History, Epistemology

I, Herodotus of Halicarnassus, am here setting forth my history [historiē], that time may not draw the color from what man has brought into being, nor those great and wonderful deeds, manifested by both Greeks and barbarians, fail of their report, and, together with all this, the reason why they fought one another.

The chroniclers [logioi] among the Persians say that it was the Phoenicians who were the cause of the falling-out . . .

§ Herodotus, The History 1

The second part of this work is devoted to the epistemology of historical knowledge. Here I want to situate this stage of my inquiry and its principal interconnections.

On one side, I consider the phenomenology of memory to be ended, making reservation for the cultural variations that historical knowledge, when integrated into individual and collective memory, may introduce into self-understanding in the mnemonic mode. A subtle combination of those features of memory we can call transhistorical and its variable expressions over the course of history will have to be taken into account when the time comes. This will be one of the themes of the hermeneutics of the historical condition to be dealt with in part 3chapter 2. Before that, however, history will have to have attained its fully autonomous status among the human sciences, following the vow that orients this middle section of my work. Then, on a second degree level of reflection, the question of the internal limits of a philosophical project that often remains tacit can be posed. This question has to do not only with the epistemological autonomy of historical research but also with the self-sufficiency of history’s own self-awareness, in accord with the favorite expression that presided over the birth of and apology for the German school of history. It is within the framework of this reflection on the limits stemming from a critical philosophy of history that the confrontation between intending the truth of history2 and the aim of that veracity or, as I shall put it, the intention of being faithful to memory (part 3chapter 1) can be brought to a good ending. Until then, the status of history as regards memory will be held in suspense without my, for all that, forbidding myself from noting along the way the resurgence of the aporias of memory in their cognitive and practical aspects, principally the aporia of the representation of an absent something that once happened, along with that of the use and abuse to which memory lends itself as actively exercised and practiced. Yet this obstinate return of the aporias of memory at the heart of historical knowledge cannot take the place of a solution of the problem of the relations between knowledge and the practice of history and the experience of lived memory, even if this solution were to present ultimately indecisive features. Nevertheless, these features will have to be painfully won on the field of battle of a reflection carried to its limits.

It remains the case that the autonomy of historical knowledge in relation to the mnemonic phenomenon remains the major presupposition of a coherent epistemology of history both as a scientific discipline and a literary one. At least this is the presupposition assumed in this middle part of this work.

I have adopted the expression the “historical”—or better “historiographical”—operation to define the field traversed by the following epistemological analysis. I owe it to Michel de Certeau in his contribution to the large-scale project edited by Jacques Le Goff and Pierre Nora under the title Faire de l’histoire.3 Beyond this, I have also adopted the broader lines of the triadic structure of Certeau’s essay, although I give them different contents on some important points. I first tried out this clear and insightful tripartite approach in an essay requested by the Institut International de Philosophie.4 Keeping this double influence in mind, I shall call the “documentary phase” the one that runs from the declarations of eyewitnesses to the constituting of archives, which takes as its epistemological program the establishing of documentary proof (chapter 1). Next I shall call the explanation/understanding [explicative/compréhensive] phase the one that has to do with the multiple uses of the connective “because” responding to the question “why?”: Why did things happen like that and not otherwise? The double term “explanation/understanding” is indicative of my refusing the opposition between explanation and understanding that all too often has prevented grasping the treatment of the historical “because” in its full amplitude and complexity (chapter 2). Finally, I shall call the “representative phase” the putting into literary or written form of discourse offered to the readers of history. If the major epistemological crux occurs in the explanation/understanding phase, it does not exhaust itself there inasmuch as it is the phase of writing that plainly states the historian’s intention, which is to represent the past just as it happened—whatever meaning may be assigned to this “just as.” It is also at this third phase that the major aporias of memory return in force to the foreground, the aporia of the representation of an absent thing that occurred previously and that of a practice devoted to the active recalling of the past, which history elevates to the level of a reconstruction (chapter 3).

At the beginning of each of these three chapters in part 2 I shall lay out the program belonging to each of these phases. Here I shall confine myself to a more specific statement of the historian’s threefold commitment.

I have proposed the word “phase” to characterize the three segments of the historiographical operation. There is no need for any equivocation concerning the use of this term. It is not a question of distinct chronological stages, but of methodological moments, interwoven with one another. As will be repeated, no one consults an archive apart from some project of explanation, without some hypothesis for understanding. And no one undertakes to explain a course of events without making use of some express literary form of a narrative, rhetorical, or imaginative character. Any idea of chronological succession must be banished from our use of the term “operative phase.” It is only in the discourse undertaken here on the moments of the unfolding of the historiographical operation that these phases become stages, successive steps in a trajectory that unrolls its linearity. We can completely avoid such equivocation regarding succession if we speak of levels, a term that evokes superposition, stacking things up. But we need to keep an eye out for another equivocation as well, that of a relation between an infra-and a super-structure, much used and abused by vulgar Marxism (which I am not confusing with Marx’s major works). Each of the three operations of the historiographical operation stands as a base for the other two, inasmuch as they serve successively as referents for the other two. In the end, I have preferred the term “phase” inasmuch as, in the absence of a chronological order of succession, it underscores the progression of this operation as having do to with the historian’s intention of a true reconstruction of the past. It is only in the third phase—as has already been suggested—that the intention to represent the truth of past things openly declares itself, through which the cognitive and practical project of history as it is written by professional historians defines itself over against memory. A third term, which I preferred in my preliminary work, is “program.” It works well to characterize the specificity of the immanent project in each step along the way. In this sense, it has an analytic privilege in regard to the other two terms. This is why I have made recourse to it whenever the accent is placed on the nature of the operations undertaken at each level.

The final word of this general orientation will be about the term “historiography.” Until recently, it designated by preference the epistemological inquiry such as we are undertaking here following its threefold rhythm. I use it, as does Certeau, to designate the very operation in which historical knowing is grasped at work. This choice of vocabulary has a major advantage that does not appear if we reserve this term for the writing phase of the operation, as suggested by the very composition of the word: historiography or history writing. In order to preserve the amplitude of the term “historiographical,” I will not call the third phase the writing of history, but instead the literary or scriptural phase, when it is a question of the exposition, presentation, or exhibiting of the historian’s intention taken in terms of the unity of its phases, that is, the present representation of absent, past things. Writing, in effect, is the threshold of language that historical knowing has already crossed, in distancing itself from memory to undertake the threefold adventure of archival research, explanation, and representation. History is writing from one end to another. And in this regard, archives constitute the first writing that confronts history, before it completes itself in the literary mode of “scripturality.” Explanation/understanding thus finds itself encased, upstream and downstream, by two writings. It gathers energy from the former and anticipates the energy of the latter.

But it is above all the setting out in writing of the historian’s knowledge starting from the upstream side of archives that gives rise to the question of confidence that cannot be answered from inside the epistemology of historical knowledge, the question of what finally becomes of the relation between history and memory. This is the question of confidence that a critical philosophy of history has the task, if not of resolving, at least of articulating and considering. But it is posed in an originary manner by the entry into writing of the historian’s knowledge. It floats as the unsaid over the whole undertaking. For me, who knows what follows, this unsaid, which will be taken up in part 3, needs to be left in suspense, in reserve, something like a methodological epoché.

To indicate this setting in reserve, in the most decidedly interrogative, skeptical way, I have chosen to place in the position of prelude a kind of parody of the Platonic myth from the Phaedrus dedicated to the invention of writing. Inasmuch as the gift of writing is held by this myth to be the antidote to memory, and therefore a kind of challenge opposed by the truth claim of history to memory’s vow of trustworthiness, it can be taken as the paradigm for every dream of substituting history for memory, as I shall consider further in the beginning of part 3. Thus, it was in order to underscore the gravity of a cultural choice from which there was no going back, that of writing history, that I have amused myself in my own fashion, with what was first due to Plato, in reinterpreting, if not rewriting, the myth his Phaedrus recounts concerning the invention of writing. The question whether the pharmakon of history-writing is remedy or poison, to take one of the propositions of the myth in the Phaedrus, will continue to accompany our epistemological inquiry as a kind of background music before breaking forth in full force on the reflective plane of the critical philosophy of history.

Why refer to myth, even in the preliminary material of a highly rational epistemological analysis? In order to confront the aporia in which every inquiry bearing on the birth, the beginning, the beginnings of historical knowledge gets lost. This perfectly legitimate inquiry, to which we owe many worthwhile works,5 rests, insofar as it itself is historical, on a kind of performative contradiction, namely, that the writing of beginnings presupposes itself as already existing in order to think of itself at its birth. We need therefore to distinguish origin from beginning. We can seek to date a beginning in a historical time scanned by chronology. This beginning is perhaps not discoverable, as the antinomies articulated by Kant in the Dialectic of the Critique of Pure Reason suggest. We can, of course, indicate something as a start in a critical treatment of testimonies, but this is not a beginning of the mode of historical thinking, if we mean by this a temporalization of common experience in a way irreducible to that of memory, even collective memory. This unassignable anteriority is that of the inscription that, in one form or another, has always accompanied orality, as Jacques Derrida has magisterially demonstrated in his Of Grammatology.6 Human beings have spaced their signs, at the same time—if this has a meaning—that they have woven them together in terms of the temporal continuity of the verbal flow. This is why the beginning of the historian’s scripturality is undiscoverable. The circular character of assigning a historical beginning to historical knowledge invites us to distinguish, at the heart of the amphibolous concept of birth, between beginning and origin. The beginning consists in a constellation of dated events, set by the historian at the head of the historical process that will be the history of history. It is toward this beginning or these beginnings that the historian of the birth of history advances by a retrospective movement that produces itself within the already constituted setting of historical knowledge. The origin is something else again: it designates the upsurge of the act of taking a distance that makes possible the whole enterprise and therefore also its beginning in time. This upsurge is always current and always already there. History continues to be born from this taking of a distance which consists in the recourse to the exteriority of the archival trace. This why we find its mark in the innumerable modes of graphism, of inscription that precede the beginnings of historical knowledge and the historian’s profession. The origin, therefore, is not the beginning. And the notion of birth conceals beneath its amphiboly the gap between the two categories of beginning and origin.

It is this aporia of birth that justifies the Platonic use of myth: the beginning is historic, the origin is mythic. It is a question, of course, of a reuse of a form of discourse appropriate to any history of beginnings presupposing themselves, such as the creation of a work, the birth of an institution, or the vocation of a prophet. Reused by the philosopher, myth presents itself as myth, in the guise of an initiation and supplement to dialectic.


History: Remedy or Poison?

I shall speak in the manner of Plato’s Phaedrus of the mythic birth of the writing of history. That this extension of the myth of the origin of writing may sound like a myth of the origin of history, thanks to rewriting, is, if I may put it this way, authorized by the myth itself, inasmuch as what is at stake is the fate of memory, even if the irony is directed in the first place at the “written discourses” of orators such as Lysias. Furthermore, it also has to do with other fabulous inventions: calculation, geometry, but also checkers and dice, which the myth compares to the invention of writing. And does not Plato indirectly include his own writing, he who wrote down and published his dialogues? But it is to true memory, genuine memory, that the invention of writing and its related drugs is opposed as a threat. How then can the debate between memory and history not be affected by this myth?

To get quickly to the point, what fascinates me, as it does Jacques Derrida, is the insurmountable ambiguity attached to the pharmakon that the god offers the king.1 My question: must we not ask whether the writing of history, too, is remedy or poison? This question, no less than that of the amphibology of the notion of birth applied to history, will not let go of us. It will spring up again in another prelude, at the beginning of part 3: Nietzsche’s second Unfashionable Observation.

Let us take up the myth: “Theuth said: ‘O king, here is something that, once learned, will make the Egyptians wiser and will improve their memory [mnēmonikōterous]; I have discovered a potion [pharmakon] for memory [mnēmēs] and wisdom [sophias]’” (274e)!2 It is the grammata that come to the fore among the potions offered by the one whom Theuth calls “the father of writing,” “the father of grammata.” Is not historiography in a certain way the heir of the ars memoriae, that artificial memory that we referred to above as a way of expressing memorization turned into an exploit? And is it not memorization rather than remembering, in the sense of a precise memory of past events, that is at issue in this narrative?3 The king readily concedes to the god the privilege of engendering the art, but he retains the right to judge what he calls its “benefit” and “harm”—just as Nietzsche will subsequently do as regards history in his second Unfashionable Observation. How does he respond to the god’s offer? “In fact, it will introduce forgetfulness into the soul of those who learn it: they will not practice using their memory [mnēmēs] because they will put their trust in writing [graphēs], which is external and depends on signs that belong to others [tupōn], instead of trying to remember [anamimnēskomenous] from the inside, completely on their own. You have not discovered a potion [pharmakon] for remembering, but for reminding [hupomnēseōs]; you provide your students with the appearance of wisdom, not with its reality” (275a).4 The verbs and nouns having to do with memory are important and differ: the god’s offer is that of a jointly held capacity—that of being “capable of remembering.” But what the king opposes to the alleged potion is instead recollection (ana-). And what he sees in the features of the potion is not memory but a hupomnēsis, a memory by default; that is, a technique offering something “certain” (saphes) and “clear” to those naïve people who believe “that words that have been written down [logous gegrammenous] can do more than remind [hupomnēsai] those who already know what the writing is about” (275c–d). Again it is memory by default (which I am proposing to call memorization) that is at issue here.

The narrative continues: writing is compared with painting (zōgraphia), whose works present themselves “as if they are alive [hōs zōnta].” We ought not to be surprised by this comparison. It imposed itself during our discussion about the imprint on the wax.5 In effect, we have passed from the metaphor of imprinting to that of writing, another variety of inscription. Therefore it is really inscription in the generality of its signification that is at issue. But it remains that the kinship with painting is perceived as disturbing (deinon, “strange”) (275d). We shall speak further of this when we confront narrative and picture at the properly literary level of historiography: the picture makes one believe in the reality through what Roland Barthes calls a “reality effect,” which, as is well known, condemns the critic to silence. This is certainly the case with “written discourse”: “it continues to signify just that very same thing forever” (275d). Yet, where is the repetitive side more clearly indicated in a nonproblematic way if not in memorized writings, learned by heart? The case turns out to be even more damning: written down once and for all, the discourse is in quest of some interlocutor, whoever it may be—one does not know to whom it is addressed. This is also the case for the historical narrative that gets written and published: it is tossed to the winds, it is addressed, as Gadamer says of Schriftlichkeit, to whomever knows how to read. There is a parallel vice: questioned, “it can neither defend itself nor come to its own support” (275e). This is certainly the case for a history book, as for any book. It has cut its ties to its speaker. What I elsewhere have called the semantic autonomy of the text here is presented as a situation of distress. The help this autonomy deprives it of can only come through the interminable work of contextualization and recontextualization that makes up reading.

But then, what quality does that other kind of discourse—“a legitimate brother of this one” (276a)—that of true memory, offer? “It is a discourse that is written down, with knowledge, in the soul of the listener; it can defend itself, and it knows for whom it should speak and for whom it should remain silent” (276a). This discourse that can defend itself before the one to whom it is well fitted is the discourse of true, happy memory, assured of being “timely” and of being capable of being shared. However, the opposition to writing is not total. The two modes of discourse remain akin, like brothers, in spite of their difference as regards legitimacy. Above all, both are written down, inscribed. But it is in the soul that the true discourse is written.6 It is this underlying kinship that allows us to say that “the written one can be fairly called an image [eidōlon]” (276a) of what is “living,” “breathing” in memory. The metaphor of life introduced above, with the painting of living beings, can thus be shifted to the fields of the sensible farmer who knows how to plant, grow, and harvest. For true memory, inscription is a kind of sowing, its true words are “seeds” (spermata). Thus we are authorized to speak of “living” writing, for this writing in the soul and for these “gardens of letters” (276d). Here, despite the kinship among these logoi, lies the gap between a living memory and a dead deposit. This remnant of writing at the very heart of memory authorizes our envisaging writing as a risk to run: “When he [the farmer] writes, it’s likely he will sow gardens of letters for the sake of amusing himself, storing up reminders for himself ‘when he reaches forgetful old age’ and for everyone who wants to follow in his footsteps, and will enjoy seeing them sweetly blooming” (276d). Forgetfulness is named for a second time. Above it was entailed by the alleged gift of writing. Now it is something undergone as a consequence of old age. But it does not lack the promise of amusement. Do we not have then a struggle against forgetfulness that preserves the kinship between “the abusive and the legitimate brother”? And, faced with forgetfulness, playfulness? Playfulness that will be welcomed by those old graybeards Nietzsche will condemn in his second Unfashionable Observation. But how serious is the game that animates those discourses that have as their object justice and as their method, dialectic! A game in which one takes pleasure, but equally a game where one is as happy as a human being can be: the just person, in effect, finds himself crowned with beauty (277a)!

The transition in terms of forgetting and games is so essential that the dialogue can elevate itself to another level, that of dialectic, where the opposition between living memory and dead deposit becomes secondary. We have moved beyond the violence of the myth that led to overstatement and entered into philosophy (278a). Discourses, to be sure, are “written in the soul,” but they bring aid to the writings that vouch for this memory which is a memory on crutches (hupomnēsis).

The case of Lysias, Socrates’ target from the beginning of this dialogue, can serve as a touchstone. The case against him is not that he writes down his discourses, but rather that these preach against art, where the art that he lacks is that of definitions, divisions, and the organization of a discourse as colorful as a multicolored soul. So long as one does not know “the truth concerning everything you are speaking or writing about” (277b), one will lack mastery of the how to use “speech artfully [to log ōn genos]” (277c) considered in terms of its full amplitude, which includes political documents. What is at issue, then, is not just epistemological, in that truth is at stake, but ethical and aesthetical, in that the question is to understand “whether it is noble or shameful [aiskhron] to give or write a speech” (277d). Why then does writing not have the “clear knowledge of lasting importance” that the myth had reserved for memory? Is this not the case with laws? The blame does not fall on writing as such but on the relation of the discourse to the just and the unjust, the bad and the good. It is with regard to this criterion that discourses “written in the soul” win out over all others and why one must bid farewell to all these others (278a).

Is this farewell also addressed to the pharmakon of the myth? We are not told. We do not learn whether philosophical discourse is capable of conjuring up the equivalent of a potion concerning which we never know whether it is healing or poisonous.

What would be the equivalent of this indecisive situation for our attempt to transpose the myth from the Phaedrus to the plane of the relations between living memory and written history? To the outcome of a prudent rehabilitation of writing and the outline of a family reunion between the bastard and the legitimate brother at the end of the Phaedrus will correspond, on our side, a stage where, on the one hand, an educated memory, illuminated by historiography, perfectly overlaps, on the other, a scholarly history capable of reanimating a fading memory and thereby, in Collingwood’s terms, in “reactualizating,” “reliving” the past. But is not this wish condemned to remain unsatisfied? In order to be fulfilled, the suspicion would have to be exorcized that history remains a hindrance to memory, just like the pharmakon of the myth, where in the end we do not know whether it is a remedy or a poison, or both at once. We shall have to allow this unavoidable suspicion to express itself again more than one time.


The Documentary Phase: Archived Memory


The initial chapter in this second part is devoted to the documentary phase of the historiographical operation, on the basis of the tripartite division of the tasks proposed above. We shall not forget that within this phase we do not have in mind chronologically distinct stages of the whole enterprise, but rather levels within a research program that are distinguished only by the distantiated epistemological gaze. This phase, taken in isolation, presents itself as a meaningful sequence whose stages lend themselves to discrete analysis. The terminus a quo is still memory grasped at its declarative stage. The terminus ad quem has the name: documentary proof. Between these two extremes unfolds a quite broad interval that I shall demarcate in the following way. We shall first pinpoint the switch to history from memory on the formal level of space and time. I shall then seek what on this level of the historiographical operation can be the equivalent of a priori forms of experience as they are determined by a transcendental aesthetics in a Kantian style: what is it that makes for a historical time and a geographical space, allowing for the fact that they cannot be articulated separately from each other? (section 1: “Inhabited Space,” and section 2: “Historical Time”).

Passing from form to content, from historical space-time to things said about the past, I shall follow the movement thanks to which declarative memory externalizes itself in testimony. I shall give all its force to the witness’s commitment in his testimony (section 3: “Testimony”). I shall dwell then awhile on the moment of inscription of testimony that is received by another. This moment is the one when things said tip from the oral field to that of writing, which history will not henceforth abandon. It is also the moment of the birth of the archive, collected, preserved, consulted. Passing through the door of archives, testimony enters the critical zone where it is not only submitted to the harsh confrontation among competing testimonies, but absorbed into a mass of documents that are not all testimonies (section 4: “The Archive”). The question will then arise about the validity of documentary proof, the first component of proof in history (section 5: “Documentary Proof”).

Considered in light of the myth from the Phaedrus, these steps taken together denote a tone of assurance as regards the well-foundedness of the confidence placed in the capacity of historiography to enlarge, correct, and criticize memory, and thereby to compensate for its weaknesses on the cognitive as much as on the pragmatic plane. The idea we shall confront at the beginning of part 3, that memory can be divested of its function of being the birthplace of history to become one of its provinces, one of the objects it studies, certainly finds its greatest backing in the self-confidence of the historian “sitting down to work,” the historian in the archives. It is good that it should be so, if only to disarm those who negate great crimes who will find their defeat in the archives. In the following stages of the historiographical operation there will be stronger reasons for not simply celebrating this victory over the arbitrary that is the glory of archival labor.

Yet we must not forget that everything starts, not from the archives, but from testimony, and that, whatever may be our lack of confidence in principle in such testimony, we have nothing better than testimony, in the final analysis, to assure ourselves that something did happen in the past, which someone attests having witnessed in person, and that the principal, and at times our only, recourse, when we lack other types of documentation, remains the confrontation among testimonies.



The impetus given the present investigation by taking up the myth from the Phaedrus leads to our organizing our reflection around the notion of inscription, whose amplitude exceeds that of writing in the precise sense of the fixation of oral expressions of discourse by a material support. The dominant idea is that of external marks adopted as a basis and intermediary for the work of memory. In order to preserve the amplitude of this notion of inscription, I shall first consider its formal conditions, namely, the mutations affecting the spatiality and temporality of living memory, whether collective or private. If historiography is first of all archived memory and if all the subsequent cognitive operations taken up by the epistemology of historical knowledge proceed from this initial gesture of archiving, the historian’s mutation of space and time can be taken as the formal condition of possibility for this gesture of archiving.

A parallel situation to the one that lies at the origin of Kant’s transcendental aesthetic associating the destiny of space with that of time is recognizable here. In passing from memory to historiography, the space in which the protagonists of a recounted history move and the time in which the told events unfold conjointly change their sign. The explicit declaration of the witness, whose profile we shall take up below, states this clearly: “I was there [j’y étais].” The use of the grammatical imperfect tense in French indicates the time, while the adverb marks the space. Together the here and there of the lived space of perception and of action, and the before of the lived time of memory, find themselves framed within a system of places and dates where the reference to the here and absolute now of lived experience is eliminated. That this double mutation can be correlated with the position of writing in relation to orality is confirmed by the parallel constitution of two sciences, geography, on the one hand, seconded by cartography (I think here of the imposing gallery of maps in the Vatican museum!), and historiography, on the other.

Following Kant’s transcendental aesthetic, I have chosen to take up the pair space/time starting from the side of space. The moment of exteriority, common to every “external mark” characteristic of writing according to the myth from the Phaedrus, then finds itself immediately underscored. What is more, the alternation of continuities and discontinuities that mark the historical mutation of these two a priori forms is then easy to decipher.

At the beginning, we have the corporeal and environmental spatiality inherent to the evocation of a memory. To make sense of this, in part 1 I opposed the worldliness of memory to its reflexive pole. The memory of having inhabited some house in some town or that of having traveled in some part of the world are particularly eloquent and telling. They weave together an intimate memory and one shared by those close to one. In memories of this type, corporeal space is immediately linked with the surrounding space of the environment, some fragment of inhabitable land, with its more or less accessible paths, its more or less easy to cross obstacles. Thinkers in the middle ages would have said that our relation to the space open to practice as well as to perception is “arduous.”

From such shared memory, we pass by degrees to collective memory and its commemorations linked to places consecrated by tradition. It is the occurrence of such experiences that first introduced the notion of sites of memory, prior to the expressions and fixations that have subsequently become attached to this expression.

The first milestone along the way of the spatiality that geography sets in parallel with the temporality of history is the one suggested by a phenomenology of “place” or “site.” We owe the former to Edward Casey, from whom I have already borrowed important insights having to do precisely with the worldliness of the mnemonic phenomenon.1 If the title chosen suggests something like a nostalgia desirous of “putting things back in their place,” it has to do with the adventure of a being of flesh and bones who, like Ulysses, is in his place as much in the places visited as upon his return to Ithaca. The navigator’s wanderings demand their right no less than does the residence of the sedentary person. To be sure, my place is there where my body is. But placing and displacing oneself are primordial activities that make place something to be sought out. And it would be frightening not ever to find it. We ourselves would be devastated. The feeling of uneasiness—Unheimlichkeit—joined to the feeling of not being in one’s place, of not feeling at home, haunts us and this would be the realm of emptiness. But there is a question of place because space is not yet filled, not saturated. In truth, it is always possible, often urgent, to displace oneself, with the risk of becoming that passerby, that wanderer, that flâneur, that vagabond, stray dog that our fragmented contemporary culture both sets in motion and paralyzes.

Investigation into what “place” signifies finds support in ordinary language, which includes expressions such as emplacement and displacement, expressions that usually come in pairs. They speak of experiences of the lived body that demand being spoken of in a discourse prior to Euclidean or Cartesian space, as Merleau-Ponty emphasized in his Phenomenology of Perception. The body, the absolute here, is the landmark for any there, be it near or far, included or excluded, above or below, right or left, in front or behind, as well as those asymmetric dimensions that articulate a corporeal typology that is not without at least implicit ethical overtones, for example, height or the right side. To these corporeal dimensions are added some privileged postures—upright, lying down—weightiness—heavy, light—orientations to front or rear, the side, all determinations capable of opposed values: active man, standing upright, someone sick and also the lover lying down, joy that awakens and arises, sadness and melancholy that lower the spirits, and so on. To these alternatives of rest and movement is grafted the act of inhabiting, which has its own polarities: reside and displace, take shelter under a roof, cross a threshold and go out. One might think here of the exploration of a house, from basement to attic, in Gaston Bachelard’s Poetics of Space.

In truth, displacements of the body and even its remaining in place cannot be spoken of, nor even thought, nor even at the limit experienced without some, at least allusive, reference to points, lines, surfaces, volumes, distances inscribed on a space detached from the reference to the here and there inherent to the lived body. Between the lived space of the lived body and the environment and public space is intercalated geometric space. In relation to it, there is no longer any privileged place but only different localities. The act of inhabiting is situated at the boundaries of lived space and geometric space. And this act of inhabiting is put in place only by an act of construction. Hence, it is architecture that brings to light the noteworthy composition that brings together geometric space and that space unfolded by our corporeal condition. The correlation between inhabiting and constructing thus takes place in a third space—if we want to adopt a concept parallel to that of the third time that I propose for the time of history, spatial localities corresponding to dates on the calendar. This third space can also be interpreted as a geometrical checkering of lived space, one of “places,” like a superimposition of “places” on the grid of localities.

As for the act of constructing, considered as a distinct operation, it brings about a type of intelligibility at the same level as the one that characterizes the configuration of time by emplotment.2 Between “narrated” time and “constructed” space there are many analogies and overlappings. Neither reduces to the fragments of the universal time and space of geometers. But neither do they oppose a clear alternative to them. The act of configuration takes place at the point of rupture and suture of two levels of apprehension: constructed space is also geometrical, measurable, and calculable space. Its qualification as a lived place superimposes itself upon and is interwoven with its geometrical properties in the same way that narrated time weaves together cosmic and phenomenological time. Whether it be fixed space or space for dwelling, or space to be traversed, constructed space consists in a system of sites for the major interactions of life. Narrative and construction bring about a similar kind of inscription, the one in the endurance of time, the other in the enduringness of materials. Each new building is inscribed in urban space like a narrative within a setting of intertextuality. And narrativity impregnates the architectural act even more directly insofar as it is determined by a relationship to an established tradition wherein it takes the risk of alternating innovation and repetition. It is on the scale of urbanism that we best catch sight of the work of time in space. A city brings together in the same space different ages, offering to our gaze a sedimented history of tastes and cultural forms. The city gives itself as both to be seen and to be read. In it, narrated time and inhabited space are more closely associated than they are in an isolated building. The city also gives rise to more complex passions than does the house, inasmuch as it offers a space for displacement, gathering, and taking a distance. There we may feel astray, rootless, lost, while its public spaces, its named spaces invite commemorations and ritualized gatherings.

It is at this point that Casey’s final reflections take on strength.3 The attraction of wild nature emerges reinforced by the opposition between the constructed and the nonconstructed, between architecture and nature. This latter does not allow itself to be marginalized. The best of civilization cannot abolish the primacy of the wilderness. The experience of the first American colonists, handed over to the two traumatic experiences of uprooting and desolation, which become legendary, returns in force with the dark moods of uprooted city dwellers that the countryside and its landscapes no longer comfort. Only those who, like Casey, aspire to the calmness and stability of house and home, can aspire to going wild in the Land, leaving an escape from the Unheimlichkeit of the wilderness, even of such a friendly setting as that envisaged by Thoreau in Walden. Even in France we have our Du Bellay and his “petit Liré.”

These incidental notes must not cover over the permanent lesson of the Odyssey, a narrative that weaves together events and places, an epic that celebrates episodes and stops along the way as much as it does the indefinitely delayed return, the return to Ithaca that is supposed to “return things to their place.” Joyce, Casey recalls, wrote in his preparatory drafts to his Ulysses: “Topical History: Places Remember Events.”4

But, to give the time of history a spatial analogue worthy of a human science we must elevate it higher on the scale of the rationalization of places. We have to move from the constructed space of architecture to the inhabited land of geography.

That geography, within the order of the human sciences, constitutes the exact guarantor of history, is still not to say much. In France, geography began by anticipating certain methodological conversions in history that will concern us below.5 Vidal de La Blache was the first, before Martonne, to react against the positivism of historicizing history and to give meaning to the notions of setting, lifestyle, and everydayness. His science was a geography in the sense that its object was above all one of “places,” “countrysides,” “of visible effects on the earth’s surface that were both natural and human” (Dosse, 15). The geometrical side of the experience of space was visualized by the cartography whose mark we shall rediscover when we consider the interplay of scales below. The human side is marked by the concepts of biological origin, cell, tissue, organism.

What was to influence the history of the Annales school was, on the one hand, the accent placed on things that were permanent, represented by the stable structures of the countryside, and, on the other, the preference for description expressed by the flourishing of regional monographs. This attachment to territory, principally rural landscapes, will find more than an echo in the Annales school with the promotion of a veritable geopolitics where the stability of landscapes and the quasi-immobility of the long time span are conjoined. Space, Braudel liked to say, was the best means of slowing down history. These spaces were in turn those of regions and those of seas and oceans. “I loved the Mediterranean passionately,” he declared in his great book where the Mediterranean was both site and hero. As Lucien Febvre wrote to Braudel: “Between these two protagonists, Philip and the interior ocean, the match is not equal” (quoted in Dosse, 108). As for the question that led to the preceding observations, that of the switch-over from the space of geographers and historians in relation to the space of lived experience, itself anchored in the range of the body and its environment, we must not focus exclusively on the break between them. Above, I referred to the schema of an alternation of ruptures, sutures, and reprises at a higher level of determinations stemming from the existential level. Geography is not geometry insofar as the land surrounded by oceans is an inhabited one. This is why geographers of the school of Vidal de La Blache speak of it as a “milieu.” But the milieu, as we learn from Canguilhem, is one pole of a debate—an Auseinandersetzung—where the living creature is the other pole.6 In this respect, the emphasis on possibility of Vidal de La Blache anticipates the dialectics of a von Uexküll and a Kurt Goldstein. And, if in Braudel’s geohistory milieu and space are taken as equivalent terms, the milieu remains that of life and civilization. “Any civilization is at bottom a space worked by men and history,” we read in The Mediterranean and the Mediterranean World in the Age of Philip II.7 And again: “What is a civilization if not the timeworn placement of a certain humanity in a certain space?” (cited by Dosse, 109). It is this mixture of climate and culture that makes up geohistory, which in turn determines the other levels of civilization, according to modes of interconnectedness that we shall discuss in the following chapter. The geopolitical perspective can be taken as “more spatial and temporal” (110), but it is so in relation to the level of institutions and events, which is that of layers built upon the geographical soil and in their turn placed under the constraint of structures of a temporal nature. I had earlier noted, in my attempt to re-narrativize Braudel’s great work, and to read it in terms of the great plot of The Mediterranean, that the first part, where space is said to be the theme, is a peopled space. The Mediterranean itself is the interior sea, a sea between inhabited or uninhabitable, hospitable or inhospitable lands. This space is the setting for the inscription of slower oscillations than those known by history.8

Braudel’s other great work, Civilization and Capitalism, 15th–18th Century, calls for similar considerations.9 What succeed each other in time are “world economies” inscribed in space but articulated in terms of places qualified by human activities and divided into concentric circles whose centers change over time. This “differential geography” (Dosse, 125) never leaves space without the interaction of exchange that binds an economy to a geography and distinguishes this latter from simple geometry.

In conclusion, from the phenomenology of “places” that beings of flesh and blood occupy, leave, lose, rediscover—in passing through the intelligibility belonging to architecture—up to the geography that describes an inhabited space, the discourse of space too has traced out an itinerary thanks to which lived spaced is turn by turn abolished by geometrical space and reconstructed at the hyper-geometrical level of the oikoumenē.10


To the dialectic of lived space, geometrical space, and inhabited space corresponds a similar dialectic of lived time, cosmic time, and historical time. To the critical moment of localization within the order of space corresponds that of dating within the order of time.

I shall not repeat my analysis of calendar time from Time and Narrative.11 My focus is different today inasmuch as it is not so much the reconciliation of the phenomenological and cosmological perspectives on time that is at issue as the transition from living memory to the “extrinsic” positing of historical knowledge. Thus it is as one of the formal conditions of possibility of the historiographical operation that the notion of a third-order time reappears.

I will limit myself to the definition that Benveniste gives of “chronicle time,” which I am calling third-order for the sake of my argument: (1) the reference of every event to a founding event that defines the axis of time; (2) the possibility of traversing the intervals of time in terms of the two opposed directions of anteriority and posteriority in relation to the zero date; and (3) the constitution of a repertory of units serving to name recurring intervals: day, month, year, and so on.

It is this constitution that we need now to place in relation to the historian’s mutation of the time of memory. In one sense, dating, as a phenomenon of inscription, is not without some connections to a capacity for dating, in an originary datability, inherent to lived experience, and singularly to a feeling of being distanced from the past and of having a sense of temporal depth. Aristotle in De memoria et reminiscentia takes for granted that simultaneity and succession characterize in a primitive manner the relations between remembered events. Otherwise, there would be no question, in the work of recalling, of choosing a starting point in order to reconstruct the interconnectedness of such events. This primitive character of a sense of intervals results from the relationship time maintains with movement. If time “has something to do with movement,” a soul is required, in order to distinguish two instants, to relate them to each other as before and after, and to evaluate their difference (heteron) and to measure the intervals (to metaxu), operations thanks to which time can be defined as the “number of motion in respect of ‘before’ and ‘after’” (Physics 4.219b). As for Augustine, who is hostile to any subordination of time to physical movement, he admires in the rhetorician the power of the soul to measure within itself lengths of time, and thereby to compare long and short syllables on the level of diction. For Kant, the notion of temporal extension makes for no difficulty. It does not result from a possibly unwarranted second-order comparison with spatial extension, but rather precedes it and makes it possible. Husserl takes the relations of time relative to its passage as inseparable a priori from “apprehensions” immanent to the inner experience of time. Finally, even Bergson, the philosopher of the durée, does not doubt that in pure memory the evoked event comes with its date. For all of them, extension appears as a primitive fact, as attested in language by questions such as “when,” “since how long ago,” “for how long,” which belong to the same semantic plane as do declarative memory and testimony. To the declaration “I was there” is added the affirmation “this happened ‘before,’ ‘during,’ ‘after,’ ‘since,’ ‘during so much time.’”

Having said this, what calendar time adds consists in a properly temporal mode of inscription, namely, a system of dates extrinsic to the events to which they apply. Just as in geographical space the places referred to the absolute “here” of the lived body and its environment become particular locations that can be inscribed among the sites that cartography maps, so too the present moment with its absolute “now” becomes a particular date among all the ones whose exact calculation is allowed for by the calendar in terms of the framework of some calendar system accepted by a more or less extended part of humanity. As concerns the time of memory in particular, the “another time” of the remembered past is henceforth inscribed within the “before that” or the dated past. Symmetrically, the “later” of expectation becomes the “when that,” marking the coincidence of an anticipated event with the grid of dates to come. Every noteworthy coincidence refers in the final analysis to those events, in chronological time, between some social event and an astrally based cosmic configuration. In the pages devoted above to the ars memoriae, we had plenty of time to take the measure of the incredible exploitation that subtle minds have given to calculations in service of an insane dream of mastery over human destiny. Those times of such exploits of memorization are no longer our own, but in many ways our lives in common remain governed by such calculation of dated conjunctions. The distinctions familiar to economists, sociologists, and political scientists, to say nothing of historians, between short term, mid-term, long term, cycle, period, and so on—distinctions to which we shall return below—are all inscribed on the same calendar time where the intervals between dated events allow themselves to be measured. The brevity of human life stands out against the immensity of indefinite chronological time.

In turn, calendar time stands out against a rising series of representations of time that cannot be reduced to what phenomenology knows as lived time. Thus Krzysztof Pomian, in L’Ordre du temps, distinguishes “four ways of visualizing time, of translating it into signs”: chronometry, chronology, chronography, and chronosophy.12 This order stems essentially from a kind of thinking that overflows that of the knowable (to use the Kantian distinction between Denken and Erkennen) within whose limits historians’ history confines itself. As thinkable, these articulations ignore the distinction between myth and reason, between philosophy and theology, between speculation and symbolic imagination. These considerations from the preface to L’Ordre du temps have a lot to say to our inquiry. We ought not to believe, for example, that historical knowledge has only collective memory for its opposite. It has also to conquer its space of description and explanation against the speculative background deployed by the problems of evil, love, and death. This is why the categories closest to the historian’s practice that Pomian considers over the course of his book—events, repetitions, ages, structures—stand out against the fourfold frame of the order of time. We can again recognize calendar or chronological time in the times of chronometry and chronology. The first of these designates the short or long cycles of time that recur, that return in cycles: day, week, month, year. The second designates the linear time of long periods: century, millennium, and so forth, whose scansion is punctuated in diverse ways by founding events and founders; cycles that take place over a number of years, such as, for example, the Greek Olympiads. These are the two kinds of time measured by clocks and calendars, with the reservation that the intervals of chronology—such as eras—have a signification that is as much qualitative as quantitative. Chronology, which is closer to the historian’s intention, knows how to order events as a function of a series of dates and names, and to order the sequence of eras and their subdivisions. But it ignores the separation between nature and history. It allows us to speak of cosmic history, of the history of the earth, the history of life where human history is just one segment. With chronography, we come to systems of notation that can go beyond the calendar. The noted episodes are defined by their relations to other episodes: a succession of unique, good or bad, joyful or sorrowful events. This time is neither cyclic nor linear, but amorphous. It is what relates the presented chronicle to the narrator’s position, before narrative detaches the told tale from its author. As for chronosophy, which will take more of our time, it exceeds the project of a critical history that has become our project. It has been cultivated by numerous families of thought that arrange times in terms of rich typologies opposing stationary time to reversible time, which may be cyclical or linear. The history we may construct of these great schemes is equivalent to a “history of history,” from which professional historians may never completely free themselves, once it is a question of assigning a significance to facts: continuity vs. discontinuity, cycle vs. linearity, the distinction of periods or eras. Once again, it is not principally the phenomenology of lived times or the exercises of popular or scholarly narratives that history confronts here, but an order of thought that ignores the sense of limits. And the categories that come from it have not ceased to construct the temporal “architecture” of “our civilization” (xiii). In this regard, the time of history proceeds as much by limiting this immense order of what is thinkable as by surpassing the order of lived experience.

It is principally on the basis of such great chronosophies of speculation on time that historical time was conquered at the price of a drastic self-limitation. I will retain from Pomian’s rich analyses only what has to do with the persistence of chronosophy on the horizon of the large categories that shape historical discourse in the phase of explanation/understanding and in that of the representation of the past, whether it be a question of “events,” “repetitions,” “ages,” or “structures” (the titles of the first four chapters of his book). These are the same categories that we have come upon more than once in our epistemological inquiry. It is good to know what excess of the thinkable they have conquered before being able to face up to the demand for truth with which history is supposed to confront the trustworthiness of memory. By chronosophy, Pomian means those large-scale periodizations of history such as those of Islam and Christianity (in Daniel and Saint Augustine) and their attempts to make them correspond with chronology. Religious and political chronosophies clash in this field. With the Renaissance appears a periodization in terms of “ages” of art and with the eighteenth century one in terms of “centuries.”

One will readily take the notion of event to be the least speculative among these developments and also as the most self-evident one. Michelet as much as Mabillon, Droysen as much as Dilthey profess with confidence the primacy of the individually determined fact. Reduced to the sphere of visibility, the event’s coming to perception would be unjustifiable. An aura of invisibility which is the past itself encircles it and hands it over to mediations that are the objects of research and not of perception. Along with the invisible, speculation comes into play and proposes a “historical typology of chronosophies” (26). In the Christian West, it is principally in terms of the opposition between profane and sacred history, on the plane of a theology of history, that the relations between the continuous and the discontinuous were conquered. We must not lost sight of this speculative history when we take up in succession the Braudelian plea for a history not based on events and the “return of the event” in the wake of the return of the political, up to the most sophisticated models pairing up event and structure.13

Would one have formed the notion of “repetitions” without the idea of a direction and a signification that were first provided by a typology on the chronosophical level?We owe to this latter the opposition between a stationary time and a nonrepeatable one, whether cyclic or linear, and, in the latter case, either progressive or regressive. It is from these large-scale orientations that the present receives a meaningful place in history as a whole. Then we speak of ages, centuries, periods, stages, epochs. Like the notion of event, that of the architecture of historical time is conquered through the disintegration of the overall time of history, from which emerges the problem of the relationships between different local times. But have we stopped adding to proposals of the type that Bernard of Chartres spoke of in talking about the “visual acuity” of dwarfs in comparison to the “greatness” of the giants upon whose shoulders they were perched? Have we renounced opposing a time of rebirth to one of darkness, spotting oscillations indicated by some cyclical phenomena, watching out for advances and retreats, extolling the return to sources, protecting taste and customs from the corruption of the cumulative effects of history? Are there no more battles between the Ancients and the Moderns? Haven’t we read and understood Vico and Turgot? The “struggle of the chronosophy of progress” (58) against the specter of philosophies of regression undoubtedly has not disappeared from our horizons. The plea for or against modernity that I shall refer to below continues to borrow from this panoply of arguments. We do not readily admit the chronosophic status of the idea of a cumulative and irreversible linear time, still familiar to professional historians. The chronosophy of cyclical time at the turn of the twentieth century suffices to recall it. And do not the cycles so dear to economists ever since the takeoff of the history of prices and economic fluctuations, with Ernest Labrousse among others, point us in the direction of a synthesis of cyclical and linear time? Even the piling up of time-spans, in the manner of Braudel, and the attempt joined to it to articulate conjuncture and event in terms of a triadic structure, poorly conceals the chronosophic residue that hides itself behind a scientific façade. In this sense, breaking away from every chronosophy, to the benefit of a certain methodological agnosticism concerning the direction of time, has not been achieved. Perhaps it is not desirable that it should be, if history is to remain interesting, that is, if it is to continue to speak of hope, nostalgia, anxiety.14

The concept of ages (Pomian, chap. 3) is perhaps the most troublesome, inasmuch as it seems to be superimposed on chronology in order to cut it into large periods. For example, in the West we continue to divide the teaching of history and even historical research into antiquity, the Middle Ages, early modernity, and the modern world. Here we recall the role that Émile Benveniste assigns to the zero point in the calculation of historical time. The birth of Christ for the Christian West, the Hegira for Islam. Yet periodizations have an even richer history as far back as Daniel’s dream as recounted in the Hebrew Bible, leading to the theory of four monarchies according to Augustine. Then we find successive quarrels between Ancients and Moderns, which play on rival periodizations. The comparison of ages of life also has its adepts, as does doubt concerning the historical replica of biological aging: Does history know an old age that does not lead to death? In truth, the concept of periods does not lend itself to a history distinct from that of cyclical or linear, or stationary or regressive conceptions. Hegel’s Philosophy of History offers in this regard an impressive synthesis of multiple ways of ordering historical time. And after Hegel, despite my vow to “renounce Hegel,” the question arises anew whether every chronosophic residue has disappeared from the use of terms such as “stages” adopted in economic history, on the plane where cycles and linear segments intersect. What is at stake is nothing less than the possibility of a history without direction or continuity. It is here, according to Pomian, that the theme of structure takes over from that of periods.15

But can one do history without periodization? I mean, not merely teach history, but produce it? According to Claude Lévi-Strauss, we would have “to spread out in space those forms of civilization which we imagined as spread out in time” (337). Were we to succeed, would this not be to remove from history any horizon of expectation, to us a concept often referred to in this work that I owe to Reinhart Koselleck? Even for Lévi-Strauss history cannot withdraw into an idea of an extended space without any horizon of expectation, inasmuch as “it is only from time to time that history is cumulative—in other words, that the numbers can be added up to form a favorable combination” (338).

The mark of the great chronosophies of the past is less easy to discern at the level of “structures,” in which Pomian sees the fourth articulation of time. I want to show its role as one phase of the historiographical operation, where the notion of structure enters into variable compositions along with those of conjuncture and event. But it is worth recalling its birth from large-scale speculations on the movement of global history. The human and social sciences have certainly given it an operative dimension. Yet the mark of its origin in speculation can still be recognized in “the split within each [of these disciplines], setting aside a few rare exceptions, into theory and history” (165). The autonomy of the theoretical in relation to the experimental was first conquered in biology, in conjunction with linguistics and anthropology. Structures are new objects, theoretical objects, endowed with a demonstrable reality or existence, in the same way that one demonstrates the existence of a mathematical object. Within the human sciences, this split between theory and history is due to Saussurian linguistics and “the simultaneous entry of theory and the object-structure into the field of the human and the social sciences” (168). Theory must deal only with atemporal entities, leaving to history the question of beginnings, developments, and genealogical trees. Here the object-structure is langue, language as a system, distinguished from parole, language as actually used, as speech. More will be said below about the happy and unhappy effects of this transposition from the linguistic domain as it has affected the historiographical incorporation of this linguistic model by those who followed Saussure—in particular, that the notions of diachrony and synchrony lose their phenomenological basis when they enter into a structural system. Conciliation between an approach based on system, as the enemy of the arbitrary, and a historical one, set out in terms of discrete events, itself becomes the object of speculation, as we see in Roman Jakobson (cf. Pomian, 174). And history as a discipline finds itself indirectly caught up in the reintegration of linguistics as a science into the space of theory as well as by the overlapping within this same space of studies of literary, and in particular of poetic language. But it is also the claim to dissolve history into a logical or algebraic combinatrics, in the name of the correlation between process and system, that the theory of history has had to deal with in the last third of the twentieth century, almost as though structuralism had given historiography a perfidious kiss of death.16 My own recourse to models stemming from the theory of action will inscribe itself within this revolt against the hegemony of structuralist models, but not without retaining something of the imprint they have exercised on the theory of history; for example, the concepts of transition as important as those of competence and performance, taken from Noam Chomsky, retailored to the scale of the relationship among the notions of agent, agency (in Charles Taylor’s sense), and structures of action, such as constraints, norms, and institutions. Equally to be rediscovered and rehabilitated are prestructuralist philosophies of language, such as that of Humboldt, which give to the spiritual dynamism of humanity and its productive activity the power to engender gradual changes of configuration. “For spirit,” proclaimed Humboldt, “to be is to act.” History was recognized in this generative dimension. Yet professional historians, who might take interest in Humboldt, cannot overlook the highly theoretical dimension of his reflections, such as those cited by Pomian: “Taken in its essential reality, language is continually changing and at each instant in the midst of some anticipatory transition. . . . In itself, language is not a work done (ergon) but an activity in the process of happening (energeia). Thus its true definition can only be genetic” (cited, 209).17

This long excursus devoted to the speculative and highly theoretical past of our notion of historical time has had a single goal, to recall to historians a number of things:

—The historiographical operation proceeds from a double reduction, that of the lived experience of memory, but also that of the multimillenary speculation on the order of time.

—The structuralism that has fascinated several generations of historians stems from a theoretical stance that, through its speculative side, is situated along the prolonging of the great theological and philosophical chronosophies, as a kind of scientific, even scientistic chronosophy.

—Historical knowledge perhaps has never, in fact, stopped dealing with these visions of historical time, when it speaks of cyclical or linear time, stationary time, decline or progress. Will it not then be the task of a memory instructed by history to preserve the trace of this speculative history over the centuries and to integrate it into its symbolic universe? This will be the highest destination of memory, not before but after history. The palace of memory, we have read in Augustine’s Confessions, not only holds the memories of events, the rules of grammar, and rhetorical examples, it also preserves theories, including those that, claiming to embrace it, have threatened to eliminate it.


Testimony takes us with one bound to the formal conditions of the “things of the past” (praeterita), the conditions of possibility of the actual process of the historiographical operation. With testimony opens an epistemological process that departs from declared memory, passes through the archive and documents, and finds its fulfillment in documentary proof.

As a first step, I shall take up testimony as such while holding in suspense the moment of inscription that is archived memory. Why this delay? For several reasons. First of all, testimony has several uses: archiving in view of consultation by historians is only one of them, beyond the practice of testimony in daily life and parallel to its judicial use sanctioned by a tribunal’s passing judgment. Furthermore, at the very interior of the historical sphere, testimony does not run its course with the constitution of archives; it reappears at the end of the epistemological inquiry at the level of the representation of the past through narrative, rhetorical devices, and images. Moreover, in some contemporary forms of deposition arising from the mass atrocities of the twentieth century, it resists not only explication and representation, but even its being placed into some archival reserve, to the point of maintaining itself at the margins of historiography and of throwing doubt on its intention to be truthful. Which is to say that in this chapter we shall follow only one of the destinies of testimony, the one sealed by its being placed into an archive and sanctioned by documentary proof. Whence the interest and importance of an attempt at an analysis of the essence of testimony as such, while respecting its potentiality for multiple uses. When we shall borrow from one or another of these uses, I shall seek to isolate those features capable of being shared among most of these uses.18

It is within the everyday use of testimony that the common core of its juridical and historical use is most easily discerned. This use brings us immediately face to face with the crucial question: to what point is testimony trustworthy? This question balances both confidence and suspicion. Thus it is by bringing to light the conditions in which suspicion is fomented that we have a chance of approaching the core meaning of testimony. In effect, suspicion unfolds itself all along the chain of operations that begin at the level of the perception of an experienced scene, continuing on to that of the retention of its memory, to come to focus in the declarative and narrative phase of the restitution of the features of the event. The untrustworthiness of witnesses has taken on scientific form within the framework of judicial psychology as an experimental discipline. One of the basic tests consists in imposing on a group of subjects the task of producing a verbal restitution of some filmed scene. This test is supposed to allow measurement of the trustworthiness of the human mind with regard to the proposed operations, whether at the moment of perception, or during that of its retention, or finally when it comes to its verbal restitution. The artifice of this test to which it is important to draw our attention is that it is the experimenter who defines the conditions of the test and who validates the reality status of the fact to be attested to. This reality status is taken for granted in the very putting together of the experiment. It is thus the gap between this reality recognized by the experimenter that is taken into consideration and measured. The implicit model of this presupposition is the undisputable trustworthiness of the camera’s eye. The results of these experiments are certainly not negligible. They have to do with the flagrant presence of distortions between the reality known in this way and the depositions made by the laboratory subjects. For me, the question is not to criticize the conclusions of these investigations as disqualifying testimony in general, but rather to call into question, on the one hand, what Dulong calls the “paradigm of recording,” that is, the video camera, and, on the other, the idea of the “disengaged observer,” a prejudice to which the experimental subjects are submitted.

This criticism of the “regulative model” of judicial psychology leads us back to the everyday practice of testimony in ordinary conversation. This approach is profoundly in agreement with the theory of action that will be brought into play in the explanatory and representative phases of the historiographical operation, and with the primacy that will be accorded to the problematic of representation in relation to action on the level of the constitution of the social bond and the identities that stem from it. The activity of testifying, grasped before the bifurcation between its judicial and its historiographical uses, then reveals the same amplitude and the same import as does that of recounting, thanks to the manifest kinship between these two activities, to which we must soon add the act of promising, whose kinship with testimony remains more concealed. Placing into an archive, on the historical side, and a deposition before a tribunal, on the judiciary one, constitute specific uses, governed on the one side by documentary proof and on the other by the passing of a judgment. The use of testimony in ordinary conversation best preserves those essential features of the fact of testifying that Dulong sums up in the following manner: “An autobiographically certified narrative of a past event, whether this narrative be made in informal or formal circumstances” (43).

Let us unpack the essential components of this operation:

1. Two sides are initially distinguished and articulated in terms of one another: on the one side, the assertion of the factual reality of the reported event; on the other, the certification or authentification of the declaration on the basis of its author’s experience, what we can call his presumed trustworthiness. The first side finds its verbal expression in the description of the experienced scene in a narration that, if it does not explicitly mention the implication of the narrator, confines itself to conveying information; the scene, so to speak, recounts itself following the distinction proposed by Émile Benveniste between narrative and discourse. There is an important nuance: This information must be taken to be important; the attested-to fact must be significant, something that renders problematic too sharp a distinction between discourse and narrative. Yet it remains the case that the factuality attested to is supposed to trace a clear boundary between reality and fiction. The phenomenology of memory early on confronted us with the always problematic character of this boundary. And the relation between reality and fiction will continue to torment us, right up to the stage of the historian’s representation of the past. Which is to say that this first component of testimony is a weighty one. It is over against this articulation that a whole battery of suspicions will take their place.

2. The specificity of testimony consists in the fact that the assertion of reality is inseparable from its being paired with the self-designation of the testifying subject.19 The typical formulation of testimony proceeds from this pairing: I was there. What is attested to is indivisibly the reality of the past thing and the presence of the narrator at the place of its occurrence. And it is the witness who first declares himself to be a witness. He names himself. A triple deictic marks this self-designation: the first-person singular, the past tense of the verb, and the mention of there in relation to here. This self-referential character is sometimes underscored by certain introductory remarks that serve as a “preface.” These kinds of assertions link point-like testimony to the whole history of a life. At the same time, the self-designation brings to the surface the inextricable opacity of a personal history that itself has been “enmeshed in stories.” Which is why the affective imprint of an event capable of striking the witness like a blow does not necessarily coincide with the importance his audience may attach to his testimony.

3. Self-designation gets inscribed in an exchange that sets up a dialogical situation. It is before someone that the witness testifies to the reality of some scene of which he was part of the audience, perhaps as actor or victim, yet, in the moment of testifying, he is in the position of a third-person observer with regard to all the protagonists of the action.20 This dialogical structure immediately makes clear the dimension of trust involved: the witness asks to be believed. He does not limit himself to saying “I was there,” he adds “believe me.” Certification of the testimony then is not complete except through the echo response of the one who receives the testimony and accepts it. Then the testimony is not just certified, it is accredited. It is this accreditation, as an ongoing process, that opens the alternative I began with between confidence and suspicion. A questioning argument can be undertaken, which the judicial psychology mentioned above supplies with well-established forms of reasons. This argument may bear on the most common conditions of bad perception, bad memory, or bad restitution. And among these must be taken into account the interval of time so favorable to what Freud in the Interpretation of Dreams calls “secondary elaboration.” It may bear in a more disturbing way on the personal qualities of the testifying subject to be habitually believed, as indicated by similar earlier occasions and the witness’s reputation. In this case, the accreditation comes down to authenticating the witness on personal terms. The result is what we call his trustworthiness, whose evaluation can be assimilated to comparative orders of magnitude.

4. The possibility of suspicion in turn opens a space of controversy within which several testimonies and several witnesses find themselves confronted with one another. In certain general conditions of communication, this space may be qualified as a public space. It is against this background that a critique of testimony is grafted to its practice. The witness anticipates these circumstances in a way by adding a third clause to his declaration: “I was there,” he says, “believe me,” to which he adds, “If you don’t believe me, ask someone else,” said almost like a challenge. The witness is thus the one who accepts being questioned and expected to answer what may turn out to be a criticism of what he says.

5. In this way, a supplementary dimension gets grafted to the moral order meant to reinforce the credibility and trustworthiness of testimony, namely, the availability of the witness to repeat his testimony. The trustworthy witness is the one who can stay steadfast about this testimony over time. This steadfastness makes testimony akin to promise-making, more precisely to the promise that precedes any promise-making, that of keeping one’s promise, of keeping one’s word. Thus testimony links up with promise-making among those acts of discourse that specify ipseity in its difference from simple sameness, the sameness of character or, better, that of one’s genetic make-up, which is immutable from the birth to the death of an individual, the biological basis of his identity.21 The witness must be capable of answering for what he says before whoever asks him to do so.

6. This stable structure of the willingness to testify makes testimony a security factor in the set of relations constitutive of the social bond. In turn, this contribution of the trustworthiness of an important proportion of social agents to the overall security of society in general makes testimony into an institution.22 We can speak here of a natural institution, even if the expression seems like an oxymoron. It is useful for distinguishing this common certification of an account in ordinary conversation from technical, “artificial” uses, consisting, in part, in the placing of things in an archive within the framework of specific institutions, and, in part, in the rule-governed taking of testimony as part of the trial process within the courtroom. I have drawn upon a parallel expression to distinguish the ordinary exercise of remembering from the memory tricks cultivated by the ars memoriae. In this way, we can oppose natural to artificial memory. What makes it an institution is, first of all, the stability of testimony ready to be reiterated, and next the contribution of the trustworthiness of each testimony to the security of the social bond inasmuch as this rests on confidence in what other people say.23 More and more, this bond of trustworthiness extends to include every exchange, contract, and agreement, and constitutes assent to others’ word, the principle of the social bond, to the point that it becomes a habitus of any community considered, even a prudential rule. First, trust the word of others, then doubt if there are good reasons for doing so. In my vocabulary, it is a question of a competence of the capable human being. The credit granted to the word of others makes the social world a shared intersubjective world. This sharing is the major component of what we can call the “sensus communis.” It is what is strongly affected when corrupt political institutions lead to a climate of mutual surveillance, of mistrust, where deceitful practices undercut the basis of confidence in language. We rediscover here, amplified to the scale of the communication structures of a society as a whole, the problematic of manipulated memory referred to above. What confidence in the word of others reinforces is not just the interdependence, but the shared common humanity, of the members of a community. This needs to be said in fine to compensate for the excessive accent placed on the theme of difference in many contemporary theories of the social bond. Reciprocity corrects for the unsubstitutability of actors. Reciprocal exchange consolidates the feeling of existing along with other humans—inter homines esse, as Hannah Arendt liked to put it. This “betweenness” opens the field to dissensus as much as to consensus. And it is dissensus that the critique of potentially divergent testimonies will introduce on the pathway from testimony to the archive. To conclude, in the final analysis, the middle level of security of language of a society depends on the trustworthiness, hence on the biographical attestation, of each witness taken one by one. It is against this background of assumed confidence that tragically stands out the solitude of those “historical witnesses” whose extraordinary experience stymies the capacity for average, ordinary understanding. But there are also witnesses who never encounter an audience capable of listening to them or hearing what they have to say.24


The moment of the archive is the moment of the entry into writing of the historiographical operation. Testimony is by origin oral. It is listened to, heard. The archive is written. It is read, consulted. In archives, the professional historian is a reader.

Before the consulted or constituted archive, there is the archiving of things.25 This brings about a break in a continuous trajectory. Testimony, we have said, gives a narrative follow-up to declarative memory. Yet narrative can be detached from its narrator, as literary criticism informed by structuralism likes to emphasize. But this does not mean that the phenomenologist is left behind. Between the saying and the said of any utterance, a subtle gap opens that allows what is stated, the saying of what is said, to pursue what we can strictly speaking call a literary career. The emplotment of a told story, moreover, reinforces the semantic autonomy of a text, whose composition in the form of a work gives it the visibility of something written.26

To these scriptural features that it shares with narrative, testimony adds specific ones having to do with the exchange structure between the one who gives and the one who receives. Thanks to a reiterable character that confers upon it the status of an institution, testimony can be taken down in writing, deposited. This deposition, in turn, is the condition of possibility of specific institutions devoted to the collecting, conserving, and classifying of documentation with an eye to its subsequently being consulted by qualified personnel. The archive thus presents itself as a physical place that shelters the destiny of that kind of trace I have so carefully distinguished from the cerebral trace and the affective trace, namely, the documentary trace. But the archive is not just a physical or spatial place, it is also a social one. It is in terms of this second angle that Michel de Certeau deals with it in the first of three snapshots of what before me he called the historiographical operation.27 To relate a product to a place constitutes, he says, the first task of an epistemology of historical knowledge: “envisaging history as an operation would be equivalent to understanding it as the relation between a place (a recruitment, a milieu, a profession or business, etc.), analytical procedures (a discipline), and the construction of a text (a literature)” (57). This idea of a social setting of production includes a critical intention aimed against positivism, a critique that Certeau shares with Raymond Aron from the period when he wrote his Introduction to the Philosophy of History (1938). But, unlike Aron, who emphasizes the “dissolution of the object,” it is not so much the subjectivity of authors or personal decisions that Certeau accentuates as the unsaid of the social status of history as an institution of knowledge. In this way he distinguishes himself as well from Max Weber who, in “Politics as a Vocation,” he asserts, “exempted” the power of scholars from the constraints of political society. In confronting this repression of the relation to the society that engenders the unspoken of “place” from which the historian speaks, Certeau, like Habermas, at the time when he was arguing for a “repolitizing” of the human sciences, denounces the appropriation of language by a subject supposed to “control” history’s discourse: “in this way historical discourse takes priority over every particular historical work, and so does the relation of this discourse to a social institution” (63).

However, it does not suffice to set historians back into society if we are to give an account of the process that constitutes a distinct object for epistemology, that is, in Certeau’s own terms, the process leading “from collecting documents to writing books” (66). The multileveled architecture of the social units that constitute archives calls for an analysis of the act of placing materials in such archives, their archiving, capable of being situated in a chain of verifying operations, whose provisory end is the establishing of documentary proof.28 Before explanation, in the precise sense of establishing answers in terms of “because,” there is the establishing of sources, which consists, as Certeau puts it so well, “in redistributing space” that had already been marked out by the collectors of “rarities,” to speak like Foucault. Certeau calls “place,” “what permits, what prohibits” (68) this or that kind of discourse within which cognitive operations properly speaking are enframed.

This gesture of setting things aside, of putting together, of collecting is the object of a distinct discipline, that of the archivist, to which the epistemology of the historiographical operation is indebted for the description of those features by which the archive breaks with the hearsay of oral testimony. To be sure, if writings constitute the principal materials deposited in archives, and if among such writings testimony by past peoples constitutes the core material, all sorts of traces can be archived. In this sense, the notion of the archive restores to the gesture of writing the full scope given to it by the myth in the Phaedrus. At the same time, every plea in favor of the archive will remain in suspense, to the degree that we do not know, and perhaps never will know, whether the passage from oral to written testimony, to the document in the archive, is, as regards its utility or its inconvenience for living memory, a remedy or a poison, a pharmakon.

I propose to set within the framework of this dialectic between memory and history what I said about the notion of an archive in Time and Narrative.29 The accent will be on those features by which the archive breaks from the hearsay of oral testimony. What first stands out is the initiative of a person or legal entity intending to preserve the traces of his or its activity. This initiative inaugurates the act of doing history. Next comes the more or less systematic organization of the material thus set aside. It consists in physical measures of preservation and in logical operations of classification stemming from the needs of a highly developed technique at the level of the archivist. All these procedures are in service of a third moment, that of consulting the materials within the limits of the rules governing access to them.30

If we consider, with all the reservations I shall speak of below, that the essential core of archival materials consists in texts, and if we want to concentrate on those of these texts that are testimonies left by contemporaries having access to this material, the change in status from spoken testimony to being archived constitutes the first historical mutation in living memory that falls under our examination. We can then speak of those written testimonies that the Phaedrus calls “written words”: “When it has once been written down, every discourse rolls everything about everywhere, reaching indiscriminately those with understanding no less than those who have no business with it, and it doesn’t know to whom it should speak and to whom it should not. And when it is faulted and attacked unfairly, it always needs its father’s support; alone it can neither defend itself nor come to its own support” (275e). In one sense, this is a good thing: like all writing, a document in an archive is open to whomever knows how to read. Therefore it has no designated addressee, unlike oral testimony addressed to a specific interlocutor. What is more, the document sleeping in the archives is not just silent, it is an orphan. The testimonies it contains are detached from the authors who “gave birth” to them. They are handed over to the care of those who are competent to question them and hence to defend them, by giving them aid and assistance. In our historical culture, the archive has assumed authority over those who consult it. We can speak, as I shall discuss further below, of a documentary revolution. In a period now taken to be outdated in historical research, work in the archives had the reputation of assuring the objectivity of historical knowledge, protected thereby from the historian’s subjectivity. For a less passive conception of consulting archives, the change in sign that turns an orphan text into one having authority is tied to the pairing of testimony with a heuristics of evidentiary proof. This pairing is common to testimony before a court and testimony gathered by the professional historian. The testimony is asked to prove itself. Thus it is testimony that brings aid and assistance to the orator or the historian who invokes it. As for what more specifically concerns history, the elevation of testimony to the rank of documentary proof will mark the high point of the reversal in the relationship of assistance that writing exercises in regard to “memory on crutches,” that hupomnēmē, or artificial memory par excellence, to which myth grants only second place. Whatever may be the shifts in documentary history—positivism or not—the documentary frenzy took hold once and for all. Allow me to mention here from a more advanced phase of contemporary discourse (to be considered below), Yerulshalmi’s dread confronted with the archival swamp, and Pierre Nora’s exclamation: “Archive as much as you like: something will always be left out.” Once freed of its disgrace and allowed arrogance, has the pharmakon of the archived document become more a poison than a remedy?

Let us follow the historian into the archives. We shall do so in the company of Marc Bloch, who was the historian who best delineated the place of testimony in the construction of the historical fact.31 That history should have recourse to testimony is not fortuitous. It is grounded in the very definition of the object of history. This is not the past, nor is it time, it is “men in time.” Why not time? First of all, because time is the setting, “the very plasma in which events are immersed, and the field within which, they become intelligible” (27–28). (In other words, as was indicated above, time as such constitutes one of the formal conditions of historical reality.) Next, because it returns as a variable among things with regard to its rhythms, as the Braudelian problematic of social times will verify. Moreover, physical nature develops in time, and in this broad sense has a history. Finally, the fascination for origins—that “idol of origins”—has to do with the direct and exclusive thematization of time. This is why the reference to human beings has to figure in the definition. But history is a matter of “men in time,” which implies a fundamental relationship between the present and the past. It is thanks to this dialectic—“understanding the present by the past” and correlatively “understanding the past by the present”—that the category of testimony comes on the scene as the trace of the past in the present. The trace is thus the higher concept under whose aegis Bloch places testimony. It constitutes the operator par excellence of an “indirect” knowledge.

Bloch carries out his examination of the relationship of history to testimony in two chapters. The first is entitled “Historical Observation,” the second “Historical Criticism.”

If we can speak of observation in history, it is because the trace is to historical knowledge what direct or instrumental observation is to the natural sciences. Here testimony figures as the first subcategory. It immediately bears the mark that distinguishes its use in history from its use in ordinary exchanges where orality predominates. It is a written trace, the one the historian encounters in the documents in an archive. Whereas in ordinary exchanges testimony and its reception are contemporary with each other, in history testimony is inscribed in the relation between past and present, in the movement of understanding the one by the other. Writing is thus the mediation of an essentially retrospective science, of a thinking “backwards.”

However, there exist traces that are not “written testimonies” and that are equally open to historical observation, namely, “vestiges of the past” (53), which are the favorite target of archeology: urns, tools, coins, painted or sculpted images, funerary objects, the remains of buildings, and so forth. By extension we can call them “unwritten testimonies,” at the risk of some confusion with oral testimonies whose fate I shall return to below.32 We shall also see testimonies divide into voluntary testimonies, meant for posterity, and those witnesses in spite of themselves, the target of indiscretion and the historian’s appetite.33 This sequence of definitions—science of men in time, knowledge by traces, written and unwritten testimonies, voluntary and involuntary testimonies—assures the status of history as a discipline and of the historian as artisan. Finally, “in the course of its development, historical research has gradually been led to place more and more confidence in the second category of evidence, in the evidence of witnesses in spite of themselves” (61). Indeed, apart from confessions, autobiographies, and other diaries, maps, secret documents, and some confidential reports by military leaders, the documents in archives for the most part come from witnesses in spite of themselves. The variety of different materials to be found in archives is in fact immense. Mastering them calls for an acquired technique, even the practice of specialized auxiliary disciplines and the consultation of different guides in order to assemble the documents necessary to research. The professional historian is someone who keeps in mind the question: “How can I know what I am about to say?” (71).34 This mental disposition defines history as “research,” following the Greek etymology of the word.

At the heart of observation, this relation to “contemporary testimony” (52)—what “others have said” preserved in archives—suffices to draw two dividing lines: the one runs between history and sociology, and the other crosses history itself, which it divides between two opposed methodological attitudes. Sociology, in Durkheim’s sense, as indifferent to time, tends to see in change a residue that it condescendingly leaves to historians. A defense of history in this regard will necessarily be a defense of the event, that privileged object of testimony, as I shall say below. (Pierre Nora’s plea in favor of a “return of the event” will stand in line with the thought of Marc Bloch.) The battle between sociology and history will be harsh and often merciless, even if Bloch can admit having learned from sociologists “to think less shoddily” (15). The second dividing line is the one that opposes a self-professed reconstructive method, owing to its active relation to traces, to one that Bloch condemns as “positivism,” the method of his own teachers, Seignobos and Langlois, whose intellectual laziness he condemns.35

The second section continuing the examination of the relationships of history to written and unwritten testimonies is that of “criticism.” This term specifies history as a science. To be sure, challenges and confrontations between human beings occur outside juridical procedures and those of historical criticism. However, only the testing of written testimony, joined with that of those other traces, the vestiges, has given rise to criticism in a sense worthy of this name. In fact, it is within the historiographical sphere that the very word criticism appeared with the sense of corroboration of what others say, before assuming the transcendental function that Kant would assign to it as critique on the level of an exploration of the limits of our cognitive faculty. Historical criticism has had to blaze a difficult trail between spontaneous credulity and Pyrrhonian skepticism in principle. And it is one that goes beyond mere common sense. We can trace the birth of historical criticism back to Lorenzo Valla’s critique of the Donation of Constantine.36 Its golden age is illustrated by three great names: the Jesuit Paperbroeck, of the Bollandists, the founder of scientific hagiography; Mabillon, the Benedictine from Saint-Maur, who founded diplomatic history; and Richard Simon, the Oratorian who marks the beginnings of critical biblical exegesis. To these three names should be added those of Spinoza with his Theologico-Political Treatise and Bayle, who posed so many doubts. Ought we also to add Descartes? No, if we emphasize the mathematical turn of the Discourse on Method, yes, if we link historians’ doubt with Cartesian methodic doubt.37 The “struggle with the document,” as Marc Bloch so well puts it, is henceforth taken as a given. Its major strategy is to examine sources in order to distinguish the true from the false, and, in order to do this, to “make speak” those witnesses who one knows may deceive themselves or lie, not in order to refute them, but “to understand them” (88).

We owe to this criticism a map or typology of “false testimonies,” to which we might compare the results of Bentham’s Treatise on Judicial Proofs, which Marc Bloch may have known, but which historians have largely improved upon.38

Bloch’s summary is exemplary. Starting from the fact of imposture, as a deliberate fraud, he moves on to the reasons for lying, mystifying, faking, that may be those of wily individuals, self-interested frauds, or those common to an age open to fabrications. Next he considers the more insidious forms of fakery: sly revisions, clever interpolations. There is room for involuntary errors and pathological errors properly speaking arising from the psychology of testimony. (One interesting remark is that the contingencies of events are more propitious for errors than are things drawn from deeply felt feelings about human fate.) Bloch does not hesitate to draw on his own experience as a soldier in the two world wars of the twentieth century in order to compare his experience as a historian, and principally a medieval one, with that of the engaged citizen, attentive to the role of propaganda, censorship, and the pernicious effects of rumor.

To this typology, Bloch grafts his section entitled “Toward a Logic of the Critical Method” (110–37). This opens a vast workspace to which many following him have contributed. At its center lies the work of comparison and its interplay of resemblances and differences. Ordinary controversy finds here an exemplary technical explication. Apart from the elementary prohibition of formal contradiction—an event cannot have happened and not have happened—the argument runs from the art of unmasking the blunders of plagiarists, from spotting what seems obviously unlikely to the logic of probabilities.39 In this regard, Bloch does not make the mistake of confusing the probability of the production of an event—what, in history, would be the equivalent of the initial equality of chances of a toss of dice? In “the criticism of evidence, almost all the dice are loaded” (126)—with the probability of a judgment concerning genuineness made by a reader in the archives. In weighing the pros and cons, doubt is an instrument of knowledge that assesses the degree of likeliness of the chosen combination. Perhaps it would be better to speak of plausibility rather than of probability. That argument is plausible that is worthy of being defended when challenged.

As has been suggested, much remains to be done regarding the validation procedures for any proof and as regards the criteria of internal and external coherence, and many have worked on this problem. Here it seems opportune to bring in a comparison drawn from Carlo Ginzburg concerning the “evidential paradigm.”40 Marc Bloch’s work, in effect, leaves unexplored the notion of a vestige, introduced with regard to archeology and quickly assimilated to the notion of unwritten testimony. But vestiges play a nonnegligible role in the corroboration of testimonies, as police work confirms, but as also does the interpretation of oral or written testimony. Ginzburg speaks here of clues and of the evidential paradigm, courageously opposed to the Galilean paradigm of the natural sciences.

Two questions arise: what are the usages of clues whose convergence authorizes bringing things together in terms of a single paradigm? Furthermore, what in fine is the relation of clues to testimony?

The answer to the first question is constructed by Ginzburg’s text. As a starting point: reference to the clever art lover—the well-known Giovanni Morelli whom Freud draws upon in his “The Moses of Michelangelo”—who made use of the examination of apparently negligible details (the shape of ear lobes) to uncover copies of original paintings. This method of drawing upon clues was the joy of Sherlock Holmes and of every author of detective novels following him. Freud recognized in it one of the sources of psychoanalysis “accustomed to divine secret and concealed things from unconsidered or unnoticed details, from the rubbish heap, as it were, of our observations.” Are not slips of the tongue, when control slips and incongruous signs escape, clues in this sense? Bit by bit, the whole of medical semiotics, with its concept of a symptom, falls under this category of a clue. In the background lies the hunter’s knowledge from earlier days, which deciphers mute tracks. After them come writings, and writing itself, concerning which Ginzburg says that “like divination, it too designated one thing through another” (104). So the whole of semiotics turns out to be based on clues. What then allows this group of disciplines to form a paradigm? Several features: the singularity of the thing deciphered, the indirect aspect of the deciphering, its conjectural character (where the term comes from divination).41 And next comes history. “All this explains why history never became a Galileian science. . . . As with the physician’s, historical knowledge is indirect, presumptive, conjectural” (106). Writing, textuality, which dematerializes orality, changes nothing, for it is once again and always individuals that the historian deals with. Ginzburg links the probabilistic character of historical knowledge to this relation to singularity.

The field opened by the evidential paradigm is immense. “Though reality may seem to be opaque, there are privileged zones—signs, clues—which allow us to penetrate it. This idea, which is the crux of the conjectural or semiotic paradigm, has made progress in the most varied cognitive circles and has deeply influenced the humane sciences” (123).

Now comes the second question: that of the place of Ginzburg’s evidential paradigm in relation to the criticism of testimony of Marc Bloch and his successors. I do not think there is room to choose between these two analyses. By encompassing historical knowledge within the evidential paradigm, Ginsburg weakens his concept of a clue, which gains in being opposed to that of written testimony. Conversely, Bloch’s treatment of vestiges as unwritten testimonies does harm to the specificity of testimony as the intermediary of memory in its declarative phase and narrative expression. The clue is noticed and decrypted; testimony is deposed and criticized. To be sure, it is the same sagacity that presides over both series of operations. But their points of application are distinct. The semiology of clues exercises its role of complement, control, corroboration in regard to oral or written testimony, to the extent that the signs that it decrypts are not verbal: fingerprints, photographic evidence, and today samples of DNA—that biological signature of the living—“testify” through their muteness. Their discourses differ among themselves in different ways than do oral collections.

Thus the benefit of Ginzburg’s contribution is to open a dialectic of clue and testimony internal to the notion of a trace and thereby to give the concept of document its full scope. At the same time, the relation of complementarity between testimony and clue comes to be inscribed in the circle of internal-external coherence that structures documentary proof.

On the one side, in effect, the notion of the trace can be taken to be the common root of testimony and clue. In this regard, its origin in hunting is significant. An animal passed by and left its track. This is a clue. But the clue by extension can be taken as a kind of writing inasmuch as the analogy of the imprint adheres originarily to the evocation of striking a letter, not to speak of the equally primitive analogy to the eikōn, written and painted, referred to at the beginning of our phenomenology of memory. Furthermore, writing is itself written down and in this way a kind of clue. Thus graphology deals with writing, its ductus, its stroke, as a form of clue. Conversely, in this interplay of analogies, clues merit being called unwritten testimonies, in the fashion of Marc Bloch. But these interchanges between clues and testimony must not prevent our preserving their different uses. In sum, the beneficiary of this operation will be the concept of document, made up of clues and testimonies, whose final amplitude rejoins the initial one of the trace.42

There remains the limit case of certain fundamentally oral testimonies, even when written in pain, whose being placed into archives raises a question, to the point of soliciting a veritable crisis concerning testimony. Essentially, it is a question of the testimonies of those who survived the extermination camps of the Shoah, called the Holocaust in English-speaking countries. They were preceded by those of the survivors of the First World War, but they alone have raised the problems I am going to discuss next. Renaud Dulong placed them at the critical point of his work, Le Témoin oculaire: “Bearing Witness from within a Life of Testimony,” this is the label under which he places a work such as Primo Levi’s Drowned and the Saved.43 Why does this genre of testimony seem to be an exception to the historiographical process? Because it poses a problem of reception that being placed in an archive does not answer and for which it even seems inappropriate, even provisionally incongruous. This has to do with such literally extraordinary limit experiences—which make for a difficult pathway in encountering the ordinary, limited capacities for reception of auditors educated on the basis of a shared comprehension. This comprehension is built on the basis of a sense of human resemblance at the level of situations, feelings, thoughts, and actions. But the experience to be transmitted is that of an inhumanity with no common measure with the experience of the average person. It is in this sense that it is a question of limit experiences. And in this way is anticipated a problem that will not find its full expression except at the end of our review of historiographical operations, that of historical representation and its limits.44 But before the limits of explanation and understanding are put to the test, those of inscription and archiving already are. This is why we may speak of a crisis of testimony. To be received, a testimony must be appropriated, that is, divested as much as possible of the absolute foreignness that horror engenders. This drastic condition is not satisfied in the case of survivors’ testimonies.45 A further reason for the difficulty in communicating has to do with the fact that the witness himself had no distance on the events; he was a “participant,” without being the agent, the actor; he was their victim. How “relate one’s own death?” asks Primo Levi. The barrier of shame is one more factor with all the others. The result is that the expected comprehension must itself be a judgment, a judgment on the fly, a judgment without mediation, absolute blame. What finally brings about the crisis in testimony is that its irruption clashes with the conquest made by Lorenzo Valla in The Donation of Constantine. Then it was a matter of struggling against credulity and imposture, now it is one of struggling against incredulity and the will to forget. Is this just a reversal of what is at issue?

Yet even Levi writes. He writes after Robert Antelme, the author of The Human Race,46 after Jean Améry, the author of Par-delà le crime et le châtiment.47 Their writings have even been written about. And I am writing here about stating the impossibility of communicating and about the unbearable imperative to testify to which, however, they do testify. What is more, these direct testimonies find themselves progressively framed, but not absorbed, by the works of historians of the present time and by the publicity of the great criminal trials whose sentences trail slowly through the collective memory at the price of a harsh dissensus. This is why in speaking of these “direct narratives,” unlike R. Dulong I do not talk of an “allergy to historiography” (Le Témoin oculaire, 219). The “allergy to explanation in general” (220), which is certain, provokes instead a kind of short circuit between the moment of testimony, at the threshold of the historiographical operation, and the moment of representation in its written expression, beyond the steps of archiving, of explanation, and even of comprehension. But it is within the same public space as that of historiography that the crisis of testimony after Auschwitz unfolds.


Let us rejoin the historian in the archives. He is their intended receiver inasmuch as the traces were conserved by an institution in view of their being consulted by those trained to do so, following the rules concerning the right of access, any delay in their being consulted varying depending on the category of documents.

At this stage arises the notion of documentary proof, which designates the part of historical truth accessible at this state of the historiographical operation. Two questions: what is proof for a document or a group of documents—and what is proved thereby?

The answer to the first question is tied to the point of articulation of the documentary phase along with the explanatory and comprehending one, and beyond this with the literary phase of representation. If a proof role can be attached to the consulted documents, it is because the historian comes to the archives with questions. The notions of questioning and of a questionnaire are thus the first ones to put in place in elaborating documentary proof. The historian undertakes research in the archives armed with questions. Marc Bloch—in his encounter with the theorists he called positivists, who I prefer to call methodologists, such as Langlois and Seignobos48—again, was one of the first to call for caution about what he took to be epistemological naïveté, namely, the idea that there could be a first phase where the historian gathered up the documents, read them, and weighed their authenticity and veracity, following which there came a second phase where he wrote them up. Antoine Prost, in his Douze Leçons sur l’histoire, following Paul Lacombe, hammers home the strong declaration: no observation without hypotheses, no facts without questions.49 The documents do not speak unless someone asks them to verify, that is, to make true, some hypothesis. Therefore there is an interdependence among facts, documents, and questions. It is the question, writes Prost, “that constructs the historical object through an original carving out from the unlimited universe of possible facts and documents” (79). He thereby rejoins Paul Veyne’s assertion characterizing the current work of historians as “an extending of the questionnaire.” What gives rise to this extension is the formation of hypotheses bearing on the place of the questioned phenomenon within the interconnections putting into play explanation and understanding. The historian’s question, Prost also says, “is not a bare question, it is an armed question that brings with it a certain idea of possible documentary sources and research procedures” (80). Trace, document, and question thus form the tripod base of historical knowledge. This irruption of the question provides an occasion for throwing a final look at the notion of document elaborated above beginning from that of testimony. Taken up by a bundle of questions, the document continues to distance itself from testimony. Nothing as such is a document, even if every residue of the past is potentially a trace. For the historian, the document is not simply given, as the idea of a trace might suggest. It is sought for and found. What is more, it is circumscribed, and in this sense constituted, instituted as document through questioning. For a historian, everything can become a document, including the debris coming from archeological excavations and other such vestiges, but in a more striking way kinds of information as diverse and mercurial as price curves, parish registers, wills, databases of statistics, and so on. Having become a document in this way, everything can be interrogated by a historian with the idea of finding there some information about the past. Among such documents are many that today are no longer testimonies. The series of homogeneous items we shall speak of in the next chapter are not even assignable to what Marc Bloch called witnesses in spite of themselves. The same characterization of the document through interrogation that applies to them holds for a category of unwritten testimonies, those recorded oral testimonies, which microhistory and the history of present times make so much use of. Their role is considerable in the conflict between the memory of survivors and already written history. These oral testimonies do not constitute documents until they are recorded. Then they leave behind the oral sphere to enter into that of writing and distance themselves in this way from the role of testimony in ordinary conversation. We can then say that memory is archived, documented. Its object ceases being a memory, in the literal sense of the word, that is, retained within a relation of continuity and appropriation in regard to some present conscious awareness.

Second question: What at this stage of the historiographical operation can be held to have been proved? The answer is clear: a fact, facts, capable of being asserted in singular, discrete propositions, most often having to do with the mentioning of dates, places, proper names, verbs that name an action or state. Here we need to be alert for one confusion, that between confirmed facts and past events. A vigilant epistemology will guard here against the illusion of believing that what we call a fact coincides with what really happened, or with the living memory of eyewitnesses, as if the facts lay sleeping in the documents until the historians extracted them. This illusion, which Henri Marrou fought against in his The Meaning of History,50 for a long time underlay the conviction that the historical fact does not differ fundamentally from the empirical fact in the experimental natural sciences. Just as, in dealing below with explanation and representation, we shall need to resist the temptation to dissolve the historical fact into narration and this latter into a literary composition indiscernible from fiction, so too we need to resist this initial confusion between a historical fact and a really remembered event. The fact is not the event, itself given to the conscious life of a witness, but the contents of a statement meant to represent it. In this sense, we should always write: the fact that this occurred. So understood, the fact can be said to be constructed through the procedure that disengages it from a series of documents concerning which we may say in return that they establish it. This reciprocity between construction (through a complex documentary procedure) and the establishing of a fact (on the basis of the document) expresses the specific epistemological status of the historical fact. It is this propositional character of the historical fact (in the sense of “fact that . . .”) that governs the mode of truth or falsity attached to the fact. The terms “true” and “false” can legitimately be taken at this level in the Popperian sense of “refutable” and “verifiable.” It is true or it is false that gas chambers were used at Auschwitz to kill so many Jews, Poles, gypsies. The refutation of Holocaust deniers takes place at this level. This is why it was important to correctly delimit this level. In fact, this qualification regarding the truthfulness of “documentary proof” will not reoccur at the levels of explanation and representation, where the Popperian characterization of truthfulness will become more and more difficult to apply.

Some may object to the use that historians make of the notion of an event, either to exile it to the margins in reason of its shortness or fleetingness, or even more because of its privileged tie to the political level of social life, while others may salute its return. Whether it is treated as suspect or as a welcome guest following a long absence, it is as the ultimate referent that the event figures in historical discourse. The question it answers is: What is one talking about when one says that something happened? Not only do I not refuse this referential status, but I will tirelessly plead in its favor throughout this work. And it is to preserve this status of the reference of historical discourse that I distinguish the fact as “something said,” the “what” of historical discourse, from the event as “what one talks about,” the “subject of . . .” that makes up historical discourse. In this regard, that assertion of a historical fact indicates the distance between the said (the thing said) and the intended reference, which according to one of Benveniste’s expressions turns discourse back toward the world. The world, in history, is past human life as it happened. This is what it is all about. And the first thing that one says is that something took place. As stated? That is the whole question. And it will accompany us to the end of the stage of representation, where it will find at least its exact formulation under the heading of “standing for” [représentance], if not its resolution. To get there, we need to leave undetermined the question of the actual relation between fact and event, and tolerate a certain indiscrimination in the employment by the best historians of these terms as standing for each other.51

For my part, I mean to honor the event by taking it as the actual referent of testimony taken as the first category of the archived memory. Whatever specification one may bring or impose subsequently on the event, principally in relation to the notions of structure and conjuncture, placing the event in third place in relation to other conjoined notions, the event in its most primitive sense is that about which someone testifies. It is the emblem of all past things (praeterita). But what is said in spoken testimony is a fact, the fact that . . . Let me be more precise. The “that” affixed to the assertion of a fact holds in reserve the intentional object that will be thematized at the end of our epistemological review under the sign “standing for.” Only a semiotics inappropriate to historical discourse undertakes to deny this referent to the profit of the exclusive pair constituted by the signifier (narrative, rhetorical, imaginative) and the signified (the statement of a fact). To the binary conception of the sign inherited from Saussurean linguistics, and perhaps already mutilated, I oppose the threefold conception of signifier, signified, and referent. Elsewhere I proposed a formula borrowed from Benveniste whereby discourse consists in someone saying something to someone about something following rules.52 In this schema, the referent is symmetrical to the speaker, that is the historian, and before him, to the witness present to his testimony.

I would like to take one last look at the relation between the starting point of this chapter—testimony—and its end—documentary proof—in terms of the mixture of light and shade projected over the whole enterprise by the myth from the Phaedrus speaking of the invention of writing. If the continuity of the passage from memory to history is assured by the notions of trace and testimony, the discontinuity tied to the effects of distantiation that we have put in place ends up at a general crisis situation within which the crisis specifically linked to the untimely testimony of the survivors of the death camps takes its place. This general crisis gives to the question of the pharmakon a precise coloration that haunts this study. What historical criticism puts in question, at the level of documentary proof, is the trustworthiness of spontaneous testimony, that is, the natural movement of having confidence in the heard word, the word of another. A true crisis is thereby opened. A crisis of belief, which authorizes taking historical knowledge for a school of suspicion. It is not just credulity that is here put in the stocks, but the initial trustworthiness of testimony. A crisis of testimony: this is the harsh way documentary history contributes to the healing of memory, of linking the work of remembering to that of mourning. But can we doubt everything? Is it not to the extent that we have confidence in some testimony that we can doubt some other testimony? Is a general crisis of testimony bearable or even thinkable? Can history cut all its lines with declarative memory? The historian no doubt will reply that history, overall, reinforces spontaneous testimony through the criticism of all testimony, that is, through the confrontation between discordant testimonies, in view of establishing a probable, plausible narrative. To be sure, but the question remains: Is documentary proof more remedy than poison for the constitutive weaknesses of testimony? It will be up to explanation and representation to bring some relief to this disarray, through a measured exercise of questioning and a reinforcing of attestation.53

If you find an error or have any questions, please email us at Thank you!