Ad quintum dicendum quod duplex est humana ratio. Una demonstrativa cogens intellectum ad consensum, et talis ratio non potest haberi de his quae fidei sunt, sed potest haberi ad evacuandum ea quae fidem esse impossibilem asserunt. Quamvis enim ea quae sunt fidei demonstrari non possint, non tamen possunt demonstrative improbari. Si autem talis ratio ad probanda ea quae sunt fidei induceretur, evacuaretur meritum fidei, quia iam assentire his non esset voluntarium, sed necessarium. Ratio autem persuasoria sumpta ex aliquibus similitudinibus ad ea quae sunt fidei inducta non evacuat fidei rationem; quia non facit ea esse apparentia, cum non fiat resolutio in prima principia quae intellectu videntur. Nec iterum meritum fidei evacuat, quia non cogit intellectum ad consensum, unde assensus remanet voluntarius.
-- Super Boethium de Trinitate, p. 1, q. 2, a. 1, ad 5.
[To the fifth argument, it] should be said that there are two kinds of human argument. One is demonstrative, forcing the assent of the intellect, and argument [sic] of this kind cannot be had of objects of the faith, but they can be had to refute those who assert that faith is impossible. For although the objects of faith cannot be demonstrated, neither can they be demonstratively disproved. If this kind of argument could be put forward to prove the objects of faith there would be no merit, since assent to them would not be voluntary, but necessary. Persuasive argument taken from things similar to what is of faith is not in conflict with the concept of faith, since it does not make them evident, since there is no resolution into the first principles to be grasped by intellect. Nor does it take away the merit of faith, because it does not force intellect to assent and thus assent remains voluntary.
-- tr. Ralph McInerny, Aquinas: Selected Writings (London: Penguin, 1998), p. 129.