Feminist empiricism in context

This presentation is part of the Feminism and Empiricism (Quinean Themes) track. In my view, feminist epistemologists and philosophers of science have offered some of the deepest challenges to traditional philosophy of science. Some of the most interesting work, in my view, has developed in the context of defending the very possibility of ‘feminist science.’ From a traditionalist point of view, after all, feminist science makes as much sense as Marxist science or Christian science. It is taken for granted that political and/or moral values are external to scientific activity. At best, such values are irrelevant to scientific success; at worst, they are impediments to it. Making room for a feminist science involves challenging the traditional view at its roots. Some feminists find a rich resource for this project in the work of W.V.O. Quine (Nelson 1990; Antony 1993; Campbell 1998; Solomon 2001). These philosophers argue that Quine’s naturalist empiricism, taken to its logical (if not Quine’s) conclusion, provides the conceptual space for a genuinely feminist empiricism. Focusing on actual cases of scientific success and failure, they are able to show that political and moral values (e.g., a commitment to gender equality) can be truth-conducive in certain circumstances (e.g., when male chauvinism is endemic). Given Quine’s holism and his naturalism, they conclude that a feminist empiricism is consistent with and, indeed, demanded by Quine’s views. I find this work compelling for substantive and strategic reasons. From a substantive point of view, I find Quine’s work, and feminist efforts to extend it, both convincing and quite general in its import. From a strategic point of view, I find this approach compelling because Quine’s work, and feminist work built upon it, addresses the traditionalists on their own ground. Traditionalists are not likely to be convinced by critique ‘from the outside.’ That said, though I am convinced for the most part by defenders of a Quinean feminist empiricism, I have reservations about Quine’s views that extend to their feminist interpretation. In particular, I believe Quine’s wholesale rejection of the analytic-synthetic distinction, and the associated rejection of a priori knowledge, is overstated (Quine 1969). Though Quine is right to assert that all sentences are potentially falsifiable – no sentence is immune to revision – all sentences are not potentially falsifiable in the same context. This point is demonstrated forcefully by Michael Friedman (2001). Friedman develops a historical sketch of the transition from Newtonian to Einsteinian physics in order to show how the development of non-Euclidean (Riemannian) geometry was a necessary precondition for the development of Einstein’s general theory of relativity; according to Friedman the general theory of relativity could not be conceived (stated) absent Riemannian geometry. From the perspective of Einsteinian physics, then, Riemannian geometry is properly viewed as (contingently) a priori: the axioms of Riemannian geometry are not open to empirical refutation in this context. In order to understand and work with the general theory of relativity, one must presuppose Riemannian geometry because Riemannian geometry is partly constitutive of an Einsteinian world-view. Thus, contrary to Quine, all sentences are not equally open to empirical refutation (in the same context). If Friedman is right, and I believe he is, Quine’s naturalist empiricism is misleading and impoverished. It is misleading because it treats of all sentences equally: all sentences are falsifiable to the same degree. Only pragmatic considerations (e.g., the principle of minimum mutilation) differentiate sentences in this respect. It is impoverished because it fails to take context seriously. For Quine, pragmatic criteria like the principle of minimum mutilation are global criteria, applying to science as a whole. These failings are shared by many followers of Quine, including feminist empiricists. The consequences here are twofold. First of all, all values are treated equally: it is entirely an empirical matter whether or not a particular political or moral commitment will turn out to be truth-conducive. Second, because local contexts are ignored, the truth-conducive character of all values is equally contingent or contingent in the same way. A commitment to gender equality, e.g., may be truth-conducive given endemic male chauvinism, but the former is entirely contingent on the latter. Of course, Friedman’s work suffers from similar problems. For one, according to Friedman the (contingently) a priori status of a sentence, or set of sentences, turns partly on its foundational character in a given historical epoch (Friedman 2001). This is, no doubt, one of the reasons Friedman emphasizes conceptual or semantic presuppositions. Following Michael Williams (2001), I resist Freidman’s (Kantian) emphasis on purely conceptual or semantic presuppositions, as well as his foundationalism. On my view, a priori presuppositions or principles are not necessarily conceptual (or semantic) in nature. Among other possibilities, a priori principles may be what Williams calls ‘methodological constraints,’ or necessary conditions of the possibility of engaging in a specific activity (Williams 2001). Furthermore, as the latter implies, the a priori or constitutive status of a principle does not necessarily turn on its foundational character, as Friedman avers. On my view, a principle may be constitutive in one context but not in another. Thus it needn’t play the same role across all or most contexts to count as constitutive. I believe that Friedman and Williams bring valuable resources to anyone interested in the role of values in science, including feminist empiricists. First of all, both authors emphasize the different roles played by sentences in different contexts: all sentences are not treated equally. Though we must, post-Quine, abandon the idea of an absolute distinction between internal and external values, Friedman and Williams both draw our attention to the import of a relativized version of this distinction. This more complex picture, in turn, illuminates the different roles that values may play in a given scientific activity. Furthermore, though the appropriateness of a value commitment in a particular context will always be contingent from an external perspective, it may well be necessary from an internal perspective. Thus, the truth-conducive character of all value commitments will not be equally contingent, or contingent in the same way. Finally, precisely because a contextual approach focuses attention on value commitments that are internally necessary (but externally contingent), this approach has strategic advantages: practitioners cannot disavow these value commitments without misrepresenting their own disciplines