In some of his books Paul Val6ry uses as epigraph a line from G6ngora, and even though he misquotes it, his mistake is a perfect illustration of my argument. "En rocas de cristal," cites Val6ry, "serpiente breve" (in crystal rocks a brief serpent). His own biases lead him to believe that those serpents must have wriggled across a static, contrasting surface, creating and destroying their letters in the process. He is searching for a background of immobility, "the crystal rocks" on which those dark games of the snakes can be performed. These lines are the beginning of "La toma de Larache," which reads "en roscas de cristal serpiente breve" (in crystal curls a brief serpent). The difference between original and citation marks the desire, the prior vision that destroys the line's reality. Jose Lezama Lima, "Luis de G6ngora's Serpent"'