When discourse analysts tell stories: what do we ‘do’ when we use narrative as a resource to critically analyse discourse?

ABSTRACT Critical discourse analysts are being pulled in two directions. On one side, in the age of validity, inter-rater reliability and evidence-based research, it can seem subversive when researchers ‘tell stories’ (rather than ‘write reports’, ‘produce findings’ or ‘demonstrate effectiveness’). On the other side, public relations departments encourage researchers to use ‘storytelling’ techniques to engage public audiences. In this paper, I draw on social and cultural theory to assume that critical discourse analyses are always already narrative. I propose that we embrace (rather than hide) the storytelling qualities of our research. Using similar strategies for all audiences, we undermine traditional power hierarchies between researchers and ‘the public’. Critical discourse analysis then performs its own critical approach. The paper also explores the potential of adopting a ‘carrier bag theory’ of storytelling (Le Guin) rather than telling ‘hero’ stories and reflects on the epistemological, relational and political work done by storytellers in critical discourse studies.


Introduction
In a recent questionnaire by the funding body of a research project, I was asked to give my project a score from 1 to 10 on the extent to which it produces 'evidence-based findings'. I believe that our project most certainly produces findings that are based on evidence. Our evidence is qualitative, critical and narrative. I suspect, however, that the funding body means something elseand something more scientisticwith the phrase 'evidence-based'.
Scholars of critical discourse studies (CDS), and most other social sciences, are today perched on the rope in a tug-of-war. Pulling in one direction are those reviewers and funders in favour of scientism (Validity, Inter-Rater Reliability, Evidence-Based-Research). These colleagues expect us to 'write reports', 'produce robust findings' and 'demonstrate' the 'impact' or 'effectiveness' of interventions. They also encourage us to use big data, or at the very least to quantify the scope of the discourse we are analysing. To 'tell stories' seem subversive, anecdotal, unscientific. Pulling in the other direction are our colleagues in the public relations department, encouraging researchers to use storytelling techniques to engage public audiences. Promoters of TED talks pull us in this direction. Over here, telling stories is about simplifying complex research for 'the public', who, it is assumed, are less able or less willing to engage with complex research findings.
If funders, reviewers and others continue to pit research against storytelling, then qualitative, critical and narrative 'evidence' loses legitimacy. Scholars' strategies of embracing and/or avoiding stories help to police the boundaries of knowledge. In this paper, I focus on what it means for CDS if we assume that reports, papers and articles are always already stories. The paper reflects on how discourse analysts tell stories and explore what happens if we tell 'carrier bag' stories rather than 'hero' stories. I suggest that storytelling in general, and carrier bag stories in particular, are well suited to the epistemological, relational and political work done by researchers who aim to critically study discourse, i.e. the ways in which language and other semiotic practices constitute knowledge, subjects and objects. 1 In this understanding, using storytelling in critical discourse analysis is a critical analytical strategy (as well as perhaps making more papers more enjoyable to read and to write). Overall, I suggest that especially in times of big data and computational methods, qualitative discourse analysts are redefining their (our) research. One approach is embracing, rather than hiding, the story-telling quality of this kind of in-depth, focused, critical research. In this way, we repurpose the rope, taking it out of a tug-of-war game and using it instead to weave stories together. As we do this, we also re-evaluate the hierarchical boundaries between scholars/researchers/funders (for whom we are supposed to write reports) and the public (to whom we are supposed to tell stories).

Knowledge and storytelling
Many scholars have investigated narrative, identifying key structural elements and highlighting the interactional practices of storytelling (see Forchtner's introduction to this special issue). I take a broad-stroke approach, in which 'narrative' and 'story' refer at a very basic level to the recounting of a series of events. Stories often involve beginning, middle and end. More specifically, a story is the recounting of a series of events involving movement or change. Science fiction writer Ursula K Le Guin, a key thinker for this paper, defines a story as 'a narrative of events (external or psychological) that moves through time or implies the passage of time and that involves change (Le Guin, 1998, p. 122f.; see also Genette, 1988, p. 19;Linde, 1993). For Le Guin, change is the 'universal aspect' of story: 'Story is something moving, something happening, something or somebody changing' (Le Guin, 1998, p. 122f.).
(Social) scientific writing very often describes change; science draws on grand and small narratives for, at the very least, legitimation (Lyotard, 1979). However, modern (social) scientists have created a conflict between 'science' and 'narrative'. Scholars such as Shirley Brice Heath (1983) and James Paul Gee (1990) have pointed to the classed and racialised impact of this apparent conflict. Gee, for instance, describes how a black working-class student became alienated from school when her teachers devalued the stories that she contributed to classroom talk. The teachers in turn had been taught to value explanatory language, stripped of story elements (Delafield, 1999;Gee, 1990).
Yet longer pieces of writing, or presentations, will invariably include some movement, some change. The classic IMRAD structure of empirical research papers has a clearly narrative structure, with beginning (Introduction), middle (Methods, Results) and end (Discussion). Mieke Bal observers that '[n]ot everything is narrative; but practically everything in culture has a narrative aspect to it, or at least can be perceived, interpreted, as narrative', including lawsuits, philosophical discourse and argumentation (Bal, 2017, p. xix). By learning that description, argumentation, exposition etc. are opposed to, rather than entwined with, story, students and emerging scholars learn to hide the narrativity of their academic writing.
In Learn to Write Badly: How to Succeed in the Social Sciences, Michael Billig sees the central problem in academic writing as the way social scientists take people out of social science. They (we) write about things (abstractions) instead of about people and actions. They favour the passive voice and use noun phrases to turn the things people do into abstract entities ('the ifications and the izations', like reification, mediatisation, nominalisation, Billig, 2013, p. 7). Social scientists have made certain validity criteria, which emerged in a particular historical time, into the foundations for evaluating knowledge production, rather than exploring open-ended, context-sensitive or 'transgressive' validity criteria which decentre the authority/privilege of the researcher (see Lather, 1993). In this process, we have learnt to write not only people in general, but also our own histories/stories out of our research publications.
There are, Billig reminds us, (economised) reasons social scientists have tended to write themselves out of their academic texts, especially as higher education is increasingly commercialised (Fairclough, 1995;Sellar, 2013). And it is the actions of people in the academic system that invites and encourages us to 'write badly' by establishing today's reward system. This includes the questionnaire that I mentioned above, asking me to evaluate my project's ability to produce evidence-based findings, and it includes the reputational status for someone who successfully coins and markets a phrase that captures a specific societal phenomenon. In abstract terms: this is the operational capitalism of research.
Billig wants us to recognise that, within this system, we are 'hacks who write for a living ' (2013, p. 12). This phrase shocked me when I read it first: is that all I am? What is our role as hacks who write for a living in marketized universities, which vie for status in the UK's REF, Germany's excellence initiatives, or global ranking systems? What kind of writing are we doing? What kind of commodity are we producing? And for whom? Then I realised that foregrounding the writing, rather than, for instance, data generation or analysis, can be a liberating perspective. The discussion returns to this in more detail. It is hinted at by Le Guin, writing in a different context: We need writers who know the difference between production of a market commodity and the practice of an art. Developing written material to suit sales strategies in order to maximise corporate profit and advertising revenue is not the same thing as responsible book publishing or authorship. (Le Guin, 2016, p. 113) Responsible publishing and authorship arguably involve positioning oneself in the writing. This article builds on the now-classic insight that knowledge is socially and culturally produced within conventions. For poststructuralist, post-qualitative, social constructionist or reflexive critical discourse analysts, knowledge work is located in the stories/histories that we bring with us to our analytical standpoint (e.g. Butler & Scott, 1992;Foucault, 1970;Haraway, 1988). For Donna Haraway, disowning the histories we bring with us to our research is one of the 'cheap tricks' of puritanical critique (Haraway, 2003, p. 89). Mary Bucholtz describes instead a 'discourse analysis that is both critical and reflexive', in which 'the analyst's choices at every step in the research process are visible as part of the discourse under investigation, and critique does not stop with social processes, whether macro-level or micro-level, but rather extends to the analysis itself' (Bucholtz, 2001, p. 166). Making the analyst's choices visible unfolds the change and movement in the story being told.
The work I have cited thus far suggests that critical discourse analysis should foreground rather than hide its storytelling features. If the overarching goal of social science is to understand human society, and if we assume that humans are storytelling beings (homo narrans), and if we agree that storytelling is becoming increasingly important across political and social media today, then 'stories as tools for making sense of human action in situated social realities are today more important than ever' (Browse et al., 2019, p. 245). Scholars have reflected on how (digital) storytelling can generate empathy and trust among readers, viewers or listeners by bringing people back in, including unheard voices, giving meaning to human action and enabling emotional engagement with the story or characters (Bourbonnais & Michaud, 2018;Liguori, 2019).
There is, however, a danger inherent in bringing people back into social science. In her 'Carrier Bag Theory of Fiction', Le Guin describes how easily a peopled story can turn into a hero's story. She sets her reflection in temperate and tropical regions in Paleolithic and Neolithic times, when the principle food stuffs were gathered rather than hunted. I quote at length: [W]hat we actually did to stay alive and fat was gather seeds, roots, sprouts, shoots, leaves, nuts, berries, fruits, and grains, adding bugs and mollusks and netting or snaring birds, fish, rats, rabbits, and other tuskless small fry to up the protein. And we didn't even work hard at it […]. The average prehistoric person could make a nice living in about a fifteen-hour work week.
Fifteen hours a week for subsistence leaves a lot of time for other things. So much time that maybe the restless ones who didn't have a baby around to enliven their life, or skill in making or cooking or singing, or very interesting thoughts to think, decided to slope off and hunt mammoths. The skillful hunters then would come staggering back with a load of meat, a lot of ivory, and a story. It wasn't the meat that made the difference. It was the story.
It is hard to tell a really gripping tale of how I wrested a wild-oat seed from its husk, and then another, and then another, and then another, and then another, and then I scratched my gnat bites, and Ool said something funny, and we went to the creek and got a drink and watched newts for a while, and then I found another patch of oats. … No, it does not compare, it cannot compete with how I thrust my spear deep into the titanic hairy flank while Oob, impaled on one huge sweeping tusk, writhed screaming, and blood spouted everywhere in crimson torrents, and Boob was crushed to jelly when the mammoth fell on him as I shot my unerring arrow straight through eye to brain.
That story not only has Action, it has a Hero. Heroes are powerful. Before you know it, the men and women in the wild-oat patch and their kids and the skills of the makers and the thoughts of the thoughtful and the songs of the singers are all part of it, have all been pressed into service in the tale of the Hero. But it isn't their story. It's his. (Le Guin, 1989, p. 165f.) It is the action heroes, the mammoth hunters, who 'spectacularly occupy the cave wall and the mind' (Le Guin, 1989, p. 165). Several theories of story or narrative situate 'conflict' or 'disruption' at the core. Overviews suggest that 'in almost every narrative of any interest, there is a conflict in which power is at stake. You might say that conflict structures narrative' (Abbott, 2014, p. 55). Guidebooks encourage us to write heroic journeys involving conflict and resolution, since these stand at the core of successful storytelling (Leitman, 2015;Olson, 2015).
While it is certainly true that many gripping Hollywood, fairy tale or other narratives are structured around conflict, Le Guin draws our attention to alternatives. Rather than focusing on the hero and his spears and arrows, Le Guin wants to tell the story of the container, the sling or net carrier, that holds the gathered products, the 'seeds, roots, sprouts, shoots, leaves, nuts, berries, fruits, and grains', the 'bugs and mollusks', the 'birds, fish, rats, rabbits, and other tuskless small fry'. As 'the thing to put things in, the container for the thing contained' (Le Guin, 1989, p. 167), the carrier bag image carries (at least) two sets of associations. First, the bag holds a myriad of different everyday things, and these everyday, apparently banal things are precisely the things that hold us together, that are worth looking at more closely, and that may gently surprise us. Second, the carrier bag image alerts us to the material out of which the bag is made. The threads weave stories together into textured fabric. A net carrier bag suggests a network of stories, connecting to one another. Focusing on the hero pushes 'the men and women in the wild-oat patch and their kids and the skills of the makers and the thoughts of the thoughtful and the songs of the singers' out of the picture.
The carrier bag also provides a way to avoid 'the danger of a single story' that Chimamanda Ngozi Adichie (2009) has highlighted. In the carrier bag theory of storytelling, there are always multiple stories, multiple truths, multiple perspectives; complexity. Complexity makes it 'impossible for the bottom line to be one single statement' (Haraway, 2000, p. 105; see Billig, 2013, p. 97).
'Discourse analysis, writes Ruth Wodak, 'stems etymologically from the Greek verb ana-lyein "deconstruct" and the Latin verb discurrere "running back and forth"' (Wodak, 2006, p. 596). 'To write is to carve a new path through the territory of the imagination, or to point out new features on a familiar route' (Solnit, 2001, p. 72). Linking back to Billig's (2013) critique of social scientists' love of abstractions, and Le Guin's (1998) understanding of story as change and movement: Stories trace the 'lines' through time and space that brought us to our current location, allowing the storyteller to foreground actions, processes and change rather than static results or abstractions. (Ingold, 2007) How critical analysts of discourse tell stories Reflecting on our role as hacks who write for a living, the carrier bag theory of writing alerts us to two different kinds of stories that discourse analysts can tell: action tales and carrier bag stories. 2 For this paper, I have selected a set of critical analysts of discourse who are currently telling these kinds of stories to provide powerful analyses of language, society and power. The subheading for this section is deliberate: I am reflecting hereinterand transdisciplinarilyon people from various fields who critically analyse discourse, not only those who see themselves as critical discourse analysts. The distinction between action tales and carrier bag stories is not set in stone, there will be overlap. My point in this section is to highlight elements of these two kinds of story. The subsequent section returns to the title question, considering the implications of what do we 'do' when we use narrative as a resource for critique.
Narrating stories around conflict and resolution: action tales 'You often hear, with regard to the alt-right or the Intellectual Dark Web or pro-Trump nationalists, that the way to avoid normalising them is to avoid responding to them, or to only respond by calling them offensive and terrible and bad', said Natalie Wynn in an interview with the New Yorker (Marantz, 2018). 'And, look, sometimes they are offensive and terrible and bad, but you don't win by saying that. You win by pointing out why they're wrong, and by making better propaganda than they do.' Wynn's YouTube persona, ContraPoints, uses seduction, parody, lavish costumes, and decadent setdesigns to critically analyse far-right discourse. She weaves a story, with herself as central narrator and an external disruption upon which she reflects, to perform this critique.
With a background in philosophy and a sprinkling of Foucault across her videos, Contra-Points engages intensively and critically with right-wing memes and issues such as incels, racism and cultural appropriation. She has been described as 'an elegant, whip-smart middle finger to the putrefying swamps of the internet' (Cross, 2018) and as bringing 'a rare skill for rational argument and emotional persuasion to one of the most vicious battlefields of the online culture wars' (Mark, 2019).
She engages, for instance, with the public intellectual Jordan Peterson's use of the phrase 'post-modern neo-Marxism' (ContraPoints, 2018). He uses this to, among other things, equate trans activists with Stalin or Mao. ContraPoints's concern is how those who follow his reactionary political agenda may target some of society's most vulnerable people, with effects on gender equality, LGBT acceptance and civil rights. She picks apart the incompatibility of the phrase itself: If Marxism is a grand narrative about humanity, and post-modernism is the very questioning of grand narratives, what can 'postmodern neo-Marxism' mean? She observes how Peterson uses the term 'to characterize the left as a unified philosophical force bent on destroying western civilization, when in fact it's a bunch of bumbling buffoons who can't stop squabbling with one another over every goddam little issue' (minute 19:45). The left only appears to be unified, she reflects, to those who are 'so far to the right that literally everyone who supports the economic and social advancement of disadvantaged groups looks like one homogeneous enemy' (minute 20:05).
With examples from Cathy Newman's interview with Peterson on Channel 4 News, Con-traPoints analyses Peterson's rhetorical strategy: He says 'something more or less uncontroversially true, while at the same time implying something controversial. For instance, he will make a claim like there are biological differences between men and women, which is obviously true. But he'll say it in the context of a conversation about the under-representation of women in government, which implies? What exactly?' (minute 20:37). In Contra-Points' analysis, it is incredibly difficult to respond to this. One can either fall into the trap of arguing against something uncontroversially true or guess at what he is implying, which allows him to claim that you are misrepresenting him.
ContraPoints is one expression of everything that Billig recommends social scientists do when they communicate with other social scientists: She brings people in, uses verbs and focuses on practices and relations rather than abstractions. In the Peterson video, she explains philosophy, modernity, Marxism, power, etc., interrupting herself halfway through to say, 'this is a lot of explaining; it's so much explaining, it's triggering gender dysphoria' (minute 14:05). Yet she uses the explanations to critique social injustice; they are an intrinsic part of the performance. She draws on classic narrative structures; her story has a clear disruptive force (Peterson), and ends with a new equilibrium. But she plays with this. She does not end with her victory. Instead, she wraps up care for social justice in a narrative of the decadent, aloof, ironic trans: 'I don't care either way. I make YouTube videos because I enjoy mood lighting and set design. So, what do you people want from me?!' (minute 26:28).
ContraPoints is only one example of a strong narrator figure who critically analyses discourse through a performance of herself. In another show, NewsBroke, Matt Lieb performs as the central narrator. That he is Jewish is part of his story. Like ContraPoints, he also speaks directly to camera, as he describes, for instance, 'how we are now somehow living in a world where people can love Israel and hate Jews at the exact same time'. He tells this apparent contradiction as a horror story. Beginning with clips showing how the USA has 'become a bit Hitlery', the video continues with Lieb's horror at the farright's support for Israel. He recounts, for instance, how they welcome a future in which all Jews have returned to Israel, how they see Israeli politics as evidence of the effectiveness of walls and profiling, and that the reason the far-right loves Israel seems to be because a key character in his story, Steve Bannon, 'doesn't hate us as much he hates the Muslims' (NewsBroke, 2018). Lieb weaves his critical analysis of today's far right into a strong narrative around his own Jewishness and connects the horror story that he is experiencing at this moment to historical figures (Hitler).
These are not stereotypical heroic journeys. Performers like ContraPoints or Matt Lieb use audio-visual resources to narrate themselves as (flawed and relatable) protagonists in their own action stories. I root for these heroes. Performative social science has used similar techniques from drama and art since the early days of the crisis of representation (Denzin, 1997;Yallop et al., 2008). These scholars have reflected on how personal narratives, i.e. narratives in which the narrator shares something from their personal lives, open up a space for readers/viewers to relate to the stories (Gergen & Jones, 2008). This relatability in turn opens the question for a 'performative critical discourse analysis' of how to maintain a balance between scholarly and aesthetic standards.
Other analysts focus more broadly on 'something moving, something happening, something or somebody changing' (Le Guin, 1998, p. 122f.). With this emphasis on 'change', I can think of no critical discourse analyses that are not stories. Some focus on conflict; they have villains and heroes, or at least, characters on different sides of a struggle for discursiveand thus political/cultural/socialdomination. Chen and Flowerdew (2019), for instance, analyse the umbrella movement protests in Hong Kong in 2014: They analyse how two opposing sides of the protracted conflict in Hong Kong comment on YouTube video clips of the protests. Their analysis of 4329 comments made by 2157 posters begins with historical analysis. The authors weave a story of what they select as important aspects of the history of Hong Kong and its relationship to mainland China. They write of 'humiliation', 'vulnerability' and 'pride', key affective issues that give meaning to this socio-historical context. We meet characters such as 'local people', 'Hongkongers', 'Mainland immigrants', 'Mainlanders', 'Beijing' and 'Chinese loyalists'. Chen and Flowerdew's focus is clear throughout: the article is about discriminatory discursive strategies; it is about Hongkongers and Mainlanders who feel strongly enough about the protests to post comments online; and it is about the changing relationship between Mainland China and Hong Kong.
I could describe Chen and Flowerdew's article as a report, as findings, as evidence; the authors say 'taxonomy'. But I choose here to focus on its storytelling. The article follows a trajectory, 'the shape of the story as a whole' which 'moves always to its end, and its end is implied in its beginning' (Le Guin, 1998, p. 124). By unfolding this trajectory, Chen and Flowerdew help us to understand the everyday discursive strategies that underwrite a socio-political conflict that was exploding as their text was published.

Tracing trajectories of change: carrier bag stories
Many well-known stories (from novels to Hollywood) revolve around conflict and resolution. Guidebooks for writers have often described this story structure as the key to gripping storytelling (see above). But, as Le Guin has observed above, it is 'hard to tell a really gripping tale of how I wrested a wild-oat seed from its husk, and then another, and then another, and then another, and then another, and then I scratched my gnat bites, and Ool said something funny, and we went to the creek and got a drink and watched newts for a while, and then I found another patch of oats'. Contesting the reduction of all successful stories to conflict, Le Guin situates conflict as just one kind of behaviour that sits alongside other behaviours that are 'equally important in any human life, such as relating, finding, losing, bearing, discovering, parting, changing' (Le Guin, 1998, p. 123). We do not need a rigid plot structure to tell a story, Le Guin continues, but we need a focus: What is the story about, who is it about and what is changing or has changed? 'Imagine that you're on holiday in Eastern Germany', begins John E. Richardson's book on British Fascism: A Discourse-Historical Analysis. 'In the rural state of Mecklenburg-Western Pomerania, to be precise. You decide to take the car out for a dayto drive around the region you're staying in and take in the sights. Perhaps you'll discover a nice place and get out the car and have a wander; maybe have lunch and something to drink. In any case, it is a nice day and you and your companion are not in any rush to be anywhere, so you decide to go and explore.' You happen upon a wooden signpost, with signs pointing towards to European towns and cities like Paris and Berlin. And you notice Breslau, Königsberg, Braunau am Inn. Why, asks Richardson, 'open a book on British fascist discourse with a fictionalised account of a holiday in Eastern Germany?' (Richardson, 2017a, p. 13f.) Because opening with a fictional account, based in a specific time and place, tells the reader what the book is about, who is it about and what is changing. The account enables Richardson to reflect on key issues at stake in fascist discourse: Who says what, with which code, under whose radar? That Braunau am Inn is Hitler's birthplace gives us a clue. The signpost sets a scene; it lets the reader visualise what we will find in the book, as the discourse analyst unpacks the trails, the paths, the relations and the things that people have done to add additional, highly politicised, meanings to apparently mundane words. It also performs that which it analyses: The scene is sunny, relaxed, pleasant. The holidaying characters are unaware that something sinister lurksas sinister things often do in horror storiesin hidden corners of the pleasant image.
Where Richardson uses fictional elements to set the backdrop for his analysis of British fascism, Peter Clough (2002) presents his entire analysis through composite stories about 'Klaus', 'Molly', 'Rob', 'Bev' and 'Lolly'. The names are not pseudonyms for Clough's research participants. He has woven each character from real details and 'symbolic equivalents' (a concept he draws from Yalom, 1991). These are 'stories which could be true, they derive from real events and feelings and conversations, but they are ultimately fictions' (Clough, 2002, p. 9, italics in original). Reading the sometimes very poignant stories, I felt that Clough achieved his goal of using stories to 'allow the report of those experiences which might otherwise not be made public by other 'traditional' tools of the trade' (2002, p. 9). 'Klaus' shows more intensely than any other kind of research on special educational needs how damaged fathers in turn damage their sons, inadvertently perhaps. 'Bev' shows how staff cope, or do not cope, in special educational needs departments behind the discourse of marketisation, efficiency and effectiveness. 'Molly' complicates the discourse about school exclusion.
In each of these cases, Clough's fictional narratives are compelling. Power is at stake, but the narrative is not structured around a heroic journey. For Clough, narrative is useful 'only to the extent that it opens up (to its audiences) a deeper view of life in familiar contexts'; his stories 'make the familiar strange, and the strange familiar' (Clough, 2002, p. 8).
'Social science fiction' is, like science fiction, set in the future. Peter Frase (2016), for instance, titles his book Four Futures: Life after Capitalism. Heidrun Allert, Anne Bruch and I locate ourselves in the future and look 'back' at the 2020s to write three fictive histories of education and technology in that era (Macgilchrist et al., 2020). Costello et al. (2020) reflect on the post-pandemic world. These 'speculative fictions' draw on elements of science fiction alongside insights from social science research. Rooted in data generated about contemporary discursive formations and/or how socio-economic structures shape the decisions people makeabout pay, relationships, education, politics, rent, climate justice, intellectual property, etc.the authors speculate on what kinds of different decisions humans around the world could make, and what futures these decisions make possible.
All three texts are, at core, critical analyses of contemporary discourse. All three are also 'generative' critical analysis, i.e. they draw on critical theories to generate new meanings and spaces of possibility (Haraway, 2000;Sedgwick, 2003). Stories arising from what some label critical discourse studies and others 'positive discourse analysis' enact an 'affirmative politics' that 'combines critique with creativity in the pursuit of alternative visions and projects' (Braidotti, 2013, p. 54; see also Bartlett, 2012;Hughes, 2018;Macgilchrist, 2016). Social science fiction showcases (nonfictional) discourse currently opening up utopian possibilities. They understand the futures they describe 'not just as ideals or self-contained utopias, but as the objects of dynamic and ongoing political projects' (Frase, 2016, p. 145).
In one of our three future scenarios we mention, for instance, 'convivial technologies', a concept in the discourse of degrowth. Convivial technologies are tools which are not designed with novelty and profitability in mind, but with a view to a more equitable, participatory, democratic, interrelated and ecologically sustainable world (Vetter, 2018). We imagine how today's apparently self-evident discourse about educational technology (edtech) might changeaway from marketisation and corporate take-overif schools and educational policymakers prioritise convivial edtech by, e.g. using open source software that students can (collaboratively) build, repair and adapt. We describe projects which see data privacy as a human right, and data ownership as intrinsic to web applications, striving against data monopolies and corporate platform centralisation. In the 2020s, we speculate, these successful interoperable platforms, heavily funded by the EU, would fulfil their promise of decentralising the power of surveillance capitalism (Zuboff, 2019).
Other analyses of discourse explore Le Guin's 'relating, finding, losing, bearing, discovering, parting', without resorting to fiction. Again, I am taking a recent piece that could be described as a report, as evidence; the authors say 'project' and 'article'. Seargeant and Tagg (2019) orient to what may be the biggest carrier bag of our day, where we scratch our gnat bites, and Ool says something funny, and we watch newts for a while: Facebook. Seargeant and Tagg steer their readers in a trajectory: First, we see how Facebook's algorithmic power manipulates users into filter bubbles. Then we begin to see how users' online actions contribute to the filter bubble effect (Seargeant & Tagg, 2019, p. 41). In contrast to contemporary work on social media or networked sociality that offers technological explanations for how the world is changing, Seargeant and Tagg bring people back into the story, as agents who co-design the contexts in which they communicate. This is not a heroic story in which we follow one main protagonist who has more power than Facebook, or the heroic analysts who struggle against corporate power (see Olson, 2015, p. 189). It is a carrier bag story of the everyday, of many people moving through time, and contributing to social change.
Perhaps the most obvious place in CDS to look for peopled storiesand, among these, especially for carrier bag storiesis at the nexus of ethnography and discourse analysis. When your own body is a research instrument, it can seem disingenuous to write yourself out of the story. In ethnographic work, discourse analysts tend to bring themselves into the trajectories of change they are writing about (Mielke, 2020;Richardson, 2017b;Wodak, 2009). Germán Canale (2019), for instance, begins his analysis of the semiotic processes through which 'policy key terms' are used to 'achieve cultural, socio-political and ideological cohesion in society by articulating meanings at different scales' by situating himself in his own school classroom in Uruguay. We can guess his age, estimate his socio-economic background, and visualise his younger self walking with classmates and teacher through the school corridors to the computer lab. We see them sitting down at the 'four or five huge wooden desks with three or four desktop computers on each of them, most of them with black and white screens' (Canale, 2019, p. 1). Canale is present in the story as he introduces us to the two educational commodities that are his focus: technology and English as a Foreign Language (EFL). His are, I suggest, carrier bag stories of everyday life in schools. His storytelling arc makes this everyday life speak to far broader issues of how, for instance, 'policy-making and policy enactment achieve cultural, socio-political and ideological cohesion in society' (2019, p. 3).
In this section, I have cherry-picked recent analyses that I enjoyed in order to illustrate my own story-focus on ways of using narrative to critically analyse discourse. Some of these are audio-visual performances of the self. Some use explicit storytelling strategies like science fiction or horror. Some tell compelling stories of relatable central protagonists journeying through conflict, change and resolution. Others are peopled, but not heroic. They move through time and involve change for a complex cast of more mundane, everyday, often silenced, characters. I want to emphasise here that when I describe these analyses as stories, I do not mean they are only stories, but that they are also stories. These critical discourse analyses embed a critical analysis of language, society and power within their stories.
Discussion: what do critical discourse analysts do when we use narrative as a resource for critique Critical discourse analysts have, as noted above, a rich tradition of acknowledging that we analyse from particular situated, mobile, complex, contradictory standpoints. It is generally agreed that our analyses do not only analyse worlds, they also make worlds (Goodman, 1978). There has been, however, far less theoretical reflection in CDS on the narrative elements of writing discourse analysis. In this section, I return to the title question, to reflect on what we are doing when we use narrative as a resource for critique. Driven by theories of storytelling and writing, I have assumed in this paper that we, as discourse analysts, always already tell stories. We do not 'find' data; 'findings' is a misnomer. We 'generate' data, 'create' connections, and 'tell' stories. This section discusses ways in which storytelling in CDS supports the overarching epistemological, relational and political goals often associated with CDS.

Epistemological work
Firstly, we are doing epistemological work. Narrating critical analysis through performing the self, or creating fiction, including speculative and social science fiction, interrupts dominant theories of knowledge, in which discourse analysts 'present findings' for scholarly audiences, and 'tell stories' for public engagement events. When we do not simplify our papers for a lay audience, nor complexify our language for a scholarly audience, but use similar strategies and styles for all audiences, we undermine hierarchical boundaries that have been drawn between researchers and 'the public'. Critical discourse analysis in this vein visibly performs its own critical approach.
ContraPoints, for instance, embeds discourse theory into a critical analysis that entertains and does not simplify. Richardson's signpost scene and Clough's composite stories are carrier bag stories of the everyday. They show the 'container', the background context within which individuals can appear as heroes; and they show 'the thing contained', the plethora of everyday, non-heroic characters, the 'beings and situations formerly relegated to the margins; the trivial events of daily life or the bruteness of a real that does not permit of inclusion' (Rancière, 2020, p. 9). Such stories often include aspects of the narrator's own life. Many scholars, however, feel uncomfortable bringing themselves into their writing. Even when colleagues in discourse studies encourage their students to embrace positionality, students are concerned about being 'too subjective'. I have also heard this from my students; their undergraduate degrees are still teaching them to value 'objectivity'. Similarly, PhD researchers have told me that their supervisors discourage them from writing in the first person: They tell them it is not scientific.
Perhaps the concern is that when we embrace storytelling, we particularise rather than categorise (Billig, 1996). Explicit storytelling elements thus bring with them different understandings of 'validity' in social science research. The call for proposals from the funding body I mentioned in the introduction was met with excitement by many qualitative or post-qualitative researchers that I know. It was the first call by a major funding body which expressly said the projects should generate insights which are 'transferable' to other settings, rather than emphasising the need for 'robust', 'generalisable' findings. The team that co-wrote the call were attentive to moves towards alternative validities. The team that updated the questionnaire and sent it to the successful projects were not. The questionnaire had been used in previous calls in which generalisability was the core marker of validity. One close colleague of mine was furious about the mismatch between the research that had been funded and the criteria in the questionnaire. This mismatch indicates the slow change in how the boundaries of epistemological legitimation are marked and policed. The point of stories is not to generalise, not to quantify, not to provide evidence. The point is to dig deep or to broaden 'the horizon of gazes at, and thoughts about, what gets called a world and the ways of inhabiting it' (Rancière, 2020, p. 9).

Relational work
This means we are also, secondly, doing relational work when we embrace (rather than hide) the storytelling qualities of our research. Stories connect things to other things, people to places, and events to histories. They also connect writers to readers. Stories expand communities of practices by providing insights on specific, local settings, and making them transferable to readers' own settings. Stories also provoke scholarly communities. The reviewers of social science fiction for a special issue of a 'regular' journal had concerns about the 'rigour' of the methodologies. For social scientists used to quantitative research, it is easy to imagine their unease: how to assess the 'robustness' of these findings? For social scientists used to qualitative research, the questions arise of which contribute the piece makes beyond just presenting vignettes: Can it speak to broader issues and thus push the critical conversation forward more than an observational or interview study could?
When we embrace, rather than hide the storytelling inherent in academic work, we thus prod conventional power hierarchies. This is, it goes almost without saying, dangerous for precariously employed early career scholars, for those aiming for tenure, or orienting to the unknown peer reviewers behind funding bodies. As Billig (2013) has pointed out, particularising rather than generalisingspecifying the particular rather than building a new school of scholarly analysisis not the easiest way to increase one's citation scores. School-building is heroic work, not carrier bag work (for the gendered implications of this, see Russ, 2018).
If we are, as Billig has said, hacks who write for a living, then we can write so that readers enjoy reading and we can write so that we enjoy writing. This insight I found liberating. Critical discourse analysts spend a great deal of time organising and participating in methods workshops. We pay far less attention to writing workshops. Hacks who write for a living are constantly making decisions of how to include, exclude, frame, foreground, code, decode, symbolise and visualise details to create compelling narratives. Hacks are attentive to how they relate to their readers. Discourse analytical hacks make their texts speak to broad societal issues, clarifying their contribution to the discourse community's conversation. Our 'products' may still be 'academic words' (Billig, 2013, p. 13), but we can choose to embed our academic words into compelling stories.

Political work
This epistemological and relational work is, thirdly, also political work. In a study on public deliberations, Polletta and Lee (2006) observed that participants used personal storytelling to voice a marginalised position. Marginalised speakers used stories to contest positions that seem to be universal but are actually the experiences of powerful dominant groups. Stories, like Klaus', Bev's or Molly's, have 'the capacity to reveal the narrow character of ostensibly shared values and neutral principles' (Polletta & Lee, 2006, p. 712; see also Liguori, 2019). This worked powerfully when participants told personal stories about cultural and memorial issues. Their stories were compelling, and they gained a hearing. But the participants were more cautious about using stories to discuss policy and finance. Storytelling worked for them as a powerful way to put forward counter-discourses, but primarily limited to the fields of culture, memory, etc. The political work of CDS thus includes telling stories in other fields, e.g. about policy-relevant issues.
At the same time, we need to do some reflexive work on the politics of our own stories. As the recognition grows that people are storytelling beings, this in turn becomes a powerful discourse. It can be used to silence other ways of being, engaging, speaking, writing. It is thus important to reflect on the politics of storytelling as a way of writing culture (Clifford & Marcus, 1986), especially since storytelling explicitly aims to be compelling, or even to activate emotions. When we tell stories from our specific location in socio-material-economic-ecological-affective spaces, we create presents, pasts, futures and possibly change. Classic analysts of narrative (e.g. Bourdieu, 1990;Linde, 1993;Todorov, 1969;White, 1973) have pointed to the tendency to make stories 'fit' a preconceived story-structure. Narratives are often made to fit a coherent, linear flow, with the storyteller smoothing over ruptures. More broadly, it is tempting to render the socio-culturally embedded process of crafting stories invisible as one crafts that story. Perhaps the carrier bag theory of writing discourse analysis opens up space to reflect on these tendencies.

Concluding thoughts
We were invited to write theory-driven papers in this special issue. In this paper, I draw on generatively critical theories (from within and outwith discourse studies) that suggest critical discourse analytical reports, papers and articles are always already also stories. When I proposed this paper for the special issue on narrative and critical discourse studies, I suggested the title, 'Narrative as an epistemological resource in critical discourse studies'. Then I thought about how my own writing performs (or does not) what I argue. The heavy nouns, the abstractions, and the lack of people in that title, and in an early draft of this paper, point to several of the ways I have 'learnt to write badly' (Billig, 2013). Comparing that to the current title, 'When discourse analysts tell stories: What do we 'do' when we use narrative as a resource to critically analyse discourse?', (hopefully) shows my process of learning to write better in this article, where 'better' means telling a story about people, doings, movement and change, where 'better' is still far from perfect, but where 'better' is, 'by definition, ideological' with 'the advantage of being overtly ideological' (Canale, 2019, p. 4).
The stories I have described show frustration (e.g. Clough), horror (e.g. NewsBroke), danger (e.g. Richardson) and hope (e.g. social science fiction). They invite readers to feel the destructive power of emerging discourse or to imagine a more liveable future after we help affirmative discourses to take over. 'Hard times are coming', writes Le Guin, 'when we'll be wanting the voices of writers who can see alternatives to how we live now, can see through our fear-stricken society and its obsessive technologies to other ways of being, and even imagine real grounds for hope' (Le Guin, 2016, p. 113).
While it is important to talk about people and actions rather than ifications and izations, I have suggested in this paper that this is only one step. If we focus on people and actions, we will tend towards stories of heroic journeys. The club-wielding hunter becomes the star. Alongside heroic action story structures, I have foregrounded carrier bag stories in this paper, stories of the container-context, of the apparently mundane, banal everyday gatherers, and of the things contained in the containers. 'It matters what stories we tell to tell other stories with' (Haraway, 2016, p. 118). These gatherers are not the autonomous human agents of the Enlightenment; we can tell stories of how their agency is connected to, e.g. the technical (algorithms), the material (computer labs; wooden signposts; mood lighting and set design), the economic (tenure, impact).
Carrier bag stories are not necessarily strategies for building a theoretical school or branding an approach. Instead they are ways of analysing and making worlds. And as they make worlds, they redefine how discourse analysis can produce socially relevant knowledge. Alongside newer computational methods and bigger data sets, critical discourse analysis can also embrace, rather than hide, the positioned story-telling quality of in-depth, focused, qualitative, critical research.
By reflecting on how (carrier bag) stories do important epistemological, relational and political work in CDS, I have suggested that we do not need to see storytelling as a PR or impact strategy for simplifying findings but as a vital approach to contemporary critical research. When we use similar storytelling strategies for all audiences, we dislocate traditional hierarchies between academic communities and the public. We repurpose the rope from the tug-of-war which pitted PR against social science. Instead, we use the rope (or the threads from the rope) to weave stories together and to weave new theories like the carrier bag theory of fictioninto CDS. Notes 1. My understanding of 'discourse' is informed by Michel Foucault, especially as read through further postfoundational theorists such as Judith Butler, Ernesto Laclau, Chantal Mouffe and Gayatri Chakravorty Spivak who explore how systems of meaningful practices produce subjects and objects; how these practices constitute versions of reality and what is widely accepted as 'knowledge' or 'common sense'. Things 'attain to existence only in so far as they are able to form the elements of a signifying system' (Foucault, 1970, p. 416). My understanding of these processes is also informed by the work of Donna Haraway, John Law and others who write of 'material semiotics'. My understanding of 'discourse analysis' is primarily informed by critical discourse studies and discursive psychology. From these scholars, I learnt that the key to a problem-driven analysis of discourse is to pay close attention to the text and talk. I hope, however, that this paper is relevant for scholars working with diverse approaches to discourse and discourse analysis. 2. Further kinds of story include, for instance, stories of collective transformation, in which collective actors and communities effect change. Kenney teaches this as an alternative to the heroic journey curriculum, which is often taught in public schools (2019).