Train Car #2 contains: Thoughts on the Poor Image

💫 - 💫 - 💫 - 💫 - 💫

Prior to the invention of photography in the 19th century, art was not reproducible. Sure, sure, we’ve always been able to paint copies of famous artworks; ancient Greece stamped standardized designs onto coins; and techniques like woodcut printing, etching, and later lithography allowed for limited reproduction. But today we can make infinite reproductions of artwork nearly indistinguishable from (at times, perfect copies of) the original.

So what’s lost? In 1935, German philosopher Walter Benjamin argued that “Even the most perfect reproduction of a work of art is lacking in one element: its presence in time and space, its unique existence at the place where it happens to be… that which withers in the age of mechanical reproduction is the aura of the work of art,” (Art in the Age of Mechanical Reproduction, Benjamin, 1935).

So artwork has some sort of “aura” that’s lost in reproduction, something inherently tied to the original, its place in history and tradition and culture, its physical presence, its accessibility or inaccessibility to the viewer, that’s lost in reproduction. But why should we buy into any of this bullshit? (Or at least, that’s what we asked Dr. Kevin Vennemann when he introduced our German philosophy class at Scripps College to the concept of an artwork’s aura in the fall of 2022).

He pointed out that we care a great deal about authenticity–that there’s a difference between the print of Van Gogh’s Starry Night that you hang in your living room and the one you wait in line to see at the MoMA, even if physically you can’t tell them apart. There’s a difference between walking along towering sections of the remains of the Berlin wall, watching the way it cuts the city in half, and looking at pictures online. This difference is the aura.

But pictures online throw a real wrench into Benjamin’s analysis. Can’t blame the guy–he died in 1940. His analysis was a first reaction to the impact of photography, the proliferation of copies among the masses, and the removal of artwork from its historical context of religion and ritual and encroach into politics instead. It’s still deeply relevant, but the internet has changed the ways that we interact with images.

In Hito Steyerl’s 2009 article In Defense of the Poor Image, she defines the poor image as, among other descriptors, “a copy in motion. Its quality is bad, its resolution substandard. It is… an itinerant image distributed for free, squeezed through slow digital connections, compressed, reproduced, ripped, remixed, as well as copied and pasted into other channels of distribution… The poor image is an illicit fifth-generation bastard of an original image. Its genealogy is dubious. Its filenames are deliberately misspelled.”

I first read Steyerl’s article in 2023, in Dr. Daniel Hackbarth’s Metropolis class. We were looking at physical and digital space; the ways that we construct cities online.

You know poor images. We’ve all dealt in them. Who hasn’t watched a pirated copy of The Last Unicorn online, or deciphered a meme through layers of jpeg compression and logos and banners added by seven different reuploaders? They’re omnipresent in our digital landscape, and I don’t think that’s a bad thing.

To Walter Benjamin, a replication can’t have an aura in the way that an original does. A poor image is defined by that lack, that hole where an aura should be, the immutable distance between it and the original, made up by downloads and uploads and redownloads and reuploads, mechanical degradation and human filtering, its position in space and time far removed from the context of the original.

I’m kind of obsessed with poor images. They’re all wild mythological creatures to me, sailboats on a rough and wild sea, tossed far across the globe every night, transformed and reshaped and dragged from port to port, never alone and never at home. They don’t have an original, or even just one “true” version. There’s no one author. They’re so tied to their particular moment in time, to the technologies and processes that created them that you can’t possibly extricate them from their context. Somehow, these low-quality, jpeg-crunched images have built an aura around themselves when one shouldn’t be possible.

The situation of the poor images “reveals the conditions of their marginalization, the constellation of social forces leading to their online circulation as poor images. Poor images are poor because they are not assigned any value within the class society of images—their status as illicit or degraded grants them exemption from its criteria. Their lack of resolution attests to their appropriation and displacement,” Steyerl writes.

🌈 - 🌈 - 🌈 - 🌈 - 🌈

Have you ever seen these things called “stimboards” online? They’re collections of images, usually gifs, with pleasing colors and visuals and motions. Lots of nostalgic aesthetics, slime videos, flashing lights and trailing colors without a clear source. They’re a particularly interesting example of poor images because of the way they proliferate online. I usually find them on Tumblr.

The gentlemanly thing to do when you post a stimboard is to leave a line of credits across the bottom, hyperlinked emojis. Like this:

🎨 - 🎨 - 🎨 - 🎨 - 🎨

It’s cute, right? The fun thing about these links is that about 25% of the time, they’re broken. They don’t lead anywhere except a dead URL. The other interesting part of them is that the links don’t usually lead back to what you’d consider the “original” source.

⭐️

Here’s a typical genealogy: In scary-caregiver’s Star Stimboard post from 2024, they link the source of this starry blanket gif back to a 2023 astrology-themed stimboard by stim-burrow, which links back to 2021 stimboard by helium-stims themed around a French children’s TV series that first aired in 1999. Finally, this post links back to the “original” source: a 2019 YouTube video called “Rubyjam Fabric - Glow Stars Blue” uploaded by an Australian fabric company. It has 176 views.

I followed this set of links in less than a minute, so why does the 2024 poster not just link back to that first video? I think it has to do with the culture of poor images online.

1. The original video is not yet a poor image; it bears very little resemblance to the gif that I see on Tumblr. It’s 18 seconds long, and it has a clear context and narrative. Someone is displaying a fabric for sale, turning and twisting it to show its properties.
   
   But the images used in these stimboards are so divorced from their origins as photographs or videos of the real world. These things are beautiful but none of them are real. They’re not real life. The hands don’t belong to people, the artists are no one, the images could be animated or AI-generated and I don’t think I would notice.
   
   In pulling just a short, zoomed-in clip of this video, the gifmaker, the 2021 poster, has transformed it. They’ve removed it from its context, made it into a bona fide poor image, valued now for its aesthetic appeal rather than its commercial properties.
   
   
2. Alright, so credit that first 2021 poster. They’re the one who clipped the video. But there’s a subtle change from that first gifset to the second in 2023; where before the fabric looked flat and real, now it’s darker and subtly glowing. It’s filtered to look like the real night sky, edited and improved and recompressed, altered before it continued its journey.
   
   And I would argue that the method of discovery is still a piece of this poor image’s genealogy, its aura. Crediting the source by which you found the gif in the first place is the courteous thing to do. The original owner of a poor image is less important than the path that it took; its aura comes not from the conditions of its creation, but the roads by which it has travelled.
   
   

I like to follow these paths back. I like the stimboards. The images are nostalgic, but the nostalgia doesn’t call back to any time other than a nebulous “childhood.” My childhood didn’t look anything like these pictures but I still yearn.

The paths take you interesting places. Instructional videos from the 1990’s. Paint stores on Instagram. An Etsy listing for slime. They lead you through weird corners of the internet you’d never have imagined existed. But often these images don’t even have a real original version anymore. You click through the source lists, post through post through post, and it ends in an uncredited gifset or a TikTok that was deleted or a Pinterest pin with no apparent source.

And even the ones where you find their source, the clipped, sped-up, filtered, 1-second gif is not the original video. Someone grabbed a second and looped it over and over and over again, freezing time, freezing an idea or feeling that’s not part of the original. The little girl in Indonesia who posted a video of her new Aquapets toy in 2013 didn’t know this was going to happen! So what’s the source? What’s the original? Does it even matter? Are these just a part of the fabric of the internet now?

I used to post my art online, back when I was a teenager in the late 2010’s. I remember the first time I ever saw one of my own pieces reposted on Instagram, just noticed by chance as I was scrolling through. I had to do a double take, a wait-that’s-me. That’s my art. Posted by a fan page that credited another fan page that said source: unknown. It drove me nuts and it happened constantly, but at least it was still my original image. Maybe a bit blurry and with the edges cut off, but you could reverse search it, or make out my little artist’s signature in the corner. It hadn’t yet taken on a life of its own, left to sail around eyes on the internet that I hadn’t intended to meet.

But I know that between 2016 and 2019 my art was spread all over the corners of the internet, from Tumblr to Instagram, Pinterest, We Heart It, Reddit, YouTube compilations, Twitter, Facebook, iFunny, reshuffled into memes and reaction videos and fandom archives; and I know that at some point an AI image generator must have scooped it up, scraped it as uncredited training data back when they were first trawling the interwebs. I know my art is built into even those earliest models from 2021, and no matter how many consent boxes I uncheck I can’t pull it out. I think that has to be the final stage of a poor image. Every time I look at an AI-generated artwork, I have to wonder if any of its pixels were once mine.

AI-generated images are sort of the ultimate problem for Walter Benjamin’s idea of the aura of artwork. They simultaneously exist perfectly without history (their artists are no one, their genealogy is lines of code), but also completely trapped within the ideas of a society that made them. AI-generated images are all kind of cliche–they can build fantastic worlds in futuristic, cyberpunk cities, or grand, ancient temples–but it’s all things you’ve almost seen before. Building only with images that already exist, generative AI necessarily cannot make something new or without cultural context. Is this context an aura?

NFT’s did their best to introduce the concept of an “original” of a digital artwork, or at least one “true” version, though I would argue that they had very little success. Take probably the most famous NFT collection–the Bored Ape Yacht Club (so sorry to dredge up pandemic-era memories, I promise I have a point). Like almost all NFT collections, each of these images is procedurally-generated, rather than individually painted by digital artists. Each has various unique features–a hat, a gold chain, a weird background color–pasted over a base image of a monkey. They’re all hosted on the Ethereum blockchain, which means that when you purchase one of these apes for a sometimes-exorbitant price, you get a smart contract that points you to a URL with your image. That’s right! You are a very special buyer and you receive a very special URL! Look, the token even has a unique serial number on the Ethereum ledger! Everyone can see that it’s your image!

This concept of digital ownership and uniqueness is so widely misunderstood and poorly defined that most people don’t even know what an NFT means. Like, it’s an image, and it’s the one true version of that image because… somebody said so? But what if I mint another NFT of the same image on a different currency’s blockchain? What if someone else mints NFT’s of the same image on the same blockchain? How the hell are we supposed to know who the “original” is?

The aura’s lost. There’s nothing there.

NFT’s haven’t really been culturally relevant since 2021 (or financially–a screenshot of the first tweet that sold for nearly $3 million back in 2021 just failed to reach even $300 in a bidding war), but their spotlight in the public eye reveals how concerned we are with what is lost without an original of an artwork–we are all concerned with aura.

So it’s something I like to think about. Can a poor image have an aura? How? Is it different from the aura of a traditional artwork? What do poor images reveal about their origin, the internet as it is today? What do they tell us? And where can I find more of them?

References

1. Walter Benjamin, “The Work of Art in the Age of Mechanical Reproduction,” (1935). https://web.mit.edu/allanmc/www/benjamin.pdf
2. Hito Steyerl, “In Defense of the Poor Image,” (2009). https://www.e-flux.com/journal/10/61362/in-defense-of-the-poor-image/