This business of "Is the moon there..." is one of the oldest issues known to mankind, and has virtually nothing to do with QM, nor with empirical science. You have your platonic ideals to Hume's extreme skepticism, with which I pretty much agree. As several have pointed out, we observe the moon all the time -- it is, from time-to-time directly visible. When it's not, its gravity has a well defined and well predicted effect on the earth; similarly with the sun, ... Thus it is prudent to assume that the moon is indeed there all the time. Astronauts have confirmed this to a point; during their time their experience told us all that the moon was solid at the surface, was round, ...
The point is that there are other possible configurations, cf. optics and potential theory, that can provide a model of the moon that accounts for what we know of it -- great homework assignment.
So why do we assume the moon is in fact a solid body, much like other planets and moons in the solar system? Its the best Occam game in town. We work with what we call common sense and intuition and assumption -- all of which have changed radically since Plato's time -- sort of like "It's the economy stupid" a memorable statement of recent political affairs.
If you really think about it, there are a huge number of assumptions required to assert that "A is really there" or "B is not there." The key one, to which virtually all of us subscribe, is that there's something out there that generates the signals which we perceive -- and there's no way to find out whether or not that's true, because we rely totally on our perceptions to live, to understand, to do science, and so forth. Over the centuries we have become highly sophisticated, or so we like to think, and we've adopted a perspective on science, made authoritative via Popper's idea of falsification -- hardly a euphonious word --, with which David Hume would agree. And, of course, to date our common assumption of an external world has survived in admirable form. It boils down to: if you cannot observe or measure something by any means, for a short or long time, then you have no empirical way to assert existence during the time of observation -- this is true of legal evidence and reasoning as well as scientific evidence and reasoning. But the assumption of continuity of existence -- including possible alterations by one means or other -- has served us well. But make no mistake: reality is a human invention, a convenient assumption beyond proof.
Schrödinger's Cat? Perhaps I'll drum myself of the respectable scientific community, but I have never understood "the problem" since I first heard of it many years ago. In no small measure, my take on this matter has been fortified by many years of day-to-day work with statistics and probability, which I've also taught at the undergraduate level. If you don't know, you don't know. Of course I work with the assumption that we have yet to encounter a cat which in, other than some mystical or a specific medical sense, simultaneously is alive and dead. Why in the world would we assume the cat's state in Schrödinger's experiment takes on a wholly different aspect than one we see in the "real" world? Cats, humans, other animals are either alive or dead -- except, perhaps in unusual medical circumstances -- are either alive or dead. There is, certainly, a probability that the cat is either alive or dead, the two allowed states. And the probability is conditioned on the state of the radioactive system, which does or does not decay. To me, standard probability theory does an Occam-like job with Schrödinger's cat.(Not dissimilar to assuming, in an unwatched horse race, that A won, because just before crossing the finishing line B stopped and tried to turn around. Arguably this could happen, but...)
Suppose you did many EPR experiments with detectors at different distances from the source, or many scattering experiments with detectors spanning the space allowed to the scattered particle. (I mean one set of counters per experiment; not multiple scattering configurations.) You, of course, will find events in different configurations, and different places. That is you will map out the density of events in the appropriate configuration space, and thus you will map out the probability of events. This approach is equally valid in the classical and quantum worlds --measure and count. (Recall that the low energy cross section for charged particle scattering, the Rutherford cross section, is equally valid classically and quantum mechanically. Granted, this scattering phenomena occupies a small portion of the space of all possible scattering experiments. Nonetheless it is suggestive that a straightforward application of basic probability to such low energy scattering is quite appropriate.)
Finally, the results discussed above, define perfectly additive measures, as do all spaces of disjoint events -- a single measurement cannot simultaneously give both a spin up and a spin down result. For a basic probability measure to exist, discrete or continuous or mixed, an additive measure is necessary and sufficient. Thus, a global perspective puts QM on a standard probability basis, and therefore allows a "knowledge interpretation", a la Sir Rudolph Peierls.
Regards,
Reilly Atkinson
If you *really* want to be politically correct, you should say that in quantum theory, the act of observation changes PERCEIVED reality, even after the fact.
.