We tested the adaptive hypothesis of psychopathy in human and computer Prisoner Dilemma (PD) matches. From a cohort of 448 male undergraduates who completed the Levenson Self-Report Psychopathy Scale (LSRP), 31 high psychopathic (PP) and 29 low PP students, sampled from upper and lower quartiles of the cohort, played a 40-round PD match in which real money was at stake. The human matches were of three kinds: Low PP versus Low PP, Low PP versus High PP, and High PP versus High PP. In computer simulations, the empirical defection probabilities from High and Low PP human players were entered in a three-player round-robin match with tit-for-tat (TFT) or a less cooperative variant. The strategies proliferated or dwindled in subsequent generations in proportion to their winnings and the matches continued until two players became extinct. In the human matches, High PP players had higher earnings in matches with Low PP and lower earnings in matches with other High PP players. In the computer matches, TFT drove both Low PP and High PP to extinction. However, High PP drove both Low PP and a less cooperative variant of TFT to extinction. Our experiments showed that PP traits may proliferate in two ways. People with low psychopathy engage in selfish behaviors also and people with high psychopathy are capable of cooperation and deter defection by means of harsh punishment. The broader significance of the in silico experiment is that randomness in social encounters may allow behavioral expressions of psychopathy to thrive over an altruistic and reciprocal strategy such as TFT.
Building similarity graph...
Analyzing shared references across papers
Loading...
Balbuena et al. (Thu,) studied this question.
www.synapsesocial.com/papers/69c08bcaa48f6b84677f9876 — DOI: https://doi.org/10.1177/14747049261435215
Lloyd Balbuena
Nathan J. Kolla
John S. Logan
Evolutionary Psychology
University of Saskatchewan
Carleton University
Swinburne University of Technology
Building similarity graph...
Analyzing shared references across papers
Loading...