🚀 go-pugleaf

In article <Aug.13.16.09.52.1992.25234@ruhets.rutgers.edu> bweiner@ruhets.rutgers.edu (Benjamin Weiner) writes:
>Andrew Mullhaupt writes:
>>You are in danger of fragmenting physics into a disconnected set of 
>>fragmentary explanations valid only in the neighborhood of legitimate
>>observations. 
>
>It's not already?

What is conservation of anything if it is conditional? The fundamental
quantities of physics depend on _universal_ applicability. You can't
leave thermodynamics out in the cold and then claim that probability is
conserved. Let's see how far we can get by deciding that parity is 'kind of'
conserved. So, physicists may think that physics is a disconnected net of
observations, but needless to say that nobody has ever directly observed
a continuous group of symmetries. So it's the issue of which blanks you
feel confident about filling in, or to belabor the point - where do you
choose to stop your explanation and start your observation.
>
>>This kind of physics has very strange versions of things like the second
>>law of thermodynamics.

>I wouldn't trust any kind of physics in which the second law of
>thermodynamics was anything other than faintly mysterious. 

But it is no more nor less mysterious than any other law of physics.

>It's
>a law of thermodynamics, anyway, induced from and valid in
>thermodynamics.  If you want to pretend to "derive" it from something
>else you wind up making equally [valid/absurd] assumptions which
>are no less mysterious - the postulate of equal a priori
>probabilities ... the ergodic hypothesis, anyone?  If someone asks
>me why you can't make a perpetual motion machine, I don't feel like
>responding "I can't prove that you can't, because I can't prove the
>ergodic hypothesis."

You can't prove anything else in physics at this rate. If we look at
physical laws, we may adopt Eddington's distinction between primary
laws (controlling the behavior of single particles) and secondary laws
such as the increase of entropy (which are applicable only to collections).
Eddington saw a loss of certainty in the primary laws as the only outcome
of the Einstein-Bohr debate. He thought that perhaps the collective laws
might someday be seen to be primary and the singular to be secondary.

It is a choice to view the direction of derivation - is thermodynamics
derived from something else, or is dynamics the 'derived' theory. Now
the Brussels school wants to put physics on a foundation of entropy
operators - and there are certainly abstract notions of thermodynamics
which need not appeal to fixed dynamics. One way to amplify this choice is
to compare the failed quest to discover the microscopic laws of physics
from macroscopic kinetic behavior. (Cf. Brush). This underscores the fact
that the collective laws are compatible with many possible dynamics, and
so it is not really sensible to view them as derivative. We can use the
example of lattice models, and even cellular automatic hydrodynamics to
put a fine point on the independence of the collective from the singular.

Where you end up, is with all the propositions of physics spread out
horizontally, with their sponsoring experiments standing next to them,
but it becomes apparent that the interpretation of some of the experiments
depends on the results of others. If we want to talk about physics as
derivative, then we have to attempt an arrangement where none of the
fundamental experiments need the results of any of the derivative ones.

Historically, thermodynamics followed dynamics, especially due to the
principle of equivalence of heat and work being seen as the validation
of a particular notion of heat. But we can quite easily arrange our
physics so that temperature comes first - which is why we need not view
thermal physics as derivative, other than through historical chance.

>I thought it was fairly well conceded by now that scientists do not
>proceed by the "scientific method" as it is usually set forth.

Yes, but I am not concerned with that. I am pointing out the fact that
the scientific method (either as practiced or theorized) does not afford
us much help in choosing how we interpret the facts which survive the
processing.

Later,
Andrew Mullhaupt