Theory

Dag_Forssell · May 5, 1993, 9:43pm

[From Dag Forssell (930505 1410)]

I have now reconsidered my thoughts on theory (930504.1120) and
(after consultation with Greg) rewritten the defining paragraphs in
"Quality of Theory" Greg helped realize - reminded me - that there
are indeed differences. Not so much in terms of thinking, as I was
concerned about yesterday, but in METHOD. I have changed the
terminology some and expanded or clarified the meanings.

The paragraph on practical applications is new. It takes the black
and white edge off the paper.

···

---------------------------------------------------------
SUMMARY:

Modern man benefits from advanced technology that could not be
imagined a few generations ago. Engineers have made unprecedented
progress in the last three centuries. The reason is generative
theoretical modeling that is different from the empirical
descriptions we all intuitively develop and live by. The nature and
power of generative theoretical modeling is not widely understood.

In the behavioral sciences, no equivalent progress has been made.
We struggle with management practices, education and counseling on
personal issues. This paper points to differences in the kind of
theory employed in different fields and indicates why Perceptual
Control Theory shows the way to rapid progress in the behavioral
sciences in coming years. Specifically: Purposeful Leadership gives
you the capability to predict the effectiveness of management
planning and action.

Empirical descriptions, Generative theoretical models and Reasoning
theory are discussed and the qualitative differences between them
explained by illustrating the difference in practical and advanced
engineering and in contemporary and future behavioral science.

EMPIRICAL DESCRIPTIONS

are descriptions of what is happening based on observations.
Explanations suggest underlying regularities; things that usually
happen "other things being equal." Testing and validation is seldom
rigorous. This leads to an intuitive or statistical recognition of
observed phenomena and empirical laws = rules for practice. Most
"common sense" understanding of our world consists of empirical
descriptions. Illustrated below, this is shown as a "surface"
approach (extending sideways ---> ).

(Figure not shown.)

GENERATIVE THEORETICAL MODELS

follow from a different and more rigorous scientific method. They
are based on logical modeling of functional relationships at levels
below the observed phenomena, followed by testing with an
expectation of 100% verification. They offer explanations of why
and how things happen. In the illustration above, a series of
explanations are shown reaching deep below the surface.

PRACTICAL APPLICATIONS

will incorporate a mix of the above. No physical science can be
without elements of empirical descriptions of material properties,
for instance. A largely empirical descriptive science may have many
pockets of insight that are of a generative nature, whether
formalized or not.

REASONING THEORIES

are sciences like mathematics, statistics, geometry, logic. These
are similar to generative theoretical modeling in that they depend
on deduction, but very different in that they start from a few
arbitrary statements, not from a tested fundamental law of nature.
Thus they are not based on first principles, but on stated
hypotheses. Reasoning theories are employed in support of empirical
descriptions and generative theoretical models, but cannot, in and
of themselves, tell us anything about nature.
-------------------------------------------------
Best, Dag

William_T_Powers2 · September 21, 1994, 5:35pm

[from Mary Powers 940921]

Bruce Buchanan 940920

You distinguish between theoretical models (involving the
invention or discovery of abstract entities) and systems models
(designed to correspond to to real world operating systems), and
place PCT in the latter camp. The difference seems to be whether
the model is universal, true always everywhere, or requires
adjustment, weighting, and improving.

I think you are simply making a distinction between scientific
theory and engineering. An electronics engineer, say (like the
one I've been married to for 38 years) drawing on the abstract
entities and laws of physics, devises a circuit diagram that by
those laws will surely work. Then he builds it. It doesn't work.
Not because of flaws in the laws of physics, but because the real
world is full of components that don't have quite the properties
of their specifications, or because there's some consequence of
their interaction that he hadn't expected, or whatever. So he
adjusts, weights, improves.

PCT exists in both worlds. The laws of control theory are as
abstract and universal as the laws of gravitation. But you are
attending to the cutting and fitting, the tweaking of loop gain
and so on, required to make a particular control system work.
It's like the pilot of a space shuttle, arriving in orbit
according to the laws of physics, having to fuss around firing
little jets to get the thing exactly where all those laws said it
should be.

* * * *

I know you think the use of the word perception at high levels is
confusing, but there is a reason for it. The conventional way of
looking at the mind distinguishes sensation, perception and
cognition. These have been considered different abstract
entities, and it is because they have been that you think it
confusing to use one word for all of them. An analogy would be
viewing metabolism, burning, and the turning yellow of last
week's newspaper as different phenomena. There are important
differences, but discovering or inventing or abstracting their
fundamental similarity, the process of oxidation, opens up a lot
of new territory. Calling it all perception (like calling it all
burning) is the first step into that new territory.

Mary P.

Dag_Forssell · December 15, 1992, 1:43am

[From Dag Forssell (921214)]

Here is a slide from a presentation I am planning. I want to highlight
the common (I perceive) bias aganist theory.

···

---------------------------------------------------------------------

What does "THEORY" mean to you?

APPROACH LEVEL OF TYPE OF
TO THEORY SCIENCE KNOWLEDGE

Type 1 Hunch, expectation Common sense / What works
based on experience. Statistical (sometimes)
Intuitive / Formal research

Type 2 Explanation, Engineering Why it works
prediction, test. science (all the time)

Type 3 Logical reasoning. Abstract Abstraction
science

---------------------------------------------------------------------
Continued to the right.

METHOD OF TIME TO PREDICTION RESULTS
LEARNING LEARN CAPABILITY

..1 Trial & error / Long Poor Spotty
data collection

..2 Create theory, Short Excellent Confident
test theory

..3 Deduction Short N/A N/A

----------------------------------------------------------------------

The idea here is to help the audience realize that when they say: "Don't
bore us with theory, show us what to do," what they object to is what
seems like useless abstractions. By contrast, a good engineering type
theory is very useful.

QUESTION:

Do you think that a mixed audience will relate to these "definitions" of
theory and level of science? Any suggestions on rewording, ever so small,
which will help will be appreciated.

Thanks, Dag

_John_Gabriel · December 15, 1992, 4:09am

[gabriel 921214 21:46]

I love 'em. And although I masquerade as a mathematician, much of what
I do might be engineering. I think every engineer I know will have
enormous recognition reflexes about this, also all the good teachers,
managers, and leaders of all kinds. Perhaps the other two of the
gang of 3 will risk their mgrs' wrath and post them on bulletin
boards (the paper kind). No use posting at Argonne. The worst
place of all to preach is among those who (mistakenly) believe
they are converted.

Marian - do you want to expound on the Hugh Dingle stages of
development of a scientific theory - from bird watching to taxonomy
to Newton, F=ma and Einstein?? Look in my .mailrc for the csgnet
alias if you do.

A small mathematician/logician's nitpick. Besides deduction, there
is induction. Both are actually predictive - Kepler's Laws
demonstrate a central acceleration proportional to the inverse
square distance with a universal constant of proportionality for
the planets observed. That's deduction. The induction is the
Gedanken experiment of putting a hypothetical planet in place
and remarking that it may be any material object.

Would you say that inferring Pythagoras' Theorem from some
simpler propositions was predictive. Moot point. Was it true
all along before anybody even mentioned the topic. Hence
Diodorus, and the stone at the bottom of the ocean that is
not, has not, and never will be seen. Or the tree in the
forest. But probably not the sound of one hand clapping.

Actually I have real trouble distinguishing between deductive
and inductive reasoning, just as I do between forward and
backward reasoning - probably one of those blind spots we all
have.

So. Bottom line. I am a bit uncomfortable about your N/A entries.
But otherwise absolutely delighted.

···

(From Dag Forssell (921214))