Systematic bias in humanities datasets: ancient and medieval coin finds in the FLAME project
This publication was co-authored together with my colleagues on the FLAME Project Lee Mordechai, Alan Stahl, and Mark Pyzyk. It draws on FLAME’s decade-long experience in the field of digital humanities to examine the biases that underpin both FLAME and digital humanities projects at large. The digital humanities revolution was undertaken with the promise that, by using digital tools like online databases and more recently AI, humanities scholars can make new and groundbreaking research that was based on objectively accumulated data. Numbers, after all, don’t lie. But the people who write them down might. After a decade of work on FLAME, we have encountered a number of biases, ancient and modern, intentional and unintentional, regional and methodological, that have so far prevented us from achieving our goal of reconstructing the economy of Late Antiquity and the Early Middle Ages. By systematically examining these phenomena, we hope to frame a discussion of such inherent biases in other digital humanities undertakings.