Abstract
The rational approach and the emotional approach coexist in the study of trust. They are too diversified to synthesize, so we start with one approach-Coleman's expected utility model of trust-and relate it to the other approach. There are two problems about his model, however. First, it does not answer why the trustor cannot expect the trustee's response for sure, Second, he does not build a formal model to support his statement that communication among trustors makes the trustee more trustworthy. We build game theoretic models to solve the above problems. We assume in the first model that there are two types of the trustee-honest and greedy-and that the trustor does not know which type is realized. The equilibria of this game reproduce the results of Coleman's formulation. Then, we build the second model with N trustors and one trustee to explain the effect of communication on trust. We add an assumption that there are communication channels among trustors through which information about past interactions between the trustors and the trustee flows. One equilibrium of this game shows that even the greedy trustee chooses to be trustworthy･ This can be interpreted as the effect of communication on trust relations.
