<p>A familiar criticism of so-called ‘assurance’ theorists is that they fail to demonstrate how testimony delivers a specifically <i>epistemic</i> reason. Whilst their accounts of the speech act of telling may do justice in various respects to our self-understanding as interlocutors – as agents who engage in the intentional act of telling and addressees who take tellers at their word – they at best explain how the act has a certain moral or pragmatic significance, but not how it is conducive to knowledge. In this essay, I show one way in which assurance theorists can meet this challenge. Drawing on Richard Moran’s elaboration of an assurance view, I argue that despite his wish to set questions of knowledge-acquisition to one side, the core ideas guiding his account suggest a certain epistemology of testimony, which I label the ‘Naïve view’. This view precisely offers an understanding of how assurances possess a distinctive epistemic significance. However, I go on to explain how both Moran’s elaborated account and Faulkner’s recent emendation to it preclude Naivety from the outset. As a result, I argue, their attempts to demonstrate how telling as an assurance has a sui generis epistemic significance are compromised. At fault in both cases is a prior commitment to a conjunctivist understanding of the epistemic value of telling. Whilst abandoning this commitment may strike most assurance theorists as a high price, it opens, I suggest, a path to rebutting their critics’ charge of epistemic irrelevance.</p>