ABSTRACT
The focal question of this essay is what if any role logical inference should be taken to play in mathematical reasoning. Special attention is given to the idea of Brouwer and Poincaré that, by its very topic-neutral character, logical inference is unsuitable for use in mathematical reasoning, and the epistemology standing behind this idea is partially developed. According to this epistemology, intuition is needed not only to supply axioms, but also to power inference.
