[From Rick Marken (970314.1800)]
OK. I've stuck my neck out once again and put PCT to the test right
out there on the net where everyone can see it. This time it's
a new demo called "The Behavioral Illusion". It's so simple, it's
embarassing; it's so profound, it's overwhelming;-)
Those who have a Java enabled browser can see the demo by pointing
their browser to:
http://www.leonardo.net/Marken/demos.html
and selecting the Behavioral Illusion demo.
I put the write-up of this demo in a new format. I am planning to re-do
all the demos in this format. But I would appreciate comments and/or
suggestions regarding the new demo itself and/or the new write up
format.
Thanks
Rick
PS. David Goldstein: I have not forgottten you. I will try to get
your paper up on the net next weekend.
from Tracy Harms (970314.1250 PST)
Gack, I didn't unsubscribe in time to avoid the temptation to reply to this!
Bill Powers (970314.0843 MST):
I think we have to be careful about accepting PCT as a meta-model, which
means only that we can accept it once and for all without further
investigation.
Not at all. No idea need be accepted with such categorical finality, and
none should. That you think such final acceptance *must* occur re.
meta-models indicates one place where you haven't done your homework. I am
sad when you can't follow my argumentation, but things like this confirm
your own suspicion that you are unfamiliar with a body of knowledge which I
rely on as I state my case. One important point therein is that
rationality succeeds without ultimate commitment. Another is that
meta-scientific ideas, although they lack the empirical testability which
you admirably call for in testing PCT, are nevertheless subject to
refinement through critical pressures involving logic and values.
[...] People like to take new ideas and go
rushing off into the distance with them, hoping to make some lucky discovery
so they can take credit for getting there first (on what grounds, I do not
know). But real progress in science relies on people who simply take the
next possible step, placing the bricks with care and making sure the joints
are sound. Boring, to some. But nothing else works.
Every step which succeeds is, tautologically, an immediate extension from
existing success. But this idea that scientists "simply take the next
possible step" erroneously presumes that such steps are apparent in the
present state of science. They are not. Insofar as science advances it is
always, *always* done by leaps into the dark. That's what is meant in
saying *bind* variation.
Tracy Bruce Harms
harms@hackvan.com
[From Bill Powers (970314.1651 MST)]
Tracy Harms (970314.1250 PST) --
Gack, I didn't unsubscribe in time to avoid the temptation to reply to >this!
Better get off the net if you're going, Tracy. There's always one more
peanut to eat.
Every step which succeeds is, tautologically, an immediate extension from
existing success. But this idea that scientists "simply take the next
possible step" erroneously presumes that such steps are apparent in the
present state of science. They are not. Insofar as science advances it
is always, *always* done by leaps into the dark. That's what is meant in
saying *blind* variation.
Sure, but you don't reorganize until all of your existing control systems
have failed to correct the error. Science is not JUST a series of leaps into
the dark. After every leap you have to explore all the implications of the
new idea before there is any reason to make a new leap, or any profit in
doing so, or any chance of landing on solid ground.
This is why science is naturally conservative and resistant to new ideas.
Before behavioral scientists can accept PCT, they must accept that their
current ways of understanding behavior have failed. But most of them say,
"No, wait, wait, there's one more variation I haven't tried; there's a
promising new line of attack that has to be explored; just get me another
grant and I'll prove that my old idea works perfectly well," and so on and
so on. They haven't reached the point of total despair, so they're not ready
for PCT. The new Phoenix can't rise until the old one has been reduced to ashes.
Best,
Bill P.
[From Bill Powers (970314.1954 MST)]
Rick Marken (970314.1800)--
OK. I've stuck my neck out once again and put PCT to the test right
out there on the net where everyone can see it. This time it's
a new demo called "The Behavioral Illusion". It's so simple, it's
embarassing; it's so profound, it's overwhelming;-)
Tried it and it's beautiful. Actually I ran it (as usual) without reading
the instructions, so when I was controlling the line's angle I decided to
keep it horizontal. That produced a ZERO angle of the line, since you were
measuring the angle from the horizontal. Of course if you happened to
measure the angle of the line relative to the vertical side of the
rectangle, the angle would have been 90 degrees, but the controlled
perception would have been the same.
When I did Expt. 1 this way, the "angle" reported is 0 degrees -- your
writeup implies that "angle" refers to the angle of the plotted points,
which is certainly not 0 degrees. It's hard to read the "shape" number
because the first digit overwrites the = sign. And how come eight
determinations but only 4 points on the plot? HAH! La cucaracha.
Maybe, to avoid the confusion about what "angle" you're talking about, you
could just say "DV = 1.0*IV" or "DV = 0.5*IV" or whatever the numbers are,
when you're talking about the graph.
A neat demo. More, more!
Best,
Bill P.