Awesome demo

From Greg Williams (920103)

Gary Cziko 930103.0240 GMT] (I got "93" on the first try!)

This is very strange indeed since the subject's responses are SIMILAR on
the two trials and yet what she saw (the cursor pencil point) during the
two trials was very DIFFERENT. How can the subject respond similarly on
two trials when what was seen (the "stimulus") was so different?

The two "responses" to two "very DIFFERENT" "stimuli" are NOT the SAME, only
SIMILAR. Skinnerians would have no problem with the fact that (even slightly)
DIFFERENT "responses" resulted from different "stimuli." And if you show them
results where successive "responses" are IDENTICAL, yet the "stimuli" in each
case are different, they will talk about "stimulus generalization" or say that
the organism can "lump" different-appearing stimuli (to the experimenter) into
ONE kind of "discriminative stimulus." But it gets even worse. If successive
"responses" are judged as different by the experimenter, they might say that
they really are all in the same "operant" set of responses.

If anyone can come up with an explanation of this which does not look like
closed-loop negative feedback model, please let us here on the CSGnet know
about it.

At the level of the observed phenomena, it is obvious that the "stimulus" in
your experiment is affected by the "response." Skinnerians have NO problems
with such situations, which they call instances of "self-stimulation." But at
the generative-model level (should any of them dare to speak thereof, lest
they lose their "Skinnerian" labels!), some of them might argue that the
"discriminative stimulus" is "middle pencil point moving (either direction)
away from the middle line" and that the (ongoing) "operant" (set of
"responses") consists of "actions to move the middle pencil back toward the
middle line." Such a generative model makes no explicit reference (no pun
intended) to postulated internal (to the organism) states. (PCTers, of
course, will immediately note the IMPLICIT reference level. Skinner used to
argue that bringing in such hypotheticals would add nothing to the "analysis
of behavior," and, worse, would tend to distract one from the data. I still
think he had a point, TO A DEGREE. Yet, by hewing that line so cautiously, he
was unable to explain the existence of particular "wants" -- to which I
quickly add that he did not WANT to explain that). Also, the typical
Skinnerian would want to claim that the person would "respond" to the
"discriminative stimuli" because of PREVIOUS reinforcements having to do with
his/her relationship to the experimenter. Regardless, Skinner's notion of
"operant" SETS of OUTPUTS, each of which result in the same OUTCOME, was a
significant step toward replacing specification of outputs with control of

As ever,