This dissertation seeks to fill two lacunae in contemporary feminist discussions of essentialism: first, a lack of critical analysis of the term "essentialism" and its cognates, and second, a paucity of feminist work that aims to develop anti-essentialist methods rather than merely presenting anti-essentialist critiques of existing feminist theories. I propose a typology of feminist essentialisms, distinguishing metaphysical, biological, linguistic, and methodological variants. I argue that methodological essentialism---understood as the practice of making false generalisations about women based on the experiences (...) and identities only of a particular group---is the most pressing political issue for feminists, and defend Elizabeth Spelman's anti-essentialist critique against its opponents. Anti-essentialism should not, however, be interpreted as disavowing the category "women" altogether, and I use Ludwig Wittgenstein's arguments in his Philosophical Investigations to articulate a form of feminist anti-essentialism. that understands similarities between women as family resemblances. This approach enables feminists to make generalisations about women that neither obscure important differences nor deminise our political efficacy. This Wittgensteinian feminism rejects the a priori and urges us to "look and see" to justify generalisations about women. I interpret this as a call for a feminist anti-essentialism that is embedded in feminist practice, and ask what "look and see" might mean for feminist research and for feminist organising against sexual violence. In chapter four, I argue that Carol Gilligan's recent work on girls' psychology in the context of race and class differences successfully responds to long-standing charges that her research is essentialist. It does not, however, fully meet the methodological challenge of anti-essentialism as it fails to acknowledge power relations embedded in research processes, which in turn shape conclusions about female identi. (shrink)