Firstly, I would like to thank our two speakers for presentations which have been thought provoking, challenging, helpful, certainly educational - in my case - and perhaps also a little worrying. I suspect that, if members of the audience have had the same diversity of reactions to what has been said, then there will be no shortage of discussion. For this reason, I will simply confine myself to one or two simple points to start that discussion. What Don kindly didn’t say in his introduction was that, before I saw the light - as a result of several years on the couch of course - I was a cognitive-behavioural therapist. I consider myself well placed therefore to be a different kind of nettle, and act as an initial voice in the discussion for non-members of the psychoanalytic community, of whom I am delighted to see there are many here this evening. My points then relate to the external functions of research and particularly to outcome research.

Firstly, I think it is very helpful that both of the speakers, in their different ways, have reminded us that the nettle which has to be grasped in relation to psychoanalysis and research is not that of deciding between approaches which we might simplistically describe as ‘objective’ and ‘subjective’. The notion that observations can be truly objective in some absolute sense, that they can occur without being influenced by the preconceptions, expectations, and frameworks of thought in which they occur, is clearly illusory. 

I found that Ron Britton’s use of chaos theory and fractals, and the analogy with predicting the weather, was also a very helpful way of thinking about the essential unpredictability of human behaviour and human thought. We cannot predict behaviour from minute to minute, let alone predict thoughts and feelings from hour to hour or week to week. 

There is a question, however, about just how far the weather analogy can be stretched. I would argue, for instance, that there are many aspects of the weather which are eminently predictable. I think I could predict with near certainty, for example, that the average temperature in Britain in August next year will be greater than the average temperature in Britain in February next year. I could also predict with near certainty that the average rainfall in the highlands of Scotland in the year 2005, say, will be greater than the average rainfall in the Sahara desert during the same year. I can predict these things because they have always been so. Summer was always hotter than winter long before we understood the fact that the earth went around the sun and that this affected the weather; and I have no doubt that ‘flat earthers’ had theories which explained the variations in the seasons. The reason we can predict these things confidently is because we know that, in general, the past tends to predict the future rather well. 

As I see it this may also be true in certain respects of human behaviour. If I take twenty people who are depressed today and another twenty who are not depressed, then I can predict with a fair degree of confidence that tomorrow, on average, those twenty depressed people will be more depressed than the ones who are not depressed today; and that this will also be true in two weeks time and probably also true in six months time. This, of course, is not the fine-grained prediction of individual elements of behaviour or thinking of the kind with which psychoanalysts are commonly concerned. This is prediction about general trends and global tendencies… but there is an argument that outcome research is more to do with this; in other words, more to do with general trends and questions, like summer being hotter than winter rather than whether or not it will rain in three days time. I suspect there are many in the audience [tonight] who would think it appropriate that psychoanalysts ask global questions about whether patients in general undergoing psychoanalysis and psychoanalytic psychotherapies are better off as a result of that experience, than those who have not undergone such treatment. Insofar as psychoanalysis is claiming to be therapeutic, in addition to consciousness raising, then there is an argument that it must grasp this nettle.

A second point that I would like to raise, this time with the external perspective of the psychotherapy researcher, and which is perhaps more of a question for Peter Fonagy’s standpoint, concerns the extent to which psychoanalysis and psychoanalytic psychotherapy can be considered as identifiable entities in their own right, independently of the individuals who practise them. Psychotherapy researchers have often challenged what is known as the ‘drug metaphor’ of psychotherapy research: the idea that a psychological therapy can be considered analogous to a drug, in that it can be applied in some direct way to a patient with a certain dosage and a certain expectation of outcome. The idea would be that one could then identify important processes (like the ingredients of a drug treatment) and expect some sort of correlation between process and outcome. There are fundamental problems with this view of psychological treatment. Some of these are conceptual; others are more down-to-earth and empirical. In the down-to-earth category is the fact that we know - from comparative therapy research - that the type of therapy patients are given seems to have very little predictive power concerning whether that treatment will be effective or not. In the language of outcome research, we might say that only a small proportion of the variance in outcome can be accounted for by treatment type. It would seem that therapist variables and the therapeutic alliance probably have a great deal more to do with outcome. Therapy therefore is something that is constructed - as it proceeds - by patient and therapist together and there must therefore be a limited extent to which we can see it as a predefined entity. This would also seem pre-eminently true of psychoanalysis. Analysts may have a shared view of their procedures, may agree on certain common components of psychoanalysis - for example the use of transference interpretations, the idea of elucidating unconscious processes - but it remains to be seen whether variations between individual analysts in these areas are sufficiently small for us to feel confident that psychoanalysis, as a set of procedures, is a sufficiently unitary entity - enough of a thing - for it to be meaningful to try to evaluate its effectiveness in global terms.