Monday, 30 January 2012

Creativity + Complexity : Don't Let The Reductionists Grind You Down!

This is an expanded version of my invited Leonardo Thinks article Creativity + Complexity = Win Win

A couple in love walking along the banks of the Seine are, in real fact, a couple in love walking along the banks of the Seine, not mere particles in motion.
Stuart Kauffman [1]

The game of real life is many things; beautiful, complex, horrific and increasingly precarious. But the rules of the game are far from being completely understood. As a field of scientific endeavour, complex systems science offers the hope of better understanding ourselves and the world around us, and producing major advances towards solving some of the worlds key environmental, cultural and social problems. The problems addressed by complex systems are recognized as being hard, as evidenced by the limitations in reductionist science to make significant advances. This is a well known problem, and Melanie Mitchell states it succinctly [2]:

In spite of its great successes explaining the very large and very small, fundamental physics, and more generally, scientific reductionism, have been notably mute in explaining the complex phenomena closest to our human-scale concerns.

Complex systems is an emerging multidisciplinary science developing new ways of researching large, highly intricate, dynamical systems in diverse areas such as biology, physics, social networks, socio-technological systems, socio-ecological systems, economics, our environment, the list goes on… [1, 2, 3, 4]. The use of complex systems and creativity, especially in music [5] and art [6] certainly has a history. So why is what I am proposing different? Put simply, it is a tale of two outcomes where art and science are a truly unified practice.

What Creativity + Complexity proposes is broad and deep research focused on a plurality of outcomes (knowledge and artefact, perhaps embedded in the same “object”) using robust methodologies. I have developed a methodology for my own knowledge making and artistic creation [7, 8], but this is not meant to be a prescription for others, although hopefully an inspiration. 

In my experience, art and science as described above is a win win scenario. The production of both successful artworks and fundamental research has a clear benefit, where art/science research is a two-way street. Technology and science are useful tools for the creative, but by the same token, the creative process is equally beneficial for driving innovation, whether scientific or technological. As Peter Cochrane puts it [9] :

Industries of the past were about process, about constrained problem-solving in a slow-moving world. But that time is long gone, and today's companies have to deal with fast-moving technology and competition, and that demands creativity and unbounded thinking.

Great, where can I go to get involved in this exciting new and bold frontier? That’s the catch, where can you go? Apart from a small, but thankfully growing number of forward thinking and forward looking organisations, the cupboard is still pretty bare (though not completely empty). The present situation is summed up nicely by Stuart Kauffman [10] :

The two cultures, science and humanities, remain firmly un-united.

Clearly though, there are people out there doing this kind of art and science, but it is still a minority sport. This situation has to change, and should be the norm rather than the exception. My assertion is that creativity and complexity, as a combined endeavour with methodological rigour, has the potential to make our current and precarious game of life a win-win situation. It is high time for the two cultures to really unite; to play and win the game. Don’t let the reductionists grind you down!


But, we need to recognise the differences, and in my view one of the key things to bear in mind with art/science interactions is this: art is not (necessarily) about progress per se i.e. there is no requirement for progress in the arts, in the sense of technological or scientific progress. To be clear, yes there is progress in the technology used to make art, but art itself may not subscribe to the traditional idea of progress (in my view at least). The scientist (and some artists) may ask for proof or empirical evidence, to which I would I would suggest an example experiment: Go to a music store and a computer store. Only in the music store can you buy a commodity that is older than, lets say, 30 years. OK, leaving aside antique shops and computer collectors on eBay etc. new product is still NEW PRODUCT, not merely a newly manufactured item and it is deemed to be progress in that field of endeavour (computing). If this were the case in music, you would not be able to buy Bach, Mozart etc. This difference is fundamental.


Algorithmic based music underwent a paradigm shift over the last two decades of the 20th century with the advent of complex systems research. Complex systems such as cellular automata (CA) produce global behaviour from rule-based interactions of simple cells. The picture at the top of this page shows a 1D CA evolving over time. At first glance it seems somewhat paradoxical to have the Kauffman quote underneath.  Now consider the picture immediately above this paragraph. This shows the behaviour of 1D CA with the time history, much more detail is apparent when we consider it from this viewpoint. In terms of analysis of CA, this is just the tip of the iceberg, there are lots of other ways of studying their evolving behaviours. The whole is much more than the sum of the parts...

CA have a distinguished and esoteric history in computer science, from its foundation to their present day influence in Artificial Life as well as numerous other important disciplines. They are fascinating objects, producing more pattern than a single human is capable of observing within their own lifetime. The different classes of behaviour they produce, whether ordered, complex or chaotic, make them interesting to artists and scientists alike. This wide variety of behaviour represents an important generative tool for the artist. A simple representation of these fundamental CA behaviours applied to music is shown in the next image. 

There is twist in the tail with CA; chaotic behaviour dominates rule space, which has serious implications for application and investigation. Obtaining a variety of pattern for free is thus a challenge to the artist and scientist alike. CA are discrete dynamical systems in terms of space, time and values assigned to cells. The set of all possible global states of these cells is termed the state space. The set of all possible rules for any particular CA architecture is termed the rule space. A concise definition of CA is given by Andrew Wuensche and Mike Lesser [11] :

A cellular automaton (CA) is a discrete dynamical system which evolves by the iteration of a simple deterministic rule.

The task of assigning behaviour to a rule is known to be undecidable, but a number of approximations have been attempted. An extensive amount of research by the CA scientific community has been conducted towards producing behaviour prediction parameters to discern the structure of rule space. Unfortunately, as the size of the CA rule space is increased the total number of rules becomes astronomical and the amount of chaotic behaviour increases dramatically. This problem continues to engage the scientific community and is the subject of much debate. 

In confronting systems of such behavioural complexity for the purpose of art, the artist is placed in a possibility space of truly vast proportions. Given that the potential for random behaviour increases with rule space, choosing CA rules at random does not represent a successful artistic strategy, unless one is actively seeking randomness. This problem has great implications for the use of CA in both scientific and generative arts practice. Pioneering composer Laurie Spiegel states this issue succinctly [12] :

But a musician's mind does not work randomly when creating, and the vast majority of truly usable musical algorithms will probably turn out to be non-random as they are discovered, explored, and put into use.

In addition, Hal Chamberlin stated that the production of algorithmic data for musical control “may be highly ordered, totally random, or somewhere in between” [13]. I approached the problem of rule space structure from an artists perspective in the context of generative music practice. All CA behaviours are deemed “interesting” to some degree as defined by the compositional application. The music practice problem in this context is to find a mixture of behaviour from the overwhelming chaos. This is in contrast, but not opposition to, the scientific approach of predicting behaviours in order to locate complexity within rule space. 

The techniques I developed are based on my own unique extensions of CA theory, to provide empirical evidence regarding rule space structure. Simply put, I applied the principles of self-organisation to the rule space, an original idea and completely different from conventional approaches. A concrete and navigable graph structure for rule space can be created using CA state space graphs, which are called attractor basins. My initial investigations were done manually, by printing out state space subtree’s and examining the resultant rule groupings. A glimpse of the process is shown below. 

Much to my surprise I discovered that CA dynamics are perfect for self-organising structure within their own rule spaces.  A brief overview is given in my short paper for Leonardo Transactions [8] and in depth details are in my PhD thesis [14]. Generative music experiments have the capability to both produce music and inspire further development of complex systems research. The discovery of a connection between state space and rule space from my research into generative music, has implications for future work in both art and science. This will hopefully encourage interdisciplinary collaboration between the arts and sciences in the area of creativity and complexity. Detailed analysis of my results is ongoing and may provide further new insights into the wilderness of rule space. The underlying notion of my rule space structure methodology benefits from its own generality, and the method of creation is not dependent on any particular aspect of musical theory, e.g. scale, mode or chord. The artistic approach taken provides an interesting and alternative method of studying rule spaces of complex systems in general, independent of musical application. 

So much for theory, its always nice to have a listen to some music, so I'll end this section with a piece called Acorn. It is realised through a hybrid of mediums, analogue/digital synthesis and old/new computer technology. The CA algorithm (a 2D system for this piece) was programmed in BBC BASIC on an ancient Acorn RISC machine. Events in the universe are mapped to synthetic speech events (allophones), tones and digital noise. Sound output is further processed by an analogue modular synthesizer (Roland System 100M & Doepfer A100). A partial view of the studio setup is shown in the image above.The final production and mixing was made by Australian electronic music legend Garry Bradbury (formerly of Severed Heads). During 2005 Acorn was premiered in concert at the Australasian Computer Music Conference in Brisbane, performed live later the same year at both the 3rd Iteration Conference in Melbourne and the Electrofringe Festival in Newcastle. In 2010 Acorn was featured on the peer reviewed DVD accompanying the Winter edition of the Computer Music Journal [15].


Art and science are vitally important to each other, and for driving innovation. Reductionism can no longer be the only game in town, and complex systems research is a much need shot in the arm for 21st Century science. Creative practice methodologies and complex systems research are a key part of moving forward to sustainable planetary goals. An exemplar of my approach to creativity and complexity was outlined, and key references for digging deeper have been provided. 

Although not discussed here, another key exemplar of my approach is the Rainwire project, a more recent work in progress since 2008. Rainwire is an art science project aimed at investigating rainfall patterns on long wire instruments through environmental sonification. The recording below is a short example excerpt provided to Leonardo Transactions as supporting material for my first Rainwire paper [16]. 

For more information on the Rainwire project please see part 2 of this blog post Creativity + Complexity Part 2 : Rainwire 

Three web pages worth checking out in relation to the Rainwire work are :


[1] Kauffman, S. A. (2008) Reinventing the Sacred: A New View Of Science, Reason, and Religion, Basic Books

[2] Mitchell, M. (2009). Complexity: A Guided Tour, Oxford University Press

[3] J. Norberg and G. Cumming, (2008) Complexity Theory for a Sustainable Future, Columbia University Press.

[4] Crutchfield, J. P. (2009) The Hidden Fragility of Complex Systems—Consequences of Change, Changing Consequences, in Cultures of Change: Social Atoms and Electronic Lives, G. Ascione, C. Massip, and J. Perello, editors, ACTAR D Publishers, Barcelona, Spain, pp98-111.

[5] Burraston, D. and Edmonds, E. (2005) Cellular Automata in Generative Electronic Music and Sonic  Art : A Historical and Technical Review. Digital Creativity 16(3) pp165-185

[6] Edmonds, E, Brown, P and Burraston, D (Eds). (2005) Generative Arts Practice. Proceedings of Generative Arts Practice Symposium 2005. Creativity & Cognition Studios Press.

[7] Burraston, D. (2011) Creativity, Complexity and Reflective Practice, In Candy, L. and Edmonds, E. eds. Interacting: Art, Research and the Creative Practitioner, Libri Publishing Ltd. Oxford.
[8] Burraston, D. (2007) Fundamental Insights on Complex Systems arising from Generative Arts Practice. Leonardo Vol 40 (4), MIT Press.

[10] Kauffman, S (2006) Beyond Reductionism: Reinventing The Sacred 

[11] Wuensche, A and Lesser, M. (1992) The Global Dynamics of Cellular Automata: An Atlas of Basin of Attraction Fields of One-Dimensional Cellular Automata. Addison-Wesley. (Available as free pdf from DDLab)

[12] Spiegel, L. (1989) Distinguishing Random, Algorithmic, and Intelligent Music, Active Sensing 1 (3), p2

[13] Chamberlin, H. (1980) Musical Applications of Microprocessors, Hayden.

[14] Burraston, D. (2006) Generative Music and Cellular Automata. PhD Thesis, University of Technology, Sydney. Available to download, along with its included CDROM at

[15] Burraston, D. (2010) Acorn. Computer Music Journal, Winter 2010 Vol 34, No. 4, pages 91-104, MIT Press (Generative sound composition on DVD)

[16] Burraston, D. (2012) Rainwire: Environmental Sonification of Rainfall, Leonardo (Forthcoming)