Martina Zirattu (Università degli Studi di Torino)
Abstract: When working on a multiple-premise and multiple-conclusion setting, in most logical systems the comma on the premise side matches the behaviour of the object language conjunction, whilst on the conclusion side it matches that of the disjunction.
However, this is not the case for Paraconsistent Weak Kleene (PWK) nor for its paracomplete dual, known simply as Weak Kleene logic K3w (Ciuni and Carrara, “Characterizing Logical Consequence in Paraconsistent Weak Kleene”, New Directions in Logic and the Philosophy of Science, 2016). Both logical systems, which I will call WK logics (WK) for short, have a three-valued semantics in which the non-classical truth-value behaves in an “infectious” way, i.e. as an absorbing element for all the operations (Kleene, Introduction to Metamathematics, 1952). It’s easy to check that Simplification (ϕ∧ψ ⊨ ϕ(ψ)) is not valid in PWK, whereas ϕ, ψ ⊨PWK ϕ(ψ). And dually, Addition (ϕ ⊨ ϕ∨ψ) isn’t valid in K3w, while ϕ(ψ) ⊨K3w ϕ, ψ. This fact shows that the metalinguistic comma alluded earlier doesn’t behave as the expected binary associative connective featured in the language (Bonzio and Baldi, “Undefinability of Standard Sequent Calculi for Paraconsistent Three-valued Logics”, 2017). This suggests that there are two distinct ways of conjoining (disjoining) formulae in WK logics: one can either use conjunction (disjunction) or the comma to gather premises (conclusions) together in PWK (K3w), resulting in a different set of validities in each case.
My objective here (developed in joint work with Agustina Borzi) is to explore a way of restoring the expected correspondence between metalanguage and object language, namely between commas and conjunction/disjunction in WK logics. With this purpose I define two logical systems, which I will call uPWK and uK3w (uWK logics, which stands for ‘uniform WK logics’), that turn out to be substructural: Left Weakening fails in uPWK, as does Right Weakening in uK3w. To study these logics, I provide sequent calculi that are sound and complete and characterize their consequence relation with respect to WK logics and variable inclusion.
The results presented in this work open some interesting new questions. In fact, the same motivation that leads me to the study of uWK logics as solutions to the puzzling mismatch between metalanguage and object language in WK logics, points to a generalization of this strategy for other logics that have the same problem.
