Joseph Manson: UCLA Department of Anthropology

Several studies have shown that, following brief interactions among strangers, subjects perform better than chance at predicting whether their co-subjects will defect in a one-shot Prisoner’s Dilemma Game (PDG). However, previous work did not explore how such predictive accuracy was possible. Theoretical work suggests that adherence to “arbitrary” norms serves a signaling function that allows individuals to assort so as to maximize coordination and cooperative efficiency. One set of norms, documented by Conversation Analysis (CA), concerns the details of face-to-face interaction, such as respecting interlocutors’ rights to complete each turn constructional unit (TCU). We hypothesized that Defection in a PDG and predictions of co-subjects’ Defections would be more likely vis-à-vis unfamiliar interlocutors who violated conversational norms at higher frequencies. We videotaped short conversations among same-sex triads of previously unacquainted university students who were naïve to the impending PDG. We then separated participants and directed each of them to (a) play a one-shot PDG with each of their two co-participants and to predict one another’s play decisions and (b) complete the Levenson Self-Report Psychopathy Scale (LSRP) along with some demographic questions. From each videotaped conversation, a mean of 3.93 min was transcribed using CA methods. Subjects chose to cooperate in 67.7% of decisions, comparable to previous studies of face-to-face interaction. However, participants performed no better than chance at predicting one another’s game play decisions. Nonetheless, personality and conversational behavior were related to one another and to objective outcomes like game play and earnings, in both expected and unexpected ways. We discuss these results with reference to the co-evolution of human cultural variation, cultural capacities and social judgment mechanisms.