Self as Apex of hierarchy

[Martin Taylor 950928 12:00]

Brian D'Agostino (950928.1020)

Brian, I have received and read your paper (once), and hope to comment
properly when I have read it more properly. My first impression is
that there are two unrelated sets of concepts: the PCT-related set and
the psychoanalytic/Jungian set. But I may change that opinion later.
The issue of appropriate use of statistics may also arise--but we'll see.


For now, I just want to comment on a bit of your interchange with Bill P.

I like Ockham's
razor, and therefore I prefer to proceed on the assumption that the
hierarchy has one apex, at least in the absence of evidence to the

I also like Ockham's razor, which leads me to believe that the hierarchy
has no apex (which is equivalent to having many). Why do I say that?

The PCT hierarchy is built on the assumption that the controlled perceptions
are scalar variables. In other words, the ONLY thing that is controlled,
anywhere in the hierarchy, is "how much of X there is in the world I perceive"
where the answer is a simple number. Now, if there is such a hierarchy
(meaning "if HPCT is a correct approximation to reality"), then for there
to be one apex, there must be a single number that describes EVERYTHING
that really matters to the individual. One single number.

The logical candidate for this is the self.

In which case, the only thing that could be controlled would be "How much
self do I perceive?" I think you mean more than that, when you refer to
"the self." Your "self" has properties, beyond magnitude.

"Self-image" must refer to a perception of "the self," if it to mean
anything like what it normally means. But it is a complex perception,
having many independent values; in HPCT, it is a set of single-valued
perceptions of the properties of "the self" such as "the self's femininity."
Where are the control systems that are controlling these single-valued
perceptions? In the classic form of PCT, any controllable perception
must be built from perceptions at lower levels, and eventually from sensory
data. Therefore, any control system that controls an aspect of the
self-image must be at a higher level than what it perceives, OR ELSE
what it perceives must be part of the environment outside the hierarchy
entirely. So, "the self" as a controllable perception cannot be the
apex of a classical HPCT hierarchy. It is either at some level below the
top, or in the environment of the perceptual hierarchy (though still
within the organism).

If it is not a controlled perception at the apex of the hierarchy,
could "the self" be a single entity that provides values for the reference
signals of the top-level control systems of the hierarchy? In theory,
perhaps. But where would the different values for the different properties
of the "self" come from? Would they be determined genetically, or by
varying body chemistry, or what? Whatever the answer, the "self" then
is reduced to a simple voltage generator, and the reference values for
properties of the self are determined by a variety of things, elsewhere
than in the control hierarchy. That's not a very satisfactory Ockham's
razor solution.

Bill P. and I have talked about the top of the hierarchy a few times on
CSG-L, and we favour different possibilities while recognizing that reality
might be quite different from either. My personal preference is to see the
biochemical intrinsic variables as the primary controlled variables, which
makes error be the only criterion variable for the reorganization control
system. Bill prefers the top level to be a set of ECUs with reference
values of zero. In his version, the reorganisation system is a separate
hierarchy having the intrinsic variables as its primary controlled variables.
Reorganization builds into the perceptual hierarchy whatever turns out to
keep the intrinsic variables under control, and the properties of the "self"
are presumably fairly near the top of the perceptual control hierarchy
that results, in most humans. I think either view is workable, and neither
(as yet) testable. But each way, perceptual control for properties of
the "self" involves no extra assumptions. Ockham's razor has nothing to
shave. The unnecessary single control system at the "apex" of the hierarchy
has been shaved away as an unnecessary and unhelpful entity.

Ockham's razor may have a sharp edge, but where it cuts depends on the barber.

That may be a cute saying, but it holds more than a grain of truth. It
depends entirely on what you take already for granted, which "entities" are
added when you attempt to describe a new phenomenon. From one viewpoint
the new phenomenon may require nothing to be added, whereas from another
it may need a whole structure of complex development. What I am saying
above is that if you buy into "classic" HPCT, Ockham's razor says that
there are multiple apices to the control hierarchy. To assert that there
is just one is to add an unnecessary entity that buys you nothing.