One of the worthy findings of brain research reported in the current issue of The Economist (”Thanks for no Memory,” November 15, 2003) is that traumatic memories are not repressed or blocked off so as to skirt anxiety, as Freud believed; rather, they are never formed. The evolutionary advantage of this is obvious: if one is better off by not remembering an event or a place, natural selection will see to it that the brain is constructed accordingly. The newspaper leaves it at that, though. It is plain that this arrangement makes sense only in the case of singular or rare traumatic experiences, which are impossible or unlikely to recur, whereas merely painful or unpleasant ones had better be committed to memory for future reference. The evolutionary advantage is obvious once again: one is better off by remembering events and places one should avoid. That is learning by another name. Therefore, the brain must correctly distinguish between traumatic experiences, on the one hand, and painful or unpleasant experiences, on the other. How does the brain do it? That remains the question.