from Tracy Harms (970309.1300 PST)
Bill Powers (970309.1011 MST)
Maybe I'm just not thinking in the right categories. Can you give me an
example of knowledge that isn't dependent on criteria of truth?Best,
Bill P.
How about the knowledge that drinking sea-water will not sate thirst?
No criteria of truth are required here.
Tracy Bruce Harms
harms@hackvan.com
[From Bruce Gregory (970309.1700 EST)]
Tracy Harms (970309.1300 PST)
How about the knowledge that drinking sea-water will not sate thirst?
No criteria of truth are required here.
Try another example. Mixing a small amount of sea-water with
fresh water slakes thirst better than "straight" water.
Bruce Gregory
from Tracy Harms (970309.1500 PST)
Bill Powers (970309.1521 MST)
Tracy Harms (970309.1300 PST) --
How about the knowledge that drinking sea-water will not sate thirst?
No criteria of truth are required here.So you're thinking of empirical observations as being independent of
criteria of truth? What about adding "... if I remember correctly, and if
past observations are an indication of the future, and if the sample in
question is in fact sea water or its equivalent, and if this sample was
taken far from a river outlet, and if my physiology has not adapted to
drinking sea water?"As far as I know, the truth of generalizations always depends on stating the
conditions under which they are expected to be true. Verifying that the
required conditions hold requires establishing still other conditions, and
so on without end. Not all conditions are verifiable, and none is verifiable
once and for all. That is ... as I understand the situation today.Best,
Bill P.
The truth of a universal statement does not depend upon stating ANYTHING.
It depends upon the way things are.
When we do state something, we can produce other statements which stand in
some logical relation to an initial statement. These include logical
dependencies such as you have pointed to. But these are not "required
conditions" if by that you mean statements of theories which must be
verified before the dependent theory can be stated. (Contrary to your
saying: We can never just say "that's true." -- Bill Powers (970309.1011
MST). We can!)
The point where we have repeatedly parted ways is the point at which I
assert that science *necessarily* advances *without aid from verification.*
(When it chances to advance at all.) What you suppose must be
misunderstanding I see as disagreement because I hear you speak as though
verification is of scientific utility. Criteria might be necessary if
verification were an important process, but it is not.
Also -- but less importantly -- what you seem to mean by criteria of truth
are to my eye not criteria, but are instead the conceptual context within
which a theory may be meaningfully stated. Yes, this context is important,
but it never decides matters of truthfulness.
Tracy Bruce Harms
harms@hackvan.com
[From Bill Powers (970309.1521 MST)]
Tracy Harms (970309.1300 PST) --
How about the knowledge that drinking sea-water will not sate thirst?
No criteria of truth are required here.
So you're thinking of empirical observations as being independent of
criteria of truth? What about adding "... if I remember correctly, and if
past observations are an indication of the future, and if the sample in
question is in fact sea water or its equivalent, and if this sample was
taken far from a river outlet, and if my physiology has not adapted to
drinking sea water?"
As far as I know, the truth of generalizations always depends on stating the
conditions under which they are expected to be true. Verifying that the
required conditions hold requires establishing still other conditions, and
so on without end. Not all conditions are verifiable, and none is verifiable
once and for all. That is ... as I understand the situation today.
Best,
Bill P.
[From Bill Powers (970309.1658 MST)]
The truth of a universal statement does not depend upon stating ANYTHING.
It depends upon the way things are.
That may be true, if the universal statement is truly universally true. But
how do you determine whether it is or not? Are you saying that any statement
which you claim to be universal is true? Don't you have to find out "the way
things are" before you can say anything about them? And do you have some way
of doing that that is infallible?
Best,
Bill P.
[From Bruce Gregory (970309.2030 EST)]
Tracy Harms (970309.1500 PST)
The truth of a universal statement does not depend upon stating ANYTHING.
It depends upon the way things are.
All philosphers are arrogant. All cows are beautiful.
Bruce Gregory
[From Bill Powers (970309.1658 MST)]
The truth of a universal statement does not depend upon stating ANYTHING.
It depends upon the way things are.That may be true, if the universal statement is truly universally true. But
how do you determine whether it is or not? Are you saying that any statement
which you claim to be universal is true? Don't you have to find out "the way
things are" before you can say anything about them? And do you have some way
of doing that that is infallible?Best,
Bill P.
Of course I'm not saying that every claim is true, I'm saying that true
claims are true. But naturally your great curiosity is which ones are the
true ones. The adequate answer is: The true ones are among the ones which
don't fail when we rely on them. We *don't* "determine" that a theory is
true, instead we evaluate problems in manner which help us decide which
theories have failed. Throw away bad theories, and the stock of theories
becomes better because it lacks those errors.
You don't you have to find out "the way things are" before you can say
anything about things -- at least, you don't have to find out more than you
have inherited bodily and culturally. That lets us say things. We make
propositions and explanations which may be true or may be false. There is
no escape from fallibility, but there is an escape from the philosophies
which only work if infallibility is obtained. What makes this
error-rejection approach so much better is that it is potent despite
falliblity.
Tracy Bruce Harms
harms@hackvan.com
[From Bill Powers (970309.2026 MST)]
Of course I'm not saying that every claim is true, I'm saying that true
claims are true. But naturally your great curiosity is which ones are the
true ones. The adequate answer is: The true ones are among the ones
which don't fail when we rely on them. We *don't* "determine" that a
theory is true, instead we evaluate problems in manner which help us
decide which theories have failed. Throw away bad theories, and the stock
of theories becomes better because it lacks those errors.
Since you can't ever determine which of this stock of theories is true, why
speak of "true theories" at all? Why not just sort theories into those that
are demonstrably wrong, those that work with so-so usefulness, and, toward
the other end, those that work so well that we would be astonished if they
ever failed under any circumstances? It seems to me that this is the limit
of what we can say about the truth of any statements, theoretical or otherwise.
We make
propositions and explanations which may be true or may be false. There is
no escape from fallibility, but there is an escape from the philosophies
which only work if infallibility is obtained. What makes this
error-rejection approach so much better is that it is potent despite
fallibility.
I agree, of course. Yet there is the nagging phenomenon of the ultra-simple
theory that seems to explain great complexities. A theory that is just as
complex as the phenomenon it explains is hardly better than a notebook full
of observations with no theory at all. We can't prove that a theory is right
because it's simple, but somehow it seems more right than a more complex
theory that only explains the same things. Is this merely an expression of a
subjective preference for simplicity, or is it linked, somehow, to finding
better and better approximations to That Which Is?
I think that the virtues of the concept of falsifiablity have led us to
neglect other phenomena of explanation, particularly the astonishing power
of simple ideas like Newton's inverse-square law, or e = mc^2 and all that.
These theories, having almost no internal detail, can be used to explain and
predict details beyond measure. There's something qualitatively different
about such theories. The idea that they are simply what is left after a
hopper-full of random statements has been shaken over a sieve until nothing
else is left seems to me inadequate. When did we eliminate the inverse
2.71773 power law, and e = mc^3? I don't believe that Newton or Einstein
ever considered any exponent other than 2. So why does it work so well? A
lucky guess?
Best,
Bill P.
From Tracy Harms (970309.2200 PST)
Bill Powers (970309.2026 MST)
Since you can't ever determine which of this stock of theories is true, why
speak of "true theories" at all? Why not just sort theories into those that
are demonstrably wrong, those that work with so-so usefulness, and, toward
the other end, those that work so well that we would be astonished if they
ever failed under any circumstances? It seems to me that this is the limit
of what we can say about the truth of any statements, theoretical or otherwise.
There is a difference between the truth we can guarantee (which is none)
and the truth which we can rely on (which is all). If we do rely on truths
and we admit our fallibility, we necessarily refer to the actual truth
above and beyond our ability to warrant those truths among our theories.
This is indicated in Don Campbell's excellent term: hypothetical realism.
[...]Yet there is the nagging phenomenon of the ultra-simple
theory that seems to explain great complexities. A theory that is just as
complex as the phenomenon it explains is hardly better than a notebook full
of observations with no theory at all. We can't prove that a theory is right
because it's simple, but somehow it seems more right than a more complex
theory that only explains the same things. Is this merely an expression of a
subjective preference for simplicity, or is it linked, somehow, to finding
better and better approximations to That Which Is?
It is more than mere subjective preference, but it is intimately tied with
this subjective preference. My guess is that in the elegance of the best
theories we see not only a glimpse of a fundamental elegance to reality at
large, even more we see something which reflects a truth about our own
qualities as knowers. As I mentioned before, knowing is always
participatory, and the presence of the knower is part of the knowledge. In
fact it was because I had already come to this way of thinking about
knowledge that I was so receptive to Gary's turn in _Without Miracles_,
Chapter 8 to the viewpoint-of-organism which is integral to PCT, and which
(in my mind, at least) distinguishes it above its competitors.
Anyway, to not stray from the topic, the simplicity of great theories may
indicate a truth about theorizing as well as truth about the nominal
subject matter. Admittedy, just saying this does not explain how they
arise or why they work so well. But those questions go well beyond the
basic topic I've been speaking to, which is the success of improving
explanation by cycles of conjecture and refutation versus the failure to
improve explanation by induction and verification.
Tracy Bruce Harms
harms@hackvan.com
-------------------------------------------------------------------------
"The world of living nature is the way the world
of non-living nature reveals itself to itself."
Peter Munz
[Hans Blom, 970310e]
(Bill Powers (970309.1658 MST))
Of course I'm not saying that every claim is true, I'm saying that
true claims are true. But naturally your great curiosity is which
ones are the true ones. The adequate answer is: The true ones are
among the ones which don't fail when we rely on them. We *don't*
"determine" that a theory is true, instead we evaluate problems in
manner which help us decide which theories have failed. Throw away
bad theories, and the stock of theories becomes better because it
lacks those errors.
I like this a lot, Tracy. It applies, mutatis mutandis, to "internal
models" if one accepts that a theory is not just true or false but
can be more or less correct, fuzzy or hazy. To quote you: The correct
models are the ones which don't fail when we rely on them. We *don't*
"determine" that an internal model is true, instead we evaluate
problems in manners which help us decide which models have failed --
by comparing the model's prediction with the actual perception. Throw
away or modify bad models, and the quality of the models becomes
better because it lacks those errors.
That is why I consider the (mostly unconscious!) building of
"internal models" to be the "science of everyday life".
Greetings,
Hans