A figure pulled from Minecraft, Minions, Fortnite, Looney Tunes, DeviantArt portraiture, the Are ya winning son meme, corporate illustration, or some other slice of the American vernacular is shown in three-quarter portrait on a field of bright colour, looking indifferent. The figure is decorated with references to popular entertainment (Harley Quinn’s baseball bat, the three-eyed fish from The Simpsons, the floating orb of water from The Last Airbender, the face mask from Mad Max, Yugi-Yo’s hair, Venom’s teeth), popular consumption (bum bags, hoodies, puffer jackets, sneakers, vapes, tinted glasses, branded headphones, gold jewellery, takeout food containers), and crypto-specific symbols (laser eyes, diamonds, and currency icons).
The image is either a vector drawing or 3d-rendered, but in either case there is little suggestion of depth; every element is evenly lit and depicted in sharp detail as if it were pressed right against the image surface. There is a love for visual detail: every hair is precisely delineated, every piece of gold is polished, and cloth is carefully draped, lasers glow.
This description covers most non-fungible tokens at the top of OpenSea (the biggest website for buying and selling NFTs) on any given day. Why do all these images look so alike?
The first level of resemblence has to do with the fact that mainstream NFTs are generally produced by an operation called layering. You start by making a list of elements like “background”, “clothes”, and “hat”. Then you produce (or pay a gig worker to produce) a set of images corresponding to each element: A few different backgrounds, some variations of your character, and some different hats. Finally you use a simple computer program to stack these layers on top of each other in random combinations, producing a set of final images. The more elements and layers you have, the more images you can produce, and the bigger your payoff will be if the collection catches on.
The practice of assembling an image or other media object from a set of individual elements isn’t unique to NFTs. It has some parallels to collage, but the more apt comparison is compositing, a processes that happens everywhere in contemporary media production. The design of any movie, video game, or other “new media object”, writes the critic Lev Manovich,
… begins with assembling a database of possible elements to be used … Throughout the design process, new elements are added to the database; existing elements are modified. The narrative is constructed by linking elements of this database in a particular order, that is by designing a trajectory leading from one element to another. The narrative is constructed by linking elements of this database in a particular order, that is by designing a trajectory leading from one element to another. On the material level, a narrative is just a set of links. 1
The most visible example of this is the film industry. When Tony Stark walks through a hangar in Avengers: Endgame, the footage of Robert Downy Jr, the 3d-model of the costume, the building, the HDRI sky, the aeroplanes in the background, and even the lens flares are only temporarily brought into the same frame — in reality they’re separate, independent assets (both in the media-industrial and financial sense of that term), ready to be re-assembled into other outputs down the line.
NFTs are a heightened, distorted version of this. Making a collection is mostly about filling the database; the linking happens almost incidentally, in the simplest way possible (randomised layering), fully automated in a few dozen lines of code.
If you want to make visually coherent images in this way you have to make the elements interchangeable; this limits the kinds of visual gestures you can make and, together with the fact that the same elements appear again and again, produces collections of very similar images. It also causes the flattened, occasionally shifting perspective and visible seams between elements that are characteristic of many NFT collections.
In both cases, the composite nature of the images isn’t a secret. Disney, who owns the Avengers assets, releases regular VFX breakdowns demonstrating the fact to their shareholders and everyone else, and OpenSea shows a list of the constituent elements next to every NFT for the same reason.
The similarity between NFTs across the field (not just inside a given collection) is an extension of this logic. A prominent selling point of many collections isn’t so much that they’re assembled from a database, but that they might themselves become one; their images re-assembled into fresh media products like comic books, toys, TV shows, video games, experiences, and merchandise.
NFT projects aren’t generally prepared to do the work to actually make any of these things (it takes hundreds of artists working 12-hour shifts to turn a database like Disney’s into a movie like Endgame, not to mention the people sewing the merchandise or staffing the theme parks), but that doesn’t matter. The idea that such labour could be performed in the future, and that you would be able to pocket the surplus by owning a piece of the database is enough to sell it.
This piece of speculation starts out in roadmaps and other marketing material, but it quickly seeps down and across into the images themselves, where it crystallises into the visual cues we’re familiar with:
- Characters are popular because it’s easy to imagine how they might appear again and again in different media products, just like Tony Stark. Visual references to existing media properties (both in the choice of character and the accessories) are designed to reinforce this connection.
- The plain backgrounds, walk cycles, soft lighting and neutral expressions reiterate the possibility that the figure is ready to be combined with real environments and real behaviours into a real product. (This is also what those gestures are designed to communicate in VFX breakdowns).
- The depictions of real-world luxury products might be read as straightforward signifiers of value (gold chains are valuable, therefore an image one is valuable as well) or, maybe more aptly, as a promise to those who buy into the speculation — you, too, are going to make it.
Incidentally, this also explains why there is such a big push to enable NFTs as avatars on social media sites like Twitter and Instagram. It’s a growth hack (people see your avatar, they buy an image from the same collection and set it as their avatar, more people see it, everyone profits), but more importantly it shores up the claim that NFTs can be composited into other media objects.
Of course, proponents are quick to point out that this is only proof-of-concept. The merchandise, video games and all the rest will be here any minute now.
It seems likely that NFTs will implode for reasons external to them: The underlying currency might collapse, they might be regulated out of existence for the environmental fallout or the widespread fraud (or both), or they might just run out of buyers. But there is a deeper argument against them: There isn’t really such a thing as a non-fungible image.
“In principle”, the philospoher Walter Benjamin wrote in 1936, “the work of art has always been reproducible”. 2 People throughout history, he argued, made images in multiples: The ancient Greeks mass-produced pictures of their rulers by striking them into coins, which they spread across the continent. Around the third century, Chinese cloth-makers began to carve images into wooden blocks, which they covered in ink and pressed against silk, leaving a coloured impression that could be repeated over and over. The technique spread, and soon artists in every town were churning out thousands of playing cards, religious icons, portraits, and scenes from nature and everyday life. Woodblock printing was joined by copperplate engraving in the 15th, lithography in the 18th, and photography in the 19th century, each time increasing the veracity and speed with which images could be reproduced.
The current stage of this development is the digital image, where even the simple act of looking entails multiple acts of reproduction. When you open an image on your computer, it’s copied from your computer’s hard drive into its working memory, parsed and translated, until a specific array of pixels in your screen is lit up to render the image. As soon as you close the window, those pixels are turned off and the picture you were looking at is destroyed, only to be produced afresh the next time you open it. When the image is stored online (as NFTs typically are), this process happens everytime anyone looks at it.
Naturally each of those coins, bolts of silk, packs of playing cards, printed photos, and arrays of pixels on your screen are as “real” and “authentic” and “valuable” as all the rest of them — there isn’t really an “original” to speak of.
It takes an enormous amount of labour to suspend belief in this fact, even temporarily. In a classic essay, the art critic John Berger describes the lengths to which the National Gallery in London has to go to maintain the notion that their version of a painting by Leonardo is in fact “the original”:
[The catalogue entry] on the “Virgin of the Rocks” is one of the longest entries. It consists of fourteen closely printed pages. They do not deal with the meaning of the image. The deal with who commissioned the painting, legal squabbles, who owned it, its likely date, the families of its owners. Behind this information lie years of research. The aim of the research is to prove beyond the shadow of a doubt that the painting is a genuine Leonardo. The secondary aim of the research is to prove that an almost identical painting in the Louvre is a replica of the National Gallery version. French art historians try to prove the opposite. 3
Similarly, Disney employs an army of copyright lawyers, image recognition software, lobbyists, a whole judicial apparatus to maintain the notion that the 3d-model of Iron Man in their database is in fact “the original”, and that everyone else only has a temporary viewing license.
NFTs are another attempt at this. People who make them recognise it’s difficult to argue that a digital image can be “original” on any material level, so they suggest a kind of authenticity-by-proxy: Buy an NFT and you get a unique entry in our special database saying you own the image. That database entry has effectively the same function as those fancy art historians and copyright lawyers: Establish authorship, keep track of provenance, authorise derivative works, mediate royalty payments, and so on.
Critics argue that this doesn’t work: There is no way of knowing, for instance, if someone who mints an NFT really made the image, and buying an NFT doesn’t really give you ownership of the image in any legally recognised form.
They’re clearly right, but if there is anything to learn from the history of image-making, it’s the notion of the attributable, ownable, dateable “original” is itself pretty shaky. Images were always produced collectively and in abundance; the recent drive (historically speaking) to enclose them for individual profit must be overcome.4
Lev Manovich (2001): The Language of New Media, p. 231 ↩︎
Walter Benjamin (1936): The Work of Art in the age of Mechanical Reproduction. Available at marxists.org/reference/subject/philosophy/works/ge/benjamin.htm ↩︎
John Berger (1972): Ways of Seeing, p. 22. Available at archive.org/details/waysofseeing00berg/page/22/mode/2up ↩︎