Many non-classical logics can be axiomatized by means of Hilbert Systems. Reasoning in Hilbert Systems, however, is extremely inefficient. Most inference methods therefore use the semantics of a logic in one kind or another to get more efficiency. In this paper a combination of Hilbert style and semantic reasoning is proposed. It is particularly tailored for cases where either the semantics of some operators is not known, or it is second-order, or it is just too complicated to handle, or flexibility in experimenting with different versions of a logic is required. First-order predicate logic is used as a meta-logic for combining the Hilbert part with the semantics part. Reasoning is done in a (theory) resolution framework. The basic method is applicable to many different (monotonic prepositional) non-classical logics. It can, however, be improved by treating particular formulae in a special way, as rewrite rules, as theory unification or theory resolution rules, even as recursive calls to a theorem prover. Examples for all these cases are presented in the paper.
[1]
Hans Jürgen Ohlbach,et al.
Semantics-Based Translation Methods for Modal Logics
,
1991,
J. Log. Comput..
[2]
Larry Wos,et al.
Experiments in Automated Deduction with Condensed Detachment
,
1992,
CADE.
[3]
R. Labrecque.
The Correspondence Theory
,
1978
.
[4]
María Manzano,et al.
Extensions of First-Order Logic
,
1996
.
[5]
Dov M. Gabbay,et al.
Towards Automating Duality
,
1994
.
[6]
R. Schmidt.
Optimised modal translation and resolution
,
1997
.
[7]
Andreas Nonnengart.
A resolution-based calculus for temporal logics
,
1995
.
[8]
Dov M. Gabbay,et al.
An Overview of Fibred Semantics and the Combination of Logics
,
1996,
FroCoS.
[9]
Dov M. Gabbay,et al.
Quantifier Elimination in Second-Order Predicate Logic
,
1992,
KR.