Re: the nature of 'construct'

Brian Gaines (
Tue, 4 Jun 1996 22:06:04 -0700

>>I'm getting to the stage myself where I feel that there's a need for a
>>formalisation of Kelly's framework. I find it intriguing that computer
>>programming plays such a large part in PCP, and there well-developed
>>formalisms for describing computer languages, there's no correpsonding
>>formalism for Kelly's Fundamental Postulate and Corollaries. Well and
>>clearly as Kelly wrote, there's too much poetry in him for precision.

I've been away at NAPCN in Banff and missed the start of this thread.

There is a nice logical development of Kelly's postulational system
based on his own "psychological geometry" as he reported it to Luria
when Kelly visited the USSR in April 1961.

Kelly's terminology for elements, constructs, postulate and corollaries
is drawn from Euclid's elements and his system is exceptionally clear
when conceived in geometrical terms.

Kelly's work predated the breakthroughs in intensional logics in the
late 60s, but his geometrical intuitions were sound because Euclidean
geometry is one model for such logics -- projective geometry is the
main formal model for modern relevance logics.

The neat thing about geometry is that it can model pure relationships
in empty space -- i.e. concepts and constructs encompassing no elements.
This is important in modelling wishes, desires, intentions, and so on --
propositional attitudes that may have no specific referent.

Mildred and I wrote a brief overview of Kelly's geometrical model
for New Psychologist -- it is up on the web site:-

and gives the relevant citations.

We show that Kelly's system generates what has come to be called a "description
logic" equivalent in power to the most refined ones developed in the artificial
intelligence literature.

The RepGrid turns out to be capable of fully representing these logics if
one generalizes "ratings" to be arbitrary "constraints", and this in turn
leads to computer tools that can represent the full richness of Kelly's

Thus the formalism does exist and arises very simply and naturally out of
Kelly's original formulation -- the only thing that was missing 40 years
ago was the math/logic to express it formally.


Dr Brian R Gaines Knowledge Science Institute
University of Calgary Calgary, Alberta, Canada T2N 1N4
403-220-5901 Fax:403-284-4707