Holistic Theories of Content and Instability

In this paper, I will defend two methodological theses, one negative and one positive, about how to develop a holistic theory of content for mental representations that avoids a problem peculiar to holistic theories, viz., the problem of content instability. The relevant debate between holists and anti-holists has focused on whether this problem provides an in principle barrier to developing a plausible holistic theory. On this front, the holists have won; defenders of holistic theories have convincingly argued that the anti-holists do not have a cogent argument from the problem of content instability to the impossibility of developing a plausible holistic theory. However, beyond this, little has been said about how to develop a holistic theory that avoids the problematic consequences of content instability; all that has been established is that it appears to be, in principle, possible to do so. This paper should contribute to making progress in this area. The two theses I will defend are about how to generate useful constraints on holistic theories so that they avoid content instability. The negative thesis of this paper is that the strategy of generating constraints suggested by the holists' response to anti-holist arguments, viz., appealing to properties of theories' determination functions, is a non-starter. The positive thesis of this paper is that the best way to develop useful stability constraints is to appeal to the explanatory role(s) that representations play in cognitive science theories.