Core Logic
Neil W. Tennant · Oxford University Press eBooks · 2017
Core Logic has unusual philosophical, proof-theoretic, metalogical, computational, and revision-theoretic virtues. It is an elegant kernel lying deep within Classical Logic, a canon for constructive and relevant deduction furnishing faithful formalizations of informal constructive mathematical proofs. Its classicized extension provides likewise for non-constructive mathematical reasoning. Confining one’s search to core proofs affords automated reasoners great gains in efficiency. All logico-semantical paradoxes involve only core reasoning. Core proofs are in normal form, and relevant in a highly exigent ‘vocabulary-sharing’ sense never attained before. Essential advances on the traditional Gentzenian treatment are that core natural deductions are isomorphic to their corresponding sequent proofs, and make do without the structural rules of Cut and Thinning. This ensures relevance of premises to conclusions of proofs, without loss of logical completeness. Every core proof converts any verifications of its premises into a verification of its conclusion. Core Logic makes one reassess the dogma of ‘unrestricted’ transitivity of deduction, because any core ‘restriction’ of transitivity ensures a more than compensatory payoff of epistemic gain: A core proof of A from X and one of B from {A}∪Y effectively determine a proof of B or of absurdity from some subset of X∪Y. The primitive introduction and elimination rules governing the logical operators in Core Logic are subtly different from Gentzen’s. They are obtained by smoothly extrapolating protean rules for determining truth values of sentences under interpretations. Core rules are inviolable: One needs all of them in order to revise beliefs rationally in light of new evidence.