According to Roberto Caso, there are two models of open science: 1. an administrative model, which accepts the existing rules and institutions and tries to mitigate intellectual monopolies and the constraints of bibliometric research evaluation; 2. a subversive model that aims to make scientific conversations public by circumventing commercial publishing and bibliometric evaluation and by reversing copyright through free licenses to create a growing scientific commons. Both Kant and Humboldt's theories of university may be read as attempts to put subversion into a modern, and therefore bureaucratic, university. I. Kant treats books as speeches, i.e. actions through which authors engage with the public, rather than as things. Thus, the purpose of his copyright is to make literature public and to acknowledge the public's right, accepting publishers' rights only if they facilitate communication between authors and the public. Kant’s Enlightenment essay therefore advocates the freedom of the public use of reason from the institutions while only hinting at their own evolution. And in his final published book, the Conflict of the Faculties, Kant argued for freedom even within the Cameralist university that employed him, by suggesting that the higher faculties of theology, law and medicine, serving the government, could be scientifically sound only if the lower faculty of philosophy, committed solely to truth, were free to criticize them by the public use of reason. The freedom and openness of science lie at the root of an unsolvable inner conflict within the university. II. About ten years later, W. von Humboldt reformed the Prussian university so that the openness of science and the fact that it would forever remain “a problem that has not yet been completely resolved“ would become foundational principles of the institution itself. In his design, it was the state that had to guarantee such freedom. However, Humboldt himself was aware that his solution was not without danger. That raises the question: if open science is identified with the free public use of reason, can science really be both open and institutional?