Stepping Stones in the Mist

Copyright © Paul Brown 2000
All Rights Reserved

This essay was written for Creative Evolutionary Systems edited by Peter Bentley and Dave Corne to be published by Morgan Kaufman in late 2000 or early 2001. It is based on the presentation I made at the First Iteration Conference in Melbourne, December 2-6, 1999.

On my approach as an artist - a disclaimer
Major Influences
Historical work - 1960’s and ‘70’s
Early computer work
Recent work
Current & future directions


This essay is a idiosyncratic and non-rigorous account of my work as an artist who has been involved in the field now known as Artificial Life for over 30 years. To give the reader some context I begin with a few opinions that define my position within the visual arts (which is far from the current mainstream) and then go on to describe early influences from the 1960’s and 70’s that have framed my involvement in the field of computational arts. This includes some examples of my work from this period. The latter part of the essay describes my working methodology and includes examples of my more recent work and ends with a some speculations about where I may go in the future.

The title is a metaphor for my self view as an artist, and individual. A long time ago I stepped off the bank of a misty river or lake and onto a line of stepping stones. Now, many years later, the stepping stones are shrouded in the mist. Those behind me are dimmed by the mists of memory and those in front are hidden by the mists of uncertainty. The one in front of me is quite clear (as is the one behind) but then they quickly fade as they progress. I have no idea what lies on the further bank, or indeed if such a shore even exists! Memories of the bank I left are now long eroded.

I only really know where I am at this moment or, perhaps, where I have just been.

On my approach as an artist - a disclaimer

Thanks to my longstanding interest in computational systems as a medium for the visual arts I have been relegated to the fringes of the arts mainstream for most of my career. The role of outsider is one that I enjoy and I was somewhat disturbed when the global art mafia appeared to be acknowledging the computer-based arts in the early 90’s. In retrospect I had no need to fear. The mainstream’s adoption of this area of work was, and is, extremely parochial, one dimensional and, dare I say, paranoid.

Work that uses computer-aided tools (productivity enhancers based on traditional tools and methods) has been adopted and, for a brief time at least, became exceptionally fashionable. However the concept of the computational metamedium (to quote Alan Kay’s term KAY84) as a unique new paradigm for the arts quickly fell prey to the ...”no skills please - we’re postmodernists” kind of rhetoric that the international contemporary arts scene use to defend their position whenever it is threatened.

The paradigm shift that is foreshadowed by the computational arts is, quite correctly, perceived by the holders of the status quo as a significant threat to their jurisdiction. They are building barricades of rhetoric to shore up the crumbling foundations of their glass menageries. As I have commented elsewhere a revolution is in process in the arts although, given the extreme conservatism of the discipline, its cliquish nature and it’s lack of any kind of quantitative foundation, I have little expectation of a resolution in the near future. When I was 20 I expected that the revolution would be over long before the time I was 30. Now I’m 52 I will be surprised if it’s resolved in my lifetime. I nevertheless remain optimistic that it will be resolved, and in my favour!

I hold dear to many unfashionable concepts in this brave new postmodern world. A favourite is my belief that the culture vultures of the arts mainstream have completely confused postmodernism with their own brand of ultra-conservative late modernist rhetoric.

Here, however, are some more productive opinions:

  • the artistic mind is a “butterfly” mind that can fly from flower to flower, from source to source, with little respect for logic or scholarship. The result is a grand synthesis formed at an meta or pre-conscious level. From this it follows that:

  • the visual arts are beyond language, beyond conscious processing, at least when they are created.

    note 1: The studio/production basis for the visual arts often means that they sit uneasily within scholastic institutions like Universities where their studio component is often undermined in favour of theory. This is partly due to economic rationalism (studio is one-to-one and expensive, theory is one-to-many and cheap) and partly due to the academic pressures of the university tenure and promotion policies (which favour theory.

    From this we can expect that:

  • theory does not (necessarily) inform creation although creation, of necessity, informs theory.

    In believing that living art (art that is still in the process of being created) is by definition beyond the linguistic mechanisms that critical theory demands I nevertheless acknowledge that dead art (like Dada which is a complete corpus and cannot now be modified) is susceptible to theoretical deconstruction and analysis. Perhaps Dada can be totally described by language though I suspect that Godel’s concept of incompleteness will apply here as it does in any defined and rigorous domain. If we want to learn about Dada it will probably help us if we look at some Dadaist artworks although I’m not convinced this is absolutely necessary. If we want to learn about a living art process them we are obliged to look at and/or interact with the work. It is the only portal for understanding that we possess.

    I also suspect that there may be analytical tools, like Charles Saunders Peirce’s Semiology or George Spencer-Brown’s Laws of Form which may have more success in their application to the living arts. My suspicions are based on the intuition that these tools may share, or overlap with, the metalinguistic domain of visual artistic creativity.

  • the medium informs the work and: skill with the medium determines the quality of the work. This is a very unpopular point of view at present and considered a legacy of high modernism’s ... “truth to the medium”. I challenge all critics to write a poem in an language with which they are not familiar that a native speaker of that language would consider as acceptable. It doesn’t have to be good, just acceptable. Random “cut-ups” and computer translations don’t count.

Major Influences

I was recommended to read Anton Ehrenzweig’s “Hidden Order of Art” soon after it was first published in 1967. As a young art student it meant little to me and it wasn’t until I had become interested in system or procedural art in the early 70’s that it made much sense. Ehrenzweig was a psychoanalyst who has been credited, by Anthony Storr, with being responsible for the major revision of the discipline that redirected it from the mainly pathological focus of Freud to a more creative and celebratory emphasis. I suspect that Ehrenzweig may himself (had he still been alive) have credited Marion Milner with this transition.

His first book “The psycho-analysis of artistic vision and hearing : an introduction to a theory of unconscious perception” is a flamboyantly unstructured and almost unreadable cry of “eureka”. “Hidden Order” if he had lived to complete it would have been his masterwork. It was published soon after his death in an unfinished format, unindexed and after a flurry of interest disappeared from view before reaching a second edition. After a long hiatus it reappeared and is still in print from University of California Press.

Ehrenzweig overcame the significant problems that had undermined Freud’s analysis of the overt content of the artwork by analysing instead its structure and, in particular, the structure of the creative process itself. This was of course a period when Abstract Expressionism - random or “subconscious” mark making - was the dominant model for the visual arts. Ehrenzweig proposes three major stages in the creative process: an initial rejection of unwanted or repressed material followed by; an “oceanic” engagement and synthesis of the material and; a final reintegration or “reification” of the material at a conscious level.

I read “Hidden Order” at one sitting and remember my excitement and agitation when I had finished it. Within a couple of hours I had devised a procedure which I though may be capable of “testing” Ehrenzweig’s hypothesis (fig. 1). It involved replacing Ehrenzweig’s initial rejection stage with a system for positioning tiles according to the output of a random number generator - a dice! This appealed to me because I was suspicious of all references to a subconscious and then, as now, was extremely sceptical of the concept of art as self expression (at least in the emotional sense of the phrase).

Figure 1:

A recent reconstruction of the first image I made using the technique devised after I had read Anton Ehrenzweig’s “Hidden Order of Art”. It used octagonal tiles and the four orientations were dictated by flicking the pages of the book and using the last digit of the page number modulo 3 - I didn’t have a dice at that time! The black squares indicate holes.

My conscious motives (as I remember them) were involved with issues that obsessed me at the time like removing myself from the work and objectifying the art making process. Issues that had, in that period of history lead to Minimalism, Conceptual Art and Art Language. In retrospect my achievement was the establishment of a personal methodology for creative production that has governed my work ever since.

Around that same time I was recommended George Spencer-Brown’s “Laws of Form”. I found it’s clinical notation intimidating and, in consequence, I read a teach-yourself format book on symbolic logic. On returning to the “Laws” I discovered my concern had been unnecessary. Spencer-Brown leads his readers step-by-step through his calculus and towards his conclusions. More recently it has been described to me as a “boundary grammar” since it deals with the ideas of distinction and of crossing.

Although I remain convinced that “Laws of Form” is one of the most important books I have ever read I am not aware of any direct influence it has had on my work. I remain surprised about this and often carry a copy around with me so I’m ready for the breakthrough when it occurs! Spencer-Brown’s clinical and precise methodology has certainly influenced me. Ironically symbolic logic which I perceived at the time as merely a spin-off or perhaps more correctly a portal to the “Laws” has had a profound and on-going influence. This is probably because of my immersion in computational methods and their foundation in logic and formal languages.

Another book which was a major influence at the time was Charles Biederman’s classic “Art as the Evolution of Visual Knowledge” (BIE48).

Since 1968 I have been fascinated by the structure of an classical Chinese text called the “I Ching or Book of changes” (WIL23). It is a two state system of broken (yin) and unbroken (yang) lines that influenced Leibniz who developed the European version of binary notation (LEI66). The basic “word” is a three bit trigram. The eight trigrams are multiplexed together to form six bit hexagrams which index the 64 chapters of the book. Via changing lines (bits that can flip from yin to yang) any chapter can change to any other and so the combinatorial permutations of the book total 4096. The book is believed to have first appeared around 1800 BCE. It is possible to interpret the book as a symbolic cosmology which derives the three dimensional cartesian universe by repeated subdivisions (the trigrams) and then populates it with agents (the hexagrams). This process is echoed in the opening stanzas of the “Tao Te Ching”:

One gives birth to two - the yin and yang
Two gives birth to three - the trigrams
Three gives birth to the myriad creatures - the hexagrams

In the 1970’s I also became aware of the work of the System Art Group. Several members taught at the Slade School of Art where I was to study from ‘77 to 79. None of them, at that time, used computers and although I cannot think of any direct influence their work has had on me it was certainly reassuring to meet others with a similar mindset!

In 1970 I read Martin Gardiner’s column “Mathematical Recreations” in “Scientific American” (GAR70) where he described John Horton Conway’s “Game of Life”. For several months I persevered with large sheets of graph paper layed out over the floor of my home. Pencil, paper and eraser were too limited and I had to wait four years until I found my “ideal” tool - the digital computer. However this initiated my fascination with cellular automata (CAs) that continues to this day.

The final influence I will mention from that brief but formative period between 1968 and 1972 is the exhibition Cybernetic Serendipity which was curated by Jasia Reichard and held in 1968 at the Institute of Contemporary Art at it’s then new premises on the Mall (REI69). It was the first historical review of artists using computers. I was fascinated and returned to London in order to spend a second day at the show. Although I was attracted to the idea of working with computers my attempts to get involved didn’t come to anything. Then the Polytechnics were formed and it was possible for me to enrol, in 1974, as a mature fine arts student at Liverpool Polytechnic (now John Moores University) and then spend most of my time in the Mathematics Dept. learning Fortran (on an ICL 1903a) and in Engineering discovering PAL3 Assembler on their DEC PDP 8. Despite this, in 1977, I was awarded an first class honours degree in fine art.

Historical work - 1960’s and ‘70’s

As a young art student at Manchester College of Art in 1965 I had the choice of being pigeon-holed as either a painter, sculptor or printmaker. I survived three years of boredom before dropping out to co-found the lightshow Nova Express which toured the North and Midlands of England for several years. In addition to the leading bands of the day like Pink Floyd, The Nice and Canned Heat, we also played with contemporary music, dance and performance groups like “Music Electronica Viva” and “Meredith Monk and The House”.

About that time John “Hoppy” Hopkins introduced me to video and, in collaboration with the musician and composer Michael Trim (who had one of the first AKS digital audio synthesisers) and an engineer (whose name I am sorry to admit I have forgotten) we made several primitive video synthesisers. These worked on principles of feedback and utilised both electronic and mechanical components (like rippled glass filters). Although most of our work was intended to be played live we did make a few tapes and one “Mandala” was included in the UK’s first major retrospective of video art - The Video Show at the Serpentine Gallery in 1974.

My years in the lightshow and video were a revelation. Contrary to my training, which had demanded “meaning”, “significance” and “context” I discovered that a simple feedback circuit or some oil and water together with a few dies and some heat could produce large scale immersive experiences that were, to me at least, a lot more attractive and interesting that the stuff on the walls and floors of the trendy galleries. A significant insight was my redundancy as the creator of these works. I could show someone else (who didn’t need to be an “artist”) how to do it. Or, possibly, I could build a machine to do it!

Being of a logical disposition I realised I needed to find a formal method of codifying and creating work of this kind as a systematic procedure. That’s why the experiment with tiles that followed my reading of Ehrenzweig was such a revelation. But my interest, some might say obsession, with tilling systems has a longer history. Back in 1967, whilst still an undergraduate student at Manchester I had produced a series of tile drawings (fig. 2) that were probably the first pieces of work that I ever made which I considered to be significant in the sense of “originality” or of being unique to me as their creator. I equate this to the idea of “personal signature” in the emotive or self-expressive arts.

Figure 2a:
a simple square tile has it’s corners labelled with the ordinal symbols 1-4 in clockwise order.

Figure 2b:
a two by two arrangement of the tile. The peripheral vertices are now labelled with the symbols 1-4 in a clockwise zigzag pattern. The internal shared central vertex repeats this pattern in an anticlockwise sense. Note the emergence of new symbol relationships at the adjacent edge vertices.

Figure 2c:
a simplification of figure 2b showing only the peripheral vertices and the zigzag pattern.

Figure 2d:
the tile from fig. 2c can now be arranged in a similar fashion to that in fig. 2b. Note how this restores the clockwise order of the symbols in the peripheral vertices. Also that the inner shared vertex repeats this order but in an anticlockwise sense. Compare the relationship of the adjacent edge vertices with those in fig. 2b.

Note also that fig 1a is a simplification of this arrangement. This demonstrates that applying the same arrangement procedure twice returns the peripheral and central vertex labels to the same state.

Figure 2e:
here we see the full expansion without simplification. It clear that this process can continue indefinitely. Every second expansion it will restore the initial conditions at the peripheral and centre vertices whist producing an expanding set of codes at the ever new intermediate vertices.

It would be ten years before I became acquainted with the work of Benoit Mandelbrot and “self similarity” and even longer before I first heard the term “emergence” used in the sense that we understand it today. However I’m now aware that back in 1967 I glimpsed these concepts in creating and studying these drawings.

Soon after I made these drawings I showed them to my lecturers who dismissed them as inconsequential and irrelevant and then suggested that I reconsider my career in the visual arts. Not long after this I dropped out and began working with the lightshow.

Six years later I enrolled in the College of Art at Liverpool Polytechnic with the express intention of learning about computers. After a few months learning FORTRAN and the graphics package Gino-F I began to develop a tile-based image generating system. Although I initially used a random number generator to drive the system I soon became dissatisfied with the simple equation of randomness with intuition. I recalled my earlier interest in “The Game of Life” and began to devise both deterministic and probabilistic CAs to create the input data for the system.

At Liverpool the painters were unsympathetic to my work and I transferred to sculpture. This department included several members of the 60’s Kinetics Group and they were very supportive and helped me develop a small digital electronics lab. I kept my head down and took my lecturers advice when they suggested I make some 3-D stuff for my final assessment.

Then in 1977 I began two years of postgraduate studio work at the Slade School of Art at University College London. The Computing and Experimental Department had been formed in 1974 by Malcomb Hughes (then head of postgraduate) and Chris Briscoe (who had begun working with computers as an undergraduate student at Portsmouth College of Art) together with an alumni endowment which helped procure a Data General Nova II minicomputer.

The late A-life pioneer Julian Sullivan was also on the staff and the late Edward Ihnatowicz, who had created the early adaptive robots “SAM - Sound Activated Mobile” and “The Senster” was a regular visitor. Harold Cohen, who was then working on the early version of his drawing automaton “Aaron” visited whenever he was in the country. Darrel Viner, who had been working with the computer graphics pioneer Dr. John Vince at Middlesex Polytechnic since 1972, was also around. The place was a magnet for artists working with computers and generative systems. Many of them were involved with automata or other procedural or rule-based systems and we were all fascinated by the area that would later be called “Artificial Life” or A-life.

I also joined the Computer Arts Society which had been founded in 1968 at Event one at the Royal College of Art. Meetings were held at the late John Lansdown’s offices in Bloomsbury Square. John became a friend and mentor who, after seeing some of my work invited me to look into the dynamic generation of unique foliage drawings for use on CAD architectural plans and perspectives. This project introduced me to Mandelbrot’s work on fractal, iterative and non-linear systems - the area that has now been dubbed “Chaos Theory” (MAN77). John published a brief overview of this work-in-progress in his “Not Only Computing - Also Art” column in 1978 (LAN78).

It’s interesting to note that by the mid to late ‘70’s the Chaos field was well populated by scientists (who were nevertheless often working “underground” to protect their career status) and by artists. By contrast the nascent A-life field was then almost exclusively the domain of artists.

Early computer work

The system that I developed then refined first at Liverpool Poly. then at the Slade School consisted of three components: a cellular automaton; a graphics interpreter and; a back-end display generator (fig. 3)

Figure 3:
Schematic of the CA graphic system

Most often I have worked with simple CAs that are identical or similar to Conway’s “Life”. The automaton is based in a regular rectangular matrix where each cell of the matrix can have one of only two states. In general these states can be symbolised as “empty” and “occupied”. Such a matrix can be represented by a one bit array where 0 = empty and 1 = occupied. This array is the current time slice. The next time slice is calculated by applying a set of rules to each cell in the matrix. These rules examine the immediate neighbours of the cell (fig. 4).

Figure 4:
Rule 1 - If the cell is occupied and it has 2 or 3 neighbours occupied then it will remain occupied in the next timeslice

Rule 2 - if the cell is empty and has 3 neighbours occupied then it will become occupied in the next timeslice

Rule 3 - otherwise the cell will be empty in the next timeslice

When the rules have been applied to all the cells in the matrix the next timeslice replaces the current timeslice and the process is repeated.

If any readers are not familiar with Conway’s “Game of Life” then Poundstone offers an excellent introduction (POU87).

Conway’s rules can be implemented in a simple look-up table:

Current Timeslice Cell

Future Timeslice Cell State if the:

No. of Neighbours occupied

Current Cell is Occupied

Current Cell is Empty




























Since each cell has 8 neighbours the neighbourhood family consists of 256 members and we can also illustrate Conway’s rules using a series of state diagrams (fig. 5)

Figure 5 a:
a diagram showing all the possible neighbourhood states of the Life CA.

As in each of these state diagrams the top row and leftmost column outside the square are four-bit nybbles that address the columns and rows of the diagram.

Figure 5 b:
a diagram showing those neighbourhood states that allow a cell that is occupied to remain occupied in the next time slice. This corresponds to rule 1.

Figure 5 c:
a diagram showing those neighbourhood states that enable a cell that is empty to become occupied in the next time slice. This corresponds to rule 2.

Note that rule 3 is implied by the empty spaces in both b and c.

Edge conditions (where a cell does not have the requisite neighbours) are often dealt with by wrapping the array so the bottom edge is considered adjacent to the top and the left edge adjacent to the right. The finite rectangular array represented becomes equivalent to the continuous surface of a torus or doughnut. This arrangement is often referred to as “wraparound”.

For every timeslice the CA produces a single bit 2-D array of data as output. This array is the input to the graphics interpreter. The graphic interpreter has a fairly simple task: it controls the way that the symbolic input array maps to an actual graphic layout.

The mapping is one-to-one. For each cell in the 2-D bit matrix there will be a corresponding tile in an equivalent 2-D graphic matrix. For single bit input the mapping is simple (fig. 6).

Figure 6 left:
Single bit mapping to two different tiles.

Figure 6 right:
Single bit mapping to the same tile rotated.

Most of my recent work uses multiple bit arrays. Although I have worked a little with CAs that operate on multiple bits I have in general preferred to derive a multiple bit solution by integrating single bit arrays over time. For example we can consider an identical cell in two timeslices Tn and Tn+1:

Cell state in Tn

Cell State in Tn+1

Anthropomorphic Interpretation






Being born






Being alive

Two-bit mapping is illustrated in figure 7 and three-bit in figure 8.

Figure 7:
At two-bit state is mapped on to a family of four tiles

Figure 8:
At three-bit state is mapped on to a family of eight tiles

Time integration imposes an important constraint. For example the three-bit code 011 can only change to 110 or 111 in the next frame. The code shifts left one bit and then the CA provides the least significant bit.

Although in my early work I often mapped bit states onto a set of different tiles (figure 6 a) I have more recently chosen to map onto families derived from rotations and mirror rotations of a single tile (figure 6 b and figure 7 and 8).

The tiles I use have patterns on them which, in the final piece, dominate the visual appearance of the work. Many viewers are, in fact, surprised to discover the underlying tile matrix and the relative simplicity of the elements that make up often complex images. The idea of complexity emerging from simplicity -or- to use the older homily “the whole is greater than the sum of the parts” has been a guiding concept behind my work for longer than I can remember. I find myself equally attracted to holism and reductionism and constantly oscillate between these two extremes.

By patterning the tiles it’s possible to explore ambiguities like those illustrated in fig. 9.

Figure 9:
In the top diagram the area tagged A is most likely to be read as the positive or foreground space whereas in the bottom one it is more likely to be read as the negative or background space.

The negative space has been hatched in both cases. In this way it is possible for the work to explore and/or reveal features like: boundary; closure; inside; outside; negative; positive; foreground; background; inversion and; crossing. Readers who are familiar with Spencer-Brown’s “Laws of Form” may now share my astonishment that there it has had so little direct influence on my work despite the fact so many similar guiding concepts are shared!

It’s also necessary for me at this point to issue another disclaimer! In exploring concepts or features like these I am not trying to understand them in the way that a cognitive scientist might attempt to do. Nor am I trying to create models or simulations that help us understand creative behaviour or perception. As an artist I am simply exploiting such concepts, exploring them and mining them. This is not to say that I don’t, at least at some level, understand them but that such understanding is not perceived by me as being of any particular relevance to the production process of my work. It is not intended to be an illustration of such concepts however it’s quite legitimate for me or for others to interpret the work in relationship to these concepts.

Recent work

In 1986 I gave up my PC-AT and the use of mainframe and minicomputers for an Apple Macintosh and this has been my preferred platform ever since. Around 1988 I became aware of a software application called VideoWorks that was subsequently renamed Director and has since become the de-facto standard for multimedia authoring worldwide. Director V7 is a powerful object-oriented fourth generation application generator that allows relatively naive users to create sophisticated products for CD-ROM and the Web. For more advanced users it supports Lingo, a well featured object oriented programming language. Since the mid 90’s it has been possible to create and support products on both the Mac and PC systems.

All of my recent time-based works have been produced using Director. These include: Infinite Permutations V1 - a one bit system, 1993/94; Infinite Permutations V2 - a two bit system, 1994/95 and; SAND LINES - another two bit system completed in 1998. During 1999 and 2000 I have been working on a new piece, provisionally titled Chromos, which will be a three bit system. Samples of these works are included in the images/timebase section of my website.

Several of these works utilise pre-computed animation tables where, for example, one member of a tile family is inbetweened to each other member (fig. 10).

Figure 10:
Page 1 of the animation table for a new work chromos where family member 1 inbetweens via 10 stages to each of the eight members of the family (including itself on line 1)

More of chromos can be seen here.

These drawings of animation tables are so interesting in themselves that I’m planning to exhibit the work in an installation which will include a large format projection of the time-based component together with wall-hanging and book format working drawings and other related material.

In all of these time-based works I process the family of tiles using the filters and image processing facilities of Adobe PhotoShop in order to both colour and texture their appearance. Although this is sometimes merely decorative or playful it can be used more significantly to anchor analogical or structural references that are otherwise contained in or implied by the work. For example this mechanism can be used to both increase and decrease ambiguity or to lock into a graphic metaphor. In Infinite Permutations V1 the colour has been specifically chosen to make it difficult for the viewer to resolve the foreground/background or negative/positive conflict inherent in the pictorial representation of the work. In SAND LINES the image processing has been selected to consolidate the stone/sand identification and create a conflict with the animation which does things that the materials represented could not.

In 1992 as professor of art & technology at Mississippi State University I was able to use an Iris ink-jet printer for the first time and was amazed at it’s resolution, surface integrity and colour fidelity. More recently Iris together with third party suppliers have introduced a variety of archival inks that can print on acid-free watercolour papers. Since 1995 I have been making limited edition prints of images that begin as timeslices of large format tile automata like those used in the time based pieces. I often get lost in an oceanic orgy of image processing whilst producing these images and occasionally even undermine or disguise the CA foundation of the work. (See the images/prints section of this website for examples).

Current & future directions

At the time of writing I have just begun a year as artist-in-residence at the Centre for Computational Neuroscience and Robotics and the School of Cognitive and Computing Sciences at the University of Sussex. For some time now I have been following work in evolutionary computational methods and have become convinced they have an important contribution to make to my future work. The thick mists shrouding these particular stepping stones make it difficult for me to predict just how these processes may relate to my practice. I’m also concerned that prediction often prejudices the outcomes of a project and want to keep an open mind.

However I can foresee one way that this technology could be integrated into my current system and it’s likely that this will be the focus of my work for the coming months.

It would involve the addition of a pre-evolved or a dynamically evolving “observer”. This should be able to analyse: the bitmap array and/or; the symbolic graphic array and/or; the “actual” raster graphic display data for each frame or timeslice. It would then be capable of dynamically modifying: the generating CA’s rules and/or; the graphic interpreter’s mapping rules and/or; the display generators image processing filters.

This would be a learning system that would be capable of evolving certain kinds of graphic behaviours (fig. 12).

figure 12:
Proposed schematic for an evolving time-based system. Compare this with figure 3.

I indicated in my opening paragraphs that the far shore of my life’s work is completely invisible or may not exist! However as I have implied in the later text I do have some long term ambitions. The main one is to contribute to the development of autonomous creative behaviour. I look forward to automata that can create artworks that a peer group of either humans or other machines will accept as legitimate creative activity.

The singular and remarkable success of artists like Harold Cohen is his achievement in externalising his own personal creative drawing behaviour. Aaron, the automaton he has created, produces 100% genuine Harold Cohen drawings. I believe that we will eventually be able to create automata that will make artworks that do not bear the signature of the creator of the system. That the system will be capable of evolving it’s own personal style.

I leave the reader to ponder the problem of just what the words “own” and “personal” might mean in this context.

My own feelings are that the every growing realm popularly referred to as cyberspace is an ecosystem and that creatures must evolve to exploit that space. Since a primary fitness characteristic would be to hide from humans (who would almost certainly try to destroy them) they may already exist! Such entities would have access to an vast repository of information from both real-time sources and from archives. I find it inconceivable that creatures with such a broad bandwidth of input “sensations” will not develop behaviours that are analogical to human art making.


I would like to thank Dr. Phil Husbands, joint coordinator of the Centre for Computational Neuroscience and Robotics and Richard Coates, Dean of the School of Cognitive and Computing Sciences at the University of Sussex for the invitation to join their program for a year as artist-in-residence.

In particular I must thank Gavin Sade of the Communication Design program of the Academy of the Arts, Queensland University of Technology who recently turned my hacked Lingo code into a series of elegant modular objects.

I am especially grateful to the Australian Commonwealth Government and the Australia Council, it's arts funding and advisory body for their award of a New Media Art’s Fellowship which will fund my work throughout 2000 and 2001.

This essay is based on a presentation I made at First Iteration, a conference arranged by Jon McCormack and Alan Dorin at Monash University, Melbourne, December 1-3, 1999 (FIR99).


BIE48 Biederman, C, “Art as the Evolution of Visual Knowledge”, Red Wing, Minnesota, 1948.

EHR65 Ehrenzweig, A, “The psycho-analysis of artistic vision and hearing : an introduction to a theory of unconscious perception” : 2 ed, Braziller, 1965.

EHR68 Ehrenzweig, A, “The hidden order of art : a study in the psychology of artistic imagination”, Weidenfeld, 1967.

FIR99 First Iteration Conference,

GAR70 Gardner, M, “Mathematical Recreations”, Scientific American 223(4), October, 1970, pp 120-123

KAY84 Kay, Alan, “Computer Software”, Scientific American, September 1984 V 251 #3 pp 41-47.

LAN78 Lansdown, John, “Only God can make a tree” in “Not only computing - also art”, Computer Bulletin, British Computer Society, September 1978.

LEI66 Leibniz, GW, "De Arte Combinatoria", 1666

MAN77 Mandelbrot, BB, “Fractals : Form, Chance and Dimension”, Freeman, 1977

POU87 Poundstone, W, “The recursive universe: cosmic complexity and the limits of scientific knowledge”, O.U.P. 1987

REI69 Reichardt,J Institute of Contemporary Arts, “Cybernetic serendipity : the computer and the arts : Special issue of : Studio International”, Studio International 1969

SPE69 Spencer-Brown, G, “Laws of Form”, George Allen and Unwin, 1969.

WIL23 Wilhelm, R (Ed., Trans. German), Baynes, CF (Trans. English), The I ching, or, Book of changes / The Richard Wilhelm translation (1923) rendered into English by Cary F. Baynes, Princeton University Press, 3rd ed 1967.

Return to the top of page