So, there's a point about philosophy of science to be made here. If you test one bronze alloy for antigravity, and then test a slightly different bronze alloy for antigravity, and so on, finding all of them to fall when dropped, how can you know when it's reasonable to stop testing bronze alloys for antigravity? Presumably, you should eventually say "this idea has already been tested," and then stop. However there will always be a new ratio of metals to try: the only way to say "this has been tested" or "this hasn't" is to establish enough of a theory to make predictions about the effect of gravity on every bronze alloy, and then trust it when it says it will not become negative (at least, trust it enough to let it rule out potential experiments from grant approval.) When we say, "a new arrangement of photons will still conserve momentum," we are performing exactly the kind of interpolation from other experiments involving photons in cavities that, previously, stopped us from making more and more bronze bars in the search for antigravity.