zurück zum Artikel

Fiction could be understand as a form of artificial life. Perhaps it is the greatest globally networked AL experiment in history. Are fictional characters alive? Are they real? Starting point for a reality check of claims in the field of AL

Matthew Taylor [1]

In an article about virtual reality Ivars Peterson observed that the idea behind such computer-mediated environments is nothing new to readers of good fiction: "Successful authors use words to depict such vivid, compelling characters and settings that it's easy to lose oneself in these purely imagined worlds" (1992, p. 8). An analogy even more apt than virtual reality would be artificial life (alife or AL), since readers of fiction are not entering a preexisting interactive environment but generating it nearly from scratch. From a mere string of words (albeit words strung together with considerable craftsmanship) readers are able to create an intensely engaging dynamic simulation in their mind.

I propose that fiction is a form of artificial life, perhaps the greatest globally networked AL experiment in human history. This assertion is not particularly remarkable or original, as it is based on self-evident properties of fiction and on well known ideas that have been advanced by Richard Dawkins, Douglas Hofstadter and others.

What may be less obvious are some of the implications that can be drawn from the comparison. Fiction presents a peculiar but I think substantial challenge to the "strong claim" in AL - that many of its digital entities are in some real sense alive. The same claim could be made for fictional characters, but is generally (and I think sensibly) not. Thus fiction might illuminate some of the ontological, epistemological, and by loose extension even ethical questions that AL has raised: What is alive? What is real? How do we know it? How can we value human life when life is defined technologically?

Though what I propose is a "common sense" view, I hope it is not construed as an "anti-AL" argument, or a call to dig in our heels against unsettling philosophical concepts from that field. I am in fact an ardent lay fan of AL research, astounded by its achievements. Yet I think a reality check (or more fittingly, an "unreality" check) is in order, if only to begin sorting out some of the provocative questions that AL has tossed our way.

Fiction as Artificial Life

Qualities that define a successful AL simulation reflect similar demands facing writers of fiction, who know that fiction will not work unless it takes on a certain "life of its own." To this end, a writer of fiction constructs a code roughly analogous to an AL program or genetic sequence. The code gets itself copied and lies in wait for a "host" whose mental activity will translate the sequence of words into spatial images and animate imagined beings. Whenever a reader becomes even mildly engaged in a work of fiction, an entire world is created - in R. G. Collingwood's terms imaginatively recreated - in their mind (1965).

Each reader's re-creation of a work of fiction will be unique in ways that the writer cannot anticipate, and fictional characters can be plausibly considered to have some measure of autonomy. Readers can, for instance, carry on fantasy conversations with characters or speculate on their life beyond the story. Fictional characters can also do real work, changing the way people live, or perhaps, like the character Uncle Tom, precipitating wars or the radical reorganization of society. Like a computer virus - among the digital entities considered closest to being a bona fide living organism - a fictional character can be threatening enough to provoke safeguards and "disinfection" programs (censorship, book burning, exiles, executions).

Fiction replicates (through print, and more recently electronic media), adapts, mutates, and evolves- albeit in ways that are more Lamarkian (inheriting acquired characteristics) than Darwinian (by natural selection). Consider a genre like detective fiction. The idea itself is a simple one: a character solving a mysterious crime. Starting with Poe's ratiocinative tales in the nineteenth century, the genre explodes with vivid characters (Dupin, Holmes, Marple, Poirot, Miagret) that remain thriving, robust components of our collective imagination. There is also an evolutionary branching into sub-genres (e.g. the "hard boiled" detective story). We see fiction, then, replicating on all levels, from a memorable phrase or speech (e.g. "Out damned spot!" or Hamlet's soliloquy), to a character or character type (e.g. Heathcliff or Holmes), to an individual work, to a technique or style (e.g. epistolary novel or stream of consciousness narrative), to a genre, or a whole class of genres (e.g. the detective story, above).

Memes

When we talk about ideas that replicate in culture we are talking about memes, a concept that was advanced by Richard Dawkins [2] and defined as "a unit of cultural transmission, or a unit of imitation" (1978, p. 206). Chris Langton defines a meme as "an idea, joke, poem or tune that spreads throughout a population by being copied again and again as people pass it on to their friends" (1992, p. xi). Erich Schultes (1992) created an interesting demonstration of a meme: a self-referential paragraph [3] that entreats us to copy it and pass it on to others. Memes have even been treated fictionally in Neal Stephenson's "Snow Crash", making that novel, like Schultes' paragraph, a meme about memes.

The landscape in which memes replicate, interact, adapt, and evolve Douglas Hofstadter calls the ideosphere, or collective cultural imagination (1985). (The "ideosphere" is to the "biosphere" what "memes" are to "genes.") Dawkins sees memes as competing for human attention, time, and memory (much the way Thomas Ray's AL creatures compete for computer memory in his Tierra [4] program) as well as things like "radio and television time, billboard space, newspaper column-inches, and library shelf-space" (1978, p. 212). This "selfish meme" seems to reach into the physical world and employ it for its own ends. In Daniel Dennett's words, "A scholar is just a library's way of making another library" (1995, p. 346). That is, we are merely the unwitting host mechanism memes use to multiply themselves.

Thus in Dennett's view, an author (or reader) is a novel's way of getting itself copied. Preposterous as this view is, it does allow us to see a fictional work as something like Erich Schultes' paragraph, a particular arrangement of words that tries to get itself replicated by virtue of its intrinsic interest or merit. A successful fictional meme could be considered one that is still getting itself copied after x generations have passed since it was first circulated. It can also be observed, and often is, that the long term reproductive success of a piece of fiction correlates pretty well with quality.

We could indeed go on and talk about the fitness landscapes on which fictional memes compete and adapt; how certain innovations in fiction change the way novels are read and hence the way subsequent novels are written, and how, in turn, these works fare in the ideosphere over the long term. Hans Robert Jauss (1970) has done exactly this, though instead of "hill climbing" or "adaptive landscapes" he called it the expansion of reader horizons. In sum, AL might contribute substantially, in ways that have only begun to be suggested here, to our understanding of creative processes and the development of literature and art in culture.

Is There a "Strong" Claim for Fiction?

I think it has been adequately demonstrated that fiction is some kind of AL. The question is what kind, or, specifically, what ontological status we should assign it. Within the field of AL itself there is no unanimous opinion about the status of its creations, but we can distinguish three major positions: the "weak" claim, the "strong" claim, and what I am calling the "extra strong" claim.

"Weak" AL: AL simulations are just that, simulations. They can help us "illuminate and understand more clearly the life that exists on earth and possibly elsewhere" (Levy, p. 5), but they are by no means "alive" in the sense that carbon-based biological life is alive.

"Strong" AL: Many AL programs are not just simulations of life but examples of life. Digital creations can reproduce, learn, adapt, behave autonomously, develop social structures, and evolve. To say that only carbon-based biology possess "life" amounts to a kind of parochialism, or, if you will, "carbocentrism."

"Extra strong" AL: Silicon-based life represents the next evolutionary leap and may likely surpass or even replace us.

Where fiction is concerned, we can happily exclude the "extra-strong" claim: fictional characters remain safely ensconced in their respective texts and in the participating imagination of readers, and are not about to step out and edge us off our niche in the biosphere.

But what about the "strong claim"? Are fictional characters in some real sense "alive"? If they aren't, what is the basis for the strong claim in AL? If they are, what would this imply for the way we determine what is "real"? Unless I am missing something important, this is the dilemma fiction presents to advocates of the strong claim.

Dawkins and Dennett clearly consider memes to be genuinely "alive" on some level, and as shown above, fictional memes - including, notably, animate characters with distinct personalities - exhibit most of the lifelike qualities of AL entities, arguably many more. Furthermore, the human mind is a far more complex information processor than any computer yet devised. If strong claim advocates denied that fictional characters were actually alive, they would be arbitrarily privileging the computer over the computationally superior human brain, and exhibiting the same kind of parochialism associated with the carbocentric position. In short, one of the chief didactic purposes of the strong claim is to show us how arbitrary and prejudiced have been our criteria for what constitutes "life." Fair enough, but do strong claimers mean business, or do they just happen, equally arbitrarily, to be fond of computers?

My guess is that strong claim advocates would be willing to grant fictional characters a degree of biological status. In essence, this is the whole idea behind memes; to bring immaterial thought into the realm of physical processes and to settle the mind-body problem once and for all in favor of body (banishing vitalism, mysticism, obscurantism, and what have you). One problem, though, is that physical processes are never "wrong" while thoughts often are - hence the scientific enterprise, which directs human thinking along correct, as opposed to incorrect, paths, typically by testing ideas against obdurate physical reality.

And this is the dilemma that comes with accepting the strong claim for fiction. We could grant fictional characters the status of life, but, as Ra smussen postulates about computer generated life, if fictional characters are alive, their reality is just as "real" as our own: the two realities "have the same ontological status" (1992, p. 769). While this may seem plausible (though certainly not uncontroversial) in a virtual world generated by computer software, we hesitate to apply it to the world of imagination. This is because we have generally found it (to put it mildly) useful to distinguish between phantasm and fact.

Yet once we accept imagined entities (e.g. fictional characters) as living beings, we are opening a Pandora's box of other, ever more bizarre claims to the status of life, and hence to the status of reality: dreams, hallucinations, etc. How will scientists debunk UFO "experiences," when the "truth" status of these experiences vies with that of the debunking? As we are saturated more and more in images, signals, symbols, simulations and simulacra, it seems to me more important than ever that we try to keep our heads screwed on straight about what is "real" and what is "not." This is not the best time for scientists (of all people) to start spacing out on us, yet the strong claim suggests exactly such an ontological and epistemological collapse.

At any rate, we seem to reach a point in this discussion where something has to give. To stick with the strong claim and exclude fictional characters is arbitrary and capricious. Yet to accept the strong claim for fictional characters seems, somewhere down the road, to introduce a weird, frankly animistic element into the epistemology of science. In short, in rejecting vitalism we ultimately have to accept animism. Take your pick: do you want spirit inhabiting body, or body inhabiting spirit?

An alternative, perhaps somewhat dispiriting, is to retreat to the "weak" claim and try to define "life" in a more satisfactory way. Lovers of literature, after all, manage to live quite happily entertaining no more than modestly "weak" claims about the ontological status of fictional characters. Perhaps fiction lovers could form support groups for recovering strong claim advocates! In the meantime, the "carbocentric" position still appears to stand. Perhaps the strong claim will have to wait until a self-replicating mechanism can be generated out of solid, physical "stuff."

But then, of course, we will have other things to worry about, namely the replacement of human (and possibly all carbon-based) life with a silicon-based reproducing mechanism. There is not room enough here to address the much more worrisome questions that come with the "extra-strong" claim in AL. The big question is obviously, Why are we willfully engineering our potential evolutionary extinction? The short answer is that there seems to be a logical inevitability here, a "slippery slope." The strong claim introduces a great deal of confusion about what is alive and what is real and how we could possibly know the difference. The "extra-strong" claim takes this one step further and finds there is no basis, other than computational excellence and some general evolutionary imperative, for preferring one kind of "life" over another.

Thus the ontological and epistemological breakdown signalled by the "strong" claim of AL realizes itself as an ethical breakdown in the "extra strong" claim. We find that there is no basis for a "human-centered" ethic when the only available principles are mechanism and Evolution (with a capital "E"). And if we can't locate a human-friendly ethic in the logic of AL, we had better start looking somewhere else.

Acknowledgements

Portions of this article were developed in a series for the Kinjo Gakuin University *Ronshu* and presented at the Thirteenth European Meeting on Cybernetics and Systems Research in Vienna, April 1996. I would like to thank my neighbor Irene Bensinger and my father Donald Taylor for help and encouragement.

References

Collingwood, R. G. The Principles of Art. Ed. Jan van der Dussen. Clarendon, Oxford, 1965.

Dawkins, Richard The Selfish Gene. Oxford, New York, 1978.

Dennett, Daniel C. Darwin's Dangerous Idea. Evolution and the Meanings of Life. Simon and Schuster, New York,1995.

Hofstadter, Douglas Metamagical Themas. Questing for the Essence of Mind and Pattern. Basic Books, New York, 1985.

Jauss, Hans Robert Literary History as a Challenge to Literary Theory, in New Literary History II, i: 7-37, 1970.

Langton, Christopher G. ed. Artificial Life III. Proceedings of the Workshop on Artificial Life Held June, 1992, in Santa Fe, New Mexico. Addison-Wesley, Reading, 1994.

Langton, Christopher G. Preface. In Langton ed., Artificial Life III. x-xxiv.

Levy, Stephen Artificial Life. The Quest for a New Creation. Penguin, London, 1992.

Peterson, Ivars Looking Glass Worlds, in Science News, 4 Jan. 1992: 8-15.

Rasmussen, Steen Aspects of Information, Life, Reality, and Physics in Christopher G. Langton et al. ed., Artificial Life II. Proceedings of the Workshop on Artificial Life Held February, 1990 in Santa Fe, New Mexico. Addison-Wesley, Reading, 1992. 767-773.

Schultes, Erich An Instance of a Replicator, in Langton ed., Artificial Life III 1.

URL dieses Artikels:
https://www.heise.de/-3445807

Links in diesem Artikel:
[1] https://www.heise.de/tp/subtext/telepolis_subtext_3513480.html?artikel_cid=3445807&row_id=1
[2] http://www.physics.wisc.edu/~shalizi/Dawkins/viruses-of-the-mind.html
[3] https://www.heise.de/tp/subtext/telepolis_subtext_3513482.html?artikel_cid=3445807&row_id=2
[4] http://www.hip.atr.co.jp/~ray/tierra/tierra.html

Copyright © 1996 Heise Medien