Kripke, Saul Aaron (1940–)

2. Identity, proper names and possible worlds

Kripke began working on these topics in the early 1960s and developed a number of positions that initially met with surprise and suspicion, but eventually became widely accepted. The most comprehensive presentation may be found in Naming and Necessity (1980), which is an edited transcript of three lectures given at Princeton in January 1970, together with a preface and addenda.

Kripke maintains that identity is a relation that holds between each thing and itself, never holds between any two things, and always holds (or fails to hold) of necessity (see Identity). The ‘necessity of identity’ is a theorem of quantified modal logic (with identity), but some philosophers nevertheless thought certain examples cast doubt upon the doctrine (see Modal logic). For example, the identity of the inventor of bifocals with the first Postmaster General can easily seem contingent. But Kripke observed that what is contingent here is that one person attained both of these (separately contingent) distinctions, and that this contingency in no way undermines the necessity of that person’s identity with himself – Benjamin Franklin would have been identical with himself even if he had never been Postmaster General or worked with lenses.

This last point reflects a fundamental Kripkean doctrine about the behaviour of ordinary proper names such as ‘Franklin’. It is that when we utter a sentence such as ‘Franklin might never have became Postmaster General’, we are describing a way in which things might have gone differently for the very person who, in actual fact, was the first Postmaster General. We are describing a ‘counterfactual situation’ or ‘possible world’ involving Franklin himself, but in which he is not Postmaster General.

Kripke does not offer a detailed ontological account of possible worlds (though he does explicitly reject the ‘counterpart’ account favoured by David Lewis (1968)). He does not see worlds as offering a useful analysis of our modal notions, but rather as reflecting pre-existing modal intuitions, and thinks that for most philosophical purposes we need no precise notion of possible worlds. It is normally sufficient to think in an everyday way of counterfactual situations or alternative possibilities, and this can help relieve any worry that we are trading on a technical notion of possible worlds without first defining it. Now invoking the terminology of worlds, the idea of the previous paragraph generalizes to the claim that an ordinary proper name always designates the same thing ‘in other possible worlds’ that it designates in the actual world. Kripke introduced the term ‘rigid designator’ for singular terms that behave in this way, allowing his thesis to be compactly expressed as the claim that proper names are rigid designators. But this rather technical-sounding formulation is intended as little more than a generalization of the hardly controversial idea that ‘Franklin invented bifocals’ and ‘Franklin might not have invented bifocals’ both concern the same person.

Now consider the sentence ‘Cicero is Tully’, which seemingly asserts a true identity and which involves two proper names. If these names are rigid designators, then each designates the same entity in every possible world – the entity that each designates in the actual world. But in the actual world both names designate the same person. It follows that they both designate this person in every possible world, and so it must be a necessary truth that Cicero is Tully. But it has often been thought that because such a fact is not knowable a priori, but rather requires empirical investigation, it could not be necessary. Kripke sees this as a fundamental error, stemming from a confusion of the epistemic notion of a prioricity with the metaphysical notion of necessity. For Kripke, ‘Cicero is Tully’ expresses an a posteriori, necessary truth (see A posteriori; A priori).

The idea that ‘Cicero is Tully’ is merely contingent is also fostered by theories of proper names in the tradition of Frege and Russell, according to which the referent of a name is determined by a definite description (or, in more recent versions, a ‘cluster’ of descriptions) associated with the name. Such theories often hold that the description gives the meaning (or sense) of the name. Because the description associated with ‘Cicero’ might be utterly different from the one associated with ‘Tully’, and because these descriptions might have distinct referents in some other world, it would be natural to think that ‘Cicero is Tully’ is contingent in much the same way that ‘The first Postmaster General is the inventor of bifocals’ is contingent. But Kripke argued that such theories of proper names are incorrect. His view is more in the spirit of J.S. Mill’s and Ruth Barcan Marcus’, according to which names are mere ‘tags’ which lack senses and serve only to pick out their referents. Kripke held that a name is explicitly linked to its referent when it is first introduced, and that subsequent uses of the name refer to that entity because they trace backwards through an appropriate causal chain of uses to that initial introduction of the name. Subsequent uses do not achieve their reference to that entity as a result of its satisfying any specific descriptions associated with the name.