Assyrian merchants meet nuclear physics. A history of early contributions from social sciences to automatic pattern detection in graphs (late 1950s–1970s)

Résumé En

The current success of “network analysis”, and the fact that its proponents are mostly computers scientists and physicists, tends to obliterate that early applications and improvements of graph analysis methods were pushed forwards by social scientists and researchers from the humanities. The research led by Jean-Claude Gardin from the 1950s, which spanned both archaeology and the emerging automation of numerical computation and documentation, are a striking example. In this paper, I focus on his work related to automated applications of graph theory to historical and ethnographic materials, that he developed from the late 1960s to the 1970s. These early work were then widely ignored both in archaeology and network analysis literature. Working from publications and archive materials this paper aims to 1) retrace and contribute to the history of pattern detection in graphs (namely cliques and cycles); 2) contribute to the history of what was called “non numerical computing” in the 1960s; 3) highlight the importance of the applications to historical data for the development of some computing methods.