We claim that the emergence of trust is best explained by relation-based arguments. After briefly surveying alternative explanations which concentrate on material payoffs both with self-centered and with other-regarding preferences, we examine theoretical discussions of cooperative and trust behavior framed in terms of attitudes, esteem and, most of all, intentions. An important implication of all these approaches is that the relational element makes human interactions different, as it is also documented by a lot of evidence produced by neuroeconomic experiments. When trust is based on relations and on the recognition of the others’ intentions, efficient outcomes are brought about by the agents’ (at least) partial disregard for the maximization of their material payoff and by heavily personalized interactions. Both these features are distinctive of the functioning of communities and the particular way how they work and solve coordination problems.