[From Bill Powers (2008.01.27.1103 MST)]
Jim Dundon 01.27.08.1246est
Two rules of PCT are
1 No just knowing, everything
has to be proven
2 No faithing, everything has
to be proven.
So. what exactly are
unobserved perceptions. Have they been proven to exist? If
so, how?
If they are sometimes
unobserved and sometimes observed how do we know they exist in the
unobserved state?
If they have never been
observed how do we know they exist at all?
OK, reasonable questions so I’ll try to answer them.
1,2: “Proving” an idea means using it to make a prediction
about something we will observe if the idea is correct, and will not
observe if it is wrong. I do this primarily by using computer simulations
because they allow making very exact predictions and detecting relatively
small departures of the simulation’s behavior from what the real person
does. There are other ways of doing this, but I’m best at using this
approach, and it is very sensitive to incorrect theorizing.
Unobserved perceptions: there are many examples of these. The way we
detect them is somewhat indirect. I can mention something like the
pressure of the seat you’re sitting in, and that perception (if not
already conscious) immediately becomes conscious unless you’re not
sitting down. If you’re not sitting down, calling attention to the
perception being described does not make you conscious of it. The mostly
likely conclusion is that the neural signals coming from the pressure
receptors were there all the time, but were not reaching your awareness
until you were encouraged to pay attention to them. If we had the
resources, we could do further tests by actually measuring the neural
impulses to verify that the sense-organs were working before you became
aware of the sensation. I haven’t done that, but the other evidence is
fairly reasonable.
Another approach is to demonstrate that a person is controlling some
environmental variable while not being aware of doing so (like keeping
your balance while walking, with all your attention on an approaching car
that might hit you). According to all we know about control processes,
control requires the presence of neural signals representing the variable
that is under control – that is, perceptual signals. We have several
ways of testing to see if a person is controlling some variable (like
disturbing it and looking for actions that oppose – or if there is no
control, fail to oppose – the disturbance). We still have to rely on the
person to tell us if the perception was consciously experienced, but at
least we can reasonably-well establish that the variable was under
control (if it was) when the person says there was no awareness of
perceiving it. We could go farther and show that if we make perception of
the variable impossible, the person can’t control it either consciously
or unconsciously. I haven’t done that in this context, though I’ve done
it in other contexts. Preventing perception of a disturbed variable
always prevents control of it.
A third approach, which I haven’t tried yet, is to measure the parameters
of control in some control task and see if they have different values
when attention is distracted to something else. It might also be possible
to detect reorganization as the parameters change, assuming there is some
kind of intrinsic error (a touchy subject, since that would require doing
something that makes the participant more than just a little
uncomfortable). Not all experiments that we can think of doing are
ethically possible to do.
Of course perceptions that we NEVER observe remain unknown and we can’t
say anything about them. I have never claimed to know of perceptions of
that kind. The answer to your final question, therefore, is “we
don’t.”
Best,
Bill P.