Paper presented at
Prof. Bo StrŒthÕs seminar ÒRewriting HistoryÓ
Writing History with Film
29 November 1999

History and the Possibility of Representing the Past

A Reflection on the TV-Film Death in the Seine

by Erik TŠngerstad

At first glance, the assertion that history is about the past looks like a tautology. For what is history, if not the representation of the past? Given a second thought, however, the relationship between history and the past turns out to be highly problematical. For in everyday language Òthe pastÓ signifies something that was, but is no more. What is, is the present, not the past. And history is. History is the present understanding of how things once were. So it is not a contradiction to say that history is always in flux Ñ it is perpetually rewritten and is therefore always changing Ñ because it is, while the past was and cannot be changed.

What does this differentiation between history and the past imply? Such a question is hard to answer, because it tends to challenge our entire common-sense view of the world, as we know it. The idea that we could have knowledge of the past is more or less taken for granted. But can we really have that kind of knowledge? When asking what we can know about time also the faculty of knowledge is put into question.

Of course, questions like these are profound and philosophical, however ÑÊor exactly therefore ÑÊthey are not often openly addressed. But because science in general, and in particular historical science, is asserting that we can have knowledge of the past, it could be relevant to make this problematic explicit. Even if the relationship between ÒisÓ and ÒwasÓ is difficult to tackle, it is, naturally, neither ÒnewÓ nor ÒunknownÓ. AugustineÕs famous dictum about the nature of time, for example, encounters this specific difficulty. ÒWhat is time then?Ó Augustine asked. ÒIf nobody asks me, I know: but if I were desirous to explain it to one that should ask me, plainly I know not.Ó[1] He then speculated whether time is a categorical unity of ÒpresenceÓ, or if time is divided into the three categories of ÒpastÓ, ÒpresentÓ and ÒfutureÓ. His conclusion was that neither he, nor any man, could settle this difficulty. Instead, he thought that knowledge of the nature of time was only possible for the ever-present almighty God to capture. Since the settling of the issue lies beyond us as timely human beings, Augustine meant that we should contemplate and obey the will of untimely God. But what if there is no God to be used as reference when reflecting on the nature of time? Would that mean that the distinction between ÒpastÓ and ÒpresentÓ would collapse into chaos, and that all forms of causality would prove to be pure grammatical structure or mere illusions?

The problem is how something here-and-now can represent something there-and-then; how one thing can be about another thing when no fixed point of reference exists. What does it mean to say that history is about the past? When trying to give an answer to this question one notices a problem in conventional historiography because traditionally it is supposed that history should be able to transcend knowledge from was to is; from past to present. But is such transcendence at all possible? This problematic is also being more troublesome by the fact that ÒhistoryÓ is sometimes regarded as identical with Òthe pastÓ, and sometimes as a representation of Òthe pastÓ. For example, Òto study historyÓ can mean both Òto study the pastÓ and Òto study understandings of the pastÓ. In the former case an identity between history and the past is presumed, while in the latter one a representation is proposed. Because the terms ÒidentityÓ and ÒrepresentationÓ are not synonymous, an unsolved problem in historiography can be assumed.

The aim of this essay is to reflect on the paradoxical relationships between ÒhistoryÓ and Òthe pastÓ, and between ÒidentityÓ and ÒrepresentationÓ in historiography. The discussion will take its point of departure in the 40-minute TV-film Death in the Seine made by Peter Greenaway in 1989.[2] The reason for this choice is that though Greenaway in this film has used the kind of source material that any academic historian would accept, his general approach differs from what is accepted as historical research. This particular work of historical writing can be used both to reflect conventional historiography as it is traditionally conceived, as well as an individual attempt to write history in an alternative mode and with alternative means (video instead of printed text). GreenawayÕs film is therefore relevant because it works both as a catalyst of an immanent crisis of conventional historiography, as well as an alternative approach to writing history.

The Problem of Representation and Identity

One way of introducing the problem of representation and identity is to bring to mind a little story. It should also immediately be mentioned that this story, in the context of this essay, is central, because it holds an explicit and prominent role in the film Death in the Seine.

A crucial year in the history of photography is 1839. In January that year L. J. M. Daguerre publicly demonstrated his invention of how to fix the images produced by a camera obscura Òso that these images are not temporary reflection of the object, but their fixed and durable impress, which may be removed from the presence of those objects like a picture or an engravingÓ, as one contemporary journal reported.[3] A few weeks later, W. Henry Fox Talbot showed his Òphotogenic drawingsÓ for the Royal Institution in London. In March the British astronomer John Herschel, a friend of TalbotÕs, coined the term ÒphotographyÓ (literally Òwriting with lightÓ), thus indicating that it was possible to produce pictures of objects as they really are in themselves, and not as depictions drawn by men.[4] Shortly before, Talbot had written: ÒI made in this way a great number of representations of my house in the country [É] And this building I believe to be the first that was ever yet known to have drawn its own picture.Ó[5]

Talbot here implicitly referred to the camera obscura, with which an image of a landscape immediately can be projected on the rear wall of an almost entirely dark room. A person in the room could then fill in the image so that a picture of the landscape could be fixed. But in TalbotÕs pictures no person had drawn the image. They had Òfixated themselvesÓ. An astonishing idea is here being formulated; objects cannot only be represented by a manmade depiction, but through photography they can make representations of themselves. Because these new pictures were understood to have been created by the objects, through their radiation, their representations would be seen as virtually identical to the things they represented. But to what extent is a photographic image identical with the object it depicts? What relationship exists between the represented and its representation? These questions run parallel with the problem of the relationship between history and the past, or for that matter, between the historical text and the historical source. To what extent can reality be represented as it really is? And can the past be represented as it really was? Can there be an identity between history and the past?

In June 1839, the Parisian inventor Hippolyte Bayard publicly exhibited his work. This event could be called the worldÕs first exhibition of photographs, because Bayard had independently developed a photographic technique. In the early autumn of that year the French government bought the rights of DaguerreÕs invention, and gave it to humanity, so that anybody could use it free of charge. However, Bayard was not compensated when the French State in this way gave away his work for nothing.[6]

Immediately after the release of the Daguerrotype, as the new invention was called, people started en masse to experiment with the new technology. Meanwhile Bayard noticed that it was Daguerre that got all the fame and fortune, while nobody really noticed his own contribution. During 1840, BayardÕs bitterness rose, which eventually led to an act of public suicide. In October 1840 he published a long letter in which he accused the French government for steeling his invention and leaving him penny-less. As a consequence he had seen no alternative to drowning himself, he wrote. Together with this letter there was a photograph made by Bayard, depicting himself as a drowned man. Of course, a dead person cannot make a self-portrait, and Bayard was not dead when he wrote:

ÒThe corpse which you see here is that of M. Bayard, inventor of the process that has just been shown to you. As far as I know this indefatigable experimenter has been occupied for about three years with his discovery. (É) The Government, which has been only too generous to Monsieur Daguerre, has said it can do nothing for Monsieur Bayard, and the poor wretch has drowned himself. Oh the vagaries of human lifeÉ!Ó[7]

What can be seen in this picture, later called A Self Portrait of a Drowned Man, is a naked body placed in an upright position in a room that could be associated with a mortuary. A straw-hat is hanging on the wall, and a vase with a flower can be seen next to the man. The straw-hat as well as the vase were things that Bayard had turned into a personal signature, and the body was a self-portrait of him, playing dead.

Interestingly enough, this scene never took place in front of a camera. The picture is a composite of three other photographic depictions, all exposed in direct sun light outdoors.[8] The supposedly dead body was alive, and the supposedly tragic scene was composed in a laboratory. This picture has been called both the first fictional, as well as the first propaganda photograph.[9] It can easily be interpreted as a provocation to the very idea that a photograph should be a perfect representation of reality as it really is, an idea that lies imbedded in the term ÒphotographyÓ (as it was once coined by Herschel).

Just to give a hint of the scope of this provocation from 1840 one example can be mentioned. 150 years later, in 1990, an entry on Daguerre was published in an encyclopaedia. Here one can read:

ÒThrough the invention of the Daguerreotype, man could, in a manner more correct and true to reality, depict himself and the world.Ó[10]

Well, is that so? Can a photographic depiction be said to be a representation more Òtrue to realityÓ than any other form of construct? Can there exist an identity between a photographic depiction and a photographed object? For example, what is the relationship between a person and the photograph on his or her ID card (card of identity)? Since nobody would be said to be identical with his or her ID card, what does it mean to say that the card represents its holder? If one answer that the relationship between an ID card and its holder is a discursive construction and that the way the card represents is holder is a culturally constructed convention, a difficult and somewhat ÒnewÓ problem is raised. Can there be other forms of relations between a motif and its representation that are not discursive? Can there be any form of identity that is not a culturally constructed convention? Can we have any knowledge of anything that is not profoundly discursive? Or is all human knowledge to be regarded as a discursive construct? One way of interpreting the above quoted encyclopaedia entry is that, however relevant this problem seems to be, there still exist tendencies to uncritically take for granted an immanent and unproblematic relationship between ÒrealityÓ and Òits representationsÓ. Moreover, there are also tendencies to call such a relationship ÒidentityÓ without further reflecting on this term.

This entire problematic was provoked already 1840 by the publication of BayardÕs photography and this picture still provokes it today. Actually, the film Death in the Seine is dedicated to precisely this picture. In the same way as BayardÕs depiction puts the representation of reality into question, GreenawayÕs film can be seen as a provocation to the notion of representation and identity in conventional historiography.

Substitution or Resemblance

In his attempt to revise the theoretical foundation of conventional historiography, F. R. Ankersmit has focused on the notion of historical representation. When doing this, he differentiates between a Òsubstitution viewÓ and a Òresemblance viewÓ of representation.[11] Basically, he argues that the difference between these two views or theoretical perspectives are that from the point of view of substitution anything can be used in order to represent anything else, while from the point of view of resemblance there must be some sort of likeness between the represented thing and the thing representing it. In the substitution view, representation is not limited to physical markers. According to the resemblance view, however, there must be physical, first and foremost visual, connection between the represented and its representation. Or to take the same exaggerated example as Ankersmit uses, according to the substitution view, a fly can represent the New York World Trade Center (of course, the reverse relation is also possible). Such a form of representation would not be legitimate in the resemblance view, since a fly is not regarded as being like the World Trade Center.

A fundamental difference between these two views or theoretical perspectives is that according to the substitution view every connection between a represented thing and its representation is a construction made up by the active interpreter, while in the resemblance view it is assumed that there is some inherent ÒlikenessÓ between the presentation and its representation. And that this ÒlikenessÓ is to be discovered by a passive analyst. What is at stake here is whether there is an essential connection that can be discovered between the represented thing and its representation, or if any such relationship of representation is invented. Hence the difficulty lies in the question of the extent to which history can be said to represent something essential that exists/existed beyond itself, or whether history can only be regarded as an inner structure of constructions based on an infinite number of other such constructions. The problem with the former position is that it short-circuits the differentiation between exists and existed Ñ it undermines temporality (this is also AnkersmitÕs critique of the resemblance view). The problem with the latter position is that it ends up in a hard-line constructivism, which in turn tends to dissolve any relationship between history and the past. In other words, the issue here is to judge between the claims made by an essentialist and a constructivist perspective.

The notion of identity is here the crucial aspect, and one can here continue to follow Ankersmit line of argument. If anything could be said to be a potential representation of anything else, then anything would also be identical with anything else, and it would not be meaningful to differentiate between representation and identity. If, on the other hand, there is first something present and latter a representation of this something, and if the identity between the presented and its representation grows stronger the more the representation resembles the original presentation, then a hierarchy will exist in which the notion of identity is placed on top of the notion of representation. This could be exemplified with the above mentioned camera obscura. If the landscape is seen as the original presentation, then the image on the wall is its first representation. The closer the picture made from this representation resembles the likeness of the original landscape, the stronger the identity between the original and its representation. As a consequence not every representation would be regarded as equally ÒgoodÓ, but the stronger the identity the better the representation. Contrarily, this idea would not be supported in the substitution view, since it would not acknowledge the condition that the notion of identity is made into the reference point with which it would be possible to measure the quality of different representations. So the fundamental difference between the two views/theories is the question whether the notion of identity should be regarded as a point of reference when estimating representations, or if it instead should be seen as a construct that is produced together with each and every representation. Ankersmit writes that the Òresemblance theory inadvertently creates such a hierarchy by presenting resemblance as the link between identity and the representationÓ, indicating that the representation holds a subordinate position in comparison to what it presents. In the substitution theory this hierarchy is absent: ÒRepresentation is the representation of identity because identity only comes into being by representation and vice versa; there is not, first, an identity, which is or could be represented next Ñ whether in agreement with certain criteria of resemblance or not Ñ no, representation and identity both come into being at one and the same time.Ó(italics in the original)[12]

Because identity and representation come into being at the same time, according to Ankersmit, the substitution view excludes the idea that one can have knowledge of an original presentation beyond the immanent and mutually constructed connection of identity and representation. This view sets out that history is a construction made on the basis of an infinite series of other constructions, whereas whatever preceded this infinite series of construction cannot be settled from a position within history. Instead, following this theoretical approach, one would find an infinite number of representations throughout history, but no original presentation. Such perspective turns to undermine the notion of presence as such, breaking up the idea of a nucleus here-and-now and turning it into an infinite number of there-and-then. The reason for this is that not even presence (as such) can be understood as an original presentation.

Contrary to the resemblance view, which takes this notion of a nucleus here-and-now as its starting point, thus assuming that history is the sum of an infinite number of such here-and-now entities, the substitution view puts the entire notion of empirical research into question. The reason for this is that empirical research as a scientific practice Ñ which is generally understood as critical reflections based on sensations experienced here-and-now Ñ cannot be intellectually defended if the notion of a nucleus here-and-now is abolished. On the other hand, Ankersmit defends the substitution theory in the face of the resemblance theory, because the former Òis the theory that remains sensitive to the difference between the represented and its representation and is capable of accounting for it, whereas this distinction is lost (...) [in] the resemblance theoryÓ.[13]

But is that really the case? What Ankersmit apparently aims to demonstrate is that when, as in the resemblance theory, the supposed likeness between an original presentation and the following representation is the foundation of identity, then the claims of conventional historiography cannot be justified, because then temporality is short-circuited. The lack of connection between presentation and representation means that there is no foundation for identity as a non-constructed quality that in any ÒnaturalÓ way could connect two or more different entities with one another. But on the other hand, because Ankersmit has argued that Òrepresentation and identity always come into being at one and the same timeÓ, also the substitution view short-circuits temporality. In practice both the resemblance and the substitution view thus undermine the presupposed connection between history and the past, because both these theoretical perspectives neglect to theorise temporality. Moreover, in both the resemblance as well as in the substitution view, every identity and every representation are the results of an act of interpretative construction made by an active interpreter. For example, the only way of bridging the gap between the notion of the past and the sources at hand is to actively produce an interpretation in which the sources are understood hermeneutically (and not experienced ÒempiricallyÓ, because within the sources there would not be any inherent content to experienced) as representations of that which one wants to study. However, it seems to be clear that what Ankersmit calls ÒidentityÓ is a constructed relation between two entities. Furthermore, he writes: ÒWhen asking a historical question we want an account, a comment on the past, and not a simulacrum of the past itself. And these are not even ideally identical things.Ó(italics in the original)[14] But even if we should want to have such an account or comment on the past (and not the simulacrum), does it not always refer back to our notion of the past, as it comes down to us through historiography, and never through Òthe past itselfÓ (to use AnkersmitÕs term)?