Wednesday, 30 April 2008

Psychodynamic Explanation in Clinical Psychology

The British Clinical Psychologist's ambition to be scientifically respectable often manifests itself in an allegiance to generalisable causal explanation and the testing of generalisable causal claims by empirical correlational means. I want to cast some doubt on the way this ambition is often realised, but not through what might be thought the obvious way - through questioning the sufficiency of correlation as an index of causation.

Richard BentallConsider the way that Richard Bentall, perhaps to the mild dismay of some of his even-more-hard-nosed empiricist colleagues, approaches the idea that paranoid delusions can serve a defensive function and thereby maintain self-esteem. (See chapters 12 & 13 of his book Madness Explained.)

Bentall's idea is that paranoid people may both over-personalise happenings (viewing them as due to agency rather than chance) and employ the defense of externally attribute negative happenings, so as to preserve their self-esteem. I want to consider the second of these ideas. The first thing that Bentall suggests is that the hypothesis would seem to suggest that paranoid patients ought to have normal self-esteem if their defense is working. Bentall notes the contradictory data on this, and (in my view rightly) concludes that 'the concept of self-esteem ... fails to capture adequately the psychological processes involved in self-representation...'. 'Clearly', he says, 'we need to turn to methods of assessing the self that more accurately reflect its dynamic nature.' And that seems right on track. For one thing, to take it that any psychological phenomenon could be mapped simply on to any behavioural phenomenon, such as a self-report, would somewhat vitiate the whole point of the former being psychological. Whilst we need to acknowledge the constitutive relation between the psychological and the behavioural, we also need to acknowledge the essential defeasibility of the criteria employed in the attribution of psychological states and processes.

Bentall notes that paranoid people may judge their actual and ideal self as fairly on a par (contrast depressed people), yet may believe that they are not the ideal selves that their parents would like them to be. (Curiously this finding is left dangling on p. 333. The obvious psychodynamic theorisation would note the way in which our (accurate or inaccurate) feelings about how our parents would like us to be (which, when accurate, may still only be unconsciously held by the parents) can constitute intrapsychic as well as interpersonal motors of defensive processes. We typically internalise these parental expectations, and then may tend to split, project, deny, the parts of our own psyche which cause us to feel hopeless about ourselves. Yet the structure of the psyche is such that we either gain our (often all-too contingent) sense of our own value from the very same introjects - hence the density and ferocity of the intrapsychic conflict - or we split the introjects into good and bad, and suffer the correlative impairment in our reality testing.)

Next Bentall notes astutely that our judgements about ourselves are more implicit than any simple self-report may capture. A la early Jung, he uses an elegant version of the stroop test to test whether there was a discrepancy between, on the one hand, reaction times to coloured negative or positive trait words, and self-descriptions of such traits on the other. (Paranoid patients, on average, indeed described themselves using both positive and negative trait words, but tended to have longer reaction times to the negative.) Various other consistent findings are presented.

Most interestingly of all (pp. 337ff.), Bentall writes about the idea that paranoid patients have consistently low self-esteem (which they attempt to mitigate with 'the' paranoid defence) as follows:

this [is] an oversimplification. The picture that emerges suggests a much more complex and dynamic relationship between attributions, different kinds of self-representations, mood and paranoid delusions, as if the paranoid patient is constantly fighting to maintain a positive view of the self, sometimes winning... but more often losing.'

Bentall continutes, making many further interesting observations and predictions, noting in particular that attributing negative thoughts, intentions and feelings to others (rather than oneself) can all too easily result in feeling that one is hated by them - which can in turn lead to low self-esteem unless one then in turn defends against this.

To my mind this point of Bentall's is extremely interesting, deserving a lot more space than he affords it. Let me come back to it in a moment, first pausing to notice what Bentall says (p. 339)about another important topic: testability:

Of course, sceptics might argue that a theory of this sort should not be trusted because it cannot be properly tested. (This is precisely the reason that many modern psychologists have rejected the theories of the early psychoanalysts [and, he could have argued, the reason why, in their assimilations of later work with simplistic readings of Freud, they reject the whole lot wholesale...]). Against this argument it can be said that some phenomena just are unpredictable (the weather is a familiar example), and establishing when this is the case is an important kind of scientific achievement. (This is one of hte main goals of the mathematical science of non-linear systems theory, more popularly known as chaos theory.)

Now I suspect I am not alone in feeling that, whilst Bentall is right to defend psychodynamic understanding against such objections, the comparison with chaos theory potentially undersells the value of the psychodynamic formulation. This is because the psychodynamic formulation provides us with a form of understanding of someone's emotional and interpersonal behaviour which is quite different from, and exceeds, that relevant to comprehending a merely causal system. What we apprehend, when we explicate a behaviour in terms of a dynamic psychological mechanism, is a relationship which we understand. Our grasp is a grasp of meaning - a meaningfulness that, I believe, is not present in the kinds of causally complex systems described by chaos theory. The meaning is a function of the behaviour's intentionality: it's directedness towards the environment - and its teleology - it's purposive character. What we apprehend is what can be called an 'inner connection' between the defence and the behaviour: we see how the latter partakes of the form or structure of the former. We don't just posit that it is an effect of it; we see not just that but how it is an expression of it, how it is, as it were, part of its own body. The increased richness of the language we use to describe the behaviour is not due simply the accumulation of knowledge of its independent causes, but rather to our increased understanding of what it itself is. We understand the behaviour better, and in and through this, better understand the person him or herself.

Bentall goes on to say that even with unpredictable systems, 'it is possible to make other kinds of predictions that can be subjected to experimental investigation', and goes on to give a nice example. But what I am considering is the possibility that - even without the possibility of making, in the empiricist mode, such predictions - we can with entitlement rest secure in our case-based or 'idiographic' understanding of individual cases in terms of their psychodynamics, even when these don't conform to the kinds of testability and 'nomothetic' generalisability expected by the empiricist epistemology of experimental psychology. It is, I want to suggest, just this understanding, this apprehension of the inner connection between motive and behaviour, that sanctions our self-ascription of psychological understanding when we are dealing with the psychodynamics of paranoid (and other) thinking and action - and not the possibility of making the 'other kinds' of predictions that can be subject to experimental investigation discussed by Bentall.

When confronted by psychodynamic hypotheses which are recalcitrant to generalisation and experimentation, the question which arises is either: "Too bad for the hypothesis" or "Too bad for the experimental method". My inclination is to vote for the latter. But the way in which an empiricist epistemology has become so heavily entrenched within clinical psychological thinking, as a kind of tacit standard for what is to count as 'scientific knowledge', makes it hard, I believe, for the pursuit of non-generalisable 'idiographic' understanding to even seem intelligible, let alone respectable.

Bentall himself works largely within an empiricist tradition, and it is this which makes his dalliance with psychodynamics unusual when compared with the approaches taken by academic colleagues working, say, at that bastion of British empiricism - the London Institute of Psychiatry. His reference to 'chaos theory' and the limits of generalisability and testability shows an awareness of the difficulties attendant on subsuming the psychodynamic hypothesis within an empiricist epistemology, although the rest of his work, and his quick textual return to 'other kinds of predictions that can be subjected to experimental investigation' appears to betray a lack of comfort when working outside of the domain of experimentalism.

In contrast with the methods by which hypotheses are validated in experimental psychology, psychodynamic hypotheses are typically reinforced by the gathering of more contextual information about the person's responses, the seeing of more internal relations of meaning, noting the patient's reactions when asked, the quality of their avowal or disavowal of the suggestion (regarding low self-esteem or what-have-you). Furthermore, psychodynamic hypotheses, when considered as generalisations rather than claims about specific incidents in the intrapsychic lives of specific patients, gain their plausibility not from their statistical prevalence, but from their identification of a structure of meaning. Once appraised of this structure we become appraised of a new 'sense' - a new understanding which we can bring to bear on any clinical material that sustains it. Having a new way of making sense of - finding intelligible - seeing meaning in - certain data is far removed from having the knowledge that there is a 68% probability that it has, in a certain population, a certain efficient cause.

Further still, there is no conclusive way of defeating a psychodynamic explanation, since any challenge to it will itself always be defeasible. But this is not to say that we are in epistemological free-fall, since the epistemic structure of the idiographic sciences has its own standards. Thus we note the internal relations between psychodynamic mechanisms and behaviours, and also their defeating conditions. It is necessary, to cast doubt on a psychodynamic hypothesis, to not merely speculate about the logical possibility of such defeating conditions being met, but to actually try to show how they are met. And of course any such attempt will itself stand only so long as no-one demonstrates its own defeating conditions are met, and so on. However much this lack of certainty may frustrate us, it may be nothing more than hubris to escape this hermeneutic circle by switching epistemology and jumping into the hypothesis-testing nomothetic methods of experimental psychology.

I said above that I'd come back to the question of the way in which defences are often self-undermining. I defend against my anxiety or shame or terror or badness, 'project it outwards', or split off anxious parts of myself in the process and locate these instead in others. The details are all too obvious from practice, although the theorisation of projective identification needs to be carefully handled, as I have argued in the previous post on An Existential Account of Projective Identification. But I have often noticed psychologists tending to assume that if one accepts the notion of psychological defences one accepts the idea that they are successful. Successful, that is, not just in bringing a very temporary relief, but in the longer term. It is however important to note however that these processes only work through operant conditioning - they are not intentional thought-through strategies on the part of the 'defender' - and so we ought not to be surprised if they are ultimately as hopeless, and self-defeating, as the kinds of strategies (avoidance, safety behaviours) which the non-dynamic psychologist might well happily sanction as clinically and theoretically valid posits. The sheer ultimate futility of many defences (to say nothing of their constant psyche-depleting nature) tells very little against the very idea of them.