Books on Amazon
Rationality/Bennett/Millikan: it seems as if he should not choose "abbreviations" as a rational human. That is, one must take into account not only positive evidence, but also negative ones.
General/formal: e.g. assuming, John believes
"Usual": if A then B and also:
"Not-(usual: if A-and-not-C, then B)"
Rational: would then follow that John would have to believe
A) "usual: if A then C" and
B) if A and C, then B. Then there are the following possible cases.
1. The only evidence for C comes from that John knows that usually, if A then C. Then he should simply pass from A to B.
2. John has independent ways to believe C due to evidence. And he encounters A, while he already has evidence for not-C.
Then, rationally, he should also believe that not-C and not conclude from A to B.
3. John has independent evidence according to which he could know C, but this time he does not know beforehand whether C.
Question: then, in order to be rational, must he first check whether C?
Millikan: Let's suppose he needs to do it.
Problem: if this again depends solely on the fact that he believes:
"Usually if D, then C", etc.
Rationality/Millikan: Problem: The more knowledge one acquires, the more he has to strive to be rational at all. Would it not be better for him to refrain from the whole checking process?
R. G. Millikan
Language, Thought, and Other Biological Categories: New Foundations for Realism Cambridge 1987