Pretty與 Charming之辨析：以語料庫與心理語言學為本的研究 - 政大學術集成

全文

(1)國立政治大學英國語文學系碩士班(英語教學組) 碩士學位論文. 指導教授：鍾曉芳. 博士. Advisor: Dr. Siaw-Fong Chung. 立. 政治大. ‧ 國. 學. Pretty 與 Charming 之辨析：以語料庫與心理語言學為本的研究. ‧. Distinguishing Pretty and Charming: A Corpus-based and Psycholinguistic Study. n. er. io. sit. y. Nat. al. Ch. i n U. engchi. 研究生：游雅嵐. v. 撰. Name: Ya-Lan Yu 中華民國一○八年六月 June 2019. . DOI:10.6814/NCCU201900111.

(2) 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. DOI:10.6814/NCCU201900111.

(3) Distinguishing Pretty and Charming: A Corpus-based and Psycholinguistic Study. A Master Thesis Presented to. 政治大. Department of English,. 立. ‧. ‧ 國. 學. National Chengchi University. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. In Partial Fulfillment of the Requirements for the Degree of Master of Arts by Ya-Lan Yu June 2019. DOI:10.6814/NCCU201900111.



(6) ACKNOWLEDGEMENT With the completion of the thesis, I would like to express my greatest gratitude to the following people. Without their assistance and support, it would be hard to accomplish the thesis. First of all, I would like to show my heartfelt gratitude to my advisor, Dr. Siaw-. 政治大 process of working on my thesis. Her supportive attitude and professional guidance 立 Fong Chung, for her patient guidance, instruction, and encouragement throughout the. ‧ 國. 學. helped me find the direction of the research, enrich its content, and sharpen my writing skills. Her constant encouragement assisted me to go through the difficulties during the. ‧. process of writing. I will never forget the words that she gave me on Christmas: Never. sit. y. Nat. stop learning, because life never stops teaching. Although my life as a student is. io. er. coming to an end, I will bear these words in heart to open a new chapter in my life. Next, I would like to give my appreciation to my committee members of the thesis. al. n. v i n Cand defense, Dr. Michael Tanangkingsing Lin, and my committee member h eDr.nShih-hui gchi U. of the proposal defense, Dr. Wen-Hui Sah. Their detailed suggestions and comments, made the research more complete. I would also thank all my teachers at the TESOL MA Program of English Department, National Chengchi University, for their abundant teaching throughout the past three years in which I develop my professional knowledge of language teaching.. DOI:10.6814/NCCU201900111.

(7) I also give my gratitude to Dr. Jin-Lan Zhang, Dr. Siaw-Fong Chung, and Dr. Yung-Ho Hung, for providing me the opportunities to conduct the psycholinguistic judgement task in their classes. With their assistance, I could further explore English language learners’ use of the target words and this makes the study more valuable. Moreover, I would like to express my sincere thanks to all the members in the corpus research lab, who generously gave me variable comments and advice on my work. With their company and friendship, the process of conducting the study was not lonely.. 政治大 the worries and joys with. I am grateful for their love, understanding, and unconditional 立 Lastly, I would like to delicate my greatest thanks to my family whom I can share. support when I need to focus on my work. They gave me the strength to face all the. ‧ 國. 學. difficulties. Without their support, I would not be able to concentrate on the study and. ‧. make this possible.. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. . DOI:10.6814/NCCU201900111.

(8) TABLE OF CONTENTS CHINESE ABSTRACT…………………………………………………………….x ENGLISH ABSTRACT…………………………………………………...………xii CHAPTER 1 INTRODUCTION ............................................................................... 1 Background Information and Motivation of the Study .............................................. 1 CHAPTER 2 LITERATURE REVIEW .................................................................... 9 2.1 Near-Synonyms .................................................................................................... 9 2.1.1 Definitions of Near-Synonyms.......................................................................... 9 2.1.2 The Variations of Near-Synonyms .................................................................. 10 2.1.3 Collocational Analysis .................................................................................... 12. 政治大. 2.2 Sub-Classification of Adjectives ........................................................................ 13. 立. 2.2.1 Syntactic Sub-Classification of Adjectives ..................................................... 14. ‧ 國. 學. 2.2.2 Semantic Sub-Classification of Adjectives ..................................................... 16 2.3 Semantic Categories of Haily and Jung (2015) .................................................. 19. ‧. 2.4 The Relation between Gender and Language .................................................... 22. sit. y. Nat. 2.5 Corpus Linguistic Studies on Gender................................................................. 30. io. er. 2.5.1 Collocational Information ............................................................................... 30 2.5.2 Keyword Analysis ........................................................................................... 32. al. n. v i n 2.6 Psycholinguistic ExperimentsC ............................................................................ 35 hengchi U CHAPTER 3 CORPUS ANALYSIS ....................................................................... 39. 3.1 Methodology of the Corpus Study ..................................................................... 39 3.1.1 The Corpus—the Corpus of Contemporary American English (COCA) ....... 39 3.1.2 The Establishment of a Coding Scheme ......................................................... 40 3.2 Results of the Corpus Study ............................................................................... 47 3.2.1 Semantic Distribution ...................................................................................... 48 3.2.2 The Distribution of Gender Roles of the Modified People ............................. 51 3.2.3 Collocational Information ............................................................................... 52 3.2.4 Comparison of ‘Pretty’ and ‘Charming’ in Different Genres in the COCA ... 60 3.2.5 Syntactic Roles of ‘Pretty’ and ‘Charming’ .................................................... 66 . DOI:10.6814/NCCU201900111.

(9) CHAPTER 4 PSYCHOLINGUISTIC JUDGEMENT TASK ................................. 69 4.1 Methods of Psycholinguistic Judgement Task ................................................... 69 4.1.1 The Design of Questionnaire........................................................................... 69 4.1.2 Grading Criteria............................................................................................... 74 4.1.3 Samples of the Questionnaires A and B .......................................................... 74 4.1.4 Participants and Procedures of Questionnaires ............................................... 77 4.1.5 Data Analysis .................................................................................................. 78 4.2 Results of the Person-Stimuli ............................................................................. 79 4.2.2 The Source of beauty ⅹ Patterns Interaction ................................................... 80. 治政大 4.3 Results of the Object-Stimuli ............................................................................. 85 立 The Source of Beauty ⅹ Patterns Interaction ............................................................ 85 4.2.3. The Source of Beauty ⅹ Patterns ⅹ Gender Interaction .................................. 83. ‧ 國. 學. CHAPTER 5 DISCUSSION…………………………………...………………….93 5.1 Major Findings in the Present Study .................................................................. 93. ‧. 5.2 Teaching Suggestions ......................................................................................... 97. sit. y. Nat. CHAPTER 6 CONCLUSION ................................................................................ 101. er. io. 6.1 Overall Summary ............................................................................................. 101. al. 6.2 Limitations and Suggestions for Future Studies .............................................. 103. n. v i n Ch REFERENCES ....................................................................................................... 105 engchi U APPENDIX ONE Questionnaire A........................................................................ 109. APPENDIX TWO Questionnaire B ....................................................................... 111. . DOI:10.6814/NCCU201900111.

(10) LIST OF TABLES AND FIGURES TABLES Table 1- 1 Four Types of Variations of Near-synonyms ……………………………....1 Table 2- 1 Haily and Jung’s (2015) Semantic Categories ............................................ 20 Table 2- 2 Semantic Preferences of Cute, ‘Pretty’ and Beautiful…………………….20 Table 3- 1 Definitions of ‘Pretty’ from Three Dictionaries.......................................... 41 Table 3- 2 Definitions of ‘Charming’ from Three Dictionaries ................................... 42 Table 3- 3 The Senses of ‘Pretty’ in the Dictionary ..................................................... 43 Table 3- 4 Types of Nouns Modified by ‘Pretty’ and ‘Charming’ ............................... 45 Table 3- 5 Coding Scheme of Person ........................................................................... 46 Table 3- 6 Coding Scheme of Object ........................................................................... 47 Table 3- 7 The Distributions of Modified Nouns for ‘Pretty’ and ‘Charming’ ............ 48 Table 3- 8 The Semantic Distribution of Person for ‘Pretty’ and ‘Charming’ ............. 49 Table 3- 9 Distribution of Concrete and Abstract Nouns Modified by ‘Pretty’ and ‘Charming’ ........................................................................................................... 50 Table 3- 10 The Semantic Distribution of the Two Adjectives Used to Describe the Attractiveness of an Object .................................................................................. 51 Table 3- 11 The Distributions of Sex Roles of Modified Nouns for ‘Pretty’ and ‘Charming’ ........................................................................................................... 52 Table 3- 12 Top 20 R1 Noun Collocates of ‘Pretty’ and ‘Charming’ .......................... 53 Table 3- 13 Top 20 Adjective Collocates of ‘Pretty’ and ‘Charming’ (within a span of eight words) ......................................................................................................... 57 Table 3-14 The Comparison of R1 Noun Collocates of the Two Adjectives in the Fictional Genre..................................................................................................... 62 Table 3- 15 The Comparison of R1 Noun Collocates of the Two Adjectives in the Spoken Genre ....................................................................................................... 65 Table 3- 16 Distribution of Syntactic Positions ........................................................... 66 Table 3- 17 The Syntactic Roles Applied in Different Categories of Modified Nouns…………………………….......................................................................67 Table 4- 1 Stimuli Used in Person-Related Modified Nouns Part of Questionnaire ... 72 Table 4- 2 Stimuli Used in the Object-Related Modified Nouns Part of Questionnaire .............................................................................................................................. 74 Table 4- 3 Background Information of Participants in Group A ................................. 77 Table 4- 4 Background Information of Participants in Group B ................................. 78 Table 5- 1 The Design of the Language Learning Task ............................................... 98. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. . DOI:10.6814/NCCU201900111.

(11) FIGURES Figure 3-1 Frequencies (Per Million) of ‘Pretty’ and ‘Charming’ in Each Genre ....... 60 Figure 4-1 Group A’s Estimated Marginal Means of Acceptability of Sentences Describing Person-Stimuli ................................................................................... 81 Figure 4-2 Overall Estimated Marginal Means of Acceptability of Sentences Describing Object-Stimuli ................................................................................... 86 Figure 4-3 Group A’s Estimated Marginal Means of Acceptability of Sentences Describing Object-Stimuli ................................................................................... 87 Figure 4-4 Group B’s Estimated Marginal Means of Acceptability of Sentences Describing Object-Stimuli ................................................................................... 89. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. . DOI:10.6814/NCCU201900111.

(12) 國立政治大學英國語文學系碩士班(英語教學組) 碩士論文提要. 論文名稱：Pretty 與 Charming 之辨析：以語料庫與心理語言學為本的研究指導教授：鍾曉芳博士研究生：游雅嵐論文提要內容：. 政治大. 在英語學習環境中，近義詞的誤用為常見之語言現象之一。學習者通常倚賴. 立. 字典提供的語意，而忽略了其語言使用環境的不同。然而，近義詞之間的差異是. ‧ 國. 學. 微小、難以察覺的。本研究旨在透過語料庫為本的方法，分析近義詞 pretty 和 charming 的描述對象。此外，為得知學習者的學習困難，本研究也進行了以語料. ‧. 庫為基礎的心理語言學實驗測試搭配詞的接受度。. Nat. sit. y. 語料庫分析結果從不同角度展現了 pretty 和 charming 之相同與相異之處。. n. al. er. io. 其一，就語意分析得知，兩者皆能用來描述人和物體的吸引力。研究也發現，男. i n U. v. 性及女性都能被描述為 pretty 或是 charming。另外，在不同的語言環境中，這兩. Ch. engchi. 個形容詞皆具備表述褒義與貶義功能。其二，從句法結構與語域（register）表現得知，這兩個形容詞都偏好出現於名詞之前來修飾名詞，且他們較常被使用於非正式語域（例如: 口語和小說）。其三，從搭配詞分析可得知兩形容詞之主要差異如下：Pretty 傾向於形容人的外表與物體外觀（例如：面容以及洋裝）；而 charming 傾向於描述內在美，這種吸引力需要花較久時間思考與感受體驗（例如：個性、書）。其四，從分析心理語言實驗結果得知，搭配詞型態（pretty-only, charmingonly, and common）影響了受試者對於句子的接受度。而在三種搭配詞型態中，受試者對於 pretty-only 句子接受度最低，且較不熟悉。而在描述物體時，受試者. x . DOI:10.6814/NCCU201900111.

(13) 似乎不確定這兩形容詞使用差異。由此可得知，受試者對於 pretty 和 charming 的描述對象並不熟悉。基於所得結果，我們提出藉由分辨被修飾名詞的種類，可以探討 pretty 和 charming 兩者間使用差異。並嘗試解釋造成近義詞學習困難的原因。本研究透過語料庫結合心理實驗研究，提供近義詞學習困難之解釋方法。此發現可應用於英語教材設計與教學方法，並被視為相關近義詞分析研究基石。關鍵字：pretty、charming、形容詞近義詞、語料庫、心理語言實驗. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. xi . DOI:10.6814/NCCU201900111.

(14) Abstract One of the commonest lexical use problems in English learning context is related to near-synonyms. Many language learners rely on dictionaries to provide the denotational meaning of a lexical item but are less aware of the subtle differences embedded in contexts. The differences between near-synonyms are hard to identify and acquire for learners. The present thesis aims to analyze a pair of near synonymous adjectives, that is, pretty and charming, through the Corpus of Contemporary American English (COCA). In addition, in order to understand the learning difficulty of learners,. 政治大. a psycholinguistic judgment task (an acceptability judgment task) based on the corpus. 立. data was also conducted.. ‧ 國. 學. The corpus analysis demonstrated the similarities and differences between the two adjectives from several perspectives. First, from the semantic analysis, pretty and. ‧. charming shared a sense of ‘describing someone or something as pleasant or attractive’.. sit. y. Nat. Both males and females can be modified by the two adjectives. Moreover, in different. al. er. io. language contexts, the two target words can be used to express negative and positive. v. n. discourse functions. Second, for syntactic position and register, pretty and charming. Ch. engchi. i n U. were more commonly shown in the attributive position, tended to pre-modify the head of a noun phrase, and in informal registers (e.g., spoken and fiction). Third, as for collocational information, the main differences between the two adjectives were as follows: pretty was found to mainly co-occur with the outer beauty of people and objects (e.g., face, dress); charming tended to describe one’s inner beauty which takes more time to experience or think (e.g., personality, book). Fourth, from the analysis of psycholinguistic judgment task, one could find that the patterns of collocates (i.e.,. xii . DOI:10.6814/NCCU201900111.

(15) pretty-only, charming-only, and common) may affect the participant’s judgment of the sentences. Among the three patterns, pretty-only gained the lowest degree of acceptability, which showed that the participants might be less familiar with the collocational use of the pretty-only pattern. Furthermore, when describing an object, the participants seemed to be uncertain about the use of the two adjectives. Based on the results, we proposed that identifying the source of beauty of their modified nouns could distinguish the uses among pretty and charming. The difference could also provide a possible reason for the difficulty of near-synonym acquisition. This. 政治大 adjectives from the integrated corpus-based and empirical findings. The findings can 立 thesis attempts to unveil the difficulties of learning a pair of near synonymous. be applied to the design of English teaching techniques and materials and may be. ‧ 國. 學. regarded as the basis of analyzing synonymous adjectives in future studies.. ‧. Keywords: pretty, charming, near synonymous adjectives, corpus, psycholinguistic. Nat. n. al. er. io. sit. y. judgement task.. Ch. engchi. i n U. v. xiii . DOI:10.6814/NCCU201900111.


(17) CHAPTER 1 INTRODUCTION Background Information and Motivation of the Study In the process of developing one’s vocabulary, language learners may encounter the difficulties in choosing an appropriate word to use from pairs of near-synonymous words. Near-synonyms are lexical pairs or sets that have similar word senses but not completely identical (Murphy, 2003, p. 155). Many language learners tend to rely on the definitions provided by dictionaries to make a lexical decision. However, it is. 政治大. inadequate to only know the definitions of words to make the decision, sometimes. 立. dictionaries may overlook the importance of providing information about the. ‧ 國. 學. appropriate context to use the particular sets of near-synonyms. The subtle differences between near-synonyms are hard to identify and to acquire for language learners, even. ‧. for native speakers of a language as well. Misuse of lexical items, particularly between. Nat. sit. y. near-synonyms, has become one of the commonest language learning problems in the. n. al. er. io. second language context (Lee & Liu, 2009, p. 205). In addition, choosing a word. i n U. v. inappropriately may cause an unwanted implication. Edmonds and Hirst’ (2002, pp.. Ch. engchi. 109-110) summarized four types of variations of near-synonyms, demonstrated in Table 1-1. Table 1- 1 Four Types of Variations of Near-synonyms Types of variations. Definitions. Examples. 1. Denotational variation. (a) To convey concepts with emphasis (b) To convey concepts indirectly by merely suggestion or implication To compare the sets of nearsynonyms involving relatively small set of dimensions. (a) enemy/ foe (b) slip/mistake. 2. Stylistic variation. inebriated / drunk. 1 . DOI:10.6814/NCCU201900111.

(18) 3. Expressive variations 4. Structural restrictions. To express the speaker’s attitude to their denotations The near-synonyms may be different from their (a) collocational choices, (b) restrictions of selection, and (c) syntactic restrictions. skinny/ slim/ slender (a) task / job (b) die/ pass away (c) ajar/ open. Note. From Edmonds and Hirst (2002, pp. 109-110) As can be seen in Table 1-1, there are four types of variations of near-synonyms. Near-synonyms can differ in the manners in which they convey a concept by their denotation of emphasis and suggestion. In some circumstances, the speakers can choose. 政治大. a more proper word to express their attitudes or emotions. On the other hand, the. 立. stylistic variation, including dialect and register, affects the appropriateness of the. ‧ 國. 學. language user’s lexical choice. As for the last category of variations among nearsynonyms, the structural restriction may be either collocational, syntactic, or selectional.. ‧. These types of variations among near-synonyms will be elaborated in detail in Chapter. sit. y. Nat. Two.. n. al. er. io. In line with Edmonds and Hirst’s study (2002), Sinclair (2004) suggested that the. i n U. v. distinctions among near-synonyms may lie in their collocational behavior and semantic. Ch. engchi. prosody. For example, Partington’s study (1998, pp. 34-39) examined how the target word, sheer, collocated with a number of items from specific semantic categories, such as ‘magnitude’, ‘volume’, ‘size’ and ‘weight’. The results indicated that sheer had a preference of co-occurring with words sharing the semantic feature of ‘magnitude’. In terms of the variance in semantic prosody, Sinclair (2004, p. 34) proposed that “[semantic prosody] has a leading role to play in the integration of an item with its surroundings. It expresses something close to the ‘function’ of the item”. For example, a phrase, the naked eyes could be interpreted as expressions of some kinds of difficulties. The semantic prosody of ‘difficulty’ may be shown by the combinations of word such 2 . DOI:10.6814/NCCU201900111.

(19) as faint, weak, difficult with see (e.g., …too faint to be seen with the naked eye …), and barely, rarely, and just with visible (e.g., …it is not really visible to the naked eye…) or by a negating ‘visibility’ or invisible itself, or it may just be hinted at by a modal verb such as can or could: (e.g., …these could be seen with the naked eye from a helicopter…) (Sinclair, 2004, pp. 33-34). Other studies also proposed that the comparison between different corpora and on genre variations may be needed to figure out the behavior of near-synonyms in different genres. Chung and Chen (2015) aimed to examine the unique and share patterns of. 政治大 adjectives were commonly found. To analyze the near-synonym pair, both the BNC 立 critical and important in academic and general use. In an academic writing, these two. (serving as a reference corpus) and BAWE (serving as a specialized corpus) were. ‧ 國. 學. adopted to examine the collocational behavior of the two target words. The results. ‧. showed that critical and important behaved differently in the two registers. The shared. y. Nat. collocates of the two adjectives were different in the BNC and BAWE under the same. er. io. sit. grammatical relation. For example, the modifiers such as clearly/especially/often were shared by the two adjectives in the BNC (i.e., general register), but only highly. n. al. Ch. important/critical was equally shared by the. engchi. v i n two U adjectives. in the BAWE (i.e.,. academic register) (p. 15). The results indicated that the two corpora contained different contexts for critical and important (p. 20). This may affect the ways of teaching these particular adjectives. In another study, Hoffman (2014) explored the differences between a group of near-synonyms—nice, kind, lovely, friendly, gorgeous, and pleasant, by a corpus analysis. The nominal collocates of the adjectives were measured by raw frequency and MI score through the Corpus of Contemporary American English. Moreover, the stylistic variations among the six adjectives were examined across the five different genres provided by the same corpus (spoken, fiction, magazine, newspaper, and academic writing). As for collocational analysis, Hoffman showed that 3 . DOI:10.6814/NCCU201900111.

(20) the language users clarify the semantic differences between the near-synonyms. For example, the results showed that lovely usually collocated with nouns denoting female human beings (p. 22); while, pleasant frequently co-occurred with terms related to smell, such as aroma and fragrance. In terms of the stylistic variations of the six adjectives, the results indicated that all of the adjectives were rarely used in more formal registers (i.e., academic writing and newspaper). In addition, researchers also focused on the syntactic variation among the nearsynonyms. Liu (2010) investigated five near-synonymous adjectives—chief, main,. 政治大 semantic and syntactic structure as well as co-occurrence information of the set of near立. major, primary, and principal via COCA. The study focused on both the internal. synonyms. For the syntactic variations among the adjectives, Liu (2010, p. 73). ‧ 國. 學. emphasized through (a) the predicative use of the adjectives and, (b) adverbs that. ‧. modify these adjectives. As previous studies had shown that the five near-synonyms. y. Nat. can be used in an attributive position only and they were rarely modified by adverbs, a. er. io. sit. few instances were found in the COCA (i.e., 95 token for major and 66 tokens for primary) showing that among the five near-synonyms, major and primary can be used. al. n. v i n Cpositions. predicative their co-occurrence U h e n gRegarding i h c. in both attributive and. with. adverbs, some instances showed that the typical degree adverbs—very and really were used to modify both major (282 tokens) and primary (18 tokens) (pp. 73-74). Overall, corpus analysis can help language users examine and understand the similarities and differences between sets of near-synonyms. Furthermore, collocational information and concordances shown in the corpus can provide a more complete context of language use which can serve as a tool to help the language users acquire knowledge of contextual use of a particular language. From the previous studies, to solve the problems of distinguishing near-synonyms and to provide the explanation resolving the difficulty of making a proper lexical 4 . DOI:10.6814/NCCU201900111.

(21) decision may be two of essential significances of language learning. Investigating the language use could also be a valuable indicator of social change and people’s attitudes. Some research has shown that using psycholinguistic experiments can demonstrate how language was constructed by social interactions, for example, through e-mail writing tasks (e.g., Thomson, Murachver, & Green, 2001). Moreover, Onem (2017) utilized photo describing tasks to reveal whether there were gender-related differences presented in native speakers of Turkish in terms of using intensive adverbs. Thus, this thesis intends to adopt an integrated approach combining corpus analysis and. 政治大 EFL learner’ s use of a pair of near-synonyms. 立. psycholinguistic experiments to help demonstrate both native English speakers’ and. ‧ 國. 學. 1.2 Significance of the Thesis. From the previous studies, to help language learners to solve the difficulties of. ‧. choosing a proper word from a set of near-synonyms may be a vital significance in the. Nat. sit. y. language acquisition. The present study intends to adopt both corpus analysis and. n. al. er. io. psycholinguistic experiment so as to distinguish the differences between the two near-. i n U. v. synonymous adjectives—pretty and charming. Firstly, corpus analysis is used to. Ch. engchi. explore the semantic distribution, collocational information, syntactic variation, and stylistic variation (i.e., the different behaviors of the adjectives in different registers) of the two target words. Corpus analysis is expected to demonstrate the language use of native speakers of English and to identify similarities and differences between pretty and charming. Secondly, a psycholinguistic experiment serves as a tool for indicating factors that might influence EFL learners’ use of the adjectives. The psycholinguistic experiment in this study contains one acceptability judgement task mainly based on the corpus data. Three main controlled variables will be considered in the task: (a) ‘gender’, (b) ‘patterns’ of modified nouns (i.e., pretty-only, charming-only, and common), and (c) 5 . DOI:10.6814/NCCU201900111.

(22) ‘source of beauty’ (i.e., inner beauty and outer beauty). We assume that these variables may affect EFL learners’ judgement about uses of pretty and charming in sentences. Furthermore, prior studies (e.g., Weitzman, Eifler, Hokada, & Ross, 1972; Tannen, 1990; Sunderland, 2006; Crabb & Marciano, 2011) have pointed out that gender representations may be obscured in the language. We also intend to investigate whether there is a relation between gender and language which can be demonstrated in the pair of near synonym—pretty and charming. Thus, the combinations of the genders of the participants and the gender of the person in the sentence will be labelled as a monitor. 政治大 demonstrated in Chapter Four). Through the integrated approach, it is expected that this 立 variable in the study (more detailed introduction about the variables will be. study can shed some lights on the language teaching about the issue of lexical. ‧ 國. 學. acquisition. Teachers could help the learners to acquire a set of near-synonyms by. ‧. providing them with different kinds of information, such as collocation, context, syntax,. y. Nat. etc. Furthermore, it may be also essential to help students to develop the ability to. er. io. sit. understand the implied meaning of language uses, such as how language may reflect one’s attitude toward different genders or one’s social expectation.. al. n. v i n C hthesis, two researchUquestions are formulated in Based on the research scope of the engchi. (a-b) below:. (a) How can corpus data help in differentiating the similarities and differences between pretty and charming? (b) How do EFL learners distinguish pretty and charming? The thesis will be arranged in the following chapters: Chapter Two will begin with a review of some studies on near-synonyms, followed by studies adopting collocational analysis to explore the differences between sets of near-synonyms. Then, some important issues of near synonymous adjectives in previous studies, such as subclassification of adjectives (i.e., syntactic and semantic differences), and semantic 6 . DOI:10.6814/NCCU201900111.

(23) categories will be discussed. The relation between language and gender will be demonstrated as well. The last section will present the studies adopting psycholinguistic experiments to explore the participants’ language use. Chapter Three will concentrate on the corpus study. We will first introduce methods of the corpus analysis including the corpus used in the study and display the way to establish a coding scheme for data classification. Then, we will demonstrate the results of corpus analysis including the overall semantic distribution of pretty and charming, the distribution of gender roles of the modified people, collocational. 政治大 (i.e., predicative and attributive), and in different registers presented in the corpus. 立 information, the different behaviors of the two adjectives in different syntactic positions. However, since the issue of how EFL learners distinguish the two adjectives cannot be. ‧ 國. 學. understood from corpus analysis, a psycholinguistic judgement task is used.. ‧. Chapter Four will present the psycholinguistic experiment which aims to find out. y. Nat. what factors may influence learners’ language use of the two adjectives. In addition, the. er. io. sit. relation between gender and language were also considered in the study. Both methodology and results of the study will be presented in Chapter Four. We will firstly. n. al. introduce the design of. v i n Ch questionnaire and the grading engchi U. criteria. Secondly, in the. following section, introduction of participants, analytic procedure, and results will be shown. Chapter Five will discuss the results of both corpus study and psycholinguistic experiment. Chapter Six will present conclusion of the study as well as some limitations of the thesis.. 7 . DOI:10.6814/NCCU201900111.


(25) CHAPTER 2 LITERATURE REVIEW The thesis aims to conduct a research on the pair of near-synonyms—pretty and charming, and compare them in terms of their semantic distribution, collocation, syntactic positions (attributive and predicative), and stylistic variations. This chapter will review the previous studies from six main related aspects: near-synonyms, subclassifications of adjectives, semantic categories, gendered language, corpus linguistic studies on gender, and psycholinguistic experiments. 2.1 Near-Synonyms. 立. 政治大. In this section, near-synonyms will be reviewed from the definitions and studies. ‧ 國. 學. adopting collocational analysis to demonstrate the distinctions between sets of near-. ‧. synonyms.. 2.1.1 Definitions of Near-Synonyms. y. Nat. io. sit. Based on the Collins Dictionary, a ‘synonym’ refers to ‘a word that has the same. n. al. er. or nearly the same meaning as another word’. However, it is true that there are few. Ch. i n U. v. synonyms which are completely equivalent and interchangeable in all contexts. Lyons. engchi. (1968, p. 447) said that “synonymy is a matter of degree; that any set of lexical items can be arranged on a scale of similarity and difference of sense”. In other words, these ‘perfect’ synonyms rarely exist (Lyons, 1968, p. 448). Pairs of words that we frequently encounter are called ‘near-synonyms’ (Taylor, 2002). The senses of near-synonyms overlap to a great degree but not completely (Murphy 2003, p. 155). Although they don’t have the exact same senses, in each member of a near-synonym pair, we can find a sense which has similar sense of its counterpart (Murphy 2003, p. 147). Cruse (1986, p. 266) also suggested that “synonyms must not only manifest a high degree of semantic overlap, they must also have a low degree of implicit contrastiveness”. Namely, the 9 . DOI:10.6814/NCCU201900111.

(26) characteristic that we used to distinguish members from near-synonyms is called ‘implicit contrastiveness’. According to the definitions, pretty and charming are both adjectives that share a certain meaning. It is essential to find a way to identify the differences between the two adjectives. In the following part, we would like to present some studies adopting collocational analyses to show the distinctions between sets of near-synonyms. 2.1.2 The Variations of Near-Synonyms Edmonds and Hirst (2002) suggested that many kinds of variations involved. 政治大. denotations. The most commonly seen difference between sets of near-synonyms lay in. 立. concepts (i.e., roles and aspects of a situation) or ideas. For example, enemy and foe,. ‧ 國. 學. they differed in the emphasis they put on. Enemy focused on stressing “antagonism that arises from a cherished hatred or a desire to harm or destroy but it may suggest nothing. ‧. much more than active or evident dislike or a habit of preying upon” (Gove, 1984, p.. Nat. sit. y. 289) as showed in example (2-1). Foe emphasized on “implying active warfare” (p.. n. al. er. io. 289) as demonstrated in example (2-2).. Ch. i n U. v. (2-1) The woodpecker is a natural enemy of insects that infest the bark of trees.. engchi. (2-2) He is the foe of all reform measures. In addition, another kind of denotational variation was related to the ways to convey meanings indirectly through implying or suggesting. For instance, slip and mistake, slip “carries a stronger implication of inadvertence or accident than mistake and often, in addition, connotes triviality” (Merriam-Webster, 1984, p. 298), shown in the sentence: the wrong date on the check was a slip of the pen. However, Edmonds and Hirst (2002, p. 110) proposed that the differences between near-synonyms in denotation can be fuzzy rather than clear-cut. The differences between groups of near-synonyms 10 . DOI:10.6814/NCCU201900111.

(27) involved many aspects of meanings, for example, “woods and forest was a complex combination of size, primitiveness, proximity to civilization, and wildness” (p. 110). Secondly, a stylistic variation was related to a comparison between near-synonyms involving relatively small set of dimensions. For example, both inebriated and drunk are used to describe ‘someone has drunk so much alcohol that they cannot speak clearly or behave sensibly’ (Collin Dictionary). The differences between them was formality— inebriated was more formal than drunk. Thirdly, some near-synonyms differed from the ways to express the speaker’s. 政治大 be described as skinny, if the speaker wants to be deprecating; while when he/ she is 立. attitude to their denotations: good thing or bad thing. For instance, the same person may. described as slim or slender, the expression is more complimentary. Furthermore, some. ‧ 國. 學. words can express the emotions of the speaker indirectly, such as daddy, dad and father,. ‧. the use of daddy expressing a stronger feeling of intimacy than dad and father.. y. Nat. In terms of the last type of variations among near-synonyms involves structural. er. io. sit. restrictions; that is, the synonyms may be different from their collocational choices, restriction of selection, and syntactic restriction. For collocational choices, in a pair of. al. n. v i n C hco-occur with certain near-synonyms, one word can word but the other one may not. engchi U This word combination could be idiomatically uses. For example, task and job, they. differed from the collocational pattern: one can face a daunting task but not * face a daunting job. Restriction of selection referred to the appropriateness of the word combinations were defined semantically. For instance, unlike die, pass away can only be used to say someone died, not for plants and animals. As for the variation in syntactic restrictions, it was derived from a syntactic subcategorization. For example, some adjectives can be only used in a predicative position (e.g., the door is ajar, * the ajar door), but some adjectives can show in both the predicative position and attributive one (e.g., the door is open, the open door). 11 . DOI:10.6814/NCCU201900111.

(28) 2.1.3 Collocational Analysis Liu (2010) examined five near-synonymous adjectives (i.e., chief, main, major, primary, and principal) by focusing on the types of nouns that they each modify in order to understand the semantic and usages patterns of these adjectives. To identify the typical types of nouns that the adjectives modify, Liu conducted two types of queries: (a) frequency and (b) Mutual Information (MI) score. MI score is used to “compare the probability of observing x [word] and y [word] together (the joint probability) with the probabilities of observing x and y independently (chance) (Church & Hanks, 1990, p.. 政治大 co-occur. The two types of queries can complement each other in identifying typical 立 23)”. An MI score of 3 or higher can be seen as a strong evidence that two words often. nouns modified by the adjectives. The frequency measurement will undervalue the. ‧ 國. 學. words that have a low general frequency but commonly co-occur with the adjectives,. ‧. while MI score overvalue the words that have a low general frequency but have a high. y. Nat. tendency to collocate with target words. Based on the results of the frequency and the. er. io. sit. MI queries, a list of top ten modified nouns in each measure was compiled. These nouns were further classified into six semantic categories including (a) ‘abstract’ (e.g. concern,. al. n. v i n road,Cdish), (c) ‘dual’ (e.g. h e n g c h i U source,. reason), (b) ‘concrete’ (e.g.. component), (d). ‘institution’ (e.g. school, corporation), (e) ‘position-title’ (e.g. executive, counsel), and (f) ‘non-position-title’ (e.g. sponsor, author) (Liu, 2010, p. 66). Then, the author conducted a multifactorial test to determine whether the distributions of the five adjectives among these types of nouns differ significantly. The results showed that all five adjectives showed tendencies to pre-modify ‘abstract’ and ‘dual’ nouns. Additionally, ‘concrete’ nouns were commonly modified by main; ‘position-title’ nouns mostly by chief; ‘non-position-title’ nouns mostly by principal; and ‘institution’ nouns almost exclusive by major and primary (p. 68). In a recent study of near-synonyms of a set of adjectives (i.e., artificial, fake, false, 12 . DOI:10.6814/NCCU201900111.

(29) and synthetic), Kamiński (2017) explored the potential usefulness of two techniques— correspondence analysis plot and collocational network—that visualize collocation preference for distinguishing the near-synonyms. Correspondence analysis (CA) converts the raw frequencies of collocations into distances and shows them in a twodimension graph. The result of CA indicates the associations with each adjective. If the adjectives are closer to each other, it means that they may share certain collocates. For example, accent, pearls, fur, and flowers are collocates shared by artificial and fake (p. 246). These two adjectives may be more similar to each other. Collocational network is. 政治大 collocation. With this technique, we can find out semantic differences between the 立 a diagram showing the links between lexical units in terms of their strength of. adjectives by identifying the collocations not shared by others (e.g. bombs, permits, and. ‧ 國. 學. accent are typical for fake). The results can also show the collocational overlap; for. ‧. instance, pitch for both synthetic and artificial, suggesting the two adjectives had a. y. Nat. similar sense. Overall, both Liu and Kamiński’s studies demonstrated how collocation. er. io. sit. analysis can serve as essential tool to distinguish pairs of near-synonyms. However, other research also proposed that the differences between near-synonyms can be found. al. n. v i n C hsyntax or in different in their behaviors in different registers. Thus, in the following engchi U section, the sub-classification of adjectives will be discussed. 2.2 Sub-Classification of Adjectives. In the section 2.2, in order to better understand the similarities and differences between near-synonymous adjectives, the sub-classification of adjectives will be discussed in two parts; namely, syntactic positions of adjectives (i.e., attributive and predicative) and semantic distinctions for adjectives (i.e., (a) stative/dynamic, (b) gradable/ non-gradable, and (c) inherent/non-inherent).. 13 . DOI:10.6814/NCCU201900111.

(30) 2.2.1 Syntactic Sub-Classification of Adjectives Adjectives were generally classified into two types—attributive adjectives and predicative adjectives. Quirk, Greenbaum, Leech, and Svartvik (1985) proposed that adjectives can be categorized either as attributive-only, predicative-only, or attributiveand-predicative. “An adjective functions attributively when it pre-modifies the head of a noun phrase” (Quirk et al., 1985, p. 417), shown in example (2-3). (2-3) a small garden In terms of predicative adjectives, they can function as a subject complement or. 治政大 (i.e., be, become, get, an object complement. They are used following copular verbs 立 look, feel, seem, and appear). The predicative adjectives are subject complement not ‧ 國. 學. only to noun phrases (example 2-4-a), but also to clauses (example 2-4-b).. ‧. (2-4). sit. y. Nat. (a) The children are happy.. n. al. (Taken from Quirk et al., 1985, p. 417). er. io. (b) Playing chess is enjoyable.. i n U. v. For predicative adjectives functioning as object complement, they often indicate the. Ch. engchi. result of the process arisen from the verb (p. 417). The results can be demonstrated for each sentence by using the verb be (as shown in the example b for each example): (2-5) (a) He pulled his belt tight. (b) His belt is tight. (2-6) (a) He pushed the window open. (b) The window is open.. (Taken from Quirk et al., 1985, p. 417) 14. . DOI:10.6814/NCCU201900111.

(31) Notice that some adjectives can appear in both attributive position and predicative position, such as beautiful shown in the following examples. (2-7) (a) The beautiful girl. (attributive position). (b) The girl is beautiful. (predicative position). (c) He found her beautiful. (predicative position) (Taken from Peters & Peters, 2000, p. 2). The adjective, beautiful, functions attributively when it occurs before the noun it. 治政大hand, beautiful can also function modifies (i.e., girl) as in example (2-7-a). On the other 立 predicatively, when it serves as a subject complement and it follows a copula (i.e., is) ‧ 國. 學. in the example (2-7-b). For the last example (2-7-c), beautiful functions predicatively. ‧. as well, and it serves as an object complement when post-modifying her in the object position (Peters & Peters, 2000, p. 2).. y. Nat. io. sit. Taking a deeper look into the characteristics of adjectives, Quirk et al. (1985 p. 432). n. al. er. further pointed out that “adjectives which were restricted to the attributive use did not. Ch. i n U. v. characterize the referent of the noun directly”. For example, an old friend of mine, old. engchi. was restricted to an attributive position and it was used to describe the person’s friendship was old rather than to describe the person was being identified as old. Similarly, the wrong candidate referred to the mistake in identifying the person as a candidate rather than the wrongness of the person. The adjective did not characterize the referent of the noun directly, and it was termed as ‘non-inherent’ (e.g., the wrong candidate and an old friend of mine). While, when adjectives characterized the referent of the noun directly, they were termed as ‘inherent’ (e.g., that old man and my friend is old) (pp. 428-429). In terms of predicative-only adjectives, they were frequently used to “describe a 15 . DOI:10.6814/NCCU201900111.

(32) temporary condition, such as the health or the lack of health of the living” (example 28). In addition, “many predicative-only adjectives can take complementation”, such as able (to+ infinitive), fond (of), afraid (that, of, about). (Quirk et al., 1985, p. 432). (2-8) He felt ill.. For the target words of this thesis, both pretty and charming can take attributive position and predicative position. The two adjectives can be used to characterize the referent of the noun directly and indirectly. It depends on what the language users intend to emphasize: a permanent characteristic or a temperate states or particular events. The. 治政 analysis of the syntactic role of pretty and charming will be大 presented in Chapter Three. 立 Overall, the syntactic sub-classification is one of essential language features of ‧ 國. 學. adjectives. It may serve as a tool to differentiate the near-synonymous adjectives.. ‧. However, the use of adjective may not only be a matter of the syntactic position only but may also be related to the meaning they convey. In the following part, the semantic. y. Nat. n. Ch. engchi. er. io. al. 2.2.2 Semantic Sub-Classification of Adjectives. sit. classification of adjectives will be illustrated.. i n U. v. Rittman, Wacholder, and Kantor (2004, p. 351) proposed that adjectives had been seen as essential indicators of subjectivity. The adjectives are frequently used to express the author’s opinions or judgements. The statement, she wrote a poem, is a factual statement. However, when an adjective, beautiful, was employed to describe the noun in the sentence: she wrote a beautiful poem, the statement includes a fact with a person’s judgement. In the study of Quirk et al. (1985, p. 434), the researchers identified three basic semantic distinctions of adjectives—(a) stative/dynamic, (b) gradable/non-gradable, and (c) inherent/non-inherent of which the last type has been exemplified in the last 16 . DOI:10.6814/NCCU201900111.

(33) section. Bruce and Wiebe’s study (1999) further sharpened the previous findings, and they suggested that the presence of adjectives in a sentence was one of vital indicators of subjectivity through distinguishing the three semantic variables. Among the three distinctions, Bruce and Wiebe (1999, p. 203) identified that the (a) stative and dynamic seemed to be the most related to subjectivity. An example of stative adjective is tall; dynamic adjective is clever. To be more precise, dynamic adjectives tended to be more indicative of subjectivity than stative adjectives. A general feature of dynamic adjectives was that they tended to evaluate an attribute that is not. 政治大 whereas, stative adjectives tended to describe a state that is almost permanent, and it is 立 always present, but it is restricted by the possessor (e.g., foolish, careful, friendly);. not under the control of the possessor (e.g., big, small, tall). Syntactically, there were. ‧ 國. 學. several ways to differentiate between stative and dynamic adjectives. A stative. ‧. adjective, such as tall, cannot be used with progressive aspect or with the imperative,. y. sit. n. al. er. io. (2-9). Nat. as in the following examples:. (a) *He’s being tall. (b) *Be tall.. Ch. i n U. v. e n g c h i (taken from Quirk, et al., 1985, p. 434). On the other hand, a dynamic adjective, such as careful, can be used in these two syntactic patterns:. (2-10) (a) He’s being careful. (b) Be careful.. 17 . DOI:10.6814/NCCU201900111.

(34) For the second distinction—gradability (i.e., gradable/non-gradable), “it is manifested through comparison” (Quirk, et al., p. 435), for example, tall, taller, tallest, beautiful, more beautiful, the most beautiful are gradable. Additionally, gradable adjectives can also be modified by intensifiers (e.g., very, so, extremely) which can be used to express the degree of intensity of the adjectives, such as very tall, so beautiful, etc. In terms of non-gradable adjectives, according to Quirk et al. (1985, p. 435), most adjectives are gradable, only some technical adjectives (e.g., atomic and hydrochloric), and adjectives denoting provenance (e.g., British) are non-gradable.. 政治大 adjectives, it has been briefly discussed in the previous section 2-2-1. In general, 立. As for the last category of the semantic distinctions: inherent/ non-inherent. inherent adjectives were said to “characterize the referent of the noun directly; whereas,. ‧ 國. 學. non-inherent ones did not” (Quirk, et al., p. 435). To illustrate the definitions, one could. ‧. see that a non-inherent adjective can be used as an extension of the basic sense of. y. Nat. modified noun shown in the example (2-11). For (2-11-a), a firm friend is “a friend. er. io. sit. whose friendship is firm” rather than a firm man or woman, and in the example (2-11b), old also characterizes the ‘friendship’ not for the referent of the noun—friend. Thus,. al. n. v i n both firm and old in these sentencesC are seen as non-inherent h e n g c h i U adjectives. (2-11) (a) A firm friend (b) My old friend. (taken from Quirk, et al., 1985, p. 435). From the above discussion, we know that adjectives can appear predicative-only, attributive-only, or with both predicate and attributive syntactic roles. In addition, adjectives with different syntactic roles may have different characteristics. Specifically, predicative-only adjectives were commonly used to describe a temporary condition, such as the health condition (e.g., He felt ill). On the other hand, attributive-only 18 . DOI:10.6814/NCCU201900111.

(35) adjectives tended to characterize the referent of nouns indirectly (e.g., an old friend of mine). From first inspection, one could know that both the target words in the study, pretty and charming, function predicatively and attributively. Still, this information is inadequate to distinguish the pair of near-synonyms. Other research indicated that analysis of semantic categories can also help to differentiate the differences between sets of near-synonyms. Thus, in the following section, Haily and Jung’s study (2015) will be elaborated. Their semantic categories in this study will be further adapted in the thesis.. 政治大. 2.3 Semantic Categories of Haily and Jung (2015). 立. Haily and Jung (2015) aimed to investigate the similarities and differences in the. ‧ 國. 學. use of cute, pretty, and beautiful by using the Bank of English (BoE). The investigation focused on five language features of the target words: frequency, collocation, semantic. ‧. preference, semantic prosody and phraseology of the adjectives. In this study, ‘semantic. Nat. sit. y. preference’ is “a type of semantic patterning where the node item collocates with words. n. al. er. io. that are semantically related” (p. 127). For example, beautiful tends to collocate with. i n U. v. words related to ‘structure’ (e.g. house, building). As for ‘semantic prosody’, it is a. Ch. engchi. feature of a string of words concerned with communicative function (p. 127).1 To explore similarities and differences in the use of the three adjectives, firstly, the researchers tried to identify the distribution of the adjectives in each sub corpus in the BoE. The results suggested that all three adjectives were more frequently found in the written corpora. These adjectives may be more frequently used in formal language than informal language. Secondly, a search of the strongest collocates in different word classes was conducted. Nouns were the strongest collocates of the three adjectives, 1. Although there is a dispute about the definitions of ‘semantic preference’ and ‘semantic prosody’, in this study, the authors adopted Sinclair (2004)’s descriptions of the two terms. In addition, in the present proposal, the author will adopt definition of semantic preference provided Sinclair (2004). 19 . DOI:10.6814/NCCU201900111.

(36) followed by verb collocates. Thirdly, in order to find the semantic preferences of the three adjectives, the researchers classified the strongest R1 and R2 noun collocates into 10 categories, showed in the following table: Table 2- 1 Haily and Jung’s (2015) Semantic Categories Categories. Examples. People Body Part Animal Art Generality Location Structure Clothing Nature Other. kid, kids, baby, babies, boy faces, face, bottoms, bum, butt animal, animals, puppy, bear, horses comedy, picture, color, music, book one, stuff, thing, things town, village, place, south, country, city, cottage, bedroom room house, home clothes, dress, dresses, patterns garden, flower, countryside, valley, scenery accent, doll, penny, game, voice. 立. 政治大. ‧ 國. 學. Note. From Haily and Jung (2015, p. 140). ‧. It was found that there were some overlapped in the semantic preferences of the. y. Nat. io. sit. adjectives (see Table 2-2): they may have a similar preference for nouns related ‘people’,. er. ‘body parts’ and ‘generality’. Moreover, cute seems to have the fewest semantic. al. n. v i n preferences since cute only frequently C hcollocates with fiveU semantic categories while engchi. pretty and beautiful have seven (p. 132).. Table 2- 2 Semantic Preferences of Cute, ‘Pretty’ and Beautiful Categories People. Body part. Animal Art Generality Location Structure Clothing Nature Other. cute. . . pretty. . .  . . . beautiful .  . .  . . .  . . . . . . . Note. From Haily and Jung (2015, p. 132) 20 . DOI:10.6814/NCCU201900111.

(37) In terms of semantic prosody, concerned with the communication function of a string of words (Sinclair, 2004, p. 34), a further investigation in concordance lines was conducted. It is suggested that although the three adjectives had a consistent discourse function of compliment the appearance of people and things, they also can serve different functions such as causing insult to males (e.g., pretty boy) or criticizing an act that is dishonest or dirty (e.g. cute trick) (p. 133). Regarding phraseology, the study examined the concordance lines in an attempt to find the patterns or phrases of the adjectives. It is reported that pretty and beautiful appeared to share similar patterns, for. 政治大 pattern for both adjectives has a sense of ‘giving pleasure’ (p. 134). For instance, their 立. example, [ADJ to-inf] with ‘observation’ verbs such as behold, look at, and see. This. watching (i.e., observation verb) in the sentence.. 學. ‧ 國. hands are beautiful to watch (BoE). We can know that the viewer feels pleasure by. ‧. Overall, it is suggested that cute, pretty, and beautiful were not completely. y. Nat. synonyms based on the analyses of five language features presented above, they may. er. io. sit. have different uses in different language contexts. The researchers also pointed out that the value of corpus linguistic research: is to help us clarify any uncertainties that. al. n. v i n language learners, languageC teacher or speakers may h e n g c h i U encounter about language use (p, 136).. In summary, from the review of some corpus-based studies, one could know that corpus can help us identify the language use of adjectives from several aspects: collocation, syntax, and semantic, etc. However, within the field of gender and language, several studies have pointed out that language use could be correlated to a person’s social attitude and reflect social expectations toward different genders (Moon, 2014; Baker, 2014; Crabb and Marciano, 2011). Therefore, in the next section, the relation between gender and language will be firstly introduced, and then a review of approaches to explore the relation between gender and language will be presented. 21 . DOI:10.6814/NCCU201900111.

(38) 2.4 The Relation between Gender and Language Within the field of gender and language, Moon (2014) identified conventional ways that women and men were frequently described in the society through investigating collocational data. The results were correlated in turn with contemporary social attitudes and were consistent with other corpus studies. Moon (2014, p. 10) proposed that the recurrent finding is that ‘females are described in terms of appearance and vulnerability more than males are, for example, beautiful, blonde, naked; whereas, males are described in terms of strength, competence and activity more than females. 政治大 language was used in Malaysian personal advertisements in order to understand the 立. such as, handsome, honest, powerful. In another study, Bakar (2014) explored how. construction of feminine and masculine identities. Both corpus linguistic techniques. ‧ 國. 學. and discourse analysis of the texts were adopted in the study. The corpus data were. ‧. collected from an on-line personal advertisement website (my. match. com). The study. y. Nat. focused on several linguistic features to attain the goal. Firstly, functionalization. er. io. sit. occurred when “people were defined in terms of an activity” (p. 749), for example their occupations or roles. Secondly, the ‘identification’ of people referred to their age,. al. n. v i n C hothers (e.g., friend, gender, kinship or work relations with parents, sisters, brothers, engchi U etc.), and physical appearances (e.g., long hair, tall, fit, etc.). Thirdly, three types of attitudinal expressions were also analyzed in the study—(a) a person’s expressions of emotion, (b) evaluation of behaviors, and (c) evaluation of things. From analyzing these linguistic features, the results showed that both typical and atypical representation of females and males were found in the study. For the typical image of males, the image of “the fit and athletic man” (Bakar, 2014, p. 751) was found in the data. Men tended to describe themselves as part of a sporting community and their sports interests. While females were commonly marked by the collaborative features of talks and social relationship. Such features of females were correlated with the “traditional social 22 . DOI:10.6814/NCCU201900111.

(39) expectation that women were supposed to be compassionate, caring, and be sensitive to other’s feelings” (Bakar, 2014, p. 754). On the other hand, some atypical images of both genders were presented in Bakar’s study as well. “The sensitive new age man” was constructed via the discourses of emotion and domestic activities. The man’s personality was emphasized in terms of caring and compassionate traits. Moreover, “the new age man” were characterized by their participations of domestic works, such as cooking, gardening, and doing house works (p. 754). While atypical representation of females was characterized as “assertive and independent”. These traits were presented. 政治大 accomplishments (e.g., I’m managing the 立. by their capacity of doing occupational activities (e.g., I am a dance instructor) and educational. company’s branding and. marketing communications efforts dealing with target audiences globally) (Bakar, 2014,. ‧ 國. 學. p. 755). Overall, previous studies might suggest that the language use can reflect. ‧. people’s social expectations of genders. However, the image of both genders keeps on. y. Nat. changing, and the roles of different genders are changeable. For instance, Bakar (2014). er. io. sit. identified that females can also be assertive and independent; males can be sensitive and caring. In the following section, the approaches to investigate the relation between. n. al. Ch. gender and language were further discussed. . engchi. i n U. v. This section will look at the different approaches used in investigating gender and. language. To summarize the development of research within this field, we will introduce the approaches in the following three parts. Firstly, Lakoff’s (1973) dominance approach is presented. In this study, Lakoff focused on the social significance of gender and tried to differentiate between the speech patterns of the two genders through investigating their preferences for words, hedges, polite language, etc. Then, we will introduce a gender difference approach which focused on the differences among male and female by categorizing each gender into groups (Baker, 2014). However, some researchers had different opinions about the issue of gender and 23 . DOI:10.6814/NCCU201900111.

(40) language. They disagreed with the dichotomy between features of male and female speech patterns. The researchers adopted gendered discourse approach which aimed to find out how language use may reflect social convention and expectation. This approach will be introduced in the last part of the section. Lakoff’s (1973) dominance approach. When talking about the relation between gender and language, Lakoff’s (1973) study Language and Woman’s Place has been seen as a prominent and pioneering study. Lakoff emphasized on the social significance of gender. The language used by men and women can represent their social positions. 政治大. and the power of discourse. In the study, two aspects of English language were. 立. investigated; that is, lexical (e.g., color-words, swear-words, etc.) and syntax (e.g., tag-. ‧ 國. 學. question, and intonation). The results indicated that language use can demonstrate the social expectation of both males and females. Men’s verbal behaviors were seen as. ‧. more powerful and prestigious; whereas, women’s language was likely to be powerless. Nat. sit. y. and subordinate (p. 80). For example, females were taught to talk like ladies, in that. n. al. er. io. their expressions of strong statement were prevented (p. 57). This phenomenon can be. i n U. v. seen in the use of a compound request (e.g., Won’t you please open the door?). In this. Ch. engchi. kind of request, the speaker left the decision up to the addressee and assumed the addressee’s answer was a negative response. Thus, the addressee might feel freer to refuse the request. As Lakoff (1973) emphasized that language use reflects one’s power and inequality of society. Thus, women’s language put themselves on a disadvantage position. Following Lakoff’s study, further studies were conducted on the gender and language issue during 1980s and 1990s. In another study, Schulz (1975) investigated the semantic change of words with respect to the role of gender in language. The study revealed that the words for women tended to develop in a derogative direction, compared to the words for men. Specifically, male terms were often associated with 24 . DOI:10.6814/NCCU201900111.

(41) positive and affirmative meaning. However, female terms easily carried negative connotations. For example, both bachelor and spinster refer to a singer adult. Whereas, bachelor refers to an unmarried man, a spinster refers to an unmarried woman beyond the marriageable age. Bachelor connoted that men were never too old to marry, but women can be. Similarly, in another pair of words the semantic degradation can also be found. Originally, lady and lord were used only for women and men of the nobility. However, in Early Modern Britain (1500-1700), there were lesser nobility in the society. Lady no longer reserved the meaning of a woman of high rank in the society. It was. 政治大 original meaning—a man who has lots of power or a man who has a high rank in the 立 generalized to refer to people in all level in the society. Whereas, lord, remained its. nobility. Overall, from the perspective of gender dominance, the differences of male. ‧ 國. 學. and female language use were related to inequality of society, where women seemed to. ‧. be subordinate to men.. Nat. sit. y. Gender difference. As one of Lakoff’s students, Tannen (1990) called for a shift. n. al. er. io. from the ‘dominance approach’ to a ‘difference approach’. She argued that “[t]he effect. i n U. v. of dominance is not always the result of an intention to dominate” (Tannen, 1990, p. 8).. Ch. engchi. In Tannen’s (1990) book, You just Don’t Understand: Women and Men in Conversation, she adopted many anecdotes and conversations to explain the frustrations of trying to communicate with the other sex. These communicative problems seemed to be a universal issue for everyone. We all have some similar experiences in our lives. To explain the causes of the communicative problems, Tannen utilized a cultural perspective. It is said that women and men grew up in different cultures which led to the development of different communicative styles. People talked to them differently, expected and accepted different ways of talking style from them. Thus, children learned the ways to speak and to communicate with others from the society and through the 25 . DOI:10.6814/NCCU201900111.

(42) interactions with others. Girls tended to maintain a close and equal relationships when they play with others; whereas, boys were likely to play in a larger and hierarchically structured groups. They focused on physical or competing activities. In this context, boys learned the ways to communicate in a dominant way. Tannen (1990, p. 9) explained that “in this world, conversations were the way to negotiate for closeness in which people try to seek and give confirmation and support, and to reach consensus.” People tried to preserve intimacy and avoid isolation in their lives. However, both the genders put the emphasis on different aspects. Females tended to pay attention on. 政治大 centers on intimacy emphasized on the connection with others, trying to minimize the 立 intimacy; while, males tended to focus on independence. A communicative style which. differences. Whereas, a communication style which centers on independence. ‧ 國. 學. emphasized on status, power, and difference (p. 31). From the different conversational. ‧. styles of the both genders, Tannen further suggested that the communicative issue. y. sit. io. er. culturally.. Nat. between males and females were related to the inherent attempts to communicate cross-. One of modern research arguing for gender differences in language is Jackson’s. al. n. v i n C hteachers’ perspective (2006) study. He explored pupils’ and about ‘ladette’ cultures in engchi U secondary schools in the United Kingdom from interview data. The term ‘ladette’ is mainly associated with females who are ‘in trouble’ or are a ‘trouble’ in the UK media. Based on the analysis of interview data, the results indicated that both teachers and pupils had similar perceptions about ‘ladettes’. They all thought that the common characteristics of ‘ladettes’ are disruptive, rude to teachers, and aggressive to other pupils. Nevertheless, in school, “discourses around ladette are complex” (p. 355). Some aspects of ‘ladettes’ behaviors were celebrated, for example, assertiveness and increased confidence. There was a line between being assertive and being arrogant. The researcher further pointed out that this line was commonly drawn in different standards 26 . DOI:10.6814/NCCU201900111.

(43) for boys and girls. Teachers were more tolerant of males’ interruptions in class. Whereas, the similar behavior by females were often regarded as big different. This double standard phenomenon may also be related to the expectation of females that they need to be ‘successful’ in class compared to males who are more commonly portrayed as ‘underachievers’ in an academic achievement (p. 355). The dominance approach advocated by Lakoff (1973) put the emphasis on the inequality of society—females seemed to be subordinate to males. Tannen (1990) proposed that the linguistic differences between both the genders were not the result of. 政治大 learned the ways to communicative from the interaction with others. However, based 立 subordination of females, but as a consequence of cultural differences. Boys and girls. on the above, Cameron (2005, p. 486) suggested that although the dominance and. ‧ 國. 學. difference approaches offered contrasting accounts, “both the approaches regarded the. ‧. linguistic differences as a matter of gender rather than sex, and both often described. y. Nat. them as the product of early socialization”. Accordingly, Cameron (2005, p. 484). er. io. sit. categorized these two big stories—Lakoff’s dominance approach and Tannen’s difference approach as ‘Modern feminist approach’. However, some researchers took. al. n. v i n an opposite position to the C modern feminist approach. h e n g c h i U They claimed that gender was. not something that everyone possesses; it was something that people perform (Butler, 1990, p. 140). From this point of view, the advocate believed that the development of male or female characteristics depended on their social environments. “Gender was determined by a complex array of social factors, rather than by biological sex alone” (Barczewska and Andreasen, 2018, p. 197). Thus, the idea of gender was constructed by social environments and it was found in both CA and discursive psychology research (Cameron, 2005). We will be presented in the next section. Gendered discourses. Sunderland (2004, pp. 20-22) proposed that the ‘gendered discourses’ concerned with the substance of gender. The language use and gender 27 . DOI:10.6814/NCCU201900111.