INDIVIDUAL EXPERIENCE AND THE FRAGMENTATION OF SOCIETIES

We use computer simulations to show how social fragmentation and conse quent social loss can result from six innocuous cognitive and behavior al assumptions. These assumptions govern individuals' choices with res pect to each other and are independent of processes associated with fr agmentation in various classical literatures. We model ''society'' as a set of individuals who encounter each other in circumstances that pe rmit potentially productive bur also risky relationships with each oth er. Individuals (1) have the option of playing Prisoner's Dilemma game s with others they encounter, but they do nor have to; (2) have prefer ences over mutual cooperation and free-riding outcomes; (3) have expec tations about each other's play; (4) have observable cateogry ''tags'' ; (5) can modify their future responses to ''tagged'' individuals base d on experience; (6) can search out partners they believe represent go od prospects for productive relationships. In these terms, the only co ndition necessary for fragmentation to emerge is that some circumstanc e makes the probability of encounters across ''tagged'' categories une qual, and a difference in sizes of categories, however slight, is suff icient to do that. Same difference in category sizes is, for practical purposes, inevitable, and thus so are pressures toward fragmentation.