Co-Production and Structural Oppression in Public Mental Health

Abstract

Co-production, in the field of mental health, aims to bring together academic and clinical researchers and those with lived experience. Often, research projects informed by this methodology involve the meeting of opposing attitudes, whether to the legitimacy of psychiatry, determinants of mental ill health, or the most appropriate interventions. This has meant that whilst some have reported positive experiences of co-production, many people with lived experience of mental ill health, sometimes referred to as ‘experts by experience’ (EbE), report harms which have taken place or been perpetuated during co-produced research projects. In the literature, nearly always, this is understood as a kind of epistemic injustice in Miranda Fricker's sense. In this paper, I argue that whilst Fricker's view does provide a plausible explanation of what's at play, we can gain more insight into the structural factors which exclude EbE by applying a framework of epistemic oppression. By highlighting the systemic and structural factors which work to keep certain knowers and their contributions out of our collective epistemic resources, we begin to understand the enormity of the task required to redress injustices in our knowledge production systems.

1. Introduction

Co-production, as a relatively nascent research methodology in public mental health, aims to bring together academic and clinical researchers and those who have lived experience of mental ill health. It is posited as a revolutionised approach to research by including those who have traditionally been researched as equal members of the research team. These research projects take place across mental health disciplines such as public health, primary care design, digital mental health interventions, etc., and bring together opposing attitudes to the legitimacy of psychiatry, to the determinants of mental ill health, and to what we ought to value. This has meant that whilst some have reported positive experiences of co-production, many people with lived experience of mental ill health, sometimes referred to as ‘experts by experience’ (EbE), report harms which have taken place or been perpetuated during co-produced research projects. In the literature these harms are almost always described as being a kind of epistemic injustice, drawing upon the work of Miranda Fricker (2007). In this paper, I'll explain how epistemic injustice provides a plausible explanation of the harms done to experts by experience in their capacity as authoritative knowers. I'm going to argue, though, that we might better understand such exclusionary and harmful practices through the lens of epistemic oppression. Epistemic oppression, in Kristie Dotson's (2014) sense, provides a distinct epistemic perspective which allows us to highlight systemic and structural factors which work to keep certain knowers, and their contributions out of our collective epistemic resources. The harms, I'll argue, are still epistemic in nature, but by extending beyond individual-level prejudice, we're able to see what would be required so that co-production is done well.

2. Epistemic Injustice: An Obvious Framework?

Oftentimes, discussions of harms perpetuated by the research environment are understood as a kind of epistemic injustice, in as much as experts by experience are often not treated as the right kind of knowers, despite being believed to have valuable knowledge of research foci. And this sense of a harm being done to someone in their capacity as a knower perhaps seems apt, given the focus of this paper. Co-production invites individuals with lived experience – understood as ‘experts by experience’, in many cases – to contribute to research from the earliest stages based upon their first-hand knowledge of the topic that is examined. Attempts to further and to deepen our knowledge of some topic or other, just are epistemic in nature.

Miranda Fricker (2007) developed what is the most widely accepted framework of epistemic injustice to explain the phenomenon of individuals, often from marginalised social positions or groups, routinely being dismissed or disbelieved. For epistemic injustice to occur, a judgement of an individual as a less credible knower (either relative to the credibility I assign to myself or compared to a competing source of information) is made, based on a prejudicial stereotype. Typically, this stereotype takes the form of a negative identity prejudice. On Fricker's view, we routinely use heuristic aids, such as stereotypes, as psychological shortcuts that aid our judgement and reasoning (Fricker, 2007). Consider the following example. If someone is invited to testify at a trial as an ‘expert witness’ and begins their testimony with a list of their academic achievements, I may assign what they say a greater level of importance than if they were a ‘bystander’, due to my prejudicial belief that academic qualification makes one particularly well suited to evidence-giving. In Fricker's terminology, I rely on my own heuristic aid concerning expert witnesses, which assigns high levels of credibility to academic ‘experts’, and take what the expert witness says at face value. What goes ‘wrong’ in cases of epistemic injustice, is that the stereotypes upon which I rely are almost always negative or ethically noxious in nature. These stereotypes are inversely correlated with judgements regarding competence and credibility. If my interlocutor belongs to a ‘group’ against whom I hold a negative identity prejudice (whether conscious or not) I may deflate or decrease the level of legitimacy I assign to what they say (Fricker, 2007).

Being perceived as ‘ill’ in any sense, whether physical or mental, can heighten susceptibility to experiencing this deflation of one's credibility as a knower. Carel and Kidd have argued that individuals with physical illnesses are more vulnerable to epistemic injustices than those in good health (Carel and Kidd, 2014, 2016). And mental ill health remains a topic subject to particularly pernicious and deeply entrenched negative stereotypes, such that vulnerabilities to being undermined or dismissed as a credible ‘knower’ are compounded (Crichton, Carel, and Kidd, 2017). Perceptions of individuals with mental ill health as being dangerous, unreliable, or irrational seem inextricably at odds with widely accepted epistemic virtues of honesty, and reliability.¹ These perceptions are not restricted to only one particular kind of psychiatric illness or diagnostic category. Decreased perceptions of the reliability of individuals with schizophrenia (Angermeyer and Matschinger, 2005; Corrigan et al., 2001), PTSD, and of mental health on a broad scale (Wahl, 1999; Guidry-Grimes, 2015) are well documented, to name only a few. Being invited to participate in research as an expert in one's own circumstance whilst simultaneously being open about having received a psychiatric diagnosis, then, may well render much of what an ‘expert by experience’ has to say vulnerable to scepticism or dismissal.

In fact, the very concept of epistemic injustice seems to have afforded some relief to those who have been treated by the psychiatric system, as one service user who now takes part in co-production explains:

I stumbled across the concept of epistemic injustice on Twitter […]. It helped me to make sense of my experiences of harm within psychotherapy and mental health services. I realised I was taught to dismiss my own knowledge and this had deeply affected my trust in myself and my confidence in what I know. From being disbelieved about my experiences to being told I should ignore what my body and mind were telling me […] epistemic injustice was everywhere […]. I have lost count of how many times I have told someone something about my mental health, only for them to turn to the ‘experts’ to confirm what I have said, as if I am an unreliable narrator of my own mind. What often happens is that service-user knowledge is only trusted if it is backed up by a researcher or professional. (Coproduction Collective, 2023)

Just as individuals have spoken of relief at realising there are established conceptual frameworks that make sense of their experiences in other areas,² the notion of epistemic injustice can be a useful tool for experts by experience to understand how and why their testimony has failed to gain the uptake intended.

I've focused here mainly on the testimonial ‘kind’ of epistemic injustice that experts by experience face as this has received the most attention and seems most apt for current discussions. What I think interesting about the application of epistemic injustice to cases where expert by experience testimony is excluded or afforded less credibility in co-production approaches to mental health research, is that the very reason individuals with lived experience are invited to participate in these spaces is because of their direct, first-hand acquaintance with mental ill health. This is not to say that the transposition of negative identity prejudices to the research context is all that surprising – we often allow biases to affect our epistemic conduct, even when we have strong intentions to the contrary – but rather, dismissal of expert by experience testimony in a research endeavour designed to include such testimony seems to be a particular instance of what Fricker calls ‘ethically bad affective investments’ (2007, p. 35).

In the domain of public mental health, the goal of co-production is not always to assess legitimacy or efficacy of existent approaches to psychiatric recovery or service design, but often to understand the etiological, psychopharmacological, or even socially driven bases of mental ill health. A holistic research picture, which is the aim of co-production, must take on board perspectives of those with experience and with lived insight. This makes the negative identity prejudices which carry over to the research context seem that much more pernicious, and make them instances of bad epistemic practice on the part of those deflating the credibility of experts by experience. Service users, carers, and those who have been subject to psychiatric treatment are known to have had, or perhaps to still have, direct acquaintance with mental ill health. Dismissal of their accounts as inconvenient, as illegitimate, or as unduly unreliable seems to go against the entire motivations for conducting this kind of research. In §3, I'll discuss in more detail the ways in which expert by experience testimony routinely fails to enter into our shared pools of knowledge.

Fricker's view, then, gives us one way of making sense of what's happening in such cases. Negative and prejudicial attitudes to mental ill health affect the heuristics upon which we all rely in judgement-making. In the context of co-production, those experts by experience who have been open about their illness or diagnosis, despite having been invited to participate in research based on their lived experience of that illness or diagnosis, routinely find that their words do not get the right sort of uptake to shift norms or to affect research conversations. The focus upon the ways in which social injustices lead to epistemic injustices captures nicely many of the facets of the discrimination faced in relation to mental health. What I think, and what I'll go on to say (in section 4), is that understanding the exclusion of expert by experience testimony or knowledge only gets us a narrow understanding of the picture. On Fricker's view, I, as an epistemic agent in my own right, am responsible for negative identity prejudices that I hold and the effects that these prejudicial beliefs have on the credibility assignments that I make. I can either rely on the faulty heuristic aids I possess and give short shrift to evidence of those I take to be members of particular categories on that basis; or, I can accept counterevidence, and revise the beliefs I hold, attempting to assign due weight to testimony of my interlocutors, and strive to be epistemically just in my interactions. This is a perfectly plausible way of making sense of how, in individual interactions, expert by experience testimony is subject to deflationary credibility assessments or is perhaps otherwise subverted. But, in the research environment, we are not only focusing upon individual interactions, and as such, the adage ‘it's not about me’ becomes relevant. The academic or clinical experts here have, presumably, committed themselves to undertaking a co-produced research project. We might expect that this means not all members of the research team hold negative identity prejudices toward those they plan to partner with, or at least that not all of the academic and/or clinical research team are making such ethically bad affective investments, repeatedly. To cast such an aspersion would be to assume that such research projects are undertaken in bad faith. And I don't think that's what's going on in these sorts of pictures. In order that research be done effectively and achieve any kind of advancement in what we take ourselves to understand or to do, there are norms, conventions, and broader requirements at play such that knowledge gains uptake and enters into shared understanding. It is not just what I, qua epistemic agent, do, but in order to advance knowledge, I must be able to affect a system level shift, or, at the very least, have my research findings enter into shared epistemic resources. Fricker's view may well accommodate this, as a series of repeated instances of credibility deficit assignment which undermine the legitimacy of knowledge offered up by experts by experience. But, in what follows, I'm going to say that making sense of the exclusion of experts by experience using an alternative framework of epistemic harm might help us to understand what's happening in these cases at a system level.

3. The Tensions at Play: Failures of Uptake

Prior to providing an alternative analysis of epistemic harm, it may be helpful to understand a little more about the tensions that are manifest in the research process. Understanding these instances, at the level of individual interaction as epistemic injustice is I think, as I've indicated, a correct appraisal. What I'm going to go on to say, though, is that we can similarly make sense of these sorts of exclusionary research practices at a broader systemic level by understanding them as epistemic oppression. The latter concept, I'll say in §4, allows us to understand why co-production as an approach to research is rife with epistemic harms. But on either view, understanding the ways in which testimony fails to gain uptake or to be received as legitimate knowledge helps to paint a fuller picture of the phenomenon this paper attempts to make sense of.

(i) Testimony received as anecdote

First, there is a purported tension between what is deemed ‘hard science’, that is, evidence-based and deriving from medical professionals, and anecdotal contributions to knowledge. Co-production, as a methodology which does not require those with lived experience to possess the same academic or clinical expertise as the research team qua accepted experts, is often seen as an ‘additional’ dimension of knowledge, bolted on to research projects rather than integral to them. Being invited to participate in research on the basis of your lived, experiential (and sometimes phenomenological) insights is often received as being invited to speak about those experiences. But experiences and recounting of thoughts and feelings on a personal level are not afforded the same level of scientific credence as the results of a pharmacological trial or a population level ethnographic study (Rose and Kalathil, 2019; Johnson and Martínez Guzmán, 2013). In fact, as Diana Rose and Jayasree Kalathil recount, in their (2019) article ‘Power, Privilege and Knowledge: The Untenable Promise of Co-Production in Mental “Health”’, individuals with lived experience are often invited to speak in the very spaces in which their ‘illnesses’ or ‘disorders’ are discussed in derogating or distressing terms by more typical academic researchers only moments later. This, they go on to say, is akin to being a subaltern in the research team and to having one's knowledge rendered unspeakable (2019).³ Even if you are sympathetic to the value of experiential knowledge and would agree that instances like those Rose and Kalathil describe are harmful and ought not to constitute the way research is done, you may also agree that present research hierarchies tend to privilege the quantitative or the ‘evidence-based’ as opposed to the qualitative or narrative kinds of knowledge (Crichton, Carel, and Kidd, 2017). ‘Lay’ research members (those who are often invited on the basis of their lived experience) are simply not afforded the power or control of research that would allow them to influence research's direction, or to challenge chosen methodologies such that their contributions ‘count’ in meaningful ways (Slade et al., 2010). Methods that are viewed as value-free and objective remain privileged over and above subjective or first-hand accounts of distress or service use (Faulkner, 2017). Knowledge that is not deemed objective and is thus value-rich or rooted in experience is – understood in a Foucauldian sense – ‘subjugated’ (Foucault, 1980; Brown and Strega, 2005, p. 11). The tendency to prefer ‘hard’ evidence over ‘soft’, then, and the habit of categorising expert by experience testimony as the latter, predisposes researchers to dismiss or distort the content of such testimony.

The harm, here, derives from experts by experience being recruited to such projects under the guise of equitable and significant involvement. Co-production is designed for that very purpose.⁴ By inviting an individual with lived experience to participate in research on the basis of their first-hand acquaintance with mental ill health, there is a (not unreasonable) expectation that their contributions will be valued. Accordingly, giving testimony relating to experience the status of ‘other’ or second-class knowledge seems particularly harmful as this testimony forms the very basis for the involvement of such individuals.⁵

(ii) Subversion of message – psychiatrisation or lacunas in understanding

Second, there is a tendency of clinical or professional members of the research team to, intentionally or otherwise, water down or otherwise subvert or pathologise the contributions of those with lived experience. This may (as we'll see later in section §4) have more to do with ingrained societal (mis)conceptions relating to the epistemic virtues, vices, and traits of those with mental ill health.

Typically, this may take the form of the individual with lived experience asserting ‘S’ in discussions, but the clinical or academic research partners on the project interpreting this as ‘S*’. Power dynamics carried over to the research environment often mean that experiential knowledge is offered up by those with lived experience but is ultimately defined by the ‘experts’ in the room. This is in many ways unsurprising when we consider the power that psychiatric professionals have over determining the rights, or removal of rights (in some cases) of their patients. Residual and engrained power dynamics transferring from the clinical setting over to the research setting may well affect which testimonies are assigned credibility, and conversely, which are not. Whoever is responsible for the writing up of meetings, the research progression, or the evidence-gathering aspects of the research process may find themselves reading into an individual's testimony a meaning that simply is not, or was not, that. And whilst this interpreting of testimony is in itself a harm, what compounds this particular harm is the way in which such testimony is routinely sanitised or made to fit dominant societal conceptions of mental ill health (Jones and Kelly, 2015). For instance, someone with lived experience might say that they ‘value the community which they have gathered around them during periods of ill health’ but this could, understood through the lens of psychiatric practice, be interpreted as ‘patient is unmotivated to stay well due to the care and attention they receive when less able to go about their day-to-day activities’. Such a subversion of meaning could be due to the habits of psychiatric practitioners in interpreting what patients say to them. This is often based upon habitual practices (not dissimilar to Fricker's heuristic aids mentioned earlier) whereby shortcuts are taken based upon experience: often patients demonstrate trait T and this could be an instance of T, despite the content of the utterance literally meaning S. Alternatively, this subversion could be due to a failure of shared frameworks of reference possessed by the expert by experience and the clinician.

As Luvell Anderson (2017) notes, often – and particularly when marginalised individuals or identities are at play – some conceptual resources simply aren't shared. Two conversational parties may believe they are talking along the same lines, but the more privileged party makes sense of what they hear based upon their understanding of the world. If this understanding doesn't quite track the understanding that the original speaker had, the end result can be that both feel as though they understand and have been understood by one another, however a gap in hermeneutic horizons leads to a lacuna in meaning or sense making (Anderson, 2017). Simply put: neither party can articulate or take away from an utterance something which they don't relate to. If a clinician or academic has not had the positive experience of finding oneself understood by a particular community in ways which differ (positively) from routine social interactions, that may not be the sense that they take from the above statement. However, they may feel as though they have understood it, through the lens of patient interactions previously had, and the expert by experience may have no reason to suppose that their words will take on a new meaning. And it is not only the words we say, and the way they are received that can affect the experience of co-production for those with first-hand acquaintance of mental ill health. Bee et al. (2015) take a similar kind of hermeneutic gap as being the cause of many failures in service user-involvement, stating that:

[…] service user involvement fails because the patients’ frame of reference diverges from that of providers. Service users and carers attributed highest value to the relational aspects but […] planning is typically operationalised as a series of practice-based activities compliant with auditor standards. (Bee et al., 2015, p.104)

Here, the hermeneutic gap relates to the most valued elements, either of a service or of the research process itself. Where the focus of academic or clinical experts is procedural, service-users involved can experience exclusion and disempowerment (Carr, 2016).

Further, individuals whose first-hand experience of psychiatric systems, or of forms of marginalisation in the social sense which are not shared by clinical or academic research partners, may be viewed as hostile or overly-critical of psychiatry as a branch of medicine. Mad Studies, and the anti-psychiatry movement, have gained a wealth of traction since the 1970s/80s and whilst criticisms of psychiatric practices are by no means restricted to these movements, they are dominant within them. Should an individual with lived experience report their believed illegitimacy of, say, psychiatric diagnosis in a research endeavour designed to interrogate the use of control and restraint in psychiatric inpatient settings, and the testimony of ‘lay’ researchers be overwhelmingly negative (as might be expected), this could be sanitised or subject to reinterpretation or subversion. The perceived hostility of those with lived experience – in that their testimony challenges accepted practices and standards – might lead to that testimony being omitted, altered, or otherwise changed (Hodge, 2005; Lewis, 2014).

Ultimately, what can be expressed is often determined by those in more traditional positions of power. The ‘rules of research’, as Marian Barnes (2002) notes, have not been transformed thus far and these rules ‘define both the way in which deliberation is conducted and who is considered to be legitimate participants in the process’ (Barnes, 2002, p. 329). Thus, traditional researchers retain the power to determine whose knowledge makes it into shared spaces and what narratives might be able to influence dominant understandings of research processes.

(iii) Duality of roles – legitimate knower, or mentally ill and lacking credence?

Thirdly, there are a multitude of roles which those with lived experience must simultaneously occupy. As Rose and Kalathil (2019) note, being positioned as an expert on the basis of one's lived experience in many ways renders one's legitimacy unstable and subject to variance on the basis of perceived mental state. As an individual with lived experience, whose lived experience is widely known in the research team, any expression of emotion or distress can serve to make one appear irrational, unwell, or unstable. And, of course, this has the effect of undermining the legitimacy of the knowledge conveyed in the same way that being perceived as mentally unwell affects the credibility or reliability you may be presumed to have in social and political spaces. Whilst the inviolability of professional knowledge is a given, or is received as a given, the degree of credibility the testimony or knowledge contributed by an individual with lived experience has depends upon the presentation of that testimony or knowledge. Becoming distressed, angry, or even appearing less animated than at previous meetings can result in the credibility your testimony is assigned being lowered, as there is an inverse (presumed) relationship between credence and legitimacy and heightened emotional states. This can also lead to what Liegghio calls ‘psychiatrization’ (Liegghio, 2013), where, like testimony subversion I outlined in (ii) above, what an expert by experience says is attributed to or presumed to be affected by their mental (ill) health. Heightened emotional state, or clinical assessment of the rationality or sanity of experts by experience can lead to either them or their testimony being pathologised.

Again, this can – almost equally – be attributed to societal attitudes to mental ill health and to the requirement for rationality in empirical science. Neither party views mental ill health as compatible with our ‘norms’ of epistemically virtuous practice. Mental ill health is viewed as in tension with rationality almost universally, and whilst it is not the goal of this paper to unpack such a tension, it is something which will be returned to in the following discussion.

4. Understanding Epistemic Harm as Epistemic Oppression

What the above scenario illustrates is a variety of ways in which the knowledge of experts by experience is simultaneously (mis)understood and undermined: either as not scientific enough, as a veiled disclosure of symptoms, or else as being lacking in authority or scientific legitimacy. These can all be explained in various ways, with various causes or reasons being pointed to, as the preceding section has shown. Whilst all of this can be explained at the level of individual interaction, I'll now move to more of a system-level view of epistemic harm, appealing to Kristie Dotson's framework of epistemic oppression (2014).⁶

4.1 Epistemic oppression: resilient systems

Dotson defines epistemic oppression as being ‘a persistent and unwarranted infringement on the ability to utilize persuasively shared epistemic resources that hinder one's contribution to knowledge production’ (2014, p. 116). Akin to the harm done to Fricker's epistemic agent in her capacity as a knower, an individual facing epistemic oppression is being undermined in her epistemic agency. However, Dotson's view specifically expands beyond the level of individual interaction and focuses upon the ability of an individual qua epistemic agent, to draw upon, contribute to, and shift shared epistemic resources (Dotson, 2014, p. 115). Whilst she acknowledges that (like instances of epistemic injustice) there are often social and political factors which undermine or impede the ability of an agent to make use of such epistemic resources, what differentiates Dotson's view from Fricker's is an account of third-order epistemic oppression. The central claim of Dotson's account of third-order epistemic oppression is that this kind of epistemic oppression is not wholly reducible to social and political factors, but ‘follows from a feature of epistemological systems themselves […] epistemological resilience’ (2014, p. 116).

Epistemological systems are, broadly speaking, systems which contain all of our epistemic habits, norms, and attitudinal beliefs. They also, following Taylor (2004) and Medina (2011), contain operative and instituted social imaginaries. Social imaginaries can be understood as shared collective understandings of what normal and desirable ways of living are. In the United Kingdom, our attitude to free healthcare as underpinned by equality of access, and the belief that one of the duties of the state is to provide public services, might be considered a social imaginary. They also help us to share common understandings of frequent dichotomies: acceptable/unacceptable beliefs or ways of living, normality and deviance, values and undesirable traits. Operative social imaginaries are those which tacitly govern our understandings and assumptions, influencing our perceptions and behaviour without us ever necessarily becoming aware of their presence. By contrast, an instituted social imaginary is more of an explicit, regulatory framework such as a legal Act, agreed terms of reference, or another codified behavioural schema; an instituted social imaginary may govern our actions and behaviour in much the same way as operative imaginaries, but we are much more cognisant of the latter (Taylor, 2004). Both operative and instituted social imaginaries partly comprise our epistemological systems. These systems, taken as a whole, affect what promotes or conversely what detracts from knowledge production (Dotson, 2014, p. 121). Whilst epistemological systems are by no means fixed or immutable, uncovering flaws in the system which governs your worldview of knowledge is a sort of meta-epistemic challenge. As such, revisions of entire epistemological systems are difficult to bring about. Dotson describes this as ‘experiencing the impossible as possible and, correspondingly, viewing the limit of one's epistemological systems that designate the possible as impossible’ (2014, p. 132). Bartunek and Moch similarly describe the incredulity one experiences when encountering the limits or drawbacks of one's own epistemological governance as being somewhat ‘mystical’ (1994, p. 28). And given the challenges associated with even identifying the limits of one's own epistemology, let alone the degree of paradigm shift required to remediate or redress injustices which are entrenched into the fabric of that system, or perhaps the imaginaries which it contains, our epistemological systems are highly resilient. Resilience, in this sense, relates to the degree of counterinformation which can be absorbed into the system itself without requiring a revision of the resources the system is comprised of (Dotson, 2014). Prudent epistemic practice, according to Medina, requires that ‘epistemic friction’ – the counterevidence mentioned above – be sought out frequently, such that epistemological systems be updated as alternatives appear and are established as credible (Medina, 2011, p. 29). Yet, when the counterevidence is such that it threatens to topple a well-established hierarchical view of knowledge, the tendency to dismiss or ignore the counterevidence, however credible it may appear, can obscure the limits of the system and be absorbed as anomalous.

Third-order epistemic oppression, then, occurs when an individual – either due to their social or institutional position, or to social and political factors which undermine their credibility – is unable to create sufficient ‘epistemic friction’ within the epistemological system that the research takes place in (Dotson, 2014). This friction would arise, should they be able to gain uptake, because of the incompatibility of their offered testimony with the system itself. However, possessing neither the power nor the epistemic virtues recognised by the system within which knowledge production is taking place, to gain uptake sufficient to make visible the limits of the dominant operative imaginary renders their knowledge incapable of entering in to the shared epistemic resources. We can thus present a range of reasons which might lead to the dismissal of testimony. Epistemic oppression, I'll explain below, occurs when (a) and either (b) or (c) apply:

1. (a) The position the utterer occupies is marginalised either socially or in a domain specific context such that their contributions are routinely met with suspicion;

   and either,

2. (b) the content of an individual utterance is at odds with commonly held beliefs or challenges norms of epistemic practice supported by the epistemological system;

   or,

3. (c) incorporating the content of the utterance into the epistemological system would render the system unstable.

If both (b) and (c) are the case, then, due to the revisions which would be required should the testimony be received as knowledge, it will likely be explained away. Instances of (a) and (b) or (a) and (c) will likely ensure that the individual contribution is delegitimised or otherwise viewed as irrelevant, misguided, or lacking in credibility. These latter scenarios relate not just to the marginalised position the individual holds as rooted solely in social and political system inequalities, but to the lack of power these individuals are perceived to possess in relation to the shifting of content of epistemological systems. I maintain that structural positionality and power are essential features of epistemic oppression, which is why (a) must be satisfied. The reasons I have stipulated that at least two of the three criteria must apply relates to what I term a ‘threshold for epistemic oppression’. To see why this threshold requires at least two of the three listed criteria, consider the following.

Taking (a) alone just gets us to an understanding of epistemic injustice, in Fricker's sense, based upon negative identity prejudices. If only (b) as a reason for dismissal of one's testimony were constitutive of epistemic oppression, then any belief or disagreement with a majority view might be considered an infringement upon your capacity as an epistemic agent. Maintaining the stability of the epistemological system as a reason for dismissal or taking some evidence less seriously alone, as in (c), doesn't necessarily constitute oppression either; the contents of such information, or who is providing it, would be required such that (c) is relevant. This is not to say, of course, that routinely reasons like (c) suffice to disregard evidence; on the contrary, I think many of us have dismissed something that doesn't ‘fit’ with all else we know to be true or that we value. As Dotson explains, epistemological systems can withstand a great deal of disruption. In actuality, I think it likely that all three of the above criteria will likely be present in most or all cases of epistemic oppression. Epistemic oppression then, occurs when information that you have, because of some fact about who you are, and the shifts to dominant epistemic resources that would be required if you were taken seriously (either due to their unsuitability in practice, or to the content of your utterance being at odds with what is commonly accepted) fail to enter into or bring about a shift in the epistemological system.

What makes this a distinctly epistemic kind of harm lies in what is required to redress the oppressive practice. As we're thinking at a system level, epistemic vices, virtues, and habits are all in play, and the norms we rely upon (which kinds of knowledge are privileged), the authorities we recognise as epistemically superior (whose knowledge is privileged), and the barriers to expanding our conceptions of good epistemic practice (what counts as good knowledge) can all be understood in distinctly epistemic terms. An individual might be marginalised based on socio-political inequalities and this may drive the misperceptions that affect the knowledge of that specific person entering into the collective domain, but the epistemic features of the epistemological system are such that challenging one's own (mis)perceptions would not suffice to redress system level epistemic oppression: our entire epistemological system would require revision.

4.2. Failure of design? Applying the framework of epistemic oppression to co-production

Thus far, then, I've explained what third-order epistemic oppression is and have given some illustration of the sorts of scenarios in which testimony might be subject to epistemic oppression – by requiring a shift or revision of epistemological systems, which an individual operating within the system is unable to bring about. What I've yet to explain is why the exclusion of experts by experience is especially well understood using this framework. What, for instance, makes our epistemological systems resistant to revision based on the testimony of individuals with lived experience of mental ill health? To begin, let's remind ourselves of the three ways I have outlined where testimony is dismissed or accorded less credibility: (i) testimony received as anecdote; (ii) subversion of message by those more powerful in the research context; and (iii) duality of roles.

In discussions of (i) I outlined the tensions between ‘hard’ and ‘soft’ science or evidence. Hard evidence is considered more robust, arising from the positivist mode of social research where fact takes precedent over value, and quests for knowledge focus upon that which is ‘invariable’ and ‘universal’ (Durkheim, 1982). As Vaditya outlines: ‘Qualities such as rationality, reason, objectivity, and impartiality are privileged over, and opposed to, irrationality, emotion, subjectivity and partiality’ (Vaditya, 2018, p. 274). Alternative forms of knowledge, such as those from the perspectives of marginalised people, were squeezed out of common accepted research practice during the growth of positivist modes of advancing knowledge (Kovach, 2005). As such, the academe, as the respected source of knowledge advancement, also tends to privilege fact and objectivity over and above opinion or values-led hypotheses. Our epistemological systems, and our operative social imaginaries, then, are established such that they prefer and uphold specific ways of doing research or arriving at new or expanded knowledge. And these constraints mean that testimony received as anecdote, such as first-hand accounts of distress, simply aren't afforded the status of ‘knowledge’ (Faulkner, 2017).

Consider also the ways in which Vaditya (2018) has characterised the qualities privileged in the research environment and how these map on to our social conceptions of mental health or, conversely, mental wellness. The epistemically more virtuous qualities of rationality, reason, objectivity, and impartiality not only indicate robustness in research, but good epistemic practices. Coincidentally, these also happen to be the very qualities presumed lacking if you happen to have lived experience of mental ill health. The negatively valenced qualities Vaditya highlights of ‘irrationality, emotion, subjectivity’ (2018, p. 274) are precisely those which are presumed inextricably linked to mental ill health. It just so happens that, according to dominant epistemologies, they also make for bad science and less than desirable epistemic practices. Being an expert by experience then, appears to bestow upon an individual in what Townley describes as ‘an epistemically disadvantageous social identity, akin to being given a version of the curse of Cassandra’ (Townley, 2003, pp.105–6). Townley is of course, not talking about research into mental ill health. But whilst in Greek mythology Apollo bestowed Cassandra with the prophetic ability to foresee the future, but simultaneously cursed her such that no one would ever believe her testimony (Townley, 2003), in cases of co-production, the phenomena at play are much less mystical. Rather, experts by experience are invited to participate in co-production because of their insights, but routinely have those insights fail to enter into the collective epistemic domain of knowledge, as that knowledge, and their very identity, fail to meet the positivist requirements of good science. This could be understood as an illustration of (a) or (b) as above, or perhaps some combination of the two. The content of testimony may be dismissed because of perceived inadequacies in the robustness of the evidence it contains because the expert by experience is not an academic or clinician themselves; lay persons cannot offer up hard scientific evidence, and testimony may be deemed too subjective. It could also be dismissed due to the perceived relationship between mental ill health and the epistemic virtues that the dominant epistemological system recognises as authoritative; irrational individuals cannot offer up rational evidence, as it were. Or, it could be that the content of expert by experience testimony is viewed, against the contents of the epistemological system, as being incompatible with accepted views and norms (in the case of anti-psychiatry, Mad studies, or other views which either disagree with or argue against curative interventions). Any of these scenarios would vastly affect when and to what degree expert by experience testimony gains uptake. I suspect it does not happen often.

In the rare circumstance where expert by experience testimony is received as truthful and reliable (though still anecdotal to a degree) but is deemed at odds with the existing and widely utilised epistemological system, it likely will still fail to gain the uptake required to act as a catalyst for revision of the shared epistemic resources into which she attempts to have her knowledge enter (Dotson, 2014, p. 130). And this, again, can be attributed to the extreme resilience dominant epistemological systems display, particularly when it comes to histories of social oppression, marginalisation, and injustice. Mental health, and psychiatric illness, have been subjugated categories in Westernised cultures throughout history, and thus anti-psychiatry or Mad Pride type views cannot be assimilated into the epistemological system without the resultant need to examine our entire worldview of mental health. The more ingrained into culture, institutions, and social understandings a view becomes, the more difficult it can be to challenge. Similarly, our research paradigms, funding processes, evaluative mechanisms, and the economy of academic education, which serve to further the successes of those who produce knowledge in accordance with the epistemic virtues and rules of the governing epistemological system, mean that conceptions of research are also entrenched (Vaditya, 2018). By taking this system-level appraisal of the harm perpetuated by co-production when experts by experience are excluded in research paradigms, and by understanding it as a kind of epistemic oppression, we're able to lay out this systemic injustice. Given the incompatibility, if my arguments here are accepted, between the variance in understandings of mental ill health with our dominant epistemologies, it is unsurprising that co-production has received criticism for dismissing the very knowledge it seeks out.

Thinking about the ways testimony is subverted, watered down, or misunderstood, as in (ii), we might also make sense of this phenomenon using Dotson's view. In her paper ‘Conceptualizing Epistemic Oppression’, she recreates the Allegory crafting an image of a row of fettered persons, facing to the left, who increase in their position of privilege from left to right. The furthest left individual is the only person able to see the remainder of the open cave, yet is also the most marginalised in terms of social position (Doston, 2014, p. 130). She has a unique position. When she attempts to share knowledge, using dominant epistemic resources she shares with others, e.g., language, conceptual frameworks, and so on, but which those further to the right have no direct experience of, she is met with ridicule. She occupies the most marginalised position in the cave hierarchy of power (Dotson, 2014, p. 130). Now her assertions may be met with mere disbelief, and those to the right of her might determine she does not occupy a social position commensurate with enough authority to enter knowledge (uncorroborated by more superior individuals) into the collective epistemic resources. Or, as I think more likely in practice, it's possible that those to the right of her attempt to make sense of what she says, based on their own experiences – their hermeneutic horizons. In either case, the content of testimony is changed or determined by others as lacking in credibility. And this is in addition to, or is perhaps compounded by, those features of the epistemological system outlined when discussing (i), which we might say are the basis for our expert by experience being situated in the position of greatest marginalisation. Similar arguments could be given for (iii). Positionality, as a non-expert research team member with expertise understood in terms of positivist qualities either exemplified by research methods or outputs, or by the individuals undertaking said research, will affect what an expert by experience is able to have taken seriously. The degree to which the knowledge she tries to share would create epistemic friction within the epistemological system will also determine whether she is able to enter what she knows into the shared epistemological resources. Whilst our conceptions of science, of research, of traditional ways of conducting these things, and our pernicious attitudes toward mental ill health remain integral parts of both the operative and the instituted social imaginaries which govern our epistemic practices, it is easy to paint a pessimistic picture of co-production.

5. Concluding Remarks: Why Appeal to System-Level Oppression?

Having explained how Dotson's epistemic oppression helps us to understand exclusionary practice in research, we might begin to question what this view offers over and above identity-based prejudice. Both pictures provide a plausible account of the sorts of factors affecting which individuals are deemed capable of knowing, and how that knowledge translates (or doesn't) to collective understanding.

On my view, Dotson's account doesn't necessarily explain what's at play more effectively, but it does help to capture two important nuances. First, as a system-based framework, epistemic oppression helps us to understand how knowledge is precluded from entering collective resources even in incredibly well-intentioned research environments. Individual researchers might be amenable to altering their research practices and may be sympathetic to views which would require significant shifts in epistemological systems that are dominant. But those individuals, too, are working within the same framework of epistemological systems, and in order that their research be taken seriously, and their outcomes be delivered in accordance with the terms of their funding, they cannot reinterpret or shift the system alone, nor can they step outside of the dominant epistemological system and continue to do research. Just as experts by experience face a series of double-binds, so too do the academic and clinical researchers. They also operate within the bounds of epistemological systems insofar as their own cognition and epistemic habits are concerned and are unlikely to have awareness of the limits of that system when it comes to the incompatibility of mental ill health with virtuous epistemic practice.

Epistemic oppression, then, gets us an understanding of an entrenched system of injustice, for which no one individual is culpable, but in which most of us are complicit. We may not even recognise this system as oppressive, and co-production in particular, as a research methodology designed to empower those with lived experience as active research partners, may be taken as a quest for epistemic justice, rather than a mode likely to perpetuate harm (Okoroji et al., 2023; Russo, 2023). And, there are accounts of co-production done well, where the issues I have laid out here are side-stepped, and experts by experience are able to participate as valued and respected partners (Faulkner et al., 2019). What this framework ‘gets us’ is an understanding of prejudices that run deeper than the level of individual interactions or personally held biases. It also goes some way to setting out the enormity of the task ahead, for to truly include individuals with lived experience in research, the epistemological systems governing research and epistemic habits in the academe would require substantial revisions. Co-production as a methodology is not one which is inherently designed to gatekeep knowledge, quite the opposite in fact. What this paper has demonstrated though, if the arguments I have given are accepted, is that for mental health in particular changes to research paradigms and social understandings of psychiatric illness are what would bring about change, rather than active participation of the typically researched in existing research culture. Epistemic oppression as an alternative to epistemic injustice merely helps us to understand how the system is stacked against those with lived experience of mental ill health.