John Stuart Mill's harm principle maintains that adult behavior cannot justifiably be subject to social coercion unless the behavior involves harm or a significant risk of harm to non-consenting others. The absence of harms to others, however, is one of the distinguishing features of many manifestations of vices such as the consumption of alcohol, nicotine, recreational drugs, prostitution, pornography, and gambling. It is therefore with respect to vice policy that the harm principle tends to be most constraining, and some current vice controls, including prohibitions on prostitution and drug possession, violate Mill's precept. In the vice arena, we seem to be willing to accept social interference with what Mill termed self-regarding behavior. Does consistency then imply that any popular social intervention into private affairs is justifiable, that the government has just as much right to outlaw skateboarding, or shag carpets, or spicy foods, as it does to outlaw drugs? In this paper I argue that advances in neuroscience and behavioral economics offer strong evidence that vices and other potentially addictive goods or activities frequently involve less-than-rational choices, and hence are exempt from the full force of the harm principle. As an alternative guide to vice policy, and following some direction from Mill, I propose the robustness principle: public policy towards addictive or vicious activities engaged in by adults should be robust with respect to departures from full rationality. That is, policies should work pretty well if everyone is completely rational, and policies should work pretty well even if many people are occasionally (or frequently) irrational in their vice-related choices. The harm and robustness principles cohere in many ways, but the robustness principle offers more scope for policies that try to direct people for their own good, without opening the door to tyrannical inroads upon self-regarding behavior
