The 14 articles in this collection assess feminist epistemology (FE) from four perspectives: its status as feminism and philosophy, its claims about scientific practice, its philosophical and political worth, and its prospects for the future. With two exceptions, Sharon L. Crasnow and Janet A. Kourany, the authors reject FE vehemently on all four grounds. Perhaps the strength of feeling expressed should not be surprising, since FE challenges established opinion about philosophy and science on many fronts. The major lapses in scholarship and logic, however, are surprising given the philosophical standing of the contributors.

The most obvious and fundamental lapse is that the authors who categorically condemn FE do not categorically condemn feminist empiricism, though the latter is correctly acknowledged at the beginning (6) to be a form of FE. Even if we grant, as feminist empiricists generally do, the serious shortcomings of the rival forms of FE, namely, feminist postmodernism and feminist standpoint theory, demonstrating these shortcomings is not sufficient reason to reject FE wholesale. Nevertheless, these authors with the notable exceptions of Crasnow and Kourany reject FE as a methodology without examining the methodology of feminist empiricism.

The second egregious lapse is the failure to recognize, let alone engage adequately, the efforts by FE supporters to resolve the problem that they label ‘the bias paradox’. Rejection of the ideal of impartiality in science because it leads to gender bias appears both to reject the ideal of impartially and at the same time to accept it. The contributors frequently raise this problem in one form or another (20, 22, 43, 60, 162, 194, 228, 229–230, 240), but nowhere do they address the discussion of it by supporters of FE. In particular, they ignore the obvious move of removing the appearance of contradiction by distinguishing between two interpretations of ‘impartiality’. In Louise Antony's well-known Reference Antony, Antony and Witt1993 paper “Quine as Feminist: The Radical Import of Naturalized Epistemology,” she contrasts (a) the ideal of trying to remove all background biases from scientific tests with (b) the ideal of preserving only background biases likely to track the truth when combined with empirical tests. Feminist epistemologists can reject the first ideal—objectivity as value-neutral inquiry—noting how the illusion that it is attainable can generate gender bias and distortions of the truth, without rejecting the second ideal—objectivity as truth-tracking inquiry—and thus avoid contradiction. Antony's paper is referenced in the bibliography, but the contributors address neither her proposed resolution nor the ensuing feminist literature about how to distinguish good biases from bad without circularity.

How are these lapses possible if we assume that the authors all endorse the ideals of fairness in argument and scholarly care that most of them find lacking in those who accept some form of FE? A clue comes when the editors say early on: “In this book we emphasize the accounts [of FE] that purport to provide a gendered methodology” (6). Does feminist empiricism not count as providing a ‘gendered methodology’? To answer we need to ask what exactly gendered methodology is. I gather that a methodology is gendered (with respect to feminism) just in case it embraces feminism but not the ideal of impartiality. But which understanding of impartiality is relevant? If it is the first of those just distinguished, then most forms of feminist empiricism, such as Anthony's, would count as gendered, since they reject that norm of impartiality. The problem of ignoring feminist empiricism in this case remains in full force. If, on the other hand, we take the condition to use the second understanding, then forms of FE in which feminist norms interact with scientific norms of justification, such as comparative empirical adequacy, would not count as providing a gendered methodology. But in the latter case some of those under severe attack for espousing a gendered methodology, such as Anne Fausto-Sterling, are among those who would not have a gendered methodology, since they accept the ideal of impartiality just when it is understood in the second sense. In short, the failure to deal with the basic ambiguities that give rise to the bias paradox and thus to use the paradox uncritically against FE throughout most of the articles is of a piece with the failure to take seriously feminist empiricism as a form of FE. They are, moreover, colossal failures no less of logic than of scholarship, making the caustic dismissal of FE on both grounds more than a little ironic.

Further lapses can be documented, but let me turn to the positive. The two pieces friendly to FE in some of its forms have much to recommend them. Crasnow argues that Longino's intersubjective (hence nonrealist) understanding of objectivity cannot give a noncircular account of what distinguishes good bias from bad; on her view only a realist understanding will do that. I have argued the same point independently in more than one place. Note, however, that this is criticism of one feminist empiricist by another, not the rejection of FE as a whole. Kourany, on the other hand, rejects realism because it does not help us to understand how science motivated by feminism could be superior normatively and empirically. While I believe that she is wrong about realism, I believe she and most supporters of FE (especially feminist standpoint theorists) are right about the interdependence of the normative and the empirical. (Contributors object on 48, 149, 164, and 222–229.) Kourany gives many examples to illustrate the interdependence. The case of heart disease (214) is particularly striking. At one time studies of coronary heart disease (CHD) were confined to large samples of men, based on the androcentric assumption that the etiology of CHD in men would be the same in women. The situation is complex because CHD kills women overall at roughly the same rate as men, and the results of the tests provided good data for assessing how men develop CHD. However, CHD develops differently in women as later tests confirmed. No one seriously doubts the objectivity of the later findings, yet the fact that CHD develops differently in women was discovered only when the androcentric assumption was challenged. Normative considerations in this case yield, in the context of suitable empirical tests, a less distorted and, in one of the senses distinguished above, a more objective scientific understanding of women's health.

There are also positive features in the other articles. Robert Klee argues with considerable plausibility that FE, qua FE, has a restricted range of application for successfully producing more accurate findings than epistemology that is not feminist. Noretta Koertege makes the shrewd point that many commentators who bash science for being too ‘positivistic’ make the same error as positivists in assuming that present empirical adequacy is the only legitimate scientific reason for preferring one hypothesis to another (58). Christina Hoff Sommers assembles good evidence to show that in the U.S. girls are doing much better at school than boys. Meera Nanda gives an illuminating account of the link between Dewey and Bhimaro Ambedkar (a mid-twentieth-century Buddhist philosopher) and its implications for the use of science to combat oppression. Susan Haack offers in one of her essays, “Staying for an Answer: The Untidy Process of Groping for Truth,” a moving account of the toils of honest inquiry. Regrettably the authors fail to see how these valuable contributions are compatible with FE, indeed support forms of FE. Their points are presented in the context of a wholesale dismissal of the very idea of FE. The truth is that this scrutiny of FE fails to be impartial in either of the understandings of impartiality distinguished above.