On the ternary relation and conditionality (Q452358)

The authors of this paper provide three philosophical explanations of Routley and Meyer's ternary relational semantics for relevant logics regarding the concept of conditionality. They consider three ways of understanding conditionality (those that can be found in the frameworks of modal, intuitionistic and conditional logics) and show that the Routley-Meyer ternary relation does make good sense in any one of them. In modal logics the conditional is thought of as asserting an absence of counterexamples. In the interpretation of the ternary relation offered for this concept of conditionality, the second and third arguments go together (\( Rx\left\langle yz\right\rangle \)). It is shown that, as in the Kripke-style semantics, where a binary relation is employed, the absence-of-counterexample pattern is actually working in the ternary relational semantics too. It is just needed that this notion be extended to pair points, which can be understood from the point of view of the theory of situated inference. Then, it can be said that a conditional \(A\rightarrow B\) is true at the point \(x\) (\(x\vDash _{1}A\rightarrow B\)) iff there is no pair point \(\left\langle yz\right\rangle \) such that \(\left\langle yz\right\rangle \) realizes the informational links of \(x\) (\(Rxyz\)) and \( \left\langle yz\right\rangle \vDash _{1}A\) (\(y\vDash A\)) but \(\left\langle yz\right\rangle \vDash _{0}B\) (\(z\nvDash B\)). That is, a conditional is true at \ point \(x\) iff there is no \(x\)-accessible counterexample. In intuitionistic logic, a conditional \(A\rightarrow B\) is a construction which takes us from a proof of \(A\) to a proof of \(B\). Then, a conditional can be understood as a function that, when applied to the antecedent, gives us the consequent. According to the authors, for a conditional \(A\rightarrow B\) to hold in a world, it is required that the function it expresses corresponds to an inference which is supported by that world. They now group together the first and second terms of the ternary relation (\(R\left\langle xy\right\rangle z\)) and argue that it is possible to think of the compound \( x;y\) (\(x\) and \(y\) are worlds) as the set of propositions resulting from \(y\) by applying to it all the inferences provided by \(x\). In the frame of relevant logics, this actually behaves as a functional application of \(x\) to \(y\). For worlds \(x\), \(y\), \(z\), \(Rxyz\) is then defined so that it means that at least everything warranted by \(x;y\) holds at \(z\). So, \(z\) is at least as strong as \(x\) and \(y\) put together. In conditional logics the conditional is conceived as describing some sort of relativized necessity. The first and third terms of the ternary relation are now grouped together (\(Rx\left\rangle y\right\langle z\)). It is shown that when the models are restricted for conditionals, the conditional-logic clause using binary relations indexed by a set of points can be replaced by the Routley-Meyer clause using a ternary relation definable from the indexed binary relations. Finally, starting from the idea that the ternary relation is a relation of relative relative possibility (\(Rxyz\) just when \(z\) is possible relative to \( y\), relative to \(x\)), a general conception of conditionality unifying the three conceptions considered in the article is discussed.