Self-Reflexive Journalism

A Corpus Study of Journalistic Culture and Community in the Guardian

Anna Marchi

• Publisher: Routledge
• Year: 2019
• eBook ISBN: 9781351714129
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This book develops a corpus-assisted approach to the study of self-reflexivity in journalism and examines the ways in which news workers and subsequently, news organizations, choose to promote an identity for themselves and the ideologies that accompany them. Using The Guardian as a case study, the volume draws on its Corpus Assisted Discourse Studies (CADS) to explore ways in which a newspaper can reflect upon itself, including how newspapers conceptualize the role of the media, how they define good vs. bad journalism, what they see as professional values, how they attempt to cement community membership amongst their readers, how they construct and project their overall identity and role as newspapers and also how they see their position within the larger community. A chapter on the book's methodological framework reflects on critical aspects of CADS, including triangulation, objectivity and subjectivity, total accountability, and replicability. CADS methods are applied in the analysis chapters, with accompanying reflections on what we learn about the strengths and also maybe about some of the limitations of corpus methodology. A summarizing chapter ties these strands together to make the case for a CADS approach to journalism and media studies and look to the future at how the digital age has shaped the journalism landscape. With its focus in extending a CADS approach to other aspects of journalism scholarship, this volume is key reading for graduate students and researchers in corpus linguistics, discourse analysis, media studies, and journalism studies.

1 Introduction

1.1. Character Is a Subtle Affair

Character is a subtle affair and has many shades and sides to it. It is not a thing to be much talked about, but rather to be felt. It is the slow deposit of past actions and ideals. It is for each man his most precious possession and so it is for that latest growth of time the newspaper.

(C. P. Scott, Manchester Guardian, 5 May 1921)

This quote comes from a piece written by the editor of the then Manchester Guardian, C.P. Scott on the occasion of the newspaper’s centennial. In the article, Scott declares the core values of the newspaper, and he goes on to say: ‘[f]undamentally it implies honesty, cleanness, courage, fairness, a sense of duty to the reader and the community’. These are the qualities that the editor expected in the newspaper along with the standard of excellence he expected from the people making it.

Commenting on Scott’s words, journalist and journalism scholar Jeff Jarvis wrote: ‘[i]f you hang around Guardian people long enough—an hour or two—you’ll hear them quote or refer to the revered long-time editor of the paper, C.P. Scott. In particular, they quote his 1921 essay’ (Jarvis, 2006), suggesting how Scott’s vision of journalism is still very much alive in The Guardian’s newsroom and also implying that the ‘Guardian people’ are very self-conscious and eager to share details about their proclaimed journalistic mission.

This book is about The Guardian and The Guardian people; in particular, it is about how The Guardian and the people creating it understand (or claim to understand) their professional role and the rules, values and a sense of journalistic community.

The relevance of journalism as a powerful agency of symbolic power—‘one of the ways society tells itself about itself’ (Dickinson, 2008: 1)—is largely recognised. News gives us knowledge about the world that is beyond our direct access and this knowledge is socially constructed through the dozens of choices that journalists regularly make, for example: what they select as news, in what order they report information, what language they use. While extensive work has been done on the linguistic analysis of news discourse, less attention has been paid to the people behind the storytelling. Research on news discourse to date has mainly focused on journalism as a product, while overlooking journalism as a process, which includes the people responsible for it. With this analysis, I wish to contribute to filling this gap.

Journalists create community (or communities) through the discourses that they circulate about themselves (Zelizer, 1993) and the boundaries and the characteristics of the ‘journalistic field’ (Bourdieu, 1999) are construed and reproduced discursively. The journalistic field, like other fields (political, social sciences, etc.), is a social universe inhabited by individuals that are cast in the invisible structure of the ‘field’. In Bourdieu’s terms, the field is a set of forces that constrain and determine the action of agents, but at the same time ‘agents react to these relations of forces … they construct them, perceive them, form an idea of them, represent them to themselves’ (Bourdieu, 2004: 30). In other words, the journalistic field is made of the ‘claims and counterclaims’ (Conboy, 2004: 4) about journalism. Through their professional tales, journalists define the boundaries of their community and set the standards of good vs. bad journalism. The meaning of being a journalist is built through daily performance and in their writing newsworkers reinforce consensus about the ideal-typical values that define what a ‘real’ (Deuze, 2005) journalist is and what journalism is or should be.

1.2. Analysing Self-Reflexive Journalism Using Corpora

‘Journalism is an intensely reflexive occupation, which constantly talks to and about itself’ (Aldridge and Evetts, 2003: 560) and the role and meaning of journalism can be studied by means of the self-reflexive traces in texts. That is, they can be detected and studied in a newspaper corpus.

This work builds methodologically and theoretically on the work done in the field of Corpus-Assisted Discourse Studies (CADS, as coined by Partington in 2004). CADS (Baker and Ellece, 2011: 24–25) covers a wide range of research activity and simply implies discourse analysis carried out in some way with the integration of corpus techniques. CADS practitioners may operate under different labels, such as corpus-based (critical) discourse analysis, corpus-based sociolinguistics or under no specific label. In this book, therefore, I will use the umbrella label CADS to refer to the whole ensemble of kin approaches, where the analysis of discourse is carried out with the assistance of corpus linguistics. Ample space will be given to epistemological and methodological discussion in chapter 3, and reflections on CADS and its practices will accompany us every step of the way, but I will provide some general background here.

1.2.1. Corpus-Assisted

Definitions of corpus linguistics¹ tend to be broad, both because the field is wide and difficult to delimit and because the term has many referents: it indicates the source of data, which is also the tool of analysis, and it is the broad label for an approach to language study. Tognini-Bonelli (2001) arguing for the ‘pre-application methodology’ status of corpus linguistics, suggests that corpus linguistics is a discipline accommodating two different methodologies, the corpus-based and the corpus-driven one. While in corpus-driven studies ‘theoretical statements are fully consistent with and reflect directly, the evidence provided by the corpus’ (Tognini-Bonelli, 2001: 84). The corpus-based approach supposedly sees the corpus as a library of texts with which to test hypotheses. Biber, 2009 endorses the distinctiveness of a corpus-driven approach which ‘differs from the standard practice of linguistics in that it makes minimal a priori assumptions regarding the linguistic constructs that should be employed for the analysis’ (Biber, 2009: 276). Yet, no research is entirely a-theoretical: ‘[a]ll analysts bring some prior understanding of language to the analysis; this is unavoidable and, in all likelihood, useful… . Since prior understanding cannot be eliminated, the question is how much that prior understanding is relied on rather than whether that prior understanding is relied on… . But this is a matter of degree, not kind. The distinction between the two approaches to corpus data is quantitative, not qualitative’ (McEnery and Hardie, 2012: 161).

Corpus linguistics is understood here as a methodology (following McEnery and Wilson, 1996; Biber et al., 1998; Meyer, 2002; McEnery et al., 2006; Gries, 2009a, 2009b, among others). This presupposes that ‘methodology’ is intended to refer to ‘a whole system of methods and principles of how to apply corpora’ (McEnery et al., 2006: 7), since, as McEnery and Gabrielatos (2006) point out, corpus linguists’ methodological practices are not uniform. Methodology is treated as a fluid concept, because methods and all stages of the research process are imbued with theory and epistemology, in the same way ‘theories are … constructed by methods’ (Hunston and Francis, 2000: 250). As Partington puts it:

To make such clear distinctions between instruments and enterprise is anachronistic. Firstly, because observation informs theory just as theory informs observation… . Secondly, … the object of observation is only tangible, in a sense only exists for an observer … through the instruments of study… . There is a sort of indivisible hermeneutic package—the observer including mind, observational instruments, observations, objects of observation. Alterations in any of the parts will affect the entire system, a process usually known as scientific advance.

(Partington, 2009: 295–296)

Consequently the distinction between corpus-based and corpus-driven is here seen as a rather fluid one, a ‘cline’, in McEnery and Gabrielatos’ (2006: 45) words, between two extremes: ‘linguists who neglect corpora … and … linguists who limit themselves to corpora’ (Johansson, 1991: 6). The space in between is filled by a constellation of corpus approaches eclectically combining intuition and rigorous use of corpus evidence. As stated earlier, these approaches go under diverse names, as varied as are the ways to enter the data and tackle questions. What differentiates corpus-based/assisted approaches (such as are found in CADS) from so-called corpus-driven ones or, more in general, from other approaches in the vast territory covered by the label corpus linguistics, is to be found in CADS’ aims rather than in its methods. So-called corpus-based/assisted research is interested in investigating language from a social as well as from a structural perspective (see Conrad, 2002: 75). CADS analysis aims at describing discourse structure, but also attempts ‘to move from the descriptive to explanatory adequacy’ (Stubbs, 2006: 34). Furthermore, in preferring the term corpus-assisted throughout this work, I second the motivation offered by Partington et al. (2013):

[i]t was felt that a term was needed not only to describe the kind of study which incorporates quantitative/statistical methods in the study of discourse types but which also emphasised the eclectic nature of the approach. That corpus techniques were only one sort amongst others and that CADS analysts employ as many as required to obtain the most satisfying and complete results, hence “corpus assisted”.

(Partington et al., 2013: 10)

CADS is traditionally described as the practice of investigating particular text types by combining the quantitative rigour of corpus linguistics with the social perspective of qualitative approaches to discourse analysis. Starting with Hardt-Mautner’s seminal paper (Hardt-Mautner, 1995a), the methodological literature in the area has amply reasoned on the ‘synergy’ (see Baker et al., 2008) between corpus linguistics and discourse analysis (Critical Discourse Analysis in particular) and corpora and discourse studies as an example of mixed methodology have most recently been re-examined by Taylor and Marchi (2018). But the dichotomy between qualitative and quantitative is misleading because one entails the other:

Qualitative distinctions and judgments … are prerequisites to quantitative measurements, the two are inseparable… . Qualitative change cannot be understood, let alone achieved, without noting the accumulation of quantities… . To consider quantification only mindless counting or number crunching is both a philosophical and strategic fallacy.

(Gerbner, 1983: 361)

Gerbner’s reference to ‘accumulation’ points to an idea largely shared by linguists and social scientists across borders, from Halliday’s ‘coding orientations’ (1985) to Bordieu’s ‘habitus’ (1984). Quantification of phenomena accounts for the pervasive power of repetition, which has peculiarly qualitative effects, as most ‘qualitative’ scholars would recognise: ‘[a] single text on its own is quite insignificant: the effects of media power are cumulative, working through the repetition of particular ways of handling causality and agency, particular ways of positioning the reader and so forth’. (Fairclough, 1989: 54).

Baker points out that there are ‘fuzzy boundaries’ (Baker et al., 2008: 296) between what we identify as quantitative and qualitative and that ‘corpus analysis shares much in common with forms of analysis thought to be qualitative’ (Baker, 2005: 36). Biber et al. (1998) remind us that corpus-based analysis depends on both quantitative and qualitative techniques, the mix is intrinsic to the methodology, actually the mix is the methodology. ‘Qualitative and quantitative are labels we attach to what we do and methods should be tools, not schools’ (Marchi and Taylor, 2018: 4) and ultimately what CADS analysts do is a qualitative analysis of quantifiable patterns. And quantification itself is necessarily underpinned by inchoate (‘qualitative’) interpretations, for example the assumption that it is more useful to look at collocates co-occurring within a span of five or less words from the node.

As Partington et al. explain in their excellent introduction to theories and practices in corpus-assisted discourse studies, CADS is best described by what it does, which points back towards the aforementioned eclectic nature of CADS and to the fundamental fluidity of its methods, flexibility of its tools and anti-reductionist (or holistic) nature of its intents (all of which will be amply discussed in chapter 3).

CADS is broadly inductive, as corpus linguistics in general is. Induction applies to both the research design and the analytic phase: questions as well as answers derive from and rely on the observation of large amounts of naturally occurring language. The researcher interrogates the data and looks for recurrent patterns on the basis of which s/he proposes generalisations. The process is thoroughly empirical, but it does not exclude intuition. A CADS analyst decides on a topic, then turns to an appropriate set of data in order to investigate that issue; the preliminary exploration of the corpus, that is the observation of the ‘boiled down extract’ (Scott and Tribble, 2006: 6) of the data (e.g. frequency lists), informs the research question and the corpus/corpora design, in a reiterative and interactive process. Throughout the process, internal, theoretical, methodological and practical matters intertwine and interact with external ones, such as personal interests, previous experience and omnivorous readings. The research question is progressively refined, and new research questions swim into view, in a double movement of focus and zoom. The bottom-up logic does not exclude deductive incursions: at times, very general research questions become hypotheses to test, which can in turn be tried on the corpus. Saying that the data can be used to corroborate or falsify (in Popper’s terms) a hypothesis, clearly does not mean that a theory is pre-imposed on the data in order to be ‘proved’.² But it simply means that all contextual (i.e. extra-textual) knowledge is a valuable resource to better understand the text, to find and refine good questions and search for answers. The idea behind it is that ‘there is room in linguistics for many approaches and combination of approaches’ (Partington, 2009: 300)

1.2.2. Discourse Studies

Discourse is notoriously an elusive concept, not only because it is defined in different ways across disciplines, but also because it is used in different ways within them. In linguistics, discourse tends to be defined in two alternative—albeit not mutually exclusive—ways: a structural one, seeing discourse in terms of units of language and a functional one, seeing it in terms of its use. An example of a structural definition of discourse is ‘language above the sentence or above the clause’ (Stubbs, 1983: 1) and a classic example of a functional one is ‘language that is doing some job in some context’ (Halliday, 1985: 10).

Beyond linguistics, discourse is central to much research in the social sciences and humanities, where Foucault’s famous definition is usually adopted: ‘practices that form the object of which they speak’ (1972: 49). Foucault himself, however, acknowledges attributing various meanings to the term:

instead of gradually reducing the rather fluctuating meaning of the word “discourse” I believe I have in fact added to its meanings: treating it as sometimes the general domain of all statements, sometimes as an individualisable group of statements and sometimes as a regulated practice that accounts for a number of statements.

(Foucault, 1972: 80)

Hall offers a useful explanation of Foucault’s discourse as ‘a group of statements which provide a language for talking about—a way of representing the knowledge about—a particular topic at a particular historical moment… . Discourse is about the...

Table of contents

• Cover
• Half Title
• Series
• Title
• Copyright
• Contents
• Acknowledgements
• 1 Introduction
• 2 Approaches to News Discourse and Journalism
• 3 Method, Corpus, Process
• 4 Newsworkers
• 5 News-Trade Taxonomy
• 6 Over There at The Guardian
• 7 Here at The Guardian
• 8 Epilogue
• References
• Appendix 1
• Appendix 2
• Index