Why Nazified Germans killed Jewish people: insights from agent-based modeling of genocidal actions

Abstract

This chapter explicates the logic of a computational agent-based model bearing on the willingness of perpetrator agents to conduct genocidal actions against Jewish people during World War II. Given realistic distributions of benefits and costs and sufficient time, as a joint consequence of these distributions and interpersonal influence the model readily creates agents who are avowed anti-Semites, Nazis, and perpetrators of the genocide, even transforming agents characterized initially by lower levels of anti-Semitism. Although many agents initially exhibit dissonance (i.e., a disjunction) between their attitudes and choices, toward the end of this period their anti-Semitic attitudes and choices become consonant (i.e., internally consistent). Experiments and parameter studies using this model indicate that different distributions of benefits and costs, changed legitimacy of authority, and different values of anti-Semitism of influential agents can modify the growth of prejudice, Nazism, and genocidal choices in these random-number-based Monte Carlo trials. The results clarify the conflicting interpretations of Goldhagen and Browning concerning the genocidal actions of a battalion of perpetrators and the role of propaganda in reducing moral costs. Six hypotheses that focus the testing of the model can be generalized creating insights about other genocides.

Citation

Smith, R.B. (2010), "Why Nazified Germans killed Jewish people: insights from agent-based modeling of genocidal actions", Dahms, H.F. and Hazelrigg, L. (Ed.) Theorizing the Dynamics of Social Processes (Current Perspectives in Social Theory, Vol. 27), Emerald Group Publishing Limited, Leeds, pp. 275-342. https://doi.org/10.1108/S0278-1204(2010)0000027012