Functional Classification of Lexical Bundles in News

Like the structural framework presented above, the functional framework in the previous chapter (see Section 5.2) is also potentially applicable to Chinese bundles in various text types. The results show that news bundles in Chinese can be adequately accommodated in this functional framework.

6.2.1 Interpersonal Bundles

The first subcategory is special interactional bundles. There are only three bundles in this subcategory: i.e., shi bu shi ‘A-not-A QUESTION’, shenme yang de ‘what kind DE’, and you ren shuo ‘have people say’. These bundles are also identified in the conversation subcorpus.

The second subcategory is epistemic stance bundles. In news writing, the writer’s epistemic stance can be expressed through modal expressions (e.g., keyi shuo shi ‘it can be said that’ conveys certainty) or the negator bu ‘not’ (e.g., bu shi hen ‘not very’ serves as a mitigator). It is found that stance bundles in news writing are all impersonal.

There are supposed to be three subcategories for interpersonal bundles. However, no attitudinal/modality bundle is identified in the news subcorpus.

6.2.2 Discourse Organizers

The first subcategory is topic introduction bundles. All the news bundles in this subcategory are presentational phrases, i.e., you yi ge/zhong/time/wei ‘have one

CLASSIFIER; there is a’.

The second subcategory is topic elaboration bundles. The additive adverb ye

‘also’ often occurs in elaboration bundles. Examples include ye you ren ‘also have people; there are also some people’, which is used to identify another group of people, and ye jiu shi ‘also just be; that is to say’, which is used to paraphrase what is just said.

The word ling ‘another’ also occurs in elaboration bundles very often (e.g., ling yi ge/zhong ‘another one CLASSIFIER’), and these bundles are used to enumerate more items.

The third subcategory is identification bundles. In news writing, identification bundles are used to highlight something as newsworthy and grab the reader’s attention.

Therefore, extremity-signaling adverbs often occur in identification bundles. The adverb zui ‘most’ is the most common one (e.g., zui da de ‘most large DE’, zui zhongyao de ‘most important DE’), and ji ‘extremely’ is also attested (e.g., ji da de

‘extremely large DE’). Identification bundles are also used after a lengthy elaboration to provide a concise summary or brief comments. Consider zhe shi yi ‘this is a’ in the

following example.

(6.1) Wo ganjuedao meiyou ren bu zhongshi, bu guanxin ziji guojia de anquan.

Suoyou de ren dou shi zai zheyang de jichu shang fabiao ziji de yijian.

Suiran you ren leguan, you ren baoliu, dan suoyou ren de chufadian dou shi yiyang. Wo juede zhe shi yi ge man kexi de xianxiang.

‘I feel that all people value and show concern for the safety of their own country. Everyone expresses personal opinions on the same basis.

Although some are optimistic and some are conservative, all people have the same point of departure. I feel that this is a very positive phenomenon.’

In (6.1), the identification bundle zhe shi yi is followed by a brief opinion (i.e., man kexi de ‘very positive DE’) and a shell noun (i.e., xianxiang ‘phenomenon’) that serves as a cohesive device to enclose the preceding discourse.

6.2.3 Referential Expressions

The first subcategory is phoric bundles. No phoric bundles in the news subcorpus feature the demonstrative na ‘that’; only zhe ‘this; these’ is attested (e.g., zhe ge wenti

‘this CLASSIFIER problem’, zhe ji nian ‘these several year; in recent years’).

The second subcategory is entity bundles. Most entity bundles in the news subcorpus are time expressions (e.g., ge yue nei ‘CLASSIFIER month within; within … months’, shang ban nian ‘up half year; the first half year’).³⁶ A time frame is

36 In the present study, high-frequency word sequences with specific numbers other than yi ‘one’ are manually excluded (see Section 3.4). However, san ge yue ‘three ^CLASSIFIER month’ and liu ge yue ‘six

CLASSIFIER month’ remain in the data. The former is a quarter, and the latter is half a year. These two are common time frames in news writing, particularly in business news.

necessary for the presentation of a news event and the interpretation of news data (McKane 2006). For example, seven joyriders were killed means little unless a time frame such as in the past year is added. Entity bundles referring to places (e.g., shijie ge di ‘world every place; around the world’, zai wanglu shang ‘on Internet up; on the Internet’) and people (e.g., de ren dou ‘DE person all; those who … all’) are also attested.

The third subcategory is attribute-specifying bundles. Some bundles in this subcategory describe intangible attributes, including types (e.g., de yi zhong ‘DE one kind; one kind of’), processes (e.g., de guocheng zhong ‘DE process middle; in the process of’), and purposes (e.g., shi wei le ‘be for LE; for the purpose of’). Other bundles describe tangible qualities, such as sizes (e.g., hen da de ‘very large DE’) and quantities (e.g., zhi you yi ‘only have one’).

The fourth subcategory is process bundles. In the news subcorpus, bundles expressing relational processes (see Section 5.2.3) are the most common (5 out of 11, 45.5%) and usually feature the copula shi (e.g., shi yi ge ‘be one CLASSIFIER’, bu shi yi

‘not be one’). In news writing, these copula-bearing bundles are thought to convey the journalist’s certain stance, i.e., showing that the information presented is seen as factual (Kaneyasu 2015).

Referential expressions also include imprecision bundles (e.g., or something like that). However, imprecision bundles are not attested in the news subcorpus.

6.2.4 Functional Distributions of Lexical Bundles

Table 6.2 presents the functional distribution of Chinese bundles in conversation.

Still, both the type distribution and the token distribution are presented. Because there are only three four-word news bundles identified in the present study, only the

structural distribution of three-word news bundles is tabulated.³⁷

As can be seen from Table 6.2, the type distribution and the token distribution display similar tendencies. For three-word news bundles, the most common functional category is referential expressions, followed by discourse organizers and interpersonal bundles. Entity bundles are the most frequent referential expressions. The high frequency of entity bundles is attributed to its close association with two common patterns ‘modifier + de + noun’ and ‘number/demonstrative + classifier/quantifier + noun’. Besides, as mentioned above, time frames are essential elements in news writing, and they are usually classified as entity bundles. As for discourse organizers, identification bundles and topic elaboration bundles are common subcategories.

Identification bundles provide a focus for something newsworthy and are often used to attract the reader’s attention. Only a marginal proportion of news bundles are interpersonal bundles, and no attitudinal/modality bundles are identified in the present study.

37 The three four-word news bundle identified in the present study are zui zhongyao de shi ‘most important ^DE be; the most important thing is that’ (identification bundle), you hen da de ‘have very large ^DE’ (attribute-specifying bundle), and mei ge ren dou ‘every ^CLASSIFIER person all’ (entity bundle).

Table 6.2(a). Functional distribution of lexical bundle types in news.

Functional category Three-word news

1. Interpersonal bundles 6 (6.9%)

(1) Epistemic stance bundles 3 (3.4%)

(2) Attitudinal/Modality stance bundles 0 (0.0%) (3) Special interactional bundles 3 (3.4%)

2. Discourse organizers 31 (35.6%)

(1) Topic introduction bundles 4 (4.6%) (2) Topic elaboration bundles 12 (13.8%)

(3) Identification bundles 15 (17.2%)