Personal Construct Theory

See http://ksi.cpsc.ucalgary.ca/articles/NewPsych92/ & WebGrid

Abstract (NewPsych92)

Personal construct psychology is a theory of individual and group psychological and social processes that takes a constructivist position in modeling human knowledge but bases this on a positivist scientific position that characterizes conceptual structures in axiomatic terms. It provides a fundamental framework for both theoretical and applied studies of knowledge acquisition and representation. This paper presents Kelly's original intuitions underlying personal construct psychology and links these to its foundational role in cognitive and computational knowledge representation.

Intro (WebGrid)

WebGrid II is an implementation for the World Wide Web of George Kelly's repertory grid technique for building conceptual models based on his Personal Construct Psychology (PCP).

WebGrid asks you to define a domain of interest, a context or purpose, and some elements or entities that are part of the domain and relevant to your purpose. It then elicits constructs from you which establish how you distinguish the elements in your domain in ways that are relevant to your purpose.

WebGrid provides a variety of methods for modeling and visualizing the relations between your constructs. It also enables you to compare your constructs with those of other people. Facilities are also included for using your conceptual model as an expert system.

What does "takes a constructivist position in modeling human knowledge but bases this on a positivist scientific position that characterizes conceptual structures in axiomatic terms" mean in English?

Too many syllables all right. But I did some papers on knowledge engineering using this stuff about 15 years ago, so I'll venture a translation: PCT suggests that human behaviour is based on a scientific method. It says that people build maps of the world and act on them, but that the dimensions of the maps are not fixed - they're constructed by theory and experiment. PCT says, basically, that to understand someone you have to elicit their terms of reference, not categorize them within your own. In this it's refreshingly different from just about every other systematic theory of psychology.

Where the axiomatic bit comes in is that PCT tries to come up with a kind of calculus that, given formal descriptions of someone's worldview, can produce a generic explanation and even prediction of their behaviour. This is to a certain extent where it falls down, too, because the formal descriptions its methods elicit are often too concrete to explain the vagaries of big sacks of protoplasm like thee and me. Still, it's a lot more flexible and expressive than the alternatives. --PeterMerel

I wonder why PCT is not mentioned in Wikipedia.