Patterns, order, and proportion Internet Encyclopedia of Philosophy, Nagarjuna
"Let us take the claim that something can be proven to be true on the basis of other facts known to be true. Suppose, to use a favorite example from the Logician Gautama, I want to know how much an object weighs. I put it on a scale to measure its weight. The scale gives me a result, and for a moment that satisfies me; I can rely on the measurement because scales can measure weight. But hold on, Nagarjuna flags, your reliance on the trustworthiness of the scale is itself an assumption, not a piece of knowledge. Shouldn’t the scale be tested too? I measure the object on a second scale to test the accuracy of the first scale, and the measurement agrees with the first scale. But how can I just assume, once again, that the second scale is accurate? Both scales might be wrong. And the exercise goes on, there is nothing in principle which would justify me in assuming that any one test I use to verify a piece of knowledge is itself reliable beyond doubt. So, Nagarjuna concludes, the supposition that something can be proven through reference to some other putative fact runs into the problem that the series of proofs will never reach an end, and leaves us with an infinite regress.... [T]he Logician might, and in fact historically did, try an alternative theory of mutual corroboration. We may not know for certain that a block of stone weighs too much to fit into a temple I am building, and we may not be certain that the scale being used to measure the stones is one hundred percent accurate, but if as a result of testing the stones with the scale I put the stones in the building and find that they work well,I have reason to rely on the knowledge I gain through the mutual corroborations of measurement and practical success. This process, for Nagarjuna, however, should not pass for an epistemologist who claims to be as strict as the Brahminical Logicians. In fact, this process should not even be considered mutual corroboration; it is actually circular. I assume stones have a certain measurable mass, so I design an instrument to confirm my assumption, and I assume scales measure weight so I assess objects by them, but in terms of strict logic, I am only assuming that this corroborative process proves my suppositions, but it in fact does nothing more than feed my preconceived assumptions rather than give me information about the nature of objects."
Nagarjuna was followed in the West by Hume, who said the same thing. It seems one has to allow some subjectivism in in order to have a rational understanding of the world.