Re: arbitrary choice of zero of potential

["John S. Denker" <jsd@MONMOUTH.COM>]

At 09:19 AM 10/17/01 -0700, John Mallinckrodt wrote:

> ...Bruce Sherwood and Ruth Chabay ......
>
> "It is customary to emphasize that only differences in potential
> energy are physically meaningful, and therefore the zero of
> potential energy is arbitrary. After 1905 we should have changed
> our tune. When particles are at rest and very far apart, their
> total energy must be equal to the sum of their rest energies,
> which means that the potential energy of interacting particles
> must go to zero at large separations. If an arbitrary constant is
> added, energy and momentum will not transform correctly between
> different reference frames. Recognizing the absolute nature of
> potential energy has far-reaching pedagogical consequences, and we
> have addressed the issues in a calculus-based introductory
> textbook to appear this summer."

Wow.  This seems to be an in-your-face assertion that gravitation is not
gauge invariant.

The usual response to such assertions goes like this:
 -- It is possible to write down non-gauge-invariant expressions and claim
that they contribute to the physics, e.g. as a driving term (source term)
in the gravitational field.
 -- On the other hand, it is possible, with a little work, to find
expressions that express everything that needs to be expressed, while
preserving gauge invariance.

For an even-handed discussion of this issue, see
  http://pancake.uchicago.edu/~carroll/notes/
in particular chapter 4
  http://pancake.uchicago.edu/~carroll/notes/four.ps
and if you don't like postscript you can track down the html version at
  http://nedwww.ipac.caltech.edu/level5/March01/Carroll3/Carroll_contents.html


At 09:38 AM 10/17/01 -0700, Roger Haar wrote:
>         Sherwood and Chabay might want to consider what
> one is required to to in QED.  That is ignore ( OK
> renormalize away ) infinite self energies.

That's another good reason for doubting the S&C claim... but I suspect that
gauge invariance can be achieved even without renormalization.