Constraints for input/output logics

From MaRDI portal





This paper continues a project of the authors [cf. ibid. 29, 383-408 (2000; Zbl 0964.03002)] to investigate what they call `input/output logics' that define operations resembling inference but in which inputs need not be included among outputs and outputs need not be reusable as inputs. These have application particularly in deontic logic and the theory of conditional norms and especially contrary-to-duty obligations, and they are connected to some forms of nonmonotonic logic. The present work investigates what happens when an input/output logic is constrained so that it does not produce excessive outputs, in particular, when it is constrained so that its output is consistent with its input. The mechanism is to limit the generators of such a logic to the maximal subsets they yield such consistent output. This is reminiscent of techniques in the theory of belief change and nonmonotonic inference. And indeed, when inputs can emerge as outputs, the `maxichoice revision' of Alchourrón and Makinson and the default logic of Poole come out as special cases. There is also a close connection to Reiter's default logic. The focus here, however, is on the general case and the effect of consistency constraints on a variety of output functions. These are examined from both a semantical point of view (broadly construed) and in terms of derivations.




Cited in
(42)








This page was built for publication: Constraints for input/output logics

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5935791)