According to the view I call `innocent Millianism', that-clauses differing only for occurrences of co-referential names provide the same contribution to the intensional profile of a belief report. It is widely believed by friends and foes of innocent Millianism alike that this approach entails either the denial of what I label a `naïve' account of belief reports, or a dismissive attitude towards our semantic intuitions. In this essay, I counter that the conjunction of innocent Millianism and the naïve view of belief reports is compatible with our intuitions of truth-conditions. In order to defend this conclusion, I defend an independently motivated approach, in which utterances endowed of the same intension may nevertheless differ in truth-conditions.