In his contribution, Prof. Rorty suggested that transcendental arguments are only useful as a means of providing self-referential ad hominem arguments against particular philosophical views, showing that those views fail to make good sense of our knowledge, such as it is, and of our dealings with the world. I do not know if Rorty is correct, for I still do not understand how to delimit transcendental arguments properly. But my argument fits his description. I will offer a self-referential ad hominem argument against the theory of (scientific) truth offered by Wilfrid Sellars and largely shared by Jay Rosenberg. In spite of its specific concern with Sellars, the argument is of general interest. It has two principal strands. One, picked up by Prof. Parsons in his comments, shows that Quine's doctrine of indeterminacy of translation is incompatible with the most interesting forms of what Sellars, Putnam and Rosenberg call Peircean realism. During the last two years I have been persuaded, rather against my will, that the indeterminacy of translation cannot be evaded. Accordingly, I hold that Peircean articulations of scientific realism cannot be adequately defended. The other major strand of the argument concerns the correction of fact-stating claims forced on us by theory replacement and by conceptual change in science. Such correction undermines the vestiges of a "growth by accretion" model of the development of science which are still to be found in Sellars's and Rosenberg's writings. In its general thrust, my argument supports Putnam's contention that all viable articulations of scientific realism will land us in some form of what he calls "internal realism."¹