This Element combines a machine learning-based approach to detect the prevalence and nature of generalization across tens of thousands of sentences from different disciplines alongside a robust discussion of potential solutions to the problem of the generalizability of textual evidence.