Abstract

According to the Quine-Putnam indispensability argument (QPIA), we should be realists about mathematics because mathematics is indispensable to science. QPIA’s reasoning can be understood in two ways. Under the confirmational analysis, QPIA argues that mathematics is confirmed as part of our best scientific theories. Under the pragmatic analysis, QPIA argues that our scientific practices implicitly assume the truth of mathematics. The usual reasons given in favour of the pragmatic analysis are that it affords advantages to proponents of QPIA by avoiding some dependencies of the confirmational analysis. This paper argues that these reasons are indecisive, because the pragmatic analysis introduces some dependencies of its own. Nevertheless, the paper also argues, there are other reasons to prefer the pragmatic analysis. QPIA is an instance of a wider class of realist arguments—indispensability arguments—and the pragmatic analysis yields a preferable overall understanding of indispensability arguments than does the confirmational analysis.