Epistemic stance taking in Chinese media discourse.

(1)

理论语言学研究第3 卷 2009 年 12 月 ISSN 1882-4226 (p) 1882-1545 (e)

1

Epistemic Stance Taking in Chinese Media Discourse

Chia-Ling Hsieh

National Taiwan Normal University

Abstract: This study inspects how Chinese epistemic modality is responsive to the participant

stance and communicative intention of the press. Results indicate predominant presence of epistemic adverbs in local news as compared with business and politics news. They are also more favored in reflective comments and quoted statements than factual descriptions. However, occurring preferences vary between epistemic subclasses. Speculative outnumbers assertive across different subject matters. Speculative also features a greater frequency than assertive as journalists narrate, comment and quote. These distributional tendencies suggest a stronger sense of stance marking carried by assertive than speculative. This in turn reflects a heavier responsibility journalists take as delivering knowledge with a higher level of commitment. It is concluded that journalists make a strategic choice of epistemic markers to attain distinct forces of stance taking. The findings substantiate the role of semantic constructs as an account for cognitive pragmatics.

Keywords: epistemic; modality; stance; news; Chinese

1. Epistemic Modality as a Stance Taking Device

Modality refers to meaning conveyed by linguistic devices that specify personal stance. Its identification allows propositions to be characterized as assertions or denials of possibility, impossibility, contingency, or necessity other than simple declarations of the fact. Of the various definitions proposed for modality, one most commonly adopted is given by Lyons (1977:452), who interprets modality as the speaker’s ‘opinion or attitude towards the proposition that the sentence expresses or the situation that the proposition describes’.1 Modality can be expressed by various linguistic resources, ranging from prosodic, lexical, and syntactic components to morphological affixes and particles that constitute an independent modal paradigm. Typological approaches also recognize this domain as legitimate being a cross-language grammatical category, similar to tense

1 _{Modality has also been understood by a number of scholars in terms of possible worlds. For example,} Kiefer (1994:2514) defines it as ‘the relativization of the validity of sentence meanings to a set of possible worlds’. See also Perkins (1983) for a construal of modality by different settings of possible worlds.

(2)

2

and aspect (e.g. Bybee and Fleischman, 1995).2 Of the typological subclasses developed, epistemic modality connotes speaker certainty on the truth of a proposition; deontic modality connotes speaker requirement for the realization of a proposition (see Lyons, 1977; Palmer, 2001; Hsieh, 2006b).3

Both within and across languages, modal expressions present noticeable variation in semantic and syntactic properties, pragmatic conditions on actual use, historical stages of grammaticalization, and interplay with other grammatical categories (Chung and Timberlake, 1985; Chafe, 1986; DeCarrico, 1986; Bybee et al., 1994). Research in western languages has given weight to theoretical dimensions on the nature of modal logic and the configuration of its corresponding representations in the formal system (e.g. Leech, 1987; Kratzer 1991; Warner, 1993; Myhill, 1995; Hoye, 1997; Krug, 2000; Barbiers et al., 2002; Facchinetti et al., 2003).4 Investigations into its pragmatic roles have taken on importance more recently, particularly in the English language over the past decade. Linguists’ primary concerns have been twofold. Some deal with the general pragmatic functions English modal auxiliaries may achieve (e.g. Klinge, 1993; Hinkel, 1995; Papafragou, 2000; Nuyts, 2001a; Verstraete, 2001). Some focus on manipulation of modal expressions under a specific context (e.g. He, 1993;Klinge, 1995; Simon-Vandenbergen, 1997; Turnbull and Saxton, 1997).

Although attracting much of western linguists’ attention, modality has been studied in a comparatively less comprehensive manner in Chinese linguistics. Up to now special stress has been laid on discovering what the very scope of Chinese modality is and how this universal conceptual category is realized by the specific syntactic constructions available to the Chinese language (e.g. Tsang, 1981; Li and Thompson, 1983; Tiee, 1985; Sanders, 1992; Lin and Tang, 1995; Tang, 2000; Hsieh, 2005, 2006a, 2006b). The pragmatic effects of Chinese modal elements are only touched upon in a small number of recent studies (e.g. Guo, 1995; Kuo, 1997; Hsieh, 2008). It could be said that Chinese modality is not yet clearly defined within a unified functional framework. The time is therefore ripe for revisiting Chinese modality on ground of past theoretical approaches and up-to-date functional perspectives.

The objective of this paper is to explore factors that motivate the use of Chinese epistemic modality by contextualizing it within newspaper reportage. Newspaper reports make up one type of media discourse that centers its communicative purpose on conveying to readers what the mass media knows or believes to be true or will come true in the read world. This can be realized by

2_{Many languages use more than one grammatical category to express modal notions. Predominant are} modal verbs, verbal affixations, and particles (Palmer, 2001).

3 _{Both terminology and taxonomy vary across studies. A distinction is also made between}

propositional modality, which expresses judgment or evidence for the factual status of the proposition (i.e. epistemic and evidential modalities), and event modality, which expresses conditioning factors that are either external or internal to the relevant individual (i.e. deontic and dynamic modalities) (Palmer, 2001).

4_{For research that addresses the issue of categorical status in the syntactic level, see McCawley} (1975).

(3)

3 journalists’ indicating the strength of certainty towards what is reported. The newspaper genre thus exhibits intensive operation of epistemic modality. The present study will look into how journalists make epistemic choices to insinuate their stance towards the truth value of the reported event. Our analysis will be based on Chinese newspaper reports drawn from China Times (中国时报), an electronic news database well-known in Taiwan.

The following discussions are divided into five parts. Section 2 outlines the formal qualities of Chinese epistemic modality and the discoursal characteristics of news reportage as suggested in previous literature. Our analysis on the linguistic level brings out semantic subdomains that will be construed as taking different effects on the pragmatic level. The next section describes how the data under consideration were obtained and classified. In section 4, distributions of Chinese epistemic modals across different news types are analyzed and quantitative comparisons are presented. Section 5 provides an explanation by relating reporters’ selection of epistemic subdomains to the degree of stance involvement. The final section recapitulates the essential points and concludes the paper with the role of stance marking in the research of modality.

2. Linguistic and Contextual Parameters

The current research concerns itself with two levels of factors: epistemic modality on the linguistic level and news discourse on the contextual level. This section will elaborate on a number of recent linguistic theories relevant to these two fields of study. The scope of linguistic elements to be discussed in the remainder of this paper will also be made explicit. These general backgrounds are directed towards paving the way to establish a pragmatic link between the choice of epistemic marking and the stance of news reporting.

2.1. The Inventory of Chinese Epistemic Adverbs

People frequently make assertions built upon partial knowledge or unreliable sources of information. All languages provide inventories of expressions to encode speaker commitment towards such inconclusive assertions. The notion of epistemic modality, generally treated in the literature as a subdomain of propositional modality, belongs to a type of assessment towards the reliability of the proposition in an utterance (see Palmer, 2001). Slightly different semantic scopes have been formulated to match the corresponding grammatical categories in typologically divergent languages. One conflict that remains unresolved is how epistemic modality should be depicted in relation to other modalities (De Haan, 1999; Dendale and Tasmowski, 2001). The number and scope of epistemic subdomains thus vary with the width of notion they take in.

(4)

4

evaluation or justification for the credibility of propositional truth (cf. Chafe, 1986; Palmer, 2001). In this sense, its propositional frame includes indication of judgment or evidence about the factual status of a given utterance. Judgmental modality denotes the degree of confidence in the reality of a proposition, ranging from weak possibility (e.g. He may be a lawyer) to strong necessity (e.g. He must be a lawyer). Bracketed under judgmental modality are three subtypes of stance coding: speculative, one encoding a state of doubt (e.g. He may be a lawyer); deductive, one indicating an inference from other observable information (e.g. He must be a lawyer); and assumptive (e.g. He will be a lawyer), one expressing a reasonable conclusion drawn from what is generally known (Palmer, 2001:24-25). Evidential modality, on the other hand, denotes the source of evidence for the reality of a proposition—whether the information is secondhand (e.g. It is said that he is a lawyer) or based on firsthand sensory experiences (e.g. I saw him studying law) (Saeed, 1997:131-133). Such references of evidence are often grouped into either quotative or sensory.5 The term ‘epistemic modality’ is also used in a narrow sense, referring only to judgmental modality (e.g. Palmer 2001). This model copes better with languages that lack a grammaticalized system for evidential modality, Chinese being a case in point (see Hsieh, 2008).6 Therefore, this paper will focus its attention on judgmental modality, and epistemic modality is henceforth used in its narrow sense.

The number of subsystems under the epistemic system also varies from language to language. In Chinese, epistemic modality is verbalized by a stock of lexical forms, including auxiliary, verbal, adverbial, and particle elements. Among them the manifestation of epistemic auxiliaries has been addressed most extensively (e.g. Tsang, 1981; Li and Thompson, 1983; Sun, 1996). This may be attributed to the influence of English linguistics, which has devoted more exhaustive treatment to modal auxiliaries than to the other modal expressions (e.g. Palmer, 1990; Facchinetti et al., 2003). In Chinese, epistemic modal auxiliaries having been widely accepted include keneng (可能) ‘may’, yinggai (应该) ‘should’, gai (该) ‘should’, hui (会) ‘will’, and yao (要) ‘be going to’ (cf. Tsang, 1981; Tiee, 1985; Tang, 2000; Hsieh, 2006a).7 One of their shared syntactic features is that they

5_{The term quotative as used by Plungian (2001) is also called reported (e.g. Aikhenvald, 2003),}

reportative (e.g. De Haan, 1999), and hearsay (e.g. Hobbs, 2003). A few languages further subdivide quotative according to the accuracy of the reported information (e.g. hearsay or direct quotation). Sensory also falls into different grammatical systems depending on what sense (e.g. sight or hearing) the sensory experience was based on (Aikhenvald, 2003).

6_{Evidential modality has been extensively studied in languages around the world for its function to} code the information source behind assertions (e.g. Chafe and Nichols, 1986; Plungian, 2001; Aikhenvald, 2003). It however remains unstudied in Chinese linguistics because in this language evidentiality is not grammaticalized into auxiliary verbs as the other types of modality (e.g. epistemic, deontic, and dynamic modalities) do (Hsieh, 2008:206).

7 _{Details differ among accounts as to the formal characterization of Chinese auxiliaries. There has been} a long controversy over whether this category exists and how best it should be delimited (Lu, 1979:41; McCawley, 1992:212-216; Sun, 1996:286). For attempts to clarify this issue, see Li and Thompson (1983), Fu and Zhou (1991), and Hu and Fan (1995:248-263).

(5)

5 always precede the main verb in a sentence, (1) provided below as an illustration.8

(1) Xuqiu keneng hui jianshao. Demand may will decrease ‘The demand may decrease.’ 需求可能会增加。

Epistemic modality can also be communicated in Chinese via lexical verbs, illustrated by cai (猜) ‘guess’, xiang (想) ‘think’, pa (怕) ‘be afraid’, xiangxin (相信) ‘believe’, huaiyi (怀疑) ‘believe’, and queding (确定) ‘be certain’ (Hsieh, 2006a), one given in (2) as an example. Another way Chinese manifests epistemic modality is via attachment of sentence-final particles (Tang, 2000), such as ba (吧) ‘probably’ and de (的) ‘indeed’, as shown by (3).9_{Given the absence of such a}

formal category in English, here Chinese particles are paraphrased into English modal adverbs closest in meaning.

(2) Changshang cai xuqiu hui jianshao. Manufacturer guess demand will decrease

‘The manufacturer guessed that the demand would decrease.’ 厂商猜需求会减少。

(3) Xuqiu hui jianshao ba. Demand will decrease Part. ‘The demand will probably decrease.’ 需求会减少吧。

Compared with other epistemic markers, Chinese features a more abundant repertoire of epistemic adverbs. They are used in various senses that as a whole cover a wider variety of epistemic meaning. Within this lexical category, a finer distinction between two subclasses appears to be justified. One is known as speculative, including adverbs that signal a lower degree of confidence towards a factual claim, viz., what Lyons (1977) terms as possibility. Another type, referred to as assertive, conveys stronger commitment to the factual status of a propositional content, which is interpreted as necessity by Lyons (1977). A contrast between the above two can be seen in (4) and

8 _{Throughout this paper, epistemic markers under discussion are underlined in examples. Abbreviations} used include Cl. (classifiers), Part. (particles), Poss. (possessive markers), and Pass. (passivization markers).

9_{Chinese particles also mark aspect such as le (了) and ne (呢) and interrogation such as ma (吗) and}

ne (呢). For more details, see Zhu (1982:207-214), Wang (1987:300-318), Li et al. (1990:260-266), Lu (1992:261-284), and Chu (1998:120-187).

(6)

6

(5). A third type of epistemic modality is realis (labeled alethic in Hsieh (2006a)), which indicates a claim as actually the case, viz. expressing emphatic affirmation of propositional truth (cf. Givón, 1984). Examples are dique (的确), zhende (真的), zhuoshi (着实), queshi (确实), dangzhen (当真), weishi (委实), and chengran (诚然), all used in a sense of ‘indeed’ or ‘really’, as shown by (6). This type of modality is excluded from consideration here because it bears a stronger similarity to evaluative modality, which some linguists identify as an independent subdomain that concerns itself with a known fact (cf. Facchinetti et al., 2003; Hsieh, 2006a).

(4) Xuqiu yexu hui jianshao. Demand perhaps will decrease ‘The demand will perhaps decrease.’ 需求也许会减少。

(5) Xuqiu biding hui jianshao. Demand certainly will decrease ‘The demand will certainly decrease.’ 需求必定会减少。

(6) Xuqiu dique hui jianshao. Demand indeed will decrease ‘The demand will indeed decrease.’ 需求的确会减少。

This paper will adopt the semantic model with a dichotomy between speculative and assertive. Our attention will be restricted to markers falling in the category of adverbs. One reason for doing so is that Chinese displays a rich inventory of such lexical items, yet their pragmatic forces have been largely ignored in the literature. Markers of this type also converge on the semantic property that the sentence where they appear is bound to express the assessment of the speaking or writing person. The assessment they communicate thus reflects journalists’ stance within the setting of newspaper reports. Confining our target of examination to epistemic adverbs thus yields a lexical family homogeneous in both semantic and syntactic selections.

(7) and (8) provide lists of Chinese epistemic adverbs to be investigated in the sections to follow. Though not meant to be exhaustive, (7) to (8) cover essentially all the epistemic modal adverbs of frequent or common emergence in Chinese newspaper reports in Taiwan. (7) provides speculative adverbs taken to mean a spectrum of possibility as denoted by English adverbs such as ‘maybe’, ‘perhaps’, ‘possibly’, ‘probably’, ‘likely’, ‘hopefully’, ‘presumably’, and ‘not necessarily’. Assertive adverbs in (8) carry a sense of necessity more or less equivalent to ‘surely’, ‘certainly’, ‘absolutely’, ‘definitely’, ‘inevitably’, ‘undoubtedly’, and ‘unquestionably’.

(7)

7 (7) Speculative:

bujiande (不见得), bujinzan (不尽然), buzhi (不致), buzhiyu (不至于), dagai (大概), gaobuhao (搞不好), huoxu (或许), kewang (可望), kong (恐), kongpa (恐怕), nanbao (难保), shuobuding (说不定), weibi (未必), yexu (也许)

(8) Assertive:

bacheng (八成), bi (必), biding (必定), biran (必然), bumian (不免), dangran (当然),

duding (笃定), jue (绝), juedui (绝对), kending (肯定), mianbuliao (免不了), nanmian (难免), shibi (势必), tieding (铁定), wuyi (无疑), xiangbi (想必), yiding (一定), zhuding (注定), zhun (准)

The epistemic adverbs as listed in (7) and (8) deserve additional explication. In addition to (7), Chinese possesses another subgroup that contains both speculative and interrogative components in their semantic content, such as ganqing (敢情), mofei (莫非), nandao (难道), nanbucheng (难不成), and qi (岂). These adverbs are used in the sense of ‘is it possible that’ or ‘could it be the case that’, as indicated by (9). For the sake of their interrogative nature, such speculative adverbs are not subsumed into the present investigation. As to (8), Hsieh (2006a:55) identifies a larger number of assertive adverbs than what we have enumerated here. Those absent from (8) are mostly literary language used in ancient China and are preserved today chiefly for highly formal occasions. Accordingly, they are not of common use in the contemporary Chinese language in Taiwan. Such assertive adverbs can be exemplified by juezan (绝然), dingzan (定然), quezan (确然), ding (定), dang (当), duan (断), liang (谅), and wen (稳), all analogous in meaning with ‘certainly’ or ‘absolutely’, as shown by (10). Such rarely used items are also removed from our list of modal adverbs.

(9) Xuqiu nanbucheng hui jianshao? Demand possible will decrease ‘Is it possible that the demand will decrease?’ 需求难不成会减少？

(10) Xuqiu dingzan hui jianshao. Demand absolutely will decrease ‘The demand will absolutely decrease.’ 需求定然会减少。

Moreover, (7) and (8) are not entirely satisfactory in that some of them seem to be verb-like, such as kong (恐), kongpa (恐怕), nanbao (难保), nanmian (难免), bujiande (不见得), gaobuhao

(8)

8

(搞不好), shuobuding (说不定), and mianbuliao (免不了). As pointed out by Hsieh (2002), a lack of morphological and syntactic distinctions as well as an intensive interaction among multifaceted grammatical factors may lead to overlap and indeterminacy across categorical boundaries. This can be illustrated by the vague and problematic discrimination between modal verbs and modal adverbs in Chinese (Hsieh, 2002). Modal markers such as kong (恐), kongpa (恐怕), nanbao (难保), and nanmian (难免) show verbal color in that they contain verbs kong (恐) ‘to dread’, pa (怕) ‘to fear’, ba (保) ‘to ensure’, and mian (免) ‘to avoid’. Difficulties also arise in the classification of bujiande (不见得), gaobuhao (搞不好), shuobuding (说不定), and mianbuliao (免不了) because the negative affix bu (不) ‘not’ seem to suggest a intrinsic connection to the verbal quality.

To deal with cases like the above, this study follows the traditional formal approach taken by Chinese linguists (e.g. Lin and Tang 1995; Tang 2000), categorizing the lexical status of modal elements on ground of their distributional characteristics. Unlike ordinary verbs, kong (恐), kongpa (恐怕), nanbao (难保), and nanmian (难免) cannot be negated by bu (不) ‘not’ or mei (没) ‘not’. Neither can they form positive-negative questions such as *kongpa-bu-kongpa (*恐怕不恐怕) or stand alone as an answer to yes-no questions.10_{As for bujiande (不见得), gaobuhao (搞不好),}

shuobuding (说不定), and mianbuliao (免不了), they do not have the positive counterparts *jiande (见得), *gaohao (搞好), *shuoding (说定), and *mianliao (免了).11_{For these reasons, markers}

mentioned above are all grouped into epistemic modals that take the form of adverbs rather than verbs. Therefore, only epistemic adverbs of general use in Taiwan newspapers are included in (7) and (8) as our target of research.

2.2. Stance Taking in News Reportage

In the area of stance taking in news reportage, topics of interest include authority and authenticity in language use. Raymond (2000) examines the authorized use of language on ground of ten hours of ‘unfolding live news coverage’, where events are observed and reported by reporters and cameramen in a helicopter. The news covering team in the helicopter instantaneously interacts with the anchor in the studio. This differs from traditional ‘live and on location’ news, where reporters simply read out descriptions written beforehand. Raymond (2000) indicates that ‘unfolding live news coverage’ is featured by the reporter’s authority coming from the first-hand information claimed. Audiences with no access to the real event thus follow what the reporter chooses to present through the camera lenses. The anchormen in the studio tend to avoid commenting straightforwardly

10_{For the formal features of Chinese verbs, see Chao (1968), Tang (1989), Chang (1994), Hu and Fan} (1995), Liu et al. (1996).

11_{Though Gaohao (搞好) and shuoding (说定) can act as verbs to mean ‘to have something well done’} and ‘to make an agreement/appointment’ respectively, their negative counterparts gaobuhao (搞不好) and shuobuding (说不定) do not have epistemic readings.

(9)

9 on what they see on the screen. Instead, they authorize the reporter by issuing inferential-like questions or using evidential markers such as ‘I saw’, ‘it seems’, and ‘it looks like’. The reporter then offers correct answers with affirmative declarative sentences that claim direct access to the same events. These are social conventions that regulate the interaction and turn-taking between the anchorman and the reporter in live news broadcasts (Drew and Heritage, 1992).

For the concept of authenticity, Leeuwen (2001) sketches a broader view with reference to how truthful, sincere, and trustworthy the examined target is perceived. Rather than being definite and stable, the norms for evaluating authenticity are socially negotiable, particularly in the discourse of mediated contexts. Leeuwen (2001) also sets new principles for being authentic in the media. Spontaneous utterances and invited guests’ talks about their life and work, for instance, are considered authentic. However, spontaneous as its performance is, speech that sounds mannered or recitative cannot be authentic. Montgomery (2001) approaches the issue of authenticity by observing broadcasting talk. He points out that the broadcaster often endeavors to create naturally occurring conversation and occasionally tries to interact with listeners by using seemingly intimate expressions. A trisection for the type of authentic talk is proposed: (1) The talk does not sound simulated or contrived; (2) the talk reflects the speaker’s real experiences; (3) the talk is true to the speaker himself or herself (Montgomery, 2001:403-404).

On the assumption that what shows on TV is likely to correspond to what happens in the real world, Scannell (2001) contends that TV allows viewers to experience events by creating possibilities of observation or participation. These events enter into views’ experiential world as they watch them on TV. The claim that experiences belong to individuals and are subjective should thus be modified. Scannell (2001) also argues that traditional TV news used to present authenticity by showing its seriousness and significance to approach news values. However, the way news values are constructed has changed—reported events now tend to be processed to appear interesting and extraordinary as if they are factual and momentous. For example, Coupland (2001) shows how a UK’s news reviewing program violates the traditional stance of news reporting. This program is featured by being individualistic, selective, and quotative in evaluating or responding to newspapers reports. It is inferential in that the quoted news is ‘extrapolated’ from an event to a broader social or political issue (Coupland, 2001:436). Local or ordinary values and conventions are highlighted in this program, as revealed by the informal broadcasting setting, with the studio, cameras, and audiences accompanying the host’s extemporaneous comments. These comments are uttered to seek to obtain the audiences’ support and thus reflect common people’s opinion or attitude. Coupland (2001) concludes that news reviewing programs spoil the truth and reality traditional TV news aims to deliver. This phenomenon can be explained in terms of the notion of ‘reauthentication’ triggered by ‘authenticity-from-below’ (Coupland, 2001:421).

(10)

10

media discourse (e.g. Anderson et al., 1994; Iedema et al., 1994; McQuail, 1994; White, 2004). It is commonly believed that the mass media undertakes the responsibility to offer righteous information with as little journalists’ bias as possible (White, 2004). Subjective opinions are expected to appear only in the editorial section. More often than not, however, news reports are to a large extent conjectural and explanatory, mingled with personal feelings and perceptions (see van Dijk, 1988; Anderson et al., 1994; McQuail, 1994). They constitute a ‘constructed reality’ (Anderson et al., 1994:51). Objectivity is therefore conceived of as ‘myth’ in the journalistic world (Anderson et al., 1994:47). Iedema et al. (1994) further argues that, since each journalist observes, weighs, and depicts the reality from his or her own perspective, the journalistic discourse is bound to involve subjectivity. It turns out that subjectivity in media texts should be assessed with respect to journalists’ confidence towards the match between their language and the reality. A linguistic device that introduces explicit subjectivity to discourse is modality. It is a language system that adds varying degrees of confidence or suspicion towards factual declaration, allowing indeterminacy between the absolute positive and negative extremes (Iedema et al., 1994).

In conclusion, previous research approaches the concept of stance taking in news reportage from different angles, including how the media exercises its authority and how the media and its viewers or readers interpret the level of authenticity and subjectivity. It can be concluded that news reporting is unlikely to be free from personal evaluation. Sometimes it even involves insidious and pervasive stance indicators, such as modal markers, as active manifestation of personal commitment to the actuality of reported events. In our later discussions, these factors will be shown as relevant to the use and choice of epistemic modality in Chinese newspaper reports.

3. Data Collection and Categorization

The data under our scrutiny were drawn from the China Times Electronic Newspaper (中时电 子报), one of the most popular electronic news databases in Taiwan.12_{With a wide variety of news}

and advertising service, the China Times has achieved a daily circulation of over one million, establishing itself as one of the most prominent and well-developed newspapers on the Taiwan island.13 After being published on the Internet in 1994, the China Times Electronic Newspaper has been acting as the leading newspaper that supplies digital news to readers of the Chinese language around the world. Its database preserves news stories issued on the China Times and its sister papers, the Commercial Times (工商时报) and the China Times Express (中时晚报), covering topics about

12_{See http://news.chinatimes.com.}

13_{A survey conducted in 2004 by the Institute of Public Opinion of Shih Hsin University shows that the}

China Times won the highest level of readership in Taiwan. Next came the Liberty Times (自由时报) and the United Daily News (联合报) on the market shares. For more information, see ABRS 8 (Asian Business Readership Survey) and ATMS (Asian Target Market Survey).

(11)

11 politics, economics, society, culture, life, arts, and literature. Featured on the website of the China Times Electronic Newspaper is an in-site search engine that allows readers to trace news articles during specific time periods. It also allows readers to define single or multiple Chinese and English keywords to view the related full texts of news.

The author of the present study used the in-site search engine provided by the China Times Electronic Newspaper website to search news reports with the keywords listed in (7) and (8) above. The reports where a given marker failed to meet the epistemic meaning were sifted out. They mainly included occurrences of a graphic form identical to those in (7) and (8) but with a reading deviant from epistemic modality. This linguistic filter ruled out, for example, reports where polysemous markers such as dagai (大概) and kending (肯定) functioned as adjectives to mean ‘approximate’ and ‘positive’ instead of as an epistemic adverb to mean ‘perhaps’ and ‘surely’ respectively. Also obviated from our dataset were reports containing markers not used in an epistemic sense, such as bacheng (八成) that served as a noun to mean ‘eighty percent’ and zhun (准) that served as a verb to mean ‘to permit’.

After the researcher completed the above linguistic filtration, the resulting news dataset subsequently underwent a contextual screening. This process was conducted by five research assistants, all undergraduate and graduate students with little or no training in linguistics. The contextual screening procedure involved data selection and categorization. It was performed following the twofold criteria below. One concerns the subject type, and the other, the function type of the news reports.

For the subject type, three primary subject matters were taken into treatment: business news, politics news, and local news. Reports on other subject matters such as weather, sports, leisure, entertainment, and health care were excluded from further inspection. So were editorials and literary creations in special columns, which were not concerned with a recently changed situation or a recent event and were not news-oriented. Business news were news that touched upon economic growth, fluctuation, or setback, as well as domestic and overseas financial and monetary affairs such as import and export status, investment tendencies, stock development, technological and industrial manufacturing, etc. (11) illustrates how the epistemic adverb kewang (可望) ‘hopefully’ is used in such context. Politics news reported on national or international political figures, political parties, political situations, government policies, and diplomatic performance. This can be exemplified by (12), where epistemic adverb bujiande (不见得) ‘not necessarily’ occurs in political reports. A last subject type taken into consideration was local news, which dealt with incidents happening around average people in Taiwan. Most of these incidents were newsworthy local activities, regional growth, accidents, crimes, and natural disasters, as exemplified by the occurrence of yiding (一定) ‘certainly’ (13).

(12)

12

(11) Qunian huoli kewang dadao yibailiushiyi yuan. last year profit hopefully reach 16 billion dollar ‘The profit of last year can hopefully reach 16 billion dollars.’ 去年获利可望达到一百六十亿元。

(12) Xian zai xuan qian qiyong de misi, minzhong bujiande anxin Lost in election before operate Poss. fantasy the public not necessarily secure qanen.

grateful

‘Lost in the “fantasy” of starting operation before the election, the public do not necessarily feel secure and grateful.’

陷在选前启用的「迷思」，民众不见得安心、感恩。

(13) Yi ta guoqude piqi gen gexing, yiding hui buji according to his past temper and personality certainly will regardless houquo chongdao meimei de banshang chuli.

consequence rush to sister Poss. class handle

‘Judged from his old temper and personality, he would certainly rush to his sister’s class to handle that regardless of any consequences.’

以他过去的脾气跟个性，一定会不计后果冲到妹妹的班上处理。

Reports that satisfied the above classifying criteria on subject type made up the effective samples for this study. Inside each effective sample, the sentence containing the epistemic marker in question was further categorized according to its communicative function into narrations, quotations, or reflections. Narrations were depictions of a reported event, including the relevant people, things, states, actions, places, time points or periods, etc. Examples of narrations are provided in (14) by the use of biding (必定) ‘certainly’. Additionally, during the course of describing an event, journalists sometimes quoted from interviews with famous or common people. These utterances were categorized in terms of their communicative function as quotations, as can be seen by the occurrence of yexu (也许) ‘perhaps’ in (15). A third component in a typical news report was journalists’ remarks or comments about the reported event. They were classified by communicative function into reflections. In the majority of cases, reflections appeared after narrations. There also existed exceptional cases where reflections appeared before or during narrations. An instance of reflections is given in (16), where kongpa (恐怕) ‘probably’ entered the scene.

(13)

13 (14) Jin Zhengri duoguo zhe chang danan hou, biding hui jinxing da

zhengsu.

Zhengri Jin avert this Cl. disaster after certainly will conduct huge purge ‘After averting this disaster, Zhengri Jin will certainly conduct a huge purge.’

金正日躲过这场大难后，必定会进行大整肃。

(15) Zhang Xiaoyan shuo, dalu zhe ci shifangchu zuihou dixian, yexu Xiaoyan Zhang say Mainland China this Cl. lift final Limit perhaps shi tingdao tade gongkai huyu.

be heard his public appeal

‘Xiaoyan Zhang said that Mainland China lifted the final limit this time perhaps because they heard his public appeal.’

章孝严说，大陆这次释放出最后底限，也许是听到他的公开呼吁。

(16) Ruquo zhongyou chichi bu neng minyinghua, weilai kongpa hui luode if CPC late not can privatized future probably will result in dakui shouchang.

great debt end

‘If Chinese Petroleum Corporation cannot be privatized any sooner, it will probably end up with great debt in the future.’

如果中油迟迟不能民营化，未来恐怕会落得大亏收场。

Each of the five assistants sifted through the collected data following the foregoing screening and categorization procedure. Every effective sample identified was located and coded for its subject and function type. Convergence among the assistants’ coding results reached 90%. Those diverging analyses were discussed and re-analyzed. Finally, the assistants arrived at consensus both on the scope of effective samples and on the assignment of each sample to one and only one subject and function type. These effective samples (from this point on referred to as samples for short) amounted to 3274 pieces of news reports in number, constituting the sample set for our statistical analysis. After the sample categorization, each epistemic marker was calculated for its occurring frequency in different subject and function types. The two subclasses belonging to the epistemic domain (i.e. speculative vs. assertive modality under judgmental modality) were then contrasted in respect to their occurring frequencies. The results of quantitative comparison are presented and interpreted in the subsequent sections.

4. Statistical Results

(14)

14

light distributional correspondence and discrepancy among different epistemic subdomains across different subject and function types identified in the preceding section. The next section will demonstrate these distributional patterns by empirical linguistic data detected in the sample set, then suggesting that these occurring tendencies are indicative of the varied strengths of stance marking embedded in the subclasses of epistemic modality.

Figure 1 presents the distributional proportions of epistemic subclasses in the sample set classified by the type of subject matters. In this figure, the labels on the horizontal axis specify the two semantic subclasses under analysis: speculative and assertive modal adverbs. The vertical axis represents the percentage of the occurrence of each epistemic subclass in the business news, politics news, and local news of the sample set.

As an example, of the 3274 pieces of news reports that constitute our entire sample set, 515 pieces involve the occurrence of speculative modal adverbs in business news, 501 pieces involve their occurrence in politics news, and 964 pieces involve their occurrence in local news. Therefore, within the total 3274 news samples observed in this study, the occurrence of speculative adverbs represents 15.73% (i.e. 515 out of 3274) in business news, 15.30% (i.e. 501 out of 3274) in politics news, and 29.44% (i.e. 964 out of 3274) in local news. The corresponding values of these percentages are plotted along the vertical axis of Figure 1. Likewise, the statistical values for the other epistemic subclass, i.e. assertive adverbs, are attained via the same calculation method.

Figure 1: Occurring frequencies of epistemic markers across subject types

Figure 2 to 4 below display the occurring frequencies of every epistemic adverb, grouped by type of subject matters into business news (see Figure 2), politics news (see Figure 3), and local news (see Figure 4). Inside each of these figures, the horizontal axis indicates the epistemic adverbs listed previously in (7) to (8), divided here into a distributional comparison between speculative and assertive modalities. The vertical axis refers to the percentage of the occurrence of a given adverb in

0% 5% 10% 15% 20% 25% 30% Speculative Assertive Business news Politics news Local news

(15)

15 the sample set. Adverbs belonging to the same epistemic subclass are sorted on the horizontal axis by percentage of occurrence in descending sequence, from left to right, numbered 1, 2, 3, 4, etc. accordingly. The epistemic adverb that each Arabic numeral stands for in the figure and the exact value of its corresponding distributional frequency are provided in Appendix 1 at the end of this paper, where adverbs are sorted from top to bottom in descending order according to their occurring frequencies.

The numerical data of the speculative adverb kewang (可望) ‘hopefully’ in Figure 2 to 4 and Appendix 1 are described below as an illustration of our statistical model. Of the 3274 pieces of news reports that make up the entire sample set, 210 pieces are attested with the use of kewang (可 望) in business news, 25 pieces with its use in politics news, and 80 pieces with its use in local news. That is, of all the news samples observed in this study, the occurrence of kewang (可望) covers 6.41% (i.e. 210 out of 3274) in business news, 0.76% (i.e. 25 out of 3274) in politics news, and 2.44% (i.e. 80 out of 3274) in local news. These percentages are presented in the columns headed by ‘Speculative’ and to the right of the item kewang in Appendix 1. Their corresponding numerical values are plotted along the vertical axis of Figure 2, 3, and 4 respectively. Because the occurring frequency of kewang (可望) (6.41%) is rated highest among all the speculative adverbs in business news, it is placed in Figure 2 to the leftmost position on the line assigned the label ‘Speculative’. Immediately following kewang (可望) are huoxu (或许) ‘perhaps’ and kong (恐) ‘probably’, which are the speculative adverbs of the second (1.47%) and the third (1.41%) highest frequency in business news. Along the same lines, ranking ninth in politics news (0.76%) and sixth in locals news (2.44%), kewang (可望) is put at the corresponding ninth and sixth position on the lines representing speculative frequencies in Figure 3 and Figure 4. The statistical data that show the presence of the remaining epistemic adverbs in the three subject types are obtained according to the above calculating mechanism.

Figure 2: Occurring frequencies of epistemic markers in business news

0% 1% 2% 3% 4% 5% 6% 7% 8% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Speculative Assertive

(16)

16

Figure 3: Occurring frequencies of epistemic markers in politics news

Figure 4: Occurring frequencies of epistemic markers in local news

Following an identical statistical model adopted in Figure 1, Figure 5 below contrasts the distributional proportions of the two epistemic subclasses categorized by the type of communicative functions. Parallel to Figure 1, the two subclasses of Chinese epistemic modals are marked in Figure 5 along the horizontal axis; the vertical axis delineates the frequency of each epistemic occurrence in narrations, quotations, and reflections among the whole sample set. For instance, of the 3274 news stories, 589 involve the use of speculative adverbs in narrations, 780 in quotations, and 964 in reflections. Consequently, of all the news stories, the use of speculative adverbs represents 17.99% (i.e. 589 out of 3274) in narrations, 23.82% (i.e. 780 out of 3274) in quotations, and 18.66% (i.e. 611 out of 3274) in reflections. The corresponding percentages are plotted along the vertical axis in Figure 5. The statistics for the assertive subclass are calculated in the same manner.

0% 1% 2% 3% 4% 5% 6% 7% 8% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Speculative Assertive 0% 1% 2% 3% 4% 5% 6% 7% 8% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Speculative Assertive

(17)

17 Figure 5: Occurring frequencies of epistemic markers across function types

The next three figures compare marker distributions in different function types: narrations in Figure 5, quotations in Figure 6, and reflections in Figure 7. The horizontal axis is labeled with epistemic adverbs grouped by semantic subclass, arranged from left to right following decreasing frequencies. The vertical axis plots occurring frequencies. Kewang (可望) ‘hopefully’ in Figure 6 is again given to illustrate our statistical scheme. It accounts for 4.49% (147/3274) in narrations, 2.60% (85/3274) in quotations, and 2.54% (83/3274) in reflections, the total sample set as the denominator. These values are indicated in Appendix 2, and plotted along the vertical axis in Figure 6 to 8. With the highest frequency in narrations, kewang (可望) stands leftmost on the line labeled ‘Speculative’ in Figure 6, followed by the second highest huoxu (或许) ‘perhaps’ (3.24%) and the third highest buzhi (不致) ‘not necessarily’ (2.23%). The remaining statistical data are derived along the same line.

Figure 6: Occurring frequencies of epistemic markers in narrations

0% 5% 10% 15% 20% 25% 30% Speculative Assertive Narration Quotation Reflection 0% 1% 2% 3% 4% 5% 6% 7% 8% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Speculative Assertive

(18)

18

Figure 7: Occurring frequencies of epistemic markers in quotations

Figure 8: Occurring frequencies of epistemic markers in reflections 5. Epistemic Qualification as Stance Marking

This section will clarify the distributional patterns presented in the preceding section and attempt to provide an explanation from a pragmatic and cognitive point of view. In particular, the preference for different types of epistemic qualification will be attributed to the varied strengths of journalists’ intention to mark personal stance. The following discussions will start with Section 5.1, a factual description on the occurring tendencies revealed by Figure 1 to 8. Subsequently, 5.2 relates the conceptual domain of stance to the semantic component of epistemic modality. Finally, 5.3 accounts for the epistemic distributions outlined in 5.1 with journalist stance taking as a factor.

0% 1% 2% 3% 4% 5% 6% 7% 8% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Speculative Assertive 0% 1% 2% 3% 4% 5% 6% 7% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Speculative Assertive

(19)

19 5.1. Distribution of Epistemic Qualification

As can be seen from Figure 1, the two epistemic subsystems display different ranges of distribution across the three subject types of news reports. The results of their distributional comparison are summarized in (17), where the symbol > is a shorthand expression for ‘attested with a greater amount of occurrence for the given subclass of epistemic modality labeled to the right’. (17) brings out a number of important facts. First, it indicates local news as the type of subject matters where Chinese epistemic modal adverbs are most densely distributed. The same point holds true for each subclass within epistemic modality. (17) also shows a distributional resemblance between speculative and assertive modality with respect to the type of news subject that ranks second and third, i.e. business news and politics news. As compared to local news, which exhibits a considerable greater amount of epistemic use, each of the two epistemic subclasses is almost equally distributed in business and politics news. This can be seen from their similar percentages specified in (17).

(17) Distributional tendencies of subject types across epistemic subclasses: Speculative: local news > business news > politics news

(29.44%) (15.73%) (15.30%) Assertive: local news > politics news > business news

(18.54%) (11.51%) (9.47%)

Another conclusion can be drawn from Figure 2 to 4. As summarized in (18), speculative adverbs all occur approximately one-third more often than assertive adverbs within all the three types of news subjects. That is, a noticeably greater amount of speculative adverbs appear in all types of news subjects in comparison with assertive adverbs. It can therefore be inferred that the occurring frequencies of epistemic adverbs appear to correlate significantly with the type of epistemic subclasses.

(18) Distributional tendencies of epistemic subclasses across subject types: Business news: speculative > assertive

(15.73%) (9.47%) Politics news: speculative > assertive

(15.30%) (11.51%) Local news: speculative > assertive

(20)

20

Figure 5 also leads to a number of important generalizations. As can be observed from (19), epistemic modal markers are most underrepresented in narrations. Moreover, epistemic subsystems occur with different frequencies across the three function types in news reports. These subsystems consistently appear less often when journalists narrate the reported event than when they reflect on it. Likewise, narrations are generally attested with fewer epistemic occurrences than quotations, which holds true for both speculative and assertive adverbs.

(19) Distributional tendencies of function types across epistemic subclasses: Speculative: quotation > reflection > narration

(23.82%) (18.66%) (17.99%) Assertive: reflection > quotation > narration

(14.32%) (14.14%) (11.06%)

Another generalization can be made from observation on Figure 6 to 8. As indicated by (20), approximately an additional one-third amount of speculative adverbs are attested in each of the three function types than assertive adverbs. Accordingly, the statistical results in (20) manifest a strong association of the distribution of epistemic markers with the subclasses they belong to.

(20) Distributional tendencies of epistemic subclasses across function types: Narration: speculative > assertive

(17.99%) (11.06%) Quotation: speculative > assertive

(23.82%) (14.14%) Reflection: speculative > assertive

(18.66%) (14.32%)

5.2. Speaker/Writer Stance in Epistemic Modality

In the subsequent discussions we will build a relationship between the use of epistemic assessment and the communicative function of newspaper reportage along the dimension of stance taking. Before that, the conceptual domain of stance in relation to modality is worth explication. The communication of personal stance in language has been the focus of much theoretical interest in the study of modal logic. It has often taken the form of a dichotomy between subjectivity and objectivity. For a more expanded conceptualization advanced by Halliday (1970) and Foley and Van Valin (1984), the term modal subjectivity refers to the expression of self, concerned with speaker/writer involvement. In this sense, epistemic judgment on the part of the speaker/writer of an utterance is

(21)

21 subjective as such because it involves speaker/writer stance, as demonstrated by the epistemic auxiliary may (e.g. He may be a lawyer), epistemic adverb possibly (e.g. He is possibly a lawyer), and epistemic adjective possible (e.g. It is possible that he is a lawyer). Judgment of this kind stands in opposition with objective modality that deals with the propositional agent in an utterance, concerned particularly with its capability, as denoted by the auxiliary can (e.g. The lawyer can speak German), and volition, as denoted by the auxiliary will (e.g. The lawyer will help the defendant). The psychological reality of the contrast between subjective and objective modalities can be supported by some grammatical characteristics in English (Verstraete, 2001). For example, subjective modal markers cannot be used in interrogations, conditionals, and constructions in certain tense. They are also located in positions farther from the main verb than objective modal markers do (Verstraete, 2001).

Another distinction is drawn by Lyons (1977:797) between subjective and objective epistemic modalities in light of speaker/writer commitment. He argues that, for an evaluation on the credibility of information, the less reliable the basis of evaluation is, the more likely the evaluation is perceived as subjective. Nuyts (2001b), on the other hand, interprets subjectivity as an epistemic qualification (or in his terminology, ‘evidential qualification’) that functions independently of epistemic modal markers. Nuyts (2001b) offers corpus data to verify that, in the case of English modal adverbs and modal auxiliaries, the values of subjectivity are in effect governed by contextual factors, while in the case of English modal adjectives, the values of subjectivity rest on structural factors. He further attributes the traditional subjectivity/objectivity opposition to the accessibility of evidence behind evaluation. In other words, the strength of stance taking depends on whether the evidence is to the knowledge of the speaking person alone or is shared by both interacting sides.

We are therefore led to the conclusion that Chinese epistemic adverbs under consideration in this study are of the type of explicit attitudinal language that involves speaker/writer stance. On this account, they infuse into news reports a sense of journalist stance. The journalist stance will be understood as how journalistic authors conceive of the influence of their language. What is at issue here is the status of epistemic modality and its subclasses in relation to the degree of journalists’ commitment, as well as whether this degree is responsive to the kind of news subjects and news functions where the commitment occurs. In what follows, these considerations will be shown to be contributive to a deeper understanding of stance involved in Chinese epistemic modality.

5.3. Interplay between Linguistic and Contextual Factors

With the aim towards informing the populace of current facts in the world, the tone of newspaper reporting is expected to stay objective. Since this feature of objectivity contradicts with the subjective nature of epistemic elements, we are justified to assume that the use of epistemic

(22)

22

modal markers should be avoided generally in news reports. However, journalists are also aware of the fact that their reporting is based on their own observation, which, as noted earlier, inevitably involves subjectivity. Therefore, when using stance taking markers such as epistemic adverbs, journalists tend to strike a balance between honesty (i.e. subjectivity) and justice (i.e. objectivity) (see Hsieh, 2008). In what follows, it will be suggested that the attempt to achieve such a balance between linguistic and contextual concerns will offer an answer to the pattern of Chinese epistemic class and its subclasses as clustered in newspaper reports.

The extent to which journalists reveal their stance may be controlled by the nature of their reports (Hsieh, 2008). For example, business news reports tend to be more objective because they are often substantiated with empirical numerical data. Another case of a higher level of objectivity is politics news reports. They are concerned with important public figures and affairs that require a greater sense of seriousness. A comparatively lower sense of objectivity can be attested in local news reports, where ordinary people and local events that do not receive much attention from the society are reported. Alternatively, the function of the news content also affects the degree to which journalists specify their stance. Narrations are objective in that, by definition, they refer to the part of news reporting where journalists present the reported event. Compared with narrations, quotations and reflections are intrinsically more subjective. Quotations are statements that represent viewpoints from interviewees, who may probably not adhere to or be aware of the media morality to stay as objective as possible. Reflections are also subjective since they are journalists’ personal comments on the reported event. Accordingly, Hsieh (2008) sets two types of reported matters apart, one inherently more objectivity-oriented, as shown by business and politics news, and the other less objectivity-oriented, as shown by local news. Another distinction can be drawn between narrations, which are objective in essence, and quotations and reflections, which manifest a higher level of stance taking than factual narrations.

Such disparity in terms of objectivity along subject types can be attested in the use of epistemic adverbs in our dataset. As already observed from (17) in the former section, epistemic adverbs are most widely distributed in local news, where the maxim of objective reporting is least strictly obeyed. In addition, as shown by (18), the speculative subclass outnumbers the assertive across all subject types. This suggests evaluation with lower certainty as more acceptable in the setting of news reporting, which in turn implies that, the weaker the journalists’ commitment is, the weaker sense of personal stance it conveys. To illustrate, (21) to (26) are samples taken from our dataset, all of which involve the use of speculative markers. (21) and (22) are taken from business news, (23) and (24) from politics news, and (25) and (26) from local news. The use of assertive markers are illustrated by (27) to (32), where (27) and (28) are extracts from business news, (29) and (30) from politics news, and (31) and (32) from local news.

(23)

23 (21) Bolanhui bujiande shi defang fazhan qiji, fanershi duyao.

exposition not necessarily be regional development opportunity but poison ‘An exposition is not necessarily an opportunity for regional development but a poison.’ 博览会不见得是地方发展契机，反而是毒药。

(22) Shangchang shang shi “jinghe guangxi”, yexu jintian business world on is competitive-cooperative relationship perhaps today hezuo, mingtian jingzheng.

cooperate tomorrow compete

‘The business world is of a “competitive-cooperative relationship”; perhaps people cooperate today and compete tomorrow.’

商场上是「竞合关系」，也许今天合作，明天竞争。

(23) Taiwan nuxing xianzaide diwei shifou reng yu nanxing you mingxiande Taiwan women current status whether still and men exist obvious chaju kong youyiwen.

gap probably in question

‘Whether there is still an obvious gap between women and men in Taiwan concerning their current social status is probably in question.’

台湾女性现在的地位是否仍与男性有明显的差距恐有疑问。

(24) Huoxu minshi gaoceng keyi yinyong zhege lizi wei benshen bianhu. perhaps FTV high executive can cite this example to oneself defend ‘Perhaps the high executives of Formosa Television can cite this example to defend themselves.’

或许民视高层可以引用这个例子为本身辩护。

(25) Yuanshan diqu de jiaotong pingjing jiang kewang huode jietao. Yuanshan area Poss. traffic jams will hopefully get resolve ‘The traffic jams in Yuanshan area will hopefully get resolved.’

圆山地区的交通瓶颈将可望获得解套。

(26) Diannao xiangqi chedi dabai rennao dagai keyi zai erlingyiling computer check completely beat human brain perhaps can in 2010 nian qianhou shixian.

year around come true

‘That computers completely beat human brain in playing check can perhaps come true around 2010.’

(24)

24

(27) Fubang jituan yiding hui ke shou dangchu dingding de heyue Fubon group certainly will strictly obey originally sign Poss. contract yiwu.

oblication

‘Fubon Group will certainly strictly obey the obligations originally signed in the contract.’ 富邦集团一定会恪守当初订定的合约义务。

(28) Zai chenggong yanlan Yidian jiaru zhenying huo, Weiruang wuyi in successfully recruit EA join league after Microsoft undoubtedly woyou gengduo de jingzheng benqian.

hold more Poss. competition capital

‘After successfully recruiting EA in its league, Microsoft undoubtedly held more capital to compete.’

在成功延揽艺电加入阵营后，微软无疑握有更多的竞争本钱。

(29) Weilai zhongda renshi Chen Shuibian yizhi biran chongfen future important personnel matter Shuibian Chen will inevitably fully zhanxian.

exert

‘Shuibian Chen will inevitably fully exert his will towards the future important personnel matters.’

未来重大人事陈水扁意志必然充分展现。

(30) Chen Shuibian zongtong lianren hou zhongguo dangju biding hui Shuibian Chen President re-elected after Mainland China authority certainly will you yiduan shiijian de guancha ji.

have some time Poss. observation period

‘After President Shuibian Chen is re-elected, the authorities of Mainland China will certainly observe him for some period of time.’

陈水扁总统连任后，中国当局必定会有一段时间的观察期。

(31) Ruguo shouji haishi danyin lingsheng de ren, na tieding bu If cell phone still single note ring tone Poss. person that definitely not shuyu E-shiddai zuqun de yi yuang.

belong E-generation group Poss. one member

‘If someone still has single-note ring tone on the cell phone, that person is definitely not a member of the E-generation.’

(25)

25 (32) Lixing yu langman pingheng, xiangbi shi Liu Xianjue de chenggong

reason and romance balance presumably be Xianjue Liu Poss. success mijue.

secret

‘The balance of reason and romance is presumably the secret of Xianjue Liu’s success.’ 理性与浪漫平衡，想必是刘先觉的成功秘诀。

Our data also lend strong empirical support to the correlation between epistemic modality and stance involvement in the aforementioned news functions. First, as observed in (19), the majority of epistemic adverbs display a broader distribution in the parts of news reports where journalists comment on the reported event or quote from interviewees. This fact conforms to the generalization that reflections and quotations are more open to subjective manipulation than narrations. Moreover, (20) indicates that the speculative subclass amounts to a greater proportion than the assertive subclass, which holds true for all three function types observed in this study. Evaluation with lower certainty thus appears more acceptable in news reporting. This again suggests that journalists’ degree of commitment correlates with their intention to mark personal stance. Specifically, assertive modality bears a stronger connection to subjectivity than speculative modality. Data illustrating the occurrence of speculative markers in our dataset are provided in (33) to (39). (33) and (34) are examples of narrations. (35), (36), and (37) are those of quotations. (38) and (39) are those of reflections. (40) to (46) give empirical evidence for the presence of assertive markers. The first two illustrate assertive markers in narrations. Those in quotations are given in the next three examples, and the last two include assertive markers used in reflections.

(33) Kezai de texing difangren weibi zhidao. Oyster Poss. nature local not necessarily know ‘About the nature of oysters, the locals do not necessarily know.’ 蚵仔的特性，地方人未必知道。

(34) Zongtongfu huoxu renwei gongtou caishi jintian jizhehui Office of the President perhaps think referendum be today press zhongdian.

conference emphasis

‘Perhaps the Office of the President thought that referendum was the emphasis of the press conference today.’

(26)

26

(35) Minjindang liwei xiaochen gaobuhao shi zongtong jiao ta nian gaozi DPP legislator joke perhaps be President ask her read script bu yao luanjiang.

not to blabber

‘A legislator of DPP joked, “perhaps it was the President who asked her not to blabber but to read the script.’

民进党立委笑称「搞不好是总统叫她念稿子不要乱讲」。

(36) Lu Xiulian shenzhi anran gaosu youren fuzongtong dagai bu shi wo. Xiulian Lu even glumly told friend Vice President perhaps not be me ‘Xiulian Lu even told her friends glumly, “perhaps the Vice President is not me.”’

吕秀莲甚至黯然告诉友人，「副总统大概不是我」。

(37) Ta ze hahadaxiao, “ni caide henmiao shuobuding ye youkeneng.” she but laugh you guess interesting perhaps also possible ‘But she laughed, “Your guess is interesting, but perhaps it is also possible.”’ 她则哈哈大笑：「你猜得很妙，说不定也有可能。」

(38) Muqianweizhi Taiwan bujiande neng zai zhelimian taodao sheme so far Taiwan not necessarily able in this earn any jingzheng yooshi.

competition edge

‘So far, Taiwan is not necessarily able to earn any competitive edge in this.’ 目前为止，台湾不见得能在这里面讨到什么竞争优势。

(39) Zhuban danwei dagai bu xiwang rang yuangdan shengqi dianli host organization perhaps not hope let New Year flag-raising ceremony chengwei xuanju zaoshi huodong.

become election campaign activity

‘The hosting organization perhaps did not hope that the New Year’s flag-raising ceremony becomes an election campaign.’

主办单位大概不希望让元旦升旗典礼成为选举造势活动。

(40) 520 taiwan wuyi jiang shi zhong wai zhumu de jiaodian. 520 Taiwan undoubtedly will be China abroad fix eyes Poss. focus ‘520 Taiwan will undoubtedly be the focus at home and abroad.’

520 台湾，无疑将是中外瞩目的焦点。

(41) Canyiyuan Kairui masaizhusaizhou duding chengwei minzhudang houxuanren. senator Kerry Massachusetts definitely become Democratic candidate ‘Senator Kerry (Massachusetts) will definitely become a candidate of the Democratic Party.’ 参议员凯瑞（马萨诸塞州）笃定成为民主党候选人。