Recommended Reading

G. Ryle (1949) The Concept of Mind, ch. 1.

Jaegwon Kim (1996) Philosophy of Mind, pp. 25-46. Westview Press.

Methodological Behaviourism

In the early 20^th C psychology was concerned with the study of phenomena of consciousness and other distinctively ‘mental’ properties in Descartes’s sense of the term. The method by which this study was pursued was that of ‘introspection’. People were trained in the techniques of introspection so as to make their discoveries ‘reliable’. But consider how such training could be conducted: how could an erroneous discovery be identified or corrected?

This reliance upon subjective data denied psychology’s scientific ambitions. A science must rely only upon objective data, repeatable experiments. The reaction was to reject the data of introspection as appropriate for a science and to restrict psychological data to just the relevant observable data of the organism. Physiological data are obviously appropriate, and behavioural data were held to be so also. Behaviour it was felt could be defined, recognised, categorised, etc. objectively. For the same reasons, psychological theories were to explain and to make predictions about behavioural phenomena only.

Note that for psychology as defined in this way the mental phenomena do not count as psychological data, but that doesn’t mean that their existence has to be denied. It is generally held, however, that psychology should not assume mental states as theoretical objects, since this would reintroduce subjective objects into the science. But is this ban a necessary consequence? Compare the hypothesis of desires to theoretical posits in other sciences. Is there a difference between the unobservability of electrons and the unobservability of thoughts? Yes, the electrons are supposed to be objective. Subjective objects are unobservable and untestable in principle rather than just as a matter of fact. But what if there were non-mental objects like quarks that were non-observable in principle? Would they be illegitimate objects too? Moreover, there are degrees of objectionability to mental objects. Desires seem more objectionable than drives or mental representations since the latter can be imagined as being epistemically inaccessible to the possessor of those objects as well as to others, and thus less like the paradigmatic problematic mental objects. Does this difference in epistemic access justify their special exclusion from theory, especially a theory that will have eventually to treat of the nature of epistemic access?

In practice methodological behaviourists forbade the hypothesis of any internal states. This cannot be justified by general considerations of the nature of science.

Problem

Given that latter ban there are difficulties in formulating useful explanations of psychological phenomena. The laws that can be constructed are no more than input-output correlations (S-R laws). These are surely insufficient for explaining the range of behaviours that can be the result of any stimulus. The solution was to make the history of the organism part of the law. But this just shifts the problem a small bit. Why is history relevant to the present response? The obvious reason is that history affects the current internal state – and it is this state that causally underlies responses.

Logical Behaviourism

Logical Positivism

Also early in the 20^th C in Europe there was a reaction to the wilder flights of metaphysics (a passage from almost any work by Heidegger/Husserl/Fichte/Hegel). Much of what was written appeared to be meaningless. The positivists looked for a criterion by which meaningful statements could be distinguished from others. They found it in the verificationist theory of meaning – the meaning of a statement is just the conditions whose verification would make that statement true. (Unfortunately, not all statements that we would want to call meaningful are included in this definition – in particular the statement of the definition itself fails to meet the criterion that the definition proposes.)

On the verificationist view, if statements about mental objects were to be meaningful they would have to determine conditions that could be verified. The only relevant conditions that can be verified seem to be behavioural and physical conditions. Therefore meaningful statements about mental objects have to determine behavioural and physical conditions that could be verified. The statement of Hempel (Kim p 29) reflects a view of this sort.

The Category Mistake

The classic statement of the behaviourist position was made by Ryle – though he himself denied it was a behaviourist position. His thesis was in reaction to the popular dualist position but he took Descartes as the archetypal dualist. The dualist sees a human existing in two realms; a public, outer, physical manifestation, and a private, inner, mental manifestation. The realms are different but they belong to what Ryle calls the same ‘category.’

Ryle calls this so-called ‘Official Doctrine’ of the mind the dogma of the ‘Ghost in the Machine’ and regards it as a logical error. The form of the error is that of a ‘category mistake’ in the use of mental terms; which is to say that the logical operations performed upon them are inappropriate to their natures. Only terms which are in the same logical category are able to have logical operations performed upon them, such as conjunction and disjunction. It is legitimate to speak of left hÿÿd and right hand gloÿÿs in a way in whÿÿÿÿit is not legitimate to speak of a left hand glove and a right hand glove and a pair of gloves. But this is just how the Official Doctrine speaks of bodies and minds. Another example Ryle uses imagines a foreigner watching a game of cricket. Cricket is famous for being associated with ‘fair play’. The observer would be able to isolate the functions of the batter, bowler, wicket-keeper, and umpire, but he might be puzzled as to which was the  member in whose charge was the provision of ‘fair play’. He would have to be told that fair play was not a function of the same type as batting, bowling, etc., that the same things could not be said of it as of those because they belonged to different logical categories. ‘Fair play’ was rather a term to be applied to the performance of the functions which constituted the observable features of the game as a judgement upon them. It is not hat ‘fair play’ does not refer to anything sensible and it is certainly not that such talk has no purpose, it is just that its purpose and use are not explained by such an hypothesis.

We are in an analogous position wrt our mental operations. Where, we ask, is the faculty to which intelligence is due? In fact, according to Ryle the intelligence epithets apply not to distinct causal entities but are descriptive of the performance of the functions which constitute the facts of our physical life. Again, it is not that they do not have a legitimate use but only that that use is not best explained in terms of the dualist theory.

When we describe people as exercising qualities of mind we are not referring to covert and inaccesible causes by which the behaviour these people comes into being, we are referring to the behaviour itself. Intelligent behaviour, for example, is behaiour which is characterized by smart-epithets whose applicatio is justified by reference to sets of testable explanatory/predictive hypotheses.

One of the consequences of the Official Doctrine which is based upon this logical error is that the ‘workings’ of other minds are closed books to us. There are no means under the Official Doctrine whereby we can be familiar with the system of causes which result in the overt behaviour which is all we can ever know of another person. On the other hand we are able to have infallible and complete knowledge of our own causes by the self-intimating nature of conscious mental events and our introspective capabilities. The recognition of the category mistake and the consequent rejection of the Official Doctrine of the mind allow us to be qualified to judge the intelligence-epithet desert of any other person quite as confidently as we judge our own, not because we can have access to their inner working but because that inner world is not there to be accessed by them either. In this way a proneness of the Official Doctrine to solipsism is avoided.

Dispositions

Suppose that the behaviourist claims that all talk of pain is no more than a way of speaking about pain behaviour. To say that Alice is in pain is to say that Alice is behaving in one of a number of ways that we have identified as being pain-behaviour. She is saying ‘ouch’, she is wincing, or she is jumping up and down on the spot, or she is gritting her teeth, or she is crying, or … Similarly, to say that Alice thinks it’s going to rain is to say that she is behaving in certain ways that will count as evidence of her ‘having that belief’. She is carrying an umbrella, hurrying home, forbidding the children to go outside to play, …

This is not a convincing way of talking about mental events or mental objects for two obvious reasons:

1.                   We think it is possible for Alice to be in pain and yet exhibit no pain behaviours. She may be a Stoic, or there may be reasons why she must not cry out in pain. Note that it is irrelevant whether her heart rate or blood pressure increases or there are other physiological signs of distress: these are not behaviours.

2.                   We think it is possible for Alice not to be in pain and yet exhibit pain behaviours. She may be just pretending.

Instead a behaviourist will talk about dispositions to behave. Thus Alice’s being in pain is her being disposed to exhibit pain behaviours, disposed to say ‘ouch’ and so on; and for Alice to be thinking that it is going to rain is for her to be disposed to act in the ways that will count as evidence of her having that belief.

Problem

Ryle compares this (Ryle p. 43) with the description of glass as brittle. ‘Brittle’ he takes to be a dispositional term that simply means disposed to shatter when dropped. A glass can be brittle even if it is not actually in the process of being broken, even if it never breaks, even if it is dropped and does not break on that occasion. But to explain the breaking of the glass in terms of its disposition-to-break seems hardly to be an explanation of its breaking at all. The definition of a disposition in terms of ‘if … then ---’ leaves the origin of this conditional quite unexamined – as a type of law statement it is what came to be sneeringly referred to as a ‘nomological dangler’. We will quite naturally ask what is it that explains the fact that glass has this disposition. The explanation we come up with is that the glass has an internal structure is easily disrupted by impacts. The facts about the internal structure allow us to predict that the glass will be brittle rather than soluble. Similarly, we are unlikely to be satisfied with an explanation of pain that stops at claiming that it is just having a disposition-to-pain-behave; we will want an explanation of how we come to have such ‘dispositions’. Of course, the most obvioius explanation is that we have internal states that constitute the causes of those behaviours.

Ontological Behaviourism

For both MB and logical (analytic, philosophical) behaviourism talk of mental phenomena is just a way of talking about behavioural phenomena. This leaves room for several attitudes towards mental objects themselves. Two are most prominent.

1.                   Behaviourist ways of speaking reflect the fact that there are no such things as mental objects.

2.                   Behaviourist ways of speaking are justified whether or not there are really mental objects.

MB would seem to be naturally associated with the latter atitude.

Behaviourism justified on logical positivist grounds would need to take the former attitude, unless, of course, one thinks that ‘meaningless’ posits are acceptable in science.

Ryle’s view would seem to lend itself to the latter attitude as well, but, in fact, his claims are such that we must assume that he takes the former.

The position that mental objects just are categories of behaviours we may call ontological behaviourism.  Note that this is quite independent of LB:

a.                   LB does not imply OB: the logical behaviourist can accept that there are inner causes, but that they can be eliminated from the discussion.

b.                   OB does not imply LB: just because an X is a Y does not mean that “X” and “Y” are identical in meaning.

Problem

The fact is that pain hurts, and that we think that we are thinking about something when we think about something. It’s possible to conceive of an imitation person – a zombie, perhaps – which behaves or is disposed to behave in all the same ways that we do or are and yet has no inner life. We conclude that there is something missing in the zombie and that behaviourism is unable to fully explain the world with which we are acquainted.

Moreover, we can conceive of two creatures that are absolutely identical in behaviouristic terms and yet are mentally quite distinct. (At least we think that we can so conceive. The classic example here is the case of ‘reversed spectrum’

Fundamental Problems with Behaviourism

Circularity – 1

All forms of behaviourism rely upon the assumption that behaviours can be defined (and recognised) independently of all notions of mental events. That is if talk of thoughts, desires, beliefs, pains, joys, etc. is to be rephrasable in principle in terms of behaviours then these behaviours have to be definable without reference to thoughts, desires, beliefs, pains, joys, etc. This turns out not to be possible.

Consider the example we gave of Alice’s belief that it is going to rain. We said that one of the behaviours that will count as evidence of her ‘having that belief’ would be for her to carry an umbrella. But this relies upon an assumption that she desires to stay dry if it rains, and a belief that an umbrella will allow her to stay dry, and that she is capable of reasoning in that fashion to come to conclusions that are actions (performing a practical syllogism). In that case we have not eliminated belief or desire attributions by our behavioural term substitution. We may try to remove those remaining beliefs and desires by saying, for example, that to say that Alice desires to stay dry is to say that she is behaving in certain ways that will count as evidence of her ‘having that desire’. She will be disposed to avoid walking in the rain, to take umbrellas when she thinks it’s going to rain, not to jump in puddles, etc… But it’s obvious that each of these behaviours is going to require further clarification in terms of further beliefs and desires. This process will never end.

This problem is not confined to human behaviour. The same argument will apply to animals, as long as we are prepared to attribute beliefs and desires (and practical reason – but that goes without saying) to those animals. When a fox appears the hen is said to be making diversionary displays, but they can only be called diversionary on the assumption that the hen recognises the fox and has certain beliefs and desires about the fox.

Circularity – 2

Behaviours are supposed to be mere ‘motions and noises’ (Ryle). So a specification of a behaviour must refer only to the purely physical aspects of the changes in the organism. When Alice reaches for the umbrella the very same physical motions can be part of behaviour that we describe as reaching for a weapon, feeling for a light switch, steadying herself, etc. depending upon whether the desire that is supposed to be being acted upon is to stay dry, defend herself, see where she is, stay upright, etc. The action becomes a part of some particular behaviour only with reference to a particular psychological state.

Indefiniteness

Moreover, any particular ‘behaviour’ will include a vast array of particular physical actions. For Alice to reach for her umbrella (and this is really only a part of a behaviour) may involve using her right hand, or her left hand, or grabbing with two fingers or three, or etc. When it comes to more general behaviours, such as (in Kim’s example) greeting, there is a commensurately greater range of possible physical configurations that can be included in the category of that behaviour. In fact, it seems clear that an infinite number of physical configurations can be included in any category, and that it is quite impossible to define a category that is sufficient to substitute for mental terms without appealing to intentional terms – and that recourse is illegitimate. 

Chauvinism

Any behaviour that is proposed as a substitution for a mental term requires a particular physical form. If we declare that pain, for example, is equivalent to a disposition to squeak, then any creature that is incapable of squeaking can not have a disposition to squeak and can therefore not be in pain. It seems that we can imagine a number of creatures who have such a physical limitation, such as an octopus, and yet that we would probably believe could feel pain. Moreover, no matter how much we expand the behavioural definition of pain it will always be possible to conceive of creatures that lack the physical capabilities required to be disposed to show that behaviour and that we would nevertheless believe could feel pain..