The problem with rationality (ie, science) as a belief is thus not logical reasoning in itself, but that classification (ie, conceptualization) is ultimately paradoxically contradictory (as Bertrand Russell demonstrated).
The problem with this problem for belief in rationality (ie, science) is that one entrance to logical reasoning (ie, assuming that kinds aren’t real) leads to ambiguity, whereas the orthogonal entrance to logical reasoning (ie, assuming that kinds are real) leads to paradoxical contradiction (which thus is inconsistent per definition by being contradictory), and thus that consistent rational thinking leads to ambiguity, because it means that belief in rationality (ie, science) in practice is belief in ambiguity. Rational thinking (ie, science) can in practice not come closer to reality than comprehending it as an ambiguity, that is, accepting that reality can be equally truthfully described by several contradictory models.
So, the problem for rational thinking (ie, science) as a belief is that it requires belief in contradictory models of reality. Better is to understand that rational thinking (ie, science) is just a tool to handle reality. It can actually accomplish things that appears like magic (in quantum mechanics), but it can never serve as a belief (contrary to what cladists and particle phycisists claim). My recommendation to rationalists is thus to explore the possibilities for rationalism (ie, science) and avoid all discussions about belief. Rationalism (ie, science) can’t prove nor falsify the statement that there is a God, but the mere existence of rational thinking (ie, science) means that there isn’t. This fact is thus not a conclusion, but a consistent assumption.