The aim of this paper is to clarify the relation between Quine's holism, the indeterminacy of translation and positivism. The interpretation I wish to present can best be brought out by contrast with a recent paper by Christopher Boorse. 1 Boorse argues that Quine's indeterminacy of translation derives from a commitment to 'semantic positivism'-the view that the meaning of any sentence is exhausted by the theory-neutral observation sentences it implies. He considers three Quinean arguments for indeterminacy and attempts to show that each appeals to semantic positivism as an implicit premiss. Since he thinks positivism is inconsistent with Quine's holistic epistemology which he accepts, Boorse feels "the indeterminacy thesis is supported by no convincing argument" (p. 387).
I think Boorse is mistaken in thinking that indeterminacy can be severed from Quine's holism. Specifically, I will try to show that Quine's argument from the underdetermination of theory does not rest on semantic positivism but rather on Quine's rejection of analyticity which, in turn, is part of Quine's general holism. Despite this, I think Boorse is right in seeing a similarity between Quine and the logical positivists. However, as I will try to show, Quine's positivism is a conclusion from indeterminacy and, hence, ultimately from his holism rather than a dubious premiss as Boorse contends. This is all the more startling and interesting, I think, when we consider Quine's reputation as a critic of positivism, for, if I am right, one result of his original criticism is a return to positivism or something very much like it on the basis of premisses quite alien to Schlick's classic version.
The underdetermination of theory may be characterized as the view that two theories may imply the same observation sentences yet still be incompatible. Two such theories would thus be tied for first place with respect to observation:'we could not choose between them without bringing in non-observational criteria, e.g. simplicity and scope. For Quine, all theories including the linguist's are underdetermined by observation in this sense. However, in the case of physics, the problem only affects our choice of Synthese 39 (1978) 317-324. All Rights Reserved.