The Historian’s Representation


With the historian’s representation we come to the third phase of the historiographical operation. It would be an error to apply to it the title of writing history or historiography. One constant thesis in this work is that history is writing through and through—from archives to historians’ texts, written, published, given for reading. The seal of writing is thus transferred from the first to the third phase, from an initial to a final inscription. The documents had their reader, the historian in his “mine.” The history book has its readers, potentially anyone who knows how to read, in fact, the educated public. By falling in this way into public space, the history book crowns the “making of histories,” leading its author back into the heart of “making history.” Pulled by the archive out of the world of action, the historian reenters that world by inscribing his work in the world of his readers. In turn, the history book becomes a document, open to the sequence of reinscriptions that submit historical knowledge to an unending process of revisions.

To underscore this phase of the historiographical operation’s dependence on some material support onto which the book is inscribed, we can speak with Michel de Certeau of scriptural representation.1 Or instead, to mark the addition of signs of literariness to the scientific criteria, we can speak of literary representation. Indeed, it is thanks to this final inscription that history indicates that it belongs to the domain of literature. This allegiance was, in fact, implicit already on the documentary plane. It becomes manifest with the coming to be of the history text. We must forget, therefore, that it is not a question of some lowering of standards through which a shift to aesthetics gets substituted for the ambition of epistemological rigor. The three phases of the historiographical operation, let us recall, do not constitute successive phases, but rather intermingled levels where only our didactic concern gives them the appearance of chronological succession.

One final word concerning vocabulary and the semantic choices that govern it. Someone may ask why I do not call this third level “interpretation,” as it might seem legitimate to do. Does not the representation of the past consist in an interpretation of the stated facts? To be sure. But, in an apparent paradox, we do not do justice to the idea of interpretation by assigning it solely to the representative level of the historiographical operation. I shall reserve for the following chapter devoted to truth in history the task of showing that the concept of interpretation has the same amplitude of application as that of truth—more precisely, that it designates a noteworthy dimension of history’s intending truth. In this sense, there is interpretation at all levels of the historiographical operation; for example, at the documentary level with the selection of sources, at the explanation/understanding level with the choice among competing explanatory models, and, in a more spectacular fashion, with variations in scale. But this will not prevent our talking of representation as interpretation when the time comes.

As for my choice of the noun “representation,” it is justified in several ways. First, it indicates the continuity of a single problematic from the explanatory to the scriptural or literary phase. In the preceding chapter we ran into the notion of representation as the privileged object of explanation/understanding, on the plane of the formation of social bonds and the identities at stake in them. We presumed that the way in which social agents understand themselves has an affinity with the way in which historians represent to themselves this connection between their represented object and social action. I even suggested that the dialectic between referring to absence and to visibility, already perceivable in the represented object, is to be deciphered in terms of the operation of representation. In a more radical fashion, the same choice of terminology allows a deep-lying connection to appear, no longer between two phases of the historiographical operation, but on the plane of the relations between history and memory. It is in terms of representation that the phenomenology of memory, following Plato and Aristotle, described the mnemonic phenomenon in that what is remembered is given as an image of what previously was seen, heard, experienced, learned, acquired. Furthermore, it is in terms of representation that what memory intends can be formulated insofar as it is said to be about the past. It is this same problematic of the icon of the past, posed at the beginning of our inquiry, that comes back in force at the end of our discussion. Historical representation follows mnemonic representation in our discourse. This is the most profound reason for choosing the term “representation” to designate the last phase of our epistemological review. Yet this fundamental correlation imposes on our examination one decisive modification in terminology: literary or scriptural representation must in the final analysis allow itself to be understood as “standing for” [représentance], the proposed variation placing the accent not only on the active character of the historical operation, but on the intended something that makes history the learned heir of memory and its foundational aporia. In this way, the fact is forcefully underscored that representation on the historical plane is not confined to conferring some verbal costume on a discourse whose coherence was complete before its entry into literature, but rather that representation constitutes a fully legitimate operation that has the privilege of bringing to light the intended reference of historical discourse.

So we have the target of this chapter. But it will be attained only with the last developments in this chapter. Before that, I shall unfold the specific resources of representation. We shall consider first the narrative forms (“Representation and Narration”).2 I explained above why it may seem as though I have put off any examination of the contribution of narrative to historical discourse. It was because I wanted to move the discussion beyond the impasse into which both the partisans and adversaries of history as narrative have led it. For the former, whom I shall call narrativists, the use of a narrative configuration is an alternative explanatory mode opposed to causal explanation; for the latter, problem-oriented history has replaced narrative history. But for all of them, narrating is equivalent to explaining. By reintroducing narrativity at the third stage of the narrative operation, we remove it not only from an inappropriate demand, but at the same time we free up its representative power.3 We shall not linger over the equation between representation and narration. And the more precisely rhetorical aspect of staging a narrative will be set aside for a distinct discussion (“Representation and Rhetoric”). The selective role for figures of style and thought in the choice of plots, the mobilizing of probable arguments within the frame of the narrative, and the writer’s concern to convince by persuading are the rhetorical resources of staging a narrative. To these solicitations by rhetorical means from the narrator correspond specific postures on the part of the reader in the reception of the text.4

A decisive step in the direction of our projected problematic will be made at the end of this chapter with the question of the relation of historical discourse to fiction (“The Historian’s Representation and the Prestige of the Image”). The confrontation between historical and fictional narrative is well known when it comes to literary forms. What is less well known is the scope of what Louis Marin, a tutelary figure in these pages, calls the “powers of the image,” which outline the contours of an immense realm which is that of the other than the real. How can that absent from present time that is the passed past not be touched by the wing of this angel of absence? Yet was not the difficulty of distinguishing the remembered from an image already the torment of the phenomenology of memory? With this specific problematic of the setting of things said about the past into images comes a distinction unnoticed until now that affects the work of representation, namely, the addition of a concern for visibility in the search for a readability proper to narration. Narrative coherence confers readability; the evoking of the referred to past gives rise to sight. From here on this whole interplay, glimpsed a first time with the represented object, will unfold itself in an explicit manner on the plane of the representation-operation that takes place between the image’s referring back to the absent thing and its self-assertion in terms of its own visibility.

This rapid overview of the major themes of this chapter allows us to see that a double effect is expected from the proposed distinctions. On the one hand, it is a question of a properly analytic undertaking aimed at distinguishing the multiple facets of the idea of historical representation in its scriptural and literary aspects; in this way we shall lay bare and unfold the diverse resources of such representation. On the other hand, it is a question of anticipating at each step along the way the major stake of this chapter, which is to discern historical discourse’s capacity for representing the past, a capacity I have named “standing for.” Under this title we shall find a designating of the very intentionality of historical knowledge that is grafted to that of mnemonic knowledge inasmuch as memory is of the past. The detailed analyses devoted to the relationship between representation and narration, between representation and rhetoric, and between representation and fiction will mark out not only a progression in the recognition of this intentional aim of historical knowledge but also a progression in the resistance to this recognition. For example, representation as narration does not simply turn naïvely toward things that happened. The narrative form as such interposes its complexity and its opacity on what I like to call the referential impulse of the historical narrative. The narrative structure tends to form a circle with itself and to exclude as outside the text, as an illegitimate extralinguistic presupposition, the referential moment of the narration. The same suspicion about the referential irrelevance of representation receives a new form under the headings of tropology and of rhetoric. Do not such figures too form a screen between the discourse and what is claimed to have happened? Do not they entrap the discursive energy within the net of the turns of discourse and thought? And is not this suspicion brought to its peak by the kinship between representation and fiction? It is at this stage that the aporia in which memory seemed to have imprisoned us, insofar as what is remembered is given as a kind of image, an icon, springs up again. How are we to preserve the difference in principle between the image of the absent as unreal and the image of the absent as prior? The entanglement of historical representation with fiction at the end of this sequence repeats the same aporia that seemed to have overwhelmed the phenomenology of memory.

Therefore it is under the sign of a progressive dramatization that the unfolding of this chapter will unfold. Challenges will continue to accompany our attestation of the intentional aim of history; this attestation will bear the indelible imprint of a protest against such suspicion, expressed by a difficulty: “And yet . . .”



My hypothesis governing the following analyses has to do with the place of narrativity in the architecture of historical knowledge. It has two sides. On the one hand, it is taken for granted that narrativity does not constitute an alternative solution to explanation/understanding, despite what the adversaries and advocates of a thesis that, to be brief, I have proposed calling narrativist curiously agree on in saying. On the other, it is affirmed that emplotment nevertheless constitutes a genuine component of the historiographical operation, but on another plane than the one concerned with explanation/understanding, where it does not enter into competition with uses of “because” in the causal or even the teleological sense. In short, it is not a question of downgrading, of relegating narrativity to a lower rank once the operation of narrative configuration enters into play with all the modes of explanation/understanding. In this sense, representation in its narrative aspect, as in the others I shall speak of, does not add something coming from the outside to the documentary and explanatory phases, but rather accompanies and supports them.

Therefore I shall first speak of what we ought not to expect from narrativity: that it fill some lacuna in explanation/understanding. French historians who have expressed their grievances through the opposition between narrative history and problem-oriented history5 and English-speaking authors who have raised the configuring act of emplotment to the rank of an explanation exclusive of causal and even teleological explanations thus in a curious way join us regarding this struggle that I propose to surpass. For in this way an alternative is created that makes narrativity sometimes an obstacle to, sometimes a substitute for explanation.

For Braudel and those close to him in the Annales school, everything turns on the sequence “event, narrative, primacy of the political” when the accent falls on decisions made by leading individuals. To be sure, no one ignores the fact that before becoming an object of historical knowledge, the event is the object of some narrative. In particular, the narratives left by contemporaries occupy a prime place among documentary sources. In this respect, Marc Bloch’s lesson has never been forgotten. The question has rather been to know whether the historical knowledge resulting from the critique of these first-order narratives in its scholarly forms is still clothed with features that belong to the narratives of every kind that nourish the art of narrating. The negative answer can be explained in two ways. On the one hand, by a highly restrictive concept of the event that the narrative is supposed to convey, which has been taken to be a minor, even marginal component of historical knowledge. The case against narrative is thus that against the event. On the other hand, before the development of narratology in the linguistic and semiotic sphere, narrative was taken to be a primitive form of discourse, both tied up with tradition, legend, folklore, and finally myth, and too little elaborated to be worthy of passing the multiple tests that mark the epistemological break between traditional and modern history. In truth, these two orders of consideration go hand in hand—an impoverished concept of event goes along with an impoverished concept of narrative. Hence the trial of the event renders superfluous a distinct trial of narrative. In fact, this trial of event-oriented history has distant antecedents. Krzysztof Pomian, for example, recalls the criticism that Mabillon and Voltaire made of a history that, they said, only taught those events that met the conditions of memory and that prevented any turn to causes and principles, hence to making known the underlying nature of human beings. However, if a worked-out writing of event-oriented history had to wait for the second third of the twentieth century, it was because in the meantime political history had occupied the foreground with its cult of what Croce called “individually determined” facts. Ranke and Michelet remain the unsurpassed masters of this style of history, where the event is held to be singular and unrepeatable. It was this conjunction between the primacy of political history and the prejudice in favor of the unique, unrepeatable event that the Annales school attacked head on. To the characteristic of unrepeatable singularity Braudel was to add the brevity that allowed him to oppose the “long time span” to the history of events. It was this fugacity of the event that, he held, characterizes individual action, principally that of political decision makers, who it had been claimed were those who made events happen. In the final analysis, the two characteristics of the singularity and brevity of the event go together with the major presupposition of so-called event-oriented history, namely, that the individual is the ultimate bearer of historical change. As for narrative history, it was taken as a mere synonym for such history. In this way, the narrative status of history did not become the object of a distinct discussion. As for the rejection of the event in the point-like sense, it is the direct consequence of the displacing of the principal axis of historical investigation from political history toward social history. In a word, it is in political, military, diplomatic, ecclesiastical history that individuals—heads of state, generals, ministers, prelates—are supposed to make history. There too reigns the event that can be assimilated to an explosion. The denunciation of the history of battles and the history of events thus constitutes the polemical underside of a plea for a history of the total human phenomenon, with however a strong accent on economic and social conditions. It was within this critical context that was born the concept of the long time span opposed to that of the event, understood in the sense of a brief duration, which I have dealt with above. The dominant intuition, we have said, was that of a fierce opposition at the heart of social reality between the instant and “time that takes a long time to unfold.” Pushing this axiom to the point of paradox, Braudel went so far as to say that “social science has something like a horror of the event.” This head-on attack against the sequence of “event, narrative, primacy of the political” found strong reinforcement with the massive introduction into history of quantitative procedures borrowed from economics and extended to demographic, social, cultural, and even spiritual history. With this development, a major presupposition concerning the nature of the historical event was called into question, namely, that as unique no event repeats itself. Quantitative history, in effect, is fundamentally a “serial history.”6

If, according to the Annales perspective, narrative as a collection of point-like events and the traditional form of cultural transmission is an obstacle to problem-oriented history, according to the narrativist school across the Atlantic, it is worthy to enter into competition with the models of explanation the human sciences share in common with the natural sciences. Far from being an obstacle to the scientific nature of history, narrative becomes its substitute. It was as confronted by an extreme demand represented by the nomological model of historical knowledge that this school of thought undertook to reevaluate narrative’s resources of intelligibility.7 This effort owed little to narratology and its claim to reconstruct the surface effects of narrative on the basis of its deep structures. The work of the narrativist school took place instead in the wake of inquiries devoted to ordinary language, and to its grammar and its logic as these function in natural languages. In this way the configuring character of narrative was brought to the fore at the expense of its episodic character, which was the only thing the Annales historians took into account. In relation to the conflict between explanation and understanding, narrativist interpretations tend to deny the relevance of this distinction insofar as to understand a narrative is thereby to explain the events that it integrates and the facts it reports. The question thus will be to know at what point the narrativist interpretation accounts for the epistemological break that has occurred between told stories and history built on documentary traces.

In Time and Narrative I summarized the successive theses of the narrativist school.8 A special place had to be given to the work of Louis O. Mink, which for a long time remained scattered in different articles before being gathered together in a posthumous book titled Historical Understanding.9 This title, which is a good statement of the central thesis of Mink’s different essays, must not mislead us. In no way is it a question of opposing explanation and understanding as in Dilthey. On the contrary, it is a question of characterizing historical explanation, as a “gathering together,” through a configuring, synoptic, synthetic act, endowed with the same sort of intelligibility as is judgment in Kant’s Critique of the Faculty of Judgment. Therefore it is not the intersubjective features of Verstehen that are emphasized here, but the function of “colligation” exercised by the narrative taken as a whole in relation to the reported events. The idea that the form of the narrative as such should be a “cognitive instrument” follows at the end of a series of increasingly precise approaches—at the price of the discovery of aporias concerning historical knowledge, aporias that only such narrativist interpretation allows to be discovered. With the passage of time, we can today praise Mink for the rigor and honesty with which he summed up these aporias. The problem is posed that will be the torment of any literary philosophy of history: what difference separates history from fiction, if both narrate? The classic answer that history alone retraces what actually happened does not seem to be contained in the idea that the narrative form has within itself a cognitive function. This aporia, which we can call that of the truth in history, becomes apparent through the fact that historians frequently construct different and opposed narratives about the same events. Should we say that some omit events and considerations that others focus on and vice versa? The aporia would be warded off if we could add rival versions to one another, allowing for submitting the proposed narratives to the appropriate corrections. Shall we say that it is life, presumed to have the form of a history, that confers the force of truth on this narrative? But life is not a history and only wears this form insofar as we confer it upon it. How, then, can we still claim that we found this form in life, our own life and by extension that of others, of institutions, groups, societies, nations? This claim is solidly entrenched in the very project of writing history. The result is that it is no longer possible to take refuge in the idea of “universal history as lived.” Indeed, what relationship could exist between this presumed unique and determined kingdom of universal history as lived and the histories we construct, when each one has its own beginning, middle, and end, and draws its intelligibility from its internal structure alone? This dilemma strikes narrative not only at its configuring level, but strikes the very notion of an event. Beyond the fact that we can question the rules governing the use of the term (was the Renaissance an event?), we can also ask if there is any meaning in saying that two historians give different narratives of the same events. If the event is a fragment of a narrative, it depends on the outcome of the narrative, and there is no underlying, basic event that escapes narrativization. Yet we cannot do without the notion of the “same event,” if we are to be able to compare two narratives dealing, as we say, with the same subject. But what is an event purged of every narrative connection? Must we identify it with an occurrence in the physical sense of this term? But then a new abyss opens between event and narrative, comparable to the one that isolates written history from history as it actually happened. If Mink undertook to preserve the commonsense belief that history is distinguished from fiction by its truth claim, this was, it seems, because he did not give up the idea of historical knowledge. In this regard, the last essay he published, “Narrative Form as a Cognitive Instrument,”10 sums up the state of perplexity in which he found himself when his work was cut off by his death. Dealing for a last time with the difference between history and fiction, he limits himself to taking as disastrous the possibility that common sense should be dislodged from its entrenched position. If the contrast between history and fiction were to disappear, both would lose their specific mark, namely, the claim to truth on the side of history and the “voluntary suspension of disbelief” on that of fiction. But Mink does not say how the distinction is to be preserved. In renouncing any resolution to the dilemma, he preferred to hold on to it as belonging to the historical enterprise per se.

Rather than playing off the adversaries and the partisans of the explanatory relevance of narrative as a configuring act against each other, it seemed to me more useful to ask about the way in which two types of intelligibility could go together, narrative intelligibility and explanatory intelligibility.11

As regards narrative intelligibility, it would be necessary to bring together the still too intuitive considerations of the narrative school and the more analytic work of narratology on the plane of the semiotics of discourse. The result is a complex notion of “narrative coherence” that must be distinguished, on the one hand, from what Dilthey called the “cohesion of a life,” in which we can recognize prenarrative features, and on the other hand, from the notion of a “causal or teleological connection (or connectedness),” arising from explanation/understanding. Narrative coherence is rooted in the former and articulated through the latter. What it itself brings is what I have called a synthesis of the heterogeneous, in order to speak of the coordination between multiple events, or between causes, intentions, and also accidents within a single meaningful unity. The plot is the literary form of this coordination. It consists in guiding a complex action from an initial situation to a terminal one by means of rule-governed transformations that lend themselves to an appropriate formulation within the framework of narratology. A logical content can be assigned to these transformations, what Aristotle characterized in his Poetics as the probable or the reasonable, the reasonable constituting the face that the probable turns toward the readers in order to persuade them; that is, to induce them to believe precisely in the narrative coherence of the told story or history.12

I want to draw on two implications of this concept of narrative coherence.

First, a properly narrative definition of the event, which subsequently will have to be connected to the definitions given it on the plane of explanation. On the narrative plane, the event is what, in happening, advances the action—it is a variable of the plot. Events that give rise to an unexpected turn are said to be sudden—“against expectation” (para doxan), says Aristotle, thinking of “theatrical effects” (peripeteiai) and “violent effects” (pathē).13 In a general way, any discordance entering into competition with concordance counts as an event. This conjunction of plot and event is open to noteworthy transpositions on the historiographical plane, which go far beyond what is called event-oriented history, which only retains one of the possibilities of the narrative event, namely, brevity combined with suddenness. There are, if we may put it this way, long time span events, indicated by the amplitude, the scope of the history recounted—the Renaissance, the Reformation, the French Revolution make up such events in relation to a period that can extend over several hundred years.

Second implication: inasmuch as the actors in a narrative—the characters—are emplotted along with the story, the notion of narrative identification, correlative to that of narrative coherence, too is open to noteworthy transpositions on the historical plane. The notion of a character constitutes a narrative operator of the same amplitude as that of an event. The characters perform and suffer from the action recounted. For example, the Mediterranean in Braudel’s great work can be taken as a quasi character of the quasi plot of the rise to power and the decline of what was “our sea” in the age of Philip II. In this regard, the death of Philip was not an event on the same scale as the plot about the Mediterranean.14

A third implication suggested by Aristotle’s Poetics would have to do with the moral evaluation of the characters, better than us in tragedy, lower or equal to us in virtue in comedy. I shall reserve this discussion for the following chapter in terms of a broader framework having to do with the relationship between the historian and the judge. We shall not, however, avoid some anticipation of this discussion when in speaking about the rhetorical categories applied to plots we shall be confronted with the question of the limits imposed on representation by events taken to be horrible, even morally unacceptable.15

I want now to propose two examples of the connecting of “narrative coherence” to “causal or teleological explanation,” corresponding to the two types of intelligibility referred to above. The solution of Mink’s dilemma and more generally of the aporia whose progression we shall follow in the remainder of this chapter hangs on the plausibility of this analysis. It would be futile to seek a direct tie between the narrative form and the events as they actually occurred; the tie can only be indirect by way of explanation and, short of this, by way of the documentary phase, which refers back in turn to testimony and the trust placed in the word of another.

My first example is suggested by the use made in the preceding chapter of the notion of an interplay of scales. Among all the kinds of syntheses of the heterogeneous that constitute emplotment, must we not take into account the narrativized course of changing scales? In fact, neither microhistory nor macrohistory work continually with a single scale. To be sure, microhistory privileges the level on interactions on the scale of a village, of a group of individuals or families. This is the level where the negotiations and conflicts unfold and the situation of uncertainly reveals itself that history makes evident. But this history does not fail also to read from bottom to top the power relations that play themselves out on another scale. The discussion about the exemplariness of these local histories played out on the lowest level presupposes the interweaving of a small-scale history with a larger-scale one. In this sense, microhistory situates itself on a sequence of changing scales that it narrativizes as it goes. The same can be said about macrohistory. In some forms, it situates itself at one determined level and does not quit it. This is the case with those operations of periodization that scan the time of history in terms of long sequences indicated by large-scale narratives. One important narrative concept proposes itself here that we have already encountered above, that of “scope,” which F. R. Ankersmit has elaborated in the context of a narrative logic whose implications concerning the relation between representation and standing for we shall discuss further below.16 The scope of an event speaks of the persistence of its effects far beyond its origin. It is correlative to the scope of the narrative itself, whose meaningful unity perdures. If we stop at this homogeneous level of a “period,” some important aspects of narrativization can be noted, among them the personalization indicated by the use of proper names (or of quasi proper names): Renaissance, French Revolution, Cold War, and so forth. The relationship of these proper names to the descriptions that in a way constitute their predicates poses the problem of a narrative logic appropriate to these strange high-level singularities, to which Ankersmit gives the name narratio. However all the narrative resources of macrohistory do not allow themselves to be reduced to effects on a single level. As Norbert Elias’s work illustrates, the effects of a system of power, such as that of the monarchical court, unfold along a descending scale down to the forms of behavior of self-control at the level of the individual psyche. In this regard, the concept of habitus can be taken as a concept of narrative transition operating all along this descending way from the higher plane of the production of meaning to the lower one of its concrete actualization, thanks to the forgetting of the cause concealed among its effects.

The second example has to do with the notion of an event. Above we recalled the narrative function as an operator on the plane of the recounted action. However, among all the attempts to define the event on the plane of explanation, we placed the accent on the one that coordinates the event with the structure and with the conjuncture, associating it with the two ideas of deviation and difference. But is it not possible to cross the logical abyss that seems to lie between these two definitions of the event? One hypothesis proposes itself: if we give its full extension to the idea of the plot as the synthesis of the heterogeneous embracing intentions, causes, and accidents, is it not up to the narrative to bring about a kind of narrative integration of the three moments—structure, conjuncture, event—that epistemology dissociates? The just-proposed idea of a narrativization of the interplay of scales suggests this, inasmuch as the three moments refer to different scales having to do as much with the level of efficacy as with that of temporal rhythms. I have found helpful support for working out this hypothesis in the work of Reinhart Koselleck in an essay titled “Representation, Event, and Structure” in his Futures Past.17 Having stated that, as separately identifiable temporal strata, structures stem in fact from description and events from the narrative, he suggests that the dynamic that interweaves them lends itself to a narrativization that makes narrative a switching point between structure and event. This integrative function of the narrative form results from the distance it takes in regard to mere chronological succession in terms of before and after, of the type venividivici. As a meaningful unity, the plot is capable of articulating structures and events within one and the same configuration. Thus, to cite an example, the evocation of a structure of domination can be incorporated into the narrative of an event such as a battle. The structure as a phenomenon of the long time span through the narrative becomes the condition of possibility of the event. We can speak here of structures in eventu whose significance is only grasped post eventum. The description of structures in the course of the narrative contributes in this way to clarifying and elucidating the events as causes independent of their chronology. What is more, the relation is reversible. Some events are taken as significant insofar that they serve as indices of social phenomena of the long time span and even seem determined by them—a trial about the right to work can illustrate in dramatic fashion social, juridical, or economic phenomena of the long time span.18 The narrative integration between structure and event thus fits with the narrative integration between phenomena situated at the different levels of the scales of endurance and efficacy. The distinction between description and narration is not thereby effaced, but if description preserves the stratification of layers, it is up to the narrative to tie them together. The cognitive relation between the two concepts is of the order of a distinction—it finds a didactic complement in their referring to each other thanks to the narrative configuration. Here the relation between structure and event is like that between the layers of time spans. Every stratification can be narratively mediated in this way.19

These two examples of the narrativization of explanatory modes at work in the historiographical operation contain two lessons. On the one hand, they show how the written forms of this operation get articulated in terms of the explanatory forms. On the other hand, they show how the intentional aim of narrative beyond its closure runs across such explanation in the direction of the reality attested to. The resistances to this passage have still to be considered.

However I do not want to leave behind the question of narrativity and its contribution to the third phase of the historiographical operation without having dealt with certain aspects of emplotment that, when joined to similar effects of other moments of the written expression of history, render the solution of the problem posed by the historical narrative’s claim to represent the past paradoxically more difficult. Along the way from representation to re-presentation, narrative has to deal with obstacles having to do precisely with the structure of the act of configuration.

The challenge in the name of a disjunction between the internal structure of a text and any extratextual reality comes from literary theory. To the extent that the fictional and the historical narrative participate in the same narrative structures, the rejection of a referential dimension by structuralist orthodoxy extends to the whole of literary textuality. This rejection is motivated by an expansion of the Saussurean model from the plane of isolated signs—like those collected into systems of a lexical type—to that of sentences and finally to that of longer textual sequences. According to this model, the relation between signifier and signified gives rise to an entity with two faces, the sign properly speaking, whose apprehension can make an exception of any relation to a referent. This exception is the work of the theoretical gaze that sets up the sign as the homogeneous theme of linguistic science. It is this bipolar model of signifier and signified, excluding any referent, that spread into every region of language accessible to a semiotic treatment. And it was in this way that a Saussurean type of narratology could apply to long textual sequences the bracketing of the referent required by the model. While the effects for fictional narrative seem discussible without being disastrous (something I discussed in The Rule of Metaphor), they have been devastating for historical narrative, whose difference from fictional narrative rests on the referential intention that runs through it and that is nothing other than the meaning of its representation. Hence I tried at the time to reconquer the referential dimension by starting from the level of the sentence as the initial unit of discourse, following the analyses of Émile Benveniste and Roman Jakobson. With the sentence, I said, someone says something to someone about something on the basis of a hierarchy of codes: phonological, lexical, syntactical, and stylistic. Saying something about something seemed to me to constitute the virtue of discourse and by extension of the text as a enchaining of sentences.20 However the problem of the referentiality of historical discourse seemed to me to need to be posed in a distinct way inasmuch as the tendency toward closure inherent in the act of emplotment is an obstacle to the extralinguistic, extratextual impulse of all referential speech through which representation becomes re-presenting.21 But, before we come to the attestation-protestation that constitutes the soul of what I call the standing for of the past,22 it will first be necessary to push as far as possible our examination of the other components of the literary phase of the historiographical operation. These add their own denying of the referential impulse of historical discourse to that emanating from the narrative configuration per se.23


It is worth the effort of granting distinct attention to the properly rhetorical dimension of the discourse of history, despite entanglement of the figures of this domain with the structures of narrative. Here we touch on a tradition that goes back to Vico and its double heritage: the plane of the description of figures of thought and discourse, called tropes—principally metaphor, metonymy, synecdoche, and irony—and that of the plea in favor of the modes of argumentation that rhetoric opposes to the hegemonic claims of logic.

The stakes in this new step of our investigation do not consist solely in broadening the field of procedures of scriptural representation, but also take into account the resistances that narrative and rhetorical configurations oppose to the referential impulse that turns historical narrative toward the past. Perhaps we shall also come upon the outlines of a counteroffensive by a certain critical realism in regard to the aestheticizing temptation that the advocates of narrative rhetoric risk surrendering to. This is what happened when the protagonists in this discussion found themselves confronted in the last decades of the twentieth century with the problem of the figuration of events that, because of their monstrosity, pushed the “limits of representation.” A part of this discussion took place in France, but it was in America that it became the occasion for the confrontation I shall allude to.

The French contribution to the debate dates from the golden age of structuralism. The methodological revolution that the French school following Ferdinand de Saussure claimed to follow brought to light aspects of the narrative code that present a close kinship with the general structural properties of the system of language, distinguished from its use in speech. The basic postulate is that the structures of narrative are homologous with those of the elementary units of language.24 The result is an extending of linguistics to the semiotics of narrative. The principal effect on the theory of narrative was an exclusion of every consideration borrowed from the literary history of a genre, uprooting the achrony of structures from the diachrony of the practice of discourse to the benefit of a logicization and a dechronologization whose steps I traced in the volume 2 of Time and Narrative.25 The implications for the historical field might never have seen the light of day in that the semiotics of narrative were applied to fiction in the wake of Vladimir Propp. All that one might deplore was the loss of the sense of the marvelous, but this was not a negligible result once the kinship of this emotion a contrario to the more frightful one that the history of the twentieth century was to unleash was taken into account. A threat directed against the referential claim of history was, however, already contained in the choice of the Saussurean model on the plane of general semiotics. Above, I referred to the consequences for the treatment of historical discourse of the excluding of the referent required by the binary constitution of the sign as signifier and signified. But for structuralism really to hit history hard it was necessary that the concern for what we can call the scientific aim of its advocates should be joined to a more polemical and ideological concern directed against the presumed humanism of a whole set of representative practices. Historical narrative found itself in the same dock with the realist novel inherited from nineteenth-century European history. Suspicion then was intermingled with curiosity, narrative history being accused particularly of producing subjects adapted to a system of power that gave them a sense of mastery over themselves, nature, and history.26 For Roland Barthes, the “discourse of history” constitutes the privileged target of this genre of criticism based on suspicion. Taking his stand on the exclusion of the referent in the linguistic field, he held against historical narrative its placing a referential illusion at the very heart of historiography. This illusion consisted in the claim that the supposed external, founding referent—that is, the time of the res gestae—was hypostatized at the expense of the signified, that is, the meaning that the historian gives to the reported facts. In this way a short-circuit was produced between the referent and the signifier, and “the discourse, meant only to express the real, believes it elides the fundamental term of imaginary structures, which is the signified” (138–39). This fusing of the referent and the signified to the benefit of the referent engenders a “reality effect” in virtue of which the referent, surreptitiously transformed into a disgraceful signifier, is clothed with the privileges of “that’s what happened.” History thus gives the illusion of finding the real that it represents. In reality, its discourse is only “a fake performative discourse in which the apparent constative (descriptive) is in fact only the signifier of the speech act as an act of authority” (139). At the end of his article, Barthes can applaud the decline of narrative history and the rise of structural history. This, in his eyes, is a veritable ideological revolution more than a change of schools: “Historical narration is dying because the sign of History is henceforth not so much the real as the intelligible” (140). It remained to spell out the mechanism of this eviction of the signified, expelled by the presumed referent. This is what a second essay, titled precisely “The Reality Effect,” undertakes to do. The key to the riddle is sought on the side of the role exercised by the notations in the realist novel and in history from the same period, that is, those “superfluous” details that contribute nothing to the structure of the narrative, to its thread of meaning. These are “insignificant intervals” in relation to the meaning imposed over the course of the narrative. To account for the reality effect, we have to begin from this insignificance. Prior to the realist novel, such notations contributed to a verisimilitude of a purely aesthetic and in no way referential character. The referential illusion consists in transforming the notation’s “resistance to meaning” into a “supposed real.” In this, there is a break between the older verisimilitude and modern realism. But, also in this, a new verisimilitude is born that is precisely this realism, by which we can understand “discourse which accepts ‘speech-acts’ justified by their referent alone” (147). This is in fact what happens in history where “‘concrete reality’ becomes the sufficient justification for speaking” (146). This argument boils down to transferring one noteworthy feature of the nineteenth-century realist novel to all historical narrative.

Here is the place to ask whether such suspicion is not wholly forged on the basis of a linguistic model that is inappropriate to historical discourse, which would be better understood in terms of alternative models for which the referent, whatever it might be, constitutes a irreducible dimension of discourse addressed by someone to someone about something. It would remain to give an account of the specificity of referentiality in the historiographical domain. My thesis is that this cannot be discerned solely on the plane of the functioning that historical discourse assumes, but that it must pass through the documentary proof, the causal and teleological explanation, and the literary emplotment. This threefold frame remains the secret of historical knowledge.27

The major contribution to the exploring of the properly rhetorical resources of historical representation remains that of Hayden White.28 It is valuable as much for the criticism it has elicited as for the relevance of this thinker’s analyses meant to expand his readers’ awareness. The discussion he stirred up about literature about the Shoah has given his propositions a dramatic dimension unattained by the French structuralists’ theses. This had to do not with a contribution to the epistemology of historical knowledge, but with a poetics that takes the imagination, more precisely, the historical imagination, as its theme. It is in this sense that it shows itself faithful to the spirit of the times and to what has been called the “linguistic turn,” inasmuch as it is through the structures of discourse that this imagination is apprehended. Therefore it is verbal artifacts that will be at issue. This detail takes nothing away from the scope of what is intended. In effect, two roadblocks get lifted. The first one has to do with the relation of history to fiction. Seen from the angle of language, historical and fictional narrative both belong to a single class, that of “verbal fictions.” All the problems tied to the referential dimension of historical discourse will be taken up starting from this classification. The second roadblock has to do with the distinction between professional historiography and the philosophy of history, at least that part of the philosophy of history that clothes itself in the form of large world-scale narratives. In this way, Michelet, Ranke, Tocqueville, Burckhardt, Hegel, Marx, Nietzsche, and Croce are set within the same framework. What is common to them is bringing the historical imagination to discourse through a form that comes from rhetoric and more precisely from the rhetoric of tropes. This verbal form of the historical imagination is the emplotment.

In Metahistory the scope of White’s gaze is made manifest in that the operation of emplotment is grasped through considering a sequence of typologies that give the enterprise the allure of a well-ordered taxonomy. But we must never lose sight of the fact that this taxonomy operates at the level of the deep structures of the imagination. This opposition between deep and surface structures was not unknown to semioticians, or to psychoanalysts. In the specific situation of verbal fictions, it allows a hierarchy of typologies rather than simply piling them on top of one another or juxtaposing them. The four typologies we are about to consider and the compositions that result from their being associated with one another must therefore be taken for matrices of possible combinations on the plane of the actual historical imagination.

White’s carrying through of this program is methodical. The major typology, which places White in Vico’s wake, the typology of plot types, crowns a hierarchy of three typologies. The first one stems from aesthetic perception. It is the dimension of the plot’s “story.” In a way somewhat like that of Mink, the organization of the story as told exceeds the simple chronology that still prevailed in chronicles and adds to the “storyline” an organization in terms of motifs that we may call inaugural, transitional, or terminal. What is important is that, as with the defenders of narrativism discussed above, the story has an “explanatory effect” in virtue of its structural apparatus alone. Here rhetoric first enters into competition with the epistemology of historical knowledge. The seriousness of the conflict is increased by two considerations. Concerning the form, as White’s recent work emphasizes, we have to say that the plot tends to make the contours of the story prevail over the distinct meanings of the events recounted, inasmuch as the accent is placed on the identifying of the configuring class in which a plot is inscribed. As for what is supposed to precede the setting into form, the rhetorician finds nothing prior to the first sketches of narrativization, other than perhaps an unorganized background, an “unprocessed historical record.” The question of the status of the factual data in relation to the initial mise en forme of the story told is left open.

The second typology has more to do with the cognitive aspects of the narrative. But, as with rhetoricians, the notion of argument is taken in terms of its persuasive capacity rather than as a purely logical demonstration.29 That there is a manner of arguing proper to narrative and historical discourse, and that this lends itself to a specific typology, constitutes an original idea, whatever may have been borrowed from disciplines other than history concerning the distinction between formalist, organic, mechanical, and contextual arguments.30

The third typology, that of ideological implications, stems rather from modes of moral and political commitment and therefore from the insertion in present practice. In this sense, it arises from what Bernard Lepetit calls the present of history. Below we shall return to the problem posed by this typology, in terms of the implication of the protagonists in events that cannot be separated from their moral charge.

Next comes the emplotment, which White takes to be the explanatory mode par excellence. He borrows his typology of four terms—romance, tragedy, comedy, satire—from Northrop Frye’s Anatomy of Criticism, in this way linking up with Vico’s rhetoric.

If we had to characterize White’s enterprise in a single word, we would have to speak as he himself does of a theory of style. Each combination among elements belonging to one or another typology defines the style of a work that one can then characterize by the dominant category.31

It is not a question of denying the importance of White’s pioneering work. With Roger Chartier, we may even regret White’s “missed encounter” with Paul Veyne and Michel Foucault, his contemporaries during the 1970s. But the idea of a deep structure of the imagination owes its indisputable fruitfulness to the tie it establishes between creativity and codification. This dynamic structuralism is perfectly plausible. Separated from the imaginary, these paradigms would only be the inert classes of a more or less refined taxonomy. Instead they are generating matrices meant to engender an unlimited number of manifest structures. In this regard, the criticism that White did not choose between determinism and free choice seems to me easily refuted. It belongs precisely to these formal matrices to open a limited space for choice. We can speak in this way of a rule-governed production, a notion that recalls the Kantian concept of the schematism, that “method for producing images.” The consequence is that the alternative objections of taxonomic rigidity and restless wandering within the space of imaginative variations miss the originality of the project, whatever hesitations or weaknesses characterize its execution. The idea that White appears to have drawn back in panic before an unlimited disorder seems to me not only inadequate but unfair, in light of the personal character of the attack that it assumes.32 The overly dramatic expression “the bedrock of order” ought not to turn our attention from the relevance of the problem posed by the idea of an encoding that functions both as a constraint and a space for invention. A place is made in this way for exploring the mediations proposed by stylistic practice over the course of the history of literary traditions. This connecting of formalism and historicity remains to be done. It belongs to a system of rules, both found and invented, to present original features of traditionality that transcend this alternative. This is what is tied up in what we call style. On the other hand, I regret the impasse that White gets caught up in in dealing with the operations of emplotment as explanatory models, held to be at best indifferent as regards the scientific procedures of historical knowledge, at worst as substitutable for them. There is a true category mistake here that engenders a legitimate suspicion regarding the capacity of this rhetorical theory to draw a clear line between historical and fictional narrative. While it is legitimate to treat the deep structures of the imaginary as common generating matrices for the creation of the plots of novels and those of historians, as is attested to by their interweaving in the history of genres during the nineteenth century, it also becomes urgent to specify the referential moment that distinguishes history from fiction. And this discrimination cannot be carried out if we remain within the confines of literary forms. Nothing is gained by outlining a desperate escape by way of a simple-minded recourse to common sense and the most traditional assertions concerning truth in history. What is required is patient articulating of the modes of representation in terms of those of explanation/understanding and, through these, of the documentary moment and its generating matrix of presumed truth—that is, the testimony of those who declare their having been there where things happened. We shall never find in the narrative form per se the reason for this quest for referentiality. The work of reconstructing historical discourse taken in terms of the complexity of its operative phases is totally absent from Hayden White’s preoccupations.

It is with regard to these aporias of the referentiality of historical discourse that the calling into question of the propositions of White’s narrative rhetoric by the horrible events placed under the sign of the “final solution” constitutes an exemplary challenge that goes beyond any textbook exercise.

This challenge has found strong expression in the notion of a “limit to representation” such as gives its title to the volume edited by Saul Friedlander, Probing the Limits of Representation.33 This phrase can designate two kinds of limits: on the one hand, a kind of exhaustion in our culture of the available forms of representation for giving readability and visibility to the event known as the “final solution”; on the other hand, a request, a demand to be spoken of, represented, arising from the very heart of the event, hence proceeding from that origin of discourse that one rhetorical tradition takes to be extralinguistic, and forbidden any sojourn in the land of semiotics. In the first case, it would be a question of an internal limit, in the second, of an external one. The problem therefore will be that of the precise articulation between these two limits. The Shoah, which is how we really should name it, at this stage of our discussion proposes for reflection both the singularity of a phenomenon at the limit of experience and discourse, and the exemplarity of a situation where not only the limits of representation in its narrative and rhetorical forms, but the whole enterprise of writing history, are open to discovery.

White’s tropology could not fail to get caught up in this.34 And in Germany itself a vast quarrel between 1986 and 1988 known as the Historikerstreit involved respected historians of the Nazi period as well as a philosopher as well known as Jürgen Habermas, over such problems as the uniqueness of Nazism and the relevance of any comparison with Stalinism, with the consistency of Hannah Arendt’s concept of totalitarianism as one of the stakes, and finally over the question of the continuity of the German nation with and beyond this catastrophe.35

It was against this charged background of questions and passions concerning the very possibility of “historicizing” (Historiserung) National Socialism and singularly “Auschwitz” that the American colloquium on the theme “History, Event, and Discourse” took place, during which Hayden White and Carlo Ginzburg set forth their opposing views about the notion of historical truth. In this way the question of the limits of representation in narrative and rhetorical form could take on the proportions of a test—a probing—of the limits of the very project of representing to oneself an event of such magnitude. Historicization and figuration—the same struggle and the same test.

In his introduction to Probing the Limits, Saul Friedlander proposes a schema whereby it is necessary to begin from the external limits of discourse in order to form the idea of the internal limits of representation. In this, he deliberately steps out of the circle that representation forms with itself. At the heart of Europe occurred an “event at the limits” (3). This event reached down to the deepest layers of solidarity among human beings: “Auschwitz has changed the basis for the continuation, the basis for the continuity of the conditions of life within history” (3). Life-within-history, not “discourse-about-history.” A truth claim arises from behind the mirror that places its exigencies on representation, which reveal the internal limits of literary genres: “There are limits to representation which should not be but can easily be transgressed” (3, his emphasis). There can be something wrong with certain representations of events (above all when the transgression is as glaring as Holocaust denial), even if we cannot formulate the nature of the transgression and are condemned to remain in a state of uneasiness. The idea of transgression in this way confers an unexpected intensity on a discussion that began on an inoffensive, if not innocent level, that of semiotics, of narratology, of tropology. The event “at the limits” brings its own opacity along with its morally “unacceptable” character (the word takes on the force of an extreme understatement)—its character of “moral offense.” Then it is the opacity of events that reveals and denounces that of language. What is more, this denunciation takes on an unexpected character at one moment of the theoretical discussion marked by what we call by convention “postmodernism,” a moment where the critique of naïve realism is at its apogee in the name of the polysemy en abîme of discourse, of the self-referentiality of linguistic constructions, which make impossible the identification of any stable reality whatsoever. What plausible response, then, can this so-called postmodernism give to the accusation of having disarmed thought in the face of the seductions of negationism?36

Confronted with Friedlander’s scheme, which proceeds from the event at its limits in the direction of the internal limits of the operation of representation, Haydon White undertakes, with an extreme honesty, to go as far as possible in the direction of the event by speaking of the rhetorical resources of verbal representation.37 But can a tropology of historical discourse link up with something like a “demand,” in the strong sense of the word in English, a truth claim, proceeding from the events themselves?

White’s essay exhibits a kind of quartering of its own discourse. On the one hand, the author increases his claim for the “inexpugnable relativity” of every representation of historical phenomena. This relativity has to be assigned to language itself, insofar as it does not constitute a transparent medium, like a mirror reflecting some presumed reality. The pair plot/trope is once again taken as the site of resistance to any return to a naïve realism. On the other hand, a suspicion grows over the course of his essay that there could be something in the event itself so monstrous as to put to flight all the modes available to representation. This something would have no name in any known class of plot types, be they tragic, comic, or whatever. Following the first direction of his thesis, White points to the roadblocks on the path to the event. It is impossible, he declares, to distinguish between a “factual statement” (singular existential propositions and arguments), on the one hand, and narrative reports, on the other. These latter will always transform lists of facts into stories. But these stories bring with them their plot types, tropes, and typologies. All we are left with is “competitive narratives,” which no argument can decide among and for which no criterion drawn from factual statements can arbitrate, once the facts are already facts of language. The principle of a distinction between interpretation and fact is thereby undermined, and the boundary between “true” and “false,” “imaginary” and “factual,” “figurative” and “literal” story falls. Applied to the events designated by the expression “final solution,” these considerations lead to the impossibility of making sense on the narrative plane of the idea of an unacceptable mode of emplotment. None of the known modes of emplotment is unacceptable a priori; nor is any one more appropriate than another.38 The distinction between acceptable and unacceptable does not stem from tropology, but proceeds from another region of our receptive capacity than that educated by our narrative culture. And, if we say with George Steiner, that “the world of Auschwitz lies outside discourse just as it lies outside reason” (cited by White, 43), whence comes the sense of the unspeakable and unrepresentable? We will not resolve the difficulty by forbidding any mode other than the literal chronicle, which would be equivalent to demanding the denarrativization of the events in question. This is only a despairing manner of setting aside every figurative addition to a literal representation of the events. This solution is despairing in the sense that it falls back on the illusions of a naïve realism that were common to the principal current of the nineteenth-century novel and to the positivist school of historiography. It is an illusion to believe that factual statements can satisfy the idea of the unrepresentable, as though facts could through the virtue of their literal presentation be dissociated from their representation in the form of events in a history; events, history, plot all go together on the plane of figuration. White pushes his argument to the point of striking with suspicion the whole enterprise of the realistic representation of reality by which Erich Auerbach characterized Western culture.39 At the end of his essay, White attempts a kind of heroic escape, by suggesting that certain modes of writing that make use of postmodernism—which he persists in calling modernist—may have a certain affinity with the opacity of the event. For example, “intransitive” writing, a notion borrowed from Roland Barthes, who compares it in turn with the “middle voice” of ancient Greek grammar. White thinks he can see this in certain of Derrida’s comments about “differance.” But, if the style of “middle-voicedness” effectively breaks with realism, what assures that it has any affinity with the “new actuality”? Was not totalitarianism modernist? Does it suffice to break with realist representation to bring language close to not only the opacity but also the inadmissible character of the “final solution”? Everything happens as though, at the end of the essay, the critique without any concession of naïve realism paradoxically contributes to reinforcing the truth claim coming from elsewhere than discourse, in such a way as to render derisory the outlines of a compromise with a realism that has become undiscoverable.

Over against White, Carlo Ginzburg makes an impassioned plea in favor not of realism but of historical reality itself as what is intended by testimony. He recalls the declaration from Deuteronomy 19:15 (which he cites in Latin): non stabit testis unus contra aliquem (“A single witness shall not suffice to convict a person of any crime or wrongdoing in connection with any offense that may be committed” [NRSV])—and compares it to the Code of Justinian: testis unus, testis nullus. With this, the title “just one witness” sounds a despairing note, as though the accumulated documents remain below the threshold of double testimony, unless by antiphrasis we point to the excess of such testimonies in regard to the capacity of plots to produce a coherent and acceptable discourse.40 His plea in favor of the reality of the historical past, akin to that of Pierre Vidal-Naquet in The Jews: History, Memory, and the Present and Assassins of Memory,41 thus bears the double aspect of an uncontestable attestation and of a moral protest that prolongs the violence of the impulse that pushed a survivor like Primo Levi to bear witness.42 It is this interweaving of attestation and protest in the case of literature about the Shoah that we need to reflect upon. If we do not want to allow this mixed status, we will not understand why and how representation must integrate into its formulation the “inadmissible” dimension of the event. But then, it is the citizen as much as the historian who is summoned by the event. And he is summoned at the level of his participation in collective memory, before which the historian is called upon to give account. But he does not do so by leaving behind the critical resources stemming from his professional competence as a historian. The historian’s task faced with events “at the limits” is not confined to the habitual hunt for falsehood that, ever since the Donation of Constantine, has become the great specialty of the discipline of history. His task extends to the discrimination among testimonies as a function of their origin: those of the survivors, the executioners,43 the spectators implied in different ways in the mass atrocities, differ. So it comes down to historical criticism to explain why we cannot write an all-encompassing history that would annul the insurmountable differences among these perspectives. This critical consideration can in particular lead to dissipating useless quarrels like the ones that contrapose the history of the everyday life of the German people, the history of the economic, social, cultural, and ideological constraints, and the history of the decision making at the summit level of the state. The notions of scale, of a choice of scale, of changes in scale can be usefully invoked here in the encounter with the confrontation between what are called “functionalist” and an “intentionalist” interpretations. As we have learned above, the very notions of fact and interpretation vary depending on the scale considered. Nor must the historian of the Shoah allow himself to be intimidated by the postulate that to explain is to excuse, to understand is to pardon. The moral judgment interwoven with historical judgment stems from another layer of historical meaning than that of description and explanation. Therefore it must not intimidate the historian to the point of leading him to censor himself.

Is it possible to be more precise about the way in which the moral judgment, signified by the expression “unacceptable,” addressed by Friedlander to forms of figuration of the event, is to be articulated on the basis of the critical vigilance of which we have just given some examples? This is the question Adorno posed when he asked, “What does coming to terms with [Aufarbeitung] the past mean?”44 We can find some help in a prudent recourse to such psychoanalytic categories as trauma, repetition, the work of memory, understood as “working through,” and, above all, in that of transference applied not to a person but to situations in which the agents of history were diversely “cathected.” Thus above I risked speaking of the use and abuse of memory and singularly of the difficulties of blocked memory. It is a comparable situation that history is confronted with in the face of events at the limit. We need to begin again here from the diversity of situations of the summoned witnesses, as was said above. It is not just a question of different points of view, but of heterogeneous investments. This is the way explored by Dominick La Capra in his contribution to Probing the Limits: aged Nazis, young Jews or Germans, and so on, are implied in different transference situations. The question thus arises whether a criterion of acceptability may be disengaged in a such a way that this or that attempted historical treatment of supremely traumatic events is capable of accompanying and facilitating the process of working through.45 The criterion in this sense is more therapeutic than epistemological. It is difficult to work with inasmuch as the historian in turn is in an indirect transference relation to the trauma by way of the testimonies that he privileges. The historian too has a problem of identification at the moment of choosing his target. This split in the transference relation confirms the hybrid position of the historian confronted with the Holocaust. As a professional scholar, he speaks in the third person, and, as a critical intellectual, in the first person. But we cannot fix once and for all the distinction between the expert and the one Raymond Aron would have called the engaged spectator.

If now we turn in the direction of the source of the demand for truth and therefore the place of the initial trauma, we need to say that this source is not in the representation but rather in the lived experience of “making history” as it is confronted in different ways by the protagonists. Following Habermas, we have said that this means “reaching a deeper level of solidarity with those bearing a human form.”46 It is in this sense that the event called Auschwitz is an event at the limits. It lies in individual and collective memory before being in the discourse of the historian. And it is from this source that the attestation-protestation arises that places the historian-citizen in a situation of responsibility as regards the past.

Must we continue to call this limit imposed on the claims to self-sufficiency of the rhetorical forms of representation external? No, if we consider the true nature of the relation of history to memory, which is that of a critical reprise, one that is internal as much as external. Yes, if we consider the origin of this claim, which is less tied to the actual use of rhetorical forms than to literary theory—whether structuralist or not—which proclaims the closure in on themselves of narrative and rhetorical configurations and announces the exclusion of any extralinguistic referent. Having said this, external and/or internal, the inherent limit of the event said to be “at the limits” prolongs its effects at the core of the representation whose limits it makes appear; that is, the impossible adequation of the available forms of figuration to the demand for truth arising from the heart of lived history. Must we then conclude the exhaustion of these forms, above all of those inherited from the naturalist and realist tradition of the nineteenth-century novel and history text? Undoubtedly, yes. But this assertion must stimulate rather than preclude the exploration of alternative modes of expression, eventually connected to other supports than just that of the printed book: drama, film, the plastic arts. We are not forbidden an ongoing search for a way to fill the gap between the representative capacity of discourse and what the event demands, even while guarding ourselves against nourishing an illusion in favor of those styles of writing Hayden White calls “modernist,” parallel to the one he condemns on the side of the realist tradition.

It follows from these considerations that to attempt to write the history of the “final solution” is not a hopeless undertaking, if we do not forget the origin of the limits in principle that affect it. It is rather the occasion to recall the trajectory that critical thought has to carry out, leading back from representation to explanation/understanding and from there to documentary work, up to the ultimate testimonies whose collection we know is broken up among the side of the executioners and that of the victims, that of the survivors, that of the different involved spectators.47

Someone may ask how the problems posed by the writing of the event “at the limits” called Auschwitz are exemplary for a general reflection on historiography. They are so insofar as they themselves are, as such, problems “at the limit.” Along our way we have encountered several illustrations of this extreme problematization: the impossibility of neutralizing the differences in position of witnesses in the interplay of scales; the impossibility of summing up in one all-encompassing history the reconstructions backed up by heterogeneous affective investments; the unsurpassable dialectic between uniqueness and incomparability at the very heart of the idea of singularity. Perhaps every singularity—turn by turn, unique and/or incomparable—is a sign of exemplarity in this double sense.


As a first approximation, referring to the iconic dimension of the historian’s representation should not introduce any large-scale adjustments to my analysis. Either it is only a question of opposing two fully constituted literary genres, or it is only a matter of accentuating certain features of narrativity I have already noted and amply commented on under the heading of the rhetorical effects that go along with emplotment.

What I want to show, however, is that with this term “image” an aporia comes to the fore that has its place of origin in the iconic constitution of memory itself.

Let us linger a moment at the level of what I have called a first approximation. The pair historical narrative and fictional narrative, as they appear as already constituted at the level of literary genres, is clearly antinomical. A novel, even a realist novel, is something other an a history book. They are distinguished from each other by the nature of the implicit contract between the writer and the reader. Even when not clearly stated, this contract sets up different expectations on the side of the reader and different promises on that of the author. In opening a novel, the reader is prepared to enter an unreal universe concerning which the question where and when these things took place is incongruous. In return, the reader is disposed to carry out what Coleridge called a “willful suspension of disbelief,” with the reservation that the story told is an interesting one. The reader willingly suspends his disbelief, his incredulity, and he accepts playing along as if—as if the things recounted did happen. In opening a history book, the reader expects, under the guidance of a mass of archives, to reenter a world of events that actually occurred. What is more, in crossing the threshold of what is written, he stays on guard, casts a critical eye, and demands if not a true discourse comparable to that of a physics text, at least a plausible one, one that is admissible, probable, and in any case honest and truthful. Having been taught to look out for falsehoods, he does not want to have to deal with a liar.48

So long as we remain in this way on the plane of constituted literary genres, no confusion is admissible, at least in principle, between the two kinds of narratives. Unreality and reality are taken as heterogeneous modes of reference. Historical intentionality implies that the historian’s constructions have the ambition of being reconstructions that more or less approach what one day was “real,” whatever the difficulties may be that are taken as resolved by what I continue to call standing for, to which the closing discussion of this chapter will be devoted. Nevertheless, despite the distinction in principle between “real” past and “unreal” fiction, a dialectical treatment of this elementary dichotomy imposes itself through the fact of the interweaving of the effects exercised by fictions and true narratives at the level of what we can call the “world of the text,” the keystone to a theory of reading.49

What I previously called the “fictionalization of historical discourse” can be reformulated as the interweaving of readability and visibility at the threshold of the historian’s representation. One is then tempted to look for the key to this imaginary structure of a new kind on the side of the rhetorical effects referred to above. Do we not call tropes those figures that not only ornament but articulate historical discourse in its literary phase? The suggestion is a good one, but it leads further than anticipated. What must be unrolled, as in examining the back side of a tapestry, is precisely the interwoven connection of readability and visibility at the level of the reception of the literary text. In fact, narrative gives itself to be understood and seen. Dissociation of the two interwoven effects is facilitated when the picture and the sequence, the descriptive stasis and the properly narrative advance, itself precipitated by what Aristotle’s Poetics calls peripeteia, particularly when it occurs through theatrical or violent effects, are separated. The historian is well aware of this alternation.50 Often it is through a set of pictures that he depicts the situation wherein the beginning of his narrative is implanted. He can end his work in the same way, unless he should choose to leave it in suspense, like Thomas Mann deliberately losing sight of his hero at the end of The Magic Mountain. The historian is not unaware of these strategies for closing a narrative, which only make sense to the eye of an educated reader thanks to an expert game of frustrating one’s usual expectations. But it is with the portrait of characters in narratives, whether they be stories from everyday life, narrative fictions, or historical fictions, that visibility decisively carries the day over readability. Here we have a constant thesis of this book: the characters in a narrative are emplotted along with the events that, taken together, make up the story told. With the portrait, distinguished from the warp of the narrative, the pairing of readability and visibility stands out most clearly.

It turns out that this pairing gives rise to noteworthy exchanges that are the sources of meaning effects comparable to those that get produced between fiction and historical narrative. We can say that an art lover reads a painting51 and that a narrator depicts a battle scene. How are such exchanges possible? Is it only when the narrative unfolds a space, a landscape, places, or when it lingers over a face, a posture, a position that a whole character gives itself as to be seen? In short, is there readability only in a polar relation to visibility—a distinction that superimposing the two extremes does not abolish? Or should we go so far as to say that in every case the narrative sets something before our eyes, gives something to be seen? This is what Aristotle had already suggested in his remarks on metaphor in Book III of his Rhetoric. Considering the “virtues of lexis” (locution, elocution), he says that this virtue consists in “setting the scene before our eyes” (1410b33). This power of the figure to set the scene before our eyes has to be linked to a more fundamental power that defines the rhetorical project considered as a whole, namely, “the faculty to discover speculatively what, in each case, is likely to persuade” (1356b25–26 and 1356a19–20). The pithanon, the “persuasive as such,” is the recurrent theme of rhetoric. To be sure, persuasion is not seduction—and Aristotle’s whole ambition was to stabilize rhetoric halfway between logic and sophistry, thanks to the connection between the persuasive and the reasonable in the sense of the probable (to eikos). This definition of rhetoric as the tekhnē of discourse likely to persuade lies at the origin of every prestige that the imagination is capable of grafting to the visibility of figures of language.52

Guided by the perplexity of the ancients, we shall take up again the broken thread of our reflections on the dialectic of presence and absence begun within the framework of a history of social representations. There we admitted that the functioning of this dialectic in the representative practice of social agents was not really made clear until it was taken up and rendered explicit through the very discourse of the historian representing to himself the representation of such social agents. The representing operation, at the level of which from here on we want to keep ourselves, will constitute not only a complement with regard to the represented object of history but also a surplus, inasmuch as the representing operation can be taken as the reflective phase of the represented object.

I propose taking as my guide here the works that Louis Marin has devoted to the prestige of the image, as he finds it lucidly fomented by the great writers of the seventeenth century to the glory of monarchical power and its incarnate figure, the king. I shall keep in reserve during the course of my reading of the Portrait of the King the question whether some instruction, concerning the relations between the justification of power and the prestige of the image, persists for the citizens of a democracy who believe themselves to have broken with singing the king’s praises, beyond what has become for them a kind of slightly exotic case.53

Marin immediately puts the accent on the force, the power, of the image substituted for something present elsewhere. It is the transitive dimension of the image that is thereby underscored in what we can call a “theory of effects” that finds strong echoes in Pascal (7). “The power-effect of representation is the representation itself.” This power-effect finds its privileged field of exercise in the political sphere, inasmuch as there power is animated by the desire for the absolute. It is the mark of the absolute imposed on power that in a way stirs up the imaginary, leading it in a fantastic direction: lacking an actual infinity, what takes its place is “the imaginary absolute of the monarch.” The king is not really king, that is, monarch, except through the images that confer upon him a reputedly real presence. Here Marin comes up with a seductive hypothesis according to which “the political imaginary and symbolism of the absolute monarch” had rediscovered the “eucharistic motif” whose central role Marin’s earlier study on the Logique de Port-Royal had demonstrated. The utterance “this is my body” governs not just the whole semiotics of the attributive proposition on the logical plane, but the discourse of power on the political one.54

The phrase “L’État c’est moi” is the political doublet of the one that consecrates the Eucharistic host.55 That this political “transposition” stems from the order of the “lure,” in line with the “fantastic” referred to by Plato in the Sophist, is known only on the basis of an external, ironic, critical discourse, which Marin sees formulated in the famous Pensées, where Pascal mercilessly demonstrates the hidden play of exchanges between the discourse of force and that of justice. Three levels of discourse are set up and practiced in this way: the implicit one in the representation at work at the heart of social practice, the explicit one of the representation articulated by the praise of power, and the one that brings to light power as representation and representation as power. Will the third discourse, which gives an anthropological dimension to the interplay of representation and power, have the virtue of setting in motion another inquiry that would bear on a comparable interplay occurring beyond the fall of the monarchy, in new projections of the king’s power? This is a question I shall reserve for later.

Whatever may be said about the political resonance of the theology of transubstantiation and the potentially blasphemous turn in such an operation, it is worth noting that the discourse of power, once it is made explicit on the plane of the historian’s representation, simultaneously assumes the two forms of narrative (evocative of some absence) and icon (the bearer of some real presence). Yet, taken together, absence and presence produce the representation of power in “the fantastic representation of the absolute monarch in his portrait and in his name” (12). “There is a dimension of narrative and of recitation in the royal portrait that is also the celebration of the king’s historical body, his monumental tomb in and through the representation of history” (12).56 Marin offers two illustrations of this double function of the representation of power. First, with the commentary on the “Projet de l’histoire de Louis XIV” addressed to Colbert by the court historian Pellisson-Fontainier,57 it is the readability of the narrative that generates visibility in a quasi “portraiture.” Second, with the treatment of the “historic medal” struck with the effigy of Louis XIV as “royal host,” it is the visibility of the portrait that generates the readability of a quasi recitative of glory.58

The “Project for a History of Louis XIV” in effect is a quite extraordinary text in that it presents to its reader’s eyes the stratagems of a yet to be written history, along with the barely concealed plan of enticing its ultimate addressee, the king, to fall into the trap of providing a royal subvention for it. The stratagem for writing history thereby laid bare comes down to a cunning use of the prestige of the image used in service of rendering praise. Another rhetoric than that of figures is made use of here, the rhetoric with an Aristotelian origin of the three genres of public discourse: the judicial genre governing pleas for judgments, the deliberative one governing political decisions, and the epideictic one (elsewhere called demonstrative) illustrated by speeches of praise and blame, the funeral oration constituting its most eloquent expression. This classification, governed less by the differences in style than by the distinction among the addressees of the discourse, vigorously takes up the rule-governed exploitation of praise discourse that, in the age of absolute monarchical power, occupies the wide-open space left vacant by the deliberative genre’s having been relegated to the back burner, it having been sacrificed to the king’s secret cabinet. Where does such praise point within the order of political power? To greatness, and to the flash of such greatness, glory. The prestige of the image made use of by the “Project for a History of Louis XIV” is put to the service of such greatness and glory. The historian’s cunning in offering such a service is first of all to anticipate the way in which power that desires to be absolute thinks: “What is the phantasmatic in and through which the politics of this desire is rationalized? What is the imaginary of absolutism, and what are the role and function of the historiographer in constituting this phantasmatic and in constructing this imaginary?” (48). The trap, if we can put it this way, lies wholly in the role played in this proposal by the eulogist: “The king must be praised everywhere but, so to speak, without praise, by a narrative of all that he has been seen to do, say, and think.” The ruse succeeds if the adulator succeeds in “enabling [the reader] to conceive of his [the king’s] greatness in all sorts of ways” (53). It is not up to the writer to speak of greatness and glory, it is up to the reader under the helpful guidance of the narrative. Among the narrative resources to be brought into play in seeking this effect within the overall field of forces are abbreviation in the narration of exploits, the brevitas so dear to Tacitus in making use of litotes, the depiction of actors and scenes, and all the simulacra of presence likely to give rise to pleasure in reading. A place of honor is to be assigned to narrative hypotyposis, that “animated and striking description” (Robert dictionary),59 which, more than any other rhetorical procedure, sets things before the eyes60 and thereby sets up the character, the event, the scene as instructive examples: “[History] puts all the great things it encounters in a better light through a noble and a more composed style, which encloses a lot in a little space with no wasted words” (Pensées, bundle V, cited in Portrait of the King, 82). This concern to show in telling is even more strongly marked in the Éloge historique du roi sur ses conquêtes depuis l’année 1672 jusqu’en 1678 by Racine and Boileau. Marin quotes these eloquent phrases from it: “Some people more particularly zealous for his glory wanted to have in their studies a précis in pictures of the prince’s greatest actions; which prompted this little work that encloses so many marvels in a very small space, so as at any time to put before their eyes at all times that which is the dearest occupation of their thoughts” (122). The monarch’s greatness stands forth before the eye as soon as the strategy of the narrative succeeds in making it appear like the archactor of the gesture.

Such in brief is the historian’s cunning, worthy of the Greek mētis described by Jean-Pierre Vernant. It consists in concealing the very panegyric project that, like the repressed, has to return in the reader’s mouth. Thus we can speak of a “simulation of history” (74) to designate of this power of representation “that the absolute needs in order to constitute itself absolutely” (75), a power targeted on extorting the panegyric from the moment of reading. What is surprising is that the author of this historical project dared to spring the trap by stating it—to the great happiness of the contemporary historiographer. Our question will be whether, with the end of the monarchy of the Ancien Regime and the transfer of sovereignty and its attributes to the people, historiography has been able to eliminate from its representation every trace of the discourse of praise. At the same time, this will be to ask whether the category of greatness and what is connected to it, glory, can disappear without leaving some trace on the horizon of the history of power. Was it reserved only to the “absolutist manner of writing the absolute history of absolutism” (88) to extract from the readability of narrative the visibility of a narrative description that would succeed in “depicting rather than recounting, in making the imagination see everything that was put on paper,” following the confession by which the author ends the “Project for a History of Louis XIV”? Has modern democracy put an end to praise of the king and the phantasmatic placed in service of this praise?61

The relation between readability and visibility is inverted by the portrait medal of the king. Or rather the exchange between readability and visibility starts from the other pole. Louis Marin can say at the beginning of his study of “The Royal Host: The Historic Medal”: “To tell the king’s history in a narrative is to show it. To show the king’s history in his icon is to tell it” (121). A chiasmus is established that makes the picture speak and the narrative show, each mode of representation finding its most specific, its ownmost effect in the domain of the other. Thus we ourselves say that one reads a painting. The medal is the most remarkable procedure of iconic representation capable of telling by showing. Unlike the drawing that illustrates a text, or even a tapestry, which most often only represents one moment of history, the medal is a portrait that, like hypotyposis, offers an abbreviation of a picture. By presenting something to be seen, the medal—a specific inscription of the king’s portrait, a metal engraving—thanks to its gold and its brilliance depicts the flash of glory. What is more, the medal, like a coin, can be shown, touched, exchanged. But above all, thanks to its hardness and the fact that it lasts, it grounds a permanence of memory by transforming the passing flash of the exploit into perpetual glory. A connection to narrative is assured by the motto inscribed on the reverse side of the figure of the king’s effigy, and in his name. It assures the potentially universal exemplarity of the virtues engraved in gold. At the center stands the name. Praise comes to the name by way of the exploits and virtues. In this way the historic medal could in its day be called a monument, like those funerary sepulchers that warn and admonish all those who had been absent from the place and time of the commemorated event. The historical medal of the king was par excellence “the monumental sign of absolute political power in the infinity of its representation” (123).

Has the time of the medal run its course, at least in the West, with the fall of absolute monarchy? Did it vanish along with the praise conveyed through the king’s narrative? Yes, undoubtedly, if we put the accent on the theological connotation that authorizes calling the medal a “royal host,” a “sacramental Host of the power of the State” (134). No, perhaps, if we grant the theme of greatness a kind of transhistorical permanence that would allow it to survive the dead glory of an absolute monarch. Do not a bit of flash, a bit of glory continue to surround the contemporary figure of the prince, even when his portrait is reduced to that of a postage stamp? Nor have medals everywhere, and once and for all, disappeared.

I have said that the narrative and iconographic representation brought about by history brings to light the representation practiced by social actors. But what is brought to light by the strategies of representation once these are said to be fomented by a phantasmatic imaginary and denounced as simulacra? Who says this?

Louis Marin’s answer in Portrait of the King is striking. It is in Pascal’s Pensées, dealing with force and justice, that the reader sees the glamour of the imagination dismantled. So it is not to the plane of the historiographical operation that the thinker of the Pensées brings his lucidity, but to that of a philosophical anthropology whose propositions abstract from every location in geographical space and historical time, even if for a discourse of a still higher degree it would be easy to take this or that pensée as having a particular time and place. But this is not how the Pensées ask to be read: the contract with the reader is here that of veracity confronted with dissimulation.62 What the famous pensées dealing with force and justice bring to light are the “effects” of the imagination that are summed up by the as yet not referred to expression of leading to belief. This “effect” is one of meaning insofar as it is one of force. Two proposals are stated by Marin:

1. Discourse is the mode of existence of an imaginary of force, an imaginary whose name is “power.”

2. Power is the imaginary of force when it is uttered as discourse of justice. (16)

On the one side, therefore, force becomes power by taking hold of the discourse of justice; on the other, the discourse of justice becomes power in standing for the effects of force. Everything takes place within the circular relation between standing for and being taken for. This is the circle of coming to believe. Here the imaginary does not designate merely the visibility of the icon that sets before the eyes the events and characters of a narrative, but a discursive power.

It is not a question here of undertaking an exegesis of those fragments that suggest placing in series the three key words “force, justice, imagination,” as though only one ordering were authorized. Sometimes they are commented on separately, sometimes in pairs, sometimes as all three at once. Therefore it is an interpretation, for all that a highly plausible one, that Louis Marin proposes in the magnificent pages that constitute the “overture” of his work under the title “The King, or Force Justified: Pascalian Commentaries.” His gathering and ordering of statements taken from the Fragments are openly oriented by a concern to dismantle the stratagems of the imagination of power. “One must have deeper motives and judge everything accordingly, but go on talking like an ordinary person” (§91). Only the pair force/justice is made use of by this text, and we could see it as a sediment of the famous assertion that “as men could not make might obey right, they have made right obey might. As they could not fortify justice they have justified force, so that right and might live together and peace reigns, the sovereign good” (§81). The justification of force can be taken as the linchpin of a whole demonstration where one takes up the titles of the just that ought to be followed and those of force that is to be obeyed, then the reversal of the apparent symmetries between force and justice. “Right without might is challenged, because there are always evil men about. Might without right is denounced” (§103). We can leave aside the question of what would be their reconciliation: “We must therefore combine right and might.” What is important for our project is the self-justifying discourse of force. At this point it is easy to introduce the no less famous fragment on the imagination.63 That, in speaking of “this master of error and falsehood,” of “the war between the senses and reason” (§44), Pascal had explicitly in mind the effects of political power is disputable. The discourse of philosophical anthropology is placed under the aegis of broader concepts such as impoverishment and vanity. In any case, taken together, fragments 44, 87, and 828 authorize, among several possible readings, treating the imagination as bringing about the process that justifies force. The imagination is itself a power—an “arrogant force” (§44). “It makes us believe, doubt, deny reason.” “It dispenses reputation . . . makes us respect and revere persons, works, laws, the great.” Other effects: “Love or hate alters the face of justice.” Also: “Imagination decides everything; it creates the beauty, justice, and happiness, which is the world’s supreme good.” What other power than the imagination would be able to clothe judges, physicians, preachers? The most eloquent of the fragments to my eyes is the one among the unclassified papers in series 31 that confronts in a striking way the “bonds of necessity” and the “bonds of the imagination”: “The bonds securing men’s mutual respect are generally bonds of necessity, for there must be differences of degree, since all men want to be on top, and all cannot be, but some can. . . . And that is where imagination begins to play its part. Until then, pure power did it, now it is power, maintained by imagination in a certain faction, in France the nobles, in Switzerland commoners, etc. So these bonds securing respect for a particular person are bonds of imagination” (§828). At this point, Pascal’s discourse is certainly one of an accusation against force without justice. It hits the “tyranny” of the power of the great, but if it also strikes the vanity of power it is because it aims well beyond politics.64

How far can the critical epistemology of the historiographical operation advance along this path pointed out by Louis Marin’s “Pascalian commentaries”? Not very far beyond its own region of competence, even if we extend this to include the order of representations linked to social practice. But far enough nevertheless if we are to find a reason, encouragement, a handhold in the supra-political dimension of anthropological discourse at the moment of posing the question whether other figures of power than that of the absolute king are capable of receiving clarification, even laterally, thanks to the enlarging of the problematic of the representation of power that the Pascalian anthropology makes possible.

Over the course of our reflection we have set down several milestones on a path that, without leaving behind the representations of power, leads in the direction of the neighborhood of post-absolutist political configurations where other forms of the prestige of the image are likely to occur, unless these be the same ones in different guise.

One word may crystallize the question: “greatness.” In fact, it belongs to the two registers of politics and anthropology. Furthermore, it is in part bound to the problematic of representation by way of the rhetoric of praise. Let us return one last time to Pascal. On the one side, greatness belongs to the same constellation as does the impoverishment for which it is the other pole in the order of contrarieties and disproportion in human beings, and as does the vanity that builds on such impoverishment: “Man’s greatness comes from knowing he is wretched: a tree does not know it is wretched. Thus it is wretched to know that one is wretched, but there is greatness in knowing one is wretched” (§114). On another side, greatness touches politics: “All these examples of wretchedness prove his greatness. It is the wretchedness of a great lord, the wretchedness of a dispossessed king” (§116). Pascal continues: “Who indeed would think himself unhappy not to be king except one who had been dispossessed?” (§117). This figure of a dispossessed king is not simple: man in general can be seen as a dispossessed king. And it is this dispossessed king that, in an astonishing fable meant for the young prince, Pascal sees “tossed up by a storm on some unknown island, where the inhabitants were having difficulty in finding their king who had been lost.” This man who finds himself resembling the lost king is “taken for him, recognized as such by all the people.” And what did he do? “He received all the respect that they wished to render to him and allowed himself to be treated as king.”65 Therefore it is a “portrait effect,” an “effect of representation” that makes the king. And in turn it is the image, dedicated to the prince, of this “drowned king,” become the “legitimate usurper,” that gives the instructive force to the epistle. Politics and anthropology conjoin in this image. At the same time, the secret is revealed of the simulative representations that underlie those greatnesses in the flesh to whom the king belongs and all those we call or that are called great.

If greatness can in this way belong to the two registers of anthropology (“man”) and politics (the “king”), it is because it contains in principle (in its known truth, like all principles, “not only through our reason but also through our heart” [§110]) an ordering and hierarchical rule. The fragment is well known about the “orders of greatness”: the greatness of the flesh, of the mind, of charity (§308). Each greatness has its degree of visibility, its luster, its flash—kings are joined with the rich and the captains among the “greatness of the flesh.”66

From these considerations comes the question by which I will end our inquiry into the prestige of the image entangled with the historian’s representation. What remains of the theme of greatness in the narration of power after the elimination of the figure of the absolute king? To ask about the possible permanence of the theme of power is at the same time to ask about the persistence of the rhetoric of praise that is its literary correlate, with its cortege of prestigious images. Has greatness abandoned the political field? And must and can historians renounce the discourse of praise with its vanities?

I will answer the first question by two comments that I shall place in no special order, in that I am concerned not to treat a problem of political philosophy with an air of expertise that I lack, a problem moreover that exceeds the competence of an epistemology of the historical operation. However the question cannot be avoided inasmuch as the nation state remains the organizing center for the ordinary referents of historical discourse, given the lack of access to a cosmopolitan point of view. In order to remain such an organizing center, must the nation state continue to be celebrated as great? This reformulation of the question leads to my first comment. I borrow it from Hegel’s philosophy of the state in the Philosophy of Right. In considering the power of the prince (§275), Hegel distinguishes three elements that stem from the constitution as a rational totality: to the universality of the constitution and the laws, and to the process of deliberation, gets added “the moment of ultimate decision, as self-determination.” In it lies “the distinctive principle of the power of the crown.” This moment is incarnated in an individual who, in monarchical regimes, is destined to the dignity of the throne by birth. However contingent this moment may be, a contingency taken up into hereditary right, it is irreducibly constitutive of state sovereignty. Someone may object that Hegel’s political thought does not get beyond the orb of the monarchical principle and in this sense from the space of absolute politics, despite his sympathies for a liberal monarchy. But Hegel is already the thinker of the modern post-revolutionary state, that is, the constitutional state in contrast to the aristocratic one. It is within these limits that the question is posed whether in a constitutional regime politics can be exempted from the moment of ultimate decision and, to put it in a word, whether it can escape entirely from the personification of power. Contemporary history seems to ratify this question. Eric Weil, in his Philosophie politique, proposes a rational framework with which to discuss this. He defines the state in formal terms: “The State is the organization of a historical community. Organized into a State, the community is capable of making decisions.”67 It is along this trajectory that a decision, within the framework of the constitution, on the basis of an administration during the stage of deliberation and execution, and by means of parliament for discussing and passing laws, poses in fine the problem of the exercise of political authority, in particular in tragic situations where the physical existence and moral integrity of the state are in danger. This is when the true homme d’État reveals himself. With this notion of a homme d’État, in a fully constitutional system, comes the Hegelian question of the prince as the incarnation hic et nunc of the “moment of ultimate decision, as self-determination.” This moment is also the moment of greatness.

Someone may object that under the figure of the homme d’État we are fraudulently reintroducing the portrait of the king. So I will offer my second comment, which will redistribute the figures of greatness in a broader social space, worthy of a Pascalian consideration of the orders of greatness because of its breadth. It was possible during the last decade of the twentieth century for a book to bear the subtitle “Economies of Greatness” and to open a new career for the idea of greatness in liaison not with the greatness of political power but with the greatness of the widest sense of justification, the demand for justice.68 In disputed situations, debates that appeal to people’s opinions must appeal to argumentative strategies meant to justify the action or to sustain the criticisms at the heart of the dispute. What is remarkable is not just that the idea of greatness should return to the sociology of action and therefore also to the history of representations, but that it does so in a pluralistic way. There are economies of greatness. For example, the legitimate forms of the common good are called great in typical situations of differences of opinion once these are legitimated by typical forms of argument. It doesn’t really matter here how the arguments are selected on the basis of some canonical text of political philosophy. Their irreducible plurality means that one is great in different ways depending on the qualifying tests that take place in the religious city, the domestic city, the city of popular opinion, the civic city, or the industrial city. For my thesis, what is important is that greatness should be taken into account by practical philosophy and in the social sciences in connection with the idea of justification as one of the ways of apprehending the common good at the heart of being-with-others. It is still very much a question of “political forms of greatness” (see Portrait of the King, 89–93), but in such an expanded sense of the term “political” that the prestige of the king in his portrait finds itself entirely exorcised by the substitution of the people and their claims to justice for the figure of the king. The return of the theme of greatness is all the more striking because of this.

This double resistance to elimination of the theme of greatness in a political philosophy centered in turn on the state and in a sociology that takes into account justified action authorizes our posing the question that will cap our inquiry concerning the prestige of the image in the praise of greatness. If the theme of greatness is inexpugnable, does this also apply to the rhetoric of praise, which, in the age of absolute monarchy, was shamelessly taken to the point of crossing the subtle line that distinguishes praise from flattery? History written by such “great” names as Ranke and Michelet cannot escape this indiscreet question. To be sure, it is not in order to judge past actions, hence to esteem them as great or not, that Ranke states that he will limit himself to events “as they actually occurred.” This principle, in which I readily read a claim to trustworthiness, was above all else the expression of a restraint, a withdrawal from the region of subjective preferences and a renouncing of selective praise. But does praise not take refuge in the vow that we read in Ranke’s Nachlass: “Every epoch is directly under God, and its value depends not on what comes from it, but its very existence itself, in its own self. . . . All generations of mankind are equally justified in the sight of God, and so must the historian view the thing.”69 The ideas of an epoch and a generation are more diffuse than those of historical individuals, yet they constitute meaningful units to which the historian’s esteem is directed, justification in God’s sight adding the seal of theology to the discretion of praise.

The case of Michelet is even more striking. Few historians have expressed admiration for the great figures among those who built France with as much freedom and jubilation. France itself never so merited being designated by its proper name as in the successive prefaces to his History of France.70 Have historians of the French Revolution, from Guizot to Furet, escaped this circle of praise? And does not being a declared thurifer suffice to exempt one?71 Is not the discreet charm of the nation state, the usual turning point to the modern era for history as made and for history as it is told, the coil spring for a restricted praise that, setting aside any trickery, repeats the admitted strategy of the “Project for a History of Louis XIV”: “The king is everywhere to be praised, but so to speak without praise, by a narrative of everything what we have seen him do, say, and think”? And does not the same vow “to draw [the epithets and magnificent praises that the king merits] from the mouth of the reader by the things themselves” continue?

This question will seem less incongruous if, in place of praise, we put blame, its contrary in the class of epideictic discourses, following the classification scheme coming down to us from the rhetorics of antiquity. Is it not extreme blame, under the litotes of the unacceptable, that stamps the “final solution” as infamous and that above led to my reflections on the “limits of representation”? Do not the events “at the limits” referred to there occupy in our own discourse the pole opposed to that of the signs of greatness used by praise? In truth, this is a disturbing symmetry that sets back to back the absolute blame inflicted on Nazi politics by our moral conscience and the absolute praise addressed by his subjects to the king in his portrait.


This concluding section is meant both to present a recapitulation of the path covered in this chapter and to open a question that surpasses the resources of the epistemology of historiography and reaches the threshold of an ontology of existence in history, for which I shall make use of the phrase “our historical condition.”

“Standing for” condenses within itself all the expectations, the exigencies, and the aporias linked to what I have elsewhere spoken of as the historian’s intention or intentionality. It indicates the expectation attached to the historical knowledge of constructions constituting reconstructions of the course of past events. Above, I introduced this relation in terms of the features of a contract between the writer and the reader. Unlike the contract between an author and a reader of fiction that rests on the double convention of suspending the expectation of any description of some extralinguistic reality and, in return, of holding the reader’s interest, the author and reader of a historical text agree that it will deal with situations, events, connections, and characters who once really existed, that is, before the narrative of them is put together, the interest or pleasure in reading coming as a kind of added surplus. The question now posed is whether, how, and to what degree the historian satisfies the expectation and promise conveyed by this contract.

I want to emphasize two complementary replies. First answer: the suspicion that the promise has not and cannot be kept is at its height in the phase of representation, at the moment when, paradoxically, the historian seemed best equipped to honor the intention of representing the past. Is not this intention the soul of all the operations placed under the heading of the historian’s representation? The second answer is that the reply to the suspicion of betrayal does not lie only in the moment of literary representation but rather in its articulation in terms of the two prior moments of explanation/understanding and of documentation, and, if we move back even further, in the articulation of history on the basis of memory.

Expectation seems at its height, with regard to the capacity of historiography to keep its contract about reading, with the phase of the historian’s representation. This representation means to be a representation of . . . If the constructions of the explanation/understanding phase aim at constituting re-constructions of the past, this intention seems to be stated and demonstrated in the representative phase. Is it not in recounting, in submitting the narrative to the turns of a style, and, crowning everything, in setting before the eyes, that one ratifies or, to take up again one of Roger Chartier’s expressions, accredits historical discourse?72 We can put it this way. What in Time and Narrative I called the “robust conviction” that animates the historian’s work is itself brought before the reader’s eyes by the literary writing that both signs and fills the contract in the three ways that run through in turn the narration, the rhetoric, and the imaginative aspect. How can historical intentionality not be at its height with the modes of writing that do not limit themselves to giving a linguistic covering to an understanding of the past that would already be wholly constituted and ready made before being invested in literary forms? Things would be much simpler if the written form of historiography were not to contribute its cognitive value, if the explanation/understanding were complete before being communicated through writing to a public of readers. But, now that we have renounced taking expression for a neutral, transparent garment thrown over a signification complete in its meaning, as Husserl maintained at the beginning of his Logical Investigations, now therefore that we are accustomed to taking thought and language as inseparable, we are prepared to listen to declarations diametrically opposed to such a setting out of play of language—that is, that in the case of the literary writing of history, narrativity adds its modes of intelligibility to those of explanation/understanding; in turn, these figures of style can be recognized to be figures of thought capable of adding a specific dimension of exhibition to the readability belonging to narratives. In short, the whole movement that carried explanation/understanding toward literary representation, and the whole internal movement of representation that displaced readability toward visibility, are both clearly meant to remain in the service of the transitive energy of the historian’s representation. Yes, the historian’s representation as such ought to testify to the historian’s ability to keep the pact with the reader.

And yet . . .

And yet we have seen the resistance that the literary form opposes to externalization in the extratextual grow with the same rhythm as does the realist impulse. The narrative form, in giving the narrative a closure internal to the plot, tends to produce a sense of an ending, even when the narrator, in misleading the readers’ expectation, undertakes to deceive them through strategies aimed at a kind of non-ending. In this way the very act of recounting comes to split off from that “real” thereby put in parentheses. An effect of the same order proceeds, we have seen, from the interplay of figures of style, to the point of rendering unclear the boundary between fiction and reality, in that these figures claim to be common to everything that is presented as a discursive tale. The paradox is at its height with the strategies meant to set things before our eyes. To the very degree that they give rise to resemblance, they are capable of supporting Roland Barthes’s criticism aimed at the “reality effect.” In this regard, in thinking of microhistory, we may congratulate the credibility effect these narratives “close to the people” engender by means of such proximity, but then, upon reflection, we may be surprised by the exoticism these descriptions give rise to which their very precision renders alien, even foreign to us. The reader finds himself in the situation of Fabrice at the Battle of Waterloo, incapable of giving any form to the very idea of a battle, still less of giving it a name under which it will be celebrated by those who will set out to place the “details” in some picture whose visibility will cloud our vision to the point of blindness. In the words of Jacques Revel, “read too closely, the image in the carpet is not easy to decipher.”73 There is another way of setting things before our eyes whose effect is to distance and at the limit to exile them. Writing on the broad scale, that which depicts historical periods, creates an effect that we can still call visual, that is, the picture of a synoptic vision. The scale of the gaze is then defined by its magnifying power, as is said of a telescope. An inverse problematic from the preceding one thereby arises from history presented in terms of large-scale features. A new kind of closure threatens, that of those grand narratives that tend to link up with sagas and foundational legends. A logic of a new kind silently is set in place, which Frank R. Ankersmit has attempted to close in on itself: the logic of narratios capable of covering vast ranges of history.74 Use of a proper name—French Revolution, “Final Solution,” and so on—is one of the distinctive signs of the circular logic in virtue of which the proper name functions as the logical subject for a whole series of attributes that develop it in terms of events, structures, persons, and institutions. These narratios, Ankersmit tells us, tend toward self-referentiality, the meaning of the proper name being given nowhere else than through this series of attributes. The result, on the one hand, is the incommensurability among narratios said to deal with the same theme and, on the other hand, the transfer to individual authors of the great controversial narratios opened by rival histories. Do we not speak of Michelet’s, Mathiez’s, Furet’s history of the French Revolution? The epistemological discussion thus finds itself carried into the field of what in the next chapter I shall call interpretation, in a limited sense, where the accent is placed on the commitment of the historian’s subjectivity. There is, after all, only one Michelet, one Furet, confronted with the unique French Revolution.75

In this way, in an unanticipated fashion, the suspicion about closure applied to small-scale narrative and that applied to larger-scale ones overlap. In one case the suspicion sets up an invisible barrier between the signifier/signified pair and the referent; in the second it opens a logical abyss between the presumed real and the cycle formed by the quasi-personified subject and the cortege of events that qualify it. In this way the literary modes said to persuade the reader of the reality, conjunctures, structures, and events set on stage become suspect of abusing the reader’s confidence by abolishing the boundary between persuasion and making believe. This slap in the face can then only give rise to a vehement reply that transforms into a protest the spontaneous attestation that the good-faith historian attaches to a well-done work. This protest rejoins in an unexpected way Ranke’s peaceable declaration whereby he proposes to report events “as they really [eigentlich] happened.”

But, then, how are we to avoid the naïveté of such a protest?

The answer seems to me to lie in the following assertion: once the representative modes supposed to give a literary form to the historical intentionality are called into question, the only responsible way to make the attestation of reality prevail over the suspicion of nonpertinence is to put the scriptural phase back in its place in relation to the preliminary ones of comprehensive explanation and documentary proof. In other words, it is together that scripturality, comprehensive explanation, and documentary proof are capable of accrediting the truth claim of historical discourse.76 Only the movement that moves back from the art of writing to the “research techniques” and “critical procedures” is capable of raising the protest to the rank of what has become a critical attestation.

Do we not nevertheless relaunch the suspicious gesture if we cite Barthes’s phrase used as an epigram by White in The Content of the Form: “Le fait n’a jamais qu’une existence linguistique”? And have I not myself in dealing with the historical fact proposed to distinguish the proposition stating “the fact that . . .” from the event itself? The critical realism professed here is forced to take another step beyond the factual proposition and to invoke the testimonial dimension of the document. Indeed, it is the force of testimony that presents itself at the very heart of the documentary proof. And I do not see that we can go beyond the witness’s triple declaration: (1) I was there; (2) believe me; (3) if you don’t believe me, ask someone else. Ought we to make fun of the naïve realism of testimony? It can be done. But this would be to forget that the seed of criticism is implanted in actual testimony,77 the critique of testimony bit by bit taking over the whole sphere of documents, up to the ultimate enigma of what presents itself under the name “trace,” as the sign-effect of its cause. I have said that we have nothing better than our memory to assure ourselves of the reality of our memories—we have nothing better than testimony and criticism of testimony to accredit the historian’s representation of the past.

I have rarely to this point pronounced the word “truth,” nor have I risked any affirmation concerning the truth in history, even though at the beginning of this work I promised to compare the presumed truth of the historical representation of the past with the presumed trustworthiness of mnemonic representation.

What does the word “truth” add to the word “representation”? A risky assertion that commits the discourse of history not only to a relationship to memory, but to one with the other sciences, both the human sciences and the natural sciences. It is in relation to the truth claims of these other sciences that history’s claim to truth makes sense. Thus the criteria qualifying this claim need to be made clear. And it is quite evidently the past itself that is the referential stake of this claim. Is it possible to define this referential stake in other terms than those of correspondence, of adequation? Or to call “real” what would correspond to the assertion of some representation? It would seem not, under the threat of renouncing the very question of truth. Representation has a vis-à-vis, a Gegenüber, to use an expression from Time and Narrative that I borrowed from Karl Heussi.78 I also took the risk of speaking of a “taking the place of” to make more precise the mode of truth proper to “standing for,” to the point of taking these two expressions as synonyms.79 But we see better what senses of the notion of correspondence are excluded when we see how this notion is made specific in relation to other uses of the term “correspondence” in other disciplines of knowledge. The so-called picture theory, which would come down to a imitation-copy, is manifestly excluded. It must be said that one is never fully free of this ghost, inasmuch as the idea of resemblance seems difficult to disengage from it without remainder. Did not Plato place the whole discussion about the eikōn in terms of an internal distinction about the mimetic arts when he distinguished between two mimetics, a properly iconic mimetic and a fantastical one? But, if the mimetic also includes the fantastical it has quite clearly to distinguish itself from repetition of the same in the form of a copy. Imitation has to incorporate a minimal heterology if it is to cover such a vast country. In any case, a narrative does not resemble the event it recounts; this has been said often enough by the most convincing narrativists. It is this minimal heterology that the Aristotelian use of mimēsis in the Poetics already satisfies. Following Aristotle, in the past I myself tried to modulate the mimetic resources of narrative discourse with the yardstick of threefold mimēsis: prefiguration, configuration, refiguration. I must admit that the notions of vis-à-vis and taking the place of or standing for constitute the name of a problem rather than that of a solution. In Time and Narrative I limited myself to proposing a “conceptual articulation” to the enigma that adequation by taking the place of constitutes.80 Through this highly metahistorical effort, I attempted to save the point of Ranke’s formula that it is not the task of history to judge the past but to show events “as they really happened.” The “as” of Ranke’s formula then designates nothing other than what I call the function of standing for. The “really” past remains then inseparable from the “as” really happened.

I have nothing to change today about this attempt to explicate the concept of taking the place of or standing for. I wish instead to apply myself to another enigma that seems to me to reside at the very heart of the relation of presumed adequation between the historian’s representation and the past. Recall that Aristotle, in his theory of memory, distinguishes the recollection (mnēmē) from the image in general (eikōn) by the mark of the formerly (proteron). We can then ask what happens to the dialectic of presence and absence constitutive of the icon when in the realm of history it is applied to the condition of the anteriority of the past in relation to the narrative that is told about it.

We can say this: the historian’s representation is indeed a present image of an absent thing; but the absent thing itself gets split into disappearance into and existence in the past. Past things are abolished, but no one can make it be that they should not have been. It is this twofold status of the past that many languages express by a subtle play of verb tenses and adverbs of time. In French we say that something no longer is (n’est plus), but has been (a été). It is not unacceptable to suggest that “avoir été” (having been) constitutes the ultimate referent intended across the “n’être plus” (being no longer). Absence thus would be split between absence as intended by the present image and the absence of past things as past in relation to their “having been.” It is in this sense that “formerly” would signify reality, but the reality of the past. At this point the epistemology of history borders on the ontology of being-in-the-world. I will call our “historical condition” this realm of existence placed under the sign of a past as being no longer and having been. And the assertive vehemence of the historian’s representation as standing for the past is authorized by nothing other than the positivity of the “having been” intended across the negativity of the “being no longer.” Here, we have to admit, the epistemology of historiographical operation reaches its internal limit in running up against the borders of an ontology of historical being.81

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