Any Theory of Rationality is normative – it describes how a rational agent should evaluate beliefs, not (only) how they in fact evaluate them.
To see how this is not a problem, we can start with the question why
we attribute normative power to the theory in the first place. We do
that because we’ve
reached
a reflective equilibrium concerning its merits – we think it is
the most reasonable account in the light of practical experience, that
is, the most useful account we know of. In other words, the
theory will reflect and articulate our practices.This means it will be a pragmatist theory of
rationality.
This Articulation, i.e. making explicit, is the actual job of a theory of rationality: The normativity was there before, it is implicit in our practices; the theory has to show how and when exactly it works to our evolutionary advantage, when and why it doesn’t, and thus help us improve upon it.
Of course this calls for different methods than in sciences that don’t deal with social practices. But under a pragmatist banner, the final aim doesn’t differ from these other sciences: To increase our chances for success.