Parallel Theories and Routine Revision in First-Order Logic

From MaRDI portal





The author recalls that in scientific investigation we not only gain beliefs but also lose and revise them. He describes this as a process of revising premises, and suggests that it may be represented using ordinary deductive logic, rather than by inventing some special non-monotonic logic. His representation consists simply of taking addition and elimination of premises in directly set-theoretical terms, i.e. as \(A\cup \{p\}\) and A-\(\{\) \(p\}\), where p is the premise to be added or eliminated. The only mathematical observation made is the perfectly trivial one that if \(A\cup B\) has n elements, then the sets that can be obtained by adding elements of B to A, and/or deleting elements of A, are just the \(2^ n\) subsets of \(A\cup B.\) Reviewer's comments: (1) The central message of the paper seems to be the methodological one, that we do not need recourse to non-monotonic logics to conceptualize the revision of sets of premises. This is perfectly true, but beside the point. What non-monotonic logic seeks to study is not just the revision of premise sets, but also the reductions in the conclusions that we feel entitled to draw, under some forms of inference, upon an increase in our premises. (2) Even on the level of the revision of premise sets under ordinary deductive logic, the author does not consider the implications of the fact that a premise set A may contain redundancies. In such a case, the process of eliminating p from A cannot reasonably be understood as simple set-theoretic removal, but rather as generation of some subset of A that fails to imply p. This gives rise to quite complex considerations, and to formal relationships that are particularly interesting when A is taken to be itself a theory, i.e. as closed under a suitable consequence operation. Some of these are studied in, for example, \textit{C. E. Alchourrón}, and \textit{P. Gärdenfors} and the reviewer, ``On the logic of theory change: partial meet contraction and revision functions [J. Symb. Logic 50, 510-530 (1985; Zbl 0578.03011)].












This page was built for publication: Parallel Theories and Routine Revision in First-Order Logic

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3773842)