MacWhinney is to be commended for reopening questions about the logical problem of language acquisition in the light of new data and models. Unfortunately his discussion is marred by technical errors, false dichotomies, and inadequate attention to detail.