The scale of diagnosed mental illness is now prodigious. In Britain, for example, some 170,000 patients are admitted each year to hospitals for various categories of “mental illness” (and another 16,000 for “mental handicap”). Mental illness patients these days are discharged quite soon, so there are only some 80,000 in hospitals at any one time. Mental handicap patients stay longer—there are almost 47,000 in hospitals at any one time. Put another way, one in twelve men and one in eight women in the United Kingdom—the proportions are similar in the United States—will now go to a hospital at some point in their lives to be treated for mental illness.[c1] Yet the medical colonization of madness is a rather recent phenomenon: only in the last two centuries has madness been regarded as a medical matter at all.[c2]
These figures are not static, reflecting as they do changing social definitions of wellness and illness, assumptions about the necessity for and most appropriate form of treatment, and so forth. Thus in recent years there have been dramatic changes in the mental hospital population. There has been an increase in the number of admissions to hospitals, but the average length of stay in hospitals has gone down. The result is a decline in the number of inpatients, that is, people confined to hospitals and regarded as unfit to leave for a period. Instead, more of those diagnosed as mentally ill are treated as outpatients outside the hospital (“within the community”—generally, that is, by their family) than previously. Perhaps the most striking example of such a change has been in Italy, which in 1978 passed a law closing all mental hospitals. From then on patients were to be treated in the community or as part of general hospital practice.
At an earlier time psychiatrists and neurologists chose to distinguish between “organic” and “functional” nervous disorders. In organic disorders there was something obviously and demonstrably wrong with the brain. There might be a lesion, or the aftermath of a stroke or toxic poisoning, or whatever. By contrast “functional” disorders—schizophrenia, the depressions, paranoia, and so on—were disorders of the mind, which could not be attributed to any obvious brain damage. We can see in this distinction a residue of the old Cartesian dualism, a split between body and mind functions. Some contemporary psychiatrists wish to maintain this position. In his numerous books polemicizing against contemporary institutional psychiatry, Thomas Szasz, for instance, argues that if schizophrenia were shown to have an associated biological dysfunction, it should be left to the compulsory medicalization of the state for treatment, but so long as it remained a disorder of the spirit with no clear biological component it should be the voluntary choice of the sufferer whether or not to visit his or her psychiatrist for paid therapy.[c3]
But such a distinction is unacceptable to the dominant, full-blooded materialism of contemporary psychiatry. If there is a disordered mind, there must be associated with it some type of disordered molecular or cellular event in the brain. Further, the reductionist argument insists that there must be a direct causal chain running from the molecular events in particular brain regions to the most full-blown manifestations of existential despair suffered by the individual.
Present-day biological psychiatry divides disorders between neuroses, such as anxiety, and psychoses, of which schizophrenia is the outstanding example and the most common form of mental illness diagnosed today. The distinction offered between neurosis and psychosis is that in the former it seems as if sufferers perceive the same “real world” as do “normal individuals” but cannot react effectively and adaptively to it. By contrast, in the psychosis the individual’s world ceases to be normal at all, at least for a considerable part of the time. Instead it is replaced by one in which large elements seem to be of the sufferer’s own making, composed of fragments of the real world seen through a multifaceted distorting mirror. To the outside observer the psychotic is seen to be suffering from hallucinations and delusions.
But such definitions are inevitably uncertain. To begin with, they rest on a judgment about the meaning of normality. This involves comparing a given individual’s behavior with that of his or her fellows in similar situations, or a person’s behavior today with that on some previous occasion. It then becomes clear that definitions of normality are themselves time—and culture—bound. Joan of Arc—who heard voices which she claimed were those of angels telling her to crown the French Dauphin and drive out the English—became a heroine of the French nation. Later, long after her death, she was made a saint. Today she would almost certainly be diagnosed as schizophrenic, even though she might be spared burning at the stake. If an individual is cast into an apathetic despair about the likelihood of the world surviving nuclear holocaust through the eighties, or a woman in a northern English town is afraid to go out of her house at night for fear of being raped or murdered, how is one to judge that these are inappropriate responses compared with those of the less sensitive majority?
Clinicians often attempt to distinguish between “exogeneous” and “endogeneous” depressions. The former, it is claimed, are precipitated by events in the world outside the individual—a bereavement or loss of job for example—but sometimes even by events such as a promotion or moving to a new house. Endogenous depressions are said to be without obvious external precipitants and may recur cyclically at regular intervals, sometimes alternating with periods of exaggerated, frenzied cheerfulness (cyclical manic depression). In our present culture, depressions are often associated with important life cycle events (e.g. postpartum or postmenopausal depression). And even exogenous depressions may seem to take on a life of their own and fail to respond to the resolution of the initial precipitating cause. Women receive a higher proportion of diagnoses of depression and anxiety than men. The typical depressive consulting her family doctor is likely to be a middle-aged housewife.
Despite the textbook neatness of the distinctions between endogenous and exogenous depression and anxiety, the likelihood is that for most sufferers the distinctions are actually unclear; and in the clinical practice of most family doctors, rough-and-ready diagnostic criteria do not allow for much subtlety. In any event, once the diagnosis has been made the question is whether to try to normalize sufferers by persuading them that their despair or anxiety is uncalled for or by dulling it with drugs. It is immediately apparent that the relationship between the diagnosis of a behavior as illness and making judgments about appropriate and normal behavior is a very close one. It is at this point that the questions of cure and of control begin to intermingle, perhaps inextricably.
The diagnosis and treatment of schizophrenia are paradigms of the determinist mode of thinking, for this is the mental disorder on which more biochemical and genetic research has been lavished than any other, the one in which claims to have discovered the cause in a particular molecule or gene have been made most extensively. It is now so widely believed that psychiatry has proved the disorder to be biological that if the case fails here, where it is strongest, it must be even weaker elsewhere. But schizophrenia is interesting from another point of view as well, for in opposition to the biologizing tendencies of medical psychiatry there has grown up a strong countermovement in recent years. Antipsychiatry, in the hands of practitioners like R. D. Laing and theorists like Michel Foucault, has gone far in the opposite direction, almost to the point of denying the existence of a disorder or group of disorders diagnosable as schizophrenia at all. Thus in the case of schizophrenia we find precisely that clash of determinisms, on the one hand biological and on the other cultural, which we discussed in general in Chapters 3 and 4 and which it is one of the purposes of our book to transcend.
If the bulk of our effort here is directed toward the biochemical and particularly the genetic explanations offered for schizophrenia, this is because at the present time these explanations are so strongly entrenched in establishment psychiatry and medicine. We emphatically do not wish in this emphasis to be tipped over into an uncritical resuscitation of dualism, or cultural determinism like that of Laing or Foucault.
Schizophrenia literally means “split mind.” The classic picture of a schizophrenic is of a person who feels in some fundamental way cut off from the rest of humanity. Unable to express emotion or interact normally or express themselves verbally in a way that is rational to most others, schizophrenics appear blank, apathetic, dull. They may complain that their thoughts are not their own or that they are being controlled by some outside force. According to the textbooks, dramatically ill schizophrenics appear not to be able to or wish to do anything for themselves—they take little interest in food, sexual activity, or exercise; they experience auditory hallucinations; and their speech seems rambling, incoherent, and disconnected to the casual listener. Some psychiatrists doubt whether schizophrenia is a single entity at all, or speak of core schizophrenia and a wider range of schizophrenia-like symptoms.
The idea of a single disease of schizophrenia may be a hangover from the nineteenth-century definition of madness—so-called dementia praecox—which preceded it. The diagnosis of schizophrenia in a patient with a given set of symptoms can vary between doctor and doctor and culture and culture. It is true that when matched and carefully controlled transnational surveys are done there is some concordance of diagnosis; however, in real life the diagnostic and prescribing practices of doctors and psychiatrists differ sharply from the more controlled procedures of clinical trials. Comparisons of figures in different countries have shown that the most frequent use of the diagnosis of schizophrenia occurs in the United States and the Soviet Union. Nonetheless, even in Britain, where it is defined in a somewhat narrower sense, up to 1 percent of the population is said to suffer from schizophrenia;[c4] and 28,000—or 16 percent—of the admissions to hospital for mental illness in 1978 were for a diagnosis of schizophrenia or its related disorders.
Faced with the complex phenomena that result in a diagnosis of schizophrenia, the biological determinist has a simple question: What is it about the biology of the individual schizophrenic that predisposes him or her toward the disorder? If no obvious gross brain difference can be found, predisposition must lie in some subtle biochemical abnormality—perhaps affecting the connections between individual nerve cells. And the thrust of the determinist argument is that the causes for these abnormalities, although they might have been environmental, are most likely to lie in the genes.
Hence the enthusiastic hunt, over many decades now, for the biochemically abnormal component in schizophrenia. How should this search be conducted? A standard pattern in the biologizing of human medicine has been to seek for experimental animals that show what appear to be analogous symptoms. Or the animals can be induced to manifest similar symptoms by damaging them in some way, infecting them, or treating them with drugs. In the case of mental disorders this approach is problematic. How could one recognize a schizophrenic cat or dog, even if the term had any meaning anyhow? Such difficulties have not entirely chilled the enthusiasm of the researchers. Experimental animals have been treated with drugs such as LSD and have been shown to become disoriented, to show abnormal fear reactions, or whatever. These may be interpreted as analogous to hallucination, and hence the effect of the drug is argued to be analogous to the assumed biochemical dysfunction in schizophrenia.
But such evidence is not very convincing, and most research is directed to a study of the biochemistry of the schizophrenic subjects themselves. Brain samples are rarely obtainable except postmortem, and so more readily accessible body materials—urine, blood, or cerebrospinal fluid—from certified schizophrenics are compared with those from control “normal” people with all the assiduity that the Roman augurs used to apply to the examination of animals’ entrails. It is assumed that any biochemical abnormality in the brain will reflect itself in the production of abnormal metabolites in the blood, ultimately to be excreted in the urine.
When such approaches were first adopted several decades ago they soon began to show up large differences in the biochemistry of hospitalized schizophrenic patients from those of normals matched for sex, age, and so forth. But these differences turned out to be artifactual; nonschizophrenic hospitalized patients showed similar differences from the normal. The differences were eventually traced to the effects of long periods of eating poor hospital diets, or to the chemical-breakdown products of drugs that had been administered to the patients—or even to excessive coffee-drinking by hospitalized patients.
Even when proper care is taken to circumvent this problem by ensuring that the subjects studied have been kept off drugs for a period, that they have the same diet as their matched controls, and so forth, there remains a general methodological problem that cannot be avoided. Even if an abnormal chemical is found in the body fluids of a diagnosed schizophrenic compared with the best-matched of controls, one cannot infer that the observed substance is the cause of schizophrenia; it might instead be a consequence. The causal argument assumes that the substance is present, and, as a result, the disorder begins. A consequential argument says that first the disorder occurs and then as a result the substance accumulates. If an individual suffers an infection from a flu virus there is a considerable increase in the antibodies present in the blood and mucus of the nose—they are the body’s defense mechanisms against the virus. The antibodies and the mucus haven’t caused the infection, and one cannot readily deduce the actual causes simply by observing such consequences.
Such problems have made yet another approach more attractive to reductionist thinking: to observe the effects of pharmacological agents—drugs—on human behavior. If a drug induces schizophrenia-like behavior—for example, auditory hallucinations—then attempts will be made to conclude that the drug interferes with a biochemical process in the normal person which is damaged in the schizophrenic. Hence, for example, there was a period in the 1960s in which attempts were made to find links between LSD and schizophrenia on the grounds that users of LSD experienced hallucinations that might be seen as analogous to those of the schizophrenic. This logic, which argues backwards from the effect of a drug to the cause of a disease (ex juvantibus logic),[c5] is plainly a risky procedure, both for the logician and for the patient. As we have emphasized in the case of L-dopa, no drug has a single site of action. Foreign chemicals introduced into the body are not magic bullets.[c6]
Yet such thinking has dominated more than thirty years of research on the biochemistry of schizophrenia, generated endless research papers, made scientific and medical reputations, and brought incidental substantial profit to the big drug firms. The history of thinking among biochemists about schizophrenia over the period is inextricably intertwined with that of the pharmaceutical industry, for which psychotropic drugs have been one of the biggest money spinners. One in five of drugs issued in the British National Health Service in 1979 was for a drug acting on the central nervous system. Hoffmann-La Roche earns nearly $1 billion a year worldwide from its sales of Valium. It is estimated that chlorpromazine, introduced in 1952 for the control of long-stay hospitalized schizophrenics and related patients, had been administered to 50 million people worldwide within the first ten years of its use.
There is still another twist to the spiral of interdependence of the drug industry and the diagnosis of mental illness. With prolonged use of drugs, a whole new range of disorders has become apparent. Substances intended to cure one problem generate another, and the growth in such iatrogenic (medically induced) disorders is serious and disturbing. This is particularly the case for the major tranquilizers like chlorpromazine. There has been a slow recognition in the last decade or so of the disorder category known as tardive dyskinesia, apparent particularly among hospitalized patients who have been long users of chlorpromazine. The symptoms, which include characteristic motor disabilities and uncontrollable gestures (for instance, movements of the mouth), do not necessarily disappear when the patient is taken off the drug. There are reports that between 10 and 40 percent of those who regularly use major tranquilizers may suffer from tardive dyskinesia, and about 50 percent of those who get the disorder will have some irreversible consequential brain damage. Nor are there at present any drugs to combat these effects, though tardive dyskinesia has become a prolific spin-off area for neurobiological research.[c7]
It would be wearisome and unnecessary to recount in detail the history of research into the biochemistry of schizophrenia over the past thirty years. Almost every biochemical substance known to be present in the brain has, within two or three years of its introduction into the biochemical dictionary, been studied for possible involvement in schizophrenia by clinical scientists with the hope of a breakthrough in their hearts and with grant money (often from drug companies) burning holes in their pockets.
We do not in any way wish to minimize the enormous difficulties faced in clinical research. The desire for a solution to the problem of schizophrenia is real and great, and the insistence on a biological mode of explanation that will enable effective drugs to be developed is part of a pressurizing culture to which clinical research is responding. Drugs that alleviate symptoms, like the use of aspirin for toothache, may be worth developing even if they tell nothing about the causes of the disorder. The multiplicity of drugs (and formulations of drugs) is an aspect of the way the pharmaceutical companies work in a field where knowledge of patent law is as important as clinical skills. The problem is that of confounding the effect of a drug with the offer of an explanation, the alleviation of suffering with a cure for the disease.
Among the claims for causative factors in schizophrenia made since the 1950s we may point to: abnormal substances secreted in the sweat of schizophrenics; injection of the blood serum of schizophrenics into other, normal subjects inducing abnormal behavior; and the presence of abnormal enzymes in red blood cells and blood proteins. Between 1955 and the present day, conflicting research reports have claimed that schizophrenia is caused by disorders in serotonin metabolism (1953); noradrenaline metabolism (1971); dopamine metabolism (1972); acetylcholine metabolism (1973); endorphin metabolism (1976); and prostaglandin metabolism (1977). Some molecules, such as the amino acids glutamate and gamma-amino-butyric acid, came into fashion in the late 1950s, fell into neglect, and now, in the 1980s, have come back into fashion once more.[c8]
Most of the substances referred to above are brain chemicals known to play a part in the transmission of nerve impulses between cells. This points to the main idea running through all such research. The notion is that in some way, in schizophrenia, messages between cells in those regions of the brain concerned with information processing and with affect become scrambled, resulting in inappropriate responses. The evidence for any and all of the various molecular disorders is based on a combination of the types of methodologies and logic described earlier. Rarely have results obtained by one group of researchers been confirmed by another group of researchers in a different group of patients. Rarely has any resolution of conflicting claims been attempted. Rarely has any concern been expressed by the enthusiastic clinical researchers that schizophrenia might be associated with many different biochemical effects, or indeed that many different types of biochemical change might lead to or be generated by the same behavioral outcomes.
The statement that the brain of a person manifesting schizophrenia shows biochemical changes compared with that of a normal person may be no more than a reaffirmation of a proper materialism that insists on the unity of mind and brain. But the ideology of biological determinism goes much deeper than this. It is, as we have reiterated, linked to an insistence that biological events are ontologically prior to and cause the behavioral or existential events, and hence to a claim that if brain biochemistry is altered in schizophrenia, then underlying this altered biochemistry must be some type of genetic predisposition to the disorder. By 1981 psychologists were claiming to be able to detect potential schizophrenics when they are only three years old—up to fifty years before the disease manifests itself. The claim, made by Venables to a meeting of the British Association for the Advancement of Science, is based on a survey of three-year-olds in Mauritius; “potentially abnormal” children were said to show “abnormal autonomic responses.”[c9]
Push the diagnosis back beyond the three-year-old and we are soon with embryo or gene. But the hunt for a genetic basis for schizophrenia goes far beyond an interest in therapy, as there is no way in which the mere demonstration of a genetic basis for the disorder would aid in its treatment.[fn:1] As we have seen, the lineage of the effort to find genetic predispositions runs back through the eugenic thinking of the 1930s and 1920s, with its belief in genes for criminal degeneracy, sexual profligacy, alcoholism, and every other type of activity disapproved of by bourgeois society. It is deeply embedded in today’s determinist ideology. Only thus can we account for the extraordinary repetitive perseverance and uncritical nature of research into the genetics of schizophrenia. Whatever such research may say about the disorder it proposes to explain, an examination of the claims of its protagonists says a very great deal about the intellectual history of our contemporary determinist society, and hence is worth analyzing in some detail.
The belief that schizophrenia has a clear and important genetic basis is now very widely held. The father of psychiatric genetics, Ernst Rüdin, was so convinced of this that, arguing on the basis of statistics collected by his co-workers, he advocated the eugenic sterilization of schizophrenics. When Hitler came to power in 1933, Rüdin’s advocacy was no longer merely academic. Professor Rüdin served on a panel, with Heinrich Himmler as head, of the Task Force of Heredity Experts who drew up the German sterilization law of 1933. -
Perhaps the most influential psychiatric geneticist in the English-speaking world was a student of Rüdin’s, the late Franz Kallmann. The blizzard of statistics published by Kallmann seemed to indicate conclusively that schizophrenia was a genetic phenomenon. From his study of a thousand pairs of affected twins, Kallmann concluded that if one member of a pair of identical twins was schizophrenic there was an 86.2 percent chance that the other would be also. Further, if two schizophrenic parents produced a child, there was a 68.1 percent chance that the child would be schizophrenic. These figures led Kallmann to argue that schizophrenia could be attributed to a single recessive gene.
The particular genetic theory espoused by Kallmann has made it possible for latter-day psychiatric geneticists to attempt a spectacular rewriting of their history. Thus, in a recent textbook the following note appears: “Kallmann’s [theory] was apparently not based solely on his data. His widow has indicated that Kallmann advocated a recessive model because he could then argue convincingly against the use of sterilization to eliminate the gene. As a Jewish refugee, Kallmann was very sensitive to this issue and afraid of the possible social consequences of his own research.”[c10] The point here is that if a disease such as schizophrenia is caused by a recessive gene, many carriers of the gene will not themselves display symptoms, Thus, sterilization merely of those who do show symptoms would be inefficient and would fail to eliminate the disease.
The picture of Kallmann as a bleeding-heart protector of schizophrenics, adjusting his scientific theories to mirror his compassion, is grotesquely false. The first Kallmann publication on schizophrenia is in a German volume edited by Harmsen and Lohse that contains the proceedings of the frankly Nazi International Congress for Population Science.[c11] There, in Berlin, Kallmann argued vigorously for the sterilization of the apparently healthy relatives of schizophrenics, as well as of schizophrenics themselves. This was necessary, according to Kallmann, precisely because his data indicated that schizophrenia was a genetically recessive disease. Two Nazi geneticists, Lenz and Reichel, rose to argue that there were simply too many apparent healthy relatives of schizophrenics to make their sterilization feasible.
The eugenicist views of Kallmann were not confined to obscure Nazi publications but were also made widely available in English after his arrival in the United States in 1936. In 1938 he wrote of schizophrenics as a “source of maladjusted crooks, asocial eccentrics, and the lowest type of criminal offenders. Even the faithful believer in … liberty would be much happier without those… . I am reluctant to admit the necessity of different eugenic programs for democratic and fascistic communities … there are neither biological nor sociological differences between a democratic and a totalitarian schizophrenic.”[c12]
The extremity of Kallmann’s totalitarian passion for eugenic sterilization was clearly indicated in his major 1938 text. Precisely because of the recessivity of the illness, it was above all necessary to prevent the reproduction of the apparently healthy children and siblings of schizophrenics. Further, the apparently healthy marriage partner of a schizophrenic “should be prevented from remarrying” if any child of the earlier marriage is even a suspected schizophrenic, and even if the second marriage is with a normal individual.”[c13]
These views of the future president of the American Society for Human Genetics are so bloodcurdling that one can sympathize with the efforts of present-day geneticists to misrepresent or to suppress them. They have not, however, suppressed the mountains of published statistics with which Kallmann attempted to prove that schizophrenia (like tuberculosis and homosexuality) was a hereditary form of degeneracy. Those figures are presented to students in today’s textbooks as the fruits of impartial science. We begin our review of the data concerning the genetics of schizophrenia with a detailed examination of Kallmann’s work, which should make clear that Kallmann’s figures cannot be regarded seriously.
The Kallmann data were collected under two very different sets of circumstances. The earlier data, published in 1938, were based upon the records of a large Berlin mental hospital. Working with records from the period 1893-1902, Kallmann made an “unambiguous diagnosis” of schizophrenia in 1,087 index cases. To make these diagnoses it was necessary to ignore “earlier diagnoses or the contemporary notes on hereditary taint conditions in the family of the patient.” Then Kallmann attempted to locate, or to acquire information about, relatives of the index cases—many of whom were long since dead. That task often involved
formidable difficulties … we were dealing with inferior people… . They sometimes escaped our search for years… . Quite a few were bad-humored … we had to overcome the suspicion with which certain classes regarded any kind of official activity…. Whenever we encountered serious opposition we found ourselves to be dealing with either officials and members of the academic world, or people with exaggerated suspicions, schizoid types, and possible schizophrenics … our private sources of information were amplified from the records of police bureaus… . In making inquiries about people already dead or living too far away, we employed … local bureaus and trusted agents.[c14]
With information gathered in this way, Kallmann felt able to diagnose the relatives of the index cases, and thus to report the probability of schizophrenia for each type of relative. The rates reported by Kallmann in this German sample are reproduced in the left-hand column of Table 8.1. The reported rates, it should be noted, were “age-corrected.” That was necessary because some of the relatives were quite young and might develop schizophrenia as they grew older. The arbitrary correction employed by Kallmann can sometimes produce rates in excess of 100 percent.
The second set of data collected by Kallmann came from a very different sample, studied in New York State. The index cases were now individuals who were schizophrenic twins who had been admitted to public mental hospitals. When Kallmann reported in 1946, there were 794 such index cases.[c15] By 1953, the number had increased to 953. There were, of course, some identical (MZ) twins, and some fraternal (DZ) twins. Thus, by obtaining information about the co-twins of index cases, Kallmann could report the probability that both members of a pair were schizophrenic. That probability is called the “pairwise concordance rate.” The age-corrected concordances were reported for different types of twins, along with the corrected morbidity rates for various types of relatives. These had been determined by collecting information about the relatives of the twin index cases. There was virtually no information given about the procedures employed in this massive study, but Kallmann wrote that “classification of both schizophrenia and zygosity were made on the basis of personal investigation and extended observation.” This obviously allowed for “contaminated diagnosis.” That is, the decision as to whether or not a co-twin was said to be schizophrenic could be influenced by the decision as to whether the twin pair was MZ or DZ and vice versa. The Kallmann 1946 data, and the even more sketchily reported data of 1953, are also presented in Table 8.1.
[t1]
These data are obviously consistent with an overwhelming genetic determination of schizophrenia—particularly the remarkable rate of 86 percent among MZ twins. Where direct comparisons can be made, the change of countries and of eras—as well as the switch to relatives of twin index cases—has had little effect on the reported figures.
The correspondence between Kallmann’s theoretical expectations and the results he discovered is sometimes quite remarkable. Thus, in 1938 Kallmann indicated that the work of earlier twin researchers suggested that schizophrenia manifested itself, even among those with the full genetic predisposition, only about 70 percent of the time.[c17] That meant, according to Kallmann’s single recessive gene theory, that 70 percent of the children of two schizophrenic parents should themselves be schizophrenic. The Kallmann data indicated that the expectation of schizophrenia in the offspring of two schizophrenics was precisely 68.1 percent. That result, of course, nicely validated Kallmann’s theory. Four other studies of the children of two schizophrenic parents suggest a risk of only between 34 and 44 percent.[c18]
Kallmann stressed repeatedly that, in his data, the “morbidity figure for the siblings … corresponds perfectly with the concordance rate for two-egg twin pairs, whose chance of inheriting a similar genotypical combination is exactly the same as that for any ordinary pair of brothers and sisters”.[c19] The same close correspondence was described as a notable finding in 1953. We shall soon see, however, that—as an embarrassment to a simple genetic theory—other investigators have not found the close correspondences of data with theory routinely detected by Kallmann.
There are many similarities between the roles of Franz Kallmann in schizophrenia research and of Cyril Burt in IQ research. The two men each believed passionately in the genetic determination of human behavior. While Kallmann fulminated against the dysgenic threat posed by schizophrenics, Burt—also a eugenicist—was deeply concerned by the threat of dysgenic reproduction by low-IQ people. The two men each gathered by far the most massive sets of data collected in their fields. The two men each failed to describe with any adequacy at all their methods and procedures. The results reported by each were incredibly consistent with simple genetic theories—far more so than the data collected by other investigators. That happy coincidence enabled Kallmann to argue for “eugenic-prophylactic measures” against the families of the mentally ill, as it enabled Burt to argue against wasting educational resources on those with low IQ scores. As we showed in Chapter 5, there is now universal agreement that Burt’s data were fraudulent and must be discarded. The same, however, is not true of Kallmann’s data. In fact, they have been defended vigorously against untoward insinuations. As Shields and his colleagues put it, these could only be made because “the abbreviated manner in which Kallmann reported his results left him more open than he would otherwise have been to criticism.”[c20]
The research conducted by others who followed Kallmann has in any event made it clear that his extraordinarily high figures cannot be repeated. The Kallmann data are still presented, unblushingly, in purportedly serious reviews of research, but they are now counterbalanced by more recent and more modest results. Perhaps the chief harm brought about by Kallmann’s deluge of incredible and poorly documented data was to create a climate in which the findings of subsequent workers seemed so reasonable and moderate that they escaped serious critical scrutiny. Thus, Kallmann’s data have faded from the body of acceptable evidence, but the belief for which he was largely responsible—that a genetic basis for schizophrenia has been clearly established—still remains powerful in and out of science.
There are basically three kinds of inquiries that attempt to demonstrate a genetic basis for schizophrenia: family studies, twin studies, and adoption studies. There is no need to spend much time on the first. The simple idea behind them is that if schizophrenia is inherited, the relatives of schizophrenics are likely to display the disease as well. Further, the more closely related a person is to a schizophrenic, the more likely it should be that the person will be affected. The problem is, of course, that these predictions would also follow from a theory that maintained that schizophrenia was environmentally produced. There is an obvious tendency for close relatives to share similar environments.
For what such data are worth, the major compilation of family studies seems to have been made by Zerbin-Rüdin.[c21] The compilation was presented to English-readers in “simplified form” by Slater and Cowie.[c22] Their table indicates, e.g., that fourteen separate studies yield a 4.38 percent expectation of schizophrenia among the parents of schizophrenic index cases. The expectation among sibs, in ten studies, was 8.24 percent; and among children, 13.31 percent in five studies. For uncles and aunts, grandchildren, and cousins the figures were all under 3 percent, but still higher than the expected 1 percent.
The exactness of these figures, however, is more apparent than real. The same basic set of studies was also summarized by Rosenthal in 1970.[c23] The relatives diagnosed in these studies, Rosenthal noted, had often been dead for many years. The studies are quite old, and methods of diagnosis and of sampling are not always spelled out. The combined figures are dominated by Kallmann’s massive samples and by data gathered by other members of Rüdin’s “Munich school.” The Rosenthal tables make clear a fact that is obscured by the Slater and Cowie summary. There are vast differences in the rates of schizophrenia reported in different studies. For parents of index cases, reported risks range from 0.2 percent (lower than in the population at large) to 12.0 percent. For sibs, the range is between 3.3 and 14.3 percent. The risk for sibs is in one study twenty-nine times larger than that for parents; but in another the risk for parents is 1½ times larger than that for sibs. These studies at best demonstrate what nobody would have contested. There is at least a rough tendency for diagnosed schizophrenia to “run in families.”[fn:2]
As described in Chapter 5, the basic logic of twin studies depends upon the fact that while MZ twins are genetically identical, DZ twins on average share (like ordinary siblings) only half their genes. Thus, if a trait is genetically determined, one would obviously expect MZs to be concordant for that trait more often than DZs. The major logical problem with twin studies is that MZ twins, who typically resemble one another strikingly in appearance, are treated much more similarly than are DZs by parents and peers. There is abundant evidence (discussed in Chapter 5) that the environments of MZs are very much more similar than those of DZs. (Twin studies typically compare concordance rates among MZs, who are always of the same sex, with concordance rates among same-sexed DZs.) The demonstration that concordance is higher among MZs does not necessarily establish a genetic basis for the trait in question. Perhaps the difference is due to the greater environmental similarity of MZs. We shall soon discuss evidence which indicates that this possibility is not at all farfetched.
Well-designed twin studies should take as their index cases all schizophrenic twins admitted to a particular hospital during a particular time period. The alternative—feasible in small Scandinavian countries, which maintain population registers—is to start with the entire population of twins and to locate index schizophrenic cases. With either technique, a number of procedural problems are inevitable. The co-twins of index cases are often dead or unavailable for personal examination. Thus, informed guesses often must be made both about whether a given pair is MZ or DZ, and whether or not the co-twin is schizophrenic. The guesses are typically made by the same person, opening the way for contaminated diagnoses. There is sometimes an effort to have blind diagnoses made of individual cases by independent judges, working from written case histories.[c25]
The case histories, however, contain selective material gathered and prepared by investigators who were not themselves “blind.” Further, the case records of those twins who have in fact been hospitalized—and their diagnoses—had been written up by doctors who questioned the ill twins in detail about possible taint in their family lines. The diagnosis of schizophrenia, as should by now be clear, is by no means a cut-and-dried affair. The fact that a person’s relative may have suffered from schizophrenia is often used to help doctors make a diagnosis.
The biases that contaminate twin studies stand out clearly from an attentive reading of the published case history materials. The very first case described by Slater in 1953 is the story of Eileen, a hospitalized schizophrenic, and of her identical twin, Fanny. Eileen had been hospitalized in 1899, “suffering from acute mania,” and died in the hospital in 1946. With Eileen as the index case, Slater’s task was to investigate the mental status of Fanny, who died, aged seventy-one, in 1938. We are told by Slater:
While still in the twenties she had a mental illness, of which no details are available… . Fanny in (1936) proved very difficult to examine … so that only the barest details were obtainable. She suppressed all mention of her own mental illness in early years, which fact was obtained from the history of her twin sister given at the time of her admission to hospital. Though there was no sign of any present schizophrenic symptoms, this suspicion and reserve are such as are commonly found as sequelae of a schizophrenic psychosis. Unfortunately, no facts are obtainable about the nature of her past mental illness, but the probabilities are very greatly in favour of it having been a schizophrenic one . . . she made a fairly complete and permanent recovery … though psychologically her reserve and lack of frankness suggest that the schizophrenia was not entirely without permanent after-effect… . According to her daughter-in-law, who had not heard of her mental illness, she led a hard life. Neither her family nor the neighbours noticed anything odd about her.[c26]
These MZ twins, according to Slater, were concordant for schizophrenia. The only evidence that Fanny had once suffered from schizophrenia was her twin’s assertion—while “suffering from acute mania” in 1899—that Fanny had had some kind of mental illness. Fanny herself, in 1936, was difficult and suppressed all mention of her illness. That lack of frankness, Slater noted, was typical of recovered schizophrenics, who otherwise appear normal. Fanny’s dead identical twin had clearly been schizophrenic. For Slater this made it obvious that Fanny’s supposed mental illness fifty years earlier had been schizophrenia. Fanny’s neighbors and family, unlike Slater and other students of the Munich school, had not the wit to detect Fanny’s schizophrenia.
Consider now the first pair of discordant DZ twins described by Gottesman and Shields in their 1972 study. Twin A was a hospitalized schizophrenic. What about Twin B? “No psychiatric history. Family unwilling for him to be contacted for Twin Investigation… . The pair differs from most in that neither twin was seen by us.” The investigators concluded that Twin B was normal; and six blind judges, pondering a case study summary prepared by the investigators, unanimously agreed that Twin B was free of psychopathology. With DZ Pair 16 of the same study, all judges again agreed that the co-twin was normal, making the pair discordant. The diagnosis of the co-twin had not been made under ideal conditions: “He refused to be seen for the Twin Investigation, remaining upstairs out of sight, but his wife was seen at the door… . He was regarded as a healthy, levelheaded, solid happy person.” That might in fact be the case—but few will agree that diagnoses of co-twins made in this way are solid or levelheaded.
Problems of this sort affect all twin studies, and that should be borne in mind as we review the results reported by various investigators. To obtain reasonable estimates of concordance rates, it seems sensible to require that a study contain at least twenty pairs of MZ and twenty pairs of same-sexed DZ twins. There have been seven such studies, and their results are summarized in Table 8.2.
The table presents raw, pairwise concordance rates, without any age correction. Two sets of rates are given for each study, one narrow and one broad. The narrow rates are based on the investigator’s attempt to apply a relatively strict set of criteria when diagnosing schizophrenia. The broad rates include as concordant cases in which one twin is described as “borderline schizophrenic” or as “schizo-affective psychosis” or a “paranoid with schizophrenic-like features.” The tabled concordance rates, it should be noted, depend upon the different investigators’ varying sets of diagnostic criteria. They have not been concocted ad hoc by us.
[t2]
The table makes clear that in all studies concordance is higher for MZ than for DZ twins. But it is also clear that the concordance reported for MZs is much higher in the three older studies than in the four more recent ones. There is in fact no overlap between the two sets of studies. For narrow concordance, the average has plunged from 56 to 26 percent for MZs; for DZs, the corresponding averages are 11 and 9 percent. For broad concordance, MZ rates have dropped from 65 to 42 percent, while the DZ rate remained at a constant 53 percent. These average values, which weight all studies equally, should not be taken too literally. The data do make clear, however, that even in genetically identical MZs environmental factors must be of enormous importance. The concordance for MZs reported by modern researchers, even under the broadest criteria, does not remotely approach the preposterous 86 percent figure claimed by Kallmann.
Those who perform such studies still claim, however, that the higher concordance observed among MZs—a unanimous finding—demonstrates at least some genetic basis for schizophrenia. We have already noted that MZs not only are genetically more similar than DZs but also experience much more similar environments than do DZs. The environmental similarity, no less than the genetic similarity, might plausibly account for the higher concordance of MZs.
There are in fact some simple and critical tests that can be made of this environmental hypothesis. There is no doubt that DZ twins experience more similar environments than do ordinary siblings. The DZ twins, however, are genetically no more alike than are ordinary siblings—they are only siblings who happen to have been born at the same time. Thus, from an environmental viewpoint—and only from such a viewpoint—we would expect concordance among DZs to be higher than among ordinary sibs. There have been a number of studies that reported rates of schizophrenia concordance among DZ twins, as well as rates among siblings of the twins. The results of all such studies are summarized in Table 8.3.
Though the reported differences are very small in the early studies, all studies agree in showing a higher concordance rate among DZs than among sibs. Within more modern studies, the difference is often statistically significant, with the risk for DZs reported as two or three times that for sibs. When we note that similarity of environment can double or triple the concordance of DZs above that of sibs, it seems entirely plausible to attribute the still higher concordance of MZs to their still greater environmental similarity.
The same kind of point can be demonstrated by comparing the concordance rates of same-sexed and of opposite-sexed DZs. Though both types of DZ twins are equally similar genetically, it is obvious that same-sexed pairs experience more similar environments than do opposite-sexed pairs. The available data, summarized in Table 8.4, again support the environmentalist expectation. There have been statistically significant differences reported by several investigators, always indicating a higher concordance among same-sexed twins. The results of the one study that appears to reverse the otherwise universal trend were not statistically significant.
Consider, finally, some implications of a finding casually reported by Hoffer and Pollin.[c35] Those authors studied the hospital records of the American war veteran twins later reported on by Allen et al. Several hundred diagnosed schizophrenic twins were located by searching through records, but the twins were not personally examined by the investigators. Thus, to determine whether a twin pair was MZ or DZ, questionnaires were mailed to all twins, asking whether they looked as much alike as two peas in a pod, whether they were confused for each other, etc. There were many occasions when only one twin of a discordant pair returned the questionnaire. When the twin returning the questionnaire had been diagnosed as schizophrenic, 31.3 percent gave answers indicating that they were MZ. When the answering twin was not the diagnosed schizophrenic, only 17.2 percent indicated that they were MZ. The difference is statistically significant, and it was produced by an unrealistically small proportion of MZs among the nonschizophrenic twins.
That is easily understandable. When you are normal and your twin is schizophrenic, you are well advised to tell twin investigators and other authorities that you are not a carbon copy of your twin—even if you really are MZs. To admit that you are the MZ twin of a schizophrenic is clearly to invite a similar diagnosis—even, perhaps, sterilization—for yourself. We recall that in all the twin studies some decisions about zygosity are made on the basis of questions put to nonaffected twins and to their relatives. With a little sensitivity to the real lives of people, we must recognize an all-too-human tendency to deny that the nonaffected MZ twins of schizophrenics really are identical. This must be still another source of error, tending to remove some discordant pairs from the MZ and into the DZ category. That, of course, artificially inflates the difference in concordance rates between MZs and DZs. There is little wonder in the fact that even psychiatric geneticists have not found twin studies to be wholly convincing, and have turned to studies of adoption. The adoption studies, in theory at least, might be able to disentangle genetic from environmental effects in a way that twin studies cannot.
The basic procedure of adoption studies is to begin with a set of schizophrenic index cases, and then to study the biological relatives from whom they have been separated by the process of adoption. Thus—at least in theory—the index case and his or her biological relatives have only genes, and not environment, in common. The question of interest is whether the biological relatives of the index cases, despite the lack of shared environments, display an increased incidence of schizophrenia. To answer that question it is necessary to compare the rate of schizophrenia among the biological relatives with the rate observed in some appropriate control group.
The adoption studies carried out in Denmark in recent years by a collaborative team of American and Danish investigators have had enormous impact. To some critics who could detect the methodological weaknesses of twin studies, the Danish adoption studies appeared to establish the genetic basis of schizophrenia beyond any doubt. The eminent neuroscientist Solomon Snyder referred to these studies as a landmark “in the history of biological psychiatry. It’s the best work that’s been done. They take out all the artifacts in the nature vs. nurture argument.”[c36] Paul Wender, one of the authors of the studies, was able to announce: “We failed to discover any environmental component… . That’s a very strong statement.”[c37] Though Wender’s total excision of environmental factors is extreme, the Danish studies have been universally accepted as an unequivocal demonstration of an important genetic basis for schizophrenia. Clearly these studies require detailed critical examination.
Though they have been described in many separate publications, there are basically two major Danish adoption studies. The first, with Kety as senior investigator, starts with adoptees as the schizophrenic index cases and examines their relatives. The second, with Rosenthal as senior investigator, starts with schizophrenic parents as index cases and examines the children whom they gave up for adoption.
The study that began with adoptees as index cases was first reported by Kety in 1968.[c38] Based on Copenhagen records, the investigators located thirty-four adoptees who had been admitted to psychiatric hospitals as adults and who could be diagnosed from the records as schizophrenics. For each schizophrenic adoptee a control adoptee who had never received psychiatric care was selected. The control was matched to the index case for sex, age, age at transfer to the adoptive parents, and socioeconomic status (SES) of the adoptive family.
The next step was to search the records of psychiatric treatment for all Denmark, looking for relatives of both the index and control cases. Those who searched the records did not know which were the relatives of index cases and which were the relatives of controls. Whenever a psychiatric record was found, it was summarized and then diagnosed blindly by a team of researchers who came to a consensus. The relatives were not at this stage personally examined.
The researchers traced 150 biological relatives (parents, sibs, or half-sibs) of the index cases, and 156 biological relatives of the controls. The first point to note is one not stressed by the authors: There were virtually no clear cases of schizophrenia among the relatives either of the index or of the control cases. To be precise, there was one chronic schizophrenic among the index relatives and one among the controls. To obtain apparently significant results the authors had to pool together a “schizophrenic spectrum of disorders.” The spectrum concept lumps into a single category such diagnoses as chronic schizophrenia, “borderline state,” “inadequate personality,” “uncertain schizophrenia,” and “uncertain borderline state.” With such a broad concept, 8.7 percent of the biological relatives of index cases and 1.9 percent of the biological relatives of controls were diagnosed as displaying spectrum disorders. There were nine biological families of index cases in which at least one spectrum diagnosis had been made, compared to only two such families among the controls. That difference is the supposed evidence for the genetic basis of schizophrenia. Without the inclusion of such vague diagnoses as “inadequate personality” and “uncertain borderline schizophrenia” there would be no significant results in the Kety study.
From the Kety data of 1968 it is possible to demonstrate that such vague diagnoses—falling within the “soft spectrum”—are not in fact associated with schizophrenia. Among the sixty-six biological families reported on in 1968 there were a total of six in which at least one “soft” diagnosis had been made.[fn:3] There was no tendency for such diagnoses to occur any more frequently in families in which definite schizophrenia had been diagnosed than in other families. However, the “soft spectrum” diagnoses very definitely tended to occur in the same families in which “outside the spectrum” psychiatric diagnoses had been made—that is, such clearly nonschizophrenic diagnoses as alcoholism, psychopathy, syphilitic psychosis, etc. There were “outside the spectrum” diagnoses in 83 percent of the families containing “soft spectrum” diagnoses, and in only 30 percent of the remaining families—a statistically significant difference. Thus it appears that the Kety et al. results depend upon their labeling as schizophrenia vaguely defined behaviors that tend to run in the same families as do alcoholism and criminality—but which do not tend to run in the same families as does genuine schizophrenia. However, it remains the case that these frowned-upon behaviors did occur more frequently among the biological relatives of adopted schizophrenics than among the biological relatives of adopted controls. What might account for such a finding?
The most obvious possibility is that of selective placement, a universal phenomenon in the real world in which adoptions in fact occur, and a phenomenon that undermines the theoretical separation of genetic and environmental variables claimed for adoption studies. The children placed into homes by adoption agencies are never placed randomly. For example, it is well known that biological children of college-educated mothers, when put up for adoption, are placed selectively into the homes of adoptive parents with higher socioeconomic and educational status. The biological children of mothers who are grade-school dropouts are usually placed into much lower status adoptive homes. Thus it seems reasonable to ask: Into what kinds of adoptive homes are infants born into families shattered by alcoholism, criminality, and syphilitic psychosis likely to be placed? Further, might not the adoptive environment into which such children are placed cause them to develop schizophrenia?
From raw data kindly made available to one of us by Dr. Kety, we have been able to demonstrate a clear selective placement effect. Whenever a record of psychiatric treatment of a relative was located by Kety’s team, notation was made about whether the relative had been in a mental hospital, in the psychiatric department of a general hospital, or in some other facility. When we check the adoptive families of the schizophrenic adoptees, we discover that in eight of the families (24 percent) an adoptive parent had been in a mental hospital. That was not true of a single adoptive parent of a control adoptee. That, of course, is a statistically significant difference—and it suggests as a credible interpretation of the Kety et al. results that the schizophrenic adoptees, who indeed had been born into shattered and disreputable families, acquired their schizophrenia as a result of the poor adoptive environments into which they were placed. The fact that one’s adoptive parent goes into a mental hospital clearly does not bode well for the psychological health of the environment in which one is reared. There is, by the way, no indication that the biological parents of the schizophrenic adoptees have been in mental hospitals at an excessive rate. That occurred in only two families (6 percent), a rate in fact lower than that observed in the biological families of the control adoptees.
The same set of subjects has also been reported on in a later paper by Kety et al.[c39] For this later work as many as possible of the relatives of index and control adoptees had been traced down personally and interviewed by a psychiatrist. The interviews were edited, and consensus diagnoses were then made blindly by the investigators. The basic picture did not change much. There were more spectrum diagnoses among relatives of index cases than among relatives of controls, although the interview procedure greatly increased the overall frequency of such diagnoses. This time, however, diagnoses of inadequate personality had to be excluded from the spectrum, since they occurred with equal frequency in both sets of relatives. The significance of the 1968 results, based on records rather than interviews, had depended upon including inadequate personality in the elastic spectrum.
Personal correspondence with the psychiatrist who conducted the interviews with relatives has revealed a few interesting details. The 1975 paper speaks only of “interviews,” but it turns out that in several cases, when relatives were dead or unavailable, the psychiatrist “prepared a so-called pseudo interview from the existing hospital records.” That is, the psychiatrist filled out the interview form in the way in which he guessed the relative would have answered. These pseudo interviews were sometimes diagnosed with remarkable sensitivity by the team of American investigators. The case of the biological mother of S-11, a schizophrenic adoptee, is one particularly instructive example.
The woman’s mental hospital records had been edited and then diagnosed blindly by the investigators in 1968. The diagnosis was inadequate personality—at that time, inside the spectrum. The 1975 paper—by which time inadequate personality is outside the spectrum—indicates that, upon personal interview, the woman had been diagnosed as a case of uncertain borderline schizophrenia—again inside the spectrum. But personal correspondence has revealed that the woman was never in fact interviewed; she had committed suicide long before the psychiatrist attempted to locate her, and so—from the original hospital records—she was “pseudo interviewed.” Perhaps the most remarkable aspect of the story, also revealed by personal correspondence, is that the woman had been hospitalized twice—and each time had been diagnosed as manic-depressive by the psychiatrists who actually saw and treated her. That is, she had been diagnosed as suffering from a mental illness unrelated to schizophrenia, and very clearly outside the schizophrenia spectrum. We can only marvel at the fact that the American diagnostcians, analyzing abstracts of these same records, were twice able to detect—without ever seeing her—that she really belonged within the shifting boundaries of the spectrum.
The Kety study has more recently been expanded to include all of Denmark (rather than merely Copenhagen). The hospital records of relatives have been searched and the results briefly referred to in a couple of publications. The relatives are also being interviewed. There have been no detailed data published or made available for the larger sample, so critical analysis is not yet possible. Though Kety asserts that results from the expanded sample confirm those earlier reported in detail, there is no reason to suppose that the more recent work is free of the invalidating flaws we have outlined above.
These results must be evaluated together with the results of a companion study reported by Rosenthal et al. using the same Danish files.[c40] This study first identified a number of schizophrenic parents who had given up children for adoption. The question is whether those children, not reared by their schizophrenic biological parents, will tend to develop schizophrenia. The control group for the index children was made up of adoptees whose biological parents had no record of psychiatric treatment. The index adoptees and the controls, when grown up, were interviewed—blindly—by a Danish psychiatrist. Based upon those interviews, decisions were made as to whether particular individuals were in or out of the spectrum of schizophrenic disorders. Countless textbooks now indicate that a higher frequency of spectrum disorders were diagnosed in the adopted children of schizophrenics than in children of normal controls. That claim is based on preliminary (and inadequately reported) accounts of the study.
The preliminary reports did claim to observe a barely significant tendency for spectrum disorders to be more frequent among the index cases. (There was only one adoptee who had ever in fact been hospitalized for schizophrenia, and the authors frankly admitted that if they had looked only for hospitalized cases of schizophrenia, “we would have concluded that heredity did not contribute significantly to schizophrenia.”)[c41] The early papers, however, are entirely vague as to when and how or by whom decisions were made about whether individual cases were in or out of the spectrum. The papers indicate merely that the interviewing Danish psychiatrist made a “thumbnail diagnostic formulation” for each interview, and that these were somehow related to whether or not the interviewee was placed into the spectrum. Personal correspondence with several of the collaborators has made it clear that the “thumbnail diagnostic formulation” of the interviewer did not specify whether the individual was in or out of the spectrum. For the early papers, that decision was made in a manner and by parties unknown.
When consensus diagnoses like those in the Kety study were reported on for the first time in 1978, it developed that there was no significant tendency for spectrum cases to occur more frequently among index subjects.[c42] Thus, despite the widely cited misleading early reports of the Rosenthal et al. study, its outcome was in fact negative.
Wender et al. added a new refinement to the Rosenthal study by reporting on a new group of twenty-eight “cross-fostered” subjects.[c43] These were adoptees whose biological parents had been normal but whose adoptive parents had become schizophrenic. The new group was added to observe whether the experience of being reared by a schizophrenic adoptive parent would produce pathology in a child. The cross-fostered children, according to Wender et al., did not show more pathology than did the control adoptees. But it is important to note that in this paper the concept of diagnosing a schizophrenia spectrum has been abandoned; instead, the Danish interviews were now being rated for “global psychopathology.” Consensus diagnoses—or any other diagnoses—of whether or not the cross-fostered children were in the schizophrenia spectrum have not appeared in any of the many papers concerned with the genetics of schizophrenia.
There is, however, an obscure paper from the Kety and Rosenthal group concerned with the characteristics of people who refuse to take part in psychological studies that contains some important and relevant information.[c44] The paper includes as an aside an incidental table (Table 14) showing the percentage of spectrum diagnoses made in each group by a Danish psychiatrist, Schulsinger. We learn from that table that fully 26 percent of the cross-fostered adoptees were diagnosed as being in the schizophrenia spectrum—a rate not significantly different from that of the index adoptees themselves. Further, that obscure table is the only place where data on an immensely relevant control group have been reported. The Danish investigators, it turns out, also interviewed (and diagnosed) a number of nonadopted children of schizophrenics, who had been reared by their mentally ill biological parents. The rate of spectrum disorder among this group did not differ from that observed among cross-fostered children. Thus, had they taken the design of their own study seriously, the investigators might have concluded that they had shown schizophrenia to be entirely of environmental origin. The cross-fostered biological children of normal parents, when merely reared by schizophrenic adoptive parents, show just as great a frequency of spectrum disorders as do the nonadopted biological children of schizophrenics. The reader may not be surprised to learn that consensus diagnoses of the nonadopted group, like consensus diagnoses of the cross-fostered group, have never been reported.
The weaknesses of the Danish adoption studies are so obvious upon critical review that it may be difficult to understand how distinguished scientists could have regarded them as eliminating all the artifacts that beset family and twin studies of nature and nurture. In fact, a team of investigators from the French National Institute of Medical Research have published, quite independently, an analysis of the Danish adoption studies that reaches the conclusion that they are gravely deficient.[c45] Perhaps one factor encouraging the usually uncritical acceptance of the investigators’ claims has been indicated by Wender and Klein in an article written for the popular magazine Psychology Today.[c46] They cite the Danish adoption study—based upon a broad concept of schizophrenia spectrum—as indicating that “for each schizophrenic there may be 10 times as many people who have a milder form of the disorder that is genetically … related to the most severe form … 8 percent of Americans have a lifelong form of personality disorder that is genetically produced. This finding is extremely important.” The importance of the finding is spelled out by Wender and Klein in the following language: “The public is largely unaware that different sorts of emotional illnesses are now responsive to specific medications and, unfortunately, many doctors are similarly unaware.” The logic, erroneous at every step, is as follows: The Danish adoption studies have shown that schizophrenia, and a number of behavioral eccentricities, are genetically produced. Since the genes influence biological mechanisms, it must follow that the most effective treatment for schizophrenia, and for behavioral eccentricity, is drug treatment. Focusing on social or environmental conditions as a cause of disordered behavior would be fruitless.
Yet any materialist understanding of the relationship of brain to behavior must recognize that even if schizophrenia were largely genetic in origin, it would in no way follow that drugs—or any biological, as opposed to social, treatment—would necessarily be the most effective therapy. Just as drugs change behavior, so will altered behavior imposed by talking therapies change brains (as indeed the latent theory behind behavior modification would itself agree). The logic of this does not depend on a belief in any more explicit integration of the biological and the social.
To reveal, as we have tried, the theoretical and empirical impoverishment of the conventional wisdom of biological determinism in relationship to schizophrenia does not then argue that there is nothing relevant to be said about the biology of the disorder, and still less does it deny that schizophrenia exists. The problem of understanding the etiology of schizophrenia and a rational investigation of its treatment and prevention is made vastly more difficult, perhaps even hopelessly tangled, by the extraordinary latitude and naiveté of diagnostic criteria. Certainly one may wonder about the relevance of biology to the diagnosis of schizophrenia either by the forensic psychiatrists of the Soviet Union or by the British psychiatrist who diagnoses a young black as schizophrenic on the basis of his use of the religious language of Rastafarianism.[c47]
Misgivings are not eased when one recalls a well-known study by Rosenhan and his colleagues in California in 1973.[c48] Rosenhan’s group of experimenters presented themselves individually at mental hospitals complaining of hearing voices. Many were hospitalized. Once inside the hospital, according to the strategy of the experiment, they declared that their symptoms had ceased. However, it did not prove so easy to achieve release. The experimenters’ claims to normality were disregarded, and most found themselves treated as mere objects by nurses and doctors and released only after considerable periods of time. A pseudo-patient who took notes in one of the hospitals, for instance, was described by nurses as showing “compulsive writing behavior.” Even more revealing, perhaps, was the drop in hospital admissions for schizophrenia in the area after Rosenhan circulated the results of the first experiment among doctors and indicated that they might be visited by further pseudo-patients in the future, although none were actually sent.
It is this sort of experience that lies behind the argument, developed in its most extreme form by Michel Foucault and his school over the last two decades, that the entire category of psychological disorders is to be seen as a historical invention, an expression of power relationships within society manifested within particular families. To simplify Foucault’s intricate argument, he claims that all societies require, a category of individuals who can be dominated or scapegoated, and over the centuries since the rise of science—and particularly since the industrial revolution of the nineteenth century—the mad have come to fill this category. In medieval times, he says, houses of confinement were built for lepers, and madness was often explained in terms of possession by demons or spirits.[c49] According to Foucault the idea of institutionalizing the mad developed during the eighteenth and nineteenth centuries after the clearing of the leper houses left a gap for new scapegoats to replace the old ones.
In this view madness is a matter of labeling; it is not a property of the individual but merely a social definition wished by society on a proportion of its population. To look for correlates of madness in the brain or the genes is therefore a meaningless task, for it is not located in the brain or the individual at all. To dismiss the suffering and the deranged behavior of the schizophrenic merely as a problem of social labeling by those who have power over those who have not seems a quite inadequate response to a complex social and medical problem. Despite Foucault’s historiography and the enthusiasm of its reception in Britain and France at the crest of the wave of antipsychiatry of the 1960s and 1970s, the actual historical account he gives of when and how asylums for the insane arose has been called into question.[c50] And by cutting the phenomenon of schizophrenia completely away from biology and locating it entirely in the social world of labeling, Foucault and his followers arrive, from a very different starting point, back in the dualist Cartesian camp, which, as we showed in Chapters 2 and 3, preceded the full-blown materialism of the nineteenth century. So much has Foucault retreated that at certain points in his argument he even seems to be ambiguous as to whether “physical” quite apart from “mental” illness exists except in the social context that proclaims it.
More modest than Foucault’s grand theorizing but nonetheless culturally determinist are the social and familial theories of schizophrenia developed by R. D. Laing.[c51] For Laing—at least the Laing of the sixties and early seventies—schizophrenia is essentially a family disorder, not a product of a sick individual but of the interactions of the members of a sick family. Within this family, locked together by the nuclear style of living of contemporary society, one particular child comes to be picked upon, always at fault, never able to live up to parental demands or expectations. Thus the child is in what Laing calls (in a term derived from Gregory Bateson) a double bind: whatever he or she does is wrong. Under such circumstances the retreat into a world of private fantasy becomes the only logical response to the intolerable pressures of existence. Schizophrenia is thus a rational, adaptive response of individuals to the constraints of their life. Treatment of the schizophrenic by hospitalization or by drugs is therefore not seen as liberation from the disease but as part of that person’s oppression.
Family context may be crucial in the development of mental illnesses such as schizophrenia, but it is clear that a larger social context is also involved. The diagnosis is made most often of working-class, inner-city dwellers, least often of middle- and upper-class suburban dwellers.[c52] To a social theorist, the argument about the social context that determines the diagnosis is clear. An example of the class nature of the diagnosis of mental illness comes from the studies of depression by Brown and Harris in 1978 in Camberwell, an inner-city, largely working-class area of London, with some pockets of middle-class infiltration.[c53] They showed that about a quarter of working-class women with children living in Camberwell were suffering from what they defined as a definite neurosis, mainly severe depression, whereas the incidence among comparable middle-class women was only some 6 percent. A large proportion of these depressed individuals, who if they had attended psychiatric clinics would have been diagnosed as ill and medicalized or hospitalized, had suffered severe threatening events in their lives within the past year, such as loss of husband or economic insecurity. The use of drugs—mainly tranquilizers—among such groups of women is clearly very high.
Biological determinism faces such social evidence with arguments that, for example, people with genotypes predisposing toward schizophrenia may drift downward in occupation and living accommodation until they find a niche most suited to their genotype. But it would be a brave biological determinist who would want to argue that in the case of the depressed housewives of Camberwell it was their genes that were at fault.[fn:4]
An adequate theory of schizophrenia must understand what it is about the social and cultural environment that pushes some categories of people toward manifesting schizophrenic symptoms; it must understand that such cultural and social environments themselves profoundly affect the biology of the individuals concerned and that some of these biological changes, if we could measure them, might be the reflections or correspondents of that schizophrenia with the brain. It may well be that, in our present society, people with certain genotypes are more likely than others to suffer from schizophrenia—although the evidence is at present entirely inadequate to allow one to come to that conclusion. This says nothing about the future of “schizophrenia” in a different type of society, nor does it help us build a theory of schizophrenia in the present. Neither biological nor cultural determinism, nor some sort of dualistic agnosticism, is adequate to the task of developing such a theory. For that, we must look to a more dialectical understanding of the relationship between the biological and the social.
[fn:1]: These words were true when we wrote them. However, reductionist science moves faster than the Gutenberg technology of book production. For if it were the case that there were schizophrenia-producing genes, then techniques that excised those abnormal genes from the genome of affected individuals and replaced them with their normal alleles would presumably prevent the expression of the disorder, If schizophrenia were a single or even a two- or three-gene defect, such techniques are not wholly beyond the reach of contemporary molecular genetics—what is sometimes called genetic engineering. There are serious research programs now under way in several laboratories to make gene libraries from schizophrenics and isolate and clone the “schizophrenic genes” with a view to studying their possible replacement. Granted the reductionist premise, the therapeutic logic would be impeccable. And if one can have schizophrenic urine, why not, indeed, schizophrenic genes?
[fn:2]: Even this modest conclusion is not unchallenged in the literature. Two studies in the United States found rates of schizophrenia among the first-degree relatives of schizophrenics which were scarcely above the rare in the general population.[c24]
[fn:3]: We here include as “soft” diagnoses the two least certain diagnoses employed by Kety et al.—their D-3 diagnosis (“uncertain borderline™) and their C diagnosis (“inadequate personality”).
[fn:4]: Brave but not impossible. In 1979 B. L. Reid and his colleagues published a paper in the Australian Medical Journal claiming that the higher incidence, of uterine cancer among working~class women was due to a factor carried in the sperm of their working-class male partners, and that the same working-class sperm had a simpler, more repetitive structure to its DNA than did that of middle-class sperm. This accounted for working-class people only being able to think simple and repetitive thoughts, unlike the complexity available to the middle classes.[c54] To such preformationist thinking, clearly, no sort of biological determinism is impossible.