A Framework for Thinking about Rationality

At the broadest level—in the academic literature as well as in an everyday sense— questions of rationality concern the normative constraints on decision-making: what we should believe, or what we should do (Kolodny 2005). The territory of rationality can be carved up in many ways (Wallace 2014), but in this chapter we will use two distinctions. The first is a distinction between procedural and substantive norms, where the former are concerned with the deliberative process by which a decision is reached, and the latter are concerned with matters of value—the ends that should be pursued. The second distinction, which cuts across the first, reflects divergent perspectives on why norms of rationality are binding—what makes these requirements required. One prominent explanation is that rational requirements are binding because accordance is essential for the pursuit of our purpose as agents (for discussion see Kolodny 2008). And one way of thinking about this purpose is in terms of getting the answers right—having true beliefs or choosing the right course of action. We will use the term epistemic to capture this perspective. This can be contrasted with a more practical orientation, focused on what makes things go well for the agent, and we will therefore refer to this as a pragmatic perspective.

There will be considerable overlap between these perspectives because in many contexts getting the answer right will be a good thing for the agent. However, they will also sometimes come apart. We illustrate this in relation to procedural requirements as these are our main focus in the chapter. From a pragmatic perspective it may not be rational for a person to hold a particular true belief if doing so would be paralyzing for them—a false belief might be more helpful in achieving their chosen goals or promoting their well-being.¹ Classic criticisms of utilitarianism—according to which the requirement to maximize utility is self-defeating because the deliberative process itself incurs a cost (Sidgwick [1874] 1907, pp. 489–490, Pettit 1991)—might be thought of as an example where these perspectives come apart. Further examples in the context of psychiatry will come to light throughout the chapter.

Is Psychiatry a Science?

The project of establishing whether psychiatry has scientific status is made difficult by the lack of agreement on necessary and sufficient conditions for a human activity to count as scientific research (e.g., Bortolotti and Heinrichs 2007). All the demarcation criteria that have been proposed so far, between science and non-science, and between science and pseudo-science, have been abandoned under the pressure of compelling counter-examples. As a result, philosophers who are interested in the status of psychiatry have stopped asking whether psychiatry is a science and have become more concerned with the scientific credentials of specific aspects of psychiatric practice, such as classification (Cooper 2009) and explanation (Murphy 2006).

As we see it, one of the central problems with classification in psychiatry is that it plays two important functions. The taxonomies proposed by the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM) and by the World Health Organization’s International Classification of Diseases (ICD) guide research into mental disorders, but also inform diagnostic categories, thus determining eligibility for treatment via national health or health insurance systems.

Potential conflicts between the aims of research and clinical practice have become obvious in the debate about the new edition of the DSM, published in May 2013. For the purposes of identifying mental disorders in a clinical setting, criteria largely based on behavioral manifestations may be appropriate, especially when no other diagnostic tools are available due to the often complex and still largely unknown etiology of many psychiatric disorders. But for the purposes of research, classification based on symptoms alone is often deemed unsatisfactory. Two weeks before the DSM-5 was published, the director of the National Institute of Mental Health, Tom Insel, claimed that the institute would no longer fund research based exclusively on DSM criteria due to problems of validity for DSM categories. Insel wrote: “Unlike our definitions of ischemic heart disease, lymphoma, or AIDS, the DSM diagnoses are based on a consensus about clusters of clinical symptoms, not any objective laboratory measure” (Insel 2013). Subsequently, Insel amended his evaluation of the DSM considerably (Insel and Lieberman 2013), but it is fair to say that his criticism about the validity of its diagnostic categories remains, and that the current direction of the NIMH is toward casting mental disorders as primarily biological phenomena.

Further questions are raised by the social and political pressures that shape classificatory and diagnostic manuals. In the 1970s and 1980s, debates surrounding homosexuality led to significant changes in the DSM. Homosexuality, which had been listed as a “sociopathic personality disturbance,” was removed from the DSM in 1973, and replaced by “ego-dystonic homosexuality” (referring to distress caused by sexual orientation), which in 1986 also disappeared. More recent debates have focused on attention deficit/ hyperactivity disorder (Koerth-Baker 2013) and on depression (Rottenberg 2013), with commentators claiming that the changes in DSM-5 will lead to an unnecessary pathologization of normal behavior.

Controversies in Classification and Diagnosis

In addition to the above complexities, there are further reasons why diagnostic criteria are more controversial in psychiatry than in other medical specialties. First, it is more common for psychiatric patients than for non-psychiatric patients to reject their specific diagnoses even when insight is present (Szasz 1974). Second, empirical evidence suggests that psychiatrists can be easily deceived to diagnose people who report false symptoms, raising concerns about the objective validation of psychiatric diagnosis, and heavy reliance on self-reports (Rosenhan 1973). Third, some disorders appear to be culturally bounded in that they are diagnosed more frequently in certain periods of time and in certain geographical areas. One interesting case is that of dissociative amnesia (Pope et al. 2007); another is that of “apathetic children” in Sweden (Godman 2013).

In addition to these challenges, symptoms do not map onto disorders in a straightforward way. Different disorders may share the same symptoms, and symptoms may be continuous with, as opposed to radically divergent from, normal patterns of behavior. Furthermore, it has been argued that a concept such as schizophrenia is “scientifically meaningless” because sharing the diagnosis does not mean having the same brain disease (Bentall 2004). Richard Bentall holds that, as such, general statements about how a person with schizophrenia is likely to behave are not going to be a reliable guide to either research or diagnosis. A similar argument might be made in relation to autism spectrum disorder (ASD), which has been introduced as a unified category in DSM-5. As for schizophrenia, genetic, environmental, psychological, and neurological causal factors have been found to contribute to ASD. The diverse behaviors that are diagnostic seem unlikely to be explained by a single “brain disease.”

All of this raises the question: How should classification and diagnosis work for mental disorders such as schizophrenia and autism? According to a strong interpretation of the medical model, we can be alerted to the presence of a pathological condition by the observation of a cluster of symptoms, but ultimately we should identify the biological markers reliably associated with that condition; and any such association should be informed by a good-enough story about how the biology causes the symptoms (Taylor 1999; Andreasen 2001). But while this model is often regarded as a regulative ideal, something to aspire to in psychiatry, given what we currently know about psychiatric conditions it seems unrealistic. As we read in the introduction to the DSM-5, for many psychiatric categories, there is insufficient information about the biological or physiological correlates. This may be because the causal mechanisms of the disorder are to some extent unknown, or because the label we use does not capture a single biologically defined disorder. Other possibilities for reinterpreting such categories will be explored below in “Rationality and Patient Autonomy.”

Irrationality as a Diagnostic Criterion

There may be a more fundamental reason why the medical model, in its strong interpretation, does not seem to suit psychiatry: namely, that we can detect pathologies of behavior only if we look closely at behavior, and find that it departs from norms of epistemic and pragmatic rationality, and accepted norms of moral conduct (see Stier 2013). The main obstacle to the reduction of psychiatry to neuroscience appears to come from the role that considerations about rationality currently play in the classification and diagnosis of mental disorders (see Pickard 2009; Broome and Bortolotti 2009 for some examples). The diagnostic criteria in the ICD and DSM rely almost exclusively on behavioral manifestations, and many of them are described as deviations from rationality. As Dominic Murphy says, the DSM offers a picture of mental disorders as “a collections of signs and symptoms” which “depend on physical processes but are not defined or classified in terms of those physical processes” (Murphy 2009, p. 107). For a variety of common disorders such as depression, schizophrenia, autism, and dementia, the symptoms are characterized in terms of deviations from rationality, especially epistemic requirements, or social norms (see Box 1 for some examples).

Is Epistemic Rationality the Mark of Mental Illness?

The fact that a variety of common disorders are chiefly characterized as breakdowns of epistemic rationality may suggest that there is a very close relationship between such deviations from rationality and mental disorder, but it would be hasty to conclude that epistemic irrationality is sufficient, or even necessary, for mental disorder. First, as cognitive and social psychologists have insisted since the 1970s, epistemic irrationality is a feature of normal cognition and is not confined to those who have a psychiatric diagnosis. People systematically fail to test simple conditional statements; are unable to determine the relative probability of statements; and routinely exhibit inconsistent beliefs and preferences (Stein 1996). These reasoning mistakes and inconsistencies are observed not only in the lab; they are also widespread in everyday decision-making, and do not spare those with professional expertise in the making of decisions involving deductive and probabilistic inferences, such as statisticians and medical doctors (see, for instance, Garb 1998).

We do not need to read the psychological literature to appreciate the extent to which human cognition is plagued by irrationality. According to norms of epistemic rationality, a belief is rational if it is well supported by, and responsive to, the available evidence. There are many instances of ordinary, non-pathological, beliefs that do not meet these standards. The racist belief that members of a certain ethnic group are violent or lazy is not obviously less irrational than the delusional belief that one’s neighbor is a spy paid by the government to follow one’s movements (persecutory delusion), or that one’s romantic partner is unfaithful (delusion of jealousy). Racist or superstitious beliefs share many of the epistemic features of delusions: they are typically false; they may conflict with other beliefs the person has; they are badly supported by the available evidence; and they are often incredibly resistant to counter-evidence (Bortolotti 2009, chapter 3).

One reason why racist and superstitious beliefs come across as less puzzling than delusions is that they are widespread in the population, whilst delusions are rare. Another reason is that the content of everyday irrational beliefs tends to be more mundane than the content of delusions. One can understand how racist beliefs are formed, but it is more difficult to make sense of the formation of many delusions if no specific reference is made to the anomalous experience people may have had prior to adopting delusional hypotheses. Delusions such as Capgras, where the person reports that a loved one has been replaced by an impostor, or Cotard, where the person claims to be dead or disembodied, seem so implausible as to involve a different sort of irrationality.

But even Capgras and Cotard delusions make some sense to those who are clinically trained. If we learn that a woman who comes to believe that her husband has been replaced by an almost identical impostor does not feel the same affective response she used to feel when she sees her husband, then her delusion becomes more understandable —her recognition of her husband is compromised. The delusional belief is still implausible, but it is not unmotivated. Moreover, whether beliefs are widespread and whether their content is puzzling to our ears seem to be contingent facts about the assumptions shared in the society in which such beliefs are reported, and do not seem necessarily to point to any significant epistemic difference.

Some authors remain unconvinced that there is continuity between familiar and more exotic forms of irrationality, and argue that pathological irrationality is more persistent and severe (e.g., Edwards 1981), or is of a different quality because it signals a more dramatic failure to be in touch with reality (e.g., Gallagher 2009). But the epistemic irrationality exhibited in the context of many psychiatric disorders is often not persistent, and, as we shall see, is not necessarily more severe than the irrationality found in the non-clinical population (Bortolotti 2013). Moreover, the assumption that familiar irrationality does not involve losing touch with reality should be more thoroughly challenged in the literature. What is the disanalogy between the irrational belief of the person affected by prejudice and the irrational belief of the person affected by delusions of persecution, in terms of reality checks? Insofar as the beliefs are irresponsive to evidence, they signal a departure from reality. The person with sexist beliefs who refuses to acknowledge evidence for the view that, say, women make good scientists, shuts herself off from the game of evidential support that all epistemically rational believers should play. These mundane failures to engage with reality may not capture our imagination as bizarre delusional hypotheses do, but, from an epistemic point of view, they are on a continuum with them.

Epistemic irrationality is not even necessary for mental illness. There is no reason to suppose that mental illness needs to manifest as a failure of epistemic rationality and, indeed, the diagnostic criteria for anxiety disorder and personality disorders do not include behaviours characterized by epistemic irrationality. Moreover, in some specific tasks, people with psychiatric diagnoses that are characterized in terms of deviations from rationality seem to be more epistemically rational than non-clinical samples. Studies investigating what has been called depressive realism suggest that people make more accurate predictions when they are depressed (for a review, see Abramson et al. 2002). This is because the statistically normal way in which predictions are made is characterized by excessive optimism. The phenomenon of depressive realism is controversial (for recent evidence against the phenomenon, see Baker et al. 2011), but similar phenomena have been observed for other diagnoses. There is evidence in autism and schizophrenia that the reasoning tendencies partially responsible for the formation of pathological beliefs, or underlying pathological behavior, can also have epistemic benefits relative to those found in the non-clinical population (Tateno 2013; De Martino et al. 2008; Owen et al. 2007). For instance, people with schizophrenia have been found to be less vulnerable to framing effects, and do not exhibit a statistically normal but procedurally irrational increased tendency to gamble when faced with a certain loss (Brown et al. 2013). Therefore, although epistemic irrationality is undoubtedly a characteristic of the behavior of people diagnosed with some mental disorders, we can have mental disorder without epistemic irrationality, and epistemic irrationality without mental disorder.

Is Pragmatic Irrationality the Mark of Mental Illness?

The above discussion questions the strength of the association between epistemic irrationality and psychiatric diagnosis. We now consider the role of pragmatic irrationality in psychiatric classification and diagnosis. In this chapter, we use the term pragmatic rationality to refer to decision-making that promotes the agent’s wellbeing or success in pursuing their goals. An alternative notion of rationality is captured by the notion of ecological rationality, which refers to decision-making strategies that promote genetic fitness, measured in terms of chances of survival and reproduction. The distinction between pragmatic and ecological rationality maps onto the distinction between psychological adaptiveness (where the goal is to increase well-being) and biological adaptiveness (where the goal is to enhance genetic fitness), which we owe to Ryan McKay and Daniel Dennett (2009). They point out that some behavior can be psychologically adaptive without being biologically so, because the relationship between well-being and genetic fitness is a complex one. Here we focus on the hypothesis according to which pragmatic irrationality—decision-making that compromises the well-being or success of an agent in pursing their goals—is the marker of mental disorder.²

An argument for the claim that pragmatic irrationality is what demarcates mental disorder could be advanced as follows. Depressive realists may have a more accurate representation of reality, and may make better predictions about future events. And people with schizophrenia may be less risk-taking when faced with a certain loss. But these epistemic advantages apparently do not translate into well-being and success from a pragmatic perspective. People with these mental disorders struggle to flourish. Mental disorder, it might be proposed, undermines the rational capacities that enable agents to navigate their environment, pursuing and achieving at least some of their goals. Indeed, this could be the difference between the person with racist beliefs and the person with delusions: maybe they are both epistemically irrational, but racist beliefs as opposed to delusions are pragmatically neutral or may even have pragmatic benefits.

The difficulty is that it is not clear, in general, that everyday instances of epistemic irrationality increase flourishing or are at least pragmatically neutral; nor that the epistemic irrationality associated with mental disorder generally undermines flourishing. It is true that commonplace over-optimistic trains of thought have been consistently shown to impact positively on mental and physical well-being (Taylor et al. 2003). But the case of people with schizophrenia taking fewer risks as a response to the prospect of certain loss does not seem to support the proposal. This uncommon epistemic rationality means that people with a diagnosis of schizophrenia are less likely to lose in gambling situations by choosing the less risky option, and thus are likely to make better decisions, arguably promoting well-being and success. Conversely, while commonplace racist and superstitious beliefs can offer comfort and increase self-esteem in the short term, they will often be socially sanctioned in the long run, leading to distress and loss of self-esteem. Thus, the hypothesis that pragmatic irrationality is the marker of mental disorder deserves closer examination. Some apparent counter-examples include the case of “successful psychotics” (who experience hallucinations and delusions but find meaning in life thanks to those psychotic experiences, and function well), and the notion of “successful psychopaths” (Board and Fritzon 2005). In both cases, it is apparently judged that a psychiatric diagnosis is appropriate—as indicated by the use of diagnostic language in these labels—despite the person’s apparent flourishing as an agent.

Interim Summary

So far, we have discussed the role of epistemic and pragmatic rationality in psychiatric classification and diagnosis, and considered whether either is the distinctive marker of mental disorder. We observed that the diagnostic criteria of many psychiatric categories make reference to cognitive or affective impairments described in the language of rationality and social functioning. We then asked whether mental disorder is demarcated by reference to the violation of epistemic norms, and concluded that this is implausible, due to: (1) the pervasiveness of epistemic irrationality in the non-clinical population; (2) the fact that epistemic irrationality does not have a different quality when it is present in people with a psychiatric diagnosis (it always involves some sort of departure from reality); and (3) the fact that on some occasions the reasoning styles and strategies of people affected by mental disorder are more conducive to epistemic rationality than the styles and strategies of the non-clinical population—a theme that will be picked up again in the sections that follow. We found it more plausible that mental disorder is demarcated by pragmatic irrationality, where decision-making results in diminished well-being or compromised success in the pursuit of one’s goals. However, we note that there are unresolved issues and potential counter-examples associated with this proposal as well.

Mental Disorder as Grounds for Interference

One approach taken in many jurisdictions holds that the presence of mental disorder authorizes intervention merely on grounds of the risk the person poses to their own health or safety. Justifications of this kind of approach refer to impaired insight in mental disorder (Saddichha 2008); to mental disorder removing autonomous control (Doyal and Sheather 2005); to the association of mental illness with global irrationality (discussed in Hewitt 2010); and to the idea that mental disorder undermines personal identity (Edwards 2010; Matthews 2000) or diachronic agency (Janssens et al. 2004). Whatever the underlying assumptions, law that bases interference on the presence of mental disorder—the so-called “status” approach—holds that a psychiatric diagnosis eliminates the need to assess the person’s decision-making ability. The diagnosis alone is taken to mean that the person is not in a position to decide for themselves, at least in relation to psychiatric treatment, and such legal structures have been widely criticized because of this feature (Department of Health 1999; Szmukler and Dawson 2011; Wildeman 2013).

In addition to this concern, criticisms have focused on the risk-based grounding of intervention that is often a feature of the status approach. This conflicts with a liberal understanding of the appropriate scope of the law, which requires that legal structures allow for considerable flexibility when it comes to the different kinds of lives people choose to live (Plant 2011). The law must allow for controversial treatment decisions to sometimes be respected, on grounds that in some cases the controversial nature of a decision will be explained by the person’s idiosyncratic commitments—their particular desires, values, projects, what makes life meaningful for them. By focusing on the risk to health and safety, such laws do not allow for any divergence of perspective on the question of what ends should be pursued. Health and safety is what matters, whether or not these are the primary concerns for the person in question.

Mental Incapacity as Grounds for Interference

In contrast, a mental capacity approach to the question of when interference is justified is held up as a much more progressive answer. For example, in England and Wales, the Mental Capacity Act 2005 (MCA) is understood to address both of the above concerns. A medical diagnosis is neither necessary nor sufficient for a finding of mental incapacity (though the person’s inability to decide must be due to an impairment in the functioning of mind or brain (s. 2(1); Department of Constitutional Affairs 2007, p. 44)). Rather, an assessment of person’s ability to understand, weigh, and use relevant information, and to express their decision, grounds decisions about the right to self-determination—what is known as a functional test.

In relation to the second concern, to significant degree assessments of mental capacity are designed to be independent of the choice the person makes. Different jurisdictions allow for the seriousness of the consequences to influence the assessment of mental capacity in different ways. When the consequences are serious, the test may be applied more stringently, or the threshold for mental capacity may be raised (Buchanan 2004).³ However, this flexibility is only supposed to go so far. It is generally understood that an adult’s treatment decisions must be respected if they have the mental capacity to make the decision, however grave the potential consequences, and there is some evidence that in England and Wales this principle is being put into practice. In 2007, Kerrie Wooltorton, a young woman with a history of mental health problems and suicide attempts, was admitted to hospital having deliberately drunk a lethal quantity of antifreeze. She refused life-saving treatment, but assented to medicine that would help make her comfortable. Her wishes were respected on grounds that she had the mental capacity to refuse life-saving treatment. She subsequently died in hospital, and the decision to respect her wishes was supported by the coroner who investigated her death (David et al. 2010; Richardson 2013).

Because of these features, a mental capacity approach has been held up as a more ethically secure basis for overriding patient autonomy. Legal intervention aims to be independent of medical diagnosis, and largely independent of the person’s choice, so allowing for divergence on the life-shaping commitments people choose to pursue. In essence, mental capacity law is understood to make moral progress because its focus is on the decision-making process. Functional tests appear to operationalize procedural norms of rationality, which, according to standard views, enable people to pursue whatever it is they want to pursue (Craigie and Coram 2013; Williams 1981). Appeal to these norms is supposed to bring impartiality to mental incapacity as a basis for interference, though the extent of the role they play in practice has been a matter of debate (see Owen et al. 2009; Holroyd 2012; Banner 2013; Freyenhagen and O’Shea 2013; Mackenzie and Rogers 2013).

However, recent developments in international human rights law and the disability rights literature have called into question the legitimacy of mental incapacity as a basis for overriding patient autonomy. We frame this challenge as questioning the impartiality of the mental capacity approach, and draw on the puzzles explored in the sections above to explore this idea. We begin by examining Ronald Dworkin’s much-discussed position on what respect for autonomy requires in the context of dementia, which provides a background against which to consider this challenge.

Respect for Autonomy and Dementia

In his book Life’s Dominion: An Argument about Abortion, Euthanasia, and Individual Freedom, Ronald Dworkin discusses the case of Margo, a woman who at 54 has advanced Alzheimer’s disease (1994, pp. 220–237). Margo is described as a woman who reads mysteries but her place in the book jumps randomly from day to day. She attends art class where she paints pretty much the same picture every time, and enjoys listening to music, happily listening to the same song over and over as if for the first time, though she does smile at a particular song which she says reminds her of her deceased husband. Despite her illness, one of Margo’s carers describes her as one of the happiest people he has known. In his words, the Alzheimer’s was leaving her “carefree” and “always cheerful” (Firlik 1991).

Dworkin asks what respect for autonomy requires if, before suffering any serious mental decline Margo had expressed a firm wish, in writing, that life-saving treatment should be withheld once she was in the advanced stages of the illness, allowing her to die. Should treatment be withheld, for example, if Margo contracts a respiratory infection that is life-threatening but could easily be treated with antibiotics? Based on an analysis of what makes it the case that we ought, generally, to respect people’s personal choices, Dworkin argues that life-saving treatment in such a case should be withheld. According to Dworkin, it is Margo’s prior wishes that should be given legal effect, because the importance of respecting autonomy “derives from the capacity it protects; the capacity to express one’s own character—values, commitments, convictions, and critical as well as experiential interests—in the life one leads. Recognizing an individual right of autonomy makes self-creation possible” (Dworkin 1994, p. 224).

The right to autonomy rests on the capacity to lead a life in this sense, and in Dworkin’s view Margo no longer has this capacity. What is essential is a person’s ability to consider their critical interests: their own ideas of what for them makes life successful, not merely enjoyable or not in the moment. Importantly, in adopting this position, Dworkin rejects the idea that respect for autonomy is required because people know best what is good for them, and therefore that welfare considerations underpin the principle. It is clear, according to Dworkin, that autonomy requires that personal choices are respected even when they are against the individual’s avowed interests (as in weakness of will and acts of self-sacrifice for another).

Among others, Agnieszka Jaworska challenged Dworkin’s conclusion about respect for autonomy in the context of dementia on the basis that he focuses on “peripheral rather than essential” elements of the capacity for autonomy (1999, p. 129; for other criticisms see Dresser 1995; Wolff 2012). Jaworska focuses instead on the possession of decision-guiding values as the foundation of the capacity that the principle of respect for autonomy aims to protect. She argues that in all but the most severe cases of dementia, the person remains a “valuer”—an agent who endorses and eschews desires, and therefore has opinions about what is good for them. In her view, what are lost in dementia are primarily the capacities for means-ends reasoning and planning that enable the person to make decisions in accordance with these values; and this loss does not justify withdrawing respect for autonomy: “An Alzheimer’s patient may be too disoriented to form a life plan or to choose specific treatment preferences, but so long as he still holds values, he is, in the most basic sense, capable of self-governance, and this fact about him commands utmost respect” (Jaworska 1999, p. 134).

The disagreement between Dworkin and Jaworska can be understood as a difference on what capacities are necessary for a person to have agency that requires legal recognition, and should be given legal effect (at least in relation to this kind of decision—Wolff points out that the answer is likely to depend on the nature of the decision: Wolff 2012). According to Jaworska, only the capacity to value is essential. For Dworkin, among other things, the agent must have the capacity to make decisions from a life overall perspective. This requirement is apparently procedural rather than substantive in nature (it does not specify what ends the agent should pursue). Nonetheless, Dworkin’s position appears to be value-laden. According to Dworkin, it is the importance of respecting a person’s capacity to live their life in their own particular way that justifies non-interference, and his conclusion therefore entails that Margo is unable to live life in her own particular way. But in a straightforward sense Margo clearly is living life in her own way. It’s just not a way that is directed from a life-overall—a life with a plan—perspective.

The importance of the procedural requirements that Dworkin holds are essential for attracting respect for autonomy is derived from his commitment to the value of human life being structured by a plan.⁴ His position on Margo’s case therefore illustrates one way in which mental incapacity as a justification for not respecting autonomy, seems, ultimately, to have value-laden roots.

Rationality and Legal Personhood

The disagreement between Dworkin and Jaworska has a resonance for more recent developments in disability rights law and literature, which have challenged the use of mental incapacity as a basis for not respecting patient autonomy. The UN Convention on the Rights of Persons with Disabilities, around which these developments are focused, is said to constitute a paradigm shift in mental health law (Richardson 2012; Bartlett 2012; Wildeman 2013). According to Article 12 of the Convention, persons with mental disabilities, including mental disorders, must be recognized before the law on an equal basis to others, and must be supported in the exercise of their legal capacity. A strong interpretation holds that recognition as a legal person, which includes the right to respect for autonomy in one’s personal affairs, should not be limited to those who meet the requirements of functional tests (Bach and Kerzner 2010; United Nations Committee on the Rights of Persons with Disabilities 2014).

In the Convention, this position is set against a background commitment to the equality, dignity, autonomy, and needs of persons with disabilities, as well as a social model of disability (preamble, section e). While medical models explain mental disability solely with reference to the individual’s mental impairment, social models point to the environmental factors that also play a role in determining whether an impairment results in a disability. The social model supports a shift away from mental incapacity as a basis for interference, to the extent that it suggests the concept of mental capacity, which is focused on the individual, should be replaced with a more socially contingent notion of decision-making ability, justifying the requirement of decision-making supports. However, the de-linking of legal capacity from mental capacity requires further justification, and these arguments have been developed in the surrounding literature.

One central moral argument given in support of this position holds that mental capacity tests set requirements that, if strictly applied, most people would fail to meet—a view that garners support from the sections above. Most prominently, Michael Bach and Gerard Quinn have questioned the degree to which people generally understand the complex health care interventions to which they consent (among other kinds of decision; Bach 2009; Quinn 2010). Bach and Quinn can be understood to rightly highlight that having mental capacity does not require full procedural rationality. It seems true the most of us will not fully understand the probabilistic results of diagnostic tests before making a related decision; and most of us place less value than we should on consequences in the far future relative to those in the near future, according to theories of rationality. Rather, functional tests are about having good enough—which in practice means statistically normal—rational capacities. The threshold for mental capacity is set so that most people have legal capacity, and therefore it is political considerations, as much as rationality considerations, that determine the requirements of mental capacity tests (Buchanan 2004).

Framed in this way, the argument against mental capacity as a basis for interference is analogous to liberal criticisms of dedicated mental health law. The proposal is that Bach and Quinn are essentially arguing that mental capacity law is unfair because the line determining whose decisions are respected is drawn on the basis of how statistically normal an irrationality is. From a liberal perspective it is wrong to draw this line on the basis that someone is not living their life in a normal way, which raises the question: why is statistical normality justified as a way of drawing the line when it concerns how people deliberate in their personal decisions?

Neurodiversity and Mental Capacity

The emerging concept of neurodiversity, which is being used to reinterpret certain psychiatric categories, can be used further to develop this idea. The central example is autism, but wherever it is applied, a neurodiversity perspective seeks to redefine a diagnostic category primarily as a set of cognitive differences. The features that are used to identify the category are understood to be atypical, but not intrinsically dysfunctional (Jaarsma and Welin 2012).

In relation to autism the call for this reassessment has been motivated, among other factors, by scientific findings that suggest autism is more accurately described as a cognitive profile or cognitive style, which confers both advantages and disadvantages, rather than straightforwardly as a disorder. The idea can be explained using the weak cognitive coherence theory of autism (Happe 1999), though it does not depend on this particular theory turning out to be true. According to the weak cognitive coherence theory, while for most people information is processed using its broader context—often at the expense of local-level information, as we will see below—for autistic people cognitive processing is focused on the local level, at the expense of more global and contextual features. Global processing means that for most people, the words of a sentence with meaning are easier to remember than a random list of words of the same length, and the perception of an object is strongly influenced by its context. This processing style confers many advantages, but also disadvantages, as seen in the difficulty that most people have remembering random lists of words; in visual illusions induced by contextual features; and in framing effects that are examples of epistemic irrationality that is widespread in the general population. On a range of these kinds of tasks there is increasing evidence that autistic people tend to outperform non-clinical controls (e.g., De Martino et al. 2008; for overview Happe 1999, Mottron 2011).

In relation to the arguments against mental incapacity as a basis for interference, this understanding of autism reinforces the concern that statistical normality more than rationality may be driving assessments of mental capacity. Tyler Cowen observes that in psychological studies where a difference in performance is observed between an autistic group and a non-clinical control group, it is often assumed that the difference is explained by a cognitive impairment associated with autism, even where the autistic group clearly outperforms controls (Cowen 2011; De Martino et al. 2008). The concern is that a parallel interpretation occurs in the context of mental capacity assessments (Mackenzie and Watts 2011). While both globally focused processing and locally focused processing confer mental incapacities, it seems plausible—likely, even—that only the incapacities associated with locally focused (autistic) processing will be identified as such, and as potentially relevant grounds for mental incapacity in the legal sense. The incapacities that are a product of globally focused processing will be accepted because they are statistically normal, however much epistemic irrationality they involve. There may even be a danger that locally focused cognitive advantages that are characteristic of autism, conferring more epistemically rational decision-making, will be interpreted as an incapacity because of their abnormal rationality relative to the general population.