Intuitionism and logical revision.

The topic of this thesis is logical revision: should we revise the canons of classical reasoning in favour of a weaker logic, such as intuitionistic logic? In the first part of the thesis, I consider two metaphysical arguments against the classical Law of Excluded Middle-arguments whose main premise is the metaphysical claim that truth is knowable. I argue that the first argument, the Basic Revisionary Argument, validates a parallel argument for a conclusion that is unwelcome to classicists and intuitionists alike: that the dual of the Law of Excluded Middle, the Law of Non-Contradiction, is either unknown, or both known and not known to be true. As for the second argument, the Paradox of Knowability, I offer new reasons for thinking that adopting intuitionistic logic does not go to the heart of the matter. In the second part of the thesis, I motivate an inferentialist framework for assessing competing logics-one on which the meaning of the logical vocabulary is determined by the rules for its correct use. I defend the inferentialist account of understanding from the contention that it is inadequate in principle, and I offer reasons for thinking that the inferentialist approach to logic can help model theorists and proof-theorists alike justify their logical choices. I then scrutinize the main meaning-theoretic principles on which the inferentialist approach to logic rests: the requirements of harmony and separability. I show that these principles are motivated by the assumption that inference rules are complete, and that the kind of completeness that is necessary for imposing separability is strictly stronger than the completeness needed for requiring harmony. This allows me to reconcile the inferentialist assumption that inference rules are complete with the inherent incompleteness of higher-order logics-an apparent tension that has sometimes been thought to undermine the entire inferentialist project. I finally turn to the question whether the inferentialist framework is inhospitable in principle to classical logical principles. I compare three different regimentations of classical logic: two old, the multiple-conclusions and the bilateralist ones, and one new. Each of them satisfies the requirements of harmony and separability, but each of them also invokes structural principles that are not accepted by the intuitionist logician. I offer reasons for dismissing multiple-conclusions and bilateralist formalizations of logic, and I argue that we can nevertheless be in harmony with classical logic, if we are prepared to adopt classical rules for disjunction, and if we are willing to treat absurdity as a logical punctuation sign.