I start from Evans’ criticism of temporalism, based on the claim that it does not “provide for the stable evaluation of utterances”. I try to show that, with suitable qualifications, assuming the possibility of evaluations yielding different truth-values at different times is not an “eccentric” move (as suggested by Evans). I briefly consider Prior’s metaphysical arguments in favour of the asymmetry between past and future and I suggest that, independently of these arguments, there are linguistic reasons in support of such an assumption. In particular, there are some future-oriented statements which (unlike past-oriented statements) are conceived of by speakers as intrinsically revisable and which require a non-monotonic characterization of the changing backgrounds of information selected by the time flow. As shown by some peculiar uses of phase adverbs like “still” and “no longer”, variability in terms of truth-value assignation is a distinctive feature of this kind of statement. But another kind of variability of truth-value assignation is detectable in the case of present or past-oriented statements: in general, by refining the notion of context, it is possible to individuate different types of propositional contents, depending on which contextual parameters are abstracted over in order to account for different needs in communicative exchanges. Thus, in the final section of the paper, a more articulated notion of context allows for a richer (preliminary) description of the propositional contents that can be associated to utterances by abstracting over the relevant parameters.