Lennart Nordenfelt's Rationality and Compulsion is a new book in the Oxford University Press book series International Perspectives in Philosophy and Psychiatry. It aims to articulate a philosophical account of action ('action theory') and then deploy this to shed light on aspects of psychopathology. In explicitly placing action (and implicitly placing values) at the heart of an analysis of mental illness, Nordenfelt's work has echoes of that of Bill (KWM) Fulford. But the style and the analysis is distinct. Nordenfelt is more of a systematic and substantial theory builder, less interested in philosophical 'diagnosis' of opposing views.

Although the first half of the book merely summarizes his approach to the philosophy of action, which is expressed more fully in two other books, it is still a substantial work of analysis and provides a clear approach to action explanation through practical syllogism. The second half develops an account of compulsion in part through critical comparison with other accounts. The book is clearly worth careful study and has something to say both to newcomers to the fields of philosophy of psychiatry more broadly and, more narrowly, to action-theoretic accounts within it. It is precisely for this reason that the rest of my review will be critical in tone. It is a book worth criticizing.

I have two main qualms about it. The first has to do more with the atmosphere, rather than the analysis itself, of the first half. Nordenfelt's description of action seems to be one in which an agent is always to some extent alienated from his or her actions.

One clue to this is what may merely be a throw-away remark [ibid: 78] that a 'want' to do F is a 'disposition to form an intention' to do F. (Actually, he suggests a want is a disposition and then in the next sentence he says that if one wants F then ipso facto one will have a disposition to form the intention, which isn't quite the identity hinted at.  I assume, however, that that is what he means.)  But on the assumption that dispositions to form intentions are often long-standing traits whose actualizations are sensitive to environmental opportunities this suggests a very wide gap between one's wants at any given time and what could be brought to consciousness at that time.  (One will have wanted things that one is disposed to select when subsequently offered them even if one did not previously know of their existence.)  More significant, however, is that I take it that wants can be the sorts of things that one offers as reasons for one's intentions.  My long-standing wants concerning pizza helped to rationalize my intentions to act when a new Pizza Express opened nearby in Kendal, for example.  But that rational answerability seems to be missing from a merely dispositional analysis.

Secondly, Nordenfelt comments at the start of the book [ibid: 4] that actions are abstract and thus cannot be observed 'with the senses'.  This is surely a startling claim, deserving more defense than it is able to get within the space of the book.  The argument for it seems to start with the idea that actions are behavior plus mental states and couple this with a Cartesian assumption that others' mental states cannot be perceived with the senses.  But it seems to me that the conclusion that others' actions cannot be observed is sufficient to cast doubt on the Cartesian assumption, rather than the other way round.

The third hint at an alienated conception of action comes from a description of understanding other people's action which flows from their perceptual inaccessibility and which I will quote at length:

What does it mean to answer the question 'what is this?' in relation to actions? Assume that two persons (A and B) witness the occurrence of a behaviour performed by a man C. Assume that they witness C is waving his hand. A asks B what is going on.
B's first task is to identify the performance as an action in the first place. The context is not legal, so B can wholly concentrate on the issue of intentionality. Does this behaviour issue from some intention of C's? B's task is one of interpretation. The answer is not self-evident since the intention in question, being a mental event, is not observable. B has to find evidence for assuming the existence of the intention. Different methods for this exist. If C is approachable for questioning one can simply ask him what he is doing. This is a recommendable procedure in many instances. However, it is not waterproof. C may lie about his intentions… A second task is therefore to put C's behaviour in a context. Where does the waving take place? Is there a human being in the neighbourhood to whom the waving can be directed? Is there anybody else doing the same thing as C?
Further observation by B may detect the following circumstances. C is standing together with a number of other people, all waving their hands, in front of a building. In a window on the upper floor of the building a man is standing looking at all the people outside. It is evident that the group outside are calling this person's attention to something. Thus, the primary question can be answered: C's behaviour is intentional. It is an action.
Preliminarily, B can now identify C's action as an action of waving. [ibid: 69-70]

This strikes me as a very odd view of understanding. If one had to do all this work merely to understand that C is a part of a crowd waving at someone, one would be a very unusual kind of person.  Now, of course, it may not be an account of everyday phenomenology but rather a reconstruction of how, if challenged, one might justify one's normally seamless understanding.  But that is not how it is presented and I fear that by starting with such an 'alienated' conception even a piece of reconstructive epistemology will be misleading.

My second main qualm concerns the analysis of compulsion in the second half of the book.  Nordenfelt makes it a condition of adequacy of any analysis of compulsion that it reflects the fact that if someone is compelled to do something then they could not avoid doing it [ibid: 141].  Thus the analysis of obsession is a good opportunity to see what the action-theoretic framework can do for understanding psychopathology.  He considers the example of a subject who obsessively checks that she has turned off a tap.

The subject believes… that a risk exists that the water is still running from the tap. At other moments… she believes that this cannot possibly be so. Subjectively, overwhelming evidence also exists that the taps have been turned off. The situation can perhaps be described as follows: the subject cannot help having a completely unjustified, indeed foolish, belief…

[I]t is primarily the belief that the tap is running that is compelling. [ibid: 177]

But there is no explanation of the tension in this case, the point that the subject believes both that the tap may be running and that it cannot be. Elsewhere, in a case involving wants as well as beliefs, Nordenfelt suggests that there are 'conceptual' limits to both irrationality and to any lack of fit between mental states and actions.

If a person genuinely wants to do F, has no conflicting want to do non-F, believes that he or she can do F, is capable of and not impeded from doing F, then he or she must for conceptual reasons at least try to do F. If A does not attempt to do F, then something must be wrong with the premises. [ibid: 117]

And earlier he expresses doubts about the psychological possibility of entertaining incoherent or conflicting beliefs:

A person who believes that both p and non-p at the same time has an incoherent belief. A logician would say that such a belief is irrational and logically forbidden. An interesting problem is whether such a belief is even psychologically possible. [ibid: 100-1]

He goes on to argue that it is psychologically impossible to entertain a belief in the incoherent conjunction but that it is possible to hold a conjunction of beliefs which turn out, on investigation, to be incoherent.  But this will not apply to the obsessional case (it is not a matter of investigation that the tap cannot be both on and off).  Thus the account of the obsessional case does not seem to advance our understanding of it very far.  Subjects of obsessional thoughts are compelled to have conflicting beliefs but we are not explicitly told very much more about the nature of that compulsion.

Perhaps the best possibility for developing an account of obsession from the materials developed in the book comes from its separation of rational and causal factors.

If no contact (or no relation whatsoever) exists between a particular want or set of beliefs, on the one hand, and an action on the other hand, it is not sensible to talk about the former being defective reasons for the latter. In order for a defective reason for F to occur, some relation must exist between the reason and the performance of F. And consider again the two plausible relations in this context, namely a causal relation and a rational relation. [ibid: 114]

Nordenfelt goes on to suggest that a causal relation can exist even when the wants and beliefs 'defectively rationalize' the action.  Further there is much talk of reasons being 'strong' or 'idle' which seems to mean causally strong or weak.  But since he argues that '[r]rationalization is a relation totally different from causation' [ibid: 97] this raises the question, which is not addressed at any length, of what connection there is between rationalizing force and causal force.  Nordenfelt seems to assume a harmony between these two factors in most cases without saying why the 'conceptual reasons' that obligate action in the case quoted above also cause it.  But this is surely the key question for a book that examines the relation between rationality and compulsion.

In fact, past experience suggests that Nordenfelt would have ready responses for these sorts of challenge and that is also part of the pleasure of the book. It gives a very strong impression of a philosopher engaged in difficult questions within the philosophy of mental health, with an ongoing research program designed to answer some issues about agency and psychopathology but writing in a way which is accessible to any potential reader with a bit of application.

© 2007 Tim Thornton

Tim Thornton, Professor of Philosophy and Mental Health, The Centre for Ethnicity and Health, University of Central Lancashire

Comment on this review