[HN Gopher] The dangers of non-empirical confirmation (2016)
___________________________________________________________________
The dangers of non-empirical confirmation (2016)
Author : kergonath
Score : 29 points
Date : 2022-08-13 16:02 UTC (6 hours ago)
(HTM) web link (arxiv.org)
(TXT) w3m dump (arxiv.org)
| RcouF1uZ4gsC wrote:
| Remember how medieval scholars would argue how many angels could
| dance on the head of a pin and craft beautiful logical arguments
| relying on their view of cosmology?
|
| These weren't dumb people. These were likely the smartest and
| best educated.
|
| Tying theories to empirical science is what keeps logic grounded
| in reality.
| marginalia_nu wrote:
| The conclusion that the types of things that can be arrived at
| empirically is preferrable to the conclusions of other means of
| reasoning, ironically enough, would not be possible to arrive
| at through empiricism alone, by virtue of is-ought.
| mannykannot wrote:
| I take your point, and now I am trying to decide if the
| effectiveness of empirical verification, in developing a
| robust consensus for conclusions so verified, is itself
| empirical evidence for preferring such conclusions, in
| domains where such verification is possible?
| svnt wrote:
| All else being equal, isn't a theory clearly more valuable
| if its empirical validation is more accessible?
|
| Stated inversely and ad absurdum, what use is a theory you
| can never prove?
| marginalia_nu wrote:
| This just moves the problem to empirically proving that
| that which is useful is preferable to that which is not.
|
| You might say, "Ah, but you can empirically prove that
| you can do more with that which is useful", which is
| again true, but then you need to empirically prove that
| being able to do more is preferable.
|
| After several more tedious steps, you end up at a
| position like "but _I_ prefer this outcome over that ",
| which may absolutely be true, but this is not something
| you've arrived at through experiments and observation,
| but is an experience in itself. In the end, the crux is
| that I can't construct an experiment that proves or
| disproves what it's like to be you (which is really a re-
| formulation of the hard problem of consciousness).
| kergonath wrote:
| I think you over-simplify the point a bit:purely intellectual
| considerations are still useful, be it Occam's razor or a bias
| towards simplicity. I think I agree with the author in that
| non-empirical factors are useless for verification. You'd think
| that this should be obvious, but apparently not (this article
| is part of some continuous intellectual guerilla warfare around
| string theory*). OTOH, they are still useful when building or
| adjusting a theory.
|
| This is really obvious when reading Einstein: relativity was
| verified _a posteriori_ , but before then there were other
| aspects that made it seem right (simplicity, elegance, the
| ability to combine several apparently incompatible theories,
| etc). It certainly was not verified in 1905, when the paper on
| special relativity was published.
|
| To take your example, arguing is fine, but the whole angel
| discussion should be proven empirically to be of any use. When
| we realise that it could not be proven, then it stops being
| science.
|
| * with the author Carlo Rovelli arguing against string theory,
| being one of the authors of the quantum loop gravitation theory
| which also aims at unifying quantum mechanics and gravitation.
| He is a great scientific vulgarisation writer as well.
| zosima wrote:
| One detail:
|
| When the special theory of relativity was published the
| problem of the aether, or rather that the speed of light was
| the same in all directions, was very well-known. This problem
| was very much based on empirical evidence.
|
| Special relativity provided a mathematical framework within
| which that could be true, and included some further testable
| hypotheses, which were then still unproven. (Time slowing
| down for fast travellers etc.)
|
| But I think it's fair to say that special relativity was
| developed due to empirical necessity.
| kergonath wrote:
| I definitely agree. But we cannot really use (only)
| already-known facts to validate a theory. A theory built
| around such a fact will explain it by design and won't have
| as much explanatory power as if the fact were verified
| after the theory is established. Of course if a theory
| cannot explain an already-known empirical fact, then it'd
| better provide a good explanation, like a different domain
| of validity.
|
| Because of this, I would not consider relativity to have
| been validated until things like the Kennedy-Thorndike
| experiment. Even if it could already reconcile the
| principle of relativity and a constant speed of light,
| which was a significant intellectual achievement.
| Explaining the Michelson-Morley experiment was great, even
| _a posteriori_.
| cycomanic wrote:
| I agree that special relativity was born out of a empirical
| question.
|
| However, that a theory explains a certain known phenomenon
| does not really increase how trustworthy it is. It is
| trivial to construct a theory that explains a certain
| phenomenon, the it is slightly less trivial (that's a bit
| of an understatement) to construct one which does not
| contradict any other known phenomena. However to
| empirically "confirm" a theory requires that the theory
| predicts a phenomenon which has not been observed and is
| empirically observed.
___________________________________________________________________
(page generated 2022-08-13 23:01 UTC)