ABSTRACT
Many researchers in social psychology and judgment and decision making focus on identifying mental mistakes and troubleshooting people’s decisions. The focus on biases and errors leads researchers to underestimate the intelligence of evolved computational systems. Here we examine this issue in the context of human friendship decisions. Social psychologists have long claimed that people’s friendship choices are surprisingly unintelligent, based on strategically irrelevant factors such as proximity, familiarity, similarity, or very simple reinforcement learning. However, this view is becoming increasingly untenable as research on many nonhuman species uncovers sophisticated computational control systems that intelligently regulate behavior in cooperative relationships. We argue, in sharp contrast, that human friendship is caused by complex computational machinery that performs a strategic alliance-building function.
