PHIL2206 Philosophy of Mind
Week 4: Behaviourism

Connections between minds and behaviour

  1. There are connections between, on the one hand, the way a thing behaves and, on the other hand, whether or not it has a mind, and which mental phenomena (if any) it exhibits.
  2. One important connection is an evidential connection: there are various ways of behaving b such that if a thing behaves in way b then that is evidence that it has a mind, perhaps even that it exhibits a certain mental phenomenon (e.g. that it believes a certain thing). Some examples:

An epistemological problem

  1. Here is a vew: Although I can get direct (i.e. non-inferential) knowledge that I have a mind and that I am exhibiting certain mental phenomena (e.g. that I feel cold), I cannot get direct knowledge that you have a mind or that you are exhibiting certain mental phenomena. Rather, if I can get any knowledge about your mind I can only get it indirectly, by making inferences from your behaviour. Ultimately, from your non-mental and publicly observable behaviour (movement, perspiration, salivation, pulse, blood pressure, and so on).

    (The inferences need not be made consciously, and might be made extremely quickly.)

  2. A concern: This means that our knowledge of other people's minds (i.e. of whether or not they have a mind, and of which, if any, mental phenomena they are exhibiting) is less than certain, prone to error.
  3. An even bigger concern: It means that knowledge of other people's minds is impossible. Why? Because we have no independent way of checking the conclusions of these inferences (we cannot, for example, directly observe that something has a mind, or that it believes a certain thing).

    An illustrative analogy: Wittgenstein's boxes. We all have a box with something inside. We can look inside our own box, but not in anyone else's. Instead, we can give each other signals about what is in our box. Can we ever know what is in each other's box? Problem: how do I know what your signals mean? You cannot tell me, because that would be just giving more signals.

  4. We don't have this problem if we can after all get direct knowledge of other things's minds. Logical behaviourism is a way of allowing that.

Logical (analytical) behaviourism

  1. (This was once a popular position but is not so anymore.)
  2. The basic idea: Every expression about minds and mental phenomena is synonymous with (i.e. has the same meaning as) an expression about non-mental publicly observable behaviour.
  3. For example, 'is in pain' means 'winces and groans' (this is too simple to be the right translation, but it gives the basic idea). So 'This thing is in pain' means 'This thing winces and groans', and when I say 'This thing is in pain' I am saying that it winces and groans - I am talking about the way it behaves.
  4. If 'is in pain' means 'winces and groans', then to be in pain is to wince and groan - they are the very same thing. If so, then since we can get direct knowledge of something's non-mental and publicly observable behaviour, we can get direct knowledge of its mind.
  5. Here is an argument for logical behaviourism that appeals to logical positivism (a view that was around in the early 20th century but is not popular these days):
  6. Objection: You claim that when I say 'This thing is in pain' I am talking about behaviour, but on the surface I don't seem to be.

    Response: Surface appearances can be misleading. When I say 'The average Australian family has 2.2 children' I seem to be talking about a thing called 'the average Australian family', but arguably I am not (plausibly, ‘The average Australian family has 2.2 children’ means ‘On average, Australian families have 2.2 children’).

Problems for logical behaviourism

  1. Here is a reason to think that 'is in pain' does not mean 'winces and groans': it is possible for a thing to be in pain without wincing and groaning (think of snakes), and for a thing to wince and groan without being in pain (think of fakers). Either of these by itself is a problem for the view.
  2. Response: Ok, so this is not the right translation. But there is some non-mental publicly observable way of behaving, b, such that 'is in pain' means 'behaves in way b', or perhaps 'is disposed to behave in way b'.
  3. But is this right? It seems that for any non-mental publicly observable way of behaving, b, it is possible for a thing to be in pain but not behave in way b, or even be disposed to behave in way b, and for a thing to behave in way b, or be disposed to behave in way b, but not be in pain. The onus is on the logical behaviourist is to come up with such a b.
  4. Maybe the task is easier if we relativize to species: for any pain-capable species, there is a certain way of behaving, b, such that, 'is of that species and is in pain' means 'behaves in way b' (or 'is disposed to behave in way b'). But this still leaves us with the unrelativized uses of 'is in pain'.
  5. It seems even harder to find translations for expressions about belief. How should we translate ‘believes that there are no kangaroos in Rome’?
  6. Remember: we cannot use expressions about linguistic behaviour. But even if we could it might not help:

Some possible retreats

  1. If logical behaviourism is false then maybe we can retreat to a weaker claim.
  2. Here are some claims, in order of decreasing logical strength:
    1. 'is in pain' means 'winces and groans'
    2. To be in pain is to wince and groan
    3. Necessarily: for all x: x is in pain iff x winces and groans
    4. Actually: for all x: x is in pain iff x winces and groans

    1 entails 2, but not vice-versa. 2 entails 3, but not vice-versa. 3 entails 4, but not vice-versa. The first claim is logical behaviourism. The second claim is sometimes called ontological behaviourism.

  3. Any of the last three weaker claims would help solve the epistemological problem we saw above (why?). Should we retreat to one of them?
  4. It seems not. The problems that we saw for logical behaviourism are also problems for the second and third claims. What about the fourth? This is not something we can verify or falsify from the armchair. But it seems unlikely that it is true. It is, in any case, not an interesting claim.
  5. Here is something else that we might say: 'pain' refers to whatever it is that is the actual cause of wincing and groaning in humans. This is not to claim that 'pain' means 'the cause of winces and groans in humans'. We fix the referet of 'pain' by appealing to a contingent fact. Like fixing the meaning of 'red' by referring to the colour of this tomato. Could this tomato have not been red?
  6. If this is right, then pain is the actual cause of wincing and groaning in humans.
  7. This might explain why there seems to be some connection between the meaning of 'pain' and behaviour - why it is a condition on knowing what 'pain' means, that one understands there is a link between being in pain and behaving in a certain way.

Scientific behaviourism

  1. Concern: appeal to introspective data is not scientific. Should only appeal to facts that are publicly observable if we want to study mental phenomena scientifically.
  2. The only things that our theories should try to explain and predict are publicly observable non-mental behavioural data. This is the only arbiter between competing theories. This is the only admissible evidence.
  3. Then we must decide whether our theories can posit other things - e.g. non-publicly observable phenomena. Or should we treat things as black boxes? The latter seems a bit harsh: does physics do this? Can we make sense of the past behaviour of something effecting its future behaviour without positing internal states?