The Swamping Problem is usually construed to be a puzzle about ultimate epistemic value. In this paper I want to show that the Swamping Problem is also a puzzle about the commensurability of epistemic values. Many replies to the Swamping Problem try to show that, besides the value of truth, knowledge produced from reliable processes also have some other value (call it the value of x). But no reply seems to take the following problem serious: namely, why we can aggregate (...) the value of truth and the the value of x—after all, they seem quite heterogeneous. (shrink)