Download this article

This is the second piece in an exchange of views between Robert Harris and George W. Brown, originally published in Medical Sociology News Volumes 5 and 6, between January 1978 and January 1979. All four parts are published here, together, for the first time.

Science and Common Sense: A Reply

George Brown
Social Research Unit
Bedford College, UK

Originally published in MSN Volume 5, Issue 3, September/October 1978

Harris’s article in the January number of Medical Sociology News criticising the London research on schizophrenia depends throughout on a dichotomy between 'common sense‘ and 'science' which are at no point defined or explained. Moreover its attack is sometimes based on the argument that the work is common sense masquerading as science but elsewhere it is seen as rejecting common sense. This makes a reply difficult and only in his last sentence do we get a hint that his doubts stem from a suspicion of quantitative expression - 'a mystifying welter of reified statistics'.

Since at heart doing science consist of a persistent effort to seek as honestly as possible answers to particular questions about the world, it is the procedures used, not the answers obtained‚ that distinguish it from common sense. Moreover science does not expect to obtain a completely valid description of the world or a final understanding of it. A scientific study merely sets out to measure the world well enough to arrive at a plausible answer to questions. Measurement has only got to he good enough to achieve this. It is therefore silly for Harris to say that he does not accept our measures 'constitute the totality of the affective components of the households' we studied. Who ever claimed that they were? We claim no more than that they are relevant and valid enough to test whether the emotional atmosphere in the house influences the course taken by a schizophrenic disorder. The statement 'insofar as Brown believes his methodology to be valid (which he does) he is denying that he may be wrong' illustrates the depth of the confusion here. The term belief is used pejoratively: the appropriate idea would he a struggle to believe. Methodology consists of procedures and arguments that can at best influence‚ positively or negatively, a confidence in one's results. This is always a matter of degree - confidence can never be absolute. To repeat‚ methodological considerations are not concerned to achieve totally accurate or comprehensive measurement (an impossible task) but to arrange measurement in such a way that‚ given a particular set of results‚ the investigator can rule out as implausible interpretations of the results other than the one he believes to be correct. There is no question of claiming at any point infallible knowledge. Indeed‚ scientific knowledge is often best viewed in economic terms: given that effective research is expensive In terms of money and time‚ is the pursuit of a question justified? Should we go on and if so, in what way? The study under discussion is the third of a series and now over twenty years since the start of the programme‚ work on the role of expressed emotion (EE) is currently being carried out in the UK‚ USA, Denmark and India. When this work is complete we will be more sure or less sure of what we know: it is unlikely that our knowledge will remain unchanged.

Harris early in his paper seems to be convinced that our conclusions are no more than common sense and‚ I infer‚ that the work need not have been done. This is a historical and an intellectual misjudgement. When I first thought in terms of the influence of home atmosphere twenty years ago there was no question of the idea’s general acceptance. There was a widespread belief in an inevitable endogenous process in schizophrenia and the extraordinarily disturbed and unusual condition of many chronic schizophrenic patients did not make this view absurd. Second‚ he ignores the plausible alternative interpretation of our results and our long struggle to rule it out. This is that there in no causal link between degree of expressed emotion (EE) and relapse: that the patients‘ disorder determines both EE and relapse and the apparent causal link is spurious:

Stacks Image 14059

To ignore such major methodological questions is to trivialise the research and to underplay grossly the need to test ideas.

Harris is also muddled about the idea of causality. He questions the relevance of our interpretation because it denies 'the volitional aspects of human action and for that reason alone must be rejected as satisfactory explanations of that behaviour.' This again posts a dichotomy the elements of which are unexplained. We, in fact‚ make clear that those involved are not without some personal influence. The patient can, of course‚ take drugs. We also demonstrate that the schizophrenic patient when returning to the 'wrong' sort of home can‚ by reducing the amount of face-to-face contact in the home‚ much reduce its adverse effects. But this is not the same as saying that a patient necessarily realises the full implication of what he or she in doing. One young woman told us how she no longer lingered after a dance to make love behind the local hall‚ as she had learned that this brought back the voices in her head. This degree of understanding in probably uncommon but it is possible‚ and one consequence of developing knowledge of aetiological processes is the chance of incensing such self-awareness. But the general point is that in sociological research we can never assume we know the degree to which 'volitional' and ’deterministic' components are involved in a situation: it in an empirical issue to be settled for every situation anew. What we can be sure about is the uselessness of a general assertion of the kind made by Harris.

The matter in inherently complex. Having plans in not the same as being able to put them into effect; putting plans in to effect in not the same as achieving them and achievement will not necessarily lead to the emotions that we hoped would flow from the successful realisation of our plans. At best we are only partially and episodically in control of our experience. It is a travesty of this view to claim it is mechanical. I should add in this context that I am at a loss to understand his discussion on page eight where he discusses patients returning to a high EE home‚ who have been seriously disturbed before admission‚ who take drugs after discharge and who do
not relapse. In interpreting this in so-called volitional terms he seems to suggest that the patient does not relapse because 'if he relapses the patient knows he will be readmitted.' This might be intended to mean that this is why they took drugs and this is the reason for them not relapsing. If so this is entirely consistent with our position; or it might be meant to suggest that patients can decide whether or not they will relapse. Our view is that schizophrenic patients do not have this control in the sense they can decide whether or not to catch a bus. But they certainly do have potential for doing things that will lessen the chance of their getting into a situation where their experience of schizophrenic symptoms gets out of control. And l have no doubt that patients to varying degrees develop and utilise such ’knowledge'. lf we assume that a phenomenon in entirely 'determined' or entirely 'voluntaristic' we will not only almost certainly be wrong, we are almost bound to rule out the development of effective social theory. Such theory in essence is about constraints and the degree to which they may be 'overcome'.

A more general point can be made. The argument illustrates a common fallacy - that of confusing procedures employed for methodological purposes with ideas held by investigators about the nature of the phenomenon they study. It is an though a scientist were accused of denying the existence of colour because he based his research on black and white photographs. Of course‚ it is possible that some are led astray - that they do doubt the existence of colour. But this cannot be used to impugn the method; the correct deduction in that we must work to obtain and retain a self-awareness of the status of our methods. Simple-minded dichotomies of the kind made by Harris obscure this vital issue.

Confusion seems in part to arise from his ideas about what we have measured. l quote: 'For him, therefore, expressed emotion (dislike) exists in and through the methodic practices by which it is recognised and described.' Once again Harris’s argument rests on the assumption that science claims to encapsulate totally its subject - a straw man. And furthermore that if one's methods play a role in shaping one's results‚ this constitutes a reason for the blanket rejection of results. That by nuking operational assumptions one will inevitably be wrong. For example‚ that using 'critical comments' as though they were equivalent to each other must lead us astray. (In fact in our earlier papers on measurement we describe an
overall measure of criticism not making this assumption - this gave much the same result in predicting relapse as the count of individual critical comments). This again implies a view that science is either right or wrong - a straw man epistemology. It also follows from such a view that only exhaustive description could avoid being wrong. Yet to invoke Garfinkel (as he does)‚ every description has to be finite and limited in its selection of reality. The world exists only insofar as we are able to develop categories that describe it. We, of course‚ still need to deal with the accuracy of our measures and whether the process of abstracting has produced casual links where there are none. And given our measures survive such tests we need to go on to question the status of any theoretical interpretations that we have made.

I will deal with the issue of measurement inaccuracy and error first. Harris notes that ‘intuitive judgements and feelings are not allowed' in our measurement of EE. We make it clear that they are. If we use observers to measure emotion there is no other way to proceed. While we systematise such judgements‚ we have never doubted that for this we use tacit knowledge and skills‚ developed and used in everyday life. There is evidence that we have managed to do this reasonably well (see Brown and Rutter‚ 1966; Rutter and Brown, 1966). Our belief that our results are not artefacts is strengthened by the measurement of expressed emotion before any relapse. I, in fact ‚ know of nothing about our measurement procedures that suggests that the links we have obtained are not casual: that returning to a certain kind of home atmosphere often leads to a relapse that would not have otherwise occurred. But ideally experimental confirmation in still required.

The second issue of the theoretical status of what we were measuring and the interpretation of the reasons for the link between EE and relapse is more open. Harris seems unaware, however, that many of the theoretical concepts and measures in the social sciences involve
dispositional concepts: constructs that indicate that a person in likely to act in a certain way given a certain set of conditions. We believe our measure of EE most likely reflects a disposition of the relative to act in a certain way towards the patient under certain conditions. While this interpretation at present is speculative‚ Harris again manages to trivialise the matter. He, for example‚ equates high EE with dislike of the patient. But we already know enough to be confident that such a general interpretation will not do. For instance‚ extensive dissatisfaction with the patient was common; and yet unless dissatisfaction was associated with seven or more 'critical comments' it was unassociated with an increased risk of relapse. This result held however marked the dissatisfaction. Moreover‚ high emotional involvement was associated with relapse irrespective of criticism or hostility.

Harris also at this point ignores our broader theory about schizophrenia - that schizophrenic patients are particularly sensitive to too little or too much stimulation. That with too little stimulation they can develop the signs of extreme withdrawal‚ poverty of speech and even muteness‚ to be seen in old style chronic mental hospital wards; and that with too much stimulation they can develop florid symptoms‚ the latter developing in response to the experience of any marked emotion including joy and excitement. Although he may not agree‚ it is misleading to reduce our work to ‘coping with dislike' without dealing with these theoretical ideas about the nature of schizophrenia.

There are many other criticisms l could make. But I will restrict myself to one more: about the nature of schizophrenia. Implicit in Harris’s interpretations is the wish to deny the existence of schizophrenia and its manifestation in characteristic florid symptoms and disordered behaviour. His opening insistence on the common sense nature of our results appears to derive from this. For him relatives who dislike the patient are more likely to define hint behaviour in a way that renders it perceivable as symptomatic‚ the patient may mimic psychiatric symptoms‚ the patient’s presence in hospital can lead to the definition of symptoms as a means of legitimising his presence. In other words any interpretation other than one accepting that schizophrenic phenomena exist‚ cause great distress and present a major medical and social problem.

Recent community studies have been‚ as far as I am aware, completely in accord about the extent and seriousness of the symptoms and handicaps of discharged schizophrenic patients. lf there is not a core of characteristic symptomatology at the heart of the disturbed behaviour of schizophrenic patients‚ I am at a loss to conceive of an explanation of what has been documented in these studies. That someone may be able to mimic a schizophrenic disorder in order to fool s hospital psychiatrist is irrelevant for this issue (and whether this has been done is in any case in some doubt - see Spitzer, 1976). My experience of schizophrenic patients has been that the majority have arrived in treatment only after persistent efforts by relatives and friends to deal with them as though there were little or nothing wrong. I would not wish to undervalue the effectiveness of such an attitude. It may help patient and family to cope. But to persist in such a view – that nothing essentially is wrong that everyday responses cannot put right - can lead to distressing‚ if not tragic consequences. Harris may well consider the implications of David Reed's account of his wife's schizophrenic illness in his book

I will make a final and more general point. Underlying invective of the kind that Harris pours on science appears to be a rejection of attempts to sum up the complexly of the human condition in abstract‚ and perhaps numerically expressed‚ principles. This is a false fear. As Toulmin has made clear‚ even for the physical sciences‚ the actual complexity of the real world can defeat any straightforward practical use of its principles. Although Newtonian and later physics gives us a satisfactory explanation for the phenomenon of tides‚ the only way to predict the tides at Southend is to go there and measure them. Likewise‚ at best we may obtain principles of relevance to the course of a schizophrenic disorder. These will be fallible not only because to some degree they are bound to be inadequate but because we cannot possibly predict (or control) the contingent factors likely to impinge on the patients' life. A patient may return to a ‘perfect’ home according to our principles, but an unexpected occurrence (say the return of the landlady’s son from sea) may transform the situation. This is why we will always need 'clinicians' to ’translate' any scientific principles we acquire; and this is why there will always be a place in the social sciences for intensive descriptions of the individual and his life. But to confuse either with the building of broader principles - and fallible knowledge - is pitifully misguided.

Stacks Image 32

Brown, G.W. and Rutter, M. (1966) The measurement of family activities and relationships: A methodological study. Human Relations, 19, 241.
Reed, David (1976) Anna. Penguin.
Rutter‚ M. and Brown, G.W. (1966) The reliability and validity of measures of family life and relationships in families containing a psychiatric patient. Social Psychiatry, 1, 38.
Spitzer, R.L. (1976) More on pseudoscience in science and the case for psychiatric diagnosis. Arch. Gen. Psychiatry‚ 33, 459-70.

Next: 'Science and Common Sense: A Rejoinder to Professor Brown' by Robert Harris