On making a choice between hypotheses

by ayushkhaitan3437

At 15:33, Peter Millican says “How can any criterion of reliable knowledge be chosen, unless we already have some reliable criterion for making that choice?”

What does this actually mean? Say I have two hypotheses- A and B. One of them is true, whilst the other is false. But I don’t know which is which. How should I choose? Maybe I should choose the one which corroborates the existing body of knowledge to a greater extent. Or I could choose the one which corroborates my personal experiences to a greater extent (my personal experiences and the existing body of knowledge would certainly overlap, but the former may or may not be a subset of the latter). Clearly, we have a problem here. On what criterion should I base my selection. This is precisely the question being discussed here.

This, I feel brings us to an even more fundamental question: How do we make choices? Let us suppose I have to choose between two identical bars, based on just sight. One is a gold bar, and the other an imitation. As they’re exactly identical, I have no reason to choose one over the other. This of course is true only after removing biases like “I should choose the bar on the left as the left is my lucky side; Today is the 13th, and I had an accident on the 13th while turning to the right, hence I should choose left, etc”. Hence, the criterion for selection, without the addition of any further knowledge, would be arbitrary.

However, when making choices between two non-identical hypotheses, we have a definite bias. I might say my textbook supports A. Or that my personal experiences support B. Now based on whether I trust my textbook more or my personal experiences and reasoning, I shall make a choice accordingly. The criterion for making a choice in such situations is hence based on evaluating biases, of one form or another, and deciding which choice I’m naturally more inclined or biased towards. A lot of this evaluation takes place implicitly in the brain in our daily lives.

This of course is assuming the choice-making process is completely rational, which it is not in humans. I could, for example, arbitrarily choose the option which the weighted sum of my biases do not support.