Pattern Recognition: What kind of relationship exists between theory and practice?

Redmond Bridgeman and Jo Law

Paper presented at Eye Site: the 2005 Art history Association Australia and New Zealand annual conference held at University of Sydney on 1 - 2 November, 2005.


We have no future because our present is too volatile[....] We have only risk management. The spinning of the given moment's scenarios. Pattern recognition

William Gibson, Pattern Recognition(2003)1

(i) Introduction

This discussion focuses on an Australian Council funded research and development project (a joint project between Redmond Bridgeman, Jo Law, and Cameron Merton) on the use of algorithms to generate screen works. First, we put forward an argument concerning the theory-practice relationship2; we then talk about aspects of the project that provide insights on this issue.3

(ii) Algorithm

Our project focuses on algorithms as conceptual models, experimentations, and strategies in developing innovative screen works. Using both 'natural' (e.g. DNA, fractal geometry) and 'artificial' (e.g. programmable languages) algorithms to act as construction agents, we investigate how pre-ordered 'sets' can be strategically used to generate meaningful patterns. By patterns we do not mean visual similarity but coherent forms that make visible the otherwise invisible relationships and formations ranging across the tele-visual, political, social, cultural, and natural.

In their most basic terms, algorithms are procedures or sets of instructions - a recipe of sorts where the principal ingredients may be numbers or data. Numbers and data are 'objects' that can be organised or grouped into Sets by a common thread (e.g. characteristics like prime, or multiples of a numeral). An algorithm can also generate sets. The best-known manifestations are the Julia Set (produced by using the simple equation of z -> z2 + c) and the Mandelbrot Set (mapping of the Julia Set).4

'Algorithms means the rules of performing complex calculation by a sequence of simpler ones,' and iterations are a necessary part of algorithms.1 Our interest lies in how algorithms can be utilised to manipulate data or building blocks (such as audiovisual images) to construct new configurations of aural and visual language. Pattern creation (and recognition) is one of the central functions algorithms perform. Their transformative ability - taking one level of abstracted information and translating it into another - enables them to play a key role in the mediation and interaction between different types of information.

The use of algorithmic generative techniques has developed as a distinct genre in electronic art. Karl Sims's highly influential early work Galapagos (1997)6, for example, inspired by Richard Dawkin's Biomorth Breeder experiment7, uses a particular understanding of evolutionary theory to produce self-replicating images that 'evolve' according to human aesthetic selection.8

Some important distinctions to Sims's rule-driven work, with its reliance upon Dawkin's ultra-Darwinism, and our aims can best be described with reference to Canadian film-maker Hollis Frampton's, film Zorn's Lemma (60min, 1970).9 This film borrowed its title from Max Zorn's Lemma, a specific type of Set. A Lemma is a proposition that is assumed to be true in order to test the validity of another proposition. The first section of film is a reading of an eighteen-century Massachusetts elementary school lesson book called The Bay State Primer. The production of sets and subsets in the second part of the film is determined by a system of substitution and progressions ordered by the 24-etter alphabet of the English language used in the primer. The third section of the film is a reading from Neo Platonist Robert Groseteste's book On Light, or the Ingression of Forms, which argues light is the infinite form from which all being takes it's form. Frampton deploys Zorn's Lemma and Groseteste's philosophy as axiomatic structures to explore how within a subset of an open system, say the infinite frames of film, order is possible. By harnessing Set Theory's potential to structure abstract representation of perception's intersecting planes in infinite combinations, Frampton hoped to bring narrative and non-narrative montage structure into new configurations using image, sound and text as primary materials. For Frampton, Zorn's Lemma only acted as a procedural guide, or perhaps new paradigmatic frame, that would open up film's inherent capacity to act as a model for how human consciousness engaged with the world.

Frampton intention was not to use Zorn's Lemma to programmatically organize data. If he did so, his film would be illustrative in the same way that Sim's work is arguably a model of Dawkin-style genocentrism. Instead, Frampton was interested in shifting the field in which film operates within. Frampton's engagement with Set Theory was part of a broader strategy that explores how a materially focused practice can engender the emergence of new and infinitely creative works. Thus, his investigation of visual and auditory patterns is not simply a combination of 'choices' picked from sets of pre-existing data, but is genuinely novel.10 His aim was to create another pattern that established a new type of filmic structure distinct from Eisenstein's theory of montage, in which, as Ian Weiss argues, a central strategy is the use of antithesis to form juxtapositions of dialectical images that in turn create a synthesis that embodies the truth of dialectical materialism.

(iii) interactivity

With Frampton's intention in mind, we now look at a common understanding of interactivity to take our discussion further. In many contemporary 'art and tech' works, we 'interact' with computing devices: a mouse, keyboard, console, pressure pad, hidden camera, or HMD helmet to produce effects: visual, audio, even olfactory. Yet, much of this type of 'interactivity' pales against the simplest of experience of, say, going for a walk or chatting with a friend. An important reason for this failure is their dependence upon a sender-receiver model of communication. Privileging stimulus-response paradigms this type of interactivity is limited. Besides conditioning the audience to learn the right type of response and teaching artists how to bring it about, such works seldom offer any more than circular games that transform the art-public into, in Buck-Morss's words, 'smart-fans in on the joke, [....] playing the game.'11 A game that can only have so many inputs or causes and these can only trigger so many outputs or predetermined endings; it cannot generate anything new.

This kind of cause-and-effect model belongs to a Newtonian or classical view of the universe. It is most famously characterised by mathematician Simon Laplace's boast that 'give me the past and present coordinates of any system and I will tell you its future'.12 Laplace's claim is known to be fallacious.13 Systems around us, including biological, social, economic, and political are open and dynamic.14 Closed systems, in contrast, are relatively rare. Scientific models are inherently closed. This is because they are descriptive in nature: they define a situation and hence are useful for understanding how things work. Nevertheless, such descriptions are an abstraction from an actual state of affairs. As an abstraction, the future of such closed systems is calculable in advance, but the living world is an open system therefore is unpredictable as it is always according to Henri Bergson 'continuously elaborating what is new.'15 If we follow a cause-effect model, we arrive at a closed system with limited possibilities for creating the new.

Bergson argued that such logic is the logic of retrospection, one that allows us to connect past events to the present like joining up dots to form a line. We misconceive this as an ability to locate a cause for every effect.16 When we attribute concepts or reasons retrospectively (for example, identifying unifying themes in art or literary movements), we should not mistake these as necessarily existing prior to our attribution. Our isolation of what is significant in the past is driven by what is important in our present. Bergson writes, 'The premonitory signs are therefore, in our eyes, signs only because we now know the course, because the course has been completed.'17 Likewise, if we assume the theory-practice relationship to be one of linear causality, we become trapped in finding undeterminable set of patterns in a sea of dots. In the worse case scenario, when practice illustrates a simplistic account of how something may occur and theory affirms its own assumptions by using practice as its evidence, the relationship between theory and practice is then defined by a circular self-affirmation.18 Theory is allowed to in-form practice by supplying the form of practice. How then are we to regard this relationship?19

We can look to the context of the event but the event is not explained by its context.20 In pursuing the 'origins' of the art event, we may find the particularity of circumstances that allows the event to emerge. These 'origins' may simply be thinking that provide the necessary context for creative practice to happen.21

Our project is concerned with the dynamics of the relationship between theory and practice. This dynamics is by nature the context of our project. The generation of new meaning and forms from algorithmic structures is a form of active chance driven engagement (as opposed to linear cause-effect interaction). In activities ranging from music re-mixing, visual sampling, open-source coding, and wiki compilation, the notion of transmogrification at the level of re-use, re-interpretation, and re-creation is pivotal. Akin to Frampton's use of Set Theory, our use of programming code is inherently coupled to the concept of creative reiteration. Such code not only responds to inputs by 'translating' them into 'actions' but it recursively reconfigures in a way that allows emergence. This idea echoes Swiss cultural theorist Vera Bulhmann's definition of complexity as necessarily involving 'the emergent phenomena and the open-mindedness of the future'.22 Buhlmann argues that creating using 'interactivity' must also provide audience with open-endedness that allows emergence of multiplicity (rather than pre-determined dead-ends). We thus borrow algorithms from the open systems of the living world that are 'continuously elaborating what is new.' 23

(iii) practice

Before showing our work-in-progress, we would just like to take a brief look at two contemporary artists' works to illustrate the diversity of such practice.

Video artist, Camile Utterback, investigates the relationship between the real, conceptual and virtual. Working at the level of computer programming and electronics she produces innovative and lyrical works that engage the viewer both on an emotional and conceptual level. Her works, generate self-reflective experiences that bring to the fore a contemporary engagement with the body in their environmental context. Utterback's work is an exemplary practice that opens up the possibility of a non-passive truly interactive experience.

They rule24, by artist/ designer Josh On, builds on the ground explored by Hans Haacke.25 On's online work allows visitors access to information concerning the board of trustees of 100 corporations (the so-called 'Fortune 100'; the 2004 version extends this information to government organizations and other public institutions). When we select a company, its board appears; we call on the board of directors, they promptly (or even obligingly) enter the picture; we further investigate each of these men's and women's memberships to other boards of trustees, these then appear; ad infinitum. After several iterations, geometry and patterns emerge. Like Haacke, On is interested in making the hidden relationships in social systems (more specifically, social geometry) perceptible.26 Pattern recognition here results from the selection of information, the structuring of data, in relation to the users' conscious (or unconscious) actions.

Our project investigates how algorithmic structures, as pre-existing ordering principles, can be deployed to transform and expand how experience is depicted. The works more explicitly explore how the relationship between sound and image can be mediated by code. It aims to instigate collaborative processes that allow programming and the material realm we inhabit to converge and produce truly creative forms.

(v) conclusion

It is futile to frame the theory-practice question within linear causality. Causes may be numerous and, as Bergson argues, are attributed retrospectively. Finding causes to effects lead us round a tautological loop. This retrospective logic prevents us from perceiving or even conceiving the possibility of something as new (reasoning that its cause must have existed in some forms prior). Instead if we focus on relations, we recognize the differences and interchange between description and actuality. More importantly, we allow the new to emerge.

It is openness to situations and the active engagement that expand art's capacity. Brian Massumi advocates an active politics that pays attention to the present and movement:

Experiencing [the] potential for change, experiencing the eventfulness and uniqueness of every situation, even the most conventional ones, that's not necessarily about commanding movement - it's about navigating movement. It's about being immersed in an experience that is already under way [...] It's more like surfing the situation, or tweaking it, than commanding or programming it. 27

We see this 'not programming' as a crucial part of creative practice.This 'not commanding' can be taken as a 'not scripting', 'not storyboarding', or 'not prescribing theory or be prescribed by theory' approach in practice. An open-ended 'make it up as it goes along' approach allows both theorists and makers to fully immerse in the work and respond to the situation of making at hand, and thus allowing new things to happen.


End notes

1. William Gibson, Pattern Recognition, Penguin/ Viking, Australia, 2003 p. 57 back

2. Yet, it has a history that needs to be unpacked rather than just invoked. I'm not going to attempt such a feat now. But one aspect of it's history illustrates the type of magic it produces. The Institutional and Art world definitions of art turned out to be, in part, simple the end of the usefulness or at least the limits of a particular tautological descriptive theories. Danto' observing that from the perspective of his own definition that art has disappeared into tautological vacuum announces the death of art and hence the death of theory. Dickies, and Danto's though are often understood as not descriptive attempts at the necessary and sufficient conditions but as proscriptive limits. A tautology definition, mistaking it's descriptive function for it's object, discovering it's own limits declares it's descriptive object in danger of disappearing. Theory is declared dead so that art can arise again. Yet the result is the triumphant return of theories that go unrecognised such as the notion of unmediated natural expression articulated by the late Romantics. Theories that have been around so long that they are hardly recognized as such. back

3. To begin with, I use the grant application process to highlight a significant problem when discussing theory and practice. In our written grant application to the now defunct New Media board, as one does, we talked about technology; we used obligatory words like: innovative, new-media, interactivity, excellence etc., words that have come to imply everything and nothing. They act as codes that set things in motion. And you tend to be rewarded to the extent you invoke the codes in a reasonably novel configuration. This is the art of 'grantsmanship'; it is more about casting a spell than it is actually saying or doing something. back

Theory with a capital T has also often operated like a spell. It is, or was, like one of those magic words that function, as Adorno argued, as a petrified formula: a sign that does not describe anything, it only invokes an abstract disposition. We are told Theory is at an end. No one seems clear about which theory, or quite when it happened. One version of its death claims the range of ideas loosely grouped under the heading of post-structuralism no longer has a productive relationship with practice. But even if the relationship between this thing called Theory, whatever it was or is, and practice has fallen by the wayside or come to the end of its usefulness, thinking goes on. And for us, one of the most important questions is critically thinking about the difference between effect (as in cause and effect) and affect, which as we describe below is what our project is essentially about.

4. As an aside, the mathematician, Roger Penrose, argues that the Mandelbrot Set is a part of nature and therefore Mandelbrot did not so much invent the set but discover it. This relates to our project in that we attempt to locate invisible patterns that may already exist. back

5. Nigel Lesmoir-Gorden, Will Rood and Ralph Edney, Introducing Fractal Geometry, Maryland (USA): Totem Books, 2001, p 55. back

6. See back

7. See back

8. Full references for both works*back

9. Allen S. Weiss, "Frampton's Lemma, Zorn's Dilemma", October, vol. 32, (Spring, 1985), p. 120. "Hollis Frampton's film Zorns Lemmais structured according to a twofold axiomatic system. The first axiom is indicated by the film's title, which refers to mathematical set theory: Max Zorn's Lemma. The maximal principle: If T is partially ordered and each linearly ordered subset has an upper bound in T, then T contains at least one maximal element. The second axiom derives from the mystical philosophy expounded by Robert Grosseteste in On light, or the Ingression of Forms, which offers a combination of neo-Platonic and Aristolian philosophy to a express a theology, ontology, and cosmology of light. A section of this text is read in the third part of the film. The mathematical axiom is operative in the alphabetical order of the text; the theological axiom is operative in the biblical content of the text. back

10. Like many complex systems in nature, simple non-linear formulae can create infinitely dissimilar and unpredictable results. back

11. Susan Buck-Morss, Thinking Past Terror: Islamsim and Critical Theory on the Left, New York and London: Verso, 2003, p. 84. back

Cited in Ziauddin Sardar and Iwona Abrams, Introducing Chaos, London: Totem Books, 1998, p. 12. back

13. We fail to predict even the positions of three bodies in motion - more famously known as the three-body problem. back

14. Chaotic systems in particular are at once deterministic and unpredictable. They can be summed up by simple algorithms but their exact outcomes cannot be predicted. back

15. Herni Bergson, The creative mind: an introduction to metaphysics, New York: Citadel Press, 1974, p 93. back

16. As also argued by Hume, the rendering of causal relationships is at best an inference (even in scientific models). If we watch the resultant action of one billiard ball hitting another, at best we can only infer one ball's movement is the direct result of another. No direct relationship is observable. back

17. Bergson, 24. This can be said to akin to 'pattern recognition'. It will also be interesting to consider this in contrast to Benjaimin's conception of origins and his constellation of history. back

18. We are not criticising theoretical positions used as underpinning frameworks for practice. back

19. Philosopher Peter Osborne recent discussion of the trajectory of Art &Language is useful here. According to Osborne Art & Language represents an extended investigation into the role of concepts and ideas in art via practice. Contemporary post-conceptual art (his term) is not about the redundancy of Conceptual art, but, amongst other things, represents the confirmation of the role of ideas and concepts in art. By their relentless investigations, Art & Language, Osborne argues, seemingly made themselves redundant by demonstrating the role of the aesthetic in art. The Post then in his terms confirms their discovery that art includes the aesthetic but is not confined to it. The question they asked, deploying theory as a critique, in the Kantian sense of the analysis of the conditions of practice, is what is the role of the concept in practice? A questioning which saw the causal relationship between theory and practice, with all its existing implicit assumptions, as the very thing that needs examining. back

20. Like a new truth that may be as perverse as a refusal to produce content for PDA's in a context where new media is seen as R&D for an abstract entity called the art industry. back

21. The Romantics saw the relationship between artist and artwork as a causal one. Their adherence of 'natural feelings' in art also ensured that ideas were banished from consideration in art practice. back

22. Vera Buhlmann, "Volatile Milieus: The Poetics of Interactivity", in Association Metaworx , Metaworx: Young Swiss Interactive, Basel, Boston & Berlin, Birkhauser , p.10. back

23. Bergson, 93. back

24. Josh On, They Rule, 2004. Retrieved from: back

25. Hans Haacke's 1971 Shalopsky et al Manhattan Real Estate Holdings, A Real time Social System, As of May 1, 1971, presented a powerful statement that greatly threatens the status quo (of the Guggenheim Museum's board of trustees) by bringing together photographs, maps and financial information concerning properties and real estate dealings in the lower East side in Manhattan. By presenting these data side by side, an image begins to emerge - namely the dubious practice of Harry Shapolsky's real-estate business in New York (Shapolsky being a well known friend of the Museum's board of trustees) back

26. A well-known example of this geometry is the 'six degrees of separation' phenomenon. back

27. Mary Zournazi, 'Navigating Movements, a conversation with Brian Massumi', Hope: new philosophies for change, Pluto Press, Annadale, Australia, 2002, 219. back