Reconstructing Schizophrenia

Richard P. Bentall, Richard P. Bentall

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781134966844
• Edition: 1
• Topic: Psychologie
• Subtopic: Histoire et théorie en psychologie

`The summaries of evidence have provided ready-made challenges to previously unquestioned medical options... the book provides a challenging update on the nature of scientific inquiry.' - British Journal of Clinical Psychology
Despite nearly one hundred years of research, very little progress has been achieved in the understanding of schizophrenic behaviour. There remains considerable uncertainty even about the fundamental features of the hypothesised illness.
Reconstructing Schizophrenia subjects the difficult concept of schizophrenia to rigorous scientific, historical and sociological scrutiny. They ask why a biological defect has been assumed in the absence of hard evidence and look at what can be done psychologically to alleviate schizophrenic symptoms. Finally, they explore what new models and research strategies are required in order to understand schizophrenic behaviour. The result is a book that provides a distinctive and critical perspective on modern psychiatric theories and which demonstrates the severe limitations of an exclusively medical approach to understanding madness.

Part One
The Concept of Schizophrenia — Problems and Prospects

Chapter One
The Non-Discovery of Schizophrenia?

Kraepelin and Bleuler reconsidered

MARY BOYLE

It might appear that the history of the concept of schizophrenia has been well documented: a number of texts (for example, Leigh 1961; Hunter and MacAlpine 1963; Jones 1972) describe the historical background against which concepts like schizophrenia were to emerge, while texts or articles which discuss 'schizophrenia', whether academic (for example, Lewis 1966; Neale and Oltmanns 1980; Strauss and Carpenter 1981) or aimed at a wider audience (for example, Wing 1978; Clare 1980) give due consideration to Kraepelin's introduction and elaboration of the concept of dementia praecox in the fifth and subsequent editions of his textbook, to his disagreements with Bleuler, and to Bleuler's introduction of the concept of schizophrenia. It will be argued here, however, that these accounts, which are usually couched in terms of Kraepelin and Bleuler having described or differentiated a form of mental illness, are best thought of as received wisdom about 'schizophrenia', and that an analysis of what Kraepelin and Bleuler actually did (as distinct from what they are said to have done) not only casts considerable doubt on the accuracy of traditional accounts but also helps us to understand some of the problems facing the modern concept of schizophrenia.

In a separate chapter of this volume, Pilgrim describes the ways in which traditional accounts of the development of psychiatry as a scientific and humanitarian enterprise have been seriously questioned. One of the most detailed and able of these critiques has been provided by Scull (1979). A number of important points arise from his and others' work, but two are of particular interest here. The first is that, in Great Britain at least, humane asylum provision was originally conceived of as a lay and not a medical enterprise and that medical dominance in asylums was achieved only after protracted political struggle. The second point is that medical dominance was won in the absence of any scientific evidence as to the efficacy of medicine in theorizing about and intervening in deviant behaviour. Taken together, these points emphasize the back-to-front nature of traditional histories of psychiatry: it was not that psychiatry developed more humane ways of dealing with the 'mentally ill' or more scientific ways of understanding 'mental illness' but that medical dominance over deviant behaviour contributed to the later widespread adoption of the idea that it should be viewed as illness. The lack of evidence for the efficacy of medical theories and practice meant that Kraepelin and Bleuler arrived at the study of deviant behaviour remarkably ill-equipped for the task: as far as can be seen, they had little awareness of the problems of describing, observing, and recording behaviour. They also approached their task with strong a priori beliefs about the kinds of patterns they would find in the behaviour of asylum inmates and the variables which controlled such behaviour. Before their work is discussed, it will be helpful to describe in more detail the theoretical models which guided them (particularly Kraepelin) and the ways in which these relate to the activities of medical and scientific researchers in general.

‘Schizophrenia’ as Disease Entity — Syndrome — Hypothetical Construct

Like all scientific research, medical research is concerned with the valid description of patterns or regularities. When patterns are suggested, it is usual for researchers to infer unobservables from them (electricity, atoms, intelligence, learning, and so on). Of particular interest here is the type of unobservable usually known in psychology as a hypothetical construct, and two conditions must be met if it is to be claimed as valid. First, it must be derived from an observed pattern; this pattern or set of events then becomes the criterion for inferring the construct. Second, the construct must lead to new observations (the construct of Down's syndrome, for example, eventually led to the observation of chromosome abnormalities). These two conditions are sometimes referred to as the necessary and sufficient conditions for claiming a construct's validity, to reflect the fact that the necessary, but not sufficient, condition for asserting the validity of a hypothetical construct is that it be derived from a pattern; the sufficient condition for claiming validity is that the construct leads to new observations. If the necessary condition is not fulfilled, i.e. if the construct is not derived from a pattern, then obviously the sufficient condition cannot be either. It is unfortunate that the language of medicine tends to obscure the process of concept formation: when new patterns are suggested, we are apt to talk of 'a new disease being discovered'. Constructs inferred from the patterns (for example, multiple sclerosis, AIDS, diabetes) then become, misleadingly, the 'name of diseases' which people are said to 'have'.

In medical research, a favoured way of demonstrating that certain events 'go together' is to show that a proposed cluster is reliably associated with (an)other independently and reliably measurable event(s), known as signs. In turn, this new cluster may be shown to be associated with other events so that an originally sparse cluster may become elaborated as research findings accumulate. As Engle and Davis (1963) have pointed out, medical classification systems contain different types of patterns which vary both in terms of the type of events they include (structural damage, presence of micro-organisms, abnormal processes, and so on) and in terms of their 'richness' - some, for example, include events thought to be antecedents; others are little more than descriptive labels. Engle and Davis have divided these proposed patterns into higher and lower 'orders of certainty' (i.e. the certainty with which new exemplars may be recognized) to reflect the fact that some of the patterns are more clearly described than others and/or show less variability across exemplars. People bitten by rattlesnakes, for example, will show a less variable picture than those said to 'have multiple sclerosis'. At the bottom of Engle and Davis's list - at the lowest order of certainty - is that pattern known as a 'syndrome', or a proposed cluster of signs and symptoms whose antecedents are unknown. This type of cluster is the very least that is required in order to claim that a pattern has been observed and it is this status which is claimed for the cluster from which the modern hypothetical construct of schizophrenia is derived.

Kraepelin, however, did not set out to describe a syndrome. Instead, he believed that his task was to discover 'disease entities'. In this, he was reflecting medical thinking of the day: the idea that there existed 'out there' natural (?God given) and discrete groupings or clusters, each with its own antecedent both necessary and sufficient to produce the cluster. Although this idea had been strongly criticized, it received, as can be imagined, strong support from Pasteur's work on infectious microorganisms. It would be unfair, however, to judge Kraepelin's work by asking whether he described a disease entity: neither he nor anyone else has ever claimed that he did so. The important question, and that which will be addressed here, is whether he described any kind of pattern. In other words, were the necessary conditions for inferring the hypothetical construct of dementia praecox ever fulfilled?

The Problem of Pattern Recognition

The sophistication of modern medical theory and research techniques is such that some new patterns are identified and elaborated relatively quickly. Perhaps because of this, because success stories are well publicized and we will never know how many patterns are missed, it is easy for us to lose sight of just how difficult is the process of identifying patterns amongst seemingly similar and overdetermined phenomena. Should we, for example, group together all instances of chest pains, coughing blood, and emaciation?; or 'wasting' of the limbs and loss of power in them? If negative answers to these seem obvious, it is only so with hindsight. It can be argued that the whole scientific enterprise is an attempt to reduce errors in this process, to impose criteria whereby putative patterns can be shown not to be chance co-occurrences. In medicine, this concern is reflected in the demand that complaints, or symptoms, be reliably associated with an independently and reliably measurable event before a new pattern is (tentatively) said to have been described.

Kraepelin was well aware of the fact that he was required to demonstrate that certain phenomena 'went together' in order to justify his concept of dementia praecox, but he faced major problems in his choice of criteria for making this demonstration. The history of the medical study of deviant behaviour in asylums had been marked by repeated failures to demonstrate reliable associations between disturbing behaviour and particular biochemical processes, i.e. a relationship between so-called symptoms and signs. Nevertheless, rapid progress in medical research, particularly the descriptions of the 'infectious diseases' and the demonstration of the link between earlier syphilitic infection and later madness, encouraged asylum doctors to believe that results were just round the corner; in the meantime, other criteria had to be used to support the claim to have observed patterns. It was, of course, important to claim that patterns had been observed, that new forms of mental illness were being discovered, lest the public begin to question the wisdom of medical dominance in asylums, or lest physicians question the validity of their alienist colleagues' claim to be part of the medical profession.

Kraepelin's chosen criterion for claiming to have observed a pattern which justified inferring dementia praecox was similarities in onset course and outcome. Thus, he suggested that the phenomena from which he had previously inferred catatonia belonged to the same cluster as did those from which he inferred dementia praecox because 'both in their development and in their origins and prognosis we find an extensive correspondence between the two forms of illness' (Kraepelin 1896: 461). He also suggested that the phenomena from which he inferred dementia paranoides should be kept separate because 'Not only does [dementia paranoides] develop on average at a rather higher age, it has a course and an outcome which is considerably different.' (ibid.: 469) Kraepelin did not favour this criterion in general; it appears to have been chosen for want of anything more satisfactory. In choosing this criterion, Kraepelin was claiming that if he could identify a group of people whose behaviour changed in a similar way at one point (onset), showed further similarity in development over time (course), and reached a similar end-point (outcome), he would be entitled to conclude that he had observed a meaningful cluster of events and to infer a new hypothetical construct.

There are, however, a number of serious problems with this criterion for imputing meaningfulness. First, although the terms 'onset', 'course', and 'outcome' are often used as if they refer to simple, discrete events, in practice they are usually summary terms for a set of complex and continuous processes, particularly when they are applied to behaviour. Thus, what is actually recorded under these headings may be quite arbitrary and will differ across researchers. Second, and more important, these criteria offer no guidance in specifying important similarities: is sudden weight loss an important similarity?; or persistent joint swelling?; or eventual paralysis? If this method is applied to a heterogeneous population, showing a wide range of both similarities and differences (as did Kraepelin's population), the likely result is unstable and ever-changing groupings as different investigators, or the same investigator at different times, apply varying criteria for deciding whether a superficial similarity is important. This is exactly the situation found in Kraepelin's writings. A third problem of this criterion is that t...