Within this literature, Williams’ (1998) Voice, Trust, and Memory most directly engaged the issue of marginalized groups in the language of representation, framing all of the classic issues of representation within the terms of the contemporary debate. “Liberal representation” of the kind descended from Locke, though promising formal equality, systematically underrepresents the historically marginalized. By treating individuals as individuals rather than as situated members of groups, Williams argues, liberal accounts of representation fail to conceptualize patterns of disadvantage that are based in group situations, and are often replicated within representative institutions. The liberal account (at least in its Lockean form) assumes a trustee relationship based on convergent majority interests, which does not in fact exist for disadvantaged groups. When such assumptions legitimate electoral systems that simply aggregate votes based on territorial constituency—particularly in the form of single-member districts—they serve to justify and stabilize existing patterns of disadvantage. For this reason, Williams argues, we need to think beyond principal-agent models of representation in which principals are presumed to be formally equal individuals. We need to understand representation as a relationship, mediated by group histories and experiences, through which relevant constituencies—particularly those related to fairness—come into existence. Finally, fair representation requires some relationship of trust between individuals and representatives,
based on shared experiences, perspectives, and interests, and this is demonstrably not present for historically disadvantaged groups within residence-based systems of representation.