DISSERTATION SEMINAR Thursday, September 22, 10:05am-12:05pm C21, Solomon Labs TITLE: Retaliation versus vigilantism: Why do we choose to punish? Victims retaliating against aggressors tend to gain the benefit of a deterrent effect against future exploitation through second-party punishment. However, research has not adequately explained the benefits behind vigilantism, in which unaffiliated third parties risk personal costs to administer punishment for an act that had no impact on their future economic well-being. Differences in costs and benefits of second-party punishment (2PP) and third-party punishment (3PP) suggest that the two punitive behaviors may respond to different contextual cues and serve different adaptive functions. However, differences between 2PP and 3PP are often blurred in the literature, with researchers taking the findings on 2PP to draw conclusions about 3PP, and vice versa. This dissertation outlines the functionally distinct roles second and third parties assume in conflict (Chapter 1), and empirically tests predictions of the proposed model (Chapter 2-3). In Chapter 1, I evaluate experimental paradigms eliciting different types of punitive responses, and examine both animal and human behavioral data on inputs and outputs characteristic of 2PP and 3PP. I propose that 2PP functions to deter, while 3PP serves as a coordination and conflict resolution device. I predict that second parties are more sensitive to inputs relevant to achieving the goal of discouraging future violations than third parties. In two studies (Chapter 2), I use fictional crime scenarios to test the effect of probability of crime detection on punitive judgment. I find that second parties are more likely to take probabilities into account than third parties when assigning punishment. In four additional studies (Chapter 3), I use behavioral economic games to explore conditions under which second and third parties seek out – or avoid – information about norm violations. I find that in the case of more serious violations, second parties show interest in details of Dictator's decision, but only when provided with an option to punish. In contrast, the presence or absence of an opportunity to punish has no effect on third parties. Taken together, these findings demonstrate that second and third parties respond differently to relevant contextual cues, further supporting the claim that punishment administered by avengers and vigilantes follow different behavioral rules.