Monday, 12 December 2011

Evaluativism versus objectivism

A draft paper hopefully to be translated into Italian for a special issue of Rivista Sperimentale di Freniatria on neuroscience.

Evaluativism versus objectivism: Is the question of facts versus values in the analysis of mental illness a factual or a value-laden question?


The degree of insight into the nature of mental illness that neuroscience can offer depends on the nature of mental illness itself: on whether it is an objective, internal matter or evaluative and socially constituted. But there has been long-standing disagreement about this. To shed light on the debate, I draw on, and refine, Zachar and Kendler’s proposed framework for debating psychiatric taxonomy. I argue that a radical externalist account of mental illness (constitutive evaluativist externalism) threatens the role of neuroscience. But, further, the disagreement between such an evaluativist position and its objectivist opposition may concern not only the nature of mental illness but also the terms of the debate. The evaluativist can argue that the disagreement itself is not a factual, but rather an evaluative, matter.


How much can we hope to learn about the nature of mental illness through neuroscientific inquiry? Here is an analogy. If mental illness is like physical illness then a scientific perspective will be as helpful in the former case as it is in the latter. But there is a long-standing debate about how similar mental and physical illness is. In his paper ‘The myth of mental illness’ Thomas Szasz, for example, claims that there is an important distinction between them.

The concept of illness, whether bodily or mental, implies deviation from some clearly defined norm. In the case of physical illness, the norm is the structural and functional integrity of the human body. Thus, although the desirability of physical health, as such, is an ethical value, what health is can be stated in anatomical and physiological terms. What is the norm, deviation from which is regarded as mental illness? This question cannot be easily answered. But whatever this norm may be, we can be certain of only one thing: namely, that it must be stated in terms of psychological, ethical, and legal concepts... [Szasz 1972: 15]

On Szasz’ view, physical illness implies deviation from the functional norms of the body, the norms governing its correct functioning. Mental illness, if it existed, would imply deviation not from bodily functional norms but from a different kind. One possibility is that they imply deviation from mental functions modelled as close analogues of bodily functions. Indeed, Szasz suggests that the norms are psychological. But he adds that they are also ethical and legal. His view is that mental illness differs from physical illness by being essentially social and evaluative.

Szasz himself does not rest content with the claim that mental illness answers to different norms from physical illness. Combining this claim with the claim that mental illness is supposed to be medically treatable and that medical interventions address bodily functional norms, he argues that there is no such thing as mental illness. Nothing could possibly satisfy these constraints. This argument, at least, can be blocked, however. Even if mental illness is defined by, or identified through, psycho-social norms, this need not imply that it is identical to, or constituted by, such deviation. It may be that the illness is the cause of the deviation such that, even though it is picked out by its characteristic effects, it is not identical to them. If so, that idea is compatible with medical intervention and so Szasz’ argument fails.

Nevertheless, if Szasz’ initial claim were true, it would still have important consequences for a neuroscientific investigation of mental illness even without his more radical sceptical conclusions. The most that a neuroscientific inquiry could find would be the causes of mental states and behaviour which, as a matter of fact, deviated from social and evaluative norms. It would not be able to investigate the nature of mental pathology as such: what makes something an illness. If mental illness is essentially socially and evaluatively constituted, then the investigation of what makes something pathological is a matter for sociological and perhaps moral inquiry, not neuroscientific inquiry. Thus Szasz’ initial claim is still of significance for understanding the connection between scientific inquiry and mental illness.

To give an example of this potential limitation, consider the experience of hearing voices. This counts as a first rank symptom of schizophrenia: a clear indication of pathology. But it is also asserted by some of those who experience it as merely an aspect of a different kind of subjectivity, in no sense intrinsically pathological. Is it pathological or not? Can neuroscience shed light on this question? Not, I suggest, if Szasz’ initial claim is true.

But is that initial claim true? One might have expected that in the 40 years since Szasz published ‘The myth of mental illness’, an agreement would have been reached as to what kind of norms are involved in a diagnosis of mental illness. But it has not been. In this paper I will explore one basic reason for that. It is not clear whether mental illness is partly constituted by values or not. But it is not even clear whether the debate about that is itself an evaluative or a factual matter. That is: is the aim of inquiry an analysis of mental illness that is true? Or is it an analysis that is good: one which fixes a desirable view of what we should mean by ‘mental illness’?

I will approach this question in the light of two distinctions from Peter Zachar and Kenneth Kendler’s recent conceptual framework for psychiatric taxonomy. In their paper ‘Psychiatric Disorders: A Conceptual Taxonomy’, Zachar and Kendler attempt to outline the many potential initial questions which have to be answered before one can draw up a taxonomy of mental illness. These questions all concern general issue of what sort of things mental illnesses are. The aim of their paper is to set an agenda for thinking about a scientific taxonomy rather than to settle it. The two distinctions with which I will be concerned are objectivism versus evaluativism and internalism versus externalism. I will take these in turn.


The first distinction is defined like this:

Is deciding whether or not something is a psychiatric disorder a simple factual matter (“something is broken and needs to be fixed”) (objectivism), or does it inevitably involve a value-laden judgement (evaluativism)? [Zachar and Kendler 2007: 558]

The example picked for objectivism may seem surprising. It may not seem to be a simple factual matter, a matter to be contrasted with an evaluation, whether something is broken and needs to be fixed. Contrast this idea with a paradigmatic objective taxonomy such as the Periodic Table in chemistry. The Periodic Table classifies on the basis of atomic number (the number of protons in the atomic nucleus). To model the example on that would require thinking of ‘needing to be fixed’ as an objective property of the layout of the world which is there anyway, like atomic number, irrespective of the values of a judging subject. It would be a property the detection of which would be enough, without complementary desires, to motivate a subject to bring about its repair. Against a stark contrast of facts and values, such an objective and yet at the same time essentially motivating property seems, using John Mackie’s term, rather queer [Mackie 1977: 38-42].

In fact, even the first element of their example is not such a simple descriptive idea. Being broken is not a simple physical property. Nor need it even supervene on (simple) physical properties since, for example, a device which is broken with respect to one function might successfully possess a different function.

These considerations would motivate an inversion of the role of the example in the definition to give, instead, this:

Is deciding whether or not something is a psychiatric disorder a simple factual matter (objectivism), or does it inevitably involve a value-laden judgement (evaluativism) (“something is broken and needs to be fixed”)?

Two things, however, make the choice of example less surprising. Firstly, outside the explicit contrast with an evaluation there is something obviously right in saying that whether something is broken and needs to be fixed is a factual matter which can be of a simple and everyday kind. Unprejudiced by neo-Humean philosophy, one would naturally say that this is the kind of thing that can be the content of a descriptive judgement. A small child viewing a freshly dropped cup might take in both that it is broken and the corresponding urgent need at a glance.

Secondly, whilst it may not have the conceptual simplicity of atomic number it more closely reflects the kind of taxonomic kinds found in psychiatry. Objectivists – as contrasted with evaluativists – will have be able to analyse such claims – broken and needs to be fixed – in value-free and objective terms. The task is fundamentally harder for objectivists than for evaluativists as the former are committed to a purely factual analysis whereas the latter allow both facts and values; they are not committed to a values-only analysis of disorder. In picking this example, Zachar and Kendler are helpfully reminding us of the challenge for objectivists.

Having now clarified this distinction, I will turn to the second one of relevance to this paper before drawing out their combined significance.

Constitutive externalism

The second distinction is summarised thus:

Should psychiatric disorders be defined solely by processes that occur inside the body (internalism), or can events outside the skin also play an important (or exclusive) defining role (externalism)? [ibid: 558]

Zachar and Kendler further characterise the distinction with the following hints. Modern psychiatry has been largely internalist and holds that events within the body are ‘critical for understanding and defining’ mental disorders [ibid: 558]. Externalists are either moderate and hold that ‘what goes on inside the head cannot be isolated from an organism’s interaction with the world’ or radical, in taking external events to be definitional, as exemplified in syndromes which are considered to be ‘reactions to harsh societal demands’ [ibid: 559].

It is helpful to draw attention to a further distinction which Zachar and Kendler do not make but which can shed light on their distinction. One can think of externalism as characterising a claim about causation or constitution. If one, plausibly, thinks that environmental factors sometimes cause mental illness then one is a causal externalist. But one may think that they cause mental illness by affecting states – perhaps neurological – within the body. If so, whilst a causal externalist, one is also a constitutive internalist. (Constitution is not quite the same thing as what defines a mental illness. Even a constitutional internalist may find it helpful to label illnesses by their causes.)

This clarification can be applied to an example of externalism that they give, the Interpersonal Model:

Contrary to any of the medical models, an interpersonal systems model is staunchly externalistic. Most fundamentally, this model views disturbed behaviour as arising from disturbed relationships. Rather than deriving from psychopathology in individuals, psychiatric disorders are seen to develop dynamically from pathology in interpersonal contexts. The notion of patients being containers of internal psychological states is minimised, whereas the view of them as persons trying to adapt to their social worlds is maximised. The context or the interpersonal system is both locus of pathology and the cause of pathological behaviour. [ibid: 562]

Most of the characterisation in this passage would fit a causal externalist but at the same time a constitutive internalist view of disorder. The fact that disturbed behaviour arises from disturbed relationships is consistent with the causation being mediated by states of the brain. Similarly, dynamic changes in response to interpersonal contexts may be dynamic changes of the brain. And there is no reason to rule out a central role for brain-mediated responses for persons adapting to social worlds. The ‘context as cause’, in the final sentence of the passage above, again exemplifies merely causal externalism.

A causal externalist but constitutive internalist view of disorder poses little threat to the insight that neuroscience can offer. If what makes something a mental illness is constituted by matters within the skull then neuroscience looks the best tool to investigate it. To generate a threat to that one needs to more radical externalism. One needs, for example, to think of the Interpersonal Model in constitutive externalist terms (and thus play up two so far neglected hints of that in the quotation). On such an account, disturbed behaviour is constituted in or by disturbed relationships. Interpersonal contexts are themselves literally pathological. (Thus, for example, family relationships do not cause pathology in a disturbed child; the relationships, rather than the child, are pathological.) The context or the interpersonal system is the locus of pathology (and thus not the cause of pathological behaviour since the interpersonal system includes the behaviour). Constitutive externalism in the philosophy of mental health is a radical view (whilst causal externalism is not). It also threatens the role of neuroscientific inquiry by locating that factors that constitute mental illness outside the body. Its role would be limited to examining those causal factors which, as a matter of fact, led to these effects not the direct investigation of illness itself.

Having now clarified the nature of both evaluativism and externalism, and suggested how constitutive forms of externalism limit the kind of insight that neuroscience can offer, I will now consider, in the next two sections, a further importance distinction: between disciplined and undisciplined evaluativism. It is this that makes the debate about mental illness particularly complex.

Constitutive evaluativist externalism

Evaluativism is a particular kind of constitutive externalism. According to it, the reason why deciding whether something is a psychiatric disorder involves a value judgement is that psychiatric disorder is constituted in part by values. (Only ‘in part’ because the values either inhere in or apply to – a distinction to which I will return – other, perhaps physical, properties.)

So, for example, according to the Szaszian view mentioned at the start of this paper, the problems that are misleadingly labelled mental illnesses are deviations from psycho-social and ethical norms: they are constituted by that deviation [Szasz 1972]. According to the ‘lost tribe’ view influenced by Laing and Foucault, madness is just another way of going on [Foucault 1989; Laing 1960]. To be mad is just to be evaluatively out of step with the rest of the community. On Bill Fulford’s more moderate picture, mental illness has to be bad for its sufferer and more specifically is bad for his or her ‘ordinary doing’ [Fulford 1989]. For Jerome Wakefield, though illness involves a supposedly factual biological dysfunction, it has also to be harmful where harm is construed as essential value-involving [Wakefield 1999]. On all of these views, the status of a condition as a mental illness is determined in part by the values in play.

Consider again the claim of some people that the experiences they have such as hearing internal voices, whilst fitting a psychiatric diagnostic category, are not pathological. On a non-evaluativist or objectivist view, this is a simple factual claim. It is true or false and, further, its truth or falsity is independent of the value judgements of the subjects of the experiences (or anyone else). But on an evaluative view, how people value experiences is a constitutive element of whether they are pathological. This raises the question of how to respond to differences of opinion about such values and the consequence of such divergence for psychiatric taxonomy.

Zachar and Kendler offer the following brief discussion of one sort of difference of value judgement.

How do we respond to historical claims that slaves who had a compulsion to run away and advocates for change in the former Soviet Union were mentally ill? An objectivist would claim that those classifications contained bad values and progress was made when those values were eliminated. Their opponents would claim that the elimination of bad values is not the same as becoming value-free, and progress has been made by adopting better values. [ibid: 558]

For an objectivist, however, the fact that a classification reflected any values (aside from the epistemic values that shaped its constructions) would be an error. Values, whether good or bad, feature merely as distortions in a classificatory scheme which should reflect the underlying facts. This mirrors the way that, in Lakatosian rational reconstructions of the history of science, social factors enter only to explain deviations from rational sensitivity to the facts [Lakatos 1970]. When all goes well, there is no need for sociological explanation. So, equally, an appeal to the presence of distorting values in the pathological construction of drapetomania is significant, for an objectivist, in pointing out the presence of values at all rather than specifically bad values.

The characterisation of the contrasting evaluativist’s response raises a further question. Talk of eliminating the bad values implicit in drapetomania suggests (though it does not strictly imply) the idea of moral or more broadly evaluative progress. It suggests that value judgements are disciplined by the attempt to reflect real values. This contrasts with a view in which nothing disciplines such judgements. What appear to be value judgements are really merely expressions of subjective preference and answer to nothing external to them. Their being right is no more than their seeming right. (This is not to downplay their seriousness or importance merely to highlight a view of their logic.) The contrast between disciplined and undisciplined evaluativism is significant in understanding the nature of mental illness. But, as I will argue in the final section, it also runs deep in why the debate about mental illness is so difficult. It is not clear whether that debate aims at truth or goodness.

Disciplined and undisciplined constitutive evaluativist externalism

On a disciplined account, psychiatric taxonomy can aim to get right the mixture, or the compound, of simple facts and values that make up the complex realm of psychopathological phenomenology. Such judgements need not merely reflect motivationally inert features of the world, as the objectivist, assumes. Nor need concepts of disorder (akin to the earlier example of what is broken) be analysed into simple factual terms in order to be accommodated in the taxonomy. But aside from these relaxations, a psychiatric taxonomy based on a disciplined evaluative account would resemble an objectivist approach in one important respect. It would aim to underpin literally true judgements. It would aim, in other words, at validity.

But an undisciplined evaluativist approach is more radical. Mental illnesses are constituted, at least in part, by matters external to the body. In addition, these matters are not features of the world, broadly construed, but rather expressions of subjectivity. If this were the correct approach to the nature of mental illness, however, it fits uneasily with the very idea of a psychiatric taxonomy. Whilst one the aims of taxonomy is validity – to cut nature at the joints – so as to enable the framing of true judgements, on an undisciplined evaluativist approach, that idea of correctness is missing.

Returning to the example of subjects who argue against the pathologising of what are conventionally taken to be pathological symptoms, this distinction is important. For disciplined evaluativists, like objectivists, their claim is a judgement that might be right or wrong and thus would inform, and be informed by, the development of a valid taxonomy. (Unlike objectivists, it is not a simple, that is value-free, factual matter.) But for an undisciplined evaluativist, this is not the case. The claim is an expression of subjectivity. This is not to downplay its importance and seriousness. But it is to suggest that its assessment is more a matter for liberal politics than empirical and more broadly academic inquiry. It is more a matter for decision (of how to act) than judgement (as to what is the case) and the recognition that psychiatric taxonomy is fundamentally the wrong tool for the job.

So far I have merely flagged two subsidiary, but still important, distinctions within Zachar and Kendler’s framework without offering a judgement as to how they might actually apply to psychiatric taxonomy. I have argued that if mental illness is best thought of according to undisciplined constitutive evaluativist externalism then it will not fit well within taxonomic thinking at all. I will end with two final thoughts which will, hopefully, shed light on such a judgement.

Firstly, might there not still be a role for taxonomy and neuroscientific inquiry even given the antecedent of that conditional? There are two immediate possibilities. An undisciplined evaluativist is committed to a fundamental ontological difference between facts and values. One might thus attempt to factor out the values from the underlying facts and develop a scientific taxonomy of merely factual elements. On this account – and by contrast with an objectivist view – what would be left would not amount to a taxonomy of illnesses but rather the factual conditions that motivate competing expressions of illness status. There are two reasons to be sceptical of such a possibility. Philosophically, the prospects for a successful analysis of value judgements into simple facts and the evaluative reactions that they prompt looks poor [see Thornton 2007: 66-67]. Practically speaking, past attempts to purge psychiatric taxonomy of evaluative elements have been unsuccessful.

The other taxonomic possibility would be to attempt to encode expressions of subjectivity without any commitment to their underlying validity: a subjective ‘hit parade’ of mental illness. The problem at root with this thought is that, in the face of disagreements about how to think about diverse experiences and with no metaphysical account of why there might ever be convergence of opinion, there seems to be no rational way to agree any single taxonomy. Pluralism would seem a politically more satisfactory response than framing a taxonomy.

The point above concerning the philosophical implausibility of factoring facts and values is a point that counts against undisciplined evaluativism. Suppose however, as a significant strain of neo-Humean moral philosophers hope, that an analysis into facts and values were possible, would that establish the truth of undisciplined constitutive evaluativist externalism about mental illness? Here a distinction between philosophical debate about moral and psychiatric values is relevant. Whilst there is disagreement about particular ethical judgements in difficult cases, there is to be sufficient agreement about the broad outline of the practices of making moral judgements to make descriptive accuracy a rational aim of meta-ethical moral philosophical debate. It seems plausible to say that Kantian deontology, utilitarianism or neo-Aristotelian moral particularism may simply be the correct description of the moral realm. But that may not be true of the debate about mental illness.

Imagine, for example, that objectivists succeeded in developing a consistent and intuitively plausible account of mental illness, reducing concepts of mental disorder to simple facts. Suppose that on this account, hearing voices turned out to be pathological. Suppose also that undisciplined evaluativists succeeded in developing a rival account on which hearing voices was not in itself pathological. How should the two accounts be assessed? One problem, of course, is that whilst the status of hearing voices is evidence one way or the other, it is contested. If one somehow knew, antecendently, its pathological status that would be a crucial test for the two accounts. But, as Neil Pickering argues, no such pre-theoretical knowledge is possible [Pickering 2006]. In fact, however, the problem goes deeper.

Setting out the debate as I have suggests that whether or not mental illness is simply factual or whether it is irreducibly evaluative – and if so of what sort – is itself a deeper level factual matter. But it is open to an undisciplined evaluativist to argue that that deeper level matter is not factual but rather, also, evaluative. It is a case of ‘values all the way down’. They can argue that we should, for reasons expressive of better subjective value, choose their model of mental illness not because it is true but because it is (evaluatively) right.

This is the consequence of adopting one of the possible positions within debate about mental illness hinted at, though not made explicit, in Zachar and Kendler’s framework. To repeat the two key claims: if one adopts a constitutive and evaluativism, that undermines the role of a neuroscientific understanding of mental illness. Neuroscience cannot explain what it is about a condition that makes it an illness. But second, the debate between such evaluativism and an objectivist counterpart need not even be as simple as a debate as to their relative truth. Evaluativists may claim that their position should be adopted because it is a more desirable view of illness, of what we should mean by illness.

In this paper, I have not argued for a particular view of the nature of mental illness, merely explored some of the options. I hope, however, that I have indicated just how deep the disagreement may go.


Foucault, M. (1989).Madness and Civilisation. London: Routledge
Fulford, K.W.M. (1989) Moral Theory and Medical Practice, Cambridge: Cambridge University Press
Lakatos, I. (1970) ‘Falsificationism and the methodology of scientific research programmes’ in I Lakatos and A Musgrave (eds)Criticism and the Growth of Knowledge, Cambridge: CUP pp91-138
Laing, R.D. (1960). The Divided Self. London: Tavistock
Mackie, J.L. (1977) Ethics: inventing right and wrong, Harmondsworth: Penguin
Pickering, N. (2006) The Metaphor of Mental Illness, Oxford: Oxford University Press
Szasz, T. (1972) The Myth of Mental Illness, London: Paladin
Thornton, T. (2007) Essential Philosophy of Psychiatry, Oxford: Oxford University Press
Wakefield, J.C. (1999) Mental disorder as a black box essentialist concept. Journal of Abnormal Psychology 108: 465-472
Zachar, P. and Kendler, K. (2007) ‘Psychiatric Disorders: A Conceptual Taxonomy’ American Journal of Psychiatry 164: 557-565