The scientific status of PCP

Brian Gaines (
Wed, 12 Jun 1996 21:48:13 -0700

The question of the scientific status of any form of psychology is a very
interesting one with rich philosophical and psychological literatures over
some 2,500 years. It is a question that has perturbed experimentalists --
Brindley's book on the eye and Poulton's on skills both devote a massive
initial section to explaining why studying peoples' perception and behavior
is virtually impossible -- then go on to do it.

One problem is that one is dealing with a system whose physical basis is still
not well understood, and also where a reduction to the physical may still
leave most questions unanswered.

Research on chaotic systems in the last twenty years has developed an
understanding that psycho-social systems are like the weather -- the unfolding
of their behavior is not determined by the initial conditions, and within
the constraints of boundary conditions many possible unfoldings are possible.

Our folk psychologies see this nondeterministic behavior as "free-will".
One interesting question is why we ascribe this to ourselves but not to
the weather. "Primitive" cultures may well ascribe it to the weather also
but the influence of "science" in popular culture is now such as to make
this unfashionable. However, it might be more perspicuous to note that
we cannot empathize with the weather or talk to it, so that ascribing it
intentions helps little in predicting or persuading it.

A complication with human psychology is that our folk psychologies
are part of our social interactions. How we model one another influences
how we behave to one another and hence our models may become reified.
Levis gives a beautiful analysis of how this is exemplified in Greek
myths and legends. Luhmann sees it as a way in which we simplify a potentially
very complex world by agreeing on conventions. Garfinkel makes it the
foundation of ethnomethodology.

The current prevalent philosophical position is Dennett's
"Intentional Stance" -- that we impute goals and intentions to others
because this is a useful stance in accounting for their behavior.
The arguments noted above suggest that this stance might be predictive,
even if it corresponds to no physically determined process, because
it is an implicit social convention.

Newell has used arguments like this to explain what he calls the
"knowledge level" -- an ideational level with no physical grounding.
His analysis is similar to Ashby's cybernetic analyses of black boxes
in terms of their behavior -- the models we obtain may be predictive
but they tell us little about the internal structure -- and they do not
tell us to what extent the behavior is determined by the internal structure
or by the external environment.

This is the scientific background from which to evaluate Kelly's personal
construct psychology. Kelly discusses these issues himself in his first chapter
which is remarkably fresh and in step with the thinking of 40 years later.
He sees his work as gnosiology (systematic analysis of conceptions
interpreting the world), positivist because of its emphasis on the
constructs through which the world his interpreted, pragmatically
empiricist, rationalist but not realist.

Pragmatic empiricism is probably the key to evaluation. He arrived at PCP
through his clinical and educational studies, and was frustrated by
psychologies that attempted to use "motivation" as a causal variable
and "learning" as something other than a descriptive term for the
phenomena of change. PCP was Kelly's basis for explaining his clients'
behavior to them in terms which they could use for the active management
of that behavior -- to see themselves not as beings to whom life was
happening but rather as beings who had made many implicit choices
and to whom those choices were still open.

There is obviously a personal value system involved is seeing this
approach as being attractive -- choice rather determinism -- control
situated in the individual, not the therapist or teacher -- any manipulation
of behavior being overt and in the hands of the person owning the behavior.
However, a value system behind a theory does not make it unscientific.
One lesson of structuralism, e.g. Foucault's "archeological" studies of the
notion of "mental illness," has been that there are value systems behind all
our sciences and that, if we do not recognize them, we may use them to
justify morally unacceptable practices.

Kelly's model for a principled analysis was Euclid's geometry which commences
with a few primitive definitions and then develops corollaries which
are a working out of those definitions applied to particular circumstances.

His fundamental postulate is definitional -- instead of a gnosiology based
on motivation, habit, learning, etc, he will build one on the way in which
people anticipate events. PCP interprets behavior in terms of anticipation
leading to a rational teleological model of human activity.

All the corollaries are also definitional. For example, the construction
corollary defines what is meant by the term "to construe." It does not
impute behavior to being the outcome of construction. It defines what the
the term construing will mean when applied to human activity.

One test of a definitional system is whether it provides a useful framework
within which to describe the world -- for Kelly, useful meant one that
was effective in therapy and teaching. If one explains a clients' problems
to them in terms of their anticipations being the source of those problems,
and they accept this, does it help them to resolve those problems?

Another test of a definitional system is whether it provides a
generative foundation for commonly adopted ontologies. Kelly's
system is very elegant. It is a natural successor to Hegel's logic
which had very similar presuppositions and argument form, and a
natural predecessor to modern intensional logics which formalize
the definitional structure.

If one wishes to go beyond this and ask whether PCP is the "best" psychology
available in some sense then one has to compare it with the competing
paradigms, behaviorism, exchange theory, cognitive science, symbolic
interactionism, and so on. None wins out -- all have some interesting
insights -- all can be interpreted in terms of one another. The PCP
interpretation of the others might seem particularly valuable because it
delineates the different constructs involved, but a social constructionist
might not find that meaningful and prefer to delineate the different
stories being told.

Thus, we inevitably return to a pluralist, post-modernist position.
There is no one correct approach to modeling the psycho-social world,
and this may be intrinsic to the chaotic, reflexive nature of that
world. There can be wrong approaches that are either out of step with
current society, or could never be in step with any human society that
we can contemplate. However, writers have invented possible worlds of social
behavior which are incredibly different and yet plausible, so what we
can contemplate goes far beyond what we have experienced.

In conclusion, it seems to me that one should read Kelly as he asks
to be read and explains very clearly in his first chapter. It would
be completely against his intentions to reify personal constructs
as being some kind of physically embodied psychological reality
that lead to behavior. It would be consistent with his statements
to see anticipation as a perspective from which to describe human
behavior, and constructs as terms in a vocabulary for that description.
It would be consistent to ask anyone who uses the PCP vocabulary
why they have found it useful in their activities, and where they
have not, and where they have modified it and extended it, and so on.

The PCP literature reflects all these issues and is by no means a
monolithic endorsement of Kelly's position. The Journal of Constructivist
Psychology has been very eclectic from the start in encouraging
submissions that are negative to aspects of PCP. The community
delights in debate. A highlight at NAPCN96 was the opening debate
between Hank Stam and Jim Mancuso which pitted social constructionism
against PCP. If a science is defined in terms of its open debate in a
reflective critical community then PCP qualifies.

However, I wonder whether the list server is a vehicle well-suited to
such a debate. This note is very short as a paper but very long as
email, and yet it would be difficult to say what has been said in
fewer words without leaving out essential steps in the argument.
We need to develop web sites also where more detailed arguments
can be made available for public scrutiny so that the list can
be effective with the shorter messages that are more normal and
socially acceptable for the medium.


Ashby, W.R. (1952). Design for a Brain. London, UK, Chapman & Hall.

Dennett, D.C. (1987). The Intentional Stance. MIT Press, Cambridge,

Foucault, M. (1962). Maladie Mentale et Psychologie. Paris, P.U.F.

Garfinkel, H. (1967). Studies in Ethnomethodology. Englewood Cliffs, New
Jersey, Prentice Hall.

Kelly, G.A. (1955). The Psychology of Personal Constructs. New York, Norton.

Levis, A.J. (1977). The formal theory of behavior. Journal of Social
Psychiatry 23(2).

Luhmann, N. (1979). Trust and Power. Chichester, UK, Wiley.

Newell, A. (1982). The knowledge Level. Artificial Intelligence 18(1) 87-127.