|Artificial Life: The Quest for a New Creation / Steven Levy|
|Reviewed by Tal Cohen||Wednesday, 26 January 2000|
True, biologists have no unanimous agreement on the definition of life (just as philosophers have no agreement on the definition of intelligence). But as Chris Langton, one of the pioneers of A-Life, once said, “any definition or list of criteria broad enough to include all known biological life will also include certain classes of computer processes, which, therefore, will have to be considered 'actually' alive.”
Steven Levy's book on artificial life provides a nice introduction to the field. It begins with a review of the field's history, from the early suggestions made by John von Neumann and John Conway's “Game of Life” all the way to current efforts.
John Conway's “Life” presents the first blooming of the field. It is a program that most programmers had written at least once, to enjoy its beauty while appreciating its simplicity. This beauty-out-of-simplicity is one of the cruxes of the whole field: bottom-up complexity. The idea is that simple mechanisms at the lower levels of design result in complex results on a higher level, often without the designer having any direct control on the higher-level results. (Just think how hard it is to plan things ahead in an initial configuration of a game of Life.)
A-life researches, as well as many biologists, believe that bottom-up complexity has a very large role in many mechanisms of natural life. For example, by using a very simple and minimalistic set of rules to guide simulated birds (“boids”) in a simulated universe, researcher Craig Reynolds was able to re-create bird flocking patterns that appear in nature. Every bird had information only about itself and the birds immediately near it, and no information about the rest of the flock; yet the resulting “emergent behavior” was that of a harmonized flock. In a similar manner, Chris Langton was able to re-create some ant behavior in his “vants” (virtual ants). Without having a “manager” ant handing out task lists every mornings, ants -- both in nature and in Langton's simulation -- are able to run a “community”.
And while many A-Life “experiments” are attempts to re-create aspects of nature, other experiments have proved to be useful tools for proving theories in biology. For example, experiments by David Ackley provide suggestive evidence (but naturally no “proof”) for the Baldwin Effect theory in evolution.
You'll never think about your Lego MindStorm creatures in the same way again.
While Levy describes many such computer experiments, readers that happen to be programmers (and I suspect they form a large portion of Levy's readership) will find that the book whets their appetite, without satisfying it: the program descriptions provide enough details to understand what the program does, but never enough details to recreate it -- which is a great miss, I believe.
In one place, though, this lack of information is a good thing. No discussion of artificial life would be complete without mentioning computer viruses, and the last chapter does just that. It begins by describing A. K. Dewdney's “Core Wars” and MARS system, where programmers wrote programs that fought each other in virtual arenas, and leads to the current reality of computer viruses, some of which, the author suspects, will never be extinct until the platform for which they are written (DOS, in most cases) is completely forgotten.
During the last chapter, Levy also discusses the suggestion of several researchers, that predict A-Life creatures will eventually inherit the earth, replacing us humans as the next step in evolution. While the discussion is academically interesting, Levy sometimes drifts into sensationalism, which is a pity.