EV, CEV and CCEV

<[Bill Leach 950627.01:53 U.S. Eastern Time Zone]

NET

Martin and I have been having a little private "fundamentals" discussion
concerning Environmental Variable, Complex Environmental Variable,
Controlled CEV, perceptions and points of view.

When we model, the Controlled Environmental Variable and the perceptual
input function are (eventually) well defined for the model. When a model
for a tracking task is created for example, it is reasonable to talk
about the CCEV with some measure of certainty (though are "word
descriptions" of what is actually happening functionally in the computer
could be wrong).

When we are talking about a real being in a real event, the "controlled"
CEV is far from certain. We use the observations of a CEV to infer a
possible controlled perception but our certainty of the CEV is always
questionable much less the subject's perception of the "object" of our
perception.

Between these two come the "hypothetical" discussions such as: dog
chasing cat, driving car, opening door, eating, drinking, "making
decisions" and influencing others. In these sorts of discussions, it is
usually necessary to posit that the exact relationship between a
perception and the EV (usually assumed to be an exact one to one) are
known and that the EV is precisely defined.

I believe that everyone one on the list is quilty of both failing to
present a situation such that it is obvious which class the discussion
belongs to and "fouling up" an otherwise good presentation by "jumping
on" the sort of assumptions necessary to even make a "hypothetical"
discussion possible.

Though not necessarily with a whole lot of success, I have been trying to
be more careful in my thinking and therefore postings with respect to
these matters, Martin perceived and "called me on" a useage error on my
part (correctly). The discussion was so useful to me that I perceive
that others might appreciate the truely outstanding summary that resulted.

Yesterday, I received a message from Martin that fits my informal
criteria for a "PCT-GEM". That is a post that addresses one or more
concepts that are vital to a PCT understanding AND is exceptional in
presentation.

With Martin's permission granted the message follows (headers removed):

Bill,

One of the things Bill Powers, correctly, harps on from time to time is
that there is no "thing" out there to which perception is linked. In
other words, there aren't any independent, physically distinct
"Environmental Variables" that we have to find by modifying our
perceptual functions until they match these pre-existing "EV" s.
Whatever we perceive, we construct from the effects of some unknowable
"out-there" factors on our sense organs. Our PIFs can perform any
neurologically feasible function on these effects (i.e. the sensory
signals), and since a PIF can be a logical function, it can perform any
function at all, provided it is not so complex as to demand more than a
brain-full of processing.

Any PIF defines a variable in the environment, but that variable is a
complex of all sorts of things about which we can know nothing. We see,
say, a rainbow. That's a simple perception, but even what little we know
about the environment (e.g. light refracts in raindrops differentially
with [respect to] wavelength) ensures that the environmental correlate of
that simple perception is complex.

It may seem redundant to say "Complex" environmental variable when all of
them are complex, but I think it is better to do that than to be tempted
into thinking of Environmental Variables as real things that can be
perceived and measured by anyone appropriately instructed. The idea is
to emphasise the constructed nature of the CEV. That it is constructed
makes it no less real.

When we come to controlled perceptions, we run into two quite different
problems: (1) it is impossible to control an environmental variable,
complex or not. One can control perceptions; (2) the survival of the
organism, and the state of the intrinsic variables (other than the
control error) depends on the environmental variables, and not on the
perceptions. It is because of the distinction between these to facts
that I like to keep the concept of the CEV distinct from that of the
perception to which it corresponds. It is because of (2) that there is
usually a fairly consistent relationship between some (complex) aspect of
the environment (as measured by instruments) and some perception--if the
perception is controlled, then that aspect of the environment is
stabilized. But there's no guarantee that controlling perception
stabilizes any aspect of the environment that is important to survival.

Bottom line: EV would be technically fine, but psychologically could
mislead one into the "reification" error. CEV as "complex" environmental
variable is an important construct to keep distinct from the controlled
variables, either the controlled perceptions or the corresponding CEVs.

In PCT, that the EV might or might not be controlled is usually central
to whatever discussion is in progress.

That's a colloquial truth, in that the control of a perception usually
means the stabilization of an environmental variable; it is the
stabilization of the EV that another observer might see, even though that
observer's CEVs are never identical to the CEVs (controlled or not) of
the person observed.

There's always the problem of the subject's viewpoint, the ECU's
"viewpoint", the observer's viewpoint, and the analyst's viewpoint. The
analyst can deal quite happily with the perceptual signal and the value
of the corresponding CEV. The observer can perceive the value of a CEV
that might correlate well with the CEV corresponding to a perception
available to the subject, but cannot perceive the subject's perception.
The ECU "has" a perception, but has no "view" of that perception. The
subject can potentially perceive with another perceptual function both
the controlled perception and a related CEV, just as the analyst can, but
for real.

They're all different. And those differences sometimes get lost in the
heat of discussion, with resulting confusion on all sides.

Martin

-bill