Literary Ethnography of Evidence-Based Healthcare: Accessing the Emotions of Rational-Technical Discourse

Introduction

1.1 Literary ethnography is a label which has been advocated before (by Van de Poel-Knottnerus and Knottnerus, 1994, in their paper Social Life Through Literature) but not widely adopted as a way to categorise social research. In this article I revisit this method to examine its use in an area of contemporary interest. I use the discourse around Evidence-Based Healthcare (EBHC) as my example, having recently investigated this topic using literary ethnography. Discourse around EBHC is explicitly concerned with technical rationality and the devaluation of emotion as knowledge and yet, as I discovered, is a discourse itself surprisingly rich with emotion. Literary ethnography is useful as a way of accounting for the process of uncovering emotional energies within rational-technical discourse, and bringing these into the conscious linguistic processes of analysis. I argue that literary ethnography, more than other types of discourse analysis and ethnography, allows the researcher to understand their capacity for subjective feeling as an asset rather than a hindrance to social investigation.

1.2 I begin by introducing the topic of EBHC, and justify thinking about it in discursive terms (as EB-discourse). I explain that EB-discourse, while being stylistically rational and technical, is quintessentially a 'human' accomplishment suited to ethnographic study. I introduce the idea of literary ethnography as a type of discourse analysis, and discuss 'analytic categories' as an approach common to different schools in discursive research. Examining literary ethnography and discourse analysis further, I contend that while the techniques of literary ethnography are comparable to those of other kinds of discourse analysis, they are particularly useful for accessing emotional currents which are buried deeply within rational-technical discourses.

1.3 In EB-discourse the distinction of rationality from emotionality is foundational. Consequently emotions become a crucial category for analysis, and literary ethnography an invaluable method for identifying emotional trends which animate the discourse. Although space restricts the examples I can give in this methodological paper I do give some, as I report the process of progressive emotional engagement with EB-discourse. I briefly discuss finding that there are emotional continuities within levels of discourse (among doctors, for example, even when they vehemently disagree), and discontinuities between levels (comparing doctors with nurses or physiotherapists, for example, discovers different emotional climates). The major finding however is that there is such an emotional dimension to this discourse which, while not simply connectable to the system of signs (through discourse analysis), is still accessible ethnographically.

1.4 The contribution closes with some further reflection on the difficulties of addressing emotion empirically, drawing on anthropological ideas, and finally suggesting (tentatively) that Brecht's dramaturgical idea of distantiation might be useful to social researchers for tracking the process of turning feelings into words. This process remains opaque, but identifying it as a problem paves the way for further consideration and research. In the context of EBHC I see an emotions approach, facilitated by literary ethnography, as a way of stepping outside the rules of discourse. This helps to open up a new space for analysing this topic which has proved impervious to creative styles of thought. An analogy can be drawn with Robert Bellah's (1991:254) critical observation that 'those who feel they are (…) most fully objective in their assessment of reality, are most in the power of deep unconscious fantasies'. EBHC means pursuing fullest objectivity, so there are insights to gain from exposing and examining its emotional roots.

Introduction to the study topic: EBHC.

2.1 In healthcare-policy and clinical literature, the acronyms EBM (evidence-based medicine) and the more general EBHC (evidence-based healthcare) refer to ways-of-doing which have been argued for and against since the early 1990s (see EBMWG 1992). Evidence-Basedness in this context means bringing to bear statistical-numerical data, derived from population studies, on decisions in individual clinical cases. Abstract statistical data is usually imagined in opposition to the expertise and experiential craft-knowledge accumulated in the minds and bodies of clinicians. Some writers described how they envisaged EBHC being procedurally enacted (eg. Haynes and Haines 1998, Rosenberg and Donald 1995, Straus and Sackett 1999), while others developed more theoretical critiques of EBHC as practical method (eg. Feinstein and Horwitz 1997, Greenhalgh 1999, Upshur 2005).

2.2 Both apparent 'sides' (those for and against EBHC) contribute to the maintenance of the debate, which is socially significant as a whole and not only for the isolable perspectives of particular groups within it (Pope 2003). I use the phrase 'EB-discourse' to refer inclusively to the set of dialogues and patterns of thought which are gathered around EBHC. EB-discourse has been well-sustained for twenty years and remains influential (Greenhalgh et al 2014) in the upkeep of healthcare as a set of social institutions believed by sociologists and anthropologists of health to have key importance for understanding modern societies. Evidence-basedness also has continuing currency in non-healthcare contexts (eg. ecology (Lamers et al 2015), education (Biesta 2010), landscape architecture (Brown and Corry 2011)), so it is important to study and understand the dynamics of this social phenomenon as a pervasive set of constructs and tensions with broad significance.

2.3 One way to approach the collective achievement of EB-discourse might be to ask how consensus is or is not established within it. For example, Harrison's (1998:26) scepticism of EBHC came partly from identifying a 'presumption of consensus where there is none'. Later, Turner (2010) gave a review of EB-literature which sought to establish coherence, or a hidden consensus to be discovered by reading across the debate. Instead, one can observe that consensus has not been necessary for the maintenance of EB-discourse: rather the continued lack of consensus in EB-discourse makes it powerful and enduring. The concern of the social-researcher of EBHC then becomes not to impose order upon this unruly discourse, but to discover the mechanisms of difference through which it continues, proliferates and repeats as people invest it with energy in different ways.

EB-discourse and discourse analysis.

3.1 Social researchers can be confident of evidence-basedness understood as 'discourse', since it exists primarily in the form of written language, academic-clinical literature in fact – a highly codified and formally-regulated literary style. These writings offer an unbroken fossil-record of EBHC, convenient for literary research. When written about, EBHC is often the main topic for discussion, with contributions from other discourses being juxtaposed and separable. EB-discourse is also self-limiting and self-referring; EB-writings are connected by reference to other EB-writings, forming a set of texts which is highly self-enclosed. This isolation fosters discursive peculiarities and unwritten rules which can be investigated.

3.2 An important part of EB-discourse, associated with the idea of 'evidence hierarchies', is a contest for the general credibility of different empirical research methods (see Mantzoukas 2008). This means that all EB-writers arguably take by implication a position within the discourse. Some, for instance (eg. Stevenson et al 2004), have investigated EBHC using the (strictly-quantitative) randomised-controlled trial techniques advocated within EBHC, aligning themselves rhetorically with a movement towards EBHC. Conversely others (eg. Traynor 2000) identified with discourse analysis as a way of putting distance between themselves and the research values associated with EBHC. That said, it is also usual to find advocacies of EBHC which are themselves conducted in qualitative methods, rather than the statistical ones advocated in EBHC (eg. Dopson et al 2003).

3.3 This raises a question of methodological reflexivity – implicitly taking up position in a methodological debate by adopting methodological sensibilities. As a sociological researcher/investigator, it is difficult to claim detachment from this debate when one's methodological allegiances are openly displayed. By and large, EB-writers have not detained themselves with this difficulty: for example many advocacies and detractions of EBM, even those arguing against the validity of Opinion, have been opinion-pieces (eg. Guyatt et al 1988, Sackett and Rosenberg 1995, Davidoff et al 1995). This tradition continues, even though the rhetorical styles of EB-discourse have become increasingly sophisticated (see eg. Djulbegovic et al 2009,Kelly and Moore 2012).

3.4 There are good practical reasons why many EB-writers adopt a freewheeling attitude to methodological self-awareness. A strictly evidence-based account of EBHC is difficult to write coherently, having to constantly make reference to itself. Against the prospect of such tail-chasing, it is simpler not to worry about dissonances between method and argument: but for analytical social researchers, solid theoretical foundations must be sought, and high standards of methodological clarity adhered to. A thoroughly-considered methodology may assist the social investigator of EBHC in differentiating from other writers who have offered their opinions assertively and repeatedly, and in seeking to escape patterns of discourse which have become well-worn. What then can be the starting point for social researchers studying EBHC?

3.5 Taking the idea of social discourse as a reliable launchpad, one is drawn to the diverse field of discourse analysis, which encompasses traditions in linguistics and conversation analysis, psychology, Marxist Realism, Foucauldian and Science Studies (see Cheek 2004, Schiffrin et al 2008). In discourse analysis the researcher must decide roughly what type of analysis to pursue, but they cannot pre-empt their discoveries, and need to remain responsive in method to the discourse they encounter. Discourses produce particular representations and constructions of reality. As the researcher reads, writes about and contributes to EB-discourse, they become subjectively involved in reproducing and contesting those representations and constructions.

3.6 The research endeavour begins then with an outline of methodological predispositions, subject to refinement and qualification in ways responsive to what is found (rather than being rigidly pre-determined, as the methods of EBHC itself may definitively be). A need remains to be clear on how arguments about EB-discourse are constructed, while replacing a static conception of research-method and research-topic with something more fluid. A study conducted on EB-discourse is also part of EB-discourse, so there is continuity and feedback between research act and research content. In EBHC there is perceived a need to label methods, to put them in order and discipline them. In researching EB-discourse there is a need to avoid such preoccupations^[1], and so to question the linearity of research-time. Stepping away from Evidence-Based values is necessary for creating conceptual space in which things about EB-discourse can be observed.

Evidence-basedness: a discourse in humanity.

4.1 The distinction between EBHC (doing healthcare in an evidence-based way) and EB-discourse (talking and writing about doing so) is important. A driving factor for EBHC is the desire to eliminate factors one might (perhaps pejoratively) think of as 'human': fallibility and error, inexplicability, unaccountability, unpredictability, variability, instinct, emotion. EB-discourse however is inescapably human: its participants write creatively, passionately and whimsically about their perceptions of the powers ascribed to quantitative Evidence. While the topic of evidence-basedness might imply a striving to transcend what is human, there is no escaping the humanity of EB-discourse; and the social analyst of EB-discourse must appreciate that they too are human, and build this into their rationale for research.

4.2 Encountering it as a health professional, it seemed to me that the significance of evidence-basedness lay in the abundance of what was being written about it, not in some other tangible reality which the mass of EB-writing was failing to capture. I made literature the raw material for my research, and took a naturalistic approach. EB-writers, I supposed, wrote not so as to be read electronically by computer programs which would chop up all the EB-writings, splice them back together or count them in different ways, but read by other humans interested in the narratives of EBHC. In this naturalistic approach, the basic activities of research are literature-seeking, reading and writing. The naturalistic approach does not import technologies to mediate the processes of reading and writing, nor does it elicit the production of custom-made data to be re-processed. Instead it takes as it finds.

4.3 This naturalistic approach to EB-discourse involves a type of immersion which is ethnographic. To be sure, some who write about EBHC also practice EBHC (or don't) in their clinical roles. Other writers imagine (or don't) what it would be like to practice EBHC, or draw upon the testimonies of practitioners. What makes these people participants in EB-discourse is purely the act of writing about EBHC. By participating, the investigator becomes one of those who experience EB-discourse as a literary form of social life. Their selfhood becomes their warrant for collecting and expressing knowledge within and about EB-discourse.

4.4 Van de Poel-Knottnerus and Knottnerus (1994) presented case-studies using literature from the realist tradition to investigate social realities which are otherwise inaccessible: realities hidden from mainstream exposure, and buried in the past. They wrote against the possibility that literary methods could be seen as incompatible with systematic and scientific enquiry. Using idioms familiar in EB-discourse, they emphasised the need to rigorously and 'systematically follow a definite procedure' (1994:69), and aim to produce guidelines for best practice. They describe literary ethnography as an 'intensive and systematic reading' in which one looks for patterns, and constructs a composite empirical portrait of a social world. This characterisation applies well to most schools in discourse analysis. Anxiety over the relationship between literary representations and social phenomena which are demonstrably 'real' can be soothed when using literature to study EBHC, in which the huge volume of academic writing assures the researcher that EBHC is something which substantively exists and has social effects. The literary ethnographer of EBHC can therefore proceed more freely to lose themselves in the wilderness of discourse, and to look for navigable pathways through it.

Analysing discourse through analytic categories.

5.1 A label of 'discourse analysis' can be attractive precisely because it is imprecise. Hodge (1989), for illustration, counsels against formalism in discourse analysis, expressing a wish for diversity which typifies the field. Foucault's discursive method (the most common reference-point for discursive theorising – see Hook 2007) being described as a 'kind of toolbox' (Foucault 1974) points the way to a diaspora of idiosyncratic styles. This jars against the aesthetic of assertively-phrased evidentialism which is foundational to EB-discourse, and which may be treated with forensic suspicion by the social analyst. Being something to be investigated, assertive evidentialism cannot comfortably be presumed in any methodological justification. Social analysts must seek a rationale to include contextual specificity, conceptual coherence (especially avoiding circularity), and the construction of arguments which are sustainable by standards which only become clear as analysis proceeds.

5.2 To put this more provocatively: the social analyst does not presume a pathway to methodological worth, but argues their approach painstakingly and thoughtfully. Like the classical anthropologist who finds themselves alone in a culture designated as 'foreign', the social researcher who reads EB-literature finds themselves alone with the discourse. By becoming a human subject within this discourse, they can hope to produce a compelling, perhaps rather personal account of how it works. In discourse analysis, and in qualitative research generally, researchers seek framing structures of meaning through which to understand their topics of study; but the process of developing these analytic categories to a point where they can be confidently trusted is difficult to codify.

5.3 This search for analytic categories is common to different schools of discursive and qualitative research. In Critical Discourse Analysis for example, Kress (1990:93) stated that any analysis will 'attempt to describe the categories which are generative in the production of discourse'. Titscher et al (2000:12) agree; 'one may say quite simply that every observation requires particular observational frameworks or categories'. In Discursive Psychology, the focus is on the everyday use of categories and constructions (Potter and Molder 2005:2), and the discursive practices through which categories are constructed (Phillips and Jorgensen 2002:108). In Discourse Theory, as presented by Howarth (2000:12), meanings are understood as effects of the inter-relational 'play of signifiers'. This protean nature of discourse compels theorists to 'modulate and articulate their concepts' to particular problems (2000:133), thus formulating flexible and responsive analytic categories.

5.4 For literary ethnography, Van de Poel-Knottnerus and Knottnerus (1994) write of themes and significant patterns in the literature which are collected together to produce 'analytic constructs' – abstracted generalisations about the phenomenon studied (1994:73-74). Like Charmaz (1990:1164) in her work on Grounded Theory, they implore the analyst towards intimate familiarity with the texts studied, to 'exhaust the search for counter-instances of patterns' (1994:72). Also like Charmaz' grounded theory, and like other methods for discourse analysis, they expect the analyst to return their newly-derived understandings to source: to verify that their explanations fit with the original texts, and if not, to revise them and seek alternative interpretations.

Doing literary ethnography.

6.1 It is possible to develop further the concept of analytic categories, and in so doing, make something of a departure from Van de Poel-Knottnerus and Knottnerus' (1994) understanding of literary ethnography. They conclude (1994:76) by reassuring their reader of the reliability and integrity of the ethnographic process, so that if properly executed, it produces an accurate portrait. Emphasis on reliability and accuracy carries a whiff of the assertive evidentialism to which the social analyst of EBHC becomes sensitive. An investigator of EBHC, a living discourse, may want their analysis to do more than produce a secure account of what is already in it: they may also hope to open up new channels for thinking about this topic.

6.2 Foucault helps to explain this desire. In The Archaeology of Knowledge (1969:38-39) he recommends that 'instead of going over with bold strokes lines that have already been sketched', discourse analysts should seek to advance towards unfamiliar territory and unforeseeable conclusions. In a discourse such as the one around EBHC, certain categories for guiding thought are securely in place, reproduced by those who have faithfully followed the terms on which debate has been set up. These are Foucault's 'lines already sketched'. Assuredly, the committed analyst will become so familiar with these lines that they can paint them in their sleep; but they might aspire to see past these lines and disclose the patterns hidden beneath. How are they to pursue this transcendence?

6.3 Reading and writing are their tools, of course, but there is more to reading and writing than first meets the eye. As Van de Poel-Knottnerus and Knottnerus (1994:72) observe, several re-readings are necessary to gain the close familiarity sought by literary ethnographers. But it is also the case, as scholars of hermeneutics argue (eg. Bauman 2010:8), that one can never read the same text twice: for reading is an active process of exchange between reader and text, and one may read the same text on different occasions and find different things of importance. The work of reading is never finished, and texts never exhausted of meaning.

6.4 Understanding develops progressively with each successive reading, and as one reads different texts in relation to each other. There is tension between reading broadly to cover a greater range of texts, and selecting particular texts to scrutinise in depth. The investigator must dig sideways (across texts), to develop a broad understanding, and also dig downwards (into texts) to excavate more valuable insights. They might recognise some texts as more significant and frequently-cited than others; then again, more obscure texts might be crucial for establishing specific points of argument. Every utterance is potentially significant: there are no accidents in discourse and everything has the potential for multiple meanings.

6.5 Similarly, writing about any topic is not a passive process where one simply reports what is 'there'. Writing always carries interpretations-already-made and uncovers new interpretations. The literary ethnographer comes to see writing as research-method in itself. As Eagleton (1996:116) says, there is something in the deed of writing, as embodied and performed behaviour in a human subject, which escapes systems of containment. Writing is a challenge to thought structures, and something which cannot be kept under control. It is done so as to see what happens; it is an empirical act, as important as reading for developing understanding. This gives rise to the technique of 'long writing'.

6.6 Rather than first generating masses of de-contextualised data from which elements can be plucked to fit into a tidy writing process, long writing generates masses of written text for editing and re-writing. A dialogue is formed between writing and reading (Bakhtin 1981). Not only is this dialogue never finished; it produces much more text than is finally presented to the reader. This unfortunately means that writing-as-research is time-consuming and inefficient, but it ensures integrity of connection between writing and the thing written-about, which is imperative.

6.7 The idea of writing as research method is not well-recognised in sociology or even necessarily in discourse analysis; but does have precedents in anthropology (see eg. Ely et al (1997) on 'writing towards understanding' and Wall's (2006) 'auto-ethnography on learning about auto-ethnography'). This is something valuable that literary ethnography can add to the techniques of discourse analysis. Instead of seeking a final and definitive meaning (as is often implicitly sought by writers in EB-discourse) the literary ethnographer seeks to open up alternative possibilities of meaning. There are no last words in this type of investigation, and always more to be said (see Bell and Gardiner 1998).

Differentiating literary ethnography from discourse analysis.

7.1 Literary ethnography, then, contributes instrumental techniques which are perhaps more extreme, but not different in kind from the painstaking methods of other well-established discursive approaches to text. Where literary ethnography subtly differs is in the manner of seeking to find a life-within-literature which cannot be brought to light except by more prolonged personal exposure. To be sure, all discourse analysis is designed to move beyond the immediate and explicit meaning of text. In giving name to this 'hermeneutics of suspicion', Ricoeur characterises the analytic approach to discourse which sees it as a carrier for meanings which can be decoded (see Josselson 2004). Discourse analysis always expresses this aim to find meaning-structures which go beyond individual texts and speakers into social contexts and constructions, which require more work to interpret (Willig 2015).

7.2 Nevertheless, the guiding notion of discourse analysis is that structures within texts, at first obscured and disguised, can be brought to light and demonstrated by reference to the arrangement of elements in the text. Ultimately discourse analysis means using technical methods for studying patterns and relationships between signifiers. In literary ethnography there is allowance for the possibility that text can carry content which, while embodied within text, may have a relationship to formal textual content which is resistant to explication. Explication could still be sought in the technicality of grammatical structures (seeOchs and Schieffelin's Language has a Heart, 1989), but for (naturalistic) empirical purposes it is just as well to think of this relation as analogous to Polanyi's (1966) account of tacit knowledge and skill in scientific practice, and to say that texts have depths which are difficult to fathom.

7.3 This property of texts is something which one cannot perceive without 'being there' in an embodied ethnographic way. It is content buried deeply beneath the structure of texts as socially-produced documents, such that the analyst has to become a feeling participant in the discourse if they are to perceive it. In the context of EBHC-literature this content can be thought of as emotional. It is an emotional life which does not find plain expression in rational-technical discourse. It is still expressed, but its expression is something which escapes, leaks out, and emerges through channels peripheral to the consciousness of writing and reading. When encountered it comes as a surprise to the discourse analyst who may pre-conceive the content of discourse of this kind. Rational-technical discourse is supposed not to be concerned with feeling but paradoxically, is found to be alive with it.

7.4 So long as texts are thought to be representations of a life more real, there is a concern of how the text reflects, constructs or distorts the reality of its subject (as for Van de Poel-Knottnerus and Knottnerus, 1994). But if texts are already thought to embody something which is more than their formal content, this anxiety fades. The text is not a mere shadow or projection of social life as lived and experienced, but is itself a chunk of social life. It does not exhaustively encompass all of social life but then, nothing does. Literary ethnography accords text the potential to be less restricted in reach than strictly-discursive analysis might suppose. The text assumes properties which go beyond the circulation of symbolic discourse, which the literary ethnographer can hope to discover.

7.5 This attempt to reveal meanings which are hidden, especially those which are outside of intentional thought and which might have the affective character of desires, fears or fantasies, savours much of psychoanalysis. There are resonances between discourse analysis and psychoanalysis which have been explored (Parker 1997) and could be explored further. Literary ethnography, pushing as it does at the boundary between meaning inside and outside of formal language, moves into psychoanalytic territory. Arguably though, literary ethnography is not properly psychoanalytic as it does not involve postulating any kind of structure to a subconscious world, whether collective or individual. It is strictly empirical. It facilitates the search for transient emotional currents which can be discerned empirically through (if not in) written language. 'Deep theorising', as Lillis' (2008) uses this term in connection with longitudinal ethnography, might be more appropriate a term than psychoanalysis for this kind of endeavour.

Emotions in EB-discourse.

8.1 The trouble with emotion, from the point of view of analysing discourse, is that its relationship to formal written language is complicated (Edwards 1999). Consider how poorly-equipped are the names used for emotions to capture and locate their meaning. For example, the word for love scarcely defines the multiple connotations of love as a social phenomenon (Seebach and Núñ ez-Mosteo 2016). Nor is any emotion-word really an adequate representation of its emotion as an embodied and experienced state of being. Sometimes it is only in hindsight, in view of their effects, that emotion-labels can be usefully assigned (as for example, when one retrospectively rationalises actions in terms of emotions: I must have been angry, you must have been scared). Also as Barbalet (2001:24) notes, 'the absence of a word for an emotion does not mean that an emotion is not (…) influential'. He alludes to an emotional life outside of language, and a culturally-instituted blindness to emotions which linguistic conventions do not recognise (ibid).

8.2 Barbalet writes also of emotions 'below the threshold of awareness' (2001:114) which, like a collective subconscious, are nonetheless effective in structuring social and individual actions. The threshold of awareness for emotions may be both culturally variable, and variable within individuals depending on nuances of context. It is difficult then for social researchers to be sure what they are talking about when they use labels for emotions, and it may be difficult for an analyst of discourses concerned primarily with issues of rationality (such as EB-discourse), to know how to respond to the visceral emotions produced in these contexts, which may seem out of place and difficult to articulate.

8.3 As Cronje and Fullan (2003) have noted, EB-discourse is concerned with rationalities of practice. Where EB-discourse is explicitly addressed to emotion, it is with the business of excluding it from the formal decision-processes of healthcare, or finding ways to incorporate it into schemata of rationality. Reading EB-discourse however, the analyst occasionally encounters 'rich points' (Agar 1996, from Lillis 2008:382) where recognisable emotions break the surface. An example of this is Miles et al's (1998:259) identification of a Scientific School who they accuse of misappropriating scientific rhetoric to advance a spurious numbers-based takeover of medicine. They subsequently refer to their disputants at points through the article, those who advocate EBHC, as the SS.

8.4 As an isolated expression of antipathy directed across the lines of debate, this nudge-and-wink linkage of EBHC to fascism might not be significant. But the analyst who has read much EB-writing, gaining close familiarity with the tenor of expressions made on both sides, can interpret it as indicative of sincere feelings which have to do all at once with distaste, disgust, fear, defiance, aggression and hatred. Such events cumulatively alert the investigator to emotion as a generative aspect of EB-discourse: to emotions in the writing of others, and in themselves as they read. To detect emotion, they cannot be just a skilled processor of symbols arranged in patterns, but must be able to feel. To participate fully in discourse, the analyst must be alert to sentience which usually is camouflaged under layers of intensely-rational argumentation.

8.5 For me, rich points of emotional disclosure gradually coalesced to create a more general sense of tension in feeling, lessening the sense of surprise when tensions breached. A key passage point was reading anthropologist Helen Lambert's (2006, 2009) comparative reflections on EBHC and anthropology. The first of these (2006) is a measured defence of EBHC, analysing the properties which afford it the flexibility to seemingly incorporate any objections into strategies which inoculate it from criticism. Upon addressing her own discipline (Lambert 2009) the ambient temperature almost seems to rise, the pace of writing to quicken, as the language turns to need: 'we need to ensure quality and veracity … we badly need to clarify standards' (2009:19). Reading this I felt that the determined 'will to evidence' was just as strong for EB-advocates outside of clinical contexts as the defiance I had encountered inside, where it was the will to resist.

8.6 Once the antennae of feeling are sensitised, emotion existing outside the thresholds of awareness can be brought into central consciousness by analysis. As something whose essence is in feeling but whose effects are discursive, I contend that emotion can only be accessed for analysis through its presence (identified not first by name, but by feeling) in discourse. To know that emotion is present and to judge its effects, the analyst must be able to detect emotion by feeling and (which is harder) transform it effectively into language. The idea of literary ethnography is crucial to this aspect of the analysis. It is only through personal immersion – through living the discourse – that this type of embodied, situated awareness, can be cultivated.

Ethnography and emotional methods.

9.1 Recent times have seen sociologists draw attention to the social significance of emotion: how emotion manifests in macro-social structures (Barbalet 2001), how emotions do inter-personal and social work (Ahmed 2004), how emotions permeate social movements (Goodwin et al 2009), how emotions are central to reflexivity (Burkitt 2012), for examples. The balance of emphasis in that discipline has been towards theorisations of emotion. For present purposes I am concerned more restrictively with how emotions can be accessed empirically, for which it is appropriate to look at anthropology and the tradition of ethnographic research.

9.2 One classic anthropological text on emotions (Lutz and White 1986) identified emotion as something of burgeoning interest, and something problematic. In it, attention is given (1986:430) to ethnographers' own emotional responses to fieldwork. While these responses have their uses, the emic-etic division between researcher and researched makes for an emotional asymmetry which has to be accounted for and managed. More recent articles show that such anxieties remain. Holland (2007:195) recognises that emotion is necessary for knowledge but seems also to resent this necessity, offering 'solutions to the pains of emotion work in the field'. Blackman (2007) uncovers a history of ethnographic research which has been hidden by researchers uncomfortable with its emotional content.

9.3 Blackman (2007) highlights some reluctance among social anthropologists to come to terms with their own capacity for emotional insight. Perhaps anthropology is a discipline in which an attachment to scientific objectivity is still difficult to escape (seeLambert 2009), and consequently the necessarily-emotional being of researchers is something which may be concealed like a mark of shame. Blackman presents a more positive and proud view of the emotions of fieldwork, but is still some way short of the position I am prospecting for. This is that the enabling properties of ethnographic research might make it not just the best way, but the only practically-feasible way, in which a researcher-analyst can hope to uncover the emotional dimensions of culture, through becoming an embodied participant.

9.4 This being the case, the researcher faces not just the problem of entry into the emotional realms of discourse, but also of exit, how to bring emotions into the realm of academic-analytical language. Interest in the emotional work of qualitative research is emerging, for example Dickson-Swift et al (2009) have articulated a concern for qualitative researchers whose work exposes them to emotional risks. Such work is welcome but does not foreground the positive properties of emotional engagement as a research tool. Indeed, academic disciplines offer few instructions on how to approach emotion as a problem in method, so it is sensible to seek other spheres in which people contend with the practical problematics of emotional expression. A useful precedent I found (that has visible linkages to social theory) was Bertolt Brecht's dramaturgy, which affords a practical framing for this kind of emotion-work.

9.5 To experience emotions in embodiment and then to recognise them in thought, involves a splitting of subjectivity between a self who feels (an acting self), and an analytical self who observes feeling and more-or-less knowingly (re)presents it in language. To conceptualise this, Brechtian theory draws upon ideas of estrangement, alienation, distantiation and 'somatic guidance' (Robinson 2008). Brecht was concerned with emotional distancing via the 'alienation effect' through which he implored actors to achieve circumspection towards the characters they portray. He wrote:

'Acting like this is healthier and in our view less unworthy of a thinking being; it demands a keen eye for what is socially important.' Brecht (trans. Willett) 1964:95, emphasis added.

9.6 Without delving deeply into Brecht's ideas (or considering the controversial consequences of drawing parallels between acting and ethnographic research as performance), I merely comment that they were useful for encouraging me to begin to keep a diary of emotions alongside my more formal discourse analysis. This diary gradually became central to my method, just as Brechtian acting makes emotion central to practice.

9.7 I began by marking passages of text and phrases which I felt were emotionally rich. I made lists of emotion-words which felt like good proxies for the feelings I encountered. I looked for emotion-words in texts, to see if they matched up with the emotions being projected. I tried to distinguish emotions in myself from emotions I could demonstrate as being in-the-text, and to work out how these emotions became perceptible from ink on paper. I looked at the emotional narratives in texts, to see how the emotional climate varied through their course. I made links between different pieces of writing which were different in explicit ways, but which showed similar emotional patterns.

9.8 In terms of reading techniques, the Brechtian view implies an approach in which particular instances of emotionality-as-experience are analysed in the context of their occurrence. The analyst engaged in this process may present particular passages of discursive text, hoping to demonstrate the emotional energy within them, and by comparing related texts, to situate different emotional currents in relation to each other. I found that patterns emerged across the literature. Quite often, participants who were most explicitly opposed to each other (for example, groups of doctors who were staunchly for or against EBM) were close together, emotionally speaking. Pronounced differences were apparent between disciplines though. Medicine, nursing, physiotherapy, health policy and sociology all showed different prevailing emotional trends.

9.9 To report these results in detail is a task for another day. For the present, the important points are these: that the analyst of discourses created amongst human subjectivities is themselves a performing human subject, and must acknowledge so in the rationale for their work; and that the researcher-subject has an emotional subjectivity which is an essential asset for accessing the hidden emotions of (rational-technical) discourse. In dialogue between emotion-as-feeling and as discursive expression, this realm of experience can be brought into consciousness.

Conclusion.

10.1 The precedent I have drawn upon for presenting literary ethnography as an investigative social-research method is Van de Poel-Knottnerus and Knottnerus' (1994), who argued a case for using realist fiction as a historical source. With those authors having laid the foundations for seeing literature as a dimension of social life, suitable for social investigation, I have taken steps towards reviving their method for use in other contexts. The concerns addressed by Van de Poel-Knottnerus and Knottnerus (1994) in their paper resonate with themes in EB-discourse: to establish the reliability, scientificity and robustness of literary research, to use it as Evidence. Consequently my recommendation to use literary ethnography as a method which allows the researcher to (as it were) see beyond those themes, requires some closing reflection.

10.2 In my view, literary ethnography can facilitate the researcher in seeking more than an accurate, painstaking reportage of a discourse. This method requires the researcher to live the discourse fully, to become personally immersed within it. Through experiencing discourse, and then achieving a separation between their feeling-self and their critical-analytical self (which is not an easy thing to do), the researcher can hope to offer a new interpretation of the discourse, and perhaps (if it is a contemporary, 'live' discourse) to change it. In the case of EB-discourse, I suggest, this change is due. I hope that recognising the emotional foundations of this rational-technical discourse is a route to transcending the dichotomies which sustain it. It remains to be seen whether this is the case, or whether this emotions-based analysis of EB-discourse becomes just another set of representations.

10.3 The major strength of literary ethnography, as I see it, it that it brings written discourse to life. It provides a structure to articulate and make sense of the emotions which appear in the context of ostensibly non-emotional, technical discourses. The social investigator can use their subjectivity, their sentient self, to record the emotional trends in the discourse as they experience them. The difficult part of this process is then to make a good-enough translation of feelings into the domain of words. Here I have envisaged the investigator to deploy a technique akin to Brechtian distantiation – re-instating a detachment from emotional states, which allows feelings to become a topic in conscious discourse. The details of this process are also a suitable topic for further research in themselves.

---

Notes

¹ Tendencies to methodological labelling and loyalty become suspicious in a way which calls to mind Foucault. 'Do not ask who I am and do not ask me to remain the same', he wrote (1969:17).

---

Acknowledgements

Thanks go to Dr Lisa Garforth, Dr Nissa Finney and Dr Sharon Leahy for their helpful advice with the writing of this article.

---

References