Data mining, also known as, Knowledge Discovery in Databases (KDD), is the creation of non-trivial knowledge suitable for action from databases of vast size and dimensionality. From the mid-1960s to the late 1990s, data mining moved from a disparaged, dubious sort of statistical work—“fishing” or “dredging”— to become what its practitioners proclaim to be an utterly transformative technology.

According to KDD advocates, traditional scientific approaches to data and the traditional competencies of scientists simply cannot keep up with the volume of data and multidimensionality possible thanks to computers. Something else is needed, something less pure, because it deals with the vast impurities of dynamic data that nearly always issues from a particular business, governmental, or scientific research goal. Using traditional and digital humanities methods, I look at how stories of technologically determined emergence were crucial to legitimizng the way data mining is authorizing the loosening and the partial abandonment of the disciplinary and epistemological values of its predecessor disciplines, statistics and machine learning.