Despite the importance of the “historiographical revolution” in Kuhn’s work, he did not carry out a specific study about it. Without a systematic investigation into it, he even affirms that the “old” historiography of science (OHS) is unhistorical, suggesting its summary disqualification in the face of his “new historiography” of science (NHS). My wider project, of which this paper is a part, is to better discuss the issue of the justification of the NHS. In this paper, I discuss the justification (and the genesis) of the OHS, focusing on Condorcet and Comte and resorting especially to Koyré. This will allow us to understand that the relation between the OHS and the NHS is a new instance of inter-theoretical incommensurability. And, indeed, that the NHS is not stricto senso a new historiography. It is the same historiography used for other disciplines (art, philosophy), which in the twentieth century begins to be applied to science as well.

The first manuscript of The Structure of Scientific Revolutions , probably written in late 1958, is available in the Kuhn Archive at MIT (Massachusetts Institute of Technology) 1 . It is the first version of Chapter 1, which is the introduction to the book, and is completely different from the version that was published. In this article, I turn to the manuscript to show that at that time Kuhn considered the comparison between the image of science and the image of art as the most appropriate way to announce his project: to change the image of science by bringing it closer to the image of art. As I try to demonstrate, this appeal to the history of art is not merely occasional. And it allows us to understand Kuhn’s intriguing retrospective statement, according to which Structure was a belated product of his discovery of the parallels between science and art. Some passages from Kuhn’s unpublished manuscript are transcribed in the article.