Thursday, August 29, 2013

The Return of Metaphysics into Analytic Philosophy

This post is based on Chapters 6 and 7 of Stephen P. Schwartz’s A Brief History of Analytic Philosophy: From Russell to Rawls (2012).

Schwartz calls Saul Kripke’s book Naming and Necessity (1980 [1972]) the “apotheosis of analytic philosophy,” because of the manner in which the book has founded a new analytic metaphysics and new insights into epistemology (Schwartz 2012: 241), even if Kripke often drew on the work of others.

The fundamentals of this new metaphysics can be summarised as follows:
(1) the modal logic of possible worlds;

(2) the new “causal” or “direct” theory of reference, applying to proper names, definite descriptions and natural kind terms;

(3) the new epistemological categories of (1) necessary a posteriori and (2) contingent a priori truth.
Despite objections from Quine (who was opposed to the existence of intensional objects), this rebirth of metaphysics began in the 1960s with developments in quantified modal logic, the logic of necessity and possibility (Schwartz 2012: 204–210). Saul Kripke played a large role in clarifying modal logic (Schwartz 2012: 212).

In the new modal logic, a necessarily true proposition p (or, in symbolic form, □P) is true in all possible worlds.

A possibly true proposition p (or, in symbolic form, ◇P) is true in at least one possible world. A thing would have a property or properties essential to it if and only if it has that property or properties in every possible world where it exists, but would have a property accidentally (or contingently), if and only if it has that property in one possible world but not another (Schwartz 2012: 214–215). It is important to note that these are metaphysical concepts of necessity, contingency and possibility, not merely semantic ones (Schwartz 2012: 215).

A “possible world” could be understood as
(1) a purely imaginary and non-real linguistic entity describing how things could have been;

(2) a real but abstract possible world (in the way numbers are often held to be real but not concrete), or

(3) a real and actual possible world but different from ours (as, for example, imagined in the multiverse hypothesis).
David Lewis adopted the view called modal realism: that infinite logically possible worlds are real and that individuals and things in those worlds exist just as concretely as actual things in our world, even though no universe is causally connected to others (Schwartz 2012: 218).

The alternative to modal realism is modal actualism: the view that only our universe actually exists, and that possible worlds are just abstract entities inside ours (Schwartz 2012: 219).

Modal realism raises issues about personal identity in other possible worlds. For example, if Nixon exists in other possible worlds with different life histories, what allows us to identify these other “Nixons” as the same man as the Nixon in our actual world? For David Lewis, there are no strict transworld personal identities, but merely counterparts in each possible world, which resemble each other to some degree (though this just raises the question of what counts as a proper counterpart!).

Others argue that individual human beings presumably have an individual essence, such as (1) being human (in a scientific sense), (2) having the same parents and birth facts, and (3) having the same DNA or genome (Schwartz 2012: 225). (Notably, Kripke, Plantinga and others strangely dismiss the problem of transworld identity as a pseudo-problem.)

Overall, most modern analytic philosophers have rejected David Lewis’s modal realism and his counterpart theory, but the metaphysical aspects of modal logic still remain (Schwartz 2012: 229).

But a further development of metaphysical ideas, via philosophy of language, was done by Saul Kripke, Hilary Putnam, and Keith Donnellan, although they drew on or developed ideas from others, such as Carnap, Ruth Barcan Marcus, Alvin Plantinga, David Lewis and Quine.

Part of this metaphysics was a new theory of reference, developed in works such as Keith Donnellan’s “Reference and Definite Descriptions” (1966), Kripke’s Naming and Necessity (1980 [1972]), and Putnam’s “The Meaning of ‘Meaning’” (1979).

Traditionally, the conjunction of properties used to describe a term was held to be its intension, and this provides the necessary and sufficient condition for deciding what the term refers to in terms of its extension (the alternative theory, associated with the later Wittgenstein, is that terms have a cluster of properties, and referents of terms have a certain number of these properties, without any one being sufficient) (Schwartz 2012: 241).

A proposition defining a thing in terms of its intension has an analytic truth, so that necessary analytic truth has a merely verbal or de dicto necessity, not a metaphysical (or de re) necessity (Schwartz 2012: 241). Therefore, in the traditional theory of sense and reference, the essence of a thing is a mere verbal or linguistic definition of it (Schwartz 2012: 241–242).

Saul Kripke began by questioning this traditional theory of reference with respect to proper names. Russell argued that proper names are really “disguised definite descriptions.” That is, each proper name has a set of descriptions which the referent of the name satisfies.

In place of this, Kripke and Donnellan proposed a new theory.

First, Donnellan argued that definite descriptions are used in two senses: in (1) an attributive sense, and (2) in a referential manner (Schwartz 2012: 243). It is possible to use a definite description in an attributive sense in which it is a subject with a predicate, without the speaker knowing the actual referent of the definite description (Schwartz 2012: 244). By contrast a direct referential use can refer to a thing independently of the descriptions.

Secondly, Kripke also argued that proper names refer independently of attached descriptions and are “rigid designators” which refer to the same individual in every possible world in which that individual exists (Schwartz 2012: 245). In all worlds, the individual to which a proper name refers need only have the properties essential to the individual and not a list of contingent properties given by definite descriptions (Schwartz 2012: 245–246).

The reference of a proper name is not determined by descriptions, according to Kripke, Putnam and Donnellan, but by the existence of many causal or historical chains of reference, passed on from speaker to speaker, going back to the time when an object was first designated with a name (Schwartz 2012: 253).

From these points, Kripke argues that identity statements using alternative names for the same thing have a necessary truth (Schwartz 2012: 246). Thus we can think of statements like the following:
(1) The morning star is the evening star.

(2) Hesperus is Phosphorus.

(3) Tully is Cicero.
Under Frege’s theory of meaning, these only have a contingent truth. But Kripke contends that any rigid designators used in a true identity statement make that statement necessarily true, and it is also necessarily true in all possible worlds where the entity exists.

Furthermore, Kripke insisted on three fundamental epistemological differences, as follows:
(1) the synthetic versus analytic distinction is a semantic difference;

(2) the notions of “necessity” and “contingency” can be understood in a metaphysical/ontological sense, and

(3) the “a priori” versus “a posteriori” distinction is an epistemological one (Schwartz 2012: 247).
That is, “necessarily true” has a sense distinct from purely verbal (or de dicto) necessity, and carries the additional metaphysical sense of “true in all possible worlds” that is itself distinct from the notion of aprioricity (Schwartz 2012: 247).

These epistemological distinctions were a landmark of recent analytic philosophy, according to Schwartz (2012: 247).

Since the epistemological concepts listed above do not coincide, Kripke presented arguments for two additional types of knowledge: the (1) necessary a posteriori truth, and (2) the contingent a priori truth (Schwartz 2012: 247).

For example, the statement “the morning star is the evening star” is necessarily true since both “rigid designators” refer to the planet Venus. Yet this was an empirical discovery, so that epistemologically it is known a posteriori. Therefore “the morning star is the evening star” is a necessary a posteriori truth (Schwartz 2012: 247).

Kripke also applied the rigid designator concept to the common nouns we call “natural kind” terms, such as “water,” “gold” and “tiger” (Schwartz 2012: 247). Gold, for instance, is the element that science has identified as having an atomic number of 79, and natural kind terms are specified by a conjunction of fundamental properties as discovered by science. And, if these properties are true properties of the natural kind, then the natural kind must have them as essential properties as a matter of nature or metaphysical necessity (Schwartz 2012: 248, 251).

Take, as an example, the difference between iron pyrite (fool’s gold) and real gold. The former has the superficial properties of gold (or many of the same concepts as that of gold in its intension), but nevertheless is not gold because of its essential chemical difference. Gold has as its natural essence the property of being the element with the atomic number of 79, and this is metaphysically necessary of gold in that gold must be like this in any possible world.

Kripke also uses the causal or historical theory of reference to explain the origin of natural kind names (Schwartz 2012: 253). The name “water” (or its equivalent in other languages) was used referentially of things familiar as water, but only modern science discovered the fundamental essence of water.

If water is truly H2O, then it is necessarily H2O in all possible worlds, and this is another synthetic necessary a posteriori truth (Schwartz 2012: 249, 251). (Problems arise when this sort of analysis is applied to natural kind types like “tiger” or “horse,” but I will skip this point.)

The upshot of this is that science can and does discover necessary truths (Schwartz 2012: 252), and scientific investigation of the fundamental atomic, chemical or biological properties or structures of some objects yields, or has already yielded, the necessary metaphysical essence of that object, in the sense that, if the object truly has that essence, it will do so in all possible worlds. And these natural essences are independent of linguistic convention, unlike mere analytic truth.

So such is the new analytic metaphysics, though it seems to be a type of metaphysics different from traditional forms.

For example, synthetic a priori knowledge does not appear in it. Nor does it seem to be fundamentally opposed to the natural sciences in the way other metaphysical systems were.

Whether it will continue to be part of future analytic philosophy is an open question.

“Saul Kripke,” Wikipedia

Naming and Necessity,” Wikipedia

“Rigid Designators,” Stanford Encyclopedia of Philosophy, 2006

“Rigid designator,” Wikipedia

“Actualism,” Stanford Encyclopedia of Philosophy, 2000 (rev. 2008)

“Modal Realism,” Wikipedia

“David Lewis,” Wikipedia

“David Lewis,” Stanford Encyclopedia of Philosophy, 2009

“Modal Logic,” Stanford Encyclopedia of Philosophy, 2000 (rev. 2009)

“Reference,” Stanford Encyclopedia of Philosophy, 2003 (rev. 2009)

Ted Parent, “Modal Metaphysics,” Internet Encyclopedia of Philosophy, 2012

Jason S. Baehr, “A Priori and A Posteriori,” Internet Encyclopedia of Philosophy, 2006

“David Malet Armstrong,” Wikipedia

Donnellan, Keith. 1966. “Reference and Definite Descriptions,” Philosophical Review 75: 281–304.

Kripke, Saul A. 1980 [1972]. Naming and Necessity (rev. edn.). Blackwell, Oxford.

Putnam, Hilary. 1979. “The Meaning of ‘Meaning,’” in Hilary Putnam, Mind, Language and Reality. Philosophical Papers. Volume 2, Cambridge University Press, Cambridge.

Schwartz, Stephen P. 2012. A Brief History of Analytic Philosophy: From Russell to Rawls. Wiley-Blackwell, Chichester, UK.


  1. "In the new modal logic, a necessarily true proposition p (or, in symbolic form, □P) is true in all possible worlds."

    If this is correct then the metaphysics are pre-Hegelian.

    1. Out of curiosity Philip Pilkington, have you seen this?

    2. Interesting. I will respond to this, I think. I see that shady Unlearningecon dude is still talking smack about me in various blog comments. Weird guy. He doesn't engage directly, ever.

  2. @Anonymous

    Here's that response if you're interested: