Many-valued logics and Suszko's thesis revisited (Q1577350)

The paper starts with an overview of some discussions around R. Suszko's thesis which states that any logic with a structural consequence operator conforming to Tarski's standard conditions (i.e. with a Polish consequence operator) is logically two-valued. In view of \textit{G. Malinowski}'s counterexample [``Inferential many-valuedness'', in: J. Woleński (ed.), Philosophical Logic in Poland, Kluwer, Dordrecht, Synthese Lib. 228, 75-84 (1994)] which consists of a many-valued logic that could not be reduced to a two-valued logic following Suszko's method, this paper aims to give a clearer picture of the matter. Startingrom a very abstract concept of logical structures which encompasses all known lologics as introduced by \textit{J.-Y. Béziau} [``Recherches sur la logique universelle'', PhD thesis, Université de Paris VII (1995; Zbl 0936.03010)], necessary and sufficient conditions are given for such a logical structure to have a complete set of bivaluations.