Since MM seems unable to argue his point about the Casimir force and zero point fluctuations in any consistent manor. Constantly using references that describe the Casimir force as arising from the fluctuations of the zero point field that he claims is ‘physically impossible’. I present the following paper for review and comment.
http://cua.mit.edu/8.422_S07/Jaffe2005_Casimir.pdf
I've been to one of Jaffe's seminars on this topic, and asked him a few questions afterwards, and in consequence I think understand his technique reasonably well. (Famous last words!) In order to calculate the energy of any system---be it (a) the vacuum or (b) the vacuum plus two conducting plates---you start writing down all of these standard field-theory loop diagrams. It turns out that the list of such diagrams is infinite, which has been well-known since Feynman. In order to calculate any measurable quantity, though, like a scattering amplitude or a mass or a force, there's a way to reorganize the equations that makes it clear that the measurable quantity will be finite.
For the Casimir force, there are two ways to approach the infinite list. Casimir started by taking the full infinite list of QED terms for the vacuum, then showing that a handful of them are removed if there are conductors present. That handful of terms represent the
component of the energy density which is affected by the plates, which gives you the force (via, I might add, p = -dE/dV) and you don't care that the remaining terms are infinite. Jaffe points out that you can do it the other way---start with the QED representation of the plates, then *start* writing the list of fields which are affected by them, which he did with some nifty scattering theory. This gets you to Casimir's same finite list, and to the same force, but by treating it this way, you can stop writing the list of terms
before you get to the infinite (but force-free) part. (This "know when to stop computing" has been built into scattering theory for a long time, and it's basically where the whole renormalization thing got started.)
This is also the same as saying that
the Casimir effect only probes the modes that couple to the plates. If you fill the vacuum with virtual neutrinos-antineutrino pairs, the plates do
not remove these energies from Casimir's list, nor add them to Jaffe's list, no matter what their total energy is, and so they do not appear in the force.
You could also say something like "The Casimir effect is just a test of cavity QED", or that "The Casimir force is just a probe of the plate-plate coupling via virtual particles", and I think those would be fair statements. But a similar statement is true of
every measurement in quantum mechanics---you only ever measure the diagrams that you couple to, never the ones you don't. In the Casimir effect, by moving the plates closer together you are coupling to
more and more diagrams, and I think in the limit of d = 0 you're coupling to "all of them" in some sense. (Or maybe it's just (1/137)^2 of them?) The diagrams which were "decoupled" and invisible at d = 1um will go ahead and couple in at d=0.5um, and so on; in a sense, this correctly-predicted d-dependence shows that these diagrams were there all along. (I could give a list of similar "it was there all along" effects from experimental particle physics.) But it doesn't tell you whether or not your computations for d=0.1nm are wrong or not---the measurement at plate separation
d doesn't tell you what is (or is not) waiting to couple in at plate separation
d/2, and in that sense it doesn't probe the full vacuum energy.
But it does probe part of the vacuum energy. And you don't need the full vacuum energy to see that P < 0 for the components we do see.
It's all very interesting, though, since scattering theory gives you the tools to calculate the Casimir forces on arbitrarily-shaped objects, which were impossible to do with other approaches. If you include enough detail about the electromagnetic properties of the plates, Jaffe's approach also makes explicit the transition between the Casimir force and the van der Waals force---which, being electromagnetic in nature, can of course be written as a bunch of QED diagrams which show up in your big list.