(urth) 5HC : Skinner, Turing

[Iorwerth Thomas]

>From: Jim Raylor <rjraylor at yahoo.co.uk>

>Pioneers of behaviourism such as Watson and Pavlov
>were attempting to put psychology onto a more
>scientific focus, by concentrating on observable
>behaviour and repeatable experiments.

Which was a worthy and probably necessary thing to do, given the excesses of 
their predecessors.  One does get the impression that the pendulum _did_ 
swing too far the other way, though.  I'm not entirely certain that the 
criteria for objectivity used in natural sciences are entirely appropriate 
for psychology, while that type of criterion was what the behaviourists 
tried to impose (not that _anything_ goes, it's just that the study of 
humans and animals and AIs may require different methodologies with 
different criteria from the study of say, quarks).  But you have to start 
somewhere, learn from your mistakes etc.

>Skinner's experiments were generally far more humane
>than those of the neurologists who sacrificed their
>subjects to measure changes of state in their brains.
>Possibly his greatest contribution to humanity is his
>demonstration that rewards are for more effective in
>conditioning behaviour than punishments.
>I am not saying he was a saint. For example, during
>the war he advocated the use of trained pigeons to
>guide missiles, perhaps though this is preferable to
>the use of Kamikaze pilots.

I'm not sure that I'd like to live in his ideal society though.

>He did create an artificially enhanced environment for
>his second daughter and spent considerable time
>observing and influencing her behaviour, which I see
>as nothing more sinister than the actions of a devoted
>parent. By all accounts she is a successful and well
>adjusted academic who continues to support her
>father's ideas. At university, the rumour was that she
>had gone mad after being treated as an experimental
>subject by her wicked father and was hospitalised!

Ah!  I thought that it might have been an exaggeration.  On the other hand, 
what I've read of Watson's 1930 child rearing handbooks probably contributed 
to this impression; by all accounts they were rather... peculiar, and 
probably didn't do kids raised by their principles much good.

>Skinner did not entirely deny the relevance of
>internal states (is the pigeon hungry?), but found
>objective work-arounds in most cases (Take a pigeon at
>95% of its previously established normal body weight).
>

I guess he'd have to, given the limits of language observed in 
behaviourism[1], but I suspect that there's an implicit connection between 
internal and external states that that manouver exploits.

>Turing's approach also focused on observable inputs
>and outputs, though he had a more mathematical and
>abstract notion of internal state than was formerly
>available.
>
Is that what's called fuctionalism?  (Just to get terminology clear).

I'll elaborate on something I said earlier.  What seems to have happened 
with behaviourism is analagous to what seems to have happened with the 
Copenhagen interpretation of Quantum Mechanics.  You have a fairly well 
defined _operational_ definition of what you're looking for (behaviour in 
psychology, particle spin - for example - in QM) and a good idea of the 
basic philosophy behind what you're doing and its limits.  Your students, 
however, aren't so careful.  Neither are the popularisers of your ideas (who 
may or may not be identical with you), largely because they're writing what 
amounts to a polemic.  At some point, the operational definition becomes an 
_ontological_ one, and then you get crass behaviorism (version [2]) and bad 
pop-sci nonsense on state-vector reduction.

I tend to automatically assume (knee-jerk prejudice) that most behaviorists 
are crass, as opposed to thoughtful (type [1]) because I'm a massive 
intellectual snob.  Oh well.

>For what it is worth, I do find the apparent
>references to these individuals in 5HC to be more than
>coincidence.

It hadn't occured to me, but there might be something in that.  I do tend to 
think that he was more of the cold, rational and amoral scientist archetype 
[1] rather than a depiction of anyone in particular though.

[1]  I won't say steriotype as there may have been one or two of this kind 
lurking around the more unpleasant reaches of 20th century research, eg. 
those motherhood experiments with monkeys, etc.     And then there's Edward 
Teller for us physicists.

Iorwerth