[From Bruce Abbott (950315.1545 EST)]
From Bill Powers (950315.0739 MST) --
From my point of view, the wrongness of the principle is only a symptom
of the basic problem. To me, the basic problem is the general approach
to learning and behavior that allows making causal statements without
reference to the internal organization of the behaving system.
This criticism is specific to EAB, and I agree. (Shhhh! They'll take my
union card away if this gets out!)
Do you really, literally, believe that there is something about a pellet
of food that can cause a change in the relationship of behavioral
actions to objects and events in the local environment? When you call a
piece of kibble a "reinforcer", that's what you're literally claiming.
From what I know of you from your writing and programming, I simply
can't accept that you believe literally in reinforcement or in
controlling stimuli or in any of that stuff.
Well, you're right, I don't, except for the part about "in any of that
stuff," because that covers a lot of territory.
And if you don't believe
literally what you're saying, why go on using that language?
First of all, it is a vocabulary with which I am familiar and comfortable.
I see no good reason for abandoning perfectly good descriptors of behavioral
phenomena (e.g., "stimulus control") and then having to convey what I'm
talking about in lengthly, awkward sentences. (Technical terms are invented
to eliminate that problem.) Third, I want to be able to "translate" our
models back into behaviorist terms in order to demonstrate to behaviorists
how PCT accounts for the phenomena to which they have applied these names.
I think it important to distinguish the descriptive from the explanatory
application of these terms. For example, "reinforcement" can refer to an
empirical observation (e.g., when I follow the rat's lever-press with
immediate access to food, and the rat has been deprived of food for awhile,
the rate of lever-pressing increases over time), or to a theoretical
explanation (e.g., following a response with a reinforcer acts to increase
the future probability of the response). I can certainly talk about
reinforcement, or extinction, or stimulus control as empirical phenomena
without necessarily subscribing to the theoretical view within which the
terms arose. I recall having recently read a provocative analysis of
classical conditioning in which Pavlovian terms like "conditioned stimulus"
were used. Sounds like something I would do--I wonder who wrote it? (;->
People
do say such things in jest, but they would be insulted to hear
themselves described as if they took their own words literally.
You mean words like "stimulus control" taken to mean literally that the
stimulus has the purpose of trying to control someone? Who would be silly
enough to do that? (;->
Regards,
Bruce