The main purpose of this paper is to apply the general idea of contextualism as a theory of knowledge attribution to the very specific case of testimony and trust characterized as being the procedure of the attribution of knowledge (and sincerity) to the informator. In the first part, I have argued in favor of evidentialism - a viewpoint that takes epistemically responsible trust as a matter of evidence. In the second part, I have considered the question of how strong an evidential basis has to be for epistemically responsible trust. I have briefly registered two main tendencies in contemporary debates about trust and testimony: (i) the non-unitary character of our trust; (ii) and the requirement for disaggregation of evidential standards. In short, I have argued in favor of the stance that any "undiscriminatory generalization" - both reductivist's one offered by J. Locke and D. Hume or anti-reductivist's offered by T. Reid - concerning epistemically responsible trust is a kind of inappropriate theoretical idealization, and that a certain theoretical reconciliation has to be offered. Finally, in the third part, I will propose trust-contextualism as the viewpoint that optimally harmonizes both our intuitive and theoretical requirements about epistemically responsible trust.