Metaphysical foundations of modal logic (W. V. Quine).

摘要

“Modal logic was conceived in sin: the sin of confusing use and mention.” So quips Quine. The stigma stuck with modal logic for a while. But by the mid-sixties, a whole cluster of mathematically elegant interpretations of modal logic became available. All are natural extensions of the classical Tarskian semantics of predicate logic. By the mid-seventies, Quine's criticisms seemed obsolete. Today, we teach the model theory of modal logic as a matter of course. Quine's “interpretive problem” is just forgotten. The purpose of my dissertation is to revamp Quine's thesis: the very foundations of modal logic—simple propositional modal logic, even before quantifiers are added—are shaky.; Quine holds that the natural interpretation of the operator “necessarily” is in terms of a full-blooded metaphysical notion of real necessity. Such a notion, Quine conjectures, is committed to the “metaphysical jungle of Aristotelian essentialism”. Hence, the intended interpretation is to be rejected. Thus Quine: so much the worse for real necessity.; An alternative comes to mind: a meta-level, formal rather than real, semantic interpretation of the operator of necessity. In the late forties, Carnap developed this line of interpretation which culminates in the aforementioned model theory (“Kripke's possible world semantics”) in the sixties. On this basis a formal interpretation is introduced: it grounds necessity in the notion of model theoretic validity. As pointed by friends of the formal interpretation (Marcus, Parsons and Kaplan), no invidious essentialist claim is verified.; But have Quine's concerns been answered? Yes and no. Technical problems regarding quantification across modal operators, substitutivity, matters of scope, and the basis of cross-world identification have indeed been solved. However, no real interpretation of “necessarily” has been provided. I argue that the seeds of such an interpretation are present in Kripke's philosophical discussion of de re necessities in Naming and Necessity. It is not model theoretic, and it does not reduce necessity to some non-modal notion, just as Quine predicted. Even more in his vein, the real interpretation is (i) metaphysically, grounded indeed in essentialist theses, and (ii) epistemologically, forced to give up Kant's ideal that all necessities are known a priori.