What you know when you know how someone behaves

Rowland Stout

Centre for Philosophy

University of Manchester

(http://les.man.ac.uk/philosophy/stout/stout.htm)

[1] In chapter 2 of The Concept of Mind, “Knowing How and Knowing That”, and especially in the section on “Understanding and Misunderstanding”, Ryle rejects two approaches to the question of the interpretation of other minds that correspond quite closely with what are now called functionalism, or theory theory, and simulation theory. There is a painful irony here that the functionalist approach to the philosophy of mind, which developed in the late 60s and 70s, has widely been regarded as completely superseding Ryle's own approach.

[2] The argument that functionalists use against Ryle is that his account of the interpretation of other minds cannot accommodate the holism of the mental, whereas functionalism can. Holism must be accommodated, so Ryle is wrong. I will argue that a Rylean approach can and should accommodate the holism of the mental. But there is a real difference between a holistic Ryle and functionalism, and with respect to this difference, the holistic Ryle is to be preferred.

[3] Lewis's classic functionalist characterization of a theory of behaviour (1972, 256) is as a cluster of platitudes of the following form:

When someone is in so-and-so combination of mental states and receives sensory stimuli of so-and-so kind, he tends with so-and-so probability to be caused thereby to go into so-and-so mental states and produce so-and-so motor response.

Such a theory constitutes an entire psychology – folk psychology. It maps mental states and stimuli on to mental states and responses. To work out what someone's state of mind is from the fact that this theory holds good of them you have to solve a huge series of enormously long simultaneous equations to determine what these mental elements must have been to account for the way the observed inputs and outputs fit into the theory.

[4] Ryle's general objection to this sort of approach is that it presupposes that the mind is some unobservable realm whose nature can only be known, if at all, by inference from knowledge of some observable stuff, namely behaviour. “[W]hen we characterise people by mental predicates, we are not making untestable inferences to any ghostly processes occurring in streams of consciousness which we are debarred from visiting; we are describing the ways in which those people conduct parts of their predominantly public behaviour” (1949, 51).

[5] Ryle's moves against the theory theory approach are characteristic of moves made latterly by simulation theorists. Both Ryle and the simulation theorists argue that the interpretation of other minds is based on knowledge how rather than knowledge that. According to Ryle, practical knowledge is more basic than theoretical knowledge. “Intelligent practice is not a step-child of theory. On the contrary theorizing is one practice among others and is itself intelligently or stupidly conducted” (1949, 26). “Understanding is a part of knowing how. The knowledge that is required for understanding intelligent performances of a specific kind is some degree of competence in performances of that kind” (1949, 51). “In making sense of what you say, in appreciating your jokes, in unmasking your chess stratagems, in following your arguments and in having you pick holes in my arguments, I am not inferring to the workings of your mind, I am following them” (1949, 61).

[6] A classic modern version of simulation theory is Jane Heal's.

“On the replicating view psychological understanding works like this. … I can imagine how my tastes, aims and opinions might change and work out what would be sensible to do or believe in the circumstances. My ability to do these things makes possible a certain sort of understanding of other people. I can harness all my complex theoretical knowledge about the world and my ability to imagine to yield an insight into other people without any further elaborate theorizing about them. Only one simple assumption is needed: that they are like me in being thinkers, that they possess the same fundamental cognitive capacities and propensities that I do. The method works like this. Suppose I am interested in predicting someone's action. … What I endeavour to do is to replicate or recreate his thinking. I place myself in what I take to be his initial state by imagining the world as it would appear from his point of view and I then deliberate, reason and reflect to see what decision emerges” (1986, 137).

[7] But Ryle clearly rejects this position too – or rather, he rejects the argument from analogy and the Verstehen tradition out of which two ideas modern simulation theory has evolved. “[T]his does not imply that the spectator or reader, in following what is done or written, is making analogical inferences from internal processes of his own to corresponding internal processes in the author of the actions or writings. Nor need he, though he may, imaginatively represent himself as being, in the shoes, the situation and the skin of the author. He is merely thinking about what the author is doing along the same lines as the author is thinking about what he is doing, save that the spectator is finding what the author is inventing” (1949, 55).

[8] Making analogical inferences from an exercise in replication is as bad as making inferences from knowledge of a psychological theory combined with observations of behaviour as far as Ryle is concerned. In both cases, inferences are being made from some observable realm to some supposedly hidden realm. Ryle rejects the idea that we must look inwards in order to be able to interpret another person. So while interpretation involves the ability to behave as someone else, it does not involve a complicated process of going into and out of different perspectives. Although Heal describes it as only one simple assumption, having to make the assumption that the people you interpret are in fundamental mental ways like you is having to make one assumption too many for Ryle.

[9] Ryle's argument is going to have no weight against someone who seems to be happy to think of the mind of another as an unobservable realm, knowledge of which must be inferred from something that is observable. This is why Ryle devotes so much of The Concept of Mind to trying to show how unhappy that way of thinking really is. But this attempt would be more persuasive if the alternative way of thinking were spelt out with more specificity. Ryle's positive claims about the interpretation of others are so unspecific that functionalists have been able to trivialize them to the point of absurdity with no fear of contradiction.

[10] A simplistic interpretation of Ryle's position concerning interpretation of other minds would be the following. We know that someone believes that it will rain by knowing how they behave (or are disposed to behave). This is knowing that they are disposed to behave in a believes-it-will-rain way. So for every mental predicate there is a corresponding way of behaving. It may seem that Ryle's claim that characterising someone by mental predicates is describing the ways in which they behave commits him to this simplistic position. But there is not obvious reason why Ryle should accept that there is just one believes-it-will-rain way of behaving. Describing a shirt as pink does not commit one to the view that for the colour pink there is one corresponding shirt. No more does describing a way of behaving using a mental predicate commit one to the view that for each mental predicate there is one corresponding way of behaving.

[11] The standard objection to this simplistic view is that there is no such thing as one believes-it-will rain way of behaving. Gilbert Harman (1973) expressed this objection clearly:

“There is no noncircular way to specify the relevant dispositions. For they are dispositions to act in certain ways given certain situations; and the relevant situations essentially include beliefs about the situation and desires concerning it. … But beliefs are dispositions to act in certain ways only given certain desires, whereas desires are dispositions to act in certain ways only given certain beliefs. A belief that it will rain will be manifested in the carrying of an umbrella only in the presence of a desire not to get wet; and the desire for money will manifest itself in acts that tend to get one money only if one believes that those acts will get one money. Since even in theory there is no noncircular way to specify relevant dispositions in pure behavioral terms, behaviorism cannot provide an adequate account of mental processes and experiences” (1973, 41).

[12] The argument is in the first instance an attack on the thought that it is possible to specify a way of behaving corresponding to an individual aspect of a subject's state of mind in isolation from the rest of their state of mind. But the conclusion might have been stronger. It is not just that we cannot specify a single way of behaving corresponding to the belief that it will rain; there is no such single way of behaving.

[13] This sort of argument is wheeled out in one form or another in almost every introduction to the philosophy of mind, where a position called philosophical behaviourism, which is associated with Ryle and is identified with this simplistic interpretation, is dispatched effortlessly.

[14] There may be some hint of the position being attacked by Harman in Ryle's example of John Doe, who is fast asleep, knowing French. “To say that this sleeper knows French, is to say that … he will cope pretty well with the majority of ordinary French-using and French-following tasks” (1949, 123-4). To this we might object that John may have a reason for concealing his ability to speak French – he is a French spy for example – and will in fact cope very badly with the majority of French-using and French-speaking tasks.

[15] But Ryle is clearly only employing this account of knowing French as a toy example to illustrate his notion of a disposition. In general Ryle is not at all committed to the idea that we can know individual aspects of someone's state of mind in isolation from the rest on the basis of knowing how they behave. On the contrary, it would be very natural to read holism into Ryle's account and say that knowledge of someone's overall state of mind consists in knowing how they are disposed to behave.

[16] There are three conceptions of a way of behaving that get used in different accounts of interpretation. According to functionalism or theory theory, we derive knowledge of someone's state of mind from a psychological theory of that person. This theory describes a way of behaving in a very broad sense. It is the person's overall way of behaving, and includes all the combinations of possible propositional attitudes that person may have and what the person would do in those circumstances. The cluster of platitudes described by Lewis describes how a person behaves in all possible circumstances. It does not correspond with a particular state of mind, but rather with a whole psychology.

[17] According to the too simplistic Ryle who is being attacked by the argument from the holism of the mental, there is one way of behaving for every different propositional attitude. There is a believing-it-will-rain way of behaving. And there is another wanting-to-stay-dry way of behaving.

[18] According to a holistic Ryle, knowledge of someone's overall state of mind is knowledge of how they behave. On this conception there is not just one way of behaving per person. Nor is there one way of behaving per propositional attitude. There are as many ways of behaving as there are possible states of mind. There is one way of behaving which is believing that it will rain at the same time as wanting to stay dry and intending to go out, etc. There is another way of behaving which is believing it will rain at the same time as wanting to get wet and believing that going out into the rain is the best way of getting wet, etc.

[19] The challenge for holistic Ryle is to explain how we are entitled to attribute individual propositional attitudes to someone. Our knowledge of overall ways of behaving is in the first instance undifferentiated, but our way of describing their state of mind is in the first instance differentiated into specific propositional attitudes. In following someone's way of behaving one knows that they are behaving this way. And such knowledge constitutes knowledge of their overall state of mind. But the “this” here picks out a way of behaving without giving any way to decompose it into its constituent attitudes. The problem is that we do not have one word for each overall type of state of mind. We communicate about the mind by referring to types of attitudes that together make up such overall states.

[20] Holistic Ryle seems to be faced with the challenge of finding what is in common between different overall states of mind and their corresponding ways of behaving in order to factor out individual propositional attitudes. This might be seized upon by functionalists as the advantage of their approach. Functionalists do not have the problem of how the interpreter splits up overall states of mind into their constituent attitudes. The functionalist starts with a psychological theory that comes with its compositional structure exposed from the beginning.

[21] But, as simulation theorists like Heal (1986) have remarked, functionalist theories of behaviour have the disadvantage of being unknowably complex. If they are really to capture an entire psychology then they must be infinitely large, since there are infinitely many possible combinations of propositional attitudes that must be included in such a theory.

[22] The only way to unify the infinite set of platitudes that must appear in a proper functionalist theory would be to invoke rationality. Invoking rationality yields the very simple theory theory, where the theory consists of just one platitude: you do what it is rational for you to do given your beliefs and desires. But such a theory looks much less like the sort of thing functionalists would be happy with. The problem is that rationality is not finitely specifiable, and certainly not as part of a theory which might count as properly scientific.

[23] This is why McDowell (1985, 391) regards Davidson's (1980, 223) requirement that interpretation must be constrained by the constitutive role of rationality as an argument against functionalism. Heal (1986) takes the requirement that the interpreter have some conception of rationality as an argument against theory theory of any sort. Their thought is the Wittgensteinian one that you can only have a conception of rationality by engaging yourself in the practice that embodies that conception. There is no place for this personal engagement in theory theory.

[24] But even acknowledging that a proper folk psychological theory of behaviour must invoke rationality does not solve the problem of how such a theory can realistically be used in the interpretation of other minds. Knowledge of the theory that incorporates the constitutive ideal of rationality would have to involve knowledge of the constitutive ideal of rationality. But being an ideal, complete knowledge of it is unrealistic, whether this is knowledge that or knowledge how. Thinking of interpretation from the point of view of a complete theory of psychology with a completed conception of rationality is assuming a sort of God's eye interpretation, and this is clearly unrealistic.

[25] One way of seeing why final completed theories of behaviour are impossible ideals is this. However good the theory, there will always be a way of redescribing the situation of the agent in a way which that theory has got no answer for. There is no designated way of categorizing different situations that a theory of behaviour should work with. Any situation can be described in an indefinite number of different ways and the discovery of a new way of describing a situation may be the discovery of a new conclusion as to what behaviour is appropriate in that situation.

[26] This entirely open-ended process of describing and redescribing the world is part of practical rationality. This is a lesson that David Wiggins (1975) and John McDowell (1979) claim can be derived from Aristotle's discussion of aesthesis. A functionalist-style completed theory of behaviour has essentially fixed on one level of description of the situation and in doing so has closed the door on this process of discovery. Since this process of discovery is part of what it is to be practically rational, no complete theory of behaviour can capture a conception of rationality; and so no complete theory of behaviour can capture an individual's overall psychology.

[27] But without knowing any infinitely large theory or having a completed conception of rationality we can interpret others. To explain this, we must think of interpretation dynamically – as always partial but on the road to improvement. This I take it is how one should think of rationality too. A completed theory of behaviour corresponding to an overall psychology is an ideal and so not a stage in the actual process of interpreting people. But knowledge of a way of behaving that one has not yet decomposed into its constituent attitudes is an entirely manageable stage in the process of developing knowledge of someone's state of mind.

[28] So let us start with an interpreter who knows a way of behaving. To begin with they only know it as this or that way of behaving. They can follow it, recognize it, and perhaps even replicate it. Probably the first thing they can do with it is share it. It seems very natural to assume that joint activity and joint attention develop first. So what the individual knows about this way of behaving is that it is how we are behaving or were behaving.

[29] Of course their knowledge is not just constituted by this fact. They do not just know something about the way of behaving; they also know the way of behaving itself. Roughly speaking, this means that they know how it is appropriate to behave in different situations given this way of behaving. According to holistic Ryle, knowing that someone is behaving in this way is knowing what state of mind they are in. But such knowledge is not yet articulated into knowledge of the component attitudes in that state of mind.

[30] Let us suppose that according to this way of behaving, if it will rain you take your umbrella if you are going outside though not if you are staying inside and if it is not going to rain you leave your umbrella whether or not you are going outside. It is now possible to work out a new way of behaving with the fact that it will rain as a given fixed point. If you know how to behave whether or not it will rain, you know how to behave on the assumption that it will rain. If you know that someone else is behaving in this restricted way, you can automatically attribute to them the belief that it will rain. And you can do that without being in a position to attribute any other beliefs or desires to them.

[31] This may appear to contradict the concession towards holism made earlier on behalf of Ryle. It now seems that there is a believes-it-will-rain way of behaving after all. But, in fact what we have here is not a way of behaving that corresponds to the belief that it will rain, but a way of behaving - one among many - that entitles you to attribute the belief that it will rain. Being able to attribute this belief on the basis of knowing this way of behaving does not mean that you are in a position to attribute that belief generally. So far you are only in a position to attribute that belief embedded in a very specific psychological context.

[32] The way of behaving we started with corresponded to a certain state of mind, but was variable with respect to various features of the situation. Just knowing this way of behaving does not give you knowledge of how to describe that state of mind. But it does give you the basis for deriving a series of more limited ways of behaving where the variable features of the situation are kept fixed. Knowledge of these more limited ways of behaving and their relationship with the initial way of behaving enables you to attribute beliefs concerning these variable features.

[33] Interpretation proceeds in a series of crabwise manoeuvres. The next step is to enrich the initial way of behaving by learning how to discriminate different kinds of situation and learning how to behave given the different possibilities for these situations. Then by holding some of these new features fixed we develop ways of attributing new beliefs.

[34] The method should apply to intentions in a parallel way. You might start off knowing a way of behaving according to which you act in one way if G is to be achieved and in another way if G is not to be achieved. Given this you know how to behave on the assumption that G is to be achieved, and if someone else is behaving in this way, then you can say that they intend to achieve G.

[35] Just how this approach should deal with the interpretation of other psychological aspects of someone's overall state of mind – their emotions for instance - need not concern us here. If the approach is plausible for beliefs and intentions, then the potted objection to a Rylean approach that it fails to accommodate the holism of the mental is refuted.

[36] In summary, any way of behaving, however inclusive and advanced, will express certain rigidities and fixed assumptions. There is no such thing as a perfectly adaptable way of behaving. Any way of behaving works with some fixed conception of how to distinguish between different situations. So no way of behaving can be entirely neutral as to the state of mind of the agent, and any way of behaving corresponds to a certain state of mind, rather than to an entire psychology.

[37] Knowing a way of behaving does not in itself give you compositionally structured knowledge of that state of mind. That is just what is hidden from you if your way of behaving takes that state of mind for granted. But knowing a way of behaving can be the basis for deriving knowledge of more limited ways of behaving that take even more for granted. Being able to see how someone behaves in these more limited ways, gives us knowledge that they are taking these things for granted themselves. But this sort of knowledge can be built up one belief at a time.

[38] Saying that someone has a certain belief is describing their way of behaving. This is not because there is a specific way of behaving corresponding to that belief. Rather it is because their way of behaving corresponding to their overall state of mind has the property of being committed to a certain assumption. When a person's way of behaving has this property we describe the person as holding that assumption themselves.

References

Davidson, D. (1980), Essays on Actions and Events, Oxford: Clarendon Press.

Harman, G. (1973), Thought, Princeton NJ: Princeton University Press.

Heal, J. (1986), “Replication and Functionalism”, in Butterfield, J. (ed.) (1986), Language, Mind and Logic, Cambridge: Cambridge University Press.

Lewis, D. (1972), “Psychophysical and theoretical identification”, Australasian Journal of Philosophy 50, 249-58.

McDowell, J. (1979), “Virtue and Reason”, The Monist 62, 331-50.

McDowell, J. (1985), “Functionalism and anomalous monism”, in Lepore, E. and McLaughlin, B. (eds.) (1985), Actions and Events, Oxford: Blackwell.

Ryle, G. (1949), The Concept of Mind, London: Hutchinson.

Wiggins, D. (1975), “Deliberation and Practical Reason”, Proceedings of the Aristotelian Society (1975-6), 29-51.

---

© Rowland Stout

EJAP is a non-profit entity. EJAP may not be printed, forwarded, or otherwise distributed for any reasons other than personal use. Do not post EJAP on public bulletin boards of any kind; use pointers and links instead. EJAP retains the right to continuously print all the articles it accepts, but all other rights return to the author upon publication.

EJAP is hosted at The University of Louisiana at Lafayette . Please send comments and questions to ejap@louisiana.edu .