Reprinted from Ingleby D (ed) Critical Psychiatry. The Politics of Mental Health. Penguin, 1981.

In this chapter I shall try to make sense of the bewildering variety of theoretical approaches which confronts anyone trying to understand the problems of psychiatry, whether as a professional helper, a seeker of help, or just an interested observer. My aim will be to show that the many conflicting viewpoints which flourish can be understood only in terms of the philosophical systems underlying them - prior beliefs about what people are, and how we should try to understand them -and that these philosophical systems are themselves based on moral or political priorities. These ideas, of course, are not new ones: they formed the main thrust of the ‘anti-psychiatry’ movement in the late 1960s. What I hope to show here is the deeper perspective that can be gained by seeing psychiatry in the context of the human sciences as a whole, and by utilizing recent developments in our thinking about these disciplines.

One cherished illusion which must be lost before we can hope to understand ‘mental illness’ is the myth which helps to keep orthodox psychiatry on the move: the belief that what we need is simply more ‘findings’ - that round the corner lies some vital new fact which will settle the arguments once and for all. It is this conviction which has inspired the prodigious volume of research that the past half-century has produced in an ever-increasing flow. But despite this profusion of findings, it is doubtful whether we understand matters any better now than we did fifty years ago: as Coulter¹ puts it, ‘the literature on mental disorders is quite out of proportion to the adequacy of our knowledge about them’. Obviously, there is a fundamental difference of opinion somewhere about what constitutes ‘adequate knowledge’; and one of the aims of this essay will be to demonstrate the error of the naïve empiricist view that knowledge simply consists in the accumulation of findings, rather like pebbles which, if stacked up in sufficient quantity, are bound to reach the sky eventually. What matters, of course, are the principles which govern the acquisition and interpretation of ‘findings’; and these principles, although they are influenced by matters of fact, are not themselves discoverable empirically - they are as much philosophical as scientific ones.

Unfortunately, in scientific circles (and especially in psychiatry), the word ‘philosophical’ has come to be a term of abuse; philosophy is seen not as a liberating and enlightening activity, but as a form of primitive pre-scientific dogmatism from whose clutches we ought to be glad to have escaped. I will not deny that something called ‘philosophy’ did (and still does) exist deserving of these strictures; but the very chaos in which psychiatry and its related disciplines finds itself shows that science can no longer afford to lean on this as an excuse for sidestepping all philosophical problems. ‘Psychology’, in Wittgenstein’s words - and he might have included psychiatry - ‘contains experimental methods and conceptual confusions’, it is because researchers have cultivated empirical at the expense of conceptual sophistication that they do not realize how serious this fault really is. Get the findings in first, and the concepts will take care of themselves, says the received wisdom - reflected in the practice of regarding only fact-finding as genuine ‘work’, in research circles. Concepts, being ‘in the mind’, are supposed to have a vague and airy quality about them, so that it would not make much practical difference what form they took.

But this metaphor is totally misleading. If one is going to picture the abstract in concrete terms, then logic is actually a ‘substance’ infinitely harder than steel; and it follows that a science whose conceptual foundations have not been properly thought out is doomed to collapse, no matter what volume of findings is stacked up above them.

What we need, then, are not more findings - we probably have all we need, if only we knew what to do with them - but a reappraisal of the kind of explanations we should be looking for, and the kind of data which would be relevant to them. The ‘great debate’ in psychiatry, which professional and disciplinary lines of demarcation have so far succeeded in keeping frozen (in the English-speaking world at least), must start with the prior questions of what kind of creatures people are, and how we should go about observing and accounting for their behaviour and misbehaviour. In what follows I shall suggest that the deep structure of disagreements about psychiatry reflects an opposition of two fundamental views on these questions: one that I shall designate as ‘positivist’, which seeks to assimilate the human sciences as closely as possible to the natural sciences, and the other an ‘interpretative’ view, which sees the subject-matter and therefore the methodology of the human sciences as sui generis.

The reader will note that my account differs from most discussions of psychiatry, in which ‘the medical model’ is contrasted with non-medical alternatives: what I want to emphasize is something which criticisms of ‘the medical model’ often obscured - namely, the fact that classical psychiatry was basically modelled on positivism, rather than on medicine per se. (Indeed, it is possible to conceive of a non-positivist approach to medicine - such as von Weiszäcker’s² ‘anthropological medicine’ - within which many of the objections to traditional medical psychiatry would disappear.)

These two fundamentally opposed frameworks (positivist and interpretative) are not simply conflicting theories, between which we might arbitrate on empirical grounds, simply according to which ‘fits the facts’ better: they correspond to what the philosopher of science Kuhn³ calls ‘paradigms’ - that is, whole systems of prejudice about what constitutes useful and respectable data, what form theories should take, what sort of language scientists should use, how they should go about their business, and so on. In short, they correspond to different mentalities. Any observer soon realizes, too, that birds of a feather flock together in groups which reinforce their own views to the exclusion of everybody else’s; to this extent, adherence to one or other paradigm is equivalent to membership of a sort of community.

This immediately raises the problem of how different mentalities can be reconciled and different communities understand each other’s customs - or, in more technical terms, the ‘commensurability of paradigms’. There is more than a figurative resemblance here to the problems arising when anthropologists from one culture try to understand people from another; and the most important lesson anthropology has to teach us about this problem is the need to respect the autonomy and internal coherence of the ‘foreign’ culture, and the impossibility of simply invalidating one mentality from the standpoint of another. Here is Evans-Pritchard⁴ writing on Azande witchcraft:

In this web of belief every strand depends on every other strand, and a Zande cannot get out of its meshes because it is the only world he knows. The web is not an external structure in which he is enclosed. It is the texture of his thought and he cannot think that his thought is wrong (p.194).

And here is Professor Sir Martin Roth, founding President of Britain’s Royal College of Psychiatrists, writing on psychiatry’s critics:

How does it come about that for Sir Martin, to criticize his profession is to abdicate both reason and humanity? One answer is suggested in Evans-Pritchard’s fine phrase: ‘he cannot think that his thought is wrong’.

But of course if this were the only answer, then the outlook for psychiatry - indeed, for humanity itself - would be dismal indeed. Though between any two mentalities a vast area of incommensurability exists, we possess nevertheless the means of detecting common ground and building bridges from the one to the other. If this were not so, there would be no sensible basis for choosing between different paradigms, and each would occupy its own self-contained, impenetrable world - allowing as ‘admissible evidence’ only that which confirms its assumptions. Despite some of the more extreme interpretations of Kuhn’s views, such a relativistic conclusion is not justified in psychiatry at least: for the different contenders do know, to a large extent, what the others are talking about. (Some of them, such as psychiatrists turned anti-psychiatrists, have actually held at one time the positions they are opposing.) Though the different paradigms are to a striking extent self-confirming and self-contained, they could logically be brought into some relation with each other: the real issues which maintain the divisions between their holders, therefore, must lie elsewhere.

These issues are not hard to find - as I have already suggested, they are questions of values or politics. Negotiation between the holders of different paradigms is difficult, not only because each paradigm uses a different conceptual system, but because each represents different interests. This view, although put forward in a very sophisticated form by the social theorist Habermas,⁶ is nevertheless at root a fairly straightforward one. If we wish to travel by bus, we will acquire timetables and an accurate watch, and be very alert to large red objects coming down the road; likewise, if we are concerned about the oppressive aspects of our society, we will take very seriously a theory which suggests that mental illness is a manifestation of them. Thus (although the connections between knowledge and interest which Habermas describes are a lot more complex than this), it is not difficult to grasp how different paradigms may be preferred on the basis of different values.

The traditional objection to bringing values into scientific debate is, of course, that the argument then collapses into a relativistic free-for-all which is the very antithesis of what science is supposed to be. However, Kuhn’s close examination of the history of science showed that the grounds on which one paradigm is preferred to another are not exclusively scientific ones: the determinants of that choice lie to a large extent outside science, in social and psychological factors. (Kuhn mentioned solidarity and the pressure to conform, but solidarity is in turn based on common interests.) This view has naturally been hotly contested, since it blurs almost to vanishing-point the distinction between scientific belief-systems and (say) religious or political ones; but as Bernstein⁷ has pointed out, Kuhn’s conclusion is only disastrous if one presupposes, as he and most scientists traditionally have, that any choice determined by social or psychological factors is necessarily an irrational one.

Habermas, on the other hand, dismisses the view that there are only two rational modes of argument appeal to facts, and appeal to logic - and claims that so far from knowledge being independent of moral values and human interests, it is actually ‘constituted’ by them. These fundamental questions of paradigm choice, therefore, are to be settled by moral reasoning, which is not a matter of brute force and emotion (as empiricists assumed it must be), but of negotiation and a common search for ideals. As Habermas’ critics have been quick to point out, this is a difficult view to uphold unless one believes in an ultimate realm of absolute moral truths; but nevertheless it is easy to see how he reaches this conclusion, and how it offers a way of rescuing science from the relativistic chaos into which Kuhn appeared to have thrown it. The questions which the first section of this chapter will serve to open up, therefore, are: What values are enshrined in positivist psychiatry? What interests does it further, and do we want the kind of society which it leads us towards?

Our search for the ‘knowledge-constitutive interests’ in psychiatry is, however, hampered by the profession’s own categorical insistence that there are none there; indeed, the fundamental aim of positivism, as set out by the man who gave it a name, Auguste Comte, was to resolve social problems on the basis of factual considerations alone. Those who object to ‘bringing politics into science’ therefore have to be shown that in the human sciences, and particularly psychiatry, politics is already there, just as it was in Comte’s own project. My first task will be to show that the view of psychiatry as ‘value-free’ is pure whitewash; only when its pretensions to objectivity have been got out of the way can we proceed to discuss what, in fact, its moral premises are, and which interests it furthers.

As I have stressed above, classical psychiatry was not merely an offshoot of the medical profession, springing up spontaneously in the soil of the clinic and the asylum, but was part of a much larger project in the history of man’s attempts to understand himself. Positivism aimed to construct human sciences on the model of natural sciences: contemporary authorities still place psychiatry firmly within this paradigm. ‘The foundations of psychiatry have to be laid on the ground of natural sciences’, say Mayer-Gross, Slater and Roth.⁸

From the beginning right up to the present day, this alliance with natural science was seen as a way of lifting psychiatry above the level of particular interests and arbitrary prejudices, and giving it an impartial and detached authority. In this section, I shall try to show that what it achieves is quite the opposite.

Underlying the positivist enterprise is, of course, one major assumption: that there are no features distinguishing human beings from the rest of nature which might necessitate the adoption of a different paradigm. I shall argue that this assumption alone is such a large and implausible one as to undermine from the start any claim to impartiality: if, as the opponents of positivism have claimed, there are such distinctive features, then the naturalistic paradigm merely serves to distort our understanding.

I shall consider in turn the two chief elements of the paradigm: its prescriptions about how to go about collecting data, and about how to construct theories. Positivism assumes, in the first case, that observations can be made objectively - that measures can be defined operationally, and applied in a precise, replicable fashion; and in the second, that theories can be constructed on the same causal, deterministic basis as in the natural sciences. Both these fundamental principles have been challenged in their embodiments across the whole range of human sciences - from Durkheimian sociology,⁹ through behaviourist psychology (Harré and Secord;¹⁰ Gauld and Shotter ¹¹), to the recent criticisms of psychiatry. Although the particular objections I shall make have all been raised before by ‘anti-psychiatrists’, the point I want to bring out is that most of them relate not just to psychiatry, but to positivism in any form.

Perhaps the sharpest difference between positivist psychiatry and its critics arises over its concept of data and how to go about collecting it. The ‘scientific’ psychiatrist aims to collect observations on his patients in the same rigorous and detached fashion as, say, the astronomer observing stars; his critics, however, are united in regarding this as a tactic which can only obscure his vision, and render useless whatever theories (biochemical, social or otherwise) are constructed out of such data. As Laing¹² caustically put it, ‘you cannot make a statistical silk purse out of a clinical sow’s ear’. Anyone who has had to deal with psychiatric case-notes will recognize the style: the patient is pinned down by a few cut-and-dried epithets, with no hint of the complex ambiguities of human conduct or of the context in which the patient acts and is observed. Constantly in the background is the example provided by ‘scientific’ medicine: the patient reduced to a set of basic variables - temperature, pulse, blood-pressure ... emotional tone, aggressivity, reality-sense.

It is no accident that the criticisms of this ideal of objectivity in psychiatry are closely paralleled by arguments going on within psychology and sociology. The notion that social facts are ‘things’ which can be described like any other natural objects has been the target of continual criticism since Durkheim first articulated it in 1895. In the next section I shall say a little more about this critical tradition; the current crisis of confidence in positivist epistemology (theory of knowledge) is largely the result of work done by Wittgensteinian analytical philosophers and by the phenomenological and ethnomethodological schools of sociology. These ideas have been applied to psychiatry by Robertson,¹³ Coulter¹⁴ and Heritage;¹⁵ their implication is that the lack of objectivity in the human sciences will never be cured simply by trying to do better - rather, it stems from the application of totally irrelevant criteria.

‘Objective description’ means many things. In the first place, it means that concepts should have an explicit and uniform usage, and that measurements should be definable in terms of the operations which give rise to them. In psychiatry, however, ‘operational definition’ is an obstacle in front of which whole armies of theoreticians have blunted their weapons - because, bunkered by their naturalistic assumptions, they have failed to take account of the distinctive problem of characterizing human behaviour and experience.

The root of this problem lies in the fact that the most useful descriptive concepts are simply not definable in terms of a finite set of observations; there are as many ways of expressing ‘anger’, for example, as there are people to get angry and situations to get angry in. None of us have much difficulty as non-scientists in recognizing anger when we see it, because we have a lifetime’s experience of how the concept is applied; common sense enables us to cope with the ‘open texture’ of such concepts. Common sense, however, does not meet positivism’s requirement of an explicit set of criteria: as the ethnomethodologist Garfinkel¹⁶ pointed out, there are no formal rules according to which a reader can reconstruct, from material classified by a researcher, the precise observations on which it is based - he has to rely on a shared background of ‘taken-for-granted’ interpretative competence. Any attempt to specify the rules relating descriptions to situations would run into endless qualifications, special clauses, and so on; this Garfinkel dubbed ‘the etcetera problem’. Faced with this problem, psychiatrists have responded in one of three main ways.

The first (which Clare,¹⁷ as a sensible middle-of-the-road practitioner, adopts) is to recognize it as a problem, and to admit that ‘clinical judgement’ is an essential ingredient of psychiatric description; this, at least, is a realistic conclusion, but ‘clinical judgement’ is a blank cheque which can be filled in with any amount of tacit biases and unwritten rules, and completely undermines psychiatry’s claim to be based on observations more solid than lay judgement. For the fact that clinical expertise is maintained by a professional clique is not, in itself, any guarantee of objectivity - more likely the reverse.

The second response to the ‘etcetera problem’ is to disguise it by offering makeshift definitions, glossing over the tacit assumptions necessary for applying them to any particular instance: most descriptions of psychiatric ‘syndromes’ are of this form (‘flatness of affect’, ‘bizarre gestures’, ‘impulsive outbursts’, ‘withdrawn manner’). Of the same ilk are psychological questionnaires asking deceptively straightforward questions of the type: ‘Are you usually happy?’, which simply lay the onus of interpretation on the subject or the coder. (Without such interpretation, the only possible answer to such questions is, of course, ‘It all depends!’)

The third ‘solution’ is that advocated by thorough-going positivists such as Eysenck: it is to deem open-textured concepts like ‘anger’ or ‘happiness’ unsuitable for scientific purposes, and to replace ‘clinical judgement’ by deliberately artificial constructs such as a person’s score on a test, or some physiological index such as skin conductance, blood pressure, or EEG readings. This strategy is familiar in the form of the slogan: ‘Intelligence is what intelligence tests measure’; it is justified by an entirely specious analogy with ‘operational definitions’ in the physical sciences, in which concepts like Mass or Force are given definitions which do not correspond to everyday usage. The fallacy is, however, that the physicist’s definition of ‘mass’ is justified by its meaningfulness and utility within a well-established theory, and the theory itself translates readily back into the language of everyday experience. None of this is true of the ‘operational definitions’ proposed in psychology; they are a perfect example of what Torgerson¹⁸ and Cicourel¹⁹ call ‘measurement by fiat’, and to pretend to ‘validate’ them by bringing them into line with the subjective judgements of psychiatrists, teachers, or whoever, is simply pulling oneself up by one’s bootlaces.²⁰

Of course, the psychiatrist may reply that within his province - that of ‘mental illness’ - these epistemological problems do not apply; what distinguishes ‘symptomatology’ from ordinary human conduct, he may say, is precisely that it does not represent the meaningful activities of human agents. He may concede that off duty, in the realm of ‘normal’ people, he is obliged like anyone else to make subjective, common-sense interpretations: but since (unlike most sociologists and psychologists) it is not in this realm that he operates, he may well think that he can evade Garfinkel’s ‘etcetera problem’.

The snag with this particular argument is that in order to know that he is not in the ‘normal’ realm, he has to make a judgement which relies even more heavily on subjective understanding than the everyday interpretations discussed above: to make a warranted ascription of insanity, he has to be sure that none of the ordinary ways of finding conduct intelligible actually works - and this he cannot do without applying each of them in turn, and thus invoking subjective judgement. To take a particular example, in order to categorize a patient as ‘aggressive’ or ‘depressive’, he has to be sure that they are not being understandably angry or sad; and to do this he must examine their behaviour in its context and apply complex cultural norms to evaluate its reasonableness. As Coulter²¹ convincingly shows, this means that insanity ascriptions are inescapably rooted in common-sense cultural understanding, and to imagine that they could be grounded in something which transcends common sense (i.e. in neutral, scientific authority) simply makes no sense.

Two points have been made so far. The first is that while it may be in principle possible to state exactly the criteria for applying physical concepts, it is in principle impossible to do so for concepts describing human activities and states of mind. Descriptions of the latter are always subjective interpretations - subjective not in the sense that there are no criteria, but that the criteria are unstated ones, lying in the culture itself. The second point is that judgements of insanity are even more dependent on cultural competence, since they assert that no conventional interpretation can be successfully applied. The first point undermines the possibility of objective description in the human sciences generally, the second applies to psychiatry in particular.

One of the ways in which critics have sought to attack psychiatry’s claim to objectivity, and supporters to defend it, is by pointing to statistics on the reliability of psychiatric diagnoses (that is, the extent to which different psychiatrists are likely to agree on the classification of cases). Published figures in fact vary widely, giving both parties ample ammunition for their case; thus, Heather²² can conclude that ‘psychiatric diagnoses are extremely unreliable’, while Clare²³ claims that they are no more so than in other branches of medicine. I shall not attempt to review the relevant statistics here, because I think there are two considerations which make this issue something of a red herring.

In the first place, reliability statistics usually relate to differential diagnosis, i.e. to decisions about the classification of a case already presumed ‘disordered’ - not about whether the patient’s behaviour is or is not understandable in the circumstances. The celebrated experiment of Rosenhahn,²⁴ in which eight perfectly sane volunteers spent several weeks in mental hospitals without their sanity being detected, suggests that such absolute diagnosis may be wildly inaccurate. In any case, high reliability does not necessarily indicate accuracy; Rosenhahn’s study is a perfect illustration of this, since all the medical staff in the hospitals studied agreed that the volunteers were insane, but all of them were wrong! Rather, what we should be talking about s the validity of diagnoses - whether psychiatrists actually detect what they claim to detect; but if, as the foregoing arguments show, there is no explicit definition of precisely what it is that they are detecting, then there is no way of demonstrating publicly that they have succeeded in detecting it.

Thus, psychiatric diagnoses can never aspire to objectivity in the natural-scientific sense, and to claim that they can is merely to conceal the tacit rules, conventions and biases which necessarily govern their application. Indeed, most diagnoses forfeit their claim to be objective descriptions for the simple reason that their basic function is not a descriptive one: in everyday psychiatric practice, a diagnosis represents an administrative decision, which is governed by many other considerations besides the actual state of the patient: the family situation, the treatment available, legal considerations, and so on. This they have in common with all such ‘official statistics’: for sociologists, the locus classicus of this problem is Durkheim’s²⁵ mistaken reliance on coroners’ verdicts as an ‘objective’ index of suicide rates. In addition, psychiatric diagnosis may serve as an emotional defence for staff and relatives: Svein Haugsgjerd has suggested that the label ‘schizophrenic’ may be used to protect hospital staff against the pain of disappointment when their efforts are seen to be fruitless. Of course, not all psychiatrists are unaware of these problems (see, for example, Wing ²⁶); but none have convincingly demonstrated a way round them, and it would certainly come as a surprise to the rest of the human sciences if psychiatrists were to succeed where all others had failed.

Another sense in which natural sciences try to be ‘objective’ is in maintaining a division between subject and object; the scientist attempts, as far as possible, to eliminate the ‘observer effect’ or ‘reactivity’ produced by his intervention. But to attempt this in the study of people is quite inappropriate. Unlike the physicist, the human scientist cannot observe in a vacuum: every scientific set-up is a social situation, which the scientist cannot avoid influencing. As Laing²⁷ convincingly showed, the Kraepelinian method of case-presentation ignores completely the effect of the situation on the patient: to this very day, patients are paraded in front of a group of observers and described not only as if they weren’t there, but also as if the observers weren’t either. Thus, for Clare to assert²⁸ that the aim of this approach is to provide neutral, theory-free descriptions, and then to characterize Laing as somehow an enemy of the truth, is extraordinarily perverse. A considerable body of recent work has shown that the same biases operate in positivist social psychology (see, e.g., Lindsay and Aronson²⁹). Eysenck³⁰ ridiculed the idea that psychoanalysis could be (in Freud’s words) ‘tested on the couch’, but it transpires that the sterile atmosphere of the laboratory is as replete with unknown forces as any consulting-room. The large literature on ‘experimenter effects’ (Rosenthal³¹) has also shown how difficult it is for the observer to avoid creating the data he needs to support his theory.

The second set of presuppositions which positivism has imposed on psychiatry concerns the form which valid explanations have to take. The only legitimate mode of explanation is assumed to be causal, on the basis (again following Durkheim) that the laws governing human life are of the same character as those governing nature.

In psychiatry, this doctrine has what I shall call a ‘strong’ and a ‘weak’ form. The ‘strong’ form, variously called the ‘faulty-machine’ or ‘disease’ model, suggests that the causal factors underlying mental illness are physiological disorders; the ‘weak’ version still invokes causal explanation, but blames the problems on psychological or environmental factors. It has become traditional (cf. Clare³²) to regard British psychiatry as ‘eclectic’ or ‘non-ideological’, because of its tolerance of both these points of view; but I shall show that they are simply two sides of the positivist coin, and that their dominance, so far from being inevitable, is in fact unjustified and obscurantist in its effects.

Let us take first the three sorts of evidence which are usually adduced in support of the ‘organic’ view, that mental illnesses have a direct physical cause. Traditionally, arguments about the a priori appropriateness of organic explanations are brushed aside with the retort that ‘they work’, so they must be appropriate; I shall try to show that large doses of faith are required to sustain a belief in their efficacy.

(i) Genetic studies. Studies of the pattern of inheritance of mental illnesses - particularly psychoses - are routinely used to bolster the claim that a genetic factor, and therefore a physical malfunction, is involved in their production. The data, however, are not as friendly to this view as they seem.

The methods by which the genetic contribution to psychiatric conditions is investigated closely parallel those used to estimate the heritability of IQ - a field in which, for some reason, arguments and criticisms have become much more sophisticated. Briefly, all the studies used as evidence for genetic transmission of ‘mental illness’ suffer from the same methodological weaknesses as their counterparts relating to ‘intelligence’. Kamin’s³³ book on IQ serves as the model for a thorough critique of this work - itself as yet unwritten, though Laing³⁴ and Jackson³⁵ have made a start. As with IQ, three sorts of data are involved.

(a) The differences in concordance rates between identical (monozygotic) and non-identical (dizygotic) twins. (The concordance rate is the probability that if one twin is affected, the other will be too.) Identical twins, of course, have all their genes in common, so that they would show a much higher degree of concordance in the case of a genetically transmitted condition: and, in many psychiatric conditions, they do. Unfortunately, this finding could come about in other ways - in the majority of studies, by biased diagnoses of zygosity of illness (because observations were not made ‘blind’), and also because the environments of identical twins likewise have far more features in common than those of non-identical twins. The idea that identical twins are intrinsically more likely to have identity-problems, and so on, seems to be unfounded (see Clare³⁶); but they do, often to a blatant extent, identify with each other more strongly than non-identical twins, and even though in some cases this might help the disturbed partner to remain sane, the net effect would still be to increase the rate of concordance. Then again, disturbance in one twin may make everyone more ready to ‘see’ disturbance in the other, which may set in motion the processes of ‘deviance amplification’ described by labelling theorists (see Section II) - a perfect example of a self-fulfilling hypothesis. Only if one starts out by rejecting all such social influences as implausible can one obtain unequivocal support for the organic model from these data.

(b) Concordance rates in separated identical twins. If these are above chance levels, then - ideally - we have a direct measure of genetic influence; conditions, however, are never in fact ideal. As in the IQ studies, samples are often biased, ‘blind’ diagnoses are sadly lacking, and the classification of twins as ‘separated’ (which is likewise seldom made ‘blind’) can become ludicrously flexible.³⁷ Furthermore, the likelihood that selective placement produces similar environments is neglected, as are age-related effects (i.e. correlations arising because twins have the same ages, and many psychiatric conditions are more likely at certain ages than at others.) Finally, no proper base rate is established from which to estimate the chance level of concordance: the circumstances surrounding such separations are never happy ones, and this fact itself may produce a higher incidence of problems in the group as a whole. All these factors combine to make the meaning of such data highly ambiguous.

(c) Incidence rates in adopted children of mental patients. Here, as with the comparisons between identical and non-identical twins, there is the possibility of environmental ‘labelling’; and as with separated identical twins, there is the likelihood of selective placement, i.e. those parents who are given the children of mental patients to adopt being less able to provide a ‘good’ environment.

What conclusion should we then draw from these studies? According to Kamin, in the case of the data on IQ, the ‘null hypothesis’ that genes play no part whatsoever emerges quite unscathed; but this is a rather biased way of putting it. We could have started from the null hypothesis that genetic influences were present, but only partial; and this assumption would survive equally well. The assumption that we can understand mental illness in genetic terms alone, however, most certainly would not.

(ii) Physiological studies. The view that the key to mental problems could be discovered by looking inside the head was well established long before the dramatic success of the organic approach in treating syphilis (in the 1920s). As Treacher and Baruch describe in Chapter 4, 76 per cent of mental patients in the year 1890 received the privilege of an autopsy to determine the state of their brain; and in the present day, the overwhelming majority of psychiatric research is devoted to neurological or biochemical aetiology. Each new physical explanation that is put forward - defects in gross anatomy, state of arousal, serum levels, or transmitter substances - reflects faithfully the current preoccupations of neurologists.

Time and again, however, such research founders on the elementary logical problems which accompany any attempt to deduce a causal influence from a correlation. Firstly, it may be that the physiological correlate of a given mental state is a product, rather than a cause, of it - for we are generally only able to perform physiological investigations after the event. (This causal link may be very indirect, as in the conditions which are found to result from the peculiar diet of mental patients.) Secondly, both the physiological and the psychological state may be produced by another factor - for instance, by the whole complex of circumstances we call a person’s ‘way of life’, with all its physical and psychological dimensions.

(iii) Physical treatments. Often it is argued that the efficacy of physical treatments (drugs, ECT, psychosurgery, etc.) is itself an indication that the original condition was physiologically determined. Two questions arise: firstly, what is the evidence for this efficacy? Many treatments only seem to be effective as long as they are believed in, either by the patient (‘placebo effect’), or the medical staff (‘selffuffilling prophecy’), or both; moreover, those which are effective may only be so because of some non-physical accompaniment, e.g. the extra care or the unconscious phantasies of punishment and reparation that may accompany trips to the operating theatre. A moment’s thought will reveal the practical and ethical constraints which make it very difficult for a researcher to eliminate these possibilities.

The more serious question is: what do we call ‘effective’? Almost all treatments have undesirable side-effects; and if ECT reduces the pain of events only by helping the patient to forget them, or if tranquillizers make people able to handle their emotions only by leaving them with no emotions to handle, then talk of a ‘cure’ becomes rather ironical. In that sense, after all, death ‘cures’ everything.

But even if the treatment does produce a real and positive effect, what does this prove? All that actually follows is that the problem could have had a physical cause; not that it did, for it is possible for a phenomenon to arise in many different ways. Thus the value of this evidence is chiefly rhetorical: to convince the sceptic that physical causes are a possibility; this role we should not deny them, but anything more is plainly unwarranted.

(We could also include under this heading ‘treatments’ which consist of a physical intervention producing illness, e.g. ‘psychotomimetic’ drugs; provided that the treatment was not self-administered (as ‘psychedelic’ drugs usually are), we can rule out reversed causation, but the mechanism of causation may still be extremely indirect; and there is widespread disagreement about how closely the induced states resemble ‘the real thing’.³⁸ The same considerations apply when we consider the similarities between certain ‘functional’ illnesses and the effects of brain damage or disease.)

*

An open-minded judge must conclude, I think, that all three types of evidence I have discussed have highly ambiguous implications: organic pathology is not established in the majority of psychiatric problems, nor is there any chance that it could be without attention to the problems I have listed above. But even where it is implicated, the ‘faulty-machine’ approach remains an inadequate one: to understand why, it is necessary to go deeper into the philosophy of explanation.

The fundamental weakness of the organic approach is that it adopts a view of causation which, despite its respectable ancestry, can be dangerously misleading. In the natural sciences, under the influence of the philosopher Hume, a ‘cause’ is regarded as any antecedent factor from which an event is highly predictable (always assuming that a third factor is not responsible for both). However, where the event depends on a number of factors - as in many physical sciences and most human ones - this usage can be very misleading, for it may tempt us to overlook the long and complicated pathway that leads from ‘cause’ to ‘effect’. This is often the case when we are discussing the effect of physiological or genetic variations on human behaviour.³⁹ To give an extreme example, people who inherit a tall physique are very prone to bump into doorways, but the bruises on their heads are not ‘caused’ by their genes, even though they may be very predictable from them. Architectural fashions, urgent appointments, and lapses of attention are just as much responsible for the problem (to say nothing of the diet which enables ‘genes for height’ to be expressed in the first place). Thus, the outcome of any physiological state depends on many factors, some of them purely matters of social convention; to take another example, the female chromosome pair (XX) does not ‘cause’ a person to have a feminine personality - even though in our society the two are highly correlated - for such a personality is the product of social responses to a person’s biological sex.

All this gives rise to what I should like to christen the ‘So What?’ problem in aetiology; for that is the only appropriate response to many findings which implicate physiological or genetic factors in human behaviour. A constitution which leads to one person being diagnosed as schizophrenic may, given a different life-history, be the basis of exceptional achievements; Heston,⁴⁰ in his study of the adopted children of schizophrenics, found that those who did not become psychiatric cases had more ‘artistic’ or ‘creative’ occupations than a comparison sample. Although, as I have shown above, the evidence implicating physical factors in mental illness is mostly very dubious, nobody in their right minds would deny that such influences may exist; the point I am trying to make here is that their discovery does not warrant the adoption of an exclusively organic approach. Thus, Siirala⁴¹ was able to write a psychoanalytic study of communication in the families of deaf children, despite the organic origin of their defect. What matters is how a person lives out their physical condition.

That physical understanding by itself is inadequate is, of course, a c1iché in ‘eclectic’ British psychiatry; but I want to show next that the narrowness of the positivist paradigm is by no means eliminated simply by introducing an alternative set of aetiological factors called ‘psychological’ or ‘environmental’. The crucial distortions reside not in the sorts of factors that are considered causal, but in the notion of ‘cause’ itself. To illustrate this, I shall now consider what I called earlier the ‘weak’ version of positivist psychiatry.

Under this heading I am including explanations of mental illness in terms of behaviourist psychology: some positivistic renderings of psychoanalytic theory, and sociological work in the Durkheimian tradition, such as the recent study of depression by Brown and Harris.⁴² ‘Behaviour therapy’ has sought to apply to human problems a theory of learning which barely fits the albino rat: mental illnesses are described as the result of an unfortunate history of conditioning. Psychoanalysis might be thought to start from a more sophisticated approach to mental functioning, but in practice it all too often degenerates into a sterile juxtapositioning of mysteriously inferred ‘mechanisms’. The general strategy of positivist sociologists working on mental illness has been the same, but with the ‘mechanisms’ located outside the patient’s head - in society - instead.

These approaches tend to be championed by those, such as social workers or clinical psychologists, who feel that psychiatry’s defects are basically due to the stranglehold of the medical profession. I shall try to show, however, that to adopt them is to jump out of the frying-pan into another frying-pan - since they all adhere to the same paradigm of explanation as the medical model itself.

Before discussing their conception of ‘causality’, we should note that these approaches have inherited from the medical model the same dubious ideal of ‘objectively’ describing human events - only here, the problems arise not merely in describing the patient’s condition, but also in characterizing the events and experiences supposed to explain it. Most environmentalist approaches start by describing the patient in much the same way that organic psychiatry does - indeed, that is hardly surprising, since they originally set out to provide ‘rival’ explanations. As in organic psychiatry, however, the inevitably subjective nature of such descriptions is denied, and the fact that clinical diagnoses do not merely serve to describe also tends to be overlooked. Even such sophisticated research as that of Brown and Harris still retains ‘illness’ as a descriptive category, when to do so is to beg the very question such research might be expected to answer (viz. how much of depression is understandable?).

When it comes to characterizing the conditions supposed to cause mental illness - be they childhood events, traumatic experiences, or social circumstances - the same misguided attempt is made to dispense with subjective meanings; here, if anything, the consequences are more serious, because it is undeniable that events affect us not for what they are, but for what they mean to us. It would seem churlish to deny that the hardships found by Brown and Harris in the environments of depressed housewives are inherently depressing - too many children, too little space, no partner, no job - but the fact is that not all are depressed by them, and in other conditions (notably wartime) these and other hardships may be associated with a dramatic fall in psychiatric complaints. Unless we allow that the subjective meaning of objective events is what influences us, we merely weaken the predictive power of environmental explanations, and leave a breach into which organic psychiatrists will eagerly step with talk of ‘constitutional predispositions’. At best, most environmental research characterizes situations in terms of the meanings they conventionally have: thus, Coulter⁴³ describes the studies of ‘schizophrenogenic’ families which preceded Laing’s as ‘an attempt to scientize what amount to a set of common-sense cultural judgements’.

Thus, environmentalist approaches are even more hamstrung by natural-scientific ideals of ‘objectivity’ than their organic predecessors. However, more relevant to the present discussion is the fact that they rely on the same basic notion of ‘cause’. In claiming that a state of mind is determined by a particular environment, they are in fact making a subtle claim about the nature of that mental state: for they imply that the person who has it is essentially not a rational agent. Their claims are about ‘what makes people do things’, rather than ‘what people do’; the point is that we typically describe being made to do something when seeking to excuse behaviour for which we wish to disown responsibility. Of course, I am not trying to suggest that one can only demonstrate one’s responsibility or agency by reacting to one’s environment in a capricious and unpredictable way (if anything, the reverse is true); the point is that an agent’s situation provides grounds or reasons for his actions, rather than causes. The distinction is crucial, for though human behaviour is indeed (for the most part) orderly, the ‘laws’ underlying it are not of the same logical type as those governing the movements of physical objects: they go hand-in-hand with agency, and in the last resort the laws themselves - unlike laws of nature - are man-made.

A discussion of the kinds of explanation which are appropriate to agents will have to wait until the second section of this chapter. The important point here is that positivist explanations essentially rule out agency; so any theorist who sets out on the quest for ‘causal factors’ or ‘aetiology’ is by no stretch of the imagination starting from a neutral or objective standpoint - he is presupposing something about the relation of conduct to its surroundings which it should be the task of research to question, not to assume. I am not, of course, arguing that people should be treated as agents a priori; this is a common point of view among linguistic philosophers and their disciples, but all it seems to me to accomplish is a redefinition of the word ‘people’. The point is that they should not be treated a priori as non-agents, which is exactly what the phoney eclecticism of orthodox psychiatry tacitly achieves.

The term reification very usefully describes what is done when the meaningful activity of agents (‘praxis’) is described as if it were the outcome of an interplay of causal forces (‘process’). I shall return to this concept later, but we may note here the ironical fact that self-reification is, according to the theorists we shall encounter in the next section, the very essence of mental illness: the patient ceases to experience his life as meaningful and himself as an agent, and as long as the doctor remains within the positivist framework he can do nothing but encourage this self-invalidation. As Foucault⁴⁴ puts it, the patient is ‘alienated in the doctor’. Should the doctor wish to practise a therapy which stresses the recovery of intelligibility and purpose within the patient, he will abruptly discover the limitations of ‘eclecticism’; for the whole of the positivist paradigm negates this aim. (Baruch and Treacher’s⁴⁵ carefully documented account of psychiatry in action shows up clearly the ‘double-bind’ within which would-be progressive psychiatrists find themselves.) But I shall argue in a moment that this invalidation of the patient’s experience and behaviour is what gives reification its key position in the performance of psychiatry’s social role.

To sum up so far: I have claimed that the basic concepts and methods of psychiatry, so far from being neutral, in fact constitute a paradigm, which I have characterized as ‘positivist’ rather than simply medical, since it is not left behind in psychological and sociological approaches. The prescriptions of this paradigm regarding both observation and theory-construction were examined, and found to be uniquely inappropriate to its subject-matter. Psychiatry’s image of itself as a science like other sciences was found to be deceptive, since (as Coulter argued) its mode of observation is not objective, but pragmatic and inherently subjective. Its mode of theorizing, while more literally borrowed from other sciences, is highly unscientific, since it begs obvious questions about the nature of the experience and behaviour it purports to explain.

This opens up an important set of questions. As long as we accepted the view of psychiatry’s supporters, that its aim was to seek the truth in the only rational way available, then we could not attempt to understand it in terms of other values and interests - just as the preKuhnian view of science (put forward by such philosophers as Karl Popper) ruled out of court any considerations of sociology or politics in scientific activity. If, however, psychiatry does contain presuppositions in Kuhn’s sense, then we may ask what influenced the choice of these particular ones. All my preceding arguments, then, have been necessary to open up the question we shall now consider.

Why should psychiatry, from the very beginning, have sought to pattern itself on the model of the natural sciences? Kuhn would probably answer this question in terms of conformity and persuasion: psychiatrists identified with the medical profession, and could not afford to isolate themselves from it by challenging its positivist assumptions - and medicine, in turn, could not alienate intellectual orthodoxy by denying the primacy of science. To invoke ‘conformity’, however, merely moves the question one stage further back: for why was the medical profession psychiatry’s reference-group, and why did natural sciences enjoy their primacy? The answer to this, I believe, can only come from an examination of psychiatry’s history and function as a social institution. It is this which determines the practice of psychiatrists, while theory merely provides an elaboration and rationalisation of it. Thus, I shall argue that positivist models not only generate techniques through which the psychiatrist can carry out his socially ordained function, but provide an essential smokescreen behind which the real nature of that function is concealed.

The social function of psychiatry can be summed up as the control of deviance; that is, the norms of mental ‘health’ and ‘illness’ are essentially matters of cultural judgement, although positivism misrepresents them as matters of empirical fact. It is probably Thomas Szasz⁴⁶ who has done most to denounce publicly the claim that psychiatry is ‘value-free’. Szasz’s analysis, however, suffers from too simple a picture of the norms against which the mental patient is deviating: ‘madness’ is not the same as ‘badness’, since it violates at a much more basic level the conventions governing thought and social relationships. What the ‘mad’ person is up to does not (ostensibly, at any rate) have a place in the everyday vocabulary of motives - a fact which led Lemert⁴⁷ to describe it as ‘residual’ deviance.

Probably the most accurate and useful account of the meaning of insanity is that offered by Coulter, whose social philosophy enables him - unlike Szasz - to see rationality itself as basically a moral concept, and ‘cognition as a moral order’.⁴⁸ However, that same philosophy also prevents Coulter from making any kind of critique of this moral order, and of psychiatry’s role in maintaining it; for him, it is ‘not intelligibly subject to doubt’. He thus follows the tenets of linguistic philosophy and ethnomethodology to their logical conclusion: a relativism which maintains that rationality, and the nature of social activities, can only be what people say they are. Thus Coulter’s analysis, though unsurpassed as a description of psychiatric practice, comes to a full stop at the very point from which I want to take off.

We may wish therefore to go back to Szasz, who does allow himself to criticize both prevailing conceptions of normality and the role of psychiatry; but his efforts are undermined again by simple-minded preconceptions. Szasz’s moral ideal is personal liberty (in the negative sense of ‘freedom from constraint’, rather than ‘being enabled to act’); psychiatrists are enemies of liberty because they seek to impose a ‘collectivist’ society, in which the needs of the group are set above those of the individual. As Sayers⁴⁹ trenchantly shows, Szasz’s philosophy is none other than our old friend from the nineteenth century, laissez-faire individualism - the ideology of ‘free enterprise’; it is, of course, impossible to set the needs of all individuals above those of society as a whole, and the opposition between individual and society which this implies is logically absurd. Individualism and collectivism can never be genuine alternatives, and conflict has to be analysed, not in terms of abstract entities like ‘individual’ and ‘society’, but in terms of particular interested parties: this the anti-Marxist Szasz is not prepared to do, so that his analysis of the politics of psychiatry comes to a dead end.

Thus, Szasz is correct in his insistence that psychiatry is concerned with ‘problems of living’, not with the maintenance of some criterion of health lying outside morality and culture; but the conflicts which give rise to these ‘problems of living’ do not stem from a contradiction between the abstractions ‘individual’ and ‘society’, but from the tension between human needs and demands and the particular social institutions (work, the family, education, ‘politics’) which are supposed to provide for them. My analysis⁵⁰ - for which I claim no originality -starts from the premise that these institutions in their present form do not represent ‘the common good’, but a particular set of interests which conceal themselves behind the notion of ‘economic progress’. Psychiatry, in turn, protects the efficient functioning of these institutions by converting the conflict and suffering that arises within them into ‘symptoms’ of essentially individual (or at best familial) ‘malfunctioning’; it thus attempts to provide short-term technological solutions to what are at root political problems.

The foundations of such an analysis were laid by Michel Foucault in his influential work, Madness and Civilisation.⁵¹ The task which medical psychiatry inherited in the nineteenth century was that of confining civilization’s misfits in a place where they could not disturb the smooth harmonious surface of social life, the chief offenders being, as Foucault shows, the ‘idle poor’. But while Foucault’s analysis still has relevance today - witness the work of Franco Basaglia described in Chapter 6 - I believe a considerable modification of his views is necessary to take into account two important changes which the medical model has led to. These are, firstly, the ‘psychiatrization’ of the whole population - whereby mental illness is not seen as confined to a particular segment of humanity, but as potentially occupying segments of everybody’s mind. Whereas previously one was either mad or sane, mental disorder is now seen as an illness which, like influenza or lumbago, can be partial and temporary. Secondly, as a necessary part of the same process, the feat of segregating mental suffering and conflict apart from ‘normal’ life is achieved not by brute force, but by ideas: for iron bars and chains are rendered unnecessary if the medical ideology can successfully define the patient’s condition as lying outside the realm of meaningful human experience.

It is here that the distinctive role of positivism emerges for, as we saw above, its whole effect is to translate what it deals with from the human order into the non-human - what I called ‘reification’. As long as a person’s actions are seen as praxis, they tend to be taken seriously - even if being taken seriously, at times, leads to the agent being bumped off. But process - the blind workings of nature - evinces no such respect: it does not need to be argued with or fought against, but can be manipulated in whatever way suits the manipulator - and if it is defined as a ‘pathological’ process, then nobody in their right mind can deny that it ought to be got rid of. Hence we see the enormous potential of medicine as an instrument of social control - a potential which is not confined to the field of psychiatry (see Illich⁵²; Waitzkin and Waterman;⁵³ Baruch and Treacher⁵⁴). Moreover, although what I called above the ‘strong’ form of positivism (the ‘faulty-machine’ model) is the ideal way of removing all blame from social institutions, the environmentalist version is hardly more threatening to the status quo: for it too denies that the patient’s response to his or her surroundings is intelligible and valid.

Thus we see that the inappropriateness of the positivist paradigm, in rational terms, is precisely what makes it so appropriate to the task of preserving existing institutions from the threat of change. Psychiatry takes on itself the responsibility for people’s pain and frustration; it confiscates their problems, redefines them as ‘illnesses’, and (with luck) exterminates the symptoms. Come unto me, all ye who labour and are heavy laden (it says), and I will give you - oblivion. As this apparatus perfects itself, so the goal of a society truly fit for human habitation recedes further and further into the future. Radical politics, and the undermining of psychiatry, are thus inseparable from each other; moreover, dismantling the enormous mystification of social life which psychiatry has built up over the last hundred years is not a task to be completed overnight. In the next section we shall examine the progress that has been made towards this goal.

As we saw in the first section, many who have sought to escape from the paradigm adhered to by traditional psychiatry have fallen short of their goal, and ended up merely by embracing a different (psycho-logical or sociological) version of positivism. Indeed, the hegemony which natural-scientific ideas have enjoyed over the last hundred years has made it hard to do otherwise, since it has become virtually taken for granted that there are no other valid forms of knowledge. Consequently the approaches we shall examine in this section are less popular, respectable and academically established than the positivism they set out to replace. Positivists themselves have regarded these approaches with a mixture of contempt and bewilderment - for, in the words of Evans-Pritchard, ‘they cannot think their thought is wrong’.

In subsuming these alternatives under a single paradigm I am, of course, doing violence to some important distinctions; and there are many authors who, because of their torn loyalties, hover uncertainly between one paradigm and the other. Nevertheless we may define the essence of the alternative paradigm simply by standing the presuppositions of positivism on their head.

Thus, contrary to positivism’s ‘naturalistic’ assumption, that there are no features distinguishing human from natural reality which might necessitate the adoption of a different paradigm, the writers in this section assume that there are such features. The crucial difference ascribed to human beings (not a unique property, perhaps, but one which emerges gradually in higher animals) is the capacity for meaningful behaviour or ‘praxis’ - the ability to intend or express; furthermore, the nature of praxis is such that it cannot be either described or explained as if it were (as Durkheim thought) a ‘thing’. In the first place, praxis cannot be described ‘objectively’, since exhaustive operational definitions of the procedures used to classify it are beyond our reach; and secondly, meaningful conduct has to be explained not in terms of its causes, but in terms of the agent’s intentions, motives, reasons, grounds, etc. Most important of all, once we abandon positivism the very distinction between describing and explaining ceases to be hard and fast, since the two processes are combined in the single act of interpretation. Since this act has such a central place, I propose to call the alternative paradigm ‘interpretative’.

The various approaches which can be subsumed under this heading differ mainly in their view of the act of ‘interpretation’. A full history of this concept would be out of place here, and in any case beyond my means, but its ancestry is a long one. Aristotle’s ‘dialectical’ (as opposed to ‘physical’) explanation implies a version of it; so too does the nineteenth-century German distinction between Geisteswissenschalten and Naturwissenschaften. ‘Historical’ explanation,¹ and Dilthey’s hermeneutics, are other versions, along with Husserl’s phenomenology, Weber’s Verstehen, and Schutz’s synthesis of the two. The Chicago-based school of Symbolic Interactionism, like English analytic philosophy, also treats in depth the problems of interpretation, and these together with Schutz’s social phenomenology all influenced ethnomethodology, with its concept of ‘accounting’ for behaviour.² Freudian psychoanalysis combines interpretative methods with some of the tenets of positivism - a paradox which, I shall argue below, is not the result merely of torn loyalties, but of a unique way of looking at people. Apart from psychoanalysts, our list is made up mainly of philosophers and sociologists; with the exception of phenomenologists such as Merleau-Ponty or Jaspers, psychologists have habitually treated interpretative approaches with disdain (though there are currently signs of a shift in attitudes).

My catalogue makes ‘interpretation’ sound a very specialized, not to say academic, activity; but this impression is seriously misleading. For interpretative approaches, in contrast to positivism, start from very basic forms of understanding - judgements we make about each other from the moment when, as children, we grasp the difference between people and. things. Interpretation is something we do all the time, even if (like M. Jourdain speaking prose) we do not realize it; so it is easier for a Laing or a Goffman to gain a popular readership than it is for a Kallmann or a Kraepelin, because their conceptual world is at root a familiar one.

Some interpretative approaches, however, stay closer to common sense than others, and along this continuum lies an important difference in outlook, which I will use as a basis for organizing the material in this section. Ethnomethodology, symbolic interactionism, phenomenology and analytic philosophy all tend to give accounts of what a person is up to in terms of what he - or his fellows - thinks he is up to; some of these writers even deny the validity of any other terms. On the other hand, both Marxism and psychoanalysis argue the need for what Habermas calls ‘depth hermeneutics’ : interpretations which actively criticize and transcend people’s own understanding of them-selves. We can see already how different interests will, as Habermas describes, inform these two approaches to interpretation; the former being concerned to improve understanding of another person’s point of view (‘practical’ interest), the latter to challenge their assumptions and thereby increase their degree of autonomy (‘emancipatory’ interest).

As applied to the problems of psychiatry, the first of these approaches may be termed ‘normalizing’;³ it starts from the assumption that common sense (or some elaborated version of it) is capable of doing justice not only to ordinary behaviour, but also to that of mental patients. (‘Anti-psychiatry’ usually incorporates some such point of view.) We will consider this position in detail first.

I have organized the material under this heading according to the particular factor by which it seeks to make behaviour intelligible. Common-sense understanding- generally tries to make sense of actions by invoking three sorts of factors: their context, their purpose, and the code of conventions that structures them; and this three-way division provides a useful set of pegs on which to hang the various ‘normalizing’ approaches. Of course, it will become apparent as we go along that every interpretation necessarily invokes all three factors; the differences are therefore merely a matter of emphasis.

(i) Contextual approaches. Stripped of its context - as it usually is in positivist psychiatry - no behaviour makes sense; many attempts to ‘normalize’ so-called mental illness have therefore sought to bring back this vital information by considering the patient’s family, the wider social situation, and the medical encounter itself. Such studies implicitly claim that anybody experiencing such situations might react in the same way. By this is meant not that the situation would probably cause illness in anybody (which is simply the ‘environmentalist’ version of positivism), but that in relation to its context, the so-called illness is in fact perfectly intelligible behaviour.

Family contexts have most often been studied in connection with schizophrenia; the psychoanalytic theory of neurosis is, of course, based on early childhood experiences, but in a rather more indirect way which we shall examine later. Of the studies of schizophrenics’ families, the well-known volume Sanity, Madness and the Family by Laing and Esterson⁴ is practically the only methodologically aware application of an interpretative approach; while there are numerous other studies, almost all are ‘interpretative by default’ - that is, they are full of subjective interpretations, but because they are so concerned to deny the fact, they never permit the reader to see who is making the interpretations and on what basis. To say this, however, is not to deny that Laing and Esterson drew heavily for their inspiration on the work of positivists such as the Palo Alto school (Bateson, Haley, Jackson, et al.), Lidz, and Wynne.⁵

In the Preface to the second edition of their book, Laing and Esterson state their claim: that ‘the experience and behaviour of schizohrenics is much more socially intelligible than has come to be supposed by most psychiatrists’. (It is, of course, irrelevant to object as Clare⁶ does that the authors are not entitled to talk about ‘schizophrenics’ if they do not believe in the reality of the illness - there is nothing to prevent us today talking about witches, even if we do not believe in the Devil.) The authors call their approach ‘existential phenomenology’, but it could perhaps better be described as a mixture of common-sense understanding and applied linguistic philosophy; their approach actually differs from the phenomenology of Jaspers, Binswanger, and the early Laing,⁷ in that they do not attribute to the patients a unique way of experiencing their situation. Rather, what warranted these patients’ withdrawn, suspicious and hostile behaviour was said to be the ‘untenable situation’ they occupied within their suffocating and oppressive families; if they seemed paranoid, it was because people really were out to get them. Any more subjective disturbance, of the kind previously characterized by Laing as ‘ontological insecurity’, revealed itself as something located not in the patient but in the ‘language-games’ played by the family as a whole. According to Pateman,⁸ ‘what these patients need is not a therapist but an epistemologist’; just as rationality, according to Wittgenstein, has its grounds not in the individual but in the social communication system as a whole, so too may its opposite, ‘thought disorder’. The patients are ‘mystified’, because they cannot learn within their family a logical way of talking, and thereby thinking, about their experience in the family (cf. also Habermas’ concept of ‘distorted communication’⁹).

Sanity, Madness and the Family does not provide a formal theory linking the observed features of family life to the schizophrenic’s ‘symptoms’; like a novel in this respect, it relies instead on the unformalizable interpretative skills of the intelligent and humane reader to bridge this gap. To have expected the authors to provide such a formal theory would be asking them to adopt a paradigm which they deliberately and advisedly rejected. In addition, provided one accepts the authors’ account of their sampling method, it is misguided to complain, as does Wing,¹⁰ of the lack of ‘normal’ controls: for the hypothesis of social intelligibility is not to be confused with that of environmental causation. The other often-repeated criticism, that the patients in this book are not ‘really’ schizophrenic, rebounds damagingly on the critics - for if eleven out of eleven randomly chosen diagnoses of schizophrenia are incorrect, the overall rate of mis-diagnosis must be getting on for 100 per cent. And if psychiatrists fail to recognize the cases in this book as typical, it merely betrays how unsympathetically and superficially they are in the habit of perceiving their own patients.

Laing and Esterson, of course, simply remove the problem one stage further back: what is the context of this crazy family situation itself? What emotional needs do these patterns of distorted communication serve, and how do they fit into the culture as a whole?

The emotional dynamics of ‘distorted’ family relationships have been studied largely in psychoanalytic terms, e.g. by Esterson¹¹ (who investigated in depth one of the cases from Sanity, Madness and the Family), Mannoni¹² and Stierlin;¹³ the projection, expelling, scapegoating and double-binding that occur in the ‘schizophrenic’ family are seen as supporting the psychological defences of individual family members - to such an extent that the consequences of abolishing them can be as disastrous as the original problem.

Such accounts, of course, tend to reduce family disorders back into the language of individual psychology, while what is much more important - and much less well understood is the relationship between both family and individual dynamics and society as a whole. Although Sanity, Madness and the Family went as far as was logically necessary to achieve its stated aim, it came as a disappointment to many readers that Laing never went beyond the level of the family in his attempt to relate ‘madness’ to its social context - or even when he did, that he expressed his diagnosis in vague and ahistorical terms (see Deleuze and Guattari¹⁴ and Jacoby¹⁵). Joan Busfield¹⁶ has made a solitary attempt to understand the ‘double-binds’ experienced by the adolescent in family life, in terms of contradictory values built into our society itself; otherwise, the only relevant writings seem to be those of the ‘Freudian Marxists’ discussed in my next section.

Study of the family context, however, may mislead not only by suggesting that the family exists in a vacuum, but also by ignoring the many other institutions which comprise a person’s social context. The assumption that only home life has emotional significance does, in fact, reflect a sort of half-truth: for, as Joel Kovel points out in Chapter 3, the modern family has had to carry the full burden of emotional involvements withdrawn from the community, as the latter fractures and falls apart under the impact of the economic forces of advanced capitalism. Certainly, most people experience only their ‘private’ lives in human terms; but psychotherapists who, in their preoccupation with the family, also forget the maddening aspects of the wider society are merely colluding with those economic forces and producing ideology, not science.

Most studies of ‘mental illness’ in its wider social context come from the environmentalist offshoot of positivism; like the corresponding family studies, they are interpretative only by default. Worse still, they often rely on crude epidemiological data such as hospital admission rates; as we saw above, such information reflects people’s mental states very indirectly, if at all. This means that most cross-cultural statistics are virtually useless, because such factors as diagnostic practice and the availability of psychiatric services vary widely between cultures. The same considerations apply to studies of the relationship between mental illness and social class, since these are virtually cross-cultural studies too; although most researchers report a strong association between poverty and (e.g.) hospital admission rates, it is very hard to know what this association means. Do these statistics tell us something about the problems which unemployed or overworked, underpaid, poorly housed and ill-nourished people actually have, or about the way society processes their complaints? Most researchers interpret the connection between economic hardship and mental illness in terms of an intervening variable called ‘stress’, but this term adds little to the explanation. If ‘stress’ is inherent in certain situations, then it is a redundant concept; while if (as is much more likely) it is a function of the meaning which the situations hold for their occupants, then we need a theory of ‘subjective meanings’, and this is precisely what no positivist psychology can offer. Postulating variables like ‘stress’ is no substitute for a properly interpretative account of the way people’s situations, as they perceive them, give them grounds for their conduct. At the level of class or nation, such studies are rare indeed: I would mention here Franz Fanon’s¹⁷ account of the vicissitudes of colonial life, and Sennett and Cobb’s study¹⁸ The Hidden Injuries of Class, which focuses on the American blue-collar worker’s situation. The ‘Freudian Marxists’ discussed later also offer their own distinctive analysis of the relations between culture and psychopathology.

Understanding people’s ‘symptoms’ in terms of their social situation is perhaps easier when we define the latter in more specific terms. Here I am returning to the idea put forward in Section I, that ‘mental illness’ is an intelligible response to conflict between people’s needs and the demands (or constraints) placed on them by their particular social roles. Although in positivistic studies of the social environment, both ‘cause’ and ‘effect’ and the link between them are defined in such a way as to conceal any such intelligibility, it can be instructive to take as a starting-point the positivists’ discovery that certain roles are associated with particular patterns of ‘symptomatology’. I shall consider four examples of this.

The first case is one we have already considered under the heading of family studies. The diagnosis of ‘schizophrenia’ tends often to be made in adolescence, i.e. during the transition from juvenile to adult roles; as we saw above, Laing and Esterson sought to demonstrate the intelligibility of such ‘symptoms’ when seen against a background of contradictory family demands, and these demands were in turn interpreted by Busfield as reflecting contradictions endemic to the adolescent’s role. But although the young schizophrenic became the cause célèbre of anti-psychiatry, I believe that this was to the detriment of the movement as a whole: the depressed housewife, the senile psychotic, and the maladjusted schoolchild do not lend themselves so readily to the part of culture-hero - nor, in general, can they afford the fees of an existential psychoanalyst - but they are far more representative of psychiatric cases. Moreover, as I shall try to indicate briefly below, they illustrate somewhat more graphically the relationship between roles and ‘symptoms’.

To start with, the high incidence of depression among housewives is a ‘symptom’ that the women’s movement has been quick to decode. Brown and Harris¹⁹ have estimated its prevalence in a London sample of working-class women at 23 per cent, and have shown that having small children, no outside employment, no stable partner, and poor housing, can render it endemic. From the feminist point of view, it is somewhat perverse to speak of ‘illness’ at all here: one might more readily imagine that there was something odd about women who were not made miserable by such burdens. This brings out very clearly the fact that deciding what constitutes a ‘warrantable’ response to one’s situation is largely and unavoidably a political decision. The work of Brown and Harris, therefore, goes a long way towards demolishing its own positivist presuppositions.

One group without a ‘liberation movement’ to voice their predicament is the elderly, who again show a characteristic pattern of mental illness; nevertheless, Robert Kastenbaum²⁰ has maintained a solitary but determined argument that the indignities and deprivations of the role which our society allocates to its members, as soon as they cease to be useful to the labour-market, are quite sufficient to warrant the anger, despair and confusion which psychiatrists routinely diagnose as ‘senile psychosis’ and so on.

Children, of course, form the other obvious group which lacks organized representation of their political interests. As I have tried to show in an earlier paper,²¹ child psychology in general tends to be heavily biased against the child; more specifically, Conrad (Chapter 3) shows how the currently fashionable diagnosis of ‘hyperkinesis’ reifies and invalidates the rebellious actions whereby schoolchildren express their boredom and frustration with their allotted roles. When hyperkinesis becomes an ‘epidemic’, as it has in the U.S.A., we should perhaps cease talking about maladjusted children, and think instead about the maladjustment of their schools and families.

Other examples could be added to this list. Maucorps²² discussed the ‘social vacuum’ responsible for many apparently bizarre symptoms in demoralized and depopulated rural areas. Haugsgjerd (Chapter 7) considers the consequences in psychological terms of the changed way of life which is forced on sectors of the Norwegian population by ‘economic forces’ - that is, the interests of capital. Jules Henry,²³ in Culture against Man, anticipated many of the analyses I have mentioned of Western family life, schooling and ageing. Despite all these examples, however, such work is yet in its infancy: I have not mentioned at all the vicissitudes of two important roles - being employed, and being unemployed.

Such research is unlikely to develop under the aegis of orthodox psychiatry, which quite literally has no use for it - since its mandate is not to change society, but individuals. True, it was a constant theme of Freud’s writings that mental illness arises from the conflicts and tensions of social life; but while he argued that the constraints giving rise to repression and neurosis were intrinsic to civilization itself, the view which we represent in this book sees these constraints (e.g. family and work roles) as social conventions, tied up with the maintenance of a particular economic system, and changeable if - and only if - that system can be changed.

Some orthodox psychiatrists, even though they may reject psychoanalysis, also see mental illness as arising from the ‘stress’ of living without certain basic human needs and comforts. Like Freud, however, these environmentalists fail to see that human needs are (to some extent, at least) a social construction: very few ‘life-events’ (to use the jargon of this school) have a fixed, determinate effect - their impact depends on the meaning ascribed to them, and this meaning is a social variable. Thus, the reason why a situation which is today experienced as oppressive may not have been so regarded a hundred years ago is that we have different expectations of life; as Kovel reminds us, capitalism itself helped to nurture precisely those needs for ‘self-realization’ and ‘autonomy’ which at the same time it thwarts. Women today would perhaps not find their situations so depressing if they had not come to regard themselves as human beings: old people would not miss love if they had not once been obliged to attach such a value to it. The ‘human nature’ being thwarted is largely second nature, so it becomes impossible to describe one pole of the contradictions as ‘inner’ and the other as ‘outer’.

The psychiatric context. As well as showing how psychiatric disturbance may arise as an intelligible response to social conditions, the ‘normalizing’ approach also has a lot to say about society’s response to the disturbance - that is, about the institution of psychiatry itself. Although this work has mostly centred on the tangible, bricks-and-mortar institution of the mental hospital, it is important to remember that the concept includes something wider - the practice of psychiatry, whether or not it takes place in a mental hospital; and to the way psychiatric ideas are incorporated within ‘common sense’ itself. For as the mental hospitals are phased out, more and more treatment takes place in the doctor’s surgery and the general hospital; but the mental patient is still just as effectively incarcerated within his role. Moreover, this role is internalized within the patient’s own thinking and that of the people around him or her, and it guides everybody’s self-interpretations, whether or not they ever become patients.

Nevertheless, it is within the mental hospital situation that the medical encounter can be most closely examined. The classic studies here are those of Erving Goffman; drawing on the sociological tradition of Symbolic Interactionism, Goffman applied the concepts of ‘total institution’, ‘degradation ceremony’, and ‘moral career’ to the asylum situation,²⁴ and described in grim and compelling detail the cognitive, rather than simply physical, pressures which kept the patient in their place. (It was these pressures which Rosenhahn’s²⁵ volunteers experienced at first hand.) Of course, it was known even to positivists (e.g. Wing and Brown²⁶) that the traditional mental hospital was not a particularly good place to get ‘better’ in.

The institution of psychiatry was also studied, in a way that transcended the hospital situation itself, within the framework of ‘labelling theory’. Scheff²⁷ argued, following the work of Balint²⁸ and others on medicine’s ‘normal cases’, that psychiatric illnesses were socially constructed, just as physical ones were; they represented schemas according to which people organized and interpreted their own and other people’s ‘sick’ behaviour. Scheff claimed, moreover, that the states supposedly characteristic of ‘mental illness’ were actually very widespread in the normal population (cf. Pasamanick²⁹); if a person was not labelled sick, these states were overlooked, but if he was, they provided further ‘proof’ that the diagnosis was correct. Thus, the diagnosis of mental illness represented a point of no return; as Lemert³⁰ and Becker³¹ had argued in their general theories of deviance, society provided a kind of positive feedback which resulted in ‘deviance amplification’, as the patient internalized other people’s judgements of him. (‘Deviance amplification’ in the opposite direction is illustrated in the famous study by Rosenthal and Jacobson,³² who found that children’s cognitive performance was improved by telling their teachers that they were ‘bright’.) The work of labelling theorists added depth to Goffman’s concept of mental illness as a ‘career’, which many psychiatrists must have taken as just a bad joke.

These studies of the mental patient’s role may help to make intelligible many aspects of his or her conduct after being diagnosed; but they do not, of course, leave us any wiser about the psychiatrist’s conduct. For Goffman, psychiatry is simply there; as we saw in Section I, Szasz’s attempts to explain what it is doing there lack cogency, and the theory which this book sets out to promote (see Chapters 2 and 4) is still very much in its infancy. But what I hope has been learned in this section is that what one thinks psychiatrists are up to depends crucially on what one thinks their patients are up to; and the latter question cannot be answered without taking an essentially political stand on what constitutes a ‘reasonable’ response to a social situation. Thus, no amount of purely empirical data will enable us to decide whether to take sides with orthodox psychiatry or its critics; and this, as Hesse³³ argues - following Habermas - may be an unavoidable feature of any social theory.

(ii) Redefinition of goals. The second tactic in the ‘normalizing’ approach is to ascribe an unrecognized purpose to the behaviour in question. Of course, this goes hand in hand with the presentation of the context of action (as well as the assumption of a code), so that no studies can be said to fall into this category alone; but two key ideas in anti-psychiatry centre chiefly on the question of purposes. The first is the idea that the ‘sick’ behaviour is a form of protest; the second, that it is a kind of self-cure.

The ‘symptom as protest’ school finds its strongest support in Goffman’s studies of asylum politics. ‘If you rob people of all customary means of expressing anger and alienation and put them in a situation where they have never had better reason for having these emotions, the natural recourse will be to seize on what remains - situational improprieties.’³⁴ Outside the asylum, this view is implicit in (for example) Laing and Esterson’s family studies; in the feminist interpretation of housewives’ depressions; and in Conrad’s views on ‘hyperkinetic’ schoolchildren. The idea gains a deeper perspective within the thinking of Sartre and Marcuse, in which the ‘dregs’ of the capitalist system - the deviants and drop-outs - constitute, in Hegelian terms, the ‘determinate negation’ of that system; hence the idea mentioned above, of ‘the schizophrenic as culture-hero’. According to David Cooper,³⁵ ‘all delusions are political declarations and all mad-men are political dissidents’.

Perhaps because it was so inadequately formulated, however, this idea became the weakest plank in the anti-psychiatric platform. It was in fact the Left who attacked it most strongly: Sedgwick,³⁶ Mitchell,³⁷ Gleiss³⁸ and Jacoby,³⁹ all objected that this ‘romanticizing’ of society’s victims distracted attention from the real need for coherent, rational political action - and didn’t help them, either. Unfortunately, these critics reverted to an astonishingly positivistic ‘us I them’ model of mental illness; the real point, surely, is not that psychiatric problems lack political significance, but that they are not effective forms of social action.

For the ‘symptom as protest’ view glosses over the differences between the kinds of behaviour that psychiatrists deal with, and conscious, socially intelligible and potentially effective forms of protest. In Morel’s original case of ‘dementia praecox’,⁴⁰ there was perhaps not much difference - the boy’s chief ‘symptom’, apparently, consisted of hatred of his father! - but most people, after all, refer themselves for treatment, and do so precisely because they do not understand what is going on. Yet rather than revert to the orthodox view of illness, as the Left has been inclined to do, what seems called for is a concept of interpretation more subtle than the common-sensical ‘normalizing’ approach can offer: and this will be the topic of our final section.

The concept of ‘symptom as self-cure’ has similar shortcomings. Laing’s suggestion that ‘catatonic’ withdrawal represents a form of meditation has an appealing simplicity; and his idea of psychosis as a ‘voyage into inner space’, ‘an attempt to overcome our normal state of appalling alienation’, is a popular one by now. Indeed, the basic idea is not all that original: Freud himself spoke of psychotic patients ‘attempting to cure themselves by becoming hysterical’,⁴¹ and the concept of controlled therapeutic regression is well established in psychoanalysis today. The key word, however, in these quotations is ‘attempt’; Freud certainly did not imagine they would succeed, and Laing offers little evidence that they do (though to be fair, the societal reaction to psychosis is so destructive that it is hard to imagine the idea ever receiving a proper test). If such symptoms are attempts at self-cure, they are neither deliberate nor effective ones; just as the psychotic ‘protest’ plays straight into the hands of the oppressors by inviting its own invalidation, so the ‘cure’ perpetuates the state of alienation by merely expressing it in a more colourful and intransigent form.

(iii) The concept of ‘code’. In appealing to a special ‘code’, the normalizing approach is claiming that the problem behaviour does not have the superficial meaning which we ascribe to it, but instead belongs within a framework which accords it quite a different meaning. Such a framework is really a form of symbolic environment, which merely shows the weakness of my classification: for this category is really an extension of the contextual approach, and the studies of schizophrenics’ families as well as Goffman’s asylum ethnography belong here as well.

The subcultural theory of deviance put forward by Becker⁴² and others suggests that much conduct is regarded as ‘sick’ simply because ‘straight’ observers do not understand the meaning that it has in its own subculture. This idea has obvious relevance in the case of diagnoses of mental illness made by professionals to whom the patient’s culture is an alien one. For example, it has been found that West Indian immigrants are commonly seen as ‘aggressive’ by English welfare officials on the basis of their intonation patterns, which have been superimposed on to English from African languages. Among the same group, the relatively frequent diagnoses of ‘psychosis’ may be due in part to English psychiatrists not appreciating that trance states, spirit possession and the like can be part of a normal life-style.

Not many problems in psychiatry, however, are likely to be amenable to solutions of this sort, where the code invoked is of a recognized and public kind. The ‘decoding’ of most symptoms calls for a rather different kind of methodology, which goes beyond the limits of common sense; the psychoanalytic approach, which alone has attempted this, invokes a ‘code’ which we fail to recognize, not because it belongs to a foreign culture, but because it is a part of our own mentality too close for comfort. This, however, is to anticipate our next section.

Having outlined the essentials of the ‘normalizing’ approach, we may now ask: how does it compare with positivism, and how successfully does it deal with the phenomena of psychiatry?

Clearly, the positivist ideals of ‘objectivity’ are not met by this form of interpretation, and indeed its exponents would claim that they never could be in any human or scientific ‘human science’. Its concepts are not operationally definable, its observations cannot be literally reconstructed, and the observer cannot pretend not to influence the situation he observes. The positivist will deplore the fact that such work draws so heavily on taken-for-granted resources of judgement and interpretation though, as we have seen, his paradigm also leans on these - while no one is looking. In addition, interpretative theories do not take the form of rigid laws of causation, but specify instead configurations of context, motive and code which make actions intelligible; in few situations could anything like a prediction be made, for the uniqueness of persons and situations is implicit in this approach. Hence, such theories do not generate an easily reproducible technology of behavioural research and intervention of the kind that positivism - falsely - promised. Moreover, to the extent that they identify certain issues as moral rather than factual questions (e.g. whether the depressed housewife is really ‘ill’ ), they do not claim to be value-free.

To the positivist, then (in Haugsgjerd’s words), such approaches represent ‘a tide of obscurantism’, while to psychotherapists and the like, terms such as ‘hermeneutic’ suggest ‘a promise that one day their kind of work will get its fair share of scientific glory’. The appeal of such theories, once the vain hope of emulating ‘hard’ science is renounced, lies in their refusal to deny the differences which we all know to exist between people and things. In consequence of this refusal, reification and individualism - the inevitable accompaniments of positivism - can be avoided: the pseudo-technological, ideological interest becomes replaced by the ‘humanistic’ one of understanding other people’s points of view. Moreover, the fact that only qualified psychiatrists could apply positivist theories and treatments gave the latter an important role in maintaining psychiatry’s exclusive hold over its territory: the ‘normalizing’ approach, by contrast, invites all and sundry to share in the task of understanding and reconciliation. As a professional ideology, then, it is a non-starter.

Yet - as Freud and Marx both held - it is possible to pay altogether too much regard to people’s self-understanding; what if they are mistaken? The attempt to rely on common-sense interpretations starts to look inadvisable when we question the authority of common sense itself. We would like to be able to emulate the positivist’s steely, sceptical gaze - even if we do not want to stray so far from conventional wisdom as to abandon, like him, the concept of interpretation itself. I shall suggest below that the shortcomings of the ‘normalizing’ approach to mental illness stem from the same fault as the shortcomings of common-sense interpretation applied to everyday life itself: namely, the inability to comprehend compulsive action, or alienated states of being.

The basic problem for any normalizing account of mental illness is a very obvious one: if the behaviour is really intelligible in common sense terms, why was it regarded as a psychiatric problem in the first place? There are, in fact, several good answers to this question. In the first place, the decision may not have been made by an open-minded, representative sample of lay people; those responsible for it may simply have wished to invalidate the behaviour in question (‘Darling, I really think you’ re ill!’), or they may not have had access to all the relevant information. The Context may have been overlooked, both because traditional psychiatry lacks any way of doing justice to it, and because the people who comprise it may not wish to be implicated. The purposes may have been ignored because of their very nature, especially if they involve an element of protest; and the codes may be overlooked because of one group’s ignorance and contempt towards another.

All this might explain why some instances of perfectly coherent behaviour get treated as symptoms of an illness; but to explain all psychiatric diagnoses in such terms seems a tall story. I think it has to be admitted that whatever sense is lent by a consideration of contexts, purposes and codes, there remains a residue in most ‘mental illnesses’ which is refractory to all ordinary procedures of understanding and empathy. After all, the ‘so what’ problem also applies to normalizing accounts, since by no means all who are placed in the situations we have described become ‘mentally ill’; rather than leave the residue to be explained by vague and immutable ‘constitutional factors’, it would be preferable to see whether the concept of interpretation itself could be modified to give it greater explanatory power.

The normalizing approach exaggerates the extent to which rational free-will operates in psychiatric conditions; they are not just, as Szasz would have it, ‘problems of living’, but a breakdown of the problem-solving ability itself - one loses one’s grip on oneself and one’s grasp of what is going on.⁴³ Perhaps the reason why anti-psychiatrists strained so hard to deny this was that they shared with positivist psychiatry - in which tradition many of them had grown up - the belief that in such conditions, the interpretative approach breaks down altogether, and one is forced back to the ‘faulty-machine’ model.

But this simple opposition of free-will and determinism will not do. To avoid it, we must transcend both common sense and positivism; and this is precisely the goal of those approaches I have labelled ‘depth hermeneutics’, the chief of which is psychoanalysis. What is required is a way of accounting for experience and behaviour in terms of meanings, but not necessarily ones which are consciously appreciated either by the agent or his fellows; and this requires a radical revision both of our conception of the person, and of the methods of the human sciences. What has to be replaced is not only the positivist myth of man as machine, but also what Marcuse⁴⁴ calls ‘the myth of autonomous man’, which interpretative theorists are equally prone to. In place of the unity of the self which is assumed both by phenomenology and common sense, we must substitute Freud’s conception of man as fragmented, self-contradictory, and alienated from his own experience. Only when this is done can the true meaningfulness of ‘mad’ behaviour become apparent - and, at the same time, the true madness of behaviour which common sense takes to be ‘sane’.

In many ways the psychoanalytic approach to mental illness seems thoroughly positivistic and out of place in this chapter. Freud’s basic model of the mind owed a lot to hydraulics; neurotic symptoms represented ‘the return of the repressed’ - that is, emotional energy which had to go somewhere after being ‘blocked’ by various obstacles. Superficially, too, psychoanalysis seems to generate aetiological hypotheses which relate childhood events to adult disorders in a thoroughly deterministic way. This positivistic image is entirely consistent with Freud’s declared aim, which was to extend the domination of science into the last territory remaining unconquered - the human mind.

In the course of this project, however, Freud was driven ineluctably far beyond the limits of scientific orthodoxy. He himself was aware of his debt to the classical and European literary heritage, and voiced increasing uncertainty about whether his work should be classified as ‘arts’ or ‘science’; this duality is also reflected among his followers. In America, where psychoanalysis was absorbed into the medical profession, a strictly positivistic reading was cultivated; analysts thus gained elite status and earning power, at the price of renouncing any claim to be more than emotional technologists (see Turkle and Kovel, Chapters 5 and 2). While the same marriage was attempted in Europe, it was never really consummated; it was the positivists who objected first to their new bedfellows, and only after being cold-shouldered by the medical and academic establishment did psychoanalysis seriously start to question its allegiances. Eysenck⁴⁵ was loud among those who objected to the alliance: psychoanalysis, he roundly declared, was nothing more than a revamped version of nineteenth-century verstehende psychology, and therefore could claim nothing in common with ‘real’ science. Subsequently, the English analyst Charles Rycroft⁴⁶ voiced agreement from the other side, claiming that the concern of psychoanalysis was with discovering ‘meanings’ and not ‘causes’; since, therefore, psychoanalysis was not trying to be a science, charges that it was unscientific would (according to Rycroft) bounce off harmlessly. (We may note, for later reference, that both authors unhesitatingly accepted the equation of ‘science’ with ‘positivism’). In Europe, where philosophers, artists and phenomenologists had given quarter to psychoanalysis, an interpretative reading of Freud had long had currency; since the 1930s, Lacan⁴⁷ had sought to wrest psychoanalysis away from positivism, a cause which Merleau-Ponty,⁴⁸ Lorenzer⁴⁹ and Ricoeur⁵⁰ also espoused. In the U.S.A., only dissident analysts such as Szasz⁵¹ and Shafer⁵² supported such a reading.

What are the grounds for calling psychoanalysis an ‘interpretative’ discipline? Chiefly, perhaps, because ‘interpretations’ are what the patient pays to get: the analyst decodes the ‘latent meaning’ of dreams, verbal slips, and symptoms. For Lacan, Lorenzer and Rycroft, the paradigm of interpretation was the reading of a text, and thus involved structural linguistics, semantics, rhetoric and poetics. Lacan’s famous dictum ‘The unconscious is structured like a language’ is perhaps the best-known expression of his view: in this statement we can discern the influence of French structuralist anthropology (Levi-Strauss, Mauss), which saw society itself as a kind of ‘text’ to be decoded.

What, however, is the force of the phrase ‘like a language’? In Lacan (as in structuralism generally), the analogy of language becomes a Procrustean bed, on which it is impossible to comfortably accommodate the full range of human action. The communication of a message is only one of the motives which can underlie an action; and while other motives may, like verbal meanings, exist only by virtue of a socially-maintained ‘vocabulary’, this does not justify cutting down the traditional concept of interpretation to a semantic sense alone. In fact, I propose to illustrate that psychoanalysis actually invokes all three of the factors which we saw utilized by the ‘normalizing’ approach: contexts, motives, and codes.

Like the layman, Freud first seeks to make actions intelligible in terms of their context; but what is distinctive about psychoanalysis is that it considers not the situation as it ‘really’ is, but as it is construed by the agent. It thus supplies the missing link in the theories we examined earlier, which resemble positivist accounts in that they tend to define the social context ‘from outside’. Moreover, Freud is not simply invoking the situation which the agent consciously believes himself to be in, but perceptions which are for the most part unconscious. Thus psychoanalysis interprets what situations mean to the patient, and this is the key to its interpretation of the patient’s behaviour; in this respect it resembles phenomenology, but with the crucial difference that the agent is not regarded as necessarily in touch with his own perceptions and intentions.

The notion of ‘unconscious experience’ (or phantasy, as the Kleinians came to refer to it) is at first sight a paradoxical one - it was thrown aside, for example, by Laing;⁵³ but I believe that resolving the paradox is merely a problem of distinguishing levels of awareness (cf. Russell’s concept of ‘logical types’). ‘Phantasy’, for a start, refers not to a sort of sideshow accompanying perceptions (though such sideshows can occur, and are very instructive to observe), but to their structure - the schemas or scenarios to which reality is unwittingly assimilated. Therefore, to call a phantasy ‘unconscious’ is merely to say that a person is unaware of the way he is construing situations, and moreover that it is so important to him emotionally to maintain this way of construing them that he will actively resist becoming aware of it.

A good illustration of this is ‘transference’ - the patient’s tendency to treat the analyst as a figure from the past; the easiest way to describe that phenomenon is to say that the patient is ‘unconsciously experiencing’ the analytic session as (for example) a feeding relationship between mother and baby. As such identifications are made in terms of past situations - however misperceived or modified - the present is always an echo of the past; thus, for Freud, the context of behaviour is never simply the present situation, but a continuum of experience stretching back to the beginning of life.

This approach obviously has great potential for making sense of apparently inappropriate behaviour, and it provides one of the chief tactics used by psychoanalysis to ‘de-reify’ so-called symptoms. A man who savagely murders prostitutes may be acting quite intelligibly, in so far as he may see in the prostitutes the mother who ‘betrayed’ him in her sexual passion for his father; that he should see them thus is itself quite unintelligible without recourse to further explanation, but to see such beliefs as a sign of abnormality is to overlook the fact that the same misrecognition provides the emotional bedrock of most marriages.

Psychoanalysis also interprets behaviour by reference to its unconscious motives - indeed, this is regarded by most as its chief activity. I have placed ‘redefinition of contexts’ first, however, since this aspect of psychoanalysis is often overlooked - but the unconscious is not simply made up of motives. In fact, to redefine the beliefs informing an action is already to redefine its motive, so that one cannot properly distinguish the theory of unconscious phantasy from the theory of unconscious motivation.

Besides the notorious concept of infantile sexuality, a more fundamental motive which does not enter into the everyday vocabulary is what Freud called the ‘pleasure principle’, which gives rise to the phenomenon of defences: a person may construct an apparently pointless way of life, with the hidden aim of avoiding certain conflicts, or keeping them out of consciousness - because the ‘pleasure principle’ regards out of sight as out of mind, and out of mind as out of existence. One may, for example, divide the world into good and bad ‘part-objects’, all for the sake of avoiding the conflict of ‘good’ and ‘bad’ feelings about the same person; or one may bind oneself to another who expresses one’s own unwanted feelings (the defence of ‘projection’). All such defences represent an attempt to restructure the world in a more comfortable pattern - or as Freud termed it ‘hallucinatory wish-fulfilment’. It is this motive which probably accounts for the phenomenon of unconscious phantasy discussed above.

The extent to which psychoanalysis invokes an unfamiliar code of meanings is rather more controversial. On the face of it, translation is the most obvious form of interpretation; the analogy was frequently used by Freud himself. Szasz and Lorenzer both spoke of the ‘protolanguage’ of symptoms, while according to Lacan,⁵⁴ the unconscious employs the literary devices of metaphor, metonymy, etc. However, as Coulter⁵⁵ and Ricoeur⁵⁶ have argued convincingly from different standpoints, it is dangerous to take this analogy with language too literally, for philosophical problems arise about the notion of ‘unconscious language’ which are even more awkward than those surrounding the concepts of unconscious experience and motivation. It will not do to represent Freud as simply offering a ‘dictionary’ of symbolic meanings; in psychoanalysis, the meaning of any symbol depends crucially on the context of its occurrence - so we are certainly not dealing with a shared language. In any case, a language is more than a set of symbols - it is also a way of using them; and nobody has succeeded in defining a distinctive syntax of the unconscious, or in explaining how such a structure could be acquired. (It could certainly not be learned, like ordinary language, by experience; how could one be ‘corrected’ in a task one is not even aware of performing?) Although useful as a metaphor, then, the idea of ‘the language of the unconscious’ is far more problematic than its devotees seem to appreciate.

The above account of psychoanalytic interpretation is, of course, incomplete and grossly oversimplified; nevertheless, it serves to indicate the kinds of praxis which analysts seek to understand, and the means they use for doing so. Actions involving unconscious experience or motivation cannot be encompassed either by common-sense understanding, or by mechanistic explanation; they are not under the control of the agent, but they are nevertheless meaningful. As long as the agent remains unaware of their meaning, they remain compulsive actions - that is, the agent does not know what he is doing, but he cannot be satisfied until he has done it. The psychoanalytic concept of ‘compulsion’ provides a way in which the term ‘illness’ may still be used meaningfully even when no organic pathology exists. Since psychoanalysis is thus concerned to enlarge the agent’s sphere of freedom, Habermas characterized its interest as ‘emancipatory’, and took it as the model of critical social theory - theory which does not simply reproduce society’s illusions about itself.

On the face of it, then, psychoanalysis seems to remedy all the defects of the ‘normalizing’ approach to mental illness; it is capable of interpreting the ‘residue’ which common sense finds unintelligible, by the apparently simple device of redefining the meaning of situation and action in subjective terms which bring the two back into coherent relationship with each other. There are two snags, however, which prevent us from rushing into the proffered embrace of psychoanalysis. One is the methodological problem, that this newly flexible concept of interpretation seems capable of encompassing everything, and therefore of explaining nothing; the other is frankly a political one, that in suggesting that mental illness is a matter of the ways in which people construe their situations, psychoanalysis tends to neglect the need to change those situations themselves. I will discuss these two problems in turn.

Methodological problems. The main objections raised by positivists against psychoanalysis have been in terms of its excessive flexibility -that it is capable of explaining everything that might happen, and is therefore unfalsifiable and unscientific (see, e.g., Popper,⁵⁷ Eysenck,⁵⁸ Borger and Cioffi⁵⁹). The instances of ‘unfalsifiability’ put forward do not, however, reflect a very intelligent reading of psychoanalysis; for example, Freud’s hypothesis that either a harsh or a lax upbringing would produce a strong super-ego is not ‘unfalsifiable’, because it excludes upbringings which are neither harsh nor lax. Specific hypotheses in psychoanalysis are falsifiable; and while it may be true to say that there is no crucial observation in face of which a psychoanalyst would be obliged to abandon his general approach, this can also be said of any other scientific system (Cosin, Freeman and Freeman⁶⁰).

The complaint that psychoanalysis lacks explicit rules governing the naming of phenomena (‘correspondence rules’) is, however, a valid one, and this vagueness can in principle be exploited to protect the theory from falsification. In Section I, I argued for the view that correspondence rules in any human science are inevitably too complex to be made explicit (the ‘etcetera’ problem); though this implied that descriptions must always involve an element of subjective judgement, it nevertheless assumed that the judgements were constrained by tacit rules - those of ‘common sense’.

Psychoanalysis, however, is on dangerous ground here: it cannot afford to lean too heavily on common sense, because that is precisely what it is busy undermining. The interpretations which psychoanalysts make are often not ones which come naturally to the layman - if they were, one would not have to pay so high a price for them; indeed, the layman is often baffled and repelled by them.

What, then, can psychoanalysis lean on? The predicament is not unique, nor is it anything to be ashamed of; it is shared by any theory which seeks to uncover systematic distortions in human awareness (e.g. Marxist theories of ‘false consciousness’). The only solution lies in the concept of ‘immanent critique’: emancipatory theories (in Habermas’ sense) can only seek to undermine one part of received wisdom by appeal to another - they must always be grounded in common sense. This is only possible if common sense contains contradictions - so that the exposure of contradictions in the realm of the ‘obvious’ must be the essence of any ‘depth hermeneutics’.

This, in fact, is an apt description of Freud’s whole project. Unlike both the positivist and the mystic - who claim privileged access to a transcendent plane of ‘reality’, from which common sense appears as an out-and-out delusion - the psychoanalyst starts from the same criteria for interpretation as anybody else; as Shafer⁶¹ convincingly illustrates, psychoanalytic understanding fades off imperceptibly into the everyday variety. Thus, Freud can only speak about illusion and compulsion by taking for granted the existence of true perception and free-will; he may draw the dividing-line between the former and the latter in a different place from the rest of us, but he cannot - and does not seek to deny that the dividing-line exists at all. All this suggests, then, that the grounds on which a psychoanalytic interpretation is based must in the last resort be of a basically familiar sort, otherwise they are not grounds at all.

To sum up so far: the view that psychoanalysis is ‘interpretative’ does not (as Eysenck thought) remove it from the realm of science - for all human sciences must employ interpretation; nor does it (as Rycroft thought) provide a secure alternative, for psychoanalysts cannot, like theatre critics, appeal tout court to the obvious. I also wish to argue that to present psychoanalysis as purely interpretative is to miss the point of the whole exercise. As well as imputing meanings to behaviour, Freud wishes to provide a causal theory of how these meanings come into existence; moreover, this causal theory is essential to the credibility of the whole system, since I do not believe that without it there can be adequate grounds for making the less obvious sorts of psychoanalytic interpretation.

In a purely interpretative theory, motives and meanings are irreducible: once an action has been made conventionally intelligible, we do not seek to go beyond the account given to ask - why this motive? why this meaning? Freud, however, does precisely that, because he is concerned with people both as subjects and as objects -both acting, and being acted upon. Biology as well as history determines the particular kind of subjects which we can be: thus, in psychoanalysis the subject is ‘decentered’,⁶² and the ‘primacy of the cogito’⁶³ is abolished. This means that any attempt to eliminate causal explanations entirely from psychoanalysis turns it into something else: as Ricoeur insists,⁶⁴ ‘This mixed discourse [of force and meaning] is the raison d’être of psychoanalysis’. Freud did not simply start off with the wrong paradigm, and correct himself as he went on -for the tension between paradigms runs through all his work; nor is Habermas⁶⁵ really justified in saying that Freud ‘misunderstood himself’. Positivistic psychoanalysts may have over-estimated the causal content of psychoanalysis, but they were nevertheless right to think it was there.

This has an important bearing on the question of methodology. It is precisely his belief in certain basic causal mechanisms - the ‘pleasure principle’, for example - that gives the psychoanalyst grounds for making interpretations which common sense alone could never arrive at. But it is his interpretations that lead him to believe in the existence of these mechanisms: hence there is a subtle interlocking relationship between the two elements of psychoanalysis’ ‘mixed discourse’, which potentially - though not inevitably - leads to circularity.

I hope I have said enough to indicate that the question we started out with - is psychoanalysis scientific? - raises vertiginous questions both about the nature of psychoanalysis, and of science - ones which serious philosophers⁶⁶ are only just beginning to confront. Moreover, the question of whether or not psychoanalysis ‘works’ - which is often adduced as evidence one way or the other - has to my mind very little bearing on the matter: quite apart from the problems of defining a psychoanalytic ‘cure’, it by no means follows that if the therapy works, the theory must be true - or vice versa (see Fisher and Greenberg⁶⁷ for an intelligent review of research on this question). All we can conclude from this section is that the methodological problems of psychoanalysis have hardly begun to be formulated, let alone solved.

Psychoanalysis and social change. Does psychoanalysis help us to understand the relationship between social conditions and mental misery, or does it merely obscure it? As we have seen, psychoanalysis helps to fill the gap in ‘normalizing’ approaches, by showing how an individual’s perception of their situation may vary; but there is a very real danger that this perception will be seen as the source of the problem, rather than the situation itself. (A true story is told of a young psychoanalytic trainee who visited a woman under treatment for persecutory phantasies concerning rats: when the front door was opened, out jumped - a rat!) Psychoanalytic theory is easily used to adapt the individual to his social role, and in this guise it may become absorbed effortlessly into the repertoire of ‘welfare’ services. Analysis for the interesting rich; drugs or ECT for the boring remainder.

This, of course, is already to a large extent the way things are; but to accept it as an inevitable consequence of psychoanalytic theory is, I think, to read the situation back to front. Freudian theory is an account par excellence of the social intelligibility of mental illness; it is the brief implicit in the therapist’s own role which leads to the social factors being taken as constants and the individual ones as variables -for the therapist has no mandate to change society. Unfortunately, since psychoanalytic theory has from the beginning been in the hands of therapists, its social dimension has become submerged and obscured.

The task of extracting from psychoanalysis implications for social change is one which Freud himself mostly shied away from. Some Marxists in the 1920s and l930s were, however, anxious to do the job for him: hence arose the corpus of ‘Freudo-Marxism’, which includes Reich, Fenichel, Fromm, Horkheimer, Adorno and Marcuse (the last four being associated with the Frankfurt School of critical theory). Since the war, the same challenge has been taken up by many others, including Habermas, Michel Schneider, Deleuze and Guattari, and Lacan.

The early thrust of this work was to show that the institutions of capitalist society were maintained by the particular personality structure which normal child-rearing practices brought about, and thus how psychic economy and political economy supported each other. Normal personality structure was regarded as largely a system of compulsions: thus psychoanalysis was applied to the supposedly sane as well as the apparently sick, a process Freud had begun in ascribing character to neurotic mechanisms, and in his analyses of religion and mass psychology. Lacan has re-interpreted the same theme in structuralist terms by speaking of the unconscious as the vehicle of ideology - in particular, of patriarchal ideology (‘le loi-du-père’).

It is not possible to discuss this work in any detail here, but some answers may be found within it to the question posed above. On the one hand, Freudian theory has shown itself uniquely appropriate to the understanding of an irrational society, since it can account for the fact that people not only fail to recognize their own exploitation, but literally become addicted to it; a key concept here is ‘internalization’ of the source of exploitation. In this way psychoanalysis goes beyond the familiar limitations of the concept of ‘oppression’. On the other hand, though, Freud’s basic assumption was that civilization was inherently repressive, though I have argued - following Reich - that this is not the only way he could have interpreted his observations.⁶⁸ Furthermore, in its treatment of the family, psychoanalysis reifies the oedipal situation into a law of nature,⁶⁹ and even Lacan seems to end up by treating patriarchy similarly. Thus, although Jacoby⁷⁰ is shrill in his denunciation of ‘revisionists’ who tamper with Freud to make his conclusions more acceptable, it is hard to see how psychoanalytic theory can acquire a fully historical dimension without a great deal of revision. And to perform this task it is essential that the theory should cease to be the monopoly of those who maintain it for therapeutic purposes alone.

My purpose in this chapter has not been just to set up a marketstall of available theories of mental illness, but to bring out one central point: that no theory is value-free, each being tied to a definite commitment about social goals.

I have tried to show that psychiatry as an institution is committed to fundamentally conservative social goals, and it is this commitment rather than ‘scientific’ considerations that determines the choice of which theories it accepts, and which it rejects. From a different moral standpoint, I have argued that theory should be explicitly, though not exclusively, interpretative; that it must transcend common-sense understanding; and that it must incorporate social structure, not as a constant but as a variable.

The last point immediately calls in question what I have just said about psychiatry’s social role. To speak of psychiatry as an ‘institution’ leads, in the end, to reifying it; for institutions exist not in theories but in history, and they are created and maintained by the activities of people. It is perfectly possible, therefore, for people to transform or abolish them: the example of Basaglia and his colleagues (Chapter 6) shows that mental health workers can - albeit not without struggle - completely redefine their allegiance and their social function, without in the process losing their specialized province. Elsewhere in this book we will return to the question of how this transformation can be accomplished.

One last point about science. In Section I, I did not bother to question whether the positivist paradigm contained, in fact, an accurate version of the methodology of the natural sciences. In a sense, however, positivism is a straw man (though still a very influential one), since what it sets out to imitate never, in fact, existed: recent philosophy of the natural sciences⁷¹ has shown that observation always involves a process of interpretation, and that theories are never completely dictated by the ‘facts’. Moreover, both cybernetics and animal biology have suggested in recent years that the special concepts supposed to be necessary for understanding human beings are relevant to the ‘natural’ realm as well. Thus, it is no longer obvious today, as it was to the positivists and their opponents, what exactly is implied by asserting or denying that man is a part of nature, and understandable in the same ways as nature.

All these considerations suggest that the acute division which has grown up between ‘scientific’ and ‘humanist’, or ‘hard’ and ‘soft’ approaches is not based on any logical distinction; political differences, as we have seen, exist, but once these are brought out into the open, the basis for a sharp division between paradigms for the study of man and of nature will probably disappear. At that time, hopefully, the analysis I have put forward in this chapter will become out of date: but until the political component of our disagreements is brought out into the open, we will never go beyond the phoney synthesis of psychiatric ‘eclecticism’.