In this story we see the return of Donovan and Powell, reflecting the core cast of characters that Asimov has built up over the stories so far. These two and robopsychologist Susan Calvin tend to lead matters – they share the spotlight in this one – supported by mathematician Peter Bogert and director-emeritus of research Alfred Lanning. Other characters come and go but it’s these five we return to again and again.
There’s not much continuity for these characters. They seem pretty static and events from one story to the next don’t seem to make much difference to them. But while the characters don’t change or develop much, the world they exist in is changing in ways that this story highlight. The previous story, Little Lost Robotintroduced the concept for the hyperatomic drive to the robot story universe. In this story, we get one step closer.
Unlike the stories we’ve had so far, ‘Escape’ has two strands to it. In the first, Susan Calvin guides U.S. Robots’ most powerful positronic brain – The Brain – through a problem that has destroyed the super-computer of their main rivals, Consolidated Robots: the design and construction of a hyperatomic drive. Rather than spitting out a plan or a bunch of calculations, The Brain goes to work like a genie in a fairy tale, summoning up an army of robot goblins to create a complete ship out of more or less nothing. The only problem is, it won’t tell its masters how to work it. Thus the second strand of the story, wherein Donovan and Powell are sent in to figure it out.
The problem at the heart of Escape is seems probably the most spurious so far. While it’s no more arbitrary than Reason or Liar!, it doesn’t have the syllogistic logic that makes those stories so pleasing. Instead, the conclusion is based on assumptions about both the laws of physics and the Laws of robotics that readers aren’t really party to until the revelation.
Like most of the other robots, The Brain comes across like a chirpy teenager with ADHD. The scientists have no control over what happens in the positronic brain, although they can make deductions and predictions using mathematics. While the understand the variables robot behaviour, they don’t seem to be able to direct the way the brains develop, except through the rarefied talking cure of robopsychology. This characteristic is probably the source of HAL 9000 as well as Marvin the Paranoid Android and the irritatingly upbeat ship’s computer Eddie in The Hitchhikers Guide to the Galaxy.
This is in contrast with a current software-based approach to AI, and a similar equation of software and mind that’s been part of SF since the 1980s. This makes the stories look quaint, and throws the robots’ weird personalities into sharp relief. These stories don’t feel speculative in the hard SF way to me any more, they’re more like elementary studies in rhetoric or syllogistic logic.
This story – and, on reflection, Little Lost Robot – feel less like this kind of philosophical puzzle story and more like an exercise in setting development and speculation. In this story Asimov’s more interested in the hyperatomic drive and how it might work than the dilemma faced by The Brain. Maybe it’s a drift of attention away from robots and on to the idea of a the galactic empire. I wonder where this approach is leading?