It's my impression that departmental language requirements are becoming extinct. And in my view that's a good thing. Students working on, say, Aristotle need to acquire a command of Greek that goes far beyond what departmental requirements would impose, and students working on, say, current debates in the metaphysics of free will probably have no occasion to use a foreign language in their research at all. In fact, students of the latter kind would be far better off devoting the time they might dedicate to learning a language to learning more modal logic, or more physics, or more psychology. And students in the former category would be better served by an intensive summer program in the languages they need, as determined by a dissertation director, given the student's projected research. The department, in other words, should have no independent role in any of this. It seems obvious to me.
But perhaps I'm overlooking some rather compelling consideration in favor of a departmental requirement?