Iconicity and viewpoint: Antonym order in Chinese four-character patterns

Language Sciences 59 (2017) 117–134

Contents lists available at ScienceDirect

Language Sciences journal homepage: www.elsevier.com/locate/langsci

Iconicity and viewpoint: Antonym order in Chinese four-character patterns Shuqiong Wu College of Foreign Languages, Southwest University of Political Science and Law, Chongqing 401120, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 March 2016 Received in revised form 23 September 2016 Accepted 28 September 2016

This paper presents an analysis of co-occurring antonym pairs in Chinese four-character patterns, known as sizige in Chinese. Drawing data from Chinese language dictionaries, this study explores the order of antonyms in four-character patterns. The antonym pairs are found to exhibit a preference for a particular order. The analysis reveals a correlation between the order of the antonyms and iconicity. More precisely, iconicity of closeness (to the prototypical speaker), iconicity of temporal sequence, and iconicity of cultural norms and values conspire in motivating the normal order of antonyms in four-character patterns. The study also examines the issue of departing from the normal antonym order and proposes that the reversibility of antonym pairs can be attributed to the speaker’s viewpoint and semantic constraints imposed by the patterns. These findings suggest that the order of co-occurring antonyms is determined by general cognitive principles. Ó 2016 Elsevier Ltd. All rights reserved.

Keywords: Antonym order Four-character patterns Iconicity Viewpoint Reversibility

1. Introduction Antonym1 co-occurrence in English has drawn considerable attention from researchers in recent years. Systematic studies have shown that English antonym pairs, which were traditionally regarded as paradigmatically related words, tend to cooccur at higher-than-chance rates in discourse (e.g. Justeson and Katz, 1991; Fellbaum, 1995; Willners, 2001; Jones, 2002, 2006). Antonym co-occurrence is also a pervasive phenomenon in Mandarin Chinese. Antonym pairs in Chinese can be juxtaposed to form antonymous compounds without any connecting word, such as daxiao (big-small ‘size’) and fumu (fathermother ‘parents’). They can also co-occur in certain lexico-grammatical patterns, such as xian lai hou dao (early come late arrive ‘first come, first served’) and ku ye bu shi, xiao ye bu shi (cry also not right, smile also not right ‘be at a loss’). This study examines the co-occurrence of antonyms in Chinese constructions as in (1). (1)

a. zhao si mu xiang (morning miss evening think ‘yearn day and night’) b. bu nan bu nü (not man not woman ‘neither fish nor fowl’) c. qing zhong dao zhi (light heavy invert put ‘put the cart before the horse’) d. dian dao hei bai (reverse black white ‘mistake black for white’)

All these constructions contain four terms, two of which are antonyms. In Chinese these constructions are known as sizige or “four-character patterns” d in their written form they appear as four signs. The antonyms may occur in different slots E-mail addresses: [email protected], [email protected]. Lyons (1977) and Cruse (1986) use the term ‘antonym’ for gradable opposites in a narrow sense. In this paper, antonym is used as a cover term for all types of lexical opposites, which follows the usage in Jones (2002), Murphy (2003), Panther and Thornburg (2012) and Jones et al. (2012). 1

http://dx.doi.org/10.1016/j.langsci.2016.09.005 0388-0001/Ó 2016 Elsevier Ltd. All rights reserved.

118

S. Wu / Language Sciences 59 (2017) 117–134

within the patterns. In (1a), the antonyms zhao ‘morning’ and mu ‘evening’ occupy the first and the third slots respectively, while in (1b), the antonym pairs nan ‘man’ and nü ‘woman’ are in the second and fourth positions. In (1c) and (1d), by contrast, the antonym pairs qing/zhong ‘light/heavy’ and hei/bai ‘black/white’ are juxtaposed. Apart from the different syntactic positions, antonym pairs in these patterns tend to occur in a particular order. Their order cannot normally be reversed freely. For instance, in 1(a), zhao ‘morning’ and mu ‘evening’ cannot be reversed. Based on these observations, this study aims to investigate the co-occurrence of antonym pairs in Chinese four-character patterns and the underlying principles governing antonym order. The term ‘four-character pattern’ was first put forward by Lu Zhiwei, who defined it as “words or expressions which contain four syllables conjoined closely” (Lu, 1956: 401). Four-character patterns are very pervasive in Mandarin Chinese. Many well-known ancient Chinese works contain a large number of four-character patterns. The Classic of Poetry, for instance, contains 7284 lines, 6724 of which are written in four-character patterns (Xia, 1985). Four-character patterns are considered to be a unique feature of Chinese language (Guo, 1979) and an important part of the Chinese lexical system (Lü, 1979). The previous studies on Chinese four-character patterns have mainly focused on their formal properties, syntactic functions and semantic meaning (e.g. Liu and Xing, 2000, 2003; Han, 2008). Few studies have been conducted on the patterns with co-occurring antonym pairs. Moreover, existing research on antonym order (e.g. Jones, 2002; Kostic, 2015a; Hsu, 2015) suggests that many factors contribute to the preferred order of antonyms, including markedness, information-processing constraints, morphology, positivity, phonology and so on. However, what remains poorly understood is how antonym order is cognitively motivated and what the cognitive factors governing antonym order are. The goal of this study is to examine the cognitive principles underlying antonym order based on a case study of Chinese four-character patterns. The paper is organized as follows. Section 2 presents previous studies on antonym co-occurrence and antonym order. Section 3 examines the preferred order of antonym pairs in the patterns based on data from Chinese language dictionaries. Section 4 provides a detailed analysis of the cognitive principles governing antonym order. Section 5 explores the reversibility of antonym pairs in four-character patterns. The last section summarizes the findings of this study. 2. Antonym co-occurrence and antonym order Antonymy has been found on both the paradigmatic and the syntagmatic axes of language (Murphy, 2006; Panther and Thornburg, 2012). As a syntagmatic relation, antonym co-occurrence has been one of the major preoccupations of antonymy research in recent decades. The first study devoted to antonym co-occurrence was conducted by Charles and Miller (1989), who used the 1,000,000-word Brown Corpus and tested the hypothesis that antonymous adjectives tend to occur in the same sentence far more frequently than expected by chance. Justeson and Katz (1991) extended Charles and Miller’s study by investigating more antonym pairs and verified their conclusion. Subsequent work has dealt with different aspects of antonym co-occurrence: frequencies of antonym co-occurrence (e.g. Fellbaum, 1995; Willners, 2001; Jones, 2002); lexicogrammatical patterns in which antonyms co-occur (e.g. Mettinger, 1994; Willners, 2001; Jeffries, 2010; Lobanova et al., 2010; Davies, 2013); discourse functions of co-occurring antonym pairs in different genres, registers and languages (e.g. Jones, 2002, 2006, 2007; Murphy and Jones, 2008; Murphy et al., 2009; Muehleisen and Isono, 2009; Lobanova et al., 2010; Hsu, 2015); interpretation of antonym co-occurrence constructions (e.g. Murphy, 2006; Panther and Thornburg, 2012; Wu, 2014), and semantic profiles of antonyms in discourse (e.g. Kosti c, 2015b; Paradis et al., 2015). All these studies on antonym co-occurrence have caused a revival of interest in antonymy in recent years. Despite this abundance of previous research on antonym co-occurrence, relatively few studies have been devoted to the order of antonyms. Jones (2002) made the first attempt to address antonym order. After examining all the 56 pairs in the 3000 sentence database, he found that most antonym pairs favor one order over the reverse order. He identified eight factors affecting the order of antonym pairs. The most important factors are morphology (the tendency of the root word to precede its derivation) and positivity (the tendency for the antonym with positive connotations to precede the one with negative connotations). The other factors include magnitude (the tendency for the antonym denoting more quantity to precede the one denoting less), chronology (the tendency of antonym order in the syntax to observe the temporal sequence of antonyms in the real world), gender (the tendency for male antonyms to precede female antonyms), phonology (the tendency for the shorter antonym to precede the longer one), idiomaticity (the preferred antonym order is attributed to the semi-idiomatic status of some phrases), frequency (the tendency of more frequently used antonyms to precede those used less frequently) and markedness (the tendency for the unmarked antonym to precede marked one). Jones also discussed the reverse order of antonym pairs and ascribed it to the syntactic distance (the closer together antonym pairs are in a given context, the greater the likelihood that they will conform to their typical order) and collocational factors (the order of antonyms is determined by the words which are collocated with them). He concludes that the rules of sequencing, unlike grammatical rules, are “there to be broken as well as obeyed” (Jones, 2002: 136). Murphy (2006) and Jones et al. (2012) have suggested that we should be cautious about specifying antonym order, but they also point out that it is necessary to account for the cases which exhibit strong preferences for a particular order. They argue that the preferred antonym sequence is the result of a number of conspiring factors. Semantic, morphological, phonological and prosodic properties of antonym pairs, frequency, markedness, information structure and conventionalization all affect our preference for antonym ordering in predictable ways. Kosti c (2015a) provided a detailed account of antonym sequence in Serbian written discourse. Based on corpus data, her study suggests that the most basic factors governing antonym sequence are frequency (the antonyms with higher frequency

S. Wu / Language Sciences 59 (2017) 117–134

119

come first), neutrality (the generic cover terms precede the more specific ones), positivity (the more positive antonyms comes first), morphology (root antonym precedes the derived one) and quantity (the antonym denoting more of a quantity precedes the one denoting less). She argues that all factors are connected with markedness in one way or another, according to which the unmarked antonym tends to precede the marked antonym. In view of the point that “the markedness criteria discussed in the theoretical literature on antonyms cannot account for the sequence of all antonymous pairs” (Kostic, 2015a: 29), the author listed another two factors: temporal and visual-spatial ordering (antonym order is iconic of extra-linguistic temporal and/or visual-spatial order) and gender (The tendency for male antonyms to appear before female antonyms). The main argument in Kosti c (2015a) is that there is a strong correlation between antonym sequence in language use and the concept of markedness. The author tried to verify whether the markedness criterion which influences the distributional asymmetries in antonymic relations also applies to explain antonym order in actual language use. This perspective is very inspiring. However, as Jones (2002: 129-130) suggested, markedness does not always determine antonym order. In most cases the unmarked member of the pair precedes the marked one. However, in some cases it is difficult to determine which member of a pair is marked and which is unmarked, such as fat/thin. Murphy (2003: 187) also pointed out that not all antonym pairs have asymmetric distribution. Therefore, I strongly agree with Jones’ view that “markedness is a symptom rather than a cause of the sequence in which antonyms appear” (Jones, 2002: 129). Markedness itself is subject to the influence of other factors. It is the overlap between factors determining markedness and factors determining antonym sequencing that misleads analysts into thinking that markedness is the principle governing antonym sequence. Kosti c argued that the definition of markedness in Jones (2002) is too narrow (implying only semantic neutrality). She adopted the view in Waugh (1982) that the unmarked/marked relationship is conceptually represented as a figure/ground relationship, where the marked antonym is the figure and the unmarked antonym is the ground. This account demonstrates that markedness itself is in need of explanation in terms of human perception and cognition. Markedness is a fascinating concept, but it is too general. As Mollin (2014: 69) said: “As markedness is a principle operating on many different linguistic levels, it cannot serve as a constraint in its own right.” Therefore, we do not think that markedness is a good explanatory notion for antonym order. Kosti c (2015a) also discussed the reversibility of some antonym pairs with five examples and concluded that the reverse order is a consequence of the contextual need. She argued that markedness can be reversed in certain contexts, which leads to the reversal of antonym order. This can account for the reversibility of antonym order governed by markedness. However, it does not apply to antonym pairs whose preferred order is governed by the other two factors (visual-spatial ordering and gender) proposed by the author, such as “early/late”. With regard to antonym order in Mandarin Chinese, Shu and Huang (2008) studied the preferred order in Chinese antonymous compounds. Their analysis has shown that the order of antonyms mainly follows the Pollyanna Principle,2 but under certain conditions this principle is subject to phonological constraints. Hsu (2015) examined antonym order in Chinese contrastive constructions3 and concluded that the broadly defined notion of positivity is the dominant factor that influences antonym order in Chinese. He proposed that positivity can be broadly defined in terms of cultural values. For instance, gender can be regarded as a factor reflecting the traditional socio-cultural bias in which men are associated with positive characters, so the male terms all precede their respective counterparts when they co-occur. This definition makes positivity an encompassing notion which lacks explanatory force. It is plausible that ‘positive’ is prototypical as against ‘negative’ and positive words are usually put before negative ones. However, it is often difficult to decide which word of antonym pairs such as soft/hard and morning/evening is to be seen as positive. Therefore, an explanation of antonym order only in terms of positivity cannot be sustained. Despite the fact that so many factors governing antonym order have been suggested in previous research, few studies on antonym order have been conducted from a cognitive point of view. A particular concern for cognitive linguists is discovering the cognitive principles that motivate linguistic structures. What are the specific cognitive principles underlying antonym ordering? How do they affect antonym order? Kosti c (2015a) has pointed the way to the fact that antonym ordering reflects general cognitive principles (figure-ground alignment). Following her suggestive findings, this study attempts to make a cognitive study of antonym order in Chinese four-character patterns. We assume that the order of antonym pairs is based on general cognitive principles that reflect our preferred ways of conceptualizing the world. 3. Antonym order in Chinese four-character patterns 3.1. Materials and methods The main aim of this study is to examine antonym order in four-character patterns and investigate the principles that govern it. This study consists of four steps. First, a search for four-character patterns with antonym pairs was conducted in the authoritative dictionaries. The first dictionary used in this study is The Chinese Four-character Patterns Dictionary (Jiang, 2000),

2 The Pollyanna Principle was put forward by Matlin and Stang (1978), which means that the positive or evaluatively more positive members are put before the negative ones. 3 According to Jones et al. (2012: 108), contrastive constructions are the lexico-grammatical frames which are used to unite and neutralize contrasting concepts. Any two words co-occurring in these constructions can get a contrastive interpretation.

120

S. Wu / Language Sciences 59 (2017) 117–134

which contains 4292 entries of four-character patterns. This is the only current dictionary which specifically contains Chinese four-character patterns. We selected all the patterns with antonym pairs and checked the order of antonyms within them. 738 patterns with 143 distinct antonym pairs were identified. To compensate for the limitations of having only one reference, we supplemented our data by using The Contemporary Chinese Dictionary4 (2012, the 6th edition). This is the most authoritative and representative Chinese language dictionary containing 69,232 entries of words and phrases. We first selected all of the four-character patterns in the dictionary. The search yielded 6334 patterns. From these patterns we identified 379 fourcharacter patterns with 122 co-occurring antonym pairs. The statistics on antonym order in four-character patterns in these two dictionaries are given in Appendices I and II. A close examination of the data collected from the above dictionaries suggest that a number of antonym pairs occur only once in four-character patterns. Antonym pairs occurring five times or more make up no more than 30% of the total number of four-character patterns, while more than 50% of the antonym pairs in two dictionaries are hapax legomena, occurring only once. In view of the fact that the infrequent use of most antonym pairs will result in poor statistics, we also used the online dictionary (http://cy.5156edu.com),5 which contains 41,833 idioms.6 Therefore, the second step is to conduct a search for fourcharacter patterns in the online dictionary by using the antonym pairs which have at least two occurrences in the above two dictionaries. The search yielded 1822 four-character patterns with antonym pairs. We excluded the antonym pairs that occur less than five times in the data. After the exclusion, the complete list of high-frequency antonym pairs in four-character patterns consists of 68 pairs.7 The frequency of normal order, the frequency of reverse order, and the binomial score of these pairs are examined. The third step is to analyze the preferred order of antonym pairs in four-character patterns. The fourth step is to explore the cognitive principles underlying the preferred antonym order. 3.2. The preferred order in four-character patterns This section will examine whether antonym pairs in four-character patterns tend to favor a particular order and how marked the tendency is. Although the frequency of antonym pairs from the first two dictionaries tends to be low, the tendency of antonym order in Chinese four-character patterns is demonstrated by the 15 most frequently used antonym pairs in the data from the two dictionaries. The search conducted in The Chinese four-character dictionary yielded 143 antonym pairs. The 15 most frequently used antonym pairs are listed in Table 1 in descending order. Columns 1 and 2 present the antonym pairs in Chinese and English; Column 3 provides the total number of patterns for each pair; Column 4 presents the number of patterns which follow the normal order (A > B); Column 5 provides the number of patterns with the reverse antonym order (B > A); Column 6 presents the percentage of the patterns in which normal order is observed, and Column 7 provides information on statistically significant preference towards one of the two possible orderings based on binomial test. As shown in Table 1, the frequency of antonym pairs with the reverse order is very low. Of the 15 most frequently occurring antonym pairs, 12 pairs never occur in reverse order. Only three pairs can co-occur in reverse order: da/xiao ‘big/small’, shang/ xia ‘up/down’ and you/wu ‘with/without’. However, da ‘big’ occurs before xiao ‘small’ in 95.2% of patterns, shang ‘up’ occurs before xia ‘down’ in 96.4% of patterns, and you ‘have’ precedes wu in 92.3% of the patterns. This indicates that these pairs are reversible but demonstrate a strong preference for one order. Judging from the results of binomial test, all 15 pairs exhibit a statistically significant preference for a particular order (p < 0.05). The 15 most frequently used antonym pairs in The Contemporary Chinese Dictionary are presented in Table 2. As seen from the percentage of normal order, seven antonym pairs achieve a 100% score. The pairs are shou/jiao ‘hand/foot’, qian/hou ‘front/ back’, shang/xia ‘up/down’, dong/xi ‘east/west’, ming/an ‘light/dark’, zuo/you ‘left/right’, and nan/nü. They occur in the normal order in every pattern in this dictionary. In addition, there are eleven pairs whose binomial scores are below 0.05. This represents a statistically significant preference for the normal sequence, which indicates that these antonym pairs show a marked preference for a certain order in four-character patterns. There are only three pairs whose binomial test score is above 0.05. They cannot be said to show a tendency for the preferred order. This provides further evidence that antonym pairs which co-occur in Chinese four-character patterns exhibit a tendency to occur in a preferred order. The statistics of antonym order for 68 high-frequency antonym pairs in the online dictionary are presented in Table 3. High-frequency antonym pairs refer to those pairs occurring more than five times. Pairs are listed in descending order of p value. As seen from the binomial scores in the table, out of the 68 pairs, 44 pairs (64.7%) show a statistically significant preference toward one of the possible orderings (p < 0.05). Among the 44 pairs, the binomial tests of the first 25 antonym pairs generate a highly significant outcome. They do not reverse their order in every pattern in the data. This indicates that the order which they favor is statistically significant to a very high degree. The remaining 24 pairs, whose p-value is above 0.05, cannot be said to have any preference towards either of the two orderings. The data presented in Table 3 demonstrates that most antonym pairs exhibit a preferred order in four-character patterns.

4

I am indebted to Prof. Renqiang Wang, who created the electronic edition of this dictionary with his project team and shared it with me. This online Idiom Dictionary was established by 5156 Middle School Team in 2005. It covers a large database and is used widely online in China (Personal communication with the editors in 2016). 6 According to Ma (1978), the boundary between four-character patterns and four-character idioms is fuzzy. In this study, four-character patterns are defined as idiomatic or non-idiomatic four-character expressions of a (more or less) fixed nature, which include four-character idioms. 7 This study deals with highly frequent antonym pairs for the pragmatic reason proposed by Mollin (2014: 72): “only the most frequent pairs allow for a reliable identification of the two possible orders in the preferred one”. 5

S. Wu / Language Sciences 59 (2017) 117–134

121

Table 1 The top fifteen antonym pairs retrieved from The Chinese Four-character Patterns Dictionary. Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

Normal order (%)

p Value

tian/di lai/qu dong/xi qian/hou nan/nü zuo/you shang/xia si/huo da/xiao ming/an gu/jin chang/duan nan/bei tou/wei you/wu

sky/earth come/go east/west front/back man/woman left/right up/down dead/alive big/small light/dark ancient/modern long/short south/north head/tail with/without

85 51 50 41 36 28 29 22 21 19 19 19 16 15 13

85 51 50 41 36 28 28 22 20 19 19 19 16 15 12

0 0 0 0 0 0 1 0 1 0 0 0 0 0 1

100% 100% 100% 100% 100% 100% 96.4% 100% 95.2% 100% 100% 100% 100% 100% 92.3%

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.003

Based on the above analysis, it can be concluded that the majority of high-frequency antonym pairs tend to favor a particular order. The following section will examine the underlying principles governing the preferred order over the reverse order. 4. The cognitive principles governing antonym order in four-character patterns The majority of antonym pairs in Chinese four-character patterns display a preferred order in our data. What principles govern the order of the antonyms in the patterns? Why do antonyms tend to be used in a particular order? Previous studies have provided evidence that linguistic forms are to a certain extent iconic with the way humans beings conceptualize the world. An acceptable formulation of the basic assumption of the iconicity principle in syntax is that “linguistic forms are frequently the way they are because, like diagrams, they resemble the conceptual structure they are used to convey” (Haiman, 1985: vii). (Diagrammatic) iconicity has been used in various functional and cognitive explanations of linguistic structures (e.g. Givón, 1985; Haiman, 1980, 1985; Dressler, 1995; Diessel, 2008). It is regarded as “an important factor in the constitution of grammar” (Langendonck, 1995: 79). A major concern in grammar is word order. Many studies have demonstrated the role of iconicity in English word order. For instance, iconic sequencing is likely to be the most frequently mentioned ordering principle accounting for binomial sequencing (e.g. Malkiel, 1959; Cooper and Ross, 1975; Benor and Levy, 2006; Mollin, 2014). As for iconicity in Chinese word order, Tai (1985) argues that the principle of temporal sequence (PTS) plays a large role in Chinese grammar. He also pointed out that “Syntactic constructions in Chinese are to a great extent iconically motivated” (Tai, 1993: 169). Wen (2001) gave a contrastive study of the English freeze “right and left” and the Chinese compound zuoyou (right-left) and concluded that the word order is mainly governed by the principle of iconicity. All these studies have signified that there is a strong correlation between iconicity and word order. Accordingly, it is hypothesized that the iconicity principle may be an important determinant of antonym order. The following will address the effect of iconicity on antonym ordering based on an analysis of Chinese four-character patterns.

Table 2 The top fifteen antonym pairs retrieved from The Contemporary Chinese Dictionary. Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

Normal order (%)

p Value

tian/di you/wu shou/jiao qian/hou da/xiao chang/duan qing/zhong shang/xia sheng/si dong/xi ming/an zuo/you yi/tong nan/nü duo/shao

sky/earth with/without hand/foot front/back big/small long/short light/heavy up/down life/death east/west light/dark left/right different/similar man/woman much/little

31 16 16 13 11 11 10 10 9 9 8 8 7 7 6

30 15 16 13 10 8 7 10 8 9 8 8 5 7 4

1 1 0 0 1 3 3 0 1 0 0 0 2 0 2

96.8% 93.3% 100% 100% 90.9% 66.7% 70% 100% 88.9% 100% 100% 100% 71.4% 100% 66.7%

0.000 0.001 0.000 0.000 0.012 0.227 0.344 0.000 0.039 0.000 0.000 0.000 0.453 0.000 0.688

122

S. Wu / Language Sciences 59 (2017) 117–134

Table 3 The 68 high-frequency antonym pairs retrieved from the online idiom dictionary. Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

p Value

dong/xi gu/jin jin/tui qian/hou shi/fei shang/xia nan/bei gong/si chang/duan nan/nü zhao/xi wen/wu zuo/you e/shan shou/jiao jin/yuan qiang/ruo sheng/si xian/hou ci/bi de/shi si/huo zong/heng tian/di zhao/mu lai/qu shi/zhong er/nü gao/di ben/mo li/wai bei/xi yin/yang jian/zhong da/xiao lai/wang shou/zu gan/shi gan/ku lao/shao tou/wei qu/shen xian/yi ming/an qing/zhong duo/shao cun/wang bao/bian xin/jiu yi/tong fei/shou shen/qian rou/gang gu/xin sheng/shu fang/yuan xiong/ji chu/ru hei/bai nei/wai you/wu mǎi/mài lao/yi fei/xing ci/xiong ruan/ying han/nuan ku/tian

east/west ancient/modern advance/retreat front/back right/wrong up/down south/north public/private long/short man/woman morning/evening civil/military left/right evil/good hand/foot near/far strong/weak life/death early/late this/that gain/loss dead/alive vertical/horizontal sky/earth morning/evening come/go beginning/end son/daughter high/low root/branch inside/outside sadness/happiness yin/yang shoulder/heel big/small come/go hand/foot dry/wet sweet/bitter old/young head/tail bend/extend danger/safety light/dark light/heavy much/little survive/perish praise/disparage new/old different/similar fat/thin deep/shallow softness/hardness old/new raw/cooked square/round calamity/blessing out/in black/white inside/outside with/without buy/sell work/rest abolish/establish male/female soft/hard cold/warm bitter/sweet

103 79 67 53 52 43 36 36 33 32 30 29 28 28 27 23 22 21 21 20 17 21 12 56 42 32 32 16 15 33 13 8 14 13 24 38 23 6 15 9 16 10 7 19 39 16 16 6 36 26 10 6 15 19 6 27 8 38 24 27 15 15 11 8 7 7 6 6

101 58 65 47 46 42 29 33 28 28 26 26 27 18 23 14 16 14 19 15 17 21 12 49 32 25 31 15 14 25 11 8 12 11 18 28 19 6 12 8 14 8 6 15 24 13 13 5 14 10 7 3 6 11 4 15 5 25 11 13 7 8 6 4 4 3 3 3

2 21 2 6 6 1 7 3 5 4 4 3 1 1 4 9 6 7 2 5 0 0 0 7 10 7 1 1 1 8 1 0 2 2 6 10 4 0 3 1 7 2 1 4 15 3 3 1 22 16 3 3 9 8 2 12 3 13 13 14 8 7 5 4 3 4 3 3

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.001 0.001 0.001 0.001 0.005 0.007 0.008 0.013 0.022 0.023 0.005 0.030 0.031 0.035 0.039 0.04 0.109 0.125 0.019 0.200 0.021 0.021 0.219 0.243 0.327 0.344 0.375 0.607 0.648 0.688 0.701 0.727 0.73 0.839 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000

S. Wu / Language Sciences 59 (2017) 117–134

123

4.1. The principle of iconicity and its sub-types The notion of iconicity has become very popular among functional and cognitive linguists, but it subsumes various meanings.8 Langendonck (1995, 2007) proposed iconicity of closeness9 as an overall principle to describe various kinds of word order iconicity. Iconicity of closeness indicates what is nearest to the speaker in a literal (physical) or in a metaphorical sense is mentioned first. His argument about closeness is virtually the same as the ‘me-first’ principle10 in Cooper and Ross (1975) and the egocentric or canonical perception in Lyons (1977). They all claim that the main criterion of semantic ordering rules which governs sequential choice is our egocentric perception of ourselves. Iconicity of closeness, like an umbrella concept, subsumes a lot of semantic factors on word order under one superordinate. However, the greatest problem with iconicity of closeness as a universal explanation is that it doesn’t hold up in every case. Iconicity of temporal sequence cannot be subsumed to iconicity of closeness. For instance, in zaowan ‘morning/evening’, zao ‘morning’ precedes wan ‘evening’ as a consequence of iconic sequencing. We cannot claim that the first element zao ‘morning’ is more closely related to the prototypical speaker than wan ‘evening’. Therefore, iconicity of temporal sequence should be listed separately, which is parallel to iconicity of closeness. In addition, iconicity of closeness does not take into account social and cultural factors. It cannot account for the tendency to order elements according to the real-life social dominance of the referents. For instance, it cannot convincingly explain the gender bias. It is implausible to argue that the prototypical speaker is male instead of female. A great deal of literature has dealt with social and cultural factors in word order. Malkiel (1959: 145) pointed out that the priorities inherent in the structure of a society influence the order of words in binomials. Quirk et al. (1985: 971) suggested that “other things being equal, the first position is given to the semantically salient or culturally dominant member”. Givón (1995: 55) put forward the pragmatic principle of linear order: “more important or more urgent information tends to be placed first in the string”. However, what constitutes more powerful and more important is not made clear in these studies. For example, it is widely accepted that the element with positive connotation tends to precede the one with negative connotation when they co-occur, but what is considered positive or negative for any particular speech community is subject to cultural construction. All these points suggest that the order of the linguistic units reflects the order of the concepts they refer to in a given culture. We lump these under the heading ‘iconicity of cultural values and norms’.11 Specifically, the order of the linguistics elements tends to correspond to the order of the concepts they denote in a specific culture. In a nutshell, the principle of iconicity involved in this study subsumes three sub-types: iconicity of closeness, iconicity of temporal sequence and iconicity of cultural values and norms. 4.2. Iconicity and antonym order in Chinese four-character patterns 4.2.1. Iconicity of closeness and antonym order Iconicity of closeness here refers to the closeness to the prototypical speaker as suggested by Langendonck (1995). It indicates that the element whose referent is closer to the prototypical speaker is prone to precede the other when they cooccur. The prototypical speaker is characterized by being “here, now, adult, male, positive, singular, living, friendly, agentive, powerful, at home, and patriotic, among other things” (Cooper and Ross, 1975: 67). The closeness to the prototypical speaker could be elaborated in a literal or metaphorical sense. In the literal sense, iconicity of closeness indicates that linguistic expressions tend to imitate our natural egocentric perceptual serialization. The constituent denoting the concept which is closer to the speaker physically will be put before the one which is farther away from the speaker. Its cognitive basis is bodily experience. Antonym pairs which follow this principle often convey spatial-temporal meaning, such as the Chinese equivalents of up/down, vertical/horizontal and front/back. Humans perceive themselves as upright beings moving over the surface of the earth. A person acting in a three-dimensional space, stands upright, has eyes in his head and not in his feet, and looks forward rather than backward. He has his principal organs of perception directed towards the region in front of him, so the elements which refer to up generally precede those which refer to down. He moves in the direction in which he is facing, so the elements which denote front often precede those which denote back. Similarly, when these antonym pairs co-occur in the patterns, front precedes back and up precedes down. In our data, in every one of 53 patterns involving qian/hou ‘front/back’, qian precedes hou, such as zhan qian gu hou (look front watch back ‘be overcautious and indecisive’). As for shang/xia ‘up/down’, 42 out of 43 patterns put shang before xia, such as hu shang hu xia (sudden up sudden down ‘up and down’) and tiao shang tiao xia (jump up jump down ‘jump up and down’).

8 Haspelmath (2008) identified eight subtypes of (diagrammatic) iconicity from the previous research: iconicity of quantity, complexity, cohesion, paradigmatic isomorphism, syntagmatic isomorphism, sequence, contiguity and repetition. 9 Givón (1995: 51) proposed the term ‘proximity principle’ to convey the idea that the formal distance between linguistic signs is determined by the conceptual distance between them. The notion ‘closeness’ here refers to the salience and accessibility to the prototypical speaker literally or mentally rather than the correspondence between the linguistic distance and the conceptual distance. 10 Cooper and Ross (1975) organized the semantic constraints on word order into nineteen categories and subsume all of them to one umbrella principle, called ‘me first’ principle. The principle says that ‘first conjuncts refer to those factors that describe the prototypical speaker’. The ‘me-first’ principle is another way of describing the phenomenon of closeness. 11 Wen (2001) mentioned iconicity of culture and convention when he contrasted the Chinese zuoyou and their English equivalents left/right, but he did not illustrate it in detail.

124

S. Wu / Language Sciences 59 (2017) 117–134

This indicates that the shang > xia order predominates over the reverse order. In 15 patterns involving gao/di ‘high/low’, there are 14 patterns in which gao appears before di, such as bu gao bu di (not high not low ‘neither high nor low’) and hu gao hu di (sudden high sudden low ‘not stable’). The priority of shang ‘up’ over xia ‘down’ can extend to certain classes of concrete referents. For example, tian/di ‘sky/earth’ can be seen as the extension of shang/xia ‘up/down’. In 56 patterns which contain tian/di, tian precedes di in 49 patterns, such as ku tian dong di (cry sky move earth ‘cry bitterly’) and hun tian hei di (dizzy sky dark earth ‘in pitch darkness’). Our data also indicate that those antonyms which refer to up generally precede those referring to down in the referents for body parts. Take tou/wei ‘head/tail and tou/jiao ‘head/foot’ for example. Tou ‘head’ tends to precede wei ‘tail’ and jiao ‘foot’ in four-character patterns, such as cong tou zhi wei (from head to tail ‘‘from the beginning to the end’), mei tou mei wei (no head no tail ‘without beginning and end’) and shu tou shu jiao (relax head relax foot ‘relax completely’). Notably, the order of these antonyms is maintained even when their meaning is transferred from the spatial domain to other domains. Spatial words in Chinese are often transferred metaphorically into the temporal domain in a systematic way (Yu, 1998: 112). For instance, qian/hou ‘front/back’ can be exploited to express time. According to Radden (2011), there are two conflicting thinking models in Chinese. On one hand, speakers of Chinese conceive of the past as lying in front and the future as lying behind. For instance, qian tian (front day ‘the day before yesterday’) and hou tian (back day ‘the day after tomorrow’), in which qian ‘front’ is used to denote the past and hou ‘back’ is used to express the future. On the other hand, Chinese speakers also conceive of the future as lying in front and the past lying behind. Examples include qian tu (front road ‘future’) and qian cheng (front journey ‘future’), in which qian ‘front’ is used to denote the future. However, our data only contain examples in which qian ‘front’ is used to denote the past and hou ‘back’ is used to denote the future. Examples include cheng qian bi hou (punish past avoid future ‘learn from past mistakes and avoid future ones’), kong qian jue hou (surpass the past extinct future ‘unprecedented’) and si qian xiang hou (think past consider future ‘consider over and over again’). Though the spatial antonym pair has been transferred into the temporal domain, the sequence of qian > hou ‘front > back’ is maintained in four-character patterns. Similarly, the spatial antonyms shang/xia ‘up/down’ can be used to express actions. For instance, in neng shang neng xia (able up able down ‘be ready to accept a higher or lower post’), shang ‘up’ and xia ‘down’ denote ‘taking a higher post’ and ‘taking a lower post’ respectively, but their order is consistent with the order when they express spatial relation. Apart from spatial antonym pairs, there are some verb antonyms in our data which embody iconicity of closeness. For example, in lai/qu ‘come/go’ and lai/wang ‘come/go’, lai ‘come’ means coming close to the speaker, and qu ‘go’ and wang ‘go’ mean being far away from the speaker. Therefore, lai is prone to be put before qu and wang when they co-occur. In our data, in 32 patterns involving lai/qu ‘come/go’, there are 25 patterns with lai preceding qu. In the 38 patterns involving lai/wang ‘come/go’, 28 patterns put lai before wang. Iconicity of closeness in the metaphorical sense indicates that the element denoting the concept which is more salient to the speaker mentally or cognitively tends to be put first. Its cognitive basis is perceptual prominence. In human experience, salient concepts tend to come to mind first. “Our perceptual apparatus is geared towards things in our immediacy and presence, things which are large and big, and things which form good gestalts, have clear boundaries and are specific instances.” (Radden and Kövecses, 1999: 46-47). Langacker (1993: 30) mentioned the following as the principle of cognitive salience: human before non-human, whole before part, concrete before abstract, and visible before non-visible. Therefore, what is more perceptible and cognitively salient is more accessible to the speaker, which leads to its being placed first in word ordering. Take positive/negative as an example: a speaker “has a positive image of himself” (Mayerthaler, 1988: 10). Positive concepts are more favored than negative ones and therefore tend to be put first. Accordingly, if one member of an antonym pair has a positive connotation, it tends to be placed first when they co-occur. In four-character patterns, the Chinese equivalents of antonym pairs such as light/dark, victory/failure, true/false, advantage/disadvantage and alive/dead all put the word with positive associations first. Another example is dimensional adjective antonyms. More of something is usually more perceptually prominent and accessible than less of something to the speaker. Therefore, the element denoting more tends to precede the element denoting less in antonym pairs. For example, the order of antonym pairs such as ‘much > little’, ‘thick > thin’, ‘long > short’, ‘big > small’ predominates over their reverse order in English. The Chinese equivalents of the above pairs also follow this order when the pairs co-occur in four-character patterns. According to our data, in every one of 21 patterns involving chang/ duan ‘long/short’, chang comes before duan. 13 out of 16 patterns involving duo/shao ‘much/little’ all put duo before shao. Similarly, in 24 patterns involving da/xiao ‘big/small’, there are 18 patterns which put da before xiao. The effect of iconicity of closeness on antonym ordering can be stated as follows:  The order of antonym pairs tends to correspond to the closeness of the concepts they denote to the prototypical speaker, whether literally or metaphorically.

4.2.2. Iconicity of temporal sequence and antonym order Iconicity of sequence means that “the order of elements in language parallels that in physical experience or the order of knowledge” (Greenberg, 1966: 103). Haiman (1980) has shown that iconicity of sequence is the most widespread iconic motivation. The most prominent aspect of iconicity of sequence is temporal ordering, which indicates that “the relative word

S. Wu / Language Sciences 59 (2017) 117–134

125

order between syntactic units is determined by the temporal order of the states which they represent in the conceptual world” (Tai, 1985: 50). In other words, iconicity of sequence means that the linguistic order of elements echoes the temporal sequence in non-linguistic reality. With regard to antonym pairs, temporally related antonyms tend to be chronologically ordered. Specifically, if the concept one antonym denotes is prone to precede the other in the real world, that order will be reflected in syntactic frames when they co-occur. Take sheng/si ‘life/death’ for example. Life precedes death temporally in the physical world, so this sequence is echoed in the linguistic order. Chinese people say sheng si yu gong (life death with share ‘go through thick and thin together’) instead of si sheng yu gong (death life with share). There are some other antonym pairs which indicate a temporal sequence, such as shizhong ‘beginning/end’, zhao/mu ‘morning/evening’, zhao/xi ‘morning/evening’, gu/jin ‘ancient/modern’ and chun/ qiu ‘spring/autumn’. In non-linguistic reality, the concepts that these pairs denote indicate an obvious temporal order, which is mirrored in the linguistic structures. This temporal pattern may account for much of the antonym ordering in four-character patterns. See some examples below: (2)

gu/jin ‘ancient/modern’: gu wang jin lai (ancient go modern come ‘ageless’) bo gu tong jin (know ancient master modern ‘well-informed’)

(3)

zhao/mu ‘morning/evening’: zhao sheng mu si (morning alive evening dead ‘ephemeral’) zhao san mu si (morning three evening four ‘indecisive’)

(4)

shizhong ‘beginning/end’: shi zhong ru yi (beginning end like one ‘be consistent’) zi shi zi zhong (from beginning from end ‘from beginning to end’)

(5)

xin/jiu ‘new/old’: bu xin bu jiu (not new not old ‘neither old nor new’), xin jiu jiao cuo (new old interlock ‘the new things and the old things interlace’)

In the above examples, the order of co-occurring antonym pairs reflects the temporal sequence of the concepts they denote in physical experience. What happens earlier in time always precedes what comes later. What is noteworthy is the pair xin/jiu ‘new/old’ in (5). New things emerge and get old with the passage of time. This sequence is mirrored in the linguistic expressions. In our data, most patterns involving xin/jiu ‘new/old’ observe the new > old sequence. However, when the speaker wishes to emphasize getting rid of old things and welcoming the new things, the default order will be reversed, such as in ci jiu ying xin (bid farewell to old welcome new ‘ring out the Old Year and ring in the New Year’) and qi jiu ying xin (discard old welcome new ‘turn over a new leaf’). Though the order of xin/jiu ‘new/old’ is reversed in these examples, it still observes the iconicity of temporal sequence, for they focus on the process of change from old to new and the creation of new things is based on getting rid of the old. Therefore, if two elements are perceived as existing in a chronological sequence, they tend to follow that sequence when they co-occur. Iconicity of order primarily concerns the order of temporally related words. Some antonym pairs, such as verbs denoting events or actions, are not temporal words, but imply a temporal relationship. The element referring to an event or action earlier in time will precede the other in linguistic expressions. Take sheng/jiang ‘rise/fall’ for example. They are not temporal words, but they indicate a temporal order in the sense that the action rise happens earlier in time than the action fall. In the physical world, every day the sun rises in the east and sets in the west, which is an invariable natural law. Therefore, when they co-occur, their order corresponds to the temporal order of the events or actions they represent. Sheng ‘rise’ tends to precede jiang ‘fall’. Examples in four-character patterns include ming sheng an jiang (open rise secret fall ‘kick someone upstairs’) and hu sheng hu jiang (sudden rise sudden fall ‘up and down’). Iconicity of temporal sequence is a very important factor in determining antonym sequence. Most of the patterns which follow this principle are symmetric constructions in our data. The principle also holds true in serial verb constructions. When two verbal phrases containing antonyms express consecutive actions, they are ordered according to their temporal order in the conceptual world. In bi shi jiu xu (avoid strong approach weak ‘be evasive’), the action bi shi ‘keep clear of the main force’ happens before jiu xu ‘strike at the weak points’. In addition to the consecutive actions, there is also resultative reading in serial verb constructions. Actions usually precede the outcome temporally. This order is mirrored in word order. For example, in sun gong li ji (harm public benefit private), the action sun gong ‘do harm to the public’ precedes the outcome li ji ‘benefit oneself’. The antonym order reflects the causal order of the concepts that they denote. Based on the above analysis, the effect of iconicity of temporal sequence on antonym ordering can be stated as follows:  The order of antonym pairs tends to correspond to the temporal sequence of the concepts they denote in the real world.

126

S. Wu / Language Sciences 59 (2017) 117–134

4.2.3. Iconicity of cultural values and norms and antonym order Iconicity of cultural values and norms indicates that the order of the linguistic elements corresponds to the order of the concepts they denote in a specific culture.12 With regard to antonym order, it means that in an antonym pair, the term that denotes a socially more highly ranked referent precedes the other when they co-occur. In the Chinese culture, the key concept of Confucianism, lunli (the ethic principle) has a deep-rooted influence on Chinese people. Lun (ethic) in Chinese means the hierarchical order. It is a shared belief that one of the foundational characteristics of Chinese culture is hierarchical relationship (Chen and Chung, 1994). Chinese culture stresses the distinction between the noble and the humble, between the superior and the inferior, between the elder and the younger. Social hierarchy is reflected in antonym order. In other words, antonym pairs tend to be ordered in accordance with the hierarchic order of their referents in a society. The antonym with higher social status tends to precede the one with lower status. This principle can account for the order of numerous antonym pairs in Chinese. The first category is gender asymmetry in antonym co-occurrence. The typical example is nan/nü ‘man/woman’. The traditional ethic that men are superior to women in Chinese culture is deeply rooted. Starting from the Han period onward, Confucians began to gradually teach women that a virtuous woman was supposed to follow the lead of the males in her family, especially the father before her marriage and the husband after she gets married. In traditional Chinese families, man is the symbol of power, status and dominance. The wife’s status is subordinate to her husband. Nowadays, the social status of women has improved greatly, but the hierarchical order is deeply rooted. Inequalities in the relative social status of males and females are reflected in language. The order of male > female in linguistic expressions accords with the dominance of men and the subordination of women. Therefore, the male > female sequence is observed in the examples such as nan/nü ‘man/ woman’ and er/nü ‘son/daughter’. In our data, 28 out of 32 patterns involving nan/nü ‘man/woman’ put nan ‘man’ before nü ‘woman’, such as gu nan gua nü (lonely man few woman ‘bachelors and spinsters’), and sheng nan yu nü (bear man raise woman ‘give birth to children’). 15 out of 16 patterns with er/nü ‘son/daughter’ pairs put er ‘son’ before nü ‘daughter’, such as you er you nü (have son have daughter ‘have kids’) and mei er mei nü (without son without daughter ‘have no kids’). The second category is age asymmetry. In Chinese tradition, respect for the elderly is considered a virtue. Old age implies more power and higher social status. The elderly are highly respected because of their experience and wisdom. Age hierarchy is reflected in antonym ordering. Chinese speakers exhibit a tendency to put the words referring to older people before the words referring to the young. For instance, in lao/shao ‘old/young’, lao/you ‘old/little’, mu/nü ‘mother/daughter’ and er/sun ‘son/ grandson’, the first member of these pairs tends to appear before the second member. This order is followed in four-character patterns such as bu lao bu shao (not old not young ‘be middle-aged’) and you lao you shao (have old have young ‘of all ages’). Hierarchical relations are not restricted to gender and age. Other instances can be found in antonyms indicating social status in a given culture. The antonym indicating higher social status precedes the antonym denoting lower social status. For example, in zhu/pu ‘master/servant’, masters definitely have higher social status than servants, so zhu ‘master’ is put before pu ‘servant’ as in ban zhu ban pu (half master half servant ‘act like a master as well as a servant’). In addition, antonym pairs tend to be ordered in accordance with a hierarchy of values inherent in the structure of a given culture. The one denoting the socially more favorable referent tends to precede the other which denotes the less favorable referent. In Chinese culture, yin yang (also yin-yang or yin and yang) is a basic principle (Min, 1995). It is used to describe how polar opposite or contrary forces are actually complementary, interconnected, and interdependent in the natural world. Natural dualities such as dark/light, female/male, low/high, cold/hot, earth/sky and passive/active are regarded as manifestations of yin yang. The relationship between yin and yang is often described in terms of sunlight playing over a mountain and in the valley. Yin (literally the ‘shady place’ or ‘north slope’) is the dark area occluded by the mountain’s bulk, while yang (literally the ‘sunny place’ or ‘south slope’) is the brightly lit portion. As the sun moves across the sky, yin and yang gradually trade places with each other, revealing what was obscured and obscuring what was revealed. “Underlying polarity of yang and yin begins with light vs. dark and extends into high vs. low, creative vs. receptive, firm vs. yielding, moving vs. resting, and feminine vs. masculine, but also into many other areas of human concern, including the sun and the moon, the weather, the parts and the body, and even the distinction between gods (all yang) and ghosts (all yin)” (Osgood and Richards, 1973: 380). The Chinese people attribute positive polarity to the yang and negative polarity to the yin. Yin is characterized as slow, soft, yielding, diffuse, cold, wet, and passive; and is associated with water, earth, the moon, femininity and nighttime. In contrast, yang is characterized as fast, hard, solid, focused, hot, dry, and aggressive; and is associated with fire, sky, the sun, masculinity and daytime. Therefore, the yang tends to be put before the yin when these concepts coordinate.13 For instance, gan ‘dry’

12 It is noteworthy that there are pan-cultural concepts in the world. Certain traditional cultural concepts in Chinese language do not lead to particular linguistic manifestations that are different from the western world. For instance, the order of male > female in Chinese is consistent with the order in English discovered by Cooper and Ross (1975). The idea of “superiority of male to female” and “the younger respecting the older” has become pan-cultural. However, some pan-cultural human concepts gain particular importance in a given culture and lead to different linguistic manifestations. 13 In Chinese compounds, the word with positive associations tends to precede the word with negative associations. However, the compound yinyang, meaning ‘dark and light; female and male; moon and sun’, is an exception, in which yin is put before yang. Scholars have proposed various explanations for this, including “linguistic convenience” (It is easier to say yinyang than yangyin), the idea that “proto-Chinese society was matriarchal”, or perhaps, since yinyang first became prominent during the late Warring States period, this term was “purposely directed at challenging persistent cultural assumptions” (Roger, 2002: 847). According to Cooper and Ross (1975), phonological constraints often outrank semantic ones, so the order of yinyang is probably phonologically motivated. The order /i/ >/æ/, originally /a/, occurs frequently in conjuncts, as in chitchat, tit for tat, zigzag, etc. This can account for the word order in yinyang, in which yin with /i/ precedes yang with /a/.

S. Wu / Language Sciences 59 (2017) 117–134

127

precedes shi ‘wet’, ming ‘light’ precedes an ‘dark’ and kuai ‘fast’ precedes man ‘slow’. This order is observed in Chinese fourcharacter patterns such as ban gan ban shi (half dry half wet), hu ming hu an (sudden light sudden dark), and shi kuai shi man (sometimes fast sometimes slow). All the examples have demonstrated that the yin yang culture has a great effect on social values, which is manifested in antonym ordering. The order of some spatial antonyms also mirrors cultural norms. Take zuo/you ‘left/right’ for example. Landsberg (1995: 74) claims that the dexterity criterion mentioned by Lyons (1977) governs the right-left linear choice that occurs ubiquitously in English. This implies a predominance of right-first usage. However, Schönefeld (2005) made a corpus study of the syntactic arrangement of right/left in English, German and Russian and concluded that the three languages show a left-first preference. He argues that the order of left and right is influenced by the culture of the people speaking the language. In our data, out of 28 patterns with co-occurring left and right, 27 patterns exhibit a left-first preference. Examples include zuo si you xiang (left think right consider ‘think over and over again’), zuo deng you deng (left wait right wait ‘wait for a long time’), and zuo pan you wen (left inquire right ask ‘inquire in detail’). In Chinese culture, the left side is regarded superior to the right side, which is attributable to one traditional aspect of Chinese custom. In China, it is preferred that the front of the house faces the north, and the back faces the south, so the left corresponds to the east side, and the right to the west side. As mentioned above, the sun rises from the east and sets in the west. This temporal sequence is mirrored in the order of east and west. East is always put before west when they co-occur in linguistic expressions. Accordingly, zuo ‘left’ tends to be put before you ‘right’. The order of zuo ‘left’ and you ‘right’ in the patterns reflects the cultural conventions in China. In a nutshell, all the above examples suggest that antonym pairs may be ordered in accordance with a hierarchy of cultural values and norms inherent in the structure of a given society. The effect of iconicity of cultural values and norms on antonym ordering can be illustrated as follows:  The order of antonym pairs tends to correspond to the order of the concepts they denote in a given culture. According to the above analysis, it can be concluded that iconicity of closeness, iconicity of temporal sequence, iconicity of cultural values and norms play an important role in antonym ordering in Chinese four-character patterns. Antonym order is found to be iconically motivated to a greater extent than has been assumed. 5. The reversibility of antonym pairs in Chinese four-character patterns Most antonym pairs in four-character patterns exhibit a preference for one particular order. However, the order of antonym pairs can be reversed in some patterns. For example, hei ‘black’ and bai ‘white’ can reverse their order in the same syntactic frame, such as yi bai wei hei (regard white as black ‘distort truth’) and yi hei wei bai (regard black as white ‘distort truth’). The reversibility of antonym order needs to be accounted for. We assume that the reversibility of these antonym pairs is attributed to the speaker’s viewpoint and the semantic meaning of the patterns. 5.1. Viewpoint and reversibility of antonym pairs Viewpoint phenomena are very pervasive in language. Langacker (1987: 122-126) proposes viewpoint as a focal adjustment with two subtypes: vantage point and orientation. A vantage point is the position from which a scene is viewed. Orientation pertains to alignment with respect to the axes of the visual field. In observing a complicated scene, we can assume different positions in relation to it. A particular choice of vantage point imposes on the scene an alignment into foreground and background. This will lead to varying perceptions of their relative proximity and salience. People tend to see most of the objects in their experience from a canonical alignment with respect to their surroundings. This viewing arrangement, common in everyday conversations, has default-case status. The default arrangement is an essential part of the conceptual substrate that supports an expression’s meaning and shapes its form. This viewpoint is reflected in conventional pictorial representations, which demonstrates the role of iconicity in determining linear word order. Departures from the default arrangement will result in non-iconic expressions. Different viewpoints explain the reversibility of antonym order in Chinese four-character patterns. Take li/wai ‘inside/ outside’ for example. The word order depends on the vantage point of the speaker. When the speaker is situated inside the container, his attention will move from the inside to the outside, which leads to the inside > outside order, such as in cong li dao wai (from inside to outside). When the speaker is outside the container, his attention moves from the outside of the container to the inside, which leads to the outside > inside word order, such as in cong wai dao li (from outside to inside). With regard to the same scene, different viewpoints result in different expressions. Similarly, the word order of da/xiao ‘big/small’ follows the principle of closeness. Da usually precedes xiao, but in some cases the order can be reversed and xiao can be put before da. For example, it is possible to say you xiao dao da (from small to big) and you da dao xiao (from big to small). This is attributed to different viewing arrangements. In the default arrangement, the vantage point is the actual location of the viewer. The big ones are more salient and accessible to the speaker, which results in the linguistic representation you da dao xiao (from big to small). However, the same scene can be observed and described from a different vantage point. When the speaker changes his vantage point, the small ones may be more salient and accessible. This construal results in the linguistics expression you xiao dao da (from small to big). Similarly, the order of jin

128

S. Wu / Language Sciences 59 (2017) 117–134

‘near’ and yuan ‘far’ is dependent on the speaker’s vantage point. It is possible to say you yuan ji jin (from far to near) and you jin ji yuan (from near to far). The departure from the default arrangement involves the relative position of the viewer. These patterns in which antonym pairs can reverse their order are mainly symmetric patterns and prepositional patterns as shown in the following. (6)

ci qi bi fu (here rise there fall) mei bai mei hei (no white no black) du lai du wang (single come single go) cong shang dao xia (from up to down) mǎi kong mài kong (buy empty sell empty)

ci fu bi qi (here fall there rise) mei hei mei bai (no black no white) du wang du lai (single go single come) cong xia dao shang (from down to up) mài kong mǎi kong (sell empty buy empty)

5.2. Semantic constraints and reversibility of antonym pairs The reversibility of antonym pairs in some four-character patterns is ascribed to the semantic meaning of the patterns. In other words, the order of antonym pairs is semantically determined to a great extent. A change of the order in which the antonyms appear creates a change in meaning. For instance, both li shao bi duo (advantage few disadvantage many ‘more disadvantages than advantages’) and li duo bi shao (advantage many disadvantage few more advantages than disadvantages) are appropriate in Chinese, but their meanings are completely opposite. The former indicates that there are more disadvantages than advantages, but the latter expresses the opposite meaning. Similarly, zhuan nu wei xi (change anger into happiness ‘chang from anger to happiness’) and zhuan xi wei nu (change happiness into anger ‘change from happiness to anger’) convey the completely different meaning. Other examples are listed below. (7)

cong wu dao you (from without to with) hua ling wei zheng (change parts into whole) fan qiang wei ruo (invert strong into weak) wai rou nei gang (outside soft inside hard)

cong you dao wu (from with to without) hua zheng wei ling (change whole into parts) fan ruo wei qiang (invert weak into strong) wai gang nei rong (outside hard inside soft)

6. Conclusion This paper explores antonym co-occurrence and the order of antonyms in Chinese four-character patterns. Based on the data from the dictionaries, the study has found that the antonym pairs in the patterns exhibit a tendency for a preferred order. The analysis reveals a correlation between the normal order of the antonyms and iconicity. More precisely, iconicity of closeness (to the prototypical speaker), iconicity of temporal sequence and iconicity of cultural values and norms conspire in motivating the order of antonyms in four-character patterns. The study also examines the issue of departing from the normal antonym order and proposes that the reversibility of antonym pairs is attributable to a change of viewpoint and semantic constraints imposed by the patterns. In contrast to the previous research on antonym order (e.g. Jones, 2002; Kosti c, 2015a; Hsu, 2015), this study made an attempt to examine the order of antonyms from a cognitive point of view. It has been found that iconicity is an important determinant of antonym order. The pervasiveness of iconicity in antonym ordering indicates that the order of co-occurring antonyms is determined by general cognitive principles. Moreover, this paper presents an analysis of antonym cooccurrence and antonym order in Chinese patterns, which may provide insight into Chinese grammar to a native Englishspeaking audience. Murphy (2003) calls for further comparative investigation into antonym use. It is hoped that this work will stimulate further research on antonym order in other languages and other types of Chinese antonym co-occurrence structures, on the idea that the cross-linguistic study of antonymy can shed further light on the role of binary opposition in human cognition and language. This study has shown that the principle of iconicity has a great explanatory value in antonym order in Chinese fourcharacter patterns, but it is by no means the only determinant. In some idiosyncratic cases, the principle is less applicable. Other factors such as prosodic factors and phonological factors also play a role in antonym ordering and may take precedence over iconicity under certain conditions (e.g. in yinyang). Moreover, the ranking of the subtypes of iconicity needs to be explored. These issues are beyond the scope of this paper but will be the focus of our future research. Acknowledgements I would like to sincerely thank three anonymous reviewers and the editor for their valuable and insightful comments. I am also grateful to the audience of the 6th UK Cognitive Linguistics Conference at Bangor University for their helpful feedback on the early version. I would like to give my gratitude to Prof. Renqiang Wang for sharing the electronic version of The Contemporary Chinese Dictionary with me. Special thanks go to my friend Michelle Canfield for providing the language help. Any remaining errors are my own.

S. Wu / Language Sciences 59 (2017) 117–134

129

Appendix I. Frequencies of antonym pairs in Chinese Four-character Dictionary.

Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

tian/di lai/qu dong/xi qian/hou nan/nü zuo/you shang/xia si/huo da/xiao ming/an gu/jin chang/duan nan/bei tou/wei you/wu xin/jiu lai/wang xian/hou er/nü nei/wai shi/fei gong/si sheng/si ci/bi gao/di ru/chu duo/shao zhao/mu gong/shou jin/yuan qing/zhong duan/xu song/jin hei/bai li/bi hao/huai jin/yin jin/tui qi/fu qiang/ruo yin/xian zao/wan shi/xu bai/sheng bao/bian bei/xi zhao/xi fan/jian gan/shi hou/bo ji/li kai/bi po/li qi/luo xi/nu xi/you xiong/ji cu/xi kuai/man lao/shao leng/re ling/zheng nan/yi nao/xiao sheng/jiang sheng/mie

sky/earth come/go east/west front/back man/woman left/right up/down dead/alive big/small light/dark ancient/modern long/short south/north head/tail with/without new/old come/go early/late son/daughter inside/outside right/wrong public/private life/death this/that high/low in/out much/little morning/evening attack/defend near/far light/heavy on/off tight/loose black/ white advantage/disadvantage good/bad gold/silver advance/retreat rise/fall strong/weak disappear/appear morning/evening real/unreal lose/win praise/disparage sorrow/joy morning/evening complicated/simple dry/wet thick/thin approach/leave open/close break/set rise/fall joy/anger hope/fear calamity/blessing thick/thin fast/slow old/young cold/hot part/whole difficult/easy anger/laugh rise/fall birth/death

85 51 50 41 36 28 29 22 21 20 19 19 16 15 13 13 13 10 10 10 10 9 9 8 7 7 7 6 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

85 51 50 41 36 28 28 22 20 20 19 19 16 15 12 9 13 10 10 10 10 8 7 8 7 5 7 6 1 1 3 4 2 2 2 3 3 1 1 2 3 3 2 2 2 2 2 2 2 2 2 2 2 2 1 1 2 2 2 2 2 1 1 1 2 2

0 0 0 0 0 0 1 0 1 0 0 0 0 0 1 4 0 0 0 0 0 1 2 0 0 2 0 0 3 3 1 0 2 1 1 0 0 2 2 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1 1 1 0 0 (continued on next page)

130

S. Wu / Language Sciences 59 (2017) 117–134

(continued ) Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

shui/xing tou/jiao tu/yang tun/tu wen/wu wen/da zhong/xi zhang/chi zhang/he bin/zhu chen/fu chen/xi cheng/hui fang/yuan fang/shou fen/he gan/ku heng/shu hun/su jin/man jin/chu ju/san peng/ma ping/zhi qie/yong qin/shu qin/zong qu/shen qu/fan qu/hui ri/ye ruan/ying san/mi mǎi/mài shang/fa shen/suo xian/dan xian/yi xian/mo ya/su yang/yi yi/tong yi/hai yin/yang you/lie cun/wang di/you chu/mo ke/zhu ku/le kua/bian lao/yi li/he li/wai li/heng li/hai ming/mie pang/shou shen/qian sheng/luo sheng/shu shuang/re tui/jiu wei/an wen/bai zhen/huan zhen/jia zheng/bi zhong/yang

sleep/wake head/foot indigenous/foreign swallow/spit civil/military question/answer Chinese/Western tension/relaxation open/close guest/host sink/float morning/evening create/destroy square/round release/capture separate/combine sweetness/bitterness horizontal/vertical meat/vegetable urgent/slow in/out assemble/scatter praise/curse flat/straight timid/brave close/distant catch/release bend/stretch go/come go/come day/night soft/hard loose/dense buy/sell reward/punishment extension/flexion salty/light danger/safety appear/disappear elegance/vulgarity promote/suppress different/similar benefit/harm yin/yang good/bad survive/perish foe/friend appear/disappear guest/host sorrow/joy praise/disparage work/rest separation/reunion inside/outside set/horizontal gain/loss appear/vanish fat/thin deep/shallow rise/fall raw/cooked cool/hot refuse/accept danger/safety classical/modern reality/fantasy true/false open/close Chinese/Western

2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

S. Wu / Language Sciences 59 (2017) 117–134

131

(continued ) Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

zhu/pu zhuan/pei zi/sun zong/heng zuo/xie zuo/wo zeng/jian zhang/luo

master/servant profit/loss son/grandson vertical/horizontal work/rest sit/lie increase/decrease rise/fall

1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0

Appendix II. Frequencies of antonym pairs in The Contemporary Chinese Dictionary. Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

tian/di you/wu shou/jiao qian/hou da/xiao chang/duan qing/zhong shang/xia sheng/si dong/xi ming/an zuo/you yi/tong nan/nü duo/shao si/huo yin/yang li/wai ci/bi lai/qu zhao/xi ci/xiong nan/bei shi/zhong shou/zu tou/wei gu/jin gao/di shi/fei xian/hou zhen/jia bai/hei xin/jiu zhao/mu fei/shou gong/si jin/tui li/bi mǎi/mài qi/luo nei/wai chen/xin ru/chu fang/yuan gan/ku ku/tian lao/shao ben/mo cun/wang de/shi e/shan

sky/earth with/without hand/foot front/back big/small long/short light/heavy up/down life/death east/west light/dark left/right different/similar man/woman much/little death/life yin/yang inside/outside this/that come/go morning/evening female/male south/north beginning/end hand/foot head/tail ancient/modern tall/short right/wrong early/late true/false white/black new/old morning/evening fat/thin public/private advance/retreat advantage/disadvantage buy/sell rise/fall in/out stale/fresh in/out square/round sweet/bitter bitter/sweet old/young root/branch survive/perish gain/loss evil/good

31 16 16 13 11 11 10 10 9 9 8 8 7 7 6 6 6 6 5 5 5 5 5 5 5 5 4 4 4 4 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2

30 15 16 13 10 8 7 10 8 9 8 8 5 7 4 6 5 6 0 4 5 5 5 5 5 5 3 4 4 4 0 2 2 3 3 3 3 3 3 3 0 1 1 1 1 1 2 2 2 2 2

1 1 0 0 1 3 3 0 1 0 0 0 2 0 2 0 1 0 5 1 0 0 0 0 0 0 1 0 0 0 3 1 1 0 0 0 0 0 0 0 2 1 1 1 1 1 0 0 0 0 0 (continued on next page)

132

S. Wu / Language Sciences 59 (2017) 117–134

(continued ) Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

fei/xing gu/xin han/nuan heng/shu heng/zhi hou/bao jian/zhong lai/wang leng/re qiang/ruo gang/rou xiu/qi yi/lao bi/ji de/yuan fu/qiong qing/zhuo shun/ni bei/kang bei/xi biao/li chen/mu cheng/bai chong/ru chu/mo dong/xia du/rou duan/lian fan/jian pi/tai fu/qi gao/ai gong/nong guang/bo han/lao hao/dai huo/fu ji/li jiang/fa ke/zhu kuan/zhai lao/tong lao/xiao le/bei lun/lei pian/quan po/li qi/fu qin/zong qu/shen ri/ye ruan/ying shen/qian shen/shu tui/jiu tun/tu wen/huo wen/da xian/yi xin/yi ya/shu yi/shun yin/guo you/lie zao/wan zeng/jian zhang/chi zhong/xi zi/nü

abolish/establish old/new cold/warm horizontal/vertical horizontal/straight thick/thin shoulder/heel come/go cold/warm strong/weak hardness/softness joy/sorrow relaxation/labor other/self good/evil rich/poor pure/turbid good/poor overbearing/servile sorrow/happiness exterior/interior morning/evening success/failure favour/disgrace appear/disappear winter/summer single/double disconnect/connect complex/simple misfortune/bliss husband/wife tall/short industry/agriculture extensive/small drought/flood good/bad weal/woe approach/leave reward/punishment guest/host wide/shallow old/young old/little happiness/sorrow this species/that class part/whole break/set rise/fall catch/release bend/stretch day/night soft/hard deep/shallow raw/cooked refuse/accept swallow/spit indifferent/enthusiastic question/answer dangerous/safe believe/doubt elegance/vulgarity benefit/harm cause/effect good/bad morning/evening increase/decrease tension/relaxation Chinese/Western son/daughter

2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

2 2 2 2 2 2 2 2 2 2 2 2 2 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

S. Wu / Language Sciences 59 (2017) 117–134

133

(continued ) Antonym pairs: A term/B term

English gloss

Total number of occurrence

Normal order (raw freq) (A > B)

Reverse order (raw freq) (B > A)

zhong/heng ku/xiao

vertical/horizontal cry/smile

1 1

1 1

0 0

References Benor, D.B., Levy, R., 2006. The chicken or the egg? A probabilistic analysis of English binomials. Language 82 (2), 233–278. Charles, W.G., Miller, G.A., 1989. Contexts of antonymous adjectives. Appl. Psycholinguist. 10, 357–375. Chen, G.M., Chung, J., 1994. The impact of Confucianism on organizational communication. Commun. Q. 42 (2), 93–105. Cooper, W.E., Ross, J.R., 1975. World order. In: Grossman, R.E., San, J.L., Vance, T.J. (Eds.), Papers from the Parasession on Functionalism. Chicago Linguistic Society, Chicago, pp. 63–111. Cruse, D.A., 1986. Lexical Semantics. Cambridge University Press, Cambridge. Davies, M., 2013. Opposition and Ideology in News Discourse. Bloomsbury Academic, New York. Diessel, H., 2008. Iconicity of sequence: a corpus-based analysis of the positioning of temporal adverbial clauses in English. Cogn. Linguist. 19 (3), 465–490. Dressler, W.U., 1995. Interactions between iconicity and other semiotic parameters in language. In: Simone, R. (Ed.), Iconicity in Language. John Benjamins, Amsterdam/Philadelphia, pp. 21–45. Fellbaum, C., 1995. Co-occurrence and antonymy. Int. J. Lexicogr. 8, 281–303. Givón, T., 1985. Iconicity, isomorphism and non-arbitrary coding in syntax. In: Haiman, J. (Ed.), Iconicity in Syntax: Proceedings of a Symposium on Iconicity in Syntax. John Benjamins, Amsterdam, pp. 187–219. Givón, T., 1995. Isomorphism in the grammatical code: cognitive and biological considerations. In: Simone, R. (Ed.), Iconicity in Language. John Benjamins, Amsterdam, pp. 47–76. Greenberg, J.H., 1966. Some universals of grammar with particular reference to the order of meaningful elements. In: Greenberg, J.H. (Ed.), Universals of Human Language. MIT Press, Cambridge, Mass, pp. 73–113. Guo, S.Y., 1979. A New Perspective on Chinese Grammar and Rhetoric. Commercial Publishing Press, Beijing. Haiman, J., 1980. The iconicity of grammar: isomorphism and motivation. Language 56 (3), 515–540. Haiman, J., 1985. Iconicity in Syntax. John Benjamins, Amsterdam. Han, C.Q., 2008. On Chinese four-character phrases with synonymous symmetrical structure and their lexicalization. Chin. Lang. Learn. 1, 57–62. Haspelmath, M., 2008. Frequency vs. iconicity in explaining grammatical asymmetries. Cogn. Linguist. 19 (1), 1–33. Hsu, C.-C., 2015. A syntagmatic analysis of antonym co-occurrences in Chinese: contrastive constructions and co-occurrence sequences. Corpora 10 (1), 47– 82. Jeffries, L., 2010. Opposition in Discourse: the Construction of Oppositional Meaning. Continuum, London/New York. Jiang, D.W., 2000. Dictionary of Chinese Four-character Patterns. Beijing Language and Culture University Press, Beijing. Jones, S., 2002. Antonymy: a Corpus-based Perspective. Routledge, London. Jones, S., 2006. A lexico-syntactic analysis of antonym co-occurrence in spoken English. Text Talk 26 (2), 191–216. Jones, S., 2007. Opposites in discourse: a comparison of antonym use across four domains. J. Pragmat. 39 (6), 1105–1119. Jones, S., Murphy, M.L., Paradis, C., Willners, C., 2012. Antonyms in English: Construals, Constructions and Canonicity. Cambridge University Press, Cambridge. Justeson, J.S., Katz, S.M., 1991. Co-occurrences of antonymous adjectives and their contexts. Comput. Linguist. 17, 1–19. Kosti c, N., 2015a. Antonym sequence in written discourse: a corpus-based study. Lang. Sci. 47, 18–31. Kosti c, N., 2015b. The textual profile of antonyms: a corpus-based study. Linguistics 53 (4), 649–675. Langacker, R.W., 1987. Foundations of Cognitive Grammar (vol. 1). Theoretical Prerequisites. Stanford University Press, Stanford. Langacker, R.W., 1993. Reference-point constructions. Cogn. Linguist. 4 (1), 1–38. Langendonck, W.V., 1995. Categories of word order iconicity. In: Landsberg, M.E. (Ed.), Syntactic Iconicity and Linguistic Freezes. Mouton de Gruyter, Berlin/ New York, pp. 79–90. Langendonck, W.V., 2007. Iconicity. In: Geeraerts, D., Cuyckens, H. (Eds.), The Oxford Handbook of Cognitive Linguistics. Oxford University Press, Oxford. Landsberg, M.E., 1995. Semantic constraints on phonologically independent freezes. In: Langsberg, M.E. (Ed.), Syntactic Iconicity and Linguistic Freezes. Mouton de Gruyter, Berlin/New York, pp. 65–78. Liu, Z.Q., Xing, H.P., 2000. The semantic symmetrical features of four-character idioms in Chinese and their effect and cognition. Chin. Teach. World 1, 77–81. Liu, Z.Q., Xing, H.P., 2003. Symmetry in phonological patterns in four-character Idioms and cognition. Lang. Teach. Linguist. Stud. 3, 48–57. Lobanova, A., Van der Kleij, T., Spenader, J., 2010. Defining antonymy: a corpus-based study of opposites by lexico-syntactic patterns. Int. J. Lexicogr. 23, 19– 53. Lu, Z.W., 1956. Parallel four-character patterns. Lang. Study 1, 45–82. Lü, S.X., 1979. The Problem of Analyzing Chinese Grammar. Commercial Press, Beijing. Lyons, G., 1977. Semantics. Cambridge University Press, Cambridge. Ma, G.F., 1978. Idioms. Inner Mongolia Publishing House, Inner Mongolia. Malkiel, Y., 1959. Studies in irreversible binomials. Lingua 8, 113–160. Matlin, M.W., Stang, D.J., 1978. The Pollyanna Principle: Selectivity in Language, Memory, and Thought. Schenkman Publishing, Cambridge. Mayerthaler, W., 1988. Morphological Naturalness. Trans. by Seidler, J.. Karoma, Ann Arbor. Mettinger, A., 1994. Aspects of Semantic Opposition in English. Clarendon Press, Oxford. Min, J.Y., 1995. The Variation between Masculine and Feminine. Chinese Social Sciences Press, Beijing. Mollin, S., 2014. The (Ir)reversibility of English Binomials. John Benjamins, Amsterdam/Philadelphia. Muehleisen, V., Isono, M., 2009. Antonymous adjectives in Japanese discourse. J. Pragmat. 41 (11), 2185–2203. Murphy, M.L., 2003. Semantic Relations and the Lexicon: Antonyms, Synonyms and Other Semantic Paradigms. Cambridge University Press, Cambridge. Murphy, M.L., 2006. Antonyms as lexical constructions: or, why paradigmatic construction is not an oxymoron. In: Schönefeld, D. (Ed.), Constructions All over: Case Studies and Theoretical Implications. Special Volume of Constructions SV1-8. http://www.constructions-online.de/articles/specvol1. Murphy, M.L., Jones, S., 2008. Antonymy in children’s and child directed speech. First Lang. 28 (4), 403–430. Murphy, M.L., Paradis, C., Willners, C., Jones, S., 2009. Discourse functions of antonymy: a cross-linguistic investigation of Swedish and English. J. Pragmat. 41 (11), 2159–2184. Osgood, C.E., Richards, M.M., 1973. From Yang and Yin to and or but. Language 49 (2), 380–412. Panther, K.-U., Thornburg, L., 2012. Antonymy in language and use. In: Brdar, M., Fuchs, M.Z., Raffaelli, I. (Eds.), Cognitive Linguistics between Universality and Variation. Cambridge Scholars Publishing, Newcastle, pp. 161–188. Paradis, C., Löhandorf, S., Van de Weijer, J., Willners, C., 2015. Semantic profiles of antonymic adjectives in discourse. Linguistics 53 (1), 153–191. Quirk, R., Greenbaum, S., Leech, G., Svartvik, J., 1985. A Comprehensive Grammar of the English Language. Longman, London.

134

S. Wu / Language Sciences 59 (2017) 117–134

Radden, G., Kövecses, Z., 1999. Towards a theory of metonymy. In: Panther, K.-U., Radden, G. (Eds.), Metonymy in Language and Thought. John Benjamins, Amsterdam/Philadelphia, pp. 17–59. Radden, G., 2011. Spatial time in the west and the east. In: Brdar, B., Omazic, M., Takac, V.P. (Eds.), Space and Time in Language. Peter Lang, Frankfurt, pp. 1– 40. Roger, T.A., 2002. Yin and Yang. In: Antonio, S.C. (Ed.), Encyclopedia of Chinese Philosophy. Routledge, pp. 846–847. Schönefeld, D., 2005. Frozen locutions-frozen dimensions: LEFT and RIGHT in English, German and Russian. In: Maeder, C., Fischer, O., Herlofsky, W. (Eds.), Outside-in–Inside-out. John Benjamins, Amsterdam/Philadelphia, pp. 241–265. Shu, D.F., Huang, J., 2008. A cognitive approach to the internal sequence and semantic changes of Chinese antonymic compounds. Foreign Lang. Teach. Res. 40 (1), 418–422. Tai, H.J., 1985. Temporal sequence and Chinese word order. In: Haiman, J. (Ed.), Iconicity in Syntax. John Benjamins, Amsterdam/Philadelphia, pp. 49–72. Tai, H.J., 1993. Iconicity: motivations in Chinese grammar. In: Eid, M., Iverson, G. (Eds.), Principles and Prediction: the Analysis of Natural Language. John Benjamins, Amsterdam/Philadelphia, pp. 153–174. Waugh, L., 1982. Marked and unmarked: a choice between unequals in semiotic structure. Semiotica 38 (3/4), 299–318. Wen, X., 2001. Exploration of the iconicity of word order. Foreign Lang. Res. 3, 90–96. Willners, C., 2001. Antonyms in Context: a Corpus-based Semantic Analysis of Swedish Descriptive Adjectives. Lund University Press, Lund. Wu, S.Q., 2014. The metonymic interpretation of Chinese antonym co-occurrence constructions: the case of you X you Y. Lang. Sci. 45, 189–201. Xia, C.C., 1985. The Linguistic Arts in Classic of Poetry. Chinese Language Publishing, Beijing. Yu, N., 1998. The Contemporary Theory of Metaphor: a Perspective from Chinese. John Benjamins, Amsterdam/Philadelphia.

Further reading Chu, C.C., 1998. A Discourse Grammar of Mandarin Chinese. Peter Lang, New York. Li, C.N., Thompson, S.A., 1981. Mandarin Chinese: a Functional Reference Grammar. University of California Press, Berkeley.