Semantics for relevance logic with identity (Q1194104)

Relevant logics have generated a number of philosophically and technically interesting areas of research. Quantified relevant logics with identity has been one of the most important of these, tying in as it does with a number of traditional and contemporary philosophical concerns. (See, for example, \textit{J. M. Dunn} [J. Philos. Logic 16, 347- 381 (1987; Zbl 0638.03003)], \textit{J. B. Freeman} [Directions in relevant logic (\textit{J. Norman} and \textit{R. Sylvan} (eds.)), 305-316 (1989; Zbl 0731.03014)], \textit{J. M. Dunn} [Knowledge representation and defeasible reasoning (\textit{H. E. Kyburg jun.} et al. (eds.)), 89-95 (1990; Zbl 0743.68024)], \textit{P. Kremer} [J. Philos. Logic 18, 349-382 (1989; Zbl 0706.03020)], \textit{J. M. Dunn} [Philos. Stud. 60, 347-381 (1990)], and \textit{J. M. Dunn} [Truth or consequences (\textit{J. M. Dunn} and \textit{A. Gupta} (eds.)), 77-95 (Kluwer, Dordrecht) (1990)]. This paper is a very worthwhile addition to that work. The author provides models for a number of different systems of quantified relevant logic with identity, various ones having different principles of transitivity of identity and/or of substitution of identicals. The models are all based upon the semantics of quantified relevant logic of \textit{K. Fine} [J. Philos. Logic 17, 27-59 (1988; Zbl 0646.03013)]. (The domain of quantification varies for different worlds/setups/theories. A universal quantification of an open formula is true at a theory \(T\) iff (roughly) \(T\) forces that open formula to be true of an arbitrary object not in the domain of \(T\).) The two principles of transitivity considered in the paper are: \[ \begin{aligned} \text{NESTED TRAN: } \quad & x=y \text{ implies } (y=z \text{ implies } x=z), \quad \text{and } \\ \text{NON-NESTED TRAN: } \quad & (x=y \text{ and } y=z) \text{ implies } x=z. \end{aligned} \] The two principles of substitution are: \[ \begin{aligned} \text{STANDARD SUB: } \quad & (x=y \text{ and } Ax) \text{ implies } Ay,\text{ and } \\ \text{WEAK SUB: } \quad & (x=y \text{ and } Ax) \text{ implies } Ay,\text{ provided } A \text{ is of degree } 0.\end{aligned} \] (A formula is of degree 0 iff it contains no occurrences of the implication connective.) The author gives a brief, but clear and balanced discussion of the relative merits and demerits of these principles and recommends that the choice between them be postponed until further investigation reveals the full extent in the differences of deductive power resulting from the different choices -- which will also vary according to the strength of the underlying relevant logic.