Theories are built by engaging in the world using imagination, tool-making, and the like. We first suppose that the universe is something like a glass sphere -

Theories are built by engaging in the world using imagination, tool-making, and the like. We first suppose that the universe is something like a glass sphere -- because we've created that. And if it is, then we derive some consequences -- if those line up, we proceed with that view until a better one comes along. Eventually after the glass sphere view is understood, we either derive contradictions with observation; or we end up unable to derive novel consequences. Here observation is essentially singular, and indeed, the rarer reason we reject a theory. We mostly reject scientific theories because of their explanatory limits, not disagreement with observation. (Rarely can we observe enough for observation even to matter.) In the case of the system of spheres, we built spinning devices on that basis and this motion -- along with the hydraulics and kinematic devices of the time -- was part of the development of an independent notion of force. With some imagination, you can start to peel away material from our creations and see certain abstract causal patterns (and the like) and you then get to, eg., the universal law of gravity. Absent this process we do not have any explanatory ideas, we cannot explain observations -- hence it takes thousands of years to get anywhere. Applying 'statistics' to do the data to arrive at statistical models is pseudoscience; it doesnt give you any account of anything. 'Statistics' doesnt own 'testing', nor does 'science' own experiment -- theologians had their experiments (prayer, say) and scientists collected data without statistical methods. What I am talking about is the 20th C. discipline of statistics, as a novel apparent 'core' to science -- this is ahistorical, and largely only true of pseudoscientific disciplines. As Ernest Rutherford said, “if you need statistics to do science, then it's not science.” It is in Rutherfod's sense of stats and of science that I speak. Not some bizarre historical back-projection by which when Aristotle analysed cases of sea creatures, "Really", he was engaged in stats. By this light i can just claim "Testing" is as owned by science. And so of course science requires testing -- it was the *scientific method* as developed by bacon and others that created the very conditions for "statistical methods" to derive from these