NEWS PRODUCTION AND ATTRIBUTION CONSTRUCTION IN CHINESE LOCAL PRESS : A PROGRESSION ANALYSIS

This research undertakes a progression analysis to examine the distinction in the ways local and national/international news is produced in Chinese local press. It takes online newswriting processes of Chinese journalists as the analytical point of entry. Based on ethnographic data collected in a Chinese local newspaper, the study finds significant differences in the production of local and national/international news in Chinese local press. It is argued that such differences are relevant to the ways attribution is constructed. Interviews with Chinese journalists reveal that they foster distinctive strategies in constructing attribution in accord with the kinds of news they work on. Additionally, the study finds that attribution constitutes significant discursive resources through which journalists exercise power in news production. only review those reports concerning themselves. Provincial Department of Publicity do not interfere with the way you report, but they will stop you from reporting certain negative events.


Introduction
This study explores the distinctive ways Chinese journalists construct local and national/international news stories in Chinese local press by taking their online writing processes as the analytical point of entry. It pays special attention to those strategies adopted by Chinese journalists to construct attribution and to the functions performed by attribution in news discourse. Attribution is defined in this paper as "any linguistic means one might use to indicate who is responsible for saying something" (Scollon 1997: 384). To analyze news production and attribution construction, this paper adopts a progression analysis (PA), which is "a computer-based, multilevel method for research on writing in the workplace (Perrin 2003: 907). Daniel Perrin further explains that "PA focuses on the situational context of the writing process (macro level), the movement of writing throughout the growing text (meso level), and the writer's consciously applied revising strategies (micro level)" (Perrin 2003: 907).
Certainly, it is important to know how journalistic attribution is constructed through that which is linguistically (semiotically) expressed in news texts. However, it is equally important to understand how attribution is displayed through that which is absent in news texts. Hence, the study attempts to fill this research gap. To do so, Section 2 outlines the procedures in collecting ethnographic data. In Section 3, the study firstly explicates the writing situations of my research participants and examines their writing progression (Section 3.1); it then investigates the application of writing strategies to construct attribution in news discourse (Section 3.2); lastly, it explores the functions of journalistic attribution in news discourse (Section 3.3). The study is concluded with a summary of findings and discussion on limitations (Section 4).
The Chinese context was selected mainly for two reasons. One is that research on professional practices in the Western contexts has been largely detailed elsewhere (e.g. Bell 1991;Cotter 2010;Cottle 2007;Cottle & Ashton 1999;Davis 2000;Gans 2004Gans /1979Jacobs 2011;Machin & Niblock 2006;Perrin 2013;Van Hout, Pander Maat & De Preter, 2011). By contrast, accesses to Chinese news institutions have been extremely limited in China, and as such we have as yet insufficient knowledge about news production in relation to journalistic practices in the Chinese context. The other reason is that of convenience because the findings to be reported in this paper are part of a larger project investigating journalistic stance in Chinese and Australian hard news reporting .

Methods
This section presents the methods adopted in collecting ethnographic data on filed sites. At first, it is necessary to briefly explain the Chinese context of newspaper publication. Basically, there are two major types of newspapers in Mainland China, one being official newspaper of the government and the committees of the Communist Party of China (CPC) at different levels (i.e. Partyorgans) while the other being called market newspaper. Both are controlled by the government, but it is not compulsory for market newspapers to report routine official activities of local authorities, unless such activities are of great importance to local citizens.
The set of ethnographic data was collected in a municipal-level Party organ in a city in the middle part of China. This newspaper was chosen for several reasons. The most obvious one was that the newspaper under investigation was one of the few available news institutions to which I could gain access. Secondly, it was a Party newspaper that was considered in China as so-called "mainstream" newspaper, and thus could by and large represent the "mainstream" newspapers in China. Thirdly, that newspaper was a daily newspaper which built its reputation on hard news reporting. Fourthly, since all newspapers published in Mainland China were subject to censorship of the government at different levels, market papers and Party papers only differ from each other in form but not in substance. Lastly, the Party paper under investigation established a press conglomerate to commercialize itself, and as such it was able to show the most recent ways of journalistic professional practices in the Chinese context. According to Blommaert & Dong (2006: 19), gaining access to the field is the first step to "understand and study the possible contexts in which your object will occur, micro as well as macro". However, due to the political nature of Chinese newspapers, accesses to newsrooms were extremely difficult to obtain. Additionally, the study is concerned with recording and analyzing journalists' newswriting processes, what requires journalists' great patience because I was not allowed to install on their computer the Inputlog tool that I adopted to record their writing. In this sense, my research participants needed to work on my computer, a matter that they found uncomfortable with. Eventually, I was successful in gaining access to the field by utilizing strategies as diverse as online carpet search and snowball strategy. By online carpet search, I mean I used Weibo, the Chinese version of Twitter, to find target journalists and sent them messages requesting their approval to be involved in my research. By snowball strategy, I mean I made use of some "guanxi" (social network) to maneuver this process. I contacted friends who might refer to me their journalist friends who might agree to either be involved in my research or be willing to recommend their fellow journalists. Eventually, five participants allowed me to log their writing processes. In accord with the ethics requirements of the project, all of them were clearly informed of ethic issues. Informed consent was obtained before interviews. They were aware that they could quit the research or refuse to answer any questions that they felt uncomfortable about at any time without giving any kind of notice to the interviewer. They were informed that their names and their affiliations would not be identified.
Datasets collected on field sites include field notes, media policy documents, keystroke logging data of newswriting processes, audio recordings of retrospective interviews after journalists' newswriting, and interviews with journalists, editors and media managers. The interviews with journalists were conducted in mandarin Chinese and were consequently transcribed by the researcher.
To record, replay and analyze journalists' newswriting processes, I deployed the tool of Inputlog (http://www.inputlog.net). According to Leijten & Van Waes (2013), Inputlog is a Microsoft Windows based logging tool that logs all types of input modes including keyboard, mouse and speech recognition, and it generates data files for pause, revision and source analyses. Since I was not allowed to install the software on journalists' working computers, it was installed on my computer. It was subsequently activated from the moment the journalists said they were ready to start writing until they filed the news story and submitted online for copy-editing. A follow-up retrospective interview with the journalist was immediately conducted afterwards.

Findings
In the following, I will report findings surrounding three key research issues. The first one is concerned with a close examination of newswriting processes with a view to comparing the ways news sources are integrated in newswriting (Section 3.1). The second one relates to how Chinese journalists attribute to news sources of different social statuses (Section 3.2). The last issue pertains to an exploration of the functions of attribution (Section 3.3). Section 3.1 is based on the analysis of writing processes, whereas Section 3.2 and 3.3 are premised on interviews with Chinese journalists.

Writing process analysis
In this section, I outline the writing processes of two journalists who report local and national/international news. The analytical focus will be on the different manners through which news is produced in general, with a view to identifying the distinctive ways of attributing to different news sources in newswriting.

Writing process of journalist T working on local news
Let us firstly examine the writing process of journalist T who worked on reporting local economic and social news. The news story was about the rising industry of local intermediary. The main news fact was obtained from the inauguration conference of the provincial association of intermediary that the journalist T and I went and attended together. We were offered pre-formulated materials, together with publications of the news event in a provincial-level Party organ. After the conference, journalist T interviewed LP and HY, directors of the provincial and municipal Administration of Industry and Commerce respectively. Before journalist T started her writing, she quickly scanned the pre-formulated reports and other materials available at her disposal. This perhaps gave her a clear idea of what to be included in the news story, and in part explained why her writing process ( Figure 1) shows no significant planning time as we would envisage in Figure 2 where she wrote a news story without pre-formulated reports. In Figure 1, the blue solid line indicates the writing process including all words produced by the writer from the beginning to the end (from zero to 3,125 words); the green solid line shows the final product of news story in terms of word count (from zero to 2,227 words). The green dotted line shows the cursor position. Figure 1 clearly demonstrates a writing process from the beginning of a text to its end, by which I mean she typed in every single word by herself as opposed to copying the text from elsewhere (see Figure 3 below in Section 3.1.2).
She spent one hour fifty-five minutes and thirty-three seconds in producing this news story with a total pause time of forty-five minutes. However, during the writing process, she was interrupted by a phone call and that took her around thirty-five minutes. Hence, the total pause time should be around ten minutes during the entire writing process. She produced a total of thirty-nine sentences amounting to 2,092 Chinese characters excluding spaces. Among them, she produced twenty-two sentences of attribution, accounting for 60 per cent of all sentences produced. That is to say, attribution constitutes a significant part of her writing, and this is further evidenced in her second newswriting.
The second example of the writing process of journalist T was shown here for purposes of comparison. The second news story produced was related to the deceiving use of scale in a local market by some shopkeepers. The journalist T and I followed officers at the municipal Bureau of Weights and Measures on the World Metrology Day (20th May 2013) to a randomly selected local market to investigate possible deceiving use of scales. We were not offered any preformulated reports, but there were some materials providing background about the World Metrology Day itself.
She spent a total of two hours and forty-seven minutes on writing the news story with a total pause time of one hour and forty-six minutes. She produced a total of 13,010 words with 9,197 words typed in (excluding spaces). Among the twenty-nine sentences in the final news text, nine are quotations from other news sources, amounting to 805 words (46 per cent of the total 1,736 words in the final product). There are clearly three stages of writing in producing the second news, as indicated by three clusters of pauses (orange dots in the diagram). The first stage is short and mainly involves the planning of structures and the writing of opening frames. The second stage is the main stage of newswriting. The last stage is the revision process in which we envisage she moved the curser back and forth in news text at issue.
In short, the writing processes of journalist T show that her newswriting is gradual and normally takes a substantial amount of time and that the writing process is often interrupted by other activities. Additionally, attribution construction occupies a significant part of news production. That said, I now move to examine the writing process of journalist Z whose daily work involved reporting or editing national/international news.

Writing process of journalist Z working on national/international news
The typical writing process of journalist Z working on editing national/international news in local press is illustrated in Figure 3 which shows clearly that he copies a text into a document (from Xinhua News Agency online news database) and that he then deleted a substantial part of the text. The experienced editors, as I observed in their workplace, normally take less than fifteen minutes to produce a news story. In fact, I have found it difficult to label those journalism practitioners working on national/international news in Chinese local press, because on the one hand they do not report news events on their own in the way journalist T does, whereas on the other hand the news stories they produce are subject to further editing. Since the study is concerned with the production of local and national/international news rather than who produces such news, the distinction will not be made between different journalism practitioners working in Chinese local press. The news story he produced was related to the notorious "Guomeimei Scandal" in relation to the Red Cross Society of China (RCSC). The lady Guomeimei, who registered her Weibo account as the commercial manager-ingeneral of the RCSC, was reportedly involved in an extramarital sex with a senior manager of the RCSC. The news story was selected from an online Xinhua database providing updated daily news happened in China and elsewhere in foreign countries. The news itself was about a press conference held by the RCSC in which reporters from Xinhua news agency challenged the ability of RCSC in handling the Guomeimei Scandal and its capacity in overseeing the organization itself. In producing the news story, the journalist Z also copied materials from other news stories provided by Xinhua news agency. He firstly opened the document but did not write a single word until one hour six minutes and twentyfive seconds later when copied a total of 4,322 words (including spaces) from the online database. He then browsed the text and started editing the text. He finished editing the news text in one hour fourteen minutes and thirty-nine seconds. That is to say, the total editing time lasted for around eight minutes and fifteen seconds, and during this time he deleted 3,107 words. The phrase he produced was limited to a frame at the end of the news text "据新华社6月14日电" (According to report by Xinhua News Agency on 14 June). I was told that this frame was a significant marker of news sources. He added that it signaled that the news story at issue was composed of several news stories from Xinhua. It is different from the frame of " 新华社6月14日电" (Xinhua News Agency on 14 June), signaling that the news at issue was copied from Xinhua without revisions. The newswriting process, as shown in Figure 3 above, documents few pauses (dots in orange), marking a striking difference from the newswriting of the journalist T as shown in Section 3.1.1. Additionally, Figure 3 shows that the journalist Z mainly deleted some parts of the original news story and did so in the last part of the newswriting process. Certainly, journalist Z is not expected to write a news story from scratch in the manner journalist T did. However, the point is that journalists in Chinese local press produce news in strikingly different ways, and that reporting on national/international news events in Chinese local press (and hence journalistic stance therein) is mediated by national news agencies.
To summarize, the above analysis indicates that news production in Chinese local press should be treated with caution in that local and national/international news are produced in different ways. Additionally, journalistic attribution is not a homogenous concept in news discourse because it is constructed differently in different kinds of news in Chinese local newspapers as well.

Attributing to different news sources
In my interviews with journalists, I found that journalists had fostered distinctive strategies in constructing attribution to news sources of different social statuses (power elites vs. ordinary citizens). In the following, I will illustrate their strategies in attributing to foreign news agencies, to Xinhua News Agency, to local authorities, and to ordinary citizens.

Attributing to foreign news agencies
Broadly, Chinese local press and its reporters are not encouraged or perhaps not allowed to attribute directly to foreign news agencies. The main concern is related to different stances foreign news agencies or newspapers undertake. Chinese journalists have internationalized this rule in their daily practices, as shown in the following excerpt of an interview intending to elicit journalists' opinions on their stance taking practices. IR: 那你们就是可以引用外媒的观点？ IR: Can you cite directly from foreign news agencies? Y: 一般不用，一般一般是不用的，因为它这个你讲的新闻立场有差异。 Y: Normally we don't. Normally, normally no, because you know that they take a different stance.
The interview with a managing editor of the newspaper at issue revealed that the restriction on citing foreign news sources was imposed by the government. That is to say, all attributions to foreign news sources must be mediated by national news agencies or national media. IR: 比如这个世界版是引自新华社还是国外通讯社？ IR: Let's take the "World" section as an example. Is this citing from Xinhua or foreign news agencies? W: 根据中宣部的规定，我们不能直接引用国外通讯社。 W: According to requirements of the Publicity Department, we cannot directly cite foreign news agencies.
The finding suggests that constraining attribution to foreign news sources has been adopted as a significant means of regulating journalistic stance in Chinese local press. All news reporting about national and international events must be mediated by national news agencies and/or other national media (e.g. Chinese Central Television). In so doing, it prevents Chinese local press from directly accessing to foreign news reporting and at the same time prevents them from channelling different stances to their readership. In this sense, the finding corroborates Scollon's (1998) argument that "attending to topics" is an important mechanism to constrain stance, in addition to attending to channel and attending to interpersonal relationship.

Attributing to Xinhua news agency
I have indicated that news reports on national and international events in Chinese local press must be mediated by national news agencies or national media. Adding to such a constraint of news topics, news from national news agencies or national media would only permit slight modification of language and style for purposes of cohesion. This is evidenced below in an interview with a senior editor. Y: It is fine as long as you do not deliberately change the intended meanings in news reports provided by Xinhua. You can edit it because you need to integrate it with other news stories. Therefore, you could adjust its language and some details so as to make the final product coherent.
Certainly, journalists in Chinese local press may choose news topics from a repertoire of news stories provided by Xinhua. However, only national media are permitted to report on national and international news events. Thereby, journalists working in local press are unable to undertake stances other than those advanced and preferred by national media. Since one principal function of national media is to reproduce dominate social values of the central government, those social values are reproduced in local press as well.

Attributing to local authorities
The primary aim of the Party organ under investigation was to promote the official stance of the government, and accordingly official events were covered in the Party organ on a daily basis. Those events constitute an important source of exclusive news to the Party organ. In reporting local political news, every journalist in Chinese local press is responsible for liaising with a designated government department so as to ensure the coverage of official events in the local newspaper. However, all local news reports must be reviewed and subsequently approved by the local government department that was the major newsmaker in the news at issue prior to publication. According to my interviewee S who was responsible for reporting local political news, once these reports were signed off, no further revision should be made except for minor typographical errors. IR: 您负责报道x部门的时政新闻，那么他们是否在语言表达或者写作风格上有 特定要求呢？ IR: As you are following news regarding specific government sectors [name of local authorities], do they require a specific kind of language or style?

S:
它的风格和其他新闻一样。他们对你的新闻报道不强加特殊的限制。但 是，他们要审核我们的报道。我们完成报道之后需要送给他们审核、报批。 S: Its style is the same as news style. They do not impose any specific constraints upon your news reporting. But they need to review our news reports. When we finish our reports, we send them [to concerned persons] to review and approve.
However, this is not to say that the government never imposes specific constraints on news reporting. On the contrary, it does so from time to time when necessary. For instance, negative local news in some cases cannot always find their way into press because the local publicity department intends to minimize the negative impact of the news, as evidenced in the following excerpt of an interview.

IR: 你认为在政府工作人员在多大程度上影响你新闻写作中的立场表达？
IR: To what extent do you think government officials may influence your expressions of stance in your news texts? Q: 他们只需要审核涉及他们部门的报道。省宣传部不会干预你的写作，但是 他们会不让你报道某些负面新闻。 Q: They only review those reports concerning themselves. Provincial Department of Publicity do not interfere with the way you report, but they will stop you from reporting certain negative events.
The above analyses have shown that attributions to foreign news agencies, Xinhua news agency and power elites are rather constrained in Chinese local press. That said, I now move to examine attributions to ordinary citizens.

Attributing to ordinary citizens
My corpus findings elsewhere forthcoming) have indicated that Chinese journalists may employ a repertoire of language resources to evaluate and engage with citizen sources. Broadly, the voice of ordinary citizens in news texts is more likely to be opened up for alternative voices, and it is ordinary citizens who are more likely to be evaluated in negative ways. However, citizen sources are often adopted by Chinese journalists as eyewitness in news reporting.
My interviews with journalists reveal that ordinary citizens are generally considered as less important than elite sources in terms of the opinions they could provide about a news event. For instance, journalist L said "views from elite sources are definitely opinions about a specific field, and therefore more authoritative and professional. As for citizen sources, how could they possibly know [such a field]?". Additionally, journalist T mentioned that she always put elite sources in late part of her news texts, because they were more important.
In short, this section has indicated that journalists in Chinese local press have fostered different strategies to construct attribution to news sources of different kinds.

Functions of journalistic attribution
The different ways of attributing to news sources of various social statuses are relevant to the functions of attribution in news discourse. One major function is to take on different social-interactive roles. According to Goffman (1974;1981), there are mainly three communicative roles: animator, author and principal. The "principal" role refers to "someone whose position is established by the words that are spoken, someone whose beliefs have been told, someone who is committed to what the words say" (Goffman 1981: 144, cited in Scollon 1997. The "author" role is defined as "someone who has selected the sentiments that are being expressed and the words in which they are encoded (Goffman 1981: 144, cited in Scollon 1997; whereas the "animator" is what Goffman calls the "sounding box" through which utterances are made (Goffman 1981: 144).
To animate others' words in news discourse is a highly ambiguous act in that it is not the case that journalists always provide a verbatim account of news sources. For instance, journalist T said she was not quoting exactly what her news source said when I asked her about the use of a colloquial word. In quoting her news source, she assumed the power of animating the utterances based on the interviews. The journalist T recalled that the reason for doing so was twofold. One was that she needed to control information flow in news texts and thus to better organize news. The other reason was that she slightly modified utterances to appeal to readers. It is argued that such entextualization is an indispensable mechanism for the construction of journalistic professional authority (e.g. Park and Bucholtz 2009: 485). By "entextualization", I mean "the process by which circulable texts are produced by extracting discourse from its original context and reifying it as a bounded object" (Park and Bucholtz 2009: 485).
In the process of newswriting, my research participants normally start writing a byline displaying his or her name at the beginning of a story or displaying sources of news. One such typical example is presented as follows:

本报讯(记者X)
Name of News Agency, by Reporter (Name of the journalists) As Ron Scollon observes "there is an ambiguity involved with being named as the author of a text. That is to say, having a text attributed to oneself may represent either the accretion of power or the denial of power" (Scollon 1997: 387). Journalist T said in the interview that bylines mainly served to fix responsibility. This is particularly important for political news reporting, because Chinese journalists are not allowed to coauthor with his/her correspondents if he/she contributes little to the concerned news text. Correspondents here mainly refer to government officials who occasionally produce a political news report on behalf of the local government. This rule is inscribed in news policy document of the newspaper under investigation (see below an example of the rule). Violation of the rule is subject to severe punishment.

三十二、记者应尊重通讯员的劳动，若对通讯员稿作了补充采访并作较大修 改的，可加署记者名，否则，不能署名。对不该署记者名而加署名的，一经 发现，将严肃处理。
Rule 32: Reporters should respect the work of correspondents. If substantial revision was made based on additional interviews, then reporters could be co-authors. Otherwise, this is not allowed. Anyone violating the rule will be severely punished.
In this sense, as Ron Scollon puts it succinctly, the power to command authorship in Chinese press could also be viewed to represent "the powerlessness of journalists in China" (Scollon 1997: 387).
The powerlessness of journalists in China is further reflected in the fact that they cannot decide whether or not power elites will be given voice. Their ability to delegate principalship or to delegate responsibility to power elites is rather constrained, although they retain animation.
Another function of attribution is to display journalistic stance. Abundant studies in the literature have demonstrated that attribution constitutes significant discursive resources through which journalists display stances in news discourse (e.g. Bednarek 2006;White 1998). Journalists do not simply reflect or mirror the objective world in news discourse, but engage in reconstructing news events through discursive resources (e.g. Beeman & Peterson 2001;Van Hout & Jacobs 2008). In this process, journalists take certain stances in line with social values that are currently operational in the society and institution at issue. For instance, based on my corpus findings elsewhere (Huan forthcoming), I found that Chinese journalists in general were more likely to open the dialogic space for alternative voices in attributing to ordinary citizens, but they were less likely to do so in attributing to power elites (e.g. government officials). One reason for this was to show alignment with the government. Certainly, stance taking practices are not limited to attribution, nor do they only function to align with the powerful. The point is that attribution may be manipulated by journalists to distance from or align with news sources in accord with different values they practice (see also Jacobs 1999). For instance, the journalist T said she would find an opposite voice when she questioned the voice at issue. T: I will then find a different voice. Then you can keep a balance, and readers can have their own judgment… If I totally disagree with him, I will not attribute to him. Furthermore, attribution may index journalistic professional identity. Attribution is traditionally considered as an indicator of objectivity of news. Hence, the use of attribution in news texts shows that journalists attempt an objective news report rather than a fiction. For instance, journalist Q believed that direct quotations would add to the objectivity of news. To directly quote news sources is considered by journalist Q as a distinctive and professional way of practicing journalism. A senior managing journalist P said that the way news sources were incorporated was important as well. She emphasized that a good journalist should not be influenced by news sources. In saying that "a good journalist will never be influenced by others (news sources)", journalist P indicates the link between various functions of attribution: commanding animation, delegating principalship, undertaking stances and displaying identity. In other words, being a good journalist, as assuming a professional identity, involves to effectively animate an utterance and to show appropriate stances, but at the same time retain the power to delegate or assume principalship to news sources.

Conclusion
The progression analysis of newswriting of journalists T and Z working on local and national/international news reporting has provided us with insights into understanding news production in the Chinese context, insights that are not available through traditional research into attribution in news products alone. The analysis shows that journalists in Chinese local press produce news in distinctive ways depending on the kind of news they report. The different ways in producing local and national/international news are relevant to the distinction in the manner attribution is constructed in news discourse. Broadly, journalists reporting local news have accesses to various news sources, whereas reporting national and international news must be mediated by national news agency and/or national media. Such ways of news production has direct and substantial impact on the strategies that journalists may employ to construct attribution. It is argued that adopting different strategies to construct attribution to news sources of different social statuses is relevant to performing various functions of attribution. Interviews with Chinese journalists, editors and media managers shows that, in addition to the display of journalistic stance and professional identity, attributions in news discourse constitute a significant mechanism through which journalists take on communicative roles of commanding animation, commanding authorship and delegating principalship.
However, the study is limited in several aspects. One limitation concerns the limited number of research participants involved in my study. The daily operation of Chinese press is supervised by the government, and as such Chinese journalists and news agencies are less willing to participate into research of the kind carried out in this study. Another limitation relates to the computer tool -Inputlogwhich I adopted to register writing processes of Chinese journalists. This tool turned out to be less useful in logging the input of Chinese characters. Such a technological limitation has prevented me to analyze detailed revisions of journalists' writing of the kind carried out by Perrin (2003;2013), Sleurs and Jacobs (2003) and Van Hout, Pander Maat & De Preter (2011). A further limitation is that I was only able to investigate writing processes of Chinese journalists in one Party organ, and in this sense the findings cannot represent professional practices of Chinese journalism practitioners in general. Perhaps the last caveat could be overcome by comparing journalistic professional practices in Chinese newspapers of different kinds (e.g. market newspaper vs. Party organs), but that falls outside the scope of the present research.