Consensus about a classif ication is def ined as agreement on a set of classes (concepts or categories) and their relations (such as generic relations and whole-part relations) for us in forming beliefs. While most research on scientif ic consensus has focused on consensus about a belief as a mark of truth, we highlight the importance of consensus in justifying shared classif icatory language. What sort of consensus, if any, is the best basis for communi-cating and reasoning with scientif ic classif ications? We describe an often-overlooked coordinative role for consen-sus that leverage agreement on how to disagree such that actors involved can still achieve one or more shared aims even when they do not agree on shared beliefs or categories. Looking forward, we suggest that investigating struc-tures and methods for coordinative consensus provides an important new direction for research on the epistemic foundations of knowledge organization.