It is reminiscent ofPierre Nora’s suggestion that physical objects and especially the written word constitute‘archival memory,’ a secondary or ‘prosthesis’ memory (Nora, 1989: 14).

When the user stores his thoughts in his own filing cabinet, these thoughts are no longer his own but those of his filing cabinet. In turn, the machine that gathers and reproduces excerpts is, and remains, a ‘black box’. It is not simply another Ego for enacting a user’s soliloquy but a true Alter Ego with whom the user communicates. Additionally, when the machine is started, the user does not simply refresh his memory; the filing cabinet actually speaks. To achieve this practical outcome, the card index must be provided with a ‘life of its own’ (Eigenleben) which should be as independent of the life of its educator as possible.30 In this sense, the card index functions as a ‘secondary memory’ in Stübel’s terms. This result raises some questions which justify the present article. Is there a socio-structural reason why such an improbability became possible? Is there a trend, in early and late modern society, toward an externalization and technologizing of social memory? And what insight can we gain into intellectual history?

In a short academic dissertation on the art of excerpts, Andreas Stübel described the card index as a ‘secondary and subsidiary memory’ (‘memoria secundaria and subsidiaria’), summing up in just three words the dilemma scholars had been struggling with for two centuries with respect to the use of commonplace books.28 As far as I know, Stübel was the first among contem-poraries to speak of secondary memory.

In fact, the methodical use of notebooks changed the relationship between natural memory and artificial memory, although contemporaries did not immediately realize it. Historical research supports the idea that what was once perceived as a memory aid was now used as secondary memory.18