Harzem Peter wrote:
Now, it is a different matter when it comes to Popper. You write:
Finally, going Popperian (i.e., sure one study can overturn a theory)
isn't really going to get much more respect. Popper's most
infliuential student, Imre Lakatos, presented example after example
of situations in the natural science in which what initially looked
like a refutation turned out to be simply motivation to adjust
(rather than reject) the theory to accomodate the new data and then
declare the theory to be stronger and broder than ever before.
I am very much for that, and think very highly of Popper. (I gave a
paper in the Centennial celebrations of Popper held in Vienna.)
What changes is the theoretical explanation, not the data on which the
revised theory had been based. Some of the prior data may be found
to be faulty and be discarded, of course. However, there is an
important exception: a single set of data can, indeed, falsify an
theory (but not the data).
Many decades of philosophical work since 1935 (when Popper wrote the
book translated in 1959 as _Logic of Scientific Discovery_), have shown
that the distinction betwen theory and data is often not so clear as
this formulation would have it. To borrow an example from Lakatos,
imagine an astronomical theory that predicts a certain phenomenon should
be found at a particular place in the sky. Astronomers train their
telescopes on the crucial coordinates and find, alas, that no such
phenomenon can be seen. Theory refuted? According to Popper, yes. But
things are not so simple. The "refuted" theoreticians argue that a dark
cloud of matter obscures the phenomenon predicted by the original
theory. A probe is sent to the place where the hypothetical cloud is
said to exist, finds evidence of it, and, further, peers beyond it to
find the phenomenon predicted by the original theory. The presumably
"refuted" theory is "revived" (because obervational "data" came with a
tacit theoretical claim that nothing is obscuring the phenomenon form
the perspective of Earth).
WVO Quine (following Pierre Duhem) provided lots of examples of data and
theory intermingling in ways that make the simplistic formula of "data"
refuting "theory" problematic. Consider: I dig. I find the bones of a
brontosaurus. Data, no? Then it turns out that the original example of
the brontosaurus (the one that gave the species its name) was actually
the skeleton of the previously discovered and named apatosaurus, mixed
with the skull of some other animal entirely. The very category
"brontosaurus" vanishes out of the paleobiological vocabulary, and my
"data" changes (Presto!) to having been the bones of an apatosaurus. The
"data" was "contaminated" with theoretical assumptions. So we have to
strip out all theory from data, right? Way back in the 19th century,
"phenomenalists" tried to strip things down to "sense data" in order to
avoid this sort of problem. (Indeed, this was exactly the reason that
Percy Bridgman invented the "operational definition" in the 1920s; an
idea that fell on its face everywhere but in psychology where it still
enjoys a kind of afterlife -- see my 192 article at
http://www.yorku.ca/christo/papers/operat.htm). The idea carried on into
the mid-20th century in the logical empiricism of people like AJ Ayer.
But it ultimately failed. One cannot even categorize data as being
"this" or "that" without bringing in a raft of theoretical assumptions.
And one cannot do science without categorizing data. Thus the dilemma
(and the reason that Popper's 1935 formulation cannot bear the weight of
explaining how science works).
Regards,
--
Christopher D. Green
Department of Psychology
York University
Toronto, ON M3J 1P3
Canada
416-736-5115 ex. 66164
[EMAIL PROTECTED]
http://www.yorku.ca/christo
======================================
---
To make changes to your subscription go to:
http://acsun.frostburg.edu/cgi-bin/lyris.pl?enter=tips&text_mode=0&lang=english
