One broadly recognised characteristic feature of (a core subset of) the self-attributions constitutive of self-knowledge is that they are ‘immune to error through misidentification’ (hereafter IEM). In the last thirty years, Evans’s notion of “identification-freedom” (Evans 1982) has been central to most classical approaches to IEM. In the Evansian picture, it is not clear, however, whether there is room for a description of what may be the strongest and most interesting variant of IEM; namely what Pryor (1999) (...) has first brought to the fore under the name “Which-object IEM”, and which I’ll prefer calling existential IEM. I argue that recent development of relativist frameworks in semantics and pragmatics, particularly in Recanati (2007a-b), (2009), (2010), (2012a), may be precisely of a nature to address this limitation. The relativist theory of IEM, and in particular its suitability to cover existential IEM, may superficially seem to stem from a rejection of the core elements of the classical identification-freedom approach. However, I hope to show that, modulo a clarification and richer un-derstanding of the relevant notions of identification and identification-freedom, the relativist theory of IEM can be seen as both pushing further and complementing Evans’s intuitions, rather than conflicting with them. (shrink)