Mémoires: for Paul de Man, by Jacques Derrida; xv & 153 pp. New York: Columbia University Press, 1986, $17.50. Discussed by R. D. Ackerman 44"|""veconstruction in America"—that was the intended topic ofthe -L/three memorable lectures that comprise Mémoires. But following the death of Paul de Man, Derrida instead gave himself over to writing about memory and friendship — for, of, and about de Man. Which is to say that Derrida also ended up writing about historical memory and deconstruction, about a deconstructive understanding of pastness and even about a deconstructive sense of future. Indeed, this intensely evocative book crucially concerns the proposed topic after all. For in pursuing the question of memory, Derrida writes as he must about the current state and stakes of deconstruction, about deconstruction in America, about the promise of deconstruction. There emerges in the late essays of de Man and in the recent work of Derrida a new affirmative emphasis that is as seductive as it is obscure. "What is love, friendship, memory, from the moment two impossible promises are involved with them?" (p. 149). Thus Derrida questions in the concluding paragraphs of Mémoires. "What are we, who are we, to what and to whom are we, and to what and to whom are we destined in the experience of this impossible promise?" What is this promise or these promises? What is the nature of deconstructive affirmation? That is the topic I wish to consider here (all too briefly), and that is also the guiding concern of Mémoires: "Underlying and beyond the most rigorous , critical, and relentless irony . . . Paul de Man was a thinker of affirmation. By that I mean— and this will not become clear immediately , or perhaps ever— that he existed in memory of an affirmation and of a vow: yes yes" (p. 21). Derrida adds elsewhere that de Man's "thought of the promise . . . [is] the most profound, most singular, 171 172Philosophy and Literature and most necessary thought; probably, too, the most difficult and most disconcerting" (p. 93). Our understanding of deconstructive affirmation remains clouded— perhaps of necessity— by the nature of the affirmation itself (or themselves ). Characteristically, Derrida directs our attention not to a particular affirmation but to the conditions underlying the possibility of all affirmation and, ultimately, to the conditions that make any act possible . (This transcendental phase of Derrida's undertaking remains in line with Kantian critique, as we will see.) He conceives deconstructive affirmation (that which makes friendship possible, for instance) as an "alliance" that is "much more ancient, resistant, and secret than all those strategic or familial manifestations of alliance that it must actually make possible and to which it is never reduced." This deconstructive yesyes is not simply active or performative, and yet it "precedes" or "underlies" occurrence in general: "we would understand nothing about what comes to pass and takes place if we did not account for this affirmation" (p. 19). If metaphysical accounts of the act ("Acts" is the title of the third lecture of Mémoires) depend upon a singular moment of revelation (an Augenblick), the deconstructive account (we know by now) begins with doubleness, a doubleness that disrupts efforts to discover a unifying bridge between being and knowing, for example, or between performative and constative uses of language or versions of rhetoric. It is this doubleness that lies behind the necessity for the double affirmative. This is why the deconstructive affirmation is a "non-active act" that "must repeat itself to itself: yes, yes. It must preserve memory; it must commit itself to keeping its own memory ... if anything is ever to come from the future. This is the law" (p. 20). The law to which deconstructive affirmation remains bound is the law of disjunction: "This disjunction is the law, the text of law and the law of the text" (p. 145). * For both Derrida and de Man, the act (or the affirmation) must be understood — the conditions of its possibility accounted for— as a writing or a text, as initially constituted by a duration or a spacing. Thus there must always be two moments or "times" of affirmation that remain intervallic, although even for de Man there...