Idolatry Machine
I recently premiered an image collage and supporting series at an art exhibit and thought it would be good to do a write up about it.
Idolatry Machine
Eidos, Greek root word for image, idea, ideal, and idol. Idolatry Machine is a conceptual series that reflects on the ways in which our capacity for image making, and abstraction more broadly, can lead us astray.
I’ve been meditating the past few years on the subject of idolatry and abstraction. This is a big theme in my work, and one of the main reasons why I have been interested in neural networks as a medium for so long. I highlight the duality of noise and abstraction in my original essay on neural style transfer.
Generative image neural networks are literal “idolatry machines” in that they are image making machines, but more broadly all neural networks are idolatry machines in that they are a compression of reality into simpler forms that we can then manipulate using math. This compression and abstraction of reality is referred to as latent space, and you can read my post explaining it, for an artist audience here.
My first piece I created on this idolatry theme was a singular piece called “The Sin of Idolatry”.
All images with their full information can be found on my public Midjourney archive here: https://www.midjourney.com/app/users/c5dfa62d-93c2-4d16-a8e6-a385946ebda5/
The full prompt for this was “The sin of idolatry, illuminated manuscript” plus the following still from Zoolander, used as an image prompt.
Haven’t seen Zoolander in forever, but it’s a good flick!
I like tossing abstract phrases to the neural network to see what it comes up with. A lot of people try to force a specific set of outcomes with the neural network, but I take a more generative approach. This is informed by my long running interest in Algorithmic Art, and specifically the artist Yvral whose work I wrote about in undergrad over a decade ago. I got over some of my views on the generative approach in my statement of ethos here.
Idolatry Machine was a phrase that came to mind as a neat juxtaposition. Idolatry is abstract. Machine is concrete, but also abstract in that it is a general category. What do the two together make?
This lead to the Idolatry Machine series:
These are vanilla outputs, just the text prompt “Idolatry Machine, illuminated manuscript”.
The phrase is evocative and ripe for variation.
Throw in the term “risograph”, a specific brand of digital duplicators manufactured by the Riso Kagku Corporation in the 80s, and you get this:
Machine, not computer, and yet it’s interesting how computer-like so many of these images end up being. Is the neural network picking up on the image etymology of the word “idolatry” and giving a screen? Or is that an emergent property of idolatry and machines generally speaking? Can’t really tell, it’s a bit of a mystery.
There’s something creepy and alien about these neural network images, especially with the earlier models. This semantically porous abstraction is part of what I find compelling about neural networks as image making tools. To quote that one famous Brian Eno bit:
“Whatever you now find weird, ugly, uncomfortable and nasty about a new medium will surely become its signature. CD distortion, the jitteriness of digital video, the crap sound of 8-bit — all of these will be cherished and emulated as soon as they can be avoided. It’s the sound of failure: so much modern art is the sound of things going out of control, of a medium pushing to its limits and breaking apart. The distorted guitar sound is the sound of something too loud for the medium supposed to carry it. The blues singer with the cracked voice is the sound of an emotional cry too powerful for the throat that releases it. The excitement of grainy film, of bleached-out black and white, is the excitement of witnessing events too momentous for the medium assigned to record them.”
This alien quality plays into the ambiguity of the idolatry theme. It brings attention to the fact that these images are not “real”, they are dreamlike, inhuman, or post-human, products of collected subjectivities into an inscrutable intersubjective artifice. They are the product of human belief, shared constructions that offer a window into our culture. They are made in our image, and are reflections of us.
Our capacity for abstraction is powerful. Without it language would be impossible. Subjectivity would not be possible. The totality of the world is too large for any single mind to hold. We necessarily reduce the totality of reality to simpler forms in order to manipulate it and put it to our own uses.
Latent space is a mathematical way to reduce vast information into relevant features. Who defines relevance? If you are a human then you might see faces where there are none.
https://old.reddit.com/r/Pareidolia/comments/1ztnxl/cloud_face_video_in_comments/
Humans project our humanity into the world around us. If I were a bat, I would experience the world completely differently. All sheep look the same to humans, but sheep know other sheep quite well.
Abstraction is a valuable tool but it is also deceptive and can leave us astray.
“Moses said to God, “If I go to the Israelites and tell them, ‘The God of your fathers has sent me to you,’ and they ask me, ‘What is his name?’ — what should I say to them?”
God said to Moses, “I am that I am.” ” — excerpt from Exodus 3, NET translation
Later Moses finds the Israelites worshiping a golden calf of their own making. When confronted, Aaron, brother of Moses says “So I said to them, ‘Whoever has gold, break it off.’ So they gave it to me, and I threw it into the fire, and this calf came out.” — excerpt Exodus 32, NET translation.
That’s a “dog ate my homework” level excuse. It just came out of the fire, all on its own! We threw a bunch of data in and look what came out!
The calf didn’t appear from nowhere. It’s a human creation, and so are these neural network models.
Idolatry Machine is an ongoing series, but when I was recently asked to exhibit at a show I decided to push things further.
All these above images were created using Midjourney, text to image, with no editing other than upscaling for print. I decided to also make a special piece for this exhibit which leads to the Idolatry Machine: Altarpiece.
I’ve wanted to experiment with photobashing techniques and neural networks for a while. I created layered “fusion” paintings with style transfer models in the past. This was a different thing though.
Control over surface quality, lighting, general color scheme etc is looser, and prone to strange unexpected outcomes (this is part of their appeal to me). But getting that to mesh harder.
The altarpiece is a formal medium that arose in western Christian Medieval Churches. These were functional works of art that served a non-commercial religious and cultural purpose. Craftspeople who painted altarpieces were regarded in the same fashion as the carpenters who built the altar itself. This lack of distinction between “fine” and “low”, or “mechanical”, or “commercial” art/craft is expounded on in more detail by Larry Shiner in his art history book “The Invention of Art”.
My primary reason for choosing an altarpiece as the subject of this collage was practical. It lent itself to the form of the medium. Altarpieces are paneled, and often compositions of multiple smaller semi-self-contained elements together. This works well with the Idolatry Machine themes, but also works well with the intended formal mission of collaging multiple text-to-image generated outputs together.
ALL IMAGES ARE FREE
All images are free. As stated in my ethos statement, I do not sell images/idols. You can download all the images in this article at full resolution here: https://cloud.aslanfrench.work/s/free-images
It cost me about 70 bucks to get the altar piece printed at Office Depot, 2 24x36 inches.
