Computer Art Image of the Month - June 2011

June 2011

  Summer Equinox (thumb)

Click the image to enlarge

Summer Equinox by Harold Cohen, 2010

Oil over pigment ink on canvas, 121.9x192.7 cms

Copyright the artist, reproduced with permission

Autonomous Art-Making Machine

Professor Harold Cohen is undoubtedly one of the greatest pioneers of the relationship of art to software architecture. He is the only artist I know who is tackling the problem of building a program that actually makes art, rather than modelling human art-making.

His AARON system may be familiar to many readers, but Harold’s latest works (shown recently at Bernard Jacobson, Cork Street) are a surprise - one that reminds us of his origins in modern European painting.

Harold established his reputation in the 1960s as a Slade School educated painter and represented Britain at the Venice Biennale in 1966. He went on to become an innovator in the use of computer code in art practice with his autonomous art making program named AARON, admired by both members of the Artificial Intelligence and fine art communities.

AARON is a decision-making program using a printer as an output device (in the early days Harold built various output machines himself). All the elements are generated from scratch by an algorithm that accumulates little ‘cells’ into different entities, depending upon the local accumulation rules. By developing a relationship with software Harold has created a truly symbiotic relationship with the computer.

In 1968 Harold was invited to the University of California San Diego where he learned FORTRAN programming. He became interested in the concept of writing a program to do some of things humans do when they make representations and thus possibly learning more about the nature of representation.

He spent two years working with Edward Feigenbaum at Stanford University’s Artificial Intelligence lab and following various early experiments, AARON began in the early 1970s. However the seeds were already planted and growing before he encountered his first computer over forty years ago. Believing he couldn’t go on inventing forever, the idea that you could use a set of rules to do the inventing attracted him.

In fact the paintings he showed in Venice in 1966 were made following his own set of predetermined rules. With computing technology he began to wonder whether it might be possible to externalise everything he knew about art making so that a machine could do it. However he tells me that in those early days the implication for ‘laypeople’ who wrote code was that they were stupid and really ought to leave it to those who knew what they were doing!

For Harold, acquiring expertise wasn’t really an option, it was a necessity - indeed it’s what he did when he was painting, so coding can be seen as an extension of this practice. In the mid-1970s an AARON drawing was downloaded over the ARPANET at the Royal College of Art in conjunction with Imperial College and became the first usage of the internet predecessor in Britain.

Harold has spent many years working out a means of encoding outside knowledge - that is knowledge of how to compose, draw and colour paintings.

Speaking in 2003, Harold explained to me that some knowledge is easy to encode, some is very difficult, or perhaps impossible: ‘What we think of as criteria for action are not nice simple things like that - they come in hierarchies, so that I have some criteria say, for example, for colour, but behind that there would be other criteria which would say what my attitude towards colour ought to be, not what colours do I like, but why do I think colour is interesting and behind that another one and so on.

So how do you give a programme criteria for modifying itself? Probably the only way would be to start by describing a personality - but it’s such a big job, I just want to make art!’ The program was originally written in C, later he changed to LISP, which he found to be far more flexible and expressive, and with which he was able to solve the challenge of creating colour in the mid-1980s.

In the past Harold has described AARON as similar to an apprentice, a talented and able assistant, which enables the creation of imagery he couldn’t have made himself or by any other means. And, until recently, the images produced with AARON were completely done by the program and printed as finished works in themselves. Harold has never edited AARON's work beyond choosing which ones to print.

In fact, it has been a deliberate decision not to provide himself with any way of modifying its output. The latest works however, of which our image this month is one, use paint applied by Harold by hand over ‘underpainting’ produced by AARON. This is the latest development in Harold’s ongoing explorations of the wider implications of how we think about images.

He explains, ‘I find myself increasingly preoccupied with the part played by the MAKING of images. Very roughly: for the whole of human history, images have been made by hand, and I suspect we are finely tuned to recognise the physical manipulation of material - the MANufacture - as first evidence of intentionality.

Conversely, we can ‘see2 elephants in clouds, but we don’t believe anyone intended to communicate ‘elephant’. And, as AARON's images were becoming simpler - fewer, larger elements - I was finding them to be lacking ... what? Can a computer program have intentions? Of course! In one sense a computer program IS the reification of an intention. But little of that intentionality seemed to survive the untouched-by-hand quality of the all-in-one-go printed images.’

So paints and brushes that had been in storage for ten years were brought out and his studio reorganised. These new works are abstracts although the shapes perhaps suggest biological forms and there is a sense that they grow organically across the canvas. The artist thinks of them as ‘existing in some imagined under-water realm, and that determines how I develop them.

I want, above all, to generate a sense of place’. Ultimately however they remain enigmatic. By what curious act of imagination does one pack a lifetime of experience as a colourist into half a dozen lines of code? Well take a look at Harold’s new brilliantly-coloured, larger than life paintings and maybe you’ll find the tantalising glimpse of an answer.

Catherine Mason is the author of A Computer in the Art Room: the origins of British computer arts 1950-80, published in 2008.

More information on the computer arts including our events programme

More about this month’s artist

Comments (2)

Leave Comment
  • 1
    jack tait wrote on 10th Jun 2011

    Enjoyed your piece on Harold. To your statement re artists who programme machines to make art could at a stretch include my machines which were built with the intention of producing art objects. Just completed a PhD on this.

    Report Comment

  • 2
    Catherine Mason wrote on 30th Jun 2011

    Yes. Thank you, Jack.

    Report Comment

Post a comment