n the French edition of Confronting Images appeared in 1990, it won immediate acclaim because of its far-reaching arguments about the structure of images and the histories ascribed to them by scholars and critics working in the tradition of Vasari and Panofsky. According to Didi-Huberman, visual representation has an "underside" in which seemingly intelligible forms lose their clarity and defy rational understanding. Art historians, he goes on to contend, have failed to engage this underside, where images harbor limits and contradictions, because their discipline is based upon the assumption that visual representation is made up of legible signs and lends itself to rational scholarly cognition.