Thursday, 20 February 2014

Psychoanalysis and the Scientific Imagination

This post is about the part of psychology that has maybe caused more disagreement than any other. Doomed and tragic, dogmatic and sometimes foolish, psychoanalysis has been derided and "killed off" multiple times over. Like a religion it inspires dogged loyalty among followers which embarrasses people of a more skeptical mindset. Nonetheless, I retain an ongoing fascination with psychoanalysis as a way of thinking.

As an undergraduate I was discouraged from even thinking about picking up a book by Freud. His work was taken as the paradigmatic example of how NOT to do psychological science. The prototypical psychology undergraduate who makes assured statements of the form "Freud has all been disproven" has not really got it right (could they even say what it would mean to disprove such a huge body of work?), but they are not wrong either. How can I be both empirically minded and interested in the metaphysical suppositions of a patrician doctor from fin de si├Ęcle Vienna?

The privilege of "Sciencey-ness":

Here's the source of one potential form of mistaken thinking about the science of therapy. Contrary to intuitions, psychoanalytic theory is no less capable than cognitive theory of leading to treatments that can be manualised and tested. We may assume Cognitive Behavioural Therapies are more appropriate for modern clinical settings because the language of "cognition"; "schema" and so forth lends an additional superficial "sciencey-ness". However, the real litmus test of a therapy is neither its sciencey-ness nor its poetic intuitive appeal, but its ability in clinical trials to effect the sort of changes it claims to be able to make.

Unfortunately, it is a shabby truth that psychoanalytically inclined clinicians have resisted seeing RCTs as a legitimate way to validate their insights. Where they have (Fonagy and Bateman's Mentalisation Based Therapy and Clarkin and colleagues Transference Focused Psychotherapy-both interventions in the field of "Borderline Personality Disorder") they have produced some promising results. Some service users have reported that DBT, with its talk of "emotion regulation" can feel infantalising and restrictive. If alternative approaches can avoid these complaints then that would seem an important advance in the direction of greater choice for a group of people who are often otherwise offered very little.

Psychoanalysis and Science:

"Basic" psychological science isn't just a set of methods for verifying theoretical statements. Although such tools for verification are the main subject matter for philosophers of science, we also need a way to generate theories and new ideas to test. This requires creativity, imaginative flair and some familiarity with what you are trying to study. Einstein's creative thought experiments are lauded by the popular imagination because they challenged the conventional wisdom about how energy worked. If we view psychoanalysis as a rich field of imagination-stimulating ideas about the mind, its utility to science comes more to the fore. Nobel prize winning brain physiologist Eric Kandel has called psychoanalysis "the most coherent and intellectually satisfying view of the mind", albeit in the context of a plea to its practitioners that they up their game in terms of scientific theory development. Another Nobel winner Daniel Kahneman tells a fascinating story here about the value of reading Freud closely to generate new ideas for experimental testing:

In thinking scientifically about the way the world works we perhaps need to start with a proliferation of theories and ideas which we only later cut down to size through empirical investigation. If nothing else, psychoanalysis has been provocative and stimulating in developing detailed ideas about what sort of things minds are.

Psychoanalysis and "the ecstasy of truth":

Somewhere beyond the deliberately limited and precise technical vocabulary of scientific psychology there lies an explanatory and expansive language of mind to which we aspire. Novelists and poets have the ultimate privilege in this domain, they deal in fiction and verse, which provides the ultimate disclaimer for saying whatever you want and sometimes landing upon something that feels deeply and wonderfully true. 

Werner Herzog explains the distinction between two different views of truth

Psychologists have a duty to write and think in a different way if they have any hope of saying things which are verifiably and usefully accurate. We are-nobly I think-purveyors of what Werner Herzog would call the "accountants' truth". The price we pay for this is that we limit our capacity to get at the fullness of "the ecstacy of truth"; the truth of what it is and means to be human. Nonetheless we all surely want, at least in some part of ourselves, to go beyond cautious psychometrics and deploy words that really get at what it is like to feel and think. Why otherwise would we have become psychologists?

Psychoanalysis is not just a theory, it is a different way of having theories; a different way of writing about people. It is also a phenomenology. When Freud differentiates mourning and melancholia by saying that in the former it is the world which has become "poor and empty" while in the latter it  is the Ego itself; when Winnicott speaks of "annihilation anxiety" or of "holding" as a metaphor for the way that a clinician can help bear disorganising feelings of dread, these sorts of descriptions would seem to bring us close to lived experience in a way that technical scientific writing cannot.

Psychoanalysis is surely a dying art, surrounded by obfuscation, obscured by shroud waving, but for me it retains a valuable intellectual core. Freud was a staggeringly good writer (arguably too good; his capacity to convince and convert people can be viewed as intellectually problematic) and he developed a beautiful and detailed view of the psyche. Unfortunately for us, the truths that are contained in the enormous corpus of psychoanalytic writing are fleeting and hard to pin down. Some might suggest that analysis and analytic training can help you grasp them more firmly. I dislike the hierarchy and authoritarian nature of psychoanalytic "received" truths, but many of its texts are poetic, imagination expanding and enriching nonetheless.


  1. Before lamenting psychoanalysis as a 'dying art', it might be worth making a few observations, not least of which is the fact that it isn't dying at all. Despite prophecies of its doom for the past century, psychoanalysis remains a considerable (if marginalised) force in the Anglophone world, and is doing pretty well in some places elsewhere (eg. Argentia, Belgium, among others). We should ask ourselves how it is that psychoanalysis has survived, if it has supposedly been 'debunked' so many times. My hypothesis is that Freud's texts will be studied long after those of Beck, Maslow, Skinner et. al. have become mere footnotes to footnotes.
    Second, to the extent that psychoanalsysis is rejected by mainstream psychology, this has little to do with science. Mainstream psychology is not science, whatever the pretensions of its practitioners. The basic ideas of psychoanalysis - a radical notion of the unconscious and subjective division, respect for the unique and non-generalisable, subjective implication in sexuality - remain as scandalous today as a century ago. When one sees the shrill protests of the anti-psychoanalysis crowd, one need only examine their tone to note that it isn't 'science' which is at stake for them.
    For this reason, Fonagy et. al. have missed the point. The dogmatic peddlars of CBT who imagine themselves to be doing 'science' will never accept psychoanalysis, irrespective of 'promising' RCTs. (The latter are a poor measure for any psychological treatment in any case). Psychoanalysts are quite correct to reject the premises of the RCT/CBT game as false. In psychoanalysis, treatment is strictly indirect (i.e. a consequence of exploration of the unconscious). Other treatments claim something similar (ACT for instance), but never take this notion seriously. If the aim of psychoanalysis is an articulation of thee unconscious then 'standardised' trials and quantitative measures are indeed a waste of time. IMO, you need to further analyse the assumptions of your 'litmus test'.
    I agree, however, that psychoanalysis addresses something in human subjectivity that is almost entirely absent elsewhere in psychology, and that this is a reason for its enduring appeal.

    1. melbournelacanian,
      I always love your trenchant posts and I hope you don't see me as part of the "shrill anti-psychoanalysis crowd"!
      Perhaps the strongest claim you make in your comment is that psychology is not a science. In reply I would say fine; I don't agree, but there are grounds for the point and I am not particularly invested in the project of claiming a privileged scientific status for the field simply for the sake of it.
      What I AM invested in is the project of a healthcare system which, as far as possible, meets some socially agreed standard of quality. This is why I do not think RCTs, Fonagy or CBT per se have "missed the point" as you claim. I think you're right to suggest that various interventions get an ideological boost from playing a spurious science-game (this is what I call "sciencey-ness") but I don't really see how we can jump from this position to the radical rejection of ANY quantitative approach to evaluating psychotherapies. The RCT litmus test is not the final answer, but it goes a long way toward the kind of assessment we need.
      What we need in addition to RCTs is a sort of qualitative critical discussion about what sort of interventions we want, what sort of care they should offer and so forth. Psychoanalysis is one way of having that conversation, but it gets into alot of trouble around legitimation of its theories. Apart from flashes of a familiar truth in various psychoanalytic writings, how are we to distinguish the claims of various theorists beyond appealing to argument from authority?

  2. My hypothesis is that Freud's texts will be studied long after those of Beck, Maslow, Skinner et. al. have become mere footnotes to footnotes.

    True, but that just proves that Freud is nonscientific. The mark of science is that it progresses and old works become obsolete. Freud is not even wrong. The Bible, the Quran, and the Tao Te Ching will be read long after Freud has been forgotten, but that doesn't make them science, or true.

  3. I don’t think that you are anti-psychoanalysis, and I can understand why somebody would want some measure of which interventions are useful in order to better allocate resources within a health system. My argument, however is that such health systems are principally political, and that science and evidence only enters into consideration in a secondary or tertiary way. The recent furore with CBT for psychosis in the UK would seem to be a case in point – lack of evidence has not stopped NICE from recommending the psych treatment with the most political clout.
    I think that some parts of psychology are more scientific than others. Some aspects of perceptual psychology, for instance, approach scientific rigour. In contrast, the branches of the discipline dealing with treatment are almost entirely unscientific, despite this being the ‘applied’ side of psychology.
    So for instance, directive, quantifiable approaches to treatment presume to know 1) what’s wrong with the suffering person; 2) the end goal where what is wrong is fixed and; 3) precisely how to get there (steps, treatment plan, ‘strategies’, etc). These are wild assumptions which have nothing to do with science. In RCTS, ‘disorders’ have to be constructed, then operationalized in a limited way, then given the myth of a ‘standardised’ treatment, and reduced to statistical outcomes. This may be an appropriate way to assess the effects of antibiotics on an infection, but from start to finish, it tells us very little of the sorts of operations involved in a psychotherapeutic treatment. It is intrinsically nonsensical. It may be unpopular, but I would call for the reinstatement of the case study as a means of learning more about treatment, with the caveat that whatever method we choose – case study or RCT – we must resign ourselves to a generalizability that is strictly limited. If we want to prioritise generalizability, we have to either throw rigour out the window (the reigning current practice) or limit our statements to very high levels of abstraction. I think that psychosis is a very interesting case in point. In psychoanalysis, there are singular examples of psychotics becoming stable with a combination of, say, psychoanalytic treatment, artistic pursuits and personal relationships, but there is no sensible way that this path to stabilisation could ever be prescribed as a generic cure.

    How we might distinguish between different psychoanalytic theories is a good question, and it might take some time to answer. I would say, though, that psychoanalysis is not just a repository of 'truths', but a powerful force for disrupting them, hence the focus on parapraxes, dreams, symptoms.

    1. We can agree that health systems are political and that they don't operate solely on the basis of available evidence. Pointing this out does not lead us away from the evidence based therapies, but should alert us to the fact that "evidence based" is a coveted term, often used unscrupulously.
      Your advocacy of case studies is well taken-you're speaking to the fact that generalised knowledge needs to be balanced against thinking about the particulars of a given case. I doubt many people would seriously argue against this view, and it finds a home in medicine and mainstream psychology in the form of "formulation". However, it is quite a jump from there to say that RCTs are "intrinsically nonsensical". Limited in the ways you describe? Absolutely, but there is less intangible stuff going on in therapy than you seem to imply. All therapists, including psychoanalysts, draw on techniques which they have a rationale for employing at any given point in the process. If we can't explain and justify these in language suitable for the training of other professionals then it's not clear what business we have using them on the unsuspecting public. Manualisation is not the same thing as telling someone precisely what to do at any given turn and it is interesting to speculate whether the problem of robotic rote-learned therapists (who surely exist) is greater than the problem of incompetents hiding behind the mask of technical psychoanalytic intervention (who also surely exist). Equally, if you can't define (and operationalise, albeit roughly) ways in which a person is currently suffering and stands to benefit from your therapy then the question is begged, what *do* you think you are doing to them?
      There is more to human subjectivity than is dreamt of in the annals of an empirical psychology, but I view the process of trying to operationalise behaviours and experiences as dynamic and contested. Mental health science is sometimes touted as an authoritative "final explanation". It's not, but it is a precise and participatory way of continuing to have the conversations we all need to have about what psychological therapies can (and can't) do.

  4. To Pnin, theories in psychology are not the precise formulae of physics or mathematics. They do not exist independently of the language used to articulate them. A formula in physics can be transmitted by somebody writing incompetently, or in a foreign language. Not so in psychology, where mediocre writing in psychology is strongly suggestive of mediocre thinking. Unlike in the humanities, Anglophone psychologists eschew the practice of close textual reading, and in my view, the discipline as a whole suffers as a result, with shoddy concepts getting halfway around the world before conceptual rigour has put its shoes on.