What does it mean to predict earthquakes?

In the historical and philosophical literature regarding scientific prediction, what constitutes a prediction is often taken for granted. The debate on prediction is focused on its role and status in the scientific endeavor. In this paper I move in a different direction by problematizing the notion of scientific prediction. I do so by focusing on earthquake prediction, an exemplary case for studying the problem of scientific prediction. More specifically, I focus on a controversy around the so-called VAN method, a method proposed by three physicists as capable of short-term earthquake prediction. This controversy was an integral part of the broader debate regarding the feasibility of earthquake prediction in the 1980s and 1990s. In reconstructing this controversy, I follow scientists in their struggle to define earthquake prediction and to legitimize themselves as experts in earthquake prediction. In other words, I follow their attempts to determine what it means to predict earthquakes.