Corresponding Author:

Moral judgments about harm vs. impurity differ in a number of non-superficial ways, as shown by dozens of studies, conducted by dozens of separate research labs, using a wide variety of methods and stimuli. Gray & Keeney (2015) attempt to explain away these differences by arguing that the “confounds ” of severity and typicality may account for them all. This comment examines the evidence for this claim. Severity and typicality are undoubtedly important factors for moral judgment, but Gray & Keeney fail to demonstrate that they account for any (much less all) of the harm-impurity differences in the literature. Correlated ratings of “harm ” and “impurity ” are redundant with severity (.93≤rs≤.97), merely tracking overall wrongness. The conclusion that harm and impurity judgments don’t meaningfully differ at all, that all the functional and cognitive differences in the literature are “illusions ” resulting from “confounds” and “sampling bias, ” is entirely unwarranted by the present studies