Eh .. Z .. , .2532; 2.7.”: . \1. res: ..,...e eckout from your record. m on or before date due. PLACE IN RETURN BOX to remove this ch TO AVOID FlNES retu MAY BE RECALLED with ear Her due date if requested. 11/00 c-jClRCIDaleDueprrm 4 HEDGES IN JAPANESE SPOKEN DISCOURSE: A COMPARISON BETWEEN YOUNGER AND OLDER SPEAKERS BY Shizuka Lauwereyns A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Linguistics and Germanic, Slavic, Asian, and African Languages 2000 ABSTRACT HEDGES IN JAPANESE SPOKEN DISCOURSE: A COMPARISON BETWEEN YOUNGER AND OLDER SPEAKERS By Shizuka Lauwereyns In the present study, hedges are defined as expressions of uncertainty, possibility, tentativeness or approximation which convey a sense of vagueness. For example, I examine hedges such as toka ’or something’, kana ’I wonder’, omou ’I think’, nanka ’like' , tabun ’probably’ and teyuuka ’or rather’ . A speaker uses these types of hedges to mitigate his/ her utterances, to show a non-committal or self—protective attitude, or to show solidarity among conversation participants. In general, the use of hedges is motivated not only by information demands (lack of knowledge), but also by sociolinguistic factors. The goal of this study is to investigate the effect of major social variables (age, sex and style) on the use of hedges in Japanese conversation so that we can understand how and why people speak vaguely using hedges. I posited three hypotheses: (1) hedges are used more often by younger speakers than by older speakers; (2) hedges are used more often by female speakers than by male speakers; (3) hedges are used more often in casual speech than in formal speech. To test these hypotheses, conversation data of 20 single sex dyads were collected from two age groups (younger and older) and from both sexes in two styles of conversation (interview and chatting wiUnafriend). AllparticipantsarenativespeakersofJapanese. The data supported hypotheses (l) and (2), but not (3). This means that the younger groups and female groups used hedges morecfitenthanthecounter—partgroups. Hedgesappearedasoften in interviews as in chats. There was also an interaction of age and sex: the younger female speakers used hedges the most. By closely examining toka ’or something’ and teyuuka ’or rather’ in context, it was found that these hedges were used differently depending on the speaker's age and the speech genre. Theyoungergroupsusedaffection—orientedhedgesmorecfitenthan the older groups. The younger groups also showed variations in the usage, and sometimes expressed playfulness, embarrassment or solidarity in the use of hedges when talking to their friends. According to the questionnaire results , the younger female speakers have a different attitude regarding the use of hedges. For example, they consider hedges a normal way of talking, fun and easy to say, whereas the older speakers show negative views, considering hedges corrupt and inappropriate or unpleasant. The present study confirmed that social factors do play important roles in the use of hedges in Japanese spoken discourse. Younger speakers exploit hedges often, and create new forms and functions of hedges when they are talking among themselves. ACKNOWLEDGMENTS First, I would like to express special thanks to my husband, Jan. This dissertation would not have been completed without his constant encouragement. Also, my deepest gratitude goes to Professor Mutsuko Endo Hudson, who gave me much valuable advice for the present study and throughout my graduate course for eight years . Thanks to her, I became interested in this field and was able to develop an attitude of analytical thinking toward language. Next,Iwishtothankothercommitteemembersandprofessors, Dr. Susan Gass, Patricia Lunn, David Dwyer and Grover Hudson, for their valuable suggestions, and especially Professor Dennis R. Preston, who always encouraged and helped me. There are many Japanese people who cooperated with me on the data collection. I would like to thank the principles, teachers, and students at Komatsugawa High School, Edogawa High School, Komagome High School, and Tokyo University of Foreign Studies. Special thanks go to Dr. Fumio Inoue and Ms. Yoshiko Onodera who helped me find participants for the preliminary and present studies. Lastly, I deeply thank my mother, father, brother, and my friends, especially Kaoru Watanabe, Mika Harland, Terumi Niitsu and Takami Oda. iv TABLE OF CONTENTS LIST OF TABLES .......................................... xiii LIST OF FIGURES ........................................... XV CHAPTER 1 INTRODUCTION 1 . 1 . General remarks ................................... 1 1.2. Goals, hypotheses and findings .................... 4 l . 3 . Outline of the dissertation ........................ 8 CHAPTER 2 REVIEW OF THE LITERATURE ON HEDGES 2 . 1 . Introduction ...................................... 9 2 . 2 . Definition of hedges ............................. 10 2 . 3 . Content—oriented and social ly—oriented hedges ..... 1 l 2 . 4 . Hedges in English ................................ 15 2 . 5 . Hedges in Japanese ............................... 16 2 . 6 . Functions of hedges ............................... 22 2 . 6 . 1 . Discoursal functions ..................... 23 2 . 6 . 2 . Semantic and informational functions ...... 26 2.6.2.1. Classical and functional approaches to hedges ............ 26 2 . 6 . 2 . 2 . Vague category identifiers ....... 30 2 . 6 . 3 . Social functions ......................... 3 1 2.6.3.1. Politeness ...................... 32 2 . 6 . 3 . 2 . Self-protection ................. 34 2.6.3.3. Solidarity ...................... 37 v 2 . 7 . Social variables ................................. 39 2 . 7 . 1 . Sex ..................................... 39 2 . 7 . 2 . Age ..................................... 42 2.7.3. Style ................................... 45 2 . 8 . Preliminary study ................................ 47 2.8.1. Hypotheses and method of the preliminary study .................... 48 2.8.2. Results of the preliminary study .................... 49 2.8.3. Conclusions from the preliminary study .................... 53 2 . 9 . Summary ......................................... 55 CHAPTER 3 HYPOTHESES AND METHODOLOGY OF THE STUDY OF HEDGES 3.1. Hypotheses ...................................... 57 3.2. Method .......................................... 63 3.2.1. Data .................................... 64 3.2.2.Participants ........... 67 3.2.3. Procedures .............................. 68 3.2.4. Units of talk ............................ 70 3 . 3 . Hedges in the present study ....................... 70 3.3.1.Listofhedges ........... ..7..............1 3.3.2. Expressions to be excluded . ............... 76 3.4. Summary... ...................................... 83 CHAPTER 4 QUANTITATIVE ANALYSIS OF JAPANESE HEDGES IN CONVERSATION 4.1. Introduction .................................... 84 4 . 2 . The results for each hedge ........................ 84 4.3. Sociolinguistic variables and the use of hedges ............................... 87 4.3.1. Age ..................................... 88 4.3.2. Sex ..................................... 92 4.3.3. Style ................................... 96 4.3.4. Comparison between the four groups and interaction among the three socialvariables ........................ 100 4.4. discussion ..................................... 104 4.4.1. Explanations of the results forthethreehypotheses ................. 104 4.4.2. Overall results excluding the use of toka’orsomething'andnanka’like’ ....... 113 4.5. Summary ........................................ 116 CHAPTER 5 TOKA (’OR SOMETHING’) AND TEYUUKA (’OR RATHER’ ) IN SPOKEN DISCOURSE 5.1. Introduction ................ . ........................ 118 5. 2 . The use of toka ( ’or something’) ........................ 118 5.2.1. General characteristics of toka in previous studies ...................... 119 5.2.2.Immediatelinguisticcontextsfortoka ........ 121 5.2.3.Basicfunctionsanddistributionsoftxka ....... 123 5.2.4. Analyses of the use of toka ................. ....127 5.2.4.1.Tokaforinexhaustivecoordination.........127 Vfi 5 . 2 . 4 . 2 . Toka for vague reference .................. 13 1 5.2.4.2. 1. Toka for uncertainty ................. 131 5 . 2 . 4 . 2 . 2 . Toka for approximation .............. 133 5.2.4.2.3. Toka for soft highlighting ........... 134 5 . 2 . 4 . 3 . Toka for vague quotation .................. 138 5.2.4.3.1. Toka for direct and indirect quotation ........................ 139 5.2.4.3.2. Toka for quoting other’s voice or speaker ’ s own voice .............. 141 5.2.4.3.3. Toka for self—quotation to avoid awkwardness or seriousness ........ 14 3 5.2.5. Summary of functions of toka and Schiffrin’s discourse model .................. 147 5 . 3 . The use of teyuuka ( ’or rather ’) ........................ 150 5 . 3 . l . Review of the literature on teyuuka ............ 15 1 5.3.2. General characteristics and immediate linguistic contexts of teyuuka ................. 152 5 . 3 . 3 . Basic functions and distributions of teyuuka . . . . 154 5 . 3 . 4 . Analyses of the use of teyuuka .................. 155 5 . 3 . 4 . 1 . Teyuuka for correction .................... 156 5.3.4.2. Teyuuka forrephrasing . . . . . . . . . . . . . . ..... .160 5.3.4.3. Teyuuka for specifying .................... 161 5 . 3 . 4 . 4 . Teyuuka for blurring ...................... 162 5. 3.4.5. Teyuuka for subtopic shift ................ 164 5 . 3 . 4 . 6 . Teyuuka for supplementary explanation ..... 1 65 5. 3 . 5 . Summary of functions of teyuuka and Schiffrin’ s discourse model . ................. 169 5.4. Summary ........................................ ......171 viii k——_—_ CHAPTER 6 RESULTS OF THE QUESTIONNAIRE ON THE USE OF HEDGES 6 . l . Introduction ......................................... 172 6 . 2 . Questionnaire results ................................. 1 72 6 . 2 . 1 . Perception of who used hedges .................. 1 74 6 . 2 . 2 . Situations when hedges are used ................ 1 81 6 . 2 . 3 . Reasons to use hedges .......................... 185 6.2.4. Discussion ................................... 190 6 . 3 . Psychological and social background of the use of hedges . . 1 93 6.4. Summary .............................................. 201 CHAPTER '7 CONCLUSION 7 . 1 . Introduction ......................................... 202 7.2. Quantitative and qualitative analysis of the use of hedges .................................... 203 7 . 3 . Behind the use of hedges ................................ 207 7 . 4 . Suggestions for future study ........................... 209 APPENDICES .............................................. 2 12 Appendix 1 . Transcription methods and conventions ......... 2 12 Appendix 2.1. Examples of hedges used in the preliminary study ....................................... 212 Appendix 2.2. Examples of hedges in Japanese writing used in the preliminary study ....... . ............. 214 Appendix 2.3. Sample questions used during the interview in the preliminary study ..................... 214 ix Appendix 2.4. Appendix 3. Appendix4.1. Appendix Appendix Appendix Appendix Appendix Appendix Appendix Appendix Appendix 4.2. 5.1. 5.5. 5.7. Appendix6.1 Appendix 6.2 Appendix6.3 Appendix6.4 List of Japanese participants in the preliminary study ........................... 216 Sample questions used during the interview in the present study ......................... 217 Examplesofhedgesusedforquestionnaire ...... 218 Questionnaire ............................... 219 Total number of words and tokens of hedges for each participant (the younger femalespeakers,jJ1chat) .................... 227 Total number of words and tokens of hedges for each participant (the younger femalespeakers,jJiinterviews) .............. 228 Total number of words and tokens of hedges for each participant (the younger malespeakers,ixxchat) ...................... 229 Total number of words and tokens of hedges for each participant (the younger malespeakers,ininterviews) ................ 230 Total number of words and tokens of hedges for each participant (the older femalespeakers,i11chat) .................... 231 Total number of words and tokens of hedges for each participant (the older femalespeakers,ininterviews) .......... ....232 Total number of words and tokens of hedges for each participant (the older malespeakers,inchat) ...................... 233 Total number of words and tokens of hedges for each participant (the older malespeakers,jjlinterviews) ................ 234 Detailed ANOVA results for all hedges . ........ 235 Post-hoc comparisons for both styles . . . . ..... 235 Post-hoccomparisonsforchat...... .......... 235 Post-hoccomparisonsforinterview..... ...... 236 x Appendix Appendix Appendix Appendix 9 . Appendix Appendix Appendix Appendix Appendix Appendix Appendix Appendix Appendix Appendix Appendix 6.5 Post—hoc comparisons between chat and interview .............................. 236 7. Detailed ANOVA results for hedges excluding toka ’or something’ and nanka ’1ike' ........... 236 Detailed ANOVA results for toka ’ or something ' ............................... 2 3 7 8. Detailed ANOVA results for nanka ’ like ’ ........ 2 37 10. Detailed ANOVA results for teyuuka ’or rather' ................................... 238 11.1 Detailed ANOVA results for answers to ”Do you use the underlined type of expressions[hedges]?" ....................... 238 11.2 Post-hoc comparisons for answers to ”Do you use the underlined type of expressions[hedges]?" ....................... 238 12.1 Detailed ANOVA results for answers to ”Usual way of speaking?” ...................... 239 12.2 Post-hoc comparisons for answers to ”Usual way of speaking?” ...................... 239 Detailed ANOVA results for answers to ”Do you think there is any gender difference?” . . 239 13.1 13.2 Post—hoc comparisons for answers to ”Do you think there is any gender difference?” . . 239 Detailed ANOVA results for answers to ”Were you able to talk naturally?" ............. 240 14. 15. Detailed ANOVA results for answers to ”Which age groups do you think use the underlined type of expressions?” . . . . . ......... 240 Detailed ANOVA results for answers to ”Should one refrain form using the underlined type of expressions depending on the occasion? " . 240 16. Detailed PEARSON results (Correlation matrix 1: ”Speaking with whom/ when do you use this type of expressions?”) ..... 241 17. Appendix 18. Appendix 19.1 Appendix 19.2 Appendix 20.1 Appendix 20.2 Appendix 21. Detailed PEARSON results (Correlation matrix 2: ”Why do you use this type of expressions?").................. ..... 241 Detailed ANOVA results for answers to ”Is there anyone around you who uses this type of expressions?". ....... . ....... . ...... 242 Post-hoc comparisons for answers to ”Is there anyone around you who uses this type of expressions?”.. ........ . ............ 242 Detailed ANOVA results for answers to "What do you think of using this type of expressionsixlconversation:Fun?"..... ..... ..243 Post—hoc comparisons for answers to ”What do you think of using this type of expressionsleconversation:Fun?" ............ 243 Detailed ANOVA results for answers to ”What do you think of using this type of expressionsiJ1conversation:Corrupt?" ........ 243 Appendix 22.1. Detailed PEARSON results (Correlation matrix 3: ”What do you think of using this type of expressions?”, the younger speakers) . . ...................... 244 Appendix 22.2. Detailed PEARSON results Appendix 23.1 Appendix 23.2 BIBLIOGRAPHY (Correlation matrix 3: ”What do you think of using this type of expressions?", the older speakers) .......................... 244 Detailed ANOVA results for answers to ”Do you think that this type of expressions are used to show closeness or solidarity?” . . . . . 245 Post-hoc comparisons for answers to ”Do you think that this type of expressions are used to show closeness or solidarity?” ..... 245 ............................................ 246 xfi Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 2.1 2.2 2.3 2.4 3.1 3.2 4.1 4.2 4.3 4.4 4.5 4.6 4.8 4.9 4.10 4.11 4.12 4.13 LIST OF TABLES Total tokens and rates of hedges (female) ......... 50 Total tokens and rates of hedges used by Group I and Group II in conversation (female) ...... 51 Tokens and rates of hedges used by Group I (female) ....................................... 52 Tokens and rates of hedges used by Group II (female) ....................................... 53 List of participants in Group I ................... 67 List of participants in Group II ................. 68 Results of each hedge by all participants ......... 85 Tokens and mean rates of hedges by age ............. 88 Tokens and mean rates of all hedges by age ......... 89 Tokens and mean rates of hedges by sex ............. 93 Tokens and mean rates of all hedges by sex ......... 94 Tokens and mean rates of hedges by style ........... 96 Tokens and mean rates of all hedges by style ....... 97 Tokens and mean rates of hedges in chats for the four groups ...... . ..................... 102 Tokens and mean rates of hedges in interviews for the four groups ............................. 102 Resultsforinteraction(ANOVA) ................. 103 Relation between the four groups in the frequency of hedges (post-hoc comparisons) ...... 109 Tokens and mean rates of hedges in chats for the four groups ............................. 1 15 Tokens and mean rates of hedges in interviews for the four groups ............................. 1 1 5 )dfi Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 4.14 5.1 5.2 5.3 5.4 5.5 5.6 Main effects and interactions of the three social variables (ANOVA, excluding toka and nanka) ............... 115 Tokens and mean rates of toka ’or something' ...... 119 Tokens of toka ’or something’ by functions in chats (younger groups) .................... 124 Tokens of toka ’or something’ by functions in chats (older groups) ....... . .............. 124 Tokens of toka ’or something’ by functions ininterviews(youngergroups) .................. 126 Tokens of toka ’or something’ by functions ininterviews(oldergroups) .................... 126 Variationsofquotativetoka .................... 138 Functions of toka and Schiffrin’s discourse model ......................................... 149 Tokensandmeanratesofteyuuka’orrather' ...... 150 Variants of teyuuka ’or rather’ ................. 153 Linguisticenvironmentforteyuuka’orrather’...153 Orientation of repair in teyuuka ’or rather’ ..... 154 Functions of teyuuka ’or rather’ ................ 155 Teyuuka as other-repair for the younger speakers ...................................... 159 Functions of teyuuka and Schiffrin’s discourse model ......................................... 170 my Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4.1 4.2 6.1 6.9 LIST OF FIGURES Use of hedges by four groups ..................... 101 Use of hedges excluding toka ’or something’ and nanka ’like’ ............................... 114 Answers to ”Do you use the underlined type of expressionsix1conversation?" .................. 175 Answersto”Usualwayofspeaking?" .............. 177 Answers to “Do you think there is any gender difference?" .................................. 178 Answers to ”Were you able to talk naturally? " ..... 179 Answers to ”Which age groups do you think use the underlined type of expressions in general?" ................................... 180 Answers to ”On what kind of occasions (casual/formal) do you use the underlined type of expressions?" .......................... 182 Answers to "Should one refrain from using the underlined type of expressions depending on the occasion?” ................................ 183 Answers to "Speaking with whom do you use the underlined type of expressions ? ” ................ l 84 Answers to ”Why do you use this type of expressions?" (YF) ............................ 186 Answers to ”Is there anyone around you who uses this kind of expressions?" .............. 187 Answers to ”What do you think of using this kind of expressions in conversation?” ....... 189 Chapter 1 Introduction 1.1. General remarks Language allows us to express ideas relatively precisely or imprecisely. .A speaker often uses hedges when he/she wants to express ideas, emotions, and attitudes imprecisely; Hedges i are defined as ”words whose meaning implicitly involves fuzziness ” (G. Lakoff 1972: 195) . More specifically, hedges are expressions of the speaker’s uncertainty (Coates 1993), ”tentativeness and possibility" (Hyland 1996), or ”imprecision or qualification” (Crystal 1997). English phrases such as I think, sort of, and perhaps are good examples of hedges (Coates 1993). The present, study investigates the use of hedges in Japanese conversation, and the relation to some social variables such as age, sex, style (speech genre). Hedges have often been associated with the mitigation effect , softening the force or directness of the utterance, and have been considered as a part of a ”politeness” strategy to:minimize the possibility of ”face threatening” in conversation (Brown and Levinson 1987 ; Holmes 1995) . Hedges are sometimes used to avoid potential disagreement or qualify the speaker’s commitment (e.g. Hyland 1996; Satake 1995; Hayashi 1997) , or to facilitate convivial conversation (n: discussion (Hyland 1996; YOnekawa 1998). In the present study, hedges are defined as expressions of uncertainty, possibility, tentativeness, and approximationwhich convey a sense of vagueness. Hedges in this study include some modal verbs, modal auxiliaries, and modal adverbs such as, respectively, to omou ’I think that’ , kamoshirenai ’may; might’ , and tabun ’probably’ . Also included are some unspecifying tags such as, toka ’or something’ , a filler nanka ’like’ , and a connective teyuuka ’or rather' . The detailed explanation of the definition and examples of hedges are presented in Chapter 2. It has been widely observed, often in a negative light, that a noteworthy feature of Japanese spoken by the younger generation is that it is unnecessarily vague with many types of hedging devices (e.g. Sakake 1995; Honna 1999; Degawa 1995; Maynard 1997; Chiba et a1. 1999) . However there is little empirical evidence to substantiate these claims. Generally, authors rely only on impressionistic observations, and present decontextualized constructed example sentences. The present study provides substantial evidence in order to verify the above observations by comparing conversations by younger and older native speakers of Japanese. The main framework adopted here is Schiffrin’s discourse model (1987) . Basically the model claims that discourse markers such as y’know, I mean, and well should be considered as multifunctional expressions, and that the functions of discourse markers emerge from the interaction among their semantic meaning, 2 sequential position, and the speaker—hearer and/or speaker- utterance relations. Schiffrin claims that understanding the function of discoursezmarkers.requires the examination.of their use from semantic, textual (sequence of utterances and turns), pragmatic, and social points of views. Although the focus of the present study is on the sociolinguistic examination of the use of hedges using quantitative analyses, hedges such as toka ’or something’ and teyuuka ’or rather’ are qualitatively examined in Chapter 5 to account for the linguistic contexts and functions using Schiffrin' s discourse model . More explanation on her model is presented in Chapter 2. As stated in Inoue (1986: 328), Japanese dialectologists have studied language differences according to age groups to investigate the process of language change. Inoue investigated sociopsychological factors of linguistic changes among the younger generations in.metropolitanflTokyo and its vicinity, and found that youth language is highly correlated with changes in the Japanese lexicon. Loveday (1986 : 305) also suggests that the linguistic behavior of students strongly influences everyday language, and ”many of their innovations eventually find their way into the standard variety". Thus studying differences between the language of youth and the language of their elders is interesting because it gives us some insights into how the language as a whole may evolve. It is hoped that the present study will help to characterize the 3 ongoinglinguisticchangesbeyondaimerelexicaldescription. It is also hoped that this study will make a contribution to the field of discourse analysis in general, and to the study of social influences on linguistic choice, as well as the system of politenessinJapaneseconversation,providingusefulinformation for Japanese language education, albeit in an indirect way. 1.2. Goals, hypotheses and findings The goal of this dissertation is to investigate the use of hedges by younger and older speakers in Japanese spoken discourse by manipulating three social variables :age, sex, and style. In Chapter 4 , I present detailed quantitative analyses of hedges used in actual single-sex spoken discourse of younger speakers as compared to that of older speakers . The data were collected from participants of a younger group, high school students (age 17 to 18), and of an older group (age 50 to 69). The hypotheses to be tested are as follows. 1. Hedges are used more often by younger speakers than by older speakers. 2. Hedges are used more often by female speakers than by male speakers. 3. Hedges are used more often in casual speech than in formal speech. Thesethreehypothesesarerelatedtothreesocialvariables; age, sex, and style (speech genre). The hypotheses are all based on 4 the general idea that the use of hedges is motivated not only by informational demands, but also by sociolinguistic factors . For example, R. Lakoff (1975), Preisler (1986), and Coates (1987) claim that hedges are characteristic of women’s language, and Okamoto (1995) states that younger female speakers often use hedges to promote solidarity among themselves . The frequency of hedges is also associated with the formal/informal distinction of conversational style (Channell 1994, Crystal and Davy 1975, and Lehrer 1975). The reasoning behind these three hypotheses is described in full in Chapter 3. While vagueness can arise from lack of knowledge, it can also arise due to the nature and the atmosphere of the social interaction, which speakers take into account when communicating (Channell 1994). The use of hedges then can be thought of as a way of packaging or presenting information in a socially vague manner. This packaging may be motivated by different purposes such as nonconfrontation and self—protection (Satake 1995) , politeness (Brown and. Levinson 1987), and solidarity (Okamoto 1995; Okamoto and Sato 1992, cited in Okamoto 1995:315). This conforms to one of the common communicational strategies, ’to construct the content of the utterance in such a way as to achieve maximum agreeableness to the recipient’ (Maynard 1987). To test the above hypotheses, conversation data from single-sex dyads were collected from two age groups (younger and older) and from both sexes in two types of conversation setting 5 (interviews with the present investigator, a stranger to the participants, and chatting with a close friend). The younger group consists of 17— to l8-year-old high school students, and the older group consists of speakers in their 50s or 60s. The participants in both groups were born and raised in Tokyo or its neighboring prefectures (Chiba, Saitama, and Kanagawa). The interview took place first. Each interview lasted about 15 minutes , during which the participants were individually asked about ’how to spend free time’ and related questions. Then followed the chat with a friend. Each pair of participants, who were close friends, were left alone in a room to have a chat about ’trips’ for about 15 minutes. Both types of conversation were recorded with a Mini Disk (a portable disk recorder) . At the end the participants were asked to fill out a questionnaire. The recorded interviews and chats were transcribed for the subsequent analyses, excluding the first 2 minutes for consistency of the level of the participants’ involvement in the conversation. The results provided support for hypotheses 1 and 2 with significant main.effects (ANOVA), meaning that the use of hedges differed according to the speakers’ age and sex. The younger speakers used hedges more often than the older speakers. For example, the younger speakers often used phrase/sentence—final hedges such as toka ’or something’, omou ’I think’, kanji ’feels like, is like’, ki ga suru ’I’ve got a feeling’; adverbs such as nanka ’like’, kekkoo ’quite’, toriaezu/ichoo ’for now, 6 tentatively’ and tabun ’probably’; and a connective teyuuka ’or rather’ . Among these hedges, toka ’or something’ and nanka ’like' showed a high frequency in chats among the younger speakers, especially among the younger female speakers. There was also an interaction between age and sex; the younger female group used hedges the most of the four groups. Hypothesis 3 (regarding style) was not supported, even reversed, in the present study. Overall, hedges were used more often in interviews than in chats . However, post—hoc comparison showed that no significant differences were found between the two styles in all four groups. In interviews, information—oriented hedges may have been stylistically motivated, whereas affection—oriented hedges were promoted in chats. By closely examining toka ’or something’ and teyuuka ’or rather’ in context, it was found that these hedges were multifunctional and their primary functions varied depending on the speaker’s age and sex, and the speech situations. General characteristics of these hedges are that they appear as vague references, and that they both indicate alternatives. Toka is used for inexhaustive listing and teyuuka is used as a repair marker. Toka and teyuuka were often used among the younger speakers to show mitigation or solidarity, to express their youthfulness , playfulness or embarrassment, or to avoid being too serious when chatting with their friends. Using a questionnaireaon what the participants thought of the use of particular hedges, it was found that the younger and older groups have different attitudes toward these hedges. The older speakers showed stronger aversion toward these hedges than did the younger speakers . Unlike other groups , the most frequent hedgeusers,theyoungerfemalespeakers,consideredthefrequent use of hedges normal, and they claimed that their use of hedges was unconscious and expressive of solidarity, and.was influenced by the people around them. The quantitative and the questionnaire results were further discussed consideringche psychologica1.and social backgrounds of the contemporary Japanese youngsters. 1.3. Outline of the dissertation The organization of remaining chapters is as follows: the relevant literature, includingrny preliminary'study, is reviewed in Chapter 2. After presenting hypotheses and methodology in Chapter 3, the quantitative results of Japanese representative hedges in interviews and chats are discussed in Chapter 4, and the use of toka ’or something’ and teyuuka ’or rather' is qualitatively examined in Chapter 5. Chapter 6 discusses the results of the questionnaire on the use of hedges, and lastly Chapter 7 concludes the present study. Chapter 2 Review of the literature on hedges 2.1. Introduction Hedges in English such as I think, sort of, maybe, like, and everything, and and stuff like that have received a great deal of attention in the fields of pragmatics and semantics (e.g. G. Lakoff 1972; Dines 1980; Ward and Birner 1993; Miller and Weinert 1995) . According to Brown and Levinson (1987) , Holmes (1995) and Coates(1987, 1993) , hedges are used to weaken the force or directness of an utterance, and they are considered to be very important expressions in comparing women’s and.men’s expression of linguistic politeness. Generally speaking, authors have stressed discoursal, semantic, and social factors as major motivations for using hedges. In this chapter, definitions and examples of hedges are presented, and then follows a review of the main studies on hedges from these three perspectives. It has been reported that age, sex, and style (speech genre) , among others , are major social variables that are closely related with the use of hedges (e.g. , Okamoto and Sato 1992; Coates 1989; Holmes 1995; Lehrer 1975) . In the sections to follow, I discuss several issues on each of the social variables. Additionally, the results of my preliminary study are presented.at the end of this chapter. 2.2. Definition of hedges In a general sense, a ”hedge" means "a fence or boundary formed by a rOW'Of shrubs or low trees planted close together”, or it could also represent a non-committal or ambiguous statement( Webster’s Third New International Dictionary 1976; The American Heritage Dictionary of the.English Language 1992). In the study of pragmatics, this sense of hedges is extended to the discussion of ”membership functions for fuzzy set" (G. Lakoff 1972), as in the definitions below. ... a ’hedge’ is a particle, word, or phrase that modifies the degree of membership of a predicate or noun phrase in a set (Brown and Levinson 1987: 145) An application in PRAGMATICS and DISCOURSE ANALYSIS of a general sense of the word (’to be non-committal or evasive’) to a range of items which express a notion of imprecision or qualification (Crystal 1997: 182. A dictionary of linguistics and phonetics, the fourth edition). Semantically hedges also express ”tentativeness and possibility” (Hyland 1996: 433) , and uncertainty (R. Lakoff 1972; Coates 1993) . When hedges are discussed in terms of social functions, they are considered as a part of a wider system of politeness (in the sense of Brown and Levinson 1987) because they ”weaken or reduce the force of an utterance" (Holmes 1995:72) and ”mitigate(s) the possible unfriendliness or unkindness of a Statement” (R. Lakoff 1975:54). Hedges are: also ‘used 'to facilitate discussion or to reduce the risk of negation by expressing the statement with caution, and diplomatic deference 10 to the addressee (Hyland 1995 ) . These functions do not only appear within the lexical information expressed by hedges, but they also come into the surface as a result of the interaction between the semantic information of a hedge and the nature and atmosphere of the social setting. In the present study, hedges are defined as expressions of uncertainty, possibility, tentativeness, and approximation, which convey sense of vagueness. The present study deals with hedging devices in which vagueness is lexically and explicitly expressed, Hedges are also considered to have social.functions snxfli as expressing' indirectness, politeness, solidarity or self-protection, and facilitating the conversation. The hedges investigated in the present study include modal auxiliaries , verbs , adjectives, adverbs, particles, and.ea conjunction. II give examples of hedges in English and Japanese in section 2.4. 2.3. Content-oriented and socially—oriented hedges Hedges are multifunctional; they can provide a range of different functions, often at the same time (Hyland 1996); for example hedges express vague categories, speaker uncertainty, caution, or interpersonal factors of the conversation participants . Consequently, as mentioned in the previous section , authors have provided different types of definitions of hedges depending on the focus of the study. 11 Generally, two aspects of hedges are addressed. One is that hedges modify the degree of precision or possibility of the proposition (content—oriented), and the other is that hedges incorporate an awareness of social factors in the interaction (socially-oriented). Hyland (1995) explains that.hedges convey both affective and referential meanings. These two approaches to the functions of hedges are parallel to the general linguistic point of view: Many linguists distinguish a referential (or descriptive, representational, or cognitive) function of language from a social function (e.g., Gumperz 1964). Along similar lines, Prince et a1 . (1982) also identify two distinct phenomena in the use of hedges: ”shields” and ”approximators". They explain that ”shields" such as I think, it seems that explicitly encode the speaker’ 5 degree of commitment to the truth or precision of what he/she is saying. Consider, for example, Prince et al.'s examples (1). (1) a. His feet were blue. b. I think his feet were blue. In (1a) , the speaker is committed to the truth of the proposition that ”his feet were blue”, while in (1b) the speaker expresses his / her uncertainty and noncommittal attitude by adding ”I think” . Sometimes the degree of the certainty is expressed by the tone or intonation over ”I think”. However, such prosodic features of hedges are beyond the scope of this study. Hedges which express fuzziness within the proposition are 12 called ”approximators"(Prince et al. ’5 example (2a) and (2b)). (2) a. He has a somewhat low interior larynx. b. His weight was approximately three point two kilograms. Both (2a) and (2b) are unshielded statements about vague propositions. Prince et al. explain that ”approximators” indicate ”fuzziness within the propositional content” and that ”shields” are correlated with ”fuzziness in the relationship between the propositional content and the speaker” (p. 85) . Thus, Prince et al’s "approximators" can be considered as content- oriented hedges, while ”shields” are socially-oriented hedges. However, as stated in Hyland (1996 ), we cannot always distinguish the two cases . Because the two functions can overlap, we can never be totally sure what the speaker’s intention really is, even when the utterance is examined in context. Consider my examples in (3). (3) a. It will probably rain tomorrow. b. I probably disagree with your idea. c. Bill will probably leave town. The hedge ’ probably’ in (3a), if uttered on a cloudy day, may be used purely based on the speaker ’ s inference about the possibility of rain. In this case, ’probably’ in (3a) falls in the group of content—oriented hedges. On the other hand, ’ probably’ in (3b) can fall in the group of socially-oriented hedges when the speaker is sure that he/she is disagreeing, but adds this hedge in order 13 to soften the tone of the rejection. These two cases are relatively clear-cut. However, ’probably’ in (3c) can be regarded as either content- or socially—oriented, or even both, depending on the context. The use of ” probably” can be based either on the speaker’s conjecture and uncertainty (content-oriented), or on social factors of the conversation (socially—oriented). In the latter case, the speaker is certain of the information but wants to avoid telling it to the addressee, knowing, for example, that the information.might hurt the addressee’s feelings. Or it can be both, content- and socially-oriented, when the speaker is not 100 % certain of the information and wants to express the lack of confidence in order to be self-protectiveror to avoid being wrong. R. Lakoff (1975: 53—54) gives a similar example, ”John is sorta short” . She states that depending on contexts, ”sorta” can be used to indicate the speaker’s uncertainty and lack of confidence about the accuracy of the statement or to mitigate the utterance for the sake of politeness. Therefore, taking the above phenomena into consideration, the present study deals with both types of hedges without a priori trying to exclude one type or another. Thus, the quantitative analysis in Chapter 4 does not distinguish between the two types . However, the difference between the referential or content— oriented hedges, and affective or socially-oriented hedges, is 14 —__¥ important for qualitative analysis , and so is addressed in Chapter 5 with the goal of understanding why and how hedges are employed. 2.4. Hedges in English G. Lakoff (1972: 195) states that, ”some of the most interesting questions are raised by the study of words to make things fuzzier or less fuzzy”. The following phrases in (4) are his examples in English (1972.: 196). (4) sort of in a real sense, kind of in an important sense, loosely speaking in a way more or less mutatis mutandis on the _ side in a manner of speaking, roughly so to say relatively a veritable somewhat a true rather real mostly a regular technically -like quintessential(1y) -ish ... In Channell ’s (1994) study of vague language in spoken and written data, she states that ordinary language leaves roomzflmr speakers to be vague and to avoid.precision and the commitment associated with it . According to her (p. 18) , there are different ways in which speakers can avoid being precise: (1) the use of vague additives (such as around and or something like that, often combined with numbers), (2) vagueness by choice of vague words (e.g., thingummy , whatsit, and loads of), and (3) vagueness by implicatureu The present study focuses on hedging expressions in Japanese which belong to her categories (1) and (2). 15 Channell (1994) also claims that the use of vague language is frequent and very customary. Kennedy (1987) provides evidence for this proposal by quantifying the frequency of the use of approximation devices in printed text (e.g. , somehow, around and approximately) for quantities and degrees, showing that the use is indeed frequent. Although his data corpus is not large enough to generalize the results , he found that overall it contained 1 , 407 different types of vague expressions, and that there were 9,135 occurrences among 63,176 running words. This means that about 14 . 46 percent of the words , or one in every seven words , is a hedge. Prince et al . ( 1982) show that in some types of conversation the use of hedges is quite frequent. They investigated conversations among doctors , and claim that vagueness is ”typical" in medical discourse. They found between 150 and 450 hedges per hour (one every fifteen seconds) in speech by doctors. 2.5. Hedges in Japanese Vague expressions are considered as one of the most noteworthy features of Japanese youth language (e.g. , Honna 1999; Maynard 1997; Yonekawa 1998; Chiba et al. 1999). Let us first examine what ”youth language” in Japanese is like. The Japanese language that young people use, called wakamonogo (’youth language' ) , is different from that of the older population (e.g. , Maynard 1997; Satake 1995; Yonekawa 1998). Yonekawa (1998: 15) defines ”youth language” in Japanese as expressions which are used 16 by both male and female younger people (ages under or around 30) with their in-group>members for entertainment or to facilitate conversation, to show solidarity, to mitigate, or to be indirect. Inoue (1986) calls youth language ”new dialect forms”. He claims that the new dialect forms (1) are used more by younger people than by older people, (2) are regarded as stylistically informal by users themselves, and (3) have different forms from those in the standard language. They are not words that will soon die out, but are typical examples of natural language'changes. According to Maynard (1997) , however, youth language is not directly indexed to the speaker’s age; the speaker’s selection of speech style depends on situational, social, and cultural factors, in addition to age. The examples in (5) on the next page are hedging devices in youth language, which are taken from the latest Japanese encyclopedic annuals, Gendaiyoogo no .kiso chishiki [Basic knowledge of current terms](1997, 1998, and 1999) and Imidas (1999), and some journal articles (e.g. Satake 1995; Niiyama & Iwamura 1998). The classifications and translations are mine. (5) a. utterance-final expressions toka ’or something' mitaina ’is like’ -tari ’and such’ (t)te iu ka ’or rather; or what should I say’ kanji ’is like; feels like' kamo(shirenai) ’may; might’ ja nai desu ka/jan? ’isn't it?’ kana ’I wonder; maybe' 17 b. noun suffixes —kee ’-type’ -teki ’-esque' -suru hito ’a person who does...’ (about the speaker him/herself) -toka ’or something; and such’ c. adverbs toriaezu ’for the time being’ ichioo ‘generally; tentatively’ kekkoo ’quite; fairly' nanka (often as a filler)’like’ Nowadays these types of vague expressions are employed frequently by students and sometimes by young ”OL"(office ladies, female office workers) and young male workers (e.g., Yonekawa 1998). Let us now see how these hedges are used in conversation. Example (6) on the next page is a conversation in youth language presented in Niiyama and Iwamura (1998: 26-27). (Translations are theirs. The glosses and emphasis in bold for relevant hedges in the present study are mine. Whether the conversation is constructed by the authors is unknown.) See Appendix 1 for transcription conventions. (6) 1 Ken: Na, kurisumasu, doo suru? hey Christmas what (k) 2 Ruri: E? sore tte, masaka futari kiri In: kurisumasu oh that QT possibly two only GEN Christmas 3 toka yacchau tte koto? or something do QT case 4 Ken: E, a, .iya... sore ‘wa, imasara da-shii...na. oh um well it TOP late—for—now BE-and IP 18 5 Ruri:Da yonee.ne,ne, sore jmui sa, nyuyasumii, nankaa, Be IP IP IP that than IP winter—vacation like 6 ryokoo toka iki-tai kanaa, mitainaa. trip or something go—want I wonder like 7 Ken: E, maji? oh really 8 Ruri: Hawai tokaa... karibu-kai tokaa... Hawaii or something the Caribbean or something 9 Ken: Oioi, kanari, goojasu-kei jan. hang on quite extravagant—type isn’t it? 10 Ruri:Yappaa, minami no shima de .mattari futari no after all south GEN island on idly two GEN 11 bakansu tte kanji? vacation QT like 12 Ken: wake, wakaran... meaning understand—NEG Ken: Hey, what shall we do for Christmas? Ruri: What? Surely you ’ re not suggesting we do something special, just the two of us? Ken: Um, well... It’s a bit late for that...isn’t it? Ruri: That’s right. Hey, you know what? Rather than that, for the winter vacation, you know, I thinkrnaybe I’d rather go on a trip or something. Ken: What, really? Ruri: Hawaii...or the Caribbean, or... Ken: Eh, hang on. That’s a bit extravagant, isn't it? Ruri: Definitely something like a laid-back vacation for two on a tropical island. Ken: I don’t get it... Such phrases as tokaa ’or something’ and kanaa ’I wonder’ in the dialogue are emphasized variants with an elongated vowel for toka and kana, respectively. Niiyama and Iwamura (1998:27) explain some of the hedges in the conversation as follows. 19 (7) . . .toka(a): added on to a sentence without adding much meaning, it has the effect of softening the tone. nankaa: An adverb used to mean nantonaku (”somehow"). . . .kanaa/mitainaa: Suffixes added to soften the tone when one is stating an opinion or thought. . . .tte kanji?: A vague expression used to avoid making a definite statement. The conversation in (6) , along with the explanations in (7) , are found in Nihongo Journal, a book for second language learners of Japanese, to help the reader understand the general meanings and the use of hedges in youth language. Also listed are teyuuka ’or rather’, ...suru hito ’a person who does...’, ichioo ’generally, tentatively’ , -kei ’-type’ , -kankei ’-related' , -hoomen ’ -direction ’ , etc . as examples of recent vague expressions used by young people. Niiyama and Iwamura claim that young people tend to avoid making a definite statement, and tend to mitigate their utterances, and that this is effective when they want to avoid their responsibilities. I have introduced typical examples of hedging expressions in Japanese which are often considered as youth language . However, this does not mean that older people do not use them at all. Some of these expressions can also be found in older speakers ' utterances, though they may not be as frequent as in those of younger speakers . The way older people use these hedges may differ as well: older speakers use them more canonically (as explained in a language dictionary), with fewer variations in usage, than 20 younger speakers. Along with the hedges in (5 ) on p. 18 , there are other typical hedging expressions in Japanese, which are presented in (8) below. Detailed explanations for each hedge are presented in Chapter 3. (8) a. Utterance-final expressions (verbs, auxiliaries, particles) ka nanka 'or something’ nado/nanka ’and so on’ deshoo/daroo ’probably' 500 ’seem; look like’ rashii 'seem; I've heard’ to omou ’(I) think that’ ki ga suru ’I’ve got a feeling’ b. Adverbial phrases taigai/daitai ’generally; about; perhaps' tabun/osoraku ‘probably; maybe’ tashika ’perhaps’ aru imi de ’in a sense’ c. suffixes -kurai/gurai; -koro/—goro; atari ’about; around' These are not particularly considered as youth language, but they are typical hedging phrases in Japanese, equivalent to hedging phrases in G. Lakoff’s (1972:196) list. I decided to include them as hedges for investigation in this study, because they seem to be important and relatively frequent in Japanese conversation. In order to compare the use of hedges by younger and older speakers and to generalize about how often their utterances are hedged, we need to examine a range of representative hedges which is not limited to youth language. 21 Among the hedges in (5) and (8), differences may be noted in the degree of uncertainty, probability, or mitigation expressed by certain hedges, and the context in which they appear. Although a discussion of this issue would be an interesting approach to examine detailed functions of particular hedges, this line of inquiry is not pursued in the present study. 2.6. Functions of hedges Language has various devices to express vagueness in communication. But why do we choose to be vague? Channell (1994 : 194), in her discussion of vague language, lists the following main communicative purposes and situations. (9) 1. to give the right amount of information 2. to deliberately withhold information 3. to use language persuasively 4. due to lexical gaps 5. due to lacking specific information 6. as a form of displacement 7. for self-protection 8. for power and politeness 9. to create informality and atmosphere 10. in women’s language Vague expressions thus play important roles in expressing the speaker's knowledge and degree of certainty (informational and/or semantic function), in negotiating social interactions with other conversation participants in an appropriate way (social function), and in coordinating texts or turns of utterances in 22 a sequence (discoursal function). In what follows, I discuss major studies on the functions of vague language from discoursal, semantic, and social points of views. 2 . 6 . 1 . Discoursal functions Channell’s (1994) list, given in (9) , shows that hedges are multifunctional . According to Schiffrin (1987 ) , multifunctionality is one of the important features of discourse markers. This section deals with Schiffrin’s discourse model (1987) in relation to the discoursal function of hedges. Schiffrin (1987) analyzes some of the English hedges, such as well, I mean, and y’know, as discourse markers. She claims that these types of discourse markers are never obligatory, that they are syntactically diverse, and that they are sequentially dependent elements, which serve as ’contextual coordinates’ (p. 326-330) . According to her (p. 328) , discourse markers have the following characteristics. (1) they are syntactically detachable from a sentence (2) they are commonly used in the initial position of an utterance (3) they have a range of prosodic contours, e. g. tonic stress and followed by a pause, phonological reduction (4 ) they operate at both local and global levels of discourse, and on different planes of discourse Schiffrin (1987) proposes a discourse model which develops the idea that ”markers allow speakers to construct and integrate multiple planes and dimensions of an emergent reality: it is out 23 of such processes that coherent discourse results" (p. 330). In her discourse model, she proposes five dimensions or levels in which discourse markers play a role: (1) the exchange structure (2) the action structure (3) the participation framework (4) the ideational structure (5) the information state She claims that discourse markers can function on some of these five different levels at the same time, and that the resulting functions, all together, contribute to overall discourse coherence. The ”exchangestructure" dealsvdth.sequentialroles of utterances, or turn—taking, e.g. adjacency pairs, questions and answers, etc. The "action structure" deals with speech acts (which action is intended and.which action precedes or follows); e.g., requesting, point making, and warning. The “participation framework” is defined as speaker—hearer relations and speaker—utterance relations, e.g. showing the speakers’ commitment to or evaluation of the proposition. The "ideational structure" deals with the organization of e.g., showing propositions,orideas,topicrelations. The”informationstate" involves the management of speaker—bearer knowledge (what the speaker and hearer know) and meta—knowledge (what the speaker and hearer know about their respective knowledge). Discourse markers can be used in more than one dimension of the framework simultaneously with one primary function (Schiffrin pp. 316—317). For example, the expression y’know 24 EZ2Z;_______________________________l—au—L______________________________l_lllllllllllljlil generally has roles on various planes, as follows. information status (Y ’know focuses on the hearer’ 3 knowledge, what the speaker and hearer share, and what is generally known.) participation status (Y’know allows the speaker to check how the discourse is progressing, and it solicits affirmation of information.) ideational structure (Y ’know marks the speaker's knowledge or information.) exchange structure ( Y ’know is used at potential participation transitions , and it marks the speaker as an information provider.) Schiffrin explains that the multifunctionality of markers is due to the fact that utterances are contextualized in more than one component of talk, and that markers are realized in different discourse slots with different linguistic properties. I contend that hedges have particularly important functions in the informational and semantic dimensions; i.e., the information status and the ideational structure in Schiffrin’s model. Hedges qualify ideas, or what speakers know on a certain topic or what speakers want to acknowledge as what they know. Hedges also play an important role in the ”participation framework” because hedges often present the speaker’s non- committal attitude toward the proposition, or because hedges can be used to soften the force of the utterance to show indirectness or politeness to the hearer. In the qualitative analysis of the use of hedges (Chapter 5) , the functions of hedges are examined according to dimensions in Schiffrin's model. 25 In what follows I discuss the informational and semantic characteristics, and the pragmatic and social characteristics of hedges separately, while presenting different views of hedges and showing examples. 2.6.2. Semantic and informational functions Among the various semantic approaches adopted in previous studies on hedges or vague language, there are basically two directions. The older approach is oriented toward theoretical implications for classical logic. The more recent approach is oriented toward functional implications in the actual use of hedging expressions, and this is the approach the present study takes. In the following sections I briefly review the two approaches, particularly, G. Lakoff (1972), Channell (1994) , and Dine (1980). 2.6.2.1. Classical and functional approaches to hedges One of the important characteristics of hedges is the indication of vagueness. Semanticians and logicians have paid a great deal of attention to vague language because of the problem . . . 1 . posed by it for semantic models of meaning . According to G. 1 In truth-conditional semantics, meaning can be described in terms of the conditions in the real world under which a sentence may be used to make a true statement (Crystal 1997). 26 Lakoff (1972) , there are predicates such as ”red”, ”tall", ”bird”, and ”bold" that are paradigmatically vague. These predicates have borderline cases or have no sharp boundary, e.g. , ”a chicken is a bird” (p. 185) . These are sentences in which it is unclear whether or not the predicate applies . Keefe and Smith (1999: 16) state that this ”lack of any sharp boundary” phenomenon is a criterion of vagueness, and they define one feature of vagueness as ”boundarylessness". Peirce (1902) also states that a predicate is vague if it can have borderline cases. In these borderline cases or boundaryless cases, the predicates are not clearly true or false; they might be both true and false, or neither true nor false. Ullmann (1962:118) claims that words are generically vague and they are never homogeneous (i.e. they are context-bound). There is even a view that it is not just words which are vague, but that all language use is vague in some way (Russell 1923), or that the world itself may be vague as well, no matter how precisely one would try to describe it (e.g. , Ullmann 1962; Keefe and Smith 1999). As G. Lakoff (1972: 195-197) states, one of the functions of hedges is to reveal the degree of category membership, and hedges offer a way for natural language to cope with borderline cases. Such a position can avoid the problem posed by classical logic . Regarding the degree of membership in the category ”bird" , 27 for example, a chicken ranks as a peripheral member while a robin is a central member (Heider 1971) . Therefore, the degree of truth in ”a chicken is a bird” is not so clear-cut, and is typically seen as ”less true” than is ”a robin is a bird”. However, if a hedge is added, as in ”a chicken is sort of a bird”, or ”in a sense, a chicken is a bird” , then the truth value of the sentence becomes apparent. G. Lakoff (1972) claims that the full range of degrees of category membership is subtle and vague, and that hedges can clarify this . This is interesting in that hedges , which typically make expressions vague, actually heighten the truth value of a sentence. According to the epistemic view, ”vagueness is a type of ignorance” (e.g. , Williamson 1992; Keefe and Smith 1999:17), and lack of knowledge leads to vagueness (Ullman 1962; Crystal and Davy 1975). Williamson (1992) claims that the main feature of a vague predicate can be described loosely as its having a fuzzy boundary, and that this is due to lack of knowledge about where those boundaries lie. To those who hold that the concept of truth values is the central issue in analyzing vague language, Keefe and Smith (1999: 18) make an objection: ”it is undeniable that meaning is intimately connected to use”. If there is a sharp boundary for a particular expression, it must be drawn in virtue of how we use the predicate. This approach to the semantics of vague language, oriented toward its actual function in conversation rather than toward its 28 theoretical implications for classical logic, is the one adhered to in the present study. My position is similar to that of Channell’s (1994) in her study of vague language . She discusses a cognitive aspect of vague language in her list of purposes and contexts of vague expressions (pp. 184-186); i.e. ”lacking specific information”. Hedges are used by speakers who lack specific knowledge concerning a particular conversation topic . She states that this use of hedges illustrates the working of the Maxim of Quality in the theory of Co-operative Principles (CP) by Grice (1975): Do not say that for which you lack sufficient evidence . Writers and speakers use ”the Maxim of Quality to guide their choice of vague and precise presentation of quantities" and ”they restrict themselves to writing or saying what they know to be true at the time” (Channel: 186) . Channell also points out that we can sometimes observe that speakers provide clear indication of their lack of knowledge, for instance, I can’t remember. . . , but. In my view, the informational aspect is one of the important motivations for the use of hedges. Speakers are vague because they don’t know or they are not sure. In addition to this informational motivation, I think that there are other occasions when speakers are vague even when in fact they have access to the relevant information. In this case, the use of hedges is prompted by some contextual needs such as styles of conversation, relationships between the speakers, topics of the conversation, 29 etc. That is, how much information the speaker provides depends on the context. In other words, the speaker is following the Maxim of quantity: Do not say more than necessary (Grice 1975). Depending on the context , the speaker may not need to , or may prefer not to provide exactly what he/she knows, and the speaker then tailors his/her contributions in the conversation in particular ways , i . e . , to give the right amount of information for the purpose of the conversation. Such socially oriented motivations are discussed in section 2.3.3., following the section on ”vague category identifiers". 2 . 6 . 2 . 2 . Vague category identifiers Channell (1994) , Dine (1980) , and Ball and Ariel (1978) also take functional approaches in their studies of expressions which refer vaguely to categories, such as or something and and stuff like that. These expressions are identified as ”tags", ”set- marking tags”, or ”vague category identifiers”. The structure of a phrase with a tag is as follows (Channell 1994:120). Exemplar + Tag bread or something Tags are combined with a variety of grammatical categories, for instance, nouns, adjectives, and subordinate clauses. According to Dine (1980 : 22) , tags serve to ”cue the listener to interpret the preceding element as an illustrative example of some more general case". She concludes that tags are used 30 _: 11111 interchangeably to relate parts to unknown wholes. Similarly, Channell (1994:143) observes that ”vague tags are understood to designate categories , either conjunctively or disjunctively, ” and that ”hearers and readers need to draw on pragmatic information in order to identify the intended vague category”. Hearers and readers use in particular: (a) the surrounding linguistic context; (b) the purpose of the text or conversation; and (c) their world knowledge. The Exemplar + Tag structure (vague category identifier) is understood as an instruction to access a category based on contextual information and/or the world knowledge. This analysis is useful in analyzing certain hedges in Japanese, such as -tari ’and such’ and toka ’or something' . However, a pragmatic or social question still remains: Why and to whom do we use such expressions? Perhaps in some cases they are simply production fillers . They can also be used to socially mitigate the force of the utterances such as requests and S tatement S . 2.6.3. Social functions There are various social and motivational factors involved in the use of hedges. These factors are usually intertwined, and the precise motivation for employing a hedge may not always be clear, as Hyland (1996) states. What is clear is that hedges are part of a wider system of politeness (Myers 1989; R. Lakoff 1975) and that to be vague is one of the verbal behaviors that result 31 rbeying the rules of politeness. Closely related to less are the concepts of solidarity and self—protection, jether they are the three most widely discussed social :ions for vague language use. These motivations are Iced by social variables, particularly the age and sex of aker and the style of conversation. The following sections the relationships among these three motivations: ess, self—protection, and solidarity. .. Politeness ne of the main motivations for speakers to employ hedges 2 their softening effect to ensure a smooth and appropriate :ion among the conversation participants. According to d Levinson (1987) , this motivation of maintaining a smooth .ion is part of the more general concept of ”politeness”. milar lines of thought, Hill et al. (1986) state that ic politeness is to constrain human interaction by hing a mutually comfortable distance and by promoting >y considering the feelings of others. In Ide (1991: 64) , fortable distance is determined by factors of social age, power relation of the participants, the formality .on, and topic of the conversation. When investigating atic functions of Japanese hedges , politeness issues play :ant role . 32 According to Brown and Levinson’s theory of politeness ), hedges are one of the politeness strategies among :sation participants to redress the threat of losing one’s '. In the notions ”face—threatening” and ”face—saving”, ' is ”something that is emotionally invested, and that can t, maintained, or enhanced, and must be constantly attended interaction” (p. 61). Face has two components (p.62). negative face: the want of every ’competent adult member ’ that his/her actions be unimpeded by others. positive face: the want of every member that his/her wants be desirable to at least others. :terances can threaten face (face—threatening) or can save face-saving). Any act of protecting face is a token of mess: an act of protecting positive face is positive mess, and an act of protecting negative face is negative less. For example, Brown and Levinson regard acts such as .n—group identity markers, seeking agreement, and showing 1y and understanding as positive politeness, whereas they acts of being indirect, giving deference, etc. as negative Less. Brown and Levinson further explain that ”alternatively, er] may choose to be vague about his own opinions, so as be seen to disagree"(ibid.: 116) . In this sense, hedges Lsed to show positive politeness, since they show intimacy, ommon ground, or mark in—group identity. Hedges can also 33 >loyed for negative politeness; speakers are nonassertive committal so as not to interfere with the addressee ’ 5 freedom :ion. The use of hedges, in some contexts, may be one way >wing their politeness to others to maintain a mutually ftable distance between conversational participants while sting their ”face" needs. Depending on the contexts, this 'table distance may require the conversation participants w either intimacy, distance, or something inbetween. R. Lakoff (1975) presents three factors which determine sness: formality, deference, and camaraderie. Among these factors, politeness in Japanese has most commonly been sed in terms of ”deference", focusing on the choice of forms in pronouns, honorifics and other lexical items. In resent study, however, I focus on the role of the aderie” factor, an aspect of positive politeness in the use ges which considers the speakers’ social attributes. 2. Self-protection Self—protection in the use of hedges is motivated by the :’s ”wanting to avoid later being shown to have said or 1 something which is not true" (Channell 1994: 184-185). :5 often want to guard against the possibility of a faulty or explanation by being vague. Self-protection is related politeness strategy of Brown and Levinson (1987) because its effects is vagueness used to avoid disagreement. This 34 if motivation, self-protection, is speaker-oriented while revious one, camaraderie, is addressee-oriented or t-oriented. The example is from Channell (p. 188): [BBS Radio 4 news: police spokesman making a statement about hijackers at Stanstead] I can tell you that approximately eleven people are helping us with our enquiries. plains that the non-round number, ”eleven", suggests that aaker knows that eleven people are involved and this is not an approximation. However, the speaker’s official on requires extreme caution to safeguard against being shown to be wrong. Prince et al. (1982) investigated doctors ’ conversations, aim that vagueness is typical in medical discourse, due to bject matter being talked about and due to caution over lizations. They found between 150 and 450 hedges per hour rery fifteen seconds) in speaking among doctors, which is sequent: Well he had Mannitol, he had Laskx, un he had Albmium, un I had to believe he was hypovolemic and he seemed to correct them awfully quickly, un and I was wondering whether there as any. . .any renal problems, but un. . .basically hard to say (italics theirs) (p. 87) ;h the speaker is using very precise medical terms, he/she 1g many ”shields”(hedges) to mitigate the statement and r commitment. Prince et al . (1982) interpret that there is ent of self-protection in this . The authors speculate that :ype of hedge is very peculiar to professionals such as 's and lawyers who have a large stake in saving their Lsional face. The use of hedges for self—protection is also common in do writing. According to Hyland (1996) , hedges are a useful rportant social communicative tool to show the writer’s n, diplomatic deference to the views of colleagues, and ion in expressing degree of uncertainty. He analyzed 26 ific articles and found that hedges are very frequent and critical role in science. He also claims that hedges signal iter’s anticipation of opposition to a proposition and are .0 reduce the risk of negation. Thus, as he states, r-oriented hedges will involve self-protection by hedging nent” (p. 438). Hyland claims that ” [r]esearch articles clearly reveal the anship between a discourse community, standards of ige, and textual representations, and it is these in ition which motivate the use of hedges" (p. 452) . He also that the precise motivation for using hedges is often :. The motivation of self—protection is closely connected ritive limitations. One hedges one’s confidence in the :y of a statement when one does not know it well. The line L a speaker’s knowledge and self-protection is often th to draw. Cognitive factors and other motivations such .teness and self-protection are all closely related. In 36 .on, solidarity is another important motivation. .3. Solidarity Ide (1982: 373) claims that ”solidarity is formed among ipants who have the same interests and responsibilities in ," and ”a common cultural, social, or geographical ound is the major factor in creating groups of solidarity” . iew of solidarity is adopted in this study. The presence ence of solidarity is also determined in part by whether a member of the in-group or the out-group (Ide 1982 , Loveday Lebra (1976) explains that the Japanese establish ties primarily on the basis of group ties. Loveday 304) adds that a ”strong sense of inward versus external tions (uchi vs. soto) fosters a deep sense of solidarity rrporate identification”. In addition, solidarity is Lly associated with informality, closeness, and rapport rpeakers, and is formed in relation to many social factors 5 age, gender, and informality (Tannen 1996). )kamoto and Sato (1992, cited in Okamoto 1995:315) claim lidarity is one of the major motivations for young females masculine expressions and hedges. They analyzed the >nship between young Japanese women’ 5 choice of speech level ir degree of intimacy. The study shows that in conversation eir close peers , the participants ”often qualified strongly .ne expressions by giggling . . .or using hedges, such as a 37 itive tte ’that’ or the expression mitai na ’like’...”, and ey elect to use them to break the norms in an attempt to orce solidarity"(p. 486). Okamoto (1995:313) also'claims sedai-hoogen ( ’generation dialect ’ ) , or wakamono no kotoba .th language’ ) , ”serves to convey an image of youthfulness, fferentiate younger from older women, and thus to establish larity”. Yoshioka (1990):claims that.high—school students think.it portant to show solidarity or build rapport in communicating [ themselves. ‘He also states that words which enhance this 2 of solidarity are easily accepted and used often by rsters, and that such expressions help youngsters to form r own culture and/or community. Solidarity is associated with politeness because it is rlly oriented and serves to preserve or create a positive :ionship and.a mutually comfortable psychological distance zen speaker and addressee. .As such, solidarity can be seen ice—saving, as opposed to face—threatening in Brown and lSOfl'S sense (1987) . In fact, speakers often sound impolite rey do not use the language of solidarity and casualness, .ng unnecessarily formal distance from his/her good friends rsual occasions. Solidarity is also intertwined with.many social variables such as sex, age, and style of the :rsation, and helps to create an appropriate atmosphere for :onversation. 38 Social variables Solidarity, self-protection and politeness are all covert rs which cannot be assessed directly but have to be inferred a speaker’s behavior in a given context. As such, their ence on the use of hedges can only be investigated by changing ontrolling aspects of the context. Which aspects of the :rsation context determine the motivational disposition of aker? This list could be long, but I focus on three social .bles that have already surfaced during the discussion of the .er’s motivational disposition, namely, sex, age, and style. .ng these variables changes conversation contexts , and allows nvestigation of the use of hedges in spoken discourse. Below of these social variables is introduced separately. 1. Sex In this section, ”Lakoff’s hypothesis” and its related .es , and issues on Japanese women ’ 5 speech are discussed. Let .rst define the difference between ”sex” and ”gender" . In resent study , the term ”sex " is used to refer to the speakers ’ reported physical differences (natural gender) , either .e or male. On the other hand, ”gender" generally includes concept of a person ’ 5 sexual orientation , [line/ feminine/ neuter, regardless of his/ her natural gender. nguistics , ”gender” is also dealt with in terms of grammatical ription, e.g. if a certain noun is masculine/feminine. Such 39 .er” issues are beyond the scope of this study. R. Lakoff (1975) discusses the relation between women’ 8 use dges and nonassertiveness . According to her, women use more as than men, and this is due to the social norm that ”asserting Lelves strongly isn’t nice or ladylike, or even feminine” (p. She associates women’s speech style with the use of certain LiStiC features which weaken or mitigate the forces of an fance, e.g. , hesitations, intensifiers and tag questions, and 1g intonations on declaratives. Her claim is called the >ff hypothesis", and later researchers aimed to validate the 1 empirically. Preisler (1986) is one of these researchers, and she -des support for the ”Lakoff hypothesis”. She conducted a r on the expression of tentativeness examining recorded arsational data from groups of four people. Coates (1989) found a relation between sex, the speaker’s sensitivity to conversation topic, and the use of hedges in single-sex >urse. Coates claims that women’s frequent use of hedges 1rs to prevent the discussion of highly sensitive topics from 3 too face-threatening. She also states that ”women’s use dressee-oriented hedges to mitigate the force of an utterance >e seen as a strength rather than as a weakness” (p. 117- She attributes men’s apparently less frequent usage of as to their choice of topic : Unlike women, men generally avoid )nal issues. In this sense, hedges are a valuable tool that 40 discussions on sensitive topics especially for women. These studies have shown that sex is indeed correlated with se of vague language, with women using hedging expressions >ften than men. This finding may be attributed to several 5. One may be that women have their own language style which racteristically soft—spoken or nonassertive with a certain e of vagueness. Another may be that vague language is an nt of power relations, particularly with respect to ness. Holmes (1995: 151) points out that ”[w]omen tend to :e attention than men to the face needs of others , especially positive face needs”. Tannen (1990) observes that men’s ge is information-centered while women’s language is 1- or rapport-centered. [t is well known that the Japanese language has distinct and male speech styles (e.g., Ide 1991; Shibamoto 1985; .992c). In general, Japanese ”women's" speech has been .erized as polite, gentle, soft—spoken, non-assertive, and tic compared to ”men’s” speech (e.g. , Ide 1982, 1991; Smith )kamoto 1995) . However, contrary to the traditional views nese women’s speech, nowadays there are also claims that among contemporary young Japanese women has become less 1e” (Okamoto 1995; Philips 1997) and that the female speech :3 become diversified (Jinnouchi 1998:52) > date there have been no empirical studies which rally associate Japanese "women’s” speech with the use of 41 —# is. Although the frequent use of hedges by the younger ese generation has been discussed in some studies (e.g., e 1995 ; Honna 1999), the explanations are generally based 9 authors’ intuition or on self-constructed sentences. A itative study of Japanese hedges based on naturally occurring is thus called for in order to investigate how their usage fluenced by social variables such as sex and age. Age Age is generally considered one of the major sociolinguistic s that influence the choice of language in Japanese sation, more so than in English. Ide (1982: 336) claims ;e in particular is involved with the rules of politeness. Japanese youth language has received a great deal of ion for more than 100 years. Yonekawa (1995, 1998) , in his Lcal review of Japanese youth language, states that Japanese .‘S were already discussing youth language 130 years ago. 1) notes that the young female students’ use of very .ial and masculine words was criticized in newspapers and es even around 1900. The features of youth language have arget of criticism, considered as something bad or corrupt . , the controversial expressions that female students used )0 years ago are regarded as ordinary or even old-fashioned s . In this sense, youth language is an interesting topic ,ying how language changes . According to Labov (1994: 47) , 42 adolescents ’ (and preadolescents ’ ) speech is the cutting edge in the progress of language change. Yonekawa also suggests that young female speakers are relatively free from the social norm in language use, and that they tend to play an important role in language changesz. We now turn to two studies related to youth language and the use of hedges. The first is Philips (1998) , which discusses the age issue. Her study investigates the relationship among the social variables age, sex and formality, with respect to the use of discourse markers such as fillers , connectives, and interactive particles. She proposes that some Japanese expressions, such as nanka ’like’ , function as fillers which are ”associated with the speaker’s non—assertiveness and uncertainty about a prior or upcoming proposition”. Nanka is one of the most frequently used fillers among young female speakers in informal conversation. None of the three social factors, however, led to statistically significant effects in the use of nanka. Tsuji (1996, 1998, 1999) investigates the latest Japanese youth language in an effort to find the social and psychological motivations in relation to the type of friendship. In his study, 253 college students were asked in questionnaires to write down 2A3 one of the features of female speech, R. Lakoff (1975) cites Jespersen ’ 8 claim that new phrases are created more by women than by men and that women’s speech is the source of language change. 43 whether they use some of the expressions typical of youth language (e.g., toka ’or something’, teyuuka ’or rather’, mitaina ’it's like’ ) . They were then also asked to choose appropriate descriptions about their friendship (e.g. , the number of friends , and if they discuss their personal problems with their friends) . Tsuji. (1999) reports a statistically' significant correlation between the use of youth language and superficial associations with friends. Superficial friendships are, according to Tsuji, relationships that are defined by particular purposes and activities, not by close constant companionships. This study reveals some aspects of language style and younger speaker’s social attitude. Tsuji states that hedges in youth language contribute to mitigating their commitments to their utterances and to the avoidance of possible conflicts in their interpersonal relationships. Tsuji's studies are interesting because he approaches youth language from social and psychological points of View in order to investigate why the younger people speak the way they do. However, his findings are based on the participants’ self-report, which may not reflect actual usage. Also no comparison was made with an older group, and sex and dialect background were not controlled. 44 2.7.3. Style One of the three important social variables is referred to as ”style” in the present study. ”Style” means speech genres or conversation situations which are related to the formality of the speech setting. The word ”style" originates with Joos (1967), and Labov (1972) further develops the topic. Joos defines speech genre as ”style”, and divides it into five classes; "intimate style”, ”casual style”, ”consultative style”, ”formal style”, and ”frozen style”. ”Intimate style" is the most casual style of the five, and the degree of formality increases in the direction of ”frozen style". In his category, the interviews in the present study fall somewhere between ”consultative style” (conversation between strangers) and ”formal style”, and the conversation with a friend falls under ”intimate style” (conversation between equals with shared knowledge). Channell (1994:191—192) states that ”vague language is associated with informal conversational settings” and that ”level of formality and giving the right amount of information are closely related". Crystal and Davy (1975) also claim that speakers can, if they choose, be more precise or imprecise according to the type Of conversation, though they do acknowledge an influence of Cognitive factors (memory loss, lack of knowledge) on the use of Vague language. They give the following reasons for being vague socially: (1) the subject of the conversation is not such that it requires precision, and an approximation or characterization 45 _——— will do; and (2) the choice of a vague item is deliberate to maintain the atmosphere. This means that the use of vague expressions is influenced by the topic and/or atmosphere of conversation. One of the most important studies related to the level of formality is Lehrer (1975) . Lehrer investigated the vocabulary which was used when talking about wine in two kinds of settings: description tasks (professional context) and communication tasks (socializing context). Lehrer found differences in expressing vagueness according to the topics and contexts of the discussion. Wine experts need to communicate precisely, whereas wine lovers do not have to be so precise. Lehrer observes that ” [w]hen a need for precision and a scientific use of language does arise, as among enologists or shippers of wine, the vocabulary can be sharpened. . . ”(p. 920). The wine lovers employ less precise language in describing wine than do the professional wine drinkers , though ”they communicate well enough for their purpose” (p. 922) . The wine lovers’ conversation is characterized as casual or intimate style, while the wine experts ' conversation is characterized as consultative or formal. Formality of the conversation thus controls the level of vagueness in language use. In a casual setting, the focus of the conversation is more on socializing and showing solidarity than on describing wines precisely. This is one way of being polite or showing ”camaraderie", as proposed by R. Lakoff (1975) , because by being vague the conversation participants are less likely to disagree 46 with each other, and so can avoid offending each other. In Japanese conversation , the major social variables of age , sex, and formality are very important because they influence the speakers ’ choice of language. However, the results of the studies to date on the use of vague expressions in Japanese remain inconclusive on the effects of age, sex, and formality, because of a lack of empirical data from actual conversations . The present study shows to what extent and in what way these major social variables influence the use of hedges in actual conversations. 2 - 8. Preliminary study I conducted a preliminary study in 1998 in Tokyo, Japan, Which investigated the use of Japanese hedges by 10 younger and l 0 older female Japanese native speakers. Although the basic design of the preliminary study is the same as that in the present Study, there are a few minor differences. ( 1) the order of the interview and the conversation between friends : The dyadic conversation was recorded before the interview in the preliminary study; the order is reversed in the present Study. (2 ) the topic of the interview: The topic during the interview, 'Youth language”, in the preliminary study was changed to ”how to spend free time” in the present study. I Will explain why these changes were made, after the quantitative results of the preliminary study are presented. 47 —'—_—'_—'_—————— 2 . 8.1. Hypotheses and method of the preliminary study The main issue that was investigated in my preliminary study is the same as in the present study; how social variables influence the use of hedges in Japanese conversation. The data were collected from: ( 1 )15-minute talk between two female friends from the same age group about "trips they made and/ or places they want to go next” or any other casual topic ( 2 ) 15-minute individual interviews on ”youth language" ( I acted as the interviewer. The interviews were with the same participants as in the casual conversation.) Right before the interview session (2), the participants were asked to read a list of example sentences of Japanese youth 1 anguage including various hedging expressions (see Appendix 2 . 1 and 2 . 2) . Then, the participants were asked questions about their use and opinions of such expressions, e.g. if they use them, if so , with whom, when, why, etc. (Appendix 2 . 3) . These data served mainly as formal style spoken data, and in addition, as metalinguistic data on the use of hedges. The participants were all female; 6 university students and 4 high—school students for Group I and 10 housewives aged between 50 and 70 for Group II (see Appendix 2 .4) . Both data sets (1) and (2) were tape-recorded and transcribed for the subsequent analysis. 48 2.8.2. Results of the preliminary study Below are the tentative answers to hypothesis 1 (regarding age) and hypothesis 3 (regarding style) obtained from my preliminary study. Included are the total number of hedge tokens and the rate (frequency per 1000 words). The hedges with two or fewer tokens in both groups and both speech styles, are excluded from the tables. Table 2. 1 presents the overall result of the use of hedges by style for each age group. In general, twice as many hedges were employed by the younger group as compared to the older group: The younger speakers used one hedge per 11 words while the older speakers employed one hedge per 22 words. The difference in the rate between two groups is much greater in conversation than in the interview: Hedges are about 2.8 times more frequent in the conversation of Group I than in the language of Group II. In interviews , both groups of participants used more hedges than they did in conversations. An analysis of variance showed that the main effect of age was statistically significant, (F( 1, 18)=45.08, MSE=464.02, p < . 01) : Group I produced more hedges than Group II . The main effect of style (speech genre) also reached statistical significance, (F(1,18)=36.79, MSE=166.24, p <.Ol): Participants produced more hedges during interviews than during conversations in both groups . 49 Table 2.1 Total tokens and rates of hedges (female, n=20)total words of utterance: 14944 (Group I), 16585 (Group II) Group 1 Group2 ANOVA token rate token rate ' F p conversation 671 81.52 233 29.00 age 45.08 .000026* interview 696 99.47 516 60.52 style 36.79 .000053* total 1367 749 interac 2.77 n.s. -tion ** n.s.>.10, *p<.05, ** age x style In contrast, the interaction between age and style was not significant (p =.109). To obtain further insights into these effects, I conducted post-hoc tests for the rates of each particular type of hedging expression. Table 2 . 2 on the next page summarizes the conversation data with the numbers in bold indicating the higher rates and significant differences between Groups I and II. It shows that the total rate of Group I was about three time as high as that toka ’or of Group II. Particularly, the rates of nanka ’like’ , something’ , teyuuka ’or rather’ , kanji ’feels like’ , mitai ’is like’, kekkoo ’quite; fairly’, and jan/janai? ’Isn’t it?’ were markedly higher for Group I. Table 2.2 Total tokens and rates of hedges used by Group I and Group II in conversation total words of utterance: 8036 (Group I), 8130 (Group II) (female, n=20) hedge Group 1(n=10) Group 2(n=10) p tokens rate tokens rate nanka ’like’ 204 25.39 22 2.71 0.0001 toka ’or something’ 208 25.88 15 1.85 0.0001 teyuuka ’or rather' 19 2.36 1 0.12 0.0001 tari ’and suchlike’ 21 2.61 15 1.85 n.s. shi ’and what's more’ 12 1.49 27 3.32 0.063 kanji ’feels like’ 27 3.36 4 0.49 0.007 mitai ’is like’ 12 1.49 1 0.12 0.013 to omou ’I think’ 18 2.24 11 1.35 n.s. deshoo ’probably’ 5 0.62 14 1.72 0.065 kamoshirenai ’It might be’ 6 0.75 7 0.86 n.s. ki ga suru 3 0.37 1 0.12 n.s. ’I’ve got a feeling’ kana/kane/kashira ’I wonder” 27 3.36 30 3.69 n.s. maa ’kind of’ 11 1.37 9 1.11 n.s. kekkoo ’quite/fairly’ 25 3.11 8 0.98 0.023 tabun ’probably’ 1 0.12 0 0.00 n.s. janai desu ka? ’Isn’t it?’ 0 0.00 6 0.74 n.s. (formal) jan/janai? ’Isn’t it?’ 58 7.22 17 2.09 0.005 (casual) desho? ’right?/you know” 14 1.74 45 5.54 0.002 total 671 81.52 233 29AM) 0.0001 n.s.>.10, Tables 2.3 and 2.4 present the number of tokens and rates of hedge usage. Only hedges which Show significant.05 As for the speech of Group II (Table 2 .4) , the frequency of hedges in the interview is about twice as high as that in conversation. The most significant differences across styles were obtained for toka ’or something ’ , to omou ’ I think that ’ , and deshoo 'probably ’ . 52 Table 2.4 Tokens and rates of hedges (female, n=10) used by Group II hedge conversation interview tokens rate tokens rate P toka ’or something’ 15 1.85 76 8.99 0.0001 teyuuka ’or rather’ 1 0.12 12 1.42 0.019 shi ’and what’s more’ 27 3.32 11 1.30 0.033 kanji ’feels like’ 4 0.49 22 2.60 0.007 mitai ’is like’ 1 0.12 7 0.83 0.045 to omou ’I think’ 11 1.35 107 12.66 0.0001 deshoo ’probably’ 14 1.72 57 6.74 0.004 kamoshirenai ’It might be’ 7 0.86 21 2.48 0.048 kana/kane/kashira ’I wonder' 30 3.69 88 10.41 0.008 desho? ’right?/you know’ 45 5.54 11 1.30 0.001 total 233 28.66 516 61.03 0.0001 . 3 . . Interestingly, the use of to omou ’ I think that ’ increased during interviews for both Groups I and II as compared to conversations. One explanation for this may be that the speaker uses to omou ’I think that’ to acknowledge that a certain opinion ’I think’ is personal and therefore may be incorrect. To omou seems to be a very common way of responding to questions in an interview. 2.8 - 3. Conclusions from the preliminary study The results of the preliminary study support hypothesis 1 (more hedges are used by younger speakers), but not hypothesis The younger speakers used a much 3 ( fewer hedges in interviews) . greater number and variety of hedges than did the older speakers . x In lnterViews the formal verSion of to omou, to omormasu, lS mostly used. 53 Unexpectedly, the data showed a higher frequency of the use of hedges in interviews than in conversations among the older speakers. However, the frequency is not as high as that of the younger speakers . It seems that the motivations for using hedges and the preferred types of hedges differ between the generations and speech contexts . This difference is investigated in Chapters 4 and 5 in the present study. The results of the preliminary study did not support hypothesis 3, concerning the influence of formality. (See Chapter 4 for the results of the present study regarding this hypothesis.) I concluded that the unexpectedly high frequency 0 f hedges in the interviews may have been caused by the particular topic chosen for the interview, ”youth language” , with which Group I I participants may not have been familiar. Even Group I participants seemed to have a hard time talking about hedges , even though they claimed to use them often. Therefore, the topic of the interview was changed to ”how to spend free time” for the Present study, which both younger and older participants may find relatively easy to talk about. This way, the cognitive demands 0f answering questions during the interview are expected to be more comparable to those made by conversing between friends . With comparable degrees of cognitive demands , I expect that an informal setting (chats among friends) leads to a higher frequency of hedges than a formal setting (interviews). 54 In the present study the interview data on the topic of "free time" serve as spoken data in a formal setting, Metalinguistic data on the use of hedges were obtained by means of questionnaires after the interview and dyadic conversation” Also, in contrast to the preliminary study, the interview was recorded first in the present study, preceding the dyadic conversation. This is because I presumed that the participants would feel more familiar and comfortable*with thernicrophone environment after the first recording. Such feelings of comfort and familiarity were deemed more important for recording casual conversation between friends than for interviews . See Chapter 3 for details of methods of data collection. 2.9. Summary In this section some of the theoretical backgrounds of the study of vague language and hedges in English and Japanese were presented. There are various aspects in the use of hedges, and due to their multifunctionality the examination is complex. Researchers in the past have approached the study of hedges from different perspectives, including cognitive (e.g., Williamson 1992), semantic (e.g., G. Lakoff 1972), sociolinguistic (e.g., Lehrer 1975; Preisler 1986; Philips 1998; Tsuji 1999), and discoursal (Schiffrin 1987). The use of hedges has also been associated with several important social aspects of the conversation contexts such as age and sex of the conversation 55 — — participants, and style of the conversation settings. However, these social variables are often investigated or discussed separately, and correlations among the variables are scarcely studied empirically. The present study takes a synthetic approach, and quantifies the influence of these social variables on the use of Japanese hedges. 56 Chapter 3 Hypotheses and Methodology of the study of hedges Chapter 2 showed that hedges are an interesting and complex area to study, and that several factors and social variables influence the use of Japanese hedges . Based on previous findings , I posit three hypotheses about the use of hedges in Japanese conversation. 3.1. Hypotheses The following are three hypotheses in the present study. 1. Hedges are used more often by younger speakers than by older speakers. 2. Hedges are used more often by female speakers than by male speakers. 3. Hedges are used more often in casual speech than in formal speech. Hypothesis 1. Hedges are used more often by younger speakers than by older speakers. Hypothesis 1 is based on the following. As mentioned in Chapter 2, age is one of the significant variables in Japanese sociolinguistics (Loveday 1986: 302). Japanese hedges such as toka ’or something’, mitaina ’is like’, te kanji ’feels like’, and tari shite ’do things like’ typically occur in youth language 57 -_f—-'_—‘ (e.g. , Jinnouchi 1997; Maynard 1997; Satake 1995) and they primarily serve to establish group membership and solidarity among the younger speakers (Okamoto 1995). Older speakers may use some hedgingwexpressions, but they are expected to use fewer variations of hedges and to use them less frequently than do younger speakers . Here I briefly discuss the issue of age grading. ”Age grading" in the present study implies that the older used hedges more often when they were young than now, but gradually decreased the usage, and that the younger will follOW‘thiS pattern. The frequent use of hedges with many variations seems a recent trend and new practice in Japanese youth language according to Japanese encyclopedic annuals , Gendaiyoogo no kiso chishiki [Basic knowledge of current terms] and Imidas , and some journal articles (e.g. Satake 1995; Niiyama & Iwamura 1998) . However, I cannot exclude the possibility of age grading in the use of hedges. The younger speakers who use some hedges often now may decrease the use when they become older. A future longitudinal study should verify this point. Younger speakers use youth language mainly with their in—group members (Yonekawa 1998), and some youngsters adopt stylistically informal forms to show their group membership, i.e. intimacy'with peer group>members (Loveday 1986). This leads us to speculate that one of the main functions of hedges is to create solidarity with an in-group addressee in a casual setting, where the main purpose of the interaction is to maintain rapport. 58 It is reasonable to assume that both younger and older speakers express their solidarity in some ways when talking to their friends, but how this is done in each age group may be different. The frequent use of certain types of hedges by younger speakers likely constitutes one of the features of youth language, which facilitates their sense of solidarity and in—group identification. This is not to say that older speakers do not try to promote solidarity and in-group identification, but rather that they might use means other than hedges to achieve such goals . For example, according to my informal observation, older people tend to often use the sentence final particle ne and the expression soo desu ne (’That’s right. ’) which mark shared information. They may also have more physical contact than younger speakers (e.g. lightly patting the addressee’s arm or shoulder) and talk about shared past memories. Here ”younger” speakers are defined as teenagers who are students, while ”older” speakers are defined as those aged over 50 . For the ”younger group" , I chose students because the language of students is often free from the norm in standard Japanese, and students are a major force in creating new words and styles (Yonekawa 1998) . Young working members are excluded because once youngsters start to take up professional roles in society, they use formal language more often than students , and so their language is more confined than that of students. For the ”older group”, I chose speakers aged over 50 years old because in general their 59 life styles are considered very distinct from those of the student population. Although it will be interesting to investigate the language of speakers from the age groups between these two groups, I limit my focus in the present study to establishing whether there is an influence from the age factor at all. This can be done most efficiently by comparing two very different groups , before further investigation of more subtle differences. Hypothesis 2. Hedges are used more often by female speakers than by male speakers. Hypothesis 2 is based on the gender-related studies on the use of vague language discussed in Chapter 2 . Women often speak more politely than men (Lakoff 1975; Okamoto 1995) , and women’s speech is less assertive than men’s speech (Smith 1992c) . Also, women’s speech is associated more strongly with frequent use of hedges than men’s speech (Lakoff 1975; Coats 1987; Preisler 1986). In Japanese sociolinguistic studies, Ide (1982, 1991) and Smith (1992c) observe that Japanese women tend to use more softening . . . 4 . . or eVidential expreSSions to make their statements polite, nonassertive, gentle, and empathetic . Previous studies thus have 4 Evidential expressions are postverbal forms/ sentence extensions that express speaker ’ s judgment or conjecture , report hearsay, present quotations, etc. 60 focused on stylistic differences in men ’ s and women ’ s conversation . Although we must be careful not to stereotypically associate the use of vague language with sex, whether women have their own styles in expressing vagueness is still an interesting topic. If the above claims are valid, women ’ 5 language should have more hedges than men’s language because hedges can function to mitigate propositions . Although it has been observed that young Japanese women ’ 5 speech nowadays is much less ”feminine” than that of older women’s (Shibamoto 1985; Okamoto 1995; Philips 1997), I speculate that, regarding hedges, young women’s speech still preserves a greater degree of nonassertiveness and softness as compared to young men’s speech. Hypothesis 3. Hedges are used more often in casual speech than in formal speech. Chapter 2 discussed that the level of formality is closely related to one of Grice ’ s conversational maxims (giving the right amount of information), and that this may influence the use of vague language (e.g., Channell 1994; Lehrer 1975) . The ”formal speech” in Hypothesis 3 refers to speech in which familiarity or solidarity is lacking among participants, and the occasions are formal(Ide 1982:372-375) . In this sense, formality depends not only on who are the speakers, but also on the structure of the 61 conversation (e.g. rigid structure of questions and answers in interviews) . In such a conversational setting, speakers are most likely to choose a formal style of language, and the use of hedges as solidarity markers may be absent. In an informal setting, the use of vague expressions can be dictated by the conversational maxim of quantity (the contribution should be as informative as is required for the current purposes of the exchange, and should not be unnecessarily informative, Grice 1975). While vagueness can arise from lack of knowledge, it can also arise from the topic or atmosphere of the conversation. For example, certain formal situations are likely to require more precise and specific speech than are informal situations. The speakers adjust the appropriate amount of information for their purposes in the communication. Thus , I explained how each of these three social variables may influence the use of hedges separately. However, these hypotheses are interrelated with each other. First, I observe that the use of hedges may be an exponent of power relations in society. Both women and teenagers are in general regarded as the groups with less power or authority in Japanese society than men or older groups. In casual conversation between equals, it is less required to express power relations in speech than in formal conversation. The lack of power is likely to lead to ”powerless" language and a self -protective or non-committal attitude . Hedges are one of such ”powerless" and/ or self-protective expressions 62 due to their fuzziness, or indirectness. Secondly, the reason that ‘women and youngsters are associated with the use of hedges is that they tend to be more free from standard social norms than men and older speakers (Lakoff 1972; Labov 1994) , as I mentioned in Chapter 2 . Women and youngsters might be playing leading roles in creating and establishing new hedging expressions in Japanese. This way, they may be enjoying breaking the norm, and creating their own styles for their entertainment purpose or for rapport-building. Oftentimes , this might be easily pursued in informal and in-group interactions. From.a different.perspective, however, it can be argued that women and youngsters have their own social norms, different from the social norms of men and the older. If the lack of knowledge is the only motivation to use hedges , then there will be no interaction between social variables and the use of hedges. Believing that social variables do have an effect on the use of hedges , I conducted the present investigation . 3.2. Method This section explains how the data were collected and analyzed. The method was designed based on the results of the preliminary study discussed in chapter 2. 63 3.2.1. Data The use of Japanese hedges is examined in relation to three social variables: age (Group I, ages 17-18 vs. Group II, ages 50-69) , sex (female vs . male), and style (casual vs . formal). Four sets of data were obtained: (1) individual interviews with participants from Group I (10 female and 10 male; ages 17-18); (2) casual dyadic conversations between friends from Group I (5 female and 5 male single-sex pairs; ages 17-18) , hereafter, referred to as ”chat”; (3) individual interviews with participants from Group II (10 female and 10 male; ages 50-69); (4) casual dyadic conversations between friends from Group II (5 female and 5 male single-sex pairs; ages 50-69) . hereafter, referred to as ”chat"; The participants for the interviews and chats are the same people for both age groups . In addition, the same participants were asked to fill out questionnaires about the use of hedges, as well as a few items of background information (see Appendix 4 . 1 and 4 .2) . First, I interviewed the participants individually to obtain formal—style speech samples, (1) and (3). The topic for all participants in both Groups I and II was ”how to spend free time” . After a few warm—up questions , the participants were asked about 9 questions for about 15 minutes (see Appendix 3). The prepared questions included yes—no questions, choice questions, and WH—questions in order to elicit various types of responses from the interviewees. I expected that the speech style would 64 5 . . . . . be formal in the interView data because the interViewer was a stranger to the participants and the style of interaction (interview style) was formal . Care was taken to maintain a natural interaction during the interviews, for example, by providing feedback and back-channel responses , and asking related questions . The recording took place in a quiet environment that was familiar to the participants such as in their class rooms or at their homes . Soon after the interview, dyadic conversation data were collected by asking a pair of participants to talk about ”trips (enjoyable trips they made, and/ or places they want to travel)" or any other familiar topic for about 15 minutes . The topic sheet was given to the participants right before the recording. Then, the pair of participants was left alone in the room. The conversation was carried out between two friends of the same sex from the same age group. For recording both the interview and chat, I used a Minidisk (a portable disk recorder) with a small two-way microphone. After the conversation, the participants filled out a questionnaire about the use of hedges. They also filled out the background information section, including items such as age, birth place, etc. Right before filling out the questionnaire, the 5 In Japanese, formality can be assessed by, among other things, the type of predicate endings: desu/masu (formal) ; ru (casual); and dearu (hyper formal, scholarly writing). 65 participants were asked to read. two short excerpts from conversations between two speakers (see Appendix 4.1). The two written conversation samples contain several kinds of hedges. Most of the hedges in the sample sheet are often regarded as typical in Japanese youth language. One of the conversation examples was based on an actual discourse between two female high-school students in my preliminary study, and the other example was based on the discourse of a male pair of high-school students. Some parts of these conversation examples were modified to simplify the discourse and to add a few varieties of hedges. The order of the two conversation samples was reversed for half of the participants. After reading the in“) conversation samples, the participants filled out the form with the multiple-Choice questions about the use of hedges: for instance, questions about whether they use these expressions, if so, with whom, when, and why (see Appendix 4 . 2 , for the complete form) . There were twenty multiple—choice questions in total and the order of the questions was counter-balanced among the participants . The purpose of this questionnaire is to collect metalinguistic data. Although admittedly there are some limitations in this type of self- reporting data, it nonetheless can yield interesting clues about the speaker’s perception of the use of hedges. 66 3.2.2. Participants The participants are all native speakers of Japanese, who were born and raised in Tokyo (n=36) or in the neighboring prefectures (Chiba, Kanagawa and Saitama, n=4) . Detailed information is presented in Tables 3 . 1 and 3 . 2 . The participants speak the Tokyo dialect which is ”the speech of the educated middle—class Tokyoites” , and has been ”considered as the standard language” in Japan (Kindaichi 1978: 60; Shibatani 1990: 186). Table 3.1 List of participants in Group I I .D. sex age hometown 1 YN female 1 7 Tokyo 2 FT female 1 7 Tokyo 3 YF female 1 8 Tokyo 4 YT female 1 8 Tokyo 5 TK female 1 8 Tokyo 6 YH female 1 8 Tokyo 7 YM female 1 8 Tokyo 8 HK female 1 8 Tokyo 9 YF female 1 8 Tokyo 1 0 MS female 1 8 Tokyo 1 1 EK male 1 8 Tokyo 1 2 ST male 1 7 Chiba 1 3 AY male 18 Tokyo 1 4 KK male 1 7 Tokyo 1 5 YM male 1 8 Tokyo 1 6 K male 18 Tokyo 1 7 MK male 1 7 Tokyo 1 8 TS male 1 7 Kanagawa 1 9 MY male 17 Tokyo 2 0 MY male 17 Tokyo 67 Table 3.2 List of participants in Group II I.D. sex age hometown 2 1 YO female 66 Tokyo 2 2 TW female 60 Tokyo 2 3 YY female 55 Tokyo 2 4 TS female 64 Tokyo 2 5 NT female 63 Tokyo/Shizuoka 2 6 SO female 62 Tokyo 2 7 EO female 68 Tokyo 2 8 TG female 65 Tokyo 2 9 SK female 50 Kanagawa 3 0 KN female 54 Tokyo 3 1 MM male 65 Tokyo 3 2 KU male 63 Tokyo 3 3 JT male 59 Tokyo 3 4 IT male 65 Tokyo 3 5 YM male 69 Tokyo 3 6 TY male 68 Saitama 3 7 Y8 (male 67 Tokyo 3 8 SK male 64 Tokyo 3 9 MD male 65 Tokyo 4 0 TF male 62 Tokyo The participants in Group I are high-school students whose schools are ranked as average in Tokyo (e.g. , Matsumura 1999 ; Imai 1999) : Tokyo Edogawa High School (public school) and Komagome High School (private school) . Group II participants consist mainly of retired men and housewives . There are 5 male and 5 female pairs in each age group, in total 40 participants . With all participants , data were obtained in both speech situations. 3.2.3. Procedures Ten-minute segments after the initial two minutes of all the interviews and conversations were fully transcribed and analyzed (see Appendix 1 for transcription conventions). The 68 initial two minutes of the interaction were excluded to make sure that all dyads were fully engaged in the conversation from the start of the transcription. Based on the transcriptions , for each participant, the number of total words and the number of occurrences of each type of hedging expressions were counted. Then the relative frequency per 1000 words (hereafter, mean rate) was calculated for each type of hedge. In total, 400 minutes of interview, and 200 minutes of conversation were transcribed and analyzed. For transcribing the utterances , I followed main ly the convention in Maynard ( 1987 : 18 ) , and therefore, not all morphemes are separated in the transcription. In counting words, each independent lexical item and function word (such as particles) were counted as one word. Inflections of independent lexical segments (e.g. negation.nai ’not’, past tense katta, and passive morpheme rareru) were not counted as a separate word. Repetitions and backchannellings were not counted when they appeared during the other participant ’ 8 turn of speaking. Analysis of variance (ANOVA) , post—hoc pairwise comparisons (Tukey’s honestly significant difference method [HSD] , see details in Ferguson and Takane 1989) , and PEARSON correlations were carried out to investigate the quantitative influences of social factors on the use of hedges (see Chapters 4, 5, and 6). 69 3.2.4. Units of talk Units of talk are a convenient and important concept for analysis since conversational interactions take place utilizing units of talk; for example, transitional periods between turns, back—channel responses, confirmation tags, etc. Maynard (1987 :23-24) refers to units of talk in Japanese as Pause-bounded Phrasal Units (PPU). According to her, PPU is defined by the phonological feature of a pause, which most often coincides with the phrase, that is , independent lexical items and junction words such as particles. PPUs are frequently followed by a ”pause— predicting tone and/ or pause—warning decreased speed, along with occasional stressed, rising intonation”(ibid. 24). In Japanese discourse, PPUs can be phrases or clauses, and PPU segmentation occurs roughly once every 1.7 seconds (ibid. 26). In the present study, I use PPU as defined by Maynard as the basic unit of analysis. In the transcription of the conversation data, the division of utterances is based on PPU. PPU is also useful especially in Chapter 6 which presents contextual analyses of the use of hedges, for example examining the relationship between utterances and turns. 3 .3 . Hedges in the present study This section presents the rationale behind my decision as to which hedges are investigated and which are not in this study. I list below hedges for investigation, and also provide 70 grammatical, semantic, and etymological information about them. 3.3.1. List of hedges In the present study I treat hedges in just one variety of Japanese, more or less standard Japanese as used by native speakers who were born and have been living in Tokyo or its neighboring prefectures (Kanagawa, Saitama and Chiba). This was done to minimize variation arising from regional differences. As G. Lakoff (1972:196) points out with his list of ”some hedges and related phenomena” (Chapter 2) , there are many possible hedging expressions . Hedges can include a broad range of lexical expressions or prosodic features, and it is sometimes difficult to draw a line between hedges and non-hedges. In this study, I limit my investigation to hedges which are lexically expressed, and which convey vagueness or uncertainty; for example, adverbs such as nanka ’like’, auxiliaries such as kamoshirenai ’may; might’ , and tags such as toka ’or something’ . The focus is placed particularly on hedges which are described in the literature as typical for Japanese youth language (e.g. Satake 1995; Niiyama and Iwamura 1998). I also include several other hedging expressions which were relatively frequently employed by both the younger and older participants in my preliminary study. The following is the list of hedges that I investigated in the present study. The explanations of each hedge are from A Dictionary of Basic Japanese Grammar (Makino and 71 Tsutsui 1986, henceforth DBJG) , The Kojien Dictionary, the fifth editions, (Shinmura 1998, hereafter KD), Supplementary Grammar Notes to _A_n Introduction t_o Modern Japanese Parts I and II (Simon 1986, hereafter SGN I and SGN II), and other sources . The hedges for investigation are classified into four major categories: utterance-final expressions, adverbs, connectives and noun suffixes. A. Utterance—final expressions (verbs, auxiliaries and particles) (1) toka ’or something; etc. ’ A combination of the quote marker to and ka ’or' . It can be used as a conjunction that lists more than one thing or as a quote marker often followed by iu ’say’. [DBJG] (2) ka nanka ’or something’ A combination of ka, a particle meaning ’or’ and nanka, a noun meaning ’something’. (3) nado; nanka ’and so on; things like...’ These are particles that indicate exemplification. ”Nanka is the informal, colloquial version of nado” [DBJG: 268] . Nanka and nado can sometimes give the sentence a derogatory meaning depending on the context. [DBJG] Nanka and nado with a derogatory meaning in a context conveying undesirability are not included in the list of hedges in this study. The Kojien Dictionary is written in Japanese. The translation is provided by the author. 72 (4) -tari; tari suru ’and such; do this, that, and others’ The inflection tari expresses an inexhaustive listing of action or states. [DBJG] "One meaning of this pattern is that one does certain things not in a sequence but in a random order. . . . One can list more than two verbs or just one verb. In either case, they are examples of various things that one does." [SGN II: 53] (5) kanji ’is like; feels like’ A noun expressing one ’s feelings caused by things, events, or persons. .Also expresses one’s impression or the atmosphere in a given situation. [KD] (6) kamoshirenai; kamo ’may; might’ ”K; and m9 are particles, and the verb shiremasen/shirenai is always in the negative. This set phrase itself is not inflected. . . ”. [SGN I: 90] This auxiliary indicates likelihood. [DBJG] Kamo can is the contracted version of kamoshiranai. (7) deshoo; daroo ’probably' Daroo is originally the informal conjecture form of the copula da, but is used as an auxiliary of conjecture. The formal version is deshoo. Deshoo/daroo indicates a higher probability than kamoshirenai ’may; might’ . But the speaker’s conjecture is not necessarily based on any evidence. [DBJG] (8) $00 ’seem; look like’ This auxiliary adjective expresses ”the speaker’s conjecture about what is going to happen or the current state of someone or something. Although this expression is based on what the speaker sees or feels, it is merely his guess and the degree of certainty in his statement is fairly low." [DBJG: 550] (9) mitai; yoo ’be like; look like’ Mitai and yoo are also auxiliary adjectives, and express the speaker’s conjecture about the likelihood of something based on what he sees or saw. The colloquial version of yoo is mitai. The degree of certainty in mi tai; yoo is higher than that in 300 ’seem; look like’ above. [DBJG] One function of this expression is to indicate the speaker’s uncertainty. [SGN II] 73 (10) rashii ’seem; I’ve heard’ Rashii is an adjective which can express the speaker ’ s uncertainty . It is interchangeable with above $00 and yoo. However rashii is used when the speaker has some personal evidence such as what he has heard or read. ”It implies the speaker’s judgment and uncertainty regarding the situation at the same time". It is equivalent to saying ’I’ve heard . . . , but I’m not sure’ or 'I’ve heard and it seems to be true that’. [SGN II: 122—123] (11) kana; kashira; kane ’I wonder’ Sentence-final particles which express the idea that the speaker wonders about. Kashira is usually used by female speakers, and kana is male speakers . Both are used in informal situations . Kane consists of the question marker ka and the confirmation or agreement marker ne. When combined, ka ne means ’I am not sure if..., but am I right?’. [DBJG] (12) to omou ’(I) think that’ ”The particle E, which indicates quotation, appears after a clause ending in the plain form and before the verbs . . . Unlike the English word that, the particle cannot be omitted. The subject of the verb omou/omoimasu is either ’I’ or ’you’ (in a question), and therefore it is not specified unless the speaker wants to contrast what ’I’ or ’you’ think with someone else’s thought or opinion” [SGN I: 64] Omoimasu is the formal form of omou. (l3) ki ga suru ’I’ve got a feeling’ Ki means ’feeling’ , and ga is a subject particle. Suru indicates that something is perceived by the speaker’s non—visual senses [DBJG]. B. Adverbial phrases (14) nanka ’like’ Nanka derives from nani ’what’ and the question particle ka. [KD] It expresses indirectness or the speaker’s vague feelings and hesitations (Hida and Asada 1994). 74 (15) toriaezu ’for the time being; for now' (16) ichioo ’generally; for the time being’ Both expressions can mean temporality. They add ”the nuance that the topic of the sentence is general rather than detailed, fleeting rather than prolonged, or temporary rather than permanent" (Suleski and Masada 1982: 8). (17) kekkoo ’quite; quite a bit’ It expresses that the degree is higher than what the speaker expected. It is an informal expression often used in casual conversation (Hida and Asada 1994). (18) taigai; daitai ’generally; approximately; perhaps’ Both taigai and daitai can be used as a noun or adverb to approximate things. When used as an adverb, taigai can also mean ’perhaps’ . [KD] (19) tabun; osoraku ’probably; maybe’ The formal version of the adverb tabun is osoraku. They are adverbs expressing probability and often used with the auxiliary daroo; deshoo ’probably' and to omou ’I think'. (20) tashika ’perhaps ’ The basic meaning of tashika is to refer to something that is certain, trustworthy, and reliable, meaning ’for sure’ and ’certainly’ . However a " further popular conversational usage of tashika carries meanings similar to the Japanese expression tabun ’probably; most likely’ . It indicates a slight uncertainty on the part of the speaker about the accuracy of the statement. . .” [Suleski and Masada 1982: 55-58]. In the present study tashika is considered a hedge when used in the latter case. (21) aru imi de ’in a sense’ Aru means ’ a certain ’ and is used when referring to things without specifying them; it only hints at their existence. Imi is a noun meaning ’meaning’. [KD] 75 C. Connective (22) teyuuka ’or rather; or what should I say’ Tteyuuka has many variations: tteyuuka, ( t )teiuka, tsuuka, teka, etc. All are derived from the quote marker to/ te ’that’ , the verb in ’say’ , and the particle ka ’or' that marks an alternative. To iu ka has become conventionalized as tteyuuka mainly to repair the speaker’s or the addressee’s utterances or to imply an alternative. D. Noun suffixes (23) -kee ’-type’ (24) -kankei ’-related’ (25) -teki; -ppoi ’—ish; -tic' These suffixes are used metaphorically by young Japanese speakers to avoid being direct. [Yonekawa 1998] (26) —kurai/gurai; -koro/goro; atari ’about; around’ Kurai/gurai approximates quantity or extent. ”Goro is different from kurai/gurai ’about’ in that the latter is used with specific quantity (time) expressions . . .”. Goro means ’approximately (with a specific point of time). [DBJG: 126-128] Atari sets an approximation for place, time, amount, and things. [KD] The above hedges can all express other meanings than vagueness or uncertainty. In the present study, these expressions are counted as hedges only when they indicate the above-mentioned meanings. The following section lists some of the expressions excluded from the present study. 3 . 3 . 2 . Expressions to be excluded At the onset of the present study a number of decisions were made as to which expressions would and would not be included among 76 hedges, because there are a number of ways to be vague. .Below I discuss briefly the linguistic forms or devices which may express vagueness, but which are not included herein. (1)Ellipsis (2)Hedges within quotations (3)Fillers (4)Features of intonation (e.g. high rising terminal, HRT, or ’half-question’[Satake 1995; Maynard 1997]) (5)Tag questions (e.g., desho? ’isn’t it?’) (1) Ellipsis One can bring about hedging effects indirectly by omitting phrases. In the case of omission, the utterance becomes vague by not completing the end of the sentence. For example, an utterance sounds vague when it ends with words such as kara ’because’ and kedo ’though’, omitting the main clause, which usually follows. I agree with Channell (1994), however in that omission qualitatively differs from additive vagueness or choice of vague words because of its unspecified character. According to Channell, such instances should be viewed not as expressing vagueness but as suppressing reference or mention. As it is difficult to figure out what is omitted or suppressed and what the speaker’s real intention is, I exclude the cases of vagueness by omission. (2) Hedges within quotations When hedges appear in quotations of what the speaker read or heard, they are not counted. For example, in English, if one 77 says ”Ken said, ’Mary is sort of cute’", the hedging expression ’ sort of ’ in the quotation is not vagueness that the speaker wants to convey him/herself. I regard hedges of this type as a mere quote, and not indicating the speaker’s intention to be vague. (3) Fillers Fillers such as eeto ’uh’, anoo “well', and.maa ’kind<1f’ are not included, although they can express vagueness due to their hesitation and softening effects . One of the major functions of fillers is to hold the floor while the speaker is searching for an appropriate expression (Maynard 1987) . They usually express no identifiable propositional meaning. Therefore, I regard the hedging aspect of fillers as peripheral to their usage, and I decided to exclude fillers from the list of hedges. One exception is nanka ’ like’ . This adverb often functions as a socially motivated filler to soften one ' s utterance (e.g., Philips 1998). I decided to include nanka ’like’ in the list of hedges because it has specific hedging functions that are derived from its original meaning. 'Unlike other fillers, nanka has an identifiable semantic meaning. It is a combination of nani ’what’ and the question marker ka. Philips summarizes the function of nanka ’like’ as indicating an approximation.of‘what the speaker has in his/her mind, which leads to the sense of uncertainty, vagueness and hesitation. 78 (4) Intonation Rising intonation patterns, often called high rising terminal (HRT) , have been described as an uncertainty sign (Lakof f 1975) . However according to Holmes (1995:103-106) , the primary function of HRT is, as in tag questions, an interactive device, ”checking out that the addressee shares the same background information as the speaker” and facilitating the addressee’s participation in the discourse. Saito (1998) also observes the same effect in ”half-question” or ”question—like intonation” in Japanese conversation. As Holmes (1995) indicates, HRT has not been researched sufficiently to prove that one of its main functions is to express uncertainty, and, therefore, I decided not to include HRT in my list of hedges. (5) Tag questions The same can be said for tag questions . The main functions of tag questions include epistemic , facilitative, challenging and softening (Holmes 1995). It is often difficult to determine whether or not certain tag questions are hedges expressing vagueness , due to their frequent facilitator roles in conversation. In the Holmes study (1995:83) women used facilitating tags more than epistemic tags. My preliminary study did include tags in the list of hedges . The tags in my preliminary study appeared to serve mainly to facilitate the interaction in the conversation. I found little 79 evidence that tags express vagueness or the speakers ’ uncertainty . The following is a typical example of the conversation between two young female friends recorded in the preliminary study. Here a speaker ’ I ’ is telling her friend ’Y ’ her opinion about the rumor that the Japanese like cameras. Tags appear in bold. (1) l I: Nihon j in ga kamera-zuki nan te uso da , Japanese SP camera like QT lie BE Tada tanni saa nanka hora nihon no saa only simply IP somehow hey Japan GEN IP koo kikai toka kekkoo saa this machine etc. quite IP ii no ga ooi janN. good NOM SP many TAG (Y: aa aa aa) yeah yeah yeah Bideo kamera toka, video camera etc. (I:) Ima hora dejitaru kamera toka aru janai, now hey digital camera etc. exist TAG (Y: 500 da ne) so BE IP Dakara, nanka sooiu no me ni tsuku kara, so like like—that NM eye catch because Motteru tte iu imeeji aru kedo, have QT say image exist though Nihonjin nanka dotchika hitotsu janN, Japanese somehow either one TAG (Y: Uh) yeah Kamera motteru ka bideo ka tte saa camera have or video or QT IP 80 9 Sonnani ippai mottenai desho7l, dakara nee. such that many have-NEG TAG so IP 10 Y: Fuun kodawari no doai ga yeah strong preference GP degree SP chigau n janN yappa. differ TAG after-all (I: Uh) yeah 1 I: It’s a lie that the Japanese like cameras. 2 It's just that, well, see, there are quite many good Japanese machines or something, aren’t there? (Y: yeah, yeah, yeah) 3 Video cameras and such. 4 Now, see, we’ve got things like digital cameras, right? (Y: That’s right) 5 So, well, that kind of thing catches people's attention. 6 There is an impression that the Japanese have (them), but, 7 the Japanese, well, own either one of them, don’t they? (Y: yeah) 8 They have either a camera or video. 9 They don’t have that many, right? So. 10 Y: Yeah, the degree of preference differs after all, doesn’t it? (I: yeah) When the speaker I presents her opinion, she uses tags four times. The speaker Y also employs a tag in her response. These tags seem to serve to catch the addressee’ s attention, rather than showing the speaker’s uncertainty. The tags mostly appear with falling intonation, and they are followed by back-channelling responses from the addressee such as ’yeah’. As Holmes (1995: 86) says ”[t]ags encourage people to talk" and there is ”the correlation between the leadership or facilitator role and a high 81 proportion of tag questions" . This is consistent with the observation by Coates (1989) who found that nearly all tags in the conversations between female friends were facilitative. Although it is not always easy to assign a primary function to a tag (whether a certain tag is epistemic or facilitative), it seems that these tag questions behave quite differently from other types of hedges to be investigated in the present study. My preliminary study shows that tags (desho? ’right?’ and jan?/janai ? ’ isn ' t it? ’ ) were used in conversation between friends more frequently'than in interviews (see'Tables 2.3 and 2.4 on p. 52—53) . This is because in interviews the style of the interaction is mostly questions—and-answers , and the participants mainly take a respondent role as an intervieweec There is not much room or need for the interviewee to facilitate the talk in an interview where the interaction is rigidly routinized. Contrarily, in conversation, the interaction is built based on both speakers’ cooperation, and the speakers often employ tags to let the conversation proceed smoothly. Thus , I came to the conclusion that, like HRT, tag questions should not be considered as pure hedges under the working definition in this study, and so I excluded them from my list. Obviously, rising intonation and tag questions are closely related to the issue of hedges, and are interesting areas to be investigated in further studies. 82 3.4. Summary My study aims to uncover facts about one of the latest phenomena in Japanese, the use of hedges by younger speakers as compared with older speakers, providing empirical evidence. In this chapter, I explained how I collected and analyzed the data, and discussed the hedges for investigation as well as about the rationale behind excluding a number of linguistic forms that have been considered as hedging expressions in other studies. In the following chapters I present the quantitative data on when Japanese native speakers use hedges and which hedges they prefer, and further discuss the pragmatic roles of hedges, which reflect the speakers’ psychological states and relative status in a: particular context. 83 Chapter 4 Quantitative analysis of Japanese hedges in conversation 4.1. Introduction This chapter presents a quantitative analysis of the use of Japanese hedges and discusses the results regarding the social variables, age, sex and style. First I will show the number of tokens, mean rates (per 1000 words), and percentages of 21 hedges which were most commonly used by all participants . Then overall data relative to each hypothesis will be presented with the results of statistical analysis (ANOVA.and.post—hoc comparison;‘Tukey's HSD method) . I will also discuss each hedge separately according to the three social variables, and speculate what caused the results . For ease of reference, I use the same order and numbering of hedges in all tables and throughout the text. 4.2. The results for each hedge Table 4.1 (next page) presents overall results of the use Of each hedging expression with the raw numbers of occurrences, the mean rates per 1000 words, and the percentages. The table Summarizes the data from all participants, in both interviews and conversations between friends. First, there appeared some Variations in the form of a few hedges. Especially with teyuuka 'Or rather’ (21) , variants such as tteiuka, tsuuka, and teka were sometimes employed (see Chapter 5 for detail). 84 Table 4.1. Results of each hedge by all participants (n=40) Jhedge gmeaning ltokens] rate* (%) phrase/sentence—final expressions (n toka or something 886 25.00 (23.98) (2) omou I think 469 12.89 (12.36) (% kana/kashira/kane I wonder 369 10.00 (9.59) (4) tari/tari suru do ...and such 312 8.33 (7.99) (5) nutai/yoo is like;look like 194 5.18 (4.97) (6) gurai/goro/atari about;around 177 4.65 (4.46) (7) kanji feels like;is like 3.14 (3.01) (8) deshoo probably 70 1.80 (1.73) (9) .nado/nanka and so on 62 1.53 (1.47) no) kamoshirenai/kamo may 37 1.11 (1.06) H1) rashii seems;I heard 33 0.98 (0.94) 02) ka nanka or something 28 0.69 (0.66) as) ki ga suru I've got a feeling 20 0.57 (0.55) H4) 500 seem;look like 18 0.54 (0.52) (15) kee ...type 12 0.32 (0.31) adverbs us) nanka like 499 14.15 (13.57) uj) kekkoo quite;fairly 150 3.93 (3.77) as) toriaezu/ichioo for now;tentatively 65 1.77 (1.70) (19) taigai/daitai generally;about 57 1.58 (1.52) (an tabun/osoraku perhaps 37 1.09 (1.05) connective (21) teyuuka gor rather 137 3.69 (3.54) (22) others 59 1.42 (1.36) (23) total 3802 104.26 (100.00) *rate = mean rate per 1000 words They are stylistic variants which are often influenced by the speaker's sex or the speech situation, but they do not differ in meaning. differences in the amount of utterances during 10 minutes. Therefore I decided to group them as one hedge. It should. also) be noted 'that there ‘were individual 85 Some participants were more talkative than others, especially some older male participants in interview situations (see Appendix 5 for detail). To even out such differences for the comparison between groups, the average tokens per 1000 words (referred as ’rate' in tables) were calculated for the analysis. In general, the result showed that, of all hedges, (l) toka ’or something’ was the most strongly preferred hedge, constituting 23.98% of the distribution of all hedges. Then follow (16) nanka ’like’ (13.57%), (2) omou ’I think’ (12.36%), (3) kana/kashira/ kane ’I wonder’ (9.59%), (4)tari/tari suru ’do ... and such' (7.99%) , and (5) mitai (na) ’is like’ (4 .97%) . These top six hedges constitute 72.46% of all the hedge occurrences. Hedges which occurred less than 10 times are not listed in the table, but are grouped together as (22)’others’ throughout the chapter. This category includes hedges such as nantonaku ’somehow’ , aru imi de ’in a sense ’ , ya nanka ’and something’ , and tashoo ‘more or less ’ . Examples of the most frequent six hedges are presented below. They are taken from the recorded conversation (henceforth, younger female speaker = YF, younger male speaker = YM, older female speaker = OF, and older male speaker = OM). (1) toka ’or something’ e.g. YF: Naraigoto mo hajimerareta shi, toka. lessons also start-could and ’And I could also start lessons, and stuff like that'. 86 (16) nanka ’ like' YF: Nanka atsukunai n da yo nanka attakai tte kanji. hot—not NM BE IP warm QT feel ’Like, it’s not hot [in Hawaii], y’know, like, it feels like warm. ' (2) omou ’I think’ YF: Zettai nihon-jin dekinai naa toka omou. absolutely Japanese can-not-do IP like ’I think like Japanese can never do it [dancing hard]. ’ (3) kana/kashira/kane ’I wonder’ YF: Uun uchi otoosan no hoo wa hataraki—sugiteru kana. ah my—family dad side TP work-too—hard ’Ah, I wonder, in my family, it’s my dad who overworks. ’ (4) tari/tari suru ’do .. and such’ 1 YM: Hoka no hito ga iru to yappa ki tsukattari other GP person SP exist if as-expected worry-about 2 shinakucha naranakute. have—to ’If there is someone else [with me on a trip], I have to worry about him, and such.’ (5) minai(na) ’is like’ YF: Nanka shinjatte 2-hiki ni natchatta mitaina. like have-died 2-dogs to have—become ’Like, my dog has died and I’ve now got two dogs, like. ’ 4.3. Sociolinguistic variables and the use of hedges In this section, the quantitative results of the use of hedges are discussed in relation to the three hypotheses. The effects of social variables (age, sex and style) on the use of hedges and the interaction effects between these variables, for the overall result as well as for the results with individual hedges , are statistically evaluated by ANOVA. The overall result showed 87 that all three variables had significant effects on the use of hedges, and that there was an interaction between age and sex. 4.3.1. Age This section presents a comparison between two age groups, group I (age: 17-18) and group II (age: 50-69) . Previously, Imade the following prediction regarding the age factor. Hypothesis 1: Hedges are used more often by younger speakers than by older speakers. Overall, this hypothesis was supported by a highly significant effect (F(1,36)=82.41, MSE=301.10, p <.00001) as shown.in Table 4.2. The younger group (Group I) used hedges twice as often as the older group (Group II). Detailed ANOVA results are given in Appendix 6.1. Table 4.2. Tokens and mean rates of hedges by age tokens mean rate P-value per 1000 words Group I (n=20) ___-£!§__§-_Y_*12________“2.3.9.9 _____________ 6. 2:211? _______ roup II (n=20) (OF & OM) 1403 34.520 .00001 The high frequency of the occurrences of hedges for the younger group is well illustrated in that the speakers in this group used approximately one hedge every 14.34 words, or one hedge per 10 seconds while the older group employed one hedge every 28 . 97 words , or per 17 seconds. The frequency in the younger group is even higher than the result in Prince et al . (1982) : one every 15 seconds 88 in medical spoken discourse. However, looking closely at the results of each hedge, the effect of age varied with the type of hedges. Table 4.3 shows the data of each hedge with indication of statistical significance . Table 4.3. Tokens and mean rates of all hedges by age lhedge Emeaning Group I Group II p phrase/sentence-final expressions token rate* token rate (n toka or something 687 20.04 199 4.96 .00001 (2) omou I think 254 7.48 215 5.42 .03470 (3) kana/kashira/kane I wonder 187 5.61 182 4.40 ns (4) tari/tari suru do ...and such 167 4.77 145 3.47 ns 5) mitai/yoo is like;look like 89 2.57 105 2.61 ns (6) gurai/goro/atari. about;around 89 2.61 88 2.05 ns (n kanji feels like;is like 70 2.15 41 0.99 .01252 (8) deshoo probably 9 0.25 61 1.55 .00119 (9) nado/nanka and so on 12 0.31 50 1.22 .02867 (10) kamoshirenai/kamo may 17 0.53 20 0.57 ns (in rashii it seems;I heard 21 0.64 12 0.35 ns u2)ka nanka or something 17 0.41 11 0.28 us (in ki ga suru I've got a feeling 15 0.43 5 0.L4 .03342 u4)soo seem;look like 17 0.52 1 0.01 .00075 (15) kee . . .type 12 0.32 0 0.00 ns adverbs u6)nanka like 406 11.67 93 2.48 .00001 (in kekkoo quite;fairly 105 2.92 45 1.02 .00754 u8)toriaezu/ichioo fOr now;tentatively 47 1.36 18 0.41 .00574 (nu taigai/daitai generally;about 28 0.94 29 0.64 ns (an tabun/osoraku perhaps 35 1.05 2 0.04 .00002 connective midteyuuka gor rather 1 86 E 2.44] 51 21.251.02167 a2)others 29 0.73 30 0.69 (2” total 2399 69.74 1403 34.52 .00001 *rate = mean rate per 1000 words, Group I = YF & YM, Group II = OF & OM 5)<.05: significant effect; ns = not significant 89 First, out of 21 hedges, about half show a significant effect of age. Especially, it is noteworthy that the younger speakers made much more use of several adverbial hedges than the older speakers : Most of the adverbs (16) nanka ’like’, (17) toriaezu/ichioo ’for now;tentatively’ , (18)kekkoo ’quite; fairly’ , and (20) tabun/osoraku ’ perhaps ;probably ’ were used more frequently by the younger speakers than by the older speakers. For the actual use of these adverbs, see below. (17) kekkoo ’quite; fairly’ (18) toriaezu ’for now; tentatively’ 1 YF: Korekara toriaezu, from-now—on 2 ima wa benkyoo o kekkoo chuushin ni yatette, now TP study DO center do ’From this point, in the meantime, I’ll be focusing on studying a bit, now.’ (16) nanka ’like’ (20) tabun ’probably' 1 YF: Nto, nanka yaritai koto ga tabun shumi tte well want-to-do thing SP hobby QT 2 yaritai koto da to omou n desu yo. want-to-do thing BE QT think NM BE IP ’Well, like, I think, things that you want to do is probably a hobby. . ., [a hobby is] things you want to do, y’know.’ For the two most frequently used hedges, (1)toka ’or something’ and (16) nanka ’like’ , the statistical analysis showed a highly significant difference between the younger and older groups (F(1,36)=124.25, MSE=36.63, p <.00001 for toka and 90 F(1,36)=58.24, MSE=28.99, p <.00001 for nanka). The use of (1)toka in Group I is about 3.5 times as frequent as that in Group II, and the use of (16) nanka is about 4.4 times more frequent than that in Group II. Considering that the tokens of toka and nanka together occupy about half of the total tokens of hedges in Group I, the high frequency of toka and nanka appears to be the major factor which caused the large statistical difference between the two groups. In addition to (1) toka ’or something', (16) nanka ’like’ and some adverbial hedges, some phrase— or sentence-final hedges were strongly preferred by Group I: (2) omou ’I think’ , (7) kanji ’feels like', (13) ki ga suru ’I’ve got a feeling’, and (14)soo ’seem; look like'. These types of hedges add to the preceding phrase or clause the vague nuance that the speaker's statement is temporary or that he/ she is not a hundred percent certain. The speakers in Group I seem not only to use hedges often, but also to exploit a great variety of these expressions, for example: (13) Id. ga suru ’I've got a feeling’ 1 FM: Go-nen mae wa a motto tomodachi to atteta 5-years before TP uh more friend with meat 2 asondeta ki ga suru. have-a—fun ’ It feels that I was meeting or having a fun with my friends more five years ago [than now].' (7) kanji ’feels like; is like’ 1 YF: Nanka faasutofuudo toka haitte, like fast-food etc. enter 91 2 Shabettee owatchattari toka, chat end-and-such or something 3 burabura machi shiteru tte kanji. rambling city doing QT ’Well, it’s like, we enter a fast-food store or something, only chat and stuff whole the time, or ramble a city.’ However, the uses of (8)deshoo ’probably’ (8) and (9)nado/nanka ’and so on’ in Group II were higher than in Group I (F(1,36)=13.18, MSE=2 .55, p= .00119 for deshoo andF(1,36)=5.08, MSE=2.05, p = .02867 for nado/nanka). Although (9)nado ’and so on’ sounds more formal than toka, nado has a very similar meaning and function as (1 ) toka ’or something’ : they usually follow a noun phrase and give it a blurred reference. Therefore, (9)nado may serve as an alternative linguistic choice to (1) toka particularly for Group II . Finally there was also a significant difference in the use of the connective (21)teyuuka ’or rather’ between Group I and II (F(1,36)=5.65, MSE=5.04, p = .02167). The detailed use of teyuuka is discussed in context in Chapter 5. 4 . 3 . 2 . Sex Regarding sex differences, it was hypothesized as follows. Hypothesis 2: Hedges are used more often by female speakers than by male speakers. The result (Table 4 . 4) shows that this prediction was correct with a highly significant effect (F(1,36)=31.69, MSE=301.10, p = .00003) . This means that in general women used far more hedges 92 than did men. This result is in accordance with R. Lakoff ’ s (1975) hypothesis that hedges are one of the characteristics of ’women’s language’. Other researchers (e.g., Coates 1993; Holmes 1995) have demonstrated based on their conversational data that Lakoff ’ s hypothesis is generally true. The present study shows that this observation also applies to Japanese women’s speech. Table 4.4. Tokens and mean rates of hedges by sex tokens mean rate P-value per 1000 words Female (n=20) 2248 63.053 ----—----——_---—-—-F-——_---—————-_— _-———-_----—-———_——- Male (n=20) 1554 41.212 .00003 As for individual hedges , significant differences according to sex were observed for six hedges (Table 4.5 on the next page). Since there were twelve hedges which showed a significant difference according to age (see previous section) , it is probably safe to say that the effect of sex is weaker than that of age. Out of the six hedges with a significant difference, there was only one hedge, (20)tabun/osoraku ’perhaps;probably’ , which occurred more often in men’s speech than in women’s speech. The other five hedges, which were used more frequently by women, are (1)toka ’or something’ , (2)0mou ’ I think’ , (4)tari/tari suru ’do . . . and such’ , (14)soo ’seem;look like’ , and (16)nanka ’1ike’ . The followings are the examples. 93 (20) tabun ’probably’ 1 YM: burabura shitari, kaimono, kana tabun. stroll do-and shopping I-wonder ’ I guess , I stroll around, go shopping and stuff, probably. ’ (1) toka ’or something’ and (14) $00 ’seem; look like' 1 YF: Koe ga zubutoi kara toka iwa omowarete soo, voice SP deep because say—PASS think-PASS ’It looks like, [I] am told... thought like ”This kid has a deep voice, so..."’ Table 4.5. Tokens and mean rates of all hedges by sex [hedge Emeaning Female Male p phrase/sentence-final expressions token rate* token rate u) toka or something 575 16.37 311 8.63 .00003 a) omou I think 283 7.95 186 4.95 .00343 (a kana/kashira/kanel wonder 195 5.43 174 4.57 us a) tari/tari suru do ...and such 192 5.21 120 3.03 .02893 a) mitai/yoo is like;look like 103 2.86 91 2.32 ns w) gurai/goro/atari about;around 77 2.02 100 2.63 ns 0) kanji feels like;is like 68 2.01 43 1.13 .05346 M) deshoo probably 23 0.63 47 1.17 ns w) nado/nanka and so on 24 0.65 38 0.88 ns u0)kamoshirenai/kamoney 23 0.70 14 0.40 us (in rashii it seems;I heard 13 0.38 20 0.60 ns (in ka nanka or something 10 0.27 18 0.41 ns (1” ki ga suru I've got a feeling 13 0.38 7 0.19 ns u4)soo seem;look like 14 0.41 4 0.13 .03547 (15 kee ...type 7 0.18 5 0.14 ns adverbs (1m nanka like 372 10.57 127 3.58 .00003 (in kekkoo quite;fairly 91 2.39 59 1.54 ns (1m toriaezu/ichioo for now;tentatively 30 0.79 35 0.98 ns u9)taigai/daitai generally;about 28 0.85 29 0.73 ns a0)tabun/osoraku perhaps 12 0.36 25 0.73 .04925 connective midteyuuka §or rather I 71 E 2.03 l 66 E 1.65 ns m2)others 24 0.64 35 0.81 ns (2” total 2248 63.05 1554 441.21 .00003 *rate = mean rate per 1000 words g><.05: significant effect; ns = not significant 94 As in the case of age (Table 4.3) , in Table 4.5 (1)toka ’or something’ and (16)nanka ’like’ showed highly significant effects (F(1,36)=32.65, MSE=36.63,p = .00003 for toka and F(1,36)=33.63, MSE=28.99, p = .00003) , which may have caused the large overall difference in the use of hedges by sex. This means that these hedges, toka and nanka, represent not only hedges of the contemporary Japanese youngsters (based on Table 4.3) , but also hedges of Japanese female speakers (based on Table 4.5). It is interesting that the female speakers used (4) tari/tari suru ’do . . .and such’ more often than did the male speakers (F(1,36)=5.06, MSE=18.71, p= .02893) . This expression is similar to (1)toka ’or something’ in that both express ”an inexhaustive listing of actions or states” (Makino and Tsutsui 1992 : 458) . (4) tari/tari suru is often combined with (1) toka ’or something’ , and they make unspecific references , as shown in the following example . (12) 1 YF: Kanarazu mushi hain no nee. always bug enter IP IP ’Bugs always get in, y’know. ’ 2 M00 nomoo to shita shunkan ni haittari toka, indeed try-to-drink QT did moment in enter ’The moment I try to drink something [juice], indeed, they get in and stuff.’ 3 Nondeinai noni haittari toka shitete, drink-not even-if enter and ’Even if I am not drinking, they still get in or something. ' The higher frequency in the use of (1) toka and (4)tari/tari suru in the female group may mean that female speakers tend to elaborate 95 their utterances by giving more examples than do male speakers, but more likely that they tend to avoid a definite answer by listing or implying other possibilities. 4.3.3. Style The last hypothesis on the use of hedges concerns style. Hypothesis 3: Hedges are used more often in casual speech than in formal speech. Although the previous two hypotheses were supported with highly significant effects, this hypothesis was not supported in the present study (Table 4.6). The result actually indicates the reversed effect; the participants used hedges more in interviews than in conversations between friends (referred to as ’chat’ in Table 4 . 6) . Thus , there was an influence of style of conversation on the use of hedges, but the interviews triggered more use of hedges than did chats (F(1,36)=12.17, MSE=207.28, p = .00164). Table 4.6. Tokens and mean rates of hedges by style tokens mean rate P-value per 1000 words ___9_E§1=__(_9_=_4192____ 1596 46 . 5 17 ---’————--—————q-——_———-—-——-—----- Interview (n=40) 2206 57.748 .00164 One of the reasons of the above finding may be suggested in Table 4 . 7 (next page) which presents the results for each hedge. The table contains seven hedges that showed a significant 96 di f f erence between /70ashira/kane ’I wonder’, ( 6 ) gurai / goro/ a tari the two styles; ’about;around’, and (2)omou ’ (18) I think’, (3)kana (4)tari/tari suru ’do ... and such’, toriaezu/ichioo ’ for now; tentatively ’ were used more frequently during interviews , vflnigle (12)ka nanka ’or something’ and (14)soo ’seem;look like’ appeared more often in conversations between friends. Talmle 4.7. Tokens and mean rates of all hedges by style Jhedge Emeaning Chat ** Interview p ;phrase/sentence-final expressions token rate* token rate (1) toka or something 456 13.16 430 11.84 ns (2) omou I think 161 4.77 308 8.12 .00031 (3) kana/kashira/kane I wonder 128 3.76 241 6.25 .00088 (4) tari/tari suru do . . .and such 58 1 . 65 254 6.59 .00001 (5) Initai/yoo is like;look like 93 2.80 101 2.38 ns (6) gurai/goro/atari aboutzaround 48 1 . 40 129 3 .25 .00005 (7) .kanji feels like;is like 44 1.38 67 1.76 ns (8) (deshoo probably 29 0.87 41 0.92 ns 6H .nado/nanka and so on 24 0.67 38 0.86 ns H0).kamoshirenai/kamo may 21 0.68 16 0.42 ns U1).rashii it seems;I heard 20 0.64 13 0.34 ns HZ).ka nanka or something 20 0.52 8 0.17 .02516 U3).ki ga suru I've got a feeling 5 0.15 15 0.42 .09004 HA).Soo seem;look like 16 0.50 0.04 .00452 US) kee ...type 9 0.24 0.08 ns adverbs (16) nanka like 272 7.82 227 6.33 ns (17) kekkoo quite;fairly 56 1.57 94 2.37 .06839 (1% inariaezu/ichioo for now;tentatively 10 0.29 55 1.48 .00103 (1% thaigai/daitai generally;about 12 0.38 45 1.20 .09542 (20) tabun/osoraku perhaps 2 0 0 . 63 1 7 0 . 46 ns connective <21) teyuuka ior rather I 66 1.93 71 1.76 ns (22) Supported Hypothesis 2: the female group use more hedges than the male group -> Supported Hypothesis 3: the participants use hedges more often in chatting than in interviews -> Not supported (Reversed) This section discusses the results and how the social factors (age, sex and style) interact with the main motivations for using hedges, taking into consideration some of the findings from previous research. I will also talk about the frequent use of toka ’or something’ and nanka ’like' among other hedges and the influence of these hedges on the overall result of the use of hedges. 4.4 . 1 . Explanations of the results for the three hypotheses First, the fact that Hypothesis 1 was supported is largely due to the frequent use of hedges by YF. YM did use hedges more frequently than CF and OM, but the difference was not as big as that of YF. Statistically, it was also proven that the use of hedges by YF was more frequent than that by YM in both styles (post-hoc comparisons, p < .01 for both chat and interview, see Appendix 6.3-6.4). The following are possible reasons why the Younger groups, especially YF, used hedges often. 104 ( 1) Toka ’or something’ and nanka ’ like’ are used as a habit . The high frequency in the use of toka ’or something' and nanka ’ like ’ by YF distinguishes their language from that of other groups . The frequent use of these two hedges among YF may actually be due to a habit. Philips (1998) explains that nanka ’like’ is often used by Japanese speakers to mark hesitation or uncertainty. Barke (2000) claims that toka ’or something’ and nanka ’like’ are frequent words of habit, especially so among Japanese female speakers . This issue is further discussed in Chapter 6 which deals with the speakers ' self—reported opinions about the use of hedges through questionnaires. In the questionnaire, most of YF agree that they use hedges unconsciously. (2 ) For entertainment purposes in youth language As Yonekawa (1998) states, younger women are relatively free from the social norm or model in language use. This is because in general they participate less in formal interactions and have fewer responsibilities in society. He also claims that younger Speakers seem more free to break standard social norms than the older people, and to enjoy creating or using youth language for fun or to express an image of youthfulness. For this reason I Spec‘llcifize that this tendency is generally stronger for younger women than younger men. The issue that younger females behave differently is further discussed in Chapter 6. The use of some of the hedges by the younger speakers may 105 be associated with ”speech play”. According to Kirshenblatt- Gimblett and Sherzer (1976: l) , speech play is a part of everyday interaction, and is defined as any local manipulation of elements and relations of language, creative of a specialized genre, code-variety, and/ or style: for example, jokes, riddles , nonsense, speech metaphor and word play . Brukman ( 1972) claims that the play element in speech play can reduce formality and bring humor. Kirshenblatt-Gimblett and Sherzer (1976) state that all members of society can engage in linguistic creativity and playfulness, and that speech play may be used to promote the formation of social groups and to stress their difference from other groups . Younger speakers may use some hedges as speech play and for fun when talking to other younger friends, and this may be one of their ways to promote solidarity. I will present examples of speech play by the younger groups in context in Chapter 5. (3) Positive politeness (promoting rapport) Younger females tend to pay more attention to protecting each other’s ’positive faces ' (see Chapter 2, Section 2.7. l. for explanation) by using hedges as in—group identity markers . They show solidarity by adopting similar ways of talking in casual settings, and by being vague with hedges, they can seek agreement or establish common ground for the conversation. They value this solidarity in feelings and opinions among conversation participants . This confirms the general observations that women 106 tend to be politer than men. (4) Negative politeness (indirectness) As Satake (1995) claims, some of the vague expressions in youth language can be characterized by indirectness. ’Negative politeness ’ (see Chapter 2 , Section 2 . 3 . 3 . 1 . ) may play an important role here, because speakers are probably afraid that they may sound too confident or too direct if they do not use hedges. While the previous three reasons are speaker-oriented motivations for the use of hedges, this motivation is rather addressee—oriented. (5)Se1f-protection (face-saving) As Tsuji (1999) states, ”superficial friendship" style by Japanese younger speakers might have some influence on the use of hedges (see Chapter 2, Section 2 .4.2. for details) . I interpret this ”superficial" style as the style of non-committal or self-protective interpersonal communication. The younger generation exploits hedges to guard against the possibility of a faulty memory or explanation, or to prevent revealing too much of themselves in order to avoid possible conflicts. This type of hedge is speaker—oriented, and helps to establish a mutually comfortable distance, which does not interfere with the conversation participants ' freedom of action. The psychological and social reasons for this ”superficial” style among youngsters are addressed in detail in Chapter 6. 107 UH Less experience and knowledge The younger generation in general has less experienceauxi knowledge than the older generation. Talking to the older participants, I had the impression that they tend to have more opinions and are used to expressing them more than younger speakers . The younger speakers , on the other hand, may resort to using hedges to express their uncertainty or the lack of knowledge. This explanation may be especially applicable for the comparison between younger and older speakers in interviews in which they are directly asked for their opinions. Thus , the motivation for using hedges can be multi-layered. What the primary motivation is depends often on the context in which the speaker is situated and on who the speaker, the bearer, and bystanders are. The main motivations discussed above are informational (speaker’s knowledge), positive gmfliteness (showing solidarity), negative politeness (being indirect cm‘ avoid sounding too confident) and self-protection. In information-centered interaction such as interviews, the informational motivation and self-protection may play more important roles than other motivations . In an affection-centered conversation between friends, however, positive politeness and the purpose of entertaining with youth language (speech play) may motivate the use of hedges more than other factors. With respect to Hypothesis 2 (women use hedges more than men), most of the motivations I described for Hypothesis 1 may 108 also hold. However, the reason why women use hedges often may be mainly explained by the general observation that women talk more politely than do men. For positive and/ or negative politeness, women tend to make use of vagueness in order to express their camaraderie or deference toward the addressee (see Chapter 2, Section 2.6.3.1. and 2.7.1.). Looking at the details , however, differences appear between the four groups and between the two styles. The table below summarizes the relation between the four groups in terms of statistical results (post-hoc tests). Table 4.11. Relation between the four groups in the frequency of hedges (post-hoe comparisons) chat YF > YM z OF z OM (YTfl== (NM) interview YF > YM == OF z OM (YNI>>(NH) total YF > YM z OF = OM (214 > (N4) YF=younger female, YM=younger male, OF=older female, OM=older male > means significant difference, = means no significant difference YF showed the highest frequency of hedge usage of all groups in both chats and interviews. The frequency for YM and OF is similar in both styles, whereas the frequency for YM is higher than that of OM in interviews. In chats, three groups, YM, OF and OM, behaved similarly. Coates (1993) associates women' s frequent use of hedges with her observation that they talk about more sensitive topics than 109 do men (see Chapter 2 , Section 2 . 4 . l . ) . However, the present study did not find such a trait; the male participants talked about personal or sensitive topics in their conversations as much as did the female participants. The older males in interviews even talked a lot about personal matters; their total number of words was the highest among all groups (Appendix 5.1-5.8). Lastly, Hypothesis 3 was rejected. This hypothesis was not supported either in my preliminary study (there was no difference between the two styles). There are a few possible explanations for why this hypothesis was rejected. Previous studies have associated vague language with informal conversation settings (e . g. , Lehrer 1975; Channell 1994) , and with solidarity (in—groupness), informality, and equality (Tannen 1996) . They have regarded solidarity as one of the major motivations for the use of hedges as well as other youth expressions (Okamoto 1995) . From these suggestions, we would expect hedges to appear more frequently in informal settings, that is, more in chats than in interviews. However, in the present study hedges were used more often in interviews than in chats. A question arises: Was the interview setting ”formal” enough? I contend that it was, or at least more formal than the chatting context, because there was no familiarity or solidarity between the interviewer and interviewee (total strangers with each other), and because the conversation format was that of an interview (with a rigid alternation between questions and answers) 110 carried out in the polite style (desu/masu endings). Ide (1982) also calls this type of conversation a ”formal conversation". However, had the interviewer been much older, the participants might have spoken more formally, and the difference between interviews and chats might have been more prominent. One of the main reasons for the unexpected results regarding style may be because during interviews more emphasis is put on information exchange than during chats. In interviews, the participants were required to provide information in response to several questions, whereas they may have been less pressured to exchange new information while chatting with a friend. Therefore, the use of information—oriented hedges may have been facilitated more in interviews than in chats. Evidence for a role of informational demands especially in interviews can be found in the types of hedges that caused a significant difference by style. For example, as I mentioned in 4.3.3 in this chapter, tari/tari suru ’do ... and such’ appeared frequently in interviews for listing representative:activities in one’s free time. When providing quantitative information or citing proper names, gurai, gortn and atari ’about; around"were often used for approximation. In searching for answers, the participants often used kana and kane ’I wonder' before their answers , as in doo kana ' I wonder how’ . Omou ’ I think’ was commonly employed to end their answers to interview questions. Additionally , kekkoo ’ quite; fairly ' and taigai/daitai 111 ’generally; about’ were often used for generalization in interviews, although they produced only a tendency toward a main effect of style (F(1,36)=3.44, MSE=3.68, p'= .06839 for kekkoo, F(1,36)=2.87, MSE=4.77, p = .09542 for taigai/daitai). In interviews the answers by the participants had to be as informative or precise as required for the purpose of the exchange , as described in Grice’s maxim of quantity. As a result, more information-oriented hedges such as the above were used in interviews to express approximation or generalization . Therefore , though it is true that informality or solidarity do trigger vagueness as explained in the literature, it is also probable that formality promotes vagueness in situations such as interviews, which require precision. This also holds in academic/scientific writing, where degrees and levels of vagueness/uncertainty are required for appropriate expression (Hyland 1996) . This type of communication style appears to be self-protective, and it is often speaker— and context-oriented. The above account for the quantitative findings concerning the three hypotheses will be further discussed in Chapter 6. Chapter 6 will deal with the analysis of the metalinguistic data on the use of hedges which were collected from the participants' self-reports. 112 4 . 4 . 2 . Overall results excluding the use of toka ‘ or something' and nanka ’like' Toka ’or something' and nanka ’like' were the two most frequent hedges in my data. Combined, these two hedges had 1385 tokens (out of 3802 total hedge tokens), constituting 36.43% of all hedge occurrences . In the age and sex comparisons , as discussed in the previous section, it was found that toka and nanka had an important role in distinguishing the younger from the older groups , and the female from the male groups . In the younger groups toka and nanka amounted to 45 . 47% of the distribution of all hedges (calculated by mean rates) , while in older groups they contributed only 21.55%. YF used toka and nanka the most frequently of the four groups; toka and nanka occupied 51.59% of all hedges for YF; 38.60% for YM; 24.62% for OF; 17.84% for OM. This high frequency of toka and nanka in the younger groups , suggests that the use of toka and nanka is highly habitual. I decided to investigate the results excluding these two hedges in order to see if the sociolinguistic factors still have any influence on the use of the other hedges. The figure on the next page demonstrates the differences among the four groups with the mean scores excluding toka and nanka tokens (see Figure 4.1 on p. 101 for comparison). 113 Figure 4.2. Use of hedges excluding toka 'or something’ and nanka ’like' (mean/1000 words) 100 90 80 - 7O . 60 - 50 40 30 20 10 O chatting E . . i D interView First, it is apparent in the figure that the bars for the younger groups shrank greatly compared to those in Figure 4. 1 (the result including toka and nanka) . Figure 4.2 also shows that the differences among the four groups for chats were reduced, though there is still a decreasing curve rightward for interviews. For the younger groups, the difference between the two styles became noticeably bigger than the one in Figure 4.1 for YF and YM. The actual numbers of tokens and mean rates are presented in Table 4.12 and 4.13 below. Surprisingly, in the chats, YM showed the lowest frequency of the four groups, though the difference was not significant statistically. These results were evaluated by ANOVA (Tables 4.14). Detailed ANOVA results are given in Appendix 7. 114 Table 4.12. Tokens and mean rates of hedges in chats for the four groups (n = 40) tokens mean rates IF 273 31.73 TM 182 21.29 OF 212 26.23 OM 201 22.92 total 863 25.54 Table 4.13. Tokens and mean rates of hedges in interviews for the four groups (n = 40) fi tokens mean rates IF 464 54.40 !M 387 44.71 OF 352 32.12 OM 346 27.06 total 1549 39.57 Table 4.14. social variables (ANOVA, Main effects and interactions of the three excluding toka and nanka) [main effect age sex style 1 |_p-value .00047 .01151 .00001 interactions age x style x style x style 1: sex age sex age x sex p-value n.s. .00018 n.s. n.s. p <.05: significant effect, n.s.= not significant As was the case with the results with all hedge tokens, there were main effects of all three social variables on the use of hedges excluding toka and nanka. The levels of significance were remarkably strong for age and style. Analyzing the data without toka and nanka expressions , both Hypotheses l and 2 were supported (F(1,36)=16.56, MSE=144.81, p = .00047 for age and F(1,36)=7.01, MSE=144 . 8 1 , .01151 for sex). And again Hypothesis 3 was p: 115 rejected, as the participants use more hedges in interviews than in chats (F(1,36)=50.35, MSE=78.19, p <.00001). A difference between these results and the previous results with toka and nanka (see Table 4.10 for comparison, on p. 101) is found in the interaction score for age and sex; the interaction turned out to be 'not significant ' in Table 4 . 14 . This means that the use of hedges by YF was less salient without toka and nanka. However, the interaction between style and age became clearer as the tendency changed to a significant effect. This implies that the use of hedges (excluding toka and nanka) by the younger groups in interviews was more frequent as compared to chats, and as compared to the older groups in either style. Thus, it was shown that even excluding the tokens of toka and nanka the younger speakers used hedges more often than the older speakers , and that women used them more frequently than men. The high frequency of the use of toka and nanka occupies an important position in the use of hedges, especially for YF. YF spoke with many hedges, and their utterances became even more uncertain, indirect, or unspecified with the frequent use of toka and nanka. 4.5. Summary The present chapter dealt with the quantitative analysis of the use of hedges, and showed that the use of hedges does differ according to the speaker's age and sex (Hypotheses 1 and 2). 116 However, the third prediction turned out to be wrong. Hedges did not appear more often in chatting with friends than in interviews with a stranger (Hypothesis 3) . I speculate that this was due to the greater informational demands in interviews, whirfl1promoted the need to use hedges to generalize, approximate, make inexhaustive lists, etc. (e.g. , gurai/goro/atari 'about; around’ and tari/tari suru ’do...and such'). The younger speakers, especially YF, showed the most frequent use of hedges in both styles. I discussed several possible explanations for this result: positive politeness (promoting rapport and solidarity); negative politeness (indirectness ) ; self-protection; and a relative lack of knowledge and experience. The quantitative results presented in this chapter will be discussed further in relation to the metalinguistic data (questionnaire results) on the use of hedges in Chapter 6. Toka ’or something' and. nanka ’like' were the :most frequently employed hedges, especially for YF. These two hedges appeared to be used as a habit. The data were reexamined excluding toka and nanka. Interestingly, this did not change the overall outcomes of the three hypotheses. In the next chapter, the use of hedges will be analyzed qualitatively in context by closely examining the most frequent hedge toka ’or something' and the connective teyuuka ’or rather’. 117 Chapter 5 Toka ('or something') and teyuuka (’or rather’) in spoken discourse 5.1. Introduction This chapter qualitatively examines the use of toka ’or something' and teyuuka ’or rather’. The usage of both toka and teyuuka are analyzed in context, and the linguistic environments , basic meanings and discoursal functions are investigated for further understanding of why and how speakers use these hedges in particular. Toka and teyuuka were chosen for analysis in the present chapter because they are two of the most important and interesting hedges in Japanese conversation . First, toka is important because it is the most frequently used hedge in my data and the difference between younger and older speakers is largest with toka. Teyuuka, too, showed a significant difference between the two groups. Though its frequency was not as high as that of toka, teyuuka is one of the interesting hedges because it is a relatively new word, and there are many variations in how it is employed. 5.2. The use of toka (’or something') In the previous chapter, it was found that the frequency of toka was highly influenced by age and sex. I repeat the quantitative result of toka from Chapter 4 in Table 5.1. 118 Table 5.1. Tokens and mean rates of toka ’or something’ (n=40) Chat interview YM YM OF OM YM YM OF OM total words 8478 8735 8238 8546 8766 8592 11083 12886 tokatokensl 244 138 43 31 218 87 70 55 toka rate 28.36 15.81 5.01 3.46 25.13 10.87 6.96 4.40 YF=younger female, YM=younger male, OF=older female, OM=older male The younger speakers, especially female speakers, employed toka frequently; Why do they use toka so often? In what follows, I further investigate the meanings and functions in discourse in order to find outwwhy toka is such.a useful linguistic means for younger speakers. 5.2 . 1. General characteristics of toka in previous studies According to Kojien (1998 : 1901) , one of the major Japanese dictionaries, there are three types of toka as shown in (1)-(3)7 . (1) listing and coordinating items e.g. ame toka yuki (toka) rain snow ’rain, snow, and others' (2) listing one example by omitting the rest, or referring to one item without specifying it e.g. koohii toka nonda coffee drank ’I drank coffee or something' (3) quoting with uncertainty e.g. kekkon shita toka marriage did ’I heard he/she got married, or something.’ This dictionary is written in Japanese. The examples are theirs . The glosses and translations are mine. 119 The first and second types of toka in (1) and (2) consist of the coordinative particles to ’and' and ka ’or' , while the third type is composed of to (quotative particle) and ka (question particle) (Kojien, p. 1901). The third type of toka in (3) often precedes a verb such as iu ’say’ or kiku ’hear' . Toka can be combined with a variety of grammatical categories such as nouns, adjectives, and verbs. Makino and Tsutsui (1992:488) give a general explanation, according to which toka is a conjunction that lists two or more items, actions, or states inexhaustively. This explanation is similar to the first type of toka in (1) in Kojien's description, which is considered as the canonical usage. The usage of toka in (2) is regarded as the more recent type (Kojien, p. 1901). English phrases such as or something, and stuff like that, etc. are similar to toka ’or something' in Japanese. As I explained in Chapter 2, these phrases are often referred to as ”tags" (e.g. Dines 1980; Ball and Ariel 1978) or as ”vague category identifiers" (Channell 1994) . Tags often refer vaguely to quantities or category. One of the important social functions of tags such as toka is to mitigate the force of speech by vaguely or indirectly telling the speaker's ideas, feelings and information (Channell 1994; Tsuj i 1999) . Channell states that the speaker is seeking to avoid ”face threatening" (in the sense of Brown and Levinson 1987). Tsuji, based on his questionnaire results, reports that toka and 1.20 teyuuka ’or rather’ are used by the Japanese young speakers to avoid conflicts in their interpersonal relationships. Sunakawa (2000) annualizes conversation data (100 minutes) which were collected from 13 native speakers of Japanese (12 women and 1 man, age unknown). Among three categories of toka similar to Kojien’s (1998: 1901), she focuses on the third type of toka, a quotation marker, and shows that toka often.precedes the verb iu (or itte) 'say' (86.7%) . She found that tokaitte, meaning lsay something like', coordinates the successive quotation of utterances by two or more speakers. Therefore she claims that tokaitte ’say something like’ has functions not only of expressing vagueness, but also of coordinating quotations. Her study shows that the functions of toka seems to have diversified. 5.2.2. Immediate linguistic contexts for toka I begin my inquiry of toka by showing in which context it appears within the utterance in order to provide a general idea of the toka usage. Tokas were used in the linguistic environments . 8 shown in the examples (4) to (9) below . The examples are from the actual conversation data that I used in the quantitative analysis in Chapter 4. Additional excerpts with toka in longer contexts are shown in the next section (5.2.4). NP: noun phrase; S: sentence; VP: verb phrase; (): optional 121 (4) [NP/Sh toka [NP/Sh (toka) ’ [NP/S]1 , [NP/S]2andothers’ e.g. Rooma toka Mirano (toka) Rome Milan ’Rome, Milan, and places like that’ (5) [NP/S] toka sooyuu no/mono ’things like [NP/S]’ e.g. okane toka sooyuu no money such one ’things like money’ (6) [NP]1 toka iu [NP]2 ’[NP]2 called [NP]1or something’ e.g. Tooyoo toka iu hoteru Tooyoo say hotel ’a hotel called Toyoo or something’ (7) [NP]toka [VP] e.g. ryokoo toka iku trip go ’(I) go on a trip or something’ (8) [S/(in)direct quotation] toka [VP] e.g. jikan ga oshii toka omou time SP precious think ’I think, like, time is precious’. (9) [NP/S] toka. e.g. naraigoto Hm) hajimerareta shi, toka. lessons also start-could and ’And I was also able to start lessons, and stuff’. The use of toka in (4) through (6) is a canonical usage for both the younger and older groups. The type in (7) is a recent usage (Kojien, p. 1901) , which is often adopted by younger speakers. Toka in (4) is used to list items inexhaustively (the second toka sometimes drops). In contrast, toka in (5) through (8) lists 122 only one representative example , and ”cue the listener to interpret the preceding element as an illustrative example of some more general case” (Dine 1980: 22). Toka in (6) is a modified version of A to iu B ( ’A called B’), where ka (question particle) is added to to (quotative particle) . All these types of toka usage express a degree of vagueness, uncertainty, or unspecification, but toka in (6) in particular indicates the speaker' 5 lack of precise memory and confidence. Toka in (9) is used utterance—finally, sometimes precedes a pause, and signals that the speaker is yielding the turn to the hearer. This toka semantically serves to blur the content of the statement preceding it, and socially mitigates the force of the utterance. 5.2 .3. Basic functions and distribution of toka I first examine the distribution of toka based on the three semantic functions, inexhaustive coordination, vague reference, and vague quotation (Kojien, p. 1901, see p. 118 in the present study), for four groups (Tables 5.2 to 5.5) and for two styles. Let us compare Table 5.2. (YF and YM) and Table 5.3. (OF and OM), which show the results in chats. First, we notice a big difference in the total toka tokens between the two groups in the chatting situation. The younger speakers ’ tokens (382 tokens) is 5.2 times as high as that of the older speakers (74 tokens). 123 Table 5.2. Tokens of toka ’or something’ by functions in chats (younger groups) functions YF YM total (1) inexhaustive 37 (15.2%) 25 (18.1%) 62 (16.7%) coordination (2)vague 80 (32.8%) 69 (50.0%) 149 (41.4%) reference (3)vague 127 (52.0%) 44 (31.9%) 171 (42.0%) quotation total 244 (100%) 138 (100%) 382 (100%) (YF=younger female, YM=younger male, n=40) Table 5.3. Tokens of toka ’or something’ by functions in chats (older groups) functions OF OM total (1) inexhaustive 20 (46.5%) 17 (54.8%) 37 (50.7%) coordination (2)vague 12 (27.9%) 9 (29.0%) 21 (28.5%) reference (3)vague 11 (25.6%) 5 (16.1%) 16 (20.9%) quotation total 43 (100%) 31 (100%) 74 (100%) (OF=older female, OM=older male, n=40) Second, a remarkable difference is found in the use of toka as ”vague quotation” in (3) . The younger groups ’ total tokens (171 tokens; 42.0%) is 10.7 times more than that of the older groups (16 tokens; 20.9%) . Especially, YF group often makes use of this type of toka (127 tokens; 52.0%), which is 11.5 times more than that of OF and 25.4 times more than OM. For YF, the use of toka as ”vague quotation" makes up more than half of the total toka tokens. This suggests that this type of usageeis common for the younger'igeneration, in. particular' YF speakers in. chatting situations. 124 The younger speakers also employ toka often for ”vague reference” as in (2) in Table 5.2. Their usage of this type (149 tokens; 41 .4%) is 7 . 1 times as frequent as that by the older groups (21 tokens; 28 .5%) . Especially YF’s tokens (80 tokens; 32 .8%) are the highest among the four groups. On the other hand, for the older groups (Table 5.3), the use of toka as ”inexhaustive coordination" in (1) seems common. It forms about half of their total toka tokens. This shows that the older groups still use toka in the canonical.way, while the younger groups tend to employ it less so. This is explained in detail with examples in the next section. Tables 5.4 and 5.5 on the following page show the results from interviews. It is interesting that in YF’s speech, tokens of toka as ”vague quotation” in (3) drastically drop from 127 tokens (52.0%; in chatting, Table 5.2) to 26 tokens (11.9%; in interview, Table 5.4). The tokens in YM’s speech also drop from 44 tokens (31. 9%) to 5 tokens (5 .7%) . This suggests that the younger groups consider this type of usage (”vague quotation”) inappropriate for the interview setting, and regand it as a casual linguistic expression among in—group members as commonly assumed (Okamoto 1995) . For them, as Okamoto states, this may be one way to express youthfulness or solidarity. In contrast, the older groups, eSpecially male speakers, rarely use toka for vaguerquotation, regardless of the setting. Table 5.4. Tokens of toka ’or something’ by functions in interview (younéger groups) functions YF YM total (1) inexhaustive 93 (42.7%) 27 (31.0%) 120 (36.9%) coordination (2)vague 99 (45.4%) 55 (63.2%) 154 (54.3%) reference (3) vague 26 (11.9%) 5 (5.7%) 31 (8.8%) quotation total 218 (100%) 87 (100%) 305 (100%) (YF=younger female, YM=younger male, n=40) Table 5.5. Tokens of toka ’or something’ by functions in interview (older groups) functions OF OM total (1) inexhaustive 53 (75.7%) 41 (74.5%) 94 (75.1%) coordination (2)vague 10 (14.3%) 13 (23.6%) 23 (19.0%) reference (3) vague 7 (10.0%) 1 (1.8%) 8 (5.9%) quotation total 70 (100%) 55 (100%) 125 (100%) (OF=older female, OM=older male, n=40) In contrast to this decrease of toka as vague quotation, in YF’s speech, there is an increase in the percentage of toka as ”inexhaustive coordination”. Similarly, for the older speakers, the tokens for ”inexhaustive coordination” increase in interview (Table 5.5) , and occupy 74.5 % of all the toka usage. In sum, one of the major differences between the two age groups is that the younger speakers’ usage is more diverse than that of the older. Especially, the younger groups often express vague quotation with toka, but mainly in chats. Older speakers use toka to coordinate items the most. 126 5.2.4. Analysis of the use of toka This section examines the use of toka in context. The excerpts from the actual conversation data are used to illustrate the discoursal functions of toka. I present some cases of toka used for vague quotations found in conversations among younger speakers. 5.2.4.1. Toka for inexhaustive coordination Using toka to coordinate items (toka coordination) is the most common usage among the older speakers. In interviews, the younger speakers also use this toka coordination often. Consider example (10) . (10) [Prior to this excerpt, 0 (older female interviewee) mentioned that she was planning to travel in Italy. Interviewer (henceforth, abbreviated as ”I” ) asked 0 where in Italy she was going9.] 1 I:Dochira desu ka, Itaria no. which BE Q Italy GP ’Where is that, in Italy?’ —>2 O:Itaria no Rooma toka Mirano toka Firentse toka, Italy GP Rome Milan Firenze ’In Italy, (I will go to) Rome, Milan, Florence, among others. —23 Ponpei t oka atchi no hoo desu . Pompeii there GP direction BE ’Pompeii and such, in the direction of those places.’ 9 See Appendix 1 for the transcription methods. 127 Here, 0 is listing cities in Italy (Rome, Milan, Florence and Pompeii) using toka. She will probably go to other places as well, but presents these places as representative examples. In such toka coordination, it is typical to combine nouns conjunctively to illustrate. This type of toka is very convenient because the speaker does not have to mention all items, yet the listed items give a general idea to the addressee. In the younger speakers’ utterances, toka coordination enjoys great flexibility, occurring not only after noun phrases, but also after verbal, adverbial, adjectival, and prepositional phrases . Toka sometimes appears to combine long utterances . The excerpt in (11) shows how sentences are combined with toka. (11) [K (YM, interviewee) is explaining to the interviewer how he manages his time in order not to get exhausted.] 1 K: Sukoshi-zutsu yoka tteyuuka yasumi no jikan 0 little by little free time or say rest GP time DO 2 doo ireru ka kangaete. how get Q think ’I think of how I can take free time, or rather, a break, here and there.’ 3 Tatoeba ma ima wa benkyoo o mein de yatteru kara, for-example well now TP study DO mainly doing because ’For example, well now, I am focusing on studying, so,’ —>4 Benkyoo shite ichijikan yattara donokurai yasumu toka, study do for-l-hour do-if how—much rest ’How much I can rest after studying for an hour, or’ 5 Ato .ie kaette shokugo ni donokurai yasumu to also home return after-meal in how-much rest if 128 —>6 choodo ii ka toka, just right Q ’Also, when I go home, how much rest is appropriate, or so, ' 7 Ato nichiyoobi wa donokurai yasunde, also Sunday TP how—much rest ——>8 Osoku made netereba ii ka toka. late till sleep-if good Q ’Also, on Sundays , how much I can rest, and how late I should get up, and stuff like that.’ 9 Sukoshi-zutsu jikan toru yooni shiteru n desu kedo ne. little—by—little time take to doing NM BE though IP ’Thing is, I am trying to have some time little by little, though.’ As K’s tatoeba ’for example’ in line 3 shows, he enumerates his examples about how he tries to find time to rest by adding toka at the end of each explanation in lines 4, 6 and 8. In such coordination, the range of toka is not limited to one sentence, but it extends across sentences. The toka in lines 4, 6 and 8 conjunctively coordinates his explanations for the statements in lines 1 to 2 and 9, sukoshi-zutsu jikan toru yooni shiteru ( ’trying to find time little by little’). This type of toka not only coordinates examples inexhaustively as in (10), but also contributes more globally to the text coherence. K’s utterances in lines 3 to 8 (support 1 to 3 below; framework according to Schiffrin 1987) are grouped as correlated idea units , which jointly provide further explanation concerning his position in lines 1 to 2 and lead to the conclusion statement in line 9. 129 Position: ’finding time to rest' (lines 1—2) ’for example’ Support 1 toka (lines 3—4) Support 2 toka (lines 5—6) Support 3 toka (lines 7—8) Conclusion: ’trying to find time’ (line 9) The above functions may also be carried out by to ’and’ (a conjunction particle). However, the difference between to and toka is that the latter has an additional unspecifying effect. Compared with to (’and’) coordination, which implies no other possibilities than coordinated items, toka coordination implies other possibilities, and thus allows the speaker to moderate his commitment to the utterance. The above characteristics show that toka has an important role in the ideational structure (coordinating idea units) in Schiffrin’s discourse model (1987)10 . Toka also has a pragmatic effect which plays a role in the exchange structure as a marker of speaker-continuation. When toka is used in the end of each unit of talk as in (11), it indicates that an upcoming utterance is still a part of a not yet completed list: it conveys that the speaker 11 has more to say . 10 Summary of the analysis of toka using Schiffrin’s discourse model is given in the later section (5.2.5). 11 When toka is used utterance finally with a falling intonation, it suggests a turn transitional point (the speaker yields his/her turn to the addressee). 130 Toka, thus, has roles in coordinating ideas inexhaustively and in serving as a speakers’ continuation. markeru The coordination often ranges over utterances in younger speakers’ speech. In general, when toka is used for coordination, it functions to enumerate representative examples to provide a general picture of the topic to the addressee. 5.2.4.2. Toka for vague reference Toka can also be used to refer only to one item instead of listing several items . In this case, toka marks that the selected item is just one among many possible representations of what the speaker has in.mind. Toka also conveys a sense of uncertainty, unassertiveness or vagueness because it indicates a possible difference between the speaker’s statement and the reality. I present three cases of toka as vague reference: (12) uncertainty, (13) approximation, and (14) to (16) soft highlighting. 5.2.4.2.1. Toka for uncertainty The example in (12) illustrates toka used to express uncertainty. (12) [K andH (YFs) are discussing where their male friend is living. H thinks he is living in Okudo, but K thinks that he is living in Shinkoiwa though she is not sure.] 1 K: Wakannai kedo ie choo dekkai n desho71, know—not but house damn big NM TAG ’I don’t know, but his house is damn big, right?’ 131 2 HzNande shitten no doko ni Okudo7l, why know Q where at Okudo ’How come you know, where, in Okudo?’ 3 K:Okudo janai yo nanka Okudo BE-not IP like ’Not Okudo, like’ 4 H:= Okudo ja e7! Okudo BE ( -not) huh ’Not Okudo, really?’ —>5 K:= Shinkoiwa toka janai no7l, Shinkoiwa TAG Q ’Isn’t it Shinkoiwa or something?’ 6 H:A 500 na no, Ah so BE IP ’Ah, is that so.’ 7 K: Wakannai kedo know—not though ’I don’t know, though.’ In K’s question in line 5, she guesses that their friend lives in Shinkoiwa. The tag janai no ’isn’t it the case?’ in this question suggests that the utterance is a request for confirmation. At the same time K is correctingH ’ s assumption that their friend lives in Okudo. The uncertain tone of voice in this tag question is due to the use of toka. Without toka, K will sound certain, and her utterance is more likely to be perceived as a repair to H’s assumption than as a request for confirmation. K's uncertainty is also explicitly expressed in line 7 with wakannai kedo (’I don’t know, though’). This type of toka has a primary function in the information state in Schiffrin’s discourse model (1987) because it concerns the state of speaker's knowledge. 132 5 . 2 . 4 . 2 . 2 . Toka for approximation The example in (13) shows the use of toka for approximation. (13) S (YF) is explaining how much time students spend to prepare for an entrance exam. 1 S: Ima demo, now even ’Even now,’ ->2 8-jikan toka yatteru ko wa yatteru mitai desu. 8-hours doing person TP doing seem BE ’Like some people are studying eight hours or so.’ Toka is used after numeral.expression 8-jikan (’8 hours’). This toka is similar to like in English as in ’I’m like six feet tall’ . Schourup (1985) claims that this type of like is equivalent to 'approximately’ . He explains that like ”can be seen as a device available to speakers to provide for a loose fit between their chosen words and the conceptual material their words are meant to reflect” (p. 42). In (13) , S is acknowledging that the number she is giving is an approximation, and that longer or shorter cases are possible. By indicating that other cases are possible, the speaker shows that he/she is aware of the lack of precision, and so reduces the risk of future objections. This type of hedge, then, is used for self-protection. Toka here has a role in the participation framework in Schiffrin’s model (1987) since it concerns speaker/ utterance relations . The speaker expresses his/ her non—committal attitude to the utterance. 133 5.2.4.2.3. Toka for soft highlighting The next type of toka is used as a soft highlighter. Consider (14a). (14a) [S (YM) is telling his opinion about Japanese‘workaholism in interview.] 1 S:Tashikani soo, certainly so ’Certainty right,’ ~22 Sarariiman toka wa hatarakisugi da to omoimasu ne. office-worker TP ‘work-too-much.BE QT ‘think IP ’I think that white collar workers work too hard.’ Sarariiman ’white collar worker’ is the topic of the sentence in line 2, and it is marked by toka, followed by the topic marker wa. Although the noun sarariiman already has a certain degree of . 12. . . . . salience in this utterance due to the topicalization by wa, toka provides an effect of soft highlighting, as I explain in the following. My analysis of toka as a soft highlighter is based on the analysis of like as ”highlighting" device in Miller and Weinert (1995). The function of toka seems similar to that of like in casual speech in English. Miller and Weinert claim that like is 1 . . . . . 2 The salience can be expressed by pitch, pOSition in the constituents, the use of a special syntactic structure or of a particle (Sperber and Wilson 1986: 203). Speakers have various reasons for making a certain information more salient. They may, for example, want to introduce important items into the discourse, or direct the listener’s attention to particular propositions. 134 a “non-contrastive focuser that can focus on new or given information or entities" which can appear in various positions (P. 379) . Underhill (1988) also states that like often marks the most important new information in the utterance. What makes toka different from other highlighters is its softening' effect. Toka Ihighlights the information in an unassertive way. This softening and/ or unassertiveness come from the lexical meaning of toka: indication of uncertainty or other possibilities. The role of toka in this sense is close to one of the functions that tte ’that ’ (a quotative particle in Japanese) has. Tte, used after a noun, has a function of marking a topic (e. g. Miura 1974; Suzuki 1998) . Topic marking by tte is , according to Miura, done in a less direct and less abrupt way. .As stated in Suzuki, tte-marked topical phrases represent information of which the speaker is psychologically distanced. This is because tte in its original usage often expresses other voices in hearsay. In Kamio's (1990, 1994) term, the tteemarked and toka-marked information do not completely belong to the speaker’s . . 13 information . According to Kamio (1990, 1994), the distinction between linguistic forms corresponds to the boundary of territories of information. For example, indirect forms such as rashii ’seem; I heard' represent information that is outside the speaker’s territory. Whether or not a piece of information belongs to the speaker’s territory decides the choice of expressions. 135 The excerpt in (14b) shows S’s concluding utterance on Japanese workaholism addressed in (14a). (14b)1 S: Ma nihon no sarariiman wa well Japan GP white-collar—worker TP 2 hatarakisugi janai ka to omoimasu ne. work-too—much TAG Q QT think IP ’Well, I think that the Japanese white collar workers maybe work too hard. ’ Here, S makes a very similar statement to that in (14a), but now without using toka. According to the statement in (14b), the speaker S is specifying sarariiman ’white collar worker’ as a workaholic group, not implying other possibilities. This suggests that toka in (14a) was not used to imply other possible professions. Toka likely serves as a soft highlighter of certain information, while still maintaining the effect of vagueness or unassertive tone of voice that toka expresses lexically. In (14a), toka was followed by the topic marker wa. However, (15) displays a case in which toka has a soft highlighting effect without the topic marker wa and subject marker ga. (15) [S (YF) is answering the interviewer’s question about the club activities at high school. She thinks that the club was fun and she felt free from restriction because there were no seniors . ] 1 I:Bukatsudoo wa doo deshita ka, kookoo jidai wa. club-activity TP how BE Q high-school days TP ’How were the club activities in high-school days?’ 136 —+2 SzAno senpai toka inakatta n desu ne, well senior exist—not NM BE IP ’Well, thing is, there weren’t seniors and such, ’ 3 Ano hai well yes , ’Well, yes,’ 4 Uchi no gakunen kara no bukatsu tteyuuka, our GP school-year form NM club or say ’The club consists only of students from.our school year, or I should say...’ 5 M00 minna sotsugyoo shichatte, already everybody graduation have-done ”Everybody already graduated,’ In line 2, S introduces the word senpai ’seniors’ to explain why club activities were fun. In line 2, senpai ’seniors’ seems to be the main exemplar of the category, or probably the only possibility because S is providing further information about senpai in the following utterances. Toka is attached to senpai, and makes it salient in the utterance without using any particle. Admittedly, the first word of an utterance usually has a certain degree of salience (Sperber and Wilson 1986), but salience is provided in a soft or unassertive way by the use of toka. As Miura stated concerning tte, I contend that toka in this type of context . . . 14 serves to mark the topic in a less direct or abrupt way . 14 According to my impression, toka marks the information in a softer or more indirect way than does tte. The comparison of toka and tte is an interesting issue to investigate. I will leave it for future study. 137 5.2.4.3. Toka for vague quotation This section concerns mainly the unconventional use of toka as it appears frequently in the speech of younger speakers. Many of these types of quotation with toka are different from what is regularly regarded as standard usage in Japanese dictionaries. I have observed that toka in vague quotation has some similarities with like in casual English. Schourup (1985:44) found that like is used to introduce direct discourse, and that this nonstandard like is very common among younger speakers. Let us first look at the distribution of quotative toka according to the verb that follows it. Table 5.6 summaries all the data from interview and chatting. Table 5.6. Variations of quotative toka (n=40) IF IM OF OM total toka iu/itte 80 30 5 2 117 toka omou 40 3 3 0 46 tokatte 24 10 2 1 37 toka 5 4 3 1 13 toka[ other verb] 4 2 5 2 13 total 153 49 18 6 226 YF-younger female, YM=younger male, OF=older female, OM=older male Toka quotation often includes iu ’say’ (or other forms such as itte ’say and’ and iwareru ’be told' ) . Omou ’think’ is also sometimes used to mark the content of the speaker ’ s thought . There are a few cases that include other verbs such as kanjiru ’feel' , kaitearu ’ be written', naru ’become’, and kangaeru ’consider’. Sometimes the verb is absent in the main clause. 138 In what follows, I present examples of toka marking direct and indirect quotations and quoting the speaker’ s hearsay (another voice) or the speaker’s own speech (self-quotation). 5 . 2 . 4 . 3 . 1 . Toka for direct and indirect quotation The following excerpt shows direct quotations of what the speaker was told (16-3,5) and an indirect quotation of the speaker’s thought (16—6). (16) [N (YF) is explaining about her part-time job at Kentucky Fried Chicken (KFC) . She was asked by her boss to move to a new KFC with him to help him.] 1 NzIe no chikaku de yatteta, (I: un) house GP nearby at doing (I: uh huh) 2 tokoro no tenchoo ga kondo sotchi ni place GP store—manager SP this—time there —23 idoo ni natte de isshoni konai toka iwarete change to became so together come—not said-PASS ’When the store manager of the place where I was doing (a part—time job) near my house was transferred to another place, I was asked like ”why don’t you come with me?”’ 4 I:= A hikinuki tte yatsu desu ne oh hiring—away QT thing BE IP ’Oh, hiring you away from the original place, right?’ "25 N:= Jikyuu 100en agete ageru kara: toka iwareta kara: , hourly-wage ¥100 raise give because said-PASS so ’I was told like, ”For you I’ll raise your hourly wage by I15100”, so. . . ’ —>6 Dattara sotchi no hoo ga otoku kana toka omotte: , if-so that GM side SP profitable I-wonder thought (I: 800 desu ne) (I: so BE IP) (I: That’s right.) 139 7 ita n desu kedo. BE NM BE though ’Then, I was thinking, that offer might be more profitable, or something.’ In the direct quotations in lines 3 and 5, toka serves as a marker of the quotation of another voice (speech fragments produced by another speaker). In line 3, N is quoting an utterance made by her boss in KFC (issho ni konai ’Why don’t you come with me?’ ) , and in line 5, she is further quoting his persuading utterance (Jikyuu lOO-en agete ageru kara: ’I ’ll raise your hourly wage by ¥100, so...’) to explain why she changed stores. To quote what she was told, N could have used the quotation particle tte ’that’ . However, she used toka instead of tte ’that ’ to indicate that the quotation is not an exact replica of words uttered by her boss but an approximation. Thus, N avoids the possibility of misrepresenting her boss’s utterance. The function of this type of toka is attributed to the participation framework in the discourse model by Schiffrin (1987): it shows a non-committal and defensive attitude toward the utterance. As Mayes (1990) and Tannen (1989) state, direct quotation is often not the exact repetition of what was said, but is the constructed dialogue by the speaker. Toka serves to make this point clear. Mayes also claims that direct quotation serves to dramatize and highlight key elements in a narrative. Quotative toka can quote not only the content of ’speech’ , but also the content of ’thought’ as shown in line 6 in (16) . Toka 140 omotte means ’ I thought something like ’ . With toka, N is reporting her own thought imprecisely. Uncertainty is also expressed by the prior phrase kana ’I wonder’ , but toka makes the utterance even more vague. Toka indicates that the quoted thought is only a representation of what she thought, and the approximation signals that there may be a better way to put it into words. 5.2 .4.3 .2. Toka for quoting other’ 8 voice or speaker’ s own voice The excerpt in (17) presents T’s short narrative, in which a previous dialogue between T and M is repeated for N (the hearer) . Tokatte (the contracted form of toka itte ’say something like’) is frequently employed in T’s speech. (17) [T (YF) is explaining to her friend N (YF) about the bicycle that T rented to another friend M (YF).] 1 N: Soredefl then? ’Then?’ 2 T:De nanka tsuita totan ni denwa kite: M ga. then like arrived immediately phone came M SP ’Then as soon as I arrived, the phone rang, M (called). ’ -+3 Ee ima nani shiten no: tokatte, ah now what doing Q ’She goes like, ”Ah, what are you doing now?”.’ 4 Korekara todokeni iku kara tte, from-now bring-back go so QT ’ And she says ”I ’m gonna bring the bicycle back to you now. ” ’ ~25 Oi tokatte moo owatta tokatte, hey already finished 'I 90 like, ”Hey”, and I go like, ”I don’t need it now."’ 141 —>6 Moo kichatta yo: kotchi tokatte, already have-come IP this—way ’ She goes like, "I am already on the way to your place.” —>7 Ee gomen ne: tokatte, oh sorry IP ’And she goes like, ”Oh sorry."’ Tokatte coordinates direct quotations of utterances by M and by the speaker herself (T). What M said to T prior to this conversation is marked by tokatte in lines 3, 6 and 7, whereas T’s direct self-quotations are in line 5. This quotative tokatte seems similar to the narrative go in English. Expressions such as ’she goes like’ or ’I go like’ fit the Japanese translation of tokatte. Narrative ’go’ usually appears in casual discourse, which is the case with tokatte. The roles of tokatte in (17) are to report what the speaker heard or said, and to coordinate the utterances in a ’ loose ’ manner. It is ’loose’ in the sense that the utterances indicate a possible slight difference between the quoted expression and what was actually heard or said. The speaker reveals his/her inability or reluctance to formulate a strictly precise version of what was heard or said. In (17) , the agent (subject) of the quoted utterances are 5 not expressed1 . In (17), however, even though there are two 15 . In Japanese, elements often drop from the utterance if they can be inferred from the context, especially in conversation. 142 voices involved in the quoted dialogue, these two voices are not distinguished from each other in expressing the subjects of tokatte, yet this does not cause confusion. Tokatte (without the subject) minimally and rhythmically presents the reconstructed interaction. What is of interest in this conversation is the self— quotation by speaker T. As Maynard (1996) states, self—quotation has an objectifying and distancing effect. T is taking on the role of reporter, and reports her own utterance as if it were made by somebody else. By introducing the voice of ’another’ , she creates a dramatizing effect. At the same time, this dramatization is delivered with a hedging attitude. In a non- committal way, tokatte tells the listener that what he/she is listening to is not exact recollection. 5 . 2 . 4 . 3 . 3 . Toka for self-quotation to avoid awkwardness or seriousness The excerpt in (18) on the next page shows similar examples of self—quotations . However, they are not quotations of previous utterances as in (17) , but quotations of on-going utterances by two younger female friends. (18) [M and K (YF) are talking about where they can go for their graduation trip together with other friends. However, their friends are not so enthusiastic about the plan.] 143 (18) —>1 KzMinna ikanai toka iidashita deshoo7l, (HHH) everybody go—not started—say TAG 'Everybody started to say like they didn't want to go.’ 2 Yabai yo bad IP ’That’s bad!’ ”23 M:=Ikenai jan toka iu can’t-go TAG say ’(I) go like ”We can’t go, then?”.’ '24 K:=Ikenai toka i u can’t-go say ’(I) go like ”we can’t go."’ —>5 M:=Okane nai kara toka itte, money have-not because say ’(I) go like ”Because we have no money."’ 6 K: Uchira de iku?! Osaka we by go Osaka ’Shall we go to Osaka by ourselves?’ 7 M:=Uwa: tte wow QT ’saying, ”Wow".’ 8 K:Uwa:, Osaka: tsutte7l, wow Osaka! say-QT ’Saying, ”Wow, Osaka"?’ —>9 M:Osaka yada yo: toka itte, Osaka dislike IP say ’I go like, ”I don’t want to go to Osaka."’ 10 K:Itchau yo moo. go IP well ’Well, I’ll go (by myself).’ Let us first examine toka iu ’go/say like’ in lines 3 and 4. K and M are quoting their own utterances and talking as if they are reporting the speech of others. This usage is unique to YFs, and 144 is. interesting' because this quotation. marker is seemingly redundant: there is no need to self-quote on-going utterances. These instances of toka iu ’go/say like’ seem to bring about playfulness (speech play) in the conversation. Maynard (1996:222) explains that self-quotation ”is used at awkward moments for the purpose of ’lightening up’", or that the speaker plays foolish to avoid being taken too seriously. In lines 3 and 4, K and M are ”lightening up" the disappointing feeling about cancellation of the trip by being playful with words. This playfulness is also expressed in the repetition style (repeating each other’s utterance) in these lines. The effect of ”lightening up” is also brought.about by thelnitigating usage of toka. Toka implies that the quoted utterance ikenai ( ’ (we) can't go’) is not exactly what the speaker wants to say, which results in an undetermined tone of voice. With toka, the speaker can avoid facing or taking the disappointing fact all too seriously. Toka itte ’go/say like’ in lines 5 and 9 is also used to mitigate the utterance, especially when difficult things have to be expressed, The utterances okane nai (’we have no money’) in line 5 and Osaka yada (’I don’t want to go to Osaka') in line 9 are potentially face-threatening‘to‘K (the addressee) because K wants to go to Osaka. Toka is used to reduce this risk by expressing non—assertiveness and by avoiding being direct. Consequently toka is, as Maynard says, used to mend the awkward situation. The following excerpt in (19) presents a similar case. 145 (19) [H and K (YF) are talking about H’s sister. K is telling H that K likes H’s sister.] 1 H:Ano ko |wa| benkyoo shinai ne, that kid TP study do-no IP ’That kid does not study,’ 2 K: |Dotchi7l| which ’Which one?’ 3 Hchhiban shita, the jyoungest ’The youngest one.’ 4 K:Suki. like ’I like her.’ —>5 Suki toka itte like say ’I go like ”I like her."’ 6 H:= Benkyoo shinai ano ko. study do-not that kid ’That kid doesn’t study.’ 7 K : A T—chan choo suki da mon . ah T-chan really like BE MN ’Ah, thing is, I really like T—chan.’ 8 H:A 50:. oh so ’Oh, really.’ —>9 K:Soo yo toka, HHHH (HHHH) so IP ’"Right", or something.’ In line 4, K says suki ’I like her’ straightforwardly. repeatedly However, there is no response from H. K then repeats suki again by quoting it with toka itte in line 5. I interpret that the speaker K feels a little awkward in this context in telling her feelings directly. Toka itte is introduced by K to express her embarrassment and at 146 the same time to avoid awkwardness by being playful. Horiuchi (1999) points out that toka ittoku ’say like’ serves to conclude a serious utterance in a funny way or to express embarrassment. All instances of toka in (19) support his point. H does not seem to know how to react to K’ s statements , and she avoids responding. This awkward interaction is rescued by K’s utterance with toka in line 9, which evokes a loud laugh from the two speakers. Thus we have seen various versions of toka quotations , which appear mainly in youth speech. Toka in quotation is used to mark the speaker’s hearsay or thought directly or indirectly. Toka implies possible minor discrepancies between the speaker ’ 5 quotes and the actual utterances or thoughts. The speaker protects him/ herself by indicating imprecision and approximations. In a narrative type of discourse, toka coordinates quotations and contributes to the text coherence. Toka is used to introduce other voices, and it is also used to self-quote the speaker’s own utterances . Self—quotation has an effect of making the utterances indirect, non-assertive, and sometimes playful, and thus lighten up the conversation atmosphere when it is awkward or embarrassing . 5 .2 . 5 . Summary of functions of toka and Schiffrin’ s discourse model In this section, I explained the basic functions of toka as a marker of inexhaustive coordination , vague reference and vague quotation. When each toka usage was closely examined, I found more 147 specific functions of toka in its particular linguistic and social context . Toka in inexhaustive listing coordinates representative items , thoughts or hearsay within an utterance or across utterances in a cohesive way. Toka as vague reference was used to express uncertainty, for approximation, or for soft highlighting, while showing the speaker’s self—protective/ non-committal attitude. Toka also marks a direct or indirect quotation, or other- or self-quotation, and at the same time implies that the information marked by toka is not complete or precise. Depending on the context, it can also display the speaker’s emotional disposition such as playfulness or awkwardness, or help avoid being too serious. Let us summarize the analysis of toka in relation to Schiffrin’s discourse model( 1987) . In general, the functions of toka as vague quotation marker operate in four planes . The primary role is on the participation framework because toka expresses the speaker ’ s orientation to the utterance and to the addressee, which in some cases implies a non—committal self-protective attitude and in other cases expresses playfulness or rapport. Toka also combines idea units (coordinating quoted propositions and utterances in a cohesive way), operating in the ideational structure. Toka is sometimes used in turn taking and floor holding (a sequential role), operating in the exchange structure. In the informational state, toka serves to organize what the speaker knows and often to imply that the knowledge is incomplete. The core 148 meaning of toka (inexhaustive coordination) is constant, but linguistic and social contexts determine in which planes of discourse toka has important roles. Thus toka is multifunctional, and some of the functions overlap with each other. Table 5.7 summarizes the specific functions of toka in relation to Schiffrin’s discourseemodell6. The major planes in which toka is operating are marked for each function by (*). Table 5.7. Functions of toka and Schiffrin’s discourse model planes oftalk specific functions (1) (2) (3) (4) (5) informa partici idea exchange action -tion -pation.-stional to express uncertainty, imprecision, lack of * * information to coordinate utterances, to mark successive * quotations sequential roles (e.g. turn- yielding, floor holding) * to express mitigation, self-protection, * * non-committal attitude to express playfulness, awkwardness, to avoid * being too serious (1) Information state (2) Participation framework (3) Ideational structure (4) Exchange structure (5) Information state 2% See Chapter 2 (p. 23) for an introduction of the model. 149 5.3. The use of teyuuka (’or rather’) Both toka ’or something’ and teyuuka ’or rather’ occur typically in spoken discourse, and are often considered as wakamonogo ’youth language’ . The Japanese encyclopedia Gendai yoogo no kiso chishiki [Basic Knowledge of Current Terms] , which is published annually in Japan, lists wakamono yoogo ’youngster ’ 8 terms’ every year, and has included toka since 1990 and teyuuka since 1992 in the list of vague expressions in youth language. The basic function of toka and teyuuka is to provide alternatives. In ”A toka B” an ”A teyuuka B”, they both present or imply options (A and/ or B) and mark the relationship between them. In general, toka coordinates them conjunctively, whereas teyuuka signals a repair. When toka and teyuuka are used at the end of an utterance, they both add a vague, unassertive or uncertain tone of voice, which creates a mitigating effect in the utterance. In Chapter 4 it was found that teyuuka was employed more often in the younger speakers ’ conversation than in that of older speakers’. Recall that only the age variable produced a main effect with a significant difference, (F(1,36)=5.65, MSE=5.04, p = .02167) . I repeat the quantitative results with teyuuka in Table 5.8. Table 5.8. Tokens and mean rates of teyuuka ’or rather Chat (N = 40) interview (N =40 ) IM IM OF OM IM IM OF OM total wordsl 8478 8735 8238 8546 8766 8592 11083 12886 tokens 32 13 8 13 22 19E 9g 21 ratio 3.84 1.34 0.99 1.54 2.44 2.14 0.86 1.60 1 50 In what follows, I present the qualitative analysis of teyuuka as a repair marker, comparing the two age groups . Differences are found in the linguistic environments in which teyuuka appears, in the orientation of teyuuka (self-repair or other-repair) , and in the functions of teyuuka. I classify six major functions of teyuuka in the present study: correction, rephrasing, specifying, blurring, subtopic shift, and supplementary explanation. 5.3.1. Review of the literature on teyuuka Teyuuka is not listed in any major Japanese dictionary because of its new or unestablished status in the Japanese lexicon, and also because teyuuka is a combination of elements, te (quotative particle), yuu (’say’) and ka (particle expressing question or doubt). Gendai yoogo no kiso chishiki (Basic Knowledge of Current Terms, 1998) explains that teyuuka is a prefacing word without any semantic meaning, and is used to start the utterance. To the best of my knowledge, no study has explored the use of teyuuka except for Yoshizawa (1998) . Yoshizawa studied the use of teyuuka in Japanese younger speakers’ conversation (recorded from 13 men and 8 women, ages around 20) . Based on the conversation data between friends, she claims that there are four major functions of teyuuka; modification (rephrasing, restating, repetition, and ending one’s utterance), background repair, replacement repair, and displaying speaker’s orientation. She 151 found that out of 98 teyuuka tokens, 69 (70.4%) are used for modification of utterances. She also points out that teyuuka as a discourse marker contributes to the text coherence because of its coordinating functions. She concludes that teyuuka indicates a difference between the utterance and the speakers ’ thoughts or feelings, and that it is used to repair the difference. I agree with Yoshizawa (1998) , who claims that teyuuka has a meaning and function similar to I mean in English. Schiffrin (1987) explains that I mean has its primary role in the participation framework and marks the speaker’s orientation or modification of his/her own talk. Schiffrin also points out two types of self-repair functions in the use cfif.I.mean. One is background repairs, with supplemental information to modify hearers’ understanding of surrounding material, and another is replacement repair, with ”substitutions of prior material” which ” switch the direction of the developing discourse to that initiated by the substitution” (pp. 300-301) . I show below how her analysis fits to the use of teyuuka. 5.3.2. General characteristics and immediate linguistic contexts of teyuuka Teyuuka consists of to (quotative particle), in (’say’), and ka (question particle). To iu (’say that’) marks that the previous phrase or proposition will be clarified next (Kojien p. 117). The main meanings of A to iu B are ’A is called B’ or ’A 152 is the same as B’ (Nihon Kokugo Daijiten [Comprehensive Japanese dictionary]1975:557) . According to my analysis, teyuuka is originally a combination of toiu and ka (a particle which marks question or doubt). Toiuka ’or should I say’ has subsequently become diversified with variations such as teyuuka and teka (see Table 5.9). Table 5.9. Variants of teyuuka ’or rather’ younger older total teyuuka/teiuka 53 (61.2%) 42 (82.4%) 95 teka/tsuka/tsuuka/chuuka 30 (34.9%) 6 (11.8%) 36 toiuka 3 (3.5%) 0 (0%) 3 toiimasuka 0 (0%) 3 (5.9%) 3 total tokens 86 (100%) 51 (100%) 137 The younger speakers use short and casual versions of teyuuka ( teka, tsuka, tsuuka, chuuka) more frequently than the older speakers. The formal variant toiimasuka is only found in the older speakers ' utterance in my data. When teyuuka follows a glottal closure immediately after the segment of the prior word, it is written as tteyuuka (with an extra /t/ in the beginning). Table 5.10 presents the immediate linguistic environments and the distribution of teyuuka in two age groups. Table 5.10. Liguistic environment for teyuuka ’or rather’ younger older total (1) ...teyuuka... 60 (69.8%) 43 (84.3%) 103 (2)1’eyuuka. . . 19 (22.1%) 0 (0%) 19 (3) ...teyuuka. 7 (8.1%) 8 (15.7%) 15 total tokens 86 (100%) 51 (100%) 137 153 The examples of each context are presented in the section of analysis (5.3.4). In Table 5.10, the form (l)’...teyuuka...’ means that teyuuka is used between phrases or sentences, (2) ’Teyuuka.. . ' means that it prefaces the utterance, and (3) ’ . . . teyuuka. ’ indicates that the utterance ends with teyuuka. Differences between the two age groups are found mainly in the tokens of (2) ’ Teyuuka. . . ’; this type of usage is found only among the younger groups. The utterance-initial teyuuka in (2) is often used to repair other’s talk. This type of repair is termed ”other-repair". In ”A teyuuka B”, A and B indicate two options and A is substituted by B. ”Other—repair” means that B is provided by another speaker (not by the speaker who provided A). On the other hand ”self— repair” means that both A and B are provided by the same speaker. The table below shows the comparison of the orientation of teyuuka . Table 5.11. Orientation of repair in teyuuka ’or rather’ younger older total self-repair. 55 (64.05%) 46 (90.25%) 101 other-repair 31 (36.0%) 5 (9.85%) 36 total 86 (100%) 51 (100%) 137 Note that other-repair is mostly employed by younger speakers. Older speakers ’ use of teyuuka is generally limited to self-repair . 5.3 .3. Basic functions and distribution of teyuuka Repair is defined as ”instances in which the speaker stops in some way in the course of producing an utterance, and then 154 repeats or replaces some part or all of it” (Hayashi 1994:77). Teyuuka shares the basic features with I mean as a marker of the speaker’s orientation and modification. Schiffrin’s (1987) two classifications of I mean (substitution and background repairs) can also be used for the analysis of teyuuka. Table 5.12 classifies teyuuka tokens according to six major functions. Table 5.12. Functions of teyuuka ’or rather’ younger older total (1) correction 35 (40.7%) 13 (29.4%) 48 (35.0%) (2) rephrasing 11 (12.8%) 12 (23.5%) 23 (16.8%) (3) specifyirLg 10 (11.6%) 9 (17.6%) 19 (13.9%) (4) blurring 9 (10.5%) 14 (27.5%) 23 (16.8%) (5) subtopic shift 5 (5.8%) 1 (2.0%) 6 (4.4%) (6) supplementary 15 (8.1%) 2 (3.9%) 17 (12.4%) explanation other(repetition) 1 (1.2%) 0 (0%) 1 (0.7%) total 86 (100%) 51 (100%) 137 (100%) The usage of teyuuka is very diverse for both age groups . The most common usage is (1) correction. In the younger speakers ’ speech, teyuuka is also used to provide (5) subtopic shift and (6) supplementary explanation. The next sections illustrate each function. 5.3.4. Analysis of the use of teyuuka Teyuuka is not yet listed in any major Japanese language dictionary. Teyuuka seems to be currently going through its 0 I 0 l7 0 O I I routinization and leXicalization process. The more teyuuka is 17 See Schourup (1985: 10-13). 155 used routinely, the more it is conventionalized and apt to be established in the Japanese lexicon. 5.3.4.1. Teyuuka for correction The example in (20) represents the most frequent function of teyuuka, correcting a part of the preceding utterance. (20) [During the interview, F (OM) is explaining that his hobby is to make potteries . He tells the interviewer that he prefers making them by hand to making them with an electric potter ’ 5 wheel . ] 1 F:Ee watashi wa moo tebine Yeah I TP indeed making—by-hand 2 shika dekinai n desu yo. can-only—do NM BE IP ’Yeah, I can’t make (potteries) any other way than by hand. ’ —>3 Dekinai tteyuuka sore ichiban suki de, can-only-do that best like BE ’Rather than I can’t, I should say that it's the way I like best.’ 4 I:Aji ga arimasu yone. class SP has IP ’It has class, you know.’ 5 F:Ee. Yes ’Yes.’ In lines 1 to 2, F first says that he cannot make potteries any other way than by tebine method ( ’by hand’ ) . However, he thinks that the expression ’cannot make’ is not appropriate. Then teyuuka is used to initiate a repair. Teyuuka here is best translated by ’or rather’ , or ’or I should say’ because it indicates 156 that something is wrong with the preceding utterance. Shika dekinai (’cannot make’) is the repairable, teyuuka signals the correction, and then the repairing segment sore ichiban suki de (’1 like it best’) follows. Teyuuka has a remedial function in this utterance. In Schiffrin’s discourse model (1987) , the primary role of this type of teyuuka , like I mean in English, is located in the participation framework because it shows the speaker’s intention of repairing the prior ideas. The secondary role is found in the ideational structure since teyuuka prefaces the expansion of the prior ideas. Additionally, because speakers' orientation to ideas is also related to knowledge about the content, teyuuka is relevant to the information state as well . These multifunctional features are characteristic of teyuuka in all its types, as shown below. The excerpt in (20) presented an example of self-repair. The excerpt in (21) presents an example of other-repair. (21) [T (YM) is asked by the interviewer if he eats out with his friends when he gets together with them.] 1 IzMinna de atsumatte donna koto suru n desu ka. everybody with gather what thing do NM BE Q ’What kind of thing do you do when everybody gets together? ’ 2 Gaishoku desu ka 7!, eat—out BE Q ’Do you eat out?’ —>3 T: Iya gaishoku ttsuuka moo sono-hen sono-hen ni, no eat-out just around-there around-there 4 tamatte suwarikonde. (I: Ee.) hang-around sit-down-for-a-while (I: yeah) ’No, not like eating out, we just hang around and sit down for a while around there..., around there.’ In lines 3 to 4, the repairable gaishoku ’eating out’ is corrected by the repairing sentence moo sono-hen sono-hen ni tamatte suwarikonde ’ just hanging around and sit down for a while around there...’. The utterance—initial Iya ’No’ in line 3 clearly indicates that gaishoku (’eating out’) is not the answer. The repairable gaishoku (’eating out’) was first uttered by the interviewer in her question in line 2. Then T corrects the repairable in his turn. This is a case of other—repair, because the repairing utterance is provided by another speaker. T could have simply used janakute ’not’ instead of teyuuka in line 3 to negate the interviewer’ s guess . However, with teyuuka, gaishoku (’eating out’) sounds as if it is not totally wrong, and thus mitigates the face-threatening act of disagreeing. The use of teyuuka by T is based on negative politeness (being nonassertive so as not to interfere with the addressee’s freedom of action). In the previous section, I mentioned three planes in Schiffrin’s discoursemodel (1987) on which teyuuka functions (the participation framework, the ideational structure, and the information state). Additionally, depending on the context, teyuuka operates in the exchange structure which concerns turns 158 and sequential roles. Teyuuka as other-repair is used across turns in the speaker-bearer interaction. Serving as a contextual coordinator, it brings coherence between two turns by different speakers. There are 36 occurrences of other-repair in the data of the present study. Out of these 36 tokens , 31 were found in the younger speakers ’ speech (see details in Table 5 . 11 on p. 153) . Table 5 . 13 below shows the distribution of teyuuka as other—repair function for both sexes and chats and interviews. Table 5.13. Teyuuka as other-repair for the younger speakers (IF and IM) II“ !M total chatting 15 7 22 interview 1 8 9 total 16 15 31 Both IF and YM speakers use other-repairs at almost the same frequency (16 for IF and 15 for YM) . However, a clear difference is found in the distribution according to speech situations. IF use teyuuka for other-repair mostly with their friends, not with the interviewer, whereas this difference is not observed in YM’s speech. This suggests that YF are more reluctant to correct other utterances in interviews. For YM, teyuuka seems to be a useful means to express disagreement in a soft way. Overall , teyuuka for other-repair is employed more often in casual conversion than in interviews. 159 5 . 3 . 4 . 2 . Teyuuka for rephrasing The excerpt in (22 ) shows a case in which the repairing phrase is a paraphrase of the repairable. (22) [M (OM) is talking about the similarity between German and Japanese with his friend U (OM) .] 1 M: Nantonaku, Nihonjin to no koo somehow Japanese with GP like-this —>2 shinjoo teyuuka kimochi ni chuuningu suru tokoro ga, feeling sentiment to tuning do point SP ’Somehow, like this, [Germans] have characteristics that are in tune with Japanese feelings or should I say sentiments.’ 3 U: Arimasu kaN, have Q ’They [Germans] have, you think?’ 4 M: Aru- aru n janai ka to omou. have have NM TAG Q QT think ’I think that they hav- have.’ In line 2 , M replaces the word shin joo ’ feeling , heart ’ with kimochi ’ feeling, sentiment, mood’ . Judging from the general meanings of the two words listed in Kojien Dictionary (1998) and Kenkyuusha ’s New Japanese-English Dictionary (1992) , these two words are synonyms , though there may be a slightly different nuance depending on the context. Teyuuka here is best translated as ’in other words’ . Why does the speaker present two similar words? This seemingly redundant repair may have some roles . For example, the speaker can show his/her slight preference between two options (the repaired one is better) and can provide more examples. The 160 speaker might have wanted to repair a phrase with a more appropriate expression, but ended with a similar word. In this case, the speaker can still show his/her intention of repair, indicating that the first word is not exactly what he/she means and he/she islookingforaabetterexpression. Withteyuuka,thespeakercan also gain time to construct his upcoming utterances. 5.3.4.3. Teyuuka for specifying Teyuukaisalsousedtospecifyone’spoint. Consider(23). (23) [N (YF) is providing examples for the types of things she has to teach to new employees at her part—time job.] 1 I:Dooitta koto shidoo suru 11 desu kaZ. what—kind-of thing instruction do NM BE Q ’What kind of thing do you teach?’ -+2 N:Uun mazu sekkyaku no kihon tteyuuka nanka, umm first wait-on-customer GP basic like ’Umm, first, the basics of waiting, or say, like’ 3 Aisatsu toka egao toka sooyuutokorokarahajimatte, greeting etc. smiling etc. such things from start ’starting from greeting, smiling and stuff like that,’ 4 Reji no uchikata toka. ... cashier GP how—to—punch ’How to use cashiers or something.’ 5 Sooyuu koto o oshiete itte:, such things DO teach keep ’1 keep teaching such things,’ The utterances in lines 3 and 4 list examples to elaborate the phrase in line 2, sekkyaku no kihon (’the basics of how to treat customers’). Teyuuka signals a turning point from the general 161 concept to its specific examples , and initiates this modification. The speaker was probably unsure whether the addressee understood the situation by only hearing sekkyaku no kihon (’the basics of how to treat customers’), and therefore she decided to add more information. This process of wondering and deciding is expressed in teyuuka with a little hesitating tone of voice. Among Schiffrin’s (1987) classifications of repairs, this type of teyuuka can be classified as background repair. Teyuuka pref aces the expansion and specification of the speakers ’ own prior idea, which is more general. The expansion and specification are subordinate asides, and the repair goes back to the point where the discourse was interrupted as the speaker N concludes in line 5, ’I keep teaching such things’. 5 . 3 . 4 . 4 . Teyuuka for blurring Above we saw teyuuka appearing utterance—medially. The excerpt in (24) presents the use of teyuuka appearing utterance-finally. Although in the examples so far teyuuka shares some features of I mean in English, this utterance—final teyuuka behaves differently from I mean. (24) [K (YM) is telling the interviewer that he does not want to take over his father’s business.] 1 KzMaa ie tsugu no wa ichiban da well take-over NM TP best BE 162 2 to omou n desu kedo yappari, QT think NM BE but after—all ’Well, although I think that it is the best thing to take over [my father’s business] after all,’ 3 Demo oya mo jibun no sukina koto yare tte but my-parents also myself GP like thing do QT 4 in fuu ni itte kureteru n de, like tell-me NM BE ’But my parents also tell me to do things that I like.’ —>5 Dakara, dakara, teyuuka, un. so so um ’80, So, or should I say..., um.’ 6 I : Naruhodo, muki fumuki arimasu yone. I-see forte—and-foible have IP ’ I see, we have things we are good at, and things we’re not, y’know.’ K uses teyuuka after his statement dakara (’so’) in line 5, and does not provide further explanation. In line 1 through 4, K implies that he does not want to take over his father’s business and that his parents are OK about it . Therefore he probably wanted to say that he won’t take over, but he leaves the point vague by using teyuuka without providing the repairing segment. This type of sentence-final teyuuka functions in a similar way as the sentence-final toka. Teyuuka implies that there is a better way to say it, and by not providing it the utterance sounds incomplete, vague or indirect. This type of teyuuka often has a blurring and softening influence on the utterance. Teyuuka in line 5 also signals a position at a turn transition. The speaker yields the turn of his speech with teyuuka, and after that, the turn is taken by the interviewer. In this sense, teyuuka 163 contributes to the textual organization, which operates in the exchange structure in Schiffrin’s discourse model (1987). 5.3.4.5. Teyuuka for subtopic shift We now turn to the use of teyuuka appearing utterance- initially to shift the subtopic as in (25). (25) [Both.S and K (YF) want to go to Disneyland abroad.(Europe and America).] 1 SzYuuro tte gaikoku n02 Euro QT abroad one ’You mean, Euro is the one abroad?’ 2 K : Gaikoku no, abroad one ’The one abroad.’ 3 S:A itte mitai, oh go try ’Oh, I wanna go.’ 4 K:Itte mitai yone go try IP ’Right, we wanna go there.’ -+5 S:=Teka kibo ga chigau yone, (K: Un) scale SB different IP (K: Right) ’Because the scale is different, you know’ —>6 K: Teka amerika no dizuniirando choo dekai n da tte, America GP Disneyland damn big NM BE QT ’Or, Disneyland in America is damn big, I heard.’ Let us examine the use of teka in line 6 (teka in line 5 concerns another function to be discussed in the next section , supplementary explanation). First, the two speakers K and S are talking about the Disneyland in Europe and agreeing that they both want to go 164 there. Then in line 6, the topic of the conversation is shifted from the European Disneyland to the one in America. This transition of subtopics within the general topic of Disneyland is initiated by the sentence—initial teka. With teka in line 6 K takes her turn to provide information (point-making) , and marks the subtopic shift. This type of teyuuka is similar to Schiffrin's replacement repairs . Replacement repairs , according to Schiffrin (1987: 301) , ’ lead forward to the ideas of upcoming discourse ’ without returning to the previous points. Instead of providing background information, the utterance initiated by teka in line 6 further develops the conversation on Disneyland. 5.3.4.6. Teyuuka for supplementary explanation As I mentioned earlier, teka in line 6 in (25) is a case of replacement repair. Teka in line 5 in (25) is an example of a background repair, as it provides a reason why S and K want to visit the Disneyland in Europe (it is huge). Both tokens of teka in (25) , appearing utterance-initially, serve also as proposition highlighters in a soft way. They are, in other words, attention-getters, which function to direct the addressee’s attention to the upcoming utterance which is the point-making statement. Both utterances by K and S in lines 5 and 6 are preceded by teka, claiming the speakers ’ turns, and creating rhythmic adjacent pairs . This kind of interaction, which involves 165 different turns and two speakers, is not found in my data of older speakers’ conversations. The following example (15) was previously shown during my analysis of toka. I present a longer excerpt below. (15) [S (YF) is answering the interviewer’s question about the club activities at high school. She thinks that the club was fun and she felt free from restrictions because there were no seniors . ] 1 IzBukatsudoo wa doo deshita ka7| kookoo jidai wa. club-activity TP how BE Q high-school days TP ’How were the club activities in high-school days?’ 2 S:Ano senpai toka inakatta n desu ne, well senior exist-not NM BE IP ’Well, thing is, there weren’t seniors and such, ’ 3 Ano hai well yes, ’Well, yes,’ -—>4 Uchi no gakunen kara no bukatsu tteyuuka, our GP school-year form NM club ’The club consisted only of students from my class, or I should say’ 5 M00 minna sotsugyoo shichatte, already everybody graduation did ”Everybody already graduated,’ 6 Hai moo zutto inakatta n de yes already for-a-long—time had—not NM BE 7 kekkoo jiyuu tte kanji datta n desu kedo, quite free QT like BE NM BE so ’Yes , we did not have (any seniors) for a long time already, so it was like we had a lot of freedom, though. ’ Here the repairable is Uchi no gakunen kara no bukatsu (’The club consisted only of students from my class’) in line 4, and this fact is further explained in lines 5 and 6. The utterances 166 in lines 5 and 6 provide supplementary explanation, and this repair is initiated with teyuuka in line 4. Then, in line 7, 8 finally gives the interviewer her main answer to the question on her club activities (she had a lot of freedom). Lines 2 to 6 provide background information, and they all lead to the conclusion in line 7. Teyuuka in line 4 organizes the background information (lines 2 to 6) as shown in the textual structure below. I: Question (line 1) S: Support teyuuka (lines 2 - 4) Supplementary support (lines 5 - 6) Position (line 7) The last excerpt in (26) illustrates another case. (26) [K (YF) hopes to get a job which deals with flowers, specifically a job in a wedding hall.] 1 KzKekkon shiki no ohana o tsukuritai n desu yo. marriage ceremony GP flower DO want-to-make NM BE IP ’The thing is, I’d like to arrange flowers for wedding halls.’ 2 I:Ano buuke desu kaZ. Well bouquet BE Q ’You mean, bouquets? ’ 3 KzBuuke toka, watashi anoo shikijoo no zenbu, bouquet etc. I well wedding-hall GN everything ’Bouquets and stuff like that, I, well, everything involving a wedding hall. ' —*4 Tsuka sooyuu no mo dezain no benkyoo shitetara such one too design GP study doing-if 5 jibun de dezain shita no de dekiru kara myself by design did one with can because 167 6 ii na: to omotte:. good IP QT think ’Or I mean, for that kind of things, if I study designing, I can work with something I design by myself, so I think it’s good.’ Tsuka in line 4 has several functions. First it initiates the explanatory utterances (lines 4 to 6) for the statement in line 1. The reason why the speaker is interested in all kinds of flower-related jobs in wedding halls is that she thinks it would be nice to study designing and then use her own designs to handle such jobs. Tsuka not only combines utterances in a coherent way (marking statement and reason relations), but textually also signals the speaker’s continuation of her utterance, and so serves as a floor-holding device. It is also used as a highlighter for her point—making. The organization of repair in ”A teyuuka B” has some variations. We have seen that repair is done at the word level (A and B are words or phrases) and at the sentence level (A and B are sentences). For example, the excerpt in (22) shows repair at the word level, while the last excerpt in (26) shows repair at the sentence level. When teyuuka is used at the sentence level, it functions to coordinate idea units in a more global way than teyuuka at the word level. Teyuuka shows that upcoming utterances are the modification of the prior utterance or segments, and thus holds the speaker’ s turn so that he/she can provide related information. 168 The relation between these utterances or turns is not expressed obviously or in an assertive way because of the semantic vagueness carried by teyuuka. In A teyuuka B, B is preferred, but A is not completely rejected. This non-committal tone of voice may originate in the semantic function of ka (question particle) which implies the speaker’s doubt. Teyuuka serves to soften the utterance and blurs the relationship between the two alternatives . 5.3.5. Summary of functions of teyuuka and Schiffrin’s discourse model In this section, I examined the functions of teyuuka. The general function of teyuuka is to repair (a part of) an utterance in a mitigated or unassertive way. Teyuuka as a repair introduces a better alternative than the prior utterance while not completely rejecting the repaired element. I classified the functions of teyuuka into six basic categories: correction, rephrasing, specifying, blurring and subtopic shift. Teyuuka was used most frequently for correction (35% of all teyuuka tokens). While older speaker used teyuuka mainly for self—repair, younger speakers used it for both other- and self—repair. Especially younger male speakers employed teyuuka as other-repair during the interviews. Younger speakers also differ from the older speakers in the use of teyuuka in the utterance-initial position. They sometimes secured their turn of speech by using teyuuka at the beginning of the turn. In addition, 169 teyuuka was used to initiate a subtopic shift and to provide supplementary explanation among the younger groups, but this was rare among the older groups. Table 5.14 summarizes the specific functions of teyuuka in relation to Schiffrin’s discourse model. The major planes in which teyuuka is operating are marked for each function by (*). Table 5.14. Functions of teyuuka and Schiffrin’s discourse model pla nes of talk specific functions a) (2) u” (4) (5) informa partici idea exchange action —tion -pation -tional as self repair (SP's provision of inclusive option, * * * explanation, correction) as other repair (other SP’s provision of inclusive option, * * * * * explanation, correction) to express uncertainty, imprecision, lack of * * information sequential roles (turn initiator, attention-getter, * turn-yielding, floor holding) to express mitigation, non-committal attitude * * subtopic shift (expansion of ideas) * SP = speaker (1) Information state (2) Participation framework (3) Ideational structure (4) Exchange structure (5) Information state 170 5.4. Summary In the present chapter, I discussed the functions of toka and teyuuka while examining their linguistic contexts closely. General characteristics of toka and teyuuka are that they appear as hedges, and that they are both used to provide alternatives as in ”A toka B” and ”A teyuuka B” . Basically, toka inexhaustively coordinates items, sentences and quotations, whereas teyuuka repairs the prior utterance in a soft or nonassertive way. Toka and teyuuka in general imply that there are other expressions, and the speaker can make his/her utteranceropen or vagueeby not specifying the other possibilities as in ”A toka” and ”A teyuuka” . This way, the speaker can be non—committal and avoid being too assertive. Younger speakers employ more variations in the usage of toka and teyuuka than do older speakers. It is interesting that they sometimes use these expressions to be playful, to express awkwardness or their preferences and to avoid being too serious. This is one way how youngsters communicate, interact and express their youthfulness, showing their solidarity or rapport among themselves. 171 Chapter 6 Results of the Questionnaire on the use of hedges 6.1. Introduction The present chapter investigates the results of the questionnaire, and presents a comparison of the findings with the actual use of hedges discussed in Chapter 4. I first discuss how the participants perceive the use of hedges : Who uses hedges , when and why. Then I briefly examine some of the important issues concerning the psychological and social background of the contemporary Japanese youngsters in order to further understand their communication style and motivations behind the use of hedges. 6 . 2 . Questionnaire results After the interview and conversation with their friends, the participants were asked to fill out a questionnaire form. The purposes of the questionnaire are to investigate what speakers think of the use of hedges, especially hedges common in conversation by Japanese younger speakers, and to compare the questionnaire results with the actual use of hedges. It is hoped that the findings with this questionnaire will provide some insights about why such hedges are used in conversation and how they function in interpersonal communication (see Niedz ielski and Preston 2000 , for a discussion of the use of metalinguistic data) . I am aware that this type of metalinguistic investigation through 172 participants’ self-report has limitations because the self- reported information does not necessarily reflect reality. Therefore, these results are considered as subsidiary information. As I explained in Chapter 3, before the questionnaire was given, all participants were first asked to read two short conversation samples in Japanese which include several hedges, e.g. toka ’or something’, nanka ’like’, kekkoo ’quite’, teyuuka ’or rather’, kanji ’feel like’, mitai ’is like’ (see Appendix 4 for the Japanese writing that was shown to the participants). These two samples are based on actual conversations between high-school students (single sex, one female and one male pairs) which were recorded in my preliminary study. The samples were chosen as they reflect the use of hedges by youngsters in conversation. Conversation samples (161) 1 A: Nankaa, Tanaka-san wa fasshon toka mo chigakuttee, like Mr./Ms.Tanaka TP fashion etc. also differ 2 Itsumo hade-kee no fuku toka kiteru tte kanji. always showy—type GP close etc. wear QT feel—like ’Like, Mr./Ms. Tanaka’s fashion and things differ, and he/she always wears like showy—type of clothes , seems like . ’ 3 B: Tey_uuka, medachi-tagari-ya na n jan? or rather person-who-wants-attention BE NM TAG ’Or rather, he/she wants to attract attention, doesn’t he/she?’ 173 (1b) 1 CzMenkyo toru no? license get Q ’Are you going to get a license?’ 2 D:Menkyo, ttemuka toritai n da kedo, license or-I-should—say want—to—get NM BE but 3 Kekkoo jikan kimatchau wake jan? a bit time be-limited case TAG 4 Kaigai toka mo ikitai shi. abroad etc. also want—to-go and 5 Nanka ima wa toriaezu ii ya, mitaina. like now TP for—now OK IP like ’A license, or I should say, I want to get it, but it restricts my schedule a bit, doesn’t it? Plus, I want to go abroad and stuff. Like, for now, it can wait, sort of.’ The participants were asked to mark an appropriate number (ranging from 1 to 5) for Likert—type questions, indicating if, when and why they use hedges like those in the sample conversations and what they think of these hedges. For the statistical analysis, ANOVA, post—hoc test (Tukey’s HSD method), and PEARSON were used to evaluate the differences. Detailed results are presented in Appendices 11-23. 6.2.1. Perception of who uses hedges The first question asked of all participants was: ”Do you use the underlined type of expressions [hedges] in conversation?” Figure 6.1 on the next page sums the answers in a bar graph displaying the mean scores for the four groups . The choices given 174 to the participants range from 1 to 5 . The higher the mean number, the more the participants agree or think that they use hedges. Figure 6.1 shows that IE (younger female speakers) most strongly agree that they use hedges (mean = 4.4). Figure 6.1. Answers to "Do you use the underlined type of expressions [hedges] in conversation?" (n=40) yb u (A) o llYes II '_I I O N U 0 OU'Il—IUWNUIOJU'Iva'IUI YF YM OF OM While most of IF agree that they use hedges like the samples often, the answers from YM (younger male speakers) are divided into either ”not often” or ”sometimes". With older groups, except for one participant in OF (older females) and one participant in OM (older males) who answered ”sometimes” , they answered that they don’t or never use such hedges. Statistical tests by ANOVA showed that there was a significant effect of age and a tendency toward an effect of sex on the perception of using hedges (F(1,36)=58.39, MSE=0.79, p <.00001 for age and F(1,36)=3.82, MSE=0.79, p =.05544 for sex, see Appendix 11 . 1 for details) . No interaction was found between age and sex. Then, post—hoc tests were conducted to evaluate the 175 differences between the groups. There was no significant difference between IF and YM nor between OF and OM (see Appendix 11.2) . All other tests showed that there were significant differences between the younger groups and the older groups (p < .01 for all contrasts). This finding is similar to the overall result of the actual use of hedges discussed in Chapter 4. I repeat Figure 4.1 from Chapter 4 below; IBoth figures show a similar pattern among the four groups, suggesting that the participants generally have a realistic perception about their use of hedges. Only YM showed a higher score for perception of the use of hedges (Figure 6.1) than the actual use of hedges (Figure 4.1). Figure 4.1. Use of hedges by four groups, (mean rate/1000 words) 100 90 - 80 . 70 . 60 - 50 ' 40 ' 30 - 20 ’ 10 ' O [ .chat I: interview J Figure 6.2 summarizes the answers to the question: ”What do you think of using this kind of expression in conversation? Usual way of speaking?”. 176 Figure 6.2. Answers to ”Usual way of speaking?" (n=40) 5 4.5 - 4 . 3.5 3 . . 2.5 I”I think it’s usual” 2 1.5 1 0.5 0 Significant differences were found between the younger and older groups (F(1,36)=8.63, MSE=0.94, p = .00584, see ANOVA results in Appendix 12 . 1 for details) and between the female and male groups (F(1,36)=8.63, MSE=0.94, I’ = .00584). There was also an interaction between age and sex (F(1,36)=6.82, MSE=0.94, p =.01256). These results indicate that YF’s score is much higher thanthoseofcmhergroups. Post—hoccomparisonsalsoshowedthat there was a significant difference between YF and other groups (p < .01 for these three contrasts), but not among YM, OF, and OM (see Appendix 12.2). While YM, OF, and OM perceive the use of hedges as deviant from the norm, this perception is not present in YF. The use of hedges may not strike YF as unusual. This is probably one of the reasons why YF use hedges more often in conversation than do other groups. Another interesting point is that YM’s answers are similar to those of OF and OM (YF > YM z OF e OM). This tendency is also found in the actual use of hedges (Figure 4.1). The fact that 177 YF behave differently from YM is consistent with the participants ’ perception shown in the following figure. Figure 6.3 shows answers to "Do you think there is any gender difference in using this type of expressions?". Figure 6.3. Answers to ”Do you think there is any gender difference?” (n=40) N n 0 n - . OWHU’lNUlelnhUIU'I [Higher scores indicate that the participants think that women use hedges more often than do men.] Overall, all four groups tend to think that.women.use hedgeSInore often than do men. ANOVA shows that there is no significant age effect (F(1,36)=0.04, MSE=0.66, p =.82610), but there is a significant sex effect (F(1,36)=4.60, MSE=0.66, p =.03668, see Appendix13.lforckmails). Post—hoctestsshowedthattherewere no significant differences between the groups (see Appendix 13 . 2) . The participants are aware or feel that the use of hedges is associated with gender, in particular with women. Figure 6.4 summarizes the answers to ”Were you able to talk to your conVersation partner naturally?". No significant 178 difference was found in ANOVA (see Appendix 14) . Overall, most of the participants thought that they were able to speak naturally . Figure 6.4. Answers to ”Were you able to talk naturally?" 5 4.5 4 3.5 3 2'; "Yes" 1.5 l 0.5 O Figure 6.5 below displays the participants ’ answers to the question: ”Which age groups do you think use the underlined type of expressions in general?”. The participants were asked to evaluate the frequency of the use of hedges by six different age groups. Higher numbers indicate higher perceived frequency of hedge usage. To illustrate these data, I used curves rather than bars because there is an ordinal relation between the dependent measures (six different age groups). It is found that all groups consider high—school students to be the group that uses hedges most often. Generally, students of junior-high school, high school and university are regarded as the frequent hedge users. Between the younger and older participants, the perception splits for groups of ”20's—30's”, ”40’s—50’s" and ”60’s—”. IF and YM show lower scores for those 179 generations than do OF and OM. This suggests that IF and YM think that the use of hedges is restricted mainly to students. Figure 6.5. Answers to "Which age groups do you think use the underlined type of expressions in general? (n=40) a} 0 LA.) 0 H O O o [\J omr—amwmwmpmm L I l 1 junior high university 20's-30’s 40’s-50’s 60’s- high school school YF 214 OF OM ANOVA tests showed that there were significant effects of age (F(1,36)=11.66, MSE=1.20, p = .00193) and perception of frequency of hedges in the six age groups (age perception, F(1,36)=229.49, MSE=0.35, g><< .00001, see Appendix 15). An interaction was also found between age and age perception (F(1,36)=4.04, MSE=0.35, p =.00204) . This shows that the younger and older groups have different perceptions on the frequency of hedge usage by the six age groups. In this section the following results were found:(1) YF think that they use hedges often, which is consistent with the finding in Chapter 4; (2) IF do not consider the use of hedges as deviant while YM, OF, and OM do; (3) the participants of all 180 four groups agree that women use hedges more often than do men, and that junior high—school, high—school and university students are the leading groups in the use of hedges. 6.2.2. Situations when hedges are used The present section deals with the answers to the question ”On what kind of occasions do you use the underlined type of expressions?". Two styles, casual and formal, and the addressee’s age group are compared. Figure 6.6 (next page) summarizes the answers to the question on whether the participant uses hedges on casual and formal occasions . Each bar in the figure displays a mean score. Higher scores indicate a higher frequency in the use of hedges. This question was asked only of the participants who answered that they use hedges in the first question (”Do you use the underlined type of expressions in conversation?"): ten YF, six YML1and only one each from OF and OM groups . For this reason, OF and OM are excluded in the figure, and only IF and YM are included. It is evident that both IF and YM think they differentiate the use of hedges depending on the formality of the conversation situations . There are remarkable differences in the mean scores between casual and formal situations for both sexes (t-tests showed a significant difference between the two styles for IE as well as for YM, p < 0.0001 for both groups). 181 Figure 6 . 6 . Answers to ”On what kind of occasions (casual/formal) do you use the underlined type of expressions?” (n=16) casual D formal Let us compare Figure 6.6 with Figure 4. 1 which was previously shown in section 6.2 . 1. These two figures, the speakers ’ perception (Figure 6 . 6) and actual use of hedges (Figure 4 . 1), show that they are quite different. According to the participants’ self-report, they think that they use hedges more often in casual situations than in formal ones. However, in actuality, IF and YM used them in interviews as often as in chats (Figure 4 . 1) . The difference between the two styles (Figure 4 . 1) was not statistically significant for either of these two groups according to post—hoc tests (see Appendix 6.5). There is additional evidence that the participants consider hedges more appropriate for casual conversations. Figure 6.7 summarizes the answers to the question: ”What do you think of using this kind of expressions in conversation: should one refrain from using them depending on the occasion?”. 182 Figure 6.7. Answers to ”Should one refrain from using the underlined type of expressions depending on the occasion ?" 5 4.5 4 3.5 3 2 1.5 l 0.5 O Not only the older groups, but also the younger groups think that they should not use hedges on certain occasions. Figure 6.6 suggests that the participants probably meant formal occasions for not using hedges . There were no significant differences among the four groups (see Appendix 16). It is interesting that all YF participants marked 5 (”strongly agree”) for this statement. Then why is there such a difference between the speakers’ perception and the actual use of hedges in conversation? I leave this discussion to the next section. The same discrepancy is also found in Figure 6.8 below. The figure shows YF's preferences of using hedges with different addressees (friends, siblings, parents, older people/ superiors, colleagues and strangers). This question was asked only of the participants who consider themselves as hedge users (question item (1 ) ) . As we can see from the figure, YF claim that they do not use hedges when talking to Older people/superiors and strangers, but only when talking to 183 their friends or siblings. Higher mean scores in the figure indicate that the participants think that they use hedges , whereas lower scores indicate that they do not think so. Figure 6.8. Answers to ”Speaking with whom do you use the underlined type of expressions?” (IF, n=10) 5 4 . 5 4 3 . 5 3 2 . 5 2 1 . 5 1 0 . 5 0 friends sibl ing 5 parents 01 der / co lle agues strange rs superiors According to PEARSON analysis, it is found that there are strong positive relationships between friends and casual situations (r2 =. 802) and between siblings and parents (r2 =. 911) , and that there are strong negative relationships between siblings and formal situations (r2 = -.844) and between parents and formal situations (r2 = -.830)18. The detailed PEARSON results are presented in Appendix 17. These relationships are reasonable because if the conversation involves close persons such as friends , siblings and parents, the style of the conversation is likely to be casual, 18 The PEARSON analysis evaluated the situation variables (casual and formal), and the addressee variables (friends, siblings, parents, superiors, colleagues and strangers) together. 184 not formal. With this result, we can group the different types of addressees . Friends , siblings , and parents belong to a group with whom YF use hedges , and they are associated with casual situations . Older people/ superiors and strangers belong to a group with whom YF refrain from using hedges, and they are associated with formal situations. However, from the actual data in conversations and interviews this perception of IF does not coincide with their actual behavior. 6.2.3. Reasons to use hedges In my preliminary study, the participants were asked regarding the motivation for using hedges, and it was found that most of the participants were not able to explain or analyze why they use hedges. It was therefore decided that in the present study, the participants are given possible reasons for why they use hedges, and asked to mark to what degree they agree with each stated reason (with a number from 1 to 5, i.e., from strongly disagree to strongly agree). Figure 6.9 on the next page summarizes the results. YF agreed most strongly with (12) to show solidarity, (13) easy to say, (17) unconsciously, and (20) influenced by people around me. A strong positive relationship is found between (13) eaSy to say and (17) unconsciously (r2 =.726, see Appendix 18 for detailed PEARSON results). This result suggests that hedges are 185 easy to use for IF and while using them often the usage may have become unconscious . It may also be the case that YF are influenced in the use of hedges by people around them who use hedges. In general, it appears that IF have no specific motivation to use hedges. Figure 6.9. Answers to ”Why do you use this type of expressions?" (IF, n =10) A l A l l L (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) N n o o o o OUIHLJINU'IwmohUIUI (11) fashion or trend (12) to show solidarity (13) easy to say (14) for fun (15) to be vague (16) to evade my responsibility (l7) unconsciously (18) adapting to the other party's language (19) to soften the tone (20) influenced by people around me (21) when talking about unfamiliar topics (22) to avoid disagreement On the other hand, YF disagreed with (16) to evade my responsibility and (22) to avoid disagreement. It means that YF negate (though not strongly) the use of hedges for self—protection . 186 It is questionable, however, whether YF are really certain that reasons (16) and (22) are false. It may just be that they did not understand reasons (16) and (22) , or that these reasons sounded too deliberate for YF. According to Figure 6. 9, most of IF speakers think that they are influenced by people around them. Also ranked high was reason (18) adapting to the other party’s language, which suggests another type of ”influence” on YF’s use of hedges. This issue of ”influence" is also supported by the results in Figure 6.10. Figure 6.10. Answers to ”Is there anyone around you who uses this kind of expressions?” (n=40) H N LA) vb o o I OU'H—‘UlNUIbJU'thIUl O 0 Y F YM OF OM Figure 6.10 shows that YF think that they are surrounded by hedge users the most, and a similar result is obtained for YM. There was a significant effect of age (F(1,36)=20.98, MSE=0.81, p = .00016, see Appendix 19.1 for detailed ANOVA results), suggesting that youngsters have more input of hedge usage in their daily lives as compared to the older groups. They listed classmates, friends at school, and their siblings as people providing such input. According to post—hoc tests, significant 187 differences were found between the younger groups and the older groups (p‘< .01 for YF-OF, YF—OMLIand YM-OF; p'< .05 for YM-OM, see Appendix 19.2 for details). OM showed a high score than did OF (p < .01) , although it was not as high as that of the younger speakers. The ANOVA result showed that there is an interaction between age and sex (F(1,36)=4.47, MSE=O.81, p =.03917). We now turn to the perceptions or views the participants have on the use of hedges. Figure 6 . 11 (next page) displays bars for each view by the four groups of participants. As previously explained, YF stand out in (31) usual way'of'speakingu They are the only group that did not show disagreement on this item. YF also stand out in (32) fun, which suggests that some of IF enjoy using hedges in conversation. This finding is supported by statistical tests. The ANOVA result showed that there was an interaction between age and sex (F( 1 , 36 )=4 . 41 , MSE=1 . 45 , p =. 04032 , see Appendix 20.1 for details). It was also found that age and sex did not influence the results (F(1,36)=1.73, MSE=1.45, p = .19473 for age, F(1,36)=0.28, MSE=1.45, p = .60870 for sex), indicating that there‘was:no significant.difference between the younger and older groups and between female and male groups. Post-hoc tests showed that there were no significant.contrasts (Appendix 20.2). Almost all participants agreed strongly with (36) one should refrain from using them on some occasions, which is also discussed in the previous section. 188 Figure 6.11. Answers to ”What do you think of using this kind of expressions in conversation?”(n=40) 5 4.5 - 4 I 3.5 - 3 . 2.5 - 2 _ EYF 1.5 - DYM 1 ' IIOF 0.5 o (31) (32) (33) ( (35) (36) (31) usual way of speaking (32) fun (33) shows solidarity (34) unpleasant/inappropriate (35) corrupt (36) one should refrain from using them on some occasions Most of the participants also think that language using hedges sounds corrupt (35). The ANOVA showed that there was no significant difference between the two age groups (F( l , 36)=0 . 25 , MSE=1.01,p=a84915)norbetweenthetwosexgroups(F11,36)=1.21, MSE=1.01, p=.27727) for (35) corrupt. See Appendix 21 for detailed ANOVA results. This perception regarding (35) has a positive relationship with (34) unpleasant/inappropriate in the older groups (I'2 =.7992, see Appendix 22.2 for details). This means that (34) and (35) can be grouped together as negative views that the older group has on the use of hedges . This negative view is very contrastive with YF’s view, (32) fun. 189 Another finding is that YM show the highest mean score for (33) showing solidarity. The ANOVA showed that there was a significant effect of age (F(1,36)=4.34, MSE=1.13, p =.04182) and sex (F(1, 36 )=8 .87, MSE=1 . 13, p =. 00531 , see Appendix 23 . 1). This indicates that the younger groups agree with the reason (33) showing solidarity more strongly than do the older groups, and the male groups than do female groups. A significant difference was found between YM and OF (p < .01) and between OF and OM (p < .05) in post-hoc test (see Appendix 23.2) In sum, YF reported that they mostly used hedges unconsciously and that using hedges in conversation was not an unusual way of speaking but fun and easy to say. The way of speaking with hedges is probably influenced by their friends or siblings, considering that IE are aware that they are surrounded by people who use hedges. On the other hand, the older groups showed more negative views on hedges than did the younger groups (e.g. hedges are unpleasant or inappropriate language). 6.2.4. Discussion In this section, two questions will be mainly discussed. One is why younger speakers use hedges often, and another is why there is a difference between the speakers’ perception and the actual use of hedges in the two different styles. Based on the participants ’ self-reports, the answer to the former question is perhaps that younger speakers, especially YF, do not have a 190 negative view on the use of hedges like the older speakers do. Rather, YF think that hedges are nothing out of the ordinary, fun and easy to say. The younger speakers also think that they use hedges to show solidarity among close people. It was found that YF are not conscious of the use of hedges and unaware of the self—protective effect hedges have in some contexts. Instead, they seem to view and use hedges more as an in-group marker, though it is not clear whether they realize this. Hedges as an in-group marker is also suggested in the younger speakers’ opinion that one should refrain from using hedges in formal situations or when talking to older people, superiors, and strangers. Hedges can be used, in their view, when talking to friends and siblings, or people who are inside the group, and in casual situations. This view is shared also with other groups. Additionally, all participants agreed that the hedges shown in the samples are typical expressions that belong to junior high school students, high school students and university students, especially female students. This indicates that the use of some hedges is stereotypically associated with the young generation of students, often with female students. In the present study, the actual use of hedges in the younger groups did not reflect their views that hedges should not be used in formal situations. There are possible explanations for this discrepancy. As discussed in Chapter 4, one is that for IF the interview with the present investigator may not have been formal 191 enough to refrain from the use of hedges. The interviewer was a stranger, older, and somebody outside of their in-group. Imi principle, then, the situation should be regarded as a formal one. However, YF possibly did not look at the situation this way. YF were likely being friendly to the interviewer (30 year old female) by not showing too much formality in their language. There were, however, distinct differences in their speech between the two styles; e.g. formal endings (desu/masu-forms) were used sentence-finally in interviews . Nevertheless , regarding the use of hedges, it is possible that YF did not make a lot of effort to avoid hedges because they thought it appropriate. Additionally, as YF claim, the use of some hedges can be unconscious or habitual, and it may be difficult for YF not to use them in their utterances unless the situation requires a great deal of formality. Another explanation for the discrepancy is found in the types of hedges they use in chats and interviews . The examination in Chapter 4 showed differences only in the frequency of hedges. However, the usage of hedges (toka ’or something' and teyuuka ’or rather’) analyzed in context, as in Chapter 5, varies slightly from context to context and exhibits different functions. For example, toka ’or something’ as a vague quotation was used more often in chats than in interviews. Also, YF used teyuuka ’or rather” to show disagreement mostly in chats. Thus, particular uses of hedges disappeared when the style was more formal. It 192 is possible that specifically these uses of hedges are the ones that the participants have in mind when they say that hedges should not be used in formal settings. Such distinctions according to style were observed only for toka and teyuuka, but a close examination of each hedge in context may reveal more subtle changes . I leave this question for future study. 6.3. Psychological and social background of the use of hedges In this section, I discuss some of the important issues in the psychology of adolescence, while considering the social backgrounds of the Japanese moderns and results from other studies . Although I am fully aware that such a discussion is largely speculative, there does exist a large academic literature on this topic that might provide some insights into the dynamics of Japanese youth language in general, and the use of hedges by younger speakers in particular (F. Inoue, personal communication). As I discussed at the end of Chapter 4, some general characteristics of being adolescent might have an influence on linguistic choices. For example, youngsters are in general more free from social norms or responsibilities than older generations , which makes it easier for youngsters to create their own or new communication styles (e.g. Yonekawa 1998) . Also, youngsters are less experienced in life, which may lead to uncertainty or 193 non—assertiveness in their statements on certain topics. In addition to these points, let us discuss two important features of adolescence: instability and avoidance personality. First, in general the term ”adolescence" refers to the period between puberty (11 to 13 years old) and adulthood (after approximately 20 years old) (e.g. Suzuki and Matsuda 1997). During this stage, the adolescent goes through big physical and psychological changes. An adolescent’s unstable personality or physical condition is often attributed to this pubertal development and the related psychological and/ or physical factors such as difficulties in accepting changes of the body image or dealing with sexual development (e.g. Suzuki and Matsuda 1997; Shimizu 1998). According to Freud 1936/1985 (translated in Sotobayashi 1985) , physical (sexual) development causes an unbalanced self and self-protective behaviors. Because of these developmental characteristics, adolescents are often thought to have an unestablished identity or a lack of self-esteem. Due to their instability or low self-esteem, they are afraid of communicating with others, relatively easily influenced by people around them, and/or sometimes self—protective (Nakanishi 2000) . They may be sensitive to trends , and easily follow what others do, though they may not be aware of these facts. It is reasonable to assume that youth language reflects such characteristics. Their fear of rejection, and of being different from others may be related to 194 the frequent use of hedges. Hedges are a convenient tool for youngsters because they enable them to avoid self-disclosure by being unassertive and non—committal in the communication. While instability in adolescence may be a universal phenomenon, the issue to follow is related more to contemporary Japanese society and its possible influence on the youth. It has been observed that the contemporary Japanese tend to have an ”avoidance personality" (e.g. Harada 1990; Sengoku 1994; Kageyama 1999). According to these authors, avoidance personality characterizes a person who avoids interpersonal contacts, self-disclosure, and/or creating a new relationship because of fear of rejection or due to inferiority complex. Typical of the avoidance personality is that the youngster is afraid of revealing him/herself and of getting hurt. As a consequence, he/ she prefers to have only superficial communication with others. Here, hedges can play an important role since they often make vague references, and convey self- protective or non—committal attitudes. It is difficult to pinpoint one crucial reason why Japanese youngsters tend to develop an avoidance personality. The cause may be found in their social backgrounds. As Kageyama (1999) claims, many youngsters are growing up without having many direct experiences and personal interactions in their social or family lives nowadays. Direct experiences include, according to Kageyama, senses of achievement and failure, and intimate 195 face—to-face interactions between close people such as fighting, arguing or reconciling. He claims that the lack of these experiences is influential in creating indifference among adolescents and eventually leads to an increase in serious crimes , as have been committed by Japanese adolescents in recent years. There are several possible social causes for this tendency. Often mentioned is that small-sized or ”nuclear” families are more common nowadays than ever before, with the youngster being an only child. Also, youngsters are often under a lot of stress or feel a heavy burden in preparing for entrance exams to schools . These social backgrounds may have deprived the youth of opportunities to experience various types of personal relationships and to fully develop social skills. The Japanese society has changed in many respects along with its economical growth in the 1980’s to the 1990’s. According to national surveys conducted by NHK (Nihon Hoosoo Kyoku [Japanese Broadcasting Bureau]) in 1973, 1978, 1983, and 1988 (introduced in the NHK Research Center 1993)., the preference of the Japanese regarding social life has gradually shifted from ”close/personal relationship with neighbors, relatives, and people at work” in 1973 to "superficial relationship” in 1988 (n = 3,853, age over 16) . Considering that the contemporary youngsters were born in the early 1980 ’5, their personality may be reflecting these characteristics of Japanese society. 196 It is also worth mentioning that recently people are living in a high-tech, information—oriented society, and influenced by the Japanese mass media (Sekiguchi 1999) . Today, it is very common for university students and high-school students (even junior—high school students) to use various computer-mediated communication channels such as keetai (cellular phone), PHS (Personal Handy—phone System), e—mail, and web-chat. Kageyama (1999) claims that these types of communication tools replace direct contact between people and they even help the youngsters to create a personal barrier in interpersonal interaction and to have superficial relationships. I agree with Chiba et al. ’s (1999) claim that the Japanese media such as TV, radio, comic books and magazines have a strong impact on youth language. Suzuki and Matsuda (1997: 157-158) also claim that these media influence the youngsters ’ way of thinking. Youngsters are daily exposed to the communication styles provided by those media . According to an NHK survey in 1996 (cited in Suzuki and Matsuda 1997: 158), students spend more than half of their free time using these media . These media can quickly display the latest trends in language use and emphasize the trends in order to receive a lot of attention from the audience. For example, some entertainment TV programs often use Japanese subtitles even for utterances in Japanese (a technique called teroppu in Japanese), and highlight the interesting phrases or pieces of information in bold. They also use many newly—created shortened 197 words (e.g., bura-pi for ’Brad Pitt’ and mini-suka for ’mini (short) skirt’). Through these media, the expressions of youth language can spread equally and quickly all over the country, and so their usage is adopted easily among youngsterslg. Lastly, why do female youngsters use hedges more often than male youngsters? Nowadays, it has been observed that the speech by younger female speakers is neutralized, or no longer characteristically ”feminine” sounding (e.g. Okamoto 1995; Philips 1997). However, in the present study, the data still support Lakoff’s hypothesis that women tend to use softening expressions such as hedges more often than do men. This suggests that less feminine language is used among female youngsters with respect to sentence-final forms and some pronouns (e.g. masculine words boku or ore for ’ I ' ) , but not yet with respect to expressing uncertainty, mitigating expressions, and/ or showing a non— committal attitude. I speculate that this fact has something to do with the stereotypical gender roles in Japanese society, or the image with 19 O I 0 There are Japanese contemporary dictionaries and annual encyclopedias which include a section of ”youth language” (e.g. Gendai yoogo no kiso chishiki [Basic knowledge of current terms] and Imidas). A list and explanations of ”youth language” and ”college women’s language" are also available on web sites (e.g. Web Imidas [Online] : http//imidas.shueisha.co.jp/koukoku/03. html; ”college women’s language”: http//www.osaka-gaidai.ac. jp/%7Ekoyano/joshidai93.htm). 198 which women are associated at home, school and the work place. In Japan there still are a strong stereotypical image of women as being subservient and dependent, and some social expectations that they appear or behave "feminine” and beautiful (T. Inoue 1999). According to a survey conducted in 1992 by the Bureau of Citizens and Cultural Affairs of the Tokyo Metropolitan Government, 46 percent of the Japanese informants (n = 3 , 524) agree that girls should be raised to be feminine and boys to be masculine. Western countries, on the other hand, show much lower percentages (e.g. 28 % for the U.S., 16 % for England, 6 % for Sweden, n z 1000). Although there may be slight differences in the definition of the words ”feminine" and ”masculine" across cultures, this survey illustrates that close to a half of the Japanese population still have different expectations for male and female children and adolescents. I speculate that this type of general social attitude and expectations concerning women and how to raise and treat girls promotes Japanese female adolescents to be non- assertive or soft-spoken. In general, it is true that the Japanese value harmony in relationships (e.g. Kindaichi 1957; Lebra 1976; Mushakoji 1976; Nakanishi 2000) . As a consequence, they prefer implicit communication styles. A public opinion poll conducted by the Agency of Cultural Affairs in Japan in 1999 shows that 74 percent 199 of the informants (all generations, n = 2, 196) prefer an implicit communication style (sensing what others want to say based on the context) over an explicit communication style (stating clearly what you want or mean). Therefore, vague speech style is not a privilege of only youngsters. Older people also prefer to use various types of vague expressions. However, some of the hedging expressions frequently used by the Japanese youngsters do catch people ’ s attention, and often negatively, and the language of youngsters is often criticized as being vague (Chiba et al. 1999) . According to the same survey by the Agency of Cultural Affairs in 1999, 85.9 percent of the informants think that the Japanese language now is ”corrupt”. When the informants were given several reasons from which to choose (up to three reasons) in a subsequent question, 62.5 percent of the informants chose ”youth language” as a reason of the corruption. This finding is similar to that with the questionnaire in the present study in which the participants were asked whether hedges sound corrupt. Furthermore in the survey, 42.4 percent of the informants say that they sometimes do not understand the new or trendy words, and 23.5 percent say that they do not understand what people from a different generation say. Considering that some of the hedges investigated in the present study represent the contemporary youth language and that the older speakers had a negative perception of hedges , they can form an obstacle in the communication and understanding among different generations . 200 6.4. Summary Through the questionnaire results in the present study, it was found that there were differences between the participants’ perception and the actual use of hedges, as well as differences in the opinions between the two age groups. Some of the hedging expressions used by youngsters are characterized as unpleasant and disapproved by the older groups . Although most of the younger female students in the present study were aware that they should consider the context‘when using hedges, in actualityy conscious control may not be as easy as it seems. Above I discussed possible psychological and social reasons why Japanese youngsters use some types of hedges often. In their developmental stage (psychological and physical) as adolescents, they do not yet have an established identity, which causes instability and uneasiness in human relationships. Additionally, their avoidance personality may prevent them from disclosing themselves in communication, which leads to superficial relationships. These characteristics of the contemporary adolescents may have been created or enhanced by some features of contemporary Japanese society such as the nuclear family, the pervasive mass media, the new communication channels , and the stereotypical gender image. 201 Chapter 7 Conclusion 7.1. Introduction The present chapter summarizes the findings concerning the use of Japanese hedges by younger and older speakers. The focus of the present study is largely threefold. First, the frequency of the use of Japanese hedges was quantitatively examined in relation to the three social variables (age, sex and style). Second, the use of toka ’or something’ and teyuuka ’or rather’ were qualitatively investigated in context to understand their main functions and differences among the younger and older speakers. Last, the speakers ’ metalinguistic knowledge about the use of hedges was discussed by examining the results of a questionnaire. The purpose of the study was to investigate the use of hedges by the younger (high-school students, n=20) and older speakers (speakers in their 50’s and 60’s, n=20) in Japanese spoken discourse. ‘By comparing the use of Japanese hedges between.two different age groups and sex groups in two different speech situations , I hoped to understand who speaks vaguely using hedges , and when and why hedges are used. It was also hoped that the present study would help us understand one of the on-going linguistic changes and the system of politeness in Japanese conversation. 202 7.2. Quantitative and qualitative analysis of the use of hedges In the present study, hedges are defined as expressions of uncertainty, possibility, tentativeness, and. approximation, which convey a sense of vagueness . For the quantitative analysis , I chose 21 Japanese hedges , which semantically express vagueness . They are for example phrase-/clause—final expressions such as toka ’or something, omou ’I think’, kana ’I wonder’, adverbs such as nanka ’like’, toriaezu/ichioo ’for now; tentatively’, kekkoo ’quite; a bit’, and a connective teyuuka ’or rather’. Depending on the context, some hedges function as information-oriented hedges, and some hedges function as affection-oriented hedges . The main motivations for using hedges are informational (speaker's knowledge), positive politeness (showing solidarity) , negative politeness (being indirect or avoid sounding too confident), self—protection or non—committal attitude , and/ or textual or sequential coordination and cohesion. In a social context, hedges play an important role in expressing the degree of the speaker’s commitment and in smoothing and facilitating the interaction. I posited three hypotheses regarding the influence of the social variables age, sex and style on the use of hedges . I repeat the results in the following. 203 1. Hedges are used more often by younger speakers than by older speakers. -> Supported 2. Hedges are used more often by female speakers than by male speakers. -> Supported 3. Hedges are used more often in casual speech than in formal speech. -> Not supported (Reversed) Overall, hedges were employed more often by the younger speakers than the older speakers, especially by younger female speakers (Chapter 4). Out of 21 hedges which were investigated in this study, 12 hedges showed a significant main effect of age (ANOVA) . It is noteworthy that in addition to some phrase— /clause—final hedges, the younger speakers made frequent.useeof several adverbial hedges: e.g. tabun ’probably’ , toriaezu/ichioo ’for now; tentatively’. The two most frequently used hedges were toka ’or something’ and nanka ’like’. They appeared most frequently in the younger female speakers' speech (51.59 % of all hedges), as compared to the younger male speakers (38.60 %), the older female speakers (24 . 62 %) , and the older male speakers (17 . 84 %) . The statistical analysis showed highly significant main effects of age and sex for both hedges. Overall , among the four groups, the younger female speakers employed hedges the most in both chats and interviews: there was a significant interaction between age and sex (ANOVA) . The younger male group, in contrast, did not use hedges as often as the younger female group, and showed a similar pattern as the older 204 female group in the chats. Hypothesis 3 was not supported in the present.study; The overall result showed that hedges were used more often in interviews than in chats. However, post-hoc comparisons showed that no significant differences in the frequency of hedges‘were found between the two situations for all four groups. Considering that this hypothesis was not supported in my preliminary study either, the general observation in the previous literature that hedges or vague expressions appear more often in a casual context than in a formal context (e.g., Lehrer 1975; Channell 1994) did not hold in this study; One of the possible reasons may be that the use of information—oriented hedges was facilitated in interviews . In interviews the participants mainly took an information—provider role in the questions-and-answers interaction, therefore emphasis was put on information exchange. During interviews the speakers frequently used hedges for listing items , approximations , and evidential expressions , e . g . tari/ tari suru ’do ... and such’, gurai, goro, and atari ’about; around’, omou ’I think’, and kana ’I wonder’. On the other hand, inucasual situations, more emphasis is put on emotional exchange in general (Lehrer 1975). IBy qualitatively examining the use of toka ’or something ’ and teyuuka ’or rather’ in context (Chapter 5) , it was found that more affection—oriented hedges were used in chats than in interview in the younger speakers’ speech. 205 The younger speakers used toka ’or something’ and teyuuka ’or rather’ slightly differently when they were chatting among themselves (Chapter 5). For example, they used toka for vague quotation more often in chats than in interviews. The younger speakers used toka to quote another’ s utterance (other—quotation) or their own utterance (self-quotation) , sometimes successively coordinating the text. With toka for vague quotation, they expressed playfulness, awkwardness and youthfulness, or they avoided being too serious in the social interaction. The frequent use of hedges, especially vague quotation, is probably one way to show solidarity for the younger speakers. Generally, toka ’or something’ and teyuuka ’or rather’ are multifunctional and their primary functions vary depending on the context. A major characteristic of these hedges is that they make vague references, and that they are both used to provide alternatives; toka as an inexhaustive coordinator and teyuuka as a repair marker. Toka and teyuuka imply that there are other possibilities, and by not mentioning the possibilities, they can leave the utterance vague or unassertive. This way, the speaker shows his/her non-committal or self—protective attitude to the addressee. Chapter 5 introduced the functions of toka ’or something’ : uncertainty marker, approximation, soft highlighter, and vague quotation. I also classified six major functions of teyuuka ’or rather’ : correction, rephrasing, specifying, blurring, subtopic 206 shift, and supplementary explanation. Teyuuka was often used for repair (self-repair and other-repair) . What makes teyuuka different from other repair expressions is that it does not completely reject the repaired segment, and so the repair is made in an unassertive or mitigated tone of voice. It was found in the present study that the younger speakers used teyuuka both for self-repair and other—repair, while the older speakers mainly used it for self—repair. By examining toka ’or something' and teyuuka ’or rather’ in context, interesting differences in the usage and linguistic environment appeared depending on the speaker ’ 5 age, sex and style of the conversation. Future study should expand this type of investigation to other hedges than these two. 7.3. Behind the use of hedges In Chapter 6, the participants’ perception about the use of hedges and the psychological and social background of the contemporary Japanese adolescents were discussed. Answers to the question ”Do you use the expressions [hedges]?" turned out to correspond to the result of the actual use of hedges analyzed in Chapter 4. An interesting finding is that the younger female speakers have different attitudes or opinions toward the use of hedges than do other groups . For example, they consider the use of hedges a usual way of talking, which was not the case with other groups. They also think that the use of hedges is fun, while the 207 older groups think it is corrupt or unpleasant . This may partially explain why the younger speakers used hedges more often than other groups. The choice of words reflects the communication style that the speaker wants to express in an interaction. Contemporary youngsters often exhibit an unstable or avoidance personality (Suzuki and Matsuda 1997; Shimizu 1998; Kageyama 1999) . Due to the avoidance personality, the youngsters tend to be afraid of communicating directly with others , and of disclosing themselves . Some of these features are caused by the (psychological and physical) developmental stage that adolescents are going through. According to Kageyama (1999), there are also some features in society which make youngsters vulnerable, too sensitive and protective. He explains that possible social causes for this tendency may be, for example, the increase of small-sized families and a heavy burden in preparing for entrance exams. It is also worth mentioning that people are living in a high—tech, information—oriented society, which may have promoted youngsters to have indirect communication styles and superficial relationships. The participants, the older participants in particular, found some of the hedges corrupt or unpleasant. As I briefly discussed in Chapter 2, it is often the case that when a language goes through a change, new forms or expressions are not welcomed in the beginning. Yet, despite such criticism, some of the new 208 forms remain in use and spread among Japanese speakers . The hedges investigated in the present study may be no exception. If adolescents’ speech plays a leading role in the progress of language change, as Labov claims (1994), the use of hedges by youngsters will influence other generations of Japanese. 7.4. Suggestions for future study The present study has some limitations. I will point out the major limitations in the following before closing my dissertation. In the present study, the focus was put mainly on the quantitative examination of the use of representative hedges in Japanese (Chapter 4), and only two hedges (toka ’or something’ and teyuuka ’or rather’) were qualitatively examined in context (Chapter 5) . Although the investigation of the two hedges have already revealed interesting differences in their usage in relation to social variables, the differences should be generalized by looking closely at other hedges as well. Second, the sample of participants was relatively small. Ideally, more data are necessary to make the findings stronger. Especially, it will be interesting for future study to examine hedge usage by other generations , between the younger and the older groups in the present study, to understand whether the age effect is discrete or continuous. 209 The present study dealt only with single—sex dyads. It would be interesting to know whether participants adjust or change their language when they talk to the opposite sex. Also, collecting data from other styles (e.g., narratives and phone conversation) than chats and interviews may reveal different characteristics of hedges . Since the use of hedges is also related to cognitive factors , we may be able to analyze more in depth when hedges appear by manipulating the informational demands e.g. requiring the participants to talk about difficult or unfamiliar topics. One should also take into consideration that the interviewer’s age and sex may influence the language of the interviewees. In the present study, the use of hedges by the younger female group was the most frequent of the four groups. However, if the interviewer was older and male, their language might have been more formal than it was with the interviewer in the present study. Lastly, to investigate the universality of this issue, a future study should also include cross-cultural comparisons on the use of hedges. It would be interesting to see what kind of hedges are used in other languages by younger speakers and older speakers. Despite these limitations, I hope that the present study was convincing in verifying earlier general observations about the use of hedges by Japanese younger speakers as compared to older 210 speakers , in showing influences of the speaker ’ 5 sex and the speech genre on the use of hedges, and in presenting likely explanations for when and why speakers use hedges in conversation. APPENDICES Appendix 1 Transcription methods and conventions 1. Japanese transliteration is given in the Hepburn style romanization. 2. Data, tables, and figures are numbered for each chapter. 3 . Participants are referred to by the initial of their first name in Chapter 5. 4 . Paralinguistic and other interactional symbols are as follows : / recognized pause . falling intonation with noticeable pause , continuing (’1ist’) intonation, with a slight rise or fall, followed by a short pause colon following a vowel indicates elongation ... omission = linked or continuing utterances without overlap || speech overlap () the hearer’s backchannelling (HHH) indicates various types of laughs rising intonation 00 5. Following abbreviations are used: the copula, be direct object particle genitive particle interaction particle indirect object nominalizer passive morpheme question particle Quotative particle subject particle tag question morpheme topic particle OW'UOEEI Far-300 HHUJIOIO’U HHQUW 95.... fig Appendix 2.]. Examples of hedges used in the preliminary study The participants are asked to read the following immediately before the interview (Appendix 2 .3 presents the questions asked 212 in the interview). sheet during the interview. the present investigator. The interview questions are about some hedges which are often found in speech by Japanese youngsters. These examples are provided so as to help the participants to have a better idea about the topic. orthography in Appendix 2.2. Examples (Hedges are underlined.) (1) (2) (3) (4) <5) (6) (7) (8) (9) toka ’or something’ e.g. I can’t get up early in the morning or something. nanka ’like’ e.g. I traveled various places and like I changed. mitaina ’is like’ e.g. If it were pilaf, I can cook it, and something like that. teyuuka ’or rather; or I should say’ e.g. That person is odd, or what should I say. e.g. Or rather, he is shy, isn’t he? kanji ’feel like; is like’ e.g. It’s like I come home about ten everyday. (da)shi ’and what's more; and suchlike' e.g. She is always doing part-time work, and so. -kei ’ -type; -1ike’ e.g. The guy is a salary man-type of person -teki ’-ish; -esque’ e.g. In my-ish case, I don’t feel like taking the trouble to go. ja nai desu ka ’Isn’t it?’ (about the speaker him/herself) e.g. I like coffee, right?. The participants are allowed to look at the These are constructed sentences by The same form is given in Japanese Appendix 2.2 Examples of hedges in Japanese writing used in the preliminary study (1)_flbéesrrmsorauahrhmxrmo t<fibhm hrhfithm assassxum fishfi i<fi5 1 2 3 4 5 2A.Lflflfldfl)?.4fl5€@hfi%fi.?&%LTV6WKEWiTOO%M%M BfiiTéw.(lfl2m3éflhfi%fi.fl®&~§®gfi3k) '®<&bum hxbfibhm <858ttgkhm W4E5 xmnorua 1 2 3 5 (6) EILODJVE 1 2 3 5 CT be W)fl$®Wfit 2 3 @)Wflfi®k°fii0fifififihAt 1 2 3 4 5 «tuna ( ) 23 moiauasquiausmramrvmo {r‘K fliifllb) 379i 0 (BUNCH) (9) 5153;117:1251 1 (ZiiftO4fifihfrf) (10) (Misti?) 1 (mreryyeae851 %091111 ( 5‘5 ‘5 & 5.5717211: 115/flak“? JZ< (917’) [O in 2C. 5") LT: 5 In 5 22111841315. (11) (12) %<1‘§§5 ELK/01125 ififififl‘é 1 2 531‘ L$H5¥1N0D¥§i b3} ° iii'fifflfi’riéfi‘ .....a [v ”is < 13% ’5 Imi/013?: “5 (13) .i%‘m‘<‘°'d‘mn\6 . 31’s 1 2 (14) 551901536 1 3 (15) 1515161223") - (inl'é‘tbo 1 2 (16) "ii-(Loam 1 3 (17) 1411:1311: - (”Him 1 2 (18)£tTiL/FHT=1C/m*7b’d’C 1 2 (19) iii-671K {$572110 1 2 (20) J’éfi-Iomeioaf: 1 2 (m)x20fi~30fi 1 2 3 4 3 (27) 4 01‘5”» 5 0R 1 2 3 4 5 (%D60fi40r 1 2 3 4 5 tom ( ) 4 shmhbkwmrfinmtherskmmrrm. %%mhm mum assasszum m5 t<34m6 (29) 1 2 3 4 5 %flmflififlfiflo ( ) 5 Tfimlfihfifiéfifi%é.%fififlhéthmxtm. cm %fiwfifi %thfi EBBttézhh fifiwhfi fifimhfl Wfii<fi5 £Ti<fli §9$5m®mtfi 1 2 3 4 5 @2)4u,m 1 2 3 4 3 an>hta-E%m#rsnrms 1 2 3 4 5 (M)$H°th I 5%.fih®§%&mtomr.EOHgfiéfioTwrfiflo e l—' (D nanka kekkoo tori aezu/ichioo l—‘F—‘NOOOOOOOI—‘OOHl—‘ONQ HI—‘bOOOOOOOONOOONO OONOHOOOWOOOfil—‘Obom l—‘bJNl-‘Ol—‘OIbOOl—‘OobOOUIO OOJONND-‘Ol—‘NOOHOOUIhNN C>f\)U'1C)OOC>l--‘C>Ol---"C>Ol---‘0.8 or <-0.8) are in bold. Appendix 18 Detailed PEARSON results (IF, n=10) Correlation matrix 2: Why do you use this type of expressions? (12) 0.667 (13) 0.248 0.000 (14) 0.639 0.575 0.514 (15) -0.269 —0.269 0.040 -0.599 (16) -0.090 -0.269 0.441 0.021 0.420 (17) 0.095 0.000 (){726 0.263 0.046 0.200 (18) 0.265 0.000 —0.072 —0.303 ().801 0.000 (19) 0.598 0.598 -0.312 0.183 -0.129 -0.129 (20) 0.000 0.000 —0.527 —0.407 0.571 0.190 (21) 0.000! 0.000 -0.430 —0.221 0.311 -0.311 (22) 0.237 0.237 -0.035 0.345 0.013 0.396 (11) <12) <13) <14) (15) (16) Extreme values (>0.8 or <-0.8) are in hold. (18) —0.167 (19) -0.445 0.615 (20) -0.607 0.320 0.211 (21) -0.165 0.062 -0.173 0.612 (22) -0.041 -0.l98 0.113 0.335 0.137 <17) <18) (19) (20) (21) Appendix 18 (cont’d) (11) fashion or trend (12) to show solidarity (13) easy to say (14) fun (15) to be vague (16) to evade my responsibility (17) unconsciously (18) adapting to the other party’s language (19) to soften the tone (20) influenced by people around me (21) when talking about unfamiliar topics (22) to avoid disagreement Appendix 19.1 Detailed ANOVA results for answers to ”Is there anyone around you who uses this type of expressions?” DF SS MS F P sweet (n=40) 39 50.4000 age 1 16.9000 16.9000 20.9793 .00016 sex 1 .9000 .9000] 1.1173 .29789 age x sex 1 3.6000 3.6mm 4.4689 .03917 error (age x sex) 36 29.0000 .8056 DF=degrees of freedom, SS=sum of squares, MS=mean squares Appendix 19.2 Post-hoc comparisons for answers to ”Is there anyone around you who uses this type of expressions ?” (n=40) Groups Q values IF IM 1.757 OF 11.125** 9.368** OM 5.855** 4.099* 5w270** IF IM OF OM ** Extreme significant values (p < .01) * Significant values (p < .05) 242 Appendix 20.1 Detailed ANOVA results for answers to ”What do you think of using this type of expressions in conversation: Fun?” DF SS MS F P subject (n=40) 39 61.5000 age 1 2.5000 2.5000 1.7241 .19473 sex 1 .4000 .4000 .2759 .60870 age x sex 1 6.4000 6.4000 4.4138 .04032 error (age x sex) 36 52.2000 1.4500 DF=degrees of freedom, SS=sum of squares, MS=mean squares Appendix 20.2 Post-hoc comparisons for answers to ”What do you think of using this type of expressions in conversation: Fun?” (n=40) Groups Q values IF‘ IM 2.626 OF 3.414 0.788 0M 1.838 0.788 1.576 IF IM OF OM ** Extreme significant values (p < .01) * Significant values (p < .05) APPENDIX 21 Detailed ANOVA results for answers to ”What do you think of using this type of expressions in conversation: Corrupt?” DF SS MS F P subject (n=40) 39 37.7750 age 1 .0250 .0250 .0248 .84915 sex 1 1.2250 1.2250 1.2149 .27727 age x sex 1 .2250 .2250 .2231 .64423 error (age x sex) 36 36.3000 1.0083 DF=degrees of freedom, SS=sum of squares, 243 MS=mean squares ,r a ’1. r / Appendix 22.1 Detailed PEARSON results (IF and IM, n=20) Correlation matrix 3: ”What do you think of using this type of expressions?” (31) (32) 0.000 (33) -0.345 -0.091 (34) 0.148 -0.193 -0.590 (35) -0.378 0.000 -0.304 0.483 (36) 0.087 0.101 -0.011 -0.373 -O.309 (31) (32) (33) (34) (35) (36) (31) usual way of speaking (32) fun (33) shows closeness or solidarity (34) unpleasant or inappropriate (35) corrupt (36) one should refrain from using them depending on the occasion Appendix 22.2 Detailed PEARSON results (OF and OM, n=20) Correlation matrix 3: ”What do you think of using this type of expressions?” (31) (32) —0.303 (33) -0.067 0.538 (34) —0.373 —0.032 -0.226 (35) -0.394 -0.204 —0.188 0.7991 (36) 0.415 0.571 0.419 -0.496 —0.456 (31) (32) (33) (34) (35) (36) Extreme values (>0.8 or <—0.8) are in bold. (31) (32) (33) usual way of speaking fun shows closeness or solidarity (34) unpleasant or inappropriate (35) corrupt (36) one should refrain from using them depending on the occasion 244 Appendix 23.1 Detailed ANOVA results for answers to ”Do you think that this type of expressions are used to show closeness or solidarity?" DF SS MS F P subject (n=40) 39 56.4000 age 1 4.9000 4.9000 4.3448 .04182 sex 1 10.0000 10.0000] 8.8670 .00531 age x sex 1 .9000 .9000 .7980 .61873 error (age x sex) 36 40.6000 1.1278 DF=degrees of freedom, SS=sum of squares, MS=mean squares Appendix 23.2 Post-hoc comparisons for answers to ”Do you think that this type of expressions are used to show closeness or sol idarity?” (n=40) Groups Q values IF‘ IM 2.084 OF 2.978 5.062** OM 0.893 1.191 3.871* IF IM OF OM ** Extreme significant values (p < .01) * Significant values (p < .05) BIBLIOGRAPHY [The translations of the Japanese titles are mine.] The Agency of Cultural Affairs (the Ministry of Education). 2000 (January). KOkugo ni kansuru yoron choosa [A survey of public opinion on Japanese language]. Aihara, Katsuo (ed. ) . 1975 . Nihon kokugo dai-jiten [Comprehensive Japanese dictionary]. Tokyo: Shoogakkan. Ball, Catherine and Mira Ariel. 1978. Or something, etc. In C. Ball and L. Matossian (eds. ) , Penn Review of Linguistics. Department of Linguistics, University of Pennsylvania, 35-45. Barke, J. Andrew. 2000 . The effect of age on the style of discourse among Japanese women. The proceedings from the 14th Pacific Asia conference on Language, Information and Computation. Tokyo. 23-33. Brown, Penelope. 1980. How and why are‘women more polite: some evidence from a Mayan community. In S. McConnell—Ginet, R. Boker, N.Furman (eds.), WOmen and language in literature and society. New York: Praeger. 111-139. Brown, Penelope and Stephen C. Levinson. [1978] 1987 . Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Brukman, J. C. 1972. Strategies of ritual interaction among the Koya of South India. Ph.D. Dissertationx ‘University of California, Berkeley. The Bureau of Citizens and Cultural Affairs (Tokyo Metropolitan Government). 1992 . Josee mondai ni kansuru kokusai hikaku choosa hookoku-sho [Report of a cross-cultural comparison of surveys on women’s issues]. Canary, Daniel J. and.Kathryn Dindia. 1998. Sex differenceseuui similarities in communication. Hillsdale: Lawrence Erlbaum. Channell, Joanna. 1994. vague language. Oxford: Oxford University Press. 246 Chiba, Mitsugi, Tamio Karato and Akihiro Kawauchi. 1999. Kokugo hyoogen no hizumi [Strains in Japanese expressions] . Tokyo: Koobundoo shuppan. Coates, Jennifer. 1987 . Epistemic modality and spoken discourse. Transactions of the philological Society 85, 100-131. Coates, Jennifer. 1989. Gossip revisited: Language in all-female groups. In Jennifer Coates and Deborah Cameron (eds.), Women in their speech communities. London: Longman. 94— 121. Coates , Jennifer. 1993 . Women, men and language. (second edition) . London: Longman. Crystal, D. and D. Davy. 1975. Advanced conversational English. London: Longman. Crystal, David. 1997. A dictionary of linguistics and phonetics (forth edition). Oxford: Blackwell Publishers. Degawa, Naoki. 1998. Gendai nihongo tanken [Explorations of modern Japanese language]. Tokyo: Shoogakkan. Dines, Elizabeth R. 1980. Variation in discourse - "and stuff like that". Language in Society 9, 13-31. Ferguson, G.A. and Y. Takane. 1989. Statistical analysis in psychology and education. New York: McGraw—Hill. Grice, H. P. 1975. Logic and conversation. In Cole, P. and Morgan, J.L. (eds.), Syntax and Semantics 9. New York: Academic Press. Gumperz, J. J. 1964. Linguistic and social interaction in two communities. American anthropologist 6, 135—153. Haiman, John. 1993 . Life, the universe, and human language (A brief synopsis) . Language Sciences 15, 293-322 . Halliday, M. A. K. 1973 . Explorations in the functions of language. London: Edward Arnold. Halliday, M. A. K. 1980. Modes of meaning and modes of expression: Types of grammatical structure and their determination by different semantic functions. In D. J. Allerton (ed.), Function and context in linguistic analysis. Cambridge: Cambridge University Press. 57—79. 247 Harada, Tadashi . 1990 . Seinen to seiji [Adolescents and politics] . In Toshio Hisayo (ed.), Henboo suru shakai to seinen no shinri [Changing society and psychology of adolescence]. Tokyo: Fukumura shuppan. Hayashi, Makoto. 1995. A comparative study of self-repair in English and Japanese conversation. Japanese and KOrean Linguistics 4, 77-93. Hayashi, Makoto. 1997. An exploration of sentence—final uses of the quotative particle in Japanese spoken discourse. Japanese and Korean Linguistics 6, 565—581. Heider, Eleanor Rosch. 1971. On the internal structure of perceptual and semantic categories. (unpublished paper) Psychology Department, University of California, Berkeley. Hida, Yoshifumi and.Hideko Asada. 1994. Gendai fukushi yoohoo jiten. [Dictionary of adverbial usage in contemporary Japanese]. Tokyo: Tokyodoo shuppan. Hill , Beverly, Sachiko Ide, Shoko Ikura, Akiko Kawasaki and Tsunao Ogino. 1986. Universals of linguistic politeness: Quantitative evidence from Japanese and American English. JOurnal of Pragmatics 10, 347-371. Holmes, Janet. 1987. Hedging, fencing and other conversational gambits: an analysis of gender differences in New Zealand speech. In Anne Pawls (ed.), WOmen and language in Australian and New Zealand society. Sydney: Australian Professional Publications. 59-79. Holmes, Janet. 1995. Women, men and politeness. London: Longman. Honna, Nobuyuki. 1999. Language and communication. In Imidas. Tokyo: Shuueesha. 1119—1124. Horiuchi, Katsuaki . 1999 . Wakamono yoogo [Youth language] . Gendai yoogo no kiso chishiki [Basic knowledge of current terms] . Tokyo: Jiyuukokuminsha. 1077-1080. Hyland, Ken. 1996. Writing without conviction? IHedging in science research articles. Applied Linguistics 17 (4), 433-454. Ide, Sachiko. 1982. Japanese sociolinguistics: Politeness and women’s language. Lingua 57, 357—385. 248 Ide, Sachiko. 1991. How and why'dO‘women speak more politely in Japanese? In Sachiko Ide and N. H. Mcgloin (eds. ) , Aspects of women's language. Tokyo: Kuroshio. 63-79. Imai, Akira. 1999. Kookoo juken annai [A guide for high-school entrance examinations]. Tokyo: Obunsha. Inoue, Fumio. 1986. Sociopsychological characteristics of users of ’new dialect forms’ . Journal of Pragmatics 10, 327-345. Inoue, Fumio. 1998. Nihongo wotchingu [Watching Japanese language]. Tokyo: Iwanami shoten. Inoue, Teruko. 1999. JOseegaku e no shootai [An invitation to women’s studies]. (4th edition) Tokyo: Yuuhikaku sensho. Jinnouchi, Masataka. 1997 . Wakamonogo [Youth language]. In Masao Gaga (ed. ) , Nihongo no mikata [Ways to examine Japanese]. (Aera Mook Asahi shinbun extra report & analysis special No.30). Tokyo: Asahi shinbunsha. 32-33. Jinnouchi, Masataka. 1998. Nihongo no ima [Japanese now]. Tokyo: Alc. Joos, Martin. 1967. The Five clocks. New York: Harcourt, Brace & World. Kageyama, Jinsuke. 1999..Kuukyona jiko no jidai [The age of the empty self]. Tokyo: NHK Books. Kamio, Akio. 1990. Joohoo no nawabari riron: gengo no kinooteki bunseki [Theory of territory of information: Functional analysis of language]. Tokyo: Taishuukan. Kamio, Akio. 1994. The theory of territory of information: The case of Japanese. JOurnal of Pragmatics 21: 67-100. Keefe, Rosanna and Peter Smith. 1999. vagueness: a reader. MA: The MIT Press. Kennedy, Graeme D. . 1987 . Quantification and the use of English: A.case study of one aspect of the learner’s taskx.Applied Linguistics 8/3: 264-286. Kindaichi, Haruhiko. 1957/1978. The Japanese language. Rutland: Charles E. Tuttle Company. 249 Kirshenblatt-Gimblett, Barbara and Joel Sherzer. 1976. Introduction. In Barbara Kirshenblatt—Gimblett (ed.), Speech play. Philadelphia: University of Pennsylvania Press. 1-16. Labov, William. 1972. Sociolinguisticupatterns. Philadelphia: University of Pennsylvania Press. Labov, William. 1994. Principles of linguistic'change. vol. 1. Cambridge: Blackwell Publishers. Lakoff, George. 1972 . Hedges: a study in meaning criteria and the logic of fuzzy concepts. In Chicago Linguistic Society Papers. Chicago: Chicago Linguistic Society. 183-228. Lakoff, Robin. 1972. The pragmatics of modality. In Chicago Linguistic Society Papers. Chicago: Chicago Linguistic Society. 229—246. Lakoff, Robin. 1975 . Language and women ’5 place. New York: Harper and Row. Lebra, T.S., 1976. Japanese patterns of behavior. Honolulu: East-West Center Press. Leech, G. 1983. Principles of pragmatics. London: Longman. Lehrer, Adrienne. 1975. Talking about wine. Language 51/4. 901—923. Loveday, J. Leo. 1986 . Japanese sociolinguistics : An introductory survey. JOurnal of Pragmatics 10, 287-326. Makino, Seiichi and Michio Tsutsui. 1986. A dictionary of basic Japanese grammar. Tokyo: The Japan Times. Masuda, Koo (ed.). 1992. Kenkyuusha’s new Japanese-English dictionary. Tokyo: Kenkyusha. Matsumoto, Yoshiko. 1985. A sort of speech act qualification:h1 Japanese: Chotto. JOurnal of Asian Culture IX, 143-159. Matsumura, Seita. 1999. KOokoo juken annai [A guide for high- school entrance examinations]. Tokyo: Gakken. Maynard, Senko. 1987. Japanese conversation: Self- contextualization through structure and interactional management. Norwood: Ablex Publishing Corporation. 250 14aynard, Senko. 1996. Multivoicedness in speech and thought representation: The case of self-quotation in Japanese. JOurnal of Pragmatics 25, 207-226. 14aynard, Senko. 1997. Japanese communication: Language and thought in context . Honolulu : University of Hawai ’ i Press . Iweyers, G. 1989. The pragmatics of politeness in scientific articles. Applied.Linguistic 10/1:1-35. rdiller, Barbara Diane. 1993. Sex and gender hierarchies. Cambridge: Cambridge University Press. Miller, Jim and Betty Birner. 1995. The semantics and pragmatics of and everything. JOurnal of Pragmatics 19, 205-214. Miller, Jim.and Regina Weinert. 1995. The function of LIKE in dialogue. JOurnal of Pragmatics 23, 365-393. Miura, Akira. 1974. ’To’ to ’tte’ [To and tte]..Nihongo kyooiku [Japanese language education] 24, 23-28. Mushakoji, K. 1976. The cultural premises of Japanese diplomacy. In Japan Center for International Exchange (ed. ) , The silent ‘power: Japan’s identity and world role. Tokyo: Simul Press. 35-49. Nagura, Toshie. 1997. Hesitations (discourse markers) in Japanese. Current Report on Japanese—language Education around the Globe 7. Nakanishi, Masayuki. 2000. Ningen kankee o manabu tame no 11- shoo [Eleven chapters to study human relationships] . Tokyo: Kuroshio shuppan. The NHK Research Center of public opinion. 1993 . Gendai nihon-jin no ishiki koozoo [The opinions of contemporary Japanese]. Tokyo: NHK Books. Niedzielski, Nancy A. and Dennis R. Preston. 2000. Folk Linguistics. New York: Mouton de Gruyter. Niimura, Tadakazu and Satomi Iwamura. 1998. Tsukaeru hanashi kotoba [Colloquialism you can use]. Nihongo JOurnal. (December). 26-27. Okamoto, Shigeko and Shie Sato. 1992 . Less feminine speech among young Japanese females. In Kira Hall, Mary Bucholtz, and Birch Moonwomon (eds.), Locating power: Processing of second Berkeley'women and language'conference. Berkeley: Berkeley Women and Language Group. 478-488. Okamoto, Shigeko. 1995. ”Tasteless” Japanese: less ”feminine" speech among young Japanese women. In K. Hall and M. Bucholts (eds.), Gender articulated: Language and the socially constructed self. New York: Routledge. 297-325. Peirce, C. S., 1902. vague. In J. M. Baldwin (ed.), Dictionary of Philosophy and psychology. New York: Macmillan. 748. Peng, Fei. 1997. Taishoo kenkyuu: Nihongo no tanoshimikata [Comparative study: A way of enjoying Japanese] . In Nihongo no mikata [Ways to examine Japanese]. Tokyo: Asahi shinbunsha 90-93. Philip, B. Gove (ed.). 1976. webster’s Third New'International Dictionary. MA: G.& C. Merrian Company Publishers. Philips, Mieko Kimura. 1997. Application of discourse analysis to second language teaching. CAJLE (Canadian Association of Japanese Language Education) conference. Philips, Mieko Kimura. 1998. .Discourse.markers in Japanese: connectives, fillers and interactional particles. Ph.D. dissertation, Michigan State University. Phillips, Susan U., Susan Steele and Christine Tanz. 1987. Language, gender and sex in comparative perspective. Cambridge: Cambridge University Press. Powell, Mava Jo. 1985. Purposive vagueness: An evaluation dimension of vague quantifying expressions. JOurnal of Linguistics 21, 31-50. Preisler, Bent. 1986. Linguistic sex roles in conversation. Berlin: Mouton de Gruyter. Prince, E. F., J. Frader, and C. Bosk. 1982. On hedging in physician-physician discourse. In R. J. di Pietro (ed.), Linguistics and the professions. Norwood: Ablex. Romaine, Susan. 1999. Communicating gender. Hillsdale: Lawrence Erlbaum. 252 Russell, Bertrand. 1923. Vagueness. Australian J0urnal of Philosophy and Psychology 1, 84-92. Saito, Miki. 1998. On question-like intonation. MA thesis, University of Tokyo. Satake, Hideo. 1995. Wakamono kotoba to retorikku [Youth language and.rhetoric]. Nihongogaku [Japanese language studies] 14, 53-60. Schiffrin, Deborah. 1987 . Discoursemarkers. Cambridge: Cambridge University Press. Schourup, Lawrence Clifford. 1985 . Common discourse particles in English conversation. New York: Garland.Publishing, Inc. Sekiguchi, Ichiro. 1999. Gendai nihon no media to komyunikeeshon kankyoo [Communicational environment in contemporary Japan]. Tokyo: Taishuukan shoten. Sengoku, Tamotsu. 1994. Masatsu kaihi no sedai [The conflict- avoidant generation]. Tokyo: PHP. Shibamoto, Janet. S.. 1985. Japanese women’s speech. New York: Academic Press. Shibamoto, Janet S.. 1987. The womanly woman: manipulation1of stereotypical and nonstereotypical features of Japanese female speech. In Susan U. Phillips, Susan Steele and Christine Tanz (eds.), Language, gender and sex in comparativegperspective. Cambridge: Cambridge University Press. 26-49. Shibatani, Masayoshi. 1990 . The language of Japan. Great Britain: Cambridge University Press. Shimizu, Hitoshi (ed. ) . 1999 . Gendai yoogo no kiso chishiki [Basic knowledge of current terms]. Tokyo: Jiyuukokuminsha. Shimizu, Masayuki. 1998 . Shishunki no kokoro [The pubertal mind] . (fourth edition) Tokyo: NHK Books. Shinmura, Izuru (ed.). 1998. Kojien Dictionary (fifth edition), Tokyo: Iwamura shoten. Simon, Endo Mutsuko. 1986. Supplementary grammar notes to An introduction Lgmodern Japanese. Part I. Part II. Ann Arbor: Center for Japanese Studies. 253 Skinner, William. 1993. Conjugal power in Tokugawa Japanese families : a matter of life or death. In Barbara Diane Miller (ed.), Sex and gender hierarchies. Cambridge: Cambridge University Press. 236-270. Smith, Janet S. 1992a. Women in charge: Politeness and directive in the speech of Japanese women. Language in Society 21, 51-82. Smith, Janet S. 1992b. Linguistic privilege: ”Just stating the facts” in Japanese. In Kira Hall, Mary Bucholtz, and Birch Moonwomon (eds . ) , Locating power: Proceedings of the Second Berkeley Women and Language Conference. Berkeley: Berkeley Women and Language Group. 540-548. Smith, Janet S. 1992c. The giving of evidence: A study of sentence extensions and gender in Japanese. JOurnal of Japanese linguistics 14, 67—89. Smith, Philip M. 1985. Language, the sexes, and society. Oxford: Blackwell. Sotobayashi, Daisaku. 1936/1985. Jiga to booee [Japanese translation of Self and.Protection, Anne Freud ]. Tokyo: Seeshin shoboo. Soukhanov, H. Ann (ed.). 1992. The American Heritage Dictionary of the English language (third edition). MA: Houghton Mifflin Company. Sperber, D. and D. Wilson. 1986. Relevance. Oxford: Blackwell. Suleski, Ronald and Hiroko Masada. 1982. Affective expressions in Japanese. Tokyo: The Hokuseido Press. Sunakawa, Chiho. 2000. ’Toka’ hyoogen no kinoo -bunpooka no sokumen kara [Functions of toka expressions -from the perspective of lexicalization]. Proceedings of the Conference of the Japanese Association of Sociolinguistic Sciences (March). 88-93. Suzuki, Kohei and Sei Matsuda. 1997 . Gendai seinen shinrigaku [The psychology of contemporary adolescents] . Tokyo: Yuuhikaku Books. Suzuki, Satoko. 1995. A study of sentence-final MITAINA. Journal of the Association of Teachers of Japanese 29 (2), 55-78. 254 Suzuki, Satoko. 1998. Tte and nanteu.Markers of psychological distance in Japanese1conversation. JOurnal of Pragmatics 29, 429—462. Tannen, Deborah. 1989. Talking voices: Repetition, dialogue, and imagery in conversational discourse. Cambridge: Cambridge University Press. Tannen, Deborah. 1990. You just don’t understand: Woman and men in conversation. New York: William Morrow. Tannen, Deborah. 1996. Gender & discourse. New York: Oxford University Press. Tsuji, Daisuke. 1996. The transfiguration of communication style in Japanese young moderns. Tokyo daigaku shakai joohoo kenkyuujo kiyoo [Journal , the center for social information studies, University of Tokyo] 51, 42-61. Tsuji, Daisuke. 1998. Keetai to ’toka’-ben — imadoki no wakamono no komyunikeeshion sutairu. [Cellular phones and ’toka’ language -communication style in Japanese young moderns] . Gakugoo 62. University of Japan. 7-11. Tsuji, Daisuke. 1999. Toka-ben no komyunikeeshion shinri. [Communication psychology in toka-language]. Proceedings of the Conference of the Japanese Association of Sociolinguistic Sciences(January). 19-24. Ullmann, S. 1962. Semantics. Oxford: Blackwell. Underhill, Robert. 1988. Like is, like, focus. American.Speech 63, 234-246. Yanagawa, Keiko. 1998. Choo nihongo kouza [Very(latest) Japanese course]. Tokyo: ASPECT Yonekawa, Akihiko. 1995. Wakamonogo no sekai [The world of youngsters]. Nihongogaku [Japanese language studies] 14, 102-110. Yonekawa, Akihiko. 1998. wakamonogo o kagaku suru [Scientific study of youth language]. Tokyo: Meijishoin. Yoshioka, Yasuo. 1990. Kookoo-see no kotoba no tokuchoo [Characteristics of the language of high-school students] . Nihongogaku [Japanese language studies] 12, 53-65. 255 Yoshizawa, Aya. 1998. Disukoosu maakaa toshite no ”tteyuuka" no kinoo bunseki [Analysis of functions of teyuuka as a discourse marker]. BA Thesis. Tokyo University of Foreign Studies. Waldron, Vincent R. and Lesley Di Mare. 1998. In Daniel J. Canary and Kathryn Dindia (eds . ) , Sex differences and similarities in communication. Hillsdale: Lawrence Erlbaum. Ward, Gregory and Betty Birner. 1993. The semantics and pragmatics of and everything. Journal of Pragmatics 19, 205—214. Williamson, Timothy. 1992. Vagueness and ignorance. Proceedings of the Aristotelian Society, Suppl. 66, 145-162.