Sascha. These are excellent comments. Thanks.
I'm sure you appreciate that it's very hard to do more than scratch the surface of this question in 10 pages and was intent on keeping it as close to math/physics as possible. There was an awful lot I wanted to put it, and I'm not happy with the actual argument itself. Also, however, this essay was for fun rather than part of my serious research: I am only a hyper-Pythagorean in my spare time!
On to your comments:
You are perfectly correct that it is mathematics doing the work here rather than physics, but that is why I included the merger between physics and mathematics. Physics would then amount to the probing of those parts of the mathematical universe that interest us humans.
On the category mistake issue. My original conclusion had Wittgenstein's stuff about letting the fly out of the bottle, essentially by dissolving the "big questions" into such simple category errors. I disagree with such objections. I didn't split into possible, necessary, existent. Existence was descried as something that can be possible or necessary. When I say existence is necessary I always mean *some kind of existing 'thing'*. We are rather constrained by language here. Of course Wittgenstein and the ordinary language brigade focused squarely on how concepts are used in practice. I thought we'd gotten beyond that straightjacketed form of philosophizing?
By the way, Nozick does not fall into this category. He expresses genuine perplexity over the problem, noting that it is compounded by the fact that anything we might use as explanatory ammunition is part of what needs to be explained, so we end up with circularity (this is the incompleteness problem I mentioned in my essay). I have a copy of his Philosophical Explanations, and what he says is:
"The question appears impossible to answer. Any factor introduced to explain why there is something will itself be part of the something to be explained, so it (or anything utilizing it) could not explain all of the something - it could not explain why there is *anything* at all." [Phil. Exp. p. 115]
Nozick views the problem as forming the absolute limit of our understanding. The point beyond which we cannot go. Incidentally, my title is based on a section title from his book! I also like his claim that: "The question cuts so deep...that any approach that stands a chance of yielding an answer will look extremely weird". That is, of course, a necessary rather than sufficient condition.
(1) On the effectiveness of math issue. It's true that there is selection bias going on here: we ignore the unsuccessful. But the successful still needs explaining. How is it that the mathematical representations we construct enable us to make often surprising successful predictions? It doesn't matter that sometimes we get unsuccessful predictions. Of course we do. Consider the prediction of the white spot in the center of the shadow cast by an opaque disc onto which light is shone, predicted by Poisson (on the basis of Fresnel's theory). The unsuccessful cases do not mitigate against cases such as this. There is a sense here in which our mathematical representation and parts of reality are isomorphic. One easy (but hard to stomach, no doubt) way is to assume that reality is itself mathematical.
The "there is existence" was supposed to be elliptical for there is the existence of *something*, but something sounded to hard and concrete. The section this comes from was certainly the worst bit. But I'd run out of space by this point and didn't really have time to revise - have a conference talk to write as well!
(2) You say: "The bare validity of math would be just the same even if no worlds were possible and nobody existed to appreciate it." That is the root of my point.
(3) When I say theory has overtaken experiment, I am referring (as you can probably tell) to quantum gravity and 'beyond the standard model'. When I say very recent, I mean within the last 100 years (I'm a historian of physics: this is recent for historians)! Riemannian geometry was not devised as a physical theory (though Riemann was aware of physical aspects). True, it was then applied in GR (which was not ahead of experiment, but based on anomalous extant data: Mercury perihelion). This seems like a case in my favour. Not sure what the black hole case is supposed to be proving here. But again, you admit that there are physical/experimental limits imposed by the scales of the new physics. That is precisely my point. If you still want to continue doing physics in the same way, then the recourse to more mathematical explorations seems inevitable.
On the quaternions/octonions as a counterexample to applicability. Certainly not true. Both find plenty of applications in physics (whether they are 'widely applicable' or not seems to be irrelevant). Two nice books are:
http://www.amazon.com/Division-Algebras-Quaternions-Mathematics-Applications/dp/0792328906/ref=sr_1_1?ie=UTF8&s=books&qid=1253180087&sr=8-1
and
http://www.amazon.com/Introduction-Octonion-Non-Associative-Algebras-Mathematical/dp/0521017920/ref=sr_1_3?ie=UTF8&s=books&qid=1253180087&sr=8-3
Also, my point was meant to be more general: even if there happen now to be structures that happen not to be applied to physics; it is hard to believe that one could not find some representational link between them and some aspect of the world - but this is, of course, impossible to prove: it's a plausibility argument.
Thank you again. These were thoughtful, intelligent questions.
Cheers,
Dean