Karl Popper

[Martin Taylor 941006 13:45]

Our internet connection was broken for a while, and apparently we now have
a low-bandwidth link until Friday night. I seem now (Oct 6) to be receiving
postings data Oct 3 and 4, but clearly there has been at least one earlier one
from Hans Blom and one from Avery Andrews that have not arrived yet. I still
have hopes of receiving them, since in recent times some postings have arrived
as much as ten days after the posting date, and eight days behind other
contemporaneous postings.

ยทยทยท

====================================

Hans Blom, 941003

On Popper:

Popper was at his best when he
could demolish another philosopher's theses, and he did, continuously.
That, at the same time, was the core of what he proposed: one million
examples pro cannot prove a theory, wheras one con falsifies it. There-
fore, was his opinion, it is the greatest honor for a theory if people try
to prove that it is wrong.

I have always thought this to be a very silly proposition. For any theory
whatever, it is EXTREMELY unlikely that it is EXACTLY true. It is so
unlikely as to suggest that there exists a true theory, which can be stated
succinctly:

There is no means whereby a human can propose a true theory.

This theory can, in principle, be falsified, by one counterexample. Is it,
in principle, possible to find a counterexample? I think not, inasmuch as
for any theory for which no counterexample has been found, a counterexample
may be found tomorrow even if none has been found in the thousand years since
the theory was proposed. Therefore, for THIS theory, no countereample
WILL be found, even though, in principle, one COULD be found.

It seems akin to Goedel's proof of the incompleteness of formal mathematics,
in spirit.

Much more to the point, as Bill Powers (941004.0700 MDT) pointed out, what
matters in a theory is not whether it is "true" of the real world, but how
well it describes and predicts what we perceive IN COMPETITION with other
theories that purport to be applicable to the same set of perceptions. The
best descriptions are usually process models ("explanations", such as PCT),
because they require fewer items of data than do more "pictorial" (shallower)
descriptions.

The whole notion of the "truth" of a theory seems vacuous. All, and I
mean ALL, theories actually proposed are false in some respect, and we
know that as solidly as we know any fact. So what is the point of trying
to "falsify" a theory? What we should be trying to do is to establish the
precision of description and the range of phenomena over which any level
of precision is maintained, while at the same time minimizing the number
of special conditions that must be described in respect of any particular
data set. Process models ("explanatory theories") that use very little
in the way of special conditions, and cover wide ranges of phenomena
with possibly good precision, are better theories than are models that
simply use a mathematical relationship to describe a set of variables
that have moderately consistent effects on data ("stochastic models" or
"statistical descriptions").

A statistical description type of theory is inherently unfalsifiable, but
such theories can be quite useful and will be used, at least until a process
model is available that is easier to use and provides better prediction. I
think that Popper's ideas have done science a great disservice. At the
least, it has encouraged the development of psychological science in a form
that uses "null hypotheses" and "significance tests")

Popper's principle of falsification means to try to reject hypotheses
using a critical and rationalistic attitude. That implies daring to criti-
cally investigate your own conclusions and enjoying others' attacks on
them. Support of your theories may flatter your ego -- but does nothing
for the progress of science.

It isn't the "principle of falsification" that makes critical appraisal
valuable. In fact, that principle is quite irrelevant since its application
can provide no information. What critical appraisal does, in science as
in everyday life, is to help define the limits of "reasonable validity"
for a theory. Every theory is false, but likewise every theory has some
range of validity, even if it be only the single set of data that encouraged
the initial proposal of the theory.

It is interesting that you mention the different evaluation people give
to supporting and critical information about a hypothesis, because only
yesterday I gave a briefing of a couple of hours duration to a group
of "conventional" psychologists--psychophysicists and social psychologists
who worked on the sleep deprivation study-- on this issue. I treated
it as a natural aspect of the workings of the "category level" of
perception, not as an aspect of the philosophy and practice of science.
It's a VERY pervasive problem, which shows up in a lot of different guises
that I want to investigate some time.

Martin