This project brings together neo-Aristotelian theory of motivational habituation and neuro-cognitive models of skill acquisition, in order to explain why it is so difficult to cultivate extended and sophisticated motivational habits that would not be so easily defeated by akrasia or other unduly situational influences. We will apply Bayesian models of social cognition to the acquisition of moral competences. Bayesian architectures have the virtue of explaining how we unconsciously minimize sophistication in order to reduce bioenergetic costs of learning. But while we prefer heuristic and narrow context-locked skills, sometimes, when environmental conditions justify it, we are able to invest in learning subtle and sophisticated patterns. Elucidating this ambivalent attitude in Bayesian terms will shed light on typical obstacles and unexplored opportunities for the cultivation of sophisticated motivational habits. The team is composed by philosophers working on the naturalization of virtue and computa-tional neuroscientists specializing in motivation.