Sage Journals: Discover world-class research

Abstract

The article revisits the canonical (Dem>Num>A>N) and non-canonical (A>Dem>Num>N & Dem>A>Num>N) prenominal patterns in Mandarin Chinese, from the perspective of the Labeling Algorithm. It shows that the syntactic distribution of adjectives are different, depending on the attributive-predicative sources and the presence or absence of the subordinator -de. The attributive adjective (with or without -de) and the de-less predicative adjective are analyzed as direct modification of the noun phrase, and it derives as a specifier of the noun, then falling into an agreement Agr projection. The non-canonical pattern undergoes AgrP-movement to the nominal left-periphery, which involves focus moving. However, the predicative adjective with -de is associated with a predicate of a reduced relative clause, and it undergoes successive focus-driven movement in the non-canonical pattern. The supporting evidence is illustrated in superlative and exclamatory adjectives as well as the CP-DP splitting hypothesis. The labeling-based explanation provides a specific picture to see the internal structure of adjectives, to look at how to package the informational constituents in the prenominal pattern, and to demonstrate the systematic parallelism between nominal and clausal domains in Mandarin Chinese.

Plain language summary

The article revisits two nominal orders in Mandarin Chinese: the canonical (Dem>Num>A>N) and non-canonical (A>Dem>Num>N & Dem>A>Num>N) patterns, from the perspective of the Labeling Algorithm . It shows that the syntactic distribution of adjectives are different, depending on the attributive-predicative sources and the presence or absence of the subordinator -de. The supporting evidence is illustrated in superlative and exclamatory adjectives as well as the CP-DP splitting hypothesis. The labeling-based explanation provides a specific picture to see the internal structure of adjectives, to look at how to package the informational constituents in the prenominal pattern.

Keywords

labeling focus left-periphery nominal-internal phrases Mandarin Chinese

Ordering Variation of Prenominal Adjectives

Greenberg proposes three ordering patterns among Demonstrative (Dem), Numeral (Num), Adjective (A) and Noun (N) in 24 possible orderings (Greenberg, 1963). The order Dem>Num>A is observed (when these elements occur) in the prenominal position, and either Dem>Num>A or A>Num>Dem is found (when these elements all occur) in the postnominal position (Greenberg, 1963: 87). Hawkins (1983) abandons the constraints of post-nominal orders and argues that every combination of post-nominal orders should be possible, given abundant counter examples under Greenberg’s formulation. However, more empirically refined pictures are discussed in Cinque (2005), which extends the above 3 orders in Universal 20 into 14 specific attested orders, when taking their occurring frequencies into consideration. In Mandarin Chinese, we can find three positions of the adjective in the prenominal pattern: the adjective right before the noun in (1a), before the demonstrative in (1b), or immediately preceding the number phrase in (1c).

(1) a. na liang-zhi maorongrong-de gou dou tai ke’ai le. that two-Cl fluffy-DE dog both too cute SFP “Those two fluffy dogs are just too cute.” Dem> Num> A> N b. maorongrong-de na liang-zhi gou dou tai ke’ai le. fluffy-DE that two-Cl dog both too cute SFP “Those two fluffy dogs are too cute.” A>Dem>Num>N c. na maorongrong-de liang-zhi gou dou tai ke’ai le. that fluffy-DE two-Cl dog both too cute SFP “Those two fluffy dogs are too cute.”Dem> A>Num>N

The order in (1a) follows Greenberg’s Universal 20, but the orders in (1b) and (1c) do not. The current literature has widely discussed the semantic distinction between (1a) and (1b) in Mandarin Chinese (Chao, 1968; Wu, 1994; C.R. Huang, 1989; C. T. J. Huang, 1982; Williams, 1998; Zhang, 2015), assuming that the pattern (1b) is a contrastively stressed modifier (Chao, 1968), with intersective and non-idiomatic interpretations (Zhang, 2015). However, there is no direct discussion of the pattern in (1c), according to my knowledge. I show that the interpretation restrictions observed in (1b) are also found in (1c) in Mandarin Chinese, and both are listed as marked prenominal patterns in Mandarin Chinese.

According to Chao (1968: 286), the pattern (2a) is contrastively stressed when it is represented by a capitalized modifier. The patterns (2b) and (2c) have the same readings of contrastive stress.

(2) a. na yi-wei DAI YANJING -de xiansheng Dem Num CL wear glasses –DE gentleman “that gentleman WEARING GLASSES” (Chao, 1968: 286; Williams, 1998:128) b. dai yanjing -de na yi-wei xiansheng wear glass-DE that Num-Cl gentleman “that gentleman wearing glasses” (Chao, 1968: 286; Williams, 1998:128) c. na dai yanjing -de yi-wei xiansheng that wearing glasses –DE Num-Cl gentleman “that gentleman wearing glasses”

In addition, the marked patterns have systematic restrictions on the intersective (3b) and non-idiomatic readings (3d).

(3) a. liang-wei naixin-de laoshi Two-Cl patient –DE teacher “two teachers, who are patient persons” (intersective) “two teachers, who teach patiently” (nonintersective) (Zhang, 2015:382) b. (naixin-de) na (naixin-de) liang-wei laoshi patient –DE that patient two-Cl teacher “two teachers, who are patient persons” (intersective) *“two teachers, who teach patiently” (*nonintersective) (Zhang, 2015:382) c. na liu ben huangse de xiaoshuo that six Cl yellow –DE novel “the six pornographic books” (idiomatic) (Zhang, 2015:383) d. (huangse-de) na (huangse-de) liu ben xiaoshuo yellow –DE that yellow –DE six Cl novel “the six yellow books” (non-idiomatic) *“the six pornographic books” (*idiomatic) (Modified from Zhang, 2015:383)

This paper will revisit the derivational mechanism of the unmarked (1a) and marked prenominal patterns (1b) and (1c) in Mandarin Chinese from the labeling algorithm, showing that prenominal adjectives have different underlying representations, depending on the attributive/predictive sources and the presence/absence of the subordinator -de in adjectives. The attributive adjectives with/without -de are associated with direct modification of the noun; the predicative adjectives without -de is related to direct modification, but the predicative adjectives with -de is derived from the predicate of a reduced relative clause. The marked patterns involve AP_COMPLEX-raising to the nominal left periphery, ending up with the direct modification, or first to the clausal left periphery then to the nominal left periphery, ending up with the predicative source of a reduced relative clause.

The paper is organized as follows. In section 2, two syntactic sources of adnominal adjectives (attributive/predicative) are characterized, and the de-contained and de-less adjectives have different underlying representations in the prenominal structure. In section 3, the theoretical mechanism of labeling algorithm (Chomsky, 2013, 2014; Rizzi, 2016) is demonstrated. In section 4, the derivations of two sources of adjectives in unmarked and marked patterns are explained from the labeling algorithm. The source of direct modification is derived from Spec-nP, undergoing AP_COMPLEX-raising to the nominal left periphery in the marked pattern. The predicative source of a reduced relative clause is generated at Spec-nP, but locating within a clausal domain, which involves AP_COMPLEX-raising to the clausal left periphery, and then to the nominal left periphery at last. In section 5, the empirical and theoretical evidence of focus raising is demonstrated. The last section concludes the study.

Characterization of Adnominal Adjectives

In this section, two syntactic sources of adnominal adjectives are introduced: attributive and predicative adjectives. Then, the semantic contrast of de-contained and de-less adjectives are differentiated in Mandarin Chinese.

Two Major Sources of Adjectives

The adjectives are traditionally divided into two types: attributive and predicative adjectives (Bolinger, 1967; Cinque, 2010; Sproat & Shih, 1988), and Chinese adjectives falls into the same distinction (Lü & Rao, 1981; Zhu, 1982). Attributive adjectives cannot be used as predicates and are usually analyzed as modifiers of the noun in the nominal phrase, but predicative adjectives have the potential to be used as a predicate in an independent copular construction (Alexiadou, 2001; Liu, 2016: 40).

(4) a. yi-ge yuanlai-de jihua one-Cl original-DE plan “one original plan” (attributive) b. yi-ge piaoliang-de wuzhe one-CL beautiful dancer “one beautiful dancer” (predicative) (5) a. *yi-ge jihua shi yuanlai-de. one-Cl plan BE original (*predicate) b. Tamen shi piaoliangde. they BE beautiful “they are beautiful.” (predicate)

The two sources of adnominal adjectives are associated with different interpretative properties. As in (6), the attributive adjective is interpreted with nonrestrictive, nonintersective, individual-level, literal or idiomatic meanings, etc; the predicative adjective is analyzed with restrictive, intersective, stage-level or individual-level, literal readings, and so on (Cinque, 2010: 2).

(6) Different interpretations of two sources of adnominal adjectives Direct Modification Source Reduced Relative Clause Source nonrestrictive reading restrictive reading non-intersective reading intersective reading modal reading of “possible” implicit RC reading of “possible” absolute (to a comparison class) reading relative (to a comparison class) reading absolute reading (of superlatives) comparative reading evaluative reading of “unknown” epistemic reading of “unknown” NP-dependent reading of “different” discourse anaphoric reading of “different” specific reading specific or nonspecific reading individual-level reading stage-level or individual-level reading literal or idiomatic reading literal reading (Cinque, 2010: 2)

The syntactic derivation of adnominal adjectives is controversially debated, with two accounts on the market: all deriving from a relative clause (Kayne, 1994; Sproat & Shih, 1988, 1991), or from different derivational positions, for example, either as modifiers of nouns or as predicates of reduced clauses (Alexiadou, 2001; Liu, 2016; Paul, 2005). Actually, Zhuang (2016) mentions the controversy of treating the -de absent attributive adjectives as involving modification at the lexical instead of phrasal level in that some de-absent adjectives can be explained at the lexical level, such as baicai“cabbage,”baijiu“wine,”xiaomai“wheat” with an idiomatic interpretation, and the lexical word as a whole does not allow any constituents to be inserted between the adjective and the noun without changing the meaning. However, there are many de- absent attributives as involving modification are allowed to insert -de without changing the meaning, such as piaoliang guniang“beautiful girl” or piaoliang-de guniang“‘beautiful girl”. Thus, we need to distinguish different types of de-absent attributive adjectives. I will show that the first view of assumption is problematic, following Paul (2005).

First, the suggestion of all adjectives from a relative clause is mainly based on the presence of subordinator -de. Sproat and Shih (1988, 1991) argued that all adjectives in Chinese are derived from relative clauses, because the subordinator -de following the adjective can be also an obligatory marker of a relative clause in Chinese, as in (7a). However, this argument is problematic, because the subordinator-de in adjectives is not always obligatory and can be deleted, as in the absent-de modification (7) (Paul, 2005: 760).

(7) a. juemi (-de) wenjian top-secret –DE document “top-secret documents” (Paul, 2005: 760) b. yi-ge fang (-de) panzi one-Cl square –DE plate “a square plate” (Paul, 2005: 760)

Second, the assumption of adjectives deriving from a predicative source is based on the fact that the adjective can function as a predicate in an independent clause (Sproat et al., 1991). In fact, we can find that a lot of adjectives, either with presence or absence of -de, are unable to constitute the predicate of a clause by itself, and must be supported by the copular shi and the particle de (Paul, 2005: 759). However, the copular shi and the particle de are prohibited in the Determiner Phrase (DP) (Paul, 2005: 759).

(8) a. zhe-ge panzi *(shi) fang *(de) This-Cl plate be square DE “This plate is square.” (Paul, 2005:760) b. Ta mai-le [_DPyi-ge (*shi) fang de panzi] 3SG buy-PERF one-Cl be square –DE plate “He bought a square plate” (Paul, 2005:760)

Considering a large sample of de-less and non-predicative adjectives, the assumption of adjectives unanimously from a relative clause is not acceptable. Instead, in this paper, I adopt that attributive and predicative adnominal adjectives are associated with different syntactic distributions: either as modifiers of a noun or as predicates of a reduced relative clause.

The Syntactic Nature of de-Contained and de-Less Adjectives

The de-contained and de-less adjectives make the two sources of adnominal adjectives more complex. In this section, following Liu (2016), I will show that all attributive adjectives are direct modifiers of the noun, but not all predicative adjectives are derived from the predicate of a reduced relative clause, depending on the presence or absence of the subordinator -de in Mandarin Chinese.

First, consider the attributive adjectives with or without -de, they both express the characteristics and properties of the noun they modify (Lü & Rao, 1981: 81; Liu, 2016: 47). In (9), the adjective zhuyao (de)“major DE” and gongtong (de)“common DE” is used to characterize the properties of the nouns jihua“plan” and pengyou“friend,” and it is associated with the interpretations of direct modification. The meaning does not change regardless of the presence or absence of the subordinator -de.

(9) a. Zhe liang-ge zhuyao (de) jihua hen shouhuanying. this two-Cl major DE plan very popular “These two major plans are very popular.” b. Women you liang-ge gongtong (de) pengyou. We have two-Cl common DE friend “We have two common friends”

Turn to the predicative adjectives in (10), the meaning changes, depending on the presence or absence of –de. In (10a), the adjective piaoliang-de“beautiful-DE” with -de expresses an intersective reading, indicating the looking of the dancer is also beautiful, and the adjective piaoliang“beautiful” without -de expressing a non-intersective reading, that is, the dancing action is beautiful. In (10b), the adjective congming-de “smart-DE” with -de has the restrictive meaning (i.e., the particular student is smart, but not all are smart), and the adjective without -de has the nonrestrictive reading (i.e., the students in the class are smart). The predicative adjectives with -de are associated with the predicate of a reduced relative clause, but those without -de have the readings of a direct modification. Thus, the predicative adjective has two underlying representations: a direct modification of a noun or a predicate of a reduced relative clause.

(10) a. Ta shi yi-wei piaoliang (de) wuzhe. she BE one-Cl beautiful DE dancer “She is dancer, who is beautiful.” (with -de) Or “She is dancer, who dances beautifully.” (without –de) (Adapted from Liu, 2006: 44) b. Ta shi ban li congming (de) xuesheng. he BE class IN smart DE student “He is a student, who is smart in the class .” (with -de) Or “He is a student, who is smart, in the class.” (without -de)

It makes sense that different syntactic distributions may lead to different interpretations in semantic interface. The assumption of the reduced relative structure explains the predicate interpretation, and the after-moving landing site corresponds to the modifier usage, ending up with the intersective readings. The assumption of the direct modification only leads to the non-intersective reading. Thus, these double association have been corroborated each other. Moreover, the sources of attributive and predicative adjectives have been extended, depending on the subordinator -de. When considering the sequential order of two sources of adjectives, it is widely accepted that the direct modification of adjectives is closer to the noun than the adjective in a reduced relative clause (Cinque, 2010). In my proposal, the general sequence of two sources of adjectives is represented in (11a), and the underlying hierarchy is demonstrated in (11b). The attributive adjective with or without -de is analyzed as the direct modification, deriving from the specifier of the noun. The predicative adjective without -de is associated with the direct modification, base generated in the specifier of the noun. The predicative one with -de is deriving from the predicate of the reduced relative clause, and may undergo an informational feature-driven moving-operation to the specifier position (and the derivational details will be specified later in section 4).

(11) a. Sequential order of two sources: [DemonstrativeP…[NumeralP…[AP_ReducedRC ]…AP_DirectMod…NP…[AP_ReducedRC ]]]] b. Underlying representation of adnominal adjective

Remnant and Phrasal Raising of AP

After introducing different sources of adjectives, let us turn to different locations of the adjective in the adnominal structure. The adnominal adjectives can occur in a non-canonical position, in addition to the regular position right before the noun. In Mandarin Chinese, the de-contained (attributive and predicative) adjectives can appear right before the number-classifier phrase or immediately precede the demonstrative, which is interpreted as foci with an emphatic effect. However, the de-less (attributive and predicative) adjectives can only occur in the canonical position: between the number-classifier phrase and the noun (12a-b).

(12) a. [ gongtong *(de)] na [gongtong *(de)] liang-ge [gongtong (de) ] pengyou dou feichang hao. common-DE that original-DE two-Cls original-DE friend both very good “Those two common friends are very good.” b. [Congming *(de)] na [congming *(de)] liang-ge [congming (de)] xiaohai shi wo-de pengyou. Smart that smart two-Cl smart child BE my friend “Those two smart children are my friends.”

The non-canonical pattern is discussed with remnant movement (Cinque, 2005) or phrasal movement (Zhang, 2015), both providing a direct derivational relation between unmarked and marked patterns, since they share a common underlying base. Based on these two methods, I will show that the derivational mechanism becomes more complex, when taking the sources of adjectives into consideration. Besides, I will explain how the focus-driven movement is involved in two non-canonical patterns.

Cinque (2005) argues that the order Dem>Num>A>N is fixed as the unmarked prenominal pattern, which is base-generated with nothing moving. In the marked prenominal pattern, a combination of NP-raising and remnant movement is necessary. If NP moves beyond AP, and then moves the remnant phrase across NumP and DemP, it will end up with the order A>Dem>Num>N. The underlying structure in line with Cinque (2005) is organized in (13).

(Cinque, 2005)

Unlike the remnant moving analysis, phrasal movement is assumed for the marked pattern in Mandarin Chinese. Zhang (2015) focuses on the marked order A>Dem>Num>N, assuming AP-raising to a position higher than that of the Demonstrative (Dem). The de-marker of the adjective (or relative modifier) is analyzed as a function head (F^H), and the numeral and the classifier are respectively generated at the specifier and the head of a functional projection UnitP (Cheng & Sybesma, 1998: 406). The de-marker and the adjective (or relative modifier) finally raise to the position higher than the demonstrative, ending up with the marked order A>Dem>Num>N. The detailed structure is formulated:

(14) [_FP Adj_k F^H_j [_DP Dem D [_UnitP Num Unit [_FP t_k t_j [_NP N]]]]] (Cheng & Sybesma, 1998: 406)

The remnant analysis (Cinque, 2005) and phrasal movement (Zhang, 2015) provide a direct derivational relation between unmarked and marked patterns, but it provides no explanation for the semantic contrast, since both unmarked and marked patterns start from the same underlying base. Based on remnant and phrasal movement, I will explain that the marked pattern involves focus moving to nominal left periphery, or first to clausal left periphery and then to nominal left periphery at last, depending on the attributive/predicative sources and the presence/absence of -de in adjectives. The derivational mechanism will be explained in the labeling algorithm.

The Labeling Mechanism of Head and Phrase

This section introduces the theoretical mechanism of Labeling Algorithm, which will be used to explain the derivational model of nominal-internal phrases. A generative grammar is generally discussed as a formal and computational system, in which the Merge operation combines two syntactic objects and Labeling is an operation to determine and label the new syntactic object. The tree of underlying representation must be labeled before it is transferred to the (PF and LF) interface, otherwise it will crash. The labeling algorithms for lexical level and phrasal level are distinguished (Rizzi, 2016).

First, in the labeling assumption, the lexical word enters the derivation as an unlabeled root but bears the feature “lex” (borrowed from Chomsky, 2013; Rizzi, 2016). The “lex” feature is contained in the head, so both the root and the functional head (e.g., v and n) are lex. The merged result (of the root and the functional head) is labeled by the functional head with lex, rather than the root with lex, because the root does not carry categorical feature, as in (15).

(15) The head-head labeling

The diagram in (13) demonstrates the labeling of the head-head merger, the feature lex is passed on to the mother category in the lexical level, but it is only constrained within the lexical level. The lex mechanism is parallel to the intermediate X’ level in generative grammar, defining the (sub)lexical and phrasal line. Beyond the lexical process, the lex feature cannot transmit to the mother category in the phrasal level, which is known as Lexical uniformity.

(16) Lexical uniformity: a lex category cannot contain non-lex material (Rizzi, 2016)

Beyond the lexical level (i.e., head-head merger), the head-phrase and phrase-phrase merger are also accounted in the labeling-based approach. In the head-phrase merger, the label is created by the closest head in a minimal configuration (Rizzi, 2016). For instance, in (17), T is the label when it merges with an Asp without any intervention; V is the label when it immediately merges with an DP; C is the label when it merges with an TP in a minimal configuration.

(17) head-phrase labeling

In the phrase-phrase merger, two alternative operations happen in the labeling algorithm (Chomsky, 2013). In the configuration {XP, YP}, if there is a shared feature between X and Y, then the shared feature is the label, which is known as a criterial feature. If no shared feature, one of the phrases must move out. The labeling algorithm takes the head of the remaining phrase as the label. In this case, the movement operation is a remedy to a structure that will crash and cannot undergo the interface transfer. Movement in this configuration falls into the anti-symmetry approach (), in which the symmetry of {XP, YP} must be broken, and one of them must move out for the labeling purpose.

(18) phrase-phrase labeling

In the labeling perspective, the motivation of the core syntactic operation (internal/external-merger) is fundamentally triggered by the labeling purpose. Any syntactic object, which cannot be labeled, is not allowed to undergo the interface transfer and will crash for linearization.

Labeling Analyses of Prenominal Patterns

In this section, the attributive and predicative adjectives in unmarked and marked prenominal patterns are demonstrated from a labeling-based approach, although they are subject to fall into the same configurational pressures. First, a nominal hypothesis is proposed for the direct modification to explain the attributive adjective (with or without -de) and the de-less predicative adjective in section 4.1, in which the adjective phrase is originated in Spec-nP, then moving to Spec-AgrP, with the overt subordinator -de as the morphological realization of functional head Agr°. The merged AP_complex (AP-Agr) undergoes movement to the informational focus packaging right before Spec-CL or further to the left periphery before DemP. Second, a clausal hypothesis is proposed for the de-containing predicative adjective in section 4.2, in which the adjective phrase is analyzed as a predicate of a reduced relative clause, and involves AP-raising to Spec-CP or further to nominal Spec-DP, ending up with the marked predicative patterns.

Derivation of Direct Modification

Analysis of the Unmarked Pattern

The (de-contained or de-less) attributive and the de-less predicative adjectives are assumed from the source of direct modification (as in section 2.2). Take the attributive adjective (19a) for example. The adjective is generated in the specifier of a noun phrase, and then moves to the specifier position of a higher functional AgrP, which is triggered by labeling requirement. The labeling-based representation is illustrated in (19c).

(19) a. na liang-ge yuanlai-de jihua DEM two-CL original-DE plan “those two original plans” b. na liang-ge congming xiaohai DEM two-Cl smart child ‘those two smart children c. Derivational mechanism of direct modification

In (19c), the √root jihua“plan” merges with the categorizing n, creating a nominal category n_lex (Chomsky, 2013, 2014; Rizzi, 2016). The adjectival phrase AP yuanlaide“original” enters the derivation, and the labeling algorithm takes the closest head n_lex in the minimal configuration. The functional Agr acts as an agreeing probe and bears uninterpretable φ-features in the nominal system, but it cannot be labeled by itself, like the Root and T in the clausal system; it must be strengthened by the movement of the subject in the specifier of its complement (following ), so the adjectival phrase AP has to move to the spec of Agr, labeling it the shared features <φ,φ><(=AgrP). The Agr is morphologically realized as the marker -de at the PF interface. Afterward, the classifier ge“Cl_lex” merges with AgrP, with the NumP liang“two” as its specifier. The idea of Num_lex in the specifier of Cl is in the line of Cheng and Sybesma (1998), in which the classifier projects a phrase, with the numeral as the specifier. The shared feature between NumP and ClP is relevant to number-feature (as part of φ-features), thus the label is created <φ,φ><(=ClP). Lastly, the phasal head D_lex enters the derivation, and the Dem phrase merges in its specifier, following Wood (2008) and Zhang (2015). In the configuration [DemP, D_lexP], the shared φ-features and specific feature [SPEC] are the label of the whole phrase. Through this derivation, the adjective immediately precedes the head noun, forming the unmarked pattern.

In the derivation, two operating hypotheses need to be noted: the AgrP-hypothesis and the M-merger hypothesis. For the AgrP-assumption, the AgrP is an agreement domain for the adjective modifier between DP and nP, based on Alexiadou et al. (2007)and Wood (2008). We can find agreement evidence between the attributive adjective and the head noun across languages, although this is not overt and shows no direct influence in Mandarin Chinese. For instance, the honorific marker in Korean si and nim are respectively attached to the adjective modifier (emha) and a head noun (sensayng), indicating the agreement relationship between the modifying adjective and the noun, as in (20a). More agreement examples are found on the (plural) number agreement tul on the genitive NP and the head noun in Korean, as in (20b). Both examples show the agreement relationship between the modifier and its head noun, which is analyzed as AgrP between the DP and NP (Alexiadou et al., 2007; Yoon, 1990; cited in Ma, 2016: 45).

(20) a. Ku emha-(si)-n sensayng-nim-i o-si-ess-ta. the strict-HON-AM teacher-HON-NOM come-HON-PST-DEC “the strict teacher came” (Ma, 2016: 45) b. Sonnim-tul-uy tochakkwangkyeng-tul guest-PL-GEN arrival scene-PL “the scenes of the guests” arrival’ (Ma, 2016: 45)

The AgrP-assumption also facilitates to distinguish the -de and de-less adjectival modifiers in Chinese nominal system. The functional head Agr provides an ideal host for the morphological realization of the subordinator -de, while a null spell-out is realized for de-less modification.

In addition, for the hypothesis of Morphological merger (M-merger), it applies between the adjective base and the subordinator, resulting in AP_COMPLEX (“AP+Agr” complex). The M-merger operation is supported by inability of inserting any constituent between the adjective and the subordinator -de, for example, yuanlai (*hao)-de jihua“original good -DE plan.”

If the Agr-hypothesis and M-merger assumption are in the correct line, the modifier should derive from a functional head Agr, with the adjective AP in its specifier position. The head-specifier relation falls into a morphological merger (M-merger) operation, creating the combined Agr_COMPLEX (AP+Agr), which does not allow any constituent to insert in between.

Analysis of the Marked Pattern

The adjectives, which appear in an atypical distribution, are widely assumed as foci, expressing emphatic interpretation, and only the de-contained adjectives are possible in the non-canonical pattern, as in (19a-b). The reordering of attributive/predicative adjectives in marked patterns, are triggered by one-step or successive cyclic AP_COMPLEX (“AP+Agr” complex) movement respectively, both of which are motivated by the [FOC] feature checking. An example of the marked attributive pattern (21a) is demonstrated in (21c).

As in (21c), the adjectival phrase AP yuanlai“original” initially enters the derivation in the specifier position of n_lexP, and then moves to the specifier node of the Agr_lex, creating the shared <φ,φ><(=AgrP) feature as its label. The AgrP-assumption and the M-merger hypothesis are adopted to analyze the syntactic distribution of adjectives (like the unmarked pattern). Following the classifier Cl and the number phrase NumP, the informational head Foc is projected with strong focus feature [FOC]. The [FOC] feature in the head Foc looks down to search its focused goal within its minimal domain, and it locates AP_COMPLEX, rather than NumP, when the adjective modifier AP_COMPLEX emphasizes the characterizing property of the thing it modifies (e.g., the plan is original, rather than recently changed). Thus, the Foc triggers AP_COMPLEX to its specifier. Then, the phasal head D_lex enters the derivation, and the Dem phrase merges in its specifier, following Wood (2008) and Zhang (2015), ending up with the marked Dem>A>Num>N pattern. In nominal phrases, in addition to the position right before the numeral phrase (i.e., ClP), the AP_COMPLEX can further successively raise up to the left periphery, immediately preceding the demonstrative phrase, which is motivated by the [FOC] feature in the functional discourse head Foc higher than the demonstrative phrase, resulting in the marked A>Dem>Num>N pattern. The focus packaging right before the Cl or before the Dem is supported by CP-DP parallelism, which theoretically permits an informational projection in the nominal left periphery.

(21) a. (yuanlai-de) na (yuanlai-de) liang-ge jihua original-DE that original two-Cl plan “those two original plans” b. (congming-de) na (congming-de) liang-ge xiaohai smart-DE that smart two-Cl child

Derivation of Reduced Relative Clause

The predicative adjective with -de in Mandarin Chinese is associated with the predicate of a reduced relative clause (as discussed in section 2.1). The adjective in the canonical pattern is derived in Spec-nP, moving to Spec-AgrP, but within a clausal structure. The AgrP (containing the adjective) first raises to the clausal left periphery, then to the nominal left periphery, as a result of successive focus-driven moving.

Analysis of the Unmarked Pattern

The adjective in (22a) is initially derived in Spec-nP, moving to Spec-AgrP, and the subordinator -de is realized in the functional head Agr. The AgrP (containing the adjective) locates within a clausal structure, as presented in (22b).

(22) a. na liang-ge congming-de xuesheng DEM two-Cl smart student “those two smart student” b. Predicative labeling in reduced relative clause

Unlike direct modification, one different operation is noted in the predicative source of a reduced relative clause. An empty PRO is assumed in the predicative source of a reduced relative clause, co-referring with Spec-RootP. The PRO assumption theoretically conforms that the adjectives (either predicative or attributive) are structurally derive in the fixed specifier of a nominal phrase (headed by either a overt noun or an empty PRO), and the predicative-attributive difference is depending on whether the AgrP (containing the adjective phrase) merges with the verbal root phrase in a clausal domain (as in the predicative adjective pattern) or merges with the classifier phrase in a nominal domain (as in the attributive adjective pattern).

In the configuration [AP, nP], LA takes the shared feature as its label, that is, nP. When Agr enters the derivation, AP moves to its specifier, taking the shared feature as its labeling, that is, AgrP. The non-phasal heads R(oot) merges into the derivation with v, creating a verbal category. The DP xuesheng“student,” base-generated in Spec-R (i.e., the Spec of R+v), is labeled as RP. Afterward, in the configuration [T, RP], with T being a non-phasal head, the subject in [Spec, RP] moves to [Spec, T], strengthening the non-phasal head. The raised subject DP xuesheng’student’ and the non-phasal head T agree in the φ-features, and labeled as such. Later, informational constituent packages into the derivation between the force phrase and the inflectional phrase, that is, between CP and TP, following the CP-split hypothesis (Rizzi, 1997). Outside the clausal projection, the classifier enters the derivation, with the numeral phrase in its specifier position. At last, the covert determiner is generated, and the demonstrative phrase enters in the nominal left periphery.

Analysis of the Marked Pattern

Successive focus moving is proposed in the marked pattern. The AP (to be precise AP_COMPLEX) is generated within the Root phrase, first undergoing raising to the clausal left periphery between CP and TP, as discussed in section 4.2.1, then to the nominal left peripheries, right before the number-classifier phrase or immediately before the demonstrative phrase.

(23) a. (congming-de) na (congming-de) liang-ge xuesheng smart DEM smart two-Cl student “those two smart students” b. Predicative labeling in reduced relative clause

The AP_COMPLEX (i.e., AP+Agr) raising to the focus projection between CP and TP is similar to the unmarked predicative pattern in section 4.2.1. The focalized constituent further moves from the clausal left periphery to the nominal domain: preceding the number-classifier phrase or further to the nominal left periphery right before DemP. The informational focus packaging in the nominal left periphery is theoretically supported by the CP-TP split hypothesis, both of which permit the focus in the left peripheries.

The syntactic analysis assumed in this paper is consistent with the differences in semantic relations. It is noted that only de-containing adjectives may appear in the noncanonical pattern. This might be relevant to the semantic relation between the adjective and the noun it modifies. The de-absent adjectives show a much closer semantic relation with the noun, and no constituent is allowed to be inserted in between; while the de-present adjectives bear a relatively loose semantic relation with the noun, and the demonstrative, the numeral and the classifier can be added in between.

Investigation of Nominal Left-Periphery Focus

Empirical Investigation of Left-Periphery Focus

The assumption that the nominal periphery is related to the emphatic interpretation and the movement of the adjective in the marked pattern involving focus is supported by superlative and exclamatory NP adjectives, which express speakers’ extreme and strong feelings (Hsu, 2014). Next, I will introduce two pieces of evidence from Hsu (2014).

First, the superlative adjective is associated with extreme and emphatic interpretation (Hsu, 2014), and it is much better to appear at the nominal periphery right before the demonstrative instead of preceding the noun, as in (24). The superlative example indicates that the nominal left periphery provides an informational focus phrase to express the extreme and exhausted readings in this position.

(24) a. Zui-gui-de na yi-liang che shi ta-de. Most-expensive-DE that one-CL car BE his “The most expensive car is his.” b. ?Na yi-liang zui-gui-de che shi ta-de. that one-CL most-expensive-DE car BE his. (Adapted from Hsu, 2014: 153)

Second, the exclamatory NP adjective usually expresses the extreme degree of unexpectedness and strong feelings of speaker, and it generally locates at the clause-initial position (Hsu, 2014). The exclamatory adjective must reverse its order to the left of the article, as in (25a-b), which is assumed involving A’-movement (Hsu, 2014). The similar phenomenon is also found in Mandarin Chinese, as in (25c-d), in which the exclamatory adjective usually appears at the nominal left-periphery.

(25) a. So beautiful a day! b. *A so beautiful day! c. Duome/Ruci meili-de yi-jian wuzi! how/so beautiful one-Cl room “How/So beautiful a room!” d. ?Yi-jian duome/ruci meili-de wuzi one-Cl how/so beautiful room (Adapted from Hsu, 2014: 154)

Both superlative and exclamatory adjectives emphasize speakers’ extreme and strong feelings, which is associated with focus in the readings. The syntactic distribution of these adjectives usually occur at the left-periphery of a nominal phrase. The above empirical evidence suggests that movement of adnominal adjectives in the left-periphery involves focalized constituents.

Theoretical Investigation of Left-Periphery Focus

The nominal left-periphery involving focus is also theoretically supported by the CP-DP parallelism hypothesis. The syntactic evidence that the informational structure can link the highest DP and the lower inflectional system (e.g., Num-Cl for numeral information, Dem for deixis information) is demonstrated by CP-DP parallelism. Parallel to split-CPs (Rizzi, 1997), the split-DP structure has been proposed by Bernstein (2001) for English and Romance, Aboh (2004) for Gungbe, and Simpson and Syed (2016) for Bangla. For instance, the Romance example in (26a) shows a basic word order, with the unmarked and neutral interpretation. However, in (26b), the post-nominal demonstrative pattern indicates a focus reading, with main stress.

(26) a. Este libro interesante this book interesting “This interesting book” (Spanish: Bernstein, 2001:2) b. El libro interesante este The book interesting this “This interesting book” (Spanish: Bernstein, 2001:2)

According to Bernstein (2001), the peripheral positions inside noun phrases are correlated with discourse interpretation. The split-DP system contains topic and focus constituents in the specifier position of left peripheries. Aboh (2004) further argued that the TopP and FocP project between the highest projection of the system (i.e., DP, parallel to CP in the clausal system), and the lower projection, which is linked to the nominal inflectional system (e.g., number, deixis). When the information packaging is directly linked the highest DP to the number information, the focus may appear right before the classifier projection. When the information packaging is directly related the highest DP to the deixis, the focus can immediately precede the Dem. These two focus packaging positions are both in the nominal-external system, and are named as higher information packaging and lower information packaging in (27).

(27) Multiple Information Packaging in nominal phrases

Applying the inflectional configuration, the reversed adjectival phrase is associated with the emphatic reading, syntactically realized before the Dem, or before the Cl. When the adjective appears right before the noun, it expresses a neutral and unmarked interpretation. Thus, the adnominal adjective can appear in three syntactic positions.

(28) Syntactic nature of attributive adjectives Distribution Domain Multi-functionality Syntactic Representation Nominal-external adjectives Higher focus packaging [_FocP Focus Dem…] Lower focus packaging [_FocP Focus Cl…] Nominal-internal adjectives Non-focus packaging [_AgrP AP_i Agr t_i N]

Concluding Remarks

A complex derivational analysis of the canonical and non-canonical prenominal patterns is discussed from the perspective of labeling algorithm. In the canonical pattern, the attributive and the de-less predicative adjective are analyzed as direct modification, but the de-contained predicative adjective is derived from the predicate of a reduced relative clause. In the non-canonical pattern, the (successive) focus raising is proposed. The focus-hypothesis is empirically supported by superlative and exclamatory adjectives and also theoretically conforms to the parallel properties between clausal CPs and nominal DPs, both of which can host the discourse constituents in the left peripheries.

Footnotes

Acknowledgements

The paper is independently written by the author

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.and/or publication of this article.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The paper is supported by Beijing Higher Education Association (MS2022079), China University of Mining and Technology- Beijing (SZ230901) (2023SKPYWF01) and Beijing Municipal Social Science Foundation (22YYC015).

Ethics statement

It is not applicable to involve the ethics statement

Research Interests

theoretical linguistics, generative syntax, language comparison

ORCID iD

Yanxiao Ma

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

References

Aboh

E. O.

(2004). The morphosyntax of complement-head sequences. Clause structure and word order patterns in Kwa. Oxford University Press.

Alexiadou

(2001). Adjective syntax and noun raising: Word order asymmetries in the DP as the result of adjective distribution. Studia Linguistica, 55, 217–248.

Alexiadou

Haegeman

Stavrou

(2007). Noun phrase in the generative perspective. Mouton de Gruyter.

Bernstein

J. B.

(2001). Focusing the right way in romance determiner phrases. Probus, 13(1), 1–29.

Bolinger

(1967). Adjectives in English: Attribution and predication. Lingua, 18, 1–34.

Chao

. (1968). A grammar of spoken Chinese. Berkeley: University of California Press.

Cheng

L. S. L.

Sybesma

(1998). Yi-wan tang, yi-ge tang: Classifiers and massifers. Tsing-Hua Journal of Chinese Studies, 28, 385–412.

Chomsky

(2013). Problems of projection. Lingua, 130, 33–49.

Chomsky

(2014). Problems of projection: Extensions. MIT.

10.

Cinque

(2005). Deriving Greenberg’s universal 20 and its exceptions. Linguistic Inquiry, 36(3), 315–332.

11.

Cinque

(2010). The syntax of adjectives: A comparative study. MIT Press.

12.

Greenberg

J. H.

(1963). Some universals of grammar with particular reference to the order of meaningful elements. In Greenberg

Joseph H.

(Ed.), Universals of grammars (pp. 73–113). MIT Press.

13.

Hawkins

J. A.

(1983). Word order universals. Academic Press.

14.

Hsu

Y. Y.

(2014). Nominal internal and external topic and focus: evidence from Mandarin. In: Proceedings of the 37th Annual Penn Linguistics Conference, 20 (1): 151–160.

15.

Huang

C.-R.

(1989). Mandarin Chinese NP de -a comparative study of current grammatical theories. Nankang.

16.

Huang

C. T. J.

(1982). Logical relations in Chinese and the theory of grammar [Doctoral dissertation]. MIT.

17.

Kayne

(1994). The antisymmetry of syntax. MIT Press.

18.

Liu

F. H.

(2016). Syntactic sources of adjectives in Mandarin Chinese. Studies in Chinese Linguistics, 37(1), 38–55.

19.

Lü

Rao

(1981). Shi lun fei wei xingrongci [On non-predicative adjectives]. Zhongguo Yuwen, 2, 81–85

20.

Paul

. (2005). Adjectival modification in Mandarin Chinese and related issues. Linguistics, 43(4), 757–793.

21.

Rizzi

(1997). The fine structure of the left periphery. In Haegeman

(Ed.), Elements of grammar. Kluwer, 281–337.

22.

Rizzi

(2016). Labeling, maximality and the head – Phrase distinction. Linguistic Review, 33(1), 103–127.

23.

Simpson

Syed

(2016). Blocking effects of higher numerals in Bangla: A Phase-based analysis. Linguistic Inquiry, 47(4), 754–763.

24.

Sproat

Shih

(1988). Prenominal adjective ordering in English and Mandarin. Proceedings of the 18th annual meeting of the North East Linguistic Society: 465–489.

25.

Sproat

Shih

(1991). The cross-linguistic distribution of adjective ordering restrictions. In Georgopoulos

Ishihara

(Eds.), Interdisciplinary approaches to languages: Essays in honor of S.-Y. Kuroda (pp. 565–592). Kluwer.

26.

Williams

(1998). Predemonstrative modifiers in Mandarin Chinese. In Dimitriadies

et al. (ed.), U. Penn working paper in Linguistics, 5. 2: 119–145.

27.

Wood

(2008). Similarities and difference between clauses and nominals comparative Syntax across theoretical approaches. https://api.semanticscholar.org/CorpusID:46521585.

28.

. (1994). Demonstratives, focus, and the interpretation of complex NPs in Mandarin Chinese, Studies in the Linguistic Sciences. In Yoon

(Ed.) Proceedings of the 5th Annual Formal Linguistics Society of Mid-America Conference, 24: 473–485.

29.

Yoon

J. Y.

(1990). Korean syntax and generalised X-bar theory [Doctoral dissertation], University of Texas.

30.

Zhang

N. N.

(2015). Nominal-internal phrasal movement in Mandarin Chinese. Linguistic Review, 32(2), 375–425.

31.

Zhu

. (1982). Yufa ji ang yi [Lectures on grammar]. Beijing: Commercial Press.

32.

Zhuang

. (2016). Ying qu fen han yu zhong de “Ju fa ci”. Zhong guo she hui ke xue bao [China social science newspaper], 957.

Revisit the Syntax of Nominal-Internal Phrases in Mandarin Chinese

Abstract

Plain language summary

Keywords

Ordering Variation of Prenominal Adjectives

Characterization of Adnominal Adjectives

Two Major Sources of Adjectives

The Syntactic Nature of de-Contained and de-Less Adjectives

Remnant and Phrasal Raising of AP

The Labeling Mechanism of Head and Phrase

Labeling Analyses of Prenominal Patterns

Derivation of Direct Modification

Analysis of the Unmarked Pattern

Analysis of the Marked Pattern

Derivation of Reduced Relative Clause

Analysis of the Unmarked Pattern

Analysis of the Marked Pattern

Investigation of Nominal Left-Periphery Focus

Empirical Investigation of Left-Periphery Focus

Theoretical Investigation of Left-Periphery Focus

Concluding Remarks

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

Ethics statement

Research Interests

ORCID iD

Data Availability Statement

References