Digital Cultural Heritage

 

The Migration of the aura through its facsimiles

(From Coover and Bartscheimer, Switching Codes)

BRUNO LATOUR AND ADAM LOWE

To Pasquale Gagliardi

Something odd has happened to Holbein’s Ambassadors at the National Gallery in London. The visitor does not immediately know how to describe her malaise. The painting is completely flat, its colors bright but somewhat garish; the shape of every object is still there but slightly exaggerated. She wonders what has happened to this favorite painting of hers. “That’s it,” she mutters. “The painting has lost its depth; the fluid dynamics of the paint has gone. It is just a surface now.” But what does this surface look like? The visitor looks around, puzzled, and then the answer dawns on her: it resembles almost exactly the poster she bought several years ago at the gallery bookshop, the poster that still hangs in her study at home. Only the dimension differs.

Could it be true? She wonders. Could they have replaced the Ambassadors with a facsimile? Maybe it’s on loan to some other museum, and so as to not disappoint the visitors, they put up this copy. Or maybe it is a projection. It’s so flat and bright it could almost be a slide projected on a screen….

Fortunately, she composes herself and does not ask the stern guard whether this most famous painting is the original or not. What a shock it would have been. Unfortunately, she knows enough about the strange customs of restorers and curators to bow to the fact that this is, indeed, the origi nal, although only in name, that the real original has been irreversibly lost, replaced by what most people like in a copy: bright colors, a shining surface, and above all a perfect resemblance to the slides sold at the bookshop, the slices shown in art classes all over the world by art teachers most often interested only in the shape and theme of a painting, not in any other marks registered in the thick surface of a work. She leaves the room suppressing a tear: the original has been turned into a copy of itself that looks like a cheap copy, and no one seems to complain, or even to notice, the substitution. They seem happy to have visited in London the original poster of Holbein’s Ambassadors!

Something even stranger happens to her, some time later, in the Salle de la Joconde in the Louvre. To get to this cult icon of the Da Vinci code, hundreds of thousands of visitors enter through two doors that are separated by a huge framed painting, Veronese’s Nozze di Cana, a dark giant of a piece that directly faces the tiny Mona Lisa, barely visible through her thick antifanatic glass. Now the visitor is really stunned. In the Hollywood machinery of the miraculous wedding, she no longer recognizes the facsimile that she had the good fortune of seeing at the end of 2007 when she was invited by the Fondazione Cini to the island of San Giorgio, in Venice. There it was, she remembers vividly, a painting on canvas, so thick and deep that you could still see the brush marks of Veronese and feel the cuts that Napoleon’s orderlies had to make in order to tear the painting from the wall, strip by strip, before rolling it like a carpet and sending it as a war booty to Paris in 1797-a cultural rape very much in the mind of all Venetians, up to this day. But there, in Palladio’s refectory, the painting (yes, it was a painting, albeit produced through the intermediary of digital techniques) had an altogether different meaning: it was mounted at a different height, one that makes sense in a dining room; it was delicately lit by the natural light of huge east and west windows so that at about 5 p.m. on a summer afternoon the light in the room exactly coincides with the light in the painting; it had, of course, no frame; and, more importantly, Palladio’s architecture merged with admirable continuity within Veronese’s painted architecture, giving this refectory of the Benedictine monks such a trompe l’oeil depth of vision that you could not stop yourself from walking slowly back and forth and up and down the room to enter deeper and deeper into the mystery of the miracle.

But here, in the Mona Lisa room, although every part of the painting looks just the same (as far as she can remember), the meaning of the painting seems entirely lost. Why does it have such a huge gilt frame? Why are there doors on both sides? Why is it hanging so low, making a mockery of the Venetian balcony on which the guests are crowded? The bride and groom, squashed into the left-hand corner, seem peripheral here, while in Venice, they were of great importance, articulating a scene of sexual intrigue that felt like a still from a film. In Paris, the composition makes less sense. Why this ugly zenithal light? Why this air-conditioned room with its dung brown polished plaster walls? In Venice, there was no air conditioning; the painting was allowed to breathe as if Veronese had just left it to dry. Anyway, here she cannot move around the painting to ponder those questions without bumping into others momentarily glued to the Joconde, their backs turned to the Veronese.

A terrible cognitive dissonance. And yet there is no doubt that this one, in Paris, is the original; no substitution has occurred, no cheating of any sort-with all its restoration, Veronese would certainly be surprised to see the painting looking as it does, but that’s different from cheating. The painting in San Giorgio, she remembers, was clearly labeled: “A facsimile.” There was even a small exhibition that explained in some detail the complex digital processes that Factum Arte, a workshop in Madrid, had used to de- then re-materialize the gigantic Parisian painting: laser-scanning it, A4 by A4, photographing it in similarly sized sections, scanning it again with white light to record the relief surface, and then somehow stitching together the digital files before instructing a purpose-built printer to deposit pigments onto a canvas carefully coated with a gesso almost identical to that used by Veronese. (Adam Lowe describes the process in an appendix to this essay.) Is it possible that the Venice version, undeniably a facsimile, is actually more original than the Paris original, she wonders? She now remembers that on the phone with a French art historian friend, she had been castigated for spending so much time in San Giorgio with the copy of the Nozze: “Why waste your time with a fake Veronese, when there are so many true ones in Venice?!” her friend had said, to which she replied, without realizing what she was saying, “But come here and see it for yourself. No description can replace seeing this original … oops, I mean, is this not the very definition of `aura’?” Without question, for her, the aura of the original had migrated from Paris to Venice: the best proof was that you had to come to the original and see it. What a dramatic contrast, she thought, between the San Giorgio Veronese and the London Ambassadors, which claims to be the original in order to hide the fact that it is now just an expensive copy of one of its cheap copies!

“But it’s just a facsimile!” How often have we heard such a retort when confronted with an otherwise perfect reproduction of a painting? No question about it, the obsession of the age is with the original. Only the original possesses an aura, this mysterious and mystical quality that no secondhand version can hope to attain. And the obsession, paradoxically, only increases as more and bet ter copies become available and accessible. If so much energy is devoted to the search for the original-for archeological and marketing reasons-it is because the possibility of making copies has never been so open-ended. If no copies of the Mona Lisa existed, would we pursue it with such energy? Would we devise so many conspiracy theories concerned with whether or not the version held under glass and protected by sophisticated alarms is the actual surface painted by Leonardo’s hand? The intensity of the search for the original, it would seem, depends on the amount of passion triggered by its copies. No copies, no original. To stamp a piece with the mark of originality requires the huge pressure that only a great number of reproductions can provide.

So, in spite of the knee-jerk reaction-“But this is just a facsimile”-we should refuse to decide too quickly the value of either the original or its reproduction. The real phenomenon to be accounted for is not the delineation of one version from all others but the whole assemblage of one-or several- original(s) together with its continually rewritten biography. Is it not because the Nile ends in such a huge delta that the century-long search for its sources was so thrilling? To pursue the metaphor, we want, in this essay, to attend to the whole hydrological complex of a river, not just its elusive origin. A given work of art should be compared not to any isolated spring but to a catchment area, a river along with its estuaries, its tributaries, its rapids, its meanders, and, of course, its hidden sources.

To give a name to this watershed, we will use the word trajectory. A work of art-no matter the material of which it is made-has a trajectory or, to use another expression popularized by anthropologists, a career (Appadurai 1986; Tamen 2001). What we will attempt to do in this essay is to specify the trajectory or career of a work of art and to move from one question that we find moot (Is it an original or merely a copy?) to another that we take to be decisive, especially in a time of digital reproduction: Is it well or badly reproduced? This second question is important because the quality, conservation, continuation, sustenance, and appropriation of the original depend entirely on the distinction between good and bad reproduction. A badly reproduced original, we will argue, risks disappearing, while a well-copied original may enhance its originality and continue to trigger new copies. Facsimiles, especially those relying on complex (digital) techniques, are thus the most fruitful way to explore the original and even to redefine what originality is.

To help shift the attention of the reader from the detection of the original to the quality of its reproduction, let us remember that the word “copy” need not be derogatory; indeed, it comes from the same root as “copious,” and thus designates a source of abundance. A copy, then, is simply a proof of fecundity. So much so that, to give a first shape to the abstract notion of the trajectory, we will call upon the antique emblem of the cornucopia, a twisted goat horn with a sharp end-the origin-and a wide mouth disgorging an endless flow of riches (all thanks to Zeus). This association should come as no surprise: to be original means necessarily to be the origin of a lineage. That which has no progeny, no heirs, is called, not original, but sterile, barren. Rather than ask, is this isolated piece an original or a facsimile? we might, then, inquire, Is this segment in the trajectory of the work of art barren or fertile?

To say that a work of art grows in originality in proportion to the quality and abundance of its copies is nothing odd: this is true of the trajectory of any set of interpretations. If the songs of the Iliad had remained stuck in one little village of Asia Minor, Homer would not be considered a (collective) author of such great originality. It is because-and not in spite-of the thousands and thousands of repetitions and variations of the songs that, when considering any copy of the Iliad, we are moved so much by the unlimited fecundity of the original. We attribute to the author (although his very existence cannot be specified) the power of each of the successive reinterpretations by saying that “potentially” all of them “were already” there in the Ur-text-which we simultaneously know to be wrong (my reading could not possibly already have been there in ancient Greece) and perfectly right, since I willingly add my little extension to the “unlimited” fecundity of this collective phenomenon called “the” Iliad. If it is indeed unlimited, it is because I and others continue to push the limit. To penetrate the poem’s “inherent” greatness, you need to bring with you all of the successive versions, adaptations, and accommodations. Nothing is more ordinary than this mechanism: Abraham is the father of a people “as numerous as the grains of sand” (and of three religions) only because he had a lineage. Before the birth of Isaac, he was a despised, barren old man. Such is the “awesome responsibility” of the reader, as Charles Peguy so eloquently said, because this process is entirely reversible: “If we stop interpreting, if we stop rehearsing, if we stop reproducing, the very existence of the original is at stake. It might stop having abundant copies and slowly disappear.”‘

We have no difficulty raising questions about the quality of the entire trajectory when dealing with dance, music, theater-the performing arts. Why is it so difficult when faced with the reproduction of a painting, a piece of furniture, a building, or a sculpture? This is the first question we want to clarify.

No one will complain on seeing a performance of King Lear, “But this is not the original!” The whole idea of “playing” Lear is to replay it. Everyone is ready to take into account the work’s trajectory, from the first presentations through the latest “revival.” There is nothing extraordinary in considering any given staging as a moment, a segment, in the career of the work of art called King Lear. No one has ever seen and no one can ever circumscribe the absolute Platonic ideal of the play. One will never see the original, as presented by Shakespeare himself, nor even the original text, but only several premieres and several dozen written versions with endless glosses and variations. As with a river, we seem perfectly happy with the anticlimactic discovery of the source of a major river in a humble spring, barely visible under the mossy grass. Even more important, spectators have no qualm whatsoever at judging a new production of the play, and the standard is clear: Is it well or badly (re) played? People may differ wildly in their opinions, some scandalized by novelties (Why does Lear disappear in a submarine?), some bored by cliched repetition, but they have no difficulty in considering that this moment in the career of all the successive King Lears-plural-should be judged on its merit and not by its mimetic comparison with the first (now inaccessible) presentation. It is what we see now, with our own eyes, on stage that counts in making our judgment, and not its resemblance to some distant, hidden Ur-event (though what we take to be the “real” King Lear remains in the background of every one of our judgments). So clearly, in the case of the performing arts at least, every new version runs the risk of losing the original-or of regaining it.

So unconstrained are we by the notion of an original that it is perfectly acceptable to evaluate a performance by saying, “I would never have anticipated this. It is totally different from the way it has been played before, utterly distinct from the way Shakespeare played it, and yet I now understand better what the play has always been about!” Some revivals-the good ones-seem to extract from the original latent traits that can now (or again) be made vivid in the minds of the spectators. The genius of Shakespeare, his originality, is thus magnified by this faithful (but not mimetic) reproduction. The origin is there anew, even if vastly different from what it was. And the same phenomenon would occur for any piece of music or dance. The exclamation “It’s so original,” attributed to a new performance, does not describe one section of its trajectory (and especially not the first Ur-version) but the degree of fecundity of the whole cornucopia. In the performing arts, the aura keeps migrating and might very well come back suddenly-or disappear altogether. Too many bad repetitions may so decrease the fecundity of a work that the original itself will be abandoned; there will be no further succession. Such a work of art dies out like a family line without offspring. Like a river deprived of its tributaries, one by one until it has shrunk to a tiny rivulet, the work is reduced to its “original” size, that is, to very little. It has then lost its aura for good.

Why is it so difficult to apply the same type of judgment to a painting or a sculpture or a building? Why not say, for instance, that Veronese’s Nozze di Cana has been reuiued thanks to a new interpretation in Venice in 2007 by Factum Arte, much as Hector Berlioz’s Les Troyens had been presented at last for the first time by Colin Davis in 1969 in Covent Garden in London (poor Berlioz had never had the money or the orchestra to play his original work in full). And yet, to claim that the Nozze di Cana has been “given again” in San Giorgio seems far-fetched. “That’s just a facsimile,” someone will immediately object. “The original is in Paris!” The intimation of fakery, counterfeiting, or betrayal would seem absurd for a piece of theater (though one might term a very bad production a sham or a travesty). Still, it seems almost impossible to say that the facsimile of Veronese’s Nozze di Cana is not about falsification but instead a stage in the verification of Veronese’s achievement, a part of the painting’s ongoing biography.

One reason for this unequal treatment is what could be called the differential of resistance among the segments of the trajectory of a visual artwork. In his much too famous essay, amid a deep fog of art-historical mysticism, it is this differential that Walter Benjamin (1968) pointed out under the name “mechanical reproduction.” In the case of the performing arts, each version is just as difficult to produce, and at least as costly, as the first. It is not because there have been zillions of presentations of King Lear that the one you are about to give will be easier to fund. The marginal cost will be exactly the same-except that the public already knows what “a King Lear” is and comes fully equipped with endless presuppositions and critical tests concerning how it should be played (a double-edged sword, as any director knows). This is why, in the performing arts, we don’t distinguish between an original and a copy but rather between successive versions of the same play: version n, version n + 1, and so on. It is also why the “real” King Lear cannot be localized (not even at the very beginning); rather, the name is given to the whole cornucopia (though each spectator may cherish special moments in his or her personal history when, thanks to exceptionally good “revivals,” the genius of Lear was “instantiated” more fully than at any time before or later). The trajectory of a performance, then, is composed of segments that are made, more or less, of the same stuff or require a similar mobilization of resources.

The situation appears to be entirely different when considering, for instance, a painting. Because it remains in the same frame, encoded in the same pigments, entrusted to the same institution, one cannot resist the impression that a reproduction is much easier to make and thus there can be no comparison, in terms of quality, between the various segments of its trajectory. The aura seems firmly attached to one version only: the autograph one. And certainly this is superficially true. No one in his right mind would deem your digital snapshot of the Nozze di Cana, palely rendered on the screen of a computer, commensurable with the sixty-seven square meters of canvas in the Louvre. If you claimed that your picture was “just as good as the original,” people would shrug their shoulders in pity, and rightly so.

But it is not because of some inherent quality of painting that we tend to create such a yawning gap between originals and copies, nor because paintings are more “material” (an opera or a play is just as “material” as pigments on canvas). The distance between version n (“the original”) and version n + i (“a mere copy”) depends just as much on the differential of efforts, costs, and techniques as on any such substantial distinction. While in performance art, the resources required are grossly homogeneous (each replay relying on the same gamut of techniques), the career of a painting or a sculpture consists of segments that are vastly heterogeneous, varying greatly in the intensity of the efforts entailed. It is this asymmetry, we wish to argue, that precludes one from saying that the Nozze di Cana in Paris has been “reprinted” or “given again” in Venice. And it is certainly this presupposition that so angered the French art historian who castigated her friend for wasting her time in San Giorgio. Hidden behind the commonsense distinction between original and copy lies a totally different process that has to do with the technical equipment, the amount of care, and the intensity of the search for the originality that goes from one version to the next. Before it can defend the quality of its reenactment, a facsimile is already discredited because of a perceived gap associated with the techniques of reproduction, a gap based on a misunderstanding of photography as an index of reality.

The proof of this claim can be obtained by showing what happens to our search for originality when we modify this differential-something that becomes easier and easier in the new digital age. Consider, by way of comparison, the copying of manuscripts. Before printing, the marginal cost of producing one more copy was identical to that of producing the penultimate one (a situation to which we are actually returning now with digital copies). Inside the scriptorium of a monastery, all exemplars were facsimiles. No copy ist would have said, this is the original, that a mere copy; distinctions were instead based on quality. Here again, the aura was able to travel and might very well have migrated to the newest and latest copy, excellently done on the best of parchments and double-checked against the best earlier sources. Following the invention of the printing press, however, the marginal cost of one extra copy became negligible compared to the time and techniques necessary to write the manuscript. An enormous distance was introduced, and rightly so, between segments of the trajectory: the autograph manuscript became the original, while the print run, from that moment on, consisted of mere copies (until of course the great art of bibliophily revealed endless subtle differences among successive printings and forensic digital analysis allowed us to date and order those copies).

There is no better proof that the ability of the aura to be retrieved from the flow of copies (or to remain stuck in one segment of the trajectory) crucially depends on the heterogeneity of the techniques used in the successive segments, than to consider what happens to the original now that we are all inside that worldwide cut-and-paste scriptorium called the web. Because there is no longer any huge difference between the techniques used for each successive instantiation of some segment of a hypertext, we accept quite easily that no great distinction can be made between one version and those that follow. We happily label successive renderings of the “same” argument version i, version 2, version n, while the notion of the author has become just as fuzzy as that of the aura-not to mention what happens to copyright royalties. Hence the popularity of collective scriptoria like Wikipedia. In effect, Benjamin confused the notion of “mechanical reproduction” with the inequality in the techniques employed along a trajectory. No matter how mechanical a reproduction is, once there is no huge gap in the process of production between version n and version n + i, the clear-cut distinction between the original and its reproduction becomes less crucial-and the aura begins to hesitate and is uncertain where it should land.

All of that might be very well, but is it possible to imagine the migration of the aura in the reproduction or reinterpretation of, say, a painting? After all, it is the contrast between the Nozze and the Ambassadors that triggered our inquiry, which would have proceeded very differently had it been limited to the performing arts. One cannot help suspecting that there is in painting, in architecture, in sculpture, in objects in general, a sort of stubborn persistence that makes the association of a place, an original, and some aura impossible to shed.

Let us first notice, however, that the difference between performing arts and the others is not as radical as it seems: a painting has always to be reproduced, that is, it is always a re-production of itself, even when it appears to stay exactly the same in the same place. That is to say, no painting remains the same in the same place without some reproduction. For paintings, too, existence precedes essence. To have a continuing substance they need to be able to subsist. This requirement is well known by curators all over the world: a painting has to be refrained, dusted, relit, sometimes restored, and it has to be re-presented, in different rooms, with different accompanying pictures, on different walls, inserted in different narratives, with different catalogs, and with changes in its insurance value and price. Even a painting that is never loaned, that survives within the same institutional setting without undergoing any heavy restoration, has a career; to subsist and be visible again, it needs to be taken care of. If it is not cared for, it will soon be accumulating dust in a basement, be sold for nothing, or be cut into pieces and irremediably lost. Such is the justification for all restorations: if you don’t do something, time will eat up that painting, as certainly as the building in which it is housed will decay and the institution charged with its care will decline. If you doubt this, imagine your precious works of art housed in the National Museum in Kabul. For a work of art to survive, it requires an ecology just as complex as that needed to maintain the natural character of a park (Western 1997).

If the necessity of reproduction is accepted, then we might be able to convince the reader that the interesting question is not so much how to differentiate the original from the facsimiles as how to tell the good reproductions from the bad. If the Ambassadors has been irreversibly erased, it is not out of negligence but, on the contrary, because of an excess of zeal in “reproducing” it. What the curators did was to confuse the obvious general feature of all works of art-to survive they have to be somehow reproduced-with the narrow notion of reproduction provided by photographic posters while ignoring many other ways for a painting to be reproduced. For instance, they could have had a perfect facsimile made, registering all the painting’s surface effects in three dimensions, then restored the copy instead of the work itself. If they had done this, they could have invited several art historians with different views to suggest different ways of restoring the copy and produced an exhibition of the results. Their crime is not to have offered a reproduction of the Holbein instead of the Holbein itself to the visitors of the National Gallery-“the Ambassadors” remains behind all successive restorations much as “King Lear” does over each of its replays, granting or withdrawing its auratic dimension depending on the merit of each instance-but to have so limited the range of reproduction techniques that they chose one of the most barren: the photograph-as if a painting were not a thick material but some ethereal design that could be lifted out of its materiality and downloaded into any reproduction without any loss of substance. In fact, a terribly revealing documentary shows the culprits restoring the Holbein, using as their model photographs, subjectively deciding what was original, what had decayed, what had been added, and imagining the painting as a series of discrete layers that could be added or removed at will-a process that resembles plastic surgery more than an open forensic investigation.

What is so extraordinary in comparing the fate of the Ambassadors with that of the Nozze di Cana is not that both rely on reproduction-this is a necessity of existence-but that the first relies on a notion of reproduction that makes the original disappear forever while the second adds originality without jeopardizing the earlier version-without ever even touching it, thanks to the delicate processes used to record it.

But, one might ask, how can originality be added? One obvious answer is: by bringing the new version back to the original location. The cognitive dissonance experienced by the visitor in the Mona Lisa room comes in part from the fact that in Palladio’s refectory every detail of the Nozze has a meaning that is entirely lost and wasted in the awkward situation provided for the version n – i in Paris. In other words, originality does not come to a work of art in bulk; it is made of different components, each of which can be interrelated to produce a complex whole. New processes of reproduction allow us to see these elements and their interrelationship in new ways. To be at the place for which it had been conceived in each and every detail is certainly to observe one aspect-one element-in what we mean by an original. On that ground, there is no question that the facsimile of the Nozze is now the original, while the version in the Louvre has lost at least this comparative advantage.

We should not, however, wax too mystical about the notion of an “original location” in the case of the Veronese, since the very refectory in which the facsimile has been housed is itself a reconstruction. If you look at photographs taken in i95o, you will notice that the original floor was gone and another had been installed at the height of the windows. The top was a theater and the basement a wood workshop-the whole space had been altered. It was rebuilt in the 195os, but the plaster and floor were wrong and the boiserie that surrounded the room and added the finishing touches to the proportion of the room was missing. In its stripped-down state, it looked more like a high Protestant space that seemed almost to laugh at the absence of Veronese’s Counter-Reformation flourish. But now there are rumors that the return of the painting-the facsimile, that is-has triggered a plan for a new restoration that will retrospectively return the space to its former glory. A facsimile of a heavily restored original, now in a new location, may cause new elements to be added to an original in its original location that is, in part, a facsimile of itself. Originality once seemed so simple …

The same is certainly true of availability. What angered the visitor to the Louvre so much was that she could not visually scan the Nozze without bumping into Mona Lisa addicts. The Veronese is so full of incident and detail that it cannot be seen without time to contemplate its meaning, its implications, and the reasons for its continued importance. What does it mean to enshrine an original, if the contemplation of its auratic quality is impossible? This, too, is an element that can be prized away and distinguished from all the others. This component of originality need not be bound to the originality of the location; the best proof of this may lie in the facsimile of the burial chamber from the tomb of Thutmosis III in the Valley of the Kings.2 It contains the first complete text of the Amduat to be used in a pharaonic tomb. The Amduat is a complex narrative mixing art, poetry, science, and religion to provide a coherent account of life in the afterworld. The tomb was never made to be visited, and the physical and climatic conditions inside it are incompatible with mass tourism. As a result, the tomb is deterioriating rapidly and glass panels have had to be installed to protect the walls from accidental damage and wear and tear. These interventions, however, change the nature of the space and inhibit both detailed study of the text and an appreciation of the specific character of the place. Exhibitions that present the facsimile and contextualize the text have now been visited by millions of people in North America and Europe. The delocalized facsimile has thus established the reasons for its continued importance, turned its visitors into a proactive force in the conservation of the tomb, and possibly become part of a long-term policy that will keep the version n – i safe but accessible to the small number of specialists who require access for continued study and monitoring. See? Each of the components that together comprise what we mean by a true original begins traveling at a different speed along the trajectory and begins to map out what we have called the catchment area of a work of art.

A third element of originality has to do with the surface features of a work. Too often, restorers make a mockery of the materiality of the original they claim to be protecting by attending only to two-dimensional shape. If there is one aspect of reproduction that digital techniques have entirely modified, it is certainly the ability to register the most minute three-dimensional aspect of a work without putting the work at risk. It is often forgotten that in its early years the British Museum used to take plaster casts of its objects; indeed, the first British Museum catalog contained a list of copies that were available and for sale. It is often forgotten because the plaster-cast collection was discarded at the end of the twentieth century, and valuable information about the surfaces of the works when they entered the museum was lost. Subsequent restorations have dramatically altered the surface and appearance of many of the objects. But beyond evidence of their earlier forms, many of the molds still contained paint that had been removed during the casting process. So, even for a work of art to be material is a question of complex trajectories. Many Venetians, when they first heard of the Nozze facsimile, conjured up images of a glossy flat surface, like that of a poster, and were horrified at the idea of this being given in reparation for Napoleon’s cultural rape of San Giorgio. They could scarcely have anticipated that the facsimile would consist of pigment on a canvas coated with gesso, “just like” Veronese had used. When it was unveiled, there was a moment of silence, then ecstatic applause and many tears. Many Venetians had to ask themselves a very difficult question: How is it possible to have an aesthetic and emotional response in front of a copy? This question is followed by another: how do we stop Venice from being flooded with bad copies without the criteria to distinguish between good and bad transformations?

Once again, digital techniques allow us to distinguish features that are being regrouped much too quickly under the generic term “reproduction.” As we have seen, exactly the same intellectual oversimplifications and category mistakes happened when Benjamin wrote about “mechanical reproduction.” Surely the issue is about accuracy, understanding, and respect-the absence of which results in “slavish” replication. The same digital techniques may be used either slavishly or originally. It depends again on which features one chooses to bring into focus and which one leaves out. The use of tiny painted dots based on photographs rather than the broader brush marks used to make the original may give the restorer more control and hide the interventions, but surely it proves that a manual reproduction might be infinitely more disputable and subjective than a “mechanical” one. The road to hell is paved with good intentions.

No doubt, it is an uphill battle: facsimiles have a bad reputation-people assimilate them with a photographic rendering of the original-and “digital” is associated with an increase in virtuality. So, when we speak of digital facsimiles, we are looking for trouble. And yet we claim that, contrary to common presuppositions, digital facsimiles are introducing many new twists into the centuries-old trajectory of works of art. There is nothing inherently “virtual” in digital techniques-and for that matter, there is nothing entirely digital in digital computers!’ The association of the digital with the virtual is entirely due to the bad habits associated with just one of its possible displays: the pretty bad screens of our computers. Things are entirely different when digital techniques are only one moment in the move from one material entity-Veronese’s Nozze version n – i in the Louvre-to another equally material entity-version n + i in San Giorgio. During this time of mass tourism and increasingly vocal campaigns for the repatriation of the spoils of wars or commerce, when so many restorations are akin to iconoclasm, when the sheer number of amateurs threatens to destroy even the sturdier pieces in the best institutions, it does not require excessive foresight to maintain that digital facsimiles offer a remarkable new handle to give to the notion of originality what is required by the new time. Since all originals have to be reproduced anyway, simply to survive, it is crucial to be able to discriminate between good and bad reproductions.

Appendix. The process used to create an accurate facsimile of Le Nozze di Cana by Paolo Caliari (called Veronese)

Adam Lowe

In the autumn of 2006, the Musee du Louvre reached an agreement with Fondazione Giorgio Cini and granted Factum Arte4 access to record Veronese’s vast painting Le Nozze di Cana. The conditions were carefully specified: the recording must entail no physical contact with the painting, all equipment must meet the highest conservation standards and be approved prior to use in the museum, no external lighting or scaffolding could be used, work could only happen when the museum was closed, and no equipment could be left in the room when the public was present. In defining each condition the safety of the painting (and the other paintings in the room) was always the decisive factor.

To record this 67.29-square-meter painting at actual size and at the highest possible resolution, Factum Arte built a non-contact color scanning sys tem that uses a large format CCD and integrated LED lights. The system records at a scale of i s i at a maximum resolution of i,2oo dots per inch (dpi). The scanning unit is mounted on a telescopic mast, which is operated by an air pump and can accurately position the scanning unit on the vertical axis. The mast has a maximum reach of 8 meters from the ground, has an ultrasonic distance sensor to ensure that the scanning head is parallel to and a uniform distance (8 centimeters) from the picture surface, and is fitted with a linear guide to position the scanning head laterally in front of the painting. This is essential to ensure that the numerous scan files can be fitted together without scale, focus, or perspective distortion.

FIGURE 31. Time-lapse photgraph of Factum Arte’s team recording the original but heavily restored painting in the Louvre. A telescopic mast and flatbed scanning system was used to record the entire surface of the painting. A 3-D scanning system was used to record the surface of the lower part of the painting. Photo: Gregoire Dupond.

The scanning head moves over the surface illuminating the area that is being recorded (figure 31). The LED light used contains no ultraviolet rays and generates minimum heat. The painting was scanned in thirty-seven columns and forty-three rows. Each capture was 22 by 30.5 centimeters with an overlap of 4 centimeters on the horizontal dimension and 7 centimeters on the vertical dimension. Each file was saved in two formats, TIFF (the working file) and JPEG (a reference file). The recording was done at 60o dpi with i6-bit color depth. During the recording of Le Nozze di Cana, 1,591 individual files were saved in TIFF format, resulting in an archive of 400 gigabytes.

FIGURE 32. Naoko Fukumaru, Factum Arte’s chief conservator, working in the Louvre with a selection of pre-prepared color sticks that are matched to the surface of the painting. Equal attention was paid to the color and the reflectivity of the surface. Photo: Gregoire Dupond.

The telescopic mast was also used for conventional photography using a Phase One H25 digital back fitted to a medium-format Hasselblad body. The Phase One records 5,488 by 4,145 pixels (22 megapixels) with a pixel size of 9 by 9 microns and 48-bit color (16 bits per channel). The photography was done in 450 sections (18 columns and 25 rows) using the ambient light in the room. For reference, a photograph of the complete painting was taken with the camera positioned on a tripod at the other end of the room, in front of the Mona Lisa, to ensure minimum distortion. The archive of photographic data consists of 593 different files totaling 41 gigabytes of data.

The lower part of the painting was recorded using a noncontact 3-D scanning system made by NUB 3D (Spain). No markers, spheres, or registration systems are fixed to the object. The average working distance is about a meter away from the surface being recorded. The NUB 3D Triple White Light Scanning System uses a mix of optical technology, 3-D topometry, and digital image processing to extract three-dimensional coordinates from the surface of an object. This technique, known as structured white-light triangulation, produces accurate measurements of the surface by analyzing the deformation in lines and pat terns of light that are projected onto the surface of an object. Multiple images are captured by an integrated camera in the measuring head, and using these the system’s integrated technology calculates a coordinated x, y, z point cloud relating to the surface of the object. About io square meters of the painting were recorded in 3-D at a resolution of between 40o and 700 microns. The scanning was done in sections of one meter square, generating an archive of about i gigabyte. Due to the tonal difference of the painted surface and the varying reflectivity of the varnish a multiexposure option was used. For the alignment and post-processing Invometric Polyworks software was used.

During the recording extensive color notes were made using a series of color sticks made on site and matched to specific points on the surface of the painting (figure 32). Bits of the color sticks were cut off and fixed into a book containing a i:i scale line drawing of the painting at locations corresponding to these points on the painting.

The first step in assembling and aligning the data, carried out while still working in the Musee du Louvre, was to preassemble all the columns using Photoshop Scripting. This resulted in a roughly assembled image of the entire painting. The final assembly, carried out in Madrid, reduced the 1,591 individual files into 185 larger units (again with overlaps). The vertical columns were used as the basic unit. Strips comprising eight or nine scans each were accurately assembled; each full column was made of five of these files. The file size of each strips is about i gigabyte.

The edges of the painting were accurately assembled in order to give an absolute reference. Three horizontal lines were established across the painting to ensure that no distortion was taking place. Four vertical lines were then stitched to these horizontals and the resulting blocks were filled in the same way, subdividing each one into horizontal and vertical areas. This avoided any compound distortion and ensured an accurate master file. The master file was broken down into manageable units (file size under 12 gigabytes) with an accurate reference to each adjoining unit. Once these units were finalized the individual scans were flattened. The painting was then divided into I-by-2-meter blocks, which served as the printing units. There are forty-four printing units.

The scanner data and the color photographic data had to be treated independently but aligned in perfect registration. The photographic data (recorded in RGB, with a color depth of 16 bits per channel) had to be mapped onto the I-by-2-meter scanner data units. The scanner data contains no distortion, but camera data always contains some lens and perspective distortion. To join these two types of information, each photograph had to be generally distorted and then locally transformed, using transparency in different blending modes depending on the nature of the data. This was mainly done using features in the picture such as brush marks or “noise” within the canvas.

FIGURE 33. During the production of the facsimile the color references recorded in the museum were essential to ensuring the accuracy of the final print. Photo: Alicia Guirao.

When more then one data set is used, the usual difficulties of color printing increase exponentially. The aim of the color adjustments was to make Factum Arte’s flatbed printer match the color sticks recorded in the Louvre (figure 33). The monitors were all calibrated to the colors that were to be printed onto the gesso-coated canvas. The color of the gesso is not a pure white so it was important to create a range from the lightest white to the darkest black. The exact matching of all the colors was then a question of trial and error, involving changes to both the digital files and the gesso mix. Printing began with the panel at the bottom left of the painting, and further changes were made until both the tone and the color of the two layers (scanner and photographic data) printed together matched the color sticks after the print had been varnished. Every change was archived and then simplified gradually resulting in a series of Photoshop “actions.” One set of actions was applied to the Phase One data and another to the Cana Scanner data. Once finalized, these actions were applied to all forty-four printing files, and small versions of each file were printed. During the recording the lighting conditions for both photography and scanning had been kept constant so these universal actions, in theory, would result in accurate color matching across all parts of the 67-square-meter canvas.

FIGURE 34. The flatbed printer specially designed by Factum Arte’s engineer Dwight Perry. Each section of the painting was printed several times in perfect registration in order to ensure correct tone and color. Photo: Alicia Guirao.

After printing the tests and comparing them with the color sticks, further local corrections were made to specific colors, mainly in areas with a lot of whites or with complex grays and blacks. These changes were made using locally applied masks to isolate specific areas of the painting.

The facsimile was printed on Factum Arte”s purpose-built flatbed printer (figure 34). This is based on an Epson Pro 9600 digital printer. The printer uses pigment inks in seven colors (cyan, light cyan, magenta, light magenta, yellow, black and light black). The bed is fixed, and the print heads move up and down the bed on linear guides. The movement of the heads is accurate to a few microns, and their height can be adjusted during printing. This made it possible to print the image in pigment onto gesso-coated canvas. The gesso coating, a mixture of animal glue, calcium carbonate, and precipitated chalk, used no metal oxides. The texture on the surface of the 16-ounce Irish flax was made from flax fibers and threads mixed with gesso. Due to its history Le Nozze di Cana has a complex and unusual surface. To reproduce this appearance, each piece of canvas was coated with a layer of animal glue, a layer of gesso and fibers, and then two layers of gesso. Acetate sheets printed with the Phase One photographic data were used, with a pin registration system, to ensure accurate placement of the texture on the prepared canvas. Each panel was then printed twice in perfect register. The first layer to be printed was the information recorded on the Phase One photographs. The second layer was the scanner data. The overprinting resulted in accurate color matching and a control of the tonal values of the painting. The entire image was divided into printing files, with io centimetres of overlap. The printed panels were varnished with a satin Golden acrylic varnish with UVLS (an ultraviolet filter).

FIGURE 35. The printed sections of the facsimile were laid out on the floor of the workshop, carefully spliced together, and glued to Alucore panels. Photo: Alicia Guirao.

FIGURE 36. A detail of the finished surface after joining but before retouching. The painting was printed onto prepared canvas in sections that were spliced together and retouched. The white line in this image is a filled join that has not yet been retouched. The horizontal join is the result of heavy restoration to repair the painting after it was cut from the wall by Napoleon’s troops. Photo: Alicia Guirao.

FIGURE 37. After retouching the spliced join is no longer visible. The retouching was carried out by Factum Arte’s conservation team in Madrid, and the finishing touches were done on-site in Venice when the painting was installed. Photo: Alicia Guirao.

F I G U R E 38. Naoko Fukumaru retouching a section of the painting in Factum Arte’s workshop in Madrid. Photo: Gregoire Dupond.

FIGURE 39. The completed facsimile installed in Palladio’s refectory at the Fondazione Giorgio Cini, Venice. Photo: Gregoire Dupond.

Ten 20-millimeter-thick Alucore panels were made, each 34o by 205.2 centimeters. When perfectly aligned in two rows of five panels they make up the whole painting. Each of the large aluminum panels was assembled from six printed canvas panels, with some of the panel overlapping the edges (figure 35). First the printed panels were laid out on the surface and perfectly aligned. They were then spliced together with irregular cuts that followed features in the painting. Straight lines were always avoided. The canvas was then fixed to the aluminum with PVA. The joins resulting from the stitching of the printed panels were then retouched by hand by a team of trained conservators.The joins were first filled with a mixture of Golden acrylic molding paste and glass microspheres, then tinted, and finally retouched using Golden acrylic paint (figures 36-38). During this process a gloss Golden varnish was used. A Golden gel was then used for the final texture and to enhance selected brushmarks and areas of impasto paint. A final coat of satin Golden varnish with UVLS was applied.

The honeycomb aluminum panels with the retouched printing were sent to Venice, where they were fixed on-site to an aluminum frame. This was done in two parts, which were then lifted into place and fitted onto the wall. The final filling, retouching, and control of the surface was done when the facsimile was in its final position and with the lighting that exists in the refectory (figure 39).

Acknowledgments

We thank the participants at the dialogue held in Venice in San Giorgio on “Inheriting the Past” for many useful conversations and especially the director of the Cini Foundation, Pasquale Gagliardi.

Notes

i. See the commentaries of Peguy in Deleuze (2005).

2. The facsimile of the tomb (in its current condition but without the elements that turn the environment into a museum) has resulted in detailed publications, in both film and book form, by the Egyptologist Erik Hornung and the psychologist Theodor Abt (Hornung and Abt 2005; Hornung et al. 2006).

3. See Lowe and Schaffer (2000); Smith (2003).

4. See http://www.factum-arte.com.

Works Cited

Appadurai, Arjun, ed. 1986. The social life of things: Commodities in cultural perspective. Cambridge: Cambridge University Press.

Benjamin, Walter. 1968. The work of art in the age of mechanical reproduction. In Illuminations, 217-51. New York: Schocken Books.

Deleuze, Gilles. 2005. Difference and repetition. Trans. Paul Patton. New York: Continuum International.

Hornung, Erik, and Theodor Abt. 2005. The dark hours of the sun: The Amduat in the tomb of Thutmose III. DVD. Madrid: Factum Arte.

Hornung, Erik, et al. 2006. Immortal pharaoh: The tomb of Thutmose III. Madrid: Factum Arte.

Lowe, Adam, and Simon Schaffer. 2000. Noise, 2000. Catalog for exhibition held simultaneously at Kettle’s Yard, the Whipple Museum of the History of Science, and the Museum of Archaeology and Anthropology, all in Cambridge, and the Wellcome Institute, London. Cambridge: Kettle’s Yard.

Smith, Brian Cantwell. 2003. Digital abstraction and concrete reality. In Impressiones. Madrid: Calcografia Nacional.

Tamen, Miguel. 2001. Friends of interpretable objects. Cambridge, MA: Harvard University Press.

Western, David. 1997. In the dust of Kilimanjaro. New York: Shearwater/Island Press.