Monday, 25 February 2013

Why Psychiatry Needs a Unified Theory of Mental Illness

Both practice and research in psychiatry, as with any other medical speciality will always be constrained by underlying philosophical and ethical assumptions. In order to describe, discuss, debate or deride any theory, we necessarily assume an ontology of the subject at hand - that is our view on the nature of things as they are. Be it implicitly or explicitly this is simply inescapable. At least to me it is.

Try explaining mental illness without alluding to its nature in some way. Its impossible. You might, rather foolishly like Guze, say its nothing more than a brain disease, or perhaps argue for some form of Freudian psychological malfunctioning or grand scale failing on the part of society. But in each instance you are assuming a particular view on the nature of the mind and mental illness. What's more is that such a view constrains the kinds of methodology that you can use. If mental illness is a brain disorder, why bother with psychology? Go straight to the molecules! If its the subconscious torments of sexual frustration, sod the CBT and whack out the Rorschach cards. You get the point. Ontology constrains methodology and treatment for that matter.

An ontology in the sense that I am using it is what Kuhn termed a paradigm; the overriding theory about the existence of nature within which your hypotheses and explanations exist. Such paradigms change and arguably lead to better more parsimonious explanations, the shift from Newtonian physics to Einstein's theory of relativity being a common example.

But there's a problem for psychiatry. The nature of the mind is a slippery one and psychiatry hasn't quite decided what exactly it is yet. This isn't the fault of psychiatrists, it simply reflects the complexity of the subject at hand. Historically, psychiatry has been grounded in a variety of paradigms from biological reductionism to Freudian psychoanalysis to the biopsychosocial model. All which have failed to offer a suitable ontology in which to explain, investigate or treat the mind and mental illness (see here).

It is interesting then, that in light of the failings of previous theories, psychiatry seems to be attempting to subvert the problem of ontology by appealing to American pragmatist philosophies of Pierce and Dewey, as Brendel explains:

Pragmatism, as I am using the term in the current context, refers to the theory that psychiatric explanations are ‘‘true’’ only insofar as they promote beneficial real-world results for individuals with mental illnesses. As a practical discipline, psychiatry is concerned more with its methodology than its ontology: by adopting a pragmatic position on explanatory models, psychiatrists do not necessarily commit themselves to a particular view on the underlying structure of the universe. Psychiatric explanations are coherent and plausible insofar as they are pragmatically useful and empirically testable in clinical settings.

One such feature of the pragmatist approach that has emerged is the application of Karl Jaspers’ Methodological Pluralism. He explains:

Instead of forcing the subject-matter into a strait-jacket of systematic theory, I try to discriminate between the different research methods, points of view and various approaches, so as to bring them into clearer focus and show the diversity of psychopathological studies. No theory or viewpoint is ignored. I try to grasp each different view of the whole and give it place according to its significance and limitations…We obtain our facts only by using a particular method. Between fact and method no sharp line can be drawn. The one exists through the other. Therefore a classification according to the method used is also a factual classification of what is, as it is for us.

Speaking to a psychiatrist recently, he informs me this seems to be the fashionable take on things for those in research and in practice. In other words, we should not concern ourselves with the difficulty (and it is very difficult) of explaining the ultimate nature of mind and mental illness but rather use all the available theories and research, picking and choosing based on their strengths and limitations, constrained by evidence based medicine.

At first glance this seems a decent suggestion and is given further stability through Kendler's argument for levels of explanation. We can explain the same phenomenon at various different levels becoming more and more abstract from the levels of particles to chemicals to biology to systems etc. Depending on what we want to explain we have to choose the correct level. It would not, for example, be particularly useful to explain the heart at the level of particle physics.

Such an approach is a sensible temporary solution to guide scientific investigation into mental illness in the absence of an overriding ontology in order to avoid the dogmatism of past paradigms such as biological reductionism. The problem, however is that it must be recognised that it can only ever be a temporary solution, since the pragmatist pluralist paradigm offers no ontology of mind to guide methodology. Whilst research can be understood within the particular level of explanation, constrained by that particular ontology, pluralism does not offer a means by which we choose or reject certain levels or indeed integrate the different levels. As such we are simply stuck with the same kind of eclecticism offered by the biopsychosocial model, by which we have scattered and often incompatible theories and research.

Pluralism is correct in emphasising the need for a focus on the different levels of explanation available but fails in its assumption that we do not need a unifying theory to tie it all together. In order to truly understand and treat mental illness we need a unified theory of mental illness to guide and constrain the kind of integrative methodology the pluralist approach entails.

Brendel, David H. "Reductionism, eclecticism, and pragmatism in psychiatry: The dialectic of clinical explanation." The Journal of medicine and philosophy 28.5-6 (2003): 563-580.

Ghaemi, Seyyed Nassir. "Paradigms of psychiatry: eclecticism and its discontents." Current opinion in psychiatry 19.6 (2006): 619-624.

Guze, Samuel B. "Biological psychiatry: is there any other kind." Psychological Medicine 19.2 (1989): 315-323.

Kendler, Kenneth S. "Toward a philosophical structure for psychiatry." American Journal of Psychiatry 162.3 (2005): 433-440.

Kuhn, Thomas S. The structure of scientific revolutions. Vol. 2. University of Chicago press, 1996.

Jaspers, Karl. General psychopathology. Vol. 2. Johns Hopkins University Press, 1997.


  1. Am I correct in inferring that that via your analogy with the search for a unified theory in physics you are implying that psychiatrists should invoke the same approach? Its all very well we saying we need a unifying theory to guide research and methodology, but in physics the research is centred around trying to find a unified theory in the first place. Do you mean you wish psychiatrists to also search for a unified theory, as in a culture change towards this kind of thinking or do you mean somebody should concoct a new theory that attempts to unify current research?

    As to clarify in physics the approach is not to tie together situational theories or opposing theories to provide direction, the goal is to find one single theory that will replace existing theories and be applicable in all cases. It can ingest and combine multiple theories if they are all consistent however then they are not treated as separate theories any more but as part of one larger theory.

    Classical mechanics (Newtonian physics) although simpler to compute and often easier to implement only provides a rough approximation to what is actually going on. Classical theories are useful on a macro scale (eg building bridges and such)but break down at a quantum level. Quantum theories provide a much higher degree of accuracy and and can account for phenomenon that make no sense within the constrains of classical mechanics (eg It would be impossible to build modern electronics using classical mechanics). You could also model how to build a bridge using quantum mechanics and actually achieve more accurate predictions than via classical methods however modelling large things from a particle scale is more computational intensive.

    You say it would make no sense to model the heart at a particle scale however this is not the point. The real achievement is the fact that you are actually able to model a heart on a particle level and it produces the the same results as modelling on a biological level. Therefore you can have confidence in your particle theory and apply it to other less well biologically understood parts of the body.

    The search for a unified theory in physics is centred around quantum theories as accurate results can be calculated at any scale, micro or macro. Instead of having different approaches and theories that work in carefully defined cases the search is to find one theory that will provide accurate predictions no matter what the set of circumstances or their complexity.

  2. Hi Brennan thanks for commenting. Before I have a go at answering, the analogy was simply meant as a demonstration of how the paradigm you are working under changes the way you conceptualise and investigate your subject matter. I think the similarities between physics and social sciences such as psychiatry stop there. As noted in the Ghaemi paper physics is too often used to demonstrate the 'ideal' scientific methodology though I do not think this should be the case, as I will explain.

    Here goes:

    1) My argument is that psychiatry needs an explanatory model of mind that unifies all the research. Any attempt to unify all the research requires a theory of/explanation of/definition of etc what mind and mental illness actually are. Of course this will be constrained within the principles of more basic sciences such as physics, (this is why there was obviously such an outrage at that shitty Bem paper which I linked you). I agree that as such this will mean that individual 'theories' or rather what Kuhn would call 'hypotheses' will thus become part of one unifying theory of mind but not one unifying theory of everything.

    2) In terms of the ultimate goal of physics, I'm not well read enough to comment but are you saying the ultimate goal is to basically provide a theory of everything? Or are you saying it is to provide a theory that will constrain all other levels of research? If if its the former I think this confuses ontology with provenance.

    The idea of levels of explanation is that there is not only one level at which to explain a phenomenon. Levels can 'supervene' on one another without conflicting but also without being reducible down the hierarchy. One such idea is that of emergent properties which presumably must have come from physics, the stronger version of which argues we can get emergent properties arising from lower level constituents that cannot be predicted simply based on those lower levels. In other words Gestalt psychologists had it right when they said 'the sum is more than the whole of its parts'.

    In this sense if the mind is an emergent property you will never ever be able to explain it at the level of particles. It will not contradict physical principles but it will emerge from them irreducibly and unpredictably. One example, and I can't remember where from, is that it is like trying to understand the meaning of the sentences in a book at the level of syntax. The meaning emerges from but is not reducible to the letters on the page.

    In the case of the heart example, which presumably is not an emergent phenomenon, modelling the heart at a particle level would be helpful for physicists but not for biologists who would need to conceptualise and explain it at the level of anatomical and physiological systems and processes. So I am not arguing psych needs to provide its own theory of everything but rather a suitable model of mind which would work within the principles of physics but at a level of explanation that serves it best. Surely this cannot be particles, even though these may well make up the human from which the mind emerges? I don't for example understand how you could explain social constructs at the level of particle physics - ever. In the paper I referenced, Kendler gives some further examples.

    As an aside there are those such as Chalmers who propose property dualism which argues there are two properties of physicalism physical and mental. This does not challenge the progress of physics, nor resort to supernatural substance dualism but rather that the physical property that physics uses as its ontology is only one of two comprising the ultimate nature of reality, the second being a mental property. Not read up on this yet and over my head but might interest you! In this latter case this would not seek to provide an alternative theory of everything but argue that such a theory would have to account for both properties of physicalism.

  3. I haven't had time to develop a real reply to this yet, but feel I should write something!

    Some minor notes: I don't like the way "pragmatic" is used by Brendel. That is perfectly fine dictionary use of the term, but doesn't connect strongly with what Peirce, Dewey, and James were up to.

    Jaspers ideas (at least as presented here) also seem kind of lame. A unified theory creates winners and losers. Surely we do need to discuss methods when considering results, but that principle does not create unification nor consensus.

    Overall, though, I think you are right that unification is needed, and I like a bunch of the stuff you say in your comment.


    P.S. About Chalmers, you say: "As an aside there are those such as Chalmers who propose property dualism which argues there are two properties of physicalism physical and mental. This does not challenge the progress of physics, nor resort to supernatural substance dualism but rather that the physical property that physics uses as its ontology is only one of two comprising the ultimate nature of reality, the second being a mental property."

    That seems all very silly. If there are two fundamentally different kinds of stuff, you have all the same problems... and calling them both "real" doesn't help anything.


    P.P.S. Just to throw a Holt quote into the mix... this is from 1915:

    We are prone, even the 'behaviorists' among us, to think of behavior as somehow consisting of reflex activities. Quite true, so far as it goes. So, too, coral reefs in the last analysis consist of positive and negative ions, but the biologist, geographer, or sea-captain would miss his point if he conceived them in any such terms. Yet we are doing the very same thing when we conceive the behavior of a man or animal in the unintegrated terms of neural process ; which means, agreeably to the bead theory, the impinging of stimulus on sense-organ, the propagation of ionization waves along a fiber, their spread among various other fibers, their combining with other similar waves, and eventually causing the lowered or heightened tonus of muscle. All this is happening. But our account has overlooked the most essential thing of all the organization of these processes.

    1. Cheers for the comment Eric. I was hoping to get your opinion since I seem to recall you were putting together a journal special on unifying psychology? When is that out?

      Do you have any links on the pragmatist approach? As I understand it, it was a move away from the abstract philsophies of Hegel etc, to make philosophy relevant again. In all honesty I don't really understand it beyond simply saying "philosophical debates are only important/useful/truthful/meaningful if they have real, practical implications for [society/research etc]" This doesn't really seem to say much beyond stating the obvious!

      In terms of psychiatry my main issue is that I don't see how we can make real explanatory advances just from shit loads of data. We need a framework to constrain a) how the data is collected and b) how it is interpreted. Since psychiatry does not have any such theory the data itself, explanations and attempts to incorporate across levels of explanation surely won't be reliable?

      I take your point re property dualism/neutral monism. I know very little about it. It just sounded like an interesting take on the mind/body problem. If I find the time for the reading, I might do a post on it soon.

      Lastly, thanks for your post on Taylor and Holt. was an enlightening read.

    2. Oye! No time to elaborate on Pragmatism now... but...

      In all honesty I don't really understand it beyond simply saying "philosophical debates are only important/useful/truthful/meaningful if they have real, practical implications for [society/research etc]" This doesn't really seem to say much beyond stating the obvious!

      Imagine if instead you said that any issue where both side lack a practical implication, both sides are identical. If there is no consequential difference, there is no difference. Period.

      That is one of a handful of ways to try to get a toe in on pragmatism.