Thoughts on Computational Creativity

There is a close relationship between computational creativity in the arts and generative art. Prior work is summarized here as to how complexity theory can serve as a theoretical context for generative art. This is extended to show how complexity theory can also illuminate discussions regarding computational creativity. Another prior proposal is also summarized. It suggests that a world-view called “complexism” can reconcile the current differences between the modern culture of science and the postmodern culture of the humanities. This line is also extended to considerations regarding computational creativity. Particular attention is extended here to the issues of authorship, progress, and aesthetic measurement in computational creativity. Finally, under artistic license, I discuss some very speculative ideas regarding computational creativity that I explore in my current artwork. 1. Definitions and Theories Some find a sense of intellectual security from the somewhat expected procedure of first defining terms and then proceeding with theoretical exposition. In a usual academic context where there is a preexisting technical language that includes commonly accepted definitions this is good standard practice. However, in an unusual discussion that breaks existing paradigms, crosses disciplines, and ultimately requires the invention of new language, it’s useful to remember that neither word meanings nor theory can stand alone. Both are rooted in the other and inseparable. It could well be that in a full 1 The difference here being similar to Kuhn’s notion of “normal science” and “extraordinary science.” [1.] Dagstuhl Seminar Proceedings 09291 Computational Creativity : An Interdisciplinary Approach http://drops.dagstuhl.de/opus/volltexte/2009/2193 account of computational creativity its definition will not be the first thing presented. It may, in fact, be the last. And that is the case here. II. Defining Generative Art via Complexity Theory Closely related to the topic of computational creativity in art is generative art. It is, among other things, where those in the arts study and put to work presumed computational creativity. And in the field of generative art some standardization of the language, and corresponding consensus around theory, has occurred. But even within generative art there is controversy. Definitions, i.e. theories, of generative art in circulation include: • Generative art involves the use of randomization in composition. • Generative art involves the use of genetic systems to evolve form. • Generative art is art that is constantly changing over time. • Generative art is created by running code on a computer. In a previous paper [2] I introduced a theory of generative art that offered what is now likely the most widely cited definition of generative art. In a more recent paper [3] I didn’t modify the underlying theory, but I did offer the following to disambiguate a few aspects of the previous definition: Generative art refers to any art practice where the artist cedes control to a system that operates with a degree of relative autonomy, and contributes to or results in a completed work of art. Systems may include natural language instructions, biological or chemical processes, computer programs, machines, self-organizing materials, mathematical operations, and other procedural inventions. This theory of generative art casts a very wide net that is independent of any particular past or future technology. By including systems such as symmetry, pattern, and tiling, generative art is theorized to be as old as art itself. This view of generative art also includes 20th century chance procedures as used by Cage [4, 5], Burroughs [6], Kelly [6], Duchamp, and others. Along with notions of symmetry, this helps to tightly bind generative art to the standard art canon rather than leaving it isolated as an awkward art world orphan. Next, if generative art turns on the use of systems, then the contemporary scientific program for the general study of systems, complexity theory, provides a context for the consideration of generative art. In brief, scientists such as Murray Gell-Mann [7] classify simple systems as being either highly ordered or highly disordered, and hold that complex systems exhibit a dynamic tension between order and disorder. This is sometimes called “effective complexity.”