Study Guide

Week Eight: Behaviourism and "Scientific Method"

Materials Assigned for the Week
The Central Point of This Week's Material
Other Concepts and Points you are Expected To Master This Week
Miscellaneous Comments and Clarifications
1. Trial-run at a Dispositional Analysis of a Mental Expression, "X wants Y"
2. Testing the Existence of God and Testing the Existence of Minds

I. Materials Assigned for the Week

Reading:

Do your reading in the following order:

Lecture 19: Logical Behaviourism
Readings: Russell, Other minds are known by analogy from one's own case, pp. 345-348
Lecture 9: Is the Existence of God Testable?
(Read this as a lecture on scientific methodology, not as a lecture on theology)
Lecture 20: Methodological Behaviourism

Exercises:

Review Questions: pp. 283-284, esp. #3, 96, 291
Problems for Further Thought: pp. 284, 291-292 esp. #1

II. The Central Point of This Week's Material

According to Behaviourists the problem with Dualism is not that arguments for non-physical minds aren't very good arguments; it's that we don't really need to drag in anything but physical behaviour in order to explain everything we wanted to explain in the first place. Mental words do not pick out special non-bodily causes of behaviour. They pick out no more than certain patterns of that outward behaviour itself. Indeed, our mental vocabulary (e.g. "X wants to drink") is properly analysed purely in terms of bodily behaviour ("if the circumstances are so and so, X does drink"). Exclusive concentration on outwardly observable behaviour provides a sounder scientific method for the psychological sciences as well.

III. Other Concepts And Arguments You Are Expected To Master This Week

The difference between mentalism and dualism- the difference is a subtle one (pg. 275) so be careful: all dualists are mentalists (dualists think that mental words refer to minds which are distinct things from bodies and that these minds are the inner causes of bodily behaviour); however, not all mentalists are dualists (e.g. identity theorists think that mental words refer to perfectly ordinary physical brains and that these physical brains are the inner causes of behaviour)
Mentalism is the idea that mental words one way or another pick out the inner causes of behaviour, whatever those inner causes may turn out to be, Cartesian or brains
Dualism is the idea that the inner causes of bodily behaviour are a completely different kind of thing than physical bodies, something immune to the Evil Genius, non-spatial and the like
The difference between logical behaviourism and methodological behaviourism- basically, behaviourism treated as providing the real meaning of mental terms versus behaviourism treated as providing a method for doing proper science
How to spell the word "behaviourism" - "behaviourism" without a "u" is the American spelling, "behaviourism" with a "u" is Commonwealth; there are lots of such pairs: "labor"/"labour", "color"/"colour", "flavor"/"flavour", "endeavor"/"endeavour", "savor"/"savour", "savior"/"saviour" and so on; either spelling is fine in 34.101 but probably not outside
What the problem of other minds is
The difference between an argument from analogy for other minds (such as Russell's pp. 345-348) and an abductive argument for other minds (such as Sober's pp. 277-278)
The difference between introspection and inspection- again
The difference between an empirical observation and a theoretical postulate- again
What a category-mistake is- be armed with several examples outside of the philosophy of mind
The difference between a cause and a disposition- in particular, the difference between treating some mental thingy-me-bob as an inner cause of behaviour versus treating it as a disposition to behave
The difference an experimental result and a research programme
The difference between a proposition being verifiable and its being falsifiable- and why being verifiable in principle is less important than being falsifiable in principle, again
What the Testability Theory of Meaning is (Lecture 9)- and its parallels with methodological behaviourism's recommendation for scientific method (Lecture 20)
What an auxiliary assumption is- though it was not assigned, you might want to compare this to the notion of a background assumption (Lecture 17, pp. 199-200)
What Chomsky's "novel behaviour" objection to methodological behaviourism is (pp. 287-289)
What Armstrong's "minimal explanation" objection to methodological behaviourism is (Box pg. 289 referring back to pg. 282)
What Sober's "environmental determinism" objection to methodological behaviourism is (pp. 290-291)

IV. Miscellaneous Comments and Clarifications

IV.A Trial-run at a Dispositional Analysis of a Mental Expression, "X Wants Y"

As always in philosophy, the best test of any pudding is in the eating. A behaviourist believes that we can happily forego Descartes' desperate search for non-physical minds because all we are really talking about when we use so-called "mentalistic" terms is purely outward behaviour anyway.

It is true that this claim trips first off of the lips of a logical behaviourist - as a claim about the real meaning of our everyday mentalistic vocabulary. But a methodological behaviourist will eventually agree too. Once we have our studies of minds firmly based on proper scientific method - in which all reference to non-physical inner causes is simply left out as methodologically ill-advised - then all our scientifically proper explanations will be in terms of purely physical bodily behaviour too. This exclusive concentration on getting everything down to outward bodily behaviour, as opposed to using the resources of definitely inner albeit physical - i.e. brain - causes, is what marks off all varieties of behaviourist theories from the theory of mind-brain identity and its spin-offs. (To come in Weeks Nine and Ten.)

So how does the behaviourism project actually fare? Can it really pull off a behaviourist reduction of what seems to be a Cartesian affair in terms of purely public, outwardly observable, physical behaviour? For instance, how well does it do with the fairly simple proposition "Tom wants a drink of water"?

Now, we know perfectly well what a dualist such as Descartes will say about such a case: The word "want" in this sentence refers to a state or process in Tom's non-physical mind. Tom alone knows whether and when he is in that state - even an Evil Genius can't take that away from him, anymore than it can take anything else mental away from him. When Tom finally does take a sip of water, that private inner mental state or process will be the cause of such outward behaviour. The bodily drinking is one thing - it's the effect. The mental wanting is quite another - it's the cause. No proper account can be given of my behaviour without both distinct components of the story being laid fully before view.

We also know perfectly well what a behaviourist wants (!) to say: The word "want" doesn't refer to any inner non-physical state in Tom's mind and can't pretend to in any properly scientific account. For a start, there just is no such state, there is only Tom's outer physical behaviour. Moreover, reference to inner causes is always scientifically suspect anyway; it amounts to little more than sticking in a miraculous "black box" whenever the going gets rough. Can the behaviourist have what s/he wants? Taste the pudding.

Try for a start the crude analysis:

"S wants to drink water" =def "S drinks water"

This won't do at all. People who want to drink water aren't always drinking water. I may want to drink water over the eight hour day, but never get the chance, and so on. Better, maybe:

"S wants to drink water" =def "S is disposed to drink water"

One problem with this is simple, though maybe it can be solved with more work (with lots more work). We have replaced the analysis of "wants" in terms of actual behaviour with an analysis of "wants" in terms of dispositions to behave. Specifically, we have replaced actual drinking with dispositions to drink But what is a disposition? And who is to say it isn't just a flash word for an inner cause we know not what, or about which we don't yet know? (Compare Sober pp. 282, 289.) One way to analyse dispositions so that at least they don't look to invoke covert inner causes is to cash them out as conditionals (see the Box pg. 15). Conditionals are "if ... then..." sentences; if we take care to ensure that we fill in the placeholders "X" and "Y" only with obviously public physical bodily behaviour and circumstances, maybe that will do. Thus:

"S wants to drink water" =def "If the circumstances were so and so, then S would drink water"

Out of the frying pan, into the fire! Maybe we haven't sneaked in anything inner, covert, non-physical here. But now how are we to separate off which circumstances are supposed to be the relevant ones? There seem to be an obvious infinity of them. For a start, I suppose we rule out as likely candidates for the "if" clause my cat sitting on the computer monitor, and the TV blaring in the next room, and the dog just running away for the umpteenth time: it would be plain mad to cash out "Tom wants to drink water" as "If the cat is on the monitor, then Tom would drink water". We want, crudely, "water-related" circumstances only. More accurately - but harder for a behaviourist to say, we want "wanting-water-related" circumstances only. Well, which ones are they? Without begging the question and going in a circle, mind.

Maybe this will start. The circumstances in which "wanting" to drink water get cashed out purely into "if ... then ..." terms had better be ones where I at least believe that the liquid before me is water. (Compare Sober, 281.) I want to drink some water. You put a cup of clear liquid before me which I believe to be acid. Those are not amongst the circumstances in which I would quaff the cup. You put a cup of clear liquid before me which I believe is good clean water. Those will be amongst the circumstances in which I would take a sip. So try:

"S wants to drink water" =def "If the circumstances are such that S believes X is water, then S would drink X"

Unfortunately, this merely puts off the Day of Reckoning. We may now possess - or be close to - a behaviourist analysis of "wants" purely in terms of outward behaviour. But now the behaviourist is forced to deliver on an analysis of "believes" purely in terms of outward behaviour. This little chore has proved surprisingly intractable. Just have a go.

"S believes X is water" =def "S is in a B-state regarding X"

But what's a "B-state" when it's at home? More pointedly, what analysis can be given of what a B-state is in purely behavioural terms? We have every reason to suspect that when the going gets rough, a B-state is going to be explainable only as a belief state, so what have we achieved? Try:

"S believes X is water" =def "S exhibits the outward behaviour which he exhibits when and only when X is water"

This won't do, because it entails that S believes X is water when and only when X really is water. But S, like everyone else, sometimes believes what is false (i.e. believes that X is water when in fact X isn't water). Back off a bit then:

"S believes X is water" =def "Under circumstances relevant to X being water, S would behave in a way appropriate to X being water"

This is not (quite) as empty as it looks- though certainly most of the work has been shoved under the rug of "circumstances relevant to" and "behaviour appropriate to". But there are easy counter-examples. Think about acting in a play, or learning your lines for a rehearsal. "Method acting" especially enjoins what's on the right side of the definition, or so I'm told. But even method actors don't have to believe the empty cup in their drinking scene contains water in order for them to act out their roles impeccably.

And so on. The details don't ever quite come together. Which means the analysis in purely behavioural terms never quite works after all. The best proof of any pudding there is.

IV.B Testing the Existence of God and Testing the Existence of Minds

Week Eight is the only occasion where I will ask you to read bits of Sober out of order. I've asked you to read Lecture 9 in-between Lectures 19 and 20. There is a reason.

A quick way to see it is to look at Sober's discussion of Skinner:

Another reason that Skinner gives for rejecting mentalistic theories is that they are "too easy". In the above example [about Jane wanting a drink of water and believing there is water in the cup before her], I attributed to Jane a combination of beliefs and desires that predict something about her behaviour. Suppose the individual doesn't reach for the water. What will I do? I'll modify my claim about what she thinks and wants. I won't abandon my assumption that Jane has beliefs and desires and that these cause her behaviour. So, apparently, the assumption of mentalism - that behaviour is caused by beliefs and desires of some sort - isn't tested by observing behaviour. Indeed, it appears that no matter what Jane does, I can always formulate a belief/desire story that is consistent with what I observe. (pg. 285, my embolding)

This is obviously not a virtue of a scientific theory, but a vice - that we will tell the same story (or the same sort of story) no matter what happens. If a theory doesn't rule something specific out, then it hardly rules anything specific in. That is, if an explanation doesn't go out on a climb sometime, about something, then it is compatible with just about anything in the world happening. So it hardly explains what does happen in the world when it happens.

We have worked a bit on this idea whenever abductive arguments have been to the fore. Now is the time to get this fundamental principle of scientific method cleared up once and for all. Sober's discussion of it, unexpectedly, is in Lecture 9. So back we go to Lecture 9. The context there is theological only in passing. The point is that theological explanations and proofs are always at least in danger of becoming vacuous or "too easy" in exactly the way Skinner is objecting to (e.g. pg. 91). Sober himself does not think that a careful proof of God or some theological explanation of some phenomenon will automatically suffer from this flaw of being consistent with anything at all that happens. But he does insist that it is a serious methodological danger always worth being clear about and wary of. And that's what Lecture 9 is all about. [In fact, Sober's own criticism of theological explanations is that too often they do not treat seriously enough the requirement that any auxiliary assumptions used must be independently established (pg. 94). Since every other kind of explanation also uses auxiliary assumptions, this too is a shared danger between theological and scientific explanations worth being clear about and wary of.]

Contents
Previous Week: Dualism and the Mind-Body Problem
Next Week: Mind-Brain Identity Theory