Evolution of mindsight, transparency and rule-rationality

Evolution of preferences models often assume that all agents display and observe preferences costlessly. Instead, we endogenize mindsight (to observe preferences) and transparency (to show preferences) as slightly costly mechanisms that agents may or may not possess. Unlike in the costless models, we show that universal rule-rationality, mindsight and transparency do not constitute an equilibrium but universal act-rationality, mind-blindness, and opaqueness do. We also find that rule-rationality, mindsight, and transparency may exist in evolved populations, albeit only in a portion of the population whose size fluctuates along an orbit around a focal point. We apply our results to Ultimatum and Trust games to explore how costly and optional mindsight may affect economic performance in interactions among evolved agents.