Modeling the Spatial Distribution and Temporal Evolution of Cultural Attributes:
Computer Simulations and Empirical Tests

 

     For more than a dozen years, I have simulated social impact in large groups of individuals who interact over time with their neighbors in social space. These simulations and their theoretical background, funded by a number of NSF grants, have been published in over three dozen chapters and articles. They show how micro-level changes in cultural attributes (e.g. attitudes, values, and behaviors) that result from individuals influencing one another lead to four forms of macro-level group organization. 1. There is a consolidation or decline in the numbers of people exhibiting minority attributes as, of necessity, the minority is more exposed to contrary pressure than the majority. 2. There is a spatial clustering of attributes as neighbors in social space become more similar to one another as a result of their contact. 3. There is an emerging correlation of attributes, because even clusters formed independently on different issues will overlap. 4. And finally, there is a continuing diversity of attributes as clustering protects minorities from becoming exposed and eliminated.
     Results of many thousands of simulations using a variety of techniques show that these consequences are extremely robust with respect to theoretical assumptions, parameter values, system characteristics, and initial conditions. Experiments exposing sets of real people to repeated messages from their neighbors in virtual space, confirm that these consequences are not simply figments of the computer’s imagination. In real groups as in simulations, minorities became reduced in size, attributes became spatially clustered, initially unrelated attributes became correlated, and minorities survived. Furthermore, the simulations, based on a few simple change rules, accurately predicted the degree of self-organization, not only according to type of attribute, but at the level of individual groups.
     These phenomena are likely to occur at several scales and may have implications for a variety of cultural and political phenomena, such as the changing geographical distribution of dialects, religious practices, and voting behavior.


Bibb Latané
Center for Human Science
http://www.humanscience.org>www.humanscience.org
latane@seafrolic.org