[bubblyworld]

You know this already, I think, but a pretty widely-used metric of utility in science is whether or not:

a) a theory makes predictions that match the results of a set of prescribed experiments

b) those results are reproducible by unrelated third parties following the same experimental guidelines

For some theories like EM we're at the point where the body of known experiments and predictions is so vast, and so many edge-cases have been probed, that we basically expect the results to generalise to the known universe.

But yeah, there's a long tail of science with a much weaker claim to "truth" or utility than this. I agree with you (I think? might have misunderstood you) that it starts to become more of a cultural or social phenomenon in these cases.

[mistermann]

> it starts to become more of a cultural or social phenomenon in these cases.

Very much part of my concern. And within that there is the phenomenon whereby the harm science causes (say, global warming") somehow doesn't count. Only the end users of the things science brought into existence bear any burden of responsibility.