Sunday, 24 August 2014

Schizophrenia as a "Disjunctive Category": Does it Matter?

Among the commonly articulated criticisms of the concept of Schizophrenia is Don Bannister's claim from the late 1960s that the diagnosis is unfit for scientific research because it is a "disjunctive category". This point recently raised its head again in the comments section of an article about the findings of the Schizophrenia Genetics Consortium (see David Pilgrim's third comment). What does it mean for us to say that Schizophrenia is "disjunctive"? Here is a quote taken from a Schizophrenia Bulletin paper by Bannister (published a few years after the BJPsych piece Pilgrim mentions):

Essentially Bannister is concerned about the fact that any given pair of people with a diagnosis of Schizophrenia can have entirely different behavioural presentations. On the face of it this looks very problematic; Schizophrenia is behaviourally defined, so it seems a little counter-intuitive that the definition can-in theory- capture radically different sorts of behaviour without any overlap among them. Bannister's "disjunctive" point therefore seems to land a blow on those who do research into Schizophrenia, and it provides critics with a nifty sounding slogan for their claims ("Schizophrenia is a disjunctive concept, so there!"), but to what extent should the "disjunctive" argument actually be a worry?

Although Schizophrenia is defined behaviourally, it has been a key assumption throughout the term's history that there is something behind that behaviour that requires explaining. Psychoanalytic theories have put more emphasis on a loss of ego boundaries and self-integration while biological theories traditionally focus on the action of neurotransmitters. Modern trauma theories put more stock in the notion that dissociation may play a role. What all these explanatory ideas have in common is that they attempt to explain the diverse range of behaviours that lead to a diagnosis.

Without any prejudice as to aetiology or mechanism therefore, we can say that the behaviour of a person with a diagnosis of Schizophrenia is not meant to be the main fact about them that determines the presence of Schizophrenia. Instead, the diagnosis is conferred when the behaviour of the patient gives the clinician some reason to hypothesise that the underlying process (aberrant dopamine signalling; disintegration of the self/other boundary) is taking place in the mind or body of the person being assessed.

We can of course argue that the diverse presentations of Schizophrenia are not in fact caused by a single underlying process, but that is an empirical matter.  If there is a single underlying process (or family of interconnected underlying processes), then the fact of Schizophrenia's being behaviourally "disjunctive" is no more interesting than the fact the fact that the same virus can lead to both diahorrea and vomiting. If there is no underlying process (or family of interconnected underlying processes) then that speaks against a single "Schizophrenia" whether or not it is "disjunctive" in Bannister's sense.

Now, just because the "disjunctive" argument may be a red herring, it does not follow that the concept of Schizophrenia can have a free pass in the clinical and scientific lexicon. There are many reasons to be dissatisfied with Schizophrenia-talk, both from a scientific perspective and from a political/social care perspective. My point is not to defend all the ideas associated with one limited reading of "Schizophrenia", nor is it to seek necessarily to preserve its use as a term. Rather it seems we should focus our attention on thinking about what sort of thing or things we really believe "Schizophrenia" refers to.

Many researchers are well aware of the contested nature of Schizophrenia and their work is about understanding how viable a category it is and what is actually going on with people who get the diagnosis. They know that the Schizophrenia of the 1960s is radically different to the Schizophrenia of the 1980s, which has in turn evolved between then and the present day. The fact that a research programme is oriented around the broad family of issues that goes under this name should not be taken the a sign of institutional myopia that many believe it to be.


  1. Your argument begs the question: WHY should one believe that characteristics A & B have the same cause as characteristics C, D, and E? The valuable point the "disjunctive" argument tries to make, I think, is that it is quite possible that there are at least two conditions that often co-occur: one that causes A and B, and another that causes C, D, and E.

    Let's say 60% of the individuals labeled schizophrenic have A through E, while 20% have A and B only and 20% have C, D, and E only. Those empirical results could be accommodated at least as well by two distinct and co-occurring disorders with different causes as by a single disorder which inexplicably produces different symptom patterns.

    It sounds like you're saying scientists are privileging their theories of what causes schizophrenia over the actual empirical evidence--the behavior of schizophrenic people. Surely this seems unscientific? First, look at the variety of theories you've cited and how much they've changed over time--the theories seem less solid, more subjective, and more based on the intellectual fashions of the day than the symptoms they are trying to explain. Second, doesn't it seem backward to have an explanation for a thing without first clearly specifying the thing to be explained? To me, it seems like putting the cart before the horse. Most importantly, science is defined by empiricism. You can theorize in any intellectual discipline; what defines science is that the standard for truth is whether your theory is supported by evidence so ultimately, it's surely the observations on the ground that should be considered most important. And in this case, wouldn't that be the pattern of behaviors of people diagnosed schizophrenic? Lastly, I don't see an easy empirical way to determine based on the behavioral evidence whether, say, a 60-20-20 patient symptom pattern split is caused by one single disorder or two different ones. Since conclusions in science have to be based on testable predictions and this doesn't seem testable, how can one strongly believe in one of these hypotheses (that there is a single cause for schizophrenia) over the other?

    It just doesn't make sense to me to dismiss such an insightful argument as a "red herring" or an excuse to coin a cool term. Am I missing something here?

  2. Hi Emily, thanks for commenting. I think we agree more than may be apparent

    Science is indeed empirical, and there are clearly empirical findings that would scotch the notion of a single-unified Schizophrenia. There may already be good reason to cleave-off some presentations and think of them as being something distinct. That is a question that seems separate from the "disjunctive" definition; having to do more with how well different groups of symptoms actually "hang together" in reality than with how Schizophrenia is defined. Don Bannister's point seems to be that it is somehow logically incoherent to have a disorder definition that contains multiple, potentially mutually-exclusive presentations. I don't see why this is so. Plenty of other disorders are "disjunctive" in his sense but nonetheless coherent. That is why I don't think it's enough to raise Bannister's argument when critiquing Schizophrenia.

    However you raise the point that certain patterns of data would undermine a single Schizophrenia; I absolutely agree. Is an interesting question when empirical findings do and do not cause difficulty for diagnostic entities. How problematic that pattern of findings (60% A-E; 20% A&B; 20% C,D&E) actually is depends on how coherent we think our theories are and on how much validation they have from other avenues of empirical inquiry. Psychiatry perhaps lacks a strongly compelling theory to group together all of Schizophrenia's presentations, though there are some interesting contenders that I don' think can be dismissed lightly either.

    Another big problem for the Schizophrenia diagnosis is the fact that it arose as a particular time in the history of psychiatry. While it was certainly somewhat empirical (in that it arose to describe actually-existing patients) it was also defined in part by the prejudices of the people who coined it. We have been lumped with a version of their concept ever since. This is surely a version of the cart-before-the-horse that you describe.