ABSTRACT
This article examines how teacher beliefs about student learning might be justified using a genetic or “nonfoundational” theory of knowledge, wherein beliefs help the knower solve problems in his/her environment. The genetic view of knowledge implies that teachers need to help learners construct their own knowledge, but it also implies that teachers (as learners) need to construct their own knowledge of student learning. These two implications are not always in harmony. The issue is further explored by considering two specific problems with teacher beliefs: that they are apt to be situated, not general, and that they are apt to be tacit, not explicit. The first problem is addressed by distinguishing between reproducible and plausible inferences, and the second problem is discussed in connection with the importance of language in knowledge growth. The upshot is a normative ideal that has teachers discussing plausible views about student learning in an explicit way.