Contradictions - Paul Cohen
When Paul Cohen is writing about his late father Harold Cohen, in a short piece called Contradictions, he speaks of the complex-simplicities, the directed-randomness, the predicted-unpredictability and the authored-autonomy which AARON (the AI driven art-machine) possessed. On the face of it, the top-layer of algorithmic processing which leads to AARON laying mark to paper, are arguably simple in their execution but Harold conflated (not confused, but conflated) the program and it’s products with his decades of complicated and deeply sophisticated thinking about the program. Historians will have to sort this out.
Paul says seperating the Cohen-AARON project into distinct causal elements is reductionist. Which differs from a scientific endeavour where results are independant of scientists. Paul and Harold both grapple with the idea that AARON is and autonomous creation, which can create infinitely if it were allowed to do so, but it was created by, and continually developed by Harold Cohen and you cannot ignore that fact. As autonomous as it may be, the fact that it is allowed certain levels of freedom/selection/choice as it operates, are all programmed in by Harold Cohen.
Harold was so close to inventing the science of art, but he failed to do so because he could not write himself out of the picture, as scientists must. There has never been an apparatus so capable of answering questions about art as AARON was, but Harold made all the answers conditional, dependant on him, like a thermometer that worked only for him.
(The Speculative (for Propositional) Art of (Harold Cohen) && OR (AARON))) - Sheldon Brown
Brown takes the opportunity, when looking at Cohen-AARON’s work, to consider where it sits in the bigger picture of art practice generally, and how it compares in what it challenges, what questions it raises and how it tackles them as other art movements have done.
the troubling of both the nature of the artist and the status of the art object due to the development of technological methods and conceptual interests, was at the core of the larger project of art in the 20th century – from Marcel Duchamp through Walter Benjamin to Saul LeWitt et al. In this context, the common thread of questions about the site of creativity takes us on a walk through the modern art era: is it in the eye of the Impressionist? the technique of the Photographer? the gesture of the Dadaist? the sub‐conscious of the Surrealist? the body of the Expressionist? the community of the Relational Aesthete? the algorithm of the Computerist? There is something a bit jarring in the neologisms of those more recent concerns, perhaps indicating some qualitative differences in their stakes.
Across the decades of his work, we can see AARON “learn” increasingly complex ideas that reflect Harold’s conceptualization of aspects of drawing and painting.
… by the end of his career, the territory explored by AARON was congruent with the territory of Harold’s interest in art.
On thinking about Harold and AARON as a stepping stone in art history which is beginning to explore the possibility space of an art tool that can extend beyond what we humanly expected/intend/comprehend.
Harold’s work is one of those bridges from one era of art operations to another. It is about what is explicitly seen but also about what is unseen, implied and imagined, as to where art might come from next.
Interview with the Artist: Harold Cohen, by Sheldon Brown
When talking about methods of AI:
How do you model your own behaviour when you don’t know anything about that behaviour?
…machines and human beings are really different entities.
The idea that the program has to understand the words in the same sense that a human being understands the words wasn’t productive.
These are all quite prevalent discussion topics in the world of AI: the Cybernetics/AI debate and the model of cognition. The AI model being what Cohen speaks of which mirrors the Cartesian duality concept of body and mind being two separate things (hardware and software).
It is interesting to hear Cohen talk about the means of producing the artwork. For Cohen it really is all about AARON, the software (the mind). The various drawing and painting machines he built over the years ultimately turned into a distraction from what he deemed the really interesting aspect of the work and that was the images produced by AARON. Cohen was disappointed when museum/gallery goers would become entranced by the robotic arm cleaning its brushes and emptying it’s paint pots, when in fact this was necessary, and no more impressive than your desktop printer. So ultimately Cohen turned to wide-format colour printing when that technology became available.
I wrote somewhere that I sometimes think I’m a prototype for the coming cyborg, not in the sense of having mechanical parts to my body – I already have that, and they don’t work very well – but in the sense of having implants in my brain that are capable of doing things that I couldn’t do with the other parts of my brain. The only difference is that my implants aren’t in my brain, they’re sitting on my desk. But I feel very connected in my relationship with the program now.
In looking at Cohen’s work with AARON (or AARON’s work with Cohen), many great questions about art, cognition, authorship, aesthetics and so on are interrogated. By Cohen creating AARON as a collaborator (which is how Cohen himself has described AARON as) it gives Cohen a means to explore territories that he could not do by himself. He comments on artists who work in a particular ‘style’ for many years, and how AARON could go on working infinitely, but only with the limited set of formal rules he currently has. AARON developed over time as Cohen did to, and the relationship was symbiotic.