ABSTRACT
In the previous chapter I argued that there is an “artifi ciality” problem in ABS. Even when the agents are assumed to be cognitive, not reactive, only generalizations/abstractions about the attributes of cognitive agents are entered into simulations. The attributes (preferences and behavioral predispositions) of agents and the rules governing their interactions are abstracted from the literature on the empirical research fi ndings in psychology, social psychology, economics, and sociology (Nan and Johnson, 2009). These generalizations do have some validity, of course, but they are only generalizations. To what extent these generalizations represent the attributes of the actual agents in a given situation-to what extent they are “realistic”—is an open question.
