Text mining is applied to 510 articles on econophysics to reconstruct the lexical evolution of the discipline from 1999 to 2020. The analysis of the relative frequency of the words used in the articles and their "visualization" allow us to draw some conclusions about the evolution of the discipline. The traditional areas of research, financial markets and distribution of wealth, remain central, but they are flanked by other strands of research-production, currencies, networks-which broaden the discipline by pushing towards a dialectical application of traditional concepts and tools drawn from statistical physics.
Keywords: correspondence analysis; lexical evolution of econophysics; text as data.