I hate to say anything bad about such a wonderful movie, but my job here is to evaluate the science, which has some definite problems. The physics are illogical: Why should a ship with artificial gravity inside "tilt," spilling all the passengers off to one side when someone turns the steering wheel? There are also fire extinguishers with improbably long CO2 charges, and a host of logic errors: How does a giant spaceship land itself when the autopilot has been shut off? How does plant life reassert itself with seeds that have spent 700 years bathing in toxic ruin? And could the plants spread themselves without the help of pollinating insects? Worst of all is the beautiful but erratic motion of bodies in the weightless vacuum of space; as one of the world's premier developers of animation physics engines, Pixar certainly knows better!
Still, these errors are delivered with a wink and a nod; as with a Road Runner cartoon, one gets the sense right away that this movie is not meant to be taken literally. It's a myth, an allegory, a cave painting for future generations to puzzle over. And on its central issue--robots with personality, and their relationship to human beings--the story finds much better footing. It seems that after 700 years of tireless duty and at least a century of total isolation, our autronic amigo Wall*E has gone a bit peculiar. He still performs his directive--gathering trash and crushing it into blocks--but on the side we find him hoarding trinkets, caring for cockroaches and dreaming of robotic companionship long after the Earth's other mechanoids have died out or moved away. Around the edges of little Wall*E's programming, he's developed a unique and distinctive personality, unlike anything his designers could have intended.
This actually makes perfect sense, because Wall*E's job requires quite a bit of cognitive flexibility. First of all, Wall*E has to locate trash of the right size and consistency to fit inside his compactor. He can't do anything with a discarded engine block (too big), a coil of rope (too flexible), or a heap of broken glass (not compressible), right? Wall*E also has to stack his blocks carefully on uneven terrain, and abandon his towers before they're tall enough to fall over. As a result, each tower looks different from the others, more like the organic spires of a termite mound than the output of an industrial robot. Wall*E also has to adapt his behavior to changing seasons and weather conditions; when the dust storms blow in, he's learned through bitter experience to seek shelter immediately. Wall*E's peers who didn't learn the lesson, or didn't learn it well enough, are all dead.
Which raises another interesting point: Wall*E was originally designed to coordinate his actions with millions of other Wall*E robots. Neither unique nor alone, they worked like ants--not only staying out of each other's way but also working cooperatively and even performing maintenance and repair operations on one another--operations that would also have to adapt over time as the number of working robots and the quality of available spare parts declined. For a robot like that, every day may be the same as the thousands before it, but every moment, every individual action requires snap judgment based on hundreds of restless variables. Wall*E has to be flexible, yes, and the only sort of computer that works well under that sort of demand is a heuristic neural network, a system developed in 1943 by Warren McCulloch and Walter Pitts at the University of Chicago, modeled loosely on the function (though not the form) of animal brains.
And the thing about that is, the "programming" of a neural net really will show a lot of variability. Oh, sure, if you drop it in a rigid environment it'll behave in a predictable way, like a rat in a maze or an office worker in a cubicle. But if you bombard it with random events and variables it wasn't specifically designed to handle, its responses will be, for lack of a better word, creative. Like a rat in the jungle or a human being on a college campus, it will learn and grow, generating new behaviors and associations while adapting and even forgetting old ones that no longer produce the desired result.
Hello (and wake up), Dolly!
Which makes sense of yet another of the movie's persistent themes: that Wall*E is able to "wake up" other robots and even humans, simply by hanging around them for a little while. This is not only a romantic notion that fits perfectly into any cartoon universe, but also a very plausible side effect of the widespread use of neural networks. Nothing is more jarring, more broadening, more disruptive to in-the-box thinking than a fellow creature who (a) behaves in conscious but unpredictable ways, (b) wants your attention and (c) won't take no for an answer. Rote responses won't get you out of an encounter like that; you really do have to think your way through it. What's going on here? Who is this guy? What does he want, and why does he want it, and how can I pay attention to my own problems with him standing in the way like that? And because the act of thinking changes the weighting and activation thresholds of your neurons, and sometimes even forces new connections to grow ... well, almost by definition you won't be the same person after the encounter that you were before it.
And so it is with robots, too. Especially broken robots, who seem to respond to Wall*E with particular interest. This also makes sense, because their assorted malfunctions have already pushed them out of their mental ruts, forcing them to think about the world in new ways. They're different, and they don't necessarily want to be "fixed" and crammed back into their old roles again. In Wall*E's capering they see an echo of their own idiosyncrasies, and a way out of the trap of their fellow robots' rigid thinking. We could learn a lot from them, which of course is the very point Stanton and Docter are trying to make. Which begs the question: Will their movie jolt our species out of its own mental rut?
Sources:
www.rottentomatoes.com: "Wall*E"
The Internet Movie Database (www.imdb.com): "Wall*E"
Encyclopedia Britannica Ultimate Reference Suite DVD, 2008 Edition: "Personality", "Neural Network"
Wil McCarthy is a rocket guidance engineer, robot designer, nanotechnologist, science-fiction author and occasional aquanaut. He has contributed to three interplanetary spacecraft, five communication and weather satellites, a line of landmine-clearing robots and some other "really cool stuff" he can't tell us about. His short writings have graced the pages of Analog, Asimov's, Wired, Nature and other major publications, and his book-length works include the New York Times notable Bloom, Amazon "Best of Y2K" The Collapsium and most recently, To Crush the Moon. His acclaimed nonfiction book, Hacking Matter, is now available as a free download.
