Why do we want our machines to seem alive?

(published Scientific American 150th anniversary issue, sept95)

The rollercoaster of technical and technological innovation is apt to disrupt a sense of historical continuity. The flashy futuristic rhetorics that surround these innovations as they are ushered into the marketplace almost always mask the presence of old and persistent motivations. These motivations shape western cultural practice on a fundamental level, informing activities in both the arts and in the sciences. The two impulses which I want to focus upon are the twin drives of mimesis and anthropmorphism. Although we consider mimesis and anthropomorphism to be concerns of the visual arts, I contend these drives cut through disciplines and find expression in the most advanced technology available at any particular historical moment. One of the most recent of these flashily clad ideas is Artificial Life, the generation of life-like behavior in the digital environment, utilising techniques which are modelled upon genetic and evolutionary processes. "Hard" A-Lifers hold that self-replicating digital organisms are alive in every sense. They argue that biology must include the study of digital life, and must arrive at some universal laws concerning "wet life" and digital life. Tom Ray, a biologist and designer of the Tierra system recently made a proposal to promote digital "biodiversity", a distributed digital wildlife preserve on the internet in which digital organisms might evolve, circumnavigating diurnally to available CPUs. He noted that "Evolution just naturally comes up with useful thingsâ" (Artificial Life IV conference, MIT 1994, personal notes) and argued that these creatures would evolve unusual and unpredictable abilities (such as good net navigation and CPU sensing abilities) and these organisms could then be captured and domesticated. According to the traditional christian view, we humans have a God-given right to harvest the products of biodiversity, the diversity of life on the planet. In a novel twist, A-Lifers seem to be harnessing the mechanism of biodiversity itself. The "seem" is critical here, it distinguishes between the "hard" and "soft" positions. Perhaps it is more prudent to say that they have built machines which simulate the functioning of a particular model of the mechanism of biodiversity.

In the arts, the drive to mimesis has most recently flourished in the field of interactive art. Since the availability of the desktop computer, a small community of artists has been exploring the possibility of a quite new artform, in which the esthetically manipulated quality was "behavior". The model for this behavior must necessarily be human behavior, the way in which humans interact with each other, or with other living things. We might model the behavior of a dog or a cat, but this would remain a human interpretation of that behavior, so the remains both mimetic and anthropmorphic. But much interactive art, along with most interactive technologies, has had a certain "pavlovian" press-the-button-and-get-the-prize quality to it, and the freedom to interact has always been within predefined pathways. Now Artificial Life techniques offer a quite new type of interactivity in which there is the potential for systems to respond in ways that have not been so explicitly defined. Unlike previous mimetc art practices, in this work the dynamics of biological systems are modeled more than their appearance. These works exhibit a new order of mimesis in which "nature" as a generative system, not an appearance, is being represented. Numerous new artworks employ biological growth algorithms, simulated ecosystems or communities, genetic algorithms, neural networks in the structure of the systems.

We are at a complex moment when digital techniques hold the promise of radical new mimetic capabilities which blur the boundaries between art and science. Simultaneously the new genetic technologies hold further "mimetic" potential. In genetic engineering, "nature" and representation, the "raw material" and the construction, become difficult to distinguish. The example of OncoMouse, the bioengineered mouse which was patented by Harvard University geneticists in 1988, and the many subsequent transgenic inventions, suggests the possibility that perhaps gene-splicing will become a new and valid artistic field, with prizes for the most esthetically pleasing transgenic forms!

If artificial life techniques are bringing new applications for biological models and metaphors into art, the entire project of Artificial Life is itself reminiscent of certain artistic enterprises of the turn of the century. The parallelism inherent in Paul Cezzane's early modernist dictum that "Art is harmony parallel to nature" is very like the goals of Artificial Life research, which Chris Langton has called "a biology of the possible". This makes Alife researchers "artists" in Cezzanes terms. Our discussion of mimesis is made complex by the intrusion of this goal of a parallel order. If the Alife researcher seeks a condition parallel to nature, this is very like the goals of the modernist artists following on from Cezzane, whose goals were not to represent the world, but to "render visible".

One of the major preoccupations of western art has been mimesis, the desire to create persuasive likeness. Although the modern period saw a move away from this idea in the fine arts toward various notions of abstraction, mimesis is the preoccupation of popular media culture: cinema, television, computer games. "Abstract" television is a rare thing indeed! For the fine arts, the prototypical mimetic moment is the story of Parrhasius and Zeuxis: "[Parrhasius] entered into a competition with Zeuxis. Zeuxis produced a picture of grapes so dextrously represented that birds began to fly down to eat from the painted vine. Whereupon Parrhasius designed so life-like a picture of a curtain that Zeuxis, proud of the verdict of the birds, requested that the curtain should now be drawn back and the picture displayed. When he realized his mistake, with a modesty that did him honour, he yielded up the palm, saying that whereas he had managed to deceive only birds, Parrhasius had deceived an artist." (Quoted in: The true vine, Representation and the Western tradition. Stephen Bann, CUP 1989 p27) Although we regard classical Greek sculpture as a high point of mimesis, at the time, I contend that the static nature of sculpture was not regarded as an esthetic requirement, it was purely a technical constraint. The Greeks stuccoed and painted their sculptures in a highly lifelike manner. Although this fact is well known in art historical circles, curiously there has been no attempt to re-polychrome the friezes of the Parthenon or the sculptures of Praxiteles. My guess is that if the greeks could have made soft fleshy sculpture, they would have. It is important to note that Hero of Alexandria was renowned for his pneumatic automata which combined static sculptural mimesis with human-like (if repetitive) movement. The famous clockwork automata of the C17th were capable of much more complex behavior than the Heros' pneumatic automata. The "scribe" by Jacquet Drosz could dip its pen and write lines of elegant script. Vaucansons famous Duck is said to have been able to flap its wings, eat, and with a characteristically duck-like wag of the tail, excrete foul smelling waste matter! It is of note, not simply that these works were contemporary with the first programmable device, the Jacquard weaving loom, but also that their behavior was constructed from mechanical "logic" much like that which Babbage used for his difference engine. We should further note that these automata were not regarded as fine art but simply as amusements.

The industrial era equipped the automaton with reliable structure and mechanism and the possibility the autonomy of untethered power sources, first steam, then electric. The image of the mechanical man became a cultural fixture, and literature was populated with a veritable army of mechanical men (and women) from pathetic representations like the tin man in the Wizard of Oz, to the mechanical girlfriend of Thomas Edison in Tommorow's Eve by de L'isle-adam, and the chillingly distopic portrayals of Mary Shelley's Frankestein, Fritz Lang's Metropolis and Karel Capek's RUR (Rossum's Universal Robots), the dramatic work in which the term "robot" originated.

It was the move into the electronic that began to offer the automaton first reflexes, then a modicum of intelligence. In the context of this historical trajectory, we must consider Artificial Intelligence as a continuation of this anthropomorphic and mimetic drive. Indeed Alan Turing defined the entire project as anthropomorphic with his test for artificial intelligence, now referred to as the "Turing Test". Simply put, this test says that if you can't tell it's not a person, then it has human intelligence. Though the project of Artificial Intelligence (as it was originally conceived) has run up against certain limitations, it has spawned other enterprises which are of great interest in this context. They are the creation of autonomous agents and artificial life, both of which perpetuate these drives. What I hope to have suggested is that certain ideas such as anthropomorphism and mimesis run very deep in western culture and these ideas continue to structure human enterprises into the present. Lack of space prohibits discussion of some of the provocative questions that arise from this suggestion. Are the mimetic and anthropomorphic drives specific to western culture? In what forms do they exist in other cultures? Nor have I conjectured at the origins of this drive. Why do we want our machines to seem alive?