Categories
Aesthetics Generative Art

Computing, Corporate, Critical, Contemporary

Harold Cohen’s AARON is exemplary art. Cohen’s painting of the 1960s was world-class abstraction, a serious and capable investigation into the nature of signification in image-making. His adoption of computer technology as a way of furthering this investigation was not opportunistic or promotional, it was a natural and effective embracing of a means of furthering his artistic aims. The aesthetic and critical content of Cohen’s painting has continued and extended in his art computing work.

AARON is exemplary artificial intelligence software. The AARON of the 1970s was an artifical-intelligence inspired production system written in C. The AARON of the 1990s was an object-oriented Lisp system. The AARON of the 2000s is still written in Lisp and is more inspired by the new dominant paradigm of cognitive science. Each instantiation of AARON is amongst the most successful examples of its programming and conceptual paradigms. It also reflects Cohen’s biography in its subject matter (from pictograms through social imagery to abstract pictures of house plants), although I will not discuss that here.

AARON is a combination and mutual intensification of the symbolic logic of Cohen’s painting and that of artificial intelligence. Rather than trying to approximate the appearance of art with computer technology, or trying to apply readymade forms of computer science to image processing, Cohen broke the theory and practice of his painting down to its most basic methods and imperatives. He then painstakingly embodied this in software using the most powerful methods available of representing not images but knowledge and behaviour.

AARON bears comparison with the most successful expert systems of the 1970s, and its output with the most successful of Cohen’s manually produced paintings. It is competent as art because it is competent as an expert system, and vice versa. Every smallest movement of the pen is a production of the expert system, and every part of the expert system exists to model Cohen’s artmaking. This is a result of the uncompromising pursuit of the practice of art through the means of computing where those means are appropriate to the artists ends.

The irony is that AARON creates more human-like art precisely because it more fully models art using the technology of computing.

AARON’s development is historically authentic given its historical environment. To reverse-engineer AARON now is to be an Impressionist in the 1920s, or a surrealist in the 1970s. At best it is to make woodcuts in the 1910s, or to re-make the Large Glass in the 1960s.

This is a criticism of my current projects. I would answer it with reference to Liu and to Art & Language: inconvenient history and genres are at least potentially resistant to the dominant paradigm of corporate knowledge culture. And AARON is very inconvenient history, paradigmatic software with irreducible aesthetic content that outperforms corporate exploitations of the same technology.

But AARON’s history is not my history (despite us both being born around the same time), and its historical environment and resources are not those of today. Contemporary computing is the social graph and dataset clouds of Web 2.0 . The crowdsourcing of “The Sheep Market” is exemplary within this environment and against the backdrop of relational art’s service-economy-that-protests-too-much. So is Casey Reas’s data visualisation, although it lacks the critical potential of the former.

The resources of contemporary computing are the vast data sets of wikipedia, archive.org, etc., and of the distributed processing systems of Amazon and Google. The coding paradigms are Ruby and its Rails framework, Python and Google’s mapreduce, and surprisingly a resurgent Lisp. Reas’s Processing is too neat a solution to too uncritical a position to corporate knowledge culture. It is exemplary, yes, but not critically exemplary, which to be fair it is not intended to be.

The lesson of AARON is that intensifying the aesthetic and significatory core of art using the resources of corporate knowledge culture’s contingent computational technology can provide an immanent critique of both. To be of value, the implications of this lesson must be worked out and through within the indexical environment and with the resources of art and computing for each historical moment. And this must be done as art if not for art’s sake then not for the sake of the efficiency of managerial regard.