Central to Paul K. Feyerabend's philosophy of science are two theses: (1) there is no standard observation language available to science; instead, observability is to be viewed as a pragmatic matter; and (2) when considering questions of empirical significance and experimental test, the methodological unit of science is a set of inconsistent theories. I argue that the pragmatic theory of observation by itself decides neither for nor against any particular specification of meaning for an observation language; and that Feyerabend's position (...) provides no decision procedure when two contending theories share no terms having the same meaning, and thus cannot be said to be logically incompatible. Also, Feyerabend's insistence upon falsification will force him to admit that there are relatively permanent facts available to all theories, or to abandon the idea of test as falsification and to conclude that scientific theories can only be accepted or rejected on the basis of non-evidential considerations. (shrink)