Re: Relational "Space"

[John Kineman <***>]

I've been trying to catch up on this thread and got this far before feeling
the urge to comment.

I agree that elements considered "real" in a view or theory do not have to
be observable, and yet I also agree with Tim that these are theoretical
"reals." On the other hand, there would then be no meaning to anything
"more" real than that, so back to Judith. In fact, most scientists call
something real if its predicted effects are confirmed (or not disproven).

Exampel: Forces are not observable, but they are a part of science and
"considered" real with regard to classical phenomena. This does not stop
someone from trying to explain where the force comes from, in terms of
something they want to call "more real." But unseen realities abound in
science. What is generally required is that the resulting phenomena can
best be explained in terms of such unseen realities, and that predictions
from such assumed realities all have confirmable results. One then tries to
devise a clever experiment specifically to test if the most critical
prediction can be shown to be wrong.

We are all too influenced by past views of reality and stories about "what
is science." Science is what works to explain things under testable
conditions. So you can propose any kind of primary structure you need to
get the results to come out right. That is what physicists are doing with
27 and 11 dimensional universes, etc. If they get the math to work so that
it conforms with observed results, then they call the 27 dimensions real
but not observable themselves. That's the way its done.

However, when complexity is introduced we lose the myth that explanations
can fully commute with nature. In that case we have to allow for
exceptions. So the 27 dimensions may be considered right if the error is
small or the exceptions not very common or interesting (to the scientist).
The test gets weaker if we do not believe in fully commuting relations with
theory -- i.e., full correspondence. In that case, we have to accept "good
enough" as a scientific test. But good enough for what? Add to that
parsimony (Occam's razor), which is NOT just simplicity (a common
misconception) but rather elegance, i.e., the least number of elements or
assumptions needed to explain a given phenomena. So Newton's theory is
parsimonious compared to alternatives for explaining classical behaviors.
It fails generality or universality (other criteria that have to be added
to parsimony) if applied to non-classical behavior because it isn't good
enough for that. So then we go to a more complicated theory and try to make
that as simple as possible and yet still good enough for the new phenomena
- i.e., elegance in explanation, not simplicity alone. If you can get rid
of aspects of a theory and have no loss in ability to explain what it was
intended to explain, that's parsimony.

Also, parsimony does not say that nature is simple or takes the simplest
course. It is about theories, not nature. It says the best theory is one
that is uncluttered with unnecessary and confusing elements, because the
least cluttered one that does the same job is easier for us to use.






At 07:49 AM 3/18/04 -0500, you wrote:
> > -----Original Message-----
> > From: ROSEN Forum [mailto:*** Behalf Of Judith
> > Rosen
> > Sent: Wednesday, March 17, 2004 9:54 PM
> > To: ***
> > Subject: Re: Relational "Space"
> >
> --snip--
> > Why does something need to be perceptible to be accepted as real,
> > when there
> > are phenomena (effects) that are being caused by it?  We should be able to
> > prove its existence indirectly, via the behavior of the effects under
> > certain circumstances. The need to "verify the existence of something with
> > one's own senses, directly" is right out of the Cartesian methodology.
> >
> > Judith
> >
>
> I disagree that we could prove its existence indirectly. Primarily for the
> reason that we could not distinguish this theory from some other different
> one which might also purport to explain the causal basis behind phenomena,
> if both theories adequately predict phenomena but both consist of
> imperceptible things. At best, all we would have is an observation that the
> phenomena behave *as if* they were generated by folding. We would not have a
> commuting modeling relation, because we could not have measurements on the
> elements which possess the hypothesized causal entailment relations.
>
> I think this is akin to the state the quantum mechanics is in (and has been
> in all along). It has equations that predict certain phenomena correctly,
> but it is unclear if elements in those equations ought to be interpreted as
> real things and real causal relations, or not.
>
> Regards,
> Tim