Rapid growth over the past two decades in digitized textual information represents untapped potential for methodological innovations in the adaptation governance literature that draw on machine learning approaches already being applied in other areas of computational social sciences. This Focus Article explores the potential for text mining techniques, specifically topic modeling, to leverage this data for large-scale analysis of the content of adaptation policy documents. We provide an overview of the assumptions and procedures that underlie the use of topic modeling, and discuss key areas in the adaptation governance literature where topic modeling could provide valuable insights. We demonstrate the diversity of potential applications for topic modeling with two examples that examine: (a) how adaptation is being talked about by political leaders in United Nations Framework Convention on Climate Change; and (b) how adaptation is being discussed by decision-makers and public administrators in Canadian municipalities using documents collected from 25 city council archives. This article is categorized under: Vulnerability and Adaptation to Climate Change > Institutions for Adaptation K E Y W O R D S climate change adaptation, governance, policy, quantitative text analysis, topic models 1 | INTRODUCTION Text-based research methods have been a cornerstone of qualitative social science methods since the 1950s (Lasswell, 1952). These approaches see documents as meaningful artifacts that can be analyzed for their thematic and semantic content (Krippendorff, 2013), and they form a core component of the climate change adaptation governance literature. In lieu of directly observable and measurable indicators such as greenhouse gas emissions, adaptation governance research relies on written records, surveys, and interviews as its primary information sources about how different actors are responding to climate change impacts. Content analysis methods are commonly applied to sources such as government reports, strategic planning documents, peer reviewed and gray literature, and media stories (Araos et al.