I make three related claims. First, certain seemingly secretive behaviours displayed by scientists and inventors are expression neither of socio-professional values nor of strategies for the maximization of the economic value of their knowledge. They are, instead, protective responses to unavoidable risks inherent in the process of publication and priority claiming. Scientists and inventors fear being scooped by direct competitors, but have also worried about people who publish their claims or determine their priority: journal editors or referees who may appropriate the claims in the manuscript they review or patent clerks who may claim or leak the inventions contained in the applications that cross their desks. Second, these protective responses point to the existence of an unavoidable moment of instability in any procedure aimed at establishing priority. Making things public is an inherently risky business and it is impossible, I argue, to ensure that priority may not be lost in the very process that is supposed to establish it. Third, I offer a brief archaeology of regimes and techniques of priority registration, showing the distinctly different definitions of priority developed by each system.