Reply to Don Hirsch

Abstract

I AM grateful to Don Hirsch for his comments, especially as they give me a chance to explain how what I have said about meaning, if unstable, is indeed intended to fall down on the relativist rather than the realist side. Hirsch has two principal objections to the family resemblance account: (1) It gives no principles for putting a limit to the accommodation of theories to each other, and so gives no justification for saying that one is good or another is wrong. (2) Because it deals with particular similarities in limited domains, it seems to imply that universalist claims for scientific theories are always incorrect. I will discuss these points first in terms of science, and then point out some differences in the application to literary theory. The dilemma Hirsch points to in (1) is exactly that of the DuhemQuine thesis for scientific theories. This assumes that the core postulates of theory are usually not directly verifiable or falsifiable by experience, and that they take their meaning from their place in the whole network of theory. Theories therefore come to the test of experience as wholes, they are underdetermined by empirical data, and in Quine's formulation, Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system.' Thus, to take Hirsch's example, Priestley was able to maintain this theory with a fair show of reasonableness, even in face of Lavoisier's experiments and arguments, by partially reidentifying cases of by regarding oxygen as dephlogisticated air, by explaining the negative weight of by a repulsion between matter particles and phlogiston, and so on. All this did not render Priestley's theory identical with Lavoisier's because, for example, oxygen was not regarded as a simple substance, but as the mixture air from which a component phlogiston had been removed.