What kind of mental state is trust? It seems to have features that can lead one to think that it is a doxastic state (cf. Adler 1994, Baier 1986, Hieronymi 2008, Keren 2014, McMyler 2011) but also features that can lead one to think that it is a non-doxastic state (cf. Baker 1987, Faulkner 2011, Jones 1996, McLeod 2011). This has even lead some philosophers to think that trust is a unique mental state that has both mind-to-world and world-to-mind direction of fit (Holton 1996), or to give up on the idea that there is a univocal analysis of trust to be had (Hardin 2004, Simpson 2012). Here, I propose that ‘trust’ is the name we give to mental states that we would think of as beliefs if belief was to be thought of in ‘pragmatist’ terms (that is, as a state posited primarily to explain agents’ actions) and belief resists ‘pragmatist’ treatment. Only such an account, I argue, can univocally account for all the diverse features of trust. As such, I also propose that the explanation of trust provides us with a case for understanding the limitations of a comprehensively ‘pragmatist’, or ‘Neo-Wittgensteinian’ conception of the mental.