What do patterns in legal language tell us about power, policy and justice? This question was at the heart of a conference on “The Fabric of Language and Law: Discovering Patterns through Legal Corpus Linguistics”, convened in March 2016 by the international research group “Computer Assisted Legal Linguistics” (CAL²) under the auspices of the Heidelberg Academy of Sciences. About forty scholars from Germany, Switzerland, Italy, Poland, Spain and the US brought together their different intellectual and disciplinary perspectives on computational linguistics and legal thinking. Concluding the conference, four legal linguistics experts – two native linguists, two native lawyers – discussed the perspectives and limitations of computer-assisted legal linguistics. Their debate, which this article faithfully reproduces, touches on some of the essential epistemological issues of interdisciplinary research and evidence-based policy, and marks the way forward for legal corpus linguistics.