Virtue ethics prescribes cultivating global and behaviorally efficacious character traits, but John Doris (Noûs 32:504–530, 1998; Lack of character: personality and moral behavior. Cambridge University Press, Cambridge, 2002) and others argue that situationist social psychology shows this to be infeasible. Here, I show how certain versions of virtue ethics that ‘go mental’ can withstand this challenge as well as Doris’ (Philos Stud 148:135–146, 2010) further objections. The defense turns on an account of which psychological materials constitute character traits and which the situationist research shows to be problematically variable. Many situationist results may be driven by impulsive akrasia produced by low-level (in some cases even perceptual), emotionally induced ignorance about one’s situation, and some may be driven by a further subtype: modal akrasia. Many subjects in the infamous Milgram experiments, e.g., seem to have recognized what the virtuous thing to do was and that they should do it, and only failed to do it because their emotions prevented them from seeing (or at least from recognizing, at the level of deliberation) that they could. If the primary constituents of character traits are higher-level mental dispositions involved in deliberation, though, then the results don’t show that these psychological materials are problematically variable.