Abstract

A steadily mounting body of social science research suggests that ascertaining a person’s conscious motives for an action may not always provide a complete explanation of why he did it. The phenomenon of unconscious bias presents a worrisome impediment to the achievement of fair equality in the workplace. There have been numerous deeply insightful articles discussing various aspects of this problem and canvassing its implications for antidiscrimination law. My purpose in this paper is to focus directly on what might be called a more naïve question: should implicit bias be a basis of disparate treatment liability under Title VII? The question might fairly be regarded as naïve insofar as any proposal for such liability would surely be unripe for present implementation, in light of serious issues pertaining to problems of proof in individual cases, not to mention intramural disputes among experts about the proper practical inferences that can be drawn from extant social science research. My interest, however, is more theoretically basic. I want to understand whether and how the notion of unconsciously biased action fits into our operative legal concept of actionable discrimination. To reach that issue, I devise a thought experiment in which I assume, first, that unconscious or implicit bias is real in a sense that I will make explicit, and second, that unconscious discrimination is provable – i.e., that the influence of implicit bias on an agent’s action is something that can, in principle, be proved in individual cases. With these assumptions, I construct an hypothetical test case that squarely raises what I regard to be the hard question for theorizing about unconscious discrimination. Should an employment action give rise to liability when that action was provably affected by the actor’s unconscious bias in respect of a statutorily protected classification, even when the actor consciously acted only on legitimate, nondiscriminatory reasons? The payoff of this thought experiment is not only a clearer picture of the theoretical commitments entailed by liability based on unconscious bias, but also a keener understanding of our currently prevailing notions of actionable discrimination.