Re: philosophy and the vacuum

[John Denker <jsd@MONMOUTH.COM>]

At 01:43 PM 9/11/00 -0500, Joel Rauber wrote:

> I am talking about a basis for logically defining a system of units.  In
> this sense we may look at the situation in a gedanken-like fashion and
> treat physics as an exact science; as one does in most gedanken experiments.

Gedanken experiments are exact only because the Gedanken-experimenter
considers only certain aspects of the problem;  all other aspects of the
problem are excluded from consideration.  To say it the other way, if one
considers the full infinite range of possible complications, one should not
expect to find nontrivial exact solutions in physics.

> > If the actual uncertainty is large enough to cause you
> > trouble in practice,
> > please explain what you are doing that requires such high accuracy!
> >
>
> My pondering isn't related to a required accuracy type of a situation; but
> rather to the definition of the unit.

You're not playing by your own rules.  You can't have that cake and eat it
too.  Either
 a) You stop asking about the effect of unknown and hitherto-unobserved
perturbations, and conduct a Gedanken-experiment using a bounded list of
phenomena, or
 b) You stop asking for infinite accuracy.

It's your choice.

> It depends on what you mean by a junk effect.  If you mean the usual sorts
> inaccuracies in laboratory experiments, like the specific examples you cite
> above, then I agree.  But I'm asking more about systematic effects, which
> can only be removed if you know what they are; and if you have an adequate
> theory available to correct for them.

There are two types of perturbations.
  1) The things I call junk effects are due to known physics, acting to an
unknown degree due to manufacturing tolerances et cetera.  It is standard
practice in metrology to reduce these as much as possible and then to
estimate the magnitude of what remains.

  2) The other type of perturbation is due to hitherto-unknown
physics.  This manifests itself when different labs attempt to replicate
each other's work and get discrepancies.  Logically and philosophically
this cannot be ruled out.  Indeed it sometimes happens!  Indeed some of the
great discoveries in physics have happened this way.

Operationally, it has been observed that the meter-stick-calibration
experiment has not suffered significant
replication-discrepancies.  Something could be discovered tomorrow, but I'm
not going to lose sleep over it.

> It also assumes you know what the reference is and can remove the "junk"
> effects relative to it.

Philosophically, that's not right.  You can make a list of known phenomena
and declare everything else out-of-bounds without specifying the details.

> Which vacuum is the idealization that the "SI gurus" intend?  So that I
> can know which "junk" effects to remove?

Operationally, if/when new nonidealities are discovered, the stakeholders
get together and decide which idealization to use.  This has actually
happened with regard to the time standard.  Originally, nobody bothered to
specify how to correct for relativistic effects in time measurements.  At
first it didn't matter.  Then clocks became good enough and portable enough
that you had to worry about the special-relativistic effects of the earth's
orbital motion, and the general-relativistic effects of the earth's
gravitational potential, the sun's gravitational potential, et cetera.  So
they had a big fight about it, trying to decide what was the appropriate
"ideal" measurement.  Last I knew, they had fought to a draw, and decided
to different time-standards according to taste, and in the (rare) cases
where it actually matters, they just specify which one they are using.

> > You asked in particular about the effects of the walls.  So stop
> > philosophizing and do the calculation yourself.
>
> For purposes of this discussion its unnecessary to carry out any
> calculations.  It was enough for me to take you at your word that their
> there are different vacuums which will yield different values for the
> "speed of light" and hence different calibrations of the length of a meter.

But I also said that it doesn't matter as long as there is general
agreement about how to account for the differences.  And there is, for all
practical purposes.
 -- You originally asked about the effect of walls.  That's been dealt
with, quantitatively.
 -- If you want to go on and ask about N-rays from the planet Krypton,
well, all that can be said is that they remain unobserved to date.

This agreement is not an accident;  the current time-standard would never
have been chosen as a standard if there had not already been plenty of
worldwide experience with atomic beams, and a marked lack of
replication-discrepancies.