The positive morpheme in Chinese and the adjectival structure
Chen-Sheng Luther Liu
Department of Foreign Languages and Literatures, National Chiao Tung University, 1001 Ta Hsueh Road, Hsinchu 300, Taiwan Received 23 July 2008; received in revised form 3 June 2009; accepted 3 June 2009
Available online 8 July 2009
Abstract
Chinese has a positive morpheme that has two allomorphs: a covert one and an overt one (i.e., the degree word hen). The former, behaving like a polarity item, only occurs in a predicate-accessible operator[-wh] domain with a structure like ½Op½-wh. . . X0
½-wh-operator½Deg P. . . Deg 0½
AP. . ., where the head X 0
, carrying the predicate-accessible operator[-wh] feature, not only introduces a predicate-accessible operator[-wh] but also licenses the occurrence of a degree phrase headed by the covert positive morpheme (i.e., Deg0), while the latter in other contexts. Having this as basis, I propose a condition on saturating Chinese gradable adjectives through which the bifurcated use of the ‘unmarked’ form of Chinese gradable adjectives can be well captured. Besides, the obligatory overt realization of a covert positive morpheme occurring in a predicate-accessible operator[-wh]domain, when the predicative adjective is substituted for by a pro-form, further implies that Chinese has an adjectival structure simpler than English. # 2009 Elsevier B.V. All rights reserved.
Keywords: Adjectival structures; Condition on saturating Chinese gradable adjectives; Polarity items; Positive morphemes; Predicate-accessible operators[-wh]; Predicate-accessible operator[-wh]domains; Unmarked forms
1. Introduction
According toZhu (1980, 1982),Lu¨ et al. (1980:11–12),Lu¨ (1984)andLiu et al. (2001), Chinese adjectives can be divided into two types: Adjectives, belonging to the first type, include non-gradable adjectives such as zhen ‘true’, jia ‘fake’, dui ‘right’, cuo ‘false’, heng ‘horizontal’, shu ‘acock’, wen ‘warm’ and zi ‘purple’, which are incompatible with any degree adverb, for example feichang ‘extremely’; the other type consists of gradable adjectives that allow degree modification. This distinction is clearly shown by the contrast below. (Also see Shi (2001:120–153) for further discussion on the distinctions between Chinese gradable and non-gradable adjectives.)
(1) a. *Ni-de da’an (*feichang) cuo. Your answer extremely wrong ‘*Your answer is extremely wrong.’ b. Na-ge nu¨haizi feichang piaoliang.
That-CL girl extremely beautiful ‘That girl is extremely beautiful.’
www.elsevier.com/locate/lingua
Available online at www.sciencedirect.com
Lingua 120 (2010) 1010–1056
E-mail address:csliu@faculty.nctu.edu.tw.
0024-3841/$ – see front matter # 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.lingua.2009.06.001
To make the murky boundary between the gradable adjective and the verb category clear,Zhu (1982:55)defines the Chinese gradable adjective in a way, as shown by (2) (Chao, 1968; Li and Thompson, 1981; Tang, 1988; Larson, 1991; McCawley, 1992; Paul, 2006).1
(2) X is a gradable adjective if and only if X can be modified by a degree adverb and X cannot take any genuine object(s).
One important syntactic characteristic of Chinese gradable adjectives, asZhu (1980, 1982)andLiu et al. (2001)
point out, is that they cannot occur as predicates unless they appear in complex forms, as the contrast between (3a) and (3b-c) illustrates.2,3
1
In addition,Zhu (1982)argues for a dichotomy of Chinese gradable adjectives into the simple and the complex one. (SeePaul, 2006for further discussion on this classification.) The former includes the mono-syllabic adjective and those whose reduplicated form is in the XXYY syllabic pattern, as shown by (i) and (ii), respectively (Lu¨ et al., 1980).
(i) da ‘big’
(ii) [[Xgan][Yjing]] ‘clean’ (ganganjingjing ‘extremely clean’)
The latter, asZhu (1982)argues, includes those in (A) the XX-er, XXYY, XYXY, and X-li-XY reduplicated pattern, (B) the X-ZZ reduplicated form, in which the ZZ component functions like a suffix, and (C) forms with adverbs of degree and in coordination, as illustrated by (iiia–g), respectively (Lu¨ et al., 1980:637–659).
Abbreviations used in this paper include: A: adjectives, CL: classifiers, CON: conjunctions, DE: the marker for modifying phrases like genitive phrases, relative clauses, and noun complement clauses in Mandarin Chinese, HEN: the degree word hen used as the overt positive morpheme, PAR: particles, and SFP: sentence final particles.
2According toDing et al. (1979),Zhu (1980),Lu¨ (1984),Shi (2001)andLiu et al. (2001), a Chinese gradable adjective always occurs as predicate
in a complex form, for example, a form with degree adverbs, and a reduplicated form, but seldom in a simple form, as the contrast between (i)–(iii) and (iv) shows.
(i) Zhangsan feichang pang. Zhangsan extremely fat ‘Zhangsan is extremely fat.’ (ii) Zhangsan pang-pang-de.
Zhangsan fat-fat-DE ‘Zhangsan is extremely fat.’ (iii) Zhangsan you gao you pang.
Zhangsan again tall again fat ‘Zhangsan is both tall and fat.’ (iv) *Zhangsan pang. (see footnote(3))
Zhangsan fat ‘Zhangsan is fat.’
The bu ‘not’ negation sentence and the contrastive focus construction are two of the limited sentence patterns where a simple adjective can occur as predicate.
(v) Zhangsan bu pang. Zhangsan not fat ‘Zhangsan is not fat.’ (vi) Zhangsan pang, Lisi shou.
Zhangsan fat Lisi thin ‘Zhangsan is fat, but Lisi is thin.’
These two sentence patterns, providing a very good empirical starting point for this study, will be detailed in section3. In order not to digress from the main theme of this study, in the rest of this paper I shall not discuss examples like (ii), which involves reduplication morphology to introduce the speaker’s subjective evaluation of the property expressed by the adjective, and (iii), where the correlative words you . . . you ‘again . . . again’, asZhu (1980:5–6)andLu¨ et al. (1980:561)argue, function as an intensifier marker to indicate the high degree of the property denoted by the adjectives connected by the correlative words (Zhu, 1980:35–40;Paul, 2006:306, 2007). The degree modification property of the correlative
(3) a. *Zhe-ke shu gao. This-CL tree tall
b. Zhe-ke shu feichang gao. This-CL tree extremely tall ‘This tree is extremely tall.’ c. Zhe-ke shu gaogao-de.
This-CL tree tall-tall-DE ‘This tree is quite tall.’
In other words, in a Chinese adjectival predicate sentence like (3b) the degree adverb (e.g., feichang ‘extremely’) is obligatory. This fact immediately brings us to the question of why the degree adverb is optional in an English adjectival predicate sentence but not in its Chinese counterpart raised by the contrast between (3a-b) and (4a-b).
(4) a. This tree is tall. b. This tree is very tall.
Extending this further, we can have this question reinterpreted as follows: Why does Chinese differ from English in that the latter allows a gradable adjective to occur as predicate in the positive form but the former does not?4To answer this question, I shall first explore the question of whether Chinese has a positive morpheme or not by having the contrast between (3a) and (4a) as the empirical starting point. The main theme which I eventually argue for is that Chinese does have a positive morpheme which has two allomorphs: an overt positive allomorph (i.e., the degree word hen) and a covert positive allomorph. The occurrence of the latter is subject to the condition on saturating Chinese gradable adjectives. More precisely, the Chinese covert positive morpheme, behaving like a polarity item, only occurs in a predicate-accessible operator[-wh]domain contained in the smallest clause that contains the adjectival predicate
and the operator with a structure like½Op½-wh. . . X0½-wh-operator½Deg P. . . Deg 0½
AP. . ., where the head X
0, carrying
the predicate-accessible operator[-wh]feature, not only introduces a predicate-accessible operator[-wh]but also licenses
the occurrence of a degree phrase headed by the covert positive morpheme (i.e., Deg0). The predicate-accessible operator[-wh]or the predicate-accessible operator[-wh]feature then coerces the covert positive morpheme to be marked,
and the marked covert positive morpheme further coerces the adjective (phrase) to be marked. An adjective (phrase) marked this way can only convey the positive degree meaning. Having this as background, I then argue that the Chinese covert positive morpheme still has to overtly realize as its overt counterpart hen even in a construction where its occurrence is licensed if the predicative adjective is substituted for by the pro-form nage ‘that’. This characteristic, as I further argue, leads us to suggest that Chinese has an adjectival structure simpler than English (Bresnan, 1973; Corver, 1997).
words you . . . you ‘again . . . again’, asZhu (1980:5–6)argues, is clearly shown by their incompatibility with non-gradable adjectives, as the ungrammaticality of (vii) illustrates.
(vii) *Na-ke juzi you cheng you he. That-CL tangerine again orange again brown
‘??The color of that tangerine is not only orange but also brown.’
3
Example (3a), which has a simple adjective as predicate, is gibberish if it is uttered in isolation, but is acceptable if it occurs as the answer for a question like (i), in which a ‘comparison’ is involved.
(i) Zhe-ke shu han na-ke shu, na-ke gao? This-CL tree and that-CL tree which-CL tall ‘As for this tree and that tree, which one is taller?’
Namely, in such kind of use the adjectival predicate gao ‘hot’ in (3a) means taller rather than tall. Since sentences with the comparison reading are not our concern here, in the remainder of this paper I shall not discuss the question raised by this kind of example: does Chinese have null comparative morphology?
4In this paper, I shall use the term ‘the positive form’ to represent a predicative adjective modified by the (covert or overt) positive morpheme, for
example, hot in coffee is hot, which has a denotation like [[DegPpos [[APhot]]]] (here the term pos represents the covert positive morpheme)
(Kennedy, 2005). So, in this paper the term ‘the positive form’ differs from the other term ‘the positive morpheme’ in usage. In addition, the term ‘the unmarked form’ is used to represent an adjectival predicate that is not modified by an overt degree term or a marked degree term, and the term ‘morpheme’ is used to replace the term ‘allomorph’ except where confusion might occur. Furthermore, in order to prevent the discussion from digressing from the main theme, I shall put aside examples containing attributive (or pre-nominal) adjectives and non-gradable adjectives.
The organization of this paper is as follows: section2begins by introducing the semantics of the positive morpheme and the distinctions between implicit and explicit comparison as basis for discussingKennedy’s (2005, 2007b)claim that Chinese has the degree word hen as the positive morpheme, and ends with some challenging data to this claim. I then start section3 by arguing for the assumption that Chinese has a covert positive morpheme and conclude this section with a descriptive generalization about the distribution of the Chinese covert positive morpheme. Pushing this descriptive generalization further, in section4I first propose the condition on saturating Chinese gradable adjectives to regulate the interpretation of the unmarked adjectival predicate in Chinese, and then provide evidence for the implication that the Chinese positive morpheme has two allomorphs: a covert and an overt one (i.e., the degree word hen). In section5, some apparent counterexamples to my assumption on the Chinese positive morpheme are pointed out first, and then I argue that these examples indeed provide strong evidence in support of the assumption that Chinese differs from English in the adjectival structure. Finally, the conclusion is reached in section6.
2. The semantics of the positive form of adjectives and implicit comparison
As is widely assumed in the formal semantics literature, gradable predicates map objects onto abstract representations of measure (i.e., scales) formalized as sets of values (i.e., degrees) ordered along some dimension (e.g., height, length, or weight). For example, Creswell (1977), von Stechow (1984), Heim (1985), and Kennedy and McNally (2005)analyze gradable adjectives as relations between individuals and degrees, assigning them denotations like (5), where expensive represents a measure function that takes an individual and returns its value, a degree on the scale associated with the adjective, so that expensive(x) represents x’s price.5
(5) [[expensive]] = ldlx.expensive(x) d
However, gradable adjectival predicates with the semantic type <d, <e, t>> do not themselves denote properties of individuals (i.e., <e, t>); therefore, we need to turn them into one with the semantic type <e, t> by having them combined with something. Asvon Stechow (1984)andKennedy and McNally (2005)suggest, this is the job of degree morphology. To put it more precisely, degree morphology saturates and restricts the degree argument of the adjectival predicate (i.e., d of <d, <e, t>>) by determining its value. At this point, one question we have to ask is how the degree argument of the positive form of gradable adjective such as expensive in (6) is saturated and restricted.
(6) The coffee in Milan is expensive.
As the first step in the discussion on the semantics of the positive form of gradable adjectives, I shall introduce the semantics of the positive morpheme as a way to bring us deep into the heart of this study. Does Chinese have a positive morpheme?
2.1. The positive morpheme
According toLewis (1970),Graff (2000),Baker (2002),Kennedy (2005:6), andKennedy and McNally (2005), the positive form of gradable adjectives, for example expensive in (6), seems to have two universal features. The first one which might be putative is that the positive form of gradable adjectives such as expensive and tall, in contrast with their comparative form (i.e., more expensive and taller), lacks overt morphology.
The second, asKennedy (2005:5)argues, is a semantic one. Most gradable adjectives have contextually dependent interpretations in the positive form (with a few important exceptions). For instance, whether example (6) is true or not depends in large part on the context in which it is uttered. To state it more concretely, sentence (6) could be judged true
5Bartsch and Vennemann (1972, 1973),Rusiecki (1985)andKennedy (1999), on the other hand, treat gradable adjectives as functions from
individuals to degrees, as shown by (i). (i) [[expensive]] = lx.expensive(x)
However, like the individual-to-degree-relation analysis, the individual-to-degree analysis can also be considered a degree-based approach to the semantics of (gradable) adjectives. And either of them assumes that gradable predicates do not themselves denote properties of individuals, and must combine with something to generate a property of individuals.
if asserted as part of a conversation about the cost of living in various Italian cities, as in (7a), but false in a discussion of the cost of living in Chicago versus Milan, as in (7b).
(7) a. In Milan, even the coffee is expensive!
b. The rents are high in Milan, but at least the coffee is not expensive!
One account for this variability, asKennedy (2005, 2007b)argues, is to assume a degree morpheme pos (i.e., a covert positive morpheme) with a denotation along the lines of (8), where s is a context-sensitive function from measure functions to degrees that returns a standard of comparison based on properties of the adjective g (such as its domain) and the context of utterance, to ‘morphologize’ the positive form of gradable adjectives (Kennedy, 2007a:17).
(8) [[Degpos]] = lglx.g(x) > s(g)
Namely, the positive form of adjectives is evaluated with respect to a context-sensitive function denoted by the covert positive morpheme: a contextual parameter (like the assignment function) that maps a measure function to a degree that represents an appropriate standard of comparison based on features of the context of utterance (what is being talked about, the interests or expectations of the participants in the discourse, etc.). Assuming this, the positive form of adjectives in (7a), for example, has a denotation like (9), which indicates that even though the denotation of the predicate is fixed, its truth conditions will vary according to the contextual features that affect the computation of s(expensive).6
(9) [[[DegPpos [APexpensive]]]] = lx.expensive(x) > s(expensive).
More importantly here, asGraff (2000)andKennedy (2005, 2007b)point out, one further fundamental semantic property shown by the positive form of gradable adjective is that it is vague; that is, the positive form due to its conventional meaning gives rise to borderline cases: objects for which it is unclear whether or not the predicate holds, meaning that borderline cases arise because of uncertainty about what exactly this degree is. This fundamental semantic property, asKennedy (2007b)suggests, is a feature of the context-sensitive function, which is constrained to return a value that counts as a significant degree of the relevant property in the context of utterance (possibly relative to a world; seeKennedy, 2005).
Kennedy (2005)further uses this semantic characteristic of the positive form to divide ‘comparison’ in natural languages into two different modes: explicit and implicit comparison. All natural languages have syntactic categories that express gradable concepts, and also have designated comparative constructions, which are used to express orderings between two objects with respect to the degree or amount to which they possess some property (Sapir, 1944). Interestingly, many languages use specialized morphology to express arbitrary ordering relations, for example the morphemes more/-er, less and as specifically for the purpose of establishing orderings of superiority, inferiority and equality in English (i.e., explicit comparison), as illustrated by (10a–c), respectively.
(10) a. Mercury is closer to the sun than Venus. b. This book is less expensive than that one. c. This book is as expensive as that one.
6Another option for the compositional semantics of the positive form of gradable adjectives within the degree-based semantics of gradable
adjectives is to assume a lexical type-shifting rule that has the same effect as the pos morpheme, as (i) shows (Chierchia, 1998; Kennedy, 2007a:16): (i) For any gradable adjective A, there is an A’ such that [[A’]] = lflx: f(x).[[A]](x),
where f is a function from individuals to truth values.
The domain restriction argument of a type-shifted adjective, asKennedy (2007a)points out, can be saturated by an explicit restriction like a for-PP, (e.g., for a Honda in (ii)), or via compositional principles which ensure that the domain restriction is ‘passed up’ to the matrix. (ii) Kyle’s car is expensive for a Honda.
Here, we simply assume the pos morpheme analysis, and leave it an open question as to which choice is the correct one. The reason why we adopt the pos morpheme option is that it makes the presentation simple and easy.
Whereas, other languages, like Samoan, take advantage of the semantics of the positive form (i.e., the inherent context sensitivity of the positive (unmarked) form) and use it as the primary means of expressing comparison by setting two objects in an adversative relation through conjunction of two positive-form adjectives that are antonymous, as (11) shows (Staseen, 1985).
(11) Ua tele le Queen Mary, ua la’itiiti le Aquitania. Is big the Queen Mary is small the Aquitania ‘The Queen Mary is bigger than the Aquitania.’
Thus, natural languages, as Kennedy (2007b) suggests, use two different modes (i.e., implicit and explicit comparison) to express comparison (Sapir, 1944).
(12) a. Implicit comparison
Establish an ordering between objects x and y with respect to gradable property g using the positive form by manipulating the context or context-sensitive function in such a way that the positive form is true of x and false of y.
b. Explicit comparison
Establish an ordering between objects x and y with respect to gradable property g using special morphology (e.g., more/-er, less, or as) whose conventional meaning has the consequence that the degree to which x is g exceeds the degree to which y is g.
These two modes of comparison (i.e., explicit and implicit comparison), asKennedy (2007b)further argues, differ from each other in the following ways: First, since the semantics of the positive form, for example [[[DegPpos [AP
expensive]]]] in (9), requires that the differential degree between expensive(x) and s(expensive) cannot be ‘crisp’ and has to be greater than some contextually determined norm, implicit comparison induced by the positive form of gradable adjectives, as expected, differs from explicit comparison in acceptability, especially in contexts involving crisp judgments (i.e., very slight differences between the compared objects) (Kennedy, 2007a:17). For example, explicit comparison in (13a) simply requires an asymmetric ordering between the degrees to which two objects possess the relevant property (i.e., the length of essays); therefore, crisp judgments are not problematic (Kennedy, 2005:11)).7
(13) Context A: Essay 1 is 10,000-words long and essay 2 is 5000-words long. a. Essay 1 is longer than essay 2.
long(e1) > long(e2)
b. Compared to essay 2, essay 1 is long. long(e1) > s[e2](long)
(14) a. Context B: Essay 1 is 10,000-words long and essay 2 is 9900-words long. Essay 1 is longer than essay 2.
long(e1) > long(e2)
b. ??Compared to essay 2, essay 1 is long. long(e1) > s[e2](long)
However, implicit comparison in (13b) requires the first novel to have a degree of length that is significant relative to the region of the scale whose lower bound is the length of the second essay; namely, the difference between the two
7Since either a positive or comparative form of English adjectives can be modified by adverbials like compared to, with respect to, and so forth, I
followKennedy’s (2007b)suggestion that compared to, with respect to and similar expressions modify the contextual parameters with respect to which the standard of comparison used to fix the extension of the positive or comparative form is evaluated.
(i) Compared to that essay, this essay is long. (ii) Compared to that essay, this essay is longer.
In other words, the semantic function of expressions like compared to is to manipulate the context relative to which the positive or comparative form is evaluated so that it only includes the argument of the adjective and the argument of compared to. So, I do not consider or treat expressions like compared to an implicit or explicit comparison marker.
degree values of length (i.e., the difference between the length of 10,000 words and the length of 5000 words), as Context A shows, must be significantly greater than some contextually determined norm and, in addition, induces a contextually given threshold specifying the degree of length that essay 1 has to exceed to be significantly long. However, in a context where very small differences in a property never count as being significant, a sentence involving implicit comparison cannot be true. For example, (14b) cannot possibly be true in Context B, which makes it an infelicitous description of such a state of affairs.
Second, implicit comparison, but not explicit comparison, generates an implicature that the positive form is false of the subject, as the contrast between (15a) and (15b) illustrates (Kennedy, 2007b).
(15) a. ??That essay is long compared to this one, and it’s already quite long. b. That essay is longer than this one, and it’s already quite long.
Third, as discussed byRotstein and Winter (2004),Kennedy and McNally (2005), andKennedy (2007b), not all gradable adjectives have context dependent standards in the positive form; for instance, adjectives like wet, open, bent, and impure are special in having positive forms in which the standard of comparison is a minimum value on the scale: x is bent is true as long as x has a non-zero degree of bend. Since the standard of comparison is not dependent on the context, I would expect a compared to constituent not to have any semantic effect on the interpretation of such adjectives; therefore, sentences containing a compared to constituent and adjectives with a positive form in which the standard of comparison is a minimum value on the scale should be infelicitous. This expectation indeed is borne out by the fact, as the contrast below indicates.
(16) a. B is more bent than A. b. ??Compared to A, B is bent.
Fourth, asKennedy (2007b)points out, composition of a measure phrase and a gradable adjective generates a predicate that is no longer context dependent; therefore, implicit comparison, as shown by the contrast between (17a) and (17b), differs from explicit comparison in that the former is impossible in a compared to construction that involves composition of a measure phrase and a gradable adjective because once a (non-explicit-comparison-denoting) adjective combines with a measure phrase, there is no standard of comparison left over to manipulate.
(17) a. ??Compared to Lee, Kim is 10 cm tall. b. Kim is 10 cm taller than Lee.
Having as background knowledge the semantics of the positive form of (English) gradable adjectives and the semantic distinctions between the implicit and the explicit comparison, now let us return to the question of whether Chinese has the positive morpheme raised by the contrast between (3a) and (4a) (repeated as (18a and b)).
(18) a. *Zhe-ke shu gao. This-CL tree tall
b. Zhe-ke shu feichang gao. This-CL tree extremely tall ‘This tree is extremely tall.’
Below, I first point out that, althoughKennedy’s (2005, 2007b)claim that the degree word hen can be considered the overt positive morpheme in Chinese is on the right track, some challenging data to his claim are still found. These data then become the central issue of section3, where a descriptive generalization about the distribution of the Chinese covert positive morpheme will be made.
2.2. Kennedy (2005): hen as the positive morpheme in Chinese
FollowingXiandai Hanyu Xuci Lishi[Examples and Explanation of the Functional Words of Modern Chinese] (1982:243–244) and Sybesma (1999:26–27), Kennedy (2005, 2007b) suggests that the degree word hen can be
considered the overt counterpart of the covert positive morpheme in Chinese.8,9I shall argue that this idea can be justified because of the following facts: First, asZhang (2002:169)points out, a predicative adjective modified by the degree word hen, for example hen qiong ‘very poor’ in (19a-b), always displays the contextually dependent interpretation as the positive form of English gradable adjectives does.
(19) a. Ta hen qiong, lian chi fan de qian dou mei you. He HEN poor even eat meal DE money all not have ‘He is poor. He even does not have money to eat meals.’ b. Ta hen qiong, lian xiao qiche dou mai-bu-qi.
He HEN poor even small car all buy-not-afford ‘He is poor. He even cannot afford a small car.’
Second,Kennedy’s (2005)claim that the degree word hen is the positive morpheme in Chinese can be adduced by the crisp judgment about borderline cases (i.e., objects for which it is unclear whether or not the predicate holds). The contextually dependent interpretations shown by the positive form of adjective, as Kennedy (2005, 2007b)argues, can be well accounted for by the delineation function. Since the delineation function is constrained to return to a value that counts as a significant degree of the relevant property in the context of utterance, a difference in acceptability is predicted when the context involves distinctions between objects based on minor but noticeable differences in degree. Here relevant to this characteristic is sentence (20), which contains a predicative adjective modified by hen, is unacceptable in scenario (21A), which involves crisp judgment, but acceptable in scenario (21B), which does not:
(20) Gen na-ke shu bi-qilai, zhe-ke shu hen gao. With that-CL tree compare-qilai this-CL tree HEN tall ‘Compared with that tree, this one is tall.’
(21) Context A: This tree is 15 meters tall while that tree is 13 meters tall. Context B: This tree is 15 meters tall while that tree is 5 meters tall.
In other words, in (20) the (implicit) comparison implied by the predicate hen gao ‘HEN tall’ requires ‘this tree’ to exceed ‘that tree’ in height by a significant amount.
Kennedy’s (2005)assumption that the degree word hen is the overt positive morpheme in Chinese not only well accounts for why (22a) is grammatical but (22b), if uttered in isolation, is not, and it also explains why the predicative adjective modified by hen in (22a) (i.e., hen gao ‘HEN tall’) displays the contextually dependent interpretation.
(22) a. Zhe-ke shu hen gao. (Gen san-ceng lou yiyang gao/Gen meiguo This-CL tree HEN tall With three-story building same tall/With American hong-shan yiyang gao).
redwood same tall
‘This tree is tall. (It is as tall as a three-story building/an American redwood.)’ b. *Zhe-ke shu gao.
This-CL tree tall
8This assumption, asKennedy (2005:6)points out, makes the first (putative) universal feature of the positive form of gradable adjectives
(i.e., being the absence of overt degree morphology) questionable.
9In Xiandai Hanyu Xuci Lishi [Examples of Explanation of the Functional Words of Modern Chinese] (1982:243–244), the word hen is
considered to have two different functions. One is that of an intensifier like English very. The other function might be paraphrased as a marker for the positive degree. As an intensifier, hen is stressed, but as a positive degree marker it is not. The reason why hen can function as the most ‘neutral positive degree marker’, asXiandai Hanyu Xuci Lishi (1982)argues, is because ‘when Chinese adjectives are used predicatively, they mostly have a contrastive meaning. (. . .) Predicative adjectives to which hen has been added lack this comparative (i.e., contrastive) sense. In this use, hen’s grammatical function is much stronger than when it serves as an intensifier’ (translated bySybesma (1999:26–27)).
Assuming that a gradable adjective is with the semantic type <d, <e, t>>, the adjective gao ‘tall’ in (22a-b) cannot itself denote properties of individuals (i.e., <e, t>). Hence, I need to turn it into one with the semantic type <e, t> by having it combined with a degree term. The assumption that the degree term hen in (22a) is the overt positive morpheme provides a natural way to explain why (22a) is grammatical but (22b), if uttered in isolation, is not. Because the degree argument of gao ‘tall’ in (22a) is saturated by the overt positive morpheme hen but that of gao ‘tall’ in (22b) is not. Additionally, given the semantics of the overt positive morpheme hen, the semantic characteristic shown by the predicative adjective modified by the degree term hen in (22a) (i.e., the predicate hen gao ‘HEN tall’ displays the contextually dependent interpretation) is naturally derived.
The examples considered so far seem to makeKennedy’s (2005)claim convincing. Yet his claim is challenged by examples like (23), in which the gradable adjective gui ‘expensive’ and pianyi ‘cheap’ both occur as predicate independently and convey the positive degree meaning only in each conjunct clause.
(23) Kafei gui, hongcha pianyi. Coffee expensive black-tea cheap
‘Coffee is expensive, but black tea is cheap.’
So, here I run into a dilemma. On the one hand, suppose Chinese does not have the covert positive morpheme which functions to turn a gradable adjective with the semantic type <d, <e, t>> into one with the property of individuals (i.e., <e, t>). I then will have a type-mismatch between gui ‘white’ and kafei ‘coffee’ in the first conjunct clause of (23), and the same also happens between pianyi ‘cheap and hongcha ‘black tea’ in the second conjunct clause. On the other hand, if I assume the existence of the covert positive morpheme in Chinese in order to explain why (23) is grammatical, then I would fail to account for the ungrammaticality of (22b).
3. The descriptive generalization about the distribution of the covert positive morpheme in Chinese
I shall start this section by arguing that in Chinese some adjectival predicate constructions indeed contain a covert positive morpheme. Then I end with the descriptive generalization that the Chinese covert positive morpheme, behaving like a polarity item, only occurs in a predicate-accessible operator[-wh]domain contained in the smallest
clause that contains the adjectival predicate and the operator.
3.1. The unmarked adjective in Chinese
According toXiandai Hanyu Xuci Lishi (1982:243–244)andSybesma (1999:26–27), Chinese adjectives differ from their European counterparts in that the latter choose the unmarked option for the positive degree but the former the unmarked option for the comparative. Namely, in European languages the comparative is morphologically marked whereas in Chinese the positive degree is marked by the most neutral ‘positive degree marker’ hen, as shown by the contrast between (24a-b) and (25a-b).
(24) a. John is taller. b. John is tall.
(25) a. (Zhangsan han Lisi, shei gao?) Zhangsan gao. (Zhangsan and Lisi who tall) Zhangsan tall
‘(As for Zhangsan and Lisi, who is taller?) Zhangsan is taller.’ b. Zhangsan *(hen) gao.
Zhangsan HEN tall ‘Zhangsan is tall.’
However, actually the language fact is more complex than whatXiandai Hanyu Xuci Lishi (1982)andSybesma (1999)indicate. AsZhu (1980:26–27)andLiu et al. (2001:196–197)point out, in Chinese it is possible for a positive-degree-denoting ‘unmarked’ gradable adjective to occur as predicate though the distribution is limited to the following
constructions: the bu ‘not’ negation sentence, the contrastive focus construction, the ma particle question, the epistemic adjectival small clause, the conditional, and sentences ending with the sentence final particle le, as shown by (26a–f), respectively (Tang, 1998:149;Huang and Li, 2008). (See footnote(1)for the distinctions between simple and complex (gradable) adjectives in Chinese.)
(26) a. Zhangsan bu gao. Zhangsan not tall
‘Zhangsan is not tall, but the possibility of Zhangsan’s being short is not excluded.’ ‘*Zhangsan is not taller.’
b. Zhangsan gao, Lisi ai. Zhangsan tall Lisi short
‘Zhangsan is tall, but Lisi is short.’ ‘*Zhangsan is taller, but Lisi is shorter.’ c. Zhangsan gao ma?
Zhangsan tall SFP ‘Is Zhangsan tall?’ ‘*Is Zhangsan taller?’
d. Zhangsan yaoshi linse dehua, jiu bu hui qing ni chi fan. Zhangsan if stingy PAR then not will invite you eat rice ‘If Zhangsan is stingy, he will not treat you to dinner.’
‘*If Zhangsna is more stingy, he will not treat you to dinner.’ e. Zhangsan xiao ni ben. (Tang, 1998; Huang and Li, 2008)
Zhangsan deride you stupid
‘Zhangsan derided you as being stupid.’ ‘*Zhangsan derided you as being more stupid.’ f. Hua hong le. (Zhu, 1980)
Flower red SFP ‘The flower got red.’ ‘*The flower got redder.’
Although the range of environments that sanction the positive-degree-denoting ‘unmarked’ adjectival predicate is not so wide, they look (totally) independent from each other. At this point, one question immediately comes out. What is the descriptive generalization about the distribution of such kind of simple gradable adjectival predicates?
In the following, I shall first argue that all of these constructions contain the covert positive morpheme, and then reach this descriptive generalization about the distribution of the Chinese covert positive morpheme. The Chinese covert positive morpheme, behaving like a polarity item, can only occur in a predicate-accessible operator[-wh]domain
contained in the smallest clause that contains the adjectival predicate and the operator.
3.2. The basic data
In this subsection, I will discuss environments that sanction the positive-degree-denoting ‘unmarked’ adjectival predicate one by one and then argue that the descriptive generalization about the distribution of the covert positive morpheme in Chinese is the following:
(27) In Chinese, the covert positive morpheme only occurs in a predicate-accessible operator[-wh]domain with
a structure like½Op½-wh. . . X0½-wh-operator½Deg P. . . Deg 0½
AP. . ., where the head X
0, carrying the
predicate-accessible operator[-wh]feature, not only introduces a predicate-accessible operator[-wh]but
also functions to license the occurrence of a degree phrase headed by the covert positive morpheme (i.e., Deg0). And this domain must be contained in the smallest clause that contains the adjectival predicate and the operator.
Based on this descriptive generalization, in section4I shall propose a licensing condition on saturating Chinese gradable adjectives to regulate the interpretation of the unmarked adjectival predicate in Chinese.
3.2.1. The bu negation sentence
The fact that the simple form of Chinese gradable adjectives can independently occur as predicates in a bu ‘not’ negation sentence like (28) immediately raises two questions related to the central issue of this paper (i.e., does Chinese have a positive morpheme?). First, how can the degree argument of the simple gradable adjectival predicate gao ‘tall’ get saturated if the negation marker bu ‘not’ is not considered a degree adverb?
(28) Zhangsan bu gao. Zhangsan not tall
‘Zhangsan is not tall, and the possibility of Zhangsan’s being short is not excluded.’
Second, as the semantic interpretation of (28) indicates, the negated adjectival predicate bu gao ‘not tall’ means anything but ‘tall’ and does not exclude the possibility of Zhangsan’s being short; this interpretation implies that what is negated in (28) cannot simply be the adjective phrase gao ‘tall’. Here, I would like to argue that the answer for the second question indeed provides a way to account for the first one.
AsGraff (2000)andKennedy (2005, 2007b)point out, in addition to the contextually dependent interpretations, the positive form of gradable adjectives shows another semantic characteristic: it establishes an ordering between objects x and y with respect to gradable property g denoted by the positive form, and g(x) must exceed g(y) by a significant amount. One option for the compositional semantics of the positive form of gradable adjectives, as Kennedy (2007a:17)suggests, is to assume a degree morpheme pos (i.e., the covert positive morpheme) with a denotation along the lines of (29) to ‘morphologize’ the positive form of gradable adjectives (Sapir, 1944; Kennedy, 2005).
(29) [[Degpos]] = lglx.g(x) > s(g)
Within (29), s is a context-sensitive function from measure function to degree which, based on properties of the adjective g and the context of utterance, further returns a value that counts as a significant degree of the relevant property in the context of utterance; namely, g(x) must exceed g(y) by a significant amount. To state it more clearly, the difference between g(x) and g(y) must be significantly greater than some contextually determined norm and, in addition, induces a contextually given threshold specifying the degree of height Zhangsan has to exceed to be significantly tall.
Given this semantic property of the positive morpheme, the semantic interpretation of (28) (i.e., Zhangsan is not tall, and the possibility of Zhangsan’s being short is not excluded) inspires us to analyze (28) as (30), in which there is a degree projection headed by the pos morpheme above the adjective phrase gao ‘tall’ (in section4.3, I shall argue that the covert positive morpheme (i.e., the pos morpheme) is in complementary distribution with its overt counterpart hen). (I will not touch on the issue of A0-to-Deg0movement at this point.)
(30) Zhangsan bu [DegPpos [APgao]].
Zhangsan not tall ‘Zhangsan is not tall.’
As clearly indicated by such a syntactic structure, what is directly negated in (30) is the degree phrase headed by the covert positive morpheme rather than the adjective phrase gao ‘tall’. So, I would expect example (28) to mean that it is not the case that Zhangsan’s height exceeds the contextually determined standard height of human beings by a significant amount. In other words, Zhangsan’s height might exceed the contextually determined standard height but the difference between the degree value of Zhangsan’s height and the standard height is not significant, and this does not exclude the possibility of Zhangsan’s being short, as the grammaticality of (31), provided by one of the anonymous reviewers, shows.
(31) Zhangsan bu gao, shijishang Zhangsan suan shi ai de. Zhangsan not tall actually Zhangsan consider is short DE
Thus, the semantic interpretation of a bu ‘not’ negation sentence containing a simple adjectival predicate implies that Chinese does have the covert positive morpheme.10
However, behind the assumption that (28) has a syntactic structure like (30) is the question of how the negation marker bu ‘not’ induces the occurrence of the covert positive morpheme. In the following, I shall argue that in Chinese the negation markers bu ‘not’ and meiyou ‘not’ are functional heads, projecting as a negative phrase, and select as complement an aspect phrase if the predicate denotes an event or is able to take an aspect marker such as -le, -guo, -zhe and zai but select a degree phrase complement if the predicate denotes a state, to which no aspect marker can attach (Ernst, 1995; Lin, 2003).11
As is widely known, Chinese has two negation markers: bu ‘not’ and meiyou ‘not’ (Wang, 1965; Chao, 1968; Li and Thompson, 1981). The negation marker bu ‘not’ is used with bare verbs and modals, while meiyou ‘not’ is used with various aspects and with accomplishment verbs, as the contrast between (32) and (33) illustrates.
(32) a. Zhangsan bu lai. Zhangsan not come ‘Zhangsan is not coming.’ b. Zhangsan bu/*meiyou hui qu.
Zhangsan not/not will go ‘Zhangsan will not go.’ (33) a. Zhangsan meiyou qu xuexiao.
Zhangsan not go school ‘Zhangsan did not go to school.’ b. *Zhangsan bu qu-le xuexiao.
Zhangsan not go-ASP school ‘Zhangsan did not go to school.’ c. *Zhangsan meiyou qu-le xuexiao.
Zhangsan not go-ASP school ‘Zhangsan did not go to school.’ d. Zhangsan meiyou qu-guo.
Zhangsan not go-ASP ‘Zhangsan has not been (there).’ e. *Zhangsan bu qu-guo.
Zhangsan not go-ASP
‘Zhangsan has not been (there).’
According toCheng et al. (1997:57–58), the same agreement restriction also holds between the negation and the verb/aspect in Chinese negative particle questions, as shown below.
(34) a. *Ta qu-le bu? He go-ASP NEG ‘Did he go?’
10As one anonymous reviewer says, this only shows that DegP can accommodate the possibility of Zhangsan’s being short. So, there is no logical
necessity that Zhangsan’s being short depends on the claimed semantics of DegP. As I have pointed out, the meaning of (30) is: It is not the case that Zhangsan’s height exceeds the contextually determined standard height of human beings by a significant amount, and the possibility of Zhangsan’s being short is not excluded. This meaning, as I have argued, can be well captured if the existence of a degree phrase headed by the covert positive morpheme is assumed for (30). If we do not make this assumption here, some other mechanism still has to be proposed to derive the meaning naturally; otherwise, the meaning of (30) cannot be well captured. Thus, there must be a logical necessity that Zhangsan’s being short depends on some mechanism that can derive the meaning of (30) straightforwardly. And DegP, as I have argued, is the candidate that fits in with (30) best.
11Lin (2003)argues, the distribution of the Chinese negation marker bu ‘not’ and mei ‘not’ is aspectually sensitive. The negation marker bu ‘not’
aspectually selects as complement a stative situation that requires no input of energy in order to obtain that situation while the negation marker mei ‘not’ aspectually selects a dynamic and bounded event as complement (Ernst, 1995).
b. *Ta qu-guo bu? He go-ASP NEG ‘Has he gone?’ c. Ta qu bu? He go NEG ‘Is he going?’ (35) a. Ta qu-le meiyou? He go-ASP NEG ‘Did he go?’ b. Ta qu-guo meiyou? He go-ASP NEG ‘Has he been (there)?’ (36) a. Ta hui/yinggai/neng qu bu?
He will/should/can go NEG ‘Will/Should/Can he go?’ b. *Ta hui/yinggai/neng qu meiyou?
He will/should/can go NEG ‘Will/Should/Can he go?’
For example, in (34a-b), the negation marker bu ‘not’ that appears as a question particle cannot appear with the perfective aspect -le or the experiential aspect -guo. In contrast, as (35a-b) indicate, we can use the negative question particle meiyou ‘not’ with these two aspect markers. The same contrast also obtains in (36a-b), which contain a modal. This leadsCheng et al. (1997)to suggest that the use of bu ‘not’ versus meiyou ‘not’ depends on the verb/aspect or modal in the sentence regardless of whether or not the negation markers are used as regular negative markers or question particles.
Given this,Cheng et al. (1997)suggest that if the agreement relation between the negation marker and the verb/ aspect shown by Chinese negative particle questions is due to a selection relation between negation and verb/aspect, this agreement phenomenon can be captured by assuming that the negation marker moves to the C position in overt syntax. The fact that agreement holding for typical negation forms also holds for negative question particles then can be naturally derived because they are in fact the same elements.Cheng et al. (1997) convincingly argue for this assumption by providing the following pieces of strong evidence: First, in cases where the matrix and embedded verbs share the same agreement requirement, ambiguity arises, as (37) illustrates.
(37) Ta yiwei [ni qu] bu? He think you go NEG
a. ‘Does he think or not think that you are going?’ b. ‘Does he think that you are going or not going?’
Although the negative question particle bu ‘not’ is in the matrix C indicating that the sentence is a matrix yes-no question, the (a) and (b) readings indicate that example (37) is ambiguous between the matrix and the embedded readings. The matrix reading arises when the negation marker bu ‘not’, as in (38a), occurs in the matrix clause while the embedded reading arises if the negation marker bu ‘not’, as (38b) shows, occurs in the embedded clause and then moves to the matrix.
(38) a. Ta bu yiwei [ni (hui) qu]. He not think you will go
‘He doesn’t think that you will go.’ b. Ta yiwei [ni bu qu].
He think you not go
Second, asCheng et al. (1997)point out, in a case where the negation marker only holds an agreement relation with the matrix verb, the sentence is not ambiguous, and the same also obtains when the agreement relation only holds between the embedded verb and the negation marker, as shown by (39)–(40), respectively.
(39) Ta yiwei [ni qu-guo] bu? He think you go-ASP NEG
a. ‘Does he think or not think that you have been (there)?’
b. *Does he think that you have been (there) or you haven’t been (there)?’ (40) Ta hui yiwei [ni qu-guo] meiyou?
He will think you go-ASP NEG
a. ‘*Will he think or not think that you have been (there)?’
b. ‘Will he think that you have been (there) or you haven’t been (there)?’
In (39), the negative question particle is bu ‘not’ and only can the matrix verb renwei ‘think’ satisfy the agreement requirement because the embedded verb qu ‘go’ has the experiential marker -guo attached to it; therefore, the question, as expected, does not have an embedded reading. On the other hand, the negative question particle in (40) is meiyou ‘not’ and only the embedded predicate can agree with it since the matrix has the modal hui ‘will’. And again, as expected, the question does not have a matrix reading.
Central toCheng et al. (1997)is the assumption that the negation marker bu ‘not’ and meiyou ‘not’ are functional heads that select as complement an aspect phrase, and there exists an aspectual selection restriction between the negation marker and the complement. Semantically, the aspectual meaning of a sentence enables us to grasp what type of situation is involved; and its viewpoint presents an event by focusing on all or part of that situation, rather as a camera lens may focus. In Chinese, viewpoints are usually indicated by overt aspect markers such as -le, zai, -guo, and -zhe. According to Smith (1997) and many others, there are at least two major viewpoint types, perfective and imperfective. Their properties are summarized below:
(41) Main types of viewpoints
Perfective viewpoints focus on the situation as a whole, with initial and final points. Imperfective viewpoints focus on part of a situation, including neither initial nor final point.
The distribution and interpretation of viewpoint indicators (i.e., aspect markers) is sensitive to the situation types and the constellation of a verb and its arguments. Situation types in Chinese, asSmith (1997)suggests, are generally distinguished as States, Activities, Accomplishments, Semelfactives, and Achievements in terms of how they differ from each other in the temporal properties of dynamism, durativity, and telicity, as summarized in (42):
(42) Basic situation types
States are static, durative (know the answer, love Mary)
Activities are dynamic, durative, atelic events (laugh, stroll in the park)
Accomplishments are dynamic, durative, telic events consisting of a process with successive stages and an outcome (build a house, walk to school, learn Greek)
Semelfactives are dynamic, atelic, instantaneous events (tap, knock)
Achievements are dynamic, telic, instantaneous events (win the race, reach the top)
So, in framing a sentence the speaker chooses situation type and viewpoint, subject to the pattern of the language. For instance, (43a) states the temporal schema of an Accomplishment: I and F indicate initial and final points, the dots indicate internal stages. (43b) states the imperfective viewpoint schema: the dots indicate internal stages of a situation. The slashes in (43c) indicate the interval actually presented in the sentence.
The slashed period represents an interval of Mary’s walking to school, an interval that includes neither the initial nor the final point. In other words, the aspectual information conveyed by a sentence is represented with a composite viewpoint and situation type temporal schema, and aspect markers function as viewpoint to present an event by focusing all or part of that situation.
However, Chinese adjectives in most cases do not take an aspect marker. If there does exist an aspectual selection restriction between the negation marker bu ‘not’ and the adjectival predicate complement, some element that plays a function analogous to what an aspect marker does to a VP has to occur in a bu ‘not’ negation sentence containing a simple adjectival predicate. Assuming this, in the following I shall argue that the degree term in an adjectival predicate sentence indeed plays such a role.
According toKennedy and McNally (2005), the distribution and interpretation of degree modifiers is sensitive to two major classificatory parameters about the scale structure of gradable adjectives: (A) whether a gradable predicate is associated with what we call an OPEN or CLOSED scale, and (B) whether the standard of comparison for the applicability of the predicate is absolute or relative to a context. For example, proportional degree modifiers are only compatible with closed-scale (or absolute limit) gradable adjectives that map their arguments onto scales with maximal and minimal elements, while non-proportional ones with open-scale (or relative) gradable adjectives, as illustrated by (44a-b)–(44a-b), taken fromKennedy and McNally (2005:355), respectively.
(44) Closed Scale Adjectives
a. The room was 100% full/empty. b. The flower was fully open/closed. (45) Lower Closed Scale Adjectives
a. ??The pipe is fully bent.
b. The pipe is now fully straight. (46) Upper Closed Scale Adjectives
a. We are fully certain about the results. b. ??We are fully uncertain about the results. (47) Open Scale Adjectives
a. ??Her brother is completely tall/short. b. ??The pond is 100% deep/shallow.
The selection restriction between degree modifiers and the scale structure of gradable adjective shown above in some sense is analogous to the selection restriction between aspect markers and the situation type of verbs. So, I have strong reason to say that the role that degree morphemes play with respect to gradable adjectives is analogous to the role that aspect markers play with respect to verbs. Thus, it is not unreasonable for us to say that the negation marker bu ‘not’ in examples like (28) selects a degree phrase as complement.
Furthermore,Lee and Pan (2001:711–713), similar toLi (1992:139)in analyzing the negation marker bu ‘not’ as an operator[-wh], argue that the negation marker bu ‘not’ is a focus-sensitive operator by pointing out that sentence (48a–d)
only differ from each other in the location of focus (Jackendoff, 1972; Jacobs, 1983; Rooth, 1992). (48) a. Lisi bu [chi]ffan, ta [zuo]ffan.
Lis not eat rice he cook rice
‘He does not eat the meal, but he cooks it.’ b. Lisi bu chi [fan]f, ta chi [mianbao]f.
Lis not eat rice he eat bread
‘He does not eat the rice, but he eats bread.’ c. Lisi bu [chi fan]f, ta [shuijiao]f.
Lis not eat rice he sleep
‘He does not eat his meal, but he sleeps.’ d. Lisi [bu]fchi fan, wo hai yiwei ta hui chi.
Lis not eat rice I yet think he will eat ‘He will not have a meal, though I expect he will.’
To describe the semantic interpretation of (48a–d) clearly, Lee and Pan (2001) incorporate Rooth’s (1992)
alternative semantics of focus into their semantic analysis of the Chinese negation marker bu ‘not’; that is, the effects of focus on semantics, for instance that in (49a), can be said to be the introduction of a set of alternatives (i.e., the focus semantic value introduced by the focus interpretation operator (i.e., [[likes [Sue]F]fshown by (49c)) that
contrasts with the ordinary semantic meaning of the VP [VPlikes [Sue]F] (i.e., [[likes [Sue]F]o= likes Sue).
So,Lee and Pan (2001:712–713)suggest that the semantic interpretations of (48a–d) can be simply represented by (50a–d), respectively.
(50) a. [[NEG [Ta chiffan]]]
Assertion: bu [[[Ta chi fan]]]
Presupposition: There is an alternative to chi ‘eat’, call it P, such that [[[Ta P fan]]]fan]]] b. [[NEG [Ta chi fanf]]]
Assertion: bu [[[Ta chi fan]]]
Presupposition: There is an alternative to fan ‘rice’, call it y, such that [[[Ta chi y]]] c. [[NEG [Ta [chi fan]f]]]
Assertion: bu [[Ta chi fan]]
Presupposition: There is an alternative to chi fan ‘eat rice’, call it P, such that [[Ta P]]. d. [[ASSERT [[Ta bufchi fan]]]]
Assertion: [[Ta bu chi fan]]
Presupposition: There is an alternative to Ta bu chi fan ‘he not eat rice’, which is [[Ta chi fan]]. Hence, it is not implausible for us to say that the negation marker bu ‘not’, being a focus-sensitive operator, carries the focus operator[-wh]feature, and a bu ‘not’ negation sentence containing a simple adjectival predicate, for example
(30) repeated as (51a), has a syntactic structure like (51b), in which the negation marker bu ‘not’ not only carries the (focus interpretation) operator feature (or introduces an operator) but also selects a degree phrase headed by the covert positive morpheme as complement.
(51) a. Zhangsan bu [DegPpos [APgao]].
Zhangsan not tall ‘Zhangsan is not tall.’
b. Zhangsan [NegPOp [[Negbu[+operator]] [DegPpos [APgao]]]].
Before moving into the next construction that sanctions the positive-degree-denoting ‘unmarked’ adjectival predicate, one point I cannot skip here is that my proposal on the bu ‘not’ negation sentence like (51a) is immediately challenged by examples in (52), where the negated adjectival predicate (e.g., bu-shufu ‘uncomfortable’) conveys a contrary reading different from what the negated predicate bu gao ‘not tall’ in (28), repeated as (53), does (Lu¨, 1984:223–229). (52) a. Zhangsan jintian shengti bu-shufu.
Zhangsan today body uncomfortable ‘Today, Zhangsan is uncomfortable.’ b. Zheli jiaotong bu-fangbian.
Here transportation inconvenient ‘Here the transportation is inconvenient.’ c. Ni zheme zuo bu-daode.
You this-way do immoral
d. Caipan bu-gongping. Referee not fair ‘The referee is unfair.’ e. Laoshi jintian bu-gaoxing.
Teacher today unhappy ‘The teacher is unhappy today.’ (53) Zhangsan bu gao.
Zhangsan not tall
‘Zhangsan is not tall, and the possibility of Zhangsan’s being short is not excluded.’
As Jesperson (1924:322), Klima (1964), Lu¨ (1984:223–229), and Teng (1985:471) point out, the distinction between the negated adjectival predicate bu gao ‘not tall’ and a contrary-reading-denoting negated adjectival predicate like bu-shufu ‘uncomfortable’ in fact corresponds to the contradictory (e.g., black vs. non-black) versus contrary (e.g., black vs. white) distinction. Furthermore,Teng (1985:471–473)argues that contradictory terms are syntactic facts while contrary terms are lexical facts because of the following syntactic and semantic distinctions. First, only contrary terms, like adjectives, can be modified by intensifiers.
(54) a. Zhangsan jintian feichang bu-shufu. Zhangsan today extremely uncomfortable ‘Today Zhangsan is extremely uncomfortable.’ b. Laoshi zuotian feichang bu-gaoxing.
Teach yesterday extremely unhappy
‘The teacher was extremely unhappy yesterday.’ (55) a. *Zhangsan feichang bu gao.
Zhangsan extremely not tall b. *Tang feichang bu tian.
Soup extremely not sweet
Second, contrary terms behave like adjectives in being able to occur as comparative predicates, as the contrast below indicates.
(56) a. Wo jintian bi zuotian (geng) bu-shufu. I today compare yesterday more uncomfortable ‘I am more uncomfortable than I was yesterday.’ b. *Zhangsan bi Lisi (geng) bu gao.
Zhangsan compare Lisi more not tall
Since a predicate in comparatives must specify a definite property, either positive or negative in meaning, and cannot refer to the absence of a certain property, which contradictory terms indicate, a contradictory term can occur in comparatives only when the negation marker bu ‘not’ precedes the comparative marker bi ‘compare’, as (57) illustrates. (57) Zhangsan bu bi Lisi gao.
Zhangsan not compare Lisi tall ‘Zhangsan is not taller than Lisi.’
Thus, along a line the same asTeng (1985:472), I suggest that the negation marker bu ‘in/un-/dis-’ of the contrary terms like bu-shufu ‘uncomfortable’ is a negative prefix whereas the negation marker bu ‘not’ of the contradictory terms, for instance bu gao ‘not tall’, is a sentential negation. The morphological operation done by the negative prefix bu- ‘in-/un-/dis-’ to the adjectival stem in cases like bu-shufu ‘uncomfortable’ is somewhat analogous to what reduplication morphology does to an adjectival stem, for example, honghong-de ‘red-red-DE’ and hong-tongtong ‘red-tongtong/quite red’. So, I would expect that the negative prefix bu- ‘in-/un-/dis-’ makes a simple adjective become a complex one that can occur as predicate independently, and the fact bears out this expectation, as shown by (52a–e).
3.2.2. The contrastive focus construction
A comparative construction semantically functions to establish orderings of superiority, inferiority, or equality between the two comparing degrees, each associated with one of the compared entities or concepts. Although the contrastive focus construction also involves ‘comparison’, ‘the comparison’ involved in this construction somewhat is relative, meaning that ‘contrast’ functions to highlight that one of the two contrasted entities or concepts has a particular quality by having it compared with the other. More concretely, ‘contrast’ (i.e., relative comparison), asLiu (2004:32)andXu (2007:48)suggest, functions to show either (A) one contrasted entity or concept has the [+A] quality while the other the [A] quality (e.g., the [+tall] quality versus the [tall] quality), (B) one contrasted entity or concept carries a quality opposite to that of the other along the same dimension (or scale) (e.g., tall versus short), or (C) one contrasted entity or concept has quality A while the other has quality B, as illustrated by (58a–c), respectively.12
(58) a. Zhangsan gao, Lisi bu gao. Zhangsan tall Lisi not tall
‘Zhangsan is tall, but Lisi is not tall.’ b. Zhangsan gao, Lisi ai.
Zhangsan tall Lisi short
‘Zhangsan is tall, but Lisi is short.’ c. Zhe-duo hua hong, na-duo huang.
This-CL flower red that-CL yellow ‘This flower is red, but that one is yellow.’
Simply put, what are contrasted in a contrastive focus construction are two different ‘categories’ of property (e.g., the [+tall] category versus the [tall] category, the tall category versus the short category, and the red category versus the yellow category) rather than two different ‘degrees’ of some property, as shown by the contrast between (59a) and (59b) in grammaticality.13
(59) a. *Zhangsan feichang gao, Lisi you-dian gao. Zhangsan extremely tall Lisi a-little-bit tall
‘??Zhangsan is extremely tall, but Lisi is a little bit tall.’ b. Zhangsan feichang gao, Lisi you-dian ai.
Zhangsan extremely tall Lisi a-little-bit short
‘Zhangsan is extremely tall, but Lisi is a little bit short.’
12
According toStaseen (1985), in languages like Samoan and Hixkaryana, comparison is effected by means of an adversative coordination of two clauses that contrast the target and standard of comparison along some dimension by using antonymous predicates or negation.
(i) Ua loa lenei va’a, ua puupuu lena. (Samoan) Is long this boat is short that
‘This boat is longer than that boat.’
(ii) Kaw-ohra naha Waraka, kaw naha Kaywerye. (Hixkaryana) Tall-not he-is Waraka tall he-is Kaywerye
‘Kaywerye is taller than Waraka.’
However, the semantic interpretation of (58b) implies that Chinese does not belong to this type of languages in expressing comparison; in other words, Chinese further differs from them in using the contrastive focus construction to highlight that one contrasted entity or concept has quality A while the other quality B rather than to convey a meaning of comparison.
13However, one might consider sentences like (i), where the degree value denoted by the degree adverb hen ‘very’ seems to be in contrast with that
denoted by the degree adverb geng ‘more’, as counterexample to my claim that what are contrasted in a contrastive focus construction are two different ‘categories’ of property rather than two different ‘degrees’ of some property.
(i) Zhangsan hen gao, Lisi geng gao. Zhangsan very tall Lisi more tall ‘Lisi is even taller than Zhangsan is.’
However, asXing (2004:216–217)argues, this kind of construction, instead of being analyzed as a contrastive focus construction, is a type of multi-clausal sentence (i.e., dijin ju ‘the increment construction’), in which there exists an incremental relation between the degree of Zhangsan’s height and that of Lisi’s. So, examples like (i) are not counterexamples to my claim.
According to the semantic property of the contrastive focus construction, it is not implausible for us to say that, to make the contrast achieved by a contrastive focus construction possible, the relevant category to which the relevant property of each contrasted entity or concept belongs must be identified first. The identification of the category of some gradable property associated with some entity or concept implies the relation between the degree of some property carried by such an entity or concept and the standard degree on the scale denoted by the same property must be identified first. According tovon Stechow (1984) andKennedy and McNally (2005), it is the degree morpheme (or the degree adverb) that does this job. So, I suggest that, in a contrastive focus construction like (58a–c), it is the contrastive function that induces the occurrence of a default degree morpheme in each conjunct clause.
More significantly, the contrast that example (58b) is felicitous under scenario (60A) rather than (60B) brings us further information to help identify the nature of the default degree morpheme occurring in a contrastive focus construction like (58b).
(60) A. Suppose the standard height for a man is 175 centimeters. Zhangsan is 185 centimeters tall and Lisi is 160 centimeters tall.
B. Suppose the standard height for a man is 175 centimeters. Zhangsan is 177 centimeters tall and Lisi is 173 centimeters tall.
As this contrast indicates, in (58b) the height of Zhangsan must exceed the contextually determined standard degree of human height by a significant amount, and the height of Lisi must be lower than the contextually determined standard degree of human height by a significant amount. This constraint leads us to suggest that the default degree morpheme in the contrastive focus construction like (58b) is the covert positive morpheme.
Furthermore, since a focus phrase, asRooth (1992)suggests, involves ordinary semantic value contributed by the focus operator and it is the contrastive focus function that licenses the occurrence of the covert positive morpheme, I suggest that a contrastive focus construction like (58b) has a syntactic structure like (61), in which there exists a focus head Foc0 carrying the operator feature. In addition, the focus head Foc0not only introduces an operator but also licenses the occurrence of a degree phrase headed by the covert positive morpheme.14,15
3.2.3. The ma particle question
According toShi (2001:260–263), a ma particle question like (62) should be distinguished from one like (63), in which the adjective occurs within the shi . . . de ‘is . . . DE’ sequence, because the former behaves the same as (64) in
14
Assuming that the focus operator licenses the occurrence of the covert positive morpheme, I would expect examples like (i)–(ii) to be grammatical and the fact bears out this expectation (The verb shi ‘is’ in (ii) functions to introduce a focus operator).
(i) Zhe-ke shu gaoF.
This-CL tree tall ‘This tree is tallF.’
(ii) Meicuo! wo shi qiongF, dan wo kao ziji guo huo.
No-mistake I is poor but I depend self live alive ‘No mistake! I am poorF, but I make a living by myself.’
15The other way to help identify the category of a property denoted by an adjectival predicate in Mandarin Chinese is adopt a construction like (i),
in which the adjective occurs in-between shi ‘is’ and the particle de (Ding et al., 1979:23). (i) Zhangsan shi gao de, Lisi shi ai de.
Zhangsan is tall DE Lisi is short DE
‘The height of Zhangsan belongs to the category of being tall, but that of Lisi the category of being short.’
Here, I would suggest that, like degree morphology, shi . . . de in sentences like (i) functions as a type-shifter in shifting a type <d, <e, t>> term into a type <e, t> term, or that shi . . . de alternatively functions to nominalize the adjective in a sense that the adjective in-between is construed as a nominal rather than an adjective. I shall leave it open for further research which one is correct.
asking about the degree of the property denoted by the adjectival predicate whereas what the latter asks about is the category to which the property denoted by the adjective occurring between shi ‘is’ and de ‘DE’ belongs.
(62) Zhe-duo hua hong ma? This-CL flower red SFP ‘Is this flower tall?’
(63) Zhe-duo hua shi hong de ma? This-CL flower is red DE SFP
‘Does the color of this flower belong to the category of being red?’ (64) Zhe-duo hua you duo hong?
This-CL flower have more red ‘How red is this flower?’
The way in which (62) differs from (63), asShi (2001:260–263)andXu (2007)clearly point out, can be shown by how their answers differ from each other; that is, the answer for a ma particle question like (62) must be one like (65a–d), which only differ from each other in the degree adverb occurring inside, rather than those like (66), which simply identifies the color category of the flower as the category of being red.
(65) a. Zhe-duo hua hen hong. This-CL flower very (or HEN) red ‘This flower is (very) red.’ b. Zhe-duo hua feichang hong.
This-CL flower extremely red ‘This flower is extremely red.’ c. Zhe-duo hua xiangdang hong.
This-CL flower quite red ‘This flower is quite red.’ d. Zhe-duo hua you-dian hong.
This-CL flower a little bit red ‘This flower is a little bit red.’
(66) Zhe-duo hua shi hong de. This-CL flower is red DE
‘The color of this flower belongs to the category of being red.’
Asvon Stechow (1984)as well asKennedy and McNally (2005)suggests, a degree adverb, for example those in (65a–d), functions to regulate the relationship between the degree of some property carried by the entity or concept modified by the adjectival predicate and the contextually determined standard degree on the scale denoted by the same property. Thus, it is not unreasonable for us to say that, in examples like (62), there exists a default degree morpheme (i.e., deg) which regulates the relationship between the degree value of this flower’s redness and the contextually determined standard degree of flowers’ redness and this relation is the target of being asked about in a ma particle question like (62). Assuming this, I suggest that example (62) has a rough structure like (67) in syntax.
(67) Zhangsan deg gao ma? Zhangsan tall SFP ‘Is Zhangsan tall?’
My assumption about the syntactic structure of a ma particle question like (62) further works together with the fact that example (67) can only have (68a) rather than (68b) as answer to provide us information for identifying the nature of this default degree morpheme.
(68) a. Zhangsan deg gao a, yi kan jiu zhidao. Zhangsan tall SFP one look then know
‘Of course. Zhangsan is tall; it is so clear. You even don’t need to look deeper.’ b. *Zhangsan deg gao a, buguo kan bu tai chulai.
Zhangsan tall SFP but look not too come-out
‘Of course. Zhangsan is tall. But if you don’t look carefully, it is hard to recognize that Zhangsan is tall.’
According toChao (1968), andLi and Thompson (1981), one important semantic (or pragmatic) characteristic of the ma particle question is that it can be functionally characterized as a request for specific information. For example, suppose you have already known that Zhangsan is tall. Some day, while talking with someone, you heard from him/her that Zhangsan is not tall, which makes you puzzled. So, you might ask question (67) to clarify the conflict and reconfirm your presupposition that the proposition denoted by Zhangsan deg gao ‘Zhangsan deg tall’ is true. As the contrast between (68a) and (68b) indicates, the answer for question (67) has to convey a sense that Zhangsan’s height must exceed the contextually sensitive standard height of human beings by a significant amount. Based on this semantic property, I suggest that the default degree morpheme deg in a ma particle question like (67) is the covert positive morpheme.
More importantly, since the question particle ma can be further considered a question operator or a functional head with the operator feature, I can say that the degree phrase in the ma particle question like (67) in fact is introduced by a head carrying the operator feature as the degree phrase is in the contrastive focus construction and the bu ‘not’ negation sentence’. So, I suggest that (67) has a more elaborated syntactic structure like (69), in which the question particle head ma, which carries an operator feature, not only introduces an operator but also introduces a degree phrase headed by the covert positive morpheme.
So, the meaning that (67) actually conveys is: Does Zhangsan’s height exceed the contextually determined standard height of human beings by a significant amount?16
3.2.4. The conditional
AsHuang and Li (2008)point out, the antecedent clause of bare (or simple) conditionals might take a simple adjective as predicate, as (70) illustrates.
(70) Zhangsan yaoshi gao dehua, Lisi jiu bu ai. Zhangsan if tall PAR Lisi then not short ‘If Zhangsan is tall, then Lisi is not short.’
Lewis (1975)argues that conditionals containing adverbs of quantification are not really conditionals but rather quantificational constructions, headed by the adverb of quantification, in which the antecedent clause functions as a restrictor on the quantifier. As for ‘bare’ conditionals where there is no operator that the if-clause could restrict,
Kratzer (1978, 1986), by taking (71a) as example, which has a slightly simplified logical form like (71b), suggests that they are implicitly modalized and the modal is usually the epistemic modal must.
16The same also obtains in the ba particle question, which, asChao (1968)andLi and Thompson (1981:309)argue, is used only in a context where
the speaker has an intent stronger than what s/he has in the ma particle question to request the hearer to reconfirm his/her presupposition, as (i) illustrates.
(i) Zhe-duo hua hong ba? This-CL flower red SFP ‘This flower is red, OK?’
