The anthropology of conflict and the Tübingen-Cologne project on conflict at the local level Hartmut Lang Institut für Völkerkunde, Universität zu Köln, Albertus-Magnus-Platz 5, D—5000 Köln 41, Deutschland The following articles are revised versions of papers first presented and discussed at a conference of the German Research Society in November 1989. All the articles have been written by members of the Tübingen-Cologne project on conflict at the local level with the exception of the articles by Marc Ross and Douglas White. The project owes much to the research of both authors. The remainder of this paper outlines the design of the Tübingen-Cologne Project. In 1985 the German Research Society invited application for a peace research pro- gram titled: The genesis of violent conflict in third world countries. Within the frame- work of this program Thomas Schweizer and the author of this paper have developed a project which has two broad aims. The first aim is to refine and test hypotheses about the causes of war on the anthropological level and that is the micro level. The second aim is to explore hitherto neglected and new areas of causation. Our project is divided into two branches. The branch we call the data base branch assembles existing data into an electronic data base, but it also will analyse these data. The field research branch consists of two workers, who did field research in Kenya among the Pokot agro-pastoralists, and in Ruanda in the urban setting of Kigali re- spectively. In the latter branch the emphasis 1s on exploring new causes, whereas a prime task of the database branche is to test the hypotheses. For testing the hypotheses we apply the methods of cross-cultural studies, using a world wide sample of cultures. The description of these cultures is contained in an- thropological field research reports, i.e. in texts. For the statistical analysis the relevant information in the texts has to be transformed into codes. When we were designing the project we were aware of a rather serious problem in this type of cross-cultural research: the problem of checking the codes for coding er- rors and validity. If the codes of a cross-cultural project are published without refer- ence and citation of the ethnographic sources, the costs of code checking are nearly as high as the costs of creating the codes anew. And even with references and citations code checking is still a rather time consuming task. Thus error checking is in these cases seriously hampered. We call a text-code data base our solution of this problem. That is, we are creating a data base that contains not only the codes but also the text slices from which the codes Zeitschrift für Ethnologie 115 (1990) 9-11 © 1992 Dietrich Reimer Verlag