Evolvable Art

Biological evolution occurs because of mutation. When individuals reproduce, they copy their genes into new cells that will become their offspring. This copying is usually not perfect: copying errors that change the genetic code are called mutations. These errors then get copied into the next generation, along with a few more new mutations. Any population of reasonable size will therefore contain genetic variation, even if it’s a simple culture of bacteria grown from a single cell. In time, some genes in a population will be completely replaced by mutant versions. This is evolution.

It’s not just living things that evolve in this way. Any system in which copies are made, and then these copies are then copied, and so on, will evolve unless the copies are always perfect. An everyday example is language. Infants inherit language from the previous generation by copying the sounds adults make, and errors can spread, changing languages through time. It’s easy to identify words in different languages that have common origins, cognates. Indeed an evolutionary tree of indo-European languages was drawn before Darwin sketched his tree showing biological evolution.

During an interdisciplinary retreat at UCL’s Blakeney Point research station, a group of artists and biologists developed a painting method using some evolutionary ideas. We all looked at the same view, across the harbour at Pinchen’s Creek. We then each drew or painted the landscape for a set time, concentrating only on the view and without looking at our work until the time was up. This classic drawing exercise is of course quite error-prone. We exchanged our work with another participant and looked at it. We then repeated the exercise, responding to what was already on the paper, adding to it or correcting it, but looking only at the view, not at the marks we were making. Then we repeated this again until everyone had worked on every piece, and each time we reduced the time allowed, so that the final “generation” only lasted 30 seconds.

The paintings have some features of biological evolution. Images were inherited from one generation to the next, and the completed works were clearly related to each other. Each contained some features from every generation; but other features were lost underneath more recent marks. Overall we found the process very interesting and the finished works were quite satisfying. The process was not an exact analogue of biological evolution, but it is easy to imagine how other aspects of evolutionary processes—random mutation, competition, natural selection—could be explored in the production of artworks.

Duncan Greig, 2023

Photos and video by James Graham Keith