Am höchsten bewertete positive Rezension
15 Personen fanden diese Informationen hilfreich
Still completely original- and totally misunderstood.
am 31. Mai 2000
When GEB was first published,the reviews and enthusiasm were endless. It's a brilliant introduction to recursion, said many. No, it's an introduction to, and demonstration of Godel incompleteness, said others. It's a demonstration of the commonality of art and science, said others. And there's something about ants near the end, but we're not sure why.
Readers today echo the same sentiments. They're all right, in their own way- but none of these views really get at what Hofsteader was trying to do. Yes, GEB is a tuorial on Godel, Bach, ants, recursion and a dozen other esoteric topics, and it's a heck of an intellectual entertainment, but Hofsteader didn't just write GEB to show prove what a clever book he could write. At the core, GEB is, first and formost, a theory of Artifical Intelligence; all the bits on Godel, recursion and combinations are just a tutorial to bring the reader up to speed for what's about to follow.
When GEB was first published, the dominant paradigm in AI was top-down; you built inference engines, programmed them with high-level knowledge about systems, and tried to get them to generalize from their. To a small minority- including Hofstader- this begged the really important questions: Where did the ability to make inferences come from in the first place? How was knowledge represented?
A few pioneers then- people like John Holland- were looking at bottom-up models in which one posited the simplest levels of an organization- the individual elements and the rules of interconnection and communication. They reasoned that that's what the brain was, so if you couldn't derive AI from a model that echoed the brain, you weren't really proving anything. It was from this perspective that GEB was written, and given the state of AI at the time, it's not surprising that most readers- even the most enthusiastic among them- totally missed the point.
Today, the bottom-up, or connectionist paradigm is gaining new respectability, and the work over the last few decades in complexity theory has given us more insight into the mechanisms of connectionism. Reading GEB in that context, not only is Hofstader's thesis much clearer, but the book appears that much more brilliant and prescient, given when it was first written.
If you've never read GEB, read it it now, and then read George Dyson's "Darwin Among the Machines", Waldrop's "Complexity", Resnick's "Turtles, Termites and Traffic James", and John Holland's "Hidden Order". If you've read GEB before, take a look at those same books and then go back and reread GEB. You'll see it in an entirely new light.