The text of the <i>Thousand and One Nights </i>has been the stuff of dreams for generations, throughout the Arab and non-Arab Orient, and since the XVIII<sup>e</sup> century throughout Western Europe, thanks to L. Galland's translation, based on a popular Egyptian edition. Since then, there have been countless translations and compilations of a legendary text that has no known authors, and those who assume it are, ironically, only copyists, compilers or translators. This is why it is presented as a palimpsest whose sources must be unraveled, authenticated versions established and the mirror effects of the stories that make it up analyzed. This is what we shall attempt to demonstrate in this study, based on the critical corpus duly established by R. Khawam. This study is divided into four successive stages: The fate of the Thousand and One Nights text, the question of manuscript and/or printed sources, an examination of the French corpora of Galland, Mardrus and Khawam, the indeterminacy of the archi-text or palimpsest, and finally the singularity of Khawam's corpus.