ABSTRACT

A fundamental misconception has pervaded the social sciences, and is also reflected in the social studies carried on in AI. It is a static, non-emergent view of micro-interactions. Social needs and capabilities are taken for granted, either because they are inherent in the structure of the situation investigated (think of game theory), or because they are built into the agents (think, again, of the strategic beliefs found in game theory and, more generally, in the theory of rationality). This statement may appear unwarranted given the impressive number of game-theoretic studies about the emergence of cooperation. However, co-operation and other positive social phenomena are considered as emergent macro-effects of micro-interactions, the latter being essentially taken for granted.