---

10. "A Self-Unsealing Poetic Text"

Poetics and Politics of Witnessing

Jacques Derrida Translated by Rachel Bowlby

The world becomes its language and its language becomes the world. But it is a world out of control, in flight from ideology, seeking verbal security and finding none beyond that promised by a poetic text, but always a self-unsealing poetic text.

MURRAY KRIEGER,A Reopening of Closure: Organicism Against Itself

[I]t is the role of art to play the unmasking role—the role of revealing the mask as mask. Within discourse it is literary art that is our lighthouse.…It would seem extravagant to suggest that the poem, in the very act of becoming successfully poetic—that is, in constituting itself poetry—implicitly constitutes its own poetic. But I would like here to entertain such an extravagant proposal.

MURRAY KRIEGER,Ekphrasis: The Illusion of the Natural Sign

Signing, sealing, divulging, unsealing. This will be about bearing witness.^[1] And about poetics as bearing witness—but testamentary witness: attestation, testimony, will. A poem can "bear witness to" a poetics. It can promise it, it can be a response to it as to a promise in a will or testament. Indeed it must, it cannot not do so. But not with the idea of applying a previously existing art of writing, nor of referring to one as to a charter written somewhere else, nor of obeying its laws like a transcendental authority, but rather by itself promising, in the act of its happening, the foundation of a poetics. Thus it would involve the poem "constituting its own poetics," as Krieger puts it, a poetics which must also, as if across its generality, become, invent, institute, offer for reading, in an exemplary way, signing it, both sealing and unsealing it, the possibility of this poem. This would come about in the event itself, in the verbal body of its singularity: at such and such a date, at the both unique and repeatable moment of a signature which opens the verbal body onto something other than itself, in the reference which carries it beyond itself, towards the other or towards the world.

As testimony to warm gratitude, I would therefore like to take a certain risk in my turn, so as to share it with Murray Krieger—the risk of "entertain[ing]

---

such an extravagant proposal." And to try it out, I would like to put to the test this experience of bearing witness. Wanting to recall places where, for over ten years, I have enjoyed living near Murray Krieger, I chose to go back over some texts by Paul Celan that I read with my students at the University of California, Irvine, for a course of seminars about witnessing. And especially about responsibility, when it involves a poetic signature, at a particular date. Hypothesis to be verified: all responsible witnessing involves a poetic experience of language.

I

Without giving up thinking about the secret in the perspective of responsibility, far from it, how is it that one has to come up against the question of witnessing (testimonium)? And why is the question of testimonium no different from that of the testamentum, of all the testaments, in other words of surviving in dying, of surviving, living on, before and beyond the opposition between living and dying?

	Aschenglorie […]
	…
	Niemand
	zeugt für den
	Zeugen
	Ash-glory […]
	…
	No one
	bears witness for the
	witness.[2]

If we want to keep for them that poetic resonance to which, already, on the page, they mean to respond, we must remember that these words come to us in German. As always, the idiom remains irreducible. This invincible singularity of the verbal body already introduces us into the enigma of witnessing, beside that irreplaceability of the singular witness which indeed may well be what this poem is speaking to us about. It thus speaks of itself, signifying itself in speaking to the other about the other, signing and designing itself in a single gesture—"sealing and unsealing itself"—or again, to quote and displace a little the words of Murray Krieger: sealing while (by, through) unsealing itself as a poetic text.

This idiom is untranslatable, ultimately, even if we translate it. These three lines resist, and resist even the best translation. What's more, they come to us at the end of a poem which, however little certainty there may be about its sense, about all its senses and all its possible meanings, it is difficult not to think of as also related, through an essential reference, to dates and events,

---

to the existence or the experience of Celan. The poet is the only one who can bear witness to these "things" which are not only "words," but he does not name them in the poem. At any rate the possibility of a secret always remains open, and this reserve is inexhaustible. It is more so than ever in the case of the poetry of Celan, who continually encrypted ("sealing, unsealing") these references. Some people have been able to bear witness—just so—to this, like Peter Szondi, friend and reader-interpreter of Celan, who shared at least some of his experiences, not that such testimony exhausts or, crucially, proves what he speaks of, far from it. This poem also remains untranslatable to the extent that it can be related to events of which the German language will have been a privileged witness, namely the Shoah, which some call by the proper name (and metonymic name—a huge problem which I leave hanging here) of "Auschwitz." The German language of this poem will have been present at everything that could destroy by fire and reduce to "ashes" (Aschenglorie, Ash-glory, is the first word of the poem, a word double and divided) existences of innumerable number—innumerably. Innumerably but also unnameably, thus incinerating, with the name and the memory, even the assured possibility of bearing witness. And since I have just said "the assured possibility of bearing witness," we will have to ask whether the concept of bearing witness is compatible with a value of certainty, of assurance, and even of knowing as such.

Ashes are that which annihilates or threatens to annihilate even the possibility of bearing witness to annihilation. Ashes are the figure of annihilation without remains, or without a readable or decodable archive. Perhaps that would lead us to think of this fearful thing: the possibility of annihilation, the virtual disappearance of the witness, but also of the capacity to bear witness, that is, of what would be the only condition of bearing witness, its only condition of possibility as condition of its impossibility—paradoxical and aporetic. When bearing witness seems assured and so becomes a demonstrable theoretical truth, when there is a piece of information or a report, a procedure of proving or even an exhibit in a trial, it risks losing its value, its sense or its testimonial status. That comes down to saying—always the same paradox, the same paradoxopoetic matrix—that as soon as it is assured, certain as a theoretical proof, a testimony cannot be assured as testimony. For it to be assured as testimony, it cannot, it must not be absolutely certain, absolutely sure and certain in the order of knowing as such. This paradox of "as," "as such" (comme tel) is the paradox we can experience—and there is nothing fortuitous about this—apropos of the secret and responsibility, of the secret of responsibility and the responsibility of the secret. How can a secret be shown as secret? To take up Murray Krieger's words again, how can a mask be revealed as mask? And in what way might a poetic work be called upon to put this strange operation to work?

So it is necessary first to hear these lines in their language, and to see

---

them in their space. Necessary out of respect for their spacing but first of all because the spaced writing of this language does not admit of translation into a simple speech, French or English. We see indications already of the poignant question of untranslatable testimony. Because it must be linked to a singularity and to the experience of an idiomatic mark, for example that of a language, testimony resists the test of translation. It thus risks not even being able to cross the frontier of singularity, if only to deliver its meaning. But what would be the worth of an untranslatable testimony? Would it be a non-testimony? And what would a testimony that was absolutely transparent to translation be? Would it still be a testimony?

Ashes, we were saying, annihilate or threaten to annihilate even the possibility of testifying to annihilation itself.

It happens that Celan's poem has as its title its own incipit. Thus its first line speaks of ashes, and it appears reasonably open to translation. It is Aschenglorie, in a single word, which André du Bouchet translates, in three words, as "Cendres-la gloire," and Joachim Neugroschel in two words, as "Ashglory." Word-for-word translation is already impossible. Infidelity has begun, and betrayal and perjury, from the very threshold of this arithmetic, with this accountability of the incalculable. The poetic force of a word remains incalculable, all the more certainly so when the unity of a word ("Aschenglorie") is that of an invented composition, the inauguration of a new body. All the more certainly so when the birth of this verbal body gives the poem its first word, when this first word becomes the verb which comes at the beginning. En arkhe en o logos [In the beginning was the word]. And if this logos was a light, for John, here it was a light of ashes. In the beginning was (the word) "Aschenglorie."

This glory of the ashes, this glory of ash, this glory which is that of the ashes but which is also of ash, in ash—and glory is minimally the light or the shining brightness of fire—here it is lighting a poem that I shall not even attempt to interpret with you. Light is also knowing, truth, meaning. Now this light is no more than ashes here, it becomes ashes, it falls into ashes, as a fire goes out. But (it is the mobile and unstable articulation of this "but" which will be important for us) ashes are also glory, they can still be re-nouned, sung, blessed, loved, if the glory of the re-noun is not reducible either to fire or to the light of knowing. The brightness of glory is only the light of knowing [connaissance] and necessarily the clarity of knowledge [savoir].

Why not even attempt to interpret this poem? I am quite willing to try to explain this limit. What matters is not what this poem means, or that it means, or that it bears witness to this or that, nor even that it names and what it names—elliptically as always. Ellipsis and caesura and the cut-off breathing presumably mean here, as always in Celan, that which seems most decisive in the body and in the rhythm of the poem. A decision, as its name indicates, always appears as [comme] interruption, it decides as [en tant que] a tearing cut.

---

What counts, then, is not the fact that the poem names some themes which we know in advance must be at the heart of a reflection on responsibility, witnessing, or poetics. What matters most of all is the strange limit between what can and cannot be determined or decreed in this poem's witnessing to witnessing. For this poem says something about bearing witness. It bears witness to it. Now in this witnessing to witnessing, in this apparent meta-witnessing, a certain limit makes meta-witnessing—that is, absolute witnessing—both possible and impossible.

Let us try to put ourselves in the region of this limit, at the passage of this line. We will be guided by a hypothesis: this line is perhaps also that line of necessary "extravagance" which Murray Krieger speaks of.

We have just alluded to some motifs which are in some way signalled by this poem, and which we know in advance intersect at the heart of the questions of responsibility, of the secret, of witnessing.

What then are these motifs? Well, one example would be the three, the figure of everything that takes itself beyond two, the duo, the dual, the couple. Now three is named twice, in the first stanza and close to the final stanza which is the one that names the Aschen (Aschenglorie, to repeat, in a single word in the first line, but Aschen-glorie, cut by a hyphen across two lines, near the end). Both times, there is a tripleness affecting both the road (Weg) and the hands (Händen), the knotted hands (we should also keep hold of the knot, the knotting of the link and the hands).

	ASCHENGLORIE hinter
	deinen erschüttert-verknoteten
	Händen am Dreiweg.

Let us quote both French and English translations; they are not wholly satisfying, but no one can teach a lesson to anyone else here, by definition:

	CENDRES-LA GLOIRE revers
	de tes mains heurtées-nouées pour jamais
	sur la triple fourche des routes.

	Ash-glory behind
	your shaken-knotted
	hands on the three-forked road.

It would also be possible to translate into French as follows:

	Gloire pour les cendres, derrière
	tes mains défaites effondrées—toutes nouées
	à la fourche des trois voies.
	[Glory for ashes, behind
	your demolished collapsed hands—all knotted
	at the fork of the three ways.]

---

I am not happy with this "pour les cendres" ("for ashes"), for it is just as much about the glory of ashes as the glory promised to ashes; and if, as I considered doing, we translated as "gloire aux cendres" ("glory to ashes"), it would be necessary to understand not only the glorification of ashes, but, as one might say of a still life, the figure of glory surrounded by ashes, on a background or an ornament of ashes. Which amounts to so many ways of noting the poetic stroke of genius in this untranslatable "Aschenglorie." Untranslatable the word remains, word for word, one word for the other, where the composed vocable does not decompose. For in the original version it is not divided, as it will be lower down, near the end, disarticulating and unifying itself to itself, this time, at the end of the line, by a strange hyphen. A hyphen of this kind is also an act of poetical memory. It points out the beginning in return; it gives a reminder of the initial undividedness of Aschenglorie:

	Aschen
	glorie hinter
	euch Dreiweg
	Händen.
	Cendres
	la gloire, revers
	de vous—fourche triple,
	mains.
	Ash
	glory behind
	your three-forked
	hands.

One could also translate another way:

	Gloire
	de cendres derrière vous les mains
	du triple chemin.
	[Glory
	of ashes behind
	you hands
	of the triple road.]

Thus "euch" (vous, you) has just replaced the "deinen" of the second line ("deinen erschüttert-verknoteten/Händen am Dreiweg," yours, your hands, the hands which are yours). The addressee of the apostrophe has gone into the plural. At any rate, it is no longer simply the same, it is no longer reducible to the being in the singular, masculine or feminine, to whom the first stanza is addressed. The two stanzas turn, they turn round, as a stanza [strophe] and an apostrophe always do. The two stanzas apostrophize more than one addressee.

---

They turn round from one to the other, they turn themselves away from the one to others.

Why point out, if not more than that, this allusion to the three, whether in connection with road (Dreiweg) or hands (Dreiweg/ Händen)? Because in fact we will soon be coming up against this motif of the third in the scene of possible/impossible testimony, of bearing witness that is possible as impossible. In its Latin etymology, the witness (testis) is someone who is present as a third person (terstis). We would have to look very closely at this to understand what it might imply. Testis has a homonym in Latin. It usually occurs in the plural, to mean "testicles." It even happens that Plautus plays on the word in Curculio, and exploits its being a homonym. Testitrahus means both complete and male, masculine. Some feminists, men or women, couldn't deprive themselves, enjoying or unjoying [s'en jouant ou sans jouer], of deriving from this an argument about the relations between a certain way of conceiving of the third person and bearing witness, on the one hand, and the chief, the head and phallocentric capital on the other. It is true that, in English, testis, testes has kept the sense of testicle—which could be an incitement to militancy.

In his Vocabulaire des institutions européennes [Dictionary of European Institutions], in the chapter on "Religion and Superstition," Benveniste analyzes a word, "superstes," which can mean "witness" in the sense of survivor: someone who, having been present then having survived, plays the role of witness. Benveniste makes an association between superstes and testis but also distinguishes them:

We can see the difference between superstes and testis. Etymologically, testis is someone who is present as a "third" (terstis) at a transaction where two persons are concerned; and this conception goes back to the Indo-European period of civilization.^[3]

As always, Benveniste analyzes the etymology by following the line of a genealogical recollection which goes back to institutions, customs, practices, arrangements. In this valuable but profoundly problematic work which, as we see, wants to be a "dictionary of Indo-European institutions," the words are selected and then placed in a network according to the institutional figures of which they are also assumed to be, precisely, the witnesses. The words bear witness to the institutions; the vocabulary attests to an institutional sense. But even if we suppose that the sense exists before and outside these words (an improbable or virtually meaningless hypothesis), it is at any rate certain that the sense does not exist without these words, which is to say without that which bears witness to it, in a sense of bearing witness which still remains highly enigmatic, but inescapable here. If the words bear witness to a usage and an institutional practice,^[4] the paradox here is concentrated in the analysis of the word testis, terstis, which attests, with regard to knowledge, thus

---

making way for assumed knowledge, to an institution or a practice, a social organization, a "conception"—that is Benveniste's word—which, he says, "goes back to the Indo-European period of civilization." In order to illustrate, in reality to establish this filiation, to prove this fact, Benveniste adds:

A Sanskrit text says: "Every time two persons are present, Mitra is there as a third"; thus the god Mitra is by nature the "witness." But superstes describes the "witness" either as the one who "subsists beyond," witness at the same time as survivor, or as "the one who holds himself to the thing," who is present there. We see now what can and must be meant theoretically by superstitio, the function of the superstes. This will be the "property of being present" as a "witness."

Benveniste's statements here open onto the larger context that we could reconstitute, in particular around superstes, the survivor determined as witness, and around testis, terstis, determined as a third. The witness is the one who will have been present. He or she will have attended, in the present, the thing of which he is witness. Every time, the motif of presence, of being-present or of being-in-presence, turns out to be at the center of these determinations. In Le Différend, a book in which the question of the witness plays a large role, Jean-François Lyotard approaches this question of the witness as third person a number of times, without reference to Benveniste or to Celan. But by privileging the example of Auschwitz and the debate around "revisionism" (which is naturally a debate about the status of testimony or bearing witness and of survival), he problematizes the idea of God as absolute witness.^[5]

Quite obviously, we must take into account an undeniable fact: like the institutions to which it is thought to refer, which it ought to reflect, represent, or incarnate, Latin semantics (testis, terstis, superstes) only denotes one etymologico-institutional configuration among others—and even one among others for "us," assuming that we can say "we" Westerners. It is not, for example, to be found in German as well. The family Zeugen, bezeugen, Bezeugung, Zeugnis, translated as "witness," "to bear witness," "testimony," "attestation," belongs to a completely different semantic network. One would be hard put, in particular, to find in it an explicit reference to the situation of the third, or even to presence. In the family of what we will not risk simply calling homonymy are all the words we have just read in the Celan poem (Zeug, Zeugen, Zeugung). Elsewhere, they also mean tool, procreation, engendering, and indeed generation—both biological and familial. After what the word "témoin" [witness] (terstis, testis) bears witness to [témoigne]by its supposed genealogy, we have what is also witnessed to by the word Zeugen in its supposed genealogy or generation, etc. If we take account of the witness as terstis superstes, as surviving third, and even as inheritor, guardian, guarantee, and legatee of the will and testament, ultimately of what has been and has disappeared, then the crossover between on the one hand a genealogical or generational

---

semantics of Zeugen, and on the other hand the semantics of terstis superstes, becomes fairly vertiginous.

Crossover of a vertiginous filiation, yes, perhaps. But it is vertigo which turns our head, vertigo in which we are going to turn and let ourselves be turned round, and not only between the tropes and stanzas of Celan.

In English, the Latin root does of course remain, with testimony and to testify, attestation, protest, testament. So it articulates together for us the two themes of survival and of witnessing, etc. But the family of witness and bearing witness is quite different, presumably opening out onto the aspect of seeing, and thus towards another semantic and poetic space in the final words of Celan's poem in translation: "No one/bears witness for the/witness." Finally—but this is where we should have begun—Greek makes no explicit reference to the third person, to surviving, to presence, or to generation: martus, marturos, the witness, who will become the martyr, the witness to faith, does not literally imply any of these values (third, surviving, presence, generation). Marturion means, to follow the institutional usage, "testimony," but also evidence or "proof."

Here we touch on a sensitive and heavily problematic distinction between testifying, bearing witness [témoignage], the act or experience of witnessing as "we"^[6] understand it, and, on the other hand, proof; between testifying and, on the other hand, theoretical-constative certainty. This is both an essential distinction and one that is in principle insurmountable. But in practice the confusion always remains possible, so fragile and easily crossed can the limit sometimes appear, and whatever language and word is used. For this is not reserved for the Greek marturion alone: the Latin testimonium—bearing witness [témoignage], giving evidence, attestation—can come to be understood as proof. So language cannot of itself alone, as a lexicon or dictionary would do, be guardian and guarantee of a usage. A pragmatic slippage from one sense to the other, sometimes in the passage from one sentence to the other, can always occur. We should ask for what necessary—not accidental—reasons the sense of "proof" regularly comes to contaminate or divert the sense of "bearing witness." For the axiom we ought to respect, it seems to me, even though it may be problematized later, is that bearing witness is not proving. Bearing witness is heterogeneous to the administration of a legal proof or the display of an object produced in evidence. Witnessing appeals to the act of faith with regard to a speech given under oath, and is therefore itself produced in the space of sworn faith ("I swear to speak the truth"), or of a promise involving a responsibility before the law, a promise always open to betrayal, always hanging on this possibility of perjury, infidelity, or abjuration.

What does "I bear witness" mean? What do I mean when I say "I bear witness" (for one only bears witness in the first person)? I mean not "I prove," but "I swear that I have seen, I have heard, I have touched, I have felt, I have

---

been present." That is the irreducibly perceptible dimension of presence and past presence, of what can be meant by "being present" and especially "having been present," and of what that means in bearing witness. "I bear witness"—that means: "I affirm (rightly or wrongly, but in all good faith, sincerely) that that was oris present to me, inspaceandtime(thus, perceptible), and although you do not have access to it, not the same access, you, my addressees, you have to believe me, because I am committed to telling you the truth, I am already committed to it, I tell you that I am telling you the truth. Believe me. You have to believe me."

The addressee of the witnessing, the witness of the witness, does not see for himself what the first witness says he has seen; the addressee has not seen it and never will see it. This direct or immediate non-access of the addressee to the object of the witnessing is what marks the absence of this "witness of the witness" to the thing itself. This absence is therefore crucial. It thus hangs on the speech or the mark of testimony inasmuch as it is dissociable from what it bears witness to: the witness is not present either, of course, present in the present to what he recalls, he is not present in the mode of perception, inasmuch as he bears witness, at the moment when he bears witness; he is no longer present, now, to what he says he was present to, to what he says he perceived; even if he says he is present, present in the present, here, now, by what is called memory, memory articulated in a language, to his having-been-present.

II

This "you have to believe me" must be understood. "You have to believe me" does not have the sense of the theoretico-epistemic necessity of knowledge. It is not presented as a convincing demonstration, the result of which is that you cannot not subscribe to the conclusion of a syllogism, to the connections of an argument, or even to the manifestation of something present. Here, "you have to believe me" means "believe me because I tell you to, because I ask it of you," or, equally well, "I promise you to speak the truth and to be faithful to my promise, and I commit myself to being faithful." In this "it is necessary to believe me," the "it is necessary," which is not theoretical but performative-pragmatic, is as determining as the "believe." Ultimately, it is perhaps the only rigorous introduction to the thought of what "to believe" might mean. When I subscribe to the conclusion of a syllogism or to the delivery of a proof, it is no longer an act of belief, even if the one who conducts the demonstration asks me to "believe" in the truth of the demonstration. A mathematician or a physicist, a historian, as such, does not seriously ask me to believe him or her. He does not appeal in the last analysis to my belief, at the moment when he presents his conclusions. "What is believing?"—what are we doing when we believe (which is to say all the time, and

---

as soon as we enter into relationship with the other): that is one of the questions that cannot be avoided when one tries to think about bearing witness.

In spite of the examples invoked to begin to make things a little clearer, bearing witness is not altogether and necessarily discursive. It is sometimes silent. It has to involve something of the body which does not have the right of speech. We should thus not say, or believe, that bearing witness is entirely discursive, thoroughly a matter of language. But we will not call something bearing witness that is not open to the order of the "comme tel," of the present or having-been-present "comme tel," "en tant que tel," of the "as such" or of this which indeed Murray Krieger stresses—as truth itself, the truth of the lie or the simulacrum, the truth of the mask—in the sentence quoted in the epigraph ("the role of revealing the mask as mask").

This "as such" is presupposed by language, unless on its side it presupposes at least the possibility of a mark, or a prelinguistic experience of the mark or the trace "as such." This is where the whole formidable problem of the apophantic opens up—of the as such, of presence and language. We will not enter into it directly here in its own right.^[7]

Whoever bears witness does not bring a proof; he is someone whose experience, in principle singular and irreplaceable (even if it can be crosschecked with others to become proof, to become conclusive in a process of verification) comes to attest, precisely, that some "thing" has been present to him. This "thing" is no longer present to him, of course, in the mode of perception at the moment when the attestation happens; but it is present to him, if he alleges this presence, as re-presented in the present in memory. At any rate, even if—something unusual and improbable—it was still contemporary to the moment of the attestation, it would be inaccessible, as perceived presence, to the addressees who receive the testimony. The witness marks or declares that something is or has been present to him, which is not so to the addressees to whom the witness is joined by a contract, an oath, a promise, by a pledge of sworn faith whose performativity is constitutive of the witnessing and makes it a pledge [gage], an engagement. Even perjury presupposes this sworn faith which it betrays. A perjurer does indeed threaten all witnessing, but this threat is irreducible in the scene of sworn faith and attestation. This structural threat is both distinct and inseparable from the finitude which any witnessing also presupposes; for any witness can make a mistake in good faith, he can have a limited, false perception, one that in any number of ways is misleading about what he is speaking about; this kind of finitude, which is just as irreducible and without which, also, there would be no place for bearing witness, is nonetheless other, in its effects, than the kind which makes it obligatory to believe and makes lying or perjury always possible. There are thus two heterogeneous effects here of the same finitude, or two essentially different approaches to finitude: the one which goes by way of error or hallucination in good faith, and the one which goes by

---

way of cheating, perjury, bad faith. Both must always be possible, at the moment of bearing witness.

But the very possibility of lying and perjury (lying being a kind of perjury) attests that for us, bearing witness, if there is such a thing, gains a sense only before the law, before the promise, the commitment. It has a sense only in regard to a cause: justice, truth as justice. Here we will merely situate this difficulty at the moment where in fact we encounter in the same word, marturion, and in a way that is not accidental, two heterogeneous meanings: 1. on the one hand, bearing witness, testimony, attestation (which belongs to the space of believing, of the act of faith, of engagement and signature; and we will constantly have to ask and re-ask ourselves, what does "believe" mean?); and 2. on the other hand, proof, certain determination, the order of knowledge. It is always the alternation between Glauben und Wissen, the title of a work of Hegel's, but also of an interminable debate between Kant and Hegel.

Whether it is phenomenological or semantic, we will not go so far as to say that in fact this distinction between testimony or bearing witness and proof " exists," in the strong and strict sense of the word. We will not go so far as to say that it holds in reality, now or at present. We are dealing here with a frontier that is both rigorous and inconsistent, unstable, hermetic, and permeable, uncrossable de jure but de facto crossed. The whole problem consists in the fact that the crossing of such a conceptual limit is both forbidden and constantly practiced. But if there is testimony and if it answers properly, incontestably, to the name and the sense intended by this name in our "culture," in the world which we think we can, precisely, inherit and bear witness to, then this testimony must not essentially consist in proving, confirming a knowledge, in assuring a theoretical certainty, a determining judgment. It can only appeal to an act of faith.

To complete this inspection of the Greek vocabulary, next to marturion, marturia thesthai refers to the action of giving a deposition: it is the attestation, the evidence set down by a witness. Martureisthai is to call to witness, to invoke witnesses, call upon as witness. A lovely example of this "call upon as witness," a sentence from the Civil Wars of the historian Appian of Alexandria, says this: "marturomenos emauton tes philotimias," "calling on myself as witness to my zeal, to my ambition, to my taste for honors." Another common translation: "calling on my conscience as witness to my ambition." Someone bears witness before the others, since he is speaking, since he is addressing himself to the others; but he calls on the others as witness to him by calling on himself first of all as witness to his being sufficiently conscious, present to himself, to bear witness before the others of what he bears witness to, to the fact that he is bearing witness and to that which he is bearing witness to—first of all before himself.

Why this translation? Why this example? Because in it we encounter one of the irreducible folds of witnessing and presence, of attendance, of attendance

---

in existence as presence: it is the fold of presence insofar as it is presence to oneself. A witness can only invoke having been present at this or that, having attended this or that, having tried out or experienced this or that, on condition of being and having been sufficiently present to himself, as such, on condition of claiming, at any rate, to have been sufficiently conscious of himself, sufficiently present to himself, to know what he is talking about. There is no longer any mask here. If there still were one, the masks would be exhibited as masks, in their mask-truth. I can only claim to offer reliable testimony if I claim to be able to testify to it before myself, sincerely, without mask and without veil, if I claim to know that what I have seen, heard, or touched, is the same as yesterday, to know what I know and mean to say what I mean to say. And thus to reveal or unveil—beyond the mask or the veil. In witnessing, presence to oneself, classic condition of responsibility, must be coextensive with presence to something else, with having-been-present to something else, and with presence to the other, for instance to the addressee of the testimony. It is on this condition that the witness can be answerable, responsible, for his testimony, as for the oath by which he commits himself to it and guarantees it. In the very concepts of perjury or lying as such it is presupposed that the liar or the perjurer is sufficiently present to himself; he has to keep the sense or the true sense of what he is dissimulating, falsifying, or betraying, present to himself, in its truth—and he has to be able, accordingly, to keep its secret. Keep it as such—and the keeping of this safekeeping is the movement of truth (veritas, verum, wahr, wahren, which means to keep; Wahrheit: the truth).

What we have here is one of the joints linking the problematics of the secret, of responsibility, and of testimony. No lie or perjury without responsibility, no responsibility without presence to oneself. This presence to oneself is of course often interpreted as self-consciousness. Under this head, bearing witness before the other would involve bearing witness before one's own consciousness; this can lead to a transcendental phenomenology of consciousness. But this presence to oneself is not necessarily the ultimate form of consciousness or of self-consciousness. It can take other forms of existence, that of a certain Dasein for instance. Think of the role (phenomenological in another sense) which the value of testimony or attestation can play in Heidegger's Sein und Zeit, especially around the passages concerning, precisely, Dasein's attestation (Bezeugung) to its originary possibility and its authenticity (Eigentlichkeit).^[8]Dasein must be able to testify about itself: here we have, in Sein und Zeit, the axiom or the testimony of the existential analytic of Dasein. From the beginning, Heidegger announces the bringing to light, the manifestation, the phenomenological presentation (Aufweis) of such a testimony (der phänomenologische Aufweis einer solcher Bezeugung), namely the phenomenology of an experience which is itself phenomenological, in other words which consists in a presentation. It is the presentation

---

of presentation, the testimony: here there is witness for the witness, testimony for the testimony.

To return now to the Celan poem, let us again stress this double reference to the attestation (Bezeugung), in "Niemand/zeugt für den/Zeugen,"as a reference to the enigmatic and recurrent figure of the three. While taking note of this crossover between the semantics of the witness and that of the "three" or the third, let us beware of being overhasty. Let us not preinterpret this co-occurrence of the two motifs in the Celan poem. Although this crossover is irreducible a priori wherever there may be a question of both the witness and the three, nothing allows us to go beyond this a priori in the reading of this poem.

The same is true for the reference to the oath. The poem names the oath and the petrified oath, that which roars deeply at the bottom of the petrified oath, of the oath of stone, of the oath become stone:

	Pontisches Einstmals: hier
	ein Tropfen,
	auf
	dem ertrunkenen Ruderblatt
	tief
	im versteinerten Schwur
	rauscht es auf.

Published translations:

	Pontique une fois: ici
	telle une goutte,
	sur
	le plat de la rame submergée,
	au profond
	du serment mué en pierre,
	sa rumeur.
	Pontic once-upon: here
	a drop
	on
	the drowned oar-blade,
	deep
	in the petrified vow,
	it roars up.

Another possible translation:

	Autrefois Pontique: ici
	une goutte
	sur
	la palme d'une rame noyée
	au fond
― 194 ―
	du serment pétrifié
	bruit.
	[Once Pontic: here
	a drop
	on
	the palm of a drowned oar
	at the bottom
	of the petrified vow
	sounds.]

Suppose that we refrain, as I would like to do here, from "commentary" on this poem. But even before doing this, in any case, and whatever the poem or its signatory means to say, whatever he means to be testifying to, one cannot not a priori link this figure of the oath to that of testimony which comes up at the end. There is no testimony without some involvement of oath (Schwur) and without sworn faith. What distinguishes an act of testifying from the straightforward transmission of knowledge, from straightforward information, from the straightforward statement or the mere demonstration of a proven theoretical truth, is that in it someone commits himself in regard to someone, by an oath that is at least implicit. The witness promises to say or to manifest something to another, his addressee: a truth, a sense which has been or is in some way present to him as a unique and irreplaceable witness. This irreplaceable singularity links the question of testifying to that of the secret but also, indissociably, to that of a death which no one can anticipate or see coming, nor give or receive in the place of the other. With this attestation, there is no other choice but to believe it or not to believe it. Verification or transformation into proof, contesting in the name of "knowledge," belong to a foreign space. They are heterogeneous to the moment peculiar to testifying. The experience of testifying as such thus presupposes the oath. It happens in the space of this sacramentum. The same oath links the witness and his addressees, for example—but this is only an example—in the scene of justice: "I swear to speak the truth, the whole truth, and nothing but the truth." This oath (sacramentum) is sacred: it marks acceptance of the sacred, acquiescence to entering into a holy or sacred space of the relationship to the other. Perjury itself involves this sacralization in sacrilege. The perjurer only commits perjury as such insofar as he keeps in mind the sacredness of the oath. Perjury, the lie, the mask, only appear as such ("the role of revealing the mask as mask") where they confirm their belonging to this zone of holy experience. To this extent at least, the perjurer remains faithful to what he betrays; he pays homage of sacrilege and perjury to sworn faith; in betrayal, he sacrifices to the very thing he is betraying; he does it on the altar of the very thing he is thereby profaning. Whence both the wiliness and the desperate innocence of him who would say: "in betraying, in betraying you, I renew the oath, I bring it back to life and I am more faithful to it

---

than ever, I am even more faithful than if I were behaving in an objectively faithful and irreproachable way while forgetting the inaugural sacramentum." For the secret that cannot be shared of the oath or perjury, for this secret that cannot even be shared with the partner in the oath, with the ally of the alliance, there is consequently only testimony and belief. Act of faith without possible proof. The hypothesis of proof does not even make sense any more. But on the other hand, because it remains on its own and without proof, this testimony cannot be authorized through a third person or through another testimony. For this witness there is no other witness: there is no witness for the witness. There is never a witness for the witness. This is also what the Celan poem might mean. It is also this that can always be alleged, inversely, by all the "revisionisms" in the world when they refuse all testimonies on the pretext that testimonies will never, by definition, be proofs. What should be the response to an allegation that might be translated like this: I can testify to this before my conscience, I am betraying you, I am lying to you, but in doing this I remain faithful to you, I am even more faithful than ever to our sacramentum? No objection can be made, nothing can be proved either for or against such a testimony. To this act of language, to this "performative" of testimony and declaration, the only possible response, in the night of faith, is another "performative" consisting of the saying or testing out, sometimes without even saying it, of an "I believe you."

How can this belief be thought? Where should we situate this faith, which does not necessarily have to take on the grand appearances of so-called religious faith? This act of faith is involved everywhere there is participation in what are called scenes of witnessing. And in truth the moment you open your mouth. The moment you open your mouth, the moment you exchange a look, even silently, a "believe me" is already involved, which echoes in the other. No lie and no perjury can overcome this appeal to belief; they can only confirm it; in profaning it, they can only confirm its invincibility.

Can this "believe" be thought? Is it accessible to the order of thought? The reason we referred to Sein und Zeit and to what it demands of Bezeugung, of the phenomenology of attestation, and precisely on the subject of the authentic being-able-to-be-oneself of Dasein, is that elsewhere Heidegger constantly excluded or at least dissociated the order of faith or belief from that of thinking or philosophy. He did this very often, but in particular in an abrupt, later statement, from "Der Spruch des Anaximander."^[9] This statement radically excludes the order of belief from that of thought in general. Heidegger then touches on a problem of translation. (I point it out because we too are caught up, right here, in the scene of translation and testimony, and of the translation of the poem by Celan on testimony, of a poem that is virtually untranslatable and which testifies on the subject of testifying.) It is for Heidegger precisely a question of the translation of a Spruch. Spruch: word of honor, sentence, judgment, decision, poem, at any rate a speech which

---

is not a theoretical and scientific utterance and which is linked to some language in a way that is both singular and "performative." Now what does Heidegger say in a passage which is also concerned, precisely, with presence (Anwesen, Praesenz), that presence which is the foundation of the classical value of testimony, and this time presence as representation, in the "representation of representing" ("in der Repraesentation des Vorstellens")? After proposing a translation for Anaximander's aphorism, Heidegger declares: "Belief has no place in the act of thinking (Der glaube hat im Denken keinen Platz)." This sentence is taken from an argument which must be reconstructed, at least in part:

We cannot prove (beweisen) translation scientifically, nor should we, in virtue of some authority, have faith in it [give it credit, believe it, glauben]. The reach of proof [understood: "scientific" proof] is too short (Beweis trägt zu kurz). Belief has no place in the act of thinking. Translation can be rethought [reflected, nachdenken] only in the thinking [im Denken] of the judgment [word of honor, Spruch: it is necessary to think the Spruch, that speech of commitment as poem, judgment, decision, commitment, to think, rethink, starting from there, the possibility of translation, and not the other way around]. But thought (das Denken) is the Dichten [the poem, poetizing, the poetical act or operation, the poetic which Krieger is perhaps speaking of in the passage quoted as epigraph—but the words "act" and "operation" are not suitable: there is something other than the activity of a subject, perhaps we should say "the event," the "coming" of the poetic—] of the truth of being (der Wahrheit des Seins) in the historical conversation [dialogue, dual language] of thinkers [geschichtlichen Zwiesprache des Denkenden].

Heidegger thus dismisses scientific proof and belief back to back, which could let it be thought that to that degree he is giving credit to a nonscientific witnessing. In this context, the believing of belief is the credulity which accredits authority, the credulity which shuts its eyes to acquiesce dogmatically in authority ("Autorität" is Heidegger's word). Heidegger extends with no less force and authority the assertion according to which believing has no place in thinking. Is this believing foreign to that which in thought itself (in particular, the thought which thinks in the Zwiesprache and holds itself in relation to the Spruch of a thinker, in the experience of thinking translation) concerns the Bezeugung, the attestation which Sein und Zeit speaks of? Is there not a belief in the recourse to attestation (Bezeugung), in the discourse which brings it into play? And in the experience of thought in general, thought as Heidegger refers to it, is there not an experience of believing which is not reducible to that credulity or that passivity before authority which Heidegger here excludes so easily from thinking? And doesn't the authority of some "believing," "making [someone] believe," "asking to believe" always necessarily insinuate itself in the invocation of a thought of the truth of being? What, in that which is not proof, holds the place of

---

this Glauben in the thought which Heidegger means to think at the very moment when he is excluding belief or faith?

III

"Raise your right hand and say ‘I swear it.’" To these words a witness must respond, when he appears before a French court. Whatever the meaning of the raised hand, it does in any case engage the visible body in the act of the oath. The same is true for the wedding ring worn on the finger. Now "Schwurfinger" means the three fingers that are raised in taking an oath. That is perhaps not unrelated to the "Händen am Dreiweg"or the "Dreiweg-Händen" which return on two occasions in the Celan poem. They are first of all tied to the tie, the knot (knoten, verknoteten). One can imagine that these "knots" are not unrelated to the ties of the oath, for instance the oath of stone the poem mentions: "im versteinerten Schwur." Secondly, they are linked to the knots of the hands (erschüttert-verknoteten Händen) and of pain (Schmerzknoten).

Stricto sensu, determined by a culture, the inherited concept of testimony implies, we were saying, some kind of oath, some swearing to a law or faith. This is the reference to the sacramentum, namely to what is at issue between the parties involved in a trial, or in a contest. This issue is entrusted, during the hearing, during the procedure known as per sacramentum, to the Pontiff. "Pontiff" is not far from Pontisches, "Pontisches Einstmals" ("Pontic onceupon") which we will have occasion to speak about again.

But that does not necessarily mean that for every testimony we have to raise our hands and swear to speak the truth, the whole truth, and nothing but the truth. That does not necessarily mean that every time we do what is called témoigner or déposer sous la foi du serment [give evidence on oath] (unter Eid bezeugen, unter Eid aussagen, or to testify or to bear witness—which almost always has of itself the weight of "attesting on oath," before the law), we do it ritually. No, but even when the scene is not formalized in this way by an institutional code of the positive law which would oblige you to observe this or that rite, there is in every testimony an implication of oath and of law.

This extension of the implied oath just mentioned may appear extraordinary and abusive, even extravagant, but I think it legitimate, meaning incontestable. Logically, it makes it obligatory to take any address to another for a testimony. Each time I speak or manifest something to another, I am testifying to the extent that even if I neither say nor show the truth, even if, behind the "mask," I am lying, hiding, or betraying, every utterance implies "I am telling you the truth, I am telling you what I think, I testify before you to that to which I testify before me, and which is present to myself (singularly, irreplaceably). And I can always be lying to you. So I am before you as before a judge, before the law or the representative of the law. As soon as I

---

testify I am before you as before the law, but, as a result, you who are my witness, you who bear witness to my bearing witness, you are also judge and arbiter, judge and party as much as judge and arbiter." We will come back to this essential possibility of the judge's becoming-witness or the witness's becoming-judge and becoming-arbiter.

I have already admitted: I will not attempt to interpret this poem. Not even its last lines:

	Niemand
	zeugt für den
	Zeugen.
	No one
	bears witness for the
	witness.

What then are we doing with this poem? And why are we quoting it? Why are we invoking its poetic force? Why are we borrowing its force when in fact, and no doubt because of this, beyond all we might decode of this poem we don't finally know to what it is bearing witness? What we are calling here the force of the poem, and first of all in its language, is what makes us have to quote it, again and again, by an irresistible compulsion. For it is quoted and requoted, it tends to be learned by heart when you know you do not know what, in the last resort, it means, when you do not even know of what or for whom and for what it is testifying. Because we do not know, even if we can know a lot and learn a lot about it, and learn a lot from it. We can "read" this poem, can desire to read, quote, and requote it, while giving up on interpreting it, or at least on going over the limit beyond which interpretation encounters both its possibility and its impossibility. What we have here is a compulsion to quote and requote, to repeat what we understand without completely understanding it, feeling at work in the economy of the ellipsis a force stronger than that of meaning and perhaps even than that of truth, of the mask which would manifest itself as mask. The reciting compulsion, the "by heart" desire, is because of this limit to intelligibility or transparency of meaning.

Is not this limit that of a crypt, and thus of a certain secret? In bearing witness for witnessing and for the witness, the poem says that there is no witness for the witness. It is presumably an indication, a descriptive statement, but also, implicitly perhaps, a prohibiting prescription: no one does in fact bear witness for the witness, no one can, it is true, but first of all because no one should. No one can, because it must not be done.

The poem bears witness. We don't know about what and for what, about whom and for whom, in bearing witness for bearing witness, it bears witness. But it bears witness. As a result, what it says of the witness it also says of itself as witness or as witnessing. As poetic witnessing. Can we not, then, here transfer

---

to witnessing, to this poetic witnessing, as to that which in all witnessing must always appear as "poetic" (a singular act, concerning a singular event and engaging a unique, and thus inventive, relationship to language), that "extravagant proposal" of Murray Krieger's: "the poem, in the very act of becoming successfully poetic—that is, in constituting itself poetry—implicitly constitutes its own poetic"?

Moreover, taken on its own, the last stanza

	Niemand
	zeugt für den
	Zeugen.
	No one
	bears witness for the
	witness.

may vacillate or pivot, it seems to turn-re-turn around the axis of its own syntax. To the point of vertigo. The "for" (für)—what does it mean?

We can offer at least three hypotheses on this subject.

1. Is it about bearing witness in favor of someone (I bear witness for you, I bear witness in your favor, I am a witness for the defense, etc.)? Zeugen für jenen does in fact generally mean to testify in favor of someone, as opposed to zeugen gegen jenen, to testify against someone.

2. Is it rather about "bearing witness for" the other in the sense of "in the place of" the other? And here invalidating this possibility, this power, this right, by recalling that no one can bear witness in the place of another, any more than anyone can die in the place of another? In this impossibility of substitution, we are testing out an alliance between death and the secret. The secret always remains the experience of bearing witness, the privilege of a witness for whom no one can be substituted, because he is, in essence, the only one to know what he has seen, lived, felt; he must thus be believed, taken at his word, at the very moment when he is making public a secret which anyway remains secret. A secret as secret. Now even if we cannot say anything definite about it, "Aschenglorie" clearly remains a poem of death and the secret. The poem survives by bearing witness, through this alliance, to the surviving of the testis as superstes.

If no one can replace anyone as witness, if no one can bear witness for the other as witness, if one cannot testify for a testimony without taking from it its worth as testimony (which must always be done in the first person), is it not difficult to identify the witness with a third person? We readily represent the third person as anyone, as a replaceable first person: the third is a singular "I" in general. In fact nothing is more substitutable but nothing is less so than an "I." The question introducing itself on the horizon is indeed that of what one calls a first person, a discourse in the first person (singular or plural, I or we). Who is the "I" of the poem? This question displaces itself; it

---

gets divided or multiplied, like the question of the signature, between the "I" of which the poem speaks, or to which the poem refers, reflexively (and which can also sometimes be a mentioned or even a quoted "I," if we want), the "I" who writes it or "signs" it in all the possible ways, and the "I" who reads it. How then is it determined, this self-referentiality, the autodeictic quality always posited or alleged by whoever says "I," thereby demonstrating, even if he is masked, that the speaker is showing himself and referring to himself? The form of this self-referential self-presentation is not only grammatical; it can be simply implied by discourses which are not conjugated in the first person of the present.

3. But there is still a third possibility: to testify "for" someone not in the sense of "in favor" or "in the place of" but "for" someone in the sense of "before" someone. One would then testify for someone who becomes the addressee of the testimony, someone to whose ears or eyes one is testifying. Then the sentence "no one bears witness for a witness" would mean that no one, no witness, bears witness before someone who is also a(nother) witness. A witness, as such, is never in a position to receive the testimony of another, nor entitled to do so. The judge or the tribunal, the representatives of the law, assumed to be neutral and objective, can certainly receive a testimony, but another witness cannot, since he is as singular and as involved as the first witness. The judge or the tribunal, the arbiters, those who judge and decide, those who conclude, are not mere witnesses; they must not, should not be only witnesses, in other words subjects who find themselves strangely in the situation of being present at or participating in that which is attested to. They would be suspected, as any witness is suspect, of being interested parties, partial subjectivities, interested, situated in the space described by the testimony. The judge, the arbiter, or the addressee of the testimony is thus not a witness: he cannot and must not be. And yet at the end of the day the judge, the arbiter, or the addressee do have to be also witnesses; they do have to be able to testify, in their turn, before their consciences or before others, to what they have attended, to what they have been present at, to what they have happened to be in the presence of: the testimony of the witness in the witness box. It is only on the basis of this testimony that they will be able to justify their judgment. The judge, the arbiter, the historian also remains a witness, a witness of a witness, when he receives, evaluates, criticizes, interprets the testimony of a survivor, for instance a survivor of Auschwitz. Whether he accepts or contests this testimony, he remains a witness of a witness. He remains a witness even if he contests the first testimony by alleging that, since he has survived, the survivor cannot be a certain and reliable witness to what happened, in particular of the existence for this purpose, a purpose of putting to death, of gas chambers or ovens for cremation—and that therefore he cannot bear witness for the only and true witnesses, those who have died, and who by definition can no longer

---

bear witness, confirm or disprove the testimony of another. In this context, "Aschenglorie" can also offer for hearing, between the words, mounting from the light of the ashes, something like a desperate sigh: no witness for the witness in this perverse situation which will permit all the judges, arbiters, historians to hold the revisionist thesis to be fundamentally indestructible or incontestable.

Although he cannot be a witness "in the box," the judge-arbiter-historian must also and further testify, if only to what he has heard attested. He must bear witness to the experience in the course of which, having been present, put in the presence of the testifying, he has been able to hear it, understand it, and can still reproduce the essence of it, etc. There would be a third person and testimony to testimony, witness for the witness.^[10]

Niemand/zeugt für den/Zeugen: "Für"is thus both the most decisive and the most undecidable word in the poem. Nothing prohibits any of these three readings. They are different, but not necessarily incompatible. On the contrary, they can accumulate their potential energy deep in the crypt of the poem, thereby giving it its force of appeal and inducing our compulsion to cite-recite it without knowing, beyond knowledge. In these three readings of "für," which intensify the three that we have not finished with, even the verb of the stanza vacillates as well. Its tense vacillates, it makes its mood and the negation to which it is subject (Niemand zeugt) vacillate with it. The present indicative can signify a fact to be stated: no one bears witness…. But as is often the case (in French too, especially when there is question of law), "no one bears witness" implies: "no one can bear witness," "no one can, has been able, and will ever be able to bear witness for the witness" (with the three possible senses of "for" that we have just invoked). And as a result, this being able, this "not being able" is displaced and translates easily into an "ought not to" or a "to have not to": no one can, which is to say no one must, no one should bear witness for the witness, replace the witness, defend the witness, bear witness before the witness, etc. One cannot and (in addition or moreover or above all) one should not bear witness for the witness, in all the senses of "for." One cannot and should not (claim to) replace the witness of his own death, for instance someone who perished in the hell of Auschwitz (but that does not mean that this poem is a poem on Auschwitz—and for the very reason that I am in the process of pointing out again, namely that no one bears witness for the witness). One neither can nor should replace (thus bear witness for) the witness of his or her own death, or the witness of others' deaths, the one who was present and survived, for instance at the hell of Auschwitz.

And yet, in its own way, the poem does bear witness to this impossibility. It attests to this prohibition which is imposed on bearing witness, in the very place where one has to go on appealing to it. This impossibility, this prohibition appears as such. Non-appearance appears (perhaps) as non-manifestation.

---

Is this possible? How? How should we understand this "perhaps"? Its possibility or its necessity?

It is a matter of death, if death is what one cannot bear witness to for the other, and first of all because one cannot bear witness to it for oneself. The survival of surviving, as place of witness and as testament, would here find both its possibility and its impossibility, its chance and its threat. It would find them in this structure and in this event.

That this is a poem on the subject of death, a poem of death, a poem which speaks death as such, can be maintained at no great risk. It can be maintained where one cannot separate questions of the secret, the crypt, and witnessing from questions of survival and death. It can also be affirmed by taking as testimony the naming of ashes, of course. There are ashes there but they are of glory. Or again, there is glory, light, fire, but already in ashes. Double possibility of the "but"—the ash, certainly, and death, but glorious; glory, certainly, but of ash and death without memory. The double possibility of this implied "but" is, indeed, implied, in the hyphen, which is now stressed at the end of the line to articulate and disarticulate the relationship between ashes and glory,

	Aschen
	glorie

(double word: we don't know which is the subject and which the predicate), and is now effaced, in a single, simple word, as in the opening ("Aschenglorie"). There too, one does not know whether the glory is of ashes or the ashes are glorious, ashes of glory. This explains du Bouchet's French translation, which prefers "Cendres-la gloire" ["Ashes-glory"] to "Gloire de cendres" ["Glory of ashes"]. "Ashes" is always in the plural, of course: ashes never gather together their dissemination and that is just what they consist in. They consist in not consisting, in losing all consistence. They have no more existence, they are deprived of any substance that is gathered together and identical to itself, of any relationship to themselves.

That is confirmed (perhaps) via the association of the Dreiweg with the Pontisches, with the petrification of the oath in its crypt, especially with the Tartar moon (Tatarenmond). There are at least two proper names in that (Pontisches and Tatarenmond) whose referent seems unavoidable. Namely, perhaps, the goddess Hecate. Here is the stanza we have not yet read:

	(Auf dem senkrechten
	Atemseil, damals
	höher als oben,
	zwischen zwei Schmerzknoten, während
	der blanke
	Tatarenmond zu uns heraufklomm,
	grub ich mich in dich und in dich.)

---

The published translations say:

	(On the perpendicular
	breath-rope, at that time,
	higher than above
	between two pain-knots,
	while the shiny
	Tartar moon climbed up to us,
	I burrowed into you and into you.)
	(Perpendiculaire, alors,
	sur cette corde le souffle,
	plus haut que le faîte,
	entre deux noeuds de douleur, cependant
	que la blanche
	lune tatare jusqu'à nous se hisse,
	je m'enfouis en toi et toi.)

It could also be translated:

	(Sur la corde verticale
	du souffle [corde vocale?] autrefois [damals, which echoes Einstmals above; once
	upon a time],
	plus haut qu'en haut,
	entre deux noeuds de douleur, pendant que
	La nue (luisante, lisse) lune tatare se haussait (s'élevait) vers nous
	je m'enfousissais (je m'enterrais, je m'encryptais) en toi et en toi.)
	[(On the vertical cord
	of the breath [vocal chord?], long ago,
	higher than on high,
	between two knots of pain, while
	the bare (shiny, smooth) Tartar moon was raising itself (rising) towards us
	I buried myself (I interred myself, I encrypted myself) into you and into
	you.)]

The name of the goddess Hecate is not pronounced. It remains, it will perhaps remain ineffaceable, beneath the surface of this poem, because of the association of the moon, the Pontic, and the three of the Dreiweg. However little one knows about the goddess Hecate, the first thing one remembers is that her most important trait is the three—and the tripleness of the way or road. She is trimorphic, she has three forms and three faces (triprosopos). She is also the goddess of the crossroads, in other words, as the name both indicates and does not indicate (quadrifurcum), of a road branching off in four rather than three directions. Of course, but apart from all the Oedipal associations which multiply with every crossroads, we know that a crossroads can be made of the crossing of two, three, or four roads, hence in three ways. Now Hecate, goddess of crossroads, is called trioditis (a word which

---

comes from triodos, three ways: it is the adjective from triodos, honored in the crossroads). She protects roads and she is polyonymous, she has lots of names. We are only selecting the features which matter to us here. An account of Hecate could be prolonged ad infinitum. For this goddess of the Dreiweg also has a privileged relationship with fire, with brightness, with burning—and so with consumption or ashes as much as with glory. Her mouth exhales fire, she is pyripnoa (breath of fire) (Atem, that word dear to the author of Atemwende, the title of the collection that includes "Aschenglorie," Atem that we also come across here again in "Atemseil"). Her hands brandish torches. The Chaldaean Oracles associate her with implacable thunderbolts and call her "flower of fire." Transporting fire from on high (think of the verticality and "höher als oben"in Celan), she is life-giving and fertile. But another chain of associations inverts these meanings and turns Hecate towards the aspect of the moon and death. Her signs and her triadic nature then couple her with Mene or Selene, the moon, the goddess of the moon—which we see appearing in Celan's poem. Some prayers to the moon invoke Hecate and Selene as one and the same goddess (three heads, crossroads, etc….): "This is why you are called Hecate of many names, Mene, you who split the air like Artemis the arrow-darter…. [I]t is from you that all proceeds and in you, who are eternal, that all comes to an end."^[11] Elsewhere, she becomes Aphrodite, universal procreator and mother of Eros both low and high "in the Underworld, the abyss and the aion (the forever, being in all times, the eternal)." Goddess of light but also of night, she keeps her festival in the crypts and the tombs. So this is also a goddess of death and the subterranean underworld, a goddess of Hades. This is the guise in which Hecate appears in Macbeth, at any rate. Apart from the general knowledge that one might have of this, we know that Celan also translated Shakespeare. What can be stressed in a completely distinctive way in the apparition of Hecate (Act 3, Scene 5) is that the three surfaces again there in the form of the three witches who meet her and speak with her ("Why, how now, Hecate! you look angerly!"). Hecate's reply is about nothing else than death ("How did you dare/To trade and traffic with Macbeth/In riddles and, affairs of death"); glory ("or show the glory of our art"); the "pit of Acheron"; the moon ("Upon the corner of the moon"); etc.

With Acheron, or the Styx, we could return to Celan's poem, to "Pontisches Einstmals," the only time that we cross the waves of the Black Sea. Because it is only once that we cross them. The "Pontic once-upon" perhaps designating the passage of death. That is also where Odysseus is only authorized to pass through a single time to go and see the dead, when he goes to consult Tiresias. At the moment of death—and to reassure themselves about their fate after death, even if they were cremated—the Greeks needed a witness.

---

They had to go by way of a trivium which would decide on the route and the place of their destination.

There would be much to say here, in particular on the subject of Odysseus, of Elpenor, his drunkenness and his oar, to which there is perhaps a reference, we will never know, in the words "ertrunkenen Ruderblatt," the blade of a drowned or drunken oar. There would be much to say as well on the subject of the vertical cord, the breath-rope (Atemseil), which perhaps, perhaps, alludes to the death of Tsvetaeva. We know what she represented for Celan. Tsvetaeva hanged herself in 1941 unwitnessed. She lived in the Tartar republic. So that the Tartar moon (Tatarenmond) may condense at least two encrypted allusions, thereby—as is most often the case—foiling the unity of reference, and thus of reading, and thus of witnessing, without however effacing the singularity of each event, of each date thereby re-lated, re-marked.

Whatever their probability or improbability, the "perhaps" of these singular references which all make appeal to dated evidence (for example, we have to know who Tsvetaeva was and who she was for Celan, and how, where, and when she killed herself, she too, like him, etc.), we can say a priori that this poem speaks of death (for which there is no witnessing), perhaps of suicide, and that the "grub ich mich in dich und in dich" may mean not only "I dig myself in, I burrow," but also "I bury myself, I encrypt myself within you inside you": graben, grub; and Grab is the tomb: you are my tomb, my own tomb, you to whom I address myself, whom I take as witness, if only to say (to you) "no one bears witness for the witness."

Beyond or before everything that could be thought, read, or said of this poem, according to the "perhaps," the probability and the act of faith which is a poetic experience, beyond or before all the possible translations, a mark remains and is here re-marked: it is a certain limit to interpretation. Ultimately, it is in all certainty impossible to put a stop to the meaning or the reference of this poem, the meaning or the reference to which it testifies or responds. Whatever one might say about it, and that can be deployed ad infinitum, there is a line. It is not marked only by the poem. It is the poem, poetics, and the poetics of the poem—which dissimulates itself by exhibiting its dissimulation as such. But it is this "as such" which turns out to be doomed to the "perhaps." Probable and improbable (possible but removed from proof), this "as such" takes place as poem, as this poem, in it, and there one cannot reply in its place, there where it is silent, there where it keeps its secret, while telling us that there is a secret, revealing the secret it is keeping as a secret, not revealing it, while it continues to bear witness that one cannot bear witness for the witness, who ultimately remains alone and without witness. In Le pas au-delà, Blanchot speaks of a "word still to be spoken beyond the living and the dead, bearing witness for the absence of attestation."^[12]

It is of this essential solitude of the witness that I would have liked to speak. It is not a solitude like just any solitude—nor a secret like just any secret.

---

It is solitude and secrecy themselves. They speak. As Celan says elsewhere, the poem, it speaks, secretly, of the secret, through the secret, and thus, in a certain way, in it beyond it: "Aber das Gedicht spricht ja! Es bleibt seiner Daten eingedenk, aber—es spricht" ("But the poem it speaks! It keeps its dates in mind, but in the end—it speaks").^[13] It speaks to the other by keeping quiet, keeping something quiet from him. In keeping quiet, in keeping silence, it is still addressing itself. This internal limit to any witnessing is also what the poem says. It bears witness to it even in saying "no one bears witness for the witness." Revealing its mask as a mask, but without showing itself, without presenting itself, perhaps presenting its non-presentation as such, representing it, it thus speaks about witnessing in general, but first of all about the poem that it is, about itself in its singularity, and about the witnessing to which any poem bears witness.

Left here to itself, in its essential solitude, in its performance or in its happening, the poetic act of the work perhaps no longer derives from the presentation of self as such.

NOTES

However, the English verb "witness" also has a transitional sense, or rather one that encompasses both meanings. When someone formally "witnesses" a signature, or some other official event, the witnessing is not only seeing the deed done (witnessing in the first sense), but also, by giving a witness's signature, attesting (the second sense) to having witnessed in the first sense. For this reason, I have sometimes translated témoignage as "witnessing," rather than "bearing witness" or "testimony," where the context makes it clear that that is the sense implied.

---