| The Magazine of the CALIFORNIA ACADEMY OF SCIENCES |
|
|
|
Horizons Virtual Evolution Thomas Ray's heart and brain are in different places. As an ecologist with 20 years' experience in Costa Rican rainforests, he is emotionally wed to preserving the vast diversity of Earth's organic life. But as a pioneer in the computer field of artificial life (A-life), he has become intellectually obsessed with digital diversity, with re- creating something like the Cambrian explosion--a "sudden" eruption of diversity about 600 million years ago--but in digital form. Ray, a professor of ecology at the University of Delaware, uses computers to probe evolutionary processes; not by employing them to crunch data, or to make evolutionary models, but by synthesizing evolution itself. In 1990, Ray impressed both his colleagues in the bio- sciences and observers in the fledgling field of A-life with a computer program called Tierra that planted simple, single-cellular "creatures" (discreet strings of computer code, or instructions) into a computer "environment." The behavioral codes composing these original creatures, called "Ancestors," are hit by random mutations--as are the DNA codes of organic, Earthly creatures. Most mutations are harmful and the creatures perish. Occasionally, a mutation causes an improvement and the new creature thrives. Thousands of generations were compressed into hours, and Ray's Ancestors diversified as they competed for the computer's energy (processing time). Ray hadn't known what to expect when he first turned Tierra on. "It fulfilled my dreams, not my expectations," he says now. Not only did the Tierran creatures resemble living things in many ways, but the patterns with which populations rose and fell also resembled those found in the fossil record: bursts of change followed by periods of stasis, and occasional mass extinctions. New species emerged, some bigger than the Ancestors, some smaller. Parasites appeared as some creatures jettisoned their replication coding, surviving by tapping the instructions of nearby hosts. Parasites and hosts coexisted until space and energy became rare, when parasites, because of their small size, proliferated until overwhelming the hosts, whose downfall caused a scarcity of replication codes. Parasite populations plummeted. One organism evolved a resistance to parasites so effective it drove the smaller creatures to extinction. Even very basic "social" or cooperative strategies emerged. The digital ecosystem fluctuated in composition, sometimes dominated by a few types of creatures, sometimes displaying a wide diversity. (Mass extinctions occurred periodically, with no help from asteroids.) Some dismissed Tierra, calling it a glorified computer game. But others said it represented a big step in the study of evolution. (In a New York Times article McGill University professor Graham Bell called Ray's program "the first logical demonstration of the validity of the Darwinian theory of evolution.") Many initially skeptical biologists, including his colleagues at the University of Delaware, were fascinated by the program's implications. Still others said Tierra represented a major step forward in evolution itself, that in the twenty-first century the center of evolutionary action on Earth will move into the digital world. To some this is exciting, but to others it is just scary. Christopher G. Langton, a Los Alamos National Laboratory A-life researcher, says "By the middle of this century, mankind had acquired the power to extinguish life on Earth. By the middle of the next century, he will be able to create it. Of the two it is hard to say which places the larger burden on our shoulders." For the student of evolutionary theory, though, the advantages of studying digital evolution are clear: every "genetic" event can be scrutinized; the evolutionary clock can be slowed down, speeded up, examined from the inside out. Ray's hope is that fundamental rules governing order and evolution in the natural world will emerge, rules that may reveal an underlying simplicity to the great complexity of nature. "It's like playing God," Ray said in 1992. But he isn't satisfied being God of a minor universe, despite the notoriety it has brought, the tenure it won him at the University of Delaware, and the scientific insights it produced. He wants more than simple digital organisms. He wants complex, roving, startling creatures, the digital equivalents of pterodactyls, hummingbirds, pumas, and hominids. In 1993, he began a two-year leave from the university and signed on with the Evolutionary Systems Department of the ATR Human Information Processing Research Laboratories, a technology think-tank in Kyoto, Japan, supported by telecommunications-industry yen. There, Ray hopes to launch a version of Tierra that will break the Cambrian barrier. The main challenge, Ray says, is creating a sufficiently complex environment. The original Tierrans were locked in a single small computer. To fabricate enough complexity in one machine, estimates Ray, would require a 300-million-dollar computer. As an alternative, Ray is forming a worldwide network of thousands of individuals and organizations who have room in their computers to donate to the evolution of artificial life forms. (Actually it will be a "virtual" or model network within the real one and using a unique language to keep the creatures from escaping and devouring the computers they occupy.) Anyone on the Internet who wants to contribute their computer's spare time to the project will be hooked into Ray's system by a sort of electronic virtual nature corridor. Ray's creatures will migrate throughout this network searching for energy, interacting, competing, and evolving. He calls his network a "digital reserve," juxtaposing it with his other major project, an "organic reserve" in the rainforest of the Sarapiqui region of northern Costa Rica. "Because of its size, topological complexity, and dynamically changing form and conditions," says Ray, "the global network of computers is the ideal habitat for the evolution of complex digital organisms." As interesting creatures evolve, so may insight into such puzzles as the origin and maintenance of sexuality, the origins of species, maximum sustainable diversity, the genetic requirements for long-term survival, and the properties of the speciation process, Ray says. But most important, scientists may get their first view into the generation of quantum increases in complexity, the "big bang" of evolution. There may be other benefits. One reason the Japanese think-tank funds Ray's work is that some of the organisms may have marketable qualities, though there is now no telling what they'll be. In fact, says Ray, better to start with an open mind than to go into the evolution business with specific product objectives. "Evolution can create complex and beautiful creatures without the help of an intelligent supervisor," says Ray, "whereas program engineers can't even create something as complex as an insect." Despite the demands of building a parallel universe, Ray remains committed to his conservation work in Costa Rica and travels there twice a year to pursue more conventional field studies. The two reserves overlap in their shared dedication to the value of living diversity, he says. Both projects, "provide a space for organisms to evolve and behave freely with a minimum of human intervention."
Gordy Slack is Associate Editor of California Wild. |
Summer 1994
|
