Deep Nostalgia, historical distance and the uncanny of the past
How a new app shows the troublesome quality of our identification with the past
In 1987, Zbigniew Rybczyński released Steps, an animated short that combines live actors with manipulated stock images that is as curious as penetrating about the specific cultural moment it was produced. In the movie, a group of North American tourists is called to try the latest invention of a Soviet-based scientist cum trickster. The experiment consists in transporting a group of people from the outside to the inside of a historical record; in this case, the famous scene of Sergei Eisenstein’s Battleship Potemkin (1925) at the Odessa steps. The group is a collection of stereotypical tropes of the North American popular culture of its time. There’s a gunslinging conservative cowboy; an African-American with cheerful clothes whose main preoccupation is guaranteeing the safety of his boombox (this is certainly the most problematic representation in the movie); a pair of obtuse White American tourists craving for the exotic and who consider everything amusing; a bored platinum blonde trophy wife concerned only with her looks, annoyed by the events happening around her. These people were inserted into the footage and respond to what happens in it, but cannot alter the course of the events and, in the end, cannot be really moved by it besides their excitement as participants in the middle of it. The one exception is the scientist himself, who appears as a black and white Russian Imperial soldier at the end of the movie. Ultimately, he is trapped inside his own invention.
Steps is full of the wit and disenchantment of late-1980s Eastern European cultural products, although Rybczyński already resided in the US at the time. It mocks both mainstream American culture and Soviet Union, assigning it quite literally to the past through the fate of the scientist behind the experiment. It also tackles the surging industry of historical tourism and the “Disneyfication” of the past. It hides, however, a deeper true: our access to the past is given through different mediatic products and shows the sign of these different media registers; by depending on this, it becomes possible to conflate the mediatized historical record with the actual historical process.
Steps came to my mind as I’ve pondered over the reactions generated by a new app, My Heritage’s Deep Nostalgia. Unlike other image manipulation apps previously available, deeply embedded both with facial recognition technology and with the transformations of the individual’s identity, such as swapping her/his gender or getting older, Deep Nostalgia makes it possible to animate old photographs, trying the recreate the semblance of movement and lifelikeness of long-deceased relatives or historical figures. On the occasion of the animation of a picture of a young W. E. B. Du Bois, I saw a comment that tried to sum up what the app provided: even if we don’t know for sure how he did it, it is as if we were looking at what he saw — that is, ourselves.
The reactions to Deep Nostalgia are a mix of amusement regarding the sudden coming to life of deceased figures and the uncanniness these images bring. Indeed, “uncanny” and “creepy” seems to be the most common reactions to the app, as least from what I could gather from the its usage to animate pictures of Mark Twain, Frederick Douglass, Du Bois or Ludwig Wittgenstein, among others. Uncanny certainly they are, as those faces’ movements show always the same lack of expression, the blankness in their eyes, and the movements of the heads from right to left seem far from natural. These characters look like fully-grown adults suddenly coming to life as newborn babies, acknowledging the world for the first time. It looks like someone being reanimated after a long cryogenic suspension, still unable to utter a phrase or articulate speech — the images, is worth remembering, are GIFs and as such do not accompany the recreation of the voice of these subjects. This is as uncanny as facing AI itself.
As uncanny as these images are, they aren’t however unheard-of. Regarding the technological reconstruction of speech, there is Belbury Poly’s song “Caermaen”, released in their 2003 album The Willows. It is crafted from the splicing up and recombination of a recording of an English folk song made on a wax cylinder in the beginning of the twentieth century. Through this work, it became possible to a “dead man [to] sing a brand-new song”, as music and culture critic Simon Reynolds wrote. Something similar gained momentum in 2016 with Star Wars’ Rogue One, in which Peter Cushing’s image was fully constructed through CGI, while Carrie Fisher was presented to look like she had the same age as in the original movies of the franchise. This showed that Hollywood could generate look-a-likes of dead actors, thus helping to foster the ethical and juridical debate surrounding the usage of the image rights in the afterlives of the original actor to whom they belonged. The combination of technologically-enable impersonations and holding the property of their movement on screen is also at the crux of the problems brought by deep fakes. In this case, the record passes as the action itself and, before its authenticity is evaluated, it has already caused damage in the real world — not unlike the movements of the stock market and its consequences upon society.
Deep Nostalgia and deep fakes have more in common than being partially homonymous. Perhaps that’s why Marina Amaral, a much respect “digital colorist”, was displeased by it. In a tweet, Amaral also considered the images “creepy”, then adding in a subtweet later deleted that they looked more like caricatures than real photos, “adding nothing to the power of the photos”.
This reaction is troublesome. There’s indeed a difference between the application of AI to “bring to life” historical picture and the research-intensive, deeply reflective practice adopted by Amaral. However, both are manipulations of the historical record, so where one should draw the line between a manipulation that “adds to the power” of the original and a manipulation that is considered only “creepy” or “uncanny”? For what reason one of these brings us closer to the past, while the other repels us away from it?
These questions bring us to what Mark Salber Philips termed “historical distance”. Historical distance is a complex combination of social discourses, epistemological practices, ethical mores, and technical features leading to how we relate to the past. It is possible, for instance, to feel close to a distant past because of its visual appeal or because of values pertaining to that society. Despite the complexity of the concept, I want to concentrate only the technical aspects of it and how they are related to historical experience.
Amaral is far from being the only “digital colorist”, in spite of being the most authorized and acclaimed that I know of. On YouTube, many amateur users are dedicated to colorizing scenes records in the black and white manner of early-twentieth century film. Besides colorization, these users also usually update the resolution of these footage to high-definition standards. This 4K take on the historical record has not come without its criticism, as some historians are keen to condemn the flattening of historical imagination.
The historical record becomes historical not only because it comprises what is depicted, but by how it depicts it, that is, the marks it carries of the time it was made — surface noise on LPs and the smell of the pages on old books are two examples of it, as well as the marks of the various imprinting techniques that made them. It’s not only the content that is historical, but the supports are too. However correct that is, the fact is that historians and colorists don’t share that same understanding of historical distance: for the former, the past matters because of its distance, and historicity is a sign of its alterity towards the present; for the latter, historical distance is a matter of identification with the past, bringing it closer to the present, “adding to its power”, to quote Amaral.
There are other instances of what could be called, following resolution theorist Rosa Menkman, the political life of resolution. It shows in Lucasfilm’s constant updating of the special effects on the Star Wars movies, as well as in the prevalence of remixes in comparison to original recordings in Spotify. These cases also show the economic power of resolution, as the constant updating ends up amounting to a new work, thus granting author rights to a third party, such as Spotify. This could also be asked of the digital colorists’ ventures, for who owns these manipulated images? The original holders, the archives that housed them, the dedicated users who crafted and updated them?
However pertinent these questions are, they lead us astray from Deep Nostalgia. I want to highlight, actually, that just as with the historian’s and the colorist’s clash over historical distance, Deep Nostalgia also underlines the uncanniness in the latter’s work, which we grew accustomed to and learned to naturalize. Both Deep Nostalgia and the digital colorist act upon the past as a mediated experience. Even if this could be said of every instance of historical inquiry, as historical investigation is always “indirect”, to remember the nineteenth-century French historian Charles Seignobos (1854–1942), both approach the past as media and conflate the past with its records. If one presupposes this, questions related to historical inquiry such as who, when, why, how, with what intentions, what where the consequences, are put aside in order to abbreviate the distance between the beholder and the portrayed. But nothing assures us that these figures moved the way they do in Deep Nostalgia’s recreations, just as the colorist’s colorizations are probabilistic, as it happens in any historical claim about the past.
The uncanniness in Deep Nostalgia is related to the realization that these figures shouldn’t be doing what they are doing. However problematic, it both reinforces and challenges Roland Barthes’ consideration of the referential nature of photography. Just as in Barthes’ case, these figures are without a shadow of a doubt, thus bridging the iconic gap that pertains to photography, but when they are animated, it is as if they weren’t because the referential aspect of the sign is overemphasized, thus becoming unnatural. Just as in Steps, the growing mediatization of the past can alter, but cannot change or move what happened. Therefore, Deep Nostalgia discloses, reinforces and destabilizes the pact of identification behind the digital reconstruction of the past, such as in the colorist’s work. Differently from the colorist’s assumption of the transparency of the images she works, Deep Nostalgia forces us to recognize the mediatic aspect of the historical record and how it shapes contemporary historical experience.
