The Social Metaphor for Distributed Processing

Abstract Social systems are asynchronous distributed processors characterized by (i) a large and variable population of small individuals and (ii) a random and changing communications architecture. Individual interactions produce global computations or processes which are immune to minor structural changes. Termite colonies are a good example. Global processes—such as the construction of a mound—are essentially unaffected by most details of colony structure. How are global computations programmed locally? What types of programs are immune to structural changes? Given the ability to program at the individual level, can social systems be viewed as general purpose computers? These and similar questions have motivated this paper. We present (i) precise definitions for social systems and related concepts, (ii) seven programs resulting in interesting global processes, (iii) theorems describing the effect of structural changes on various global processes, and (iv) some questions for further research.