Schizophrenia research embodies a microcosm of the vexing problems that confront the behavioural sciences, but particularly those disciplines concerned with psychopathology. We are ignorant of the means to prevent schizophrenia because we continue to be ignorant about its aetiology. Despite recognizable descriptions of the syndrome in ancient Hindu treatises (c. 1400 b.c.) and 76 years after its designation as dementia praecox by Kraepelin (1896), we are still grappling with such basic issues as when and how to diagnose Eugen Bleuler's (1911) ‘group of schizophrenias' (cf. Katz, Cole, and Barton, 1968). Despite brilliant advances in molecular biology, neurochemistry, and brain-behaviour phenomena generally, we cannot pinpoint any necessary biological defect in all or most schizophrenics. Despite selfless expenditures of time and energy by gifted psychotherapists and sophisticated social science efforts, we cannot specify any necessary life experience, either at the level of the family or of a culture, common to all or most schizophrenics. While there is obvious merit in casting the problem in an interactionist framework, aetiology still defies an easy solution because we must then isolate what element(s) in the genotype interact with what element(s) in the internal and/or external environment (as well as when and how) to produce the phenotype we recognize as a schizophrenic one.

It has long been my wish to review the early history of the Maudsley Laboratories, which were my working home for many years. While I was on their active staff, however, I was forced to postpone this plan in favour of more pressing problems.

In this paper I shall discuss certain problems which arise in the theory of diagnosis and argue that there is a case for applying historical methods, as distinguished from scientific methods, as a possible approach towards solving them; and attempt also to show that such an approach may have value for generating research as well as being of practical use in clinical work.

It is ironic but not unexpected that, at a time when psychiatry is moving out of isolation and into closer proximity with other disciplines in general hospitals, the relationship of psychiatry with medicine should be called into question. The anti-medical critique of psychiatry represents one approach within a wider movement which has assumed international proportions and adopts a critical or derogatory stance towards psychiatry's methods, aims and social role; it is anti-medical, anti-therapeutic, anti-institutional, and anti-scientific, either by expressed aim or implicitly through the dogmatic, hortatory, diffuse and inconsistent character of its utterances.

The aim of this social-psychiatric and personal follow-up examination has been to illustrate the long-term prognosis of Norwegian youth suffering from psychiatric disorders which initially presented during peacetime compulsory military service. The follow-up examinations were conducted personally, and a control group of presumed healthy men was also followed. Sund (1968, 1970) has made a systematic comparison between the patient group and the control group concerning prognosis. Here, therefore, the main emphasis will be on a comparison of the courses followed by clinically different diagnostic sub-group. Eitinger (1950) showed the necessity of undertaking personal follow-up examinations in order to map out the prognosis for this kind of patient. Prognostic studies outside Scandinavia have been reported by Ginzberg et al. (1959), by Glass et al. (1956) and by Plag and Arthur (1965). However, the periods of observation in these studies have been short; moreover, adjustment to the military system was the objective, and the cases were not personally examined in a follow-up. It has been difficult to find prognostic studies with sufficiently long observation periods and with personal follow-up to serve as an adequate basis of comparison with our material. To a large extent, therefore, we have chosen to see the development of our patients in relation to our knowledge of psychiatric reactions in the rest of the population.

In an earlier paper (Maxwell, 1972) the difficulties met in a dimensional approach to the description of symptomatology were discussed. In that study a representative sample of patients was considered under the three broad diagnostic categories, neurotics, affective psychotics and schizophrenics, and it was seen that all had a basic core of symptoms of the type generally referred to as ‘neurotic’, which lent themselves fairly well to a dimensional description and which indeed accounted for the greater part of the covariability of the symptoms in all three categories. But in addition the schizophrenics, and to some extent the affective psychotics, showed further symptoms of the type generally referred to as ‘psychotic’, though these occurred in a patchy and less general way than did the ‘neurotic’ symptoms. As a consequence these symptoms were not amenable to a dimensional description, nor did they lead to clear-cut typologies, as Everitt, Gourlay and Kendell (1971) have already shown.

In 1955 an epidemic occurred among the staff of a London Teaching Hospital group. The nature of the epidemic remains uncertain. One view is that the illness was a viral encephalomyelitis (Medical Staff Report, 1957); an alternative is that it was a manifestation of anxiety spreading through a population consisting largely of young women (McEvedy and Beard, 1970). This paper reports the results from the follow-up study carried out in 1968–69 on the nuclear group affected by the epidemic and on a matched series of controls.

Recent family studies of manic-depressive psychosis have emphasized the role of genetics in the aetiology of this bipolar illness (3, 8). However, the mode of genetic transmission is still unknown, the main controversy being between major gene and polygenic inheritance. Furthermore, it is not yet evident whether bipolar illness constitutes a homogeneous entity or whether it may be subdivided into different genetic subgroups. Mendlewicz, Fieve, Rainer, and Fleiss (1) recently produced some evidence that bipolar psychosis can be differentiated into two subgroups on the basis of family history data. Two matched samples of 30 patients each were studied, distinguished by the presence or absence of bipolar illness in their first degree relatives. The patients with a positive family history (FH+) in first degree relatives showed earlier onset of illness and more psychotic symptoms occurring in the manic phase. Alcoholism, if present, was of an episodic pattern. In patients with a negative family history (FH—), there was a later onset of illness; psychotic symptoms occurred usually in the depressive phase; and alcoholism, when present, tended to be chronic.

Weight change and sleep disturbance are well documented features of psychiatric illness. Complaint about difficulty in getting off to sleep is a common feature of neurotic illness, especially anxiety states. Early morning waking has for many years been regarded by some as a clinical feature—almost a diagnostic pointer—for ‘endogenous depression‘. Motor restlessness is also commonly described as a feature of this disorder. However, other clinicians have concluded that such depressive illness is merely one end of the spectrum of the depressive response. In particular, McGhie (1966) has shown that sleep disturbances of all kinds, as reported by the patients themselves, may occur in a variety of psychiatric illnesses. For instance he found that early morning waking was reported in 16 per cent of non-depressed psychiatric patients as distinct from 17 per cent of patients presenting with depression. Moreover, Hinton (1963) has shown that early morning waking when carefully measured does not specifically characterize those types of depressive illness described by others as ‘endogenous', but that it can occur in all types of depressive illness. He also reported his observation that sleep disturbance amongst depressed patients is mostly related to the phenomenon of agitation. Agitated patients lost more sleep, most apparent in the latter part of the night. On the other hand Carney et al. (1965), in their study of over 100 patients with depression, found that certain symptoms clustered into two groups reflecting the usual concepts of ‘endogenous' and ‘neurotic’ depression respectively. Early morning waking and weight loss were important characteristics of their group of patients designated ‘endogenous' depression, and were regarded as important indicators for them of the value of ECT. There was a positive inter-symptom correlation of 0.234 between weight loss (7 lb.+) and early waking. Most studies do not suggest a causal link between weight change and sleep disturbance, but by inference regard these features as being more integral aspects of an underlying biological factor which also promotes the disturbed mood.

If, at this distance in time, neither Mapother's nominalism nor Golla's vitalism-holism are at all satisfying, this is because so much has happened to the philosophy of science in the interval. The writings of Sir Karl Popper have had a revolutionary influence; and the important work he has done over thirty years or so is readily available in The Logic of Scientific Discovery (LSD, 1968) and Conjectures and Refutations (CR, 1969). Popper starts from the position that no logical basis for the accumulation of knowledge can be laid on inductive processes (and he also maintains that this is not the way the mind of the scientist works). From observing crows, and seeing that one after another they are all black, one cannot conclude that all crows are black. In general terms, it is not possible to proceed logically from single existential (‘there are …’) statements to universal statements; and an epistemological process that based itself on induction would be rocky at its very foundations. How is it then that science progresses, making mistakes from time to time, but recovering from them, providing for itself a firmer foundation for further work, anchoring it more deeply and extending it more widely? This could only be if the basic logical processes were not, in fact, inductive, as the nineteenth century scientist supposed. The search for an acceptable basic principle came down to finding a criterion of demarcation which would distinguish the empirical sciences from other branches of learning. Popper found it in the principle of refutability.

Ever since Kraepelin introduced his schizophrenic-manic-depressive dichotomy in 1896, this classification of the functional psychoses has enjoyed virtually unanimous acceptance. However, the frequency with which the differentiation of schizophrenia from manic-depressive psychosis seems to present problems in clinical practice inclines one to question whether Kraepelin's classification was entirely adequate. The object of this paper will be to try and highlight some of the reasons for the existing confusion and tentatively to suggest possible means of clarifying the situation.

From beginnings in imaginative literature and philosophy, the study of abnormal personality was taken up by clinicians who founded their investigations on the methods of observation and intuition. On the basis of case notes, experience and recollection, the pioneering psychiatrists constructed descriptive classifications which served as an initial taxonomy and enabled diagnostic and therapeutic work to proceed.

In 1965, when this study began, there were no clear guidelines about the optimum duration of in-patient treatment for alcoholics. Davies et al. (1956) had reported that at the Maudsley Hospital the aim was for a ‘short stay’ of between two and three months—but some individuals stayed for up to ten months. The policy at Warlingham Park Hospital as described by Glatt (1955) was for an in-patient stay of at least two months, though ‘on the whole about three months is the optimum period’.

There has been a reawakening of interest in the classification of psychiatric patients. Clinicians and researchers alike have realized that it is impossible to evaluate methods of treatment, determine aetiology, or measure the course of illness of psychiatric disorders without adequate methods for diagnosis. On the other hand, it has been shown that conventional clinical methods for making psychiatric diagnoses are of distressingly low reliability, except for the broadest categories, and have only marginal relationships to such criteria of validity as common aetiology, common response to treatment, and common prognosis (Baldessarini, 1970; Beck et al., 1962; Jenkins, 1966). Klein (1967) has shown that the usual clinical diagnoses, because of low validity, can actually obscure important relationships between types of psychopathology and such crucial variables as response to treatment.

Previous studies (McConaghy, 1969, 1970a; McConaghy et al., 1972) evaluated the efficacy of three widely different forms of aversion therapy in treating homosexuality. The methods investigated were aversion-relief, apomorphine therapy, and avoidance conditioning.

The first three cases of tardive dyskinesia were described in a brief communication by Schonecker (1957), and in the following ten years some 600 new cases were reported in 37 papers. Since the literature up to 1967 has been reviewed by Ayd (1967a) and myself (1968a), this communication will deal mainly with the publications of the last five years. I will also report a few personal observations of the last two years, hitherto unpublished.

Several researchers, among them Penrose (1944), Slater and Woodside (1951), Pond et al. (1963), Kreitman (1962, 1968) and others, have found that psychiatric disturbance occurs in both members of marital pairs more often than can be expected by chance. It appears that symptoms in the spouse are more evident when it is the husband who is the patient, and that neurosis rather than psychosis is associated with higher levels of disturbance in the spouse (Kreitman, 1962, 1964).

For the study reported in the accompanying paper (Willems et al., 1973), a number of variables, thought to be of prognostic relevance in alcoholism, were selected and categorized in a rank order. These can be regarded as ordinal scales, since in every case (a) each category differs from the others, and (b) each is related to the others. In each case there can be said to be an underlying continuum and in all cases a higher number indicates, in our opinion, a less ‘desirable’ state.

Hypochondriasis, though recognized as common and important, has not been accorded the study that its ubiquity and prominence might be thought to require. It is not long since anecdotes and uncontrolled observations masqueraded as knowledge.