Putnam, Hilary (1926–2016)

2. Realism

While sharing the logical positivists’ respect for natural science (see Logical positivism §2), Putnam criticized fundamental non-realist tenets of their philosophy--the verificationist theory of meaning (which replaces truth with warranted assertability), reductionism (which sees all truths as reducible to truths of physics) and conventionalism (which claims that necessary truths originate in human conventions). Defending realism, the view that there are objective, mind-independent truths, he argued that realism is the only philosophy that does not render the success of science a miracle (see Realism and antirealism; Scientific realism and antirealism §3). His argument from the success of science is presented as structurally similar to hypothetico-deductive arguments within science: realism provides the best explanation for the success of science and the practice of scientists in the same way that atoms, genes, and black holes provide the best explanations for observable phenomena. This argument constitutes an inference to the best explanation, that is, an inference purporting to infer the truth of a claim from the fact that it provides the best explanation of a particular class of phenomena. Putnam's argument for realism, also known as the 'no miracle' argument, enhanced the popularity of realism among philosophers of science, but has also been criticized as purely formal: Realism has no empirical import beyond that of its alternatives, a desideratum scientific hypotheses must meet. Whereas scientific hypothesis such as the existence of electrons or genes have empirical implications that can be tested, no such implications follow from the truth of realism.

Putnam’s criticism of conventionalism is developed in "An Examination of Grünbaum’s Philosophy of Geometry" (1963, reprinted in 1975a) and "The Refutation of Conventionalism" (1975, reprinted in 1975b). On the Reichenbach–Grünbaum conception, the core of the transition from Newtonian to relativistic mechanics is a new definition of the spacetime metric. The definition chosen is a matter of convenience, not truth (see Conventionalism §1). According to Putnam, however, meaning change is only part of the story; theoretical concepts have explanatory import, and must be anchored in a theory that meets both empirical and non-empirical constraints. Since adopting a different metric has observable implications, there is, actually, much less freedom than the conventionalist alleges. Putnam raises a similar objection to Quine’s celebrated thesis regarding the indeterminacy of translation (see Quine, W.V. §9). Quine argued that translation is underdetermined by speakers' pronouncements in the same way that science is underdetermined by observation. Putnam responded that underdetermination, much like conventionalism in science, is an illusion created by considering an unreasonably limited set of constraints. Once we recognize coherence, simplicity, and so on, as constraints on translation or theory-construction, the freedom granted to the scientist and to the translator is much more limited than conventionalism and the indeterminacy thesis allow.

There is, however, one feature of conventionalism that Putnam upholds—the existence of empirically equivalent theories and descriptions. There are, for example, different formulations of classical mechanics—the Newtonian formulation and the Lagrange-Hamilton formulation--which are equivalent in terms of their empirical import. Equivalent theories of this kind are translatable into one another and thus, do not differ in truth value. Preference for one of them on account of its being simpler or more elegant is not a conventional stipulation of truth, but rather preference for a particular way of expressing truth.