Depression, rational identity and the educational imperative: concordance-finding in tricky diagnostic moments

It is well-documented, within most medical and much health psychology, that many individuals find diagnoses of depression confusing or even objectionable. Within a corpus of research and practical clinical guidance dominated by the social-cognitive paradigm, the explanation for resistance to a depression diagnosis (or advice pertaining to it) within specific interactions is bordering on the canonical; patients misunderstand depression itself, often as an output of an associated social stigma that distorts public knowledge. The best way to overcome corollary resistance in situ is, logically thus, taken to be a clarification of the true (clinical) nature of depression. In this paper, exploring the diagnosis of depression in UK primary care contexts, the social-cognitive position embedded in contemporary medical reasoning around this matter is critically addressed. It is firstly highlighted how, even in a great deal of extant public health research, the link between an individual holding “correct” medical knowledge and being actively compliant with it is far from inevitable. Secondly, and with respect to concerns around direct communication in clinical contexts, a body of research emergent of Discursive Psychology and Conversation Analysis is explored so as to shed light on how non-cognitive concerns (not least those around the local interactional management of a patient’s social identity) that can inform the manner in which ostensibly “tricky” medical talk plays-out in practice, especially in cases where a mental illness is at stake. Finally, observations are drawn together in a formal Discursive Psychological analysis of a small but highly illustrative sample of three cases where a depression diagnosis is initially questioned or disputed by a patient in primary care but, following further in-consultation activity, concordance with the diagnosis is ultimately reached—a specific issue hitherto unaddressed in either DP or CA fields. These cases specifically reveal the coordinative attention of interlocutors to immediate concerns regarding how the patient might maintain a sense of being an everyday and rational witness to their own lives; indeed, the very act of challenging the diagnosis emerges as a means by which a patient can open up conversational space within the consultation to address such issues. While the veracity of the social-cognitive model is not deemed to be without foundation herein, it is concluded that attention to local interactional concerns might firstly be accorded, such that the practical social concerns and skills of practitioners and patients alike might not be overlooked in the endeavour to produce generally applicable theories.


Introduction
I n contemporary psychological research, particularly in the clinical and health domains, there is a widely-reported concern that many individuals find diagnoses of depression troubling or even objectionable, and will often show some form of active resistance when the diagnosis is made (Van Voorhees et al., 2005;Highfield et al., 2010;Wimsatt et al., 2015). Within this body of literature, the explanation for such resistance and the proposed mechanism for overcoming it are bordering on the canonical. Ultimately, resistance is taken to be an output of the patient in some way misunderstanding the fundamental nature or implications of depression itself, often as an output of a broader social stigma attached to the condition (Van Voorhees et al., 2006;Corrigan and Wassel, 2008). Consequently, and logically, the proffered solution stems from a singular proposition: "education is the key" to backfilling or correcting that patient's knowledge, given that it is likely (and respectively) incomplete or distorted (Tylee and Jones, 2005;Tanney et al., 2012;Simmons et al., 2015). As evidenced in many guidelines for frontline clinical practice, it is taken that education of this order should then facilitate the resistant patient's acceptance of their diagnosis (Petit and Sederer, 2006;Simmons et al., 2015), such that positive moves forward can be made in terms of treatment and care.
While intuitive on many levels, given the above, the core understanding of depression-related diagnostic activity embedded in this broadly social-cognitive orthodoxy is not unproblematic. A range of work in the discursive and interactional sciences has, to date, highlighted how the reduction of any clinical interaction principally to matters of information transfer and processing tends to obscure the complex social contexts inhabited by clinicians and patients, and also the practical social skills of the interactants (Silverman, 1997;Stivers, 2006;Miller, 2013). Indeed, it is often acknowledged in clinical directives around depression diagnosis themselves that factors such as culture, use of language, ad hoc social skills and personal understandings can play important roles in frontline medicine (Tylee et al., 1996;Tylee and Jones, 2005; National Institute for Clinical Excellence, 2016). In terms of direct clinical research and guidance, however, attention is rarely accorded to how the particular interpersonal actions involved in delivering diagnostic news, in resisting (or not) it and in reaching a subsequent understanding (or not) are actually "done" in given cases. Consequently, generalizing explanations for resistance, and recommendations for reaching resolution, tend to prevail.
Given the above, the core aim of this article is to offer a Discursive Psychological (henceforth DP, see Edwards, 2012;Wiggins, 2013;Miller and Benkwitz, 2016;Patterson and Fyson, 2016) perspective on the phenomenon resistance to a depression diagnosis in frontline clinical work. In the service of this, and following a review of currently influential social-cognitive research on the issue (as typified in Tylee and Jones, 2005;Cook and Wang, 2011;Wang et al., 2015;Wimsatt et al., 2015), and allied clinical guidelines (for example, Petit and Sederer, 2006;National Institute for Clinical Excellence, 2016), the assumptions and recommendations therein are then explored with reference to pertinent literature on interaction in healthcare contexts, and particularly during diagnosis and diagnostic reception, emerging from research in DP and the closely-related discipline of Conversation Analysis (henceforth CA, see Maynard, 2006;Peräkylä, 2006;Miller, 2013;Sikveland et al., 2016). Key matters are then drawn together in a formal empirical DP examination of a small but highly illustrative original data set arising from a wider study of the diagnosis of depression in United Kingdom (UK) primary care settings. Herein, three pertinent consultations are examined in which new depression diagnoses are delivered, and subsequently met with troublesome receptions from patients-a specific matter hitherto empirically unexplored in either the DP or CA domain. By paying particular attention to the nuanced activities of clinicians and patients during these consultations, a description is provided of some of the particular, non-cognitive concerns than can inform and shape interaction during these "tricky" clinical moments. Not least among these are matters relating to how patients' social identities are coordinatively constituted by speakers as a part of the interactions themselves (Antaki and Widdicombe, 1998a;Miller and Benkwitz, 2016) and, most importantly, how ostensive challenges to the diagnosis itself can be functional devices for a patient to open up conversational space within the consultation where these matters might be addressed. All emergent matters are finally addressed in terms of a reflection on issues relevant for research and practice around the clinical handling of nominally difficult topics in general, and the diagnosis of depression in particular.
Depression, diagnosis and resistance: social cognition and silver bullets As noted above, it is well noted in much extant medical research, and emergent guidelines, that patients will often display a range of inferably resistant activities following receipt of a depression diagnosis in primary care (Tylee et al., 1996;Van Voorhees et al., 2005;Cook and Wang, 2011;Wang et al., 2015;National Institute for Clinical Excellence, 2016). Alongside a series of analogously troublesome behaviours surrounding depression in frontline healthcare 1 , a patient questioning and/or disputing their diagnosis in situ, or indeed advancing an outright rejection thereof, is generally taken to indicate one of two problematic states of knowledge regarding depression itself. The first is a simple lack of information regarding the complexities of the condition. Clinical depression, in its most commonly diagnosed forms, is defined in terms of a constellation of core symptoms (American Psychiatric Association, 2013), not all of which areby any general inference-psychological in character 2 . Indeed, directives within the UK's National Health Service (henceforth NHS) actively separate these symptoms into psychological, physical and social domains, as schematised from National Institute for Clinical Excellence (2016) guidelines in Fig. 1.
While a patient's report of the more obviously somatic aspects herein (such as sleep perturbations, fatigue, appetite changes and so forth) are often sufficient trigger for a clinician to suspect a case of depression, their co-occurrence is far more likely to lead that patient to expect a clearly somatic diagnosis (Tylee and Jones, 2005;Van Voorhees et al., 2005;Wang et al., 2015). Moreover, patients often interpret (and therefore report) some depression-relevant psychological states as simple upshots of their somatic symptoms or social circumstances (Wolpert, 2001;Wang et al., 2015). To feel low, or enjoy specific activities less than one has in the past, for example, when one is constantly fatigued, struggling at work and/or unable to sleep hardly seems illogical. Thus, without knowledge of the formal (medical) character of depression, the act of having simply relayed a set of ostensibly somatic matters, and in some cases their "natural" psychological consequences, leads many patients to find a diagnosis of something commonly known to be a mental illness highly confusing and/or upsetting (Paykel et al., 1997;National Institute for Clinical Excellence, 2016).
The second explanation given for resistance to a diagnosis of depression in this body of work relates to a relatively stable prejudicial belief about depression held by some-a social stigma (Corrigan and O'Shaughnessy, 2007;Cook and Wang, 2011). Wolpert (2001, p.223) famously describes the outputs of this nominal stigma thusly: The shame and stigma associated with depression can prevent those with the illness admitting they are ill … . There is also the stigma of taking antidepressant medication which is perceived as mind altering and addictive. Stigma may also cause somatic symptoms as it is more acceptable to talk of stomach ache and fatigue than mental problems.
So rather than simply having a lack of knowledge, these individuals are subject to an entrenched version of the wrong knowledge. Within current clinically oriented research, this order of understanding, particularly prevalent among older adults (Connery and Davidson, 2006), is generally taken to be itself based on internalized information relating to depression, and to depressed people, which is (a) negative, (b) factually-limited and/ or distorted, and (c) globalizing (Tanney et al., 2012;Simmons et al., 2015;Wimsatt et al., 2015). To this extent, clinical guidelines in the United Kingdom actively acknowledge obstacles that healthcare practitioners may encounter when diagnosing or treating a patient with depression, cautioning that General Practitioners (henceforth GPs) in particular should "…be aware that stigma and discrimination can be associated with a diagnosis of depression." (National Institute for Clinical Excellence, 2016, p. 11).
Whatever the proposed social and/or psychological root of a patient's resistance to a depression diagnosis might be, however, the general solution remains the same in all such accounts: the best way to overcome resistance is to propositionally inform the patient of the true character of their condition, that is, to demonstrate that it is a genuine medical disorder, like any other (Barney et al., 2011;Griffiths and Crisp, 2013). This model is schematised in Fig. 2.
The essential value of providing the general public-in person or otherwise-with a clear account of current medical knowledge is both a laudable and essential goal, particularly in a climate where social and "alternative" media can facilitate the rapid spread of prospectively dangerous disinformation about medicine (Rosselli et al., 2016). This aspect of the broad framework addressed above is in no way disputed herein. Rather, deemed as problematic are two of the assumptions that underpin the explanatory character of the model detailed above: (a) That lacking and/or socially-distorted knowledge is the likely root of behaviours that imply a resistance to (or rejection of) a depression diagnosis (Nieuwsma and Pepper, 2010;Barney et al., 2011;Griffiths and Crisp, 2013), and; (b) That furnishing members of the public with more and/or more accurate information about depression is the clearest mechanism through which resistance can be overcome, and positive steps forward taken (Petit and Sederer, 2006;National Institute for Clinical Excellence, 2016).
Regarding (a), it is widely accepted in much contemporary healthcare research (including social-cognitive work itself) that resistance to a medical opinion need not output directly from issues of knowledge-deficit about the medical matter itself. Rather, where a diagnosis or piece of advice conflicts with a deeply-held (often religious) conviction, then resistance can equally be an output of an individual having made an active choice that-while inferably irrational from a purely instrumental perspective-cannot be taken to directly evidence a lacking or distorted understanding of the medical facts    (Koenig, 2004). Connectedly, and regarding (b), meanwhile, there is a commonly voiced complaint in public health literature that individuals being well-informed of a medical truth and using it as a catalyst for a change in behaviour seldom square-up as well as they theoretically should (Gray et al., 2006). For example, full knowledge of the prospective dangers of smoking is indicated as the primary motivation for quitting in only a small minority of cases; far more commonly, it a personally-relevant health event (that is, experience of a negative health outcome for that individual, or a close friend or relative) that actually triggers practical action (Gallus et al., 2013). In this respect, there is limited evidence to suggest that a transfer of medically correct information, at the broader public level at least, will inherently determine a positive response to it. Despite broader healthcare literature highlighting how contingent concerns at the local and personal levels can be hugely influential in an individual's approach to medical information, however, the core social-cognitive reliance on a "deficit model" (McNeil, 2013) of knowledge around depression as an explicatory mechanism for resistant behaviour remains largely uncontested within practice-facing research and guidance. This, it is not unreasonable to suppose, is connected to the status of depression itself as a "mental illness," a category of diagnoses that is wellevidenced to be less well-understood by the general public than many somatic conditions, and sometimes seen as the medicalization of moodiness, irrational behaviour or even laziness (Tylee and Jones, 2005;Barney et al., 2011;Cook and Wang, 2011). It is important to note that the prospectively troublesome status of public knowledge about depression, although widely thought to have been improving for some time (Pescosolido et al., 2010), is not at stake in this paper any more than the value of transparent information. Indeed, it is not proposed herein that the broad social-cognitive model detailed above might not apply absolutely in some cases. Rather, it is the contention here the relationship between evidence and inference should be considered in particular, rather than general, circumstances. A set of observations emerging from the traditions of DP and CA regarding interpersonal action in direct clinical contexts themselves (Antaki et al., 2007c;Finlay et al., 2008;Miller, 2013), particularly those relevant to the practical business of giving and receiving diagnosis (Frankel, 2001;Maynard, 2004;Peräkylä, 2006;Ijäs-Kallio et al., 2012), give us cause to reflect upon the complex issues that might be indicated within any specific case of diagnosis/resistance, as outlined below.
DP, CA and discursive action Herein, findings emerging from studies in DP and CA are addressed as generally compatible, given that they both address language-use as a concurrently constructed and constructive medium (Potter and Hepburn, 2005), the character of which is not reducible to what speakers "think," but is situationally accomplished with reference to how they infer it will be heard within specific sequences of talk (Miller and Grimwood, 2015). For the purposes of this section, such a stance is relatively unproblematic. It is important, however, to be mindful that while both approaches are ultimately rooted in Harvey Sacks' original lectures on conversational forms, first delivered in the 1960s (Sacks, 1992a, b), CA has remained a relatively consistent approach since this foundation, while DP has a more differentiated character. Accounts of the specific convergence and divergence within and between the disciplines are available in a number of sources (for example, Wooffitt, 2005;Miller, 2012) but, as Wiggins (2017, 37) highlights, the most significant point of difference is that while: …CA is primarily focused on the social organisation of talk and how people make sense of each other in interaction, DP has an explicitly constructionist focus. That is, DP is concerned with the way in which categories are produced and performed in discourse; with the versions of reality that are invoked and made available for social and psychological actions. So, while CA might examine how talk is socially organised, DP would examine how this particular version of talk is socially organised. [Original emphasis] Pertinent findings from the two traditions are reviewed here, thus, with respect to their implications for an understanding of talk within diagnosis, and broader talk around mental illness categories, to illuminate the specific matter at hand.
Diagnosis, discourse and social identity. Diagnosis is, from any point of view, something that generally is "done" by one individual to another; the bearer of the diagnostic news (a GP, radiographer, oncologist and so forth) will, in effect, be ascribing new attributes to the social identity of their patient at the local level (Maynard, 2004;Miller, 2013). Within the social-cognitive tradition, this process is taken to be an essentially stable one (Tucker, 2009). Where a diagnosis of depression is made, for example, its known implications for the patient's identity become correctly or incorrectly cognitively placed by the patient, and generally serve as logical drivers of subsequent activity. So, a patient misunderstanding of the character of depression (as the wrong thing, or just a universally bad thing) tends to result in resistant behaviour (Griffiths and Crisp, 2013;Simmons et al., 2015;Wang et al., 2015). Should they cognitively place depression as a "real illness" rather than embarrassing/stigmatised state, and/ or the right thing with which to be diagnosed given the symptoms reported, on the other hand, they will likely accept the diagnosis without caveat (Cook and Wang, 2011).
This approach to the relationship between knowledge and action in context is, on many levels, highly intuitive. A wealth of work in the interactional sciences has, however, demonstrated how individuals in practical discursive circumstances, including clinicians and patients, tend to manage matters of social identity in ways that do not easily reduce to matters of cognition (Widdicombe and Wooffitt, 1995;Heritage and Robinson, 2006;Wiggins, 2009;Miller and Benkwitz, 2016). Take, for example, the baseline business of delivering a diagnosis. Clinicians can rarely be seen to simply announce a diagnosis as transparent "news" to be ingested, following a process of objective symptom analysis. Rather, the material of the diagnosis is itself generally and observably framed with sensitivity to prior activity within the consultation, and the inferred impact it might have upon the patient's sense of immediate self (Stivers, 2006). With reference to this, Silverman (1997) and Peräkylä (2006) highlight that the way in which a news-delivery sequence in clinical practice is formulated as "good" or "bad" largely proceeds from a situated assessment of the patient's contextual expectations. Where diagnostic news is anticipated to be taken badly, specific communicative strategies are used to soften the blow, not least among which is the use of "expressive caution" (Silverman, 1997) -a set of interpersonal techniques including a hesitant approach, lowered voice, and stretched-out delivery involving a prediagnostic raft of reassurances regarding how the physical and social impacts of the (as yet undelivered) news is not necessarily a worst-case-scenario. These simultaneously preclude the diagnosis being heard as a "shock," via pre-preparing the patient for the prospect of bad news, while also making available that the clinician understands the matter is indeed sensitive (local ARTICLE PALGRAVE COMMUNICATIONS | DOI: 10.1057/palcomms.2017.33 empathy). Where the clinician's assessment of the interactional context aligns with that of the patient, and actions proceed in this way, then news is generally received amicably (if not necessarily cheerfully) by that patient. If misjudged, however, this order of interpersonal approach can actually engender the exact interpersonal difficulty it was designed to avoid (McLeod, 1994;Miller et al., 2017); in short, the patient can take the clinician's act of treating an issue cautiously as being implicative of the news itself as something shameful. The reverse is also true. Where the clinician directly and unproblematically addresses matters that the patient actually finds contextually sensitive, the latter will often withdraw from the interaction, withholding cooperation and/or rejecting advice (Silverman, 1997). Thus, the sensitivity of an issue or topic in medical interaction is to some extent "… constituted by the very act of talking about it cautiously and discreetly." (Bergmann, 1992, 154). As Goffman (1963) famously notes, social stigma is not something that necessarily pre-exists a context; rather it can also emerge as a consequence of matters therein being handled as stigmatised (or not) within the to-andfro of meaningful interaction 3 . At a more directly sequential level, and as Maynard (1997Maynard ( , 2004 comprehensively explores, clinicians will often work over a series of questions and answers so as to infer a patient's stance on how given news might be taken before they deliver it. Such perspective-display sequences 4 provide the clinician with contextual resources from which to infer whether (or not) a softening of the form of the news itself, or a mitigation of the stated consequences, might be necessary to avoid a difficult response from the patient. Perhaps more importantly, however, it also allows for a clinician to co-implicate the patient into the diagnostic formulation, incorporating their own assertions (or even exact words). In doing so, the patient's agency within the immediate context is foregrounded (Maynard, 2004); a local case is made for how the diagnosis is a direct upshot of the patient's own interactional work, and accountability for the material of the diagnosis is thereby shared between both participants. In these circumstances, acceptant (concordant) responses are very often engendered, even when the news itself is by any account "bad".
This indicates that local issues around social identity (that is, the impact of identity ascriptions within the immediate contexts of their ascription) often appear to be of greater concern to clinicians and patients alike than the global cognitive impacts of someone becoming known as, for example, depressed. While the use of perspective display (or expressive caution) within a consultation can make little inherent difference to the wider social circumstances of the patient ultimately being known to "have" any given condition, the precise manner of diagnostic formulation remains inferred by practitioners to be a critical element of patient-management in situ (Potter and Hepburn, 2005;Antaki et al., 2007a;Goicoechea, 2013;Miller et al., 2017). Were this not the case, then it would be largely unnecessary for a clinician to address news-delivery with reference to its local import, nor would this likely have any meaningful impact upon the manner in which patients receive their diagnoses.
Although this is possibly counter-intuitive within cogntive psychology, work on social interaction across a range of contexts has demonstrated how individuals often shape their activity first and foremost with respect to local issues of "face" (Goffman, 2003) and immediate social credibility (Miller and Grimwood, 2015). A broad range of work on medical interaction (Speer and Parsons, 2007;Antaki, 2008;Goicoechea, 2013) has robustly demonstrated that patients in a range of healthcare settings pervasively attend primarily to local interactional concerns rather than notional knowledge-structures, or easily defined social positions. In a study of upper respiratory tract infections for example, Ijäs-Kallio et al. (2012), in line with the work of Maynard (2004), highlight how apparent comfort or discomfort with diagnosis among the patients in their corpus (the latter expressed in questioning or challenging the clinician's opinion, that is, "non-concordance") has little to do with those patients' propositional understandings of the diagnosed condition, but is usually an outcome of the degree to which they can interpret the GP's diagnosis itself as following logically from the sequential unfolding of questions asked and the answers given during that specific interaction. In short, the formatting of the consultation itself needs to confirm that the patient is a full participant in its outcome; i.e. cooperation from the patient is broadly contingent upon a locally-generated sense of their own identity as fully agential. Similarly, Miller (2013) illustrates how GPs in primary care, when trying to establish if a patient has suicidal or selfharming ideation, need to direct a careful series of questions (each contingent upon the given answers) that gradually align the patient's local social identity with that of someone who "might perfectly reasonably think that way." Where the patient's identity is successfully aligned in this manner, and despite the apparently thorny topic, a question about suicidal ideation is answered (positively or negatively) without problem. Where it is not, and the question appears sequentially out-of-place, the patient then withdraws cooperation within the consultation. Here, cooperation is contingent upon a local sense of the patient's identity being both agential and rational.
This latter work, in particular, gives us cause to reflect not only upon the specific business of diagnostic news-delivery, but also upon broader issues arising from ascribing a "mental" illness as an attribute of an individual's identity. Much of the clinically oriented literature previously reviewed proceeds from a central assumption that diagnoses such as depression are more inherently difficult for patients to accept in situ than those of more obviously somatic conditions, as they are less well understood and/or more subject to social stigma (Wolpert, 2001;Barney et al., 2011;Griffiths and Crisp, 2013;Wang et al., 2015). There is a wealth of well-cited research in the CA and DP domains that addresses interaction in which a mental illness category is announced, applied or implied (Antaki et al., 2007a, b;Finlay et al., 2008;Horne and Wiggins, 2009;Tucker, 2009). Like the work on diagnostic interaction discussed above, this body of literature does not dismiss the influence of broader knowledges relating to the illness categories themselves, but rather highlights how the manner in which knowledges are themselves actively addressed tends to arise initially in terms of their local (rather than global) consequences. Horne and Wiggins (2009), for example, in a study of online suicide support forums, intricately demonstrate the difficulties experienced by individuals in getting their own suicidal ideation taken seriously in a context where others are making (and therefore usually evaluating) similar claims. Rather than simply and neutrally providing an account of how they think or feel, thus, members of the chat-groups remain demonstrably sensitive to local inferential concerns around how their own claims might be read (and their local social identity thus interpreted). Foremost among these is the matter that if one is: …not suicidal enough and you may be treated as 'just' depressed; 'too' suicidal and it may be challenged if you do not carry through your actions. (Horne and Wiggins, 2009: 170) In a broader DP study of schizophrenia, meanwhile, Tucker (2009) explores how individuals who have been diagnosed with the condition subsequently make sense of their diagnosis. The emergent findings highlight how the concerns of the specific individuals regarding schizophrenia, in terms of how they are relayed during an interview, are similar to those arising within Miller's (2013) study of questioning around suicidal inclination. Effectively, so long as the (respectively) interviewer or GP handles their co-interlocutor's own activity as a full and rational contribution to the design of questions and reaching of conclusions, then the apparent stigma of the matter often goes almost unrecognized. This does not mean, once again, that the emergent category does not potentially carry social "weight". Rather, that the implications of categorization are best firstly explored in terms of their immediate import for the face-to-face context itself.
The extant body of interactional research detailed above indicates, above all, that the concerns of patients in clinical interactions stem primarily from their interest in being (at the very least) constituted as capable participants in the local interaction where otherwise they might be seen as passive recipients of institutional labels, particularly where given diagnosis/advice implies an inferably mental illness (Antaki, 2007;Antaki et al., 2007c;Finlay et al., 2008;Miller, 2013). In the specific terms of depression diagnosis, then, we might expect that practitioners and patients alike might be empirically seen to orient to how the categorization of the latter might affect the particular interaction (Peräkylä, 1998;Antaki et al., 2007b;Miller, 2013). This is (once again) not to propose that no patients, during clinical interaction in primary care, can behave on the basis of the incomplete or globalizing ideas about depression proposed in the conventional social-cognitive model outlined above (Van Voorhees et al., 2005Cook and Wang, 2011;Wang et al., 2015). Rather, it is the contention here that the active manner in which depression is diagnosed by a medical professional, and the way in which this diagnosis is received and then handled, might firstly be explored in terms of the microdynamics of empirically available evidence, and with respect to the concerns of the individuals involved.
The DP and CA materials above, then, regarding the practical business of giving and receiving diagnoses in clinical contexts, and talking about mental illness categories across a range of contexts, render relevant three key research questions regarding the data-set reported below, specifically addressing what are conventionally deemed "troublesome" receptions 5 of a new depression diagnosis in primary care (that is, those where the patient actively challenges the medical opinion given, see Cook and Wang, 2011): 1. How do the clinicians formulate a new depression diagnosis, and with reference to what prior patient activity? 2. What order of activity does the patient's resistance make relevant from the clinician? 3. How is concordance reached, or not reached, and as a consequence of what, in each case?

Methods
The following analytic sections employ the contemporary interactional-sequence model of DP approach advocated by Wiggins (2017), and as executed in line with the work of Miller and Benkwitz (2016) and Patterson and Fyson (2016). Ideally, the core epistemology of the approach, a non-cognitive empirical stance on social interaction, is largely evident in the discussion of pertinent research advanced above, though a more formal ontological statement on the position can be found, however, in Potter (2010) 6 . DP was chosen over CA as the core approach, given the focus on interaction about a specific issue (depression/resistance), rather upon the broader nature of clinical interaction itself (Wooffitt, 2005;Wiggins, 2017). In the service of this investigation, the data addressed, in the form of recorded doctorpatient primary care consultations, were collected exclusively in UK primary care settings. Analytic focus herein falls, and as noted, specifically upon a specific consultations where a patient either contests or questions a diagnosis of depression immediately after that diagnosis is delivered, actions characteristically defined as Ethical considerations. With full approval from the pertinent NHS ethics committee, all patients at participating surgeries across the duration of the data collection period were asked if they might permit their consultations to be audiorecorded, such that they might work with a research project about potentially improving doctor-patient communication. In this way, new depression diagnoses could be captured, and recordings not relevant to the purposes of the project could be deleted securely at source. Recordings were collected (and, where relevant, deleted) by the GPs themselves, with no researchers present, to minimize disruption to the medical process. Before initially signing the consent form, participants were advised of the steps to preserve their anonymity (that is, that all dates, places and names would be removed from transcripts, and that no details of the location or timing of the broader project would be made publicly available), and were also made aware that there would be a further 2 week "cooling-off" period subsequent to their consultation, during which they could withdraw their contribution if they saw fit.
Participants. Across the data collection period, N = 15 new diagnoses of depression were collected, and recordings of the full consultations in which they arose were then returned. Within this corpus, nominally unproblematic transitions between diagnostic delivery and talk-about-treatment were not a universal feature. Instead, in a fifth of cases (N = 3), the patient raised issues about the diagnosis itself that prompted the GP to engage in a visible (re)negotiation of what depression is, means or implies, until some form of concordance was finally reached.
Analysis. All data were subject to multiple hearings to transcription; transcription itself was then executed using full Jeffersonian conventions (see Hepburn and Bolden, 2013, and Box 1). Analysis proceeded in line with the systematic principles outlined in exacting detail by Wiggins (2017), regarding how conversational data should be interpreted within the DP tradition. The core aspects of these are summarized by Miller and Benkwitz (2016, 43): Firstly, any example of spoken discourse is concurrently a constructed phenomenon, and a constructive one. Secondly, talk is itself actionoriented. It is used in the service of a variety of interpersonal activities… Thirdly, talk is always situated. It is produced and understood according to its position within a broader sequence of discourse.
Throughout the following analyses below, specific sections of talk are located using a notation that indexes the extract and line; thus, [E2/L6] indicates something said in extract 2, line 6. Also, therein, "Dx" designates the GP and extract number (x) and "Px" the patient and extract number (x).

Findings and Discussion
Consider now E1, in which the following activity is visible: 1. A diagnosis of depression is delivered by the GP; 2. That diagnosis is directly questioned/challenged by the patient; 3. A period of negotiation around the meaning and or implications of the diagnosis ensues; 4. A concordance on the veracity of the original diagnosis is reached; 5. The consultation proceeds to a discussion of treatment.

Extract 1
[P1 has previously discussed feeling low, exhausted and "empty". Immediately before this extract, D asked P if he has ever felt low enough to harm himself. P has politely but firmly denied this.] 1. D1: okay (.) right (.) then 2.
( or low and just not well (.) like you said 25.
(.5) 28. D1: and I think the best explanation here is a:: little bit of depression 29.
(1.0) 30. P1: right (.) ukay ((consultation continues to discussion of treatment options)) The immediate pre-diagnostic turns, in this case 7 , do not evidence any of "scoping" work by the D1 consistent with that described by Maynard (1991Maynard ( , 1997; rather, diagnostic questioning segues directly into diagnostic delivery. Of particular interest here, however, is the formatting of the diagnosis itself [E1/L1-7]. Note how the diagnostic news is delayed within the turn, mitigated (as "not serious"), and the talk itself is subject to a discernible series of hesitations. Such dispreferred action turn-shapes (Sacks, 1992b, p. 414), when produced in everyday talk, typically demonstrate an orientation by the speaker to the possibility that the news they are about to break (or question they are about to ask) may not be well-received by their co-interlocutor. Moreover, and as noted above, expressive caution in medical news-delivery sequences often heard by a patient as foreshadowing of a "difficult" disclosure (Silverman, 1997;Miller, 2013). Given this interactional property, tentatively formulated diagnoses can therefore be actively used by a doctor to forearm the patient for this very outcome, at a point where the consultation might normally proceed directly to a discussion of treatment options (ten Have, 1995; Peräkylä, 2006). In short, framing the diagnosis here as uncertain (or hesitant): (a) Is indicative of D1 anticipating that there are insufficient contextually-available resources to suggest that the patient will find the diagnosis unproblematic, and; (b) Opens up space for P1 to ask questions about a "delicate" diagnosis in a way that framing it as an incontrovertible fact does not. Regarding the former, the P1 can reasonably initiate dialogue about the doctor's opinion. With respect to the latter, anything other than an agreement would likely be hearable as a direct challenge to D1's contextual authority, with potentially damaging consequences for "social solidarity" (Silverman, 1997) within the consultation itself.
So, in this case at least, we might see the patient's ostensibly resistant activity to their diagnosis of depression as to some extent an invited phenomenon, rather than direct evidence of an inherently "negative attitude," or incomplete/distorted knowledge.
Equally noteworthy here is the manner in which the initial uptake on the diagnosis is accomplished by P1. No immediate contest is provided, but rather a minimal utterance in E1/L9 ("erm?") is the first response. As Heath (1992) observes, a very common way in which a patient can signal unproblematic accordance with the diagnosis provided is via a silence or "marked acknowledgement" 8 (Heritage and Sefi, 1992) in the first available turn thereafter, thereby displaying (or at least implying) agreement with, or deference to, its validity.
By withholding response, patients not only provide the doctor with the opportunity of developing the consultation as they so wish, but preserve the objective, scientific, and professional status of the diagnosis or medical assessment; the silence or acknowledgement operating retroactively to underscore the significance of the practitioner's opinion of the condition. (Heath, 1992, 262) However, neither silence, nor marked acknowledgement, is the outcome of the first turn after the diagnosis here; rather a minimal utterance is used by the patient in E1/L9 ("erm?") which ultimately leads to a series of negotiations about the diagnosis itself. Heritage (1984) similarly demonstrates how the production of this order of particle routinely projects further talk on the part Box 1: Jeffersonian transcription notation (.5) The number in brackets indicates a time gap in seconds (that is, in this instance, five tenths).

(.)
A dot enclosed in brackets indicates a pause in the talk of less than two tenths of a second. ·hh A dot before an "h" indicates an in-breath by the speaker. More h's indicate a longer breath. hh An "h" indicates an out-breath. More h's indicate a longer breath.

(( )) A description enclosed in double brackets indicates a non-verbal activity. -
A dash indicates a sharp cut off of the prior word or sound. : Colons indicate that the speaker has drawn out the preceding sound or letter. More colons indicate a greater degree of "stretching" of the sound. ( ) Empty brackets indicate the presence of an unclear fragment in the recording.

(guess)
The words within a single bracket indicate the transcriber's best guess at an unclear fragment. .
A full stop indicates a stopping fall in tone, not necessarily the end of a sentence. , A comma indicates a continuing intonation. ?
A question mark indicates a rising inflection, not necessarily a question. * An asterisk indicates a "croaky" pronunciation of the immediately following section. ↑↓ "Up" and "Down" arrows represent a rising or falling intonation, respectively. CAPITALS With the exception of proper nouns, capital letters indicate a section of speech louder than that surrounding it. º º Degree markers indicate that the talk they encompass was noticeably quieter than that surrounding it. underline Indicates speaker emphasis Thaght A "gh" indicates a guttural pronunciation in the word. 4 o "More than" and "less than" signs indicate that the section of talk they encompass was noticeably quicker than surrounding talk. = "Equals" indicates contiguous utterances. [ Square brackets between adjacent lines of concurrent speech indicate the onset and end of a spate of overlapping talk.
of the producer; a freestanding "oh" (like an "erm") can, for example, be both used (and heard) as an indicator of "withholding" an issue, or a reluctance to elaborate it. Moreover, and as clearly evident in E1/L11, it often precipitates further work from the co-interlocutor in terms of clarification. This activity is similarly present in extract 2: (1.0) 6. P2: snuthing rea::lly (.5) jus: (.) um 7.
her doctor down at [CONFIDENTIAL] university but she used to ring up and (.) well (.) not 11. say anythin (.5) sit there in silence or (.) she:d burst into tears for no reason 12.
(.) and keep saying that she was gonna (.) you know (.) hurt herself and 13.
( Herein, the material of the diagnosis is not only marked by D2 as "delicate" through mechanisms highly comparable to those noted above, but D2's own activity, like that of D1, explicitly coopts P2 and P1 respectively [E1/L5; E2/L2] into the outcomes. As such, and as outlined above, co-accountability for the diagnosis itself is embedded within the diagnostic turn; conversantly, P2 uses the "oh" particle (Heritage, 1984) to open up conversational space in the way that P1 used "erm." From a DP perspective, however, it is the subsequent activity in both cases that is of significant import.
Ultimately, in the cases above, whether either patient was indeed psychologically predisposed to object to the diagnosis of depression or not, the cautious formatting of the diagnostic turn provides them both with (a) a reason to question or challenge it (that is, the GP has himself treated the diagnosis as a "sensitive" one) and (b) an opportunity to do so. In both extracts so far explored, this is then exactly what proceeds to occur. Rather than volunteer the explicit assent or dissent which is now conditionally relevant, for example, P2 makes available further reluctance to produce it at all, downgrading its importance with the claim that the issue being withheld is "nothing really" [E2/L6]. While this ultimately constitutes a weak acceptance of the diagnosis, at best, it serves an important moral function within this interaction. As a number of studies have described, active resistance to a diagnosis is relatively unusual in any form (Frankel, 1984;Heath, 1992;Peräkylä, 1998Peräkylä, , 2006. Indeed, for P2 to explicitly contradict D2's judgment in this case would run a range of local, contextual risks, not least in providing for an interpretation of his own character as that of an unreasonable, disruptive or aggressive person (Miller, 2013). By displaying a hearable deference to proper conduct in the consultation via the overt restraint of the turn, P2 works to uphold D2's contextual authority and, reflexively, negate inferential damage to his own identity. In E2/L8, D2 then issues a further encouragement token and P2 finally produces an autobiographical account as further evidence of such deference, as a dutiful (yet reluctant) response to D2's second request. P2 thereby distributes accountability for its production between both speakers. This mirrors the activity in E1/L16-17 very closely; herein, P1 also finds an initial means of calling the diagnosis into question without directly contradicting D1.
Being depressed and being "ordinary" as primary concerns in initial non-concordance. It is particularly noteworthy that in both cases above, the patients' subsequent diagnosis-related discourse not only makes relevant further explanatory action from the GPs, but also concurrently provides a warrant for the patients having not simply agreed with the diagnosis in the first place. In E1/L13-16, P1-while markedly acknowledging the potential veracity of the depression diagnosis-provides a framework for an alternative reading of the symptoms he has previously made relevant (that is, he is "just a little run down"-itself a reformulation of D1's own diagnostic turn). This is grounded by aligning depressed persons with persons who are potentially "suicidal" [E1/L15], and then disassociating himself from suicidal activity. In E2/L9-12, meanwhile, P2 ventures an autobiographical account through which to problematize the diagnosis, directly contrasting his sister's experience of depression with his own: well (.) it's jus that my sister was (.) ahh. she got told she was depressed by her doctor down at CONFIDENTIAL university but she used to ring up and (.) well (.) not say anythin (.5) sit there in silence or (.) she:d burst into tears for no reason (.) and keep saying that she was gonna (.) you know (.) hurt herself and.hhhh … well (.) I'm not like that In both cases, adopting a social-cognitive stance, we could potentially propose that activity of this order exposes false beliefs around the character of depression itself; to be considered depressed, one must have some fairly extreme symptoms (respectively suicidal ideation or self-harming threats) and, where these symptoms are not present, the patient might not be considered depressed. We might also propose, with particular reference to E2, that this false belief is rooted in unpleasant personal experience with a family member. Notwithstanding the fact that individuals who have personal experience of depression (first-or second-hand) tend to understand more about the condition itself, not less (Link et al., 1999), it is observable that P2's account is finely constructed to achieve this sharp contrast between his sister's condition and his own.
Firstly, the protagonist in P2's narrative is explicitly identified in terms of a familial category; she is P2's sister. As with all mobilized social identity categories, the specific selection demonstrably attends to inferential concerns arising from account itself (Antaki and Widdicombe, 1998b;Antaki et al., 2007b;Miller and Benkwitz, 2016). In this case, a sister is somebody about whom the speaker has rights to possess reliable information, and to be concerned. Secondly, the contextual relevance of the story is established; his sister was told she was depressed by a doctor, just as P2 has been. Her identity as university student is also occasioned, which makes further available a number of inferences pertaining to her general age, level of intelligence, lifestyle and so forth. Moreover, since P2's own student identity has itself been a salient topic throughout the ARTICLE PALGRAVE COMMUNICATIONS | DOI: 10.1057/palcomms.2017.33 consultation, he makes relevant a further equivalence between his case and that of his sister. In E2/L10, however, the specific description of his sister's depressed behaviour is prefaced with the contrast marker "but." In this way, D2 is furnished with resources from which to infer that this is where the similarities between the two cases end.
The important point here is that the account is not just an account of irrational or unusual behaviour, but is organized to accomplish this exact character (see Smith, 1978). Firstly, "not saying anything", upgraded to "sitting in silence," is not, in itself, particularly bizarre. Indeed, in many contexts such behaviour would be decidedly appropriate. In this account, however, the action is located within the broader context of having "rung up," and actually being the caller makes relevant the obligation of having something to say. While people may (and do) regularly call to say nothing of substance, "normal" people do not generally ring up to say nothing at all. Similarly, crying is not in itself a particularly unusual activity. Note, however, the construction here: "bursting into tears." This is a rather more abrupt formulation than, say, "starting to cry." The behaviour is framed as an uncontrolled one for which, P2 explicitly claims, there was "no reason." All emotion displays or claims are, however, subject to situated entitlements (Edwards, 1999). For a strong emotional reaction such as "bursting into tears" to be heard as normal, a proportionate stressor needs to be discerned. By stripping his sister's behaviour not only of an appropriate reason, but of any reason at all, P2 discursively disaligns it entirely with activity which is "ordinary", reasonable or rational.
The final part of the account, meanwhile, emphasizes the more explicitly persistent character of a more hearably dangerous behaviour; the sister would "keep saying" that she would hurt herself [E2/L12]. This utterance serves to discursively embed self-harming intent (or an active ideation, at the very least) in the general sense of depression being worked-up here and, reflexively, to mark an implicit point of contrast between P2's own case and that of his sibling. In this account, his sister makes (a) active, (b) repeated and (c) unambiguous threats regarding self-harm.
In E2/L9-12, thus, P2 produces a factual account which, in its entirety, makes available a particular understanding of the character of depression itself, grounded in the details of his own concrete personal experience. He then marks the whole narrative as "not like" his own behaviour [E2/L16]. This marks an explicit, situated disalignment between himself and his (depressed) sister, thereby making available an inferable asymmetry between his own symptoms and depression itself. As such his resistance to the diagnosis is achieved indirectly, circumventing any hearably direct challenge to D2's authority as diagnosisprovider. To these ends, P2 mitigates this final claim by producing a body of evidence for making that very claim. Moreover, and embedded in this activity, is some complex, situated identity-achievement; P2 works consistently to construct an ordinary, everyday and "rational" identity for himself. He organizes his account so as to emphasize its factuality and relevance, constructs his argument a logical deduction and, by working-up the sheer irrationality of his sister's behaviour within the account, provides D2 with a contrast-framework from which P2 can be judged as very normal by comparison. Moreover, he does all this while sidestepping any direct challenges to D2's authority, and upholds the local social solidarity throughout. In this respect he does what Harvey Sacks (1984) famously terms "being ordinary". He orients the listener to the everydayness of his own actions, character and reasoning process and also avoids any activity that could potentially seem contextually unusual or irrational (such as directly challenging a GP's opinion during a consultation, and so forth).
In the light of this, it is also possible to reflect upon the activity of P1 in E1/L13-16, who appears to request that D1 consider an alternative candidate diagnosis: that he is just "a bit run down." Rather than take his apparent conflation of depression and suicide as a transparent representation of a false belief about depression (likely stigma-related), we might instead consider this a similar contrast-formulation that encourages a view of P1 as relatively normal in the circumstances. In short, it is another exercise in doing "being ordinary".
"Rational" identity in action. Thus far it has been demonstrated, in both E1 and E2, that the diagnostic turns and the immediate receptions thereof are clearly sensitive to a range of prospective inferences regarding the patients' social identities, not least those pertaining to the patients' states of mind (Horne and Wiggins, 2009;Tucker, 2009;Antaki, 2013), that a diagnosis of depression can make situationally relevant: 1. The formatting of the diagnostic turns constitute depression as a "delicate" object within the consultations, but also open up space for patient-initiated questions. 2. The patients, using a variety of techniques, employ this discursive space to assert the ordinariness of their cases in comparison to some more extreme formulations of what depression "might be." 3. They achieve (2) without making direct challenges to the local authority of the GPs, thereby preserving the core orderliness of the consultation itself.
These coordinative activities, thus, provide cause to consider that the post-diagnosis activities of P1 and P2 do not transparently represent any form of attempt to resist, obstruct or alter their diagnoses at all. Rather, by doing what they do in the way that they do it, the patients carefully open up further discursive spaces within their respective consultations that might not otherwise have been available (Peräkylä, 2006;Ijäs-Kallio et al., 2012), such that the local implications of that diagnosis might be further addressed. In short, they invite the physicians to collaborate in affirming the ordinary identities proposed within their ostensibly non-concordant activity. In both cases, this is precisely what is forthcoming in the GPs' responses to the patients' concerns [E1/L18-28; E2/L18-20], and the consultations subsequently progress unproblematically to discussions of treatment options. At the core of each is an explicit acknowledgement of how everyday the patient's case own (and thereby social identity) apparently is; the patient is locally constituted as a normal person with depression, rather than a depressed person (ten Have, 1995;Miller, 2013). In E1, this involves D1 recognizing (and echoing) the correctness of P1's own analysis; that he does just indeed seem "low" and "run down". In E2, D2 does this by implicitly and explicitly recognizing P2's own agency and lucidity; this demonstrates an acute sensitivity on the part of D2 to both the contextual implications of the diagnosis itself, and also to those of P2's story about his sister. In neither of these cases does the GP's interactional work in establishing concordance lend to an easy interpretation of the data as evidencing an attempt to "educate" the patient, as is typically proposed within contemporary clinical guidance (Petit and Sederer, 2006;National Institute for Clinical Excellence, 2016). Propositional education is, by any common-sense definition, premised in part on the effective transfer of novel information to correct incomplete or distorted knowledge. Consider, however, the relevant turns in E1 [L18-28].
feeling really down and hopeless (.5) an a lot of the (.) time yes it is (.) 20.
or low and just not well (.) like you said 25.
(.5) 28. D1: and I think the best explanation here is a:: little bit of depression The pertinent point here is that the information "released" by D1 in this particular description of what depression really implies cannot be considered new unless we are to attribute P1 a very poor memory indeed. It has been established previously (less than a minute ago within this very consultation) that P1 does not have suicidal thoughts. A diagnosis of depression has nevertheless been made and, moreover, D1 also explicitly references earlier stages of the consultation in which P1 had defined his own symptoms: "low and just not well, like you said" [E1/L24]. As such, it is readily available from local context that these things, both of them consistent with the state of being "run down", are symptoms of depression.
In E3, further discursive action of this general order is visible. In this case, the diagnostic turn is once again delivered highly cautiously, with D3 explicitly co-implicating P3's own perspective into its cautious and explicitly modalised, rather than factual ["I'd say"-E3/L2] shape, a feature that Maynard (1992) notes is often present in clinical formulations where a "bad news" receipt is anticipated.
I'd say (.) given everything you've to:ld me (.) that you have a bit of 3. a (.) depression 4.
(.5) 8. P3: I was just going to say that (.) um (.) I still go out and have fun (.) and it's only 9.
sometimes that I get (.) upset and silly (.) yesterday I had a brilliant day and.hh 10.
(.) so today I was wondering if I should come to see you (.5) at all 11. D3: you did the right thing (.) depression can be a (.) sporadic thing (.) it won't stop 12.
you being happy all the time 13.
(1.0) 14. P3: so (.) a bit of depression (.) then? 15. D3: yes ((continues)) Herein [E3/4], P3 initiates a section of overlapping talk to which she herself orients as an interruption (via an apology) within the same turn. D3, however, releases task and topic control (ten Have, 1995) and hands the floor to P3 herself, retroactively legitimizing her activity, and opening up space for her to proceed. She then provides an account of her recent behaviour, which foregrounds the relative normality of most of her life and the relative scarcity of her episodes of "getting upset and silly" [E3/ L8], and further downgrades the hearable severity of the general situation with an account of her reasoning process regarding the need for a GP at all [E3/L9]. Like that of P1 and P2, her discourse here is demonstrably sensitive to prior activity in the consultation (in which she has discussed her emotional state extensively), the implications of the diagnostic category and the manner in which that diagnosis was delivered. D3 responds to these by (a) acknowledging the good sense of her decision to see him, and (b) affirming that depression can indeed be (by implication "in this case") sporadic. The consultation then proceeds.
In all three extracts explored in this paper, this broad model of discursive action would seem to provide a stronger account of activities than the social-cognitive orthodoxy. This is not, for one moment, to suggest that there are, therefore, not cases in which something more akin to the documented localized education model does operate. Instead, the point is to begin with a closer exploration of the sense of what individuals are actually and cooperatively doing, rather than begin with an assumption. From this point of view, it is visible that the need for a patient to be seen as a "rational" teller of their own story-a "credible witness" (Zimmerman, 1992)-can come to the fore when the trustworthiness of that story is prospectively under threat from the stereotypical implications of being seen as a person with a mental illness, or the corollary concern that they might consequently be viewed as an unreliable witness to a report of any other circumstances. Nevertheless, in the three cases addressed above, even these issues are managed within the sequential unfolding of the interaction, and with respect to the practical business of opening up discursive space without, for example, directly challenging a clinician's opinion.
Finally, it is imperative to clarify that none of the above is designed to propose that, even in the extracts examined, there is not some order of "education"-addressing misunderstandings around depression-potentially going on. P1's explanation that depression is "not just about feeling really down and hopeless," for example, or P3's clarification that depression "can be a sporadic thing" might well serve that function. In each case, the GP is clearly explaining that one can still be depressed without displaying severe symptoms. In so doing, however, they actively preserve their patients' claims to normalcy, while nonetheless sustaining the position that those patients are mildly depressed. We might, thus, argue that the "education model" is operating, albeit in a limited way, in all of these cases. It is not, however, the primary concern in how the tricky clinical moments are resolved; education-talk itself is visibly a device used here in the service of preserving the patients' ordinary identities.

Conclusion
The analytic work presented in this paper has explored interactional issues underpinning situated patient resistance to diagnoses of depression (that is, to question or dispute the diagnosis in the first available turn). The first, and most fundamental, of these is that such activity can be an invited phenomenon, arising from the GPs' formulations of depression diagnoses as (potentially, at least) "bad news." It would certainly appear, in the three cases above, that the GPs make an assumption (either from experience, or received medical wisdom, or both) that the diagnoses they are delivering are likely to be met with negativity. The consequence of this is that they actually create the interactional space for the diagnoses to be questioned. By endeavouring to ease the passage of the diagnosis, they engender the exact result that the structure of the diagnostic turn anticipates. As such, at the local level, the notion that diagnoses of depression are contextually troublesome is potentially ARTICLE PALGRAVE COMMUNICATIONS | DOI: 10.1057DOI: 10. /palcomms.2017 further perpetuated. This activity is, however, borne of highly constructive function.
A key matter addressed in relation to these consultations relates to why, having taken the time to question their depression diagnoses, these patients are so apparently ready to then go on to accept them within a few turns. The analysis has, in these cases, questioned the efficacy of two of the more conventional psychological explanations for this kind of ostensive volte-face; namely, underlying (or explicit) power differentials between GP and patient (Wodak, 1996;Wileman et al., 2002), and the silver bullet model of instant education (Griffiths and Crisp, 2013). Indeed, on closer inspection of the data, what at first seem to be acts of uncertainty and/or negativity in diagnostic delivery transpire to be highly practical strategies on the part of the GPs to open up the floor, encouraging the patients to negotiate with them what the diagnosis does or does not mean in this case. Literature from both discursive and clinical domains converges on the assertion that, for a diagnostic phase to progress into a cordial discussion of treatment options, it is highly desirable that both GP and patient reach some manner of accord with respect to the character of that diagnosis (Heath, 1992; National Institute for Clinical Excellence, 2016). By "doing hesitancy" and then relinquishing their own task and topic control within the diagnostic phase, the GPs in the cases above allow for exactly such a co-construction of dialogue around the potentially awkward diagnosis of depression.
This renders germane the second core issue to emerge from the analysis, relating to what Harvey Sacks (1984) terms doing "being ordinary." In the business of questioning their diagnoses, each of the patients actively and concurrently works to build a selfidentity that is inferably rational and "normal." Orienting to a range of inferential issues arising from being situationally categorized as a depressed person, and also from their own activity in having questioned a diagnosis at all, they build accounts (grounded in a variety of occasioned details from their own everyday lives) that are intricately designed to forestall their selves (or their recent behaviour) being viewed as non-rational or, indeed, obstructive on account of either. In their subsequent explanations, the GPs produce turns that are sensitive to these concerns, collaboratively renegotiating the implications of the diagnosis itself. The patients then actively accept their diagnoses under these terms, with explicit endorsement from the GPs; in short, a mutually satisfactory concordance on the nature of that diagnosis is found (Ijäs-Kallio et al., 2012).
Fundamental to the analyses herein, therefore, have been illustrations of the constructive functionality of apparently dysfunctional talk by medical professionals (Silverman, 1997;Miller, 2013). For example, the widely-cited UK guidebook "Depression in General Practice" (Tylee et al., 1996) proposes that when talking about depression with a patient, GPs should be as "direct" as possible. The diagnostic turns above display activity that is far from direct, but it is also difficult to sustain an interpretation of its as evidencing a "lack of confidence" or indeed "poor practice" when taking into account the positive interactional outcomes. In these cases at least, therefore, by drawing upon tacit social skills rather than normative frameworks, the GPs do not visibly exercise "power" in reaching concordance, nor educate the patients out of their depression-based stigmas with a silver bullet of new knowledge, but instead open a discursive space within which a process of further ad-hoc activity can take place, to demonstrably constructive ends.  Jungbluth and Shirk, 2009); and a non-compliance with treatment regimens, particularly where these involve anti-depressant medication (Sirey et al., 2001;Lu and Roughead, 2012). 2 Note: The currently dominant international guideline on the classification and diagnosis of depression, the Fifth Edition of the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM-5), presents an intricate set of nuanced disorders built on the description of a Major Depressive Episode (American Psychiatric Association, 2013). Since this paper relates to frontline primary care practice, however, adapted variants relevant to this specific domain (for example, National Institute for Clinical Excellence, 2016) are used as key touchstones throughout. 3 At the most basic level, for an alcoholic to talk about their condition to a room full of teetotallers is likely to involve a different order of embarrassment or awkwardness to the act of addressing fellow alcoholics. This is, to a great extent, the very founding principle of support groups-that some discursive spaces are "safer" than others even when the topic is the same. 4 Maynard (1991) himself illustrates that these are not unique to medical interaction, but are primordially a property of everyday talk used particularly around potential "bad news" disclosures. 5 Notwithstanding Stivers (2006) work on the "grey" ways in which diagnosis can be received; resistance here, in terms of a critique of the dominant social-cognitive model, is defined in terms consistent with that approach. See below. 6 Associated investigative issues around this broad discursive paradigm (as compared to, for example, speech-act theory and semiotics) are also outlined in Miller and Grimwood (2015) 7 And also those discussed subsequently. As such, the immediately prior turns are not presented here. 8 Typically "okay," "yes" or "alright."