JACQUES DERRIDA’S OF GRAMMATOLOGY famously announces “the end of the book,” but it is hard not to think of it as quite a book, and especially perhaps in the Anglo-American context, as Derrida’s magnum opus. Although another of Derrida’s three 1967 books had already appeared in English,[1] the Grammatology translation in 1976 — with its somewhat manifesto-like quality, its focus on apparently less specialized and more broadly familiar material (Ferdinand de Saussure, Claude Lévi-Strauss, Jean-Jacques Rousseau), and its 100-page preface by a then relatively unknown Gayatri Chakravorty Spivak — became the book to have read across the humanities. It launched “deconstruction” in America.

Derrida makes a startling claim in the Grammatology: to focus on the apparently marginal and secondary issue of writing raises problems serious enough to overrun all the conceptual resources of the then triumphant “human sciences” (and their model of scientificity provided by structural linguistics), in addition to those of history in general and indeed philosophy itself. All these disciplines share presuppositions that a hard look at the question of writing radically unsettles.

Like other seismic events of thought, Derrida’s insight is quite simple, yet in its very simplicity hard to grasp. Identities in general (of whatever kind, at whatever level) arise out of difference, but difference is not itself any identity or indeed any thing at all. It is not that there are first things, and then differences and relations between them: the “things” emerge only from the differences and relations, which have an absolute priority, and that emergence is never complete. It’s that insight that led to the neologism différance. In the beginning is différance, which means that there is no simple beginning or origin. And the différance never ends, which means that there is no simple end. Derrida’s simple claim, then, is that nowhere ever is there anything simple.

For many readers at the time, the most accessible way into this thinking was Derrida’s account of Saussure, and more especially his radicalization of Saussure’s own insight that “language is a system of differences without positive terms.” On this view of language, objects and meanings do not come first, only subsequently to be named and referred to via some conventional linguistic means. Rather, they are from the start involved in the play of difference which alone affords them any kind of identifiability and identity. Things are what they are only by bearing the trace of what they are not. It flows from this insight that thinking itself is always caught up in webs or weaves of traces, and that it needs to engage with those traces if it is going to be able to think at all. That engagement involves, Derrida makes clear, alongside a question of language, a question of temporality and a question of the relation to the other. For short, we can say that this means thinking proceeds essentially by reading. But just because of the trace-structure, reading can no longer be conceived as retrieving a content (a signified, in the structuralist jargon) from the text being read, and must be thought of quite differently. The Grammatology is also a meticulous ongoing “methodological” reflection on what it is doing as it reads in this different way.

By thinking and reading in this way, we are always transgressing the limits of language itself. Once signifieds and referents are identifiable only through the trace-structure, then language has neither inside nor outside: everything in general is what it is only through indefinite referrals to other “things,” which themselves refer on again. Experience in general is differential, made possible by the trace that cannot itself be directly experienced. Being is trace-being. This is not a claim merely about language, however important language remains. It would be tempting to say it is an ontological claim, except that the trace, being no thing or object at all, cannot be held within the terms of ontology (something that has been conveniently forgotten by many more recent “realisms”), whence Derrida’s later half-serious proposal of a hauntology.

Derrida’s simple insight, with its almost unimaginably, fractally complex implications, is difficult to stay with, and, especially since his death in 2004, intellectual fashion has tended to bypass its complexity and settle back into the more familiar terms of science, ontology, and, especially, history, as though the mere passage of time could make it go away. This new, “40th anniversary edition” of the translation of Derrida’s perceived magnum opus is, then, extremely timely in its untimeliness, and, we might hope, has a chance of shaking up our disciplinary habits all over again.

In 1997 a so-called “corrected” but in fact largely unchanged edition of the Grammatology was released, but only now, 40 years on, has the translation itself been quite extensively revised by Spivak herself. One hopes that this might set a precedent for more revised editions of Derrida translations, many of which, often done by enthusiastic readers whose French was very far from native, are much more inaccurate than the original Grammatology translation. In this case, the volume has also been bulked up for its anniversary from 360 to 441 pages by the adoption of a slightly less crowded page layout (which is welcome), and the addition of what I am tempted for some reason to call two appurtenances: a new Introduction by Judith Butler (pp. vii-xxiv), and a new Afterword by Spivak herself (pp. 345-68). There is something definitely monumentalizing about the publication of this volume in contrast to the 1976 original, when Spivak was just an almost unknown translator: three big names now appear on the cover above “40th Anniversary Edition | Newly Revised Translation.” Not only “revised,” but “newly revised,” hot off the presses, eager to be an event all over again. But anniversaries are complex and often embarrassing events, and this one is no exception: at least in all that is new in it, the new edition is in fact a big disappointment. Far from enhancing Derrida’s book or even reflecting the progress that has been made in understanding deconstruction since 1976, the Introduction, the Afterword, and — most importantly — the revisions to the translation all represent significant steps backward.


Given the very substantial and ambitious nature of Spivak’s original Preface, which is reprinted here and stands up rather well 40 years later (reminding the reader of a time before her more demanding, disconcerting, and idiosyncratic “late style,” here exemplified in the Afterword), I assume that the decision to add a further Introduction by Judith Butler was motivated by a desire to give the old book some more contemporary credentials than it might otherwise seem to have. Perhaps the Press imagined no harm could possibly come from associating Butler’s name with Derrida and Spivak. Whatever the motive, and whatever sequence of editorial oversights one can only imagine followed from it, the decision has produced an unfortunate result: the added Introduction is, as a matter of fact, riddled with vagueness, inaccuracies, misunderstandings, and plain errors. It does Spivak’s (and, more importantly, Derrida’s) work a real disservice, badly scrambling in advance the access a new generation of readers might otherwise have had to Derrida’s book. Luckily for such new readers, perhaps, the Introduction is very hard to follow, and is perhaps most likely to be skipped in favor of Spivak’s original Preface or (better still) Derrida’s own text.

Let me try not to dwell unduly on the detail of the Introduction. It doesn’t actually start so badly: reflecting on the issue of translation itself, Butler draws quite helpfully on Derrida’s lecture “What is a Relevant Translation?” to set up important problems about originals and origins, tries to find Derrida doing something other than Hegel in the perception that origins are always in a sense derivative of their derivatives, and registers Derrida’s thought that translation harbors an inevitability of failure and even ruin. The first real sign of trouble comes when Butler quotes Derrida as claiming, “There is nothing outside the trace.” No source is given for this quotation, and pour cause, either here or when she repeats it a page later (“As we have seen, he writes that ‘there is nothing outside the trace’”) because Derrida never wrote that sentence. Nothing serious, we might charitably concede, comes of this very curious but all in all minor lapse for the famous il n’y a pas de hors-texte, “there is nothing outside of the text” (translation unchanged in the new edition, and still with brackets offering the alternative “there is no outside-text” and providing the French); we might even begin to hope that it would presage Butler’s avoidance of the usual erroneous understanding of the word “text” in that passage to refer to writing in the common sense.

The real trouble is yet to come. When she actually begins to describe the book to which this is the Introduction, Butler immediately goes wrong, claiming quite falsely that it was Saussure’s Course in General Linguistics that “provocatively proposed that a future science of grammatology might include linguistics only as a subfield,” and going on,

In the course of defining language as a system of signs, Saussure proposed that writing might be the way to understand the more general structures of language. Distinguished from speech, writing includes phonic and nonphonic elements, and so offers a more capacious approach to language.

I can only assume that this rather extraordinary sequence comes not from any reading of Saussure himself, but from a hasty look at p. 55 of the Grammatology, where, having entertained the suggestion — one that Saussure really did make — about a general science of semiology including linguistics as a “subfield,” (albeit a subfield that would remain the “general pattern” for semiology more generally), it is, quite explicitly, Derrida and not Saussure who is provocatively (though only provisionally) replacing “semiology” with “grammatology” in Saussure’s formulation. It is of course Derrida and emphatically not Saussure who thinks that “writing might be the way to understand the more general structures of language,” not really because it would be more “capacious” (a signature Butlerian adjective), and not really either because “it includes phonic and nonphonic elements” (assuming we even understand what that really means), but because of the structure of the trace, which can be brought out most perspicuously and effectively via the standard view of writing as “(graphic) signifier of the (phonic) signifier.”

Things do not get any better as Butler ploughs on into an extraordinarily inaccurate account of Saussure’s notion of the sign and how that notion differs from that of his predecessors, who, Butler incorrectly claims, thought there was a necessary link between sign and referent — she might have looked at Aristotle (“conventional”), or at Locke (who even uses the word “arbitrary” (Essay Concerning Human Understanding, III, II, §8)); the difference with Saussure in fact lies not in a distinction between necessary and arbitrary, but in the radicality with which “arbitrary” is being thought — and thought radically enough, “arbitrary” entails difference, and difference entails the trace.

These specific problems in the account of Saussure aside (there are many more), the root cause of Butler’s difficulties in the Introduction seems to be that she shares the widespread misapprehension — that 40 years really might have put to rest — that Derrida is essentially talking about discourse when he uses terms like “writing” or “text,” or even that he is proposing a philosophy of language in the form of a “grammatology.” Maybe when the book was first published and the general haze of structuralism and post-structuralism was harder to see through it seemed as though Derrida could have been talking about nothing else. But the book is pretty clear from the outset that it is taking a certain “inflation” of language or sign-talk (“inflation itself”) as an object of suspicion and, indeed, of deconstruction, and that it is primarily concerned to bring out the conditions of impossibility of any grammatology.[2] However dense the comments on Heidegger in the Grammatology (clarified recently by publication of Derrida’s 1964-5 seminar), and however difficult the pages about the trace in the Saussure chapter may remain, Derrida is explicit that the thought of the trace breaks down not only the idea that language is made up of signs (which might be a good reason to look to Frege and analytic philosophy, or perhaps Chomskyan linguistics rather than structuralism, but no more than that), but also the idea that “language” can be thought of as set over against something else (a “world” of “things” or “objects” or “referents”). Derrida is rather clearly and energetically deconstructing, and certainly not endorsing, any such view of language and its others. Différance does not primarily emerge because of a delay between a referent and a discourse about that referent, as Butler suggests, but affects the very definitions of referent and discourse. Already in the opening chapter of the book Derrida writes, not merely of the relationship of speech and writing but indeed of language in general and writing, that “the concept of writing exceeds and comprehends that of language.” The claim is that the apparently unlimited extension or inflation of the concepts of language and writing, initially conceived logocentrically as the transport of a meaning ideally separable from the means of that transport, leads to an overflow exceeding language in anything like its “normal” senses, and exceeding by the same token the predicates (“human,” “intentional,” “meaningful,” and so on) that have traditionally been tied to that concept. Writing or text in Derrida’s sense is not discourse or any other recognizable determination of language, but the beginning of the in-determination of language into the absolute generality of the trace-structure. This is what commits him to the still widely misunderstood claims that the trace has to be thought “before” distinctions such as those between man and world, nature and culture, human and animal, being and beings, life and death, and even animate and inanimate. Butler’s essentially Foucaldian understanding of language as discourse — however interesting it might be on its own account and within its limits — will always fall short of the thought of the trace.


Perhaps this is a good moment to begin to talk about the translation itself. At a crucial and difficult moment in the account of the trace from the Saussure chapter, Derrida writes the following in his original French: “On ne peut penser la trace instituée sans penser la rétention de la différence dans une structure de renvoi où la différence apparaît comme telle et permet ainsi une certaine liberté de variation entre les termes pleins” (De la grammatologie, my italics). “Retention of difference in a structure of reference,” as Spivak translates the italicized phrase in both editions. There might be good reason to translate renvoi here not as “reference,” but as “referral,” if only to avoid the specific sense that “reference” has in the philosophy of language, where it is in any case ambiguous between the action or process of referring to something, and the something thus referred to. Derrida’s renvoi is not yet reference in either of these senses (for which French would happily use référence), in that the whole point of the différance view of language is that — like the famous letter that always might not arrive and that, even when it does, remains haunted by the possibility of its non-arrival — it never does really arrive at either signified or referent (and to that extent is not made up of signs, and indeed not even of signifiers, at all). Referral refers on and refers on again, endlessly. A trace is always the trace of a trace. Renvoi never winds up in reference — it is precisely the différance of reference, so it would probably be best to use the plausible term “referral” here.

Regret that Spivak did not change this particular translation from one edition to the next leads one to wonder what did in fact get changed. Her new Afterword provides some hints about this — none of them particularly reassuring. See for example the quite incomprehensible rationale for reverting to the first-person plural pronoun in the new translation (p. 367), where the argument goes: in the Grammatology Derrida uniformly uses a standard academic “nous” that was initially translated as “I”; in his late work, by contrast, he seems to differentiate the “we” into “I,” “you” and “them” — in order to respect that differentiation, the “I” has been changed back to “we” in the revised Grammatology, where however no such differentiation in fact obtains. This rationale produces quite a few inconsistencies in the translation, where the change of “I” back to “we” is not fully carried out, and at least one funny moment, when Spivak in the Afterword quotes from the old first-person singular version of the text to have Derrida “producing the problems of critical reading” and “often embarrassing myself in the process,” whereas the revised text has “often embarrassing ourselves in the process.” The question of exactly who and how many we are to be embarrassed here, and why, is produced by the fact that en nous y embarrassant in French really does not refer to the affect of embarrassment at all, and simply means getting caught or tangled up in the process of producing those problems.

Butler, incidentally, also thinks that this I/we issue is important: quoting (now, confusingly enough, with page-reference to the older edition, p. 99, which corresponds to pp. 107-8 in the new one) from the “Introduction to the Age of Rousseau” (as the older edition had it[3]), she misconstrues (as does Spivak in both editions) a sequence that both she and Spivak nonetheless recognize to be methodologically important. This passage is used by Spivak to make a point about “cathecting proper names” that recurs several times in her Afterword (but not in Derrida himself), and explains why Derrida, implicitly negotiating his way here between Foucault (no — ages, epochs, or epistemes do not exhaustively determine texts) and Deleuze (no — we can’t simply “create” new concepts), will focus so closely on Rousseau and his theory of writing. The passage gives five concessive reasons for adopting this procedure, all the while recognizing that it is problematic: we do not really think that the proper names of authors or movements can be used to designate causes, origins, identities, or structural coherences, but 1) texts are not for all that simply effects of structures; 2) all available concepts for thinking the articulation of a text and a historical structure are caught up in the metaphysical closure we are trying to think outside; 3) we have no other concepts for this and will be unable to produce any while we’re still caught in that closure; 4) any prospect we have of making progress here depends on first questioning the internal structure of texts as symptoms; 5) this prior “internal” reading is the only way we can determine the belonging of texts to that metaphysical closure. From these five reasons, nous en tirons argument pour (“we draw from this an argument in favor of […]” or, a little more freely, “we find in this grounds to […],” Spivak has a slightly indeterminate “we draw our argument from them”) to single out Rousseau and his theory of writing for this kind of preliminary reading. Butler silently simplifies this whole sequence, notably having the “from them” refer only to the concepts in point 2 above rather than to the whole argumentative sequence, and then makes a meal of it being all about the “we” rather than the “I,” before tailing off into obscurity again:

As one can see by my insertion [she has supplied the French to show Derrida’s use of the pronoun “nous”], there is no reference to a first-person “I” but rather a “we” who questions and draws out one text from another [it is anybody’s guess what this “drawing one text out from another” is describing in this immediate context]. Is there an implication that some “we” is in this bind together, breaking open a metaphysical closure that has preempted not only what can be written and understood, but the interconnected (and contingent) limits of writing and intelligibility?

To which the answer, here at least, has to be no.


Translation is a laborious and painstaking job at the best of times (as indeed is checking a translation against the original text and explaining what might have gone wrong). In her Introduction, Butler suggests (perhaps not intending it literally) that Spivak actually re-translated the whole book (“she translated it forty years ago, and now again […],” and while it would be inaccurate to say that the new edition is quite such a radical revision as that, there are indeed many changes — seemingly on every page — compared to the earlier version. Butler, just before the rather Olympian judgment that, “It is rare to find such a combination of nuance and erudition in any translation of this magnitude,” writes that readers “are solicited [remember this verb, which will return to embarrass us in just a moment] to inhabit Derrida’s language through the excellent translation that Gayatri Chakravorty Spivak has so ably and carefully provided once again,” leaving it entirely unclear why any revision was necessary at all.

Revisions to the original translation certainly were necessary: let’s look at some further examples to get a sense of what has changed in the new edition. Here, for example, is a famous passage, in fact the only block quote to appear in Butler’s Introduction, where it reads:

The movements of deconstruction do not destroy structures from the outside. They are not possible and effective, nor can they take accurate aim, except by inhabiting those structures. Inhabiting them in a certain way, because one always inhabits, and all the more when one does not suspect it. Operating necessarily from the inside, borrowing all the strategic and economic resources of subversion from the old structure, borrowing them structurally, that is to say without being able to isolate their elements and atoms, the enterprise of deconstruction always in a certain way falls prey to its own work.” (1976 and 97 editions, p. 24, quoted by Butler p. xx, who adds “est emportée” in parentheses after “falls prey”.)

Butler actually refers this passage to p. 25, which, in the new edition, is indeed what corresponds to p. 24 in the old, from which the passage is taken. Why the Introduction should be quoting from the earlier edition, while giving the page-reference to the “newly revised translation” is already quite mysterious. Especially when, turning to page 25 in the new edition, we find that Spivak has indeed revised the translation quite a bit compared to the version quoted by Butler. The passage now reads:

The movements of deconstruction are not interested in [solliciter] structures from the outside. They are not possible and effective, they cannot take accurate aim, except by inhabiting those structures. Inhabiting them in a certain way, because one always inhabits, and all the more when one does not suspect it. Operating necessarily from within, borrowing from the old structure all the strategic and economic resources of subversion, borrowing them structurally, that is to say without being able to isolate their elements and atoms, the enterprise of deconstruction is always in a certain way defeated by its own work.

This quite radically modified version (in which the verb solliciter goes from “destroy,” which we shall see it can (kind of) mean, to “be interested in,” which it really cannot) is not all, for Spivak is moved to quote the “same” passage again in her new Afterword, “newly revising” it again quite significantly, and this time giving:

The movements of deconstruction do not put a strain on [solliciter] structures from the outside. They are not possible and effective, they do not focus their strikes, except by inhabiting these structures. Inhabiting them in a certain way, for one always inhabits and more so yet when one does not suspect it. Operating necessarily from the inside, borrowing from the old structure the strategic and economic resources of subversion, borrowing them structurally, that is to say without being able to isolate their elements and atoms, the enterprise of deconstruction is always in a certain way swept away by its own work.

She immediately adds, in fact rather unhelpfully: “The new edition translates solliciter, ajuster leurs coups, and emporter differently in this passage.” Does she mean that the new edition translates those terms differently compared to the first edition? Or that the new edition translates them differently from the (apparently newer still) translation she has just given in the Afterword? And what of the other variations across the three versions, not to do with these three specific moments? All this is not made any clearer when, four pages later, she quotes again the first sentence of this passage, this time in the old version as quoted by Butler, “The movements of deconstruction do not destroy structures from the outside.”

Of these three versions, one has to say that the one that appears in the body of the newly revised edition is, by some distance, the least satisfactory. “Solliciter” is something of a signature term of Derrida’s, and, as the recently-published 1964-5 seminar on Heidegger confirms, is closely related to the thought of deconstruction itself. In that seminar, which of course Spivak could not have known when first translating the book (though at a pinch she might have seen the French edition published in the Fall of 2013 before writing her Afterword), Derrida sometimes uses “solliciter” to capture the Heideggerian sense of Destruktion. But he had also used it, and thematized it quite explicitly, in texts Spivak certainly had read when she wrote her initial Translator’s Preface, such as the essay “Force et signification” that opens L’écriture et la différence, and the famous essay “La différance” that opens Marges. Alan Bass (whose translations slightly postdate Spivak’s original Grammatology, and might themselves be due for some revision one day) chooses to keep the cognate terms (“solicit,” “solicitation”), with a note to the common etymology it shares with the French, which Derrida himself regularly glosses as a shaking movement (citare) of the whole (sollus) or the totality, and this is a decision followed by the translator of the 1964-5 seminar. (Spivak herself, elsewhere in the new edition, maintains the 1976 translation of sollicitation as “undoing,” with a translator’s note duly referring to Writing and Difference, but then changes that earlier “undoing” to “shaking up” on the very next page, and quotes this modified version in the Afterword.) So, by the standard of Derrida’s own descriptions, the new edition’s “interested in” in the passage Butler highlights in its earlier version hardly begins to capture what is at stake here. Nor, quite, does the “destroy” of the first edition’s version of this passage, which would need to be more explicitly linked to Heideggerian Destruktion for the reader to grasp the rather specialized sense here. But what seems clear in all this confusion is that, of the three different versions offered in this volume, it is the last, of entirely indeterminate status, from the Afterword, that is the most accurate, at least as regards this sentence. And indeed as regards the “emportée par” that Butler was moved to provide in French, signaling a difficulty with the original translation, and that “swept away” captures quite nicely, and certainly more accurately than the “defeated” used in the body of the text. (The translation of “ajustent leurs coups” as “take accurate aim” in the first two versions seems a little preferable to the slightly unidiomatic “focus their strikes” of the third, but nothing very important hangs on that.) The reader remains perplexed, of course, as to why the Afterword would offer a further revision of what one imagines will, unfortunately enough, become the “official” text now given on p. 25, rather than simply replacing that version with the newer one.

This is only one of the more surprising examples of changes in the new edition, that often, to say the least, do not seem to offer any clear improvement on the 1976 version. Here, for example, is a challenging sentence from near the beginning of chapter 1:

La dévaluation même du mot « langage », tout ce qui, dans le crédit qu’on lui fait, dénonce la lâcheté du vocabulaire, la tentation de séduire à peu de frais, l’abandon passif à la mode, la conscience d’avant-garde, c’est-à-dire l’ignorance, tout cela témoigne.

In 1976/97, the translation ran as follows:

The devaluation of the word “language” itself, and how, in the very hold it has upon us, it betrays a loose vocabulary, the temptation of a cheap seduction, the passive yielding to fashion, the consciousness of the avant-garde, in other words — ignorance — are evidences of this effect.

This is not easy to construe syntactically: one has to work quite hard to make the “how” mean something like “the fact of how,” and assume that the first dash should be displaced to before the “in other words” to make some sense of it. Even then it is not so easy to grasp what “this effect” is, in the absence of any obvious reference for the anaphoric “this,” which has no equivalent in the French. So it seems plausible that Spivak might want to make some revisions in the new edition, and indeed she does. Here is the new version:

The devaluation even of the word “language,” and everything, by way of the credit given to it, announces the looseness of its vocabulary, the temptation to seduce on the cheap, the fashionable passive abandon, the consciousness of the avant-garde, in other words its ignorance, all of this gives evidence.

It’s unclear how much progress has been made here: if anything, this is even harder to construe as a grammatical sentence. It seems as though the tout ce qui… dénonce is calling for a relative “that” that does not appear in the translation. And compared to the earlier version, “the consciousness of the avant-garde, in other words its ignorance” is difficult to bring into any real focus. And “the fashionable passive abandon” is now a straightforward misreading. Not to speak of the maintenance of the awkwardness of “gives evidence” replacing the equivalently awkward “evidences” of the earlier version. (In both cases Spivak does at least correctly recognize the possible link of témoigner to courts of law, and is not led astray into the realm of false friends, as she is elsewhere whenever Derrida uses the noun évidence, which means “evidence” only in the sense that something is or seems obvious, évident, and is usually best translated as “self-evidence”: see examples on pp. 14 and 18, and a funny moment where des évidences qui semblent depuis toujours aller de soi comes out in 1976 as “evidence which always seems self-evident” (p. 28) and in 2016 as “evidence which seems forever self-evident.”). Let me offer another possible translation of this sentence: “The very devaluation of the word ‘language,’ everything that, in the credit given to that word, betrays looseness of vocabulary, the temptation of a cheap seduction, passive giving-in to fashion, a sense of being avant-garde (in other words, ignorance) — all of this bears witness.”

If, then, there are no clear improvements in some of these changes, there are other cases where the new edition actually introduces wholly new errors that were simply not there in the first version. For example, still in Chapter 1 (the only one I have checked in any detail against the French), Derrida argues in a preliminary way that writing, traditionally described as “signifier of the signifier,” has gone beyond its traditional determination and begun to comprehend language itself, overflowing its traditional bounds and in so doing effacing them. All signifieds are now drawn into the play of referrals (renvois: Spivak, as discussed earlier on, has “references” in both editions), and this makes of speech or voice merely a guise or disguise of a more primordial writing (in the new edition Spivak unaccountably — but quite deliberately (see Afterword, p. 368) — adds the word “vowel” in brackets at least a dozen times where Derrida uses the word voix, which in fact never means “vowel,” for which French has the perfectly good word voyelle). Starting a new paragraph, Derrida proceeds as follows:

Ces déguisements ne sont pas des contingences historiques qu’on pourrait admirer ou regretter. Le mouvement en fut absolument nécessaire, d’une nécessité qui ne peut comparaître, pour être jugée, devant aucune autre instance.

In the first version, Spivak has the following, which simply pretends the words comparaître and pour are not there, but manages to hold on to the general sense in spite of that omission:

These disguises are not historical contingencies that one might admire or regret. Their movement was absolutely necessary, with a necessity which cannot be judged by any other tribunal.

Presumably wanting to account for the words she skipped, Spivak now has, introducing a blunder other translators of Derrida have also been known to make:

These disguises are not historical contingencies that one might admire or regret. Their movement was absolutely necessary, with a necessity that cannot compare, in order to be judged, to any other instance.

But comparaître really is not the same verb as comparer: far from meaning “compare,” comparaître means to “compear,” to appear before the law, before a tribunal, so it should be “[…] with a necessity that cannot appear, in order to be judged, before any other authority.”

This is still, perhaps, a touch specialized, and there were those missing words to account for after all. But other revisions alter perfectly reasonable solutions from the earlier edition for no apparent reason at all. For example, a little later, in normal everyday French, Derrida writes of “tout ce qui peut donner lieu à une inscription en général,” which Spivak had translated only a little loosely in 1976 as “all that gives rise to an inscription in general” (p. 9: more literally, “all that can give rise […]”). Now, perhaps overreading the lieu and wanting to make something thematic of it, she puts the unidiomatic, and indeed misleading, “all that can make room for an inscription in general.”

Spivak is, then, not shy about revising her earlier translation, has done so often, and often for the worse. Some random sampling of the rest of the translation suggests that Chapter 1 is certainly not an outlier in this respect: for example on p. 72, the defensible “but if anticipation were privileged” (for mais à privilégier l’anticipation) is changed to the incomprehensible “anticipation privileging.” On p. 171, in a paragraph where the awkward “opinionatedly” is maintained as a translation of opiniâtrement (“stubbornly” or “obstinately” are simple and good solutions here), the (admittedly a little awkward) verb arraisonner, from maritime vocabulary where it means to stop a ship and ascertain its origin and destination, is changed from the (defensible) “arrest” (“called to account” or even in an effort to hang onto the raison embedded in it, “asked to state its reasons” might be preferable) to the incomprehensible verb “show-cause.” On p. 162, still about reason, the tricky phrase la raison raisonnable, sinon raisonneuse, originally translated a bit obscurely as “the reason which is reasonable, if not reasoning,” becomes, very strangely, “reasonable, if not resonating [raisonneuse] reason.” (“Raisonneuse” means something like ratiocinating in the sense of quibbling or splitting hairs, and may resonate (résonner) phonetically with raisonner, but does not for all that mean “resonating.”) Most strikingly, perhaps, in the famous “Question of Method” section, the translation of the important claim that malgré quelques apparences, le repérage du mot supplément n’est ici rien moins que psychanalytique has been catastrophically changed from the earlier “in spite of certain appearances, the locating of the word supplement is here not at all psychoanalytical,” to the entirely misleading “the location of the word supplement is nothing short of psychoanalytic.” The distinction between rien moins que (“anything but”) and rien de moins que (“nothing less than”) is a known writer’s and translator’s pitfall, and Spivak here goes from the right reading to the wrong in a way that might sum up the entire story of the new edition. Many translations from Rousseau are equally hard to understand: where the earlier version had, “Forced to provide for winter, people living under such conditions have to establish some sort of convention among themselves in order to help each other,” which does lose the stylistic flourish of Rousseau’s twice using voilà in the original, the revision again makes it almost incomprehensible: “Forced to provide for winter, people living under such conditions have to help each other see how they are constrained to establish some sort of convention among themselves.” The last sentence of the Afterword defends in general modifications of published translations of the authors Derrida cites, in the name of Derrida himself and Spivak’s “private piety in memoriam.” My point here is of course not against the fact of modifying existing translations, nor against private piety, but against modifying those translations for the worse.

This does not of course mean that the original translation was not in need of revision: but if Spivak has often “corrected” passages that were not really in need of correction, she has elsewhere left many problems untouched. Opening the book entirely at random on p. 122, for example, I find “knowledgeably put in place” maintained for savamment mis en place (here savamment means, rather, “skillfully” or “adroitly”), and irruption éthnographique is still translated as “anthropological eruption” (an irruption is a bursting in, not a bursting out); opening it at random again on p. 236, pour n’être pas simplement apparentes, ces contradictions is still translated as “in order to be not merely apparent,” whereas it should be “while not being merely apparent” or “although not merely apparent.” There are also many recurrent decisions that would need to be revised: for example, appartenance, belonging (as in the belonging of the concept of the sign to metaphysics), is at least a dozen times still translated as “appurtenance,” which (clever post-translation points about the supplement aside) it really does not mean: on every page there are inaccuracies and infelicities that one might, indeed, have hoped to see corrected here.

Indeed. The word “indeed” is, usually, the best solution for the standard French en effet, which confirms a prior statement or expectation (Il a dit qu’il serait là, et en effet il était là, “He said he would be there, and indeed he was there”), whereas en fait, usually best rendered as “in fact” (“Il a dit qu’il serait là, mais en fait il n’était pas là,” “he said he would be there, but in fact he was not there”) often contradicts an expectation. Spivak never once translates en effet as “indeed,” preferring the standard false friend “in effect,” “in fact” or even “as a matter of fact”: where she does use “indeed” on p. 96 (“and, indeed, one cannot say,” unchanged from the earlier edition), she is translating “on ne peut même pas dire […],” better rendered as “one cannot even say […].” We could make similar points about other idiomatic phrases, such as sans doute, which in Modern French is almost always used concessively and prepares for a following mais (just as English “no doubt” usually is not proclaiming absence of doubt at all, but making a preliminary concession preparing for an important “but,”) so that translating sans doute as “without a doubt” is almost always too affirmative.


Spivak, we said, is not shy about revising her own earlier version, often for the worse, and she is not shy about revising others’ translations either. In her curious Afterword — which is curious from its very first sentence: “The Translator’s Preface to this book, written in 1972–73, gives a general background to Jacques Derrida, who at the time was at the beginning of a brilliant career and tried to say a few things about the book Of Grammatology,” but maybe the whole editing process has to take some responsibility here — where among other things she draws a number of parallels and contrasts between the Grammatology and Derrida’s last published book Rogues, she is moved to modify not only her own translation of the important quote about the “movements of deconstruction” discussed earlier, and to quote her own translations inconsistently elsewhere,[4] but she also modifies (almost) every quotation she takes from Pascale-Anne Brault and Michael Naas’s translation of Rogues [5]. Of course it is common enough to need on occasion to modify a published translation — even one’s own translation — depending on the specific focus of discussion or simply the discovery of inaccuracies. Spivak might even, I suppose, have simply said she herself had retranslated all the passages from Rogues if that’s what she preferred to do for some reason. But as things stand, her persistent use of “translation modified” can only give the impression that the Brault-Naas translation is somehow almost always unsatisfactory. Let’s check a few of those modifications. On p. 355, all that she has done to a passage from their translation is remove a single comma. Elsewhere (see for example p. 361, where the page-reference is in fact to the French edition, though nothing signals that fact), one assumes that Spivak has in fact simply translated the passage herself. Sometimes the modifications seem fairly anodyne (and to that extent unnecessary): sometimes, however, as was the case with Spivak’s self-modifications, they simply introduce problems and errors that the Brault-Naas version had avoided. For example, on p. 363, having provided an incorrect bracketed insert (she has “implications [of Muslim theocracies] in the geopolitics […],” whereas it should be “the implication [in the singular: Brault and Naas have “the role it plays”] [of Saudi Arabia] in the geopolitics […]”), she goes on to provide an apparently meaningless sequence: “On the other side, all the nation-states deeply linked, if not in their constitution, at least in their culture, one time to Jewish (there’s only one, Israel) or Christian […],” where I can only assume (a little incredulously, it is true) that she misread une foi juive, which is what Derrida wrote, as une fois juive, and never untangled the result (Brault and Naas have, quite accurately, “On the other side, all the nation-states fundamentally linked, if not in their constitution at least in their culture, to a Jewish faith (there’s only one, Israel) or Christian faith […].” A little later, quoting Derrida describing as “categorically political” (he writes expressément politique, translated by Brault and Naas as “explicitly political”), she prefers to translate Derrida’s call for une interprétation de l’héritage coranique qui y fasse prévaloir, comme du dedans, les virtualités démocratiques as “an interpretation of the Quranic heritage that will show the prevalence, as from inside, the democratic virtualities […].” Quite apart from making syntactic sense on its own account only if we read “the democratic virtualities” quite improbably as in apposition to “the prevalence” (which is in any case not what the French says), this also misreads the sense of the subjunctive fasse as though it were a future tense, and the meaning of the verb faire prévaloir, which does not mean to show any prevalence, but to bring out, to have or make prevail. Brault and Naas have, much more accurately and fluently, “an interpretation of the Koranic heritage that privileges, from the inside as it were, the democratic virtualities[…].”

Apart from these unfortunate “corrections,” the Afterword, which is, as I mentioned, much harder to follow than the original Translator’s Preface from 1976, offers a number of claims about Derrida’s thinking and its history from the Grammatology to Rogues. As always, Spivak’s heart here is absolutely in the right place, as she establishes a continuity of concern in Derrida about ethnocentrism, for example, and finds interesting ways of describing the reading practices more or less inaugurated in the Grammatology (“funky,” for example, in the musical sense of the term). But many strange things also happen along the way, including (this quite on a par with Butler’s blunder about Saussure and grammatology) a claim to provide Kant’s definition of the transcendental via a quotation that is in fact Kant’s definition of the (confusingly perhaps, but importantly) quite different notion of the transcendent. Even what Spivak’s own Critique of Postcolonial Reason called “a scrupulous travesty” of Kant might not be able to justify an error such as this.


Towards the end of her Introduction, Butler picks up a thought with which she began, about what she now calls “two groups of skeptics”:

the first to ask whether the text is readable (in any language); the second, whether the text is fairly translated into English. Those who question its general readability are often considered to be unwilling to do the work. Those who quarrel about its translation tend to do a great deal of work, and want more work to be done. The first group tends to defend its own metaphysical grid of intelligibility; the second wants to make sure that the challenge to that very grid of intelligibility is delivered in precise textual ways. Police functions can flare up on either side.

Butler thinks she can appeal to Benjamin to rise above what she presents here as an “entrenched battle,” and indeed she has already surreptitiously dialectized and sublated it by making it seem on the one hand as though these two groups are immediately antagonistic opposites, and on the other as though they really basically come down to the same thing (the police). But doing the work and wanting the work to be done as well as it can be done are not, intrinsically, “police functions” at all, nor do they intrinsically constitute “quarrelling.” Translation is, of course, impossible, an experience of the impossible, as Derrida would be the first to point out, and involves a principle of ruin, as Butler recalls. But that does not mean that all translations are equivalently unsatisfactory, or that mistakes cannot often be identified and corrected. Nor does it mean that Derrida could not complain (in the recent 1964-5 Heidegger course, for example), that a given mistranslation can transform the text into a bouillie pour les chats (p. 35) or, as the translator chose to render it, a pig’s breakfast.


[1] La voix et le phénomène, translated by David Allison as Speech and Phenomena (Evanston: Northwestern University Press, 1973), is a detailed reading of Husserl’s theory of the sign. The book was retranslated by Leonard Lawlor as Voice and Phenomenon (Evanston: Northwestern University Press, 2010).

[2] “On what conditions is a grammatology possible? Its fundamental condition is certainly the shaking up [sollicitation] of logocentrism. But this condition of possibility turns into a condition of impossibility. In fact it risks destabilizing the concept of science as well. Graphematics or grammatography ought no longer to be presented as sciences; their goal should be exorbitant when compared to grammatological knowledge.” (p. 80)

[3] The revised edition uses the more literal but less idiomatic “Epoch of Rousseau,” justified a little confusingly in the Afterword by reference to a supposedly “‘spatial’ sense of the reducing-out technique proposed by Edmund Husserl, called epokhē.” And adds, culinarily, but perhaps forgetting the need for further clarification, “In cooking this is called ‘rendering’. One might say that this is a rendered Rousseau, as well as the era of Rousseau” (p. 353).

[4] So that the passage beginning “the names of authors or of doctrines…” beginning at the bottom of p. 353 is actually a direct quotation from Derrida, as is clearer when it reappears later in the Afterword (p. 360), with however, an inconsistency about the use of singular or plural first-person pronouns to which I alluded above.

[5] Brault and Naas are seasoned Derrida scholars and translators, and are part of the Derrida Seminars Translation Project (DSTP) team, formed in 2007, along with Geoffrey Bennington, Peggy Kamuf, Elizabeth Rottenberg, and David Wills, to oversee translation of Derrida’s previously unpublished courses and seminars. DSTP has written collectively to Johns Hopkins University Press asking that this new edition of the Grammatology be withdrawn, pending appropriate revision.


Geoffrey Bennington is the Asa G. Candler Professor of Modern French Thought at Emory University. He is the author of several books on Derrida and translator of many others by him, and he is co-editor of The Seminars of Jacques Derrida series.