From text to intertext. The Macron palimpsest in the revealing of Artificial intelligence

Intertext is the condition for the interpretation of any text. However, its materialization, its explicitation, its implementation have always seemed difficult for textual linguistics. This contribution proposes a materialization of the intertext within "numerical reflexive corpora", and calls in an original way artificial intelligence (deep learning, convolutional model) and logometry to systematically explore the intertext thus materialized. For this study, we postulate that the Elysian corpus since 1958 constitutes the intertext of Emmanuel Macron's speeches, from which the president draws consciously or unconsciously to construct his speeches. As an example, we thus automatically bring to light the grammatical imprints of Giscard in Macron's speech, and the lexical borrowings that the new president concedes to de Gaulle.