What is a theory? LO3659

John O'Neill (jao@itd0.dsto.gov.au)
Tue, 7 Nov 95 17:10:43 +1100

It is often postulated that we reason about the world using theories of
how we expect the world to behave. In cognitive science these theories are
called mental models. In computer science (and mathematics?) these
theories are called models or simulations.

One question I have is what is a theory?

In mathematics, a theory consists of a set of axioms, and rules of
inference. Another way of viewing a theory is a representation of
observations, assumptions and conclusions. Theories are also viewed as a
description and explanation of behaviour observed in the world. Finally,
theories are conceptual models that involve using variables and processes
to describe the world.

Each of these descriptions sound similar in some respects, however, the
people who have suggested these various descriptions refuse to accept that
the mathematical definition of axioms and inference covers all other
descriptions - unfortunately they haven't been able to articulate to me
WHY they're different.

The second question I have is how do we construct new theories?

Learning (to me) is all about the construction of new theories to deal
with situations in the real world. These new theories are often
constructed as a result of invalidating assumptions in existing theories.
For example, Einstein's theory of relativity arose by firstly invalidating
the assumption that the speed of light was constant as assumed by
Newtonian physics.

Questions here include:
- what are the "generic" components of theories
- how do we reuse these components across theories
- how do we know whether we can/cannot use these components (i.e. the scope
or context of a component)

In organisations, I see strategic planning as being one mechanism for
developing new theories. The question I have is what is more important,
the theory developed, or the analysis that produces the theory?

If the analysis is more important, how do you capture this?

Questions, questions. The sad part is that I'm sure that _someone_ in
philosophy has "solved" these problems, and that people like John Warfield
and Jeff Conklin and probably most of you on this list have at least tried
to solve similar problems (actually John goes a step further and has
developed mindbugs that are theories about how people don't develop new
theories from mankind's total knowledge perspective). And that leads me to
the corporate memory issue of how do you reuse knowledge created by other
... but that's a different thread :->

John O'Neill
DSTO C3 Research Centre, Australia
email: jao@itd.dsto.gov.au