currently’ while the [Numpl-Cl-NP] sequences in (26) cannot.
Second, the [Numsg-Cl-NP] sequence in [Psg-Numsg-Cl-NP] can accommodate modification by evaluative adverbials, in which respect the [Numpl-Cl-NP] sequence in [Ppl-Numpl-Cl-NP] behaves differently, again. As shown in (27a) and (27b) respectively, the [Numsg-Cl-NP] sequence can be modified by the evaluative adverbial mingming ‘evidently’ whereas the [Numpl-Cl-NP] cannot.
The above observations, particularly the highly ‘predicational’ nature exhibited by the apposition [Numsg-Cl-NP] (i.e. being compatible with temporal and evaluative adverbials), may lead one to question whether the appositive [Psg-Numsg-Cl-NP] should be treated as a clause rather than as a nominal phrase. A closer scrutiny suggests the answer should be no. A strong piece of evidence comes from the fact that the appositive [Psg-Numsg-Cl-NP] can grammatically serve as the object of a verb that necessarily requires a nominal complement. For instance, [Psg-Numsg-Cl-NP] can serve as the object of the verb renshi ‘acquainted with’, which needs to be human-denoting, as in (28a), or as the direct object that bears the thematic role of ‘recipient’ in the double object construction, as in (28b).
Further note that the appositive [P-Num-Cl-NP] as a whole is a referring expression, i.e. referring to the referent denoted by the P. Given this, based on the hypothesis that referentiality is structurally encoded in the functional head D (e.g. Abney Reference Abney1987; Zamparelli Reference Zamparelli2000; Alexiadou, Haegeman & Stavrou Reference Alexiadou, Haegeman and Stavrou2008), the present paper considers that both the attributional and identificational [P-Num-Cl-NP] should correlate with DP in the underlying structure. In the following subsection, a detailed syntactic analysis will be presented for the appositive [P-Num-Cl-NP].
4.2 A syntactic analysis of the appositive [P-Num-Cl-NP]
In dealing with the syntax of appositives, a key issue needed to be addressed is the structural relationship between the anchor and the apposition. In the existing analyses of Chinese appositives (see Section 2), the anchor P and the apposition [Num-Cl-NP] have been treated as (i) the D head and its complement (Del Gobbo Reference Del Gobbo1999; Huang et al. Reference Huang, Li and Li2009), (ii) conjuncts of the coordination construction (S. Tang Reference Tang2010), or (iii) the specifier and the complement of a DP projection (Hong & Shi Reference Hong and Dingxu2012). Upon a closer scrutiny, nevertheless, none of the approaches is problem-free. Specifically, first, the head-complement analysis lacks sound empirical evidence which shows a categorial-selection relationship between the P and the [Num-Cl-NP]. For example, given the appositives like ta yi ge xuesheng ‘he as a student’ or tamen san ge xuesheng ‘they three students’, it is not clear how the numeral classifier expression yi ge xuesheng/san ge xuesheng would be categorially required by the pronoun ta/tamen as its complement. Second, the coordination analysis encounters a challenge from the fact that the P and the [Num-Cl-NP] cannot be reversed (e.g. ta yi ge xuesheng ‘he as a student’ versus *yi ge xuesheng ta ‘*a student as him’) whereas ordinary conjuncts can (e.g. meili shanliang ‘pretty and kind’ and shanliang meili ‘kind and pretty’). Last, the specifier-complement analysis warrants external stipulations to explain why the referent(s) of the whole DP construction would crucially depend on the content of the specifier (i.e. the anchor P) rather than on that of the complement (i.e. the apposition [Num-Cl-NP]), and thus theoretically, it is also not an optimal account for appositives.
Departing from the prior attempts, the present study adopts a ‘supplement’ analysis to account for the syntactic status of the apposition with respect to the anchor. This analysis is based on a crosslinguistically attested fact that the anchor and the apposition are not equivalent both semantically and syntactically, with the apposition exhibiting a rather ‘subordinated’, ‘non-headedness’ status with respect to the anchor, in view of which many linguists adopt the term ‘supplements’ to call appositions (Huddleston & Pullum Reference Huddleston and Pullum2002; Potts Reference Potts2005; Loock & O’Connor Reference Loock and O’Connor2013). Taking English for illustration,Footnote 37 first, it is the anchor but not the apposition that can always independently determine the referentiality of the appositive construction. As shown in (29), the referent of the appositive construction is independently determined by the anchor Mr Campbell, while the apposition a lawyer serves to provide additional background information about the man under discussion:
Second, it is the anchor, not the apposition, that is part of the main clause predication (Loock & O’Connor Reference Loock and O’Connor2013). This can be demonstrated by the fact that normally the anchor and the apposition cannot be reversed freely, as shown below:
Third, it is only the anchor, but not the apposition, that triggers agreement with the main verb of the sentence, as shown below:
In the recent literature, such a ‘supplement’ status of the apposition has been accounted for in that the apposition adds a ‘secondary message’ – which features a predicational relationship underlyingly – to convey a meaning like ‘[anchor] is [apposition]’ (e.g. Potts Reference Potts2005; Loock Reference Loock2007; Heringa Reference Heringa2012; Loock & O’Connor Reference Loock and O’Connor2013; Liu Reference Liu2016). More specifically, it is considered that the apposition functions as a supplement to the anchor by adding a predicate about the anchor via a null copula with the meaning of ‘be’. This makes the appositive construction ‘a short and simple spell-out of a fairly complex structure’, enabling the speaker to ‘add a complete message about an element in the sentence he pronounces, without having to use a new sentence and therefore without spelling out a subject or a verb’ (Heringa Reference Heringa2012: 570).
Under this viewpoint, at the structural level, Heringa (Reference Heringa2012) makes a pioneering attempt to assume that such a subordinated ‘secondary message’ expressed by appositives does not stem from a predicational syntax between the anchor and the apposition directly; rather, it is derived via a ‘bridging’ pro. With the technical details irrelevant to the linguistic phenomenon discussed in this paper aside,Footnote 39 it is postulated that the apposition is underlyingly predicative of a silent pronoun pro, and that the pro is co-indexed with the preceding anchor P, as visualized by a simplified version of Heringa’s proposal as given in (32). An important advantage of this treatment is that it is able to capture the apparent clausal properties exhibited by appositives (as shown in (35)) without assigning a clause-level projection to the whole construction, which thus fares well with the ‘predicational’ properties of the apposition on the one hand and the ‘referring’ nature of the whole appositive construction on the other.
To approach the syntax of the appositive [P-Num-Cl-NP] in Mandarin Chinese, the present paper also treats the apposition [Num-Cl-NP] as a supplement to the anchor P. This is crucially motivated by the fact that the referent of the appositive [P-Num-Cl-NP] is independently determined by the anchor P, while the apposition [Num-Cl-NP] primarily serves discourse-level functions such as adding background information for making evaluations, or facilitating referent identification via anchoring down a group of people. We also follow the line of Heringa (Reference Heringa2012) and others to analyze the appositive [P-Num-Cl-NP] as underlyingly encoding a ‘secondary message’ of predication,Footnote 41 as depicted below:
However, we slightly depart from Heringa (Reference Heringa2012) in that we do not adopt a coordination analysis for the syntax of appositives, due to the consideration as given at the beginning of this subsection. Instead, to structurally represent the supplement status of the apposition with respect to the anchor, we side with Potts (Reference Potts2005) and Del Gobbo (Reference Del Gobbo2017) in assuming a syntax of adjunction for appositives, analyzing the apposition as syntactically adjoined to the anchor, as depicted below (to be implemented shortly):
With respect to the categorial nature of the XP in (34), as for the case of the attributional appositive [P-Num-Cl-NP], recall from Section 4.1 that the property-denoting [Num-Cl-NP] allows for modification by temporal and evaluative adverbials. Interestingly, this is akin to what has been found with English appositives (Heringa Reference Heringa2012). As exemplified in (35), the English appositions can accommodate temporal adverbials like once and currently, and evaluation-oriented adverbials such as obviously and unfortunately. This has motivated Heringa to argue that the underlying structure of the English apposition contains TP and CP projections, i.e. the functional layers hosting tense- and evaluation-related adverbs, respectively.
In view of the commonality in terms of modification between the English appositions and the property-denoting apposition [Num-Cl-NP] in Mandarin Chinese, the present study hypothesizes that for the attributional appositive [P-Num-Cl-NP], the [Num-Cl-NP] structurally projects as high as into the CP layer. Accordingly, the whole appositive construction underlyingly correlates with the structure ‘[DP [DP P] [CP Num-Cl-NP]]’. Further, regarding the internal structure of the CP correlating with [Num-Cl-NP], the present paper incorporates Heringa’s insight, with our own amendments, of integrating the secondary, predicational message of appositives into the structural representation.Footnote 42 It is postulated that the apposition [Num-Cl-NP] underlyingly starts out forming a null-headed Copula Phrase (CopulaP) with a silent pronoun pro, and that the pro is co-indexed with the preceding anchor P; the null Copula head encodes a meaning of ‘be’. The [Num-Cl-NP] contained in the attributional appositive [P-Num-Cl-NP], as a property-denoting phrase, projects into NumP (but not further into the individual-denoting DP; see Li Reference Li1998), and the associated CopulaP continues projecting into TP and CP, i.e. the layers hosting temporal and evaluative adverbials, respectively. Finally, the CP is adjoined to the DP projection of the anchor P as a supplement. See (36) for visualization:
As for the case of the identificational appositive [P-Num-Cl-NP], in the same vein, it is assumed that the apposition [Num-Cl-NP] starts out forming a CopulaP with a silent pronoun pro, with the pro co-indexed with the preceding anchor P. The [Num-Cl-NP] here, as an individual-denoting expression, correlates with DP. Given that the [Num-Cl-NP] in the identificational appositive [P-Num-Cl-NP] cannot accommodate temporal and evaluative adverbials, the CopulaP is assumed to be directly adjoined to the anchor P without projecting into TP/CP. Accordingly, the whole construction underlyingly correlates with the structure ‘[DP [DP P] [CopulaP Num-Cl-NP]]’. See (37) for the syntax:
To recap, this subsection proposed a dichotomous analysis for the syntax of the appositive [P-Num-Cl-NP]. It is first assumed that both the attributional and identificational appositive [P-Num-Cl-NP] involve a syntax of adjunction (Potts Reference Potts2005; Del Gobbo Reference Del Gobbo2017). Then, to syntactically represent the predicational relationship between the anchor and the apposition, a CopulaP analysis is adopted, under which the apposition [Num-Cl-NP] occurs as the predicate of a silent pro and the pro is co-indexed with the anchor P underlyingly (Heringa Reference Heringa2012). The asymmetries observed between the attributional and identificational appositive [P-Num-Cl-NP] are captured by proposing that the CopulaP in the case of the attributional apposition projects as high as into the CP layer, whereas the CopulaP in the case of the identificational apposition does not project any further.
4.3 Consequences and further discussion
The dichotomous analysis of the appositive [P-Num-Cl-NP] proposed above has several welcome consequences.
First, this analysis can help to explain the observed distinctions between the attributional and identificational appositives in syntax, semantics, and pragmatics in a straightforward way. To be concrete, syntactically, given that the identificational appositive [Num-Cl-NP] underlyingly does not project into TP/CP whereas the attributional appositive does, it can be accounted for straightforwardly that the former cannot, whereas the latter can, accommodate sentential modifiers such as temporal and evaluative adverbials. Semantically, the observed condition on the licensing of the NP in the attributional appositive [P-Num-Cl-NP] can be well understood from the perspective that the [Num-Cl-NP] in this case serves the function of ascribing a property about the referent concerned, which in turn lays a crucial ground for the speaker to make a subjective evaluation regarding the referent in question. Given this, it follows that only the NP that could be well associated with an evaluation-related interpretation would be allowed to occur in the attributional [Num-Cl-NP], in contrast with the NP in the identificational [Num-Cl-NP]. Finally, regarding the pragmatic condition on the use of the attributional appositive [P-Num-Cl-NP], recall the assumption that a silent proform of the anchor P and the apposition [Num-Cl-NP] underlyingly form a copula clause which syntactically projects as high as into CP. Given the assumption that CP is an articulated functional projection (the ‘Split-CP’ hypothesis; Rizzi Reference Rizzi and Haegeman1997) that encodes a wide range of discourse properties including the speaker’s point of view or attitude (e.g. Cinque Reference Cinque1999; Speas & Tenny Reference Speas, Tenny and Di Sciullo2003; Hill Reference Hill2007, Reference Hill, Aelbrecht, Haegeman and Nye2012; Haegeman & Hill Reference Haegeman, Hill, Folli, Sevdali and Truswell2013; Wiltschko Reference Wiltschko2014), it follows straightforwardly that from a syntax-discourse interface perspective, the use of this construction, which underlyingly correlates with a CP projection, could be relevant to expressing speaker-oriented evaluations.
Second, the present analysis offers a way out for the contrast between the attributional and identificational appositive [P-Num-Cl-NP] in terms of nominal ellipsis and topicalization. Recall from (7)–(10) that only the identificational appositive [P-Num-Cl-NP], but not the attributional appositive [P-Num-Cl-NP], permits nominal ellipsis and topicalization. Under the present analysis, this can be well explained from the perspective of discourse conditions on nominal ellipsis and topicalization. Concretely, as for nominal ellipsis, it involves discourse linking in that the elided site presupposes the construction of a discourse set which includes (i) the entities referred to by the nominal phrase containing the ellipsis site and (ii) the entities referred to by the nominal phrase containing the antecedent of the elided element (López Reference López2000, Reference López2009; Merchant Reference Merchant2001; Elbourne Reference Elbourne2008; Cornilescu & Nicolae Reference Cornilescu and Nicolae2010). As for topicalization, it is generally agreed that a felicitous topic must at least be referential (see Gundel & Fretheim Reference Gundel, Fretheim, Horn and Ward2004 and the references therein). Compare (38a) and (39a) for illustration, which involve referential and non-referential [Num-Cl-NP], respectively:
To understand the contrast between the two (a) examples in nominal ellipsis, observe that a discourse set can be appropriately retrieved in (38a) (i.e. a set of books), which is composed by the books referred to by the nominal phrase containing the elided shu ‘book’ on the one hand (i.e. the three books ‘I’ bought) and the books referred to by the nominal phrase containing the antecedent of the elided shu ‘book’ on the other (i.e. the two books he bought). As such, the discourse condition on ellipsis is fulfilled. By contrast, the noun laoshi ‘teacher’ in (39a) is used as a property-denoting nominal (i.e. the property of ‘being a teacher’) and not referential; consequently, a discourse set cannot be appropriately constructed based on laoshi here. With respect to the asymmetry between the two (b) examples in topicalization, this can be straightforwardly attributed to the fact that shu ‘book’ in (38b) is D(iscourse)-linked to discourse referents whereas laoshi ‘teacher’ in (39b) is property-denoting, due to which the referentiality discourse condition on topicalization is respected in the former but violated in the latter.
Now recall from Section 3.3 that the [Num-Cl-NP] in the identificational appositive [P-Num-Cl-NP] is individual-denoting whereas the [Num-Cl-NP] in the attributional appositive [P-Num-Cl-NP] is property-denoting. Given this, the discrepancy between the two in nominal ellipsis and topicalization can be straightforwardly explained in that the NP in the former can, whereas the NP in the latter cannot, construct a discourse set or be D-linked to discourse referents, whereby the discourse conditions on nominal ellipsis and topicalization can be satisfied in the former case but not in the latter case.
Third, the present analysis fares well with the contrast between the attributional and identificational appositive [P-Num-Cl-NP] in the (non-)licensing of demonstratives. Observe that the identificational appositive [P-Num-Cl-NP] allows a demonstrative to intervene between the P and [Num-Cl-NP], while this is not allowed for the attributional appositive [P-Num-Cl-NP], as given below:Footnote 43
Under the present analysis, this can be straightforwardly accounted for in that the underlying structure correlating with the individual-denoting [Num-Cl-NP] involves a DP layer (see (37)). As a consequence, the demonstrative, an element standardly assumed to occur at the layer of DP (Abney Reference Abney1987; Zamparelli Reference Zamparelli2000; Leu Reference Leu2008; among many others), can be introduced into this construction. On the contrary, given that the underlying structure of the property-denoting [Num-Cl-NP] lacks a DP layer (see (36)), no syntactic position is available for the demonstrative, hence the incompatibility of the demonstrative here.
A final remark needs to be made before closing this subsection. Strictly speaking, the close correlation between the singular versus plural nature and the attributional versus identificational type of appositives as exhibited in this paper is better viewed as a strong tendency rather than as an absolute grammar rule. This is because under some cases, it might not be entirely impossible for one to utter the sentences as those given in (42): here [Ppl-Numpl-Cl-NP] is used as an attributional appositive (evidenced by the presence of the temporal adverbial cengjing ‘once’ in front of [Numpl-Cl-NP]), and [Psg-Numsg-Cl-NP] is intended as an identificational appositive (evidenced by the licensing of a demonstrative in front of [Num-Cl-NP], an operation only applicable to identificational appositives):
To understand the above examples, the present study proposes that it is feasible to consider the seemingly ‘exceptional’ identificational use of [Psg-Numsg-Cl-NP] as well as the attributional use of [Ppl-Numpl-Cl-NP] observed on some occasions as reflecting a variation of the speaker’s interpretation about the apposition [Num-Cl-NP] concerned.Footnote 46 Taking (42) for illustration, specifically, in the context of (42a), it is pragmatically possible for the speaker to intend the [Numpl-Cl-NP] liang ge name wujuwushu de hao-pengyou ‘two such unfettered good friends’ as a property-denoting expression, which underpins the speaker’s surprise about the fact that the two friends are now so courteous to each other. As a result, under this scenario, the [Ppl-Numpl-Cl-NP] can qualify as an attributional appositive construction, hence the licensing of temporal modification like cengjing ‘once’. For the context of (42b), given that the [Numsg-Cl-NP] yi ge yanyuan ‘one actor’ could help the hearer more easily identify who ‘he’ is in the given context (i.e. the one and only actor performing the show), the speaker may use it as an expression serving the function of facilitating referent identification. Therefore, the [Psg-Numsg-Cl-NP] could be classified under the type of identificational appositives and consequently can allow for the occurrence of the demonstrative zhe ‘this’ in front of [Numsg-Cl-NP].
However, even though the possibility is not entirely excluded for [Psg-Numsg-Cl-NP] to be used as an identificational appositive construction or for [Ppl-Numpl-Cl-NP] to be used as an attributional appositive construction in some occasions, the systematic differences observed between the attributional and identificational appositive [P-Num-Cl-NP] across the levels of syntax, semantics, and pragmatics warrants the need to maintain a dichotomous formal analysis for Chinese appositives, because all the asymmetries exhibited between [Psg-Numsg-Cl-NP] and [Ppl-Numpl-Cl-NP] as shown throughout Sections 3.1, 3.2, and 4.1 are absolute and can be unambiguously testified, independent of the specific context of utterance. If blurring the singular versus plural dichotomy at the structural level, one would not be able to come up with an explanatorily adequate analysis to account for these distinctions in a formal, systematic way.
Now one may raise a question as to why generally it is [Psg-Numsg-Cl-NP], but not [Ppl-Numpl-Cl-NP], that appears to be possessed with idiosyncratic characteristics relevant to expressing evaluative meanings, as has been illustrated in Section 3 . While we do not have a conclusive answer to offer for the moment, we suggest that this possibly is another instantiation of the subjective use of the nominal phrase composed by [yi ‘one’-Cl] in Mandarin Chinese. It has been well-noted that nominal constructions composed by the [yi ‘one’-Cl] sequence exhibit a strong trend of subjectivization in their use (see Yan Reference Yan2016 and the references therein). A representative example is the [hao ‘good’-yi ‘one’-Cl-NP] construction, which is used to express the speaker’s exclamation. Note that this construction can only accommodate the singular numeral yi but not plural numerals, as shown below:
Interestingly, a similar number constraint has also been observed for the so-called attributive qualitative binominal noun phrase in English, such as a fool of a policeman, a simpleton of a judge, that nitwit of a doctor (Den Dikken Reference Den Dikken2006). Notice that the nominals contained in this construction also cannot be plural, as illustrated below:
While the fundamental reason behind this pattern is still unclear, it seems that presumably there might be some number-related constraints crosslinguistically regarding whether a nominal expression could be, or tends to be, used to express subjective evaluations. For the purpose of the present study, it suffices to understand the notably strong tendency of the appositive [Psg-Numsg-Cl-NP] to exhibit an evaluation-related use as being consistent with a general rule of subjectivization on the use of [yi ‘one’-Cl] in Chinese (see Yan Reference Yan2016).Footnote 47
5. Conclusion
This paper investigates the appositive [P-Num-Cl-NP] in Mandarin Chinese and proposes an attributional versus identificational dichotomy for this construction. For the attributional appositive [P-Num-Cl-NP], (i) the [Num-Cl-NP] is property-denoting and allows for temporal and evaluative modification, (ii) the nominal contained denotes an identity regarding which certain contextually or conventionally held evaluations/judgments can be well implied in the current utterance, and (iii) the whole construction is used to convey speaker-oriented evaluations regarding the characteristics or behaviors of the referent of the P. For the identificational appositive [P-Num-Cl-NP], differently, (i) the [Num-Cl-NP] is individual-denoting and resists temporal and evaluative modification, (ii) there are no particular semantic requirements on the nominal contained, and (iii) the appropriate use of [P-Num-Cl-NP] is not subject to evaluation-related pragmatic conditions.
Based on this, this paper develops a dichotomous syntactic analysis for the appositive [P-Num-Cl-NP]. Upon treating the apposition as a supplement to the anchor (Huddleston & Pullum Reference Huddleston and Pullum2002; Potts Reference Potts2005; Loock & O’Connor Reference Loock and O’Connor2013), it is assumed that for both types of [P-Num-Cl-NP], the extended projection associated with the apposition [Num-Cl-NP] is syntactically adjoined to the anchor P. It is further hypothesized that the [Num-Cl-NP] starts out forming a CopulaP with a pro that is co-indexed with the anchor P (see Heringa Reference Heringa2012). Then, the asymmetries observed between the attributional and identificational appositive [P-Num-Cl-NP] are captured in that for the former, the CopulaP projects into CP, with the anchor [Num-Cl-NP] correlating with NumP; whereas for the latter, the CopulaP does not project further, and the [Num-Cl-NP] contained correlates with DP.
The present study fills a long existing research niche in the literature with respect to the heterogeneity of Chinese appositives. Taking the appositive [P-Num-Cl-NP] as a representative case, this paper is intended as one of the first to argue for a non-unified analysis for Chinese appositives. For a comprehensive understanding of different appositive constructions in Mandarin Chinese as well as across languages, needless to say, more work needs to be done in the future.
