This paper has its origins in a comparison between two mathematics texts. On the one hand, Cardan’s Ars magna (1545) is indecipherable today, and yet it is very representative of the mathematical XVIth century. On the other hand, Descartes’s Géométrie is the very first directly readable text in the history of mathematics. After 1637, the text of the Géométrie was certainly modified and improved, but in its original form the text had already acquired the constituent features of its present form. How and why did such drastic changes, reflected by the differences between the two mathematics texts, take place ? I suggest historical and epistemological answers to these questions, as well as an analysis of Descartes’s role regarding the following three basic points in the establishment of symbolic writing : the representation of powers (Cartesian exponents), then the representation of equality, and, finally, the representation of aggregation. I briefly describe one of the previous symbolic systems, called « diophanto-cossic », with all of its contradictions. The community adopted Descartes’s system instead, and it still remains in use today. Moreover, it was, chronologically speaking, the first. Actually, the Cartesian exponential for powers appeared for the first time in rule XVI of Descartes’s Regulae.