These "control reaction times" could be a useful diagnostic
for validating such relationships in the hierarchy.
You bet yer boots they could!
And I'll bet that was one point of the paper you cited. I remember reading
it back then in 1991 or so, when you made it available, but I don't
remember that point, so your memory is doing better than mine.
Shouldn't we verify what the typical times are and use them
as a tool of research?
Absolutely. What's holding "us" back?
Touche.
Thank you for not commenting on this gaffe:
Depends on whether sequence is above or below program level.
I was thinking of sequence relative to categorization. Duh!
It might be hard to get "reaction time" data for categorization. It's not
easy, at least for me, to imagine control of category in the same way that
I control sequence, configuration, event, etc. It does not seem that I have
a reference and produce outputs to "make it so." It seems rather that I
recognize something as an instance of a symbol that I can control in a
sequence or program. If I am mistaken, no action of mine can make it so, I
am just mistaken. One of the reasons Martin's suggestions (991022 10:19)
seem very sensible to me. I'll try to reply to that separately, though time
is tight.
It might be hard to get "reaction time" data for categorization. It's not
easy, at least for me, to imagine control of category in the same way that
I control sequence, configuration, event, etc.
Category control is quite easy to implement. What you need is an
experimental setup in which stimuli change so as to exemplify different
categories, with the subject having a means of affecting the stimuli (which
are also subject to disturbance). If the stimuli change so that "round"
changes to "square", it will take a little time for the subject to
recognize that there has been a change, and take the action that will
return the category to the goal-state (this is quite aside from other
properties of category perception such as hysteresis).
To separate this level from lower levels, it would be necessary to have
several different sets of stimuli that would exemplify each category. This
would allow causing changes which altered, say, the configuration being
seen without altering the category to which it belongs -- picture of duck
to picture of chicken rather than picture of duck to picture of alligator.
The subject would not be required to correct changes that changed only the
configuration without changing the category. This is probably a good
general experimental principle for investigating different levels of
perception, if we can learn to apply it. I'm sure it's been done already.
It does not seem that I have
a reference and produce outputs to "make it so." It seems rather that I
recognize something as an instance of a symbol that I can control in a
sequence or program.
I think this is largely because people like us think so incessantly in
terms of categories that we don't even realize when we're doing it. Have
you ever sorted through a junk-drawer look for "a button"? This is not like
looking for a specific configuration, because buttons come in all sizes and
shapes, and colors, and hole configurations, and textures. You're looking
for "a" button, not "the" button -- that is, any object of the category
"button." Any time you're looking for _any thing of a specified class_
you're looking for a category, not a configuration. The same generally
applies whenever you use an indefinite article before the name of a noun or
verb (a chopping motion). "Buy me a puppy" the little boy says. He's not
referring to an object, but to a category of objects. Buy me anything that
I can perceive as belonging to the class "puppy".