Our aim is to present here a proposal of counterfactual truth value evaluation, based on a notion of relevance, which we compare to classical counterfactual evaluation. Informally we demand that the truth value of the counterfactual (or conditional) implication p > q should depend on the existence of a deduction of q that requires the truth of p, in some appropriate nearest world. We believe that relevance of counterfactuals is important in A.I. and illustrate that with examples.
Unable to display preview. Download preview PDF.
- 1.Lewis, D., Counterfactuals, Harvard University Press, Cambridge, MA. Revised edition, 1986.Google Scholar
- 2.Ginsberg, Matthew L., Counterfactuals, Artificial Intelligence 30 (1986), pp. 35–79.Google Scholar
- 3.Ginsberg, Matthew L., Smith, David E., Reasoning about Action I:A Possible Worlds Approach Artificial Intelligence 35, pp. 165–195, 1988.Google Scholar
- 4.Nute, Donald, Conditional Logic, in Handbook of Philosophical Logic Vol. 2, pp. 387–439, (eds. D. Guenthner and D. Gabbay), D. Reidel, 1986.Google Scholar
- 5.Gärdenfors, P., Conditionals and changes of belief, Acta Philos. Fennica 30 (1978), pp. 381–404.Google Scholar
- 6.Anderson and Belnap, Entailment: The Logic of Relevance and Necessity, Princeton University Press, 1975.Google Scholar
- 7.Read, S., Relevant Logic, Basil Blackwell, 1988.Google Scholar
- 8.Routley, F.R. et al. Relevant Logics and their rivals, Atascadero, California, 1982.Google Scholar
- 9.Dunn, J.M., Relevant Logic and Entailment, in Handbook of Philosophical Logic, Vol 3, pp. 177–224, (eds. D. Guenthner and D. Gabbay), D. Reidel, 1986.Google Scholar