“I was born in the Age of Machines” – Hollis Frampton
A few weeks ago, I saw a post on a subreddit about a bizarre YouTube account called Unfavorable Semicircle. According to a website devoted to the videos:
Unfavorable Semicircle is a YouTube Channel that since the 30th of March 2015 has been uploading strange videos. Generally the videos:
- Are 5 seconds long
- Have a man’s voice saying a letter or number
- Have a solid background with a colour pixel in the frame
The distorted noise and abstract images are creepy on their own, but the most unsettling part of Unfavorable Semicircle was that it was uploading videos at an inhuman rate, going from one every 10 minutes to three per-minute.
A few days later the BBC did a story about Redditors who were devoting much time and energy to try and solve the mystery. You can read about some of their theories in the article (it was also discussed on last week’s episode of one of my favorite podcasts, Mysterious Universe) but the most popular were variations on some kind of intelligence agency theme; it was a numbers station, a recruitment puzzle, a test of some sort. The theory I am partial to, mentioned on the US website, is that it’s an algorithmic test, along the lines of WebDriver Torso. Shortly after the BBC story ran, nearly a year after the account began uploading, it was removed form Youtube.
As a fan of the strange, urban legends, and the occult, I was drawn to Unfavorable Semicircle on its weirdness-value. But as someone in screen studies, the videos interested me in another way. To phrase it as a question, what happens when a computer (or AI) makes a structuralist film?
Unfavorable Semicircle’s Magnum Opus is the 25-minute long “LOCK,” which has been archived and reuploaded below:
The film looks like a lost, 1970s structuralist classic.
Now let’s take a look at one of Webdriver Torso’s films:
This looks a lot like a Mondrian painting. In both cases, we have an algorithm making what resembles modernist art.
Filmmaker Hollis Frampton utilized mathematical theorems to analyze and create films. In his essay “A Pentagram for Conjuring Narrative” he introduces several of them. The content isn’t important to this discussion, but what is significant is that he uses the word “algorithm.”
In his essays and lectures, the algorithm takes on metaphysical importance:
A waterfall is not a “thing,” nor is a flame of burning gas. Both are…stable patterns of energy determining the boundaries of a characteristic sensible “shape” in space and time….you and I are semistable patterns of energy, maintaining in the very teeth of entropy a characteristic shape in space and time.
The computer is a tool for manipulating symbols, in the most generalized sense. It seemed reasonable, since film or video or sound are symbolic systems in precisely the same way that mathematics or the natural languages are, to pursue further the implications of a tool that must be perceived as totally generalized in the fields where it had not been used or understood by experienced practitioners…
What is most important, though, in attempting to use this new tool is that it rigorously enforces a precise understanding of the fundamental nature of the symbols and systems to be operated upon. To state very roughly a fundamental theorem in computer science, the generation of an algorithm that performs a process requires a rigorous and detailed understanding of that process…
The attraction and power then, of computers is that they afford means to test and exercise our understanding of what we are doing with any tool, what we mean when we say that we make something, what we mean when we say we know what we are doing.
-Hollis Frampton, from “About the Digital Arts Lab,” 1982
In his book Dreams of Chaos, Visions of Order, James Peterson lists how what he calls “the minimal strain” of avant-garde filmmaking, called by others structuralist film, was incorporated into the Modernist art project through 3 schema.
- From Annette Michelson, the phenomenological schema: that films are presented as “direct perception” and “the film is interpreted as the embodiment of some fundamental feature of consciousness”
- From John Cage, the art-process schema. In this schema, the film is seen as displaying how it was made, or the viewer “begins by reconstructing the process that the artist used to produce the work,” compares those to “reigning artistic norms” and reads the work as a comment on the “repressive nature” of those norms.
- From Clement Greenberg, the anti-illusion schema, in which modern artists “reject representation in favor of an examination of the features of their media.”
You might have missed it earlier, but I referred to Webdriver Torso as a filmmaker. Is this possible? Can algorithms be artists? A lot of film theory deals with the question/problem of the camera and the artist; can a filmmaker be an artist if the camera does the work of creation? But these algorithms present another wrinkle. If we consider Webdriver Torso and Unfavorable Semicircle’s works as movies suddenly it challenges definitions of art in general, and these specific schema.
- Can they be considered as perception or features of consciousness if they are made by a machine that does not have consciousness or perceive in an understandable way to us?
- What does it mean if the process by which these films are made are to complex for the human mind to comprehend?
- How can we conceptualize media if these computers are unaware that they are creating media? What does it mean if, according to one theory, Unfavorable Semicircle’s videos are the visualizations of an AI being developed? If these are visual metaphors of processes that are not, in the conception of the creator, media?
In a 1978 interview, Frampton gave this vision of the future:
I think it’s fair to say that within two years we will have intelligent toasters…if you want a really smart toaster, let alone a Waring blender or an intelligent frying pan that will flip the omelet at exactly the right time, you are going to have it. And in fact…it’s probably going to be difficult to have anything else quite soon. And I suspect that that is going to have social consequences and consequences with regard to the fate of the counter-machine of language–the image machine–at least as far-reaching as video has, as broadcast television has, and probably more so eventually.
This, like most of his work Frampton is slightly joking here, vision of the future was about 30 years early and recalls a Rick and Morty scene. But given his interests, it’s odd that Frampton doesn’t take the next step: instead of an intelligent frying pan, an intelligent camera. Then again, in these cases, there isn’t even a camera. So we are left with cinema created without a camera and without the involvement of a human being. However, the possibility that these images are metaphors for other processes, the byproducts of attempts at conceiving problems related to purpose is itself a metaphor by which we also understand art and cinema: that we create these works as metaphorical attempts to understand our own place in life. In this case, cinema may not definable by its elements, its creation, its creators, or even intended use, but rather by its relation, it’s adjacency to activities of existence. The machine is trying to figure out what it is and can do. These videos are the result. Humans are trying to figure out what we are and what we can do; art, and our videos, are the result.