PeterDonis said:
Your definition of "unreasonable" is your personal opinion. Evidently my personal opinion differs from yours. But PF discussions should not be based on personal opinions. Discussions of what predictions various models make are fine. Claims that one model is "unreasonable" are not.
...
By this criterion, lots of nonsense theories are "viable", such as Russell's teapot.
That empirical viability includes a lot of nonsense theories is simply fact. You cannot get rid of it (except with Orwellian renaming). Logical consistency allows even much more nonsense, nonetheless it remains a useful criterion (say, against MWI and similar nonsense). And that's why you have to allow the use of other criteria to get rid of the nonsense.
So what do you propose? Either you restrict yourself to empirical predictions and their agreement with observation, and reject arguments going beyond this as "personal opinions" with no place in science, then Russel's teapot is fine.
I allow them, and accept the usefulness of criteria like simplicity, predictive power, compatibility with fundamental principles like causality and realism, which go, indeed, beyond viability.
PeterDonis said:
If different theoretical models make exactly the same predictions, so that there is no way to distinguish them experimentally, then the usual scientific criterion that is applied is Occam's razor.
The problem with Occam's razor is that it is not that well-defined as it should be. Say, the spacetime interpretation adds a whole dimension to what actually exists. But somehow the theory which uses only a three-dimensional space, where the past no longer exists and the future not yet exists, is rejected using Occam's razor. IMHO that's complete nonsense, but that's the problem with Occam's razor that one can reinterpret it in such a way that it somehow sounds acceptable.
Then, don't forget that "exactly the same predictions" is a more an urban myth than reality. Say:
EPR realism + SR spacetime interpretation ===> Bell's inequality for spacetime-separated events.
EPR realism + SR pref. frame interpretation =/=> Bell's inequality for spacetime-separated events.
There are other examples. So, there is the Wallstrom objection against some realistic interpretations of QM, which is that the realistic "interpretations" are, in fact, different theories which make different predictions (namely they exclude some QM solutions).
PeterDonis said:
If different theoretical models make different predictions, but the difference is not testable with our current technology and powers of observation, then we are in much the same position as above (i.e., most scientists will use Occam's razor) until our technology or our powers of observation improve to the point where we can test the different predictions.
That's the problem of modern physics. The new devices which have given the SM have reached their limit, and all those speculations beyond the SM are evaluated in a completely unprofessional way. Because the evaluation with the standard methods, namely experiments, gives nothing, nada, nitschewo, nix. And the professional discussion of other ways to evaluate theories has been essentially suppressed as evil "metaphysics" or "philosophy". As a consequence, what is used is bad philosophy. Unprofessional because professional philosophy is already much better than the philosophy used by physicists.
And if we would follow your philosophy, there would be no discussion about philosophy in the future too, and as a consequence, I guess, the philosophy used to evaluate theories beyond the SM would remain as bad and unprofessional as it is today.