Mental Illness, Medicine, Malingering and Morality: The Myth of Mental Illness vs The Myth of Free Will

Thomas Szasz, Psychiatry: The Science of Lies New York: Syracuse University Press, 2008

The notion that psychiatric conditions, including schizophrenia, ADHD, depression, alcoholism and gambling addiction, are all illnesses ‘just like any other disease’ (i.e. just like smallpox, malaria or the flu) is obvious nonsense. 

If indeed these conditions are to be called ‘diseases’, which, of course, depends on how we define ‘disease’, they are clearly diseases very much unlike the infections of pathogens with which we usually associate the word ‘disease’. 

For this reason, I had long meant to read the work of Thomas Szasz, a psychiatrist whose famous (or perhaps infamous) paper, The Myth of Mental Illness (Szasz 1960), and book of the same title, questioned the concept of mental illness and, in the process, rocked the very foundations of psychiatry when first published in the 1960s. I was moreover, as the preceding two paragraphs would suggest, in principle open, even sympathetic, to what I understood to be its central thesis. 

Eventually, I got around to reading instead Psychiatry: The Science of Lies, a more recent, and hence, I not unreasonably imagined, more up-to-date, work of Szasz’s on the same topic.[1]

I found that Szasz does indeed marshal many powerful arguments against what is sometimes called the disease model’ of mental health

Unfortunately, however, the paradigm with which he proposes to replace this model, namely a moralistic one based on the notion of ‘malingering’ and the concept of free will, is even more problematic, and less scientific, than the disease model that he proposes to do away with.  

Physiological Basis of Illness 

For Szasz, mental illness is simply a metaphor that has come to be taken altogether too literally. 

Mental illness is a metaphorical disease; that, in other words, bodily illness stands in the same relation to mental illness as a defective television stands to an objectionable television programme. To be sure, the word ‘sick’ is often used metaphorically… but only when we call minds ‘sick’ do we systematically mistake metaphor for fact; and send a doctor to ‘cure’ the ‘illness’. It’s as if a television viewer were to send for a TV repairman because he disapproves of the programme he is watching” (Myth of Mental Illness: p11). 

But what is a disease? What we habitually refer to as diseases are actually quite diverse in aetiology. 

Perhaps the paradigmatic disease is an infection. Thus, modern medicine began with, and much of modern medicine is still based on, the so-called ‘Germ theory of disease’, which assumes that what we refer to as disease is caused by the effects of germs or ‘pathogens’ – i.e. microscopic parasites (e.g. bacteria, viruses), which inhabit and pass between human and animal hosts, causing the symptoms by which disease is diagnosed as part of their own life-cycle and evolutionary strategy.[2]

However, this model seemingly has little to offer psychiatry. 

Perhaps some mental illnesses are indeed caused by infections. 

Indeed, physicist-turned-anthropologist Gregory Cochran even controversially contends that homosexuality (which is not now considered by psychiatrists as a mental illness, despite its obviously biologically maladaptive effects – see below) may be caused by a virus

However, this is surely not true of the vast majority of what we term ‘mental illnesses’. 

However, not all physical diseases are caused by pathogens either. 

For example, developmental disorders and inherited conditions are also sometimes referred to as diseases, but these are caused by genes rather than germs. 

Likewise, cancer is sometimes referred to as a disease, and, while some cancers are indeed sometimes caused by an infection (for example, cervical cancer is usually caused by HPV, a sexually transmitted virus), many are not. 

What then do all these examples of ‘disease’ have in common and how, according to Szasz, do so-called mental illnesses differ conventional, bodily ailments? 

For Szasz, the key distinguishing factor is an identified underlying physiological cause for, or at least correlate of, the symptoms observed. Thus, Szasz writes: 

The traditional medical criterion for distinguishing the genuine from the facsimile – that is, real illness from malingering – was the presence of demonstrable change in bodily structure as revealed by means of clinical examination of the patient, laboratory tests on bodily fluids, or post-mortem study of the cadaver” (Myth of Mental Illness: p27) 

Thus, in all cases of what Szasz regards as ‘real’ disease, a real physiological correlate of some sort has been discovered, whether a microbe, a gene or a cancerous growth. 

In contrast, so-called mental illnesses were first identified, and named, purely on the basis of their symptomology, without any understanding of their underlying physiological cause. 

Of course, many diseases, including physical diseases, are, in practice, diagnosed by the symptoms they produce. A GP, for example, will typically diagnose flu without actually observing and identifying the flu virus itself inside the patient under a microscope. 

However, the existence of the virus, and its causal role in producing the symptoms observed, has indeed been demonstrated scientifically in other individuals afflicted with the same or similar symptoms. We therefore recognise the underlying cause of these symptoms (i.e. the virus) independently from the symptoms they produce. 

This is not true, however, for mental illnesses. The latter were named, identified and diagnosed long before there was any understanding of their underlying physiological basis. 

Rather than diseases, we might then more accurately call them syndromes, a word deriving from the Greek ‘σύνδρομον’, meaning ‘concurrence’, which is usually employed in medicine to refer simply to a cluster of signs and symptoms that seem to correlate together, whether or not the underlying cause is or is not understood.[3]

Causes and Correlates 

The main problem for Szasz’s position is that our understanding of the underlying physiological causes of psychiatric conditions – neurological, genetic and hormonal – has progressed enormously since he first authored The Myth of Mental Illness, the paper and the book, at the beginning of the 1960s. 

Yet reading ‘Psychiatry: The Science of Lies’, published in 2008, it seems that Szasz’s own position has advanced but little.[4]

Yet psychiatry, and psychology, have come a long way in the intervening half-century. 

Thus, in 1960, American psychiatry was still largely dominated by Freudian Fruedian psychoanalysis, a pseudoscience roughly on a par with phrenology, of which Szasz is rightly dismissive.[5]

Of particular relevance to Szasz’s thesis, the study of the underlying physiological basis for psychiatric disorders has also progressed massively.  

Every month, in a wide array of scientific journals, studies are published identifying neurological, genetic, hormonal and other physiological correlates for psychiatric conditions. 

In contrast, Szasz, although he never spells this out, seems to subscribe to an implicit Cartesian dualism, whereby human emptions, psychological states and behaviour are a priori assumed, in principle, to be irreducible to mere physiological processes.[6]

Szasz claims in Psychiatry: The Science of Lies that, once an underlying neurological basis for a mental illness has been identified, it ceases to be classified as a mental illness, and is instead classed as a neurological disorder. His paradigmatic example of this is Alzheimer’s disease (p2).[7]

Yet, today, the neurological correlates of many mental illnesses are increasingly understood. 

Nevertheless, despite the progress that has been made in identifying physiological correlates for mental disorders, there remains at least two differences between these correlates (neurological, genetic, hormonal etc.) and the recognised causes of both physiological and neurological diseases. 

First, in the case of mental illnesses, the neurological, genetic, hormonal and other physiological correlates remain just that, i.e. mere correlates

Here, I am not merely reiterating the familiar caution that correlation does not imply causation, but also emphasizing that the correlations in question tend to be far from perfect, and do not form the basis for a diagnosis, even in principle. 

In other words, as a rule, few such identified correlates are present in every single person diagnosed with the condition in question. The correlation is established only at the aggregate statistical level. 

Moreover, those persons who present the symptoms of a mental illness but who do not share the physiological correlate that has been shown to be associated with this mental illness are not henceforth identified as not truly suffering from the mental illness in question. 

In other words, not only is diagnosis determined, as a matter of convenience and practicality, by reference to symptoms (as is also often true for many physical illnesses), but mental illnesses remain, in the last instance, defined by the symptoms they produce, not any underlying physiological cause. 

Any physiological correlates for the condition are ultimately incidental and have not caused physicians to alter their basic definition of the condition itself. 

Second, the identified correlates are, again as a general rule, multiple, complex and cumulative in their effects. In other words, there is not one single identified physiological correlate of a given mental illness, but rather multiple identified correlates, often each having small cumulative effects of the probability of a person presenting symptoms. 

This second point might be taken as vindicating Szasz’s position that mental illnesses are not really illnesses. 

Thus, recent research on the genetic correlates of mental illnesses, as recently summarized by Robert Plomin in his book Blueprint: How DNA Makes Us Who We Are, has found that the genetic variants that cause psychiatric disorders are the exact same genetic variants which, when present in lesser magnitude, also cause normal, non-pathological variation in personality and temperament. 

This suggests that, at least at the genetic level (and thus presumably at the phenotypic level too), what we call mental illness is just an extreme presentation of what is normal variation in personality and behaviour. 

In other words, so-called mental illness simply represents the extreme tail-end of the normal bell curve distribution in personality attributes. 

This is most obviously true of the so-called personality disorders. Thus, a person extremely low in empathy, or the factor of personality referred to by psychometricians as agreeableness, might be diagnosed with anti-social personality disorder (or psychopathy). 

However, it is also true for so-called other mental disorders. For example, ADHD (attention deficit hyperactivity disorder) seems to be mere medical jargon for someone who is very impulsive, with a short attention span, and lacking self-discipline (i.e. low in the factor of personality that psychometricians call conscientiousness) – all traits which vary on a spectrum across the whole population. 

On the other hand, clinical depression, unlike personality, is a temporary condition from which most people recover. Nevertheless, it is so strongly predicted by the factor of personality known to psychometricians as neuroticism that psychologist Daniel Nettle writes: 

Neuroticism is not just a risk factor for depression. It is so closely associated with it that it is hard to see them as completely distinct” (Personality: p114). 

Yet calling someone ‘ill’ because they are at the extreme of a given facet of personality or temperament is not very helpful. It is roughly equivalent to calling a basketballer ‘ill’ because he is exceptionally tall, a jockey ‘ill’ because he is exceptionally small, or Albert Einsteinill’ because he was exceptionally intelligent

Mental illness or Malingering?

While Szasz has therefore correctly identified problems with the conventional disease model of mental health, the model which he proposes in its place is, in my view, even more problematic, and less scientific, than the disease model that he has rightly rejected as probematic and misleading. 

Most unhelpful is the central place given in his theory to the notion of malingering, i.e. the deliberate faking of symptoms by the patient. 

This analysis may be a useful way to understand the nineteenth century outbreak of so-called hysteria, to which Szasz devotes considerable attention, or indeed the modern diagnosis of Munchausen syndrome, which again involves complaining of imagined or exaggerated physical symptoms. 

It may also be a useful way to understand the recently developed diagnosis of chronic fatigue syndrome (CFS, formerly ME), which, like hysteria, involves the patient complaining of physical symptoms for which no physical cause has yet been identified. 

Interestingly from a psychological perspective, all three of these conditions are overwhelmingly diagnosed among women and girls rather than men and boys. 

However, malingering may also be a useful way to understand another psychiatric complaint that was primarily complained of by men, albeit for obvious historical reasons – namely, so-called ‘shell shock’ (now, classed as PTSD) among soldiers during World War One.[8]

Here, unlike with hysteria and CFS, the patient’s motive and rationale for faking the symptoms in question (if this is indeed what they were doing) is altogether more rational and comprehensible – namely, to avoid the horrors of the trenches (from which women were, of course, exempt). 

However, this model of ‘malingering’ is clearly much less readily applicable to sufferers of, say, schizophrenia

Here, far from malingering or faking illness, those afflicted will often vehemently protest that they are not ill and that there is nothing wrong with them. However, their delusions are often such that, by any ordinary criteria, they are undoubtedly, in the colloquial if not the strict medical sense, completely fucking bonkers. 

The model of malingering can, therefore, only be taken so far. 

Defining Mental Illness? 

The fundamental fallacy at the heart of psychiatry is, according to Szasz, the mistaking of moral problems for medical ones. Thus, he opines: 

Psychiatrists cannot expect to solve moral problems by medical methods” (Myth of Mental Illness: p24). 

Szasz has a point. Despite employing the language of science, there is undoubtedly a moral dimension to defining what constitutes mental illness. 

Whether a given cluster of associated behaviours represents just a cluster of associated behaviours or a mental illness is not determined on the basis of objective scientific criteria. 

Rather, most American psychiatrists simply regard as a mental illness whatever the Diagnostic and Statistical Manual of Mental Disorders (DSM) of the American Psychiatric Association classifies as a mental disorder. 

This manual is treated as gospel by psychiatrists, yet there is no systematic or agreed criteria for inclusion within this supposedly authoritative work. 

Popular cliché has it that mental illnesses are caused by a ‘chemical imbalance’ in the brain.  

Certainly, if we are materialists, we must accept that it is the chemical composition of the brain that causes behaviour, pathological or otherwise. 

But on what criteria are we to say that a certain chemical composition of the brain is an ‘imbalance’ and another is ‘balanced’, one behaviour ‘pathological’ and one ‘normal’? 

The criteria on which we make this judgement is, as I see it, primarily a moral one.[9]

More specifically, mental illnesses are defined as such, at least in part, because the behavioral symptoms that they produce tend to cause suffering or distress either to the person defined as suffering from the illness, or to those around them. 

Thus, a person diagnosed with depression is themselves the victim of suffering or distress resulting from the condition; a person diagnosed with psychopathy, on the other hand, is likely to cause psychological distress to those around them with whom they come into contact. 

This is a moral, not a scientific, criterium, depending as it does on the notion of suffering or harm

Indeed, it is not only a moral question, but it is also one that has, in recent years, been heavily politicized. 

Thus, gay right activists actively and aggressively campaigned for many years to have homosexuality withdrawn from the DSM and reclassified as non-pathological, and, in 1974, they were successful.[10]

This campaign may have had laudable motives, namely to reduce the stigma associated with homosexuality and prejudice against homosexuals. Yet it clearly had nothing to do with science and everything to do with politics and morality. 

Indeed, homosexuality satisfies many criteria for illness.[11]

First, it is, despite some ingenious and some not so ingenious attempts to show otherwise, obviously biologically maladaptive. 

Whereas the politically correct view is that homosexuality is entirely natural, normal and non-pathological variation of normal sexuality, from a Darwinian perspective, this view is obviously untenable. 

Homosexual sex cannot produce offspring. Homosexuality therefore involves a maladaptive misdirection of mating effort, which would surely strongly selected against by natural selection.[12]

Homosexuality is therefore best viewed as a malfunctioning of normal sexuality, just as cancer is a kind of malfunctioning of cell growth and division. In this sense, then, homosexuality is indeed best viewed as something akin to an illness. 

Second, homosexuality shows some degree of comorbidity with other forms of mental illness, such as depression.[13]

Finally, homosexuality is associated other undesirable life-outcomes, such as reduced longevity and, at least for male homosexuals, a greater lifetime susceptibility to various STDs.[14]

Yet, just as homosexuals successfully campaigned for the removal of homosexuality from the DSM, so trans rights campaigners are currently embarking on a similar campaign in respect of gender dysphoria

The politically correct consensus today holds that an adult or child who claims to identify of the opposite ‘gender’ to their biological sex should be encouraged and supported in their ‘transition’, and provided with hormone therapy, hormone blockers and sex reassignment surgery, as requested. 

This is roughly the equivalent of, if a person is mentally ill and thinks they are Napoleon, then, instead of telling them that they are not Napoleon, instead we provide them with legions with which to invade Prussia. 

Moving beyond the sphere of sexuality, some self-styled ‘neurodiversity’ activists have sought to reclassify autism as a normal variation of mental functioning, a claim that may seem superficially plausible in respect of certain forms of so-called ‘high functioning autism’, but is clearly untenable in respect of ‘low functioning autism’. 

Yet, on the other hand, there is oddly no similar, high-profile campaign to reclassify, say, anti-social personality disorder (ASPD) or psychopathy as a normal, non-pathological variant of human psychology. 

Yet psychopathy may indeed be biologically adaptive at least under some conditions (Mealey 1995). 

Yet no one proposes treating ASPD as normal or natural variation in personality, even though it is likely just that. 

The reason that there is no campaign to remove psychopathy from the DSM is, of course, because, unlike homosexuals, transexuals and autistic people, psychopaths are hughly disproportionately likely to cause harm to innocent non-consenting third-parties. 

This is indeed a good reason to treat psychopathy and anti-social personality disorder as a problem for society at large. However, this is a moral not a scientific reason for regarding it as problematic. 

To return to the question of disorders of sexuality, another useful point of comparison is provided by paedophilia

From a purely biological perspective, paedophilia is analogous to homosexuality. Both are biologically maladaptive because they involve sexual attraction to a partner with whom reproduction is, for biological reasons, impossible.[15]

Yet, unlike in the case of homosexuality, there has been no mainstream political push for paedophilia to be reclassified as nonpathological or removed from the Diagnostic and Statistical Manual of Mental Disorders of the AMA.[16]

The reason for this is again, of course, obvious and entirely reasonable, yet it equally obviously has nothing to do with science and everything to do with morality – namely, whereas homosexual behaviour as between consenting adults is largely harmless, the same cannot be said for child sexual abuse.[17]

Perhaps an even better analogy would be between homosexuality and, say, necrophilia

Necrophilic sexual activity, like homosexual sexual activity, but quite unlike paedophilic sexual activity, represents something of a victimless crime. A corpse, by virtue of being dead, cannot suffer by virtue of being violated.[18]

Yet no one would argue that necrophilia is a healthy and natural variation on normal human sexuality. 

Of course, although numbers are hard to come by due to the attendent stigma, necrophilia is presumably much less common, and hence much less ‘normal’, than is homosexuality. However, if this is a legitimate reason for regarding homosexuality as more ‘normal’ than is necrophilia, then it is also a legitimate reason for regarding homosexuality itself as ‘abnormal’, because homosexuality is, of course, much less common than heterosexuality.

Necrophile rights is, therefore, the reductio ad absurdum of gay rights.[19]

Medicine or Morality? 

The encroachment of medicine upon morality continues apace, as part of what Szasz calls the medicalization of everyday life Thus, there is seemingly no moral failing or character defect that is not capable of being redefined as a mental disorder. 

Selfish people are now psychopaths, people lacking in willpower and with short attention spans now have ADHD

But if these are simply variations of personality, does it make much sense to call them diseases? 

Yet the distinction between ‘mad’ and ‘bad also has practical application in the operation of the criminal justice system. 

The assumption is that mentally ill offenders should not be punished for their wrongdoing, but rather treated for their illness, because they are not responsible for their actions. 

But, if we accept a materialist conception of mind, then all behaviour must have a basis in the brain. On what basis, then, do we determine that one person is mentally ill while another is in control of his faculties?

As Robert Wright observes: 

“[Since] in both British and American courts, women have used premenstrual syndrome to partly insulate themselves from criminal responsibility… can a ‘high-testosterone’ defense of male murderers be far behind?… If defense lawyers get their way and we persist in removing biochemically mediated actions from the realm of free will, then within decades [as science progresses] the realm will be infinitesimal” (The Moral Animal: p352-3).[20]

Yet a man claiming that, say, high testosterone caused his criminal behaviour is unlikely to be let off on this account, because, if high testosterone does indeed cause crime, then we have good reason to lock up high testosterone men precisely because they are likely to commit crimes.[21]

Szasz wants to resurrect the concept of free will and hold everyone, even those with mental illnesses, responsible for their actions. 

My view is the opposite: No one has free will. All behaviour, normal or pathological, is determined by the physical composition of the brain, which is, in turn, determined by some combination of heredity and environment. 

Indeed, determinism is not so much a finding of science as its basic underlying assumption and premise.[22]

In short, science rests on the assumption that all events have causes and that, by understanding the causes, we can predict behaviour. If this were not true, then there would be no point in doing science, and science would not be able to make any successful predictions. 

In short, criminal punishment must be based on consequentialist utilitarian considerations such as deterrence, incapacitation and rehabilitation rather than such unscientific moralistic notions as free will, just deserts and blame.[23]

A Moral Component to All Medicine? 

Szasz is right, then, to claim that there is a moral dimension to psychiatric diagnoses. 

This is why psychopathy is still regarded as a mental disorder even though it is likely an adaptive behavioural strategy and life history in certain circumstances (Mealey 1995). 

It is also why homosexuality is no longer regarded as a mental illness, despite its obviously biologically maladaptive consequences, yet there is no similar campaign to remove paedophilia from the DSM. 

Yet what Szasz fails to recognise is that there is also a moral element to the identification and diagnosis of physical illnesses too. 

Thus, physical illnesses, like psychiatric illnesses, are called illnesses, at least in part, because they cause pain, suffering and impairment in normal functioning to the person diagnosed as suffering from the illness. 

If, on the other hand, an infection did not produce any unpleasant symptoms, then the patient would surely never bother to seek medical treatment and thus the infection would probably never come to the attention of the medical profession in the first place. 

If it did come to their attention, would they still call it a disease? Would they expect time and resources attempting to ‘cure’ it? Hopefully not, as to do so would be a waste of time and resources. 

Extending this thought experiment, what if the infection in question, not only caused no negative symptoms, but actually had positive effects on the person infected.   

What if the infection in question caused people to be fitter, smarter, happier, kinder and more successful at their jobs? 

Would doctors still call the infection a ‘disease’, and the microscopic organism underlying it a ‘germ’? 

Actually, this hypothetical thought experiment may not be entirely hypothetical. 

After all, there are indeed surely many microorganisms that infect humans which have few or negligible effects, positive or negative, and with which neither patients nor doctors are especially concerned. 

On the other hand, some infections may be positively beneficial to their hosts. 

Take, for example, gastrointestinal microbiota (also known as gut microbiota). 

These are microorganisms that inhabit our digestive tracts, and those of other organisms, and are thought to have a positive beneficial effect on the health and functioning of the host organism. They have even been marketed as probiotics and good bacteria in the advertising campaigns for certain yoghurt-like drinks. 

Another less obvious example is perhaps provided by mitochondrial DNA

In our ancient evolutionary history, this began as the DNA of a separate organism, a bacterium, that infected host organisms, but ultimately formed a symbiotic and mutualistic relationship with us, and now plays a key role in the functioning of those organisms whose distant ancestors it first infected. 

In short, all medicine has a moral dimension.  

This is because medicine is an applied, not a pure, science. 

In other words, medicine aims not merely to understand disease in the abstract, but to treat it. 

We treat diseases to minimize human suffering, and the minimization of human suffering is ultimately a moral (or perhaps economic, since doctors are paid, and provide a service to their patients), rather than a purely scientific, endeavour. 


[1] Although this post is a review of Thomas Szasz’s Pyschiatry: The Science of Lies, readers may note that many of the quotations from Szasz in the review are actually taken from his earlier, more famous book, The Myth of Mental Illness, published some several decades previously. By way of explanation, while this essay is a review of Szasz’s Psychiatry: The Science of Lies, I listened to an audiobook version of this book, and do not have access to a print copy. It was therefore difficult to find source quotes from this book. In contrast, I own a copy of The Myth of Mental Illness, but have yet to read it in full. I thought it more useful to read a more recent statement of Szasz’s views, so as to find out how he has dealt with recent findings in biological psychiatry and behavioural genetics. Unfortunately, as I discuss above, it seems that Szasz has reacted to recent findings in biological psychiatry and behavioural genetics hardly at all, and includes few if any references to such developments in his more recent book.

[2] Thus, proponents of Darwinian medicine contend that many infections produce symptoms such as coughing, sneezing and diarrhea precisely because these symptoms facilitate the spread of the disease through contact with the bodily fluids expelled, hence promoting the pathogens’ own Darwinian fitness or reproductive success.

[3] For example, the underlying physical cause of chronic fatigue syndrome (CFS) is not fully understood. On the other hand, the underlying cause of acquired immunodeficiency syndrome (AIDS) is now understood, namely HIV infection, but, presumably because it involves increased susceptibility to many different infections, it is still referred to as a syndrome rather than a disease in and of itself.

[4] Indeed, according to Szasz himself, in an autobiographical interlude in ‘Psychiatry: The Science of Lies’, he had arrived at his opinion regarding the scientific status of psychiatry even earlier, when first making the decision to train to become a psychiatrist. Indeed, he claims to have made the decision to study psychiatry and qualify as a psychiatrist precisely in order to attack the field from within, with the authority which this professional qualification would confer upon him. This, it hardly needs to be said, is a very odd reason for a career choice.

[5] Attacking modern psychiatry by a critique of Freud is a bit like attacking neuroscience by critiquing nineteenth century phrenology. It involves constructing a straw man version of modern psychiatry. I am reminded in particular of Arthur Jensen’s review of infamous charlatan Stephen Jay Gould’s discredited The Mismeasure of Man, which Jensen titled The debunking of scientific fossils and straw persons, where he described Gould’s method of trying to discredit the modern science of IQ testing and intelligence research by citing the errors of nineteenth-century phrenologists as roughly akin to “trying to condemn the modern automobile by merely pointing out the faults of the Model T”.

[6] In The Myth of Mental Illness, Szasz, writes: 

There remains a wide circle of physicians and allied scientists whose basic position concerning the problem of mental illness is essentially that expressed in Carl Wernicke’s famous dictum: Mental diseases are brain diseases’. Because, in one sense, this is true of such conditions as paresis and the psychoses associated with systemic intoxications, it is argued that it is also true for all other things called mental diseases. It follows that it is only a matter of time until the correct physicochemical, including genetic, bases or cause’, of these disorders will be discovered. It is conceivable, of course, that significant physicochemical disturbances will be found in some mental patients and in some conditions now labeled mental illnesses. But this does not mean that all so-called mental diseases have biological causes, for the simple reason that it has become customary to use the term mental illness to stigmatize, and thus control, those persons whose behavior offends society—or the psychiatrist making the diagnosis” (The Myth of Mental Illness: p103). 

Yet, if we accept a materialist conception of mind, then all behaviours, including those diagnostic of mental illness, must have a cause in the brain, though it is true that the same behaviours may result from quite different neuroanatomical causes.
It is certainly true that the concept of mental illness has been used to “stigmatize, and thus control, those persons whose behavior offends society”. So-called drapetomania provides an obvious example, albeit one that was never widely recognised by physicians, at least outside the American South. Another example would be the diagnosis of sluggish schizophrenia to& institutionalize political dissidents in the Soviet Union. Likewise, psychopathy (aka sociopathy or anti-social personality disorder) may, as I argue later in this post, have been classified as a mental disorder primarily because the behaviour of people diagnosed with this condition does indeed “offend society” and arguably demand the “control”, and sometimes detention, of such people.
However, this does not mean that the behaviours complained of (e.g. political dissidence, or anti-social behaviours) will not have neural or other physiological correlates. On the contrary they undoubtedly do, and psychologists have also investigated the neural and other physiological correlates of all behavours, not just those labelled as ‘mental illnesses’.
However, Szasz does not quite go so far as to deny that behaviours have physical causes. On the contrary, in The Myth of Mental Illness, hedging his bets against future scientific advances, Szasz acknowledges:

I do not contend that human relations, or mental events, take place in a neurophysiological vacuum. It is more than likely that if a person, say an Englishman, decides to study French, certain chemical (or other) changes will occur in his brain as he learns the language. Nevertheless, I think it would be a mistake to infer from this assumption that the most significant or useful statements about this learning process must be expressed in the language of physics. This, however, is exactly what the organicist claims” (The Myth of Mental Illness: p102- 3). 

Here, Szasz makes a good point – but only up to a point. Whether we are what Szasz calls ‘organicists’ or not, I’m sure we can all agree that, for most purposes, it is not useful to explain the decision to learn French in terms of neurophysiology. To do so would be an example of what philosopher Daniel Dennett, in Darwin’s Dangerous Idea, calls ‘greedy reductionism’, which he distinguished from ‘good reductionism’, which is central to science.
However, it is not clear that the same is true of what we call mental illnesses. Often it may indeed be useful to understand mental illnesses in terms of their underlying physiological causes, including for therapeutic reasons, since understanding the physiological basis for behaviour that we deem undesirable may provide a means of changing these behaviours by altering the physical composition of the brain. For example, if the hormone serotonin is involved in regulating mood, then manipulating levels of serotonin in the brain, or their reabsorption may be a way of treating depression, anxiety and other mood disorders. Thus, SSRIs and SNRIs, which are thought to do just this, have been found to be effective in doing just this.
However, for other purposes, it may be useful to look at a different level of causation. For example, as I discuss in a later endnote, although it may be scientifically a nonsense, it may nevertheless be useful to cultivate a belief in free will among some psychiatric patients, since it may encourage them to overcome their problems rather adopting the fatalistic view that they are ill and there is hence nothing they can do to improve their predicament. Szasz sometimes seems to be arguing for something along these lines.

[7] In The Myth of Mental Illness, as quoted in the preceding endnote, Szasz also gives as examples of behavioural conditions with well-established physiological causes “paresis and the psychoses associated with systemic intoxications(The Myth of Mental Illness: p103).

[8] I hasten to emphasize in this context, lest I am misunderstood, I am not saying that Szasz’s model of ‘malingeringis indeed the appropriate way to understand conditions such as hysteria, Munchausen syndrome, chronic fatigue syndrome or shell shock – only that a reasonable case can be made to this effect. Personally, I do not regard myself as having a sufficient expertise on the topic to be willing to venture an opinion either way.

[9] Of course, we could determine whether a certain composition and structure of the brain is ‘balanced’ ‘imbalanced’ on non-moralistic, Darwinian criteria. In other words, if a certain composition/structure and the behaviour it produces is adaptive (i.e. contributes to the reproductive success or fitness of the organism) then we could call it ‘balanced’; if, on the other hand, it produces maladaptive behaviour we could call it ‘imbalanced’. However, this would produce a quite different inventory and classification of mental illnesses than that provided by the DSM of the APA and other similar publications, since, as we will see, homosexuality, being obviously biologically maladaptive, would presumably be classified as an ‘imbalance’ and hence a mental illness, whereas psychopathy, since it may well, under certain conditions, be adaptive, would be classed as non-pathological and hence ‘balanced’. This analysis, however, has little to do with mental illness as the concept is currently conceived.

[10] Oddly, Szasz himself is sometimes lauded by some politically correct-types as being among the first psychiatrists to deny that homosexuality was a mental illness. Yet, since he also denied that schizophrenia was a mental illness, and indeed rejected the whole concept of ‘mental illness’ as it is currently conceived, this is not necessarily as ‘progressive’ and ‘enlightened’ a view as it is sometimes credited as having been.

[11] Here, a few caveats are in order. Describing homosexuality as a mental illness no more indicates hatred towards homosexuals than describing schizophrenia as a mental illness indicates hatred towards people suffering from schizophrenia, or describing cancer as an illness indicates hatred towards people afflicted with cancer. In fact, regarding a person as suffering from an illness is generally more likely to elicit sympathy for the person so described than it is hatred.
Of course, being diagnosed with a disease may involve some stigma. But this is not the same as hatred.
Moreover, as is clear from my conclusion, I am not, in fact, arguing that homosexuality should indeed be classified as a mental illness. Rather, I am simply pointing out that it is difficult a frame a useful definition of what constitutes a ‘mental disorder’ unless that definition includes moral criteria, which are necessarily extra-scientific. However, in the final section of this piece, I argue that there is indeed a moral component to all medicine, psychiatry included.
Of course, as I also discuss above, there are indeed some moral reasons for regarding homosexuality as undesirable, for example its association with reduced longevity, which is generally regarded as an undesirable outcome. However, whether homosexuality should indeed be classed as a ‘mental disorder’ strikes me as debatable and also dependent on the exact definition of ‘mental disorder’ adopted.

[12] If homosexuality is therefore maladaptive, this, of course, raises the question as to why homosexuality has not indeed been eliminated by natural selection. The first point to make here is that homosexuality is in fact quite rare. Although Kinsey famously originated the since-popularized claim that as many as 10% of the population are homosexual, reputable estimates using representative samples suggest less than 5% of the population identifies as exclusively or preferentially homosexual (though a larger proportion of people may have had homosexual experiences at some time, and the ‘closet factor’ makes it possible to argue that, even in an age of unprecedented tolerance and indeed celebration of homosexuality, and even in anonymous surveys, this may represent an underestimate due to underreporting).
Admittedly, there has recently been a massive increase in the numbers of teenage girls identifying as non-heterosexual, with numbers among this age group now slightly exceeding 10%. However, I suspect that this is also as much a matter of fashion as of sexuality. Thus, it is notable that the largest increase has been for identification as ‘bisexual’, which provides an convenient cover by which teenage girls can identify with the so-called ‘LGBT+ community’ while still pursuing normal, healthy relationships with opposite-sex boys or men. The vast majority of these girls will, I suspect, grow up to have sexual and romantic relationships primarily with members of the opposite sex.
Yet even these low figures are perhaps higher than one might expect, given that homosexuality would be strongly selected against by evolution. (However, it is important to remember that, when homosexuals were persecuted and hence mostly remained in the ‘closet’, homosexuality would have been less selected against, precisely because so many gay men and women would have married members of the opposite sex and reproduced if only to evade accusations of homosexuality. With greater tolerance, however, they no longer have any need to do so. The liberation of homosexuals may therefore, paradoxically, lead to their gradual disappearance through selection.)
A second point to emphasize is that, contrary to popular perception, homosexuality is not especially heritable. Indeed, it is rather less heritable than other behavioural traits about which it is much less politically correct to speculate regarding the heritability (e.g. criminality, intelligence).
If homosexuality is primarily caused by environmental factors, not genetics, then it would be more difficult for natural selection to weed it out. However, given that exclusive or preferential homosexuality would be strongly selected against by natural selection, humans should have evolved to be resistant to developing exclusive or preferential homosexuality under all environmental conditions that were encountered during evolutionary history. It is possible, however, environmental novelties atypical of the environments in which our psychological adaptations evolved are responsible for causing homosexuality.
For what it’s worth, my own favourite theory (although not necessarily the best supported theory) for the evolution of male homosexuality proposes that genes located on the X chromosome predispose a person to be sexually attracted to males. This attraction is adaptive for females, but maladaptive for males. However, since females have two X chromosomes and males only one and therefore any X chromosome genes will find themselves in females twice as often as they find themselves in males, any increase in fitness for females bearing these X chromosome genes only has to be half as great as the reproductive cost to males for the genes in question to be positively selected for.
This is sometimes called the ‘balancing selection theory of male homosexuality’. However, perahps more descriptive and memorable is Satoshi Kanazawa’s coinage, ‘the horny sister hypothesis’.
This theory also has some support, in that there is some evidence the female relatives of male homosexuals have a greater number of offspring than average and also that gay men report having more gay uncles on their mother’s than their father’s side, consistent with an X chromosome-linked trait (Hamer et al 1993; Camperio-Ciani et al 2004). Some genes on the X chromosome have also been linked to homosexuality (Hamer et al 1993; Hamer 1999).
On the other hand, other studies find no support for the hypothesis. For example, Bailey et al (1999) found that rates of reported homosexuality were no higher among maternal than among paternal male relatives, as did McKnight & Malcolm (1996). At any rate, as explained by Wilson and Rahman in their excellent book Born Gay: The Psychobiology of Sexual Orientation:

Increased rates of gay maternal relatives might also appear because of decreased rates of reproduction among gay men. A gay gene is unlikely to be inherited from a gay father because a gay man is unlikely to have children” (Risch et al 1993) (Born Gay: p51).

[13] Gay rights activists assert that the only reason that homosexuality is associated with other forms of mental illness is because of the stigma to which homosexuals are subject on account of their sexuality. This has sometimes been termed the ‘social stress hypothesis’, ‘social stress model’ or ‘minority stress model’. There is indeed statistical support for the theory that the social stigma is indeed associated with higher rates of depression and other mental illnesses.
It is also notable that, while homosexuality is indeed consistently associated with higher levels of depression and suicide, conditions that can obviously be viewed as a direct response to social stigma, I am not aware of any evidence suggesting higher rates of, say, schizophrenia among homosexuals, which would less obviously, or at least less directly, result from social stress. However, I tend to agree with the conclusions of Mayer and McHugh, in their excellent review of the literature on this subject, that, while social stress may indeed explain some of the increased rate of mental illness among homosexuals, it is unlikely to account for the totality of it (Mayer & McHugh 2016).

[14] Yet, in describing the life outcomes associated with homosexuality, as undesirable, I am, of course, making am extra-scientific value judgement. Of course, the value judgement in question – namely that dying earlier and being disproportionately likely to contract STDs is a bad thing – is not especially controversial. However, it still illustrates the extent to which, as I discuss later in this post, definitions of mental illnesses, and indeed physical illnesses, always include a moral dimension – i.e. diseases are defined, in part, by the fact that they cause suffering, either to the person afflicted, or, in the case of some mental illnesses, to the people in contact with them.

[15] Indeed, from a purely biological perspective, homosexuality is arguably even more biologically maladaptive than is paedophilia, since even very young children can, in some exceptional cases, become pregnant and even successfully birth offspring, yet same-sex partners are obviously completely incapable of producing offspring with one another.

[16] Indeed, far from there being any political pressure to remove paedophilia from the DSM of the AMA, as ocurred with homosexuality, there is instead increasing pressure to add hebephilia (i.e. attraction to pubescent and early-post-pubescent adolescents) to the DSM. If successful, this would probably lead to pressure to also add ‘ephebophilia’ (i.e. the biologically adaptive and normal male attraction to mid- to late-adolescents) to the DSM, and thereby effectively pathologize and medicalize, and further stigmatize, normal male sexuality.

[17] Of course, homosexual sex does have some dangers, such as STDs. However, the same is also true of heterosexual sex, although, for gay male sex, the risks are vastly elevated. Yet other perceived dangers result from only from heterosexual sex (e.g. unwanted pregnancies, marriage). Meanwhile, the other negative life outcomes associated with homosexuality (e.g. elevated risk of depression and suicide) probably result from a homosexual orientation rather than from gay sex as such. Thus, a celibate gay man is, I suspect, just as likely, if not more likely, to suffer depression than is a highly promiscuous gay man.
Yet, while gay sex may be mostly harmless, the same cannot, of course, be said for child sexual abuse. It may indeed be true that the long-term psychological effects of child sexual abuse are exaggerated. This was, of course, the infamous conclusion of the Rind et al meta-analysis, which resulting in much moral panic in the late-1990s (Rind et al 1998). This is especially likely to be the case when the sexual activity in question is consensual and involves post-pubertal, sexually mature (but still legally underage) teenagers. However, in such cases the sexual activity in question should not really be defined as ‘child sexual abuse’ in the first place, since it neither involves immature children in the biological sense, nor is it necessarily abusive. Yet, it must be emphasized, even if child sexual abuse does not cause long-term psychological harm, it may still cause immediate harm, namely the distress experienced by the victim at the time of the abuse.

[18] Of course, one might argue that the relatives of the deceased may suffer as a result of the idea of their dead relatives’ bodies being violated. However, much the same is also true of homosexuality. So-called ‘homophobes’, for example, may dislike the idea of their adult homosexual sons having consensual homosexual sex. Indeed, they may even dislike the idea of unrelated adult strangers being allowed to have consensual homosexual sex. This was indeed presumably the reason why homosexuality has been criminalized and prohibited in so many cultures across history in the first place, i.e. because other people were disgusted by the thought of it. However, we no longer regard this sort of puritanical, disapproval other people’s private lives as a sufficient reason to justify the criminalization of homosexual behaviour. Why then should it be a reason for criminalizing necrophilia?

[19] Other similar thought experiments involve the prohibitions on other sexual behaviours such as zoophilia and incest. In both these cases, however, the case is morally more complex, in the case of zoophilia on account of whether the animal participant suffers harm or has consented, and, in the case of incest, because of eugenic considerations, namely the higher rate of the expression of deleterious mutations among the offspring of incestuous unions.

[20] Indeed, the courts, in both Britain and America, have been all too willing to invent bogus pseudo-psychiatric diagnoses in order to excuse women, in particular, for culpability in their crimes, especially murder. For example, in Britain, the Infanticide Acts of 1922 and 1938 provide a defence against murder for women who kill their helpless new-born infants where “at the time of the act… the balance of her mind was disturbed by reason of her not having fully recovered from the effect of giving birth to the child or by reason of the effect of lactation consequent upon the birth of the child”. In terms of biology, physiology and psychology, this is, of course, a nonsense, and, of course, no equivalent defence is available for fathers, though, in practice, the treatment of mothers guilty of infanticide is more lenient still (Wilczynski and Morris 1993).
Similarly, in both Britain and America, women guilty of killing their husbands, often while the latter was asleep or otherwise similarly incapacitated, have been able to avoid being a murder conviction by claiming to have been suffering from so-called ‘battered women syndrome’. There is, of course, no equivalent defence for men, despite the consistent finding that men are somewhat more likely to be the victim of violence from their female intimate partners than women are to have been victimized by their male intimate partners (Fiebert 2014). This may partly explain why men who kill their wives receive, on average, sentences three time as long as women who kill their husbands (Langan & Dawson 1995).

[21] Of course, another possibility might be some form of hormone therapy to reduce the offender’s testosterone. Actually, it must be acknowledged that whether testosterone is indeed correlated with criminal or violent behaviour is the subject of some dispute. Thus, Alan Mazur, a leading researcher in this area, argues that testosterone is not associated with aggression or violence as such, but rather only with dominance behaviours, which can also be manifested in non-violent ways. For example, a high-powered business tycoon is likely to be high in social dominance behaviours, but relatively unlikely to commit violent crimes. On the other hand, a prisoner, being of low status, may be able to exercise dominance only through violence. I am therefore giving the example of high testosterone only as a simplified thought experiment.

[22] Of course, one finding of science, namely quantum indeterminism, complicates this assumption. Ironically, while determinism is the underlying premise of all scientific enquiry, nevertheless one finding of such enquiry is that, at the most fundamental level, determinism does not hold.

[23] Nevertheless, I am persuaded that there may be some value in the concept of free will, after all. Although it is a nonsense, it may, like some forms of religious belief, nevertheless be a useful nonsense, at least in some circumstances.
Thus, if a person is told that there is no free will, and that their behaviours are inevitable, this may encourage a certain fatalism and the belief that people cannot change their behaviours for the better. In fact, this is a fallacy. Actually, determinism does not suggest that people cannot change their behaviours. It merely concludes that whether people do indeed change their behaviours is itself determined. However, this philosophical distinction may be beyond many people’s understanding.
Furthermore, if people are led to believe that they cannot alter their own behaviour, then this may become something of a self-fulfilling prophecy, and thereby prevent self-improvement.
Therefore, just as religious beliefs may be untrue, but nevertheless serve a useful function in giving people a reason to live and to behave prosocially and for the benefit of society as a whole, so it may be beneficial to encourage a belief in free will in order to encourage self-improvement, including among the mentally ill.


Bailey et al (1999) A Family History Study of Male Sexual Orientation Using Three Independent Samples, Behavior Genetics 29(2): 79–86. 
Camperio-Ciani (2004) Evidence for maternally inherited factors favouring male homosexuality and promoting female fecundity, Proceedings of the Royal Society B: Biological Sciences 271(1554): 2217–2221. 
Fiebert (2014) References Examining Assaults by Women on Their Spouses or Male Partners: An Updated Annotated Bibliography, Sexuality & Culture 18(2):405-467. 
Hammer et al (1993) A linkage between DNA markers on the X chromosome and male sexual orientation, Science 261(5119):321-7.  
Hammer (1999) Genetics and Male Sexual Orientation, Science 285(5429): 803. 
Langan & Dawson (1995) Spouse Murder Defendants in Large Urban Counties, U.S. Department of Justice Office of Justice Programs, Bureau of Justice Statistics: Executive Summary (NCJ-156831), September 1995. 
Mayer & McHugh (2016) Sexuality and Gender Findings from the Biological, Psychological, and Social Sciences, New Atlantis 50: Fall 2016. 
McKnight & Malcolm (2000) Is male homosexuality maternally linked? Evolution and Gender 2(3):229-252. 
Mealey (1995) The sociobiology of sociopathy: An integrated evolutionary model. Behavioral and Brain Sciences, 18(3): 523–599.
Rind et al(1998). A Meta-Analytic Examination of Assumed Properties of Child Sexual Abuse Using College Samples, Psychological Bulletin 124 (1): 22–53.
Risch et al (1993) Male Sexual Orientation and Genetic Evidence, Science 262(5142): 2063-2065. 
Szasz 1960 The Myth of Mental Illness. American Psychologist, 15, 113-118. 
Wilczynski & Morris (1993) Parents Who Kill their children, Criminal Law Review, 31-6.