Quite correct. If we don't have the prerequisites to use something, then we don't use it, because we can't. It doesn't follow that we do nothing, or that by doing something else that is adapted to the reality of the inference problem that faces us, we need to change academic or professional fields.No, we cannot lower or modify a valid method if we don't have enough material from which to arrive at supportable answers.
We have to accept that we don't know, and that there are some things we'll never know. OK, I've accepted the obvious, now what? There is nothing "lowering" about a resolution to do as much as possible with the materials that are actually available.
What we do instead may resemble some other method, or it may not. Some methods have the property referred to by the adjective "anytime." Anytime methods apply the same operations to whatever amount of resources are available to the user of the method. Bayes rule is an anytime inference method, for example.
Yes, I mean hypotheses are not assumptions.Do you mean that hypotheses are not assumptions? If so, it seems to me that the sort of hypotheses you are talking about are untestable, and if so, merely pointless speculation.
Your follow on: how do you know what is "untestable?" Untested so far, fine; you personally can't figure out a test that's acceptable to you right now, that's fine, too. Untestable? Apart from logic-puzzle hypotheses ("This hypothesis is false," referring to itself, for instance), potential for testing is contingent, and not properly a property of the hypothesis, but rather a relationship among the hypothesis, the world, and the cleverness of the community that might attempt a test.
The language of necessity ("require") has little place in contingent reasoning. You use that language a lot.Explanations (hypotheses) require some verification, substantiation, from the data being investigated or they are mere speculation and not serious inquiry.
Explanations can only be improved by some verification, substantiation, from data. In the meantime, they may well be the subject of serious inquiry (including the effort needed to locate data, which effort may be guided by attending to the more promising of the proposed explanations). The alternative to certainty most certainly is not speculation.
There is a spectrum of justifiable confidence, and the role of data is to move hypotheses along the spectrum. O Happy Day when there is enough data to move one hypothesis to the Promised Land of Godfrey-approval. For most things, though, there is only more or less, and with that, a fair amount of disagreement among persons about which is more and which is less.