臺灣閩南話與事實相關的情態詞之功能研究

全文

(1)國立台灣師範大學英語學系碩士論文 Master Thesis Graduate Institute of English National Taiwan Normal University. 臺灣閩南話與事實相關的情態詞之功能研究 Truth-related Expressions in Taiwanese Southern Min. 指導教授：張妙霞博士 Advisor: Dr. Miao-Hsia Chang 研究生：蔡婉怡 Student: Wanyi Tsai. 中華民國一○一年七月 July, 2012.

(2) 摘要. 本研究從言談及語用的觀點探討台灣閩南語與事實相關的副詞，研究根據十二個小時的自然語料，對「真 e」「真正」「有影」「誠實」「實在」「正經」，文中的研究分為三類：事實描述、表達情意及人際互動。事實描述方面，是用來確定一個事件和事實的真實性、用來和虛構的事件作為對比、或強調自己言論的真實性、去除聽者對說者描述的疑慮。表達情意方面，是用來強調說者的評價性言論的強度。雖然事實描述也會產生強調的功能，但是表達情意詞的強調，其真實性是不能由上下文推斷出來或被證實的。表情達意功能可以用來增加評價的強度，或者提高述語的程度，特別的是事實相關詞如「實在有影」也可用來表達苦惱或憤怒、甚至和特殊句型 toh 或 khah 搭配後，有表示讓步的意味。人際互動方面，可用來表達對說者言論的評價，例如：表示驚訝或者同意，和動詞「講」形成搭配字可以成為一種規避詞，說者在發表可能會接受挑戰或者較誇大的言論之前，使用的溝通技巧。臺灣閩南話的事實相關詞和英文及中文事實相關詞相似，已經發展成多於描述事實的用法，常常用來增強評價的強度，甚至事實相關詞本身獨立存在就可以代表一種評價，臺灣閩南話事實相關詞的發展和中英文不謀而合，而閩南話事實詞的表現由於構句或者詞性較少的限制，而有更多的變化值得探索。. 關鍵字: 關鍵字閩南語、事實相關、有影、實在、正經、真正、真 e、誠實. i.

(3) Abstract. This thesis explores six TREs in TSM and their various functions in TSM conversation. The discourse-pragmatic functions are categorized into three: truth-describing, expressive, and interactive. Truth-describing TREs are employed to assert an event or a fact, make a contrast with an unreal scenario, justify the speaker's claim, and negate any doubts from the listener toward the statement made earlier. In contrast to descriptive TREs' truth-reporting function, expressive TREs are used to express subjective and evaluative meanings. The truth-value of the proposition can not be verified from the discourse. They serve emphasizer, upgrading, exasperation, and concessive functions. Emphasizer TREs are used to strengthen the force of the predicate. Upgrading TREs elevate the degree of the gradable predicate. Expressive TREs can express a person's exasperation such as sit-chai u-iann. What's more, collocating with limiting marker toh and comparative khah, TREs do not boost the meaning; instead, they become a way of showing concession. Interactive TREs can serve as a responsive token or a prefacing device. A free-standing TREs plus a UFP can especially be used as a responsive token. They are used to express emotions such as surprise or unexpectedness or express agreement. When a TRE collocating with a verb of saying kong, it serves like a hedge, prefacing a statement containing counter-expectedness of the interlocutor. TREs in TSM, similar to TREs in English and Mandarin Chinese, are more than truth-related or emphasizing meanings. They themselves become an evaluation. TREs in TSM has developed with more versatility due to less syntactic constraints, which deserve further investigation. Keyword: TSM, truth, u-iann, sit-tsai, tsing-king, tsiann-sit, tsin-e, tsin-tsiann. ii.

(4) Acknowledgements I would like to express my deepest gratitude to my advisor Professor Miao-Hsia Chang whose warm encouragement and constant support have made my MA degree possible. Without her teaching, research, and training, I would not have been able to appreciate the fascinating aspects of natural conversation, especially pragmatics, discourse analysis and Taiwanese Southern Min. I also owe her a great deal for her enlightening and invaluable discussions on Saturday mornings, which she should spend with her loving family; her precious calls when I was frustrated by work and thesis-writing at the same time; her patient supervisions and revisions during the thesis writing. I am greatly indebted to my committee members: Professor Cherry Ing Li, Professor Yung-O Biq for their critical comments and insightful suggestions on the general theory, the terminology, Romanization, and the examples of the present work, which have not only helped me to improve my understanding but to clarify the issues that would otherwise have been overlooked. I would also like to take this opportunity to thank Professor Doris C.-Y. Chen, Professor Jen-I Li, Professor Joy Wu, Professor Miao-Ling Hsieh, Professor John Kwock-Ping Tse, Professor Hsueh-O Lin, Professor Shu-Kai Hsieh for instruction, precious suggestions and training during the years of my graduate study. As my work on this thesis progressed, I have been lucky to have my classmates Yi Wang, Aaron Wu, Stella Liu, Carolin Kuo, Hui-Chuan Lee, Elvis Yang, and friends Robin Hsin, Chi-Fang Lee, ZiHui Liao, Li-Ying Tseng, Moryn Mao, Ching-Ya Kung, Jane Chung, Wan-Jhen Yan, Bartleby Hsui at National Taiwan Normal University giving me concern and companionship. Finally, my biggest gratitude is dedicated to my parents and family. My parents. iii.

(5) are the source of my data recording. Without their help, the work of data collection would be a lot harder. Also, their unconditional love always backs me up and spurs me to pursue my career. Despite all the obstacles along the road, they always have faith in me. Without their support, the completion of this thesis would be impossible.. iv.

(6) Table of Contents Chinese Abstract............................................................................................................i English Abstract ..........................................................................................................ii Acknowledgements.....................................................................................................iii Table of Contents.........................................................................................................v List of Abbreviations...................................................................................................vi Transcription Conventions .......................................................................................vii List of Tables and Figures .......................................................................................viii Chapter One Introduction 1.1 Motivation and Aim……………………………………………………………….1 1.2 Data & Method……………………………………………………………………4 1.3 Organization of the Thesis………………………………………………………...6 Chapter Two Literature Review 2.0 Introduction……………………………………………………………………….7 2.1 Epistemic Stance Adverbials……………………………………………………...7 2.2 Intensifiers/Emphasizer…………………………………………………………..10 2.3 Stance/Discourse Stance…………………………………………………………13 2.4 Truth-related Expressions in English and Mandarin Chinese……………………17 2.4.1 Real; really………………………………………………………………..17 2.4.2 truly……………………………………………………………………….23 2.4.3 Zhen; zhende……………………………………………………………...26 2.4.4 Shizai……………………………………………………………………...33 2.5 Truth-related Expressions in TSM……………………………………………….38 2.6 Chapter Summary………………………………………………………………..39 Chapter Three Analysis of TREs in TSM 3.0 Introduction………………………………………………………………….…...41 3.1 Truth-describing Functions…………………………………………...………….43 3.2 Expressive Functions……………………………………………………………..68 3.3 Interactive Functions...........……………………………………………………...99 3.4 Summary and Discussion.....................................................................................113 Chapter Four Conclusion 4.1 Summary of the Analysis………………………………………………………..118 4.2 Limitation for Further Research………………………………………...………120 References.................................................................................................................121 Appendix ..................................................................................................................125 v.

(7) List of Abbreviations The abbreviations used in the glosses are as follows: ASP ASSC CL COP DM DISP NEG. aspect marker associative morpheme e classifier copula si (是) discourse marker disposal marker kah(給) negative morpheme. NOM PN RT. nominalizer e proper noun reactive token. UFP KAH TOH. utterance-final particle the morpheme kah (甲) the morpheme toh (就). vi.

(8) Transcription Conventions Transcription is based on the Du Bois et al. (1993) system. The transcription notations appearing in the examples used in this thesis are as follows: [] :. speaker overlap speaker identity. -. ,. truncated intonation unit truncated word final intonation continuing intonation. \ / _. falling pitch rising pitch level pitch. ^ ... ... ... primary accent (N) long pause medium pause short pause. = (0) @ (Hx). lengthening latching laughter exhalation. (H) (TSK) % <@ @> <Q Q> <F F>. inhalation alveolar click glottal stop laugh quality quotation quality fast tempo. <A A> <D D>. gradually increasing in speed gradually decreasing in speed. <H H> <L L> <MRC MRC> <X X> X. high pitch low pitch each word distinct and emphasized uncertain hearing indecipherable syllable. <L2 L2> <J J>. code switching from Taiwanese to Mandarin code switching from Taiwanese to Japanese vii.

(9) List of Tables and Figures Table 1. The TSM mini-corpus Table 2. The shift from manner adverb to discourse marker (Lenker 2007:94 Table1.) Table 3. Frequency of TREs in the database Table 4. Pragmatic functions of different categories of TREs.. Figure 1. The stance triangle Figure 2. Mechanism of Delexicalization (Lorenz 2002:148) Figure 3. From emphasizer to exasperation. viii.

(10) Chapter One Introduction. 1.1 Motivation and Aim In daily conversation, people make and receive comments on facts or entities. During the interaction, the speakers subjectively express their own ideas or evaluate a proposition or an entity. Therefore, the notion of subjectivity (Lyons 1994) and evaluation (Thompson and Hunston 2000) are often associated with and referred to in the study of conversation. Subjectivity is defined as “the speaker’s expression of himself or herself in the act of utterance” (Lyons 1994:13). Evaluation has operationalized as “the broad cover term for the expression of the speaker or writer’s attitudes or stance towards, viewpoints on, feelings about the entities or proposition that he or she is talking about” (Thompson and Hunston 2000:5). Both subjectivity and evaluation represent the two important elements in the definition of “stance.” Biber and Finegan (1989:92) define stance as “the lexical and grammatical expressions of attitudes, feelings, judgments, or commitment concerning the propositional content of message.” Therefore, it is clear that stance lies in the lexical and grammatical expressions. Among previous studies of stance, adverbials have proven to be a rich source of various types of epistemic, attitudinal, and style stances (Biber and Finegan 1988, Conrad and Biber 2000). Modals (Thompson and Hunston 2000) and evaluative adjectives and nouns (Hunston and Sinclair 2000) are also well documented and investigated. With regard to epistemic stance, or the degree of commitment/certainty. of. the. speaker/writer,. complement. clauses. and. complement-taking predicates (Field 1997) have frequently been the focus of research.. 1.

(11) In this study, we will examine the truth-related expressions 1 in Taiwanese Southern Min (TSM henceforth). Truth-related expressions (TREs henceforth) are used to show the speaker’s commitment and certainty of the proposition. Therefore, they belong to the epistemic stance (Biber et al. 1999). There has been a rich literature on the TREs in English (Quirk et al. 1985; Biber et al. 1999; Tagliamonte and Ito 2002; Lorenz 2002; Paradis 2003; Paradis and Bergmark 2003; McCarthy 2003; Lenker 2007) and in Mandarin Chinese (Chen 2010; Li 2003; Wu 2010; Chen 2008). However, linguists put little emphasis on TREs in TSM. Sia (2008) conducts a corpus study of the “intensity devices” in TSM to view their distribution in different genres. TREs in TSM and their uses can only be found from dictionary definitions. The previous studies seem not to include the discussion of their stance-taking and discourse-pragmatic functions. In fact, there are various ways to show the speaker’s commitment or comment on the actuality of the statement in TSM. In example (1), the TRE 有影 u-iann ‘truly, really’ takes the scope over the whole proposition and it is used to comment on the actuality of the following statement. Besides 有影 u-iann, 正經 tsing-king, 真正 tsin-tsiann, and 誠實 tsiann-sit can also be used to comment on the actuality. In example (2), 正經 tsin-king and 真正 tsin-tsiann are used to explicitly point out the truth of the state of 醉 tsui ‘being drunk’. 真實 Tsiann-sit collocating with 講 kong ‘say’ can also be used to show the speaker’s commitment to the following utterance the speaker wants to convey, as shown in example (3). The use of a TRE plus the verb of saying kong is very similar to frankly speaking or to tell the truth, which reveals the speaker’s stance.. 1. We use the term truth-related expressions because the data we examine contain different syntactic categories, including adverbials, epistemic predicates, and some truncated elements in the clause. 2.

(12) (1) A: 咱. 大家. 有. 當時仔,. lan ta-ke u we everyone have →A: 有影禮拜時仔來. tang-si-a sometimes 仁愛路. u-iann le-pai-si-a lai. jin-ai-loo. really. Sunday. [行. 袂. kiann kiann bue. come Jin-Ai Road [mm]. D:. 行. walk walk. 䆀] bai. NEG bad. ‘A: It is really not bad that we take some time to take a walk on Jin-Ai Road on Sunday. D: Yes.’ (2) →M: 咱進前. 嘛. 有. 講. 過. 你. lan tsin-tsing ma u kong kue li we before also have say PERF you 啉甲醉 e 時陣, lim kah tsui e si-tsun, drink KAH drunk ASSC time →M: 真正醉的時陣你 tsin-tsiann tsui. e. sit-tsun. li. 若. 正經. na if. tsing-king really. 就— toh. really drunk ASSC time you then M: 路邊插咧睏一醒 a.\ loo-pinn tshah le road-side insert PERF. khun tsit-tshenn-a sleep a-wake UFP. F: hm==!\ M: 啊結果. 伊%,_. ah kiat-ko DM result →F: 伊路邊. i he 真正. [插. tsin-tsiann. tshah lo-khi. i. loo-pinn. he road-side really →M: [真正真正. 落去@]@@@@. insert into 給插. tsin-tsiann kah M: 但是 Tan-si but. 落去].\. tshah lo-khi. 去. really 給. DISP insert into 插落去鐵支路.. khi go. kah DISP. tshah lo-khi insert into. thih-ki-loo railroad. ‘ M: We previously talked about that if you are drunk seriously, really drunk, stop the car at the roadside and take a nap. 3.

(13) F: Yes. M: He turned out to... F: He really [stopped at the roadside] M: [Really stopped] M: But to stopped at the side of railroad.’ (3) →B: 但是 Tan-si. 彼陣仔. 阮，誠實. hit-tsun-a. guan, tsiann-sit kong. but that time I A: hum.\ RT B: 阮毋敢執行. truly. 講. 乎,_ honnh. speak UFP. la.. guan m-kann tsip-hing la I NEG-dare enforce UFP ‘ B: But at that time I, truly speaking.. A: Yes. B: I daren’t enforce.’ The following research questions will be addressed in this study: (a) What are the discourse-pragmatic functions of the TRE tsin-e, tsin-tsiann, tsiann-sit, sit-tsai, tsing-king and u-iann in TSM? (b) How are the functions similar to and different from one other? (c) What role do the collocational and situational contexts of these TREs play in the interpretation of the meanings?. 1.2 Data & Method The study investigates spoken data. The database contains a TSM mini-corpus, which is twelve hours and fifteen minutes hours of tape-recorded spontaneous TSM. The database includes 83 episodes. The types of interaction include natural face-to-face conversation (5hr.46’), telephone conversations (3hr.11’), TV drama (39’40”), TV talk show (1hr.41’), and radio call-in (8’59”) and radio interviews (48’07”), as shown in Table 1. The TV drama is a rather spontaneous talk without 4.

(14) detailed scripts because the speakers do not follow the Mandarin written scripts word by word.. Type of interaction. Subtype. Face-to-face. TV drama. Non-face-to-face. Length. Total Length. 39’40”12. TV talk show. 1hr.41’08”00. Natural conversation. 5hr.46’24”44. Radio interview. 48’07”06. Radio call-in. 08’59”00. Telephone conversation. 8hr.55’19”62. 3hr.20’12”44. 3hr.11’13”44 12hr. 15’32”06. Table 1. The TSM mini-corpus The spoken data were recorded with cassettes or digital recorders, which were transcribed into intonation units (IU) according to the transcription systems developed by Du Bois et al. (1993). The Romanization of Taiwanese in this study generally follows Taiwanese Romanization2 system released by Ministry of Education in 2008, without marking the tone for ease of typing. Tone number will be marked when there’s an ambiguity. When necessary, some examples from 台語文語詞檢索 Tai-gu-bun Gu-su Kiam-sek (Taiwanese Concordancer)3 will be used to support our analysis. The thesis conducts a synchronic study on the spoken TREs in TSM. The approach adopted in the present study is conversation analysis (CA) (Sacks et al. 1974; Psathas 1995). CA is an empirical approach to examine the natural occurring data and conduct a bottom-up analysis. That is, no researcher’s priori assumption is involved and the result or generalization is all data-driven.. 2. 台灣閩南語羅馬字拼音方案. The Taiwanese Romanization System [ ] is released by Ministry of Education in 2008. It unifies different schools of Taiwanese Romanization system to provide teachers and students with a more unified system to teach and learn Taiwanese. See Appendix. 3 Taiwanese Concordancer (TC henceforth) is designed by Un-GiaN Iunn [ ] in 2003. http://210.240.194.97/TG/Concordance/form.asp. 楊允言. 5.

(15) 1.3 Organization of the Thesis The organization of this thesis is as follows. A brief introduction of the study is given in Chapter 1. Chapter 2 reviews previous works on the TREs and some related studies. Chapter 3 discusses the semantics and pragmatic uses of the TREs tsin-e, u-iannn, tsin-tsiann, tsiann-sit, tsing-king and sit-tsai in our corpus. A further discussion and conclusion will be given in Chapter 4.. 6.

(16) Chapter Two Literature Review. 2.0 Introduction In order to explore the truth-related expressions more comprehensively, it is necessary to give a thorough review of the previous studies. According to reference grammar (Biber et al. 1999:549), the truth-related expressions, conveying the level of certainty, belong to the expressions of epistemic stance. Most of them are adverbials. Many of the truth-related adverbials often have the intensifying uses when it is a pure adverb. Therefore, we will first review the studies of epistemic stance adverbials in Section 2.1 and intensifiers of languages in general in Section 2.2. As the use of truth-related expressions generally involves the subjective attitude and stance-taking toward the general fact, Section 2.3 reviews theories of stance in terms of the participants and message involved in stance-taking.. Section 2.4 follows with. previous studies which discuss the versatile functions of English and Chinese truth-related expressions. Due to the lack of literature on truth-related expressions in TSM, some dictionary definitions and discussion in reference grammar will be included in Section 2.5. Section 2.6 presents a summary of this chapter.. 2.1 Epistemic Stance Adverbials Truth-related expressions often appear in the form of an epistemic stance adverbial4, which has different terminologies in the previous literature (Givon 1993, Biber et al. 1999, Quirk et al. 1985). Givon (1993) terms the adverbials that convey the speaker’s attitude toward the truth, certainty or probability of the proposition as. 4. Adverbials here are a broader category, which includes one-word adverbs or phrases having the function of adverbials. 7.

(17) epistemic adverbials. He points out that epistemic adverbials are heterogeneous in form, appearing most commonly as either one-word expressions, one-word expressions derived with –ly, or prepositional phrases (Givon 1993:74) and the semantic scope of epistemic adverbials ranges over the entire proposition. Besides the use of conveying speaker’s attitude toward the proposition as Givon (1993) has pointed out, Biber et al. (1999) refer to the truth-related adverbials as epistemic stance adverbials, adding the notion of stance. The truth-related adverbials fall in one of the subtypes of Epistemic stance adverbials (Biber et al. 1999), which can be used to convey doubt and certainty (e.g. no doubt, certainly, probably), actuality and reality (e.g. in fact, really, truly), source of knowledge (e.g. apparently, evidently, according to Mr. Kandil), limitation (e.g. in most cases, typically), viewpoint or perspective (e.g. in our view, from our perspective), and imprecision (e.g. like, sort of, roughly). Biber et al. (1999) further point out that there is a frequent ambiguity--a linguistic element is a stance adverbial or a circumstantial adverbial of extent/degree. For example, the epistemic stance adverb really is particularly difficult to disambiguate. When the adverb really appears in the initial or final position, some instances seem to have an epistemic stance meaning ‘in reality’ or ‘in truth’ as shown in (1a) and (1b). When really appears medially, it can have an epistemic meaning with propositions that concern absolute characteristics such as being alive in (1c). (1) a. Really you’ve noticed the difference? b. I had no choice really. c. Was Molly ever really alive alive-oh?. (Biber et al. 1999:857-858). However, when really is in the medial position with gradable propositions, determination of the meaning can become even more difficult. In the following. 8.

(18) examples, really can either have the stance meaning or can be interpreted as intensifying the scalar property of a verb or adjective, with the approximate meaning ‘very (much).’ Even after they examine the wider context, it’s still not easy to clarify which meaning the speaker/writer intended to convey in such cases as (2a-c). Bondi (2008:33) also has a similar observation that it is difficult to “distinguish the writer’s judgment about the certainty, reliability, and limitations of the proposition from her/his attitude toward proposition’s content.” (2) a. It’s really wonderful. b. Susie’s really excited about that backyard. c. The numbers really took off in the late 1890s.. (Biber et al. 1999:858). In addition to the various syntactic positions of epistemic stance adverbials in an utterance, Biber et al. (1999:551) point out that adverbials can stand alone as structurally unconnected elements, and even as complete utterances. In some cases, the adverbials are related to elliptic forms of previous sentences. Adverbials as utterances can also serve to express or emphasize agreement and concurrence and can be used as questions. Often, these adverbials are stance adverbials of actuality or style of speaking as shown in (3).. (3) A: You can still vote if you lost it. B: Really? A: They should have your name on the roster.. (Biber et al. 1999:551). Some of the above examples illustrate that stance adverbs can serve not only to pose questions, but also answer them as indicated by (4).. 9.

(19) (4) A: It’s warm isn’t it? By the radiator. B: Probably.. (Biber et al. 1999:551). Sometimes, the interpretation of a linguistic element as a stance adverbial or as a response form is not clear-cut. For instance, really can be used as a backchannel like I see, which is stronger in indicating a high degree of interest in what the previous speaker had to say compared to the affirmative markers like mm, uh huh, and mhm.. (5) A: She’s real good, the people like her real well over at the hospital. B: Really? To summarize, epistemic stance adverbials appearing in different syntactic positions have different semantic meanings and pragmatic functions. Epistemic stance adverbials of actuality can be either an element in a clause or standing alone as utterances to convey the speaker’s agreement or interest/attention. Thus, it is interesting to explore whether TREs in TSM share similar features.. 2.2 Intensifiers/Emphasizer Truth-related expressions can be treated either as epistemic stance adverbials or extent/degree circumstance adverbials. They fall in an ambiguous situation especially when the epistemic stance adverbial is in the medial position before a scalar proposition (Biber et al. 1999). Extent/degree circumstance adverbials, as indicated by Biber et al. (1999:780) can tell the extent to which a proposition holds, answering questions such as “‘how much/many?” and “to what extent?” They can either amplify the intensity of the clause proposition, sometimes called amplifiers or intensifiers, or lower the intensity, also called diminishers. Truth-related stance adverbials usually belong to the intensifying uses since they are often used to emphasize the force of the 10.

(20) actuality of the statement. Intensifiers generally are believed to be linguistic elements which convey the degree of the quality of the item. An intensifier is defined by Quirk et al. (1985:445) as a scaling device, which co-occurs with a gradable adjective. However, intensifiers are variously referred to in literature. Stoffel (1901) uses the term “intensive adverbs;” Bolinger (1972:18) terms them “degree words” with no separation from downtoners. Quirk et al. (1985:445) distinguish two subsets of intensifiers: amplifiers, which ‘scale upwards from an assumed norm,’ such as completely and very much, and downtoners, which ‘scale downwards from an assumed norm,’ such as partly and hardly. Furthermore, amplifiers can be divided into maximizers and boosters. Maximizers ‘denote the upper extreme of a scale,’ and boosters ‘denote high degree, a high point on the scale,’ whose examples are provided in example (6). Therefore, we may conclude that intensifiers do not only refer to an increase in intensification that is expressed, but also denote that there is a scale of intensity involved. Also, Paradis (1997:12) further points out that intensifiers are indefinite and pertain to extent since they don’t have specific measurements as quantification. (6) (from Quirk et al. 1985:590-591) • Maximizers: absolutely, altogether, completely, entirely, extremely, fully, perfectly, quite, thoroughly, totally, utterly • Boosters: badly, bitterly, deeply, enormously, far, greatly, heartily, highly, intensely, much, severely, so, strongly, terribly, violently, well. The truth intensifiers have the scaling effect when they are used with gradable verbs as in (7) and adjectives as in (8).. (7) a. He really likes her. [‘He likes her very much’] b. I indeed appreciate your help. [‘I greatly appreciate your help’] 11.

(21) (8) a. It was really funny. b. He’s dangerous indeed.. (Quirk et al. 1985:586). However, sometimes the truth adverbials do not require the modified element to be gradable (Quirk et al. 1985:583). They have a reinforcing and emphatic effect with non-gradable verbs too, as shown in (9).. (9) a. He really was there. b. She indeed sat next to them.. (Quirk et al. 1985:586). Quirk et al. (1985:583) especially term the adverbials expressing the semantic role of modality as emphasizers, which have a reinforcing effect on the truth value of the clause or part of it, as shown in (10).. (10) a. He really may have injured innocent people. b. He may really have injured innocent people. c. He may have really injured innocent people.. (Quirk et al. 1985:583). Examples in (10) can be paraphrased respectively as follows:. (11) a. It is really possible that he has injured innocent people. b. It is possible that it is really true that he has injured innocent people. c. It is possible that he has really [i.e. actually, indeed, certainly] injured innocent people. c’. It is possible that he has really [i.e. seriously, to a severe extent] injured innocent people. (Quirk et al. 1985:583). It is shown that really in (11a) and (11b) is a pure emphasizer since really is used to modify the whole proposition. However, when it moves closer to the predicate, it will imply a high degree of injury as shown in (11c’) as well as the assertion of certainty 12.

(22) as in (11c). Quirk et al. (1985:583) classify two common groups of emphasizers as shown in (12). Adverbials in (12a) are employed to make a comment on what is being said as being true. Adverbials in (12b) are used to convey the speaker’s assertion, in which his or her words are the literal truth. (12) a. actually, certainly, clearly, definitely, indeed, obviously, plainly, really, surely, for certain, for sure, of course b. frankly, honestly, literally, simply, fairly <esp. BrE>, just In sum, in most cases, truth-related adverbials, when collocating with gradable verbs or adjectives, have an intensifying function. When they are in the sentence-initial position or placed between the subject and predicate, truth-related adverbials are often used to emphasize the truth value and make an assertion toward the proposition, which are termed as emphasizers by Quirk et al. (1985). From the above examples, we can see that intensification is a concept that refers to not only the expression but also an indexing of the speaker’s perspective (Athanasiadou 2007:555). It can be viewed as a kind of stance-taking. In the following section, we will review some studies related to stance.. 2.3 Stance/Discourse Stance The truth-related expressions illustrate a type of evaluation since they are used to comment on the certainty of a proposition (Givon 1993, Biber et al. 1999). As argued by Du Bois (2007), evaluation is a means to stance-taking. Berman (2004) and Scheibman (2007) also discuss the elements of stance-taking in discourse, which provides a good theoretical foundation in our investigation of stance in truth-related expressions. Du Bois (2007) proposes the notion of the “stance triangle.” There are three participants in an act of stance-taking and they interact with each other. He. 13.

(23) points out that the most salient and widely recognized one is evaluation. It is defined as “the process whereby a stance-taker orients to an object of stance and characterizes it to have some specific quality or value” (Du Bois 2007:142). In example (13), the stance predicates horrible, ideal, and nasty are used to evaluate something, i.e., the stance object. (13) a. PETE: That’s horrible. b. LANCE: (TSK) That’s ideal. c. KEVIN: @that’s @nasty.. If the subject of an utterance is a human, a stance-taker, followed by an affective predicate, the utterances are often referred to as indexing affective stance (Besnier 1993, Haviland 1991, Maynard 1993, Ochs 1996, Shoaps 2002). As shown in (14a), the speaker I positions himself in a position of an affective scale such as glad. An epistemic scale is a position that the speakers put themselves on. In (14b), the speaker presents himself as ignorant.. (14) a. LANCE: I am glad. b. DAN: …I don’t know.. From the above examples, we know that there must be one stance-taker and a stance object. The two elements combined with the role of addressee constitute the tri-act of stance triangle. The following is the definition (Du Bois 2007:163) and the stance model:. (15) Stance is a public act by a social actor, achieved dialogically through overt communicative means, of simultaneously evaluating objects, positioning subjects (self and others), and aligning with other subjects, with respect to any salient dimension of the sociocultural field. 14.

(24) Figure 1. The stance triangle. The model tries to clarify how these relations are constituted through the stance acts of evaluating objects (objective), positioning subjects (subjective), and aligning with other subjects (intersubjective).The stance triangle provides a framework for people to understand the socio-cognitive relations (objective, subjective, intersubjective) that are present in all dialogic interaction. Three elements in this model provide us with important cues to interpret the use of truth-related expressions since we are going to explore the collocational and situational uses of the expressions. It is necessary to examine from Subject1, Subject2, and Object since they are all meaningful in the interpretation of the data. Du Bois (2007) has claimed that stance is a tri-act. Berman (2004:107-108) also discusses discourse stance involving three related dimensions. First, orientation explains the relation between three elements in text production and interpretation: sender (speaker or writer), text (narration or exposition), and recipient (hearer or reader). Second, attitude is treated in three aspects: epistemic, deontic, and affective. An epistemic attitude expresses a relation between the speaker/writer and the possibility, certainty, or evidence for the speaker/writer’s belief about the truth of the proposition; a deontic attitude entails a viewpoint of judgment or evaluation; while an 15.

(25) affective attitude concerns the speaker/writer’s emotions with respect to a given state of affairs (Berman 2004:107). Berman (2004) claims that the three distinctions form a continuum—from the more objective, abstract, and universalistic epistemic attitudes via socio-culturally determined deontic attitudes, to the most subjective reactions and personal feelings. Third, generality describes how relatively general or specific reference is to people, places, and times that are pointed out in the text (Berman 2004:107). This dimension is closely dependent on the two other dimensions of discourse stance. Berman (2004) argues that ‘stance’ is an essentially relational concept, since it involves the interaction between text as discourse and sender/receiver as speaker–writer or hearer–reader. In this sense, stance can be thought of as an inclusive notion of the three. Scheibman (2007:113) proposes that stance might be classified as three aspects: individual, interactive, and sociocultural. The aspect of individual is construed as relevant to a speaker’s position in discourse. The individual stance taking involving evaluation and expressiveness can be framed as subjective. Interactive aspect is relevant to local discourse activities and sociocultural is relevant to general beliefs of people as members of communities. The interactive discourse activities and the sharing beliefs forming the collaborative expression of evaluation and attitude can be referred to as intersubjective. The truth-related expressions are used to show the speaker’s comment on the proposition, which belongs to the epistemic stance. However, the speaker may not always be the subject of the sentence. Therefore, to judge the stance, we need to have the notion of stancetaker and the proposition as a shared object that people are going to evaluate. Du Bois’s (2007) stance model and Berman’s (2004) three aspects of discourse stance provide this study with a more comprehensive system to examine the collocational uses, which will supplement our study of TREs in TSM. The three 16.

(26) aspects come into play in a conversational interaction. Different linguists have different aspects and terminologies in stance; however, they all present the core meaning of stance in that the subjective evaluation and intersubjective interaction contribute the stance. Although they all provide three elements of stance, they focus on different levels. Du Bois (2007) specifically points out the three participants in the collaboration of stance. Berman (2004) discusses stance from the orientation of the speaker, listener; the difference of the attitude; and the generality of the referents in the proposition. Scheibman (2007) argues that stance can be classified as subjective and intersubjective.. 2.4 Truth-related Expressions in English and Mandarin Chinese This section will examine previous studies on truth-related expressions in English and Mandarin Chinese. Previous literature on truth-related expressions mainly discusses the adverbial uses. Truth adverbials real/really and truly in English will be reviewed in Section 2.4.1 and Section 2.4.2 respectively. Previous studies of zhen/zhende in Mandarin Chinese will be discussed in Section 2.4.3.. 2.4.1 Real; really Real, though not having recognizable –ly suffix, also has its adverbial use. It is referred to as zero form by Tagliamonte and Ito (2002). This use is often associated with colloquial/non-standard languages (Quirk et al. 1985). On the other hand, really has received a lot of scholarly attention in various aspects. Some studies describe the various uses and functions in different syntactic positions (Quirk et al. 1985, Lorenz 2002, Paradis 2003). Some observe the use of really in different social groups (Paradis 2003, Paradis and Bergmark 2003). The availability for different positions might result from grammaticalization. Therefore, many linguists investigate the 17.

(27) phenomena of grammaticalization using really as examples (Lorenz 2002). The grammaticalizational path can be approximately from the truth reading, the intensifying use, to stancetaking element. Quirk et al. (1985) discuss the syntactic features of really. Really, like other adverbials, can appear clausal-initially, clausal-medially, and clausal-finally. Quirk et al. (1985) especially point out that it can precede a negated verb phrase as in (16a) and lie within the scope of clause negation as in (16b). The different positions yield two different interpretations. Really can also appear in a question as in (17a) and with imperatives as in (17b). (16) a. I really don’t know him. [‘The real truth is that I don’t know him,’ i.e. ‘I don’t know him at all’] b. I don’t really know him. [‘It’s not the real truth that I know him,’ i.e. ‘I don’t know him well’] (17 ) a. Do they really want him to be elected? b. Make an effort this time; but really make an effort. Unlike Quirk et al.’s (1985) listing the possible syntactic positions, Paradis (2003) discusses the semantic meanings of really and categorizes really into three different readings: truth-attesting of proposition as in (18), subjective emphasis of situation as in (19), and reinforcement of scalar property as in (20). (18) Really, they are quite strange.. [truth-attesting of proposition]. (19) I really appreciate your support. (20) They are really nice.. [subjective emphasis of situation] [reinforcement of scalar property]. The truth-attesting really takes scope over a proposition, which is used to assert something that may be true or false. Truth attesting really can appear in any adverbial positions in an utterance such as in the initial position as (21a), in the final position as (21b), and in the medial position as (21c). The role of really explicitly points out the 18.

(28) truth of a proposition, especially when the listener does not expect the fact. For example, in (21c), the listener did not expect Sue and Bill to buy the farmhouse. (21) a. Really that’s quite good. b. She loves me really. c. Sue and Bill really bought the farmhouse they had been dreaming of.. Emphasizing really, unlike the truth attesting really, can only be placed adjacent to an adjective or a predicate that denotes a situation that is attitudinal in character and expresses the speaker’s emphasis. In the three examples, emphasizing really takes scope over a situation type: “in sad” in (22a), “can’t be bothered” in (22b), and “like” in (22c). Paradis (2003) further terms really as a marker of epistemic subjectivity.. (22) a. No, that’s sad, that really is definitely. b. I meant to be going to the choir tonight but I really can’t be bothered. c. I do actually really like singing.. Degree reinforcing really enhances the degree of the scalar property denoted by the adjectives. For example, really reinforces the degree of “rudeness” in (23a), “length” in (23b), and “personality” in (23c). The gradable property of the modified element distinguishes the degree reinforcing really from the emphasizing really. Paradis (2003) further claims that degree reinforcing really can be replaced by very in all the cases. However, degree reinforcing really is not a mature degree modifier. It still takes the truth value of the proposition into consideration. Very, on the other hand, is a degree reinforcer both in affirmatives and in questions (Paradis 2003; Paradis 1997:19-21).. (23) a. Hugh is apparently really rude about everyone especially when he gets drunk. b. …will be one big nuclear war which will last really long and finally end the world. 19.

(29) c. He thinks he is really good.. What’s more, Paradis and Bergmark (2003) make a clear investigation on the abundant uses of really in the informal conversation among teenagers. They argue that the interpretation of really is pragmatically conditioned by the speaker’s wish to back up his or her expressions with judgment of truth. Although the aim of Paradis and Bergmark’s (2003) study is to explore the conceptual representations of three different interpretations of really rather than their discourse/pragmatic functions, they still provide a clear classification of the various senses of really. Especially for the third use, reinforcement of scalar property indirectly proves that really is an intensifier since it is used to push the property to the high point of the scale. The truth-attesting function corresponds to the epistemic stance use, which is used to ‘show levels of certainty or doubt’ or ‘comment on the reality or actuality of a proposition’ (Biber 1999:557). The various readings of really, according to Lenker (2007), is a process of grammaticalization. She proposes that adverbs with a basic meaning ‘truly’ undergo a recurrent and regular semantic cline from clausal-internal adverbial to sentence adverbials and discourse particle, which conforms to the Adverbial Cline Internal Adverb > Sentence Adverb > Discourse Particle suggested by Traugott (1995). Furthermore, Partington (1993:182) observes that the modal-to-intensifier shift of really is not only a diachronic but also a synchronic phenomenon. While really in (24a-24b) maintains its modal use, used across the whole scope of sentence, really is far more likely to be perceived as an intensifier as shown in (24c): (24) a. Really, I could hear her thinking. b. Aw, don’t rub it in. Ah fell awfu’. I do really. c. When the horsetail is really hot, wrap it up… 20.

(30) (All cites in Partington 1993:182). However, Partington (1993) does not provide an account for the synchronic various uses of modal really and intensifier really. Lenker (2007) uses conversational maxims and invited inference to explain the changes. We will turn to her study in the review of truly. Lorenz (2002) conducts a corpus-based study to explore the mechanisms of delexicalization in adjective intensification in the case of really. Really is the most likely candidate to follow the grammaticalization path of very and become the next fully grammaticalized adjective intensifier from the MODAL resource. Lorenz (2002:148) provides a diagram and depicts how the delexicalization works as shown in Figure 2. The three vertical arrows EVALUATIVE, COMPARATIVE, and MODAL represent three clines of delexicalization; through SEMANTIC FEATURE COPYING, they will gradually feed into the inventory of the SCALAR set. COMPARATIVE. EVALUATIVE. MODAL. comparatively. awfully. clearly. relatively. horribly. doubtless(ly). especially. disgustingly. certainly. particularly. terrifically. virtually. extraordinarily. fabulously. seemingly. SEMANTIC absolutely?. really?. terribly?. FEATURE COPYING. SCALAR very, rather, almost, … slightly, greatly, fully… Figure 2. Mechanism of Delexicalization (Lorenz 2002:148). As for the case of really, Lorenz (2002:153) observes that really is an original truth-attesting item from the MODAL resource with varying scope and focus. If really 21.

(31) is a truth emphasizer, it will be semantically identical with the replacement of other MODAL adverbs such as clearly and truly. Or if really is used as an intensifier, it can be replaced by very, or extremely in the SCALAR set (Lorenz 2002:155). The motivation for the grammaticalization of really is claimed by Lorenz (2002:153) not only due to its frequent uses but also its features in young, informal, and hence dynamic usage. In addition to truth adverbials as part of the clause, really also stands alone in a responsive turn. Really pointed out by McCarthy (2003) is the most frequent response item in the combined corpus of CANCODE5 and Northern American spoken sample of the Cambridge International Corpus. McCarthy (2003) discusses the “small” interaction response tokens such as really in everyday conversation. Really can be used in the non-minimal responses without expanded content, acting as a response token that invites continuation by the previous speaker, or indicates confirmation before the talk can continue or before the full affective reaction occurs such as “No you’re joking” in (25) (McCarthy 2003:51). (25) [social chat among students] A: Yeah. And there’s there’s a thing on, there’s a erm blackboard in the erm not blackboard sorry chalkboard. Can’t call it blackboard any more. It’s not PC. B: Really? A: Yeah. B: No you’re joking. A: No I’m not joking. It’s a chalkboard.. (McCarthy 2003:51). Diani (2008:312) claims really in the interrogatives like (25) can serve as a comment that the hearer inserts in order to show he understands and is paying attention to which is being said, as a sort of ‘backchannel’.. 5. CANCORD stands for Cambridge and Nottingham Corpus of Discourse in English, which is used as the database for British English in McCarthy’s (2003) study. 22.

(32) In summary, real/really appearing in different syntactic positions has the truth meaning in different scopes of sentence. Really mainly acts as a sentence adverb in initial position, expressing the speaker’s judgment/attitude toward the truth value of the proposition. In the medial position, really then may be interpreted as an intensifier limiting its scope to gradable adjectives or verb phrases that follows. What’s more, really is often used as a response token. Therefore, many studies are conducted to examine the development of the various meanings. Lenker (2007) argues the grammaticalization of really conforms to Adverbial Cline (Traugott 1995) and semantic-pragmatic tendencies (Traugott and Dasher 2002). Lorenz (2002) provides a delexicalization model. An application of Gricean Conversational Maxim and Inference is also applied to account for the grammaticalization (Lenker 2007). Although the truth-related expressions in TSM in this study may not entirely be adverbials, the discussion of the various meanings of really related to the syntactic positions shed some light on the analysis of various uses of truth-related expressions in TSM. The review of the grammaticalization development of really and the delexicalization model also offer an account for the dynamic versatile uses of truth-related expressions.. 2.4.2 Truly Compared to studies of really, literature on truly is rather small.. Lenker (2007). conducts an investigation of the diachronic development of truth-related adverbials in the case of truly. She examines the various functions of truth-related adverbial soþlice in Old English, soothly and forsoothe in Middle English and truly in Present-day English. It is proposed that adverbs with a basic meaning ‘truly’ undergone a development following the Adverbial Cline Internal Adverb > Sentence Adverb > Discourse Particle suggested by Traugott (1995). Lenker offers an account from a 23.

(33) pragmatic perspective and communicative principles, in particular Grice’s Maxims of Quality. Soplice in Old English and soothly in Middle English are employed in several co-existing functions. At the phrasal level, they can be used as manner adverbs or intensifiers. Combined with verbs of communication, they stress the assertion of a proposition by explicitly pointing out its truth value. At the sentence level, the adverbs are used to introduce the voice of the narrator and allow the speaker to convey his/her evaluative/subjective expressions. At the discourse level, adverbs lose almost all of their propositional meaning ‘truth’ as a discourse marker. They mark the beginnings of episodes, serving as an episode boundary marker. At the more local level, the discourse-maker function marks the end of a quotation, serving as a highlighting device. Similar functions are also found in forsooth in Middle English. Lenker (2007) points out that forsooth is mainly used as an emphasizer or intensifier and frequently used in both positive and negative answers to questions. It also functions as a discourse marker, collocating with conjunctions such and and but to indicate the end of an episode. Finally, truly in Present Day English, which was developed from treowlice in Old English, also has the similar functions. Originally, it is a manner adverb, meaning ‘faithfully, loyally,’ and by 13th century, it becomes semantically epistemic, indicating the speaker’s subjective evaluation. Finally, it can also be used as a discourse marker, collocating with conjunctions to function as a highlighting device or a marker of personal opinion. Table 2 is the summary of the recurrent shift of the adverbs discussed above.. 24.

(34) Table 2. The shift from manner adverb to discourse marker (Lenker 2007:94 Table1.) manner adjunct >. sentence adverb >. discourse marker. medial position. (commonly) initial position. (commonly) initial position. syntactically obligatory. syntactically detachable ‘I tell you truly’. syntactically detachable no/vague meaning. (lexical synonym). ‘truly speaking’ ‘It is true that’. pragmatic functions: highlighting function: episode boundary marker, markers of personal opinion. Lenker (2007) also argues that the developmental cline of the adverbs share the pragmatic-semantic tendencies proposed by Traugott and Dasher (2002:281): non-subjective content scope within proposition truth-conditional. > subjective > content-procedural. > intersubjective > procedural. > scope over proposition >. > scope over discourse non-truth conditional. In addition to the Adverbial Cline and Pragmatic-semantic tendencies, Lenker (2007) provides her own account with the application of Gricean Maxims. She claims that the development of the epistemic modal adverbials from the manner adverbs to a discourse marker all comes from the propositional meaning of the adverbs, i.e., indicating the veracity or the factuality of a proposition. According to Gricean Maxim of Quality and Maxim of Quantity (Grice 1975:26-27), people do not say what they believe to be false and do not say for which they lack adequate evidence. However, the use of these adverbs seems to flout the two maxims. Lenker (2007) proposes that the expressions denoting the veracity or factuality of a proposition are used when the speaker/writer or listener/readers have doubt toward the proposition, when the speaker/writer wants to express the commitment to the truth of the proposition for fear 25.

(35) of the non-acceptance of the interlocutors, when the speakers want to be co-operative in negative contexts, or when the speakers want to hedge a proposition such as in I’m not sure whether it is true but…. Lenker (2007) further indicates that people make sense with the truth-related adverbs indicating more than the propositional meaning by making inference from the context. On the epistemic level, the listeners conceptualize the truth-intensifying adverbs as emphasizers when the speakers want to emphasize the assertion and actuality of his/her proposition. When the epistemic meaning on the epistemic level does not make sense in the context, the listeners will then turn to the textual level, conceiving the adverbs as discourse markers. Lenker (2007) provides a pragmatic account for the fact that the propositional meaning, the truth-intensifying reading, and discourse marker of truth-related expressions co-occur in conversation. How people decide from three readings to apply in the current conversation is through application of the Gricean Conversational Maxims and Inference. Her account is quite insightful, which provides a further perspective on the semantic change in addition to a grammaticalization cline.. 2.4.3 Zhen; zhende While the most frequent use of truth adverbials in English is real/really, the truth-related adverbials in Mandarin Chinese can be zhen and zhende. Previous studies of zhen(de) investigate the polysemy and the use of zhen and zhende and the semantic changes. For example, Chen (2010) examines zhende in a diachronic approach, discussing the grammaticalization of zhende. Wu (2010) makes a clear taxonomy of different levels of zhende in the spoken data and also explores the grammaticalization of zhende. Li (2003) both synchronically describes the meaning and use of zhen and zhende and the classification of polysemous meanings and uses of zhen and zhende and diachronically examines the development of zhende. 26.

(36) Wu (2010) categorizes zhen and zhende as having descriptive, expressive, and interactive meaning. First, the descriptive meaning of zhen/zhende is to verify something as real/true, in which the zhende is the adjectival use. (26) a. 高市. 警方. 偵破. 假. 真. 車禍. gaoshi jingfang zhenpo jia chehuo PN police detect fake car-accident. 強盜. 案. zhen qiangdao an real robbery case. ‘Police of Gaoxiong detected a case that appeared to be a car accident but was actually a robbery.’ b. 我用的是 shi. 真的. 玫瑰. 花瓣. zhende. meigui. huaban. rose. petal. wo. yong. de. I. use. NM C/F real. ‘‘What I used was real rose petal.’. (Wu 2010:35). Second, the expressive meaning is the adverbial use to show the speaker/writer’s subjective assessment toward a statement. It can be used in a hypothetical clause as in (27a) or an interrogative clause, implying the speaker’s doubt of the validity of a statement as in (27b). Also, zhende can appear in an affirmative clause, in which the description is less believable or exaggerated. Therefore, zhende is used to emphasize the truth. In other cases, it is simply used whenever the speaker/writer feels the need to show his/her emphasis as in (27c). (27) a. 如果連戰. 真. ruguo lianzhan zhen if 外. PN 交. 有. 和. 墨國. 總統. 會面,. you. he. moguo zongtong huimian,. really have with PN president meet 部會通知他向新聞界. 說明. wai jiao bu hui tongzhi ta xiang xinwen jie shuoming foreign affair ministry will notify him to news filed explain ‘If Lianzhan really met the president of Mexico, the Ministry of Foreign Affair would notify him to explain to the press.’ b. 藉著嚴刑峻法, 真能改善 jiezhe yan xing jun through severe punishment strict 27. fa, law. zhen neng gaishan really can improve.

(37) 治安. 嗎?. zhian ma? security UFP? ‘Can public security really be improved through strict law and severe punishment?’ c. 我們無話不談, 真的與兄弟一般。 women wu hua bu tan, zhende yu xiongdi yiban we NEG speech NEG talk, really with brothers alike ‘We talk about everything. We’re really like brothers.’ (Wu 2010:36-39). Third, when zhende is used independently of a clause, it serves interactive functions. It can be used in a question to ask the truth value of a statement (28a), an answer to a question about the truth of a statement (28b), a confirmation about the truth of a statement (28c), or a response token signaling the receipt of information (28d).. (28) a. B: [<@她是= 藍心湄@>].\ [<@ta shi= lanxinmei@>].\ she C/F PN ‘She is Lanxinmei’ A: [她就是藍心]-[ta jiushi lanxin]-she just PN ‘She is Lanxin’ →A: ...真的真的嗎?/ ... zhende ma?/ really. UFP. ‘Really?’ A: ...這麼誇張. 嗎?/. ... zheme kuazhang ma?/ this exaggerated UFP ‘Is this really so?’ b. B: 喔! 可是我是 o! keshi wo shi. (Wu 2010:44). 覺得女生. 的. juede nusheng de 28. 沒. 什麼. 好. mei. sheme hao. 看的， kan de.

(38) UFP but. I. C/F feel. girl. NM NEG what. good. look NM. ‘But I feel that the ones for girls are not very good looking.’ →A: 真的嗎? zhende ma? really UFP ‘Really?’ A: 不會， buhui, NEG ‘No’ A: 其實. 你. 可以去看，. qishi. ni. keyi qu. kan,. actually you can go look ‘Actually, you can go have a look.’ A: 我覺得不錯， wo juede bu cuo, I feel NEG bad ‘I think they’re not bad.’ c.. B: 102. (Wu 2010:45). 公斤？. 102 gongjin? 102 Kg ‘102 Kg?’ A: 對啊。 dui a right UFP ‘Right.’ B: 真的假. 的。. zhende jia. de. really fake NM ‘Really?’ →A: 真的啊。 zhende really. a UFP. ‘Really.’ d. B: 女生. 肥胖. (Wu 2010:46) 超過. 四十歲. nuxing feipang chaoguo sishi. sui 29. 然後. 什麼，. ranhou sheme,.

(39) female fat. over. forty. year-old then. ‘Fat women over forty years old and’ B: 生過很多小孩這都是. what 高危險. 因子，. sheng guo hen duo xiaohai zhe dou shi gao weixian yinzi, give-birth EXP very many kids this all C/F high dangerous factor ‘having given birth to many kids, these are all dangerous factors.’ B: 不知道為什麼。 bu. zhidao weisheme. NEG know why ‘Nobody knows why’ →A: 真的喔， zhende o, really. UFP. ‘Really’ A: 那怎麼. 辦. na zenme ban then what do. 呢？ ne? UFP. ‘Then what can be done?’. (Wu 2010:47). Wu (2010) gives a clear classification on the polysemous meanings of zhende based on different levels of conceptualization. The three meanings of zhende concern the syntactic positions or types of the clause, i.e., affirmative or non-affirmative clauses, also the interaction with the interlocutors. Since Wu’s (2010) study focuses on the lexicalization of zhensi, he only describes the polysemous meanings with examples but he does not provide the frequency counts of each uses, which might truly reveal the synchronic uses of zhende. Li (2003) investigates zhen and zhende comprehensively, including discussing the polysemy, syntactic distribution, different uses of zhen and zhende and investigating the grammaticalizational process of zhende in particular. For zhen, she discovers that it’s unsuitable to classify zhen as an adjective and an adverb. Instead, three uses of zhen can be concluded: modal adverb (zhen1), distinctive adverb (zhen2), and state adverb (zhen3). Zhen1 is used to modify adjectives to strengthen the intensity 30.

(40) of the degree such as 姚明真高 yao-ming zhen kau ‘Ming Yao is really tall.’ The use often appears in an exclamation sentence, carrying strong subjective emotion. For example, 你刻得真好! ni ke-de zhen hau! ‘you carve it very well!’ Zhen2 is a distinctive adverb. By Zhu’s (1999) definition, a distinctive adverb is a bound form that only appears in front of nouns or nominative de. Therefore, zhen2 modifies a noun phrase to distinguish it from something fake such as 真朋友 zhen pengyou ‘real friends’. Zhen3, as a state adverb, modifies a predicate to describe the state of an event such as 他們真打起來 ta-men zhen da qi-lai ‘they really hit each other’. As for zhende, Li (2003) proposes two types: nominal (zhende1) and adverbial (zhende2). Zhende1 is the combination of zhen2 and nominative de. Because of the nominal feature, it can be used as a subject and object, the typical position for nouns. For example, zhende in 真的假不了 zhende jia bu liao ‘the real things can be not fake’ is used as a subject. Zhende in 這會兒成真的, 你就麻煩囉 zhehuier cheng zhende, ni jiu mafan luo ‘if now it turns out to be real, you will be in trouble’ is used as an object. Zhende1 is often used as the object of the verb shi to form a predication shi zhende. The subject can be a noun phrase or a clausal subject. For example, the subject in 太太的國籍是真的 tai-tai de guo-ji shi zhende ‘the woman’s nationality is real’ is a noun phrase; the subject in 舅舅發財了是真的 Jiu-jiu fa-le-cai shi zhende ‘That the uncle becomes rich is real’ is a clausal subject. Zhende2 is an adverbial, which has more flexibility of syntactic positions than zhende1. It can be put between the subject and the predicate as (29a) or in front of a clause as (29b). Or it can appear independently as a clause itself, which can be put clause-initially as (29c) and clause-finally as (29d). Or it can appear without the existence of another clause as (29e).. 31.

(41) (29) a.要是爸. 真的. 跟著. 大伯. 去了，她可怎麼辦呢？. yaoshi ba zhende genzhe dabo qu if father really follow uncle go. le, ta ke zeme ban ne ASP she can what do UFP. ‘If Father really went with Uncle, what can she do?’ b. 真的他得哭一場！ zhende ta really he. de have-to. ku cry. ‘Really, he has to cry.’ c. 真的，我真是真的 zhende wo zhenshi Really I really 熬憔悴了。. yi chang one CLF. 不忍. 看. 您. bu ren NEG bare. kan nin look you. 這麼 zheme this. ao qiaocui le pallid PERF ‘Really, I really can’t look you become this pallid.’ d. 幹咱們這行兒的，別成家，. 真的！. gan zamen zhe hang er de, bie cheng jia, zhende do our this business, NEG become family, really! ‘People who do our job don’t get married. Really!’ e. 問: 真的。真的？答: 真的 Wen: Zhende? Ask: Really?. Da: Answer:. Zhende. Really.. ‘Ask: Really? Answer: Really.’. (Li 2003:17-18). Li (2003) also points out the pragmatic functions of zhende2. In (30a), zhende2 is used to show the speaker’s assertion and subject emphasis toward the proposition. However, in (30b), the emphasizing effect scales down. Zhende2 is used to link the clause 去過北京 qu guo beijing ‘went to Beijing’ to the previous clause 說要去北京 shuo yao qu beijing ‘said he will go to Beijing’ to make an assertion.. (30) a. 我 wo I. 真的. 很. 喜歡. 這. 部. 電影！. zhende hen. xihuan zhe. bu. dianying. really. like. CLF. movie. very. this. ‘I really like this movie.’. 32. (Li 2003:21).

(42) b. 他說要. 去北京,. ta shuo yao qu he say want go. 今年他真的. 去過. beijing , jin nian ta zhende qu Beijing, this year he really go. 北京. 了。. guo beijing le PERF Beijing UFP. ‘He said he will go to Beijing. This year, he really went to Beijing.’ Chen (2010) is motivated by Li’s (2003) discussion on zhende and examines the grammaticalization process of zhende. She claims that there are three stages for zhende1 to be grammaticalized. The first stage is the appearance of the construction …shi zhende in the colloquial contexts. The second stage is that the subject of the shi zhende shifts from the nominal phrases to the predicational phrases from the Qing Dynasty. Due to the shift to a clausal subject, the subject of the predication shi zhende becomes longer. Therefore, at the end of Qing Dynasty, shi is gradually bleached, having zhende left. The adverbial zhende is actually grammaticalized and used as an individual clause in the 1920s. Zhende2 can be used in various positions, sentence-internally, sentence-initially, and even sentence-finally. It can not only used as a question in a conversation but also used in the declarative sentences to express the speaker’s assertion. The reviews on truth-related expressions in Mandarin Chinese actually are useful in the analysis of the similar expressions in TSM, since Mandarin Chinese and TSM are both dialects of Chinese. It’s also interesting to explore the dialectal differences.. 2.4.4 Shizai Shizai like zhende is also one of often-discussed truth-related expressions in Mandarin Chinese. Many of its discussions also focus on the development from a truth-attesting adverbial to an adverbial that expresses a subjective attitude. For example, Chen (2008) examines shizai from the syntactic, semantic, and pragmatic perspective. Wu (2010) also discuss the various uses of shizai in three aspects:. 33.

(43) descriptive, expressive, and interactive level. Chen (2008:25) points out that shizai in the real language data appears in three contexts: First, shizai is put before a degree adverb tai ‘too’ plus an adjective or a verb, such as 實在太緊張了 shizai tai jinzhang le ‘really too nervous.’ Second, it is put before a verbal complement, which can be either a complement of degree predicate such as 實在累壞了 shizai lei huai le ‘really exhausted’ or a complement of state 實在是搞得我沒有面子 shizai gao de wo meiyou mianzi ‘really make me lose face’. Third, shizai precedes the negator 不 bu ‘not’ such as in 實在不是個小數目 shizai bushi ge xiao shumu ‘really is not a small number’ or other negated forms. In these contexts, Chen (2008) illustrates the semantic and pragmatic connotations of shizai by claiming that sentence containing shizai involves the speaker’s subjective feelings and shizai is used to emphasize the psychological process that people try repeatedly and confirm finally. Chen (2008) only illustrates the three contexts of shizai. Wu (2010) further categorizes three types of meanings of shizai. The descriptive meaning of shizai is to describe an entity (concrete or abstract) that is “sturdy and full of content” (Wu 2010:72) as shown in (31) and a trait of person as “honest and down to earth” (Wu 2010:73) as shown in (32). (31) a.一碗. 八寶冰. yi wan babaobing one CLF shaved-ice 實在。實在. 只要. 二十五元，. 料多. 又. zhi yao ershiwu yuan liao duo you only cost twenty-five dollars content much and. shizai SHIZAI ‘One bowl of shaved ice only costs twenty-five dollars. It has much and compact content.’ b. 這句口號，絕 zhe ju. kouhao, jue. (Wu 2010:72) 不是. 空的，. 而. bushi. kong de,. er. 34.

(44) this CLF slogan absolutely not 相當實在。實在 xiangdang very. empty NM, but. shizai SHIZAI. ‘This slogan is absolutely not empty, but very substantial/practical.’ (Wu 2010:74) (32) 林炳坤就像這個菜頭一樣實在，實在 lingbingkun jiu xiang zhe ge PN 值得. caitou. yiyang shizai,. just like this CLF turnip same 大家選他當立委。. zhide dajia xuan ta dang worthy everyone elect him be. SHIZAI. liwei legislator. ‘Linbingkun is very reliable, just like this turnip is sturdy. It is worthwhile for everyone to elect him as a legislator.’ (Wu 2010:73) The expressive meaning of shizai is used to subjectively express the speaker’s attitude toward a statement as shown in (33). In this example, shizai represents the speaker’s strong assertion toward the statement the speaker has made. The scope of this use ranges from the whole proposition. (33) 這實在. 不是他所. 能. 負責的。. zhe shizai bushi ta suo neng this SHIZAI NEG he SUO can. fuze de hold-responsibility EVI. ‘This really is not something he can be responsible for.’. (Wu 2010:75). However, when shizai precedes a gradable predicate, its intensifying scope merely focuses on the quality it modifies as shown in (34), in which shizai reinforces the degree of 太辛苦 tai xinku ‘too tiring’.. (34) 中小學生 zhong xiao xue sheng. 天天. 要. 考試，實在. 太. tiantian. yao. kaoshi, shizai. tai. middle elementary school student everyday must test 辛苦。 xinku 35. SHIZAI. too.

(45) tiring ‘Middle school and elementary school students have to take tests every day. Their work is really tiring.’. (Wu 2010:77). The interactive meaning of shizai, according to Wu (2010:78) is manifested in the form of collocating with verbal verb 說 shuo ‘speak.’ For example, shuo-shizai-de ‘frankly speaking’ is used interactively to inform the listeners that some unexpected or negative statement will be made in the following as shown in (35). In example (35), shizai is used as a hedge of the following negative evaluation toward Taiwan.. (35). 說實在的，說實在的. 本人. shuoshizaide,. benren guoqu. SHUSHIZAIDE, I 因為過去台灣 yingwei guoqu because past 禁忌太多，. 過去. 並非. past 的. 的. 朋友，. bingfei Taiwan de. pengyou,. NEG 政治. friend 民主，. PN 不. ASSC 夠. Taiwan de zhengzhi bu gou PN ASSC politics NEG enough 壓力大，有政治犯，. jinji tai duo, yali da, you taboo too much, pressure big, have 而且言論受到限制。 erqie also. 台灣. yanlun shoudao speech subject-to. minzhu, democratic,. zhengzhi fan, political prisoners,. xianzhi. control.. ‘Frankly, I was not a friend of Taiwan in the past, because Taiwan was not democratic enough, had too many taboos, was with much pressure, punished political prisoners, and controlled speech.’. (Wu 2010:79). Besides the three types of meanings of shizai, Wu (2010) also points out the lexicalization of shizai plus the copula/focus marker shi to address the speaker’s intensity on his/her evaluation denoted by the following speech. The evaluation followed by shizaishi can be overtly and covertly expressed as shown in (36) and (37) respectively.. 36.

(46) (36) H: ..對不起.\ duibuqi.\ sorry ‘Sorry’ D: @@ →H: <@這實在是=@>,_ 實在 <@ zhe shizaishi=@>,_ this SHIZAISHI ‘This is really’ H: 這個=,_ zhege=,_ this ‘this’ →H: 這個節目 zhege jiemu this. 實在是. 太. 爛. 了,\. shizaishi. tai. lan. le,\. too. awful. CRS. program SHIZAISHI. ‘This program is really too awful.’ (37) →A: 老師. 也. 在. 玩. (Wu 2010:87). 實在是實在. laoshi ye zai wan shizaishi teacher also DU play SHIZAISHI ‘The teacher is also playing (computer games). (You are) really (misbehaving).’ A: 那還說我， na hai shui wo that still say I ‘then how can you still say that I’ A: 有問題， you have. wen-ti problem. ‘have a problem’. (Wu 2010:90). Chen (2008) and Wu (2010) provide insightful analysis of the truth-relation expressions shizai in Mandarin Chinese. Since we will examine the TSM equivalent sit-tsai, the previous studies offer useful classification and discussion on the interpretation of the semantic meanings and pragmatic functions of sit-tsai.. 37.