It is reasonable to propose a new task for mathematical philosophy: not to seek indubitable truth but to give an account of mathematical knowledge as it really is – fallible, corrigible, tentative and evolving, as is every other kind of human knowledge.
(Hersh, 1979, page 43)