The increased availability of textual data in digital formats alongside a proliferation of computational tools has led international law scholars to expand significantly their methodological toolkit in recent years, with many having found inspiration in quantitative methods of text analysis. Such methods enable scholars to identify patterns in large quantities of text and challenge pre-existing assumptions about various aspects of international law, how it is developed, practiced, interpreted and applied. Yet, subjecting legal texts to quantitative methods of analysis also comes with notable shortcomings. When treating texts as abstract data in accordance with the logic of data science, researchers often fail to appreciate texts as means of human interaction and communication, thereby diverting attention from terms and meaning to algorithms and statistics. The aim of this article is to offer a lay of the land of the burgeoning empirical legal scholarship that adopts text analytical approaches in its study of international law, to exhibit what gets lost when insights from linguistics are ignored, and what is gained when they are not.