Abstract

In §1 of this paper I will present the two criteria, which I will call respectively the coextensionality and the recognitional criteria of synonymy. An established tendency in the literature is to ascribe to Frege only the recognitional criterion, discounting the coextensionality criterion as inconsistent with some of his other views. My aim in the paper will be to contribute to the reversal of this tendency. First, I wish to show that the recognitional criterion is flawed in a way that makes it unacceptable as a criterion of synonymy and that Frege was aware of this. I will argue, second, that in Frege’s writings a modified version of the coextensionality criterion finds better support than the recognitional criterion. And I will suggest, third, that although there is an inconsistency between some of Frege’s views and the coextensionality criterion, this inconsistency can be explained as a prominent feature of his philosophy, a tension inherent in his goals and views. Finally, in a brief appendix, I will focus on the way in which Frege applied the two criteria in the practice of linguistic analysis, and I will argue that, contrary to the accepted wisdom, within this practice the two criteria are compatible with each other.