The concepts of “collective” and “social” are routinely confused, with claims about collective facts and their necessity justified by evidence that involves only social or interactional facts. This is the case with Harry Colllins’ argument for tacit knowledge as well. But the error is deeply rooted in the history of philosophy, in the notion of shared presuppositions popularized by neo-Kantianism, which confused logical claims of necessity with factual claims about groups. Claims of this neo-Kantian kind have difficulties shared by Collins’s argument for Collective Tacit Knowledge. The alternative, a fully social account of our capacities of understanding and interpretation, avoids these difficulties and allows for a model of tacit knowledge that serves the supposed explanatory needs that the “collective” part of Collins’s account purports to meet.