Sadness Expressions in English and Chinese

Corpus Linguistic Contrastive Semantic Analysis

Ruihua Zhang

• Publisher: Bloomsbury Academic
• Year: 2014
• ISBN:
  9781472506610
• Edition: 1
• Topic: Lingue e linguistica
• Subtopic: Semantica linguistica

Winner of the Tianjin Social Science Outstanding Achievement Award. This book reports on the contrastive-semantic investigation of sadness expressions between English and Chinese, based on two monolingual general corpora and a parallel corpus. The exploration adopts a unique theoretical approach which integrates corpus-linguistic theories on meaning (as a social construct, usage and paraphrase) with a corpus-linguistic lexical model. It employs a new complex but workable methodology which combines computational tools with manual examination to tease meaning out of corpus evidence, to compare and contrast lexical items that do not match up neatly between languages. It looks at sadness expressions both within and across languages in terms of three corpus-linguistic structural categories, i.e. colligation, collocation and semantic association/preference, and paraphrase (both explicit and implicit) to capture their subtle nuances of meaning, disclose the culture-specific conceptualisations encoded in them, and highlight their respective cultural distinctiveness of emotion. By presenting multidisciplinary original work, Sadness Expressions in English and Chinese will be of interest to researchers in corpus linguistics, contrastive lexical semantics, psychology, bilingual lexicography and language pedagogy.

1

Introduction

The issues of universality versus cultural specificity of emotions have long been debated among psychologists and biologists. Biologists tend to believe that emotions are innate and universal because they have biological bases. The dominant view among psychologists, mainly based on some psychological experiments, is that human beings have some basic emotions, which are universal. The cultural specificity of emotions cannot be discussed without recourse to the cultural specificity of language, which has often been overlooked by psychologists. To what extent does language shape the way we perceive things and therefore influence our emotions? Whorf (1956) tells us that the way we cut the world into objects and actions is not a true reflection of reality but a division imposed upon us by our language. Although the strong version of the Sapir-Whorf hypothesis—language entirely determines thought and cognition—is probably not tenable, more and more linguists have begun to support its weak version, i.e. language does influence our thoughts to some extent. According to this hypothesis, the emotional lexicon of a language constrains the thoughts of its speakers about feelings. Each language has its own set of emotion words to characterize emotional experience, but apparently there are more distinct human emotions than those words can describe (Wierzbicka 1992a). A language has not lexicalized a particular emotion because its culture has not found it worthy of a specific word; it does not mean that its speakers have never experienced that emotion or cannot perceive it as a recognizable feeling (Wierzbicka 1992a). This is why we reject the Sapir-Whorf hypothesis in its strong form. However, the embracing of its weak version means that the ways in which the speakers of a language characterize, discuss and communicate their emotional experience are largely shaped by the ready-made emotional lexicon in their language.

‘Emotion words reflect, and pass on, certain cultural models, and these models, in turn, reflect and pass on values, preoccupations, and frames of reference of the society (speech community) within which they have evolved’ (Wierzbicka 1999: 32). The close relation between language and culture can be explained by the newly rising enterprise—cognitive linguistics,¹ represented by the works of Lakoff and Johnson (1980), Lakoff (1987), Langacker (1987) and the like. Its basic assumption is that human language cannot be viewed as an abstract system isolated from socio-psychological settings in which it is used; on the contrary, it reflects the way people experience the world. In other words, a wide variety of factors such as perception, reasoning, the nature of the body, the emotions, memory, social structure, and cognitive development, etc. determine the structural characteristics of language. Langacker (1999: 16) reveals the correlation between language and culture:

… the specific conventions of a given language are culturally transmitted through social interaction. Language is an essential instrument and component of culture, whose reflection in linguistic structure is pervasive and quite significant.

At the outset of communication, speakers of a language need to characterize the objects and events in their world. If we reject the innate hypothesis, their characterizations are unavoidably influenced by the way in which they perceive and construe things around them. Such characterizations are then encoded into various layers of their language and culturally transmitted from generation to generation; of course, they have been gradually enriched in the later development of the language. Hence, the organization of a language is shaped by the way in which its speakers perceive the world and conceptualize the phenomena around them. This explains, in large measure, why different languages have different structures and lexicons. The lexicalization of a concept is to ‘present it as an established category of human thought’ (Fillmore 2003: 258). As the established categories of human thought, the lexicon of a language will no doubt reflect its cultural features, including its emotional lexicon. Semantically similar emotion words in different languages might overlap. The comparison of the emotion lexicon of a language with that of another can therefore suggest similarities and differences in conceptualizations of emotion.

In lexical contrastive semantics, one of the thorniest problems is how to establish equivalence of lexical items across languages. One common belief would be that equivalent lexical items in different languages can be decomposed into identical sets of features. This sort of decomposition presupposes the existence of primitive or universal concepts, which can serve as tertium comparationis to establish lexical equivalence across languages. However, this will bring numerous problems. Krzeszowski (1990) presents a detailed analysis of this difficulty (see Section 3.2). It seems that the establishment of equivalence of lexical items across languages, by decomposing words into primitive concepts, is more than difficult to achieve. This is because the idea of primitive concept itself is problematic: we have to express it in a natural language, where the great majority of words are polysemous and ambiguous. Traditionally, their meanings in the dictionary have to rely on other words there, which in turn will unavoidably lead to circularity. Meaning is far more subtle than a componential analysis based on discrete features can account for, and far more complex than any print dictionary can describe. To carry out a cross-linguistic lexical study, the better choice is to start with ‘a perceived or assumed similarity’ between contrasted items (James 1980: 168) rather than to seek primitive or atomic units. What we can do is to refine the ‘initial assumptions of similarity’ (Altenberg and Granger 2002: 16). The analysis is only to ‘add explicitness, precision, perhaps formalization’ (Chesterman 1998: 58); it may also provide ‘added information, added insights, added perception’ (Chesterman 1998).

Most of the traditional contrastive lexical studies are incomplete and inadequate, for they are not supported by empirical quantitative data and fail to present linguistic information on frequency of use, collocations, colligations, semantic preferences/associations of the examined items. Their lexical descriptions have been largely based on individual researchers’ introspection, dictionaries or scanty empirical evidence. Due to the inadequacy of previous contrastive lexical studies, there has been a growing tendency to base such studies on corpora.

The methodology of corpus linguistics is based on scientific principles. It relies on authentic texts and computational tools to detect and classify linguistic phenomena; the datasets and computer programs are open to scrutiny by the peer community; the results they produce are reproducible. This approach to lexical studies aims to arrive at the meaning of a word by observing its external linguistic behaviour in various contexts rather than looking for its primitive units intuitively. Such a practice can be justified by Firth’s (1957) dictum that a word shall be known by the company it keeps. In keeping with Firth’s observation, Levin’s (1993: 1) central claim that the behaviour of a verb is to a large extent determined by its meaning can also be extended to other classes. Thus, the semantic information encoded in the word can be revealed by examining its linguistic behaviour. By using corpus tools, we can read its concordance and examine its collocates. As long as we have sufficient instances, anything significant about the node word will be disclosed. Hence, to know the meaning of a word does not necessarily follow that we have to decompose it. Computer corpus linguistics will tell us in exactly what way collocation is related to meaning. Corpus-assisted contrastive lexical studies will show that distinctions of semantically similar words in the same language or in different languages can be drawn without decomposing them.

1.1 Aims and objectives

Emotions are not innate but constructed in various cultural contexts through the process of socialization (e.g. Santangelo 2003). Different cultural groups conceptualize their emotional experience in different ways, as they categorize other human actions and thoughts (e.g. D’Andrade 1995; Palmer 1996; Strauss and Quinn 1996). The conceptualizations emerging at the cultural level of cognition are referred to as ‘cultural conceptualizations’ (Sharifian 2003, 2008), which are negotiated and renegotiated by its members through time and across generations (Sharifian 2003, 2008). The cultural conceptualizations of emotions are encoded in the emotional lexicon of a language, which reflect the ideas, beliefs and cultural models of emotions of its speech community. Analysis of the emotional lexicon of a language helps reveal the concepts available to its speakers through which they categorize and describe their emotional experience. Where lexicalized in a single word, the presence of a word implies the existence of the concept it expresses. Language-specific emotion words might overlap semantically across languages. By comparing and contrasting the semantic features of the emotion words in a language and their equivalents in another, we can know what they share and how they differ.

Sadness has been proposed as one of the basic emotions by many psychologists. However, what are considered basic emotions in the Chinese culture are markedly different from those proposed for the English language. The disparity between sadness concepts across these two cultures seems to be more significant, as compared with happiness, anger, etc. The primary goal of this book is to investigate the general concept of ‘sadness’ within the English and Chinese contexts by examining lexical sets of sadness expressions. The exploration will be conducted by applying corpus-linguistic theories (Teubert 2005, 2010; Teubert and Cermakova, 2004) on meaning (as a social construct, usage and paraphrase), i.e. meaning is language use and it is constructed and negotiable through paraphrases in discourse, and by employing Hoey’s (2005) or adapting Sinclair’s (1996, 2004) corpus-linguistic lexical model, i.e. to look at lexical items in terms of colligations, collocations and semantic preferences/associations.

In this book, I will describe the colligational properties of sadness expressions and formulate their semantic preferences/associations on the basis of their collocates, compare and contrast the semantically similar items within and across languages to capture their semantic similarities and differences, uncover the respective cultural distinctiveness, and explain why they exhibit such differences from a cultural perspective. Their semantic information will become vivid when we look at the empirical language data in corpora. As long as our corpora are large enough, our quantitative tools will yield sufficient data, on the basis of which the exact meanings of these sadness expressions and the finer distinctions between them can be accurately captured. By comparing the semantic features of sadness expressions between Chinese and English, their emotive content will be revealed and the cultural difference of emotion between English and Chinese can be highlighted.

The sadness expressions to be examined include (cf. Section 7.1): sorrow, grief, unhappy, sad, heartbreak, mourn, doleful, woeful, woebegone, depression, melancholy, gloomy, upset, dejected, disheartened, despondent, bitter, anguish, agony in English, and bēitòng, bēishāng, bēiāi, shāngxīn, nánguò, yōushāng, bù gāoxìng, bú kuàilè, bù yúkuài, xīnsuì, āidào, chóuméikǔliǎn, jǔsàng, yāyì, qíngxùdīluò, yìyù, yōuyù, shānggǎn, yīnyù, mènmènbúlè, huīxīn, sàngqì, chuítóusàngqì, huīxīnsàngqì, xīnhuīyìlěng, tòngkǔ in Chinese. The above English items will be looked at together with their cognate items. For instance, sad will be examined along with sadness, sadden, sadly. Finally, an equivalence network between English and Chinese sadness expressions and a summary of their semantic similarities and differences across languages will be presented.

1.2 Methodology

There has been an increasing awareness that cross-linguistic studies cannot rely on introspection or dictionaries, but must be firmly based on naturally occurring language used in a variety of situations. With the aid of computer technology, we can observe semantic similarities and differences across languages objectively; we can identify clearly what elements they share with similar expressions from the other language and what elements are specific to their own language. Basing my research on corpora ensures that my investigation leads to tenable results and a more accurate description of English and Chinese sadness expressions.

Discussions of emotions can be cleared from the charge of being impressionistic by means of an examination of empirical data. To retrieve reliable information on lexical use, large corpora are needed, particularly for less frequent words. Sufficient instances of the examined item from various texts are required to give a reliable picture about its use. The data resources of my investigation will be the Chinese corpus developed at the Centre for Chinese Linguistics, Peking University, the Bank of English and the Babel English–Chinese Parallel Corpus of Fiction compiled at Peking University. This corpus-based approach is embedded in a corpus-linguistic theoretical framework. The purpose of incorporating a parallel corpus into this methodology is to base cross-linguistic lexical links on quantitative data.

This analysis will involve, by means of examining their collocates, describing the colligations and semantic preferences/associations of the English and Chinese sadness expressions respectively. Colligations ‘provide an objective view from the outside. They describe the observable behaviour of a social group’ (Stubbs 2009: 125). According to Stubbs, semantic preferences/associations ‘provide a subjective view from the inside’ and they can only be identified intuitively (p. 125). Sometimes it is not an easy task to label them adequately. Once this is achieved, I will then look for real paraphrases (explicit paraphrases) and typical examples (implicit paraphrases, see Section 5.2.2) that contain information contributing significantly to revealing meaning. The final step of the analysis is to compare and contrast cross-linguistically and qualitatively the meanings of the semantically similar expressions within the same group on the basis of the previous corpus findings and provide a cultural explanation for the differences.

1.3 Significance

Psychological approaches to emotions have led to a number of important insights. Nevertheless, such studies have the disadvantage of seeming less objective because they are either based on introspection or the subjects studied are unlikely to behave in a natural way under observation. On the other hand, the corpus-based approach to the study of emotions offers a number of advantages. First, there is no ‘observer’s paradox’; in other words, since the study is based on texts written independently, the results will be less distorted by the subjects. Secondly, it is easier to achieve accuracy and reliability in such an analysis because the number of texts to be examined is much larger than can be looked at in psychological studies.

This study extends the corpus-linguistic framework to analyse and compare lexical meaning in depth. It proposes a unique theoretical approach that integrates corpus-linguistic theories on meaning (as a social construct, usage and paraphrase) at the macro level with lexical items at the micro level. It also develops a new complex methodology, which combines computational tools with manual examination to tease meaning out of corpus evidence, to compare and contrast lexical meaning within and across languages. Hopefully, this will provide fresh insights into how developments in corpus-linguistic theories as well as in the application of computer corpus technology can advance the field of lexical contrastive linguistics. By so doing, we can have a nuanced description of similarities and differences that epitomizes the descriptive potential of the corpus-linguistic approach to contrastive lexical semantics. This book will be the first large-scale corpus-based examination of the ‘sadness’ expressions. By analysing a large quantity of contrastive corpus data, I will uncover differences in the use of emotion words, and a large part of this book will be devoted to explaining them. I will show to what extent the differences can be explained through cultural differences. This is a relatively recent field of enquiry and investigations are still few and far between. My own contribution will hopefully offer new insights not yet described in the literature. In addition, the corpus-linguistic approach can throw new light on psychology, contrastive lexical studies, bilingual lexicography and language pedagogy.

1.4 Summary

This book explores, with reference to large corpora, the semantic disparities between sadness expressions in English and Chinese by applying the corpus-linguistic approach. It aims to probe into the cultural reasons that caused these differences, based on the corpus findings. Chapter 1 gives the rationale, objectives, methodology and significance of this research. It is grounded in three research areas: emotions, contrastive lexical semantics and corpus linguistics, so Chapter 2 addresses topics of emotions in psychology and emotion research in linguistics, Chapter 3 deals with contrastive lexical semantics, focusing on the frequently discussed issues of lexicalization and decomposition, and Chapter 4 is concerned with corpus linguistics, parallel corpora and cross-linguistic research. Chapter 5 explicates the theoretical framework on which the research is based and Chapter 6 deals with its methodology. Chapter 7 mainly reports, with great detail, on the corpus findings from the semantic-contrastive analysis of English and Chinese sadness expressions. The corpus-based analysis is also complemented by a contrastive-qualitative analysis of metaphorical sadness expressions in English and Chinese. Chapter 8 discusses possible implications of this research for psychology, contrastive lexical semantics, bilingual lexicography and language pedagogy. The final chapter summarizes the research findings, discusses the limitations of this research and envisages the further development of cross-linguistic lexical studies with the aid of the corpus-linguistic approach in the near future.

2

Emotions

2.1 Emotions in psychology

There has been a heated debate about whether emotions are culture-specific or universal among anthropologists, biologists and psychologists. In 1872, Darwin addressed the behavioural manifestations of emotions in his The Expression of the Emotions in Man and Animals (1872/1965). Much of the research on emotion in the field of psychology maintains his tradition by documenting universal aspects of human emotions. The basic emotion theory has drawn the most attention, which argues that there are 5–7 basic emotions common to all human beings. Proponents of this theory have used facial-expressions recognition studies as their central source of evidence. Th...