There is an error – or infelicity – of translation in the English version of Grammatology which is so deeply anticipatory (if not performative) of the future of deconstruction in the United States as to deserve exemplary status. It concerns nothing less than the origin of language. Rousseau, it will be recalled, imagines passion's first utterance as a woman tracing with a stick (or “baguette”) the outline of her beloved's shadow on the ground. What sounds, he suggests, could ever match that inscription? Derrida, as well he might, reinscribes that “mouvement de baguette” as a chapter heading in his book. Here, now, is the English rendering of Rousseau: “How she could say things to her beloved, who traced his shadow with such pleasure! What sounds might she use to render this movement of the magic wand?” In the translation, it is the beloved who appears to be tracing his – own – shadow. The origin of language, that is, appears to have been lost in the translation. But the translation seems to convey its own myth: in between languages, the woman appears to have deferred to her beloved, to have handed him the stick or wand and invited him to speak (or trace) her love for him to himself. Communication has been lost, the inscription of difference (i.e., Derrida's argument) has been botched as the beloved is made to speak the other's admiration of himself…

It is tempting to view this sequence as a kind of primal scene of literary adulation – an obliteration of difference by deference (to refer to my own attempt at a mistranslation in the title of these remarks, although it is not nearly as eloquent as the case just quoted).